Earth and Planetary Science Letters 341–344 (2012) 255–267

Contents lists available at SciVerse ScienceDirect

Earth and Planetary Science Letters journal homepage: www.elsevier.com/locate/epsl

Letters

Planar defects as Ar traps in trioctahedral micas: A mechanism for increased Ar retentivity in phlogopite A. Camacho a,n, J.K.W. Lee b, J.D. Fitz Gerald c, J. Zhao b, Y.A. Abdu a, D.M. Jenkins d, F.C. Hawthorne a, T.K. Kyser b, R.A. Creaser e, R. Armstrong c, L.W. Heaman e a

Department of Geological Sciences, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2 Department of Geological Sciences and Geological Engineering, Queen’s University, Kingston, Ontario, Canada K7L 3N6 c Research School of Earth Sciences, The Australian National University, Canberra, ACT 0200, Australia d Department of Geological Sciences and Environmental Studies, Binghamton University, PO Box 6000, Binghamton, NY 13902-6000, USA e Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2R3 b

a r t i c l e i n f o

abstract

Article history: Received 24 December 2011 Received in revised form 22 May 2012 Accepted 26 May 2012 Editor: B. Marty Available online 20 July 2012

The effects of planar defects and composition on Ar mobility in trioctahedral micas have been investigated in samples from a small marble outcrop (  500 m2) in the Frontenac Terrane, Grenville Province, Ontario. These micas crystallized during amphibolite-facies metamorphism at  1170 Ma and experienced a thermal pulse  100 Ma later at shallow crustal levels associated with the emplacement of plutons. 87Rb/86Sr ages of the phlogopites range from  950 to  1050 Ma, consistent with resetting during the later thermal event. The same phlogopites however, give 40Ar/39Ar ages between  950 and 1160 Ma, spanning the age range of the two thermal events. This result is intriguing because these micas have undergone the same thermal history and were not deformed after peak metamorphic conditions. In order to understand this phenomenon, the chemical, crystallographical, and microstructural nature of four mica samples has been characterized in detail using a wide range of analytical techniques. The scanning electron microscope (SEM), electron microprobe (EMP), and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) data show that the micas are chemically homogeneous (with the exception of Ba) and similar in composition. The Fourier transform infrared spectroscopy and Mossbauer results show that the M sites for three of the micas are dominated by divalent cations and the Fe3 þ /(Fe2 þ þFe3 þ ) ratio for all four phlogopites ranges from 0.10 to 0.25. The stable-isotopic data for calcite indicate that this outcrop was not affected by hydrothermal fluids after peak metamorphism. No correlation between chemical composition and 87Rb/86Sr and 40Ar/39Ar age or between crystal size and 40Ar/39Ar age is observed. The only major difference among all of the micas was revealed through transmitted electron microscope (TEM), which shows that the older 1M micas contain significantly more layer stacking defects, associated with crystallization, than the younger micas. We propose that these defect structures, which are enclosed entirely within the mineral grain may serve as Ar traps and effectively increase the Ar retentivity of the mineral. As this phenomenon has not been previously documented in micas, this may have significant implications for the interpretation of 40Ar/39Ar ages of minerals which have similar defect structures. & 2012 Elsevier B.V. All rights reserved.

Keywords: Ar retentivity diffusion Grenville Province ionic porosity stable isotopes transmitted electron microscopy

1. Introduction Biotite forms under a wide range of temperature and pressure conditions, and has become one of the most commonly dated minerals by the 40Ar/39Ar technique used to reconstruct the cooling history of metamorphic terranes (McDougall and Harrison, 1999). In some cases, however, interpreting the geological significance of biotite 40Ar/39Ar plateau ages has proven

n

Corresponding author. Tel.: þ1 204 474 7413; fax: þ1 204 474 7623. E-mail address: [email protected] (A. Camacho).

0012-821X/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.epsl.2012.05.041

difficult as the mineral can yield anomalously old dates that cannot be interpreted without additional geochronological or geological information (e.g., Foland, 1983). These old dates are typically attributed to the incorporation of excess 40Ar (40ArE) during cooling or deformation. In addition, deformation induces microstructures and defects which have the potential to behave as intra-grain fast-diffusion pathways (e.g., Lee, 1995; Hodges et al., 1994; Hodges and Bowring, 1995). Stacking disorder, twinning and polytypism are common microtextures in micas. Although much of our knowledge of these phenomena was established using X-ray diffraction (Ross et al., 1966), transmission electron microscopy (TEM) has proven

256

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

very effective for characterizing such structural features (e.g., Bell and Wilson, 1977; Fregola and Scandale, 2011; Xu and Veblen, 1995), particularly at the unit-cell scale (e.g., Noe and Veblen, 1999). Structural defects are commonly associated with the growth of mica crystals, although defects are seldom considered in studies of Ar release from micas. In one notable exception, Kramar et al. (2003) examined the effects of stacking faults in muscovite, associated with splitting of dislocations formed during deformation long after crystal growth. They found that Ar loss could be directly correlated with the density of extended defects. Here, we present integrated geochronologic and crystallographic data for a range of coexisting trioctahedral micas in a single marble outcrop from the Grenville Province in Canada that show a 40Ar/39Ar plateau age difference 4100 Ma between samples. Because the micas come from the same outcrop and have therefore experienced the same geologic history, this outcrop serves as a natural laboratory to examine the influence of chemical composition and crystallographic defects on Ar transport in micas. This age difference is surprising considering that the rocks have experienced the same thermal history. Because there may be several factors contributing to this result, including the presence of microstructures, grain size, 40ArE (excess argon from fluid inclusions or in the crystal structure), localized thermal pulses associated with magmatism, and chemical composition (e.g., Baxter et al., 2002; Onstott et al., 1989; Villa and Puxeddu,

1994), we have used a variety of analytical techniques to explore this enigmatic age difference.

2. Geological setting The Grenville Province covers an area approximately 1900 km in length  400 km in width, and is bounded by the Superior craton in the northwest and the Appalachian Mountains in the southeast (Fig. 1). This province preserves tectonically stacked slices of Neoarchean, Paleoproterozoic and Mesoproterozoic rock assemblages resulting from the Mesoproterozoic continental collision associated with the assembly of the supercontinent Rodinia (Hoffman, 1991). It is one of the best examples of an ancient exhumed continental-collision zone and records a protracted tectonic history on one of the continental margins preceding the collision (Carr et al., 2000; Ludden and Hynes, 2000). The southwestern Grenville province is traditionally subdivided into three lithotectonic belts of metamorphic rocks: the Central Gneiss Belt, the Central Metasedimentary Belt (CMB), and the Central Granulite Terrane (Supplementary material supplAFig. 1; Davidson, 1986; Hanmer et al., 2000). Several terranes have been identified within the CMB, including the Bancroft, Elzevir, Sharbot Lake and Frontenac terranes, located respectively from northwest to southeast. Each of the

80°W

60°W

CANADA

LEGEND

bay Buck bay

Churchill Province Superior Province

Green Symbols

60°N

Labrador

le vil en ce Gr rovin P

Ontario Québec

Geological contact: inferred Trace of layering Scale

N Synform and axial trace

40°N

500 km

USA

Road 0

1

2 km

Skootamatta intrusive suite (>1050 -1075 Ma)

White lake

Granodiorite - granite: undeformed

Felsic granitoids (>1100 Ma < 1250 Ma) Frontenac intrusive suite (>1150 -1175 Ma)

Tichborne monzonite: undeformed/gneissic Anorthositic gabbro, gabbroic anorthosite: generally gneissic

Quartofeldspathic gneiss: quartz-biotite-perthite-plagioclase

O1

Thirty Island lake

Pyroxene gneiss (mainly derived from calc-silicate rocks) Quartofeldspathic gneiss: quartz-biotite-feldspar ± garnet; ± cordierite; ± hypersthene or ± sillimanite Marble ± calc-silicates

Fig. 1. Geological map of the Glendower vicinity, Tichborne area (Easton, 2001; Wynne-Edwards, 1965). Inset: sketch map of North America showing the relation between the Grenville and older Archean Superior Province (after Hoffman, 1989).

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

terranes has distinctive lithologic and magmatic associations, and they are together considered to represent elements of a larger arc or arc system that was subsequently amalgamated (Carr et al., 2000). The Frontenac terrane contains marbles in association with metapelites, quartzites, amphibolites, and orthogneisses, all metamorphosed to upper-amphibolite facies. U/Pb dating of detrital zircon from quartzite indicates that the original deposition of sediments is younger than 1300 Ma (Sager-Kinsman and Parrish, 1993). The Frontenac terrane underwent its peak regional metamorphism at low pressure ( r6 kbar) and high temperature ( Z750 1C) from ca. 1180 to 1160 Ma (U/Pb zircon and titanite ages; Carr et al., 2000; Mezger et al., 1993; van Breemen and Davidson, 1988). 40Ar/39Ar cooling ages for hornblende from gneisses and amphibolites range 1125–1104 Ma, suggesting that this area has not been significantly affected by regional metamorphic events since ca. 1120 Ma (Cosca et al., 1991; van der Pluijm et al., 1994). This is also suggested by the presence of undeformed and unmetamorphosed 1160 Ma diabase dykes in the region (Carr et al., 2000). Plutonic rocks in the Frontenac terrane fall into two major age groups, 1180–1150 Ma of the Frontenac suite, and ca. 1090– 1065 Ma of the Skootamatta granitoid suite (Davidson and van Breemen, 2000). The Skootamatta suite was emplaced some 100 Ma later than the Frontenac suite during the extensional period after the amalgamation of the Frontenac-Adirondack and Composite Arc Belts (suppA Fig. 1). Apart from the introduction of small plutons of the Skootamatta suite, very little seems to have happened in the Frontenac terrane after ca. 1150 Ma as it was at structurally high levels. 2.1. Sample site The studied outcrop (O1; Fig. 1) is located near Godfrey, Ontario, northeast of the village of Verona, about 35 km by road northwest of the city of Kingston (suppA Fig. 1). This outcrop, which lies within the Frontenac terrane of the CMB, is exposed along a road cut on both sides of the road and is approximately 60 m in length and 8 m in height. The vertical exposure is excellent as the surfaces are not masked by sand or vegetation. The main lithologic units present are marble, pyroxenite, syenite and pegmatite, and the contacts between neighboring units are generally sharp. Marble is the predominant rock-type in outcrop O1, with a white-gray weathered surface and pink-white fresh surface. In general, this unit is quite homogeneous and is usually medium- to coarse-grained. Thin laminations, caused by a concentration of phlogopite, pyroxene and clinohumite, define the well-developed foliations with strikes ranging 070–1001 and dips ranging 75– 851S. A detailed description of the marble in thin section as well as, modal abundances and the cathodoluminscence images can be found in Supplementary material (B). As part of a larger study, four samples were selected from two locations in the outcrop, location 1 (samples 1A and 1B) and location 2 (samples 1J and 1K). These locations were chosen because each contains neighboring marble units with phlogopite of different color (suppA Fig. 2) and chemical composition. In addition, no structural discontinuities or cross-cutting relations were observed along lithological boundaries.

3. Analytical results Several analytical techniques were used to characterize the micas. The following experimental details are described in the following papers: electron microprobe analysis: Oberti et al.

257

¨ (2010); Mossbauer Spectroscopy: Abdu et al. (2008); Fourier transform infrared spectroscopy: Kovacs et al. (2010); crystalstructure refinement: Hawthorne and Simmons (2010); 87Rb/86Sr analysis: Creaser et al. (2004); diffusion experiments: Sharma and Jenkins (1999). Supplementary material (C) contains the experimental techniques for laser ablation inductively coupled mass spectrometry, 40Ar/39Ar and 40Ar/36Ar analysis, U–Pb, stable isotopes, and transmission electron microscopy. 3.1.

40

Ar/39Ar geochronology

Sixty-three single euhedral grains of phlogopite from samples 1A, 1B, 1J, and 1K were carefully hand-picked to be free of alteration and inclusions based on stereomicroscope appearance, and analyzed by the 40Ar/39Ar method. The grain sizes of 41 crystals were in the range 0.7–2.0 mm (large) and the remaining 22 were in the range 125–150 mm (small). Analysis of the small crystals typically involved two steps, and they were analyzed to determine whether there was a correlation between grain size and age. All results are summarized in Table 1 and descriptions of the age spectra are presented in Supplementary material (D). All large crystals exhibit age spectra having plateau segments representing 450% of 39Ar released, with MSWD values ranging between 0.1 and 3.1 (Table 1). Weighted mean Ca/K and Cl/K values, derived respectively from 37Ar/39Ar and 38Ar/39Ar, are in accord with the electron-microprobe data (suppE Table 1). The anomalously old apparent 40Ar/39Ar ages in the lowtemperature steps of micas from samples 1A and 1K correlate with high Cl/K ratios (Figs. 2a and d), which indicate the presence of 40ArE in minor alteration phases (Section 3.2). However, these phases have no bearing on the ages of the higher-temperature steps (i.e. plateau segments) used to define plateau ages, and as they do not affect the geological significance of the plateau ages, they will not be considered further. Micas from all samples are characterized by flat 40Ar/39Ar age profiles (Fig. 2). The 40Ar/39Ar age distribution from micas in a single sample is greater than the analytical uncertainty for each analysis; however, with the exception of sample 1J, each sample generally gives a mean age that is within 730 Ma (2s) of all analyses (Table 1). In addition, the small phlogopite crystals of all four samples give 40Ar/39Ar ages that are within error of the plateau ages of the large crystals and therefore indicates that there is no relation between the 40Ar/39Ar plateau age and grain size (suppA Fig. 3). More fundamentally, there is a significant difference in the plateau ages between 1A and 1K (ca. 1140–1160 Ma) and 1B and 1J (ca. 1020–1050 Ma) which cannot be easily explained from a detailed interpretation of the geochronological and petrological data alone. Thus, to explain these age variations, the samples were further characterized using various analytical techniques discussed below. 3.2. Scanning electron microscopy (SEM) The Ca/K and Cl/K ratios suggest that the mica from samples 1A and 1K contains another minor phase that degasses during the low-temperature steps, and contains high Ca and Cl (giving elevated 40Ar/39Ar ages that have no geological significance). Consequently, SEM was used to search for any traces of such material. Backscattered electron (Fig. 3 and suppA Fig. 4) images show that these micas are, for the most part, pristine. The only exception found was that parts of grain edges rarely show a thin (1 mm) layer of a phase that is low in K (suppA Fig. 4), which is likely a smectite-group mineral and a product of local, late-stage alteration.

258

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

Table 1 Summary of 40Ar/39Ar age data. Note: MSWD’s greater than unity indicate either an underestimation of the analytical errors, or the presence of non-analytical scatter. 1A

Integrated age (Ma)

7(2r) (Ma)

Large grains (900–1250 mm in diameter) Grain 1 1172 Grain 2 1159 Grain 3 1178 Grain 4 1163 Grain 5 1157 Grain 6 1154 Grain 7 1174 Grain 8 1143

Plateau age (Ma)

7 (2r) (Ma)

39 Ar release (%)

MSWD

2 5 2 2 4 3 2 5

1177 1162 1163 1157 1154 1145 1170 1142

2 5 2 3 4 3 2 6

50.35 89.07 76.79 80.22 90.34 65.87 90.58 84.82

1.07 1.94 0.46 0.98 1.96 1.36 0.22 1.44

13 6 8 Smalll

1171 1178 1164

12 5 8

97.20 99.70 100.00

0.32 0.95 1.19

7(2r) (Ma)

Plateau age (Ma)

7 (2r) (Ma)

39 Ar release (%)

MSWD

Large grains (1100–1500 mm in diameter) Grain 1 1047 Grain 2 1029 Grain 3 1040 Grain 4 1025 Grain 5 1032 Grain 6 1038 Grain 7 1028 Grain 8 1024

2 1 1 2 3 2 2 2

1047 1027 1039 1026 1033 1039 1029 1025

2 1 1 2 3 2 2 2

100.00 88.14 96.20 93.77 94.84 97.91 97.86 91.53

1.40 0.81 0.89 0.49 0.86 0.38 1.17 1.43

(550–1000 mm in diameter) Grain 9 1014 Grain 10 1034 Grain 11 1030

4 2 2

1014 1035 1034

4 2 2

99.86 93.80 71.32

0.74 0.23 1.04

7 8 6 10 10 11 14 Smalll

1017 1074 1047 1029 1036 1070 1073

7 8 5 9 9 11 14

93.40 100.00 98.60 99.30 98.00 100.00 100.00

0.36 1.45

Plateau age (Ma)

7 (2r) (Ma)

39 Ar release (%)

Small Grain Grain Grain

grains (125–150 mm in diameter) 9 1175 10 1178 11 1164 Large

Mean age (Ma) Std dev (Ma)

1159 12

1B

Integrated age (Ma)

Small Grain Grain Grain Grain Grain Grain Grain

grains (125–150 mm in diameter) 12 1020 13 1074 14 1044 15 1030 16 1038 17 1070 18 1073 Large

Mean age (Ma) Std dev (Ma)

1031 9

1J

Integrated age (Ma)

1171 7

1049 23 7(2r) (Ma)

MSWD

Large grains(1100–1450 mm in diameter) Grain 1 1017 Grain 2 1057 Grain 3 997 Grain 4 1052 Grain 5 1035 Grain 6 981 Grain 7 1044 Grain 8 1042 Grain 9 1010

2 3 3 3 3 2 3 2 2

1021 1054 997 1052 1035 981 1044 1042 1011

2 4 3 3 3 2 3 2 2

90.37 82.75 94.99 98.90 99.04 97.53 98.52 98.41 92.70

2.22 3.23 1.02 1.10 1.42 1.11 0.97 1.33 0.47

(700–950 mm in diameter) Grain 10 1044 Grain 11 978 Grain 12 1043 Grain 13 989

2 2 1 3

1045 978 1040 990

2 3 2 1

98.61 96.67 79.09 90.08

1.06 1.79 0.82 2.75

9 5 5 4

1063 1029 1059 1026

8 4 5 4

99.83 97.39 100.00 100.00

Small Grain Grain Grain Grain

grains (125–150 mm in diameter) 14 1062 15 1024 16 1059 17 1026

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

259

Table 1 (continued ) Integrated age (Ma)

7(2r) (Ma)

Plateau age (Ma)

7(2r) (Ma)

39 Ar release (%)

18 19 20 21

1049 1041 1053 1033 Large

5 5 6 8 Smalll

1049 1041 1053 1033

5 5 6 8

100.00 100.00 100.00 100.00

Mean age (Ma) Std dev (Ma)

1022 28

1044 14

1K

Integrated age (Ma)

Plateau age (Ma)

7(2r) (Ma)

39 Ar release (%)

1J

Grain Grain Grain Grain

7(2r) (Ma)

MSWD

MSWD

Large grains(950–1200 mm in diameter) Grain 1 1132 Grain 2 1161 Grain 3 1117 Grain 4 1131 Grain 5 1126

4 2 2 2 2

1146 1160 1115 1130 1126

5 2 2 2 2

83.38 93.09 75.59 96.06 96.72

3.15 1.06 1.24 1.29 1.21

(700–950 mm in diameter) Grain 6 1129 Grain 7 1155 Grain 8 1143 Grain 9 1146

2 3 3 2

1129 1155 1143 1136

2 3 3 3

100.00 99.96 97.16 51.86

0.60 1.32 0.88 0.94

5 5 8 5 Smalll

1144 1155 1186 1134

5 5 8 9

33.95 90.51 81.55 99.25

Small Grain Grain Grain Grain

grains (125–150 mm in diameter) 10 1129 11 1150 12 1178 13 1132 Large

Mean age (Ma) Std dev (Ma)

1034 16

1155 23

3.3. Electron microprobe analysis (EMP) The chemical composition of mica from the four samples was analyzed by electron microprobe (suppE Table 1). Line traverses parallel and perpendicular to (001) in phlogopite from rim to core on individual crystals indicate that the grains are generally homogeneous in major-element composition (Fig. 4). However, the cores of phlogopite from sample 1K are elevated in Ba relative to the rims (range 0.1–0.7 wt% Ba; Fig. 4) and inversely correlated with a corresponding decrease in K concentration (8.4–8.7 wt% K). On the basis of Mg#[Mg2 þ /(Fe2 þ þMg2 þ )], the micas are close to the phlogopite end-member. Micas of samples 1A, 1J, and 1K have the M sites filled with Fe2 þ and Mg2 þ (trioctahedral subgroup; Deer et al., 1992). Mica from sample 1B has significant Tschermak substitution (Al2Mg  1Si  1), which results in M-site vacancies (Foster, 1960). The deficiency of divalent cations (Ca2 þ and Ba2 þ ) at the interlayer sites suggests few interlayer vacancies. As the concentration of OH  cannot be determined by EMP, it was calculated based on stoichiometry. The calculated F#[F  / (F  þCl  þOH  )] indicates that sample 1B differs from the others by having a lower F# (suppE Table 1). 3.4. Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) Isotope-dilution analysis typically does not provide information about the spatial distribution of isotopes or trace elements within a crystal or across grain boundaries. Consequently, 87 Sr/86Sr ratios and trace elements concentrations were measured in profile across single grains using LA-ICPMS. Profiles across

phlogopite grains show that Li, Rb and Sr concentrations are uniform (suppA Fig. 5), as for the major elements. However, Ba varies significantly across the grains (suppA Fig. 5), and contents are highest in the cores of large grains. The 87Sr/86Sr ratios in calcite show that the grain boundary of calcite has the same 87 Sr/86Sr ratio as the core, and despite variations in grain size, there is no intergrain variation (suppA Fig. 6).

3.5. M¨ ossbauer spectroscopy ¨ The Mossbauer spectra of the mica samples are shown in suppA Fig. 7. They consist of two strong absorption peaks located at 0.1 mm/s and 2.4 mm/s due to Fe2 þ , a weak absorption at  1.0 mm/s representing the high velocity peak of octahedrally coordinated Fe3 þ (VIFe3 þ ), and a shoulder at  0.4 mm/s (observed only for samples 1A and 1K) due to the high-velocity peak of tetrahedral Fe3 þ (IVFe3 þ ). The low-velocity peaks for VIFe3 þ and IV Fe3 þ are overlapping with that of Fe2 þ , giving rise to the strong peak at   0.1 mm/s. Thus, the spectra of samples 1B and 1J are fit to a quadrupole-splitting-distribution (QSD) model having two generalized sites, one for Fe2 þ (with two overlapping Gaussian components) and the other for VIFe3 þ (with one Gaussian component). For the spectra of samples 1A and 1K, an additional QSD-site is added to the model for IVFe3 þ (with one Gaussian component). To account for absorber-thickness and instrumental broadening, the Lorentzian linewidth of the symmetric elemental doublets of the QSD was allowed to vary during the spectrum-fitting procedure (Rancourt, 1994). The QSD fit parameters of our samples ¨ (suppE Table 2) are in accord with previously reported Mossbauer

260

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

0.010

0.03

1200

1300 1100 1200 0.005

1000

900

0

20

40

60

0.000 100 0.010

80

1200

800

0

20

40

60

80

0.01 100 0.0035

Cl/K

Apparent 40Ar/39Ar Age (Ma)

1100

0.02

1000

1300 1100 0.0025 1200 0.005

1000

800

0.0015

1100

900

0

20

40

60

0.000 100

80

1000

0

20

40

60

80

0.0005 100

Cumulative 39Ar (%) Fig. 2. Representative 40Ar/39Ar release spectra for step-heated single phlogopite grains from marbles in Outcrop 1. Steps shaded in black represent 40Ar/39Ar apparent age and steps shaded in gray represent Cl/K ratios. Thickness of steps represents 2s errors.

1A

Phl

Phl

Ap

1B

Crn Crn

250μm

1K

500μm Crn Phl

1J

Phl Crn

500μm

500μm

Fig. 3. Backscattered electron images of euhedral phlogopite crystals in textural equilibrium with the surrounding calcite. Abbreviations: phlogopite, phl; calcite, crn; apatite, ap.

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

0.5

1A

0.4

20

261

1A

Mg

15

0.3 Na

0.2

K

5

0.1 0.0

10

F Fe

Ba 0 50 100 150 200 250 300 350 400 450

0.5

1B

0.4

0

0 50 100 150 200 250 300 350 400 450

20

1B

15

Mg

0.3 10

0.2

Fe 5

Concentration (Wt%)

0.1 0.0

K

Ba Na 0

200

400

600

800

0.5

1000 1J

0.4

0

F 0

200

400

600

800

1000

20

1J

Mg

15

0.3 Na

0.2 0.1 0

Ba 0

200

400

600

800

1.0

1000 1K

0.8

10

K

5

F

0

Fe 0

200

20

400

600

800

1000 1K

Mg

15

0.6 10

0.4

Na

0 rim

F

5

0.2 0.0

K

Fe

Ba 100 200 300 400 500 600 700 rim

0

0 rim

100 200 300 400 500 600 700 rim

Distance (µm) Fig. 4. Electron microprobe profiles for Ba, Na, Mg, Fe, K and F (wt%) across representative phlogopite crystals. Error bars are 7 2s.

data for trioctahedral micas (Rancourt and Ping, 1991). The ¨ Mossbauer parameters are given in suppE Table 2. Assuming equal recoil-free fractions for Fe2 þ and Fe3 þ , the ¨ Fe3 þ ( ¼ VIFe3 þ þ IVFe3 þ ) Mossbauer areas for samples 1A, 1B, 1J, and 1K translate to Fe3 þ /(Fe2 þ þFe3 þ ) ratios of 0.1470.04, 0.1670.01, 0.1070.01, and 0.2570.04, respectively (suppE Table 2). These values were used to recalculate FeO and Fe2O3 wt% from (FeO)tot wt% in the EMP data (suppE Table 1). The amounts of Fe3 þ (pfu) in the tetrahedral and octahedral sites ¨ were calculated by multiplying the Mossbauer relative areas of IV Fe3 þ and VIFe3 þ , respectively, by the total amount of Fe (pfu) obtained by EMP (suppE Table 1). 3.6. Fourier transform infrared spectroscopy analysis (FTIR) Fourier-transform infrared analysis (FTIR) was also done on phlogopite from all four samples to provide information on the

vacancy distribution. In mica-group minerals, the vibrational frequency of the hydroxyl group (OH) is determined by the M cations to which the (OH) group is bonded (Farmer et al., 1971). Hydroxyl absorption over the range of 3750–3540 cm  1 reflects the configuration of the three octahedral sites within a mica unitcell (Rancourt, 1994). In the 3750–3640 cm  1 absorption band (light-gray area in suppA Fig. 7), the OH is coordinated to all three M cations; in the 3640–3540 cm  1 band (dark-gray area in suppA Fig. 8), the OH is coordinated to two M cations, with one M site vacant. Phlogopite from samples 1A and 1K have two peaks in the range 3750–3640 cm  1, one intense peak at 3710 cm  1 and a weak peak at 3667 cm  1. Phlogopite from sample 1B has three intense peaks, one at 3704 cm  1 and the remaining two peaks at 3618 cm  1 and 3591 cm  1. The intense peak at 3710 cm  1 of 1A and 1K indicates that the (OH) group is coordinated by three divalent M cations (expressed

262

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

as M2 þ M2 þ M2 þ , where M2 þ represents Mg2 þ or Fe2 þ ). The lack of any significant peaks in the range 3640–3540 cm  1 of samples (1A, 1J, 1K) suggests that there is essentially no Al3 þ or Fe3 þ at the M sites (Farmer et al., 1971), which is in accord with the Fe3 þ ¨ content (0.01 apfu) obtained by combining Mossbauer and EMP data. The one intense peak in 1B at 3704 cm  1 indicates that the M sites are dominated by divalent cations. Because peaks shift to lower frequencies (lower wave numbers) with increasing octahedral Fe2 þ content (Farmer et al., 1971), 1B should have more Fe2 þ than 1A (3710 cm  1 of 1A vs. 3704 cm  1 of 1B), in accord with ¨ the combined EMP and Mossbauer data (suppE Tables 1 and 2). The two peaks at 3618 cm  1 and 3591 cm  1 in 1B suggest that the M sites are occupied by M2 þ Al3 þ and one vacant site. The shoulder/peak in the 3560–3540 cm  1 region can be assigned to the octahedral configuration M2 þ Fe3 þ and vacancy (Farmer et al., 1971; suppE Table 2). The FTIR spectra of phlogopites from sample 1J most resemble those from samples 1A and 1K. In addition, peaks observed in the 3640–3540 cm  1 range for 1B spectra are also seen in the spectra of some grains of sample 1J, but to a lesser extent. The weak 3620 and 3590 cm  1 FTIR peaks observed in spectra of sample 1J (octahedral configuration M2 þ Al and vacancy) may suggest the presence of a small amount of octahedral Al in this sample.

3.7. Crystal-structure refinement Single-crystal X-ray diffraction (XRD) experiments and corresponding structural refinements were obtained from phlogopite from all four samples, giving unit-cell parameters (suppE Table 3), and atom coordinates and site populations (suppE Table 3). The results are similar with crystallographic data of trioctahedral micas reported in the literature (Farmer et al., 1971). The substitution of Fe3 þ for Fe2 þ at both the tetrahedral (T) and octahedral (M) sites is very minor, as indicated also by the ¨ Mossbauer and FTIR results. The calculated ionic porosity of the interlayer site (Zi; Dahl, 1996) of the four micas differs by only 0.24% (suppE Table 3).

3.8. U/Pb analysis U–Pb analysis of zircon from sample 1B was obtained by SHRIMP to determine whether the older mica 40Ar/39Ar plateau ages are older than the age of metamorphism. Zircon grains are typically ameboid to subround. Cathodoluminescence imaging shows that the majority of zircon grains are sector zoned or recrystallized (suppA Fig. 9). Uranium contents range between 30 and 189 ppm, with a narrow range of Th/U ratios of 0.13–0.36 (although, one analysis gave a Th/U ratio of 0.62; suppE Table 4). All 20 zircon U–Pb analyses combine to give a single concordant population with a 207Pb/206Pb age of 116678 Ma (MSWD¼0.26; suppA Fig. 9). U–Pb analyses of apatite from five marble samples were also obtained by ID-TIMS to determine whether the outcrop experienced thermal effects associated with emplacement of the nearby 1070 Ma Tichborne pluton (Skootamatta suite). The apatite samples consist of medium-size ( 40.2 cm) clear to pale blue, euhedral crystals in a calcite matrix. U/Pb data from 4 to 5 grains (suppE Table 5) show relatively low concentrations of U (  10 ppm) and moderate levels of radiogenicity due to high concentrations of common Pb. The 206Pb/238U ages of the five apatite samples range  1000–1030 Ma, suggesting that the outcrop was thermally perturbed by the intrusion.

3.9.

87

Rb/86Sr analysis

Trioctrahedral micas are commonly considered to have similar Ar and Sr closure temperatures (Armstrong et al., 1966). Consequently, 87Rb/86Sr analyses of minerals were done on the four marble samples (suppE Table 6) to determine whether there is any consistency in the ages between the two isotopic systems. The calcites have negligible Rb ( o0.08 ppm), high Sr (47000 ppm), and 87Sr/86Sr ratios of 0.704, which are lower than other Grenvillian-aged marbles (0.705; Shields and Veizer, 2002), indicating that the marbles may not be sedimentary in origin. Phlogopite has high 87Rb/86Sr ratios ( 50–150) and 87 Sr/86Sr ratios that are quite variable  1.48–4.92. The Rb/Sr isochron age of each sample has been determined by pairing calcite and phlogopite, except for 1A, which also includes diopside. Sample 1B yields the youngest 87Rb/86Sr age of 99075 Ma, and the ages of samples 1A, 1J and 1K are slightly older at  1030 Ma,  1025 Ma and  1040 Ma, respectively (suppE Table 6). 3.10. Stable-isotope analysis Small portions of calcite grains from samples collected along a traverse perpendicular to layers 1J and 1K were analyzed for C and O isotopes to determine whether the system remained closed after peak metamorphism at  1170 Ma. The d18O values of  17% indicate that the carbonate units were originally sedimentary rocks (Hoefs, 1975) and despite the different mineral assemblages the narrow but constant range (Fig. 5) suggests that the O isotopic composition of the rocks was equilibrated. Differences in d13C values between layers 1J (  2.2 to  2.0) 1K and (  1.8 to  0.5) probably reflect original isotopic variations, and are also consistent with minimal transfer of fluids or components across the contact zone. The d13C variations near the contact of the two layers (Fig. 5) are symptomatic of a diffusion profile and may reflect syn or post-metamorphism transfer of carbon in a dry system where the water/rock ratio is low (i.e. rock dominated) as high quantities of fluid are required to homogenize carbon (Labotka et al., 2004). These variations between the carbon and oxygen isotopic systems suggest that the rocks have not been affected by subsequent low-temperature hydrothermal events (Labotka et al., 2004). Phlogopite dD values across all four samples range from  70% to 68% and d18OSMOW range from þ 12.5 to þ 14.4% (suppE Table 7). 3.11. Transmission electron microscope characterization of phlogopite Ion-milled samples of 1A, 1B, 1J and 1K were prepared for TEM investigation. In TEM, only a few twins were recognized and some grains showed no twinning at all. In general, the grains also contained very few dislocations, consistent with the slight deformation of a few grains observed using polarized light microscopy. However, at least one grain in TEM showed some zones near one end with arrays of dislocations marking the beginning of subgrain wall formation. All grains showed prominent planar features parallel to (001) running without terminations (Fig. 6). These features have effects that can also be recognized both in electron diffraction patterns, some of which show streaking parallel to cn (Fig. 7), and in images with (00l) banding appearing when diffraction contrast was strong. However, it is clear from the diffraction patterns and bright-field images that the four specimens examined have some distinctive characteristics with respect to these planar stacking features. In particular, grains from 1A and 1K have a significantly

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

18.0

δ18O (SMOW)

17.5

1A

1J

Contact zone

1K

17.0

263

1B

200 μm

200 μm 1K

1J

16.5

16.0 0.0 1J

Contact zone

1K -0.5

200 μm

200 μm

Fig. 6. Bright-field TEM images showing higher density of planar defects in phlogopites from samples 1A and 1K relative to phlogopites from samples 1B and 1J.

δ13C (PDB)

-1.0

-1.5 000 -2.0 1A -3.0 -15

-10

-5

0 5 Distance (cm)

10

15

20

1B

Fig. 7. TEM diffraction patterns of phlogopite showing poorly defined diffraction patterns in sample 1A relative to 1B.

Fig. 5

higher density of these stacking features than 1B and 1J (Fig. 6). An attempt was made using high-resolution lattice imaging to identify the type of disorder but this was not conclusive. In the crystals with higher density of planar features, there was clear evidence for polytypes, seen both in the (00l) lattice images (Fig. 8) and by closely spaced spots in diffraction patterns along the cn direction which elsewhere, show only reflections spaced by ˚ as expected for 1M phlogopite. 10 A, The streaking in the diffraction patterns from 1A (Fig. 7) and 1K affects only reflections with ka3n; this is fully consistent with rotational disorder, which is commonly observed and has been well-characterized by Ross et al. (1966) and Bell and Wilson (1977), from which it is interpreted that such disorder can arise due either to polytpism or ‘true’ stacking defects. Because the (001) planar features extend across grains which are effectively undeformed, we interpret this to indicate that such stacking disorders formed at the time of crystal growth. Therefore, there are significant structural differences between the individuals in each of the two-mica pairs chosen for study – viz.

20 nm

10 nm

Fig. 8. TEM lattice image of phlogopite showing (a) more higher-order polytypes ˚ circled area) in sample 1A (view along [010]) (repetition in [010] every 20 or 30 A; relative to (b) view along [ 7310].

1A is different from 1B, and 1J is different from 1K – persisting from the time of their growth. Significantly, the TEM results show that phlogopites with the younger 40Ar/39Ar apparent ages (1B and 1J) have a lower density

264

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

of planar defects (number of dislocations seen in an image of 2.0 and 2.8, respectively) than phlogopites with the older 40 Ar/39Ar apparent ages (1A and 1K; number of dislocations seen in an image of 7.7 and 7.3, respectively). 3.12. Diffusion experiments Diffusion experiments were subsequently designed to investigate the effect of the density of these planar defects on Ar diffusion. Micas from the four samples were placed inside open platinum capsules and heated in an internally-heated gas vessel at 350 1C for 194–200 h in a 2 kbar Ar atmosphere. In addition, mica from samples 1A and 1B were heated for 341 h at 400 1C in an Ar atmosphere at 2 kbar pressure. To monitor any signs of chemical reaction associated with dehydration or changes in the redox state of the mica crystals, detailed BSE imaging and ¨ Mossbauer spectroscopy were done on the micas before and after the diffusion experiments. The results show that the crystals were unaffected during the experiments (suppE Table 2; suppA Fig. 7). The micas were then incrementally step-heated in a vacuum and the released Ar was analyzed to determine whether it had an atmospheric or radiogenic signature, or a combination of both. The Ar step-heating results of single crystals from 1A, 1J and 1K that were heated in the diffusion experiments to 350 1C are very similar and show that 40Aratm is released before 40Arn (suppA Fig. 10). These micas release  10% of their cumulative 40Arn in the temperature range 900–1150 1C, with the remainder released between 1150 and 1200 1C. Mica from sample 1B has an initial release of 40Aratm at low temperature ( o650 1C), but in the higher-temperature steps both 40Aratm and 40Arn are released simultaneously (suppA Fig. 10). This difference in the pattern of Ar release between mica 1B and the others is thought to be related to the fact that mica 1B has M-site vacancies. Micas that were heated to 400 1C show similar patterns of Ar release as the micas from the lower-temperature diffusion experiment (suppA Fig. 11).

4. Discussion Before discussing the implications of the 40Ar/39Ar data, it is important to discuss Ar loss or gain in both open and closed systems as this has implications as to how Ar can be distributed throughout a geological system. In an open system, the reservoir of Ar surrounding the minerals is either infinitely low (Ar diffuses out) or high (Ar diffuses in). Fluids are generally released through metamorphic reactions during metamorphism or released during the cooling of magmas (Kretz, 2001; Munoz, 1984). Either of these types of fluids may be rich or poor in halogens and noble gases (such as Ar), thus acting as sources or sinks for 40Ar, which can diffuse in or out of mica crystals. In contrast, in a closed system no 40ArE or radiogenic Ar (40Arn produced by radioactive decay) is lost, but can be redistributed amongst the various minerals within the system. During a thermal event, the 40Arn in a closed system is expected to accumulate in the grain-boundary network (Kelley, 2002), effectively becoming 40ArE, until the concentration in the grain boundaries eventually equilibrates with the Ar concentration in the mineral(s); this equilibrium is generally controlled by the most potassic minerals (e.g., Foland, 1979). Consequently, the transport of Ar along grain boundaries in closed, fluid-poor systems is considered to be only a few centimetres over several million years (e.g., Scaillet, 1996). In a closed system, all K-bearing minerals lose 40Arn to the grain boundaries (Kelley, 2002), but coarse-grained minerals may lose only a small proportion of their

40 Arn whereas fine-grained minerals may become almost entirely depleted. As the concentration of Ar in the grain boundary increases, the K-poor minerals will re-equilibrate with this 40ArE. The marble outcrop (O1) from the Grenville Province contains coexisting trioctahedral micas that yield 40Ar/39Ar plateau ages that differ by 4100 Ma, which is surprising considering that the rocks have experienced the same thermal history. The outcrop experienced two thermal events: the first (upper-amphibolite facies metamorphism) event took place in the mid-crust at  1170 Ma (based on the U/Pb zircon age from the marbles) while the second event took place  100 Ma later (based on the U/Pb apatite age from the marbles), associated with the emplacement of the Skootamatta suite plutons at shallow crustal levels (Davidson and van Breemen, 2000). The following geological scenarios consistent with these events can be envisaged: Scenario 1: All phlogopite crystallized during the second thermal event at  1070 Ma. This is the only scenario that involves the presence of 40ArE. Scenario 2: A first generation of phlogopite crystallized during amphibolite-facies metamorphism (Phl 1) at  1170 Ma with the terrane cooling slowly to explain the range of 40Ar/39Ar ages in the phlogopites. Scenario 3: A first generation of phlogopite crystallized during amphibolite-facies metamorphism (Phl 1) with the terrane cooling rapidly to explain the 40Ar/39Ar age of  1170 Ma in some phlogopite. The second thermal event at 1070 Ma is a shortlived spike that caused either complete resetting of some of the phlogopite 40Ar/39Ar ages, or new crystallization of a second generation of phlogopite (Phl 2). Consequently, the principal issue is to determine if the 40 Ar/39Ar ages represent crystallization, cooling, or resetting, or have no geological significance due to 40ArE. Scenario 1 is the most obvious one to consider because it appears to easily explain the old ages in 1A and 1K, which would therefore reflect the incorporation of 40ArE. In this situation, 40ArE is introduced into the system either through external fluids (open system) or from high-K-bearing minerals within the rock volume (closed system). However, in attempting to rationalize this scenario, some problems occur. In an open-system regime, the mica 40 Ar/39Ar age variation between each sample pair (1A–1B and 1J–1K) can be potentially reconciled if fluids with different 40ArE concentrations infiltrated along lithologies with contrasting permeability (e.g., Baxter et al., 2002). Crustal fluid events subsequent to peak metamorphism would readily modify the carbonate composition (chemical and isotopic) along units, cracks and grain boundaries by either diffusive or recrystallization processes (e.g., Lewis et al., 1998). Integration of the oxygen and carbon isotope analysis (Fig. 5), 87Sr/86Sr ratios of the core and rims (suppA Fig. 6) in calcite, and cathodoluminescence imaging (suppB Fig. 1), however, suggests that fluids did not channel along cracks and grain boundaries within the marbles after peak metamorphism. Hence, there is no evidence to suggest either fluids of different compositions or permeability differences between the different lithologies. Problems also arise when considering a closed system because the only K-bearing mineral in the marbles from O1 is phlogopite, which is consequently the only contributor of Ar into the grainboundary network. In other words, there is no opportunity for redistribution of 40Arn in a single-phase system. Another possibility is if calcite was a source of 40ArE. However, it is difficult to explain how it could affect some samples and not others within the same outcrop. Whether dealing with open or closed system behavior, the amount of 40ArE which is incorporated in a mineral is largely controlled by the number of defects in the crystal structure (Lee, 1995; Hodges et al., 1994; Hodges and Bowring, 1995). The common thought is that samples with a higher density

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

of defects (such as 1A and 1K) incorporate more 40ArE and hence, yield older 40Ar/39Ar ages. The problem with this hypothesis for our samples is that the Ar diffusion experiments (Section 3.12) indicate that these micas do not uptake 40ArE more readily than the phlogopites with less defects (1B and 1J). In fact, 1B mica seems to incorporate greater concentrations of Ar than the other micas (Table 2), which would suggest that this sample should have yielded the oldest 40Ar/39Ar age, whereas it does not. In addition, the 40Ar/39Ar ages of 1A and 1K are not older than the age of metamorphism ( 1170 Ma), which seems very fortuitous if there was uptake of 40ArE. Let us now consider what would happen under scenario 2. According to closure-temperature theory (Dodson, 1973), larger grains (with effective diffusion dimensions represented by their diameters in the absence of microstructures) should preserve older ages in slow-cooling terranes than smaller grains, and this has previously been interpreted to have occurred throughout the Grenville province (Cosca et al., 1992, 1991). Harrison et al. (1985) and Wright et al. (1991) suggested that an effective diffusion radius for all micas was 200 mm and should thus be independent of grain sizes larger than this. However, more recent studies have consistently shown that 40Ar/39Ar ages of biotite are strongly dependent on grain size, and there is no crystallographic evidence supporting an effective diffusion dimension smaller than the grain size (e.g., Camacho et al., 2005; Dempster, 1986; Onstott et al., 1989). Moreover, our Ar data show that there is no correlation between 40Ar/39Ar age and grain size (suppA Fig. 3). Therefore, the hypothesis of a slowly cooling terrane from 1170 to 1070 Ma is not supported. In scenario 3, the phlogopite that crystallized at  1170 Ma (Phl 1) should show a range of ages dependent on grain size. In order to preserve the  1160–1170 Ma ages in Phl 1 and because we do not observe a correlation between grain size and 40Ar/39Ar age (as discussed above), the terrane would have had to cool rapidly after amphibolite-facies metamorphism at  1160– 1170 Ma. However, the new generation of mica growth at 1070 Ma (Phl 2) should overprint preexisting fabrics, but there is no field or petrographic evidence to support this interpretation. Thus, at face value, it appears that none of the scenarios above can adequately explain the geochronological data. It is also worthwhile to consider other models which have been postulated to account for varying Ar retentivities in K-bearing minerals. Differences in chemical composition have been suggested to explain why some micas seem to be more retentive (i.e. have higher closure temperatures) than others. For example, Mg-rich micas are considered more retentive than their Fe-rich counterparts (Giletti, 1974; Grove and Harrison, 1996; Harrison et al., 1985; Norwood, 1974). In addition, F and Cl substitution for hydroxyl has been proposed to increase Ar retentivity in micas (Grove and Harrison, 1996). Let us first consider whether the 40Ar/39Ar ages are correlated with Mg#. Phlogopite from sample 1B has the lowest Mg# (  0.79) whereas micas from samples 1A, 1J and 1K have a similar Mg# (  0.92–0.96). Consequently there is no relation between Table 2 Total 40Aratm (moles) incorporated into phlogopite during the 350 1C diffusion experiments. Total 40Ar* (moles) is also shown and is an indicator of the relative size of the crystals analyzed.

1A 1B 1J 1K

SUM 40Aratms (moles)

Sum 40Ar* (moles)

2.00E  11 1.39E  10 8.62E  12 1.96E  12

1.93E  11 1.66E  11 7.79E  12 7.27E  12

265

40 Ar/39Ar age and Mg#. Similarly, no association is observed between halogens and 40Ar/39Ar age (suppE Table 1). Dahl (1996) attempted to elucidate the chemical compositional control on argon mobility in micas by focusing on the ‘‘ionic porosity’’ of the interlayer site (Zi). This model assumes that the large interlayer site, in comparison to the small dimension of the c axis, is the primary site in the mica structure for the atomic transport of argon (i.e. diffusion occurs preferentially parallel to the T–O–T layer rather than perpendicular to this layer). The chemical and X-ray data of our micas were used to calculate the interlayer ionic porosity for each sample. The results show that the calculated Zi values for all our mica samples are very similar with the variation among them being only 0.17% (suppE Table 3). Moreover, the micas from samples 1A and 1K have the lowest and highest Zi values respectively, although they both give the oldest 40 Ar/39Ar ages. If the variation in Ba concentration from core to rim is considered for sample 1K, Zi only varies from 47.25 to  47.50, respectively. This implies that mica from this sample should yield younger ages in the low-temperature steps (degassing from the grain rim) and older ages in the high-temperature steps (degassing from the grain core), which is not observed. In addition, the other samples contain only small variations in concentration of Ba, which do not affect the Zi. A more detailed discussion of the effect of chemical composition and Zi on argon retentivity in micas from O1 will be forthcoming in a subsequent paper. The most significant and consistent difference between the two pairs of micas (1A vs 1B and 1J vs 1K) is highlighted by the TEM data (Fig. 6), which show that a higher density of planar defects is associated with the phlogopites that yields the oldest ages (samples 1A and 1K). At this stage, we can only speculate on why such a correlation might exist. Defects come in a variety of forms and dimensions. Linear defects such as screw and edge dislocations, and volume defects such as micropores, are generally considered to enhance the rate of Ar transport through a solid (e.g., Kelley, 2002; Lee, 1995; Parsons et al., 1999). Essentially, such defects allow Ar to migrate through a crystal at a rate much faster than through the crystal structure itself (i.e., volume diffusion). In the field of materials science, diffusion along dislocations has been well-studied and is commonly termed ‘‘pipe diffusion’’. In addition, planar defects may or may not enhance Ar diffusion, depending on the nature of the planar boundary. For example, the boundary of an exsolution lamellae in K-feldspar is planar, but if the lamellae are coherent, such that the lattice spacings across its boundary are similar enough that any differences can be fully accommodated by elastic lattice strains, it is highly unlikely that such a boundary could serve to enhance Ar diffusion (Parsons et al., 1999). If the boundary is incoherent, however, then it is possible that such a boundary could serve as a path for the rapid transport of diffusing species, and it is generally accepted that grain-boundary diffusion is much faster than volume diffusion (e.g., Joesten, 1983; Farver and Yund, 1991, 1995). In all cases where extensive defects are postulated to affect diffusion rates, it is generally assumed that such defects enhance diffusion. In addition, to be effective diffusive conduits, the defects must connect to the environment ‘‘external’’ to the crystal. Consider the hypothesis that a crystal with a high dislocation density would lose Ar to its surroundings much faster than one with a much lower defect density; this is only true if the dislocations or dislocation network has at least one path that allows the Ar in the defects to escape readily out of the crystal. However, what if the defects are not connected to the external environment, but remain internally contained within the crystal? In this instance, the defects would not serve as pathways for rapid diffusion but might, instead, serve as Ar traps (Arnaud and Kelley,

266

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

1997). In the Helium isotopic system, Shuster et al. (2006) developed a model whereby a void can decrease the ‘‘effective’’ diffusivity of a mineral and the mechanism is summarized as follows. Consider the situation where a gas is preferentially partitioned into the void and becomes locally ‘‘trapped’’ when it diffuses into it. Once in the void, this gas must overcome the energy barrier required to penetrate back into the crystalline matrix. This additional energy is associated with partitioning of the gas between the defects and the solid lattice, and will therefore cause the effective diffusion coefficient for the material to decrease. Although we propose an analogous mechanism for Ar, it is important to emphasize that we are not proposing a ‘‘higher’’ closure temperature for phlogopite. The closure temperature concept only applies to diffusion out of a homogeneous crystal upon cooling, and this is certainly not the situation here. The higher density of defects in 1A and 1K is correlated with older ages and therefore offers a potential explanation for the observed pattern of 40Ar/39Ar ages; if these defects act as Ar traps, then micas with the higher defect density should be more retentive than those with a lower density. The notion of defects serving as Ar traps is rare in the published literature (Kelley, 2002; Arnaud and Kelley, 1997). In our case, the higher density of planar growth defects in samples 1A and 1K is such that they may comprise a significant volume of the crystal. As such, if they do serve as Ar traps, then a significant amount of Ar could potentially be ‘‘stored’’ in these defects and the net result could be that the planar defects make the mica seem to be more retentive than expected. Based on this discussion, consider now a fourth geological scenario, distinct from the three previously discussed. The micas all crystallized at ca. 1150 Ma, although some developed a high density of these planar defects (i.e., 1A and 1K) while others did not (i.e., 1B and 1J). At 1050 Ma, the intrusion of the Skootamatta Suite at high crustal levels generates a regional thermal pulse. For those micas with few planar defects (1B and 1J), they undergo Ar loss as expected, such that their 40Ar/39Ar age is reset to ca. 1050 Ma. For those micas with a higher density of defects (1A and 1K), the 40Arn might simply migrate and become trapped in the defects and thus, retained in the crystal, effectively preserving the crystallization age. Our data are consistent with this scenario. The detailed step-heating 40Ar/36Ar experiments show that most of the 40Arn ( 490%) is released in the temperature range 1150– 1300 1C (suppA Fig. 10) and that the defects are not acting as reservoirs or fast diffusion pathways for extraneous Ar to diffuse into the crystal lattice. Our results are important from the geochronological perspective that this phenomenon (in which defects could serve as Ar traps and increase the apparent Ar retentivity of a crystal) has not previously been recognized as playing a significant role in the interpretation of the ages of rocks. Although we must emphasize that this mechanism cannot be proven at this time, it does seem to offer the most consistent for the age results for trioctahedral micas that we have recorded in this paper. We would encourage much more work in this area.

5. Conclusions Phlogopites from a small marble outcrop in the Frontenac Terrane, Grenville Province, Ontario give 40Ar/39Ar apparent ages that are no older than the age estimate for regional metamorphism yet differ by  100 Ma. This age discrepancy cannot be explained by the effects of grain size, composition, percent vacancy in the interlayer site or regional-thermal histories. The stable-isotope data indicate that the rocks equilibrated in the

presence of fluids with similar isotopic compositions. Consequently, it is difficult to envision how these fluids could have different 40ArE concentrations. The TEM results show that phlogopite with a low density of planar defects give younger 40Ar/39Ar apparent ages than phlogopite with a higher density of planar defects. The Ar diffusion experiments show that for the most part, micas with high and low density of planar defects micas show similar patterns of Ar release—the one exception being a mica that has M-site vacancies. We tentatively consider that the Ar retentivity of a mica may be enhanced by planar defects acting as traps, such that the crystallization age of the mineral can be retained, even after the occurrence of subsequent thermal events.

Acknowledgments AC would like to acknowledge the support of an URGP University of Manitoba grant as well as a University of Manitoba start-up grant. JKWL would like to acknowledge the support of an NSERC Discovery grant as well as an NSERC Major Facilities Access (MFA) grant. We would also like to thank Dr. Doug Archibald for his generous assistance in the Queen’s University Ar laboratory as well as Thomas Ullrich for doing the Ar analysis (University of British Columbia) on the single small crystals, and ¨ Marnie Forster for the use of the Ar laboratory at the ANU. Jorg Hermann is thanked with help with the FTIR. We also appreciate the assistance of Stephen Eggins and Jon Woodhead with the LAICPMS analysis.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.epsl.2012.05.041.

References Abdu, Y.A., Annersten, H., Ericsson, T., Hawthorne, F.C., 2008. High-temperature ¨ cation ordering in olivine: an in situ Mossbauer study of synthetic (Mg0.55Fe0.45)2SiO4. Hyperfine Interact. 186, 99–103. ¨ Armstrong, R.L., Jager, E., Eberhardt, P., 1966. A comparison of K–Ar and Rb–Sr ages on Alpine biotites. Earth Planet. Sci. Lett. 1, 13–19. Arnaud, N.O., Kelley, S.P., 1997. Argon behaviour in gem-quality orthoclase from Madagascar: experiments and some consequences for 40Ar/39Ar geochronology. Geochim. Cosmochim. Acta 61, 3227–3255. Baxter, E.F., DePaolo, D.J., Renne, P.R., 2002. Spatially correlated anomalous 40 Ar/39Ar ‘‘age’’ variations in biotites about a lithologic contact near Simplon Pass, Switzerland: a mechanistic explanation for excess Ar. Geochim. Cosmochim. Acta 66, 1067–1083. Bell, I.A., Wilson, C.J.L., 1977. Growth defects in metamorphic biotite. Phys. Chem. Miner. 2, 153–169. Camacho, A., Lee, J.K.W., Hensen, B.J., Braun, J., 2005. Short-lived orogenic cycles and the eclogitization of cold crust by spasmodic hot fluids. Nature 435, 1191–1196. Carr, S.D., Easton, R.M., Jamieson, R.A., Culshaw, N.G., 2000. Geologic transect across the Grenville orogen of Ontario and New York. Can. J. Earth Sci. 37, 193–216. Cosca, M.A., Essene, E.J., Kunk, M.J., Sutter, J.F., 1992. Differential unroofing within the Central Metasedimentary Belt of the Grenville orogen: constraints from 40 Ar/39Ar thermochronology. Contrib. Mineral. Petrol. 110, 211–225. Cosca, M.A., Sutter, J.F., Essene, E.J., 1991. Cooling and inferred uplift/erosion history of the Grenville orogen, Ontario: constraints from 40Ar/39Ar thermochronology. Tectonics 10, 959–977. Creaser, R.A., Grutter, H.S., Carlson, J., Crawford, B., 2004. Macrocrystal phlogopite Rb–Sr dates for the Ekati property kimberlites: evidence for multiple intrusive episodes during Paleocene and Eocene time. Lithos 76, 399–414. Dahl, P.S., 1996. The crystal-chemical basis for Ar retention in micas: inferences from interlayer partitioning and implications for geochronology. Contrib. Mineral. Petrol. 123, 22–39. Davidson, A., 1986. New interpretations in the southwestern Grenville Province, in: Moore, J.M., Davidson, A., Baer, A.J. (Eds.), The Grenville Province. Geological Association of Canada, Special Paper 31, pp. 61–74. Davidson, A., van Breemen, O., 2000. Late Grenvillian granite plutons in the central metasedimentary belt, Grenville Province, southeastern Ontario. In: Geological

A. Camacho et al. / Earth and Planetary Science Letters 341–344 (2012) 255–267

Survey of Canada, Current Research 2000-F5; Radiogenic Age and Isotopic Studies. Deer, W.A., Howie, R.A., Zussman, J., 1992. An Introduction to the Rock-forming Minerals, second ed. John Wiley, New York. Dempster, T.J., 1986. Isotope systematics in minerals: biotite rejuvenation and exchange during Alpine metamorphism. Earth Planet. Sci. Lett. 78, 355–367. Dodson, M.H., 1973. Closure temperature in cooling geochronological and petrological systems. Contrib. Mineral. Petrol. 40, 259–274. Easton, R.M., 2001. Precambrian Geology, Tichborne area; Ontario Geological Survey, Preliminary map P 3442, scale 1:50 000. Farmer, V.C., Russell, J.D., McHardy, W.J., Newman, A.C.D., Ahlrichs, J.L., Rimsaite, J.Y., 1971. Evidence for loss of protons and octahedral iron from oxidized biotites and vermiculites. Mineral. Mag. 38, 121–137. Farver, J.R., Yund, R.A., 1991. Measurement of oxygen grain boundary diffusion in natural, fine-grained, quartz aggregates. Geochim. Cosmochim. Acta 55, 1597–1607. Farver, J.R., Yund, R.A., 1995. Grain boundary diffusion of oxygen, potassium and calcium in natural and hot-pressed feldspar aggregates. Contrib. Mineral. Petrol. 118, 340–355. Foland, K.A., 1983. 40Ar/39Ar incremental heating plateaus for biotites with excess argon. Chem. Geol. 41, 3–21. Foland, K.A., 1979. Limited mobility of argon in a metamorphic terrain. Geochim. Cosmochim. Acta 43, 793–801. Foster, M.D., 1960. Layer charge relations in the dioctrahedral and trioctrahedral micas. Am. Mineral. 45, 383–398. Fregola, R.A., Scandale, E., 2011. A 94-layer long period polytype. Am. Mineral. 96, 172–178. Giletti, B.J., 1974. Studies in diffusion I. Argon in phlogopite mica. In: Geochemical Transport and Kinetics. Carnegie Institute Washington Publication, pp. 107– 115. Grove, M., Harrison, T.M., 1996. 40Arn diffusion in Fe-rich biotite. Am. Mineral. 81, 940–951. Hanmer, S., Corrigan, D., Pehrsson, S., Nadeau, L., 2000. SW Grenville Province, Canada: the case against post-1.4 Ga accretionary tectonics. Tectonophysics 319, 33–51. Harrison, T.M., Duncan, I., McDougall, I., 1985. Diffusion of 40Ar in biotite: temperature, pressure, and compositional effects. Geochim. Cosmochim. Acta 49, 2461–2468. Hawthorne, F.C., Simmons, W.B., 2010. The crystal structure of zigrasite, MgZr(PO4)2(H2O)4, a heteropolyhedral framework structure. Mineral. Mag. 74, 567–575. Hodges, K.V., Hames, W.E., Bowring, S.A., 1994. 40Ar/39Ar age gradients in micas from a high-temperature–low-pressure metamorphic terrain: evidence for very slow cooling and implications for the interpretation of age spectra. Geology 22, 55–58. Hodges, K.V., Bowring, S.A., 1995. 40Ar/39Ar thermochronology of isotopically zoned micas; insights from the southwestern USA Proterozoic orogen. Geochim. Cosmochim. Acta 59, 3205–3220. Hoefs, J. 1975. Geochemistry of Stable Isotopes. Angewandte Chemie-International Edition in English, vol. 14, pp. 75–79. Hoffman, P.F., 1989. Precambrian geology and tectonic history of North America. In: Bally, A.W., Palmer, A.R. (Eds.), The Geology of North America – An Overview. Geological Society of America, Decade of North American Geology, pp. 447–512. Hoffman, P.F., 1991. Did the breakout of Laurentia turn Gondwanaland inside-out? Science 252, 1409–1411. Joesten, R., 1983. Grain growth and grain-boundary diffusion in quartz from the Christmas Mountains (Texas) contact aureole. Am. J. Sci. 283A, 233–254. Kelley, S.P., 2002. Excess argon in K–Ar and Ar–Ar geochronology. Chem. Geol. (Isot. Geosci. Sect.) 188, 1–22. Kovacs, I., Hermann, J., O’Neill, H.St.C., Hauri, E., 2010. Site-specific infrared O–H absorption coefficients for water substitution into olivine. Am. Mineral. 95, 292–299. Kramar, N., Cosca, M.A., Buffat, P.A., Baumgartner, L.P., 2003. Stacking Faultenhanced Argon Diffusion in Naturally Deformed Muscovite, vol. 220. Geological Society London Special Publications, pp. 249–260. Kretz, R., 2001. Oxygen and carbon isotopic composition of Grenville marble, and an appraisal of equilibrium in the distribution of isotopes between calcite and associated minerals, Otter Lake area, Quebec, Canada. Can. Mineral. 39, 1455–1472.

267

Labotka, T.C., Cole, D.R., Riciputi, L.R., Fayek, M., 2004. Diffusion of C and O in calcite from 0.1 to 200 MPa. Am. Mineral. 89, 799–806. Lee, J.K.W., 1995. Multipath diffusion in geochronology. Contrib. Mineral. Petrol. 120, 60–82. Lewis, S., Holness, M., Graham, C., 1998. An ion microprobe study of marble from Naxos, Greece: grain-scale fluid pathways and isotopic equilibration during metamorphism. Geology 26, 935–938. Ludden, J., Hynes, A., 2000. The lithoprobe Abitibi–Grenville transect: two billion years of crust formation and recycling in the Precambrian Shield of Canada. Can. J. Earth Sci. 37, 459–476. McDougall, I., Harrison, T.M., 1999. Geochronology and Thermochronology by the 40 Ar/39Ar Method, second ed. Oxford University Press, New York. Mezger, K., Essene, E.J., van der Pluijm, B.A., Halliday, A.N., 1993. U–Pb geochronology of the Grenville orogen of Ontario and New York: constraints on ancient crustal tectonics. Contrib. Mineral. Petrol. 114, 13–26. Munoz, J.L., 1984. F–OH and Cl–OH exchange in micas with applications to hydrothermal ore deposits. In: Bailey, S.W. (Ed.), Micas. The Mineralogical Society of America, Washington, DC, pp. 469–494. Noe, D.C., Veblen, D.R., 1999. HRTEM analysis of dislocation cores and stacking faults in naturally deformed biotite crystals. Am. Mineral. 84, 1925–1931. Norwood, C.B., 1974. Radiogenic Argon Diffusion in the Biotite Micas. Brown University, Providence, Rhode Island. Oberti, R., Boiocchi, M., Hawthorne, F.C., Robinson, P., 2010. Crystal structure and crystal chemistry of fluoro-potassic-magnesio-arfvedsonite from Monte Metocha, Xixano region, Mozambique, and discussion of the holotype from Quebec, Canada. Mineral. Mag. 74, 951–960. Onstott, T.C., Hall, C.M., York, D., 1989. 40Ar/39Ar thermochronometry of the Imataca complex, Venezuela. Precambrian Res. 42, 255–291. Parsons, I., Brown, W., Smith, J., 1999. 40Ar/39Ar thermochronology using alkali feldspars: real thermal history or mathematical mirage of microtexture? Contrib. Mineral. Petrol. 136, 92–110. ¨ Rancourt, D.G., 1994. Mossbauer spectroscopy of minerals. I. Inadequacy of Lorentzian-line doublets in fitting spectra arising from quadrupole splitting distributions. Phys. Chem. Miner. 21, 244–249. Rancourt, D.G., Ping, J.Y., 1991. Voigt-based methods for arbitrary-shape static ¨ hyperfine parameter distributions in Mossbauer spectroscopy. Nucl. Instrum. Meth. Phys. Res. B58, 85–97. Ross, M., Takeda, H., Wones, D.R., 1966. Mica polytypes: systematic description and identification. Science 151, 191–193. Sager-Kinsman, E.A., Parrish, R.R., 1993. Geochronology of detrital zircons from the Elzevir and Frontenac terranes, Central Metasedimentary Belt, Grenville Province, Ontario. Can. J. Earth Sci. 30, 465–473. Scaillet, S., 1996. Excess 40Ar transport scale and mechanism in high-pressure phengite: a case study from an eclogitized metabasite of the Dora-Maira nappe, western Alps. Geochim. Cosmochim. Acta 60, 1075–1090. Sharma, A., Jenkins, D.M., 1999. Hydrothermal synthesis of amphiboles along the tremolite–pargasite join and in the ternary system tremolite–pargasite– cummingtonite. Am. Mineral. 84, 1304–1318. Shields, G., Veizer, J., 2002. Precambrian marine carbonate carbonate isotope database: Version 1.1. Geochem. Geophys. Geosyst. 3, 12. Shuster, D.L., Flowers, R.M., Farley, K.A., 2006. The influence of natural radiation damage on helium diffusion kinetics in apatite. Earth Planet. Sci. Lett. 249, 148–161. van Breemen, O., Davidson, A., 1988. U–Pb zircon ages of granites and syenites in the Central Metasedimentary Belt, Grenville Province, Ontario. Geological Survey of Canada Paper 88-2, pp. 45–50. van der Pluijm, B.A., Mezger, K., Cosca, M.A., Essene, E.J., 1994. Determining the significance of high-grade shear zones by using temperature-time paths, with examples from the Grenville orogen. Geology 22, 743–746. Villa, I.M., Puxeddu, M., 1994. Geochronology of the Larderello geothermal field: new data and the ‘‘closure temperature’’ issue. Contrib. Mineral. Petrol. 115, 415–426. Wright, N., Layer, P.W., York, D., 1991. New insights into thermal history from single grain 40Ar/39Ar analysis of biotite. Earth Planet. Sci. Lett. 104, 70–79. Wynne-Edwards, H.R., 1965. Geology of Tichborne (east half) Map-area, Ontario. Geological Survey of Canada Paper 64-56, 5 pp. Xu, H.F., Veblen, D.R., 1995. Periodic and nonperiodic stacking in biotite from the Bingham Canyon porphyry-copper deposits. Clays Clay Miner. 43, 159–173.