Selective enrichment of the degradation products of organophosphorus nerve agents by zirconia based solid-phase extraction

Journal of Chromatography A, 1218 (2011) 6612–6620

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Selective enrichment of the degradation products of organophosphorus nerve agents by zirconia based solid-phase extraction Pankaj K. Kanaujia, Deepak Pardasani, Vijay Tak, Ajay K. Purohit, D.K. Dubey ∗ Vertox Laboratory, Defence Research and Development Establishment, Gwalior, India

a r t i c l e

i n f o

Article history: Received 9 June 2011 Received in revised form 26 July 2011 Accepted 28 July 2011 Available online 6 August 2011 Keywords: Solid-phase extraction Phosphonic acids Carboxylic acids Zirconia Nerve agents HybridSPE

a b s t r a c t Selective extraction and enrichment of nerve agent degradation products has been achieved using zirconia based commercial solid-phase extraction cartridges. Target analytes were O-alkyl alkylphosphonic acids and alkylphosphonic acids, the environmental markers of nerve agents such as sarin, soman and VX. Critical extraction parameters such as modifier concentration, nature and volume of washing and eluting solvents were investigated. Amongst other anionic compounds, selectivity in extraction was observed for organophosphorus compounds. Recoveries of analytes were determined by GC–MS which ranged from 80% to 115%. Comparison of zirconia based solid-phase extraction method with anion-exchange solidphase extraction revealed its selectivity towards phosphonic acids. The limits of detection (LOD) and limit of quantification (LOQ) with selected analytes were achieved down to 4.3 and 8.5 ng mL−1 , respectively, in selected ion monitoring mode. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Retrospective identification of chemical warfare agents (CWAs) and related chemicals in environmental matrices requires efficient sample preparation. It has extensively attracted attention of researchers because of high degree of toxicity associated with CWAs and need of ensuring compliance to Chemical Weapons Convention (CWC) [1–5]. CWC is an international treaty, administered by the Organization for Prohibition of Chemical Weapons (OPCW) and prohibits production, storage and use of CWAs other than the purposes not prohibited in convention [2,5,6]. Compliance of CWC is ensured by the OPCW through on-site inspection of declared and suspected sites; and off-site analysis of collected samples in designated laboratories [2,3]. Organophosphorus nerve agents belong to the most lethal class of CWAs; they comprise G-agents such as (tabun, sarin, soman and cyclosarin) and V-agents (VX) [7]. In the environment, water hydrolyzes these nerve agents initially to alkyl alkylphosphonic acids and eventually to alkylphosphonic acids as shown in Fig. 1 [8]. As a result, these phosphonic acids (PAs) are considered as environmental markers (signatures) of their corresponding nerve agent

∗ Corresponding author at: Vertox Laboratory, Defence Research and Development Establishment, Jhansi Road, Gwalior 474 002, India. Tel.: +91 751 2233488; fax: +91 751 2341148. E-mail addresses: [email protected], [email protected] (D.K. Dubey). 0021-9673/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2011.07.091

because their identification indicate prior presence of nerve agents. Identification of PAs in environmental samples can be achieved by mass spectrometry coupled to gas chromatography (GC–MS) or liquid chromatography (LC–MS). However, extraction and enrichment of PAs from environmental samples is prerequisite before subjecting to GC–MS or LC–MS analysis. Hence, development of extraction methods of degradation products of nerve agents from environmental matrices has evinced attraction of several workers [9–13]. PAs are acidic degradation products of nerve agents with their pKa 1 and pKa 2 values of ∼2.5 to ∼7.0, respectively; hence they tend to remain in their anionic state in neutral water (pH ∼7). Anion-exchange solid-phase extraction (SPE) is highly suitable for extraction and enrichment of such chemicals from water [14,15], but undesired organic anions (as carboxylic acids) are also likely to be co-extracted with PAs. Other sample preparation protocols for extraction of these PAs from water and aqueous extracts of soil involve removal of interfering cations (by passing through cation-exchange SPE cartridges) followed by evaporation of water with subsequent derivatization and analysis by gas chromatography mass spectrometry (GC–MS) [9,16]. With this method, removal of neutral interferents such as polyethylene glycols is not possible. In addition, molecularly imprinted polymers have been used for their extraction from water with good recoveries; the prerequisite of this method is change of sample matrix from aqueous to organic phase, as strong hydrogen bonding of water molecules interferes with interaction of analytes [17,18]. Recently, metal oxides such as TiO2 and ZrO2 and others have been investigated as selective extraction materials for several

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620 O R

P

O OR1

hydrolysis

R

X Nerve Agents

P

O OR1

hydrolysis

R

P

OH

OH

OH

O-alkyl alkylphosphonic acids

Alkylphosphonic acids

R = CH3, C2H5, i-C3H7, n-C3H7 R1= C1-C10 alkyl and cycloalkyl Sarin: R=CH3, R1=i-C3H7, X=F;

VX: R=CH3, R1=C2H5, X=SCH2CH2N(R)2

Fig. 1. Hydrolytic pathways of nerve agents.

classes of organic compounds [19–23]. In particular, zirconia’s surface chemistry, chemical and thermal stability have drawn attention of researchers to explore it as a potential chromatographic phase [24]. Zirconia has been studied as a material for selective extraction/removal of phospholipids and other phosphorus containing endogenous components in biological matrices [19–23]. Zirconia strongly exhibits Lewis acid properties and interacts with Lewis bases present in the sample. Phospholipids (phosphates) being strong Lewis base interact with zirconium by donating electrons in its empty d-orbitals and therefore can be effectively removed from biological fluids by this mechanism. Besides, there are a number of investigations in which this property of zirconia has been exploited for the extraction of small organic species. Li et al. extracted melamine residues in dairy products employing zirconia hollow fibre sorptive micro extraction exploring hydrogen bonding between Zr–OH and H2 N– of melamine [25]. Du et al. reported SPE of organophosphate pesticides at zirconia nanoparticles modified electrode followed by stripping voltammetric detection [26]. In a similar study, zirconia nanoparticles based electrochemical immunosensor was developed for the detection of phosphorylated acetyl cholinesterase [27]. Preparation and characterization of zirconia hollow fibre was reported by Xu and co-workers with its successful application in the micro extraction of pinacolyl methylphosphonic acid (degradation marker of soman) [28]. However, it required the cumbersome preparation of hollow fibre, and lack of commercial availability restricted its application. Commercial SPE cartridges named HybridSPETM available with Supelco, Sigma–Aldrich have found large number of applications in the method development and recovery optimization for biomolecules. Its packed bed of proprietary adsorbent comprise zirconia coated silica particles and is primarily applicable to reduce phospholipids that cause ionization suppression in mass spec-

6613

trometric analysis of biological samples containing phospholipids as endogenous interferents. A number of reports are available in the literature involving use of HybridSPE in removal of phospholipids background from urine, plasma and other biomatrices [19,20,23,29,30]. Thus considering the selective adsorption of phosphorus compounds on zirconia, inadequacy of anion- and cation-exchange based solid-phase extraction to remove anionic, cationic and neutral interferents and commercial availability of zirconia based SPE cartridges, we envisaged their use as selective extractants for enrichment of PAs from water samples; for which the mechanistic outline is presented in Fig. 2. In this study, we report the selective extraction of various PAs from water containing various saturated carboxylic acids (CAs) as background. For this purpose, commercially available SPE cartridges (HybridSPE) were used which exhibited selectivity towards organophosphorus nerve agent’s markers. It remained non-selective towards a range of basic, neutral and acidic background compounds. Various extraction parameters were optimized during the course of study and validity of the method for aqueous samples was demonstrated. 2. Experimental 2.1. Materials The analytical grade solvents were obtained from Merck Specialties Pvt. Ltd. (Mumbai, India). The names, structures and abbreviations of acidic degradation products of nerve agents (PAs) used in optimizing various extraction parameters are O-isopropyl methylphosphonic acid (IMPA), O-propyl ethylphosphonic acid (PrEPA), O-sec-butyl ethylphosphonic acid (BEPA), O-pentyl methylphosphonic acid (PMPA), O-pentyl isopropylphosphonic acid (PIPA) and ethylphosphonic acid (EPA). Three additional PAs namely O-cyclohexyl methylphosphonic acid (CHMPA), methylphosphonic acid (MPA) and isopropylphosphonic acid (IPA) were used during internal validation. All the PAs were synthesized in house as per the reported procedure [31]. Carboxylic acids (CAs) which were used as background chemicals throughout the study included hexanoic acid (HA), 5-phenyl valeric acid (PVA), octanedioic acid (ODA) and decanedioic acid (DDA). Heptanoic acid (HPA) and octanoic acid (OA) were used during internal validation. All the carboxylic acids were obtained commercially from Sigma Aldrich Chemicals Pvt. Ltd. (Banga-

Fig. 2. Zirconia mediated selective extraction of organophosphorus degradation products of nerve agents.

6614

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

O

O

O

P O

P O

P O

OH

OH

O-isopropyl methylphosphonic acid (IMPA)

O

O-sec-butyl ethylphosphonic acid (BEPA)

O

O

P O

P O

P O

OH

OH

OH O-cyclohexyl methylphosphonic acid (CHMPA)

O-pentyl isopropylphosphonic acid (PIPA)

O-pentyl methylphosphonic acid (PMPA)

O

O

O

P OH

P OH

OH

OH

P OH OH

Methylphosphonic acid (MPA)

O

OH

O-propyl ethylphosphonic acid (PrEPA)

Ethylphosphonic acid (EPA)

Isopropylphosphonic acid (IPA)

O OH

Hexanoic acid (HA)

OH

O

OH

HO

O Heptanoic acid (HPA)

HO O 5-Phenyl valeric acid (PVA)

O

Octanoic acid (OA)

O

OH

Nonanoic acid (NA) O

O

OH

HO

HO Octanedioic acid (ODA)

Decanedioic acid (DDA)

Fig. 3. Structure and abbreviations of analytes and interferents (phosphonic and carboxylic acids, respectively) selected for the study.

lore, India). The structures of all PAs and CAs are presented in Fig. 3. N,O-bis(trimethylsilyl) trifluoroacetamide (BSTFA), pentadecane (internal standard), polyethylene glycol (PEG-300) and HybridSPETM Precipitation SPE tubes (30 mg, 1 mL) were purchased from Sigma Aldrich Chemicals Pvt. Ltd. (Bangalore, India). Formic acid, citric acid, sodium sulfate, calcium chloride and potassium fluoride were purchased from Merck Specialties Pvt. Ltd. (Mumbai, India). The SampliQ silica based strong anion-exchange SPE cartridges (Si-SAX) (200 mg, 3 mL and 0.6 mequiv. g−1 ) were obtained from Agilent Technologies (Milwaukee, WI, USA). All aqueous solutions were prepared in triply distilled deionized water. 2.2. Instrumentation The quantitative analyses were performed with an Agilent 6890N gas chromatograph equipped with a model 5975 inert XL mass selective detector (Agilent Technologies, Milwaukee, WI, USA). An SGE BPX5 capillary column with 30 m length × 0.25 mm i.d. × 0.25 ␮m film thickness was used. The GC oven temperature was programmed from 80 ◦ C (hold for 2 min) to 250 ◦ C at 15 ◦ C min−1 then at 40 ◦ C min−1 to a final temperature of 300 ◦ C (hold for 2 min). Helium at a flow rate of 1.2 mL min−1 was used as carrier gas under constant flow mode. The transfer line temperature was fixed to 280 ◦ C. Samples were analyzed in splitless mode at an

injection temperature of 250 ◦ C. Injection volume was kept at 1 ␮L with a 5 ␮L micro syringe. EI source was kept at 230 ◦ C with 70 eV ionization energy and the quadrupole temperature at 150 ◦ C. Quantitation studies were performed in selected ion monitoring (SIM) mode; the ions selected for monitoring the trimethylsilyl derivatives of analytes were 153 for IMPA, PMPA and CHMPA; 167 for PrEPA and BEPA; 181, 225, 239 and 253 for PIPA, MPA, EPA and IPA, respectively; 73 for ODA and DDA; 75 for HA, HPA, OA and PVA. 2.3. Standards and spiking solutions Single analyte stock standard solutions of all the selected PAs and CAs were prepared in acetonitrile at a concentration of 1 mg mL−1 separately in a screw capped vial with PTFE septa. The stock standard solution of internal standard (pentadecane) was prepared in toluene at a concentration of 1 mg mL−1 . For optimization of extraction parameters, intermediate working standards of PAs and CAs were prepared in acetonitrile separately at 500 ␮g mL−1 . Water sample was prepared by adding appropriate amounts of both the intermediate working standards to get the final concentration of PAs and CAs at 1 ␮g mL−1 and 10 ␮g mL−1 , respectively. Stock solutions of PEG-300, Na2 SO4 and CaCl2 were prepared at 100 mg mL−1 separately. During preparation of samples, stock solutions of PEG, Na2 SO4 and CaCl2 were diluted to

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

6615

get their individual concentration at 1 mg mL−1 . For determining internal validity of the developed method, two different spiking concentrations of PAs were used, i.e. 0.1 and 2.0 ␮g mL−1 . 2.4. Solid-phase extraction procedure All extractions were performed using VisiprepTM SPE vacuum manifold (Supelco, Bellefonte, PA, USA). For optimization of extraction parameters, sample volume was kept constant at 1 mL and spiking concentrations of PAs and CAs were kept at 1 ␮g mL−1 and 10 ␮g mL−1 , respectively. Conditioning of HybridSPE cartridge was performed only with 1 mL water followed by sample loading under gravity. Prior to loading, 2% formic acid was added to the water sample. In cases where PEG-300 and metal salts (Na2 SO4 and CaCl2 ) were present in the sample, a rinsing step with 2 mL water was performed after loading. Afterwards, the cartridge was eluted with 1 mL of 10% ammonium hydroxide (NH4 OH) solution in methanol under mild vacuum at a flow rate of 0.5 mL min−1 . The eluate was evaporated in a rotary evaporator to dryness and residue was re-dissolved in 100 ␮L acetonitrile. Reconstituted sample was derivatized with BSTFA to convert the analytes into their corresponding trimethylsilyl (TMS) derivatives. TMS derivatives are volatile and GC amenable. Derivatives were prepared as per standard procedure [9,10,14]. Briefly, 100 ␮L BSTFA was added to vial, it was sealed thereafter and heated at 70 ◦ C for 1 h; cooled to room temperature and analyzed with GC–MS after adding internal standard. For method validation and comparison, the sample was processed with the procedures discussed above and compared with silica based strong anion-exchange (Si-SAX) cartridge. Extraction with Si-SAX was performed in accordance with the procedure reported earlier [14]. Recoveries of analytes were calculated from the ratios of peak areas of analytes to internal standard, compared in control versus spiked samples. Control was processed in parallel to the sample and prepared by adding analytes in post extract aliquot exactly at the concentration corresponding to the maximum theoretically possible recovery. All extractions were performed in triplicate and the values shown in figures are the mean of triplicate runs. After extraction, cartridge was always discarded; however, it was observed that recycling the used cartridges for fresh extractions did not result in lowering of analyte recoveries. 3. Results and discussion Although ‘HybridSPE Precipitation’, as the name suggests, has been devised primarily for bioanalytical applications for minimizing ionization suppression due to phospholipids in mass spectrometric analysis, however, with equal efficacy and selectivity, we made use of this adsorbent in cartridge form for the selective enrichment of trace amounts of organophosphorus degradation markers of nerve agents from water. During environmental analysis, removal of excessive background from matrix interferents is of prime concern and SPE being most versatile approach has been remarkably exploited. Selection of appropriate sorbent for SPE is the backbone of extraction and discrimination between analytes and background is attained. In a challenging situation, similar chemical nature of analytes and background often result in poor optimization of recovery of target analytes which necessitates selective extraction/removal of matrix components. During the preliminary screening, we found HybridSPE helps in making a distinction between PAs and CAs by virtue of Lewis acid–base interaction strength. By fine tuning the pH of adsorption and desorption, specificity in retention of phosphonate anion (from PAs) on zirconia (adsorbent) was increased significantly. HybridSPE adsorbent contains zirconia, surface of which is charged and consists of oxygen atoms having negative charge

Fig. 4. Effect of formic acid addition on the recoveries of phosphonic acids (A); and carboxylic acids (B) spiked at 1 ␮g mL−1 and 10 ␮g mL−1 , respectively.

and the positive charge gets accumulated on zirconium atoms. Thus, electrophilicity of zirconium along with its vacant d-orbitals imparts Lewis acidity in it. In a real world water sample, several nucleophiles may be expected such as chloride, sulfate, carboxylate, and phosphate which are also Lewis bases and interact with zirconia with varying strength. In this study, we hypothesized a scenario where phosphonates (target analytes) are present collectively at 1 ␮g mL−1 with carboxylates (matrix interferents) at exactly 10folds higher concentration. Phosphonate being stronger Lewis base than carboxylates forms coordination complex with zirconia more strongly than carboxylates. In turn, hydroxide being strongest of all these bases was used to displace retained Lewis bases on zirconia. Thus lowering the concentration of OH− ions by reducing the pH, interactions of other bases may be enhanced with zirconia surface. Subsequent increase in the pH of eluent can dislodge the retained analytes. This principle was used for retention and elution of PAs from zirconia. During the course of study, eluting solvent, rinsing solvent, amount and nature of matrix modifier were optimized. The internal validation was performed with the help of tap water spiked at two different concentrations and external validation with water sample supplied during an official OPCW proficiency test. 3.1. Selection of modifier HybridSPE considerably reduces levels of residual phospholipids contaminations in biological samples and is evident by a large

6616

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

Fig. 5. Recoveries of phosphonic and carboxylic acids spiked at 1 and 10 ␮g mL−1 , respectively, after the addition of fluoride anions (potassium fluoride) (A and B) and; citrate anions (citric acid) (C and D) in the sample just before loading.

number of research reports discussed earlier [20–24]. In our study, since the target phosphonates, by virtue of their strong nucleophilicity interact with zirconia, carboxylates present in the sample at higher concentrations are likely to co-retain. To overcome this, formic acid was added to the sample prior to loading and became essential part of sample before processing. The principle emphasizing the use of modifier is their relatively strong Lewis basicity than carboxylates, hence, modifier binds competitively with available uncoordinated zirconium minimizing retention of carboxylates. Modifier is weaker Lewis base than phosphonates (or phosphates), so do not affect their retention. Few modifiers have been reported and recommended by supplier and mainly include formate and citrate. In this study, we added varying concentrations of formate anions (formic acid) (v/v) to the sample, initially to optimize its amount, and then compared with other Lewis bases such as citrate (citric acid) and fluoride (KF) anions which are stronger than carboxylates but weaker than phosphonates. Retained analytes were eluted with NH4 OH after washing with 2 mL water; details of optimization of eluent are discussed later in Section 3.2. Fig. 4 shows optimization of formic acid concentrations in loading samples. Formic acid was added to the samples just before loading and its concentration was varied from 0.25% to 5%. Almost 95% to complete removal of carboxylates was observed with 2% and 5% formic acid without affecting the recoveries of PAs. With 1% formic acid, 85–92% of dibasic CAs were removed. In the presence of water, the surface of zirconia gets converted to hydroxyls (Zr–OH). The addition of formic acid through sample results in lowering the pH of sample to nearly 2.5–3, which causes protonation of Zr–OH which in turn is responsible for the increase in the Lewis acidity of zirconium [19]. The coordination sites of zirconium are occupied with water, hydroxyl ions and hydronium ions depending upon the pH, and

form coordination complexes with Lewis bases like borate, sulfate, carboxylates, phosphates and fluoride [19]. For this reason, individual effect of formate, citrate and fluoride on the recoveries of CAs and PAs were investigated. Fig. 5 shows the effect of addition of fluoride (KF) and citrate anions (citric acid) on recoveries of PAs and CAs. It is clear from Fig. 5A that recoveries of CAs were less than 2% with just 0.25% KF (w/v), but at the same concentration of KF, loss of PAs was relatively less. Further increase in KF concentration to 2% and 5% caused complete loss of PAs except EPA whose recovery was 21% (Fig. 5A and B). Similarly, increase in citric acid addition to sample from 0.25% to 1% caused decrease in the recoveries of CAs from 4% to less than 2% (Fig. 5C). Loss of around 60–85% was seen in recoveries of all the PAs with increase in citrate anion concentration from 0.25% to 5% (Fig. 5D). With these observations we concluded that amongst these three modifiers (formate, citrate and fluoride), maximum recoveries of PAs and removal of CAs were observed with 2% formic acid in samples prior to loading. Rinsing of cartridges with 2% aqueous formic acid solution after loading of sample (containing no modifier) resulted in the marginal loss of PAs when spiking concentration of CAs was raised to 100 ␮g mL−1 (data not shown). This observation substantiated our finding that PAs and CAs if loaded collectively in the absence of modifier, competitively retain over zirconia with different strength of interactions and selectivity can be imparted in extraction and enrichment of PAs in the presence of CAs. 3.2. Selection of factors affecting elution of analytes: volume of eluting solvent and flow rate As the water sample, containing 1 ␮g mL−1 and 10 ␮g mL−1 of PAs and CAs, respectively, was percolated down the HybridSPE cartridge, PAs preferentially retained over sorbent followed by

Washing solvent Phosphonic acids Spiking (1 ␮g mL−1 )

Without Formic acid in sample

With 2% Formic acid in sample

Carboxylic acids Spiking (10 ␮g mL−1 )

Polyethylene glycols Spiking (1 mg mL−1 )

IMPA

PrEPA

EPA

BEPA

PMPA

PIPA

HA

PVA

ODA

DDA

DEG

TEG

PEG

HEG

No washing 1 mL water 2 mL Water 5 mL water 1 mL methanol 2 mL methanol

114.3 (6.4) 104.3 (5.9) 114.9 (5.1) 111.2 (6.2) 117.5 (6.2) 99.9 (6.4)

113.5 (5.9) 109.0 (5.3) 114.9 (6.0) 111.0 (6.4) 115.0 (7.0) 108.3 (6.3)

74.5 (5.1) 79.0 (5.1) 76.8 (5.6) 79.6 (5.5) 80.4 (6.0) 84.9 (6.3)

117.1 (5.2) 111.8 (5.0) 113.1 (5.1) 109.8 (5.6) 118.0 (5.4) 115.4 (5.2)

117.7 (5.6) 115.1 (5.4) 115.1 (5.7) 116.4 (4.8) 118.9 (4.1) 115.2 (5.3)

96.7 (4.5) 99.9 (4.7) 108.4 (4.8) 96.5 (5.0) 110.0 (5.4) 103.0 (5.6)

23.1 (6.5) 17.9 (5.4) 12.4 (6.2) 6.9 (6.5) 12.5 (6.8) 7.1 (6.0)

34.5 (3.7) 27.5 (3.5) 17.6 (4.4) 1.8 (4.5) 17.5 (5.0) 12.3 (5.2)

101.2 (5.5) 105.6 (5.1) 104.9 (4.3) 104.6 (5.3) 98.9 (6.4) 102.1 (6.2)

103.3 (6.1) 105.3 (6.6) 101.5 (6.7) 103.8 (7.0) 99.9 (6.5) 104.9 (6.4)

13.2 (5.2) 2.4 (4.6) 2.0 (4.6) 1.6 (5.2) 3.5 (5.0) 1.9 (4.9)

34.9 (7.0) 0.8 (6.8) 0.6 (6.4) 0.2 (5.1) 0.9 (4.8) 0.4 (3.8)

25.3 (5.6) 1.0 (5.4) 0.7 (5.6) 0.2 (5.1) 1.0 (5.8) 0.5 (4.9)

41.7 (6.4) 1.1 (6.0) 0.8 (5.0) 0.3 (5.1) 1.1 (5.3) 0.6 (4.9)

No washing 1 mL water 2 mL water 5 mL water 1 mL methanol 2 mL methanol

90.1 (5.4) 96.7 (5.3) 99.2 (5.0) 95.6 (4.9) 82.3 (6.1) 63.5 (6.0)

99.8 (5.5) 90.5 (5.6) 102.9 (5.9) 101.8 (5.0) 92.9 (6.7) 75.5 (6.4)

74.6 (6.6) 76.2 (6.5) 69.3 (7.1) 68.6 (7.0) 66.3 (6.3) 62.7 (6.5)

98.9 (4.5) 95.8 (4.3) 98.9 (4.7) 98.2 (5.6) 87.4 (5.4) 75.8 (5.8)

95.2 (7.1) 97.2 (6.8) 99.4 (6.2) 97.1 (5.7) 82.1 (4.8) 82.7 (7.8)

99.8 (5.8) 97.4 (5.1) 102.1 (4.5) 106.6 (4.8) 93.6 (4.9) 90.6 (4.1)

4.1 (3.9) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)

8.7 (3.6) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)

14.1 (3.9) 2.2 (5.5) 3.0 (6.1) 3.7 (6.3) 2.2 (5.9) 1.9 (4.6)

14.1 (6.9) 3.6 (5.9) 3.5 (5.7) 4.0 (4.4) 2.2 (5.1) 1.8 (5.0)

5.9 (4.7) 2.0 (3.8) 1.9 (4.1) 1.5 (4.0) 1.7 (4.2) 1.3 (3.8)

24.1 (6.1) 0.4 (6.2) 0.6 (5.2) 0.4 (4.9) 0.9 (4.8) 0.0 (3.9)

31.0 (6.5) 0.6 (6.5) 0.6 (6.2) 0.2 (5.0) 0.1 (5.0) 0.0 (0.0)

37.1 (5.8) 1.1 (5.9) 0.9 (5.4) 0.2 (5.0) 0.9 (5.2) 0.0 (0.0)

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

Fig. 6. Optimization of volume of eluting solvent (10% ammonium hydroxide in methanol).

CAs (which got retained over remained sites which are generally Zr–OH− and free Si–OH). To check the retention of nucleophiles other than phosphonates (carboxylates in this case), modifier (2% formic acid) was added to the sample prior to loading (discussed in Section 3.1). The presence of CAs in effluent (from column) indicated that they were not retained on the column. After loading, rinsing step was performed with optimized solvents (results discussed Section 3.3) followed by the elution of the retained analytes. The cartridges were eluted with varying volumes of 10% NH4 OH solution in methanol. Fig. 6 shows the percentage mean recoveries of PAs obtained with 0.6, 1.25, 2.5, 5 and 10 mL of eluting solvent. Quantitative recoveries of all the PAs (except EPA) were obtained with 1–1.25 mL of eluting solvent with a moderate loss in EPA, whose recovery was around 83% with 1 mL eluting solvent (Fig. 6A). Recoveries of all the selected CAs were found to be less than 8% with 1 mL eluent. It shows that CAs were not retained effectively on the sorbent (Fig. 6B). In the absence of modifier (HCOOH), however, recoveries of all the CAs were sufficiently high which indicated that CAs were indeed retained but not eluted with selected solvent. It has been discussed in detail in Section 3.1. One mL of eluting solvent (10% NH4 OH in methanol) resulted in excellent recoveries of all PAs, hence was considered adequate for further optimizations of other critical variables involved in extraction process. To see the effect of flow rates of eluting solvents through cartridges, 0.5, 1, 1.5 and 2 mL min−1 flow rate were applied and confirmed to cause no decrease or increase in the recoveries for any of the studied analytes (PAs as well as CAs). We finally concluded that the retention (between zirconia and PAs/CAs) were

Table 1 Effect of washing (rinsing) solvent on the recoveries of phosphonic acids and matrix interferences (carboxylic acids and polyethylene glycols). Values given against the analytes are their percentage mean recoveries with relative standard deviations in parenthesis.

Abbreviations: DEG – diethylene glycol; TEG – tetraethylene glycol; PEG – pentaethylene glycol and HEG – hexaethylene glycol. 6617

6618

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

A

Abundance

1

TIC: DIREVAP.D\DATASIM.MS

6.169

4000000

2 7.183

3500000 3000000 2500000

IS

2000000

9.368

1500000 1000000 500000 Time-->

B

0 5.50

PIPA (TMS)

PMPA (TMS)

8.957

8.098

6.00

6.50

7.00

Abundance

7.50

8.00

8.50

9.00

9.50

IS

TIC: HSPE.D\DATASIM.MS

9.360

1600000 1400000

BEPA (TMS)

1200000 1000000 800000

PrEPA (TMS)

6.863

6.224

PMPA (TMS)

PIPA (TMS)

7.920

8.871

600000 400000

2 7.141

200000

Time-->

0 5.50

6.00

6.50

7.00

7.50

8.00

8.50

9.00

9.50

Fig. 7. Comparison of SIM chromatograms of extraction carried through HSPE (B) and direct evaporation (A) with spiking concentrations of phosphonic acids and carboxylic acids were kept 1 and 100 ␮g mL−1 , respectively. 1: Heptanoic acid (TMS); 2: octanoic acid (TMS); IS: internal standard.

largely Lewis acid–base interactions (primary interactions) and the equilibrium was not influenced by the flow rate of eluting solvents, hence throughout the recovery optimization of PAs, 0.5 mL min−1 flow rate of solvents through cartridge was considered optimum. 3.3. Effect of matrix interferences on analyte recoveries Besides CAs, effects of other matrix interferents commonly encountered with real water samples were also investigated. For this, water samples containing PAs at 1 ␮g mL−1 were fortified with 10, 1000 and 1000 ␮g mL−1 of CAs, PEG-300 and metal salts (Na2 SO4 and CaCl2 ), respectively. Thus water samples now contained three other anions besides phosphonates. Under the optimized elution volume of 1 mL of 10% NH4 OH solution in methanol, two set of experiments were performed; first, the sample (1 mL) was processed without modifier (formic acid) and the second sample was processed with 2% formic acid as modifier. After the loading, rinsing of HybridSPE cartridge with varying volumes of water and methanol was performed. Mean percentage recoveries of PAs, CAs and PEGs are presented in Table 1. It is evident from data that modifier significantly affects the CAs retention on zirconia. The presence of formic acid in sample lowered the CAs recovery considerably in the eluate from 23%, 34%, 101% and 103% to 4%, 8.6%, 14% and 14.2% for HA, PVA, ODA and DDA, respectively, even without performing the washing step. When washing with 1 mL water was performed, HA, PVA, ODA and DDA were significantly reduced in the eluate of sample containing 2% formic acid. In this sample HA and PVA were completely

removed, ODA and DDA were reduced to 3.0% and 3.5%, respectively. On the other hand, Recoveries of HA, PVA, ODA and DDA in the eluate of sample with no modifier were 12.4%, 17.6%, 104.8% and 101.5%, respectively, with 2 mL water washing of cartridge (optimization of modifier is described in Section 3.1). However, with 2 mL water washing and sample with modifier, excellent recoveries of PAs were still observed. Recoveries of IMPA, PrEPA, EPA, BEPA, PMPA and PIPA under these conditions were 99.2%, 102.9%, 69.3%, 98.9%, 99.42% and 102.1%, respectively. Marginally lower recoveries of dibasic EPA were observed throughout the study and could not be increased further. Washings with 1 and 2 mL methanol did not result in any significant reduction in the recoveries of CAs; instead, it lowered recoveries of most of the PAs marginally. PEG-300 generally consists of diethylene glycol (DEG), tetraethylene glycol (TEG), pentaethylene glycol (PEG) and hexaethylene glycol (HEG) which were almost completely removed to less than 2% with just 2 mL of water washings and were unaffected with the presence of modifier in sample. Similarly, anions derived from salts (Na2 SO4 and CaCl2 ) also did not interfere at all with any PAs and CAs in their extraction and their respective recoveries remained unaffected; it is attributed to the weak Lewis acid–base interactions between zirconium and anions (Cl− and SO4 2− ) (data not shown). These finding helped us to summarize that owing to strong Lewis base character phosphonates were selectively retained on zirconia cartridge in comparison to carboxylates, PEGs and metal salts even when these were present at much higher concentration. Effect of much higher concentration of CAs and CaCl2 (1 mg mL−1 ) over PAs (1 ␮g mL−1 ) was further investigated and discussed in Section 3.4.

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

A

Abundance

1.6e+07

6619

TIC: HSPE.D\DATA.MS

IMPA (TMS)

MPA (bis-TMS) *

5.545 5.902

1.4e+07 1.2e+07

IS

1e+07

9.364

8000000 6000000

EPA (bis-TMS) *

4000000

6.618

2000000 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.50

Time-->

B Abundance 1.8e+07 1.6e+07

IMPA (TMS) 5.573

MPA (bis-TMS) *

TIC: SAX.D\DATA.MS

5.919

1.4e+07 1.2e+07

IS

1e+07 8000000 6000000 4000000

9.366

EPA (bis-TMS)* 6.631

2000000 Time-->

5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.50

Fig. 8. Analysis of 26th OPT water sample (code: 811) which contained IMPA at 10 ␮g mL−1 by HybridSPE (A) and SampliQ Si-SAX (B); peak at Rt 9.3 min in (A) and (B) is of pentadecane (internal standard for chromatographic analysis). *Method internal standard.

3.4. Comparison of method The addition of modifier to the samples hardly matters when exogenous nucleophilic background in environmental sample (such as carboxylates, borates, and sulfates) is present at lower concentration. At very high concentrations, however, retention on zirconia might get dominated by these species (applicable in the analysis of CWAs and related chemicals from environmental samples). To simulate this scenario, 1 mL water sample containing 1, 1000 and 1000 ␮g mL−1 of PAs, CAs and CaCl2 was processed with HybridSPE under optimized conditions of washing, modifier addition and elution. As discussed in previous section, anions arising from metal ions (Cl− and SO4 2− ) do not interfere in the extraction of PAs with HybridSPE, the effect of higher concentrations of CAs on the enrichment and detection of PAs is demonstrated in Fig. 7. Total ion chromatogram (TIC) of sample which was directly evaporated and analyzed after derivatization without any clean-up was compared with TIC of sample processed with HybridSPE (shown in Fig. 7A and B, respectively). It is evident that two saturated GC signals in Fig. 7A at Rt 6.16 and 7.18 corresponding to heptanoic and octanoic acid (trimethylsilyl derivatives) spiked at 1000 ␮g mL−1 completely masked two analytes (PrEPA and BEPA at Rt 6.22 min and 6.86 min, respectively, Fig. 7B). By comparing Fig. 7A and B, it is important to note that the presence of CAs at very high concentration to PAs significantly affected retention profiles of other analytes. With the HybridSPE cartridge, cleanup and selectivity towards PAs in the presence of CAs is clearly demonstrated. Samples sent by the OPCW in official proficiency tests (OPTs) are blind samples to participating laboratories and provide an opportunity for further validation of the newly developed method.

26th OPT conducted during October 2009, included a water sample (code: 811) which contained IMPA (immediate degradation marker of highly toxic nerve agent sarin) at concentration of 10.2 ␮g mL−1 . Since IMPA is one of the analytes used in our optimization study, hence this sample (1 mL) was processed with the HybridSPE method under the optimized extraction parameters. The recovery of IMPA was compared with that obtained with Si-SAX under optimized conditions [12] (Fig. 8A and B). The SIM chromatogram obtained from the analysis of 1 mL water sample (code: 811) extracted separately through HybridSPE and Si-SAX are given in Fig. 8. The recoveries of IMPA with HybridSPE and Si-SAX were 98.1 ± 4.5 and 101.1 ± 4.1, respectively. The obtained results were in good agreement with the optimized results in the case of HybridSPE and previously optimized protocol for Si-SAX [12]. In addition, two aqueous samples prepared using tap water at two different concentrations (0.1 and 2.0 ␮g mL−1 ) by two persons were extracted and analyzed in triplicate to ensure the internal validity of the developed method. The results of extraction of PAs (three additional analytes included for internal validation were MPA, IPA and CHMPA) is shown in Table 2. The recoveries of all the PAs were in accordance with the values optimized in this study with the general observation of moderate recoveries of dibasic PAs (MPA, EPA and IPA) and excellent recoveries of monobasic PAs (IMPA, PrEPA, BEPA, PMPA, PIPA and CHMPA). 3.5. Quantification and limits of detection Under the optimized loading, elution, rinsing and modifier conditions, the repeatability, reproducibility, limits of detection (LODs) and limits of quantification (LOQs) were determined for all PAs

6620

P.K. Kanaujia et al. / J. Chromatogr. A 1218 (2011) 6612–6620

Table 2 Percentage recoveries of analytes (RSD in parenthesis) during internal validation of the method under optimized extraction conditions with limits of detection and quantification (obtained with 1 mL sample) in selected ion monitoring mode. S. no.

1. 2. 3. 4. 5. 6. 7. 8. 9. a b

Analytes

IMPA MPA PrEPA EPA BEPA IPA PMPA PIPA CHMPA

Internal validation (recovery %) Spiking 0.1 ␮g mL−1

Spiking 2.0 ␮g mL−1

98.6 (5.6) 87.9 (5.1) 100.3 (6.0) 73. 9 (6.2) 97.8 (5.8) 61.2 (5.3) 109.7 (6.9) 110.4 (7.0) 111.9 (6.0)

93.3 (5.0) 78.5 (5.2) 88.1 (4.9) 59.4 (6.0) 84.1 (5.6) 58.8 (5.2) 109.5 (5.9) 107.2 (6.7) 102.8 (7.2)

Limit of detection (ng mL−1 )a

Limit of quantification (ng mL−1 )b

r2

4.3 5.2 8.0 6.8 5.9 7.3 6.6 5.1 11.0

8.5 9.0 15.0 10.0 9.5 10.0 10.0 8.0 19.0

0.9885 0.9750 0.9867 0.9619 0.9666 0.9813 0.9915 0.9901 0.9882

At signal to noise ratio (S/N) of 3:1. At signal to noise ratio (S/N) of 10:1.

(including three additional analytes included during internal validation). Table 2 shows the LOD and LOQ values of all the PAs calculated with 1 mL sample in selected ion monitoring mode. The analyses were performed in triplicate and the values of recoveries of studied analytes given in Figs. 4–8, and Tables 1 and 2 are percentage mean values. Reproducibility of extraction for all analytes was observed for three consecutive days. The relative standard deviation (RSDs) ranged from 3.6% to 7.1% for all analytes. The linearity in detector response for all analytes was studied in the concentration range of 0.005–50 ␮g mL−1 (in SIM) and showed good correlation with r2 ranging from 0.9619 to 0.9915. Limits of detection were calculated as minimum concentration of analyte which gave the chromatographic peak at signal to noise (S/N) ratio of 3:1 and determined in selected ion monitoring mode (Table 2). It is evident from the LOD and LOQ values, which are much less than 1 ␮g mL−1 in SIM mode, that the sensitivity of the developed method is in accordance with the requirement of verification analysis of CWC [1–5]. 4. Conclusion In this study, for the first time we have demonstrated that zirconia containing HybridSPE material is useful in the selective enrichment of phosphonic acids from aqueous matrix. The selected phosphonic acids are environmental markers of highly toxic nerve agents, and with the use of HybridSPE, they can be selectively extracted from real world aqueous samples; hence the study is important in light of verification analysis of CWC. Application of the developed method to extract IMPA from external aqueous samples resulted in recovery of around 98%. The optimized extraction parameters include use of 2% formic acid, washing of cartridges with 2 mL water and elution with 1 mL of 10% ammonium hydroxide solution in methanol. Under the optimized conditions, recoveries of all PAs were in the range of 80–115%. The extraction efficiency of HybridSPE for PAs is comparable with anion-exchange solid-phase extraction; however, the only advantage with HybridSPE is its selectivity towards PAs. Use of zirconia in selective extraction of organophosphorus compounds has been made easy due to the commercial availability of HybridSPE cartridges. References [1] M. Mesilaakso, Chemical Weapons Convention Chemical Analysis Sample Collection, Preparation and Analytical Methods, Wiley, Chichester, 2005. [2] Criteria for Designation of Laboratories by OPCW-C-I/DEC.61, Organization for Prohibition of Chemical Weapons, The Hague, 22 May 1997.

[3] J. Hendrikse, in: M. Mesilaakso (Ed.), Chemical Weapons Chemical Analysis: Sample Collection, Preparation, and Analytical Methods, Wiley, Chichester, 2005, p. 89. [4] E.W.J. Hooijschuur, A.G. Hulst, Ad.L. De Jong, L.P. De Reuver, S.H. Van Krimpen, B.L.M. Van Baar, E.R.J. Wils, C.E. Kientz, U.A.Th. Brinkman, Trends Anal. Chem. 21 (2002) 116. [5] Convention on the Prohibition of the Development, Production, Stockpiling and Use of Chemical Weapons and their Destruction, Technical Secretariat of the Organization for Prohibition of Chemical Weapons, The Hague, 1997, Accessible through internet http://www.opcw.org. [6] W. Krutzsch, R. Trapp, A Commentary of CWC, Martinus Nijhoff, Dordrecht, 1994. [7] R.L. Maynard, in: T.C. Marrs, R.L. Maynard, F.R. Sidell (Eds.), Chemical Warfare Agents: Toxicology and Treatment, 2nd ed., Wiley, 2007. [8] S.S. Talmage, N.B. Munro, A.P. Watson, J.F. King, V. Hauschild, in: T.C. Marrs, R.L. Maynard, F.R. Sidell (Eds.), Chemical Warfare Agents: Toxicology and Treatment, 2nd ed., Wiley, 2007. [9] M.L. Kuitunen, in: M. Mesilaakso (Ed.), Chemical Weapons Convention Chemical Analysis Sample Collection, Preparation and Analytical Methods, Wiley, Chichester, 2005, p163. [10] D.K. Dubey, D. Pardasani, A.K. Gupta, M. Palit, P.K. Kanaujia, V. Tak, J. Chromatogr. A 1107 (2006) 29. [11] H.S.N. Lee, C. Basheer, H.K. Lee, J. Chromatogr. A 1124 (2006) 91. [12] J. Riches, I. Morton, R.W. Read, R.M. Black, J. Chromatogr. B 816 (2005) 251. [13] V. Tak, D. Pardasani, P.K. Kanaujia, D.K. Dubey, J. Chromatogr. A 1216 (2009) 4319. [14] P.K. Kanaujia, D. Pardasani, A.K. Gupta, R. Kumar, R.K. Srivastava, D.K. Dubey, J. Chromatogr. A 1161 (2007) 98. [15] S.A. Fredriksson, L.G. Hammarstrom, L. Henriksson, H.A. Lakso, J. Mass Spectrom. 30 (1995) 1133. [16] M. Rautio, Recommended Operating Procedures for Sampling and Analysis in the Verification of Chemical Disarmament, Ministry for Foreign Affairs of Finland, Helsinki, 1994. [17] M.Z. Hui, L. Qin, Anal. Chim. Acta 435 (2001) 121. [18] S.L. Moullec, A. Begos, V. Pichon, B. Bellier, J. Chromatogr. A 1108 (2006) 7. [19] V. Pucci, S. Di Palma, A. Alfieri, F. Bonelli, E. Monteagudo, J. Pharm. Biomed. Anal. 50 (2009) 867. [20] A. Gonzalvez, B. Preinerstorfer, W. Lindner, Anal. Bioanal. Chem. 396 (2010) 2965. [21] H. Wan, J. Yan, L. Yu, X. Zhang, X. Xue, X. Li, X. Liang, Talanta 82 (2010) 1701. [22] H.K. Kweon, K. Hakansson, Anal. Chem. 78 (2006) 1743. [23] H.J. Cha, N.H. Kim, E.K. Jeong, Y.C. Na, Bull. Korean Chem. Soc. 31 (2010) 2322. [24] J. Nawrocki, M.P. Rigney, A. McCormick, P.W. Carr, J. Chromatogr. A 657 (1993) 229. [25] J. Li, H.Y. Qi, Y.P. Shi, J. Chromatogr. A 1216 (2009) 5467. [26] D. Du, X. Ye, J. Zhang, Y. Zeng, H. Tu, A. Zhang, D. Liu, Electrochem. Commun. 10 (2008) 686. [27] G. Liu, J. Wang, R. Barry, C. Petersen, C. Timchalk, P.L. Gassman, Y. Lin, Chem.-Eur. J. 14 (2008) 9951. [28] L. Xu, H.K. Lee, Anal. Chem. 79 (2007) 5241. [29] L. Silvestro, M.C. Gheorghe, I. Tarcomnicu, S. Savu, S.R. Savu, A. Iordachescu, C. Dulea, J. Chromatogr. B 878 (2010) 3134. [30] W. Zeng, Y. Xu, M. Constanzer, E.J. Woolf, J. Chromatogr. B 878 (2010) 1817. [31] G.M. Kosolapoff, L. Mair, Organic Phosphorus Compounds, vol. 7, Wiley, New York, 1976, p. 9, 22.