Diagenetic alteration of a Mesozoic fluvial gold placer deposit, southern New Zealand

Ore Geology Reviews 83 (2017) 14–29

Contents lists available at ScienceDirect

Ore Geology Reviews journal homepage: www.elsevier.com/locate/oregeo

Diagenetic alteration of a Mesozoic fluvial gold placer deposit, southern New Zealand Gemma Kerr, Kirstine Malloch, Kat Lilly, Dave Craw ⇑ Geology Department, University of Otago, PO Box 56, Dunedin 9054, New Zealand

a r t i c l e

i n f o

Article history: Received 20 October 2016 Received in revised form 11 December 2016 Accepted 15 December 2016 Available online 20 December 2016 Keywords: Alluvial Gold Clay Smectite Authigenic Groundwater

a b s t r a c t Gold paleoplacers become progressively more affected by diagenetic processes with age and burial. Mesozoic paleoplacer deposits in southern New Zealand display intermediate stages of diagenetic transformation compared to little-affected Late Cenozoic paleoplacers and strongly-affected Paleozoic and Precambrian paleoplacers. The Mesozoic (Cretaceous) diagenesis resulted in near-pervasive alteration, cementation and lithification of the paleoplacer. Lithic clasts and matrix have been extensively altered to illite, ferrous iron-bearing smectite-vermiculite, and kaolinite, and the cement consists mainly of clays and calcite. Diagenetic pyrite, marcasite, vivianite, and Mn oxide also contributed to cementation. Alteration occurred under near-surface (<500 m depth) conditions with groundwater that had circumneutral pH, high alkalinity, and elevated dissolved K, Mg and Ca. Detrital albite remained unaffected by alteration. Detrital gold has been variably dissolved and redeposited, with widespread formation of gold overgrowths on the 1–10 lm scales, with 1–3 wt% Ag. Gold mobility was driven by reduced sulphur complexes in the low redox, high pH diagenetic environment. The overgrowth gold locally contributed to cementation of fine clastic grains, and has intergrown with diagenetic clays and Mn oxide. Postdiagenetic oxidation of the paleoplacer deposit has transformed much of the pyrite to ferric oxyhydroxide and deposited some ferric oxyhydroxide coatings on gold. These oxidation processes have had only minor effects on gold mobility and textures. Hence, the low redox conditions of diagenetic gold mobility were distinctly different from those typically associated with oxidation-related supergene gold mobility. Diagenesis can affect economics of paleoplacer mining by hindering rock disaggregation during processing, coating gold particles with secondary minerals, and increasing the clay content of the deposit, all of which can lower the efficiency of gold recovery. Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction Most economic placer gold deposits occur in young fluvial sediments that were deposited in the late Cenozoic or in active river systems (Boyle, 1979; Garnett and Bassett, 2005). However, the detrital gold within these placer deposits has commonly been recycled from older sedimentary rocks, some of which contained placer gold accumulations as well (Henley and Adams, 1979; Boyle, 1979; Patyk-Kara, 1999; Garnett and Bassett, 2005; Craw, 2010, 2013; Chapman and Mortensen, 2016). Gold in some of these older deposits, commonly called paleoplacers, have in turn been recycled from even older sedimentary rocks (Boyle, 1979; Henley and Adams, 1979; Patyk-Kara, 1999; Garnett and Bassett, 2005; Lowey, 2006; Craw, 2010; Frimmel, 2014; Frimmel and Hennigh, 2015). Hence, paleoplacer deposits have significance as sources ⇑ Corresponding author. E-mail address: [email protected] (D. Craw). http://dx.doi.org/10.1016/j.oregeorev.2016.12.018 0169-1368/Ó 2016 Elsevier B.V. All rights reserved.

of gold and can be locally economic in their own right (Garnett and Bassett, 2005). Most paleoplacer deposits occur in uplifted, deformed, and partially eroded older sedimentary rocks, and their hosting sedimentary sequence has typically been extensively disrupted and removed by younger geological events (Lindgren, 1911; Henley and Adams, 1979; Patyk-Kara, 1999; Lowey, 2006; Craw, 2013). Consequently, economic Phanerozoic paleoplacers older than Cenozoic are relatively rare although some minor occurrences are known in Mesozoic sequences (e.g., Trumbull et al., 1992, Jurassic; Leckie and Craw, 1995, Cretaceous) and Paleozoic sequences (e.g., Jennex et al., 2000, Carboniferous; Dewitt et al., Cambrian). The most famous ancient paleoplacers are those of the Archean and Paleoproterozoic (Mossman and Harron, 1983; Minter et al., 1993; Frimmel et al., 2005; Frimmel, 2014; Frimmel and Hennigh, 2015). Of these, the Witwatersrand deposits have been the most productive of gold and therefore have received abundant mineralogical study (Reimer and Mossman, 1990; Minter, 1999;

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Frimmel et al., 2005; Heinrich, 2015). However, post-depositional transformations have largely overprinted the original detrital mineralogy of the Archean and Paleoproterozoic placers and at least some of the gold may have been introduced during later fluid flow events, either diagenetic, metamorphic or hydrothermal (Mossman and Harron, 1983; Phillips, 1988; Law and Phillips, 2005). All paleoplacers have undergone some degree of postdepositional alteration, and these diagenetic effects are most pronounced in older (pre-Cenozoic) paleoplacers (Garnett and Bassett, 2005). Mesozoic and Paleozoic paleoplacers are typically well lithified, which has facilitated their preservation (Boyle, 1979; Dewitt et al., 1986; Jennex et al., 2000; Garnett and Bassett, 2005). However, there is a gap in knowledge of the diagenetic processes that have affected these relatively rare older paleoplacers, because of the rarity of such deposits and the past focus on the economic and sedimentological aspects, rather than post-depositional mineralogy and textures (Dewitt et al., 1986; Leckie and Craw, 1995; Jennex et al., 2000; Garnett and Bassett, 2005). In this paper, we present mineralogical and geochemical information on diagenetic processes that have occurred in a Mesozoic gold placer in New Zealand. The hosting sedimentary rocks have preserved a record of diagenesis that has cemented and lithified the original fluvial gravels and turned them into hard scarp-forming conglomerates. The diagenetic processes were driven by water-rock interaction in the sedimentary rocks, and detrital gold was one of the minerals modified by these processes. Hence, this deposit provides a window into the post-depositional processes that progressively transform young placers into ancient paleoplacers. 2. General setting 2.1. Regional geology The Otago placer goldfield of southern New Zealand (Fig. 1) has yielded > 8 million ounces of gold, and is one of the giant placer

15

goldfields of the circum-Pacific tectonic zone (Williams, 1974; Henley and Adams, 1979; Craw, 2010, 2013). The goldfield was the site of one of the large circum-Pacific 19th Century gold rushes. The Otago gold rush started in 1861 at a site subsequently called Gabriels Gully (Fig. 1) after the initiator of the gold rush, Gabriel Read. The gold at this site was obtained from a modern stream, but it was soon traced to a source in nearby lithic and lithified Cretaceous sedimentary rocks, the Blue Spur Conglomerate (Fig. 1) from which the gold was being recycled into the modern river. Some subsequent mining activity extracted gold from those Cretaceous sedimentary rocks and similar correlatives in the general vicinity. This paper focuses on the Blue Spur Conglomerate, which is the oldest gold-bearing sedimentary unit in the Otago goldfield and one of the few pre-Cenozoic Phanerozoic gold placers of economic interest in the world. The basement for the goldfield is the Mesozoic metasedimentary Otago Schist belt and adjacent Paleozoic-Mesozoic terranes (Fig. 1). The Otago Schist was metamorphosed in the Jurassic and Early Cretaceous, and late metamorphic mineralisation processes emplaced some orogenic gold deposits, including the world-class Macraes deposit (Fig. 1; Mortensen et al., 2010). The varying structural and metamorphic levels of the schist basement (Fig. 1) were exhumed between Early and Late Cretaceous (Mortensen et al., 2010). Exhumation was partly driven by regional extensional tectonics, beginning in the Early Cretaceous, and involved differential motion on a network of normal faults (Bishop and Turnbull, 1996; Deckert et al., 2002; Mortensen et al., 2010). A second generation of orogenic gold mineralisation accompanied this regional extension (Mortensen et al., 2010). Cretaceous erosion along the normal fault scarps yielded localised deposits of lithic debris that were up to several kilometres thick in places (Bishop and Laird, 1976; Craw, 2010). The sedimentary rocks described in this study, the Blue Spur Conglomerate, accumulated adjacent to one of these normal faults, the Tuapeka Fault (Fig. 1). The Blue Spur Conglomerate is a member of the regionally extensive Taratu Formation (Fig. 1), a nonmarine to marginal

Fig. 1. Location map for the Otago placer goldfield on Otago Schist basement. Principal localities for Blue Spur Conglomerate mentioned in the text are indicated along the Tuapeka Fault system (from Bishop and Turnbull, 1996). Stratigraphic setting for the Blue Spur Conglomerate (as outlined in text) is indicated at right.

16

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

marine sedimentary sequence of Late Cretaceous age (Harrington, 1958; Bishop and Turnbull, 1996). Deposition of the Taratu Formation was partly controlled by normal faults, and the formation is up to 1 km thick near the eastern coast (Fig. 1; Harrington, 1958; Bishop and Turnbull, 1996). The formation rests unconformably on either schist basement or older fault-related lithic conglomerates (Fig. 1). Most of the Taratu Formation consists of quartz pebble conglomerates and thick coal seams, with subordinate sandstones and siltstones. However, fault-related lithic conglomeratic facies occur throughout the sequence, and the Blue Spur Conglomerate is one such deposit (Harrington, 1958; Bishop and Turnbull, 1996; Els et al., 2003). 2.2. Present remnants of Blue Spur Conglomerate The Blue Spur Conglomerate now occurs as small (<1 km2 surface area) remnants in fault angle depressions along the Tuapeka Fault, which is a regional scale multi-stranded normal fault system on the southwestern side of the Otago Schist belt (Fig. 1; Bishop and Turnbull, 1996; Els et al., 2003; Craw, 2010). The conglomerate rests unconformably on schist basement, and has been tilted and truncated by subsequent fault motion along the Tuapeka Fault (Els et al., 2003). The northern end of the Tuapeka Fault system has been reactivated as reverse faults in the Cenozoic, but the southeastern end has not been reactivated and this has allowed preservation of the Blue Spur Conglomerate remnants. The remaining occurrences of the Blue Spur Conglomerate are <100 m thick. Stratigraphic reconstruction of the area suggests that the conglomerate was unconformably overlain by Eocene quartz pebble conglomerates, remnants of which occur in the area (Fig 1; Harrington, 1958; Bishop and Turnbull, 1996), although this relationship is not preserved at any outcrops. Likewise, regional stratigraphy shows that the area was inundated by marine transgression in the middle Cenozoic (Landis et al., 2008), although the resultant deposits are not preserved in areas near the Tuapeka Fault. The total thickness of the overlying sedimentary pile that included the Blue Spur Conglomerate is inferred to have been <500 m based on nearby preserved sequences (Bishop and Turnbull, 1996). Total gold production from the Blue Spur Conglomerate is not known with any accuracy because records from the early days of the gold rush are haphazard. Also, the distinction was rarely made between production from the conglomerate itself, and gold from nearby streams that was presumably recycled from the conglomerate. Production estimates range from 0.5 to 1.5 million ounces (Els et al., 2003). A more reliable local estimate comes from a small area of conglomerate near Waitahuna, which was mined extensively in the late 19th and early 20th centuries and yielded >100 000 oz (Barnett, 2016). These estimates indicate that the Blue Spur Conglomerate was a major source of mined placer gold in the Otago goldfield, and was substantially larger than most of the individual placer deposits that have been mined over the past 160 years (Williams, 1974). Mining of the conglomerate was hindered by the relatively low gold contents of parts of the deposit, and the diagenetic cementation that is the topic of this paper. 3. Methods Outcrops of the Blue Spur Conglomerate have been variably created and destroyed by mining activity since 1861. Remaining exposures comprise a mixture of oxidised and relatively unoxidised rock, but little totally unoxidised rock is present. For this study, we obtained rock samples from the least oxidised exposures in order to examine the effects of diagenesis prior to oxidation. Minerals from some of this material have been described in general by

Craw et al. (1995), based on previous mineralogical characterisations (Craw, 1984, 1994). Original metamorphic mineralogy of nearby Otago Schist basement rocks have been characterised by Brown (1967). Mineralogy of the oxidation processes was examined in material from variably oxidised outcrops and samples. Most material for this study was obtained from a modern mining operation at Waitahuna (Fig. 1) that has exposed new outcrops and processed large volumes of conglomerate. This mining operation provided access to a wide range of gold particles and heavy mineral concentrates. Rock specimens from the Blue Spur Conglomerate were examined by standard light microscopic methods in polished thin sections. Heavy mineral concentrates, including gold, were examined initially with a stereoscopic light microscope. Representative hand-picked heavy mineral and gold particles in mine concentrates were mounted in epoxy resin blocks and polished for examination in incident light and scanning electron microscopy (SEM). Internal textures of gold particles were examined in polished sections after etching with 50% aqua regia for 5 min. This etching is necessary to remove the soft gold polishing layer and expose the internal structure. In particular, etching enhances grain textures by preferentially dissolving grain boundaries. However, the etching process causes some dissolution of other minerals inclusions, such as clays, and locally enhances gold porosity and polishing scratches. Also, etching may leave a sub-micron residue of insoluble material. Hand-picked gold particles from mine concentrates were mounted on carbon tape for examination with a Zeiss Sigma VP SEM and its energy dispersion analytical attachment (University of Otago Centre for Electron Microscopy). Most SEM images were obtained in electron backscatter mode, and some were obtained as secondary electron images. Operating voltage was 15 or 30 kV, and this caused some clay minerals to appear translucent while gold is uniformly electron-opaque. Electron backscatter diffraction (EBSD) images of grain structure were obtained from the same SEM that is fitted with a field emission gun, and operated at 30 kV, with sample tilt of 70°. Instrumentation and methods are described in more detail by Little et al. (2015). Images were obtained from flat etched gold particle surfaces in a regular stepped grid across portions of gold particles. Crystallographic orientations were imaged on a phosphor screen, recorded digitally, and automatically indexed for gold using AZTEC software (Oxford Instruments). The EBSD images of gold show well resolved crystallographic structures for coarse-grained gold, but portions of fine-grained gold images (micron to submicron scales) remained unresolved. Compositions of clay minerals were determined by spot analyses and element mapping with the SEM. The SEM beam diameter of 3 lm is greater than the width of most clay mineral laminae in the complexly intergrown clay masses. Hence, analytical results are only semiquantitative but provide ranges of major elemental variations. Element mapping and spot analyses of Ag contents of gold particle exterior surfaces was conducted with the same SEM equipment. These Ag analyses are semiquantitative as well, because of rough and irregularly oriented faces under the electron beam. Similarly, polished sections of gold that have been etched are rough and may have a residue coating. Therefore, resultant Ag analyses have an uncertainty of 1 wt% Ag, and detection limit is near to 1 wt% Ag.

4. Blue Spur Conglomerate lithology and mineralogy The conglomerate is dominated by angular to subangular poorly sorted lithic clasts that range from centimetre to metre scales (Fig. 2a). Most lithic clasts are greenschist facies and subgreenschist facies schist fragments that were derived from the

17

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Fig. 2. Blue Spur Conglomerate at outcrop and microscopic scales. (a) Outcrop photograph of poorly sorted lithic conglomerate with angular locally-derived schist clasts and rounded greywacke (Gwke) and quartz (Qtz) clasts in cemented matrix. Characteristic blue-green colour is a result of alteration of Fe-Mg phyllosilicates to ferrous ironbearing smectite-vermiculite. (b) Photomicrograph (plane polarised transmitted light) of clay-rich matrix and cement with blue-green smectite-vermiculite (SV) locally oxidised to brown clay, and colourless illite and/or illite-smectite (Ill). Clasts are epidote (Ep), albite and quartz. (c) SEM backscatter image of clay matrix and finer grained clay cement around detrital quartz clasts. (d) SEM-generated element maps (Al, Fe, Mg, K) of clay matrix and cement, showing the chemical variability around the principal clay end-members. (e) SEM backscatter image of anhedral diagenetic pyrite forming a cement. Darker grey rims on some pyrite are Fe oxyhydroxide.

immediately underlying basement and the basement on the adjacent Tuapeka Fault scarp. The conglomerate is crudely stratified with 1–10 m scale beds that are variably clast-supported and matrix-supported. The matrix consists of sand and silt clasts derived from the same local basement sources. The schist fragments in the conglomerate were unoxidised and unweathered at the time of erosion and deposition, and clast mineralogy was essentially the same as the fresh basement, with unaltered primary metamorphic ferrous iron-bearing minerals such as pyrite and chlorite (Table 1). Finer grained beds (sandstone and siltstone) consist almost entirely of schist-derived clasts, and some of these

contain minor carbonaceous debris. However, coal beds are absent from this portion of the Taratu Formation. In addition to the local basement-derived debris, there are variable proportions of rounded greywacke cobbles and quartz pebbles that were derived from more distal sources (Table 1). The quartz pebbles were derived mainly from metamorphic segregations in the highest grade Otago Schist basement to the north or east of the Tuapeka Fault zone (Fig. 1) and have undergone considerable transport with physical and chemical decomposition of most associated minerals. This quartz pebble source dominated the sediment supply for the rest of the Taratu Formation. Some quartz pebbles

Table 1 Summary of the source rock types and minerals that form the Blue Spur Conglomerate, and the secondary minerals that have formed via diagenesis and subsequent oxidation. Basement

Detrital minerals

Diagenetic minerals

Oxidation minerals

Gold

Distal & local orogenic veinhosted Au

Flakes, equant, rounded, rough, angular, <50 lm to> 1 mm

Minor remobilisation; localised intergrowths with Fe oxyhydroxide

Underlying schist source

Angular & subangular subgreenschist & lower greenschist facies schist clasts

Distal sources

Rounded greywacke cobbles; rounded quartz pebbles

Quartz, albite, muscovite, chlorite, stilpnomelane, pumpellyite, actinolite, calcite, pyrite, hematite, epidote, apatite, titanite, zircon Kaolinite, magnetite, ilmenite, pyrite (oxidised) pseudomorphs, garnet, pumpellyite

Nano- & micro-particles, crystals; nano- & micro-plates, overgrowths; cement, intergrowths Calcite Smectite-vermiculite Kaolinite Vivianite Pyrite Marcasite Mn oxide

Minor calcite Smectite Kaolinite Fe oxyhydroxide Mn oxide?

18

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Fig. 3. Blue Spur Conglomerate gold textures. (a) Typical gold particle shapes, including flakes, equant rounded particles, and angular particles. (b) A thin gold flake that has undergone some dissolution to make a ragged morphology. Flake is accompanied by detrital magnetite (Mt) and ilmenite (Ilm), and a diagenetic pyrite pseudomorph (PP). (c) Reverse side of flake in b, showing diagenetic calcite (Cc) cement and Mn oxide stain that causes detrital quartz grains (Q) to adhere to the surface. (d) Incident light view of a polished section through a gold grain (etched with aqua regia) showing internal grain structure and diagenetic overgrowth margin.

may have been derived from scattered gold-bearing vein systems as well (Fig. 1). The greywacke cobbles were derived from low grade portions protoliths of the Otago Schist belt that occur to the south and west of the Tuapeka Fault zone (Fig. 1). Mineralogy of the locally-derived conglomerate clasts is typical of low grade metasedimentary schist and is dominated by quartz, albite, phengitic muscovite, stilpnomelane and chlorite, with a range of accessory minerals (Table 1). The distally-derived greywacke cobbles have similar mineralogy. Minor detrital kaolinite is typical of the quartz pebble conglomerates elsewhere in the Taratu Formation. Detrital heavy minerals are disseminated through the conglomerate and locally concentrated at internal degradational unconformities. The heavy mineral suite (Table 1) includes minerals derived from the local basement and from more distal sources. Detrital gold is highly variable in morphology and is dominated by rounded equant particles, but includes distallyderived thin flakes and some locally-derived angular particles (Fig. 3a, b; Table 1). 5. Diagenesis and oxidation of Blue Spur Conglomerate 5.1. Principal alteration minerals Diagenetic alteration of the conglomerate was focussed on lithic clasts derived from the underlying schist basement. Alteration was almost pervasive through even the largest clasts, and the finer grained matrix has been extensively altered to form part of the cement for the conglomerate (Fig. 2a–e). Quartz and albite in the schist clasts have remained largely unaltered by these processes,

and detrital quartz and albite have retained their variably rounded and angular detrital shapes (Fig. 2b). In contrast, phyllosilicates in the schist clasts and matrix have been extensively altered throughout the conglomerate, in schist and greywacke clasts and matrix (Fig. 2b–d). This alteration occurred mainly by direct replacement of phyllosilicate grains in clasts by clay minerals with a wide range of compositions (Figs. 2d, 4a). In addition, some cementation of the matrix has occurred as these clay minerals have infilled pore spaces between clasts in the matrix (Fig. 2c). This latter clay formation was accompanied by widespread calcite cementation (Figs. 2a; 3a, b) via remobilisation of metamorphic calcite from the detrital debris. Compositional variations in the altered phyllosilicate grains occur because of interlayering of diagenetic clay laminae on the 1–10 lm scale. Phengitic muscovite has been strongly but variably illitised, with some of this alteration proceeding to form kaolinite laminae (Fig. 4a). The most prominent diagenetic clay mineral in the conglomerate is green smectite-vermiculite, which has almost entirely replaced metamorphic chlorite and stilpnomelane in the clasts and matrix (Fig. 2b–d). This clay mineral has variable proportions of Fe2+ and Fe3+ (Craw, 1984), and variable amounts of potassium in its structure (Figs. 2d, 4a). Another abundant clay mineral is Mg-bearing interstratified illite-vermiculite (Fig. 4a), which has a distinctive 12 Å structure (Craw, 1994). The most intensely altered chlorite, stilpnomelane, and muscovite grains have been replaced by kaolinite, although this extreme alteration is relatively rare. Nevertheless, both illitisation and smectitevermiculite alteration yield clay mineral compositions that generally converge towards Al-rich kaolinitic compositions (Figs. 2d, 4a).

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

19

Fig. 4. Mineralogy and geochemistry of diagenetic alteration. (a) Aluminium, Mg and K trends for diagenetic alteration of detrital metamorphic phyllosilicates to illite, smectite-vermiculite and kaolinite. Basement metamorphic mineral compositions are generalised from Brown (1967). (b) Outcrop photograph showing diagenetic alteration of metamorphic Fe-Mg phyllosilicates (MnO 0.5 to 1 wt%) to low-Mn green clays, releasing black Mn oxide on to fracture surfaces. (c) Incident light image of euhedral and subhedral diagenetic pyrite in conglomerate matrix. (d) SEM backscatter image of a polished section of a pyrite pseudomorph, showing octahedral boxwork structure with some fine intergrown gold (white). (e) Sulphur isotope data (from Tostevin et al., 2016) for diagenetic sulphides in the Taratu Formation, in relation to principal S reservoirs in the Waitahuna area.

5.2. Accessory minerals Primary metamorphic chlorite typically contained 0.5 wt% MnO and primary stilpnomelane typically contained >1 wt% MnO (Brown, 1967), whereas diagenetic smectite-vermiculite contains as little as 0.1 wt% MnO. Excess Mn expelled during clay alteration was deposited as black Mn oxide coatings within the diagenetically altered rocks, especially on the surface of quartz particles (Fig. 3c) and quartz-rich fracture surfaces (Fig. 4b). Post-diagenetic oxidation has caused some alteration of smectite-vermiculite to iron oxyhydroxide, brown smectite and minor kaolinite, and may have remobilised early-formed Mn oxides (Table 1). Clay alteration and cementation of the conglomerate was accompanied by formation of diagenetic pyrite and marcasite. Pyrite crystallisation dominated and pyrite is widespread and coarse-grained (millimetre scale), whereas marcasite is relatively rare and fine-grained (0.1 mm). Pyrite grains are commonly euhedral, and have locally overgrown clastic material and diagenetic clay minerals (Figs. 2e, 4c). Most pyrite grains have subsequently been oxidised to iron oxyhydroxide, but retain their cubic shape as well-formed pseudomorphs (Figs. 3b; 4d). Sulphur isotopic ratios for diagenetic pyrite and marcasite from conglomerates exposed in coal mines elsewhere in the Taratu Formation have a wide range of values, from < 20‰ to >+20‰, reflecting diverse diagenetic processes involving marine sulphur, basement sulphur, and bacteriogenic processes (Fig. 4e).

Minor amounts of vivianite (Fe3[PO4]2
8H2O) are intergrown with the diagenetic clay minerals in the cemented matrix of the conglomerate. The vivianite occurs as irregularly oriented stubby needles, and clusters of needles, that are 1–10 lm long and 1 lm thick. The vivianite has formed from dissolution of detrital apatite that commonly occurred as an accessory mineral in schist clasts and the matrix of the conglomerate (Table 1). The ferrous iron-bearing vivianite has formed with the mixed Fe2+-Fe3+ smectite-vermiculite clay under the relatively low redox conditions of diagenesis. The effects of subsequent oxidation on vivianite stability are not apparent in material examined in this study, as the mineral is too fine-grained to determine if it has changed colour with oxidation of ferrous to ferric iron.

6. Gold diagenesis 6.1. Larger scale features Diagenetic transformations of detrital gold particles have occurred at the 1–20 lm scale and has involved both dissolution and reprecipitation of gold (Fig. 3d). Hence, at larger scales the particles retain most of their original detrital textures (Figs. 3a–c; 5a–e). Detrital surfaces retain abundant evidence of abrasion and hammering during transport at the 0.1–1 mm scales (Figs. 3a–c; 5a–e). Hammering has made flakes of distal gold particles

20

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Fig. 5. Contrasting gold morphology in Blue Spur Conglomerate. (a) Angular particle has gold intergrown with quartz. (b) Rounded equant particle with cavities filled with diagenetic clay and stained with iron oxyhydroxide (Fe). (c) Irregular thin flake (top), rounded equant particles (centre and bottom). Lower particle has almost complete black Mn oxide coating with minor brown Fe oxyhydroxide on top. (d) SEM backscatter image of an angular crystalline gold particle (white) with intergrown quartz (black) and diagenetic clay (dark) infilling cavities. (e) Surface of a rounded equant gold particle with diagenetic clay in cavities and extensive platy gold overgrowths. (f) Iron oxyhydroxide coating (Fe) with complex and delicate intergrowths of gold, on the surface of an equant gold particle. A clastic grain is embedded in diagenetic overgrowth gold.

(Figs. 3b, 5c), and formed delicate protrusions on many flakes and more locally derived particles (Fig. 5c, d). Rounded surfaces are almost ubiquitous, even on the most angular, locally-derived particles that still contain intergrowths of hydrothermal quartz (e.g., Fig. 5a and d). Likewise, some internal crystal structures of detrital grains are still preserved from the basement vein sources (Figs. 3d; 5d). Diagenetic changes have been superimposed on all the above primary and detrital surface features to some extent at the 0.1–1 mm scales. Dissolution of gold has caused thinning of some flakes, so that they appear ragged near their edges with dissolution cavities or holes (Fig. 3b). In contrast, cavities in particles, which were formed by physical and/or chemical processes, have been filled with diagenetic clays (Fig. 5b, d, e) and clay almost completely coats some particles (Fig. 6a). Clay-filled cavities appear recessive on most gold particles because some clay was washed out during the water-based extraction procedure. The clays and adjacent gold surfaces have commonly been stained with black manganese oxide, brown iron oxyhydroxide, or a combination of both (Fig. 5b, c). Calcite cement also coats some gold particles (Fig. 3a, c). 6.2. Gold overgrowths At the ten micron scale, gold overgrowths consist of plates with a range of shapes that have coalesced to form rims on the gold particles (Fig. 5e, f). Internally, these rims are commonly porous, with micron scale vermiform cavities that contain diagenetic clays. The rim zones consist of finer grained gold (1–20 lm) than the central

portions of the grains that commonly have >50 lm crystals (Figs. 3d; 6b, c). There is a strong contrast in silver contents between the coarse-grained cores (3–9 wt% Ag) and the finer grained rims (<3 wt% Ag), as quantified in Figs. 6 and 7. The higher Ag contents of coarse grains typically, but not invariably, result in darker appearance in backscatter SEM images (Figs. 6c; 8a), and these differences are also apparent in incident light (Fig. 3d). There are commonly several generations of superimposed overgrowth layers on the exterior surfaces, and these can be distinguished by topographic steps in the overgrowths, morphological contrasts between different overgrowth generations, and/or chemical compositions (Figs. 5e, f; 7a–c). Some overgrowth generations are abundantly pockmarked with micron-scale pores that are partially filled with diagenetic clays (Figs. 7b, d; 8b). In detail, the porous example in Fig. 7a consists of overlapping plates with sub-micron thickness, and the gold in these overgrowths has 1 wt% Ag (Fig. 7b, c; left). In contrast, overgrowths with smooth surfaces coat parts of this same particle, and these smooth overgrowths are also made up of multiple generations (Fig. 7b, c; right). The smooth-surfaced overgrowths have marginally higher Ag contents, 3 wt%, than the porous overgrowths (Fig. 7b, c). Gold overgrowths have locally formed a cement for adjacent fine detrital clasts in the matrix of the conglomerate. The overgrowth gold has penetrated between clasts and filled some pore spaces (Figs. 9b, 10a, b). As a result of this cementation, sedimentary clasts commonly adhere to the outside of gold particles (Fig. 10c), or are fully enclosed (Fig. 9b). In detail, this cementing gold consists of coalesced micro- and nano-particles which are locally crystalline, including dodecahedra and hexagonal plates

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

21

Fig. 6. SEM images of textures and Ag contents of gold particle rims viewed in polished section etched with aqua regia. (a) Silicon element map showing an almost complete clay mineral coating. (b) Ag element map of the same particle as (a), showing Ag-rich core (dashed line) and an inclusion of Fe oxyhydroxide after pyrite. (c) Backscatter image of the centre of a particle, showing contrasting grain sizes (dotted boundary) and Ag contents. (d) Ag element map of a portion of particle in c. (e) Silver content variations along transect X–Y in c.

(Fig. 9b). Some of this fine-grained gold cement has in turn coalesced to form multi-generational gold plates that almost totally enclose some clasts (Fig. 9d). There is a diffuse boundary between the exterior manifestations of gold overgrowths, as described above, and the internal textures of gold particles (Fig. 8b–f). The overgrowth gold is extremely finegrained (lm to nm scale), and grains are equant and unstrained, as shown with EBSD images (Fig. 8c–f). Immediately inside the overgrown particles, there is a rim zone in which remnants of original coarse (>10 lm) grains can be recognised (Fig. 8c). These original grains commonly show internal evidence of deformation-induced variations in crystallographic orientations, and there has been extensive recrystallisation of these grains to form variably oriented fine-grained domains that overprint the primary textures (Fig. 8c–f). This rim zone with complex grain structural overprinting surrounds the core zone in which primary grain textures are preserved (Figs. 6c; 8a). It is not possible to resolve differences in Ag contents of overgrowths from those of the fine-grained, variably recrystallised, rim zones, and all this gold has 1–3 wt% Ag at the analytical scale of the SEM. However, there is generally a sharp change to higher Ag contents in the cores (Figs. 6 and 7). The most prominent internal indicator of overgrowth gold in this finegrained zone is the presence of clays filling vermiform cavities (following section) and the presence of rare included detrital clasts (Fig. 8b). 6.3. Relationships between gold and diagenetic clays Gold overgrowths have partially enclosed diagenetic clay within cavities on particle surfaces (Figs. 9d; 10a, b; 11a–d). This has resulted in complex intergrowths of gold and clays at the

micron scale at the interface between gold particles and the adjacent diagenetic cement (Figs. 10a–d, 11a–d). Overgrowth gold protrudes from the gold particles into the clay accumulations and included fine detrital clasts (Fig. 10b, c). Within the clay cement, gold forms micron scale delicate lacy networks and platy crystals amongst the clay particles, and gold nanoparticles are scattered through the clay patches (Figs. 10d, 11a–d). Manganese oxide occurs with gold within the clay cement and as coatings on clastic grains (Fig. 9b, c). The Mn oxide has apparently formed coevally with gold cement and has locally been overgrown by later cementing gold (Fig. 10b, c). Rare vivianite occurs within the clay cement patches, and is intimately intergrown with cementing gold and gold nanoparticles (Fig. 11a–c). The vivianite needles are irregularly oriented within the clay patches and are most closely associated with clay crystals rather than gold. However, gold has locally overgrown some vivianite needles within the clay (Fig. 11c). 6.4. Gold in iron oxyhydroxide Iron oxyhydroxide is widespread through the Blue Spur Conglomerate as a result of post-diagenetic oxidation, but there has been only minor chemical mobility of gold during this oxidation (Table 1). Iron oxyhydroxide coatings typically occur on pyrite grains (Fig. 2e), and pseudomorphs of diagenetic pyrite can contain minor amounts of gold that have intergrown with the iron oxyhydroxide (Fig. 4d). Iron oxyhydroxide coatings on gold particles are relatively rare compared to the common Mn oxide coatings. Most such Fe-rich coatings are incipient stains superimposed on to pre-existing diagenetic minerals such as Mn oxide and clay minerals (Fig. 5b, c). There has been minor oxidation-related

22

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Fig. 7. SEM views of textures of multigenerational diagenetic gold rims. (a) General view of particle surface, showing gold overgrowths that have included some detrital silicates (dark). At least two distinct overgrowth generations are visible, with contrasting smooth and pock-marked surfaces. (b) Silver map of the large box in a, showing the generally higher Ag contents of the smooth-surfaced overgrowth generation. (c) Close view of the boundary between distinct overgrowth textures, as in small rectangle in a, showing thin platy overgrowths on the porous (clay-filled) surface on left, and stepped smooth surface on right. Contrasting Ag contents are indicated.

remobilisation of diagenetic overgrowth gold, leading to complex intergrowths of iron oxyhydroxide and gold (Fig. 5f), but such textures are rare. These complex intergrowths include micron scale gold plates and amorphous gold protrusions, and some nanoparticulate gold distributed through the iron oxyhydroxide. 7. Discussion 7.1. Inferred geochemical environment The Blue Spur Conglomerate has been extensively cemented and lithified (Fig. 2a), so there is no remaining evidence of the water that caused the diagenetic alteration of the rocks. However, some inferences of the geochemical environment can be made by examination of similar modern environments in which water is interacting with debris derived from Otago Schist basement. We have chosen two possible modern analogues that involve groundwater passing through immature lithic schist debris. The Taieri basin (Fig. 1; Litchfield et al., 2002) is a faulted depression into which abundant Pleistocene schist debris has been deposited by fluvial and debris flow processes. A large debris dam at Macraes gold mine (Fig. 1; Craw, 2000) has been constructed from fresh schist waste rock, and groundwater passing through this dam was analysed in the early life of the mine (1990s). Water compositions from both these analogues provide some insights into the likely water compositions in the Blue Spur Conglomerate as diagenetic cementation occurred. The most prominent feature of waters that have interacted with schist debris is the rapid increase in dissolved Ca and associated alkalinity from dissolution of metamorphic calcite (Fig. 12a). This calcite dissolution ensures that the pH of waters remained circumneutral. This is a widespread feature in the Otago Schist belt

that occurs in all river waters almost instantly when rain water reaches the rocks (Jacobson et al., 2003), and has resulted in localised deposition of calcite from groundwater (Craw and Lilly, 2016). The abundant cementing calcite in the Blue Spur Conglomerate attests to the same processes occurring during diagenesis, which implies circumneutral to alkaline pH during rock cementation. Chlorite, stilpnomelane and muscovite are other metamorphic minerals that readily interact with groundwater to release labile components, and dissolved Mg and K have risen accordingly (Fig. 12b). In particular, smectitic clay alteration of chlorite has occurred rapidly (time scale of years) in the Macraes debris dam (Craw, 2000), and has been a common natural weathering phenomenon in the area (Brown, 1967). Stilpnomelane is rare in the Taieri basin fluvial sediments and absent from the Macraes mine rocks, but that mineral has similar responses to weathering processes as chlorite (Brown, 1967). Illitisation of muscovite has occurred in parallel with chlorite alteration, but to a lesser extent (Fig. 12b). These processes are similar to, albeit less advanced than, the diagenetic processes in the Blue Spur Conglomerate (Fig. 4a). 7.2. Redox conditions and gold mobility during diagenesis Dissolved sulphur levels are typically between 10 and 100 mg/L in Otago Schist groundwaters (Craw, 2000; Litchfield et al., 2002; Craw and Lilly, 2016). This dissolved sulphur has been derived from dissolution of metamorphic pyrite (Table 1) and input of marine aerosols in rain (Youngson, 1995; Craw and Lilly, 2016; Tostevin et al., 2016). Speciation of that dissolved sulphur at circumneutral pH is controlled by redox conditions in the aquifers, as is dissolved Fe. Consequently, saturated gravels below the water

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

23

Fig. 8. Internal grain structure of gold particles in polished section (etched with aqua regia). (a) Strip view across a gold particle, with enhanced contrast to show differing grain sizes in core and rim, with differing Ag contents as annotated. (b) Porous rim zone of a gold particle, showing vermiform clay inclusions in fine-grained overgrowth gold, in which grain texture is indicated by mottled appearance. A detrital apatite grain has also been overgrown by gold. (c) EBSD image (500 nm spot spacing) showing crystallographic orientations in a transect from interior (top) to edge (bottom) of a gold particle, with a thin overgrowth layer. Original interior coarse grain texture is apparent. Dark diagonal lines are polishing scratches. (d) Same area as c, viewing perpendicular crystallographic orientations and showing widespread recrystallisation to smaller grains in the interior. (e) EBSD image (300 nm spot spacing) of the edge of a gold particle, showing fine-grained overgrowth gold (bottom) merging with elongate recrystallised grains (top). (f) EBSD image (100 nm spot spacing) of a portion of the edge of a gold particle, showing fine-grained overgrowth gold near bottom and finegrained recrystallised gold at top. Most black areas are gold but are unresolved crystallographically.

table typically have authigenic pyrite and/or marcasite, whereas more oxidised conditions have yielded iron oxyhydroxide and dissolved sulphate (Tostevin et al., 2016). Some of this iron sulphide deposition was primarily driven by inorganic processes and the resultant sulphur isotopic ratios reflect the principal source reservoir (Fig. 4e). There may have been a minor component of sulphide reduction mediated by bacteria in the Taratu Formation (Fig. 4e) but the bacterial effect is distinctly less pronounced than in nearby Pliocene paleoplacers (Falconer et al., 2006). The abundant diagenetic pyrite and detrital magnetite in the Blue Spur Conglomerate (Fig. 3b) attests to the low redox environment of the diagenetic processes, as does the presence of vivianite (Table 1; Fig. 11) and the presence of ferrous iron in the diagenetic smectite (Craw, 1984). Preservation of detrital albite is a common feature of lowredox groundwater alteration conditions in the Otago Schist, whereas under oxidising conditions albite has undergone some dissolution to release dissolved Na into groundwater, and in older oxidised alteration zones albite has been extensively altered to kaolinite (Craw, 1994, 2000). Post-diagenetic uplift and erosion allowed incursion of more oxygenated groundwater into the Blue Spur Conglomerate, and the diagenetic pyrite has been extensively oxidised to iron oxyhydroxide (Table 1). Likewise, the ferrous iron in smectitic clay has been variably oxidised to cause iron oxyhydroxide staining of some diagenetic clay patches (Figs. 2b, 5b). Oxidation of clay caused transformation of smectite-vermiculite to other related clays (Table 1), although these are too fine-grained to fully characterise. However, the widespread Mn oxide in diagenetic cements and stains on quartz is commonly separate from the iron oxyhydroxide and occurs with ferrous iron-bearing minerals (Table 1; Figs. 4b;

10b). The Mn oxide appears to be primarily a product of low redox diagenesis rather than later oxidation, although some later remobilisation with iron oxyhydroxide may have occurred as well (Fig. 5c). The varying redox conditions inferred in the previous paragraph can be modelled to quantify relative mineral stabilities during diagenesis of the Blue Spur Conglomerate (Fig. 13a, b). Diagenesis occurred under low redox conditions where pyrite was stable in the presence of detrital magnetite (Fig. 13a). This was apparently near to, but below, the sulphate stability boundary where pyrite and ferrous iron-bearing smectites (saponite and nontronite in these models) can coexist with ferrous iron-bearing vivianite (Fig. 13a, b). Amorphous manganous hydroxide was formed with this low-redox iron mineral assemblage, and this supports the above inferences of a high pH during diagenetic alteration (Fig. 13a). Illite and smectite (beidellite) formed from phengite and chlorite in groundwaters with compositions similar to the modern analogues, as modelled in Fig. 13c. These models used end-member compositions whereas the actual diagenetic processes yielded a range of clay mineral compositions with solid solution and interlayering (Figs. 2d, 4a). 7.3. Gold mobility Our textural and mineralogical observations show that there is a close spatial association between diagenetic gold overgrowths and ferrous iron-bearing minerals such as smectite-vermiculite clays and vivianite (Table 1; Figs. 6–11). These observations suggest that gold dissolution and overgrowths, and the cementation that accompanied diagenesis, occurred under relatively low redox

24

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Fig. 9. Textural relationships between diagenetic gold overgrowths and detrital silicate grains. (a) SEM backscatter image of diagenetic gold (white) forming cement and overgrowths around and over silicate gains (dark). (b) Close view of box in a, showing gold cementing silicates. Box highlights gold microcrystals: a dodecahedron at top left, and a hexagonal plate at lower right. (c) Gold particle with overgrowth gold that has partially encapsulated numerous detrital silicate grains (arrowed). (d) Detailed structure of gold overgrowths on the surface of a gold particle, showing several generations of elongate and equant plates.

Fig. 10. Backscatter SEM images of textural relationships between gold (white) and diagenetic clay (dark). (a) Clay fills a cavity between protrusions of gold particle surface. Gold surface has been etched during diagenesis by dissolution that has exposed and enhanced the internal grain structure. (b) Clay is intergrown with gold overgrowths, and these authigenic minerals have partially encapsulated detrital quartz and albite. Manganese oxide staining of a quartz grain in square is enlarged with enhanced contrast in c. (d) Complex overgrowths and veinlets of gold are intimately mixed with clay.

conditions in which reduced sulphur and ferrous iron dominated (Fig. 13a, b). Gold is soluble in low-redox circumneutral groundwaters, such as those described above, primarily as reduced sulphur

complexes such as Au(HS)2 (Webster, 1986; Heinrich, 2015). We suggest that this is the most likely chemical process that drove the diagenetic gold mobilisation.

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

25

Fig. 11. SEM images of overgrowths on gold particles, showing textural relationships of diagenetic gold, clays, and vivianite crystals (V). (a) Platy gold overgrowths (white) have cavities filled with clay, micro- and nano-particulate gold, and vivianite needles. (b) Intergrowths of nano- and micro- particulate gold, clay, and vivianite. (c) Closer view of similar material to b, showing gold that has overgrown vivianite crystals (in box). (d) Crystalline gold plates and small hexagonal gold crystals (centre) have overgrown clay on a gold particle.

Gold is also soluble as metastable thiosulphate complexes under more oxidised circumneutral pH conditions (Webster, 1986; Heinrich, 2015; Craw and Lilly, 2016). Thiosulphate ions can form temporarily during sulphide mineral oxidation (Rimstidt and Vaughan, 2003; Melashvili et al., 2015), and this probably occurred during post-diagenetic oxidation of pyrite in the Blue Spur Conglomerate (Fig. 4c). Gold remobilisation under relatively oxidising conditions, in which sulphides have been transformed to iron oxyhydroxide, is widespread in young (Miocene-Pleistocene) paleoplacers in the Otago placer goldfield (Hesson et al., 2016; Craw and Lilly, 2016). Much of the remobilised gold is intimately dispersed through, or overgrown on, iron oxyhydroxides as nano- and micro-particulate and crystalline intergrowths. Similar textures do occur in the Blue Spur Conglomerate (Fig. 5f) but they are notably rare compared to the almost ubiquitous diagenetic gold textures that formed under low redox conditions. The processes, chemistry, and associated mineralogy of the low redox gold mobility in the Blue Spur Conglomerate contrasts strongly with the more common oxidation-dominated supergene processes that occur in many gold deposits around the world (Stoffregen, 1986; Gao et al., 1995; Yesares et al., 2014; Craw et al., 2015). Detrital gold particles in many placer deposits have Ag-depleted rims, typically 10 lm wide, that are a result of processes occurring during fluvial transport (Groen et al., 1990; Knight et al., 1999a,b). This Ag depletion may have occurred as a result of deformation-induced recrystallisation, as has been inferred for the Blue Spur Conglomerate gold (Fig. 8). This recrystallisation and Ag leaching may have occurred during transport, after deposition, or at both of these times (Groen et al., 1990; Knight et al., 1999a,b). There is a diffuse boundary between diagenetic overgrowth gold and possible relict detrital particle rims in the Blue Spur Conglomerate, and all this gold is extremely fine-grained

(lm to nm scale; Fig. 8). Hence, the distinction between overgrowth gold and original detrital gold has become indistinct. The vermiform porous structures in gold particle rims (e.g., Figs. 7, 8b) probably occur within gold overgrowths that have enclosed vermiform diagenetic clays, as has been observed on younger gold paleoplacers in the Otago goldfield (Youngson and Craw, 1993; Falconer and Craw, 2009). On the other hand, relict coarse gold grains can be detected close to some particle edges (Fig. 8c), attesting to only limited post-depositional changes in these particles. It is likely that the effects of both detrital deformation and subsequent diagenesis have merged to form the finegrained low-Ag rims of particles (Figs. 6c; 8a, b). These processes have led to extensive recrystallisation of the particle rims to finegrained low-Ag gold, with localised addition of 5–20 lm gold overgrowths. 7.4. Implications for ancient paleoplacers The processes and results of diagenetic alteration of a gold placer in lithic sedimentary rocks, as documented above, are summarised in Fig. 14a–c. Clearly, gold cannot be created by these processes, but there have been significant changes in shape of gold particles because of dissolution and reprecipitation of gold. These processes occurred at the 10 lm scale and smaller, and so the significance of these changes to coarse gold particles is trivial. However, the smaller the gold particles, the greater the relative significance of these processes. For example, a 10 lm overgrowth on a 1 mm diameter spherical particle would change the volume of that particle by only 6%, but the same overgrowth would change the volume of a 100 lm particle by 33%, and a 50 lm particle would change volume by 73%. These effects are minima when considering the convoluted shape of many particles and their overgrowths (Figs. 3d, 14c), but associated dissolution of parts of

26

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

Fig. 12. Geochemical plots for groundwaters that have interacted with lithic debris that has been derived from Otago Schist basement and are similar to Blue Spur Conglomerate. Fluvial gravels are Pleistocene (Litchfield et al., 2002), and the debris dam is at Macraes mine (Fig. 1; Craw, 2000).

particle rims (Figs. 3b; 10a; 14b) lowers the relative contributions of the overgrowths. Nevertheless, the effects of these processes on size and shapes of placer gold particles can be significant for finegrained particles such as much of the gold in the Blue Spur Conglomerate (Fig. 3a). Diagenetic transformation of paleoplacers became more prominent with age. Pliocene paleoplacers in the giant Klondike goldfield of Yukon Territory, Canada have significant clay alteration and cementation, although lithic clasts have remained largely intact (Fig. 14d; Lowey, 2002, 2006). Placer gold in the Klondike goldfield is notable for its lack of overgrowths and the resultant utility of gold characteristics for linking placer gold to sources (Knight et al., 1999a,b; Chapman et al., 2011). However, some gold from the Pliocene paleoplacers has been extensively internally recrystallised, leaving only relict cores similar to those seen in Blue Spur Conglomerate (Fig. 14e; Knight et al., 1999a). Eocene paleoplacers of the giant Californian goldfield (Lindgren, 1911) have undergone considerable diagenetic alteration and kaolinitisation of lithic clasts (Fig. 14f; Mulch et al., 2006). In parallel with this alteration, diagenetic pyrite growth has been extensive in low-redox parts of the sedimentary pile, and gold from these parts of the sequence has convoluted overgrowths on rounded detrital surfaces (Fig. 14g; Marsh et al., 2014). More thorough post-depositional alteration and lithification of sedimentary hosts of paleoplacers have occurred in the Witwater-

Fig. 13. Model geochemical and mineralogical relationships (at 25 °C and circumneutral pH, from Geochemists Workbench) for low redox species (green areas) relevant to diagenesis of the Blue Spur Conglomerate, with water compositions (blue boxes) inferred from mineralogy and the modern groundwater compositional analogues (e.g., Fig. 10). (a) Manganese and iron minerals. (b) Magnesium, Fe and P mineral relationships, in comparison to post-diagenetic oxidation (brown). (c) Calcium and K silicate mineral stabilities.

srand goldfield of South Africa, resulting in recrystallisation of any lithic matrix (Fig. 14h; Phillips, 1988; Law and Phillips, 2005). Remnants of detrital gold are fine-grained and have convoluted outlines (Fig. 14i; Minter et al., 1993; Minter, 1999) similar to some

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

27

Fig. 14. Diagenetic alteration of gold placer deposits. (a) and (b) Cartoon summary of alteration of the Cretaceous Blue Spur Conglomerate of New Zealand (this study), with (c) an example of the resulting modified gold particle texture. (d) Incipient diagenetic alteration in Pliocene White Channel Gravel placer, Yukon, Canada. Basement schist fragments (B) have begun to disintegrate and the matrix has weak clay alteration. Q = quartz pebbles. (e) Outline of a gold particle from White Channel Gravel (cores dotted), from image in Knight et al. (1999a). (f) Advanced diagenetic clay alteration of basement clasts (B = granite, schist) and initial cementation of matrix in Eocene quartz pebble conglomerate placer, California. (g) Outline of a gold particle from the Eocene placer (traced from an image collected by Marsh et al. (2014)). (h) Archean quartz pebble conglomerate from Witwatersrand goldfield, South Africa, with extensively recrystallised and cemented matrix. P = pyrite. (i) Outline of a presumed relict Archean detrital gold particle (sketched from photograph in Minter (1999)).

in the Blue Spur Conglomerate (e.g., Fig. 3d) and the Californian paleoplacers (e.g., Fig. 14g; Marsh et al., 2014). In addition to these apparent diagenetic changes to the Witwatersrand paleoplacers and their gold, there is also evidence for substantial remobilisation and/or addition of gold to these sedimentary rocks, possibly via hydrothermal fluids (Law and Phillips, 2005). 7.5. Economic implications Mining of placer gold deposits involves (i) disaggregation of the hosting rocks, (ii) liberation of the gold from those rocks, and (iii) concentration of the gold, typically by gravity-driven settling in water. These processes are readily achieved with modern placer deposits, where the hosting sediments are unconsolidated. However, paleoplacers became progressively more consolidated and lithified by diagenetic processes with increasing age (Fig. 14). Hence, the three principal processing stages (above) are more difficult and less efficient for older and diagenetically lithified paleoplacers. Modern placers and young paleoplacers are easily processed because individual clasts retain their original structural integrity, disaggregation involves separation of intact clasts around their original outer surfaces, and the sediment will break up into its original clastic constituents (Fig. 14d). As diagenesis proceeded, alteration of clasts and cementation of the intervening matrix decreased, or even reversed, the contrast in physical competence between clasts and matrix, so that the rocks break across clasts

and cemented matrix (Figs. 2a; 14f, h). Rock disaggregation can require additional effort during processing under these circumstances. The well-lithified and largely recrystallised Archean paleoplacers (e.g., Fig. 14h) require extensive crushing and grinding, and few clasts break around their depositional margins. The Blue Spur Conglomerate described in this study represents a transitional stage in diagenetic lithification whereby some clasts, such as quartz and greywacke, are readily separated around original clastic margins but altered schist clasts have similar competence to the cemented matrix and the rock commonly breaks uniformly across clasts and matrix instead (Fig. 2a). Diagenetic cementation necessitated blasting and/or crushing of some Blue Spur Conglomerate during historic mining, but also permitted underground mining of richer portions (Barnett, 2016). This alteration and cementation resulted in mineralogical incorporation of gold into the diagenetic rock structure, including coatings by secondary minerals (Figs. 3, 5 and 10), or even encapsulation (Figs. 4c; 8b). Liberation and separation of this gold is then more difficult because composite particles of gold with these other minerals have lower density than gold particles alone, and mineral coatings resist separation by chemical processes such as cyanidation or mercury amalgamation. The alteration associated with early stages of diagenesis (Figs. 4a; 14f) increased the clay content of the paleoplacers, and that fine-grained clay is readily mobilised in suspension in processing water (Druzbicka and Craw, 2012). The suspended clays increase the density of the processing water and this can further

28

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29

inhibit the efficiency of gravity separation of gold. The resultant suspended sediment also has potential downstream environmental issues as well (Druzbicka and Craw, 2012). Advanced diagenetic processes have produced more coherent rocks with coarser grained phyllosilicates, such as pyrophyllite in the Witwatersrand deposits (Fig. 14h; Phillips, 1988; Law and Phillips, 2005), and the suspended sediment issues are less significant.

insights provided by Ross Barnett. SEM images were obtained at the Otago Centre for Electron Microscopy. Brent Pooley created useful sections of difficult specimens. Discussions with Donna Falconer, Jon Lindqvist, and John Youngson helped us to develop some ideas expressed herein. Constructive reviews by Hartwig Frimmel and Jim Mortensen improved the presentation of the ms. References

8. Conclusions The Cretaceous Blue Spur Conglomerate of southern New Zealand provides a useful view into diagenetic alteration processes that have affected paleoplacer gold deposits. Diagenetic effects increased in intensity with time, and the amount of diagenesis of Blue Spur Conglomerate is intermediate between largely unaltered Late Cenozoic paleoplacers and well-lithified pre-Mesozoic paleoplacers that include the well-known Archean deposits such as Witwatersrand. The Blue Spur Conglomerate is moderately well lithified and cemented by diagenetic processes so that it forms erosion-resistant cliffs and historic mining at times involved some blasting and crushing to liberate the gold. Diagenetic alteration of paleoplacers like the Blue Spur Conglomerate can affect mine economics by decreasing the efficiency of host rock disaggregation, gold liberation, and gold concentration. Further, fine-grained diagenetic minerals can raise the suspended sediment content of processing waters and affect downstream water quality at mine sites. Diagenetic alteration of the Blue Spur Conglomerate was dominated by in situ transformation of phyllosilicates in lithic clasts and matrix to clay minerals. Basement-derived metamorphic chlorite and stilpnomelane in clasts were initially altered to ferrous iron-bearing smectite-vermiculite, and phengitic muscovite altered to illite. More intense alteration yielded interstratified illite-vermiculite and ultimately kaolinite. Manganese liberated from altered phyllosilicates formed Mn oxide that is intergrown with clay minerals and coats some surfaces. Clay minerals and Mn oxide combined with diagenetic pyrite, marcasite, vivianite, and calcite to cement the conglomerate. Detrital gold occurs as flakes, equant rounded particles, and angular particles typically between 20 lm and 1 mm across. Gold was mobilised by reduced sulphur complexes during diagenesis by low redox groundwater with circumneutral pH and elevated alkalinity and dissolved K, Mg and Ca. This led to localised dissolution of gold surfaces and widespread redeposition of gold overgrowths with varying Ag contents (1–3 wt%). The gold overgrowths were intimately intergrown with ferrous iron-bearing minerals including smectite-vermiculite, vivianite and Fe sulphides. Associated recrystallisation of interior rims of particles to fine-grained low-Ag gold has left relict cores of the original detrital gold. Subsequent oxidation of the deposit after uplift and partial erosion has transformed much of the Fe sulphides to Fe oxyhydroxide, and formed localised Fe oxyhydroxide coatings on gold and clays. Only minor gold remobilisation occurred during this oxidation process. Hence, the diagenetic gold mobilisation was distinctly different from that associated with most supergene gold mobilisation, which is dominantly an oxic process closely associated with Fe oxyhydroxides. Acknowledgements Funding for this study was provided by the Marsden Fund administered by the Royal Society of New Zealand. Alan Roberts provided excellent logistical support and hospitality at the Waitahuna Gully mine, and Barry MacDonell provided useful discussions on geological aspects of the site. We appreciate the historical

Barnett, R., 2016. Gold Entrepreneurs: The Norwegian Party at Waitahuna Gully, Otago. Lawrence Athenaeum and Mining Institute, Lawrence, New Zealand, p. 216. Bishop, D.G., Laird, M.G., 1976. Stratigraphy and environment of deposition of the Kyburn Formation (Cretaceous), a wedge of coarse terrestrial sediments in Central Otago. J. R. Soc. New Zealand 6, 55–71. Bishop, D.G., Turnbull, I.M., 1996. Geology of the Dunedin area. GNS Science 1:250 000 Geological Map, Lower Hutt, New Zealand, Sheet 21 + 52 pp. Boyle, R.W., 1979. The geochemistry of gold and its deposits: Geol Surv Canada Bull 280, 579 pp. Brown, E.H., 1967. The greenschist facies in part of eastern Otago, New Zealand. Contrib. Mineral Petrol. 14, 259–292. Chapman, R.J., Mortensen, J.K., LeBarge, W.P., 2011. Styles of lode gold mineralization contributing to the placers of the Indian River and Black Hills Creek, Yukon Territory, Canada as deduced from microchemical characterization of placer gold grains. Mineral Deposita 46, 881–903. Chapman, R.J., Mortensen, J.K., 2016. Characterization of gold mineralization in the Northern Cariboo Gold District, British Columbia, Canada, through integration of compositional studies of lode and detrital Gold with historical placer production: a template for evaluation of orogenic gold districts. Econ. Geol. 111, 1321–1345. Craw, D., 1984. Ferrous-iron-bearing vermiculite-smectite series formed during alteration of chlorite to kaolinite, Otago Schist, New Zealand. Clay Miner. 19, 509–520. Craw, D., 1994. Contrasting alteration mineralogy at an unconformity beneath auriferous terrestrial sediments, central Otago, New Zealand. Sed. Geol. 92, 17–30. Craw, D., 2000. Water-rock interaction and acid neutralization in a large schist debris dam, Otago, New Zealand. Chem. Geol. 171, 17–32. Craw, D., 2010. Delayed accumulation of placers during exhumation of orogenic gold in southern New Zealand. Ore Geol. Rev. 37, 224–235. Craw, D., 2013. River drainage reorientation during placer gold accumulation, southern New Zealand. Mineral Deposita 48, 841–860. Craw, D., Lilly, K., 2016. Gold nugget morphology and geochemical environments of nugget formation, southern New Zealand. Ore Geol. Rev. 79, 301–315. Craw, D., Smith, D.W., Youngson, J.H., 1995. Formation of authigenic Fe2+-bearing smectite-vermiculite during terrestrial diagenesis, southern New Zealand. NZ J. Geol. Geophys. 38, 151–158. Craw, D., MacKenzie, D.J., Grieve, P., 2015. Supergene gold mobility in orogenic gold deposits, Otago Schist, New Zealand. NZ J. Geol. Geophys. 58, 123–136. Deckert, H., Ring, U., Mortimer, N., 2002. Tectonic significance of Cretaceous bivergent extensional shear zones in the Torlesse accretionary wedge, central Otago Schist, New Zealand. NZ J. Geol. Geophys. 45, 537–547. Dewitt, E., Redden, J.A., Wilson, A.B., Buscher, D., 1986. Mineral resource potential and geology of the Black Hills National Forest, South Dakota and Wyoming. USGS Bull 1580, 140 pp. Druzbicka, J., Craw, D., 2012. Turbidity development and dissipation in paleoplacer gold deposits, southern New Zealand. Environ. Earth Sci. 68, 1575–1589. Els, B.G., Youngson, J.H., Craw, D., 2003. Blue Spur Conglomerate: auriferous Late Cretaceous fluvial channel deposits adjacent to normal fault scarps, southeast Otago, New Zealand. NZ J. Geol. Geophys. 46, 123–139. Falconer, D.M., Craw, D., 2009. Supergene gold mobility: a textural and geochemical study from gold placers in southern New Zealand. Econ. Geol. Special Publ. 14, 77–93. Falconer, D.M., Craw, D., Youngson, J.H., Faure, K., 2006. Gold and sulphide minerals in Tertiary quartz pebble conglomerate gold placers, Southland, New Zealand. Ore Geol. Rev. 28, 525–545. Frimmel, H.E., 2014. A giant mesoarchean crustal gold-enrichment episode; possible causes and consequences for exploration. Econ. Geol. Special Publ. 18, 209–234. Frimmel, H.E., Hennigh, Q., 2015. First whiffs of atmospheric oxygen triggered onset of crustal gold cycle. Mineral Deposita 50, 5–23. Frimmel, H.E., Groves, D.I., Kirk, J., Ruiz, J., Chesley, J., Minter, W.E.L., 2005. The formation and preservation of the Witwatersrand goldfields, the world’s largest gold province. Econ. Geol. 100th Anniv Vol., 769–797. Gao, Z.L., Kwak, T.A.P., Changkakoti, A., Hussein, E., Gray, J., 1995. Supergene ore and hypogene nonore mineralization of the Nagambie sediment-hosted gold deposit, Victoria, Australia. Econ. Geol. 90, 1747–1763. Garnett, R.H.W., Bassett, N.C., 2005. Placer Deposits. Econ. Geol. 100th Anniv Vol, 813–843. Groen, J.C., Craig, J.R., Rimstidt, R.D., 1990. Gold-rich rim formation on electrum grains in placers. Canadian Mineral 28, 207–228. Harrington, H.J., 1958. Geology of the Kaitangata Coalfield. NZ Geol Surv Bull 59, 131 pp.

G. Kerr et al. / Ore Geology Reviews 83 (2017) 14–29 Heinrich, C.J., 2015. Witwatersrand gold deposits formed by volcanic rain, anoxic rivers and Archaean life. Nat. Geosci. 8, 206–209. Henley, R.W., Adams, J., 1979. On the evolution of giant gold placers. Trans. Inst. Min. Metall. 88, B41–B50. Hesson, M., Stewart, J., Stephens, S., Kerr, G., Craw, D., 2016. Gold nuggets in proximal placers, Old Man Range, Central Otago. In: Christie, A.B. (Ed.), Mineral Deposits of New Zealand: Exploration and Research, Australasian Institute of Mining and Metallurgy Monograph, vol. 31, pp. 359–366. Jacobson, A.D., Blum, J.D., Chamberlain, C.P., Craw, D., Koons, P.O., 2003. Climatic and tectonic controls on chemical weathering in the New Zealand Southern Alps. Geochim. Cosmochim. Acta 67, 29–46. Jennex, L., Murphy, J., Anderson, A., 2000. Post-orogenic exhumation of an auriferous terrane: the paleoplacer potential of the Early Carboniferous St. Marys Basin, Canadian Appalachians. Mineral Deposita 35, 776–790. Knight, J.B., Morison, S.R., Mortensen, J.K., 1999a. Lode and placer gold composition in the Klondike District, Yukon Territory, Canada: implications for the nature and genesis of Klondike placer and lode gold deposits. Econ. Geol. 94, 635–648. Knight, J.B., Morison, S.R., Mortensen, J.K., 1999b. The relationship between placer gold particle shape, rimming, and distance of fluvial transport as exemplified by gold from the Klondike district, Yukon Territory, Canada. Econ. Geol. 94, 635–648. Landis, C.A., Campbell, H.J., Begg, J.G., Mildenhall, D.C., Paterson, A.M., Trewick, S.A., 2008. The Waipounamu erosion surface: questioning the antiquity of the New Zealand land surface and terrestrial fauna and flora. Geol. Mag. 145, 173–197. Law, J., Phillips, G.N., 2005. Hydrothermal replacement model for Witwatersrand gold. Econ. Geol. 100th Anniv Vol., 799–811. Leckie, D.A., Craw, D., 1995. Westerly derived Early Cretaceous gold paleoplacers in the Western Canada foreland basin, southwestern Alberta: tectonic and economic implications. Can. J. Earth Sci. 32, 1079–1092. Lindgren, W., 1911. The Tertiary gravels of the Sierra Nevada of California: US Geol Surv Prof Paper 73, 226p. Litchfield, N.J., Craw, D., Koons, P.O., Edge, B., Perraudin, E., Peake, B.M., 2002. Geology and geochemistry of groundwater within the Taieri Basin, east Otago, New Zealand. NZ J. Geol. Geophys. 45, 481–497. Little, T.A., Prior, D.J., Toy, V.G., Lindroos, Z.R., 2015. Relationship of quartz LPO in mylonites near the Alpine fault, New Zealand to quartz content and recrystallization by grain-boundary migration. J. Struct. Geol. 81, 59–77. Lowey, G.W., 2002. White Channel Gravel alteration revisited. In: D.S. Emond, L.H. Weston and L.L. Lewis (Eds.), Yukon Exploration and Geology 2001, Exploration and Geological Services Division, Yukon Region, Indian and Northern Affairs Canada, p. 147–162. Lowey, G.W., 2006. The origin and evolution of the Klondike goldfields, Yukon, Canada. Ore Geol. Rev. 28, 431–450. Marsh, E.E., Craw, D., Goldfarb, R.J., Alpers, C.N., Bowell, R.J., Lowers, H.A., 2014. Eocene paleoplacer gold, Sierra Nevada foothills, California: Characteristics and potential source(s). GSA Annual Meeting, Vancouver, British Columbia, Paper 96–11.

29

Melashvili, M., Fleming, C., Dymov, I., Matthews, D., Dreisinger, D., 2015. Equation for thiosulphate yield during pyrite oxidation. Miner. Eng. 74, 105–111. Minter, W.E.L., 1999. Irrefutable detrital origin of Witwatersrand gold and evidence of eolian signatures. Econ. Geol. 94, 665–670. Minter, W.E.L., Goedhart, M., Knight, J., Frimmel, H.E., 1993. Morphology of Witwatersrand gold grains from the basal reef: evidence for their detrital origin. Econ. Geol. 88, 237–248. Mortensen, J.K., Craw, D., MacKenzie, D.J., Gabites, J.E., Ullrich, T., 2010. Age and origin of orogenic gold mineralisation in the Otago Schist belt, South Island, New Zealand: Constraints from lead isotope and 40Ar/39Ar dating studies. Econ. Geol. 105, 777–793. Mossman, D.J., Harron, G.A., 1983. Origin and distribution of gold in the Huronian Supergroup, Canada: the case for Witwatersrand-type paleoplacers. Precambrian Res. 20, 543–583. Mulch, A., Graham, S.A., Chamberlain, C.P., 2006. Hydrogen isotopes in Eocene river gravels and paleoelevation of the Sierra Nevada. Science 313, 87–89. Patyk-Kara, N.G., 1999. Cenozoic placer deposits and fluvial channel systems on the Arctic shelf of Siberia. Econ. Geol. 94, 707–720. Phillips, G.N., 1988. Widespread fluid infiltration during metamorphism of the Witwatersrand goldfields: generation of chloritoid and pyrophyllite. J. Metamorph. Geol. 6, 311–332. Reimer, T.O., Mossman, D.J., 1990. Sulfidization of Witwatersrand black sands: from enigma to myth. Geology 18, 426–429. Rimstidt, J.D., Vaughan, D.J., 2003. Pyrite oxidation: a state-of-the-art assessment of the reaction mechanism. Geochim. Cosmochim. Acta 67, 873–880. Stoffregen, R., 1986. Observations on the behavior of gold during supergene oxidation at Summitville, Colorado, USA, and implications for electrum stability in the weathering environment. Appl. Geochem. 1, 549–558. Tostevin, R., Craw, D., van Hale, R., Vaughan, M., 2016. Sources of environmental sulfur in the groundwater system, southern New Zealand. Appl. Geochem. 70, 1–16. Trumbull, R.B., Morteani, G., Li, Z., Bi, H., 1992. Gold Metallogeny in the Sino-Korean Platform. Springer-Verlag, Berlin, p. 202. Webster, J.G., 1986. The solubility of Au and Ag in the system Au–Ag–S–O2–H2O at 25 °C and 1 atm. Geochim. Cosmochim. Acta 50, 245–255. Williams, G.J., 1974. Economic Geology of New Zealand, Austral Inst Min Metall Monograph vol. 4, 490p. Yesares, L., Saez, R., Nieto, J.M., Almodovar, G.R., Cooper, S., 2014. Supergene enrichment of precious metals by natural amalgamation in the Las Cruces weathering profile (Iberian Pyrite Belt, SW Spain). Ore Geol. Rev. 58, 14–26. Youngson, J.H., 1995. Sulphur mobility and sulphur-mineral precipitation during early Miocene-Recent uplift and sedimentation in Central Otago, New Zealand. NZ J. Geol. Geophys. 38, 407–417. Youngson, J.H., Craw, D., 1993. Gold nugget growth during tectonically induced sedimentary recycling, Otago, New Zealand. Sed. Geol. 84, 71–88.