Variation of heavy metal speciation during the pyrolysis of sediment collected from the Dianchi Lake, China

2.1 Materials

The sediment samples were collected from the Caihai section of the Dianchi Lake. The sediment was freeze-dried, ground, and passed through a 2 mm sieve. Four pretreated sediment powder samples were put on aluminum foil and pyrolyzed in a muffle furnace for 4 h at 200 °C, 300 °C, 400 °C, and 500 °C, respectively. Nitrogen gas was used to purge air out of the muffle furnace. After 4 h of pyrolysis, the furnace was cooled to room temperature, but nitrogen gas was still used to prevent the oxidation of the pyrolyzed sediment. These sediments were noted as B2, B3, B4, and B5 for the temperatures of 200 °C, 300 °C, 400 °C, and 500 °C, respectively.

All reagents used were of analytical grade; deionized water was supplied from a Millipore Milli-Q system. All glassware and plastic containers were previously cleaned with 15% nitric acid and rinsed thoroughly with deionized water.

2.2 Characterization of samples

Elemental (C, H, N, S, O) compositions were analyzed by an elemental analyzer (MicroCube Elementar, Germany). The surface area was measured with N₂ by a surface area analyzer (Autosorb-1C, Quantachrome). The pH was measured in sediment and TTSs extracts (solid/deionized water ratio 1:5 w/v) by a pH meter (Leici PXSJ-216).

2.3 Sequential extraction procedure

Sequential extraction was performed by using the BCR three-step procedure (Cuong and Obbard, 2006), which was proposed by the European Community Bureau of Reference (now the Standards, Measurements and Testing Program). The residual fraction was designed as step four. Four replicates were applied for sequential extraction, in which three replicates were used to determine the heavy metal concentrations and the fourth was used for elemental analysis.

Step one: Briefly, 0.5 g of dry sample of the original sediment or TTSs was put into a 40 mL centrifuge tube. Twenty mL of 0.11 mol/L acetic acid was added to the tube and shaken at 90 rpm for 16 h at 25 °C. The extract was separated from the solid phase through centrifugation at 2000 rpm for 15 min. The supernatant was decanted into a 40 mL centrifuge tube then placed at 4 °C. Twenty mL of de-ionized water was added to the residue and shaken again for 15 min then centrifuged for 15 min at 2000 rpm. The second supernatant was decanted without loss of the solid residue.

Step two: Twenty mL of 0.5 mol/L hydroxylamine hydrochloride adjusted to pH 2.0 was introduced into the solid residue from the first step in the centrifuge tube. The tube was then shaken for 16 h at 90 rpm at 25 °C. The extraction procedure described previously was followed.

Step three: Five mL of 8.8 mol/L hydrogen peroxide (pH 2.0) was slowly added, in small aliquots, into the solid residue from step two in the centrifuge tube. The solid residue was digested at room temperature for 1 h and occasionally shaken by hand. The tube was placed in a water bath heated to 85 °C for 1 h. The reagent in the tube was evaporated to almost dryness then cooled down. Another 5 mL of hydrogen peroxide reagent was added to the tube and again was evaporated to near dryness. After cooling, 25 mL of 1.0 mol/L ammonium acetate (pH 2.0) was added to the tube and the tube was shaken for 16 h at 25 °C. The extract was separated from the solid residue as previously described.

Step four: Five mL of aqua regia was introduced to the remaining solid residue from step three; the aqua regia was evaporated to near dryness at 95 °C in a water bath. After cooling, small aliquots of 25 mL of HNO₃ were added to the solid residue and the extract was separated from the solid residue as previously described.

The metals recovered from the first step (F1) are possibly co-precipitated with carbonates or are exchangeable. The metals from the second step (F2) are associated with Fe/Mn oxides and the third step (F3) is associated with organic matter and sulfides. The residual step (F4) precipitates primary and secondary minerals, which have trace heavy metals within their crystal structures (Gleyzes et al., 2002). The heavy metal concentrations for Cd, Cu, Pb, and Zn in extracts from each step were determined by an atomic absorption spectrometer (Varian AA240FS).

2.4 Risk assessment code (RAC)

The risk of heavy metals released from sediment and TTSs was evaluated by the risk assessment code (RAC), which is defined as the fraction of metals that areexchangeable or bound to carbonates (F1) (Jain, 2004). The RAC classifications are used to evaluate the risks of the four heavy metals.

2.5 Statistical analysis

The principal component analysis (PCA) was carried out using SPSS v13.0 in order to identify similar behaviors of metals (Devesa-Rey et al., 2010; Passos et al., 2010).

3.1 The efficiency of the sequential extraction methods

In order to check the efficiency of sequential extraction methods, an internal check was used by comparing the sum of four fractions (acid-soluble + reducible + oxidizable + residual) with total metal concentrations from aqua regia digestion (Table 1). The recovery was calculated using the equation: Recovery = (F1 + F2 + F3 + F4)/(total concentrations from aqua regia digestion) × 100%. The overall sum of four fractions was in good agreement with total metals concentration and the recovery range was 70.4–127.5%, which means the BCR methods are reliable.

3.2 Total heavy metal concentration

Total heavy metal concentrations of sediment are presented in Table 1. Total concentrations in sediment decreased in the order of Zn > Cu > Pb > Cd. Zn was in the highest concentration at 546.1 mg/kg and Cd was in the lowest concentration at 27.4 mg/kg. Table 2 illustrates the background concentrations of shale and soil in the Dianchi Lake area. The total concentrations of Cd, Pb, and Zn were higher than the background concentrations, whereas, the total concentration of Cu was within the background range. The higher levels of Zn may be the result of urban runoff, which has an elevated Zn concentration from the corrosion of galvanized materials and car washes (Sorme and Lagerkvistb, 2002). The total heavy metal contents of TTSs also show that pyrolysis concentrated the heavy metals (Cantrell et al., 2012). During sediment pyrolysis, volatile matter of total sediment mass was lost in gas phases but the heavy metal mass of the sediment may not be vaporized at pyrolysis temperature below 750 °C (Kistler et al., 1987).

3.3 TTS characterization

3.3.2

3.3.2 Surface area and pH

The surface area of TTSs increased with an increase in temperature (Fig. 1b). Specifically, the surface area of B2 was 7.98 m²/g and that of B5 was 37.8 m²/g. The surface areas of TTSs are much smaller than that of plant residues such as wood charcoal, sugarcane bagasse, and wheat residue (Shinogi and Kanri, 2003; Chun et al., 2004). The composition of parent biomass rather than biochar production methods determined the surface area of biochar (Cao and Harris, 2010). Less surface area of TTSs may be the result of the abundance of mineral fractions.

Close

Figure 1 Change in pH (a), and specific surface area (b) in sediment and TTSs as a function of pyrolysis temperature.

Export to PPT

The pH of TTSs generally increased with increasing pyrolysis temperature (Fig. 1a). The pH value slightly decreased when the sediment was pyrolyzed at 200 °C, because the cellulose and hemicelluloses in sediment decomposed at 200 °C and formed phenolic compounds and organic acids (Abe et al., 1998). After pyrolysis temperature increased to 300 °C, alkali salts separated from the organic matrix and increased the pH. The samples of B3–B5, which had higher pH, can be used as amendments to neutralize soil acidity of red soil in the Yunnan Province.

3.4 The occurrence of heavy metals in Dianchi Lake sediment

As illustrated in Fig. 2, Cd speciations in the sediment were dominant in the fraction associated with Fe/Mn oxides (64.6%) and then in the fraction associated with carbonate (35.4%). The fraction associated with F3 and F4 could not be detected. Tessier et al. (1996) reported that at high lake pH values, Cd²⁺ tended to more dominantly associate with Fe oxyhydroxides rather than with organic matter. The pH values of the water in the Dianchi Lake were 8.41–9.12 (Mo et al., 2007), thus Cd associated strongly with F2. The strong association of Cd to carbonate may be contributed to the similarity of the ionic radii of Cd (0.097 nm) and Ca (0.099 nm). When calcium ions precipitate with carbonates, the cadmium ions diffuse into the calcite crystal lattice as a camouflaged element and co-precipitates with carbonates (Korfali and Jurdi, 2011). The residual fraction of Cd is not detected, which means that Cd is not bound to silicates. Cd may come from anthropogenic sources such as atmospheric deposition or industrial effluents.

Close

Figure 2 Fractionation of Cd, Cu, Pb, and Zn in sediment and TTSs.

Export to PPT

Fig. 2 shows that Cu speciation in sediment followed the order F2 > F3 > F4 > F1. The predominant fraction was associated with Fe/Mn oxides (88%), then with organic matter and sulfides (5.8%), residual fractions (4.4%), and carbonates (1.8%), respectively. As mentioned before, at high lake pH values, Cu²⁺ tended to dominantly associate with Fe oxyhydroxides rather than with organic matter. Much higher concentrations are found in bioavailable fractions (F2 + F1 + F3) than that of F4, which suggests Cu may also come from anthropogenic sources.

In sediment, the chemical form of Pb followed the order F2 > F4 > F3 > F1. The predominant fraction was associated with Fe/Mn oxides (82.3%), then residual fractions (8.4%), organic matter and sulfides (5.9%), and carbonates (3.4%). Again, at high lake pH values, Pb²⁺ also tended to associate with Fe oxyhydroxides rather than with organic matter.

The content of Zn in sediment followed the order F2 > F1. Zn speciation was dominant in the fraction associated with Fe/Mn oxides (51.7%) and followed by the fraction associated with carbonate (48.3%). The stronger association of Zn ions to Fe/Mn oxides may also be due to high lake pH values. The residual fraction (F4) of Zn is not detectable, which confirms that Zn may come from anthropogenic sources.

Comparing the results obtained from Chaohu Lake, China (Xu et al., 2008), except for F3 and F4 of both Cd and Zn, all other fractions concentrations of Cd, Cu, Pb, and Zn in the Dianchi Lake are much higher than that of the Chaohu Lake. This means that the heavy metals Cd, Cu, Pb, and Zn heavily pollute the Dianchi Lake.

3.5 Heavy metals speciation in TTSs

The fractions of heavy metals in TTSs are given in Table 4. The percentages of heavy metals with respect to the sums of the four fractions in sediment and TTSs are shown in Fig. 2.

3.6 Risk assessment code

The metals separating in F1 are thought to be the most labile fraction and pose a greater risk to the environment (Jain, 2004). When the percentage of F1 is less than 1%, the sediment poses no risk to the environment. Percentages of 1–10% pose low risk while 11–30% is considered medium risk, and 31–50% is high risk to the environment. Any percentage over 50% means the sediment poses very high risk and is considered dangerous since heavy metals can easily move into the food chain (Jain, 2004). Fig. 3 shows the RAC applied to the sediment and TTSs. The high content of Zn and Cd in the F1 of all five samples represents that Zn and Cd can be easily removed if environmental conditions such as pH change. The lower content of Cu and Pb in the F1 (<10%) shows these metals pose low risk. Fig. 3 also shows that B4 has the lowest risk for all four heavy metals, due to the lowest percentage of the F1 fraction. Thus, pyrolysis temperature of 400 °C will be the best choice temperature to treat the sediment of the Dianchi Lake.

Close

Figure 3 Risk assessment code (RAC) was applied to Cd, Cu, Pb, and Zn in sediment and TTSs.

Export to PPT

3.7 Principal component analysis

In order to better understand the geochemical behavior of cadmium, copper, lead, and zinc in different fractions in TTSs, PCA was applied to analyze the data of Table 4. For F1 (Fig. 4a), two components were extracted from PCA, which accounts for about 72.7% and 24.6% of total variance. Four main groups are easily visualized. For F2 (Fig. 4b), two components were extracted from PCA, which accounts for about 48.3%, 36.1%, respectively, of total variance. Two components were used to plot Fig. 4b and four groups are easily visualized. The first group included Cd and Zn, the second group included Cu, and the third group included Pb. For F3 (Fig. 4c), two components were extracted from PCA, which accounted for about 73.6% and 24.4% of the total variance. Two components were used to plot Fig. 4c and four groups are easily visualized.

Close

Figure 4 Principal component analysis applied to TTSs fractions. Groupings identified by analysis are circled. PCA-F1 (a), PCA-F2 (b), PCA-F3 (c), PCA-Bioavailable fraction and F4 (d).

Export to PPT

Finally, for the bioavailable fractions (F1 + F2 + F3) and the non-bioavailable fraction (F4), two components were extracted from PCA, which accounted for about 64.9% and 31.6% of the total variance. Two components were used to plot Fig. 4d and six groups are seen. Because Cd and Zn belong to 2B groups in the periodic table of elements (Moore et al., 2011), they are always grouped together in PCA analysis.