- Email: [email protected]
Contribution to vapor generation-inductively coupled plasma spectrometric techniques for determination of sulfide in water samples
Talanta 80 (2010) 1777–1781
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Contribution to vapor generation-inductively coupled plasma spectrometric techniques for determination of sulfide in water samples a ˇ Jiˇrí Cmelík , Jiˇrí Machát b,∗ , Vítˇezslav Otruba a , Viktor Kanicky´ a a b
Department of Chemistry, Faculty of Science, Masaryk University, Kotláˇrská 2, Brno, Czech Republic Research Centre for Environmental Chemistry and Ecotoxicology, Faculty of Science, Masaryk University, Kamenice 3, Brno, Czech Republic
a r t i c l e
i n f o
Article history: Received 9 July 2009 Received in revised form 23 September 2009 Accepted 8 October 2009 Available online 5 November 2009 Keywords: Hydrogen sulfide Sulfide Vapor generation ICP-OES Chemical interference
a b s t r a c t Vapor generation-inductively coupled plasma-optical emission spectrometry was used for the determination of sulfide in water samples preserved by the addition of a zinc acetate and sodium hydroxide solution. Hydrogen sulfide and acid-volatile sulfides were transformed, by acidification, to a gaseous phase in a vapor generator and subsequently detected by inductively coupled plasma optical emission spectrometry. Compounds interfering with iodometric titration and spectrophotometric determination were examined as potential chemical interferents. The proposed method provides results comparable to iodometric titration in the tested concentration range 0.06–22.0 mg L−1 . Limit of detection for the determination of hydrogen sulfide by this method is 0.03 mg L−1 . © 2009 Elsevier B.V. All rights reserved.
1. Introduction Sulfide concentration in a water sample is an important parameter for the assessment of spring, surface or waste water pollution [1]. There are several established methods for the determination of sulfide and/or hydrogen sulfide such as iodometric titration, spectrophotometric determination, potentiometry or ion chromatography [2–5]. Some drawbacks of these methods (insufficient limit of detection and their laboriousness) led to the development of their modification and to the establishment of completely new methods. The application of inductively coupled plasma (ICP) spectrometry hyphenated to a vapor generator (VG) for sulfide determination is one such innovative approach. Inductively coupled plasma-mass spectrometry hyphenated to vapor generation (VG-ICP-QMS) was used for the determination of sulfide in water samples in ref. [6]. The quadrupole analyzer of the mass spectrometer suffers from insufficient spectral resolution hence the measurement of sulfur is strongly interfered by polyatomic ions mainly from molecular ion O2 + . Oxygen forming these interfering ions arises particularly from water when nebulizing aqueous solutions. Generation of water-free gaseous hydrogen sulfide into the ICP results in the suppression of polyatomic interference. Further elimination of O2 + can be achieved by using a
∗ Corresponding author. Tel.: +420 549 495 684; fax: +420 549 492 840. E-mail address: [email protected] (J. Machát). 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.10.022
collision-reaction cell (CRC) in mixed H2 /He mode. In this work, optimization of the CRC parameters was primarily solved to obtain a satisfactory LOD. The LOD obtained for VG-ICP-MS determination of sulfide in water was 2 g L−1 . Inductively coupled plasma-optical emission spectrometry (ICP-OES) can also be used for sulfur detection. Measurement of the main sulfur analytical emission lines in vacuum UV range are not usually subject to spectral interference as the ICP emission spectrum in this region is relatively low in emission lines. In the case of ICP-OES, two techniques of sample introduction, both taking full advantage of vapor generation, can be used. A vapor generator with a gas/liquid phase separator enables introduction of the gaseous phase and thus only volatilized sulfide is detected [7]. The advantage of this technique is the elimination of spectral and non-spectral interferences in the ICP resulting from matrix separation. Nebulization of the reaction mixture (sample and acid) is another concept for the introduction of a volatile analyte allowing sensitive determination of sulfide as well as sulfate, latter being introduced into the ICP as an aerosol. Both these techniques were applied by [7,8]. Limits of detection for sulfide using VG and aerosol were 6 and 5 g L−1 , respectively, with an axially-viewed ICP. Hydrogen sulfide and sulfide concentration in water can be influenced by volatilization and/or oxidation; hence preservation of the samples is necessary in the sampling step. Preservation can be provided, for example, by adjustment of the pH to 8.5–9 with sodium hydroxide and ascorbic acid [9], or by zinc or cadmium acetate with subsequent alkalinization [10] and storage at a low
1778
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
temperature (1–5 ◦ C). Sodium hydroxide solution (2 mol L−1 ) and ascorbic acid (0.2 mol L−1 ) was used in ref. [6] for sample treatment by mixing the same volumes of both water sample and buffer. Separation of the gaseous analyte from liquid waste is achieved in a vapor-generator, thus high concentrations of sodium in a sample does not present a problem. In the case of nebulization of a reaction mixture [7,8] into the ICP, high Na concentrations in the solution lead to plasma instability. In this case an ammonium/ammonium chloride buffer was successfully used for alkalinization. Both ICP methods were successfully applied to spiked and natural water samples [6–8] and to the determination of acid volatile sulfide in marine sediment [6]. Results of these techniques were compared to those of a potentiometric method. A vapor generator with a phase separation – ICP-OES method was also applied in our work for the determination of sulfide in water samples. The addition of zinc acetate/sodium hydroxide was used for the sample preservation with subsequent dissolution of the precipitate by adding EDTA prior to the measurement. Considering the lack of study of chemical interference in the abovementioned works [6–8], we have dealt with the effect of some ions in the vapor generation of hydrogen sulfide. Our method was applied to real spring water samples with sulfide concentrations in the range 0.06–22 mg L−1 and the results were compared with the results from iodometric titration following distillation separation of hydrogen sulfide from the sample. 2. Experimental 2.1. Instrumentation An optical emission spectrometer with inductively coupled plasma Jobin-Yvon 170 Ultrace (JY-Horiba, Longjumeau, France) and hydride generator HG-1 (Labtech Ltd., Brno, Czech Republic) were used in this work. The schematic of the vapor generation system can be found in [11]. Operating conditions are shown in Table 1. Standard alumina injector (i.d. 2.5 mm) in the JY demountable plasma torch was replaced by a quartz injector because of the significant memory effects with the alumina injector. Similar effects were observed with iodine vapor generation as presented by Niedobova et al. [11]. 2.2. Chemicals Chemicals of reagent grade quality (Lach-Ner, Neratovice, Czech Republic) and deionized water were used throughout the experiments. Zinc acetate, sodium hydroxide, di-sodium EDTA (dihydrate) and hydrochloric acid (6 mol L−1 ) were used for the preservation of the samples and following the generation of hydrogen sulfide. Chemical interferences were examined using the following chemicals: sodium nitrite, nitrate, sulfate, sulfite, Table 1 Operation conditions of ICP spectrometer and vapor generator. ICP spectrometer Forward power ICP (W) Plasma gas flow (L min−1 ) Carrier gas flow (L min−1 ) Sheath gas flow (L min−1 ) Monochromator purging (Ar) (L min−1 ) Plasma viewing
1200 12.0 0.5 0.5 3.0 Radial
Vapor generator Carrier gas flow (L min−1 ) Sample flow rate (mL min−1 ) Acid flow rate (mL min−1 )
0.5 1.0 0.5
bicarbonate, copper (II) sulfate pentahydrate, iron (II) sulfate heptahydrate, and iron (III) chloride hexahydrate. Standard sulfide solutions were prepared by dissolving an appropriate amount of sodium sulfide (nonahydrate) in sodium hydroxide solution (0.1 mol L−1 ) and by subsequent dilution. These calibration solutions were standardized by iodometric titration because of the poorly defined content of sulfide in sodium sulfide and its instability. In the case of distillation separation of hydrogen sulfide from the real samples, glycerol was used in a trap solution of sodium hydroxide [2]. Iodine volumetric solution (0.01 mol L−1 ) was prepared by precisely weighing potassium iodate and the excess of potassium iodide after acidification by sulfuric acid. Sodium thiosulfate solution (0.01 mol L−1 , prepared from sodium thiosulfate pentahydrate) was standardized by titration of an iodine solution using starch as an indicator. 2.3. Samples The proposed method was applied to spring water samples with different concentrations of sulfide which were preserved immediately after sampling. The preservation was performed by adding of 1 mL 10% (m/m) zinc acetate and 0.5 mL 4 mol L−1 sodium hydroxide to 100 mL of water sample. For VG-ICP-OES determination powdered Na2 EDTA·2H2 O (about 200 mg) was added to each sample and mixed thoroughly to dissolve the precipitate in the sample prior to the analysis. The amount of Na2 EDTA·2H2 O depends on the amount of the precipitate formed by the preservation. The preservation was performed in the laboratory during the development of the method and during the investigation of chemical interferences. Real samples were preserved immediately after the sampling procedure. Real samples of spring waters (100 mL) were sampled from the inspection wells in the region around Hodonín city (Southern Moravia, Czech Republic). The inspection wells are located near petroleum and exploratory wells where a high concentration of sulfate and variable concentrations of hydrogen sulfide can be found, probably as a result of microbial reduction of sulfate. Reference values of sulfide concentration in the samples were obtained by iodometric titration with a separation step [2]. Concentration of potential interferents was determined by an outside analytical laboratory. Sulfate in samples A, B and G was found to be in range 450–500 mg L−1 and for other samples 60–100 mg L−1 . Concentrations of other ions in all samples was as follows: NO3 − < 5 mg L−1 , NO2 − < 0.1 mg L−1 , Fe (total) < 1 mg L−1 , Cu < 5 g L−1 . The following pH values of original samples were measured: samples A, B and G (6.9), C (8.2), D, E, F (7.4), and H (7.1). 2.4. Determination of sulfide by VG-ICP-OES method Determination of sulfide in water samples was realized by the continuous generation of gaseous hydrogen sulfide to the ICP source. Generation and separation of hydrogen sulfide from a sample has been performed in a hydride generator HG-1 with siphon gas/liquid separator. Hydrochloric acid (6 mol L−1 ) was applied for acidification and releasing of hydrogen sulfide from the sample. An ICP spectrometer was used for measuring the emission intensity of S(I) 180.676 nm line in radial mode. Linearity of the calibration curve was checked in the concentration range of sulfide 4–40 000 g L−1 . Interference effects of some potential interferents were checked for the VG-ICP-OES method. The compounds (ions) which interfere with iodometric and spectrophotometric determination (based on methylene blue reaction) were selected as potential interferents. Ions were added to model samples with a constant concentration of sulfide (100 g L−1 ) at the following concentration levels: 10, 100 and 1000 mg L−1 (for sulfate and bicarbonate), 1, 10 and 100 mg L−1
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
Fig. 1. Logarithmic form of calibration graph (sulfur emission line S(I) 180.676 nm) for the determination of S2− in the concentration range 4–40 000 g L−1 . Values below 80 g L−1 (black squares) are not included in linear regression.
(for nitrite, nitrate, sulfite, Fe3+ , Fe2+ ) and 0.1, 1 and 10 mg L−1 for Cu2+ . The concentration range for the interferents was selected with respect to their concentration in the real spring, surface or waste water samples. 3. Results and discussion 3.1. Method development Determination of hydrogen sulfide and sulfide in water samples through vapor generation ICP optical emission spectrometry is based on the separation of hydrogen sulfide and the detection of sulfur. Hydrogen sulfide dissolved and/or fixed in sulfides is released from a sample by acidification. Separation of hydrogen sulfide from the liquid phase is provided by a gas/liquid phase separator and then the gaseous phase is introduced to plasma where the atomization and excitation of sulfur atoms is performed. Subsequently optical emission is registered by optical emission spectrometry. The intensity of emission is proportional to the concentration of sulfidic sulfur in the sample. Linearity of the calibration curve with our VG system was tested in the concentration range 4–40 000 g L−1 S2− for two sulfur emission lines S(I) 180.676 nm and S(I) 181.978 nm. Sensitivities of both sulfur emission lines were comparable thus only emission line S(I) 180.676 nm was used for further measurements. The calibration curve was found to be linear in the range 80–40 000 g L−1 (Fig. 1). Determination of concentration below about 80 g L−1 was found to be influenced by a nonzero blank value which is caused by the presence of impurities, such as hydrogen sulfide or sulfur dioxide, in argon, the atmosphere and probably in the chemicals. Limit of detection for sulfide, calculated as 3 of blank measurement (n = 10), was 30 g L−1 . The blank samples were prepared with deionized water by the preservation method described below. LOD value of our configuration is about five-times higher when compared to other works [7,8] as axially-viewed ICP was also applied in these works. 3.2. Sample preservation and treatment The real samples have to be preserved immediately after sampling otherwise changes in sulfide concentration caused by oxidation or vaporization can be observed. For the preservation, alkalinization with a reducing agent [9] or the addition of a heavy metal salt (Zn, Cd) with subsequent alkalinization [10] can be used. Hydrogen sulfide and sulfides are precipitated as ZnS or CdS and hence are protected from oxidation and/or vaporization. Excess of
1779
a heavy metal cation is precipitated as hydroxide. When pumping such a sample for analysis by continuous flow vapor generation, vigorous mixing of this mixture is necessary to ensure a constant input of sample and, as a result, constant analyte signal. Dissolution of the formed precipitate by di-sodium EDTA salt was used as an alternative technique. The amount of EDTA must be sufficient for dissolution of all the sulfide and hydroxide precipitate and depends on its amount. In our case 200 mg of Na2 EDTA·2H2 O was sufficient for 100 mL of water sample preserved with 1 mL 10% (m/m) zinc acetate followed by mixing the contents for 5 min. Dissolution of the precipitate results in a homogenous solution or very fine dispersed colloid particles and, as a consequence, repeatability (RSD) of the measurement is significantly improved. Results of analysis of two different samples by both methods, with and without dissolution of the precipitate, are shown in Table 2. Results of both methods are comparable in terms of accuracy. However, repeatability of the measurement is improved in the case of precipitate dissolution and there is no need for sample mixing during analysis. 3.3. Interference study Spectrophotometric determination of sulfide, iodometric titration and other methods can be interfered by a family of compounds. Therefore the effect of known interferents and some major anions in natural waters on the vapor generation technique was investigated. The influence of the following ions was tested: nitrite, nitrate, sulfate, bicarbonate, sulfite and the effect of Fe2+ , Fe3+ and Cu2+ cation. The effect of the interferents was tested with preserved samples of sulfide with subsequent dissolution of precipitate by Na2 EDTA. Sulfate does not interfere with sulfide determination as it does not form volatile species after acidification and thus is separated in a gas/liquid separator as liquid waste (see Table 3). Nitrate in the sample does not interfere in the investigated range (up to 100 mg L−1 ). A weak suppression of the sulfide signal could be observed with a higher bicarbonate concentration. Concentration of bicarbonate over 100 mg L−1 can frequently be found in natural water and, moreover, the concentration of dissolved free carbon dioxide must be taken into account since it forms (bi)carbonate with an alkaline reagent (hydroxide) used for the sample preservation. The mechanism of this effect was not investigated in this study. Significant positive interference was observed in the case of sulfite. Even the lowest concentration 1 mg L−1 caused a significant increase of the sulfur signal. The signal appearing at 1 and 100 mg L−1 of sulfite corresponds to 0.09 and 16.4 mg L−1 of sulfide, respectively. This positive interference is induced by volatilization of sulfur dioxide formed by acidification of the sulfite. Hydrogen sulfide and sulfur dioxide cannot be distinguished in the detector as ICP-OES is an element-specific detector. On the other hand, this fact can be employed in the determination of sulfur dioxide and/or sulfite, for example, in wine samples [12]. Conversely, the determination of sulfide at lower concentrations of sulfite can be influenced by a mutual reaction of both gases (ions in acidic solution) producing elemental sulfur as described in [13]. Due to this fact we can Table 2 Influence of precipitate dissolution on repeatability of results. Sample
Dissolutiona
Sulfide concentration (mg L−1 )
RSD (%)
B
+ − + −
1.14 ± 0.02 1.2 ± 0.2 3.5 ± 0.1 3.4 ± 0.4
2 13 3 11
D
The value ± denotes standard deviation of measured concentration (n = 3), RSD = relative standard deviation a +: Determination with precipitate dissolution, −: determination without precipitate dissolution.
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
1780
Table 3 Relative VG-ICP-OES signal of sulfide (100 g L−1 ) in the presence of potential chemical interferents. cinterferent (mg L−1 )
0.1 1 10 100 1000
Relative S signal (%) SO4 2−
HCO3 −
NO3 −
NO2 −
SO3 2−
Fe3+
Fe2+
Cu2+
– – 104 ± 3 102 ± 3 102 ± 3
– – 98 ± 3 95 ± 4 88 ± 3
– 102 ± 3 104 ± 3 101 ± 3 –
– 103 ± 3 75 ± 3 19 ± 3 –
– 190 ± 3 1760 ± 60 16400 ± 800 –
– 99 ± 3 101 ± 3 90 ± 3 –
– 105 ± 3 94 ± 3 68 ± 4 –
86 ± 3 40 ± 3 10 ± 3 – –
The value ± denotes standard deviation of relative signal (n = 3).
speculate that, in such cases, the ICP signal of the volatile sulfur species would decrease as elemental sulfur is not vaporized under these conditions. The amount of volatile sulfur species entering the plasma will depend on the molar ratio of sulfide and sulfite and, in the case of an equimolar ratio (2 mol sulfide + 1 mol sulfite), the signal should reach the minimum. The excess of sulfite will result in production of sulfur dioxide and a subsequent increase in the emission signal. However, interferences of sulfite at the equimolar sulfide to sulfite ratio and below were not investigated. This interference can be significant in the case of waste water from the pulp and paper industry where both sulfite and sulfide can be found. Their mixture is however stable only in an alkaline environment for the reasons given above. Negative interference was also observed for nitrite at concentrations over 1 mg L−1 . Oxidative properties of nitrite in acidic solution result in an oxidation of sulfide to elemental sulfur and signal suppression. However, nitrite concentration in spring water should be far below this limit which is not the case of waste water. For Fe2+ and Fe3+ ions, the signal suppression can be observed at Fe concentrations over 10 mg L−1 . For Fe3+ at 100 mg L−1 signal suppression is about 10% which is an acceptable measure of error. The same concentration of Fe2+ results in 40% decrease in signal. In the case of Cu2+ only 0.1 mg L−1 results in 10% decrease of signal, 1 mg L−1 Cu2+ results in even more than 50% decrease. Formation of insoluble sulfides causes signal suppression in the cases of Fe and Cu as these sulfides do not react with hydrochloric acid quantitatively in the hydrogen sulfide generation step. Other cations forming stable, acid resistant sulfides can be expected as interfering ions for this technique. On the other hand, the proportion of sulfur related to metallic insoluble sulfides is considered to be non-toxic in an aquatic environment. The distillation separation step for the determination of this type of sulfide requires a reaction performed at higher temperatures with over a prolonged time sufficient for almost the complete sulfide decomposition. The same technique can be used for determination of acid-volatile sulfide (AVS) in sediments [6].
Table 4 Results of the real samples of water with different concentration analyzed by VGICP-OES method and iodometric titration with distillation step.
3.4. Application to real samples
1. Sulfur-containing species forming volatile compounds in acidic conditions (sulfite demonstrated here, predictably thiosulfate, dithionate and volatile organic sulfur-containing species). 2. Oxidants oxidizing sulfide to non-volatile sulfur. 3. Heavy metal ions forming insoluble sulfides.
The VG-ICP-OES method was applied to a series of spring water samples from the inspection wells in the region Hodonín (South Moravia, Czech Republic). Results from the VG-ICP-OES method and iodometric titration with distillation separation step are presented in Table 4. Results of both methods were tested by the Student’s t-test (˛ = 0.05, critical value t˛ = 4.303). Because of the low power of this test (due to low n = 3), results were also tested using the pair t-test with t = 0.988 (˛ = 0.05, critical value t˛ = 2.365). Both these tests revealed non-significant statistical differences between the values from LOD to 22 mg L−1 . Routine elemental analysis of samples indicated that concentrations of Fe and Cu, as the main cationic interferents, were below maximal acceptable limits resulting from the study in Section 3.3. These limits can be determined as 10 and 0.1 mg L−1 for total Fe and Cu, respectively. Hence concentrations of both these
Sample
Iodometric titration (mg L−1 )
VG-ICP-OES (mg L−1 )
t-test
A B C D E F G H
0.06 ± 0.01 1.05 ± 0.07 3.10 ± 0.05 3.7 ± 0.1 6.2 ± 0.1 6.5 ± 0.2 7.2 ± 0.1 20.3 ± 0.8
0.06 ± 0.01 1.14 ± 0.02 3.18 ± 0.05 3.5 ± 0.1 6.2 ± 0.1 6.4 ± 0.3 7.26 ± 0.08 22 ± 2
0 1.748 1.600 1.500 0.571 0.513 0.588 1.070
The value ± denotes standard deviation of measured concentration (n = 3).
elements in real samples did not affect the accuracy of sulfide determination. 4. Conclusion The applicability and reliability of the vapor generation technique with ICP detection for sulfide determination has been demonstrated in previous works and in our recent work. These methodologies have certain advantages compared to the methods usually applied for the determination of sulfide, including better LOD, lower time and chemical consumption and higher sample throughput, especially when compared to iodometric titration following distillation separation. The main disadvantage of these instrumental methods is the need for standardized calibration solutions resulting from the inaccessibility of defined chemicals and the instability of sulfide solution. In our work, applicability of sulfide preservation by co-precipitation with zinc hydroxide was demonstrated. Dissolution of precipitate by EDTA prior to VG-ICP-OES analysis results in a homogenous mixture and improves measurement repeatability. As a chemical reaction is involved in the vapor generation step, chemical interferences can be expected with this method. Generally, three groups of chemical interferents can be mentioned:
These interferences will definitely be observed in the VG technique with experimental setups, gas-phase separation or reaction mixture nebulization, and ICP-OES or ICP-MS detection. Acknowledgement Authors would like to thank the Ministry of Education, Youth and Sports of the Czech Republic for financial support in project ˇ registration number MSM0021622412 and also Dr. L. Durd’ová (Academy of Sciences of the Czech Republic) for providing samples and information on their composition.
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
References [1] P. Patnaik, A Comprehensive Guide to the Hazardous Properties of Chemical Substances, Wiley, New York, 1999. [2] Czech National Regulation 830530/31, Chemical and physical analysis of surface water, Determination of sulfide and hydrogen sulfide. [3] N. Gilboa-Garber, Anal. Biochem. 43 (1971) 129–133. [4] E. Schmidt, A. Marton, J. Hlavay, Talanta 41 (1994) 1219–1224. [5] S-H. Ohira, K. Toda, J. Chromatogr. A 1121 (2006) 280–284. [6] M. Colon, M. Iglesias, M. Hidalgo, Spectrochim. Acta B 62 (2007) 470–475. [7] M. Colon, M. Iglesias, M. Hidalgo, J.L. Todoli, J. Anal. At. Spectrom. 23 (2008) 416–418. [8] M. Colon, J.L. Todoli, M. Hidalgo, M. Iglesias, Anal. Chim. Acta 609 (2008) 160–168.
1781
[9] US EPA, SW-846, Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, Method 9215 Potentiometric Determination of Sulfide in Aqueous Samples and Distillates with Ion-Selective Electrode, 1996. [10] US EPA, SW-846, Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, Method 9030b – Acid-Soluble and Acid-Insoluble Sulfides: Distillation, 1996. [11] E. Niedobova, J. Machat, V. Otruba, V. Kanicky, J. Anal. At. Spectrom. 20 (2005) 945–949. [12] J. Cmelik, J. Machat, E. Niedobova, V. Otruba, V. Kanicky, Anal. Bioanal. Chem. 383 (2005) 483–488. [13] C. Chambers, A.K. Holliday, Modern Inorganic Chemistry, Butterworth & Co., London, 1975.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Contribution to vapor generation-inductively coupled plasma spectrometric techniques for determination of sulfide in water samples a ˇ Jiˇrí Cmelík , Jiˇrí Machát b,∗ , Vítˇezslav Otruba a , Viktor Kanicky´ a a b
Department of Chemistry, Faculty of Science, Masaryk University, Kotláˇrská 2, Brno, Czech Republic Research Centre for Environmental Chemistry and Ecotoxicology, Faculty of Science, Masaryk University, Kamenice 3, Brno, Czech Republic
a r t i c l e
i n f o
Article history: Received 9 July 2009 Received in revised form 23 September 2009 Accepted 8 October 2009 Available online 5 November 2009 Keywords: Hydrogen sulfide Sulfide Vapor generation ICP-OES Chemical interference
a b s t r a c t Vapor generation-inductively coupled plasma-optical emission spectrometry was used for the determination of sulfide in water samples preserved by the addition of a zinc acetate and sodium hydroxide solution. Hydrogen sulfide and acid-volatile sulfides were transformed, by acidification, to a gaseous phase in a vapor generator and subsequently detected by inductively coupled plasma optical emission spectrometry. Compounds interfering with iodometric titration and spectrophotometric determination were examined as potential chemical interferents. The proposed method provides results comparable to iodometric titration in the tested concentration range 0.06–22.0 mg L−1 . Limit of detection for the determination of hydrogen sulfide by this method is 0.03 mg L−1 . © 2009 Elsevier B.V. All rights reserved.
1. Introduction Sulfide concentration in a water sample is an important parameter for the assessment of spring, surface or waste water pollution [1]. There are several established methods for the determination of sulfide and/or hydrogen sulfide such as iodometric titration, spectrophotometric determination, potentiometry or ion chromatography [2–5]. Some drawbacks of these methods (insufficient limit of detection and their laboriousness) led to the development of their modification and to the establishment of completely new methods. The application of inductively coupled plasma (ICP) spectrometry hyphenated to a vapor generator (VG) for sulfide determination is one such innovative approach. Inductively coupled plasma-mass spectrometry hyphenated to vapor generation (VG-ICP-QMS) was used for the determination of sulfide in water samples in ref. [6]. The quadrupole analyzer of the mass spectrometer suffers from insufficient spectral resolution hence the measurement of sulfur is strongly interfered by polyatomic ions mainly from molecular ion O2 + . Oxygen forming these interfering ions arises particularly from water when nebulizing aqueous solutions. Generation of water-free gaseous hydrogen sulfide into the ICP results in the suppression of polyatomic interference. Further elimination of O2 + can be achieved by using a
∗ Corresponding author. Tel.: +420 549 495 684; fax: +420 549 492 840. E-mail address: [email protected] (J. Machát). 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.10.022
collision-reaction cell (CRC) in mixed H2 /He mode. In this work, optimization of the CRC parameters was primarily solved to obtain a satisfactory LOD. The LOD obtained for VG-ICP-MS determination of sulfide in water was 2 g L−1 . Inductively coupled plasma-optical emission spectrometry (ICP-OES) can also be used for sulfur detection. Measurement of the main sulfur analytical emission lines in vacuum UV range are not usually subject to spectral interference as the ICP emission spectrum in this region is relatively low in emission lines. In the case of ICP-OES, two techniques of sample introduction, both taking full advantage of vapor generation, can be used. A vapor generator with a gas/liquid phase separator enables introduction of the gaseous phase and thus only volatilized sulfide is detected [7]. The advantage of this technique is the elimination of spectral and non-spectral interferences in the ICP resulting from matrix separation. Nebulization of the reaction mixture (sample and acid) is another concept for the introduction of a volatile analyte allowing sensitive determination of sulfide as well as sulfate, latter being introduced into the ICP as an aerosol. Both these techniques were applied by [7,8]. Limits of detection for sulfide using VG and aerosol were 6 and 5 g L−1 , respectively, with an axially-viewed ICP. Hydrogen sulfide and sulfide concentration in water can be influenced by volatilization and/or oxidation; hence preservation of the samples is necessary in the sampling step. Preservation can be provided, for example, by adjustment of the pH to 8.5–9 with sodium hydroxide and ascorbic acid [9], or by zinc or cadmium acetate with subsequent alkalinization [10] and storage at a low
1778
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
temperature (1–5 ◦ C). Sodium hydroxide solution (2 mol L−1 ) and ascorbic acid (0.2 mol L−1 ) was used in ref. [6] for sample treatment by mixing the same volumes of both water sample and buffer. Separation of the gaseous analyte from liquid waste is achieved in a vapor-generator, thus high concentrations of sodium in a sample does not present a problem. In the case of nebulization of a reaction mixture [7,8] into the ICP, high Na concentrations in the solution lead to plasma instability. In this case an ammonium/ammonium chloride buffer was successfully used for alkalinization. Both ICP methods were successfully applied to spiked and natural water samples [6–8] and to the determination of acid volatile sulfide in marine sediment [6]. Results of these techniques were compared to those of a potentiometric method. A vapor generator with a phase separation – ICP-OES method was also applied in our work for the determination of sulfide in water samples. The addition of zinc acetate/sodium hydroxide was used for the sample preservation with subsequent dissolution of the precipitate by adding EDTA prior to the measurement. Considering the lack of study of chemical interference in the abovementioned works [6–8], we have dealt with the effect of some ions in the vapor generation of hydrogen sulfide. Our method was applied to real spring water samples with sulfide concentrations in the range 0.06–22 mg L−1 and the results were compared with the results from iodometric titration following distillation separation of hydrogen sulfide from the sample. 2. Experimental 2.1. Instrumentation An optical emission spectrometer with inductively coupled plasma Jobin-Yvon 170 Ultrace (JY-Horiba, Longjumeau, France) and hydride generator HG-1 (Labtech Ltd., Brno, Czech Republic) were used in this work. The schematic of the vapor generation system can be found in [11]. Operating conditions are shown in Table 1. Standard alumina injector (i.d. 2.5 mm) in the JY demountable plasma torch was replaced by a quartz injector because of the significant memory effects with the alumina injector. Similar effects were observed with iodine vapor generation as presented by Niedobova et al. [11]. 2.2. Chemicals Chemicals of reagent grade quality (Lach-Ner, Neratovice, Czech Republic) and deionized water were used throughout the experiments. Zinc acetate, sodium hydroxide, di-sodium EDTA (dihydrate) and hydrochloric acid (6 mol L−1 ) were used for the preservation of the samples and following the generation of hydrogen sulfide. Chemical interferences were examined using the following chemicals: sodium nitrite, nitrate, sulfate, sulfite, Table 1 Operation conditions of ICP spectrometer and vapor generator. ICP spectrometer Forward power ICP (W) Plasma gas flow (L min−1 ) Carrier gas flow (L min−1 ) Sheath gas flow (L min−1 ) Monochromator purging (Ar) (L min−1 ) Plasma viewing
1200 12.0 0.5 0.5 3.0 Radial
Vapor generator Carrier gas flow (L min−1 ) Sample flow rate (mL min−1 ) Acid flow rate (mL min−1 )
0.5 1.0 0.5
bicarbonate, copper (II) sulfate pentahydrate, iron (II) sulfate heptahydrate, and iron (III) chloride hexahydrate. Standard sulfide solutions were prepared by dissolving an appropriate amount of sodium sulfide (nonahydrate) in sodium hydroxide solution (0.1 mol L−1 ) and by subsequent dilution. These calibration solutions were standardized by iodometric titration because of the poorly defined content of sulfide in sodium sulfide and its instability. In the case of distillation separation of hydrogen sulfide from the real samples, glycerol was used in a trap solution of sodium hydroxide [2]. Iodine volumetric solution (0.01 mol L−1 ) was prepared by precisely weighing potassium iodate and the excess of potassium iodide after acidification by sulfuric acid. Sodium thiosulfate solution (0.01 mol L−1 , prepared from sodium thiosulfate pentahydrate) was standardized by titration of an iodine solution using starch as an indicator. 2.3. Samples The proposed method was applied to spring water samples with different concentrations of sulfide which were preserved immediately after sampling. The preservation was performed by adding of 1 mL 10% (m/m) zinc acetate and 0.5 mL 4 mol L−1 sodium hydroxide to 100 mL of water sample. For VG-ICP-OES determination powdered Na2 EDTA·2H2 O (about 200 mg) was added to each sample and mixed thoroughly to dissolve the precipitate in the sample prior to the analysis. The amount of Na2 EDTA·2H2 O depends on the amount of the precipitate formed by the preservation. The preservation was performed in the laboratory during the development of the method and during the investigation of chemical interferences. Real samples were preserved immediately after the sampling procedure. Real samples of spring waters (100 mL) were sampled from the inspection wells in the region around Hodonín city (Southern Moravia, Czech Republic). The inspection wells are located near petroleum and exploratory wells where a high concentration of sulfate and variable concentrations of hydrogen sulfide can be found, probably as a result of microbial reduction of sulfate. Reference values of sulfide concentration in the samples were obtained by iodometric titration with a separation step [2]. Concentration of potential interferents was determined by an outside analytical laboratory. Sulfate in samples A, B and G was found to be in range 450–500 mg L−1 and for other samples 60–100 mg L−1 . Concentrations of other ions in all samples was as follows: NO3 − < 5 mg L−1 , NO2 − < 0.1 mg L−1 , Fe (total) < 1 mg L−1 , Cu < 5 g L−1 . The following pH values of original samples were measured: samples A, B and G (6.9), C (8.2), D, E, F (7.4), and H (7.1). 2.4. Determination of sulfide by VG-ICP-OES method Determination of sulfide in water samples was realized by the continuous generation of gaseous hydrogen sulfide to the ICP source. Generation and separation of hydrogen sulfide from a sample has been performed in a hydride generator HG-1 with siphon gas/liquid separator. Hydrochloric acid (6 mol L−1 ) was applied for acidification and releasing of hydrogen sulfide from the sample. An ICP spectrometer was used for measuring the emission intensity of S(I) 180.676 nm line in radial mode. Linearity of the calibration curve was checked in the concentration range of sulfide 4–40 000 g L−1 . Interference effects of some potential interferents were checked for the VG-ICP-OES method. The compounds (ions) which interfere with iodometric and spectrophotometric determination (based on methylene blue reaction) were selected as potential interferents. Ions were added to model samples with a constant concentration of sulfide (100 g L−1 ) at the following concentration levels: 10, 100 and 1000 mg L−1 (for sulfate and bicarbonate), 1, 10 and 100 mg L−1
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
Fig. 1. Logarithmic form of calibration graph (sulfur emission line S(I) 180.676 nm) for the determination of S2− in the concentration range 4–40 000 g L−1 . Values below 80 g L−1 (black squares) are not included in linear regression.
(for nitrite, nitrate, sulfite, Fe3+ , Fe2+ ) and 0.1, 1 and 10 mg L−1 for Cu2+ . The concentration range for the interferents was selected with respect to their concentration in the real spring, surface or waste water samples. 3. Results and discussion 3.1. Method development Determination of hydrogen sulfide and sulfide in water samples through vapor generation ICP optical emission spectrometry is based on the separation of hydrogen sulfide and the detection of sulfur. Hydrogen sulfide dissolved and/or fixed in sulfides is released from a sample by acidification. Separation of hydrogen sulfide from the liquid phase is provided by a gas/liquid phase separator and then the gaseous phase is introduced to plasma where the atomization and excitation of sulfur atoms is performed. Subsequently optical emission is registered by optical emission spectrometry. The intensity of emission is proportional to the concentration of sulfidic sulfur in the sample. Linearity of the calibration curve with our VG system was tested in the concentration range 4–40 000 g L−1 S2− for two sulfur emission lines S(I) 180.676 nm and S(I) 181.978 nm. Sensitivities of both sulfur emission lines were comparable thus only emission line S(I) 180.676 nm was used for further measurements. The calibration curve was found to be linear in the range 80–40 000 g L−1 (Fig. 1). Determination of concentration below about 80 g L−1 was found to be influenced by a nonzero blank value which is caused by the presence of impurities, such as hydrogen sulfide or sulfur dioxide, in argon, the atmosphere and probably in the chemicals. Limit of detection for sulfide, calculated as 3 of blank measurement (n = 10), was 30 g L−1 . The blank samples were prepared with deionized water by the preservation method described below. LOD value of our configuration is about five-times higher when compared to other works [7,8] as axially-viewed ICP was also applied in these works. 3.2. Sample preservation and treatment The real samples have to be preserved immediately after sampling otherwise changes in sulfide concentration caused by oxidation or vaporization can be observed. For the preservation, alkalinization with a reducing agent [9] or the addition of a heavy metal salt (Zn, Cd) with subsequent alkalinization [10] can be used. Hydrogen sulfide and sulfides are precipitated as ZnS or CdS and hence are protected from oxidation and/or vaporization. Excess of
1779
a heavy metal cation is precipitated as hydroxide. When pumping such a sample for analysis by continuous flow vapor generation, vigorous mixing of this mixture is necessary to ensure a constant input of sample and, as a result, constant analyte signal. Dissolution of the formed precipitate by di-sodium EDTA salt was used as an alternative technique. The amount of EDTA must be sufficient for dissolution of all the sulfide and hydroxide precipitate and depends on its amount. In our case 200 mg of Na2 EDTA·2H2 O was sufficient for 100 mL of water sample preserved with 1 mL 10% (m/m) zinc acetate followed by mixing the contents for 5 min. Dissolution of the precipitate results in a homogenous solution or very fine dispersed colloid particles and, as a consequence, repeatability (RSD) of the measurement is significantly improved. Results of analysis of two different samples by both methods, with and without dissolution of the precipitate, are shown in Table 2. Results of both methods are comparable in terms of accuracy. However, repeatability of the measurement is improved in the case of precipitate dissolution and there is no need for sample mixing during analysis. 3.3. Interference study Spectrophotometric determination of sulfide, iodometric titration and other methods can be interfered by a family of compounds. Therefore the effect of known interferents and some major anions in natural waters on the vapor generation technique was investigated. The influence of the following ions was tested: nitrite, nitrate, sulfate, bicarbonate, sulfite and the effect of Fe2+ , Fe3+ and Cu2+ cation. The effect of the interferents was tested with preserved samples of sulfide with subsequent dissolution of precipitate by Na2 EDTA. Sulfate does not interfere with sulfide determination as it does not form volatile species after acidification and thus is separated in a gas/liquid separator as liquid waste (see Table 3). Nitrate in the sample does not interfere in the investigated range (up to 100 mg L−1 ). A weak suppression of the sulfide signal could be observed with a higher bicarbonate concentration. Concentration of bicarbonate over 100 mg L−1 can frequently be found in natural water and, moreover, the concentration of dissolved free carbon dioxide must be taken into account since it forms (bi)carbonate with an alkaline reagent (hydroxide) used for the sample preservation. The mechanism of this effect was not investigated in this study. Significant positive interference was observed in the case of sulfite. Even the lowest concentration 1 mg L−1 caused a significant increase of the sulfur signal. The signal appearing at 1 and 100 mg L−1 of sulfite corresponds to 0.09 and 16.4 mg L−1 of sulfide, respectively. This positive interference is induced by volatilization of sulfur dioxide formed by acidification of the sulfite. Hydrogen sulfide and sulfur dioxide cannot be distinguished in the detector as ICP-OES is an element-specific detector. On the other hand, this fact can be employed in the determination of sulfur dioxide and/or sulfite, for example, in wine samples [12]. Conversely, the determination of sulfide at lower concentrations of sulfite can be influenced by a mutual reaction of both gases (ions in acidic solution) producing elemental sulfur as described in [13]. Due to this fact we can Table 2 Influence of precipitate dissolution on repeatability of results. Sample
Dissolutiona
Sulfide concentration (mg L−1 )
RSD (%)
B
+ − + −
1.14 ± 0.02 1.2 ± 0.2 3.5 ± 0.1 3.4 ± 0.4
2 13 3 11
D
The value ± denotes standard deviation of measured concentration (n = 3), RSD = relative standard deviation a +: Determination with precipitate dissolution, −: determination without precipitate dissolution.
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
1780
Table 3 Relative VG-ICP-OES signal of sulfide (100 g L−1 ) in the presence of potential chemical interferents. cinterferent (mg L−1 )
0.1 1 10 100 1000
Relative S signal (%) SO4 2−
HCO3 −
NO3 −
NO2 −
SO3 2−
Fe3+
Fe2+
Cu2+
– – 104 ± 3 102 ± 3 102 ± 3
– – 98 ± 3 95 ± 4 88 ± 3
– 102 ± 3 104 ± 3 101 ± 3 –
– 103 ± 3 75 ± 3 19 ± 3 –
– 190 ± 3 1760 ± 60 16400 ± 800 –
– 99 ± 3 101 ± 3 90 ± 3 –
– 105 ± 3 94 ± 3 68 ± 4 –
86 ± 3 40 ± 3 10 ± 3 – –
The value ± denotes standard deviation of relative signal (n = 3).
speculate that, in such cases, the ICP signal of the volatile sulfur species would decrease as elemental sulfur is not vaporized under these conditions. The amount of volatile sulfur species entering the plasma will depend on the molar ratio of sulfide and sulfite and, in the case of an equimolar ratio (2 mol sulfide + 1 mol sulfite), the signal should reach the minimum. The excess of sulfite will result in production of sulfur dioxide and a subsequent increase in the emission signal. However, interferences of sulfite at the equimolar sulfide to sulfite ratio and below were not investigated. This interference can be significant in the case of waste water from the pulp and paper industry where both sulfite and sulfide can be found. Their mixture is however stable only in an alkaline environment for the reasons given above. Negative interference was also observed for nitrite at concentrations over 1 mg L−1 . Oxidative properties of nitrite in acidic solution result in an oxidation of sulfide to elemental sulfur and signal suppression. However, nitrite concentration in spring water should be far below this limit which is not the case of waste water. For Fe2+ and Fe3+ ions, the signal suppression can be observed at Fe concentrations over 10 mg L−1 . For Fe3+ at 100 mg L−1 signal suppression is about 10% which is an acceptable measure of error. The same concentration of Fe2+ results in 40% decrease in signal. In the case of Cu2+ only 0.1 mg L−1 results in 10% decrease of signal, 1 mg L−1 Cu2+ results in even more than 50% decrease. Formation of insoluble sulfides causes signal suppression in the cases of Fe and Cu as these sulfides do not react with hydrochloric acid quantitatively in the hydrogen sulfide generation step. Other cations forming stable, acid resistant sulfides can be expected as interfering ions for this technique. On the other hand, the proportion of sulfur related to metallic insoluble sulfides is considered to be non-toxic in an aquatic environment. The distillation separation step for the determination of this type of sulfide requires a reaction performed at higher temperatures with over a prolonged time sufficient for almost the complete sulfide decomposition. The same technique can be used for determination of acid-volatile sulfide (AVS) in sediments [6].
Table 4 Results of the real samples of water with different concentration analyzed by VGICP-OES method and iodometric titration with distillation step.
3.4. Application to real samples
1. Sulfur-containing species forming volatile compounds in acidic conditions (sulfite demonstrated here, predictably thiosulfate, dithionate and volatile organic sulfur-containing species). 2. Oxidants oxidizing sulfide to non-volatile sulfur. 3. Heavy metal ions forming insoluble sulfides.
The VG-ICP-OES method was applied to a series of spring water samples from the inspection wells in the region Hodonín (South Moravia, Czech Republic). Results from the VG-ICP-OES method and iodometric titration with distillation separation step are presented in Table 4. Results of both methods were tested by the Student’s t-test (˛ = 0.05, critical value t˛ = 4.303). Because of the low power of this test (due to low n = 3), results were also tested using the pair t-test with t = 0.988 (˛ = 0.05, critical value t˛ = 2.365). Both these tests revealed non-significant statistical differences between the values from LOD to 22 mg L−1 . Routine elemental analysis of samples indicated that concentrations of Fe and Cu, as the main cationic interferents, were below maximal acceptable limits resulting from the study in Section 3.3. These limits can be determined as 10 and 0.1 mg L−1 for total Fe and Cu, respectively. Hence concentrations of both these
Sample
Iodometric titration (mg L−1 )
VG-ICP-OES (mg L−1 )
t-test
A B C D E F G H
0.06 ± 0.01 1.05 ± 0.07 3.10 ± 0.05 3.7 ± 0.1 6.2 ± 0.1 6.5 ± 0.2 7.2 ± 0.1 20.3 ± 0.8
0.06 ± 0.01 1.14 ± 0.02 3.18 ± 0.05 3.5 ± 0.1 6.2 ± 0.1 6.4 ± 0.3 7.26 ± 0.08 22 ± 2
0 1.748 1.600 1.500 0.571 0.513 0.588 1.070
The value ± denotes standard deviation of measured concentration (n = 3).
elements in real samples did not affect the accuracy of sulfide determination. 4. Conclusion The applicability and reliability of the vapor generation technique with ICP detection for sulfide determination has been demonstrated in previous works and in our recent work. These methodologies have certain advantages compared to the methods usually applied for the determination of sulfide, including better LOD, lower time and chemical consumption and higher sample throughput, especially when compared to iodometric titration following distillation separation. The main disadvantage of these instrumental methods is the need for standardized calibration solutions resulting from the inaccessibility of defined chemicals and the instability of sulfide solution. In our work, applicability of sulfide preservation by co-precipitation with zinc hydroxide was demonstrated. Dissolution of precipitate by EDTA prior to VG-ICP-OES analysis results in a homogenous mixture and improves measurement repeatability. As a chemical reaction is involved in the vapor generation step, chemical interferences can be expected with this method. Generally, three groups of chemical interferents can be mentioned:
These interferences will definitely be observed in the VG technique with experimental setups, gas-phase separation or reaction mixture nebulization, and ICP-OES or ICP-MS detection. Acknowledgement Authors would like to thank the Ministry of Education, Youth and Sports of the Czech Republic for financial support in project ˇ registration number MSM0021622412 and also Dr. L. Durd’ová (Academy of Sciences of the Czech Republic) for providing samples and information on their composition.
ˇ J. Cmelík et al. / Talanta 80 (2010) 1777–1781
References [1] P. Patnaik, A Comprehensive Guide to the Hazardous Properties of Chemical Substances, Wiley, New York, 1999. [2] Czech National Regulation 830530/31, Chemical and physical analysis of surface water, Determination of sulfide and hydrogen sulfide. [3] N. Gilboa-Garber, Anal. Biochem. 43 (1971) 129–133. [4] E. Schmidt, A. Marton, J. Hlavay, Talanta 41 (1994) 1219–1224. [5] S-H. Ohira, K. Toda, J. Chromatogr. A 1121 (2006) 280–284. [6] M. Colon, M. Iglesias, M. Hidalgo, Spectrochim. Acta B 62 (2007) 470–475. [7] M. Colon, M. Iglesias, M. Hidalgo, J.L. Todoli, J. Anal. At. Spectrom. 23 (2008) 416–418. [8] M. Colon, J.L. Todoli, M. Hidalgo, M. Iglesias, Anal. Chim. Acta 609 (2008) 160–168.
1781
[9] US EPA, SW-846, Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, Method 9215 Potentiometric Determination of Sulfide in Aqueous Samples and Distillates with Ion-Selective Electrode, 1996. [10] US EPA, SW-846, Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, Method 9030b – Acid-Soluble and Acid-Insoluble Sulfides: Distillation, 1996. [11] E. Niedobova, J. Machat, V. Otruba, V. Kanicky, J. Anal. At. Spectrom. 20 (2005) 945–949. [12] J. Cmelik, J. Machat, E. Niedobova, V. Otruba, V. Kanicky, Anal. Bioanal. Chem. 383 (2005) 483–488. [13] C. Chambers, A.K. Holliday, Modern Inorganic Chemistry, Butterworth & Co., London, 1975.