Lithos 284–285 (2017) 1–10

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r .com/ locate / l i thos

Magmatic recharge buffers the isotopic compositions against crustalcontamination in formation of continental flood basalts

Xun Yu, Li-Hui Chen ⁎, Gang ZengState Key Laboratory for Mineral Deposits Research, School of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China

⁎ Corresponding author.E-mail address: chenlh@nju.edu.cn (L.-H. Chen).

http://dx.doi.org/10.1016/j.lithos.2017.03.0270024-4937/© 2017 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 2 August 2016Accepted 27 March 2017Available online 08 April 2017

Isotopic compositions of continental flood basalts are essential to understand their genesis and to constrain thecharacter of their mantle sources. Because of potential crustal contamination, it needs to be evaluated if and towhich degree these basalts record original isotopic signals of their mantle sources and/or crustal signatures.This study examines the Sr, Nd, Hf, and Pb isotopic compositions of the late Cenozoic Xinchang–Shengzhou(XS) flood basalts, a small-scale continental flood basalt field in eastern China. The basalts show positivecorrelations between 87Sr/86Sr and 143Nd/144Nd, and negative correlations between 143Nd/144Nd and176Hf/177Hf, which deviate from compositional arrays of crustal contamination and instead highlight variationsinmagmatic recharge intensity andmantle source compositions. The lava samples formed by high-volumemag-matic recharge recorded signals of recycled sediments in themantle source,which are characterized bymoderateBa/Th (91.9–106.5), excess 208Pb/204Pb relative to 206Pb/204Pb, and excess 176Hf/177Hf relative to 143Nd/144Nd.Thus, we propose that magmatic recharge buffers the original isotopic compositions of magmas against crustalcontamination. Identifying and utilizing the isotope systematics of continental flood basalts generated by highvolumes of magmatic recharge are thus crucial to trace their mantle sources.

© 2017 Elsevier B.V. All rights reserved.

Keywords:Continental flood basaltsMantle heterogeneityMagmatic rechargeCrustal contamination

1. Introduction

Continental flood basalts are thought to be the volcanic manifesta-tions of large volumes of magma derived from Earth's mantle includingmantle compositionswith variable elemental and isotopic compositions(e.g., Hooper et al., 2007; Sobolev et al., 2009, 2011; Zhang et al., 2006).Understanding the compositional record of continental flood basaltsis thus crucial for a better understanding of mantle compositionalvariations and ultimately the dynamics that lead to these variations.Radiogenic isotope systematics of melts derived from a mantlesource can help provide clues on the formation mechanisms(e.g., asthenosphere versus continental lithosphere, identifying theproperties of the recycled crustal materials in mantle sources) ofcontinental flood basalts (e.g., Chung and Jahn, 1995; Hooper et al.,2007; Sharma et al., 1992; Xiao et al., 2004). However, crustal contami-nation commonly obscures the primary compositions and needs to beconstrained (Carlson et al., 1981; Cox, 1980; Devey and Cox, 1987;Ramos et al., 2005; Silver et al., 2006). It remains uncertain and debatedwhether the radiogenic isotopic compositions of continental floodbasalts record elemental and isotopic heterogeneities of the mantlesource or if this information is lost during their passage through

and/or temporary storage in the crust (Arndt et al., 1993; Brandonet al., 1993; Meshesha and Shinjo, 2007; Yu et al., 2015b).

To recognize effects of crustal contamination during magma evolu-tion, previous studies have examined correlations between SiO2 andincompatible element ratios (e.g., Ba/Th and Ce/Pb) and isotopic ratios(e.g., 87Sr/86Sr and 143Nd/144Nd) (e.g., Arndt and Christensen, 1992;Sharma et al., 1991; Xiao et al., 2004). This approach is based on theprinciple assumption that continental flood basalt compositions areessentially mixtures of two end-members: mantle-derived melts andcontinental crust. Thus, ideally, mixing between these two end-members yields a mixing line indicating crustal contamination, other-wise insignificant crustal contamination is suggested (e.g., Li et al.,2015; Xu et al., 2005). However, mixing or contamination by crustalmaterials in mantle sources can also form geochemical correlations asmentioned above. Besides of source mixing or crustal assimilation andfractional crystallization (AFC), another kind of magmatic processesfor continental flood basalts is magmatic recharge in magma chambers(e.g., Camp and Roobol, 1989; Devey and Cox, 1987; Kumar et al., 2010).Magmatic recharge means continuous or episodic magma replenish-ment in the long-livedmagma chambers, which are situated above a re-gion undergoing relatively continuous magmatic fluxing (e.g., Bohrsonand Spera, 2003; Davidson and Tepley, 1997; Lee et al., 2014; O'Haraand Mathews, 1981; Spera and Bohrson, 2002; Yu et al., 2015b).Although magmatic recharge commonly occurs in magma chambersthat produce flood basalts (Camp and Roobol, 1989; Devey and Cox,

2 X. Yu et al. / Lithos 284–285 (2017) 1–10

1987; Kumar et al., 2010; Song et al., 2006; Yu et al., 2015a, 2015b) andlayered intrusions (e.g., Bohrson and Spera, 2003; Howarth and Prevec,2013; Maydagan et al., 2014; Pang et al., 2009; Wark et al., 2007), therole of magmatic recharge has not been addressed adequately in previ-ous studies. Magmatic recharge represents a plausiblemechanism of di-luting the signal of crustal contamination and resulting in moreprimitive magma and the composition of the mantle-derived rechargemagmamay vary with time (Yu et al., 2015b). Therefore, it is importantto understand possible temporal variations in mantle-derived magmasupply in addition to processes and compositional evolution in the crust.

To quantify magmatic recharge and crustal contamination when anisotopic study is conducted, radiogenic isotopic compositions of Sr, Nd,Hf, and Pb of the Xinchang–Shengzhou (XS) flood basalts, a small-scale continental flood basalt field in Zhejiang Province, southeasternChina (Fig. 1a), were analyzed and evaluated. The XS flood basalts areknown to have been influenced by magma chamber processes(i.e., crustal contamination and magmatic recharge) in the lower conti-nental crust (Yu et al., 2015a). Consequently, the XS flood basalts aresuitable for investigating the impact of magmatic recharge on theisotope systematics of continental flood basalts. Combined with majorand trace elemental compositions (Yu et al., 2015a), the isotopiccompositions of the XSflood basalts give clues to the interplay of crustal‘contamination’, magmatic ‘recharge’, andmantle ‘heterogeneity’ (here-after termed the CRH model) and possible short-term variations. It isproposed that continuous and significant recharge of the basalticmagma chamber can suppress the isotopic variations by crustal

Fig. 1.Distribution of Cenozoic basalts in eastern China (a) and in the Zhejiang Province (b). ThLishui–Yuyao fault (B), and the Zhenhai–Wenzhou fault (C). The studied basalts crop out north aof Yu et al. (2015a). The indicated K–Ar and Ar–Ar ages of the basalts are fromHo et al. (2003) asymbols stand for high-SiO2 basalts sampled from south and north of the Lishui–Yuyao fault, rereferred to the web version of this article.)The whole map is modified from Yu et al. (2015a).

contamination, thereby isotopic compositions of flood basalts with sig-nificant volume recharge can provide constraints on the compositionalvariations of the mantle.

2. Background review

Eastern China is composed, from north to south, of the XingmengBlock, the North China Craton, the Yangtze Craton, and the CathysiaBlock. Cenozoic basalts that have erupted from monogenetic volcanoesor as small-scale flood basalts are widely distributed in eastern China(e.g., Liu et al., 1992). According to the geophysical observations,subducted paleo-Pacific oceanic slab is stagnant at themantle transitionzone covering the distribution area of Cenozoic basalts of eastern China(Huang and Zhao, 2006) (Fig. 1a). Geochemical studies furthersuggested that the formation of these Cenozoic basalts has closerelationship with the stagnant slab which contributed recycled crustand sediment materials (e.g., Li et al., 2015; Sakuyama et al., 2013; Xuet al., 2012; Xu, 2014).

In the Cathysia Block, small-scale flood basalts occur in the coastalareas of Zhejiang, Fujian, Guangdong, and Hainan provinces. Amongthem, XS flood basalts crop out in Zhejiang Province of the northeasternCathysia Block (Fig. 1a). The main body of the XS flood basalts iscomposed of high-SiO2 alkaline/tholeiitic basalts (here we definedbasalts with SiO2 N 47.5 wt.% and positive correlation between SiO2 andNa2O + K2O as high-SiO2 basalts while basalts with SiO2 b 47.5 wt.%and negative correlation between SiO2 and Na2O + K2O as low-SiO2

reemajor wrench faults in the Zhejiang Province are the Jiangshan–Shaoxing fault (A), thend south of the surface trace of the Lishui–Yuyao fault (B). The sampling locations are thosend Yu et al. (2015a).White symbols stand for low-SiO2 alkaline basalts, and blue and greenspectively. (For interpretation of the references to color in this figure legend, the reader is

Fig. 3.Primitivemantle-normalizedmulti-elementdiagramofXinchang–Shengzhoufloodbasalts.Theprimitivemantle values are fromMcDonoughand Sun(1995). The trace elemental dataof basalts are from Yu et al. (2015a).

3X. Yu et al. / Lithos 284–285 (2017) 1–10

basalts (Fig. 2)). Negative correlation of Ar–Ar ages versus latitudes ofhigh-SiO2 basalts (Fig. 2B in Yu et al., 2015a) suggests that the floodbasalts were first emplaced in the south (Xinchang), and then movednorthward continuously to the north (Shengzhou), across the Lishui–Yuyao trans-lithospheric fault from9.4Ma to 3.0Ma (Fig. 1b). It indicatesthat the magmatism lasted more than six million years and the volcaniccenter was migrating northwards (Ho et al., 2003; Yu et al., 2015a).Besides of low SiO2 contents and negative correlation of SiO2 versusNa2O + K2O (Fig. 2a), low-SiO2 alkaline basalts (including nephelinite,basanite and minor alkaline olivine basalt) are characterized with highMgO contents (N8.5 wt.%), which are distinct from high-SiO2 alkaline/tholeiitic basalts (Fig. 2b). These low-SiO2 basalts were erupted sporadi-cally from 10.1 Ma to 3.0 Ma (Yu et al., 2015a) and marginally to theregion of flood basalt field (Fig. 1b).

Overall, low-SiO2 alkaline basalts are more enriched in highlyincompatible elements than high-SiO2 alkaline/tholeiitic basalts(Fig. 3). In a primitive mantle-normalized multi-element plot, low-SiO2 alkaline basalts are characterized by pronounced negative K, Pb,Zr, Hf, and Ti anomalies and positive Nb, Ta, and Sr anomalies. In com-parison, high-SiO2 alkaline/tholeiitic basalts show variable K anomaly,less degree of negative Pb, Zr, and Hf and positive Sr anomalies thanlow-SiO2 alkaline basalts (Fig. 3). Among high-SiO2 basalts, high-SiO2

alkaline basalts are more enriched in highly strong incompatibleelements and light rare earth elements (LREEs) than high-SiO2 tholeiiticbasalts.

According to the correlations of MgO versus Ni, CaO/Al2O3

and Rayleigh fractional calculations for SiO2 versus TiO2 (Fig. S1 ofYu et al., 2015a), the low-SiO2 alkaline basalts are characterized byminor fractionation of olivine while the high-SiO2 alkaline/tholeiiticbasalts are characterized by 10% to 30% fractionation of olivine andclinopyroxene. As most of the basalts are located on the mixing trendformed by nephelinite and the local juvenile lower continental crust inthe plot of Sm/Nd versus Ce/Pb, both kinds of basalts are proposed tobe geochemically affected by crustal contamination in the lower

Fig. 2. Plots of SiO2 versus Na2O+ K2O (a) and MgO (b), latitude versus Differentiation Index aSiO2 alkaline basalts, and trend two (2) and trend three (3) highlight high-SiO2 alkaline and tholestimated from the relationship between latitude and age (Age=−15.303× latitude+ 457.2)for low-SiO2 alkaline basalts. The blue and green symbols stand for high-SiO2 flood basalts fromtholeiitic basalts. Closed symbols represent data of the present study while open symbols rinterpretation of the references to color in this figure legend, the reader is referred to the web

continental crust (Yu et al., 2003, 2015a). All high-SiO2 basalts, haveevolved compositions (e.g., MgO b 8%) and the later erupted basaltsare always more evolved than the former erupted. For example, theShengzhou high-SiO2 alkaline/tholeiitic basalts have lower MgOcontents and higher Differentiation Index than the Xinchang high-SiO2

alkaline/tholeiitic basalts in Fig. 2. The above observations confirm thatmagma chamber processes are necessary in the formation of XS floodbasalts (Yu et al., 2015a).

According to the variations of fractional crystallization sensitivemajor elemental ratios (e.g., (Na2O + K2O)/MgO; Fig. 2C and D of Yuet al., 2015a) and Differentiation Index (Fig. 2c; Thornton and Tuttle,

nd age (c) for the Xinchang–Shengzhou flood basalts. (a) Trend one (1) stands for the loweiitic trends. (c) The axis of age is only valid for high-SiO2 alkaline/tholeiitic basaltswhich isfor high-SiO2 alkaline/tholeiitic basalts in Fig. 2B from Yu et al. (2015a).White circles standXinchang and Shengzhou. Among them, triangle and square symbols stand for alkaline andepresent published data for high-SiO2 alkaline/tholeiitic basalts (Ho et al., 2003). (Forversion of this article.)

4 X. Yu et al. / Lithos 284–285 (2017) 1–10

1960) relative to the latitude (and the basalt age), we found thatthe differentiation degree of magma stagnated for about six millionyears from 9.4 Ma to 3.5 Ma. Subsequently, the differentiationre-accelerated until the final eruption of this flood basalt field, whichhas been explained as response to the decrease of recharge intensityand magma temperature (Yu et al., 2015a). In addition, the low-SiO2

alkaline basalts erupted sporadically around the margin of the floodbasalt body among the eruption interval of high-SiO2 alkaline/tholeiiticbasalts verifying the continuous partial melting of the mantle source.Thus a magmatic recharge and magma chamber growth model wasproposed to explain the northward-directed formation and youngingof the high-SiO2 alkaline/tholeiitic basalts of XS flood basalts (Yu et al.,2015a), where magmatic recharge plays an important role in holdingthe degree ofmagma evolution unchanged and controlling the eruptioncenter migration.

In this study, we further subdivided the high-SiO2 alkaline/tholeiiticbasalts of the XS basalt field into the Xinchang high-SiO2 basalts in thesouth and the Shengzhou high-SiO2 basalts in the north, relative to theposition of the Lishui–Yuyao trans-lithospheric fault (Fig. 1), as theShengzhou high-SiO2 basalts are more evolved than the Xinchanghigh-SiO2 basalts (Fig. 2).

3. Methods

Here, newly measured Sr, Nd, Pb, and Hf isotopic compositions arereported for 26 samples. All data are shown in Table S1. Isotopic analy-ses were performed at the State Key Laboratory for Mineral DepositsResearch, School of Earth Sciences and Engineering at NanjingUniversity.Basaltic samples were first crushed into gravel-size chips. Clean chipswere then pulverized in a corundummill for isotopic analysis. All chem-ical digestion and separation were carried out in a Class 100 ultra-cleanlaboratory and the mass spectrometric analyses were performed in aClass 1000 clean laboratories. All reagents used for leaching, dissolutionand separation were twice-distilled, extra-pure grade. All dilutionswere made using ≥18.2 MΩ cm−1 de-ionized water, and all labwarewas acid-washed prior to use.

Isotope ratios of Sr, Nd were measured on a TRITON (ThermoFinnigan) thermal ionization mass spectrometer (TIMS) in static modewith relaymatrix rotation on singleW (Sr) or double Re (Nd) filaments.100 mg basaltic powder was leached for 12 h in warm 2.5 N HCl,and dissolved in a hot HF–HNO3 mixture followed by ion exchangeprocedures. Sr and Nd isotopic compositions were normalized to86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219. Detailed analytical pro-cedures for Sr and Nd isotopes can be seen elsewhere (Pu et al., 2005).During the analytical session for our samples, the measured values forthe NBS987 Sr standard and JNdi-1 Nd standard were 0.710239 ±0.000002 for 87Sr/86Sr and 0.512128 ± 0.000004 for 143Nd/144Nd,respectively. Measured 87Sr/86Sr and 143Nd/144Nd values for the basalticstandard BCR-2 were 0.705018 ± 0.000003 and 0.512624 ± 0.000004,respectively (reference values from Weis et al. (2006) are 0.705019 ±0.000016 and 0.512634 ± 0.000012).

Pb isotopic compositions were obtained using a Neptune plus(Thermo Finnigan) multi-collector inductively coupled plasma massspectrometer (MC-ICP-MS). 200 mg basaltic powder was leached for12 h in warm 2.5 N HCl before dissolved in HF–HNO3 acid mixture at120 °C for more than 36 h. After evaporation to dryness, all sampleswere repeatedly taken up and dried with concentrated HNO3 withtraces of HF in order to break CaF bonds. Finally, the samples weredissolved in HBr–HNO3 acid mixture. Pb was separated from the rockmatrix by chromatographic extraction using an anion exchange resin(200–400 mesh Bio-RAD resin). The chemical procedure of Pb separa-tion from natural rock samples consists of two steps. In the first step, alarge column was used to remove most of the elements other than Pb.In the second step, a small column was used to remove trace contami-nants of elements in the Pb fraction and organic materials derivedfrom resin in the first column. A HBr–HNO3 mixture is used to remove

elements other than Pb. Detailed chemical separation procedures aregiven by Kuritani and Nakamura (2002). Thalliumwas added to correctmass bias during the measurement. Measured Pb isotopic ratios werecorrected for instrumental mass fractionation by reference to replicateanalyses of the standard NIST-981. Detailed measurement proceduresare given by White et al. (2000). Measured values for the NIST-981 Pbstandard are 16.9318 ± 0.0003 for 206Pb/204Pb, 15.4858 ± 0.0003 for207Pb/204Pb, and 36.6819 ± 0.0008 for 208Pb/204Pb, respectively. Inaddition, international standards BHVO-2 and BCR-2 were also testedin this method as unknown samples. Measured values for the BHVO-2Pb standard are 16.6315 ± 0.0010 for 206Pb/204Pb, 15.4878 ± 0.0009for 207Pb/204Pb, and 36.1756± 0.0023 for 208Pb/204Pb. Measured valuesfor the BCR-2 Pb standard are 18.7881 ± 0.0005 for 206Pb/204Pb,15.6196 ± 0.0004 for 207Pb/204Pb, and 38.8005 ± 0.0010 for 208Pb/204Pb (reference values can be referred to Weis et al., 2006).

Hf isotopic data were obtained using a Neptune plus (ThermoFinnigan) multi-collector inductively coupled plasma mass spectrome-ter (MC-ICP-MS). 100 mg basaltic powder was leached for 12 h inwarm 2.5 N HCl and dissolved in 15 ml Teflon beakers in an HF–HClO4

acid mixture at 120 °C for more than 5 days. After evaporation todryness, all samples were dried at 200 °C in order to break CaF bonds.Finally, the samples were dissolved in 3 N HCl. Hafnium was separatedfrom the rock matrix by ion exchange procedures using an Eichrom®Ln-Spec resin. The detailed analytical procedure for the Hf isotopicmeasurement can be seen elsewhere (Yang et al., 2010). Hf isotopiccompositions were normalized to 179Hf/177Hf = 0.7325. The sampleresults were then normalized to the 176Hf/177Hf value of 0.282160(Vervoort et al., 1999) using the daily average of the JMC 475 Hfstandard. The JMC 475 Hf standard analyzed during the session of thesample analyses gave an average value of 176Hf/177Hf = 0.282157 ±0.000005. In addition, international standards BHVO-2 and BCR-2were also used as reference standards. Measured values for BHVO-2and BCR-2 were 0.283082 ± 0.000004 and 0.282857 ± 0.000006(reference values can be referred to Weis et al., 2007).

The total procedure blanks for Sr, Nd, Hf, and Pb isotopes were lessthan 100 pg, 60 pg, 50 pg, and 50 pg, respectively.

4. Results

The low-SiO2 alkaline basalts have variable isotopic compositions forSr, Nd,Hf and Pb, and the 87Sr/86Sr ratio ranges from0.70328 to 0.70435,εNd from+5.2 to+7.0, εHf from+7.7 to+9.2, and the 207Pb/204Pb ratiofrom 15.520 to 15.589. The basalts show positive correlations for Nd–Hfand Pb–Pb isotope systematics, and a negative correlation for Sr–Ndisotope systematics (Fig. 4). In comparison, the high-SiO2 alkaline/tholeiitic basalts show lower 87Sr/86Sr ratios at a given εNd value andlower εNd at a given εHf value (87Sr/86Sr = 0.70359–0.70434, εNd =+3.5 to +5.4, εHf = +6.2 to +9.1; Fig. 4a, b). For a given 206Pb/204Pbratio, high-SiO2 alkaline/tholeiitic basalts show higher 207Pb/204Pb and208Pb/204Pb (207Pb/204Pb = 15.560–15.619; Fig. 4c, d). Within thehigh-SiO2 alkaline/tholeiitic basalts, the Xinchang high-SiO2 basaltsand the Shengzhou high-SiO2 basalts form two distinct trends in Pb–Pb isotope systematics, characterized by higher 207Pb/204Pb and208Pb/204Pb at a given 206Pb/204Pb ratio (Fig. 4c, d). The XS flood basaltsthus form three groups with distinct major and trace element composi-tions (Yu et al., 2015a) and with distinct Sr, Nd, Hf, and Pb isotopiccompositions (Figs. 2 and 4).

5. Discussion

5.1. Crustal contamination in magma chambers

For theXSflood basalts, a trend ofmixing between aMORB&OIB-likemantle component and a lower continental crust component has beeninferred on the basis of Ce/Pb versus Sm/Nd (Fig. 3 in Yu et al., 2015a).The lower continental crust component is represented by the granulitic

Fig. 4. Plots of 87Sr/86Sr versus εNd and εNd versus εHf, and 206Pb/204Pb versus 207Pb/204Pb, 208Pb/204Pb for the Xinchang–Shengzhou flood basalts. In (a) and (b), crustal contaminationtrends are modeled by mixing between nephelinite and lower crustal granulite from southeastern China (Yu et al., 2003, 2007). 2 sigma standard errors for Sr, Nd, and Hf isotopiccompositions are shown. Parameters for simulation can be seen in Table S2. In (c) and (d), the 2 sigma standard error of each samples is smaller than the size of the symbols.

5X. Yu et al. / Lithos 284–285 (2017) 1–10

xenoliths reported in the basalts from southeastern China, which arecharacterized by low Ce/Pb b 15.0 (Yu et al., 2003). The high-SiO2

alkaline/tholeiitic basalts have low ratios of Ce/Pb (13.8–22.1) andhigh Sm/Nd ratios (0.24–0.29), which are comparable to the juvenilelower continental crust component as represented by the granuliticxenoliths with Ce/Pb b 15.0, and Sm/Nd N 0.30 (Yu et al., 2003). Thelow-SiO2 alkaline basalts, in contrast, show moderate Ce/Pb ratiosbetween 20.7 and 33.1, possibly suggesting that these basalts werenot as contaminated as the high-SiO2 basalts (Yu et al., 2015a).

The isotopes (Sr, Nd, and Hf) also indicate that both low-SiO2

alkaline basalts and high-SiO2 alkaline/tholeiitic basalts were affectedby crustal contamination (Fig. 4a, b). In the study area, the juvenilelower continental crust as represented by granulitic xenoliths andgranulitic batholith is characterized by high 87Sr/86Sr and low εNd, aswell as low εHf values (Table S2; Xu et al., 2007; Yu et al., 2003, 2007)(Fig. 4a, b). Thus contamination by lower continental crust yields anegative correlation of the Sr–Nd and a positive correlation of theNd–Hf isotope systematics of the basalts (Fig. 4a, b). According to theirdifferent isotopic arrays, high-SiO2 alkaline/tholeiitic basalts are lessconsistent with such crustal contamination than the low-SiO2 alkalinebasalts (Fig. 4a, b).

5.2. Mantle heterogeneity

5.2.1. Carbonated source with depleted isotopic compositionsAs the isotopic variations in low-SiO2 alkaline basalts indicate crustal

contamination, the most depleted isotopic values, represented by twonephelinite samples, may approach or record the primary isotopiccompositions of their mantle sources, i.e., a source with low 87Sr/86Srratios as well as high εNd and εHf values (Fig. 4a, b). Low-SiO2 alkalinebasalts are characterized by enriched in incompatible elements, highCaO/Al2O3 ratios (N0.8); strongly negative Zr, Hf, and Ti anomalies

(Hf/Hf* = 0.58–0.76 and Ti/Ti* = 0.80–0.97); and super-chondriticZr/Hf ratios (41.3–48.1). A carbonated mantle source was proposed toexplain such carbonatite-like geochemical features of nephelinitesand basanites elsewhere in eastern China (e.g., Huang et al., 2015;Zeng et al., 2010). Addition of carbonatitic components can enrich thedepleted mantle in incompatible elements besides of Zr, Hf, and Ti.Thus melting of such carbonated source generate melts with geochem-ical features similar to carbonatites (e.g., Dasgupta et al., 2007), whichexplains the key geochemical features of low-SiO2 alkaline basaltsfrom the XS volcanic field. Recently, this proposal has been supportedby prevalent low δ26Mg and high δ66Zn values of Cenozoic alkalinebasalts from eastern China, because carbonated mantle only can begenerated by the contribution of recycled carbonates (Li et al., 2016;S.-C. Liu et al., 2016; S.-A. Liu et al., 2016). Thus, the above observationssuggest a carbonated component with depleted isotopic compositionsin the asthenosphere source of the studied flood basalts (Yu et al.,2015a).

5.2.2. Crustal contamination or source heterogeneityThe isotopic compositions of the Xinchang high-SiO2 alkaline/

tholeiitic basalts deviate from themixing trend of crustal contaminationformed by the low-SiO2 alkaline basalts (Fig. 4a, b). Their unusualpositive correlation between 87Sr/86Sr and εNd, and negative correlationbetween εNd and εHfwhich are inconsistent with the crustal contamina-tion need to be explained (Fig. 4a, b). The decoupling relationshipbetween different isotopic systems has been attributed to recycledoceanic crust (Lassiter et al., 2003; Simonsen et al., 2000; Xu et al.,2012), recycled pelagic sediments (Blichert-Toft et al., 1999; Chauvelet al., 2008; Kuritani et al., 2011) or recycled lower continental crust(Chen et al., 2009; Zeng et al., 2011).

To better understand the origin of the isotopic co-variations of high-SiO2 alkaline/tholeiitic basalts, Δ87Sr/86Sr and ΔεHf for the XS flood

Fig. 5. Plots of latitude versusΔ87Sr/86Sr and ΔεHf for the Xinchang–Shengzhou flood basalts. Δ87Sr/86Sr refers to the relative deviation from themixing trend formed by low-SiO2 alkalinebasalts in Sr–Nd systematics as calculated byΔ87Sr/86Sr= (87Sr/86Sr− (895.35− εNd) / 1263.5) × 1000. ΔεHf refers to the relative deviation from themantle array in Nd–Hf systematics(ΔεHf = εHf − 1.55εNd − 1.21; Vervoort et al., 2011). 2SE for Δ87Sr/86Sr and ΔεHf are also presented. Symbols are the same as in Fig. 2.

6 X. Yu et al. / Lithos 284–285 (2017) 1–10

basalts were calculated, estimating the relative deviation from themixing trends formed by low-SiO2 alkaline basalts in Sr–Nd systematicsand the mantle array in Nd–Hf systematics (Vervoort et al., 2011)(Δ87Sr/86Sr = (87Sr/86Sr − (895.35 − εNd) / 1263.5) × 1000; ΔεHf =εHf − 1.55εNd − 1.21). Here the mixing trend formed by low-SiO2

alkaline basalts in Sr–Nd systematics represents the trend of crustalcontamination while the ΔεHf values are calculated relative to the Nd–Hf trend formed by global samples (the mantle array; Vervoort et al.,2011). The Xinchang high-SiO2 basalts show Δ87Sr/86Sr values between−0.85 and −1.91, and ΔεHf between −1.19 and +1.38, showinghigher degrees of deviation than the Shengzhou high-SiO2 basalts,which show Δ87Sr/86Sr between −0.54 and −1.34, and ΔεHf between−3.04 and −1.32 (Fig. 5). It indicates, besides crustal contamination,an alternative component, e.g. a kind of melts with distinct isotopiccompositions from the carbonated mantle endmember, is needed toexplain the isotopic variations of high-SiO2 basalts. Therefore, Xinchanghigh-SiO2 basalts record not only crustal contamination, but also thesource heterogeneity.

5.2.3. Recycled sediments in the sourceIntraplate basalts with positive ΔεHf values can be interpreted to

indicate recycled old lower continental crust (Chen et al., 2009; Zeng

Fig. 6. Plots of εHf versus Ba/Th and Th/La ratios for the Xinchang–Shengzhou flood basalts. Da2007). Data for GLOSS and average sediments of thewest Pacific aremodified fromPlank and Lafrom Sun and McDonough (1989) and Workman and Hart (2005). Symbols are the same as in

et al., 2011) or recycled pelagic sediments (Chauvel et al., 2008; Choiet al., 2014) in mantle sources. Here, a contribution of recycled oceanicsediments, rather than recycled lower continental crust, may explainthe elemental and isotopic signatures of Xinchang high-SiO2 basalts.Sediments are usually characterized by moderate Ba/Th and highTh/La ratios (Plank and Langmuir, 1998; Plank et al., 2007) (GLOSS:Ba/Th = 112.3, Th/La = 0.24; average sediments from the west PacificOcean: Ba/Th = 105.6, Th/La = 0.15). As Ba/Th and Th/La ratios arenot commonly affected by partial melting and fractional crystallization(Kuritani et al., 2011; Plank, 2005), they are considered as ideal toolsto trace the source differences besides identifying the signal of crustalcontamination. As lower continental crust in the study area is character-ized by high Ba/Th and low Th/La ratios (Yu et al., 2003) (Ba/Th N 470,Th/La b 0.08), and low εHf, contamination of lower continental crustcan generate negative εHf–Ba/Th (Fig. 6a) and positive εHf–Th/La corre-lations (Fig. 6b). Accordingly, contamination by lower continental crustcan explain the geochemical trends observed for the low-SiO2 alkalinebasalts and the Shengzhou high-SiO2 basalts. However, the Xinchanghigh-SiO2 basalts show largely invariant Ba/Th (91.9–106.5) and Th/La(0.145–0.150) ratios, which are comparable to those of averagesediments from the west Pacific, and we thus infer that they comprisea significant component of low Ba/Th and high Th/La sediments

ta for the lower continental crust from southeastern China are referred to Yu et al. (2003,ngmuir (1998). Trends for primitivemantle (PM) and depletedmantle (DM) are estimatedFig. 2.

Fig. 7. Simplified three endmembers mixing model for Pb isotopes. Endmembers ofancient (~1.5 Ga) and recent sediments are derived from Kuritani et al. (2011) (ancientsediments: 206Pb/204Pb = 16.757, 207Pb/204Pb = 15.478, Pb = 18 ppm; recentsediments: 206Pb/204Pb = 18.827, 207Pb/204Pb = 15.672, Pb = 18 ppm). The startingendmember is the depleted mantle material (DMM) from Workman and Hart (2005)with assumed Pb content (206Pb/204Pb = 18.275, 207Pb/204Pb = 15.486, Pb = 1 ppm).Symbols are the same as in Fig. 2.

7X. Yu et al. / Lithos 284–285 (2017) 1–10

(Fig. 6). This inference is consistent with the deduction that recycledcrustal material forms part of the mantle source of alkaline basaltsfrom the study area as was inferred on the basis of Os isotopes(187Os/188Os= 0.15–0.90) (Li et al., 2015). Here Pb isotopic systematicsare used to discover the nature of the sediments qualitatively (Fig. 7).We cited the ancient (~1.5 Ga) and recent sediments endmembersfrom Kuritani et al. (2011) (ancient sediments: 206Pb/204Pb = 16.757,207Pb/204Pb = 15.478, Pb = 18 ppm; recent sediments:206Pb/204Pb = 18.827, 207Pb/204Pb = 15.672, Pb = 18 ppm). TheDMM (depleted MORB mantle) is regarded as the starting endmemberfor high-SiO2 alkaline/tholeiitic basalts. Therefore, three endmembersmixing lines could be modeled (Fig. 7), where we can see that therecycled crustal materials are mainly composed of recent sediments.

Sediments characteristically have higher Rb/Sr, and lower Sm/Ndand Lu/Hf ratios than depleted mantle (Plank and Langmuir, 1998;Workman and Hart, 2005). Accordingly, sediments represent reservoirswith high 87Sr/86Sr, and low 143Nd/144Nd and 176Hf/177Hf ratios (Plank

Fig. 8.Modeling results for mantle source mixing and crustal contamination using Sr, Nd, andmantle material and recycled sediment from the west Pacific Ocean (Plank and Langmuir, 19simulated as in Fig. 4. Symbols are the same as in Fig. 2.

and Langmuir, 1998; Plank et al., 2007; Vervoort et al., 2011). Directmixing between depleted mantle and recycled sediments, however,cannot generate the observed isotopic co-variations for Xinchanghigh-SiO2 basalts (Fig. 4a, b). Here, we propose instead a two-stagemixing model that explains the isotopic characteristics of the Xinchanghigh-SiO2 basalts (Fig. 8). During the first stage, recycled oceanicsediments (~2%) mixed with the depleted asthenospheric mantle(~98%) and produced mantle sources with relatively enriched isotopicsignatures (lower εNd but higher εHf values than carbonated compo-nents in mantle sources) (“enriched asthenosphere” in Fig. 8). Duringthe second stage, melts derived from these enriched isotopic mantlesources mixed with 20% to 30% crustally contaminated low-SiO2 meltsrepresented by the low-SiO2 alkaline basalts (Fig. 8). The high-SiO2

Xinchang basalts represent the final products of this two-stage mixing.Fig. 8 and Table S2 present the inferred mixing trends and parametersin detail.

5.2.4. Isotopic response for the drop off of magmatic rechargeCompared to Xinchang high-SiO2 basalts, Shengzhou high-SiO2

basalts show lower 206Pb/204Pb ratios (Fig. 4) and lower Δ87Sr/86Srand ΔεHf values (Fig. 5). Since Xinchang high-SiO2 basalts andShengzhou high-SiO2 basalts are distributed along the two sides of theLishui–Yuyao fault, reasonably, their difference in isotopic compositionscan be generated by the chemical heterogeneity of the lithosphere,e.g. lower continental crust as the contaminants. However, such aproposal is not supported by the mixing trend of the Shengzhou high-SiO2 basalts on the plots of εHf versus Ba/Th and Th/La ratios (Fig. 6).On the contrary, they can be generated by mixing between Xinchanghigh-SiO2 basalts and lower crustal materials from study area. This isconsistent with their lower 206Pb/204Pb ratios because old lowercontinental crust is characterized by their unradiogeneic Pb isotopicfeatures (e.g., Liu et al., 2008). Therefore, Shengzhou high-SiO2 basaltsreflect a more crustal contaminated feature than Xinchang high-SiO2

basalts due to the deceleration of magmatic recharge (Yu et al., 2015a).In summary, a heterogeneous mantle source has been recorded in

the XS basalts. The Xinchang high-SiO2 basalts indicate a mantle sourcewith more enriched isotopic signatures than that of the low-SiO2

alkaline basalts (Fig. 8). In addition to crustal contamination, they thusrecord local and small-scale mantle heterogeneity.

5.3. Magmatic recharge buffers isotope systematics against crustalcontamination

To better understand the relationship between isotope systematics,and the spatial and temporal distribution of the XS flood basalts, a

Hf isotope systematics. The source mixing trend is modeled by mixing between depleted98; Vervoort et al., 2011; Workman and Hart, 2005). The crustal contamination trend is

Fig. 9.Cartoon illustrates themagma chamber processes to explain the formation ofXSflood basalts. (a) Primitive low-SiO2meltswere trapped in the lower crust and formed smallmagmareservoirs which were contaminated by the crust. The erupted magmas produced the low-SiO2 alkaline basalts. (b) The small magma reservoir was recharged by melts from mantlesources which were enriched by recycled sediments, resulting into an increase in chamber size. The erupted magmas produced the Xinchang high-SiO2 basalts. (c) With the decreaseof the magmatic recharge, the strengthening of the crustal contamination renewed. The erupted magmas then produced the Shengzhou high-SiO2 basalts.

8 X. Yu et al. / Lithos 284–285 (2017) 1–10

three-stage-evolution model is proposed: infancy (Eq. (5-1)), growth(Eq. (5-2)), and decline (Eq. (5-3)), as follows.

Low‐SiO2 alkaline meltsCarbonated mantleð Þ

þCrust contamination→

Low‐SiO2 alkaline basaltsSmall magma chambersð Þ

ð5� 1Þ

Low‐SiO2 alkaline meltsþ High‐SiO2meltsSmall magma chambersð Þ Enriched mantleð Þ

recharge

→ Xinchang high‐SiO2 basaltsEnlarged magma chambersð Þ

ð5� 2Þ

Fig. 10.Continuously erupted sequence fromthe Emeishan large igneous province divided into uselected to infer the role of magmatic recharge. Nb/La and εNd were selected to detect the signThe indicated average lower continental crust composition is from Rudnick and Gao (2003).

Xinchang high‐SiO2meltsþ CrustEnlarged magma chambersð Þ

contamination→

Shengzhou high‐SiO2 basaltsDeclined magma chambersð Þ

ð5� 3ÞDuring the first stage, partial melting of carbonated mantle generat-

ed low-SiO2 alkaline melts. Some of these primitive low-SiO2 meltswere trapped in the lower continental crust, where they formed smallmagma reservoirs and were contaminated by crust (Fig. 9a). Theerupted magmas produced the low-SiO2 alkaline basalts with negativecorrelations in Sr and Nd isotopic compositions and positive correla-tions in Nd and Hf isotopic compositions (Eq. (5-1)). In the secondstage, the small magma chambers were recharged from the mantle

pper and lower parts (data fromSonget al. (2006)).MgO, Ni and (Na2O+K2O)/MgOwereal of crustal contamination.

9X. Yu et al. / Lithos 284–285 (2017) 1–10

source beneath by continuously partial melting (Fig. 9b), resulting in anincrease in chamber size via melt supply from mantle sources whichwere enriched by recycled sediments (Eq. (5-2)). Subsequently, theerupting magmas produced the Xinchang high-SiO2 basalts (Figs. 5, 6and 8), while the residual magma chambers were growing northwardsas evidenced by coupled basalt distributions, age correlations and ele-mental variations (Yu et al., 2015a). In this stage, the unusual Sr–Ndand Nd–Hf istopic correlations were generated (Fig. 4). The final stageis characterized by waning magmatic recharge and renewed strength-ening of the signal of crustal contamination in the magma chambers(Fig. 9c). The magmas with stronger signal of crustal contaminationerupted as the Shengzhou high-SiO2 basalts (Eq. (5-3)).

We therefore suggest that mixing of variable volumes of juvenilerecharge magma and crustal components is recorded by the elementaland isotopic compositions of the XS flood basalts, but we have alsoinferred that the XS flood basalt compositions reflect variable isotopiccompositions of their mantle sources. To further test whether otherflood basalts also record the isotopic characteristics of their mantlesources, a continuous eruption sequence from the Emeishan largeigneous province (ca. 260 Ma) in southwestern China was examined(Song et al., 2006) (Fig. 10). Song et al. (2006) suggested that thebasaltic rocks from this sequence are the result of magma chamberprocesses (i.e., crustal contamination). Here, we use MgO, Ni and(Na2O + K2O)/MgO to identify magmatic recharge. When the contentsof MgO and Ni increase and the ratios of (Na2O + K2O)/MgO remainstagnant, it suggests magmatic recharge (Yu et al., 2015a). In Fig. 10,the stratigraphically lower part of the sequence (below 400 m depth)showsstrongoscillations in theMgO,Ni contents and(Na2O+K2O)/MgOand Nb/La ratios, which suggests that the basalts experienced strongfractional crystallization. In the stratigraphically upper part of thesequence (above 400 m depth), the contents of these elements and el-emental ratios are less variable, suggesting that the evolution of basalticmagma chambers was controlled by continuous magmatic recharge.Meanwhile, samples from the upper section show higher Nb/La ratiosand εNd than the lower section basalts (Fig. 10), indicating a lower con-tribution from crustal contamination. Such phenomena can also be ob-served from other continuously erupted sequences from Emeishan LIPslike Heishitou section (e.g., Qi and Zhou, 2008). This suggests that theCRH model for the continental flood basalts proposed in the presentstudy can further be used to globally evaluate the isotopic compositionsof large igneous province basalts.

6. Conclusion

New Sr, Nd, Hf, and Pb isotope geochemical data combined withinsights from previous whole-rock elemental and age data indicate thatthe XS flood basalts, exposed in eastern China resulted from multi-stage and multi-component mixing, involving crustal contamination,magmatic recharge and partial melting of a heterogeneous mantle(CRH). Acceleration ofmagmatic recharge resulted in isotope systematicsof flood basalts evolving from a contamination-dominated signature to aheterogeneous-source-dominated signature. While crustal contamina-tion is common in continental flood basalts, large-volume magmaticrecharge (Camp and Roobol, 1989; Devey and Cox, 1987; Kumar et al.,2010; Song et al., 2006; Yu et al., 2015b) is key for flood basalts topreserve the geochemical features of their mantle sources. Accordingly,the results of this work provide the basis for applying the isotopesystematics of continental flood basalts in tracing their mantle sources,even in cases in which crustal contamination has been significant.

Acknowledgement

We thank Wei Pu, Tao Yang, Lie-Wen Xie, Yue-Heng Yang, andHai-Zhen Wei for providing the technical support. Cin-Ty Lee, SaskiaErdmann, Wen-Rong Cao, Yuan Li, and Jian-Qiang Liu are thanked forthe discussions. We are grateful to Dr. Kung-Suan Ho for providing the

location of his samples. Two anonymous reviewers are acknowledgedfor their constructive and helpful comments. This study was supportedby NSFC (grants 41430208, 41688103 and 41202039).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.lithos.2017.03.027.

References

Arndt, N.T., Christensen, U., 1992. The role of lithospheric mantle in continental floodvolcanism: thermal and geochemical constraints. Journal of Geophysical Research97, 10967–10981.

Arndt, N.T., Czamanske, G.K., Wooden, J.L., Fedorenko, V.A., 1993. Mantle and crustalcontributions to continental flood volcanism. Tectonophysics 223, 39–52.

Blichert-Toft, J., Frey, F., Albarede, F., 1999. Hf isotope evidence for pelagic sediments inthe source of Hawaiian basalts. Science 285, 879–882.

Bohrson, W.A., Spera, F.J., 2003. Energy-constrained open-system magmatic processes IV:geochemical, thermal and mass consequences of energy-constrained recharge,assimilation and fractional crystallization (EC-RAFC). Geochemistry, Geophysics,Geosystems 4. http://dx.doi.org/10.1029/2002GC000316.

Brandon, A.D., Hooper, P.R., Goles, G.G., Lambert, R.S.J., 1993. Evaluating crustal contami-nation in continental basalts: the isotopic composition of the Picture Gorge Basalt ofthe Columbia River Basalt Group. Contributions to Mineralogy and Petrology 114,452–464.

Camp, V.E., Roobol, M.J., 1989. The Arabian continental alkali basalt province: part I.Evolution of Harrat Rahat, Kingdom of Saudi Arabia. Geological Society of AmericaBulletin 101, 71–95.

Carlson, R.W., Lugmair, G.W., Macdougall, J.D., 1981. Columbia River volcanism: the ques-tion of mantle heterogeneity or crustal contamination. Geochimica et CosmochimicaActa 45, 2483–2499.

Chauvel, C., Lewin, E., Carpentier,M., Arndt, N.T.,Marini, J.C., 2008. Role of recycled oceanicbasalt and sediment in generating the Hf–Nd mantle array. Nature Geoscience 1,64–67.

Chen, L.-H., Zeng, G., Jiang, S.-Y., Hofmann, A.W., Xu, X.-S., Pan, M.-B., 2009. Sources ofAnfengshan basalts: subducted lower crust in the Sulu UHP belt, China. Earth andPlanetary Science Letters 286, 426–435.

Choi, H.-O., Choi, S.H., Yu, Y., 2014. Isotope geochemistry of Jeongok basalts, northernmostSouth Korea: implications for the enriched mantle end-member component. Journalof Asian Earth Sciences 91, 56–68.

Chung, S.-L., Jahn, B.-M., 1995. Plume–lithosphere interaction in generation of theEmeishan flood basalts at the Permian–Triassic boundary. Geology 23, 889–892.

Cox, K.G., 1980. A model for flood basalt vulcanism. Journal of Petrology 21, 629–650.Dasgupta, R., Hirschmann, M.M., Smith, N.D., 2007. Partial melting experiments of

peridotite + CO2 at 3 GPa and genesis of alkalic ocean island basalts. Journal ofPetrology 48, 2093–2124.

Davidson, J.P., Tepley III, F.J., 1997. Recharge in volcanic systems: evidence from isotopeprofiles of phenocrysts. Science 275, 826–829.

Devey, C.W., Cox, K.G., 1987. Relationships between crustal contamination andcrystallisation in continental flood basalt magmas with special reference to theDeccan Traps of the Western Ghats, India. Earth and Planetary Science Letters 84,59–68.

Ho, K.-S., Chen, J.-C., Lo, C.-H., Zhao, H.-L., 2003. 40Ar–39Ar dating and geochemicalcharacteristics of late Cenozoic basaltic rocks from the Zhejiang–Fujian region, SEChina: eruption ages, magma evolution and petrogenesis. Chemical Geology 197,287–318.

Hooper, P.R., Camp, V.E., Reidel, S.P., Ross, M.E., 2007. The origin of the Columbia Riverflood basalt province: plume versus nonplume models. Geological Society ofAmerica Special Papers 430, 635–668.

Howarth, G.H., Prevec, S.A., 2013. Trace element, PGE, and Sr–Nd isotope geochemistry ofthe Panzhihua mafic layered intrusion, SW China: constraints on ore-forming pro-cesses and evolution of parent magma at depth in a plumbing-system. Geochimicaet Cosmochimica Acta 120, 459–478.

Huang, J.-L., Zhao, D.-P., 2006. High-resolution mantle tomography of China andsurrounding regions. Journal of Geophysical Research 111. http://dx.doi.org/10.1029/2005JB004066.

Huang, J., Li, S.-G., Xiao, Y.-L., Ke, S., Li, W.-Y., Tian, Y., 2015. Origin of low δ26Mg Cenozoicbasalts from South China Block and their geodynamic implications. Geochimica etCosmochimica Acta 164, 298–317.

Kumar, K.V., Chavan, C., Sawant, S., Raju, K.N., Kanakdande, P., Patode, S., Deshpande, K.,Krishnamacharyulu, S.K.G., Vaideswaran, T., Balaram, V., 2010. Geochemical investi-gation of a semi-continuous extrusive basaltic section from the Deccan Volcanic Prov-ince, India: implications for themantle andmagma chamber processes. Contributionsto Mineralogy and Petrology 159, 839–862.

Kuritani, T., Nakamura, E., 2002. Precise isotope analysis of nanogram-level Pb for naturalrock samples without use of double spikes. Chemical Geology 186, 31–43.

Kuritani, T., Ohtani, E., Kimura, J.-I., 2011. Intensive hydration of the mantle transitionzone beneath China caused by ancient slab stagnation. Nature Geoscience 4,713–716.

Lassiter, J., Blichert-Toft, J., Hauri, E., Barsczus, H., 2003. Isotope and trace elementvariations in lavas from Raivavae and Rapa, Cook-Austral islands: constraints on the

10 X. Yu et al. / Lithos 284–285 (2017) 1–10

nature of HIMU- and EM-mantle and origin of mid-plate volcanism in FrenchPolynesia. Chemical Geology 202, 115–138.

Lee, C.-T., Lee, T.-C., Wu, C.-T., 2014. Modeling the compositional evolution of recharging,evacuating, and fractionating (REFC) magma chambers: implications for differentia-tion of arc magmas. Geochimica et Cosmochimica Acta 143, 8–22.

Li, Y.-Q., Ma, C.-Q., Robinson, P.T., Zhou, Q., Liu, M.-L., 2015. Recycling of oceanic crust froma stagnant slab in the mantle transition zone: evidence from Cenozoic continentalbasalts in Zhejiang Province, SE China. Lithos 230, 146–165.

Li, S.-G., Yang,W., Ke, S., Meng, X., Tian, H., Xu, L., He, Y., Huang, J., Wang, X.-C., Xia, Q., Sun,W., Yang, X., Ren, Z.-Y., Wei, H., Liu, Y., Meng, F., Yan, J., 2016. Deep carbon cyclesconstrained by a large-scale mantle Mg isotope anomaly in eastern China. NationalScience Review, nww070.

The K–Ar age and tectonic environment of Cenozoic volcanic rocks in China. In: Liu, R.-X.,Chen, J.-Z., Sun, J.-Z., Li, D.-M., Liu, R.X. (Eds.), The Age and Geochemistry of CenozicVolcanic Rock in China. Seismology Publ., Beijing, pp. 1–43 (in Chinese).

Liu, Y.-S., Gao, S., Kelemen, P.B., Xu,W.-L., 2008. Recycled crust controls contrasting sourcecompositions of Mesozoic and Cenozoic basalts in the North China Craton.Geochimica et Cosmochimica Acta 72, 2349–2376.

Liu, S.-C., Xia, Q.-K., Choi, S.H., Etienne, D., Li, P., Liu, J., 2016a. Continuous supply ofrecycled Pacific oceanic materials in the source of Cenozoic basalts in SE China: theZhejiang case. Contributions to Mineralogy and Petrology 171, 100.

Liu, S.-A., Wang, Z.-Z., Li, S.-G., Huang, J., Yang, W., 2016b. Zinc isotope evidence for alarge-scale carbonated mantle beneath eastern China. Earth and Planetary ScienceLetters 444, 169–178.

Maydagan, L., Franchini, M., Chiaradia, M., Dilles, J., Rey, R., 2014. The Altar porphyryCu–(Au–Mo) deposit (Argentina): a complex magmatic-hydrothermal system withevidence of recharge processes. Economic Geology 109, 621–641.

McDonough, W.F., Sun, S.-S., 1995. The composition of the earth. Chemical Geology 120,223–253.

Meshesha, D., Shinjo, R., 2007. Crustal contamination and diversity of magma sourcesin the northwestern Ethiopian volcanic province. Journal of Mineralogical andPetrological Sciences 102, 272–290.

O'Hara, M.J., Mathews, R.E., 1981. Geochemical evolution in an advancing, periodicallyreplenished, periodically tapped, continuously fractionated magma chamber. Journalof the Geological Society of London 138, 237–277.

Pang, K.-N., Li, C., Zhou, M.-F., Ripley, E.M., 2009. Mineral compositional constraints onpetrogenesis and oxide ore genesis of the late Permian Panzhihua layered gabbroicintrusion, SW China. Lithos 110, 199–214.

Plank, T., 2005. Constraints from thorium/lanthanum on sediment recycling at subductionzones and the evolution of the continents. Journal of Petrology 46, 921–944.

Plank, T., Langmuir, C.H., 1998. The chemical composition of subducting sediment and itsconsequences for the crust and mantle. Chemical Geology 145, 325–394.

Plank, T., Kelley, K.A., Murray, R.W., Stern, L.Q., 2007. Chemical composition ofsediments subducting at the Izu–Bonin trench. Geochemistry, Geophysics,Geosystems 8. http://dx.doi.org/10.1029/2006GC001444.

Pu, W., Gao, J.-F., Zhao, K.-D., Lin, H.-F., Jiang, S.-Y., 2005. Separation method of Rb–Sr,Sm–Nd using DCTA and HIBA. Journal of Nanjing University (Natural Sciences) 41,445–450 (in Chinese with English abstract).

Qi, L., Zhou, M.-F., 2008. Platinum-group elemental and Sr–Nd–Os isotopic geochemistryof Permian Emeishan flood basalts in Guizhou Province, SW China. Chemical Geology248, 83–103.

Ramos, F.C., Wolff, J.A., Tollstrup, D.L., 2005. Sr isotope disequilibrium in Columbia Riverflood basalts: evidence for rapid shallow-level open-system processes. Geology 33,457–460.

Rudnick, R.L., Gao, S., 2003. Composition of the continental crust. Treatise onGeochemistry3, 1–64.

Sakuyama, T., Tian, W., Kimura, J.-I., Fukao, Y., Hirahara, Y., Takahashi, T., Senda, R., Chang,Q., Miyazaki, T., Obayashi, M., Kawabata, H., Tatsumi, Y., 2013. Melting of dehydratedoceanic crust from the stagnant slab and of the hydrated mantle transition zone:constraints from Cenozoic alkaline basalts in eastern China. Chemical Geology 224,247–271.

Sharma, M., Basu, A.R., Nesterenko, G.V., 1991. Nd–Sr isotopes, petrochemistry, and originof the Siberian flood basalts, USSR. Geochimica et Cosmochimica Acta 55, 1183–1192.

Sharma, M., Basu, A.R., Nesterenko, G.V., 1992. Temporal Sr-, Nd- and Pb-isotopic varia-tions in the Siberian flood basalts: implications for the plume-source characteristics.Earth and Planetary Science Letters 113, 365–381.

Silver, P.G., Behn, M.D., Kelley, K., Schmitz, M., Savage, B., 2006. Understanding cratonicflood basalts. Earth and Planetary Science Letters 245, 190–201.

Simonsen, S., Neumann, E.-R., Seim, K., 2000. Sr–Nd–Pb isotope and trace-elementgeochemistry evidence for a young HIMU source and assimilation at Tenerife (CanaryIsland). Journal of Volcanology and Geothermal Research 103, 299–312.

Sobolev, A.V., Krivolutskaya, N.A., Kuzmin, D.V., 2009. Petrology of the parental melts andmantle sources of Siberian trap magmatism. Petrology 17, 253–286.

Sobolev, S.V., Sobolev, A.V., Kuzmin, D.V., Krivolutskaya, N.A., Petrunin, A.G., Arndt, N.T.,Radko, V.A., Vasiliev, Y.R., 2011. Linking mantle plumes, large igneous provincesand environmental catastrophes. Nature 477, 312–316.

Song, X.-Y., Zhou, M.-F., Keays, R.R., Cao, Z.-M., Sun, M., Qi, L., 2006. Geochemistry of theEmeishan flood basalts at Yangliuping, Sichuan, SW China: implications for sulfidesegregation. Contributions to Mineralogy and Petrology 152, 53–74.

Spera, F.J., Bohrson, W.A., 2002. Energy-constrained open-system magmatic processes 3.Energy-constrained recharge, assimilation, and fractional crystallization (EC-RAFC).Geochemistry, Geophysics, Geosystems 3. http://dx.doi.org/10.1029/2002GC000315.

Sun, S.-S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts:implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J.(Eds.), Magmatism in the Ocean Basins. Geol. Soc. Spec. Publ. 42, pp. 313–345.

Thornton, C.P., Tuttle, O.F., 1960. Chemistry of igneous rocks—part 1, differentiation index.American Journal of Science 258 (9), 664–684.

Vervoort, J.D., Patchett, P.J., Blichert-Toft, J., Albarède, F., 1999. Relationships betweenLu–Hf and Sm–Nd isotopic systems in the global sedimentary system. Earth andPlanetary Science Letters 168, 79–99.

Vervoort, J.D., Plank, T., Prytulak, J., 2011. The Hf–Nd isotopic composition of marinesediments. Geochimica et Cosmochimica Acta 75, 5903–5926.

Wark, D.A., Hildreth, W., Spear, F.S., Cherniak, D.J., Watson, E.B., 2007. Pre-eruptionrecharge of the Bishop magma system. Geology 35, 235–238.

Weis, D., Kieffer, B., Maerschalk, C., Barling, J., de Jong, J., Williams, G.A., Hanano, D.,Pretorius, W., Mattielli, N., Scoates, J.S., 2006. High-precision isotopic characterizationof USGS reference materials by TIMS and MC-ICP-MS. Geochemistry, Geophysics,Geosystems 7. http://dx.doi.org/10.1029/2006GC001283.

Weis, D., Kieffer, B., Hanano, D., Nobre Silva, I., Barling, J., Pretorius, W., Maerschalk, C.,Mattielli, N., 2007. Hf isotope compositions of US Geological Survey referencematerials. Geochemistry, Geophysics, Geosystems 8. http://dx.doi.org/10.1029/2006GC001473.

White, W.M., Albarède, F., Télouk, P., 2000. High-precision analysis of Pb isotope ratios bymulti-collector ICP-MS. Chemical Geology 167, 257–270.

Workman, R.K., Hart, S.R., 2005. Major and trace element composition of the depletedMORB mantle (DMM). Earth and Planetary Science Letters 231, 53–72.

Xiao, L., Xu, Y.-G., Mei, H.-J., Zheng, Y.-F., He, B., Pirajno, F., 2004. Distinctmantle sources oflow-Ti and high-Ti basalts from the western Emeishan large igneous province, SWChina: implications for plume–lithosphere interaction. Earth and Planetary ScienceLetters 228, 525–546.

Xu, Y.-G., 2014. Recycled oceanic crust in the source of 90–40 Ma basalts in Northand Northeast China: evidence, provenance and significance. Geochimica etCosmochimica Acta 143, 49–67.

Xu, Y.-G., Ma, J.-L., Frey, F.A., Feigenson, M.D., Liu, J.-F., 2005. Role of lithosphere–asthenosphere interaction in the genesis of Quaternary alkali and tholeiitic basaltsfrom Datong, western North China Craton. Chemical Geology 224, 247–271.

Xu, X.-S., O'Reilly, S.Y., Griffin, W.L., Wang, X.-L., Pearson, N.J., He, Z.-Y., 2007. The crust ofCathaysia: age, assembly and reworking of two terranes. Precambrian Research 158,51–78.

Xu, Y.-G., Zhang, H.-H., Qiu, H.-N., Ge, W.-C., Wu, F.-Y., 2012. Oceanic crust componentsin continental basalts from Shuangliao, Northeast China: derived from the mantletransition zone? Chemical Geology 328, 168–184.

Yang, Y., Zhang, H., Chu, Z., Xie, L., Wu, F., 2010. Combined chemical separation of Lu, Hf,Rb, Sr, Sm and Nd from a single rock digest and precise and accurate isotope determi-nations of Lu–Hf, Rb–Sr and Sm–Nd isotope systems using Multi-Collector ICP-MSand TIMS. International Journal of Mass Spectrometry 290, 120–126.

Yu, J.-H., Xu, X.-S., O'Reilly, S.Y., Griffin, W.L., Zhang, M., 2003. Granulite xenoliths fromCenozoic Basalts in SE China provide geochemical fingerprints to distinguish lowercrust terranes from the North and South China tectonic blocks. Lithos 67, 77–102.

Yu, J.-H., O'Reilly, Y.S., Wang, L.-J., Griffin, W.L., Jiang, S.-Y., Wang, R.-C., Xu, X.-S., 2007.Finding of ancient materials in Cathaysia and implication for the formation ofPrecambrian crust. Chinese Science Bulletin 52, 13–22.

Yu, X., Chen, L.-H., Zeng, G., 2015a. Growing magma chambers control the distributionpattern of small-scale flood basalts. Scientific Reports 5. http://dx.doi.org/10.1038/srep16824.

Yu, X., Lee, C.-T., Chen, L.-H., Zeng, G., 2015b. Magmatic recharge in continental floodbasalt: insight into Chifeng igneous province in Inner Mongolia. Geochemistry,Geophysics, Geosystems http://dx.doi.org/10.1002/2015GC005805.

Zeng, G., Chen, L.-H., Xu, X.-S., Jiang, S.-Y., Hofmann, A.W., 2010. Carbonated mantlesources for Cenozoic intra-plate alkaline basalts in Shandong, North China. ChemicalGeology 273, 35–45.

Zeng, G., Chen, L.-H., Hofmann, A.W., Jiang, S.-Y., Xu, X.-S., 2011. Crust recycling in thesources of two parallel volcanic chains in Shandong, North China. Earth and PlanetaryScience Letters 302, 359–368.

Zhang, Z.-C., Mahoney, J.J., Mao, J.-W., Wang, F.-S., 2006. Geochemistry of picritic andassociated basalt flows of the western Emeishan flood basalt province, China. Journalof Petrology 47, 1997–2019.