Establishing adequate conditions for mercury determination in environmental samples by INAA

Mercury (Hg) is a toxic element released into the environment mainly by anthropic activities. Consequently, the improvement for Hg determination in environmental samples is of great interest. Instrumental Neutron Activation Analysis (INAA) is considered an adequate method to determine several elements. However, Hg determination by INAA is often hampered by its volatility, which causes losses, depending on the local temperature in the reactor where the sample is irradiated. The aim of this study was to establish adequate irradiation conditions in the IEA-R1 reactor for Hg determination in environmental samples by INAA. The following parameters were evaluated: irradiation time, container for irradiation and spectral gamma ray interferences. For the study, aliquots of certified reference materials (CRMs) and tree bark samples were irradiated together with Hg synthetic standard at the IEA-R1 nuclear research reactor. Gamma ray activities of 197 Hg and 203 Hg were measured in a spectrometer coupled to a HPGe detector. Obtained results indicated that polyethylene capsules or envelopes can be used as container for sample irradiation and the Hg impurities in these containers were negligible. Irradiation time of one hour was adequate for Hg determination and in long irradiations of 8 h problems of spectral interference of 198 Au and 75 Se were observed. In addition, Hg loss during the irradiation of 1 h and after irradiation was not observed. Quality control of Hg results, obtained in the CRMs analyses using one hour of irradiation, indicated good precision and accuracy with HORRAT < 2 and |Z score| < 2. The experimental conditions established in this study were applied to tree bark samples. Detection limits in these analyses were between 0.14 and 1.9 μg g -1 .


INTRODUCTION
Mercury (Hg) is a toxic element that causes adverse effects on human health and on the environment, arousing much concern everywhere due to its volatility, persistence and bioaccumulation [1].
Its toxicity depends on its chemical form, concentration, route of exposure and the exposed individual's vulnerability [2]. Mercury is emitted into the atmosphere by various anthropic activities, such as coal-fired power plants, mining activities, waste incineration and industrial processes used in foundries and cement production [3]. Consequently, it is of great interest to improve analytical methods for the quantification of this element in environmental samples.
Instrumental neutron activation analysis (INAA) is considered an advantageous technique for element determination of environmental contaminants, since it is a non-destructive technique, that requires minimal manipulation of the sample and provides low detection limits [4,5]. However, some elements are not determined with good accuracy by this technique, as in the case of Hg.
In the analysis of Hg by INAA, two radioisotopes that can be measured for its quantification are formed: 197 Hg (77.34 keV with half-life of 64.16 h) and 203 Hg (279.20 keV with half-life of 46.61 d) [6]. However, the main difficulty in the determination of Hg by INAA is the loss of this element during irradiation due to a combination of electrolytic action creating elemental Hg, recoil due to the (n,γ) reaction, and thermal heating [7]. The use of quartz ampoules as irradiation container would avoid the Hg loss [8], but it is not adequate due to the difficulty in transferring the irradiated sample from the ampoule to a counting container. In addition, the ampoule becomes quite radioactive due to the activation of the sodium present in it.
The aim of this study was to establish adequate conditions for Hg determination in environmental samples by INAA in IEA-R1 reactor. In order to obtain reliable results, the following experimental parameters were evaluated: irradiation time, type of container for sample irradiation and the problem of the spectral interferences. For this study, environmental certified reference materials and tree bark samples, used as environmental biomonitor, were analyzed.

Environmental samples
Tree bark samples of the species Poincianella pluviosa (Sibipiruna) and Tipuana tipu (Tipuana) were collected in several sites of São Paulo Metropolitan Region (SP, Brazil). The collection was carried out at a height from 1.5 to 2.5 m from topsoil using a stainless steel knife. When the bark was wet, it was previously dried at 40 ºC using an oven with forced air circulation. For the analyses, the samples were carefully cleaned using a dental brush with soft nylon bristles in order to remove the foreign materials and then the surface layer of the tree bark, with a thickness less than 3 mm, was grated using a titanium grater. Finally, the samples were ground to a fine powder using a vibratory micro mill with an agate mortar (Vibratory Micro Mill Pulverisette 0, FRITSCH), and then placed in plastic vials that were stored in a desiccator.
Percentages of residual humidity of these powder bark samples, determined as described for CRMs, varied from 4.9 to 12.5 %.

Preparation of synthetic standard of mercury
Certified standard solution of Hg acquired from Spex CertiPrep USA was used to prepare synthetic standard. From the stock standard solution, a diluted solution with Hg concentration of 136.5 µg mL -1 was prepared. To avoid Hg loss in the synthetic standard, thioacetamide solution was also used. Thioacetamide minimizes Hg losses by volatilization, since the sulfur present in its structure reacts with Hg, leading to the formation of HgS [9,10]. A 7.8 mg mL -1 thioacetamide solution was prepared by dissolving 195.7 mg of thioacetamide p.a. from Merck with purified water, and diluting to 25 mL.
Aliquots of 50 μL of thioacetamide solution were pipetted on each small sheet of nº. 40 Whatman filter paper with the size 1.5 cm x 6.0 cm. The sheets were placed in a desiccator for drying the aliquots at room temperature. Then, aliquots 50 μL of diluted Hg solution were pipetted on each sheet, which remained again in a desiccator for drying at room temperature. The sheets were folded and placed in plastic envelopes. The plastic (polyethylene) foils used to prepare these envelopes were previously cleaned using diluted nitric acid solution and purified water, and finally drying at room temperature. The micropipette used was previously checked in relation to its calibration.
To prepare Hg standard in the type W polyethylene capsule (Vrije Univesiteit, Amsterdam, The Nertherlands), a small sheet of Nº. 40 Whatman filter paper was cut and placed inside the capsule.
Then, 50 μL of thioacetamide solution were pipetted on the small sheet placed inside the capsule, which remained in a desiccator for drying at room temperature. After that, 50 μL of Hg diluted standard solution were pippeted on this sheet, which remained again in a desiccator for drying at room temperature. The Hg synthetic standards were kept in a refrigerator until its use in the analyses.

Irradiation, measurement and calculations
Aliquots of each tree bark sample or certified reference material were weighted (150 to 200 mg) in polyethylene envelopes. Each envelope containing the sample or standard was wrapped with aluminum foil. This set of samples and standard was placed in an aluminum irradiation device called "rabbit". Irradiation was performed at the IEA-R1 nuclear research reactor for 1 or 8 h, under a thermal neutron flux varying from 4.2 x 10 12 n cm 2 s -1 to 4.6 x 10 12 n cm 2 s -1 .
In the case of irradiations using polyethylene capsules, aliquots (150-200 mg) of each material were weighed in the capsules, occupying about 2/3 of the total volume of the capsule. Each capsule con-taining sample or standard was wrapped with aluminum foil and then the set of samples and standard was placed into the "rabbit".
After adequate decay time of 1 d, for samples irradiated for 1 h, and of 3 d, for samples irradiated for 8 h, samples and standard were placed individually in stainless steel planchets for gamma ray activity measurements. In the case of irradiations using polyethylene capsules, samples and standard were placed individually in plastic counting tubes.
The induced gamma ray activity measurements were carried out using a hyperpure germanium detector from Canberra (model GC3020) connected to a digital spectrum analyzer (model DSA 1000).

Treatment of data
To evaluate the mass fraction results of Hg in the CRMs in relation to their accuracy and precision, the arithmetic mean of the results, standard deviation, relative standard deviation, HORRAT value, relative error and standardized difference (Z score) were calculated. In addition, detection limits and quantification limits of Hg were obtained for the analyzed environmental samples of tree barks.

Relative standard deviation and HORRAT value
The precision of the results was evaluated according to Wood [12], who considers that this parameter is related to the level of concentration (or mass fraction) of the analyte in the sample and it is defined by the Horwitz equation [13] given by: Where RSD H is the value of Horwitz and c is the analyte's mass fraction. The precision of the method was verified by the HORRAT value given by [12]: (2) Where RSD is the relative standard deviation obtained experimentally.

Standardized difference or Z score value
To evaluate the accuracy of the CRMs results, Z score values were calculated using Equation 2, where X lab is the mass fraction obtained in the laboratory, X ref , the mass fraction of the certificate and SD mod , the modified standard deviation, obtained by Equation 3 [14].
Where SD is the standard deviation of the measurements performed in the laboratory and u Xref is the combined uncertainty of the reference value of the certificate. The result is classified as satisfactory or unsatisfactory according to the following criteria [14]: If | Z score | ≤ 2, the result is considered satisfactory; If 2 <| Z score | <3, the result is considered uncertain; If | Z score | ≥ 3, the result is considered unsatisfactory.

Detection limits and quantification limits
Detection limits and quantification limits were calculated using Currie's criterion [15], which defines the detection limit (DLT) and the quantification limit (QLT) as: Where DLT and QLT are, respectively, detection limit and quantification limit in counting rates, BG is the background and LT is the counting time. Once the DLT and QLT values were obtained, the detection limits (DL) and quantification limits (QL) in element mass fraction were obtained by the comparative method.

Analysis of Hg impurities in the polyethylene envelopes and in the Nº. 40 Whatman filter paper sheets
To evaluate the presence of Hg in the polyethylene envelopes, used for sample irradiation, empty envelopes were analyzed using the same experimental conditions used in the sample analyses.
The same procedure used for determination of Hg in the polyethylene envelopes was applied for the analysis of the Nº. 40 Whatman filter paper sheets, used as support in the preparation of synthetic standard. For this study, filter paper samples were cut in the same size as the sheets used for preparing the synthetic standard.

Irradiation using different polyethylene containers for Hg determination
To Where A and A 0 are the counting rates for decay times t and t = 0, respectively and t 1/2 is the halflife.

Selection of irradiation time
Mercury losses during the irradiation were evaluated by comparing the Hg mass fractions obtained and the certified value.
Aliquots of the CRM IAEA-085 Human Hair and Hg synthetic standard placed in polyethylene envelopes were irradiated for 1 and 8 h and the gamma ray activities were measured for different decay times. In order to evaluate the Hg loss after irradiation, the counting rates of Hg standard obtained for different decay time were corrected for zero decay time. The irradiated Hg standard was kept in the counting room at temperature of 20 ºC throughout the period in which the counts were performed.

Sheets
Mercury impurities in the polyethylene envelopes used in the irradiations were not detected, and detection limits of 0.007 and 0.10 μg per envelope were obtained by measuring 197 Hg and 203 Hg, respectively.
The determination of Hg impurity in the Nº. 40 Whatman filter paper sheet (used in the preparation of synthetic standard) indicated that this element is present at very low mass fractions. In the anal ysis of a filter paper sheet, measuring 1.5 cm x 6.0 cm and weighting 0.12030 g, Hg mass fraction of 0.0702 ± 0.0056 μg g -1 or a quantity of 0.0084 μg per sheet was obtained. Since the Hg mass in the synthetic standard is 6.8 μg, the presence of this element in the filter paper sheet was considered negligible.

Irradiation Using Different Kinds of Polyethylene Containers for Hg Determination
This study was carried out in order to choose a most adequate container for sample irradiation, avoiding Hg loss by volatilization. The Hg loss in the envelope and in the capsule during 1 h of irradiation was verified by the analysis of aliquots of the CRM IAEA-085 Human Hair, with certified value of Hg mass fraction of 23.2 ± 0.8 µg g -1 [17]. By measuring 197 Hg and 203 Hg for different decay times, results of Hg mass fraction were obtained, and they varied from 23.23 ± 0.11 to 26.75 ± 0.40 µg g -1 , for sample irradiated in polyethylene capsule, and from 23.58 ± 0.14 to 25.536 ± 0.087 µg g -1 , when polyethylene envelope was used. These results indicate that there were no Hg losses during the irradiations from both containers, since the obtained results for mass fraction are very close to the certified value.
The Hg loss after irradiation was also verified by measuring synthetic standard of Hg for different decay times. The counting rates obtained were corrected for the same decay time (t = 0) from the measurements obtained in different decay times. Tables 1 and 2    Based on these results, polyethylene envelopes were select for sample and standard irradiations for Hg determination due to its ease of acquisition and cheaper cost.

Mercury Determination Using Different Irradiation Times
The adequate irradiation time for Hg determination without causing its loss was chosen by analyz-  [6]. Table 3 shows the results obtained in the study of Hg loss after irradiation for Hg standard irradiat-   Tables 4 to 7 show the results obtained in the analyses of certified reference materials NCR DOLT- Mercury was not detected in the tree bark samples analyzed in this study, so detection limits and quantification limits for Hg determinations were calculated by the measurement of 197 Hg and 203 Hg radionuclides ( Table 8). The bark samples analyzed were probably not collected in regions polluted by Hg emission into the atmosphere. According to Martín et al [21], this element is naturally present in the environment at very low concentrations (0.005 to 0.06 ng m -3 in air). Results showed in Table 8 indicate that detection limits obtained for 197 Hg gamma ray peak are lower than those obtained for 203 Hg. -0.74 -0.87 1.1 a. X ± SD: Arithmetic mean and standard deviation; n: number of determinations; RSD: relative standard deviation; RE: relative error; b. dt = decay time.  a. X ± SD: Arithmetic mean and standard deviation; n: number of determinations; RSD: relative standard deviation; RE: relative error. b. dt = decay time.

CONCLUSION
From the results obtained in the analyses of polyethylene envelope and of the Nº. 40 Whatman filter paper sheet, it can be concluded that Hg is present in these materials at very low levels and it was considered negligible in the analyses.
Results obtained using different containers for irradiation demonstrated that polyethylene capsules and polyethylene envelopes can be used for Hg determination by INAA. Both containers showed no Hg loss. In this study, polyethylene envelope was chosen due to its ease of acquisition and cheaper cost.
Irradiation time of 1 h was chosen due to the problem of gamma ray spectral interference of 198 Au and 75 Se radionuclides. In addition, high activities of 24 Na, formed by the longer irradiation time, may affect the results of Hg determinations. The high activity of 24 Na masked the less intense activities of 197 Hg. The decay time of 6 d, in general, presented more precise and accurate results.
The results obtained for certified reference materials indicated good agreement with the certificate values, demonstrating the accuracy of the data. Results obtained also presented good precision and accuracy, evaluated by the HORRAT values (HORRAT < 2), and by Z score obtained values (|Z score| < 2). The exception was for Hg determination in the analysis of CRM BCR 186 Pig Kidney using decay time of 5 d.
Mercury was not detected in tree bark samples. The detection limits obtained for tree bark analysis varied from 0.07 and 1.9 μg g -1 , and the quantification limits were between 0.22 and 5.8 μg g -1 . Results obtained for Hg determination in tree barks indicated that there is no emission source of this element in the region where they were collected.