Metal mobility in hydrothermal fluids: insights from ab ...Metal Mobility in Hydrothermal Fluids:...

__________________________________

Metal Mobility in Hydrothermal Fluids:

Insights from ab initio Molecular Dynamics Simulations

__________________________________

YUAN MEI

B.Eng. (2007), East China University of Science and Technology, Shanghai, China

This thesis is submitted for the degree of Doctor of Philosophy in

School of Earth and Environmental Sciences at

The University of Adelaide

November 2013 Adelaide, Australia

i

Table of Contents Abstract ............................................................................................v!

Declaration .....................................................................................vii!

Acknowledgements .........................................................................ix!

List of publications .........................................................................xi!

Chapter 1 ........................................................................................1!1.1! Metal mobility in hydrothermal fluids ......................................................................3!1.2! Experimental approaches ..........................................................................................6!1.3! Advances of ab initio MD simulation in Geochemistry............................................9!

1.3.1! Quantum Mechanics (QM) calculations...........................................................10!1.3.2! Classical Molecular Dynamics (MD) ...............................................................13!1.3.3! Ab initio (First principles) MD and free energy calculation ............................15!

1.4! Molecular simulation and experimental data interpretation....................................18!1.5! Research objectives .................................................................................................19!1.6! Thesis organization .................................................................................................20!1.7! References ...............................................................................................................21!

Chapter 2 ......................................................................................33 2.0 Abstract ..................................................................................................................35

2.1 Introduction .............................................................................................................36

2.1.1 Molecular understanding of metal transport in hydrothermal fluids................36

2.1.2 Experimental studies of Cu–Cl and Cu–HS/H2S complexes............................36

2.1.3 Molecular dynamics simulations of metal complexes in hydrothermal

geochemistry.....................................................................................................37

2.1.4 Aims..................................................................................................................37

2.2 Computational methods ..........................................................................................38

2.2.1 Car–Parrinello molecular dynamics simulations..............................................38

2.2.2 Ab initio thermodynamic integration and free energy calculations .................38

2.2.3 Correction of standard state and calculation of formation constants................40

2.3 Results .....................................................................................................................41

2.3.1 Ab initio molecular dynamic simulations.........................................................41

ii

2.3.2 Ab initio thermodynamic integration and free energy calculations..................43

2.3.3 Geometries of metastable Cu(I) complexes forming during distance constrained

MD....................................................................................................................46

2.4 Discussion ...............................................................................................................47

2.4.1 MD simulations vs. experiments ......................................................................47

2.4.2 Entropy and complex formation .......................................................................49

2.4.3 Thermodynamic properties...............................................................................49

2.4.4 Geological implications ....................................................................................51

2.5 Acknowledgement...................................................................................................51

2.6 References ...............................................................................................................52

Chapter 3 ......................................................................................55 3.0 Abstract ..................................................................................................................57

3.1 Introduction .............................................................................................................57

3.2 Methods...................................................................................................................58

3.2.1 Static quantum mechanical (QM) calculations.................................................58

3.2.2 Ab-initio molecular dynamics (MD) simulations.............................................58

3.2.3 Spectroscopic calculations................................................................................59

3.3 Results .....................................................................................................................59

3.3.1 Static quantum mechanical calculations...........................................................59

3.3.2 Ab-initio molecular dynamics (MD) simulation ..............................................59

3.3.3 Spectroscopic properties...................................................................................61

3.4. Discussion and conclusion ........................................................................................62

3.4.1 Static/COSMO vs. ab initio MD.......................................................................62

3.4.2 S3! as a ligand in gold transport........................................................................62

3.4.3 Geological implications ....................................................................................62

3.5 Acknowledgement...................................................................................................64

3.6 References ...............................................................................................................64

Chapter 4 ......................................................................................67!4.0! Abstract ...................................................................................................................70!4.1! Introduction .............................................................................................................71!

4.1.1! Metal transport in hydrothermal fluids and aims of this study.........................71!4.1.2! Experimental studies of Zn(II)-chloride speciation..........................................73!4.1.3! Computational chemistry studies of Zn(II)-Cl speciation ................................74!

iii

4.2! Methodology ...........................................................................................................75!4.2.1! Ab initio molecular dynamics simulation.........................................................75!4.2.2! Ab initio thermodynamic integration ...............................................................77!

4.3! Results .....................................................................................................................79!4.3.1! Ab initio molecular dynamic simulations.........................................................79!4.3.2! Ab initio thermodynamic integration and free energy calculations .................90!

4.4! Refitting solubility data for Zn(II)-Cl complexes ...................................................95!4.4.1! Data sources and methods ................................................................................95!4.4.2! Fitting procedure...............................................................................................96!4.4.3! Speciation of Zn(II)-Cl based on solubility experiments .................................99!

4.5! Discussion .............................................................................................................103!4.5.1! Speciation of Zn(II)-Cl complexes .................................................................103!4.5.2! Ab initio MD vs experiments .........................................................................106!

4.6! Acknowledgement.................................................................................................108!4.7! References .............................................................................................................108!

Chapter 5 ....................................................................................115!5.0! Abstract .................................................................................................................118!5.1! Introduction ...........................................................................................................119!

5.1.1! Controls on metal solubility in hydrothermal fluids.......................................119!5.1.2! Complexing and hydration of Cu(I) and Au(I) in hydrothermal brines and

vapors..............................................................................................................121!5.1.3! Computational chemistry studies of metal speciation ....................................122!5.1.4! Aim of study ...................................................................................................124!

5.2! Method: Ab initio molecular dynamics simulations .............................................125!5.2.1! Computational methods..................................................................................125!5.2.2! Choice of model systems for comparison with experimental studies ............129!

5.3! Results ...................................................................................................................130!5.3.1! Cu(I) and Au(I) complexing as a function of solution density and temperature ..

........................................................................................................................130!5.3.2! Hydration numbers of Na, Cl, Cu and Au ......................................................138!5.3.3! Na-O and Cl-O bond distances.......................................................................140!

5.4! Discussion .............................................................................................................141!5.4.1! The nature of neutral Cu(I) and Au(I) chloro-complexes...............................141!

iv

5.4.2! Density dependence of the charged or neutral complexes..............................142!5.4.3! Density dependence of hydration of ions and metal complexes.....................146!5.4.4! Translational entropy and the density dependence of hydrothermal reactions.....

........................................................................................................................148!5.5! Acknowledgement.................................................................................................150!5.6! References .............................................................................................................150!

Chapter 6 ....................................................................................159!6.1! d10 transition metal ions complexation..................................................................161!6.2! Metal complexation in mixed ligand solutions .....................................................162!6.3! Thermodynamic properties ...................................................................................162!6.4! Current limitation ..................................................................................................163!6.5! Perspective and future work..................................................................................165!6.6! Reference...............................................................................................................166!

Appendix A...................................................................................169!

Appendix B...................................................................................215!

Appendix C...................................................................................239!

Appendix D...................................................................................255!

Appendix E...................................................................................259!

Appendix F ...................................................................................263!

Appendix G ..................................................................................267!

Appendix H ..................................................................................271!

Appendix I ....................................................................................275!

v

Abstract Aqueous fluids are an important medium for transporting metals in the Earth’s crust, and are

responsible for the formation of many ore deposits. The nature and thermodynamic

properties of metal complexes in hydrothermal fluids plays a key role in controlling

elemental mobility and mineral solubility in natural and man-made systems. The bulk of our

knowledge on metal complexation in hydrothermal fluids originates from experimental

studies. Experimental studies at extreme conditions (i.e. high temperature and pressure) are

challenging; they can be carried on only over limited P-T-x conditions, and require an

accurate speciation model for interpretation. Molecular dynamics (MD) simulations are

coming of age for studying metals in hydrothermal processes; the simulations can support

the interpretation of experiments; explore conditions beyond the range over which

experiments are available; and provide a molecular-level understanding of hydrothermal

metal mobility.

In this thesis, ab initio (first principles) molecular dynamics (MD) simulations based on

density functional theory were conducted to predict the stochiometries and geometries of

Cu(I), Au(I) and Zn(II) complexes in solutions with different ligands (Cl–, H2O/OH–, HS–

/H2S, S3–) at temperatures and pressures ranging from ambient to hydrothermal/magmatic

conditions. The important complexes related to metal transport in fluids with different

temperatures, pressures and ligand concentrations were simulated. The simulations

accurately reproduce the identities and geometries of metal complexes derived from

experimental studies, where available. The ab initio MD also demonstrates novel complexes

which have not yet been observed by experiments (i.e. CuCl(HS)–, AuS3(HS)–, AuS3(OH)–,

AuS3(H2O)).

vi

The thermodynamic properties of metal-ligand association/dissociation reactions of Cu(I)-

Cl-HS and Zn(II)-Cl complexes were investigated by distance-constrained MD simulations

using thermodynamic integration. The predicted equilibrium constants (logK) for the ligand

substitution reactions at high temperature (i.e. >= 300 ˚C) show good agreement (within 1-2

log units) with the experimental values. Although the slow kinetics at lower temperatures

(i.e. < 200 ˚C) leads to a decrease in the accuracy of the predicted logKs, MD simulations

can still reproduce the trends of the change of metal mobility successfully. The predictions

of the stoichiometry and thermodynamic properties demonstrate the potential of MD

simulations in studying metal mobility in hydrothermal fluids.

vii

Declaration I certify that this work contains no material which has been accepted for the award of any

other degree or diploma in any university or other tertiary institution and, to the best of my

knowledge and belief, contains no material previously published or written by another

person, except where due reference has been made in the text. In addition, I certify that no

part of this work will, in the future, be used in a submission for any other degree or diploma

in any university or other tertiary institution without the prior approval of the University of

Adelaide and where applicable, any partner institution responsible for the joint-award of this

degree.

I give consent to this copy of my thesis when deposited in the University Library, being

made available for loan and photocopying, subject to the provisions of the Copyright Act

1968.

The author acknowledges that copyright of published works contained within this thesis

resides with the copyright holder(s) of those works.

I also give permission for the digital version of my thesis to be made available on the web,

via the University’s digital research repository, the Library catalogue and also through web

search engines, unless permission has been granted by the University to restrict access for a

period of time.

Signature: Date: 18-Nov-2013

ix

Acknowledgements My heart is always full of gratitude when I look back to the past three years and a half that I

spent on my PhD study. I was lucky to get into this exciting and inter-disciplinary project

which fits my academic background but also poses a challenge to my existing knowledge

and experience. I have achieved more than what I expected from a PhD, not only in the

academic research, but also in the benefits of diverse cultural exchange from different

background, as well as the charming personalities of my supervisors and colleagues.

I am greatly grateful to all my supervisors, Prof. Joël Brugger, Prof. David Sherman and Dr.

Weihua Liu, for taking me into the mysterious geochemistry world. With them I have

experienced systematic training from the laboratory to the field, from molecular dynamics

simulation to geochemical modelling, and from scientific paper writing to conference

presentation. Their board knowledge, countless curiosity and the endless passion into

science encouraged me to explore the unknown world. This thesis would not be completed

without their infinite patience and unreserved support.

I am thankful to my colleagues based in South Australian Museum, Dr. Barbara Etschmann,

Prof. Allan Pring, Dr. Fang Xia (now in CSIRO Clayton), Dr. Pascal Grundler (now in PSI

Switzerland) and Dr. Frank Reith for their helpful suggestions. I am also thankful to Dr.

Denis Testemale in ESRF/CNRS and Dr. Stacey Borg in CSIRO for sharing ideas. It was

great time to study in Adelaide, many thanks to School of Earth and Environmental

Sciences for providing administrations, and South Australian Museum for office and

laboratories. Thanks a lot to the School of Earth Sciences in University of Bristol for

organising my study with Prof. David Sherman in Bristol, UK.

x

I would like to acknowledge eResearchSA (Adelaide), iVEC (Perth) and BlueCrystal

(Bristol, UK) for the access of high performance computing facilities. I also appreciate

Australian Synchrotron and ESRF for provision of synchrotron XAS beamtime and lab

facilities. This thesis has been examined by Prof. Julian Gale and Prof. Terry Seward, whose

helpful reviews and comments are greatly appreciated.

I appreciate Australian government and the University of Adelaide for providing IPRS

Scholarship and CSIRO for the Minerals Down Under Flagship Postgraduate Scholarship to

support my PhD study. This project is also supported by Australian Research Council (ARC)

under discovery project DP0878903.

Last but not least, I thank all my families in China for their support and encouragements.

Particularly, to my dear husband Dr. Yuan Tian for accompanying me in Adelaide. Yuan’s

solid support from both life and study enable me to finish this thesis on time at better quality.

Thanks for Yuan’s constant love during my whole PhD.

xi

List of publications This PhD thesis is of publication format. Four papers constructed this PhD thesis, including

TWO PUBLISHED PAPERS

1 Yuan Mei*; David M Sherman; Weihua Liu; Joël Brugger*, Complexation of gold in

S3–-rich hydrothermal fluids: Evidence from ab-initio molecular dynamics simulations,

Chemical Geology, 2013, 347, 34-42

2 Yuan Mei; David M Sherman; Weihua Liu; Joël Brugger, Ab initio molecular

dynamics simulation and free energy exploration of copper(I) complexation by chloride

and bisulfide in hydrothermal fluids, Geochimica et Cosmochimica Acta, 2013, 102, 45-

64

ONE SUBMITTED MANUSCRIPT:

3 Yuan Mei; David M Sherman; Weihua Liu; Joël Brugger, The effect of solution

density on ion hydration and metal complexation: ab initio molecular dynamics

simulation of Cu(I) and Au(I) in chloride brines (25-1000 °C, 1-5000 bar), accepted

pending revisions in Geochimica et Cosmochimica Acta

ONE MANUSCRIPT PREPARED FOR SUBMISSION

4 Yuan Mei; David M Sherman; Weihua Liu; Joël Brugger, thermodynamic properties of

Zn-Cl complexation from ab initio MD simulation, to be submitted to a geochemistry

journal.

OTHER PUBLICATIONS RELATED TO THIS THESIS:

SUBMITTED MANUSCRIPT

A Yuan Tian; Barbara Etschmann; Yuan Mei; Pascal Groundler; Denis Testemale; Yung

Ngothai; Joël Brugger, Speciation and thermodynamic properties of Manganese (II)

xii

chloride complexes in hydrothermal fluids: in situ XAS study, accepted pending

revisions in Geochimica et Cosmochimica Acta

REFEREED JOURNAL PAPERS

B Yuan Tian; Barbara Etschmann; Weihua Liu; Yuan Mei; Denis Testemale; Brian

O'Neil; Nick Rae; David Sherman; Yung Ngothai; Stacey Borg; Joël Brugger,

Speciation of Nickel (II) chloride complexes in hydrothermal fluids: in situ XAS study,

Chemical Geology, 2012, 334, 345-363

C Weihua Liu, Stacey Borg, Barbara Etschmann, Yuan Mei, Joël Brugger, An XAS study

of speciation and thermodynamic properties of aqueous zinc bromide complexes at 25–

150 °C, Chemical Geology, 2012, 298-299, 57–69

REFEREED CONFERENCE ABSTRACTS

D Yuan Mei, David M Sherman, Weihua Liu and Joël Brugger, Speciation and

thermodynamic properties of d10 transition metals: insights from ab-initio Molecular

Dynamics simulations, 34th International Geological Congress (IGC) 2012, August

2012, Brisbane, Australia

E Yuan Mei, David M Sherman, Joël Brugger, Weihua Liu, Zn-Cl Complexation in

Magmatic-Hydrothermal Solutions: Stability Constants from Ab initio Molecular

Dynamics, Goldschmidt Conference 2012, June 2012, Montreal, Canada

F Yuan Mei, David M Sherman, Joël Brugger, Weihua Liu, Ab initio molecular

dynamics simulation of copper(I) complexation in chloride/sulfide fluids, Goldschmidt

Conference 2011, August 2011, Prague, Czech Republic

G Weihua Liu, Barbara Etschmann, Denis Testemale, Yuan Mei, Jean-Louis Hazemann,

Kirsten Rempel, Harald Müller, Joël Brugger, Which Ligand is the most Import for

Gold Transport in Hydrothermal Fluids? An in situ XAS Study in Mixed-Ligand

Solutions, Goldschmidt Conference 2013, August 2013, Florence, Italy

xiii

H Weihua Liu, Stacey Borg, Barbara Etschmann, Yuan Mei, Denis Testemale, David

Sherman and Joël Brugger, Molecular-level understanding of metal transport in

hydrothermal ore fluids: in situ experiments and ab initio molecular dynamic

simulations, 34th International Geological Congress (IGC) 2012, August 2012,

Brisbane, Australia

I David M Sherman, Yuan Mei, Metal Complexation in Hydrothermal Fluids: Insights

from Ab initio Molecular Dynamics, Goldschmidt Conference 2011, August 2011,

Prague, Czech Republic

1

_____________________________________

Chapter 1

Introduction

_____________________________________

CHAPTER 1. INTRODUCTION

3

1.1 Metal mobility in hydrothermal fluids

Most of the World’s metals resources (Fe, Mn, Au, Cu, Zn, Co, Pb, U, Mo, etc) are mined

from hydrothermal ore deposits. Metals concentrations form as a result of the flow of

aqueous fluids within the Earth Crust, and the ability of these fluids to re-distribute metals

(Seward and Barnes, 1997). Understanding the mobility of metals in hydrothermal fluids

underlies our knowledge of not only natural systems, such as the formation of ore deposits

in the Earth’s Crust and water-rock interactions in geothermal systems, but also industrial

processes such as corrosion in power plants, hydrometallurgy, and material synthesis and

catalysis.

Ore-forming fluids cover wide ranges of temperature (25 to >600 °C), pressure (0.1 to >500

MPa), and compositions (from pure water to complex composition with up to >50 mol%

salt such as NaCl, KCl, CaCl2), and it is known that metals are dissolved and transported in

hydrothermal fluids by forming aqueous complexes with various ligands (e.g. Cl–, HS–, NH3,

OH–, CH3COO–; review in Seward and Barnes, 1997). In addition, the coordination

geometry of aqueous complexes can also change as a function of temperature, pressure, and

fluid composition; this will affect the stability of the nature and stability of the dominant

complexes, which in turn affects the solubility of metals in the fluids. Therefore, in order to

understand the behavior of metals in hydrothermal fluids and provide thermodynamic data

for numerical modeling of metal mobility in various natural and man-made hydrothermal

systems, it is crucial to understand coordination and stability of aqueous metal complexes in

hydrothermal fluids at different T, P and salinity conditions.

Based on available field and laboratory observations, sulfur and chlorine species are

considered to play a prominent role in metal dissolution, transport and deposition in

hydrothermal fluids (Seward and Barnes, 1997). For instance, it is recognized that copper


4

(I)-chloride (e.g., Crerar and Barnes, 1976; Hemley et al., 1992; Seyfried and Ding 1993;

Xiao et al., 1998; Liu W et al., 2001; Hack and Mavrogenes, 2006) and copper (I)-

hydrosulfides complexes (e.g., Crerar and Barnes, 1976; Mountain and Seward, 1999; 2003)

are most likely responsible for copper transport in aqueous fluids; Au(I) hydrosulfide

complexes dominate Au transport in most hydrothermal environments (Seward, 1973;

Stefánsson and Seward, 2004; review in Williams-Jones et al., 2009). Another sulfur ligand,

the blue trisulfur ion S3–, was found in S-rich solutions at P > 0.5 GPa and T > 250 °C

(Giggenbach 1971; Pokrovski and Dubrovinsky, 2011). Quantum chemical calculations and

molecular dynamics simulations show that this ligand is likely to play an important role in

Au and Cu transport (Tossell, 2012; Mei et al., 2013b; Chapter 3 of this thesis).

The formation of metal complexes in hydrothermal fluids is affected by many factors,

including temperature, pressure and chemical properties of fluids (i.e., ligand concentration,

and redox and pH conditions). Temperature and pressure affect complexing mainly by

changing the properties of water as a solvent (Seward and Barnes, 1997; Seward and

Driesner, 2004). The structure of water (hydrogen bonding network; clustering) changes

significantly as temperature and pressure change; as a result, the macroscopic properties of

water (i.e. dipole moment, dielectric constant (DC), ion product constant (Kw)) change,

which affects the solvation, stability, stoichiometry, and coordination of metal complexes in

solution. The effect of temperature in metal complexation is illustrated by considering the

stability of cobalt-chloride complexes. Co(II) complexes show a decrease in coordination

number from six- (octahedral) to four-fold (tetrahedral) with increasing temperature (Liu W

et al., 2011a). Increasing NaCl concentration has a similar effect as increasing temperature:

the predominant Co(II)-Cl complex changes from octahedral CoCl(H2O)5+ to tetrahedral

CoCl42–. As a result of this change in coordination, cobalt solubility shows a complex

evolution with temperature for example (Figure 1.1, Liu W et al., 2011a). Also note that


5

speciation models based on extrapolations from room-temperature experiments would lead

to poor predictions, since Co(II) speciation at T > ~200˚C is dominated by tetrahedral

complexes that are not stable at room temperature.

Figure 1.1 Solubility of CoS2-CoS in 3 m NaCl solutions as a function of temperature (Liu

W et al., 2011a) (magenta for Co, green for Cl, red for O and pink for H).

Pressure also is expected to affect coordination chemistry. This is illustrated for example by

Suleimenov et al. (2004)’s UV-Vis study of aqueous Ni chloride complexes in supercritical

brines showed significant blueshift of measured spectra with change of pressure from 300 to

900 bar at constant temperature, which indicated the change from tetrahedral to octahedral

coordination of nickel species. The redox and pH conditions can also control metal

solubility because they affect the nature of the ligands. This is especially true in the case of

sulfur. HS– is the predominant S species in reduced, alkaline conditions, and the stable range

of pH and redox changes with temperature (Figs 1.2a,b). Further complexity of sulfur

geochemistry is introduced by the possible stability of “exotic” S species under high

pressure, such as the polysulfur ions S3– (Pokrovski and Dubrovinsky, 2011; Figure 1.2c).

Heating and/or decompression of hydrothermal fluids can lead to phase separation, which

also changes the complexation of metals and leads to the re-distribution of metals in the

0.001

0.01

0.1

1

10

100

Co

Con

cent

ratio

n (p

pm)

300200100Temperature (!C)

"#$%&'()'*+(*%+,-%$'.(+'#//'0('$*%1-%$

"#$%&'()'()/2'0(0/3


6

vapor and liquid phases (Williams-Jones and Heinrich, 2005; Pudack et al., 2009). In the

past decades, numerous experimental and theoretical studies have been conducted to

investigate metal chloride and bisulfide complexes under hydrothermal conditions. In the

following sections some relevant studies will be briefly reviewed, and the aims of this PhD

project will be outlined.

Figure 1.2 pH-redox diagrams of sulfur species (0.1 m SO42–) at 25 °C, 1 bar (a), 300 °C,

500 bar (b) and 350 °C, 5k bar (c) calculated using the Geochemist’s Workbench (Bethke,

2008) using the data in Suleimenov and Seward, (1997) for (a,b) and Pokrovski and

Dubrovinsky, (2011) for (c).

1.2 Experimental approaches

Many experimental studies have been conducted to investigate the stability and coordination

structure of aqueous metal complexes under hydrothermal conditions relevant to

hydrothermal ore systems (Ulmer and Barnes, 1987; McKibben and Williams, 1989;

Hemley et al., 1992; Seward and Barnes, 1997; Liu W et al., 2008). These studies used a

variety of experimental methods, including solubility, spectroscopy using Ultraviolet-

Visible-Near Infrared (UV-Vis-NIR) and X-ray radiations, and potentiometric approaches.

Solubility experiments have provided us with the bulk of the knowledge about metal

behavior under hydrothermal conditions relevant for the formation of ore deposits (e.g.,

McKibben and Williams, 1989; Wood and Samson, 1998; Fleet and Knipe, 2000; Zezin et

!"# !$# !%#


7

al., 2007; Liu W et al., 2008). These experiments measure directly the concentrations of

dissolved metals at equilibrium with a known mineral (or mineral assemblage) in solutions

of known and well-constrained chemistry. Series of measurements conducted under varying

conditions (e.g., P,T, ligand concentration, pH) can be interpreted to recover the

stoichiometries and stability constants of metal complexes. These quantitative

thermodynamic properties, gathered from simplified experimental systems, can be used to

model metal transport in complex hydrothermal fluids (Crerar et al., 1985; Ulmer and

Barnes, 1987).

Solubility measurements provide direct information on the concentrations of metals that can

be carried in a specific fluid under given P,T conditions, but they do not provide a direct

insight into the molecular-level structure of the aqueous complexes responsible for metal

mobility. On the other hand, spectroscopic methods, such as UV-Vis spectroscopy and in

particular X-ray Absorption Spectroscopy (XAS) are sensitive to aspects of the electronic

structure of the aqueous complexes, and are hence capable of providing information on the

coordination structure of the species in solution. For example, when forming a metal

complex, the d-orbitals of transition metal ions split, and these changes in d-orbital energies

are reflected in the electronic spectra (e.g. UV-Vis) (Ulmer and Barnes, 1987; Janes and

Moore, 2004). As metal complexes form with specific geometries that reflect the ligand-

field splitting, the measured UV-Vis spectra allow to study the coordination geometry of

these complexes.

Understanding of the coordination geometry of complexes is essential for the setup of the

correct speciation model to interpret experimental data, because this geometry limits the

possible stoichiometries of the complexes. A typical example is the speciation of copper (I)

in chloride brines. By 2007, the consensus arising from a large body of experimental work

was that CuCl(aq) and CuCl2– were the main Cu complexes in brines, with CuCl4

3– stable at


8

high salinity and low temperature (reviewed in Brugger et al., 2007). However, synchrotron-

based XAS studies (Fulton et al., 2001; Brugger et al., 2007) showed that CuCl43– is not

stable in solution under any conditions, a surprising result given that the CuCl43– moiety is

found in many solids that crystallize from salty solutions. Instead, low temperature

(<200 °C), high salinity solutions contain trigonal planar CuCl32–, while under all other

conditions linear complexes (CuCl2- and CuCl(H2O)(aq) predominate. These results were

cross-proofed independently by an ab initio MD study (Sherman, 2007). Based on this new

understanding of the coordination chemistry of Cu(I) chloro complexes as a function of

temperature and salinity, Brugger et al. (2007) re-interpretation earlier UV-Vis (Liu W et al.,

2002) and solubility (Xiao et al., 1998; Liu W et al., 2001) data. This enabled the

construction of a self-consistent thermodynamic model, which greatly improved our

understanding of Cu transport in hydrothermal fluids (Liu W et al. 2008).

Synchrotron-based XAS has proved to be a powerful tool to study metal species at

molecular level (review in Brugger et al., 2010). In particular, the development of

spectroscopic autoclaves for in situ XAS measurement of solutions up to supercritical

conditions has made it possible to directly determine the structure of aqueous complexes at

high PT (e.g., Seward et al., 1996; Fulton et al., 2001; Testemale et al., 2005, Liu et al.,

2007; Tian et al., 2012). In the past decade a number of in situ studies have investigated the

geometries, speciation and thermodynamic properties of aqueous transition metal complexes

at elevated temperatures and pressures, such as Zn(II)-Cl (Mayanovic et al., 1999; Liu W et

al., 2007), Fe(II)-Cl (Liu W et al., 2007; Testemale et al., 2009), Co(II)-Cl (Liu W et al.,

2011a), Mn(II)-Br (Chen et al., 2005a,b), Ni(II)-Cl (Hoffmann et al., 1999); Zn(II)-Br (Liu

W et al., 2012a; Appendix C of this thesis), Cd(II)-Cl (Bazarkina et al., 2010); Cu(I)-Cl

(Brugger et al., 2007; Liu W et al., 2008); Cu(I)-HS (Etschmann et al., 2010); Au(I)-Cl

(Pokrovski et al., 2009a); Au(I)-HS (Pokrovski et al., 2009b); Ag(I)-Cl (Pokrovski et al.,


9

2013). Synchrotron XAS based studies have shed light on the detailed molecular-level

understanding of these systems, and refine or confirm the interpretation of the available

experimental data.

Experimental studies have provided the majority of the available structural and

thermodynamic data that underpin our understanding of metal mobility under hydrothermal

conditions. However, there are still some limitations of experimental methods used for the

determination of structural information: 1) The PT condition of in situ experiments are

limited by technical difficulties, such as corrosion and mechanical properties to withstand

high PT. 2) Every experimental method has its intrinsic limitations. For example, UV-Vis

only provide indirect information about the structure of the metal complexes; and XAS is

not able to see the second coordination shell, and ligands with similar numbers of electrons

can be difficult to distinguish (e.g., Cl– and HS–). 3) It would be desirable to have

independent crosscheck and molecular models to help interpret experimental data. With the

advances of high-performance computing techniques, molecular dynamics simulation has

come to age as an alternative and complementary approach to study metal speciation at

molecular level in hydrothermal fluids under and beyond experimental conditions.

1.3 Advances of ab initio MD simulation in Geochemistry

The theories and approaches of molecular simulations are demonstrated briefly in Figure 1.3,

which shows the theories, various approaches and output information in those methods. In

the past decades, numerous studies based on Quantum Mechanics (QM) calculations and

Molecular Dynamics (MD) simulations have been conducted, which provided knowledge of

speciation, thermodynamic properties and spectroscopic properties of mineral surface,

aqueous metal complexes and melting systems (review in Sherman, 2001).


10

Figure 1.3 Theories and approaches of molecular simulation methods.

1.3.1 Quantum Mechanics (QM) calculations As shown in Figure 1.3, quantum mechanics (or first principles) calculations are based on

solving the Schrödinger equation (Eq. 1.1).

!! = "! (1.1)

where H is the Hamiltonian operator, ! is the wavefunction, E is the total energy of the

system. In QM calculations, E is the internal energy and the thermodynamic properties can

be predicted via statistical mechanics if all the values of E for all the possible states of the

system were known. However, for all but the simplest system, the Schrödinger equation

does not have an analytic solution. In quantum chemistry, the computation of the energy and

the wavefunction of an average-size molecule is a formidable task that is alleviated by the

Born–Oppenheimer (BO) approximation (Born and Oppenheimer, 1927). Density functional

theory (DFT) was developed in the physics community and for many years was neglected

by computational chemists (Thomas, 1927; Parr and Yang, 1989). In the past decade,

however, improvements in the formalism have allowed electronic structure calculations on

very large systems and it is now a standard tool for inorganic and physical chemists (Lee et

al., 1988; Parr and Yang, 1989). DFT is an exact theorem with unknown solution so the

=

Structure Physical properties Thermodynamics Kinetics Spectroscopy …

F = ma

Quantum Mechanics

Forcefield

Empirical Potential

Classical Molecular Dynamics

Ab initio/First principle Molecular Dynamics

e.g., CPMD

Interatomic Potential

Static calculation, @ 0 Kelvin, gas phase

+ Continuum Model

e.g., COSMO @ T in solution

Approximation (Hartree-Fock or DFT)

+


11

Kohn-Sham approximation (Kohn and Sham, 1965) has to be used for most practical

calculations. The Kohn–Sham equation is the Schrödinger equation of a fictitious system of

non-interacting particles (typically electrons) that generate the same density as any given

system of interacting particles (Kohn and Sham, 1965; Parr and Yang, 1989). Most

geochemical problems are of a sufficient complexity to warrant the use of density functional

theory.

Transition metal ions in group BII in the periodic table (e.g., Zn2+, Cd2+, Hg2+) were

extensively investigated by QM, as the fully occupied d-orbitals in those metal ions renders

electronic structure calculations much easier than for metals with unpaired electrons. For

example, Tossell (1991) calculated the structures, stabilities and spectroscopic properties of

Zn(II)-Cl complexes in aqueous solutions by Hartree-Fock methods. Using the same

methods, Butterworth et al. (1992) studied Cd(II)-Cl, Cd(II)-Br, Zn(II)-Cl species in

aqueous solution, and Tossell and Vaughan (1993) Zn(II) and Cd(II) bisulfide aqueous

complexes. Asthagiri et al., (2004) investigated the hydration structure of aqueous dications

of alkaline earth and Zn2+ by quantum chemical calculations between the ion and its first-

shell water molecules and used a dielectric continuum model to supply for outer-shell

contribution. These calculations gave the following optimized geometries: six-fold

structures for Zn2+, Cd2+, Hg2+ aqua complexes (e.g., Zn(H2O)62+, Cd(H2O)6

2+, Hg(H2O)62+)

and four- or six-fold structure for Zn2+, Cd2+, Hg2+ chloride complexes (e.g., Zn(H2O)2Cl2,

Cd(H2O)4Cl2, Cd(H2O)Cl3). The structural properties are in good agreement with

experimental studies by X-ray diffraction (Ohtaki et al., 1974, 1976; Tabata and Ozutsumi,

1992) and XAS (D’Angelo et al., 2002; Liu W et al., 2007; Bazarkina et al., 2010;

Migliorati et al., 2011). QM calculations also indicate a hydration number of 6 for Cl–, and

the loss of hydration water of Cl– explained why forming complex is favored by increasing

of entropy (Sherman, 2001; 2007). Cations in group BI (Cu+, Ag+, Au+) with fully occupied


12

d-orbitals share similar electronic structures as ions in group BII, but the QM calculations of

their aquo species (Feller et al., 1999) gave different structures, with two-fold Cu(H2O)2+

and Au(H2O)2+, and three-fold (Ag(H2O)3

+) complexes being stable.

To simulate the complexation for transition metals ions with open-shell configurations (e.g.,

Mn, Ni, Fe), spin-unrestricted calculation is usually needed and the electronic configuration

is defined by multiplicity (i.e. to distinguish wavefunctions that only have different

orientation of angular spin momenta, defined as 2S+1, where S is the angular spin

momentum). By comparing the structural properties calculated at certain spin configuration,

the d-orbital configuration related to coordination chemistry can be understood. For

example, the high-spin ground state of Fe3+ aqua complex was confirmed by calculated the

electronic structure of Fe(H2O)63+ (Harris et al., 1997). Electronic structure study of the

manganese minerals rhodochrosite (MnCO3) indicates a high spin (5 unpaired electrons)

Mn2+ configuration (Sherman, 2009), which has been adopted to optimize geometries of

Mn-Cl and Mn-Br complexes and finally cross-proofed by synchrotron XAS measurement

(Appendix A in this thesis). Although there are still uncertainties of DFT in assessing the

spin-state splitting for transition metals (Harvey, 2004), DFT can be an inexpensive and

useful method to predict spin-states and related geometric and energetic properties. QM was

also employed for investigating s- (e.g., Ca, Mg, Sr by Feller, 1999) and p- (e.g., Sb by

Tossell, 1994) block metals of geological interest.

Quantum mechanics calculations relate to the ideal gas phase, or use a “Polarizable

Continuum Model” to represent the solvent (e.g., COnductor-like Screening MOdel

(COSMO), Klamt and Schuurmann, 1993 in Figure 1.3). Simulation of aqueous solutions is

challenging because of the dipole and hydrogen bond of water molecules, and the properties

of water changes dramatically with changing of T and P (i.e. dielectric constant, Debye and

Hückel, 1923). In molecular dynamics simulations (MD), water molecules are used to


13

represent the solvation. Two approaches based on different level of theory - classical MD

and ab initio (or first principles) MD were applied to simulation aqueous systems at variable

temperatures and pressures.

1.3.2 Classical Molecular Dynamics (MD)

In classical MD, the atomic interactions are described using empirical inter-atomic

potentials, which normally consists of a long-range Coulombic force and a short-range

repulsive/attractive force (Sherman, 2001). For example, Equation (1.2) is widely employed

to describe the interaction Uij between ions i and j:

(1.2)

where qi and qj are the charges of ions i and j; #ij and $ij are the potential parameters; rij is

the distance between ions i and j. The potential parameters for ions i and j can be calculated

by Equation (1.3) with known parameters for a single ion:

!

" ij = " i" j and

!

" ij =" i +" j

2 (1.3)

In Equation (1.2), Lennard-Jones potential (first term of Eq. 1.2) describes short-range

interactions (Verlet, 1967) and Coulombic term (second term of Eq. 1.2) describes long-

range interactions.

For the simulations of aqueous fluids, water molecules are significant but also difficult to

describe accurately. A number of rigid water molecule models (i.e. O-H bond length and H-

O-H bond angle are fixed) have been developed for classical MD simulations (e.g., MCY by

Matsuoka et al., 1976; TIPS2 and TIP4P by Jorgensen 1982; Jorgensen et al., 1983; SPC

and SPC/E by Berendsen et al. 1987). The SPC (Simple Point Charge) and SPC/E

(Extended Simple Point Charge) (Berendsen et al., 1987) water models were used in many

!

Uij = 4" ij# ij

rij

$

% & &

'

( ) )

12

*# ij

rij

$

% & &

'

( ) )

6+

,

- -

.

/

0 0

+qiq j

rij


14

studies because of their simplicity and computational efficiency. For example, Driesner et al.

(1998) investigated the change of ion pairing, hydration and bond distances as functions of

temperature, pressure and fluid density in NaCl solutions from ambient to supercritical

conditions. Sherman and Collings (2002) used the SPC/E model for water together with

some empirical values from Smith and Dang (1994) for Lennard-Jones potentials to

simulate the nature of concentrated sodium chloride and water mixtures under various

temperatures, pressures and salinities and the MD results showed good correspondence to

the experimental solution densities and phase diagram. A few previous studies have used

classical MD to predict the physical and chemical properties of NaCl fluids, such as

densities, ion pairing and ion hydration, conductance, phase diagram, and dielectric

constants (Oelkers and Helgeson, 1993; Smith and Dang, 1994; Cui and Harris, 1995;

Driesner et al., 1998; John, 1998; Lee et al., 1998; Sherman and Collings, 2002; Chialvo and

Simonson, 2003).

The advantage of classical MD is that large scale simulations can be conducted to

investigate the chemical and physical processes over relatively long times (i.e. several

nanosecond) and for large simulation sizes (i.e. thousand to million particles; more than one

phase), which can help to explain many important questions in geochemistry such as

mineral formation, crystal growth, phase separation, surface absorption, etc. For example,

the calcite-water interface and the nucleation of calcium carbonate have been investigated

by force-field and molecular dynamics simulations (Raiteri et al., 2010; Raiteri and Gale,

2010). A new force-field model was developed, that accurately predicted the

thermodynamics of aqueous CaCO3 system, and led to the development of a nonclassical

model for the nucleation of CaCO3 nanoparticles via a liquid-like phase (Demichelis et al.,

2011).


15

Classical MD has been applied with great success to predict the properties of solutions of

alkali and alkaline earth metals, for which electrostatic interactions are predominant

(Sherman, 2001). However, to describe the interatomic interactions involving transition

metals like Cu, Ni, or Au (Sherman, 2010), more complicated models which beyond the

simple pairwise model of classical MD are required. For such systems, the interatomic

interactions need to be described by complexed force-fields models, or quantum

mechanically. The development of plane-wave pseudopotential methods for electronic

structure calculations enables us to conduct ab initio MD simulation to investigate transition

metals.

1.3.3 Ab initio (First principles) MD and free energy calculation

Ab initio MD simulations treat atomic interactions following the principles of quantum

mechanics using density functional theory (Fig. 1.3). For example, in the Car-Parrinello

method (Car and Parrinello, 1985), the electronic wavefunctions were incorporated into the

MD scheme by defining the extended Lagrangian (Eq. 1.4):

!

L = 12i

" µ d 3r ˙ # i$%

2+ 1

2I" M I

˙ R I2 + 1

2v" µ& ˙ ' &

2 (E[{#i},{Ri},{'&}] (1.4)

The first term implements the constraint that the one-electron Kohn–Sham orbitals (

!

" i) are

orthonormal with Lagrange multipliers; the second term is the classical kinetic energy of the

nuclei with mass M and coordinates R; the third is the kinetic energy of the wavefunctions

with fictitious mass µ; the fourth term is the Kohn-Sham total energy that is a function of the

electronic charge density (as function of the orbitals

!

" i), nuclear positions (Ri) and external

constraints on the system (%&). Current CPU capacity allows us to calculate systems for

simulation time up to several dozen picoseconds with hundreds of atoms (discussed in Mei

et al., 2013a;b; Chapter 2,3,4 in this thesis).


16

Over the past decade, ab initio MD has been employed to predict the aqueous speciation of

transition metals with d10 electronic structures in hydrothermal fluids. For instance, Harris et

al. (2003) determined Zn(II) speciation in chloride-rich brines (Cl– concentration up to

7.4 m) at 25 °C and 300 °C, which indicates stable octahedral structures with one or two

chloride ions complexed at low temperature and tetrahedral clusters with three or four

chloride ions. Sherman (2007) explored the speciation of Cu(I) chloro-complexes and found

that the linear CuCl2– complex predominates at hydrothermal conditions instead of the

tetrahedral CuCl43– complex proposed by the UV-Vis study (Liu W et al., 2002). This result

was confirmed independently by the XAS study of Brugger et al. (2007). Since Sherman

(2007)’s seminal study, more speciation and geometry studies of transition metals in

hydrothermal fluids by ab initio MD achieved good agreement with experiments (Table 1.1).

Table 1.1 Bond distances of metal complexes - comparing MD with XAS

Bond distance (Å)¶ Metal complexes MD XAS

CuCl2–

2.15 (Sherman, 2007) 2.13 (Mei et al., 2013a; Chapter 2)

2.152(7) (Brugger et al., 2007) 2.12–2.13 (Fulton et al., 2000a,b)

Cu(HS)2– 2.16 (Mei et al., 2013a; Chapter 2) 2.149(9) (Etschmann et al., 2010)

AuCl2– 2.27-2.32 (Chapter 5) 2.267(4) (Pokrovski et al., 2009a)

Au(HS)2– 2.36 (Liu X et al., 2011b) 2.29(1) (Pokrovski et al., 2009b)

¶ Distances of metal (Cu, Au) and ligand (Cl or S)

However, these qualitative studies have a limited practical impact on the understanding of

the mobility of transition metals in hydrothermal fluids. To predict metal transport in

complex natural and engineered environments, we need to know the thermodynamic

properties of the related metal complexes. This is a crucial challenge for MD simulations. In

principle, thermodynamic properties could be derived by observing equilibria among


17

aqueous complexes over periods of times much larger than the ligand exchange reaction

rates. Since these rates lies in the ns to "s range for the aqua- and chloride complexes of

transitions metals (Table 11.4 in Burgess 1978; Sharps et al., 1993), such calculations are

currently out of reach of ab initio MD methods. Two techniques were developed in the last

few years, namely, metadynamics (Laio and Parrinello, 2002; Alessandro and Francesco,

2008) and thermodynamic integration (Sprik, 1998; Sprik and Giovanni, 1998), to get

around this problem and to calculate the free energy surface of chemical reactions related to

forming metal complexes (Sherman, 2010). Metadynamics can be used both for

reconstructing the free energy of complexes and for accelerating rare events, eliminating the

need to conduct long simulations (Laio and Parrinello, 2002; Alessandro and Francesco,

2008). Thermodynamic integration is employed to calculate the free energy of reaction with

predefined reaction coordinates (e.g., change of bond distance, angle, coordination number,

etc; Sprik 1998; Sprik and Giovanni, 1998). In geochemistry, Van Sijl et al. (2010)

conducted metadynamics calculation to explore the free energy surface of Ti(IV) aqua

complexes at 300 K and 1000 K by constraining the coordination number of Ti-O, which

indicates a stable 5-fold Ti(IV) aqua complex at room temperature an 6-fold aqua complex

at 1000K. The study of the stability of Ag(I)-Cl complexes (Liu X et al., 2012b)

qualitatively explored the free energy surface of dissociation reactions from 3-fold AgCl32–

to 2-fold AgCl2– complex, indicating a unstable trigonal plannar AgCl3

2– at room

temperature. It could be extremely time-consuming to obtain a converged free energy

surface for calculating thermodynamics properties quantitatively. The mechanism of growth

and dissolution of barite surface was studied by metadynamics and the free energy surface

was obtained by a sequence of umbrella sampling (Stack et al., 2012). With predefined

reaction path, thermodynamic integration with coordination number constraint was

employed to investigate the hydration mechanisms of Zn2+ (Liu X et al., 2011c), Al3+ (Liu X


18

et al., 2010a) and Cu2+ (Liu X et al., 2010b). The complexation and free energies of aqua-

and chloride-complexes of U(VI) were studied by Bühl and coworkers (Bühl et al., 2006,

2008; Bühl and Golubnychiy, 2007) using distances constraint thermodynamic integration

by fixing the bond distances of uranium–chloride and uranium–oxygen. The free energy

studies mentioned above show the potential of predicting thermodynamic properties by MD

simulation. These studies show that thermodynamic integration can provide quantitative to

semi-quantitative information about the energetics of aqueous complexes over wide ranges

of P,T and solution compositions. Because the association/dissociation reactions of metal

complexes can be described as the change of metal-ligand distances, in this thesis, we used

distance constraint thermodynamic integration to calculate the free energy of Cu-Cl-HS

(Mei et al. 2013a; Chapter 2) and Zn-Cl (Chapter 4) complexes under hydrothermal

conditions.

1.4 Molecular simulation and experimental data interpretation

In geochemistry, molecular simulations are also used to interpret experimental results,

especially in processing synchrotron-based XAS data. Palmer et al. (1996) developed a

method to generate the EXAFS spectra directly from molecular dynamics trajectories and

applied it to strontium chloride solutions. Classical MD was used to support the

interpretation of first-shell structures of Ni2+-aqua- and -chlorocomplexes in supercritical

fluids from EXAFS data (Wallen et al. 1998; Hoffmann et al. 1999). Pasquarello et al. (2001)

studied the structure of the hydrated Cu(II) complex by neutron diffraction and first-

principles MD, and claimed a fivefold coordination of the first solvation shell of the Cu(II)

aqua ion. Peacock and Sherman (2005) measured the EXAFS spectra of Copper (II)

sorption onto goethite, hematite and lepidocrocite to investigate the coordination structure

of inner-sphere complexes as well as the iron oxide surface, and then interpret the EXAFS


19

data by DFT-based quantum mechanical calculation. Similar DFT calculations were also

conducted to optimized geometries of Ni(II) and Mn(II) complexes for simulating XANES

spectra and comparing with experiments (Tian et al., 2012; Appendix A, B of this thesis).

Ab initio MD was also combined with hydrothermal experiments of Ag-Cl complexes

(Pokrovski et al., 2013). With the advances in computational facilities and theoretical

algorithms, MD simulation is increasingly used to explore hydrothermal fluids at the

molecular level (Sherman, 2010). The combination of MD and synchrotron-based XAS

provides an unprecedented detailed view of metal speciation in geochemical systems. The

resulting increase in confidence of the speciation model also improves the confidence in the

quantitative thermodynamic interpretation of the experimental data.

1.5 Research objectives

The aim of this project was to use molecular simulations to obtain molecular-level

understanding of metal mobility in hydrothermal fluids, in particular, to understand the

complexation and stability of gold(I), copper(I) and zinc(II) in chloride and bisulfide rich

fluids over a wide range of temperatures and pressures. More specifically, the main

objectives of the project are:

1. Calculate metal (Cu(I), Au(I), Zn(II)) speciation in hydrothermal chloride- and

bisulfide-rich fluids, and address the significance of these ligands in transporting

these metals in hydrothermal fluids.

2. Test the hypothesis that S3– is a good ligand for Au(I), and predict the nature and

geometry of the most likely complexes to form;

3. Predict the formation constants for Cu(I) and Zn(II) complexes using

thermodynamic integration in MD simulations, and compare them to existing

experimental data;


20

4. Study the nature of second-shell complexing for Cu(I) and Au(I) chloride

complexes, and the change of their speciation and hydration as a function of fluid

density in conditions ranging from room-temperature to low density magmatic

hydrothermal vapours;

5. Obtain a better understanding of gold, copper and zinc mobility in hydrothermal

fluids.

1.6 Thesis organization

The main body of this thesis consists of four chapters (chapters 2-5). Chapter 2

demonstrates the ab initio MD study of copper(I) complexation by chloride and bisulfide in

hydrothermal fluids, and assesses the ability of thermodynamic integration to provide

thermodynamic properties for the ligand formation reactions. Chapter 3 describes the

complexation of gold in S3– rich hydrothermal fluids by MD simulation. Chapter 4

investigates the mobility of zinc in chloride-rich fluids via MD, and uses this information to

reinterpret existing solubility experiments and provide self-consistent thermodynamic

properties for Zn(II) chlorocomplexes. Chapter 5 reveals the effect of solution density on

ion association and hydration, and the role of configurational entropy in driving changes in

speciation and complex free energies as a function of P and T. The overall conclusions and

future work perspectives are discussed in Chapter 6.

Some other publications relevant to this thesis are included in the Appendixes. In Appendix

A, QM was applied to optimize the geometry of Mn(II)-Cl and Mn(II)-Br clusters, and

classical MD was conducted to reveal the effect of second shell water in XANES spectra. In

Appendixes B-C, I conducted DFT calculation to optimize the geometry of Ni(II)-Cl and

Zn(II)-Br complexation and applied it to ab initio XANES calculations. Appendixes D-I are


21

peer-reviewed conference abstracts that have been presented in international conference as

oral presentation by myself (D, E, F) and my co-supervisors (G, H, I).

1.7 References Alessandro, L. and Francesco, L. G., 2008. Metadynamics: a method to simulate rare events

and reconstruct the free energy in biophysics, chemistry and material science.

Reports on Progress in Physics 71, 126601.

Armellini, F. J. and Tester, J. W., 1993. Solubility of sodium chloride and sulfate in sub-and

supercritical water vapor from 450–550 °C and 100–250 bar. Fluid Phase Equilibria

84, 123-142.

Asthagiri, D., Pratt, L. R., Paulaitis, M. E., and Rempe, S. B., 2004. Hydration structure and

free energy of biomolecularly specific aqueous dications, including Zn2+ and first

transition row metals. Journal of the American Chemical Society 126, 1285-1289.

Bazarkina, E. F., Pokrovski, G. S., Zotov, A. V., and Hazemann, J.-L., 2010. Structure and

stability of cadmium chloride complexes in hydrothermal fluids. Chemical Geology

276, 1-17.

Berendsen, H. J. C., Grigera, J. R., and Straatsma, T. P., 1987. The missing term in effective

pair potentials. The Journal of Physical Chemistry 91, 6269-6271.

Bethke, C. M., 2008. Geochemical and biogeochemical reaction modelling. Cambridge

Univ. Press, Cambridge, UK.

Born, M. and Oppenheimer, R., 1927. Zur quantentheorie der molekeln. Annalen der Physik

389, 457-484 (in German).

Brugger, J., Etschmann, B., Liu, W., Testemale, D., Hazemann, J. L., Emerich, H., van Beek,

W., and Proux, O., 2007. An XAS study of the structure and thermodynamics of

Cu(I) chloride complexes in brines up to high temperature (400 °C, 600 bar).

Geochimica et Cosmochimica Acta 71, 4920-4941.

Bühl, M. and Golubnychiy, V., 2007. Binding of Pertechnetate to Uranyl(VI) in Aqueous

Solution. A Density Functional Theory Molecular Dynamics Study. ChemInform 38

Bühl, M., Kabrede, H., Diss, R., and Wipff, G., 2006. Effect of Hydration on Coordination

Properties of Uranyl(VI) Complexes. A First-Principles Molecular Dynamics Study.

Journal of the American Chemical Society 128, 6357-6368.


22

Bühl, M., Sieffert, N., Golubnychiy, V., and Wipff, G., 2008. Density Functional Theory

Study of Uranium(VI) Aquo Chloro Complexes in Aqueous Solution. The Journal of

Physical Chemistry A 112, 2428-2436.

Burgess, J., 1978. Metal ions in solution. Published by Ellis Horwood Ltd. Chichester and

distributed by John Wiley and Sons, Chichester and New York. 481 pp.

Butterworth, P., Hillier, I. H., Burton, N. A., Vaughan, D. J., Guest, M. F., and Tossell, J. A.,

1992. Calculations of the structures, stabilities, Raman spectra, and NMR spectra of

cadmium halide hydroxide (CdCln(OH2)a2-n, CdBrn(OH2)a

2-n), and zinc chloride

hydroxide (ZnCln(OH2)a2-n) species in aqueous solution. The Journal of Physical

Chemistry 96, 6494-6500.

Car, R. and Parrinello, M., 1985. Unified Approach for Molecular Dynamics and Density-

Functional Theory. Physical Review Letters 55, 2471-2474.

Chen, Y., Fulton, J., and Partenheimer, W., 2005a. A XANES and EXAFS Study of

Hydration and Ion Pairing in Ambient Aqueous MnBr2 Solutions. Journal of

Solution Chemistry 34, 993-1007.

Chen, Y., Fulton, J. L., and Partenheimer, W., 2005b. The Structure of the Homogeneous

Oxidation Catalyst, Mn(II)(Br-1)x, in Supercritical Water:# An X-ray Absorption

Fine-Structure Study. Journal of the American Chemical Society 127, 14085-14093.

Chialvo, A. and Simonson, J., 2003. Aqueous NaCl pair association from liquidlike to

steamlike densities along near-critical isotherms. The Journal of Chemical Physics

118, 7921.

Collins, C. R., Sherman, D. M., and Ragnarsdottir, K. V., 1999. Surface Complexation of

Hg2+ on Goethite: Mechanism from EXAFS Spectroscopy and Density Functional

Calculations. Journal of Colloid and Interface Science 219, 345-350.

Crerar, D. A. and Barnes, H., 1976. Ore solution chemistry; V, Solubilities of chalcopyrite

and chalcocite assemblages in hydrothermal solution at 200 degrees to 350 degrees

C. Economic Geology 71, 772-794.

Crerar, D. A., Susak, N., Borcsik, M., and Schwartz, S., 1978. Solubility of the buffer

assemblage pyrite+ pyrrhotite+ magnetite in NaCl solutions from 200 to 350 °C.


Crerar, D. A, Wood, S., Brantley, S., and Bocarsly, A., 1985. Chemical controls on

solubility of ore-forming minerals in hydrothermal solutions. Canadian Mineralogist

23, 333-352.


23

Cui, S. and Harris, J., 1995. Solubility of sodium chloride in supercritical water: A

molecular dynamics study. The Journal of Physical Chemistry 99, 2900-2906.

D’Angelo, P., Benfatto, M., Della Longa, S., and Pavel, N. V., 2002. Combined XANES

and EXAFS analysis of Co2+, Ni2+, and Zn2+ aqueous solutions. Physical Review B

66, 064209.

Debye, P. and Hückel, E., 1923. The theory of electrolytes. I. Lowering of freezing point

and related phenomena (In German). Physikalische Zeitschrift 24, 185–206.

Demichelis, R., Raiteri, P., Gale, J. D., Quigley, D., and Gebauer, D., 2011. Stable

prenucleation mineral clusters are liquid-like ionic polymers. Nature

communications 2, 590.

Driesner, T., Seward, T. M., and Tironi, I. G., 1998. Molecular dynamics simulation study

of ionic hydration and ion association in dilute and 1 molal aqueous sodium chloride

solutions from ambient to supercritical conditions. Geochimica et Cosmochimica

Acta 62, 3095-3107.

Etschmann, B. E., Liu, W., Testemale, D., Müller, H., Rae, N. A., Proux, O., Hazemann, J.

L., and Brugger, J., 2010. An in situ XAS study of copper(I) transport as

hydrosulfide complexes in hydrothermal solutions (25-592 °C, 180-600 bar):

Speciation and solubility in vapor and liquid phases. Geochimica et Cosmochimica

Acta 74, 4723-4739.

Feller, D. F., Glendening, E. D., and De Jong, W. A., 1999. Structures and Binding

Enthalpies of M+ (H2O)(n) Clusters, M= Cu, Ag, Au. Journal of Chemical Physics

110, 1475-1491.

Fleet, M. E. and Knipe, S. W., 2000. Solubility of Native Gold in H-O-S Fluids at 100–

400 °C and High H2S Content. Journal of Solution Chemistry 29, 1143-1157.

Fritz, J. J., 1981. Representation of the solubility of copper(I) chloride in solutions of

various aqueous chlorides. The Journal of Physical Chemistry 85, 890-894.

Fulton, J. L., Darab, J. G., and Hoffmann, M. M., 2001. X-ray absorption spectroscopy and

imaging of heterogeneous hydrothermal mixtures using a diamond microreactor cell.

Review of Scientific Instruments 72, 2117-2122.

Fulton, J. L., Hoffmann, M. M., and Darab, J. G., 2000a. An X-ray absorption fine structure

study of copper(I) chloride coordination structure in water up to 325 °C. Chemical

Physics Letters 330, 300-308.

Fulton, J. L., Hoffmann, M. M., Darab, J. G., Palmer, B. J., and Stern, E. A., 2000b.

Copper(I) and Copper(II) Coordination Structure under Hydrothermal Conditions at


24

325 °C:# An X-ray Absorption Fine Structure and Molecular Dynamics Study. The

Journal of Physical Chemistry A 104, 11651-11663.

Giggenbach, W., 1971. Blue solutions of sulfur in water at elevated temperatures. Inorganic

Chemistry 10, 1306-1308.

Hack, A. C. and Mavrogenes, J. A., 2006. A synthetic fluid inclusion study of copper

solubility in hydrothermal brines from 525 to 725 C and 0.3 to 1.7 GPa. Geochimica

et Cosmochimica Acta 70, 3970-3985.

Harris, D., Loew, G. H., and Komornicki, A., 1997. Structure and Relative Spin-State

Energetics of [Fe(H2O)6]3+: A Comparison of UHF, Møller-Plesset, Nonlocal DFT,

and Semiempircal INDO/S Calculations. The Journal of Physical Chemistry A 101,

3959-3965.

Harris, D. J., Brodholt, J. P., and Sherman, D. M., 2003. Zinc complexation in hydrothermal

chloride brines: Results from ab initio molecular dynamics calculations. The Journal

of Physical Chemistry A 107, 1050-1054.

Harvey, J., 2004. DFT Computation of Relative Spin-State Energetics of Transition Metal

Compounds, Principles and Applications of Density Functional Theory in Inorganic

Chemistry I. Springer Berlin Heidelberg.

Heinrich, C. A. and Seward, T. M., 1990. A spectrophotometric study of aqueous iron(II)

chloride complexing from 25 to 200 °C. Geochimica et Cosmochimica Acta 54,

2207-2221.

Hemley, J. J., Cygan, G. L., Fein, J. B., Robinson, G. R., and D'Angelo, W. M., 1992.

Hydrothermal ore-forming processes in the light of studies in rock-buffered systems;

I Iron-copper-zinc-lead sulfide solubility relations. Economic Geology 87, 1-22.

Hoffmann, M. M., Darab, J. G., Palmer, B. J., and Fulton, J. L., 1999. A transition in the

Ni2+ complex structure from six-to four-coordinate upon formation of ion pair

species in supercritical water: An X-ray absorption fine structure, near-infrared, and

molecular dynamics study. The Journal of Physical Chemistry A 103, 8471-8482.

Janes, R., Moore, E., and Abel, E. W., 2004. Metal-ligand bonding. Royal Society of

Chemistry, Cambridge.

John P, B., 1998. Molecular dynamics simulations of aqueous NaCl solutions at high

pressures and temperatures. Chemical Geology 151, 11-19.

Jorgensen, W. L., 1982. Revised TIPS for simulations of liquid water and aqueous solutions.

The Journal of Chemical Physics 77, 4156.


25

Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W., and Klein, M. L., 1983.

Comparison of simple potential functions for simulating liquid water. The Journal of

Chemical Physics 79, 926.

Klamt, A. and Schuurmann, G., 1993. COSMO: a new approach to dielectric screening in

solvents with explicit expressions for the screening energy and its gradient. Journal

of the Chemical Society, Perkin Transactions 2, 799-805.

Kohn, W. and Sham, L. J., 1965. Self-Consistent Equations Including Exchange and

Correlation Effects. Physical Review 140, A1133-A1138.

Laio, A. and Parrinello, M., 2002. Escaping free-energy minima. Proceedings of the

National Academy of Sciences 99, 12562-12566.

Lee, S., Cummings, P., Simonson, J., and Mesmer, R., 1998. Molecular dynamics

simulation of the limiting conductance of NaCl in supercritical water. Chemical


Lee, C., Yang, W., and Parr, R. G., 1988. Development of the Colle-Salvetti correlation-

energy formula into a functional of the electron density. Physical Review B 37, 785-

789.

Liu, W., Borg, S., Etschmann, B., Mei, Y., and Brugger, J., 2012a. An XAS study of

speciation and thermodynamic properties of aqueous zinc bromide complexes at 25–

150 °C. Chemical Geology 298–299, 57-69.

Liu, W., Borg, S. J., Testemale, D., Etschmann, B., Hazemann, J.-L., and Brugger, J., 2011a.

Speciation and thermodynamic properties for cobalt chloride complexes in

hydrothermal fluids at 35–440 °C and 600 bar: An in-situ XAS study. Geochimica


Liu, W., Brugger, J., Etschmann, B., Testemale, D., and Hazemann, J. L., 2008. The

solubility of nantokite (CuCl(s)) and Cu speciation in low-density fluids near the

critical isochore: An in-situ XAS study. Geochimica et Cosmochimica Acta 72,

4094-4106.

Liu, W., Brugger, J., McPhail, D. C., and Spiccia, L., 2002. A spectrophotometric study of

aqueous copper(I)-chloride complexes in LiCl solutions between 100 °C and 250 °C.


Liu, W., Etschmann, B., Foran, G., Shelley, M., and Brugger, J., 2007. Deriving formation

constants for aqueous metal complexes from XANES spectra: Zn2+ and Fe2+ chloride

complexes in hypersaline solutions. American Mineralogist 92, 761-770.


26

Liu, W., McPhail, D. C., and Brugger, J., 2001. An experimental study of copper(I)-chloride

and copper(I)-acetate complexing in hydrothermal solutions between 50 °C and

250 °C and vapor-saturated pressure. Geochimica et Cosmochimica Acta 65, 2937-

2948.

Liu, X., Lu, X., Jan Meijer, E., and Wang, R., 2010a. Hydration mechanisms of Cu2+: Tetra-,

penta- or hexa-coordinated? Physical Chemistry Chemical Physics 12, 10801-10804.

Liu, X., Lu, X., Meijer, E. J., Wang, R., and Zhou, H., 2010b. Acid dissociation

mechanisms of Si(OH)4 and Al(H2O)63+ in aqueous solution. Geochimica et

Cosmochimica Acta 74, 510-516.

Liu, X., Lu, X., Wang, R., and Zhou, H., 2012b. Silver speciation in chloride-containing

hydrothermal solutions from first principles molecular dynamics simulations.

Chemical Geology 294–295, 103-112.

Liu, X., Lu, X., Wang, R., Zhou, H., and Xu, S., 2011b. Speciation of gold in

hydrosulphide-rich ore-forming fluids: Insights from first-principles molecular

dynamics simulations. Geochimica et Cosmochimica Acta 75, 185-194.

Loucks, R. R. and Mavrogenes, J. A., 1999. Gold Solubility in Supercritical Hydrothermal

Brines Measured in Synthetic Fluid Inclusions. Science 284, 2159-2163.

Mayanovic, R. A., Anderson, A. J., Bassett, W. A., and Chou, I., 1999. XAFS

measurements on zinc chloride aqueous solutions from ambient to supercritical

conditions using the diamond anvil cell. Journal of Synchrotron Radiation 6, 195-

197.

McKibben, M. A. and Williams, A. E., 1989. Metal speciation and solubility in saline

hydrothermal fluids; an empirical approach based on geothermal brine data.

Economic Geology 84, 1996-2007.

Mei, Y., Sherman, D. M., Liu, W., and Brugger, J., 2013a. Ab initio molecular dynamics

simulation and free energy exploration of copper(I) complexation by chloride and

bisulfide in hydrothermal fluids. Geochimica et Cosmochimica Acta 102, 45-64.

Mei, Y., Sherman, D. M., Liu, W., and Brugger, J., 2013b. Complexation of gold in S3–-rich

hydrothermal fluids: Evidence from ab-initio molecular dynamics simulations.

Chemical Geology 347, 34-42.

Migdisov, A. A., Williams-Jones, A., and Suleimenov, O., 1999. Solubility of chlorargyrite

(AgCl) in water vapor at elevated temperatures and pressures. Geochimica et



27

Migliorati, V., Mancini, G., Chillemi, G., Zitolo, A., and D'Angelo, P., 2011. Effect of the

Zn2+ and Hg2+ Ions on the Structure of Liquid Water. The Journal of Physical

Chemistry A 115, 4798-4803.

Mountain, B. W. and Seward, T. M., 1999. The hydrosulphide/sulphide complexes of

copper(I): experimental determination of stoichiometry and stability at 22 °C and

reassessment of high temperature data. Geochimica et Cosmochimica Acta 63, 11-29.

Mountain, B. W. and Seward, T. M., 2003. Hydrosulfide/sulfide complexes of copper(I):

Experimental confirmation of the stoichiometry and stability of Cu(HS)2! to elevated

temperatures. Geochimica et Cosmochimica Acta 67, 3005-3014.

Oelkers, E. H. and Helgeson, H. C., 1993. Multiple ion association in supercritical aqueous

solutions of single electrolytes. Science 261, 888.

Ohtaki, H., Maeda, M., and Ito, S., 1974. X-ray diffraction studies of aqueous solutions of

cadmium perchlorate and sodium tetraiodocadmate. Bulletin of the Chemical Society

of Japan 47, 2217-2221.

Ohtaki, H., Yamaguchi, T., and Maeda, M., 1976. X-ray diffraction studies of the structures

of hydrated divalent transition-metal ions in aqueous solution. Bulletin of the

Chemical Society of Japan 49, 701-708.

Palmer, B. J., Pfund, D. M., and Fulton, J. L., 1996. Direct Modeling of EXAFS Spectra

from Molecular Dynamics Simulations. The Journal of Physical Chemistry 100,

13393-13398.

Pan, P. and Susak, N. J., 1989. Co(II)-chloride and -bromide complexes in aqueous

solutions up to 5 m NaX and 90 °C: Spectrophotometric study and geological

implications. Geochimica et Cosmochimica Acta 53, 327-341.

Parr, R. G. and Yang, W., 1989. Density-functional theory of atoms and molecules. Oxford

university press.

Pasquarello, A., Petri, I., Salmon, P. S., Parisel, O., Car, R., Toth, E., Powell, D. H., Fischer,

H. E., Helm, L., and Merbach, A., 2001. First Solvation Shell of the Cu(II) Aqua Ion:

Evidence for Fivefold Coordination. Science 291, 856-859.

Peacock, C. L. and Sherman, D. M., 2005. Copper(II) sorption onto goethite, hematite, and

lepidocrocite: A surface complexation model based on ab initio molecular

geometries and EXAFS spectroscopy. Geochimica et Cosmochimica Acta 69, 5141-

5142.

Pokrovski, G. S. and Dubrovinsky, L. S., 2011. The S3– ion is stable in geological fluids at

elevated temperatures and pressures. Science 331, 1052-1054.


28

Pokrovski, G. S., Roux, J., Ferlat, G., Jonchiere, R., Seitsonen, A. P., Vuilleumier, R., and

Hazemann, J.-L., 2013. Silver in geological fluids from in situ X-ray absorption

spectroscopy and first-principles molecular dynamics. Geochimica et Cosmochimica

Acta 106, 501-523.

Pokrovski, G. S., Tagirov, B. R., Schott, J., Bazarkina, E. F., Hazemann, J.-L., and Proux,

O., 2009a. An in situ X-ray absorption spectroscopy study of gold-chloride

complexing in hydrothermal fluids. Chemical Geology 259, 17-29.

Pokrovski, G. S., Tagirov, B. R., Schott, J., Hazemann, J.-L., and Proux, O., 2009b. A new

view on gold speciation in sulfur-bearing hydrothermal fluids from in situ X-ray

absorption spectroscopy and quantum-chemical modeling. Geochimica et


Pudack, C., Halter, W. E., Heinrich, C. A., and Pettke, T., 2009. Evolution of Magmatic

Vapor to Gold-Rich Epithermal Liquid: The Porphyry to Epithermal Transition at

Nevados de Famatina, Northwest Argentina. Economic Geology 104, 449-477.

Ruaya, J. and Seward, T., 1986. The stability of chlorozinc (II) complexes in hydrothermal

solutions up to 350 C. Geochimica et Cosmochimica Acta 50, 651-661.

Seward, T. M., 1976. The stability of chloride complexes of silver in hydrothermal solutions

up to 350 °C. Geochimica et Cosmochimica Acta 40, 1329-1341.

Seward, T. M., 1973. Thio complexes of gold and the transport of gold in hydrothermal ore

solutions. Geochimica et Cosmochimica Acta 37, 379-399.

Seward, T. M. and Barnes, H. L., 1997. Metal transport by hydrothermal ore fluids. In:

Barnes, H. L. (Ed.), Geochemistry of Hydrothermal Ore Deposits. Wiley, New York.

Seward, T. M., Henderson, C. M. B., Charnock, J. M., and Dobson, B. R., 1996. An X-ray

absorption (EXAFS) spectroscopic study of aquated Ag+ in hydrothermal solutions

to 350 °C. Geochimica et Cosmochimica Acta 60, 2273-2282.

Sharps, J. A., Brown Jr, G. E., and Stebbins, J. F., 1993. Kinetics and mechanism of ligand

exchange of Au (III), Zn(II), and Cd(II) chlorides in aqueous solution: An NMR

study from 28–98°C. Geochimica et Cosmochimica Acta 57, 721-731.

Sherman, D. M., 2001. Quantum Chemistry and Classical Simulations of Metal Complexes

in Aqueous Solutions. Reviews in Mineralogy and Geochemistry 42, 273-317.

Sherman, D. M., 2007. Complexation of Cu+ in Hydrothermal NaCl Brines: Ab initio

molecular dynamics and energetics. Geochimica et Cosmochimica Acta 71, 714-722.


29

Sherman, D. M. and Collings, M. D., 2002. Ion association in concentrated NaCl brines

from ambient to supercritical conditions: results from classical molecular dynamics

simulations. Geochemical Transactions 3, 102-107.

Smith, D. E. and Dang, L. X., 1994. Computer simulations of NaCl association in

polarizable water. The Journal of Chemical Physics 100, 3757-3766.

Sprik, M., 1998. Coordination numbers as reaction coordinates in constrained molecular

dynamics. Faraday Discussions 110, 437-445.

Sprik, M. C., Giovanni, 1998. Free energy from constrained molecular dynamics. Journal of

Chemical Physics 109, 7737-7744.

Stack, A. G., Raiteri, P., and Gale, J. D., 2012. Accurate Rates of the Complex Mechanisms

for Growth and Dissolution of Minerals Using a Combination of Rare-Event

Theories. Journal of the American Chemical Society 134, 11-14.

Suleimenov, O. M., 2004. Simple, compact, flow-through, high temperature high pressure

cell for UV-Vis spectrophotometry. Review of Scientific Instruments 75, 3363-3364.

Suleimenov, O. M. and Seward, T. M., 2000. Spectrophotometric measurements of metal

complex formation at high temperatures: the stability of Mn (II) chloride species.


Suleimenov, O. M. and Seward, T. M., 1997. A spectrophotometric study of hydrogen

sulphide ionisation in aqueous solutions to 350 °C. Geochimica et Cosmochimica

Acta 61, 5187-5198.

Tabata, M. and Ozutsumi, K., 1992. Equilibrium and EXAFS Studies of Mercury (II)

Porphyrin in Aqueous Solution. Bulletin of the Chemical Society of Japan 65, 1438-

1444.

Testemale, D., Argoud, R., Geaymond, O., and Hazemann, J. L., 2005. High pressure/high

temperature cell for X-ray absorption and scattering techniques. Review of Scientific

Instruments 76, 043905.

Testemale, D., Brugger, J., Liu, W., Etschmann, B., and Hazemann, J. L., 2009. In-situ X-

ray absorption study of Iron (II) speciation in brines up to supercritical conditions.


Tian, Y., Etschmann, B., Liu, W., Borg, S., Mei, Y., Testemale, D., O'Neill, B., Rae, N.,

Sherman, D. M., Ngothai, Y., Johannessen, B., Glover, C., and Brugger, J., 2012.

Speciation of nickel (II) chloride complexes in hydrothermal fluids: In situ XAS

study. Chemical Geology 334, 345-363.


30

Tossell, J. A., 1991. Calculations of the structures, stabilities, and Raman and zinc NMR

spectra of zinc aqua chloride ZnCln(OH2)a2-n species in aqueous solution. The

Journal of Physical Chemistry 95, 366-371.

Tossell, J. A., 1994. The speciation of antimony in sulfidic solutions: a theoretical study.


Tossell, J. A., 2012. Calculation of the properties of the S3! radical anion and its complexes

with Cu+ in aqueous solution. Geochimica et Cosmochimica Acta 95, 79-92.

Tossell, J. A. and Vaughan, D. J., 1993. Bisulfide complexes of zinc and cadmium in

aqueous solution: Calculation of structure, stability, vibrational, and NMR spectra,

and of speciation on sulfide mineral surfaces. Geochimica et Cosmochimica Acta 57,

1935-1945.

Thomas, L. H., 1927. The calculation of atomic fields. Mathematical Proceedings of the

Cambridge Philosophical Society 23, 542-548.

Ulmer, G. C. and Barnes, H. L., 1987. Hydrothermal experimental techniques. John Wiley

and Sons, New York.

Van Sijl, J., Allan, N. L., Davies, G. R., and van Westrenen, W., 2010. Titanium in

subduction zone fluids: First insights from ab initio molecular metadynamics

simulations. Geochimica et Cosmochimica Acta 74, 2797-2810.

Verlet, L., 1967. Computer "experiments" on classical fluids. I. Thermodynamical

properties of Lennard-Jones molecules. Physical Review 159, 98.

Wallen, S. L., Palmer, B. J., and Fulton, J. L., 1998. The ion pairing and hydration structure

of Ni2+ in supercritical water at 425 °C determined by X-ray absorption fine

structure and molecular dynamics studies. The Journal of Chemical Physics 108,

4039-4046.

Williams-Jones, A. E., Bowell, R. J., and Migdisov, A. A., 2009. Gold in Solution. Elements

5, 281-287.

Williams-Jones, A. E. and Heinrich, C. A., 2005. 100th Anniversary Special Paper: Vapor

Transport of Metals and the Formation of Magmatic-Hydrothermal Ore Deposits.


Wood, S. and Samson, I., 1998. Solubility of ore minerals and complexation of ore metals in

hydrothermal solutions. Reviews in Economic Geology 10, 33-80.

Xiao, Z., Gammons, C. H., and Williams-Jones, A. E., 1998. Experimental study of copper(I)

chloride complexing in hydrothermal solutions at 40 to 300 °C and saturated water

vapor pressure. Geochimica et Cosmochimica Acta 62, 2949-2964.


31

Zeng, Y., Wang, F., and Ai, R., 1997. Solubility of gold in the SiO2-NaCl-H2O system at

300 °C and 50 MPa. Science in China Series D: Earth Sciences 40, 485-490.

Zezin, D. Y., Migdisov, A. A., and Williams-Jones, A. E., 2007. The solubility of gold in

hydrogen sulfide gas: An experimental study. Geochimica et Cosmochimica Acta 71,

3070-3081.


H2O–H2S vapour at elevated temperature and pressure. Geochimica et



32

33

_____________________________________

Chapter 2 Ab initio Molecular Dynamics Simulation and Free Energy

Exploration of Copper (I) Complexation by Chloride and

Bisulfide in Hydrothermal Fluids

Yuan Mei1, 2, 3, David M Sherman2, Weihua Liu3 and Joël Brugger1,4,*

1 School of Earth and Environmental Sciences, The University of Adelaide, Adelaide, SA 5005,

Australia

2 Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK

3 CSIRO Earth Science and Resource Engineering, Clayton, VIC 3168, Australia

4 South Australian Museum, North Terrace, SA 5000, Australia

Geochimica et Cosmochimica Acta, 2013, 102, 45-64.

(Copyright of this paper belongs to Elsevier Ltd.)

_____________________________________

34

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' *4' /&/"/('%(2$5+2#-'67&#%/5.'./%+2#"/(&'#&6')-$$'$&$-87'$902(-#"/(&'()'5(00$-:;<'5(%02$9#"/(&'

47'5,2(-/6$'#&6'4/.+2)/6$'/&',76-(",$-%#2')2+/6.'

3+42/5#"/(&'!"#"+.''

'

3+42/5#"/(&'=$"#/2.'>$/?'@A?'!,$-%#&?'=A'>A?'B/+?'CA?'#&6'D-+88$-?'EA?'FGHIA'*4'/&/"/('%(2$5+2#-'67&#%/5.'

./%+2#"/(&'#&6')-$$'$&$-87'$902(-#"/(&'()'5(00$-:;<'5(%02$9#"/(&'47'5,2(-/6$'#&6'4/.+2)/6$'/&'

,76-(",$-%#2')2+/6.A'J$(5,/%/5#'$"'K(.%(5,/%/5#'*5"#'HGF?'LMNOLA'

'

!"#$%&'(%)#&*+"#*%),'D7' ./8&/&8' ",$' !"#"$%$&"' ()' *+",(-.,/0?' $#5,' #+",(-' 5$-"/)/$.' ",#"' ",$/-' ."#"$6' 5(&"-/4+"/(&' "(' ",$' 0+42/5#"/(&' /.' #55+-#"$' #&6' ",#"'0$-%/../(&'/.'8-#&"$6')(-'",$'0+42/5#"/(&'"('4$'/&52+6$6'/&'",$'5#&6/6#"$P.'",$./.A'

Q#%$'()'K(N*+",(-' @+#&'>$/'

K(&"-/4+"/(&'"('",$'3#0$-' =$R$2(0$6' 5(%0+"#"/(&#2' %$",(6.?' 5(&6+5"$6' #22' %(2$5+2#-' ./%+2#"/(&.A' D+/2"' 8$(5,$%/5#2'%(6$2'#&6'S-("$'%#&+.5-/0"A'

!/8&#"+-$'

'

=#"$' FLNE+2NHI'

'

Q#%$'()'K(N*+",(-' =#R/6'>'!,$-%#&'

K(&"-/4+"/(&'"('",$'3#0$-' *../."$6'S/",'6$R$2(0/&8'5(%0+"#"/(&#2'%$",(6.'#&6'%#&+.5-/0"'$R#2+#"/(&A'

'

!/8&#"+-$'

'

=#"$' FLNE+2NHI'

'

Q#%$'()'K(N*+",(-' C$/,+#'B/+'

K(&"-/4+"/(&'"('",$'3#0$-' *../."$6'S/",'4+/26/&8'8$(5,$%/5#2'%(6$2'#&6'S-("$'0#-"'()'",$'%#&+.5-/0"A'

'

!/8&#"+-$'

'

=#"$' FLNE+2NHI'

'

Q#%$'()'K(N*+",(-' E(T2'D-+88$-'

'K(&"-/4+"/(&'"('",$'3#0$-' *../."$6'S/",'4+/26/&8'8$(5,$%/5#2'%(6$2'#&6'%#&+.5-/0"'$R#2+#"/(&A'*5"/&8'#.'5(--$.0(&6/&8'#+",(-A'

!/8&#"+-$'

'

=#"$' FLNE+2NHI'

'

CHAPTER 2. Cu-Cl-HS COMPLEXATION

35

A Mei, Y., Sherman, D.M., Liu, W. & Brugger, J. (2012) Ab initio molecular dynamics simulation and free energy exploration of copper(I) complexation by chloride and bisulfide in hydrothermal fluids. Geochimica et Cosmochimica Acta, v. 102, pp. 45-64

NOTE:

This publication is included on pages 35-54 in the print copy of the thesis held in the University of Adelaide Library.

It is also available online to authorised users at:

http://dx.doi.org/10.1016/j.gca.2012.10.027

55

_____________________________________

Chapter 3

Complexation of Gold in S3--Rich Hydrothermal Fluids:

Evidence from ab-initio Molecular Dynamics Simulations

Yuan Mei1, 2,*, David M. Sherman3, Weihua Liu2 and Joël Brugger1,4,*

1Tectonics, Resources and Exploration (TRaX), School of Earth and Environmental Sciences, The

University of Adelaide, Adelaide, SA 5005, Australia

2CSIRO Earth Science and Resource Engineering, Clayton, VIC 3168, Australia

3School of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK

4South Australian Museum, North Terrace, SA 5000, Australia

Chemical Geology, 2013, 347, 34-42.

(Copyright of this paper belongs to Elsevier Ltd.)

_____________________________________

56

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4(%02$5#"/(&' ()' 6(27' /&' !89:-/;,' ,<7-(",$-%#2' )2+/7.=' >?/7$&;$' )-(%' #@:/&/"/(' %(2$;+2#-'

7<&#%/;.'./%+2#"/(&.'

3+@2/;#"/(&'!"#"+.''

'

3+@2/;#"/(&'A$"#/2.' B$/C' DEC' !,$-%#&C' AE' BEC' F/+C' GEC' #&7' H-+66$-C' IEC' JKL8E' 4(%02$5#"/(&' ()' 6(27' /&' !89:-/;,',<7-(",$-%#2')2+/7.='>?/7$&;$')-(%'#@:/&/"/('%(2$;+2#-'7<&#%/;.'./%+2#"/(&.E'4,$%/;#2'M$(2(6<'8NOC'8N:NJE''

!"#$%&'(%)#&*+"#*%),'H<' ./6&/&6' ",$' !"#"$%$&"' ()' *+",(-.,/0C' $#;,' #+",(-' ;$-"/)/$.' ",#"' ",$/-' ."#"$7' ;(&"-/@+"/(&' "(' ",$' 0+@2/;#"/(&' /.' #;;+-#"$' #&7' ",#"'0$-%/../(&'/.'6-#&"$7')(-'",$'0+@2/;#"/(&'"('@$'/&;2+7$7'/&'",$';#&7/7#"$P.'",$./.E'

Q#%$'()'4(:*+",(-' D+#&'B$/'

4(&"-/@+"/(&'"('",$'3#0$-' A$?$2(0$7' ;(%0+"#"/(&#2' %$",(7.C' ;(&7+;"$7' #22' %(2$;+2#-' ./%+2#"/(&.E' H+/2"' 6$(;,$%/;#2'%(7$2'#&7'R-("$'%#&+.;-/0"E'*;"/&6'#.';(--$.0(&7/&6'#+",(-E'

!/6&#"+-$'

'

A#"$' JN:I+2:L8'

'

Q#%$'()'4(:*+",(-' A#?/7'B'!,$-%#&'

4(&"-/@+"/(&'"('",$'3#0$-' *../."$7'R/",'7$?$2(0/&6';(%0+"#"/(&#2'%$",(7.'#&7'%#&+.;-/0"'$?#2+#"/(&E'

'

!/6&#"+-$'

'

A#"$' JN:I+2:L8'

'

Q#%$'()'4(:*+",(-' G$/,+#'F/+'

4(&"-/@+"/(&'"('",$'3#0$-' *../."$7'R/",'@+/27/&6'6$(;,$%/;#2'%(7$2'#&7'%#&+.;-/0"'$?#2+#"/(&E'

'

!/6&#"+-$'

'

A#"$' JN:I+2:L8'

'

Q#%$'()'4(:*+",(-' I(S2'H-+66$-'

'4(&"-/@+"/(&'"('",$'3#0$-' *../."$7'R/",'@+/27/&6'6$(;,$%/;#2'%(7$2'#&7'%#&+.;-/0"'$?#2+#"/(&E'*;"/&6'#.';(--$.0(&7/&6'#+",(-E'

!/6&#"+-$'

'

A#"$' JN:I+2:L8'

'

CHAPTER 3. Au-S3 COMPLEXATION

57

,

NOTE:



http://dx.doi.org/10.1016/j.chemgeo.2013.03.019

A Mei, Y., Sherman, D.M., Liu, W. & Brugger, J. (2013) Complexation of gold in S3-rich hydrothermal fluids: evidence from ab-initio molecular dynamics simulations. Chemical Geology, v. 347, pp. 34-42

CHAPTER 3. Au-S3 COMPLEXATION

66

67

_____________________________________

Chapter 4

Zinc mobility in chloride-rich hydrothermal fluids: insights

from ab initio molecular dynamics simulation

Yuan Mei1, 2,*, Weihua Liu2, David M. Sherman3 and Joël Brugger1,4,*






_____________________________________

68

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4/&5' %(6/2/"7' /&' 5,2(-/8$9-/5,' ,78-(",$-%#2' )2+/8.:' /&./;,".' )-(%' #6' /&/"/(' %(2$5+2#-' 87&#%/5.'

./%+2#"/(&'

3+62/5#"/(&'!"#"+.''

'

3+62/5#"/(&'<$"#/2.' =$/>'?@>' A/+>'B@>'!,$-%#&>'<@'=@>' #&8'C-+;;$->' D@>' 4/&5'%(6/2/"7' /&' 5,2(-/8$9-/5,',78-(",$-%#2')2+/8.:' /&./;,".' )-(%' #6' /&/"/(' %(2$5+2#-' 87&#%/5.' ./%+2#"/(&>' 3-$0#-/&;' )(-' .+6%/../(&' "(' #';$(5,$%/."-7'E(+-&#2@''

'

!"#$%&'(%)#&*+"#*%),'C7' ./;&/&;' ",$' !"#"$%$&"' ()' *+",(-.,/0>' $#5,' #+",(-' 5$-"/)/$.' ",#"' ",$/-' ."#"$8' 5(&"-/6+"/(&' "(' ",$' 0+62/5#"/(&' /.' #55+-#"$' #&8' ",#"'0$-%/../(&'/.';-#&"$8')(-'",$'0+62/5#"/(&'"('6$'/&52+8$8'/&'",$'5#&8/8#"$F.'",$./.@'

G#%$'()'H(9*+",(-' ?+#&'=$/'

H(&"-/6+"/(&'"('",$'3#0$-' <$I$2(0$8' 5(%0+"#"/(&#2' %$",(8.>' 5(&8+5"$8' #22' %(2$5+2#-' ./%+2#"/(&.@' C+/2"' ;$(5,$%/5#2'%(8$2'#&8'J-("$'%#&+.5-/0"@'

!/;&#"+-$'

'

<#"$' KL9D+29MN'

'

G#%$'()'H(9*+",(-' B$/,+#'A/+'

H(&"-/6+"/(&'"('",$'3#0$-' O$9)/""$8' $P0$-/%$&"#2' 8#"#>' #../."$8' J/",' 6+/28/&;' ;$(5,$%/5#2' %(8$2' #&8' J-("$' 0#-"' ()' ",$'%#&+.5-/0"@''

!/;&#"+-$'

'

<#"$' KL9D+29MN'

'

G#%$'()'H(9*+",(-' <#I/8'='!,$-%#&'

H(&"-/6+"/(&'"('",$'3#0$-' *../."$8'J/",'8$I$2(0/&;'5(%0+"#"/(&#2'%$",(8.'#&8'%#&+.5-/0"'$I#2+#"/(&@'

'

!/;&#"+-$'

'

<#"$' KL9D+29MN'

'

G#%$'()'H(9*+",(-' D(Q2'C-+;;$-'

'H(&"-/6+"/(&'"('",$'3#0$-' O$9)/""$8' $P0$-/%$&"#2' 8#"#>' #../."$8' J/",' 6+/28/&;' ;$(5,$%/5#2' %(8$2' #&8' %#&+.5-/0"'$I#2+#"/(&@'

!/;&#"+-$'

'

<#"$' KL9D+29MN'

'

CHAPTER 4. Zn(II)-Cl COMPLEXATION

69

Zinc mobility in chloride-rich hydrothermal fluids:

insights from ab initio molecular dynamics simulation

Yuan Mei1,2, Weihua Liu2*, David M Sherman3, Joël Brugger1,4*

1 Tectonics, Resources and Exploration (TRaX), School of Earth and Environmental

Sciences, The University of Adelaide, Adelaide, SA 5005, Australia

2 CSIRO Earth Science and Resource Engineering, Clayton, VIC 3168, Australia

3 Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK

4 South Australian Museum, North Terrace, SA 5000, Australia

*Corresponding authors: [email protected], [email protected]


70

4.0 Abstract

The speciation and thermodynamic properties of metal complexes in hydrothermal

solutions plays a key role in controlling element mobility and mineral solubility in natural

and engineered systems. Ab initio molecular dynamics simulations were conducted to

investigate the stoichiometry, speciation and thermodynamic properties of Zn(II) chloride

complexes from ambient to hydrothermal-magmatic conditions up to 600 °C, 2 kbar. MD

simulations gave an octahedral geometry for the Zn(II)-aqua complexes (Zn(H2O)62+) at

25 °C, 1 bar. In contrast, tetrahedral complexes (ZnCln(H2O)4-n2-n; n=1-4) formed when

chloride bonded to Zn(II). Dehydration of the [ZnCl3–] complex was also observed with

increasing temperature, resulting in the trigonal planar ZnCl3– complexes becoming more

important at 600 °C. As Zn(II) is a small ion with a high charge, the ligand exchange rate

(!s) of Zn(II)-Cl is far beyond the time scale we can simulate in MD (ps); metastable

species may persist for lifetimes beyond the simulation time. Therefore, thermodynamic

integration is necessary to evaluate the relative stabilities of different Zn(II)-Cl complexes.

However, thermodynamic integration based on constraint (Zn-Cl bond distances) MD

simulations enable us to calculate energetic properties and the free energy of binding

between Zn(II) and Cl–.

The properties for the formation of Zn(II)-Cl complexes predicted at 200, 350 and

600 °C via ab initio thermodynamic integration show the same trends compared to those

obtained via the refitting of existing experimental data. The predicted stability constants

(log K) for ZnCl2 and [ZnCl3–] at 350 °C are in excellent agreement with refitted data

(within one log unit), which shows the potential of MD simulation in studying metal

complexation in hydrothermal conditions. The experimental data and the ab initio

thermodynamic properties both confirm the stability of the high order ZnCl42- complex even

at T > 400 °C, and the increase stability of [ZnCl3-] at T 600 °C.


71

4.1 Introduction

4.1.1 Metal transport in hydrothermal fluids and aims of this study

Most of the world’s resources of transition metals such as Zn, Pb, Cu, Au, Ag, Co or Fe

come from hydrothermal ore deposits. The speciation of metal complexes in hydrothermal

brines plays a key role in controlling metal mobility and mineral solubility in natural and

man-made hydrothermal environments. Chlorine is one of the main elements involved in the

complexing of metals in hydrothermal systems; therefore, understanding metal speciation in

Cl-bearing hydrothermal fluids is required for modeling metal transport in natural

hydrothermal fluids (Seward and Barnes, 1997; Brugger et al. 2010).

Experimental measurements of metal speciation in hydrothermal fluids is a major challenge,

however, and semi-empirical equations of state (e.g., the HKF[1] model, Tanger and

Helgeson, 1988) are extensively used to estimate thermodynamic properties of aqueous

complexes over wide P,T ranges. These methods allow to extrapolate experimental data

collected over a limited P,T range.

In recent years, computational chemistry approaches such as molecular dynamics (MD)

simulation are playing more significant roles in studying metal complexation in

hydrothermal fluids. MD simulations help to understand speciation and support

interpretation of experimental data qualitatively (Sherman, 2010). Taking advantage of the

development of high performance computing facilities, ab initio (or first principles) MD

simulations based on density functional theory (DFT) are coming of age in studying metal

speciation and thermodynamic properties of hydrothermal fluids. In the last few years, ab

initio MD has been employed to predict the speciation and thermodynamic properties of

transition metals complexes with different ligands (e.g., Zn(II)-Cl (Harris et al., 2003),


72

Cu(I)-Cl (Sherman, 2007), Au(I)-HS (Liu X et al., 2011), Ag(I)-Cl (Liu X et al., 2012;

Pokrovski et al., 2013), Cu-HS-Cl!(Mei et al., 2013a), Au(I)-HS/OH/S3 (Mei et al., 2013b),

which demonstrates the potential of ab initio MD in exploring metal behavior in

hydrothermal fluids.

MD simulations can provide quantitative thermodynamic properties through methods such

as metadynamics (Laio and Parrinello, 2002) or thermodynamic integration (Sprik 1998;

Sprik and Giovanni, 1998). Recently, Mei et al. (2013a) demonstrated that thermodynamic

integration can be used to provide accurate predictions of Cu(I) transport in chloride-

hydrosulfide solutions.

In this study, we use ab initio molecular dynamics simulations to improve our

understanding of Zn(II) transport in chloride brines up to magmatic hydrothermal conditions

(to 600 °C, 2 kbar). Zn(II) was chosen because a number of experimental studies show

significant discrepancies in the predicted speciation at elevated temperature. In addition,

MD simulations are particularly well suited for studying Zn(II) complexes as the lower

computing costs were requried because of its well-defined electronic, which makes it

amendable to computation intensive techniques such as thermodynamic integration.

Specifically, this study aims to 1) understand the nature and stoichiometry of Zn(II)-Cl

complexes, and in particular explore the stability of the high order ZnCl42– complex at high

temperature; 2) calculate the Gibbs free energies of reactions for the formations of different

Zn(II)-Cl complexes; 3) provide new values of thermodynamic properties for the simulation

of Zn(II) transport that are consistent with existing experimental data as well as the results

from the MD simulations.


73

4.1.2 Experimental studies of Zn(II)-chloride speciation

Several experimental studies have provided thermodynamic properties for Zn(II)-Cl

complexes at high temperature since the 1980s (e.g., Ruaya and Seward, 1986; Bourcier and

Barnes, 1987; Wesolowski et al., 1998). Ruaya and Seward (1986) measured the solubility

of AgCl(s) in aqueous Zn(II)-HCl solutions from 100-300 °C at water vapor-saturated

pressure, and provided formation constants for Zn-Cl species based on the decreased AgCl

solubility due to complexing between Zn and chloride when HCl (0.3 - 3.5 m) was added to

0.1 m Zn solutions. Therefore their data are dependent on the reliability of the properties of

Ag–Cl complexes, which have been improved in the past decades (Zotov et al., 1995;

Pokrovski et al., 2013). Bourcier and Barnes (1987) measured solubility of zincite (ZnO;

200-350 °C) and smithsonite (ZnCO3; 100 and 150 °C) in 0-5 m NaCl solutions where pH

was buffered by fixing pCO2(g). Wesolowski et al. (1998) also measured the solubility of

zincite in acidic chloride solutions and provided the formation constants for ZnCl+ and

ZnCl2(aq) at 200 °C. These three studies have similar values for the formation of ZnCl2(aq),

however their log K values differ by more than one log unit for the formation of ZnCl+,

ZnCl3– and ZnCl4

2–. Figure 4.1 shows the changes of Zn(II)-Cl speciation as a function of

temperature and Cl– concentration calculated from different sources. Ruaya and Seward

(1986)’s study (Fig. 4.1a) and the SUPCRT database (Fig. 4.1c, Johnson et al., 1992)

indicate predominant ZnCl42– over a wide range of temperature in Cl– rich fluids. In contrast,

Bourcier and Barnes (1987) (Fig. 4.1b) suggested that ZnCl42– is a minor species above

100 °C, with ZnCl3– being the predominant species in high-temperature brines.

This dichotomy is also reflected by spectroscopic studies. Several in-situ Raman (Buback,

1983) and XAS (Anderson et al., 1998; Mayanovic et al., 1999; Bassett et al., 2000) studies

suggest that ZnCl42– predominates in highly saline fluids at high temperature. For example,

the diamond anvil cell EXAFS study of Mayanovic et al. (1999) concluded that ZnCl42– is


74

the predominant complex in a 1 m ZnC12/6m NaCl solution at 800 MPa from 25 to 660 °C,

while ZnC12(H2O)2(aq) was predominant in a 2 m ZnC12 solution over the same T range.

Similarly, XAS studies of natural hypersaline (36 wt% NaCl equiv) fluid inclusions

indicated that the dominant complex was tetrahedral ZnCl42– at temperatures up to 430 °C

(Anderson et al., 1998). Liu W et al. (2012a) also confirmed the stability of the ZnBr42–

complex up to 150 °C. In contrast, a recent XAS study of Cd-chloride complexation (20-

450 °C, 600 bar) by Bazarkina et al. (2010) found no evidence for the stability of CdCl42– at

high temperature. Consequently, they doubt that ZnCl42– exists at elevated temperature,

pointing to the difficulty of distinguishing between the O and Cl ligands via XAS for first-

row transition metals.

Figure 4.1 Comparison of Zn(II)-Cl speciation in chloride solutions as function of

temperature and Cl– concentration from (a) Ruaya and Seward (1986), (b) Bourcier and

Barnes (1987) and (c) SUPCRT database (Johnson et al., 1992).

4.1.3 Computational chemistry studies of Zn(II)-Cl speciation

A number of quantum mechanical studies have investigated the nature, geometry, and

qualitative relative stabilities of ZnCln(H2O)a2-n complexes in aqueous solution. Based on

static calculations, Tossell (1991) showed ZnCl42– to be tetrahedral with bond distances of

2.31 Å, compared with the value of 2.30 Å by X-Ray diffraction study of aqueous Zn-Cl

solution (Kruh and Standley, 1962). The classical MD study of Zn-Cl (Harris et al., 2001)

!" !#$% !# !$% & $% #

%&

#&&

#%&

"&&

"%&

'&&

()*+,+-(.

/011

/0-(1/0-("2,34

/0-('.

/0-(5..

!" !#$% !# !$% & $% #

%&

#&&

#%&

"&&

"%&

'&&

()*+,+-(.

/011

/0-(1

/0-("2,34

/0-(5..

SUPCRTBourcierRuaya

!" !#$% !# !$% & $% #&

%&

#&&

#%&

"&&

"%&

'&&

()*+,+-(.

6+27-4

/011

/0-(1

/0-("2,34

/0-('.

/0-(5..

(a) (b) (c)


75

suggested the existence of both ZnCl3(H2O)– (predominant) and ZnCl42– at 600 °C and

2k bar. Harris et al. (2003) investigated the speciation and geometries of Zn(II)-Cl

complexes at 25 and 300 °C using ab initio MD simulation, and demonstrated stable

octahedral [ZnCln(H2O)6-n]2-n (n=1,2) complexes at low temperature, and tetrahedral

[ZnCln(H2O)4-n]2-n (n=3,4) complexes in concentrated Cl– brines (up to 7.4 m Cl–) over the

temperature range of 25-300 °C. A recent MD study has revised the calculations, and

proposed that ZnCl42– exists in 3 m NaCl at 625 °C and 15 kbar (Sherman, 2010). However,

due to the limitation of computing capacity, the simulations were only conducted for around

one picosecond, too short to reach the equilibrium state, as the ligand exchange rate for

Zn(II)-Cl are in the magnitude of ~300 µs at 25 °C (Sharps et al., 1993). The exchange rate

for Zn(II)-H2O is faster than Zn(II)-Cl (~30 ns, Sharps et al., 1993), but the magnitude is

still outside the computational capabilities for ab initio MD. To overcome this limitation,

Liu X et al. (2011) used thermodynamic integration with a coordination number constraint

to investigate the hydration mechanism of Zn2+. The free energy profile of the aqua-zinc(II)

complexes indicates the change of coordination number from 6 at room temperature to 4-5

at 620-1000 K. In natural fluids, Zn(II) concentrations correlate with chlorinities (Yardley

2005), and chloride complexes are expected to be the main form of Zn(II) in many ore-

forming fluids. A central aim of this study is to use thermodynamic integration to estimate

the properties of Zn(II)-Cl complexes over a wide range of P,T conditions.

4.2 Methodology

4.2.1 Ab initio molecular dynamics simulation

In this study, ab initio MD simulations were performed by using the Car-Parrinello (CP)

molecular dynamics code CPMD (Car and Parrinello, 1985). Car-Parrinello molecular


76

dynamics simulations implement density functional theory using a plane-wave basis set and

pseudo-potentials for the core electrons plus the nucleus. The PBE exchange-correlation

functional (Perdew et al., 1996) was employed with a cutoff of gradient correction 5$10-5.

Lin et al. (2012) showed that the energy profiles for liquid water calculated by PBE agree

very well with higher-level ab initio calculations (MP2, CCSD). Plane-wave cutoffs of

25 Ry (340.14 eV) were used together with Vanderbilt ultrasoft pseudo-potentials in CPMD

package generated using the valence electron configuration 3d104s2 for Zn (Laasonen et al.,

1993). Molecular dynamics simulations were conducted in the NVT ensemble (Sherman,

2007). A time-step of 3 a.u. (0.073 fs) was used to stabilize the simulations. Temperatures

were controlled by the Nosé thermostat for both the ions and electrons. The target fictitious

kinetic energies (keyword EKINC in CPMD code) were obtained by taking the converged

value of a 10,000-steps’ simulation with no defined Nosé thermostat for electrons. Fictitious

electron mass of 400 a.u. (3.644$10–28 kg) was used to obtain convergence of the energy of

the total CP-Hamiltonians.

The initial atomic configurations of each simulation were generated by classical MD using

the SPC/E potential for water (Berendsen et al., 1987; Smith and Dang, 1994) and

approximate pair potentials derived from finite cluster calculations for Zn-O and Zn-Cl

(Harris et al., 2001). As listed in Table 4.1, MD simulations No. 1-4 were conducted with

1 Zn2+, 2 Na+, 4 Cl– and 55 H2O in the simulation box corresponding to the fluids of 4 molar

Cl–. To investigate the stability of Zn(II)-Cl complexes with different number of chloride

ligands, five simulations were performed at each T-P condition with the initial

configurations of ZnCln2-n (n = 0, 1, 2, 3, 4). Simulation 5 was conducted with 10 Cl– and

initial configuration of ZnCl3(H2O)– and ZnCl42– to test the stability of tetrahedral ZnCl4

2–

complex in chloride rich fluids.


77

Table 4.1 Solution composition, temperature, pressure and density of Zn(II)-Cl simulations

Job No. Solution composition T

(˚C) P

(bar) Box size

(Å) Density (g/cm3)

1 1 Zn, 2 Na, 4 Cl, 55 H2O 25 1 12.193 1.14 2 1 Zn, 2 Na, 4 Cl, 55 H2O 200 14 12.658 1.02 3 1 Zn, 2 Na, 4 Cl, 55 H2O 350 500 13.235 0.89 4 1 Zn, 2 Na, 4 Cl, 55 H2O 600 2000 13.848 0.78 5 1 Zn, 8 Na, 10 Cl, 55 H2O 600 2000 14.040 0.96

Periodic boundary conditions were used to eliminate surface effects. The computing costs of

ab initio simulations were discussed in Mei et al. (2013a); consequently, in this study,

simulation boxes with 172-184 atoms were chosen to provide manageable computation

times while enabling the simulation of realistic solution compositions. MD simulations were

performed from ambient to hydrothermal-magmatic conditions (Table 4.1). The fluid

densities were chosen to correspond to the equation of state of NaCl fluids at the same ionic

strength at the pressure and temperature of interest (Driesner, 2007; Driesner and Heinrich,

2007). All simulations were run for more than 10 picoseconds. To obtain the time average

of the geometric and stoichiometric information, radial distribution functions (RDF) of Zn-

Cl and Zn-O pairs and their integrals (reflecting the time averaged coordination number)

were calculated by using VMD (Humphrey et al, 1996).

4.2.2 Ab initio thermodynamic integration

Ab initio MD simulations gave qualitative information about the relative stability of some

species. However, since the ligand exchanges (timeframe of "s for Zn-Cl and ns for Zn-H2O)

cannot be observed during the short MD simulation (e.g., less than 10 ps), the distribution of

species with close energies cannot be retrieved from these calculations. To estimate the

stability of Zn(II)-Cl complexes quantitatively, thermodynamic integration (Resat, 1993;

Sprik and Ciccotti, 1998) was employed to evaluate energetic properties.


78

As described in our previous study (Mei et al., 2013a), to calculate the binding free energy

between Zn2+ and Cl–, a series of Zn-Cl distance constraint calculations were performed by

constraining Zn-Cl distances along predefined reaction paths. The constraint force was

recorded during the simulation time by sampling possible configurations of Zn(II)-Cl

complexes and the surrounding solvent and ions at each distance-constraint, and the mean

constraint force f(r) at fixed distances r was obtained by computing the average of the

constraint force. Here, f(r) is the force necessary to maintain the chloride ion at a distance r

from the Zn2+ ion, where r varies from ‘infinity’ (i.e., negligible interaction between Zn2+

and Cl– ions) to a bonded state (~2.2 Å). In each series of distance-constrained calculation,

the simulation parameters were kept identical in the constrained and unconstrained

simulations. All the constrained simulations were calculated for more than 5 ps, including

0.7 ps for stablization (Bühl et al., 2006; 2008). Then the mean forces f(r) as a function of

constrained Zn-Cl distances (r) from ~2.2 Å to 5 Å were obtained. The change in free

energy for the reaction was derived by integrating f(r) with respect to the constrained

distance (r) (Bühl et al., 2006; 2008; Sprik and Ciccotti, 1998):

(4.1)

To test the stability of ZnCl42– at 600 °C, the free energy of the association reaction

[ZnCl3–] + Cl– = [ZnCl4

2–] (4.2)

was calculated. As ZnCl42– dissociates to [ZnCl3

–] at 600 °C according to unconstrained MD

simulations, all Zn-Cl distances were constrained in this series of calculation (Mei et al.,

2013a). Three of the Zn-Cl distances were constrained at the distance 2.3 Å, the other Zn-Cl

distance was constrained along the reaction path, with values from 2.3 to 5.0 Å. Free energy

surface of Zn-Cl association reactions were calculated at the temperature of 200 °C, 350 °C

and 600 °C.

!

"Aa# b = $ f (r) dra

b%


79

As the free energy surface of Zn-Cl association reactions were calculated by thermodynamic

integration, the stability constants of different Zn-Cl species could be derived. We

conducted calculations at constant volume, so that the Helmholtz free energies were

obtained. Similarly to Mei et al. (2013a), the Gibbs free energies of the reaction were

approximated by assuming:

'rG = 'Aa!b (4.3)

After activity correction, the formation constants were calculated from

(4.4)

4.3 Results

4.3.1 Ab initio molecular dynamic simulations

Ab initio MD simulations of Zn(II)-Cl complexes were conducted at different temperatures

and pressures with varying initial configurations. At ambient conditions (25 °C, 1 bar), the

number of Cl– complexed to Zn2+ (from 0 to 4 in the initial configuration; simulations 1a-e)

does not change during the first 14 ps. MD simulation 1a was started with an octahedral

Zn(H2O)62+ structure (Fig. 4.2a); the calculations reveal steady state Zn-O distances of

2.13 Å and a Debye-Waller factor of 0.023 Å. In the following, we use Debye-Waller

factors to provide uncertainties on bond distance (Campbell et al, 1999). Observation of the

dynamic Zn-Cl distances confirms that Cl- did not complex to Zn2+ during the simulation

(Fig. 4.3a). Simulation 1b was started with an octahedral structure, ZnCl(H2O)5+; two water

molecules dissociated after 4.3 ps, resulting in the formation of a tetrahedral complex

ZnCl(H2O)3+, with a Zn-Cl distance of 2.25±0.010 Å and Zn-O distances of 1.99±0.005 Å

(Fig. 4.2c). The Zn-O bond distances are shorter in ZnCl(H2O)3+ than in Zn(H2O)6

2+, as

!

"rG# = $RT lnK#


80

expected for a transition from octahedral to tetrahedral geometry. In simulations 1c, 1d and

1e, Zn(II)-Cl complexes kept the same structure as the initial configuration with no Cl-Cl

exchange (Fig. 4.3). The Zn-Cl and Zn-O bond distances are listed in Table 4.2. The

dynamic Zn-Cl distances shown in Figure 4.3 indicate that the species Zn(H2O)62+

(Fig. 4.2a), ZnCl(H2O)3+ (Fig. 4.2c), ZnCl2(H2O)2 (Fig. 4.2d), ZnCl3(H2O)– (Fig. 4.2e),

ZnCl42– (Fig. 4.2f) can be (meta)stable at 25 °C, 1 bar over ps time lengths, reflecting the

slow kinetics of the ligand exchange reactions between Zn2+ and Cl– ions. In terms of

geometry, the only stable octahedral complex was the Zn(H2O)62+ aqua-complex (Fig. 4.2a).

Once Cl– complexes to Zn2+, the structure changes from octahedral to tetrahedral within a

few ps (Figs. 4.2c,d,e,f).

Figure 4.2 Snapshot of Zn(II)-Cl complexes

!"

#$

%&

'() '*) '+) ',)

'-) '.) '/)


81

Table 4.2 Geometrical details of Zn(II)-Cl complexes by MD simulations

T, P Job No.

Simulation time (ps)

Initial configuration Final species dZn-Cl

(Å) dZn-O (Å)

1a 14.51 Zn(H2O)62+ Zn(H2O)6

2+ - 2.13

1b 14.51 ZnCl(H2O)5+ ZnCl(H2O)3

+ 2.25 1.99

1c 14.51 ZnCl2(H2O)20 ZnCl2(H2O)2

0 2.27 1.99

1d 14.51 ZnCl3(H2O)– ZnCl3(H2O)– 2.26 2.05

25 ˚C, 1 bar

1e 14.51 ZnCl42– ZnCl4

2– 2.30 -

2a 14.51 Zn(H2O)62+ ZnCl(H2O)3

+ 2.23 2.00

2b 14.80 ZnCl(H2O)5+ ZnCl2(H2O)2

0 2.24 2.01

2c 14.51 ZnCl2(H2O)20 ZnCl2(H2O)2

0 2.24 2.03

2d 15.53 ZnCl3(H2O)– ZnCl3(H2O)– 2.25 2.03

200 ˚C, 14 bar


2– 2.30 -

3a 14.51 Zn(H2O)62+ ZnCl(H2O)3

+ 2.21 2.02

3b 14.51 ZnCl(H2O)5+ ZnCl(H2O)3

+ 2.22 2.03

3c 14.51 ZnCl2(H2O)20 ZnCl2(H2O)2

0 2.23 2.06

3d 14.51 ZnCl3(H2O)– ZnCl3(H2O)–

& ZnCl3– (minor) 2.28 2.11

350 ˚C, 500 bar


2– 2.29 -

4a 14.62 Zn(H2O)62+ ZnCl2(H2O)2

0 2.20 2.14

4b 16.52 ZnCl(H2O)5+ ZnCl2(H2O)2

0 2.21 2.07

4c 17.63 ZnCl2(H2O)20 ZnCl3(H2O)–

& ZnCl3– (major) 2.22 2.21

4d 17.63 ZnCl3(H2O)– ZnCl2(H2O)20 2.22 2.08

600 ˚C, 2000 bar

4e 17.49 ZnCl42– ZnCl3(H2O)–

& ZnCl3– (major) 2.23 2.32


82

Figure 4.3 Dynamic distances of Zn-Cl in the simulations at 25 ˚C with different initial

configurations

12

10

8

6

4

2

014121086420

25˚C

Zn(H2O)62+

12

10

8

6

4

2

014121086420

25˚C

ZnCl(H2O)3+

ZnCl(H2O)4+

ZnCl(H2O)5+

4

3

2

114121086420

25˚C

ZnCl42-

12

10

8

6

4

2

014121086420

25˚C

ZnCl2(H2O)2

12

10

8

6

4

2

014121086420

25˚C

ZnCl3(H2O)-


Zn-C

l dis

tanc

es (Å

)

(a)(a)

(b)

(c)

(d)

(e)


83

Simulation 2a shows that at 200 °C, 14 bar (saturation pressure), the octahedral aqua

complex Zn(H2O)62+ becomes unstable. Zn(H2O)6

2+ first loses two water molecules, and

then one Cl– complexed to Zn2+ and replaced one of the remaining 4 waters to form the

tetrahedral complex ZnCl(H2O)3+ (Fig. 4.2c, Fig. 4.4a). Ligand (Cl–) association and water

dissociation reactions are also observed in simulation 2b, which was started with the initial

species ZnCl(H2O)5+ (Fig. 4.4b). One water molecule dissociated at the beginning of the

MD run, resulting in the ZnCl(H2O)4+ complex. This species lasted for ~1 ps until one water

molecule dissociated and tetrahedral ZnCl(H2O)3+ formed. As shown in Figure 4.4b, after

~7 ps, one Cl– ligand replaced one water molecule and ZnCl2(H2O)20 (Fig. 4.2d) formed as

the final species, with Zn-O and Zn-Cl distances of 2.01±0.012 Å and 2.24±0.007 Å,

respectively. In simulations 2c, 2d and 2e, similarly to simulation 1, the number of chlorides

remained the same as in the initial configuration during the whole simulation: ZnCl2(H2O)20

(Fig. 4.4c), ZnCl3(H2O)– (with 4% of ZnCl3–, Figs. 4.2f, 4.4d) and ZnCl4

2– (Figs. 4.2g, 4.4e).

The Zn-Cl and Zn-O bond distances are listed in Table 4.2. The MD simulations at 200 °C

indicate that rather than the unstable Zn(H2O)62+ aqua ion, tetrahedral Zn(II)-Cl complexes

with 1-4 Cl– can exist in chloride-rich fluids. The MD simulations at 200 °C also show

faster kinetics than simulations at room temperature, as Cl-H2O ligand exchange reactions

could be observed in 15 ps simulations (Fig. 4.4b).


84


configurations.

12

10

8

6

4

2

014121086420

200˚C

ZnCl2(H2O)2 ZnCl(H2O)3+

ZnCl(H2O)4+

12

10

8

6

4

2

014121086420

200˚C

ZnCl2(H2O)2

12

10

8

6

4

2

014121086420

200˚C

Zn(H2O)62+

ZnCl(H2O)3+

4

3

2

114121086420

200˚C

ZnCl42-

12

10

8

6

4

2

014121086420

200˚C

ZnCl3(H2O)-

ZnCl3-

Zn-C

l dis

tanc

es (Å

)


(a)

(b)

(c)

(d)

(e)


85

At 350 °C, 500 bar, the octahedral aqua complex Zn(H2O)62+ was also found to be unstable

in a 4 m Cl solutions, similarly to the 200 °C case. Three water molecules in Zn(H2O)62+

dissociated gradually and one Cl– ligand complexed to Zn2+ to form the tetrahedral

ZnCl(H2O)3+ complex (Fig. 4.5a) with intermediate species of Zn(H2O)4

2+ (Fig. 4.2b). In

simulations 3b and 3c (Figs. 4.5b, 4.5c), the number of Cl– does not change during the

simulation, and the tetrahedral geometries are maintained throughout the simulations. In

simulation 3d with the starting structure of ZnCl3(H2O)–, the number of Cl– does not change

during the simulation, but sometimes one water molecule left the complex to form trigonal

ZnCl3– (22% of [ZnCl3

-] complex). Simulation 3e (Fig. 4.5e) shows that ZnCl42– is stable at

350 °C in a 4 m Cl solution, with an average Zn-Cl bond distance of 2.29±0.014 Å. MD

simulations under hydrothermal conditions (200 and 350 °C) indicate that the Zn2+-aqua

complex exists as tetrahedral Zn(H2O)42+ (Fig. 4.2b), but not as octahedral complex in Cl-

poor fluids; this species is not stable in Cl-rich fluids as water molecules will be replaced by

chloride. Four tetrahedral complexes (ZnCln(H2O)4-n2-n; n=1,2,3,4) were observed in Zn-Cl

fluids.


86


configurations.

12

10

8

6

4

2

014121086420

350˚C

ZnCl3(H2O)- & ZnCl3-

12

10

8

6

4

2

014121086420

Zn(H2O)42+ &

Zn(H2O)52+ ZnCl(H2O)3

+

350˚C

4

3

2

114121086420

350˚C

ZnCl42-

12

10

8

6

4

2

014121086420

350˚C

ZnCl(H2O)3+

12

10

8

6

4

2

014121086420

350˚C

ZnCl2(H2O)2

Zn-C

l dis

tanc

es (Å

)


(a)

(b)

(c)

(d)

(e)


87

MD simulations at 600 °C, 2 kbar were conducted to investigate the behavior of Zn(II)-Cl

complexes under magmatic hydrothermal conditions. As seen in Figure 4.6, the kinetics of

ligand-exchange reactions at 600 °C is faster than at lower temperatures. In simulation 4a,

the initial octahedral complex Zn(H2O)62+ changed to tetrahedral Zn(H2O)4

2+ (Fig. 4.2b)

quickly, then two Cl– complexed to Zn2+ by successive replacement of H2O ligands, forming

the neutral complex ZnCl2(H2O)20 (Fig. 4.2d). Simulation 4b also gave a stable

configuration of ZnCl2(H2O)20 (Fig. 4.6), with one more Cl– complexed to Zn2+ in ~5 ps.

Simulation 4c was started with ZnCl2(H2O)20, and the predominant species are tetrahedral

ZnCl3(H2O)– (Fig. 4.2e, 44%) and trigonal planar ZnCl3– (Fig. 4.2f, 56%) with one more

chloride complexing to zinc after about 5 ps (Fig. 6c). Simulation 4d contained ZnCl3–,

ZnCl3(H2O)– and ZnCl2(H2O)20 complexes, with several ligand exchanges observed during

17 ps. Simulation 4e shows that ZnCl42– may be not stable under these condition (Fig. 4.6e),

since one chloride dissociated to from tetrahedral ZnCl3(H2O)– (44%) and trigonal planar

ZnCl3– (56%).

The dehydration of the [ZnCl3–] complex with increasing temperature, i.e. the increase in

the relative stability of the ZnCl3– form versus the ZnCl3(H2O)– form, as a function of

temperature is summarized in Figure 4.7. To calculate the average number of waters bonded

to [ZnCl3–] (n of ZnCl3(H2O)n

–), a distance cut-off of 2.75 Å was selected; the uncertainties

reflect the change in hydration within a distance of ±0.25 Å (i.e. 2.75±0.25 Å). Hydration

numbers decrease from 1 to 0.98, 0.79, and 0.44 at 25 °C, 200 °C, 350 °C, and 600 °C,

respectively.


88


configurations.

12

10

8

6

4

2

01614121086420

600˚C

ZnCl42- ZnCl3(H2O)- & ZnCl3

-

12

10

8

6

4

2

01614121086420

600˚C

ZnCl3- ZnCl2(H2O)2

ZnCl3(H2O)-

ZnCl3-

ZnCl2(H2O)2

12

10

8

6

4

2

01614121086420

ZnCl2(H2O)2

600˚C


12

10

8

6

4

2

01614121086420

600˚C

ZnCl(H2O)3+ ZnCl2(H2O)2

12

10

8

6

4

2

014121086420

600˚C

ZnCl(H2O)3+

ZnCl2(H2O)2

Zn(H2O)42+


Zn-C

l dis

tanc

es (Å

)

(a)

(b)

(c)

(d)

(e)


89

Figure 4.7 Radial distribution function (RDF, blue curves) and coordination number (CN,

red curves) of Zn-O of ZnCl3(H2O)n– complexes in selected MD simulations. The error of

No is calculated within distance width of 0.5 Å.

1.0

0.5

0.06420

2

1

0

(e) 600˚C, Sim_4e, 13.2 ps

1.0

0.5

0.06420

2

1

0

(d) 600˚C, Sim_4c, 12.5 ps

2.0

1.5

1.0

0.5

0.06420

4

3

2

1

0

(c) 350˚C, Sim_3d, 14.5 ps

Zn-O

Rad

ial D

istri

butio

n Fu

nctio

n (R

DF)

Zn-O C

oordination Num

ber (CN

)

Zn-O Distances (Å)

3.0

2.0

1.0

0.06420

4

3

2

1

0

(b) 200˚C, Sim_2d, 15.5 ps

4.0

3.0

2.0

1.0

0.06420

4

3

2

1

0

(a) 25˚C, Sim_1d, 14.5 ps

CN RDF


90

In Cl-rich fluids (simulation 5, Cl– concentration of 10 m), when starting with an initial

configuration of ZnCl3(H2O)–, tetrahedral ZnCl3(H2O)– and trigonal planar ZnCl3–

predominate with short presence (~2 ps) of tetrahedral ZnCl42– (Fig. 4.8a). The simulation

started with tetrahedral ZnCl42– gives shows the ZnCl4

2–, ZnCl3(H2O)– and ZnCl3– species

over the 27.4 ps simulation time, with at least one observable Cl-Cl ligand exchange

(Fig. 4.8).

Figure 4.8 Dynamic distances of Zn-Cl in the simulations at 600 ˚C in 10 m Cl– solutions

with different initial configurations (a) ZnCl3(H2O)–, (b) ZnCl42–.

4.3.2 Ab initio thermodynamic integration and free energy calculations

MD simulations results listed in the previous section gave us a basic idea about the

speciation of Zn(II)-Cl complexes from ambient to hydrothermal-magmatic conditions; the

simulations also help to understand qualitatively the stability of certain Zn(II)-Cl complexes

12

10

8

6

4

2

02826242220181614121086420

ZnCl42-

600˚C


12

10

8

6

4

2

02826242220181614121086420


ZnCl42-

600˚C

Zn-C

l dis

tanc

es (Å

)


(a)

(b)


91

under different conditions. In principle, formation constants for the Zn(II) complexes could

be derived from the ratio of abundance of the complexes in solutions where at least two

complexes co-exist. In order to achieve this, the ligand exchange reactions must happen at a

rate much faster than the simulation time. However, in the MD simulations described above,

the constant number of complexed chloride in many simulations (e.g., 1a-e, 2c-e, 3b-e)

suggests a slow kinetics of chloride-exchange reactions. The time scale of MD simulation is

small comparing with the exchange rate of Cl– ligand with Zn2+ in aqueous solution at room

temperature (~300 µs for Cl–, Sharps et al., 1993). Although the kinetics at high temperature

is faster and ligand exchange can be observed in some simulations (e.g., 2a-b, 3a, 4a-e, 5a-

b), the statistics of ratio of different species is not good enough to calculate the stability of

the complexes. Consequently, we used thermodynamic integration to predict the energetic

properties and help to estimate the relative stabilities of Zn(II)-Cl complexes (semi)-

quantitatively.

Ab initio thermodynamic integration was applied at 200, 350 and 600 °C to obtain the free

energy of ligand exchange reactions. Using the same solution composition as listed in

Table 4.1 (No. 2-4), constrained MD simulations were performed by fixing Zn-Cl distances.

Figure 4.9 shows an example of calculating the free energy of reaction (4.7) (substitution of

a water ligand by a chloride) through thermodynamic integration.

Zn(H2O)3Cl+ + H2O = Zn(H2O)42+ + Cl– (4.7)

The Helmholz free energy (!Aa!b) of reaction (4.7) at 350 °C was obtained by integrating

the constraint mean force with respect to the constrained Zn-Cl distances. The force is near

zero (5.25 kJ‧Å-1‧mol-1) at 2.20 Å - the distance corresponding to the equilibrated Zn-Cl

bond distance in the MD simulation. With increasing Zn-Cl distance, an external force must

be applied in order to maintain a given Zn-Cl distance because of the attraction between

Zn2+ and Cl–. The maximum absolute value of the constraint force is reached at a distance of


92

2.6 Å (-106.1 kJ‧Å-1‧mol-1), then the absolute magnitude of the force decreases with

increasing Zn-Cl distances. The constraint force becomes zero again at a distance of ~3.2 Å,

and becomes slightly positive in the range of 3.2-4.3 Å, which results from the outer

solvation shell and reflects the activation barrier for the ion exchange reaction. Beyond the

Zn-Cl distance of 4.5 Å, the force between Zn2+ and Cl– is negligible. Considering the size

of the simulation box (13.235 Å at 350 °C), the distances of 4.5-5.0 Å can be recognized as

“safe” distance beyond which the interaction between Zn2+ and Cl– ligand can be neglected

(Mei et al., 2013a).

Figure 4.9 Constraint mean force and the integral of reaction (4.7) at 350 ˚C, 500 bar

Integration of the mean constraint force along the reaction coordinates gives a free energy

difference of +39.7 kJ‧mol–1 for reaction (4.7). The free energies for the stepwise formation

of [ZnCln2-n] (n=1,2,3,4) complexes were calculated at 200, 350 and 600 °C (Fig. 4.10).

!"#$

!"$$

!%$

!&$

!'$

!#$

$

#$

'$

&$

()$')(')$*)(*)$#)(#)$+,!-./01234,56/789

#)#$/8

#)&$/8

#)%$/8

()$$/8

:/;/+,7</+,7<#:9*-.=/=/<# #:9'

#=/=/-.!!"#

!$#

!%#

!&#

>//7?@/8!"/AB.!"

!"9


93

Figure 4.10 Free energy surface of Zn-Cl dissociation reactions.

At 200 °C, positive energies of 22 to 24 kJ‧mol–1 were obtained the substitution of a

chloride by a water in the tetrahedral ZnCl(H2O)3+, ZnCl2(H2O)2(aq) and ZnCl3(H2O)–

complexes, which reflects the strong tendency of Zn(H2O)42+ to react with Cl– to form

!"

#"

$"

%"

&"

'"

("

"

)("

$*"%*$%*"&*$&*"'*$'*"

+++,-./'01&234+4+/'0+5+,-./'01%

'4+4+23)+

+++,-./'01'23'+4+/'0+5++,-./'01&2344+23)+

+++,-./'0123&)+4+/'0+5++,-./'01'23'+4+23

)+

+++,-23%')+4+/'0+5++,-./'0123&

)+4+23)++

.61++#""+!2

$"

%"

&"

'"

("

"

)("$*"%*$%*"&*$&*"'*$'*"

.71+++&$"!2

+++,-./'01&234+4+/'0+5+,-./'01%

'4+4+23)+

+++,-./'01'23'+4+/'0+5++,-./'01&2344+23)+

+++,-./'0123&)+4+/'0+5++,-./'01'23'+4+23

)+

+++,-23%')+4+/'0+5++,-./'0123&

)+4+23)++

%"

&"

'"

("

"

)("$*"%*$%*"&*$&*"'*$'*"

.81++'""!2

+++,-./'01&234+4+/'0+5+,-./'01%

'4+4+23)+

+++,-./'01'23'+4+/'0+5++,-./'01&2344+23)+

+++,-./'0123&)+4+/'0+5++,-./'01'23'+4+23

)+

+++,-23%')+4+/'0+5++,-./'0123&

)+4+23)++

9+.:;<=>31

,-)23+?@AB8-6CA+.D1


94

tetrahedral [ZnCln(H2O)4-n]n-2 (n=1,2,3) complexes. The positive free energies of

12.3 kJ‧mol–1 for the ZnCl42– complexes also indicates the preference of forming 4-Cl

tetrahedral complex in the simulated 4 m chloride solution, but the energy gained in

replacing the last water is not as strong as for the previous ([ZnCln(H2O)4-n]n-2; n=1,2,3)

steps.

The free energies for step-wise formation constants at 350 °C show similar trends as at

200 °C: tetrahedral [ZnCln(H2O)4-n]n-2 (n=1,2,3,4) complexes are preferred, with strong

tendency of Zn2+ to react with 1-3 Cl–. The hydration change for the ZnCl3– and

ZnCl3(H2O)– complexes is neglected at 600 °C as the solvent exchange is fast (less than

1 ps). As shown in Figure 4.10c, Zn(II)-Cl complexes of [ZnCln(H2O)4-n]n-2 (n=1,2,3,4) are

also favored at 600 °C, but the relatively smaller energies of stepwise formation for

ZnCl3(H2O)– and ZnCl42– indicate that those two species are not as stable as ZnCl(H2O)3

+

and ZnCl2(H2O)2(aq); these results are consistent with the unconstrained MD shown in

Figure 4.6.

The results of the distance constraint thermodynamic integration for the Zn-Cl association

reactions are listed in Table 4.3. Three versions of Gibbs free energies of reaction are listed:

'rG(P,T) is the Gibbs free energy of the reactions calculated from thermodynamic

integration (i.e., in a 4 m Cl solution), 'rG(,c(P,T) is the Gibbs free energy with

concentration correction (i.e., infinite dilution at the P,T of the simulation), and 'rG((Pr,Tr)

is the standard state Gibbs free energy under Pr = 1 bar, Tr = 25 °C (Mei et al., 2013a). The

errors of 'rG(P,T) are calculated from the standard deviation of the force integral within the

distance range of 4 to 5 Å. log K((P,T) values of the reactions which form different Zn(II)

complexes were calculated according to Equation (4.4).


95

Table 4.3 Gibbs free energy of reaction and stability constants of the Zn-Cl association

reactions

Reaction T (˚C) 'rG(P,T) (kJ/mol)

'rG(,c(P,T) (kJ/mol)

'rG((Pr,Tr) (kJ/mol) logK!(P,T)

Zn2+ + Cl– = ZnCl+ 200 -23.2±1.6 -14.4 -23.4 2.58±0.18 ZnCl+ + Cl– = ZnCl2(aq) 200 -23.3±1.6 -15.8 -19.5 2.16±0.18 ZnCl2(aq) + Cl– = ZnCl3

– 200 -22.4±1.9 -16.9 -17.4 1.92±0.21 ZnCl3

– + Cl– = ZnCl42– 200 -12.3±3.2 -12.3 -8.5 0.94±0.35

Zn2+ + Cl– = ZnCl+ 350 -39.7±1.0 -28.0 -62.4 5.23±0.08 ZnCl+ + Cl– = ZnCl2(aq) 350 -32.3±2.7 -22.4 -40.0 3.35±0.23 ZnCl2(aq) + Cl– = ZnCl3

– 350 -37.9±1.7 -30.7 -31.7 2.66±0.14 ZnCl3

– + Cl– = ZnCl42– 350 -15.0±3.9 -15.0 -11.3 0.95±0.33

Zn2+ + Cl– = ZnCl+ 600* -67.4±3.1 -51.1 -98.8 5.91±0.19 ZnCl+ + Cl– = ZnCl2(aq) 600 -66.1±2.6 -52.2 -76.5 4.58±0.16 ZnCl2(aq) + Cl– = ZnCl3

– 600 -39.9±2.4 -29.8 -31.3 1.87±0.14 ZnCl3

– + Cl– = ZnCl42– 600 -21.2±4.7 -21.2 -15.4 0.92±0.28

*bdot parameter at 500 ˚C

4.4 Refitting solubility data for Zn(II)-Cl complexes

4.4.1 Data sources and methods

In this section, we attempted to reinterpret the available solubility data of Ruaya and Seward

(1986) (R&S) and Bourcier and Barnes (1987) (B&B) to get a consistent thermodynamic

dataset for zinc chloride complexes that can describe both datasets. We aim to reconcile

both studies by fitting both datasets using a unified model and use the MD results to choose

the speciation model. Further impetus for refitting the solubility measured by R&S and

B&B is provided by the availability of improved thermodynamic properties for some of the

solids and aqueous species required in the interpretation of the experimental solubilities

(Table 4.4). These new data may affect the refined log K data for the Zn(II) chloride

complexes.


96

Table 4.4 Species included in the solubility calculations.

Species Source New data

ZnO ZnCO3 AgCl

Zincite (ZnO(s)) Wesolowski et al., 1998 Yes % Smithsonite (ZnCO3(s)) Preis et al., 2000 Yes % Chlorargyrite (AgCl(s)) Zotov et al., 1995 Yes % HCl(aq) Tagirov, 1997 Yes % % % ZnCli

(2-i) Fitted in the analysis % % % Zn(OH)+ and Zn(OH)2(aq) Bénézeth et al., 2002 Yes % % AgCl(aq), AgCl2

–, AgCl32– Zotov et al., 1995 Yes %

NaCl(aq) Svenjensky et al. 1997 No NaCO3

- Smith and Martell, 1976 No % % NaHCO3(aq) Smith and Martell, 1976 No % % CO2(g) Pankratz, 1982 No % % H2CO3(aq) No % % Basis species are: Ag+, Cl-, H+, HCO3

-, Na+, Zn2+ % species are used in the model

4.4.2 Fitting procedure

The fitting was conducted according to the procedure illustrated in Figure 4.11.

Figure 4.11 Schematic view of the procedure used to fit the solubility data of B&B and

R&S simultaneously.


97

Table 4.5 Modified Ryzhenko-Bryzgalin (MRB) equation of state parameters for Zn-Cl complexes fitted in this study (Ryzhenko et al., 1985; Shvarov and Bastrakov, 1999).

Reaction pK25˚C A B

ZnCl+ = Zn2+ + Cl– -0.46 0.50 210.60 ZnCl2(aq) = Zn2+ + 2Cl– -0.40 2.60 -554.20 ZnCl3

– = Zn2+ + 3Cl– -3.20 1.10 731.90 ZnCl4

2– = Zn2+ + 4Cl– -0.10 3.50 -1020.70

(i) A thermodynamic model was set-up to calculate the solubility in each experiment, based

on full distribution of species calculations. The model includes the species listed in

Table 4.5. The relevant mass action, mass balance and charge balance equations were

solved using an implementation of the EQBRM code in Matlab (Anderson and Crerar, 1993;

Brugger et al., 2007). As in previous studies (e.g., Liu et al., 2001), an extended Debye-

Hückel expression (Helgeson, 1969) was used to calculated individual molal activity

coefficient for charged species:

, (4.8)

where A! and B! are the Debye-Hückel solvent parameters taken from Helgeson and Kirkham

(1974);

!

˙ a n is the distance of closest approach of ion n, and is given a value of 5 Å for

divalent and trivalent ions, and 4.0 Å for monovalent ions, except for H+ and Zn2+ which are

given values of 9 and 6 Å, respectively (Kielland 1937); is the effective ionic strength

using the molal scale; ! is a mole fraction to molality conversion factor; and B!,SALT is the

extended-term parameter (b-dot coefficient) for NaCl (B&B) and HCl (R&S) solutions

taken from Helgeson et al. (1981).

(ii) Initially, each isothermal dataset was fitted individually ((1) in Fig. 11). At each

temperature, the formation constants of the four possible Zn(II) chloride complexes (i.e.,

!

log(" n ) = #A"Zn

2I 1 / 2

1+ B" ˙ a nI 1 / 2 + B" ,SALT I +$

!

I


98

ZnClx2-x, x = 1-4) were optimized using the nonlinear least squares simplex method (Nelder

and Mead, 1965), so as to minimize the difference between the solubilities calculated using

the model (

!

yi ) and the experimental solubilities (

!

yi0). In order to provide the same

weighting for the two experimental studies (irrespective of the number of individual

solubility measurements in each dataset), the minimized residual function R was defined as:

!

R = 1nexp

log(yi ) " log(yi0)( )2

i

nexp

#exp

R& S /B&B

# (4.9)

Maps of the residual R were calculated as a function of log K (ZnClx2-x) around the

optimized values, and plotted in two dimensions to examine the uncertainties in the derived

log K values. These uncertainties are estimated as log K ranges around optimized values at

an approximate confidence level of 95% for non-linear regression (e.g., cuprite solubility

experiments of Liu et al. 2001). To calculate the 95% confidence level for the isothermal fits,

we used the relationship (Draper and Smith, 1998):

!

R95% = Rmin 1+p

n " pF(p,n " p,1"#)

$ % &

' ( ) (4.10)

where p is the number of degrees of freedom, n the number of independent measurements,

and F is the F-distribution; for p = 4, n ~40, " = 0.05, F = 2.6, and

!

R95% " Rmin #1.3.

The residual maps obtained for the 200 °C dataset are shown in Figure 4.12; similar maps

were obtained at each temperature. At 200 °C, only minimum values for the formation

constants of ZnCl42– and ZnCl2(aq) can be given; the range of values for the formation of

ZnCl3– changes dramatically as a function of the ZnCl4

2- value. Only the formation constant

of ZnCl2(aq) is fixed within ~1.5 log unit.


99

Figure 4.12 The residual maps of fitting the 200 °C solubility data

(iii) In order to minimize the uncertainty in the values of log K (ZnClx2-x) resulting from the

large number of possible species, we performed a new analysis in which all datasets (all

temperatures) were included ((2) in Fig. 4.11). In this approach, instead of fitting log K’s at

each temperature, we optimized the parameter for the 4 Zn(II) chlorocomplexes in the

modified Ryzhenko-Bryzgalin (MRB) equation of state (Ryzhenko et al., 1985; Shvarov and

Bastrakov, 1999):

!

pK(P,T ) =298.15T

pK(298.15K ,1bar)+ B(P,T ) A+BT

"

# $

%

& '

B(P,T ) = 11.0107 logKw (P,T )(T 298.15K logKw (298.15K,1bar)( ) (4.11)

A and B are empirical parameters, and Kw(P,T) is the dissociation of water. Ki(25 °C,1bar)

for ZnCl+ and ZnCl42– were fixed to the values of 0.46 and 0.10, respectively, from Liu et al.

(2007); all the other EOS parameters were allowed to vary during the least square fit.

4.4.3 Speciation of Zn(II)-Cl based on solubility experiments

The results of the new fitting are shown in Figure 4.13, and the MRB parameters in

Table 4.5. In general the calculated and experimental data are within error. Note that fits

using the HCl0 data from Ruaya and Seward (1987) provides better fits at high Cl

LogK

(ZnC

l 42–)

LogK(ZnCl2(aq)) = 0

LogK(ZnCl +) LogK(ZnCl3–)

LogK

(ZnC

l 42–)

LogK(ZnCl +) LogK(ZnCl3–)

Red volume: 95% con!dence level

0

0.04

0.06

0.08

0.2

0.22

0

0.2

0.3

0.4

0.6

0.7

LogK(ZnCl2(aq)) = -3.93

!"# !$#


100

concentrations, although the refined properties and speciations are very similar. This may be

due to the fact that Ruaya and Seward (1987) used a similar chemical system (AgCl(s)) to

measure HCl association.

Figure 4.13 Results of the new fitting of the solubility data of B&B and R&S. Red circles

represent experimental data and the lines represent fitted data.

At 100 °C, with the best fit for ZnCl42– and maximal values for other species, the speciation

shows the increased predominance of ZnCl42– as chloride concentration increases

(Fig. 4.13a). At 150 °C, ZnCl42– still predominates in high chloride concentration solution,

and ZnCl3– becomes more important in solutions of ~1-2 m Cl– (Fig. 4.13b). At 200 °C, as

shown in Fig. 4.13c, ZnCl3– becomes predominant in most solutions, while ZnCl4

2– acts as

the second important species at chloride concentrations > 1 m. Our fitting of 350 °C data

gives different results for B&B and R&S (Fig. 4.13d). The fitting based on B&B data

suggests predominant species of ZnCl3– at lower chloride concentration (~ less than 1 m)

10 100 10110

10

10

Zn_t

ot [m

]

10 100 1010

0.4

0.6

0.8

1

10010

10

10

Ag_t

ot [m

]

R&S1986 @ 100

1000

0.4

0.6

0.8

10 100 10110

10

10B&B1988, smithsonite @ 150

10 100 1010

0.4

0.6

0.8

10010

10

10

10R&S1986 @ 150

1000

0.4

0.6

0.8

10 100 10110

10

10

10 100 1010

0.4

0.6

0.8

10010

10

10

1000

0.4

0.6

0.8

10010

10

10

10

1000

0.4

0.6

0.8

1

10010

10

10

100

1000

0.4

0.6

0.8

1

Zn

–

4

!"# !$# !%# !&#


101

and ZnCl42– at high chloride concentration (~ more than 1 m), while the fitting results of

R&S data indicates ZnCl3– predominates in all chloride concentration ranges.

The formation constants from the solubility studies of Ruaya and Seward (1986) and

Bourcier and Barnes (1987) are compiled in Table 4.6, and compared with the properties

resulting from our new fit and MD simulations. The new Ryzhenko model for each Zn(II)-

Cl species based on our fitting is plotted as a function of temperature (Fig. 4.14). The log K

values from different studies show the same trend as change of temperature. It is noticed

that fitted values at 350 °C are very close to log K obtained by MD simulation (log K value

of 8.57 for ZnCl2(aq) and 11.23 for ZnCl3–). The excellent agreement at 350 °C shows the

potential of MD simulation in applying to the prediction of log K at high temperature.

Despite the over-fitting problem, our fitting agree with the original studies in that the

ZnCl42– is an important species in highly saline solutions at low temperature, and ZnCl3

– is

predominant species at 350 °C. This is also generally in agreement with the MD results.


102

Table 4.6 The log formation constants for zinc(II) chloride complexes

T (˚C)

Zn2+ + Cl– = ZnCl+

Zn2+ + 2Cl– = ZnCl2

0 Zn2+ + 3Cl– = ZnCl3

– Zn2+ + 4Cl– = ZnCl4

2– Reference

1.80±0.01 1.92±0.01 1.36±0.01 2.04±0.01 R&S* 1.2±0.1 1.9±0.1 2.3±0.1 1.4±0.1 B&B* 100

1.02 (<1.7) <1.7 <1.7 1.01 (0.8-

1.2) Refit§

2.89±0.01 2.96±0.01 2.02±0.02 3.21±0.01 R&S* 2.1±0.2 3.0±0.1 3.8±0.1 2.7±0.2 B&B* 150 2.13 (1.9-2.3) <3 3 (2.7-3.2) 1.97 (<2.5) Refit§

4.01±0.01 3.98±0.02 3.00±0.10 4.23±0.01 R&S* 3.1±0.3 4.3±0.2 5.2±0.2 4.4±0.3 B&B* 2.5±0.1 4.2±0.1 W&B&P 2.58 4.74 6.66 7.6 MD§

200

2.26 (1.8-2.6) <3.7 4.42 (4.2-

4.6) 3.19 (<3.7) Refit§

4.4±0.3 5.6±0.2 6.7±0.2 6.0±0.2 B&B*¶ 250¶ 4.96 (4.8-

5.1) <5.8 7.16 (7.0-7.3)

5.94 (5.5-6.2) HCh*¶

275¶ 5.82±0.02 6.73±0.02 - 7.54±0.02 R&S*¶ 5.7±0.4 7.2±0.2 8.1±0.2 7.4±0.2 B&B*

300 5.6 (4.0-6.0) <7.2 8.1 (8.0-8.3) Refit§ 8.00±0.10 9.54±0.10 - - R&S* 7.0±0.6 9.3±0.5 9.3±0.5 7.7±0.6 B&B* 5.22 8.57 11.23 12.18 MD§ 350

<6.3 9.4 (8.8-9.8)

11.4 (11.1-11.7) <10.5 Refit§

600¶ 5.91 10.49 12.36 13.28 MD§


103

Figure 4.14 LogKs as function of temperature

4.5 Discussion

4.5.1 Speciation of Zn(II)-Cl complexes

MD simulations of Zn-Cl solutions show the stability of the 6-fold aqua-species Zn(H2O)62+

at room temperature, in agreement with previous MD studies (Harris et al., 2003; Liu X et

al., 2011). When one Cl– complexes to Zn2+, however, the MD simulations gave the 4-fold

tetrahedral complex ZnCl(H2O)3+ from ambient to hydrothermal conditions, different from

the octahedral ZnCl(H2O)5+ complex at 25 °C demonstrated by Harris et al. (2003). As

shown in Figure 1b, two waters are released and the Zn(II) complex changes from

ZnCl(H2O)5+ to ZnCl(H2O)3

+ in ~4 ps, so the simulation by Harris et al. (2003) conducted

for only one picosecond is not long enough to observe the steady state of Zn-Cl at room

temperature.

12

10

8

6

4

2

040035030025020015010050

(c) ZnCl3–

14

12

10

8

6

4

2

040035030025020015010050

(d) ZnCl42–

10

8

6

4

2

040035030025020015010050

(b) ZnCl20

8

6

4

2

040035030025020015010050

(a) ZnCl+

R_S; B_B; Refit; new_Ryz; MD; Poly(new_Ryz)

Temperature (˚C) Temperature (˚C)

logK

logK


104

There were some discrepancies in the interpretation of the experimental studies by Ruaya

and Seward (1986) and Bourcier and Barnes (1987) at 300 °C: ZnCl3– is the important

species according to Bourcier and Barnes (1987) (Fig. 4.15a), and ZnCl42– (no ZnCl3

–)

predominates in the study by Ruaya and Seward (1986) (Fig. 4.15b). Our new fitting results

(Fig. 4.15c) indicate that ZnCl42– is the most important species in most solutions ([Cl–] >

0.05 m).

Figure 4.15 Speciation plot at 300˚C from different datasets

The change of Zn(II)-Cl speciation with temperature is also calculated using the

thermodynamic properties from the MD simulations. The MD simulations suggest that both

ZnCl3– and ZnCl4

2– are important species in many solutions (Fig. 4.16) from 200-600 °C. At

200 °C, ZnCl3– accounts for more than 40% of Zn(II) in [Cl–] of ~0.02-0.2 m, and ZnCl4

2–

predominates in [Cl–] of > 0.2 m (Fig. 4.16a); while ZnCl3– predominates at 350 °C over

most of the Cl concentration range, together with ZnCl42– preferred when [Cl–] > 1 m

(Fig. 4.16b). Similar trends are found at 600 °C, with ZnCl20 more important at lower [Cl–]

of <0.02 m (Fig. 4.16c). For [ZnCl3–], the speciation also changes with temperature as the

first-shell hydration (i.e. number of water n in first shell of ZnCl3(H2O)n– complex)

decreases with increasing temperature (Fig. 4.7).

100

80

60

40

20

0

x10-6

0.001 0.01 0.1 1

B&B 300˚C 100

80

60

40

20

0

x10-6

0.001 0.01 0.1 1

R&S 300˚C 100

80

60

40

20

0x1

0-6

0.001 0.01 0.1 1

New fit 300˚C

Concentration of Zn2+

ZnCl+

ZnCl20

ZnCl3_

ZnCl42–


105

Figure 4.16 Speciation plot at 200, 350, 600˚C from MD simulations

The refit of experiment data at low temperature (& 350 °C, Figs. 4.13, 4.15c) also suggests

the predominance of the ZnCl42– species. A similar complex, ZnBr4

2–, has been confirmed

by XAS study at temperature of 25-150 °C in Br-brines (Liu W et al., 2012b). MD results

are also fully consistent with the high stability of the charged ZnCl42– complex even at

elevated temperature.

Other transition metals (Cd, Hg) in the same group as Zn(II) share the same close shell d-

electronic configuration (d10), yet the structures and coordinations of these complexes are

different. For example, a maximum number of 3 Cl– coordinates to Cd(II) in hypersaline

solutions (Bazarkina et al., 2010); and Hg2+ has both four-fold tetrahedral and two-fold

linear complexes (Huheey et al., 1983). The size of the M2+ (M=Zn, Cd, Hg) ions increases

(Zn2+ < Cd2+ < Hg2+), and the energy difference between the (n-1)d-(n)s-(n)p orbitals of

these ions decreases with increasing ion size (Orgel, 1958; Fisher and Drago, 1975). This

leads to different hybridization of (n-1)d-(n)s-(n)p orbitals. The large energy separation

between the 3d and 4s orbitals in Zn2+ favors the sp3 hybridization of Zn(II), and accounts

for the high stability of the symmetric tetrahedral ZnCl42– and ZnBr4

2– structures (Huheey et

al., 1983). In contrast, the close energy of the 5d and 6s orbitals in Hg(II) allows d-s

hybridization and the formation of linear complexes (e.g., HgCl2, Orgel, 1958; Huheey et al.,

1983). The predominance of CdCl2(aq) and CdCl3– (Bazarkina et al., 2010) may relate to the

100

80

60

40

20

0

x10-6

0.001 0.01 0.1 1

MD 600˚C 100

80

60

40

20

0

x10-6

0.001 0.01 0.1 1

MD 350˚C 100

80

60

40

20

0

x10-6

0.001 0.01 0.1 1

MD 200˚C

Concentration of Zn2+

ZnCl+

ZnCl20

ZnCl3_

ZnCl42–


106

hybridization of 4d5s5p orbitals favoring distorted tetrahedral complexes such as

CdCl3(H2O)–.

4.5.2 Ab initio MD vs experiments

The formation constants generated using MD and thermodynamic integration are usually

close to those derived from the experiments. In particular, the MD results at 350 °C for

ZnCl2(aq) and ZnCl3– are in excellent agreement with the experimentally derived log Ks

(within one log units, Table 4.6). However, at lower temperature (25 °C and 200 °C), the

discrepancies become larger (up to 2 log units). Because the kinetics of solvent exchange

(~30 ns for H2O, Table 11.4 in Burgess 1978) and ligand exchange for Zn(II) complexes is

slow at low temperature (~300 µs for Cl–, Sharps et al., 1993), it is possible that MD

simulation does not sample all the configurations at lower temperature. At high temperature,

because the kinetics is faster than at low temperature, the simulated structure of the species

and energy of ligand exchange is closer to the equilibrium state; therefore the calculated

log Ks are more reliable.

As detailed in section 5.1, ab initio MD simulations are able to accurately predict, to a

semiquantitative level, the nature and relative stabilities of the predominant Zn(II)

complexes up to high pressures and temperatures. In practice, thermodynamic properties for

these complexes are required to simulate (i) the gradients in minerals solubilities occurring

through complex processing (e.g., fluid mixing, fluid-rock interaction, cooling, phase

separation) and (ii) the amount of Zn that can be transported under specific conditions. In

many studies, gradients are the key result of reactive transport calculations, since these

determine, for example, where ore deposition or scaling will occur. Approximations of

amounts (mass balance) within one order of magnitude are accurate enough in many

applications. In order to test the ability of the thermodynamic properties obtained from the


107

MD simulations to provide useful information for reactive transport modeling, we

extrapolated MRB parameters from the MD results. These MD-based properties were then

used to calculate the mineral solubilities in the experiments of Ruaya and Seward (1986)

and Bourcier and Barnes (1987) (Fig. 4.17).

Figure 4.17 Fit of the experimental data of Bourcier and Barnes (1987) and Ruaya and

Seward (1987) based on parameters extrapolated from the MD results. Red circles represent

experimental data and the lines represent fitted data.

The MD-derived properties reproduce accurately the gradients in ZnCO3(s) and ZnO(s)

solubilities with [Cl–] measured by Bourcier and Barnes (1987) at 100-350 °C. At

temperatures of 100, 150, 200 °C, the amount of total dissolved Zn predicted by MD-based

properties is up to one order of magnitude difference from experiments data; at 350 °C, this

difference is significantly smaller (factor ~2). The overestimated Zn(II) solubility may be

caused by the overestimation of the stability of ZnCl42–, as MD gives relatively high value

10 100 10110

10

10

100

Zn_t

ot [m

]

10 100 1010

0.4

0.6

0.8

1

10010

10

10

10

Ag_t

ot [m

]

R&S1986 @ 100

1000

0.4

0.6

0.8

10 100 10110

10

10

100

10 100 1010

0.4

0.6

0.8

10010

10

10

10

10 100 10110

10

10

100

10 100 1010

0.4

0.6

0.8

10010

10

10

1000

0.4

0.6

0.8

10010

10

100

1000

0.4

0.6

0.8

1

10010

10

10

100

1000

0.4

0.6

0.8

1

Zn

–

4

1000

0.4

0.6

0.8

!"# !$# !%# !&#


108

of log Ks for ZnCl42– (Table 4.6). For R&S data, the Ag solubilities calculated using the

MD-retrieved parameters shows good agreement (within experimental uncertainty) at high

[Cl–] of > 1 m. It is noticeable that the new fit for both B&B and R&S are very close to the

experimental data at 350 °C.

4.6 Acknowledgement

Research funding was provided by the Australian Research Council (ARC) to JB

(DP0878903), and the Minerals Down Under Flagship to WL. We are grateful to T. Seward

and W. Bourcier for discussing their experimental solubility data. The MD calculations

were supported by iVEC through the use of advanced computing resources located in Perth,

Australia, and the computational facilities of the Advanced Computing Research Centre in

University of Bristol, UK. This paper is part of Yuan Mei’s PhD thesis. YM acknowledges

the University of Adelaide for IPRS scholarship and CSIRO Minerals Down Under Flagship

for a scholarship top-up.

4.7 References Anderson, A. J., Mayanovic, R. A., and Bajt, S., 1995. Determination of the local structure

and speciation of zinc in individual hypersaline fluid inclusions by micro-XAFS.

The Canadian Mineralogist 33, 499-508.

Anderson, A. J., Mayanovic, R. A., and Bajt, S., 1998. A microbeam XAFS study of

aqueous chlorozinc complexing to 430 degrees C in fluid inclusions from the

Knaumuehle granitic pegmatite, Saxonian granulite massif, Germany. The Canadian

Mineralogist 36, 511-524.

Anderson, G. M. and Crerar, D. A., 1993. Thermodynamics in geochemistry: The

equilibrium model. Oxford University Press New York.

Bazarkina, E. F., Pokrovski, G. S., Zotov, A. V., and Hazemann, J.-L., 2010. Structure and


109


276, 1-17.

Bénézeth, P., Palmer, D. A., Wesolowski, D. J., and Xiao, C., 2002. New measurements of

the solubility of zinc oxide from 150 to 350 C. Journal of Solution Chemistry 31,

947-973.

Berendsen, H. J. C., Grigera, J. R., and Straatsma, T. P., 1987. The missing term in effective

pair potentials. The Journal of Physical Chemistry 91, 6269-6271.

Bourcier, W. L. and Barnes, H. L., 1987. Ore solution chemistry; VII, Stabilities of chloride

and bisulfide complexes of zinc to 350 degrees C. Economic Geology 82, 1839-1863.

Brugger, J., 2007. BeerOz, a set of Matlab routines for the quantitative interpretation of

spectrophotometric measurements of metal speciation in solution. Computers &

geosciences 33, 248-261.

Brugger, J., Pring, A., Reith, F., Ryan, C., Etschmann, B., Liu, W. H., O'Neill, B., and

Ngothai, Y., 2010. Probing ore deposits formation: New insights and challenges

from synchrotron and neutron studies. Radiation Physics and Chemistry 79, 151-161.

Bühl, M., Kabrede, H., Diss, R., and Wipff, G., 2006. Effect of Hydration on Coordination

Properties of Uranyl(VI) Complexes. A First-Principles Molecular Dynamics Study.

Journal of the American Chemical Society 128, 6357-6368.

Bühl, M., Sieffert, N., Golubnychiy, V., and Wipff, G., 2008. Density Functional Theory

Study of Uranium(VI) Aquo Chloro Complexes in Aqueous Solution. The Journal of




Draper, N. R. and Smith, H., 1998. Applied regression analysis. Wiley, New York.

Driesner, T., 2007. The system H2O–NaCl. Part II: Correlations for molar volume, enthalpy,

and isobaric heat capacity from 0 to 1000°C, 1 to 5000bar, and 0 to 1 XNaCl.


Driesner, T. and Heinrich, C. A., 2007. The system H2O–NaCl. Part I: Correlation formulae

for phase relations in temperature–pressure–composition space from 0 to 1000 °C, 0

to 5000bar, and 0 to 1 XNaCl. Geochimica et Cosmochimica Acta 71, 4880-4901.




Acta 62, 3095-3107.


110

Harris, D. J., Brodholt, J. P., Harding, J. H., and Sherman, D. M., 2001. Molecular dynamics

simulation of aqueous ZnCl2 solutions. Molecular Physics 99, 825-833.


chloride brines: Results from ab initio molecular dynamics calculations. Journal of


Helgeson, H. C., 1969. Thermodynamics of hydrothermal systems at elevated temperatures

and pressures. American Journal of Science 267, 729-804.

Helgeson, H. C. and Kirkham, D. H., 1974. Theoretical prediction of the thermodynamic

behavior of aqueous electrolytes at high pressures and temperatures; II, Debye-

Hückel parameters for activity coefficients and relative partial molal properties.

American Journal of Science 274, 1199-1261.

Helgeson, H. C., Kirkham, D. H., and Flowers, G. C., 1981. Theoretical prediction of the

thermodynamic behavior of aqueous electrolytes by high pressures and temperatures;

IV, Calculation of activity coefficients, osmotic coefficients, and apparent molal and

standard and relative partial molal properties to 600 degrees C and 5kb. American

Journal of Science 281, 1249-1516.

Humphrey, W., Dalke, A., and Schulten, K., 1996. VMD: visual molecular dynamics.

Journal of molecular graphics 14, 33-38.

Kielland, J., 1937. Individual activity coefficients of ions in aqueous solutions. Journal of

the American Chemical Society 59, 1675-1678.

Kruh, R. F. and Standley, C. L., 1962. An X-Ray diffraction study of aqueous zinc chloride

solutions. Inorganic Chemistry 1, 941-943.

Laasonen, K., Pasquarello, A., Car, R., Lee, C., and Vanderbilt, D., 1993. Car-Parrinello

molecular dynamics with Vanderbilt ultrasoft pseudopotentials. Physical Review B

47, 10142-10153.

Lin, I. C., Seitsonen, A. P., Tavernelli, I., and Rothlisberger, U., 2012. Structure and

Dynamics of Liquid Water from ab Initio Molecular Dynamics—Comparison of

BLYP, PBE, and revPBE Density Functionals with and without van der Waals

Corrections. Journal of Chemical Theory and Computation 8, 3902-3910.

Liu, W., Borg, S., Etschmann, B., Mei, Y. and Brugger, J., 2012. An XAS study of

speciation and thermodynamic properties of aqueous zinc bromide complexes at 25-

150 °C. Chemical Geology, 298-299, 57-69.


111

Liu, W., Etschmann, B., Foran, G., Shelley, M., and Brugger, J., 2007. Deriving formation

constants for aqueous metal complexes from XANES spectra: Zn2+ and Fe2+ chloride

complexes in hypersaline solutions. American Mineralogist 92, 761-770.




2948.

Liu, X., Lu, X., Wang, R., and Meijer, E. J., 2011. Understanding hydration of Zn2+ in

hydrothermal fluids with ab initio molecular dynamics. Physical Chemistry


Liu, X., Lu, X., Wang, R., and Zhou, H., 2012. Silver speciation in chloride-containing



Martell, A. E. and Smith, R. M., 1976. Critical stability constants, Vol. 4: Inorganic

complexes. Plenum Press.

Mayanovic, R. A., Anderson, A. J., Bassett, W. A., and Chou, I., 1999. XAFS

measurements on zinc chloride aqueous solutions from ambient to supercritical

conditions using the diamond anvil cell. Journal of Synchrotron Radiation 6, 195-

197.




Mei, Y., Sherman, D. M., Liu, W., and Brugger, J., 2013b. Complexation of gold in S3--rich



Nelder, J. A. and Mead, R., 1965. A simplex method for function minimization. The

computer journal 7, 308-313.

Pankratz, L. B., 1982. Thermodynamic properties of elements and oxides. Volume 672 of

Bulletin United States. Bureau of Mines.

Perdew, J. P., Burke, K., and Ernzerhof, M., 1996. Generalized Gradient Approximation

Made Simple. Physical Review Letters 77, 3865-3868.




112


Acta 106, 501-523.

Preis, W., Königsberger, E., and Gamsjäger, H., 2000. Solid-solute phase equilibria in

aqueous solution. XII. Solubility and thermal decomposition of smithsonite. Journal

of solution chemistry 29, 605-618.

Resat, H. M., Mihaly, 1993. Studies on free energy calculations. I. Thermodynamic

integration using a polynomial path. The Journal of Chemical Physics 99, 6052-6061.

Ruaya, J. and Seward, T., 1986. The stability of chlorozinc (II) complexes in hydrothermal

solutions up to 350 C. Geochimica et Cosmochimica Acta 50, 651-661.

Ruaya, J. R. and Seward, T. M., 1987. The ion-pair constant and other thermodynamic

properties of HCl up to 350°C. Geochim. Cosmochim. Acta 51, 121-130.

Ryzhenko, B., Bryzgalin, O., Artamkina, I. Y., Spasennykh, M. Y., and Shapkin, A., 1985.

An electrostatic model for the electrolytic dissociation of inorganic substances

dissolved in water. Geochem. Intern 22, 138-144.











Shvarov, Y. V. and Bastrakov, E., 1999. HCh: a software package for geochemical

equilibrium modelling. User’s Guide Record/25 (Australian Geological Survey

Organisation, Canberra, 1999).



Smith, R. M. and Martell, A. E., 1976. Critical stability constants, Vol. 4: inorganic

complexes. Plenum Press.

Sprik, M., 1998. Coordination numbers as reaction coordinates in constrained molecular

dynamics. Faraday Discussions 110, 437-445.


113

Sprik, M. and Ciccotti, G., 1998. Free energy from constrained molecular dynamics.

Journal of Chemical Physics 109, 7737-7744

Sverjensky, D., Shock, E., and Helgeson, H., 1997. Prediction of the thermodynamic

properties of aqueous metal complexes to 1000 °C and 5 kb. Geochimica et


Tagirov, B. R., Zotov, A. V., and Akinfiev, N. N., 1997. Experimental study of dissociation

of HCl from 350 to 500 °C and from 500 to 2500 bars: Thermodynamic properties

of HCl°(aq). Geochimica et Cosmochimica Acta 61, 4267-4280.

Tossell, J. A., 1991. Calculations of the structures, stabilities, and Raman and zinc NMR

spectra of zinc aqua chloride ZnCln(OH2)a2-n species in aqueous solution. The

Journal of Physical Chemistry 95, 366-371.

Wesolowski, D. J., Bénézeth, P., and Palmer, D. A., 1998. ZnO Solubility and Zn2

Complexation by Chloride and Sulfate in Acidic Solutions to 290 °C with In-Situ pH

Measurement. Geochimica et Cosmochimica Acta 62, 971-984.

Yardley, B.W.D. (2005) Metal concentrations in crustal fluids and their relationship to ore

formation. Economic Geology, 100, 613-632.

Zotov, A. V., Kudrin, A. V., Levin, K. A., Shikina, N. D., and Var’yash, L. N., 1995.

Experimental studies of the solubility and complexing of selected ore elements (Au,

Ag, Cu, Mo, As, Sb, Hg) in aqueous solutions, Fluids in the Crust. Springer.


114

115

_____________________________________

Chapter 5

Metal complexation and ion hydration in low density

hydrothermal fluids:

ab initio molecular dynamics simulation of Cu(I) and Au(I)

in chloride solutions (25-1000 °C, 1-5000 bar)

Yuan Mei1, 2,*, Weihua Liu2, David M. Sherman3 and Joël Brugger1,4,*






_____________________________________

116

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4(&'#..(5/#"/(&'#&6'%$"#2'5(%02$7#"/(&' /&' 2(8'6$&./"9',96-(",$-%#2' )2+/6.:'#;' /&/"/('%(2$5+2#-'

69&#%/5.'./%+2#"/(&'()'<+=4>'#&6'*+=4>'/&'5,2(-/6$'.(2+"/(&.'=?@ABCCC'D<E'BA@CCC';#->'

3+;2/5#"/(&'!"#"+.''

'

3+;2/5#"/(&'F$"#/2.' G$/E'HIE'J/+E'KIE'!,$-%#&E'FI'GIE'#&6'L-+MM$-E'NIE' 4(&'#..(5/#"/(&'#&6'%$"#2'5(%02$7#"/(&' /&'2(8' 6$&./"9' ,96-(",$-%#2' )2+/6.:' #;' /&/"/('%(2$5+2#-' 69&#%/5.' ./%+2#"/(&' ()'<+=4>' #&6' *+=4>' /&'5,2(-/6$'.(2+"/(&.' =?@ABCCC'D<E'BA@CCC';#->E'!+;%/""$6' "('O$(5,/%/5#'$"'<(.%(5,/%/5#'*5"#E'+&6$-'-$P/$8''

'

!"#$%&'(%)#&*+"#*%),'L9' ./M&/&M' ",$' !"#"$%$&"' ()' *+",(-.,/0E' $#5,' #+",(-' 5$-"/)/$.' ",#"' ",$/-' ."#"$6' 5(&"-/;+"/(&' "(' ",$' 0+;2/5#"/(&' /.' #55+-#"$' #&6' ",#"'0$-%/../(&'/.'M-#&"$6')(-'",$'0+;2/5#"/(&'"(';$'/&52+6$6'/&'",$'5#&6/6#"$Q.'",$./.I'

R#%$'()'<(A*+",(-' H+#&'G$/'

<(&"-/;+"/(&'"('",$'3#0$-' F$P$2(0$6'5(%0+"#"/(&#2'%$",(6.'#&6'5(&6+5"$6'#22'%(2$5+2#-'./%+2#"/(&.I'L+/2"'M$(5,$%/5#2'%(6$2'#&6'8-("$'%#&+.5-/0"I'

!/M&#"+-$'

'

F#"$' ?SAN+2ABT'

'

R#%$'()'<(A*+",(-' K$/,+#'J/+'

<(&"-/;+"/(&'"('",$'3#0$-' *../."$6'8/",';+/26/&M'M$(5,$%/5#2'%(6$2'#&6'%#&+.5-/0"'$P#2+#"/(&I'

!/M&#"+-$'

'

F#"$' ?SAN+2ABT'

'

R#%$'()'<(A*+",(-' F#P/6'G'!,$-%#&'

<(&"-/;+"/(&'"('",$'3#0$-' *../."$6'8/",'6$P$2(0/&M'5(%0+"#"/(&#2'%$",(6.'#&6'%#&+.5-/0"'$P#2+#"/(&I'

'

!/M&#"+-$'

'

F#"$' ?SAN+2ABT'

'

R#%$'()'<(A*+",(-' N(U2'L-+MM$-'

'<(&"-/;+"/(&'"('",$'3#0$-' *../."$6'8/",';+/26/&M'M$(5,$%/5#2'%(6$2'#&6'%#&+.5-/0"'$P#2+#"/(&I'*5"/&M'#.'5((-$.0(&6/&M'#+",(-I''

!/M&#"+-$'

'

F#"$' ?SAN+2ABT'

CHAPTER 5. CHARGE OR NEUTRAL COMPLEXES

117

Metal complexation and ion hydration in low density

hydrothermal fluids:

ab initio molecular dynamics simulation of Cu(I) and

Au(I) in chloride solutions (25-1000 °C, 1-5000 bar)

Yuan Meia,b, Weihua Liub, David M Shermanc and Joël Bruggera,d

a Tectonics, Resources and Exploration (TRaX), School of Earth and Environmental

Sciences, The University of Adelaide, Adelaide, SA 5005, Australia

b CSIRO Earth Science and Resource Engineering, Clayton, VIC 3168, Australia

c School of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK

d South Australian Museum, North Terrace, SA 5000, Australia

*Corresponding author: [email protected]

US English


118

5.0 Abstract Low-density supercritical fluids are suspected of being able to transport metals, but it is

unclear what the speciation/complexation would be in such conditions. In this work, we

used ab initio molecular dynamics simulations to investigate the complexation and ion

association of Au+ and Cu+ in NaCl fluids as a function of solution density, from ambient to

supercritical conditions (to 1000 ˚C, 5000 bar). Cu(I) and Au(I) form distorted linear

complexes with two chloride ligands (i.e., CuCl2– and AuCl2

–) in subcritical chloride brines.

We have discovered that these charged complexes remain in high density supercritical fluids

even at high temperature; however, with decreasing density, these complexes become

progressively neutralized by ion association with Na+ to form low-charge (NanCuCl2)n-1 and

(NanAuCl2)n-1 complexes. In these species, the Na+ ion is very weakly bonded in the outer

coordination sphere, resulting in highly disordered structures and fast (few picoseconds

comparing to exchange rate of ns to "s for the aqua- and chloride complexes of transitions

metals) exchange among coordinated and solvent Na+ ions. Yet, thermodynamic models to

predict the solubility of metals in low-density magmatic or metamorphic fluids must

account for these species. In addition, we found that the number of water molecules (i.e., the

hydration number) surrounding the Cu+, Au+, Na+ and Cl– ions decreases linearly when fluid

density decreases, supporting empirical thermodynamic models that correlate the stability

constants of complexation reactions with solvent density. The traditional Born-Model

description explains the ion association as resulting from the decreased dielectric constant of

the solvent. However at a molecular level, the increase in translational entropy associated

with ion dehydration is the main contributor to changes of the solvation properties of

aqueous fluids at high P, T.

Keywords: molecular dynamics, hydrothermal fluids, coordination spheres, hydration,

aqueous metal complexes, translational entropy, gold and copper deposits.


119

5.1 Introduction

5.1.1 Controls on metal solubility in hydrothermal fluids

Investigating metal speciation and solubility of minerals in hydrothermal fluids is important

for understanding the mobility of metals in geological fluids (e.g., leading to the formation

of hydrothermal ore deposits) and in engineered systems (e.g., geothermal production wells;

hydrometallurgy; corrosion in power plants). To accurately model the dissolution, transport

and deposition of metals from hydrothermal fluids, one needs to know the identity,

stoichiometry and thermodynamic properties of aqueous metals species under wide ranges

of temperature, pressure and fluid composition. In popular numerical modeling packages,

changes in free energy of the complexes are calculated using semi-empirical equations,

based on electrostatic theory and empirical correlations with bulk solvent properties such as

density and Dielectric Constant (DC; e.g., Debye and Hückel 1923; Helgeson and Kirkham,

1974). For metal transport under hydrothermal conditions, much of our understanding relies

on mineral solubility data, conducted under limited ranges of conditions, and providing

mostly a macroscopic understanding. Over the past 30 years, these data have been

complemented by a growing number of in situ spectroscopic studies, in particular X-ray

absorption spectroscopy (XAS), that provide us with a molecular-level understanding of

metal speciation over a wide range of conditions (reviews in Brugger et al. 2010; Seward

and Driesner 2004). In addition, with the advance of supercomputing technology and

computational chemistry, molecular dynamics simulations are also providing us with a

deeper insight into the molecular-level structures, dynamics, and energetics of fluids and

solutes (e.g., Driesner, 1998; Liu et al., 2011b; Mei et al., 2013a,b; Sherman, 2007, 2010).

The description of solution chemistry and mineral solubility in hydrothermal systems relies

on the concept of aqueous metal complexes and their coordination structure. A coordination


120

complex (also called ‘inner sphere complex’, or ‘contact ion pair’) consists of an ion

surrounded by bound molecules known as ligands (e.g., Co(II)Cl42-). In general, a central

atom and its ligand are bonded by covalent bonds. Depending on the nature and strength of

these bonds, coordination complexes can have more or less rigid geometries. For example,

for divalent transition metals, UV-Vis and XAS studies have shown that with increasing

temperature, the metal chloride complexes show a transition from octahedral to tetrahedral-

like structures, with a reduction in the number of coordinated waters and an increase in the

number of halide ligands (e.g., Bazarkina et al. 2010; Hoffman et al., 1999; Liu et al., 2007;

2011a; 2012a; Susak and Crerar, 1985; Testemale et al., 2009; Tian et al., 2012). An outer-

sphere complex (also known as ‘solvent-shared ion-pair’) denotes a complex between a

solvated complex and an anion or cation (e.g., [Co(III)(NH3)6]Cl3(aq), where 6 NH3 are

bonded in the inner sphere and 3 Cl- in the outer sphere; House 2012). The interactions

between the central atom and the ions in the outer sphere are primarily electrostatic.

Outer-sphere complexes are often difficult to characterize experimentally under

hydrothermal conditions, but the formation of outer-sphere complexes is one of the major

factors affecting metal solubilities with increasing temperature and decrease in fluid

densities. In general, the formation of neutral complexes at high temperature or low fluid

density (e.g., vapor) is explained in terms of low DC of the solvent (e.g., Barnes 1997). In

fluids with low DC, charged species become rather unstable, and the formation of outer-

sphere complexes (e.g., CuCl2– + Na+ = CuCl2Na(aq)) provides a mechanism that can

account for high metal contents in low DC hydrothermal fluids (Zajacz et al., 2011).

Water itself can act as a ligand, forming aqua ions and coordination complexes (e.g.,

octahedral Co(H2O)62+; tetrahedral CoCl2(H2O)2(aq); Liu et al. 2011a). Coordination

complexes are associated through hydrogen bonding with other water molecules in a

secondary solvation shell. This second solvation shell is usually highly disordered, but it can


121

play a first order role in controlling mineral solubility. For example, the XAS study by Liu

et al. (2008) showed that the solubility of CuCl(s) (nantokite) near the critical density of

water at 420 °C, 290-400 bar is explained with a [H2O-Cu-Cl](aq) complex hydrated by

approximately two additional water molecules in the second coordination shell. In water

vapor, the fugacity of water (

!

f H2O) appears to control metal transport and mineral solubility

(Archibald et al., 2001; 2002; Williams-Jones et al., 2002); hence the hydration of ions is

important in understanding the metal speciation and transport for both vapors and

supercritical fluids.

5.1.2 Complexing and hydration of Cu(I) and Au(I) in hydrothermal brines

and vapors

The Cu(I) and Au(I) ions share a d10 electronic structure, and linear halide and hydrosulfide

complexes of these two ions prevail over a wide range of temperature, pressure and salinity.

For Cu(I)-Cl complexes, early solubility studies (Crerar et al., 1978; Hemley et al., 1992;

Seyfried and Ding 1993) suggested a neutral CuCl(aq); however in situ XAS measurements

of natural and synthetic fluid inclusions (Berry et al., 2006; Mavrogenes et al., 2002) and

model solutions (Brugger et al., 2007; Fulton et al., 2000a,b), backed up by molecular

dynamic simulations (Mei et al., 2013a; Sherman, 2007), all pointed to linear species

(CuCl2–) being predominant up to 600 °C. The CuCl(aq) complex also has a linear structure

([H2O-Cu-Cl](aq)), with an oxygen atom from a water molecule in the first shell (Fulton et

al., 2000a,b; Liu et al. 2008; Sherman 2007). Similarly, the predominant Au species in

chloride solutions at elevated temperatures also has a linear structure (AuCl2–; Pokrovski et

al., 2009; Stefánsson and Seward, 2003).

In the past decade, a considerable number of experiments have been carried out to

investigate Cu(I) and Au(I) solubility in water vapor, and partitioning between


122

liquid/vapor/melt (e.g., Archibald et al., 2001, 2002; Etschmann et al., 2010; Frank et al.

2011; Lerchbaumer and Audétat 2012; Liu et al., 2008; Pokrovski et al., 2005; Rempel et

al., 2012; Simon et al., 2005, 2006; Zajacz et al., 2010, 2011). In particular, a few studies

have investigated the identity and hydration of Cu(I) and Au(I) species in the vapor and low

density supercritical phases. Archibald et al. (2001) conducted Au solubility experiments in

subcritical, HCl-bearing water vapor at 300-360 °C and 144 bar, and suggested five

(300 °C) and three (360 °C) water molecules around AuCl(g). A sister study of Cu(I)

solubility (Archibald et al., 2002) found 7.6 and 6.1 waters around a Cu3Cl3(aq) cluster at

320 and 280 °C, respectively. Liu et al. (2008) concluded that the solubility of nantokite

(CuCl(s)) in supercritical water across the critical isochore (density of 0.19–0.42 g/cm3) at

420 °C, 290-400 bar, is explained by a [H2O-Cu-Cl](aq) complex with two extra water

molecules in the second shell. Zezin et al. (2007; 2011) suggested AuS(H2O)n in H2S-water

with n = 2.3 at 300 °C and 2.2 at 365 °C. Zajacz et al., (2010; 2011) conducted solubility

experiments to investigate Au and Cu transport in magmatic volatile phases (1000 °C,

1500 bar). They suggested on the basis of solubility measurements backed by static

quantum chemical calculations that Au(I) and Cu(I) are transported as neutral species in

high T, low density supercritical-magmatic fluids, due to the formation of outer-sphere

complexes with alkali ions (e.g., neutral complexes NaAuCl2(aq), NaAu(HS)2(aq) and

NaCuCl2(aq)). Such outer sphere alkali complexing increased the solubility of Au and Cu in

high-T chloride and/or H2S bearing vapors by a factor of up to over an order of magnitude

compared to that in alkali-chloride free fluids (Zajacz et al., 2010; 2011).

5.1.3 Computational chemistry studies of metal speciation

A few previous studies have used classical MD to predict the physical and chemical

properties of NaCl fluids, such as densities, ion pairing and ion hydration, conductance,


123

phase diagram, and DC (Chialvo and Simonson, 2003; Cui and Harris, 1995; Driesner et al.,

1998; Giberti et al., 2013; Hassan, 2011; John, 1998; Lee et al., 1998; Oelkers and

Helgeson, 1993; Sherman and Collings, 2002; Smith and Dang, 1994; Zahn, 2004). In

classical MD, the atomic interactions are described using empirical inter-atomic potentials

(e.g., Lennard-Jones potentials; Verlet, 1967). Classical MD has been applied with great

success to predict the properties of solutions of alkali and alkaline earth metals, for which

electrostatic interactions are predominant. For example, Driesner et al. (1998) investigated

the change of ion pairing, hydration and bond distances as functions of temperature,

pressure and fluid density in NaCl solutions from ambient to supercritical conditions.

However, to investigate complexation and hydration of transition metals such as Cu(I) and

Au(I), interatomic interactions cannot be accurately accounted for by simple pairwise

models (Sherman, 2010). One option is to use ab initio MD, in which atomic interactions

are treated following the principles of quantum mechanics. Density functional theory (DFT)

is a popular approximation method for solving the Kohn-Sham equation. For Cu(I) and

Au(I), recent studies have shown that static DFT calculations provide good agreement with

experiments for the structure (geometry and bond distances) of Cu(I) and Au(I) chloride

complexes (Pokrovski et al. 2009; Zajacz et al. 2011). However, such “static” quantum

chemical calculations are conducted in the ideal gas phase, and are not able to sample the

configuration degrees of freedom as required in simulation of condensed fluids as a function

of pressure and temperature. In the present study, we used the ab initio Car-Parrinello MD

method, which solves the classical equations of motion as in classical MD, but treats the

inter-atomic interactions quantum mechanically using DFT (Car and Parrinello, 1985). In

the past few years, ab initio MD has been employed to predict the aqueous speciation and

hydration number of transition metals with d10 electronic structure; for example, Zn(II)-Cl

(Harris et al., 2003), Cu(I)-Cl (Sherman, 2007), Au(I)-HS (Liu et al., 2011b), Ag(I)-Cl (Liu


124

et al., 2012b; Pokrovski et al., 2013), Cu-HS-Cl (Mei et al., 2013a), Au(I)-HS/OH/S3 (Mei

et al., 2013b). These studies report excellent agreement between the predicted nature and

geometry of the complexes and available experimental studies. In a breakthrough study, Mei

et al. (2013a) also demonstrated that thermodynamic properties for important Cu(I) chloride

and bisulfide complexes can be derived from ab initio MD simulations; the properties

compare well with the experimental ones, and, most importantly, enable accurate

predictions of the mineral solubility gradients as a function of P, T and solution

composition. Altogether, these recent studies show that ab initio MD methods have come of

age for the simulation of transition metals in aqueous fluids over a wide P, T range.

Accurate MD simulations provide a unique tool to address the lack of systematic

investigation of the dependence of complexation and hydration of these species on

macroscopic parameters such as fluid density, over conditions varying from liquid-like to

vapor-like. The MD results allow us to explore the molecular-level mechanisms affecting

Cu(I) and Au(I) chloride complexing and solvation as a function of changes in temperature

and solution density.

5.1.4 Aim of study

This paper uses ab initio molecular dynamics (MD) simulations to quantify the effects of P

and T (as the main parameters affecting solution density and DC of the solvent) on the

complexation and hydration of metal complexes. We chose Cu(I) and Au(I) chloride

complexes because of the large amount of experimental and MD studies available on these

systems; the high level of agreement between experimental and MD results demonstrates

the accuracy of the MD methods used in this study. In addition, Cu(I) and Au(I) form inner

sphere [MCl2–] coordination complexes that are stable over wide ranges in P, T and solution

composition, providing an ideal model system to study the effect of the complexation and


125

hydration within the outer coordination sphere as a function of T and P, in the liquid phase

as well as in the supercritical phase, with solution densities ranging from liquid-like

(1.2 g/cm3) to vapor-like (0.1 g/cm3). Specifically, we want to investigate 1) if and under

what conditions the linear CuCl2– and AuCl2

– chloride complexes are actually paired with

another cation (e.g., Na+) to form a neutral outer sphere complex; and 2) if and how the

hydration numbers of Au(I) and Cu(I) (as well as Na+ and Cl–) change with fluid density

and temperature.

5.2 Method: Ab initio molecular dynamics simulations

5.2.1 Computational methods

Ab initio MD simulations of Cu(I) and Au(I) complexes at different densities (i.e., P, T

conditions) were performed using the Car-Parrinello MD code “CPMD” (Car and

Parrinello, 1985), which implements DFT using a plane-wave basis set and pseudo-

potentials for the core electrons plus the nucleus. The PBE exchange correlation-functional

(Perdew et al., 1996) was employed with a cutoff of gradient correction 5$10-5. Plane-wave

cutoffs of 25 Ry (340.14 eV) were used together with Vanderbilt ultrasoft pseudo-potentials

in CPMD package (Laasonen et al., 1993). A time-step of 3 a.u. (0.073 fs) was used to

stabilize the simulations. The MD simulations were conducted in an NVT ensemble (i.e.

constant composition, volume and temperature). Temperatures were controlled by a Nosé

thermostat for both the ions and electrons. Most simulations contained four Cl–, one Au+ or

one Cu+, and three Na+ atoms to balance the charge in the solution. The pressures of the

simulated systems were evaluated using the equation-of-state of NaCl solutions as

implemented in the SOWAT code (Driesner, 2007; Driesner and Heinrich, 2007) based on a

total equivalent NaCl salinity of 4 molal. To obtain time averages of the geometric and

stoichiometric information, radial distribution functions (RDF) and their integrals (related to


126

coordination number) for different atom pairs were calculated by VMD (Humphrey et al,

1996).

Although the formal Cu or Au concentrations are one molal, the simulations represent dilute

solutions with respect to Cu or Au (no Cu-Cu or Au-Au interaction), because there is only

one Cu or Au atom in each box. Most of the simulation systems listed in Tables 5.1 and 5.2

are under-saturated with respect to NaCl(s), except for some low-density fluids for which

the NaCl concentrations exceed the solubility limit. For instance, the solubility of NaCl(s)

reaches only 1.04 molal at 1000 °C, 1500 bar. In theory, the chloride concentration needs to

be diluted by adding more water in the simulation box to represent the real fluid. But in

practice, as discussed by Mei et al. (2013a), increasing the number of particles increases the

simulation time dramatically, causing simulations to become prohibitively costly in terms of

computing time.

Table 5.1 Simulation details and geometrical properties of Cu-Cl complexes in a box

containing 1 Cu+, 3 Na+, 4 Cl- and 55 H2O. Time step is 3 a.u. (0.0726 fs). All the

simulations were preformed for more than 29 picoseconds (400,000 steps)

T (˚C)

P (bar)*

Box size (Å)

Density (g/cm3)

Simulation time (ps) Stoichiometry¶

dCu-Cl (Å)

dNa-O

(Å) dCl-O (Å)

Cl-Cu-Cl angle(˚)

25 1 12.110 1.18 42.81 [CuCl3]–2 2.21 2.38 3.16 -

300 500 13.034 0.95 29.86 [CuCl2Na0.26]–0.74 2.11 2.35 3.22 164.6(7.7)

300 1000 12.886 0.98 29.02 [CuCl2Na0.13]–0.87 2.12 2.34 3.21 163.5(9.9)

300 2000 12.683 1.03 29.02 [CuCl2Na0.09]–0.91 2.12 2.35 3.24 164.7(8.4)

500 600 14.658 0.67 29.02 [CuCl2Na0.25]–0.75 2.12 2.35 3.26 160.3(11.7)

500 1000 14.030 0.76 29.02 [CuCl2Na0.28]–0.72 2.12 2.34 3.26 161.2(10.1)

500 2000 13.434 0.87 29.86 [CuCl2Na0.16]–0.84 2.11 2.35 3.24 161.1(11.3)

1000 1500 19.290 0.29 29.75 [CuCl2Na0.68]–0.32 2.11 2.35 3.32 153.6(14.9)

1000 2500 15.609 0.55 29.60 [CuCl2Na0.31]–0.69 2.14 2.33 3.34 157.0(12.3)

1000 5000 14.128 0.74 34.32 [CuCl2Na0.41]–0.59 2.13 2.33 3.31 157.1(13.7)

* Evaluated from the equation-of-state of NaCl fluids at the same ionic strength using the SoWat code (Driesner, 2007; Driesner and Heinrich, 2007).

¶ Based on the number of Na atoms within 4 Å of a Cu atom.


127

In this study, an ab initio MD simulation of the Cu-Cl system with 173 particles (55 H2O,

1 Cu+, 3 Na+, 4 Cl–, reflecting [Cl–] of 4 molal) at a density of 0.29 g/cm3 (box size 19.29 Å)

required ~750 CPU hours per picosecond, while a simulation involving 341 particles (111

H2O, 1 Cu+, 3 Na+, 4 Cl–, reflecting [Cl–] of 2 molal) cost more than 3,500 CPU hours for

one picosecond.

Table 5.2 Simulation details and geometrical properties of Au-Cl complexes in a box with

1 Au+, 3 Na+, 4 Cl- and 55 H2O. Time step is 3 a.u. (0.0726 fs). All the simulations were

preformed for more than 29 picoseconds (400,000 steps)

T (˚C)

P (bar)*

Box size (Å)

Density (g/cm3)

Simulation time (ps) Stoichiometry¶

dAu-Cl (Å)

dNa-O

(Å) dCl-O (Å)

Cl-Au-Cl angle(˚)

300 500 13.034 1.05 32.65 [AuCl2Na0.21]–0.79 2.28 2.35 3.25 169.5(5.4)

500 600 14.658 0.74 32.43 [AuCl2Na0.29]–0.71 2.27 2.35 3.27 166.7(7.2)

500 2000 13.434 0.96 32.65 [AuCl2Na0.24]–0.76 2.29 2.35 3.27 167.1(7.1)

1000 1500 19.290 0.32 31.93 [AuCl2Na0.50]–0.50 2.32 2.34 3.32 159.2(10.7)

1000 2500 15.609 0.61 33.38 [AuCl2Na0.30]–0.70 2.30 2.33 3.32 162.9(8.8)

1000 5000 14.128 0.82 32.65 [AuCl2Na0.32]–0.68 2.30 2.34 3.32 163.8(9.6)

* Evaluated from the equation-of-state for NaCl fluids at the same ionic strength using the SoWat code (Driesner, 2007; Driesner and Heinrich, 2007).

¶ Based on the number of Na atoms within 4 Å of a Au atom.

Two simulations at the same density (Cu-Cl system at the density of 0.29 g/cm3) but with

different numbers of water molecules (55 vs. 111) were conducted to test the impact of box

size on the results. The Cu-Na and Cl-Na pair distribution functions (Fig. 5.1a) show

different intensities because in the simulation with the larger box, the Cu+, Cl– and Na+ ions

are diluted compared with the MD of the smaller box. However, the RDF integrals showing

the number of Na+ surrounding the Cu+/Cl– ions for both simulations overlap, indicating a

small effect of the box size and dilution of the salt concentration from 4 to 2 molal.

Similarly, the hydration numbers of Na+ and Cl– in the simulations with different box sizes

agree well (Fig. 5.1b). Since the number of oxygens is about the same in a unit box (same


128

solution density), the interactions of water with Cl- and Na+ are expected to be independent

from the box size. The slight differences in the integrals of RDF result in calculated

hydration number that are still very close (error <0.2) within a distance of 3 Å for Na-O and

3.6 Å for Cl-O (see 5.3.3 for choosing distance cut-off); these differences represent the

magnitude of the uncertainty for the hydration number in our calculations, related mainly to

the finite calculation time. Since the simulation of the smaller box is able to represent the

features in the bigger box, we conducted simulations with 55 H2O. To observe the behavior

of the charge-balanced ions (Na+) in the second-shell, long runs (> 20 ps) are necessary,

since the atoms in the second-shell are more disordered than those in the first-shell.

Figure 5.1 Small box vs. large box: the differences on RDF (left axes, solid curves) and

coordination number (right axes, dashed curves).

4

3

2

1

0

RD

F

54321Pair distance (Å)

8

7

6

5

4

3

2

1

0

Hydration num

ber

Na-O Cl-O

(b)

80

60

40

20

0

RD

F

654321

5

4

3

2

1

0

Coordination num

ber

Cu-Na Cl-Na

thick lines: small box (1Cu + 3Na + 4Cl + 55H2O) thin lines: large box (1Cu + 3Na + 4Cl + 111H2O)

(a)


129

5.2.2 Choice of model systems for comparison with experimental studies

In addition to the simulations aimed at exploring systematically the effect of solution

density on the complexation and hydration of Au(I) and Cu(I) chlorocomplexes (Tables 5.1

and 5.2), five simulations were conducted at different T, P corresponding to specific

experimental studies for direct comparison (Table 5.3).

Table 5.3 Simulation details for runs aimed at comparing with experimental studies.

No. Simulation box T (˚C)

P (bar)‡

Box size (Å)

Density (g/cm3) Stoichiometry dCu/Au-Cl

(Å) dCu/Au-O

(Å) Angle

(˚)

1* 1Cu, 3Na, 4Cl, 55H2O 1000 1500 19.290 0.29 [CuCl2Na0.7]–0.3 2.11 - 155.4(13.4) 2¶ 1Cu, 1Cl, 55H2O 420 290 20.830 0.20 CuCl(H2O)0 2.09 1.93 164.1(8.6) 3¶ 1Cu, 1Cl, 55H2O 420 400 15.808 0.46 CuCl(H2O)0 2.10 1.94 163.1(8.8)

4§ 1Au, 1Cl, 55H2O 340 139§ 27.186 0.10' AuCl(H2O)0 2.23 2.11 168.9(6.0)

5** 1Au, 1Cl, 55H2O 340 139 13.694 0.79 AuCl(H2O)0 2.24 2.08 168.9(6.1)

Reference: *Zajacz et al., 2011; ¶Liu et al., 2008; §Archibald et al., 2001, vapor phase; **Archibald et al., 2001, liquid phase ‡Pressures evaluated from equation-of-state of NaCl fluids at the same ionic strength using the SoWat code (Driesner, 2007; Driesner and Heinrich, 2007). ' Evaluated from the equation-of-state of the vapor phase using the pure water density of 0.082 g/cm3; the density of the AuCl solution is 0.10 g/cm3.

MD Simulation of Cu(I) chloride complexes at a density of 0.29 g/cm3 at 1000 °C,

corresponding to a pressure of 1500 bar (No. 1) were performed to compare with the

experimental study by Zajacz et al. (2011). Simulations 2 and 3 were conducted at 420 °C

and pressures of 290 and 400 bar for comparison with the nantokite solubility, XANES and

EXAFS data of Liu et al. (2008). The fluid densities shown in Table 5.3 are slightly higher

than those listed by Liu et al. (2008), because the MD simulation were conducted at higher

salt concentrations (1 molal Cu and 1 molal Cl) compared with real fluids. It is also notable

that as the simulation box contained only one Cu and one Cl atoms, Cu-Cu and Cl-Cl

interactions are negligible, reflecting the relatively dilute nature of the solutions studied by

Liu et al. (2008). Simulation 4 was performed to compare with the experimental study by

Archibald et al. (2001) of Au chloride solubility in water vapor (340 °C, 139 bar); a solution


130

density of 0.10 g/cm3 was chosen by using the pure water density of 0.082 g/cm3 and adding

one Au and Cl atoms to the simulation box (Table 5.3, No. 4). Simulation 5 was conducted

to represent the liquid phase at the same temperature of 340 °C; the density was chosen at

0.79 g/cm3 according to equation-of-state of NaCl fluids at the same ionic strength

(Driesner, 2007; Driesner and Heinrich, 2007). All these simulations were conducted for

more than 29 ps (400,000 simulation steps).

5.3 Results

5.3.1 Cu(I) and Au(I) complexing as a function of solution density and

temperature

MD simulation of Cu-Cl (Table 5.1) at 25 °C gave a three-fold triangular planar complex

CuCl32– with the Cu-Cl distances of 2.21 Å, in accordance with previous experimental

(Brugger et al., 2007; Liu et al., 2008) and theoretical studies (Sherman 2007; Mei et al.,

2013a). At elevated T and P, two-fold distorted linear complexes MCl2– (M=Cu/Au) were

found in simulations of Cu-Cl and Au-Cl solutions, with the Cu-Cl distances of 2.11-2.14 Å

(Table 5.1) and Au-Cl distances of 2.27-2.32 Å (Table 5.2), in good agreement with

previous experimental studies: Cu-Cl distance of 2.152(7) Å by Brugger et al. (2007) and

2.12-2.13 Å by Fulton et al. (2000a,b); Au-Cl distance of 2.27-2.28 Å by Pokrovski et al.

(2009).

The configurations of MCl2– (M=Cu/Au) complexes with surrounding Na+/Cl– ions were

investigated along the simulation time at different P, T conditions to test if and how the

MCl2– complexes form neutral species as a function of fluid density and temperature. The

results show that charged species (i.e., MCl2–) predominate in high-density fluids (

≥ 0.7 g/cm3), even in high-temperature supercritical fluids. In low density fluids, however,


131

the probability of forming the neutral species MNaCl2(aq) and larger clusters such as

MNanClm1+n-m increases. Selected snapshots of the Cu-Cl simulation at 1000 °C, 1500 bar

(density 0.29 g/cm3, Table 5.1; Table 5.3, No. 1) reveal that the neutral CuNaCl2(aq) species

does not display a well-characterized geometry (e.g., configurations at 10.97 and 15.8 ps;

Figs. 5.2a,b). The simulation also occasionally shows charged species and larger clusters;

for example at 24.5 ps (Figs. 5.2c,d) the CuCl2– cluster is weakly linked to a NaCl(aq)

cluster. Over the next 0.7 ps, this cluster evolves into a CuNa2Cl3(aq) cluster (Figs. 2d-g). A

similar situation is found for the Na+ ions in this solution; the neutral cluster NaCl(aq) is

present at 10.97 ps (Fig. 5.2a), but at 15.8 ps the Na2Cl2(aq) cluster formed (Fig. 5.2b). At

24.5 ps, the simulation contains a Na2Cl+ cluster in addition to a NaCuCl2(aq) cluster.

Figure 5.2 Cu+ and Na+ local structures at 1000 ˚C, 1500 bar (0.29 g/cm3). (a-c) Selected

snapshots of simulations. (d-g) Example of dynamic transformation of Cu(I) ion pairs.

!"#!!!"#$%"&'(!$(!)*+,-!%"./0$.12!3$45!6!2!0$.1!3$45

!$#!!!"#$%"&'(!$(!)7+8!%"./0$.12!3$45!6!0$2.12!3$45

!%#!!!"#$%"&'(!$(!29+7!%":./.12;<=0$.1!3$45!6!0$2.16

!"

#$

#%

&'()*+,- &'(.)+,- &'(/0+,- &)(&1+,-!"# !$# !%# !&#


132

The dynamic distances between the Cu and Na atoms as a function of simulation time also

indicate the extent of Na+ pairing with CuCl2–. The evolutions of Cu-Na distances and

instantaneous coordination numbers for three simulations at three different T, P conditions,

with fluid densities ranging from 0.29 to 1.03 g/cm3 are shown in Figure 5.3.

Figure 5.3 Cu-Na distance vs simulation time (left) and instantaneous coordination number

(distance cutoff = 4 Å) together with running average as a red line (right).

At a density of 0.29 g/cm3 (1000 °C, 1500 bar; Figs. 5.3a,b), there is at least one Na+ ion at

a distance of ~3 to 5 Å from Cu+. The closest Na+ ions exchange at a very fast rate (longest

residence time < 15 ps) comparing with the exchange rates which lie in the ns to "s range

for the aqua- and chloride complexes of transitions metals (Table 11.4 in Burgess 1978;

Sharps et al., 1993). For short periods (< 3 ps), two or three Na+ ions get close to Cu+,

forming large clusters (e.g., Fig. 5.2g). The running average of instantaneous coordination

16

14

12

10

8

6

4

2

0

Cu-

Na

dist

ance

(Å)

302520151050

(c)

14

12

10

8

6

4

2

0302520151050


(e)

18

16

14

12

10

8

6

4

2

0302520151050

(a)3.0

2.5

2.0

1.5

1.0

0.5

0.0

302520151050

(b)

0.70

2.0

1.5

1.0

0.5

0.0

302520151050

(d)

0.26

2.0

1.5

1.0

0.5

0.0

302520151050


(f)

0.10

1000˚C, 1500 bar 0.29 g/cm

3

500˚C, 600 bar 0.67 g/cm

3

Instantaneous coordination number

300˚C, 2k bar 1.03 g/cm

3


133

number gave an average number of 0.70 Na+ surrounding Cu+ (within 4 Å). In contrast, at

the density of 0.67 g/cm3 (Figs. 5.3c,d), Cu+ and Na+ only bind together occasionally, with

an average of 0.26 Na+ within 4 Å of Cu+. In the simulation at 300 °C, 2000 bar (Figs. 3e,f),

with a density of 1.03 g/cm3, there is less chance for Na+ to complex to Cu+, which is also

confirmed by the instantaneous coordination number (0 is predominant, as shown in

Fig. 5.3f).

Figure 5.4a shows the Cu-Na RDF at different temperatures and densities (solid lines),

together with the integrals of the selected distribution functions, representing the number of

Na+ surrounding Cu+ (NNa(Cu)) (dashed lines). There is a clear trend that the height of the

RDF peak and NNa(Cu) decrease with increasing density. At a density of 0.29 g/cm3

(1000 °C, 1500 bar), there is a large peak between 2.5 and 4.5 Å, and NNa(Cu) is close to one

at the distance of ~4.5 Å. This shows the predominance of the neutral species (e.g.,

CuNaCl2(aq)), although the width of the RDF shows that Na+ is only weakly bound. The

distribution functions of Cu-Na agree with the distance plots shown in Figure 5.3b,

indicating there are more neutral complexes in low-density fluids and more charged

complexes in high-density fluids.

For Au+, the Au-Na RDF and their integrals indicate that an analogous neutral species,

NaAuCl2(aq), predominates at low density (0.32 g/cm3), with an average of 0.5 Na+ within

4 Å of Au+ (Table 5.2). The number of Na+ surrounding Au+ decreases to 0.21 with an

increase of density to 1.05 g/cm3.


134

Figure 5.4 Radial distribution functions and the integrals of Cu-Na, Cl-Na, Na-O, Cl-O

pairs at different solution densities (see Table 5.1 for temperatures and pressures).

In addition to the neutral MNaCl2(aq) (M=Cu,Au) complexes, the MD simulations also

show that Na+ and Cl– form a number of neutral ion pairs in low-density fluids (e.g.,

!"#

!"$

%"#

%"$

$"#

$"$

&'()*+,-

#"$."$/"$!"$%"$&'()*01234567*89:

%$

;

<

=

>

#

.

/

!

%

$

?@0A431B5*5CD

E7A*BF*&'

**$"!;*GH6D/

**$"##*GH6D/

**$">=*GH6D/

**$"=.*GH6D/

**$"=>*GH6D/

**$"<=*GH6D/

**$";#*GH6D/

**$";<*GH6D/

**%"$/*GH6D/

**%"%<*GH6D/

.$

/#

/$

!#

!$

%#

%$

#

$

&'(I4*+,-

#"$."$/"$!"$%"$&'(I4*01234567*89:

!"$

%"#

%"$

$"#

$"$

ICDE7A*BF*I

4*2CAABC5070*E@*&'*

***$"!;*GH6D/

***$"##*GH6D/

***$">=*GH6D/

***$"=.*GH6D/

***$"=>*GH6D/

***$"<=*GH6D/

***$";#*GH6D/

***$";<*GH6D/

***%"$/*GH6D/

***%"%<*GH6D/

>

#

.

/

!

%

$

I4()*+,-

."$/"$!"$%"$I4()*01234567*89:

%$

;

<

=

>

#

.

/

!

%

$

?@0A431B5*5CD

E7A*BF*I4

***$"!;*GH6D/

***$"##*GH6D/

***$">=*GH6D/

***$"=.*GH6D/

***$"=>*GH6D/

***$"<=*GH6D/

***$";#*GH6D/

***$";<*GH6D/

***%"$/*GH6D/

***%"%<*GH6D/

%!

%$

<

>

.

!

$

&C(I4*+,-

>#./!%

&C(I4*01234567*89:

!"$

%"#

%"$

$"#

$"$ICDE7A*BF*I

4*2CAABC5070*E@*&C

**$"!;*GH6D/

**$"##*GH6D/

**$">=*GH6D/

**$"=.*GH6D/

**$"=>*GH6D/

**$"<=*GH6D/

**$";#*GH6D/

**$";<*GH6D/

**%"$/*GH6D/

**%"%<*GH6D/

84:

8E:

86:

80:


135

Figs. 5.2a,b). Since there are 3 Na+ and 4 Cl– in the simulation box, it would be difficult to

show each Cl-Na distances. Here we calculating the RDF of Cl-Na pairs and the number of

Na+ surrounding Cl– (NNa(Cl)). The RDF peaks of both Cl-Na pairs decrease with an increase

in density, and NNa(Cl) decreases, too (Fig. 5.4b). At 1000 °C, 1500 bar, NNa(Cl) is ~1,

corresponding to the NaCl(aq) species shown in Figure 5.2. In contrast, the Cl-Na RDF peak

almost disappears at room temperature, consistent with the predominance of the Na+ and Cl–

aqua ions. The classical MD simulations of NaCl solutions of Driesner et al. (1998) showed

that aqua ions account for 63.3% Na or Cl at room temperature (density of 1.02 g/cm3); at

380 °C (density of 0.55 g/cm3) this proportion decreases to 12.8%, neutral ion pairs (NaCl0,

Na2Cl20, Na3Cl3

0, Na4Cl40) contribute 49.8% and charged clusters (Na2Cl+, NaCl2

–, Na3Cl2+,

Na2Cl3–, Na4Cl3

+, Na3Cl4–) contribute 37.4% to total solute species. Hence, our ab initio

calculations confirm the results of earlier classical calculations in emphasizing the

importance of the neutral NaCl(aq) species, together with short-lived, larger clusters for

understanding the properties of high temperature brines (Driesner et al., 1998).

To further illustrate the relationship between fluid density and the number of Na+

surrounding Cu(I) and Au(I) complexes (NNa(Cu), NNa(Au)), as well as Na-Cl ion pairing

(NNa(Cl)), quantitative values of NNa(X) were obtained from RDF plots. One problem in

defining NNa(X) is to choose an appropriate distance cutoff. The distance cutoff is usually set

at the shoulder of the integral of the RDF plot, and can vary slightly with temperature (e.g.,

Driesner et al., 1998). In this study, we chose to fix the distance cutoff for all temperatures,

to avoid bias in the temperature trends. A cutoff distance at 3 Å was set for the Cl-Na pair

according to Figure 5.4b, where the shoulder in Cl-Na RDF is clearly seen. For the Cu-Na

distribution function, the RDF peak is not very sharp and the data are affected by significant

noise, making it difficult to define the cutoff distance. Yet we need to find a safe cutoff

distance that includes the possible bonding Cu-Na pairs but excludes the random Na


136

movement and the impact of different box size. As shown in Figure 5.4a, the RDF peak

appears at ~3 Å for low densities and ~3.5 Å for high densities. A cutoff distance of 4 Å

was chosen to calculate both NNa(Cu) and NNa(Au).

Plots of number of Na+ surrounding Cu+ or Cl– as a function of density (Fig. 5.5a) show

good linear relationships, with correlation coefficients (R2) of 0.84 for NNa(Cu) and 0.90 for

NNa(Cl). A similar correlation holds for Au, with correlation coefficients of 0.87 for NNa(Au)

and 0.97 for NNa(Cl) (Fig. 5.5b). As the density of NaCl-H2O fluids is well constrained (e.g.,

equation-of-states by Duan et al., 1995; Driesner 2007; Driesner and Heinrich 2007), the

correlations versus solution density were significantly better than the correlation with the

DC of pure water.


137

Figure 5.5 The number of Na surrounding Cu/Cl (a) and Au/Cl (b), and the hydration

number of Na/Cl/Cu (c) and Na/Cl/Au (d) as function of solution density

!"#

!"$

!"%

!"&

!"!

'()*+,-./-'0-1(,,.(23425-6(768-9

:

;"&;";;"!!"<!"#!"=!"$!">!"%!"?

@+214AB-957C)?:

- DC-E-%"!-F--DC-E-?"!-F

!"#

!"=

!"$

!">

!"%

!"?

!"&

!";

'()*+,-./-'0-1(,,.(23425-G(768--9

-:

;";;"!!"<!"#!"=!"$!">!"%!"?@+214AB-957C)?:

- --DC-E-%"!-F- ---DC-E-?"!-F

#

=

$

>

%

?

&

;

HB3,0A4.2-2()*+,-./-'076876(

;"&;";;"!!"<!"#!"=!"$!">!"%!"?

@+214AB-957C)?:

-DC-E-?"!-F--DC-E-?"$-F-DC-E-%">-F

=

$

>

%

?

&

;

HB3,0A4.2-2()*+,-./-'07687G(

;";;"!!"<!"#!"=!"$!">!"%!"?@+214AB-957C)?:

- --DC-E-?"!-F- ---DC-E-?"$-F- --DC-E-%">-F

!"#

!$#

!%#

!&#


138

5.3.2 Hydration numbers of Na, Cl, Cu and Au

The number of water molecules surrounding Na (Hyd(Na)) and Cl (Hyd(Cl)) increases with

increasing solution density (Figs. 5.4c,d). For example, there are ~5 waters surrounding Na

at a density of 1.18 g/cm3, but the number decreases to ~2 at a density of 0.29 g/cm3

(Figs. 5.4c). The hydration numbers of Na+ and Cl– in Au-Cl solutions gave similar results.

To calculate the hydration of Na+ and Cl–, distances of 3 Å and 3.6 Å, respectively, were

chosen according to the RDF plot (Figs. 5.4c,d). The hydration numbers of Cu+ and Au+

were calculated with a cutoff distance of 4.5 Å. The calculated hydration numbers are

plotted as a function of solution density in Figs. 5.5(c,d). Good linear correlations between

the hydration number of Na+, Cl–, Cu+ and density were obtained. For Cu-Cl fluids, linear

correlation coefficient R2 of 0.97, 0.99 and 0.99 for Hyd(Na), Hyd(Cl) and Hyd(Cu) were

obtained (Fig. 5.5c); and Au-Cl solutions gave R2 of 0.95, 0.99 and 0.99 for Hyd(Na), Hyd(Cl)

and Hyd(Au) , respectively (Fig. 5.5d).

The simulations of 1 Cu, 1 Cl and 55 H2O at temperature of 420 °C and densities of 0.20

and 0.46 g/cm3 (No. 2,3 in Table 5.3) gave the predominant complex of CuCl(H2O)(aq),

consistent with the interpretation of the experiments (Liu et al., 2008). In the calculation of

the hydration number of Cu+ and Cl–, the first shell water in the CuCl(H2O)0 complex was

excluded. The hydration numbers of Cu+ and Cl– both increase with increasing density

(Fig. 5.6). Specifically, there are 2.86 water molecule surrounding CuCl(H2O)0 at a density

of 0.20 g/cm3 (400 °C, 290 bar, vapor-like fluids); this hydration number increases by ~2

(4.63 water molecules) at a density of 0.46 g/cm3 (400 °C, 400 bar). The hydration number

of Cl– increased from 0.92 (0.20 g/cm3) to 1.39 (0.46 g/cm3).


139

Figure 5.6 Instantaneous hydration numbers of Cu and Cl at different T, P conditions.

The simulation of Au-Cl in both vapor and liquid phases (Table 5.3, No. 4,5) gave a

predominant species of AuCl(H2O)0. Figure 5.7 shows the instantaneous hydration numbers

of Au+ and Cl–, excluding the first shell water in the AuCl(H2O)0 complex. In simulation 5,

which represents the AuCl(H2O)0 complex in a liquid phase with a density of 0.79 g/cm3,

MD gave a hydration number of six for the second-shell (within 4.5 Å of Au; Fig. 5.7a). In

contrast, in a vapor phase (equivalent pure water density of 0.082 g/cm3; AuCl fluid density

of 0.10 g/cm3) at identical P, T, simulation 4 gave a hydration number of two in the second-

shell (Fig. 5.7c), i.e. a total hydration number of three. The hydration of chloride also

decreased dramatically from liquid to vapor phase (from 1.9 to 0.5, Figs. 5.7b,d). Based on

measurements of Au(s) solubility in HCl-bearing steam, Archibald et al. (2001) concluded

that the total hydration number for the Au(I) complex decreased from 5 to 3 as temperature

increased from 300 to 360 °C at constant pressure, agreeing with our conclusion that the

hydration of Cu/Au complexes decreased with decreasing density (since at the same

5

4

3

2

1

0

20151050

1.39

(d)

4

3

2

1

0

2520151050

0.92

(b)7

6

5

4

3

2

1

0

2520151050

2.86

(a)

9

8

7

6

5

4

3

2

120151050

4.63

(c)

Simulation time (ps) Simulation time (ps)

Instantaneous hydration number of C

l

400˚C, 290 bar 0.20 g/cm

3

Inst

anta

neou

s hy

drat

ion

num

ber o

f Cu

400˚C, 400 bar 0.46 g/cm

3


140

pressure, density decreases with increasing temperature). Archibald et al. (2001) obtained a

total hydration number of four at 340 °C, compared to three in the MD simulations.

Figure 5.7 Instantaneous hydration numbers of Au and Cl in coexisting liquid phase and

vapor phases at 340 °C

5.3.3 Na-O and Cl-O bond distances

Whereas the hydration numbers of the Na+ and Cl– ions are correlated with solution density,

the Na-O and Cl-O bond distances are correlated with temperature. The Na-O distances

decrease from 2.38 Å at 25 °C to ~2.33 Å at 1000 °C (Tables 5.1 and 5.2). On the other

hand, Cl-O distances increase from 3.16 Å at 25 °C to ~3.32 Å at 1000 °C. The same trends

were discovered by Driesner (1998) by classical MD simulations, in which Na-O distances

decreased by 0.03 Å (from 2.26 to 2.23 Å) and Cl-O distances increased by 0.02 Å (from

3.22 to 3.24 Å) over a temperature increase from 27 to 317 °C.

Simulation time (ps) Simulation time (ps)

Instantaneous hydration number of C

lInst

anta

neou

s hy

drat

ion

num

ber o

f Au

6

5

4

3

2

1

0

302520151050

2.11

(c)3

2

1

0

302520151050

0.49

(d)340˚C, 139 bar

vapor phase, 0.10 g/cm3

5

4

3

2

1

0

302520151050

1.89

(b)11

10

9

8

7

6

5

4

3

2302520151050

5.94

(a)340˚C, 139 bar

liquid phase, 0.79 g/cm3


141

5.4 Discussion

5.4.1 The nature of neutral Cu(I) and Au(I) chloro-complexes

A few studies have demonstrated experimentally that the solubility of Cu and Au in low

density, high temperature supercritical fluids (1000 °C, 1500 bar, 0.29 g/cm3) is dependent

upon the nature and concentration of alkali ions, a fact that can be explained by the

formation of neutral species such as MCl2Na(aq) (M=Au,Cu) (Zajacz et al. 2010, 2011).

Zajacz et al. (2011) also report static quantum chemical calculations for linear Cu(I)

chloride and bisulfide complexes, suggesting that the NaCuCl2(aq) and NaCuClHS(aq)

species are inner sphere complexes. These calculations relate to the ideal gas phase, or use a

“Polarizable Continuum Model” to represent the solvent, which is characterized by its DC

(Cossi et al., 1996). The calculations are performed for selected complex stoichiometries,

which have been observed in the MD calculations in this study. Both static and MD

calculations present the geometry of CuNaCl2 with Na occurs sideways from the center Cu

atom of the linear CuCl2– complex (Fig. 5.2ab) with close agreement of the Cu-Na distance

(2.73 Å by static calculations vs 3.00 Å by MD according to Fig. 5.4a). The results of these

calculations suggest that an ion pair with Na+ in a well-defined position explain the

solubility data.

The static calculations, however, could not predict the disorder and residence time of this

configuration neither that other fundamentally different geometries appear with a

statistically significant likelihood, while the picture arising from our ab initio MD

simulations provided more details: i) explicit inclusion of the water molecules in MD

simulations results in CuCl2– having a distorted linear geometry, consistent with the

experimental measurements, even in the absence of Na+ within the first or second

coordination shells (see also Mei et al. 2013a). This shows that the hydration has a large


142

effect on the complex geometry, so that MD simulations with explicit water molecules

provide a better prediction of the structures than the static calculations based on continuum

solvent models. ii) Although the MD simulations show the predominance of a neutral

MCl2Na(aq) ion pair in low density, high temperature solutions at conditions similar to the

experiments of Zajacz et al. (2010, 2011), the nature of the complex differs fundamentally

from the picture suggested by static quantum mechanical calculations. The Na+ ion is only

weakly bonded (weak electrostatic bonding), as shown by the large range in M-Na

distances, the lack of definite coordination geometry for the neutral ion pair, and extremely

fast exchange rate of the bonded Na+ (< 15 ps) (Figs. 5.2, 5.3). These exchange rates are fast

relative for example to the H2O residence times for divalent aqua-ions, with range from

0.2 ns for Cu2+ to 30 µs for Ni2+ (Table 11.4 in Burgess 1978). Therefore we suggest that

Na+ is loosely bounded in the outer coordination sphere. The MD simulations also

demonstrate that atoms bounded at the outer sphere (or second shell) are unlikely to be

detected by EXAFS studies (e.g., Brugger et al., 2007; Fulton et al., 2000a,b; Liu et al.,

2001, 2008 for Cu), because of the highly disordered nature of the Cu-Na bond.

5.4.2 Density dependence of the charged or neutral complexes

When dealing with metal speciation in hydrothermal fluids, one of the fundamental

assumptions is that in supercritical fluids, neutral species will predominate over charged

species, because of a dramatic decrease in the DC of water with increasing temperature

(Barnes, 1997). Our ab initio MD simulations accurately reproduce the expected trend

towards a predominance of neutral clusters over charged complexes with decreasing fluid

density and DC. This trend has been demonstrated before for NaCl brines on the basis of

classical MD studies (e.g., Driesner et al., 1998; Sherman and Collings, 2002), showing an

increase in the stability of large clusters in supercritical fluids. For example, the Na2Cl2(aq)


143

cluster is important in NaCl fluids at temperature of 380 °C and density of 0.55 g/cm3

(Driesner et al., 1998).

Our systematic ab initio calculations reveal a negative linear dependency of the degree of

formation of the neutral species MCl2Na(aq) with solution density. Charged complexes

(e.g., CuCl2– and AuCl2

–) are predominant at high densities (> 0.7 g/cm3), even at high

temperature (Fig. 5.5), and neutral complexes are predominant at low densities. These

simulations confirm the interpretation of in situ experimental XAS data accumulated in the

last decade (e.g., Brugger et al., 2007; Fulton et al., 2000a, 2000b; Liu et al., 2008 for Cu),

which emphasize the stability of the linear [CuCl2–] structure over a wide range of

temperatures and pressures. In general, the majority of solubility data is also consistent with

a predominance of CuCl2– (e.g. Liu et al., 2001; Var'yash, 1992; Xiao et al., 1998). For low-

density fluids, the available data also agree to the predominance of neutral species (Zajacz et

al., 2011). Similarly for the Au(I)-chloride system, the MD results compare well to the

available experimental data that suggest that AuCl2– predominates in high-density fluids

(e.g, Pokrovski et al., 2009; Seward, 1973; Stefánsson and Seward, 2003), while

NaAuCl2(aq) is predominant in low density supercritical fluids (Zajacz et al., 2010).

Our MD results have showed that the MD simulations are capable of revealing the loosely-

bonded atoms in the outer coordination sphere of a complex, the key to understand the

neutrality of the species in the high-temperature, low-density fluids. In contrast, in situ

spectroscopic methods such as XAS failed to detect the second shell (Liu et al., 2008).

Because of the fast exchange of Na+ in the NaMCl2(aq) complexes, the ratio of NaMCl2(aq)

to MCl2– species is well constrained by the MD calculations. The linear relationship

observed as a function of fluid density (Figs. 5.5a,b) further allows to extrapolate the MD

results to different P,T conditions. We used these facts to calculate Gibbs free energies of

reaction (5.1) (

!

"rGNaMCl2 (aq )# ) for the formation of the neutral complexes:


144

Na+ + MCl2– = NaMCl2(aq) (M=Cu, Au) (5.1)

The mass action equation for reaction (5.1) is:

!

K ="NaMCl2 (aq )

"Na + # "MCl2$

=[NaMCl2 (aq)][MCl2

$]•% NaMCl2 (aq )"Na +%MCl2$

(5.2)

The Gibbs free energy of reaction (5.1) is:

!

"rGNaMCl2 (aq)# = $RT lnK = $RT ln

[NaMCl2 (aq)] % &NaMCl2 (aq)[MCl2

$ ] %'Na+ % &MCl2$ (5.3)

where "i and !i are the activity and activity coefficients of species i, respectively;

!

[NaMCl2 (aq)] = NNa(M ) , and

!

[MCl2"] = 1" NNa(M ) ;

!

NNa(M ) values are obtained from the

linear relationships shown in Figure 5.5. The activity of Na+ (

!

"Na + ) and the activity

coefficients

!

"MCl2– and

!

"NaMCl2 (aq ) (Tables 5.4, 5.5) were calculated based on distribution of

species calculations performed using the HCh software (Shvarov and Bastrakov 1999). The

Gibbs free energies of formation from the elements for NaMCl2(aq) (

!

" fGNaMCl2 (aq)#

) was

retrieved using the Gibbs free energy of formation of Na+ and MCl2– (Equation 5.1) taken

from Johnson et al. (1992) (Na+), Liu and McPhail (2005) (CuCl2–), and Akinfiev and Zotov

(2001) (AuCl2–). These

!

" fGNaMCl2 (aq)#

values were used to regress Helgeson-Kirkham-

Flowers (HKF) equation of state parameters (Tanger and Helgeson, 1988) using the OptimB

software (Shvarov, 1993). These HKF parameters are listed in Table 5.6, and can be used to

estimate the degree of association in high density (( 0.4 g/cm3) fluids up to 1000 ˚C, 5 kbar.


145

Table 5.4 Fitting of thermodynamic properties of the reaction Na+ + CuCl2

– = NaCuCl2(aq).

T (˚C)

P (bar)

ρ g/cm3

NNa(Cu) (Fit)

!

"CuCl2#

!

"Na +

!

"NaCuCl2 (aq )

!

"rG (kJ/mol) logK

25 1 1.18 0.01 0.64 1.74 0.91 12.0 -2.1 300 500 0.95 0.16 0.31 0.62 0.92 0.367 -0.033 300 1000 0.98 0.14 0.34 0.72 0.92 2.21 -0.20

300 2000 1.03 0.11 0.37 0.85 0.91 4.90 -0.45 500 600 0.67 0.35 0.03 0.028 0.93 -41.1 2.8

500 1000 0.76 0.29 0.14 0.17 0.93 -18.1 1.2 500 2000 0.87 0.22 0.25 0.39 0.92 -6.26 0.42

1000 2500 0.55 0.43 0.12 0.076 0.94 -45.7 1.9 1000 5000 0.74 0.30 0.21 0.21 0.93 -23.5 0.97

Table 5.5 Fitting of thermodynamic properties of the reaction Na+ + AuCl2– = NaAuCl2(aq)

T (˚C)

P (bar)

ρ g/cm3

NNa(Au) (Fit)

!

"AuCl2#

!

"Na +

!

"NaAuCl2 (aq )

!

"rG (kJ/mol)

logK

25 1 1.31 0.11 0.67

1.66 0.91 5.65 -0.99 300 500 1.05 0.20 0.34

0.61 0.92 -0.578 0.053

300 1000 1.09 0.19 0.37

0.71 0.92 0.919 -0.084 300 2000 1.14 0.17 0.038 0.83 0.91 2.69 -0.25 500 600 0.74 0.31 0.04

0.029 0.93 -38.5 2.6

500 1000 0.84 0.28 0.16

0.17 0.93 -16.8 1.1 500 2000 0.96 0.24 0.28

0.39 0.92 -6.32 0.43

1000 2500 0.61 0.36 0.14

0.081 0.94 -41.0 1.7 1000 5000 0.82 0.29 0.23

0.21 0.93 -21.2 0.87


146

Table 5.6 Equation-of-state parameters and standard partial molal properties (HKF

parameters) for the NaCuCl2(aq) and NaAuCl2(aq) complexes regressed from the MD

results.

Parameters NaCuCl2(aq) NaAuCl2(aq)

!

" f G Pr ,Tr

0 (Cal mol–1) -117956 -98036

!

S Pr ,Tr

0 (Cal mol–1 K–1) 72.166 54.218

a1 (Cal mol–1 bar–1)$10 -8.4582 -5.9176

a2 (Cal mol-1)$10–2 93.5847 94.2553

a3 (Cal mol–1 bar–1) -33.1580 -17.7597

a4 (Cal mol-1)$10–4 -6.6478 -6.6755

c1 (Cal mol–1) -1.2859 29.2502

c2 (Cal K mol–1) $10–4 12.7302 30.5724

!

"Pr ,Tr (Cal mol–1) $10–5 -0.0006 0.0384

5.4.3 Density dependence of hydration of ions and metal complexes

The number of water molecules (i.e., the hydration number) surrounding Cu+, Au+, Na+ and

Cl– ions decreases linearly when fluid density decreases (Figs. 5.5c,d). The decrease of

hydration number with the decrease of density has been demonstrated in previous

experimental (Archibald et al., 2001, 2002) and theoretical studies (e.g., Driesner et al.,

1998; Hemley et al., 1992; Seward and Driesner, 2004; Sherman, 2007). Note that in

contrast to the density dependence of the hydration number, the Na-O and Cl-O distances

change with temperature rather than density (see section 5.3.3). The decrease of Na-O

distances with increasing temperature may result from the faster kinetics at high temperature

increasing the attractive force of Na+ and negative charged oxygen in H2O (recognized as

O2– ions). Similarly, the increasing kinetics at high temperature increases the repulsive force

of Cl– and O2– ionic pairs, which lead to the increase of Cl-O distances.


147

There are numerous experimental studies on the solubility of NaCl(s) and hydration of NaCl

ion pairs in water vapor (see reviews of Palmer et al., 2004). In these studies, a linear trend

was established between the solubility of NaCl(s) and the density of the water, with the

intersection of Y axis as concentration constants, and the slope being interpreted as

hydration numbers (e.g., Armellini et al., 1993; Palmer et al., 2004):

CNaCl = logKc + n H2O log #, (5.4)

where CNaCl is the concentration of NaCl in the fluid, n is the number of hydration waters

and # is the density of water. Equation (5.4) is valid for an ideal gas. This approach has been

used to interpret the solubility of Ag(I) and Cu(I) in subcritical water vapor (Archibald et

al., 2002; Migdisov et al., 1999; Williams-Jones et al., 2002) and in supercritical water near

the critical isochore (Liu et al., 2008). Equation (5.4) implies that the hydration number is

independent of density at a given temperature. Our results and other MD studies (Driesner et

al., 1998; Lee et al., 1998; see reviews in Seward and Driesner, 2004) all suggest that

hydration numbers should decrease with fluid density, and therefore the hydration number

implied from the regression of experimental salt solubility (Equation 5.4) is not accurate at a

given temperature, although it may indicate a first-order estimation. For example, our MD

results of the hydration number of Cu are 1.6 at 290 bar and 2.5 at 420 bar at 400 °C,

comparable to the experimental values (2.8 including the water complexes to CuCl, Liu et

al., 2008). Similarly the MD results of the Au-Cl system in vapor phase (hydration number

of 3 at 340 °C) are also comparable to the experimental values (hydration number of 5 at

300 °C to 3 at 360 °C) of Archibald et al. (2001).


148

5.4.4 Translational entropy and the density dependence of hydrothermal

reactions

It is well documented that the free energies of reactions involving aqueous complexes often

vary in a linear fashion with water density (as related to pressure) at fixed temperature

(Anderson et al., 1991; Dolej) and Manning, 2010; Manning 1998). Our MD data provide a

simple explanation for this empirical observation, by revealing a linear correlation (i.e.,

hydration number ~ a + b*s) between the solution density and the hydration number of Cu(I)

and Au(I) complexes that show no evidence for significant change in stoichiometry and

geometry over the investigated P, T range. Consequently, the empirical correlation between

Log K and water density can be explained by assuming that the change in configurational

entropy is proportional to the change in hydration number:

!

"rG = "rH #T"rS= "rH #T(a + b$s)= #RT lnK

so : lnK ="rH #Ta +Tb$

RT~ $ @ constant T

(5.5)

Since the hydration numbers of Na and Cl decline with the decrease of density (increase of

volume), the Cl/Na hydration ions may have smaller volume at high density, and the volume

would increase upon release of the hydration waters to solution, e.g.,

(5.6)

To explain how the charged complexes control the solubility and metal (e.g., Cu, Au)

transport in supercritical and fluids, We considered the 1st shell hydration for the following

chemical reaction:

CuNaCl2 + m H2O ! CuCl2– + Na•(H2O)m

+ (5.7)


149

At high density, Na+ attracts more water to form hydration ions (m ( 3 when density ( 0.9),

so the preferred reaction direction is to form the charged complexes CuCl2–. At low density,

the hydration water of Na ions is released (m & 2 when density & 0.5 g/cm3), so the

preferred reaction direction is to form neutral complex CuNaCl2(aq). The entropy of the left

side in Equation (5.7) is higher than the right side because it is more disordered. The loss of

hydration water of Cl and Na may contribute to the formation of large CuNanClm1+n-m

clusters because of the smaller steric effects and smaller ion volumes at lower densities.

Archibald et al. (2001; 2002) suggested that the solubilities of Au and Cu decreased with

increasing temperature as a result of decreasing hydration number, which is consistent with

the results from our MD simulations. In fluids with low chloride concentrations, the less

hydrated Cl– ions (i.e., hydration of 0.5 at fluid density of 0.1 g/cm3, Table 5.3, No. 4) at

lower density may affect metal solubility. In low-density chloride-rich fluids, Zajacz et al.

(2010) noticed that at a constant Cl (NaCl + HCl) concentration of 0.75 m, gold solubility

increased with the increasing of NaCl/(NaCl+HCl) ratio to 0.5, and then decreased with

higher ratio of NaCl. The solubility experiment by Zajacz et al. (2010) provides the

evidence of dominance of the AuNaCl2 comolex, which is proofed by our MD simulations

show that the stronger ion pairing and existence of neutral NaCl (aq) species.

The traditional Born-Model description explains the increase in ion association with

decreasing solution density as resulting from the decrease in dielectric constant of the

solvent (Born 1920; Seward and Barnes, 1997). A molecular description, however, explains

the ion association as resulting from the increase in translational entropy of products relative

to the reactants. The translation entropy of the products is greater than that of the reactants

since the hydration waters surrounding Na+ are liberated. Because of the increase in

translational entropy, the ion association is favored by temperature. Moreover, the entropy


150

of liberated water molecules will increase with decreasing fluid density; consequently, the

ion association is favored by decreasing fluid density.

5.5 Acknowledgement

Research funding was provided by the Australian Research Council (ARC) to JB

(DP0878903), and the Minerals Down Under Flagship to WL. The MD calculations were

supported by iVEC through the use of advanced computing resources located in Perth,

Australia, and the computational facilities of the Advanced Computing Research Centre in

University of Bristol, UK. This paper is part of Yuan Mei’s PhD thesis. YM acknowledges

the University of Adelaide for IPRS scholarship and CSIRO Minerals Down Under Flagship

for a scholarship top-up.

5.6 References Akinfiev, N. N. and Zotov, A. V., 2001. Thermodynamic description of chloride,

hydrosulfide, and hydroxo complexes of Ag(I), Cu(I), and Au(I) at temperatures of

25-500°C and pressures of 1-2000 bar. Geochemistry International 39, 990-1006.

Anderson, G. M., Castet, S., Schott, J., and Mesmer, R. E., 1991. The density model for

estimation of thermodynamic parameters of reactions at high temperatures and

pressures. Geochimica et Cosmochimica Acta 55, 1769-1779.

Archibald, S. M., Migdisov, A. A., and Williams-Jones, A. E., 2001. The stability of Au-

chloride complexes in water vapor at elevated temperatures and pressures.


Archibald, S. M., Migdisov, A. A., and Williams-Jones, A. E., 2002. An experimental study

of the stability of copper chloride complexes in water vapor at elevated temperatures

and pressures. Geochimica et Cosmochimica Acta 66, 1611-1619.

Armellini, F. J. and Tester, J. W., 1993. Solubility of sodium chloride and sulfate in sub-and

supercritical water vapor from 450–550 °C and 100–250 bar. Fluid Phase Equilibria


151

84, 123-142.

Barnes, H. L., 1997. Geochemistry of Hydrothermal Ore Deposits. Wiley, New York. 795

pp.

Bazarkina, E. F., Pokrovski, G. S., Zotov, A. V., and Hazemann, J. L., 2010. Structure and


276, 1-17.

Berry, A. J., Hack, A. C., Mavrogenes, J. A., Newville, M., and Sutton, S. R., 2006. A

XANES study of Cu speciation in high-temperature brines using synthetic fluid

inclusions. American Mineralogist 91, 1773-1782.

Born, M., 1920. Volumen und Hydratationswärme der Ionen. Zeitschrift für Physik 1, 45-48.

Brugger, J., Etschmann, B., Liu, W., Testemale, D., Hazemann, J. L., Emerich, H., van Beek,

W., and Proux, O., 2007. An XAS study of the structure and thermodynamics of

Cu(I) chloride complexes in brines up to high temperature (400°C, 600bar).


Brugger, J., Pring, A., Reith, F., Ryan, C., Etschmann, B., Liu, W.H., O'Neill, B., and

Ngothai, Y. (2010) Probing ore deposits formation: New insights and challenges

from synchrotron and neutron studies. Radiation Physics and Chemistry 79, 151-161.

Burgess, J., (1978). Metal Ions in Solution. Published by Ellis Horwood Ltd, Chichester and

distributed by John Wiley and Sons, Chichester and New York. 481 pp.



Chialvo, A. and Simonson, J., 2003. Aqueous NaCl pair association from liquid-like to

steam-like densities along near-critical isotherms. The Journal of Chemical Physics

118, 7921.

Cossi, M., Barone, V., Cammi, R., and Tomasi, J., 1996. Ab initio study of solvated

molecules: a new implementation of the polarizable continuum model. Chemical


Crerar, D. A., Susak, N., Borcsik, M., and Schwartz, S., 1978. Solubility of the buffer

assemblage pyrite + pyrrhotite + magnetite in NaCl solutions from 200 to 350 °C.


Cui, S. and Harris, J., 1995. Solubility of sodium chloride in supercritical water: A

molecular dynamics study. The Journal of Physical Chemistry 99, 2900-2906.

Debye, P. and Hückel, E., 1923. The theory of electrolytes. I. Lowering of freezing point

and related phenomena (In German). Physikalische Zeitschrift 24, 185–206.


152

Dolejs, D. and Manning, C. E., 2010. Thermodynamic model for mineral solubility in

aqueous fluids: theory, calibration and application to model fluid-flow systems.

Geofluids 10, 20-40.

Driesner, T., 2007. The system H2O–NaCl. Part II: Correlations for molar volume, enthalpy,

and isobaric heat capacity from 0 to 1000°C, 1 to 5000bar, and 0 to 1 XNaCl.


Driesner, T. and Heinrich, C. A., 2007. The system H2O–NaCl. Part I: Correlation formulae

for phase relations in temperature–pressure–composition space from 0 to 1000°C, 0

to 5000bar, and 0 to 1 XNaCl. Geochimica et Cosmochimica Acta 71, 4880-4901.




Acta 62, 3095-3107.





Acta 74, 4723-4739.

Frank, M. R., Simon, A. C., Pettke, T., Candela, P. A., and Piccoli, P. M., 2011. Gold and

copper partitioning in magmatic-hydrothermal systems at 800 degrees C and 100

MPa. Geochimica et Cosmochimica Acta 75, 2470-2482.

Fulton, J. L., Hoffmann, M. M., and Darab, J. G., 2000a. An X-ray absorption fine structure

study of copper(I) chloride coordination structure in water up to 325 °C. Chemical


Fulton, J. L., Hoffmann, M. M., Darab, J. G., Palmer, B. J., and Stern, E. A., 2000b.

Copper(I) and Copper(II) Coordination Structure under Hydrothermal Conditions at

325 °C:# An X-ray Absorption Fine Structure and Molecular Dynamics Study. The

Journal of Physical Chemistry A 104, 11651-11663.

Giberti, F., Tribello, G. A., and Parrinello, M., 2013. Transient Polymorphism in NaCl.

Journal of Chemical Theory and Computation 9, 2526-2530.

Hack, A. C. and Mavrogenes, J. A., 2006. A synthetic fluid inclusion study of copper

solubility in hydrothermal brines from 525 to 725 C and 0.3 to 1.7 GPa. Geochimica



153




Hassan, S. A., 2011. Microscopic mechanism of nanocrystal formation from solution by

cluster aggregation and coalescence. The Journal of chemical physics 134, 114508.

Helgeson, H. and Kirkham, D. H., 1974. Theoretical prediction of the thermodynamic

behavior of aqueous electrolytes at high pressures and temperatures: II. Debye-

Hückel parameters for activity coefficients and relative partial molal properties.

American Journal of Science 274, 1199-1261.

Hemley, J., Cygan, G., Fein, J., Robinson, G., and d'Angelo, W., 1992. Hydrothermal ore-

forming processes in the light of studies in rock-buffered systems; I, Iron-copper-

zinc-lead sulfide solubility relations. Economic Geology 87, 1-22.

Hoffmann, M. M., Darab, J. G., Palmer, B. J., and Fulton, J. L., 1999. A transition in the

Ni2+ complex structure from six- to four-coordinate upon formation of ion pair

species in supercritical water: An X-ray absorption fine structure, near-infrared, and

molecular dynamics study. Journal of Physical Chemistry 103, 8471-8482.

House, J. E., 2012. Inorganic Chemistry. Elsevier Academic Press (second edition). 848 pp.

Humphrey, W., Dalke, A., and Schulten, K., 1996. VMD: visual molecular dynamics.

Journal of Molecular Graphics 14, 33-38.

John P, B., 1998. Molecular dynamics simulations of aqueous NaCl solutions at high

pressures and temperatures. Chemical Geology 151, 11-19.

Johnson, J. W., Oelkers, E. H., and Helgeson, H. C., 1992. SUPCRT92: A software package

for calculating the standard molal thermodynamic properties of minerals, gases,

aqueous species, and reactions from 1 to 5000 bar and 0 to 1000˚C. Computers

Geosciences 18, 899-947.

Laasonen, K., Pasquarello, A., Car, R., Lee, C., and Vanderbilt, D., 1993. Car-Parrinello

molecular dynamics with Vanderbilt ultrasoft pseudopotentials. Physical Review B

47, 10142-10153.

Lee, S., Cummings, P., Simonson, J., and Mesmer, R., 1998. Molecular dynamics

simulation of the limiting conductance of NaCl in supercritical water. Chemical


Lerchbaumer, L. and Audétat, A., 2012. High Cu concentrations in vapor-type fluid

inclusions: An artifact? Geochimica et Cosmochimica Acta 88, 255-274.


154

Liu, W., Borg, S., Etschmann, B., Mei, Y., and Brugger, J., 2012a. An XAS study of

speciation and thermodynamic properties of aqueous zinc bromide complexes at 25–

150 °C. Chemical Geology 298–299, 57-69.

Liu, W., Borg, S. J., Testemale, D., Etschmann, B., Hazemann, J.-L., and Brugger, J., 2011a.

Speciation and thermodynamic properties for cobalt chloride complexes in

hydrothermal fluids at 35–440°C and 600bar: An in-situ XAS study. Geochimica et


Liu, W., Brugger, J., Etschmann, B., Testemale, D., and Hazemann, J.-L., 2008. The

solubility of nantokite (CuCl(s)) and Cu speciation in low-density fluids near the

critical isochore: An in-situ XAS study. Geochimica et Cosmochimica Acta 72,

4094-4106.

Liu, W., Etschmann, B., Foran, G., Shelly, M., Foran, G., and Brugger, J., 2007. Deriving

formation constants for aqueous metal complexes from XANES spectra: Zn(II) and

Fe(II) chloride complexes in hypersaline solutions. American Mineralogist 92, 761-

770.

Liu, W. and McPhail, D. C., 2005. Thermodynamic properties of copper chloride complexes

and copper transport in magmatic-hydrothermal solutions. Chemical Geology 221,

21-39.




2948.

Liu, X., Lu, X., Wang, R., and Zhou, H., 2012b. Silver speciation in chloride-containing



Liu, X., Lu, X., Wang, R., Zhou, H., and Xu, S., 2011b. Speciation of gold in

hydrosulphide-rich ore-forming fluids: Insights from first-principles molecular

dynamics simulations. Geochimica et Cosmochimica Acta 75, 185-194.

Manning, C. E., 1998. Fluid composition at the blueschist - eclogite transition in the model

system Na2O-MgO-Al2O3-SiO2-H2O-HCl. Schweizerische Mineralogische Und

Petrographische Mitteilungen 78, 225-242.

Mavrogenes, J. A., Berry, A. J., Newville, M., and Sutton, S. R., 2002. Copper speciation in

vapor-phase fluid inclusions from the Mole Granite, Australia. American

Mineralogist 87, 1360-1364.


155




Mei, Y., Sherman, D. M., Liu, W., and Brugger, J., 2013. Complexation of gold in S3–-rich



Migdisov, A. A., Williams-Jones, A., and Suleimenov, O., 1999. Solubility of chlorargyrite

(AgCl) in water vapor at elevated temperatures and pressures. Geochimica et


Oelkers, E. H. and Helgeson, H. C., 1993. Multiple ion association in supercritical aqueous

solutions of single electrolytes. Science 261, 888.

Palmer, D. A., Simonson, J., and Jensen, J., 2004. Partitioning of electrolytes to steam and

their solubilities in steam. Aqueous systems at elevated temperatures and pressures:

New York, Elsevier, 409-439.

Perdew, J. P., Burke, K., and Ernzerhof, M., 1996. Generalized Gradient Approximation

Made Simple. Physical Review Letters 77, 3865-3868.

Pokrovski, G. S., Roux, J., Hazemann, J.-L., and Testemale, D., 2005. An X-ray absorption

spectroscopy study of argutite solubility and aqueous Ge(IV) speciation in

hydrothermal fluids to 500 °C and 400 bar. Chemical Geology 217, 127-145.

Pokrovski, G. S., Tagirov, B. R., Schott, J., Bazarkina, E. F., Hazemann, J.-L., and Proux,

O., 2009. An in situ X-ray absorption spectroscopy study of gold-chloride

complexing in hydrothermal fluids. Chemical Geology 259, 17-29.




Acta 106, 501-523.

Rempel, K. U., Liebscher, A., Meixner, A., Romer, R. L., and Heinrich, W., 2012. An

experimental study of the elemental and isotopic fractionation of copper between

aqueous vapour and liquid to 450 °C and 400 bar in the CuCl-NaCl-H2O and CuCl-

NaHS-NaCl-H2O systems. Geochimica et Cosmochimica Acta 94, 199-216.

Seward, T. M., 1973. Thio complexes of gold and the transport of gold in hydrothermal ore

solutions. Geochimica et Cosmochimica Acta 37, 379-399.




156

Seward, T. and Driesner, T., 2004. Hydrothermal solution structure: Experiments and

computer simulations. In: Palmer, D. A., Fernández-Prini, R., and Havey, A. H.

Eds.), Aqueous Systems at Elevated Temperatures and Pressures. Elsevier.

Seyfried, W. and Ding, K., 1993. The effect of redox on the relative solubilities of copper

and iron in Cl-bearing aqueous fluids at elevated temperatures and pressures: An

experimental study with application to subseafloor hydrothermal systems.





Shvarov, Y. V., 1993. UT-HEL: IBM Computer Code for Calculating the HKF revised

Equation of State Parameters for Aqueous Species. Moscow State University.

Shvarov, Y. and Bastrakov, E., 1999. HCh: a software package for geochemical modelling.

User's guide. AGSO record 1999/25, 60 pp.



Sherman, D. M., 2010. Metal complexation and ion association in hydrothermal fluids:

insights from quantum chemistry and molecular dynamics. Geofluids 10, 41-57.




Simon, A. C., Pettke, T., Candela, P. A., Piccolli, P. M., and Heinrich, C. A., 2006. Copper

partitioning in a melt-vapor-brine-magnetite-pyrrhotite assemblage. Geochimica Et


Simon, A. C., Frank, M. R., Pettke, T., Candela, P. A., Piccoli, P. M., and Heinrich, C. A.,

2005. Gold partitioning in melt-vapor-brine systems. Geochimica Et Cosmochimica

Acta 69, 3321-3335.



Stefánsson, A. and Seward, T. M., 2003. Stability of chloridogold(I) complexes in aqueous

solutions from 300 to 600 °C and from 500 to 1800 bar. Geochimica et



157

Susak, N., and Crerar, D.A., 1985. Spectra and coordination changes of transition metals in

hydrothermal solutions: Implications for ore genesis. Geochimica et Cosmochimica

Acta 49, 555-564.

Tanger, J. C. I. and Helgeson, H. C., 1988. Calculation of the thermodynamic and transport

properties of aqueous species at high pressures and temperatures: Revised equations

of state for the standard partial molal properties of ions and electrolytes. American

Journal of Science 288, 19-98.

Testemale, D., Brugger, J., Liu, W., Etschmann, B., and Hazemann, J. L., 2009. In-situ X-

ray absorption study of Iron (II) speciation in brines up to supercritical conditions.


Tian, Y., Etschmann, B., Liu, W., Borg, S., Mei, Y., Testemale, D., O'Neill, B., Rae, N.,

Sherman, D. M., Ngothai, Y., Johannessen, B., Glover, C., and Brugger, J., 2012.

Speciation of nickel (II) chloride complexes in hydrothermal fluids: In situ XAS

study. Chemical Geology 334, 345-363.

Var'yash, L. N., 1992. Cu(I) complexing in NaCl solutions at 300 and 350°C. Geochemistry

International 29, 84-92.

Verlet, L., 1967. Computer "experiments" on classical fluids. I. Thermodynamical

properties of Lennard-Jones molecules. Physical Review 159, 98.

Williams-Jones, A. E., Migdisov, A. A., Archibald, S. M., and Xiao, Z., 2002. Vapor-

transport of ore metals. Water-rock interaction: a tribute to David A. Crerar:

Geochemical Society Special Publication 7, 279-305.

Xiao, Z., Gammons, C. H., and Williams-Jones, A. E., 1998. Experimental study of copper(I)

chloride complexing in hydrothermal solutions at 40 to 300°C and saturated water

vapor pressure. Geochimica et Cosmochimica Acta 62, 2949-2964.

Zahn, D., 2004. Atomistic Mechanism of NaCl Nucleation from an Aqueous Solution.

Physical Review Letters 92, 040801.

Zajacz, Z., Seo, J. H., Candela, P. A., Piccoli, P. M., Heinrich, C. A., and Guillong, M.,

2010. Alkali metals control the release of gold from volatile-rich magmas. Earth and

Planetary Science Letters 297, 50-56.

Zajacz, Z., Seo, J. H., Candela, P. A., Piccoli, P. M., and Tossell, J. A., 2011. The solubility

of copper in high-temperature magmatic vapors: A quest for the significance of

various chloride and sulfide complexes. Geochimica et Cosmochimica Acta 75,

2811-2827.


158


hydrogen sulfide gas: An experimental study. Geochimica et Cosmochimica Acta 71,

3070-3081.


H2O–H2S vapour at elevated temperature and pressure. Geochimica et


159

_____________________________________

Chapter 6

Conclusion ____________________________________

CHAPTER 6. CONCLUSION

161

The major contribution of this PhD thesis is to demonstrate the potential of ab initio

molecular dynamics simulations in investigating speciation, geometries, stoichiometries and

thermodynamic properties of transition metal complexes responsible for the transport of

these metals in hydrothermal fluids.

6.1 d10 transition metal ions complexation

Ab initio MD conducted in this thesis predicted the complexation of Cu+, Au+ and Zn2+ ions

in chlorine- and sulfur-rich fluids successfully. These ions share a similar electronic

configuration, in that their d-orbitals are all fully occupied (closed shell). MD simulations

are particularly well suited for studying d10-metal complexes because of well-defined

electronic configurations and lower computing costs for spin-restricted systems. In

simulations of Zn-Cl, ab initio MD predicted the stability of tetrahedral [ZnCln(H2O)4-n]2-n

complexes from ambient to hydrothermal conditions. Tetrahedral complexes are very

common in transition metal ions; complexation and geometry can be explained by crystal-

field theory (Janes et al., 2004). The fully occupied electronic configuration of Zn2+ results

in no crystal-field stabilization energy (CFSE); tetrahedral complexes are favored by large

ligands such as Cl–, Br– and I– with such d10 ions. However, despite the fact that metal ions

in group BI (e.g., Cu+, Au+) share the same electronic configuration as the Zn2+ ion, their

complexes have different coordination and structures. For Au+ and Cu+ complexes, ab initio

MD gave linear ML2– complexes (e.g., CuCl2

–, Cu(HS)2–, AuCl2

–, Au(S3)2–, Au(S3)(HS)–,

etc) in most conditions (except CuCl32– in hypersaline fluids at low temperature (<200˚C)).

The linear aqueous species ML(H2O)0 are also observed as intermediate structures (Chapter

2, Mei et al., 2013a) or in solutions with low ligand concentrations (Chapter 5). Orgel (1958)

suggested that the stability of the linear configuration for these ions results from the


162

hybridization among (n-1)d orbitals (

!

dz 2 ) and ns and np (pz) orbitals that have similar

energies in these metals (review by Huheey et al., 1983).

Hence, the structural properties obtained from ab initio MD are in good agreement with the

available experimental studies, and are consistent with the theory of complex formation.

6.2 Metal complexation in mixed ligand solutions

In most experimental studies of metal mobility in hydrothermal fluids, a single ligand (in

addition to H2O as water acts as solvent) was included in each experiment because 1) simple

systems help to understand the fundamentals of individual ligand-forming reactions;

2) adding more constraints makes data interpretation easier and provides more precise

results. However, in real fluids several ligands may compete for the metals. For example,

both Cl– and HS– play important role in hydrothermal fluids, but few studies investigated the

transport of metals in solutions containing both Cl– and HS– mixture (e.g., Etschmann et al.,

2010). MD provides a molecular-level understanding of metal complexation and is well

suited for the study of mixed ligand systems. For example, we studied Cu(I) complexation

in Cl– and HS– fluids in Chapter 2 and Au(I) complexation in high pressure sulfur-rich fluids

with the ligands S3–, HS– and OH– in Chapter 3. These studies succeeded in identifying

novel metal complexes (e.g., mix-ligand complex Cu(HS)Cl–) and in providing

thermodynamic properties for some of these species. Hence, ab initio MD can expand our

knowledge of metal behavior in real fluids, and provide critical information to guide the

design of experimental studies aimed at testing the MD predictions.

6.3 Thermodynamic properties

One of the most significant contributions of this thesis is to demonstrate the potential of ab

initio MD for calculating thermodynamic properties for the formation of transition metal


163

complexes under hydrothermal conditions (Chapter 2, 4; Mei et al., 2013a). The approach of

distance constraint thermodynamic integration employed in this thesis enables us to

measure the force and free energy of ligand-association/dissociation reactions directly at

molecular level. For the simulations at high temperatures (i.e. ( 300 ˚C), the thermodynamic

properties obtained by ab initio MD are in good agreements with experimental values (i.e.

with in 1-2 log units). Because of the fast kinetics at high temperatures, it is more likely that

the system will be approaching equilibrium at high temperature within the time scale of ab

initio MD (picoseconds) for the complexes studied. Although larger uncertainties appear at

low temperatures, the MD results are still comparable with the experiments, and, most

importantly, enable to reproduce the mineral solubility gradients as a function of P, T and

solution composition. As experimental studies designed to obtain thermodynamic properties

at high T, P are highly demanding and may have significant uncertainties (Chapter 4), ab

initio MD is coming of age for obtaining thermodynamic properties and validating and

guiding experimental studies.

6.4 Current limitations

The ab initio MD simulations demonstrated in this thesis show the potential of

computational approaches for studying metal mobility in hydrothermal fluids. However, the

current methods still have significant limitations. One major limitation is that current

computational resources limit the scale of the simulations. For example, ab initio MD

simulations in this thesis were conducted for ~200 particles and up to ~40 ps. Although

current performance of ab initio MD has improved significantly compared with the time

scale of a few picoseconds in the past (e.g., Harris et al., 2003; Sherman, 2007), it is still

difficult to reproduce the equilibrium of many chemical reactions in such a short time

(picoseconds). As the kinetics of ligand exchange reactions is faster at high temperatures,


164

the predictions of speciation and thermodynamic properties gave better agreements

compared to experiments (i.e. ( 300 ˚C, Mei et al., 2013a; Chapter 2; Chapter 4), but the

simulations at lower temperature (i.e. < 200 ˚C) still need to be improved. The size of the

box (or the number of particles) also affects the accuracy of the simulation. As ab initio MD

simulations with only a small number of particles (i.e. magnitude of several hundreds) are

affordable currently (Chapter 5), the simulation systems at this scale reflect a solution with

high concentrations of metal and ligands (several molal), which bring significant errors in

accounting for activity coefficients to get standard state properties (infinite dilution).

Another source of uncertainty comes from the computational methods. As there is no

analytical solution for the Schrödinger equation (Chapter 1, Equation 1.1), the accuracy of

the MD simulation depends on the approximation methods used to describe the

wavefunction. For example, the pseudo-potentials used in the Car-Parrinello methods is not

as accurate as methods such as CCSD(T) (coupled-cluster perturbation theory with single,

double and triple excitations from the Hartree–Fock determinant), which currently

represents the “gold standard” of molecular quantum chemistry (Tossell, 2012). However,

the accurate CCSD(T) method is only affordable for small molecule calculations, and it is

not (yet) feasible to conduct simulations for a large and disordered system (e.g., metal in

hydrothermal fluid) at such a high level of theory. Instead, the pseudo-potentials,

particularly ultra-soft pseudo-potential with the Car-Parrinello method used in this thesis

gave reasonably accurate results within affordable CPU hours: this method reproduces

accurately the geometry of the complexes, and provides quantitative to semi quantitative

thermodynamic properties. Simulation of mineral solubility based on these properties

provide predictions of solubility within an order of magnitude of experimental data, and

accurately reproduce solubility gradients as a function of P, T, and ligand concentrations.

The latter feature is key to simulations of metal transport in natural and engineered systems.


165

All of these issues will become less predominant in the future with the development of more

efficient simulation methods and improvements in computational capacity (e.g., use of

GPU). Hence, we expect that the capacity of ab initio MD to reproduce real solutions and

get more reliable results will improve quickly, and this method will become a pillar to

underpin our understanding of metal mobility in hydrothermal systems.

6.5 Perspective and future work

MD simulations in this thesis focus on the metal-ligand complexation in aqueous solutions.

MD simulations will play an important role in exploring key aspects of the geochemistry of

hydrothermal systems and deciphering the complicated chemistry that leads, for example, to

the formation of ore deposits.

The acidity of aqueous solutions is an essential variable in controlling metal dissolution and

precipitation. As chloride and bisulfide are two important ligands in metal transport, the

effect of pH on these ions needs to be well understood. Experimental studies show that the

pKa of H2S(aq) changes with changes of T, P (Suleimenov and Seward, 1997), while there

is a controversy of the pKa of HCl in hydrothermal fluids (Frantz and Marshall, 1984;

Tagirov et al., 1997). A DFT based MD method opens the way to calculating pKa with a

high level of accuracy (Sulpizi and Sprik, 2008; 2010).

Redox potential is another important part of metal mobility as some metals exist in the Earth

as different valences (e.g., Cu(I) and Cu(II); Fe(II) and Fe(III); U(IV) and U(VI)). Redox

potentials can also be calculated by quantum mechanical simulations (e.g., Cheng et al.,

2009). The computational results can be used to interpret previous existing data (for

example, in Chapter 4), cross-proof experimental results, and evaluate the gradient by

making pH-redox diagrams (i.e. change of speciation as a function of pH and redox,

Figure 1.2).


166

Many of the Earth systems include more than one phase, for example, the mineral-water

interface (solid-liquid phase), volcanic gas and lava (vapor-melt), phase separation upon

cooling or decompression when fluids rise in the crust (liquid-vapor). The ab initio MD

methods used in this thesis are too computationally intensive to simulate systems with

different phases. One realistic approach is combining the classical MD and ab initio MD:

for the large scale simulations, the simulation of individual phases will be treated as ab

initio MD, and these results will be used to construct the parameters of force-field classical

MD (Demichelis et al., 2012). Then classical MD will be employed to give a global view

about molecular properties at the scale of thousands of particles and several nanoseconds for

systems containing both liquid and vapor phases.

6.6 Reference Demichelis, R., Raiteri, P., Gale, J. D., Quigley, D., and Gebauer, D., 2011. Stable

prenucleation mineral clusters are liquid-like ionic polymers. Nature

communications 2, 590.





Acta 74, 4723-4739.

Frantz, J. D. and Marshall, W. L., 1984. Electrical conductances and ionization constants of

salts, acids, and bases in supercritical aqueous fluids; I, Hydrochloric acid from 100

degrees to 700 degrees C and at pressures to 4000 bars. American Journal of Science

284, 651-667.





167

Huheey, J. E., Keiter, E. A., Keiter, R. L., and Medhi, O. K., 1983. Inorganic chemistry:

principles of structure and reactivity. Harper & Row New York. page 464-467

Janes, R., Moore, E., and Abel, E. W., 2004. Metal-ligand bonding. Royal Society of

Chemistry, Cambridge.




Mei, Y., Sherman, D. M., Liu, W., and Brugger, J., 2013b. Complexation of gold in S3--rich



Orgel, L. E., 1958. 843. Stereochemistry of metals of the B sub-groups. Part I. Ions with

filled d-electron shells. Journal of the Chemical Society (Resumed) 0, 4186-4190.



Suleimenov, O. M. and Seward, T. M., 1997. A spectrophotometric study of hydrogen

sulphide ionisation in aqueous solutions to 350 °C. Geochimica et Cosmochimica

Acta 61, 5187-5198.

Sulpizi, M. and Sprik, M., 2008. Acidity constants from vertical energy gaps: density

functional theory based molecular dynamics implementation. Physical Chemistry


Sulpizi, M. and Sprik, M., 2010. Acidity constants from DFT-based molecular dynamics

simulations. Journal of Physics: Condensed Matter 22, 284116.

Tagirov, B. R., Zotov, A. V., and Akinfiev, N. N., 1997. Experimental study of dissociation

of HCl from 350 to 500°C and from 500 to 2500 bars: Thermodynamic properties of

HCl°(aq). Geochimica et Cosmochimica Acta 61, 4267-4280.

Tossell, J. A., 2012. Calculation of the properties of the S3! radical anion and its complexes

with Cu+ in aqueous solution. Geochimica et Cosmochimica Acta 95, 79-92.


168

169

_____________________________________

Appendix A

Speciation and thermodynamic properties of Manganese (II) chloride complexes in hydrothermal fluids: in situ XAS study

_____________________________________

170

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' !0$4/#"/(&' #&5' ",$-%(56&#%/4' 0-(0$-"/$.' ()' 7#&8#&$.$' 9::;' 4,2(-/5$' 4(%02$<$.' /&'

,65-(",$-%#2')2+/5.='!"#$!%&'>*!'."+56'

3+?2/4#"/(&'!"#"+.''

'

3+?2/4#"/(&'@$"#/2.' 1/#&A' BCA' D".4,%#&&A' ECA' 7$/A' BCA' F-+&52$-A' 3CA' 1$."$%#2$A' @CA' G#H$%#&&A' ICJKCA' D22/(""A' 3CA'L8(",#/A'BCA'E-+88$-A'ICA#!0$4/#"/(&'#&5' ",$-%(56&#%/4'0-(0$-"/$.'()'7#&8#&$.$' 9::;'4,2(-/5$'4(%02$<$.' /&' ,65-(",$-%#2' )2+/5.=' !"# $!%&' >*!' ."+56C' 9+&5$-' -$M/$N' )(-' F$(4,/%/4#'O(.%(4,/%/4#'*4"#;'

!"#$%&'(%)#&*+"#*%),'E6' ./8&/&8' ",$' !"#"$%$&"' ()' *+",(-.,/0A' $#4,' #+",(-' 4$-"/)/$.' ",#"' ",$/-' ."#"$5' 4(&"-/?+"/(&' "(' ",$' 0+?2/4#"/(&' /.' #44+-#"$' #&5' ",#"'0$-%/../(&'/.'8-#&"$5')(-'",$'0+?2/4#"/(&'"('?$'/&42+5$5'/&'",$'4#&5/5#"$P.'",$./.C'

L#%$'()'3-/&4/0#2'*+",(-'9O#&5/5#"$;' B+#&'1/#&'

O(&"-/?+"/(&'"('",$'3#0$-' @$./8&$5' #&5' 0$-)(-%$5' $<0$-/%$&".A' /&"$-0-$"$5' #&5' 0-(4$..$5' >*!' 5#"#A' #&5' N-("$'%#&+.4-/0"C'

'

!/8&#"+-$'

'

@#"$' QRJI+2JST'

'

L#%$'()'O(J*+",(-' E#-?#-#'D".4,%#&&'

O(&"-/?+"/(&'"('",$'3#0$-' *../."$5'N/",'$<0$-/%$&".A'#&5'%#&+.4-/0"'$M#2+#"/(&C'

!/8&#"+-$''

@#"$' QRJI+2JST'

'

L#%$'()'O(J*+",(-' B+#&'7$/'

O(&"-/?+"/(&'"('",$'3#0$-' 3$-)(-%$5'@U1'4#24+2#"/(&.A'#&5'N-("$'",$'@U1'0#-"'()'",$'%#&+.4-/0"C'

!/8&#"+-$'

'

@#"$' QRJI+2JST'

'

L#%$'()'O(J*+",(-' 3#.4#2'F-+&52$-'

O(&"-/?+"/(&'"('",$'3#0$-' *../."$5' N/",' .#%02$' %#"$-/#2' .6&",$./.' #&5' ."-+4"+-#2' %$#.+-$%$&".A' #&5' %#&+.4-/0"'$M#2+#"/(&C'

!/8&#"+-$'

'

@#"$' QRJI+2JST'

'

171

!

!

Name of Co-Author Denis Testemale

Contribution to the Paper Assisted with experiments, and manuscript evaluation.

Signature

Date 24-Jul-13

Name of Co-Author Jean-Louis Hazeman

Contribution to the Paper Assisted with experiments.

Signature

Date 24-Jul-13

Name of Co-Author Peter Elliott

Contribution to the Paper Assisted with sample material structural measurements, and manuscript evaluation.

Signature

Date 24-Jul-13

Name of Co-Author Yung Ngothai

Contribution to the Paper Assisted with manuscript evaluation.

Signature

Date 24-Jul-13

Name of Co-Author Joël Brugger

Contribution to the Paper Assisted with experiments and experimental design, and manuscript evaluation.

Acting as corresponding author.

Signature

Date 24-Jul-13

APPENDIX A

173

Speciation and thermodynamic properties of Manganese (II)

chloride complexes in hydrothermal fluids: in situ XAS study

Yuan Tian1, 2, Barbara Etschmann2, 3, Yuan Mei2, 3, 4, Pascal V. Grundler2, 3†, Denis

Testemale5, Jean-Louis Hazeman5, Peter Elliott2, Yung Ngothai1, Joël Brugger2, 3*

1. School of Chemical Engineering, The University of Adelaide, Adelaide 5000,

South Australia, Australia

2. Division of Mineralogy, South Australian Museum, Adelaide 5000, South

Australia, Australia

3. Tectonics, Resources and Exploration (TRaX), School of Earth and

Environmental Sciences, The University of Adelaide, Adelaide 5000, South

Australia, Australia

4. CSIRO Earth Science and Resource Engineering, Clayton, Victoria, 3168,

Australia

5. Institut Neel, Département MCMF, 38042 Grenoble, France; and FAME

beamline, ESRF, 38043 Grenoble, France

† Present address: Laboratory for Nuclear Materials, Paul Scherrer Institute, CH-

5232 Villigen PSI, Switzerland

* Corresponding author: [email protected]

APPENDIX A

174

Abstract The speciation of Mn(II) in acidic brines under a wide range of conditions (30-550 ˚C,

600 bar, 0.100 - 10.344 m chloride and 0.110 - 2.125 m bromide) was investigated using in

situ X-ray Absorption Spectroscopy (XAS). Increasing temperature and/or salinity results in

a structural change of the Mn(II) complexes from octahedral to (distorted) tetrahedral.

Octahedral species predominate at room temperature within the whole salinity range and

persist up to ~400 ˚C in low salinity solutions (mCl < 1 m), and tetrahedral species become

significant above 300 ˚C. A combination of EXAFS refinements, Density Functional Theory

calculations and ab initio XANES simulations shows that at temperatures ( 400 ˚C, the

highest order chlorocomplex predominating in high salinity solutions (mCl > 3 m, Cl:Mn

ratio > 53) is MnCl3(H2O)-, and that a lower order chlorocomplex, MnCl2(H2O)2(aq), is the

predominant species in low salinity solutions (mCl < 0.5 m, Cl:Mn ratio < 10). A similar

result was also found in Mn bromide solutions: MnBr3(H2O)- and MnBr2(H2O)2(aq) are the

dominant species at 500 ˚C in high salinity solutions (e.g., 2.125 m, Br:Mn ratio = 33.73)

and in low salinity solutions (e.g., 0.110 m, Br:Mn ratio = 2.04), respectively. XANES

spectra of Mn(II) chloride solutions were used to retrieve formation constants of

MnCl2(H2O)2(aq) and MnCl3(H2O)- at 600 bar. The speciation and thermodynamic model of

this study are consistent with previous solubility and UV-Vis spectroscopic studies.

Keywords: XAS, speciation, thermodynamic properties, Mn(II) chloride complexes,

coordination change, EXAFS, XANES, hydrothermal fluids.

APPENDIX A

175

1 Introduction

Manganese is the second most abundant transition element in the Earth’s crust. It is highly

mobile in hydrothermal systems, occurring in high concentrations in geothermal waters,

with values up to 59 ppm Mn recorded in Mid Oceanic Ridge fluids with Cl concentrations

close to that of seawater (~0.5 m Cl; James et al., 1995) and values up to 1900 ppm in

hypersaline brines (>5 m Cl; Salton Sea, McKibben and Williams, 1989). In general, Mn

concentrations in natural fluids increase with increasing temperature and increasing chlorine

concentrations, and hypersaline magmatic brines can carry in excess of one wt% Mn

(Yardley, 2005). The formation of much of the world’s Mn resources is related to shallow

hydrothermal circulation and interaction with surface waters (Brugger and Gieré, 1999;

Brugger and Meisser 2006; Cornell and Schütte, 1995; Roy, 1992), but due to its mobility,

Mn is an important metal in many hydrothermal processes (e.g. epithermal deposits; Davies

et al., 2008).

Accurate modeling of the transport and deposition of Mn under hydrothermal conditions is

required to improve predictive mineral exploration and to increase the efficiency of mineral

and metallurgical processing techniques (e.g., Seward and Driesner, 2004; Brugger et al.,

2010). Understanding Mn transport and deposition relies on our knowledge of the aqueous

complexes responsible for the mobility of Mn in hydrothermal fluids and on the availability

of thermodynamic properties for each of these species as a function of pressure, temperature,

and fluid composition (Seward and Barnes, 1997). Mn is widely accepted to exist mainly in

the form of the Mn(II) aqua ion and Mn(II) chloro complexes in hydrothermal waters

(Gammons and Seward, 1996; Yardley, 2005), and a number of studies have been carried

out to investigate Mn(II) aqueous speciation over a wide range of conditions using different

techniques (Table 1).

APPENDIX A

176

At temperatures to 300 ˚C, there is good agreement that the Mn2+ aqua ion, [Mn(H2O)6]2+

(e.g., Koplitz et al., 1994) and low order octahedral chlorocomplexes (mainly MnCl+)

predominate in acidic waters containing up to ~1 m Cl. MnCl+ was identified on the basis of

potentiometric (Libu+ and Tialowska, 1975), electron spin resonance (Wheat and Carpenter,

1988), solubility (Gammons and Seward, 1996) and spectrophotometric (Libu+ and

Tialowska, 1975; Suleimenov and Seward, 2000) measurements. In highly saline solutions

(Cl >> 2 m) and with increasing temperature (150-300 ˚C), octahedral MnCl2(aq) becomes an

important complex; however no evidence for a higher order Mn(II) chloride complex, e.g.,

MnCl3- and MnCl4

2- was found in solutions up to 300 ˚C (Gammons and Seward, 1996;

Suleimenov and Seward, 2000). Boctor (1985) found that Mn(II) speciation was dominated

by MnCl2(aq) in supercritical solutions with low Cl:Mn ratio (2-7) over the temperature range

400 to 700 ˚C at 1 and 2 kbar. At higher Cl:Mn ratios (Cl:Mn >> 7), Uchida et al. (1995,

2003) and Uchida and Tsutsui (2000) concluded that MnCl3- was the predominant species in

2 mole/(kg H2O) (m) NaCl solutions at similar T-P conditions.

XAS studies of Mn complexing with halide ions at ambient conditions confirmed the

existence of a series of six-fold octahedral Mn(II) species [MnXn(H2O)(6-n)](2-n)+ (X = Cl and

Br; n=0, 1, 2) in solutions containing 0.10 – 4.95 m MnCl2, 0.10 – 5.2 m MnBr2 and 0.05 – 6

m MnBr2 (Beagley et al., 1991; Chen et al., 2005a). Currently, there is no XAS

characterization of the structure of Mn(II) chloroaqua complexes in aqueous solutions under

hydrothermal conditions, but Chen et al. (2005b) reported a structural transition of Mn(II)

bromoaqua complexes from octahedral at room-T, 1 bar to tetrahedral under supercritical

conditions (400 ˚C, 310 bar, 0.4 - 1.2 m Brtot). The highest order complex identified was

MnBr3(H2O)-, similarly to the MnCl3(H2O)- complex found in concentrated 2 m NaCl

solutions at high temperature (Uchida et al., 1995, 2003; Uchida and Tsutsui, 2000).

APPENDIX A

177

Table 1 Previous studies of aqueous Mn(II) speciation under ambient and hydrothermal

conditions.

T, P and composition Species identified Reference

Solubility

400 – 700 °C at 1 and 2 kbar, Cl:Mn ratio = 2 – 7 MnCl2(aq) Boctor,1985

25 – 275 °C at Psat, 4 < pH < 10 Mn2+, MnHCO3+, MnCO3(aq),

Mn(OH)CO3, Mn(OH)2(aq) Wolfram and Krupp, 1996

25 – 300 °C at Psat, 0.01-6.0 m HCl and Cl:Mn ratio ≥ 2 Mn2+, MnCl+, MnCl2(aq)

Gammons and Seward, 1996

400 – 600 °C at 1 kbar, 2 m NaCl MnCl3- Uchida et al., 1995

300 – 800 °C at 1 kbar, 0.5 – 1 kbar at 600 °C, 2 m NaCl MnCl2(aq), MnCl3

- Uchida and Tsutsui, 2000

300 – 800 °C at 1 kbar, 2 m NaCl MnCl2(aq), MnCl3- Uchida et al., 2003

Electron Spin Resonance spectroscopy

20 °C at 1 bar, 50 ppb – 15 ppm Mn in marine and lacustrine pore waters

Mn2+, MnCl+, MnSO4(aq), MnCO3(aq), Mn(HCO3)+ Carpenter, 1983

50 – 170 °C at 1 – 100 bar, & 1.0 m Cl Mn2+, MnCl+, MnSO4(aq) Wheat and Carpenter, 1988

Potentiometry and UV spectroscopy

25 °C at 1 bar, 0.01m Cl Mn2+, MnCl+ Libu+ and Tialowska, 1975

UV-Vis spectroscopy

25 – 250 °C at Psat, 0.9 m Mn(ClO4)2 with 0.01 m HClO4

Mn(H2O)62+ Koplitz et al., 1994

25 – 300 °C at Psat, 0.022 – 0.80 m Cl Mn2+, MnCl+, MnCl2(aq)

Suleimenov and Seward, 2000

XAS spectroscopy Ambient conditions, 0.1 m (n-

C4H9)NXO4 (X=Cl, Br) in hexamethylphosphoric triamide

solution

MnCl+, MnCl2(aq), MnCl3-, MnBr+,

MnBr2(aq), MnClBr(aq), MnCl2Br- Ozutsumi et al., 1994

Ambient conditions, 0.10 – 4.95 m MnCl2, 0.10 – 5.2 m MnBr2

Mn(H2O)62+, MnCl(H2O)5

+, MnCl2(H2O)4(aq), MnBr(H2O)5

+, MnBr2(H2O)4(aq)

Beagley et al., 1991

Ambient conditions, 0.05 – 6 m MnBr2

Mn(H2O)62+, MnBr(H2O)5

+, MnBr2(H2O)4(aq)

Chen et al., 2005a

25 – 400 °C at 1 – 310 bar, 0.4 – 1.2 m Br

and Cl:Mn ratio = 2 – 6

Mn(H2O)62+, MnBr(H2O)3

+, MnBr2(H2O)2(aq), MnBr3(H2O)- Chen et al., 2005b

This study provides a molecular-level understanding of Mn(II) chloride complexing in

chloride solutions (0.100 m & Cltot & 10.344 m) by in situ XAS analysis over a wide

APPENDIX A

178

temperature range (30 to 550 ˚C, 600 bar). The experimental results are supported by

Density Functional Theory (DFT) calculations. This study aims to: i) map the effects of

temperature and salinity on the coordination changes of Mn(II) chloroaqua complexes;

ii) identify whether high order tetrahedral Mn(II) chloroaqua complexes exist at elevated

temperatures and determine the structure (stoichiometry and geometry) of the predominant

Mn(II) complexes by EXAFS refinements, DFT calculations and ab initio XANES

simulations; iii) establish a Mn(II) speciation model that is consistent with the available

experimental data (e.g., Suleimenov and Seward, 2000).

2 Materials and methods

2.1 Experimental samples

Millipore® doubly deionized water and analytical grade chemicals (MnCl2•4H2O(s),

MnBr2•4H2O(s), NaCl(s), NaBr(s), HCl(aq), HBr(aq), LiCl(s)) purchased from Sigma-Aldrich

were used without further treatment. All sample solutions were prepared by dissolving

MnCl2•4H2O(s) or MnBr2•4H2O(s) in Millipore® doubly deionized water, slightly acidified

with HCl/HBr to prevent hydrolysis of Mn2+ (Table 2). Accurately weighed amounts of

NaCl(s)/NaBr(s) were added to prepare a series of solutions with the desired chloride/bromide

concentration. LiCl(s) was used to achieve the highest chloride concentration (solution 8,

10.344 m Cltot). Distribution of species calculations performed using the HCh package

(Shvarov and Bastrakov, 1999) confirm that the hydrolysis of Mn2+ can be neglected

throughout the temperature range in this study based on the thermodynamic properties of

Mn(OH)+ from Shock et al. (1997) and of Mn(OH)2(aq) from Wolfram and Krupp (1996).

APPENDIX A

179

Table 2 List of sample solutions.

Solution No. Mn [m] NaCl [m] HCl [m] Total Cl [m] Cl:Mn molar ratio

1 0.050 0 0.0003 0.100 2.00

11 0.050 0.054 0.0026 0.157 3.14

6 0.050 0.154 0.0026 0.257 5.14

13 0.051 0.411 0.0026 0.516 10.12

7 0.053 0.941 0.0027 1.050 19.81

12 0.056 2.001 0.0043 2.117 37.80

9 0.058 2.959 0.0028 3.078 53.07

14 0.061 3.982 0.0028 4.107 67.33

3 0.064 5.000 0.0029 5.131 80.17

Solution No. Mn [m] LiCl [m] HCl [m] Total Cl [m] Cl:Mn molar ratio

8 0.141 10.045 0.0167 10.344 73.36

Solution No. Mn [m] NaBr [m] HBr [m] Total Br [m] Br:Mn molar ratio

5 0.054 0 0.0016 0.110 2.04

15 0.105 0.990 0.0043 1.204 11.47

4 0.063 1.997 0.0016 2.125 33.73

The oxygen- and moisture-sensitive solid compound tetraethylammonium

tetrachloromanganate(II), (NEt4)2MnCl4(s), which contains tetrahedral [MnCl4] moieties

(Cotton et al., 1962; Mahoui et al. 1996), was synthesized for use as a solid standard. Details

of the synthesis procedures, and crystal structure solution and refinement are given in the

Supplementary Information (SI). For standards containing Mn in octahedral coordination,

rhodochrosite (MnCO3) from the Sweet Home Mine, Colorado (South Australian Museum

collection number G9806; [MnO6] moieties; Maslen et al., 1995) and MnCl2•4H2O(s)

(Sigma-Aldrich; [MnCl2(H2O)4] moieties; Zalkin et al., 1964; El Saffar and Brown, 1971)

were used. The identity of these standards was confirmed via X-ray powder diffraction.

APPENDIX A

180

2.2 XAS Measurements

XAS measurements were conducted at beamline BM-30B (FAME) at the European

Synchrotron Research Facility (ESRF) (see SI for details). The spectroscopic cell developed

by the Laboratoire de Cristallographie, CNRS, Grenoble was used for the XAS

measurements (Testemale et al., 2005). The autoclave was equipped with three 1.5 mm

thick beryllium windows that allow collection of fluorescence and transmission signals

concurrently up to a pressure of 750 bar. The sample solution was enclosed inside a glassy

carbon tube, and the pressure was transmitted to the sample by two glassy carbon pistons

using He as a medium. The Mn contents of the high purity beryllium (Brush Wellman grade

PF60) were & 0.01 wt%, and baseline XAS data show that the Mn signals from the Be

windows and scattering off the autoclave components are negligible.

The solution temperature was calibrated using XAS to measure the density of water as a

function of the thermocouple temperature at 600 bar (Etschmann et al., 2010; Borg et al.,

2012). The mass attenuation coefficients from Chantler (1995) (8.55 keV) and the water

densities tabulated in the NIST database (Lemmon et al., 2000) were used in the

calculations. The temperatures used throughout this paper are the calibrated temperatures,

with estimated precisions better than ± 5 ˚C.

2.3 Extended X-ray Absorption Fine-Structure (EXAFS) data analysis

EXAFS fluorescence data were analyzed by the HORAE package (Ravel and Newville,

2005), using the procedure outlined in Etschmann et al. (2010). EXAFS calculations were

performed using FEFF8 (Ankudinov et al., 1998). Based on the FLUO

(http://www.aps.anl.gov/~haskel/fluo.html) and BOOTH (Booth and Bridges, 2003)

algorithms, self-absorption is negligible, and therefore raw data were used in the XANES

and EXAFS analysis. A ‘spike’ at about 6596 eV (~3.6 Å-1 in k-space), present on all Mn(II)

spectra (feature F in Figures 1-3), is attributed to a multielectron excitation (KMII,III

APPENDIX A

181

transition; Chen et al., 2005a). Removal of this feature had negligible effect on the EXAFS

refinements, thus the raw spectra were analyzed without further correction. All EXAFS

refinements were performed based on k2-weighted data in R-space using a Hanning window.

2.4 Density Functional Theory (DFT) calculations

DFT calculations were performed using the Amsterdam Density Functional program,

ADF2010.02 (Te Velde et al., 2001) to optimize the geometries for the proposed Mn(II)

complexes. The basis functions are localized Slater-type orbitals. The Vosko-Wilk-Nusair

parameterization (Vosko et al., 1980) was used for Local Density Approximation, the

exchange functional of Perdew-Burke-Ernzerhof (Perdew et al., 1996) generalized gradient

approximation was used to calculate the exchange-correlation energy (Parr and Yang, 1989).

For all atoms, an uncontracted triple-zeta basis set was used with polarization functions

(Van Lenthe and Baerends, 2003). Frozen core orbitals (Mn: 1s, 2s, 2p; Cl: 1s, 2s, 2p; Br: 1s,

2s, 2p, 3s, 3p and O: 1s) were applied during the calculations to reduce the computational

time (Peacock and Sherman, 2004; Te Velde et al., 2001). All calculations were done using

the spin-unrestricted formalism to account for the five unpaired 3d-electrons of Mn(II)

(Sherman, 2009). The Conductor-like Screening model (COSMO) (Klamt and Schuurmann,

1993; Pye and Ziegler, 1999) was employed to account for long-range solvation in aqueous

solutions under hydrothermal conditions (Sherman, 2007; 2010; Tossell 2012). In COSMO

models, each atom is surrounded by a sphere of radius Ra = 1.17 $ RavDW (where Ra

vDW is the

van der Waals radius of the atom) as optimized by Klamt et al. (1998) in order to define a

molecular cavity; radii of 2.1 Å for Mn, 1.72 Å for O, 1.3 Å for H, and 2.05 Å for Cl and Br

were used. Outside of this cavity, the solvent is represented as a dielectric continuum. A

dielectric constant of 10 and a solvent radius of 1.3 Å were applied to represent

hydrothermal fluids at 500 ˚C, 600 bar (Fernandez et al., 1997).

APPENDIX A

182

2.5 Ab initio X-ray Absorption Near Edge Structure (XANES)

simulations

Ab initio XANES simulations were conducted using the FDMNES program (Joly, 2001)

following the method described in recent studies of metal complexing (Brugger et al., 2007;

Testemale et al., 2009; Etschmann et al., 2010; Liu et al., 2011, 2012; Borg et al., 2012;

Tian et al., 2012; Tooth et al., 2012). FDMNES calculates the photo-absorption cross

section using two different methods: i) multiple scattering theory, which uses the Muffin

Tin (MT) approximation to evaluate the inter-atomic potentials; and ii) the Finite Difference

Method (FDM), which avoids the limitations of the MT approximation by allowing a free

potential shape (Joly, 2001). To compare to the experimental spectra, the calculated raw

data were convoluted to account for the broadening of the features caused by experimental

resolution, core-hole life-time, inelastic plasmon interactions with the photoelectron, and

Fermi energy. Details of convolution parameters are provided in the SI.

3 Qualitative analysis of XAS spectra

3.1 Effect of temperature

The Mn K-edge XAS spectra of the two solutions containing the lowest and highest Cl

concentrations (S1: 0.10 m Cl and S8: 10.34 m Cl) are shown in Figure 1 as a function of

temperature at 600 bar. With increasing temperature, the changes in XANES spectra include:

i) the intensity of the pre-edge (feature A) increases; ii) the shoulder (feature B) at the lower

energy side of the white line becomes more prominent; iii) the intensity of the white line

(feature C) decreases and its position shifts towards lower energy; and iv) reduced intensity

of the oscillation in the 6570 to 6610 eV range (features E and F). This evolution is similar

to that observed for Mn2+ in bromide solutions (Chen et al., 2005a, b) and for some other

divalent transition metals in halide solutions (e.g., Fe2+ in chloride solutions, Testemale et

al., 2009; Co2+ in chloride and bromide solutions, Liu et al., 2011; Ni2+ in chloride and

APPENDIX A

183

bromide solutions, Hoffman et al., 1999, Tian et al., 2012; Cd2+ in chloride solutions,

Barzakina et al., 2010), and reflects a structural transition from octahedral to tetrahedral.

Figure 1 XANES spectra of S1 (0.100 m Cl) and S8 (10.344 m Cl) as a function of

temperature from 30 to 550 oC (a and b). The growth of the pre-edge peak at about 6540 eV

is shown in the insets upon heating. EXAFS spectra in k-space of S1 and S8 with increasing

temperature (c), and R-space spectra of of S1 and S8 at two extreme temperatures (30 and

550 oC) compared with those of Mn solid reference compounds – MnCO3(s) and

(NEt4)2MnCl4(s).

APPENDIX A

184

The transition from octahedral to tetrahedral–like structures occurs at higher temperatures

for S1 than for S8 (Figure 1a). For solution S1, only small spectral changes can be observed

in the XANES and EXAFS spectra from 25 ˚C to 200 ˚C (Figure 1a, c), implying that

octahedral complexes are largely dominant over this temperature range. Spectra of S1

change dramatically upon heating from 200 to 500 ˚C, but show only subtle systematic

changes upon further heating to 550 ˚C. In contrast, the spectra of solution S8 change

rapidly from 30 to 300 ˚C, but evolve slowly at temperatures above 300 ˚C (Figure 1b, d),

which suggests that Mn speciation in this solution is dominated by a single species at

T > 300 ˚C. The Fourier-transformed spectra reflect the radial distribution function around

Mn2+ ions, and the main peaks for MnCO3(s) and (NEt4)2MnCl4(s) correspond to six Mn-O

and four Mn-Cl bonds, respectively (Figure 1d). The centroids of the main peaks of the S1

and S8 solutions are close to those of MnCO3(s) at 30 ˚C, and move to larger R values at

550 ˚C, close to (NEt4)2MnCl4(s). The increase of bond distances to neighboring ligands

upon heating is related to a chlorination process, which results in the replacement of H2O by

Cl- in the first shell of Mn2+ associated with a structural transition from six-coordination to

four-coordination.

The XANES pre-edge feature is strongly indicative of the local geometry of Mn(II) (Chen et

al., 2005a, b; Farges, 2005; Chalmin et al., 2009). The pre-edge peak is due to the 1s,3d

transition that is forbidden in symmetrical sites; a small pre-edge in octahedrally

coordinated environments is related to a weak electric quadrupole caused by distortion and

p-d hybridization (e.g., MnCl2•4H2O(s); feature A in Figure 2a; Shulman et al., 1976; Westre

et al., 1997). Intense pre-edges are characteristic of electric dipole transitions, such as those

allowed in tetrahedral symmetry (de Groot et al., 2009; Shulman et al., 1976; Westre et al.,

1997; Yamamoto, 2008). Hence the pre-edge data are consistent with octahedral complexes

APPENDIX A

185

dominating at low temperature (Figure 2a) and tetrahedral complexes at high temperature

(Figure 2c).

Figure 2 XANES spectra of Mn(II) chloride solutions as a function of salinity from 0.100 to

10.344 m Cl at (a) 100 oC, (b) 300 oC and (c) 500 oC,. The growth of pre-edge peak at

around 6540 eV with increasing chloride concentration are shown in insets. XANES spectra

of Mn reference compounds – MnCl2.4H2O(s) and (NEt4)2MnCl4(s) are also shown in (a) and

(c) for comparison. All features are labelled in (b,c), same labels apply to (a).

APPENDIX A

186

3.2 Effect of salinity

Mn K-edge XANES spectra are shown as a function of salinity at fixed temperatures (100,

300 and 500 ˚C in Figure 2). The effect of increasing salinity from 3.078 to 10.344 m at

100 ˚C is small (Figure 2a): it results in a slight decrease of white line intensity, an energy

shift (1.6 eV) of the white line peak to lower energy and the growth of a band at ~6575 eV.

The XANES spectra of the highest salinity solution (S8) resemble that of MnCl2•4H2O(s)

(Figure 2a), containing octahedral MnCl2(H2O)4 moieties. Hence, octahedral species

dominate Mn speciation at low temperature throughout the whole salinity range (Farges,

2005; Chalmin et al., 2009).

The effect of salinity is more pronounced at elevated temperatures, in particular at 300 ˚C

(Figure 2b). S1 has a typical octahedral spectrum, while the spectrum of S8 is characteristic

of tetrahedral-like coordination. At 500 ˚C, Mn exists in tetrahedral-like complexes, with

XANES spectra similar to that of (NEt4)2MnCl4(s) (Figure 2c), and no spectral changes occur

upon further heating to 550 ˚C (Figures 1a, b). At ( 500 ˚C, the solutions can be categorized

into two groups based on the shape of the white line (feature C) and intensities of the

features B and E: the low salinity (S1, S11, and S6) and high salinity groups (S14, S3, and

S8). In particular, feature E is strongest for (NEt4)2MnCl4(s), is significant for solutions of

the high salinity group, but is absent for the low salinity group (Figure 2c). These

differences suggest that two tetrahedral species with a different ratio of water to chloride

ligands dominate Mn(II) speciation in low and high salinity solutions at high temperature. A

similar situation was found for Co(II) chlorocomplexes at 440 ˚C, with tetrahedral

CoCl2(H2O)2(aq) and CoCl42- complexes dominating in low and high salinity solutions,

respectively (Liu et al., 2011).

APPENDIX A

187

3.3 XANES spectra of bromide solutions

The spectral evolution of the Mn(II) bromide solutions follows a similar trend to that of the

Mn-Cl solutions (Figure 3). Two isosbestic points at ~6565 and 6583 eV in the XANES

spectra of both chloride and bromide solutions support the structural transition from

octahedral to tetrahedral-like as a function of both temperature and salinity (Figures 1-3).

The structural transition for Mn(II) complexes occurs at higher temperature in Br than in Cl

solutions for similar halide concentrations, as it did for Co(II) (Liu et al., 2011). At 400 ˚C,

Linear Combination Fits (LCF) in Athena, (Ravel and Newville, 2005; 6519 to 6589 eV)

show that in the dilute solutions S5 (0.110 m Br) and S1 (0.100 m Cl), tetrahedral species

account for 57% and 71%, respectively; in the more concentrated solutions S4 (2.125 m Br)

and S12 (2.117 m Cl), tetrahedral species account for 63% and 80%, respectively. Note that

at 500 ˚C the solutions with the lowest and highest Br contents (0.110 - 2.125 m Br) show

only slight differences and a similar feature E (Figure 3c).

APPENDIX A

188

Figure 3 XANES spectra of S5 (0.110 m Br) and S4 (2.125 m Br) as a function of

temperature from 30 to 500 oC (a and b). The representative individual XANES spectra are

shown in (c) with the pre-edge region shown in the inset.

4 EXAFS refinements

EXAFS analysis constrained the coordination number and bond distances for Mn(II)

chloride and bromide complexes. An S02 value of 0.68 was obtained by fitting the EXAFS

data of (NEt4)2MnCl4(s) and solution S1 (0.050 m MnCl2) at 30 ˚C and 1 bar (SI, section

APPENDIX A

189

4.1). This value is similar to that of 0.72 used in the study of Mn-bromide complexing by

Chen et al. (2005a, b), and was used in all subsequent EXAFS refinements.

The -E0 parameter accounts for a misalignment of energy between theoretical calculations

and experiments (Kelly et al., 2008). Metal-ligand distances are strongly correlated with

-E0 (Bunker, 2010), and in some cases assigning different -E0 to different ligands can

result in more physically significant fits (Kelly et al., 2008). In this case, using a single -E0

to fit the 500 ˚C data for S1 (0.100 m Cl), assuming a coordination number of four

(Section 3), resulted in poor agreement (R-factor > 0.50, an order of magnitude higher than

refinements reported in Table 3) and in a Mn-O distance of 2.27 Å, far longer than the room

temperature octahedral Mn-O distance of 2.16 Å, but close to the tetrahedral Mn-Cl distance

of 2.31 Å, suggesting that O and Cl are interchangeable during the fitting. Therefore,

different -E0 values were used for the Mn-Cl and Mn-O scattering shells. In order to limit

the number of fit parameters, the -E0 values were determined from fitting the standards

(NEt4)2MnCl4(s) [-E0(Mn-Cl) = 6.1(7)] and Mn(H2O)62+ [solution S1 at 30 ˚C, 1 bar;

-E0(Mn-O) = -0.9(7)]. The two -E0 values were allowed to vary within the error range

obtained on the standards for each individual fit; a similar approach was used by Mayanovic

et al. (2002, 2009).

4.1 Mn-Cl solutions

4.1.1 Solutions with one dominant coordination geometry

XANES analysis showed that octahedral complexes dominate Mn(II) speciation at low-T

and tetrahedral complexes at high-T (Figure 2). The EXAFS fit of S1 at 30 ˚C (Table 3 and

Figure 4) confirms that the octahedral Mn(H2O)62+ complex (6.1(4) water) is the

predominant species in chloride-poor solutions (e.g., Chen et al., 2005a), and the number of

water ligands for solutions containing 0.100 – 3.078 m Cl at 30 ˚C were around 6 within

error (SI, Table S4). Attempts to add one chloride into the first octahedral shell resulted in a

APPENDIX A

190

statistically worse fit (i.e. the change in reduced $2 for the two fits is greater than two

standard deviations, as defined by Kelly et al., (2008; equation 19). The refined Mn-O bond

distance of 2.16(1) Å is in good agreement with the literature value of 2.169(7) Å (Chen et

al., 2005a) for the Mn(II) hexaaqua complex in a 0.05 m MnBr2 solution at 25 ˚C and 1 bar.

With increasing salinity, the presence of chloride within the octahedral complex was

detectable for solutions with chloride concentrations ( 4.107 m at 30 ˚C, ( 2.117 m at

100 ˚C, and all chloride concentrations at T ( 200 ˚C (SI). The refined Mn-O and Mn-Cl

bond lengths for solution S8 at 100 ˚C are 2.17(2) Å and 2.49(2) Å, respectively, in

excellent agreement with the averaged Mn-O lengths of 2.20(2) Å and Mn-Cl length of

2.49(2) Å of MnCl2•4H2O(s) (Zalkin et al., 1964; El Saffar and Brown, 1971) and of MnCl2

solutions (Beagley et al., 1991).

Figure 4 Experimental (blue dashed line) and fitted (red solid line) EXAFS spectra of S1

(0.100 m Cl) and S8 (10.344 m Cl) at 30 ˚C and 500 ˚C in (a) k-space and (b) R-space. All

fits are shown with k2 weighting.

APPENDIX A

191

Table 3 Summary of EXAFS refinements of selected Mn chloride and bromide solutions at

extreme temperatures at 600 bar.

Mn-O interaction Mn-Cl interaction T

(˚C) !E0 (eV) NO RMn-O

(Å) "O

2(Å2#10-

3) !E0 (eV) NCl RMn-Cl

(Å) "Cl

2(Å2#10-

3)

R-range k-range R-

factor $2

S1, 0.100 m Cl, Cl:Mn molar ratio = 2.00

30 -1.0 6.1(4) 2.16(1) 4.6(1.0) - - - - 1-4.3 2.2-11 0.016 16.84

100 -1.6 6.2(4) 2.15(1) 6.5(1.3) - - - - 1-4.3 2.2-10.5 0.019 14.64

500 -1.6 2.1(3) 2.13(5) 9.6(6.2) 6.8 1.9(3) 2.33(1) 7.4(3.0) 1-3 2.2-11 0.015 17.94

550 -1.6 2.0(6) 2.15(10) 13(16) 6.8 2.0(6) 2.32(2) 6.2(4.2) 1-3 2.2-11 0.034 10.72


30 -1.1 5.1(4) 2.16(1) 5.6(1.1) 6.8 0.9(4) 2.53(3) 7(5) 1-4.5 2.2-11 0.022 12.18

500 -1.6 1.0(4) 2.14(10) 5.9(1.0)* 6.5 3.0(4) 2.37(2) 5.9(1.0)* 1-3 2.2-10 0.013 8.27


30 -0.3 4.4(5) 2.18(2) 6.1(1.3) 5.4 1.6(5) 2.51(2) 5.5(2.8) 1-4.7 2.2-11 0.024 12.28

100 -1.6 4.0(4) 2.17(2) 7.5(1.4) 5.4 2.0(4) 2.49(2) 7.9(2.3) 1-4.3 2.2-11 0.019 14.51

500 -1.6 1.0(3) 2.13(5) 7.3(1.0)* 6.2 3.0(3) 2.37(1) 7.3(1.0)* 1-3 2.2-11 0.016 16.13

550 -1.6 1.0(4) 2.12(10) 8.1(1.0)* 6.2 3.0(4) 2.37(2) 8.1(1.0)* 1-3 2.2-11 0.013 12.32

S5, 0.110 m Br, Br:Mn molar ratio = 2.04

30 -1.7(8) 6.4(5) 2.15(1) 5.1(1.3) -1.7(8) † - - - 1-4.3 2.2-11.0 0.028 10.57

500 -1.8(2.6) 2.2(1.1) 2.14(3) 16(12) -1.8(2.6) † 1.7(7) 2.43(1) 6.1(3.4) 1-3 2.2-10.5 0.026 5.52

S4, 2.125 m Br, Br:Mn molar ratio = 33.73

30 0.8(1.3) 5.6(8) 2.18(2) 6(3)* 0.8(1.3) † 0.2(4) 2.40(21) 6(3)* 1-4.3 2.2-9.8 0.039 14.99

500 -3.3(2.9) 1.3(5) 2.14(4) 9.8(2.9)* -3.3(2.9) † 3.0(1.1) 2.43(2) 9.8(2.9)* 1-3 2.2-10.5 0.059 5.96

* Debye-waller factor set to be identical for Mn-O and Mn-Cl interactions.

† -E0 set to be identical for Mn-O and Mn-Br interactions.

Please see detailed EXAFS refinements of all sample solutions from 30 to 550˚C in the supplementary

information.

The coordination number of Mn was constrained to be four for all EXAFS refinements of

solutions at 500 ˚C and 550 ˚C (Section 3). For the weakly saline solutions (0.100 & mCltot &

0.257), the numbers of water and chloride were both around two, revealing predominance of

the MnCl2(H2O)2(aq) complex. At high salinity, the best fit was MnCl3.0(4)O1.0(4) for all

solutions with ≥5.131 m Cltot (Table 3), consistent with the MnBr3.4(1.6)O0.9(4) average

ligation identified by Chen et al. (2005b). The refined Mn-Cl bond length was 2.37(1) Å for

solution S8 at 500 ˚C (Figure 4; Table 3), identical to that of 2.37(1) Å for (NEt4)2MnCl4(s);

the fitted Mn-O bond length was 2.13(5) Å, within error of the room-T bonds length of

APPENDIX A

192

2.18(2) Å for S8 and 2.16(1) Å for S1. EXAFS refinement using a MnCl4 model resulted in

a statistically worse fit.

4.1.2 Solutions with mixed octahedral and tetrahedral complexes

Based on the XANES data, most chloride solutions contain mixtures of octahedral and

tetrahedral complexes at intermediate temperatures (200 – 450 ˚C). As each complex has

different Mn-O and Mn-Cl distances, Cl/O ratio, and Debye-Waller factors, parameter

under-determination is a problem. To reduce the number of fit parameters, EXAFS data at

intermediate temperatures for each solution were analyzed assuming that they consist of a

mixture of the species that exist at low (100 ˚C; octahedral coordination) and high

temperatures (500 ˚C; tetrahedral coordination). For each solution, the fraction of tetrahedral

and octahedral species was derived from XANES data via LCF (SI, section 4.2, Table S5).

These fractions were fixed in the EXAFS refinements and the Mn-Cl and Mn-O bond

lengths and corresponding Debye-Waller factors were fitted, and represent average values

of octahedral and tetrahedral complexes in each solution. A summary of EXAFS

refinements of all Mn chloride solutions (30 – 550 ˚C) can be seen in the SI (Table S4).

A contraction of the Mn-Cl bond length with increasing temperature for each solution is

noticeable. For example, the representative octahedral Mn-Cl bond length of 2.51(2) Å at

30 ˚C is around 0.14 Å longer than a typical tetrahedral Mn-Cl bond length of 2.37(2) Å at

500 ˚C (Solution S8, Table 3). The Mn-O bond distances of low salinity solutions (0.100 –

0.516 m) remain constant (2.16(1) Å) from 30 to 300 ˚C. This is inconsistent with the ~3%

increase in the Mn-O bond length for Mn(H2O)62+ (Cl-free perchlorate solutions, 25-250 ˚C)

interpreted by Koplitz et al. (1994) on the basis of UV-Vis spectroscopy. This discrepancy

may result from the omission of small amounts of tetrahedral complexes at high temperature

in Koplitz et al.’s analysis. For most other solutions there seems to be a small contraction of

APPENDIX A

193

the Mn-O bond distances from room-T to 500 ˚C (~0.05 Å), although this is not statistically

significant due to the relatively large errors at high temperature. The changes in O and Cl

ligation numbers as a function of temperature further illustrates the coordination change

from octahedral at 30 and 100 ˚C to tetrahedral at 450 and 500 ˚C for both S11 and S14,

with the structural transition happening at higher temperature (300–450 ˚C) for S11

compared to 200–400 ˚C for S14 (Figures 5a, b). At 300 and 400 ˚C, increasing salinity

causes a decrease in coordination number and an increase in the number of Cl ligands

(Figure 5c, d).

Figure 5 Number of oxygen (water molecule; red circles with error bars) and chloride (blue

triangles with error bars) ligands in the first shell of Mn2+ as a function of temperature for (a)

S11 (0.157 m Cl) and (b) S14 (4.107 m Cl) and as a function of chloride concentration (in

log scale) at (c) 300 ˚C and (d) 400 ˚C based on the EXAFS refinements (Table 3) and the

predictions (dashed lines) based upon the thermodynamic properties in Table 5.

APPENDIX A

194

4.2 Mn-Br solutions

Mn-Br EXAFS spectra were fitted following a similar strategy to the Mn-Cl solutions,

except for the use of a single -E0 for each refinement as it is easier to distinguish between

the Mn-O and Mn-Br scattering paths. The radial distribution function shows two distinct

peaks corresponding to the presence of O and Br within the first shell of Mn for solutions S5

(0.110 m Br) and S4 (2.125 m Br) at 500 ˚C (Figure 6b). The refined Mn-O bond length of

2.15(1) Å for S5 at 30 ˚C is identical to the bond lengths in the counterpart chloride

solutions, and with those refined by Chen et al. (2005a, b). At 500 ˚C, the best fit model of

MnBr3.0(1.1)O1.3(5) for solution S4 (2.125 m Br) is similar to the MnBr3.4(1.6)O0.9(4) model

identified by Chen et al. (2005b) at 400 ˚C and 310 bar (Table 3). The best fit model of

MnBr1.7(7)O2.2(1.1) for the low salinity solution S5 at 500 ˚C is also consistent with

MnBr1.9(2.4)O1.7(1.3) in a 0.2 m MnBr2 solution at 400 ˚C and 310 bar by Chen et al. (2005b).

The Mn-O and Mn-Br bond lengths refined in this study (Table 3) for tetrahedral Mn(II)

complexes are within error of those derived by Chen et al. (2005b). There is no obvious Mn-

O bond length contraction upon heating, but the Mn-Br bond length contraction expected

for a change from octahedral to tetrahedral coordination was observed: for solution S15, the

Mn-Br bond length of 2.58(5) Å) at 100 ˚C is 0.13 Å longer than the typical tetrahedral Mn-

Br bond length of 2.45(2) Å obtained at 500 ˚C.

APPENDIX A

195

Figure 6 Experimental (blue dashed line) and fitted (red solid line) EXAFS spectra of S5

(0.110 m Br) and S4 (2.125 m Br) at 30 ˚C and 500 ˚C in (a) k-space and (b) R-space. All

fits are shown with k2 weighting.

5 DFT calculations

The results of geometry optimization by DFT for the proposed Mn(II) complexes are

summarized in Table 4, with detailed structural information listed in the SI (Table S7). The

optimized Mn-O and Mn-Br bond distances were generally longer than the refined EXAFS

distances, while Mn-Cl distances were slightly shorter than the EXAFS distances (e.g.,

2.16 Å for Mn-O by EXAFS at 30 ˚C in 0.100 m Cl solution and 2.21 Å for DFT; 2.37 Å for

Mn-Cl by EXAFS at 500 ˚C in 10.344 m Cl solution and 2.34 Å for DFT). These

discrepancies are in-line with the expected accuracy of the DFT calculations (e.g., Bühl and

Kabrede 2006). The most important result from the DFT calculations is that the trends in the

variation of the experimental bond lengths are accurately reproduced: a small contraction

APPENDIX A

196

occurs when geometry changes from octahedral to tetrahedral, and the Mn-O, Mn-Cl and

Mn-Br bond lengths increase with increasing number of halide ligands within the same

coordination geometry.

Table 4 Bond distances for the manganese (II) clusters optimized by DFT. Where several

identical ligands are present, the value represents the average distance, with the standard

deviation given in parentheses.

Mn(II) Cluster Geometry Mn-O (Å) Mn-Cl (Å)

Mn(H2O)62+ Octahedral 2.21(2) -

MnCl(H2O)5+ Octahedral 2.26(1) 2.366

MnCl2(H2O)4 (aq) (cis) Octahedral 2.30(2) 2.429(2) MnCl2(H2O)4 (aq) (trans) Octahedral 2.287(6) 2.44(4)

MnCl(H2O)3+ Tetrahedral 2.129(1) 2.255

MnCl2(H2O)2 (aq) Tetrahedral 2.178(8) 2.291(1)

MnCl3(H2O)- Tetrahedral 2.232 2.338(5)

MnCl42- Tetrahedral - 2.381

MnBr(H2O)5+ Octahedral 2.25(1) 2.549

MnBr2(H2O)4 (aq) (cis) Octahedral 2.29(2) 2.599(4)

MnBr2(H2O)4 (aq) (trans) Octahedral 2.264(8) 2.64(13)

MnBr(H2O)3+ Tetrahedral 2.129(1) 2.414

MnBr2(H2O)2 (aq) Tetrahedral 2.17(2) 2.458(2)

MnBr3(H2O)- Tetrahedral 2.221 2.501(7)

MnBr42- Tetrahedral - 2.549

6 Ab initio XANES simulations

6.1 Simulations for solid standards

Calculations for Mn(II) solid standards were performed to assess the ability of XANES

simulations to reproduce the spectral features of known Mn(II) structures (Figure 7). The

compounds contain octahedral and tetrahedral moieties: octahedral [MnO6] in MnCO3(s),

APPENDIX A

197

octahedral [MnCl2(H2O)4] in MnCl2•4H2O(s), and tetrahedral [MnCl4] in (NEt4)2MnCl4(s).

For MnCO3(s), calculation with a radius of 6 Å using the FDM method reproduced the key

spectral features well (Figure 7a). Note that the raw spectrum (dotted line) reproduced well

the region of feature F, but the convolution process (solid line) smoothes out the sharp band

at ~6596 eV. This 6596 eV band is not due multielectron excitations, since multielectron

excitations are not taken into account in the calculations. For MnCl2•4H2O(s) and

(NEt4)2MnCl4(s), the calculated spectra with a radius of 3 Å using the FDM method was in

good agreement with the experimental spectrum except for a missing Feature B.

Calculations using larger radii (6 Å for MnCl2•4H2O(s) by FDM; 7 Å for (NEt4)2MnCl4(s) by

MT approximation) improved the agreement of feature B (Figure 7b, c). Therefore,

feature B for MnCl2•4H2O(s) and (NEt4)2MnCl4(s) is strongly affected by contributions from

beyond the first shell. Feature B is related to the 1s,4p electronic transition (Chen et al.,

2005a, b), and hence reflects a strong interaction between metal cation and the surrounding

ligands. Overall, the agreement for the standards inspires confidence that ab initio XANES

simulations can be used to explore the structure of Mn(II) in solution.

APPENDIX A

198

Figure 7 Experimental and calculated (both raw and convoluted) XANES spectra for Mn

model compounds: (a) MnCO3(s), (b) MnCl2.4H2O(s), and (c) (NEt4)2MnCl4(s). Open circles

are experimental spectra; the calculated raw spectra (‘raw’) and the convoluted spectra

(‘conv’) with different calculation radius using different methods (FDM: Finite Difference

Method, MS: multiple scattering theory) are shown. All features are labelled in (c), and

labels reemphasize only specific features in (a,b).

APPENDIX A

199

6.2 Simulations of aqueous complexes

The calculated XANES spectra based on DFT-optimized geometries for a series of

octahedral species with increasing Cl substitution exhibit a consistent trend with the spectral

evolution observed experimentally at low temperature with increasing salinity (Figure 8a):

the increase in pre-edge peak (inset in Figure 8a) and slight decrease in white line intensity,

the shift of white line to lower energy, and reduced oscillation of features D and E were

correctly reproduced. The inclusion of H atoms in the calculations affects mainly the

intensity of the pre-edge peak: the calculated pre-edge for the [Mn(H2O)6] moiety was over-

estimated compared to that for [MnO6] (Figure 8a). As these calculations were based on a

single rigid configuration and the H atoms are disordered (Testemale et al., 2009), the

overall effect of the H atoms on the XANES spectrum is expected to be small (e.g., Liu et

al., 2012). For tetrahedral species, increasing the number of Cl ligands from one to four

resulted in the following systematic changes in the calculated XANES spectra:

i) progressive growth of feature E, ii) decreased intensity of feature D, and iii) shifted

feature D to lower energy leading to a slimmer white line. These trends were consistent with

the evolution of experimental spectra at 500 ˚C when salinity increased from 0.100 to

10.344 m (Figures 2c and 8b). In particular, the intensity of feature E strongly depends on

the chloride ligation number of the predominant species: the more Cl- complexed to Mn2+,

the larger the bump is. Hence, comparison with the feature E for (NEt4)2MnCl4(s), which is

more intense than those of the aqueous solutions at 500 ˚C (Figure 2c), suggests that the Cl

ligation number of the aqueous solutions at 500 ˚C is less than four. In addition, the

intensity of features C and E for [MnCl4] are higher compared to the experimental spectrum

of S8 at 500 ˚C and feature C is broader and the intensity of feature E in [MnCl(H2O)3] did

not match that of the experimental spectrum of S1 at 500 ˚C (Figure 8b). Therefore, the best

agreement with the experimental spectra of S1 and S8 at 500 ˚C is obtained by the

APPENDIX A

200

calculated spectra of [MnCl2(H2O)2] and [MnCl3(H2O)], respectively. The XANES spectra

of the bromide solutions S5 (0.110 m Br) and S4 (2.125 m Br) at 500 ˚C were well

reproduced by the calculations for the [MnBr2(H2O)2] and [MnBr3(H2O)] clusters,

respectively, although the shape and intensity of feature C and E are less distinguishable

than in the counterpart Mn chloride complexes (Figure 8b). The main disagreement between

experimental and calculated spectra for tetrahedral complexes is a double peak for the white

line on the calculated spectra, instead of the experimental broad asymmetrical peak

(Figure 8b). The reason could be that no thermal and structural disorders were taken into

account for the XANES calculations; the absence of the second hydration shell may also

contribute to this discrepancy (Figure S3b in SI).

Figure 8 Experimental XANES spectra of some representative Mn solutions and calculated

XANES spectra of Mn-Cl-O clusters based on DFT-optimized structures (see text for

details). The calculated raw spectra were labelled as ‘raw’.

APPENDIX A

201

Overall, the XANES simulations confirm the result of the EXAFS analysis that stepwise

substitution of H2O by Cl- occurs within the octahedral coordination sphere around Mn2+

with increasing chlorinity at low temperatures (30 – 100 ˚C), and that MnCl2(H2O)2(aq) and

MnCl3(H2O)- predominate Mn speciation in low and high salinity solutions at elevated

temperatures (i.e., 450 – 500 ˚C).

7 Discussion: Mn(II) speciation in chloride brines

7.1 Nature of Mn(II) chlorocomplexes

The combination of XANES analysis (Section 3), EXAFS refinements (Section 4) and

DFT/ab initio XANES simulations (Sections 5/6) presents a complete picture of the

coordination changes of Mn(II) chloride complexes from ambient to supercritical conditions.

At low temperature, chloride anions replace some water groups within an octahedral

complex with increasing Cl- concentration. At 10.344 m Cltot, 1.6(5) Cl- at 30 ˚C and 2.0(5)

Cl- at 100 ˚C are bonded to Mn2+ (Table 3), indicating that a series of octahedral complexes

are present, e.g., MnCl(H2O)5+ and MnCl2(H2O)4(aq). Higher order octahedral complexes

such as MnCl3(H2O)3- and MnCl4(H2O)2

2- may exist in highly concentrated chloride

solutions from room-T to around 200 ˚C, but these complexes are minor species because

octahedral complexes transform to tetrahedral quickly upon heating above 100 ˚C and they

were not detected in our system (salinity: 0.100 – 10.344 m Cl) or in the UV-Vis

experiments of Suleimenov and Seward (2000). Thus, the octahedral complexes

Mn(H2O)62+, MnCl(H2O)5

+, and MnCl2(H2O)4(aq) were included in our final model.

The role of solvent on determining the coordination geometry of the Mn(II) halogenide

complexes was illustrated by Ozutsumi et al. (1994), who identified a series of five-fold

halogenated Mn(II) complexes in a hexamethylphosphoric triamide solution at room

temperature (dielectric constant ~30). In contrast, in Cl- or Br-bearing aqueous solutions,

APPENDIX A

202

Mn(II) complexes exhibit a coordination change from octahedral to tetrahedral upon heating

and/or increasing salinity. Chen et al. (2005b) identified a family of four-coordinated

species: MnBr(H2O)3+, MnBr2(H2O)2(aq) and MnBr3(H2O)-, but found no evidence for the

fully halogenated complex MnBr42-. Similarly, the highest order chloride complex we can

identify by combining XANES and EXAFS data was not MnCl42- but MnCl3(H2O)-. Hence,

the available data suggest that MnCl42- is only a minor species for Mn transport in brines up

to 10 m Cltot. Similarly, MnCl(H2O)3+ is only a minor species in very dilute solutions, and

Mn(II) speciation in Cl-bearing solutions at elevated temperatures (≥ 450 ˚C) consists

mainly of two tetrahedral complexes: MnCl2(H2O)2(aq) and MnCl3(H2O)- with the ratio of

the two species strongly depending on the total Cl- concentration and the Cl:Mn molar ratio.

Similar to the results for the Mn-Cl system, the EXAFS analysis for the Mn-Br system

confirm that: i) tetrahedral Mn(II) bromoaqua complexes dominate Mn speciation at 500 ˚C

and 600 bar; ii) MnBr2(H2O)2(aq) and MnBr3(H2O)- are the predominant species in solutions

with low and high Br concentrations, respectively; and iii) MnBr3(H2O)- is the highest order

Mn(II) bromocomplex that could be identified in all bromide solutions of this study.

Throughout all temperature and salinity ranges, no evidence of Mn hydroxide complexes

has been found and Mn speciation is controlled by halide complexes only, consistent with

speciation calculations conducted using the thermodynamic properties of Mn(OH)+ from

Shock et al (1997) and Mn(OH)2(aq) regressed from solubility data of Wolfram and Krupp

(1996) (Table 6).

7.2 Thermodynamic analysis

Normalized fluorescence XANES spectra of Mn(II) chloride solutions with a chlorinity

range of 0.100 - 5.131 m were used to derive formation constants for Mn(II) chloride

complexes following the method described in previous studies (Liu et al., 2007, 2011, 2012;

APPENDIX A

203

Testemale et al., 2009; Etchmann et al., 2011; Tian et al. 2012) using the BeerOz package

(Brugger 2007).

Three octahedral complexes and two tetrahedral complexes were included in the Mn

speciation model: Mn(H2O)62+, MnCl(H2O)5

+, MnCl2(H2O)4(aq), MnCl2(H2O)2(aq) and

MnCl3(H2O)-. In this analysis, we assume that the MnCl+ and MnCl2(aq) species identified in

previous studies at low temperatures (& 300 ˚C) have an octahedral geometry. Modified

Ryzhenko-Bryzgalin (MRB) parameters (Ryzhenko et al., 1985) for MnCl(H2O)5+ and

MnCl2(H2O)4(aq) were regressed using the optimC program (Shvarov, 2008) from the

following datasets and weighting scheme (Figures 10a, b; Table 6). For MnCl(H2O)5+, the

data derived from the solubility experiments of Gammons and Seward (1996) were given a

weighting of 0.5 at 25 ˚C and 50 ˚C, because of difficulties in equilibrating solubility

experiments at low temperature. The remaining data points (Libu+ and Tialowska 1975;

Carpenter 1983; Wheat and Carpenter 1988; Gammons and Seward, 1996; Suleimenov and

Seward, 2000) were given the weighting of 1.0. For MnCl2(H2O)4(aq), the UV-Vis data of

Suleimenov and Seward (2000) and the solubility data of Gammons and Seward (1996)

were given weightings of 1.0 and 0.8, respectively, to slightly emphasize the newest results.

The formation constants of NaCl(aq) and HCl(aq) were taken from Sverjensky et al. (1997)

and Tagirov et al. (2001), respectively.

As XANES simulations confirmed that the experimental XANES spectra of S1 at 30 ˚C, S3

at 100 ˚C, S8 at 100 ˚C, S1 and S8 at 500 ˚C were representative of the octahedral and

tetrahedral species Mn(H2O)62+, MnCl(H2O)5

+, MnCl2(H2O)4(aq), MnCl2(H2O)2(aq) and

MnCl3(H2O)-, respectively, the spectra of these species were fixed from their corresponding

experimental spectra. The idea is to limit the number of unknown parameters in the model

in order to retrieve formation constants for the two tetrahedral complexes MnCl2(H2O)2(aq)

and MnCl3(H2O)-. As tetrahedral complexes are unstable at low temperatures (30 and

APPENDIX A

204

100 ˚C), only the last four datasets (200 - 450 ˚C) were used for XANES fitting. The

formation constants for octahedral species were extrapolated from the MRB equation of

state, and fixed during the analysis. The refined formation constants for MnCl2(H2O)2(aq) and

MnCl3(H2O)- are listed in Table 5. The uncertainties in the fitted log K values were

estimated at the 90% confidence level based on residual maps using the F-distribution

factors given by Draper and Smith (1998).

Table 5 Formation constants of the aqueous species at 600 bar used in thermodynamic

analysis

Species 30 ˚C 100 ˚C 200 ˚C 300 ˚C 400 ˚C 450 ˚C References

HCl(aq) -0.78 -0.96 -0.33 0.79 2.33 3.51 Tagirov et al. (1997)

NaCl(aq) -0.83 -0.54 -0.04 0.61 1.57 2.45 Sverjensky et al. (1997)

MnCl(H2O)5+ 0.15 0.86 2.04 3.39 5.34 7.02

Libu+ and Tialowska (1975) Carpenter (1983)

Wheat and Carpenter (1988) Gammons and Seward (1996)

Suleimenov and Seward (2000)

MnCl2(H2O)4(aq) -5.66 -1.73 1.64 3.98 6.59 8.72 Gammons and Seward (1996) Suleimenov and Seward (2000)

MnCl2(H2O)2(aq) - - 1.04 (<1.55) 4.05(25) 7.52(15) 10.6(5) Fitted in this study

MnCl3(H2O)- - - 0.80 (<1.17) 3.76(25) 7.60(18) 11.2(3) Fitted in this study

The calculated Mn(II) speciation from 200 to 450 ˚C is shown in Figure 9. At 200 ˚C,

octahedral species predominate Mn speciation; at 300 ˚C tetrahedral species represent less

than 30% of all species from 0 to 0.8 m Cl, consistent with the findings of Suleimenov and

Seward (2000); at 400 ˚C octahedral species still persist in low salinity solutions but

tetrahedral species become predominant with increasing Cl concentration; and at 450 ˚C the

octahedral species are negligible, and only two tetrahedral species exist with the fraction of

APPENDIX A

205

each species depending on the total Cl concentration. The average ligation numbers

calculated using the properties in Table 5 are in good agreement with the EXAFS refined

stoichiometries (Figure 5), providing additional evidence that the refined formation

constants of MnCl2(H2O)2(aq) and MnCl3(H2O)- are valid.

Figure 9 Speciation as a function of chlorinity based on the refined thermodynamic

properties in Table 5 at (a) 200 °C, (b) 300 °C, (c) 400 and (d) 450 °C and 600 bar.

To further test the validity of the thermodynamic model presented here, we checked the

compatibility of the properties of MnCl2(H2O)2(aq) and MnCl3(H2O)- with the data obtained

by Boctor (1985) and Uchida and Tsutsui (2000) at higher temperatures and pressures. In

this analysis, we assume that the MnCl2(aq) and MnCl3- species identified in these two

solubility studies at 1 and 2 kbar from 400 ˚C to 800 ˚C have a tetrahedral geometry. The

equilibrium constants for the reaction MnSiO3(s) + HCl(aq) = MnCl2(aq) + SiO2(s) + H2O at 1

and 2 kbar from 400 to 700 ˚C (Boctor, 1985) were converted to cumulative formation

APPENDIX A

206

constants for MnCl2(aq) using the thermodynamic properties of MnSiO3(s) from Robie et al.

(1978), HCl(aq) from Tagirov et al (1997) and SiO2(s) from Helgeson et al. (1978).

Attempting to fit all the data using MRB or HKF resulted in unrealistic pressure parameters,

i.e. the extrapolations predicted that the speciation was much more pressure sensitive in the

0-1000 bar region than experiments suggest, making the predictions 1-2.5 orders of

magnitude higher than the experimental data at 1 kbar (Figure 10c). Thus, in the final

analysis, the 1 kbar dataset of Boctor (1985) was discarded, and MRB parameters of

MnCl2(H2O)2(aq) were regressed from the combination of our data (weighting of 1.0; Table 5)

and the 2 kbar solubility data from Boctor (1985) (weighting of 0.5 applied for data at 400 –

500 ˚C; and 1.0 for 550 – 700 ˚C). The 1 kbar equilibrium constants for the reaction

MnCl2(aq) + Cl- = MnCl3- (Uchida and Tsutsui 2000) were converted to cumulative

formation constants for MnCl3- by extrapolating the formation constants of MnCl2(aq) to 1

kbar at 400 – 600 ˚C using the newly regressed MRB parameters of MnCl2(H2O)2(aq) (Table

6). Similarly, MRB parameters of MnCl3(H2O)- were regressed from the combination of our

data (Table 5) and the 1 kbar solubility data from Uchida and Tsutsui (2000) by equal

weighting. The refined formation constants of MnCl3(H2O)- at 600 bar in this study are

consistent with the 1 kbar data derived from Uchida and Tsutsui (2000) (Figure 10d). The

refined MRB parameters of the four Mn species are listed in Table 6 and calculated

formation constants of MnCl2(H2O)2(aq) and MnCl3(H2O)- at 150 – 800 ˚C from water

saturated pressure (Psat) to 2 kbar are shown in Table 7. Overall, the speciation and

thermodynamic model of this study are not only self-consistent but also compatible with

previous studies (e.g., Boctor, 1985; Gammons and Seward 1996; Suleimenov and Seward

2000; Uchida and Tsutsui 2000).

APPENDIX A

207

Table 6. MRB parameters (A(zz/a) and B(zz/a)) for Mn(OH)2(aq), MnCl(H2O)5+,

MnCl2(H2O)4(aq), MnCl2(H2O)2(aq) and MnCl3(H2O)-.

Species log K(298K) A(zz/a) B(zz/a)

Mn(OH)2(aq) 9.450 1.806 -748.90

MnCl(H2O)5+ 0.242 1.276 -191.18

MnCl2(H2O)4(aq) -5.293 0.263 963.32

MnCl2(H2O)2(aq) -6.936 1.000 679.76

MnCl3(H2O)- -4.627 2.303 -295.96

Figure 10 Formation constants (log K) of (a) MnCl(H2O)5+ determined by experimental

data of Libu+ and Tialowska (1975), Carpenter (1983), Wheat and Carpenter (1988),

Gammons and Seward (1996) and Suleimenov and Seward (2000); (b) MnCl2(H2O)4(aq)

determined by experimental data of Gammons and Seward (1996) and Suleimenov and

Seward (2000); (c) MnCl2(H2O)2(aq) determined by this study and derived from 1 and 2 kbar

data of Boctor (1985); (d) MnCl3(H2O)- determined by this study and derived from the

1 kbar data of Uchida and Tsutsui (2000). The solid lines represent fits to these

experimental data using the modified Ryzhenko–Bryzgalin (MRB) model (Ryzhenko et al.,

1985) with MRB parameters listed in Table 6.

APPENDIX A

208

Table 7 Formation constants for MnCl2(H2O)2(aq) and MnCl3(H2O)- from 150 to 800 ˚C at

water saturated pressure (Psat), 500 bar, 1000 bar and 2000 bar based on the properties in

Table 6. Note: MnCl2(H2O)2(aq) and MnCl3(H2O)- are unstable at low temperatures, and only

data from 150 to 800 ˚C are listed.

T (˚C) Psat 500 bar 1000 bar 2000 bar

MnCl2(H2O)2(aq)

150 -0.31 -0.79 -1.20 -1.87

200 1.58 1.00 0.54 -0.18

250 3.36 2.60 2.02 1.20

300 5.32 4.15 3.37 2.38

350 8.27 5.88 4.68 3.44

400 - 8.23 6.05 4.42

450 - 12.44 7.56 5.36

500 - - 9.27 6.28

600 - - 12.81 8.06

700 - - - 9.69

800 - - - 11.08

MnCl3(H2O)-

150 -0.44 -0.74 -0.99 -1.40

200 1.18 0.78 0.46 -0.03

250 2.89 2.31 1.88 1.26

300 4.89 3.93 3.29 2.49

350 7.91 5.83 4.78 3.69

400 - 8.43 6.41 4.90

450 - 13.03 8.26 6.12

500 - - 10.42 7.37

600 - - 15.19 9.94

700 - - - 12.49

800 - - - 14.85

7.3 Comparison with Fe(II) chloride complexing Like other first row divalent transition metals (e.g., Fe, Co, Ni, Zn, Cu; Collings et al. 2000;

Brugger et al. 2001; Liu et al. 2007, 2011, 2012; Testemale et al. 2009; Tian et al., 2012),

the stability of tetrahedral complexes is promoted both by increasing temperature and

APPENDIX A

209

chlorinity. Compared with Fe(II) chloride complexation, the octahedral to tetrahedral

structural transition occurs at higher temperature for Mn(II) complexes (Testemale et al.,

2009). The highest order chloride complex identified in this study is the tri-chloro mono-

aqua complex MnCl3(H2O)- with MnCl42- being unstable through all T-P-salinity range. In

contrast, the fully chlorinated tetrahedral complex [FeCl4]2- has high stability, being the

dominant species in concentrated brines at elevated temperature (Testemale et al., 2009).

Tetrahedral Mn chloroaqua species (MnCl2(H2O)2(aq) and MnCl3(H2O)-) are largely

responsible for the high mobility of Mn in Cl-rich fluids in the Earth’s crust.

Acknowledgement: This research was undertaken on the FAME beamline at ESRF, France. The authors thank

the ESRF-CRG for the beamtime, and Australian International Synchrotron Access Program for travel funding.

Research funding was provided by the Australian Research Council to J.B. (grant DP0878903). The authors

thank Prof. David Sherman for his help on DFT calculations. The authors thank Dr Yves Joly for his

suggestions on the XANES simulations. The FDMNES calculations were supported by iVEC through the use

of the EPIC advanced computing resource located in Perth, Australia. Y.T. acknowledges the University of

Adelaide for ASI scholarship.

References Ankudinov, A.L., Ravel, B., Rehr, J.J., Conradson, S.D., 1998. Real-space multiple-scattering calculation and

interpretation of X-ray-absorption near-edge structure. Physical Review B 58, 7565-7576.

Beagley, B., McAuliffe, C.A., Smith, S.P.B., White, E.W., 1991. EXAFS studies of aqueous solutions of

manganese(II) chloride and bromide: structural variations with concentration and interactions with

solvent. Journal of Physics: Condensed Matter 3, 7919.

Boctor, N.Z., 1985. Rhondonite solubility and thermodynamic properties of aqueous MnCl2 in the system

MnO-SiO2-HCl-H2O. Geochimica et Cosmochimica Acta 49, 565-575.

Booth, C. H. and Bridges, F., 2003. Improved self-absorption correction for fluorescence measurements of

extended X-ray absorption fine-structure. The 12th international conference on X-ray absorption fine

structure (XAFS XII). Physica Scripta T115, 202–204.

Borg, S., Liu, W., Etschmann, B., Tian, Y., Brugger, J., 2012. An XAS study of molybdenum speciation in

hydrothermal chloride solutions from 25–385 ˚C and 600 bar. Geochimica et Cosmochimica Acta 92,

292-307.

Brugger, J., 2007. BeerOz, a set of Matlab routines for the quantitative interpretation of spectrophotometric

measurements of metal speciation in solution. Computers & Geosciences 33, 248-261.

APPENDIX A

210

Brugger, J., Black, J., McPhail, D.C., and Spiccia, L., 2001. Complexation of metal ions in brines: application

of electronic spectroscopy in the study of the Cu(II)-LiCl-H2O system between 25 and 90 ˚C.


Brugger, J., Etschmann, B., Liu, W., Testemale, D., Hazemann, J.L., Emerich, H., van Beek, W., Proux, O.,

2007. An XAS study of the structure and thermodynamics of Cu(I) chloride complexes in brines up to

high temperature (400 ˚C, 600 bar). Geochimica et Cosmochimica Acta 71, 4920-4941.

Brugger, J. and Gieré, R., 1999. As, Sb, and Ce enrichment in minerals from a metamorphosed Fe-Mn deposit

(Val Ferrera, Eastern Swiss Alps). Canadian Mineralogist 37, 37-52.

Brugger, J., Pring, A., Reith, F., Ryan, C., Etschmann, B., Liu, W., O'Neill, B., Ngothai, Y., 2010. Probing ore

deposits formation: New insights and challenges from synchrotron and neutron studies. Radiation

Physics and Chemistry 79, 151-161.

Brugger, J. and Meisser, N., 2006. Manganese-rich assemblages in the Barrhorn Unit, Turtmanntal, Central

Alps, Switzerland. Canadian Mineralogist, 44, 229-248.

Bunker, G., 2010. Introduction to XAFS: A Practical Guide to X-ray Absorption Fine Structure Spectroscopy,

1st ed. Cambridge University Press, New York.

Bühl, M., Kabrede, H., 2006. Geometries of Transition-Metal Complexes from Density-Functional Theory.

Journal of Chemical Theory and Computation 2, 1282-1290.

Chalmin, E., Farges, F., Brown, G., 2009. A pre-edge analysis of Mn K-edge XANES spectra to help

determine the speciation of manganese in minerals and glasses. Contributions to Mineralogy and

Petrology 157, 111-126.

Chantler, J., 1995. Theoretical form factor, attenuation, and scattering tabulation for Z = 1–92 from E = 1–10

eV to E = 0.4–1.0 MeV. Journal of Physical and Chemical Reference Data 24, 71–591.

Chen, Y., Fulton, J., Partenheimer, W., 2005a. A XANES and EXAFS study of hydration and ion pairing in

ambient aqueous MnBr2 solutions. Journal of Solution Chemistry 34, 993-1007.

Chen, Y., Fulton, J.L., Partenheimer, W., 2005b. The structure of the homogeneous oxidation catalyst,

Mn(II)(Br-1)x, in supercritical water: an X-ray absorption fine-structure study. Journal of the

American Chemical Society 127, 14085-14093.

Carpenter, R., 1983. Quantitative electron spin resonance (ESR) determinations of forms and total amounts of

Mn in aqueous environmental samples. Geochimica et Cosmochimica Acta 47, 875-885.

Collings, M. D., Sherman, D. M., and Ragnarsdottir, K. V., 2000. Complexation of Cu2+ in oxidised NaCl

brines from 25 C to 175 ˚C : results from in situ EXAFS spectroscopy. Chemical Geology 167, 65-73.

Cotton, F. A., Goodgame, D. M. L., Goodgame, M., 1962. Absorption Spectra and Electronic Structures of

Some Tetrahedral Manganese(II) Complexes. Journal of the American Chemical Society 84(2), 167-

172

Cornell, D. H. and Schütte, S. S., 1995. A volcanic-exhalative origin for the world's largest (Kalahari)

Manganese field. Mineralium Deposita 30, 146-151.

Davies, A. G. S., Cooke, D. R., and Gemmell, J. B., 2008. Hydrothermal breccias and veins at the Kelian gold

mine, Kalimantan, Indonesia: Genesis of a large epithermal gold deposit. Economic Geology 103,

717-757

APPENDIX A

211

de Groot, F., Vanko, G., Glatzel, P., 2009. The 1s X-ray absorption pre-edge structures in transition metal

oxides. Journal of Physics: Condensed Matter 21, 104207.

Draper, N. R., and Smith, H., 1998. Applied Regression Analysis. John Wiley & Sons, New York.

El Saffar, Z.M., Brown, G.M., 1971. The structure of manganese dichloride tetrahydrate: a neutron diffraction

study. Acta Crystallographica Section B 27, 66-73.

Etschmann, B.E., Black, J.R., Grundler, P.V., Borg, S., Brewe, D., McPhail, D.C., Spiccia, L., Brugger, J.,

2011. Copper(I) speciation in mixed thiosulfate-chloride and ammonia-chloride solutions: XAS and

UV-Visible spectroscopic studies. RSC Advance 1, 1554-1566.

Etschmann, B.E., Liu, W., Testemale, D., Müller, H., Rae, N.A., Proux, O., Hazemann, J.L., Brugger, J., 2010.

An in situ XAS study of copper(I) transport as hydrosulfide complexes in hydrothermal solutions (25-

592 ˚C, 180-600 bar): speciation and solubility in vapor and liquid phases. Geochimica et


Farges, F., 2005. Ab initio and experimental pre-edge investigations of the Mn K-edge XANES in oxide-type

materials. Physical Review B 71, 155109.

Fernandez, D. P., Goodwin, A. R. H., Lemmon, E. W., Sengers, J. M. H. L., and Williams, R. C., 1997. A

Formulation for the Static Permittivity of Water and Steam at Temperatures from 238 K to 873 K at

Pressures up to 1200 MPa, Including Derivatives and Debye-Hückel Coefficients. Journal of Physical

and Chemical Reference Data 26, 1125-1166.

Gammons, C.H., Seward, T.M., 1996. Stability of manganese (II) chloride complexes from 25 to 300 ˚C.


Helgeson, H.C., Delany, J.M., Nesbitt, H.W., 1978. Summary and critique of the thermodynamic properties of

rock - forming minerals. American Journal of Science 278A, 1-229.

James, R. H., Elderfield, H., and Palmer, M. R., 1995. The chemistry of hydrothermal fluids from the broken

spur site, 29˚N Mid-Atlantic Ridge. Geochimica et Cosmochimica Acta 59, 651-659.

Joly, Y., 2001. X-ray absorption near-edge structure calculations beyond the muffin-tin approximation.

Physical Review B 63, 125120.

Kelly, S. D., Hesterberg, D., and Ravel, B., 2008. Analysis of soils and minerals using X-ray absorption

spectroscopy. In Methods of Soil Analysis. Part 5. Mineralogical Methods. SSSA Book Series No. 5.

(eds. A.L. Ulery and L.R. Drees). pp. 387–463.

Klamt, A., Jonas, V., Bürger, T., and Lohrenz, J. C. W., 1998. Refinement and Parametrization of COSMO-RS.

The Journal of Physical Chemistry A 102, 5074-5085.

Klamt, A. and Schuurmann, G., 1993. COSMO: a new approach to dielectric screening in solvents with

explicit expressions for the screening energy and its gradient. Journal of the Chemical Society, Perkin

Transactions 2, 799-805.

Koplitz, L.V., Kim, K., McClure, D.S., 1994. Temperature Dependence of 10Dq for Aqueous

HexaaquomanganeseHexaaquomanganese(II). Inorganic Chemistry 33, 702-704.

Libu+, Z., Tia.owska, H., 1975. Stability and nature of complexes of the type MCl+ in aqueous solution

(M=Mn, Co, Ni, and Zn). Journal of Solution Chemistry 4, 1011-1022.

Lemmon, E. W., McLinden, M. O. and Friend, D. G. 2000. Thermophysical properties of fluid systems. In

NIST Chemistry WebBook, NIST Standard Reference Database (eds. W. G. Mallard and P. J.

APPENDIX A

212

Linstrom). National Institute of Standards and Technology, Gaithersburg, MD. Available from:

http://webbook.nist.gov.

Liu, W., Etschmann, B., Foran, G., Shelley, M., Brugger, J., 2007. Deriving formation constants for aqueous

metal complexes from XANES spectra: Zn2+ and Fe2+ chloride complexes in hypersaline solutions.

American Mineralogist 92, 761-770.

Liu, W., Borg, S.J., Testemale, D., Etschmann, B., Hazemann, J.-L., Brugger, J., 2011. Speciation and

thermodynamic properties for cobalt chloride complexes in hydrothermal fluids at 35–440 ˚C and

600bar: An in-situ XAS study. Geochimica et Cosmochimica Acta 75, 1227-1248.

Liu, W., Borg, S., Etschmann, B., Mei, Y., Brugger, J., 2012. An XAS study of speciation and thermodynamic

properties of aqueous zinc bromide complexes at 25-150 ˚C. Chemical Geology 298–299, 57-69.

Mahoui, A., Lapasset, J., Moret, J., Saint Gregoire, P., 1996. Bis(tetraethylammonium) Tetrachlorometallates,

[(C2H5)4N]2[MCl4], where M = Hg, Cd, Zn. Acta Crystallographica Section C 52, 2671-2674.

Maslen, E.N., Streltsov, V.A., Streltsova, N.R., Ishizawa, N., 1995. Electron density and optical anisotropy in

rhombohedral carbonates. III. Synchrotron X-ray studies of CaCO3, MgCO3 and MnCO3. Acta

Crystallographica Section B 51, 929-939.

Mayanovic, R.A., Anderson, A.J., Bassett, W.A., Chou, I.M., 2009. The structure and stability of aqueous

rare-earth elements in hydrothermal fluids: New results on neodymium(III) aqua and chloroaqua

complexes in aqueous solutions to 500 ˚C and 520 MPa. Chemical Geology 259, 30-38.

Mayanovic, R.A., Jayanetti, S., Anderson, A.J., Bassett, W.A., Chou, I.M., 2002. The Structure of Yb3+ Aquo

Ion and Chloro Complexes in Aqueous Solutions at Up to 500 ˚C and 270 MPa. The Journal of


McKibben, M. A. and Williams, A. E., 1989. Metal speciation and solubility in saline hydrothermal fluids - an

empirical-approach based on geothermal brine data. Economic Geology 84, 1996-2007

Newville, M., Ravel, B., Haskel, D., Rehr, J.J., Stern, E.A., Yacoby, Y., 1995. Analysis of multiple-scattering

XAFS data using theoretical standards. Physica B 208, 154-156.

Ozutsumi, K., Abe, Y., Takahashi, R., Ishiguro, S.-i., 1994. Chloro and Bromo Complexation of the

Manganese(II) Ion and Solvation Structure of the Manganese(II), Iron(II), Cobalt(II), Nickel(II),

Copper(II), and Zinc(II) Ions in Hexamethylphosphoric Triamide. The Journal of Physical Chemistry

98, 9894-9899.

Parr, R. G. and Yang, W., 1989. Density-functional theory of atoms and molecules. Oxford University Press,

USA, New York.

Peacock, C. L. and Sherman, D. M., 2004. Copper(II) sorption onto goethite, hematite and lepidocrocite: a

surface complexation model based on ab initio molecular geometries and EXAFS spectroscopy.


Perdew, J. P., Burke, K., and Ernzerhof, M., 1996. Generalized Gradient Approximation Made Simple.

Physical Review Letters 77, 3865-3868.

Pye, C. C. and Ziegler, T., 1999. An implementation of the conductor-like screening model of solvation within

the Amsterdam density functional package. Theoretical Chemistry Accounts: Theory, Computation,

and Modeling (Theoretica Chimica Acta) 101, 396-408.

APPENDIX A

213

Ravel, B., Newville, M., 2005. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption

spectroscopy using IFEFFIT. Journal of Synchrotron Radiation 12, 537-541.

Robie, R.A., Hemingway, B.S., Fisher, J.R., 1978. Thermodynamic properties of minerals and related

substances at 298. 15 K and 1 bar (105 pascals) pressure and at higher temperatures, USGS-Bulletin-

1452.

Roy, S., 1992. Environments and processes of manganese deposition. Economic Geology 87, 1218-1236.

Ryzhenko, B. N., Bryzgalin, O. V., Artamkina, I. Y., Spasennykh, M. Y., and Shapkin, A. I., 1985. An

electrostatic model for the electrolytic dissociation of inorganic substances dissolved in water.

Geochemistry International 22, 138-144.

Seward, T.M., Barnes,H.L.,1997. Metal transport by hydrothermal ore fluids. In: Barnes,H.L. (Ed.),

Geochemistry of Hydrothermal Ore Deposits.Wiley, New York, pp. 435–486

Seward, T.M., Driesner, T., 2004. Hydrothermal solution structure: experiments and computer simulations. In:

Palmer, D.A., Fernández-Prini, R., Harvery, A.H. (Eds.), Aqueous Systems at Elevated Temperatures

and Pressure. Elsevier Academic Press, pp. 149–182.

Sherman, D. M., 2007. Complexation of Cu+ in Hydrothermal NaCl Brines: Ab initio molecular dynamics and

energetics. Geochimica et Cosmochimica Acta 71, 714-722.

Sherman, D. M., 2009. Electronic structures of siderite (FeCO3) and rhodochrosite (MnCO3): Oxygen K-edge

spectroscopy and hybrid density functional theory. American Mineralogist 94, 166-171.

Sherman, D. M., 2010. Metal complexation and ion association in hydrothermal fluids: insights from quantum

chemistry and molecular dynamics. Geofluids 10, 41-57.

Shock, E.L., Sassani, D.C., Willis, M., Sverjensky, D.A., 1997. Inorganic species in geologic fluids:

Correlations among standard molal thermodynamic properties of aqueous ions and hydroxide

complexes. Geochimica et Cosmochimica Acta 61, 907-950.

Shulman, G.R., Yafet, Y., Eisenberger, P., and Blumberg, W.E., 1976. Observations and interpretation of X-

ray absorption edges in iron compounds and proteins. Proceedings of the National Academy of

Sciences of the United States of America 73, 1384–1388.

Shvarov, Y. V., 2008. HCh: New potentialities for the thermodynamic simulation of geochemical systems

offered by windows. Geochemistry International 46, 834-839.

Shvarov, Y.V., Bastrakov, E.N., 1999. HCh, a software package for geochemical equilibrium modeling: user’s

guide. Record 1999/25. Australian Geological Survey Organisation.

Suleimenov, O.M., Seward, T.M., 2000. Spectrophotometric measurements of metal complex formation at

high temperatures: the stability of Mn(II) chloride species. Chemical Geology 167, 177-192.

Sverjensky, D.A., Shock, E.L., Helgeson, H.C., 1997. Prediction of the thermodynamic properties of aqueous

metal complexes to 1000 ˚C and 5 kb. Geochimica et Cosmochimica Acta 61, 1359-1412.

Tagirov, B. R., Zotov, A. V., and Akinfiev, N. N., 1997. Experimental study of dissociation 685 of HCl from

350 to 500 ˚C and 500 to 2500 bars; thermodynamic properties of HCloaq. Geochimica et


Testemale, D., Argoud, R., Geaymond, O., Hazemann, J.L., 2005. High pressure high temperature cell for X-

ray absorption and scattering techniques. Review of Scientific Instruments 76, 043905.

APPENDIX A

214

Testemale, D., Brugger, J., Liu, W., Etschmann, B., Hazemann, J.-L., 2009. In-situ X-ray absorption study of

iron(II) speciation in brines up to supercritical conditions. Chemical Geology 264, 295-310.

Te Velde, G., Bickelhaupt, F. M., Baerends, E. J., Fonseca Guerra, C., van Gisbergen, S. J. A., Snijders, J. G.,

and Ziegler, T., 2001. Chemistry with ADF. Journal of Computational Chemistry 22, 931-967.

Tian, Y., Etschmann, B., Liu, W., Borg, S., Mei, Y., Testemale, D., O'Neill, B., Rae, N., Sherman, D. M.,

Ngothai, Y., Johannessen, B., Glover, C., and Brugger, J., 2012. Speciation of nickel (II) chloride

complexes in hydrothermal fluids: In situ XAS study. Chemical Geology 334, 345-363.

Tooth, B., Etschmann, B., Pokrovski, G., Testemale, D., Hazemann, J.-L., Grundler, P. and Brugger, J. 2012.

Bismuth speciation in hydrothermal fluids: an X-ray absorption spectroscopy and solubility study.

Geochimica et Cosmochimica Acta, 101, 156-172.

Tossell, J. A., 2012. Calculation of the properties of the S3! radical anion and its complexes with Cu+ in

aqueous solution. Geochimica et Cosmochimica Acta 95, 79-92.

Uchida, E., Goryozono, Y., Naito, M., Yamagami, M., 1995. Aqueous speciation of iron and manganese

chlorides in supercritical hydrothermal solutions. Geochemical Journal 29, 175-188.

Uchida, E., Haitani, T., Suetsugu, T., Kashima, T., Tsutsui, K., 2003. Experiments on the Element Distribution

between the Granodiorite JG-1a and 2M NaCl Hydrothermal Solution at Temperatures of 300 to 800

˚C and a Pressure of 1 kb. Resource Geology 53, 155-161.

Uchida, E., Tsutsui, K., 2000. Cation Leaching from the Basalt JB–1a by 2M NaCl Hydrothermal Solutions.

Resource Geology 50, 93-102.

Van Lenthe, E. and Baerends, E. J., 2003. Optimized Slater-type basis sets for the elements 1–118. Journal of

Computational Chemistry 24, 1142-1156.

Vosko, S. H., Wilk, L., and Nusair, M., 1980. Accurate spin-dependent electron liquid correlation energies for

local spin density calculations: a critical analysis. Canadian Journal of Physics 58, 1200-1211.

Westre, T.E., Kennepohl, P., DeWitt, J.G., Hedman, B., Hodgson, K.O., Solomon, E.I., 1997. A multiplet

analysis of Fe K-edge 1s → 3d pre-edge features of iron complexes. Journal of the American

Chemical Society 119, 6297-6314.

Wheat, C.G., Carpenter, R., 1988. MnCl+ and MnSO4 association constants to 170 ˚C. Journal of Solution

Chemistry 17, 467-480.

Wolfram, O., Krupp, R.E., 1996. Hydrothermal solubility of rhodochrosite, Mn (II) speciation, and

equilibrium constants. Geochimica et Cosmochimica Acta 60, 3983-3994.

Yamamoto, T., 2008. Assignment of pre-edge peaks in K-edge X-ray absorption spectra of 3d transition metal

compounds: electric dipole or quadrupole? X-Ray Spectrometry 37, 572-584.

Yardley, B. W. D., 2005. Metal concentrations in crustal fluids and their relationship to ore formation.


Zalkin, A., Forrester, J.D., Templeton, D.H., 1964. The Crystal Structure of Manganese Dichloride

Tetrahydrate. Inorganic Chemistry 3, 529-533.

215

_____________________________________

Appendix B

Speciation of Nickel (II) chloride complexes in

hydrothermal fluids: in situ XAS study

_____________________________________

216

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' !0$4/#"/(&'()'5/46$2'7889'4,2(-/:$'4(%02$;$.'/&',<:-(",$-%#2')2+/:.='!"#$%&''>*!'."+:<'

3+?2/4#"/(&'!"#"+.''

'

3+?2/4#"/(&'@$"#/2.' 1/#&A' BCA' D".4,%#&&A' ECA' F/+A' GCA' E(-HA' !CA' I$/A' BCA' 1$."$%#2$A' @CA' JK5$/22A' ECA' L#$A' 5CA'!,$-%#&A' @CICA' 5H(",#/A' BCA' M(,#&&$..$&A' ECA' N2(O$-A' PCA' E-+HH$-A' MCA' QRSQC' !0$4/#"/(&' ()'&/46$2'7889'4,2(-/:$'4(%02$;$.'/&',<:-(",$-%#2')2+/:.='8&'./"+'>*!'."+:<C'()*+%,-.#/*0.012'TTUA'TUVWTXTC''

!"#$%&'(%)#&*+"#*%),'E<' ./H&/&H' ",$' !"#"$%$&"' ()' *+",(-.,/0A' $#4,' #+",(-' 4$-"/)/$.' ",#"' ",$/-' ."#"$:' 4(&"-/?+"/(&' "(' ",$' 0+?2/4#"/(&' /.' #44+-#"$' #&:' ",#"'0$-%/../(&'/.'H-#&"$:')(-'",$'0+?2/4#"/(&'"('?$'/&42+:$:'/&'",$'4#&:/:#"$Y.'",$./.C'

5#%$'()'3-/&4/0#2'*+",(-'7P#&:/:#"$9' B+#&'1/#&'

P(&"-/?+"/(&'"('",$'3#0$-' @$./H&$:' #&:' 0$-)(-%$:' $;0$-/%$&".A' /&"$-0-$"$:' #&:' 0-(4$..$:' >*!' :#"#A' #&:' Z-("$'%#&+.4-/0"C'

'

!/H&#"+-$'

'

@#"$' QUWM+2WST'

'

5#%$'()'P(W*+",(-' E#-?#-#'D".4,%#&&'

P(&"-/?+"/(&'"('",$'3#0$-' *../."$:'Z/",'$;0$-/%$&".A'#&:'%#&+.4-/0"'$O#2+#"/(&C'

'

!/H&#"+-$''

@#"$' QUWM+2WST'

'

5#%$'()'P(W*+",(-' G$/,+#'F/+'

P(&"-/?+"/(&'"('",$'3#0$-' *../."$:'Z/",'$;0$-/%$&".A'#&:'%#&+.4-/0"'$O#2+#"/(&C'

!/H&#"+-$'

'

@#"$' QUWM+2WST'

'

5#%$'()'P(W*+",(-' !"#4$<'E(-H'

P(&"-/?+"/(&'"('",$'3#0$-' *../."$:'Z/",'$;0$-/%$&".A'#&:'%#&+.4-/0"'$O#2+#"/(&C'

!/H&#"+-$'

'

@#"$' QUWM+2WST'

'

217

!

!

Name of Co-Author Yuan Mei

Contribution to the Paper Performed DFT calculations, and wrote the DFT part of the manuscript.

Signature

Date 24-Jul-13


Contribution to the Paper Assisted with experimental design, and manuscript evaluation.

Signature

Date 24-Jul-13

Name of Co-Author Brian O’Neil

Contribution to the Paper Assisted with experimental design, and manuscript evaluation.

Signature

Date 24-Jul-13

Name of Co-Author Nick Rae


Signature

Date 24-Jul-13

Name of Co-Author David Sherman

Contribution to the Paper Supported DFT calculations, and manuscript evaluation.

Signature

Date 24-Jul-13

Name of Co-Author Yung Ngothai

Contribution to the Paper Assisted manuscript evaluation.

Signature

Date 24-Jul-13

218

!

!

Name of Co-Author Bernt Johannessen


Signature

Date 24-Jul-13

Name of Co-Author Chris Glover


Signature

Date 24-Jul-13


Contribution to the Paper Assisted with experiments and experimental design and manuscript evaluation.

Acting as corresponding author.

Signature

Date 24-Jul-13

APPENDIX B

219

A Tian, Y., Etschmann, B., Liu, W., Borg, S., Mei, Y., Testemale, D., O'Neill, B., Rae, N., Sherman, D.M., Ngothai, Y., Johannessen, B., Glover, C. & Brugger, J. (2012) Speciation of nickel (II) chloride complexes in hydrothermal fluids: in situ XAS study. Chemical Geology, v. 334, pp. 345-363

NOTE:




APPENDIX B

238

239

_____________________________________

Appendix C

An XAS study of speciation and

thermodynamic properties of aqueous zinc

bromide complexes at 25–150 °C

_____________________________________

240

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' *&'4*!'."+56'()'.0$7/#"/(&'#&5'",$-%(56&#%/7'0-(0$-"/$.'()'#8+$(+.'9/&7':-(%/5$'7(%02$;$.'

#"'<=>?=@'AB'

3+:2/7#"/(&'!"#"+.''

'

3+:2/7#"/(&'C$"#/2.' D/+E'FGE'H(-IE'!GE'J".7,%#&&E'HGE'K$/E'LGE'#&5'H-+II$-E'MGE'<@?<G'*&'4*!'."+56'()'.0$7/#"/(&'#&5' ",$-%(56&#%/7' 0-(0$-"/$.' ()' #8+$(+.' 9/&7' :-(%/5$' 7(%02$;$.' #"' <=>?=@' ABG' B,$%/7#2'N$(2(I6'<OP><OOE'=QRSOG'

!"#$%&'(%)#&*+"#*%),'H6' ./I&/&I' ",$' !"#"$%$&"' ()' *+",(-.,/0E' $#7,' #+",(-' 7$-"/)/$.' ",#"' ",$/-' ."#"$5' 7(&"-/:+"/(&' "(' ",$' 0+:2/7#"/(&' /.' #77+-#"$' #&5' ",#"'0$-%/../(&'/.'I-#&"$5')(-'",$'0+:2/7#"/(&'"(':$'/&72+5$5'/&'",$'7#&5/5#"$T.'",$./.G'

U#%$'()'B(R*+",(-' F$/,+#'D/+'

B(&"-/:+"/(&'"('",$'3#0$-' C$./I&$5' $;0$-/%$&".E' 7(&5+7"$5' $;0$-/%$&".' #&5' V-("$' 0#-"' ()' ",$' %#&+.7-/0"G' *7"/&I' #.'7(--$.0(&5/&I'#+",(-G'

!/I&#"+-$'

'

C#"$' <WRM+2R?X'

'

U#%$'()'B(R*+",(-' !"#7$6'H(-I'

B(&"-/:+"/(&'"('",$'3#0$-' *../."$5'V/",'$;0$-/%$&".E'#&5'V-("$'0#-"'()'",$'%#&+.7-/0"G'

!/I&#"+-$' C#"$' <WRM+2R?X'

'

U#%$'()'B(R*+",(-' H#-:#-#'J".7,%#&&'

B(&"-/:+"/(&'"('",$'3#0$-' *../."$5'V/",'$;0$-/%$&".E'#&5'V-("$'0#-"'()'",$'%#&+.7-/0"G'

!/I&#"+-$''

C#"$' <WRM+2R?X'

'

U#%$'()'B(R*+",(-' L+#&'K$/'

'B(&"-/:+"/(&'"('",$'3#0$-' 3$-)(-%$5'CY1'7#27+2#"/(&.E'#&5'V-("$'./%+2#"/(&'0#-"'()'",$'%#&+.7-/0"G'

!/I&#"+-$'

'

C#"$' <WRM+2R?X'

'

U#%$'()'B(R*+",(-' M(Z2'H-+II$-'

B(&"-/:+"/(&'"('",$'3#0$-' *../."$5'V/",'$;0$-/%$&".'#&5'$;0$-/%$&"#2'5$./I&'#&5'V-("$'0#-"'()'",$'%#&+.7-/0"G'

'

!/I&#"+-$'

'

C#"$' <WRM+2R?X'

APPENDIX C

241

NOTE:




A Liu, W., Borg, S., Etschmann, B., Mei, Y. & Brugger, J. (2012) An XAS study of speciation and thermodynamic properties of aqueous zinc bromide complexes at 25-150°C. Chemical Geology, 298-299, pp. 57-69

APPENDIX C

254

255

_____________________________________

Appendix D

Speciation and thermodynamic properties of

d10 transition metals: insights from ab-initio

Molecular Dynamics simulations

_____________________________________

256

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' !0$4/#"/(&' #&5' ",$-%(56&#%/4' 0-(0$-"/$.' ()' 578' "-#&./"/(&' %$"#2.9' /&./:,".' )-(%' #;</&/"/('

=(2$4+2#-'>6&#%/4.'./%+2#"/(&.'

3+;2/4#"/(&'!"#"+.''

'

3+;2/4#"/(&'>$"#/2.'=$/?'@A?'!,$-%#&?'>A'=A?'B/+?'CA?'#&5'D-+::$-?'EA?'F87FA'!0$4/#"/(&'#&5'",$-%(56&#%/4'

0-(0$-"/$.'()'578'"-#&./"/(&'%$"#2.9'/&./:,".')-(%'#;</&/"/('=(2$4+2#-'>6&#%/4.'./%+2#"/(&.?'GH",'

I&"$-&#"/(&#2'J$(2(:/4#2'K(&:-$..'LIJKM?'*+:+."'F87F?'D-/.;#&$?'*+."-#2/#A'

'

!"#$%&'(%)#&*+"#*%),'D6' ./:&/&:' ",$' !"#"$%$&"' ()' *+",(-.,/0?' $#4,' #+",(-' 4$-"/)/$.' ",#"' ",$/-' ."#"$5' 4(&"-/;+"/(&' "(' ",$' 0+;2/4#"/(&' /.' #44+-#"$' #&5' ",#"'0$-%/../(&'/.':-#&"$5')(-'",$'0+;2/4#"/(&'"(';$'/&42+5$5'/&'",$'4#&5/5#"$N.'",$./.A'

O#%$'()'K(<*+",(-' @+#&'=$/'

K(&"-/;+"/(&'"('",$'3#0$-' >$P$2(0$5' 4(%0+"#"/(&#2' %$",(5.' #&5' 0$-)(-%$5' %(2$4+2#-' ./%+2#"/(&A' D+/2"' :$(4,$%/4#2'%(5$2?'Q-("$'%#&+.4-/0"'#&5'0-$.$&"$5'/&'4(&)$-$&4$A'

!/:&#"+-$'

'

>#"$' FH<E+2<7G'

'

O#%$'()'K(<*+",(-' >#P/5'='!,$-%#&'

K(&"-/;+"/(&'"('",$'3#0$-' *../."$5'Q/",'5$P$2(0/&:'4(%0+"#"/(&#2'%$",(5'#&5'%#&+.4-/0"'$P#2+#"/(&A'

!/:&#"+-$'

'

>#"$' FH<E+2<7G'

'

O#%$'()'K(<*+",(-' C$/,+#'B/+'

K(&"-/;+"/(&'"('",$'3#0$-' *../."$5'Q/",';+/25/&:':$(4,$%/4#2'%(5$2'#&5'%#&+.4-/0"'$P#2+#"/(&A'

!/:&#"+-$'

'

>#"$' FH<E+2<7G'

'

O#%$'()'K(<*+",(-' E(R2'D-+::$-'

'K(&"-/;+"/(&'"('",$'3#0$-' *../."$5'Q/",';+/25/&:':$(4,$%/4#2'%(5$2'#&5'%#&+.4-/0"'$P#2+#"/(&A'

!/:&#"+-$'

'

>#"$' FH<E+2<7G'

'

APPENDIX D

257

A Mei, Y., Sherman, D., Liu, W. & Brugger, J. (2012) Speciation and thermodynamic properties of d10 transition metals: insights from ab-initio molecular dynamics simulations. International Geological Congress(IGC), August, Brisbane, Australia

NOTE:

This publication is included on page 257 in the print copy of the thesis held in the University of Adelaide Library.

APPENDIX D

258

259

_____________________________________

Appendix E

Zn-Cl Complexation in Magmatic-

Hydrothermal Solutions: Stability Constants

from Ab initio Molecular Dynamics

_____________________________________

260

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4&562' 6(%02$7#"/(&' /&' 8#9%#"/:5' ;<=-(",$-%#2' !(2+"/(&.>' !"#?/2/"<' 6(&."#&".' )-(%' *?' /&/"/('

8(2$:+2#-'@<&#%/:.'

3+?2/:#"/(&'!"#"+.''

'

3+?2/:#"/(&'@$"#/2.' 8$/A' BCA' !,$-%#&A' @C' 8CA' D-+99$-A' ECA' #&=' F/+A' GCA' HIJHC' 4&562' 6(%02$7#"/(&' /&' 8#9%#"/:5';<=-(",$-%#2' !(2+"/(&.>' !"#?/2/"<' 6(&."#&".' )-(%' *?' /&/"/(' 8(2$:+2#-' @<&#%/:.A' K(2=.:,%/="'6(&)$-$&:$'HIJHA'E+&$'HIJHA'8(&"-$#2A'6#&#=#'

!"#$%&'(%)#&*+"#*%),'D<' ./9&/&9' ",$' !"#"$%$&"' ()' *+",(-.,/0A' $#:,' #+",(-' :$-"/)/$.' ",#"' ",$/-' ."#"$=' :(&"-/?+"/(&' "(' ",$' 0+?2/:#"/(&' /.' #::+-#"$' #&=' ",#"'0$-%/../(&'/.'9-#&"$=')(-'",$'0+?2/:#"/(&'"('?$'/&:2+=$='/&'",$':#&=/=#"$L.'",$./.C'

M#%$'()'6(5*+",(-' B+#&'8$/'

6(&"-/?+"/(&'"('",$'3#0$-' @$N$2(0$=' :(%0+"#"/(&#2' %$",(=.' #&=' 0$-)(-%$=' %(2$:+2#-' ./%+2#"/(&C' D+/2"' 9$(:,$%/:#2'%(=$2'#&='O-("$'%#&+.:-/0"C'

!/9&#"+-$'

'

@#"$' HP5E+25JQ'

'

M#%$'()'6(5*+",(-' @#N/='8'!,$-%#&'

6(&"-/?+"/(&'"('",$'3#0$-' *../."$='O/",'=$N$2(0/&9':(%0+"#"/(&#2'%$",(='#&='%#&+.:-/0"'$N#2+#"/(&C'

!/9&#"+-$'

'

@#"$' HP5E+25JQ'

'

M#%$'()'6(5*+",(-' E(R2'D-+99$-'

'6(&"-/?+"/(&'"('",$'3#0$-' *../."$='O/",'?+/2=/&9'9$(:,$%/:#2'%(=$2'#&='%#&+.:-/0"'$N#2+#"/(&C'

!/9&#"+-$'

'

@#"$' HP5E+25JQ'

'

M#%$'()'6(5*+",(-' G$/,+#'F/+'

6(&"-/?+"/(&'"('",$'3#0$-' *../."$='O/",'?+/2=/&9'9$(:,$%/:#2'%(=$2'#&='%#&+.:-/0"'$N#2+#"/(&C'

!/9&#"+-$'

'

@#"$' HP5E+25JQ'

'

APPENDIX E

261

Goldschmidt 2012 Conference Abstracts

Mineralogical Magazine | www.minersoc.org

Zn-Cl Complexation in Magmatic-Hydrothermal Solutions: Stability

Constants from Ab initio Molecular Dyamics

YUAN MEI1,2*, DAVID M SHERMAN3, JOËL BRUGGER1, WEIHUA LIU2

1School of Earth and Environmental Sciences, The University of Adelaide, Adelaide, SA 5005, Australia,


3Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK

[email protected] (* presenting author)

The speciation of metal complexes in hydrothermal brines plays a key role in controlling the mobility and solubility of minerals in natural and man-made fluids. Experimental measurements of metal speciation is a major challenge, however, and semi-empirical equations of state (e.g., the HKF[1] model) have been needed to estimate thermodynamic properties of hydrothermal fluids. Now, however, we can apply Ab initio (quantum mechanical) molecular dynamics simulations, based on density functional theory, to predict the speciation of metal complexes in hydrothermal fluids as a function of temperature, pressure and fluid composition. Using thermodynamic integration and metadynamics techniques, these simulations can yield stability constants for metal complex formation at conditions that are experimentally inaccessible.

In this study, we investigated the species of zinc chloride

complexes via ab initio Car-Parrinello Molecular Dynamics (CPMD) simulations for ZnCl2-NaCl-H2O system with Cl- concentration of 4 m from ambient to hydrothermal-magmatic conditions. At both 25˚C and 350˚C, the MD simulations indicate that Zn-Cl complex changed from initial structure octahedral ZnCl(H2O)5

+ to tetrahedral ZnCl(H2O)3

+ after 4 picosecond (ps). However, the ligand change of chloride cannot be observed. Since Zn2+ has higher charge, the binding between Zn2+ and Cl- is stronger, the ions exchange can be hardly observed via short MD simulation (< 10 ps). Thermodynamic integration is a realistical approach to evaluate entropic properties. To calculate the binding free energy between Zn2+ and Cl-, thermodynamic integration based on constraint CPMD simulations were conducted by constraining the Zn-Cl bond distances. The constraint systems reached the state of equilibrium within 1 ps. The mean forces of forming Zn-Cl bond at the different distances from 2 Å to 5 Å were calculated over 5 ps, and the change in free energy was derived by integrating the mean force vs distance. The integral gave binding free energies of -6.75 kJ/mol at 25˚C, 1 bar and -75.52 kJ/mol at 350˚C, 500 bar for reaction Zn2+ + Cl- = ZnCl+. The predicted logKs of that reaction are 1.18 at 25˚C and 6.08 at 350˚C, in good agreement with the experimental values (0.20 and 6.87, respectively, from revised HKF model[1]). Having established our methodology, we will now investigate ZnCl complexation at the extreme temperatures of magmatic-hydrothermal fluids.

[1] Sverjensky et al., (1997) Geochim. Cosmochim. Acta, 61, 1359–

1421

APPENDIX E

262

263

_____________________________________

Appendix F

Ab initio molecular dynamics simulation of

copper(I) complexation in chloride/sulfide

fluids _____________________________________

264

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' *4'/&/"/('%(2$5+2#-'67&#%/5.'./%+2#"/(&'()'5(00$-89:'5(%02$;#"/(&'/&'5,2(-/6$<.+2)/6$')2+/6.'

3+42/5#"/(&'!"#"+.''

'

3+42/5#"/(&'=$"#/2.' >$/?' @A?' !,$-%#&?' =A' >A?' B-+CC$-?' DA?' #&6' E/+?' FA?' GHIIA' *4' /&/"/(' %(2$5+2#-' 67&#%/5.'./%+2#"/(&' ()' 5(00$-89:' 5(%02$;#"/(&' /&' 5,2(-/6$<.+2)/6$' )2+/6.?' J(26.5,%/6"' K(&)$-$&5$' GHII?'*+C+."'GHII?'3-#C+$?'KL$5,'M$0+42/5'!

"#$%&'!(&)$'*+#$*&),!B7' ./C&/&C' ",$' !"#"$%$&"' ()' *+",(-.,/0?' $#5,' #+",(-' 5$-"/)/$.' ",#"' ",$/-' ."#"$6' 5(&"-/4+"/(&' "(' ",$' 0+42/5#"/(&' /.' #55+-#"$' #&6' ",#"'0$-%/../(&'/.'C-#&"$6')(-'",$'0+42/5#"/(&'"('4$'/&52+6$6'/&'",$'5#&6/6#"$N.'",$./.A'

O#%$'()'K(P*+",(-' @+#&'>$/'

K(&"-/4+"/(&'"('",$'3#0$-' =$Q$2(0$6' 5(%0+"#"/(&#2' %$",(6.' #&6' 0$-)(-%$6' %(2$5+2#-' ./%+2#"/(&A' B+/2"' C$(5,$%/5#2'%(6$2?'R-("$'%#&+.5-/0"'#&6'0-$.$&"$6'/&'5(&)$-$&5$A'

!/C&#"+-$'

'

=#"$' GSPD+2PIT'

'

O#%$'()'K(P*+",(-' =#Q/6'>'!,$-%#&'

K(&"-/4+"/(&'"('",$'3#0$-' *../."$6'R/",'6$Q$2(0/&C'5(%0+"#"/(&#2'%$",(6'#&6'%#&+.5-/0"'$Q#2+#"/(&A'

!/C&#"+-$'

'

=#"$' GSPD+2PIT'

'

O#%$'()'K(P*+",(-' D(U2'B-+CC$-'

'K(&"-/4+"/(&'"('",$'3#0$-' *../."$6'R/",'4+/26/&C'C$(5,$%/5#2'%(6$2'#&6'%#&+.5-/0"'$Q#2+#"/(&A'

!/C&#"+-$'

'

=#"$' GSPD+2PIT'

'

O#%$'()'K(P*+",(-' F$/,+#'E/+'

K(&"-/4+"/(&'"('",$'3#0$-' *../."$6'R/",'4+/26/&C'C$(5,$%/5#2'%(6$2'#&6'%#&+.5-/0"'$Q#2+#"/(&A'

!/C&#"+-$'

'

=#"$' GSPD+2PIT'

'

APPENDIX F

265

Goldschmidt Conference Abstracts

Mineralogical Magazine www.minersoc.org

1447

Ab initio molecular dynamics simulation of copper(I) complexation

in chloride/sulfide fluids YUAN MEI1,2,3, DAVID M SHERMAN2, JOËL BRUGGER1

AND WEIHUA LIU3 1School of Earth and Environmental Sciences, The University

of Adelaide, Adelaide, SA 5000, Australia

2Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK

3CSIRO Earth Science and Resource Engineering, Clayton, VIC 3168, Australia Chloride and hydrosulfide are the primary ligands

believed to control the transport of copper in hydrothermal fluids. Recent studies of Cu complexation in hydrothermal Cl-, HS- solutions have been done using X-ray Absorption Spectroscopy (XAS). However, coordination numbers have a large uncertainty and are strongly correlated with Debye-Waller factors; moreover, it is very difficult to distinguish between chloride and sulfur ligands. Ab initio molecular dynamics simulations based on density functional theory enable us to interpret EXAFS results and, potentially, predict stability constants of metal complexes.

In this study, we investigated the species of copper(I) complexes via ab initio Car-Parrinello Molecular Dynamics simulations for copper(I) solutions with different hydrosulfide/chloride ratios at 500bar and 600K. Calculations were done using Vanderbilt ultrasoft pseudopotentials and the PBE exchange-correlation functional.

In the absence of Cl-ligands, copper forms a Cu(HS)2-

complex with a Cu-S bond length of 2.17 Å (vs. an expt. value of 2.15 Å); moreover, the S-Cu-S bond angle is ~162°, in excellent agreement with experiment (150-160°). In the presence of excess chloride, however, we find that Cu forms previously unknown Cu(HS)Cl- and (minor) CuCl2(HS)-2 complexes. Such complexes would be difficult to resolve from CuCl2

- or Cu(HS)2- using EXAFS. We also explored the

complexation of Cu in a low density (0.29 g/cm3), high T (1273K) fluid (vapour). Here, we find that Cu forms Cu(HS)2

- (not the neutral CuHS, as expected). We tentatively suggest that charged complexes may be significant in high temperature, low density fluids.

Ultimately, we hope to predict stability constants of metal complexes. To this end, we are testing metadynamics and thermodynamic integration with respect to metal-ligand distances or coordination numbers. Using these techniques, we estimate the free energy difference between CuCl2

- + HS- and CuCl(HS)- + Cl- to be ~40 KJ/mol.

APPENDIX F

266

267

_____________________________________

Appendix G

Which Ligand is the most Import for Gold

Transport in Hydrothermal Fluids? An in situ

XAS Study in Mixed-Ligand Solutions _____________________________________

268

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4(2$5+2#-62$7$2'+&8$-."#&8/&9'()'%$"#2'"-#&.0(-"'/&',:8-(",$-%#2'(-$')2+/8.;'/&'./"+'$<0$-/%$&".'

#&8'#='/&/"/('%(2$5+2#-'8:&#%/5'./%+2#"/(&.'

3+=2/5#"/(&'!"#"+.''

'

3+=2/5#"/(&'>$"#/2.'?/+@'AB@'C".5,%#&&@'DB@'1$."$%#2$@'>B@'4$/@'EB@'F#G$%#&&@'HB@'I$%0$2@'JB@'4K$22$-@'FB@'#&8'

D-+99$-@'HB@'LMNOB'A,/5,'2/9#&8'/.'",$'%(."'/%0(-"')(-'9(28'"-#&.0(-"'/&',:8-(",$-%#2')2+/8.P'*&'

/&'./"+'Q*!'."+8:'/&'5(%0$"/"/7$'./"+#"/(&.@'R(28.5,%/8"'LMNO@'*+9+."'LMNO@'S2(-$&5$@'T"#2:B'

!"#$%&'(%)#&*+"#*%),'D:' ./9&/&9' ",$' !"#"$%$&"' ()' *+",(-.,/0@' $#5,' #+",(-' 5$-"/)/$.' ",#"' ",$/-' ."#"$8' 5(&"-/=+"/(&' "(' ",$' 0+=2/5#"/(&' /.' #55+-#"$' #&8' ",#"'0$-%/../(&'/.'9-#&"$8')(-'",$'0+=2/5#"/(&'"('=$'/&52+8$8'/&'",$'5#&8/8#"$U.'",$./.B'

V#%$'()'W(6*+",(-' A$/,+#'?/+'

W(&"-/=+"/(&'"('",$'3#0$-' >$./9&$8'#&8'0$-)(-%$8'$<0$-/%$&".@'X-("$'%#&+.5-/0"'#&8'0-$.$&"$8'/&'5(&)$-$&5$B'

!/9&#"+-$'

'

>#"$' LY6H+26NO'

'

V#%$'()'W(6*+",(-' D#-=#-#'C".5,%#&&'

'W(&"-/=+"/(&'"('",$'3#0$-' *../."$8'X/",'$<0$-/%$&".'#&8'%#&+.5-/0"'$7#2+#"/(&B'

!/9&#"+-$''

>#"$' LY6H+26NO'

'

V#%$'()'W(6*+",(-' >$&/.'1$."$%#2$'

'W(&"-/=+"/(&'"('",$'3#0$-' *../."$8'X/",'$<0$-/%$&".'#&8'%#&+.5-/0"'$7#2+#"/(&B'

!/9&#"+-$'

'

>#"$' LY6H+26NO'

'

V#%$'()'W(6*+",(-' E+#&'4$/'

W(&"-/=+"/(&'"('",$'3#0$-' 3$-)(-%$8'4>'5#25+2#"/(&.'#&8'#../."$8'X/",'%#&+.5-/0"'$7#2+#"/(&B'

!/9&#"+-$'

'

>#"$' LY6H+26NO'

'

'

269

!

!

Name of Co-Author Jean-Louis Hazemann


Signature

Date 24-Jul-13

Name of Co-Author Kirsten Rempel


Signature

Date 24-Jul-13

Name of Co-Author Harald Müeller


Signature

Date 24-Jul-13


Contribution to the Paper Assisted with experiments and manuscript evaluation.

Signature

Date 24-Jul-13

!"#$%&'%(&)*+,-&. /$"0)1&+23%4"5$#"00

(&0,.26+,2&0%,&%,-$%7"8$. *3323,$9%:2,-%$;8$.2#$0,3%"09%#"0+3<.28,%$=">+",2&0?

@2A0",+.$ B",$ CD)/+>)EF

!"#$%&'%(&)*+,-&. G2.3,$0%H$#8$>

%(&0,.26+,2&0%,&%,-$%7"8$. *3323,$9%:2,-%$;8$.2#$0,3%"09%#"0+3<.28,%$=">+",2&0?

@2A0",+.$ B",$ CD)/+>)EF

!"#$%&'%(&)*+,-&. 4".">9%IJ$>>$.

%(&0,.26+,2&0%,&%,-$%7"8$. *3323,$9%:2,-%$;8$.2#$0,3%"09%#"0+3<.28,%$=">+",2&0?

@2A0",+.$ B",$ CD)/+>)EF

!"#$%&'%(&)*+,-&. /&K>%L.+AA$.

%(&0,.26+,2&0%,&%,-$%7"8$. *3323,$9%:2,-%$;8$.2#$0,3%"09%#"0+3<.28,%$=">+",2&0?

@2A0",+.$ B",$ CD)/+>)EF

APPENDIX G

270

Goldschmidt2013 Conference Abstracts

!!!"#$%&'()*")'+,-./012"11324#$%#5+"6217"288"9"16

1627

Which ligand is the most important for gold transport in hydrothermal

fluids? An in situ XAS study in mixed-ligand solutions

WEIHUA LIU1, BARBARA ETSCHMANN2,4, DENIS TESTEMALE5, YUAN MEI1,3,

JEAN-LOUIS HAZEMANN5, KIRSTEN REMPEL6, HARALD MÜLLER7 AND JOËL BRUGGER2,3

1 CSIRO Earth Science and Resource Engineering, Australia, ([email protected])

2 South Australian Museum, Australia 3 School of Earth and Environmental Sciences and 4School of

Chemical Engineering, University of Adelaide, Australia 5Institut Néel, Département MCMF and FAME beamline,

ESRF, France 6Department of Applied Geology, Curtin University, Perth,

Australia 7ESRF, Grenoble, France

Gold transport and deposition in hydrothermal ore fluids is dependent on the identity and stability of predominating aqueous gold complexes. Gold(I) bisulfide (e.g., Au(HS)2

-) and in some instances Au(I) chloride complexes are widely acknowledged to account for Au transport in ore fluids.

This study investigates the potential of the unconventional ligands Br- and NH3 to increase Au mobility. This was achieved by determining the predominant Au species in hydrothermal fluids with binary mixed ligands (Br- - Cl-, Br- - HS-, HS- - NH3), and measuring their structural properties using in situ Synchrotron X-ray Absorption Spectroscopy (XAS). The capacity of XAS to follow the progress of ligand exchange reactions was demonstrated at room temperature, where the Au(III)Br4

- complex was found to predominate in mixed Br-/Cl- solutions (Br-/Cl- = 0.1-1), with average ligand numbers derived from XAS in good agreement with a recent UV-Vis study (Usher et al.., 2009, Geochim. Cosmochim. Acta 73, 3359-3380). At temperatures up to 400 oC and at 600 bar, the XAS measurements show that Au(I) – HS- complexes are the only stable Au species in mixed HS-/Br- and HS-/NH3 fluids (HS-/Br- = 0.1; HS-/NH3 = 0.2), indicating that hydrosulfide is the most important ligand for Au transport in the hydrothermal fluid under our experimental conditions, i.e., hydrosulfide complexes outcompete bromide and ammine complexes in S-bearing fluids. These results are comparable to solubility and speciation calculations based on the available thermodynamic data.

271

_____________________________________

Appendix H

Molecular-level understanding of metal transport

in hydrothermal ore fluids: in situ experiments

and ab initio molecular dynamic simulations _____________________________________

272

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4(2$5+2#-62$7$2'+&8$-."#&8/&9'()'%$"#2'"-#&.0(-"'/&',:8-(",$-%#2'(-$')2+/8.;'/&'./"+'$<0$-/%$&".'

#&8'#='/&/"/('%(2$5+2#-'8:&#%/5'./%+2#"/(&.'

3+=2/5#"/(&'!"#"+.''

'

3+=2/5#"/(&'>$"#/2.'?/+@'AB@'C(-9@'!B@'D".5,%#&&@'CB@'4$/@'EB@'1$."$%#2$@'>B@'!,$-%#&@'>B'4B@'#&8'C-+99$-@'FB@'

GHIGB'4(2$5+2#-62$7$2'+&8$-."#&8/&9'()'%$"#2'"-#&.0(-"'/&',:8-(",$-%#2'(-$')2+/8.;'/&'./"+'

$<0$-/%$&".'#&8'#='/&/"/('%(2$5+2#-'8:&#%/5'./%+2#"/(&.@'JK",'L&"$-&#"/(&#2'M$(2(9/5#2'

N(&9-$..'OLMNP@'*+9+."'GHIG@'C-/.=#&$@'*+."-#2/#B'

'

!"#$%&'(%)#&*+"#*%),'C:' ./9&/&9' ",$' !"#"$%$&"' ()' *+",(-.,/0@' $#5,' #+",(-' 5$-"/)/$.' ",#"' ",$/-' ."#"$8' 5(&"-/=+"/(&' "(' ",$' 0+=2/5#"/(&' /.' #55+-#"$' #&8' ",#"'0$-%/../(&'/.'9-#&"$8')(-'",$'0+=2/5#"/(&'"('=$'/&52+8$8'/&'",$'5#&8/8#"$Q.'",$./.B'

R#%$'()'N(6*+",(-' A$/,+#'?/+'

N(&"-/=+"/(&'"('",$'3#0$-' >$./9&$8'#&8'0$-)(-%$8'$<0$-/%$&".@'S-("$'%#&+.5-/0"'#&8'0-$.$&"$8'/&'5(&)$-$&5$B'

!/9&#"+-$'

'

>#"$' GK6F+26IJ'

'

R#%$'()'N(6*+",(-' !"#5$:'C(-9'

'N(&"-/=+"/(&'"('",$'3#0$-' *../."$8'S/",'$<0$-/%$&".'#&8'%#&+.5-/0"'$7#2+#"/(&B'

!/9&#"+-$'

'

>#"$' GK6F+26IJ'

'

R#%$'()'N(6*+",(-' C#-=#-#'D".5,%#&&'

'N(&"-/=+"/(&'"('",$'3#0$-' *../."$8'S/",'$<0$-/%$&".'#&8'%#&+.5-/0"'$7#2+#"/(&B'

!/9&#"+-$''

>#"$' GK6F+26IJ'

'

R#%$'()'N(6*+",(-' E+#&'4$/'

N(&"-/=+"/(&'"('",$'3#0$-' 3$-)(-%$8'4>'5#25+2#"/(&.@'#&8'#../."$8'S/",'%#&+.5-/0"'$7#2+#"/(&B'

!/9&#"+-$'

'

>#"$' GK6F+26IJ'

'

273

!

!



Signature

Date 24-Jul-13

Name of Co-Author David M Sherman

Contribution to the Paper Supported MD calculations.

Signature

Date 24-Jul-13


Contribution to the Paper Assisted with experiments and manuscript evaluation.

Signature

Date 24-Jul-13

APPENDIX H

274

A NOTE:

This publication is included on page 274 in the print copy of the thesis held in the University of Adelaide Library.

A Liu, W., Borg, S., Etschmann, B., Mei, Y., Testemale, D., Sherman, D. & Brugger, J. (2012) Molecular-level understanding of metal transport in hydrothermal ore fluids: in situ experiments and ab initio molecular dynamic simulations. International Geological Congress(IGC), August, Brisbane, Australia

275

_____________________________________

Appendix I

Metal Complexation in Hydrothermal Fluids: Insights from Ab Initio Molecular Dynamics

_____________________________________

276

!"#"$%$&"'()'*+",(-.,/0' 1/"2$'()'3#0$-' 4$"#2'5(%02$6#"/(&'/&'789-(",$-%#2':2+/9.;'<&./=,".')-(%'*>'<&/"/('4(2$?+2#-'@8&#%/?.'

3+>2/?#"/(&'!"#"+.''

'

3+>2/?#"/(&'@$"#/2.' !,$-%#&A'@B'4BA'#&9'4$/A'CBA'DEFFB'4$"#2'5(%02$6#"/(&' /&'789-(",$-%#2':2+/9.;' <&./=,".')-(%'

*>' <&/"/(' 4(2$?+2#-' @8&#%/?.A' G(29.?,%/9"' 5(&)$-$&?$' DEFFA' *+=+."' DEFFA' 3-#=+$A' 5H$?,'

I$0+>2/?''

!"#$%&'(%)#&*+"#*%),'J8' ./=&/&=' ",$' !"#"$%$&"' ()' *+",(-.,/0A' $#?,' #+",(-' ?$-"/)/$.' ",#"' ",$/-' ."#"$9' ?(&"-/>+"/(&' "(' ",$' 0+>2/?#"/(&' /.' #??+-#"$' #&9' ",#"'

0$-%/../(&'/.'=-#&"$9')(-'",$'0+>2/?#"/(&'"('>$'/&?2+9$9'/&'",$'?#&9/9#"$K.'",$./.B'

'

L#%$'()'5(M*+",(-' @#N/9'4'!,$-%#&'

5(&"-/>+"/(&'"('",$'3#0$-' @$N$2(0$9'?(%0+"#"/(&#2'%$",(9.'#&9'0$-)(-%$9'%(2$?+2#-' ./%+2#"/(&A'O-("$'%#&+.?-/0"'#&9'

0-$.$&"$9'/&'?(&)$-$&?$B'

!/=&#"+-$'

'

@#"$' DPMQ+2MFR'

'

L#%$'()'5(M*+",(-' C+#&'4$/'

5(&"-/>+"/(&'"('",$'3#0$-' *../."$9'O/",'?(%0+"#"/(&#2'%$",(9'9$N$2(0/&='#&9'9#"#'?(22$?"/&=B'

!/=&#"+-$'

'

@#"$' DPMQ+2MFR'

'

'

'

'

APPENDIX I

277

Goldschmidt Conference Abstracts

Mineralogical Magazine www.minersoc.org

1855

Metal complexation in hydrothermal fluids: Insights from ab initio

molecular dynamics DAVID M. SHERMAN1* AND YUAN MEI1,2

1School of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK (*correspondence: [email protected])

2School of Earth and Environmental Sciences, University of Adelaide, Adelaide SA 5000, Australia Complexation of metals by Cl- and HS- ligands in

hydrothermal fluids is a fundamental process in the evolution of the Earth’s crust and the formation of ore deposits. Current thermodynamic models of complexation equilibria under hydrothermal conditions depend on extrapolations of experimental data using equations of state based on the Born model of solvation. Thermodynamic parameters for aqueous species are often provisional estimates based on systematic correlations between fundamental properties such as entropy, volume, ionic radius etc. Computational molecular simulations, however, can be used to test current thermodynamic models, predict metal speciation, and even estimate thermodynamic properties. For a condensed fluid, molecular dynamics simulations can be used to sample the configurational degrees of freedom in order to predict properties as a function of pressure and temperature. Simulations of dilute solutions, however, require very large systems (1000’s of atoms) and very long (> 1 ns) simulation times; such calculations are only practical by treating the atomic interactions using classical two- or three-body interatomic potentials. However, classical potentials seem to be unreliable for describing metal-ligand interactions, especially for transition metals and metalloids such as Sn+2, Au+3, Cu+2 and Cu+. ‘Ab initio molecular dynamics’ treats the molecular motions classically but the atomic interactions quantum mechanically. Although these simulations are only practical for systems with 100’s of atoms over short times (< 100 ps), they are giving fundamental new insights on metal speciation in hydrothermal fluids. Here, we describe simulations of Cu, Zn, Sn, Au, and Ni in NaCl- and HS-bearing aqueous fluids up to 350 °C. We show that predicted structures and speciation are in close agreement with experiment. Based on our simulations, we propose that the major driving force for metal complexation in hydrothermal fluids is the change in translational entropy between reactants and products. Entropies and free energies of complex formation can be estimated using thermodynamic integration and metadynamics. Applications of these techniques to Zn-Cl and Cu-Cl-HS stability constants will be presented.

APPENDIX I

278

Metal mobility in hydrothermal fluids: insights from ab ...Metal Mobility in Hydrothermal Fluids:...

Documents

Transcript of Metal mobility in hydrothermal fluids: insights from ab ...Metal Mobility in Hydrothermal Fluids:...