Introduction

As, Cd, and Pb are toxic elements that are harmful to human health. Cu, Cr, and Mn are either the essential elements for human health or toxic elements. Zn is an essential element present in all human organs, tissues, fluids, and secretions, and intake and supplementation of Zn is thought to have a beneficial effect (Jomova and Valko 2011).

As, Cd, Cr, and Pb can enter the food chain via plants and vegetables. Food consumption has been identified as the major pathway of human exposure to pollution, and considered to account for more than 90% compared with inhalation and dermal contact (Loutfy et al. 2006). High accumulation and translocation of toxic metals in the edible parts of plants and vegetables were reported in spinach, radish, tomato, chili, and cabbage growing in agricultural fields, and their cultivation is unsafe due to possible transfer to the food chain and consequent health hazards (Tiwari et al. 2011). Transfer capability of the elements to the food chain via plants is determined by both phytoavailability of the elements in soil and their uptake by plants. A correlation between As content in the vegetables and the bioavailable As in soil was observed (Kar et al. 2013). A linear relationship was found between Cd phytoavailability and the CaCl2 extracted Cd/DOC (dissolved organic matter) level (Tang et al. 2017). An approach of assessing phytoavailability was the single extraction procedure (Meers et al. 2007; Pinto et al. 2015). The modified BCR approaches were applied to assessing phytoavailability (Ferri et al. 2015; Tang et al. 2017). BCR was thought as a useful approach to predict mobility of the trace elements (Pueyo et al. 2008). By the modified BCR, Cd, Mn, and Zn were extracted primarily in water and weak acid extraction fractions, but it was thought of as poor predictors of phytoavailability (Ferri et al. 2015). Indeed, the significant correlation between phytoavailability and the extractable fractions was not always observed (Tang et al. 2017). The correlation between soil extractable metal concentrations and plant metal concentrations was usually poor (Ferri et al. 2015; Oyeyiola et al. 2011). No extraction procedure proved to be able to accurately predict the phytoavailability (Fedotov et al. 2012).

The contrast findings imply that it is still a challenge to correlate the extractable fractions of the elements in soil with phytoavailability and uptake of plants. Firstly, not only are the extractable fractions of elements in soil but also the types and families of plants are needed to be considered in such correlations. Secondly, the species and transformation of the elements in soil and rhizosphere may play a key role in the phytoavailability of the elements and their uptake by plants, since speciation of elements in plants is important for understanding the mechanics of elemental uptake and biotransformation in plants (Shen 2014). Thirdly, agricultural activities should make impacts on the phytoavailability and bioaccumulation of the elements.

The phytoavailability of heavy metals for uptake by plants is associated with their sorption and desorption by soil components (Tahervand and Jalali 2016). Lead(II) binds to iron (hydr)oxides, such as ferrihydrite (Nyquist and Greger 2009), or organic matter, such as carboxylate groups (Wang et al. 2006). By forming such complexes with oxygen-containing ligands and singly coordinated OH groups of (hydr)oxides, Pb(II) is strongly adsorbed to soil organic matter (soil pH < 6) and iron (hydro)oxides (soil pH > 6), and will result in low mobility and plant availability of Pb2+ (Gustafsson et al. 2011). On the other hand, it was also thought that a strong relationship between dissolved Pb and DOC was suspected, yet only a weaker correlation was observed (Linde et al. 2007). The sorption and retention of Cd, Cu, and Pb are mainly controlled by cation exchange capability, pH, and Mn oxides’s content (Tahervand and Jalali 2016). Formation of organometallic complexes in soil solution has a strong influence on metal phytoavailability by restricting Pb ions from playing their chemical roles (Shahid et al. 2011; Shen and Song 2017). Though Pb uptake decreased in the presence of river fulvic acid with respect to noncomplexed Pb, total adsorbed Pb was higher than predicted from Pb2+ activities (Slaveykova et al. 2003). Some reports indicated that metal uptake was favored by humic substances because of the high presence of carboxylic acids (Salati et al. 2010), and humic acids contributed to the mobilization of Pb (Laborda et al. 2008). It was reported that humic acids increased the phytoavailability of Pb and Cd to wheat plants with increased extractable metals (Khan et al. 2008). Pb bio-uptake was higher in the presence of river fulvic acid or riverine organic matter (Slaveykova et al. 2003). Thus, the role of organic matter in the phytoavailability of heavy metals for uptake by plants is still needed to conduct a deeper investigation.

Uptake of toxic elements by vegetables poses the potential health risks via dietary consumption of foods. The toxic elements in minerals, such as Cd and Pb in pyrite, pyrrhotite, and chalcopyrite, have potential health risks (Zhuang et al. 2009). In city and industrial areas, Cr was the major pollutant with the highest health quotient (Liu et al. 2013; Zhao et al. 2018). On the basis of an indoor dust investigation, Pb posed the highest health risk (Kurt-Karakus 2012). In a Pb-Zn mine, Pb, As, and Cd were the major risks (Li et al. 2014). In a case of crops with wastewater irrigation, Mn, Cd, and Pb had a health risk index > 1 and indicated a possible human health risk via food crops (Khan et al. 2013). All of these results reveal that health risks exist in areas polluted by mining, industrial, and anthropogenic activities. Obviously, the potential health risks need to be assessed. The health risk index (Khan et al. 2013) or health risk quotient (THQ) (Bortey-Sam et al. 2015) were used to make a quantitative estimation of the risks. THQ is the ratio of the daily intake of metals to the oral reference dose. When the ratio is < 1, none of the exposed population would be likely to experience obvious adverse effects, and if the ratio is > 1, the inhabitants will suffer a relatively high health risk (Bortey-Sam et al. 2015; Kurt-Karakus 2012; Zhuang et al. 2009). High concentrations of Cd and Pb were reported in irrigation water and crops, and the health risk indexes of Cd and Pb were > 1, which poses a possible health risk (Khan et al. 2013). Therefore, the objective of this study is to quantitatively obtain information on the bioaccumulation of metal(loid) elements in the vegetables; to investigate phytoavailability of the elements in soil; to explore the effects of the elemental species, the families of vegetables, and agricultural activities on phytoavailability, bioaccumulation, and plant uptake; and to assess the health risks of the toxic elements in this area.

Materials and methods

Sampling and sample analysis

The investigation site is located in an area near a Pb-Zn-Mn mine (Fig. 1S in supplemental material), which has been in operation for > 50 years. Silver (Ag), gold (Au), Mn, Pb, sulfur (S), and Zn are its major extracts. The residential houses and vegetable fields are around the mine. The mining activities have introduced pollution into agricultural soils and irrigation systems (Luo et al. 2014).

Fig. 1
figure 1

Means and SD of BAFs of Cd and Zn in various vegetables. Spinacia oleracea L. (Spinach) and Amaranthus tricolor (three-colored amaranth) are the plants of Amaranthaceae family, in which the mean of BAFs of Cd and Zn is over 3.0 and 0.6, respectively. Daucus carota L. (Carrot), Apium graveolens (celery), and Coriandrum sativum L. (coriander) belong to Apiaceae family, in which the mean of BAFs of Cd and Zn is over 1.2 and 0.3, respectively

Full size image

Vegetables, plants, rhizosphere soil, field soils, and irrigation water were sampled together. In total, 204 vegetable samples were taken from the agricultural fields and 205 soils samples were taken from the landscape, agricultural fields, residential yards, traffic sites, and industrial areas around the mine.

The plant samples were thoroughly washed using a brush and tap water, and the process was repeated three times with distilled water. The cleaned plants were wrapped in pieces of gauze and left to dry at room temperature. The plant samples were powdered to 76 μm. An average conversion factor was calculated for a ratio of fresh and dry plants on the basis of 34 fresh and dry vegetable samples.

For analysis of plant samples, a set of containers was used, in which an inner polytetrafluoroethene container was within an outer steel container; 0.5000 ± 0.0005 g of powdered sample was weighed into the inner polytetrafluoroethene container, then 5.00 ml of HNO3 and 1.00 ml of H2O2 were added, and the samples kept overnight. The containers were heated at 70–80 °C until yellow smoke containing nitrogen-oxygen components vanished and the solution was almost dry. Next, 5.00 ml of HNO3 and 2.00 ml of H2O2 were added to the inner containers. The steel containers were moved into an oven and kept at 150 °C for 4 h. After the container was taken out and cooled to room temperature, 50 ml of the distilled water was added into the inner container quantitatively; 1 ml of the solution containing the dissolved plant sample was diluted to a volume of 5 ml with a 5% acid solution. The solution was analyzed by inductively coupled plasma atomic emission and mass spectrometry (ICP-AES, Thermo IRIS Advantage, and ICP-MS, Thermo X-series). For quality control, four certified reference materials, GBW10010 (Rice), GBW10015 (Spinach), GBW10015 (tea leaf), and GBW10020 (citrus leaf), were used, and 20% of the plant samples were analyzed in duplicate in addition to the analysis of blank samples. Pb isotopes were determined by Agilent-7500a ICP-MS. Certified reference material, NBS981, was used to monitor the analytical quality. It was found that the determination of Cd in the geochemical samples by ICP-MS was affected by zirconium (Zr) and molybdenum (Mo) oxides and hydroxides, and the application of membrane dissolvation could possibly reduce the interference (WANG et al. 2017). In addition, another challenge of analyzing the polluted soils is the lack of soil certified reference materials with high concentrations of heavy metals. A certified reference material taken from the agricultural soil surrounding the mining districts was reported, in which the concentrations of As, Cd, and Pb was 297, 3.09, and 971 mg/kg, respectively (Tian et al. 2017).

XANES experiments were conducted in the 15 U1 and 14 W1 beamline of the Shanghai Synchrotron Radiation Facility. A double crystal monochromator of Si (111) was applied to achieve an energy resolution of 2.0 × 10−4, and a K-B mirror was used in a focusing optics system to obtain an ultimate beam size of 2 × 2 μm2. Due to low Pb concentrations in the samples, a defocused beam size of 100 × 100 μm2 was used for the XANES measurements. X-ray fluorescence spectra were collected using a Vortex-90EX Si drift detector in both fluorescence mode (samples) and combined fluorescence/transmission mode (standards) measurements.

Sequential extraction

On the basis of the modified BCR sequential extraction procedure (Sutherland 2010; Sutherland and Tack 2003), five fractions were used in the study using the sequential extraction procedure: (1) water soluble, 1.0 g soil sample was extracted with 20 ml distilled water for 16 h at 22 °C; (2) weak acid extractable, 1.0 g soil sample was extracted with 40 ml acetic acid (0.11 M HOAc); (3) reducible, 40 ml of hydroxylamine hydrochloride was added to the washed sample; (4) oxidizable, the sample was dissolved in 10 ml of H2O2 in a sealed bottle; and (5) residual, 0.1000 g residual was weighted into a Teflon container and HCl, HNO3, HClO4, and HF acids (3, 2, 1, and 5 ml) were added to solve it.

The sum of the elemental concentrations in the five fractions were compared with the elemental concentrations in soil obtained by XRF, and the average relative errors were 3.0, 3.0, 10.2, 14.5, and 9.2% for As, Cd, Cu, Pb, and Zn, respectively. The recovery (the ratio of the sum and the concentration) ranged from 92 to 100, 91–102, 101–121, 107–127, and 100–108% for As, Cd, Cu, and Zn, respectively. The results by this study are comparable to the precision and accuracy of 10 and ± 15% by other researchers (Anju and Banerjee 2011; Fleming et al. 2013; Marguı́ et al. 2004). For example, the recoveries by Tokahoglu and Kartal (2004) were 76–102%. The precision by professional laboratories was between 3.7 and 4.6% for the most precise, and 32.6 and 32.9% for the least precise, and the highest precision was associated with the acid extractable step for Cd (6.2%) and the reducible step for Pb (6.3%), while the lowest precision was associated with the oxidizable step for Pb (19.0%) and the acid extractable step for Cr (15.6%) (Sutherland 2010). The higher the concentrations, the greater the analytical precision.

In the study, water and weak acid extractable fractions were included as the soluble fraction, and the sum of the water, weak acid, reducible, and oxidizable extractable fractions were included as the extractable fractions.

Data analysis

Bioaccumulation factors

Bioaccumulation factors (BAFs) were expressed either on a dry weight basis of plants (McAloon and Mason 2003) or a fresh weight basis (Judy et al. 2012). In the present study, BAFs were calculated by the ratios of element concentrations in plants (fresh weights, FW) and soils. BAFs were calculated using the following equation:

$$ \mathrm{BAF}=\frac{C_{\mathrm{vegetable}}}{C_{\mathrm{soil}}} $$
(1)

where Cvegetable represents the heavy metal concentrations in dry vegetables, and Csoil, the heavy metal concentrations in soils.

Statistics

Pearson’s product correlation coefficients were used in the study for statistics analysis, and a two-tailed test of significance was used with a confidence level 0.95. Linear combination and fitting algorithm, and principle component analysis were used to extract the information on the species of Pb in the investigated samples by XANES and Athena software (Ravel and Newville 2005).

Estimated daily intake

The estimated daily intake (EDI, μg kg−1 person−1 day−1) was calculated using the following equation:

$$ \mathrm{EDI}=\frac{C_{\mathrm{FreshVegetable}}{M}_{\mathrm{ingested},}}{B_{\mathrm{w}}} $$
(2)

where CFreshVegetable is the concentration of heavy metals in fresh vegetables (μg g−1, fresh weight), Mingested, the mass of daily vegetable ingested per person per day, or rate of vegetable ingestion (301 g day−1 person−1, carbohydrate), and Bw, the average body weight (62.8 kg, a mean of male (66.2 kg) and female (57.3 kg), respectively. The mass of daily vegetable ingested and the average body weight were taken from the report on nutrition and chronic diseases of Chinese residents (2015).

Target hazard quotient

The target hazard quotient (THQ) is defined as follows:

$$ \mathrm{THQ}=\frac{F_{\mathrm{expfrq}}{D}_{\mathrm{expdr}}{C}_{\mathrm{FreshVegetable}}{M}_{\mathrm{ingested}}}{R_{\mathrm{fDs}}{B}_{\mathrm{w}}T}\times {10}^{-3} $$
(3)

where Fexpfrq is the exposed frequency (365 days per year), Dexpdr, exposure duration (70 years, average lifetime), RfDs, the oral reference dose (mg kg−1 day−1), and T, the averaging exposure time for non-carcinogens (365 days year−1 multiplied by number of exposure years, 70 years in this study). When the same exposed frequency and exposure duration as the average exposure time for non-carcinogens are used in formula (4), i.e., Fexpfrq × Dexpdr = T, the THQ formula is simplified as the following:

$$ \mathrm{THQ}=\frac{C_{\mathrm{FreshVegetable}}{M}_{\mathrm{ingested}}}{R_{\mathrm{fDs}}{B}_{\mathrm{w}}}\times {10}^{-3}=\frac{\mathrm{EDI}}{R_{\mathrm{fDs}}}\times {10}^{-3} $$
(4)

Results

Concentrations of metal and metalloid elements in vegetables

In the vegetables collected from the vegetable fields, the maximum concentrations (Wt., mg/kg, fresh weight) were observed in celery for As (1.38) and Pb (8.94), three-color amaranth for Cd (1.14), coriander for Cr (29.83), soybean stem for Cu (8.50), spinach for Mn (76.88), and radish root for Zn (81.68), respectively. The highest values of As, Cd, Cr, and Pb were 3, 4, 60, and 30 times higher than the maximum allowable concentrations (MACs) of contaminants in foods (GB2762-2012 2012), respectively. The Pb concentrations in all samples of coriander, celery, Chinese radish, and leek exceeded the MACs. Their means are shown in Table 1.

Table 1 Means of concentrations of the elements in the vegetables (Wt, mg∕kg FW 1)) and bioaccumulation factor (BAF)
Full size table

The concentrations of metal(loid) elements in vegetables followed the order: (1) for As, coriander > celery > Chrysanthemum coronarium (C. coronarium) > spinach and leek; (2) for Cd, spinach > coriander > celery > leek; (3) for Cr, coriander > soybean > leek > C. coronarium; and (4) for Pb, coriander > celery > C. coronarium > spinach. Generally, coriander, celery, and spinach contained high concentrations of As, Cd, and Pb. They are very popular vegetables in the area of the study.

Bioaccumulation of elements in vegetables

The bioaccumulation of elements by plants varied with plant species and the elemental properties. In the investigated plants, the maximum BAFs were observed in coriander for As (0.23), Cr (1.79), and Pb (0.42); three-color amaranth for Cd (25.10); leek for Cu (0.64); spinach for Mn (0.65); and radish root for Zn (14.69), respectively. The bioaccumulation of the elements in the popular vegetables is as follows: Cd > Zn > Cu > Mn > Cr > Pb > As. Among the investigation elements, BAF of As was lowest, ranging from 0.001 to 0.23.

On the basis of mean BAFs (Table 1) and the number of the collected samples (not less than five), in the current investigation, bioaccumulation of the elements by the vegetables in the present investigation followed the order: for As, C. coronarium > coriander > leek > spinach > celery and radish > carrot and B. chinensis; for Cd, spinach > coriander > carrot > celery > B. chinensis > leek > radish > C. coronarium; and for Pb, C. coronarium > leek > coriander > radish > spinach > celery > carrot > B. chinensis. The means of BAFs of Cd were over 1.0 in 7 vegetables, accounting for 64%, and much higher than the others. Coriander showed high bioaccumulation of As, Cd, and Pb, being always in the top three of the investigation vegetables.

Bioaccumulation is decided by the plant species and the element properties. Spinach and three-colored amaranth belong to the Amaranthaceae family and their means and standard deviation (SD) of BAFs of Cd were over 3.0, the highest among all of the investigated vegetables, as shown in Fig. 1. Next were carrot, celery, and coriander, belonging to Apiaceae family, and their BAFs of Cd were over 1.0. Both the Amaranthaceae and Apiaceae families have high bioaccumulation of Cd. In the elemental periodic table, Zn is in the same group as Cd. The high BAFs were also observed in spinach and three-colored amaranth of the Amaranthaceae family, as well as in carrot, celery, and coriander of the Apiaceae family. This indicates that both Cd and Zn can accumulate in the plants belonging to both Amaranthaceae and Apiaceae families. This evidence reveals the fact that the plants in the same family have the same bioaccumulation trend of the elements which are in the same group but are a different period.

Extractable fractions of elements in soils and bioaccumulation in plants

Extractable fractions of As

Phytoavailability of elements in soils was determined mainly by the chemical species. Most of As in the whole collected soil samples occurred in the residual fraction, accounting for 58.6–99.2%. In the non-vegetable soils, the concentrations of As in the water soluble fraction were only from 0.055 to 0.99 mg/kg, and As in the weak acid soluble fraction from 0.12 to 0.91 mg/kg, as shown in Table 2. In the vegetable soils, the concentrations of As in the water soluble fraction ranged from 0.55 to 1.46 mg/kg, and As in the weak acid fraction was from 2.02 to 6.19 mg/kg.

Table 2 Mean of elemental concentrations in sequential extraction fractions in soil (mg∕kg)
Full size table

The concentrations of As in the vegetables did not correlate with any of the sequential extraction fractions nor with pH. None of the correlation coefficients was higher than 0.5. However, the concentration of As in B. chinensis and celery were moderately correlated with the soil As. The concentrations of As in C. coronarium was strongly correlated with soil As (r = 0.81, p < 0.05). Partly and moderately, As in the vegetables was correlated with Mn oxides, but not with Fe oxides.

Extractable fractions of Cd

In the whole collected soil samples, Cd was mainly in soluble and reducible fractions and accounted for 35.1 and 35.5%, respectively. In the non-vegetable soils, Cd in the two fractions accounted for 30.3 and 43.8%, respectively. In the vegetable soils, Cd in the soluble fraction increased and accounted for 46.3%, and the reducible Cd accounted for 36.8%. Cd in the extractive fractions accounted for 89.7% in total. However, Cd in a sewage sludge sample from an ore dressing plant contained a very low soluble fraction (3.20%).

In the investigation, correlations were observed between plant Cd and the extractable Cd (r > 0.5, p < 0.05). The plant Cd correlated with the Cd in soluble fraction (r = 0.75, p < 0.05). Cadmium in C. coronarium strongly correlated with soil Cd (r = 0.99, p < 0.05). Cd in B. chinensis and celery moderately correlated with soil Cd. High extractable fractions of Cd was also reported by other researchers. For example, the acid soluble (39.41%) and the reducible (30.44%) fractions were reported in soils around a Pb-Zn mine (Anju and Banerjee 2011). Cadmium in exchangeable and carbonate fractions accounted for 47.53% in soil (Wang et al. 2016a). Soil Cd is highly available (Sterckeman et al. 2009) and tends to be more readily available to plants than many other heavy metals (Shaheen 2009).

Extractable fractions of Cu, Pb, and Zn

In the soil samples, most of the Cu was in the residual fraction with a proportion of 57.7%. Pb was predominantly present in the reducible fraction, accounting for 64.6%. Zinc was mainly in reducible and residual fractions, accounting for 36.03 and 36.02%, respectively. The results are consistent with the reports by other studies. For example, the reducible fraction of Pb accounted for 34.24–77.77% in a Pb-Zn mine area (Anju and Banerjee 2011) and 60.65–64.97% in historically contaminated orchard soil (Fleming et al. 2013).

Typically, the exchangeable and weak acid soluble amount of Pb in soils was low, such as < 0.20% (Marguı́ et al. 2004), 4.7–5.4% (Fleming et al. 2013), or 0.4–36.02% (Anju and Banerjee 2011). In this study, the soluble fraction of Pb accounted for only 1.65%. In the non-vegetable soils, the Pb concentrations in the water soluble fraction ranged from 0.02 to 4.47 mg/kg. In the vegetable soils, however, the Pb concentrations in the water soluble fraction increased and ranged from 0.36 to 6.87 mg/kg, and the reducible fraction of Pb increased to account for 78.5%.

In the sample taken from the sewage sludge in the drain outlet of an ore dressing plant, most of the Pb was found in the residual fraction, as well as in the soluble fraction, accounting for 62.9 and 20.7%, respectively. The dressing procedures changed species of Pb, and resulted in a much higher extractable fraction of the toxic element, which may bring more soluble Pb into ambient soils and stronger risks to ecosystems.

Discussion

Phytoavailability and bioaccumulation of elements

In this investigation, a linear relationship between the bioaccumulation factor of metal(loid)s and their soluble and extractable fractions was found. The relative ratio of the soluble and extractable fractions to the total fractions was first calculated. By comparing the means and SD of the element BAFs, as shown in Fig. 2, the bioaccumulation of As, Cd, Cu, Pb, and Zn was correlated with their relative soluble and extractable proportion in the total fractions based on the non-rhizosphere soil. A linear relationship and a linear multiple function were obtained among the means of BAFs of the five elements and the relative soluble and extractable proportions as follows:

$$ \mathrm{BAF}=0.02517{f}_{\mathrm{soluble}}-0.03132,r=0.87 $$
(5)
$$ \mathrm{BAF}=0.02878{f}_{\mathrm{soluble}}+0.00584{f}_{\mathrm{reducilbe}}-0.00732{f}_{\mathrm{oxidizable}}-0.1425,r=0.97 $$
(6)
Fig. 2
figure 2

Relative ratio (%) of the soluble and extractable fraction to the whole fractions in the vegetable soil, and the means and SD of BAFs of the elements

Full size image

Here, adjusted R-squares were used since the fitting was from a small sample. The linear correlation is shown in Fig. 2S in the supplemental material. The equations imply that the bioaccumulation of the elements by plants is proportional to the amount of the soluble and reducible fractions, and inversely proportional to the oxidizable fraction.

The relationship reveals that the bioaccumulation of the elements in plants and the phytoavailability of the elements in soils are linearly correlated with their relative soluble and extractable proportion, rather than their absolute concentrations. In addition, the soluble proportion of the elements is in the order Cd > Zn > Cu > As > Pb, the extractable order Cd > Pb > Zn > Cu > As, and the bioaccumulation Cd > Zn > Cu > Pb > As. Thus, the phytoavailability of the elements in soils follows the order Cd > Zn > Cu > Pb > As. The phytoavailability of the elements in soils was generally determined by the soluble fraction, and partly by the extractable fraction.

Effects of soil pH on Cd phytoavailability

Soil pH has effects on Cd phytoavailability and a dual effect can be observed. A negative correlation between plant Cd uptake and soil pH (pH 4.5–7.2) was reported (Delcastilho and Chardon 1995). However, it has also been reported that total Cd uptake at pH = 4 was about one fourth of the uptake at pH = 6 (Sterckeman et al. 2011) In the nettle, only Cu and Zn, not Cd, were negatively correlated with pH (Boshoff et al. 2014). In fact, different soils had various responses to acidification treatment and a different optimum pH may exist (Wang et al. 2006).

In the current investigation, Cd in non-leaf vegetables was not correlated with soil pH. Most of the Cd in the vegetables was not correlated with Fe and Mn oxides in soils, as shown in Table 3. However, Cd in the leafy vegetables increased with pH increment at pH 6.61–7.71, with a correlation of r = 0.63 (p < 0.05), as shown in Fig. 3S in the supplemental material. In the study area, pH in the irrigation water system ranged from 7.3 to 9.9, and the soils mainly consisted of clay and carbonate (Luo et al. 2014). Moreover, a complex of a metal with the dissolved organic matter (DOM) was identified (Luo et al. 2016).

Table 3 Correlation coefficients of heavy elements (p < 0.05)
Full size table
Fig. 3
figure 3

Correlation of Pb in B. chinensis L. and in soil (r = 0.77, p < 0.05). When two outliers of the soil samples taken from the vegetable pad near traffic roads were removed, a strong correlation (r = 0.96, p < 0.05) was observed

Full size image

Indeed, lowering the pH increased phytoavailable Cd and enhanced metal uptake (Wang et al. 2006) due to the increase in Cd solubility with proton concentration (Alloway 1995). However, the contrary evidence is that protons in solutions compete with Cd for the binding site on the cell wall and the plasma membrane (Sterckeman et al. 2011). Moreover, most of the sites in the apoplasm are weak acids with pK values similar to that of polygalacturonic acid (Alam et al. 2003). Therefore, the competition between proton and Cd always exists. The competition results in a dual effect of pH on Cd phytoavailability and plant uptake (Wang et al. 2011). Finally, the dual effects would result in a balance. In addition, Cd phytoavailability is affected by the contents of Ca, clay, and DOM in soils. Cadmium is taken up through Ca channels, and both Cd and Ca ions compete for this absorption site (Lindberg et al. 2004).

Effects of Pb species in soils on the phytoavailability

In the current investigation, soil Pb is dominant in the reducible fraction, accounting for 74.75–84.34%. The concentrations of Fe and Mn oxides in the vegetable soils ranged from 3.28 to 4.35% and 0.08 to 1.26%, and the correlation coefficients of Pb-Fe and Pb-Mn oxides were 0.38 and 0.81 (p < 0.05), respectively. Pb in most of the vegetables was moderately correlated with Mn and Zn oxides in soils, but not with Fe oxides, as shown in Table 3. This suggests that Pb is mainly bound to Mn oxides, as well as Zn oxides. In the sequential extraction procedure, the reducible fraction is an implication of an element which is bound to Fe or Mn oxides (Fleming et al. 2013). Though the reducible fraction was the major species of Pb in the investigated soils, it is notable that Pb in plants was not correlated with the reducible fraction nor with any single fraction or pH. No significant effects of the extractable fraction were observed on the phytoavailability of Pb in soils. Pb in most of the vegetables was correlated with soil Pb. The correlation coefficients of five vegetables ranged from 0.51 to 0.86. When two soil outliers from the vegetable pad near busy roads were removed, a strong correlation (r = 0.96, p < 0.05) was observed between vegetable Pb and soil Pb, as shown in Fig. 3. The correlation between soil Pb and plant Pb was also confirmed by other researchers. For example, Pb content in rice is positively correlated with soil Pb (Wang et al. 2016b), and Pb uptake is significantly correlated with pseudo-total content based on aqua regia digestion, as well as with all aggressive extraction procedures (Meers et al. 2007).

The Pb isotope fractions of materials depend only on the sources of minerals or ores in samples, and have no correlation with any anthropogenic activities. Thus, four natural Pb isotopes, 204Pb, 206Pb, 207Pb, and 208Pb, were used to trace the sources of Pb in the research. The ratios of Pb isotopes in the plants collected from the investigated area ranged from 1.126 to 1.170 (206Pb∕207Pb) and 2.088 to 2.163 (208Pb∕206Pb). The Pb data from the plants formed an ellipse in which one end was within the range of irrigation water and another connected with soil samples, which indicates a connection bridge between soil and water, as shown in Fig. 4. In that bridge, there was a mixing between aerosol and gasoline exhaust particles. Pb isotopes in the plants overlapped with Pb minerals. The results reveal that Pb in plants originate mainly from the vegetable soils correlated with Pb minerals and the irrigation water, partly from aerosols and vehicle exhaust particles.

Fig. 4
figure 4

Pb isotopes in the collected samples. Lead isotopes in the plants formed an ellipse, one end being within the range of water, another connected with soils. Pb isotopes in parts of plants overlapped with Pb minerals. In the figure, Minerals means Pb minerals, VehicleExhaust means gasoline and diesel vehicle exhaust particles, VegetableSoils means soils taken from the vegetable fields, Aerosols means Aerosols and dust particles, Water means water samples from water well and irrigation water system, Plant means plants and vegetables

Full size image

The more the soluble components of Pb in soils, the more phytoavailability of it in soils. The oxidizable fraction of Pb in the collected vegetable soils, which was related with organic matter and sulfides, accounted for 4.90%. The average soluble fraction accounted for 1.64% in the total fractions. In a non-rhizosphere loose and dried powder sample from the LP region (M_5_LPSoils), two major components, Pb3(PO4)2 and Pb5(PO4)3Cl, were identified by XANES, which accounted for 51 and 49%, respectively, with an R-factor of 0.002932, as shown in Fig. 5. No soluble component was detectable. When the sample was scanned and measured in XRF mode, a significant XRF peak was detected in one spot. The spot (LPSoils_special) was then measured by XANES. In the micro area, hydrocerussite Pb(CO3)2Pb(OH)2, Pb5(PO4)3Cl, and Pb(NO3)2 were detectable, accounting for 53, 29, and 18%, respectively. The total insoluble components in the micro non-rhizosphere soils accounted for 82%. In a rhizosphere soil sample (M_2_NJSoils), a Pb-FA complex was found by XANES. Three components were identified as the Pb-FA complex, pyromorphite Pb5(PO4)3Cl, and Pb(NO3)2, accounting for 60, 22, and 18%, respectively, with an R-factor of 0.0125. In the micro size of 100 × 100 μm2, the amount of Pb-FA and the soluble Pb compound (Pb(NO3)2 were much larger than those found by the modified BCR with a large amount of soils. In the rhizosphere soil, Pb organic complex and soluble Pb accounted for 78%. Compared with non-rhizosphere soil and soluble fraction, the rhizosphere soil increased the phytoavailability of Pb by providing a large quantity of Pb-FA complex and soluble Pb in soils, and favored Pb uptake by plants.

Fig. 5
figure 5

XANES Spectra of Pb L3 edge in the reference compounds, the collected soils, and the plants taken from the fields

Full size image

Effect of agricultural soils on elemental phytoavailability

Soil treatments with biosolids (Lavado et al. 2005) and biochar (Moreno-Jiménez et al. 2016) have effects on bioavailability of elements in soils. In the current investigation, it was observed that the agricultural activities changed the extractable properties and therefore phytoavailability of the elements in soils. In the soil samples collected from hills, industrial areas, and gas station sites (non-agricultural soils), the residual fraction of As accounted for 90.1–99.2% in the total five fractions. In the vegetable soils, the residual fraction of As decreased and only accounted for 58.6–77.3%. Moreover, the soluble fractions of As increased from 0.03–3.0% in the non-agricultural soils to 3.5–8.7% in the vegetable soils, as shown in Fig. 6. The extractable fractions of As in agricultural soils (22.7–41.4%) was much larger than those (0.84–9.86%) in the non-agricultural soils. Generally, the residual fraction of As decreased in agriculture soils, and the extractable and soluble fractions increased significantly.

Fig. 6
figure 6

Water and weak acid extractable fractions of elements in different soils. In the figure, -Veg means soils from vegetable fields, -Hill means soils from hill site, and -Ind means soils from industrial sites

Full size image

The proportion of the residual fraction to sum fractions of Cd, Cu, and Zn in non-agricultural soils ranged from 4.4 to 42.5%, 52.4 to 79.1%, and 33.3 to 51.8%, respectively. The residual proportion of them in agricultural soils decreased to 7.4–15.0, 28.3–58.3, and 11.9–36.6%, respectively. The proportion of the soluble fraction of them in non-agricultural soils ranged from 13.9 to 51.8, 1.3 to 4.6, and 5.5 to 19.0%, respectively. In the agricultural soils, the soluble proportion increased to 41.5–54.5, 3.4–11.3, and 20.0–45.4%, respectively. The average proportion of the whole extractable fractions of them in agricultural soils was 89.7, 50.9, and 73.5%, respectively, and those in non-agricultural soils was 78.7, 32.8, and 58.5%. Thus, the soluble and extractable fractions of Cd, Cu, and Zn in the agricultural soils were much higher than those in the non-agricultural soils.

The residual fraction of Pb in the non-agricultural soils accounted for 7.0–48.1% in the total fractions. Its residual proportion in the agricultural soils decreased to 7.6–18.8%. The average proportion of the reducible fraction of Pb was 59.7% in non-agricultural soils, and in the agricultural soils, it was 78.5%. From Eq. (6), the same fact can also be observed that the bioaccumulation of the elements in plants is proportional to the amount of the reducible fraction. Generally, the proportion of the extractable fractions of Pb increased from 68.0% in non-agricultural soils to 85.0% in agricultural soils. However, the average proportions of the soluble fractions in the agricultural and non-agricultural soils were 1.6 ± 0.6 and 2.7±3.9%, respectively, and there was no significant difference between them.

In general, the agricultural activities and plant growing increased the extractable fractions and phytoavailability of As, Cd, Cu, Pb, and Zn by decreasing their residual fraction in soils and raising the soluble or extractable fractions.

Assessment of human health risks

In the investigation area, the concentrations of As, Cd, Cr, and Pb in parts of the popular vegetables were over MACs, as described in the results section. They may enter the human body by a routine diet of the vegetables. An assessment on a potential health risk was made on the basis of EDI of the vegetables and THQ.

The estimated daily intake was calculated and compared, as shown in Table 1S in the supplemental materials. Coriander, three-colored amaranth, and celery were the top three vegetables of intakes of toxic elements.

When the THQ is over 1.0, the vegetable containing that element poses a potential risk to human health. The larger the THQ, the higher is the health risk. In this study, as shown in Table 4 and in Fig. 4S(a)–(g) in the supplemental materials, the maximum THQs of As in celery and coriander were over 20, and the maximum THQ of Pb in celery was over 10. The THQs of As, Cd, and Pb in parts of the vegetables were between 5.0 and 10.0. The maximum THQs of Mn were over 1.0 in four vegetables.

Table 4 Target hazard quotient (THQ) of elements in edible vegetables
Full size table

The average THQs of As and Pb in radish, spinach, celery, leek, coriander, C. coronarium, three-colored amaranth, lettuce, and pea were between 1.0 and 5.0, accounting for 69% in 13 vegetables. The vegetables posed binary healthy risks. The average THQs of Cd in spinach, coriander, and three-colored amaranth were over 1.0 and below 3.0, accounting for 23%. Thus, spinach, coriander, and three-colored amaranth had a ternary blended health risks with As, Cd, and Pb. In the investigated agricultural fields near the Pb-Zn mine, therefore, celery, coriander, spinach, three-colored amaranth, C. coronarium, and leek presented the highest potential health risks. The health risks of the elements ranked in order of As > Pb > Cd > Mn > Cu > Zn > Cr.

In addition, a correlation between THQ and BAF was found in parts of the investigated elements. The correlation coefficients of Cd and Cr between THQ and BAF were 0.92 and 0.93 (p < 0.05), respectively. The medium correlations were found between THQ and BAF of Mn (r = 0.64, p < 0.05) and Zn (r = 0.55, p < 0.05). The correlation between THQ and BAF implied that elements in soils posed a direct effect and risk for human health by the biogeochemical chains.

Conclusions

The vegetable fields near the Pb-Zn mine have been polluted by over 50 years of mining activities, and the concentrations of As, Cd, Cr, and Pb in parts of the vegetables grown in the fields exceed the maximum allowable concentrations. The assessment by THQ reveals that the pollution of metal(loid) elements pose a serious risk to human health for the residents living in the area.

Both Cd and Zn can accumulate in plants of both the Amaranthaceae and Apiaceae families. This evidence reveals a new fact that the plants in the same family have the same trend of bioaccumulation of the elements in the same group in the periodic table of chemical elements but are in a different period.

The concentrations of As in the vegetables were not correlated with any of the sequential extraction fractions. The plant Cd was correlated with the Cd in the soluble fraction and extractable fractions in the soils. Pb in most of the vegetables was correlated with soil Pb. The results by XANES reveal that the phytoavailability of Pb increases in the rhizosphere soil with a large quantity of Pb-FA complex and soluble Pb by plant root activities.

The bioaccumulation of the elements in plants and the phytoavailability of the elements in soils were determined by their soluble fraction, as well as by the extractable fraction. A linear relationship was found between the means of BAFs of As, Cd, Cu, Pb, and Zn and their relative soluble proportions in soils. The bioaccumulation of the elements by plants was proportional to the amount of the soluble and reducible fractions, and inversely proportional to the oxidizable fraction.

The agricultural activities and plant growing increased the soluble and extractable fractions and phytoavailability of As, Cd, Cu, Pb, and Zn by decreasing their residual fraction in soils and raising the soluble or extractable fractions.

Since the potential health risks in the area are very real, as shown by the study, the focus of research into the effects of heavy metals on human health from the food chain will be the next step in this investigation.