当前位置：
X-MOL 学术
›
Lithosphere
›
论文详情
Our official English website, www.x-mol.net, welcomes your
feedback! (Note: you will need to create a separate account there.)
Coexisting Late Cenozoic Potassic and Sodic Basalts in NE China: Role of Recycled Oceanic Components in Intraplate Magmatism and Mantle Heterogeneity
Lithosphere ( IF 1.8 ) Pub Date : 2020-06-30 , DOI: 10.2113/2020/8875012 Ming Lei 1, 2, 3 , Zhengfu Guo 1, 2, 3 , Wenbin Zhao 1, 2, 3 , Maoliang Zhang 1, 2, 3, 4 , Lin Ma 1, 2, 3
Lithosphere ( IF 1.8 ) Pub Date : 2020-06-30 , DOI: 10.2113/2020/8875012 Ming Lei 1, 2, 3 , Zhengfu Guo 1, 2, 3 , Wenbin Zhao 1, 2, 3 , Maoliang Zhang 1, 2, 3, 4 , Lin Ma 1, 2, 3
Affiliation
This study presents an integrated geochemical study of the Wudalianchi-Erkeshan potassic basalts and Halaha sodic basalts of NE China, and uses these data to further our understanding of the petrogenetic relationships between the coeval potassic and sodic basalts in this region. The potassic basalts with high concentrations of K2O have arc-like trace-element compositions and enriched Sr-Nd-Hf isotopic compositions with unradiogenic 206Pb/204Pb values (16.77–16.90). In contrast, the sodic basalts with high concentrations of Na2O have OIB-like trace-element compositions and depleted Sr-Nd-Hf isotopic compositions with radiogenic 206Pb/204Pb values (18.27–18.40). These data suggest that the potassic and sodic basalts were derived from mixed depleted mid-ocean-ridge basalt mantle (DMM) and enriched mantle source end-members, where the enriched end-members are ancient sediment for the potassic basalts and Pacific oceanic crust for the sodic basalts. The combined geophysical and geochemical data indicate that these two enriched end-members are located in the mantle transition zone. We propose that partial melting of upwelling asthenospheric mantle comprising ambient DMM and recycled materials shifting from the ancient sediment to the Pacific oceanic crust could have produced the coeval potassic and sodic basalts in NE China. The proposed mantle sources for the potassic and sodic basalts indicate that the upper mantle beneath NE China was highly heterogeneous during late Cenozoic.Continental alkali basalts including potassic basalts (K2O/Na2O>1) and sodic basalts (Na2O/K2O>1) are particularly important because they preserve geochemical features that likely reflect the nature of their mantle source (e.g., [1]). Previous studies have suggested that potassic basalts are the result of the low degree of melting of phlogopite-bearing peridotite (or pyroxenite) in the subcontinental lithospheric mantle (SCLM) or asthenospheric mantle ([2–9], 2016), whereas sodic basalts are generally the result of decompression melting of asthenospheric mantle or mantle plumes [10]. According to Bonin [11, 12], potassic basalts are usually formed in postcollisional or postorogenic settings, whereas sodic basalts are normally formed within continental and oceanic lithosphere associated with rift systems, hot spots, or mantle plumes (e.g., [13]).Intriguingly, coeval spatially and temporally related potassic and sodic basalts are also reported in some places, including the Basin-and-Range Province (e.g., the Rio Grande Rift) [14–17], the East African Rift [18], and the Hong’an-Dabie orogen, China [19–21]. Previous studies have proposed two main models to explain the coeval potassic and sodic basalts as follows: (1)In the lithospheric thinning or delamination model, potassic basalts are formed by the partial melting of a previously enriched lithospheric mantle, whereas sodic basalts are presumed to be formed by the decompressional melting of sublithospheric material (asthenospheric mantle or mantle plume) when the local lithospheric mantle is thinning [14, 15, 17, 18]. Alternatively within this model, potassic basalts could be produced by the melting of a delaminated lithospheric mantle veined by a mica-bearing, Al-poor assemblage at high pressure (great depth), whereas sodic basalts could be generated by the melting of a delaminated lithospheric mantle veined by an aluminous amphibole-bearing assemblage at low pressure (shallow depth) [16](2)In the recycled crustal material model [19–21], potassic basalts are viewed as resulting from the partial melting of metasomatites that were produced by a reaction between mantle-wedge peridotite and recycled continental crust, whereas sodic basalts are considered as resulting from the partial melting of metasomatites that were produced by the reaction between mantle-wedge peridotite and recycled oceanic crustIn the lithospheric thinning or delamination model, potassic basalts are formed by the partial melting of a previously enriched lithospheric mantle, whereas sodic basalts are presumed to be formed by the decompressional melting of sublithospheric material (asthenospheric mantle or mantle plume) when the local lithospheric mantle is thinning [14, 15, 17, 18]. Alternatively within this model, potassic basalts could be produced by the melting of a delaminated lithospheric mantle veined by a mica-bearing, Al-poor assemblage at high pressure (great depth), whereas sodic basalts could be generated by the melting of a delaminated lithospheric mantle veined by an aluminous amphibole-bearing assemblage at low pressure (shallow depth) [16]In the recycled crustal material model [19–21], potassic basalts are viewed as resulting from the partial melting of metasomatites that were produced by a reaction between mantle-wedge peridotite and recycled continental crust, whereas sodic basalts are considered as resulting from the partial melting of metasomatites that were produced by the reaction between mantle-wedge peridotite and recycled oceanic crustLate Cenozoic intraplate volcanic rocks are widely in and around the Songliao Basin and occur along the Yilan-Yitong and Fushun-Mishan faults, NE China (e.g., [22]). Coeval potassic and sodic basalts have also been extensively reported in NE China (e.g., [23–34]). However, the mantle sources of these potassic and sodic basalts are still unresolved, with various proposals for their sources including (1) SCLM metasomatized by delaminated ancient lower continental crust or recycled ancient sediment [26, 27, 31, 33–35], (2) interaction between asthenospheric (or carbonated asthenospheric) mantle and enriched (or carbonated) lithospheric mantle [28, 36], (3) asthenospheric mantle enriched by delaminated ancient lower continental crust or recycled oceanic materials (oceanic crust and/or sediment; [37–42]), (4) interaction between depleted lithospheric mantle and recycled ancient subducted sediment [24, 25, 29], (5) depleted mid-ocean-ridge basalt (MORB) mantle (DMM) [23, 32], and (6) ancient primitive mantle with recycled oceanic materials [30]. Recent studies have focused on the petrogenetic relationship among these (ultra)potassic basalts and have proposed that the geochemical variations (e.g., K2O/Na2O and Rb/Nb ratios) of these potassic basalts might result from melt-lithosphere interaction [24, 25]. However, another important issue is the petrogenetic relationship of the coeval potassic and sodic OIB basalts in NE China, which is still poorly constrained.For this study, we have conducted an integrated investigation of olivine, whole-rock major- and trace-element, and Sr-Nd-Pb-Hf radiogenic isotopic compositions of potassic basalts from the Wudalianchi-Erkeshan volcanic field and sodic basalts from the Halaha volcanic field of NE China. We combine our new data with previously published data in these areas to (1) constrain the mantle source of the potassic basalts and sodic basalts, (2) evaluate the petrogenetic relationship of these two suites of basalts, and (3) characterize the upper mantle beneath NE China.The Xing’an-Mongolia Orogenic Belt is the eastern segment of the Central Asian Orogenic Belt, which lies between the Siberia and Baltica cratons to the north and the Tarim and North China cratons to the south (e.g., [43]). NE China lies within the eastern portion of the Paleozoic Central Asian Orogenic Belt (Figure 1(a)). From Paleozoic to Mesozoic, NE China has experienced the amalgamation of several microcontinental blocks (e.g., Xing’an, Songliao, and Jamusi) along suture zones [44, 45]. Since the Late Jurassic, the tectonic history of NE China has been dominated by the Paleo-Pacific plate, as evidenced by the Jurassic-Cretaceous accretionary complexes along the eastern Eurasian plate [46]. During the Cenozoic, NE China was in a continental extension setting due to the Pacific slab rollback and trench retreat [47], which probably resulted in asthenospheric upwelling and led to continental intraplate volcanism in NE China [48–50].The Wudalianchi volcanic field and the adjacent Erkeshan volcanic field, which were located on the Northern margin of the Songliao Basin in the Xing’an-Mongolia Orogenic Belt (Figure 1(b)), are known for producing highly potassic basalts (e.g., [51]). Previous studies indicated that the Wudalianchi and Erkeshan volcanic activities mainly occurred in the middle Pleistocene (0.56–0.13 Ma) and recent (1719–1721 AD) periods (e.g., [22, 31]).The Halaha volcanic field lies in the center of the Greater Xing’An Mountains (Figure 1(b)). The magmatism of the Halaha volcanic field was mainly distributed above the valley of Halaha River, Chaoer River, Chai River, and Dele River forming a low lava platform [32]. Previous studies have shown that the volcanic activities of the Halaha volcanic field erupted over a short period from 2.30 to 0.16 Ma [23, 52].Sixteen basalts of the Wudalianchi-Erkeshan volcanic fields and fourteen basalts of the Halaha volcanic field were sampled in this study. The Wudalianchi-Erkeshan basalts have typical porphyritic texture and contain 10% phenocrysts, which are primarily olivine and clinopyroxene. The matrix primarily comprises olivine, clinopyroxene, and some plagioclase (Figures 2(a) and 2(b)). The Halaha basalts also show typical porphyritic texture and contain 10%–20% phenocrysts, which are primarily olivine and minor clinopyroxene. The matrix primarily comprises olivine, clinopyroxene, and minor oxide minerals (Figures 2(c) and 2(d)). The descriptions of locations and ages for the studied samples are summarized in Supporting Information Table S1.Major-element analyses of olivine were performed using a JEOL JXA-8230 electron microprobe. The precision of all analyzed elements are better than 5%. Whole-rock major-element contents were determined using a wavelength X-ray fluorescence (XRF) spectrometer. The analytical uncertainties on major elements are generally better than 5%. Whole-rock trace-element analyses were performed by ICP-MS using a PerkinElmer Sciex ELAN 6000 instrument. For most of the trace elements, analytical precision and accuracy are better than 5%. Whole-rock Sr-Nd-Pb-Hf isotope analyses were performed with a Micromass Isoprobe MC-ICP-MS. The 87Sr/86Sr ratio obtained for the NBS SRM 987 standard was 0.710288±28 (2σ). The 143Nd/144Nd ratio obtained for the Shin Etsu JNdi-1 standard was 0.512115±7 (2σ). The 176Hf/177Hf ratio obtained for the JMC 475 standard was 0.282224±0.000019 (2σ). The measured Sr, Nd, and Hf isotope ratios were normalized to 86Sr/88Sr=0.1194, 146Nd/144Nd=0.7219, and 179Hf/177Hf=0.7325, respectively. Measured Pb isotopic ratios were corrected for mass fractionation based on repeated analyses of international standard NBS981. The mean 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb ratios of NBS981 were 16.932±6 (2σ, n=6), 15.484±6 (2σ, n=6), and 36.677±18 (2σ, n=6), respectively. Detailed analytical methods and the isotopic ratios of geological reference materials (BCR-2, BHVO-2) measured during the analytical procedure are given in Supporting Information Text 1 and Table S2, respectively.The results of whole-rock major- and trace-element concentrations and Sr-Nd-Pb-Hf isotopic compositions for the studied samples are given in Tables 1 and 2, respectively. Olivine compositions are presented in Supporting Information Table S3.In the total alkali versus silica diagram (TAS), the Wudalianchi-Erkeshan samples plot in the fields of basaltic trachyandesite and phonotephrite to tephriphonolite, whereas Halaha samples plot from the basalt to trachybasalt fields (Figure 3(a)). In the K2O versus Na2O diagram, the Wudalianchi-Erkeshan and Halaha basalts belong to potassic and sodic basalts, respectively (Figure 3(b)). The Wudalianchi-Erkeshan potassic basalts have lower MgO (5.9–6.7 wt.%), Cr (140–240 ppm), and Ni (104–160 ppm) contents than those of Halaha sodic basalts, which are characterized by higher MgO (9.8–11.9 wt.%), Cr (370–600 ppm), and Ni (231–433 ppm) contents.In the chondrite-normalized REE diagrams (Figures 4(a) and 4(b)), the extent of fractionation between LREE and HREE of the Wudalianchi-Erkeshan potassic basalts (La/YbN=42.9‐62.5) is higher than those of the Halaha sodic basalts (La/YbN=10.3‐14.9). In the primitive-mantle-normalized incompatible trace element spidergrams, the Wudalianchi-Erkeshan potassic basalts are characterized by significantly negative high-field-strength element (HFSE) (e.g., Nb and Ta) anomalies and pronounced positive anomalies in the large ion lithophile elements (LILE) (e.g., Rb and Pb) (Figure 4(c)). In contrast, the Halaha sodic basalts show Ba enrichment and positive Nb-Ta anomalies, resembling the typical OIBs (Figure 4(d)).The Wudalianchi-Erkeshan potassic basalts have 87Sr/86Sr values from 0.705214 to 0.705630, 143Nd/144Nd values from 0.512299 to 0.512321, and 176Hf/177Hf values from 0.282513 to 0.282609 within the EM1 (enriched mantle 1) ranges in the Sr-Nd-Hf diagrams. The Halaha sodic basalts have less radiogenic Sr (87Sr/86Sr of 0.703613–0.704223) but more radiogenic Nd (143Nd/144Nd of 0.512894–0.512928) and Hf (176Hf/177Hf of 0.282626–0.283100) isotopic ratios, comparable to OIBs (Figures 5(a)–5(c)).The Pb isotopic ratios of the Wudalianchi-Erkeshan potassic basalts (206Pb/204Pb=16.77‐16.90, 207Pb/204Pb=15.43‐15.45, and 208Pb/204Pb=36.76‐36.92) were less radiogenic than those of Halaha sodic basalts (206Pb/204Pb=18.27‐18.40, 207Pb/204Pb=15.52‐15.54, and 208Pb/204Pb=38.20‐38.45). In the 207Pb/204Pb versus 206Pb/204Pb diagram (Figure 5(d)), the Wudalianchi-Erkeshan potassic basalts plot above the 4.55 Ga geochron line, while the Halaha sodic basalts lie above the North Hemisphere Reference Line (NHRL) defined by Hart [53], respectively.In general, basaltic lavas contain two forms of olivine: (1) host magma-derived phenocrysts and (2) mantle-derived xenocrysts. The latter generally contain low concentrations of CaO (<0.1 wt.%), whereas olivine phenocrysts generally have high CaO contents (>0.1 wt.%; e.g., [54]). The majority of olivine xenocrysts are anhedral and may contain kink banding, whereas olivine phenocrysts are generally either euhedral or subhedral and are often free of kink banding (e.g., [55]; Kamenetsky et al., 2006). These criteria indicate that all of the olivine in the Wudalianchi-Erkeshan potassic basalts are phenocrysts, as they contain elevated concentrations of CaO (>0.1 wt.%) and possess either a euhedral or subhedral shape (Figure S1a). In comparison, the Halaha sodic basalts contain two groups of olivine. Group one olivines contain low concentrations of CaO (<0.1 wt.%) and are angular with kink-banded extinction patterns, suggesting that they may be xenocrysts (Figures S1b and S1c). Group two olivines contain high concentrations of CaO (>0.1 wt.%) and are euhedral, suggesting that they may be magmatic phenocrysts derived from the host magmas (Figure S1d).The olivine phenocrysts in the Wudalianchi-Erkeshan potassic basalts have a range of Fo (68.8–87.4), MnO (0.14–0.49 wt.%), CaO (0.11–0.35 wt.%), and NiO (0.09–0.39 wt.%) values. They have weak normal zoning that is characterized by Fo values and NiO concentrations that decrease, and they have CaO and MnO concentrations that increase from core to rim (Figure S2). Group one olivine xenocrysts within the Halaha sodic basalts have high Fo values (86.4–92.7) and contain low concentrations of CaO (0.01–0.84 wt.%; Figure S2b). In comparison, group two olivine phenocrysts within these basalts have a wider but slightly lower range of Fo values (72.0–89.6) and contain variable MnO (0.11–0.46 wt.%), CaO (0.10–0.36 wt.%), and NiO (0.11–0.40 wt.%) values. Group two olivine phenocrysts are also weakly normally zoned with Fo values and NiO concentrations decreasing and CaO and MnO contents increasing from core to rim (Figure S2).In order to better constrain the nature of the mantle source of the studied basalts, we first assess the possible effects, if any, of low-pressure processes such as crustal contamination and fractional crystallization on the chemical compositions of these basalts. Based on the following lines of evidence, we suggest that the crustal contamination was not significant during their generation. Firstly, both the potassic and sodic basalts display limited variations in the 87Sr/86Sri and εNdi isotopic ratios with decreasing MgO contents (Figures 6(a) and 6(b)). Secondly, the Nb/U ratios (40–64) of potassic and sodic basalts are similar to oceanic basalts (47±10) (Figure 6(c), [56]) and higher than continental crust ratios (Nb/U=6.2) [57], indicating that these basalts did not suffer significant continental contamination. Thirdly, mantle xenoliths are present in both the potassic and sodic basalts (e.g., [58, 59]), indicating that the host magmas ascended rapidly without significant interaction with the continental crust. Lastly, U-Th disequilibrium data show that the potassic basalts have strong 230Th excesses. This observation also indicates that the potassic basalts cannot suffer from significant crustal contamination. This is because assimilation of crustal rocks (230Th/238U=1.0) would reduce the extent of 230Th excesses of potassic basalts [34]. The Os isotopic data of the potassic basalts further indicate that they might suffer relatively minor amounts (3.5%) of lower crust contamination [35]. The sodic basalts display positive Nb and Ta anomalies in the primitive-mantle-normalized incompatible-element spidergrams (Figure 4(d)), also suggesting negligible contamination from the continental crust.The Wudalianchi-Erkeshan potassic basalts in this study have low values of MgO (5.9–6.7 wt.%), indicating that their parental magmas might have experienced a variable degree of differentiation. The positive correlations between Mg# versus Ni, CaO, and CaO/Al2O3 suggest that they have experienced fractional crystallization of olivine and clinopyroxene (Figures 7(a)–7(c)). Plagioclase fractionation is insignificant, as demonstrated to some degree by no positive correlation of Al2O3 with Mg# (Figure 7(d)) in concert with the absence of negative Eu and Sr anomalies in the REE patterns (Figures 5(a) and 5(c)).The Halaha sodic basalts in this study have relatively high MgO (9.8–11.9 wt.%) contents, suggesting insignificant crystal fractionation of the sodic basalts. The positive correlation between Mg# and Ni values indicates that the parental magmas of sodic basalts underwent olivine fractionation (Figure 7(a)). The relatively constant concentrations of CaO, CaO/Al2O3, and Al2O3 with a decreasing Mg# value indicate insignificant fractionation of clinopyroxene and plagioclase (Figures 7(b)–7(d)).Primary magmas of basaltic lavas can be used as probes of their mantle sources (e.g., [60]). Experimental studies have shown that basaltic melts can be generated from partial melting of peridotite (e.g., [61, 62]), pyroxenite (e.g., [63, 64]) and carbonated peridotite and/or pyroxenite [65, 66]. Here, the least-fractionated samples, which record only olivine fractionation, were used to calculate the primary-melt compositions of both the Wudalianchi-Erkeshan potassic basalts and the Halaha sodic basalts. The detailed methods and results are given in Supporting Information Text 2 and Table S4, respectively.Partial melting of carbonated mantle (peridotite and/or pyroxenite) commonly generates basaltic melts with high Zr/Hf ratios and large negative Zr and Hf anomalies in multielement variation diagrams [67–71]. These characteristics are not present with either the Wudalianchi-Erkeshan potassic basalts or the Halaha sodic basalts, which are characterized by low Zr/Hf values (38–43 for both potassic and sodic basalts; Table 1) and no negative Zr and Hf anomalies (Figures 4(c) and 4(d)). This suggests that both the potassic and sodic basalts may not have been derived from sources dominated by carbonated peridotite or pyroxenite material.Previous researches have suggested that potassic and sodic basalts within NE China were derived from peridotite-dominated sources (e.g., [23, 26, 30, 32]). This conclusion is consistent with the calculated primary magma compositions of the potassic and sodic basalts, which have the values of FeOT, Al2O3, and TiO2 similar to melts derived from partial melting of peridotite (Figures 8(a)–8(c)). However, the following lines of evidence argue that the mantle sources of the potassic and sodic basalts also contain pyroxenite: (1) Compared with basaltic magmas derived from pure peridotite sources, both the potassic and sodic basalts have low CaO concentrations at a given MgO value (Figure 8(d); [72]). (2) The potassic and sodic basalts have higher Fe/Mn but lower Fe/Zn ratios than melts derived from the partial melting of normal mantle peridotite (Figure 9; e.g., [73, 74]). (3) As olivine is the first silicate mineral to crystallize from mantle-derived magmas, olivine geochemistry can also provide insights into the mantle source of basalts (e.g., [75]). The olivine phenocrysts in the potassic and sodic basalts have higher Ni contents and Fe/Mn ratios, but lower Mn and Ca contents at a given Fo than those expected for olivine crystallized from melting of pure peridotite (Figure 10; [72]). We therefore propose that the mantle lithologies of the potassic and sodic basalts were mixed or came from hybrid sources containing both peridotite and pyroxenite.Both the Wudalianchi-Erkeshan potassic basalts and the Halaha sodic basalts define mixing trends between depleted and enriched end-members within Sr-Nd-Pb-Hf isotopic plots (Figure 5). The depleted end-member for the late Cenozoic basalts in NE China is commonly thought to be DMM (e.g., [33]). However, studies indicate that a depleted component may also be present within the deep mantle, which could contribute to the production of oceanic basalts (e.g., [76–79]). As all modern terrestrial lavas have 142Nd/144Nd values that are about 18 ppm higher than those in ordinary chondrites (e.g., [80–82]), ancient primitive mantle is argued to have been both chondritic and nonchondritic [81, 83, 84]. It has been further suggested that this ancient primitive mantle has superchondritic 142Nd/144Nd values and represents an early precursor for all modern terrestrial mantle reservoirs (e.g., [85]). The so-called “FOZO,” which has also been defined as “PREMA” [86], “PHEM” [87], and “C” [88, 89], is a less primitive but ubiquitous reservoir within the mantle that is thought to be a mixture of recycled components (e.g., oceanic crust) and early isolated ancient primitive mantle [90]. Thus, it is highly likely that both ancient primitive mantle and FOZO could provide the depleted components for the production of basalts (e.g., [85, 90–92]). Based on radiogenic isotopic data (e.g., Sr-Nd-Pb) from late Cenozoic basalts, previous studies have suggested that the depleted end-member in the mantle source of late Cenozoic basalts in NE China could be ancient primitive mantle material or FOZO [30, 93]. However, it is not a simple task to discriminate between DMM, ancient primitive mantle, and FOZO reservoirs simply by using radiogenic isotopic values (e.g., Sr-Nd-Pb) alone (Figure S3). A more definitive discrimination requires the use of noble-gas isotopes. DMM-derived basalts (MORBs) have nearly constant 3He/4He values (8 Ra, where Ra is the atmospheric 3He/4He ratio; [94]), whereas basalts derived from relatively primordial ancient primitive mantle or FOZO reservoirs should have higher 3He/4He values (>30 Ra; [85, 91, 95]; Garapić et al., 2015).The low 3He/4He values (<8 Ra) of late Cenozoic basalts and associated mantle xenoliths within NE China ([93] and references therein) suggest that these magmas were not derived from a deep-seated mantle plume. This is consistent with a lack of geophysical evidence of a mantle plume in this region during the late Cenozoic (e.g., [96]). These observations suggest that the depleted component in the mantle source of the late Cenozoic basalts (e.g., the Wudalianchi-Erkeshan potassic basalts and the Halaha sodic basalts) is most likely DMM rather than ancient primitive lower mantle or FOZO.The nature of the enriched end-members in the mantle source of these basalts remains contentious, with proposed sources including delaminated ancient lower crust [35, 39], recycled SCLM [31, 33, 34], and recycled oceanic materials [24–27, 29, 30, 37, 38, 40, 41]. In the following section, we focus mainly on the enriched end-members in the mantle sources of Wudalianchi-Erkeshan potassic basalts and Halaha sodic basalts.The Wudalianchi-Erkeshan potassic basalts have high Ba, Sr, and LREE contents as well as depletion in HFSEs and extremely unradiogenic Pb isotopic compositions (206Pb/204Pb=16.77‐16.90), indicating that the enriched component in the mantle source has a clear EM1-like signature. Previous studies have proposed that this signature originated mainly from the delaminated ancient lower continental crust of NE China [35]. However, it should be noted that the average K2O content of the lower continental crust is only ~0.61 wt.% [57]. Accordingly, it is difficult to conceive that lower continental crust with such a low K2O content could modify the asthenospheric mantle and then produce the Wudalianchi-Erkeshan potassic basalts with high K2O contents (4.6–5.8 wt.%). Also, the mantle xenoliths hosted in the volcanic rocks with Paleoproterozoic model ages argue against delamination of the lower crust having occurred in this area during the Cenozoic [97]. Therefore, the EM1-like signature of the Wudalianchi-Erkeshan potassic basalts did not originate from the delaminated ancient lower continental crust.The ancient metasomatized lithospheric mantle and recycled ancient sediment residing in the mantle transition zone (410–660 km depth) have been the two competing origins of the EM1-like signature of the Wudalianchi-Erkeshan potassic basalts [26, 27, 29, 31, 33, 34, 41]. This is because both of them could account for the high K2O, Sr, and Pb contents (Figure 4(b)), Ba/La (18–26) and Ba/Th (220–352) values (Figure 11), and unradiogenic Pb isotopic compositions of potassic basalts. Based on the following two main observations, we propose that the EM1-like signature of the Wudalianchi-Erkeshan potassic basalts is most likely derived from recycled ancient sediment residing in the mantle transition zone: (1) Many mantle xenoliths from NE China show that the lithospheric mantle had moderately depleted Sr-Nd-Hf isotopic compositions [98–101], which could not have produced the EM1-type isotopic characteristics observed in the Wudalianchi-Erkeshan potassic basalts. Actually, this observed EM1 signature is found not only in NE China but also in many other volcanic fields in the North China Craton (e.g., [70, 102, 103]). Geochemical, geophysical and petrological data have indicated that the lithospheric mantle beneath the present-day North China Craton has been mostly removed and replaced by a hot, thin, juvenile SCLM [104–111], which could not have provided the observed EM1 signature in the Cenozoic basalts of the North China Craton. Although it has been discovered that there might have been a small portion of Archean lithospheric mantle beneath NE China (e.g., Keluo, [97]) and NCC (e.g., Hebi, Fanshi, [112, 113]), many studies have proven that these ancient mantle peridotites have been refertilized by asthenosphere-derived melts and their EM1 compositions might have been diluted (e.g., Zhang et al., 2009, [114–116]). Besides, even if these ancient lithospheric mantle peridotites were characterized with an EM1 signature, it appears to be volumetrically unrealistic that the small portion of ancient lithospheric mantle beneath several local places could be the source of EM1 basalts, which were widely distributed in NE China and NCC during the Cenozoic (e.g., [51, 117]). (2) The low 206Pb/204Pb values (16.77–16.90) of the Wudalianchi-Erkeshan potassic basalts indicate that their source should have very low 238U/204Pb values (as low as ~2), which would severely retard the subsequent production of radiogenic Pb isotopes in a prolonged (2–3 Gyr) period [118, 119]. The estimated high water contents (up to 4.5 wt.%) of the parental magmas of the Wudalianchi-Erkeshan potassic basalts require that their mantle source should be highly hydrated [41, 120]. This above discussion implies that the mantle source of potassic basalts must be hydrated and isolated for a long period (2–3 Gyr) from mantle convection. The Archean fluid-metasomatized SCLM could satisfy the long-term isolation of low-μ material that is required to generate an EM1-like signature (e.g., [31]). However, because the addition of water into the SCLM would have significantly decreased its viscosity [121, 122], the hydrated SCLM cannot remain isolated for a long period of time from the convective upper mantle [122]. In contrast, the mantle transition zone is likely to be the slab “graveyard” (e.g., [123]) and the recycled sediment can be isolated for a long period of time (~2.0 Ga) in the mantle transition zone (e.g., [124]). In addition, as the mantle transition zone has a considerably high water content [125], the recycled ancient sediment from the mantle transition zone could add large amounts of water to the mantle source of potassic basalts. Therefore, we propose that the recycled ancient sediment from the mantle transition zone is the most likely origin of the EM1 signature of the Wudalianchi-Erkeshan potassic basalts.The Halaha sodic basalts have positive Nb and Ta anomalies and yield Ce/Pb and Nb/U ratios of 18–25 and 46–64, respectively, resembling typical OIBs (Figure 4(d); [56]). Previous studies indicate that the likely mantle sources of OIBs are lithospheric mantle with amphibole-bearing metasomatic veins, asthenospheric mantle, or a mantle plume [126–130].As we argue in Section 5.3, there is no evidence that supports the presence of a deep mantle plume beneath NE China during the late Cenozoic. Thus, the mantle plume model cannot explain the mantle source of the Halaha sodic basalts. The model involving a recycled metasomatized lithospheric mantle proposes that the hornblendite veins that formed during the percolation and differentiation of volatile-bearing melts within the lithospheric mantle could be a candidate source of OIBs [127, 128, 131]. However, metasomatized minerals such as amphibole and phlogopite are not observed in the mantle peridotite xenoliths (including harzburgites and lherzolites) hosted in the Halaha sodic basalts [59]. In addition, if the high Nb-Ta contents were contributed by the Nb-rich minerals (e.g., rutile) in the mantle lithosphere, they could also host Zr-Hf, thereby producing positive Zr-Hf anomalies (e.g., [42]), which are not observed in the multielement variation diagram for the Halaha sodic basalts (Figure 4(d)). Therefore, the recycled lithospheric mantle is unlikely to be the enriched component in the mantle source of the Halaha sodic basalts.It is likely, therefore, that the Halaha sodic basalts originated from the asthenospheric mantle. However, these sodic basalts are enriched in LILEs and LREEs (Figures 4(c) and 4(d)) and have enriched Sr-Nd-Hf-Pb isotopes (Figure 5) compared to MORBs. These differences indicate that the normal asthenospheric mantle cannot have served directly as the source of these OIBs. Previous studies suggest that the recycled oceanic crust is a common enriched component in the mantle sources of OIBs (e.g., [126, 129]). Based on the following lines of evidence, we favor the recycled oceanic crust as the most likely enriched component in the mantle source of the Halaha sodic basalts. As slab dehydration occurred during the subduction of the oceanic crust in the rutile stability field, fluid-mobile elements (Th and U, as well as Ba, Rb, Cs, and Sr) were transported to peridotite in the mantle wedge, while HFSEs (e.g., Nb, Ta, and Ti) remained in the residual oceanic crust (e.g., [132]). Fluid-fluxed mantle-wedge peridotites were generated above the subducting slab, and their partial melting produced oceanic arc basalts (OABs) that are enriched in fluid-mobile elements (LILE, LREE, and Th) but depleted in Nb, Ta, and Ti [133, 134]. Compared with OABs, the residual oceanic crust after dehydration is always characterized by high Ta/UN, Nb/ThN, Nb/U, and TiO2/Al2O3 ratios (e.g., [135, 136]). Thus, the Halaha sodic basalts with high Ta/UN (>1), Nb/ThN (>1), Nb/U (46–64), and TiO2/Al2O3 (0.16–0.18) ratios were most likely derived from a mantle source that contained a contribution from the recycled oceanic crust with the breakdown of rutile (Figure 12). In addition, the high values of δ18O in the Halaha sodic basalts (cpx up to 8.57‰; [38]) also require the recycled oceanic crust and/or sediment to be involved in their mantle source. This is because the recycled oceanic crust and/or sediment could have high δ18O values as a result of undergoing low-temperature (<350°C) water-rock interaction [137–139]. A high-velocity representative of the stagnant segment of the subducted Pacific plate in the mantle transition zone beneath eastern China has been imaged by P and S wave tomography [96, 140, 141]. According to the geophysical data, the western edge of the stagnant Pacific slab has reached the mantle transition zone beneath the Daxing’Anling-Taihangshan Gravity Lineament in eastern China, and the Halaha volcanic field is located close to the west of the Daxing’Anling-Taihangshan Gravity Lineament. Therefore, the subducted Pacific oceanic crust in the mantle transition zone most likely provided the enriched component in the mantle source of the Halaha sodic basalts. Other volcanic fields, including the Jining, Abaga, and Wulanhada basalts to the west of the Daxing’Anling-Taihangshan Gravity Lineament, are also considered to be closely associated with the deep subduction of the Pacific oceanic plate [48–50, 142]. The Halaha sodic basalts have high Ba/La (14–19) and Ba/Th (102–143) ratios, suggesting that their mantle source might have a contribution from recycled sediment (Figure 8). However, the low 143Nd/144Nd and 176Hf/177Hf ratios and negative Nb and Ta anomalies of recycled subducted sediment do not match those of the Halaha sodic basalts (e.g., [143, 144]), indicating that the contribution of subducted sediment to the mantle source of the Halaha sodic basalts is negligible. Therefore, we propose that the enriched component in the mantle source of the Halaha sodic basalts is Pacific oceanic crust with negligible sediment, which also resided in the mantle transition zone.In summary, the enriched component (ancient sediment) in the mantle source of the Wudalianchi-Erkeshan potassic basalts is probably related to the stagnation of an ancient subducted slab, whereas the enriched component (Pacific oceanic crust) in the mantle source of the Halaha sodic basalts is likely linked to the stagnation of the recently subducted Pacific oceanic slab. This conclusion is illustrated by the plots of Ba/Th ratios versus 143Nd/144Nd and 206Pb/204Pb ratios (Figure 13), which also suggest that the enriched end-members are ancient sediment for the potassic basalts and Pacific oceanic crust for the sodic basalts.In several areas in the Basin-and-Range Province and the East African rift, late Cenozoic temporospatially related potassic and sodic basalts were generated from enriched lithosphere and sublithosphere (asthenospheric mantle or mantle plume), respectively [14, 15, 17, 18]. The transition from potassic to sodic basalts in these areas was interpreted as a direct response to lithospheric extension. In detail, initial extension triggered melting of the enriched lithosphere to produce the potassic basalts, which was followed by the decompressional melting of upwelling sublithospheric mantle to produce the sodic basalts. Alternatively, both the potassic and sodic basalts were produced by partial melting of the delaminated enriched lithosphere [16]. The potassic basalts were developed at higher pressures from deeper fragments of foundered lithosphere veined by a mica-bearing, Al-poor assemblage, whereas the sodic rocks were generated at lower pressures from shallower fragments of foundered lithosphere veined by an aluminous amphibole-bearing assemblage. As discussed in Section 5.4, the mantle sources for the Wudalianchi-Erkeshan potassic basalts and Halaha sodic basalts are suggested to be the asthenospheric mantle with recycled oceanic components without the involvement of the lithospheric mantle. There is no direct evidence of lithospheric mantle delamination having occurred in the study area during the late Cenozoic. Thus, the late Cenozoic coeval potassic and sodic basalts in NE China cannot have been formed in the above way.Coeval potassic and sodic basalts have also been reported in the Hong’an-Dabie orogen in eastern China. The change from potassic basalts (mafic intrusive rocks) to sodic basalts (mafic dikes) in the Hong’an-Dabie orogen has been attributed to mantle sources with distinct recycled crustal materials (oceanic crust versus continental crust) [19–21]. In the following discussion, we suggest that partial melting of the mantle source with different recycled oceanic materials (ancient sediment versus Pacific oceanic crust) could have produced the coeval Wudalianchi-Erkeshan potassic basalts and Halaha sodic basalts.Experimental studies have shown that partial melting of fertile peridotite fluxed by hydrous sediment can generate (ultra)potassic magmas [145]. This hydrous sediment is generally enriched in K2O, LILEs, and Sr-Nd-Hf radiogenic isotopic compositions but depleted in HFSEs (e.g., [143, 144]). The partial melting of a mantle source with recycled sediment would generate (ultra)potassic rocks with such features. The Pb isotopic composition of (ultra)potassic rocks is strongly affected by the duration of isolation of the possible sources, where short isolation causes samples to plot to the right of the Geochron (Figure 5(d)), and long isolation (i.e., from ancient subduction events) causes samples to plot to the left of the Geochron as a result of the significant retardation of 206Pb/204Pb ratios [124]. For example, (ultra)potassic rocks from the western Mediterranean and southern Tibet are interpreted as having been derived from a mantle source with recently recycled sediment and plot to the right of the Geochron (Figure 5(d); [5–9, 146]). In contrast, the (ultra)potassic rocks from Western Australia, Leucite Hills, and Gaussberg plot to the left of the Geochron and are interpreted as having been derived from the mantle transition zone with ancient sediment that was isolated for more than ~2 Gyr (Figure 5(d); [124, 147–149]). The Wudalianchi-Erkeshan potassic basalts have 206Pb/204Pb ratios similar to those of the (ultra)potassic basalts from Western Australia. Thus, we interpret that this isotopic signature is likely to have been introduced to the deep mantle during an ancient subduction event (i.e., ancient sediment).Because rutile is a key mineral for storing HFSEs, partial melting of the oceanic crust with the breakdown of rutile will give rise to felsic melts in which Nb and Ta are enriched [150, 151]. Meanwhile, the oceanic crust could have high Na2O/K2O values and depleted the Sr-Nd-Hf radiogenic isotopic compositions (e.g., [144, 152]). Thus, partial melting of the mantle source with recycled oceanic crust with the breakdown of rutile could generate sodic basalts with OIB-like trace-element distribution patterns (i.e., enrichment of LILEs and LREEs but with no depletion of HFSEs) and depleted radiogenic Sr-Nd-Hf isotopic signatures. Experimental studies have further shown that partial melting of a mixed source of oceanic crust and fertile peridotite could generate sodic basalts with OIB-like characteristics (e.g., [153, 154]). The Pb isotopic system of sodic basalts with OIB-like characteristics is sensitive to variations in the age of the recycled oceanic crust and can be divided into basalts of typical HIMU islands such as St. Helena, Tubuaii, and Mangaia (206Pb/204Pb=20.5‐22; [155, 156]) and basalts of young HIMU islands such as La Palma and El Hierro (Canary Islands) (206Pb/204Pb=18.8‐20; Thirlwall et al., 1997; [157, 158]). The Pb isotopic difference between the typical and young HIMU island sources depends on the age of the recycled materials and the presence of a small amount of sediment (e.g., [159]). The typical HIMU basalts with high 206Pb/204Pb values indicate that oceanic crust was recycled into their mantle source at least 2 Gyr ago (Hofmann, 1988; [160]). In contrast, the young HIMU basalts with relatively low 206Pb/204Pb values require recycled oceanic crust to have been incorporated into their mantle more recently [159, 161, 162] because the recently recycled oceanic crust in the mantle would have had insufficient time to produce high Pb isotopic ratios. If a source contained oceanic crust without the long time-integrated ingrowth of Pb isotope systems, a resulting basaltic melt that is less enriched in radiogenic Pb isotopes is possible. The Halaha sodic basalts have radiogenic 206Pb/204Pb (18.27–18.40) values that are slightly lower than those of the young HIMU basalts (18.8–20.0), suggesting that the recycled oceanic crust in their mantle source is related to a recent subduction event (Pacific oceanic crust).We therefore propose that partial melting of the ambient DMM with recycled ancient sediment would produce the Wudalianchi-Erkeshan potassic basalts, whereas partial melting of the ambient DMM and recycled Pacific oceanic crust would develop the Halaha sodic basalts. The transition from the potassic to the sodic basalts in NE China could be a result of the difference in the recycled components incorporated into their mantle sources, namely recycled ancient sediment and recent Pacific oceanic crust, respectively.The deep-Earth cycling of fluids [125, 163, 164] due to the presence of water-enriched wadsleyite and ringwoodite in the mantle transition zone [125, 165–167] may cause Rayleigh-Taylor instabilities and generate wet plumes [168, 169]. Once wet plumes are generated, they could transport oceanic materials (ancient sediment and recent Pacific crust) from the mantle transition zone into the overlying upper asthenospheric mantle (Figure 14(a)). When these recycled oceanic materials were transported into the asthenospheric mantle, they could melt preferentially to the ambient peridotitic mantle owing to the recycled oceanic components having a lower solidus than peridotite [170, 171]. The recycled melt derived from oceanic components could react with mantle peridotite and convert it to pyroxenite by forming orthopyroxene at the expense of olivine (e.g., [154, 172, 173]). This is consistent with the mantle lithologies inferred in this study for both potassic and sodic basalts in NE China, which have mixed sources of peridotite and pyroxenite. The recycled ancient sediment would react with the asthenospheric mantle peridotite in generating the EM1-like pyroxenite, providing the enriched component for Wudalianchi-Erkeshan potassic basalts (Figure 14(b)), while the recycled Pacific oceanic crust would react with the asthenospheric mantle peridotite to generate young HIMU-like pyroxenite, providing the enriched component for Halaha sodic basalts (Figure 14(b)). Therefore, the asthenospheric mantle beneath NE China was not homogeneous and these two types of pyroxenite (EM1-like and young HIMU-like) may be distributed as blobs, veinlets, or streaks within the upper asthenospheric mantle, resembling a “plum-pudding” structure (e.g., [174]). During the Cenozoic, NE China was in a continental extensional setting due to the rollback of the Pacific slab and associated trench retreat [47]. In this extensional setting, partial melting of the upwelling heterogeneous asthenospheric mantle comprising ambient DMM and the two types of pyroxenite produced the Wudalianchi-Erkeshan potassic basalts and Halaha sodic basalts, respectively (Figures 14(c) and 14(d)).The asthenospheric component is commonly considered to have been depleted by previous melting events but subsequently homogenized by convective stirring (e.g., [86]). However, the deduced mantle sources (ambient DMM, two types of pyroxenite) for the potassic and sodic basalts indicate that the upper asthenospheric mantle beneath NE China is highly heterogeneous. Recent studies of early Cenozoic basalts in the Songliao Basin have concluded that chemical heterogeneity was present in the asthenospheric mantle beneath this basin [37, 42]. These combined results show that the upper asthenospheric mantle beneath NE China may have been temporospatially heterogeneous during the Cenozoic.Various lines of evidence suggest that the upper asthenospheric mantle is indeed incompletely homogenized on a global scale. First, MORBs in the Indian and the southern Atlantic oceans have distinctive isotopic compositions with relatively high 87Sr/86Sr and 208Pb/204Pb values for a given 206Pb/204Pb value, referred to as the Dupal anomaly [53, 175]. The Dupal anomaly was initially considered to imply the existence of hemispheric-scale fertile geochemical heterogeneity in the southern hemisphere upper asthenospheric mantle. Subsequent studies established that the Dupal anomaly signature was also present along the mid-ocean ridges of the northern Atlantic [176, 177] and Arctic [178] oceans. The Dupal anomaly requires an enriched component to be dispersed in the upper asthenospheric mantle. This enriched component has been suggested as being either delaminated recycled continental crust and/or ancient SCLM during the early rifting stages of Gondwana (e.g., [178–183]) or lower continental crust and recycled pelagic sediment together with altered oceanic crust from a deep mantle plume, possibly at the edge of a large low-shear-velocity province (e.g., [184–187]). Second, decoupling of Hf and Nd isotopes in some MORBs indicates that the asthenospheric mantle has a component of ancient refractory domains, which were probably characterized by an extremely enriched Hf isotopic composition (e.g., [188]). The observed very depleted 187Os/188Os (<0.12) and εHf (up to 104) compositions found in some abyssal peridotites [189–192] are considered as solid evidence for the presence of such ancient refractory mantle domains in the upper asthenospheric mantle. These ancient residual mantle domains in the upper asthenospheric mantle are argued to be either ancient residual oceanic lithosphere (ReLish) or SCLM reservoirs that underwent previous depletion events followed by carbonatite-type metasomatism [188, 193–195].The heterogeneity of the upper asthenospheric mantle can thus be attributed to at least three components: a normally depleted MORB mantle (DMM), an ancient residual lithospheric mantle (ReLish or SCLM), and enriched recycled oceanic components as inferred in the present study (Figure 15). Therefore, the asthenospheric mantle is not a well-stirred and chemically homogenized mantle but could be highly heterogeneous on both local (e.g., NE China) and global scales. As a result, using MORBs or abyssal peridotite for estimating the average isotope compositions of the upper asthenospheric mantle needs to be undertaken with caution.The late Cenozoic Wudalianchi-Erkeshan potassic basalts and Halaha sodic basalts were coeval in NE China. The potassic and sodic basalts originated mainly from DMM and an enriched end-member. The enriched end-member in the mantle source of the potassic basalts has an EM1-like signature with relatively low 143Nd/144Nd, 176Hf/177Hf, and 206Pb/204Pb contents and negative Nb and Ta anomalies but high Ba/La and Ba/Th values, whereas the enriched end-member in the mantle source of the sodic basalts has high Ta/UN, Nb/ThN, radiogenic 143Nd/144Nd, 176Hf/177Hf, and 206Pb/204Pb values. We suggest that ancient sediment and recent Pacific oceanic crust in the mantle transition zone are the respective enriched end-members in the mantle sources of the potassic and sodic basalts. As wet plumes ascend through the upper mantle, they may bring up preexisting recycled materials (ancient sediment and Pacific oceanic crust) from the mantle transition zone into the above asthenospheric mantle. These recycled materials could melt preferentially relative to the ambient peridotitic mantle to produce silicic melts, which would react with mantle peridotite to produce the two kinds of mantle pyroxenite (EM1-like and young HIMU-like pyroxenite).The partial melting of the upper asthenospheric mantle comprising the ambient DMM and these two kinds of pyroxenite produced the potassic and sodic basalts in NE China, respectively, which was triggered by a rollback of the deeply subducted Pacific slab during the late Cenozoic. The change of the recycled component from ancient sediment to Pacific oceanic crust within the upper asthenospheric mantle beneath NE China played an important role in producing the coeval potassic and sodic basalts.The authors declare that they have no conflicts of interestWe would like to thank Xinyu Wang, Hongxia Yu, Xirong Liang, and Le Zhang for their help with the experiments. This research was financially supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB26000000), the Key Research Project of Frontier Sciences of the Chinese Academy of Sciences (QYZDY-SSW-DQC030), and grants from the Natural Science Foundation of China (41702361 and 41572321).The supporting information presents the detailed analytical methods (Text S1); the calculated methods of the primary melts of potassic and sodic basalts (Text S2); the backscattered electron images of olivines in potassic and sodic basalts (Figure S1); the (a) NiO (wt.%) versus Fo numbers, (b) CaO (wt.%) versus Fo numbers, and (c) MnO (wt.%) versus Fo numbers for potassic and sodic basalts (Figure S2); (a) 143Nd/144Nd versus 87Sr/86Sr and (b) 207Pb/204Pb versus 206Pb/204Pb for DMM, FOZO, and the ancient primitive mantle (Figure S3); the sample locations, ages, and dating methods for potassic and sodic basalts (Table S1); the results of Sr-Nd-Hf-Pb isotopes for standard reference materials (Table S2); the olivine compositions of potassic and sodic basalts (Table S3); and the results of primary magmas for potassic and sodic basalts (Table S4).
更新日期：2020-09-01
京公网安备 11010802027423号