Thiol-functionalized silica microspheres for online preconcentration and determination of mercury species in seawater by high performance liquid chromatography and inductively coupled plasma mass spectrometry

Abstract

Thiol-functionalized silica microspheres (SH–SiO₂) were prepared from aminosilica by usage of 3-mercaptopropionic acid for online mercury preconcentration. The adsorbent offered adsorption capacities of 27.4 mg g⁻¹ for Hg²⁺, 62.1 mg g⁻¹ for MeHg and 59.6 mg g⁻¹ for EtHg. The enrichment could be directly performed without any column preconditioning, where commonly coexisting ions in seawater and sample pH (1.0–9.0) as well as the sampling flow rate caused no significant adverse effects. The adsorbed Hg-analytes (Hg²⁺, MeHg and EtHg) could be readily eluted (less than 20 s) from the surface of SH–SiO₂ with 60 μL of 100 mM L-cysteine (Cys) + 0.1 M HCl. Hence the enrichment factors of 830 for Hg²⁺, 916 for MeHg and 883 for EtHg were obtained using 5 mL sample in a 35 s enrichment procedure. The detection limits for Hg²⁺, MeHg and EtHg were 0.019, 0.013 and 0.014 ng L⁻¹, respectively, and the relative standard deviations of peak height and peak area were all below 4% for 5 ng L⁻¹. Mercury speciation in ten seawater samples was then analyzed by the proposed method, giving rise to satisfactory spiking recoveries (90–102%).

---

1. Introduction

Mercury has become a global environmental concern, especially in the form of methylmercury (MeHg), by virtue of global transport and the biogeochemical cycle.¹ The toxicity and bioavailability of mercury are species-specific. It was reported that organomercuric compounds are generally more toxic than inorganic mercuric species and elemental mercury.² The earth's oceans supply human beings with hundreds and thousands of different kinds of seafood. Considering their high bioaccumulation and biomagnification in the food chain, the amounts of mercury species in seawater are vital to the quality of seafood. Besides, mercury speciation analysis in seawater is beneficial to further understand the biogeochemical cycling of mercury.² The development of accurate and sensitive analytical methods for ultra-trace mercury speciation analysis in seawater is hence highly desirable.

Analytical methods for mercury speciation are predominately based on hyphenated techniques, combining chromatographic separation with atomic spectroscopic detection, such as atomic fluorescence spectroscopy (AFS), atomic absorption spectroscopy (AAS), inductively coupled plasma atomic emission spectroscopy (ICP-AES) and mass spectrometry (ICP-MS), which were comprehensively reviewed by Gao et al.³ The selection of high performance liquid chromatography (HPLC) avoids derivatization by gas chromatography (GC) and arduous fabrication of the interface by capillary electrophoresis (CE). The use of ICP-MS not only offers ultra-sensitive element-specific detection capabilities over other atomic spectroscopies, but also eliminates vapor generation required by AFS and hence reduces environmental pollution. The coupling of HPLC to ICP-MS has been adopted for mercury speciation analysis in seawater,^4–9 offering detection limits downscaled to several to several tens ng L⁻¹ for various mercury forms (without any preconcentration).

Mercury species in seawater samples are normally at trace and even ultra-trace levels. For instance, the concentration of total mercury is below 1 ng L⁻¹ in the open ocean water.¹⁰ In these cases, a preliminary preconcentration step is preferred. Prevailing preconcentration techniques, such as liquid–liquid micro-extraction,^11–15 cloud point extraction,^16,17 and solid phase (micro-) extraction (SPE and SPME)^{5–7,18–22} have been utilized for the enrichment of mercury species. Besides, mercury preconcentration was also achieved by other strategies including stir bar sorptive extraction,²³ single drop micro-extraction^24,25 and flow injection displacement sorption.²⁶ Among them, solid phase (micro-) extraction is most attractive since it allows high-yield preconcentration and feasibility of online hyphenation with other methods (HPLC, etc.) and automation for sample preparation.²⁷ As the heart of SPE, absorbents for mercury enrichment are paid much attention to, ranging from reversed-phase materials such as C₁₈ ^{5,6,18,19,21,22} to cation⁷ and anion⁹ exchange columns. However, column preconditioning before preconcentration was usually performed, which not only produced some amounts of waste but also decreased the analytical throughput. Besides, the enriching process might be interfered and deteriorated by some factors such as sample pH and coexisting ions.

Many researchers made their efforts to improve preconcentration through new and modified materials. Many sulfur-containing materials including sodium diethyldithiocarbamate immobilized polyurethane foam,²⁰ thiol-rich polyhedral oligomeric silsesquioxane,²⁸ 2-(3-(2-aminoethylthio)-propylthio)-ethanamine modified silica gel,²⁹ a 3-mercaptopropyltrimethoxysilane coated capillary³⁰ and thiol-functionalized metal–organic framework nanocomposite³¹ were fabricated to selectively adsorb various mercury species. In addition, ethylene glycol bis-mercaptoacetate modified 3-(trimethoxysilyl)-1-propanethiol coated magnetic nanoparticles,³² thiodiethanethiol modified silica coated magnetic nanoparticles³³ and 2-(2′-benzothiazolylazo)-p-cresol functionalized polystyrene–divinylbenzene resin³⁴ were also utilized. Besides, other materials such as oxidized³⁵ and silver modified magnetic³⁶ carbon nanotubes, silver and gold nanoparticle membrane filters,³⁷ magnetic gold nanoparticle microspheres,³⁸ N-(pyridin-2-ylmethyl)ethenamine coated magnetic nanoparticles,³⁹ tetradecyl(trihexyl)phosphoniumchloride immobilized polystyrene–divinylbenzene resin⁴⁰ and 1,3-bis(2-ethoxyphenyl)triazene modified octadecyl silica membrane disks⁴¹ were also prepared for mercury preconcentration. However, to the best of our knowledge, the online application of preconcentration and determination of mercury species by SPE and HPLC-ICP-MS using thiol-functionalized materials has not been reported.

In this work, we synthesized thiol-functionalized silica microspheres (SH–SiO₂) based on condensation reaction between amino-terminated silica microspheres and 3-mercaptopropionic acid. Then the preconcentration column was fabricated by packing SH–SiO₂ microspheres in a 10 mm demountable hollow stainless steel guard column. The new absorbent could directly retain mercury species in the flow-through sample due to the reaction between mercurial compounds and the mercapto groups, and thus the preconditioning step was eliminated. The enriched mercury species were eluted out of the preconcentration column and then separated by a C₁₈ column and followingly detected by ICP-MS. The preconcentration conditions including the type, concentration and volume of the eluant, the sample volume and flow rate, and matrix effect were investigated. To validate this method, the analysis of a certified reference material of seawater (GBW(E) 080042) and spike recovery test were also performed. Finally, the proposed method was applied for speciation analysis of mercury in real seawater samples.

2. Materials and methods

2.1. Chemicals and reagents

All reagents of analytical or chromatographic grade were used. Throughout the experiment ultrapure water with a resistivity of 18.2 MΩ cm was used, which was newly prepared from a Milli-Q Plus water purification system (Millipore, Bedford, MA, USA) just before use to inhibit the possible mercury contamination during the storage. 3-Mercaptopropionic acid from Aladdin Chemistry Co., Ltd (Shanghai, China) and aminosilica microspheres (5 μm, ≈5% m m⁻¹ amino group) from Borui Bonded Chromatography Co., Ltd (Tianjin, China) were used in the fabrication of SH–SiO₂. N,N′-Dicyclohexylcarbodiimide, N-hydroxybenzotriazole and dichloromethane as the solvents for in the fabrication of SH–SiO₂ were also purchased from Aladdin Chemistry. Methylmercury chloride (≥95%) and ethylmercury chloride (≥95%) were both purchased from Alfa Aesar (A Johnson Matthey Company, MA, USA) to prepare stock standard solutions of MeHg and EtHg at 1000 mg L⁻¹ (as Hg) in methanol. Then they were stored in amber glass bottles and kept at 4 °C in the dark. A stock standard solution of 1000 mg L⁻¹ Hg²⁺ in 5% nitric acid was obtained from National Standard Material Center (GSBG 62069-90, Beijing, China). Demountable hollow stainless steel guard columns (4.6 mm i.d. × 10 mm long) were supplied by Dikma Technologies Inc. (Beijing, China). Work standard mixture solutions containing the three mercury species were prepared by successive dilution of the above stock solutions in the ultrapure water. 1000 mg L⁻¹ bismuth stock solution prepared by dissolving bismuth(III) nitrate pentahydrate from Sigma-Aldrich (St. Louis, MO, USA) in 1% HNO₃ was employed to prepare the internal standard solution (5 μg L⁻¹).⁴² L-Cysteine (Cys) was purchased from Sigma-Aldrich for preparing the mobile phase (10 mM Cys at pH 8.0) and the eluant (100 mM Cys in 0.1 M HCl). High-purity nitric and hydrochloric acid (Jiangyin Chemical Reagents, Jiangyin, China) was used in the experiment. Extreme attention must be paid to all reagents for obtaining low blank mercury values, e.g., 100 mM Cys in 0.1 M HCl was checked before everyday use in term of mercury concentration. Once contaminated, they were prepared again and even re-purchased from the suppliers. A certified reference material (CRM) of seawater (GBW(E) 080042) from National Standard Material Center (Beijing, China) was used to validate the accuracy for mercury speciation analysis in seawater. Before use, all the utensils were soaked in 10% HNO₃ for 24 h, and rinsed thoroughly in ultrapure water. All solutions were filtered through membranes of 0.45 μm pore size before analysis.

2.2. Fabrication of mercaptopropionic modified SiO₂

The SH–SiO₂ microspheres were fabricated according to Varghese et al.⁴³ with minor modification (Fig. 1a). N,N′-Dicyclohexylcarbodiimide (2.32 g, 11.3 mmol) and N-hydroxybenzotriazole (1.73 g, 11.3 mmol) were added into a solution of 3-mercaptopropanoic acid (0.8 mL, 9.4 mmol) in dichloromethane (18 mL) in a flat-bottom flask and stirred for 1 h at room temperature. Then 0.6 g aminosilica microsphere (2.1 mmol –NH₂) was added. The reaction mixture was left overnight for 24 h at room temperature. After the completion of the reaction, the product was sequentially washed with dichloromethane, ethyl acetate, methanol, ultrapure water and methanol, and collected by centrifugation at 3000 rpm for 5 min. SH–SiO₂ (0.1 g) was dispersed in a mixture of dichloromethane (12 mL) and methanol (8 mL) under ultrasonication for 5 min. The suspension was then packed downward into a demountable hollow stainless steel guard column under 350 bar for 30 min using MeOH as the displacement liquid. The packed SPE column was then washed with methanol and ultrapure water at a flow rate of 2 mL min⁻¹ for 1 h.

Fig. 1 Scheme demonstrating of the fabrication of SH–SiO₂ (a) and mercury preconcentration by SH–SiO₂ (b).

2.3. Instrumentation

The system comprising of HPLC separation and ICP-MS detection with online mercury preconcentration by SiO₂–SH packed column was similar to our previous work.⁹ Chromatographic separation of mercury species was carried out on a home-assembled HPLC system, which consisted of a reversed-phase C₁₈ column (Diamonsil (2) C₁₈, 5 μm, 4.6 mm i.d. × 150 mm long, Dikma Technologies Inc., Beijing, China), a high pressure pump (Pump 1) with a 0.01–5 mL min⁻¹ flow rate range (Model PU-985, Jasco, Japan) and a six-port injection valve (Valve 1) with a 60 μL sample loop (Rheodyne 7175, Rheodyne, LP, Rohnert Park, CA, USA). The preconcentration system was composed of another high pressure pump (Pump 2) with a 0.001–10.000 mL min⁻¹ flow rate range (Model LC-20AT, Shimadzu (Suzhou) Co., Ltd, Suzhou, China), another Rheodyne six-port injection valve (Valve 2) with a 5 mL sample loop (Rheodyne 7175, Rheodyne, LP, Rohnert Park, CA, USA), a Valco six-port valve (Valve 3) (Valco Instruments Co. Inc., Houston, Texas, USA) and a SH–SiO₂ packed column (5 μm, 4.6 mm i.d. × 10 mm long, Dikma Technologies Inc., Beijing, China). The SH–SiO₂ packed column substituted for the sample loop of the Valco valve for online preconcentration. An argon ICP-MS (X Series^‖, Thermo Fisher Scientific Inc., USA) was operated in the time-resolved analysis mode to detect mercury at a mass isotope of 202. The outlet of the reversed-phase C₁₈ column was directly connected to the concentric nebulizer (TR-30-A1, Meinhard Glass Products, USA) of ICP-MS via PTFE tubing (150 mm long, 0.25 mm i.d.) with appropriate fittings. The separation column was equilibrated with the mobile phase at a flow rate of 1.5 mL min⁻¹ for at least 0.5 h before separation. Adjustment of pH was accomplished under the assistant of a HI 98128 pH-meter (Hanna Instrument, Italia). The room temperature of the HPLC-ICP-MS system was kept at 25 °C. 5 μg L⁻¹ bismuth (²⁰⁹Bi) was added into the mobile phase to serve as the internal standard. The optimized ICP-MS and HPLC operating conditions were summarized in Table 1.

Table 1 Operating conditions of the HPLC-ICP-MS system

Parameters	Value
ICP-MS system
RF power, W	1200
Sampler cone (orifice diameter, mm)	1.1
Skimmer cone (orifice diameter, mm)	0.9
Cooling gas, L min^−1	13.02
Auxiliary gas, L min^−1	0.75
Nebulizer gas, L min^−1	0.88
Isotope monitored	^202Hg, ^209Bi
Dwell time, ms	100
Data acquisition mode	Time resolved analysis
Acquisition time, min	9
HPLC system
Analytical columns	Diamonsil (2) C18 (5 μm, 150 mm × 4.6 mm i.d.)
Mobile phase	10 mM Cys at pH 8.0
Flow rate of mobile phase, mL min^−1	1.5
Sample loop	SiO2–SH packed column
Preconcentration system
Preconcentration column	SiO2–SH packed column (5 μm, 10 mm × 4.6 mm i.d.)
Sample volume, mL	5
Sample flow rate, mL min^−1	10
Eluant	0.1 M Cys and 0.1 M HCl
Eluent volume, μL	60

---

Scanning electron microscopy (SEM) images were conducted by an FEI Sirion SEM instrument with an EDAX energy dispersive X-ray spectrometer (FEI, Netherlands). The surface images of the graphene coating were obtained at 25 kV. IR spectra were collected on a Thermo Nicolet iS5 Fourier transform infrared (FT-IR) spectrometer (Thermo Fisher Scientific Inc., USA). XRD measurements were conducted on a D8 ADVANCE X-ray diffraction system (Bruker, Swiss). Centrifugation in the experiment was performed using a centrifuge (Model 80-1, Jiangsu Zhenji Instruments Co., Ltd, Jintan, Jiangsu, China).

2.4. Preconcentration procedure

SPE preconcentration of mercurial compounds was performed as shown in Fig. 1b according to our previous work with slight modification.⁹ After a standard or sample was loaded into the 5 mL sample loop of Valve 2, Pump 2 was switched on to feed the carrier (ultrapure water) at a flow rate of 10 mL min⁻¹. Mercury species were trapped on the column via the reaction with mercapto groups. After 35 s, Pump 2 was switched off and the sample loop of Valve 1 was filled with the eluant (100 mM Cys + 0.1 M HCl). When Valves 1 and 3 were both changed to the “Inject” position, Pump 1 was then switched on at 1.5 mL min⁻¹. Mercury species trapped on the column were eluted out and then analyzed by HPLC-ICP-MS. After 8 min, Pump 1 was switched off. The above procedure was repeated for another injection.

2.5. Sample collection

Ten seawater samples were collected from the surface water area along the coast of Taizhou, Zhejiang Province, China. They were stored in an ice-filled box and immediately transported to the laboratory for immediate analysis. The seawater samples were directly analyzed after filtration through membranes of 0.22 μm pore size. The certified reference material was 10-fold diluted with ultrapure water prior to analysis due to high concentrations of mercury species. The spike recovery tests for all seawater samples were performed by spiking 100 ng L⁻¹ for Hg²⁺ and 10 ng L⁻¹ for MeHg and EtHg.

3. Results and discussions

3.1. Characterization of SH–SiO₂

According to the FT-IR spectra in Fig. 2A, the most characteristic features were the adsorption bands corresponding to the C=O of carboxyl stretching at 1635 cm⁻¹ and the amide II of the –C–N stretching and –C–N–H deformation at 1580 cm⁻¹, which confirmed the formation of amide bond. Besides, the increased intensities of the –CH₂ stretching vibration asymmetrically at 2930 cm⁻¹ and symmetrically at 2850 cm⁻¹ also indicated the binding of 3-mercaptopropionyl group to the aminosilica surface. The XRD patterns were then used to evaluate the preparation of SH–SiO₂ (Fig. 2B), which demonstrated the existence of the characteristic signal at 2θ = 20–25° for SiO₂. As shown by the SEM micrographs of aminosilica and SH–SiO₂ (Fig. 2C), the smooth surface of aminosilica turned rough, indicating the modification with mercapto group. The EDX images were further used to evaluate the preparation of SH–SiO₂ (Fig. 2D). Compared with the aminosilica (a), the peak of carbon was significantly enhanced and a new peak of sulfur was observed for SH–SiO₂ (b), indicating the successful immobilization of 3-mercaptopropionyl group on the surface of aminosilica.

Fig. 2 (A) IR spectra of aminosilica, 3-mercaptopropionic acid and SH–SiO₂ microspheres. (B) XRD patterns of aminosilica and SH–SiO₂. (C) SEM micrographs of aminosilica and SH–SiO₂. (D) EDX images of aminosilica and SH–SiO₂.

The adsorption capacities of SH–SiO₂ toward Hg²⁺, MeHg and EtHg were investigated. 500 mL standard solution individually containing 20 mg L⁻¹ Hg²⁺, MeHg and EtHg and 100 mL ultrapure water were successively flowed through the SH–SiO₂ packed column at 10 mL min⁻¹ and the waste was collected. The captured mercury species were then eluted out by 100 mL of the eluant (100 mM Cys and 0.1 M HCl) and washed by 100 mL ultrapure water. Both the two eluates were collected and combined. Because of the high concentrations of mercury species in the resulting solution, 25 mL of this solution (the combined eluates) was diluted to 500 mL with ultrapure water. Concentrations of mercurial compounds in the waste and eluate were determined by the proposed HPLC-ICP-MS. The summary amounts of the residual mercury in the waste and the retained mercury in the eluate made up 91–95% of the original mercury in the standard solutions. The maximum capacities of SH–SiO₂ were found to be 27.4 ± 2.6 mg g⁻¹ for Hg²⁺, 62.1 ± 4.8 mg g⁻¹ for MeHg and 59.6 ± 5.5 mg g⁻¹ for EtHg. It demonstrated its promising prospect for the removal of mercury and clean-up of contaminated water. The difference in the adsorption capacities of mercury species could be rooted in the fact that single mercury atom of Hg²⁺ was bounded to two –SH groups whereas MeHg and EtHg were both complexed with mercapto group one to one.²⁸ Considering the high capacity of SH–SiO₂ for mercury species, the maximum sample volume for mercury preconcentration by the proposed SH–SiO₂ packed column was as high as 2700 L providing that total concentrations of the three mercurial compounds was about 1 μg L⁻¹. In comparison with other adsorbents (see Table S1†), SH–SiO₂ exhibited moderate adsorption characteristics. For instance, the proposed adsorbent offered better adsorption performance than most silica-based adsorbents and other organic materials. However, the capacities of SH–SiO₂ were significantly inferior to those of thiol modified metal–organic framework and nanoparticles, which were characterized by ultra-high surface area. Since the amino density on the surface of the aminosilica microspheres was no more than 4%, the obtained SH–SiO₂ had relatively low thiol density, hence restricting the adsorption capacity. In the near future, SH–SiO₂ with denser –SH groups are anticipated to be fabricated from aminosilica containing higher content of –NH₂ groups.

3.2. Optimization of the preconcentration conditions

Generally, mercurial compounds have high affinities for the sulfur-ligand (for instance mercapto group and thione functionalities). As shown in Fig. 1b, the SH–SiO₂ packed column could adsorb mercury species from standards or water samples owing to the binding reaction of mercapto group with mercury. Hence the SH–SiO₂ packed column could be directly utilized for the enrichment of mercury species without any preconditioning. To efficiently elute out the trapped target analytes (Hg²⁺, MeHg and EtHg) from the preconcentration column, three sulfur-containing compounds, namely, Cys, thiourea and 2-mercaptoethanol, individually served as the eluant. The best elution with the highest intensities and narrowest width was obtained with Cys as the eluant. Hence Cys was selected for the following experiment. Since a competitive complexation of Cys with mercury species against mercapto groups on the column took place during the elution, Cys concentration may affect the elution of mercury species. Its concentration was then optimized in the range of 5–200 mM (Fig. S1†). As the Cys concentration raised from 5 to 100 mM, the peak height of the eluting profile for each mercury species was gradually increased. The peak heights nearly levelled off on further increasing the concentration to 200 mM. Therefore, 100 mM Cys was selected. It is well-known that the pH value has an important effect on the complexation. Hydrochloric acid was added into 100 mM Cys to decrease pH and thus the concentration of HCl was estimated in the range of 0.01–0.5 M (Fig. S2†). The peak intensities were significantly enhanced with increasing HCl concentration from 0.01–0.1 M whereas the ascendance became very slow from 0.1 to 0.5 M. As a result, 0.1 M HCl was chosen to add into the 100 mM Cys solution. Followingly, the volume of the eluant was optimized in the range of 1–200 μL (Fig. S3†). The peak intensities were gradually increased with the eluant volume from 1 to 5 μL. Minute ascendances (by 13–19%) were observed on further increasing the eluant volume to 60 μL and nearly stable peak intensities were obtained with continuously increasing to 200 μL. Although the eluant volume above 5 μL was sufficient for the rapid elution of enriched mercury species, 60 μL of 100 mM Cys + 0.1 M HCl was yet employed in the elution step to confirm no residual of mercury species to affect the next preconcentration trial.

The sample pH on mercury preconcentration was also optimized in a pH range of 1.0–9.0 with 5 ng L⁻¹ Hg-species standard solution as the sample. The experimental results demonstrated that the enrichment factors (EFs) were independent of the sample pH value over the tested range. Since the pH values of the collected seawater samples were variable between 4.1–7.9, they were directly analyzed without pH adjustment. The influence of the sample flow rate on the enrichment factors of mercury species was evaluated in the range of 1–10 mL min⁻¹. No significant variation (<3.1%) of the EFs of Hg²⁺, MeHg and EtHg was found with increasing the sample flow rate from 1 to 10 mL min⁻¹. It indicated the independent relationship between the preconcentration efficiency and the sample flow rate. Therefore, the preconcentration procedure was performed at a flow rate of 10 mL min⁻¹ (the maximal flow rate of the pump). In this case, the consumed time in the preconcentration step was reduced to 35 s. However, it was probable to further reduce the preconcentration time with higher sample flow rates (for instance by usage of preparative high-pressure pumps with 10–100 mL min⁻¹) since the back pressure of the preconcentration column was just 84 bar at 10 mL min⁻¹. Then the effect of the sample volume on the enrichment factors of mercury species was further investigated in the range of 0.5–50 mL with 5 ng L⁻¹ Hg-species standard solution as the sample. A good linear relationship between the EFs and the sample volume from 0.5 mL to 50 mL was observed (Fig. S4†). It indicated that lower detection limits could be achieved by using more sample volume. Nevertheless, it was to some extent subjected to a loss in the analytical throughput. In this study, a sample volume of 5 mL was considered as a good compromise between the enrichment factor and the analytical throughput.

The interferences of commonly coexisting ions in real water samples to the enrichment of the target Hg-analytes were also investigated. 5 ng L⁻¹ each of the three mercury species was prepared in the matrices containing various amounts of interfering ions (50 g L⁻¹ of Na⁺, 5000 mg L⁻¹ of K⁺, Mg²⁺, Ca²⁺, Ba²⁺, Al³⁺, Sr²⁺, SO₄²⁻, CO₃²⁻, NO₃⁻, 2000 mg L⁻¹ of Br⁻ and I⁻, 1000 mg L⁻¹ of Cu²⁺, Pb²⁺, Ni²⁺, Zn²⁺ and Co³⁺, and 50 g L⁻¹ of Cl⁻), and then they were analyzed with the online preconcentration procedure. The intensities of Hg²⁺, MeHg and EtHg prepared in these matrices were about 90–103% of those prepared in ultrapure water, which meant no significant interference from coexisting ions in the online enrichment method. Since the concentrations of the above mentioned coexisting ions in the environmental water samples are normally below the above tested concentrations, the enrichment of mercury species in water samples could be directly performed by SH–SiO₂ without any sample pretreatment. It was worth noting that the selectivity of SH–SiO₂ was to some extent better than that of the commercial strong anion exchange particles.⁹

3.3. Analytical figures of merit of the online SH–SiO₂ preconcentration HPLC-ICP-MS system

A mobile phase of 10 mM Cys (pH 8.0) was selected for mercury speciation analysis according to the work by Jia et al.⁷ Baseline separation could be achieved within 8 min with 10 mM Cys (pH 8.0) as the mobile phase. Typical chromatograms of 5 ng L⁻¹ Hg-mixture standard solution without and with preconcentration, and 5 μg L⁻¹ Hg-mixture standard solution without online preconcentration by HPLC-ICP-MS with this mobile phase are demonstrated in Fig. 3a–c. The EFs of the analytes were calculated as the ratios of the concentrations with and without online preconcentration (with 5 ng L⁻¹ Hg-mixture standard solution as the sample), which were 830 for Hg²⁺, 916 for MeHg and 883 for EtHg. Under the optimized conditions, a series of 0.2, 1.0, 5.0, 20, 50 and 200 ng L⁻¹ standard mixture solutions were analyzed in ten replicates for 5 ng L⁻¹ and triple replicates for the other standards by the proposed method. Linear calibration curves were obtained from peak areas of these standard solutions. The detection limits were defined as the concentrations yielding a signal equal to three times the noise and provided in Table 2. The retention time, repeatabilities of peak height and peak area based on ten chromatograms of 5 ng L⁻¹ Hg-mixture standard solution are also shown in Table 2. The relative standard deviations of peak heights and peak areas for mercury species were all below 4% while the repeatabilities of migration time and peak width were superior to 2%. These results proved satisfactory precisions of the proposed method.

Fig. 3 Chromatographic separations of Hg²⁺, MeHg and EtHg in 5.0 ng L⁻¹ without (a) and with (b) online preconcentration, in 5 μg L⁻¹ Hg-species standard solution without online preconcentration (c), and in Seawater 1 (d) and GBW(E) 080042 (e) with online preconcentration by HPLC-ICP-MS. The inset graphs were the full scope chromatograms of Seawater 1 (d) and GBW(E) 080042 (e).

Table 2 Retention time, detection limit, linearity, repeatability and enrichment factor of mercury species

Analyte	Retention time (tR ± SD, min)	Linear range (ng L^−1)	R	Detection limit (ng L^−1)	Enrichment factor	Repeatability (RSD%, n = 10)
						Peak height	Peak area
Hg^2+	1.95 ± 0.03 (1.5)	0.2–200	0.9990	0.019	830	2.2	3.1
MeHg	3.18 ± 0.04 (1.3)	0.2–200	0.9988	0.013	916	2.7	3.4
EtHg	7.42 ± 0.07 (0.9)	0.2–200	0.9993	0.014	883	3.0	3.8

---

The proposed method offered three benefits over the established HPLC methods with online or offline preconcentration (see Table S1†). Firstly, the detection limits of this work were superior to those of most methods, which could be ascribed from high enrichment factors of SH–SiO₂ and powerful detection capacity of ICP-MS. However, the LODs for EtHg was a little inferior to the value by Jia et al.⁷ due to low sample volume. Furthermore, owing to more active adsorption sites of the commercial anion exchange column (quaternary ammonium) than those of SH–SiO₂ (thiol group), slight superior LODs over the proposed method were achieved by our previous work.⁹ Secondly, the preconcentration time of the proposed method was the lowest (0.7 min), which indicated observable enhancement of the analytical throughput. It could be rooted in that the high affinities of SH–SiO₂ to mercury species eliminated column preconditioning before preconcentration. Thirdly, the aqueous cysteine solution of 10 mM served as the mobile phase, which was quite suitable to the ICP-MS instrument. Furthermore, water samples could be directly analyzed by the proposed method without any pretreatment because of the independent relationship between SH–SiO₂ and sample pH plus coexisting ions.

3.4. Determination of mercury species in water

The feasibility of the proposed method was demonstrated by mercury speciation analysis of ten seawater samples. Concentrations of mercury species in these samples are shown in Table 3. Fig. 3d demonstrates a typical chromatogram of Seawater 1. Hg²⁺ was found the dominant mercury species in seawater samples and its percentage in total mercury was above 83%. The recovery tests (spiked with 100 ng L⁻¹ for Hg²⁺ and 10 ng L⁻¹ for MeHg and EtHg) were performed and also provided in Table 3. The recoveries varying from 90 to 102% validated the accuracy of the proposed method. Analysis of the certified reference material of seawater (GBW(E) 080042) for total mercury as well as its spike recovery test was then carried out to validate the accuracy of the proposed method. Hg²⁺ and MeHg were the predominant mercury species in the material from Fig. 3e while EtHg accounted for at most 1% of mercury forms was also found in the CRM. Contents of Hg²⁺ and MeHg in the GBW(E) 080042 agreed well with the determined values by Jia et al.⁷ Total mercury content of 0.95 ± 0.03 μg L⁻¹ in the seawater (GBW(E) 080042) equal to the summary amount of the three detected mercury species was in good accordance with the certified value (1.00 ± 0.06 μg L⁻¹). As there is no commercially available certified reference material of water-based matrix for mercury speciation, comparison of the results of the proposed analytical method with those of a contemporary method was recommended to determine accuracy by ICH guideline.⁴⁴ Mercury speciation in these marine water samples was thereby analyzed by online mercury preconcentration by strong anion exchange column and HPLC-ICP-MS detection based on our previous work⁹ and the results are also listed in Table 3. Most of the values determined by the proposed method were in good accordance (95–105%) with the contemporary method though a slight bias (88–125%) was observed for values below 20 ng L⁻¹. All results indicated the proposed method could be used for accurate determination of ultra-trace mercury species in the unpolluted environmental water.

Table 3 Analytical results for mercury speciation analysis in seawater samples (n = 3)

Sample	Proposed method						Previous HPLC-ICP-MS^a
	Hg^2+		MeHg		EtHg		Hg^2+	MeHg	EtHg
	Conc.^b (%)	R^c (%)	Conc. (%)	R (%)	Conc. (%)	R (%)	Conc. (%)	Conc. (%)	Conc. (%)
a Online mercury preconcentration by strong anion exchange column and HPLC-ICP-MS determination according to our previous work.^9b Concentration in ng L^−1.c Recovery (%, n = 3).d Not detected.
GBW(E) 080042	606 ± 16 (2.6)	96 ± 2	332 ± 14 (4.2)	98 ± 3	9.1 ± 0.1 (1.6)	95 ± 2	626 ± 19 (3.0)	348 ± 11 (3.1)	9.8 ± 0.4 (3.6)
Seawater 1	69.2 ± 1.6 (2.3)	94 ± 3	2.1 ± 0.1 (4.3)	98 ± 2	ND (−)^d	90 ± 5	70.5 ± 2.7 (3.8)	2.3 ± 0.1 (4.8)	ND (−)^d
Seawater 2	129.2 ± 3.5 (2.7)	93 ± 2	1.6 ± 0.1 (4.4)	99 ± 2	ND (−)	92 ± 4	134.4 ± 4.7 (3.5)	2.0 ± 0.1 (5.2)	ND (−)
Seawater 3	134.5 ± 3.9 (2.9)	93 ± 2	10.3 ± 0.3 (2.7)	101 ± 3	ND (−)	91 ± 4	129.7 ± 4.4 (3.4)	11.2 ± 0.5 (4.4)	ND (−)
Seawater 4	140.8 ± 4.4 (3.1)	94 ± 2	10.6 ± 0.3 (2.6)	102 ± 3	ND (−)	91 ± 4	134.3 ± 4.7 (3.5)	11.5 ± 0.5 (4.5)	ND (−)
Seawater 5	22.6 ± 0.8 (3.5)	98 ± 4	ND (−)	99 ± 3	ND (−)	94 ± 3	23.5 ± 0.8 (3.6)	ND (−)	ND (−)
Seawater 6	163.2 ± 4.0 (2.5)	91 ± 2	13.7 ± 0.4 (3.0)	96 ± 3	17.6 ± 0.6 (3.6)	92 ± 5	155.6 ± 4.8 (3.1)	15.1 ± 0.5 (3.6)	18.2 ± 0.7 (3.8)
Seawater 7	42.2 ± 0.9 (2.1)	92 ± 3	ND (−)	94 ± 2	ND (−)	94 ± 5	44.2 ± 1.9 (4.2)	ND (−)	ND (−)
Seawater 8	182.1 ± 4.8 (2.6)	92 ± 2	12.0 ± 0.4 (3.3)	96 ± 2	14.9 ± 0.5 (3.5)	98 ± 4	178.0 ± 5.7 (3.2)	10.6 ± 0.4 (3.9)	14.2 ± 0.6 (4.3)
Seawater 9	71.2 ± 2.2 (3.2)	95 ± 3	ND (−)	98 ± 2	ND (−)	96 ± 4	72.6 ± 3.0 (4.2)	ND (−)	ND (−)
Seawater 10	140.1 ± 4.3 (3.1)	94 ± 3	ND (−)	95 ± 2	ND (−)	94 ± 4	146.1 ± 4.5 (3.1)	ND (−)	ND (−)

---

4. Conclusions

Thiol-functionalized silica microspheres were synthesized for online SPE preconcentration of mercury species to improve the analytical performance of HPLC-ICP-MS. Owing to the presence of thiol groups, the adsorption of mercurial compounds by SH–SiO₂ was nearly independent of pH and commonly coexisting ions, showing its potential in direct enrichment of real water samples with any pretreatment. The new adsorbent also possessed favorable adsorption capacities for Hg²⁺, MeHg and EtHg, indicating its promising prospect for the removal of mercury and clean-up of contaminated water. The enrichment factors of Hg²⁺, MeHg and EtHg around 900-fold were obtained using 5 mL sample in a 35 s enrichment procedure. The density of thiol group at the surface of aminosilica microspheres would be further increased and higher adsorption capacities and better enrichment results are anticipated in the near future.

Acknowledgements

The authors acknowledge the financial support from the National Natural Science Foundation of China under project nos 21275038 and 21405030, the Analysis and Measurement Foundation of Zhejiang Province under project nos 2014C37094 and 2014C37100, and the Scientific Research Foundation of Zhejiang Provincial Department of Education under project no. (Y201430445).

References

1. N. E. Selin, Annu. Rev. Environ. Res., 2009, 34, 43–63.
2. T. Syversen and P. Kaur, J. Trace Elem. Med. Biol., 2012, 26, 215–226.
3. Y. Gao, Z. M. Shi, Z. Long, P. Wu, C. B. Zheng and X. D. Hou, Microchem. J., 2012, 103, 1–14.
4. A. Castillo, A. F. Roig-Navarro and O. J. Pozo, Anal. Chim. Acta, 2006, 577, 18–25.
5. W. R. L. Cairns, M. Ranaldo, R. Hennebelle, C. Turetta, G. Capodaglio, C. F. Ferrari, A. Dommergue, P. Cescon and C. Barbante, Anal. Chim. Acta, 2008, 622, 62–69.
6. Y. G. Yin, M. Chen, J. F. Peng, J. F. Liu and G. B. Jiang, Talanta, 2010, 81, 1788–1792.
7. X. Y. Jia, D. R. Gong, Y. Han, C. Wei, T. C. Duan and H. T. Chen, Talanta, 2012, 88, 724–729.
8. X. P. Chen, C. Han, H. Y. Cheng, Y. C. Wang, J. H. Liu, Z. G. Xu and L. Hu, J. Chromatogr., A, 2013, 1314, 86–93.
9. H. Y. Cheng, C. L. Wu, L. H. Shen, J. H. Liu and Z. G. Xu, Anal. Chim. Acta, 2014, 828, 9–16.
10. K. Leopold, M. Foulkes and P. Worsfold, Anal. Chim. Acta, 2010, 663, 127–138.
11. F. Moreno, T. García-Barrera and J. L. Gómez-Ariza, J. Chromatogr., A, 2013, 1300, 43–50.
12. L. Xia, B. Hu and Y. Wu, J. Chromatogr., A, 2007, 1173, 44–51.
13. Z. B. Gao and X. G. Ma, Anal. Chim. Acta, 2011, 702, 50–55.
14. X. Y. Jia, Y. Han, X. L. Liu, T. C. Duan and H. T. Chen, Spectrochim. Acta, Part B, 2011, 66, 88–92.
15. X. Y. Jia, Y. Han, C. Wei, T. C. Duan and H. T. Chen, J. Anal. At. Spectrom., 2011, 26, 1380–1386.
16. H. T. Chen, J. G. Chen, X. Z. Jin and D. Y. Wei, J. Hazard. Mater., 2009, 172, 1282–1287.
17. J. G. Chen, H. W. Chen, X. Z. Jin and H. T. Chen, Talanta, 2009, 77, 1381–1387.
18. J. Margetínová, P. Houserová-Pelcová and V. Kubáň, Anal. Chim. Acta, 2008, 615, 115–123.
19. T. Hashempur, M. K. Rofouei and A. R. Khorrami, Microchem. J., 2008, 89, 131–136.
20. J. S. dos Santos, M. de la Guárdia, A. Pastor and M. L. P. dos Santos, Talanta, 2009, 80, 207–211.
21. B. R. Vermillion and R. J. M. Hudson, Anal. Bioanal. Chem., 2007, 388, 341–352.
22. Y. K. Tsoi, S. Tam and K. S. Y. Leung, J. Anal. At. Spectrom., 2010, 25, 1758–1762.
23. R. Ito, M. Kawaguchi, N. Sakui, N. Okanouchi, K. Saito, Y. Seto and H. Nakazawa, Talanta, 2009, 77, 1295–1298.
24. F. Pena-Pereira, I. Lavilla, C. Bendicho, L. Vidal and A. Canals, Talanta, 2009, 78, 537–541.
25. E. M. Martinis and R. G. Wuilloud, J. Anal. At. Spectrom., 2010, 25, 1432–1439.
26. L. M. Dong, X. P. Yan, Y. Li, Y. Jiang, S. W. Wang and D. Q. Jiang, J. Chromatogr., A, 2004, 1036, 119–125.
27. B. Buszewski and M. Szultka, Crit. Rev. Anal. Chem., 2012, 42, 198–213.
28. W. Wang, M. Chen, X. Chen and J. Wang, Chem. Eng. J., 2014, 242, 62–68.
29. M. Puanngam, P. K. Dasgupta and F. Unob, Talanta, 2012, 99, 1040–1045.
30. X. H. Bai and Z. F. Fan, Microchim. Acta, 2010, 170, 107–112.
31. M. Sohrabi, Microchim. Acta, 2014, 181, 435–444.
32. M. H. Mashhadizadeh, M. Amoli-Diva, M. R. Shapouri and H. Afruzi, Food Chem., 2014, 151, 300–305.
33. A. Beiraghi, K. Pourghazi and M. Amoli-Diva, Anal. Lett., 2014, 47, 1210–1223.
34. V. A. Lemos and L. O. dos Santos, Food Chem., 2014, 149, 203–207.
35. B. Parodi, A. Londonio, G. Polla, M. Savio and P. Smichowski, J. Anal. At. Spectrom., 2014, 29, 880–885.
36. C. G. Yuan, J. Wang, W. Zhai, Y. Zhang, Y. Zhang, J. Li and Q. Zhang, Int. J. Environ. Anal. Chem., 2013, 93, 1513–1524.
37. N. Panichev, M. M. Kalumba and K. L. Mandiwana, Anal. Chim. Acta, 2014, 813, 56–62.
38. W. B. Zhang, C. X. Sun and X. A. Yang, Anal. Methods, 2014, 6, 2876–2882.
39. E. Najafi, F. Aboufazeli, H. R. Lotfi Zadeh Zhad, O. Sadeghi and V. Amani, Food Chem., 2013, 141, 4040–4045.
40. L. B. Escudero, R. A. Olsina and R. G. Wuilloud, Talanta, 2013, 116, 133–140.
41. M. K. Rofouei, A. Sabouri, A. Ahmadalinezhad and H. Ferdowsi, J. Hazard. Mater., 2011, 192, 1358–1363.
42. M. M. Santoyo, J. A. L. Figueroa, K. Wrobel and K. Wrobel, Talanta, 2009, 79, 706–711.
43. O. P. Varghese, D. A. Ossipov and J. Hilborn, Polym. Prepr., 2009, 50, 169–170.
44. ICH Harmonized tripartite guideline, Validation of analytical procedures: text and methodology, 2005, Q2(R1), p. 9.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra13941k

‡ The first two authors contributed equally to this work.

---