Robust ultrasound assisted extraction approach using dilute TMAH solutions for the speciation of mercury in ﬁ sh and plant materials by cold vapour atomic absorption spectrometry (CVAAS)

A simple and rapid ultrasound assisted extraction (UAE) protocol with dilute solutions of tetramethylammonium hydroxide (TMAH) for the speciation of mercury in ﬁ sh and plant tissues was developed as an alternative to conventional methods which require intensive treatments. The main operational parameters, such as extractant concentration (TMAH), sonication time and amount of sample, were optimized using BCR ERM-CE 464 (tuna ﬁ sh) and mercury loaded coriander powder, an in-house reference material, taken as representatives of ﬁ sh and plant tissues, respectively. Quantitative extractions of the inorganic mercury (iHg) and methyl mercury (MeHg) species were obtained using 8 mL of 2% TMAH with a sonication time of 5 min for a sample weight of <0.5 g. After sonication, the supernatant obtained upon centrifugation was directly used for the determination of iHg by cold vapour atomic absorption spectrometry (CVAAS). Inorganic mercury was determined using SnCl 2 as a reducing agent, while total mercury was determined after the oxidation of methyl mercury (MeHg) with KMnO 4 solution. Amount of organic mercury, basically MeHg, was obtained by using the di ﬀ erence. The analytical results were in good agreement with the certi ﬁ ed reference values of iHg, MeHg and total mercury at a 95% con ﬁ dence level. The method was further validated through the analysis of additional certi ﬁ ed reference materials: BCR CE-463 (tuna ﬁ sh), IAEA-350 ( ﬁ sh homogenate), BCR-60 (aquatic plant Lagarosiphon major) and BCR-482 (lichen). The detection limit of the overall procedure was found to be 0.014 m g g (cid:1) 1 for both inorganic and methyl mercury species.


Introduction
Mercury is a global pollutant and highly toxic among heavy metals because of its persistence, long range transport potential and bioaccumulation in the environment. Mercury is introduced in the environment mainly as elemental mercury (Hg 0 ), inorganic mercury (iHg) and organic mercury species as a result of both natural and anthropogenic activities from where it reenters the human food chain. [1][2][3][4][5][6][7] More than 2500 tons of mercury is emitted annually from global anthropogenic sources, which are signicantly contributing to elevated levels of mercury. It has been known that organomercury compounds, particularly methyl mercury (MeHg), are 50-100 times more toxic than inorganic mercury species. 8 These two are the common and predominant forms of mercury generally found in biological and environmental samples such as sh tissues and plant matrices. [9][10][11][12][13] Because of the accumulative properties and adverse toxic effects of mercury species even at ultra-trace levels, its accurate determination in sh and plant samples is very important for environmental protection and food safety.
As a consequence, considerable efforts and progress have been made in the development of sensitive and accurate sample preparation methods for the determination of total mercury and its speciation analysis in environmental and biological samples. [14][15][16][17][18][19][20][21][22][23][24][25][26] The most frequently used approaches for the extraction of mercury species from sh and plant samples are based on microwave [27][28][29][30] or ultrasound 31,32 assisted alkaline or acid leaching and solid phase extraction. 33,34 Despite excellent sensitivity and selectivity, most of the abovementioned approaches suffer from major limitations that include laboriousness of the procedures, use of high amount of acids along with complexing agents, lack of acceptable efficiency and long time consumption.
TMAH and formic acid reagents have been extensively used as the most appropriate tissue solubilizers for various biological samples prior to the analysis of various elements including mercury and its speciation. [35][36][37][38][39][40] Among these two solubilizers, the alkaline solubilization with TMAH offers a simple and rapid approach for the preparation of a homogenized sample solution, which is a distinct advantage over conventional slurry preparation methods. Therefore, several methods for the determination of iHg and MeHg species using TMAH have been developed and reported in the literature. [35][36][37][38][41][42][43] However, sample solutions produced aer solubilisation with TMAH are cloudy and emit an unpleasant odour as well that requires adequate ventilation. The use of dilute TMAH solutions can minimize the odour, but quantitative extraction of the species of interest may be affected.
Nowadays, there have been signicant developments in green analytical methodologies aimed to reduce the amount of toxic chemical reagents as well as simplify and accelerate experimental procedures. [44][45][46] In this context, the ultrasound assisted extraction (UAE) approach can be an excellent alternative to minimize the abovementioned limitations of conventional extraction procedures. 47,48 Being a clean technology, ultrasound energy has already been well exploited for a number of analytical applications such as speeding up solid-liquid extraction of the elements/species of interest for the determination of total-element contents and speciation analysis, remediation, organic synthesis and a number of other analytical and industrial applications. [49][50][51][52][53] Based on these facts, ultrasound assisted extraction protocol was utilized in the present work for the speciation of mercury in sh and plant materials using dilute TMAH solutions.
The most commonly and widely used techniques employed for the determination of mercury species in a great variety of matrices, including sh and plant tissues with and without applying chromatographic separation, are cold vapour atomic absorption spectrometry (CVAAS) 54 and atomic uorescence spectrometry (CVAFS). 55 In the present study, CVAAS was selected for Hg determination because of its high sensitivity, absence of spectral interferences, relatively low operational costs and simplicity as well as rapidity.
The main objective of the work is to develop a simple, efficient and green analytical methodology for the determination of t-Hg, iHg and indirectly MeHg without the use of a chromatographic separation aer treatment with dilute TMAH solutions with the aid of ultrasound probe energy, which is suitable for both sh and plant tissues. Mercury loaded coriander powder (representative of samples of plant origin) and BCR CRM 464 (tuna sh; representative of sh) were used for optimization experiments. Aer extraction using optimized conditions, the concentration of iHg and tHg were determined using CVAAS aer employing KMnO 4 treatment for the oxidation of organic mercury species to inorganic mercury. A closed microwave digestion procedure based on the use of dilute nitric acid solutions and H 2 O 2 was utilized for the dissolution of the test samples for the subsequent determination of total mercury by CVAAS.

Instrumentation
High intensity probe sonicator for ultra-sound assisted extraction. Extractions were performed using a 750 W power and 20 kHz frequency high intensity probe sonicator equipped with a 6 mm Ti probe (Sonic Vibra Cell, Sonics and Materials Inc., CT, USA, Model: VCX 750). According to the manufacturer's recommendation, the amplitude of the ultrasonic processor for the ultrasonic vibrations at the probe was set at the maximum allowable limit of 40%. Pre-cleaned polypropylene centrifuge tubes of 50 mL capacity (Tarson) were used as vessels for sonication experiments. Aer sonication, all the extracts were centrifuged at 8000 rpm (REMI Instruments Pvt. Ltd, Mumbai, India) for about 5 min for the rapid separation of the solid-liquid mixture.
Microwave digestion system for total decomposition of samples. A microwave digestion system (CEM Mars 5, Matthews, NC, USA) was used for the mineralization of the test samples for the determination of total mercury. The sample carousel was capable of holding 10 PTFE digestion vessels (XP-1500 Plus) with a capacity of 100 mL each, which also includes a control vessel tted with a ber optic temperature sensor and a pressure transducer for controlling the microwave program, and it is capable of withstanding a pressure of 500 psi and temperatures up to 260 C.
Determination of mercury and its species. Mercury was determined by cold vapour atomic absorption spectrometry (CVAAS) using a mercury analyzer (Model MA 5840E, Electronics Corporation of India Ltd, Hyderabad, India). The information of organic and inorganic forms of mercury could also be obtained with the same instrumentation by changing reducing agents with different reducing powers. SnCl 2 is known to reduce only Hg 2+ to Hg 0 , whereas NaBH 4 is capable of reducing both iHg and MeHg to elemental mercury, albeit with different sensitivities.

Reagents and materials
All the chemicals used in this work were at least of AR grade. High-purity water with a resistivity of >18 MU cm, used for the preparation of standards and samples and for the cleaning of vessels, was produced using a Milli-Q high purity water system, located in class 100 area of the ultra-trace analysis laboratory of this centre. Dilute solutions of TMAH, prepared from stock solution (25% in methanol, Aldrich, USA), was used as the extractant. Tin(II) chloride (SnCl 2 ) (5%, w/v) used as the reducing agent was prepared by dissolving an appropriate amount of SnCl 2 $2H 2 O (Merck, India) in HCl and diluting with water. Sodium borohydride (NaBH 4 ) (Merck, Darmstadt, Germany) (1%, w/v) was prepared fresh daily by dissolving an appropriate amount of solid in 0.3% (w/v) NaOH solution. A carrier solution of 10% HCl was used along with SnCl 2 or NaBH 4 for the reduction of mercury. Inorganic mercury standard solution (1000 mg L À1 ) in 5% HNO 3 (SD Fine-Chem Ltd, Mumbai, India) traceable to NIST 3133 was used as a stock standard. A methyl mercury (CH 3 Hg + ) stock standard solution (100 mg L À1 , Hg as MeHg) was prepared from methyl mercury iodide (Aldrich) by dissolving an appropriate amount of the solid in methanol and making up to the required volume with high purity water. All the stock standard solutions were stored in a refrigerator at 4 C and protected from light. Working standard solutions were prepared just before use by the appropriate dilution of the stock standard solutions.
The following certied reference materials (CRMs) were analysed to evaluate the developed method; Lichen-482 from Community Bureau of Reference (BCR), BCR-60 (Lagarosiphon major, aquatic plant), European reference Materials (ERM) CE-463 and 464 (tuna sh) and sh homogenate IAEA-350. All the solid reference materials were used as received, without further grinding and sieving.
Preparation of mixture of iHg and MeHg loaded coriander material (laboratory reference material). In most of the certied reference materials (CRMs) either inorganic or methyl mercury is found to be at much higher concentrations relative to the other species. In particular, CRMs of plant origin containing high levels of Hg are scarce. To the best of our knowledge, no reference material is available, which is certied for higher contents (ppm) of both iHg and MeHg for the validation of methods for plant and sh samples. Another issue is the large quantity of reference material required for optimization experiments. In view of this, a coriander sample loaded with a known content of mercury (iHg/MeHg separately) and a mixture of iHg and MeHg at high ppm level was prepared in the laboratory for use in the optimization experiments related to the samples of plant origin. In the present work, we have chosen coriander material (Coriandrum sativum) (common edible plant in every household), for the preparation of in-house reference material because of its availability, ease of preparation, high uptake capacity for mercury species and cost effectiveness. In our earlier studies, the sorption capacities for iHg and MeHg were determined to be $24 mg g À1 and $7 mg g À1 , respectively. 56 A large quantity of coriander plants was obtained from the local market and thoroughly washed with water to remove all the adhering soil particles. The whole plant (roots, stem and leaves) was cut into small pieces and dried at 50 C in a conventional heating oven, ground in a planetary ball mill (Fritz, Germany) and sieved to get a particle size of #100 mm. Aer this step, about 10 g of the powdered coriander was placed in a glass beaker containing 200 mL of high purity water spiked with 100 mg of iHg and MeHg individually (designated as Cori-iHg and Cori-MeHg, respectively) such that the amount of mercury species in the coriander compounded to about 10 mg g À1 Hg in the solution.
The mixture was continuously stirred for about 1 h for quantitative sorption and to facilitate uniform loading of spiked mercury. Aer shaking, the mixture was separated by centrifugation (8000 rpm for 5 min) and the supernatant was drained. Then, the sorbent was initially allowed to dry at room temperature and then dried in a conventional heating oven at $40 C to remove the residual moisture. Then, the dried sample was nely ground and sieved to obtain 200-400 mesh size particles. In another set of experiments, both iHg and MeHg (10 mg each) were loaded on coriander powder (weight of coriander 10 g) using a procedure similar to the one described above such that the total amount of mercury in coriander was about 20 mg g À1 (Cori-iHg-MeHg). In all the cases, the supernatant was analysed for the determination of residual mercury by CVAAS, and the results indicated the absence of mercury.
Microwave-assisted digestion procedure using diluted acids for the determination of total mercury. It is still usual to digest the samples by adding large amounts of concentrated mineral acids, which leads to the generation of large volumes of toxic wastes. At present, considering the excessive use of concentrated acids, environment-friendly strategies are being implemented without impairing analytical performance aiming toward greener sample preparation methods. 57,58 In this context, dilute solutions of HNO 3 in the presence of the auxiliary reagent H 2 O 2 have been successfully developed and used in the complete digestion of bio-environmental samples for the determination of total mercury. The efficacy of the proposed extraction procedure was evaluated aer the decomposition of the test materials with the closed microwave-assisted acid digestion procedure as described below.
For total Hg determination, an accurately weighed aliquot ($200 mg) of the target materials was placed in the PTFE microwave digestion vessel to which 1 mL of concentrated HNO 3 and 1 mL of H 2 O 2 followed by 2 mL of high purity water were added. Aer closing, the vessels were clamped within a support module and placed inside a microwave-assisted digestion system. A microwave program consisting of the following steps was used: (i) temperature was ramped to 100 AE 2 C in 5 min (pre-digestion step); (ii) temperature was ramped to 200 AE 5 C in 10 min and held there for 10 min and, (iii) zero power for 20 min (cooling step). Aer cooling, the resultant clear sample digests were quantitatively transferred from the PTFE vessel to another pre-cleaned tube and diluted to the desirable volume with water depending on the concentration of mercury. Aer suitable dilution, all the sample solutions were analysed by CVAAS aer VG of mercury using SnCl 2 and/or NaBH 4 for the determination of the total mercury present in each CRM. The corresponding process blank solutions were also subjected to the same procedure in the absence of the sample.
Ultrasound-assisted extraction procedure. For the extraction of iHg and MeHg species by ultrasound energy, accurately weighed aliquots ($200 mg) of the selected CRMs were placed in polypropylene (PP) centrifuge tubes (50 mL volume) and 8 mL of the desired extractant (2% v/v TMAH) solution was added. Then, the sample-extractant mixture was sonicated for the chosen sonication time and amplitude settings. Aer sonication, the supernatant was separated from the solid phase by centrifugation for about 5 minutes at 8000 rpm. The known volume of the supernatant was then transferred to another pre-cleaned PP tube. The resultant solutions aer suitable dilution were analysed for iHg and tHg by CVAAS as described below.
A known amount of iHg standard/sample solution was added into a reaction vessel (of CVAAS system) containing $5 mL of 10% HCl carrier solution. The reaction mixture was vigorously stirred for the desired length of time (1-3 min) in a closed environment before passing Hg 0 vapors to the quartz cell of the AAS system for quantication. One part of the split samples was analysed for the determination of iHg using SnCl 2 as the selective reducing agent. To determine total Hg, it was necessary to add an appropriate amount of KMnO 4 solution to the other part of the split sample for oxidation treatment in the presence of 5% HNO 3 to convert MeHg to iHg, which was followed by its determination by CVAAS using SnCl 2 or NaBH 4 as the reducing agent. The concentration of methyl mercury was calculated as the difference between the total and iHg values.
Corresponding process blanks (with and without oxidative treatment) were also prepared in the same way without taking any sample material. Three different aliquots of each sample were used for the extraction process. All the analytical measurements were run in triplicate for each sample solution. With each series of extractions, a blank was also prepared and measured in parallel to determine the cross-contamination of mercury. Quantications of the mercury species in the test samples are based on a 5-point calibration graph obtained with the standards of mercury in the concentration range of 0 (analytical blank)-100 ng mL À1 prepared using process blank solutions containing 2% TMAH and TMAH-extracted blank coriander sample solutions. These calibration plots were compared with those of pure aqueous standards of mercury to test the matrix effects, if any. The standard addition method was also applied to determine other possible interferences, if any.
To test the volatility of the mercury species under ultrasound-assisted extraction conditions, a set of experiments was carried out, in which standards containing known amounts of mercury species (iHg and MeHg) prepared in 8 mL of the optimized extractant solution (2% TMAH) were subjected to the proposed ultrasound-assisted extraction procedure as in the case of the samples. The resultant solutions (aer suitable dilution) were analysed for the determination of the species of mercury by CVAAS. Calibration plots were also obtained with these processed standard solutions and compared with the plots obtained for pure aqueous mercury standards.
Aer applying the ultrasound-assisted extraction process, the extraction efficiency at each step was tested by calculating the percentage recovery of the test mercury species in the samples using the following equation:

Results and discussion
The main concerns in the quantitative extraction of the mercury species from solid matrices (in this case sh and plant tissues) should be the efficiency, volatility, inter-species conversion, contamination and amount of reagents required. Extraction methods based on the use of ultrasound energy usually do not require intensive conditions such as high temperatures, pressures or concentrated acids. Based on this fact, the present work was initiated using dilute solutions of TMAH with the aid of ultrasound energy for the speciation of mercury. As mentioned in earlier sections, inorganic and methyl mercury (MeHg) species are the two common and predominant forms generally found in various biological and environmental samples. Therefore, the present study was focussed on the determination of only inorganic and methyl mercury species.
Initially, a series of experiments were carried out to optimize these variables for quantitative recovery of both iHg and MeHg. Mercury loaded coriander powder (representative of samples of plant origin) and BCR CRM 464 (tuna sh; representative of sh tissue) were used for the optimization experiments. In the case of the sh representative sample, the concentration of iHg was very low (represented only 2.3% of the total Hg concentration); thus, the level of iHg was increased using standard addition to evaluate the stability of both iHg and MeHg species during the USE process. Accordingly, $0.2 g of ERM-CE464 was spiked with 100 mL of iHg standard (from 10 mL mL À1 stock standard), to which extraction solvent TMAH was added. Aer each extraction step, the percentage recovery of both iHg and MeHg was determined during the method development.

Total mercury determination
Different volumes of HNO 3 and H 2 O 2 , irradiation times and microwave power settings of CEM microwave system were tested to ensure the total recovery of Hg. In each case, $200 mg of the solid sample was taken and digested using the microwave program described in the earlier section. The addition of a mixture of 1.5 mL HNO 3 , 1 mL of H 2 O 2 and 2.5 mL of water greatly improved the efficiency of digestion, providing a clear solution and quantitative recovery of mercury from the CRMs, selected in this work. The reduction of mercury was carried out using NaBH 4 with a concentration of 2% w/v for the subsequent determination of total mercury by CVAAS. The results obtained with the digestion performed with the proposed procedure were found to be in good agreement with the certied values (recoveries higher than 98%). The use of diluted HNO 3 in the presence of H 2 O 2 was proven to be a feasible and recommendable sample digestion procedure complying with the green chemistry recommendations.

Speciation analysis of mercury
The ultrasound-assisted extraction of mercury species may not be equally effective for all the solid samples; thus, maximizing the extraction yield requires the process variables to be optimized for each specic matrix (in this case, plant and sh matrices). The extraction efficiency of ultrasound energy is essentially governed by various parameters including extractant concentration (TMAH), sonication time and amount of sample. Therefore, these variables were individually optimized to achieve quantitative recovery of both the mercury species, while the others were kept constant.

Optimization of concentration of TMAH
As mentioned in the earlier sections, TMAH is strongly alkaline, soluble in aqueous media, stabilizes volatile elements and does not require heating or only requires gentle heating, and it is thus promising for the speciation analysis of mercury. In the present work, dilute solutions of TMAH were used to test its efficacy as an extractant to achieve the quantitative extraction of the mercury species from plant and sh tissues with the aid of ultrasound energy. Based on the results obtained from various preliminary experiments, different concentrations of TMAH in the range of 0.5-3% were chosen for two representative materials, keeping the other parameters (sonication time: 5 min, volume of extractant: 8 mL and amount of sample: $200 mg) constant. An extractant volume of 8 mL was chosen in all the optimization experiments so that the required number of replicates could be performed without the exhaustion of the sample solution.
As a compromise between sensitivity and reagent consumption, 5% w/v SnCl 2 in 10% v/v HCl solution was chosen as the reducing agent for the determination of iHg, while 2% w/ v NaBH 4 and 5% v/v HNO 3 was chosen as the optimum condition for tHg determination in the nal TMAH-sample extracts aer oxidation treatment with KMnO 4 . Fig. 1a and b shows the effect of the concentration of TMAH on the extraction efficiency of the iHg and MeHg species from BCR CRM 464 and mercury loaded coriander representative materials. As shown in these gures, extraction efficiency (i.e., recovery of Hg species from the solid matrix) with water (in the absence of TMAH) was very low (<10%), while the efficiency of TMAH for the extraction of both iHg and MeHg increased with the concentration of TMAH up to 2% and reached a plateau in the concentration range of 2-5%, which is the highest studied concentration. As seen from Fig. 1a and b, the optimum concentration of TMAH was found to be about 1.5% for the quantitative extraction (>95%) of the two selected mercury species from BCR-464, while 2% of TMAH was required for the mercury loaded coriander material (which had a plant origin). In general, sh tissues are so as compared to plant tissues, and thus sh tissue requires a lower concentration of TMAH for the complete extraction of the species of interest.
Both iHg and MeHg species show similar extraction behaviour with quantitative recoveries between 95% and 102% when dilute solutions of TMAH were used as the extractant. Aer sonication, the colour of the nal extractant solution resembled the original colour of the powdered sample. The effect of ultrasound energy on the stability of the Hg species was also studied using the two representative materials by analyzing TMAH-extracted solutions at different time intervals. These studies clearly indicate that aer carrying out UAE with 2% TMAH, the two tested mercury species remained stable even aer standing for a week in the laboratory at room temperature. The TMAH concentration of 2% v/v was adopted for further extraction experiments to make it suitable to both sh and plant tissues.

Optimization of sonication time
The sonication time of the sample is an important parameter because the dose of ultrasound sonication received by the matrix and extractant mixture determines the extent of cavitation phenomena followed by the efficiency of extraction. The sonication time of 5 min or less is usually reported when ultrasonic probes are used for solid liquid extraction. Fixing ultrasound amplitude (40%), extractant concentration (2% v/v TMAH), extractant volume (8 mL) and sample weight ($200 mg), we investigated the inuence of sonication time on the extraction of the Hg species in the range of 1-6 min. In both the sh and the coriander representative samples, the extraction efficiency of the two Hg species increased from 45% to $98% as the sonication time increased from 1 to 4 min and remained almost constant in time interval of 5-7 min. The results obtained from these studies indicated that the sonication time of 4 min was found to be sufficient for the quantitative extraction of mercury species from both the representative materials, which is advantageous to obtain a high sample throughput. A sonication time of 5 min was thus selected as the optimum for further optimization studies because the species recovery was highly reproducible.

Evaluation of KMnO 4 concentration and reaction time for MeHg oxidation
First, the concentrations of KMnO 4 and HNO 3 were optimized for the quantitative conversion of MeHg to Hg 2+ followed by CVAAS determination. This oxidation treatment was performed before adding a reducing agent for VG of mercury. As mentioned above, iHg was determined using SnCl 2 as the selective reducing agent, whereas tHg was determined aer the oxidation of organic mercury to iHg through reaction with KMnO 4 followed by reduction to elemental mercury. A variety of oxidizing agents, viz., H 2 O 2 , KMnO 4 , K 2 Cr 2 O 7 and K 2 S 2 O 8 , in combination with strong acids (such as HCl and HNO 3 ), UV and microwave irradiation have been extensively used for the oxidation of organic mercury to iHg followed by the determination of tHg. In the present work, KMnO 4 was selected to decompose the organomercury species (predominantly MeHg in this case) due to its ease of preparation, stability and low mercury blank. KMnO 4 also promotes the efficient stabilization of mercury in solution until analysis. 41 Because the extraction of the mercury species was carried out using a 2% TMAH solution, it is necessary to add HNO 3 along with KMnO 4 to acidify the sample digest for the rapid oxidation of the organomercury species. Methyl mercury loaded coriander sample (Cori-MeHg) and tuna sh (BCR-CE 464) were taken as representatives for optimizing the concentration of HNO 3 and KMnO 4 required for the quantitative conversion of CH 3 Hg + to Hg 2+ . Aer a general speciation procedure, a sample volume of 0.5 mL was taken for the optimization studies. To optimize the composition of HNO 3 and KMnO 4 , a factorial (two factors, three levels) experimental design approach was applied and the conversion efficiency of MeHg at each level of treatment was estimated. Based on the results obtained from various preliminary experiments, a mixture of 4.5 mL of 0.02% w/v KMnO 4 and 5% v/v HNO 3 (added to a reaction vessel of CVAAS containing 0.5 mL of the TMAH-extracted sample) was selected as the base level for the two representative materials (the upper and lower levels were obtained using a difference of AE0.01% for KMnO 4 and AE2.5% for HNO 3 ). The mixture was stirred for about 1 min, and then the reducing agent was added for the determination of mercury by CVAAS. At each optimization step, the corresponding solutions were employed as blanks.
From Fig. 2, it can be seen that the conversion efficiency of MeHg varied signicantly with different concentrations of KMnO 4 and HNO 3 added to the TMAH-extracted coriander sample solution. From these studies, it was observed that the best efficiency of conversion was obtained with a mixture of 4.5 mL of 0.02% w/v KMnO 4 and 5% v/v HNO 3 for 0.5 mL of the sample solution. This is believed to be a result of the efficient conversion of MeHg to Hg 2+ in the standard and samples as well as due to the stabilization of mercury in the standard/sample solution in its oxidized form. Similar results were obtained for the sh representative sample, and thus the data obtained for sh sample are not shown here.   In the case of the MeHg standard, the addition of a mixture of 4.5 mL of 0.01% w/v KMnO 4 and 5% v/v HNO 3 allowed quantitative conversion to iHg, whereas the conversion efficiency was only 70-80% for TMAH-extracted sample solutions. Thus, it was inferred that more oxidizing agent is required for test samples in comparison with the MeHg standard solution because of the presence of other sample components, which competed with the MeHg species during the oxidation process. This may be mainly due to the consumption of a major part of KMnO 4 by the sample matrix, thus reducing the availability of the oxidizing agent for the oxidative conversion of CH 3 Hg + to Hg 2+ . Based on these results, a mixture of 4.5 mL of 0.02% w/v KMnO 4 and 5% v/v HNO 3 was added to the reaction vessel (of CVAAS) containing 0.5 mL of the sample solution prior to the reduction to elemental mercury. However, for treating a higher volume of TMAH-extracted sample solutions (>0.5 mL) (depending on the concentration of MeHg), an increased amount of KMnO 4 solution is required for the quantitative conversion.
Aer optimizing the concentration of the oxidizing agent KMnO 4 , it was necessary to optimize the reaction time (stirring time) required for a complete oxidation of the CH 3 Hg + to Hg 2+ in the tested samples. Based on a series of experiments, a reaction time of 1 min was chosen as the optimum because the recovery of mercury was quantitative and the mercury signal was highly reproducible. No signicant improvement in sensitivity could be obtained with longer reaction periods (1-3 min). Thus, a reaction time (i.e., stirring time) of 1 min was used in all the subsequent experiments.
Tao et al. 36 used L-cysteine and KMnO 4 for the determination of iHg and tHg, respectively. They added L-cysteine to sample solutions to liberate iHg from protein-bound mercury or other molecules in the TMAH-extracted solutions. In this work, however, the addition of L-cysteine did not enhance the iHg level, indicating that the reducing agent (SnCl 2 or NaBH 4 ) alone was found to be sufficient (without the need of L-cysteine) for the quantitative recovery of iHg in the sample solutions aer UAE using dilute TMAH ($2%) solutions.

Figures of merit
The whole analytical procedure proposed for the speciation of mercury in plant and sh tissues is schematically presented in Fig. 3. Calibration curves were obtained across the concentration range of 0 (analytical blank)-100 ng mL À1 for iHg and MeHg species prepared in different solvent media (aqueous, 2% TMAH and TMAH-extracted solutions of blank coriander powder). Analytical response characteristics of iHg and MeHg species spiked in different solvent media are presented in Table  1. In all the cases, the correlation coefficients were >0.995. The slopes of the calibration curves corresponding to Milli-Q water, 2% TMAH solutions and TMAH-extracted sample solutions spiked with iHg and MeHg did not differ signicantly, showing no matrix effect in the TMAH medium, and thus demonstrating the efficacy of the developed UAE procedure using dilute solutions of TMAH. This allows the use of the aqueous standard calibration curve for quantication purposes. As both the external and standard addition approaches provided comparable results, all mercury measurements were subsequently carried out using only the external calibration method. The analytical results of the mercury loaded coriander sample and various CRMs together with the certied/reference values are presented in Table 2 and 3, respectively. The determined values for total iHg obtained by both UAE and MAD digestion methods agree with the certied values (at 95% condence level). The organic mercury concentration, calculated as the difference between the total and iHg values, also agrees with the certied MeHg concentration. This demonstrates that most of the organic mercury obtained by the arithmetic difference is mainly MeHg. The detection limit of the method, determined as the concentration corresponding to three times the standard deviation of the blank, was 0.014 mg g À1 based on 0.4 g of the sample and 8 mL of the extractant solution. The precision, evaluated as the relative standard deviation (RSD%), was better than 10% in most of the cases.
The proposed analytical procedure markedly reduces the concentration of TMAH required for extraction by more than 10 times as compared to the reported solubilisation methods and the time needed for sample preparation (total 10 min including centrifugation time). In addition, keeping the number of analytical steps to a minimum considerably reduces the sources of analytical errors.

Conclusions
An effective analytical method based on the use of a dilute TMAH (2%) solution as the extractant with the aid of ultrasound energy for the speciation analysis of mercury by CVAAS in plant and sh tissues was developed. The developed extraction procedure and Hg-species determination was validated by the analysis of various certied reference materials. Aer ultrasound-assisted extraction, the TMAH-extracted sample solutions were directly analysed for iHg by CVAAS, while tHg was determined aer oxidation with a solution of KMnO 4 . This method also provides very important information on the toxic organomercury content, mainly MeHg (determined as the difference between tHg and iHg) in sh and plant tissues without handling a highly toxic methyl mercury standard. If a sample contains other organic species, such as phenyl mercury and dimethyl mercury, then the present method shall be suitable only for the identication of the inorganic and organic forms of mercury. The developed method can not only signicantly reduce the sample preparation time, but also provide quantitative recoveries (in the range of 95-102%) and preserve the integrity of the species. In addition, extra reagents (such as L-cysteine) and concentrated reagents (TMAH) are not required for the determination of iHg and total mercury. In the proposed UAE approach, the speciation analysis of mercury was achieved without using any chromatographic technique, requiring only an ultrasound probe and CVAAS instruments. The main features of the present UAE method are as follows: no matrix separation, reduction in time and solvent consumption, easy implementation, efficacy, reproducibility and safety of the procedure.