Вы находитесь на странице: 1из 7

Available online at www.sciencedirect.


Journal of Chromatography A, 1174 (2007) 1319

Determination of nitrate esters in water samples Comparison of efciency of solid-phase extraction and solid-phase microextraction
adal b , Ale s Eisner a , Petra Bajerov a a, , Karel Ventura a V era Je zov a a , Jan Skl

University of Pardubice, Faculty of Chemical Technology, Department of Analytical Chemistry, 532 10 Pardubice, Czech Republic b Research Institute of Industrial Chemistry, Explosia a.s., 530 50 Pardubice Semt n, Czech Republic Available online 30 August 2007

Abstract This paper deals with comparison of efciency of extraction techniques (solid-phase extraction, SPE and solid-phase microextraction, SPME) used for extraction of nitrate esters (ethyleneglycoldinitrate, EGDN and nitroglycerin, NG), representing the rst step of the method of quantitative determination of trace concentrations of nitrate esters in water samples. EGDN and NG are subsequently determined by means of high-performance liquid chromatography with ultraviolet detection (HPLC-UV). Optimization of SPE and SPME conditions was carried out using model water samples. Seven SPE cartridges were tested and the conditions were optimized (type of sorbent, type and volume of solvent to be used as eluent). For both nitrate esters the limit of detection (LOD) and the limit of quantication (LOQ) obtained using SPE/HPLC-UV were 0.23 g mL1 and 0.70 g mL1 , respectively. Optimization of SPME conditions: type of SPME bre (four bres were tested), type and time of sorption/desorption, temperature of sorption. PDMS/DVB (polydimethylsiloxane/divinylbenzene) bre coating proved to be suitable for extraction of EGDN and NG. For this bre the LOD and the LOQ for both nitrate esters were 0.16 g mL1 and 0.50 g mL1 , respectively. Optimized methods SPE/HPLC-UV and SPME/HPLC-UV were then used for quantitative determination of nitrate esters content in real water samples from the production of EGDN and NG. 2007 Elsevier B.V. All rights reserved.
Keywords: Nitrate ester; EGDN; NG; SPE; SPME; HPLC

1. Introduction Contamination of water and soil is a global environmental problem and it is important to nd analytical techniques to cope with this challenge. When contaminants are present in natural samples, their concentration is usually limited to trace levels. Therefore, as the rst step of their quantitative determination the extraction technique is used, capable of isolating the analyte and removing matrix interferences. Widely used is the combination of modern extraction techniques with chromatographic systems. Nitrate esters have been manufactured, stored, tested and used worldwide over the past centuries. This means that nitrate esters entered the environment and became soil and water contaminants. The Czech Republic, as nitrate esters producer, has to monitor the quality of the waters leaving current or former nitrate esters production plants or the places of military training areas

Corresponding author. Tel.: +420 466037088; fax: +420 466037068. E-mail address: petra.bajerova@upce.cz (P. Bajerov a).

and ammunition stores. Also, there exists a risk of contamination of surface or underground waters. Ethyleneglycoldinitrate (EGDN) and nitroglycerin (NG) are typical components of various explosive materials and represent one group of environmental contaminants. EGDN is transparent, colourless, liquid explosive, less sensitive than NG, used mainly in mixtures with NG for low-freezing dynamites. NG is used in the production of commercial explosives as e.g. dynamites, and as the component of multi-base propellants. In low concentrations, NG functions as a drug to cure cardiovascular diseases. Well skin absorbable NG is however toxic. It is colourless, odourless viscous liquid explosive, very sensitive to manipulation. EGDN and NG are relatively soluble in water (EGDN more than NG). Toxic and environmental effects of EGDN are similar to those of NG. However, because of higher EGDN vapour pressure, the exposure to EGDN results in more acute symptoms and clinical manifestations [1]. Extraction methods SPE and SPME are powerful techniques, widely used in analysing various food, biological and environmental matrices. A few references and publications were found

0021-9673/$ see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.08.053


V. Je zov a et al. / J. Chromatogr. A 1174 (2007) 1319

concerning the use of these methods for extraction of explosive compounds, namely nitramines [28] and aromatic nitrocompounds [26,8,9]. Only, a minimal number of publications exists, dealing with utilization of these extraction methods for determination of nitrate esters [1012]. SPE is a very simple way how to extract the molecules of interest from complex aqueous matrices. The results are comparable with those of classical extraction as e.g. liquidliquid extraction [2]. SPE has been successfully applied to the extraction of some types of nitramines and aromatic nitrocompounds in seawater, river water or well water samples [4,6]. SPME is extraction technique for analysing liquid, solid and gas samples. Basic component of SPME system is the silica bre coated with polymer sorbent (stationary phase), placed in SPME holder. Several studies describe the coupling of SPME with gas chromatography (GC) for determination of some nitramines and aromatic nitrocompounds [3,9,10,13]. Utilization of SPME/GC coupling is more common than SPME/HPLC, as the latter one requires the use of SPME/HPLC adapter. SPME/GC determination of nitrate esters is, however, problematic because of thermal instability and sorption in the analytical system [14,15]. SPME/HPLC determination of nitramines and aromatic nitrocompounds in environmental water samples (four seawater and two groundwater) was described in [4]. Three SPME bres were tested for isolation of some nitramines and aromatic nitrocompounds from water. The results were compared with SPE/HPLC as SPE is used in routine determinations of explosive compounds in water. It was found that if compared with SPE, the SPME detection limits were ten times higher. This paper describes utilization of SPE/HPLC-UV and SPME/HPLC-UV for quantitative determination of trace concentrations of nitrate esters content in water samples. The efciency of SPE and SPME was compared. The aim of these experiments was to compare the efciency of extraction techniques at the analysis of water samples contaminated with EGDN and NG. Optimization of various extraction parameters is necessary to achieve the best extraction recoveries. 2. Experimental 2.1. Chemicals and reagents

2.2.2. Real water samples Real water samples used were from the production of nitrate esters (Explosia, Pardubice, Czech Republic). The wastewaters leaving production process enter the reactor to be pre-treated. Substantial decrease in nitrate esters content is achieved by means of continuous reduction with iron (steel microballs) in neutral medium [16]. This way pre-treated waters are discharged to biological wastewater treatment plant. Real water samples were taken on the outlet of pre-treatment reactor in three various time intervals (T1, T2, T3). 2.3. SPE procedure Various polymeric sorbents and solvents were tested to compare their extraction properties in the isolation process of EGDN and NG from water samples. SPE procedure was tested using various types of cartridges and several sorbents: Supelco Discovery DSC-18LT-1000 mg/6 mL, monomer octadecyl, Supelco DSC-C8-500 mg/3 mL, octyl monomer, and Supelco DSC-PH500 mg/3 mL, phenyl monomer (Supelco Bellefonte, PA, USA); three cartridges Phenomenex (Torrance, CA, USA) Strata X (500 mg/6 mL, surface-modied styrene-divinylbenzene polymer), Strata XC (500 mg/6 mL, 33 m caption mixed polymer) and Strata C18 E (500 mg/3 mL, octadecyl). The last cartridge tested was Waters Oasis HLB (60 mg/3 mL, divinylbenzene and N-vinylpyrolidone) (Waters, Milford, MA, USA). SPE process was conducted on Visiprep SPE Vacuum manifold system Supelco (Bellefonte, PA, USA). The general sample preparation scheme for SPE is: conditioning of the SPE columns, followed by sample loading and nally elution of the analytes off the solid-phase columns. SPE cartridges were conditioned immediately prior to each extraction experiment by means of successive washing with methanol and de-ionized water. 200 mL aliquot of model water sample, spiked with 3 g mL1 of each nitrate ester, was transferred to SPE cartridge for extraction. Nitrate esters were eluated from the SPE cartridge with solvent (methanol, acetonitrile or acetone). The eluate was concentrated under a stream of nitrogen to the nal volume of 1 mL with 20 L of which being injected into HPLC system. 2.4. SPME procedure

Nitrate esters standards (Supelco, Bellefonte, PA, USA) were dissolved in methanol. All solvents (methanol, acetonitrile, acetone) were from Riedel-de Ha en, SigmaAldrich (Seelze, Germany). De-ionized water used for model samples was prepared with Demiwa 5-roi purication system (Watek, Lede c nad S azavou, Czech Republic) and by Ultra Clear UV system (SG Water Nashua, NH, USA). 2.2. Water Samples 2.2.1. Model water samples Model water samples were prepared from de-ionized and drinking water. EGDN and NG were spiked into 200 mL of deionized or drinking water in the concentration 3 g mL1 of each compound.

Solid phase microextraction was used as a reference method for evaluation of the results obtained by SPE in terms of quantication limits and accuracy. The manual SPME holder used and bres, coated with various polymers: polydimethylsiloxane/divinylbenzene (PDMS/DVB, 65 m lm thickness), Carbowax/divinylbenzene (CW/DVB, 65 m lm thickness), Carboxene/polydimethylsiloxane (CAR/PDMS, 75 m lm thickness), Carbowax/templated resin (CW/TPR, 50 m lm thickness), as well as SPME/HPLC interface desorption reservoir were from Supelco, Bellefonte, PA, USA. Before its rst use, each bre was conditioned in mobile phase for 30 min. SPME process was carried out in 8 mL glass vials with magnetic stirring (500 rpm), using aliquots of water samples (4 mL) containing 20 g mL1 of both nitrate esters and addi-

V. Je zov a et al. / J. Chromatogr. A 1174 (2007) 1319


tions of water solution (10%) of sodium chloride. Nitrate esters were desorbed in desorption chamber of SPME/HPLC interface. To minimize the possibility of carry-over effect the bre was washed with methanol for 5 min before the next extraction. 2.5. HPLC-UV system HPLC analyses were carried out using chromatographic system with UV-detector (Ecom, Prague, Czech Republic). Separation of nitrate esters was performed in a column LiChrospher 100 RP-18e, Merck (Darmstadt, Germany). Security Guard C18 Phenomenex was used as well. Isocratic mobile phase acetonitrile:water (60:40, v/v) was used for separation. The analyses were conducted at ow rate of 0.5 mL min1 and column temperature was kept at 25 C. Twenty microlitres sample was injected through six-port valve into chromatographic column and nitrate esters eluted within 10 min. UV-detector wavelength was set at 210 nm. Quantication was carried out using calibration curve and peak area measurements. 2.6. Method validation Calibration curves were created ranging from 0.7 to 7 g mL1 (0.7, 1, 3, 5 and 7) for EGDN and NG in case of SPE extraction and 0.5 to 7 g mL1 (0.5, 1, 3, 5 and 7) in case of SPME extraction, respectively. The calibration curves consisted of ve different concentrations of analyte per 1 mL of water. Each concentration point was determined in ve replicates. Curves were prepared by adding varying concentrations of analytes to water. The precision and accuracy of the method were evaluated intra- and inter-day by analysis of six replicates of three concentrations, including the limit of quantication (LOQ) as the lowest point of calibration curve, middle and highest point. LOQ was set as the lowest amount of an analyte in a sample that can be determined quantitatively with suitable precision and accuracy. The accuracy of the method was determined as percent error, while precision was evaluated by intra- and inter-day relative standard deviation. All parameters were in acceptable range of 15%. 3. Results and discussion 3.1. SPE optimization 3.1.1. Extraction recoveries with various cartridges Selection of proper SPE cartridge with the most suitable sorbent material plays an important role in achieving high and reproducible recovery of contaminants. Several types of SPE sorbents from various producers were tested for extraction of nitrate esters from water samples. Methanol was used as elution solvent. The best extraction for both compounds was obtained with Waters Oasis HLB and Strata X cartridges (Fig. 1). The other cartridges produced high NG extraction recoveries and only low EGDN yields. Out of all SPE cartridges tested, Waters Oasis HLB copolymer cartridges were chosen to be used for further testing.

Fig. 1. Extraction recoveries obtained with tested SPE cartridges.

3.1.2. Elution with different solvents The recovery of organic compounds by SPE is highly dependent on the polarity of eluents. Acetone, acetonitrile and methanol as eluents were tested for the elution of EGDN and NG from Waters Oasis HLB and Strata X cartridges. Comparison of these two SPE cartridges and elution with all solvents is shown in Fig. 2. The results show the best extraction recoveries for Waters Oasis HLB cartridges for all solvents and for both nitrate esters. Acetone and methanol produced almost identical recoveries (between 70% and 86% for EGDN and 75% and 89% for NG). The best elution recoveries were obtained with acetonitrile (90% for EGDN and 92% for NG). Acetonitrile as elution solvent in combination with Waters Oasis HLB cartridge was chosen for the analyses of real water samples and measurements of limit of detection (LOD). 3.2. SPME optimization In order to develop a direct-SPME procedure for the analysis of EGDN and NG in water samples, several parameters related to the extraction and desorption processes were evaluated (selection of SPME coating, effect of sorption and desorption mode and effect of temperature).

Fig. 2. Comparison of various elution solvents.


V. Je zov a et al. / J. Chromatogr. A 1174 (2007) 1319

Fig. 3. Comparison of recoveries with various SPME bres.

3.2.1. Selection of SPME coating Four SPME bre coatings were tested for isolation of nitrate esters from water samples. The bre coating dominates the recoveries of analytes and inuences sorption time. Optimization of SPME conditions for determination of nitrate esters was accomplished using aliquots of de-ionized water (4 mL) spiked with both nitrate esters (20 g mL1 ). Each bre (CW/TPR, CAR/PDMS, PDMS/DVB and CW/DVB) was immersed into spiked de-ionized water for 15 min at 50 C. The other sorption and desorption conditions were as follows: direct sampling, magnetic stirring static, and desorption in the desorption chamber for 10 min. Comparison of recoveries with these bres is shown in Fig. 3. The best results were obtained with PDMS/DVB bre. 3.2.2. Effect of sorption mode Inuence of sorption mode on the efciency of extraction was evaluated using stirred de-ionized water spiked with both nitrate esters. Conditions for direct and headspace sorption were kept constant in all experiments. Both types of sorption were conducted at 50 C with PDMS/DVB bre in vessels capped with a Teon-lined septum. Sorption time was limited to 15 min and time for desorption to 10 min in all cases. Extraction efciencies at direct and headspace sorption are compared in Fig. 4(a). Better sorption of analytes was achieved when SPME bre was exposed directly to water sample. EGDN and NG are not volatilized and this is why the extraction efciencies at headspace sorption were lower than in case of direct sorption. The direct sorption also showed to be the most suitable extraction mode for nitrate esters. SPME is an equilibrium process, in which analytes are distributed between the sample matrix and SPME stationary phase. This is the reason why the sorption time is necessary to be optimized. The inuence of extraction time was measured by immersing the PDMS/DVB bre into spiked de-ionized water for 170 min. In all cases, 4 mL of spiked de-ionized water, magnetic stirring, sorption at 50 C and desorption time 10 min were used. Adsorption prole was dened as the function of time (Fig. 4(b)). Rapid rise in extraction efciency for NG occurred in the interval 130 min and after that time the efciency remained constant. In case of EGDN extraction efciency increased with time (110 min) and then remained constant. Extraction time

Fig. 4. (a) Comparison of direct and headspace sorption. (b) Optimization of sorption time.

30 min proved to be sufcient to obtain quantitative extraction. This extraction time is a reasonable compromise between a good peak area and acceptable extraction time. 3.2.3. Effect of desorption mode Desorption period (the period during which the nitrate esters are eluated from bre coating by means of mobile phase in the SPME/HPLC-UV interface) was optimized after the sorption mode optimization. PDMS/DVB bre was immersed in stirred spiked de-ionized water at 50 C for 15 min. SPME bre was then placed in SPME/HPLC-UV interface. Desorption times varied within 1 and 20 min. The results (Fig. 5(a)) indicated fast desorption and 5 min desorption time was then used for all other SPME/HPLC-UV experiments. Efciency of static desorption with optimized desorption time was then compared with the efciency of dynamic desorption (the analytes are immediately injected into the column) that was, however, lower than that achieved with static desorption (Fig. 5(b)). 3.2.4. Effect of temperature Temperature plays an important role in the process of extraction, because it inuences the mass transfer rates and partition coefcients of the analytes. Extraction temperature prole 3070 C for PDMS/DVB bre was tested under these conditions: 4 mL stirred spiked de-ionized water, magnetic stirring, sorption 15 min and static desorption 10 min. NG peak area

V. Je zov a et al. / J. Chromatogr. A 1174 (2007) 1319


Fig. 7. (a) Calibration curves for SPE extraction. (b) Calibration curves for SPME extraction. Fig. 5. (a) Optimization of desorption time. (b) Comparison of dynamic and static desorption.

tion of sodium chloride to increase the ionic strength, and magnetic stirring. 3.3. Validation Optimized extraction conditions (SPE and SPME) were used for method validation. A ve-point calibration curve was created as described in Section 2. Good linearity was observed for both compounds in mentioned ranges (r2 = 0.99). The calibration curves resulting from analyses and the corresponding equations are shown in Fig. 7(a) for SPE extraction and Fig. 7(b) for SPME extraction. Typical HPLC separation of EDGN and NG with UV detection is shown in Fig. 8. The limit of detection was arbitrarily

increased with increasing temperature and the amount of EGDN sorbed decreased somewhat with increase in extraction temperature from 50 to 70 C (Fig. 6). Decreased EGDN sorption with increasing temperature is caused by the decrease of the distribution constant with increasing temperature. Temperature 50 C was chosen as a compromise to achieve good extraction efciencies for both methods and for both nitrate esters. In summary, the optimized extraction conditions were established to be: PDMS/DVB bre, 30 min sorption and 5 min static desorption time, 50 C temperature, addition of 10% water solu-

Fig. 6. Effect of temperature.

Fig. 8. HPLC separation of EDGN and NG with UV detection.


V. Je zov a et al. / J. Chromatogr. A 1174 (2007) 1319

Table 1 Concentrations of nitrate esters in model water samples spiked with 3 g/mL of each nitrate ester Analyte Measured concentrations (g/mL) SPE/HPLC-UV Spiked de-ionized water EGDN NG 2.85 0.13 2.80 0.11 Spiked drinking water 2.91 0.08 2.83 0.09 SPME/HPLC-UV Spiked de-ionized water 2.68 0.12 2.74 0.10 Spiked drinking water 2.63 0.11 2.69 0.13

Table 2 Concentrations of nitrate esters in real water samples taken in three time intervals Sample Measured concentrations (g/mL) SPE/HPLC-UV EGDN T1 T2 T3 5.17 0.11 6.97 0.21 6.27 0.19 NG 4.13 0.25 1.94 0.23 4.60 0.14 SPME/HPLC-UV EGDN 4.72 0.17 6.72 0.17 5.83 0.32 NG 3.91 0.22 1.68 0.11 4.12 0.30

set at 1/3 of the limit of quantication. The limit of quantication is the lowest concentration that can be determined in a sample with acceptable precision under the stated operational conditions of the method. The LODs for EGDN and NG obtained on the basis of the calibration curve were 0.23 g mL1 , and LOQs were 0.7 g mL1 for both nitrate esters in case of SPE extraction, LODs for EGDN and NG were 0.16 g mL1 and LOQs were 0.50 g mL1 in case of SPME extraction, respectively. The obtained LOD values were at the adequate levels for the identication of target nitrate esters in analysed water samples. Nitrate esters from 10 different water samples were extracted and analysed for the determination of the specicity of the method. No interference was observed in the region of the retention times of the individual analytes. 3.4. Comparison of SPE and SPME efciencyapplication to model water samples Two hundred millilitres of de-ionized or drinking water were spiked with EGDN and NG in concentration level 3 g mL1 of each nitrate ester. The solutions were analysed by means of SPE/HPLC-UV as well as SPME/HPLC-UV under optimized conditions, and the efciency was compared (Table 1). 3.5. Comparison of SPE and SPME efciencyapplication to real water samples Real water samples were analysed by SPE/HPLC-UV and SPME/HPLC-UV under optimized conditions. EGDN and NG were detected in all analysed samples and their concentrations are shown in Table 2. 4. Conclusion Efciency was compared of SPE and SPME at extraction of EGDN and NG from water samples. By means of model water

samples important parameters SPE and SPME were optimized and optimal measurement conditions for both extraction techniques were dened. Subsequent quantitative determinations were performed using SPE/HPLC-UV and SPME/HPLC-UV. When comparing with the other tested SPE cartridges, the highest extraction efciency was achieved with Water Oasis HLB. The best eluent proved to be acetonitrile. SPME bre with polydimethyl/siloxane/divinylbenzene coating was found to be the most suitable for extraction of nitrate esters from water samples. The combination of direct sorptiondynamic desorption proved to be the best to obtain high extraction efciency. For both extraction methods, the detection limits found were in g mL1 range with SPME detection limit being higher. If comparing SPE and SPME extraction efciency of SPE was higher and SPME was faster. The SPME method has the advantage of being rapid and organic solvent-free. The results of the study proved that SPE and SPME techniques, in combination with HPLC-UV, are suitable for quantitative determination of low concentrations of nitrate esters in water samples. The SPE and SPME procedures, both precise and accurate, satisfy nitrate esters producers requirements. Finally, the SPE approach is a further successful application in the quantication of two important nitrate esters in the context of analysis of esters in water samples; the SPME method, giving results comparable to those obtained with the SPE method. Acknowledgements The experiments were performed thanks to nancial support from the Ministry of Education, Youth and Sports of the Czech Republic (project MSM0021627502) and the Grant Agency of 203/05/2106). the Czech Republic (GACR References
[1] J. Yinon, Forensic and Environmental Detection of Explosives, Wiley, Chichester, 1999. [2] J. Feltes, K. Levsen, D. Volmer, M. Spiekermann, J. Chromatogr. A 518 (1990) 21. [3] T.F. Jenkins, D.C. Leggett, P.H. Miyares, M.E. Walsh, T.A. Ranney, J.H. Cragin, V. George, Talanta 54 (2001) 501. [4] F. Monteil-Rivera, Ch. Beaulieu, S. Deschamps, L. Paquet, J. Hawari, J. Chromatogr. A 1048 (2004) 213. [5] A. Halasz, C. Groom, E. Zhou, L. Paquet, C. Beaulieu, S. Deschamps, A. Corriveau, S. Thiboutot, G. Ampleman, C. Dubois, J. Hawari, J. Chromatogr. A 963 (2002) 411. [6] M. Smith, G.E. Collins, J. Wang, J. Chromatogr. A 991 (2003) 159.

V. Je zov a et al. / J. Chromatogr. A 1174 (2007) 1319 [7] D.A. Cassada, S.J. Monson, D.D. Snow, R.F. Spalding, J. Chromatogr. A 844 (1999) 87. [8] Solid Phase Microextraction Application Guide, T396098, 5th ed., Supelco, Bellefonte, PA, 2004. [9] E. Psillakis, N. Kalogerakis, J. Chromatogr. A 938 (2001) 113. [10] S. Calderara, D. Gardebas, F. Martinez, Forensic Sci. Int. 137 (2003) 6. [11] J. Skl adal, V. Je zov a, A. Eisner, M. Kouba, K. Ventura, New Trends in Research of Energetic Materials 8, University of Pardubice, Pardubice, 2005, p. 789.


[12] W. Tomaszewski, V.M. Gunko, J. Skubiszewska-Zi eba, R. Leboda, J. Colloid Interface Sci. 266 (2003) 388. [13] S.-A. Barshick, W.H. Griest, Anal. Chem. 70 (1998) 3015. [14] D.D. Fetterolf, in: J. Yinon (Ed.), Forensic Applications of Mass Spectrometry, CRC Press, Boca Raton, FL, 1995, p. 215. [15] D.T. Burns, R.J. Lewis, Anal. Chim. Acta 307 (1995) 89. [16] J. Skl adal, M. Kouba, M. Mudru nkov a, M. Slez ak, New Trends in Research of Energetic Materials 9, University of Pardubice, Pardubice, 2006, p. 723.