Talanta 140 (2015) 189–197
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Selective and trace determination of monochloramine in river water by chemical derivatization and liquid chromatography/tandem mass spectrometry analysis Said Kinani a,n, Stéphany Layousse a, Bertille Richard a, Aziz Kinani a,b, Stéphane Bouchonnet b,1, Astrid Thoma c, Frank Sacher c,2 a Laboratoire National d'Hydraulique et Environnement (LNHE), Division Recherche et Développement, Electricité de France (EDF), 6 Quai de Watier, 78401 Chatou Cedex 01, France b Laboratoire de Chimie Moléculaire, Ecole Polytechnique, 91128 Palaiseau, France c DVGW – Technologiezentrum Wasser (TZW), Karlsruher Strasse 84, 76139 Karlsruhe, Germany
art ic l e i nf o
a b s t r a c t
Article history: Received 11 December 2014 Received in revised form 16 March 2015 Accepted 22 March 2015 Available online 30 March 2015
Monochloramine (MCA) may enter the aquatic environment through three main sources: wastewater treatment plant effluents, industrial effluents and thermal power plant wastes. Up to date, there are no available data about the concentration levels of this chemical in river water due to lack of appropriate analytical methods. Therefore, sensitive and selective analytical methods for monochloramine analysis in river water are required to evaluate its environmental fate and its effects on aquatic ecosystems. Thus, in this study we describe a highly specific and sensitive method for monochloramine determination in river water. This method combines chemical derivatization of monochloramine into indophenol followed by liquid chromatography coupled to electrospray ionisation–tandem mass spectrometry (LC–ESI–MS/MS) analysis. Two precursor-to-product ion transitions were monitored (200-127 and 200-154) in positive ionisation mode, fulfilling the criteria of selectivity, in accordance with the European Legislation requirements (decision 2002/657/EC). Ion structures and fragmentation mechanisms have been proposed to explain the selected transitions. Linearity range, accuracy and precision of the method have been assessed according to the French method validation standard NF T90-210. Detecting the derivatized monochloramine (indophenol) in Multiple Reaction Monitoring (MRM) mode provided a limit of quantification of 40 ng L 1 equivalent monochloramine. Applied to Loire river water (France), the developed method occasionally detected monochloramine at concentrations less than 300 ng L 1, which could be explained by punctual discharges of water containing active chlorine upstream of the sampling point. Indeed, it is widely reported in the literature that the addition of chlorine to water containing ammonia (e.g., wastewater effluents and river water) may result in the instantaneous formation of monochloramine. The proposed method is a powerful tool that can be used in environmental research (e.g., assessment of environmental fate and generating of ecotoxicological data) as well as in research studies concerning the evaluation of water disinfection efficiency; but it is not currently appropriate for routine use in industrial applications given the complexity of the procedure, the instability of indophenol and the use of certain toxic reagents. & 2015 Elsevier B.V. All rights reserved.
Keywords: Monochloramine River water Chemical derivatization Liquid chromatography/tandem mass spectrometry Ecotoxicology and environmental fate Ecotoxicological risk assessment
1. Introduction
n
Corresponding author. Tel.: þ 33 1 30 87 91 13. E-mail addresses: [email protected] (S. Kinani), [email protected] (S. Bouchonnet), [email protected] (F. Sacher). 1 Tel.: þ33 1 69 33 48 05. 2 Tel.: þ49 721 9678 156. http://dx.doi.org/10.1016/j.talanta.2015.03.043 0039-9140/& 2015 Elsevier B.V. All rights reserved.
The term “inorganic chloramines” refers to a group of three chemical substances formed by a reaction between chlorine and ammonia in an aqueous medium: monochloramine (MCA, NH2Cl), dichloramine (DCA, NHCl2), and trichloramine (TCA, NCl3). Concentration and speciation of these species primarily depend on two parameters: pH value and chlorine to nitrogen ratio (Cl2/N). Only MCA has industrial uses in the water treatment sector, in
190
S. Kinani et al. / Talanta 140 (2015) 189–197
particular as a biocide for inactivating pathogenic microorganisms and viruses to prevent the propagation of water-borne infectious diseases [1–4]. Monochloramine may be also formed as unwanted by-products of water chlorination in the presence of ammonium ions (e.g. in swimming pools or polluted natural water) [5]. Compared to chlorine, monochloramine is less reactive towards the organic matter present in water and therefore produces total organic halogen and regulated disinfection by-products in lower quantities, especially trihalomethanes (THM) and haloacetic acids (AHA) [6,7]. Nevertheless, MCA has been reported to generate higher levels of nitrogenous organic halogenated by-products such as haloacetonitriles, cyanogen halides, and haloacetamides than chlorination [6,8–10]. It has also been reported that disinfection of iodide-containing waters with monochloramine or by chlorination in the presence of ammonia favours the formation of iodinated organic compounds such as iodo-THM and iodo-AHA [11–13]. In water treatment processes, monochloramine as a secondary disinfectant is currently a well-established technology and even recommended by the WHO [14]. It is used in several industrialised countries such as Canada, Japan, USA, the United Kingdom and Australia. According to estimates, around 30% of drinking water plants in the USA use MCA, and a further 8–12% intend to adopt it as a treatment [15]. In France and Germany, monochloramine is currently banned as a disinfectant for water intended for human consumption. It is, however, authorised for the disinfection of water in industrial cooling systems. It is an example used by Electricité de France (EDF) to disinfect the cooling circuits in some of its thermal power plants located alongside rivers. Monochloramine may enter the aquatic environment through three main sources: wastewater treatment plant effluent, industrial effluent and thermal power plant waste. There is therefore a need to investigate whether the aquatic ecosystem is being affected by effluents containing residual MCA. A report on Canadian environmental quality standards for chlorine species was published in 1999 and lists the main ecotoxicological effects of MCA [16]. The lowest acute toxicity values given in this report are between 14 and 82 mg L 1 for fish, and between 16 and 78 mg L 1 for invertebrates. Thus, in order to assess monochloramine impact on the aquatic environment, there is a need for a sensitive and specific analytical method to quantify this compound at trace concentrations in river water. Quantifying trace levels of inorganic chloramines poses a major challenge, since the available analytical methods, including standardised and benchmark methods, are not sensitive and specific enough to detect these compounds in river water. Scientific articles report the use of two types of methods: those involving colorimetric, titrimetric or amperometric analysis, and those involving chromatography and membrane introduction mass spectrometry (MIMS) [17]. DPD (N,N-diethyl-phenylenediamine) colorimetry is currently the most widely used method. This standardised method (EN ISO 7393) is easy to use and suitable for field measurements, but has two main limitations: its lack of MCA specificity and insufficient sensitivity (LOQ 430 mg eq. Cl2 L 1) for the concentrations expected to be found in river water (o1 mg eq. Cl2 L 1) [18]. To overcome the lack of specificity of DPD colorimetry, other colorimetric methods using the Berthelot reaction have been developed. They are based on the principle of measuring the intensity of the indophenol colour produced when monochloramine and phenol react in an alkaline medium and in the presence of a catalyst. Due to the toxic and noxious nature of phenol, several compounds are now being used as alternatives. Although specific, these methods are not sensitive enough to measure trace levels of monochloramine (LOQ 4 50 mg eq. Cl2 L 1). DPD and ferrous ammonium sulphate (FAS) titration is another standardised method routinely used by water processing laboratories. However, it is unsuitable for field use and its performance is
similar to that of DPD colorimetric testing. Amperometric titration with sodium thiosulphate (Na2S2O3) or phenylarsine oxide (PAO, C6H5AsO) is also a widely used standardised method, especially for the in situ quantification of free and combined chlorine. This method's limit of quantification is similar to that of titrimetric and colorimetric methods, but it has better selectivity. Liquid chromatography (LC) and membrane introduction mass spectrometry (MIMS) methods have been developed to overcome the limitations posed by these traditional methods. Very few methods employing liquid chromatography without prior chemical derivatization of the chloramines have been developed. The detection is performed using UV photometry or electrochemical detection, which vastly reduces the sensitivity and selectivity of the measurements. In order to improve the sensitivity and selectivity, two pre-column derivatization agents have been suggested: 5-(dimethylamino) na phthalene-1-sulphonic acid (or dansyl sulphonic acid, DANSO2H) and 2-mercapto-benzothiazole (2-MBTZ). However the reactions involving those two reagents present various limitations in terms of complexity, reaction duration and/or low yield [19,20]. Membrane introduction mass spectrometry (MIMS) was used to analyse inorganic chloramines for the first time by Kotiaho et al. [21]. The principle of this technique is a selective introduction of molecules through a membrane placed between the sample and the ion source of a mass spectrometer. MIMS has the advantage of being selective and compatible with on-line and real-time analyses without any prior sample preparation stage [22]. However, the hydrophobic nature of the membrane promotes the diffusion of apolar molecules and limits that of polar and semi-polar compounds. This discrimination is a major limitation of the method for analysing monochloramine, since the best limit of quantification obtained to date is 0.2 mg L 1 [23,24]. Therefore, due to the lack of sensitive and specific analytical methods, there are no quantified data on the monochloramine concentration levels in river water. Liquid chromatography coupled with mass spectrometry (LC–MS) is a highly sensitive and specific technique for the quantification of polar and semi-polar molecules at trace concentrations. However, up to date, this technique has not been tested as a potential method for monochloramine analysis, probably due to its low molar mass (51.5 g mol 1). We report on the first attempts to develop a measurement method based on a combination of chemical derivatization of monochloramine into indophenol and LC–MS/MS quantitation of the later. This analytical method was assessed in river water according to the French standard NF T90-210 and successfully used to measure the concentration levels of monochloramine in the water of the river Loire (France) within the framework of research activities.
2. Materials and methods 2.1. Reagents and standard solutions 2.1.1. Reagents and solvents All reagents and solvents used to produce and analyse monochloramine were of analytical quality. Acetonitrile (HPLC grade) and potassium iodide were obtained from Scarlau (Spain) and Carl Roth (Karlsruhe, Germany) respectively. Ammonium chloride (purity 4 99.5%), sodium hypochlorite (15% Cl2), sulphuric acid (purity 4 95.0%), N,N-diethyl-p-phenylenediamine (DPD, purity 4 99.0%), lithium citrate tribasic tetrahydrate (purity 4 99.0%) and ammonium formate (purity 4 99.9%) were purchased from Sigma-Aldrich (Steinheim, Germany). Monopotassium phosphate (purity 4 99.5%), disodium hydrogen phosphate dehydrate (purity 4 99.5%), ethylenediaminetetraacetic acid (EDTA, purity 4 99.0%), potassium iodate (purity 4 99.0%),
S. Kinani et al. / Talanta 140 (2015) 189–197
191
sodium nitroferricyanide dehydrate (purity 4 99.0%), phenol (purity 4 99.5%) and thymolindophenol sodium (purity 4 30.0%) were acquired from Merck (Darmstadt, Germany). Demineralised water was used to prepare the reagents and the monochloramine standard solutions. 2.1.2. Preparing the monochloramine standard solutions in demineralised water MCA is unstable in an aqueous solution. There are therefore no commercially-available standard solutions and it must be prepared and calibrated using a reference analysis method. MCA standard solutions were obtained by diluting a 1 g L 1 stock solution of MCA, prepared by reaction between an ammonium chloride solution and sodium hypochlorite. The stock solution was prepared in three steps, using the following protocol. In the first step, 398.5 mg ammonium chloride (NH4Cl) was dissolved in 250 mL demineralised water. 100 mL of this solution (S1) was taken, the pH adjusted to 9.1 using a sodium hydroxide solution (2 M) and then refrigerated at þ4 °C for at least 30 min before use. In the second step, a sodium hypochlorite solution (S2) was prepared by diluting a commercial bleach solution (approximately 15% by mass), titrated beforehand using DPD colorimetry, in ultrapure water. The volume of commercial bleach solution was calculated to ensure a mass ratio of Cl2 to N of 4.8 in the S3 solution. The final volume was adjusted to 100 mL using demineralised water (S2). This solution was then refrigerated at þ4 °C for at least 30 min before use. In the third step, the monochloramine stock solution (S3) was prepared by carefully combining the pre-prepared S1 and S2 solutions. The synthetic monochloramine was titrated using DPD colorimetry as described in standard ISO 73932 [25]. The MCA stock solution can be kept for up to 24 h at þ4 °C, out of direct light. In order to ensure the accuracy of the concentration, DPD titration was used before any liquid chromatographic assay. 2.1.3. Preparing the indophenol standard solutions in river water For a standard solution with a defined concentration of MCA equivalents, a solution with a tenfold higher concentration of monochloramine was chemically derived using the method described in Section 2.2. Then, 1 mL of the derivatization product was added to a 10 mL volumetric flask and the river water to be analysed was added up to the mark. 2.2. Chemical derivatization – the Berthelot reaction 2.2.1. Principle of chemical derivatization reaction For this study, it was decided to use chemical derivatization due to the need to convert monochloramine into a chemical product with a higher molecular weight that could be analysed with high sensitivity using LC–MS/MS. The chemical derivatization explored is based upon the Berthelot reaction, also known as “indophenol method”, which is widely used for the determination of ammonia and ammonium ions in water [26–28]. This reaction is illustrated in Fig. 1. The Berthelot reaction depends on many factors, the most decisive ones being the type of chemical derivatizing agent (phenol-type reagents), temperature, reaction time, pH and catalyst [27,29]. The advantage of chemical derivatization using the Berthelot reaction comes from its specificity to MCA. This is due to the presence of two exchangeable hydrogen atoms in the monochloramine molecule, which is not the case for dichloramine, trichloramine or organic chloramines [29]. Harp [30], Lee et al. [18] and Tao et al. [29] have shown that the chemical derivatization of MCA into indophenol is little affected by the presence of amines and organic chloramines. Due to the toxicity of phenol, several alternative phenolbased reagents have been tested for the determination of ammonia
Fig. 1. Schematic representation of the derivatization reaction of monochloramine by phenol to form indophenol.
and ammonium ions: o-phenylphenol ((1,1′-biphenyl)-2-ol) by Kanda [31], 1,2-dihydroxybenzene by Harp [30], thymol by Hata et al. [32] and Moliner-Martinez et al. [33], salicylic acid by Tao et al. [29] and 1-naphthol by Shoji et al. [34,35]. To our knowledge only phenol, 2-hydroxyphenol and salycilate have been tested for analysing monochloramine [29,30,36]. Despite its toxicity, phenol remains the most commonly used reagent. In the presence of sodium nitroprusside, monochloramine reacts fully with phenol in just a few minutes, compared to other reagents which have much slower reaction times. There has been extensive research into the Berthelot reaction. Since its initial development, several modifications to enhance efficacy have been made. However, optimal operating conditions vary from one study to another, as shown by the comparisons carried out by [26]. 2.2.2. Derivatization procedure A chemical derivatization protocol based on the one described by [30] was used. A sample volume of 20 mL was placed in a 25 mL beaker containing 0.5 g lithium citrate to avoid any precipitation of the alkaline earth hydroxides during the chemical derivatization. The mixture was stirred until the sodium citrate crystals had fully dissolved, then 0.1 mL sodium nitroferricyanide (5 g L 1) and 1.2 mL phenol (100 g L 1), both prepared fresh each day, were added. The pH of the solution was adjusted to a value between 12.0 and 12.2 using a 5 M sodium hydroxide solution. The mixture was transferred to a 25 mL volumetric flask which was then filled up to the mark with deionized water. After 20 min at room temperature (20–25 °C), the reaction product (indophenol) was measured using spectrophotometry at 635 nm wavelength, or by LC–MS/MS after or without solid phase extraction. All samples (about 1 mL) were post-derivatization filtered through a 0.45 mm cellulose acetate membrane syringe filter (Whatman) before LC– MS/MS analysis. 2.2.3. Sample preparation In order to increase the sensitivity of the method, a sample preparation protocol involving solid phase extraction (SPE) after chemical derivatization (post-SPE) was tested. Several SPE cartridges were tested. The best results were obtained with the LiChrolut polymeric cartridge (200 mg, 6 mL, Merck Millipore). SPE was performed with the following parameters: (1) cartridge preconditioned with 5 mL MeOH followed by 5 mL demineralised water; (2) percolation by slowly aspirating 10 mL of the sample at a flow rate of approximately 0.5 mL min 1; (3) vacuum drying for 3 min to eliminate any traces of water; and (4) elution of indophenol with 6 mL MeOH. The eluate was dried by adding a few milligrams of anhydrous sodium sulphate (Na2SO4) and then filtered through a 0.45 mm cellulose acetate membrane syringe filter (Whatman). The filtrate (MeOH) was evaporated to dryness in a gentle current of nitrogen, and then dissolved in 200 mL of a H2O:ACN mixture (90:10, v/v) for LC–MS/MS analysis. The concentration factor is 50.
192
S. Kinani et al. / Talanta 140 (2015) 189–197
2.3. Equipment and analytical methods HPLC–MS analyses were conducted using an Agilent 1200 Infinity chromatography detector (Agilent Technologies, Waldbronn, Germany) coupled to an API 4000 QTrap triple quadrupole mass spectrometer (Applied Biosystems MDS SCIEX, Toronto, Canada). Data were collected and processed using the Analyst software (version 1.6.0, AB Sciex). The chromatographic separation was performed at 40 °C using a Gemini C18 reversed phase column (250 mm 2 mm, particle size 5 μm) from Phenomenex. After various optimisation tests, the mobile phase flow rate and the injection volume were set at 0.4 mL min 1 and 20 μL respectively. The mobile phase used for the chromatographic separation was a binary mixture comprising a mobile phase A (20 mM ammonium formate in water) and a mobile phase B (20 mM ammonium formate in a 90:10, v/v mixture of ACN:H2O), delivered at a constant flow rate. The gradient began with 90% phase A and decreased linearly over 7 min to 0%. The system was kept in isocratic elution mode for 5 min and then brought back to initial conditions in one minute. At the end of each test, the chromatography column was re-equilibrated for 10 min. The API 4000 QTrap is a hybrid mass spectrometer that can operate as a triple quadrupole (TQ) or a linear ion trap (trap); it was used as a triple quadrupole for this study. The Atmospheric Pressure Chemical Ionisation (APCI) and ElectroSpray Ionisation (ESI) modes were individually optimised and compared in both positive and negative ionisation modes. Positive mode ESI provided the best ion yield and was therefore selected for the rest of the study. Ultrapure nitrogen, produced using a laboratory nitrogen generator (Gengaz, France), was used as spray gas, desolvation gas, curtain gas and collision gas. The spray voltage, pressures of the spray gas, desolvation gas and curtain gas were set at 5.5 kV, 80, 60 and 10 psi respectively. The temperatures of the source and the desolvation nitrogen were set at 80 and 500 °C respectively. The Q1 and Q3 quadrupoles had unit-mass resolution. The nitrogen gas collusion induct dissociation (CID) was regulated at a pressure of 4 mTorr. The mass spectrometer was calibrated with infusion of a polypropylene glycol solution (2 mM in ACN) at a flow rate of 10 mL min 1. In this study, spectrophotometry was used for two different purposes: firstly to analyse the commercial bleach solution and the prepared monochloramine solutions, and secondly to monitor the kinetics of the phenol-induced chemical derivatization of the monochloramine. In each case, the samples were analysed using a CADAS 200 spectrophotometer (HACH LANGE Company). The measurements were carried out using containers with a 5 cm optical path length.
3. Results and discussion 3.1. Chemical derivatization 3.1.1. Indophenol formation kinetics The first step of method development and optimisation involved assessing the kinetics of the phenol-induced derivatization of monochloramine. The reproducibility of the results was confirmed by performing measurements from three monochloramine solutions with concentrations between 0.5 and 0.7 mg L 1 equivalent MCA. The measurements were performed by spectrophotometry at a wavelength of 635 nm. For each solution, the results were normalised against the maximum absorbance registered. Fig. 2 shows changes in the normalised absorbance over a 6 h period. The results in Fig. 2 show that the absorbance, typical for indophenol formation, develops gradually before levelling out
Fig. 2. Reaction kinetic of derivatization of monochloramine by phenol at room temperature.
after 15 min, and then remains stable for at least 6 h after the reagents have been added. It can therefore be concluded that, at room temperature, monochloramine is fully derived into indophenol after a 15 min reaction time. This finding is consistent with several other studies [37–39]. A reaction time of 20 min was retained for the study. 3.1.2. Temporal stability of indophenol Monochloramine is an unstable molecule in river water. In order to measure it, an adequate protocol for sample preservation usable directly on site is required. An interesting alternative and a way to overcome the problem of instability is to chemically derivatize monochloramine immediately after sampling to convert it into a stable product. However, special and sometimes restrictive precautions are needed when using hazardous reagents outside the laboratory, in order to ensure the safety of the personnel taking the samples. In the selected approach, the monochloramine was derivatized into indophenol, a product whose temporal stability needed to be verified. Indophenol stability in water has been tested during 9 days. Four preservation conditions were tested: A. In daylight and at laboratory temperature (18 °C); B. in darkness and at laboratory temperature; C. refrigerated at 4 °C; and D. frozen at 18 °C. A monochloramine solution with a concentration between 0.7 and 1.0 mg/L was prepared and chemically derivatized using the protocol described in Section 2.2. For each of the four preservation conditions, the stability of indophenol was verified in triplicate, for each storage time. The samples were kept in transparent glass vials. Before being analysed, the samples stored at 4 °C and 18 °C were placed in a water bath for about 30 min to bring them to room temperature. The measurements were performed by spectrophotometry at a wavelength of 635 nm. In order to assess the stability of the indophenol according to different preservation conditions, the results were normalised against the average absorbance of the solutions prior to storage (t0 ¼20 min). Fig. 3 presents the results. Storage time has been represented using a logarithmic scale. A 10% uncertainty interval was applied to the value recorded from the samples before storage (after a derivatization time of 20 min) in order to take account for the experimental error relating to the use of spectrophotometry. The results revealed no significant loss of indophenol after 24 h storage, regardless of the storage conditions. After 24 h, the indophenol concentration decreased in the samples exposed to daylight (Fig. 2A), by approximately 20%, 30% and 40% after 2, 3, and 4 days' storage respectively. The samples stored in darkness demonstrated a clear improvement in indophenol stability over time (Fig. 2B–D). Unlike the samples stored in daylight, indophenol remained stable for 48 h post-derivatization. Light therefore seems to affect the stability of indophenol in water. For the samples stored in darkness, the kinetics were similar
S. Kinani et al. / Talanta 140 (2015) 189–197
regardless of temperature ( 18 °C, 4 °C and 18 °C). Temperature is therefore not a determining factor for the stability of indophenol in water. The results show that it is possible to mediate the instability of monochloramine by chemically converting it into indophenol, immediately after sampling. Storage time should not exceed 48 h, provided the samples are stored in darkness and at less than 18 °C. Following these tests, all samples were stored in amber glass vials in order to limit exposure to daylight as much as possible. The indophenol stability was also assessed at 1 mg L 1 equivalent MCA by monitoring the residual concentration of monochloramine using the LC–MS/MS method described below. The results are very similar to those obtained above. 3.2. LC–MS/MS analysis 3.2.1. MS/MS transitions Two transitions were chosen to trace the presence of indophenol in Multiple Reaction Monitoring (MRM) mode, both of
193
which are between the pseudo-molecular MH þ ion and the most abundant fragment ions: m/z 200-127 (quantification transition) and m/z 200-154 (confirmation transition). The declustering potential (DP), collision energy (CE) and collision cell exit potential (CXP) values were optimised for each transition and are given in Table 1. The mechanisms suggested to explain the formation of these ions are presented in Fig. 5. The mechanism proposed for the formation of the m/z 154 ion involves consecutive eliminations of water and carbon monoxide, following cyclization of the system to reduce its entropy. There is likely to be an analogue dissociation pathway in which the loss of CO precedes that of H2O, as confirmed by the presence of m/z 182 [MH þ –H2O] and m/z 172 [MH þ –CO] ions in the CID spectrum in Fig. 4. Given the odd m/z ratio and the nitrogen rule, the m/z 127 ion is either an ion with an even number of electrons but without nitrogen atom, or an ion with an odd number of electrons [40]. The first hypothesis is highly unlikely given the chemical structure of indophenol. The second hypothesis is supported by the presence of an m/z 183 ion
Fig. 3. Temporal stability of indophenol in various storage conditions (A. In daylight and at laboratory temperature (18 °C); B. in darkness and at laboratory temperature; C. refrigerated at 4 °C; and D. frozen at 18 °C).
Table 1 MRM transitions and optimised electrospray and CID conditions for the measurement of indophenol and its internal standard thymolindophenol. Compound
R.T.a (min) Precursor ion (m/z) Product ion (m/z) Declustering potential (DP) (V)
Indophenol
7.9
200.1b 200.1
127.0b 154.0
Thymolindophenol (IS) 8.9
219.9b 256.1
219.9b 108.0
a b
R.T.: retention time. Transitions in bold were used for quantitation.
Collision energy (CE) (V)
Collision exit potential (CXP) (V)
76 76
39 37
20 10
71 71
21 29
16 6
194
S. Kinani et al. / Talanta 140 (2015) 189–197
in the CID spectrum in Fig. 4 corresponding to [MH þ –HO]. The elimination of a hydroxyl radical on the aromatic cycle is a mechanism commonly found with high-energy CID [41]. Since the elimination of ethylene is impossible from m/z 175, the transition m/z 155-m/z 127 has been proposed as resulting from the successive losses of acetylene and dihydrogen. 3.2.2. LC–MS/MS method performance 3.2.2.1. Monochloramine identification – selectivity. European Commission Decision 2002/657/EC on the development of analytical methods, recommends following at least two MRM transitions when
identifying an analyte. The directive also requires the use of tolerance criteria for the variation of the relative intensities of the quantification and confirmation ions. For the method being developed here, the tolerance allowed for variation in the relative abundance of the m/z 154 ion compared to that of the m/z 127 ion was 20%, insofar as the intensity of the m/z 154 ion represents more than 50% (56.3%) of that of the m/z 127 ion. The abundance of the m/z 154 ion must therefore be 45.6–67.0% of the m/z 127 ion in order for the indophenol to be identified. The relative tolerance for the retention time (7.03 min) was 2.5%, which therefore needs to be between 6.86 and 7.20 min. The indophenol identification parameters (retention
Fig. 4. Dissociation spectrum for the MH þ ion of indophenol in positive ESI–MS (collision energy 40 V).
Fig. 5. Dissociation pathways suggested for the formation of m/z 154 and m/z 127 ions from protonated indophenol. The added proton is in bold print in the MH þ ion structure.
S. Kinani et al. / Talanta 140 (2015) 189–197
195
time for relative ion abundances) were established using the arithmetic means of the results of 10 successive injections of an indophenol solution obtained using derivatization of river water with MCA at a concentration of 0.5 μg L 1. Solutions with concentrations ranging from 40 to 1000 ng L 1 were injected to assess the stability of the area-under-curve ratios for the m/z 127 and m/z 154 ions based on MCA concentration. The values remained within the tolerance limits (45.6–67.0%) for all concentrations tested. 3.2.2.2. Evaluation of matrix effects. Matrix compounds present in river water may disrupt processes occurring during electrospray ionisation, causing suppression or enhancement of the signal response in LC–MS/MS. Several methods to assess this phenomenon have been reported in the literature [42,43]. In this study, the matrix effect was evaluated by comparing the chromatographic peaks areas of indophenol prepared by dilution with river water with corresponding solutions prepared in ultrapure water at equivalent concentration (0.5 mg L 1 equivalent MCA). The average matrix factor value, calculated as the ratio between the chromatographic peaks areas of indophenol diluted with river water and those measured with indophenol prepared in ultrapure water, was about 0.92, which indicates a minor ionisation suppression. Thymolindophenol was used as internal standard and allowed the effective correction for the losses associated with this phenomenon. 3.2.2.3. Monochloramine quantification. The performance of the developed LC–MS/MS method was assessed in river water following the procedure described in the French standard NF T90-210. The assessment criteria were response function, limit of quantification and correctness. The correctness profiles method (accuracy and precision) was used to test the validity of the method. To do this, the acceptance criteria were first defined, and then tests were conducted to check the validity of the analytical method using the pre-determined criteria. Quantification involved internal calibration with thymolindophenol, added as the internal standard at a concentration of 200 ng L 1. The response function was assessed over a concentration range of 40–1000 ng L 1 equivalent MCA. Five calibration curves, each plotted with seven concentration levels, were produced over five different days. The calibration curves were obtained by plotting the ratio between the area under the chromatography peak for indophenol and that of the internal standard, depending on the theoretical concentration of MCA. Several mathematical functions were adjusted and compared. The results of the comparison showed the linear regression model as being the most suitable for describing the relationship between concentration and analytical response. The correlation coefficients were all higher than 0.995. The linearity of the LC–MS/MS method was proven since the concentrations that were re-calculated using obtained calibrations curves were all within the acceptance range of 720% for 40 ng L 1 and 7 15% for 100–1000 ng L 1. The Limit of Quantification (LOQ) can be defined as the smallest concentration of an analyte that can be quantified to a given degree of accuracy. In the absence of any regulatory requirements as to the accuracy of the limit of quantification for monochloramine, a maximum permitted deviation (MPD) of 760% was used, as recommended by standard NF T90-210. The LOQ was first estimated by LC–MS/MS analysis of a solution of indophenol at a concentration of 100 ng L 1 equivalent monochloramine. Next, five solutions obtained by spiking river water with indophenol at a concentration equivalent to the pre-supposed LOQ were prepared over five different days, and analysed by LC–MS/MS. Each solution was tested twice. The resulting limit of quantification was 40 ng L 1. As an example, Fig. 6 shows the chromatogram obtained from injecting 20 μL of an indophenol solution in river water at 40 ng L 1 equivalent MCA.
Fig. 6. Chromatographic peak of indophenol in river water at 40 ng L 1 equivalent MCA.
Tests were conducted combining the MCA analysis procedure with the SPE sample preparation protocol described in Section 2.2.3. The main advantage of SPE is the purification of the sample, which considerably limits the risks of premature contamination of the mass spectrometer source and reduces ion suppression and other matrix effects. In addition, the SPE protocol that was used reconcentrates the sample thirtyfold (approximately 60% yield) which potentially lowers the method's limit of quantification to a theoretical value of 1.3 ng L 1. However, the purification/preconcentration stage was ruled out for two main reasons. Firstly, the lack of repeatability of the results; secondly, the over-complication and lengthening of the protocol. The limit of quantification obtained without SPE is sufficient given the environmental factors involved, and the noted absence of any negative matrix effects justifies omitting a relatively complicated pre-concentration stage which moreover is a source of contamination and error. The correctness of a measurement method refers to how well the concentrations obtained from a wide range of measurements correlate to the reference concentrations. Correctness takes into account the total errors of the result i.e. systematic errors (precision) as well as random errors (accuracy). Ideally, correctness should be verified using samples with a certified reference value and a maximum permitted deviation (MPD) based on regulatory requirements, official standards or even specifications given by the client or laboratory itself. In the absence of any certified reference material for MCA in water, river water samples spiked with indophenol were used. The correctness study was performed over five separate days for each of the following three concentrations of indophenol: 40, 100 and 500 ng L 1 equivalent MCA. On each day of analysis, each concentration was measured twice. The MPD levels were determined for these three concentrations. The correctness of the LC–MS/MS method was proven since the results were within the chosen acceptance limits, namely 760% for 40 ng L 1 and 720% for 500 and 1000 ng L 1. 3.3. Analysis of river water samples The LC–MS/MS method that was developed was used to analyse water samples taken from the river Loire (France) for research purposes. Samples were collected in duplicate, in amber borosilicate glass vials, as recommended by ISO 5667-3 [44]. The samples were immediately placed in a cool box containing coldpacks to keep the temperature between 4 and 8 °C during transport to the laboratory. The MCA levels measured in each of the different samples are given in Table 2. The results show occasional monochloramine presence in the river Loire. The concentrations measured in the samples did not exceed 300 ng L 1, well below the detection thresholds of current
196
S. Kinani et al. / Talanta 140 (2015) 189–197
Table 2 Monochloramine levels in samples of Loire river water (France). Sampling date Time between sampling and chemical derivatization (min)
MCA concentration (ng L 1)
01/10/2012 02/10/2012 03/10/2012 08/10/2012 12/10/2012 15/10/2012 16/11/2012 07/11/2012 11/06/2013 13/06/2013 25/09/2013
oLQ 48 oLQ 298 54 195 88 oLQ oLQ oLQ oLQ
50 45 90 60 50 55 55 55 0 0 0
LQ (limit of quantitation) ¼ 40 ng/L as monochloramine.
analytical methods (LQ o50 mg L 1 equivalent MCA). Initially, the chemical derivatization was performed in the laboratory for safety reasons. Due to the MCA instability in river water and the time gap between sampling and derivatization, the concentrations have therefore probably been underestimated as regards actual levels in the river. There are two possible hypotheses to explain the origin of the monochloramine found in the river Loire at the sampling point in question. First, monochloramine could have been released into the river by power plants located upstream of the sampling point. However, the samples which contained monochloramine at concentrations higher than the LOQ were taken outside of the monochloramine disinfection treatment application periods (no MCA treatment was applied in October and November 2012 in the power plants situated upstream of the sampling point). The second hypothesis is that the monochloramine was formed in situ by so-called “precursors”. Indeed, one of the elements needed for the formation of monochloramine, ammonium ions, is found in the river Loire at concentrations of up to 0.1 mg L 1. Water effluent containing active chlorine released into the river upstream of the sampling point (e.g. run-off tap water, water from purification plants, running fluids from certain industrial installations, etc.) could trigger the oxidation of the ammonium ions into monochloramine and explain the low amounts of monochloramine found in the river. However, further studies will be needed to pinpoint the origin of the monochloramine.
4. Conclusion This work introduces a new approach to the measurement of monochloramine in surface water. By combining the derivatization of MCA into indophenol and the detection of indophenol by LC–MS/MS in the MRM mode, a limit of quantification of 40 ng L 1 could be achieved, a limit which could be further reduced by including a SPE reconcentration/purification stage. Among all the methods reported to date in the literature, this is the only one sensitive enough to analyse MCA at the trace-levels likely to be found in river water. The applicability of the method was confirmed through the analysis of several water samples from Loire River (France). The monochloramine was found at concentrations ranging from below LOQ to 298 ng L 1. The monochloramine decay at low environmental relevant concentrations can differ significantly from that at higher concentrations. The developed method removes the main obstacle against the study of the environmental fate of monochloramine at low concentrations. Its use will provide new data needed for the establishment of the degradation kinetics of monochloramine at
low concentrations and thus participate to a better assessment of its impact on the environment. The developed method can also be used to measure monochloramine in others types of water, such as tap water, groundwater or wastewater. The method has already been used for our research activities. However, it is currently not suitable for routine use in industrial applications given the complexity of the procedure, the instability of indophenol and the use of certain toxic reagents. Due to MCA instability in river water, the time between sample collection and analysis in the laboratory must be kept to an absolute minimum. Due to the lack of commercially available standard for indophenol, this compound was prepared by chemical derivatization of a titrated solution of monochloramine. In the absence of an internal standard to check the success of the chemical derivatization procedure, this step must be repeated each day of analysis for all samples, including calibration standards, quality control samples and real samples. In addition, the analysis of quality control samples derived at the same time as real samples should precede and follow the analysis of each series of real derivatized samples.
Acknowledgement This research was conducted in partnership between Electricite de France (EDF), Ecole Polytechnique/CNRS (National Center for Scientific Research) and Water Technology Centre (TZW) of applied research of the German Waterworks Association (DVGW). A part of the financial support for this work was provided by EDF Research & Development which the authors are thankful.
References [1] J.L. Kool, J.C. Carpenter, B.S. Fields, Lancet 353 (1999) 272–277. [2] B.R. Kim, J.E. Anderson, S.A. Mueller, W.A. Gaines, A.M. Kendall, Water Res. 36 (2002) 4433–4444. [3] A.M. Kahler, T.L. Cromeans, J.M. Roberts, V.R. Hill, Water Res. 45 (2011) 1745–1751. [4] S.-H. Chien, I. Chowdhury, M.-K. Hsieh, H. Li, D.A. Dzombak, R.D. Vidic, Water Res. 46 (2012) 6508–6518. [5] C.T. Jafvert, R.L. Valentine, Environ. Sci. Technol. 26 (1992) 577–586. [6] C. Tian, R. Liu, T. Guo, H. Liu, Q. Luo, J. Qu, Sep. Purif. Technol. 102 (2013) 86–93. [7] K. Doederer, W. Gernjak, H.S. Weinberg, M.J. Farré, Water Res. 48 (2014) 218–228. [8] H.S. Weinberg, S.W. Krasner, S.D. Richardson, A.D. Thruston, The Occurrence of Disinfection By-Products (DBPs) of Health Concern in Drinking Water: Results of a Nationwide DBP Occurrence Study, EPA/600/R-02/068. 〈http://www.epa. gov/athens/publications/reports/EPA_600_R02_068.pdf〉. [9] S.W. Krasner, H.S. Weinberg, S.D. Richardson, S.J. Pastor, R. Chinn, M.J. Sclimenti, G.D. Onstad, A.D. Thruston, Environ. Sci. Technol. 40 (2006) 7175–7185. [10] X. Yang, C. Fan, C. Shang, Q. Zhao, Water Res. 44 (2010) 2691–2702. [11] Y. Bichsel, U. von Gunten, Environ. Sci. Technol. 34 (2000) 2784–2791. [12] G.H. Hua, D.A. Reckhow, J. Kim, Environ. Sci. Technol. 40 (2006) 3050–3056. [13] S.D. Richardson, F. Fasano, J.J. Ellington, F.G. Crumley, K.M. Buettner, J.J. Evans, B.C. Blount, L.K. Silva, T.J. Waite, G.W. Luther, A.B. Mckague, R.J. Miltner, D.E. Wanger, M.J. Plewa, Environ. Sci. Technol. 42 (2008) 8330–8338. [14] WHO, Guideline for Drinking Water Quality, 2011. [15] J.G. Pressman, W.H. Lee, P.L. Bishop, D.G. Wahman, Water Res. 46 (2012) 882–894. [16] Loi canadienne sur la protection de l'environnement (1991), Inorganic Chloramines, 2001, ISBN: 0-662-0-662-85047-5. [17] S. Kinani, B. Richard, Y. Souissi, S. Bouchonnet, Trends Anal. Chem. 33 (2012) 55–67. [18] W. Lee, P. Westerhoff, X. Yang, C. Shang, Water Res. 41 (2007) 3097–3102. [19] F.E. Scully, A.B. Maxwell, Environ. Health Perspect. 46 (1984) 111–116. [20] M.T. Lukasewycz, C.M. Bieringer, R.J. Liukkonen, M.E. Fitzsimmons, H. F. Corcoran, S. Lin, R.M. Carison, Environ. Sci. Technol. 23 (1989) 196–199. [21] T. Kotiaho, A.K. Lister, M.J. Hayward, R.G. Cooks, Talanta 38 (1991) 195–200. [22] R.A. Ketola, T. Kotiaho, M.E. Cisper, T.M. Allen, J. Mass Spectrom. 37 (2002) 457–476. [23] C. Shang, E.R. BlatchleyIII, Environ. Sci. Technol. 33 (1999) 2218–2223. [24] L.S. Riter, Z. Takáts, R.G. Cooks, Analyst 11 (2001) 1980–1984.
S. Kinani et al. / Talanta 140 (2015) 189–197
[25] ISO 7393-2, Water Quality-Determination of Free Chlorine and Total Chlorine – Part 2: Colorimetric Method Using N,N-diethyl-1,4-phenylenediamine, for Routine Control Purposes, 2000. [26] A. Aminot, D.S. Kirkwood, R. Kérouel, Mar. Chem. 56 (1997) 59–75. [27] Q.P. Li, J.-Z. Zhang, F.J. Millero, D.A. Hansell, Mar. Chem. 96 (2005) 73–85. [28] M. Okamura, K. Fujinaga, Y. Seine, S.J. Honda, Anal. Chem. 365 (1999) 467–469. [29] H. Tao, Z.-L. Chen, X. Li, Y.-L. Yang, G.B. Li, Anal. Chim. Acta 615 (2008) 184–190. [30] D.L. Harp, Water Environ. Res. 72 (2000) 706–713. [31] J. Kanada, Water Res. 29 (1995) 2746–2750. [32] N. Hata, I. Kasahara, S. Taguchi, Anal. Sci. 18 (2002) 697–699. [33] Y. Moliner-Martínez, R. Herráez-Hernández, P. Campíns-Falcó, Anal. Chim. Acta 534 (2005) 327–334. [34] T. Shoji, E.J. Nakamura, Flow Inject. Anal. 26 (2009) 37–41. [35] T. Shoji, E. Nakamura, Anal. Sci. 26 (2010) 779–783. [36] T.-H. Nguyen, L. Mugherli, C. Rivron, T.-H. Tran-Thi, Sens. Actuators B 208 (2015) 622–627.
197
[37] P.J. Rommers, J. Visser, Analyst 94 (1969) 653–658. [38] A. Daridon, M. Sequeira, G. Pennarun-Thomas, H. Dirac, J.P. Krog, P. Gravesen, J. Lichtenberg, D. Diamond, E. Verpoorte, N.F. de Rooij, Sens. Actuators B 76 (2001) 235–243. [39] K.T. Lau, S. Edwards, D. Diamond, Sens. Actuators B 98 (2004) 12–17. [40] F.W. McLafferty, Interpretation of Mass Spectra, 3rd ed., University Science Books, Mill Valley, California (1980) 33. [41] E.M. Thurman, I. Ferrer, O.J. Pozo, J.V. Sancho, F. Herdandez, Rapid Commun. Mass Spectrom. 21 (2007) 3855–3868. [42] A. Kruve, I. Leito, K. Herodes, Anal. Chim. Acta 651 (2009) 75–80. [43] C. Ghosh, C.P. Shinde, B.S. Chakraborty, J. Chromatogr. B 893 (2012) 193–200. [44] IS0 5667-3, Water Quality – Sampling – Part 3: Preservation and Handling of Water Samples, 2012.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Selective and trace determination of monochloramine in river water by chemical derivatization and liquid chromatography/tandem mass spectrometry analysis Said Kinani a,n, Stéphany Layousse a, Bertille Richard a, Aziz Kinani a,b, Stéphane Bouchonnet b,1, Astrid Thoma c, Frank Sacher c,2 a Laboratoire National d'Hydraulique et Environnement (LNHE), Division Recherche et Développement, Electricité de France (EDF), 6 Quai de Watier, 78401 Chatou Cedex 01, France b Laboratoire de Chimie Moléculaire, Ecole Polytechnique, 91128 Palaiseau, France c DVGW – Technologiezentrum Wasser (TZW), Karlsruher Strasse 84, 76139 Karlsruhe, Germany
art ic l e i nf o
a b s t r a c t
Article history: Received 11 December 2014 Received in revised form 16 March 2015 Accepted 22 March 2015 Available online 30 March 2015
Monochloramine (MCA) may enter the aquatic environment through three main sources: wastewater treatment plant effluents, industrial effluents and thermal power plant wastes. Up to date, there are no available data about the concentration levels of this chemical in river water due to lack of appropriate analytical methods. Therefore, sensitive and selective analytical methods for monochloramine analysis in river water are required to evaluate its environmental fate and its effects on aquatic ecosystems. Thus, in this study we describe a highly specific and sensitive method for monochloramine determination in river water. This method combines chemical derivatization of monochloramine into indophenol followed by liquid chromatography coupled to electrospray ionisation–tandem mass spectrometry (LC–ESI–MS/MS) analysis. Two precursor-to-product ion transitions were monitored (200-127 and 200-154) in positive ionisation mode, fulfilling the criteria of selectivity, in accordance with the European Legislation requirements (decision 2002/657/EC). Ion structures and fragmentation mechanisms have been proposed to explain the selected transitions. Linearity range, accuracy and precision of the method have been assessed according to the French method validation standard NF T90-210. Detecting the derivatized monochloramine (indophenol) in Multiple Reaction Monitoring (MRM) mode provided a limit of quantification of 40 ng L 1 equivalent monochloramine. Applied to Loire river water (France), the developed method occasionally detected monochloramine at concentrations less than 300 ng L 1, which could be explained by punctual discharges of water containing active chlorine upstream of the sampling point. Indeed, it is widely reported in the literature that the addition of chlorine to water containing ammonia (e.g., wastewater effluents and river water) may result in the instantaneous formation of monochloramine. The proposed method is a powerful tool that can be used in environmental research (e.g., assessment of environmental fate and generating of ecotoxicological data) as well as in research studies concerning the evaluation of water disinfection efficiency; but it is not currently appropriate for routine use in industrial applications given the complexity of the procedure, the instability of indophenol and the use of certain toxic reagents. & 2015 Elsevier B.V. All rights reserved.
Keywords: Monochloramine River water Chemical derivatization Liquid chromatography/tandem mass spectrometry Ecotoxicology and environmental fate Ecotoxicological risk assessment
1. Introduction
n
Corresponding author. Tel.: þ 33 1 30 87 91 13. E-mail addresses: [email protected] (S. Kinani), [email protected] (S. Bouchonnet), [email protected] (F. Sacher). 1 Tel.: þ33 1 69 33 48 05. 2 Tel.: þ49 721 9678 156. http://dx.doi.org/10.1016/j.talanta.2015.03.043 0039-9140/& 2015 Elsevier B.V. All rights reserved.
The term “inorganic chloramines” refers to a group of three chemical substances formed by a reaction between chlorine and ammonia in an aqueous medium: monochloramine (MCA, NH2Cl), dichloramine (DCA, NHCl2), and trichloramine (TCA, NCl3). Concentration and speciation of these species primarily depend on two parameters: pH value and chlorine to nitrogen ratio (Cl2/N). Only MCA has industrial uses in the water treatment sector, in
190
S. Kinani et al. / Talanta 140 (2015) 189–197
particular as a biocide for inactivating pathogenic microorganisms and viruses to prevent the propagation of water-borne infectious diseases [1–4]. Monochloramine may be also formed as unwanted by-products of water chlorination in the presence of ammonium ions (e.g. in swimming pools or polluted natural water) [5]. Compared to chlorine, monochloramine is less reactive towards the organic matter present in water and therefore produces total organic halogen and regulated disinfection by-products in lower quantities, especially trihalomethanes (THM) and haloacetic acids (AHA) [6,7]. Nevertheless, MCA has been reported to generate higher levels of nitrogenous organic halogenated by-products such as haloacetonitriles, cyanogen halides, and haloacetamides than chlorination [6,8–10]. It has also been reported that disinfection of iodide-containing waters with monochloramine or by chlorination in the presence of ammonia favours the formation of iodinated organic compounds such as iodo-THM and iodo-AHA [11–13]. In water treatment processes, monochloramine as a secondary disinfectant is currently a well-established technology and even recommended by the WHO [14]. It is used in several industrialised countries such as Canada, Japan, USA, the United Kingdom and Australia. According to estimates, around 30% of drinking water plants in the USA use MCA, and a further 8–12% intend to adopt it as a treatment [15]. In France and Germany, monochloramine is currently banned as a disinfectant for water intended for human consumption. It is, however, authorised for the disinfection of water in industrial cooling systems. It is an example used by Electricité de France (EDF) to disinfect the cooling circuits in some of its thermal power plants located alongside rivers. Monochloramine may enter the aquatic environment through three main sources: wastewater treatment plant effluent, industrial effluent and thermal power plant waste. There is therefore a need to investigate whether the aquatic ecosystem is being affected by effluents containing residual MCA. A report on Canadian environmental quality standards for chlorine species was published in 1999 and lists the main ecotoxicological effects of MCA [16]. The lowest acute toxicity values given in this report are between 14 and 82 mg L 1 for fish, and between 16 and 78 mg L 1 for invertebrates. Thus, in order to assess monochloramine impact on the aquatic environment, there is a need for a sensitive and specific analytical method to quantify this compound at trace concentrations in river water. Quantifying trace levels of inorganic chloramines poses a major challenge, since the available analytical methods, including standardised and benchmark methods, are not sensitive and specific enough to detect these compounds in river water. Scientific articles report the use of two types of methods: those involving colorimetric, titrimetric or amperometric analysis, and those involving chromatography and membrane introduction mass spectrometry (MIMS) [17]. DPD (N,N-diethyl-phenylenediamine) colorimetry is currently the most widely used method. This standardised method (EN ISO 7393) is easy to use and suitable for field measurements, but has two main limitations: its lack of MCA specificity and insufficient sensitivity (LOQ 430 mg eq. Cl2 L 1) for the concentrations expected to be found in river water (o1 mg eq. Cl2 L 1) [18]. To overcome the lack of specificity of DPD colorimetry, other colorimetric methods using the Berthelot reaction have been developed. They are based on the principle of measuring the intensity of the indophenol colour produced when monochloramine and phenol react in an alkaline medium and in the presence of a catalyst. Due to the toxic and noxious nature of phenol, several compounds are now being used as alternatives. Although specific, these methods are not sensitive enough to measure trace levels of monochloramine (LOQ 4 50 mg eq. Cl2 L 1). DPD and ferrous ammonium sulphate (FAS) titration is another standardised method routinely used by water processing laboratories. However, it is unsuitable for field use and its performance is
similar to that of DPD colorimetric testing. Amperometric titration with sodium thiosulphate (Na2S2O3) or phenylarsine oxide (PAO, C6H5AsO) is also a widely used standardised method, especially for the in situ quantification of free and combined chlorine. This method's limit of quantification is similar to that of titrimetric and colorimetric methods, but it has better selectivity. Liquid chromatography (LC) and membrane introduction mass spectrometry (MIMS) methods have been developed to overcome the limitations posed by these traditional methods. Very few methods employing liquid chromatography without prior chemical derivatization of the chloramines have been developed. The detection is performed using UV photometry or electrochemical detection, which vastly reduces the sensitivity and selectivity of the measurements. In order to improve the sensitivity and selectivity, two pre-column derivatization agents have been suggested: 5-(dimethylamino) na phthalene-1-sulphonic acid (or dansyl sulphonic acid, DANSO2H) and 2-mercapto-benzothiazole (2-MBTZ). However the reactions involving those two reagents present various limitations in terms of complexity, reaction duration and/or low yield [19,20]. Membrane introduction mass spectrometry (MIMS) was used to analyse inorganic chloramines for the first time by Kotiaho et al. [21]. The principle of this technique is a selective introduction of molecules through a membrane placed between the sample and the ion source of a mass spectrometer. MIMS has the advantage of being selective and compatible with on-line and real-time analyses without any prior sample preparation stage [22]. However, the hydrophobic nature of the membrane promotes the diffusion of apolar molecules and limits that of polar and semi-polar compounds. This discrimination is a major limitation of the method for analysing monochloramine, since the best limit of quantification obtained to date is 0.2 mg L 1 [23,24]. Therefore, due to the lack of sensitive and specific analytical methods, there are no quantified data on the monochloramine concentration levels in river water. Liquid chromatography coupled with mass spectrometry (LC–MS) is a highly sensitive and specific technique for the quantification of polar and semi-polar molecules at trace concentrations. However, up to date, this technique has not been tested as a potential method for monochloramine analysis, probably due to its low molar mass (51.5 g mol 1). We report on the first attempts to develop a measurement method based on a combination of chemical derivatization of monochloramine into indophenol and LC–MS/MS quantitation of the later. This analytical method was assessed in river water according to the French standard NF T90-210 and successfully used to measure the concentration levels of monochloramine in the water of the river Loire (France) within the framework of research activities.
2. Materials and methods 2.1. Reagents and standard solutions 2.1.1. Reagents and solvents All reagents and solvents used to produce and analyse monochloramine were of analytical quality. Acetonitrile (HPLC grade) and potassium iodide were obtained from Scarlau (Spain) and Carl Roth (Karlsruhe, Germany) respectively. Ammonium chloride (purity 4 99.5%), sodium hypochlorite (15% Cl2), sulphuric acid (purity 4 95.0%), N,N-diethyl-p-phenylenediamine (DPD, purity 4 99.0%), lithium citrate tribasic tetrahydrate (purity 4 99.0%) and ammonium formate (purity 4 99.9%) were purchased from Sigma-Aldrich (Steinheim, Germany). Monopotassium phosphate (purity 4 99.5%), disodium hydrogen phosphate dehydrate (purity 4 99.5%), ethylenediaminetetraacetic acid (EDTA, purity 4 99.0%), potassium iodate (purity 4 99.0%),
S. Kinani et al. / Talanta 140 (2015) 189–197
191
sodium nitroferricyanide dehydrate (purity 4 99.0%), phenol (purity 4 99.5%) and thymolindophenol sodium (purity 4 30.0%) were acquired from Merck (Darmstadt, Germany). Demineralised water was used to prepare the reagents and the monochloramine standard solutions. 2.1.2. Preparing the monochloramine standard solutions in demineralised water MCA is unstable in an aqueous solution. There are therefore no commercially-available standard solutions and it must be prepared and calibrated using a reference analysis method. MCA standard solutions were obtained by diluting a 1 g L 1 stock solution of MCA, prepared by reaction between an ammonium chloride solution and sodium hypochlorite. The stock solution was prepared in three steps, using the following protocol. In the first step, 398.5 mg ammonium chloride (NH4Cl) was dissolved in 250 mL demineralised water. 100 mL of this solution (S1) was taken, the pH adjusted to 9.1 using a sodium hydroxide solution (2 M) and then refrigerated at þ4 °C for at least 30 min before use. In the second step, a sodium hypochlorite solution (S2) was prepared by diluting a commercial bleach solution (approximately 15% by mass), titrated beforehand using DPD colorimetry, in ultrapure water. The volume of commercial bleach solution was calculated to ensure a mass ratio of Cl2 to N of 4.8 in the S3 solution. The final volume was adjusted to 100 mL using demineralised water (S2). This solution was then refrigerated at þ4 °C for at least 30 min before use. In the third step, the monochloramine stock solution (S3) was prepared by carefully combining the pre-prepared S1 and S2 solutions. The synthetic monochloramine was titrated using DPD colorimetry as described in standard ISO 73932 [25]. The MCA stock solution can be kept for up to 24 h at þ4 °C, out of direct light. In order to ensure the accuracy of the concentration, DPD titration was used before any liquid chromatographic assay. 2.1.3. Preparing the indophenol standard solutions in river water For a standard solution with a defined concentration of MCA equivalents, a solution with a tenfold higher concentration of monochloramine was chemically derived using the method described in Section 2.2. Then, 1 mL of the derivatization product was added to a 10 mL volumetric flask and the river water to be analysed was added up to the mark. 2.2. Chemical derivatization – the Berthelot reaction 2.2.1. Principle of chemical derivatization reaction For this study, it was decided to use chemical derivatization due to the need to convert monochloramine into a chemical product with a higher molecular weight that could be analysed with high sensitivity using LC–MS/MS. The chemical derivatization explored is based upon the Berthelot reaction, also known as “indophenol method”, which is widely used for the determination of ammonia and ammonium ions in water [26–28]. This reaction is illustrated in Fig. 1. The Berthelot reaction depends on many factors, the most decisive ones being the type of chemical derivatizing agent (phenol-type reagents), temperature, reaction time, pH and catalyst [27,29]. The advantage of chemical derivatization using the Berthelot reaction comes from its specificity to MCA. This is due to the presence of two exchangeable hydrogen atoms in the monochloramine molecule, which is not the case for dichloramine, trichloramine or organic chloramines [29]. Harp [30], Lee et al. [18] and Tao et al. [29] have shown that the chemical derivatization of MCA into indophenol is little affected by the presence of amines and organic chloramines. Due to the toxicity of phenol, several alternative phenolbased reagents have been tested for the determination of ammonia
Fig. 1. Schematic representation of the derivatization reaction of monochloramine by phenol to form indophenol.
and ammonium ions: o-phenylphenol ((1,1′-biphenyl)-2-ol) by Kanda [31], 1,2-dihydroxybenzene by Harp [30], thymol by Hata et al. [32] and Moliner-Martinez et al. [33], salicylic acid by Tao et al. [29] and 1-naphthol by Shoji et al. [34,35]. To our knowledge only phenol, 2-hydroxyphenol and salycilate have been tested for analysing monochloramine [29,30,36]. Despite its toxicity, phenol remains the most commonly used reagent. In the presence of sodium nitroprusside, monochloramine reacts fully with phenol in just a few minutes, compared to other reagents which have much slower reaction times. There has been extensive research into the Berthelot reaction. Since its initial development, several modifications to enhance efficacy have been made. However, optimal operating conditions vary from one study to another, as shown by the comparisons carried out by [26]. 2.2.2. Derivatization procedure A chemical derivatization protocol based on the one described by [30] was used. A sample volume of 20 mL was placed in a 25 mL beaker containing 0.5 g lithium citrate to avoid any precipitation of the alkaline earth hydroxides during the chemical derivatization. The mixture was stirred until the sodium citrate crystals had fully dissolved, then 0.1 mL sodium nitroferricyanide (5 g L 1) and 1.2 mL phenol (100 g L 1), both prepared fresh each day, were added. The pH of the solution was adjusted to a value between 12.0 and 12.2 using a 5 M sodium hydroxide solution. The mixture was transferred to a 25 mL volumetric flask which was then filled up to the mark with deionized water. After 20 min at room temperature (20–25 °C), the reaction product (indophenol) was measured using spectrophotometry at 635 nm wavelength, or by LC–MS/MS after or without solid phase extraction. All samples (about 1 mL) were post-derivatization filtered through a 0.45 mm cellulose acetate membrane syringe filter (Whatman) before LC– MS/MS analysis. 2.2.3. Sample preparation In order to increase the sensitivity of the method, a sample preparation protocol involving solid phase extraction (SPE) after chemical derivatization (post-SPE) was tested. Several SPE cartridges were tested. The best results were obtained with the LiChrolut polymeric cartridge (200 mg, 6 mL, Merck Millipore). SPE was performed with the following parameters: (1) cartridge preconditioned with 5 mL MeOH followed by 5 mL demineralised water; (2) percolation by slowly aspirating 10 mL of the sample at a flow rate of approximately 0.5 mL min 1; (3) vacuum drying for 3 min to eliminate any traces of water; and (4) elution of indophenol with 6 mL MeOH. The eluate was dried by adding a few milligrams of anhydrous sodium sulphate (Na2SO4) and then filtered through a 0.45 mm cellulose acetate membrane syringe filter (Whatman). The filtrate (MeOH) was evaporated to dryness in a gentle current of nitrogen, and then dissolved in 200 mL of a H2O:ACN mixture (90:10, v/v) for LC–MS/MS analysis. The concentration factor is 50.
192
S. Kinani et al. / Talanta 140 (2015) 189–197
2.3. Equipment and analytical methods HPLC–MS analyses were conducted using an Agilent 1200 Infinity chromatography detector (Agilent Technologies, Waldbronn, Germany) coupled to an API 4000 QTrap triple quadrupole mass spectrometer (Applied Biosystems MDS SCIEX, Toronto, Canada). Data were collected and processed using the Analyst software (version 1.6.0, AB Sciex). The chromatographic separation was performed at 40 °C using a Gemini C18 reversed phase column (250 mm 2 mm, particle size 5 μm) from Phenomenex. After various optimisation tests, the mobile phase flow rate and the injection volume were set at 0.4 mL min 1 and 20 μL respectively. The mobile phase used for the chromatographic separation was a binary mixture comprising a mobile phase A (20 mM ammonium formate in water) and a mobile phase B (20 mM ammonium formate in a 90:10, v/v mixture of ACN:H2O), delivered at a constant flow rate. The gradient began with 90% phase A and decreased linearly over 7 min to 0%. The system was kept in isocratic elution mode for 5 min and then brought back to initial conditions in one minute. At the end of each test, the chromatography column was re-equilibrated for 10 min. The API 4000 QTrap is a hybrid mass spectrometer that can operate as a triple quadrupole (TQ) or a linear ion trap (trap); it was used as a triple quadrupole for this study. The Atmospheric Pressure Chemical Ionisation (APCI) and ElectroSpray Ionisation (ESI) modes were individually optimised and compared in both positive and negative ionisation modes. Positive mode ESI provided the best ion yield and was therefore selected for the rest of the study. Ultrapure nitrogen, produced using a laboratory nitrogen generator (Gengaz, France), was used as spray gas, desolvation gas, curtain gas and collision gas. The spray voltage, pressures of the spray gas, desolvation gas and curtain gas were set at 5.5 kV, 80, 60 and 10 psi respectively. The temperatures of the source and the desolvation nitrogen were set at 80 and 500 °C respectively. The Q1 and Q3 quadrupoles had unit-mass resolution. The nitrogen gas collusion induct dissociation (CID) was regulated at a pressure of 4 mTorr. The mass spectrometer was calibrated with infusion of a polypropylene glycol solution (2 mM in ACN) at a flow rate of 10 mL min 1. In this study, spectrophotometry was used for two different purposes: firstly to analyse the commercial bleach solution and the prepared monochloramine solutions, and secondly to monitor the kinetics of the phenol-induced chemical derivatization of the monochloramine. In each case, the samples were analysed using a CADAS 200 spectrophotometer (HACH LANGE Company). The measurements were carried out using containers with a 5 cm optical path length.
3. Results and discussion 3.1. Chemical derivatization 3.1.1. Indophenol formation kinetics The first step of method development and optimisation involved assessing the kinetics of the phenol-induced derivatization of monochloramine. The reproducibility of the results was confirmed by performing measurements from three monochloramine solutions with concentrations between 0.5 and 0.7 mg L 1 equivalent MCA. The measurements were performed by spectrophotometry at a wavelength of 635 nm. For each solution, the results were normalised against the maximum absorbance registered. Fig. 2 shows changes in the normalised absorbance over a 6 h period. The results in Fig. 2 show that the absorbance, typical for indophenol formation, develops gradually before levelling out
Fig. 2. Reaction kinetic of derivatization of monochloramine by phenol at room temperature.
after 15 min, and then remains stable for at least 6 h after the reagents have been added. It can therefore be concluded that, at room temperature, monochloramine is fully derived into indophenol after a 15 min reaction time. This finding is consistent with several other studies [37–39]. A reaction time of 20 min was retained for the study. 3.1.2. Temporal stability of indophenol Monochloramine is an unstable molecule in river water. In order to measure it, an adequate protocol for sample preservation usable directly on site is required. An interesting alternative and a way to overcome the problem of instability is to chemically derivatize monochloramine immediately after sampling to convert it into a stable product. However, special and sometimes restrictive precautions are needed when using hazardous reagents outside the laboratory, in order to ensure the safety of the personnel taking the samples. In the selected approach, the monochloramine was derivatized into indophenol, a product whose temporal stability needed to be verified. Indophenol stability in water has been tested during 9 days. Four preservation conditions were tested: A. In daylight and at laboratory temperature (18 °C); B. in darkness and at laboratory temperature; C. refrigerated at 4 °C; and D. frozen at 18 °C. A monochloramine solution with a concentration between 0.7 and 1.0 mg/L was prepared and chemically derivatized using the protocol described in Section 2.2. For each of the four preservation conditions, the stability of indophenol was verified in triplicate, for each storage time. The samples were kept in transparent glass vials. Before being analysed, the samples stored at 4 °C and 18 °C were placed in a water bath for about 30 min to bring them to room temperature. The measurements were performed by spectrophotometry at a wavelength of 635 nm. In order to assess the stability of the indophenol according to different preservation conditions, the results were normalised against the average absorbance of the solutions prior to storage (t0 ¼20 min). Fig. 3 presents the results. Storage time has been represented using a logarithmic scale. A 10% uncertainty interval was applied to the value recorded from the samples before storage (after a derivatization time of 20 min) in order to take account for the experimental error relating to the use of spectrophotometry. The results revealed no significant loss of indophenol after 24 h storage, regardless of the storage conditions. After 24 h, the indophenol concentration decreased in the samples exposed to daylight (Fig. 2A), by approximately 20%, 30% and 40% after 2, 3, and 4 days' storage respectively. The samples stored in darkness demonstrated a clear improvement in indophenol stability over time (Fig. 2B–D). Unlike the samples stored in daylight, indophenol remained stable for 48 h post-derivatization. Light therefore seems to affect the stability of indophenol in water. For the samples stored in darkness, the kinetics were similar
S. Kinani et al. / Talanta 140 (2015) 189–197
regardless of temperature ( 18 °C, 4 °C and 18 °C). Temperature is therefore not a determining factor for the stability of indophenol in water. The results show that it is possible to mediate the instability of monochloramine by chemically converting it into indophenol, immediately after sampling. Storage time should not exceed 48 h, provided the samples are stored in darkness and at less than 18 °C. Following these tests, all samples were stored in amber glass vials in order to limit exposure to daylight as much as possible. The indophenol stability was also assessed at 1 mg L 1 equivalent MCA by monitoring the residual concentration of monochloramine using the LC–MS/MS method described below. The results are very similar to those obtained above. 3.2. LC–MS/MS analysis 3.2.1. MS/MS transitions Two transitions were chosen to trace the presence of indophenol in Multiple Reaction Monitoring (MRM) mode, both of
193
which are between the pseudo-molecular MH þ ion and the most abundant fragment ions: m/z 200-127 (quantification transition) and m/z 200-154 (confirmation transition). The declustering potential (DP), collision energy (CE) and collision cell exit potential (CXP) values were optimised for each transition and are given in Table 1. The mechanisms suggested to explain the formation of these ions are presented in Fig. 5. The mechanism proposed for the formation of the m/z 154 ion involves consecutive eliminations of water and carbon monoxide, following cyclization of the system to reduce its entropy. There is likely to be an analogue dissociation pathway in which the loss of CO precedes that of H2O, as confirmed by the presence of m/z 182 [MH þ –H2O] and m/z 172 [MH þ –CO] ions in the CID spectrum in Fig. 4. Given the odd m/z ratio and the nitrogen rule, the m/z 127 ion is either an ion with an even number of electrons but without nitrogen atom, or an ion with an odd number of electrons [40]. The first hypothesis is highly unlikely given the chemical structure of indophenol. The second hypothesis is supported by the presence of an m/z 183 ion
Fig. 3. Temporal stability of indophenol in various storage conditions (A. In daylight and at laboratory temperature (18 °C); B. in darkness and at laboratory temperature; C. refrigerated at 4 °C; and D. frozen at 18 °C).
Table 1 MRM transitions and optimised electrospray and CID conditions for the measurement of indophenol and its internal standard thymolindophenol. Compound
R.T.a (min) Precursor ion (m/z) Product ion (m/z) Declustering potential (DP) (V)
Indophenol
7.9
200.1b 200.1
127.0b 154.0
Thymolindophenol (IS) 8.9
219.9b 256.1
219.9b 108.0
a b
R.T.: retention time. Transitions in bold were used for quantitation.
Collision energy (CE) (V)
Collision exit potential (CXP) (V)
76 76
39 37
20 10
71 71
21 29
16 6
194
S. Kinani et al. / Talanta 140 (2015) 189–197
in the CID spectrum in Fig. 4 corresponding to [MH þ –HO]. The elimination of a hydroxyl radical on the aromatic cycle is a mechanism commonly found with high-energy CID [41]. Since the elimination of ethylene is impossible from m/z 175, the transition m/z 155-m/z 127 has been proposed as resulting from the successive losses of acetylene and dihydrogen. 3.2.2. LC–MS/MS method performance 3.2.2.1. Monochloramine identification – selectivity. European Commission Decision 2002/657/EC on the development of analytical methods, recommends following at least two MRM transitions when
identifying an analyte. The directive also requires the use of tolerance criteria for the variation of the relative intensities of the quantification and confirmation ions. For the method being developed here, the tolerance allowed for variation in the relative abundance of the m/z 154 ion compared to that of the m/z 127 ion was 20%, insofar as the intensity of the m/z 154 ion represents more than 50% (56.3%) of that of the m/z 127 ion. The abundance of the m/z 154 ion must therefore be 45.6–67.0% of the m/z 127 ion in order for the indophenol to be identified. The relative tolerance for the retention time (7.03 min) was 2.5%, which therefore needs to be between 6.86 and 7.20 min. The indophenol identification parameters (retention
Fig. 4. Dissociation spectrum for the MH þ ion of indophenol in positive ESI–MS (collision energy 40 V).
Fig. 5. Dissociation pathways suggested for the formation of m/z 154 and m/z 127 ions from protonated indophenol. The added proton is in bold print in the MH þ ion structure.
S. Kinani et al. / Talanta 140 (2015) 189–197
195
time for relative ion abundances) were established using the arithmetic means of the results of 10 successive injections of an indophenol solution obtained using derivatization of river water with MCA at a concentration of 0.5 μg L 1. Solutions with concentrations ranging from 40 to 1000 ng L 1 were injected to assess the stability of the area-under-curve ratios for the m/z 127 and m/z 154 ions based on MCA concentration. The values remained within the tolerance limits (45.6–67.0%) for all concentrations tested. 3.2.2.2. Evaluation of matrix effects. Matrix compounds present in river water may disrupt processes occurring during electrospray ionisation, causing suppression or enhancement of the signal response in LC–MS/MS. Several methods to assess this phenomenon have been reported in the literature [42,43]. In this study, the matrix effect was evaluated by comparing the chromatographic peaks areas of indophenol prepared by dilution with river water with corresponding solutions prepared in ultrapure water at equivalent concentration (0.5 mg L 1 equivalent MCA). The average matrix factor value, calculated as the ratio between the chromatographic peaks areas of indophenol diluted with river water and those measured with indophenol prepared in ultrapure water, was about 0.92, which indicates a minor ionisation suppression. Thymolindophenol was used as internal standard and allowed the effective correction for the losses associated with this phenomenon. 3.2.2.3. Monochloramine quantification. The performance of the developed LC–MS/MS method was assessed in river water following the procedure described in the French standard NF T90-210. The assessment criteria were response function, limit of quantification and correctness. The correctness profiles method (accuracy and precision) was used to test the validity of the method. To do this, the acceptance criteria were first defined, and then tests were conducted to check the validity of the analytical method using the pre-determined criteria. Quantification involved internal calibration with thymolindophenol, added as the internal standard at a concentration of 200 ng L 1. The response function was assessed over a concentration range of 40–1000 ng L 1 equivalent MCA. Five calibration curves, each plotted with seven concentration levels, were produced over five different days. The calibration curves were obtained by plotting the ratio between the area under the chromatography peak for indophenol and that of the internal standard, depending on the theoretical concentration of MCA. Several mathematical functions were adjusted and compared. The results of the comparison showed the linear regression model as being the most suitable for describing the relationship between concentration and analytical response. The correlation coefficients were all higher than 0.995. The linearity of the LC–MS/MS method was proven since the concentrations that were re-calculated using obtained calibrations curves were all within the acceptance range of 720% for 40 ng L 1 and 7 15% for 100–1000 ng L 1. The Limit of Quantification (LOQ) can be defined as the smallest concentration of an analyte that can be quantified to a given degree of accuracy. In the absence of any regulatory requirements as to the accuracy of the limit of quantification for monochloramine, a maximum permitted deviation (MPD) of 760% was used, as recommended by standard NF T90-210. The LOQ was first estimated by LC–MS/MS analysis of a solution of indophenol at a concentration of 100 ng L 1 equivalent monochloramine. Next, five solutions obtained by spiking river water with indophenol at a concentration equivalent to the pre-supposed LOQ were prepared over five different days, and analysed by LC–MS/MS. Each solution was tested twice. The resulting limit of quantification was 40 ng L 1. As an example, Fig. 6 shows the chromatogram obtained from injecting 20 μL of an indophenol solution in river water at 40 ng L 1 equivalent MCA.
Fig. 6. Chromatographic peak of indophenol in river water at 40 ng L 1 equivalent MCA.
Tests were conducted combining the MCA analysis procedure with the SPE sample preparation protocol described in Section 2.2.3. The main advantage of SPE is the purification of the sample, which considerably limits the risks of premature contamination of the mass spectrometer source and reduces ion suppression and other matrix effects. In addition, the SPE protocol that was used reconcentrates the sample thirtyfold (approximately 60% yield) which potentially lowers the method's limit of quantification to a theoretical value of 1.3 ng L 1. However, the purification/preconcentration stage was ruled out for two main reasons. Firstly, the lack of repeatability of the results; secondly, the over-complication and lengthening of the protocol. The limit of quantification obtained without SPE is sufficient given the environmental factors involved, and the noted absence of any negative matrix effects justifies omitting a relatively complicated pre-concentration stage which moreover is a source of contamination and error. The correctness of a measurement method refers to how well the concentrations obtained from a wide range of measurements correlate to the reference concentrations. Correctness takes into account the total errors of the result i.e. systematic errors (precision) as well as random errors (accuracy). Ideally, correctness should be verified using samples with a certified reference value and a maximum permitted deviation (MPD) based on regulatory requirements, official standards or even specifications given by the client or laboratory itself. In the absence of any certified reference material for MCA in water, river water samples spiked with indophenol were used. The correctness study was performed over five separate days for each of the following three concentrations of indophenol: 40, 100 and 500 ng L 1 equivalent MCA. On each day of analysis, each concentration was measured twice. The MPD levels were determined for these three concentrations. The correctness of the LC–MS/MS method was proven since the results were within the chosen acceptance limits, namely 760% for 40 ng L 1 and 720% for 500 and 1000 ng L 1. 3.3. Analysis of river water samples The LC–MS/MS method that was developed was used to analyse water samples taken from the river Loire (France) for research purposes. Samples were collected in duplicate, in amber borosilicate glass vials, as recommended by ISO 5667-3 [44]. The samples were immediately placed in a cool box containing coldpacks to keep the temperature between 4 and 8 °C during transport to the laboratory. The MCA levels measured in each of the different samples are given in Table 2. The results show occasional monochloramine presence in the river Loire. The concentrations measured in the samples did not exceed 300 ng L 1, well below the detection thresholds of current
196
S. Kinani et al. / Talanta 140 (2015) 189–197
Table 2 Monochloramine levels in samples of Loire river water (France). Sampling date Time between sampling and chemical derivatization (min)
MCA concentration (ng L 1)
01/10/2012 02/10/2012 03/10/2012 08/10/2012 12/10/2012 15/10/2012 16/11/2012 07/11/2012 11/06/2013 13/06/2013 25/09/2013
oLQ 48 oLQ 298 54 195 88 oLQ oLQ oLQ oLQ
50 45 90 60 50 55 55 55 0 0 0
LQ (limit of quantitation) ¼ 40 ng/L as monochloramine.
analytical methods (LQ o50 mg L 1 equivalent MCA). Initially, the chemical derivatization was performed in the laboratory for safety reasons. Due to the MCA instability in river water and the time gap between sampling and derivatization, the concentrations have therefore probably been underestimated as regards actual levels in the river. There are two possible hypotheses to explain the origin of the monochloramine found in the river Loire at the sampling point in question. First, monochloramine could have been released into the river by power plants located upstream of the sampling point. However, the samples which contained monochloramine at concentrations higher than the LOQ were taken outside of the monochloramine disinfection treatment application periods (no MCA treatment was applied in October and November 2012 in the power plants situated upstream of the sampling point). The second hypothesis is that the monochloramine was formed in situ by so-called “precursors”. Indeed, one of the elements needed for the formation of monochloramine, ammonium ions, is found in the river Loire at concentrations of up to 0.1 mg L 1. Water effluent containing active chlorine released into the river upstream of the sampling point (e.g. run-off tap water, water from purification plants, running fluids from certain industrial installations, etc.) could trigger the oxidation of the ammonium ions into monochloramine and explain the low amounts of monochloramine found in the river. However, further studies will be needed to pinpoint the origin of the monochloramine.
4. Conclusion This work introduces a new approach to the measurement of monochloramine in surface water. By combining the derivatization of MCA into indophenol and the detection of indophenol by LC–MS/MS in the MRM mode, a limit of quantification of 40 ng L 1 could be achieved, a limit which could be further reduced by including a SPE reconcentration/purification stage. Among all the methods reported to date in the literature, this is the only one sensitive enough to analyse MCA at the trace-levels likely to be found in river water. The applicability of the method was confirmed through the analysis of several water samples from Loire River (France). The monochloramine was found at concentrations ranging from below LOQ to 298 ng L 1. The monochloramine decay at low environmental relevant concentrations can differ significantly from that at higher concentrations. The developed method removes the main obstacle against the study of the environmental fate of monochloramine at low concentrations. Its use will provide new data needed for the establishment of the degradation kinetics of monochloramine at
low concentrations and thus participate to a better assessment of its impact on the environment. The developed method can also be used to measure monochloramine in others types of water, such as tap water, groundwater or wastewater. The method has already been used for our research activities. However, it is currently not suitable for routine use in industrial applications given the complexity of the procedure, the instability of indophenol and the use of certain toxic reagents. Due to MCA instability in river water, the time between sample collection and analysis in the laboratory must be kept to an absolute minimum. Due to the lack of commercially available standard for indophenol, this compound was prepared by chemical derivatization of a titrated solution of monochloramine. In the absence of an internal standard to check the success of the chemical derivatization procedure, this step must be repeated each day of analysis for all samples, including calibration standards, quality control samples and real samples. In addition, the analysis of quality control samples derived at the same time as real samples should precede and follow the analysis of each series of real derivatized samples.
Acknowledgement This research was conducted in partnership between Electricite de France (EDF), Ecole Polytechnique/CNRS (National Center for Scientific Research) and Water Technology Centre (TZW) of applied research of the German Waterworks Association (DVGW). A part of the financial support for this work was provided by EDF Research & Development which the authors are thankful.
References [1] J.L. Kool, J.C. Carpenter, B.S. Fields, Lancet 353 (1999) 272–277. [2] B.R. Kim, J.E. Anderson, S.A. Mueller, W.A. Gaines, A.M. Kendall, Water Res. 36 (2002) 4433–4444. [3] A.M. Kahler, T.L. Cromeans, J.M. Roberts, V.R. Hill, Water Res. 45 (2011) 1745–1751. [4] S.-H. Chien, I. Chowdhury, M.-K. Hsieh, H. Li, D.A. Dzombak, R.D. Vidic, Water Res. 46 (2012) 6508–6518. [5] C.T. Jafvert, R.L. Valentine, Environ. Sci. Technol. 26 (1992) 577–586. [6] C. Tian, R. Liu, T. Guo, H. Liu, Q. Luo, J. Qu, Sep. Purif. Technol. 102 (2013) 86–93. [7] K. Doederer, W. Gernjak, H.S. Weinberg, M.J. Farré, Water Res. 48 (2014) 218–228. [8] H.S. Weinberg, S.W. Krasner, S.D. Richardson, A.D. Thruston, The Occurrence of Disinfection By-Products (DBPs) of Health Concern in Drinking Water: Results of a Nationwide DBP Occurrence Study, EPA/600/R-02/068. 〈http://www.epa. gov/athens/publications/reports/EPA_600_R02_068.pdf〉. [9] S.W. Krasner, H.S. Weinberg, S.D. Richardson, S.J. Pastor, R. Chinn, M.J. Sclimenti, G.D. Onstad, A.D. Thruston, Environ. Sci. Technol. 40 (2006) 7175–7185. [10] X. Yang, C. Fan, C. Shang, Q. Zhao, Water Res. 44 (2010) 2691–2702. [11] Y. Bichsel, U. von Gunten, Environ. Sci. Technol. 34 (2000) 2784–2791. [12] G.H. Hua, D.A. Reckhow, J. Kim, Environ. Sci. Technol. 40 (2006) 3050–3056. [13] S.D. Richardson, F. Fasano, J.J. Ellington, F.G. Crumley, K.M. Buettner, J.J. Evans, B.C. Blount, L.K. Silva, T.J. Waite, G.W. Luther, A.B. Mckague, R.J. Miltner, D.E. Wanger, M.J. Plewa, Environ. Sci. Technol. 42 (2008) 8330–8338. [14] WHO, Guideline for Drinking Water Quality, 2011. [15] J.G. Pressman, W.H. Lee, P.L. Bishop, D.G. Wahman, Water Res. 46 (2012) 882–894. [16] Loi canadienne sur la protection de l'environnement (1991), Inorganic Chloramines, 2001, ISBN: 0-662-0-662-85047-5. [17] S. Kinani, B. Richard, Y. Souissi, S. Bouchonnet, Trends Anal. Chem. 33 (2012) 55–67. [18] W. Lee, P. Westerhoff, X. Yang, C. Shang, Water Res. 41 (2007) 3097–3102. [19] F.E. Scully, A.B. Maxwell, Environ. Health Perspect. 46 (1984) 111–116. [20] M.T. Lukasewycz, C.M. Bieringer, R.J. Liukkonen, M.E. Fitzsimmons, H. F. Corcoran, S. Lin, R.M. Carison, Environ. Sci. Technol. 23 (1989) 196–199. [21] T. Kotiaho, A.K. Lister, M.J. Hayward, R.G. Cooks, Talanta 38 (1991) 195–200. [22] R.A. Ketola, T. Kotiaho, M.E. Cisper, T.M. Allen, J. Mass Spectrom. 37 (2002) 457–476. [23] C. Shang, E.R. BlatchleyIII, Environ. Sci. Technol. 33 (1999) 2218–2223. [24] L.S. Riter, Z. Takáts, R.G. Cooks, Analyst 11 (2001) 1980–1984.
S. Kinani et al. / Talanta 140 (2015) 189–197
[25] ISO 7393-2, Water Quality-Determination of Free Chlorine and Total Chlorine – Part 2: Colorimetric Method Using N,N-diethyl-1,4-phenylenediamine, for Routine Control Purposes, 2000. [26] A. Aminot, D.S. Kirkwood, R. Kérouel, Mar. Chem. 56 (1997) 59–75. [27] Q.P. Li, J.-Z. Zhang, F.J. Millero, D.A. Hansell, Mar. Chem. 96 (2005) 73–85. [28] M. Okamura, K. Fujinaga, Y. Seine, S.J. Honda, Anal. Chem. 365 (1999) 467–469. [29] H. Tao, Z.-L. Chen, X. Li, Y.-L. Yang, G.B. Li, Anal. Chim. Acta 615 (2008) 184–190. [30] D.L. Harp, Water Environ. Res. 72 (2000) 706–713. [31] J. Kanada, Water Res. 29 (1995) 2746–2750. [32] N. Hata, I. Kasahara, S. Taguchi, Anal. Sci. 18 (2002) 697–699. [33] Y. Moliner-Martínez, R. Herráez-Hernández, P. Campíns-Falcó, Anal. Chim. Acta 534 (2005) 327–334. [34] T. Shoji, E.J. Nakamura, Flow Inject. Anal. 26 (2009) 37–41. [35] T. Shoji, E. Nakamura, Anal. Sci. 26 (2010) 779–783. [36] T.-H. Nguyen, L. Mugherli, C. Rivron, T.-H. Tran-Thi, Sens. Actuators B 208 (2015) 622–627.
197
[37] P.J. Rommers, J. Visser, Analyst 94 (1969) 653–658. [38] A. Daridon, M. Sequeira, G. Pennarun-Thomas, H. Dirac, J.P. Krog, P. Gravesen, J. Lichtenberg, D. Diamond, E. Verpoorte, N.F. de Rooij, Sens. Actuators B 76 (2001) 235–243. [39] K.T. Lau, S. Edwards, D. Diamond, Sens. Actuators B 98 (2004) 12–17. [40] F.W. McLafferty, Interpretation of Mass Spectra, 3rd ed., University Science Books, Mill Valley, California (1980) 33. [41] E.M. Thurman, I. Ferrer, O.J. Pozo, J.V. Sancho, F. Herdandez, Rapid Commun. Mass Spectrom. 21 (2007) 3855–3868. [42] A. Kruve, I. Leito, K. Herodes, Anal. Chim. Acta 651 (2009) 75–80. [43] C. Ghosh, C.P. Shinde, B.S. Chakraborty, J. Chromatogr. B 893 (2012) 193–200. [44] IS0 5667-3, Water Quality – Sampling – Part 3: Preservation and Handling of Water Samples, 2012.
