1 Introduction

Strontium isotope analysis is an important tool for understanding weathering processes and the global carbon cycle. For example, 87Sr/86Sr analysis is often used to research the oceanic Sr isotope budget (Taylor and Lasaga 1999; Dessert et al. 2001; Allègre et al. 2010; Li and Elderfield 2013) and for end-member analysis in ecosystem studies (Gosz et al. 1983; Capo et al. 1998; Pett-Ridge et al. 2009; Meek et al. 2016). That rocks and minerals have the same 87Sr/86Sr ratios as those of their weathering products is regularly assumed when calculating the contribution of different rocks/minerals to weathering solutions (Chung et al. 2009; Rengarajan et al. 2009).

It is generally recognized that basalt crystallizes rapidly, with a low degree of evolution and little magmatic differentiation; therefore, the Rb/Sr and 87Sr/86Sr ratios among different minerals in basalt should be similar. Similarly, the chemical weathering rate of basalt is very fast (Dessert et al. 2003), suggesting that there is no differentiation of Sr isotopes in the weathering processes. As a result, the 87Sr/86Sr ratios of rivers should be similar to those of bedrock in basaltic basins.

However, we found that the 87Sr/86Sr ratios of river water were higher than those of basaltic rocks in Xuyi, China, which is a pure Cenozoic basalt area. Our research suggests that rainwater and dust are not the reason for the difference of ratios between basalt and basin water; rather, much higher 87Sr/86Sr ratios in plagioclase than in bulk basalt might account for the difference.

2 Samples and methods

A total of 22 water samples were collected from 3 rivers, 2 reservoirs, a spring, and 2 wells in Xuyi, China during three periods: January, May, and June of 2015. All water samples were filtered through 0.22 μm Millipore filter, acidified to pH < 2 with ultrapure nitric acid, and stored at a temperature below 5 °C. Two fresh basaltic rock samples were collected and washed under ultrasonic wave to remove clay, then dissolved in an HF–HNO3 mixture and analyzed for Sr isotope composition. A total of 3 kg of basalt were grinded to 2–5 mm and leached with pH = 3 hydrochloric acid. Leached solutions were collected for major element and strontium isotopic ratio analyses. About 150 pieces of plagioclase, quartz, Ca-montmorillonite, augite, and spinel crystals, as well as matrix, were selected from the broken and sieved basalt under binoscope by hand, and then dissolved in HF–HNO3 mixture for Sr isotope analysis.

Major element and strontium concentrations in the water samples were measured in filtered and acidified water with an inductively coupled plasma atomic emission spectrometer (ICP-AES, iCAP-6300) in the Key Laboratory of Surficial Geochemistry, Ministry of Education, Nanjing University. Strontium in the solution was separated using Sr-Spec resin and isotopic ratios were determined on a Neptune Plus multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS) also at the Key Laboratory of Surficial Geochemistry. 87Sr/86Sr was normalized to 86Sr/88Sr = 0.1194. Reproducibility and accuracy of strontium isotopic analyses were periodically checked by running the Standard Reference Material NBS 987, yielding a long-term mean 87Sr/86Sr value of 0.710265 ± 13 (2σ external standard deviation, n = 19). The international seawater standard IAPSO was used to verify the chemical procedure. Measurements of ten replicates yielded a mean 87Sr/86Sr value of 0.709194 ± 22 (2σ external standard deviation, n = 10). The analytical blank was <1 ng for strontium.

3 Results and discussion

The Sr isotope data for all samples have been plotted in Fig. 1 (detailed data are omitted in this paper). The 87Sr/86Sr ratios of field water—including river, reservoir, spring, and well water—ranged from 0.706255 to 0.707471 with a mean value of 0.706726 (n = 22). The 87Sr/86Sr ratios of the two fresh rock samples were 0.704052 and 0.703948 with a mean value of 0.704000—much lower than the field water. The 87Sr/86Sr ratios of the in-house weathering solution ranged from 0.708081 to 0.708339 with a mean value of 0.708203 (n = 17). The 87Sr/86Sr ratios of plagioclase, Ca-montmorillonite, quartz, augite, and matrix were 0.712897, 0.707761, 0.707585, 0.707192, and 0.703943, respectively. Spinel is insoluble in HF and HNO3, and is dysgeogenous; therefore we did not measure its Sr isotopes. Plagioclase had the highest 87Sr/86Sr ratio; the other mineral samples had 87Sr/86Sr ratios similar to those of the field water samples. The matrix sample had a 87Sr/86Sr ratio very similar to that of the fresh basaltic rocks, which might suggest that matrix and whole rocks have the same composition.

Fig. 1
figure 1

Strontium isotope data of all samples

Full size image

The average 87Sr/86Sr ratio of field water (0.706726) was 0.002726 higher than that of fresh basaltic rocks (0.704000), a difference much larger than analysis error. It is generally recognized that rainwater and dust can have significant influence on a river’s composition (Jin et al. 2011; Wu et al. 2013; Sherman et al. 2015; Clauer and Semhi 2016); therefore, higher 87Sr/86Sr ratios of field water might be due to dust and rainwater. However, the fact that the 87Sr/86Sr ratios of the in-house weathering solution (without rainwater or dust inputs) were higher than fresh basaltic rocks, suggests that dust and rainwater aren’t the reason for the 87Sr/86Sr difference between basalt and basin water.

Among the selected four minerals (not considering matrix or spinel), plagioclase had the highest 87Sr/86Sr ratios. The 87Sr/86Sr ratios of the other three minerals were lower than those of the in-house weathering solutions; quartz, Ca-montmorillonite, and augite could not, therefore, have caused the high 87Sr/86Sr ratios of the in-house weathering solution. Mostly likely, plagioclase was primarily responsible for the high 87Sr/86Sr ratios of field water and of the in-house weathering solution, while quartz and Ca-montmorillonite more likely precipitated from the pore solution of basalt, which would explain the similar 87Sr/86Sr ratios among quartz, Ca-montmorillonite, and field water. The 87Sr/86Sr ratio of augite is close to that of field water, but lower than that of the in-house weathering solution, which might suggest that augite is not related to the difference of 87Sr/86Sr ratios between basalt and basin water.

While it is generally argued that the 87Sr/86Sr ratios of all minerals in basalt are similar as a result of the rock’s rapid crystallization, this work shows that different minerals in basalt have different 87Sr/86Sr ratios. We have the opinion that plagioclase has suffered from low-temperature hydrothermal alteration, increasing the 87Sr/86Sr ratios of plagioclase and augite.

In the field of weathering research, the 87Sr/86Sr ratios are widely used to analyze the end-member contribution of various sources. It is usually assumed that the 87Sr/86Sr ratios of basalt contribute to a weathering basin are equal to those of the basaltic rocks themselves (Chung et al. 2009; Rengarajan et al. 2009). Our work found that there is a big difference in 87Sr/86Sr ratios of basalt and basaltic basin water, which suggests that the 87Sr/86Sr ratios of the basaltic rocks do not represent the basalt end-member.

4 Conclusion

There is a big difference in 87Sr/86Sr ratios of basalt and basin water in Xuyi, China. Our work found that the high 87Sr/86Sr ratios of plagioclase, caused by low-temperature hydrothermal alteration, are the main reason for this difference. This finding tells us that when analyzing the end-member contribution in a mixed-rock basin including basalt, we should take this difference in account.