On-line SPE-LC/MS/MS to Detect Organonitrogen and Triazine Pesticides at 10ng/L in Drinking Water

Transcription

1 application note On-line SPE-LC/MS/MS of Pesticides On-line SPE-LC/MS/MS to Detect Organonitrogen and Triazine Pesticides at 10ng/L in Drinking Water API 3200 LC/MS/MS System Overview The feasibility of on-line Solid Phase Extraction (SPE) coupled to Liquid Chromatography with tandem Mass Spectrometry (LC/MS/MS) to detect organonitrogen and triazine pesticides at concentrations of 10ng/L in drinking water was studied. Sensitivity, reproducibility, linearity and recovery in tap water samples and water samples spiked with humic acid were investigated to evaluate sample preparation efficiency and possible matrix effects. Introduction Current legislation for drinking waters in the UK requires a lower Limit of Quantitation (LOQ) for pesticides of 100ng/L (ppt) with a Coefficient of Variation (CV) below 12.5% and a Limit of Detection (LOD) of 20ng/L. 1 In order to achieve these criteria, traditional methods usually use different sample pre-treatments for each pesticide class, such as different procedures, e.g. SPE, to concentrate the analytes and remove matrix components. After extraction and specific chemical derivatization processes several different analytical techniques are often needed to perform the actual detection and determination of pesticides at these low levels. If the analytes are thermally stable pesticides, then they are analyzed by Gas Chromatographic (GC) techniques whilst thermally labile compounds are usually determined by Liquid Chromatography (LC). This paper describes an automated SPE sample clean-up and pre-concentration step utilizing large volume injection (5mL of untreated water samples) prior to LC/MS/MS analysis. With this approach, all investigated 26 pesticides are first trapped on an SPE cartridge and then back flushed onto a reverse phase LC column. The pesticides are then separated on the analytical column using a binary LC gradient and are determined on the API 3200 LC/MS/MS System in Multiple Reaction Monitoring (MRM).

2 Trapping the analytes on an on-line SPE cartridge has three major advantages over methods employing direct on column injection: 1. Analytes can be rapidly enriched on the SPE column, thus enabling detection at lower LOD. 2. Matrix components, such as humic acid or inorganic salts are washed away prior to the analytical column. The longevity of the column life is therefore extended whilst maximizing assay robustness. Table 1. LC gradient used for separation of organonitrogen and triazine pesticides Step Total Time (min) A (%) B (%) The reproducibility of the analysis is improved by automation and thus minimizing human errors during sample preparation. Table 2. MRM transitions with compound dependant MS/MS parameters for each detected pesticide Compound Q1 Q3 DP in V CE in V Dwell Time (ms) Ametryn Atraton Atrazine Desmetryn Epoxiconazole EPTC Fenpropidin Fenpropimorph Flusilazole Flutriafol Metazachlor Pendimethalin Prometon Prometryn Propachlor Propazine Propiconazole Propyzamide Simazine Simetryn Tebuconazole Terbuthylazine Terbutryn Triadimefon Tri-allate Trietazine

3 Figure 1. MRM transitions illustrating differences in the relative sensitivities of 10ng/L Terbutryn (S/N = 498), Atrazine (S/N =65) and for EPTC (S/N =9) calculated with peak-to-peak algorithm Experimental HPLC 5mL of tap water were loaded onto a Phenomenex Strata C18 on-line SPE cartridge using a CTC HTC PAL autosampler (fitted with a 5mL syringe, a 5mL loop and a 32 x 10mL sample rack). The trap cartridge was fitted into load position A of a high pressure, Valco 10 port switching valve. An Agilent 1200 isocratic pump was used to load the sample at 3mL/min onto the trapping cartridge. After 2.5 minutes the valve was switched to position B where an Agilent 1200 Binary pump was used to provide the gradient elution of pesticides. For the separation of pesticides an Agilent Zorbax 1.8µm 4.6x50mm with a flow rate of 600µL/min at ambient temperature was used. The gradient profile is shown in Table 1 where mobile phase A was 2mM ammonium formate with 0.1% formic acid in water and mobile phase B was 2mM ammonium formate with 0.1% formic acid in methanol. After 14.5 minutes the switching valve was switched back to the loading position to analyze the next sample. Local tap water without any pretreatment was spiked with pesticides and injected directly onto the SPE-LC/ MS/MS system. To study matrix effects commercially available humic acid (Sigma-Aldrich) was dissolved in 0.2% formic acid at a level of 10 parts per million (ppm) and pesticides were spiked into this matrix. Mass Spectrometry An API 3200 LC/MS/MS System with a Turbo V Source and Electrospray Ionization probe was used for all experiments. Curtain gas 25psi Gas1 70psi Gas2 30psi CAD gas 4 Temperature 700 C IonSpray voltage 5500V

4 Table 3. Retention times, Signal-to-Noise, reproducibility and linearity data for each detected pesticide in tap water without the use of internal standards Compound Retention time S/N of 10ng/L in Tap Water Standard Deviation of 10ng/L (n=6) %CV of 100ng/L (n=6) r Value for Calibration Line (10-500ng/L) Ametryn Atraton Atrazine Desmetryn Epoxiconazole EPTC Fenpropidin Fenpropimorph Flusilazole Flutriafol Metazachlor Pendimethalin Prometon Prometryn Propachlor Propazine Propiconazole Propyzamide Simazine Simetryn Tebuconazole Terbuthylazine Terbutryn Triadimefon Tri-allate Trietazine Detected MRM transitions with compound dependant parameters which were optimized using Analyst Software Quantitative Optimization for each pesticide are displayed in Table 2 [Declustering Potential (DP) and Collision Energy (CE)]. During any chromatographic measurement small drifts in retention time will inevitably be caused by gradual column degradation or small changes in temperature and mobile phase composition in between running samples. When incorporating time windows (periods or segments) into the chromatogram a small change in retention may result in pesticides falling outside of the window and therefore going undetected during the MS/MS acquisition. This phenomena becomes more profound as the number of compounds being monitored and/or windows within a given experiment increases. Therefore, the ion path of the API 3200 System using a Linear Accelerator (LINAC ) Collision Cell was set to constantly monitor all pesticides throughout the entire chromatogram. In addition with this approach, it is possible to subsequently add more pesticides to the suite of compounds screened for without the prior need to investigate the elution profile of the added analytes.

5 Figure 2. Calibration lines of the weakly ionizing organonitrogen pesticide EPTC (top) with r=0.9993, and the strongly ionizing triazine pesticide Terbuthylazine (bottom) with r= over a linear range of 10ng/L to 500ng/L Results and Discussion The analyte sensitivity varied markedly from pesticide to pesticide and the corresponding S/N (peak-to-peak) ratio for the pesticides varied from 498:1 (Terbutryn) down to 9:1 (EPTC), see Figure 1 and Table 3. Therefore, this qualifies the need to use 5mL of sample for SPE so as to include weakly ionizing compounds, such as EPTC, Flutriafol and Trietazine within the same sample injection. Reproducibility was assessed by the analysis of tap water samples which had been spiked at two levels 10ng/L and 100ng/L (the LOQ needed for drinking water monitoring). Table 3 shows excellent standard deviations and coefficients of variation (%CV) for all pesticides. Linearity was studied over a concentration range form 10 to 500ng/L. Correlation coefficients were found to be greater than for all calibration lines when a linear fit with 1/x weighting was applied as presented in Figure 2 and Table 3. No internal standards were used when reproducibility and linearity were investigated.

6 Figure 3. Carry over assessment using an injection of a 500ng/L standard of Fenpropidin and EPTC followed by an injection of blank water directly afterwards (corresponding MRM transitions for the same compounds are shown for the two injections) To further test the robustness of the method 70 injections were made of the same 100ng/L standard of Terbutryn over a 20 hour period which gave a CV of 6.6%. A 500ng/L standard was injected followed by a blank water sample to assess possible carry over using the developed method. Figure 3 shows a typical result for 2 pesticides and indicates that carry over is less than 0.2%. Figures 4 shows examples of Total Ion Chromatograms (TIC) obtained from the direct injection of different types of spiked water. In these examples it can be seen that there is a very low effect on the intensity of the response observed when tap water and water containing a high level of humic acid is compared to distilled water. The recoveries of all pesticides in tap water ranged from 77 to 115% (Table 4). Also when pesticides were spiked into water containing 10ppm of humic acid there was a decrease in sensitivity but recoveries compared to distilled water were still over 67%. The effect of humic acid was to decrease recovery in comparison to tap water samples by typically less than 15% (the only exception being Tri-allate with 20%). Conclusion The method described provides a simple, robust and cost effective approach for the routine analysis of organonitrogen and triazine pesticides in potable waters by on-line SPE-LC/MS/ MS. The use of on-line SPE eliminates the additional cost and time required for traditional SPE methods and offers improved robustness and reproducibility. Limits of Detection (LOD) have been calculated using both peak-to-peak measurements and also by calculating the LOD as 4.65 times the standard deviation of a 10ng/L spike. Both calculations show that all the pesticides

7 Figure 4. Total Ion Chromatograms of a 100ng/mL spike of organonitrogen and triazine pesticides in distilled water (top), tap water (middle) and 10ppm humic acid (bottom) to assess ion suppression effects tested can be detected below 10ng/L with good CV below 8% at 100ng/L and excellent linearity from 10 to 500ng/L. The sensitivity of the method was only slightly effected by the type of water sample analyzed and has been shown to be sensitive enough for all the pesticides in potable waters. In water samples with a high total organic carbon, such as humic acid at 10 parts per million, some ion some suppression results but it was less than 20%. The suitability of the application of the API 3200 LC/MS/MS System for the routine analysis of these pesticides has therefore been demonstrated at the required levels and precisions. Currently, the application of this technique to the determination of other pesticides such as acidic herbicides and organophosphorus pesticides is being investigated. Author Stephen J. Lock (Applied Biosystems, Warrington, Cheshire, UK) Reference 1 A manual of Analytical Quality control for the water industry (NS30), Drinking Water Inspectorate, UK

8 Table 4. Recovery data at 100ng/L for each detected pesticide in tap water and water spiked with 10ppm humic acid Compound Peak Area of 100ng/L in Distilled Water Peak Area of 100ng/L in Tap Water Peak Area of 100ng/L in Water with 10ppm Humic Acid Recovery (%) from Tap Water Recovery (%) from Water with 10ppm Humic Acid Effect of Humic Acid on Recovery (%) Ametryn 2.68E E E Atraton 2.68E E E Atrazine 1.53E E E Desmetryn 4.32E E E Epoxiconazole 3.87E E E EPTC 1.57E E E Fenpropidin 1.03E E E Fenpropimorph 4.59E E E Flusilazole 2.59E E E Flutriafol 6.93E E E Metazachlor 5.19E E E Pendimethalin 1.24E E E Prometon 6.82E E E Prometryn 8.61E E E Propachlor 2.59E E E Propazine 3.97E E E Propiconazole 2.18E E E Propyzamide 1.31E E E Simazine 1.52E E E Simetryn 4.87E E E Tebuconazole 2.09E E E Terbuthylazine 9.68E E E Terbutryn 2.45E E E Triadimefon 1.50E E E Tri-allate 4.13E E E Trietazine 2.74E E E For Research Use Only. Not for use in diagnostic procedures. Applera Corporation is committed to providing the world s leading technology and information for life scientists. Applera Corporation consists of the Applied Biosystems and Celera Genomics businesses. Copyright 2007, Applera Corporation and MDS Inc. All rights reserved. Information subject to change without notice. Applera, Applied Biosystems and AB (Design) are registered trademarks of Applera Corporation or its subsidiaries in the US and/or certain other countries. API 3200 and Turbo V are trademarks and Analyst and LINAC are registered trademarks of Applied Biosystems/MDS SCIEX, a joint venture between Applera Corporation and MDS Inc. All other trademarks are the sole property of their respective owners. Printed in the USA. 07/2007 Publication 114AP63-01 Headquarters 850 Lincoln Centre Drive Foster City, CA USA Phone Toll Free International Sales For our office locations please call the division headquarters or refer to our Web site at