Analysis of depth-diameter relationship of craters around oceanus procellarum area

0. INTRODUCTION

Emeishan large igneous province (ELIP) is widely distributed in three provinces of Sichuan, Yunnan, and Guizhou in southwestern China, with its erupted area exceeding 2.5 × 10⁵ km² (Fig. 1, Chung and Jahn, 1995). In the ELIP, the continental flood basalts are the dominant volcanic rocks, even though a few picrites, rhyolites, basaltic andesites and pyroclastics are locally exposed. In the past decades, involving geochemistry, isotopic compositions (e.g., Sr-Nd-Pb-Os-Lu-O-Mg) and zircon U-Pb dating of these basalts, have been reported (Fu et al., 2021; Ji et al., 2021; Tian et al., 2017; Zhong et al., 2014; Lai et al., 2012; Shellnutt and Jahn, 2011; Yan et al., 2010; Fan et al., 2008; Li et al., 2008; Qi et al., 2008; Xu Y G et al., 2008, 2001; Xu J F et al., 2007; Zhou et al., 2006, 2002; Xiao et al., 2004; Song et al., 2001). Xu et al. (2001) and He et al. (2010) divided the basalts from the ELIP into high-Ti and low-Ti basalts in terms of their bulk TiO₂ contents, Ti/Y ratios and a linear relationship (TiO₂ = -0.08MgO + 2.91). Except for Ti, both of them also display distinct geochemical and Sr-Nd-Pb isotopic compositions. For instance, the high-Ti basalts have lower Mg^#, (⁸⁷Sr/⁸⁶Sr)_i, CaO, MgO, and higher Sm/Yb ratios, ε_Nd(t), and Total FeO, K₂O, Na₂O contents relative to those in low-Ti basalts (Liu et al., 2020; Tian et al., 2017; Jiang et al., 2009; Xiao et al., 2004; Song et al., 2001; Xu et al., 2001). These compositional variations suggest that they might be affected by different magmatic processes. Xu et al. (2001) proposed that the geochemical heterogeneity in both low-Ti and high-Ti basalts might be derived from the same mantle plume, possibly accompanied by variable extents of contamination. Xu et al. (2007) further advocated that some subcontinental lithospheric mantle (SCLM) materials have greatly contributed to the formation of the high-Ti basalts during the magma ascent. In a study of the basalts from the Panzhihua area, partial melting and crustal contamination appeared to be two major processes that controlled the compositional diversity of the high-Ti and low-Ti basaltic magmas (Shellnutt and Jahn, 2011). Nevertheless, Lai et al. (2012) believed that the basalts with high Ti/Y ratios from the Yunnan and Guizhou provinces probably originated from the SCLM of the Yangtze Block with minor contamination of crustal materials or metasomatic veins of the lithosphere heated by the plume. Recently, Li et al. (2016) also claimed that crustal contamination exerted a limited effect on the genesis of the Permian basalts in the Sichuan Province on the basis of their geochemical and Sr-Nd isotopic features. Therefore, the role of crustal contamination, needs to be re-appraised in order to promote a better understanding of the genesis of the ELIP.

Figure  1.  Simplified regional geological map of the Emeishan large igneous province (modified from Shellnutt and Pham, 2018). Sample locations are identified.

DownLoad: Full-Size Img PowerPoint

As an incompatible element, Ba is readily incorporated into magma during the partial melting of the mantle. Compared to the mantle, crustal rocks are usually abundant in Ba. The average Ba content in the primitive mantle (PM) is about 6.9 ppm (Palme and O'Neill, 2014; McDonough and Sun, 1995), which is obviously lower than that in continental crusts. The latter is as high as 456 ppm, as reported by Rudnick and Gao (2014). Furthermore, it is generally believed that both low-temperature (e.g., mineral precipitation and adsorption) and high-temperature (e.g., crystal-melt separation) processes can lead to Ba isotopic fractionation (Deng et al., 2021; Gou et al., 2020; Gong et al., 2019; Böttcher et al., 2012; von Allmen et al., 2010). The δ^138/134Ba values in the PM are homogeneous; however, crustal materials have a wide range from -0.63‰ to 0.47‰ (An et al., 2020; Li et al., 2020; Nan et al., 2018; Nielsen et al., 2018). Accordingly, the Ba isotopic system is expected to be a sensitive tracer of crust recycling and crust-mantle interactions (Li et al., 2020). Furthermore, Nielsen et al. (2018) reported that the Ba isotopic discrepancy in global mid-ocean ridge basalts (MORB) could probably be attributed to the variable input of sedimentary materials into the mantle, making it a reliable tool for tracking the genesis of basaltic rocks.

In this study, based on the whole-rock geochemical compositions, we present new Ba isotopic evidence of the low-Ti and high-Ti continental flood basalts from the Lijiang, Miyi and Emeishan regions in order to discuss their petrogenesis. This study further provides isotopic constraints on the role of crustal contamination in the magmatic evolution of basaltic melts from the mantle.

1. GEOLOGICAL BACKGROUND AND SAMPLE DESCRIPTION

The ELIP is extensively distributed along the western margin of the Yangtze Block, SW China, with significant variations in the thickness of volcanic sequences (several hundred meters in the east and nearly 5 km in the west) (Xiao et al., 2004; Song et al., 2001; Xu et al., 2001; Chung and Jahn, 1995). The detailed geology of the ELIP has been described in previous studies (Xu et al., 2013, 2004, 2001; Ali et al., 2010, 2005; Fan et al., 2008; Wang et al., 2007; He at al., 2006, 2003; Song et al., 2006; Zhang et al., 2006, 2004; Zhou et al., 2006). Here, we offer a brief description as follows: the ELIP comprises abundant continental flood basalts, minor picrites, basaltic andesites, rhyolites, trachytes, as well as associated mafic-ultramafic and syenitic intrusions (Fig. 1, Xu et al., 2001; Chung and Jahn, 1995). The continental flood basalts are exposed in a large area of over 2.5 × 10⁵ km² (Fig. 1, Chung and Jahn, 1995). The mineral assemblage includes plagioclase, clinopyroxene, olivine, basaltic glass, and minor amounts of magnetite, and ilmenite. In particular, the aphyric, porphyritic, and amygdaloidal basalts dominate in the ELIP (Fu et al., 2021; Tian et al., 2017; Jiang et al., 2007; Wang et al., 2007; Xu et al., 2001). According to their bulk TiO₂ contents and Ti/Y ratios, the continental flood basalts from the ELIP can be classified into two groups: high-Ti basalts (TiO₂ > 2.5, Ti/Y > 500) and low-Ti basalts (TiO₂ < 2.5, Ti/Y < 500). These two groups show distinct petrologic and geochemical characteristics (Tian et al., 2017; You and Liu, 2014; Xu Y G et al., 2013, 2001; Xu J F et al., 2007). The low-Ti basalts contain more clinopyroxenes and less magnetites, whereas the high-Ti basalts are dominated by plagioclases with minor amounts of clinopyroxene phenocrysts; the high-Ti basalts have lower Mg^# (0.32-0.61) and (⁸⁷Sr/⁸⁶Sr)_i (0.704 9-0.706 4), but higher Sm/Yb ratio, ε_Nd(t) (-0.71-1.5) and γOS(t) (-1.4- -0.8) compared to the low-Ti basalts (Mg^# = 0.52-0.64, (⁸⁷Sr/⁸⁶Sr)_i = 0.706 3-0.707 8, ε_Nd(t) = (-6.74- -0.3), γOS(t) = 6.5). Regarding spatial distribution, the inner zone, comprises low-Ti and high-Ti basalts, however, the outer zone, is merely confined to high-Ti basalts (Tian et al., 2017). As shown above, the geology, geophysics, stratigraphy, sedimentology, petrology, geochemistry and Sr-Nb-Pb isotope signatures of the ELIP supported that a mantle plume potentially made a great contribution to the generation of the Emeishan volcanism (He et al., 2006, 2003; Zhang et al., 2006, 2004; Xu et al., 2004, 2001).

Eighteen continental flood basaltswere collected from Lijiang, Miyi and Emeishan in the ELIP for detailed analyses (Fig. 1). Based on petrographic observations, the basalts mainly comprised plagioclases, clinopyroxenes, olivines and a few opaque minerals (e.g., magnetite). Most of the basaltic samples are relatively fresh with a little secondary alteration. In terms of their bulk TiO₂ contents and Ti/Y ratios, high-Ti basalts were only observed in Miyi and Emeishan, whereas low-Ti basalts dominantly distributed in Lijiang (Section 3.1). The basalts are mostly cryptocrystalline, and some high-Ti samples typically exhibit porphyritic textures in the presence of plagioclase phenocrysts (Fig. 2).

Figure  2.  Photographs (a), (b) and petrographic features (c), (d) of the basalts from the ELIP. Pl. Plagioclase.

DownLoad: Full-Size Img PowerPoint

2. METHODS

2.1 Major and Trace Elements

The whole-rock major oxides (e.g., MgO, SiO₂, TiO₂, Al₂O₃, K₂O, etc.) of the studied basalts (wt.%) were measured by gravimetry (wet chemistry) and X-ray fluorescence (XRF) at University of Science and Technology of China, Hefei (USTC). The concentrations of trace elements (including rare earth elements, REE) were analyzed by inductively coupled plasma mass spectrometer (ICP-MS) at Key Laboratory of Crust-Mantle Materials and Environments, Chinese Academy of Sciences (CAS), Hefei. Two standard materials (BCR-2 and BHVO-2) were used for quality monitoring. The uncertainties are better than 1.0% for major elements and 5.0% for trace elements.

2.2 Barium Isotope

Barium isotopic compositions of the basalts were performed at CAS Key Laboratory of Crust-Mantle Materials and Environments, University of Science and Technology of China. More details about the procedure of sample digestion, chemical separation and purification have been already published by Nan et al.(2018, 2015) and An et al. (2020). During dissolution and purification processes, the basalt powders were dissolved in double-distilled high purity acids (HF, HNO₃ and HCl) in Teflon screw-top beakers with 18.2 MΩ/cm ultra-pure water. After that, barium was purified from the matrix by cation exchange column with AG50W-X12 resin (200-400 mesh, Bio-Rad, USA). The twice purification process is performed twice, firstly using a column with 2 mL resin followed by the second column with a 0.5 mL resin. The 3 mol/L HCl was utilized to elute matrix elements, and then 3 mol/L HNO₃ was added to collect barium. The procedural blank was 4.2 ng, which is negligible (< 1‰) compared with the amount of Ba analyzed from the samples (1-5 μg).

Barium isotopic compositions were conducted using a Neptune Plus multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS). ¹³⁵Ba-¹³⁶Ba double-spike method with anoptimal ratio of 1.72 (m/m; Rudge et al., 2009) was utilized to correct for instrumental mass discrimination. Signals of ¹³¹Xe, ¹³⁴Ba, ¹³⁵Ba, ¹³⁶Ba, ¹³⁷Ba, and ¹⁴⁰Ce were collected simultaneously by L4, L2, L1, C, H1, and H3 Faraday cups, respectively. The background signals of ¹³⁷Ba (< 0.005 V) were negligible relative to the sample signals (approximately 7 V). Relative to NIST Standard Reference Material (SRM) 3104a, Ba isotopic composition is defined as

For comparison, all published δ^137/134Ba values have been recalculated to δ^138/134Ba values by multiplying the measured δ^137/134Ba by a factor of 1.33 (Pretet et al., 2015). The three-isotope plot is shown in Fig. 3. The average δ^138/134Ba values of reference materials (BCR-2 and BHVO-2) in this study are consistent with the uncertainty reported in previous studies (Table 1, An et al., 2020; Li et al., 2020; Bullen and Chadwich, 2016; Nan et al., 2015).

Figure  3.  Barium three-isotope plot of the basalts and reference standard materials, defining a δ^138/134Ba-δ^137/134Ba fractionation line with R² = 0.998 1.

DownLoad: Full-Size Img PowerPoint

Table  1.  Barium isotopic compositions of reference materials in this study

Standard	Description	δ138/134Ba (‰)	2SD	References
BCR-2	Basalt, Columbia River, USA	0.05	0.04	Li et al. (2020)
		0.06	0.04	An et al. (2020)
		0.06	0.03	Nan et al. (2015)
		0.07	0.02	This study
BHVO-2	Basalt, Hawaii, USA	0.06	0.03	Nan et al. (2015)
		0.06	0.03	Bullen and Chadwich (2016)
		0.05	0.03	This study

 | Show Table

DownLoad: CSV

3. RESULTS

3.1 Geochemistry

Except for only one sample (EM001-1), most of the studied basalts have relatively low LOI (loss on ignition) contents varying from 1.42 wt.% to 3.67 wt.%. According to the bulk TiO₂ contents and Ti/Y ratios (Xiao et al., 2004; Xu et al., 2001, Fig. 4), basalts from the Lijiang region had low TiO₂ contents (1.24-1.85) with Ti/Y ratios of 269-429, whereas the Miyi and Emeishan basalts had higher TiO₂contents and Ti/Y ratios (TiO₂ = 3.33-4.87, Ti/Y = 519-897). Moreover, the high-Ti basalts display larger compositional variations of SiO₂ (42.64 wt.%-54.00 wt.%) and MgO (3.73 wt.%-7.49 wt.%), relative to low-Ti basalts (SiO₂ = 49.55 wt.%-51.92 wt.%, MgO = 4.60 wt.%-6.12 wt.%, Fig. 5). The former does not show any linear correlations between MgO and other major oxides (e.g., Na₂O, SiO₂, TiO₂, total Fe₂O₃, Fig. 5, Table S1).

Figure  4.  The elemental diagrams of TiO₂ versus SiO₂ and Sm/Yb. The classification criterion of TiO₂ is from Xu et al. (2001).

DownLoad: Full-Size Img PowerPoint

Figure  5.  (a) SiO₂ vs. MgO; (b) TiO₂ vs. MgO; (c) Na₂O vs. MgO; (d) Fe₂O₃^T vs. MgO diagrams of the high-Ti and low-Ti basalts.

DownLoad: Full-Size Img PowerPoint

As shown in the primitive mantle-normalized spider diagrams (Figs. 6a-6b), the basalts generally show negative Nb and Sr anomalies, even though they are variably enriched in Ta with respect to neighboring elements. The high-Ti samples contain high Sm/Yb ratios (3.62-4.98) and Nb/U (20.23-31.32) ratios (Figs. 4b and 7a). Furthermore, almost all samples have LREE-enriched chondrite-normalized REE patterns comparable to the average ocean-island basalts (OIB) with obviously negative Eu anomalies (Figs. 6c-6d). In addition, the high-Ti basalts contain higher total REE contents (∑REE = 222 ppm -420 ppm) and stronger LREE/HREE differentiation ((La/Yb)_N = 8.5-17.9) than that of the low-Ti basalts (∑REE = 112 ppm -202 ppm, (La/Yb)_N = 4.4-8.2). Both high-Ti and low-Ti basalts were commonly enriched in large ion lithophile elements (LILEs, e.g., Ba, Th and U).

Figure  6.  Primitive mantle-normalized spider diagrams of trace elements and chondrite-normalized REE patterns for the low-Ti basalts and high-Ti basalts. The normalizing values and OIB are from Sun and McDonough (1989).

DownLoad: Full-Size Img PowerPoint

Figure  7.  (a) δ^138/134Ba vs. Nb/U; (b) δ^138/134Ba vs. Th/Ta; (c) δ^138/134Ba vs. Th/Nb; (d) δ^138/134Ba vs. Th/Yb for the Emeishan basalts.

DownLoad: Full-Size Img PowerPoint

3.2 Barium Isotopic Compositions

Generally, the δ^138/134Ba values of the basalts are highly heterogeneous, ranging from -0.38‰ to 0.38‰ with a mean of 0.026‰ ± 0.38‰ (2SD, n = 16, Table 2). The two samples had the lowest δ^138/134Ba values of -0.38‰ and -0.33‰, respectively. As for the low-Ti basalts, their δ^138/134Ba values vary from -0.33‰ to +0.23‰ with an average of -0.02‰ ± 0.40‰ (2SD, n = 4). The high-Ti samples show a similar range of Ba isotopic compositions from -0.38‰ to +0.38‰ with their average value of 0.038‰ ± 0.36‰.

Table  2.  Barium isotopic composition of Emeishan flood basalts from Lijiang, Miyi and Emeishan areas

Locality	Types	Sample	Ba (ppm)	δ137/134Ba (‰)	2SD	δ138/134Ba (‰)	2SD
Lijiang	Low-Ti	WKB7	711	0.02	0.05	0.01	0.04
		WKB9	396	0.19	0.00	0.23	0.02
		PCB1	384	-0.26	0.01	-0.33	0.01
		PCB2	170	-0.03	0.03	-0.04	0.02
		Replicate		-0.03	0.03	-0.03	0.03
Miyi	High-Ti	MY1	402	0.29	0.00	0.38	0.01
		MY2	1 028	-0.28	0.02	-0.38	0.03
		MY4	619	0.09	0.00	0.13	0.02
		MY5	468	0.14	0.01	0.18	0.03
		MY6	527	0.05	0.02	0.06	0.03
		MY7	365	0.19	0.00	0.26	0.04
Emeishan	High-Ti	EM001-1	272	0.03	0.04	0.04	0.05
		EM001-2	250	0.03	0.00	0.04	0.01
		EM002-1	523	0.08	0.02	0.11	0.01
		EM002-2	564	0.05	0.02	0.07	0.05
		EM002-3	495	0.15	0.01	0.19	0.01
		EM004-2	517	0.01	0.02	0.01	0.02
		EMSB001	479	-0.02	0.01	-0.03	0.01

 | Show Table

DownLoad: CSV

4. DISCUSSION

4.1 Possible Mechanisms for the Compositional Variations of the Basalts

The geochemical and isotopic compositions of basalts can be strongly influenced by partial melting, fractional crystallization, weathering alterations and crustal contaminations (Gong et al., 2019; You and Liu, 2014; Zhu et al., 2013; Hou et al., 2011; Jiang et al., 2007; Xu et al., 2007). In this study, the Ba isotopic compositions of the flood basalts were highly heterogeneous (-0.38‰ to 0.38‰, Figs. 8 and 9), indicating that petrogenesis is more complex than previous thoughts.

Figure  8.  δ^138/134Ba (‰) vs. Ba (ppm) diagram of the studied basalts. The δ^137/134Ba value of the upper mantle (PM) is from Nan (2017).

DownLoad: Full-Size Img PowerPoint

Figure  9.  Compiled δ^138/134Ba values of the studied basalts from the ELIP and other reservoirs from literatures (Li et al., 2020; Nan et al., 2018; Nielsen et al., 2018; Nan, 2017).

DownLoad: Full-Size Img PowerPoint

4.1.1   Weathering alteration

Except for only one sample (EM001-1), all analyzed basalts contain relatively low LOI contents (1.42 wt.%-3.67 wt.%), reflecting that weathering alterations pose a limited impact on their geochemical components. Sample EM001-1 has the highest LOI (7.13 wt.%) and median δ^138/134Ba value (0.038‰). Because of its highly fluid-mobile ability, barium preferentially enters into low-temperature fluids during weathering processes, such as serpentinization. However, it is uncertain that whether the weathering process could cause Ba isotopic fractionation within basalts. Thus, the sample EM001-1 is not considered in the following discussion.

4.1.2   Partial melting

Owing to a highly incompatible coefficient (D_solid/melt) of 0.000 12 during mantle melting (Workman and Hart, 2005), a large proportion of Ba (99%) is preferentially incorporated into melts despite 1% partial melting (Nan et al., 2018; Nielsen et al., 2018). Xu et al. (2001) proposed that the high-Ti basalts were formed via a partial melting of 1.5% within the garnet stability field, whereas the low-Ti magmas were generated by a greater extent of melting (16%). Accordingly, almost all Ba were separated from the mantle and entered into the basaltic magmas after partial melting. Therefore, this process, does not result in significant Ba isotopic fractionation, i.e., the δ^138/134Ba values of the primitive basaltic melts were likely dependent on those in the mantle source.

4.1.3   Fractional crystallization

Previous studies have demonstrated that the high-Ti and low-Ti basalts underwent fractional crystallization during basaltic magma ascent, supporting the abundance of plagioclase or clinopyroxene phenocrysts as well as other geochemical fingerprints (Cheng et al., 2019; Yang et al., 2018; Tian et al., 2017; Song et al., 2008; Hao et al., 2004). The continental flood basalts, especially high-Ti samples, are no exception. The high-Ti basalts contain numerous plagioclase phenocrysts (Fig. 2), supporting the possibility that they might have experienced significant fractionation of plagioclases. The negative Eu and Sr anomalies are also considered to be a consequence of plagioclase fractional crystallization (Fig. 6, Huang et al., 2014). Even so, the geochemical proxies of fractional crystallization (e.g., solidification index, SI = 100 × MgO/(MgO + FeO + Fe₂O₃ + Na₂O + K₂O), Liu et al., 2020 and references therein) did not coincide with the measured δ^138/134Ba values (Fig. 10a), suggesting that the variations in Ba isotopic compositions of the basalts are independent of plagioclase fractionation crystallization due to an elevated degree of fractionation corresponding to a decreased SI value (Liu et al., 2020, and references therein). Similarly, the lack of linear correlations between SI and Eu or Sr further provides additional evidence for this possibility (Fig. 10b).

Figure  10.  (a) δ^138/134Ba vs. solidification index (SI) and (b) Sr contents vs. SI diagrams of the basalts.

DownLoad: Full-Size Img PowerPoint

4.1.4   Crustal contamination

The high-Ti and low-Ti basalts collected from the Emei-shan, Miyi and Lijiang regions have wider ranges of δ^138/134Ba values (-0.38‰- +0.38‰) relative to the typical ones of the typical ocean island basalts (-0.11‰- +0.20‰, Nan, 2017) and PM (0.05‰ ± 0.06‰, Li et al., 2020, Fig. 8). In view of a fact that Ba isotope is indeed homogeneous in the upper mantle (Li et al., 2020), such isotopic diversity suggests probable contamination of some crustal materials either in magmatic sources through lithosphere-plume interactions (e.g., He et al., 2010; Fan et al., 2008), or during the basaltic magmas en route to the surface. If the contamination process took place in their magmatic chambers via crust-mantle interaction, both the high-Ti and low-Ti basalts possibly originated from the same sources because of their similar δ^138/134Ba values. This possibility can be ruled out owing to their distinct Ce/Y, Sm/Y (Fig. 11b), Th/Nb as well as Zr/Nb ratios (e.g., Li et al., 2016). Also, Gd/Yb ratio can be utilized to estimate the depth of partial melting (e.g., He et al., 2010). The Gd/Yb ratios of the high-Ti basalts (3.57-4.94) are obviously greater than the low-Ti basalts (2.06-2.33), reflecting a larger melting depth of the high-Ti basalts. In addition, two different mantle sources responsible for these two types of basalts have been well demonstrated in previous studies (e.g., Xu J F et al., 2007; Yan et al., 2007; Xu Y G et al., 2001; Xiao et al., 2004).

Figure  11.  (a) Nb/U vs. Th/Nb, (b) Ce/Y/ vs. Sm/Y, and (c) (Nb/Th)_PM vs. (Th/Yb)_PM for the Emeishan basalts (modified from Qi and Zhou, 2008). OIB and PM data are from McDonough and Sun (1995) and the upper crust data are from Rudnick and Gao (2014).

DownLoad: Full-Size Img PowerPoint

The ratios of incompatible elements of basalts, such as Nb/U, Th/Ta, Nb/La, Th/Nb, Ce/Y, and Sm/Y, is indicative of crustal contamination (e.g., Fu et al., 2021; Liang et al., 2021; Tan et al., 2019; Li H B et al., 2017; Lai et al., 2012; Li C S et al., 2012; Zi et al., 2011; He et al., 2010; Jiang et al., 2009, 2007), in that these ratios are not easily affected by fractional crystallization (Song et al., 2006). In practice, crustal contamination commonly results in low Nb/La, Nb/U and high Th/Ta, Th/Nb, Th/Yb (e.g., Fu et al., 2021; Li et al., 2017; Qi and Zhou, 2008; Wang et al., 2007; Song et al., 2006). The studied basalts exhibit a similar compositional feature (Table 3), which proves crustal contamination played a significant role in the petrogenesis of the basalts. Another evidence comes from the (Th/Yb)_PM versus (Nb/Th)_PM diagram (Fig. 11c), in which all samples are plotted along the trend line of crustal contamination. Crustal contamination also accounts for the negative Nb anomalies shown in the PM-normalized spider diagrams of trace elements (Figs. 6a-6b) and Ba isotopic deviation from the PM (Fig. 9). What is peculiar to note that their Th/Nb (0.13-0.60), Nb/La (0.56-1.29), Nb/U (6.58-31.32), La/Sm (3.09-5.21) and Th/Yb (1.10-5.76) ratios are between those of PM (McDonough and Sun, 1995) and continental crust (Rudnick and Gao, 2014, Table 3). This strongly indicates that the overlying continental crusts might be the most potential source of contamination materials. More importantly, these low-Ti basalts have lower Nb/U, Sm/Y, Nb/Y ratios as well as Nb and Ta abundances with respect to the high-Ti basalts (Fig. 11). Thus, the low-Ti basalts had undergone higher degrees of crustal contamination. Figure 7 further illustrates that the basalts containing more amounts of contaminated materials (e.g., low-Ti basalts) seem to display a nearly linear correlation between their δ^138/134Ba values and geochemical proxies of crustal contamination (e.g., Th/Ta, Th/Yb, Nb/U and Th/Nb).

Table  3.  The elemental ratios of the basalts from ELIP

Magma types	High-Ti basalts			Low-Ti basalts			PMc	OIBd	UCCe
	Ranges	Rangesa	Rangesb	Ranges	Rangesa	Rangesb
Ti/Y	519-897	449-979	528-682	351-531	149-531	269-430	813	593	182
Nb/La	0.76-1.06	0.78-1.16	0.62-1.07	0.56-1.29	0.81-0.99	0.61-1.81	1.04	1.30	0.39
Th/Yb	1.84-3.75	0.11-0.15	0.31-3.22	1.10-2.95	1.35-1.52	0.45-2.88	0.17	1.85	5.36
Nb/U	20.23-31.32	28.75-38.00	20.67-46	6.58-21.96	19.8-23.3	8.00-50.00	33.95	47.05	4.44
Th/Nb	0.13-0.21	0.11-0.15	0.07-0.24	0.14-0.60	0.17-0.18	0.06-0.4	0.12	0.08	0.88
La/Sm	3.09-5.00	2.78-4.55	3.32-4.72	3.35-5.21	3.28-4.04	2.43-4.23	1.55	3.70	6.60
Th/Ta	1.10-2.84	1.56-2.07	1.23-3.35	1.73-5.76	2.41-2.55	1.67-6.55	2.07	1.48	11.67
a Data from He et al. (2010); b data from Xu et al. (2001); c PM. primitive mantle; d OIB. ocean-island basalts, OIB and PM data are from McDonough and Sun (1995); e UCC. upper continental crust, the upper continental crust data are from Rudnick and Gao (2014).

 | Show Table

DownLoad: CSV

In conclusion, crustal contamination is the most likely interpretation for the Ba isotopic variation in basalts, rather than partial melting, fractional crystallization and weathering alteration.

4.2 Petrogenesis of the High-Ti and Low-Ti Basalts

It is generally believed that the mantle plume-derived melts played a decisive role in the genesis of the ELIP during the Late Permian-Early Triassic (Fu et al., 2021; Liu et al., 2020; Shellnutt and Pham, 2018; Li H B et al., 2016; Deng et al., 2014; Li C S et al., 2012; Ali et al., 2010, 2005; He Q et al., 2010; He B et al., 2006, 2003; Zhang et al., 2006; Chung and Jahn, 1995, and references therein), even though a few scholars still insisted a non-mantle plume model for the genesis of the ELIP (e.g., Liu et al., 2015; Thompson et al., 2001; Huang et al., 1992). Regardless of whether ELIP has a genetic affinity with a plume, the wider ranges of Ba isotopic compositions of the Emeishan basalts (-0.38‰- +0.38‰) in contrast to that of the PM prove that they have been variably contaminated by crustal materials. As corroborated by their geochemical and Ba isotopic features, the basaltic magma was initially generated via partial melting of the mantle, and subsequently overprinted by crustal contamination of the overlying continental materials during the ascent. Evidence from studies on their Sr-Nd-Pb-Hf and Re-Os isotopes also confirms this process (e.g., Fu et al., 2021; Zhang, 2019; Shellnutt and Jahn, 2011; He et al., 2010; Xu J F et al., 2007; Xiao et al., 2003; Xu Y G et al., 2001). More importantly, the low-Ti basalts had undergone a higher degree of crustal contamination relative to the high-Ti basalts. It is well consistent with their spatial distributions that the high-Ti basalts are located in the periphery of the ELIP and the low-Ti basalts are within the center (Tian et al., 2017). Due to their higher mantle potential temperature (T > 1 550 ℃), the hotter, low-Ti basaltic magmas from the plume center could assimilate more crustal materials easily than those cooler, high-Ti basalts (T < 1 500 ℃) during eruption (Liang et al., 2021; Xu et al., 2001). This is also reinforced by additional elemental data of the basalts from Lijiang, Dongchuan, Binchuan, and Miyi of the ELIP (Table 3, He et al., 2010; Xu et al., 2001).

5. CONCLUSIONS

(1) The δ^138/134Ba values of the Emeishan continental flood basalts are highly heterogeneous, ranging from -0.38‰ to 0.38‰, which reflects significant contributions from variable contamination of crustal materials rather than other geological processes, such as partial melting, fractionation crystallization and weathering alteration.

(2) The elemental ratios of incompatible element simply that the low-Ti basalts from Lijiang experienced higher degrees of crustal contamination than the high-Ti basalts from Miyi and Emeishan during the ascent of basaltic magmas to the surface.