Boron isotope systematics of a fossil hydrothermal system from the Troodos ophiolite, Cyprus: Water–rock interactions in the oceanic crust and subseafloor ore deposits

Chemical Geology 396 (2015) 61–73

Contents lists available at ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

Boron isotope systematics of a fossil hydrothermal system from the Troodos ophiolite, Cyprus: Water–rock interactions in the oceanic crust and subseafloor ore deposits Kyoko Yamaoka a,⁎, Seiya Matsukura b,1, Tsuyoshi Ishikawa c, Hodaka Kawahata b a b c

Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi, Tsukuba, Ibaraki 305-8567, Japan Atmosphere and Ocean Research Institute, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8564, Japan Kochi Institute for Core Sample Research, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), 200 Monobe-otsu, Nankoku, Kochi 783-8502, Japan

a r t i c l e

i n f o

Article history: Received 19 June 2014 Received in revised form 22 December 2014 Accepted 23 December 2014 Available online 3 January 2015 Editor: Michael E. Böttcher Keywords: Boron isotope Ophiolite Oceanic crust Ore deposit Hydrothermal alteration

a b s t r a c t We determined concentrations and isotopic compositions of boron in a complete section of the hydrothermally altered Cretaceous oceanic crust of the Troodos ophiolite. The boron content and δ11B value for each lithological section are: pillow lava (3.8–206.8 μg/g, 63 μg/g average; δ11B = +0.17‰ to +15.6‰, +8.1‰ average), sheeted dike complex (0.6–18.0 μg/g, 4.0 μg/g average; δ11B = +3.3‰ to +10.6‰, +6.0‰ average), and plutonic complex (0.3–8.4 μg/g, 1.7 μg/g average; δ11B = − 1.7‰ to + 18.5‰, + 4.5‰ average). These boron contents are higher than the estimated original igneous values throughout the oceanic crust, indicating uptake of boron from seawater and hydrothermal fluid at temperatures ranging from ≤50 °C to N300 °C. Although our boron data for the Troodos ophiolite are generally consistent with those for the Oman ophiolite of similar age, the distinctly low δ11B values of the lower gabbro section in the Troodos ophiolite (b+3‰) suggest reaction with 11 B-depleted fluid at a very small water/rock ratio. The boron content of the bulk oceanic crust (12.3 μg/g) estimated for the Troodos ophiolite is relatively high as a result of strong boron enrichment in the pillow lava section, which underwent prolonged seafloor weathering. Despite these differences, the weighted average δ11B value of the bulk oceanic crust (+ 7.6‰) is similar to that of the Oman ophiolite (+ 7.9‰). We also analyzed the boron isotope geochemistry of a subseafloor hydrothermal stockwork sulfide deposit in the Troodos ophiolite to investigate its formation processes. In contrast to the normal upper oceanic crust, the δ11B values of the rocks below the ore body decrease with increasing depth and have large negative values (−6‰) in the highly altered uppermost dike section. These low δ11B values are coupled with high boron contents (2.5–17 μg/g) and high and uniform 87Sr/86Sr ratios (0.7064 average), and are unlikely to have resulted from interactions with fluids at a small water/rock ratio. These characteristics are better explained by interaction of ore-forming hydrothermal fluids with oceanic crust that had previously been enriched in boron through hydrothermal alteration at low temperatures. These observations demonstrate that boron and boron isotopes are useful for quantitative evaluation of fluid-related processes with multiple stages, including petrogenesis of hydrothermal ore deposits. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Boron is a powerful tracer for understanding various igneous and fluid-related geochemical processes. Because of its high incompatibility in magmatic systems and its high mobility in fluids, boron is strongly concentrated in surface reservoirs (5 to 500 μg/g) compared to the upper mantle (b1 μg/g: Chaussidon and Jambon, 1994). The boron

⁎ Corresponding author. E-mail address: [email protected] (K. Yamaoka). 1 Present address: Fukuhoku Works, Saibu Gas Co., Ltd., 2-9-118 Higashihama, Higashi-ku, Fukuoka 812-0055, Japan.

http://dx.doi.org/10.1016/j.chemgeo.2014.12.023 0009-2541/© 2014 Elsevier B.V. All rights reserved.

isotope ratio (11B/10B) varies widely in natural systems, and each geochemical reservoir has its particular boron isotopic composition: for example, δ11Bseawater = + 39.5‰, δ11Bupper mantle = − 7‰, and δ11Bmean continental crust = − 10.5‰ (Spivack and Edmond, 1987; Chaussidon and Albarède, 1992; Chaussidon and Jambon, 1994). Three major fluid-related processes that affect the oceanic crust are hydration by hydrothermal alteration at mid-ocean ridges, seafloor weathering, and dehydration during subduction at convergent margins. Boron isotope data have been applied to elucidate these processes. The strong boron enrichment and high δ11B generally observed in the upper oceanic crust reflect uptake of seawater-derived boron during seafloor weathering and hydrothermal alteration at various temperatures (e.g., Spivack and Edmond, 1987). Boron enrichment and high δ11B in

62

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

arc magmas are interpreted as the contribution of boron-rich fluids or melts liberated from the subducting oceanic slab to the mantle wedge (e.g., Morris et al., 1990; Palmer, 1991a; Ishikawa and Nakamura, 1994; Tonarini et al., 2001). For quantitative understanding of these fluid-related processes, the spatial distribution of boron content and boron isotopes in the altered oceanic crust is essential information. However, boron data from ocean drilling in the intact oceanic crust is restricted to lava and sheeted dike sections of the upper crust (Spivack and Edmond, 1987; Ishikawa and Nakamura, 1992; Smith et al., 1995). Recently, vertical profiles of boron content and δ11B value in an oceanic crustal section from the uppermost pillow lava to the lowermost gabbro were reported for the Oman ophiolite, which represents Cretaceous oceanic lithosphere formed at a fast-spreading ridge system (Yamaoka et al., 2012). These results showed that the hydrothermally altered gabbro section is considerably enriched in 11B, and the δ11B value estimated for the bulk oceanic crust (+7.9‰) is much higher than a previous estimate (+3.7‰) that was based on composite data from oceanic crust in different settings (Smith et al., 1995). Profiles of boron concentration and δ11B have the potential to vary in different oceanic crusts because the mode of hydrothermal alteration reflects the influences of tectonic structure associated with different spreading rates. Therefore, studies of the boron isotopes in oceanic crustal sections require examples from varied tectonic settings. Seafloor hydrothermal systems also produce sulfide ore deposits. Most massive sulfide deposits form a mound on the seafloor immediately adjacent to the stockwork zone, indicating precipitation of metal from vent fluids emitted from the seafloor. The TAG active hydrothermal mound on the Mid-Atlantic Ridge is an example of such exhalative-type ore deposits (Humphris et al., 1995; Hannington et al., 1998). Cyprus-type massive sulfide deposits, found in many ophiolites, are regarded as fossil examples of ore-forming processes in modern hydrothermal systems. However, some of the largest deposits in ophiolites consist only of stockwork ore, with little or no massive sulfide. These stockwork-type ore deposits are thought to have formed below the seafloor from the subsurface mixing of cold seawater with upwelling hydrothermal fluids (Cann et al., 1987; Herzig and Friedrich, 1987). Difficulties in observing this setting have hampered studies of stockwork-type ore deposits in the modern ocean (Alt et al., 1986; Alt, 1995). Although boron isotopes have been studied in hydrothermal vent fluids (Palmer, 1991b; You et al., 1994; James et al., 1995), no boron isotope data are available for the underlying altered oceanic crust associated with ore solution. Boron isotope studies of ore deposits will improve our knowledge of the ore-forming processes associated with seafloor/subseafloor hydrothermal activity. The Troodos ophiolite is a fragment of oceanic lithosphere that formed during the Cretaceous. Although most ophiolites have critical differences from modern oceanic crust, they offer unique opportunities to investigate a complete sequence of the oceanic crust. In the Cyprus Crustal Study Project in the 1980s, the International Crustal Research Drilling Group recovered several cores from the Troodos ophiolite (Cyprus Crustal Study Project Initial Report, 1987, 1989, 1991). The whole vertical sequence of the oceanic crust, from extrusive rocks to the basal plutonic complex, was recovered from holes CY1 and CY4. Samples from hole CY2A, which penetrated a mineralized stockworktype ore deposit (Agrokipia B) within pillow lava and underlying sheeted dikes, have led to many detailed mineralogical and geochemical studies (e.g., Herzig and Friedrich, 1987; Thy, 1987; Baragar et al., 1989; Bednarz and Schmincke, 1989; Alt, 1994). In this study, we determined concentrations and isotope compositions of boron in rock samples from these drill cores and from well-characterized outcrops in the Troodos ophiolite. The objectives of this paper are (1) to describe the distribution of boron and boron isotopes in the hydrothermally altered Troodos oceanic crust and (2) to elucidate the origin of subseafloor stockwork-type ore deposits from boron isotope systematics.

2. Geological settings The Troodos ophiolite, on the island of Cyprus in the eastern Mediterranean Sea, is a fragment of Cretaceous oceanic lithosphere (Fig. 1). The ophiolite formed around 90–92 Ma according to U–Pb ages from zircon in plagiogranites (Mukasa and Ludden, 1987). Although the various ophiolites in the eastern Mediterranean region formed in different tectonic settings (Robertson, 2002), geochemical studies on fresh volcanic glasses have concluded that the Troodos ophiolite was formed by seafloor spreading above a supra-subduction zone (Robinson et al., 1983; Rautenschlein et al., 1985). Some structural features suggest that the Troodos oceanic crust formed at a slow-spreading center (Varga and Moores, 1985; Dilek et al., 1990), whereas the paleomagnetic studies suggest that it formed at an intermediate- to fast-spreading center (Allerton and Vine, 1987). It preserves a complete sequence of the oceanic crust, consisting of pillow basalts, sheeted dike complex, and gabbroic and ultramafic cumulates, overlying depleted and tectonized mantle harzburgites. On the basis of glass compositions, two distinct lava suites have been recognized on the northern flank of the Troodos: lower tholeiite lavas (high-Ti series) and upper depleted tholeiite lavas (low-Ti series) (Robinson et al., 1983; Schmincke et al., 1983). The plutonic rocks were formed from multiple magma chambers (Malpas et al., 1989). Although it is suggested that low-Ti lower gabbroic and websteritic cumulates formed from an intrusive magma chamber during an off-axis event, the relations between magma systems of cumulate and volcanic rocks are still unclear (Baragar et al., 1989; Browning et al., 1989; Thy et al., 1989). The sampling area of this study was not affected by late-stage boninitic magmatic activity that occurs along the east–west trending Arakapas fault zone in southern Cyprus, which is interpreted as a fossil transform fault (Simonian and Gass, 1978; Flower and Levine, 1987). Emplacement of the ophiolite began with a 90° anticlockwise rotation between Maastrichtian and early Eocene time, and episodic uplift has continued from the early Miocene to the present (Robertson and Woodcock, 1979; Moores et al., 1984). Emplacement-related metamorphism is limited to late-stage filling of fractures and faults (Gillis and Robinson, 1990). 3. Samples The samples used in this study are core samples from holes CY1 (35°02′54″N, 33°10′46″E, 475 m core length), CY2A (35°02′40″N, 33°08′55″E, 689 m core length), and CY4 (34°54′06″N, 33°05′38″E, 2263 m core length) as well as samples from well-characterized outcrops along the Akaki River (Fig. 1). Samples from holes CY1 and CY4 and the Akaki River represent oceanic crust that underwent typical hydrothermal alteration near the spreading axis, whereas samples from hole CY2A represent a mineralized stockwork sulfide deposit (Agrokipia B) developed in the lower pillow lava section and the underlying hydrothermal upflow zone. Hole CY1 penetrated an upper series of low-Ti pillow lava (Fig. 1). It is divided into two of the alteration zones defined by Gillis and Robinson (1988): (1) the Seafloor Weathering Zone (SWZ, 0–275 m) characterized by pervasive alteration (≤50 °C, high water/rock ratio N 100) and abundant smectite, calcite, Fe-hydroxides, and K-feldspar, and (2) the LowTemperature Zone (LTZ, 275–475 m) characterized by large variations in the pervasiveness of alteration (≤100 °C, water/rock ratio N20) and abundant smectite, celadonite, zeolite, and calcite. The development of secondary minerals tends to be controlled by the porosity of the rock (Gillis and Robinson, 1991). Hole CY4 penetrated the lower sheeted dike complex (0–675 m), the upper gabbro and sporadic dikes (675–1330 m), and the layered lower gabbro and ultramafic (websteritic) cumulates (1330–2263 m: Fig. 1). Geophysical data suggest that the petrologic Moho is about 500 m deeper than the bottom of CY-4 (Robinson, 1989). The dikes consist of interfingering low-Ti and high-Ti series rocks altered to greenschistfacies mineral assemblages containing albite, chlorite, actinolite, quartz,

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

Sediment Pillow lava Sheeted dike complex 35o05’

63

Mediterranean Sea Nicosia

Larnaca Paphos

CY2A (Agrokipia B) CY1 Agrokipia A

Limassol

Ak ak iR

ive r

N

35o00’

CY4 34o55’

0

5

10 km

33o10’

33o20’

Subseafloor ore deposit

Oceanic crust 0

sediment pillow lava

33o30’

CY1 (475 m)

CY2A (689 m)

Akaki River

Depth (km)

1

2

sheeted dike complex

plagiogranite intrusive contact

CY4 (2263 m)

gabbro 3

4

layered ultramafic Moho mantle peridotite

Fig. 1. Simplified geologic map of the Troodos ophiolite showing locations of drill holes CY-1, CY-2A, and CY-4 and the Akaki River (modified from Bednarz and Schmincke, 1989), and a schematic columnar section of the Troodos ophiolite. Inset shows location of geologic map in the island of Cyprus with the Troodos ophiolite in gray.

and minor epidote (Baragar et al., 1989). In the plutonic sequence underlying the sheeted dikes, the core crossed a major intrusive contact at about 1330 m depth that separates upper fine- to medium-grained gabbro from lower medium- to coarse-grained gabbroic and websteritic cumulates (Thy, 1987). The plutonic rocks have been variably altered at amphibolite-facies conditions. According to Alt (1994), the uppermost part of the gabbros (650–850 m) is intensively hydrothermally altered and consists of Ca-plagioclase and hornblende with minor amounts of sphene, clinozoisite, epidote, and quartz (70–100% recrystallized), whereas the underlying upper gabbro (850–1330 m) and the lower gabbro and websteritic cumulates (1330–2263 m) are less altered and consist of hornblende, minor sphene, epidote, and clinozoisite (10–70% recrystallized).

The rock samples from the Akaki River consist of hydrothermally altered pillow basalt from the lower part of the extrusive sequence and dolerite from the upper part of the sheeted dikes. Their estimated stratigraphic depth below the crust–sediment interface ranges from 300 to 1000 m (Fig. 1). Hole CY2A penetrated volcanic rocks (0–529 m) and underlying dikes (529–689 m: Fig. 1). The top of the core is about 150 m below the crust–sediment interface. The upper 70 m consists of low-Ti basaltic andesite, below which high-Ti andesite–dacite series rocks are dominant (Bednarz et al., 1987). The core can be divided into three alteration zones, with alteration temperatures estimated from the occurrence of secondary minerals as follows: (1) zone A, pillow lava weakly

64

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

Table 1 Boron and niobium contents and δ11B isotope data for rocks in the Troodos ophiolite. Hole/site

Sample number

Rock type

CY1

Z-1 Z-2 Z-3 Z-4 Z-5 Z-6 Z-6V Z-7 Z-8 Z-9 Z-9B Z-10B Z-304 Z-302A Z-301A Z-302B Z-301B Z-288 Z-287 Z-286 Z-284 Z-282 Z-280 Z-275 Z-226 Z-224 Z-245 Z-52 Z-53 Z-54 Z-55 Z-58 Z-59 Z-61 Z-63 Z-66 Z-73 Z-74 Z-75V Z-76 Z-80 Z-81 Z-86 Z-87 Z-89V Z-93A Z-96 Z-98 Z-100 Z-101 Z-104 Z-105A Z-110 Z-116A Z-116B Z-117V Z-118B Z-121A Z-121B Z-122A Z-122B Z-128 Z-129 Z-130 Z-133A Z-136A Z-136C Z-137B Z-140 Z-142 Z-149 Z-149V Z-150 Z-155

Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Dolerite dike Gabbro Dolerite dike Dolerite dike Dolerite dike Dolerite dike Gabbro Gabbro Gabbro Dolerite dike Plagiogranite Zeolite vein Plagiogranite Gabbro Plagiogranite Plagiogranite Gabbro Gabbro Plagiogranite Gabbro Gabbro Plagiogranite Gabbro Plagiogranite Gabbro Gabbro Gabbro Zeolite vein Gabbro Gabbro

Akaki River

CY4

Core depth (m) 13.05 58.12 89.89 157.42 190.50 250.80 250.80 282.80 361.60 376.79 376.79 471.80

16.50 24.90 34.80 46.80 83.65 102.69 121.87 148.86 187.65 298.40 308.26 313.38 336.54 383.00 401.60 485.91 486.76 529.49 620.22 679.06 729.10 773.23 779.37 833.35 849.36 864.00 917.55 917.55 925.42 933.90 959.70 959.70 970.16 970.16 1013.64 1021.37 1027.10 1045.48 1102.10 1102.10 1107.04 1147.44 1159.32 1255.60 1255.60 1296.87 1386.03

δ11B (‰)

Stratigraphic depth below pillow–sediment interface (m)

Nb (μg/g)

B (μg/g)

13.05 58.12 89.89 157.42 190.50 250.80 250.80 282.80 361.60 376.79 376.79 471.80 300 400 400 450 450 500 550 600 700 750 800 900 950 950 1000 1016.50 1024.90 1034.80 1046.80 1083.65 1102.69 1121.87 1148.86 1187.65 1298.40 1308.26 1313.38 1336.54 1383.00 1401.60 1485.91 1486.76 1529.49 1620.22 1679.06 1729.10 1773.23 1779.37 1833.35 1849.36 1864.00 1917.55 1917.55 1925.42 1933.90 1959.70 1959.70 1970.16 1970.16 2013.64 2021.37 2027.10 2045.48 2102.10 2102.10 2107.04 2147.44 2159.32 2255.60 2255.60 2296.87 2386.03

1.44 0.895 0.812 0.843 0.725 0.653 0.689 0.640 0.809 0.791 0.781 0.762 0.896

118 120 79.7 207 47.3 71.2 42.3 33.3 59.5 4.92 3.83 28.7 5.73

0.791 1.34

7.86 4.88

1.12

9.10

1.16 0.553 2.91 2.19 1.80 0.678 1.21 1.15 0.948 0.215 1.77 2.18 0.932 2.05 0.898 0.882 2.12 0.333 1.67 1.76 1.30 1.23 1.27 0.104

7.04 7.72 18.0 1.65 3.63 1.36 1.17 2.05 4.36 0.596 11.1 2.19 3.54 2.26 4.49 2.92 2.19 1.75 2.86 1.45 4.97 1.42 1.52 0.822

0.299

1.89

6.1

0.989 1.34

2.01 2.93

10.6 9.5

1.38

8.40

4.5

1.24 1.88

1.12 9.84

1.00 0.097 0.536

2.92 2.28 5.41

18.5

1.68 0.105

3.13 0.842

3.3

1.00

2.24

2.05 0.069 0.088 0.696 0.183 0.026 0.192

3.27 0.332 1.11 1.09 29.0 0.296 0.585

0.2 6.8 3.5

14.9 7.7 15.6

4.3

8.2

4.7

7.3

3.7

3.4

9.2

6.3

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

65

Table 1 (continued) Hole/site

CY2A

Sample number

Rock type

Core depth (m)

Stratigraphic depth below pillow–sediment interface (m)

Z-157 Z-159A Z-160 Z-166 Z-171 Z-172 Z-174 Z-176A Z-180 Z-182 Z-183 Z-184 Z-187 Z-190 Z-191A Z-191B Z-192 Z-11 Z-14 Z-15 Z-17 Z-20 Z-22 Z-23 Z-23B Z-26A Z-26B Z-27 Z-30 Z-32 Z-35 Z-37 Z-38 Z-40 Z-42 Z-45 Z-47 Z-47B Z-49

Gabbro Gabbro Gabbro Gabbro Gabbro Gabbro Gabbro Gabbro Gabbro Gabbro Gabbro Websterite Websterite Websterite Websterite Websterite Websterite Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Red jasper Sulfide Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Pillow basalt Dolerite dike Dolerite dike Dolerite dike

1429.23 1471.67 1480.69 1582.88 1665.77 1684.06 1703.22 1717.10 1775.18 1861.24 1903.35 1934.04 2046.60 2150.70 2187.65 2187.65 2217.97 10.83 111.20 123.55 151.75 175.30 234.50 255.12 255.12 272.56 272.56 273.45 289.90 305.13 369.94 397.56 407.79 461.39 536.98 606.25 666.37 666.37 680.42

2429.23 2471.67 2480.69 2582.88 2665.77 2684.06 2703.22 2717.10 2775.18 2861.24 2903.35 2934.04 3046.60 3150.70 3187.65 3187.65 3217.97 160.83 261.20 273.55 301.75 325.30 384.50 405.12 405.12 422.56 422.56 423.45 439.90 455.13 519.94 547.56 557.79 611.39 686.98 756.25 816.37 816.37 830.42

hydrothermally altered at N 50 to ~135 °C (0–154 m) characterized by smectite + celadonite + zeolite ± quartz ± calcite; (2) zone B, stockwork sulfide mineralization (154–300 m) containing chlorite + quartz + illite + sulfide as dominant secondary minerals, altered at maximum temperatures of about 250 °C under high water/rock ratios; and (3) zone C, lower greenschist-facies alteration (300–689 m) with the secondary mineral assemblage chlorite + albite + quartz + epidote + sulfide + sphene ± calcite. The absence of actinolite suggests that alteration temperatures at the base of the hole did not markedly exceed 300 °C (Cann et al., 1987; Herzig and Friedrich, 1987). Massive, silicified pyrite ore occurs as intercalations in zone B, whereas mineralization in zone C is mainly restricted to fissures and veins. The sulfide ore mineral assemblage consists predominantly of pyrite, variable amounts of sphalerite, less chalcopyrite, and minor but widespread pyrrhotite (Herzig and Friedrich, 1987). 4. Analytical procedures Rock samples were cut with a rock cutter and crushed into millimeter-sized chips that were then handpicked and cleaned ultrasonically in Milli-Q purified water. The cleaned rock chips were dried at 110 °C and then powdered using an alumina ceramic mill. To minimize contamination of boron from the experimental environment, all chemical operations in the boron analysis were done in a clean room at the Kochi Core Center, and an evaporation apparatus equipped with a boron-free ULPA filter was used for the evaporation of the sample solutions. Boron and niobium contents were measured using a quadrupole inductively coupled plasma mass spectrometer (ELAN DRC II, Perkin Elmer, USA) and procedures detailed by Nagaishi and Ishikawa (2009).

Nb (μg/g)

B (μg/g)

δ11B (‰)

0.113 0.046 0.051 0.099

1.26 0.316 1.08 1.00

2.9

0.066

0.719

-1.7

0.038 0.030 0.027 0.023 0.003 0.003 0.070 1.38 1.24 1.23 0.963 1.01 0.748 0.033 0.013 0.015 0.030 1.32 0.869 1.73 1.39 0.985 2.92 1.11 1.13 0.747 0.974 0.666 1.60

1.08 0.588 0.830 0.835 2.03 1.70 2.21 6.24 25.7 31.7 21.0 14.9 5.09 1.65 2.50 2.68 0.919 8.18 8.77 13.2 7.85 6.56 7.52 17.0 8.03 6.03 3.40 2.47 12.2

2.2

2.4

3.4 3.5 4.8

4.0

3.5 0.6 −1.2 −2.7 −6.4 −5.7

Typically, ~ 30 mg of rock powder was decomposed with a HCl-HFmannitol mixture and evaporated to dryness at 60 °C overnight. The dried sample was then dissolved in an internal standard solution containing 400 ng/ml Be and 10 ng/ml In and Re. The analytical reproducibility of boron and niobium determinations using this technique was better than ±5% RSD and ±4% RSD, respectively. In samples with low levels of boron (0.15 μg/g), the analytical reproducibility was ±8% RSD. The procedures used for the boron isotopic analysis are basically the same as those described by Yamaoka et al. (2012). Briefly, rock powder samples were decomposed with a HCl-HF-mannitol mixture and converted to the chloride form. Boron was separated from other elements using two-step ion-exchange chromatography: first major cations were removed using 0.6 ml of cation-exchange resin (Bio-Rad, AG 50WX12, 200–400 mesh) in H+ form, and then boron was purified using 0.3 ml of anion-exchange resin (Bio-Rad, AG 1-X4, 200–400 mesh) in F− form. Boron was purified twice more using a smaller anionexchange column (30 μl). Boron isotope ratios were then determined by a Cs2BO+ 2 -graphite method with sample preheating (Ishikawa and Nagaishi, 2011) using a thermal ionization mass spectrometer (TRITON, ThermoFinnigan, Germany). Cs2BO+ 2 ions (m/z = 308 and 309) were detected by static multi-collection. Boron isotopic composition is expressed in the δ11B notation relative to NIST SRM 951 boric acid thus: δ11B (‰) = [(11B/10B)sample/(11B/10B)SRM 951 − 1] × 1000. In this study, 19 separate analyses of SRM 951 yielded 11B/10B = 4.05292 ± 0.00042 (2 SD). The determined δ11B of standard JB-3 was +6.69 ± 0.24‰ (2 SD), which is in agreement with previously reported values (+6.5 to +6.72‰: Ishikawa and Nakamura, 1994; Ishikawa and Tera, 1997; Yamaoka et al., 2012). The total procedural blank of boron was 660–1040 pg, which is negligible with regard to the isotopic results.

66

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

5. Results 5.1. Holes CY1 and CY4 and Akaki River outcrop samples The boron abundances of the oceanic crust in the Troodos ophiolite decrease with increasing stratigraphic depth (Table 1 and Fig. 2). Strong enrichment of boron in the pillow lava section (3.8–206.8 μg/g, 63.2 μg/g average) is consistent with the previously reported boron contents of hole CY1 (Bergeron, 1989). The boron contents in the sheeted dikes (0.6–18.0 μg/g, 4.0 μg/g average) were lower than those in the pillow lava, in good agreement with previous studies (Smith et al., 1995). Although the uppermost gabbro was relatively enriched in boron (2.9–8.4 μg/g), most of the gabbro and websterite had low boron contents (0.3–2.2 μg/g) compared to upper sequences. The average boron content of the plutonic section was 1.7 μg/g. The boron contents of plagiogranite samples (1.1–3.3 μg/g) were indistinguishable from those of the adjacent upper gabbro. The zeolite veins were greatly enriched in boron (9.8–29.0 μg/g) relative to their host gabbroic rocks. In contrast to the boron content, the δ11B values of the Troodos ophiolite samples showed no clear trend with stratigraphic depth in the oceanic crust (Fig. 2). The δ11B values were + 0.17‰ to + 15.6‰ for pillow basalt, +3.3‰ to +10.6‰ for sheeted dikes, and −1.7‰ to +18.5‰ for gabbro. This contrasts strikingly with the Oman ophiolite, in which the δ11B values systematically increase with depth. Although the δ11B values of the pillow lava and sheeted dikes from the Troodos ophiolite are similar to those from the Oman ophiolite, gabbro from the lowest 2500 m at Troodos had distinctly low δ11B values ranging from −1.7‰ to +2.9‰. 5.2. Hole CY2A Basalts altered at low temperature (zone A) are relatively enriched in boron (6.2–31.7 μg/g with an average value of 21.1 μg/g; Fig. 3). Mineralized basalt and dolerite that was altered at high temperature (zones B and C) had lower boron contents (0.92–17.0 μg/g) with no clear trend with stratigraphic depth. The δ11B values of basalt from the upper 450 m of hole CY2A ranged from + 2.4‰ to + 4.8‰, whereas those of basalt and dolerite below 450 m showed a distinct decrease with increasing depth (Fig. 3). The highly altered dolerite (Z-49) and the sulfide-enriched part of the dolerite (Z-47) had δ11B values as low as − 6‰. There was no correlation between the boron contents and boron isotopic compositions of rocks from hole CY2A. 6. Discussion 6.1. Igneous boron contents and δ11B values of the Troodos ophiolite rocks Fresh mid-ocean ridge basalt (MORB) has boron contents generally lower than 1 μg/g (Spivack and Edmond, 1987; Chaussidon and Jambon, 1994). Because boron and niobium have similar incompatibilities during the partial melting and fractional crystallization processes, fresh MORB has relatively uniform B/Nb ratios of approximately 0.3 (Ryan and Langmuir, 1993; Ryan et al., 1996). However, geochemical characteristics of Troodos volcanic glasses have been reported to be different from those of MORB and similar to those observed in primitive island arc or back-arc basin settings (Rautenschlein et al., 1985). Unfortunately, the boron contents and δ11B values of fresh Troodos glasses have not been determined. Although a precise estimation is difficult, B/Nb ratios of altered rocks can provide constraints on the original, igneous boron content. During hydrothermal alteration of oceanic crust, seawater-derived boron is incorporated into secondary minerals even at high temperatures (Yamaoka et al., 2012). Because niobium content is not affected by hydrothermal alteration, altered rocks should have higher B/Nb ratios than original fresh rocks (Fig. 4). Assuming that the B/Nb ratio of unaltered Troodos volcanic rock does not exceed the

lowest value observed in altered doleritic dikes (0.8), the average niobium content of altered basalt and dolerite (1.1 ± 0.6 μg/g) signifies an estimated original boron content lower than 0.88 ± 0.48 μg/g. Given the high incompatibility of boron, its abundance should be much lower in cumulate rocks than in extrusive rocks. The initial boron content of unaltered cumulate rocks would appear to be lower than 0.19 ± 0.33 μg/g (1 SD) given the average niobium content (0.24 ± 0.41 μg/g) of gabbro and websterite from hole CY4. We may gain confidence in these results from the fact that boron contents and B/Nb ratios of rocks from the Troodos ophiolite are indistinguishable from those from the Oman ophiolite (Fig. 4). Estimating the original δ11B values of Troodos rocks is more difficult. The δ11B values of oceanic basalt glasses showed a restricted range of − 7.4 to + 0.6‰, with no significant difference between N-MORB, E-MORB, BABB, and OIB (Spivack and Edmond, 1987; Chaussidon and Jambon, 1994). In contrast, the δ11B values of arc lavas show a wide range from − 7 to + 7‰ (Palmer, 1991a; Ishikawa and Nakamura, 1994; Ishikawa and Tera, 1997, 1999; Smith et al., 1997; Ishikawa et al., 2001; Rosner et al., 2003). However, the δ11B values of arc rocks systematically decrease with decreasing B/Nb ratios and become lower than 0‰ at B/Nb b 2 (Ishikawa and Nakamura, 1994; Ishikawa and Tera, 1997; Ishikawa et al., 2001; Rosner et al., 2003). Therefore, if we assume that B/Nb b 0.8 for the unaltered basalt, the δ11B value of the rock would be indistinguishable from MORB. Given these considerations, we assumed that the δ11B values of fresh rocks in the Troodos ophiolite were not very different from −4‰ of average MORB. 6.2. Boron geochemistry of hydrothermally altered oceanic crust Pillow basalts from both the Troodos and Oman ophiolites show elevated boron contents, but the boron enrichment is greater in Troodos than in Oman (Fig. 2). Boron enrichment in pillow basalt reflects the effective uptake of seawater boron into clay minerals such as smectite during low-temperature alteration or seafloor weathering (Thompson and Melson, 1970; Donnelly et al., 1980; Spivack and Edmond, 1987). Although it is suggested that the Tethyan oceanic crust in both ophiolites formed in Late Cretaceous time (Tilton et al., 1981; Mukasa and Ludden, 1987), the cold seawater circulation might vary in intensity with seafloor topography, resulting in longer exposure to seawater and more intense alteration beneath morphological highs (Gillis and Robinson, 1988; Bednarz and Schmincke, 1989). The decrease in boron content from the upper to lower pillow lavas reflects the difference between seafloor weathering by cold seawater and lowtemperature hydrothermal alteration at depth (Bergeron, 1989). Similar distribution patterns have been observed for potassium, barium, and rubidium in holes CY1 and 1A (Gillis and Robinson, 1991). Higher alteration temperatures lead to weaker boron partitioning into altered basalt and lower water/rock ratios lead to a lesser degree of alteration (Ishikawa and Nakamura, 1992). However, boron contents in the lower sheeted dikes exceed the estimated original values, indicating that boron is incorporated into rocks to some extent during hightemperature hydrothermal alteration, even at N 300 °C, as observed in the Oman ophiolite (Yamaoka et al., 2012: Fig. 2). In contrast to the distinctly high δ11B values in the gabbro section of the Oman ophiolite (+ 7.3‰ to + 18.6‰, + 13.1‰ average), the Troodos gabbros have δ11B values (−1.7‰ to +9.5‰, +4.5‰ average) that are lower than or comparable to those of the upper crust, with the exception of one upper gabbro sample (+18.5‰; Z-122B). Boron isotopic fractionation between fluids and minerals depends on temperature and on the proportions of different boron coordinations in the interacting fluid and mineral phases (Palmer and Swihart, 1996; Peacock and Hervig, 1999; Williams et al., 2001; Hervig et al., 2002; Wunder et al., 2005). Boron is predominantly 10B-enriched and tetrahedrally coordinated in most silicate minerals, whereas 11B-enriched trigonally coordinated boron predominates in the fluid phase (Palmer and Swihart, 1996; Schmidt et al., 2005). The temperature-dependent

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

Troodos ophiolite (This study)

(km) 0

67

Oman ophiolite (Yamaoka et al., 2012)

(km) 0

1

1

2

2

3

3

0.1

1

10

B (µg/g)

100 -5

0

10

20

δ11 B (‰)

4

Pillow lava Dolerite dike Gabbro and ultramafic Plagiogranite Data from Bergeron (1989)

5

0.1

1

10

B (µg/g)

100 -5

0

10

20

δ11 B (‰)

Fig. 2. Depth profiles of boron content and boron isotopic composition of oceanic crustal sections in the Troodos ophiolite (holes CY1 and CY4 and Akaki River outcrops, this study) and in the Oman ophiolite (Yamaoka et al., 2012). Vertical lines indicate igneous values.

boron isotopic fractionation between mineral and fluid is approximated by the equation Δ11Bmineral–fluid = −10.69 (1000 / T [K]) + 3.88, based on experimental studies (Wunder et al., 2005). Although the boron isotopic fractionation between basalt and seawater has not been determined, this equation may be adequate because the main fractionation effect takes place between trigonally coordinated boron in fluids to tetrahedrally coordinated boron in minerals. Yamaoka et al. (2012) attributed the ~ 20‰ increase in δ11B values with increasing stratigraphic depth in the Oman ophiolite to a decrease in Δ11Bmineral–fluid with increasing alteration temperatures, arguing on the basis of a negative correlation of δ11B with δ18O values. On the other hand, low δ11B values in the lower gabbro of the Troodos ophiolite (−1.7‰ to +2.9‰) are inconsistent with this scenario, nor can they be explained by retrograde alteration at low-temperatures because the gabbro consists of amphibolite-facies mineral assemblage and has no late lowtemperature minerals such as calcite and zeolite. Alternatively, low δ11B values can be produced in rocks by interaction with 11B-depleted fluids. The δ11B value of fluids should decrease during interactions at small water/rock ratios because the proportion of rock-derived light boron is greater. The O, S, and Sr isotope data also suggest that deep gabbros have weaker seawater signatures in the Troodos ophiolite than in the Oman ophiolite (e.g., Bickle and Teagle, 1992; Alt, 1994; Alt and Teagle, 2000). We evaluated the effect of water/rock ratio with a simple model calculation in which a body of rock interacts with multiple batches of fluid at a very small water/rock mass ratio of 0.01. Each batch of fluid reaches equilibrium with the rock and then is removed from the system. Assuming equilibrium in a closed system, the boron concentration in the rock after interaction with each batch of fluid is   i i Cr ¼ DB C r þ RC f =ðR þ DB Þ;

ð1Þ

where Cir, Cif, and Cr are the boron concentrations of the initial rock, initial fluid, and rock after reaction, respectively; R is the water/rock mass ratio; and DB is the rock/fluid bulk distribution coefficient of

boron (see Appendix A). The boron isotopic composition of the altered rock can likewise be calculated as 11=10

  i 11=10 i i 11=10 i Br ¼ α C r B r þ RC f B f =ðαCr þ RC f Þ;

ð2Þ

where 11/10Bir, 11/10Bif, and 11/10Br are the 11B/10B ratios of the initial rock, initial fluid, and rock after reaction, respectively, and α is the isotopic fractionation factor of boron between rock and fluid. For the initial rock, we assumed the estimated values of fresh Troodos basalt (Cir = 0.88 μg/g and δ11B = − 4‰). For the initial fluid, we assumed the values of modern seawater (Cif = 4.5 μg/g B and δ11B = +39.5‰; Spivack and Edmond, 1987) or the hydrothermal fluid (Cif = 9 μg/g B and δ11B = +25‰) as estimated from the altered dolerite dikes of the Oman ophiolite (Yamaoka et al., 2012). For water–rock interaction at 300 °C, we assumed values of DB = 0.1 (Yamaoka et al., 2012) and α = 0.986 (Wunder et al., 2005). Fig. 5 shows the calculated boron concentrations and δ11B values of rock and fluid as a function of total water/ rock ratio. At total water/rock ratio less than 0.1, fluids can reach boron concentrations higher than 6 μg/g and δ11B values lower than 20‰, and interaction with such fluids can produce the δ11B values observed in the Troodos lower gabbro (Fig. 5). Thus, δ11B values lower than +3‰ in the lower gabbro can be explained by interaction with significantly 11B-depleted fluids produced by water–rock interactions at very small water/rock ratios. 6.3. Boron inventory of oceanic crust in the Troodos and Oman ophiolites Table 2 lists the estimated boron inventory of the oceanic crust in the Troodos ophiolite, of slow- to intermediate-spreading ridge origin, and the Oman ophiolite (Yamaoka et al., 2012) of fast-spreading ridge origin. The boron inventory (the boron abundance in a rock column with a basal area of 1 cm2) was estimated from the thickness of each sequence by assuming a uniform rock density of 2.8 g/cm3. The boron inventory of the pillow lava section is much greater in the Troodos ophiolite (8.8 g/cm2) than in the Oman ophiolite (1.3 g/cm2).

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73 fresh basalt

0

fresh basalt

Cretaceous seawater

Hole CY2A Zone A

UPL

68

Low-grade alteration (>50 - ~135 )

Zone B Silicified stockwork (~220 - ~250 ) LPL

Core depth (m)

200

400 Zone C Lower greenschist facies alteration (~260 - ~300 ) SDC

600

0

20

40 -8

-4

0 11B

B (µg/g)

4

8 0

(‰)

50

Sr (µg/g)

100

0.702

0.704 87Sr

0.706

0.708

/ 86Sr

Fig. 3. Depth profiles of boron content, boron isotopic composition, strontium content, and 87Sr/86Sr ratios of hole CY2A (subseafloor Agrokipia B ore deposit) in the Troodos ophiolite. Strontium data are from Herzig and Friedrich (1987), Rommel and Friedrichsen (1987), and Kawahata and Scott (1990).

The boron inventory of the pillow lava strongly depends on the period of exposure to seawater, as shown in DSDP Holes 417/418 drilled into 110 Myr crust formed at the Mid-Atlantic Ridge (Smith et al., 1995). Boron isotope systematics of hydrothermal vent fluids from the Mid-Atlantic Ridge showed that N50% of seawater boron is taken up at low temperatures on slow-spreading ridges, as opposed to b 10% on fast-spreading ridges in the Pacific (James et al., 1995). Alt et al. (2010) showed that in upper oceanic crust formed at fast spreading rates, alteration halos are less abundant than in basement formed at intermediate spreading rates, reflecting more rapid sealing of basement by sediment at fast spreading rates. If the Troodos ophiolite formed at a slow- to intermediate-spreading ridge, the higher boron inventory of pillow lava may reflect the influence of a difference in spreading rate. It is noteworthy that because the average boron contents of the sheeted dike and plutonic complex sections of the Troodos ophiolite are very close to those from the Oman ophiolite, the smaller boron inventories of the Troodos sheeted dikes (1.6 g/cm2) and plutonic complex (0.62 g/cm2) result directly from their smaller thicknesses. Nevertheless, the total boron inventory of the Troodos ophiolite (11.0 g/cm2) is twice that of the Oman ophiolite (5.2 g/cm2), reflecting the high boron inventory of the pillow lava section. The average δ11B values of the pillow lava (+ 8.1‰) and sheeted dike complex (+6.0‰) of the Troodos ophiolite are similar to those of the Oman ophiolite. Although the average δ11B value of the plutonic complex is notably lower in the Troodos ophiolite (+ 4.5‰) than in

the Oman ophiolite (+ 13.1‰), the thinner plutonic section in the Troodos ophiolite has a smaller effect on the total boron isotope budget relative to the Oman ophiolite. The weighted average of δ11B for the bulk oceanic crust in the Troodos ophiolite (+7.6‰) is indistinguishable from that in the Oman ophiolite (+7.9‰). The altered oceanic crust moves away from the spreading axis and ultimately subducts at convergent margins as a part of oceanic slab. It has been suggested by many workers that boron liberated from downgoing slab through dehydration or melting is incorporated in the mantle source for arc magmas (e.g., Morris et al., 1990; Palmer, 1991a; Ishikawa and Nakamura, 1994). Thus, differences in the distribution of boron and boron isotopes in the subducting oceanic slab may affect the δ11B value of the slab-derived component incorporated into arc magmas. Yamaoka et al. (2012) noted that boron liberated from the gabbro section of the subducting slab may be highly 11B-enriched, given the high δ11B value of hydrothermally altered gabbro (+ 13‰) in the Oman ophiolite. If only the δ11B value of the deep gabbro section is concerned, the δ11B value of the slab-derived component would be much lower in the case of the Troodos ophiolite. Nevertheless, the weighted average δ11B for the bulk oceanic crust in the Troodos ophiolite (+7.6‰) is similar to that in the Oman ophiolite (+7.9‰), and distinctly higher than the previous estimate of + 3.7‰ (Smith et al., 1995). However, data from ophiolites should be applied to the global mid-ocean ridge system with caution because there are fundamental differences between hydrothermal alteration in ophiolites and in oceanic crust (Alt and Teagle, 2000).

1000 crystal fractionation ↑ hydrothermal alteration

6.4. Two-stage hydrothermal alteration in a subseafloor ore deposit

↑

100

B (µg/g)

↑

10 Troodos (this study) Pillow lava Dolerite Gabbro Oman (Yamaoka et al., 2012) Pillow lava Dolerite Gabbro

1 fresh basalt fresh gabbro

0.1 0.1

1

10

100

1000

B/Nb Fig. 4. Diagram of B/Nb ratio vs. boron content for hydrothermally altered rocks from the oceanic crust in the Troodos ophiolite (this study) and the Oman ophiolite (Yamaoka et al., 2012).

Depth profiles of boron content and δ11B values in hole CY2A are clearly distinct from those in hole CY1 and the Akaki River outcrops (Figs. 2 and 3). Boron contents of the upper part of the pillow basalt (lithologically lower pillow basalt) from zone A (low-grade alteration) in hole CY2A (6.2–31.7 μg/g, 21.1 μg/g average) are consistent with those from the lower pillow basalt in hole CY1 (3.8–59.5 μg/g, 22.7 μg/g average). The δ11B values of zone A basalt (+ 2.4‰ and +3.4‰) are also consistent with those of CY1 and Akaki River basalts altered at temperature lower than 150 °C (+ 0.17 to + 7.7‰). Their relatively low 87Sr/86Sr ratios (~0.7042: Fig. 3) are explained by hydrothermal alteration at water/rock ratios of ~2 (Rommel and Friedrichsen, 1987). On the other hand, the boron contents of altered rocks from zone B (silicified stockwork) and zone C (lower greenschist-facies alteration) are much lower than in zone A (0.92–17.0 μg/g, 7.2 μg/g average; Fig. 3).

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

The high and uniform 87Sr/86Sr ratios (~ 0.7059) in zones B and C indicate pervasive recrystallization and equilibrium with upwelling hydrothermal fluids having elevated 87Sr/86Sr ratios (0.7047–0.7059; Bickle and Teagle, 1992). Although the δ11B values of rocks in zone B (+3.5‰ to +4.8‰) are similar to those in zone A, in zone C they clearly decrease with increasing depth. The highly altered dolerite (Z-49) and the sulfide-enriched part of the dolerite (Z-47B) have large negative δ11B values (−5.7‰ and −6.4‰, respectively), as low as the average of igneous δ11B value of − 4‰, even though their boron contents (12 μg/g and 2.5 μg/g) are much higher than the estimated original boron content of b 0.88 μg/g. Such low δ11B values have not previously been reported for hydrothermally altered oceanic crustal rocks. The different distribution patterns of boron and boron isotopes between CY2A and the CY1–Akaki River group suggest that different hydrothermal alteration processes affected the two groups of rocks, which is consistent with mineralogical and S isotope data (Alt, 1994). In a system under chemical equilibrium, low δ11B values of altered rock can be explained by large isotope fractionation during lowtemperature alteration or by equilibrium with low-δ11B fluid formed at very small water/rock ratios. Low-temperature alteration is unlikely in zone C because the rocks have been altered to the lower greenschist facies (~300 °C) and contain no secondary minerals formed at low temperatures. The stockwork sulfide mineralization in zone B is interpreted to have formed below the seafloor, where upwelling hot (250–300 °C) hydrothermal fluids cooled and mixed with cold seawater circulating in the overlying rocks (Herzig and Friedrich, 1987). Alteration at very small water/rock ratios is also unlikely, because the rocks of zone C

10

Boron concentration (µg/g)

8

fluid 6

4

2

rock

0

40

fluid

δ 11B (‰)

30

20

rock 10

0

CY4 lower gabbro 0

0.2

0.4

0.6

0.8

1

Total W/R ratio Fig. 5. Model calculation illustrating boron contents and δ11B values of fluid and rock during water–rock interaction at 300 °C. Solid and dashed lines indicate interactions with hydrothermal fluid (9 μg/g B and δ11B = +25‰) and seawater (4.5 μg/g B and δ11B = +39.5‰), respectively. Boron content of 0.88 μg/g and δ11B value of −4‰ are assumed for initial values of rock. Hatched area represents the range of lower gabbro from hole CY4.

69

are pervasively altered and have relatively high boron contents (2.5–17 μg/g) (Fig. 5). In addition, the high and uniform 87Sr/86Sr ratios of the altered rocks from zones B and C (0.7059 ± 0.0004, 1SD) indicate interaction with seawater-dominated hydrothermal fluids at high water/rock ratios (Bickle and Teagle, 1992). If we assume that the rocks of zones B and C have been at equilibrium with hydrothermal fluid at 300 °C, the δ11B values of the hydrothermal fluid can be estimated to be + 8.4‰ to + 19.6‰, using α = 0.986 (Wunder et al., 2005). Vent fluids from modern hydrothermal systems along mid-ocean ridges have δ11B values ranging from + 24.3‰ to + 32.6‰, and such 11 B-depleted fluids have never been observed except in sedimenthosted hydrothermal sites (Spivack and Edmond, 1987; Palmer, 1991b; James et al., 1995). Therefore, no single instance of hydrothermal alteration is adequate to explain the extremely low δ11B values coupled with high boron content observed in zone C of the Agrokipia B ore deposits. Alternatively, repeated interaction of hydrothermal fluid with weathered oceanic crust can explain these boron isotope characteristics. In modern hydrothermal sites on the Mid-Atlantic Ridge, the low Eu anomaly and high Cs/Rb ratio of the vent fluids suggest that the hydrothermal fluids reacted with a component of weathered basalt (Campbell et al., 1988; Palmer and Edmond, 1989). Although there are no obvious faults around the Agrokipia ore deposit, lava piles in the area display highly fractured structures (Adamides, 1987). It is plausible that the oceanic crust underwent low-temperature weathering before the subseafloor Agrokipia B deposit formed. To examine this hypothesis, we carried out a model calculation on a two-stage process of hydrothermal alteration. In the first stage, the oceanic crust was altered at low temperature and the boron content of the rock increased by uptake of boron from circulating seawater. Taking zone A as representing oceanic crust altered in the first stage, we assigned its boron values (20 μg/g B and δ11B = +3‰) to the altered rocks formed in the first stage. In the second stage, upwelling hydrothermal fluid penetrated the B-enriched altered basalt and interacted with it at high temperature. Alteration products from the first stage should be especially susceptible to interaction with fluids as they are largely distributed in veins and cracks that form conduits for ascending fluids. Using Eqs. (1) and (2), and the same factors and fluid compositions as the single-stage model shown in Fig. 5, we modeled the change of boron concentration and δ11B value as the boron-enriched altered rock interacted with multiple batches of fluid at 200 °C and 300 °C (Fig. 6). In contrast to the singlestage model, the δ11B value of rock decreases with increasing total water/rock ratio because the δ11B value of the fluid decreases with the addition of isotopically light boron derived from the boron-enriched rock. The modeled δ11B value is as low as − 6‰ with a water/rock ratio of 0.2 (at 200 °C) or 0.1 (at 300 °C). The boron concentrations of the rock interacting with fluid at these same water/rock ratios are 11 μg/g and 8 μg/g, respectively, consistent with the average boron content of the rocks from zone C (8.4 μg/g). The high 87Sr/86Sr ratios of rocks from zones B and C would have been acquired during both stages of alteration because strontium isotope exchange is not dependent on temperature. From these considerations, we conclude that the Agrokipia B ore deposit was emplaced within oceanic crust that had been hydrothermally altered at low temperatures, in an earlier stage, by the mixing of circulating seawater and upwelling high-temperature hydrothermal fluids (Fig. 7). We note that water/rock ratios around 0.6 (at 200 °C) and 0.3 (at 300 °C) in the second alteration stage could produce rock δ11B values as low as −16 to −18‰ (Fig. 6). Thus, boron isotope systematics can be a sensitive indicator for hydrothermal alteration that takes place in multiple stages. 7. Conclusions We determined depth profiles of boron content and boron isotopic composition in an oceanic crustal sequence of the Troodos ophiolite,

70

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

Table 2 Boron inventory, average boron content, and δ11B value of the oceanic crust. Troodos ophiolite (this study)

Pillow lava Sheeted dike complex Plutonic complex Total Weighted average a

Oman ophiolite (Yamaoka et al., 2012)

Thickness (km)

B inventory (g/cm2)

[B] (μg/g)

δ11B (‰)

Thickness (km)

B inventory (g/cm2)

[B] (μg/g)

δ11B (‰)

0.5 1.4 1.3 3.2

8.8 1.6 0.62 11.0

63 4.0 1.7

8.1 6.0 4.5a

0.6 1.7 3.0 5.3

1.3 2.5 1.4 5.2

7.9 5.3 1.7

5.5 6.3 13.1

12.3

7.6

3.6

7.9

Except for high δ11B value of the upper gabbro (+18.5‰; Z-122B).

which represents hydrothermally altered oceanic crust that formed in Cretaceous time together with the Oman ophiolite. The boron contents of the pillow lava section in the Troodos ophiolite (3.8–206.8 μg/g, 63.2 μg/g average) are notably higher than those of the Oman ophiolite, reflecting a longer period of seafloor weathering, whereas the average boron contents of the underlying sheeted dikes (4.0 μg/g) and the plutonic complex (1.7 μg/g) are similar in the two ophiolites. Elevated boron contents, even in lower oceanic crust, strongly suggest the incorporation of seawater-derived boron into deep crustal rocks that underwent hydrothermal alteration at high temperatures (N300 °C). Although the distribution of δ11B values in the Troodos oceanic crust is generally consistent with that in the Oman ophiolite, the lower gabbros in Troodos show distinctly low δ11B values (−1.7‰ to +2.9‰), which can be explained by reaction with hydrothermal fluids at very small water/rock ratios (b0.1). Despite some differences in patterns and

processes of hydrothermal alteration, the average δ11B value for the bulk oceanic crust in the Troodos ophiolite (+ 7.6‰) is consistent with that in the Oman ophiolite. We also determined the boron content and boron isotopic composition of rocks from the stockwork-type sulfide deposit (Agrokipia B) in the Troodos ophiolite. In contrast to the δ11B distribution in hydrothermally altered normal upper oceanic crust, the δ11B value in the mineralized zone clearly decreases with increasing depth and reaches unusually low value around − 6‰ despite the boron enrichment at that depth (2.5–17 μg/g). Considering their high and uniform 87Sr/86Sr ratios (0.7064 average), the only plausible mechanism for producing such low δ11B values is the interaction of hydrothermal fluid with oceanic crust that was enriched in boron through previous low-temperature alterations. This study demonstrates that boron abundance and boron isotopic composition of altered rocks, combined with strontium isotopic

200oC

300oC

Boron concentration of rock (µg/g)

20

15

CY2A Zone C

CY2A Zone C

CY2A Zone C

CY2A Zone C

10

5

0 20

δ 11 B of rock (‰)

10

0

-10

-20

-30 0

0.2

0.4

0.6

Total W/R

0.8

1

0

0.2

0.4

0.6

0.8

1

Total W/R

Fig. 6. Model calculation illustrating boron content and δ11B value of rock during water–rock interaction at 200 °C (left) and 300 °C (right). Solid and dashed lines indicate interactions with hydrothermal fluid (9 μg/g B and δ11B = +25‰) and seawater (4.5 μg/g B and δ11B = +39.5‰), respectively. Initial boron content and δ11B of rock (after the first stage of alteration) are 20 μg/g and +3‰, respectively. Hatched zone represents the range of zone C in hole CY2A.

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

71

Stage 1 Agrokipia A ore deposit

Black smoker

Seafloor Pillow lava

Low-T alteration Sheeted dike complex

Seawater circulation

High-T hydrothermal fluid

Stage 2 Agrokipia A ore deposit

Warm fluid Seafloor Zone A

Seawater circulation

Pillow lava Zone B Zone C

Sheeted dike complex

Agrokipia B ore deposit

High-T hydrothermal fluid

Fig. 7. Schematic diagram of two-stage alteration process to form a subseafloor ore deposit (Agrokipia B) near a spreading axis. In the first stage, boron contents and δ11B values of pillow lava and the uppermost sheeted dikes increase as secondary minerals form during low-temperature alteration. In the second stage, boron contents and δ11B values of the rocks decrease by interaction with upwelling high-temperature fluid (zone C) as mixing of high-temperature fluid and circulating seawater results in precipitation of sulfide minerals within the lower pillow lava (zone B). The formation of the exhalative massive sulfide deposit (Agrokipia A) is unspecified.

data, are useful for identifying and quantifying multistage water–rock interactions, including petrogenesis of hydrothermal ore deposits.

Appendix A Assuming chemical equilibrium, the rock/fluid bulk distribution coefficient of boron (DB) is defined by

Acknowledgments The rock samples from Cyprus were obtained in 1986, when H. Kawahata was a postdoctoral research fellow with Prof. S. D. Scott at the University of Toronto. We are grateful to Prof. Scott for providing an opportunity for H. Kawahata to spend several weeks in Cyprus. We also thank Dr. G. Constantinou and Dr. A. Panayiotou of the Geological Survey of Cyprus for providing information on the Troodos ophiolite and giving us access to several ore deposits. Technical support for the ICP-MS and TIMS analyses by J. Matsuoka and K. Nagaishi are highly appreciated. This study was supported by JSPS Grants-in-Aid for Scientific Research 23840056 and 24340127. We also acknowledge the constructive comments by Dr. Jeffrey C. Alt and one anonymous reviewer.

DB ¼ Cr =C f

ðA:1Þ

where Cr and Cf are the boron concentrations in the rock and fluid, respectively. The boron contents of altered rocks shown in Figs. 4 and 6 are based on a mass balance expressed by i

i

Cr þ RC f ¼ C r þ RC

f

ðA:2Þ

where Cir and Cif are the boron concentrations of the rock and fluid before the interaction, respectively, and R is the water/rock mass ratio.

72

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73

From Eqs. (A.1) and (A.2), we obtain the equation for calculating the boron concentration of rock after the interaction with fluid:   i i Cr ¼ DB C r þ RC f =ðR þ DB Þ:

ðA:3Þ

In terms of the boron isotopic composition, the isotopic fractionation factor of boron (α) is defined by α¼

11=10

11=10

Br =

Bf

ðA:4Þ

where 11/10Br and 11/10Bf are the 11B/10B ratios of the rock and fluid, respectively. The boron isotopic compositions of altered rocks shown in Figs. 4 and 6 are based on a mass balance expressed by   11=10 11=10 Br þ RC f B f =ðCr þ RC f Þ Cr     i 11=10 i i 11=10 i i i B r þ RC f B f = C r þ RC f ¼ Cr

ðA:5Þ

where 11/10Bir and 11/10Bif are the 11B/10B ratios of the rock and fluid before the interaction, respectively. From Eqs. (A.2), (A.4), and (A.5), the boron isotopic composition of rock after the interaction with fluid can be calculated by the following equation: 11=10

 11=10  i i i 11=10 i Br ¼ α Cr Br þ RC f B f =ðαCr þ RC f Þ:

ðA:6Þ

In the model, we assumed interaction of rock with multiple batches of fluid, each with a very small, constant water/rock mass ratio R of 0.01. Each of the batches of fluid reaches equilibrium with the rock and then is removed from the system. Total water/rock ratio is given as the sum of the R values for each of the batches of fluid. References Adamides, N.G., 1987. Diverse modes of occurrence of Cyprus sulphide deposits and comparison with recent analogues. In: Robinson, P.T., Gibson, I.L., Panayiotou, A. (Eds.), Cyprus Crustal Study Project: Initial Report, Holes CY-2 and 2a. Geological Survey of Canada Paper 85-29. Geological Survey of Canada, Ottawa, pp. 153–168. Allerton, S., Vine, F.J., 1987. Spreading structure of the Troodos ophiolite, Cyprus: some paleomagnetic constraints. Geology 15, 593–597. Alt, J.C., 1994. A sulfur isotopic profile through the Troodos ophiolite, Cyprus: primary composition and the effects of seawater hydrothermal alteration. Geochim. Cosmochim. Acta 58, 1825–1840. Alt, J.C., 1995. Subseafloor processes in mid-ocean ridge hydrothermal systems. Geophys. Monogr. 91, 85–114. Alt, J.C., Teagle, D.A.H., 2000. Hydrothermal alteration and fluid fluxes in ophiolites and oceanic crust. GSA Special Paper 349, 273–282. Alt, J.C., Honnorez, J., Laverne, C., Emmermann, R., 1986. Hydrothermal alteration of a 1 km section through the upper oceanic crust, deep sea drilling project hole 504B: mineralogy, chemistry, and evolution of seawater-basalt interactions. J. Geophys. Res. 91, 10,309–10,335. Alt, J.C., Laverne, C., Coggon, R.M., Teagle, D.A.H., Banerjee, N.R., Morgan, S., Smith-Duque, C.E., Harris, M., Galli, L., 2010. Subsurface structure of a submarine hydrothermal system in ocean crust formed at the East Pacific Rise, ODP/IODP Site 1256. Geochem. Geophys. Geosyst. 11, Q10010. http://dx.doi.org/10.1029/2010GC003144. Baragar, W.R., Lambert, M.B., Baglow, N., Gibson, I.L., 1989. Sheeted dikes from CY4 and surface sections: Troodos ophiolite. In: Gibson, I.L., Malpas, J., Robinson, P.T., Xenophontos, C. (Eds.), Cyprus Crustal Study Project: Initial Report, Hole CY-4. Geological Survey of Canada Paper 88-9. Geological Survey of Canada, Ottawa, pp. 69–106. Bednarz, U., Schmincke, H.-U., 1989. Mass transfer during sub-seafloor alteration of the upper Troodos crust (Cyprus). Contrib. Mineral. Petrol. 102, 93–101. Bednarz, U., Sunkel, G., Schmincke, H.-U., 1987. The basaltic andesite–andesite and the andesite–dacite series from the ICRDG Drill Holes CY-2 and CY-2a. I. Lithology, petrology and geochemistry. In: Robinson, P.T., Gibson, I.L., Panayiotou, A. (Eds.), Cyprus Crustal Study Project: Initial Report, Holes CY-2 and 2a. Geological Survey of Canada Paper 85-29. Geological Survey of Canada, Ottawa, pp. 183–204. Bergeron, M., 1989. Distribution and behavior of boron in the oceanic crust. Can. J. Earth Sci. 26, 782–790. Bickle, M.J., Teagle, D.A.H., 1992. Strontium alteration in the Troodos ophiolite: implications for fluid fluxes and geochemical transport in mid-ocean ridge hydrothermal systems. Earth Planet. Sci. Lett. 113, 219–237. Browning, P., Roberts, S., Alabaster, T., 1989. Fine scale modal layering and cyclic units in ultramafic cumulates from the CY-4 borehole, Troodos ophiolite: evidence for an open system magma chamber. In: Gibson, I.L., Malpas, J., Robinson, P.T.,

Xenophontos, C. (Eds.), Cyprus Crustal Study Project: Initial Report, Hole CY-4. Geological Survey of Canada Paper 88-9. Geological Survey of Canada, Ottawa, pp. 193–220. Campbell, A.C., et al., 1988. Chemistry of hot springs on the Mid-Atlantic Ridge. Nature 335, 514–519. Cann, J.R., Oakley, P.J., Richards, H.G., Richardson, C.J., 1987. Geochemistry of hydrothermally altered rocks from Cyprus Drill Holes CY-2 and CY-2A compared with other Cyprus stockworks. In: Robinson, P.T., Gibson, I.L., Panayiotou, A. (Eds.), Cyprus Crustal Study Project: Initial Report, Holes CY-2 and 2a. Geological Survey of Canada Paper 85-29. Geological Survey of Canada, Ottawa, pp. 87–102. Chaussidon, M., Albarède, F., 1992. Secular boron isotope in the continental crust: an ion microprobe study. Earth Planet. Sci. Lett. 108, 229–241. Chaussidon, M., Jambon, A., 1994. Boron content and isotopic composition of oceanic basalts: geochemical and cosmochemical implications. Earth Planet. Sci. Lett. 121, 277–291. Dilek, Y., Thy, P., Moores, E.M., Ramsden, T.W., 1990. Tectonic evolution of the Troodos ophiolite within the Tethyan framework. Tectonics 9, 811–823. Donnelly, T.W., Thompson, G., Salisbury, M.H., 1980. The chemistry of altered basalts at Site 417. Deep Sea Drilling Project Leg 51. Init. Rep. DSDP 51–53pp. 1319–1330. Flower, M.F.J., Levine, H.M., 1987. Petrogenesis of a tholeiite–boninite sequence from Ayios Mamas, Troodos ophiolite: evidence for splitting of a volcanic arc? Contrib. Mineral. Petrol. 97, 509–524. Gillis, K.M., Robinson, P.T., 1988. Distribution of alteration zones in the upper oceanic crust. Geology 16, 262–266. Gillis, K.M., Robinson, P.T., 1990. Patterns and processes of alteration in the lavas and dykes of the Troodos ophiolite, Cyprus. J. Geophys. Res. 95, 21523–21548. Gillis, K.M., Robinson, P.T., 1991. Alteration of the ICRDG CY-1 and CY-1A drill cores, Troodos ophiolite: mineralogical and geochemical studies. In: Gibson, I.L., Mapas, J., Robinson, P.T., Xenophontos, C. (Eds.), Cyprus Crustal Study Project: Initial Report, Holes CY-1 and 1a. Geological Survey of Canada Paper 90-20. Geological Survey of Canada, Ottawa, pp. 41–60. Hannington, M.D., Galley, A.G., Herzig, P.M., Petersen, S., 1998. Comparison of the TAG mound and stockwork complex with Cyprus-type massive sulfide deposits. In: Herzig, P.M., Humphris, S.E., Miller, D.J., Zierenberg, R.A. (Eds.), Proc. ODP, Sci. Results vol. 158, pp. 389–415. Hervig, R.L., Moore, G.M., Williams, L.B., Peacock, S.M., Holloway, J.R., Roggensack, K., 2002. Isotopic and elemental partitioning of boron between hydrous fluid and silicate melt. Am. Mineral. 87, 769–774. Herzig, P.M., Friedrich, G.H., 1987. Sulfide mineralization, hydrothermal alteration and chemistry in the drill Hole CY-2A, Agrokipia, Cyprus. In: Robinson, P.T., Gibson, I.L., Panayiotou, A. (Eds.), Cyprus Crustal Study Project: Initial Report, Holes CY-2 and 2a. Geological Survey of Canada Paper 85-29. Geological Survey of Canada, Ottawa, pp. 103–138. Humphris, S.E., et al., 1995. The internal structure of an active sea-floor massive sulphide deposit. Nature 377, 713–716. Ishikawa, T., Nagaishi, K., 2011. High-precision isotopic analysis of boron by positive thermal ionization mass spectrometry with sample preheating. J. Anal. At. Spectrom. 26, 359–365. http://dx.doi.org/10.1039/c0ja00060d. Ishikawa, T., Nakamura, E., 1992. Boron isotope geochemistry of the oceanic crust from DSDP/ODP Hole 504B. Geochim. Cosmochim. Acta 56, 1633–1639. Ishikawa, T., Nakamura, E., 1994. Origin of the slab component in arc lavas from across-arc variation of B and Pb isotopes. Nature 370, 205–208. Ishikawa, T., Tera, F., 1997. Source, composition and distribution of the fluid in the Kurile mantle wedge: constraints from across-arc variations of B/Nb and B isotopes. Earth Planet. Sci. Lett. 152, 123–138. Ishikawa, T., Tera, F., 1999. Two isotopically distinct fluid components involved in the Mariana arc: evidence from Nb/B ratios and B, Sr, Nd, and Pb isotope systematics. Geology 27, 83–86. Ishikawa, T., Tera, F., Nakazawa, T., 2001. Boron isotope and trace element systematics of the three volcanic zones in the Kamchatka arc. Geochim. Cosmochim. Acta 65, 4523–4537. James, R.H., Elderfield, H., Palmer, M.R., 1995. The chemistry of hydrothermal fluids from the Broken Spur site, 29°N Mid-Atlantic Ridge. Geochim. Cosmochim. Acta 59, 651–659. Kawahata, H., Scott, S.D., 1990. Water–rock interaction in pillow lava and sheeted dike complex of Troodos ophiolite, Cyprus — the fossil subseafloor sulfide ore deposit. Geochem. J. 24, 349–356. Malpas, J., Brace, T., Dunsworth, S.M., 1989. Structural and petrologic relationships of the CY-4 drill hole of the Cyprus Crustal Study Group. In: Gibson, I.L., Malpas, J., Robinson, P.T., Xenophontos, C. (Eds.), Cyprus Crustal Study Project: Initial Report, Hole CY-4. Geological Survey of Canada Paper 88-9. Geological Survey of Canada, Ottawa, pp. 39–68. Moores, E.M., Robinson, P.T., Malpas, J., Xenophontos, C., 1984. A model for the origin of the Troodos Massif, Cyprus and other Mideast ophiolites. Geology 12, 500–503. Morris, J.D., Leeman, W.P., Tera, F., 1990. The subducted component in island arc lavas: constraints from Be isotopes and B–Be systematics. Nature 344, 31–36. Mukasa, S.B., Ludden, J.N., 1987. Uranium–lead isotopic ages of plagiogranites from the Troodos ophiolite, Cyprus, and their tectonic significance. Geology 15, 825–828. Nagaishi, K., Ishikawa, T., 2009. A simple method for the precise determination of boron, zirconium, niobium, hafnium and tantalum using ICP-MS and new results for rock reference samples. Geochem. J. 43, 133–141. Palmer, M.R., 1991a. Boron-isotope systematics of Halmahera arc (Indonesia) lavas: evidence for involvement of the subducted slab. Geology 19, 215–217. Palmer, M.R., 1991b. Boron isotope systematics of hydrothermal fluids and tourmalines: a synthesis. Chem. Geol. 94, 111–121. Palmer, M.R., Edmond, J.M., 1989. Cesium and rubidium in submarine hydrothermal fluids: evidence for recycling of alkali elements. Earth Planet. Sci. Lett. 95, 8–14.

K. Yamaoka et al. / Chemical Geology 396 (2015) 61–73 Palmer, M.R., Swihart, G.H., 1996. Boron isotope geochemistry: an overview. In: Grew, E.S., Anovitz, L.M. (Eds.), Boron: Mineralogy, Petrology and Geochemistry vol. 33. Mineral. Soc. Am, Washington, DC, pp. 709–744. Peacock, S.M., Hervig, R.L., 1999. Boron isotopic composition of subduction-zone metamorphic rocks. Chem. Geol. 160, 281–290. Rautenschlein, M., Jenner, G.A., Hertogen, J., Hofmann, A.W., Kerrich, R., Schmincke, H.-U., White, W.M., 1985. Isotopic and trace element composition of volcanic glasses from the Akaki Canyon, Cyprus: implications for the origin of the Troodos ophiolite. Earth Planet. Sci. Lett. 75, 369–383. Robertson, A.H.F., 2002. Overview of the genesis and emplacement of Mesozoic ophiolites in the Eastern Mediterranean Tethyan region. Lithos 65, 1–67. Robertson, A.H.F., Woodcock, N.H., 1979. The Mamonia Complex, S.W. Cyprus: the evolution and emplacement of a Mesozoic continental margin. Geol. Soc. Am. Bull. 90, 651–665. Robinson, P.T., 1989. Hole CY-4 of the Cyprus Crustal Project: background and objectives. In: Gibson, I.L., Malpas, J., Robinson, P.T., Xenophontos, C. (Eds.), Cyprus Crustal Study Project: Initial Report, Hole CY-4. Geological Survey of Canada Paper 88-9. Geological Survey of Canada, Ottawa, pp. 1–4. Robinson, P.T., Melson, W.G., O'Hearn, T., Schmincke, H.-U., 1983. Volcanic glass compositions of the Troodos ophiolite, Cyprus. Geology 11, 400–404. Rommel, U., Friedrichsen, H., 1987. The alteration chemistry of CCSP Hole CY-2a: a stable and radiogenic isotope study. In: Robinson, P.T., Gibson, I.L., Panayiotou, A. (Eds.), Cyprus Crustal Study Project: Initial Report, Holes CY-2 and 2a. Geological Survey of Canada Paper 85-29. Geological Survey of Canada, Ottawa, pp. 169–181. Rosner, M., Erzinger, J., Franz, G., Trumbull, R.B., 2003. Slab-derived boron isotope signatures in arc volcanic rocks from the Central Andes and evidence for boron isotope fractionation during progressive slab dehydration. Geochem. Geophys. Geosyst. 4, 9005. http://dx.doi.org/10.1029/2002GC000438. Ryan, J.G., Langmuir, C.H., 1993. The systematics of boron abundances in young volcanic rocks. Geochim. Cosmochim. Acta 57, 1489–1498. Ryan, J.G., Leeman, W.P., Morris, J.D., Langmuir, C.H., 1996. The boron systematics of intraplate lavas: implications for crust and mantle evolution. Geochim. Cosmochim. Acta 60, 415–422. Schmidt, C., Thomas, R., Heinrich, W., 2005. Boron speciation in aqueous fluids at 22 to 600 °C and 0.1 MPa to 2 GPa. Geochim. Cosmochim. Acta 69, 275–281. Schmincke, H.-U., Rautenschlein, M., Robinson, P.T., Mehegan, J.M., 1983. Troodos extrusive series of Cyprus: a comparison with oceanic crust. Geology 11, 405–409.

73

Simonian, K.O., Gass, I.G., 1978. Arakapas fault belt, Cyprus: a fossil transform fault. Geol. Soc. Am. Bull. 89, 1220–1230. Smith, H.J., Spivack, A.J., Staudigel, H., Hart, S.R., 1995. The boron isotopic composition of altered oceanic crust. Chem. Geol. 126, 119–135. Smith, H.J., Leeman, W.P., Davidson, J., Spivack, A.J., 1997. The B isotopic composition of arc lavas from Martinique, Lesser Antilles. Earth Planet. Sci. Lett. 146, 303–314. Spivack, A.J., Edmond, J.M., 1987. Boron isotope exchange between seawater and the oceanic crust. Geochim. Cosmochim. Acta 51, 1033–1043. Thompson, G., Melson, W.G., 1970. Boron contents of serpentinites and metabasalts in the oceanic crust: implications for the boron cycle in the oceans. Earth Planet. Sci. Lett. 8, 61–65. Thy, P., 1987. Magmas and magma chamber evolution, Troodos ophiolite, Cyprus. Geology 15, 316–319. Thy, P., Schiffman, P., Moores, E.M., 1989. Igneous mineral stratigraphy and chemistry of the Cyprus Crustal Study Project drill core in the plutonic sequences of the Troodos ophiolite. In: Gibson, I.L., Malpas, J., Robinson, P.T., Xenophontos, C. (Eds.), Cyprus Crustal Study Project: Initial Report, Hole CY-4. Geological Survey of Canada Paper 88-9. Geological Survey of Canada, Ottawa, pp. 147–185. Tilton, G.R., Hopson, C.A., Wright, J.E., 1981. Uranium–lead isotopic ages of the Samail ophiolite, Oman, with applications to Tethyan ocean ridge tectonics. J. Geophys. Res. 86, 2763–2775. Tonarini, S., Leeman, W.P., Ferrara, G., 2001. Boron isotopic variations in lavas of the Aeolian volcanic arc, South Italy. J. Volcanol. Geotherm. Res. 110, 115–170. Varga, R.J., Moores, E.M., 1985. Spreading structure of the Troodos ophiolite, Cyprus. Geology 13, 846–850. Williams, L.B., Hervig, R.L., Holloway, J.R., Hutcheon, I., 2001. Boron isotope geochemistry during diagenesis: Part I. Experimental determination of fractionation during illitization of smectite. Geochim. Cosmochim. Acta 65, 1769–1782. Wunder, B., Meixner, A., Romer, R.L., Wirth, R., Heinrich, W., 2005. The geochemical cycle of boron: constraints from boron isotope partitioning experiments between mica and fluid. Lithos 84, 206–216. Yamaoka, K., Ishikawa, T., Matsubaya, O., Ishiyama, D., Nagaishi, K., Hiroyasu, Y., Chiba, H., Kawahata, H., 2012. Boron and oxygen isotope systematics for a complete section of oceanic crustal rocks in the Oman ophiolite. Geochim. Cosmochim. Acta 84, 543–559. You , C.-F., Butterfield, D.A., Spivack, A.J., Gieskes, J.M., Gamo, T., Campbell, A.J., 1994. Boron and halide systematics in submarine hydrothermal systems: effects of phase separation and sedimentary conditions. Earth Planet. Sci. Lett. 123, 227–238.