ASPECTS OF THE SUPERGENE
GEOCHEMISTRY OF COPPER,
NICKEL AND BISMUTH
Meagan E. Clissold BAppSc(Hons), UWS
This thesis is submitted for the degree of
Doctor of Philosophy
in the University of Western Sydney
July 2007 ACKNOWLEDGEMENTS
I wish to sincerely thank my supervisor Professor Peter Williams for sharing his wealth of knowledge in chemistry and mineralogy. His enthusiasm, guidance and support through out this study are greatly appreciated. I would also like to thank my co-supervisor Professor Peter Leverett, for his invaluable insight in many fields, especially crystallography. Jim Sharpe is thanked for his assistance with fieldwork and for sharing his experience in mineralogy. David Colchester is acknowledged for determining the optical properties of gillardite.
Auzex Resources Limited is thanked for permission to work on the Wolfram pipe deposit. I am grateful for the assistance and interest of company staff, particularly Dr. Roger Mustard, Kevin Chard and Ed Hammond. Northparkes Mines is thanked for permission to work on their deposits. Allan Ryan of Northparkes
Mines provided details concerning groundwater analyses. Professor Erik Melchiorre
(California State University, San Bernardino) is thanked for stable isotope data.
Valuable technical assistance was provided by Malcolm Tiddy and Wayne
Higginbotham. I would like to thank the people who have helped and encouraged me throughout this work, especially Adam McKinnon, Jacqueline, Sarah, Suzie, Sandy and Tanya. I thank my fellow students, my friends, and family for their love and support.
ii DECLARATION
The work presented in this thesis is, to the best of my knowledge and belief, original except as acknowledged in the text. I hereby declare that I have not submitted this material in any form for a degree or diploma at any university or other institution of tertiary education.
……………………………
Meagan Elizabeth Clissold
July, 2007
iii TABLE OF CONTENTS
Page
SUPPORTING PUBLICATIONS……………………...…………...viii
ABSTRACT…………………………………………………………... ix
Chapter 1: INTRODUCTION………………………………...... 1 1.1 AIMS OF THIS STUDY………………………………………………... 4 1.1.1 Northparkes study………………………………..……………. 4 1.1.2 Lavendulan study………………………….…….………...... 5 1.1.3 Kingsgate Bi study……………….……………..………..……. 6
Chapter 2: GEOCHEMICAL EVOLUTION OF THE OXIDISED ZONES OF THE COPPER ORE BODIES AT NORTHPARKES, NEW SOUTH WALES, AND ITS RELEVANCE FOR GEOCHEMICAL EXPLORATION……………………………….... 8 2.1 SETTING………………………………………………………………... 8 2.2 EXPERIMENTAL……………………….…………………………….. 11 2.2.1 Instrumentation………………………………………………. 11 2.2.2 Synthesis …………………………………………………….. 12 2.2.3 Solution studies………………………………………………. 12 2.3 GEOCHEMISTRY OF DEVELOPMENT OF THE OXIDISED ZONES AT NORTHPARKES…………………………………………….. 15 2.3.1 Phosphate minerals…………………………………………... 15 2.3.2 Atacamite, malachite and azurite…………………………….. 25 2.4 GEOCHEMICAL DISPERSION OF COPPER FROM THE OXIDISED ZONES AT NORTHPARKES…………………….…………. 28
iv Chapter 3: THE GEOCHEMISTRY OF FORMATION OF LAVENDULAN…………………………………………..………….. 32 3.1 BACKGROUND……………………………………….……………… 32 3.2 EXPERIMENTAL…………………………………….……………….. 34 3.2.1 Powder X-ray diffraction (XRD)………………………...... 34 3.2.2 Total reflectance X-ray fluorescence (TRXRF) spectrometry……………………………………...……... 34 3.2.3 Scanning electron microscopy (SEM)……………...………... 35 3.2.4 Electron microprobe analysis…………………….…………... 35 3.2.5 Atomic absorption spectroscopy (AAS)……....……………... 35 3.2.6 Lavendulan synthesis ………………………………………... 36 3.2.7 Solution studies……………………………..………………... 36 3.3 GEOCHEMISTRY OF MINERAL FORMATION…………………… 39 3.3.1 Phase relationships…………………………………………… 39 3.4 SOLID SOLUTION PHENOMENA INVOLVING LAVENDULAN AND RELATED MINERALS………………………….. 47
Chapter 4: GILLARDITE, Cu3NiCl2(OH)6, A NEW MINERAL FROM THE 132 NORTH DEPOSIT, WIDGIEMOOLTHA, WESTERN AUSTRALIA……………………….………………….. 54 4.1 HISTORY…………………………………...…………………………. 54 4.2 OCCURRENCE………………………………….…………………….. 55 4.3 CHEMICAL COMPOSITION………………………………….……... 56 4.4 PHYSICAL AND OPTICAL PROPERTIES………………………….. 57 4.5 SINGLE-CRYSTAL X-RAY STRUCTURE………………………….. 58 4.6 X-RAY POWDER DIFFRACTION…………………………………… 62 4.7 RELATIONSHIP OF GILLARDITE TO HERBERTSMITHITE AND OTHER POLYMORPHS……………………………………..……... 62
v Chapter 5: THE BISMUTH MINERALS OF THE KINGSGATE, NEW SOUTH WALES, DEPOSITS AND THE SUPERGENE DISPERSION OF BISMUTH……………….. 66 5.1 INTRODUCTION…………………………………………...………… 66 5.2 EXPERIMENTAL METHODS.………………………..…….………... 69 5.2.1 Powder X-ray diffraction (XRD)…………...………………... 69 5.2.2 Scanning electron microscopy (SEM)…….……..…………... 69 5.2.3 Electron microprobe analysis………………………………… 70 5.3 A SURVEY OF SECONDARY Bi AND Mo MINERALS FROM AUSTRALIAN COLLECTIONS………………………………………….. 71 5.4 THE Bi-Mo DEPOSITS OF KINGSGATE…..………..……….……... 76 5.4.1 Primary mineralogy………………..…………...……………. 80 5.4.1.1 Bismuth………………………….…………………………. 82 5.4.1.2 Bismuthinite……………………..…………………………. 83 5.4.1.3 “Cannizzarite”………………….…………………………... 86 5.4.1.4 Cosalite………………………...…………………………... 86 5.4.1.5 Galenobismutite……………………………………..……... 86 5.4.1.6 Ikunolite…………………………………………….……… 86 5.4.1.7 Joséite………………………………………….…………… 88 5.4.1.8 Kobellite……………………………………..……………... 88 5.4.1.9 Molybdenite………………………………………………... 93 5.4.1.10 Ag-rich Pb-Bi sulfides…………………….……………… 93 5.5 THE “LOST” MINES OF KINGSGATE……………...………………. 97 5.5.1. Minerals from Maurer’s Claim…………...………………... 103 5.6 A GEOCHEMICAL MODEL FOR BISMUTH IN THE SUPERGENE ENVIRONMENT……………………...…………………. 107 5.6.1. The pH regime of the supergene zone at Kingsgate……….. 107 5.6.2. A model for Bi solubility and dispersion…………………... 110 5.6.3 Exploration implications and previous exploration campaigns in the Kingsgate region…………………………...... 119
vi REFERENCES……………………………………………………... 123
APPENDIX…………………………………………………………. 144 Table A.1 COMICS output……..………………………………… 145 Table A.2 Observed and calculated structure factors for gillardite. 146 Table A.3 Crystallographic Information File (CIF) for gillardite… 151
vii SUPPORTING PUBLICATIONS
Clissold, M.E., Leverett, P. and Williams, P.A. (2003) Gaspéite-magnesite solid
solutions and their significance. CRC LEME Regional Regolith Symposia,
2003. CRC LEME, 78-79.
Clissold, M.E., Leverett, P. and Williams, P.A. (2005) Chemical mineralogy of the
oxidized zones of the E22, E26 and E27 ore bodies at Northparkes, New
South Wales. In: Roach, I.C. (Ed.), Regolith 2005 - Ten Years of CRC LEME,
Proceedings. CRC LEME, 55-58.
Clissold, M.E., Leverett, P., Williams, P.A., Hibbs, D.E. and Nickel, E.H. (2007)
The structure of gillardite, Cu3NiCl2(OH)6, from Widgiemooltha, Western
Australia: the Ni-analogue of herbertsmithite. The 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, 21-24.
viii ABSTRACT
The solution geochemical conditions associated with the development of supergene copper mineralisation in the E22, E26 and E27 deposits at Northparkes,
New South Wales, have been explored. Determination of a stability constant for sampleite [NaCaCu5(PO4)4Cl·5H2O], a conspicuous species in the upper oxidised zone of E26, has led to an understanding of the differences between the three deposits in terms of the influence of groundwater geochemistry on their mineralogical diversity. Modelling of copper dispersion from the three deposits using current ground water compositions as proxies for past solution conditions has shown that the elevated chloride concentrations associated with E26 have negligible influence on total dissolved copper concentrations over a wide pH range. The results are discussed with respect to applications in exploration geochemistry for the discovery of new ore deposits in the region.
Determination of a stability constant for lavendulan
[NaCaCu5(AsO4)4Cl·5H2O], the arsenate isomorph of sampleite, suggests that solid solution between lavendulan and sampleite is likely to be extensive and this has been established by reference to mineral compositions from a number of deposits.
Activity-activity phase diagrams have been developed to explain the common mineral associates of lavendulan and differences between the analogous phosphate and arsenate systems. With respect to the occurrence of lavendulan in the oxidised zone of the Widgiemooltha 132 N ore body, Western Australia, its crystal chemistry explains why Ni does not substitute for Cu in the lattice. This is despite Ni being abundantly available in the deposit and substituting freely into other copper-based minerals. The substitution of Ni for Cu was explored in a study of supposedly Ni-rich paratacamite, Cu2Cl(OH)3, from the deposit. It transpires that much of this is a new
ix mineral, gillardite, Cu3NiCl(OH)6, the isomorph of herbertsmithite, Cu3ZnCl(OH)6.
The nature of gillardite was thoroughly investigated and the mineral was approved as a new species by the International Mineralogical Association. A high resolution single-crystal X-ray structure of gillardite has been completed. In addition, the substitution of Ni in simple carbonate lattices has been explored as gaspéite, NiCO3,
Ni-rich magnesite, MgCO3, and calcite, CaCO3, are all common species in the oxidised zone of the Widgiemooltha 132 N deposit.
Attention was subsequently focussed on the geochemistry of the element Bi, with special reference to deposits of the Kingsgate region, New South Wales. This study has led to a modern assessment of the Mo-Bi deposits in the area and new Bi sulfosalts from the Wolfram pipe at Kingsgate are described. A survey of secondary
Bi minerals from a host of deposits has led to the development of a model for the dispersion of Bi in the supergene environment, which will have widespread applications in exploration geochemistry where Bi is used as a pathfinder element.
Calculations of aqueous Bi species in equilibrium with bismite, Bi2O3, bismoclite,
BiOCl, and bismutite, Bi2O2CO3, over a wide pH range show that the element is very insoluble under ambient oxidising conditions. It is noted that the results of previous geochemical exploration campaigns in the region will have to be reassessed.
x Chapter 1
INTRODUCTION
When primary sulfide deposits that formed deep in the Earth’s crust under comparatively reducing conditions are later exposed to oxidising conditions above the water table, a series of reaction takes place. Many minerals that were stable in the primary environment become thermodynamically unstable in the secondary environment and react. In general the process is one of oxidation and for the most part elements achieve their highest oxidation states (Williams, 1990). New assemblages of secondary minerals can be formed in the oxidised zone above the water table and just which species form depends on the local chemical environment.
In terms of pH and redox potential, two key physicochemical parameters that describe the geochemical setting of aqueous systems, a remarkably wide range of conditions may develop. These span the whole range of the redox stability of water as defined by equations (1.1) and (1.2) (Appelo and Postma, 1993; Garrels and
Christ, 1965).
+ - O2(g) + 4H (aq) + 4e ⇋ 2H2O(l) (1.1)
+ - H (aq) + e ⇋ 1/2H2(g) (1.2)
The range of achievable pH conditions has its own special limits. Oxidation of pyrite,
3+ FeS2, and the subsequent hydrolysis of Fe (aq) ions generates sulfuric acid as shown in equations (1.3) to (1.5).
1 2+ 2- + FeS2(s) + 7/2O2(g) + H2O(l) → Fe (aq) + 2SO4 (aq) + 2H (aq) (1.3)
2+ + 3+ Fe (aq) + 1/4O2(g) + H (aq) → Fe (aq) + 1/2H2O(l) (1.4)
3+ + Fe (aq) + 2H2O(l) ⇋ FeOOH(s, goethite) + 3H (aq) (1.5)
Subsequent evaporation can give rise to solutions that are up to 6 F (formal) in sulfuric acid. The mineralogical evidence for this is apparent from a consideration of the Fe2O3-SO3-H2O system (Posnjak and Merwin, 1922). The mineral kornelite,
HFe(SO4)2·4H2O, found in certain oxidised assemblages derived from pyrite-rich sulfides, can only crystallise from solution under ambient temperatures and pressures when sulfuric acid concentrations are between about 5 and 7 F. On the basic side of neutral, the limits of the natural environment are more constrained. Carbonate buffering [CO2(g)-CaCO3(s)-H2O(l)] serves to limit the pH of aqueous solutions to
8.3 at 298.2 K and 105 Pa (Appelo and Postma, 1993; Garrels and Christ, 1965).
Higher pH values, up to ca 12.5 are much more rarely met as the result of the hydrolysis of Fe-poor silicates, e.g. the hydrolysis of fayalite as shown in equation (1.6).
2+ o - Mg2SiO4(s) + 4H2O(l) ⇋ 2Mg (aq) + Si(OH)4 (aq) + 4OH (aq) (1.6)
This serves to explain the enormous diversity of secondary mineral assemblages (Anthony et al., 1990, 1995, 1997, 2000, 2003) that form under ambient surface or near-surface conditions. The lower temperatures permit the crystallisation of phases (stable or metastable) that would be unstable at higher temperatures.
In general the oxidised equivalents of primary minerals are much more soluble in aqueous solution and their elements are dispersed from oxidising sulfide
2 ore bodies. This gives rise to a classical zoning (Williams, 1990) from the surface down to the water table characterised by a gossan outcrop composed of chemical resistates and oxyhydroxides of Fe and Mn. Beneath this a leached zone is often encountered, in which base metal ions have been depleted, followed by the so-called
“oxidised zone” where secondary base metal minerals are concentrated. At the water table conditions become reducing and certain elements, notably Cu, Ag and Au, become enriched (the supergene enriched zone). This latter zone has long been recognised and targeted in terms of high grade ores of a number of elements
(Emmons, 1917).
The detection of the dispersed elements in waters and the regolith forms the basis of geochemical exploration methods used to discover new sources of ores.
Modern methods have their origins in the work of Soviet geoscientists starting around 1930 and then taken up by their western counterparts. Hawkes and Webb
(1962) give a fine review of this early work. It is now highly unusual for any exploration program not to employ geochemical methods. Nevertheless, despite some eight decades of research, the chemistry of element dispersion in the secondary environment is not fully understood and for some elements information is virtually non-existent. To substantiate this claim, one could point to current research effort centred in the Cooperative Research Centre for Landscape Environments and
Mineral Exploration in Australia (http://crcleme.org.au/). Key areas for research for many elements are:
• How they are transported in solution?
• What is their chemical speciation?
• What controls secondary mineral formation, adsorption onto other mineral
surfaces and other processes involving mobilisation and concentration?
3 1.1 AIMS OF THIS STUDY
This thesis explores some aspects of the chemistry of Cu, Bi and Ni in the supergene zone near oxidising ore bodies. It builds on work of others to develop some original contributions to the knowledge of the geochemistry of the above and related elements found in such settings. An underlying theme concerns the apparent simple notion that each secondary mineral found near oxidising systems reflects the chemistry of the solutions from which it crystallised. Thus, by reading the
“geochemical signature”, past geochemical events can be reconstructed using appropriate thermodynamic and equilibrium data. An equilibrium approach was brought to bear on selected systems, to reach a surprising number of conclusions and outcomes. The methodology is not immediately obvious, and so first the problems tackled are outlined and the geochemical outcomes of the study are highlighted.
1.1.1 Northparkes study
A geochemical study of the oxidised zones of porphyry-style Cu-Au deposits at Goonumbla (Northparkes), New South Wales, was undertaken. Both primary and secondary assemblages at Goonumbla have been thoroughly described in the past
(Crane et al., 1998; Heithersay et al., 1990; McLean et al., 2004; Tonui et al., 2002).
In the three ore bodies that have been exploited, E22, E26 and E27, upper oxidised zones are dominated by the secondary Cu(II) phosphate minerals libethenite,
Cu2PO4(OH), and pseudomalachite, Cu5(PO4)2(OH)4, and, uniquely to E26, the otherwise rare mineral sampleite, NaCaCu5(PO4)4Cl·5H2O. The reason for the abundance sampleite in E26 has been traced to its emplacement under rather more saline (NaCl) conditions than those in other two deposits. The determination of a
o value for the standard free energy of formation (ΔfG ) of sampleite at 298.2 K was a
4 key to understand the chemical evolution of the secondary copper mineralisation of the region. By using current ground water compositions as proxies for ground waters of the past it became possible to reconstruct the geochemical history of the mineralising supergene environment of previous times. In particular, an assessment of copper dispersion in the different (saline) settings helped to discern whether they influence dissolved copper loadings. Little difference is evident from modelling aqueous solution chemistries, which implies that the supergene geochemical
“footprint” of the systems is independent of ground water compositions. The system was compared with other deposits rich in secondary copper phosphates (Chapman et al., 2005; Crane et al., 2001) to provide an explanation for the origin of the gypsum that blankets the Goonumbla deposits provided. The conclusions of the study have been incorporated in the geochemical exploration strategy of Northparkes
Mines.
1.1.2 Lavendulan study
Lavendulan, NaCaCu5(AsO4)4Cl·5H2O, is the arsenate isomorph of sampleite
o (Giester et al., 2007). A value for ΔfG of lavendulan at 298.2 K has also been determined. The value suggests that solid solution between lavendulan and sampleite is likely to be extensive and this was established by reference to mineral compositions from a number of deposits. The incorporation of Pb for Ca in the lattice was explored. In terms of known mineral species, the CuO-As2O5-H2O system is more complex than the CuO-P2O5-H2O system and this was investigated in terms of the different mineral associates of lavendulan reported in the literature (Birch, 1990;
Giester et al., 2007; Guillemin, 1956; Kleeman and Milnes, 1973; Kokinos and Wise,
1993; Nickel et al., 1994; Ryall and Segnit, 1976; Segnit, 1976). An interesting
5 feature of the extent of the stability field of lavendulan in relation to that of sampleite is that lavendulan is restricted by conditions that give rise to a common associate, conichalcite, CaCuAsO4(OH). Conichalcite has no phosphate counterpart. However, lavendulan is found more commonly than sampleite as a result of the generation of arsenate during the oxidation of common As-bearing sulfide ores.
In regard to lavendulan in the Widgiemooltha 132 N deposit, Western
Australia (Nickel et al., 1994), its crystal chemistry (Giester et al., 2007) explains why Ni does not substitute for Cu in the lattice. This is despite Ni being abundantly available in the deposit and substituting freely in magnesite, MgCO3, calcite, CaCO3, and other secondary copper minerals. An assessment of the energetics of these systems was made.
The substitution of Ni for Cu was further appraised in a study of Ni-rich paratacamite, Cu2Cl(OH)3, space group R3 (Fleet, 1975), from the Widgiemooltha
132 N deposit (Nickel et al., 1994). The material is now a new mineral, approved by the IMA as gillardite (Colchester et al., 2007). Gillardite is rhombohedral, space group R3 m, and has an ideal stoichiometry of Cu3NiCl2(OH)6. Its single-crystal X- ray structure was determined (Clissold et al., 2007) and it is the Ni-analogue of herbertsmithite, Cu3ZnCl2(OH)6 (Braithwaite et al., 2004).
1.1.3 Kingsgate Bi study
Attention was focussed on the geochemistry of the element Bi, with special reference to deposits of the Kingsgate region, New South Wales. This led to a modern assessment of the Mo-Bi deposits situated south of the Yarrow River
(Andrews, 1916; Garretty, 1943, 1944a,b), and new Bi sulfosalts from the Wolfram pipe on the main field are described. A survey of the relative frequency of occurrence
6 of secondary Bi minerals from a host of deposits was made to determine which phases are important in developing a model for the dispersion of Bi in the supergene environment. For the first time, a coherent chemical explanation for Bi mobility has been outlined, based on the solution chemistry of the element and the stabilities of bismite, Bi2O3, bismutite, Bi2O2CO3, and bismoclite, BiOCl.
The ramifications of the model in terms of past exploration geochemical campaigns (MacDuff and Snow, 1971; MacDuff and Zerwick, 1971a,b, 1972) are discussed and ways to elaborate it outlined. The model will have widespread applications in exploration geochemistry where Bi is used as a pathfinder element
(Hale, 1981; Lang and Baker, 2001; Leverett et al., 2004; Lueth, 1999). The usefulness of the model is reflected in its adoption in both regional and local geochemical exploration in the New England region of New South Wales by
Auzex Resources Limited.
7 Chapter 2
GEOCHEMICAL EVOLUTION OF THE OXIDISED
ZONES OF THE COPPER ORE BODIES AT
NORTHPARKES, NEW SOUTH WALES, AND ITS
RELEVANCE FOR GEOCHEMICAL EXPLORATION
2.1 SETTING
Geopeko Wallsend Operations Ltd discovered several porphyry copper-gold deposits at Northparkes in central New South Wales (Figure 2.1.1) in 1977, during a regional reconnaissance drilling survey. They are associated with finger-like bodies of 430-440 Ma quartz-monzonite porphyries, which intrude the Ordovician age
Goonumbla Volcanics (Heithersay et al., 1990). Disseminated, fracture- and vein- controlled mineralisation occurs in the monzonite and the volcanics, but is focussed on quartz stock work veining associated with the monzonite. The major primary minerals are pyrite, bornite and chalcopyrite, together with economically significant native gold. Minor amounts of primary digenite, chalcocite, galena, sphalerite, covellite, tennantite, gersdorffite, calaverite, and other tellurides and selenides are also present. Sphalerite and galena occur in quartz veins associated with a later faulting event. Quartz is the only significant gangue mineral. Both primary and secondary ores from three pipe-like bodies, E22, E27 and E26 North (E26) have been worked for copper and gold by Northparkes Mines since 1994.
8
Figure 2.1.1. Location of the Northparkes deposits, NSW (ca 32o 55′S, 148o02′E).
Outcrop is poor over the three deposits, which are covered by a mixture of clays, some of which are in situ weathering products, and transported overburden
(Arundell, 2004; Heithersay et al., 1990). The base of oxidation extends up to 80 m from the surface, with strong kaolinisation of the monzonite from approximately 5 to
45 m below surface. There is little evidence of gossan development over the deposits, owing to the disseminated nature of the primary mineralisation. The region has experienced a prolonged and episodic cycle of weathering. Saprolite and transported sediments over E22 and E27 have been dated using paleomagnetic methods and two significant weathering events have been identified, one in the Carboniferous at around 325 Ma, and another in the Cainozoic, from about 60 Ma and presumably
9 continuing to the present (O'Sullivan et al., 2000; Pillans et al., 1999). The latter is one of two major periods of comparatively recent weathering around Australia, the other occurring around 12-16 Ma (McQueen et al., 2002; Pillans, 2004); the latter reference concentrates on the weathering history of the neighbouring Cobar region.
As a result of weathering, E22, E26 and E27 deposits display well-developed oxidised zones, with intense leaching of the upper sections giving rise to pervasive kaolinite development, especially in E22 and E27 (Crane et al., 1998; McLean et al.,
2004; Tonui et al., 2002). Upper oxidised zones are dominated by the secondary
Cu(II) phosphate minerals libethenite (Cu2PO4(OH)) and pseudomalachite
(Cu5(PO4)2(OH)4) and, uniquely to E26, sampleite (NaCaCu5(PO4)4Cl·5H2O). In
E26, this phosphate-rich zone came closer to the surface and the three phosphate minerals are common in material removed from the top of the ore body prior to block caving operations. Beneath the phosphates a zone dominated by malachite
(Cu2CO3(OH)2), azurite (Cu3(CO3)2(OH)2) and chrysocolla (ca CuSiO3·nH2O), the former carrying coarse secondary gold, gave way at depth to a thin native copper- cuprite (Cu2O)-chalcocite (Cu2S) supergene enriched zone. Chrysocolla is less abundant in E26. Again, E26 is exceptional in that the formation of the secondary
Cu(II) carbonates was preceded by extensive deposition of atacamite (Cu2Cl(OH)3).
A controlling factor in the development of the very different oxidised zone in E26 to those in E22 and E27 is groundwater composition; this is reflected in turn by the compositions of present ground waters (McLean et al., 2004), with those associated with E26 being much more saline (NaCl).
This research has focussed on the reconstruction of the solution conditions associated with the formation of oxidised base metal ores and the control that they may exert on the dispersion of metallic elements and hence on geochemical
10 anomalies in the regolith. A study of the New Cobar deposit (Leverett et al., 2004) showed that this approach could be applied to explain how elements were differentially mobilised during oxidation of sulfides and the relative extent of their transport away from primary ore zones. This study extends this approach to the deposits at Northparkes and discusses its application in exploration geochemistry.
Comparison is made with the oxidised zone uncovered in the main pit at
Girilambone, New South Wales (Chapman et al., 2005; Fogarty, 1998; Shields,
1996). In order to model the environment of phosphate mineral deposition in E26, a stability constant for sampleite at 298.2 K was determined. This bears on the development of the secondary Cu(II) phosphates in an acid, saline environment.
2.2 EXPERIMENTAL
2.2.1 Instrumentation
Powder X-ray diffraction (XRD) measurements were carried out using a
Philips PW1925-20 powder diffractometer (Cu Kα radiation with pure Si as internal standard). Complementary analyses, where appropriate, were carried out using a
JEOL JXA 840 microanalyzer scanning electron microscope (SEM) equipped for energy-dispersive X-ray analysis (EDAX), or a total reflectance X-ray fluorescence
(XRF) instrument of in-house design using an optically flat quartz sample support,
Mo Kα radiation and an Amptek XR-100CR Peltier-cooled X-ray detector. Electron microprobe analyses were carried out using a JEOL JXA-8600 electron microprobe analyzer equipped with wavelength dispersive spectrometers, analysing from Be to
U, and a Kevex energy dispersive spectrometer, analysing from Na to U. Atomic absorption spectrophotometry (AAS) measurements (Cu) were carried out using a
Perkin Elmer AAnalyst 100 instrument using matched standards consisting of
11 -3 CuCl2·2H2O in 0.1 mol dm NaCl prepared using pre-calibrated glassware. A
Radiometer PHM 93 meter fitted with an IJ44 combination electrode was used for pH measurements.
2.2.2 Synthesis
Pure sampleite was synthesised as follows. CaCl2 (0.1110 g, 1 mmol),
CuCl2·2H2O (0.8520 g, 5 mmol), Na2PO4 (0.5678 g, 4 mmol) and NaCl (5.843 g, 0.1 mol) were added to a Teflon-lined Parr acid digestion bomb together with 75 cm3 of deionised water. The bomb was heated to 100oC for 48 hours and then cooled slowly to room temperature. The product was collected at the pump, washed with water, then acetone, and sucked dry. Use of lesser amounts of NaCl resulted in mixtures of sampleite and libethenite. Gram quantities of sampleite were easily obtained using the above procedure. No other phases were detected in the product using XRD and satisfactory analyses corresponding to the stoichiometric formula were obtained.
2.2.3 Solution studies
Preliminary studies showed that sampleite dissolves incongruently in aqueous
10-3 mol dm-3 HCl solution to give mixtures of sampleite and libethenite. However, the mineral does dissolve congruently in aqueous 10-3 mol dm-3 HCl solution that is also 0.1 mol dm-3 in NaCl. Excess solid sampleite was equilibrated in several sealed flasks with 100 cm3 of aqueous 10-3 mol dm-3 HCl plus 0.1 mol dm-3 NaCl. The flasks were placed in a shaking water bath and maintained at 25 ±0.1oC. Parallel experiments were conducted to monitor pH changes as the solutions reached equilibrium with sampleite. No pH change was observed after two weeks, but solutions were kept in the water bath for a further ten weeks. At this time, the pH of
12 each solution was recorded and the sample filtered through a 0.2 m fibreglass disc prior to measurement of total dissolved Cu. For the six experimental runs undertaken, total dissolved Cu concentrations (x 10-4 mol dm-3) with the equilibrium pH values in parentheses were 6.07(4.551), 5.95(4.550), 6.18(4.550), 6.25(4.551),
6.18(4.551) and 6.01(4.550). This leads to a solubility of sampleite,
-4 -3 NaCaCu5(PO4)4Cl·5H2O, of 1.22 ±0.03 x 10 mol dm at 298.2 ±0.1 K and a pH of
4.550 ±0.001.
In order to calculate a stability constant for sampleite, a species distribution calculation using the average solubility of the mineral was carried out, together with similar calculations that took into account the error in the solubility. Species incorporated in the COMICS calculations (Perrin and Sayce, 1967) were Na+, Ca2+,
2+ 3- 2- - o - + + o + Cu , PO4 , HPO4 , H2PO4 , H3PO4 , Cl , CuOH , CaOH , CaHPO4 , CaH2PO4 ,
+ o + CuCl , CuHPO4 , and CaH2PO4 . Stability constants (Martell and Smith, 1982;
Smith and Martell, 1976, 1989) were corrected to an ionic strength of I = 0.102 mol dm-3 (shown in Table 2.2.3.1) using the Davies modification of the Debye-Hückel equation, taking γ for neutral species to be equal to unity, at 298.2 K (lgγi = -
2 ½ ½ 0.5085zi [I /(1 + I ) –0.3I], where zi is the charge on species i. Calculated
Table 2.2.3.1 Equilibrium constants for complex species when I = 0.1 used in determining species distributions. Species Log K CuOH+ -8.02 CaOH+ -13.02 2- HPO4 11.85 - H2PO4 18.67 o H3PO4 20.78 o CaHPO4 13.73 + CaH2PO4 19.64 CuCl+ -0.03 o CuHPO4 15.17 + CuH2PO4 19.98 Values calculated from Martell and Smith (1982); Smith and Martell (1976, 1989).
13
+ 2+ 2+ - equilibrium concentrations of the free aquated ionic species Na , Ca , Cu , H2PO4 and Cl- were 0.1001, 1.212 x 10-4, 5.503 x 10-4, 4.537 x 10-4 and 0.1011 mol dm-3, respectively, for the average solubility of sampleite at pH 4.550. These values, together with lgγ± = -0.108 and lgγ2± = -0.430 (I = 0.102 mol dm-3) were used in turn to calculate KH, with respect to the equations below, where a(i) is the activity of species i (a(i) = m(i)γ(i), with m(i) being the concentration of species i).
+ + 2+ 2+ NaCaCu5(PO4)4Cl·5H2O(s) + 8H (aq) ⇋ Na (aq) + Ca (aq) + 5Cu (aq)
- - + 4H2PO4 (aq) + Cl (aq) + 5H2O(l) (2.1)
+ 2+ 2+ 5 - 4 - + 8 KH = a(Na )a(Ca )a(Cu ) a(H2PO4 ) a(Cl )/a(H ) (2.2)
At 298.2 K, lgKH is equal to –2.42 ±0.47. From this result,
o -1 ΔfG (NaCaCu5(PO4)4Cl·5H2O,s,298.2K) = -6342.8 ±7.4 kJ mol . With respect to this quantity, values for the free energies of formation of component ions and water
- were taken from Robie and Hemmingway (1995), except for H2PO4 (aq). The value
- -1 used for H2PO4 (aq) at 298.2 K (-1130.6 ±0.9 kJ mol ) was derived from a value for
3- the corresponding PO4 ion (Robie and Hemingway, 1995), together with the first and second proton association constants given by Smith and Martell (1976). The total error in ΔfGo for sampleite was calculated by the addition of the errors for individual species (Robie and Hemingway, 1995) and the calculated error in log KH. It is noted that no error is listed for ΔfGo for Ca2+(aq) and this is estimated to be 0.1 kJ mol-1.
14 2.3 GEOCHEMISTRY OF DEVELOPMENT OF THE OXIDISED
ZONES AT NORTHPARKES
2.3.1 Phosphate minerals
The oxidised zones at Northparkes developed as a result of a series of discontinuous events. Secondary Cu(II) phosphates are present in the upper sections of all three oxidised zones and represent an early phase of mineralisation.
Thermodynamic data are available for libethenite, pseudomalachite and cornetite
(Cu3PO4(OH)3), although the last is not observed at Northparkes (Magalhães et al.,
1986, 1988), and the conditions under which they form has been explored (Crane et al., 2001). Thermodynamic data for associated phosphate minerals have been listed in Table 2.3.1, along with similar data for relevant dissolved species.
Figure 2.3.1 shows that libethenite, common to all three oxidised zones at
Northparkes, can only form with elevated activities of dissolved phosphate species and at rather low pH values at 298.2 K. The diagram was constructed using a value for a(Na+) equal to that of the chloride ion, and 10 times that of the calcium
Table 2.3.1. ΔfGº (298.2 K) for minerals and dissolved species used in geochemical modelling calculations. Species Formula ΔfGº(kJmol-1) Sourcea Cu3(PO4)2 -2051.3 1 Libethenite Cu2PO4(OH) -1224.4 ± 3.2 1 Pseudomalachite Cu5(PO4)2(OH)4 -2832.8 ± 5.0 1 Cornetite Cu3PO4(OH)3 -1603.0 ± 2.8 1 Conichalcite CaCuAsO4(OH) -1471.7 ± 3.3 1 Atacamite Cu2Cl(OH)3 -667.55 1 Sampleite NaCaCu5(PO4)Cl·5H2O -6342.8 ± 7.4 2 Cu2+ +65.1 3 Ca 2+ -553.6 3 Na+ -261.5 3 Cl- -131.2 3 - H2PO4 -1130.61 4 a[1] Williams (1990); [2] this study; [3] Robie and Hemingway (1995); [4] Robie and Hemingway (1995); Smith and Martell (1976).
15 23456789pH
Sampleite - -6 NaCl saturation a (H2PO4 )= 10
Atacamite 0 - -5 a (H2PO4 )= 10
- -4 a (H2PO4 )= 10 ) - -2 (Cl a lg
Pseudomalachite Cornetite
-4 Libethenite
-6
Figure 2.3.1. Relationships between the secondary copper(II) minerals at 298.2 K. The diagram was constructed using a value for a(Na+) equal to that of the chloride ion, and 10 times that of the calcium ion (see text). While an atacamite field is - expressed at the lowest H2PO4 activity, its crystallisation would only occur at untenably high Cu2+ activities (it is soluble at low pH).
ion (approximately in line with the current groundwater geochemistry associated with E26; see below). Thermochemical data for water and dissolved ions were taken from the literature (Martell and Smith, 1982; Robie and Hemingway, 1995; Smith and Martell, 1976, 1989).
In order to mobilise sufficient phosphate for this to occur, mineralising solutions must be acidic enough to begin to dissolve accessory fluorapatite
(Ca5(PO4)3F). Data for fluorapatite (Robie and Hemingway, 1995) indicate that this is only feasible under acidic conditions, most probably associated with the oxidation
16 of sulfides, and the hydrolysis of Fe(III) to form goethite, as illustrated in equation
(2.3) below for the reaction of pyrite.
+ 2- FeS2(s) + 15/4O2(g) + 5/2H2O(l) → FeOOH(s) + 4H (aq) + 2SO4 (aq) (2.3)
As protons were consumed by reaction with, for example, feldspars to give clays (see equations below; similar reactions may be written for micas and other silicate phases, with montmorillonite forming as an intermediate), particularly kaolinite, leached copper and phosphate would react to form libethenite and pseudomalachite in turn.
The latter phase is stable over a much wider range of pH and phosphate activities and is relatively abundant in all the Northparkes upper oxidised zones. Additional evidence for the neutralisation of acidic phosphate-bearing solutions with time is found in the crystallisation of late secondary hydroxylapatite on libethenite and pseudomalachite in E22 and E27 (Crane et al., 1998).
+ + 3KAlSi3O8(s) + 12H2O(l) + 2H (aq) → KAl3Si3O10(OH)2(s) + 6Si(OH)4(aq) + 2K (aq) orthoclase illite (2.4)
+ + 2NaAlSi3O8(s) + 9H2O(l) + 2H (aq) → Al2Si2O5(OH)4(s) + 4Si(OH)4(aq) + 2Na (aq) albite kaolinite (2.5)
+ 2+ CaAl2Si2O8(s) + H2O(l) + 2H (aq) → Al2Si2O5(OH)4(s) + Ca (aq) anorthite kaolinite (2.6)
In addition, the release of dissolved silicic acid during kaolinisation of the leached zones at Northparkes as protons are consumed is mirrored in the deposition of secondary quartz endomorphs enveloping libethenite (Crane et al., 1998). Formation
17 of secondary barite (BaSO4) early in the paragenetic sequence in E22, E26 and E27
(Crane et al., 1998; McLean et al., 2004) is further evidence of a low pH regime at this stage. While the source of Ba2+ ions probably also involves the decomposition of feldspars and micas, transport of significant amounts of barium and sulfate ions in solution at ambient temperatures requires conditions acidic enough to form
- appreciable concentrations of HSO4 , together with ionic strengths high enough to
2+ 2- have an effect on lowering the activity coefficients of Ba (aq) and SO4 (aq)
(Blount, 1977). A subsidiary mechanism to generate locally high activities of dissolved phosphate species and copper ion could have involved the well-understood adsorption of Cu2+ and phosphate on iron and manganese oxides (Arai and Sparks,
2001; Post, 1999; Sparks, 2001). It is noted in this connection that much of the libethenite and pseudomalachite of the three deposits is associated with goethite and
Mn-rich oxides at the base of the leached zones (particularly in E22 and E27). As poorly-crystalline precursors of goethite and the like recrystallised to more stable phases, copper and phosphate would be released from cryptocrystalline surfaces to react essentially in situ. Nevertheless, the emplacement of libethenite and pseudomalachite requires a low pH regime, under which secondary phases such as malachite, azurite and atacamite are freely soluble.
In conjunction with the above phenomena, extensive crystallisation of sampleite in the oxidised zone of E26 took place (Figure 2.3.2). For this to occur, both elevated phosphate and chloride activities are necessary. Figure 2.3.1 also illustrates the conditions that must be satisfied for sampleite to form at 298.2 K, using the stability constant for sampleite reported above and that for atacamite from
- -4 the literature (Smith and Martell, 1976; Williams, 1990). For a(H2PO4 ) = 10
- (H2PO4 is the dominant phosphate solution species at pH values between about 2 and
18 - -2 - -6 - 7), a(Cl ) must be greater than 10 and with a(H2PO4 ) = 10 , necessary a(Cl ) values need to approach those associated with saturated NaCl solutions. The rarity of sampleite (Anthony et al., 2000) is thus seen as a consequence of the rather unusual solution conditions required for its crystallisation. In this connection it is possible to
- estimate both pH and a(H2PO4 ) conditions that prevailed during the sampleite mineralisation event. As noted above, dissolved phosphate species are derived from fluorapatite, which dissolves incongruently to form fluorite (Misra, 1999) as shown in the equation below.
+ 2+ - 2Ca5(PO4)3F(s) + 12H (aq) ⇋ CaF2(s) + 9Ca (aq) + 6H2PO4 (aq) (2.7)
o -1 At 298.2 K, ΔrG = +38.1 kJ mol and lgKH = -6.673
2+ 9 - 6 + 12 (KH = a(Ca ) a(H2PO4 ) /a(H ) ) for the reaction based on the data of Robie and
Hemingway (1995). As a first approach, it is convenient to ignore activities and to treat the reaction in terms of concentrations. From stoichiometric relationships,
2+ - - [Ca ] = 9/6[H2PO4 ], and the lgK expression can thus be recast as lg[H2PO4 ] = -
- 0.551 – 0.8 pH. For pH = 3, 4, 5 and 6, lg[H2PO4 ] = -2.95, -3.75, -4.55 and -5.35, respectively. These results are quite consistent with those that may be abstracted
- from Figure 2.3.1. To preclude formation of atacamite, a(H2PO4 ) would need to be of the order of 10-5 and the pH thus less than 5. For libethenite to form, as observed in the zone of sampleite crystallisation (Figures 2.3.2 and 2.3.3), the pH must locally have fallen below 3. Figure 2.3.1 indicates an activity of Cl- of the order of 0.1 for the above conditions. However, a more realistic set of solution parameters can be deduced on the basis of current groundwater geochemistry. This is because Ca2+ is contributed, of course, from sources other than fluorapatite alone.
19 500 µm
Figure 2.3.2. Secondary electron image of sampleite and libethenite from Northparkes E26 deposit, NSW.
500 µm
Figure 2.3.3. Secondary electron image of libethenite from Northparkes E26, NSW.
20 Taking note of the absence of sampleite in E22 and E27, it is apparent that present day groundwater compositions reflect mineralogical differences in the oxidised copper assemblages at Northparkes. Table 2.3.2 lists data for three high- quality analyses of ground waters associated with E22 and E26. It is particularly noteworthy that the two samples from E26 have chloride concentrations over a magnitude greater than the sample from E22. Ca2+ activities have been calculated
- from the data and are listed in Table 2.3.2. Calculation of values of a(H2PO4 ) can be made directly using the KH value for the acid dissolution of fluorapatite. For sample
- E26P118, lga(H2PO4 ) = -3.86, -5.86, -7.86 and -9.86 for pH 3, 4, 5 and 6, respectively. Corresponding values for sample E26P119 are
- lga(H2PO4 ) = -3.73, -5.73, -7.73 and -9.73, for the same pH conditions, and for
- sample E22P157, lga(H2PO4 ) = -3.07, -5.07, -7.07 and -9.07, respectively. Similar corrections for activity coefficients give values of lga(Cl-) of -0.62 and -0.60 for E26 and -1.67 for E22 (Table 2.3.2). Reference again to Figure 2.3.1 indicates that with
- -5 respect to the former an a(H2PO2 ) of less than 10 in the presence of suitable amounts of Na+ and Ca2+ will cause the crystallisation of sampleite, whereas for the
- -4 E22 case, a(H2PO2 ) must be somewhat greater than 10 . Evidently, this higher value was not achieved and sampleite is thus absent in both E22 and E27 deposits.
Furthermore, the calculations fix the pH of mineralising solutions in E26 during the phosphate phase at between 3 and 4. This low pH regime is quite consistent with the oxidation of sulfides in a poorly buffered environment and serves to explain other features of the deposits at Northparkes. Leaching of the upper oxidised zones is extensive and the phosphate minerals are not accompanied by any other Cu(II)
21 Table 2.3.2. Groundwater analyses (concentrations in mol dm-3 except for pH) for E22 and E26 deposits.a Sample numberb E26P118 E26P119 E22P157 pH 7.25 6.85 7.80 carbonate negc neg 0.00002 bicarbonate 0.00338 0.00439 0.00336 chloride 0.32550 0.34581 0.02651 sulfate 0.03643 0.03940 0.00588 potassium 0.00046 0.00049 0.00015 sodium 0.26925 0.28447 0.01544 calcium 0.01884 0.02445 0.00474 magnesium 0.04422 0.04566 0.00823 Σ(+) 0.39583 0.42518 0.04153 Σ(-) 0.40176 0.42900 0.04163 I 0.50 0.54 0.06 γ± 0.737 0.726 0.806 γ2± 0.295 0.278 0.432 lg a(Cl-) -0.62 -0.60 -1.67 lg a(Ca2+) -2.255 -2.167 -2.697 aAnalytical data originally reported in ppm; data are reliable to three significant figures and extended values are only the result of calculations to change concentration units. bE26P118 and E26P119 samples from beneath the E26 deposit and E22P157 from E22. cNot detectable.
phases to a significant extent. The reason for this is that the latter are simply too soluble at this pH to be precipitated. To illustrate this phenomenon, reference is made to equilibrium conditions between sampleite and pseudomalachite, in the equation below.
+ 2+ NaCaCu5(PO4)4Cl.5H2O(s) ⇋ Cu5(PO4)2(OH)4(s) + Na (aq) + Ca (aq)
- - + 2H2PO4 (aq) + Cl (aq) + H2O(l) (2.8)
- Use of the data of Table 2.3.2 gives for E26 lga(H2PO4 ) = ca -4, and a pH of ca 3.1.
In turn, a(Cu2+) is derived from (2.2) as being about 0.03. Inspection of the data of
Table 2.3.2 for this level of dissolved copper indicates a deficiency in anion concentrations and these must be accounted for. However, it is possible to deduce what these would have been, and it leads to a somewhat surprising outcome. First,
22 extra chloride would be a component, now incorporated in subsequent atacamite mineralisation. More significant, however, would be sulfate released during the oxidation of sulfides, together with the copper in solution that was to some extent incorporated in the Cu minerals of the oxidised and supergene enriched zones. As protons reacted with host rocks, particularly feldspars, to form clays, the overall exchange process was of protons for potassium, sodium and, particularly, calcium ions [c.f. equations (2.4) to (2.6)]. A feature of the profiles at Northparkes is the pervasive development of gypsum at a level somewhat below the water table
(Heithersay et al., 1990). This then is seen as a natural consequence of the weathering of both sulfides and host rock silicates during the development of the oxidised zone. It is noted that the layer of gypsum impregnation is not geologically controlled, but parallels surface topography (Heithersay et al., 1990). Gypsum is rather slow to nucleate from supersaturated solution (see for example, Abdel-Aal et al., 2004; Lancia et al., 1999; van der Ende, 1991) and its precipitation at lower levels reflects this, at least in part.
As aqueous phosphate activities fell (with an accompanying and necessary rise in pH), solution conditions evolved to be consistent with those associated with the continuing formation of pseudomalachite and, with the essential exhaustion of available phosphate, subsequent crystallisation of atacamite in E26. Pseudomorphs of atacamite, chrysocolla and, more rarely, pseudomalachite after sampleite are noted in the oxidised zone of E26 (McLean et al., 2004) and lend further support to the evolution of oxidising conditions with time. The basic copper(II) phosphates are rather slow to dissolve at ambient temperatures (Crane et al., 2001) and their preservation is no doubt due to this in part.
23 The fact that both high chloride and phosphate activities play key roles in the formation of sampleite is a key to understanding its absence in other deposits where the more common secondary Cu(II) phosphates libethenite and pseudomalachite are found. At Girilambone, New South Wales, abundant secondary libethenite and pseudomalachite mineralisation is associated with supergene mineralisation in the main pit (Chapman et al., 2005) and sampleite is absent. High-quality analyses of some 13 groundwater samples adjacent to the main pit and three satellite deposits
(Northeast, Hartmans, and Larsens) are available (Elvy, 1998). Full speciation calculations for these give a(Cl-) values ranging from 0.055 to 0.156. Values of
- a(H2PO4 ), however, are much lower, ranging from less than measurable to
4.39 x 10-7. In line with Figure 2.3.1, the absence of sampleite in the deposits is readily explained. Sampleite from the type locality at Chuquicamata, Chile (Cook,
1978), has an association similar to that found at Northparkes. Accompanying phases are atacamite (Figure 2.3.4), libethenite and jarosite. The former is indicative of high chloride activities and the latter two of an acidic environment. These observations are
Figure 2.3.4. Sampleite associated with atacamite from Northparkes E26 deposit.
24 quite consistent with the conclusions drawn above, based on the determined stability constant for sampleite.
2.3.2 Atacamite, malachite and azurite
Atacamite deposition in E26 took place after the phosphate suite had developed and was emplaced higher in the oxidised zone than secondary copper carbonates. In some places, atacamite is associated with malachite and azurite, but much material containing the secondary carbonates contains no halide minerals.
Instead, it is often associated with cuprite and native copper (McLean et al., 2004), indicating that it came from the lower oxidised zone. It is noteworthy that only traces of clinoatacamite, the thermodynamically stable polymorph of composition
Cu2Cl(OH)3, are present at Northparkes. This is a consequence of the inhibition of nucleation of clinoatacamite when excess Cl-(aq) is present (Pollard et al., 1989).
Under the conditions that prevailed at Northparkes in E26, the metastable polymorph persisted, in line with present groundwater compositions (Table 2.3.2). Development of atacamite is the consequence of a second secondary Cu mineralisation event, when dissolved phosphate species had essentially been exhausted, as pH rose when protons reacted with feldspars and the like. The development of atacamite versus malachite or azurite is dictated by the prevailing pH of mineralising solutions, the availability of chloride ions, and variation of the partial pressure of CO2(g). Figure 2.3.5 illustrates relationships between atacamite, malachite and brochantite at 298.2 K, at a
-3.5 p(CO2) value of 10 , corresponding to atmospheric concentrations. Atacamite encroaches on the stability fields progressively as a(Cl-) increases. With respect to
E26, for a value of lga(Cl-) = -0.6 (Table 2.3.2), brochantite can only form when
2- -1 a(SO4 ) is greater than about 10 . This is consistent with the data of Table 2.3.2, in that brochantite is very rare in E26 (McLean et al., 2004).
25 23456789pH 0
Brochantite
-2
) -4 2- 4 (SO
a Malachite/ Azurite
lg -6
-8 Atacamite
10-4 10-3 10-2 10-1 10-0.6 a (Cl-) -10
Figure 2.3.5. Relationships between atacamite, malachite and brochantite at 298.2 K, -3.5 with a p(CO2) value of 10 (corresponding to atmospheric concentrations).
Figure 2.3.6 illustrates stability relations between atacamite, malachite and azurite with varying partial pressures of CO2(g) at 298.2 K. Boundaries are between
-1.36 atacamite and malachite up to a p(CO2) value of 10 . This p(CO2) is the equilibrium partial pressure at 298.2 K for reaction given below, when malachite
-1 coexists with azurite. At p(CO2) = 10 , azurite is the stable secondary carbonate.
3Cu2CO3(OH)2(s) + CO2(g) ⇋ 2 Cu3(CO3)2(OH)2(s) + H2O(l) (2.9)
26 345678pH 0
Atacamite
-1 Azurite
-2 ) -
-1 (Cl -3 10 a -1.36
lg 10
-4
10-2.5 Malachite
-5
10-3.5
p (CO2) -6
Figure 2.3.6. Relationships between atacamite, malachite and azurite with varying partial pressures of CO2(g) at 298.2 K. Boundaries are drawn with respect to -1.36 atacamite and malachite up to a p(CO2) value of 10 . At p(CO2) values higher than 10-1.36, azurite is the stable secondary carbonate.
Primary mineralisation at Northparkes contains small amounts of carbonate (Tonui et al., 2002), but a likely source for the considerable amounts fixed in malachite-azurite mineralisation is apparent from stable carbon measurements of azurite from E27. A value of 13C = -11.23 indicates a significant biological input (Melchiorre and
Enders, 2003). With pH ranging of between 4 and 8, and lga(Cl-) = -0.6 (Table
-3.5 -1 2.3.2), as p(CO2) increased from, say, 10 to 10 , conditions pass progressively from those that would give rise to atacamite to those associated with malachite and azurite formation in turn. This simple geochemical picture serves to explain how all of the major copper(II) minerals developed in E26 (and E22 and E27); in addition, it
27 illustrates why phases common in other settings are not seen. While continuously evolving chemical conditions may have prevailed during the passage from atacamite to carbonate deposition, this too may have been episodic. It is not possible at this stage to relate the phosphate, chloride and carbonate events to the major paleoweathering ages for Northparkes. However, the overall phosphate → chloride
→ carbonate sequence is clearly indicated by the phase relationships described above, together with field relationships (McLean et al., 2004).
2.4 GEOCHEMICAL DISPERSION OF COPPER FROM THE
OXIDISED ZONES AT NORTHPARKES
Leverett et al. (2004) described a method for estimating the extent of element mobility in ground waters at the New Cobar deposit, based on a reconstruction of groundwater compositions. This was possible in the light of the secondary mineral assemblage at the deposit. As far as Northparkes is concerned, present day groundwater compositions are evidently good proxies for previous groundwaters as far as major cations and anions are concerned, as they are consistent with a formal description of the development of the oxidised zones using equilibrium approaches; this has been dealt with in detail above. The approach here was to derive a set of starting parameters for minor aqueous species for the chloride and carbonate events, taken separately, and to use the current groundwater compositions to derive total dissolved copper concentrations in equilibrium with the assemblages. This follows the New Cobar approach, but in this case is much simpler because copper is the only base metal present in any significant amount. As far as the phosphate mineralising event is concerned, the low pH values associated with it indicate that significant copper was dispersed. For the atacamite and malachite/azurite events, solution
28 chemistry is more subtle at higher pH values and calculations were carried out to determine effects on copper dispersion of these two styles of mineralisation.
Calculations of total dissolved copper in equilibrium with azurite, malachite and atacamite were carried out for pH values of 5, 6 and 7, and for p(CO2) values of 10-3.5 (atmospheric), 10-2.5 (typical soil gas levels) and 10-1.36 [equilibrium pressure of CO2(g) at 298.2 K with respect to equation (2.9)]. It is noted that both secondary carbonates were abundant in all lower oxidation zones (Crane et al., 1998;
McLean et al., 2004). Stability constant data for dissolved carbonate species,
+ o 2- + o CuHCO3 , CuCO3 , Cu(CO3)2 , CuCl and CuSO4 , and malachite and azurite were taken from the literature (Martell and Smith, 1982; Smith and Martell, 1976, 1989;
Williams, 1990; Zirino and Yamamoto, 1972). The concentrations of these species and of Cu2+(aq) were calculated using the ground water concentrations of Table 2.3.2 for all other species. Corrections to activities were made in the usual way and total copper concentrations summed. Results are listed in Table 2.4.1. For the E26 calculations, reference to Figure 2.3.6 indicates that the stable phase to be taken into
-1.36 account is atacamite in all cases except when pH = 7 and p(CO2) = 10 , when it is malachite or azurite. In the case of the E22 sample, azurite is the stable phase except
-1.36 - at pH = 7 and p(CO2) = 10 , when again it is malachite or azurite. At p(CO2) = 10
3.5 and pH = 5, total dissolved copper concentrations for the two E26 and the E22 ground water analyses are 3.66 x 10-3, 3.74 x 10-3 and 4.07 x 10-2, respectively. At
-1.36 p(CO2) = 10 and pH = 7, when azurite is stable in all cases, total dissolved copper concentrations for the two E26 and the E22 ground water analyses are 3.26 x 10-6,
3.29 x 10-6 and 2.86 x 10-6, respectively. There is little evidence for increased total copper in solution as a result of elevated Cl-(aq) activities associated with E26. More copper is in solution associated with malachite at lower pH values (E22), as
29 Table 2.4.1. Calculated total dissolved copper loads (mol dm-3) for the aqueous solutions of Table 2.3.2, calculated at pH values of 5 to 7 and varying p(CO2) (see text). The calculations assume equilibrium conditions with respect to malachite, azurite or atacamite in line with stabilities indicated by Figure 2.3.6. ______[Cu]tot______p(CO2) pH E26P118 E26P119 E22P157 10-3.5 5 3.66x10-3 3.74x10-3 4.07x10-2 10-3.5 6 1.16x10-4 1.18x10-4 4.07x10-4 10-3.5 7 3.75x10-6 3.83x10-6 4.28x10-6 10-2.5 5 3.66x10-3 3.74x10-3 1.35x10-2 10-2.5 6 1.17x10-4 1.19x10-4 1.35x10-4 10-2.5 7 3.60x10-6 3.74x10-6 2.04x10-6 10-1.36 5 3.69x10-3 3.76x10-3 3.51x10-3 10-1.36 6 7.59x10-5 7.89x10-5 4.05x10-5 10-1.36 7 3.29x10-6 3.33x10-6 2.89x10-6
expected, and greater dispersion will have taken place under conditions where lower
Cl- activities prevailed and atacamite could not form. This result seems counter- intuitive at first in that CuCl+(aq) is expected to contribute to copper mobility, but it simply reflects solubility relationships for the carbonate minerals; acid reacts with carbonate and significant amounts of malachite and azurite can only form at higher pH values.
It is concluded that significant amounts of copper were dispersed from all three deposits during the acidic weathering event associated with secondary copper phosphate formation. In E22 (and presumably in E27), copper dispersion was greater than in E26 until approximately neutral pH was attained. Subsequently, there has been negligible difference in the relative extent of copper mobilisation from the three deposits. Taken together, these facts indicate that the total solution load of copper associated with weathering Northparkes-style mineralisation is largely independent of ground water salinity. Furthermore, the significant amounts of copper mobilised during the acidic weathering event, and those associated with later development of malachite-azurite-atacamite zones (particularly at lower pH values) is expected to generate a rather diffuse geochemical signature for the element, and one that would
30 have a considerable lateral extent. It is noted that Tonui et al. (2002) reported highly anomalous copper values in saprolite immediately adjacent to and overlying mineralisation in the E22 and E27 deposits. The results of the present work indicate that Northparkes style mineralisation should be detectable using geochemical exploration methods at some distance, in view of the reconstruction of the chemical conditions responsible for the development of supergene copper mineralisation in the regolith above E22, E26 and E27 deposits. This may have use in exploration strategies adopted in the region, especially with respect to blind mineralisation that has some weathering. The careful work of Tonui et al. (2002) with respect to appropriate sampling media in this setting should also be taken into account. Their conclusion that saprock is the preferred sampling medium at Northparkes is well- founded in the light of extensive leaching of copper in the upper oxidised zones.
31 Chapter 3
THE GEOCHEMISTRY OF FORMATION OF
LAVENDULAN
3.1 BACKGROUND
The mineral lavendulan was originally described by Breithaupt (1837) as a hydrous cobalt arsenate contaminated by nickel and copper from the Galliläische
Wirtschaft mine, Annaberg, Germany. It was subsequently found at Jáchymov,
Czech Republic by Vogl (1853) and San Juan, Chile (Goldsmith, 1877). Uncertainty about its identity crept into the literature when Foshag (1924b) published findings that showed the material from Jáchymov was optically different to that from Chile.
The Chilean material gave a chemical formula of Na3(Cu,Ca)3(AsO4)2(OH)3·H2O, leading Foshag to rename it freirinite (Foshag, 1924a). However, powder X-ray studies by Guillemin (1956) showed that lavendulan and frerinite from both
Jáchymov and San Juan gave identical diffraction patterns. Analysis of the mineral from Chile and synthetic material gave the formula Na(Cu,Ca)6(AsO4)4Cl·4-5H2O
(or NaCaCu5(AsO4)4Cl·5H2O, its true composition). The mineral was shown to be isostructural with sampleite, NaCaCu5(PO4)4Cl·5H2O.
The ‘type’ specimen has recently been analysed and described as an amorphous mixture containing As and Zn with minor Cu, Co and S (Ondruš et al.,
2006). Thus the ‘type’ specimen bears no resemblance to established lavendulan.
Odruš et al. (2006) described the new mineral lemanskiite as a tetragonal dimorph of pseudo-tetragonal lavendulan, and the Ca analogue of zdenekite,
32 NaPbCu5(AsO4)4Cl·5H2O (Chiappero and Sarp, 1995). In describing lemanskiite they noted that some specimens thought to be lavendulan were in fact lemanskiite.
This included specimens of minerals from Talmessi in Iran, San Juan in Chile and
Postrama Murcia in Spain. Powder X-ray measurements of material form Jáchymov and Laurion in Greece confirmed the presence of lavendulan, although lavendulan from Laurion (Figure 3.2.1) transformed to lemanskiite several days after grinding.
However, Geister et al. (2007) suggested that lemanskiite may be a partly dehydrated and tetragonal or nearly tetragonal analogue of lavendulan and unequivocally established the isomorphous structural relationship between lavendulan and sampleite.
Figure 3.2.1. Lavendulan from Laurion, Greece. Field of view 660 m.
33 In order to adequately understand the mode of formation of lavendulan in relation to sampleite and other copper(II) arsenate minerals, its stability constant was determined at 298.2 K. This is described below, together with the chemical conditions that give rise to it and its congeners.
3.2 EXPERIMENTAL
3.2.1 Powder X-ray diffraction (XRD)
Powder X-ray diffraction (XRD) was commonly used for identification of minerals and synthetic phases. Measurements were carried out using a Philips
PW1925-20 powder diffractometer operating at 40 kV and 30 A using Ni-filtered
Cu Kα radiation ( = 1.5405 Å). Pure Si was used as internal standard when unit cell parameters were to be determined. Detected X-rays were recorded between 5 and 75°
2θ, with a step size of 0.02° and a rate of 1.2° per minute. Identification of phases present in diffraction patterns was effected using Diffraction Technology Data
Processing Software (Traces Version 6) with JCPDS-ICCD database files.
3.2.2 Total reflectance X-ray fluorescence (TRXRF) spectrometry
When no image capture was necessary, complementary, but qualitative, elemental analysis was carried out using a total internal reflectance XRF instrument of in-house design. The instrument is mounted on a Philips PW1729 generator operating at 40 kV and 35 mA to produce Mo Kα radiation ( = 0.7107 Å). A small amount of sample was placed on an optically flat quartz sample support and X-rays were detected with an Amptek XR-100CR Peltier-cooled X-ray detector. Energy profiles were created and analysed using Moran Scientific software.
34 3.2.3 Scanning electron microscopy (SEM)
Scanning electron microscopy was carried out using a JEOL 840
Microanalyzer, fitted with a thin window energy dispersive detector, controlled by
Moran Scientific acquisition and processing software. Energy dispersive spectrometry (EDS) was used for qualitative/semi-quantitative elemental analysis.
The system was operated at an accelerating voltage of 15 keV and a beam current of
20 nA. SEM was also employed for high magnification secondary electron (SE) image capture. Samples were coated with carbon prior to analysis to prevent charging.
3.2.4 Electron microprobe analysis
Quantitative elemental analysis was carried out using a JEOL JXA 8600
Superprobe equipped with three wavelength dispersive spectrometers, analysing from Be to U, and a Kevex energy dispersive spectrometer analysing from Na to U controlled by Moran Scientific software. Samples were embedded in epoxy resin, ground flat and polished with 1 m and 0.25 m diamond pastes before being carbon coated. Samples were analysed at an accelerating voltage of 15 keV and a beam current of 20 nA. Standards used for each element and minimum detection limits are given in Table 3.2.1.
3.2.5 Atomic absorption spectroscopy (AAS)
AAS measurements (Cu) were carried out with a Perkin Elmer AAnalyst 100
-3 instrument using matched standards. Standards of CuCl2·2H2O in 0.1 mol dm NaCl were prepared using pre-calibrated glassware. Measurements of pH were taken using a Radiometer PHM 93 meter fitted with an IJ44 combination electrode.
35 Table 3.2.1 Element standards and minimum detection limits (MDL) for electron microprobe analyses. Element Standard MDL (wt%) Na Jadeite 0.019 K Orthoclase 0.013 Ca Fluorite 0.019 Pb Lead sulfide 0.123 Cu Cu metal 0.022 P Apatite 0.028 As Gallium arsenide 0.048 Cl Thallium chloride 0.017 Cd Cadmium metal 0.228 Ni Ni metal 0.153 Zn Zn metal 0.239 Mg Olivine 0.021
3.2.6 Lavendulan synthesis
Pure lavendulan was synthesised as follows. CaCl2 (0.1110 g, 1 mmol),
CuCl2·2H2O (0.8520 g, 5 mmol), Na2HAsO4 (1.2487 g, 4 mmol) and NaCl (5.843 g,
0.1 mol) were added to a Teflon-lined Parr acid digestion bomb together with 75 cm3 of deionised water. The bomb was heated to 100oC for 48 hours and then cooled slowly to room temperature. The product was collected at the pump, washed with water, then acetone, and sucked dry. Use of lesser amounts of NaCl resulted in mixtures of lavendulan and olivenite. Gram quantities of lavendulan were easily obtained using the above procedure. No phases other than lavendulan were detected in the product using XRD and satisfactory analyses corresponding to the stoichiometric formula were obtained.
3.2.7 Solution studies
Preliminary studies showed that lavendulan dissolves incongruently in aqueous 10-3 mol dm-3 HCl solution to give mixtures of lavendulan and olivenite.
However, the mineral does dissolve congruently in aqueous 10-3 mol dm-3 HCl
36 solution that is also 0.1 mol dm-3 in NaCl. Excess solid lavendulan was equilibrated in several sealed flasks with 100 cm3 of aqueous 10-3 mol dm-3 HCl and 0.1 mol dm-3
NaCl. The flasks were placed in a shaking water bath and maintained at 25 ±0.1oC.
Parallel experiments were conducted to monitor pH changes as the solutions reached equilibrium with lavendulan. No pH change was observed after two weeks, but solutions were kept in the water bath for a further six weeks. At this time, the pH of each equilibrated solution was recorded and the sample filtered through a 0.7 m fibreglass disc. Measurement of total dissolved Cu was carried out using AAS. For the seven experimental runs undertaken, total dissolved Cu concentrations (x 10-4 mol dm-3) with equilibrium pH values in parentheses were 5.635 (4.511), 5.537
(4.522), 5.635 (4.523), 5.870 (4.525), 5.870 (4.528), 5.503 (4.526) and 5.503 (4.530).
This leads to a solubility of lavendulan, NaCaCu5(AsO4)4Cl·5H2O, of 1.24 ±0.03 x
10-4 mol dm-3 at 298.2 ±0.1 K and a pH of 4.524 ±0.006.
In order to calculate a stability constant for lavendulan, species distribution calculations using the solubility of the mineral were carried out for each of the experimental runs. Species incorporated in the COMICS calculations (Perrin and
Sayce, 1967) are shown in Table 3.2.1.
Table 3.2.1 Equilibrium constants for complex species when I = 0.1 used in determining species distributions. Species Log K CuOH+ -8.02 CaOH+ -13.02 2- HAsO4 11.95 - H2AsO4 17.60 0 H3AsO4 19.74 0 CaHAsO4 13.94 + CaH2AsO4 18.78 CuCl+ -0.03 0 CuHAsO4 15.38 + CuH2AsO4 19.12
37 Stability constants (Martell and Smith, 1982; Smith and Martell, 1976, 1989) were corrected to an ionic strength of I = 0.102 mol dm-3 using the Davies modification of the Debye-Hückel equation, taking γ for neutral species to be equal
2 ½ ½ to unity, at 298.2 K (lgγi = -0.5085zi [I /(1 + I ) –0.3I], where zi is the charge on species i). Calculated equilibrium concentrations of the free aquated ionic species
+ 2+ 2+ - - -4 -4 Na , Ca , Cu , H2AsO4 and Cl were 0.1001, 1.217 x 10 , 5.160 x 10 ,
3.980 x 10-4 and 0.1011 mol dm-3, respectively, for the average solubility of lavendulan at pH 4.524 (for all data, see Table A.1 in the Appendix). These values, together with lgγ± = -0.108, lgγ2± = -0.430 and lgγ3± = -0.968 (I = 0.102 mol dm-3) were used in turn to calculate KH, with respect to equations 3.1 and 3.2 below, where a(i) is the activity of species i (a(i) = m(i)γ(i), with m(i) being the concentration of species i).
+ + 2+ 2+ NaCaCu5(AsO4)4Cl·5H2O(s) + 8H (aq) ⇋ Na (aq) + Ca (aq) + 5Cu (aq) +
- - 4H2AsO4 (aq) + Cl (aq) + 5H2O(l) (3.1)
+ 2+ 2+ 5 - 4 - + 8 KH = a(Na )a(Ca )a(Cu ) a(H2AsO4 ) a(Cl )/a(H ) (3.2)
At 298.2 K, log KH is equal to -2.54 ±0.08. From this result, the standard free energy
o of formation of lavendulan is ΔfG (NaCaCu5(AsO4)4Cl·5H2O,s,298.2 K) = -4839.1
±6.2 kJ mol-1. With respect to this quantity, values for the free energies of formation of component ions and water were taken from Robie and Hemingway (1995), except
- - -1 for H2AsO4 (aq). The value used for H2AsO4 (aq) at 298.2 K (-753.9 ±0.12 kJ mol )
3- was derived from a value for the corresponding AsO4 ion together with the first and second proton association constants given by Smith and Martell (1976)and Martell
38 and Smith (1982). The total error in ΔfGo for lavendulan was calculated by the addition of the errors for individual species (Robie and Hemingway, 1995) and the
o 2+ calculated error in log KH. It is noted that no errors are listed for ΔfG for Ca (aq)
- -1 and H2AsO4 (aq); these are estimated to be 0.1 and 0.12 kJ mol , respectively.
3.3 GEOCHEMISTRY OF MINERAL FORMATION
3.3.1 Phase relationships
The chemistry of formation of a number of copper(II) arsenates, carbonates and chlorides have been described in the literature, but for a large number of arsenate species reliable thermodynamic properties are yet to be determined. For geochemical modelling purposes, thermodynamic properties (stability constants) of common
Cu(II) arsenate minerals are available and these, together with the presently determined value for lavendulan have been used to construct relevant phase diagrams. Stability field diagrams show the chemical conditions under which the various minerals crystallise from aqueous solutions and can explain observed paragenetic sequences. Thermodynamic data for associated arsenate minerals have been listed in Table 3.3.1, along with similar data for relevant dissolved species. It should be noted here that errors in values of ΔfGº (298.2 K), derived from uncertainties for component dissolved species, cancel out in calculations describing the equilibrium conditions between a pair of stable minerals.
A stability field diagram for lavendulan and associated copper(II) arsenate minerals is shown in Figure 3.3.1. Using the thermodynamic data from Table 3.3.1, boundaries for mineral pairs were calculated for aCa2+ = 10-5, the predominant
- arsenate species being H2AsO4 when pH is between 2.1 and 7.0. The diagram shows the change in the lavendulan stability field when the chloride activity is increased
39 Table 3.3.1. ΔfGº (298.2 K) for minerals and dissolved species used in geochemical modelling calculations. Species Formula ΔfGº (kJ mol-1) Referencea
Atacamite Cu2Cl(OH)3 -667.55 1 Lammerite Cu3(AsO4)2 -1300.8 1 Olivenite Cu2AsO4(OH) -846.4 ± 1.6 1 Cornwallite Cu5(AsO4)2(OH)4 -2057.9 ± 4.1 1 Clinoclase Cu3AsO4(OH)3 -1211.2 ± 2.7 1 Conichalcite CaCuAsO4(OH) -1471.7 ± 3.3 1 Lavendulan NaCaCu5(AsO4)4Cl·5H2O -4839.1 ± 6.2 2 Cu2+ +65.1 ± 0.1 3 Ca 2+ -553.6 3 Na+ -261.5 3 Cl- -131.2 3 - H2AsO4 -753.89 4 a[1] Williams (1990); [2] this study; [3] Robie and Hemingway (1995); [4] Martell and Smith (1982); Smith and Martell (1976, 1989).
234567pH 0
Cornwallite
-2 Olivenite Clinoclase
Lammerite ) 2+
-4 (Cu a lg
Lavendulan Conichalcite -6
a (Cl-) 10-3 10-2 2+ -5 10-1 a (Ca ) = 10 -8
Figure 3.3.1. Stability fields for lavendulan and selected Cu(II) arsenates relative to 2+ 2+ -5 - -3 aCu and pH, when aCa = 10 . Boundaries are drawn for aCl = 10 (solid lines), - -2 - -1 aCl = 10 (dotted lines) and aCl = 10 (dashed lines).
40 from 10-3, corresponding to weakly saline groundwater, to 10-1, corresponding to strongly saline groundwater (Williams, 1990). It is not unreasonable to consider chloride activities of these magnitudes as very saline grounds waters can be present in oxidising sulphide ores. In fact, a large number of chloride minerals including lavendulan are a feature of oxide zones located in arid and semi-arid areas (see below), or alternatively associated with saline groundwaters such as was the case with sampleite at Northparkes, as discussed previously.
The rarity of lavendulan compared with the more common copper(II) arsenates can be readily explained with reference to Figure 3.3.1. When the activity of chloride is high and the pH is low, lavendulan is stable over a wide range of copper activities. As activity of chloride increases from weakly saline to strongly saline conditions, the stability field of lammerite is impinged upon by that of lavendulan until it is completely eliminated at a chloride activity of 10-1. At localities such as the south pit of the Gold Hill mine, Utah, lavendulan is found in association with other Cu(II) arsenates including olivenite, cornwallite and conichalcite
(Kokinos and Wise, 1993). While lavendulan requires a low pH regime to form, its association with these minerals (particularly cornwallite) indicates a wide range of pH values associated with the complete assemblage. Elsewhere in the same deposit clinoclase is also known to form, indicating yet higher pH values were ultimately achieved.
The activity of calcium in the mineralising system greatly affects the stability of lavendulan and the other Cu(II) arsenates. Figure 3.3.2 is a stability field diagram for the copper(II) arsenates constructed in the same fashion as Figure 3.3.1, but with an increased activity of Ca2+ of 10-3. This results in conichalcite significantly encroaching on the stability field of lavendulan, and also those of olivenite,
41 234567pH 0
Olivenite Cornwallite
Lammerite Clinoclase -2 Lavendulan ) 2+
-4 (Cu a
lg Conichalcite
-6 10-3
10-2 a (Cl-) 10-1 a (Ca2+) = 10-3 -8
Figure 3.3.2. Stability fields for lavendulan and selected Cu(II) arsenates relative to 2+ 2+ -3 - -3 aCu and pH, when aCa = 10 . Boundaries are drawn for aCl = 10 (solid lines), - -2 - -1 aCl = 10 (dotted lines) and aCl = 10 (dashed lines).
cornwallite and clinoclase. As calcium is readily available in groundwaters associated with oxidising orebodies, it is therefore not unexpected that conichalcite is much more common than lavendulan in copper arsenate assemblages.
As predicted by the above stability field diagrams, olivenite and conichalcite are close associates of lavendulan in many deposits (Anthony et al., 2003). The association of lavendulan, olivenite and conichalcite may be seen as being the result of subtle fluctuations in the activities of Cl-, Ca2+ and Cu2+. An example of this is reflected by the supergene mineralisation of the Dome Rock deposit, South Australia, from which occurrences of lavendulan, conichalcite, olivenite and other arsenate minerals have been reported (Kleeman and Milnes, 1973; Ryall and Segnit, 1976;
42 Segnit, 1976). Pseudomorphs of conichalcite after olivenite are found, indicating that conichalcite is later in the paragenetic sequence (Ryall and Segnit, 1976). Olivenite is rare in the deposit, with some pseudomorphs of conichalcite after olivenite still retaining an olivenite core (Segnit, 1978). With reference to Figures 3.3.1 and 3.3.2 this may be considered the result of increasing activity of calcium ions, or increasing activity of copper ions, with constant pH. The low pH necessary for lavendulan formation is also confirmed by the presence of the mineral scorodite, FeAsO4·2H2O, from this deposit (Segnit, 1976). Scorodite can only be formed at pH values less than those which serve to stabilise goethite, FeOOH.
Lavendulan has been observed in a number of deposits associated with scorodite or its aluminium analogue mansfieldite, AlAsO4·2H2O. These minerals are indicative of very acidic conditions of formation; to form scorodite a pH of less than
2 is required (Zhu and Merkel, 2001). XRD analysis of material from the El Guanaco mine, Taltal, Chile, showed lavendulan to be associated with mansfieldite (Figure
3.3.3). The occurrence of dark blue green crystals of lammerite associated with the lavendulan from the same deposit (Geister et al., 2007) is also indicative of low pH conditions (vide supra).
Figure 3.3.4 is a stability field diagram showing relationships between selected Cu(II) arsenates and atacamite at 298.2 K. Atacamite is used as a proxy for all Cu(II) chloride minerals with a Cu2Cl(OH)3 stoichiometry. The diagram clearly shows that high arsenate activities and relatively low pH values are necessary for lavendulan to crystallise from mineralising solutions. In addition, the combination of very low pH and very high copper and arsenate activities required for the formation of lammerite goes some way to explaining its rarity. The formation of the arsenates cornwallite and clinoclase is precluded by atacamite and conichalcite at the
43
Figure 3.3.3. Lavendulan associated with minor mansfieldite from El Guanaco, Taltal, Chile. Field of view is 5cm.
23456789pH 0 Lavendulan
-2 Lammerite ) -
4 -4 Conichalcite
Olivenite AsO 2 (H a -6 lg
-8 a (Cl-) -2 Atacamite 10 10-1 100 -10 Figure 3.3.4. Stability fields for lavendulan and selected Cu(II) arsenates relative to - 2+ -5 - 0 a(H2AsO4 ) and pH, when aCa = 10 . Boundaries are drawn for aCl = 10 (solid - -1 - -2 lines), aCl = 10 (dotted lines) and aCl = 10 (dashed lines).
44 levels of calcium and chloride on the diagram. The association of lavendulan with olivenite, Ni-rich paratacamite (an isomorph of atacamite) and gillardite
(Cu3NiCl2(OH)6; see following Chapter) in the Widgiemooltha 132N deposit (Nickel et al., 1994) is indicative of the variations in the activities of arsenate and chloride in the oxidised zone that are necessary to develop the assemblage.
It is possible to compare relationships between the analogous phosphates of the species considered above. Although not as common as their arsenate analogues, sampleite and the Cu(II) phosphates have been identified from numerous locations including Northparkes, N.S.W., as discussed in Chapter 2. Mineralising solutions in the supergene zone are generally more likely to be rich in arsenate rather than phosphate. This is the result of the oxidation of common primary assemblages that contain arsenopyrite, FeAsS, löllingite, FeAs2, or other As-bearing sulfosalts, which are frequently associated with complex base metal deposits. Phosphate, on the other hand, needs to be mobilised from minerals such as apatite, which may be somewhat remote from the oxidising sulfide body (Williams, 1990). Figure 3.3.5 is a stability field diagram showing relationships between atacamite, sampleite and other Cu(II) phosphates at 298.2 K. The diagram was constructed using a value for a(Na+) equal to that of the chloride ion, and 10 times that of the calcium ion. Sampleite, like lavendulan, occupies an area of the stability field diagram requiring elevated chloride activities. As was the case for the arsenates lammerite and olivenite, the stability fields of libethenite and pseudomalachite are eliminated as the activity of chloride increases. An obvious difference is the absence of a phosphate analogue of conichalcite, which is unknown naturally or as a synthetic phase. As such, the
45 23456789pH 0
Sampleite -2
Libethenite a (Cl-)=10-2 ) - -4 -1 4 a (Cl-)=10 PO 2 a (Cl-)=100 (H a -6 Pseudomalachite lg
Cornetite
-8
Atacamite
-10
Figure 3.3.5. Stability fields for sampleite and selected Cu(II) phosphates and - - 0 chlorides relative to a(H2PO4 ) and pH. Boundaries are drawn for aCl = 10 (solid - -1 - -2 lines), aCl = 10 (dotted lines) and aCl = 10 (dashed lines).
stability field of sampleite is not restricted in the same way as lavendulan by rising calcium activities. In fact, an increase in the activity of calcium serves to make the stability field for sampleite even larger.
At present, sampleite is known from 16 localities and lavendulan from 103
(www.mindat.org). This simply reflects the relative availability of arsenate and its protonated species from a wider range of primary deposits. Lower pH regimes are inevitably associated with oxidising As-rich primary phases. A simple examination of the mass balances involved with the oxidation of pyrite versus arsenopyrite bears
46 this out. Equations 3.3 to 3.5 describe the oxidation of these two phases with goethite acting as a repository for Fe(III).
2+ 2- + FeS2 + 7/2O2 + H2O → Fe + 2SO4 + 2H (3.3)
2+ 2- 3- + FeAsS + 13/4O2 + 3/2H2O → Fe + SO4 + AsO4 + 3H (3.4)
2+ + Fe + 1/4O2 + 3/2H2O → FeOOH + 2H (3.5)
The oxidation of one mol of pyrite and one mol of arsenopyrite give four and five mol of protons, respectively. In effect, if saline (chloride) groundwaters are associated with an oxidising Cu- and As-rich primary deposit, the conditions likely to give rise to lavendulan are in a sense inevitable.
3.4 SOLID SOLUTION PHENOMENA INVOLVING
LAVENDULAN AND RELATED MINERALS
Giester et al. (2007) proposed a reclassification of the lavendulan group, to include zdenekite, lemanskiite, mahnerite and shubnikovite, as well as richelsdorfite, bleasdaleite, andyrobertsite, calico-andyrobertsite(-1M) and calico-anydyrobertsite(-
2O). Little information is available in the scientific literature about solid solution and extent of substitution between members of the lavendulan group. It would be reasonable to expect that lavendulan and sampleite would exhibit solid solution at
3- 3- least to some extent, particularly as the PO4 and the AsO4 ions have the same charge and shape, about the same radius (238 and 248 pm, respectively) and practically the same conjugate acid pKa values at 298.2 K, as shown in equations 3.6 to 3.8 below, ignoring state terms (Martell and Smith, 1982).
47 - + H3XO4 ⇋ H2XO4 + H pKa = 2.148 (P), 2.24 (As) (3.6)
- 2- + H2XO4 ⇋ HXO4 + H pKa = 7.199 (P), 6.96 (As) (3.7)
2- 3- + HXO4 ⇋ XO4 + H pKa = 12.35 (P), 11.50 (As) (3.8)
The isomorphous minerals lavendulan, sampleite and zdenekite crystallise in space group P21/n (Giester et al., 2007; Zubkova et al., 2003). On the basis of the close chemical similarities above, it is likely that solid solution involving arsenate and phosphate in lavendulan and sampleite would display close to ideal behaviour.
This is supported by the close match of their stability constants. Theoretical considerations for this have been outlined by Williams (1990). An equation (3.9) can be written for phosphate-arsenate substitution as shown below.
- - NaCaCu5(AsO4)Cl·5H2O(s) + 4H2PO4 ⇋ NaCaCu5(PO4)Cl·5H2O + 4H2AsO4 (3.9)
The calculated value of ΔrG°(298.2K) for this reaction is 3.2 kJ mol-1, a rather low number with respect to the cited free energy of formation data for the species, and corresponds to an equilibrium constant of 0.27. Thus, phosphate or arsenate may substitute in the appropriate lattice more or less freely depending on their availability in mineralising solutions. A larger equilibrium constant would be necessary to limit arsenate-phosphate anion substitution in the sampleite-lavendulan pair (Williams, 1990).
Solid solution between the end members is indeed known from several localities. Lavendulan with 21 mol% substitution of phosphate for arsenate was reported from the Dome Rock mine, South Australia (Kleeman and Milnes, 1973).
Microprobe analyses of lavendulan from several other localities are shown in Table
3.4.1.
48 Table 3.4.1 Microprobe analyses (wt%) of lavendulan by locality. Chemistry Ideal Dome Rocka San Juanb Widgiemoolthac Na2O 2.92 3.3 3.1 2.84 CaO 5.28 5.7 5.8 4.84 CuO 37.45 38.2 36.3 37.46 CoO 0.03 As2O5 43.28 36.4 44.8 42.2 P2O5 6.0 2.39 Cl 3.34 5.5 3.5 3.26 H2O 8.48 6.0 7.3 7.74 -O = Cl 0.75 1.2 0.8 0.73 aKleeman and Milnes (1973), bAnthony et al. (2003), cthis work, average of 6 analyses, with H2O calculated by difference.
Analyses of Broken Hill sampleite (pictured in Figure 3.4.1) show that lead substitution for calcium can also occur (Birch, 1999), suggesting possible solid solution between sampleite and zdenekite. At the other end of the spectrum, sky-blue spherules of zdenekite, closely resembling lavendulan, have been reported by Birch
(1990) from Block 14 with up to 20 mol% calcium replacing lead (Birch, 1999).
Calcium replacing lead in zdenekite from the Cap Garonne mine, France (the type locality) with a ratio of Pb:Ca = 88:12 has been reported, with a possible restricted solid solution between lavendulan and zdenekite put forward by Chiappero and
Figure 3.4.1. Sampleite associated with cerussite, pyromorphite and libethenite from Block 14 open-cut, Broken Hill, N.S.W. Field of view is 4mm.
49 Sarp (1995). However, it is not possible to comment in the same way on Pb-Ca substitution, in the absence of stability constant data for zdenekite. It is noted in passing that successful attempts have been made to synthesise zdenekite using the same methods as employed for sampleite and lavendulan, but substituting PbCl2 for
CaCl2. However, the zdenekite obtained was always contaminated by mimetite,
Pb5(AsO4)3Cl. The yield of mimetite could be reduced by increasing the amounts of
NaCl used in the synthesis, but it was never eliminated, even under conditions of halite saturation. Zdenekite must then form at very high activities of chloride, and perhaps at lower arsenate activities. Its association with mimetite at Broken Hill is noted (Birch, 1999).
The lavendulan in the Widgiemooltha 132 N deposit, Western Australia occurs with a wide range of secondary nickel minerals (Nickel et al., 1994). Electron microprobe analyses of lavendulan from this deposit in the present study indicate substitution of phosphate for arsenate at levels of between 4 and 10 mol%. Although substitution of Ni2+ in the lavendulan lattice might be expected given the high activity of nickel in the mineralising system at the 132 N deposit, analyses show no substitution for copper. An examination of the crystal structure of lavendulan
(Giester et al., 2007) reveals that all Cu atoms in the lattice are 5-coordinate in Jahn-
Teller distorted sites. This coordination geometry is evidently unsuited to the requirements of the Ni(II) ion. Nickel does, however, substitute for copper in other copper minerals of the deposit, including paratacamite. Nickel et al. (1994) reported nickeloan paratacamite from Widgiemooltha 132 N containing up to 17.4 mol% Ni.
During this study, electron microprobe analyses of material from this locality, together with a single crystal X-ray structure, reveal a Cu:Ni ratio of 3:1 (Clissold et al., 2007). The Cu site is strongly Jahn-Teller distorted, and the Ni site is regular
50 with respect to Ni-O bond lengths but with a slight angular distortion. The resulting new mineral, gillardite, Cu3NiCl2(OH)6, is discussed in detail in Chapter 4.
Nickel-rich magnesite and dolomite are conspicuous in the upper section of the carbonate zone at Widgiemooltha 132N (Nickel et al., 1994). Figure 3.4.2 shows gillardite associated with gaspéite, NiCO3, and gypsum. The behaviour of members of the gaspéite-magnesite solid solution series in the supergene zone can be described effectively using a very simple equilibrium thermodynamic approach
(Clissold et al., 2003). This is possible because of the similarity between the radii of the Ni(II) and Mg(II) ions and their compatibility with regular octahedral coordination sites. It should be noted that there is no smooth transition between regions of different composition in samples from the 132 deposit. Rather, a number of abrupt phase changes are observed (Figure 3.4.3). Electron microprobe analyses show that the bright core in Figure 3.4.3 is gaspéite, the light edge dolomite, the dark edge magnesite and the thin bright line gaspéite. In particular, Ni-rich areas are most pronounced when they are embedded in material carrying more Mg. The deposition sequence observed is Ni-rich > Ca- and Mg-rich → Mg-rich → Ni-rich → Mg-rich.
Values of lgKsp at 298.2 K are -7.505, -8.25 and -8.35 for gaspéite, magnesite and calcite, respectively (Robie and Hemingway, 1995; Wallner et al., 2002). The simple thermodynamic model that can be used to describe the compositions of the carbonates analysed rests upon a consideration of the relative solubility products of the solid carbonate phases and their interconversions. In this connection the solid solution is assumed to behave ideally.
First for the magnesite-gaspéite pair, the following may be derived using recently determined thermodynamic data at 298.2 K (Wallner et al., 2002). For the reaction below (3.10), lgK is equal to -0.75.
51
Figure 3.4.2. Gillardite associated with gaspéite and gypsum from Widgiemooltha 132 N deposit. Field of view is 4mm.
Figure 3.4.3. Backscattered electron image of gaspéite, dolomite and magnesite (right to left) Widgiemooltha 132 N deposit. Field of view is 350 m.
52
2+ 2+ MgCO3(s) + Ni (aq) ⇋ NiCO3(s) + Mg (aq) (3.10)
This implies that for equal substitution into the lattice, the activity of Ni2+ ion only needs to be about 5 times greater than that of the Mg2+ ion. In other words, given a geochemical setting in which magnesite can form, any available Ni2+ will be easily incorporated into the magnesite lattice. Other relevant equilibrium data are tabulated in Table 3.4.2.
Table 3.4.2. LgKSP values for carbonate minerals used in calculations of relative stabilities of species at I = 0 (298.2K) a Species Formula LgKsp Reference 2+ 2- Gaspéite NiCO3(s) ⇋ Ni (aq) + CO3 (aq) -7.505 1, 2 2+ 2- Magnesite MgCO3(s) ⇋ Mg (aq) + CO3 (aq) -8.25 3 2+ 2- Calcite CaCO3(s) ⇋ Ca (aq) + CO3 (aq) -8.39 3 a[1] Martell and Smith (1982); Smith and Martell (1976, 1989);[2] Wallner et al. (2002); [3] Robie and Hemingway (1995).
Similar conclusions are reached for the substitution of Ni in calcite. For reaction 3.11;
2+ 2+ CaCO3(s) + Ni (aq) ⇋ NiCO3(s) + Ca (aq) (3.11)
log K is equal to -0.89 and the corresponding ratio of Ca2+/ Ni2+ activities amounts to approximately 0.14. The salient geochemical conclusion that may be drawn is that Ni is preferentially and selectively immobilised in the carbonate suite under equilibrium conditions.
53 Chapter 4
GILLARDITE, Cu3NiCl2(OH)6, A NEW MINERAL FROM
THE 132 NORTH DEPOSIT, WIDGIEMOOLTHA,
WESTERN AUSTRALIA
4.1 HISTORY
The 132 North deposit, is situated 5 km north-northwest of the hamlet of
Widgiemooltha, Western Australia (31° 30' S, 121° 34' E), about 80 km south of the city of Kalgoorlie (Figure 4.1.1). An open cut exposed an extensive oxidised zone developed on a komatiite-hosted sulfide deposit. The main sulfide minerals present were pyrrhotite, pentlandite, pyrite and chalcopyrite (Marston et al., 1981; McQueen,
1981). A remarkable array of secondary Ni and Cu minerals was found in the oxidised zone and these have been thoroughly documented by Nickel et al. (1994).
Among these, specimens of “nickeloan paratacamite”, with compositions reaching
Cu3NiCl2(OH)6, were noted. In light of the report of the new species herbertsmithite,
Cu3ZnCl2(OH)6 (Braithwaite et al., 2004), it seemed likely that this material was also a new species. This has proved so and the material is described below as gillardite.
The new mineral and its name were approved by the Commission on New Minerals and Mineral Names, IMA (IMA 2006-041). The holotype is housed in the Gartrell
Collection, specimen number 8774, of the Department of Earth and Planetary
Sciences, Western Australian Museum, Perth, WA. The mineral is named in honour of Professor Robert David Gillard (1936-), formerly of the Department of Chemistry,
Cardiff University, Wales, UK, in recognition of his contributions to inorganic chemistry, especially in the field of coordination chemistry.
54
Figure 4.1.1. Location of the Widgiemooltha 132 N deposit.
4.2 OCCURRENCE
Aggregates of equant, rhombohedral crystals up to 0.5 mm in size in a silicified ferruginous gossan were examined. Clusters of dark green gillardite and smaller, paler bottle-green clinoatacamite crystals are scattered across the ferruginous gossan matrix (Figure 4.2.1). Nickel et al. (1994) reported “nickeloan paratacamite” crystals up to several mm in size from throughout the carbonate zone of the oxidised profile. Associated secondary minerals in the deposit are gaspéite, gypsum, magnesite, carrboydite, motukoreaite, glaukosphaerite, hydrohonessite, retgersite, pyrolusite, huntite, aragonite, pecoraite, dolomite, kambaldaite,
55
Figure 4.2.1. Section of the holotype specimen of gillardite. Individual crystals range up to 0.3 mm in size.
annabergite, azurite, lavendulan, nepouite, nullaginite, olivenite, otwayite, pharmacosiderite, reevesite, takovite, widgiemoolthaite and kambaldaite in a silicified ferruginous gossan (Nickel et al., 1994).
4.3 CHEMICAL COMPOSITION
A crystal cluster of gillardite was embedded in epoxy resin, polished and carbon-coated. Twelve spot analyses were carried out using a Jeol JXA 8600 electron microprobe operating with an accelerating voltage of 20 keV, beam current of 20 nA, and with a 3 m beam diameter. No Zn was detected in any analysis. The sample used for microprobe analysis was somewhat unstable in the beam. A single TGA analysis (TA Instruments SDT 2960) for H2O was performed. Analytical results are given in Table 4.3.1. No other elements than those reported were detected (Jeol JXA-
840 SEM equipped with a light element detector EDS system). The empirical
56 Table 4.3.1. Analytical data for gillardite. Constituent wt% Range Probe Standard CuO 55.6 54.2-58.6 Cu NiO 15.3 14.3-17.4 Ni CoO 0.2 <0.1-0.4 Co FeO 0.1 <0.1-0.3 FeS2 Cl 17.3 15.9-18.8 TlCl H2O 13.1 Less O=Cl -3.9 Total 97.7
formula (based on 2Cl pfu) is (Cu2.865Ni0.840Co0.011Fe0.006)Σ3.722Cl2(OH)5.960.
Normalisation of the metal distribution to 4 metal ions pfu gives
(Cu3.08Ni0.90Co0.01Fe0.01)Σ4.00Cl2(OH)5.96. Spot analyses show variation of Cu:Ni ratios and metal occupancies range from (Cu3.135Ni0.853Co0.012) to (Cu2.922Ni1.058Co0.020).
Other analyses are given in Nickel et al. (1994). The simplified formula is thus
Cu3.1Ni0.9Cl2(OH)6 or Cu3NiCl2(OH)6.
4.4 PHYSICAL AND OPTICAL PROPERTIES
Equant, rhombohedral crystals showing the forms {1011}, {02 2 1}, {0001} and {1010} (probable) are dark green in colour and larger crystals are nearly black
(Figure 4.4.1). No twinning was observed. Gillardite is non-fluorescent, has a green streak and is transparent with a vitreous lustre. Mohs hardness is 3, fracture is splintery and uneven, and the tenacity is brittle. No parting was observed, but cleavage is good on {1011}. The calculated density is 3.76 g cm-3, from the single- crystal structure analysis. Gillardite is uniaxial (+), with ω = 1.836 ± 0.002, = 1.838
± 0.002 (white light). No dispersion or pleochroism was observed. The Gladstone-
Dale compatibility index, calculated using the empirical formula
(Cu3.08Ni0.90Co0.01Fe0.01)Σ4.00Cl2(OH)5.96 and unit-cell parameters from the single-
57 crystal structure analysis is -0.032 (excellent). After dissolution of gillardite in 6M
HNO3, reaction with dimethylglyoxime (DMGH) 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 adjustment of pH with ammonia gives a white precipitate of AgCl. The mineral decomposes with complete loss of water at 300oC.
Figure 4.4.1. Crystal drawing of gillardite.
4.5 SINGLE-CRYSTAL X-RAY STRUCTURE
Data collection was performed on a Bruker SMART CCD diffractometer at
273(2) K with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). Data were corrected for Lorentz, polarisation and absorption effects, the latter using
SADABS (Sheldrick, 1996). Maximum and minimum apparent transmission factors were 1.225 and 0.850, respectively. The structure was solved by direct methods
58 using SHELXS97 (Sheldrick, 1997a) and refined by least-squares procedures using
SHELXL97 (Sheldrick, 1997b). The locations of the metal (Cu and Ni) and chlorine atoms were found to be in similar positions to the analogous atoms in the R3m substructure of paratacamite reported by Fleet (1975). In line with the results of the electron-microprobe analyses, which indicate a Cu:Ni ratio of close to 3:1, the divalent copper fully occupies the MO4Cl2 site, whereas nickel occupies the non- tetragonally distorted MO6 site. This is also consistent with the metal ion distribution reported for herbertsmithite (Braithwaite et al., 2004), with which gillardite is isomorphous. A difference map then revealed the position of the hydroxyl oxygen atom, and in a similar position to that found for one of the two disordered oxygen atoms in the paratacamite substructure (Fleet, 1975). An isotropic refinement gave
R1 = 0.025 and a subsequent difference-map clearly revealed the hydrogen atom position of the hydroxyl group, ca 0.83 Å from the oxygen atom. The hydroxyl group is probably oriented in a similar fashion in herbertsmithite, but as no atom coordinates were listed (Braithwaite et al., 2004) this can only be assumed. The complete structure then refined smoothly with anisotropic displacement parameters for the non-hydrogen atoms to an R1 of 0.011. A final refinement was performed using a mixed occupancy for the nickel atom position, in line with the analytical results, (Ni0.901Cu0.081Co0.012Fe0.004). This resulted in no significant change to the structure. Final R1 and wR2 values were 0.011 and 0.027, respectively, for 448
2 2 unique reflections with I > 2σ(I). The weighting scheme used was w = 1/(σ Fo +
2 2 2 0.081P +1.700P), where P = (Fo + 2Fc )/3, as defined by SHELXL97 (Sheldrick,
1997b). Crystal data and structure refinement details are given in Table 4.5.1. Final coordinates of the atoms are listed in Table 4.5.2, anisotropic displacement parameters in Table 4.5.3, and selected bond-lengths and angles in Table 4.5.4.
59
Table 4.5.1. Crystal data and structure refinement details for Cu3NiCl2(OH)6. Chemical formula Cu3.081Ni0.903Co0.012Fe0.004Cl2O6H6 Mr 422.66 Crystal system Rhombohedral Space group R3 m Unit-cell dimensions (Å) a = 6.8364(1), c = 13.8459(4) V (Å3) 560.41(2) Z 3 Density (Calc.) (Mg m-3) 3.76 (Mo Kα) (mm-1) 11.6 F(000) 609.2 Crystal size (mm) 0.22 x 0.21 x 0.23 θ range for data collection (°) 3.74-40.33 Index ranges -11≤h≤12, -11≤k≤12, -24≤l≤22 Reflections measured 8443 Independent/observed reflections 454/448 (Rint = 0.022) Data/parameters 454/20 Goodness-of-fit on F2 1.062 Final R indices [I > 2.0σ(I)] R1 = 0.0112 (wR2 = 0.0274) Final R indices (all data) R1 = 0.0114 (wR2 = 0.0274) -3 Largest diff. peak/hole (e Å ) 0.63/-0.53
Table 4.5.2. Final atomic coordinates and equivalent isotropic displacement 2 parameters (Å ) for Cu3NiCl2(OH)6 (U(eq) is defined as one third of the trace of the orthogonalised Uij tensor). x/a y/b z/c U(eq) Ni(1) 0 0 ½ 0.0043(1) Cu(2) ½ 0 0 0.0054(1) Cl 0 0 0.1933(1) 0.0073(1) O 0.2070(1) -x 0.0624(1) 0.0073(12) H 0.1440(11) -x 0.0858(14) 0.022(6)
Table 4.5.3. Anisotropic displacement parameters (Å2) for the non-hydrogen atoms of Cu3NiCl2(OH)6 (the anisotropic displacement factor exponent takes the form - 2 2 2π [h a*2U11 + ... + 2hka*b*U12]) U11 U22 U33 U23 U13 U12 Ni(1) 0.0049(1) U11 0.0031(1) 0 0 0.5U11 Cu(2) 0.0053(1) 0.0049(1) 0.0058(1) 0.0011(1) 0.5U23 0.5U22 Cl 0.0080(1) U11 0.0059(1) 0 0 0.5U11 O 0.0067(2) U11 0.0087(3) -0.0011(1) - U23 0.0034(2)
60 o Table 4.5.4. Selected bond-lengths (Å) and angles ( ) for Cu3NiCl2(OH)6, compared with those in herbertsmithite (Braithwaite et al., 2004) and the subcell of paratacamite (Fleet, 1975) herbertsmithite gillardite Fleet (1975) M(1) = Zn M(1) = Ni M(1) = Cu M(2) = Cu M(2) = Cu M(2) = Cu M(1) – O 2.119(1) 2.0791(8) 2.11b M(2) – O 1.985(1) 1.9812(4) 1.98b M(2) – Cl 2.779(1) 2.7665(3) 2.78b O – M(1) – O a 102.93(3), 77.07(3), 180 O – M(2) – O a 98.34(5), 81.66(5), 180 O – M(2) –Cl a 97.81(2), 82.19(2), 180 O – H a 0.82(1) 1.14b H – Cl a 2.26(1) 1.95b O – Cl 3.071 3.049(8) 3.07b O – H – Cl ~166 162(2) 165.7b aNot given. bAverage distances in paratacamite with respect to split sites in space group R3 m.
No reflections corresponding to the paratacamite superlattice are present in the diffraction pattern of the gillardite structure. The OH-Cl system of hydrogen bonds supplements the Ni-O and Cu-O bonds in linking layers of metal-oxygen polyhedra, with an O-Cl distance of 3.049 Å. A selection of bond lengths and angles is given in Table 4.5.4, together with analogous values (where given) reported for herbertsmithite and the substructure of paratacamite, for comparison. The unit cell of gillardite is slightly smaller than that of herbertsmithite (a = 6.834(1), c = 14.075(2)
Å). This is consistent with the smaller radius of the Ni2+ ion, 0.69 Å, as compared to the Zn2+ ion, 0.74 Å, for coordination number 6 (Shannon, 1976).
A table of structure factors is given in the appendix (Table A.2). The
Crystallographic Information File (CIF) for gillardite (CSD-415857) is available in
Appendix Table A.3.
61 4.6 X-RAY POWDER DIFFRACTION
Powder X-ray diffraction data were recorded using a Philips PW1925-20 powder diffractometer (Ni-filtered Cu Kα radiation with pure Si as internal standard;