40Ar/39Ar and (U–Th)/He – 4He/3He geochronology of landscape evolution and channel iron deposit...

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 117 (2013) 283–312

40Ar/39Ar and (U–Th)/He – 4He/3He geochronologyof landscape evolution and channel iron deposit genesis

at Lynn Peak, Western Australia

Paulo M. Vasconcelos a,⇑, Jonathan A. Heim a,1, Kenneth A. Farley b,Hevelyn Monteiro a, Kathryn Waltenberg a

a School of Earth Sciences, The University of Queensland, Brisbane, Queensland 4072, Australiab Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA

Received 10 March 2011; accepted in revised form 28 March 2013; available online 9 April 2013

Abstract

(U–Th)/He geochronology of authigenic goethite cements from the Lynn Peak channel iron deposit (CID), HamersleyProvince, Western Australia, reveals a history of mineral precipitation ranging from ca. 33 to 14 Ma. Massive goethites fromnearby weathering profiles at Roy Hill North, a possible source of detrital material during the aggradation of the Lynn Peakchannels, yield (U–Th)/He results as old as ca. 64 Ma. The combination of (U–Th)/He geochronology with incremental out-gassing 4He/3He studies on proton-irradiated samples reveals that Lynn Peak goethites host radiogenic 4He in low retentivity(LRD) and high retentivity (HRD) domains and that the HRDs account for most of the sample mass and have lost very littleof their original 4He over geologic time. Such high retentivity is especially notable given the goethites were collected from thesurface, where they were subject to significant heating by solar irradiation. Minor contamination by detrital fragments ofpotentially 4He-rich primary phases (e.g., rutile, ilmenite, zircon) occurs in some samples. Fortunately, the 4He/3He methodpermits characterization of this extraneous 4He component, which is small (<10 wt.% of the total 4He in the goethite) and canbe corrected out in estimating the goethite formation age. These results indicate that the Lynn Peak channel was alreadyaggraded and undergoing goethite cementation by ca. 33 Ma.

The history of aggradation and channel cementation independently measured through 40Ar/39Ar geochronology is consis-tent with that obtained from the (U–Th)/He and 4He/3He record. Laser incremental-heating 40Ar/39Ar geochronology ofdetrital and authigenic Mn oxides, primarily cryptomelane (KMn8O16�xH2O), from the same locality in the Lynn Peak chan-nel reveals that detrital oxides are older than ca. 44 Ma (and as old as ca. 65 Ma) and authigenic oxides are younger than ca.35 Ma and as young as ca. 16 Ma. Authigenic cryptomelane precipitation and channel cementation occurred throughout theMiocene, with a particularly strong period at around 20 Ma. The 40Ar/39Ar geochronological results suggest that regionalweathering profiles, developed before 44 Ma and possibly as early as ca. 65 Ma, were incised and partially eroded in the44–35 Ma interval. Semi-arid conditions promoted the transport of large volumes of sediments, aggrading the regional drain-age system. Metasomatic reactions in the aggraded channels caused the ferruginization and, in some places, manganesereplacement of the CIDs; goethite and K–Mn oxide cementation continued throughout the Oligocene and Miocene. Post-Miocene aridification contributed to the preservation of the cemented channel sediments, forming some of the largest readilymineable iron ore deposits on earth.� 2013 Elsevier Ltd. All rights reserved.

0016-7037/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.gca.2013.03.037

⇑ Corresponding author. Tel.: +61 7 3365 3001 (HOD Office), +61 7 3365 2297 (Office), +61 7 3365 7454 (Lab); fax: +61 7 3365 1277.E-mail address: [email protected] (P.M. Vasconcelos).

1 Present address: Rio Tinto Exploration – Project Generation Group, 1 Research Avenue, Melbourne, Victoria 3083, Australia.


mailto:[email protected]


284 P.M. Vasconcelos et al. / Geochimica et Cosmochimica Acta 117 (2013) 283–312

1. INTRODUCTION

Channel iron deposits (CIDs) in the Hamersley Prov-ince, northwestern Australia, are ferruginized fluvial paleo-channel sediments that contain pisolitic or pisoidalgoethite–hematite iron mineralization (58–60% Fe,�0.05% P, 1–2% Al2O3, 3–5% SiO2, �10% loss on ignition)(Ramanaidou et al., 2003). This style of iron deposit is con-sidered unique to Western Australia; the closest knownanalogues are oolitic ironstone deposits in Kazakhstan(Morris, 1988). Measured reserves and indicated resourcesof CIDs in Australia total >10 billion tonnes (Dalstraet al., 2009), and they currently account for almost 30%of total iron production in the country. In addition to theirgreat economic importance, CIDs preserve the most com-prehensive and possibly continuous record of Cenozoicweathering, erosion, and sedimentation in northwesternAustralia. Reconstructing the formation of CIDs providesunique information on the tectonic and climatic historiesof this region, information not easily obtainable by othermethods. It also provides insight into the long-term land-scape evolution of continental landmasses, where stronglyoxidized and undated sedimentary deposits potentially pre-serve a rich geological history.

The genesis of CIDs, primarily focused on the econom-ically important Robe River and Yandicoogina (Yandi)deposits, has been investigated by Morris et al. (1993)(and references therein), Ramanaidou et al. (2003), Stoneet al. (2002), and Morris and Ramanaidou (2007), but thenature of the original sediments, the timing and mecha-nisms of channel aggradation, and the timing and processescontrolling ferruginization (goethite cementation) of chan-nel sediments remain unresolved (Morris et al., 2007;Vasconcelos et al., 2007).

Pollens and spores preserved in basal claystones of theYandi CID tentatively date the initiation of channelaggradation at the Eocene–Oligocene transition (MacP-hail and Stone, 2004). (U–Th)/He dating of authigenicgoethite cementing detrital iron oxides and rock frag-ments at Yandi reveals ages ranging from �18 Ma nearthe surface to 5 Ma towards the bottom of the channel(Heim et al., 2006), indicating that the channel had com-pletely aggraded and was undergoing goethite cementa-tion by at least the Early Miocene (Heim et al., 2006).Goethite (U–Th)/He ages decrease progressively withchannel depth, which Heim et al. (2006) interpret as evi-dence that goethite precipitated at the groundwater-atmo-sphere interface during deepening of the water table, aprocess driven by aridification of Western Australiathroughout the Neogene. Heim et al. (2006) also proposethat the oldest results in the upper parts of the channel(ca. 18–12 Ma) represent the products of partial recrystal-lization of early goethite cements during more humid cli-mates in the Miocene, and interpret channel aggradationto have occurred much earlier. By extrapolating the agevs. depth trends from the bottom towards the surfaceof the channel, Heim et al. (2006) estimate that aggrada-tion occurred at ca. 40–35 Ma and suggest that ferrugini-zation of the Yandi channel sediments may have startedimmediately after aggradation. They also suggest that

excursion towards more humid conditions in the Miocenepromoted the partial recrystallization of the upper partsof the channel, partially resetting the He ages of a signif-icant proportion of the near-surface goethite. Morris andRamanaidou (2007) and Morris et al. (2007) disagreewith this interpretation and instead suggest that the chan-nels hosting the CIDs aggraded relatively rapidly in theMiddle Miocene, after which the channels were cementedby goethite. Since the Middle Miocene, the deposits havebeen partially depleted in iron by late-stage groundwaterinteractions (Morris and Ramanaidou, 2007). To resolvethis discrepancy and derive a more complete history ofchannel aggradation and cementation for Western Aus-tralian CIDs, we sampled another ferruginized channelin the Pilbara region – the Lynn Peak channel – and con-ducted a detailed geochronological investigation (Fig. 1).

We carried out geochronological investigations onsupergene iron oxyhydroxides, building on and refining ap-proaches pioneered by Strutt (1908) and more recently re-vived and further developed by Lippolt et al. (1998),Pidgeon et al. (2004), and particularly Shuster et al.(2005). We also applied the approach of Vasconceloset al. (1992) in dating supergene Mn oxides. The work pre-sented here is as much about the genesis of CIDs as it isabout the systematics of applying geochronological toolsto study weathering and erosion in long-lived landscapes.

Ferruginized sediments and weathered crusts are com-mon surficial features in long-lived landscapes, particularlyin the southern hemisphere cratons. Applying geochronol-ogy to these surficial deposits is challenging, and weatheringgeochronology by the 40Ar/39Ar and (U–Th)/He methodsoffers a unique opportunity to date these landscapes and de-rive time-calibrated histories of weathering, erosion, sedi-mentation, re-weathering, providing insights into thetectonic and climatic histories of stable cratons. Weatheringgeochronology by the 40Ar/39Ar method, first applied inMinas Gerais (Vasconcelos et al., 1992) and Carajas, Brazil(Vasconcelos et al., 1994), is now widely used in such stud-ies (Ruffet et al., 1996; Henocque et al., 1998; Dammeret al., 1999; Van Niekerk et al., 1999; Vasconcelos,1999a,b; Hautmann and Lippolt, 2000; Feng and Vasconce-los, 2001, 2007; Li and Vasconcelos, 2002; Vasconcelos andConroy, 2003; Carmo and Vasconcelos, 2004, 2006; Colinet al., 2005; Arancibia et al., 2006; Spier et al., 2006; Liet al., 2007; Beauvais et al., 2008; Bissig and Riquelme,2010). The (U–Th)/He method has great potential to com-plement results obtained by 40Ar/39Ar geochronology(Shuster et al., 2005, 2012; Heim et al., 2006). It also ex-pands the application of weathering geochronology to min-erals devoid of K. When used in tandem, the two methodsmay provide a more complete weathering history than thatobtained by either method separately. However, the miner-alogical, paragenetic, and analytical difficulties in weather-ing geochronology cannot be underestimated, and the co-occurrence of detrital and authigenic Fe and Mn oxyhy-droxides at Lynn Peak offers a unique opportunity to testthe two methods at the same site. Comparing results fromtwo independent techniques applied to independent mineralsystems is essential in assessing the reliability of the historyof weathering obtained from the two methods.

Fig. 1. A digital elevation model (a) and topographic cross-section (b) illustrate sites subject to weathering geochronology investigation (a –Metawandy Valley; b – Mt Wall; c – Marandoo; d – Rhodes Ridge; e – Yandicoogina; f – Lynn Peak; g – Roy Hill North) by our group. Adetailed digital elevation model (c), geological map (map by J. Heim, modified from Department of Mines and Petroleum, Western Australia,1987. SF51-09 Balfour Downs WA Geological Map, 1:250 K scale) (d), and topographic cross-sections (e, f) illustrate location and relativeelevations of the (e) Roy Hill North and (f) Lynn Peak localities investigated in this study. The two sites are located at relatively low elevationsnorth of the Fortescue River plain and south of the Chichester Range, a possible source for the detrital material in the Lynn Peak CID. Thenearby occurrence of Marra Mamba BIF and the Balfour Formation (d) identify possible sources of Mn oxide blocks and Mn cements in theLynn Peak CID.

P.M. Vasconcelos et al. / Geochimica et Cosmochimica Acta 117 (2013) 283–312 285

1.1. Geology

The Lynn Peak CIDs (LP-CIDs) (22.50�S, 120.09�E,Fig. 1) are a sub-economic system of channel iron depositslocated north of the Fortescue River floodplain, in the farnortheast fringes of the Hamersley Province, approximately90–100 km ENE of the Yandi CID (Fig. 1a,b). The domi-nant lithologies in this area are subhorizontal and gentlyfolded and metamorphosed Archaean banded iron-forma-tions (BIF) and shales of the Hamersley Group and Ar-chaean metamorphosed clastic sediments, dolomites,jaspillites, and cherts of the Fortescue Group (Trendallet al., 2004 and references therein). Topography is domi-nated by rolling hills ranging from �450 to 470 m elevation(Figs. 1 and 2a). Unconsolidated and partially consolidatedcolluvial and alluvial deposits blanket the hills. Isolatedoccurrences of CIDs, including the Lynn Peak occurrencesampled in this study, crop out in the headwaters and trib-utaries of the ephemeral Kilkinbah Creek catchment(Fig. 1f). These CIDs constitute part of an ancient palaeo-

channel system that either drained northeasterly into theDavis River CIDs or southward into a now eroded paleo-channel that once drained into the Fortescue River system(Fig. 1). South of Lynn Peak, lateritic weathering profilesdeveloped on the Marra Mamba BIF form a chain of ridgesand mesas, 500–550 m in elevation, whose summits define arelatively flat surface (Figs. 1 and 2a). These weatheringprofiles may represent the remnants of the postulatedHamersley Surface (Campana et al., 1964), and they are alikely source for the sediments that aggraded the Lynn Peakchannel.

The LP-CIDs are predominantly composed of coarsesand to fine gravel-size fragments of ferruginized clay, goe-thite and hematite; iron-rich pisoliths; and ferruginized(hematite and goethite) wood fragments (�1–10 mm)(Figs. 2 and 3). This detrital material is cemented by authi-genic brown to black vitreous goethite (Figs. 2c and 3a).Continuous layers surrounding one or more detrital frag-ments indicate that goethite precipitated in situ as a chem-ical cement surrounding clastic grains (Figs. 2c and 3a).

(e)

2 cm

(c)

(f)

Auth Gt

Det Hem

Fe-Wood

(b)

Det Mn OxBlocks

Mn OxCemented CID

(d)

2 cm

Auth Cml

GtPis

Det Mn Ox

(a)

Fig. 2. (a) The landscape at Lynn Peak is characterized by �450–470 m elevation rolling hills (foreground) surrounded by 500–550 m deeplyweathered plateaus background. (b) The Lynn Peak channel contains strongly ferruginized and Mn cemented sediments and detrital blocks,which, when sectioned and polished (c, d), reveal detrital hematite, goethite pisoliths, ferruginized wood, or Mn oxide blocks completelycemented and partially replaced by authigenic goethite or cryptomelane. Since authigenic cryptomelane post-dates goethite cementation, theage of cryptomelane imposes a minimum age for channel aggradation and ferruginization. At Roy Hill North, large blocks of elongated andcurved fibrous goethite crystals (e, f) provide information on the possible ages of detrital material.


Two hand specimens of LP-CID were selected for (U–Th)/He geochronology (Table 1).

At Roy Hill North, we sampled massive blocks com-posed of extremely large (> 10 cm long) fibres of goethite(Figs. 2e, f and 3d). The blocks were deposited and cemen-ted into outwash (colluvium?) derived from erosion ofweathering profiles in the Chichester Ranges (Fig. 1e, f), adrainage divide between the Fortescue River plain andthe Lynn Peak drainage. These weathering profiles repre-sent possible source areas for the Lynn Peak alluvialsediments.

A particularly important geological feature at LynnPeak is the presence of subhorizontal outcrops of the

Hamersley Group Marra Mamba BIF, which are locallydeeply weathered and enriched in supergene iron and man-ganese deposits (Fig. 1d). In addition, outcrops of the Bal-four Formation (Fig. 1d), a manganiferous shale, ensuresthat sediments and groundwaters in this region are enrichedin manganese. Therefore, it is not surprising that we findlarge colluvial Mn oxide blocks loose at the surface orembedded and cemented into the Lynn Peak CIDs(Fig. 2b,d). In addition to detrital Mn oxides, parts of theLynn Peak CID are strongly cemented by authigenicMn oxide cements (Figs. 2b, d and 3e, f), probably theresult of Mn2+ influx into the paleodrainage system fromlocal groundwater seepage. Authigenic Mn oxide cements

(a)

(c)

Gt

Hem

Fe-Wood

Gt

Pis

Pis Pis

Gt

Gt

Gt

(d)

Hem

Gt

e)

Gt

Gt

Gt Gt

Gt

Gt

Gt

Cml

Cml

Gt Gt

Gt Gt

Cml

(e) 100 µm

Gt CmlCml

Pst

Gt (g) (h)

Gt

Cml

(f)

(b)

Gt1

Pis1

Pis2

Pis3Pis4

Gt2

Fig. 3. Reflected-light photomicrograph (a) illustrates detrital hematite (Hem), pisoliths (Pis), and ferruginized wood (Fe-Wood) cemented byauthigenic goethite (Gt) precipitated in the CID. The continuity of the goethite cement around several adjacent detrital grains confirms thatthe goethite cement sampled for geochronology precipitated in situ after channel aggradation. (b) Several generations of authigenic goethite(Gt1 and Gt2) form at the expense of partially dissolved detrital iron oxides (Pis1, 2, 3, and 4). (c) Visually pure Lynn Peak goethite grainsselected for geochronology often contain bands of authigenic hematite (Hem). Pure crystalline banded goethite fragments from the interior oflarge masses, as these grains from Roy Hill (d), were our preferred material for (U–Th)/He geochronology. Reflected-light photomicrograph(e) and SEM backscattered image (f) illustrate cryptomelane (Cml) cements filling cavities and partially replacing pisoliths, detrital goethitegrains, and authigenic goethite cements. Some of the Mn oxide grains recovered from the channels are relatively pure (h), while other grains(g) contain several generations of authigenic minerals (pyrolusite – Pst; goethite – Gt; and more than one generation of cryptomelane – Cml),explaining the complex 40Ar/39Ar incremental-heating spectra obtained for these samples.


Table 1Sample location and brief description.

Sample No. Eastinga Northinga Grain N� Mineral Depth Host rock Paragenesis

LynP-02-08-C 201,584 7,509,294 a, b, c Goe Surface Colluvium Colluvial float containin iron-rich detrital material and pisolites cemented bygoethite or CID

LynP-02-09-A1 201,146 7,508,708 a, b, c Goe Surface CID Banded vein of sub-vitr us – vitreous goethite crosscutting a block of channeliron deposit




LynP-02-09-A5 201,146 7,508,708 a, b, c Goe Surface CID Goethite replacing iron ich detrital material and cementing pore spaceRoy-02-02-B – – a, b Goe Surface Colluvium Massive goethite blocks ith unusually long crystals at the surfaceRoy-02-02-C – – a, b Goe Surface Colluvium Massive goethite blocks ith unusually long crystals at the surfaceLynP-02-02-B 201,303 7,508,750 2897 Cml Surface CID Mn oxides replacing iro detrital material and early goethite cements of CID;

late-stage pyrolusite and cryptomelane partially replacing early Mn oxides andfilling void spaces

LynP-02-02-A4 201,303 7,508,750 2900 Cml Surface CID Mn oxides replacing iro rich detrital materials, pisolites, and goethite cementsof CID

LynP-02-07-A1 201,303 7,508,750 2915/4487 Cml Surface CID Mn oxides replacing iro rich detrital materials, pisolites, and goethite cementsof CID

LynP-02-07-B 201,303 7,508,750 4489 Cml Surface CID Mn oxides replacing iro rich detrital materials, pisolites, and goethite cementsof CID

LynP-02-05-3 201,146 7,508,708 2912 Cml Surface CID Mn oxides replacing iro rich detrital materials, pisolites, and goethite cementsof CID

LynP-02-06-m 201,146 7,508,708 2914 Cml Surface CID Mn oxides replacing iro rich detrital materials, pisolites, and goethite cementsof CID

LynP-05-01 201,146 7,508,708 3894/4488 Cml Surface CID Mn oxides replacing iro rich detrital materials, pisolites, and goethite cementsof CID

LynP-02-03-B 201,751 7,509,582 2911 Cml Surface Colluvium Late-stage massive Mn xides filling pore space and replacing goethite ofcolluvium

LynP-02-01-A 201,303 7,508,750 2896 Cml Surface CID Mn-rich pisolites and M oxides replacing pisolites and goethite cements ofCID; banded goethite fi ing pore space

LynP-02-08-A 201,584 7,509,294 3884 Cml Surface Colluvium Colluvial float containin Mn oxides replacing iron-rich detrital material,goethite cements, and w od fragments

a Coordinates in AGD 84.

288P

.M.

Vasco

ncelo

set

al./G

eoch

imica

etC

osm

och

imica

Acta

117(2013)

283–312

g

eo

eo

eo

eo

-rww

n

n-

n-

n-

n-

n-

n-

o

nllgo


– mostly cryptomelane – partially replace detrital Fe andMn oxides, wood fragments, and authigenic goethite(Figs. 2d and 3e,f). Since authigenic Mn oxides replacedetrital and authigenic phases, including authigenic goe-thite cements, Mn oxide precipitation ages impose con-straints on the minimum age for channel aggradation andferruginization.

Other important geological features in the area includethe presence of calcretes, aeolian sand dunes, and a sub-stantial alluvium–colluvium cover that suggests that theLynn Peak area has been arid for a significant portion ofthe recent past (Fig. 1d). Geomorphologically, hills of dee-ply weathered Marra Mamba BIF south, west, and north-west of the area sampled in this study provide suitablesources for the detrital iron and manganese oxides presentin the Lynn Peak deposit and also distributed in the morerecent colluvial and alluvial deposits (Fig. 1).

2. METHODS

Samples were sectioned into �1–2 cm-thick slabs andpolished to enhance discrimination of the various genera-tions of goethite (Fig. 2c,d). Polished thin sections for thesamples were studied by a combination of transmitted-and reflected-light microscopy, scanning electron micros-copy (SEM) with energy dispersive X-ray analyses (EDX),and electron microprobe (EMP) analysis to determine themineralogy and paragenesis of the datable phases and thenumber of generations of goethite present in each sample.

Suitable phases for dating were recovered using a high-precision diamond wafer saw or a 5 mm ID diamond drillbit. All sub-samples were crushed and ultrasonicated inwater and ethanol. Five to ten grains of goethite from eachsample were mounted in epoxy blocks, polished, and inves-tigated by SEM (JEOL 6460 LA) and EMP (JEOL JX-8200) analyses at the University of Queensland Centre forMicroscopy and Microanalysis (UQ-CMM). Additionalgrains from each sample were selected for geochronology.

Synchrotron X-ray diffractometry confirmed the miner-alogy and crystallinity of selected goethite grains. Severalgrains of goethite, weighing between 50 and 100 mg, werepowdered in ethanol with �10 wt.% NIST 640c silicon stan-dard. The powdered goethite-silicon mixtures from eachsample were loaded into low background borosilicate Mul-ler glass capillaries (0.3 mm-diameter and 0.01 mm wallthickness) for synchroton X-ray investigation.

Several loose Mn oxide blocks were collected at the sur-face of the Lynn Peak channel (Fig. 2b). Blocks of Mn-ce-mented channel sediments were also retrieved (Fig. 2d).Hand specimens visually enriched in Mn oxides were sec-tioned into �1–2 cm-thick slabs and polished to exposethe complex mineralogical assemblage and to enable mi-cro-sampling pure concentrations of Mn oxide cements(Figs. 2d and 3e–h). Manganese oxides from detrital blocksor precipitated as authigenic cements surrounding goethite,hematite, and ferruginized rock fragments (Figs. 2d and 3e–h) were recovered with a 5 mm ID diamond drill bit. Drillcores were crushed to <2 mm, ultrasonicated in distilledwater for 30 min to remove adsorbed clay-size particles,ultrasonicated again in absolute ethanol, and dried under

heat lamps. Manganese oxide grains (Fig. 3g, h) from eachsample were mounted in epoxy blocks, polished, and inves-tigated by SEM (Joel 6460 LA) and EMP (JOEL JX-8200),while representative aliquots from each sample were se-lected for geochronology.

2.1. Synchrotron X-ray analyses

Synchrotron X-ray analyses were performed at the Aus-tralian National Beamline Facility (ANBF) at the PhotonFactory, National Laboratory for High Energy Physics(KEK), Tsukuba, Japan. The ANBF diffractometer or“BIGDIFF” is a large (0.573 m radius) Debye–Scherrercamera, equipped with a cylindrical cassette which holdsimaging plates covering 320� in 2h, and an eight-position cap-illary sample stage located at the centre of the camera (Gar-rett et al., 1995; Sabine et al., 1995; Cookson et al., 1998;Creagh et al., 1998). The whole diffractometer is enclosedand evacuated to less than 0.01 Torr to minimize air scatter-ing. Each pattern was acquired using a 0.7 A monochromat-ed beam in an exposure P10 min under vacuum. Exposuretime varied depending on bean intensity, which decayedexponentially over 12-h cycles. Data was collected on Fujiimage plates simultaneously at �45� to 90� 2h. The imageplates were scanned in a Fuji BAS200 scanner. A digital im-age of the exposure was converted to a logarithmic intensityscale from calibrated pixel brightness values. The 2h scalewas calibrated using radioactive fiducial marks and datawere extracted to a linear scale by integrating the area acrossan 8 mm-wide strip centred on the X-ray beam axis.

The synchrotron X-ray data were analysed by the Riet-veld refinement method (Rietveld, 1969) using Topas aca-demic V4 (Coelho, 2006). The background was refinedusing a 1/x function and a Chebyshev polynomial linearinterpolation function. A pseudo-voight function wasincorporated into the refinement to account for a broad,diffuse peak that resulted from diffractions from the capil-lary; the contribution of this peak to the diffracted signalwas identified and defined by analysing an empty capillarytube during one of our experiments. A lorentzian convolu-tion applied to the peak at full width half maximum(FWHM) was used to describe crystallite size. Cell param-eters, peak displacement, full width at half maximum, crys-tallite size, scale factor, linear absorption coefficient of thepacked sample capillary and percent amorphous contentwere all refined. The isotropic thermal parameters werefixed at the values measured for goethite single-crystalsdetermined by Gualtieri and Venturelli (1999).

2.2. (U–Th)/He analyses

(U–Th)/He geochronology was carried out at Caltech.Five to six goethite fragments from each sub-sample wereinserted individually into 1 mm ID � 2 mm long Pt cap-sules and heated with a Nd-YAG laser; the released gaseswere cryofocused and cryopurified, and then analysed usingisotope dilution quadrupole mass spectrometry followingprocedures detailed in Farley (2002) and Shuster et al.(2005). After complete degassing, the Pt capsules weretransferred to Teflon beakers containing 230Th and 235U


spikes, and goethite was dissolved in concentrated HCl for12 h at 90 �C. This solution was dried down and the result-ing salts redissolved in 5% HNO3. These spiked solutionswere analysed for 238U, 232Th, and 147Sm on a Thermo Ele-ment I double-focusing ICP-MS. (U–Th)/He ages werecomputed from the measured He, U, and Th assuming sec-ular equilibrium among all intermediate daughters. We de-tected no significant amount of 147Sm in these (nor in anyother) goethites. Grains selected for dating were recoveredfrom the interior of large goethite masses, obviating theneed for a-ejection correction. Most (U–Th)/He ages are re-ported with ±10% uncertainty, which accounts for analyti-cal uncertainties in U, Th, Sm, and He determinations, andthe uncertainty associated with possible 4He diffusive lossand/or 4He contribution from possible contaminants (seebelow).

Shuster et al. (2005) show that laser heating extractionof Pt-encapsulated goethite grains at temperatures of�1150 �C is suitable for complete He outgassing whilemaintaining quantitative retention of U and Th for (U–Th)/He dating. Heim et al. (2006) successfully used thesesame experimental conditions. However, our early attemptsto date Lynn Peak goethites yielded scattered (U–Th)/Hedates (36.9 ± 3.7, 31.7 ± 3.2, and 24.0 ± 2.4 Ma) for grainsfrom a single generation. These unexpected results raisedthe possibility that Lynn Peak goethite grains underwentU loss by volatilization during laser heating. To test thishypothesis, we extracted He from equivalent aliquots ofPt-encapsulated goethite grains (sample LynP-02-09-A2)at laser heating temperatures of 925, 945, 960, 1000, 1040,1070 and 1160 �C. One to three experiments were per-formed at each temperature. Complete He extraction wasconfirmed by fusing, at �1500 �C, one grain previouslyheated for He extraction at 925 �C; He isotopes releasedat this temperature were below background levels.

Since our petrographic studies revealed the possiblepresence of detrital contaminants (e.g., rutile and zircon)in the goethite grains used in geochronology, we quantifiedthe 4He budget from these contaminants. Multiple grains ofgoethite from the same batch used for 4He/3He experimentswere weighed and dissolved in ultra-pure 10% HCl in Tef-lon beakers. Solutions were heated to approximately100 �C for several hours. Goethite rapidly dissolved in thesolution, leaving a residue of clear and white crystals. Theresidue was washed in milli-Q water and dried down. Theresidual powders were wrapped in aluminium foil capsulesand outgassed in a resistance furnace. Hot blanks and stan-dards were measured before and after sample analysis. Theexperiment was repeated by dissolving additional goethitegrains in 10% HCl and 10% HF at 100 �C, and then adding10% HNO3. After digestion, the walls of the vials wererinsed with milli-Q water, which was transferred into alu-minium capsules, dried down and the residue analysed byresistance furnace.

A major question regarding the suitability of supergenegoethite for geochronology is whether samples collected atthe earth’s surface, and possibly subject to solar heatingand brush fires during their geological histories, retainradiogenic 4He quantitatively. To address this question,we assessed the He retention of our goethite samples

through 4He/3He experiments (Shuster et al., 2004, 2005;Heim et al., 2006). This method creates a uniform distribu-tion of 3He via spallation reactions, and 4He/3He ratiosmeasured during stepped outgassing are used to character-ize the 4He distribution in the sample. Selected goethitegrains were bombarded with 220 MeV protons at theNortheast Proton Therapy Centre, and the irradiatedgrains were incrementally heated using a low temperatureextraction cell described by Farley et al. (1999). The re-leased gas fractions were analysed for 3He and 4He by sec-tor field mass spectrometry following the procedures ofShuster et al. (2004, 2005). After low temperature Heextraction (25–400 �C), the remaining 4He and 3He contentswere determined by transferring the grains to a laser lineand heating them to ca. 1500 �C.

2.3. 40Ar/39Ar analyses

Three to 10 Mn oxide grains (1–2 mm) from each samplewere loaded into Al-disks together with Fish Canyon sani-dine (28.201 ± 0.046 Ma, Kuiper et al., 2008) neutron fluxmonitors, using the geometry illustrated in Vasconceloset al. (2002), and irradiated at the TRIGA reactor at theOregon State University B-1 CLICIT Facility for 14 h.Two to three grains from each sample were analysed usingthe incremental-heating laser 40Ar/39Ar method at the UQ-AGES (University of Queensland – Argon Geochronologyin Earth Sciences) facility, following procedures detailed inVasconcelos (1999a,b) and Vasconcelos et al. (2002).

Before each grain of Mn oxide, a series of three full-sys-tem blanks and one air pipette were analysed. Raw analyt-ical results were collected and interpreted using MassSpecv. 7.527 (Deino, 2009). Isotope ratios were obtained bypropagating to time zero the concentration for each isotope(40Ar, 39Ar, 38Ar, 37Ar, and 36Ar) as measured by eightpeak-hoping cycles (isotope refitting). Full-system blankswere refit using parabolic, parabolic, linear, linear, and lin-ear regressions for 40Ar, 39Ar, 38Ar, 37Ar, and 36Ar, respec-tively. 40Ar, 38Ar, and 36Ar from the air pipettes were refitthrough parabolic regressions, and isotope discriminationwas calculated for the system based on the average40Ar/36Ar measured for each group of five air pipettes,and assuming an 40Ar/36Ar atmospheric ratio of298.56 ± 0.31 (Lee et al., 2006). Individual gas fractionsfor the Mn oxide grains were refit through the same regres-sion procedure used for the blanks, and the values mea-sured were corrected for full-system blanks interpolatedfrom measurements made immediately before and aftereach grain. All results are reported with 2r errors, andage calculations include errors in J determination. J factorswere calculated for each irradiation disk through the anal-ysis of 15 individual crystals of Fish Canyon sanidine flu-ence monitor (age of 28.201 ± 0.046 Ma, Kuiper et al.,2008), and J values and errors are reported in EA 4. Plateauages are defined as two contiguous steps containing morethan 50% of the total 39Ar released by the grain, and whoseerrors overlap within 2r as defined and justified inVasconcelos (1999a). Plateau age errors are calculated asthe standard error of the weighted mean, but if the MSWDis greater than one, the errors are calculated as SEM � sqrt


(MSWD). The errors are 2r and include the error in J. Inte-grated ages are calculated by the isotopic recombination ofall steps for each grain. When a sample did not reach a pla-teau as defined above, but produced a relatively flat seg-ment in the age spectrum that contained a significantamount (>30–40%) of the nucleogenic gas fraction, we de-fined the flat segment as a “pseudo-plateau” and calculatedan age (‘forced-plateau age’, identified by ‘ ’ in the dia-grams) for the segment. This approach is useful when com-paring replicate grains. The probability–density plot(ideogram) contains all steps analysed and reported in EA4, independently of the quality of the results, as explainedand justified in Vasconcelos and Conroy (2003).

3. RESULTS

3.1. Goethite petrographic relations, mineralogy, and mineral

chemistry

Petrographic studies reveal that goethite in CIDs occurs aspisoliths, ferruginized clay and wood, and as authigenic cementbinding and partially replacing detrital phases (Figs. 2c and 3a–c). We specifically targeted the authigenic goethite cement forgeochronology. Texturally continuous goethite bands envelop-ing detrital grains (Figs. 2c and 3a) suggest that authigenic goe-thite precipitated in empty cavities or during partial dissolutionor replacement of detrital phases. Etched surfaces at the inter-face between authigenic goethite growth bands and detritalgrains (Fig. 3a, b) reveal that the solution from which goethiteprecipitated was corrosive to the detrital phases. Precipitationof late-stage goethite at the expense of previously precipitatedgoethite (Fig. 3b) also indicates several periods of influx of solu-tions out of equilibrium with the mineralogical assemblages inthe channels.

Goethite samples from Roy Hill North are unusuallylarge and massive (Figs. 2e, f and 3d), with elongated curvedfibrous crystals exceeding 10 cm in length. These goethitesare extremely pure (Fig. 3d) and appear to have precipitatedin empty cavities formed by weathering of relatively solubleFe-minerals (siderite?, pyrite?, etc.). Partial recrystallizationis also detectable as thin cross-cutting veinlets.

Goethites from Lynn Peak are not as pure as those fromRoy Hill North. Initial petrographic observations indicatedthat, in addition to goethite, the grains selected for geochro-nology also contained minor hematite (as revealed by thestoichiometry of the EMPA results), Al-goethite, Ti-oxides,and a silica phase (opal?) (Figs. 4 and 5). These phases werenot detected by synchrotron X-ray diffractometry, suggest-ing that they are present in very low abundances. In addi-tion, after detecting some extraneous 4He component inour 4He/3He experiments (as described below), and suspect-ing that this 4He was derived from He-rich contaminants,we carried out detailed SEM and EMPA on grains repre-sentative of the four micro-drilled growth bands in sampleLynP-02-09. The SEM observations (Fig. 4) illustrate twoimportant features in the grains: the presence of hematiteintergrown with goethite in all the grains; and the presenceof occasional mineral contaminants (rutile, ilmenite, andzircon) in grains from three out of the four samples. Thegoethite textures are also very complex, revealing an intri-

cate pattern of goethite-replaced detrital grains surroundedby authigenic goethite cements. Electron microprobe tra-verses across these grains confirm the presence of hematiteand Ti-rich phases (Figs. 4 and 5, EA 1).

3.2. Goethite crystallinity and purity

Synchrotron XRD analyses reveal that Lynn Peak goe-thite samples selected for apparent visual purity are, withinthe limits of the XRD detection, composed of pure goethite.The XRD traces show relatively narrow, sharp, and sym-metrical peaks indicating a high degree of crystallinity(EA 3). Rietveld refinement of the X-ray diffraction pat-terns reveals the unit-cell parameters in Table 2. Synchro-tron XRD, however, was not sufficiently sensitive toidentify the presence of minor contaminants (rutile, zircon).Since these phases may potentially affect the (U–Th)/He re-sults, the high-resolution electron microscopy and EMPAanalyses described above are necessary steps to ascertainwhether grains subjected to dating may host mineralinclusions.

3.3. Mn oxide petrographic relations, mineralogy, and

mineral chemistry

Field relations at Lynn Peak clearly reveal the presenceof two types of Mn oxides in the CIDs: detrital Mn oxide(mostly cryptomelane) blocks and grains, and authigenicMn oxide (cryptomelane and minor pyrolusite) cements(Figs. 2d and 3e–h). Detrital Mn oxides occur as large(>10 cm-diameter) massive blocks scattered at the CID sur-face or cemented into the CID. Detrital Mn oxide grains(<1 m) also occur embedded into the CID, where they arepartially replaced by authigenic Mn oxides or authigenicgoethite (Figs. 2d and 3e, f). Authigenic Mn oxide cementpartially or completely replaces detrital and early authi-genic Mn and Fe oxides and hydroxides, clearly postdatingchannel aggradation and subsequent ferruginisation of thechannel sediments (Figs. 2d and 3e, f).

Reflected-light and scanning electron microscopy con-firm that massive cryptomelane is the main phase in thedetrital fragments. Cryptomelane is also the main authi-genic cement, while minor pyrolusite also occurs. Early gen-erations of authigenic colloform cryptomelane, sometimesvery porous (Figs. 3e–h), replace detrital Mn and Fe oxihy-droxide and previously precipitated authigenic goethite ce-ments (Figs. 2d and 3e, f). Most samples contain more thanone generation of authigenic Mn oxide (Figs. 2d and 3e–g).Late-generation (Figs. 2d and 3e, f) colloform cryptomela-ne fills cavities formed by dissolution of goethite or early-generation Mn oxides. Detrital Mn oxides and distinct gen-erations of authigenic Mn oxides coexist in close spatialproximity (Fig. 2d). This suggests that a single 1–2 mmgrain may potentially host several generations of crypto-melane (Fig. 3g), posing difficulties for geochronology.Electron microprobe analyses (Fig. 6) reveal that grains se-lected for geochronology have K and Ba contents consis-tent with a hollandite-group composition (�0.5–5.2 wt.%K, 0–2.2 wt.% Ba) (full results provided in EA 2). The elec-tron microprobe results also show the intimate associationbetween Fe and Mn oxyhydroxides in the channel deposits.

Fig. 4. The mechanism by which some authigenic goethite precipitates in the LP-CID, i.e., the partial dissolution of detrital phases, the partialleaching of more soluble ions (e.g., Ca++, Mg++, Na+) and the local reprecipitation of Fe3+- and Al3+-oxides and SiO2 gives rise to impureand often mineralogically complex goethites. Despite great effort in selecting hand samples containing massive vitreous goethite that appearedpure and clean of contaminants under visual inspection (Fig. 2), high-resolution microscopy reveals a very different pattern. Goethite grainscontain partially replaced primary phases, hematite (Hem) bands – distinguished based on the higher reflectance in BSE imaging – andresistate phases, such as zircon (Zrn), ilmenite (Ilm), and rutile (Rt).


3.4. (U–Th)/He geochronology

Initial attempts to date Lynn Peak goethites revealedunexpected results. Scattered (U–Th)/He dates initially ob-tained for a single generation of goethite from sampleLynP-02-09 led us to suspect that either the sample hadundergone variable and erratic 4He loss during its geologi-cal history, or that the lack of age reproducibility might bedue to U and/or Th loss upon heating. The latter suggestionis strongly supported by the observation that lased samplesyielded higher Th/U ratios than unlased grains of the samematerial (for a very similar result in titanite, see Reiners andFarley, 1999). We evaluated possible He loss during thesample’s history by conducting 4He/3He diffusion experi-ments. We also tested for possible U and/or Th loss duringlaboratory heating by conducting a systematic study of theage and U,Th concentrations obtained after laser-heatingextractions at different temperatures.

3.4.1. Quantifying U loss during laser heating

Results of laser He extractions at different temperaturesare summarized in Fig. 7 and Table 3. Grains that under-went He extraction at 940 to 1040 �C yield reproducibleHe ages (32.3 ± 1.9 to 29.2 ± 1.8 Ma, Average = 30.9, St.Dev. = 1.0, n = 7) and a narrow range of Th/U ratios(3.3 ± 0.2 to 2.3 ± 0.1). A grain that underwent He extrac-tion at 1080 �C shows a 12% increase in the measured age(34.5 ± 1.7 Ma) and a noticeable increase in Th/U, now�4. Three replicate He extractions at 1120 �C show much

older and widely varying ages and higher Th/U values(Fig. 7). These results suggest that Lynn Peak goethites un-dergo U loss upon heating beyond ca. 1050 �C. The mech-anisms leading to U loss are unknown: U may be volatilizedand lost to the sample chamber; or U may react with the Ptcapsule and become insoluble during HCl dissolution of thegoethite grain for U and Th analysis. To obviate U lossesand unreliable ages, we performed all subsequent (U–Th)/He dating experiments at laser heating temperatures of945 ± 50 �C.

3.4.2. (U–Th)/He geochronology of authigenic goethites

After determining that Lynn Peak goethites were suscep-tible to U loss if heated above 1050 �C, we discarded all pre-vious results and concentrated our geochronological effortson two samples, LynP-02-08-C, a loose fragment of CID;and LynP-02-09, a large CID block with a distinct lami-nated “vein” of vitreous goethite cross-cutting detritalphases (Fig. 8, Table 4). Results for 15 grains extractedfrom five distinct areas in LynP-02-09 (Fig. 8) yield agesranging from 32.7 ± 3.3 to 14.5 ± 1.5 Ma. Three samples(LynP-02-09-A1, -A2, and -A3, Fig. 8, Table 4) were mi-cro-drilled from a visually pure continuous band of vitreousgoethite precipitated in the 3 cm wide (>10 cm long) “vein”

within the ferruginized CID sediments. Three grains fromeach of the three samples yield the exact same average(U-Th)/He results: 30.7 ± 3.1, 30.7 ± 3.1, and 30.7 ±3.1 Ma. Statistically reproducible average Th/U values(2.32 ± 0.05, 2.49 ± 0.29, and 2.43 ± 0.13) for the three

Fig. 5. (a–d) Electron microprobe traverses across the four grains illustrated in Figure 4 reveal the presence of hematite (analysis where the Fecontents plot above the black line). Element correlation diagrams (e–g) reveal Al contamination of goethite samples and the presence of Al, Si,and Ti phases intergrown with goethite. Contamination with Si, Al, and Ti is typical of goethite precipitated by the iron metassomatism andpartial replacement of detrital minerals in the channels. Full results are presented in AE1.


Table 2Unit cell parameters for goethite samples used in geochronology.

Sample Unite cell parameters Reliabilityfactors

Crystallite size Comments

a-Axis(A)

b-Axis(A)

c-Axis(A)

Unit cellvolume(A3)

Rp Rwp a-Direction(A)

b-Direction(A)

c-Direction(A)

General size(A)

CAPGoeB1 4.607 9.960 3.020 138.633 11.6 13.1 1407 63,241 1130Not resolved

Crystallite size using Scherrer equationextended to SPH

CAPGoeY1 4.602 9.959 3.021 138.514 8.90 9.36 2700 5666 49452904

Crystallite size using Scherrer equationextended to SPH

LynP-02-09-A1a 4.589 9.894 3.001 136.268 10.8 11.6 306 228 252290

Unit cell volume indicates that goethiteis Al-rich

LynP-02-09-A1b 4.586 9.887 3.000 136.014 11.5 12.4 352 204 183260


LynP-02-09-A1c 4.608 9.944 3.020 138.387 8.07 8.89 201 340 439277

A good fit and SPH resolves crystallitesize

LynP-02-09-A2a 4.600 9.928 3.013 137.613 11.1 11.9 330 385 364366


Roy-02-02-C3 (fit1) 4.615 9.963 3.023 138.985 15.3 15.7 Full structural fitCrystallite size could not be resolvedthrough the Scherrer equation or its SPHextension. Crystallites are too large

Roy-02-02-C3 (fit 2) 4.615 9.963 3.023 138.977 13.3 14.1 Le Bail fitCrystallite size could not be resolvedthrough the Scherrer equation.Crystallites are too large

294P

.M.

Vasco

ncelo

set

al./G

eoch

imica

etC

osm

och

imica

Acta

117(2013)

283–312

Fig. 6. Electron microprobe traverses across CID grains shows aninverse correlation between Fe and Mn (a), consistent with theparagenetic relationship observed where late-stage Mn oxidecements replace early goethite cements. The increase in K and Bawith increasing Mn contents (b) is consistent with petrographicevidence showing that cryptomelane is the major Mn oxide presentin our samples. The relatively high Si and Al contents identified inthe EMPA (c) suggest that these elements were not completelyflushed from the system during the weathering reactions thatprecipitated the Fe and Mn cements. Some of the Al occurs in solidsolution in goethite, but some microinclusions of clays, supergenequartz, or amorphous silica may also be present. The high Al and Sicontents are consistent with the interpretation that the Fe- and Mn-cements precipitated during the dissolution and partial replacement(Figs. 2–4) of detrital phases. Full results are presented in AE2.


samples are consistent with petrographic observations indi-cating that they represent a single generation of goethite(Table 4, Fig. 8). Another sample, LynP-02-09-A4, recov-ered from the innermost band in the “vein” (Fig. 8), yieldsa slightly younger average (U–Th)/He age of 28.0 ± 2.8 Maand distinct Th/U (2.02 ± 0.14), consistent with the accre-tionary precipitation of goethite from the wall towardsthe centre of the cavity. Three goethite grains analysed fromanother area (A5) in the sample – banded goethite cement-ing the wall of a large cavity possibly formed by roots ororganisms – yield significantly younger and less reproduc-ible results (20.5 ± 2.1, 14.5 ± 1.5, and 21.8 ± 2.2 Ma).Three goethite grains from the other CID hand sample(LynP-02-08-C, Table 4) yield (U–Th)/He ages(26.1 ± 2.6, 18.5 ± 1.9, and 18.9 ± 1.9 Ma) similar to theages obtained for A5 in sample LynP-02-09. The variabilityin ages for these two samples may actually reflect variousgoethite generations closely intergrown, as suggested bypetrographic analysis of individual grains (Fig. 3b).

The Roy Hill North goethites yield significantly older re-sults (Table 4). Most grains were analysed in duplicate, and10 of the 12 grains yield ages in the 66–60 Ma interval(average 63.1 ± 2.1 Ma), within error of the 40Ar/39Ar re-sults obtained for some of the detrital Mn oxide at LynnPeak. Two grains yield results (53.5 ± 5.35 and28.2 ± 2.82 Ma) not in the 66–60 Ma interval. We interpretthese younger results as recrystallised areas in the sample,identifiable by textures (microveins) observed during SEManalysis. These younger results also correspond to youngergenerations of goethite (this study) and Mn oxides mea-sured at Lynn Peak. We interpret the older results as indic-ative of the mineral precipitation age for this massivegoethite sample from Roy Hill North.

Shuster et al. (2005) and Heim et al. (2006) show,through incremental-heating 4He/3He experiments on pro-ton-irradiated samples, that supergene goethite quantita-tively retains �80–95% of ingrown 4He over geologicaltime. These results confirm that the (U–Th)/He method issuitable for dating supergene goethite, but they also suggestthe need for correcting goethite (U–Th)/He ages for modest4He diffusive losses that may have occurred during the sam-ple’s geological history. The samples studied by Shusteret al. (2005) and Heim et al. (2006) were collected fromthe subsurface, shielded from solar radiation, which mayaccount for their near-quantitative 4He retention. The sam-ples dated in the present study, on the other hand, are sur-face samples, and so may have undergone greater 4He loss.4He/3He experiments provide one way to detect and quan-tify these possible losses (Shuster et al., 2005).

3.5. 4He/3He incremental-heating analysis

To detect and quantify 4He loss and to determine thereliability of our (U–Th)/He ages, we performed incremen-tal-heating experiments on three proton-irradiated aliquotsfrom the same samples that we used for geochronology(LynP-02-09-A1, -A2, and -A3). The results are summa-rized in Table 5 and illustrated in Figs. 9 and 10. 4He/3Heresults can be interrogated in a variety of ways, and eachapproach provides slightly different information. There is

Fig. 7. Helium laser extraction temperatures are plotted against He age on the Y1 axis and Th/U ratios on the Y2 axis. The error on the Th/Uratio is �5% at the 2r confidence interval. Note that U, Th, and He concentrations were not determined for experiments performed at lasertemperatures of 1120 �C, only their relative abundances. Complete helium extraction of the grains was achieved at all temperatures. Uraniumand/or Th losses during heating, presumably by volatilization, will lead to erroneously older He ages and be recorded as a fractionation in theTh/U ratio, unless loss of these chemically distinct elements is in equal proportions (Farley, 2002).

Table 3(U–Th)/He agess and trace element contents for several aliquots of a goethite sample heated at progressively higher temperature.

Aliquot Temperature (�C) ±(�C) Age (Ma) ±(Ma) U (ppm) Th (ppm) He (nmol/g) Mass (lg) Th/U

05PXJ 920 20 31.83 3.18 42.90 128.15 12.66 5.73 2.9905PWN 945 20 30.23 3.02 38.32 86.93 9.67 6.53 2.2705PWO 945 20 31.13 3.11 36.34 102.19 10.23 9.04 2.8105PWP 945 20 30.73 3.07 35.72 85.20 9.33 10.12 2.3905PXE 960 20 29.19 2.92 41.89 107.37 10.67 3.80 2.5605PXF 1000 20 32.34 3.23 48.17 130.44 13.89 4.68 2.7105PXI 1040 20 30.55 3.05 48.73 162.57 14.47 5.72 3.3405PXG 1080 20 34.45 3.44 34.14 136.16 12.41 4.74 3.9905PPV 1120 40 60.77 6.08 * * * * 6.8005PPW 1120 40 66.22 6.62 * * * * 10.1105PPX 1120 40 56.18 5.62 * * * * 12.82

* Not measured.


also considerable redundancy in these presentations. Forthe sake of better understanding the 4He/3He method as ap-plied to supergene goethite, we will illustrate our results inthree different ways and explore the information providedby each approach.

3.5.1. 4He/3He and Rstep/Rbulk vs. RF 3He profiles and the

relative proportions of low and high retentivity domains

In the 4He/3He vs. RF 3He presentation, the 4He/3He ra-tio measured in each step of the incremental degassingexperiment is plotted as a function of the cumulative frac-tion of the total 3He in the sample released up to andincluding that step. In this space, plots for the three samples(Fig. 9a–c) are very similar. The samples yield high 4He/3Heratios in the initial steps; with increased degassing the4He/3He ratios drop sharply, and then ascend towards aplateau that constitutes most of the He in the sample. Inall three samples, the plateau is terminated in the last stepby a significant 4He/3He increase. The three samples reach4He/3He plateaus at RF 3He �0.22, 0.06, and 0.18, and with

values of �400, �240, and �300 for -A1, -A2, and -A3respectively. The distinct plateau ratios must reflect distinct3He concentrations in these samples since we will show be-low that they are coeval, have similar U and Th contents,and have undergone similar amounts of 4He diffusive lossduring their geological histories. Variable 3He concentra-tions arise from spatial variability in the proton fluenceand/or energy in the irradiation assembly. The sampleswere located several cm apart in the assembly, so this de-gree of variability is not surprising, nor does it imply vari-ations at the mm scale of each individual sample.

To facilitate comparison among samples that differ intheir concentrations of either natural 4He or synthetic3He, the measured 4He/3He ratios in each step (Rstep) canbe normalized to the 4He/3He ratio in the entire sample(Rbulk) and again plotted against RF 3He (Fig. 9d–f). Thisplot is almost identical to Fig. 9a–c, differing only by a con-stant factor on the y-axis. However the samples now plotalmost identically relative to each other. In either presenta-tion there are three distinct regions of the spectrum seen in

Fig. 8. Scanned image of CID block illustrating a laminated and visually pure goethite “vein” and other areas in the sample dated by the (U–Th)/He method. The results obtained for grains micro-drilled from a single band (A1, A2, and A3) are reproducible, while the results for aninner band in the “vein” (A4) is younger. A thin band of goethite coating a large empty cavity (A5), apparently associated with a root zone oran insect burrow, yields distinctly younger results.


all three samples: the region of high 4He/3He ratios at thestart of the experiment, the region including the rise up toand including the uniform 4He/3He ratio plateau, and thestep up to higher ratios at the end.

The excess of 4He with respect to 3He in the very lowtemperature steps in our samples differs markedly from pre-vious observations (Shuster et al., 2005; Heim et al., 2006).The low temperature steps reveal the presence of excess 4He(with respect to 3He) hosted in a very low retentivity reser-voir interrogated right at the beginning of the experiment.This reservoir could be the interstitial space between goe-thite crystallites; poorly crystallized goethite; goethite crys-tals damaged during sample preparation or protonirradiation; or a combination of two or more of these envi-ronments. While the 3He in this reservoir must be synthetic,the 4He may be either trapped atmospheric He or radio-genic He either diffused or ejected from the crystallites. Thisreservoir corresponds to the low retentivity domain de-scribed by Shuster et al. (2005), but the present samplescontain more 4He in this domain than samples studied pre-viously. Assuming this reservoir or domain is depletedwhen the 4He/3He ratios begin to rise, and also that its3He has been quantitatively retained since irradiation,Fig. 9 indicates that the low retentivity reservoir corre-sponds to a small fraction of the mass of each sample (lessthan 1% for samples LynP-02-9-A1 and -A2, and less than4% for sample -A3). This reservoir has even less of the 4Hein all three of the samples. Thus its existence has little effecton the ages obtained for the samples.

The second region corresponds to the high retentivitydomains described by Shuster et al., 2005. Its rise-fol-lowed-by-plateau is consistent with a domain that has expe-rienced partial diffusive He loss. The rise in 4He/3He ratiooccurs more rapidly for sample LynP-02-09-A2, suggestingthat it is more He retentive than the other two samples. Theplateau segment in all three samples indicates a region inthe grains with uniform 4He/3He and therefore complete

He retention. The plateau corresponds to �65%, �88%,and �65% of the 3He in samples A1, A2, and A3, respec-tively (Fig. 9).

The increase in 4He/3He in the last step for the threesamples has not previously been observed. A large propor-tion of both He isotopes released in the last steps mustcome from goethite because complete He extraction fromgoethite only occurs at temperatures exceeding ca. 500–900 �C and our incremental-heating experiments only reachca. 300 �C. The increase in 4He/3He in these steps demandthe existence of a reservoir that is both more He retentiveand richer in 4He than the reservoir that degasses to formthe plateau. This reservoir may represent goethite crystal-lites that are the same age as the bulk of the sample butare both more retentive and richer in 4He than the majorityof the sample. Alternatively, this reservoir could be a min-eral inclusion within goethite, such as the small fragmentsof zircon, rutile, and ilmenite detected by SEM and EMPanalysis (Figs. 4 and 5). To assess the role of mineral con-taminants in the He budget of the goethite, we quantifiedthe potentially “extraneous” 4He carried in zircon, rutile,and ilmenite. Since the contaminant grains are a very smallmass fraction of the sample and 3He generated by protonbombardment is proportional to mass (Shuster and Farley,2004), the 3He contribution from these phases is negligible.Assuming that all the “extraneous” 4He is derived fromcontaminants, we can estimate their 4He contribution byquantifying the amount of excess 4He with respect to 3Hefor the last step, assuming that the 4He/3He ratios shouldbe ca. 400, 240, and 300 – the plateau values (Fig. 9a–c) –for LynP-02-09-A1, -A2, and -A3, respectively. The “extra-neous 4He” is 41%, 20%, and 63% of the total 4He liberatedin the last step, and 9%, 2%, and 11% of the bulk 4He con-tents for samples LynP-02-09-A1, -A2, and -A3, respec-tively (Table 5). These percentages correspond to�0.5 nmol/g of excess 4He.

Table 4(U–Th)/He geochronology results for Lynn Peak (run at 950 �C) and Roy Hill (run at 1050 �C) goethite samples.

Sample number Age (Ma) ± U (ppm) Th (ppm) He (nmol/g) Mass (ug) Th/U

LYN-9-A1g:la 32.4 3.2 34.9 81.9 9.6 4.2 2.34LYN-9-A1h:la 29.2 2.9 41.4 93.4 10.1 5.0 2.26LYN-9-A1i:la 30.4 3.0 38.6 90.8 9.9 4.9 2.35Mean 30.7 38.3 88.7 9.9 2.32

LYN-9-A2f:la 30.2 3.0 38.3 86.9 9.7 6.5 2.27LYN-9-A2g:la 31.1 3.1 36.3 102.2 10.2 9.0 2.81LYN-9-A2h:la 30.7 3.1 35.7 85.2 9.3 10.1 2.39Mean 30.7 36.8 91.4 9.7 2.49




LYN-8-Cf:las 26.1 2.6 21.8 26.8 4.0 10.2 1.23LYN-8-Cg:las 18.5 1.8 23.4 33.8 3.2 8.4 1.45LYN-8-Ch:las 18.9 1.9 30.0 31.1 3.9 5.6 1.04Mean 21.2 25.1 30.6 3.7 1.24

Roy-02-02-B1(a) 65.5 6.6 0.14Roy-02-02-B1(b) 64.8 6.5 0.13

Roy-02-02-B2(a) 62.0 6.2 0.09

Roy-02-02-B4(a) 28.2* 2.8 0.13Roy-02-02-B4(b) 63.4 6.3 0.10

Roy-02-02-B5(a) 61.3 6.1 0.12Roy-02-02-B5(b) 62.4 6.2 0.23

Roy-02-02-B6(a) 62.1 6.2 0.12Roy-02-02-B6(b) 59.3 5.9 0.30

Roy-02-02-B7(b) 63.7 6.4 0.11

Roy-02-02-C1(a) 53.5* 5.3 0.15Roy-02-02-C1(b) 66.1 6.6 0.36

Roy-02-02-C4(a) 64.0 6.4 0.22Roy-02-02-C4(b) 72.1 7.2 0.22

Mean 63.9 6.4

* Not included in mean calculation.


The 4He content of possible contaminants (rutile, ilmen-ite, and/or zircon) was determined directly by analysing Hefrom HCl-insoluble residues obtained by dissolving aknown mass (�50 mg) of goethite from each of the threesamples. These experiments yielded between 0.005 and0.025 nmol of radiogenic helium per gram of starting goe-thite. While this is an easily measurable amount, it is a tinyfraction of the inferred excess. Thus if the apparent excessof 4He implied by the 4He/3He spectra arises from mineralcontaminants, either they are soluble in HCl or were notcompletely recovered by our experiments, or have ejectedthe majority of their He into the surrounding goethite.

In summary, the 4He/3He spectra (Fig. 9) demonstratethat the Lynn peak goethites host most (>60% and up to

90%) of the radiogenic 4He and proton-induced 3He in atight and He-retentive reservoir; that these reservoirs haveonly suffered partial helium losses; and that some minorremnant contaminants may add a component of extraneoushelium to the samples. These two later phenomena indicatethe need to correct the bulk He ages, in opposite directions.While Fig. 9 can be used to make the corrections, an alter-native presentation makes these corrections more readilyinterpretable.

3.5.2. 4He/3He age spectra

A 4He/3He spectrum can be converted into an age spec-trum (Shuster and Farley, 2004), which is especially usefulfor computing formation ages of samples that have

Table 54He/3He incremental-heating results.


Color code: Gray = LRD, black = HRD, bold black = last step.


experienced relatively modest He loss or gain. Assumingthat U and Th are uniformly distributed, a “step age” canbe calculated using the 4He/3He ratio of an incremental-heating step and the (measured) bulk (U–Th)/He age,according to the equation:

\Step Age" ¼ ðRstep=RbulkÞmeasured=ðRstep=RbulkÞa-ref

� bulk 4He age

The “a-ref” ratio accounts for the effects of a-ejection onthe distribution of natural radiogenic helium, an effect mostpronounced at mineral surfaces. Because we dated goethitegrains from the interior of large masses, and away fromsample surfaces, a-ejection is presumably balanced byimplantation and there is no requirement to incorporatethis scaling factor. As such, the step age reduces to:

\Step Age" ¼ ðRstep=RbulkÞmeasured � bulk4He age

The basic idea here is that, while bulk ages are sensitive toboth diffusive loss and the presence of extraneous He (andthus may need to be corrected for these phenomena), theincremental-heating procedure may isolate gas from a por-tion of the sample unaffected by these phenomena. In a4He/3He age spectrum, the model ‘step age’ is plottedagainst cumulative 3He much like in a 40Ar/39Ar age spec-trum. For samples that contain a series of steps that yieldrelatively flat segments at mid to high temperatures, a pla-teau age may be determined. By analogy to 40Ar/39Ar dat-ing, this plateau is most simply explained by degassing of asubstantial mass of sample unaffected by both loss and

extraneous gas. Plateau ages (Fig. 10) are calculated fromthe inflection point, where contiguous gas fractions repre-senting more than 50% of the 3He released from the sampleyield step-ages that are indistinguishable from the mean ageat the 95% confidence interval (2r). Plateau ages for thethree grains analysed in this study are 31.4 ± 1.8,30.0 ± 1.0, and 29.9 ± 0.9 Ma. Coincidentally these valuesare indistinguishable from the bulk ages (30.7 ± 3.1 Mafor all three samples), indicating that the loss and excesscomponents are nearly identical in size.

3.6. 40Ar/39Ar geochronology

40Ar/39Ar incremental-heating spectra for all grains, cor-rected for interfering isotopes, full-system blanks and massdiscrimination, are displayed in Figs. 11 and 12, and the fullanalytical results are provided in EA 4.

The incremental-heating spectra in Figs. 11 and 12 illus-trate that most grains do not yield simple spectra with well-defined plateau ages. The analysis of at least two or three rep-licate grains from each sample, often yielding very similarspectra, indicates that the complex results are indeed a min-eralogical/paragenetic feature of the samples and not pooranalysis. Complex paragenesis and mineral intergrowthsare confirmed by petrographic observations (Fig. 3).

Nevertheless, two major patterns emerge from the re-sults. Firstly, the incremental-heating spectra for grains ex-tracted from colluvial Mn oxide blocks yield resultsconsistently older than ca. 44 Ma (Fig. 11). For example,

Fig. 9. Incremental 4He/3He analysis for three distinct Lynn Peak goethite grains are illustrated as 4He/3He vs. RF 3He (a–c) and Rstep/Rbulk

vs. RF 3He diagrams (d–f).


sample LynP-02-03-B2 (Grains 2911-01, -02, and -03) yieldsthree very similar climbing spectra defining high-tempera-ture minimum ages of 65.1 ± 0.9, 65.3 ± 1.1 and64.7 ± 0.3 Ma (Fig. 11). This pattern in incremental-heat-ing spectra suggests the coexistence of multiple generationsof Mn oxides (Henocque et al., 1998; Vasconcelos, 1999a).The climbing spectra suggest the precipitation of Mn oxidesat ca. 65 Ma and partial recrystallization of this Mn oxideduring subsequent weathering events. Another detritalMn oxide sample, LynP-08-A, yields two compatible pla-teau (43.9 ± 1.0 Ma) and forced-plateau (44 ± 2 Ma)(Fig. 11) ages, imposing minimum constraints for the ageof the weathering profile from which the detrital blockswere derived. If the Lynn Peak channel was aggraded with

previously weathered material from nearby weathering pro-files, the channels were aggraded after 44 Ma.

Some incremental-heating spectra for grains micro-drilled from Mn oxides cementing CID blocks, such asthose illustrated in Figs. 2d and 3e, f, yield complex climb-ing patterns reaching ages as old as ca. 50 Ma (Fig. 11).Although the samples were recovered from blocks cemen-ted into the Lynn Peak CID, these blocks were particularlyenriched in detrital Mn oxide grains as illustrated inFig. 2d. Since it is impossible to determine whether an opa-que Mn oxide grain submitted for geochronology containsonly one or more generations of Mn oxides (Fig. 3g), weinterpret the results for these grains as indication of mixedgenerations of K–Mn oxides (detrital and authigenic) in the

Fig. 10. Incremental-heating 4He/3He apparent age (or Rstep/Rbulk) vs. cumulative fraction of 3He released spectra reveal that the three grainsanalysed for sample LynP-02-09 yield compatible plateau ages. The spectra also reveal the LRDs contain “excess He” in the initial steps,similarly to excess Ar detected in 40Ar/39Ar incremental-heating analysis. The rising steps identify reservoirs that underwent 4He loss bydiffusion during the geological history of the sample. The plateau steps identify the bulk of the homogeneous He-retentive goethite that hostsmore than 60% of the 3He and 4He in the grain. Finally, the last step represents the remainder of the goethite that did not degas at 300 �C; itmay also reveal the presence of an “extraneous” 4He reservoir, such as minor amounts of He-retentive contaminants in the samples.


1–2 mm grain dated. This is not uncommon (as illustratedin Figs. 2d and 3g) and in the present case unavoidable, de-spite our effort in micro-sampling each generationseparately.

Finally, 26 grains from samples composed entirely ofauthigenic Mn oxides cements yield ages that are youngerthan ca 35 Ma (Fig. 12). At first inspection, one might con-clude that the analytical results for the authigenic oxides areof poor quality, and that the complex spectra (Fig. 12) forsome grains reflect analytical problems (uneven sampleheating, improper gas cleaning, etc.). This is clearly notthe case, since some grains investigated under the exactsame analytical conditions do yield proper plateau ages(e.g., grains 2897-01, -02, and -03; 2915-03, Fig. 12). Basedin our experience at UQ-AGES, where we have analysed inexcess of 3000 petrographically well characterized Mn oxide

grains by 40Ar/39Ar incremental-heating spectra, the com-plexity in the spectra for Lynn Peak samples results fromthe mineralogical heterogeneity documented in the grains.As discussed above, a single 2-mm grain may contain sev-eral coexisting generations of Mn oxides (Fig. 3g) varyingin age, composition, and crystallinity. Coexisting genera-tions of Mn oxides, displaying distinct retentivities andAr liberation temperatures, will result in the release of gaseswith different amounts of atmospheric, radiogenic, andnucleogenic Ar isotopes at each incremental-heating step.Variable proportions of Mn oxide crystals from each gener-ation outgassing at particular heating steps will result inrandom and unpredictable mixtures of gases from reser-voirs of different ages, yielding complex spectra, such asthe ones observed in this study. Petrographic observationsconfirm the intimate coexistence of different generations

Fig. 11. 40Ar/39Ar incremental-heating spectra reveal that colluvial Mn oxide blocks, composed primarily of cryptomelane, are older than ca.44 Ma and were probably derived from ancient regional weathering profile as old as 65 Ma, as indicated by the minimum ages for three grains(2911-01, -02, and -03) analysed for sample LynP-02-03-B2. The ascending spectra for the three grains suggest that the detrital Mn oxidesunderwent some recrystallization after deposition in the channels. Incremental-heating spectra for partially recrystallized detrital Mn oxidesreveal complex patterns ascending towards the minimum ages of the detrital phases. Apparent age plateaus in ‘xxx’ denote forced-plateaus, asdefined in the text.


of Mn oxides (Fig. 3e–h). SEM/EDS and electron micro-probe analyses (Figs. 3e–h and 6) confirm variations incompositions, such as variable K and Ba contents, in thedifferent generations of Mn oxides, that should result in

variable Ar retentivities for the coexisting phases(Vasconcelos et al., 1994).

The spectra reveal additional evidence that thegrains contain more than one generation of Mn oxides.

Fig. 12. Representative 40Ar/39Ar incremental-heating spectra for authigenic Mn oxide cement reveal that detrital material in the Lynn PeakCID was undergoing partial cementation by Mn oxides at least as early as 35–32 Ma, suggesting that sedimentation precedes this time. Thecomplex incremental-heating spectra reveal a complex paragenesis, where several generations of authigenic minerals may coexist in a singlesample or 1–2 mm grain, as illustrated in Fig. 3g. (Apparent age plateaus in ‘ ’ denote forced-plateaus).


Fig 12. (continued)


For example, some grains (e.g., 4488-01, -02, and -03; 4489-01, -02, and -03, Fig. 12) yield reproducible hump-shapedspectra, defining low-age plateau-like segments at low T,climbing towards older results at middle T, and reachingadditional low-age plateau-like segments at high T. Thistype of spectrum is common in samples composed of mixedgenerations of Mn oxides with distinct Ar retentivities(Vasconcelos, 1999b). Some samples (LynP-02-02-A4,grains 2900-01, -02, and -03, Fig. 12) yield plateaus or pseu-do-plateau segments with distinct ages (31 ± 4, 16.6 ± 1.1,23 ± 5 Ma), revealing different generations of Mn oxidescoexisting in a single hand specimen. Despite the complex-

ity in the spectra, the reproducibility of results obtained forvarious grains from a same sample, for most samples, sug-gests that the 40Ar/39Ar laser incremental-heating data re-cord ages of mineral precipitation or, in the worst casescenario, mixed authigenic mineral ages.

In summary, the 40Ar/39Ar results reveal that CID sam-ples contain detrital blocks that still record the ages of theweathering profiles from which they were eroded; containpartially reset detrital Mn oxides; and also contain severalgenerations of authigenic Mn oxides. Detrital Mn oxidesare older than 44 Ma; authigenic Mn oxides are youngerthan ca. 35 Ma; and the main generation of authigenic


Mn oxides is ca. 20 Ma. The robustness of this interpreta-tion becomes more apparent when comparing 40Ar/39Arwith (U–Th)/He results (Fig. 13), as discussed below.

4. DISCUSSION

This study illustrates the systematic approach –combined optical and electron microscopy, electronmicroprobe and XRD analysis, and detailed incremen-tal-heating noble gas isotope analysis in natural andirradiated samples – necessary to extract informationfrom the weathered profiles that cover large parts ofthe Australian (and other cratonic) landscape. It also

Fig. 13. (a) A probability-density plot for all Lynn Peak Mn Oxide 40Ar/3

older than ca. 44 Ma, while authigenic oxides (black) are younger thanHistogram for (U–Th)/He ages of authigenic goethites from Lynn Peakauthigenic Mn oxides from the same CID. (U–Th)/He results for goethiweathering profiles at the source areas for Lynn Peak sediments, reveal aPeak.

provides information on the approaches necessary forthe successful application of weathering geochronologyto the study of complex and long-lived weathering pro-files, where detrital and authigenic generations of a dat-able mineral may coexist. It also contributes to therefinement of the (U–Th)/He and 4He/3He methods asapplied to weathering processes. Finally, this investiga-tion provides a robust geochronological dataset that per-mits determination of a minimum age for the Lynn Peakand possibly other CIDs in Western Australia, sheddingnew light on the mechanisms controlling erosion, channelaggradation, and goethite cementation during the forma-tion of Western Australian CIDs.

9Ar results (full results in EA 4) show that detrital oxides (gray) areca. 35 Ma, suggesting a possible age for channel aggradation. (b)(gray) show that the ages are compatible with those obtained for

te grains from Roy Hill (gray), interpreted to represent the ages ofges compatible with those obtained for detrital Mn oxides at Lynn


4.1. Petrography and mineral chemistry

Field and hand specimen observations, optical and SEMpetrography, and electron microprobe analyses reveal thatsupergene goethite has a complex history. CID samplescontain detrital and several generations of authigenicgoethites (Figs. 2, 3, and 8). Authigenic minerals also showcontrasting textures and purity. Some authigenic goethitesand cryptomelane appear to be very pure, devoid of con-taminants, and precipitated directly from solution intoempty cavities (Figs. 2 and 3). Other authigenic goethitesand cryptomelane contain mineral contaminants and ap-pear to form by Fe- or Mn-metasomatism and incompletereplacement of primary phases (Figs. 2, 3, 5, and 8). Thesedistinct precipitation mechanisms affect the final composi-tion and crystallinity of the grains used in geochronology;they also potentially affect the geochronological results.For example, goethites resulting from the incompletereplacement of primary phases are clearly identifiable bytheir high Th/U values (Table 4) and by the presence of res-istates (Fig. 4). These samples are prone to “extraneous4He” contents, which fortunately may be identified andquantified by 4He/3He analysis (Figs. 9 and 10). As longas the “extraneous 4He” is small (<10%), goethite precipi-tated by partial replacement of primary phases can be suc-cessfully dated by the (U–Th)/He method.

All goethite grains from Lynn Peak are relatively en-riched in Th (Table 4) compared with goethite from weath-ering profiles directly overlying banded iron-formations(Heim, 2007, p. 170) and with the goethites from Roy HillNorth. Petrographic investigation shows that goethites inCIDs form by the partial dissolution and replacement ofdetrital sediments, which includes fragments of bandediron-formations, but also detrital phases derived from otherlithologies (Figs. 2, 3, and 8). Limited solubility in mostgroundwaters should result in Th inheritance from detritalmaterial during goethite cementation and replacement: asdetrital phases dissolve, their burden of Th and other mod-erately insoluble elements is at least partially precipitated orsorbed onto the surface of growing goethite crystals (Lang-muir and Herman, 1980). This interpretation is consistentwith electron microprobe analysis of goethite grains fromsample LynP-02-09 which, in addition to Th (Table 4), alsoshow enrichment in Al, Si, and Ti (Fig. 5), elements likelyderived from dissolution of detrital phases in the CID. Aslong as the Th-bearing phase undergoes complete dissolu-tion during weathering and ferruginization, the inheritanceof Th derived from detrital minerals does not precluderetrieving reliable goethite precipitation ages. In contrast,the low Th/U values in Roy Hill goethite suggest that thelarge goethite crystals in those samples form by the com-plete dissolution of primary Fe-bearing minerals (carbon-ates or sulfides) or the influx of large quantities ofdissolved iron into cavities, mechanisms distinct from thoseassociated with the precipitation of goethite at the LynnPeak channels.

The most troubling mineralogical observation from ageochronological perspective is the presence of fragmentsof detrital rutile, ilmenite, and zircon in some of our goe-thite samples (Fig. 4). The likely sources for these detrital

phases are Archean in age, and their presence could con-ceivably invalidate the (U–Th)/He results. Detrital phasesmay contribute to anomalously old goethite He ages bothby containing a large amount of He generated since the Ar-chean, and also, because we do not dissolve these phasesduring goethite digestion, by ejecting “parentless” alphaparticles into the goethite since its growth in the Cenozoic.Our observations demonstrate that these mineral contami-nants certainly do not dominate the 4He budget in our sam-ples. If they did, the reproducibility of results for distinctgrains from the same sample (Figs. 4 and 8, Table 4) wouldbe unlikely, unless each goethite fragment contained thesame abundance of contaminants, an improbable scenario.In addition, quantification, through 4He/3He analysis, ofthe 4He possibly contributed by contaminants clearly re-veals that no more than ca. 10% of the total 4He could havecome from phases other than goethite (Figs. 9 and 10). Thesmall amount of 4He in the contaminants probably reflectstheir small grain sizes, since crystal fragments smaller thanca. 20 lm would lose most of their 4He by a-ejection. Thesmall amounts of 4He detected in the insoluble residues ofthe goethite dissolution experiments also confirm that mostof the 4He in our samples comes from authigenic goethitesand not detrital phases, supporting the conclusions above.Lastly, the excellent agreement between the (U–Th)/Heand independent 40Ar/39Ar ages for the same system iscompelling evidence that our geochronological results re-flect authigenic mineral precipitation ages.

Similarly, the close coexistence of detrital and authigenicMn oxides combined with petrographic evidence for morethan one generation of Mn oxides pose problems whenusing these samples in 40Ar/39Ar geochronology.

4.2. 4He/3He and (U–Th)/He geochronology of goethite

(U–Th)/He dating of authigenic goethite (18 grains fromtwo distinct hand specimens) reveals a protracted history ofmineral precipitation for the Lynn Peak CID. Incremental-heating experiments on representative goethites confirmtheir high He retentivity and suitability for geochronologyby the (U–Th)/He method. Additional geochronology onmassive goethite (14 grains from one sample) from a nearbyweathering profile at Roy Hill North (Fig. 1a,b) providesinformation on possible sediment sources for the aggrada-tion of the local drainage and formation of the Lynn PeakCID. (U–Th)/He ages for authigenic goethite cementsrange from 32.7 ± 3.3 to 14.5 ± 1.5 Ma (Table 4, Fig. 13),revealing that the Lynn Peak channel was undergoing ironcementation by ca. 33 Ma and indicating that channelaggradation occurred before this time. Textural evidence– replacement of detrital and early goethite cement bylate-stage cements – suggests that the weathering processesassociated with the ferruginizaton of the CIDs at LynnPeak did not lead to complete dissolution of previous gen-erations of authigenic cements or detrital oxides (Figs. 2–4).Our field observations reveal that under extremely wetconditions, such as those prevailing in the Amazon regiontoday, detrital iron oxyhydroxides eroded from weatheringbanded iron-formations are dissolved very close to thesource area and do not accumulate as sediments in local


drainage systems. Therefore, we conclude that Mioceneweathering in the Hamersley province did not occur underextremely humid conditions, but it certainly took place un-der a much wetter climate than at present to account for theeffective leaching of other elements (Al, Si, Ca, Mg, etc.)and to promote large-scale iron cementation of the channelsediments.

The source of detrital material and the mechanisms ofchannel aggradation that led to CIDs in Western Australiaare unique and still puzzling. Detrital BIF, hematite, andgoethite fragments suggest rapid erosion of previouslyformed weathering profiles. (U–Th)/He ages for goethitegrains from a weathering profile at Roy Hill North(Figs. 1a,b and 2e,f), representative of possible sources ofdetrital material for the Lynn Peak CIDs, yield results rang-ing from 72–60 Ma and clustering at 63.9 ± 6.4 Ma(n = 12), consistent with the existence of an ancient andpossibly extensive weathering blanket covering the higherelevation source areas (Hamersley Surface?) in the region.Unfortunately, we did not attempt to analyse detrital Fe-oxyhydroxides in the Lynn Peak deposit, which preventsus from inferring, from direct (U–Th)/He dating of detritalphases, the ages of the weathering profiles that supplied thedetrital material for the Lynn Peak channel.

The U retentivity of Lynn Peak goethites during vacuumheating clearly differs from that of goethites dated in the pre-vious studies by Shuster et al. (2005) and Heim et al. (2006).Mineralogical analyses, however, reveal that goethites ana-lysed in this study have similar unit-cell parameters, degreesof crystallinity, and relatively similar chemistry to goethitesanalysed in the previous studies. Uranium behaviour duringlaser heating may be dependent on whether and to what de-gree U is structurally bound in the goethite or loosely ad-sorbed on surface sites. Until we understand U distributionand behaviour in goethite, it may be necessary to performindependent tests on new samples prior to (U–Th)/He datingto identify an extraction temperature that ensures completeHe release and quantitative retention of U and Th.

Finally, the CIDs at Lynn Peak are particularly usefulbecause, unlike most CIDs, they contain blocks of detritalMn oxides and also contain authigenic Mn oxide cementssuitable for independent dating by the 40Ar/39Ar method(Vasconcelos et al., 1994). Laser incremental-heating40Ar/39Ar analysis of detrital cryptomelane blocks (fivegrains from two samples), partially recrystallised detritalgrains (six grains from two samples), and authigenic Mnoxide cements (26 grains from nine samples) permits inde-pendently constraining the maximum and minimum agesof aggradation and CID cementation at Lynn Peak andtesting the accuracy of the results and interpretations ob-tained with the (U–Th)/He and 4He/3He methods.

4.3. 40Ar/39Ar geochronology of Mn oxides

Field and petrographic observations reveal two groupsof Mn oxides at Lynn Peak: an older detrital group, pre-sumably derived by erosion, transport, and deposition ofsupergene phases from a previously formed weathering pro-file; and a younger authigenic group, composed of numer-ous generations of Mn oxide cements precipitated in situ

by the partial dissolution of detrital phases, the partialmobilization of Mn (and K+, Ba2+, etc.) in solution, andthe reprecipitation of authigenic Mn oxides (Figs. 2 and3). Influx of paleogroundwater enriched in Mn2+, Ba2+,and K+ probably resulted from weathering of nearby lithol-ogies, such as the Marra Mamba BIF or units from the Bal-four Formation (Fig. 1d). The dissolution of these elementsfrom source areas nearby and its reprecipitation in thechannels suggest moderately wet conditions to promoteeffective weathering and leaching of the source areas, butconditions dry enough to prevent the complete removal ofthe dissolved elements, in solution, from the system.

Weathering geochronology by the 40Ar/39Ar methodclearly indicates that detrital oxides are older than 44 Ma,while authigenic oxides are younger than ca. 35 Ma(Fig. 13). These results date the aggradation of the LynnPeak CID at the 44–35 Ma interval. It also reveals thatthe region was covered by a more widely distributed ancientweathering profile, as old or older than 65 Ma. This weath-ering profile, partially eroded, may have been the weather-ing blanket covering the Hamersley Surface (Campanaet al., 1964). Partial erosion of that blanket in the 44–35 Ma interval provided the detrital material now presentin the CIDs.

4.4. Comparison between 40Ar/39Ar and (U–Th)/He

geochronology results

(U–Th)/He results for goethites from Lynn Peak areconcordant with the 40Ar/39Ar results for Mn oxides(Fig. 13). Authigenic goethites reveal that the channels wereundergoing iron cementation at ca. 33 Ma (Fig. 13), indi-cating that channel aggradation occurred before this time.Channel cementation processes were particularly strong inthe 33–30 Ma period, as recorded both in (U–Th)/He agesfor authigenic goethites and the Mn oxide 40Ar/39Ar resultsfor authigenic Mn oxides (Fig. 13). Another strong periodof authigenic goethite precipitation at ca. 20 Ma is also asignificant interval of Mn oxide precipitation (Fig. 13).

The combined 40Ar/39Ar and (U–Th)/He results alsoprovide compatible ages for the weathering profiles fromwhich sediments in the Lynn Peak CID were derived. (U–Th)/He ages for goethite grains from a weathering profileat Roy Hill North, a possible source of detrital materialfor the Lynn Peak CIDs, yield results ranging from 66–60 Ma and clustering at 63.5 ± 1.7 Ma (n = 9), within errorof the 40Ar/39Ar results for the oldest detrital Mn oxideblocks at Lynn Peak (Fig. 13). In summary, the combined40Ar/39Ar and (U–Th)/He dataset shows a compatible his-tory of mineral precipitation and sediment sources for theLynn Peak CID.

4.5. Comparison between Lynn Peak and Yandi CID results

The youngest goethite dated at Lynn Peak(14.5 ± 0.9 Ma) is considerably older than the youngestgoethites dated at Yandi by Heim et al. (2006). However,it is important to remember that all samples dated for theLynn Peak channels are surface samples. At Yandi, Heimet al. (2006) only measured young goethite cement ages at


the middle to the bottom of the channel. Whether the ab-sence of young goethites at Lynn Peak simply reflects depthfrom which the samples were collected remains to be re-solved. However, the (U–Th)/He results showing that thesurface of the Lynn Peak CID was already undergoing ironcementation by ca. 33 Ma are entirely compatible with theinterpretation by Heim et al. (2006) for the Yandi CID.At Yandi, CID goethite cement becomes younger withdepth into the profile. Propagation of the age vs. depthtrend would suggest an original age of ca. 35–40 Ma forthe original surface, which lead Heim et al. (2006) to con-clude that channel aggradation did occur in the Eocene–Oligocene transition, as proposed by MacPhail and Stone(2004). Both the goethite (U–Th)/He and Mn oxide40Ar/39Ar results for Lynn Peak are compatible with thatinterpretation.

4.6. Weathering, erosion, and landscape evolution at

Hamersley

Pollen assemblages collected from basal clays (MunjinaMember) from the Yandi CID have been interpreted as evi-dence that channel aggradation in the Hamersley Provincestarted sometime in the Late Eocene or Early Oligocene(MacPhail and Stone, 2004), and it was initiated in responseto drastic changes in global climate and ocean circulationpatterns. Iron cementation of 50 to 70 m of Fe-rich sedi-ments is attributed to increasingly warm and seasonallywet conditions in the Late Oligocene and Early Miocene(MacPhail and Stone, 2004).

Our (U–Th)/He results show that ferruginization of theCIDs was already ongoing at ca. 33 Ma (Fig. 13) and pos-sibly before, suggesting that channel aggradation predates33 Ma. Our 40Ar/39Ar results show that authigenic crypto-melane replacing early goethite cements may be as old as35 Ma (Fig. 12), suggesting that channel aggradation andthe initial stages of channel cementation preceded this time.Channel aggradation probably resulted from erosion ofpreviously formed and more widely distributed regionalweathering profiles, probably associated with the postu-lated Mesozoic Hamersley Surface. Aggradation of CIDsmost probably initiated in response to global climatechange (transition from greenhouse to icehouse conditions)before 35 Ma, possibly reflecting crustal reorganization andAustralia’s accelerated northward drift away from Antarc-tica. Tectonic collisions as Australia moved northward,such as those detected for the 26–23 Ma period (Kneselet al., 2008), could also have contributed to uplift, wide-spread erosion, and channel aggradation. Unfortunately,evidence for active tectonism in northwestern Australia inthe 45–35 Ma period is missing.

Whether climate change or tectonic reactivation playedthe dominant role in regional erosion and channel aggrada-tion in the Hamersley region, an important factor is that Fecementation of the channel sediments must have occurredrapidly after aggradation to permit the fossilization ofwood fragments in these channels (Heim et al., 2006; Mor-ris and Ramanaidou, 2007). Our geochronological resultsindicate that goethite cementation was ongoing at ca.33 Ma and continued throughout the Miocene (Fig. 13).

Ferruginization of channel sediments, leading to thegeneration of the goethite-cemented CIDs, probably oc-curred rapidly after channel aggradation, when acid-reduc-ing groundwater, saturated in organic acids generated bythe decay of detrital plant material, partially dissolveddetrital minerals, and goethite reprecipitated from Fe2+-rich solutions, possibly aided by microorganisms, at thegroundwater-atmosphere interface. The proposed processis similar to mechanisms interpreted for the genesis ofbog iron deposits (Crerar et al., 1979), except for the factthat in the CIDs iron came in large part from the detritalmaterial already in the channels.

Transition towards aridity in Australia in the Neogenewould have caused the regional groundwater table to sub-side, promoting deepening of the oxidation–reduction frontand the precipitation of goethite cements at progressivelygreater depths in the channels (Heim et al., 2006). A returnto warm and wet (perhaps seasonal) climatic conditions inthe Middle Miocene (McGowran et al., 2000) would havecaused groundwater tables to rise, once again raising theoxidation–reduction front to shallower depths, promotingincreased Fe2+ dissolution, transport, and reprecipitationof Fe3+-oxihydroxide near the surface of the channels. Par-tial recrystallization of Fe and Mn oxides at the Lynn PeakCID surface, revealed by (U–Th)/He and 40Ar/39Ar results,and also observed at Yandi (Heim et al., 2006), supportsthis interpretation. Finally, deepening of the water tablein the late Miocene, when Australia underwent a pro-nounced transition towards aridity (Truswell and Harris,1983; Frakes, 1999; McGowran et al., 2000; Vasconceloset al., 2008), promoted the preservation of the now mostlyfossilized CID deposits.

Presently, most CIDs in the Hamersley Province are par-tially or almost completely saturated by groundwater. Yet,none of the goethite samples dated from Yandi or Lynn Peakrecord recent mineral precipitation. This could reflect thestrongly seasonal climate and the absence of abundant vege-tation cover, which would preclude the saturation of thegroundwater with organic acids, a pre-requisite for Fe reduc-tive-dissolution and reprecipitation. On the other hand, thepresence of incipient ferruginization directly underlying re-cent drainage systems suggests that some weathering andpost-depositional alteration of CIDs (dissolution and recrys-tallization of detrital and authigenic cements) may have oc-curred more recently, a hypothesis that can be tested usingweathering geochronology by U-series dating.

Finally, the preservation of older than 33 Ma ferrugi-nized channel sediments at the present land surface revealsthat erosion rates in this part of Australia have been veryslow throughout the Neogene, otherwise these depositswould have been incised and eroded away, as they fill thesame valleys as the modern drainage system (Fig. 1). Theclimatic and tectonic conditions conducive to such low ero-sion rates for such an extended period of time still eludeexplanation.

5. CONCLUSIONS

New goethite (U–Th)/He results from the Lynn Peakchannel deposit confirm that CIDs of the Hamersley


Province appear to have completely aggraded by ca. 33 Ma.Authigenic Mn-oxide cements in the same channel depositssuggest that aggradation may have been as early as 35 Ma.Channel aggradation occurred by erosion of previouslyformed ancient – older than 64 Ma – weathering profiles,during rapid erosion under semi-arid conditions. Ferrugini-zation of the channel sediments must have started during orimmediately after channel aggradation to permit fossiliza-tion of vegetation fragments deposited in the channel.Excursion towards more humid climates in the Early-Mid-dle Miocene promoted the partial dissolution of detritalphases and early-stage authigenic cements, and partialcementation by late-stage authigenic cements.

The application of weathering geochronology to investi-gate these systems depends on the He retentivity of supergeneFe-oxyhydroxides. This study shows that supergene goethitecan be extremely retentive of He, surviving at surface condi-tions for extended periods, even when exposed to potentiallyintense solar heating, as in Western Australia. But even whensuitable supergene goethite exists, paragenetic complexities(several generations of the same phase intimately intergrown,presence of small inclusions of U- and Th-bearing contami-nants) may add difficulties to the interpretation of the isoto-pic results. Interpreting these results requires high-resolutionmineralogical characterization, including XRD, SEM, andEMPA, without which geochronological interpretation re-mains tentative. Finally, this study demonstrates that (U–Th)/He dating of supergene minerals must be carried out un-der strictly controlled laboratory heating conditions; other-wise, U and Th may be lost from the grains during Heextraction rendering the results inaccurate.

ACKNOWLEDGMENTS

We thank Graham Broadbent and RT Exploration for logisticsupport; UQCMM staff (Ron Rash) for help during microanalysis;Peter Colls for sample preparation; Janet Sisterson for help withproton-irradiation, and Lindsey Hedges for help with (U–Th)/Heanalyses, and Albert Mostert for interpretation of synchrotron re-sults. This project was partly funded by ARC DP0666925, UQ-APA scholarships to J.A.H. and K.W., and an UQ-AGES scholar-ship to H.M. Gilles Ruffet, and four anonymous reviewers pro-vided useful suggestions for improving the original submission.Peter Reiners provided excellent editorial handling and his sugges-tions were valuable in improving this manuscript.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.gca.2013.03.037.

REFERENCES

Arancibia G., Matthews S. J. and Perez de Arce C. (2006) K–Arand 40Ar/39Ar geochronology of supergene processes in theAtacama Desert, Northern Chile: tectonic and climatic rela-tions. J. Geol. Soc. 163, 107–118.

Beauvais A., Ruffet G., Henocque O. and Colin F. (2008) Chemicaland physical erosion rhythms of the West African Cenozoic

morphogenesis: the 40Ar/39Ar dating of supergene K–Mnoxides. J. Geophys. Res. – Earth Surf. 113, F04007.

Bissig T. and Riquelme R. (2010) Andean uplift and climateevolution in the southern Atacama Desert deduced fromgeomorphology and supergene alunite-group minerals. Earth

Planet. Sci. Lett. 299, 447–457.Campana B., Burns W. G., Hughes F. E., Muceniekas E. and

Whitcher I. G. (1964) Discovery of the Hamersley iron deposits(Duck Creek-Mount Pyrton-Mount Turner area). Proc. Aust.

Inst. Min. Metal. 210, 1–30.Carmo I. O. and Vasconcelos P. M. (2004) Geochronological

evidence for pervasive Miocene weathering, Minas Gerais,Brazil. Earth Surf. Proc. Land. 29(11), 1303–1320.

Carmo I. O. and Vasconcelos P. M. (2006) 40Ar/39Ar geochronol-ogy constraints on late Miocence weathering rates in MinasGerais, Brazil. Earth Planet. Sci. Lett. 241(1–2), 80–94.

Coelho A.A. (2006) TOPAS-Academic version 4.Colin F., Beauvais A., Ruffet G. and Henocque O. (2005) First

40Ar/39Ar geochronology of lateritic manganiferous pisolites:implications for the Palaeogene history of a West Africanlandscape. Earth Planet. Sci. Lett. 238, 172–188.

Cookson D. J., Hunter B. A., Kennedy S. J. and Garrett R. F.(1998) Multiple-wavelength powder diffraction using imagingplates at the Australian National Beamline. J. Synchrotron

Radiat. 5, 926–928.Creagh D. C., Foran G. J., Cookson D. J., Garrett R. F. and

Johnson F. (1998) An eight-position capillary sample spinningstage for the diffractometer at BL20B at the Photon Factory. J.

Synchrotron Radiat. 5, 823–825.Crerar D. A., Knox G. W. and Means J. L. (1979) Biogeochemistry

of bog iron in the New Jersey pine barrens. Chem. Geol. 24,111–135.

Dalstra H. J., Gill T., Faragher A., Scott B. and Kakebeeke (2009)Channel iron deposits – a major new district around Caliwin-gina Creek, Central Hamersley Ranges, Western Australia.Proceedings Iron Ore 2009, AusIMM, Perth.

Dammer D., McDougall I. and Chivas A. R. (1999) Timing ofweathering-induced alteration of manganese deposits in Wes-tern Australia: evidence from K/Ar and 40Ar/39Ar dating. Econ.

Geol. 94, 87–108.Deino A. (2009) Mass Spec Version 5.527, Program Documentation.

Berkeley Geochronology Center. <http://www.bgc.org>.Farley K. A. (2002) (U–Th)/He dating: techniques, calibrations,

and applications. In Reviews in Mineralogy and Geochemistry:

Noble Gases in Geochemistry and Cosmochemistry (eds. D.Porcelli, C. J. Ballentine and R. Wieler). Mineralogical Societyof America, Washington, pp. 819–844.

Farley K. A., Reiners P. W. and Nenow V. (1999) An apparatus forhigh precision helium diffusion measurements from minerals.Anal. Chem. 71, 2059–2061.

Feng Y. X. and Vasconcelos P. M. (2001) Quaternary continentalweathering geochronology by laser-heating 40Ar/39Ar analysisof supergene cryptomelane. Geology 29, 635–638.

Feng Y. X. and Vasconcelos P. M. (2007) Chronology ofPleistocene weathering processes, southeast Queensland, Aus-tralia. Earth Planet. Sci. Lett. 263, 275–287.

Frakes L. A. (1999) Evolution of Australian Environments, Flora of

Australia. Commonwealth Scientific and Industrial ResearchOrganisation Publishing, Melbourne, pp. 163–203.

Garrett R. F. et al. (1995) Powder diffraction using imaging platesat the Australian national beamline facility at the photonfactory. Rev. Sci. Instrum. 66, 1351–1353.

Gualtieri A. F. and Venturelli P. (1999) In situ study of thegoethite–hematite phase transformation by real time synchro-tron powder diffraction. Am. Mineral. 84, 895–904.



http://refhub.elsevier.com/S0016-7037(13)00205-6/h0005



















































http://www.bgc.org



























Hautmann S. and Lippolt H. J. (2000) 40Ar/39Ar dating of centralEuropean K–Mn oxides – a chronological framework ofsupergene alteration processes during the Neogene. Chem.

Geol. 170, 37–80.Heim J. A. (2007) Geochronology of weathering and landscape

evolution, Hamersley Iron Province, Australia. Ph. D. thesis,The University of Queensland, p. 214.

Heim J. A., Vasconcelos P. M., Shuster D. L., Farley K. A. andBroadbent G. C. (2006) Dating palaeochannel iron ore by (U–Th)/He analysis of supergene goethite. Geology 34, 173–176.

Henocque O., Ruffet G., Colin F. and Feraud G. (1998) 40Ar/39Ardating of West African lateritic cryptomelanes. Geochim.

Cosmochim. Acta 62, 2739–2756.Knesel K. M., Cohen B. E., Vasconcelos P. M. and Thiede D. S.

(2008) Rapid change in the drift of the Australian Plate recordscollision with the Ontong Java Plateau. Nature 254, 754–756.

Kuiper K. F., Deino A., Hilgen F. J., Krijgsman W., Renne P. R.and Wijbrans J. R. (2008) Synchronizing rock clocks of earthhistory. Science 320, 500–504.

Langmuir D. and Herman J. S. (1980) The mobility of thorium innatural waters at low temperatures. Geochim. Cosmochim. Acta

44, 1753–1766.Lee J.-Y., Marti K., Severinghaus J. P., Kawamura K., Yoo H.-S.,

Lee J. B. and Kim J. S. (2006) A redetermination of the isotopicabundances of atmospheric Ar. Geochim. Cosmochim. Acta 70,4507–4512.

Li J. W. and Vasconcelos P. M. (2002) Cenozoic continentalweathering and its implications for palaeoclimate: evidencefrom 40Ar/39Ar geochronology of supergene K–Mn oxides inMt. Tabor, central Queensland, Australia. Earth Planet. Sci.

Lett. 200, 223–239.Li J. W., Vasconcelos P. M., Duzgoren-Aydin N., Yan D. R.,

Zhang W., Deng X. D., Zhao X. F., Zeng Z. P. and Hu M. A.(2007) Neogene weathering and supergene manganese enrich-ment in subtropical South China: an 40Ar/39Ar approach andpaleoclimatic significance. Earth Planet. Sci. Lett. 256, 389–402.

Lippolt H. J., Brander T. and Mankopf N. R. (1998) An attempt todetermine formation age of goethites and limonites by(U.Th)–4He dating. Neues Jahrb. Mineral. Monatsh. 200, 505–528.

MacPhail M. K. and Stone M. S. (2004) Age and palaeoenviron-mental constraints on the genesis of the Yandi channel irondeposits, Marillana Formation, Pilbara, northwestern Austra-lia. Aust. J. Earth Sci. 51, 497–520.

McGowran B. et al. (2000) Australasian palaeobiogeography: thePalaeogene and Neogene record. In Palaeobiogeography of

Australasian Faunas and Floras (ed. A. J. T. Wright).Association of Australasian Palaeontologists, Canberra, pp.405–470.

Morris R. C. (1988) The goethitic channel ores of the HamersleyProvince. A Pilot Study of the “Robe Pisolite”: CSIROExploration Geoscience. Report No. MG70R.

Morris R. C. and Ramanaidou E. R. (2007) Genesis of the channeliron deposits (CID) of the Pilbara region, Western Australia.Aust. J. Earth Sci. 54, 733–756.

Morris R. C., Ramanaidou E. and Horwitz R. C. (1993) ChannelIron Deposits of the Hamersley Province: CSIRO Explorationand Mining Restricted Report 399R.

Morris R. C., Kneeshaw M. and Ramanaidou E. R. (2007) DatingPaleochannel Iron Ore by (U–Th)/He Analysis of SupergeneGoethite, Hamersley Province, Australia: COMMENT. GSAElectronic.

Pidgeon T. T., Brander R. and Lippolt H. J. (2004) Late Miocene(U–Th)/He ages of ferruginous nodules from lateritic duricrust,Darling Range, Western Australia. Aust. J. Earth Sci. 51, 901–909.

Ramanaidou E. R., Morris R. C. and Horwitz R. C. (2003)Channel iron deposits of the Hamersley Province, WesternAustralia. Aust. J. Earth Sci. 50, 669–690.

Reiners P. W. and Farley K. A. (1999) He diffusion and (U–Th)/Hethermochronometry of titanite. Geochim. Cosmochim. Acta 63,3845–3859.

Rietveld H. M. (1969) A profile refinement method for nuclear andmagnetic structures. J. Appl. Crystallogr. 2, 65–71.

Ruffet G., Innocent C., Michard A., Fbraud G., Beauvais A.,Nahon D. and Hamelin B. (1996) A geochronological40Ar/39Ar and 87Rb/87Sr study of K–Mn oxides from theweathering sequence of Azul, Brazil. Geochim. Cosmochim.

Acta 60, 2219–2232.Sabine T. M., Kennedy B. J., Garrett R. F., Foran G. J. and

Cookson D. J. (1995) The performance of the Australianpowder diffractometer at the photon-factory, Japan. J. Appl.

Crystallogr. 28, 513–517.Spier C. A., Vasconcelos P. M. and Oliveira S. M. B. (2006)

40Ar/39Ar geochronological constraints on the evolution oflateritic iron deposits in the Quadrilatero Ferrıfero, MinasGerais, Brazil. Chem. Geol. 234(1–2), 79–104.

Shuster D. L. and Farley K. A. (2004) 4He/3He thermochronom-etry. Earth Planet. Sci. Lett. 217, 1–17.

Shuster D. L., Farley K. A., Sisterson J. and Burnett D. S. (2004)Quantifying the diffusion kinetics and spatial distributions ofradiogenic 4He in minerals containing proton-induced 3He.Earth Planet. Sci. Lett. 217, 19–32.

Shuster D. L., Vasconcelos P. M., Heim J. A. and Farley K. A.(2005) Weathering geochronology by (U–Th)/He dating ofgoethite. Geochim. Cosmochim. Acta 69, 659–673.

Shuster D. L., Farley K. A., Vasconcelos P. M., Balco G.,Monteiro H. S., Waltenberg K. and Stone J. O. (2012)Cosmogenic 3He in hematite and goethite from Brazilian“canga” duricrust demonstrates the extreme stability of thesesurfaces. Earth Planet. Sci. Lett. 329–330, 41–50.

Stone M. S., George A. D. and Kneeshaw M. (2002) Stratigraphyand sedimentary features of the Tertiary Yandi Channel irondeposits, Hamersley Province, Western Australia. AusIMM

Bull. 6, 36–41.Strutt R. J. (1908) Helium and radioactivity in rare and common

minerals. Proc. R. Soc. London A 80, 572–594.Trendall A. F., Compston W., Nelson D. R., De Laeter J. R. and

Bennett V. C. (2004) SHRIMP zircon ages constraining thedepositional chronology of the Hamersley Group, WesternAustralia. Aust. J. Earth Sci. 51, 621–644.

Truswell E. M. and Harris W. K. (1983) The Cainozoic palaeo-botanical record in arid Australia; fossil evidence for the originof an arid-adapted flora. In Evolution of the Flora and Fauna of

Arid Australia (eds. W. R. Barker and P. J. M. Greenslade).Peacock Pubs. Frewville, S.A., Australia, pp. 67–76.

Van Niekerk H. S., Gutzmer J., Beukes N. J., Phillips D. andKiviets G. B. (1999) An 40Ar/39Ar age of supergene K–Mnoxyhydroxides in a post-Gondwana soil profile on the Highveldof South Africa. S. Afr. J. Sci. 95, 450–454.

Vasconcelos P. M. (1999a) 40Ar/39Ar geochronology of supergeneprocesses in ore deposits. In Application of Radiogenic Isotopes

to Ore Deposit Research and Exploration. Reviews in Economic

Geology (eds. D. Lambert David and J. Ruiz). Society ofEconomic Geologists, Boulder, CO, United States, pp. 73–113.

Vasconcelos P. M. (1999b) K–Ar and 40Ar/39Ar geochronology ofweathering processes. Annu. Rev. Earth Planet. Sci. 27, 183–229.

Vasconcelos P. M. and Conroy M. (2003) Geochronology ofweathering and landscape evolution, Dugald River valley, NWQueensland. Australia. Geochim. Cosmochim. Acta 67, 2913–2930.




















































































































































Vasconcelos P. M., Becker T. A., Renne P. R. and Brimhall G. H.(1992) Age and duration of weathering by 40K–40Ar and40Ar/39Ar analysis of potassium–manganese oxides. Science

258, 451–455.Vasconcelos P. M., Becker T. A., Renne P. R. and Brimhall G. H.

(1994) Direct dating of weathering phenomena by K–Ar and40Ar/39Ar analysis of supergene K–Mn oxides. Geochim.

Cosmochim. Acta. 58, 1635–1665.Vasconcelos P. M., Onoe A. T., Kawashita K., Soares A. J. and

Teixeira W. (2002) 40Ar/39Ar geochronology at the Instituto deGeociencias, USP: instrumentation, analytical procedures, andcalibration. An. Acad. Bras. Cienc. 74(2), 297–342.

Vasconcelos P. M., Heim J. A., Shuster D. L., Farley K. A. andBroadbent G. C. (2007) Dating Paleochannel Iron Ore by (U–Th)/He Analysis of Supergene Goethite, Hamersley Province,Australia: RESPONSE. GSA Electronic.

Vasconcelos P. M., Knesel K. M., Cohen B. E. and Heim J. A.(2008) Geochronology of the Australian Cenozoic: a history oftectonic and igneous activity, weathering, erosion, and sedi-mentation. Aust. J. Earth Sci. 55, 865–914.

Associate editor: Marc Norman

























Date post:	11-Dec-2016
Category:	Documents
Upload:	kathryn
View:	216 times
Download:	3 times

40Ar/39Ar and (U–Th)/He – 4He/3He geochronology of landscape evolution and channel iron deposit...

Documents