Optimization of pretreatment procedure for MeHg determination in sediments and its applications
Abstract
Methylmercury (MeHg) in sediment is difficult to be determined due to its low concentration and binding com- pounds like sulfide and organic matter. Moreover, wet sediment samples have been suggested to behave differently from certified reference materials in MeHg analysis. Optimal pretreatment procedure for MeHg determination in sediments has not been ascertained and whether the procedure could apply to sediment samples with complex matrix merits further research. This work firstly compared recovery results of five pretreatment procedures for MeHg determination using ERM-CC580. Using the optimal pretreatment procedure, recovery results were analyzed in different sediment samples after manipulation of moisture content, organic matter, and acid volatile sulfide. The procedure using CuSO4/HNO3 as leaching solutions and mechanical shaking as extraction method was proved to produce the most satisfactory recovery results (100.67 ± 6.75%, mean ± standard deviation). And when moisture content varied from 20 to 80%, average recovery results in sediment samples ranged from 100 to 125%. Furthermore, before and after the manipulation of organic matter or acid volatile sulfide, spiking recovery results varied little and were all within acceptable limit (85~105%). Therefore, the procedure of CuSO4/HNO3-mechanical is proposed as a universal pretreatment method for MeHg determination in sediment samples with various characteristics.
Keywords : Methylmercury . Sediment . Pretreatment . Moisture content . Organic matter . Acid volatile sulfide
Introduction
Mercury (Hg), a toxic metal, is highly persistent and able to travel globally (Poulain and Barkay 2013). In surface waters, inorganic mercury can be methylated to methylmercury (MeHg) (Krabbenhoft and Sunderland 2013), whose content corresponds with changes of mercury inputs (Harris et al. 2007). As a lipophilic and protein-binding neurotoxin, MeHg can pose an even severer threat to human beings after bioaccumulation and biomagnification through food chain (Clarkson and Magos 2006). Dietary intake, especially sea food, is the dominant pathway for general populations to be exposed to MeHg (Jiang et al. 2006). The element of Hg occurs naturally but can be anthropogenically introduced into the environment by industrial activities like mining and smelting. It is suggested that anthropogenic perturbations to global mercury cycle have tripled the mercury content of sur- face waters than that of pre-industrial times (Lamborg et al. 2014). In aquatic system, sediment is both an important sink and source of Hg and also a potential hotspot for Hg methyl- ation (Ullrich et al. 2001). Maximum Hg methylation rate usually occurs at sediment-water interface, leading to the con- siderable MeHg content in surface sediment (Gilmour et al. 1992; Lambertsson and Nilsson 2006). As a result, accurate and feasible analysis of MeHg concentrations in sediment is of great necessity for environmental risk assessment.
In sediment matrices, MeHg is difficult to be isolated due to binding compounds like sulfide and organic substances, especially humic substances (Horvat et al. 1993). Moreover, MeHg concentrations in sediment are rather low, usually as nanogram per gram (Caricchia et al. 1997). With gas chroma- tography and cold vapor atomic fluorescence spectrometry becoming the well-acknowledged analysis system for MeHg (Mao et al. 2008), there is still some dissent over the pretreat- ment procedures, including leaching solutions (Kodamatani et al. 2017a; Liang et al. 2004; Liu et al. 2015), extraction solutions (Liang et al. 1996; Maggi et al. 2009), and extraction methods (Bloom et al. 1997; Gu et al. 2013; He et al. 2015). Thus, it is necessary to specifically optimize the pretreatment procedure for MeHg determination in sediment samples.
Recently, researchers mostly choose dry sediment samples to determine MeHg, either after air-dried (Mikac et al. 1999; Qiu et al. 2005) or freeze-dried (Hoggarth et al. 2015; Meng et al. 2015), but there are still others using wet samples directly for analysis (Mikac et al. 1999; Yu et al. 2012). Also, owing to their different moisture content and chemical compositions, practical sediment samples have been suggested to behave dif- ferently from certified reference materials (CRMs) in MeHg determination (Liang et al. 2004). Whether the optimal pretreat- ment procedure could be applied to sediment samples with distinct characteristics merits further research. Typically, re- searchers would collect worldwide sediment samples to testify their analytical methods, which is quite demanding and lacks truth values (Kodamatani et al. 2017a; Liang et al. 2004; Maggi et al. 2009). In this work, instead of collecting a diversity of sediment samples, we manipulated certain chemical composi- tions of sediment physically to present different features. Then, we spiked methylmercury chloride (MeHgCl) standard solu- tion to the manipulated sediment and analyzed the recovery results.
The objectives of this work are to evaluate the optimal pretreatment procedure for MeHg determination in sediment and then to study the applicability of this procedure to practi- cal sediment samples with different characteristics relative to MeHg determination. For this aim, five pretreatment proce- dures were compared, including CuSO4/HNO3 as leaching solutions with mechanical shaking or manual shaking as ex- traction methods (short as CuSO4/HNO3-mechanical and CuSO4/HNO3-manual), KBr/H2SO4/CuSO4 as leaching solu- tions with mechanical shaking or manual shaking as extrac- tion methods (short as KBr/H2SO4/CuSO4-mechanical and KBr/H2SO4/CuSO4-manual), and KOH/CH3OH as leaching solutions with mechanical shaking as extraction method (short as KOH/CH3OH). And practical sediment samples were ma- nipulated physically to achieve different content of moisture, organic matter (OM), and acid volatile sulfide (AVS). Recovery results of MeHg analysis using the optimal pretreat- ment procedure were compared in sediment samples with and without manipulation.
Materials and methods
Reagents
The following reagents were used for the pretreatment proce- dures for MeHg determination in sediments: 65% nitric acid (Merck, Germany), 36% hydrochloric acid (Gaoheng, Beijing Institute of Chemical Reagents, China), copper sulfate pentahydrate (Sinopharm Chemical Reagent Co., Ltd., China), dichloromethane (J.T.Baker® Chemicals, USA), 98% sulfuric acid (Sinopharm Chemical Reagent Co., Ltd., China), potassium bromide (Sigma-Aldrich, USA), potassium hydrox- ide (Sigma-Aldrich, USA), methanol (LiChrosolv®, Merck, Germany), sodium tetraethylborate (Strem Chemicals Inc., USA), citric acid monohydrate (Sigma-Aldrich, USA), sodium citrate dihydrate (Sigma-Aldrich, USA).
Certified reference materials included ERM-CC580 (MeHg content, 75.5 ± 3.7 ng g−1 Hg, European Reference Materials, Institute for Reference Materials and Measurements, Belgium) and GSD-10 (THg content, 0.28 ± 0.03 μg g−1, GBW07310, IGGE, China). Standard solutions included methylmercury chloride standard solution (65.5 ± 2.5 μg g−1 Hg, GBW08675, National Institute of Metrology, China). ERM-CC580 and GSD-10 were kept at 4 °C in dark. MeHgCl standard solution was diluted to 10.0 mg L−1 Hg by ultra-pure water (18.2 MΩ, Millipore, Darmstadt, Germany) and stored in dark. MeHgCl working solution was stepwise diluted by ultra-pure water when it would be used.
Sediment pretreatment procedures
Around 0.25 g of ERM-CC580 was weighed into a 50 mL polypropylene centrifuge tube (Corning, USA) for each treat- ment. The pretreatment procedures evaluated were as follows. All experiments were performed in triplicate.
CuSO4/HNO3 leaching, CH2Cl2 extraction, mechanical/manual shaking
1.5 mL of 2 mol L−1 CuSO4 and 7.5 mL of 25% (v/v) HNO3 were added to the 50-mL centrifuge tubes with ERM-CC580 and waited 1 h for MeHg to be leached out thoroughly. To realize extraction, 10.0 mL of CH2Cl2 was added to each tube and the mixture was shaken mechanically (with a reciprocat- ing shaker) at 350 r min−1 for 1.5 h (He et al. 2004) or man- ually for 0.5 h (Gu et al. 2013). Different lengths of the ex- traction time were applied according to the reported proce- dures. After leaching and extraction, these tubes were centri- fuged at 3000 r min−1 for 15 min. Then, the mixture was filtered with phase separators (Whatman, GE Healthcare Life Sciences, UK) and the organic phase with MeHg was kept. 4.0 mL of the organic phase was added to tubes with around 20 mL distilled water and 2~3 pieces of boiling stones (Saint-Gobain Performance Plastics, France) in them. These tubes were heated at 65 °C for 6 h to remove organic solvent. After heating, each sample was brought to 20.0 mL with ultra- pure water. 200.0 μL of the extract was pipetted into 40-mL amber glass vials (Agilent Technologies, USA) for MeHg analysis by the MERX-M Automatic Methylmercury System (Brooks Rand Laboratories, USA) following USEPA method 1630 (USEPA 2001).
KBr/H2SO4/CuSO4 leaching, CH2Cl2 extraction, mechanical/manual shaking
5.0 mL of 18% (m/v) KBr dissolved in 5% (v/v) H2SO4 and 1.0 mL of 1 mol L−1 CuSO4 were added to centrifuge tubes with about 0.25 g ERM-CC580 and waited 1 h. Then, 10.0 mL of CH2Cl2 was added to each tube and the mixture was shaken mechanically at 350 r min−1 for 1.5h or manually for 0.5 h. The following procedure and analytical method were the same as CuSO4/HNO3 procedure.
KOH/CH3OH leaching, heating
This pretreatment procedure was based on the published pro- cedure with several improvements (Liang et al. 1996). 3.0 mL of 25% (m/v) KOH/CH3OH was added to centrifuge tubes with around 0.25 g ERM-CC580. After heating at 75 °C for 3 h, 10.0 mL CH2Cl2 and 2.3 mL concentrated HCl were slowly added to each tube. These tubes were shaken at 300 r min−1 for 10 min. The following procedure and analyt- ical method were the same as CuSO4/HNO3 procedure.
Practical sample collection and analysis
Practical surface sediment samples were collected from Xingfu Reservoir, Qingnian Reservoir, and Taihu Lake in 2016. Spatial distribution of the sampling sites is shown in Fig. 1. The two reservoirs are located in Wanshan City of Guizhou Province, southwest of China. Even though they were constructed mainly for drinking and irrigation use, these reservoirs suffered severe mercury pollution from local mer- cury mining area (Du et al. 2016; Li et al. 2009; Qiu et al. 2009). In addition, sediment samples of Taihu Lake (Jiangsu Province) were collected from two lake regions, Meiliang Bay and Zhushan Bay. Taihu Lake is a eutrophic lake with mild mercury pollution and a high organic matter content (Guo 2007; Wang et al. 2012). After collection, all sediment sam- ples were transferred to the lab instantly. Around 200 g sedi- ment samples were separated and centrifuged at 3000 r min−1 for 15 min to extract pore water. Then, the pore water samples were filtered through a 0.22-μm syringe filters (ANPEL Laboratory Technologies (Shanghai) Inc., China) for further analysis. The remaining sediment samples were lyophilized to achieve constant weight and then grounded and homogenized to a size of 200 meshes per inch. All sediment samples pre- pared were stored in amber glass vials with Teflon lids at 4 °C.
Analysis of pore water samples
Sulfate (SO 2−) concentrations of filtered pore water samples were determined by Ion Chromatography (IC6200, WAYEAL, China). The samples were separated using an an- ion column (IC SI-52 4E, 4 mm ID × 250 mm) with the eluent (3.6 mmol L−1 Na2CO3) flow rate of 0.8 mL min−1 and col- umn temperature of 45 °C (Liu et al. 2016). Concentrations of total iron and ferrous iron (Fe2+) were determined using 1,10- phenanthroline method with a UV-visible spectrophotometer (Shanghai Sunny Hengping, 756PC, China) (Tamura et al. 1974). Concentration of ferrous iron (Fe3+) could be obtained by subtracting the concentration of Fe2+ from total iron. Total mercury (THg) concentrations in pore water samples were determined with MERX-T Automatic Total Mercury System (Brooks Rand Laboratories, USA) following USEPA 1631, Revision E (USEPA 2002).
Analysis of sediment samples
The concentrations of total carbon (TC) and total nitrogen (TN) in sediment samples were determined with an elemental ana- lyzer (Elementar, Vario EL III, Germany). The determination of TC and TN was performed in triplicate. Sediment moisture content was measured using a weight loss method with a ly- ophilizer. THg content of lyophilized sediment samples was determined by Leeman mercury analyzer (Leeman Labs Hydra II C, USA) according to the USEPA 7473 (USEPA 2007). MeHg concentrations of the samples were determined according to the predetermined optimal pretreatment proce- dure. Analysis of MeHg content was performed in triplicate.
Manipulation of sediment characteristics
The lyophilized surface sediment samples with different concen- trations of moisture, organic matter, and acid volatile sulfide were achieved through physical methods. Then, we immediately de- termined the MeHg concentrations in the manipulated samples using the predetermined optimal pretreatment procedure. The MeHg concentration analysis experiments were performed in triplicate.
Moisture content
Considering the high background concentrations of MeHg in Xingfu Reservoir and Qingnian Reservoir, sediment samples there were suitable for investigating whether the optimal pretreat- ment procedure could apply to sediment with different moisture content. Lyophilized surface sediment samples from Xingfu Reservoir and Qingnian Reservoir were mixed with different spiked into the slurries to form ~ 8 ng g−1 MeHg. TOC content of samples before and after the removal was determined by the elemental analyzer (Schumacher 2002).
Acid volatile sulfide
Fresh sediment samples collected from Meiliang Bay and Zhushan Bay in Taihu Lake were purged with N2 in order to produce higher levels of AVS (Lee et al. 2000a, b). Specifically, about 50 g sediment slurries were reduced by purging N2 at 300 mL min−1 for 3 days. Sediment samples before and after manipulation were prepared for AVS analysis. As for recovery test, MeHgCl standard solutions were spiked into the slurries to form ~ 8 ng g−1 MeHg. AVS content in sediment samples was determined using Bpurge-and-trap^ method along with methylene blue spectrophotometry (Allen et al. 1993; Lasorsa and Casas 1996). Then, AVS con- tent in wet sediment samples was normalized to dry sediment weight following Eq. (2).
Quality control and statistical analysis
For THg analysis in sediment samples, we used GSD-10 as certified reference material and measured analytical blanks for quality control. The average THg concentra- tion we measured was 279.99 ± 0.02 ng g−1 (mean ± SD, n = 6), which agreed well with the certified value (0.28 ± 0.03 μg g−1). The detection limit for THg was 7 ng Hg in terms of absolute mass. For MeHg analysis, the detection limit was 10 pg Hg in terms of absolute mass. Analytical blanks were lower than detection limit. The linear range is from 5 to 800 pg. All glassware used was cleaned with distilled water three times, soaked in 10% (v/v) HNO3 for at least 48 h, washed with distilled water three times, and finally heated at 500 °C for 2 h before use.
Statistical analysis was performed using SPSS 24.0 soft- ware. The difference among recovery results of each proce- dure was assessed by an independent t test. Analysis of vari- ance (ANOVA) was applied to evaluate the significant differ- ence of means. Significance probabilities (p) were calculated and difference was declared significantly for p < 0.01 in the current work.All mercury-containing waste was properly disposed as hazardous waste.
Results and discussion
Optimization of pretreatment procedure using certified reference material
As the pretreatment procedures were to be applied to MeHg analysis in bulk sediment samples, accuracy, op- erability, and security were considered comprehensively. Therefore, five procedures were selected according to the recent publications involving MeHg analysis in sediment (Kodamatani et al. 2017a; Wang et al. 2018; Yin et al. 2018). The detailed leaching and extraction procedures of five pretreatment procedures using ERM- CC580 are listed in Table 1 and the recovery results are illustrated in Fig. 2. As shown in Fig. 2, CuSO4/HNO3- mechanical procedure had the most satisfying recovery (100.67 ± 6.75%, mean ± SD) among five procedures tested.
With the same leaching solutions, recovery results of mechanical shaking did not differ significantly from those of manual shaking (p > 0.1). Specifically, CuSO4/ HNO3 being leaching solvents, extraction efficiency of manual shaking (0.5 h, 116.70 ± 7.33%) was higher than that of mechanical shaking (350 r min−1 for 1.5 h) and to some extent exceeded the optimum value (100%). Yet, with KBr/H2SO4/CuSO4 being leaching solvents, mechanical shaking could produce higher and relatively more accurate recovery results (82.60 ± 7.87%). This fluctuation might be caused by the inadequate stability and repeatability of manual shaking, which usually were the consequences of individual’s difference in strength. After considering the accuracy and reproducibility, me- chanical shaking was selected instead of manual shak- ing. In addition, KOH/CH3OH could produce decent recovery results as well (86.33 ± 7.95%). However, this procedure required heating as leaching method for 3 h and back-extraction for 6 h, which was rather time-con- suming. And all with mechanical shaking, CuSO4/HNO3 as leaching solvents displayed higher leaching efficiency than KBr/H2SO4/CuSO4 and KOH/CH3OH. As a strong oxidizing acid, HNO3 has a strong ability to destroy the strong embedded sites of MeHg and sediment (Hammerschmidt and Fitzgerald 2001; Liang et al. 2004). The reasons for other procedures producing low- er recovery results might be due to their inadequate leaching abilities or interference with sediment matrix (Horvat et al. 1993; Liang et al. 2004; Tseng et al. 1997).
Generally, recovery results of this work accorded with the reported studies (Table 2). Our recovery results using CuSO4/HNO3 (100.67 ± 6.75%) with mechanical shaking agreed well with He’s results (97.8 ± 10.2%), but the shaking frequency was not described in their work (He et al. 2004). So, this work further clarified and established the pretreatment procedure of CuSO4/ HNO3. Moreover, the pretreatment procedure using KBr/H2SO4/CuSO4 as leaching solutions in this work produced a bit lower recovery results than reported re- sults (~ 100%) (Gu et al. 2013; Kodamatani et al. 2017a). In Kodamatani’s both method C and D, they transferred certain amounts of CH2Cl2 (in the lower lay- er of the mixture) to deionized distilled water (Kodamatani et al. 2017a). It was possible to carry inorganic mercury (in the upper layer of the mixture) as well, which could be methylated to MeHg artifacts dur- ing back-extraction periods (Bloom et al. 1999). In this work, phase separators were used to avoid inorganic Hg to move into the CH2Cl2 phase. Therefore, the differ- ence between the separation methods might result in the minor distinction of the recovery results. As to Gu’s pretreatment procedure, even though the average recov- ery (104 ± 15%) was similar to the result of the current procedure, the standard deviation of theirs is somewhat higher than other procedures (from 2.56 to 9.09%). This might be related to the potential instability of manual shaking. In addition, pretreatment procedure using KOH/CH3OH as leaching solutions in this work pro- duced slightly lower recovery results than reported re- sults (100.18 ± 2.56%) (Liang et al. 1996). Yet, Liang’s procedure was dependent on heating process but varia- tions of heating efficiencies of different heaters would bring about difficulty in repeating.
Thus, after comparing the recovery results, efficiency and reproducibility of the five pretreatment procedures, CuSO4/ HNO3-mechanical procedure was determined as the optimal pretreatment procedure in the present work. And its applica- bility to sediment samples with different characteristics would be further examined.
Analysis of practical samples
As shown in Table 3, characteristics of surface sediment sam- ples from different sampling sites varied greatly. Sediment samples from reservoirs of Guizhou Province were rich in THg content (especially Qingnian Reservoir), which might be due to their short distance from Wanshan mer- cury mine area. However, THg concentrations in Meiliang Bay and Zhushan Bay from Taihu Lake were much lower (< 1/20) than Guizhou, indicating the mild Hg disturbance by human activities as mentioned be- fore. Moreover, MeHg concentrations in Xingfu Reservoir and Qingnian Reservoir (over 2 ng g−1) were comparatively higher than those from Meiliang Bay and Zhushan Bay (under 0.8 ng g−1). Therefore, according to the detection limit, lyophilized surface sediment from Xingfu Reservoir and Qingnian Reservoir was suitable for the verification of the application of the optimal pretreatment procedure to sediment with different mois- ture content, without being spiked with MeHgCl stan- dard solution.
From the analysis of pore water in fresh sediment samples, concentrations of SO42−, Fe2+, and Fe3+ were different in Xingfu Reservoir and Qingnian Reservoir (Table 4). Thus, sediment samples from Xingfu Reservoir and Qingnian Reservoir could represent two different water bodies.
In addition, MeHg concentrations in sediment sam- ples from Meiliang Bay and Zhushan Bay were very low and would be under detection limit after mixing with water. They were considered for investigating the influence of OM and AVS on the optimal procedure with spiking MeHgCl. After mixture with water and subsampling, the final MeHg concentration in the Taihu sediment samples would be under 0.008 ng, which was far lower than the content of spiking stan- dard solution (0.08 ng). Thus, sediment samples in Meiliang Bay and Zhushan Bay were suitable to be analyzed on whether the pretreatment procedure would still apply to sediment samples after the manipulation of OM and AVS.
Application to sediment with various moisture content
Compared with soil (~ 40%), surface sediment tends to have high levels of moisture content (~ 70%). However, moisture in sediment might affect MeHg determination unpredictably. On the one hand, Hg methylating micro- organisms prefer moist and warm conditions, which might induce higher MeHg content during the pretreat- ment procedure with heating process included (Kodamatani et al. 2017b). On the other hand, as wet sediment samples were usually more viscous, it is diffi- cult to obtain a homogenous subsample. But whether the moisture content affects the predetermined optimal pretreatment procedure for MeHg analysis remains un-HNO3-mechanical procedure to wet sediment samples, different aliquots of ultra-pure water were added to ly- ophilized sediment samples to produce sediment sam- ples with a variety of moisture content.
The recovery results of wet sediment samples with various moisture content are illustrated in Fig. 3. With moisture content varying from 20 to 80%, recovery re- sults in both reservoirs were mostly within acceptable range (100~125%). Specifically, sediment samples from Xingfu Reservoir could produce slightly higher recovery results than Qingnian Reservoir. Somehow, all the re- covery results were above 100%, which might be caused by the original deviation in MeHg determination of lyophilized sediment samples. The reason for CuSO4/ HNO3-mechanical procedure producing satisfying recov- ery might be that HNO3 could destroy the bond of MeHg and moisture in wet sediment samples. So it would enable MeHg to be leached out (Liang et al. 2004).
Generally according to the results, CuSO4/HNO3-me- chanical procedure could be applied to MeHg determi- nation in wet sediment samples with various moisture content.
Application to sediment with distinct organic matter content
Sediment organic matter, like humic substances, was able to bind MeHg so strongly that MeHg became dif- ficult to be leached out completely (Caricchia et al. 1997; Schartup et al. 2012). As organic matter in sedi- ment or soil samples are difficult to be measured direct- ly, we used the content of TOC to represent the level of organic matter. Sediment samples with low or high or- ganic matter were achieved by heating or not. Content of TOC and TN before and after the removal is shown in Table 5. After being muffled for 8 h, the TOC con- tent in surface sediment samples dropped markedly (from over 1% to less than 0.5%) compared with TN content. The results indicated that after the removal, sediment samples could be used as contrasts containing low organic matter in comparison with the original sed- iment samples. The MeHgCl spiking recovery results in these comparison groups were analyzed to determine whether the predetermined optimal pretreatment proce- dure could apply to sediment samples rich or lacking in organic matter.
The recovery results are illustrated in Fig. 4, and all the results in wet sediment samples were converted to the ratio of MeHg content in lyophilized samples fol- lowing Eq. (1). As shown in Fig. 4, despite the varia- tion in TOC content, the spiking recoveries of MeHg did not differ a lot. Before the removal of organic mat- ter, while the TOC content was higher than 1%, the recovery results in both surface sediment samples of Taihu Lake were near 100%. Then, after the removal, as the TOC content dropped sharply, the recoveries in Meiliang Bay dropped a little but were still within ac- ceptable limit (> 85%). In Zhushan Bay sediment samples, the recovery results increased to around 100% after the removal. Overall speaking, the removal of the organic matter did not influence the spiking recovery results of Taihu surface sediment samples remarkably.According to the spiking recovery results above, the pretreatment procedure using CuSO4/HNO3 as leaching solutions with mechanical shaking as extracting method applies to sediments samples with distinct content of organic matter.
Application to sediment with distinct acid volatile sulfide content
Sulfide, especially AVS, is the most reactive phase for most metals in sediment, Hg and MeHg included (Lee et al. 2000b; Rickard and Morse 2005). The content of AVS is able to reflect the sulfide that can bond with MeHg tightly (Zhu et al. 2017). After the manipulation of AVS, the reduced sediment samples had the AVS content over 2 μg g−1, while the content of the original sample were lower than 1.2 μg g−1 (dry weight) (Table 6). The increase of AVS content in reduced sed- iment samples was consistent with the trend of Lee’s (Lee et al. 2000b). As a result, sediment samples purged with N2 could be used as contrasts to the original sed- iment samples which are rich in sulfide.
After spiking the slurries with MeHgCl standard so- lution, the recovery results can help decide whether the optimal pretreatment procedure could apply to sediment samples with different AVS content. The recovery re- sults are illustrated in Fig. 5. Before manipulation, while the AVS content was comparatively low, the re- covery results were around 92%. After the manipulation, the recoveries increased a little (by the ratio of 8.9% and 2.3% respectively) with the increase of AVS con- tent. Generally speaking, all the recovery results were near 100% within the mentioned range of AVS content. Therefore, the pretreatment procedure using CuSO4/ HNO3 as leaching solutions with mechanical shaking as extraction method could accurately determine MeHg content in sediment samples with a variety of AVS content.
Conclusions
The current work compared recovery results of five pre- treatment procedures for MeHg analysis in sediment samples using ERM-CC580. And the procedure using CuSO4/HNO3 as leaching solutions with mechanical shaking as extraction method produced the most satisfy- ing recovery result, which was 100.67 ± 6.75% in average. In addition, moisture content in sediment sam- ples (from 20 to 80%) had little influence on the ana- lytical performance of the optimal pretreatment proce- dure. Considering the strong complexation ability of or- ganic matter and sulfide with MeHg, sediment samples were manipulated physically to produce a relatively wide range of TOC (from ~ 0.3 to ~ 1.5%) and AVS (from ~ 1.1 to ~ 2.7 μg g−1, dry weight) content. And the spiking recovery results were mostly around 100%, indicating that the optimal pretreatment procedure was able to produce satisfactory results for MeHg determi- nation in sediment samples with various properties.
There is possibility that spiking recovery tests may not entirely reflect the real performance of MeHg in sediment samples. It is likely that natural compounds may bind with MeHg so tightly that they cannot be extracted easily (Qian et al. 2002). Even so, spiking recovery tests have been carried out in plenty of work to prove the accuracy of their analytical methods when the background MeHg concentration was under detec- tion limit (Heyes et al. 2004; Horvat et al. 1993; Liang et al. 2004). Also, inevitably, the manipulation of one characteristic in sediment may bring about changes to other characteristics as well. However, con- sidering the manipulation methods used are mainly through physical instead of chemical means, the proce- dures can be well-controlled (Lee et al. 2000b). And using the same sediment samples in MeHg recovery tests could eliminate the influence of other irrelevant characteristics, like background MeHg content and par- ticle size. Still, if time and energy permit, researchers should collect sediment samples as various as NG25 possible.