SUBSTITUTION IN BASIC

SECONDARY Cu(II) CHLORIDE

MINERALS

Matthew Joseph Sciberras BSc (Hons), UWS

This thesis is submitted for the degree of

Doctor of Philosophy

in the University of Western Sydney

April 2013

“The important thing in science is not so much to obtain new facts as to discover new ways of thinking about them” ― Sir Lawrence Bragg

This thesis is dedicated to Annina

i

TABLE OF CONTENTS

ACKNOWLEDGMENTS ...................................................................................................... iv

STATEMENT OF AUTHENTICATION ............................................................................. v

LIST OF ABBREVIATIONS ................................................................................................ vi SUPPORTING PUBLICATIONS ....................................................................................... vii ABSTRACT .......................................................................................................................... viii CHAPTER 1 - INTRODUCTION ......................................................................................... 1

1.1 THE BASIC Cu(II) CHLORIDE MINERALS ......................................................................... 2 1.2 PARATACAMITE AND THE SUBSTITUTED PHASES ...................................................... 4 1.3 PHASE IDENTIFICATION ....................................................................................................... 7 1.4 SYNTHESIS AND STABILITY ............................................................................................... 10

CHAPTER 2 – CRYSTALLOGRAPHIC STUDIES ......................................................... 12

2.1 THE TEMPERATURE-RELATED REVERSIBLE PHASE TRANSFORMATION BETWEEN PARATACAMITE AND HERBERTSMITHITE .............................................. 12 2.1.1 INTRODUCTION ............................................................................................................... 12 2.1.2 THE STRUCTURE OF PARATACAMITE ....................................................................... 12 2.1.3 SAMPLE AND ANALYSIS ............................................................................................... 14

2.1.3.1 Single-crystal X-ray diffraction ................................................................................................... 15 2.1.4 RESULTS ............................................................................................................................ 24 2.1.5 DISCUSSION ...................................................................................................................... 28

2.1.5.1 The (2+2+2) Jahn-Teller distortion ............................................................................................ 29 2.1.5.2 Origin of the phase transition in type paratacamite .................................................................... 32

2.2 THE SINGLE-CRYSTAL STRUCTURE OF Mg- AND Ni-ANALOGUES OF PARATACAMITE ..................................................................................................................... 33 2.2.1 INTRODUCTION ............................................................................................................... 33 2.2.2 SAMPLES AND ANALYSIS ............................................................................................. 33

2.2.2.1 Optical and physical properties .................................................................................................. 35 2.2.3 CRYSTALLOGRAPHY ..................................................................................................... 35

2.2.3.1 Sample 64041 (Mg-rich) ............................................................................................................. 35 2.2.3.2 Sample WAM M365.2003 (Ni-rich) ............................................................................................. 37

2.2.4. X-RAY POWDER DIFFRACTION .................................................................................. 38 2.2.5 DISCUSSION ...................................................................................................................... 38

2.2.5.1 Interlayer cation distribution ...................................................................................................... 38 2.2.5.2 The (2+2+2) Jahn-Teller distortion ............................................................................................ 43

2.2.6. NEW MINERALS .............................................................................................................. 46 2.3 THE SINGLE-CRYSTAL X-RAY STRUCTURE OF THE Co ANALOGUE OF

HERBERTSMITHITE FROM SALAR GRANDE, IQUIQUE PROVENCE, CHILE ....... 47 2.3.1 INTRODUCTION ............................................................................................................... 47 2.3.2 SAMPLES AND ANALYSIS ............................................................................................. 47

2.3.2.1 Optical and physical properties .................................................................................................. 48

ii

2.3.3 CRYSTALLOGRAPHY ..................................................................................................... 48 2.2.4 X-RAY POWDER DIFFRACTION ................................................................................... 49 2.3.4 DISCUSSION ...................................................................................................................... 49 2.3.5 A NEW MINERAL ............................................................................................................. 52

2.4 THE COMPOSITION-DEPENDENT STRUCTURAL TRANSFORMATION SERIES OF THE PARATACAMITE GROUP ............................................................................................. 53 2.4.1 INTRODUCTION ............................................................................................................... 53 2.4.2 SAMPLES AND ANALYSIS ............................................................................................. 53 2.4.3 CRYSTALLOGRAPHY ..................................................................................................... 54 2.4.4 RESULTS AND DISCUSSION .......................................................................................... 56

CHAPTER 3 – RAMAN SPECTROSCOPY ...................................................................... 67

3.1 RAMAN SPECTROSCOPY OF NATURAL SINGLE-CRYSTALS ................................... 67 3.1.1 INTRODUCTION ............................................................................................................... 67 3.1.2 SAMPLES AND METHODS ............................................................................................. 68

3.1.2.1 Single-crystal X-ray diffraction ................................................................................................... 69 3.1.2.2 Raman spectroscopy .................................................................................................................... 69

3.1.3 RESULTS AND DISCUSSION .......................................................................................... 71 3.1.3.1 Single-crystal X-ray diffraction ................................................................................................... 71 3.1.3.2 Raman spectroscopy .................................................................................................................... 72 3.1.3.3 Metal-anion framework vibrations .............................................................................................. 75 3.1.3.4 M–O–H deformation ................................................................................................................... 77 3.1.3.4 O–H stretching ............................................................................................................................ 77 3.1.3.5 Phase identification ..................................................................................................................... 78

3.2 RAMAN SPECTROSCOPY OF THE SYNTHETIC Cu4-xMx(OH)6Cl2 SUBSTITUTION SERIES ........................................................................................................................................ 80 3.2.1 INTRODUCTION ............................................................................................................... 80 3.2.2 SAMPLES AND METHODS ............................................................................................. 80

3.2.2.1 Powder X-ray diffraction ............................................................................................................. 81 3.2.2.2 Composition ................................................................................................................................ 81 3.2.2.3 Raman spectroscopy .................................................................................................................... 81

3.2.3 RESULTS AND DISCUSSION .......................................................................................... 82 3.2.3.1 Powder X-ray diffraction ............................................................................................................. 82 3.2.3.2 Raman spectroscopy .................................................................................................................... 84 3.2.3.3 Metal–anion framework vibrations ............................................................................................. 84 3.2.3.4 M–O–H deformation and O–H stretching regions ...................................................................... 87 3.2.3.5 H bonding .................................................................................................................................... 93 3.2.3.6 The transformation series ............................................................................................................ 95

CHAPTER 4 – SUBSTITUTION AND ACTIVITY .......................................................... 97

4.1 THERMODYNAMICS OF SUBSTITUTION IN CLINOATACAMITE ............................ 97 4.1.1 INTRODUCTION ............................................................................................................... 97 4.1.2 SAMPLES AND METHODS ............................................................................................. 98

4.1.2.1 Powder X-ray diffraction ............................................................................................................. 98 4.1.2.2 Clinoatacamite synthesis ............................................................................................................. 98 4.1.2.3 Synthesis of cation-substituted phases ......................................................................................... 99 4.1.2.4 Solution calculations ................................................................................................................... 99

4.1.3 RESULTS AND DISCUSSION ........................................................................................ 101

iii

4.2 DISTRIBUTION COEFFICIENTS FOR HERBERTSMITHITE AND GILLARDITE . 104 4.2.1 INTRODUCTION ............................................................................................................. 104 4.2.2 SAMPLES AND METHODS ........................................................................................... 104 4.2.3 RESULTS AND DISCUSSION ........................................................................................ 104

CHAPTER 5 – CONCLUSIONS ....................................................................................... 109

5.1 NEW MINERALS .................................................................................................................... 109 5.2 A REVERSIBLE R𝟑𝟑� TO R𝟑𝟑�m PHASE TRANSFORMATION .......................................... 109 5.3 THE (2+2+2) DISTORTION OF M(2) IN PARATACAMITE ........................................... 110 5.4 COMPOSITION-INDUCED STRUCTURAL TRANSFORMATIONS ............................ 111

5.4.1 CRYSTALLOGRAPHIC STUDIES ................................................................................. 111 5.4.1 RAMAN SPECTROSCOPY ............................................................................................. 112

5.6 SOLID–SOLUTION STUDIES .............................................................................................. 113 5.7 DEFINITION OF PARATACAMITE ................................................................................... 114

REFERENCES .................................................................................................................... 115

APPENDIX – CRYSTALLOGRAPHIC INFORMATION FILES .............................CD

iv

ACKNOWLEDGMENTS

The work contained in this thesis was conducted through collaboration between

laboratories in the University of Western Sydney (UWS), Sydney, NSW, Australia, and the

University of Hamburg (UHH), Hamburg, Germany. I am grateful to my supervisors

Professors Peter Williams and Peter Leverett for their support and guidance over the years.

They both frequently devoted long hours of invaluable discussion with me over the course of

this project. The majority of the single-crystal measurements and all Raman data were

collected by myself with collaborative supervisors at UHH, Professor Jochen Schlüter, Dr

Thomas Malcherek and Dr Boriana Mihailova. I am grateful to each of them for the time they

devoted to me and for their advice.

Dr Mark Welch of the Natural History Museum, London, UK, is thanked for

supplying the single-crystal X-ray data from the type specimen of paratacamite and for

compiling some of the Tables and Figures in Chapter 2.1. Thanks go to Dr Anthony R.

Kampf of the Natural History Museum of Los Angeles County, Los Angeles, CA, USA, for

supplying single-crystal X-ray data and compiling optical, morphological and physical data

for both the new Mg-analogue of paratacamite and the Co-analogue of herbertsmithite.

Professor David E. Hibbs of the University of Sydney, Sydney, NSW, Australia, is thanked

for providing single-crystal X-ray data from the Ni-analogue of paratacamite. Dr Peter J.

Downes of the Western Australian Museum, Welshpool, WA, Australia, is thanked for

locating suitable specimens for study. Jim Sharpe is also thanked for loaning specimens of

gillardite and herbertsmithite for analysis. Simon Hager and Stefanie Heidrich are thanked for

assistance with electron microprobe work. I thank Dr Jason Reynolds, Tim Murphy and

Adam Roper for hours of scientific discussion.

I acknowledge support from Deutscher Akademischer Austausch Dienst (DAAD) for

a scholarship in the program Research Grant for Doctoral Candidates and Young Academics

and Scientists A/11/93939, for a research stay in UHH. UWS is also acknowledged for a

post-graduate studies grant.

My family has always given their continuing support and guidance throughout my

endeavours and I am grateful to them. Special thanks go to Annina Schulz for her

unconditional support and advice throughout the years.

v

STATEMENT OF AUTHENTICATION

This thesis contains work that, to the best of my knowledge and belief, is original except

where due acknowledgment appears in the text. I declare that material in this thesis has not

been submitted in any form for a degree or diploma at any university or institution of tertiary

education.

..............…………………

Matthew Joseph Sciberras

April 2013

vi

LIST OF ABBREVIATIONS

AAS atomic absorption spectrophotometry BAV bond-angle variance CIF crystallographic information file D distribution coefficient ΔGf

ɵ standard Gibbs free energy of formation ε refractive index of the extraordinary ray FWHM full-width at half-maximum γ activity coefficient IR infra-red K equilibrium constant n refractive index NPD non-positive-definite ω refractive index of the ordinary ray pfu per formula unit PXRD powder X-ray diffraction QE quadratic elongation RI fluid refractive index fluid Tc critical temperature WDS wavelength dispersive spectroscopy XRD X-ray diffraction Z atomic number

vii

SUPPORTING PUBLICATIONS Welch, M.D., Sciberras, M.J., Leverett, P., Williams, P.A, Schlüter, J. and Malcherek, T.

(2013) A temperature-induced reversible transformation between paratacamite and herbertsmithite. Physics and Chemistry of Minerals, Submitted.

Kampf, A.R., Sciberras, M.J., Leverett, P., Williams, P.A., Malcherek, T., Schlüter, J., Hibbs,

D.E., Downes, P.J., Welch, M.D. and Dini, M (2013) The single-crystal structure of Mg- and Ni-analogues of paratacamite. In Preparation.

Kampf, A.R., Sciberras, M.J. and Williams, P.A. (2013) The single-crystal X-ray structure

of the Co-analogue of herbertsmithite from Salar Grande Provence, Chile. In Preparation.

Sciberras, M.J., Leverett, P., Williams, P.A., Malcherek, T., Schlüter, J., Hibbs, D.E.,

Welch, M.D. and Downes, P.J. (2013) The composition dependent structural transformation series of the paratacamite group. In Preparation.

Sciberras, M.J., Leverett, P., Williams, P.A., Welch, M.D., Malcherek, T., Schlüter, J. and

Mihailova, B. (2013) Raman spectroscopy of the substituted basic Cu(II) chloride phase transformations, Part 1: Reference spectra. In Preparation.

Sciberras, M.J., Leverett, P., Williams, P.A., Malcherek, T., Schlüter, J. and Mihailova, B.

(2013) Raman spectroscopy of the substituted basic Cu(II) chloride phase transformations, Part 2: Raman spectroscopy of the synthetic Cu4-xMx(OH)6Cl2

substitution series. In Preparation.

ABSTRACT

viii

ABSTRACT

This thesis reports results from a comprehensive crystallographic and spectroscopic

investigation of natural and synthetic samples of the basic Cu(II) chloride minerals, with

focus on substitution phenomena in the group. A series of composition-induced phase

transformations occur in the group. They are possibly anatacamite P1�→ clinoatacamite P21/n

→ paratacamite R3� → herbertsmithite R3�m, when Zn is the dominant substituting cation in

the formula Cu4-xZnx(OH)6Cl2. The role of paratacamite in this series is poorly understood.

It was originally described with the formula Cu2(OH)3Cl, but it is likely its structure is

stabilised by the presence of essential Zn. An analogous series with Ni substitution to the

end-member R3�m phase known as gillardite, which is isostructural with herbertsmithite, also

occurs. Based on the group theory, two series of space group symmetries are possible, P1�→

R3� → R3�m and P1�→ C2/m → P21

A crystal from the type specimen of paratacamite (British Museum specimen

BM86958), with composition Cu

/c → R3�m. These possibilities are explored through an

analysis of the inherent structural changes related to compositional effects.

3.71Zn0.29(OH)6Cl2, was analysed by single-crystal X-ray

diffraction at several temperatures (100, 200, 300, 353, 393 and 423 K). Its structure at 300 K

is confirmed in space group R3� for the unit cell a ≈ 13.6, c ≈ 14.0 Å, with a pronounced

substructure corresponding to a′ ≈ ½a and c′ ≈ c, in space group R3�m (analogous to the

structure of herbertsmithite). Paratacamite undergoes a reversible phase transformation to the

R 3� m substructure at elevated temperatures. This establishes that paratacamite is

thermodynamically stable at 300 K for the composition studied. The loss of the superstructure

at elevated temperature indicates that the substitution phenomenon is statistical between both

interlayer metal sites, rather than being preferential at M(1) as has been suggested in the

literature. It is suggested that the observed (2+2+2) octahedral configuration at M(2) is a

consequence of a superimposition of non-tetragonally elongated Zn(OH)6 octahedra with

dynamic (4+2) Jahn-Teller distorted Cu(OH)6

In the course of this investigation the single-crystal X-ray structure of two new

analogues of paratacamite were determined. One is an Mg-rich specimen from the Quebrada

Mine, Camerones, Chile, Cu

octahedra occupying two orientations.

3(Mg,Cu)(OH)6Cl2, and the other is a Ni-rich specimen from the

Carr Boyd Rocks Mine, Western Australia, Australia, Cu3(Ni,Cu)(OH)6Cl2. The supercell

analogous to that reported for paratacamite was identified and the structure was solved in

space group R3�. Both analogues exhibit a substructure with a' ≈ ½a, c' ≈ c in space group

ABSTRACT

ix

R3�m. They are the first examples of naturally occurring substituted paratacamite congeners to

be reported. Substitution phenomena in the Mg analogue is confirmed as being statistical in

nature by refinement of the site scattering factors of interlayer sites M(1) and M(2). The

substitution behaviour in the Ni analogue, as well as paratacamite containing from Zn from

the holotype specimen was assumed to be statistical throughout this investigation, but the

possibility remains that Zn and Ni preferentially occupy one of the interlayer sites.

Furthermore, the single-crystal X-ray structure of naturally occurring Cu3(Co,Cu)(OH)6Cl2

Raman spectroscopy analyses, supported by single-crystal X-ray diffraction of

samples exhibiting a range of compositions, have revealed several trends associated with

variation in composition. An examination of both natural and synthetic samples indicate that

the transformation series proceeds as P1�→ P2

from the Torrecillas Mine, Salar Grande provence, Chile, is reported with unit cell parameters

and structure analogous to that of herbertsmithite.

1/c → R3�m, with decreasing Cu2+ content in

Cu4-xMx(OH)6Cl2. The composition-induced changes in the paratacamite Raman spectrum

with high interlayer Cu2+ content, suggests a distortion towards that of anatacamite. With

excess substitution for interlayer Cu2+, the structure converges with that of the R 3� m

aristotype. This corresponds to the space group series P1�→ R3� → R3�m, with decreasing Cu2+

Finally, the synthetic series of Zn- and Ni-substituted members was explored in order

to determine the behaviour of the solid state activity coefficient (γ). In clinoatacamite, Zn

content as described above. The stability of paratacamite appears to be dependent on the type

of substituting cation.

2+

occupation exhibits non-ideal behaviour for dilute solid solutions and γ > 1. With increasing

Zn content, γ approaches unity near the composition Cu3.80Zn0.20(OH)6Cl2. The Ni-

substitution series shows regular behaviour of γ with dilute solid solutions, and γ < 1. The

distribution of γ values with increasing Ni content indicates non-ideal behaviour. Distribution

coefficients for the synthetic rhombohedral series were calculated and these demonstrate the

incongruent nature of the dissolution of the phases in aqueous media.

CHAPTER 1

1

CHAPTER 1 - INTRODUCTION

The substitution of ions in a crystalline substance can induce significant changes in its

thermodynamic stability and physical characteristics. A good example of this is illustrated by

the discovery of high-Tc superconductors by Bednorz and Müller (1986). They prepared

polycrystalline samples with Ba2+ substituting for La3+ in the compound BaxLa5-xCu5O5(3-y

This work reports the results of a crystallographic and spectroscopic study of

substitution phenomena that take place in the basic Cu(II) chloride minerals. One particular

issue concerning the differentiation of rhombohedral members of the transformation series

was investigated using both naturally occurring and synthetic materials. A major outcome of

this thesis is the determination of an unambiguous method for the identification of the

mineral paratacamite, originally described as Cu

),

x = 1 and 0.75, y > 0, and reported percolative superconductivity in the 30 K range. This

result accelerated research into the properties of substitution into particular compounds and

led to the discovery of superconducting materials with a critical temperature above 100 K

(Batlogg, 1991). Ion substitution can alter a material’s stability field with respect to other

phases. This can lead to a series of structural transformations that exhibit a compositional

dependency. One such series of transformations involving members of the basic Cu(II)

chloride minerals is investigated in this work. Members of the basic Cu(II) chlorides have a

unique geometrically frustrated magnetic state derived from an exotic structural feature called

a kagomé lattice (Schores et al., 2005). The full realisation of a functional kagomé lattice is

generated by complete doping of a diamagnetic cation in place of a Cu(II) ion located

between the kagomé layers. This results in the suppression of long range magnetic order and

spin freezing down to at least 0.05 K (Han et al., 2012). Such materials may allow a better

understanding of the nature of superconductivity and may also provide some help in the

development of quantum computers (Colman et al., 2008, 2010).

2(OH)3

This investigation has resulted in the discovery of new naturally occurring members

of the substituted basic Cu(II) chlorides, which are either analogues of paratacamite or

Cl by Smith (1906). This work settles

a dispute in the literature that has lasted for over 100 years, since paratacamite was

discredited as a mineral by Ungemach (1911). The structure of paratacamite is confirmed in

space group R3� on the unit cell reported by Frondel (1950) and Fleet (1975), and the origin of

particular anomalous octahedral distortions inherent in this structure has been elucidated.

CHAPTER 1

2

herbertsmithite, Cu3ZnCl2(OH)6. The thermodynamics of cation substitution in the group

have been explored using solution methods for the Cu2+/Zn2+ and Cu2+/Ni2+

1.1 THE BASIC Cu(II) CHLORIDE MINERALS

system.

The formula Cu2(OH)3Cl is known to exist in the four polymorphous minerals

atacamite (orthorhombic), anatacamite (triclinic), botallackite and clinoatacamite (both

monoclinic), with structures reported by Wells (1949), Malcherek and Schlüter (2009),

Hawthorne (1985), and Grice et al. (1996), respectively. In natural settings, Cu2(OH)3Cl

commonly forms in the oxidised zones of base metal ore deposits in arid climates (Bandy,

1938). However, they are not limited to such locations. The combination of conditions that

promote or inhibit the nucleation and growth of one particular polymorph over another has

been investigated and can be used to identify prevailing environmental conditions at the time

of formation (Skarkey and Lewin, 1971; Woods and Garrels, 1986b; Pollard et al., 1989).

The Cu2(OH)3Cl polymorphs are well known as corrosion products of copper, bronze and

brass artefacts (Dei et al., 1998; Scott, 2000). Minerals in this group have also been identified

as alteration products of mineral pigments (azurite, Cu3(CO3)2(OH)2, or malachite,

Cu2(CO3)(OH)2

A series of minerals related to the Cu

) used in wall paintings from ancient times (Scott, 2000).

2(OH)3Cl polymorphs are characterised by

solid-solution phenomena in the group. Substitution for Cu2+ is known to occur commonly

with divalent Zn, Ni, Co, Mn, Fe and Mg ions, though some examples in the basic Cu(II)

chloride group are not yet known in Nature (Feitknecht and Maget, 1949a, b; de Wolff, 1953;

Kracher and Pertlik, 1983; Jambor et al., 1996; Braithwaite et al., 2004). The most

extensively studied solid-solution series in the group occurs between clinoatacamite and the

end-member composition Cu3Zn(OH)6Cl2, known as herbertsmithite (Jambor et al., 1996;

Braithwaite et al., 2004). An analogous situation exists for Ni2+ substitution and leads to the

mineral gillardite, Cu3Ni(OH)6Cl2, which is isostructural with herbertsmithite (Colchester et

al., 2007; Clissold et al., 2007). The mineral paratacamite was originally considered to be a

polymorph of Cu2(OH)3Cl with rhombohedral symmetry (Frondel, 1950). Recent studies

have established that the rhombohedral phase is stabilised by partial substitution of Cu2+ by

another divalent cation with comparable ionic radius, such as Zn2+ or Ni2+ (Jambor et al.,

1996). It was suggested by these authors that clinoatacamite undergoes a composition-

dependent phase transformation to paratacamite. However, the thermodynamic stability of

paratacamite is not known and it is possible that much of the previous data was collected

from mixtures of the basic Cu(II) chlorides. Additional minerals related to the substituted

CHAPTER 1

3

members of the group are kapellasite, trigonal, P3�m1, which is a dimorph of herbertsmithite

(Kraus et al., 2006), and haydeeite, which is the Mg analogue of kapellasite (Malcherek and

Schlüter, 2007). Crystallographic details concerning the basic Cu(II) chlorides minerals are

given in Table 1.1.

Table 1.1. Crystallographic data for the Cu2(OH)3Cl polymorphs and related minerals.

Mineral Crystal system SG a or α b or β c or γ Ref. Atacamite orthorhombic Pnma 6.030(2) Å 6.865(2) Å 9.120(2) Å 1 Cu2(OH)3Cl Botallackite monoclinic P21/m 5.717(1) Å 6.126(1) Å 5.636(1) Å 2 Cu2(OH)3

Anatacamite triclinic P1� 9.1646(9) Å 9.2029(8) Å 9.2102(8) Å 3 Cu

Cl 93.07(1)°

2(OH)3Cl 95.858(6)° 96.290(7)° 96.507(2)° Clinoatacamite monoclinic P21/n 6.157(2) Å 6.814(3) Å 9.105(5) Å 4 Cu2(OH)3

Paratacamite rhombohedral R3� 13.654(5) Å 14.041(6) Å 5 Cu

Cl 99.65(4)°

2(OH)3Cl*

Herbertsmithite rhombohedral R3�m 6.834(1) Å 14.075(2) Å 6 Cu

3Zn(OH)6Cl

Gillardite rhombohedral R3�m 6.8364(1) Å 13.8459(4) Å 7 Cu

2

3Ni(OH)6Cl

Kapellasite trigonal P3�m1 6.300(1) Å 5.733(1) Å 8 Cu

2

3Zn(OH)6Cl

Haydeeite trigonal P3�m1 6.2733(4) Å 5.7472(5) Å 9 Cu

2

3Mg(OH)6Cl2

[1] Parise and Hyde (1986); [2] Hawthorne (1985); [3] Malcherek and Schlüter (2009); [4] Jambor et al. (1996); [5] Fleet (1975); [6] Braithwaite et al. (2004); [7] Clissold et al. (2007); [8] Krause et al. (2006); [9] Malcherek and Schlüter (2007). *

The apparent compositional-induced phase transformations that occur in this group

have not been fully characterised and aspects of the thermodynamics involved with

substitution are unknown. A high degree of contradictory literature on the group, particularly

concerning paratacamite, has marred references to other members (Jambor et al., 1996). This

is most likely due to the structural similarity of paratacamite with the other related minerals

(Malcherek and Schlüter, 2009). In addition, mixed reports on the true composition of

Not the true ideal composition (vide infra).

CHAPTER 1

4

paratacamite have complicated a formal description (Smith, 1906; Frondel, 1950; Kracher

and Pertlik, 1983; Braithwaite et al., 2004). These issues necessitate a complete

reinvestigation of the group with emphasis on the determination of a formal description of

paratacamite.

1.2 PARATACAMITE AND THE SUBSTITUTED PHASES

Paratacamite was first reported by Smith (1906), working on material from the

Herminia and Generosa mines, Sierra Gorda, Chile, and the Bolaco mine, San Cristóbal,

Chile. It was noted that the composition was the same as that of atacamite, based on chemical

analysis of 0.5132 g of material from the type specimen. The pseudo-cubic nature of the

crystals is better described as being rhombohedral or pseudo-rhombohedral and twinning in

the samples studied was ubiquitous. In addition, optical characters were not consistent with

the apparent morphology and only minute crushed fragments showed extinction between

crossed nicols. A few crystals were biaxial and indices of refraction were found to be nearly

the same in all orientations. The mean refractive index (589 nm) was approximately 1.846,

which is similar to the mean for atacamite. Type material was preserved in the Natural

History Museum, London (BM86958). Recasting of the original analysis of material from the

Generosa mine gave CuO 73.96, Cl 15.97, H2O 14.10, O ≡ Cl -3.61, total 100.42 wt%. The

empirical formula based on four anions pfu is Cu1.92Cl0.93O3.07H3.23, and the simplified

formula is thus Cu2(OH)3Cl; the absence of Zn is noteworthy.

Frondel (1950) showed that a specimen of paratacamite (USNM 95146), sourced from

the same material as that used by Smith (1906) had the same powder X-ray characteristics as

two specimens from Remolinos, Vallinar, Chile and Sierra Gorda, Chile (Harvard University

collection specimens 97523 and 82883, respectively) and to corrosion products formed by sea

water from copper and brass sheets. Recasting the analysis of the Remolinos material gave

CuO 74.26, Cl 16.29, H2O 13.13, O ≡ Cl -3.68, total 100.00 wt%. The empirical formula

based on four anions pfu is Cu1.97Cl0.97O3.03H3.08, with the simplified formula Cu2(OH)3Cl;

again, the absence of Zn is noted. Crystals were stated to be rhombohedral, a = 13.65,

c = 13.95 kX, with a marked pseudo-cell (a′ = ½a, c′ = c). Both the Remolinos and type

material from Sierra Gorda (BM86958) were found to be uniaxial (+), with ω = 1.843,

ε = 1.849 (Remolinos) and ω = 1.842, ε = 1.848 (Sierra Gorda). Extinction was described as

undulant or patchy. Many grains on both specimens were found to be biaxial with 2V up to

50o and exhibiting strong dispersion, r > v.

CHAPTER 1

5

Some of the above data were confirmed by Jambor et al. (1996) in their description of

clinoatacamite. Thirteen grains on specimen USNM 95146 were found to contain no Zn and

four of those selected for X-ray studies showed each to be clinoatacamite. Other grains on the

same specimen were described as being “zincian paratacamite” and were found to be biaxial

(–). Clinoatacamite itself has indices of refraction > 1.8, is biaxial (–) with strong dispersion,

r << v, 2V = 75o and nonpleochroic. On the basis of the above, Jambor et al. (1996) inferred

that the “biaxial paratacamite” of Frondel (1950) was clinoatacamite. Furthermore, a

subsequent analysis of BM86958 (Kracher and Pertlik, 1983) showed the analysed material

to contain 2.45 wt% Zn, confirmed in another analysis by Jambor et al. (1996). The report of

a “new” mineral, “anarakite”, (Cu,Zn)2(OH)3Cl, by Adib and Otteman (1972) must also be

considered. This phase was reported to be biaxial (+), α = 1.842, γ = 1.849 and 2V = 40o. The

empirical formula was determined as (Cu1.65Zn0.35)(OH)2.99Cl1.05. Analysis of similar

material (William Pinch collection 508) gave a range of compositions varying between

(Cu1.70Zn0.30)(OH)2.98Cl1.02 and (Cu1.86Zn0.14)(OH)2.99Cl1.01 and optical character uniaxial (–

), n > 1.8 with some grains being biaxial, 2V ca 5o (Jambor et al., 1996). Single-crystal X-ray

analysis of one grain gave a hexagonal cell with a = 6.832, c = 14.042 Å and longer

exposures gave evidence of the supercell (a′ = 2a, c′ = c). Originally, Adib and Otteman

(1972) reported a monoclinic cell with a = 11.901, b = 6.830, c = 10.162 Å and β = 112.87o,

but Kracher and Pertlik (1983) noted that the synthetic monoclinic phase, Cu2(OH)3Cl, of

Oswald and Guenter (1971) is compatible with these parameters by a transformation of

[1�00/01�0/102]. Jambor et al. (1996) further showed that the cell of Oswald and Guenter

(1971), space group P21/a, a = 11.83(1), b = 6.822(3), c = 6.166(5) Å and β = 130.62(3)o can

be reduced to a = 6.166, b = 6.805, c = 9.112 Å and β = 99.71o, space group P21/n, that of

clinoatacamite, a = 6.144, b = 6.805, c = 9.112 Å and β = 99.55o, space group P21/n. It is

certainly remarkable that this can be transformed to a pseudo-rhombohedral cell with

a = 13.610, b = 13.626, c = 14.031 Å, α = 89.47, β = 90.00 and γ = 119.96o

One particular problem with all of the above concerns the fact that most studies have

involved specimens which contain at least two different basic Cu(II) chlorides. The discovery

of three further related minerals has helped to clarify the situation. Herbertsmithite, ideally

Cu

, as compared to

the above cell of Frondel (1950) and that reported for the single-crystal structure of

paratacamite (Fleet, 1975), R 3� , a = 13.654(5), c = 14.041(6) Å, with a pronounced

substructure, R3�m, a being halved.

3Zn(OH)6Cl2, was described using material from the Kali Kafi and Chah Khouni mines,

Anarak, Iran, and the Los Tres Presedentes mine, Sierra Gorda, Chile, and is uniaxial (–)

CHAPTER 1

6

(Braithwaite et al., 2004). Its structure is that of the R3�m rhombohedral subcell described by

Fleet (1975) for paratacamite, with a = 6.834(1), c = 14.075(2) Å, containing a single

crystallographic site which links sheets of Jahn-Teller distorted octahedra with the

composition [Cu2(OH)4Cl2]2–. Shortly thereafter, the nickel analogue gillardite,

Cu3Ni(OH)6Cl2, uniaxial (+), R 3� m, a = 6.8364(1), c = 13.8459(4) Å, was described

(Colchester et al., 2007; Clissold et al., 2007). This R3�m structure is considered to be the

aristotype for the group (Malcherek and Schlüter, 2009) and is the only structure from the

abovementioned group that possesses an ideal kagomé arrangement of Cu2+ ions (Figure 1.1).

Finally, the structure and properties of anatacamite from the La Vendida mine, Sierra Gorda,

Chile, with composition Cu1.97Ni0.03(OH)2.99Cl1.01

Figure 1.1. A comparison of the kagomé planes in herbertsmithite (Braithwaite et al., 2004), paratacamite (Fleet, 1975), clinoatacamite (Grice et al., 1996), and anatacamite (Malcherek and Schlüter, 2009). Blue spheres are Cu

, was reported (Malcherek and Schlüter,

2009, 2010) This mineral is structurally virtually indistinguishable from clinoatacamite or

paratacamite. Optically, anatacamite is biaxial, but no refractive indices were measured.

2+ ions. Image adapted from Malcherek and Schlüter (2009).

CHAPTER 1

7

1.3 PHASE IDENTIFICATION

Structural relationships in the group were explored by Malcherek and Schlüter (2009).

They have shown that atomic displacements from their ideal position in the R3�m aristotype

model are responsible for generation of the lower symmetry structures. For this reason, the

authors suggested that only single-crystal methods would be able to distinguish the minerals

from each other. The set of supercell reflections characteristic of paratacamite is reportedly

very weak in intensity (Fleet, 1975; Jambor et al., 1996). Long count times using single-

crystals are generally required to detect an unambiguous set of reflections corresponding to

the supercell, while the subcell data set, corresponding to the aristotype structure, is high in

intensity.

Table 1.2 gives relevant powder X-ray data sets. The correspondence of the data for

“anarakite”, zincian paratacamite, clinoatacamite, anatacamite and herbertsmithite is

remarkable. Braithwaite et al. (2004) also reported powder X-ray data for paratacamite

obtained from samples 14 and 17 of specimens WHP 593/374 and HMx/III/182, respectively,

both originally discovered at Anarak, Iran. The authors noted the similarity of this data to

powder patterns determined for herbertsmithite samples 22 and 24 from specimens obtained

from Sierra Gorda, Chile. It is not therefore surprising that single-crystal data sets for

anatacamite can be indexed to P1�, P21/n, R3� and R3�m cells with varyingly acceptable values

of Rint (Malcherek and Schlüter, 2009). However, accurate measurements of the Irel = 100,

5.432 Å (1�11 + 11�1 + 1�1�1) reflection may be useful and it is possible to discern differences

between the powder X-ray diffraction (PXRD) data for anatacamite and herbertsmithite. The

most pronounced distinction between powder data in Table 1.2 can be made for identification

of clinoatacamite by observation of the strong Irel = 60, 2.266 Å (220) and Irel

However, much of the previous work on these minerals was made using powder X-

ray methods, which presents a problem regarding interpretation of this data. Table 1.3 lists

various naturally occurring and synthetic materials as well as the method used to identify

them. The majority of phases have been examined PXRD methods. The identification of an

R 3�𝑚𝑚 paratacamite supercell by Oswald and Feitknecht (1964) and the refinement of a

= 50, 2.243 Å

(004) reflections. Other reflections may be used if the data collected has relatively high

resolution and intensity. The identification of the paratacamite supercell using powder X-ray

data alone cannot be achieved. Powder X-ray data of these minerals may exhibit considerable

deviation influenced by solid solution effects.

CHAPTER 1

8

supercell for R3� paratacamite reported by Chu et al. (2011), both using powder X-ray

methods, appears anomalous considering the above information.

Table 1.2. Powder X-ray data for structurally related Cu2(OH)3

65 5.476 100 5.48 100 5.47 5.472 𝟏𝟏�01 100 5.432 55 5.466 <5 5.03* 15 4.697 40 4.69 30 4.68 4.677 101 10 4.657 14 4.702 5 4.52 <5 4.54 4.532 110 3 4.537

Cl minerals. “anarakite”a zincian paratacamitea clinoatacamitea anatacamiteb herbertsmithitec

Irel dmeas Irel dmeas Irel dmeas dcalc hkl Irel dmeas Irel dmeas

10 3.429 15 3.424 20 3.406 3.409 1�12 4 3.395 5 3.423 3.407 020 1 3.028 <5 3.019 3.006 112 20 2.901 30 2.896 40 2.887 2.892 1�21 31 2.889 11 2.899 2.882 1�03 100 2.755 75 2.759 60 2.767 2.771 211 100 2.764 15 2.739 20 2.726 70 2.742 2.739 013 94 2.747 2.736 2�02 13 2.730 10 2.713 2.714 022 10 2.342 10 2.343 20 2.339 2.339 202 2 2.342 4 2.346 70 2.263 65 2.263 60 2.266 2.266 220 56 2.257 36 2.266 50 2.243 2.244 004 10 2.215 <5 2.210 5 2.208 2.209 2�13 2 2.210 10 2.042 2.045 2.049 3�01 4 2.040 10 2.035 10 2.035 10 2.027 2.027 123 6 2.027 <5 1.930 1.940 310 1 1.934 15 1.904 10 1.901 5 1.907 1.906 301 5 1.905 10 1.895 1.894 213 7 1.895 20 1.824 1.824 3�03 20 1.817 25 1.817 20 1.817 1.818 2�31 21 1.812 13 1.820 10 1.807 1.809 033 1.806 1�05 <5 1.751 1.748 231 3 1.745 1 1.752 30 1.708 35 1.708 50 1.704 1.705 𝟐𝟐�24 29 1.702 18 1.709 1.704 040 <5 1.661 <5 1.664 1.664 321 1 1.664 5 1.630 5 1.626 1.628 2�23 2 1.626 3 1.631 <5 1.601 1.601 141 1 1.606 1.600 3�14 2 1.600 1 1.599 10 1.514 7 1.509 5 1.516 1.518 400 1.517 4�02 4 1.513 15 1.496 10 1.494 10 1.504 1.503 224 4 1.494 3 1.496 1.501 4�11 10 1.487 1.489 233 1.484 125 <5 1.471 5 1.471 1.471 3�05 1 1.472 1.467 143 5 1.447 5 1.445 1.447 241 1 1.448 1.446 2�42 5 1.420 1.422 3�33 1.418 323 10 1.384 15 1.380 10 1.387 1.386 4�22 6 1.381 5 1.377 1.377 242 5 1.376 10 1.362 10 1.368 1.370 026 4 1.363 1.368 4�04

CHAPTER 1

9

Table 1.2 continued 5 1.356 1.357 044 3 1.354 5 1.350 <5 1.346 1.348 234 2 1.351 1.347 051 <5 1.311 1.313 3�16 10 1.270 10 1.271 1.271 422 3 1.268 5 1.272 5 1.260 1.261 206 1.260 017 15 1.244 1.247 4�15 3 1.245** 1.245 3�43 1.243 305 1.243 250 1.243 2�51 1.240 053 aJambor et al. (1996). bMalcherek and Schlüter (2010). cBraithwaite et al. (2004). *This line was attributed to the presence of trace amounts of atacamite by Jambor et al. (1996) and is the Irel = 100 reflection in that mineral. **Plus six extra lines to d = 1.010 Å.

Table 1.3. Natural and synthetic compounds of the basic chlorides and the analytical method used for identification. Space group Mineral name Method(s) used Ref. Cu2(OH)3Cl P1� Anatacamite Single-crystal XRD 1 Cu(OH)Cl P21/a Belloite Single-crystal XRD 2, 3 Cu2(OH)3Cl P21/c Clinoatacamite PXRD, single-crystal XRD 4 α-Cu2(OH)3Cl P21/m Botallackite PXRD, single-crystal XRD 5, 6 δ-Cu2(OH)3Cl Pnam Atacamite PXRD, single-crystal XRD 5, 7 β-Mn2(OH)3Cl Pnam Kempite PXRD, single-crystal XRD 5, γ-Fe2(OH)3Cl Pnam Hibbingite Single-crystal XRD 8 β-Mg2(OH)3Cl Pnam PXRD 5 γ-Cu2(OH)3Cl R3�𝑚𝑚 Paratacamite PXRD 5 R3� Paratacamite PXRD, single-crystal XRD 9, 14 Cu3Zn(OH)6Cl2 R3�m Herbertsmithite PXRD, single-crystal XRD 10 Cu3Ni(OH)6Cl2 R3�m Gillardite PXRD, single-crystal XRD 11 β-Co2(OH)3Cl R3�m PXRD 5, 15 β-Fe2(OH)3Cl R3�m PXRD 5 α-Cu3Zn(OH)6Cl2 P3�m1 Kapellasite Single-crystal XRD 12 α-Cu3Mg(OH)6Cl2 P3�m1 Haydeeite Single-crystal XRD 13 α-Co2(OH)3Cl P3�m1 PXRD 5 α-Fe2(OH)3Cl P3�m1 PXRD 5 α-Mn2(OH)3Cl P3�m1 PXRD 5 α-Ni2(OH)3Cl P3�m1 PXRD 5 α-Mg2(OH)3Cl P3�m1 PXRD 5 [1] Malcherek and Schlüter (2009); [2] Schlüter et al. (2000); [3] Effenberger (1984); [4] Jambor et al. (1996); [5] Oswald and Feitknecht (1964); [6] Hawthorne (1985); [7] Wells (1949); [8] Saini-Eidukat et al. (1995); [9] Fleet (1975); [10] Braithwaite et al. (2004); [11] Clissold et al. (2007); [12] Krause et al. (2006); [13] Schlüter and Malcherek (2007); [14] Chu et al. (2011); [15] de Wolf (1953)

CHAPTER 1

10

There are more R3�m phases than analogues for paratacamite. Considering the long

history of the mineral and the commonality of substitution effects in the group, it is likely that

some new phases have been mistaken for paratacamite or other members. This is certainly

true for clinoatacamite, herbertsmithite and gillardite, and may also apply to anatacamite.

1.4 SYNTHESIS AND STABILITY

Oswald and Feitknecht (1964) reported end-member M2(OH)3Cl compositions of

substituted synthetic phases for several structure types, where M2+ = Co, Fe, Ni, Mg or Mn.

Naturally occurring phases more commonly exhibit the general formula Cu4-xMx(OH)6Cl2,

where 0 ≤ x ≤ 1 and M is a divalent cation such as those above. There are several

reproducible synthetic methods available in the literature for the basic copper chlorides, save

for the mineral anatacamite (Feitknecht and Maget, 1949a; Sharkey and Lewin, 1971; Pollard

et al., 1989; Jambor et al., 1996). Although atacamite is not the thermodynamically stable

phase at 298.2 K, Sharkey and Lewin (1971), and Pollard et al. (1989) reported that it may be

stabilised by addition of excess Cl- (> 0.4 mol dm-3) to the reaction solution. A similar

stabilisation of the atacamite phase is achieved by addition of Ca2+ ions to CuCl2 solutions

(Garrels and Stine, 1948). In contrast, higher concentrations of Cu2+ promote formation of

clinoatacamite (incorrectly reported as paratacamite) (Sharkey and Lewin, 1971; Pollard et

al., 1989). Sharkey and Lewin (1971) suggested that the formation of transient CuCl+(aq),

which becomes a significant species in solution at around pH 4 if the concentration of CuCl2

is high (> 0.001 F), either promotes atacamite formation or inhibits paratacamite (actually

clinoatacamite) crystallisation. Botallackite may be synthesised from a wide range of solution

conditions, provided the concentration of Cu2+

Prior to Jambor et al. (1996), the order of thermodynamic stability of the Cu

is low enough so as to not force initial

crystallisation of clinoatacamite. The mineral must be isolated from solution and dried

quickly after crystallisation as it rapidly decomposes to the more thermodynamically stable

phases (Pollard et al., 1989). These authors remarked on the solution conditions required to

stabilise one particular polymorph and concluded that no simple explanation based on

speciation could account for the recrystallisation and nucleation phenomena.

2(OH)3Cl

polymorphs at 298 K was established by several authors as paratacamite (actually

clinoatacamite) > atacamite > botallackite, which coincided with the same order of hydrogen

bond strengths in the minerals (Garrels and Stine, 1948; Garrels and Dreyer, 1952; Oswald

and Feitknecht, 1964; Walter-Levy and Goreaud, 1969; Sharkey and Lewin, 1971, 1972;

Woods and Garrels, 1986b; Pollard et al., 1989). Values of ΔGfɵ at 298.2 K for paratacamite,

CHAPTER 1

11

atacamite and botallackite have been reported as -1341.8, -1335.1 and -1322.6 kJ.mol-1,

respectively (Woods and Garrels, 1986a; Pollard et al., 1989). The ΔGfɵ value for

paratacamite reported by Woods and Garrels (1986a) is adopted here as the value for

clinoatacamite because the composition and synthetic approach used by these authors is

consistent with that for the latter (Jambor et al., 1996). There are no data in the literature for

the stabilities of any substituted basic Cu(II) chloride mineral. Presumably, identifying

suitable end-member compositions for paratacamite have complicated matters.

Jambor et al. (1996) reported the synthesis of paratacamite from a range of solution

concentrations of CuCl2–ZnCl2, and CuCl2–Zn(NO3)2. They also reported a synthesis for

Co-rich clinoatacamite and paratacamite with a wider phase transition boundary as compared

with the Cu–Zn series. Their attempts to produce a similar synthesis for Ni-rich

clinoatacamite and paratacamite resulted in mixed Ni(OH)2 and Cu2(OH)3

Cl phases.

However, phase identification was made using PXRD and some validation of these methods

is warranted.

All of the above methods produce a microcrystalline product which can only be

analysed by powder methods which limits the effectiveness of phase identification in the

group. Recently, a synthetic method for the growth of large single-crystals of these materials

was reported (Schores et al., 2005) and developed further by Chu et al. (2011) and Han et al.

(2011). The method takes advantage of a recrystallisation process under hydrothermal

conditions with a temperature gradient across the reaction vessel. The synthesis reportedly

takes in excess of 10 months before the Ostwald ripening of “large” crystals is sufficient.

Using this method, single-crystals of paratacamite have been reported by Schores et al.

(2005), Wulferding et al. (2010), Chu et al. (2010), Chu (2011), and Han et al. (2011) for

characterisation of their antiferromagnetic properties. However, single-crystal X-ray data on

paratacamite reported by these authors is entirely consistent with the unit cell and symmetry

of herbertsmithite. Two additional authors, Mendels et al. (2007) and Chu et al. (2011),

reported analyses of synthetic single-crystals of paratacamite that were identified using

powder X-ray diffraction. The kagomé planes in paratacamite are non-ideal and interpretation

of magnetic susceptibility data, Raman spectroscopy, etc., made on a particular material is

influenced by a prior knowledge of its structure.

CHAPTER 2.1

12

CHAPTER 2 – CRYSTALLOGRAPHIC STUDIES 2.1 THE TEMPERATURE-RELATED REVERSIBLE PHASE TRANSFORMATION BETWEEN PARATACAMITE AND HERBERTSMITHITE 2.1.1 INTRODUCTION

The structure of paratacamite was determined by Fleet (1975) in space group R3�

assuming the formula Cu2(OH)3Cl. Notably, paratacamite has a strong subcell with a' = 1/2a

and c' = c, in space group R3�m. This subcell corresponds to the structure of herbertsmithite

and gillardite, Cu3M(OH)6Cl2, M2+ = Zn, Ni, respectively (Braithwaite et al., 2004; Clissold

et al., 2007). At ca 1/3 occupation of Zn or Ni in the interlayer structure of clinoatacamite a

rhombohedral phase assumed to be paratacamite is stabilised (Jambor et al., 1996). It is

believed that the substituted basic Cu(II) chloride minerals are related through a series of

composition-dependant phase transformations, which are possibly anatacamite P 1� →

clinoatacamite P21/n → paratacamite R3� → herbertsmithite R3�m. According to group theory,

two different space group chains are possible; P1�→ C2/m → P21/c → R3�m, and P1�→ R3� →

R3�m. These possibilities have been discussed by Malcherek and Schlüter (2009) and the

space group chain P1�→ P21/c (P21

2.1.2 THE STRUCTURE OF PARATACAMITE

/n) → R3�m was suggested as the most likely order of

transformations with increasing substitution for Cu.

Aspects of the structure of paratacamite have remained enigmatic since its description

by Fleet (1975), principally because of the highly unusual rhombic distortion of one of the

two interlayer metal positions (Cu2). There has been no additional structural data reported for

paratacamite, although numerous reports of its occurrence have entered the literature (Smith

1906; Jambor et al., 1996; Pring et al., 1987; Braithwaite et al., 2004). This crystallographic

investigation was made using material from the type specimen of paratacamite to elucidate

the origin of the unusual octahedral distortion mentioned above and to explore the mineral’s

thermal stability.

The structure paratacamite was solved by Fleet (1975) assuming the nominal formula

Cu2(OH)3Cl, using a crystal from the type specimen (British Museum sample BM86958)

originating from the Generosa Mine, Sierra Gorda, Chile. In the R3�m subcell, the sole

interlayer Cu ion is coordinated by six symmetry-related OH– ions from adjacent sheets

CHAPTER 2.1

13

(Figure 2.1.1). The O atom is disordered over two positions. For O1, Cu–O1 = 2.041 Å, all

trans HO–Cu–OH angles are constrained to be 180o and the cis HO–Cu–OH angle is 103.5o.

For O2, Cu–O2 = 2.335 Å, all trans HO–Cu–OH angles are constrained to be 180o and the

cis HO–Cu–OH angle is 105.3o

.

Figure 2.1.1. The coordination environment of the disordered Cu(OH)6

Both geometries are highly unusual for six-coordinate Cu(II) with OH

group between the sheets in the R3�m subcell of Fleet (1975); Cu atoms appear as blue spheres, O atoms appear as red spheres. The unit cell is outlined.

– ligands. Fleet

(1975) also reported that O1 and O2 (of Figure 2.1.1) were fractionally occupied (0.76(9) and

0.24(9), respectively). Thus the representation of Figure 2.1 may be viewed as a

superimposition of four short Cu–O1 bonds (the higher occupancy) and two long Cu–O2

bonds. The occupancies should be 0.67 and 0.33, respectively. The result is a space group-

imposed superimposition of three different orientations of the common (4+2) Jahn-Teller

distorted Cu(II) geometry. The Cu1 site of the R3�m substructure is of course constrained to

be split in the R3� structure and the coordination spheres of the interlayer Cu sites are depicted

in Figure 2.1.2. Cu1 is bonded to six equivalent OH– groups (O4) with Cu1–O4 = 2.121 Å.

Trans HO–Cu1–OH angles are constrained to be 180o and the cis HO–Cu1–OH angle is

105.4o. Cu2 is bonded twice (trans) to three crystallographically-independent OH– ions with

all trans HO–Cu2–OH angles equal to 180o. Cu2–O1 = 1.933, Cu2–O2 = 2.186 and Cu2–O3

= 2.204 Å, respectively, and O1–Cu2–O2 = 102.8, O1–Cu2–O3 = 101.8 and O2–Cu2–O3 =

105.1o, or their supplements, respectively.

CHAPTER 2.1

14

Figure 2.1.2. The coordination environment of the Cu(OH)6 groups between the sheets in the R3� cell of Fleet (1975); Cu atoms appear as blue spheres, O atoms appear as red spheres. The unit cell is outlined.

Again, the geometry for Cu1 is unacceptable for Cu(II) with six OH– ligands. Cu2

possesses a compressed tetragonal and rhombically distorted coordination sphere which,

although not unprecedented in the literature, is confined to about 10 copper complexes

containing chelating ligands and only three containing unidentates or bridging bidentates

(Hathaway, 1987). The geometry is unprecedented for Cu(II) complexes containing only OH–

or H2

2.1.3 SAMPLE AND ANALYSIS

O ligands. Whilst the geometry of Cu1 can be rationalised by the occupation of some

Zn in this site, the true composition of the type crystal used must be considered as unknown.

The paratacamite crystal taken from the type specimen (BM86958) and used in the

single-crystal XRD study was attached to a glass slide using Crystalbond after the X-ray

diffraction experiments were finished. The crystal was positioned so that a flat surface was

parallel to the glass slide in order to minimise differential absorption of the electron beam

during analysis. Electron microprobe analyses were made using a CAMECA SX100

microprobe operated in WDS mode with an accelerating voltage of 20 kV, 10 nA and a 20

μm spot size. Analytical data are given in Table 2.1.1. The average of five analyses gave the

empirical composition Cu3.73Zn0.29Cl1.95O6.05H6, calculated based on eight anions pfu. The

Cu1

O4

O2

O3

Cu2

O1

CHAPTER 2.1

15

sample is somewhat unstable in the beam, owing to the low analytical total and Cl- content;

electron-beam-induced Cl- migration is the most likely cause (Stormer et al., 1993). The

structure of paratacamite is composed of sheets of composition Cu3Cl2(OH)62– linked by M2+

ions lying between them (Fleet, 1975). Thus, the empirical formula may be written as

Cu3.73Zn0.29(OH)6Cl2. Normalisation based on four cations pfu gives Cu3.71Zn0.29(OH)6Cl2.

Table 2.1.1. Electron microprobe analyses of paratacamite. Range Average* Empirical Normalised (wt%) (wt%) composition composition** CuO 66.42–67.76 67.27(51) 3.73 3.71 ZnO 4.47–6.09 5.33(64) 0.29 0.29 NiO 0–0.04 0.02(2) - - MgO 0–0.02 < 0.01 - - FeO 0–0.02 < 0.01 - - CoO 0–0.01 < 0.01 - - Cl 15.6–15.85 15.69(11) 1.95 2.00 H2

2.1.3.1 Single-crystal X-ray diffraction

O 12.24 6.00 6.00 O≡Cl -3.55 Total 97.00 *Standard deviation of the average value is in parentheses. **Compositions were normalised to Σ(cations) = 4.00.

A crystal of paratacamite, from specimen BM86958, of dimensions 0.14 × 0.13 ×

0.10 mm, was attached to a non-diffracting, amorphous carbon fibre for X-ray diffraction

experiments. Data were collected on a XcaliburE four-circle diffractometer equipped with an

Eos 1K CCD detector. A Cryojet cryoheater with a liquid-nitrogen supply was used for

variable-temperature experiments at 100, 200, 300, 353, 393 and 443 K. The tip of the nozzle

of the cryoheater was positioned to within 7 mm of the crystal. Monochromatic Mo Kα

radiation (λ = 0.71073 Å) at 45 kV and 40 mA was used for all experiments. Pure ω scans

with a 1° frame-width and a 40 s frame-time were used. The data collection strategy was

determined from a 30 min initial experiment and collection of the full data set lasted 25

hours. For the non-ambient experiments, nitrogen flow-rates of 6 L min-1 onto the sample and

4 L min-1 for the shield flow were used. Temperatures were found to be within 2 K of

nominal throughout each experiment. A 15 min thermal equilibration time was used before

data collection began. A sphere of data was collected to 34° θ, with 100% completion for

data up to 30° θ. Intensity data were corrected for Lorentz, polarisation and absorption

(multiscan) effects and converted to structure factors using the program CrysAlis RED

CHAPTER 2.1

16

(Agilent, 2012). Unit cell parameters were calculated from reflections having Fobs > 7σ(Fobs).

Structure solution was made by direct methods and refinement used using the program

SHELX-97 (Sheldrick, 2008).

The unit cell metric of paratacamite for all datasets is hexagonal (trigonal).

Examination of the unconstrained triclinic cells did not indicate deviation from this metric

within two standard deviations of all cell parameters. Examination of “pseudo-precession”

photographs reconstructed using CrysAlis PRO

There is a gradual reduction in the intensities of superlattice reflections from the 100

to 353 K datasets. Systematic absences for all datasets are consistent with space groups R3�m,

R3m, R3�, R3 and R32. Structures related to paratacamite, such as herbertsmithite, have R3�m

symmetry. This space group was tested for the superstructure in order to elucidate whether

the occurrence of this unusual rhombic distortion of the M(2) coordination sphere is due to

the choice of incorrect space group. However, structure solution in R3�m did not reach

convergence for all datasets in which superlattice reflections were present. There is enough

information from the superlattice reflections to discriminate clearly between and R3�m and

R3�.

(Agilent, 2012) for hk0, h0l and 0kl sections

were made for each data set collected. A set of strong subcell reflections giving a ~ 6.85

c ~ 14 Å is evident. In the hk0 sections at temperatures between 100 and 353 K, a series of

weak reflections at half integer positions of h and k, corresponding to a superlattice with

a’ ≈ 2a and c’ ≈ c, are present. The hk0 sections derived from the 393 and 443 K datasets

lacked these superlattice reflections. The hk0 reconstructions of the initial 300, 353, 393 and

return to 300 K data sets depicting the loss and return of these superlattice reflections can be

seen in Figure 2.1.3.

Structure solution in R3� found all four Cu and the two Cl atoms. All four O atoms

appeared after the first 10 cycles of least-squares refinement. All four H atoms were found

after a further 10 cycles of weighted least-squares. All non-H atoms were refined

anisotropically. However, in the 100 K refinement two O atoms (O1 and O3) had non-

positive-definite (NPD) displacement parameters, and the 200 K refinement had one non-

positive-definite O atom (O1). NPD behaviour for the 100 and 200 K refinements appears

when H atoms are refined. It is a recurrent feature of refinements of R3� paratacamite that one

O atom (O1) has a Ueq of about half those of the other three. As temperature decreases this

atom inevitably becomes NPD. Data obtained at 393 and 443 K, which did not display any

superlattice reflections, was solved in space group R3�m. Structure solution found only two

CHAPTER 2.1

17

Cu atoms, one Cl and one O atom. A single H atom was located on a difference map. All

non-H atoms were refined anisotropically and yielded acceptable values.

Figure 2.1.3. A reconstruction of the pseudo-precession hk0 diffraction patterns of paratacamite at various temperatures. Many superlattice reflections are evident in the diffraction patterns of the two 300 K datasets, whereas in the 353 K pattern they are very weak (indicated by arrows). Superlattice reflections are absent from the 393 K pattern and are recovered on return to 300 K.

Neutral scattering factors for Cu, O, Cl and H were taken from International Tables

for Crystallography, Volume C (1992). Information relating to data collections and relevant

structure refinement details at each temperature is given in Table 2.1.2. Data relating to an R3�

and R3�m refinement for the 443 K data set are displayed for comparison. Atom coordinates

and equivalent-isotropic displacement parameters are shown in Table 2.1.3. Anisotropic

displacement parameters are given in Table 2.1.4 and selected bond lengths and angles in

Table 2.1.5.

Table 2.1.2. Data collection and structure refinement details of paratacamite at each experimental temperature.

100 K R3�

200 K R3�

300 K initial R3�

353 K R3�

393 K R3�m

443 K R3�

443 K R3�m

300 K return R3�

Unit cell a (Å) 13.6245(7) 13.6311(6) 13.6440(4) 13.6558(3) 6.8394(1) 6.8396(3) 6.8396(3) 13.6495(4) c (Å) 14.011(1) 14.018(1) 14.0354(7) 14.0428(5) 14.0716(2) 14.0834(7) 14.0834(7) 14.0359(7) Volume (Å3) 2252.3(2) 2255.7(2) 2262.8(1) 2267.9(1) 570.052(15) 570.55(4) 570.55(4) 2264.7(1) θ range (°) 2.99–34.35 2.99–34.51 2.99–34.38 2.98–34.45 3.73–33.23 3.73–34.31 3.73–34.31 2.98–34.39 µ (mm-1) 11.958 11.958 11.958 11.958 11.722 11.712 11.958 11.958 F(000) 2448 2451 2451 2451 612 612 2451 2451 Limiting indices -20 ≤ h ≤ 21 -21 ≤ h ≤ 21 -20 ≤ h ≤ 21 -20 ≤ h ≤ 21 -10 ≤ h ≤ 10 -10 ≤ h ≤ 10 -10 ≤ h ≤ 10 -20 ≤ h ≤ 21

-21 ≤ k ≤ 20 -21 ≤ k ≤ 21 -20 ≤ k ≤ 21 -21 ≤ k ≤ 20 -10 ≤ k ≤ 10 -10 ≤ k ≤ 10 -10 ≤ k ≤ 10 -20 ≤ k ≤ 20

-22 ≤ l ≤ 21 -22 ≤ l ≤ 21 -22 ≤ l ≤ 21 -21 ≤ l ≤ 21 -21 ≤ l ≤ 21 -21 ≤ l ≤ 21 -21 ≤ l ≤ 21 -21 ≤ l ≤ 21

Reflections 18909 19161 18984 18883 4191 4452 4452 18951 Unique 2032 2041 2039 2031 297 519 318 2036 I > 2σ(I) 1648 1578 1554 1391 292 504 313 1520 Rint 0.0388 0.0369 0.0302 0.0237 0.0189 0.0211 0.0215 0.031 Data/restraints/parameters 2032/0/88 2041/0/88 2039/0/88 2031/0/88 297/0/19 519/0/25 318/0/19 2036/0/88 R1 [Iobs > 2σ(Iobs)] 0.0366 0.0353 0.0304 0.0267 0.0135 0.0162 0.0162 0.0314 R1 (all data) 0.0456 0.0472 0.0434 0.0436 0.0140 0.0170 0.0167 0.0466 wR2 [Iobs > 2σ(Iobs)] 0.0781 0.0783 0.0622 0.0639 0.0348 0.0399 0.0427 0.0592 wR2 (all data) 0.0863 0.0872 0.0706 0.0747 0.0351 0.0402 0.0429 0.0678 GoF 1.064 1.053 1.084 1.084 1.095 1.181 1.193 1.059 max. shift/σ 0.001 0.001 0.001 0.001 0.002 0.000 0.000 0.001 Δρmax, Δρmin (e. Å-3) 1.54, -2.08 0.90, -1.96 0.93, -1.48 0.43, -1.45 0.41, -0.45 0.39, -0.65 0.43, -0.65 0.71, -1.57

CH

APTER

2.1

18

CHAPTER 2.1

19

Table 2.1.3. Atomic coordinates and isotropic displacement parameters (Å2) at each experimental temperature. Ue q = 1/3(U11 + U22 + U33

) 100 K (R𝟑𝟑�)

200K (R𝟑𝟑�)

x/a y/b z/c Ueq

x/a y/b z/c Ueq M1 0 0 0.5 0.00529(14)

0 0 0.5 0.00762(14)

M2 0.5 0.5 0.5 0.00507(10)

0.5 0.5 0.5 0.00752(10) M3 0.41416(3) 0.32839(2) 0.33148(2) 0.00659(9)

0.41437(2) 0.32881(2) 0.33136(2) 0.00919(9)

M4 0.41089(2) 0.57757(2) 0.33340(2) 0.00598(9)

0.41105(2) 0.57773(2) 0.33331(2) 0.00851(9) Cl1 0 0 0.19376(8) 0.00779(18)

0 0 0.19375(8) 0.01111(18)

Cl2 0.50221(5) 0.50221(5) 0.19358(5) 0.00731(12)

0.50208(5) 0.50207(5) 0.19365(5) 0.01114(13) O1 0.55579(15) 0.61962(15) 0.40094(13) 0.0040(3)

0.55615(14) 0.61986(13) 0.40094(12) 0.0059(3)

O2 0.55923(18) 0.43244(17) 0.39444(18) 0.0151(4)

0.55941(17) 0.43275(16) 0.39438(17) 0.0172(4) O3 0.36341(17) 0.42816(16) 0.38470(17) 0.0122(4)

0.36360(17) 0.42837(15) 0.38526(16) 0.0145(4)

O4 0.06828(17) 0.12686(19) 0.39442(19) 0.0151(4)

0.06814(16) 0.12662(18) 0.39418(18) 0.0165(4) H1 0.579(5) 0.662(4) 0.428(4) 0.027(9)

0.585(4) 0.678(4) 0.432(3) 0.031(8)

H2 0.583(4) 0.400(4) 0.421(4) 0.027(9)

0.586(4) 0.402(4) 0.416(4) 0.031(8) H3 0.316(4) 0.401(5) 0.409(4) 0.027(9)

0.310(4) 0.398(4) 0.409(3) 0.031(8)

H4 0.097(4) 0.182(5) 0.417(4) 0.027(9)

0.096(4) 0.183(4) 0.412(4) 0.031(8)

Initial 300 K (R𝟑𝟑�)

353 K (R𝟑𝟑�)

x/a y/b z/c Ueq

x/a y/b z/c Ueq

M1 0 0 0.5 0.00925(12)

0 0 0.5 0.01156(12) M2 0.5 0.5 0.5 0.00915(9)

0.5 0.5 0.5 0.01136(9)

M3 0.41479(2) 0.32965(2) 0.33163(2) 0.01104(8)

0.41542(2) 0.33083(2) 0.33215(2) 0.01340(8) M4 0.41191(2) 0.57858(2) 0.33339(2) 0.01054(8)

0.41333(2) 0.58003(2) 0.33331(2) 0.01313(8)

Cl1 0 0 0.19388(7) 0.01415(17)

0 0 0.19401(6) 0.01690(16) Cl2 0.50169(4) 0.50172(4) 0.19380(4) 0.01385(11)

0.50115(4) 0.50113(4) 0.19402(4) 0.01688(11)

O1 0.55691(13) 0.62045(13) 0.40031(11) 0.0091(3)

0.55855(12) 0.62191(12) 0.39903(11) 0.0145(3) O2 0.56019(15) 0.43349(14) 0.39478(15) 0.0181(4)

0.56109(13) 0.43444(13) 0.39475(13) 0.0201(3)

O3 0.36533(15) 0.42946(14) 0.38686(14) 0.0174(4)

0.36767(14) 0.43139(13) 0.38950(12) 0.0211(3) O4 0.06743(15) 0.12656(16) 0.39447(15) 0.0182(4)

0.06613(13) 0.12654(14) 0.39475(13) 0.0199(3)

H1 0.588(3) 0.679(3) 0.426(3) 0.031(7)

0.591(3) 0.683(3) 0.427(3) 0.039(7) H2 0.591(3) 0.399(3) 0.420(3) 0.031(7)

0.593(3) 0.403(3) 0.419(3) 0.039(7)

H3 0.308(3) 0.399(3) 0.410(3) 0.031(7)

0.308(3) 0.402(3) 0.414(2) 0.039(7) H4 0.098(3) 0.185(4) 0.414(3) 0.031(7)

0.098(3) 0.188(3) 0.420(3) 0.039(7)

393 K (R𝟑𝟑�m)

443 K (R𝟑𝟑�m)

x/a y/b z/c Ueq

x/a y/b z/c Ueq

M1 0 0 0.5 0.01171(11)

0 0 0.5 0.01321(12) M2 0.5 0 0 0.01383(9)

0.5 0 0 0.01528(10)

Cl1 0 0 0.19405(5) 0.01763(14)

0 0 0.19427(6) 0.01992(16) O1 0.20681(13) 0.79319(13) 0.06158(10) 0.0202(3)

0.20681(14) 0.4136(3) 0.06157(11) 0.0215(3)

H1 0.137(3) 0.275(7) 0.085(2) 0.044(10)

0.141(3) 0.283(7) 0.085(2) 0.035(10)

Final 300 K (R𝟑𝟑�)

x/a y/b z/c Ueq

M1 0 0 0.5 0.00956(13) M2 0.5 0.5 0.5 0.00938(9) M3 0.41484(2) 0.32970(2) 0.33166(2) 0.01134(8) M4 0.41197(2) 0.57866(2) 0.33330(2) 0.01081(8) Cl1 0 0 0.19382(7) 0.01459(18) Cl2 0.50172(5) 0.50171(5) 0.19383(4) 0.01409(11) O1 0.55718(14) 0.62082(13) 0.40016(12) 0.0097(3) O2 0.56011(16) 0.43349(15) 0.39451(16) 0.0185(4) O3 0.36528(16) 0.42955(15) 0.38708(15) 0.0184(4) O4 0.06734(15) 0.12657(17) 0.39450(16) 0.0179(4) H1 0.584(4) 0.678(4) 0.432(3) 0.037(7) H2 0.588(4) 0.403(4) 0.421(3) 0.037(7) H3 0.308(4) 0.401(4) 0.410(3) 0.037(7) H4 0.102(4) 0.191(4) 0.418(3) 0.037(7)

CHAPTER 2.1

20

Table 2.1.4. Anisotropic displacement parameters (Å2

) at each experimental temperature 100K (R𝟑𝟑�)

U11 U22 U33 U23 U13 U12 M1 0.00394(19) 0.00394(19) 0.0080(3) 0 0 0.00197(9) M2 0.00434(18) 0.00356(18) 0.0072(2) 0.00047(13) 0.00043(13) 0.00189(14) M3 0.00676(15) 0.00250(14) 0.00931(17) -0.00133(10) -0.00052(11) 0.00143(11) M4 0.00309(14) 0.00266(14) 0.01012(17) 0.00007(10) 0.00029(10) -0.0001(1) Cl1 0.0069(2) 0.0069(2) 0.0096(5) 0 0 0.00344(12) Cl2 0.0057(2) 0.0063(2) 0.0096(3) -0.00014(18) 0.00001(18) 0.00268(19) O1 0.0032(7) 0.0003(7) 0.0047(8) 0.0030(5) 0.0026(6) -0.0021(6) O2 0.0128(9) 0.0055(8) 0.0246(12) 0.0015(7) -0.0102(8) 0.0028(7) O3 0.0067(8) 0.0001(7) 0.0227(11) -0.0049(7) 0.0079(7) -0.0036(6) O4 0.0051(8) 0.0140(10) 0.0259(13) -0.0133(9) -0.0025(8) 0.0044(7)

200 K (R𝟑𝟑�)

U11 U22 U33 U23 U13 U12 M1 0.00689(19) 0.00689(19) 0.0091(3) 0 0 0.00344(9) M2 0.00752(18) 0.00615(17) 0.0088(2) 0.00086(13) 0.00073(13) 0.00340(14) M3 0.00932(15) 0.00500(14) 0.01214(17) -0.00197(10) -0.00072(10) 0.00275(11) M4 0.00558(13) 0.00511(14) 0.01329(17) 0.00034(10) 0.00000(10) 0.00151(10) Cl1 0.0107(3) 0.0107(3) 0.0119(5) 0 0 0.00536(13) Cl2 0.0096(2) 0.0109(2) 0.0126(3) -0.00014(18) 0.00006(18) 0.00484(19) O1 0.0058(6) 0.0017(6) 0.0068(7) 0.0032(5) 0.0024(6) -0.0007(5) O2 0.0137(8) 0.0076(8) 0.0279(12) 0.0016(7) -0.0098(8) 0.0035(7) O3 0.0099(8) 0.0015(7) 0.0254(11) -0.0041(7) 0.0085(7) -0.0022(6) O4 0.0081(8) 0.0135(9) 0.0273(12) -0.0118(8) -0.0030(7) 0.0050(7)

Initial 300 K (R𝟑𝟑�)

U11 U22 U33 U23 U13 U12 M1 0.00966(17) 0.00966(17) 0.0084(3) 0 0 0.00483(8) M2 0.01034(16) 0.00876(16) 0.00834(17) 0.00082(12) 0.00083(12) 0.00476(13) M3 0.01130(13) 0.00738(12) 0.01351(15) -0.00222(9) -0.00072(9) 0.00398(10) M4 0.00873(12) 0.00815(12) 0.01415(15) 0.00087(9) -0.00009(9) 0.00377(10) Cl1 0.0146(2) 0.0146(2) 0.0132(4) 0 0 0.00730(12) Cl2 0.0135(2) 0.0151(2) 0.0128(2) -0.00010(17) -0.00005(17) 0.00695(19) O1 0.0086(6) 0.0051(6) 0.0102(7) 0.0011(5) 0.0007(5) 0.0009(5) O2 0.0156(8) 0.0108(7) 0.0263(10) 0.0021(6) -0.0089(7) 0.0054(6) O3 0.0137(7) 0.0057(6) 0.0278(10) -0.0009(6) 0.0111(7) 0.0012(6) O4 0.0112(7) 0.0161(8) 0.0266(10) -0.0112(7) -0.0028(7) 0.0064(6)

353 K (R𝟑𝟑�)

U11 U22 U33 U23 U13 U12 M1 0.01281(16) 0.01281(16) 0.0091(3) 0 0 0.00641(8) M2 0.01321(15) 0.01205(15) 0.00883(16) 0.00051(10) 0.00056(10) 0.00633(12) M3 0.01356(12) 0.01054(12) 0.01542(14) -0.00200(8) -0.00066(8) 0.00552(9) M4 0.01229(12) 0.01185(12) 0.01564(14) 0.00109(8) -0.00048(8) 0.00632(9) Cl1 0.0182(2) 0.0182(2) 0.0143(4) 0 0 0.00911(11) Cl2 0.0175(2) 0.0186(2) 0.0144(2) -0.00037(15) -0.00026(15) 0.00888(18) O1 0.0138(6) 0.0103(6) 0.0175(7) -0.0029(5) -0.0014(5) 0.0045(5) O2 0.0173(7) 0.0144(7) 0.0277(9) 0.0042(6) -0.0075(6) 0.0073(6) O3 0.0186(7) 0.0117(6) 0.0310(9) 0.0032(6) 0.0140(6) 0.0061(6) O4 0.0145(7) 0.0177(7) 0.0280(9) -0.0111(6) -0.0042(6) 0.0084(6)

CHAPTER 2.1

21

Table 2.1.4 Continued

393 K (R𝟑𝟑�m)

U11 U22 U33 U23 U13 U12 M1 0.01347(15) 0.01347(15) 0.0082(2) 0 0 0.00674(7) M2 0.01373(12) 0.01172(14) 0.01536(14) 0.00196(10) 0.00098(5) 0.00586(7) Cl1 0.0195(2) 0.0195(2) 0.0138(3) 0 0 0.00976(10) O1 0.0167(4) 0.0167(4) 0.0262(6) -0.0054(3) 0.0054(3) 0.0078(5)

423 K (R𝟑𝟑�m)

U11 U22 U33 U23 U13 U12 M1 0.01523(16) 0.01523(16) 0.0092(2) 0 0 0.00762(8) M2 0.01515(13) 0.01285(15) 0.01707(16) 0.00221(10) 0.00111(5) 0.00642(8) Cl1 0.0220(2) 0.0220(2) 0.0157(3) 0 0 0.01101(11) O1 0.0180(4) 0.0191(7) 0.0279(6) 0.0106(6) 0.0053(3) 0.0096(3)

Final 300 K (R𝟑𝟑�)

U11 U22 U33 U23 U13 U12 M1 0.01017(18) 0.01017(18) 0.0083(3) 0 0 0.00508(9) M2 0.01064(17) 0.00897(17) 0.00840(18) 0.00077(13) 0.00078(13) 0.00480(14) M3 0.01164(13) 0.00791(13) 0.01361(15) -0.00210(10) -0.00072(10) 0.00423(10) M4 0.00912(13) 0.00850(13) 0.01432(15) 0.00085(10) -0.00021(10) 0.00403(10) Cl1 0.0153(3) 0.0153(3) 0.0133(4) 0 0 0.00763(13) Cl2 0.0138(2) 0.0153(2) 0.0129(2) -0.00022(18) -0.00045(18) 0.0071(2) O1 0.0093(7) 0.0046(6) 0.0117(7) 0.0010(5) 0.0014(6) 0.0008(6) O2 0.0161(8) 0.0114(8) 0.0268(11) 0.0026(7) -0.0088(7) 0.0059(7) O3 0.0141(8) 0.0064(7) 0.0301(11) -0.0014(7) 0.0115(8) 0.0016(6) O4 0.0118(8) 0.0167(9) 0.0253(11) -0.0116(8) -0.0033(7) 0.0071(7) The anisotropic displacement factor exponent takes the form: -2π2[h2a*2U11

+...+ 2hka*b*U12

]

Table 2.1.5. Selected bond lengths (Å) and angles (°) for paratacamite at each experimental temperature Temperature 100 K 200 K Initial 300 K 353K 393K 423 K Final 300 K Sapce group R3� R3� R3� R3� R3�m R3�m R3� Interlayer

M1(1)–O4(1)* x6 2.106(3) 2.107(2) 2.106(2) 2.104(2) 2.106(2) 2.106(2) 2.106(2) O4(1)–M1(1)–O4(1) 103.91(9) 104.10(9) 104.02(8) 103.91(7) 103.88(6) 103.77(7) 103.99(8) M2–O1 x2 1.980(2) 1.983(2) 1.996(2) 2.022(2)

2.001(2)

M2–O2 x2 2.104(3) 2.105(2) 2.103(2) 2.105(2)

2.106(2) M2–O3 x2 2.282(2) 2.276(2) 2.249(2) 2.204(2)

2.247(2)

O1–M2–O2 101.88(8) 101.99(8) 102.24(7) 102.70(6)

102.27(7) O1–M2–O3 104.36(8) 104.20(7) 104.15(7) 103.95(6)

104.05(7)

O2–M2–O3 106.03(8) 105.91(7) 105.59(7) 105.08(6)

105.58(7) Intralayer

M3(2)–O3(1) (x6) 1.955(2) 1.957(2) 1.957(2) 1.963(2) 1.9840(7) 1.9860(9) 1.960(2) M3–O2 1.973(2) 1.974(2) 1.979(2) 1.979(2)

1.977(2)

M3–O2 1.976(2) 1.976(2) 1.979(2) 1.981(2)

1.979(2) M3–O1 1.993(2) 1.994(2) 1.996(2) 1.995(2)

1.997(2)

M3(2)–Cl2(1) (x2) 2.7347(7) 2.7412(7) 2.7499(6) 2.7585(6) 2.7820(5) 2.7817(6) 2.7502(7) M3–Cl2 2.8177(7) 2.8124(7) 2.8065(6) 2.7962(6)

2.8066(7)

O3(1)–M3(2)–O2(1) 177.10(10) 177.29(9) 177.72(8) 178.44(7) 180.00(8) 180.00(10) 177.78(9) O2–M3–O1 178.72(9) 178.76(8) 178.80(7) 178.95(7)

178.89(8)

O2–M3–O2 97.93(15) 97.99(14) 98.10(12) 98.14(10)

98.08(13) O2(1)–M3(2)–O1(1) 80.92(9) 80.86(9) 80.73(8) 80.85(7) 81.73(9) 81.79(12) 80.87(8) O3(1)–M3(2)–O1(1) 96.57(8) 96.64(8) 97.17(7) 97.66(7) 98.27(9) 98.21(12) 97.03(8) Cl2(1)–M3(2)–Cl2(1) 176.78(2) 176.92(2) 177.45(2) 178.26(2) 180.00(2) 180.00(3) 177.48(2) Cl2(1)–M3(2)–O2(1) 96.16(7) 96.17(7) 96.29(6) 96.70(5) 97.50(3) 97.41(4) 96.40(6) Cl2(1)–M3(2)–O1(1) 84.28(6) 84.18(5) 83.86(5) 83.42(4) 82.50(3) 82.59(4) 83.85(5) Cl2–M3–O3 97.70(7) 97.51(6) 97.43(6) 97.36(5)

97.39(6)

CH

APTER

2.1

22

Temperature 100 K 200 K Initial 300 K 353K 393K 423 K Final 300 K Space group R3� R3� R3� R3� R3�m R3�m R3� M4–O3 1.940(2) 1.943(2) 1.953(2) 1.966(1)

1.955(2)

M4–O4 1.970(2) 1.967(2) 1.971(2) 1.976(2)

1.972(2) M4–O4 1.981(2) 1.981(2) 1.984(2) 1.984(2)

1.984(2)

M4–O1 1.998(2) 2.001(2) 1.998(2) 1.994 (2)

2.000(2) M4–Cl1 2.7735(9) 2.7760(8) 2.7767(8) 2.7774(7)

2.7786(8)

M4–Cl2 2.7780(7) 2.7777(7) 2.7790(6) 2.7779(6)

2.7790(7)

O3–M4–O4 175.86(10) 176.23(9) 176.67(8) 177.72(7)

176.83(9) O3–M4–O4 97.05(10) 97.15(10) 97.47(9) 97.74(8)

97.43(9)

O3–M4–O1 83.61(9) 83.49(8) 82.94(8) 82.35(7)

83.03(8) O4–M4–O1 176.84(8) 176.71(8) 177.24(7) 178.01(6)

177.24(8)

O4–M4–O1 97.45(9) 97.51(9) 97.85(8) 98.16(7)

97.75(8) Cl1–M4–Cl2 176.13(2) 176.26(2) 176.84(2) 177.81(1)

176.88(2)

Cl1–M4–O1 94.52(6) 94.53(5) 94.96(5) 95.64(4)

94.95(5) Cl1–M4–O3 101.42(6) 101.16(6) 100.67(5) 99.66(5)

100.52(6)

Cl1–M4–O4 82.51(7) 82.41(7) 82.50(6) 82.51(5)

82.49(6) Cl1–M4–O4 82.32(7) 82.18(6) 82.28(6) 82.38(5)

82.29(6)

Other

O2(1)–Cl1(1) 3.065(2) 3.068(2) 3.069(2) 3.073(2) 3.079(1) 3.077(2) 3.072(2) O1–Cl2 3.057(2) 3.058(2) 3.066(2) 3.071(2)

3.064(2)

O3–Cl2 3.080(2) 3.077(2) 3.080(2) 3.077(2)

3.078(2) O4–Cl2 3.064(2) 3.071(2) 3.073(2) 3.075(2)

3.074(2)

*Atom labels without parentheses correspond to R3� structures while atoms labels in parentheses correspond to R3�m structures.

CH

APTER

2.1

23

CHAPTER 2.1

24

2.1.4 RESULTS

Figure 2.1.4 displays the variation in unit cell parameters with temperature. Both a

and c parameters increase non-linearly up to 393 K. At 443 K the a parameter is unchanged

from 393 K and c shows an additional increase with higher temperatures.

The M(1) interlayer site lies at 3�m and is coordinated octahedrally by O(4) atoms. The

M(2) interlayer site is coordinated by O(1), O(2) and O(3) atoms in a tetragonally elongated

and rhombically distorted octahedral environment, which may be described as a (2+2+2)

Jahn-Teller distortion and is discussed further below. The intralayer M(3) and M(4) sites are

both coordinated to four short equatorial O atoms and two long axial Cl atoms, which is a

typical (4+2) Jahn-Teller distorted configuration for octahedrally coordinated Cu2+. The

polyhedral environments of M(1) and M(2) octahedra in R3� paratacamite at 300 K are shown

in Figure 2.1.5. The M(1) octahedron shares edges with triplets of M(4) octahedra above and

below it. The M(2) octahedron shares edges with two triplets above and below it, each

comprising one M(4) and two M(3) octahedra. Mirror planes of the R3�m superstructure are

broken primarily by the (2+2+2) distortion of the M(2) site. Small rotations (~ 2°) of the M(1)

and M(2) out of these mirror planes is also apparent.

The variation of M–O and M–Cl bond-lengths with temperature is shown in Figure 2.1.6.

Several significant observations can be made from this Figure. First, the M(1)O6 octahedron

is essentially unaffected by heating. The M(2)O6 octahedron shows a marked change with

heating that involves convergence of M(2)–O(1) and M(2)–O(3) bond-lengths with those of

M(2)–O(2) and M(1)–O(4). The M(3)O4Cl2 octahedron has a differential response between

100 and 353 K in which the two M(3)–O(2) bonds show small increases in length while the

M(3)–O(1) and M(3)–O(3) bonds, which are quite different, rapidly converge at the

transformation. The M(3)–Cl(2) bonds converge steadily towards the transformation. Finally,

the M(4)O4Cl2 octahedron, which initially had very different M(4)–O(1) and M(4)–O(3)

bond-lengths, converged smoothly towards the transformation. The two M(4)–O(4) bonds

lengths do not change significantly with heating or cooling. One of them is almost unchanged

at ~ 1.980 Å while the other bond shows a minor contraction at 200 K, followed by a steady

lengthening until convergence is achieved at the transformation. The two M(4)–Cl bonds

have very similar lengths and remain almost constant throughout heating.

CHAPTER 2.1

25

Figure 2.1.4. Variation in unit cell parameters with temperature. The open circle is the cell parameter upon return to 300 K.

Figure 2.1.5. Polyhedral environments of M(1) and M(2) interlayer octahedra of R 3� paratacamite. Cl atoms shown as green spheres. O atoms are omitted for clarity. M(1)-M(4), M(2)-M(3) and M(2)-M(4) shared edges are O–O.

CHAPTER 2.1

26

Figure 2.1.6. Variations of M–O and M–Cl bond lengths of paratacamite on heating from 100 to 443 K. Error bars on bond lengths are shown when these are larger than data symbols. Lines connecting data points are intended as guides to the eye only, and it should be noted that the convergence shown at 393 K relates to a transition between 353 and 393 K. Variation of the M(1)–O(4) bond-length is shown as a dashed line.

Figure 2.1.7 shows the variation with temperature of the volumes of M(1), M(2), M(3)

and M(4) octahedra calculated using the program Xtaldraw (Downs and Hall-Wallace, 2003).

The interlayer M(1)O6 and M(2)O6 octahedra contract on heating while the intralayer

M(3)O4Cl2 and M(4)O4Cl2 octahedra expand.

Figure 2.1.8 shows the bond-angle variances (BAV) of all four octahedra and the

quadratic elongation (QE) of the M(2) octahedron as a function of temperature. These

geometrical parameters were calculated using the formulation of Robinson et al. (1971) as

implemented in the program Xtaldraw. The quadratic elongation of M(1), M(3) and M(4)

octahedra is invariant with temperature, displaying values of 1.054, 1.074 and 1.075,

respectively. In contrast, the behaviour of the quadratic elongation of the M(2) octahedron

shows a smooth, non-linear decrease to the phase transition between 353 and 393 K.

CHAPTER 2.1

27

Figure 2.1.7. Volume changes of M(1), M(2), M(3) and M(4) octahedra with temperature.

Figure 2.1.8. Bond-angle variance (BAV) of M(1), M(2), M(3) and M(4) octahedra and

quadratic elongation (QE) of M(2) with temperature.

CHAPTER 2.1

28

The intralayer M(3) and M(4) octahedra show typical (4+2) Jahn-Teller distortion for

Cu2+

2.1.5 DISCUSSION

, their response to heating is a shortening of bonds that does not significantly change

their degree of angular deviation. Because the M(1) octahedron has equal M(1)–O(4) bond-

lengths, the distortion registered by QE is due entirely to deformation of the O–M–O bond-

angle, which at 76° differs considerably from 90°. The higher QE values of the M(2)

octahedron in the R3� structure are due to the three very different bond lengths of the (2+2+2)

configuration and deformation of the O–M–O bond angles. On heating, the QE value of the

M(2) octahedron converges smoothly to 1.054 at the transformation.

The structures of paratacamite from 100 to 353 K reported here are analogous to that

of the R3� superstructure first determined by Fleet (1975), although with all four H atoms also

located. At 393 and 443 K, the structure is that of the average substructure, again similar to

Fleet (1975). However, refinements of this substructure in R3� and R3�m gave comparable

results (Table 2.1.2) and there is no clear evidence of a split O position reported by Fleet

(1975). The disappearance of superlattice reflections between 353 and 393 K, together with

clear discontinuities in cell parameters and their smooth non-linear variations from 100 and

353 K are consistent with a structural phase transformation. The re-appearance of superlattice

reflections on cooling from 443 to 300 K demonstrates that the transformation is reversible.

The lattice parameters and structure for the high temperature phase correspond very well to

those of the related mineral herbertsmithite, a = 6.834(1) and c = 14.075(2) Å in space group

R3�m (Braithwaite et al., 2004) The structures determined at 300 and 393 K are depicted in

Figure 2.1.9. The transformation of paratacamite to herbertsmithite between 353–393 K

involves a four-fold reduction in unit cell volume and corresponding change in point group

from C3i (3� ) to D3d (3�m). To achieve the transformation, the most pronounced atomic

displacements occur with the O positions (Table 2.1.3). From the geometrical changes of M

octahedra as a function of temperature (Figures 2.1.5 to 2.1.8), it is clear that the M(1)

octahedron is a rigid, almost invariant, feature of the R3� paratacamite structure. The M(2)

octahedron changes considerably on heating as it becomes progressively less distorted

leading up to convergence upon the M(1) configuration at the transition. In contrast to the

different behaviour of M(1) and M(2) interlayer octahedra, the M(3) and M(4) intralayer

octahedra experience comparable expansions. Structurally, it appears that the variable nature

of the M(2) octahedron is a pivotal feature of paratacamite.

CHAPTER 2.1

29

Figure 2.1.9. Polyhedral structure representation for the initial 300 K R3� (left) and 393 K R3�m (right) structures. M(3) and M(4) octahedra are blue, M(1) octahedra are orange, M(2) octahedral are yellow, Cl atoms are displayed as green spheres, O atoms are pictured as red spheres, H atoms appear as white spheres. The unit cell is outlined. Both structures are viewed down the a axis. Anisotropic thermal ellipsoids calculated with probability at 85%.

2.1.5.1 The (2+2+2) Jahn-Teller distortion

The M(2)O6 coordination environment of paratacamite displays a tetragonally

elongated and rhombically distorted octahedron, which may be described as the rare (2+2+2)

Jahn-Teller distortion (Burns and Hawthorne, 1996). The superstructure for type paratacamite

cannot be refined in space group R3�m, but refines very well in R3�. In an R3�m superstructure,

determined in another study on Mg-rich paratacamite (Chapter 2.2), two of the four O atoms

in 18f, O(1) and O(3), merge into one lower symmetry position at 36i. This changes the M(2)

coordination to a slightly tetragonally compressed octahedron with four equivalent equatorial

O atoms, effectively removing the rhombic distortion. However, this structure does not

account for all of the observable superstructure reflections and the model did not support

other paratacamite data sets, such as those in this study. Two possibilities are considered to

explain the origin of the (2+2+2) distortion of the M(2) coordination sphere.

CHAPTER 2.1

30

First, the composition of the paratacamite crystal studied is very close to

Cu3.75Zn0.25(OH)6Cl2. Unfortunately, the X-ray scattering factors of Cu (Z = 29) and Zn

(Z = 30) are too similar to be able to distinguish between them for site occupancy refinement.

In all paratacamite-related phases with R3�m subcells, such as gillardite and herbertsmithite,

the minor substituent cation occurs within the interlayer site. On this basis, it can be argued

that the substructure recorded by XRD implies that Cu and Zn occupy the same site, M(1)

and are long-range disordered. A corollary of this reasoning is that Cu and Zn may be ordered

in the superstructure, where Zn preferentially occupies M(1) and Cu occupies M(2), M(3) and

M(4) because thermally-induced disordering of Zn over M(1) and M(2) sites may be unlikely

at these low temperatures. Based on the composition determined here, Zn would be confined

to M(1) which induces the observed uniform coordination geometry of this octahedron.

Therefore, the (2+2+2) coordination sphere of the M(2) octahedron, which would be fully

occupied by Cu2+, could be generated by long-range crystal structure constraints which

distort the common (4+2) Jahn-Teller configuration. A (2+2+2) Jahn-Teller distortion (sensu

stricto) from Cu2+ octahedra with OH-

The various possible Jahn-Teller configurations were investigated by Burns and

Hawthorne (1996) for a host of Cu(II) bearing compounds. It was concluded by these authors

that the occurrence of a (2+2+2) distortion may be explained by a dynamic Jahn-Teller effect

as the octahedron continually shifts between two configurations of the (4+2) Jahn-Teller

geometry. As a consequence, the detection of a (2+2+2) octahedral coordination sphere using

X-ray diffraction is a result of a time lapse average position of the atoms involved. Figure

2.1.10 displays the M(2)O

ligands is unprecedented.

6 coordination environment in the structure of paratacamite before

and after the structural transformation. The analogous interlayer position, M(1), in the 393 K

structure is also displayed and all atoms are pictured with anisotropic displacement ellipsoids.

The maximum principal axis of the trans O(2) and O(3) ellipsoids are subparallel with the

direction of the M(2)–O bond axis in paratacamite. In contrast, the O(1) displacement

ellipsoid appears as a slightly compressed sphere with its maximum principal axis almost

perpendicular to the M(2)–O(1) bond axis. Burns and Hawthorne (1996) suggested that a

more or less spherical displacement ellipsoid indicates a static bond. However, the

M(2)–O(1) bond length, which is the shortest of the coordination sphere, increases

significantly with an increase in temperature. Heating to 353 K yielded similar anisotropic

displacement values and orientation for all O atoms. After the transformation to the R3�m

CHAPTER 2.1

31

structure, the maximum principal axis of the single crystallographic O displacement ellipsoid

is again subparallel to the M(1)–O bond direction.

Figure 2.1.10. The M(2) octahedral coordination environment of paratacamite during the heating and cooling cycle. Heavy atoms are displayed with anisotropic displacement ellipsoids (probability 85%). Arrows point in the direction of atomic displacement (trans atoms) with respect to the previous temperature.

Considering a statistical distribution of Zn between both interlayer sites in

paratacamite, a superimposition of 25% occupancy of Zn in M(2) at 9d (Zn–O ca 2.1 Å as in

herbertsmithite; Braithwaite et al., 2004) with 75% Cu (Jahn-Teller distorted as in

clinoatacamite with Cu–O ca 2.29, 2.05, and 1.99 Å; Grice et al., 1996) gives the average

detectible bond lengths of 2.24, 2.06 and 2.02 Å, which are similar to those of the initial and

final 300 K structures in Table 2.1.5. The relatively spherical anisotropic O(1) ellipsoid may

indicate that two different orientations of Jahn-Teller distorted Cu(OH)6 octahedra occur in

this site. The O(1) atoms would become the pivotal short Jahn-Teller bond of both

orientations while the other two directions alternate between the long and short bond

distance. The elongated anisotropic thermal ellipsoids of the O atoms involved would

therefore be the result of a distribution of two orientations of either a dynamic or static (4+2)

Jahn-Teller distortion of Cu(OH)6

Similarly, the O(4) atom, which makes up the non-tetragonally distorted M(1)

octahedra in the R3� structure, and the O(1) atom of the R3�m structure, show a degree of

anisotropy with the maximum principal axis of the ellipsoid directed subparallel with the

bond. By the same reasoning, with 25% occupancy of Zn in M(1) at 3b, accompanied by

three different orientations of the common (4+2) Jahn-Teller distortion (25% occupancy of

octahedra. The anisotropic O(2) and O(3) thermal

ellipsoids and variable nature of the M(2)–O(2) and –O(3) bond lengths with changes in

temperature suggest that this octahedron is dynamically distorted.

Initial 300 K

Final 300 K

353 K 393 K

O3

O1

O2

O1 O3

O1

O2

O3

O1

O2

R3� R3�

R3�

R3�m

CHAPTER 2.1

32

the three orientations) the average detectible bond length for M(1) would be ca 2.1 Å, using

the same values as before. This compares well with the 300 K data in Table 2.1.5. The rigid

nature of this octahedron with changes in temperature might indicate that the proposed three

orientations of the Jahn-Teller effect are static, rather than dynamic. If the M(1) octahedron is

dynamically distorted then the three orientations of the Jahn-Teller effect would be

energetically equivalent.

2.1.5.2 Origin of the phase transition in type paratacamite

The change in cell parameters and polyhedral behaviour with heating are consistent

with a steady convergence upon a new structure above 353 K. The loss of the paratacamite

superstructure above this temperature may indicate that the superstructure reflections are a

direct result of atomic displacements from the R 3� m aristotype structure, particularly

concerning O atoms that constitute the M(2)O6

The reversible nature of this phase transformation establishes that paratacamite is a

thermodynamically stable phase below ca 353 K, for the composition Cu

octahedron. The fact that the superstructure in

R 3�m would not refine, but succeeded in R 3� indicates that the superlattice reflections,

although much weaker than the sub-lattice reflections, contain enough information associated

with the (2+2+2) distorted M(2) site to enable the correct structure to be identified.

The driving force behind the phase transition is a temperature-induced reduction of

the M(2) octahedral rhombic distortion as it converges upon the configuration of the

temperature invariant M(1) coordination sphere. The responses of intralayer M(3) and M(4)

octahedra are primarily determined by deformation of the M(2) octahedron, with which they

share edges. Assuming a statistical distribution of Zn between these sites as well as the

proposed model of superimposed dynamically Jahn-Teller distorted octahedra with non-

tetragonally distorted octahedra, it may be inferred that there is an energy activation barrier

associated with the generation of three configurations of equally occupied (4+2) Jahn-Teller

distorted octahedra, such as the configuration observed at M(1).

3.71Zn0.29(OH)6Cl2.

Finally, the high temperature transformation is in agreement with the proposed space group

chain P1�→ R3� → R3�m. Although there was no evidence of a triclinic distortion at low

temperatures, the presence of significant amounts of non-tetragonally distorted Zn(OH)6

octahedra may be restricting a transformation to P1�. It must also be stated that the nature of

Cu-Zn ordering in type paratacamite remains to be determined.

CHAPTER 2.2

33

2.2 THE SINGLE-CRYSTAL STRUCTURE OF Mg- AND Ni-ANALOGUES OF PARATACAMITE

2.2.1 INTRODUCTION

Paratacamite reportedly occurs in up to 134 localities world-wide, according to

Mindat.org. Clinoatacamite coexists on the type specimen of paratacamite (Jambor et al.,

1996) and is the thermodynamically stable Cu2(OH)3Cl polymorph at 298 K [cited as

paratacamite in Pollard et al. (1989)]. The apparent compositional dependence on the

rhombohedral structure type, determined by Jambor et al. (1996) and confirmed by

Braithwaite et al. (2004), indicates that paratacamite is not, sensu stricto, a polymorph of

Cu2(OH)3Cl. The discovery that paratacamite is thermodynamically stable at ambient

temperatures for the composition Cu3.71Zn0.29(OH)6Cl2 (Chapter 2.1) suggests that other

analogues of paratacamite are likely to exist. The cations Zn2+, Co2+, Fe2+, Ni2+, Mg2+ and

Mn2+ are known to substitute for Cu2+ in the structure of synthetic clinoatacamite and

herbertsmithite (Oswald and Feitknecht, 1964). Although many of the descriptions of

paratacamite lack compositional data (Smith, 1906; Jambor et al., 1996; Pring et al., 1987),

the possibility exists that some of these analyses were made concerning samples of new

materials exhibiting contrasting compositions. In validation of this, the current investigation

led to the discovery of two new, naturally-occurring members of the paratacamite group

characterised by extensive Mg2+ and Ni2+ substitution in the interlayer sites, giving the

general formula Cu3(Mg,Cu)(OH)6Cl2 and Cu3(Ni,Cu)(OH)6Cl2

2.2.2 SAMPLES AND ANALYSIS

, respectively. They are

isomorphous with paratacamite. This section reports a crystallographic investigation of both

new paratacamite congeners and provides their physical and optical data.

A specimen exhibiting members of the basic Cu chlorides was obtained from a small

deposit located about 5 km NE of the village of Cuya in the Camarones Valley, Arica

Province, Chile (approximately 19°08'14"S 70°08'49"W). The material occurs in association

with anhydrite, atacamite, chalcopyrite, copiapite, dolomite, epsomite, haydeeite, hematite,

magnesite and quartz. The specimen has been retained in the collections of the Natural

History Museum of Los Angeles, Los Angeles, California, USA (specimen numbers 64041,

64042 and 64043).

Another specimen, originally from the Carr Boyd Rocks Mine, Western Australia,

Australia (30°04'S 121°37'E), and retained in the Gartrell Collection of the Western

CHAPTER 2.2

34

Australian Museum under specimen number WAM M365.2003 (formerly G3520), was

obtained on loan. Electron microprobe analyses were made using a JEOL 8600 electron

microprobe operated in WDS mode with an accelerating voltage of 15 kV, 20 nA and 3 μm

beam diameter.

Analytical data are given in Table 2.2.1. Twenty-three electron microprobe analyses

of the Camerones sample gave the empirical formula Cu3.49Mg0.62Mn0.01Ni0.01Cl1.75O6.25H6

and the average of eight spots of the Carr Boyd Rocks sample gave the empirical formula

Cu3.25Ni0.70Co0.02Cl2.06O5.94H6.

Table 2.2.1. Electron microprobe analyses of material in this study. Camerones - sample 64041 Range Average* Empirical Normalised (wt%) (wt%) composition composition** CuO 59.98–77.87 69.34(4.74) 3.49 3.38 NiO 0–0.37 0.10(0.09) 0.01 0.01 MnO 0.03–0.37 0.17(0.08) 0.01 0.01 MgO 3.58–10.54 6.29(2.25) 0.62 0.60 CoO 0–0.15 0.08(0.05) - - Cl 14.59–16.32 15.47(0.51) 1.75 2.00 H2O 13.52 6.00 6.00 O≡Cl -3.50 Total 101.47 Carr Boyd Rocks - sample WAM M365.2003 Range Average* Empirical Normalised (wt%) (wt%) composition composition** CuO 61.16–64.37 62.43(0.98) 3.25 3.27 NiO 11.21–13.95 12.61(1.11) 0.70 0.71 MgO 0–0.13 0.03(0.05) - - CoO 0.24–0.38 0.31(0.05) 0.02 0.02 Cl 17.25–17.78 17.60(0.20) 2.06 2.00 H2

Both samples were rather unstable in the probe, even with reduced beam current, thus

accounting for the higher analytical totals and lower Cl values obtained for the Camerones

sample; low Cl values are attributed to electron-beam-induced Cl

O 13.04 6.00 6.00 O≡Cl -3.98 Total 102.04 *Standard deviation of the average value are in parentheses. **Compositions were normalised to Σ(cations) = 4.00

- migration (Stormer et al.,

1993). The structure of both samples are composed of sheets of composition Cu3Cl2(OH)62–

linked by M2+ ions lying between them. Therefore, the empirical formula may be

written as (Cu3.49Mg0.62Mn0.01Ni0.01)Σ4.13Cl2(OH)6, for the Camerones sample, and

CHAPTER 2.2

35

(Cu3.25Ni0.70Co0.02)Σ3.97Cl2(OH)6, for the Carr Boyd Rocks sample. Normalisation to

Σ(cations) = 4 gives (Cu3.38Mg0.60Ni0.01Mn0.01)(OH)6Cl2 and (Cu3.27Ni0.71Co0.02)(OH)6Cl2,

respectively. Elements other than those displayed in Table 2.2.1 were below detection limits.

H2

2.2.2.1 Optical and physical properties

O was calculated by stoichiometry from the results of the crystal structure analysis,

because of the paucity of available material.

The Mg-rich Camerones sample examined forms steep rhombohedra with the forms

{201} and {001} as well as thick and thin prisms due to twinning. It is non-fluorescent. Its

colour is green and it has a light green streak and is transparent with a vitreous lustre. Its

Mohs hardness is ~3. It is brittle and has perfect cleavage on {201}. No parting was

observed. Calculated density is 3.532 g cm-3 based on the empirical formula and compares

well with measured density of 3.50(2) g cm-3 by flotation using Clerici solution. Optically, it

is uniaxial (–), ω and ε > 1.8 (crystals decompose rapidly in RI fluids > 1.8). Pleochroism:

O (bluish green) > E (green), slight. It is readily soluble in cold, dilute HCl.

The Ni-rich Carr Boyd Rocks sample forms aggregates of equant, rhombohedral

crystals showing the forms {101}, {021} and {001}. It is dark green with a green streak and a

vitreous lustre. It is non-fluorescent. Mohs hardness is 3. It is brittle and has a good cleavage

on {201}. It has no observable parting and an uneven fracture. Its calculated density from the

empirical formula is 3.735 g cm-3 and its measured density by flotation in Clerici solution is

3.70 g.cm-3. Optically, it is uniaxial (–), ω and ε > 1.8 (reacts rapidly with RI fluids > 1.8);

nav. = 1.81 using the Gladstone-Dale relationship for the empirical formula. There was no

dispersion and it is non-pleochroic. Chemical tests reveal that after dissolution in 6 M HNO3,

reaction with dimethylglyoxime and excess NH3 gives a heavy precipitate of Ni(DMGH-1)2.

Reaction of the acidic solution with potassium mercuric thiocyanate gives pale, yellow-green

rosettes of copper mercuric thiocyanate crystals. Addition of AgNO3 solution and raising the

pH with ammonia gives a white precipitate of AgCl.

2.2.3 CRYSTALLOGRAPHY

2.2.3.1 Sample 64041 (Mg-rich)

Data was collected at 293(2) K using a single-crystal on a Rigaku R-Axis Rapid II

curved imaging plate microdiffractometer with monochromatic MoKα radiation. A strong

substructure was identified with cell dimensions a = 6.8441(8) and c = 14.025(1) Å. A series

of weak reflections at half integer positions of h and k were identified to suggest that the true

CHAPTER 2.2

36

unit cell was a = 13.689(1) and c = 14.025(1) Å, in the hexagonal setting. Refinement of the

subcell converged smoothly to the structure of herbertsmithite with R1 = 0.0240 and

wR2 = 0.0592 for 239 reflections with Fobs > 4σ(Fobs

Initial refinement was made in space group R3�m. The structure was solved by direct

methods and resulted in atom coordinates and topology analogous to those of paratacamite

(Fleet, 1975). The R3� structure of paratacamite has four O atoms each in an 18f position. The

R3�m refinement merges O(1) and O(3) of the R3� structure into a single position located on a

mirror plane. This results in three crystallographically independent O atoms, two of which are

in 18f positions and the third in the lower symmetry 36i position. The rhombic distortion

exhibited by the M(2) octahedron of R 3� paratacamite is effectively removed by the

incorporation of a lower symmetry equatorial O position. However, the R3�m refinement did

not account for the full set of observable reflections. Therefore, the full data set was refined

in space group R3� based on the atom coordinates of paratacamite, established by Fleet (1975)

and confirmed in Chapter 2.1.

Initially, refinement was made with the M(1) site at 3b filled with Mg, and the M(2),

M(3) and M(4) sites occupied completely by Cu. However, when the heavy atoms were

refined anisotropically unacceptable values of displacement parameters were obtained. The

X-ray scattering factors of Mg (Z = 12) and Cu (Z = 30) are sufficiently different to allow

refinement of site occupancies between these two metals in M(1) and M(2). Initial refinement

of site occupancies resulted in an equivalent distribution of Mg (63% occupancy) at both

M(1) and M(2) and gave a significant improvement to the residuals. The ratio of Cu:Mg was

slightly higher than that determined from the analytical composition. The composition was

then set to that obtained by analysis while maintaining the statistical distribution of Mg; the

trace amounts of Ni and Mn present were arbitrarily placed at the M(1) site. Refinement

converged with equivalent isotropic displacement parameters for the M(1) and M(2) sites

remaining the same within error. The final distribution of Mg between M(1) and M(2) was

60% occupancy each. The coordinates of all four H atoms were determined from a difference

map and each was given a soft distance constraint of 0.80 ± 0.02 Å with its respective O

atom.

The final refinement converged to R

). Based on systematic absences for both

data sets, the space groups R3�m, R3m, R3�, R3 and R32 were possible.

1 = 0.0387 and wR2 = 0.1106, for 480 unique

reflections with Fobs > 4σ(Fobs). The weighting scheme used was w = 1/[σ2(Fo) 2 + (0.0423P)2

+ 14.34P] where P = [max(0, Fo)2 + (2Fc)2

]/3 as defined by SHELX–97 (Sheldrick, 2008).

CHAPTER 2.2

37

2.2.3.2 Sample WAM M365.2003 (Ni-rich)

A single-crystal from sample WAM M365.2003 was analysed at 296(2) K, using a

Bruker Smart 1000 CCD diffractometer with graphite-monochromatized MoKα radiation. A

similar approach was taken with refinement of the data set collected, as described above. A

rhombohedral supercell with parameters of a = 13.665(4), c = 13.915(4) Å, and a strong

subcell of dimensions a = 6.843(1), c = 13.935(3) Å was found. Data reduction to the subcell

and refinement based on the atom coordinates of gillardite resulted in R1 = 0.0168 and

wR2 = 0.0443 for 189 reflections with Fobs > 4σ(Fobs), but the true structure was reached by

indexing on the supercell. The proposed alternate R3�m model described above was refined

based on the data obtained from this crystal. This resulted in a similar structure exhibiting

three O atoms, one of which is in the 36i position. Again, this removed the rhombic distortion

of the M(2) coordination environment. However, anisotropic refinement of all heavy atoms

resulted in some becoming NPD and the H atoms could not be located. Additionally, the R3�m

refinement does not account for the full data set of observable reflections. Again, the correct

structure emerged from space group R3� for the full data set with atom coordinates analogous

to that of paratacamite. All H atoms were located in a difference map and fixed with soft

constraints to a uniform distance of 0.85 ± 0.02 Å from their respective O atom. All non-H

atoms were allowed to refine with anisotropic displacement parameters yielding acceptable

values. Refinement of M(1) and M(2) site occupancies was made based on a statistical

distribution of Cu and Ni, for the analytical composition (Cu3.27Ni0.71Co0.02)Cl2(OH)6. On

the basis of similar isotropic displacement parameters, both interlayer sites converged with

71% Ni occupancy and the small amount of Co distributed evenly between them. It should be

noted that the true distribution of Cu and Ni between M(1) and M(2) could not be determined

due to the similarity of X-ray scattering factors of Ni (Z = 28) and Cu (Z = 29). The

possibility remains that the distribution of Ni is preferentially at M(1), with excess occupying

M(2), as suggested elsewhere in the literature (Grice et al., 1996; Braithwaite et al., 2004).

The final structure converged with R1 = 0.0227 and wR2 = 0.0648 for 314 reflections

with Fobs > 4σ(Fobs). The final weighting scheme used was w = 1/[σ2(Fo) 2 + (0.0407P)2 +

2.33P] where P = [max(0, Fo)2 + (2Fc)2

]/3 as defined by SHELX–97 (Sheldrick, 2008).

Structure refinement details of both samples are given in Table 2.2.2. Final atom

coordinates and anisotropic displacement parameters are listed Tables 2.2.3 and 2.2.4 and

selected bond lengths and angles in Table 2.2.5.

CHAPTER 2.2

38

2.2.4. X-RAY POWDER DIFFRACTION

X-ray powder diffraction data for the Camerones sample were recorded using the

same diffractometer and radiation noted above for the single-crystal work. Observed d

spacing and intensities were derived by profile fitting using the JADE 2010 software package

(Materials Data, Inc, 2011). Unit cell parameters refined from the powder data, from whole

pattern fitting, are a = 13.667(2), c = 14.011(2) Å and V = 2266.5(4) Å3. X-ray powder

diffraction data for the Carr Boyd Rocks sample were recorded at room temperature using a

Bruker D8 Advance diffractometer (Ni-filtered CuKα1 radiation with pure Si as internal

standard; λ = 1.5406 Å). Unit cell parameters refined from the powder data are a = 13.667(2),

c = 13.908(5) Å and V = 2256.0(7) Å3

2.2.5 DISCUSSION

. A comparison of the X-ray powder data is in Table

2.2.6.

2.2.5.1 Interlayer cation distribution

The structures reported here are analogous to those of paratacamite, originally

reported by Fleet (1975) and confirmed in Chapter 2.1. The premise that paratacamite is

stable only with some essential non-Jahn-Teller distorting cation in the interlayer position

(Jambor et al., 1996), is supported by this analysis. Both the Camerones and Carr Boyd

Rocks sample display significant substitution for Cu. In both cases the amount of Cu that has

been replaced is in excess of ½ the interlayer total, which exceeds the limit proposed by

Braithwaite et al. (2004) on the stability of paratacamite. However, the compositional

stability field of paratacamite may be influenced by the type of substituting cation.

It has been suggested by Grice et al. (1996) and Braithwaite et al. (2004) that the

M(1) position preferentially occupies the substituting cation and that this type of ordering is

responsible for the appearance of the superstructure reflections. Refinement of Cu-Mg site

occupancies, in sample 64041, based on their X-ray scattering factors suggests a statistical

distribution with 60% Mg occupancy of both sites. The formula may be recast to represent

the relative occupancies as Cu3[(Mg0.15Cu0.08Ni0.01Mn0.01)(Mg0.45Cu0.30)]Σ1.00(OH)6Cl2

For the Ni-analogue, sample WAM M365.2003, even though the true distribution of

metals could not be determined from the X-ray scattering values of these cations, Ni is in

, with

the small amount of Ni and Mn placed in M(1). A truly statistical dispersion of the small

amount of Ni and Mn between M(1) and M(2) results in rounding errors. The actual

placement of these cations is not significant.

CHAPTER 2.2

39

such excess as to dominate both M(1) and M(2) sites whether a statistical or preferential

distribution is in fact the case. Nevertheless, the site occupancies were manually adjusted

based on a statistical distribution of Ni, reflecting equivalent proportions (71% occupancy) in

3b and 9d, and resulted in convergence of the isotropic thermal parameters, within error. The

formula may be recast as Cu3[(Ni0.178Cu0.067Co0.005)(Ni0.533Cu0.203Co0.015)]Σ1.00(OH)6Cl2

with the small amount of Co distributed in both interlayer sites.

Table 2.2.2. Crystal data and structure refinement details of samples in this study 64041 WAM M365.2003

Formula weight 1210.31 1270.76 Temperature (K) 293(2) 296(2) Wavelength (Å) 0.71073 0.71073 Crystal system Trigonal Trigonal Space group R3� R3� Unit cell dimensions a (Å) 13.689(1) 13.665(2)

c (Å) 14.025(1) 13.915(2) Volume (Å3 2275.8(3) ) 2250.2(11) Z 12 12 Calculated density (g cm-3) 3.532 3.751 Absorption coefficient (mm-1 10.144 ) 11.635 F(000) 2325 2439 Crystal size (mm) 0.11 x 0.06 x 0.01 0.21 x 0.15 x 0.10 Theta range for data 3.38 to 30.46° 2.26 to 28.17° Limiting indices -19 ≤ h ≤ 19 -17 ≤ h ≤ 17

-19 ≤ k ≤ 19 -18 ≤ k ≤ 17

-19 ≤ l ≤ 19 -18 ≤ l ≤ 17

Reflections /unique 9733/1547 5782/1168 R 0.0296 int 0.0236 Completeness to theta 30.46° 99.9% 28.23° 93.8% Refinement method Full-matrix Full-matrix

least-squares on F least-squares on F2

Data/restraints/parameters 2

1547/4/89 1168/4/87 Goodness-of-fit on F2 0.999 0.943 Final R [Iobs > 2σ(Iobs)] R 0.0387 1 0.0227

wR 0.0974 2 0.0648 R indices (all data) R 0.0864 1 0.0600

wR 0.1106 2 0.0882 Δρmax, Δρmin (e.Å-3 0.785 and -0.687 ) 0.619 and -0.924

CHAPTER 2.2

40

Table 2.2.3. Final atom coordinates and anisotropic displacement parameters* (Å2) of sample 64041 in space group R3�. x/a y/b z/c Ueq U11 U22 U33 U23 U13 U12 M1 0 0 0.5 0.0117(7) 0.0121(8) U11 0.0109(20) 0 0 0.5U11 M2 0.5 0.5 0.5 0.0120(3) 0.0129(8) 0.0125(8) 0.0098(7) 0.0004(7) 0.0003(7) 0.0057(7) M3 0.4162(1) 0.3322(1) 0.3329(1) 0.0115(2) 0.0109(3) 0.0099(3) 0.0135(4) -0.0022(3) 0.5U23 0.0050(3)

M4 0.4151(1) 0.5818(1) 0.3333(1) 0.0117(2) 0.0108(3) 0.0106(3) 0.0140(4) 0.0012(3) -0.0011(3) 0.0054(2) Cl1 0 0 0.1941(2) 0.0161(6) 0.0168(7) U11

0.0148(16) 0 0 0.5U11 Cl2 0.5006(1) 0.5004(1) 0.1944(1) 0.0147(3) 0.0158(6) 0.0162(7) 0.0122(6) 0.0005(6) 0.0001(6) 0.0080(6) O1 0.5613(3) 0.6246(3) 0.3962(4) 0.0137(10) 0.0097(19) 0.0105(20) 0.0185(28) -0.0041(18) -0.0017(18) 0.0032(16) O2 0.5625(3) 0.4353(3) 0.3956(4) 0.0133(10) 0.0121(19) 0.0117(20) 0.0165(31) 0.0019(17) -0.0017(17) 0.0064(16) O3 0.3700(3) 0.4342(3) 0.3940(3) 0.0155(10) 0.0117(20) 0.0110(19) 0.0219(29) 0.0006(2) 0.0046(20) 0.0042(17) O4 0.0645(3) 0.1269(3) 0.3953(4) 0.0138(11) 0.0111(19) 0.0126(20) 0.0173(32) -0.0059(19) -0.0007(17) 0.0057(16)

H1 0.5922(52) 0.6814(39) 0.4263(50) 0.0206 H2 0.5879(56) 0.4014(53) 0.4208(50) 0.0199 H3 0.3035(18) 0.4016(51) 0.3953(57) 0.0233 H4 0.0912(54) 0.1837(39) 0.4247(51) 0.0207

*The anisotropic displacement factor exponent takes the form: -2π2[h2a*2U11 +...+ 2hka*b*U12].

Ueq = 1/3(U11 + U22 + U33

Table 2.2.4. Final atom coordinates and anisotropic displacement parameters* (Å

).

2) of sample WAM M365.2003 in space group R3�. x/a y/b z/c Ueq U11 U22 U33 U23 U13 U12 M1 0 0 0.5 0.0106(5) 0.0121(8) U11

0.0076(12) 0 0 0.5U11 M2 0.5 0.5 0.5 0.0109(2) 0.0128(8) 0.0123(7) 0.0071(5) 0.0003(6) 0.0009(6) 0.0059(6) M3 0.4167(1) 0.3335(1) 0.3332(1) 0.0132(2) 0.0139(5) 0.0125(5) 0.0130(5) -0.0023(3) -0.0008(3) 0.0065(4) M4 0.4165(1) 0.5833(1) 0.3334(1) 0.0126(2) 0.0133(5) 0.0126(4) 0.0121(5) 0.0010(3) -0.0003(3) 0.0066(4) Cl1 0 0 0.1938(1) 0.0169(7) 0.0192(11) U11 0.0122(17) 0 0 0.5U11 Cl2 0.5002(2) 0.5000(2) 0.1935(1) 0.0162(3) 0.0193(10) 0.0186(10) 0.0115(6) 0.0009(8) -0.0003(9) 0.0100(9) O1 0.5631(5) 0.6261(5) 0.3953(4) 0.0138(12) 0.0154(32) 0.0094(27) 0.0193(31) 0.0011(22) -0.0012(22) 0.0083(25) O2 0.5639(5) 0.4380(5) 0.3956(4) 0.0168(12) 0.0142(30) 0.0234(34) 0.0144(28) 0.0006(23) -0.0053(23) 0.0105(27) O3 0.3742(5) 0.4376(5) 0.3945(4) 0.0191(13) 0.0143(31) 0.0212(34) 0.0255(31) 0.0014(25) 0.0074(25) 0.0118(29) O4 0.0638(4) 0.1263(5) 0.3958(4) 0.0158(13) 0.0089(27) 0.0193(34) 0.0160(31) -0.0050(22) -0.0015(21) 0.0046(26) H1 0.5900(69) 0.6947(27) 0.4117(56) 0.0207 H2 0.5962(67) 0.4056(66) 0.4236(59) 0.0253 H3 0.3073(32) 0.4093(74) 0.4158(61) 0.0286 H4 0.0922(72) 0.1876(48) 0.4289(61) 0.0237

*The anisotropic displacement factor exponent takes the form: -2π2[h2a*2U11 +...+ 2hka*b*U12].

Ueq = 1/3(U11 + U22 + U33).

CHAPTER 2.2

41

Table 2.2.5. Selected bond lengths (Å) and angles (°) of samples 64041 and WAM M365.2003.* Locality Camerones Carr Boyd Rocks Sample 64041 WAM M365.2003 Interlayer M1–O4 x6 2.103(5) 2.083(6) O4–M1–O4 180 180 O4–M1–O4 103.4(2) 103.1(2) M2–O1 x2 2.074(4) 2.081(6) M2–O2 x2 2.101(5) 2.086(6) M2–O3 x2 2.141(5) 2.091(6) O1–M2–O1 180 180 O2–M2–O2 180 180 O3–M2–O3 180 180 O1–M2–O2 103.0(2) 103.6(2) O1–M2–O3 103.4(2) 103.8(2) O2–M2–O3 103.5(2) 103.8(2) Intralayer M3–O1 1.987(4) 1.979(6) M3–O2 1.987(4) 1.970(6) M3–O2 1.992(4) 1.991(6) M3–O3 1.991(5) 1.978(6) M3–Cl2 2.765(2) 2.768(2) M3–Cl2 2.784(2) 2.772(3) O1–M3–O2 179.5(2) 179.9(3) O1–M3–O2 81.6(2) 81.4(2) O2–M3–O2 98.1(3) 98.7(4) O2–M3–O3 179.3(2) 179.5(3) O2–M3–O3 82.6(2) 81.4(2) Cl2–M3–Cl2 179.26(5) 179.9(1) Cl2–M3–O2 97.7(1) 97.5(2) Cl2– M3–O2 82.8(1) 82.4(2) M4–O1 1.989(4) 1.981(6) M4–O3 1.986(4) 1.968(6) M4–O4 1.985(4) 1.975(6) M4–O4 1.988(4) 1.989(6) M4–Cl1 2.778(2) 2.765(3) M4–Cl2 2.777(2) 2.775(2) O1–M4–O4 179.5(2) 179.8(3) O3–M4–O4 179.5(2) 179.4(3) O3–M4–O4 97.9(2) 98.5(2) O4–M4–O4 82.0(2) 81.6(4) O4–M4–O1 97.9(2) 98.4(2) Cl1–M4–Cl2 178.98(4) 179.91(9) Cl1–M4–O3 98.0(1) 98.1(2) Cl2–M4–O4 98.2(1) 97.4(2) *The M3 and M4 sites are Cu.

CHAPTER 2.2

42

Table 2.2.6. X-ray powder diffraction data (Å) for the Mg- and Ni-analogues of paratacamite. Mg-analogue Sample 64041

Ni-analogue Sample WAM M365.2003

Iobs dobs dcalc Icalc hkl

Iobs dobs dcalc Icalc 87 5.469 5.4597 100

0 2 1

81 5.445 5.446 67

26 4.686 4.6749 37

0 0 3

13 4.637 4.636 13 10 4.535 4.5268 15

2 0 2

8 4.505 4.507 2

3 3.425 3.4222 2

2 2 0

6 3.414 3.417 5 5 3.018 3.0177 8

0 2 4

2 3.001 2.998 3

34 2.904 2.8996 53

4 0 1

21 2.894 2.894 32 100 2.762 2.7614 95

2 2 3

100 2.751 2.750 51

2.7299 19

0 4 2

8 2.342 2.3375 14

0 0 6

5 2.318 2.318 16 81 2.265 2.2634 91

4 0 4

65 2.254 2.253 100

2.2123 3

4 2 1�

2 2.209 2.208 3

1 2.134 2.1341 3

4 2 2

1 2.128 2.129 <1 15 2.037 2.0372 22

0 4 5

4 2.026 2.027 18

2 2 6

1 1.921 1.918 1

10 1.898 1.8980 13

0 2 7

5 1.884 1.884 18 26 1.819 1.8199 32

6 0 3

14 1.815 1.815 35

5 0 5

1 1.798 1.803 <1

2 1.751 1.7505 3

2 4 5�

2 1.744 1.743 3 34 1.710 1.7111 42

4 4 0

9 1.708 1.708 57

1.6811 2

2 0 8

10 1.660 1.6598 10

4 0 7

2 1.650 1.650 8 4 1.631 1.6328 7

6 2 1

1 1.630 1.630 6

1.6068 8

4 4 3

1.603 3

6 1 4�

1 1.602* 1.602 <1

8 1.604 1.6006 5

2 6 2

1.597 2

1.5583 1

0 0 9

1 1.543 1.545 <1

10 1.508 1.5089 12

0 4 8

3 1.500 1.499 16 13 1.493 1.4934 15

2 4 7

4 1.486 1.485 12

7 1.472 1.4736 8

0 8 1

2 1.471 1.471 9 5 1.448 1.4498 7

8 0 2

1.4183 1

2 6 5 3 1.417 1.4182 4

2 2 9

19 1.380 1.3807 26

4 4 6

4 1.376 1.375 17 11 1.364 1.3649 12

0 8 4

4 1.362 1.361 19

1.3648 1

0 2 10

3 1.352 1.3535 5

6 4 1�

2 1.351 1.351 5

1.3350 1

6 4 2

4 1.310 1.3102 5

8 0 5 *Not used for unit cell determination.

CHAPTER 2.2

43

2.2.5.2 The (2+2+2) Jahn-Teller distortion

Both paratacamite congeners described here exhibit rhombic distortion of the M(2)

coordination environment (Table 2.2.5). It was suggested in Chapter 2.1 that this distortion

may be derived from a superimposition of Jahn-Teller distorted Cu(OH)6 octahedra with non-

tetragonally distorted M(OH)6

The M(2)–O bond lengths of both the Mg- and Ni-analogues of paratacamite reported

here appear consistent with the above assumption (Table 2.2.5). The lower proportion of Cu

octahedra, which assumes a statistical distribution of cations

between both interlayer sites for all paratacamite congeners.

2+

octahedra at M(2) result in a smaller (2+2+2) distortion. This may indicate that composition

directly influences the (2+2+2) bond length distribution at M(2) and also that the near

equivalence of M(2)–O bond lengths in sample WAM M365.2003 suggests that it is close to

the upper limit of compositional stability for the R3� phase.

Figure 2.2.1 displays the M(2)O6 coordination environment with atoms pictured with

anisotropic displacement ellipsoids. The maximum principal axis of the O(2) and O(3)

displacement ellipsoids are subparallel with the direction of the M(2)–O bond axis (Figure

2.2.1). A Ni occupancy of 71% at M(2) (Ni–O ca 2.08 Å as in gillardite, Clissold et al., 2007)

with 29% Cu, split into two orientations of Jahn-Teller distorted octahedra, 2/3 and 1/3

occupancy of each orientation (Jahn-Teller distorted as in clinoatacamite with Cu–O ca 2.29,

2.05, and 1.99 Å, Grice et al., 1996) and ignoring the contribution from the small amount of

Co, gives an average bond lengths of 2.11, 2.08 and 2.07 Å. These values are similar to the

data in Table 2.2.5. Similarly, in the Camerones sample a Mg occupancy of 60% at M(2)

(Mg–O ca 2.09 Å as in synthetic Mg-substituted herbertsmithite, composition

Cu3.25Mg0.75(OH)6Cl2, Chu et al., 2010) with 40% Cu, split into two orientations of Jahn-

Teller distorted octahedra, again with 2/3 and 1/3 occupancy of each orientation (Cu–O Jahn-

Teller distortion as described above) results in bond lengths of 2.13, 2.09 and 2.07 Å. Again,

the averages compare well with the observed bond lengths in Table 2.2.5.

The O(1) isotropic displacement value of the Ni-analogue is smaller than the other O

atoms and appears relatively spherical, which is consistent with previous observations for

paratacamite (Fleet 1975; Chapter 2.1). Additionally, the M(2)–O(1) bond is the shortest of

the coordination sphere which is in line with the previous suggestion that this atom is the

pivotal short bond of a dynamic Jahn-Teller distortion (Chapter 2.1). The long M(2)–O(3)

bond would suggest O(3) is in a more stable energetic configuration (Burns and Hawthorne,

1996). In the Mg analogue, the O(1) isotropic displacement value is equivalent to those of the

other O atoms (Table 2.2.3).

CHAPTER 2.2

44

Figure 2.2.1. The (2+2+2) Jahn-Teller distortion of the M(2)O6

The difference in observed and calculated bond lengths, by the means set out above,

may indicate that three orientations of (4+2) Jahn-Teller distorted octahedra occur with a

dynamic interchange between them, the relative occupancy of each orientation being

dependent upon temperature. The occupancy of each orientation would therefore contribute

differently towards the observed (2+2+2) distortion. These results suggest that both

congeners are near the upper limit of substitution for the stability of the R3� structure.

octahedral coordination environment for the Mg- and Ni-rich analogues with the paratacamite structure. Atoms are represented with anisotropic displacement ellipsoids.

The quadratic elongation (QE) and bond-angle variance (BAV) for M(1) and M(2)

were calculated for these samples using the formulation of Robinson et al. (1971) as

implemented in the program VESTA (Momma and Izumi, 2008). The QE value for the M(2)

coordination environment was calculated for the Mg analogue as 1.0497 at 293 K and the Ni

analogue as 1.0511 at 296 K. The distortion present in these analogues is smaller than that

observed in paratacamite, as described in Chapter 2.1, with QE = 1.060 at 300 K. This is

consistent with the expected trend resulting from the higher proportion of non-tetragonally

distorted M(OH)6 octahedra occupying the M(2) site in these new analogues. Based upon the

results obtained in Chapter 2.1, the QE of M(1) is temperature-invariant. However, it may

vary with changes in composition. The phase transformation between the R3� and R3�m

structures, in terms of compositional instability, would also be represented by a merger of the

parameters of the M(2) coordination sphere with that of the M(1) octahedron. The difference

in QE values of M(1) and M(2) for each sample may provide insight into where they are

Ni-analogue

O1

O2

O3

Mg-analogue

O1

O3

O2

CHAPTER 2.2

45

positioned in their respective compositional stability field. Figure 2.2.2 displays the QE and

BAV values for M(1) and M(2) octahedra of type paratacamite in Chapter 2.1 and the two

analogous as a function of composition. Paratacamite from the type specimen with a

composition of Cu3.71Zn0.29(OH)6Cl2, displays the greatest difference in QE values between

the interlayer octahedra. Divergence of these values could indicate a distortion that eventually

destabilises the R3� structure. The QE values of M(1) and M(2) octahedra of the Mg analogue

are nearly equal and the thermal parameters between O(1), O(2), O(3) and O(4) are very

similar. Since the R 3� to R 3� m phase transformation involves a convergence of the

configuration of M(1) and M(2) octahedra, the Mg analogue appears to be close to the limit of

compositional stability for the substitution of Cu2+ by Mg2+

Figure 2.2.2. Quadratic elongation (QE) and bond-angle variance (BAV) for the M(1) (circles) and M(2) (triangles) coordination environment of paratacamite from the BM86958 type specimen at 300 K, x(Zn) = 0.29, the Mg analogue from sample 64041 at 293 K, x(Mg)

in paratacamite. The Ni analogue

shows some deviation in QE and BAV values despite its similar M(2)–O bond lengths. The

distortion may therefore be influenced by some degree of long range crystal structure

constraints. Nevertheless, based on the similarity of the M(1)–O and M(2)–O bond lengths

the mineral also appears to be approaching its upper limit of compositional stability.

= 0.60 and the Ni analogue from sample WAM M365.2003 at 296 K, x(Ni) = 0.71. The composition (x) is based on the formula Cu4-xMx(OH)6Cl2

.

CHAPTER 2.2

46

2.2.6. NEW MINERALS

The structures determined for samples WAM M365.2003 and 64041 are analogous to

that of paratacamite, originally reported by Fleet (1975) and confirmed in Chapter 2.1 using

material from the BM86958 type specimen. Optically they differ from the other related

minerals, because they are both uniaxial (–), ω and ε > 1.8. This is similar to the reports of

“anarakite” by Adib and Otteman (1972). Herbertsmithite is uniaxial (–), ω = 1.825 and

ε = 1.817 (Braithwaite et al., 2004). Their crystal form and colour make them visually

virtually indistinguishable from herbertsmithite, gillardite or paratacamite. In addition, the

X-ray powder data of these samples displays a high degree of correspondence with those for

herbertsmithite, gillardite, anatacamite and type-paratacamite which makes it difficult to

distinguish them from the group. Both a paratacamite-type supercell and a herbertsmithite-

type subcell may be refined from the observed powder data. Differences in observed and

calculated intensities of the X-ray diffraction data suggest that both samples are influenced by

preferred orientation effects. Single-crystal methods are the best way to distinguish these

phases from other members of the basic Cu(II) chloride group.

Separate submissions to the International Mineralogical Associations’

(IMA) Commission on New Minerals, Nomenclature and Classification (CNMNC) have been

made to have both samples recognised as new species on the basis of the Dominant

Constituent Rule (Hatert and Burke, 2008). A suffix-based nomenclature based on the

dominant interlayer cation is proposed in the submissions to signify that they are analogues of

paratacamite. The name “paratacamite-(Ni)” is proposed for the Ni-analogue under the

submission IMA 2013-013. The name “p

aratacamite-(Mg)” is proposed for the Mg-analogue

under the submission IMA 2013-014.

CHAPTER 2.3

47

2.3 THE SINGLE-CRYSTAL X-RAY STRUCTURE OF THE Co ANALOGUE OF HERBERTSMITHITE FROM SALAR GRANDE, IQUIQUE PROVENCE, CHILE

2.3.1 INTRODUCTION

This section describes the single-crystal X-ray structure of a new naturally occurring

basic Cu(II) chloride with dominant Co occupation of the interlayer position,

Cu3(Co,Cu)(OH)6Cl2

2.3.2 SAMPLES AND ANALYSIS

, and which is isomorphous with herbertsmithite and gillardite.

The specimen used in this study (Natural History Museum of Los Angeles, Los

Angeles, California, USA, catalogue number 64031) originated from the Torrecillas Mine,

Salar Grande, Iquique Provence, Tarapacá Region, Chile (approximately 20°58'13''S

70°8'17''W). A polished sample of the specimen was analysed using a JEOL 8600 electron

microprobe in WDS mode. To prevent volatilisation of the sample in the focussed beam, an

accelerating voltage of 15 kV, 20 nA, and 3 μm beam diameter was used. Table 2.3.1

displays the electron microprobe analysis of nine spots. The data show the crystal

was quite zoned. The empirical composition from the average data is

Cu3(Co0.43Cu0.40Mn0.17Ni0.07Mg0.01)∑1.08Cl1.87O6.13H6, calculated based on eight anions pfu.

The sample was unstable in the probe, accounting for the low Cl content (Stormer et al.,

1993). As is the case with other samples analysed in Chapters 2.1 and 2.2, the

structure is composed of sheets of composition Cu3Cl2(OH)62– linked by M2+ ions lying

between them. Therefore the empirical formula may be written as

(Cu3.40Co0.43Mn0.17Ni0.07Mg0.01)Σ4.08Cl2(OH)6. Normalising this formula to Σ(cations) = 4

gives Cu3(Co0.42Cu0.33Mn0.17Ni0.07Mg0.01)(OH)6Cl2.

Table 2.3.1. Electron microprobe analyses of 64031 Range Average* Empirical Normalised (wt%) (wt%) composition composition** CuO 56.44–67.65 62.80(3.43) 3.40 3.33 CoO 5.64–9.85 7.49(1.33) 0.43 0.42 NiO 0.51–2.77 1.23(0.71) 0.07 0.07 MnO 0.78–3.68 2.72(0.97) 0.17 0.17 MgO 0–0.12 0.06(0.04) 0.01 0.01 Cl 14.58–15.80 15.40(0.37) 1.87 2.00 H2O 12.57 6.00 6.00 O ≡ Cl -3.48 Total 98.79 *Standard deviation of the average value is in parentheses. **Compositions were normalised to Σ(cations) = 4.00.

CHAPTER 2.3

48

2.3.2.1 Optical and physical properties

The crystals occur as green rhombs, no larger than 1 mm across. These rhombs are

sometimes stacked along the c axis in finger-like, parallel growths or form V-shaped twins.

The sample is green with a light green streak and transparent with a vitreous lustre. The

crystals are non-fluorescent. Mohs hardness is ~3. The crystals are brittle with a perfect

cleavage on {101} and no parting was observed. The fracture is conchoidal. The calculated

density is 3.695 g cm-3, based on the empirical formula and the measured density by flotation

in Clerici solution is 3.64(2) g cm-3

2.3.3 CRYSTALLOGRAPHY

. The crystals are uniaxial (–), ω and ε > 1.8 (reacts

rapidly with RI fluids > 1.8) and pleochroic with O (bluish green) > E (slightly yellowish

green). Chemical tests indicate that the mineral is readily soluble in cold, dilute HCl.

A small, tabular crystal was mounted on a Rigaku R-Axis Rapid II curved imaging

plate microdiffractometer with monochromated MoKα radiation and analysed at 293(2) K. A

unit cell of a = 6.8436(6) and c = 14.064(1) Å was determined from the data set. No

superstructure reflections were observed. The structure was solved by direct methods in space

group R3�m using SHELXS-97 (Sheldrick, 2008). The heavy atom positions were located in

analogous positions to those found in herbertsmithite. Of the two metal positions identified,

the M(2) position at 9e (0.5, 0, 0), which is bonded to four equatorial OH- and two axial Cl-

ligands in a typical (4+2) Jahn-Teller distorted geometry, has been well established to ideally

be fully occupied by Cu (Braithwaite et al., 2004; Clissold et al., 2007; Fleet, 1975). The

M(1) site at 3b (0, 0, 0.5), which is bonded to six symmetry equivalent O atoms in a slightly

angularly distorted octahedron, exhibits the full extent of Cu substitution.

The same approach as in previous studies was followed here with Cu occupying the

M(2) site and excess being placed in M(1) along with the average values for Co, Mn, Mg and

Ni. The M(1) site metal composition of (Co0.42Cu0.33Mn0.17Ni0.07Mg0.01) shows that Co is the

dominant cation, but occupies less than 50% of the position due to the minor but significant

co-occupancy of Mn and trace amounts of Ni and Mg. A difference map revealed the

position of the single crystallographic O and corresponding H atom. The H atom was fixed at

0.90 ± 0.03 Å from the O atom. Refinement converged with anisotropic displacement

parameters for all non-H atoms. Final refinement gave R1 and wR2 of 0.0226 and 0.0552,

respectively, for 183 reflections with Fobs > 4σ(Fobs). The weighting scheme used was

w = 1/[σ2(Fo) 2 + (0.0386P)2 + 2.82P] where P = [max(0, Fo)2 + (2Fc)2]/3 as defined by

SHELX-97 (Sheldrick, 2008).

CHAPTER 2.3

49

Structure refinement details are given in Table 2.3.2. Final atom coordinates and

anisotropic displacement parameters are in listed Table 2.3.3 and selected bond lengths and

angles in Table 2.3.4.

2.2.4 X-RAY POWDER DIFFRACTION

The X-ray powder diffraction pattern was measured using the same instrumentation

and radiation described above for single-crystal measurements. Unit cell parameters refined

from the powder data, using the JADE 2010 software package (Materials Data, Inc, 2011)

with whole pattern fitting, is a = 6.8383(9), c = 14.081(2) Å and V = 570.2(1) Å3

2.3.4 DISCUSSION

. Powder

diffraction data are listed in Table 2.3.5.

The structure determined here, with dominant Co in the interlayer position, is

analogous to those described by Braithwaite et al. (2004) for herbertsmithite, Clissold et al.

(2007) for gillardite and Fleet (1975) for the paratacamite substructure and extends upon the

known substitution series for naturally occurring members of the paratacamite group. The

composition, however, is far from the ideal Cu3Co(OH)6Cl2 formula, which is expected

to possess a unit cell somewhat larger than the determined dimensions of a = 6.8436(6) and

c = 14.064(1) Å. This would be indicative of the increased influence of high spin Co2+ with

an effective ionic radius of 0.74 Å for a six coordinate environment, being slightly larger than

that of Cu2+ with 0.73 Å and in line with previous analyses of herbertsmithite and gillardite

unit cells (Braithwaite et al., 2004; Clissold et al., 2007). Oswald and Feitknecht (1964)

reported the unit cell dimensions of synthetic Co2(OH)3Cl, with R3�m symmetry as a = 6.84

and c = 14.50 Å.

The H–Cl distance was determined as 2.32 Å, which is longer than that reported for

gillardite (2.26 Å) by Clissold et al. (2007). Whilst the true H position is not accurately

represented by data collected using X-rays, the O–Cl distance is a more reliable estimate of

the strength of H bonding. In the structure reported here an O–Cl distance of 3.079(3) Å was

found and is significantly longer than that reported for gillardite (3.049(8) Å) and

herbertsmithite (3.071 Å) (Clissold et al., 2007; Braithwaite et al., 2004). The H bond

network provides additional support between the sheets of Jahn-Teller distorted

[Cu(OH)4Cl2], and is weaker when Co is the dominant substituting cation. By this inference,

the strongest H bonding would be expected to occur in the structure of gillardite.

CHAPTER 2.3

50

Table 2.3.2. Crystal data and structure refinement details for 64031. Cation normalised composition Cu3(Co0.42Cu0.33Mn0.17Ni0.07Mg0.01)Cl2O6H

Space group R3�m Unit cell dimensions a = 6.8436(6) Å c = 14.0637(10) Å Volume 570.42(8) Å3 Z 3 Density (calculated) 3.694 g cm-3 Absorption coefficient 11.214 mm-1 F(000) 607 Crystal size 0.15 x 0.10 x 0.09 mm Theta range for data collection 3.73 to 27.32°. Index ranges -8 ≤ h ≤ 8, -8 ≤ k ≤ 6, -18 ≤ l ≤ 18 Reflections collected 3409 Independent reflections 183 [R

6 Formula weight 422.98 Temperature 293(2) K Wavelength 0.71075 Å Crystal system Hexagonal

int = 0.0203] Completeness to theta=27.32° 99.5 % Max. and min. transmission 0.432 and 0.284 Refinement method Full-matrix least-squares on F2 Data/restraints/parameters 183/1/20 Goodness-of-fit on F2 1.099 Final R [Iobs > 2σ(Iobs)] R1 = 0.0226, wR2 = 0.0552 R indices (all data) R1 = 0.0226, wR2 = 0.0552 Δρmax, Δρmin (e. Å-3) 1.263 and -0.508

Table 2.3.3. Final atom coordinates (x104) and equivalent isotropic displacement parameters (Å2 x103). Ueq = 1/3(U11 + U22 + U33). x/a y/b z/c Ueq U11 U22 U33 U23 U13 U12 Co* 0 0 0.5 0.0138(3) 0.012(1) U11 0.009(1) 0 0 0.5U11 Cu 0.5 0 0 0.0124(3) 0.013(1) 0.012(1) 0.015(1) 0.002(1) 0.001(1) 0.006(1) Cl 0 0 0.19407(8) 0.0164(3) 0.018(1) U11 0.015(1) 0 0 0.5U11 O 0.2064(2) -0.2064(2) 0.0612(2) 0.0155(5) 0.015(1) U11 0.019(1) -0.002(1) U23 0.007(1) H 0.135(3) -0.135(3) 0.074(4) 0.063(20) *The assigned occupancy for the Co site is (Cu0.33Co0.42Mn0.17Ni0.07Mg0.01). Table 2.3.4. Selected bond lengths [Å] and angles [°] for 64031. Interlayer Intralayer M1–O 2.114(3) M2–O 1.983(1) O–M1–O 76.1(1) M2–Cl 2.7821(9) O–M1–O 103.9(1) O–M2–O 97.9(2) O–M1–O 180 O–M2–O 82.2(2) O–M2–Cl 82.36(6) Cl–M2–Cl 180

CHAPTER 2.3

51

Table 2.3.5. X-ray powder diffraction data for sample 64031. Iobs dobs dcalc Icalc hkl

Iobs dobs dcalc Icalc hkl 90 5.469 5.4616 100 1 0 1

7 1.666 1.6629 5 0 2 7 18 4.701 4.6879 18 0 0 3

5 1.632 1.6327 7 1 3 1

10 4.557 4.5318 3 0 1 2

7 1.606 1.6072 4 2 2 3 8 3.420 3.4218 5 1 1 0

1.6006 2 3 1 2

5 3.032 3.0239 2 1 0 4

11 1.513 1.5119 9 2 0 8 22 2.905 2.8997 25 0 2 1

1.5106 2 3 0 6

100 2.766 2.7638 84 1 1 3 15 1.497 1.4957 13 2 1 7

2.7308 19 2 0 2

4 1.472 1.4735 4 4 0 1

12 2.348 2.3440 12 0 0 6

4 1.450 1.4498 6 0 4 2 66 2.269 2.2659 70 0 2 4 6 1.422

1.4214 4 1 1 9 7 2.215 2.2122 5 2 1 1

1.4192 1 3 1 5

11 2.041 2.0401 11 2 0 5

23 1.383 1.3830 2 1 2 8

1.9338 1 1 1 6

1.3819 17 2 2 6

13 1.906 1.9027 12 1 0 7

8 1.364 1.3684 1 1 0 10 26 1.822 1.8205 27 0 3 3

1.3654 9 4 0 4 3 1.754 1.7523 4 1 2 5

4 1.352 1.3534 5 3 2 1

33 1.711 1.7109 32 2 2 0 3 1.311 1.3109 3 0 4 5

1.6854 1 0 1 8

There is a high correspondence of peak positions between the X-ray powder data

(Table 2.3.5) of this sample and those reported for herbertsmithite, gillardite and paratacamite

(Braithwaite et al., 2004; Clissold et al., 2007; Jambor et al., 1996). Comparative data for

64031, herbertsmithite and gillardite are reported in Table 2.3.6. Optically and

crystallographically it would be difficult to distinguish a crystal of herbertsmithite with its Co

analogue unless these methods were combined with chemical analyses.

This new analogue represents an intermediate composition along the solid solution

series of rhombohedral members described by Jambor et al. (1996) and Braithwaite et al.

(2004). These authors suggested that the monoclinic structure of clinoatacamite is stable with

less than 1/3 interlayer occupancy of a substituting cation. They determined from synthetic

material that the clinoatacamite structure destabilises at about 5 wt% Co substitution.

However, in their work two-phase mixtures of rhombohedral Co2(OH)3Cl and monoclinic

Cu2(OH)3Cl were reported for synthetic material. An end-member of Cu3Co(OH)6Cl2

is

proposed for the Co analogue of herbertsmithite described here, based on knowledge of other

members of the group. The lower end-member composition for this phase is unknown.

CHAPTER 2.3

52

Table 2.3.6. Comparative data for 64031, herbertsmithite and gillardite. 64031* herbertsmithite** gillardite** Formula CoCu3Cl2(OH)6 ZnCu3Cl2(OH)6 NiCu3Cl2(OH)6 Space group R3�m R3�m R3�m a (Å) 6.8383(9) 6.834(1) 6.8364(1) c (Å) 14.081(2) 14.075(2) 13.8459(4) Z 3 3 3 Optical character uniaxial (–) uniaxial (–) uniaxial (+) *(Cu3.33Co0.42Mn0.17Ni0.07Mg0.01)Σ4.00Cl2(OH)6. **Close to stoichiometrically pure.

2.3.5 A NEW MINERAL

A submission to the IMA CNMNC has been made with relevant data for this new Co

mineral to have it established as a separate species under proposal IMA 2013-011.

CHAPTER 2.4

53

2.4 THE COMPOSITION-DEPENDENT STRUCTURAL TRANSFORMATION SERIES OF THE PARATACAMITE GROUP

2.4.1 INTRODUCTION

It was shown in Chapter 2.1 that a reversible structural transformation from

paratacamite R3� to herbertsmithite R3�m occurs at elevated temperatures. This transformation

is in line with the predicted space group chain P1�→ R3� → R3�m, discussed by Malcherek and

Schlüter (2009). These authors also suggested that a series of compositionally related

structural transformations occurs from anatacamite to clinoatacamite and herbertsmithite,

described by the space group chain P1�→ P21/c (P21/n) → R3�m. Earlier investigations of the

phase transformations by Jambor et al. (1996) and Braithwaite et al. (2004) indicated that

paratacamite is an intermediate phase between clinoatacamite and herbertsmithite. The

former authors reported that the transformation from clinoatacamite to a rhombohedral phase

occurs over a range of compositions between ca Cu3.80Zn0.20(OH)6Cl2 and

Cu3.67Zn0.33(OH)6Cl2. Braithwaite et al. (2004) suggested that the stability field of

paratacamite extents between the compositions ca Cu3.67Zn0.33(OH)6Cl2 and

Cu3.50Zn0.50(OH)6Cl2

This crystallographic investigation of naturally occurring samples from the series was

completed to elucidate the compositional boundary between the R3� and R3�m structures in

terms of Zn and Ni substitution.

. The composition determined for paratacamite in Chapter 2.1 is in line

with the observations made by these authors. However, the recent discovery of Mg- and Ni-

bearing analogues of paratacamite with a composition significantly greater than 50%

interlayer occupancy of the substituting cation (Chapter 2.2) has indicated that the

compositional stability field of paratacamite may be different than expected.

2.4.2 SAMPLES AND ANALYSIS

Specimens of the basic Cu(II) chlorides were obtained from the Mineralogical

Museum, Hamburg, and from several private collections for compositional and

crystallographic analysis. The sample names and localities are reported in Table 2.4.1.

Two different electron microprobes were used, a JEOL 8600 electron microprobe for

samples originating from 132 N nickel mine, Widgiemooltha, Western Australia, and a

Cameca SX 100 electron microprobe for the remaining samples. Both microprobes were

operated in WDS mode with an accelerating voltage of 15 kV, a specimen current of 20 nA

CHAPTER 2.4

54

and focussed beam. Table 2.4.1 also lists the empirical formulae determined from these

analyses. The simplified formula, based on Σ(cations) = 4, for each sample was used in the

structural refinement and is reported as follows: CB03, Cu3.61Ni0.39(OH)6Cl2;

CB07, Cu3.51Ni0.49(OH)6Cl2; G8502, Cu3.12Ni0.88(OH)6Cl2; G8568,

Cu3.11Ni0.88Co0.01(OH)6Cl2; G7751, Cu3.09Ni0.90Co0.01(OH)6Cl2; MD166-3,

Cu3.65Zn0.35(OH)6Cl2; MM02, Cu3.61Zn0.39(OH)6Cl2 and MD166-2, Cu3.36Zn0.64(OH)6Cl2

2.4.3 CRYSTALLOGRAPHY

.

Crystals of Ni-bearing specimens from the 132 N deposit G8502, G8568, and G7751,

were measured at 293(2) K using a Bruker Smart 1000 CCD diffractometer with graphite-

monochromated MoKα radiation. The remaining samples from the Carr Boyd Rocks mine,

the Murrin Murrin mine, and the San Francisco mine, CB03, CB07, MM02, MD166-2 and

MD166-3 were analysed at 294(2) K on a Nonius Kappa CCD diffractometer with MoKα

radiation. All data sets were corrected for absorption, Lorentz and polarisation effects. Final

unit cell dimensions were determined by a least-squares refinement of the full data sets and

all structure refinements were made using SHELXL (Sheldrick, 2008) based on atom

coordinates reported for analogous phases (Braithwaite et al., 2004; Clissold et al., 2007).

Special attention was given to the identification of weak reflections characteristic of

paratacamite at half integer positions of h and k. Of some 20 crystals analysed of five

specimens from four localities the superlattice reflections of paratacamite were never

identified. Unit cell parameters determined for each sample indicated the paratacamite

substructure (Table 2.4.2). All samples containing Ni as the substituting cation possessed cell

axes analogous to those of gillardite (a ~ 6.8, c ~ 13.9 Å). Along the compositional series

analysed, the c axis showed the greatest variation, decreasing from 13.936(2) to 13.848(2) Å

as Cu is replaced by Ni. The cell dimensions of sample G7751 are a = 6.8421(8) and

c = 13.848(2) Å, and the composition Cu3(Ni0.90Cu0.09Co0.01)(OH)6Cl2, compare well with

the unit cell reported for holotype gillardite, a = 6.8364(1) and c = 13.8459(4) Å,

Cu3(Ni0.903Cu0.081Co0.012Fe0.004)(OH)6Cl2

, by Clissold et al. (2007). Similarly, Zn-bearing

samples exhibited unit cell parameters related to herbertsmithite (a ~ 6.8, c ~ 14.1 Å). The

range detected expressed the varying contribution of Zn content, increasing from 14.046(9) to

14.062(4) Å, as Zn content increases.

Table 2.4.1. Electron microprobe analyses of material in this study. *Average (above), range (below) (wt%) Sample Spots CuO ZnO NiO MgO CoO MnO Cl H2O** O≡Cl Total Empirical formula CB03 100 67.29(0.93) - 6.75(0.69) - - 0.01(0.02) 16.12(0.14) 12.59 -3.64 102.76 (Cu3.63Ni0.39)∑4.02Cl1.95O6.05H6.00 65.64–70.59 4.54–7.70 0–0.05 15.78–16.42 CB07 8 65.79(2.29) - 8.71(1.63) - - 0.07(0.04) 16.70(0.13) 12.77 -3.77 100.27 (Cu3.50Ni0.49)∑3.99Cl2.00O6.00H6.00 62.99–69.65 5.17–10.12 0–0.12 16.44–16.84 G8502 8 60.81(0.41) - 16.19(0.96) 0.06(0.05) 0.15(0.07) - 17.23(12) 13.28 -3.89 103.83 (Cu3.11Ni0.88)∑3.99Cl1.98O6.02H6.00

59.92–61.16 14.93–17.45 0–0.14 0.06–0.24 17.04–17.46 G8568 12 60.25(1.98) - 16.01(1.40) 0.02(0.03) 0.25(0.07) - 17.40(0.26) 13.20 -3.93 103.20 (Cu3.10Ni0.88Co0.01)∑3.99Cl2.01O5.99H6.00 56.60–64.86 13.92–18.49 0–0.11 0.12–0.40 17.09–17.95 G7751 16 59.11(2.21) - 16.32(1.33) 0.02(0.03) 0.24(0.10) - 17.58(0.22) 13.10 -3.97 102.40 (Cu3.07Ni0.90Co0.01)∑3.98Cl2.05O5.95H

MD166-3 15 68.10(0.52) 6.65(0.11) - - - - 16.27(0.21) 12.63 -3.68 99.97 (Cu

6.00 55.96–62.27 14.74–19.05 0–0.10 0.11–0.49 17.24–18.00

3.67Zn0.35)∑4.02Cl1.97O6.03H6.00 67.39–69.27 6.44–6.80 15.98–16.85 MM02 100 66.76(2.41) 7.32(1.67) - - - - 16.66(0.26) 12.59 -3.77 99.56 (Cu3.61Zn0.39)∑4.00Cl2.02O5.98H6.00 62.03–71.94 4.56–11.29 16.17–17.51 MD166-2 40 61.42(0.86) 11.93(0.83) - - - - 16.57(0.26) 12.46 -3.74 98.64 (Cu3.35Zn0.64)∑3.99Cl2.03O5.97H6.00

59.96–64.91

*Fields with a dash (-) represent elements not detected. **H2O content was calculated based on 8 anions pfu.

9.57–13.84 16.23–17.34 CB03 Carr Boyd Rocks Mine, Western Australia, Australia CB07 Carr Boyd Rocks Mine, Western Australia, Australia G8502 132N nickel mine, Widgiemoothla, Western Australia, Australia G8568 132N nickel mine, Widgiemoothla, Western Australia, Australia G7751 132N nickel mine, Widgiemoothla, Western Australia, Australia MD166-3 San Francisco Mine, Sierra Gorda, Chile MM02 Murrin Murrin mine, Western Australia, Australia MD166-2 San Francisco Mine, Sierra Gorda, Chile

CH

APTER

2.4

55

CHAPTER 2.4

56

The reported unit cell for herbertsmithite is a = 6.834, c = 14.075 Å for material of

end-member composition Cu3Zn(OH)6Cl2

2.4.4 RESULTS AND DISCUSSION

(Braithwaite et al., 2004) and is in line with the

composition versus unit cell relationship determined here. These results are also in accord

with the variation in cell parameters reported for synthetic rhombohedral Zn-bearing

members of the basic Cu(II) chlorides by Jambor et al. (1996). Due to the absence of any

superlattice reflections and the similarity of these unit cells with those reported for

herbertsmithite and gillardite, structural refinements were made in space group R3�m for all

data sets. All structures were refined based on the atom coordinates established by

Braithwaite et al. (2004) and Clissold et al. (2007) and converged to acceptable residuals and

anisotropic thermal parameters establishing their identity as either herbertsmithite or

gillardite. Structure refinement details can be found in Table 2.4.2. Atom coordinates and

anisotropic thermal parameters can be found in Table 2.4.3. Selected crystallographic data are

given in Table 2.4.4.

The compositional range determined for herbertsmithite and gillardite single-crystals,

Cu3.65Zn0.35(OH)6Cl2–Cu3.36Zn0.64(OH)6Cl2 and Cu3.61Ni0.39(OH)6Cl2–Cu3.13Ni0.87(OH)6Cl2,

respectively, indicates that the R3�m structure can exist down to the monoclinic–rhombohedral

transition zone determined by Jambor et al. (1996), between ca Cu3.75Zn0.25(OH)6Cl2 to

Cu3.66Zn0.34(OH)6Cl2. Schores et al. (2005) reported X-ray structural data for synthetic

single-crystals of Zn-bearing paratacamite, produced by hydrothermal methods. Although, all

structure refinements by these authors were made on the R3�m subcell. The authors did not

mention the presence of any superlattice reflections and their data are in complete agreement

with those for herbertsmithite. The range of compositions studied by these authors is

Cu3.67Zn0.33(OH)6Cl2–Cu3Zn(OH)6Cl2

The description of paratacamite by Fleet (1975) was unfortunately lacking a chemical

analysis. Braithwaite et al. (2004) reported numerous electron microprobe analyses from

various specimens and suggested that paratacamite was stable up to a composition of ca

Cu

, and supports these observations.

3.44Zn0.56(OH)6Cl2 (Sample 17 in their paper). However, they did not report any single-

crystal data on paratacamite and their figured IR spectra are very similar to those of

herbertsmithite. Crystallographic data from the structural refinements are displayed in Table

2.4.4 and compared with data from the paratacamite R3�m substructure and from the literature.

CHAPTER 2.4

57

Table 2.4.2. Crystal data and structure refinements of samples in this study Sample MD166-3 MM02 MD166-2 Normalised formulaa Cu3.65Zn0.35Cl2O6H6

Cu3.61Zn0.39Cl2O6H6 Cu3.36Zn0.64Cl2O6H

Space group R3�m R3�m R3�m Unit cell dimensions a (Å) 6.835(4) 6.839(7) 6.8347(9) c (Å) 14.046(9) 14.052(4) 14.062(4) Volume (Å

6

Formula weight 427.75 427.82 428.28 Temperature (K) 294(2) 294(2) 294(2) Wavelength (Å) 0.71073 0.71073 0.71073 Crystal system trigonal trigonal trigonal

3) 568.3(6) 569.2(8) 568.87(19) Z, Calculated density (g cm-3) 3, 3.750 3, 3.744 3, 3.750 Absorption coefficient (mm-1) 11.885 11.880 11.976 F(000) 613 613 614 Crystal size (mm) 0.11 x 0.09 x 0.08 0.24 x 0.20 x 0.16 0.25 x 0.20 x 0.15 Theta range for data 3.74 to 34.98° 3.73 to 34.95° 3.73 to 34.98° Limiting indices -10 ≤ h ≤ 10 -10 ≤ h ≤ 10 -10 ≤ h ≤ 9 -11 ≤ k ≤ 11 -10 ≤ k ≤ 10 -10 ≤ k ≤ 11 -21 ≤ l ≤ 22 -22 ≤ l ≤ 22 -22 ≤ l ≤ 22 Reflections/unique 3714/339 4024/340 3797/340 R int 0.0369 0.0290 0.0289 Completeness to theta 34.98° 99.7 % 34.95 100.0 % 34.97 100.0 % Refinement method Full-matrix Full-matrix Full-matrix least-squares on F2 least-squares on F2 least-squares on F2 Data/restraints/parameters 339/1/18 340/1/19 340/1/19 Goodness-of-fit on F2 1.326 1.322 1.415 Final R indices [I > 2σ(I)] R1 0.0153 0.0191 0.0192 wR2 0.0337 0.0491 0.0466 R indices (all data) R1 0.0172 0.0204 0.0197 wR2 0.0340 0.0495 0.0469 Δρmax, Δρmin (e.Å-3

) 0.818 and -0.636 0.555 and -0.525 0.495 and -1.274

Table 2.4.2. Continued Sample CB03 CB07 G8502 G8568 G7751 Normalised formulaa Cu3.61Ni0.39Cl2O6H6 Cu3.51Ni0.49Cl2O6H6 Cu3.12Ni0.88Cl2O6H6 Cu3.11Ni0.88Co0.01Cl2O6H6 Cu3.09Ni0.90Co0.01Cl2O6H

Space group R3�m R3�m R3�m R3�m R3�m Unit cell dimensions a (Å) 6.8376(8) 6.841(4) 6.8403(8) 6.8407(9) 6.8421(8) c (Å) 13.936(2) 13.944(5) 13.852(2) 13.846(2) 13.848(2) Volume (Å

6

Formula weight 425.24 424.74 422.91 422.81 422.71 Temperature (K) 294(2) 294(2) 293(2) 293(2) 293(2) Wavelength (Å) 0.71073 0.71073 0.71073 0.71073 0.71073 Crystal system trigonal trigonal trigonal trigonal trigonal

3) 564.27(11) 565.1(5) 561.30(12) 561.10(17) 561.42(11) Z, Calculated density (g cm-3) 3, 3.754 3, 3.744 3, 3.753 3, 3.754 3, 3.751 Absorption coefficient (mm-1) 11.717 11.666 11.622 11.616

Crystal size (mm) 0.22 x 0.18 x 0.15 0.15 x 0.11 x 0.08 0.18 x 0.20 x 0.20 0.08 x 0.10 x 0.10 0.10 x 0.10 x 0.14 Theta range for data 3.74 to 34.97° 3.74 to 34.99 ° 3.74 to 28.16° 3.74 to 28.23° 3.74 to 28.27° Limiting indices -10 ≤ h ≤ 10 -10 ≤ h ≤ 11 -9 ≤ h ≤ 8 -8 ≤ h ≤ 8 -8 ≤ h ≤ 9 -10 ≤ k ≤ 10 -11 ≤ k ≤ 11 -8 ≤ k ≤ 8 -8 ≤ k ≤ 8 -8 ≤ k ≤ 7 -22 ≤ l ≤ 22 -22 ≤ l ≤ 21 -17 ≤ l ≤ 17 -15 ≤ l ≤ 17 -18 ≤ l ≤ 18 Reflections/unique 8365/336 3755/338 1462/186 1481/187 1450/189 R

11.603 F(000) 611 611 609 609 609

int 0.0343 0.0290 0.0254 0.0202 0.0218 Completeness to theta 34.97° 99.7% 34.99° 100.0% 28.16° 96.9 % 28.23° 96.4 % 28.27° 95.9% Refinement method Full-matrix Full-matrix Full-matrix Full-matrix Full-matrix least-squares on F2 least-squares on F2 least-squares on F2 least-squares on F2 least-squares on F2 Data/restraints/parameters 336/1/18 338/1/19 186/1/19 187/1/19 189/1/19 Goodness-of-fit on F2 1.279 1.221 1.394 1.325 1.290 Final R indices[I > 2σ(I)] R1 0.0159 0.0139 0.0297 0.221 0.231 wR2 0.0385 0.0327 0.0786 0.0569 0.0568 R indices (all data) R1 0.0166 0.0151 0.0297 0.222 0.234 wR2 0.0387 0.0330 0.0786 0.0570 0.0571 Δρmax, Δρmin (e.Å-3) 0.558 and -0.759 0.444 and -0.611 0.609 and -2.449 0.467 and -1.741 0.576 and -1.5810

a

CH

APTER

2.4

The normalised formula used in the structure refinements was made to ∑(cations) = 4.

58

CHAPTER 2.4

59

Table 2.4.3. Atomic coordinates, isotropic (Å2) and anisotropic (Å2) displacement parameters for samples of herbertsmithite and gillardite. x y z U(eq) U11 U22 U33 U23 U13 U12 *MD166-2 M1 0 0 0.5 0.0091(1) 0.0116(1) U11 0.0061(2) 0 0 0.0053(1) M2 0.5 0 0 0.0091(1) 0.0085(1) 0.0077(1) 0.0107(1) 0.0021(1) 0.5U23 0.0039(7) Cl 0 0 0.1943(1) 0.0130(1) 0.0143(2) U11 0.0104(2) 0 0 0.0072(1) O 0.2059(1) -0.2059(1) 0.0613(1) 0.0113(2) 0.0100(3) U11 0.0143(4) -0.0016(2) -U23 0.0052(4) H 0.1385(18) -0.1385(18) 0.0850(19) 0.0204

*MM02 M1 0 0 0.5 0.0084(1) 0.0097(2) U11 0.0058(2) 0 0 0.0049(1) M2 0.5 0 0 0.0092(1) 0.0088(1) 0.0076(1) 0.0109(2) 0.0016(8) 0.5U23 0.0038(1) Cl 0 0 0.1940(1) 0.0125(1) 0.0136(2) U11 0.0102(3) 0 0 0.0068(1) O 0.2064(1) -0.2064(1) 0.0614(1) 0.0136(2) 0.0111(4) U11 0.0182(6) -0.0033(2) -U23 0.0054(4) H 0.1445(15) -0.1445(15) 0.0849(23) 0.0257

*MD166-3 M1 0 0 0.5 0.0083(1) 0.0091(1) U11 0.0065(2) 0 0 0.0046(6) M2 0.5 0 0 0.0090(1) 0.0079(1) 0.0070(1) 0.0118(1) 0.0020(1) 0.5U23 0.0035(1) Cl 0 0 0.1941(1) 0.1258(12) 0.0133(2) U11 0.0112(3) 0 0 0.0066(1) O 0.2062(1) -0.2062(1) 0.0612(1) 0.0119(2) 0.0098(3) U11 0.0163(5) -0.0022(1) -U23 0.0049(4) H 0.1405(12) -0.1405(12) 0.0850(18) 0.0179 **CB03 M1 0 0 0.5 0.0064(1) 0.0075(1) U11 0.0042(2) 0 0 0.0038(1) M2 0.5 0 0 0.0079(1) 0.0077(1) 0.0066(1) 0.0089(1) 0.0017(1) 0.5U23 0.0033(1) Cl 0 0 0.1935(1) 0.0110(1) 0.0124(1) U11 0.0083(2) 0 0 0.0062(1) O 0.2071(1) -0.2071(1) 0.0622(1) 0.0115(2) 0.0099(4) U11 0.0148(5) -0.0027(2) -U23 0.0049(4) H 0.1410(13) -0.1410(13) 0.0875(19) 0.0173 **CB07 M1 0 0 0.5 0.0085(1) 0.0089(1) U11 0.0078(1) 0 0 0.0045(6) M2 0.5 0 0 0.0103(7) 0.0094(1) 0.0082(1) 0.0129(1) 0.0017(1) 0.5U23 0.0041(1) Cl 0 0 0.1936(1) 0.0134(1) 0.0141(1) U11 0.0120(1) 0 0 0.0070(1) O 0.2071(1) -0.2071(1) 0.0621(1) 0.0141(2) 0.0113(1) U11 0.0194(4) -0.0030(2) -U23 0.0054(3) H 0.1427(14) -0.1427(14) 0.0870(21) 0.0499

**G8502 M1 0 0 0.5 0.0028(4) 0.0033(5) U11 0.0018(6) 0 0 0.0016(2) M2 0.5 0 0 0.0034(3) 0.0034(4) 0.0029(4) 0.0041(5) 0.0011(2) 0.5U23 0.0015(2) Cl 0 0 0.1933(1) 0.0057(4) 0.0066(6) U11 0.0048(8) 0 0 0.0033(3) O 0.2072(3) -0.2072(3) 0.0625(2) 0.0054(6) 0.0054(10) U11 0.0065(12) -0.0002(6) -U23 0.0034(13) H 0.1214(24) -0.1214(24) 0.0922(34) 0.0073

**G8568 M1 0 0 0.5 0.0038(3) 0.0042(3) U11 0.0028(5) 0 0 0.0021(2) M2 0.5 0 0 0.0046(2) 0.0043(3) 0.0042(3) 0.0051(4) 0.0012(1) 0.5U23 0.0021(2) Cl 0 0 0.1933(1) 0.0066(3) 0.0075(4) U11 0.0050(6) 0 0 0.0037(2) O 0.2072(2) -0.2072(2) 0.0623(2) 0.0065(5) 0.0062(7) U11 0.0077(10) -0.0003(4) -U23 0.0036(9) H 0.1281(23) -0.1281(23) 0.0901(33) 0.0301 **G7751 M1 0 0 0.5 0.0033(3) 0.0034(3) U11 0.0032(5) 0 0 0.0017(2) M2 0.5 0 0 0.0041(2) 0.0033(3) 0.0031(3) 0.0058(3) 0.0013(2) 0.5U23 0.0015(1) Cl 0 0 0.1933(1) 0.0066(3) 0.0070(4) U11 0.0059(6) 0 0 0.0035(2) O 0.2073(2) -0.2073(2) 0.0625(2) 0.0056(5) 0.0052(8) U11 0.0075(9) -0.0003(4) -U23 0.0033(9) H 0.1271(21) -0.1271(21) 0.0897(29) 0.0166 The anisotropic displacement factor exponent takes the form: -2π2[h2a*2U11

+...+ 2hka*b*U12]. Ueq = 1/3(U11 + U22 + U33). *herbertsmithite. **gillardite.

CH

APTER

2.4

Table 2.4.4. Unit cell data and selected bond lengths and angles of the paratacamite substructure in space group R3�m. Interlayer Unit cell parameters M1–O O–M1–O M2–O M2–Cl O–M2–O O–M2–Cl O–Cl Sample cations M(x) a (Å) c (Å) (Å) cis (°) (Å) (Å) cis (°) cis (°) 1Paratacamite* Cu > Zn# (-) 6.827(5) 14.041(6) 2.11 (-) 1.98 2.78 (-) (-) 3.07 2BM86958* Cu > Zn 0.29 6.8247(1) 14.0298(4) 2.102(2) 103.99(7) 1.9774(9) 2.7774(6) 98.25(11) 97.59(7) 3.072(1) 3MD166-3 Cu > Zn 0.35 6.835(4) 14.046(9) 2.112(2) 103.77(7) 1.982(1) 2.778(1) 97.77(8) 97.59(5) 3.073(2) 3MM02 Cu > Zn 0.39 6.839(7) 14.052(4) 2.109(2) 103.78(6) 1.983(2) 2.781(2) 97.94(9) 97.56(5) 3.074(2) 3MD166-2 Zn > Cu 0.64 6.8347(9) 14.062(4) 2.114(1) 103.67(5) 1.9838(6) 2.7778(6) 97.62(7) 97.49(3) 3.072(1) 4Herbertsmithite Zn > Cu 1 6.834(1) 14.075(2) 2.119(1) (-) 1.985(1) 2.779(1) (-) (-) 3.071 3CB03 Cu > Ni 0.39 6.8376(6) 13.936(2) 2.088(1) 103.31(5) 1.9827(6) 2.7735(5) 98.42(8) 97.66(3) 3.060(1) 3CB07 Cu > Ni 0.49 6.841(4) 13.944(5) 2.089(1) 103.36(5) 1.983(1) 2.775(1) 98.46(7) 97.69(4) 3.063(2) 5WAM M365.2003* Ni > Cu§ 0.73 6.843(1) 13.935(3) 2.088(2) 103.39(9) 1.982(1) 2.775(8) 98.48(13) 97.75(5) 3.064(2) 3G8502 Ni > Cu 0.88 6.8403(8) 13.852(2) 2.077(3) 102.93(14) 1.983(2) 2.768(1) 98.48(19) 97.80(8) 3.051(3) 3G8568 Ni > Cu§ 0.89 6.8407(9) 13.846(2) 2.079(2) 102.99(10) 1.981(1) 2.7673(9) 98.43(14) 97.89(6) 3.053(2) 3G7751 Ni > Cu§ 0.91 6.8421(8) 13.848(2) 2.077(2) 102.94(10) 1.983(1) 2.7676(9) 98.53(14) 97.85(6) 3.053(2) 6Gillardite Ni > Cuᵠ 0.90 6.8364(1) 13.8459(4) 2.0791(8) 102.93(3) 1.9812(4) 2.7665(3) 98.34(5) 97.81(2) 3.049(8) 564041* Mg > Cu 0.60 6.8441(8) 14.025(1) 2.104(3) 103.33(10) 1.988(1) 2.7764(9) 97.96(15) 97.49(6) 3.069(2)

764031 Co > Cu¥ 0.67 6.8436(6) 14.064(1)

The composition (x) corresponds to the formula Cu2.114(3) 103.92(11) 1.983(1) 2.782(1) 97.87(17) 97.67(7) 3.079(2)

4-xMx(OH)6Cl2; (-) not given. 1Average distances with respect to split sites in space group R3�m of the paratacamite substructure from Fleet (1975); 2Paratacamite from Chapter 2.1 at 300 K; 3This study; 4Braithwaite et al. (2004); 5 WAM M365.2003 and 64041 are the Ni and Mg analogues of paratacamite at 296 K and 293 K, respectively, from Chapter 2.2; 6Clissold et al. (2007); 7Chapter 2.3; #True composition must be considered unknown. ¥Also contains a small amount of Mn and trace Ni and Mg. §

60

Also contains trace Co. ᵠContains minor Co and trace Fe. *R3� superstructure is considered to be the true structure..

CH

APTER

2.4

CHAPTER 2.4

61

The paratacamite subcell structure is an average representation of the full structure.

Substructure crystallographic data for paratacamites in previous chapters were refined after

data reduction of the full set of structure factors and appear in Table 2.4.4. Samples

containing Zn2+

There is a small contraction of the M–O bond lengths for both metal sites with

decreasing Zn content. All cis O–M–O angles displayed show a corresponding increase along

the series, the most pronounced being associated with the M(1) position. All trans angles of

the subcell structure are constrained to be 180°. The trends are generally reversed when Ni

as the primary substituting cation show that both the hexagonal a and c axes

decrease towards the monoclinic–rhombohedral transformation boundary, in line with the

observations of powdered material in Jambor et al. (1996).

2+

is the dominant substituting cation. The c axis length increases by ~ 0.1 Å with decreasing Ni

content. Along the same compositional trend the cis O–M–O angle displayed for both M(1)

and M(2) gradually increase, with the most pronounced change existing in the cis O–M(1)–O

angle. For Zn-bearing samples, there is no significant change in the O–Cl bond distance with

changes in composition. The Ni-bearing samples do exhibit some variation in the O–Cl bond

distance with composition, but it is not significant. Data from the paratacamite R3�m subcell

structure are generally consistent with trends observed for strict aristotype structure samples.

The average subcell structure of the Mg-analogue of paratacamite appears consistent

with variation attributed to the difference in ionic radius of the cations. The effective ionic

radius of six-coordinate Mg2+ (0.72 Å) is only marginally smaller than that of Cu2+ and Zn2+

(0.73 Å and 0.74 Å, respectively), but is larger than six-coordinate Ni2+ (0.69 Å) (Shannon,

1976). The Co-rich analogue of herbertsmithite (sample 64031), which has a relatively large

unit cell, would be influenced to some degree by the presence of Mn2+ (0.83 Å) which is

significantly larger than Co2+ (0.745 Å), in a six-coordinate environment (Shannon, 1976).

The corresponding strain tensor of the aristotype unit cell was calculated for samples

listed in Table 2.4.4. The scalar strain of the transformed paratacamite subcell was also

determined. According to the crystallographic data in Table 2.4.4, the paratacamite

substructure offers a good comparison with the aristotype. Therefore, the corresponding unit

cell strain observed for this substructure should also be comparable with the compositional

trends observed for the aristotype. The tensor components for the hexagonal setting can be

determined from the following equations.

CHAPTER 2.4

62

𝑒𝑒11 = 2𝑎𝑎 sin γ𝑎𝑎𝑜𝑜(3)1/2 −1 (1)

𝑒𝑒22 = 𝑏𝑏𝑎𝑎𝑜𝑜

−1 (2)

𝑒𝑒33 = 𝑐𝑐 sin α sin β∗

𝑐𝑐𝑜𝑜 −1 (3)

𝑒𝑒23 = 𝑐𝑐 cos α2𝑐𝑐𝑜𝑜

(4)

𝑒𝑒13 = − 𝑐𝑐 sin α cos β∗

2𝑐𝑐𝑜𝑜 (5)

𝑒𝑒12 = 1𝑎𝑎𝑜𝑜(3)1/2 �𝑎𝑎 cos γ + 𝑏𝑏

2� (6)

Strictly speaking, the equations listed above can be used to determine the strain

component of a lattice distortion in the hexagonal setting transformed from a lower symmetry

structure. Because all structures examined here naturally exhibit a hexagonal unit cell,

a = b ≠ c, α = β = 90° and γ = 120°, calculation of the strain tensor can be simplified

considerably.

𝑒𝑒11 = 𝑒𝑒22 = 𝑎𝑎𝑎𝑎𝑜𝑜

−1 (7)

𝑒𝑒33 = 𝑐𝑐𝑐𝑐𝑜𝑜

−1 (8)

𝑒𝑒23 = 𝑒𝑒13 = 𝑒𝑒12 = 0 (9)

The above equations are derived from Carpenter et al. (1998) and discussed in the

context of this mineral series by Malcherek and Schlüter (2009). The unit cell reported by

Braithwaite et al. (2004) for herbertsmithite was used for reference values in the calculation

giving 𝑎𝑎𝑜𝑜 = 6.834 and 𝑐𝑐𝑜𝑜 = 14.075 Å. A reference unit cell used for gillardite is from

Clissold et al. (2007) with 𝑎𝑎𝑜𝑜 = 6.8364 and 𝑐𝑐𝑜𝑜 = 13.8459 Å. The composition

(Cu3.081Ni0.903Co0.012Fe0.004)(OH)6Cl2 was reported for holotype gillardite and is not ideal as

a reference for the lattice parameters expected for Cu3Ni(OH)6Cl2

The trace amount of lattice Co present in some of the gillardite samples is not

expected to contribute significantly to the unit cell volume. The scalar strain and calculated

tensor components can be found in Table 2.4.5. The distortion of the aristotype unit cell

. However, it does exhibit

the smallest lattice volume and highest substitution of the available gillardites in the literature

and this study. Calculations were made using the unit cell parameters in Table 2.4.4 for the

Zn- and Ni-bearing samples.

CHAPTER 2.4

63

increases towards the monoclinic–rhombohedral transformation zone as the interlayer Cu

content increases.

The strain for both chemical systems is quite small across the entire series, but

increases much more rapidly for Ni-bearing samples. This might be due to the greater

difference in ionic radius between six-coordinate Cu2+ and Ni2+, versus Zn2+. Figure 2.4.1

displays the scalar strain data plot against composition. The subcell of paratacamite

(BM86958) shows the greatest strain of all Zn-bearing samples. The upper compositional

limit proposed for the stability of clinoatacamite, at x ~ 0.33, appears to be a pivotal

composition in terms of the aristotype unit cell strain. Extrapolation of the trend for Zn-

bearing samples indicates that the Zn composition of holotype paratacamite examined by

Fleet (1975), with a scalar strain of 0.0028 associated with the subcell, is between ca

Cu3.70Zn0.30(OH)6Cl2 and Cu3.67Zn0.33(OH)6Cl2

Sample Zn

.

Table 2.4.5. Scalar stain and strain tensor components for the aristotype unit cell

x#

Sample Ni

𝑒𝑒11 𝑒𝑒22 𝑒𝑒33 �∑𝑒𝑒𝑖𝑖𝑖𝑖2

Paratacamite* (-) -0.0010 -0.0010 -0.0024 0.0028 BM86958* 0.29 -0.0014 -0.0014 -0.0032 0.0037 MD166-3 0.35 0.0001 0.0001 -0.0021 0.0021 MM02 0.39 0.0007 0.0007 -0.0016 0.0019 MD166-2 0.64 0.0001 0.0001 -0.0009 0.0009

x 𝑒𝑒11 𝑒𝑒22 𝑒𝑒33 �∑𝑒𝑒𝑖𝑖𝑖𝑖2

CB03 0.39 0.0002 0.0002 0.0065 0.0065 CB07 0.49 0.0007 0.0007 0.0071 0.0071 5WAMM365.2003* 0.71 0.0010 0.0010 0.0064 0.0066 G8502 0.88 0.0006 0.0006 0.0004 0.0009 G8568 0.89 0.0006 0.0006 0.0 0.0009 G7751 0.91 0.0008 0.0008 0.0002 0.0012 *The true unit cell is the paratacamite supercell. #The composition relates to the formula Cu4-xMx(OH)6Cl2

The QE and BAV values for the M(1) octahedra of the R3�m aristotype structure were

calculated for Zn- and Ni-bearing material in this study. The data are displayed in Figure

2.4.2. Both the QE and BAV values for herbertsmithite and gillardite samples show

significant changes that can be related to composition. The single representative QE and

BAV value determined from the paratacamite (BM86958) R3�m subcell structure has the

highest distortion of all samples. This sample has a composition from within the monoclinic–

rhombohedral transition zone. Zoning of Jahn-Teller distorted Cu(OH)

; (-) not known.

6 octahedra would be

CHAPTER 2.4

64

pronounced. This level of distortion may be indicative of the lower limit of compositional

stability for paratacamite. With increasing Zn content, both QE and BAV values decrease to a

minimum for compositions above x ~ 0.6 and are unaffected by increased Zn content.

Similarly, gillardite samples show a significant and reproducible decrease for both QE and

BAV values with excess Ni content. However, the increase in these values appears to be

sharper and occurs at a composition x > 0.7. The R3�m subcell structure of the Ni-analogue of

paratacamite gives comparable QE and BAV values with samples displaying lower Ni-

content.

Figure 2.4.1. The scalar strain of material used in this study. The composition x applies to the formula Cu4-xMx(OH)6Cl2 where M = Zn (blue triangles) or Ni (red squares). Filled markers are samples of paratacamite and open markers are herbertsmithite or gillardite. The dotted lines mark the proposed compositional transformation zone between monoclinic and rhombohedral members determined by Jambor et al. (1996).

Holotype paratacamite of Fleet (1975) has QE and BAV values associated with the

interlayer octahedron of the average subcell structure of 1.053 and 207.64 (degrees2

The difference in trend evolution of QE and BAV values between the Zn- or Ni-

bearing aristotype structure may be attributed to the difference in crystal-chemical behaviour

of these cations. These results show that the distortion exhibited by M(1) varies with changes

in composition in the aristotype structure. It may be inferred that the analogous interlayer

position in the paratacamite superstructure (M1), which is invariant with temperature, varies

with composition. Therefore, it is likely that the Zn- and Ni-bearing samples of paratacamite

),

respectively. Extrapolation of the trends in Figure 2.4.2 indicate a compositional range in

agreement with that suggested from the scalar strain results described above.

CHAPTER 2.4

65

would have a different set of end-members. This would certainly be true of other

paratacamite congeners as well.

Figure 2.4.2. Quadratic elongation (QE) and bond-angle variance (BAV) of M(1) octahedra in herbertsmithite and gillardite (open shapes) and in the paratacamite R3�m substructure (filled shapes). Compositional error bars are smaller than the size of the symbol.

The range of compositions that the R3�m aristotype structure can exhibit spans across

most of the rhombohedral series. The paratacamite samples described with greater than 50%

interlayer occupancy of the substituting cation (samples WAM M365.2003 and 64041)

indicate that the R3� supercell may also exist across much of the substitution series. One must

consider also the multitude of structural refinements for the R3�m aristotype structure with

end-member or near end-member stoichiometry from the literature (Clissold et al., 2007;

Braithwaite et al., 2004; Chu et al., 2011; Chu et al., 2010; Han et al., 2011; Chu 2011;

Wulferding et al., 2010; Schores et al., 2005). This may be taken as an indication that the

aristotype structure is thermodynamically stable near the end-member composition

Cu3M(OH)6Cl2. As the presence of Cu2+ becomes significant in the interlayer the R 3�

CHAPTER 2.4

66

structure may become metastable. Based on the quantifiable distortion of the interlayer

position in the aristotype structure, the substituting cation defines the range of stability (or

metastability) for the phase. This implies that paratacamite and herbertsmithite are part of an

Ostwald series which exists between the compositional end-members for the stability of

paratacamite. However, the end-members for paratacamite are unknown. Within the stability

field of paratacamite, the conversion from the R3�m to R3� structure may be very slow at

ambient temperatures. In addition, the particular conditions which promote the nucleation and

growth of the aristotype structure may serve to inhibit the nucleation and growth of R3�

domains.

CHAPTER 3.1

67

CHAPTER 3 – RAMAN SPECTROSCOPY 3.1 RAMAN SPECTROSCOPY OF NATURAL SINGLE-CRYSTALS

3.1.1 INTRODUCTION

In the previous Chapter the potential series of composition-dependent structural

transformations in the group of basic Cu(II) chlorides were investigated, using single-crystal

X-ray diffraction methods on natural materials. The results from this study indicate that the

aristotype structure of herbertsmithite can exist with compositions approaching the

monoclinic–rhombohedral transition zone. The availability of natural single-crystals, that

exhibit the full range of compositions, imposes limitations on a comprehensive

crystallographic study. Previous characterisations of these phase transformations have been

reported, using powder and single-crystal X-ray diffraction as well as IR spectroscopy

(Jambor et al., 1996; Braithwaite et al., 2004; Malcherek and Schlüter, 2009). However, as

previously discussed, it is not possible to definitively differentiate the PXRD patterns of

anatacamite, paratacamite and herbertsmithite from each other. Additionally, the figured IR

spectra for naturally occurring paratacamite and herbertsmithite in Braithwaite et al. (2004)

appear very similar. Therefore, some of the above studies require validation. Most of the

available data in the literature is based on powdered samples of these minerals. Irrespective

of the above, data pertaining to paratacamite are insufficient to allow an assessment of the

minerals compositional stability field. There is a need for an unambiguous method for

differentiation of the phases, particularly the rhombohedral members, when the only sample

available for analysis is unsuitable for single-crystal work.

Raman analyses of natural and synthetic members of the series are reported in the

literature. Data for anatacamite, clinoatacamite, paratacamite and herbertsmithite can be

found in Frost et al. (2002), Downs (2006), Wulferding et al. (2010), Chu et al. (2011), Liu et

al. (2011), Bertolotti et al. (2012) and de Vries et al. (2012). However, some inconsistencies

exist between the spectra of the same mineral cited by separate references, principally

concerning paratacamite.

As discussed previously, the work of Jambor et al. (1996), Braithwaite et al. (2004)

and Malcherek and Schlüter (2009), suggests the order of chemically-induced

transformations is possibly anatacamite (triclinic, P1�) → clinoatacamite (monoclinic, P21/n)

→ paratacamite (rhombohedral, R 3� ) → herbertsmithite (rhombohedral, R 3� m), with

CHAPTER 3.1

68

increasing substitution of Cu by Zn. This study was made to characterise the solid solution in

terms of Zn and Ni substitution and to elucidate the compositional transition boundaries

between the rhombohedral R3� and R3�m members. Here, new reference spectra using oriented

single-crystals of each mineral in the series are reported and the potential for using Raman

spectroscopy as a non-ambiguous method for differentiation of the phases is discussed. The

results of Raman spectroscopy on polycrystalline synthetic material from the solid solution

series is presented in the next section of this Chapter.

3.1.2 SAMPLES AND METHODS

Single-crystals of anatacamite, clinoatacamite and herbertsmithite were obtained from

specimens housed in the Mineralogical Museum, Hamburg. Anatacamite (specimen MD

199), clinoatacamite (specimen MD 311) and herbertsmithite (specimen MD 166) originated

from the La Vendida mine, the Santa Catalina mine, and the San Francisco mine,

respectively, Sierra Gorda District, Antofagasta Region, Chile. The crystal of paratacamite

examined in Chapter 2.1 was retained in the collections of the Natural History Museum,

London. Therefore new single-crystals were obtained from the same British Museum,

London (BM 86958) type specimen, which originated from the Generosa mine, Sierra Gorda

District, Antofagasta Region, Chile.

The chemical composition of anatacamite, clinoatacamite and herbertsmithite was

determined by electron microprobe analyses after single-crystal X-ray diffraction and Raman

measurements were completed. Measurements were made using a Cameca SX 100

microprobe operated in wavelength dispersive mode. Due to the volatile nature of the

samples in the electron beam, an accelerating voltage of 15 kV and a specimen current of 20

nA was used. Analyses are in Table 3.1.1. The average of 15 analyses for the crystal of

herbertsmithite gave the composition Cu3.46Zn0.64(OH)6Cl2. For clinoatacamite, the average

of 10 analyses gave Cu3.86Zn0.12Ni0.02(OH)6Cl2. The average of 20 analyses for anatacamite

gave the composition Cu3.99Ni0.01(OH)6Cl2. The compositions were normalised to ∑(cations)

= 4.

In order to preserve the crystal of paratacamite, its composition was determined from

the position of the 510 cm-1 Raman mode, which shifts linearly to lower wavenumbers with

increasing Zn content (vide infra, Chapter 3.2). The composition of paratacamite examined

here is estimated to be Cu3.80Zn0.20(OH)6Cl2.

CHAPTER 3.1

69

Table 3.1.1. Electron microprobe analyses of samples in this study. *Average (above), range (below) (wt%) Sample CuO ZnO NiO Cl H2O O≡Cl Total anatacamite 73.75(0.47) - 0.24(0.04) 16.20(0.14) 12.53 -3.66 99.07 72.87–74.77 - 0.13–0.31 15.93–16.44 clinoatacamite 72.21(0.49) 2.30(0.24) 0.35(0.07) 15.29(0.10) 12.50 -3.46 99.20 71.61–73.42 1.91–2.72 0.27–0.47 15.18–15.49 1herbertsmithite 61.42(0.86) 11.93(0.83) - 16.57(0.26) 12.46 -3.74 98.64 59.96–64.91 9.57–13.84 - 16.23–17.34 *Empirical composition from the average data was calculated based on 8 anions pfu: anatacamite, Cu4Ni0.01Cl1.97O6.03H6; clinoatacamite, Cu3.93Zn0.12Ni0.02Cl1.87O6.13H6; herbertsmithite, Cu3.35Zn0.64Cl2.03O5.97H6. The normalised composition is described in the text. 1

3.1.2.1 Single-crystal X-ray diffraction

Herbertsmithite used in Chapter 2.4 (sample MD 166-2).

A single-crystal was attached to the tip of a glass fibre and analysed using X-ray

diffraction methods to determine its identity, phase purity and orientation with respect to the

glass fibre. Analyses were made using a Nonius KappaCCD diffractometer with Mo Kα

radiation. In each case the true unit cell was transformed to a pseudo-hexagonal cell,

approximating the aristotype lattice for the group, a ≈ 6.8, c ≈14.0 Å. This pseudo-cell was

used as the basis for orientation of each crystal for collection of polarised Raman data.

3.1.2.2 Raman spectroscopy

Raman data were collected using a Horiba Jobin-Yvon T64000 triple grating

spectrometer equipped with an Olympus BH41 microscope and a mounted 50x objective.

Spectra were collected in backscattering geometry using an Ar+ laser (514.5 nm) with a

resolution of 2 cm-1

Ideally, the pseudo-hexagonal (aristotype) c axis was oriented perpendicular to the

wavevector of the incident photon k

, at a temperature of 294 K. Measurements were taken with parallel and

crossed polarised light for each orientation of the crystal. The signal-to-noise ratio was

improved by averaging the intensities of 15 acquisitions. Raman polarizability tensors for

each phase analysed are reported in Table 3.1.2.

i, such that the polarisation of the incident light Ei, was

along the c axis direction. Parallel polarised and crossed polarised spectra with the

polarization of the scattered light Es parallel and perpendicular the Ei, respectively, were

measured. The crystal was then rotated by 90° in the plane of the c axis and measurements

were made in both polarisations.

CHAPTER 3.1

70

Table 3.1.2. Raman tensors and active modes for minerals structurally related to paratacamite Anatacamite Clinoatacamite Paratacamite Herbertsmithite Symmetry triclinic monoclinic hexagonal hexagonal Space group P1� P21/n R3� R3�m Point group Ci (1�) C2h (2/m) C3i (3�) D3d (3�m) Ag Ag Ag A1g a d e a d · a · · a · · d b f d b · · a · · a · e f c · · c · · b · · b Bg 1Eg Eg,1 Active · · e c d e c · · Raman · · f d -c f · -c d tensors e f · e f · · d · 2Eg Eg,2

d -c -f · -c -d -c -d e -c · · -f e · -d · · Atom Active Cu 18Ag 3Ag + 3Bg 6Ag + 6Eg Raman Cl 12Ag 3Ag + 3Bg 4Ag + 4Eg A1g+Eg modes O 36Ag 9Ag + 9Bg 12Ag + 12Eg 2A1g + 3Eg H 36Ag 9Ag + 9Bg 12Ag + 12Eg 2A1g + 3E

Using Porto notation for the scattering geometry, A

g Total modes 102 48 68 12

1g modes are allowed to be

observed in parallel polarised 𝑌𝑌’(ZZ)Y’ and 𝑌𝑌’(XX)Y’ spectra and should be absent in crossed

polarised 𝑌𝑌’(XZ)Y’ and 𝑌𝑌’(ZX)Y’ spectra, where Z is along the rhombohedral c axis, X is

perpendicular to c and to ki and Y’ is along ki

Selecting a suitable crystal face for Raman analysis was different between samples

and is described in Table 3.1.3. For herbertsmithite the incident light E

.

i was directed along

the y axis (Y) and for clinoatacamite it was along the x axis (X). The crystal faces on the

anatacamite and paratacamite samples were non-ideal with respect to the hexagonal c axis.

The (111) face of anatacamite, which translates to the (011) plane of the pseudo-hexagonal

cell, was analysed. Therefore, the pseudo-hexagonal c axis was oriented slightly down into

the plane of rotation. The crystal of paratacamite was a well-formed rhomb and only

displayed faces along (hk1) and (hk1) directions. The (011) face of paratacamite was chosen

CHAPTER 3.1

71

and corresponds to the (024) plane in the aristotype unit cell. This directed the hexagonal c

axis about 45° into the plane of rotation.

Table 3.1.3. Porto notation with respect to the aristotype unit cell for samples in this study. Anatacamite clinoatacamite paratacamite herbertsmithite Face analysed* (011) (210) (024) (120) Orientation 1 𝑌𝑌′(AA)Y’ 𝑋𝑋(ZZ)X 𝑌𝑌’(Z’Z’)Y’ 𝑌𝑌(ZZ)Y 𝑌𝑌’(BA)Y’ 𝑋𝑋(Y’Z)X 𝑌𝑌’(XZ’)Y’ 𝑌𝑌(X’Z)Y Rotation 45° 90° 90° 90° Orientation 2 𝑌𝑌’(XX)Y’ 𝑋𝑋(Y’Y’)X 𝑌𝑌’(XX)Y’ 𝑌𝑌(X’X’)Y 𝑌𝑌’(Z’X)Y’ 𝑋𝑋(ZY’)X 𝑌𝑌’(Z’X)Y’ 𝑌𝑌(ZX’)Y Planes X = (016) Y’ = (001) X = (026) X‘ = (001) Z’ = (210) Z = (010) Z’ = (210) Z = (100) A** B** *All planes describe directions in the aristotype unit cell.

3.1.3 RESULTS AND DISCUSSION

**Arbitrary orientation with respect to the aristotype unit cell.

3.1.3.1 Single-crystal X-ray diffraction

The unit cells determined for each phase are reported in Table 3.1.4 and are consistent

with those described in the literature (Fleet, 1975; Malcherek and Schlüter, 2009; Grice et al.,

1996; Braithwaite et al., 2004).

Diffraction patterns of anatacamite, clinoatacamite and paratacamite showed

evidence of twinning. A review of the literature revealed that this is consistent with previous

reports on these materials (Smith, 1905; Frondel, 1950; Malcherek and Schlüter, 2009). A

series of weak superlattice reflections was identified in the diffraction pattern of the

paratacamite single-crystal. However, an analysis of the metrics for the full data set suggested

a deviation towards monoclinic symmetry. Post refinement of 34086 reflections using the

program EVAL15 (Schreurs et al., 2010) resulted in a best fit to a C centred monoclinic unit

cell with dimensions a = 12.2920(9), b = 13.6339(9), c = 9.1263(6) Å and β = 99.597(4)°.

The residual obtained for this setting, of 0.23881, is almost half of the residual determined for

the corresponding hexagonal unit cell of paratacamite with 0.40415 obtained from 34378

reflections. Structural refinement based on the hexagonal cell resulted in convergence to the

paratacamite structure with no better than R1 ~ 0.06. Structure refinement based on the

monoclinic unit cell in space group C2/m was attempted but did not converge adequately to

any realistic model. Transformation to the anatacamite unit cell parameters, a = 9.1257(8),

CHAPTER 3.1

72

b = 9.1757(5), c = 9.1785(6) Å, α = 95.923(3), β = 96.428(3) and γ = 96.390(4)°, and

subsequent structure refinement in P1� based on atom coordinates for anatacamite (Malcherek

and Schlüter, 2009) yielded an improvement to the residuals, but not all H atoms could be

located. The origin of this metric deviation is not entirely clear.

Based on the diffraction pattern obtained, the crystal examined does match the

expected paratacamite pattern. The BM 86958 type specimen contains crystals of both

clinoatacamite and paratacamite (Braithwaite et al., 2004), but no other basic Cu(II) chlorides

have been reported on it. The diffraction patterns of clinoatacamite and paratacamite are

easily distinguishable (Jambor et al., 1996). On this basis, the crystal examined is confirmed

as being paratacamite and it was therefore retained for Raman analysis. The slight deviation

in unit cell metrics may be caused by the excess of Jahn-Teller distorted CuO6 octahedra in

the interlayer sites, inherent for the composition Cu3.80Zn0.20(OH)6Cl2

3.1.3.2 Raman spectroscopy

.

Table 3.1.4. Unit cell parameters for material in this study.* Anatacamite Clinoatacamite Paratacamite Herbertsmithite a (Å) 9.159(4) 6.115(6) 13.6349(7) 6.835(4) b (Å) 9.163(3) 6.822(6) Unit cell* c (Å) 9.167(4) 9.164(1) 14.029(3) 14.046(9) α (°) 96.27(2) β (°) 96.37(2) 99.50(7) γ (°) 96.27(4) *For orientation in the Raman laser, unit cells were transformed from the original parameters to the pseudo-hexagonal setting approximating the herbertsmithite cell.

Raman spectra are displayed in Figure 3.1.1. Several features are apparent from a

comparison of the collected spectra. First, there is an increase in complexity of the detectable

modes from herbertsmithite to anatacamite, which is in agreement with previous reports by

Chu et al. (2011) and Bertolotti et al. (2012). Secondly, the Raman spectrum of paratacamite

displays a resemblance to those of clinoatacamite and anatacamite.

The spectra can be divided into three regions separating modes of low, mid and high

frequency. Table 3.1.5 lists the positions of modes determined from each sample. The Raman

spectrum of paratacamite displays 25 modes and three additional weak and broad signals at

~200, 263 and 740 cm-1. The weak signals in these samples between 200 and 270 cm-1 are

most likely due to crystallographic disorder (Wulferding et al., 2010; Frost et al., 2002).

Mode assignments have been discussed for minerals in this group by Frost et al. (2002),

Wulferding et al. (2010) and Liu et al. (2011).

CHAPTER 3.1

73

CHAPTER 3.1

74

Figure 3.1.1. Raman spectra collected from herbertsmithite, Cu3.46Zn0.64(OH)6Cl2, paratacamite, Cu3.80Zn0.20(OH)6Cl2, clinoatacamite, Cu3.86Zn0.12Ni0.02(OH)6Cl2, anatacamite, Cu3.99Ni0.01(OH)6Cl2. The Porto notation is described in the text. Table 3.1.5. Raman shift for minerals in the Cu4-xMx(OH)6Cl2 substitution series determined from this study.* Anatacamite Clinoatacamite Paratacamite Herbertsmithite 86 w 89 w 96 w 94 w 94 w 111 w 111w 118 s 119 s 118 s 122 s 141 s 144 s 143 s 147 s 168 w 167 w, br 168 w ~ 197 w, br ~ 200 w, br ~ 262 w, br ~ 263 w, br ~ 231 w, br 331w 370 s 365 s 368 s 363 s 419 w 419 s 418 w, br 402 m 446 w 446 w, br 445 w, br, sh 513 s 511 s 510 s 503 s 579 br 580 w, br 574 w, br ~755 w, br ~720 w, br ~740 w, br 705 s 801 br 800 w, br 802 m 870 s 871 w, sh 870 m 896 m 892 m 895 m 931 m 932 s 929 m 950 w, sh 943 s 973 m 971 m 971 m 978 w, br, sh 3309 w, br 3311 w, br, sh 3307 s 3317 m, br, sh 3314 w, sh 3354 s 3350 s 3351 s 3342 w, br, sh 3373 w, sh 3376 w, br, sh 3374 s 3401w, br 3400 m 3402 m 3404 s 3440 s 3441 s, sh 3440 s 3448 w, sh 3448 w, sh 3448 w, br 3511 w, br *s = strong, m = medium, w = weak, br = broad, sh = shoulder.

The original Raman analysis of paratacamite by Burgio and Clark (2001) was

suggested by Frost (2003) to have been made on a sample of clinoatacamite. A comparison of

peak positions for clinoatacamite and paratacamite in this study with those reported by

Burgio and Clark (2001) confirm this suggestion.

CHAPTER 3.1

75

The spectrum for Ni-rich paratacamite in Frost et al. (2002) appears similar to that of

herbertsmithite measured by Wulferding et al. (2010), Chu et al. (2011) and in this study

(Figure 3.1.1). Neither Burgio and Clark (2001) nor Frost et al. (2002) reported

crystallographic data in support of their Raman analysis. The frequency shift of related modes

between Ni-rich paratacamite in Frost et al. (2002) and herbertsmithite could be due to solid

solution effects caused by the difference in crystal-chemical behaviour of Ni2+ and Zn2+

.

3.1.3.3 Metal-anion framework vibrations

For the aristotype structure of herbertsmithite, 12 Raman modes and one weak signal

at ~231 cm-1, which is probably due to crystallographic disorder (Wulferding et al., 2010),

are observed. This corresponds exactly to the number of expected modes

(Table 3.1.2).The strong peaks at ca 120 and 145 cm-1 in each sample and low intensity

modes at ca 90, 110 and 168 cm-1 in paratacamite, clinoatacamite and anatacamite are from

metal-anion framework vibrations. The orientations 𝑌𝑌'(Z’Z’)Y’ of paratacamite and 𝑌𝑌'(AA)Y’

of anatacamite (Figure 3.1.1) exhibit modes with weak intensity at ca 110 and 168 cm-1,

respectively. In herbertsmithite, the peaks at 122 and 147 cm-1 show intensity in parallel

polarised spectra but not in crossed polarised measurements. Both of these correspond to A1g

Several peaks appear between 300 and 500 cm

modes. -1 in the spectrum of paratacamite,

clinoatacamite and anatacamite. Herbertsmithite has only two observable modes in this

region at 393 and 402 cm-1. They display a distinct difference in behaviour in parallel and

crossed polarisation for both orientations. The 393 cm-1 line is attributed to an Eg mode

because it has no intensity contribution in 𝑌𝑌(ZZ)Y orientation. The 402 cm-1 line is assigned

to an A1g mode because its intensity contribution occurs only in 𝑌𝑌 (ZZ)Y and 𝑌𝑌 (XX)Y

orientations. At odds to this analysis, Wulferding et al. (2010) assigned both of these peaks to

Eg modes based on the detection of a four-fold modulation in intensity with rotation of the

crystal in the crystallographic xy plane. Modes in this region were assigned to M–Cl

stretching by Frost et al. (2002) for atacamite and Ni-rich paratacamite (most likely

gillardite). Site symmetry analysis of the phonon modes for the R3m structure (Table 3.1.2)

indicates that Cl in herbertsmithite participates to two Raman-active modes, ΓCl = A1g + Eg,

which is in agreement with this analysis. Anatacamite and paratacamite have more than one

crystallographically independent Cl- ions with several different M–Cl bond lengths.

Clinoatacamite only has one independent Cl- ion which is bonded to three cations with

CHAPTER 3.1

76

different M–Cl bond lengths. These differences are in agreement with the increase in modes

observed between 300 and 450 cm-1 for these lower symmetry structures. Many of the modes

associated with M–Cl stretching in anatacamite, clinoatacamite and paratacamite are not

observed. The peaks at ~450 cm-1 appear with broad bandwidths. Therefore, it is likely that

the unobserved M–Cl modes exist at near equivalent frequencies.

All modes between 500 and 800 cm-1 are attributed to M–O stretching. At 502 cm-1 in

herbertsmithite a strong mode with A1g behaviour is shifted to higher wavenumbers in the

other samples and is accompanied by an additional weak mode at ca 580 cm-1

At 705 cm

. It is likely that

this is due to the additional Raman active O atoms of the lower symmetry structures. -1 a strong mode is apparent in the spectrum of herbertsmithite, but does not

appear in the other samples. In 𝑌𝑌’(XZ)Y’ and 𝑌𝑌’(ZX)Y’ orientations this peak displays Eg

mode behaviour. In direct contrast to this it appears in parallel 𝑌𝑌’(ZZ)Y’ and disappears in

parallel 𝑌𝑌’(XX)Y’ polarisation. This peak was assigned as an A1g mode by Wulferding et al.

(2010). A compositional dependence on this mode with its frequency shift was reported by de

Vries et al. (2012). They indicated that as Cu2+ is replaced by Zn2+, it moves linearly to lower

wavenumbers. Based on their linear regression equation, a composition of

Cu3.31Zn0.69(OH)6Cl2 is calculated which is comparable to the analytical composition

determined here of Cu3.36Zn0.64(OH)6Cl2

The structure of paratacamite and herbertsmithite possess an interlayer octahedron

bonded to six symmetry-equivalent O atoms with a bond length of ~2.11 Å (Fleet, 1975;

Braithwaite et al., 2004; Chapter 2.1). In the aristotype structure the interlayer M(OH)

.

6

octahedra are oriented such that the three-fold axis is tilted 45° from the c axis. The Cl–O

distance and O–Cl–O angle is symmetry-constrained. The distortion of this octahedron

increases in each progressive lower symmetry structure effectively removing the symmetrical

bending and stretching vibrations of the aristotype structure. The H atoms are weakly bonded

to a Cl- ion, which also influences their vibrational freedom. The deviation of the Cl- ion in

each structure from the ideal aristotype position is minimal (Malcherek and Schlüter, 2009).

The greatest deviation occurs with the O atom position. Therefore, the mode at 705 cm-1 in

herbertsmithite is most likely due to symmetrical M–O6 stretching. It was discussed in

Chapter 2.1 and 2.2 that the M(1) octahedron of paratacamite may be composed of a

superimposition of non-tetragonally distorted Zn(OH)6 octahedra with equal proportions of

three orientations of (4+2) Jahn-Teller distorted Cu(OH)6 octahedra, assuming a statistical

distribution of these cations between M(1) at 3b and M(2) at 9d. The composition of

paratacamite in this study, Cu3.80Zn0.20(OH)6Cl2, shows that Cu significantly dominates the

CHAPTER 3.1

77

interlayer. A truly statistical distribution of Zn2+ between the 3b and 9d positions would result

in the composition (Cu0.20Zn0.05)(Cu0.60Zn0.15)Cu3(OH)6Cl2. Therefore, the absence of an

equivalently strong M–O6 stretching mode in the spectrum of paratacamite at ~705 cm-1 may

be due to its limited Zn content. Paratacamite does contain a weak and broad mode at ~740

cm-1, which also appears in the spectrum for anatacamite and clinoatacamite but shifted to

different frequencies (Table 3.1.5). Each crystal examined contains some amount of

substitution of Cu2+ in the interlayer site. Therefore, these weak modes may be due to

removal of the Jahn-Teller distortion on a localized scale as Cu2+ is replaced by Zn2+ or Ni2+.

Confirmation of this might be obtained by comparing the Raman spectra in this study

with the spectrum of the new Ni or Mg analogues of paratacamite, described in Chapter 2.2.

Attempts were made to find another single-crystal of suitable size from both type specimens

(WAM M365.2003 and 64041, respectively) for Raman analysis, but were unsuccessful.

Excess Ni or Mg in the interlayer of these analogues increases the proportion of non-

tetragonally distorted M(OH)6 octahedra which may increase the Raman signal at ~750 cm-1

3.1.3.4 M–O–H deformation

.

Modes between 800 and 1000 cm-1 have been assigned to M–O–H bending by

Liu et al. (2011). The intense signal at 943 cm-1 in herbertsmithite is attributed to an Eg

mode. There is additional complexity of phonon modes in this region for paratacamite,

clinoatacamite and anatacamite. Changes in the orientation and polarisation reveal the

positions of several weak modes in the spectra of these minerals. In clinoatacamite a strong

mode at 932 cm-1

3.1.3.4 O–H stretching

shows asymmetry due to peak overlap (Figure 3.1.1). Anatacamite and

paratacamite both exhibit a mode at this approximate frequency. This mode appears with

significantly reduced intensity and no apparent asymmetry in the Raman spectrum of

stoichiometrically pure clinoatacamite reported by Liu et al. (2011) and Bertolotti et al.

(2012). Therefore, it is likely that modes in this region are influenced by Cu substitution.

The most pronounced difference between each structure in this series is the number of

crystallographically independent O and H atoms and the distortion associated with each

M(OH)6 octahedron. Many of the peaks in the O–H stretching regions show asymmetry

which is likely due to a significant amount of peak overlap. Figure 3.1.2 shows the peak

fitting in the O–H stretching region in order to elucidate the shape and position of modes.

CHAPTER 3.1

78

The intensity scale of Figure 3.1.2 has been modified from the original to better depict

the shape of low intensity modes. There is a high degree of correspondence with peak

position in this region owing to the structural relationships. The two strong signals in

herbertsmithite, Figure 3.1.2 (d), at 3374 and 3404 cm-1 behave as an Eg and A1g

Figure 3.1.2. Peak fitting of the Raman spectra in the O–H stretching region between 3250 and 3500 cm-1 for (a) anatacamite, (b) clinoatacamite, (c) paratacamite and (d) herbertsmithite. The Porto notation is described in the text.

mode,

respectively. Weak signals at these approximate frequencies are identified in the spectrum of

the other members. Bertolotti et al. (2012) and Liu et al. (2011) reported the Raman spectrum

of stoichiometrically pure clinoatacamite with the absence of these modes. Therefore, they

are most likely generated by the reduction of localised Jahn-Teller distortions in the interlayer

by the occupation of Zn. These modes exist in paratacamite but their relative intensity is

smaller, perhaps due to misalignment of the crystal.

3.1.3.5 Phase identification

This Raman analysis suggests that the M–O–H bending and O–H stretching regions

offer the greatest possibility for phase differentiation. Additionally, the increased complexity

CHAPTER 3.1

79

of modes below 100 cm-1 in anatacamite and paratacamite can also be used to differentiate

these phases from clinoatacamite and herbertsmithite. The compositional stability field of

each phase is quite different. No anatacamite crystal has ever been identified with greater

than trace amounts of a substituting cation. The triclinic structure might be unstable with

significant interlayer substitution by Zn2+ or Ni2+ etc. (Malcherek and Schlüter, 2009).

Therefore, if Raman measurements are combined with knowledge of composition of the

sample, accurate phase identification between all members of the substitution series can be

made.

It is likely that prior to Jambor et al. (1996), Braithwaite et al. (2004) and Malcherek

and Schlüter (2009), many analyses of paratacamite may have been made on clinoatacamite,

herbertsmithite, anatacamite or their analogues. In the spectrum of Ni-rich paratacamite

reported by Frost et al. (2002), the presence of a strong mode at 732 cm-1 is consistent with

the aristotype spectral pattern, therefore, their sample was most likely a crystal of gillardite.

Likewise, the Raman spectrum of a synthetic single-crystal of paratacamite measured by Chu

et al. (2011), with the composition Cu3.66Zn0.34(OH)6Cl2, contains a pronounced mode at

~700 cm-1

and a spectroscopic profile consistent with that established for herbertsmithite.

The results obtained here for paratacamite must be validated using additional material.

The crystal could not be ideally oriented for polarisation measurements and the origin of the

metric distortion towards a C centred monoclinic unit cell is not entirely clear. It may be

possible that this metric deviation is related to the low amount of Zn present in the structure.

The crystal could be at the limit of stability of Zn content for trigonal symmetry.

Nevertheless, Raman data obtained for this crystal is distinct when compared to the other

minerals in the series, and suggest that its structure possesses a unique distortion from the

aristotype model.

CHAPTER 3.2

80

3.2 RAMAN SPECTROSCOPY OF THE SYNTHETIC Cu4-xMx(OH)6Cl2

SUBSTITUTION SERIES

3.2.1 INTRODUCTION

Almost all methods for the synthesis of the substituted basic Cu(II) chlorides produces

a micro-polycrystalline (powdered) product, which is a consequence of the nucleation rate

and the conditions that promote accelerated Ostwald ripening. Only recently a reproducible

method for the growth of “large” single-crystals of the substitution series was reported

(Schores et al., 2005; Chu et al., 2011; Han et al., 2011). Consequently, much of the data in

the literature regarding the composition-dependent phase transformations of clinoatacamite,

paratacamite and herbertsmithite was derived from powdered samples (Jambor et al., 1996;

Braithwaite et al., 2004). The limitations of PXRD for differentiation of the rhombohedral

members of the group are discussed in Chapters 1 and 2. Phase identification using single-

crystals can be conveniently made by X-ray diffraction methods but the synthetic method

mentioned above reportedly takes an excess of 10 months before crystals of sufficient size

have formed (Han et al., 2011). Therefore the production of samples exhibiting the full range

of compositions through a trial-and-error basis is impractical. The reference spectra reported

in Chapter 3.1 for anatacamite, clinoatacamite, paratacamite and herbertsmithite may provide

the means for accurate phase identification from powdered samples.

Braithwaite et al. (2004) suggested that the transformation from paratacamite (R3�) to

herbertsmithite (R 3� m) takes place near the composition Cu3.50Zn0.50(OH)6Cl2

To identify the subtle structural changes leading towards the proposed paratacamite

(R3�) and herbertsmithite (R3�m) phase transformation, polycrystalline samples displaying a

broad range of Cu-Ni and Cu-Zn compositions were analysed by Raman spectroscopy.

. Results

obtained in Chapter 2 on natural crystals from the substitution series suggest that under

ambient conditions the R3�m structure is persistent down to a composition approaching the

monoclinic–rhombohedral transition zone determined by Jambor et al. (1996).

3.2.2 SAMPLES AND METHODS

Synthetic powdered members of the Zn and Ni series of Cu4-xMx(OH)6Cl2, were

prepared after a modification of the method of Jambor et al. (1996). Solutions were prepared

in a round bottomed flask with the appropriate metal ratio, while maintaining the

concentration of CuCl2 at 0.02 M in 100 cm3 of MilliQ water. The composition of the

CHAPTER 3.2

81

product obtained was adjusted on a trial-and-error basis by adjustments to the solution metal

ratio. The solution was heated under reflux and an eqimolar amount of standardised 0.050 M

NaOH was added. The solution was refluxed for two to five days in order to improve

crystallinity and, in some cases, to ensure that a transient atacamite impurity had

decomposed. The cooled product was collected by suction filtration through Whatman GF/F

fibre glass filter paper, washed with MilliQ water and acetone, and dried at the pump.

3.2.2.1 Powder X-ray diffraction

Samples were identified by PXRD methods with either a Phillips PW1825-20 powder

diffractometer (CuKα radiation, λ = 1.5406 Å, 40 kV and 30 mA), or a Bruker D8 Advance

powder diffractometer (CuKα radiation, λ = 1.5406 Å, 40 kV and 40 mA). Pure Si was used

as internal standard. Patterns were collected between 5–70° 2θ using a step size of 0.02° and

1 step sec-1.

The X-ray diffraction pattern of clinoatacamite was differentiated from the aristotype

pattern by identification of characteristic peak splitting, such as the Irel = 60, 2.266 Å (220),

and Irel

3.2.2.2 Composition

= 50, 2.243 Å (004) reflections. Diffraction patterns consistent with the paratacamite

substructure (powder diffraction file 50-1558), were indexed based on the herbertsmithite

unit cell (a ~ 6.8 and c ~14.1 Å) for Zn-substituted members, or the gillardite unit cell (a ~

6.8 and c ~ 13.9 Å) for Ni-substituted members.

The program PowderCell (Kraus and Nolze, 1996a, b) was used to index the two sets

of samples and unit cell parameters were refined using the least squares software LAPOD

(Langford, 1973).

The solid state metal ratio in each sample was determined by atomic absorption

spectrophotometry (AAS), after dissolution of some of the solid in a small quantity of

aqueous HNO3

3.2.2.3 Raman spectroscopy

. All analyses were made using a GBC 1000 spectrophotometer (air-acetylene)

with appropriate standards for Cu, Zn, and Ni. The composition was calculated based on four

cations pfu (Table 3.2.1).

Unpolarised Raman data were collected from the polycrystalline samples using the

same experimental conditions and instrumentation described in the previous section of this

Chapter.

CHAPTER 3.2

82

Table 3.2.1. Composition of powdered samples examined.

Composition Solution (ppm) Solid phase Solution (ppm) Solid phase Sample Cu Zn x(Cu) x(Zn) Cu Ni x(Cu) x(Ni) Monoclinic samples 1 71.40 0.65 3.96 0.04 71.20 0.08 4.00 0.00 2 87.80 3.00 3.87 0.13 60.40 0.66 3.95 0.05 3 87.40 4.60 3.81 0.19 67.36 1.62 3.90 0.10 Rhombohedral samples 4 55.72 2.88 3.81 0.19 52.00 3.83 3.70 0.30 5 66.00 4.60 3.75 0.25 67.78 5.57 3.67 0.33 6 59.00 4.40 3.73 0.27 60.38 6.10 3.61 0.39 7 55.60 5.00 3.68 0.32 32.46 4.13 3.52 0.48 8 50.60 5.20 3.64 0.36 49.90 6.84 3.48 0.52 9 48.20 5.00 3.63 0.37 53.14 7.88 3.45 0.55 10 60.60 6.60 3.62 0.38 47.18 7.12 3.44 0.56 11 49.00 5.60 3.60 0.40 62.98 10.00 3.41 0.59 12 56.40 6.86 3.58 0.42 81.80 14.80 3.34 0.66 13 46.70 5.90 3.56 0.44 47.40 9.40 3.29 0.71 14 48.60 6.86 3.52 0.48 66.28 14.58 3.23 0.77 15 51.20 7.80 3.48 0.52 68.60 17.40 3.14 0.86 16 44.60 9.00 3.34 0.66 47.6 13.2 3.08 0.92 17 43.80 12.90 3.11 0.89

3.2.3 RESULTS AND DISCUSSION

3.2.3.1 Powder X-ray diffraction

In all cases, the PXRD pattern revealed a phase pure product consistent with either the

paratacamite substructure or clinoatacamite. Tables 3.2.2 and 3.2.3 display the unit cell

parameters and compositions for material used in this study. Phase differentiation of samples

within the monoclinic–rhombohedral transition zone by PXRD methods was not ideal. Peak

broadening exhibited in some samples masked the unambiguous identification of the

characteristic clinoatacamite diffraction lines. The crystallinity of each sample had a large

influence on the identification of the characteristic (220) and (004) Bragg peaks of

clinoatacamite. To avoid the analysis of potential mixtures of the phases, only those samples

where unambiguous differentiation could be made were retained for Raman analysis.

The first phase to exist in both the Ni and Zn series, produced by the synthetic

method described, was confirmed as clinoatacamite. The triclinic structure of anatacamite

was suggested by Malcherek and Schlüter (2009) to be stable with only minimal amounts of

Cu substitution and that it should represent the stable compositional end-member,

CHAPTER 3.2

83

Cu2(OH)3Cl, for the series. The calculated powder pattern for anatacamite shows that its

characteristic superstructure reflections are too weak to be observed in a standard PXRD

experiment. The observable peak profile appears more consistent with that of the aristotype

structure.

Phase identification between clinoatacamite and anatacamite samples using PXRD

should involve the same approach as differentiation of monoclinic and rhombohedral

members, described above (subsection 3.2.2.1). It may be the case that clinoatacamite with

composition Cu2(OH)3Cl, is the thermodynamically stable phase at elevated temperatures, as

suggested by Malcherek and Schlüter (2009). Therefore, the identification of clinoatacamite

in all samples with minimal substitution is consistent with its synthesis at elevated

temperatures.

Table 3.2.2. Monoclinic unit cell parameters for synthetic Cu4-xMx(OH)6Cl2, for M = Zn and Ni, identified as clinoatacamite. a (Å) b (Å) c (Å) β (°) x(Zn) 0.04 6.158(3) 6.813(2) 9.123(5) 99.64(4) 0.13 6.162(3) 6.811(2) 9.117(5) 99.54(3) 0.19 6.152(3) 6.814(3) 9.134(7) 99.66(5) x(Ni) 0.00 6.162(2) 6.820(2) 9.119(4) 99.67(3) 0.05 6.155(5) 6.809(4) 9.118(5) 99.61(4) 0.10 6.139(8) 6.812(8) 9.139(6) 99.64(6) Table 3.2.3. Hexagonal unit cell parameters of synthetic Cu4-xMx(OH)6Cl2, for M = Zn and Ni, identified as the paratacamite substructure. x(Zn) a (Å) c (Å) x(Ni) a (Å) c (Å) 0.19 6.833(2) 14.054(7) 0.30 6.832(3) 13.998(8) 0.25 6.838(1) 14.052(3) 0.33 6.839(2) 13.964(9) 0.27 6.832(2) 14.046(6) 0.39 6.831(2) 13.940(1) 0.32 6.837(2) 14.049(4) 0.48 6.840(2) 13.926(7) 0.36 6.835(2) 14.046(6) 0.52 6.837(2) 13.951(9) 0.37 6.837(2) 14.044(4) 0.55 6.840(2) 13.939(7) 0.38 6.839(1) 14.054(5) 0.56 6.838(3) 13.942(8) 0.40 6.837(1) 14.047(3) 0.59 6.837(2) 13.922(7) 0.42 6.835(2) 14.037(4) 0.66 6.843(2) 13.908(6) 0.44 6.837(2) 14.049(7) 0.71 6.842(2) 13.913(5) 0.48 6.842(1) 14.045(3) 0.77 6.844(1) 13.917(5) 0.52 6.834(1) 14.060(2) 0.86 6.838(2) 13.918(8) 0.66 6.8403(9) 14.077(2) 0.92 6.842(3) 13.918(9) 0.89 6.842(3) 14.067(6)

CHAPTER 3.2

84

3.2.3.2 Raman spectroscopy

The range of compositions examined by Raman spectroscopy was designed to

identify composition related structural changes leading up to the phase transformation

between the R3� and R3�m structures. Figures 3.2.1 and 3.2.2 show Raman spectra of the

synthetic powdered samples for the Zn and Ni series, respectively. The spectrum of samples

with x < 0.20 is in agreement with the reference spectrum for clinoatacamite. Samples with

x ˃ 0.50 d isplays a spectroscopic profile consistent with that of the herbertsmithite single-

crystal measured, as well as the spectra reported by Frost et al. (2002) and Wulferding et al.

(2010). Between these compositions, the intensity ratios appear significantly different. Modes

within this range generally match the frequency positions of the reference clinoatacamite and

herbertsmithite single-crystals. This is most evident in the M–O–H deformation and O–H

stretching regions.

For rhombohedral Zn-bearing samples measured with compositions

x < 0.36, the intensity ratios between modes in the O–H stretching region approach those of

the clinoatacamite samples (x = 0.04, 0.13, 0.19 of Figure 3.2.1). In addition, the

characteristic mode of the herbertsmithite reference spectrum, at ~715 cm-1 appears in all

powdered samples identified a hexagonal unit cell by PXRD. In the Ni substitution series,

this mode appears at higher frequencies (730–740 cm-1) and is consistent with the equivalent

mode identified in Ni-rich paratacamite at 732 cm-1

3.2.3.3 Metal–anion framework vibrations

by Frost et al. (2002).

Figure 3.2.3 displays peak fitting of modes between 250 and 650 cm-1 in a sample of

polycrystalline clinoatacamite. These modes are most likely derived from Cu–Cl and

Cu–O stretching. As lattice Zn or Ni content increases the mode at ~360 cm-1 increases in

intensity. Both of the modes at ~360 cm-1 and ~510 cm-1 shift to a different frequency with

changes in composition (Figure 3.2.4).

The mode at ~510 cm-1 shifts linearly with an increase in the Zn content up to a

composition of x ~ 0.66. The position of the spurious data point at x = 0.19 in the Zn-series of

Figure 3.2.4 may have been influenced by structural defects, such as polytypic features,

giving rise to intense H bonding (vide infra).

CHAPTER 3.2

85

Figure 3.2.1. Raman spectra between 50–1100 cm-1 and 2850–3600 cm-1 for synthetically prepared samples of Cu4-xZnx(OH)6Cl2

Figure 3.2.2. Raman spectra between 50–1100 cm

. The value of (x) is presented next to the spectrum. Spectrum of note marked by an asterisk (*).

-1 and 2850–3600 cm-1 for synthetically prepared samples of Cu4-xNix(OH)6Cl2. The value of (x) is presented next to the spectrum. Spectrum of note marked by an asterisk (*).

*

*

CHAPTER 3.2

86

Figure 3.2.3. Peak fitting in the spectral range between 250–650 cm-1 for clinoatacamite of composition Cu3.96Zn0.04(OH)6Cl2

Figure 3.2.4. A comparison of data for samples between the spectroscopic range between 250 and 650 cm

.

-1

For the compositional range examined, excluding the low frequency data point at x =

0.19 and the data point at x = 0.89, a linear regression equation was determined as �̅�𝑣 R1

. Filled symbols are samples of clinoatacamite and open symbols are samples exhibiting the paratacamite substructure, as identified by PXRD.

=

513.22-16.67x with R2 = 0.95, where x is the content of Zn in the formula Cu4-xZnx(OH)6Cl2.

Using this method a composition of Cu3.80Zn0.20(OH)6Cl2 was estimated for the single-

crystal of paratacamite examined in Chapter 3.1. The analogous phonon mode in the single-

crystal of herbertsmithite (Chapter 3.1) exists at 503 cm-1 and results in a calculated

composition of Cu3.49Zn0.61(OH)6Cl2. In the clinoatacamite single-crystal, the mode is

CHAPTER 3.2

87

shifted to 511 cm-1 and calculates to a composition of Cu3.86Zn0.13(OH)6Cl2. The small

amounts of Ni detected by the microprobe in clinoatacamite and anatacamite are expected to

have little influence on the position of this mode. These calculated compositions are all in

agreement with the microprobe results (Chapter 3.1, Table 3.1.1). Although no Zn was

detected in the sample of anatacamite the mode at 513 cm-1

A linear fit to data from the analogous mode in the Ni series (Figure 3.2.4) can also be

made but its accuracy is reduced. The equation �̅�𝑣 R 1

indicates x = 0.01, which is also

in agreement with the amount of substitution determined by the microprobe results. Previous

microprobe analyses of material from the Generosa mine type specimen of paratacamite have

indicated that Zn is the only significant substituting cation (Kracher and Pertlik, 1983;

Jambor et al., 1996; Braithwaite et al., 2004). Therefore, the accuracy of this method for the

composition determination of the sample of paratacamite is validated. However, it must be

noted that larger quantities of lattice Ni or another substituting cation could have a significant

effect on the behaviour of this mode.

= 512.45 -9.3692x, R2 = 0.88, was

determined from the full range of data displayed in Figure 3.2.4 for the ~510 cm-1 mode. The

�̅�𝑣 R1 intercept at ~513 cm-1 is consistent from both data sets for a product with no substitution

for Cu. The difference in slope exemplifies the contrasting effect that the substituting cation

has on the propagation of this mode.

Both the Zn and Ni series show the loss of the ~446 cm-1 mode in the clinoatacamite

samples by a composition of ca Cu3.67M0.33(OH)6Cl2. In the Ni series, there is a sharp

increase in the mode bandwidth (FWHM data) of the ~510 and 368 cm-1

3.2.3.4 M–O–H deformation and O–H stretching regions

modes over the

monoclinic to rhombohedral transformation boundary (0.20 < x < 0.33). This is followed by a

sudden decrease in mode bandwidth with greater substitution. A sharp drop in the bandwidth

occurs in the Zn series at x ~ 0.50 which is followed by a sequential increase in bandwidth

values of samples with higher Zn contents.

The strongest mode in the O–H stretching region of the herbertsmithite reference

spectrum at ~3400 cm-1 appears in the clinoatacamite powdered samples with a composition

x > 0.05 (Figures 3.2.1 and 3.2.2). This mode disappears from the clinoatacamite spectrum as

the composition approaches that of the Cu2(OH)3Cl end-member. This is in line with the

observations of Bertolotti et al. (2012) and Liu et al. (2011) on stoichiometrically pure

clinoatacamite. At this nominal composition, the spectrum of clinoatacamite is very similar to

that of anatacamite (Bertolotti et al., 2012; Chapter 3.1). For samples after the monoclinic to

CHAPTER 3.2

88

rhombohedral structural transformation, the intensities of the ~3370 and 3400 cm-1 modes

increase significantly. Simultaneously, a decrease in intensity of the modes of the

clinoatacamite spectrum at ~3310, 3350 and 3440 cm-1 occurs. In addition, the mode at

~940 cm-1 in the Zn series suddenly increases in intensity at composition at x = 0.52. The Ni

series exhibits a significant increase in the intensity of this mode at x ~ 0.50.

Mode deconvolutions of representative spectra in the O–H stretching region are

displayed in Figure 3.2.5. The three different spectroscopic profiles from the Ni and Zn series

are represented and show particular differences. In Figures 3.2.5 (a) and (d), the modes at

~3340 and 3450 cm-1 appear more intense in the Zn series than in the Ni series. The intense

modes at ~3370 and 3400 cm-1 are suppressed with a significant amount of Cu occupying the

structure, as represented in Figures 3.2.5 (b) and (e). The modes at 3362 and 3394 cm-1 of the

Ni series (Figure 3.2.5 a), which correspond to the strongest O–H stretching modes of the

reference herbertsmithite spectrum, are shifted by about 10 cm-1 to lower frequencies with

respect to their positions in the Zn samples. Consequently, the intense mode at 3354 cm1 in

clinoatacamite of Figure 3.2.5 (c) masks the true position and intensity of the weak emergent

mode at ~3360 cm-1

Figure 3.2.5. Raman peak fitting in the O–H stretching region between 3250–3500 cm

until the composition x ~ 0.50.

-1 for synthetic Ni- and Zn-bearing material of composition (a) Cu3.23Ni0.77(OH)6Cl2, (b) Cu3.61Ni0.39(OH)6Cl2, (c) Cu3.90Ni0.10(OH)6Cl2, (d) Cu3.34Zn0.66(OH)6Cl2, (e) Cu3.65Zn0.35(OH)6Cl2 and (f) Cu3.81Zn0.19(OH)6Cl2. Samples (a), (b), (d), and (e) give an X-ray powder pattern consistent with the paratacamite substructure. Samples (c) and (f) give an X-ray powder pattern matching clinoatacamite.

CHAPTER 3.2

89

It was suggested in Chapter 2 that the incorporation of Zn or Ni in these phases would

remove local interlayer Jahn-Teller distortions in favour of a non-tetragonally elongated

coordination sphere. This can be seen in Figure 3.2.6 which displays the ratio between the

sum of intensities of the two primary rhombohedral O–H stretching modes (~3370 and

3400 cm-1) and the combined set of intensities in the O–H stretching region (3200 to

3600 cm-1

) against the composition.

Figure 3.2.6. Integrated intensity ratio between the sum of primary O–H stretching modes of herbertsmithite or gillardite and the total sum of intensities of the O–H stretching region between 3300 and 3500 cm-1 (𝐼𝐼~3370 𝑐𝑐𝑚𝑚−1 + 𝐼𝐼~3400 𝑐𝑐𝑚𝑚−1 )/∑ 𝐼𝐼(𝑣𝑣�O−H ) where �̅�𝑣O−H are the observable O–H stretching modes. Filled symbols are samples of clinoatacamite and open symbols are samples exhibiting the paratacamite substructure, identified by PXRD.

The ratio is at its minimum when no substitution of Cu has taken place in

clinoatacamite and the modes at ~3370 and 3400 cm-1 do not exist. As Zn or Ni substitution

takes place and these modes become apparent the integrated area ratio increases. For the Zn

series there is a relatively linear relationship between composition and the intensity ratio of

these modes up to x ~ 0.70. To reiterate, the outlier at x = 0.19 may have been influenced by

strong H bonding effects (vide infra).

The Ni series exhibits a similar trend which increases at a reduced rate above

x ~ 0.50. As described above the intense mode at ~3355 cm-1 in clinoatacamite masks the true

position and intensity of the weak gillardite mode ~3360 cm-1 until the composition x ~ 0.50.

For this reason the ~3360 cm-1 mode could not be modelled for compositions below x ~ 0.50

and consequently resulted in a decrease in the intensity ratio of rhombohedral samples with

0.30 < x < 0.45 in Figure 3.2.6.

CHAPTER 3.2

90

Figure 3.2.7. Composition-induced changes in spectroscopic mode frequency and intensity for Zn-bearing members of Cu4-xZnx(OH)6Cl2. Filled symbols are samples of clinoatacamite and open symbols are samples exhibiting the paratacamite substructure, identified by PXRD.

CHAPTER 3.2

91

Figure 3.2.8. Composition-induced changes in spectroscopic mode frequency and intensity for Ni-bearing members of Cu4-xNix(OH)6Cl2

Figures 3.2.7 (1a) and (1b) show the ~3310 and ~3320 cm

. Filled symbols are samples of clinoatacamite and open symbols are samples exhibiting the paratacamite substructure, identified by PXRD.

-1 modes shift to higher

frequencies and decrease in intensity with increasing Zn content. A corresponding trend is

observed for the modes at ~3440 and ~3450 cm-1 in Figures 3.2.7 (4a) and (4b). By a

composition of x ~ 0.66 the spectrum cannot be modelled with two modes in this region. This

is because of mode coalescence as they form one broad peak at an intermediate frequency.

The gradual decrease in intensity of the mode at ~3310 cm-1 results in its complete

suppression at compositions with x > 0.50. A similar pattern of mode evolution is seen with

the Ni substitution series in Figure 3.2.8. Samples with composition x > 0.50 indicate a

suppression of the mode at ~3310 cm-1 and complete loss of this feature by a composition of

CHAPTER 3.2

92

x ~ 0.60. At x > 0.50, the mode at ~3360 cm-1, seen in Figures 3.2.8 (2a) and (2b), is no

longer significantly affected by additional Ni substitution. The most intense band in this

region at ~3394 cm-1 undergoes a sudden increase in frequency and decrease in intensity for

samples with x > 0.80, seen in Figures 3.2.8 (3a) and (3b). This is also observed in the Zn

series for the peak at ~3400 cm-1 in Figures 3.2.7 (2a), (2b), (3a) and (3b).

The propagation of peak bandwidths for the primary aristotype O–H stretching modes

at ~3370 and 3400 cm-1 is different between Zn- and Ni-bearing samples (Figure 3.2.9). A

sharp increase in the bandwidth occurs for the ~3370 cm-1 mode in the Ni series over the

monoclinic–rhombohedral transition. To reiterate, this Raman signal in the Ni series is the

contribution of two pronounced modes which occur at near equivalent frequencies. The

FWHM data for the ~3370 cm-1 signal therefore shows the composition when contribution

from each signal has reached its maximum. This feature is not observed in the analogous

mode of the Zn series which increases in bandwidth until x ~ 0.50. The trend observed in the

~3400 cm-1 occurs with the opposite direction between Zn and Ni samples. In both sets of

samples, the FWHM data for the ~3370 cm-1

Figure 3.2.9. Bandwidths (FWHM) of the O–H stretching modes at ~3370 and at 3400 cm

mode displays a significant change in trend for

x > 0.50. The change in FWHM data for samples with x > 0.50 is minimal.

-1. Filled symbols are samples of clinoatacamite and open symbols are samples exhibiting the paratacamite substructure, identified by PXRD.

CHAPTER 3.2

93

Many of these effects may be explained by the response of substitution of the

interlayer Jahn-Teller distorted Cu2+ octahedra with the Zn2+ ion versus the Ni2+ ion. The

occupation of excess Zn2+ in the interlayer (x > 0.50) has a greater effect on the propagation

of Raman phonon modes than the corresponding compositions with excess Ni2+. This may be

a consequence of chemical-induced local pressure on the coordination environment by the

incorporation of a larger cation, Zn2+ 0.74 Å versus Cu2+ 0.73 Å and Ni2+

3.2.3.5 H bonding

0.69 Å, in a six-

coordinate environment (Shannon, 1976)

Several samples studied by Raman spectroscopy show the presence of two moderately

intense peaks at ~2905 and 2965 cm-1 (Figures 3.2.1 and 3.2.2). The samples were

remeasured at a higher wavelength (785 nm) to eliminate fluorescence as a possible origin of

these peaks. This signal appeared in each measurement and is therefore not due to

fluorescence. They appear in four synthetic Zn-bearing samples (x ~ 0.2, ~ 0.33, ~ 0.4 and

~ 0.6) and only strongly in one Ni-substituted sample (x ~ 0.33). They are present as broad,

low intensity peaks in most of the Ni series when x < 0.6. Above this composition they have

not been observed. Interestingly, these low intensity peaks were identified from a natural

single-crystal of herbertsmithite from the San Francisco mine with a composition of

Cu3.65Zn0.35(OH)6Cl2, and which was used for test measurements. In this data set they were

slightly shifted to lower frequencies compared with the synthetic samples. However, when

the crystal was rotated these peaks were no longer detectable. Subsequent readjustments of

the crystal failed to reproduce the results. Peaks in this region have been detected in

anatacamite (RRUFF-ID R100198) and intensely in claringbullite (RRUFF-ID R110007)

(Downs, 2006). It is likely that they are due to strong H bonding. An analysis of the effect of

composition on the structure of herbertsmithite indicates that an increase in the Zn content is

followed by a slight decrease in the O–Cl distance, from 3.073 to 3.071 Å for compositions x

= 0.35 to 1 (Chapter 2.4). The change in O–Cl distance in gillardite with increasing Ni

content is more pronounced with values ranging from 3.060 to 3.049 Å, for x = 0.39 to 0.90

(Chapter 2.4). Each H atom is directed towards the nearest Cl-

There is a pronounced structural relationship between claringbullite, herbertsmithite

and gillardite. The structure of claringbullite is depicted in Figure 3.2.10. Claringbullite

ion. This decrease in O–Cl

bond length could potentially reduce the H–Cl bond distance. Therefore, because the

additional modes are only observed in samples with dominant interlayer Cu, it is more likely

that they are caused by another structural phenomenon.

CHAPTER 3.2

94

possesses sheets of Jahn-Teller distorted Cu32+(OH)6Cl2 octahedra (M1) arranged in a

kagomé lattice (Burns et al., 1995) similar to those found in herbertsmithite and gillardite.

While in herbertsmithite and gillardite subsequent sheet layers are rotated by 60° giving an

ABABA arrangement, the layers in claringbullite are stacked without rotation. This causes the

interlayer Cu(OH)6 coordination to adopt a distorted trigonal prismatic environment. In

addition, the Cu of the interlayer site is disordered over three positions. The interlayer also

contains an independent Cl(2) site which does not bond to a metal cation; rather it accepts six

H bonds from nearest neighbour OH- groups. In addition, the Cl(2) site can exhibit a mixed

occupancy of Cl- and OH-

Figure 3.2.10. A polyhedral representation of the structure of claringbullite. O atoms are red spheres and H atoms small white spheres.

. The corresponding O–Cl(2) distance is 2.747(2) Å, resulting in a

short H–Cl(2) distance of 1.81(4) Å (Burns et al., 1995).

Claringbullite is thermodynamically unstable at ambient temperatures and may

recrystallise ultimately to clinoatacamite under certain conditions (Pollard et al., 1990). In the

presence of Zn2+ ions, claringbullite recrystallises to herbertsmithite (Schores et al., 2005).

The corresponding basal spacings in claringbullite, herbertsmithite and gillardite are 4.592,

4.692 and 4.615 Å, respectively. It may be possible that some claringbullite-type layers occur

in certain crystals of herbertsmithite and gillardite and these may be responsible for the

generation of the intense H bonding Raman peaks. This type of polytypism could be

dependent on composition and recrystallisation conditions. Frost (2003) reported the Raman

CHAPTER 3.2

95

spectrum of claringbullite with the absence of these H bonding peaks. However, the Cl-/OH-

composition of the Cl(2) site was not reported. The ratio of Cl- and OH- may be a contributing

factor to the strength of H bonding that occurs, as OH- is a better acceptor of H bonds than Cl-

(Lutz et al., 1994). The Raman spectrum of claringbullite displays similar mode frequencies

with clinoatacamite and anatacamite reported here. An intense mode at 511 cm-1 was

suggested by Frost (2003) to be due to Cu–Cl stretching. In the samples showing strong peaks

at ~2905 and 2965 cm-1 some additional mode overlap is apparent particularly for the ~510

cm-1 mode which displays asymmetry when the proposed H bonding peaks are most intense

(x = 0.19 of Figure 3.2.1). This possibility may be validated by an investigation of the Raman

spectrum of claringbullite with variable Cl-/OH-

3.2.3.6 The transformation series

in the Cl(2) site.

The spectroscopic profile of synthetic polycrystalline samples examined in this study

corresponds to mode positions established from clinoatacamite and herbertsmithite single-

crystals in the previous section of this Chapter. The evolution of modes over the monoclinic–

rhombohedral transition zone combined with the appearance of the characteristic

herbertsmithite mode at ~710 cm-1, indicates that the transformation series is P21/n → R3�m.

Changes observed in mode position, bandwidth and intensity are most likely due to structural

changes on a local or mesoscopic scale, related to the reduction of interlayer Jahn-Teller

distortions by the substitution of Cu in these structures. The absence of paratacamite across

the series of compositions examined is unexpected. It was established in Chapter 2.1 that the

paratacamite structure is thermodynamically stable at ambient temperatures for the

composition Cu3.71Zn0.29(OH)6Cl2. Ni-bearing paratacamite described in Chapter 2.2 has the

composition Cu3(Ni0.71Cu0.27Co0.02)(OH)6Cl2

Two possibilities present themselves. First, all samples in this study were

synthesised at elevated temperatures under similar solution conditions which may serve to

kinetically stabilise the R3�m structure. Transformation from R3�m to R3� may be very slow

within the compositional stability field of paratacamite. However, the reversible

transformation from the R 3� to R 3� m structure, while certainly temperature related, is

apparently fast (Chapter 2.1). Alternatively, the paratacamite reference crystal examined in

Chapter 3.1 of Cu

, which indicates that its stability field may be

quite different to the Zn-bearing congener.

3.80Zn0.20(OH)6Cl2, is most likely very close to the lower limit of Zn

content for stability of the rhombohedral structure. The apparent deviation in the unit cell

metric towards a lower symmetry cell would suggest that significant distortions induced by

CHAPTER 3.2

96

excess interlayer Cu are present. This may explain why its Raman spectrum appears similar

to that of anatacamite. The M–O–H deformation and O–H stretching regions of the

paratacamite reference crystal examined in Chapter 3.1 suggest that paratacamite distorts

towards the anatacamite (P1�) structure, rather than to clinoatacamite (P21/n). It is clear that

substitution effects play a significant role in the appearance of the Raman spectrum of these

minerals. For this reason, the presence of excess Zn, Ni or other cations in the interlayer

position of paratacamite may have a profound influence on its spectroscopic profile. The

spectrum of stoichiometrically pure clinoatacamite is also similar to that of anatacamite.

Therefore, these results suggest that two lines of composition-induced structural

transformations occur in the basic Cu(II) chloride minerals, one involving clinoatacamite as

P1� → P21

/n → R3�m and the other involving paratacamite as P1�→ R3� → R3�m, both with

increasing substitution of interlayer Cu. This suggestion is in line with the group theory

discussion in Malcherek and Schlüter (2009). Both transformation series may well be

influenced by kinetic effects.

CHAPTER 4.1

97

CHAPTER 4 – SUBSTITUTION AND ACTIVITY

4.1 THERMODYNAMICS OF SUBSTITUTION IN CLINOATACAMITE

4.1.1 INTRODUCTION

The transformation from clinoatacamite to herbertsmithite or gillardite (Chapter 3.2)

involves end-members whose compositions are difficult to quantify. Their compositions will

depend on such factors as the nature of the substituting cation, solution pH and temperature.

The end-members of these minerals represent pure phases by definition. The

substitution of a cation in a solid substance may therefore be considered as the mixing of

various amounts of one end-member with the other. In terms of activity and concentration

relationships, a phase exhibiting a composition between the end-members is no longer

considered as being pure and can be treated as a solid solution, where the activity of

dissolved components can be modelled (Garrels and Christ, 1965). This allows an insight into

the behaviour of cation mixing in the solid state and may elucidate aspects of the

thermodynamics of this system.

In the case examined here, substitution of Cu2+ by other ions into clinoatacamite leads

to instability of the monoclinic structure in favour of a rhombohedral phase. Substitution

phenomena in the basic Cu(II) chlorides is explored for synthetic members from both the Zn

and Ni substitution series. The effect of substitution on the stability of clinoatacamite has

been assessed and discussed. Since the lower compositional end-member for herbertsmithite

or gillardite has not been defined it would only be possible to determine a value for the Gibbs

free energy of formation of the pure phase Cu3M(OH)6Cl2. With the synthetic conditions

employed in this study, the formation of this end-member requires in excess of 4 M ZnCl2 or

NiCl2

solutions to produce it (Jambor et al., 1996). Under such conditions the inherently high

ionic strength of the solution makes calculation of species distributions extremely difficult.

Therefore, distribution coefficients for cation mixing in herbertsmithite and gillardite were

determined from the available data and discussed below.

CHAPTER 4.1

98

4.1.2 SAMPLES AND METHODS

4.1.2.1 Powder X-ray diffraction

All samples produced in this study were identified by powder X-ray diffraction

methods using a Bruker D8 Advance powder diffractometer (Cu Kα radiation, λ = 1.5406 Å,

40 kV, 40 mA), with pure Si as internal standard. Traces were taken between 5–70° 2θ, step

size 0.02°, 1 step sec-1. Clinoatacamite was identified by the presence of the Irel = 60, 2.266 Å

(220), and Irel

4.1.2.2 Clinoatacamite synthesis

= 50, 2.243 Å (004) reflections. The remaining samples gave a powder pattern

consistent with the paratacamite substructure (powder diffraction file 50–1558) and were

indexed based on the herbertsmithite unit cell (a ~ 6.8 and c ~14.1) or the gillardite unit cell

(a ~ 6.8 and c ~ 13.9), for Zn- or Ni-bearing samples, respectively, as described in Chapter

3.2. Samples were indexed using the program PowderCell (Kraus and Nolze, 1996a, b) and

unit cell parameters were refined using the least squares software LAPOD (Langford, 1973).

Samples of end-member clinoatacamite were prepared in a similar way to the method

of Jambor et al. (1996). An aqueous solution containing 0.07 M CuCl2 in 50 cm3 of milliQ

water was made. The solution was heated under reflux and titrated with standardised 0.05 M

NaOH to an equivalent amount of total Cu. Under these conditions the concentration of Cl-

present partially inhibits the nucleation of clinoatacamite and promotes the formation of

atacamite. These minerals are part of an Ostwald Step Rule series (Ostwald, 1897) with

clinoatacamite being the thermodynamically stable phase. To promote the decomposition of

the atacamite impurity, the solutions were heated under reflux for six days.

Whilst under reflux, a small amount of water was retained in the condenser. In order

to obtain the volume of the reaction solution, the total mass of the reaction vessel was

carefully measured before and after reflux. The mass of NaOH added was determined from

its density and volume. The flasks were allowed to cool to room temperature prior to

recording the final mass.

Each flask was equilibrated in a thermostatted water bath at 298.2 K over an

additional five days. The pH of each solution was measured and the product was separated

from the filtrate by suction filtration through fibre glass filter paper (Whatman GF/F,

0.7 μm). The filtrate was retained and stored in polyethylene sample bottles for further

analysis. The solid product was washed with MilliQ water then acetone and dried at the

pump. Six repeats were made in order to calculate the error in lg K.

CHAPTER 4.1

99

4.1.2.3 Synthesis of cation-substituted phases

To derive the variation in distribution coefficients (D) and solid-state activity

coefficients (γ) with increasing substitution of Zn and Ni, a series of solutions were prepared

similar to those described above. To reduce the ionic strength of the solution and the amount

of Cl-, the concentration of CuCl2 was lowered to 0.01 M in 200 cm3 of milliQ water. To

these solutions an amount of ZnCl2 or NiCl2 was added while maintaining the concentration

of CuCl2 at 0.01 M. The range of Cu2+:M2+ mol ratios were 1:0.5 to 1:25 when M = Zn and

1:0.5 to 1:17 when M = Ni. It was found that purely substituted clinoatacamite could only be

obtained for Cu:Zn ratios of between 1:0.5 and 1:4, and Cu:Ni ratios between 1:0.5 and 1:3.

With Zn or Ni in excess of these ratios the solid gave an X-ray powder diffraction pattern

consistent with that of the paratacamite subcell (powder diffraction file 50-1558). Based on

results obtained in Chapter 3.2, these samples were treated as having the aristotype structure

and are discussed in the following section of this Chapter. The crystallinity of samples with a

composition Cu4-xMx(OH)6Cl2

4.1.2.4 Solution calculations

between x ~ 0.20 and 0.30 was generally quite poor, as

indicated by broad peaks and a reduced signal. This limited the accuracy of monoclinic or

rhombohedral phase differentiation for those particular samples by powder X-ray diffraction.

In addition, the product was composed in part of crystallites smaller than the retention pore

size of 0.7 μm for the filter paper used, which did not allow separation of the dispersed solid

phase from the filtrate. For some of these samples, centrifugation followed by subsequent

decantation of the filtrate was attempted, but not successful.

Total metal concentrations were determined by atomic absorption spectrometry

(AAS) using a GBC100 spectrophotometer (air-acetylene), from the filtrate and solids after

dissolution in dilute nitric acid. Species distributions at 25°C were calculated (Table 4.1.1)

using the program COMICS (Perrin and Sayce, 1967). The species modelled and their

cumulative lg K values at 298.2 K are listed in Table 4.1.2.

The concentrations of other hydrolysed species of higher nuclearity at the pH

measured for these solutions are negligible, as determined by trial calculations, and can be

omitted from the calculation. The only hydrolysed species retained for the calculation were

CuOH+ and ZnOH+, which became significant for solutions with a pH > 4.6. Significant

solution species with chloride as a ligand were ZnCl+, ZnCl2°, CuCl+, and NiCl+

Equilibrium constants at 298.2 K of the species were iteratively corrected to the

solution’s ionic strength using the extended Debye-Hückel equation for solutions with an

.

CHAPTER 4.1

100

ionic strength I < 0.1 (lg γi = -0.5085zi2[I½/(1 + I½)], where zi is the charge of species i), and

the Davies modification of the Debye-Hückel equation for solutions with I ˃ 0.1

(lg γi = -0.5085zi2[I½/(1 + I½) - 0.3I]).

Table 4.1.1. Species concentrations from COMICS calculations for clinoatacamite Cu4(OH)6Cl2. Sample Concentration (mol dm-3) Na+ Cl- Cu2+ CuCl+ CuCl2 CuOH+ 1 1.583 x10-2 3.775 x10-2 1.074 x10-2 4.346 x10-4 8.414 x10-7 2.571 x10-6 2 1.583 x10-2 3.699 x10-2 1.037 x10-2 4.108 x10-4 7.793 x10-7 2.539 x10-6 3 1.585 x10-2 3.737 x10-2 1.054 x10-2 4.222 x10-4 8.090 x10-7 2.692 x10-6 4 1.584 x10-2 3.801 x10-2 1.085 x10-2 4.420 x10-4 8.620 x10-7 2.670 x10-6 5 1.583 x10-2 3.766 x10-2 1.070 x10-2 4.316 x10-4 8.334 x10-7 2.668 x10-6 6 1.582 x10-2 3.808 x10-2 1.091 x10-2 4.453 x10-4 8.498 x10-7 2.611 x10-6

Table 4.1.2. Equilibrium constants used in COMICS calculations Species lg K ref.* CuCl+ 0.40 1 CuCl2° -0.71 2 CuOH+ -7.6 1 NiCl+ -0.35 3 ZnCl+ 0.43 1 ZnCl2° 0.61 1 ZnCl3

- 0.5 1 ZnCl4

2- 0.2 1 ZnOH+ -9.0 1 NaOH -14.3 1 *[1] Smith and Martell (1976); [2] Long and Angino (1977); [3] Lee and Nam (2009).

The value of γ for neutral species was taken as unity. Final values are listed in Table

4.1.3. The equilibrium concentrations of free Cu2+(aq) and Cl-

Cu

(aq), as well as the measured

pH were used to calculated the equilibrium constant K for end-member clinoatacamite with

respect to its doubled formula, using equation (4.2) below, where a(i) is the activity of

species i. The activity was calculated with the equation a(i) = m(i)γ(i), where m(i) is the molal

concentration of species i.

4(OH)6Cl2 + 6H+ ⇋ 4Cu2+(aq) + 2Cl-(aq) + 6H2

𝐾𝐾clinoatacamite = 𝑎𝑎(Cu 2+)4 𝑎𝑎(Cl−)2

𝑎𝑎(H+)6 (4.2)

O(l) (4.1)

CHAPTER 4.1

101

The activity of cation-substituted clinoatacamite is no longer unity. The effect of

substitution on its stability was investigated by calculating the changes in its solid state

activity coefficient. Equations (4.3) and (4.4) describe the changes to the chemical

equilibrium for cation substitution in clinoatacamite. The solid state activity coefficients (γ)

for clinoatacamite with Zn and Ni substituted in the lattice were calculated from the

equilibrium constant of clinoatacamite, (Kclinoatacamite), the equilibrium constant (K) for each

new solution and the mole fraction of cation substituted clinoatacamite (N), using equation

(4.5). The value of K of each new solution was determined using equation (4.2). Only the

composition of the interlayer position was used to determine the mole fraction (N), where, for

example, (Cu0.90Zn0.10)Cu3(OH)6Cl2

𝐾𝐾clinoatacamite = 𝑎𝑎(Cu 2+)4 𝑎𝑎(Cl−)2

𝑎𝑎(H+)6 𝑎𝑎clinoatacmaite (4.3)

gives N = 0.90.

𝑎𝑎clinoatacamite = 𝑁𝑁clinoatacmaite γclinoatacamite R

γclinoatacamite = 1��𝐾𝐾clinoatacamite /�𝑎𝑎(Cu 2+)4 𝑎𝑎(Cl−)2

𝑎𝑎(H +)6 � �𝑁𝑁clinoatacamite �R

(4.4)

4.1.3 RESULTS AND DISCUSSION

(4.5)

A value of lg K at 298.2 K for end-member clinoatacamite was determined as

13.22 ±0.11 from the average of six repeats (Table 4.1.3). To calculate the standard Gibbs

free energy of formation for clinoatacamite, values of ΔGfɵ for Cu2+(aq) = 65.1 ±0.1,

Cl-(aq) = -131.2 ±0.1 and H2O(l) = -237.1 ±0.1, were used (Robie and Hemingway, 1995).

Calculation using these values and the lg K determined for clinoatacamite gave a value for

the Gibbs free energy of formation, ΔGfɵ(clinoatacamite, 298.2 K) = -1349.61 ±1.83 kJ mol-1.

Woods and Garrels (1986a), reported a ΔGfɵ value for clinoatacamite (reported incorrectly as

paratacamite), Cu4(OH)6Cl2 as -1341.8 kJ mol-1 calculated using the equilibrium between

brochantite, Cu4(OH)6SO4

The compositions determined and equilibrium values used for the calculation of lg K

for Zn- and Ni-substituted clinoatacamite are given in Tables 4.1.4 and 4.1.5, respectively.

For a solid solution approaching end-member composition (N > 0.90), the value of γ can be

and clinoatacamite. This compares reasonably well with the value

in this study.

CHAPTER 4.1

102

used to indicate the behaviour of the substituting cation as a solute interacting with in a pure

substance.

Table 4.1.3. Data used for the calculation of the clinoatacamite, Cu4(OH)6Cl2, stability constant. [Cu2+] [Cl-] I pH γ(2+/-) γ(1+/-) lg K (x10-2) (x10-2) (x10-2) 1.074 3.775 4.849 4.259 0.4295 0.8095 13.180 1.037 3.699 4.818 4.259 0.4304 0.8100 13.106 1.054 3.737 4.790 4.277 0.4313 0.8104 13.255 1.085 3.801 4.885 4.271 0.4284 0.8090 13.271 1.070 3.766 4.836 4.277 0.4299 0.8097 13.281 1.091 3.808 4.899 4.259 0.4279 0.8088 13.208 Table 4.1.4. Equilibrium constants and solid state activity coefficients for the substitution of Cu2+ by Zn2+ in clinoatacamite. N [Cu2+] [Zn2+] [Cl-] I pH γ(2+/-) γ(1+/-) lg K γclinoatacamite (x10-3) (x10-3) (x10-2) (x10-2) 0.96 5.129 1.529 3.115 3.781 4.565 0.4665 0.8265 13.73 3.35 0.93 5.268 3.224 3.492 4.341 4.557 0.4460 0.8172 13.74 3.53 0.89 5.215 6.564 4.168 5.346 4.507 0.4150 0.8026 13.43 1.83 0.87 5.168 9.734 4.828 6.318 4.475 0.3903 0.7904 13.23 1.18 0.86 4.955 14.56 5.803 7.755 4.481 0.3605 0.7749 13.20 1.11 0.83 4.993 12.17 5.294 7.010 4.474 0.3752 0.7826 13.17 1.03

Table 4.1.5. Equilibrium constants and solid state activity coefficients for the substitution of Cu2+ by Ni2+ in clinoatacamite. N [Cu2+] [Ni2+] [Cl-] I pH γ(2+/-) γ(1+/-) lg K γclinoatacamite (x10-3) (x10-3) (x10-2) (x10-2

In this case, the pure substance is the doubled formula Cu

) 0.97 5.121 7.485 4.272 5.533 4.466 0.4099 0.8001 13.15 0.88 0.95 5.107 10.97 4.975 6.583 4.446 0.3843 0.7874 13.03 0.68 0.94 5.142 14.86 5.775 7.775 4.428 0.3602 0.7747 12.94 0.56 0.92 5.331 19.16 6.689 9.139 4.404 0.3372 0.7620 12.86 0.47 0.90 2.672 22.34 5.899 8.400 4.584 0.3491 0.7687 12.70 0.33 0.87 2.663 21.76 5.780 8.222 4.577 0.3522 0.7703 12.65 0.31

4(OH)6Cl2 end-member. If

the value of γ is near unity across the range of possible substitution, then the solid solution

may be classified as ideal (Garrels and Christ, 1965). The system may be classed as regular

when the value of γ approaches unity as N approaches unity but shows divergence when

N < 0.90. It can be seen from Figure 4.1.1 that the value of γ for the Zn-substitution series in

clinoatacamite begins well above unity when N > 0.90. With higher amounts of Zn

substitution the value of γ decreases rapidly, but non-linearly, towards unity. The distribution

CHAPTER 4.1

103

of γ values suggests that the solid solution behaves irregularly for compositions approaching

the end-member Cu4(OH)6Cl2 composition. As the mol fraction N approaches 0.80, i.e., the

composition Cu3.80Zn0.20(OH)6Cl2, γ appears to behave ideally. The distribution of γ values

when Ni2+

Figure 4.1.1. The solid-state activity coefficients for Zn-substituted (circles), and Ni-substituted (triangles) clinoatacamite. The mol fraction N is described in the text, and the composition x is based on the formula Cu

is the substituting cation appears quite different. For compositions very close to

unity the value of γ is also close to unity, but is below 1. Increased Ni substitution results in a

decrease in γ.

4-xMx(OH)6Cl2

The behaviour of Zn

. The calculated error in the Zn series solid state activity coefficient is smaller than the size of the symbol.

2+ and Ni2+ solid solution in clinoatacamite is quite different.

These effects may be attributed to the difference in crystal-chemical behaviour for that cation

in terms of suitability for the substitution of Cu2+. There are subtle differences in the radii of

six-coordinate Ni2+ and Zn2+ (0.69 and 0.74 Å, respectively; Shannon, 1976). In addition, the

cell volumes of herbertsmithite and gillardite are noticeably different (V = 569.3 and

560.41(2) Å3, respectively; Braithwaite et al., 2004; Clissold et al., 2007). Both substitution

series can be readily produced by solution methods, but the formation pure Ni-bearing

clinoatacamite, gillardite, and perhaps paratacamite, takes longer than the Zn-bearing

congeners. Jambor et al. (1996) reported mixtures of Ni(OH)2 and Cu2(OH)3Cl using similar

solution composition to those described above, but with significantly reduced reaction times

(2–3 hours).

CHAPTER 4.2

104

4.2 DISTRIBUTION COEFFICIENTS FOR HERBERTSMITHITE AND GILLARDITE

4.2.1 INTRODUCTION

Paratacamite has not been identified using the synthetic method described above.

Thus, either paratacamite exists within a narrow range of compositions, or is slow to form.

The phase identified to exist across almost all of the rhombohedral series is herbertsmithite or

gillardite, in Cu/Zn or Cu/Ni solutions, respectively. It is noted that the reversible phase

transformation between paratacamite and herbertsmithite or gillardite makes it certain that the

experiments were not complicated by the possible presence of metastable solids. No

definitive evidence has been collected that suggests what the lower end-member composition

is for either herbertsmithite or gillardite. Without knowledge of this end-member the

equilibrium constant for the pure phase cannot be determined. In order to understand some

aspects of the conditions of formation of the rhombohedral phase, distribution coefficients

(D) were calculated.

4.2.2 SAMPLES AND METHODS

The method used to produce synthetic samples is described in the previous section of

this Chapter. The species distribution at equilibrium at 298 K (Tables 4.2.1 and 4.2.2) was

calculated using COMICS and the appropriate aqueous species in Table 4.1.2. Solution

aqueous concentrations, pH, adjusted ionic strength and distribution coefficients are given in

Tables 4.2.3 and 4.2.4 for the Zn and Ni systems, respectively.

4.2.3 RESULTS AND DISCUSSION

For the range of solutions reproduced here, it was found that Ni2+ effected a more

efficient uptake into the solid phase than Zn2+. This is evident from the higher ratio of

solution Zn2+ required to produce a solid phase with equivalent composition to the Ni-bearing

congener. This is most likely due to the stability of additional aqueous Zn species, such as

ZnCl20, ZnCl3

- and ZnCl42- which become significant with a ZnCl2 concentration > 0.09 M.

This reduces the available free Zn2+(aq) concentration and influences the solid phase metal

ratio. It can be seen in Table 4.2.3 that there are inconsistencies with the solution composition

versus the solid phase composition when Zn2+ is substituting for Cu2+.

Table 4.2.1. Species concentrations from COMICS calculations for the Zn substitution series in Cu4-xZnx(OH)6Cl2. Hydrolysed species and are negligible and were omitted. Composition Concentration (mol dm-3) x(Cu) x(Zn) Na+ Cl- Zn2+ Cu2+ ZnCl+ ZnCl2

0 ZnCl3- ZnCl4

2- CuCl+ 0.96 0.04 1.758 x10-2 3.115 x10-2 1.529 x10-3 5.129 x10-3 5.996 x10-5 1.911 x10-6 4.621 x10-6 1.067 x10-9 1.877 x10-4 0.93 0.07 1.758 x10-2 3.492 x10-2 3.224 x10-3 5.268 x10-3 1.354 x10-4 4.727 x10-6 1.281 x10-7 3.394 x10-9 2.064 x10-4 0.89 0.11 1.758 x10-2 4.168 x10-2 6.564 x10-3 5.215 x10-3 3.070 x10-4 1.250 x10-5 4.045 x10-7 1.309 x10-8 2.276 x10-4 0.87 0.13 1.774 x10-2 4.828 x10-2 9.734 x10-3 5.168 x10-3 4.921 x10-4 2.269 x10-5 8.502 x10-7 3.260 x10-8 2.438 x10-4 0.86 0.14 1.789 x10-2 5.803 x10-2 1.456 x10-2 4.955 x10-3 8.447 x10-4 4.575 x10-5 2.061 x10-6 9.720 x10-8 2.683 x10-4 0.83 0.17 1.772 x10-2 5.294 x10-2 1.217 x10-2 4.993 x10-3 6.444 x10-4 3.184 x10-5 1.309 x10-6 5.631 x10-8 2.467 x10-4 0.69 0.31 9.369 x10-3 4.160 x10-2 1.359 x10-2 2.171 x10-3 6.196 x10-4 2.461 x10-5 7.947 x10-7 2.626 x10-8 9.240 x10-4 0.67 0.33 9.127 x10-3 6.994 x10-2 2.730 x10-2 2.218 x10-3 1.663 x10-3 1.013 x10-4 5.500 x10-6 3.350 x10-7 1.261 x10-4 0.62 0.38 9.555 x10-3 5.605 x10-2 2.028 x10-2 2.368 x10-3 1.085 x10-3 5.477 x10-5 2.414 x10-6 1.125 x10-7 1.183 x10-4 0.59 0.41 9.401 x10-3 7.853 x10-2 3.108 x10-2 2.216 x10-3 2.385 x10-3 1.708 x10-4 1.041 x10-5 6.803 x10-7 1.587 x100.52 0.48 9.114 x10

-4 -3 8.973 x10-2 3.645 x10-2 2.248 x10-3 3.053 x10-3 2.441 x10-4 1.700 x10-5 1.299 x10-6 1.757 x10-4

0.51 0.49 9.127 x10-3 6.158 x10-2 2.308 x10-2 2.419 x10-3 1.326 x10-3 7.114 x10-5 3.401 x10-6 1.824 x10-7 1.298 x10-4 0.43 0.57 9.136 x10-3 1.011 x10-1 4.168 x10-2 2.289 x10-3 3.843 x10-3 3.384 x10-4 2.655 x10-5 2.338 x10-6 1.970 x10-4 0.42 0.58 9.743 x10-3 1.667 x10-1 7.139 x10-2 2.235 x10-3 9.675 x10-3 1.342 x10-3 1.737 x10-4 2.641 x10-5 2.826 x10-4 0.41 0.59 9.733 x10-3 1.458 x10-1 6.184 x10-2 2.402 x10-3 7.502 x10-3 9.100 x10-4 1.030 x10-4 1.370 x10-5 2.719 x10-4 0.38 0.62 9.820 x10-3 1.567 x10-1 6.623 x10-2 2.433 x10-3 8.500 x10-3 1.108 x10-3 1.348 x10-4 1.926 x10-5 2.892 x10-4 0.34 0.66 1.020 x10-2 2.060 x10-1 8.855 x10-2 2.332 x10-3 1.416 x10-2 2.426 x10-3 3.879 x10-4 7.287 x10-5 3.481 x10-4 0.32 0.68 1.020 x10-2 1.831 x10-1 7.837 x10-2 2.383 x10-3 1.140 x10-2 1.736 x10-3 2.467 x10-4 4.120 x10-5 3.234 x10-4 0.31 0.69 1.019 x10-2 1.932 x10-1 8.284 x10-2 2.339 x10-3 1.271 x10-2 1.995 x10-3 2.992 x10-4 5.393 x10-5 3.349 x10-4 0.28 0.72 1.019 x10-2 2.108 x10-1 9.072 x10-2 2.279 x10-3 1.484 x10-2 2.542 x10-3 4.159 x10-4 8.180 x10-5 3.480 x10-4 0.28 0.72 1.020 x10-2 2.230 x10-1 9.603 x10-2 2.221 x10-3 1.662 x10-2 3.013 x10-3 5.215 x10-4 1.085 x10-4 3.589 x10-4 0.18 0.82 1.021 x10-2 2.421 x10-1 1.042 x10-1 2.275 x10-3 1.958 x10-2 3.853 x10-3 7.240 x10-4 1.635 x10-4 3.990 x10

-4

105

CH

APTER

4.2

CHAPTER 4.2

106

Table 4.2.2. Species concentrations from COMICS calculations for the Ni substitution series in Cu4-xNix(OH)6Cl2. Hydrolysed species and are negligible and were omitted. Composition Concentration (mol dm-3) x(Cu) x(Ni) Na+ Cl- Ni2+ Cu2+ NiCl+ CuCl+ 0.97 0.03 1.718 x10-2 4.272 x10-2 7.485 x10-3 5.121 x10-3 1.135 x10-4 2.138 x10-4 0.95 0.05 1.717 x10-2 4.975 x10-2 1.097 x10-2 5.107 x10-3 1.936 x10-4 2.426 x10-4 0.94 0.06 1.718 x10-2 5.775 x10-2 1.486 x10-2 5.142 x10-3 2.976 x10-4 2.708 x10-4 0.92 0.08 1.718 x10-2 6.689 x10-2 1.916 x10-2 5.331 x10-3 4.342 x10-4 3.035 x10-4 0.90 0.10 8.365 x10-3 5.899 x10-2 2.234 x10-2 2.672 x10-3 4.570 x10-4 1.373 x10-4 0.87 0.13 8.371 x10-3 5.780 x10-2 2.176 x10-2 2.663 x10-3 4.362 x10-4 1.372 x10-4 0.67 0.33 8.274 x10-3 1.333 x10-1 5.856 x10-2 2.562 x10-3 2.586 x10-3 2.713 x10-4 0.54 0.46 8.264 x10-3 1.008 x10-1 4.251 x10-2 2.895 x10-3 1.452 x10-3 2.484 x10-4 0.50 0.50 8.276 x10-3 1.186 x10-1 5.125 x10-2 2.741 x10-3 2.059 x10-3 2.641 x10-4 0.48 0.52 8.298 x10-3 1.296 x10-1 5.667 x10-2 2.649 x10-3 2.432 x10-3 2.727 x10-4 0.47 0.53 8.376 x10-3 1.301 x10-1 5.656 x10-2 2.916 x10-3 2.436 x10-3 3.013 x10-4 0.41 0.59 8.266 x10-3 1.551 x10-1 6.869 x10-2 2.800 x10-3 3.529 x10-3 3.295 x10-4 0.37 0.63 8.458 x10-3 2.198 x10-1 9.881 x10-2 3.051 x10-3 7.192 x10-3 4.858 x10-4 0.33 0.67 8.423 x10-3 2.299 x10-1 1.036 x10-1 3.018 x10-3 7.706 x10-3 5.026 x10-4 0.29 0.71 8.468 x10-3 2.353 x10-1 1.062 x10-1 2.927 x10-3 8.087 x10-3 4.990 x10-4 0.27 0.73 8.545 x10-3 1.888 x10-1 8.448 x10-2 2.827 x10-3 5.282 x10-3 3.957 x10-4 0.24 0.76 8.450 x10-3 2.457 x10-1 1.111 x10-1 2.897 x10-3 8.832 x10-3 5.157 x10-4 0.23 0.77 8.507 x10-3 2.120 x10-1 9.522 x10-2 2.957 x10-3 6.683 x10-3 4.540 x10-4

Table 4.2.3. Data used for the calculation of distribution coefficients in the rhombohedral Cu-Zn solid solution series. Composition* [Cu2+] [Zn2+] [Cl-] I pH D x(Cu) x(Zn) (x10-3) (x10-2) (x10-1) (x10-1) (x10-2) 0.69 0.31 2.171 1.359 0.4160 0.5740 4.708 0.7180 0.67 0.33 2.218 2.730 0.6994 0.9950 4.598 0.4002 0.62 0.38 2.368 2.028 0.5605 0.7870 4.616 0.7157 0.59 0.41 2.216 3.108 0.7853 1.118 4.585 0.4955 0.52 0.48 2.248 3.645 0.8973 1.284 4.591 0.5693 0.51 0.49 2.419 2.308 0.6158 0.8710 4.653 1.007 0.43 0.57 2.289 4.168 1.011 1.451 4.601 0.7280 0.42 0.58 2.235 7.139 1.667 2.406 4.586 0.4323 0.41 0.59 2.402 6.184 1.458 2.102 4.592 0.5589 0.38 0.62 2.433 6.623 1.567 2.251 4.575 0.5994 0.34 0.66 2.332 8.855 2.060 2.975 4.547 0.5112 0.32 0.68 2.383 7.837 1.831 2.642 4.566 0.6461 0.31 0.69 2.339 8.284 1.932 2.788 4.550 0.6285 0.28 0.72 2.279 9.072 2.108 3.044 4.535 0.6460 0.28 0.72 2.221 9.603 2.230 3.221 4.531 0.5947 0.18 0.82 2.275 10.42 2.421 3.498 4.523 0.9946 *Based on the cation composition of the total interlayer metal position

CHAPTER 4.2

107

Table 4.2.4. Data used for the calculation of distribution coefficients in the rhombohedral Cu-Ni solid solution series. Composition* [Cu2+] [Ni2+] [Cl-] I pH D x(Cu) x(Ni) (x10-3) (x10-2) (x10-1) (x10-1) (x10-2

Figure 4.2.1. Distribution coefficients for synthetic rhombohedral Zn- and Ni-members of the basic Cu(II) chloride series. The composition x, applies to the formula Cu

) 0.67 0.33 2.562 5.856 1.333 1.945 4.431 2.155 0.54 0.46 2.895 4.251 1.008 1.462 4.397 5.801 0.50 0.50 2.741 5.125 1.186 1.726 4.355 5.348 0.48 0.52 2.649 5.667 1.296 1.889 4.376 5.064 0.47 0.53 2.916 5.656 1.301 1.896 4.372 5.814 0.41 0.59 2.800 6.869 1.551 2.266 4.296 5.866 0.37 0.63 3.051 9.881 2.198 3.217 4.168 5.258 0.33 0.67 3.018 10.36 2.299 3.365 4.157 5.915 0.32 0.68 2.978 7.807 1.756 2.566 4.245 8.106 0.29 0.71 2.927 10.62 2.353 3.445 4.150 6.748 0.27 0.73 2.827 8.448 1.888 2.761 4.227 9.048 0.24 0.76 2.897 11.11 2.457 3.597 4.153 8.257 0.23 0.77 2.957 9.522 1.551 3.102 4.210 10.40 *Based on the cation composition of the total interlayer metal position

4-xMx(OH)6Cl2.

CHAPTER 4.2

108

The distribution coefficients listed in Table 4.2.3 and 4.2.4 are all in the order of

ca 10-2

. These values are displayed in Figure 4.2.1 against the solid state composition of the

substituting cation. The spread of D values for Ni-bearing samples is more consistent across

the range of solution compositions examined. However, these data show that the dissolution

of herbertsmithite and gillardite is incongruent.

CHAPTER 5

109

CHAPTER 5 – CONCLUSIONS

The studies reported in this thesis on substitution phenomena in the basic Cu(II)

chloride minerals have several implications for the group, particularly concerning analyses of

synthetic material. The combination of single-crystal X-ray and Raman spectroscopy

measurements has elucidated several aspects associated with the structural transformations

inherent in the group.

5.1 NEW MINERALS

The structure of paratacamite was confirmed in space group R 3� based on the

supercell reported by Frondel (1950) and Fleet (1975), using material from the Generosa

mine type specimen (BM86958). An investigation of specimens of the basic Cu(II) chloride

minerals revealed two new congeners of paratacamite. One is from the Camarones Valley,

Arica Province, Chile and is Mg-analogue, corresponding to the general formula

(Mg,Cu)Cu3(OH)6Cl2, unit cell a = 13.689(1) and c = 14.025(1) Å, space group R3�, and a

pronounced subcell with a' ≈ ½a, c' ≈ c, space group R3�m. The second is from the Carr Boyd

Rocks Mine, Western Australia, Australia, and is a Ni-analogue with the general formula

(Ni,Cu)Cu3(OH)6Cl2, unit cell a = 13.665(4), c = 13.915(4) Å, space group R3� , and a

pronounced subcell as described above. In addition, a new analogue of herbertsmithite was

identified from the Torrecillas Mine, Salar Grande, Iquique Provence, Tarapacá Region,

Chile, with dominant interlayer Co, giving the general formula (Co,Cu)Cu3(OH)6Cl2

5.2 A REVERSIBLE R𝟑𝟑� TO R𝟑𝟑�m PHASE TRANSFORMATION

, unit

cell a = 6.8436(6) and c = 14.064(1) Å, space group R3�m, and no observable superstructure.

The new R 3� congeners of paratacamite display a rhombically distorted M(2)

coordination environment, similar to that reported by Fleet (1975). Variable temperature

single-crystal X-ray studies (between 100–423 K) of paratacamite from the type specimen

have revealed that the M(2) octahedron undergoes a systematic reduction in distortion with

increasing temperature. The long M(2)–O(3) bond decreases with a simultaneous increase in

length of the short M(2)–O(1) bond while the M(2)–O(2) bond length remains stable at ~2.1

Å. The non-tetragonally elongated M(1) octahedron is temperature invariant, maintaining a

uniform M(1)–O bond length of ~2.1 Å. Between 353 and 393 K, the configuration of M(2)

merges with that of M(1) and the superstructure reflections disappear. The diffraction pattern

CHAPTER 5

110

obtained at 393 and 423 K is that of the R3�m substructure. Upon cooling back to 300 K, the

superstructure reflections return. This result is consistent with a reversible structural

transformation from R 3� → R 3� m and establishes that the R 3� superstructure is

thermodynamically stable at 300 K.

5.3 THE (2+2+2) DISTORTION OF M(2) IN PARATACAMITE

The above studies have allowed the origin of the (2+2+2) distortion of the M(2)

octahedron, which is a characteristic feature of the structure of paratacamite, to be elucidated.

First, the Mg congener of paratacamite displays a statistical distribution of interlayer Mg

between M(1) and M(2) (60% occupancy each), determined from the site X-ray scattering

factors. With the assumption that all paratacamite congeners exhibit a similar statistical

distribution of interlayer cations, the rhombic distortion of M(2) is most likely generated by

the superimposition of non-tetragonally elongated M(OH)6 octahedra with two or three

orientations of the common (4+2) Jahn-Teller distorted Cu(OH)6 octahedra. The relative

occupancies of each orientation appear to the dependent upon temperature. In addition, the

variable nature of the M(2) distortion with temperature and the observed O atom anisotropic

thermal ellipsoids suggests that this site is dynamically Jahn-Teller distorted. The observed

(2+2+2) distortion would therefore be the consequence of time lapsed averaged positions of

all atoms involved. In addition, it can be inferred that the temperature invariant M(1)

environment is similarly composed of a superimposition on non-tetragonally elongated

M(OH)6 octahedra with three orientations of equally occupied, static (4+2) Jahn-Teller

distorted octahedra. The O atom displacements which would result from the abovementioned

superimposition of M(OH)6 and Cu(OH)6

The superstructure reflections of type paratacamite reduce in intensity with

increasing temperature as the distortion of the M(2) coordination environment is reduced.

Therefore, the loss of these reflections between 353 and 393 K indicate that the

superstructure is derived from atomic displacements, particularly concerning the O atoms

associated with the M(2) environment, rather than cation ordering at M(1) as it has been

suggested in the literature (Grice et al., 1996; Braithwaite et al., 2004). The proposed O atom

octahedra would therefore be disordered over the

structure. The relative occupancies of each O atom would be defined by the proportion of

each orientation of (4+2) Jahn-Teller distorted octahedra which itself is dependent upon

temperature.

CHAPTER 5

111

disorder is the most likely explanation as to why the paratacamite superstructure reflections

are so weak in intensity.

5.4 COMPOSITION-INDUCED STRUCTURAL TRANSFORMATIONS

5.4.1 CRYSTALLOGRAPHIC STUDIES

A composition-dependent phase relationship from the R3� to R3�m structure (Jambor

et al., 1996; Braithwaite et al., 2004) should be characterised by a similar systematic

reduction in M(2) distortion. Both new Mg- and Ni-analogues of paratacamite display greater

than 50% interlayer occupancy of the substituting cation and confirm this hypothesis. The QE

and BAV values calculated for both interlayer octahedra and the similarity of the M(2)–O

bond lengths in the new paratacamite analogues suggests that they are near the upper limit of

compositional stability for the R 3� superstructure. Significantly, single-crystal X-ray

diffraction analyses of several other naturally occurring samples have established that the

R3�m phase can exist with a composition near the monoclinic–rhombohedral transition zone

defined by Jambor et al. (1996). The R3�m structure was reported from synthetic single-

crystals of composition between Cu3.67Zn0.33(OH)6Cl2 and Cu3Zn(OH)6Cl2

Trends associated with the QE and BAV values of the R3�m structure interlayer

octahedron, with changes in composition, indicate that the distortion present in type

herbertsmithite (Braithwaite et al., 2004) and type gillardite (Clissold et al., 2007) is at a

minimum for Zn and Ni substitution, respectively. With an increase in the interlayer Cu

content in gillardite, a sharp increase in the distortion of the M(1)O

by Schores et al.

(2005), and is in agreement with this study of natural samples.

6 octahedron occurs

between ~Cu3.25Ni0.75(OH)6Cl2 and Cu3.15Ni0.85(OH)6Cl2. This distortion remains high and

relatively stable with increasing Cu content. The QE and BAV values from the R 3�m

substructure of the Ni analogue of paratacamite are in support of this. The distortion observed

in the interlayer octahedra of herbertsmithite remains at a minimum when Zn in is excess of

at least ~Cu3.40Zn0.60(OH)6Cl2. With greater Cu substitution the distortion increases non-

linearly towards the monoclinic–rhombohedral transition zone. Based on this behaviour it is

likely that herbertsmithite would become metastable for compositions below

~Cu3.40Zn0.60(OH)6Cl2. Similarly, gillardite is expected to become metastable for

compositions below ~Cu3.25Ni0.75(OH)6Cl2 or Cu3.15Ni0.85(OH)6Cl2

The QE and BAV values determined from the R 3�m substructure of holotype

paratacamite are 1.053 and 207.64 (degrees

.

2), respectively. Based on the trends observed in

CHAPTER 5

112

herbertsmithite samples, the holotype crystal of paratacamite used by Fleet (1975) has a

composition between ~Cu3.70Zn0.30(OH)6Cl2 and Cu3.67Zn0.33(OH)6Cl2

5.4.1 RAMAN SPECTROSCOPY

. Similarly, the trend

observed in the unit cell strain of herbertsmithite samples with decreasing Zn content

suggests that holotype paratacamite, with a scalar strain of 0.0028, has a composition in

agreement with the above.

The end-members for paratacamite cannot be defined from the above information,

but the above results strongly suggest that they will be different depending upon the type of

cation substituting for Cu.

Much of the above was confirmed using Raman spectroscopy on oriented natural

single crystals and synthetic polycrystalline samples of the Zn- and Ni-bearing members. The

Raman spectrum of paratacamite was reported from a confirmed single crystal from the

BM86958 type specimen. Several trends related to structural changes induced by composition

are observed in the synthetic series. The spectrum of end-member clinoatacamite appears

similar to anatacamite. The incorporation of Zn or Ni in the monoclinic structure induces the

appearance of additional modes in the M–O–H deformation and O–H stretching regions

which systematically increase in intensity with increasing substitution for Cu. After the

monoclinic to rhombohedral transformation these additional modes become the predominant

feature of these regions as the peaks related directly to the clinoatacamite spectrum

sequentially decrease in intensity.

The appearance of a single mode with moderate intensity at ~705 cm-1 in samples

exhibiting compositions in excess of ~Cu3.80M0.20(OH)6Cl2 appears to be characteristic of

the aristotype structure. These results suggest that the transformation series examined

proceeds from P21/n → R3�m. The similarity of the clinoatacamite and anatacamite Raman

spectra of stoichiometrically pure material may indicate that low lattice impurities allow

clinoatacamite to distort towards the triclinic structure. The proposed series of space group

symmetries with increasing substitution for Cu is therefore P1�→ P21/n → R3�m. A structural

conversion from clinoatacamite to anatacamite was not observed in these synthetic samples.

Clinoatacamite is the thermodynamically stable phase for the formula Cu2(OH)3Cl which

suggests that either anatacamite is a transient phase in the Ostwald series from botallackite to

clinoatacamite, or the nucleation phenomena of the triclinic phase requires a specific

combination of factors to proceed.

CHAPTER 5

113

The nature of paratacamite in the composition-induced transformation series is not

fully understood. Several factors must be considered. The composition of paratacamite used

in this Raman analysis, ~Cu3.80Zn0.20(OH)6Cl2, is most likely at the limit of stability for the

rhombohedral phase. This is evident from the observed unit cell metric distortion towards a

lower symmetry structure. At this composition the Raman spectrum is distinct when

compared with those of clinoatacamite and herbertsmithite. However, there are significant

correlations in terms of mode position and complexity between paratacamite and anatacamite.

Changes in composition influence the appearance of the Raman spectra of herbertsmithite

and clinoatacamite, and by inference paratacamite as well. Based on the above, paratacamite

is involved in the series of transformations along the series P1�→ R3� → R3�m, with increasing

substitution for Cu.

From the observed trends in these materials, it is predicted that the Raman spectrum

of paratacamite with Zn content between ~Cu3.70Zn0.30(OH)6Cl2 and ~Cu3.60Zn0.60(OH)6Cl2,

or with Ni content between ~Cu3.70Ni0.30(OH)6Cl2 and Cu3.25Ni0.75(OH)6Cl2

5.6 SOLID–SOLUTION STUDIES

, will exhibit

stronger correlation with the aristotype spectrum.

In synthetic studies, there has been no evidence to suggest that paratacamite forms

from the procedures enlisted. The transformation from clinoatacamite to herbertsmithite or

gillardite involves end-members that are difficult to quantify. In addition, it is likely that the

end-members are dependent upon the nature of the substitution cation.

Calculation of the solid state activity coefficients (γ) show that the behaviour of Zn-

substitution in clinoatacamite is non-ideal for compositions approaching Cu2(OH)3

Cation mixing phenomena which affect the structure wide distribution of non-

tetragonally distorted octahedra influence the required primary order parameters that drive

the transformation. The same may be considered for paratacamite, where a specific set of

solution conditions promote the nucleation and growth of R3� domains. This study also

Cl.

However, the spread of γ values with increasing Zn content displays a rational distribution as

it approaches unity near the monoclinic–rhombohedral transformation. In contrast, Ni-

substitution shows ideal behaviour with small Ni contents in clinoatacamite. The deviation of

γ from unity with increasing Ni content suggests non-ideal behaviour. The distribution

coefficients determined from solution studies of rhombohedral members reinforce the

incongruent nature of dissolution of herbertsmithite and gillardite.

CHAPTER 5

114

suggests that the nature of the substituting cation will produce a different set of end-members

for these phases.

5.7 DEFINITION OF PARATACAMITE

Paratacamite possesses the structure determined by Fleet (1975) and is confirmed in

this study. The mineral crystallises in space group R3�, with hexagonal unit cell parameters

a ~13.6 and c ~14.0 Å, and a pronounced substructure with a′ ≈ ½a and c′ ≈ c, space group

R3�m. It is highly likely that the structure of paratacamite is stabilised by the presence of some

essential cation with an ionic radius comparable to that of Cu2+, such as Zn2+, Ni2+ or Mg2+.

The end-member composition of each paratacamite congener is probably different and

dependent upon the nature of the substituting cation. The substituting cation either

preferentially orders at the interlayer M(1) position, or is statistically distributed between both

interlayer M(1) and M(2) octahedra. The total amount of interlayer substitution may be taken

to define the correct nomenclature, rather than the actual distribution. Nevertheless, if either

one of the interlayer sites is dominated by an ion other than Cu2+, this would constitute a new

mineral species according to the dominant constituent rule (Hatert and Burke, 2008).

Paratacamite (sensu stricto) has a Cu dominant interlayer with the formula

(Cu,M)Cu3(OH)6Cl2, where M is a divalent metal ion. Other congeners, with total interlayer

dominance of the substituting cation corresponding to (M,Cu)Cu3(OH)6Cl2

The paratacamite R3� supercell and R3�m subcell structures are related though both

temperature and compositional factors. Accordingly, both of these parameters must be

considered when describing either structure. It must be noted that the superlattice reflections

of paratacamite become very weak in intensity near the R3� → R3�m transformation and are

easily overlooked.

, should be

defined by the dominant cation M with the substituent referenced in the name within

parentheses, e.g. paratacamite-(Ni).

115

REFERENCES Adib, D. and Ottemann, J. (1972) Ein neues mineral, (Cu,Zn)2(OH)3Cl, aus der Kali-Kafi

mine, Provinz Anarak, Zentral Iran. Neues Jahrbuch für Mineralogie, Monatshefte, 335-

338.

Agilent (2012). CrysAlis PRO and CrysAlis RED. Agilent Technologies, Yarnton, England.

Bandy, M.C. (1938) Mineralogy of three sulphate deposits of Northern Chile. American

Mineralogist, 23, 669-760.

Batlogg, B. (1991) Physical properties of high-Tc superconductors. Physics Today. 44, 44-50.

Bednorz, J.G. and Müller, K.A (1986) Possible High Tc superconductivity in the Ba – La –

Cu – O system. Zeitschrift fur Physik B:Condensed Matter. 64, 189-193.

Bertolotti, G., Bersani, D., Lottici, P.P., Alesiani, M., Malcherek, T. and Schlüter, J. (2012)

Micro-Raman study of copper hydroxychlorides and other corrosion products of bronze

samples mimicking archaeological coins. Analytical and Bioanalytical Chemistry, 402,

1451-1457.

Braithwaite, R.S.W., Mereiter, K., Paar, W.H. and Clark, A.M. (2004) Herbertsmithite,

Cu3Zn(OH)6Cl2, a new species, and the definition of paratacamite. Mineralogical

Magazine, 68, 527-539.

Burgio, L. and Clark, R.J.H. (2001) Library of FT-Raman spectra of pigments, minerals,

pigment media and varnishes, and supplement to existing library of Raman spectra of

pigments with visible excitation. Spectrochimica Acta Part A, 57, 1491-1521.

Burns, P.C., Cooper, M.A., and Hawthorne, F.C. (1995) Claringbullite: a Cu2+ oxysalt with

Cu2+ in trigonal-prismatic coordination. Canadian Mineralogist, 33, 633-639.

Burns, P.C. and Hawthorne, F.C. (1996) Static and dynamic Jahn-Teller effects in Cu2+

oxysalt minerals. The Canadian Mineralogist. 34, 1089-1105.

Carpenter, M.A., Salje, E.K.H. and Graeme-Barber, A. (1998) Spontaneous strain as a

determinant of thermodynamic properties for phase transitions in minerals. European

Journal of Mineralogy, 10, 621–691.

Chu, S. (2011) Magnetic properties of geometrically frustrated polymorphic crystals of

Cu4-xMgx(OH)6Cl2. Journal of Physics: Conference Series, 273, 012123.

Chu, S., McQueen, T.M., Chisnell, R., Freedman, D.E., Müller, P., Lee, Y.S. and Nocera,

D.G. (2010) A Cu2+ (S = ½) kagomé antiferromagnet: MgxCu4-x(OH)6Cl2. Journal of the

American Chemical Society, 132, 5570-5571.

116

Chu, S., Müller, P., Nocera, D.G. and Lee, Y.S. (2011) Hydrothermal growth of single

crystals of the quantum magnets: clinoatacamite, paratacamite and herbertsmithite.

Applied Physics Letter, 98, 092508.

Clissold, M.E., Leverett, P. and Williams, P.A. (2007) The structure of gillardite, the Ni-

analogue of herbertsmithite, from Widgiemooltha, Western Australia. Canadian

Mineralogist, 45, 317-320.

Colchester, D.M., Leverett, P., Clissold, M.E., Williams, P.A., Hibbs, D.E. and Nickel, E.H.

(2007) Gillardite, Cu3NiCl2(OH)6, a new mineral from the 132 North deposit,

Widgiemooltha, Western Australia. Australian Journal of Mineralogy, 13, 15-18.

Colman, R.H., Ritter, C. and Wills, A.S. (2008) Towards perfection: Kapellasite,

Cu3Zn(OH)6Cl2, a new model S = ½ kagomé antiferromagnet. Chemistry of Materials,

20, 6897-6899.

Colman, R.H., Sinclair, A. and Wills, A.S. (2010) Comparisons between haydeeite, α-

Cu3Mg(OD)6Cl2, and kapellasite, α-Cu3Zn(OD)6Cl2, isostructural S = ½ kagomé

magnets. Chemistry of Materials, 22, 5774-5779.

Dei, L., Ahle, A., Baglioni, P., Dini, D. and Ferroni, E. (1998) Green degradation products of

azurite in wall paintings: Identification and conservation treatment. Studies in

Conservation, 43, 80-88.

Downs, R.T. (2006) The RRUFF Project: an integrated study of the chemistry,

crystallography, Raman and infrared spectroscopy of minerals. Program and Abstracts of

the 19th General Meeting of International Mineralogical Association in Kobe, Japan.

O03-13.

Downs, R.T. and Hall-Wallace, M. (2003) The American Mineralogist Crystal Structure

Database. American Mineralogist, 88, 247-250.

Effenberger, H. (1984) Verfeinerung der Kristallstruktur von Kupfer(II)-hydroxichlorid,

Cu(OH)Cl. Monatshefte für Chemie, 115, 725-730.

Feitknecht, W. and Maget, K. (1949a) Über Doppelhydroxide und basische Doppelsalze:

über basische Doppelchloride des Kupfers. Helvetica Chimica Acta, 32, 1653-1667.

Feitknecht, W. and Maget, K. (1949b) Über Doppelhydroxyde und basische Doppelsalze.

Mischphasen von Kupfer-Nickelhydroxychloriden. Helvetica Chimica Acta, 32, 1667-

1674.

Frondel, C. (1950) On paratacamite and some related copper chlorides. Mineralogical

Magazine, 29, 34-45.

117

Frost, R.L. (2003) Raman spectroscopy of selected copper minerals of significance in

corrosion. Spectrochimica Acta, Part A, 59, 1195-1204.

Frost, R.L., Martens, W., Kloprogge, J.T. and Williams, P.A. (2002) Raman spectroscopy of

the basic copper chloride minerals atacamite and paratacamite: implications for the study

of copper, brass and bronze objects of archaeological significance. Journal of Raman

Spectroscopy, 33, 807-806.

Fleet, M.E. (1975) The crystal structure of paratacamite, Cu2(OH)3Cl. Acta

Crystallographica, B31, 183-187.

Garrels, R.M. and Christ, C.L. (1965) Solutions, Minerals and Equilibria. Harper and Row,

New York.

Garrels, R.M. and Dreyer, R.M. (1952) Mechanism of limestone replacement at low

temperatures and pressures. Geological Society of America Bulletin, 63, 325-380.

Garrels, R.M. and Stine, L.O. (1948) Replacement of calcite by atacamite in copper chloride

solutions. Economic Geology, 43, 21-30.

Grice, J.D., Szymański, J.T. and Jambor, J.L. (1996) The crystal structure of clinoatacamite, a

new polymorph of Cu2(OH)3Cl. Canadian Mineralogist, 34, 73-78.

Han, T.H., Helton, J.S., Chu, S., Nocera, D.G., Rodriguez-Rivera, J.A., Broholm, C. and Lee,

Y.S. (2012) Fractionalized excitations in the spin-liquid state of a kagomé-lattice

antiferromagnet. Nature, 492, 406-410.

Han, T.H., Helton, J.S., Chu, S., Prodi, A., Singh, D.K., Mazzoli, C., Müller, P., Nocera, D.G.

and Lee, Y.S. (2011) Synthesis and characterisation of single crystals of the spin-1/2

kagome-lattice antiferromagnets ZnxCu4-x(OH)6Cl2. Physical Review, B 83, 100402.

Hatert, F. and Burke, E.A.J. (2008) The IMA–CNMNC dominant-constituent rule revisited

and extended. Canadian Mineralogist, 46, 717-728.

Hathaway, B.J. (1987) Copper. In: Wilkinson, G., McCleverty, J.A. and Gillard, R.G. (eds)

Comprehensive Coordination Chemistry. Pergamon Press, Oxford, Volume 5, pp. 534-

774.

Hawthorne, F.C. (1985) Refinement of the crystal structure of botallackite. Mineralogical

Magazine, 49, 87-89.

Jambor, J.L., Dutrizac, J.E., Roberts, A.C., Grice, J.D. and Szymański, J.T. (1996)

Clinoatacamite, a new polymorph of Cu2(OH)3Cl, and its relationship to paratacamite

and “anarakite”. Canadian Mineralogist, 34, 61-72.

118

Kracher, A. and Pertlik, F. (1983) Zinkreicher Paratacamit, Cu3Zn(OH)6Cl2, aus der

Herminia Mine, Sierra Gorda, Chile. Annalen des Naturhistorischen Museums in Wien,

85/A, 93-97.

Kraus, W., Bernhardt, H.-J., Braithwaite, R.S.W., Koltisch, U. and Pritchard, R. (2006)

Kapellasite, Cu3Zn(OH)6Cl2, a new mineral from Lavrion Greece, and its crystal

structure. Mineralogical Magazine, 70, 329-340.

Kraus, W. and Nolze, G. (1996a) Powdercell – a program for the representation and

manipulation of crystal structures and calculation of the resulting X-ray powder patterns.

Journal of Applied Crystallography, 29, 301-303.

Kraus, W. and Nolze, G. (1996b) Powdercell 1.8 and Windows version 1.0. Federal Institute

for Materials Research and Testing (BAM), Berlin.

Langford, J.I. (1973) Least-squares refinement of cell dimensions from powder data by

Cohen’s method. Journal of Applied Crystallography, 6, 190-196.

Lee, M.S. and Nam, S.H. (2009) Chemical equilibria of nickel chloride in HCl solution at

25°C. Bulletin of the Korean Chemical Society, 30, 2203-2207.

Liu, X., Meng, D., Zheng, X., Hagihala, M. And Guo, Q. (2011) Mid-IR and Raman spectral

properties of clinoatacamite-structure basic copper chlorides. Advanced Materials

Research, 146-147, 1202-1205.

Long, D.T. and Angino, E.E. (1977) Chemical speciation of Cd, Cu, Pb, and Zn in mixed

freshwater, seawater, and brine solutions. Geochimica et Cosmochimica Acta, 41, 1183-

1191.

Lutz, H.D., Beckenkamp, K. And Möller, H. (1994) Weak hydrogen bonds in solid

hydroxides and hydrates. Journal of Molecular Structure, 322, 263-266.

Malcherek, T. and Schlüter, J. (2007) Cu3MgCl2(OH)6 and the bond-valence parameters of

the OH-Cl bond. Acta Crystallographica, B63, 157-160.

Malcherek, T. and Schlüter, J. (2009) Structures of the pseudo-trigonal polymorphs of

Cu2(OH)3Cl. Acta Crystallographica, B65, 334-341.

Malcherek, T. and Schlüter, J. (2010) Anatacamite from La Vendida mine, Sierra Gorda,

Atacama desert, Chile: a triclinic polymorph of Cu2(OH)3Cl. Neues Jahrbuch für

Mineralogie, Abhandlung, 187, 307-312.

Materials Data, Inc (2011). JADE 2010. Materials Data Incorporated, Livermore, California,

USA.

119

Mendels, P., Bert, F., de Vries, M.A., Olariu, A., Harrison, A., Duc, F., Trombe, J.C., Lord,

J.S., Amato, A. and Baines, C. (2007) Quantum magnetism in the paratacamite family:

towards an ideal kagomé lattice. Physical Review Letters, 98, 077204.

Momma, K. and Izumi, F. (2008) VESTA: a three-dimensional visualization system for

electronic and structural analysis. Journal of Applied Crystallography, 41, 653-658.

Ostwald, W.Z. (1897) Studien über die Bildung und Umwaldung fester Korper. 1.

Abhandlung: Übersattigung und Überkaltung. Zeitschrift für Physikalische Chemie, 22,

289-330.

Oswald, H.R. and Guenter, J.R. (1971) Crystal data on paratcamite γ-Cu2(OH)3Cl. Journal of

Applied Crystallography, 4, 530-531.

Oswald, H.R. and Feitknecht, W. (1964) Über Hydroxidhalogenide Me2(OH)3Cl, -Br, -J

zweiwertiger Metalle (Me = Mg, Ni, Co, Cu, Fe, Mn). Helvetica Chimica Acta, 47, 272-

289.

Parise, J.B. and Hyde, B.G. (1986) The structure of atacamite and its relationship to spinel.

Acta Crystallographica, C42, 1277-1280.

Perrin, D.D. and Sayce, I.G. (1967) Computer calculation of equilibrium concentrations in

mixtures of metal ions and complexing species. Talanta, 14, 833-842.

Pollard, A.M., Thomas, R.G. and Williams, P.A. (1989) Synthesis and stabilities of the basic

copper(II) chlorides atacamite, paratacamite and botallackite. Mineralogical Magazine,

53, 557-563.

Pollard, A.M., Thomas, R.G. and Williams, P.A. (1990) Connellite: stability relationships

with other secondary copper minerals. Mineralogical Magazine, 54, 425-430.

Pring, A., Snow, M.R. and Tiekink, E.R.T. (1987) Paratacamite from South Australia.

Transactions of the Royal Society of South Australia. 3, 127-128.

Robie, R.A. and Hemingway, B.S. (1995) Thermodynamic properties of minerals and related

substances at 298.15 K and 1 bar (105 pascals) pressure and at higher temperatures.

U.S. G.P.O., Washington and Denver, CO.

Robinson, K., Gibbs, G. V. and Ribbe, P. H. (1971) Quadratic elongation: A quantitative

measure of distortion in coordination polyhedral. Science, 172, 567-570.

Saini-Eidukat, B., Kucha, H. and Keppler, H. (1994) Hibbingite, γ-Fe2(OH)3Cl, a new

mineral from the Duluth Complex, Minnesota, with implications for the oxidation of Fe-

bearing compounds and the transport of metals. American Mineralogist, 79, 555-561.

Schlüter, J., Klaska, K.H. and Gebhard, G. (2000) Belloite, Cu(OH)Cl, a new mineral from

Sierra Gorda, Antofagasta, Chile. Neues Jahrbuch für Mineralogie, Monatshefte, 67-73.

120

Schlüter, J. and Malcherek, T. (2007) Haydeeite, Cu3MgCl2(OH)6, a new mineral from the

Haydee mine, Salar Grande, Atacama desert, Chile. Neues Jahrbuch für Mineralogie –

Abhandlungen, 184, 39-42.

Schores, M.P., Nytko, E.A., Bartlett, B.M. and Nocera, D.G. (2005) Structurally perfect S =

½ kagomé antiferromagnet. Journal of the American Chemical Society, 127, 13462-

13463.

Schreurs, A.M.M., Xian, X. and Kroon-Batenburg, L.M.J. (2010) EVAL15: a diffraction data

integration method based on ab initio predicted profiles. Journal of Applied

Crystallography, 43, 70-82.

Scott, D.A. (2000) A review of copper chlorides and related salts in bronze corrosion and as

painting pigments. Studies in Conservation, 45, 39-53.

Shannon, R.D. (1976) Revised effective ionic radii and systematic studies of interatomic

distances in halides and chalcogenides. Acta Crystallographica. A32, 751-767.

Sharkey, J.B. and Lewin, S.Z. (1971) Conditions governing the formation of atacamite and

paratacamite. American Mineralogist, 56, 179-192.

Sharkey, J.B. and Lewin, S.Z. (1972) Thermochemical properties of the copper(II)

hydroxychlorides. Thermochimica Acta, 3, 189-201.

Sheldrick, G.M. (2008) A short history of SHELX. Acta Crystallographica. A64, 112-122.

Smith, G.F.H. (1906) Paratacamite, a new oxychloride of copper. Mineralogical Magazine,

14, 170-177.

Smith, R.M. and Martel, A.E. (1976) Critical Stability Constants. Volume 4. Inorganic

Complexes. Plenum Press, New York.

Stormer, J.C. Jr., Pierson, M.L., Tacker, R.C. (1993) Variation of F and Cl X-ray intensity

due to anisotropic diffusion in apatite during electron microprobe analysis. American

Mineralogist, 78, 641-648.

Ungemach, H. (1911) L'atacamite. Bulletin de la Société Française de Minéralogie, 34, 148-

216.

de Vries, M.A., Wulferding, D., Lemmens, P., Lord, J.S., Harrison, A., Bonville, P., Bert, F.

and Mendels, P. (2012) Extension of the zinc paratacamite phase diagram: Probing the

effect of spin vacancies in an S = ½ kagome antiferromagnet. Physical Review, B 58,

014422.

Walter-Levy, L. and Goreaud, M. (1969) Sur la formation des chlorures basiques cuivriques

en solution aqueuse de 25 à 200°C. Bulletin de la Société Chimique de France, 8, 2623-

2634.

121

Wells, A.F. (1949) The crystal structure of atacamite and the crystal chemistry of cupric

compounds. Acta Crystallographica, 2, 175-180.

de Wolff, P.M. (1953) The crystal structure of Co2(OH)3Cl. Acta Crystallographica, 6, 359-

360.

Woods, T.L. and Garrels, R.M. (1986a) Phase relations of some cupric hydroxy minerals.

Economic Geology, 81, 1989-2007.

Woods, T.L. and Garrels, R.M. (1986b) Use of oxidized copper minerals as environmental

indicators. Applied Geochemistry, 1, 181-187.

Wulferding, D., Lemmens, P., Scheib, P., Röder, J., Mendels, P., Chu, S., Han, T. and Lee,

Y.S. (2010) Interplay of thermal and quantum spin fluctuations in the kagome lattice

compound herbertsmithite. Physical Review, B 52, 144412.