Introduction

Seafloor hydrothermal activity is the key intermediary for material and energy exchange between the ocean and the crust (Corliss et al., 1979). For nearly 50 years, abundant geochemical and mineral criteria and indicators are successfully used in recognizing hydrothermal activity, such as the metalliferous sediment index MSI = Al/(Al + Fe + Mn) vs. Fe/Ti ratio (Boström et al., 1969; Boström, 1973), high U/Fe ratios, and low Th/Fe ratios (Mills et al., 1993; Dias et al., 2010). High concentrations of Cu, Zn, Fe, Mn, and other hydrothermal-derived elements in sediments can indicate the hydrothermal components’ input (e.g., Fe content higher than 5% indicates hydrothermal input on a carbonate-free basis [cfb]) (Lisitzin et al., 1996, 1978). Typical element and mineral combinations, such as Fe–Si ± S-rich and Mn–Fe–Si-rich phases, sulfide (2–4 wt.%), and barite derived from buoyant plume particulates and black smoker particle, can be enhanced in proximal sediments (35–200 m from the active vent), whereas Fe–Mn oxyhydroxides and authigenic Fe–rich montmorillonite can be enhanced in distal sediments (up to 10–15 km) (Kuksa et al., 2021; Hrischeva and Scott, 2007).

Rare earth elements (REEs) have been widely used for tracing the hydrothermal geochemical process owing to their similar geochemical behavior, redox state, and ion radius (Henderson, 1984; Dubinin, 2004). The hydrothermal components in deep-sea sediments related to the oceanic hydrothermal system have a dual origin, weathered massive sulfide edifices, and precipitation of particulate matter from overlying plumes (Hodkinson and Cronan, 1995; Kuksa et al., 2021). In metalliferous sediments, most sulfide particles are typically mass wasting and completely oxidized. On the other hand, hydrothermal plumes generated by hydrothermal activity contain abundant hydrothermal components that generally precipitate in adjacent sediments (Mills and Elderfield, 1995). This results in variable REE composition and patterns, which significantly differ with background sediments (Chavagnac et al., 2005; Li et al., 2016). Therefore, the REE characteristics of sediments could be also used in identifying hydrothermal activity (Chavagnac et al., 2005).

Previous researches have shown that variations in REEs in seafloor sediments are primarily controlled by mineral and geochemical composition (Zhang et al., 2017; Arraes-Mescoff et al., 2001) and grain size (Mccave and Hall, 2006). In general, the enrichment of REEs in seafloor sediments follows the particle size control law, whereby REEs are prominently enriched in clay, especially fine-grained components (particularly clay- and silt-sized grains) (Cullers et al., 1987, 1988; Condie and States, 1991); however, the occurrence of REEs in coarse-grain minerals, particularly heavy minerals and bioclasts, would break this law (Dou et al., 2012). In mid-ocean ridges (MORs), the composition and distribution patterns of REEs sediments are usually influenced by the following: (i) scavenging by Fe–Mn hydroxide-oxides in hydrothermal plumes and the subsequent precipitation, which is controlled by the distance from and degree of the hydrothermal activity (Kato et al., 2011; German et al., 1993; Mills and Elderfield, 1995); (ii) the dilution of calcareous and siliceous biogenic components from biological activities (Pattan and Jauhari, 2001; Yasukawa et al., 2016); (iii) contamination from terrestrial debris (Kato et al., 2011); (iv) type of sediment and the grain size of the sediment (Menendez et al., 2017; Fan et al., 2015; Ranasinghe et al., 2002); and (v) the local redox environment (Argese et al., 1992). Such complex factors result in difficulties in analyzing the REE features of sediments with different grain sizes and, thus, the interpretation of REE patterns for the identification of hydrothermal activities (Agarwal et al., 2020; Mitra et al., 1994; Sands et al., 2012).

Recent studies have shown that abundant hydrothermal activity occurs on the magma-starved ultraslow-spreading Southwest Indian Ridge (SWIR) (Tao et al., 2014). Studies on surface sediments of the SWIR show that these sediments mainly comprise calcite ozze with minor local debris, eolian dust, Fe–Mn (hydr)oxides, and occasinally hydrothermal components (Jia et al., 2019; Li et al., 2016; Liao et al., 2018, 2019; Lin et al., 2018; Agarwal et al., 2020). The metalliferous sediments show high contents of Cu, Zn, and Fe, mostly negative Ce and positive Eu anomalies (Liao et al., 2018). In addition, the contribution of hydrothermal activity to the sediments was found to be related with their distances and also be affected by local mafic and ultramafic rocks (Chen et al., 2021; Pan et al., 2018). The great distance of the SWIR from the continent reduces terrigenous content to a minimum, providing a favorable condition for the study of the REE composition of MOR sediments with different grain sizes. We systematically analyzed the REE characteristics of sediments from > 830 to < 74 μm grain-sized samples collected from areas proximal, at a moderate distance, and distal to hydrothermal fields (HFs) on the SWIR. We analyzed the sediment source and the main factors controlling the REE characteristics, and we discussed the favorable grain size for hydrothermal activity identification and hydrothermal contribution as determined by this analysis. Importantly, we emphasize that the implications of the results obtained by this study are restricted in the studied segments while not necessarily to be considered as unified model applicated in MOR sediments.

Geological background

The ultraslow-spreading SWIR (full spreading rate of 14–16 mm/a) forms a boundary between the Antarctic and African plates and has a total length of around 8000 km (Dick et al., 2003). The water depth varies along the axis of the SWIR, reaching a maximum of about 5000 m. The axial rift is rugged and cut by a series of north–south striking transform faults. The oceanic crust is on average ~ 4 km thick, which is far thinner than the global average of ~ 7 km (Baker and German, 2004). The topography, the mantle Bouguer anomaly, and a magnetic anomaly suggest that the axial magmas here are characterized by lateral migration and centralized proliferation at specific locations, probably attributable to local magma accretion or the Marion and Crozet hot spots (Fig. 1a; Cannat et al., 2008; Sauter et al., 2004; Yang et al., 2017).

Fig. 1
figure 1

a Tectonic location and b topography of the study area. The topography data in (a) were from the General Bathymetric Chart of the Oceans (GEBCO, 2008), and data in (b) were collected by multibeam survey

Full size image

The terrain in the study area, located between the Indomed and Gallieni transition faults, is rough, with the water depth on the north side of the axis being deep but relatively shallow on the south side (Fig. 1b). The seafloor is mainly composed of exposed basalts, with localized serpentinized peridotite, gabbro, and pyroxenite (Zhou and Dick, 2013). Modern sediments are relatively rare and restricted to localized depressed areas (Liao et al., 2018). Currently, several HFs have been confirmed in the study area, including the Longqi-1, Suye (previously Longqi-3), Donglongjing-2, Yuhuang-1, and Chunxi-1 (Tao et al., 2014). The Longqi-1 field, the first active field to be discovered on an ultraslow-spreading ridge, is characterized by deep circulating hydrothermal fluids controlled by detachment faults (Tao et al., 2020, 2012). The other fields have been reported as inactive.

Samples and methods

Samples

Surface sediment samples were collected from six stations located proximal, at a moderate distance, and distal to the confirmed HFs via TV grab (Fig. 1b and Table 1). The water depths at the sampling stations ranged 1494–3268 m, which is less than the calcium carbonate compensation depth. These sediments were primarily composed of calcareous ooze with small amounts of volcanic glass or basaltic debris.

Table 1 Location and descriptions of the study sediment samples
Full size table

Methods

Mineralogy

The samples were sieved into nine categories before being oven-dried at 60 °C (i.e., > 20, 20–40, 40–60, 60–80, 80–100, 100–120, 120–160, 160–200, and < 200 mesh corresponding to > 830, 380–830, 250–380, 180–250, 150–180, 120–150, 96–120, 74–96, and < 74 μm, respectively). The finest grained samples (< 200 mesh) were first elutriated in distilled water to prevent the loss of clay minerals.

The sieved sediments were observed in hand specimen, microscopically, and by smear microscopy in the Key Laboratory of Submarine Geosciences, Second Institute of Oceanography in Hangzhou, China. Because rock debris > 2 mm is not conducive to smearing, only grain sizes < 2 mm were used for the smear observations and analyses. The microscopic observation used a polarizing microscope (Eclipse E600 Pol; Nikon Corporation, Japan) with stereoscopic and digital functions. The petrological characteristics of the samples were observed and analyzed, including mineral type, composition, and grain size. The mineralogical and biological detrital components, including foraminifera, were identified by consulting a sedimentological reference map (Terry and Chilingar, 1955).

Major and trace element analyses

Due to the small number of samples available after sieving, we only analyzed the major elements of the unsieved bulk samples. The major elements were analyzed at the ALS Laboratory Group in Guangzhou, China, by an X-ray fluorescence spectrometer (Magix PW2424; Panalytical, the Netherlands). The sediments were ground to < 200 mesh (74 μm) in an agate mortar. After drying, 0.6 g of the powdered sample was weighed and completely mixed with 6 g of a cosolvent containing lithium and transferred to a platinum crucible and heated to 1000 °C for approximately 1 h. After melting, each sample was left to cool in a flat glass sheet. Then, 0.5-g sample powders were weighed, placed in a porcelain crucible, and heated to 1000 °C for 4 h to determine the loss on ignition. During the analysis, three standard samples (i.e., GBW07105, NCSDC47009, and SARM-4) were also processed for data accuracy monitoring, and duplicate samples were used for quality inspection. The data measurement error was within 5%.

The trace and REEs were analyzed at the ALS Laboratory Group by inductively coupled plasma mass spectrometry (ICP-MS) (7700x; Agilent Company, USA). Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was used to detect elements with high contents, that is, Ni, Cr, Sc, Co, and V. For the digestion, ~ 250 mg sample powders were weighed in Teflon beakers and mixed with high purity HNO3, HF, and HClO4. After digestion, each sample was heated in an oven for > 24 h at 180 °C and then transferred to an electric heating plate heated to 180 °C until the samples evaporated to wet salt status. The residues were dissolved in dilute HCl, and the final solution was analyzed by either ICP-MS or ICP-AES. The results were adjusted for spectral inter-element interference. Two standard samples (i.e., GBM908-10 and MRGeo08) were processed for data quality control. The data measurement error was within 10%.

Results

Mineral composition

Microscopic observations indicated that the sediments were mainly calcareous oozes, comprising calcareous nannofossils (> 50%, in volume) and foraminifera shells (10–50%) with minor debris. This was consistent with XRD analysis that showed the majority of the minerals were calcite (81.2–96.8 wt.%), followed by aragonite (1.32–24.0 wt.%) with minor quartz (0.3–1.2 wt.%), albite (0.6–2.4 wt.%), halite (1.7–2.8 wt.%), and clay minerals of kaolinite (0.4–15.4 wt.%) (Chen et al., 2013). The detrital content of the proximal sediments was significantly higher than that of the moderate distance and distal sediments. The > 250 μm proximal sediments had high local debris contents, being mainly basaltic or altered basaltic debris, with additional minor clastic serpentinite, mica, and asbestos (Fig. 2a–c). The minor debris was pale yellow with a blue surface, probably the result of the partial oxidization of sulfide minerals (Fig. 2b). All the < 250 μm sediments exhibited comparable mineral compositions, with high calcareous biological components and clay minerals (90–95%, in volume) and minor rock debris (5–10%) (Fig. 2d–g). The calcareous components were mainly calcareous nannofossils (> 50%) and foraminifera shells and fragments (10–50%) (Fig. 2h and i), which were mostly white to yellowish-white with some brownish-yellow Fe staining.

Fig. 2
figure 2

Photomicrographs of the sediments, plane-polarized light. a The proximal sediment S24, > 830 μm sediments primarily composed of basalt debris, volcanic glass, and sponge spicules; b the proximal sediment S15, mineralized rock debris in the > 830 μm sediments showing sulfides oxidized to covellite; c the proximal sediment S24, 380–830 μm sediments primarily composed of altered basalt debris, pyroclastic debris, and minor serpentinized peridotite debris; d the proximal sediment S24, 250–380 μm sediments with significant foraminifera content and some Fe-stained brownish-yellow sediments, indicating oxidation; e the moderate sediment S19, 120–150 μm sediments with considerable foraminifera content and minor basaltic debris; f the moderate sediment S19, 74–96 μm sediment with compositions similar to 120–150 μm sediments; g the distal sediment S05, magnified 180–250 μm sediments with significant contents of clear and white-colored foraminifera; h abundant foraminifera shells and fragments with rare debris (smear photo, the proximal sediment S24); and i abundant foraminifera shells and fragments (smear photo, the moderate sediment S16). Bas, basalt; Sp, sponge; Cv, covellite; Slf, sulfide; Ol, olivine; foram, foraminifera.

Full size image

Geochemistry

Major elements in the unsieved sediments

The unsieved bulk sediments from the six stations had high CaO contents (41.43–51.51 wt.%; average 47.15 wt.%; Table 2), consistent with being calcareous oozes. The SiO2, Al2O3, and TiO2 contents were generally low, ranging from 2.27 to 11.65, 0.67 to 3.07, and 0.03 to 0.14 wt.%, respectively. The distal sediments exhibited high CaO and MnO contents (average 50.87 wt.%, 0.073 wt.%) but relatively low TFe2O3, Al2O3, and TiO2 contents (average 0.68, 0.81, and 0.04 wt.%, respectively) compared with the proximal and moderate distance samples (average 2.48, 1.69, and 1.78 wt.%, respectively), consistent with the relatively low debris contents. However, the moderate distance sediments exhibited relatively high SiO2, Al2O3, TiO2, and MgO contents with an average value of 11.07, 3.02, 0.12, and 2.85 wt.%, respectively.

Table 2 Major element composition of the bulk sediments (wt.%)
Full size table

Distribution of trace elements according to sediment grain size

We analyzed 52 sieved-sediment samples, and according to the distribution of the trace elements, the sediments were divided into three groups: > 250 μm, 74–250 μm, and < 74 μm (Fig. 3, Table 3). In general, all the > 250-μm sediments had the highest Cu, Zn, Fe, and Mn contents, Fe/Mn ratio, and the lowest Cu/Fe and Zn/Fe ratios, whereas the < 74 μm sediments displayed the opposite trends.

Fig. 3
figure 3

Trace element distribution according to sediment grain size. 1 =  > 830 μm, 2 = 380–830 μm, 3 = 250–380 μm, 4 = 180–250 μm, 5 = 120–150 μm, 6 = 150–180 μm, 7 = 96–120 μm, 8 = 74–96 μm, and 9 =  < 74 μm. The bright red, turquoise, and dark blue dots represent the proximal, moderate distance, and distal sediments, respectively

Full size image
Table 3 Trace and rare earth elements composition of different grain-sized sediments from different distances to hydrothermal vent
Full size table

For the proximal sediments, the > 250-μm samples contained high Cu, Zn, Fe, and Mn (average 458 μg/g, 167 μg/g, 5.47 wt.%, 0.16 wt.%, respectively) (Fig. 3a–d), and the Fe/Mn ratio decreased with decreasing grain size (Fig. 3g). The Cu content and Cu/Fe ratio were also relatively high in the 74–250 μm (average 513 μg/g, 0.026) and < 74 μm (average 549 μg/g, 0.034) sediments (Fig. 3c and e), implying a high contribution of hydrothermal components. With decreasing grain size, the Cu/Fe and Zn/Fe ratios gradually increased (Fig. 3e–f), while the Fe/Mn ratio decreased (Fig. 3g).

In the moderate distance sediments, the > 250 μm samples showed lower Cu, Zn, Fe, and Mn contents (average 114 μg/g, 51 μg/g, 4.41 wt.%, and 0.07 wt.%, respectively). The Cu/Fe and Zn/Fe ratios exhibited decreasing trends comparable to the proximal sediments, while the Fe/Mn ratios exhibited the opposite trend. The Ni, Co, Fe contents, and Fe/Mn ratio were also high in the 74–250 μm (average 239 μg/g, 337 μg/g, 3.19 wt.%, 40.75, respectively), and the Zn, Fe contents, and Fe/Mn ratio were low in the < 74 μm sediments (average 33 μg/g, 1.52 wt.%, 25.31, respectively).

In the distal sediments, except sample S05 with high content of Fe and Mn contents, the > 250 μm samples had relatively low Cu, Zn, and Fe contents (average 21 μg/g, 17 μg/g, 0.59 wt.%, respectively), high Mn, Ba, Sr, and Pb contents (average 0.08 wt.%, 153 μg/g, 1361 μg/g, and 10 μg/g, respectively), and low Cu/Fe and Fe/Mn ratios (average 0.004 and 12.95) and exhibited a similar Fe/Mn ratio distribution to the moderate distance sediments. The Cu, Zn, Fe, Mn contents, and Fe/Mn ratio were also relatively low in the 74–250 μm (average 18 μg/g, 15 μg/g, 0.34 wt.%, 0.03 wt.%, 11.37, respectively) and low in the < 74 μm sediments (average 29 μg/g, 34 μg/g, 0.50 wt.%, 0.07 wt.%, 9.64).

Distribution of REEs according to sediment grain size

The REE contents (ΣREE) of the sediments were relatively low, ranging from 15.21 to 85.72 μg/g (average 29.99 μg/g), except for sample S05 (293.93 μg/g; Table 3). Overall, the sediments showed light REE (LREE) enrichment features and REE distribution patterns comparable to the background sediments that collected > 60 km away from the ridge axis (Fig. 4), except for the proximal > 250 μm samples and one moderate distance > 250 μm sample (Fig. 4a and b). The sediments showed various Eu and Ce anomalies ranging from 0.66 to 1.12 (average 0.79) and 0.47 to 1.79 (average 0.78), respectively. The > 250 μm sediments exhibited positive and negative Ce and Eu anomalies, ranging from 0.67 to 1.12 and 0.53 to 1.79, respectively. However, the 74–250 and < 74 μm samples all showed negative δEu and δCe anomalies, ranging from 0.66 to 0.75 and 0.66 to 0.95 and 0.49 to 1.09 and 0.51 to 0.88, respectively. The majority of the < 250 μm proximal and moderate distance samples had less negative δEu and δCe anomalies compared to the background sediments (0.71 and 0.57, respectively) (Fig. 4). The δCe and δEu values gradually decreased with decreasing grain size. The ratios of (La/Yb)N, (La/Sm)N, and (Gd/Yb)N in the > 250 μm distal sediments were all remarkably higher (average 6.36, 3.65, and 1.32, respectively; Table 3), indicating significant LREE enrichment.

Fig. 4
figure 4

Chondrite-normalized REE distribution diagrams based on sediment grain size (Taylor and Mclennan, 1985; Boynton, 1984). BS and HTS data were obtained from Liao et al. (2018), hydrothermal plume data from Edmonds and German (2004), SLF data were obtained from Cao et al. (2012), and terrigenous dust data were obtained from Wu et al. (1991). The bright red, turquoise, and dark blue dots represent the proximal distance, moderate distance, and distal sediments, respectively. BS, background sediments; HTS, hydrothermal sediments; SLF, sulfide

Full size image

Discussion

Provenance and composition of the sediments

Much research has shown that MOR sediments mainly consist of calcareous sediments, clastic materials, and hydrothermal components (Cave et al., 2002; Li et al., 2016; Radke et al., 2011). Detrital materials in MOR sediments are mainly derived from terrestrial dust and pyroclastics (Dymond, 1981). Li et al. (2016) found terrestrial dust in sediments sampled about 100 km away from the study area; therefore, a dust component of our studied sediments cannot be excluded. In addition, the sediments from the study area were plotted close to basalt on the Mg–Al–Fe diagram (cfb) (Fig. 5a), implying that mafic debris dominated the sediment content. However, the moderate distance and distal samples showed a relatively high Mg content, implying relatively high ultramafic debris components (Fig. 5a). The occurrence of mafic and ultramafic components in the studied sediments was also supported by the moderate and distal sediments exhibit low Ti/Fe, high Ni/Fe, and low Cu/Fe, Zn/Fe, U/Fe, and ΣREE/Fe ratios that are comparable with local serpernites (Fig. 6), while the proximal sediments show high Ti/Fe, low Ni/Fe, high Zn/Fe, and low ΣREE/Fe ratios comparable with local basalts (Fig. 6a, b, d, f). These results suggest the debris in the moderate and distal sediments is mostly of ultramafic origin while is mostly of mafic origin in the proximal sediments.

Fig. 5
figure 5

a Mg–Al–Fe (cfb) and b 0.1 Fe–Mn–10 Cu diagrams of different grain-sized sediments from different distances to hydrothermal vent in the study area. Database: serpentinite (Chen et al., 2015), basalt (Yang et al., 2017), Rainbow sediments (Chavagnac et al., 2005), and Endeavor sediments (Hrischeva and Scott, 2007). The large solid circles in (a) represent the bulk sediment, and the bright red, turquoise, and dark blue dots represent the proximal, moderate distance, and distal sediments, respectively

Full size image
Fig. 6
figure 6

Fe vs. a Ti, b Ni, c Cu, d Zn, e U, f ΣREE diagrams of different grain-sized sediments from different distances away from confirmed hydrothermal field in the study area. Database: serpentinite (Chen et al., 2015; Zeng et al., 2012) and basalt (Yang et al., 2017). The large solid circles represent the bulk sediment samples, while the bright red, turquoise, and dark blue dots represent the proximal, moderate distance, and distal sediments, respectively. The red dashed line represents fitting line of the proximal sediments, while the blue dashed line represents fitting line of the moderate and distal sediments

Full size image

To further explore the relative contributions of local debris and terrestrial dust in the study area sediments, we used the discriminant function (DF). The DF value was calculated as follows:

$$\mathrm{DF}=|\left({\mathrm{C}}_{1\mathrm{X}}/{\mathrm{C}}_{2\mathrm{X}}\right)/\left({\mathrm{C}}_{1\mathrm{L}}/{\mathrm{C}}_{2\mathrm{L}}\right)-1|$$

where (C1X/C2X) represents the ratio of two elements from the study area sediments and (C1L/C2L) represents the ratio of two elements from a possible source (Li et al., 2017; Wang et al., 2016; Xu et al., 2008). A DF value < 0.5 indicated possible origins, and the lower the DF values are, the closer their relationships are (Li et al., 2017; Wang et al., 2016; Xu et al., 2008). The relative stability of REEs during most geological processes provides a favorable condition for the identification of debris origin. Generally, mafic debris is characterized by heavy REE (HREE) enrichment, while terrigenous clasts and their weathering products are characterized by LREE enrichment. Therefore, Lu/Yb and Sm/Nd ratios were chosen to represent local volcanic (DF1) and dust (DF2) origins, respectively.

In general, the sediments exhibited significantly lower DF1 than DF2 values (Table 4), implying that local debris was the main component. With increasing grain size, the average DF1 value increased from 0.04 to 0.07, and the average DF2 values decreased from 0.40 to 0.10, suggesting decreasing local debris and increasing terrestrial dust components in the finer sediments. The DF2 values for the < 250 μm proximal, moderate distance, and distal samples were all < 0.5, probably attributable to terrestrial dust absorbed by the different grain-sized calcareous ooze. The proximal sediments exhibited mostly lower DF1 values than the moderate distance and distal sediments, except the > 250 μm moderate sediments, suggesting that the sediments adjacent to hydrothermal activities contained more local debris components. This is consistent with MOR hydrothermal activities being mainly controlled by magmatic or tectonic activities. The < 74 μm distal sediments had significantly lower DF2 values than DF1 values, suggesting that the abundance of terrestrial dust significantly increased in these sediments. However, the proximal and moderate sediments had significantly lower DF1 values than DF2 values, indicating local debris was still dominant in these sediments, which is supported by their REE patterns were similar to those of basalt and hydrothermal sediments (Fig. 4a–d).

Table 4 The discriminant function (DF) values of the studied sample
Full size table

Previous studies indicate that sediments on MORs with decreasing Fe/Ti and increasing Al/(Al + Fe + Mn) (typically > 0.4) can be considered less hydrothermal inputs whereas with increasing Fe/Ti and decreasing ratios (typically < 0.1) can be considered as enrichment in hydrothermal metals (Boström, 1973). In the Fe/Ti vs. Al/(Al + Fe + Mn) plot (Fig. 7), the proximal sediments exhibited relatively high Fe/Ti ratio and Fe contents with an average value of 35.55 wt.% and 12.06 wt.%, respectively, and relatively low Al/(Al + Mn + Fe) (average 0.27) compared with those of moderate and distal sediments (average 23.10 wt.%, 6.22 wt.%, and 0.46, respectively), indicating a high hydrothermal composition in proximal sediments (Dias and Barriga, 2006), while the moderate and distal sediments might be less affected by the hydrothermal plumes. In addition, former studies have shown that ore-forming elements are precipitated in sediments during hydrothermal plume dispersal, whereby Cu and Zn precipitate in proximal sediments and Fe and Mn are distally dispersed (Edmonds and German, 2004; Hrischeva and Scott, 2007). The oxidation of Mn in a plume is slower than that of Fe, meaning that Mn is dispersed farther because it exists longer in a dissolved state in a hydrothermal plume after jetting (German et al., 1991). Accordingly, the 0.1 Fe–Mn–10 Cu diagram shows that the proximal sediments are relatively Cu enriched, while the moderate distance and distal sediments are relatively Fe and Mn enriched, respectively (Fig. 5b). Furthermore, the < 250 μm proximal samples, particularly the < 74 μm samples, contained relatively higher Cu contents than the > 250 μm sediments, suggesting hydrothermal component enrichment in the 74–250 and < 74 μm proximal sediments. In addition, there are probably also mass-wasted sulfides in the proximal sediment, especially 74–250 μm sediments, as supported by the higher U/Fe ratios than those of moderate and distal sediments (Fig. 6e; German and Sparks, 1993; Mills and Elderfield, 1995). This observation is consistent with previous studies found up to 6–13% of pyrite and sphalerite grains in proximal metalliferous sediments and up to 70% of > 500 μm coarse-grained sediments comprising sulfides as a part of reworked basalts in the Longqi-1 hydrothermal field on SWIR (Jia et al., 2019). The relatively high Fe contents of the > 250 μm proximal and high Fe, Ni, and Cr contents of moderate distance sediments were probably due to high contents of local basement debris, while the high Cu and Fe contents of the 74–250 μm and < 250 μm moderate distance sediments were probably the result of Fe–Mn (hydr)oxides, local debris, and the hydrothermal composition Fe–Mn (hydr)oxides, and hydrothermal components were discussed separately for convenience in the main text, though Fe–Mn (hydr)oxides in MOR are considered fallouts of the neutral buoyant plume). Therefore, the > 250 μm distal samples S05 showed high Fe and Mn contents (Fig. 3a–b), revealing the existence of Fe–Mn (hydr)oxides (Hrischeva and Scott, 2007).

Fig. 7
figure 7

(a) Fe/Ti vs. Al/(Al + Fe + Mn) plot of sediments from different distances to hydrothermal vent in study area, base map according to Dias and Barriga (2006). The bright red, turquoise, and dark blue solid circles represent the proximal, moderate distance, and distal sediments, respectively

Full size image

Systematic analyses of REEs

 > 250 μm sediments

Despite the differing sampling distances from the confirmed HFs, the samples generally showed high ΣREE contents (Fig. 8a) and various ΣLREE/ΣHREE ratios (Fig. 8b) and δEu and δCe anomalies (Fig. 8c,d). These features probably resulted from the complicated mineral compositions, evidenced by our micro-observations. The REE content of pelagic calcite sediments is controlled by the substitution of Ca2+ by REE in the lattices of carbonate minerals, which is directly controlled by the concentration of REEs in seawater (Zhong and Mucci, 1995). Thus, the REE content of the original sediments was relatively low, and the relatively high REE contents of the samples could be attributed to the greater abundance of mafic debris in these sediments. This is consistent with the micro-observations and the REE distribution patterns being comparable to mafic rocks (Fig. 4a–b). In addition, the proximal sediments showed high ΣREE contents, implying high portions of local debris that probably caused by more intensive volcanic and tectonic activities in these areas, which favor the development of hydrothermal activity. Moreover, the sediments exhibited decreasing ΣREE contents and decreasing δEu and δCe with decreasing grain size, which reflected decreasing local debris content and was consistent with most of the samples exhibiting increasing ΣLREE/ΣHREE ratios.

Fig. 8
figure 8

a ΣREE, b ΣLREE/ΣHREE, c δCe, d δEu, e (La/Yb)N, f (La/Sm)N, and g (Gd/Yb)N in sediments with different grain sizes. 1 =  > 830 μm, 2 = 380–830 μm, 3 = 250–380 μm, 4 = 180–250 μm, 5 = 120–150 μm, 6 = 150–180 μm, 7 = 96–120 μm, 8 = 74–96 μm, and 9 =  < 74 μm. Calculating formula: δEu = Eu N/(0.5Sm N + 0.5Gd N); δCe = Ce N/(0.5La N + 0.5Pr N) (Bau and Dulski, 1996; Bolhar et al., 2004). The bright red, turquoise, and dark blue dots represent the proximal, moderate distance, and distal sediments, respectively

Full size image

However, the > 250 μm distal sample S05 showed a markedly high ΣREE content and ΣLREE/ΣHREE ratios and a positive δCe that decreased with decreasing grain size, and the > 250 μm moderate distance sample S16 showed positive δEu and δCe values (Fig. 4c,e). Studies have shown that Fe–Mn oxides of hydrothermal origin generally feature negative or no Ce anomalies, whereas hydrogenetic Fe–Mn oxides typically display positive Ce anomalies (Elderfield et al., 1981; Kuhn et al., 1998). The positive Ce anomalies in the distal and moderate distance samples were in agreement with this interpretation, probably caused by the mixing of strongly absorbent Fe–Mn hydroxide, consistent with the high Mn contents of these samples (Fig. 5b). A positive δEu in MOR sediments is generally considered a signal of hydrothermal components (Meyer et al., 2013; Wang et al., 2014a, b). However, the occurrence of plagioclase in volcanical debris may also result in such signals, which is consistent with the high pyroclastics debris content in these samples. Furthermore, the moderate sample S16 exhibits REE patterns of LREE/HREE (0.80) lower than proximal S15 (0.83), which contains high mafic debris. Considering the high MgO contents (3.79%) in this sample, we suggest ultramafic debris occurs. Thus, it was concluded that the REE features of the > 250 μm sediments were mainly controlled by the local debris from a host of mafic and ultramafic debris, hydrothermal components, and the Mn oxide content of the samples.

74–250 μm sediments

These sediments showed relatively uniform ΣREE contents (Fig. 8a), probably implying that the volcanic debris was no longer the main REE host, as evidenced by the REE characteristics being comparable with background sediments (Fig. 4a) and having consistently, significantly higher ΣLREE/HREE ratios than those of the > 250 μm sediments (except for sample S05). Importantly, most of the distal and moderate distance sediments exhibited slightly decreasing ΣLREE/ΣHREE (Fig. 8b), (La/Yb)N, (La/Sm)N, and (Gd/Yb)N ratios (Fig. 8e–g) with decreasing grain size, indicating the relative enrichment of HREEs in finer sediments. Because the mafic and ultramafic debris was characterized by HREE enrichment and Fe–Mn oxides favor the absorption of LREEs, these features could be explained by increasing contamination with local debris.

The δCe values showed a generally increasingly negative anomalies with decreasing grain size in all the sediments (Fig. 8c). It could be attributed to the sediments basically inherit the seawater REE signature by adsorption of REE onto hydrothermal Fe oxyhydroxides that often display negative Ce anomalies (German et al., 1990; Edmonds and German, 2004), while diagenetic possible contamination of biogenic carbonate (foram) ooze does not change the overall REE pattern significantly in the result of their adsorbing REE from seawater (but with much lower concentration). The δEu distribution was more complicated, with both increasing and decreasing features observed in the proximal and moderate distance samples (Fig. 8d). Studies have shown that a positive Eu anomaly is generally the result of the mixing of hydrothermal components (Meyer et al., 2013; Wang et al., 2014a, b). However, plagioclase in basaltic debris can also cause positive Eu anomalies. Thus, the variation in the observed δEu values may reflect the contributions of multiple factors in these sediments, which is supported by the proximal sediments of all grain sizes exhibiting comparable DF1 values.

In conclusion, the REE characteristics of the 74–250 μm sediments were resulted from multiple controlling factors, including the hydrothermal components, calcareous ooze, local mafic, and ultramafic debris.

 < 74 μm sediments

The < 74 μm sediments exhibited higher ΣREE contents and slightly higher ΣLREE/ΣHREE ratios than those of the 74–250 μm sediments. They also exhibited the most negative δEu and δCe (Fig. 8c,d), except for distal sample S05. These features probably resulted from the abundant clay minerals and silty sand in these sediments, which was supported by the lowest DF2 values (Table 3) and the high absorption capacity of clay minerals for REEs (Zhao et al., 2010). The significantly negative δEu and δCe were probably due to the influence of biogenic carbonate and seawater. However, the < 74 μm distal sample S05 showed slightly positive δCe anomalies, probably owing to higher Mn oxide contents (Fig. 8c, Fig. 3b), because hydrogenetic Fe–Mn oxides are typically characterized by positive Ce anomalies (Elderfield et al., 1981; Kuhn et al., 1998). Clay minerals itself may also influence the REE features of these sediments. Studies have shown terrigenous dust has slightly positive or no Ce anomalies, negative Eu anomalies, and substantial LREE enrichment features (Gallet et al., 1996; Honda et al., 2004). These features are consistent with the slightly higher ΣLREE/ΣHREE ratios and slightly positive δCe anomalies in sample S05. Overall, the REE contents of the < 74 μm sediments were mainly controlled by the clay mineral, Mn oxide, and the terrigenous dust contents.

Implication for the identification of hydrothermal activity

Previous studies have shown variations in the Y/Ho ratio for different sources of MOR sediments. Terrigenous rocks and chondrites have approximately constant Y/Ho ratios between 26 and 28 (Bolhar et al., 2004); hydrothermal fluids generated on MORs show Y/Ho ratios of 25–28, comparable to those of basalt; and low-temperature fluids related to white chimneys have Y/Ho ratios between 30 and 35 (Bau and Dulski, 1999). Seawater and sediments have Y/Ho ratios between 44 and 72 (Bau and Dulski, 1996), with Fe–Mn crusts having lower Y/Ho ratios than sediments because the absorption capacity of Ho is greater than that of Y in Fe–Mn oxides (Bau and Dulski, 1996, 1999). In this study, most < 250 μm proximal and moderate sediments show Y/Ho ratios < 30 and fall within the range of terrigenous or volcanic sediments, indicating the influence of debris in these coarse-grained components. However, 74–250 μm (28.08–40.00) and < 74 μm (33.16–38.80) sediments exhibit Y/Ho ratios comparable with hydrothermal sediments (~ 33–40), indicating that the fine-grained component contains a more hydrothermal source with the influence of pelagic sediment. The < 74 μm sediments have considerably higher ΣREE and similar Y/Ho with eolian dust, indicating a higher contribution of eolian dust in these sediments.

As described, the sediments contain abundant local debris, hydrothermal, and dust components, and the REE features of the > 250 μm, 74–250 μm, and < 74 μm sediments were primarily controlled by the local debris, hydrothermal, and clay minerals, respectively. Thus, based on the abovementioned possible control factors, we propose that 74–250 μm sediments are the most feasible for indicating potential hydrothermal activity. This conclusion is evident in the δEu–(Gd/Yb)N diagram (Fig. 9), which illustrates the relative abundance of the hydrothermal and debris components (δEu values) and the distance away from the HFs (LREE enrichment, (Gd/Yb)N). The sediments plot as three groups, especially in the > 250 μm sediments. The 74–250 μm proximal and moderate distance sediments show lower (Gd/Yb)N and higher δEu than the distal sediments, indicating greater hydrothermal contributions (Fig. 9). Furthermore, the < 74 μm sediments (particularly the distal sediments) plot in the same area as the 74–250 μm sediments, which is consistent with the higher silt and clay mineral compositions, can adsorb more LREEs.

Fig. 9
figure 9

δEu vs. (Gd/Yb)N indicating the feasible grain size for the identification of hydrothermal contribution. The orange arrow represents increasing hydrothermal contribution. The transparent orange-shaded range represents the REE parameter change of most of the < 250 μm sediments. The bright red, turquoise, and dark blue dots represent proximal, moderate distance, and distal sediments, respectively. Database: BS and HTS (Liao et al. 2018), basalt (Yang et al., 2017), terrigenous dust (Gallet et al., 1996)

Full size image

Conclusions

The sediments sampled from the study area comprised primarily of calcareous carbonates, followed by local basic and ultrabasic debris, hydrothermal components from an overlying plume with mass-wasted sulfide debris in the proximal sediments, Fe–Mn (hydr)oxides, and minor terrigenous dust. Grain size of the studied sediments is closely related to mineral and geochemical composition of sediment provenance. The > 250 μm sediments, particularly the proximal and moderate sediments, contained relatively high amounts of local mafic, ultramafic debris, and hydrothermal components with high REE, Cu, Zn, Fe, Ti, V, Mg, Ni, Cr, and Mn contents; the 74–250 μm sediments contained relatively high amounts of hydrothermal components and local debris with high Cu, Zn, Fe, U contents, U/Fe ratio, and Fe/Mn ratio; the < 74 μm sediments contained relatively high amounts of Fe–Mn (hydr)oxides and terrestrial dust with high Mn and LREE contents. With increasing distance from the HFs, the contribution of hydrothermal components decreased.

The REE features of the different grain-sized sediments were controlled by multiple factors, including mineral sources and distance from the HFs. The > 250 μm sediments showed REE features comparable to local basement rocks, particularly the proximal sediments, which implied that the formation of the HFs was accompanied by intensive volcanic/tectonic activity. The REE features of these sediments were primarily controlled by local mafic and ultramafic debris, hydrothermal components, and Fe–Mn (hydr)oxide contents. The 74–250 μm sediments showed LREE enrichment, negative δCe, and negative δEu features, which were probably controlled by the local debris, calcareous ooze, and hydrothermal components. The < 74 μm sediments contained high amounts of hydrothermal component, Mn oxides, and terrigenous dust contents. We proposed that the REE characteristics of the 74–250 μm sediments are the most feasible for the identification of hydrothermal activity in studied segments.