You are on page 1of 1380

Solutions that meet your demands for

food safety testing

Excellent choices for food applications

Table of Contents
The table of contents on the right has been linked to the individual sections in this compendium. Click on the text to jump to a specific section. If you prefer to search for words or phrases used in any of the enclosed applications notes, click on the Search button below to open Acrobats Search window.

Solution Segments
Pesticides and Residues
Pesticides and Herbicides Antibacterial Drug Residues

Contaminants
Acrylamides Food Packaging PCBs Toxins Trace Metals

Search

Natural Compounds and Additives


Additives Amino Acids Carbohydrates Fats and Oils Flavors and Fragrances Miscellaneous Proteins Vitamins

Bioanalysis
Applications

Productivity Tools
Applications

Index

Pesticides & Residues Pesticides and Herbicides

> Return to Table of Contents > Search entire document

Replacing Multiple 50-Minute GC and GC-MS/SIM Analyses with One 15-Minute Full-Scan GC-MS Analysis for Nontargeted Pesticides Screening and >10x Productivity Gain Application
Food Safety

Author
Chin-Kai Meng and Mike Szelewski Agilent Technologies 2850 Centerville Road Wilmington, DE 19808

Introduction
To have a plentiful food supply, most fruits and vegetables are treated with pesticides (insecticides, fungicides, herbicides, etc.) to protect primarily against insects, molds, and weeds. Therefore, in order to ensure food safety, the food supply is frequently monitored for pesticide residues. Nowadays, the pesticide monitoring is expanding beyond food, for example, to botanical dietary supplements. The analytical challenge to monitor (identify and quantify) trace multiresidues requires an effective and universal extraction and analysis method for maximum productivity and efficiency. Up to now, the trade-off in analysis has been between sensitivity and confirmation. Element-selective gas chromatograph (GC) detectors, such as the flame photometric (FPD), electron capture (ECD), electrolytic conductivity (ELCD), and halogen selective (XSD) detectors, provide excellent selectivity and sensitivity; however, they lack the capability to identify. On the other hand, mass spectrometry (MS) is capable of identifying an analyte by fullscan library match or multiple target and qualifier ion ratios from selected ion monitoring (SIM). However, MS sometimes lacks the selectivity to

Abstract
Pesticide analysis of fruits and vegetables requires finding trace-level residues in complex matrices. Up to now, the typical trade-off is between sensitivity and confirmation. Therefore, multiple injections are needed for screening and confirmation using gas chromatography with mass spectrometry (GC-MS) or GC-MS in combination with GC with element-selective detection. With the recent introduction of hardware and software tools, for example, capillary flow three-way splitter, trace ion detection, and deconvolution, a 15-minute fast analysis can match the results obtained from three injections of approximately 50 minutes each. A table comparing the results from the Food and Drug Administration/Center for Food Safety and Applied Nutrition (FDA/CFSAN) procedure using the traditional multi-instrument approach and the new Agilent single injection approach shows that Agilent's fast analysis is capable of finding all the target analytes in less than one-tenth of the current FDA/CFSAN total analysis time.

find target analytes in a complex matrix full of interferences and chemical background. Analyte spectra are sometimes overwhelmed by similiar ions contributed from the coextractives in the matrix that prevent the analyte of interest from being identified or confirmed. The compromise and the typical approach are to use selective GC detector(s) to flag potential target analytes and use MS SIM for confirmation. For instance, many laboratories screen food samples for semivolatile pesticides using the ECD or ELCD (or XSD) for organohalogen, FPD or pulsed FPD (PFPD) for organophosphorus, and NPD for nitrogen-containing targets [1 5]. Any found targets are further confirmed by GC-MS/SIM. In addition, other procedures have used GC-MS/SIM entirely for the screening of pesticides in foods [6 8]. In most of these procedures, multiple injections are needed to identify hundreds of compounds at the detection limit in the low parts-per-billion (ppb) levels. To improve the efficiency and increase the productivity of screening for all of these pesticides, the challenge is to reduce the GC-MS or the combination of GC and GC-MS analysis times. There are several hundred pesticides typically used in the world, and each country has its own pesticide tolerance levels for different agricultural commodities. This presents another analytical challenge in multiresidue monitoring: to develop a nontargeted procedure to identify pesticides at trace levels in different food matrices. These challenges are met by the recent introduction of hardware and software tools, including GCMS, capillary flow three-way splitter, trace ion detection, and deconvolution reporting software (DRS). The splitter allows multiple GC as well as MS signals to be acquired from a single injection for productivity gains (from three injections down to one). Trace ion detection minimizes noise on the signal and DRS separates target analyte ions from matrix background ions. Several sample extracts were analyzed by the current Food and Drug Administration/Center for Food Safety and Applied Nutrition (FDA/CFSAN ) multiple injection process and this new Agilent pesticide system. With DRS, the demand for chromatographic resolution is minimum; therefore, the Agilent system was running the analysis at a 3x faster speed (one-third the analysis time) to further increase productivity. A table comparing the

results from the current FDA/CFSAN multi-instrument approach and the new Agilent single-injection approach shows that not only is Agilents fast analysis capable of finding all the target analytes, but it is also accomplished in just one-tenth of the current FDA/CFSAN total analysis time.

Experimental
Sample Preparation Sample extracts of fresh produce were prepared by FDA based on modifications of the QuEChERS protocols [9, 10]: Homogenize 1 to 2 kg of sample 15 g sample + 15 mL 1% AcOH/ACN, homogenized Add 6 g MgSO4 and 2.5 g NaOAc, shake vigorously for 1 minute, and centrifuge Transfer ~15 mL + 0.5 g C-18 + 1.2 g MgSO4, shake, and centrifuge for 5 min at 3,000 rpm Transfer ~12 mL + 0.4 g PSA + 0.2 g GCB + 1.2 g MgSO4, vortex Add 4 mL toluene, shake, and centrifuge Transfer 6 to 8 mL, evaporate and bring to volume with toluene, add I.S. Add MgSO4, vortex and centrifuge, transfer to ALS vials GC and GC-MS analysis Sample preparation of dried ginseng powder is similar to that used for fresh produce, but smaller sample sizes (2 g) were used [11]. Capillary Flow Three-Way Splitter One of the capillary flow devices is a three-way splitter, which consists of two half plates bonded together (diffusion bonding) to form a plate with the etched flow channels inside. The splitter is only 6.5 cm tall and 3 cm wide and is mounted on the side of the oven wall (see Figure 1). The low thermal mass minimizes cold spots and peak broadening. All capillary flow devices use metal column ferrules, have extremely low dead volumes, are inert, and do not leak, even after many oven cycles.

To MSD

To FPD To ECD

Aux EPC in

Column in

different inlets and outlets. Aux pressure can be either an inlet (for the splitter flow restrictors connected to different detectors) or an outlet (for the analytical column). A graphical user interface makes the configuration easy to set up. Once all the columns and restrictors are configured, the backflush can be executed easily. Backflush Traditional bakeout step for removing late eluters could be very time consuming, or even as long as the analysis time depending on the matrix. Backflush is a simple technique to remove high boilers from the column faster and at a lower column temperature to cut down analysis time and increase column lifetime. Capillary flow devices (in this case, a three-way splitter) also provide backflush [13, 14] capability. Backflush is a term used for the reversal of flow through a column such that sample components in the column are forced back out the inlet end of the column. By reversing column flow immediately after the last compound of interest has eluted, the long bake-out time for highly retained components can be eliminated. Therefore, the column bleed and ghost peaks are minimized, the column will last longer, and the MS ion source will require less frequent cleaning. The split vent trap may require replacement more frequently than usual. Figures 2 and 3 are two screen shots from the MSD ChemStation software, providing a summary of the backflush operation. In Figure 2, the column and three restrictor dimensions and respective detectors are shown (the setup came from the column configuration section). For MSD, the user can choose the vacuum pump installed on the system. This information will be used to calculate if the backflush is within the system flow limits. By clicking on the Evaluate button, the screen shown in Figure 3 appears, listing the maximum flow for each detector and the void volumes for a certain backflush time. In this example, Aux pressure is at 60 psi, inlet is at 1 psi, and oven is at 280 C. The backflushing flow is shown to be 8.66 mL/min, and the void time is shown to be 0.16 min. Therefore, backflushing for 2.5 minutes will send 15.6 void volumes through the column. This is useful for developing the backflush method. Figures 2 and 3 simplify the setup and development of a backflush method.

Etched flow channel inside two diffusion bonded plates Capillary tube connection via metal ferrule

Figure 1.

Flow diagram of the three-way splitter. The picture shows the splitter mounted on the oven wall.

The three-way splitter enhances productivity by splitting column effluent proportionally to multiple detectors: MSD, dual flame photometric detector (DFPD) and micro-electron capture detector (ECD). Therefore, two GC detector signals can be acquired together with the MS data (both SIM and scan signals if desired) from one injection [12]. The exit end of the analytical column is installed into one of the four ports on the splitter using a metal ferrule. The other three ports are connected to three detectors via restrictors (deactivated capillary tubing) of varying diameter and length to set the split ratio among the three detectors. Restrictors are sized for 1:1:0.1 split ratio in favor of MSD and DFPD (ECD has 1/10 of the flow to MSD), with similar hold-up times. The splitter uses auxiliary (Aux) electronic pneumatics control (EPC) for constant pressure makeup flow. The makeup gas (Aux pressure 6) at the splitter is fixed at 3.8 psi to maintain the split ratio throughout the run. This multisignal configuration provides full-scan data for library searching, SIM data for trace analysis, DFPD (phosphorus or sulfur mode), and ECD data for excellent selectivity and sensitivity from complex matrices. The trade-off is the decrease of analyte concentration in any detector due to the flow splitting and the additional makeup gas from the splitter. An analyte would have similar retention times in all three detectors. Therefore, the GC data can be used in two ways: first, to confirm the presence of target analytes found by the MSD deconvolution reporting software (DRS), and second, to highlight potential target compounds to be further confirmed by MSD. With the new 7890A GC software, up to six columns/ restrictors can be configured/assigned to

Figure 2.

Backflush setup in ChemStation.

Figure 3.

Automated backflush calculations in ChemStation.

Deconvolution

Another useful feature in Figure 3 is the warning, shown as highlighted yellow cells. In this example, setting the backflush pressure to 60 psi sends more than the allowable flow (60 mL/min) to the FPD. Therefore, the backflush pressure setting and the actual flow value to FPD are shown in yellow as warnings. Although the system will accept the setup, the high flow may cause consequences in the analysis, for example, flameout. Trace Ion Detection Trace ion detection [15] is a filtering algorithm to smooth peaks. This filtering is an advanced form of averaging used to remove the noise riding on the signal. The implications from TID are typically a slight loss in peak height and some peak broadening. The default setting in ChemStation for TID is off. It should be turned on for any analysis that uses deconvolution and has more than six sampling points across a peak. TID provides better signal-to-noise ratios and helps deconvolution to confirm target compounds as shown in the Results section. Deconvolution In GC/MS, deconvolution is a mathematical technique that separates overlapping mass spectra into deconvoluted spectra of the individual components. Figure 4 is a simplified illustration of this process. Here, the total ion chromatogram (TIC) and apex spectrum are shown on the left. In a complex matrix, a peak may be composed of multiple overlapping components and matrix background ions; therefore, the apex spectrum is actually a composite of these constituents. A mass spectral library search would give a poor match, at best, and certainly would not identify all of the individual components that make up the composite spectrum. The deconvolution process groups ions whose individual abundances rise and fall together within the spectrum. The deconvolution process first corrects for the spectral skew that is inherent in quadrupole mass spectra and determines a more accurate apex retention time of each chromatographic peak. As illustrated in Figure 4, deconvolution produces a cleaned spectrum for each overlapping component. These individual spectra can be library searched with a high expectation for a good match. Deconvolution significantly reduces chromatographic resolution requirements, allowing much shorter analysis times.

TIC and spectrum


TIC

Deconvoluted peaks and spectra


Component 1 extracted spectrum

Component 2 extracted spectrum

Component 3 extracted spectrum

Library search each component to identify

Figure 4.

Deconvolution process of three overlapped peaks.

Agilent Deconvolution Reporting Software (DRS) utilizes the AMDIS deconvolution program from the National Institute of Standards and Technology (NIST), originally developed for trace chemical weapons detection in complex samples [16]. DRS presents the analyst with three distinct levels of compound identification: (1) ChemStation, based on retention time and four ion agreement; (2) AMDIS, based on cleaned spectra full spectral matching and expected retention time window as a qualifier; and (3) NIST05 search using a >163,000compound library [17, 18]. In this application, both the ChemStation quantitation database and the AMDIS library have the same 926 entries. These entries include pesticides, numerous metabolites, endocrine disruptors, important PCBs and PAHs, certain dyes, synthetic musk compounds, and several organophosphorus fire retardants [18]. The AMDIS software, shipped with the NIST05 Library CD-ROM, is also capable of deconvoluting selected ion monitoring (SIM) data [19], while previous AMDIS revisions were not. Testing has shown that proper compound identification requires four ions per compound. All Agilent DRS databases are retention time locked and have both full-scan and SIM libraries for AMDIS. Instrument Method The system used for this study consists of an Agilent 7890A GC with split/splitless inlet, a threeway splitter, ECD, DFPD, and 5975 MSD. For a detailed description of SIM/scan and the splitter system configuration, please refer to the experimental section of reference [12]. See Table 1 for hardware detail and settings.

Table 1. Gas Chromatograph, Mass Spectrometer, and Three-Way Splitter Operating Parameters GC Injector Syringe size Injection volume Solvent A wash Solvent B wash Sample wash Sample pump Plunger speed Inlet Mode Inlet temperature Pressure Agilent Technologies 7890A with 240V fast oven option Agilent Technologies 7683 10 L 1 L 1 (pre), 3 (post) 1 (pre), 3 (post) 0 4 Fast EPC split/splitless Splitless 250 C ~24.4 psi (chlorpyrifos methyl RT locked to 5.531 min, 3x speed) constant pressure mode 50.0 mL/min 2 min 3 mL/min Switched Off Helium Helix double taper liner, deactivated, p/n 5188-5398 C /min 75 9 24 13.96 min 1.0 min 2.5 min 280 C Agilent Technologies HP-5MS, p/n 19091S-431 15.0 m 0.25 mm 0.25 m Constant pressure RT locked to chlorpyrifos methyl at 5.531 min, 3x analysis speed 3.5 mL/min Aux pressure 6 3.8 psi (Aux EPC pressure to three-way splitter), helium gas 280 C 2.5 min 1 psi 60 psi (column outlet) ECD 300 C 60.0 mL/min Nitrogen 20 Hz Dual FPD 250 C Final (C) 70 150 200 280 Hold (min) 0.67 0 0 3.33 Hydrogen flow Air flow Const Col + Makeup Make gas type Lit offset Data rate Transfer line AUX Thermal 1 AUX Pressure 6 Gas type Initial pressure Backflush pressure MSD Tune file Mode Solvent delay EM voltage Low mass High mass Threshold Samples Scans/sec Quad temp Source temp Three-way splitter Pressure Split ratio MSD restrictor DPFD restrictor ECD restrictor Flow to MSD Flow to DFPD Flow to ECD Makeup (Aux 6) Software GC/MSD ChemStation MS Libraries Agilent part number G1701EA (version E.01.00 or higher) NIST05a mass spectral library (Agilent part number G1033A) Agilent RTL Pesticide and Endocrine Disruptor Libraries (926 entries) in Agilent and AMDIS formats (part number G1672AA) Automated Mass Spectral Deconvolution and Identification Software (AMDIS_32 version 2.65 Build 116.66) NIST MS Search (version 2.0d or greater) (comes with NIST'05a mass spectral library Agilent part number G1033A) Agilent part number G1716AA (version A.03.00 or higher) 75.0 mL/min 100.0 mL/min 60.0 mL/min Nitrogen 2.00 20 Hz 250 C MSD transfer line, 280 C Three-way splitter Helium 3.8 psi 60 psi Agilent Technologies 5975C MSD Atune.u Scan 1.50 min Atune voltage 50 amu 550 amu 0 2 2.91 150 C 230 C Agilent 7890A Option 890, installed during factory assembly 3.8 psi (Aux pressure 6 setting) 1:1:0.1 MSD:DFPD:ECD 1.444 m 0.18-mm id deactivated fused silica tubing, p/n 160-2615-10 0.532 m 0.18-mm id deactivated fused silica tubing, p/n 160-2615-10 0.507 m 0.10-mm id deactivated fused silica tubing, p/n 160-2635-1 3.43 mL/min (at 70 C), 1.53 mL/min (at 280 C) 3.43 mL/min (at 70 C), 1.53 mL/min (at 280 C) 0.343 mL/min (at 70 C), 0.153 mL/min (at 280 C) 3.19 mL/min (at 70 C), 1.52 mL/min (at 280 C)

Purge flow Purge time Septum purge flow Septum purge mode Gas saver Gas type Liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Runtime Oven equilib time Post-run time Post-run temperature Column Length Diameter Film thickness Mode

Nominal initial flow Outlet Outlet pressure Backflush (post-run) Oven Time Inlet Aux pressure 6 Front detector Temperature Const col + makeup Make gas type Data rate Back detector Temperature 6

Deconvolution software Library searching software

Deconvolution reporting software

Results and Discussion


Backflush Example Blank runs, made after separate milk analyses with different backflush (BF) times, are shown in Figure 5. The top TIC is a blank run after a milk extract analysis stopped at 42 minutes and the system backflushed for 1 minute. The next TIC is a blank run after another milk extract analysis stopped at 42 minutes and backflushed for 2 minutes and so on for the other five TICs. It is interesting to confirm graphically that the latest eluters disappeared from the TIC earliest in backflushing.
100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0

Trace Ion Detection Figure 6 compares the signals when TID is on and off. Visually, it is obvious that TID smoothes the noise riding on top of the signal. When TID was on, Atrazine was successfully identified by AMDIS. When TID was off, Atrazine was not found by AMDIS and resulted in a false negative. Figure 7 compares TID on and off for two different analytical conditions of the same ginseng extract. On the right, the fast (3x) analysis was a 1-L splitless injection with TID on. The analyte Diazinon was found by AMDIS with a peak width less than 5 seconds. On the left side, the normal (1x)

BF for 1 min

BF for 2 min

BF for 3 min

BF for 4 min

Note: late eluters were backflushed out first

BF for 5 min

BF for 6 min

BF for 7 min
10.00 20.00 30.00 40.00 50.00

Column is clean
60.00 70.00 80.00

Figure 5.

Differences in blank runs as the result of seven different backflush times. Atrazine found by AMDIS

TID on

TID off

Atrazine not found by AMDIS Figure 6. The power of deconvolution with TID for atrazine, from AMDIS. 7

analysis was a 5-L cold splitless injection using a PTV with TID off. The 9-second-wide Diazinon was not found by AMDIS, also a false negative. Both examples show that TID is a very useful feature for trace target analysis. Benefits of TID: Improves the signal-to-noise ratio AMDIS is more thorough in identifying components, resulting in fewer false positives Improves library match quality Improves area repeatability, resulting in more reliable quantitation DRS and Splitter Figures 8 and 9 are DRS reports for ginseng and peach extracts with pesticides highlighted (for a detailed explanation of the report, please refer to references [17] and [20]). Figures 10 and 11 show simultaneously collected GC and MS signals (RT locked) for the corresponding ginseng and peach extracts from a three-way splitter. The presence of the GC peaks from the ECD and FPD (P) helps confirm the targets reported by DRS. Each run is finished at 15 minutes using the 3x speed and a 240V oven. With deconvolution, less peak resolution is required for compound identification. A 4minute backflush is added after the run to make sure that the column is clean to maintain the next runs locked RTs for all peaks.
PTV, 1x speed, 5 L injection No TID, Diazinon not found (false negative)

Deconvolution Figures 12 through 15 show the results from AMDIS. There are three spectra for each target compound found by AMDIS. The top window shows the spectrum (scan) from the TIC. This is the only spectrum that would be available for library searching without deconvolution obviously quite useless. The middle window shows the deconvoluted spectrum and the bottom window is the target compounds spectrum in the library. The compound confirmation can be done easily and with confidence by visually comparing the bottom two spectra. The power of deconvolution is appreciated while comparing the top two spectra (the raw scan and the spectrum hidden in the raw scan). It is easy to further confirm the hits found by deconvolution. In Figure 9, four pesticides found by AMDIS in the peach extract have a match factor of about 80 or lower. The four pesticides are Cabaryl, Captan, Propiconazole, and Fenbuconazole. A SIM method of these compounds was set up to analyze the peach extract. By selecting the proper AMDIS library (full-scan or SIM), DRS can process full-scan as well as SIM data files [19]. Figure 16 is the DRS report of the peach SIM analysis. The high match factor (99 or higher) and the small RT difference of all targets found by AMDIS confirm the presence of all compounds.

Splitless, 3x speed, 1 L injection with TID, Diazinon found

Figure 7.

The power of deconvolution with TID for diazinon.

Figure 8.

DRS report for ginseng with pesticides highlighted.

Figure 9.

DRS report for peach with pesticides highlighted.

Ginseng
1e+07 8000000 6000000 4000000 2000000 2.00 3.00 4.00 3.524 5.00 6.00 7.00 8.00 8.699 9.00 10.00 9.413 11.00 12.00 13.00

TIC

9e+07 8e+07 7e+07 6e+07 5e+07 4e+07 3e+07 2e+07 1e+07

Tetrachloro-m-xylene (ISTD) Chlorthaldimethyl


4.501 6.248 2.00 1.660 3.00 4.00 3.665 5.00 6.00 7.005 7.00 8.00

8.290

7.783

8.744 8.992 9.093 9.059 9.265 9.528 9.646 9.784

ECD Azoxystrobin

9.00

10.00

11.00

12.00

13.00

6000000 5000000 4000000 3000000 2000000 1000000

Tributylphosphate (ISTD)
9.413

Diazinon
4.831 9.313 7.641

FPD (P)

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

11.00

12.00

13.00

Figure 10. Simultaneous display of MSD and GC selective detector signals for ginseng.

2200000 1800000 1400000 1000000 600000 200000 2.00 3.5e+07 3e+07 2.5e+07 2e+07 1.5e+07 1e+07 5000000 2.00 5000000 4000000 3000000 2000000 1000000 2.00 3.00 4.00 9.358 4.773 5.00 6.00 7.00 8.00 9.00 10.00 9.826 9.910 2.703 4.696 3.932 3.00 4.00 3.641 1.838 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00

Peach TIC

11.00

12.00

13.00

Tetrachloro-m-xylene (ISTD)
3.519

Carbaryl
5.640 5.00

7.093 6.680

Endosulfan(alpha) 7.556 Phosmet 9.538 Captan


9.081 8.00 9.00 9.523 10.559 10.00 11.00

ECD

12.382 12.00 13.00

6.00

7.00

Tributylphosphate (ISTD)

Phosmet FPD (P)

11.066 11.00 12.00 13.00

Figure 11. Simultaneous display of MSD and GC selective detector signals for peach. 10

Scan at 12.299 min

Deconvoluted/extracted spectrum

Library spectrum Azoxystrobin

Figure 12. Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of azoxystrobin found in ginseng, from AMDIS.

11

Scan at 5.615 min

Deconvoluted/extracted spectrum

Library spectrum Carbaryl

Figure 13. Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of cabaryl found in peach, from AMDIS.

12

Scan at 10.776 min

Deconvoluted/extracted spectrum

Library spectrum Fenbuconazole

Figure 14 Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of fenbuconazole found in peach, from AMDIS.

13

Scan at 8.934 min

Deconvoluted/extracted spectrum

Library spectrum Endosulfan sulfate

Figure 15 Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of endosulfan sulfate found in tomato, from AMDIS.

14

Figure 16. DRS report from the SIM analysis of peach. Please refer to reference 18 for the explanation of the fictitious CAS number assigned to Propiconazole-II (999048032).

Comparison of Incurred Samples The current approach at FDA/CFSAN is to find a wide suite of organohalogen and organophosphorus pesticide residues. This requires four injections (GC-MS/SIM and GC-ELCD for organohalogen and GC-MS/SIM and GC-FPD for organophosphorus screening) of approximately 50-minutes runtime each (total runtime = 200 minutes). Table 2 shows that FDA found several target compounds in three extracts as well as quantitation
Table 2.

results from both GC and MS. In comparison, using the new tools (splitter, TID, and deconvolution) found as many target compounds and a few more in just one short (15-minute) full-scan analysis. The three-way splitter was used to get selective GC signals (ECD and FPD) for confirmation purposes. Due to column effluent splitting to three detectors (1:1:0.1), the MSD is getting less than half of the amount injected. FDA/CFSAN GC and GC/MS/SIM analyses for organohalogen monitor-

Comparison of the Agilent Pesticide System Results with the FDA Results Agilent DRS (full scan/TID) FDA (FPD, ELCD, SIM) Diazinon (FPD, SIM) GC-FPD or ELCD 25 3 ppb GC-MS/SIM 25 2 ppb

Ginseng

Diazinon Chlorthal-dimethyl Azoxystrobin Carbaryl Captan Endosulfan (alpha) Phosmet Propiconazole I and II Fenbuconazole Chlorothalonil Endosulfan (alpha) Endosulfan (beta) Endosulfan sulfate 1 15-min injection (splitter) found these

Peach

Phosmet (FPD, SIM)

320 37

230 23

Tomato

Chlorothalonil (ELCD, SIM) Endosulfan (alpha) (ELCD, SIM) Endosulfan (beta) (ELCD, SIM) Endosulfan sulfate (ELCD, SIM) 2 50-min injections found these

205 10 16 2 34 4 14 2

153 47 26 4 47 5 21 6

FDA quant results

15

ing found endosulfan sulfate at 14/21 ppb (pg/L) in tomato. Agilent MSD/DRS also found this compound in full-scan mode with less than half the amount reported by FDA/CFSAN available at the MSD due to the split. Several other target compounds that did not contain any halogens or the organophosphorus moeity at the low ppb concentrations were also identified by DRS, such as carbaryl (C12H11NO2) in peach and azoxystrobin (C22H17N3O5) in ginseng. The two FDA/CFSAN procedures for organohalogen and organophosphorus pesticides never would have been able to detect these additional nitrogen-containing pesticides. This shows that deconvolution of data acquired with TID is capable of identifying compounds below 10 pg on column in full-scan mode.

Department of Food and Agriculture, Fresenius J. Anal. Chem, 1991, 339, 376383 2. J. Cook, M. P. Beckett, B. Reliford, W. Hammack, and M. Engel, Multiresidue Analysis of Pesticides in Fresh Fruits and Vegetables Using Procedures Developed by the Florida Department of Agriculture and Consumer Services, J. AOAC Int, 1999, 82, 14191435 3. G. E. Mercer, Determination of 112 Halogenated Pesticides Using Gas Chromatography/ Mass Spectrometry with Selected Ion Monitoring, J. AOAC Int, 2005, 88, 14521462 4. G. E. Mercer and J. A. Hurlbut, A Multiresidue Pesticide Monitoring Procedure Using Gas Chromatography/Mass Spectrometry and Selected Ion Monitoring for the Determination of Pesticides Containing Nitrogen, Sulfur, and/or Oxygen in Fruits and Vegetables, J. AOAC Int, 2004, 87, 12241236 5. USDA Pesticide Data Program Analytical Methods: http://www.ams.usda.gov/science/pdp/ Methods.htm 6. J. Fillion, R. Hindle, M. Lacroix, and J. Selwyn, Multiresidue Determination of Pesticides in Fruit and Vegetables by Gas ChromatographyMass-Selective Detection and Liquid Chromatography with Fluorescence Detection, J AOAC Int, 1995, 78, 12521266 7. S. Nemoto, K. Sasaki, S. Eto, L. Saito, H. Sakai, T. Takahashi, Y. Tonogai, T. Nagayama, S. Hori, Y. Maekawa, and M. Toyoda, Multiresidue Determination of 110 Pesticides in Agricultural Products by GC/MS(SIM), J. Food Hyg. Soc, Japan, 2000, 41, 233241 8. G. F. Pang, C. L. Fan, Y. M. Liu, Y. Z. Cao, J. J. Zhang, X. M. Li, Z. Y. Li, Y. P. Wu, and T. T. Guo, Determination of Residues of 446 Pesticides in Fruits and Vegetables by Three-Cartridge SolidPhase Extraction Gas Chromatography-Mass Spectrometry and Liquid ChromatographyTandem Mass Spectrometry, J. AOAC Int, 2006, 89, 740771 9. M. Anastassiades, S. J. Lehotay, D. Stajnbaher, and F. J. Schenck, Fast and Easy Multiresidue Method Employing Acetonitrile Extraction/Partitioning and Dispersive Solid-Phase Extrac-

Conclusions
The trade-off in trace-level pesticide residue analysis is sensitivity versus confirmation. Therefore, the common practice is to use element-selective GC detectors to screen the extracts and use MS/SIM to confirm hits found by GCs. This can take as many as four injections to have a complete residue analysis from a sample extract. Recent introduction of hardware and software tools, which include the capillary flow three-way splitter, trace ion detection, and deconvolution reporting software, can increase productivity dramatically. With deconvolution the demand for chromatographic resolution is lowered; therefore, the Agilent system can run the analysis at a 3x faster speed to further increase productivity. A single-injection approach even at the 3x fast speed can replace the three-injection approach. A table comparing the results from the current FDA/CFSAN multi-instrument approach and the new Agilent single-injection approach shows that not only is Agilents fast analysis capable of finding all the target analytes, but it can also do it in just one-tenth of the current FDA/CFSAN total analysis time.

References
1. S. M. Lee, M. L. Papathakis, H. C. Feng, G. F. Hunter, and J. E. Carr, Multipesticide Residue Method for Fruits and Vegetables: California

16

tion for the Determination of Pesticide Residues in Produce, 2003, J. AOAC Int, 86:412431 10.S. J. Lehotay, K. Matovsk, and A.R. Lightfield, Use of Buffering and Other Means to Improve Results of Problematic Pesticides in a Fast and Easy Method for Residue Analysis of Fruits and Vegetables, 2005, J. AOAC Int, 88:615629 11.J. W. Wong, M. K. Hennessy, D. G. Hayward, A. J. Krynitsky, I. Cassias, and F. J. Schenck, Analysis of Organophosphorus Pesticides in Dried Ground Ginseng Root by Capillary Gas Chromatography-Mass Spectrometry and -Flame Photometric Detection, J. Agric. Food Chem, 2007, 55, 11171128 12.Chin-Kai Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Agilent Technologies publication, 5989-3299, July 2005 13.Chin-Kai Meng, Improving Productivity and Extending Column Life with Backflush, Agilent Technologies publication, 5989-6018EN, December 2006 14.Matthew Klee, Simplified Backflush Using Agilent 6890 GC Post Run Command, Agilent Technologies publication, 5989-5111EN, June 2006 15.Randy Roushall and Harry Prest, The 5975C Series MSDs: Method Optimization and Trace Ion Detection, Agilent Technologies publication, 5989-6425EN, March 2007 16.http://chemdata.nist.gov/mass-spc/amdis/ explain.html

17. Philip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD Using Deconvolution Reporting Software, Agilent Technologies publication, 5989-1157EN, May 2004 18.Philip L. Wylie, Screening for 926 Pesticides and Endocrine Disruptors by GC/MS with Deconvolution Reporting Software and a New Pesticide Library, Agilent Technologies publication, 5989-5076EN, April 2006 19.Mike Szelewski and Chin-Kai Meng, New Features of Deconvolution Reporting Software Revision A.02, Agilent Technologies publication, 5989-4159EN, November 2005 20.Bruce Quimby and Mike Szelewski, Screening for Hazardous Chemicals in Homeland Security and Environmental Samples Using a GC/MS/ ECD/FPD with a 731 Compound DRS Database, Agilent Technologies publication, 5989-4834EN, February 2006

Acknowledgements
The authors would like to thank Jon Wong, FDA/CFSAN, for the sample extracts and results used in this study as well as valuable feedback on this application.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

17

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA December 18, 2007 5989-7670EN

Reducing Analysis Time Using GC/MSD and Deconvolution Reporting Software

Application
Food and Flavors

Authors
Mike Grady, Steve Morrison, and Bob Deets Campbell Soup Company Campbell Place Camden, NJ 08103 USA Mike Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA

Abstract
Analyzing a complex matrix can be accomplished using multiple specific detectors, but a significant time savings is realized using GC/MSD. Using an available Retention Time Locked database with Deconvolution Reporting Software (DRS) adds a second expert opinion. GC/MSD with DRS provides the fastest methodology to the fewest number of false positives/negatives with the greatest confidence in the results.

dual flame photometric for phosphorus and sulfur (DFPD). Most of the analytes, depending on their molecular formula, could be run on two of these specific detectors for confirmation. However, that would take twice the time. A second choice is to run each sample on unlike columns of differing polarities to the same type detector. Both of these approaches, using ECD, NPD, and/or DFPD, involve more sample handling/tracking, usually twice the number of runs and subsequent data interpretation, and still only have retention time as an identifier. Some methodologies require final confirmation by GC/MSD. Also, it is very difficult to analyze for hundreds of compounds using a nonMS method due to peak overlaps. Campbell Soup Company has been moving from GC-specific detector analyses to GC/MSD analyses for the above-mentioned reasons. A significant time savings can be realized using GC/MSD. It is imperative, however, that data quality be maintained while increasing productivity. Agilent Technologies recently introduced Deconvolution Reporting Software (DRS) for use with a GC/MSD system [1]. DRS automatically combines the results from the MSD ChemStation Quantitation with AMDIS Deconvolution and NIST Search into an easy-to-read report. Using an available Retention Time Locked database reduces methods development for DRS and speeds data comparison among labs. Positives found by normal quantitation are confirmed, and the analyst is directed to further target compound hits to verify. The analyst

Introduction
Analyzing a complex matrix can be accomplished using multiple specific detectors such as electron capture (ECD), nitrogen-phosphorus (NPD), and

need not spend the time to review hundreds of possible compounds, which could take hours per sample. GC/MSD with DRS provides the fastest methodology to the fewest number of false positives/negatives with the greatest confidence in the results. It typically takes less than 2 minutes to process a DRS Report. The purpose of this application note is to show the use of GC/MSD with DRS on complex samples in Campbells Research Lab. It is also intended to show a time savings while maintaining data quality.

The HP-5MS column was used by Agilent to develop the original method and is run in constant pressure mode. Constant pressure methods can be precisely scaled, or sped up, for faster analyses. The retention times of 927 analytes have been recorded on this column. The system is Retention Time Locked to methyl chlorpyrifos at 16.596 min. The primary benefit of RTL for a food laboratory is the ability to maintain retention times after clipping or changing the column. Other benefits include: 1) constant quantitation database and integration events times; 2) switching group times remain constant for laboratories performing SIM analyses; 3) multi-site laboratories can easily compare data; 4) commercially available RTL databases can be used. Additional information is available at www.agilent.com/chem, with application notes detailing the numerous benefits of RTL. The injector parameters are modified from the default Fast Injection to Variable. This allows matching the injector parameters to the PTV-SV requirements. Most importantly, the injection speed is slowed to accommodate the evaporation of the solvent during injection. The 5973N MSD had been upgraded to an inert source and Performance Electronics. The inert source has shown improved response and less degradation for active compounds. Performance electronics minimize noise at faster Scan sampling rates and allow shorter dwell times and more ions/group for SIM acquisitions. A sampling rate of 2 and a scan range of 35-500 yields 3.12 scans/sec. This results in at least 10 data points across the earliest narrow peaks that are 0.055 min wide. This is a good number of points for both quantitation and deconvolution while minimizing noise.

Experimental
Instrument Operating Parameters The instrument operating parameters are listed in Table 1. These conditions may have to be optimized for use in another laboratory. A Programmable Temperature Vaporizing inlet (PTV) is used in the solvent vent mode (SV). The sample is injected at or below the boiling point of the solvent, in this case 50 C. Solvent is evaporated through the split vent line with helium at 200 mL/min for 0.3 min. At 0.5 min, the PTV is rapidly heated to 320 C, transferring the analytes, with minimum solvent, onto the column. The column is held at the initial temperature of 70 C during this process. The PTV-SV allows larger volumes of sample to be injected, 10 L for this study, versus the typical 1 L for this column. The PTV inlet liner, 5183-2037, is multi-baffled and deactivated. It does not contain glass wool, which could contribute to active compound degradation. This liner has sufficient capacity to accommodate the 10 L injection volume

Table1. GC Inlet Mode

Gas Chromatograph and Mass Spectrometer Conditions Agilent Technologies 6890N EPC PTV Solvent vent C/min 600 50 On 50 C 10.00 min (On) On 17.98 psi (On) 0.30 min 200.0 mL/min 0.0 psi 400.0 mL/min 1.00 min 403.9 mL/min On 20.0 mL/min 2.00 min Helium Agilent PTV Liner part# 5183-2037 120 V C/min 25 3 8 41.87 min 0.5 min 325 C Agilent Technologies HP 5 MS, part# 19091S-433 30.0 m 0.25 mm 0.25 m Constant Pressure 17.98 psi 1.9 mL/min Front MSD Vacuum Next C 70 150 200 280 Hold min 2.00 0.00 0.00 10.00 Next C 50 320 200 Hold min 0.50 3.00 0.00

RTL Front injector Sample washes Sample pumps Injection volume Syringe size PreInj Solv A washes PreInj Solv B washes PostInj Solv A washes PostInj Solv B washes Viscosity delay Plunger speed Injection speed Draw speed Dispense speed PreInjection dwell PostInjection dwell MSD Upgrades Solvent delay Low mass High mass Threshold Sampling Scans/sec Quad temperature Source temperature Transfer line temp Tune Type Calibration Standards

System Retention Time Locked to methyl chlorpyrifos at 16.596 min

Temperature ramp Initial Ramp 1 Ramp 2 Cryo Cryo use temperature Cryo timeout Cryo fault Pressure Vent time Vent flow Vent pressure Purge flow Purge time Total flow Gas saver Saver flow Saver time Gas type PTV liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Total run time Equilibration time Oven max temperature Column Length Diameter Film thickness Mode Pressure Nominal initial flow Inlet Outlet Outlet pressure

0 3 10.00 L 25.0 L 0 1 2 2 1s Variable 50.00 L/min 600.00 L/min 1000.00 L/min 0.00 min 0.00 min Agilent Technologies 5973N Inert source and Performance Electronics 4.00 min 35 amu 500 amu 50 2 3.12 150 C 230 C 280 C Autotune

Prepared from certified reference standards available from ChemServe and Crescent Chemical Company. All standards were corrected for purity.

Extraction Procedure An appropriate amount of commodity is weighed, typically 10-15 grams. Surrogates and, if necessary, fortification standards (spike) are added. The commodity is extracted with 1% acetic acid in acetonitrile, centrifuged [2], and passed through an SPE cartridge [3]. Analytes are eluted from the cartridge using acetonitrile/toluene. A 1-gram volume equivalent is taken from the eluant and internal standard(s) is added. The extract is brought to near dryness and solvent exchanged into ethyl acetate for GC/MSD analysis.

trace quantities (< 0.02 ppm) to 0.48 ppm. In most cases, the trace quantities were not found by the GC/MSD standard quantitation using 3 qualifier ion identification. As a verification of the methodology, a second sample of strawberries and tomatoes were each fortified with six analytes at the 0.1 ppm level, together with the Aldrin SS/IS. The analyses results for these spiked samples are shown in Table 3. There are excellent recoveries for most analytes. Responses for two analytes in tomato, atrazine, and permethrin were higher than expected and could be due to matrix enhancement during injection. Time constraints did not allow for further investigation. Duplicate results are also shown in Table 3. The chlorothalonil in tomato at 0.08 ppm compares favorably with the 0.09 ppm found in the first sample. The same is true for captan at 0.47 ppm in the duplicate versus 0.48 ppm in the original sample. The GC/MSD system was calibrated for more than 50 compounds. Using DRS, the analyst can get a verification of the presence of those 50 compounds together with an automated expert second opinion of other compounds that may be present. This second opinion is in two distinct parts. First, the deconvolution of the complex TIC with subsequent matching of clean spectra to a database is provided. Second, the matching of these clean spectra to an independent database, in this case the NIST05a library of > 163,000 compounds.
Celery Gr pepper Strawberry Tomato

Results
Six commodities - apples, lettuce, carrots, celery, green peppers, strawberries, and tomatoes - were purchased at local supermarkets. They were extracted and analyzed using the GC/MSD conditions described earlier. Aldrin was added to each during the extraction process and acts as both a surrogate standard (SS) and an internal standard (IS). The datafiles were processed using Agilents Deconvolution Reporting Software. DRS automatically combines the results from the MSD ChemStation Quantitation with AMDIS Deconvolution and NIST Search into an easy-to-read report. The results are shown in Table 2. Each of the samples showed at least one residue, ranging from
Table 2. Market Basket Commodity Results Commodity Apple Lettuce Compound Table R.T. Methamidopho Acephate Diphenylamine Chlorothalonil Carbaryl Metalaxyl Malathion Isodrin Thiabendazole Captan Phosmet Permethrin Cyfluthrin Cypermethrin Fenvalerate Addional DRSonly compounds
* First R.T. of multiple isomers.

Carrot

5.655 7.690 10.516 14.784 16.806 17.337 18.800 20.031 20.939 21.227 28.504 31.369* 32.218* 32.690* 34.271*

t t 0.02 0.02

0.04 0.23 t t t t 0.09

0.02 0.48 0.03 t t t t 8 0 1 0.03 t

13

10

t = Trace quantity

Table 3.

Spiked and Duplicate Commodity Results, ppm Spikes Strawberry 0.15 0.11 0.11 0.13 0.09 0.14 0.15 Tomato 0.25 0.11 0.15 0.19 0.13 0.26 0.25 Duplicates 0.08

Compound Table R.T. Atrazine 13.159 Lindane 13.461 Carbaryl 16.806 Linuron 18.187 Parathion 19.275 Permethrin I 31.369 Permethrin II 31.550 Chlorthalonil Captan 14.784 21.227

0.47

A portion of the DRS Report for celery is shown in Figure 1. Chlorothalonil was found at 14.823 minutes at 0.02 ppm. AMDIS verified chlorothalonil with a 97 match eluting only 2.7 seconds from its expected RTL time. NIST search further verified chlorothalonil with a 93 match, as the third hit out of the top 100 hits. Aldrin and permethrin are similarly verified with permethrin below the normal reporting limit. Malathion was found by AMDIS and verified by NIST. It was not found by ChemStation because one of three qualifier ions was out of range. AMDIS mitigates this problem because deconvolved spectra are cleaned of interferences and full spectrum matching is used. It would be nearly impossible to identify the analytes of interest in the presence of > 650 individual components in the celery without deconvolution. The TIC for celery is shown in Figure 2.

For all of the commodities tested, DRS verified the presence of all the calibrated peaks found by the ChemStation. Most laboratories only calibrate a fixed number of compounds, say 50-100, as it is not practical to calibrate for all 927. At the same time, these laboratories are interested in identifying other compounds that may be present in samples. Numerous uncalibrated compounds were identified using the 927 compound DRS database. The number of these additional compounds is shown on the last line in Table 2, excluding phthalates, cresols, and sulfur. When these are important to the laboratory, the GC/MSD system can, of course, be calibrated using additional standards. If an estimate of the amount is needed, or if a standard is unavailable, an average response factor can be used. The DRS database provides an average response factor for all 927 compounds. In contrast to the above methodology is the use of multiple element specific detectors such as ECD, NPD, and DFPD. Most of the analytes could be run on two of these specific detectors, but that would take twice the time. Also, it is very difficult to analyze for 927 compounds using a non-MS method due to peak overlaps. A second choice is to run each sample on unlike columns of differing polarities to the same type detector. Both of these approaches, using ECD, NPD, and/or DFPD, involve more sample handling/tracking, usually twice the number of runs, and still only have retention time as an identifier.

MSD Deconvolution Report Sample Name: celery Data File: C:\msdchem\1\Data\sjm03.D Date/Time: 02:30 PM Friday, Jan 19 2007 The NIST library was searched for the components that were found in the AMDIS target library.
R.T. 14.823 18.5992 18.7799 31.3134 Figure 1. CAS # 1897456 309002 121755 52645531 Compound Name Chlorothalonil Aldrin Malathion Permethrin I Agilent ChemStation Amount (ppm) 0.02 1 0.03 Match 97 99 59 71 AMDIS R.T. Diff Sec 2.7 4.3 -1.2 -3.3 NIST Reverse Match Hit Num. 91 93 47 53 3 1 1 5

DRS report for celery.

www.agilent.com/chem

Aldrin

Chlorothalonil Permethrin Malathion

10

12

14

16

18

20

22

24

26

28

30

32

Figure 2.

Market basket celery total ion chromatogram.

Conclusions
Analyzing a complex matrix can be accomplished using multiple specific detectors, but a significant time savings is realized using GC/MSD. Many more analytes can be determined simultaneously using mass spectra for confirmation. Using an available Retention Time Locked database reduces methods development and speeds data comparison among labs. Deconvolution Reporting Software adds a second expert opinion. Positives found by normal quantitation are confirmed, and the analyst is directed to further target compound hits to verify. GC/MSD with DRS provides the fastest methodology to the fewest number of false positives/negatives with the greatest confidence in the results.

2. Fast and Easy Multiresidue Method Employing Acetonitrile Extraction/Partitioning and Dispersive Solid-Phase Extraction for the Determination of Pesticide Residues in Produce, M. Anastassiades, S. J. Lehotay, D. Stajnbaher, F. J. Schenck. (2003) J.AOAC Int. 86, 412-431. 3. Analytical Methods for Residual Compositional Substances of Agricultural Chemicals, Feed Additives, and Veterinary Drugs in Food, Japan Department of Food Safety, Ministry of Health, Labour and Welfare.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA May 14, 2007 5989-6677EN

References
1. Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Philip L. Wylie, Michael J. Szelewski, and Chin-Kai Meng, Agilent Technologies Pub # 5989-1157EN.

Rapid Analysis of Herbicides by Rapid Resolution LC with Online Trace Enrichment Application

Environmental

Author
Michael Woodman Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
Environmental and Food Safety agencies are constantly updating methods to improve detection limits and to resolve interfering compounds. One particular method, EPA 555, is used for the analysis of chlorinated phenoxy acid herbicides in drinking water. A mandated trace enrichment step significantly impacts the ease of use and reliability of the method. The method uses 5-m analysis columns and online trace enrichment. The variation here uses small ZORBAX 3.5- and 1.8-m RRHT columns and an autoSPE (Solid Phase Extraction) cartridge with an automated switching valve mounted in the column compartment. Combined with sample introduction via direct injection to the autoSPE cartridge, instead of the loading pump specified in the EPA method, we dramatically reduce the overall analysis time and virtually eliminate the potential of sample cross-contamination.

disposable SPE cartridges, which require an elution step into a vial prior to analysis, online SPE assures 100% sample transfer to the analysis column and dramatically increases sensitivity by increasing the analyte mass delivered to the column. In EPA Method 555, 20 mL of drinking water is loaded through a pump to an SPE cartridge mounted on a high-pressure switching valve on the HPLC system. Because few, if any, autosamplers can inject this large volume, the sample must be pumped onto the cartridge. Contamination of the loading pump with prior samples is always a concern, and adequate flushing and blank runs become an important part of the overall method procedure. To reduce the sampling volume sufficient for available automatic preparative samplers, without losing sensitivity in the method, it is necessary to reduce the analysis column size while preserving resolving power. Ancillary benefits of using smaller columns generally include reduced analysis time and solvent consumption, and greater compatibility with ionization sources in mass spectrometers. If the ratios of their column length to particle size are equal, columns are considered to have equal resolving power. Significant reductions in column volume can be made by reducing the length and/or internal diameter of the column. In the latter case, the flow rate would normally be reduced as in Equation 1.

Introduction
Trace analyte detection in relatively clean matrices is an excellent application for online SPE procedures. Compared to manually loading samples with

Flowcol. 1

Diam.column2 Diam.column1

= Flowcol. 2

(eq. 1)

little improvement is observed. In the subsequent examples, we will see the results associated with the calculations discussed above. Experimental Conditions See figure 1 for configuration.
System Agilent 1200 Series Rapid Resolution LC consisting of: G1379B micro degasser G1312B binary pump SL G1312A binary pump with solvent selection valve option, or G1354A quaternary pump G1367C HiP ALS autosampler SL, and G2258A Dual Loop Prep autosampler 5 ml G1316B Thermostatted column compartment SL with 6- or 10-port 2-position switching valve G1315C UV/VIS diode array detector (DAD) SL, flow cell as indicated in individual chromatograms ChemStation 32-bit version B.02.01 Columns Agilent ZORBAX SB-C18, 4.6 250 mm, 5 m Agilent ZORBAX SB-C18, 3.0 150 mm, 3.5 m Agilent ZORBAX SB-C18, 2.1 80 mm, 1.8 m Agilent ZORBAX SB-Aq, 4.6 12.5 mm, 5 m Mobile phase conditions Organic solvent: Aqueous solvent: Gradient conditions Gradient slope: 7.8 or 2.3% per column volume, as indicated. See individual chromatograms for flow rate and time Acetonitrile 25 mm phosphoric acid in Milli-Q water

The combined effect of reduced length and diameter contributes to a reduction in solvent consumption. We normally scale the injection mass to the size of the column and a proportional injection volume would be calculated from the ratio of the void volumes of the two columns, multiplied by the injection volume on the original column, as in Equation 2 below.
Volumecolumn2 Volumecolumn1

Inj. vol.col. 1

= Inj. vol.col. 2

(eq. 2)

Short columns packed with small particle sizes are typically operated at high linear velocities. The increase in elution speed will decrease absolute peak width and may require the user to adjust data acquisition rates and reduce signal filtering parameters. This will ensure that the chromatographic separation is accurately recorded in the acquisition data file. For gradient elution separations, where the mobile phase composition increases through the initial part of the analysis until the analytes of interest have been eluted from the column, successful method conversion to smaller columns requires that the gradient slope be preserved. We can express the gradient slope as in Equation 3.

% Gradient slope =

(End% Start%) #Column volumes

Sample
(eq. 3)

Note that the use of % change per column volume rather than % change per minute frees the user to control gradient slope by altering gradient time and/or gradient flow rate. A large value for gradient slope yields very fast gradients with minimal resolution, while lower gradient slopes produce higher resolution at the expense of increased solvent consumption and somewhat reduced sensitivity. Longer analysis time may also result unless the gradient slope is reduced by increasing the flow rate, within acceptable operating pressure ranges, rather than by increasing the gradient time. Resolution increases with shallow gradients because the effective capacity factor, k*, is increased. Much like in isocratic separations, where the capacity term is called k', a higher value directly increases resolution. The effect is quite dramatic up to a k value of about 510, after which
2

EPA 555 Group A chlorinated phenoxy acid herbicides (picloram, chloramben, dicamba, bentazon, 2,4-D, dichlorprop, 2,4,5-TP, acifluorfen), 100 g/mL in methanol or diluted to 20 ng/L (20 ppb) in reagent water acidified with 25 mm phosphoric acid.

Results
The separation was initially performed via direct injection of concentrated standard on a 4.6 250 mm, 5-m ZORBAX SB-C18 column, thermostatted to 25 C, using conditions referenced in US EPA method 555 (Figure 2). The described trace enrichment procedure using pump A as the loading pump was performed (Figure 3). The method was then scaled in flow and time for exact translation to a 3.0 150 mm 3.5-m column (Figure 4) using 5-mL trace enrichment injection. Finally, a 2.1 80 mm 1.8-m configuration (50-mm plus 30-mm columns in series) is used to demonstrate the feasibility of this separation under conditions for trace enrichment requiring less than 1.5-mL injection. (Figure 5)

Load/Wash position

Elute/Analyze position

2 Position/6 Port valve

Figure 1.

Trace enrichment autoSPE scheme.

Figure 1 shows the schematic placement of modules and columns in the system. The A pump is the loading pump in case of volumes exceeding the 5-mL capacity of the G2258A Dual Loop Autosampler, thus pump A uses one line for sample and a second line for the aqueous eluent, 25 mm phosphoric acid. If direct injection from the autosampler is used, pump A is delivering 25 mm phosphoric acid. If the A pump is fitted with a degasser, the sampling line should bypass the degasser module to minimize contamination with sample solutions. To conduct sampling through the A pump, the valve should be in position B while the new sample is flushed through the A pump. Then switch the valve to the A position and load the required 20 mL sample volume. The analysis

begins when the valve is returned to the B position, at which time the sampling line on the A pump would be flushed with reagent water or the next sample, as appropriate. Figures 2 and 3 show the standard separation by direct injection and pumped trace enrichment, respectively. With column regeneration steps, this results in a total analysis time of 60 minutes. Translation of the gradient to the 3.0- 150-mm column requires a reduction in flow rate, due to the smaller diameter, and a reduction in gradient time because of the shorter column length. The resulting analysis is reduced from 60 to 36 minutes and solvent consumption is proportionately reduced from 60 mL to 15.5 mL.

100

80

12.679 - Picloram

14.436 - Chloramben

mAU 120

Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 25 C Gradient: 25 mM H3PO4, ACN, 10% to 90% ACN in 30 min Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1 mL/min Group A compounds: 1 L of 100 g/mL Total analysis time: 60 min Detection: UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default)
18.895 - Bentazon 17.845 - Dicamba 24.678 - Acifluorfen

60

21.104 - Dichlorprop

19.480 - 2,4-D

40

13.303

0 10 12 14 16 18 20 22 24 26 28 min

Figure 2.

Gradient separation of herbicides on 4.6 mm 250 mm, 5 m ZORBAX SB-C18.

17.653

20.238

29.609

20

23.095 - 2,4,5-TP

14.380 - Chloramben

12.557 - Picloram

mAU 350 300 250 200 150 100

18.871 - Bentazon

19.414 - 2,4-D

21.063 - Dichlorprop

17.779 - Dicamba

23.050 - 2,4,5-TP

24.667 - Acifluorfen

13.194

17.607

18.439

0 10 12 14 16 18 20 22 24 26 28 min

Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 25 C Gradient: 25-mM H3PO4, ACN, 10% to 90% ACN in 30 min Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1 mL/min Group A compounds: 20 mL of 20 g/L trace enrichment Total analysis time: 60 min Detection: UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default) Figure 3. Trace enrichment (20 mL) of 20-ppb solution on 4.6 250 mm 5-m ZORBAX SB-C18.

4.094

120 100 80 60 40 20 0 2

5.118

mAU

Conditions: EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 3 mm 150 mm, 3.5 m Column temp: 25 C Gradient: 25 mm H3PO4, ACN, 10% to 90% ACN in 18 min Gradient slope: 7.8% ACN/column volume Analysis flow rate: 0.43 mL/min Group A compounds: 5 mL of 20 g/L (20 ppb) Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 36 min
7.870

7.073

8.221

9.211

21.720

10.352

27.714

50

4.462

6.642

10

11.555

12

14

14.041 min

Figure 4.

Trace enrichment (5 mL)of 20-ppb solution on 3.0 150 mm, 3.5-m ZORBAX SB-C18.

The last peak in Figure 4 is missing due to a valve timing error that was not detected until sometime after the lab work was completed. Peak 8 was not eluted from the trace enrichment column before
4

the valve switched offline for regeneration and equilibration. Note the baseline shift that occurs after peak 7, not seen in other autoSPE examples.

29.595 29.786

1.532

700 600 500 400 300 200

1.672

mAU

Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18, 2.1 mm 80 mm (50mm + 30mm in series), 1.8 m Column temp: 50 C Gradient: 25-mM H3PO4/ACN, 10% to 90% ACN in 2.7 min 7.8% ACN/column volume Analysis flow rate: 0.72 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Sample: Aged 1 L 100 g/mL Total analysis time: 6 min
2.048 2.723 2.626 2.830

2.161 2.206

2.385

1.643

1.121

0 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25 min

Figure 5.

High-speed gradient separation of herbicides on 2.1 80 mm, 1.8-m ZORBAX SB-C18.

In Figure 5 we see the combination of highest speed and resolution, using the full capability of the 1200 Rapid Resolution LC. Operating pressure was, at the maximum point, about 520 bar. We maintain comparable resolution to the original 4.6 250 mm, 5-m method, a 60-minute run time, with an analysis time of only 6 minutes.

Conclusion
As is the case for many existing methods, it is both possible and practical to modernize this method to improve throughput and overall performance. Here we see the potential for a 10-fold increase in analysis speed and elimination of the loading pump scheme found in the original method. Approximately 1.3 mL of sample solution, injected to the autoSPE setup using the 2.1 80 mm configuration, is all that is needed to replace the 20-mL injection previously loaded through the pump. This approach can greatly improve productivity and ensure minimal sample cross-contamination.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

2.310

3.025

100

2.577

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA March 30, 2007 5989-5176EN

Improving Productivity and Extending Column Life with Backflush

Application Brief
Chin-Kai Meng

All Industries

A previous application note [1] has shown that multiple GC signals and MS signals can be acquired from a single sample injection. When a 3-way splitter is connected to the end of a column, column effluent can be directed proportionally to two GC detectors as well as the MSD. This multi-signal configuration provides full-scan data for library searching, SIM data for quantitation, and element selective detector data for excellent selectivity and sensitivity from complex matrices. The system used in this study consists of a 7683ALS, a 7890A GC with split/splitless inlet, 3-way splitter, ECD, dual flame photometric detector (DFPD), and a 5975C MSD. Figure 1 shows four chromatograms from a single injection of a milk extract. The synchronous SIM/scan feature of the 5975C MSD provides data useful for both screening (full scan data) and quantitation (SIM data). DFPD provides both P and S signals without the need to switch light filters. Noticeably in the full scan TIC in Figure 1, a significant number of matrix peaks were observed after 32 minutes. It is not uncommon to add a bake-out oven ramp to clean the column after analyzing complex samples. The bake-out period is used to quickly push the late eluters out of the column to be ready for the next injection. Therefore, it is common to use a higher oven temperature than required for the analysis and an extended bake-out period at the end of a normal
Full scan TIC

Highlights
Backflush a simple technique to remove high boilers from the column faster and at a lower column temperature to cut down analysis time and increase column lifetime. The milk extract example shows that a 7-minute 280 C backflush cleaned the column as well as a 33-minute 320 C bake-out. The cycle time was reduced by more than 30%. Using backflush, excess column bleed and heavy residues will not be introduced into the MSD, thus reducing ion source contamination.

SIM

ECD

DFPD(P)

10

15

20

25

30

35

40

Figure 1.

Four chromatograms collected simultaneously from a single injection of a milk extract.

www.agilent.com/chem
over program to clean out the column, which adds to the cycle time and shortens the column lifetime. Adding the bake-out period to the milk extract analysis, additional matrix peaks were observed even up to 72 minutes, while target compounds already eluted before 42 minutes. This means that 30 minutes were lost in productivity for each injection. Backflush [2] is a simple technique to drastically decrease the cycle time by reversing the column flow to push the late eluters out of the inlet end of the column. Late eluters stay near the front of the column until the oven temperature is high enough to move them through the column. When the column flow is reversed before the late eluters start to move down the column, these late eluters will take less time and at a lower oven temperature to exit the inlet end of the column. There are many benefits in using backflush: Cycle time is reduced (no bake-out period, cooling down from a lower oven temperature) Column bleed is reduced (no high-temperature bake-out needed), resulting longer column life Ghost peaks are eliminated (no high boilers carryover into subsequent runs) Contamination that goes into the detector is minimized, which is especially valuable for the MSD (less ion source cleaning) of 320 C. A column effluent splitter or QuickSwap is required to do the backflush.

References
1. Chin-Kai Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Agilent Application Note, 5989-3299EN, July 2005. 2. Matthew Klee, Simplified Backflush Using Agilent 6890 GC Post Run Command, Agilent Application Note, 5989-5111EN, June 2006.

Acknowledgement
Milk extract is courtesy of Dr. Steven Lehotay from USDA Agricultural Research Service in Wyndmoor, Pennsylvania, USA.

Figure 2 shows three total ion chromatograms from the Agilent 7890A GC/ 5975C MSD. The top chromatogram is a milk extract analysis with all the target compounds eluted before 42 minutes (over program goes to 280 C). However, an additional 33-minute bake-out period at 320 C was needed to move the high boilers out of the column. This bake-out period was almost as long as the required time to elute all target compounds. The middle chromatogram is the same milk extract analysis stopped at 42 minutes with a 7-minute backflush post-run at 280 C added to the analysis. The bottom chromatogram is a blank run after the backflushing was completed. The blank run shows that the column was very clean after backflushing. The example shows that a 7-minute backflush cleaned the column as well as a 33-minute bake-out. The milk extract example in Figure 2 illustrates the backflush technique in reducing cycle time and column bleed. The cycle time was reduced by more than 30% and the column was kept at 280 C, without going to the bake-out temperature

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

It took an additional 33 min and heating the column to 320 C to remove these high boilers

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA December 26, 2006 5989-6018EN

Run stopped at 42 min and backflushed at 280 C for 7 mins Blank run after backflushing showing the column was clean
5 10 15 20 25 30 35 40 45 50 55 60 65 70 min

Figure 2.

Three total ion chromatograms comparing the results with and without backflush.

Using RTL and 3-Way Splitter to Identify Unknown in Strawberry Extract Application Brief
Chin-Kai Meng

Food Safety and Environmental

Fruit and vegetable extracts are usually very complex to analyze. It is common to use the very selective GC detectors, for example NPD, ECD, and FPD, to look for trace pesticide residues in the extracts. Mass spectrometry is most often used to confirm the hits from GC detectors. A previous application note [1] describes a GC/MS system with a three-way splitter added to the end of the column. The column effluent could be split three ways to two GC detectors and the MSD. The splitter system is therefore capable of providing up to four signals (two GC signals, SIM, and full-scan chromatograms) from a single injection. The combination of element selective detectors, SIM/scan, and Deconvolution Reporting Software (DRS) makes a very powerful pesticide analysis system [2]. The trade-off is the decrease of analyte concentration in any detector due to the flow splitting at the end of the column. The system used for this study consists of an Agilent 7890A GC with split/splitless inlet, a three-way splitter, ECD, dual flame photometric detector (DFPD), and 5975C MSD. Figure 1 shows chromatograms from 2 separate injections (each injection provides two GC signals) of the same strawberry extract without any hardware or filter changes. All of the target compounds were found and confirmed by DRS, GC, and MS signals except the unknown peak at about 41 minutes. The peak shows responses from ECD, DFPD(S) and DFPD(P). However, no peak was observed in the MS full-scan signal. This makes it difficult to confirm the unknown peak using the full-scan TIC. Since the analysis was retention time locked, it is therefore possible to find potential matches by examining the RTL pesticide database (part number G1672AA). The unknown compound, containing electron-capturing atoms (for example, Cl or O), P, and S atoms, would have a target retention time inside the

Highlights
Splitter+an inert, easy-to-use capillary flow technology that splits column effluent to multiple detectors (for example, MSD, DFPD, and ECD). The splitter configuration provides a comprehensive screening and quantitative system. By combing RTL, element-selective detector chromatograms, and the RTL pesticide database, a trace level pesticide residue was identified without the full-scan mass spectrum.

TIC

ECD

DFPD (S)

DFPD (P)

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

Figure 1.

Unknown compound detected by GC signals not found in strawberry extract TIC.

Table 1.

Compound List Extracted from the RTLPest3.tab File CAS 117337196 191242 3383968 60044260 83794 Mol form C15H15CIFN3O3S2 C22H12 C16H20O6P2S3 C12H4Br6 C23H22O6 Mol wt 403.9 276.3 466.5 627.6 394.4 R.T. 39.10 39.13 40.74 40.93 41.70 Target Ion 403 276 466 308 192 Q1 56 277 125 468 191 Q2 405 138 93 148 394 Q3 232 275 109 154 177

Name Fluthiacet-methyl Benzo[g,h,i]perylene Temephos PBB 169 hexabrombiphenyl Rotenone

41 \ 0.5-minute window (that is, 40.5 to 41.5 min) in the database, if it is included in the database. Table 1 is a portion of the RTLPest3.tab file1 opened in Microsoft Excel. The compound temephos at locked retention time 40.74 min meets all the criteria for the unknown peak. To further confirm peak identity, extracted ion chromatograms (EICs) of the four major ions of temephos were plotted. Figure 2 shows EICs of target ion and three qualifiers (ions 466, 125, 93, and 109 from Table 1) of temephos. Although the ion intensities were weak, the noticeable presence of all four ions at 40.9 min helped to confirm that the unknown peak was temephos.

1. The RTLPest3.tab file is created in the C:\Database directory while executing the Tools\List Screen Database command (in MSD Enhanced Data Analysis software) and selecting the RTLPest3.scd from the C:\Database directory.

200 100 0

Ion 466

200 100 0

Ion 125

200 100 0

Ion 93

200 100 0

Ion 109

36

37

38

39

40

41

42

43

44

Figure 2.

EICs of target ion 466 (temephos) and three qualifier ions.

References
1. Chin-Kai Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Application Note, 5989-3299, July 2005. 2. Mike Szelewski and Bruce Quimby, New Tools for Rapid Pesticide Analysis in High Matrix Samples, Application Note, 5989-1716, October 2004.

Acknowledgement
Strawberry extract is courtesy of Dr. Steven Lehotay from USDA Agricultural Research Service in Wyndmoor, Pennsylvania, USA.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Copyright 2006 Agilent Technologies All Rights Reserved. Reproduction, adaptation, or translation without prior written permission is prohibited, except as allowed under the copyright laws. Printed in the USA December 13, 2006 5989-6007EN

Automated Screening of 600 Pesticides in Food by LC/TOF MS Using a Molecular-Feature Database Search Application

Food Safety

Authors
E. Michael Thurman and Imma Ferrer Pesticide Residue Research Group University of Almera 04120 Almera, Spain Jerry A. Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

the compounds to <0.5 mg/kg for 95% of the compounds. Strengths of the MF algorithm include rapid screening of hundreds of compounds at sensitive levels in minutes compared to a manual approach of hours to days.

Introduction
Contamination of foodstuffs with pesticides always means a risk to the consumer. Fetuses, infants, and children, a particularly sensitive group, are currently protected by a limit of 0.01 mg/kg in baby food by European Union legislation [1]. Maximum residue limits (MRLs) should be set at the lowest possible level with the aim of setting products on the markets without any measurable residues. However, this is often difficult because of the use of fungicides for transport and storage of fruit and vegetables, as well as the use of insecticides for crop protection. The analysis of pesticides in baby food has involved both GC/MS and LC/MS methods. Classically, GC/MS has been used to look for volatile pesticides, especially the chlorinated insecticides and herbicides, and LC/MS has been used to monitor the more polar insecticides, herbicides, and fungicides. GC/MS methods rely on both full scan and selected ion monitoring. Recently the use of reverse search methods for GC/MS has made it possible to search large NIST pesticide libraries in minutes [2] and has made screening quite simple for pesticides amenable to GC/MS methods.

Abstract
Searching a database using a molecular feature (MF) algorithm was developed for the screening of 600 pesticides and degradates in extracts of food by liquid chromatography time-of-flight mass spectrometry in positive ion mode with full-scan accurate mass spectra. The database search works by compiling the accurate mass of the ions detected and identified as real compound chromatographic peaks without ion extraction and compares them with the monoisotopic exact masses of the compounds in the database. The screening criteria consisted of 5 ppm accurate mass window, 0.2 minute retention time window, and a minimum area count of 1,000 counts (signal-to-noise ratio of ~10:1). The limit of detection and retention time was determined for 100 of the 600 compounds and varied from <0.01 mg/kg for 34% of

Unfortunately, similar reverse-search methods have not been available for LC/MS for two reasons. First, the single quadrupole and triple quadrupole methods do not operate in full scan mode for pesticide screening because of a lack of sensitivity. Second, standardization and reproducibility of CID fragmentation energy for broad usage has been difficult, making common spectral libraries unavailable. Recently, the use of LC/TOF MS has shown that it has the capability and sensitivity to obtain full scan spectra with accurate masses of less than 1 ppm [3] suitable for database searching [4].

to 100% B in 30 minutes at a flow rate of 0.6 mL/min Agilent 6210 LC/MS TOF with dual spray electrospray source Positive ESI, Capillary 4000 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Fragmentor 190 V, skimmer 60 V, Oct DC1 37.5 V, OCT RF V 250 V Reference Masses: mass range (m/z) 121.0509 and 922.0098, resolution: 9500 500 at m/z 922.0098, m/z 50 to 1,000, Reference A Sprayer 2 is constant flow rate during the run

Experimental
Vegetable and Fruit Extraction (QuEChERS) QuEChERS is the acronym for the extractionmethod, which stands for quick, easy, cheap, effective, rugged, and safe. It is a method that is widely receiving acceptance for rapid extraction of pesticides in food [5, 6]. Weigh 15 g of a previously homogenized sample into a 40-mL Teflon centrifuge tube. Add 15 mL of acetonitrile (containing 1% acetic acid), add 6 g of anhydrous MgSO4 and 2.5 g of NaAc63H2O (sodium acetate trihydrate) and shake the sample vigorously for 1 min by using a vortex mixer at maximum speed or hand shaking. Centrifuge for 3 minutes at 3,700 rpm. Take 5 mL of supernatant into a 15-mL tube, add 250 mg of PSA adsorbent and 750 mg of MgSO4, and vortex and shake for 20 s. Then centrifuge again for 3 min at 3700 rpm. Transfer 1.0 mL into LC/MS vial. The 1.0 mL supernatant is then evaporated to dryness and brought back up in 8/92% methanol/water for LC/MSD TOF and ion trap analysis. Direct analysis of the fruit and vegetable extracts were analyzed by injecting 50 L. Nonfortified samples were analyzed directly at this same point by LC/MS TOF. LC/MS TOF Methods LC pumps: Agilent 1100 binary pumps, injection volume 50 L with standard Agilent 1100 ALS Column: ZORBAX Eclipse XDB 4.6 150 mm C-8, 5-micron (p/n 993967-906) Mobile Phase A = 0.1% formic acid in water, and B = acetonitrile, gradient began with 5 minutes isocratic at 10% B followed by a linear gradient

Results and Discussion


Molecular Feature Database Search Theoretical monoisotopic exact masses of compounds based on their molecular formula were calculated using an Excel spreadsheet and put into csv (comma separated values) format for use by the TOF software of the Agilent LC/TOF MS system for 600 pesticides known to ionize by positive ion electrospray. The csv file is created in the TOF software by an Excel spreadsheet tool called Formula DB Generator. The csv file is then searched automatically by the LC/TOF MS instrument at the completion of the sample run and a report generated on compounds that were found in the database. Search criteria include ppm mass tolerance (5 ppm), retention time window (0.2 minutes) if available, and minimum peak height count, which is called the compound threshold (1,000 counts or a signal-to-noise ratio of ~10:1 or 0.06% relative volume), and adducts and neutral fragment losses. The search routine is called a molecular feature extractor and is software recently available on the Agilent LC/TOF MS (November 2005). The molecular feature extractor finds all ions in an LC/MS TOF data file representing ions of real compounds in the sample analyzed. Noise and other extraneous ions are excluded. The resulting list of ions are then searched against the csv database using the chosen criteria and ions found are then tabulated from the full scan spectrum and checked for accuracy and retention time against the database. The molecular feature approach is more suited to large libraries because of the ease of operation and the quickness for which the search is done. Thus,

each ion of interest in the database is not extracted from the sample file as in a reverse search. This procedure requires much more time with LC/TOF MS data files collected in profile mode. From this point, confirmation is carried out manually by checking the positive screens for retention time match and fragment ions (if present). The sample may be reanalyzed at a higher fragmentor voltage to check for fragment ions and confirmed by authentic standard analysis. Limit of Detection The limit of detection (LOD) was determined in several matrices, including spiked food samples and solvent extracts for 100 compounds (Table 1). These compounds consist of the major classes of pesticides that are commonly used in the United States and Europe for treating crops of fruits and vegetables. The limit of detection was based on an accurate mass of less than 3 ppm and the appearance of the correct accurate mass of A+1 and A+2 isotopic signatures. The LOD was determined at various levels for pesticide work, including the EU regulation of 0.01 mg/kg for baby food, and 0.05, 0.1, and 0.5 mg/kg for various food levels, depending on pesticide and crop type. The LOD was equal to or less than 0.01 ppm for 33 compounds and less than or equal to 0.05 for 60 compounds (60%). The 0.05 mg/kg LOD is also a critical one for food monitoring of banned substances or controlled compounds. The LOD for 95% of the compounds was equal to or less than 0.2 mg/kg for food. Only six compounds were found

to be insensitive with LODs of 0.5 mg/kg for food. The insensitive compounds were promecarb and aldicarb, which are two carbamate pesticides that fragment easily in the electrospray source and give low abundance for the MH+ ion. Likewise, malathion oxon and dimethoate are two organophosphate pesticides that fragment easily. Thus, these compounds could be more sensitively detected by the use of the more abundant fragment ion rather than the MH+. For example, Figure 1 shows the mass spectrum of dimethoate. The MH+ ion is not the major fragment ion of the spectrum, and in fact has an intensity about three to four times less than the m/z 124.9819 ion. Furthermore, it must be taken into consideration that the LOD is affected by the matrix for two reasons. One is suppression of ionization and the second is interfering ions of nearly identical mass. Suppression of ionization has been tested in our previous work for some of these pesticides in food [7], including pepper, broccoli and tomato, melon, orange, and lemon, so we have experience on which matrices are most difficult. For example, Figure 2 shows the matrix chromatogram for pesticide-free pepper, which gives a complicated chromatogram. The MF database search identified ~3,000 compounds of which none were pesticides. These peaks contained signal-to-noise ratio of 10:1 or greater and present an extremely difficult matrix for which to find ions, especially at trace levels. Spiked food extracts of these difficult matrices were used to establish LOD for the compounds shown in Table 1. An example printout is shown in Figure 3. The report contains formula, compound,

SH P H3CO OCH3 O C5H13NO3PS2+ Exact mass: 230.0069 S H N CH3

Figure 1.

Dimethoate mass spectrum showing low intensity of MH+ ion and importance of using characteristic fragment ions to lower LODs on some compounds of low intensity, specifically, m/z 124.9819 ion.

Figure 2.

Blank pepper sample showing complexity of the sample ~3000 accurate mass peaks were detected in this sample at signal to noise of 10:1 or greater.

accurate mass of the neutral molecule, error in mDa and error in ppm, retention time error in minutes, and a description (specifically, fungicide). The mass spectrum of the MH+ and isotope signature of the compound are also shown, which is useful for a quick check and partial confirmation of the formula, especially since most pesticides show an interesting A+2 ion from a halogen or sulfur atom. A+2 Ions and Empirical-Formula Confirmation Because the MF database search is a screening program, a formula is not unequivocally confirmed at this point with a mass accuracy window of 5-ppm. The molecular formula (not molecular identification) may be confirmed by the A+2 isotopic signature of the compound. For example, 70% of the 600-compound database contains either S, Cl, or

Br, which give the isotopic cluster of the A+2 ions. The accurate mass of the isotope along with the intensity profile can then be used as a first level confirmation of the empirical formula. Thus, this adds a great deal of confidence to the screening data of the accurate mass but of course does not yet meet the standards for identification of the molecule, a topic discussed later. For example, let us examine the isotopic signature of imazalil in a pear extract (Figure 4). The measured mass of the MH+ was 297.0564 and the chlorine 37 isotope was 299.0533. Thus the difference in mass is 1.997 mass units, which is the mass defect of a chlorine 37 atom relative to the chlorine 35 atom that has been replaced. Furthermore, the intensity of the A+2 peak is about 2/3 of the A peak, which is consistent with two chlorine atoms in the molecule. This is further established by the presence of an A+4 peak at 301.0503. Thus, these

Limits of Detection for Pesticides in Food Samples with Retention Time and Accurate Mass Accurate mass Compound Ret time Formula molecule Atrazine 21.1 C8H14N5Cl 215.0938 Azoxystrobin 24.0 C22H17N3O5 403.1168 Benalaxyl 26.8 C20H23NO3 325.1678 Buprofezin 27.2 C16H23N3OS 305.1562 Cyanazine 22.0 C9H13N6Cl 240.0890 Diazinon 27.6 C12H21N2O3PS 304.1010 Difenconazole 26.4 C19H17Cl2N3O3 405.0647 Isomer 26.6 C19H17Cl2N3O3 405.0647 Difenoxuron 21.3 C16H18N2O3 286.1317 Dimethomorph 22.2 C21H22NO4Cl 387.1237 Fenamiphos 23.9 C13H22NO3PS 303.1058 Imazalil 18.0 C14H14N2OCl2 296.0483 Imazapyr 20.0 C13H15N3O3 261.1113 Imazaquin 20.0 C17H17N3O3 311.1270 Irgarol 21.2 C11H19N5S 253.1361 Irgarol metabolite 17.0 C8H15N5S 213.1048 Isoproturon 21.3 C12H18N2O 206.1419 Mebendazole 18.2 C16H13N3O3 295.0957 Metolachlor 25.6 C15H22NO2Cl 283.1339 Metribuzin 15.0 C8H14N4OS 214.0888 Nicosulfuron 17.0 C15H18N6O6S 410.1009 Prochloraz 23.0 C15H16Cl3N3O2 375.0308 Prometon 16.6 C10H19N5O 225.1590 Prometryn 19.0 C10H19N5S 241.1361 Propazine 23.0 C9H16N5Cl 229.1094 Propiconazole 25.9 C15H17CI2N3O2 341.0698 Isomer 26.1 C15H17CI2N3O2 341.0698 Simazine 18.8 C7H12N5Cl 201.0781 Spinosyn A 20.9 C41H65NO10 731.4608 Spinosyn D 21.9 C42H67NO10 745.4765 Spiroxamine 19.6 C18H35NO2 297.2668 Isomer 19.7 C18H35NO2 297.2668 Terbuthylazine 23.4 C9H16N5Cl 229.1094 Terbutyrn 20.4 C10H19N5S 241.1361 Triflumazole 25.9 C15H15ClF3N3O 345.0856 Acetamiprid Acetochlor Alachlor Bensultap Bromuconazole Carbaryl Carbendazim Carbofuran Cartap Chlorfenvinphos Cyproconazole Cyromazine Deethylatrazine Deisopropylatrazine Dichlorvos Dimethenamide Dimethoate Diuron 16.3 23.0 23.0 21.4 23.8 21.3 6.2 20.4 3.1 26.5 23.4 2.9 15.3 12.1 20.0 24.0 16.3 21.0 C10H11N4Cl C14H20NO2Cl C14H2ONO2Cl C17H21NO4S4 C13H12N3OCl2Br C12H11NO2 C9H9N3O2 C12H15NO3 C7H15N3O2S2 C12H14Cl3O4P C15H18N3OCl C6H10N6 C6H10N5Cl C5H8N5Cl C4H7Cl2O4P C12H18NO2SCl C5H12NO3PS2 C9H10N2OCl2 222.0672 269.1183 269.1183 431.0353 374.9541 201.0790 191.0695 221.1052 237.0606 357.9695 291.1138 166.0967 187.0625 173.0468 219.9459 275.0747 228.9996 232.0170

Table 1.

LOD food mg/kg 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Table 1.

Limits of Detection for Pesticides in Food with Retention Time and Accurate Mass (Continued) Ret time 21.8 15.7 15.7 19.2 22.5 16.9 21.2 23.5 12.1 18.7 11.9 19.1 28.6 24.0 25.6 29.0 3.7 17.7 4.5 18.5 6.0 20.6 20.4 25.0 11.5 28.6 14.0 24.1 17.4 24.8 27.3 17.0 27.5 11.4 18.5 23.8 15.0 25.0 29.2 18.8 27.2 14.6 25.4 25.0 27.6 24.4 28.2 23.0 3.2 Formula C11H15NO2S C9H12N2O C9H10N5O2Cl C13H18N2O2 C10H19O6PS2 C10H19O6PS2 C15H21NO4 C11H15NO2S C5H10N2O2S C9H11ClN2O C11H15CIN4O2 C14H18N2O4 C11H15BrCIO3PS C12H17NO2 C11H14NOCl C14H21NOS C10H7N3S C10H9N4SCl C5H11NS3 C7H14N2O2S C7H14N2O3S C11H13NO4 C10H13N2OCl C14H13N3O2SF4 C8H15N5O C17H8N2O3Cl2F8 C10H10N4O C6H11N2O4PS3 C11H15NO4S C9H17NOS C10H14NO5PS C9H9NOCl2 C13H9CI3N2O C7H14N2O4S C9H13N2O2Br C13H12N3OCI2Br C11H23NOS C14H9N2O2ClF2 C21H11N2O3ClF6 C7H5N2O3Cl2F C16H8N2O3Cl2F6 C11H10N2OCl2 C13H13N3O3Cl2 C13H19N3O4 C14H6N2O2Cl2F4 C9H8CI3NO2S C7H7CI3NO3PS C23H30O4 C5H13NO6S4 Accurate mass molecule 225.0823 164.0950 255.0523 234.1368 330.0361 330.0361 279.1471 225.0823 162.0463 198.0560 270.0884 278.1267 371.9351 205.1341 211.0764 251.1344 201.0361 252.0236 181.0054 190.0776 206.0725 223.0845 212.0716 363.0665 197.1277 509.9784 202.0855 301.9619 257.0722 187.1031 291.0330 217.0061 313.9780 222.0674 260.0160 374.9541 217.1500 310.0321 488.0362 253.9661 459.9816 256.0170 329.0334 281.1376 379.9742 298.9341 320.8950 370.2144 310.9626 LOD food mg/kg 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.5 0.5 0.5 0.5

Compound Ethiofencarb Fenuron Imidacloprid Lenacil Malathion 1 Malathion 2 Metalaxyl Methiocarb Methomyl Monuron Nitenpyram Oxadixyl Profenfos Promecarb Propachlor Prosulfocarb Thiabendazole Thiacloprid Thiocyclam Aldicarb Aldicarb sulfoxide Bendiocarb Chlorotoluron Flufenacet Hydroxyatrazine Lufenuron Metamytron Methidathion Methiocarb sulfone Molinate Parathion ethyl Propanil Triclocarban Aldicarb sulfone Bromacil Bromuconazole Butylate Diflubenzuron Flufenoxuron Fluroxypyr Hexaflumuron Imazalil degradate Iprodione Pendimethalin Teflubenzuron Captan Chloropyrifos methyl Spiromesifen Thiosultap

Figure 3.

Example of a report from the molecular feature database search.

data are excellent for the confirmation of the molecular formula for imazalil (however not for the identification of imazalil). The mass accuracy was 0.8 mDa off or 2.7 ppm error. Thus, this is an example of how one confirms the database screening of a pesticide molecular formula. The database report prints the isotopic signature and accurate masses to help for the manual screening of the database hits. When the formula does not contain an A+2 atom (that is, consists of only C, H, N, and O), then formula confirmation is not readily possible by this method. The reason being that only an A+1 peak is present and this peak is dominated by the carbon 13 signal of the molecule, which in itself is not enough for formula confirmation. Thus, for these pesticides (about 30% of the database) more data are required (specifically, retention time or fragment ion). Screening of Fruits and Vegetables Table 2 shows the results of screening six fruit and vegetable samples from a nearby grocery store (apple, pear, tomato, potato, pepper, and cucumber) and one commercial brand of olive oil for the 600 pesticides in the MF database search. The MF database search found from 617 to 2,681 accurate mass peaks in the sample chromatograms. The least complicated sample matrix was the tomato with 617 peaks, and the apple was the most complex sample with 2,681 peaks. The sensitivity of the MF database search was set at a signal-to-noise ratio of 10:1. The quantity of peaks found approximately doubles with decreasing the signal-to-noise

ratio from 20:1 to 10:1. The value of 10:1 is chosen in order to obtain good values for the isotopic signature of the compound, which is the A+1 and A+2 isotope signatures with the maximum instrument sensitivity. The accuracy window of the MF database search is set at 5 ppm in order to be well within the mass accuracy of the LC/TOF MS system, which typically operates at less than 3 ppm and often at 1 to 2 ppm or approximately 0.3 mDa [3]. The number of pesticides found in the 5 ppm mass window of these samples varied from 8 to 41 compounds. The only criterion to be included in this match was that the MH+ ion was within 5 ppm of the database value. Thus, as an example, the pepper sample, which had 2,402 peaks, had only 41 of these peaks that met the 5-ppm accuracy window (Table 2). Of these 41 peaks only three formulae were confirmed based on the correct isotope signature and retention time match (for compounds not containing an A+2 isotope), which was checked not only in the printout of the automated database match, but also by manual confirmation of the data file. The confirmation of pesticides varied from no detections in the potato sample, one pesticide in olive oil, three pesticides in pepper and tomato, and five pesticides in the cucumber and apple. The most common compound found in the fruit and vegetable samples was imazalil, which is a postharvest fungicide used for transport and storage of fruits and vegetables before their sale. Other compounds included organophosphate insecticides, such as diazinon, phosmet, and malathion and the
7

Cl

+1.997
H2C

A+2
N NH

Cl

1 37Cl

+1.997

1 35Cl A+4

Figure 4.

Isotope cluster of the m/z 297 ion and imazalil structure in a pear extract.

oxon of malathion, which is a pesticide degradate. The insect growth regulator buprofezin was found in a tomato sample, as was thiophanate methyl and carbendazim, both fungicides. The software originally used for this work did not address saturated peaks and compounds at high concentrations such as imazalil in the pear sample and buprofezin in the tomato sample were incorrectly identified. The latest software handles saturated peaks and correctly identified these compounds. The accuracy of all confirmed samples had an absolute-value average of 0.3 mDa or 1.2 ppm and a standard deviation of 0.25 mDa and 1.0 ppm, respectively (Table 2 ). The absolute-value average for retention time match was 0.07 minutes and standard deviation of 0.09 minutes. Thus, the windows chosen for the database are chosen with enough margin of error to find 99% of the compounds based on two standard deviations of the mean for mass accuracy and retention time.

Screening of Baby-Food Samples Of the 100 tested compounds, 33 met the baby-food screening level of 0.01 mg/kg. The only compound detected by the database search was a trace level of imazalil in the puree of pear, banana, and orange. Furthermore, examination of the label shows that lemon juice was also used in the baby food, which was a possible source for the imazalil. The, concentration though was approximately 0.0005 mg/kg, which is considerably less than the levels for baby-food safety of 0.01 mg/kg. Imazalil is easily screened because it contains the characteristic A+2 chlorine signature with two chlorine atoms. The error in identification was 0.3 mD or 1 ppm. None of the other compounds of the database was detected. Confirmation was not possible on this sample because of the low signal of the fragment ion and the low concentration of the suspected imazalil (0.0005 mg/kg). The sample was screened as safe, though, based on the health limit of 0.01 mg/kg for baby food. Approximately 10 different baby-food samples have been

Table 2.

Screened Pesticides in Food Samples Using the MF Database Peaks screened 2681 Pesticide matches < 5 ppm 12 Pesticides confirmed LC/TOF MS Imazalil Imazalil degradate Iprodione Fluqinconazole Difenoconazole Terbuthylazine (Deisopropylatrazine) Imazalil Diazinon Buprofezin Buprofezin (Benzthiazuron) Carbendazim Thiophanate methyl Thiabendazole Malathion isomer 1 Malathion isomer 2 Malathion oxon Imazalil Imazalil Carbendazim Imazalil degradate Phosmet None 0.2 0.04 0.22 0.25 0.1 0.21 0.51 0.31 1 0.1 0.7 0.8 0.3 1.1 2 1 0.01 0.05 0.03 0.08 0.05 0.05 0.03 0.04 0 0 Error mDa 1 0.22 0.05 0.74 0.06 0.3 0.11 1 Error ppm 3.9 0.7 0.1 1.8 0.2 1 0.3 3.3 Retention time error (min.) 0.08 0.12 FALSE negative 0 FALSE positive 0

Sample Apple

Olive oil Pepper

1678 2402

10 41

0.09 0.07 0.12 0.1

0 0

0 1

Tomato

617

0 1

Cucumber

1619

17

Pear

1209

14

Potato

1150

11

Mean Standard deviation

0.30 0.25

1.20 1.00

screened, including a variety of brand-name vegetables and fruits and, fortunately, no positive detections have been found for the pesticides in the MF database search with the exception of imazalil shown above. The baby food samples represent the most difficult samples to screen because of the low LODs required. Weakness and Strengths of MF Database Search The only weakness of the database is the loss of mass accuracy because of interferences in the matrix. This problem is easily solved by the addition of a second ion (fragment ion) or a sodium or ammonium adduct ion for added confidence from matrix interference. The use of accurate mass LC/TOF MS combined with database searching is a powerful example of a new wave of monitoring technology for identification of pesticides in food and water. The use of

classical fragmentation libraries with comparison of fragmentation patterns is not needed in LC/TOF MS. Rather the use of molecular formula and the calculated accurate mass, especially when combined with one or two fragment ions of accurate mass, gives the identity of compounds without the worry of the intensity of the fragment ions and how this may vary from instrument to instrument and matrix to matrix. For example, the Molecular Hunter Software works in conjunction with the Molecular Feature Database using the .mhd files to link ions in groups according to their exact retention time (matching within 0.005 minutes). Therefore, it is possible to find and differentiate the fragment ions of a pesticide, such as imazalil, from the background ions of the matrix. Thus, it is the view of the authors that a large problem in LC/MS libraries is on the verge of being solved with the use of accurate mass

www.agilent.com/chem
databases for pesticide screening in food with a molecular feature algorithmic approach using the MH+ ion and a major fragment ion.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

References
1. PAN Europe position on the European Commission Proposal for a Regulation of the European Parliament and of the Council on maximum residue levels of pesticides in products of plant and animal origin COM(2003) 117 final, 2003/0052 (COD). 2. Phillip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN. 3. Imma Ferrer, E. M. Thurman, 2005, Measuring the mass of an electron by LC/TOF MS: A study of twin ions, Analytical Chemistry, 77, 33943400. 4. E. Michael Thurman, Imma Ferrer, 2005, Identification of unknown pesticides in food using both LC/MSD TOF and ion trap MSn, Agilent Technologies, publication 5989-1924EN. 5. M. Anastassiades, S. J. Lehotay, D. Stajnbaher, and F. J. Schenck, Fast and Easy Multiresidue Method Employing Acetonitrile Extraction/Partitioning and Dispersive Solid-Phase Extraction for the Determination of Pesticide Residues in Produce, (2003) Journal of AOAC International, 86:412431. 6. S.J. Lehotay, K. Matovsk, A.R. Lightfield, Use of Buffering and Other Means to Improve Results of Problematic Pesticides in a Fast and Easy Method for Residue Analysis of Fruits and Vegetables, (2005) Journal of AOAC International, 88:615629. 7. Imma Ferrer, Juan F. Garcia-Reyes, M. Mezcua, E. M. Thurman, A. R. Fernandez-Alba, 2005, Multiresidue pesticide analysis in fruits and vegetables by liquid chromatography time-offlight mass spectrometry, J. Chromatography A, 1082, 8190.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA October 26, 2006 5989-5496EN

Multiresidue Analysis of 100 Pesticides in Food Samples by LC/Triple Quadrupole Mass Spectrometry Application

Food Safety

Authors
Imma Ferrer and E. Michael Thurman Pesticide Residue Research Group University of Almeria Almeria, Spain Yanyan Fang, Paul Zavitsanos, and Jerry A. Zweigenbaum Agilent Technologies, Inc. USA

Introduction
In recent years, the established regulations regarding the maximum residue limits (MRLs) in commodities have become more and more stringent. The European Union (EU) has set new directives for pesticides at low levels in vegetables in order to meet health concerns. For fruits and vegetables intended for production of baby food, an MRL of 10 g/kg is applicable for all pesticides, and compounds without a stated regulation also have the lowest MRLs at 10 g/kg. The low MRLs have encouraged the development of more sensitive analytical methods to meet the requirements in complex samples. In this sense, liquidchromatography tandem-mass spectrometry (LC-MS-MS) with triple quadrupole in multiple reaction monitoring (MRM) mode has become, so far, the most widely used technique for the monitoring and quantitation of pesticides in food, as reported extensively in the literature. On the other hand, high-resolving power mass spectrometric techniques, such as time-of-flight mass spectrometry (TOF-MS), have been applied recently for screening purposes as well. Nevertheless, the simplicity of methodologies using triple quadrupole as a detection technique, together with the low limits of detection achieved and the MS/MS capability make this technique a valuable tool for routine

Abstract
An analytical methodology for confirming the presence of a group of 100 pesticides in vegetable and fruit samples was developed using the Agilent G6410AA Triple Quadrupole Mass Spectrometer (QQQ). One transition per parent compound was monitored in a single chromatographic run containing two time segments. The sensitivity obtained meets the maximum residue levels (MRLs) established by the European Union regulation for food monitoring programs. The analytical performance of the method was evaluated for different types of fruit and vegetables + orange, tomato, and green pepper + showing little or no matrix effects. Linearity of response over two orders of magnitude was demonstrated (r > 0.99). This study is a valuable indicator of the potential of the QQQ for routine quantitative multiresidue analysis of pesticides in vegetables and fruits.

monitoring programs established in regulatory official laboratories. The easiness of use is sometimes an essential for these types of regulatory agencies, which lack the high-skilled personnel required for more sophisticated techniques such as TOF-MS. Triple quadrupole technology is not new in the sense that it needs to be validated for monitoring purposes and its basis is already wellestablished for routine analysis. Our study in this report is one of the first of its kind to examine the new Agilent Triple Quad for the analysis of pesticides in fruit and vegetables. This topic was chosen because of the relevance of these compounds and their significant use on food commodities. The sensitivity of the QQQ easily meets the levels required by the regulations on pesticides in food.

LC/MS/MS Instrumentation
LC Conditions Column: Agilent ZORBAX Eclipse XDB C-8, 4.6 mm 150 mm, 5 m, (p/n 993967-906). 25 C A = 0.1% formic acid in water B= Acetonitrile 0.6 mL/min 10% B at 0 min 10% B at 5 min 100% B at 30 min 1-5 L

Column temperature: Mobile phase: Flow-rate: Gradient:

Injection volumes: MS Conditions Mode:

Experimental
Sample preparation Pesticide analytical standards were purchased from Dr. Ehrenstorfer (Ausburg, Germany). Individual pesticide stock solutions (around 1,000 g/mL) were prepared in pure acetonitrile or methanol, depending on the solubility of each individual compound, and stored at 18 C. From these mother solutions, working standard solutions were prepared by dilution with acetonitrile and water. Vegetable samples were obtained from the local markets. Blank vegetable and fruit extracts were used to prepare the matrix-matched standards for validation purposes. In this way, two types of vegetables and one fruit (green peppers, tomatoes, and oranges) were extracted using the QuEChERS method already described in a previous application [1]. The vegetable extracts were spiked with the mix of standards at different concentrations (ranging from 2 to 100 g/kg) and subsequently analyzed by LC/MS/MS.

Positive ESI using the Agilent G6410AA Triple Quadrupole Mass Spectrometer Nebulizer: 40 psig Drying gas flow: 9 L/min V capillary: 4000 V Drying gas temperature: 350 C Q1 resolution: Unit Q2 resolution: Unit Fragmentor voltage: 70 V Collision energy: 525 V MRM: 1 transition for every compound as shown in Table 1 Dwell time: 15 msec

Results and Discussion


Optimization of LC/MS/MS conditions A preliminary study of the optimal MRM transitions for every compound was carried out by injecting groups of analytes (around 10 analytes in one chromatographic run) at a concentration level of 10 g/mL. Various collision energies (5, 10, 15, 20, and 25 V) were applied to the compounds under study. The optimum energies were those that gave the best sensitivity for the main fragment ion and, as a general rule, left about 10% of parent compound in the spectra, and they were selected as optimum ones. Only one fragment ion was chosen as the most abundant product ion for every target compound. Results are shown in Table 1.

Table 1.

Analytical Conditions and Limits of Detection (LOD) for Each of the Compounds Tested Retention time (min) 2.7 2.7 3 4.5 6.4 6.6 7.9 10.8 11 11.2 11.5 11.9 12.5 13.9 14.5 14.8 14.8 15.4 15.5 15.7 16 16.4 16.9 17 17.2 17.2 17.5 17.8 17.9 17.9 18 18 18.4 18.5 18.5 18.6 18.7 18.9 19.4 19.5 19.5 19.6 19.7 20 20.1 20.3 20.3 20.4 20.4 20.4 20.5 Protonated molecule [M+H]+ 167 312 150 182 207 192 202 223 271 198 163 174 262 203 165 188 256 230 223 226 214 258 411 253 297 296 213 312 279 255 202 199 235 241 166 298 221 215 213 242 242 222 224 732 202 254 216 280 287 207 432 Product ion (m/z) 125 232 105 137 89 160 175 148 225 156 88 132 234 175 72 146 209 199 126 184 158 122 182 126 159 264 89 284 219 209 132 72 153 214 109 144 109 187 72 200 186 165 167 142 145 198 174 248 123 72 290 Collision energy 20 10 15 10 5 15 25 5 10 15 5 15 15 15 15 15 10 5 15 20 15 5 15 15 15 20 10 20 10 10 15 10 10 10 5 15 15 15 15 20 15 10 5 5 5 15 15 10 15 15 15 LOD (pg) 10 90 10 8 9 5 10 50 7 3 4 18 8 8 2 4 7 7 6 4 0.8 6 6 3 7 2 10 15 10 120 5 2 20 70 2 10 10 5 3 2 1 2 2 12 2 0.1 0.3 5 5 1 6 3

Compound name Segment 1 Cyromazine Thiosultap Cartap Thiocyclam Aldicarb sulfoxide Carbendazim Thiabendazole Aldicarb sulfone Nitenpyram Hydroxyatrazine Methomyl Deisopropylatrazine Imazapyr Metamitron Fenuron Deethylatrazine Imidacloprid Dimethoate Acetamiprid Prometon Irgarol metabolite Methiocarb sulfone Nicosulfuron Thiacloprid Imazalil Mebendazole Aldicarb Imazaquin Oxadixyl Fluroxypyr Simazine Monuron Lenacil Cyanazine Metolcarb Spiroxamine Dichlorvos Metribuzin Chlorotoluron Prometryn Terbutryn Carbofuran Bendiocarb Segment 2 Spinosad A Carbaryl Irgarol 1051 Atrazine Metalaxyl Difenoxuron Isoproturon Bensultap

Table 1.

Analytical Conditions and Limits of Detection (LOD) for Each of the Compounds Tested (Continued) Retention time (min) 20.5 20.7 20.7 21.3 21.6 21.7 21.9 22.2 22.5 22.6 22.7 22.8 23 23 23.2 23.2 23.2 23.3 23.3 23.7 23.7 24.1 24.6 24.7 24.8 24.9 24.9 24.9 25 25.1 25.2 25.3 25.4 25.5 25.8 26.2 26.4 26.5 26.7 26.8 26.9 27.1 27.6 27.9 28 28.5 28.7 29.2 29.7 Protonated molecule [M+H]+ 233 746 226 388 212 388 376 218 292 226 230 376 304 303 404 318 300 276 208 376 188 311 330 342 331 342 284 346 270 270 364 406 406 359 326 292 315 461 306 305 381 322 373 511 252 489 218 282 336 Product ion (m/z) 72 558 107 301 170 301 308 162 70 169 174 159 217 145 372 160 264 244 151 159 126 158 245 159 127 159 252 278 238 224 194 251 251 155 294 236 162 158 201 169 158 212 303 158 91 158 57 212 236 Collision energy 15 5 5 20 10 20 10 15 10 5 15 20 15 5 10 5 10 10 10 20 10 10 10 20 5 20 10 10 10 10 5 20 20 10 5 10 15 10 10 15 15 15 10 10 15 10 10 5 15 LOD (pg) 5 100 5 11 1 8 6 10 6 15 0.3 6 0.7 5 0.4 2 50 1 5 6 5 9 8 5 5 5 2 7 8 8 5 4 4 8 5 9 8 7 1 1 22 15 7 10 2 6 2 5 30

Compound name Diuron Spinosad D Ethiofencarb Dimethomorph isomer 1 Propachlor Dimethomorph isomer 2 Prochloraz Propanil Cyproconazole Methiocarb Terbutylazine Bromuconazole isomer 1 Fenamiphos Methidathion Azoxystrobin Phosmet Captan Dimethenamide Promecarb Bromuconazole isomer 2 Molinate Diflubenzuron Iprodione Propiconazole isomer 1 Malathion Propiconazole isomer 2 Metolachlor Triflumizole Alachlor Acetochlor Flufenacet Difenoconazole isomer 1 Difenoconazole isomer 2 Chlorfenvinphos Benalaxyl Parathion ethyl Triclocarban Hexaflumuron Buprofezin Diazinon Teflubenzuron Chlorpyrifos methyl Profenofos Lufenuron Prosulfocarb Flufenoxuron Butylate Pendimethalin Trifluralin

The MRM transitions were included in the method with a dwell time of 15 msec, and two different time segments were recorded in the chromatographic run (each one of them containing about half of the pesticides studied). Figure 1 shows the chromatogram corresponding to 100 pg on column for all the compounds studied. Extracted ion chromatograms are overlaid for each one of the target analytes according to their respective protonated molecule and product ion MRM transition. Linearity and Limits of Detection Linearity was evaluated by analyzing the standards solutions at five different concentration levels in the range 2 to 100 pg on column. As an

example, the calibration curve generated for atrazine is shown in Figure 2. As it can be observed in this figure, the linearity of the analytical response across the studied range is excellent, with a correlation coefficient of 0.998. Similar results were obtained for the rest of the compounds analyzed. The limits of detection (LOD) were estimated from the injection of standard solutions at concentration levels corresponding to a signal-to-noise ratio of about 3. The results obtained are included in Table 1 as well. The best limits of detection were obtained for the triazines (from 100 fg to 2 pg on column) and the highest limits of detection were for fluoroxypyr and spinosad D (above 100 pg).

x105 1.6 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

+ MRM (282.0 212.0) mix100_100 pg_5May.d

12

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Abundance vs. acquisition time (min)

Figure 1.

Product ion chromatograms of a mix of 100 pesticides (concentration: 100 pg on column).

500000 400000 Area 300000 200000 100000 0 0

y = 3828.6x % 1790.3 R2 = 0.9988

20

40

60 Amount injected (pg)

80

100

120

Figure 2.

Calibration curve for atrazine using a linear fit with no weighting and no origin treatment.

www.agilent.com/chem
Application to Vegetable Matrices To confirm the suitability of the method for analysis of real samples, matrix-matched standards were analyzed in three different matrices + green pepper, tomato, and orange + at two different concentration levels (10 and 100 g/kg). Figure 3 shows the analysis of a green pepper spiked with the pesticide mix at 10 g/kg (10 pg on column). As it can be observed in two of the MS/MS extracted product ion chromatograms, for dimethoate and azoxystrobin, compounds can be easily identified in these complex matrices due to the selectivity of the MRM transitions, thus fulfilling the regulation limits imposed by the EU directives. In general, the LOD obtained meet the requirements regarding the MRLs imposed by the existing European regulations.

Reference
1. Imma Ferrer and E. Michael Thurman, Determination of Fungicides in Fruits and Vegetables by Time-of-Flight and Ion Trap LC/MS (2005) Agilent Technologies, publication 5989-2209EN www.agilent.com/chem.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

x104 1 1.8 1.4 1 0.6 0.2 0 x103 1 3 2 1 0 x103 5 4 3 2 1 0 1

+ TIC MRM (** **) mix100_10pg_green pepper.d

12

+ MRM (230.0 199.0) mix100_10pg_green pepper.d

12

(a)

+ MRM (404.0 372.0) mix100_10pg_green pepper.d

12

(b)

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Abundance vs. acquisition time (min)

Figure 3.

MRM chromatogram of a spiked green pepper sample at 10 g/kg. Product ion chromatograms for (a) dimethoate and (b) azoxystrobin.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA August 9, 2006 5989-5469EN

Improving the Effectiveness of Method Translation for Fast and High Resolution Separations Application

Author
Michael Woodman Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
The increased availability of sub-2-micron (STM) columns and increased demand for methods friendly to mass spectrometers has led to strong trend toward conversion of existing HPLC methods to smaller diameter and smaller particle size columns. While the conversion is a simple mathematical exercise requiring the scaling flow rates, gradient times and injection volumes, many users observe less than perfect results. Here we look closely at the problem and propose calculations that improve the speed and/or resolution in a more predictable and beneficial way.

Simplistically, a column of 250-mm length and containing 5-m particles can be replaced by a 150-mm length column packed with 3-m particles. If the ratio of length to particle size is equal, the two columns are considered to have equal resolving power. Solvent consumption is reduced by L1/L2, here about 1.6-fold reduction in solvent usage per analysis. If an equal mass of analyte can then be successfully injected, the sensitivity should also increase by 1.6-fold due to reduced dilution of the peak as it travels through a smaller column of equal efficiency. LC/MS (Liquid Chromatography/Mass Spectrometry) ionization sources, especially the electrospray ionization mode, have demonstrated greater sensitivity at lower flow rates than typically used in normal LC/UV (UltraViolet UV/VIS optical detection) methods, so it may also be advantageous to reduce the internal diameter of a column to allow timely analysis at lower flow rates. The relationship of flow rate between different column diameters is shown in Equation 1.
Flowcol. 1 Diam.column2 Diam.column1
2

Introduction
Methods developed on older columns packed with large 5- or 10-m particles are often good candidates for modernization by replacing these columns with smaller dimension columns packed with smaller particle sizes. The potential benefits include reduced analysis time and solvent consumption, improved sensitivity and greater compatibility with mass spectrometer ionization sources.

= Flowcol. 2

(eq. 1)

The combined effect of reduced length and diameter contributes to a reduction in solvent consumption and, again assuming the same analyte mass can be injected on the smaller column, a proportional increase in peak response. We normally scale the injection mass to the size of the column,

though, and a proportional injection volume would be calculated from the ratio of the void volumes of the two columns, multiplied by the injection volume on the original column.
Inj. vol.col. 1 Volumecolumn2 Volumecolumn1 = Inj. vol.col. 2 (eq. 2)

For isocratic separations, the above conditions will normally result in a successful conversion of the method with little or no change in overall resolution. If one wishes to improve the outcome of the method conversion, though, there are several other parameters that should be considered. The first of these parameters is the column efficiency relative to flow rate, or more correctly efficiency to linear velocity, as commonly defined by van Deemter [1] and others, and the second is the often overlooked effect of extracolumn dispersion on the observed or empirical efficiency of the column. Van Deemter observed and mathematically expressed the relationship of column efficiency to a variety of parameters, but we are most interested here in his observations that there is an optimum linear velocity for any given particle size, in a wellpacked HPLC column, and that the optimum linear velocity increases as the particle size decreases. Graphically, this is often represented in van Deemter plots as shown in Figure 1, a modified version of the original plot [2]. In Figure 1 we observe that the linear velocity at which 5-m materials are most efficient, under the conditions used by the authors, is about 1 mm/sec. For 3.5-m materials the optimum linear velocity is about 1.7 mm/sec and has a less distinct opti0.02

mum value, suggesting that 3.5-m materials would give a more consistent column efficiency over a wider flow range. For the 1.8-m materials, the minimum plate height, or maximum efficiency, is a broad range beginning at about 2 mm/sec and continuing past the range of the presented data. The practical application of this information is that a reduction in particle size, as discussed earlier, can often be further optimized by increasing the linear velocity which results in a further reduction in analysis time. This increase in elution speed will decrease absolute peak width and may require the user to increase data acquisition rates and reduce signal filtering parameters to ensure that the chromatographic separation is accurately recorded in the acquisition data file. The second important consideration is the often overlooked effect of extracolumn dispersion on the observed or empirical efficiency of the column. As column volume is reduced, peak elution volumes are proportionately reduced. If smaller particle sizes are also employed there is a further reduction in the expected peak volume. The liquid chromatograph, and particularly the areas where the analytes will traverse, is a collection of various connecting capillaries and fittings which will cause a measurable amount of bandspreading. From the injector to the detector flow cell, the cumulative dispersion that occurs degrades the column performance and results in observed efficiencies that can be far below the values that would be estimated by purely theoretical means. It is fairly typical to see a measured dispersion of 20 to 100 L in an HPLC system. This has a disproportionate effect on the smallest columns and smallest particle sizes, both of which are expected to yield the smallest

Plate height (mm)

0.015

0.01

0.005

5.0 m SB-C18 3.5 m SB-C18 1.8 m SB-C18

Lin. vel. 4.6 mm 3 mm 2.1 mm 1 mm

mm/sec mL/min mL/min mL/min mL/min

1 0.7 0.3 0.14 0.033

2 1.4 0.6 0.29 0.066

3 2.1 0.9 0.44 0.1

4 2.8 1.2 0.58 0.133

5 3.5 1.5 0.73 0.166

Figure 1.

van Deemter plot with various flow rates and particle sizes.

possible peak volumes. Care must be taken by the user to minimize the extracolumn volume and to reduce, where practical, the number of connecting fittings and the volume of injection valves and detector flow cells. For gradient elution separations, where the mobile phase composition increases through the initial part of the analysis until the analytes of interest have been eluted from the column, successful method conversion to smaller columns requires that the gradient slope be preserved. While many publications have referred to gradient slope in terms of % change per minute, it is more useful to express it as % change per column volume. In this way, the change in column volume during method conversion can be used to accurately render the new gradient condition. If we think of each line of a gradient table as a segment, we can express the gradient by the following equation:

Experimental Conditions
System Agilent 1200 Series Rapid Resolution LC consisting of: G1379B micro degasser G1312B binary pump SL G1367C autosampler SL, with thermostatic temperature control G1316B Thermostatted column compartment SL G1315C UV/VIS diode array detector SL, flow cell as indicated in individual chromatograms ChemStation 32-bit version B.02.01 Columns Agilent ZORBAX SB-C18, 4.6 mm 250 mm, 5 m Agilent ZORBAX SB-C18, 3.0 mm 150 mm, 3.5 m Mobile phase conditions Organic solvent: Aqueous solvent: Gradient Conditions Gradient slope: 7.8% or 2.3% per column volume, as indicated. See individual chromatograms for flow rate and time Acetonitrile 25 mm phosphoric acid in Milli-Q water

% Gradient slope =

(End% Start%) #Column volumes

(eq. 3)

Sample Standard mixture of chlorinated phenoxy acid herbicides, 100 g/mL in methanol

Note that the use of % change per column volume rather than % change per minute frees the user to control gradient slope by altering gradient time and/or gradient flow rate. A large value for gradient slope yields very fast gradients with minimal resolution, while lower gradient slopes produce higher resolution at the expense of increased solvent consumption and somewhat reduced sensitivity. Longer analysis time may also result unless the gradient slope is reduced by increasing the flow rate, within acceptable operating pressure ranges, rather than by increasing the gradient time. Resolution increases with shallow gradients because the effective capacity factor, k*, is increased. Much like in isocratic separations, where the capacity term is called k', a higher value directly increases resolution. The effect is quite dramatic up to a k value of about 5 to 10, after which little improvement is observed. In the subsequent examples, we will see the results associated with the calculations discussed above.

Results
The separation was initially performed on a standard 4.6 250 mm, 5-m ZORBAX SB-C18 column thermostatted to 25 C (Figure 2) using conditions referenced in US EPA Method 555. The method was then scaled in flow and time for exact translation to a 3.0 150 mm, 3.5-m column (Figure 3). Solvent consumption is reduced from 60 mL to 15.5 mL per analysis. The separation was then re-optimized for faster separation with the identical slope, 7.8%, by increasing the flow rate from 0.43 to 1.42 mL/min, and proportionately reducing the gradient time (Figure 4). Finally, increased resolution is demonstrated by keeping the original times used in Figure 3 with the increased flow rate (Figure 5). This yields a gradient with identical time but a reduced slope of 2.3%. The increased resolution of peaks 4 and 5 is readily apparent. The conditions in Figure 4, 7.8% slope at increased linear velocity on 3.0 150 mm, 3.5-m material, yield a separation with comparable resolution to the original 4.6 250 mm method, but with only a 12-minute total analysis time. This is excellent for

12.557

mAU 350 300 250 200 150 100 50 0

14.380

17.779

18.871

19.414

21.063

23.050

24.667

13.194

12.5

15

17.5

17.607

20

22.5

25

27.5

Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 25 C Gradient: 10% to 90% ACN vs. 25 mM H3PO4 Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1 mL/min Group A Compounds Total analysis time: Detection: Figure 2. 60 min UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default)

Gradient separation of herbicides on 4.6 250 mm 5-m ZORBAX SB-C18.

800 700 600 500 400 300 200

9.990

mAU

Conditions: EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 3.0 mm 150 mm, 3.5 m Column temp: 25 C Gradient: 25 mm H3PO4/ACN, 0% to 90% ACN in 18 minutes Gradient slope: 7.8% ACN/column volume Analysis flow rate: 0.43 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 36 min.
12.061 12.831 16.314

8.781

14.106

13.046

15.317

15.786

17.081

0 8 10 12 14 16 18 min

Figure 3.

Gradient separation of herbicides on 3.0 150 mm, 3.5-m ZORBAX SB-C18.

13.854

100

18.348

9.120

29.595 min

400

300

2.674

3.011

mAU

Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18, 3.0 mm 150 mm, 3.5 m Column temp: 25 C Gradient: 25 mM H3PO4/ACN, 10% to 90% ACN in 5.4 min. Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1.42 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 12 min.

3.620

3.850 3.919

4.240

200

2.780

0 2 2.5 3 3.5 4 4.5 5 5.5 min

Figure 4.

High speed gradient separation of herbicides on 3.0 150 mm, 3.5-m ZORBAX SB-C18.

3.964

mAU 400 350 300 250

Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18, 3.0 mm 150 mm, 3.5 m Temp: 25 C Gradient: 25 mM H3PO4/ACN, 10% to 90% ACN in 18 min. Gradient slope: 2.3% ACN/column volume Analysis flow rate: 1.42 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 36 min.
6.793 7.583 11.291 10.257 11.465 8.905

150 100

0 4 6 8 10 12 min

Figure 5.

Reduced slope gradient separation of herbicides on 3.0 150 mm, 3.5-m ZORBAX SB-C18.

12.692

4.093 4.253

50

7.912

10.056

200

4.933

4.743

100

4.611

4.914

www.agilent.com/chem
high throughput screening and quantitation of a large number of samples. Figure 5, with the gradient slope reduced to 2.3%, results in a high-resolution separation with a calculated R value of 3.3 vs. the standard 3.0 150 mm separation value of 1.9, for the critical pair seen in Figure 5 at 7.5 to 8 minutes. In Table 1 the column has been replaced with a low dead volume connecting union in a system fitted with 0.12-mm id capillary tubing at all points of sample contact. A 1-L injection of dilute actone
Table 1. Volumetric Measurements of Various Flow Cells Elution volume (L) 11 14 Half height width (L) 5 6 5 Sigma width (L) 12 18

Conclusion
Careful analysis of the existing gradient conditions, coupled with an awareness of the need to accurately calculate new flow and gradient conditions can lead to an easy and reliable conversion of existing methods to new faster or higher resolution conditions. In addition, awareness of extracolumn dispersion, especially with small and high resolution columns, will ensure good column efficiency which is critical to a successful translation of the method.

References
1. J. J. van Deemter, F. J. Zuiderweg, A. Klinkenberg, Chemical Engineering Science 1956, 5, 271289 2. The Influence of Sub-Two Micron Particles on HPLC Performance, Agilent Technologies, application note 5989-9251EN, May 2003

Flow cell New SL 2 L 3 mm Micro 6 mm 1.7 L (n = 2) Semi-micro 6 mm 5 L (n = 2) Standard 10 mm 13 L New SL 10 mm 13 L

13

6.5

18.5

26 27

11 11

26 25

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

is made to determine the bandspreading contribution of the system, with various flow cells. Multiple flow cells were tested, and the average result reported, where possible. The elution volume summarizes the total volume of all tubing in the system. While the absolute volume from the 2-L to the 13-L flow cells is 11 L, we observe an increase of 15 to 16 L because of the larger diameter inlet tubing integral to the larger volume flow cells.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA August 9, 2006 5989-5177EN

Determination of 44 Pesticides in Foodstuffs by LC/MS/MS Application

Food Safety

Authors
Masahiko Takino Agilent Technologies, Inc. Hachioji, Tokyo Japan Toshitsugu Tanaka Kobe Institute of Health Department of Food Chemistry Chuo-ku, Kobe Japan

Abstract
A sensitive and selective analytical method for the determination of 44 pesticide residues in several foodstuffs using the Agilent G6410AA Triple Quadrupole Mass Spectrometer (QQQ) was developed. This method use two different sample preparation methods followed by LC/MS/MS (liquid chromatography/tandem mass spectrometry). The limits of detection for all pesticides were less than 10 ng/mL in foodstuff. The sensitivity of QQQ easily met the maximum residue limits (MRLs) of all investigated pesticides in Japan Food Hygiene Law.

safety. In recent years, the established regulations regarding MRLs in commodities have been more and more stringent. In Japan, the positive list system was introduced this year, and MRLs have been set for over 500 pesticides in all foodstuffs. This new system sets different MRLs for each pesticide within each food group. Typically, the MRLs range from 0.01 to 3 g/g depending on the commodities and pesticides. The low MRLs fostered the development of more sensitive analytical methods to meet the requirements of complex samples. In this sense, liquid chromatography/tandem mass spectrometry (LC/MS/MS) with QQQ in multiple reaction monitoring (MRM) mode has become so far the most widely used techniques for the quantitation of polar pesticides in food. MRM mode provides for more specific detection in a complex matrix such as food. In this work, 44 pesticides (Tables 1 and 2) are analyzed in two separate runs with sample analytical conditions. The sensitivity requirements set by the positive lost system for these pesticides are easily met.

Experimental
Chemicals The acetonitrile was of LC/MS grade from Wako Pure Chemical Ind (Japan). Toluene, acetone, nhexane, formic acid, sodium chloride, and anhydrous sodium sulfate were of analytical grade from Wako Pure Chemical. All SPE cartridges were purchased from Spelco Japan (Japan). Pesticide standards were obtained from Hayashi Pure Chemical (Japan).

Introduction
Pesticides are widely used in agricultural practices. The main application can be classified in production and post-harvest treatment of agricultural commodities for transport purposes. In this sense, production agriculture comprises the main category of use of pesticides subject to control requirements and, therefore, maximum residue levels (MRLs) have been fixed to assess food

Sample Preparation
Extraction

LC/MS/MS Instrument The LC/MS/MS system used in this work consists of an Agilent 1100-series vacuum degasser, binary pump, well-plate autosampler, thermostatted column compartment, and the Agilent G6410 Triple Quadrupole Mass Spectrometer with an electrospray ionization source (ESI). The objective of the method development was to obtain a fast and sensitive analysis for quantifying pesticides in fruits and vegetables. For chromatographic resolution and sensitivity, different solvents and columns were optimized. It was found that a simple solvent system using water, acetonitrile, formic acid, formic acetate, and a 1.8-m particle size C18 column would work very well.
LC Conditions Instrument: Column: Column temp: Mobile phase: Agilent 1100 HPLC ZORBAX Extend C18, 100 mm 2.1 mm, 1.8 m (p/n 728700-902) 40 C A = 0.1% formic acid +5 mM ammonium formate in water B= Acetonitrile 10% B at 0 min, 80% B at 30 min 0.2 mL/min 5 L Agilent 6410 QQQ Positive ESI 10 L/min 50 psig 350 C 4000 V m/z 100 to 550 Variable 100 V Shown in Tables 1 and 2 Shown in Tables 1 and 2

Vegetable and fruit samples were obtained from the local markets. A sample of 10 to 500 g was chopped in a food processor to obtain thoroughly mixed homogenates. A 20-g portion of sample homogenate was weighed in a 200-mL PTFE centrifuge tube. Then 50 mL of acetonitrile was added and blended in a Polytoron. The extract was then filtered by applying vacuum. The filtrate was collected and the residue was re-extracted with 20 mL of acetonitrile. The filtrates were combined in a 100-mL volumetric flask and made up to volume with acetonitrile. A 20-mL portion of the extract was transferred into a PTFE centrifuge tube, and 10 g of NaCl and 20 mL of 0.5 M phosphate buffer (pH 7.0) were added to the extract followed by shaking for 5 min. Five grams of anhydrous Na2SO4 were added to the acetonitrile layer obtained after salting out. After removing anhydrous Na2SO4, the extract was evaporated to dryness by rotary evaporator (water bath temperature did not exceed 40 C). The residue was dissolved in 2 mL of acetonitrile-toluene (3:1).
Cleanup

Group 1 - The extract was loaded into a GCB/amino propyl SPE cartridge (500 ng/500 mg) preconditioned with 10 mL of acetonitrile-toluene (3:1). The 20 mL of acetonitrile-toluene (3:1) was further added to the SPE cartridge. All eluate was collected and evaporated by rotary evaporator. The residue was dissolved in 4 mL of methanol. Group 2 - The extract was loaded into a silica gel SPE cartridge (500 mg) preconditioned with 10 mL each of methanol, acetone, and n-hexane (10 mL of methanol, 10 mL of acetone, and 10 mL of n-hexane, total volume is 30 mL). The 10 mL of acetone-triethylamine-n-hexane (20:0.5:80) was further added to the SPE cartridge. All eluate was discarded. The 20 mL of acetone-methanol (1:1) was applied and the eluate was collected and evaporated by rotary evaporator. The residue was dissolved in 4 mL of methanol. Standard Preparation Stock solutions of individual pesticides were prepared in methanol at 1 g/mL. Serial dilutions using methanol produced a range of standard mixture solutions at 0.001 g/mL to 1 g/mL. The blank matrix residues were fortified with a mixture of pesticides studied at 10 ng/g.
2

Gradient: Flow rate: Injection vol: MS Conditions Instrument: Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: Scan: Fragmentor: MRM ions: Collision energy:

LC/MS/MS Method Quantitative analysis was carried out using MRM mode with time program. The parameters of MRM transition are shown in Tables 1 and 2.

Table 1. No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

Data Acquisition Parameters of MRM Transitions of Each Pesticide in Group 1 Pesticides Thiabendazole Thiamethoxam Clothianidin Chloridazon Imidacloprid Dimethirimol Oxycarboxine Thiacloprid Azamethiophos Ferimzone(E) Ferimzone(Z) Phenmedipham Azinphos-methyl Simeconazole Isoxaflutol Pyriftalid Tridemorph Methoxyfenozide Chromafenozide Fenoxycarb Naproanilide Butafenacil Cyazofamide Anilofos Pyrazolate Benzofenap Cyflufenamid Indoxacarb Clomeprop Cloquincet-mexyl Furathiocarb Lactofen Tralkoxydim RT (min) 5.018 6.16 7.83 8.19 8.39 8.8 11.02 11.03 12.87 13.21 13.7 17.77 17.9 18.5 18.7 18.7 19.21 20.06 20.57 20.63 21.27 21.55 21.7 22.5 23.5 24 24.3 24.37 24.78 24.8 25.7 26.3 26.7 Molecular weight 201 291 249 221 255 209 267 252 324 254 254 317 318 293 359 318 297 312 394 301 291 491 324 367 438 430 412 527 372 335 365 478 329 Precursor ion (m/z) 202 292 250 222 256 210 268 253 325 255 255 318 132 294 360 319 298 313 175 302 292 492 325 368 439 431 413 528 373 336 383 479 330 Product ion (m/z) 175 211 169 104 209 171 175 126 183 124 132 136 77 70 251 139 130 149 141 88 171 331 108 199 173 105 241 150 299 238 195 344 284 Collision energy(V) 20 5 10 10 20 20 10 20 10 20 20 20 15 15 15 20 15 20 20 15 10 20 10 10 15 20 20 15 5 15 15 15 10

Table 2. No 1 2 3 4 5 6 7 8 9 10 11

Data Acquisition Parameter of MRM Transitions of Each Pesticide in Group 2 Pesticides Flumetsulam Thidiazuron Imazaquin Thifensulfuron-methyl Florasulam Forchlorfenuron-methyl Clorasulam-methyl Diclosulam Fomesafen Triflusulfuron-methyl Haloxyfop RT (min) 9.96 11.95 12.25 12.89 13.75 14.63 16.41 16.83 18.27 19.29 19.67 Molecular weight 325 220 311 387 359 247 429 405 438 492 361 Precursor ion (m/z) 326 221 312 388 360 248 430 406 456 493 362 Product ion (m/z) 129 102 267 167 129 129 398 161 344 264 316 Collision energy (V) 20 10 20 10 20 10 10 20 10 15 15

Results and Discussion


Optimization of MRM Transitions Determination of the optimal MRM transitions for each pesticide was carried out using full scan mode followed by product ion scan mode using two pesticide standard mixtures at 1 g/mL. TICs of these standard mixtures in full scan mode and product ion scan mode are shown in Figures 1 and 2. The mass spectrum of each pesticide by full scan mode exhibited protonated molecular ions; [M+H]+ as the base peak ion except azinphos-methyl, furathiocarb, and fomesafen, which exhibited fragment ion and ammonium adduct ion [M+NH4]+. These ions were selected as precursor ions for MRM mode. It was possible to generate individual product ion MS/MS spectrum of each pesticide by using multiple acquisition and time programming mode. As shown in Tables 1 and 2, 10 time segments for 33 pesticides in group 1 and 7 time segments for 11 pesticides in group 2 were used for MRM mode.

Total ion chromatograms of pesticide standard mixture corresponding to the minimum MRL value for pesticides (10 ng/mL) are shown in Figure 3. These show excellent signal-to-noise (S/N) ratios for all pesticides. The limit of detection (LOD) for each pesticide was determined using an S/N ratio of 3 with an MRM chromatogram of each pesticide at 1 ng/mL (see Table 3). To evaluate the linearity of the calibration curves, various concentrations of pesticide standard solutions ranging from 0.001 ng/mL to 1 ng/mL were analyzed. As shown in Table 3, the linearity was very good for all pesticides with correlation coefficients (r2) greater than 0.998 The matrix effect of this method was investigated by using orange, apple, potato, and cabbage extracts spiked with pesticide standards at 10 ng/mL. Typical MRM chromatograms of orange extract are shown in Figures 4 and 5. The other chromatograms of apple, potato, and cabbage extract are shown in Figure 6. There was not additional peak from sample matrix in all food when compared with the pesticide standard mixture. These results indicate that MRM mode has very high selectivity.

x107 1.2

(A)

6 18 19, 20 24 4 2 3 11 9 15,16 13 21 17 22 23

29,30

7,8 0.8 1

10

12

31 27, 28 26 25 32 33

0.4

14

1 x106 6.5

9 6

11

13

15

17

19

21

23

25

27

29

31

Abundance versus acquisition time (min)

(B)
9 10 12 19, 20 18 21 22 23 21 23 33 25 27, 28 26 25 32 27 29 31 29, 30 24 31 7, 8

4.5

2.5 1 0.5 1 3 5

5 34

11

13

14

15, 16 17

11

13

15

17

19

Abundance versus acquisition time (min)

Figure 1.

TIC of 33 pesticides standard in full scan mode (A) and product ion scan mode (B) at 1 g/mL.

x107 4.0

(A)

6 3.0 2.0 1.0 1 2 4 5 7 8 9

10

11

10

11

12 3

13

14

15

16

17

18

19

20

21

22

23

Abundance versus acquisition time (min) x106 3.0

(B)
4 5 6 10

2.0

1.0

7 8 9

11

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Abundance versus acquisition time (min)

Figure 2.

TIC of 11 pesticides standard in full scan mode (A) and product ion scan mode (B) at 1 g/mL.

Furthermore, the change on the peak intensity of each pesticide by sample matrix was calculated by comparing with the peak intensity of pesticide standards. As these results show in Table 4, the relative intensity of each pesticide ranged from 91 to 116%. Thus, matrix effect such as ion suppression may be insignificant and it was possible to use external standards instead of matrix matched standards. The repeatability of each pesticide in orange extract is also shown in Table 4, and the RSD of each pesticide was in the range from 1.7 to 5.9%.

7,8 5000 4000 3000 2000 1 1000 2 4 3 5 7 6 10 11 9 11 13 15 17 14 13 15,16 17 19 22 23 24 21 26

(A)

12 9 18 19,20,21 25 29,30

28 27 25

31 32 33 27 29

23

Abundance versus acquisition time (min) 6 16000 12000 1 8000 4000 23 7 9 1 3 5 7 9 11 13 15 17 19 11 21 23 25 27 29 8

(B)
5 4 10

Abundance versus acquisition time (min)

Figure 3.

TIC of 33 pesticide standards (A) and 11 pesticides standard (B) at 10 ng/mL in MRM mode.

Table 3.

Linearity and LOD of 44 Pesticide Standard Solutions r2 0.9999 0.9992 0.9999 0.9993 0.9995 0.9989 0.9993 0.9991 0.9988 0.9993 0.9995 0.9993 0.9997 0.9992 0.9991 0.9988 0.9991 0.9996 0.9994 0.9992 0.9989 0.9969 0.9977 LOD (ng/mL) <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.34 0.53 <0.1 <0.1 <0.1 <0.1 <0.1 1.21 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 No 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Pesticides Methoxyfenozide Chromafenozide Fenoxycarb Naproanilide Butafenacil Cyazofamide Anilofos Pyrazolate Benzofenap Cyflufenamid Clomeprop Indoxacarb Quinclorac-methyl Furathiocarb Lactofen Tralkoxydim r2 0.9993 0.9992 0.9988 0.9993 0.9994 0.9987 0.9991 0.9990 0.9982 0.9993 0.9993 0.9991 0.9988 0.9987 0.9987 0.9992 LOD (ng/mL) 0.55 0.49 <0.1 <0.1 <0.1 0.43 <0.1 0.51 0.49 0.43 0.61 1.04 0.63 <0.1 1.10 0.52

No Pesticides Group 1 1 Thiabendazole 2 Thiamethoxam 3 Clothianidin 4 Chloridazon 5 Imidacloprid 6 Dimethirimol 7 Oxycarboxine 8 Thiacloprid 9 Azamethiophos 10 Ferimzone(E) 11 Ferimzone(Z) 12 Phenmedipham 13 Azinphos-methyl 14 Simeconazole 15 Isoxaflutol 16 Pyriftalid 17 Tridemorph Group 2 1 Flumetsulam 2 Thidiazuron 3 Imazaquin 4 Thifensulfuron-methyl 5 Florasulam 6 Forchlorfenuron-methtyl

7 8 9 10 11

Clorasulam-methyl Diclosulam Fomesafen Triflusulfuron-methyl Haloxyfop

0.9987 0.9989 0.9989 0.9992 0.9995

<0.1 <0.1 0.32 <0.1 0.19

25000 15000

1
m/z = 202>>>175

700 300

13
m/z = 132>>>77

550 350

2
m/z = 292>>>211

900 500

14
m/z = 294>>>70

300 200

3
m/z = 250>>>169

2400 1200

15
m/z = 360>>>251
10 11 12 13 14 15 16 17 18 19 20

700 300

4
m/z = 222>>>104

250 150 50

16
m/z = 319>>>139

300 100

24

5 m/z = 256>>>209

17
m/z = 298>>>130

12

800 400

2500

18
m/z = 313>>>149

m/z = 210>>>171

1500

10

11

12

13 900

7
2500 1500

19
m/z = 175>>>141

m/z = 268>>>175

500

4000 2000

8
m/z = 253>>>126

900 600

20
m/z = 302>>>88

3000

9
m/z = 325>>>183
7 8 9 10 11 12 13 14 15 16

1100 500

21
m/z = 292>>>171

1000

300 200 100 400 200

10
m/z = 255>>>124

500 300

22
m/z = 492>>>331

11
m/z = 255>>>132

300 200

23
m/z = 325>>>108

3000 1000

12
m/z = 318>>>136
10 11 12 13 14 15 16 17 Retention time (min) 18 19 20

2200 1000

24
m/z = 368>>199
15 16 17 18 19 20 21 Retention time (min) 22 23 24

Figure 4.

MRM of 33 pesticides in orange extract spiked at 10 ng/mL. (Continued)

400 200

25
m/z = 439>>>173

1000 400

31
m/z = 383>>>195

200 80

26
m/z = 431>>>105

120 60

32
m/z = 479>>>344

180 100

27
m/z = 413>>>241
15 16 17 18 19 20 21 22 23 24

300 100

33
m/z = 330>>>284
21 22 23 24 25 26 Retention time (min) 27 28 29

800 400

28
m/z = 373>>>299

45 25

29
m/z = 528>>>150

3500 1500

30
m/z = 336>>>238
21 22 23 24 25 26 Retention time (min) 27 28 29

Figure 4.

MRM of 33 pesticides in orange extract spiked at 10 ng/mL.

6000 2000

1
m/z = 326>>>129

6000 4000

7
m/z = 430>>>398

7500 2500

2
m/z = 221>>>102

8000 4000

8
m/z = 406>>>161

7500 2500

3
m/z = 312>>>267

800 400

9
m/z = 456>>>344

20000 10000

4
m/z = 388>>>167

20000 10000

10
m/z = 493>>>264

11
7500 2500

5
m/z = 360>>>129

2000 1000

m/z = 362>>>316
16 17 18 19 Retention time (min) 20

6
10000 5000 1

m/z = 248>>>129
2 3 4 5 6 7 8 9 10 11 12 13 14 15

Retention time (min)

Figure 5. 8

MRM of 11 pesticides in orange extract spiked at 10 ng/mL.

7, 8 5000

12 9 13 15,16 10 11 14 17 22 24 29, 30 18 19, 20, 21 23 25 28 26 27 31 32 33

(A)
5 1 2 6 34

3000

1000

5000

(B)

3000

1000

5000

(C)

3000

1000 1 3 5 7 9 11 13 15 17 Retention time (min) 6 4 1 8 2 3 5000 7 9 11 19 21 23 25 27 29

15000 10000

(D)

10

15000 10000 5000

(E)

15000

(F)

10000

5000

9 11 13 Retention time (min)

15

17

19

Figure 6.

TIC of spiked at 10 ng/mL.

Table 4.

Relative Intensity of Each Pesticide in Sample Extracts Relative intensity(%) Orange* Cabbage 105 (3.2) 101 103 (2.1) 98 106 (2.9) 101 105 (3.3) 102 (1.7) 103 (4.6) 106 (3.7) 104 (3.1) 93 (4.6) 116 (4.1) 96 (5.3) 90 (2.1) 104 (4.4) 102 (2.7) 97 (4.1) 92 (3.1) 96 (2.8) 97 (3.4) 99 (2.1) 91 (4.3) 102 (2.6) 93 (3.5) 102 (2.7) 103 (4.7) 108 (5.2) 108 (3.4) 109 (2.6) 105 (4.2) 105 (4.1) 102 (1.8) 100 (3.7) 101 (3.3) 97 (2.6) 104 (4.8) 105 (3.1) 106 (2.9) 99 (3.1) 101 (4.4) 94 (3.9) 95 (3.3) 99 (5.9) 97 (4.1) 108 (4.8) 106 97 107 102 104 90 109 99 103 102 104 103 99 102 96 105 97 114 92 105 101 111 110 105 107 104 104 109 111 110 101 100 112 106 103 104 102 101 111 114 Apple 116 104 109 101 102 103 104 106 94 102 100 104 106 108 104 104 103 100 102 98 104 87 103 103 98 105 100 106 104 105 105 102 156 102 100 116 103 100 97 96 95 104 110 Potato 107 105 112 109 104 108 106 108 84 112 104 110 110 103 93 97 101 111 101 103 114 95 107 97 108 101 111 104 105 101 112 117 104 113 101 113 109 108 142 107 109 108 124

No Pesticides Group 1 1 Thiabendazole 2 Thiamethoxam 3 Clothianidin 4 5 6 7 8 9 Chloridazon Imidacloprid Dimethirimol Oxycarboxine Thiacloprid Azamethiophos

10, 11 Ferimzone(E,Z) 12 Phenmedipham 13 Azinphos-methyl 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 28 30 31 32 33 Simeconazole Isoxaflutol Pyriftalid Methoxyfenozide Chromafenozide Tridemorph Fenoxycarb Naproanilide Butafenacil Cyazofamide Anilofos Pyrazolate Benzofenap Cyflufenamid Indoxacarb Clomeprop Cuinclorac-methyl Furathiocarb Lactofen Tralkoxydim

Group 2 1 Flumetsulam 2 Thidiazuron 3 Imazaquin 4 5 6 7 8 9 10 11 Thifensulfuron-methyl Florasulam Forchlorfenuron-methyl Clorasulam-methyl Diclosulam Fomesafen Triflusulfuron-methyl Haloxyfop

*( ): RSD,% calculated based on five replicates within one day

10

Conclusions
The multiresidue method by LC/MS/MS described here was suitable for the determination of 44 pesticides in a variety of food samples due to its high sensitivity and high selectivity. Another advantage of this method is that ion suppression was not observed for all food samples studied. Thus, it may eliminate the need for matrix-matched standards, which make analysis more tedious for samples from different origins. For more details concerning this application, please contact masahiko_takino@agilent.com

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

11

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA August 8, 2006 5989-5459EN

Screening for 926 Pesticides and Endocrine Disruptors by GC/MS with Deconvolution Reporting Software and a New Pesticide Library Application Note

Food and Environmental

Authors
Philip L. Wylie Agilent Technologies, Inc. 2850 Centerville Rd. Wilmington, DE 19808-1610 USA

endocrine disrupters in about two minutes per sample. Deconvolution helps identify pesticides that are buried in the chromatogram by co-extracted materials. The new database was compared to the smaller one for the DRS analysis of 17 surface water samples. With the new database, DRS found 99 pesticides, metabolites, fire retardants, and related contaminants that were not contained in the original RTL Pesticide and Endocrine Disruptor Library.

Abstract
An updated and greatly expanded collection of mass spectral libraries has been introduced, replacing Agilents RTL Pesticide Library and DRS pesticide solution. The new library contains 926 pesticides, endocrine disruptors, and related compounds 359 more than the original library. Included are all compounds specified for GC/MS analysis in the new Japanese Positive List regulations. All compounds have locked retention times that can be accurately reproduced using an Agilent GC/MS system with the ChemStation's Retention Time Locking software. The new Database can be used as a standard GC/MS library for compound identification or with Agilent's Screener software for identifications based upon retention time and mass spectral matching. The greatest benefit accrues when these libraries are used with Agilents new version of Deconvolution Reporting Software (part number G1716AA version A.03.00). This solution allows one to screen GC/MS files for all 926 pesticides and

Introduction
Several years ago Agilent Technologies introduced Retention Time Locking (RTL) for gas chromatography (GC) and GC with mass spectral detection (GC/MS). RTL software makes it possible to reproduce retention times from run-to-run on any Agilent GC or GC/MS, in any laboratory in the world, so long as the same nominal method and GC column are used (1). Since any laboratory can reproduce retention times generated in another, it is possible to create mass spectral libraries that contain locked retention times. By locking their method to the published database, users can screen GC/MS files for all of the librarys compounds. Hits are required to have the correct retention time as well as the correct spectrum, which eliminates many false positives and gives more confidence in compound identifications (2).

More recently, Agilent introduced Deconvolution Reporting Software (DRS) that incorporates mass spectral deconvolution with conventional library searching and quantification. DRS results from a marriage of three different GC/MS software packages: 1) The Agilent GC/MS ChemStation, 2) The National Institute of Standards and Technology (NIST) Mass Spectral Search Program with the NIST 05 MS Library, and 3) The Automated Mass Spectral Deconvolution and Identification System (AMDIS) software, also from NIST. The original DRS software was intended to be a comprehensive solution for pesticide analysis and, therefore, included the mass spectra (in AMDIS format) and locked retention times for 567 pesticides and suspected endocrine disrupters (3). Recently, Agilent introduced an updated and greatly expanded Pesticide and Endocrine Disruptor Database (part number G1672AA) that now contains 926 entries. This represents the addition of 359 new compounds to the original library. At the same time, Agilent introduced a new version of the DRS software (part number G1716AA version A.03.00) that can be used with any Agilent-provided or user-developed DRS library. Pesticide and Endocrine Disruptor Database Contents The G1672AA Pesticide and Endocrine Disruptor Database contains virtually all GC-able pesticides, including those introduced very recently. In addition, the database includes numerous metabolites, more endocrine disruptors, important PCBs and PAHs, certain dyes (for example, Sudan Red), synthetic musk compounds, and several organophosphorus fire retardants. This new database includes: A conventional mass spectral library for use with Agilent GC/MS ChemStations

A screener database for use with Agilents powerful screener software that is integrated into the GC/MS ChemStation Locked Retention Times for all 926 compounds that any Agilent 5975 or 5973 GC/MS user can reproduce in their laboratory Files for use with Agilents G1716AA (A.03.00) Deconvolution Reporting Software An e-method that can be loaded into Agilents G1701DA (version D.02.00 SP1 or higher) with instrument parameters for acquiring GC/MS files and analyzing the data with DRS. These parameters are listed in Table 1. Example files Application notes On November 29, 2005, the Japanese Government published a Positive List system for the regulation of pesticides, feed additives, and veterinary drugs. Maximum Residue Limits (MRL) have been set for 758 chemicals while 65 others have been exempted from regulation. Fifteen substances must have no detectable residues. Other agricultural chemicals not mentioned have a uniform MRL of 0.01 ppm (4). This new regulation is scheduled to take effect on May 29, 2006. Of the pesticides in the Japanese Positive List, 265 are to be analyzed by GC/MS. The new G1672AA Pesticide library contains mass spectra and locked retention times for all of these compounds. Thus, a laboratory could screen for all 265 positive list compounds and several hundred more pesticides in just 13 minutes after the GC/MS run.

Experimental
Table 1 lists the instrumentation, software, and analytical parameters used by Agilent for pesticide analysis. Depending upon the desired injection volume, a PTV inlet or split/splitless inlet can be used.

Table 1.

Instrumentation and Conditions of Analysis Agilent 6890N Agilent 7683 Injector and AutoSampler Agilent PTV operated in the solvent vent mode or Split/Splitless Agilent 30 m 0.25 mm 0.25 m HP-5MSi (part number 19091S-433i) Helium in the constant pressure mode Chlorpyrifos-methyl locked to 16.596 min (nominal column head pressure = 17.1 psi) 70 C (2 min), 25 C/min to 150 C (0 min), 3 C /min to 200 C (0 min), 8 C /min to 280 C (1015 min) Temp program: 40 C (0.25 min), 1600 C/min to 250 C (2 min); Vent time: 0.2 min; Vent flow: 200 mL/min; Vent pressure: 0.0 psi; Purge flow: 60.0 mL/min; Purge time: 2.00 min 15 L (using a 50-L syringe) Agilent 5975 inert Atune.u Scan (or SIM with SIM DRS library) 50550 u 230, 150, and 280 C, respectively 4.00 min Autotune voltage

Gas Chromatograph Automatic Sampler Inlet Column Carrier gas Retention time locking Oven temperature program PTV inlet parameters Injection volume Mass Selective Detector Tune file Mode Scan range Source, quad, transfer line temperatures Solvent delay Multiplier voltage Software GC/MSD ChemStation Deconvolution Reporting Software Library searching software Deconvolution software MS Libraries

Agilent part number G1701DA (version D02.00 sp1 or higher) Agilent part number G1716AA (version A.03.00) Deconvolution Reporting Software NIST MS Search (version 2.0d or greater) (comes with NIST '05 mass spectral library Agilent part number G1033A) Automated Mass Spectral Deconvolution and Identification Software (AMDIS_32 version 2.62 or greater; comes with NIST '05 mass spectral library Agilent part number G1033A) NIST 05 mass spectral library (Agilent part number G1033A) Agilent RTL Pesticide and Endocrine Disruptor Libraries in Agilent and NIST formats (part number G1672AA)

Results and Discussion


DRS, which has been described in preceding papers (3,5,6), can be summarized as follows: Three separate, but complimentary, data analysis steps are combined into the DRS. First, the GC/MS ChemStation software performs a normal quantitative analysis for target pesticides using a target ion and up to three qualifiers. An amount is reported for all calibrated compounds that are detected. For other compounds in the database, an estimate of their concentration can be reported based upon an average pesticide response factor

that is supplied with the DRS software. The DRS then sends the data file to AMDIS, which deconvolutes the spectra and searches the Agilent RTL Pesticide Library using the deconvoluted full spectra. A filter can be set in AMDIS, which requires the analytes retention time to fall within a userspecified time window. Because RTL is used to reproduce the RTL database retention times with high precision, this window can be quite small typically 1020 seconds. Finally, the deconvoluted spectra for all of the targets found by AMDIS are searched against the 147,000-compound NIST mass spectral library for confirmation; for this step, there is no retention time requirement.

This approach was rapidly adopted by many laboratories because of its ability to identify pesticides in complex chromatograms containing high levels of co-extracted interferences. Indeed, the solution proved to be so useful that users began to create their own DRS libraries (7). Therefore, the DRS was unbundled from the pesticide database so that it could be used with any agilent-provided or user-created database. The original 567-compound RTL Pesticide Library (G1049A) included pesticides, a few metabolites, and most of the GC-amenable endocrine disruptors that were known at the time. The new version of the library includes many more pesticides, endocrine disruptors, and metabolites. This update also contains important compounds from other classes of contaminants that have been found in food and water supplies. Included are eighteen polychlorinated biphenyls (PCBs), four polybrominated biphenyls (PBBs), several polynuclear aromatic hydrocarbons (PAHs), several organophosphorus fire retardants, three important toxaphene congeners, and three Sudan dyes.

Advantages of Deconvolution Figure 1 shows a screen from AMDIS that illustrates the power of this deconvolution software. The white trace in Figure 1A is the total ion chromatogram while the other three are extracted ions of a deconvoluted peak (a component in AMDIS terminology). Note that the TIC and extracted ions are not scaled to each other and this component is actually obscured by co-eluting compounds. Figure 1B juxtaposes the deconvoluted component spectrum (white) with the complete undeconvoluted spectrum (black). Clearly, this component is buried under co-eluting peaks that would ordinarily obscure the analyte. Figure 1C shows that the deconvoluted peak (white spectrum) is a good library match for norflurazon (black spectrum). The locked retention time for norflurazon in the RTL Pesticide Database is 26.933 min, which is just 2.3 seconds away from its observed RT in this chromatogram. Confidence in peak identifications is greatly enhanced by the combination of spectral deconvolution and locked retention time filtering.

Figure 1.

AMDIS screen showing the identification of norflurazon. A) The total ion and extracted ion chromatograms where norflurazon elutes. B) The deconvoluted component spectrum (white) juxtaposed with the spectrum at 26.972 min (black). C) The deconvoluted component matched to the library spectrum of norflurazon.

Surface Water Analysis - Revisiting an Earlier Study In an earlier study, a comparison was made between Agilents DRS and conventional pesticide analysis (3). The California Department of Food and Agriculture (CDFA) provided data files for 17 surface water extracts that had been analyzed in their laboratory. Since the GC/MS chromatograms were locked to the Agilent pesticide method, it was possible to analyze these data files using DRS without having to re-run the samples. The original DRS analysis was made using the 567-compound RTL Pesticide Database. For comparison, these same data files were re-analyzed using the new 926-compound RTL Pesticide Database. The chromatogram (Figure 2) and the DRS report (Figure 3) from one of these samples are shown below.

Excluding phthalates, seven new compounds (shown with bold type in Figure 3) were identified using the 926-compound database: 4-chlorophenyl isocyanate (a phenylurea herbicide metabolite); 3,4-dichlorophenyl isocyanate (diuron metabolite); tris(2-chloroethyl) phosphate (a fire retardant); caffeine (a stimulant); Cyprodinil (a fungicide); desmethyl-norflurazon (a metabolite of norflurazon, an herbicide); and tris(2-butoxyethyl) phosphate (a fire retardant). Although caffeine is not generally considered to be dangerous, it is included in the database because it has been found frequently in sewage effluent and in numerous waterways together with a various pharmaceuticals and pesticides (8).

2800000 2600000 2400000 2200000 2000000 Abundance 1800000 1600000 1400000 1200000 1000000 800000 600000 400000 200000 5.00 10.00 15.00 20.00 Time 25.00 30.00 35.00 40.00 TIC: E02-557.d\data.ms

Figure 2.

Chromatogram of a surface water extract that was analyzed by DRS using the new RTL Pesticide and Endocrine Disrupter Database. The results of this analysis are shown in Figure 3.

MSD Deconvolution Report Sample Name: E02-557 Data File: C:\MSDChem\1\DATA\CDFA surface water data\E02-557.d Date/Time: 11:24 AM Tuesday, Apr 4 2006 The NIST library was searched for the components that were found in the AMDIS target library. Agilent ChemStation amount (ng) AMDIS match 62 84 99 84 93 67 63 RT Diff (sec.) 3.2 1.8 3.1 2.0 2.1 1.7 7.7 62 1.29 85 98 86 96 83 88 79 95 80 90 97 90 99 90 69 2.2 2.6 2.6 3.0 0.7 1.4 1.0 1.3 1.6 3.2 0.4 1.5 0.4 0.7 0.1 70 87 87 94 89 75 98 65 4.5 1.5 0.5 0.8 3.3 0.3 1.9 71 10 1 69 79 94 83 83 90 1 2 1 1 1 1 3 84 92 88 90 74 86 78 83 74 88 90 84 94 87 2 1 2 1 1 2 1 1 1 4 1 1 1 1 1 NIST reverse match 48 86 95 85 89 84 Hit number 1 2 1 1 2 2

RT 4.4689 4.4689 4.8840 6.3879 6.8357 7.6988 7.9342 8.1112 8.1112 8.941 9.7903 10.0019 10.7109 10.9684 11.6491 12.9326 13.4309 13.7478 15.4048 15.9474 16.5988 17.3653 18.4213 18.9214 20.5633 20.5633 26.4247 26.9700 26.9992 27.3984 28.0127 29.6537 33.9298 33.9298 13.739 Figure 3.

Cas # 106445 0000 104121 102363 759944 95761 131113 25013165 0000 29878317 134623 84662 119619 126738 1582098 122349 115968 1517222 58082 84695 5598130 7287196 84742 51218452 121552612 76470252 23576241 27314132 85687 51235042 78513 117817 84764 0000

Compound name 4-Methylphenol 3-Carbobenzyloxy-4-ketoproline 4-Chlorophenyl isocyanate Diuron Metabolite [3,4-Dichlorophenyl isocyanate] EPTC 3,4-Dichloroaniline Dimethylphthalate Butylated hydroxyanisole 7-Methoxy-2,2,4,8-tetramethyltricyclo [5.3.1.0(4,11)]undecane Tolyltriazole [1H-Benzotriazole, 4-meth-] N,N-Diethyl-m-toluamide Diethyl phthalate Benzophenone Tributyl phosphate Trifluralin Simazine Tris(2-chloroethyl) phosphate Phenanthrene-d10 Caffeine Diisobutyl phthalate Chlorpyrifos Methyl Prometryn Di-n-butylphthalate Metolachlor Cyprodinil 9,9-Dimethoxy-9-sila-9, 10-dihydroanthracene Norflurazon, DesmethylNorflurazon Butyl benzyl phthalate Hexazinone Tris(2-butoxyethyl) phosphate Bis(2-ethylhexyl)phthalate Di-n-nonyl phthalate Phthalic acid, 3,4-dichlorophenyl propyl ester Phenanthrene-d10

DRS report from the analysis of a surface water sample. The compounds shown in bold type were found by the new RTL Pesticide Database but not the original one because these compounds were not included.

For this sample, the ChemStation identified only tolyltriazole at 8.941 min, but AMDIS did not confirm this assignment, nor could it be confirmed manually. Butylated hydroxyanisole was tentatively identified by AMDIS with a low match value, but the retention time is off by 7.7 seconds which is considerably more than most other hits. This compound is not in the NIST library so it could not be confirmed. The ChemStation method used for this analysis required that all three qualifier ions fall within 20% (relative) which is a rigorous requirement for such a complex sample. This explains why so few compounds were found by the ChemStation. Cyprodinil (20.563 min) was identified by AMDIS but the NIST library search failed to confirm its presence. The next line shows that the best NIST library match is an anthracene derivative that is nothing like cyprodinil. This result was obtained when AMDIS was configured to use uncertain peaks as shown in Figure 4. When this feature is

turned off in DRS Compound Identification Configuration, the best NIST library hit for this spectrum is, indeed, cyprodinil. When a compound's identity is ambiguous, as with cyprodinil, it may be useful to perform the DRS search both ways and compare the results. In the comparison described earlier (3), DRS was able to identify all 37 pesticides found by the CDFA chemist. However, DRS completed the task for all 17 samples in about 20 minutes compared to ~8 hours for the manual procedure (Table 2). Moreover, DRS identified one false positive in the CDFA report and found 34 additional pesticides and related compounds. Using the new 926-compound Database, it took 32 minutes to analyze all of the samples and DRS was able to find an additional 99 pesticides, metabolites, fire retardants, and related compounds (Table 2).

Figure 4.

DRS configuration screen for the method called Tri_Pest. When the box labeled Use Uncertain Peaks is checked, AMDIS will use uncertain peaks for library searches. When unchecked, AMDIS ignores uncertain mass spectral peaks. Sometimes, this can affect the quality of a library match.

Table 2.

Comparison of the Results Obtained by Screening 17 Surface Water Extracts Using Traditional Methods (CDFA) and Using DRS With Two Different Databases the G1049A With 567 Compounds and the G1672AA With 926 Entries Agilent DRS (Original G1049A database) Same 37 +34 more 0 20 minutes Agilent DRS (G1672 AA database) Same 37 +99 more 0 32 min

CDFA Targets found (not counting ISTD) False positives Processing time 37 1 ~8 hrs (ChemStation only)

Handling Stereoisomers Many pesticides have multiple stereoisomers with virtually identical mass spectra. For example, cyfluthrin has four diastereomers arising from its three chiral centers. It is very difficult and sometimes impossible to determine the elution order of these isomers and most analysts report them as a sum of the isomer amounts. Agilents G1049A RTL Pesticide database arbitrarily assigned each isomer a Roman numeral with I for the earliest eluting isomer, II for the next, and so on. The same Chemical Abstracts Service number (CAS #) was assigned to all of the isomers. Generally, it was a CAS # for the compound with unstated stereochemistry. This caused some incompatibility with AMDIS as explained below. AMDIS software differentiates among compounds using a chemical identification number. The easiest and most consistent approach is to use each compound's CAS #. The default setting for AMDIS is to allow each CAS # to be used only once when analyzing a GC/MS data file. While this seems logical, it requires that each database entry have a different CAS #. It is possible to allow multiple hits per compound by checking the box in AMDIS found in the drop down menu under Analyze/ Settings/Identif. However, this allows multiple peaks to be assigned the same compound name.

In the new RTL Pesticide Database (G1672AA), the Roman numeral designations remain and the first isomer in the series is given its genuine CAS #. Subsequent isomers in the series are given unique, but fictitious CAS #s generated by Agilent. The compound's real CAS # appears in braces after the compound name. For example, the cyfluthrin isomers are entered into the database as shown in Table 3.

Table 3.

RT 32.218 32.359 32.477 32.536

Method for Listing Compounds with Multiple Stereoisomers in the New G1672AA RTL Pesticide Database Compound name* CAS #** Cyfluthrin I Cyfluthrin II {CAS # 68359-37-5} Cyfluthrin III {CAS # 68359-37-5} Cyfluthrin IV {CAS # 68359-37-5} 68359-37-5 999028-03-4 999029-03-7 999030-03-4

* In a series, the earliest eluting isomer is identified with I and is assigned its legitimate CAS #. Subsequent isomers are assigned unique, but fictitious CAS #s (see footnote **). Their actual CAS # is put in braces behind the compound name. **Cyfluthrin I has been given it's genuine CAS #. Cyfluthrin II-IV have been given unique numbers that can be distinguished from actual CAS numbers because they all have six digits before the first hyphen (9 total) and all begin with the series 999.

Figure 5 shows how permethrin was identified in a spinach sample using both databases with AMDIS configured to allow one hit per compound. Using the older 567-compound database (G1049A) only one permethrin isomer was identified because its CAS # could be used only once. With the new format used in the 926-compound RTL Pesticide Database (G1672AA), both isomers of permethrin were identified. Not surprisingly, the NIST library search found no hits with the same fictitious CAS # assigned to permethrin II. So, the software printed the best match on the following line. This compound, a cyclopropanecarboxylic acid derivative, is a permethrin isomer. So long as the NIST library search is turned on in DRS, it will always print another line after reporting a compound with a fictitious CAS #. Note that these fictitious CAS #s always contain 9 digits and begin with 999. A)
Agilent ChemStation amount (ng) AMDIS match 88 RT Diff (sec.) 3.9 NIST reverse match 91 Hit number 3

RT 31.6158

Cas # 52645531

Compound name Permethrin II

B)
Agilent ChemStation amount (ng) AMDIS match 78 65 RT Diff (sec.) 2.6 3.5 95 1 NIST reverse match 81 Hit number 3

RT 31.4127 31.6088 31.6088

Cas # 52645531 999046036 51877748

Compound name Permethrin I Permethrin II {CAS # 52645-53-1} Cyclopropanecarboxylic acid, 3-(2,2-dichlorovinyl)-2,2-dimethyl-, (3-phenoxyphenyl)methyl ester, (1R-trans)-

Figure 5.

A) A single isomer of permethrin was identified by DRS using the G1049A 567-compound database when AMDIS was not allowed to use multiple hits per compound. B) Two permethrin isomers are identified by DRS with the G1672AA 926-compound database under the same circumstances.

Conclusions
The new G1672AA RTL Pesticide and Endocrine Disruptor library contains substantially more target analytes than its predecessor. With the addition of 359 new compounds, it is the most comprehensive library of its type available today. Many new pesticides, metabolites, and endocrine disruptors were added along with important PCBs, PBBs, PAHs, synthetic musk compounds, Sudan dyes, and organophosphorus fire retardants. The database contains all of the analytes specified for GC/MS analysis in the new Japanese Positive List regulations. When combined with the complete DRS solution, one can screen GC/MS data files for all 926 compounds in about two minutes per sample. This is the fastest, most comprehensive, most accurate, and least tedious method for screening food and environmental samples for these compounds.

5. C. P. Sandy, A Blind Study of Pesticide Residues in Spiked and Unspiked Fruit Extracts using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1564EN, www.agilent.com/chem 6. C. Lesueur and M. Gartner, Routine Identification and Quantification of Pesticide Multiesidues in Fruit and Vegetable Samples with Full Scan, SIM, and Deconvolution Reporting Software, 2005 Ernhrung/Nutrition, 29 (11) 466471 7. X. Ping, C.-K. Meng, and M. Szelewski, Building Agilent GC/MSD Deconvolution Reporting Libraries for any Application, Agilent Technologies, publication 5989-2249EN, www.agilent.com/chem 8. Large-scale studies of the occurrence and distribution of new contaminants in the environment Reconnaissance studies, USGS, Contaminant Occurrence Studies, http://toxics.usgs.gov/topics/reconnaissance_ studies.html

References
1. V. Giarocco, B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E, www.agilent.com/chem 2. H. Prest, P. Wylie, K. Weiner, and D. Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the 6890/5973 GC/MSD System, Agilent Technologies, publication 5968-4884E, www.agilent.com/chem 3. P. L. Wylie, M. J. Szelewski, C.-K. Meng, C. P. Sandy, Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN, www.agilent.com/chem 4. Introduction of the Positive List System for Agricultural Chemical Residues in Foods Department of Food Safety, Ministry of Health, Labour and Welfare http://www.mhlw.go.jp/english/topics/foodsafety/positivelist060228/introduction.html

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Acknowledgments
The author wishes to thank Dr. G. Kempe of the Landesuntersuchungsanstalt Sachsen, Institut, Chemnitz, Germany for his help in acquiring much of the data for this library update. The author also thanks Dr. Mark Lee and Mr. Steve Siegel of the California Department of Food and Agriculture for providing the surface water extract data files.

10

Appendix A

Lists of Compounds in Databases


1,2,4-Trichlorobenzene 1,2-Dibromo-3-chloropropane 1,3,5-Tribromobenzene 1,3-Dichlorbenzene 17a-Ethynylestradiol 1-naphthalenol 2-(1-naphthyl)acetamide 2-(2-Butoxyethoxy)ethyl thiocyanate 2-(Octylthio)ethanol 2,3,4,5-Tertrachloronitrobenzene 2,3,4,5-Tetrachlorophenol 2,3,4,6-Tetrachlorophenol 2,3,5,6-Tetrachlorophenol 2,3,5,6-Tetrachloro-p-terphenyl 2,3,5-Trichlorophenol 2,3,5-Trimethacarb 2,3,6-Trichloroanisole 2,3,7,8-Tetrachlorodibenzofuran 2,3,7,8-Tetrachlorodibenzo-p-dioxin 2,4,5,6-Tetrachloro-m-xylene 2,4,5-T methyl ester 2,4,5-Trichloroaniline 2,4,5-Trichlorophenol 2,4,5-Trichloro-p-terphenyl 2,4,5-Trimethylaniline 2,4,6-Tribromoanisole 2,4,6-Tribromophenol 2,4,6-Trichloroanisole 2,4,6-Trichlorophenol 2,4-D methyl ester 2,4-D sec-butyl ester 2,4-DB methyl ester 2,4'-Dichlorobenzophenone (2,4'-Dicofol decomposition product) 2,4-Dichlorophenol 2,4-Dichlorophenyl benzenesulfonate 2,4-Dimethylaniline 2,4-Dimethylphenol 2,6-Dichlorobenzamide 2,6-Dichlorobenzonitrile 2,6-Dimethylaniline 2-[3-Chlorophenoxy]propionamide 2-Chlorophenol 2-Ethyl-1,3-hexanediol 2-ethyl-6-methylaniline 2-Hydroxyestradiol 2-Methyl-4,6-dinitrophenol 2-Methylphenol 2-Nitrophenol 2-Phenoxypropionic acid 3,4,5-Trimethacarb 3,4-Dichloroaniline 3,5-Dichloroaniline 3-Aminophenol 3-Chloro-4-fluoroaniline 3-Chloro-4-methoxyaniline 3-Chloroaniline 3-Hydroxycarbofuran 3-Indolylacetonitrile 3-Trifluormethylaniline 4,4'-Dichlorobenzophenone 4,4'-Oxydianiline 4,6-Dinitro-o-cresol (DNOC) 4-Aminodiphenyl 4-Bromoaniline 4-Chloro-2-methylaniline 4-Chloro-3-methylphenol 4-Chloroaniline 4-Chlorophenyl isocyanate 4-Isopropylaniline 4-Methylphenol 4-Nitrophenol 4-Nonylphenol 5,7-Dihydroxy-4'-methoxyisoflavone 9,10-Anthraquinone Acenaphthene Acenaphthylene Acephate Acequinocyl acetamiprid Acetochlor Acifluorfen methyl ester Aclonifen Acrinathrin Alachlor Aldrin Allidochlor Ametryn Amidithion Aminocarb Amitraz Amitraz metabolite [Methanimidamide, N(2,4-dimethylphenyl)-N'-methyl-] Ancymidol Anilazine Aniline Anilofos Anthracene Aramite I Aramite II {CAS # 140-57-8} Atraton Atrazine Atrazine-desethyl Azaconazole Azamethiphos Azibenzolar-S-methyl Azinphos-ethyl Azinphos-methyl Aziprotryn metabolite [2-Amino4-isopropylamino-6-methylthio1,3,5-triazine] Aziprotryne Azobenzene Azoxybenzene Azoxystrobin Barban Beflubutamid Benalaxyl Benazolin-ethyl Bendiocarb Benfluralin 11

Benfuracarb Benfuresate Benodanil Benoxacor Bentazone Bentazone methyl derivative Benthiocarb Benzene, 1,3-bis(bromomethyl)Benzenesulfonamide Benzidine Benzo(a)anthracene Benzo(a)pyrene Benzo[b]fluoranthene Benzo[g,h,i]perylene Benzo[k]fluoranthene Benzophenone Benzoximate metabolite Benzoylprop ethyl Benzyl benzoate b-Estradiol BHC alpha isomer BHC beta isomer BHC delta isomer BHC epsilon isomer Bifenazate metabolite (5-Phenyl-o-anisidine) Bifenox Bifenthrin Binapacryl Bioallethrin Bioallethrin S-cyclopentenyl isomer Bioresmethrin Biphenyl Bis(2,3,3,3-tetrachloropropyl) ether Bis(2-butoxyethyl) phthalate Bis(2-ethylhexyl)phthalate Bisphenol A Bitertanol I Bitertanol II {CAS # 55179-31-2} Boscalid (Nicobifen) Bromacil Bromfenvinphos-(E) Bromfenvinphos-(Z) Bromobutide Bromocyclen Bromophos 12

Bromophos-ethyl Bromopropylate Bromoxynil Bromoxynil octanoic acid ester Bromuconazole I Bromuconazole II {CAS # 116255-48-2} Bufencarb Bupirimate Buprofezin Butachlor Butafenacil Butamifos Butoxycarboxim Butralin Butyl benzyl phthalate Butylate Butylated hydroxyanisole Cadusafos Cafenstrole Caffeine Captafol Captan Carbaryl Carbetamide Carbofuran Carbofuran-3-keto Carbofuran-7-phenol Carbophenothion Carbosulfan Carboxin Carfentrazone-ethyl Carpropamid Carvone Cashmeran Cekafix Celestolide Chinomethionat Chloramben methyl ester Chloranocryl Chlorbenside Chlorbenside sulfone Chlorbicyclen Chlorbromuron Chlorbufam Chlordecone Chlordene, trans-

Chlordimeform Chlorethoxyfos Chlorfenapyr Chlorfenethol Chlorfenprop-methyl Chlorfenson Chlorfenvinphos Chlorfenvinphos, cisChlorfenvinphos, transChlorflurecol-methyl ester Chlormefos Chlornitrofen Chlorobenzilate Chloroneb Chloropropylate Chlorothalonil Chlorotoluron Chlorpropham Chlorpyrifos Chlorpyrifos Methyl Chlorthal-dimethyl Chlorthiamid Chlorthion Chlorthiophos Chlorthiophos sulfone Chlorthiophos sulfoxide Chlozolinate Chrysene Cinerin I Cinerin II Cinidon-ethyl cis-Chlordane Clodinafop-propargyl Clomazone Cloquintocet-mexyl Coumaphos Crimidine Crotoxyphos Crufomate Cyanazine Cyanofenphos Cyanophos Cyclafuramid Cycloate Cyclopentadecanone Cycluron

Cyflufenamid Cyfluthrin I Cyfluthrin II {CAS # 68359-37-5} Cyfluthrin III {CAS # 68359-37-5} Cyfluthrin IV {CAS # 68359-37-5} Cyhalofop-butyl Cyhalothrin I (lambda) Cyhalothrin (Gamma) Cymiazole Cymoxanil Cypermethrin I Cypermethrin II {CAS # 52315-07-8} Cypermethrin III {CAS # 52315-07-8} Cypermethrin IV {CAS # 52315-07-8} Cyphenothrin cisCyphenothrin trans- {CAS # 39515-40-7} Cyprazine Cyproconazole Cyprodinil Cyprofuram Cyromazine d-(cis-trans)-Phenothrin-I d-(cis-trans)-Phenothrin-II {CAS # 260002-80-2} Dazomet DDMU [1-Chloro-2,2-bis(4'-chlorophenyl)] Decachlorobiphenyl Deltamethrin Demephion Demeton-S Demeton-S-methylsulfon Desbromo-bromobutide Desmedipham Desmetryn Dialifos Di-allate I Di-allate II {CAS # 2303-16-4} Diamyl phthalate Diazinon Diazinon-oxon Dibenz[a,h]anthracene Dicamba Dicamba methyl ester Dicapthon Dichlofenthion Dichlofluanid

Dichlofluanid metabolite (DMSA) Dichlone Dichlormid Dichlorophen Dichlorprop Dichlorprop methyl ester Dichlorvos Diclobutrazol Diclocymet I Diclocymet II {CAS # 139920-32-4} Diclofop methyl Dicloran Dicrotophos Dicyclohexyl phthalate Dicyclopentadiene Dieldrin Diethatyl ethyl Diethofencarb Diethyl dithiobis(thionoformate) (EXD) Diethyl phthalate Diethylene glycol Diethylstilbestrol Difenoconazol I Difenoconazol II {CAS # 119446-68-3} Difenoxuron Diflufenican Diisobutyl phthalate Dimefox Dimepiperate Dimethachlor Dimethametryn Dimethenamid Dimethipin Dimethoate Dimethomorph-(E) Dimethomorph-(Z) {CAS # 110488-70-5} Dimethylphthalate Dimethylvinphos(z) Dimetilan Dimoxystrobin Di-n-butylphthalate Di-n-hexyl phthalate Diniconazole Dinitramine Di-n-nonyl phthalate Dinobuton

Dinocap I Dinocap II {CAS # 39300-45-3} Dinocap III {CAS # 39300-45-3} Dinocap IV {CAS # 39300-45-3} Di-n-octyl phthalate Dinoseb Dinoseb acetate Dinoseb methyl ether Dinoterb Dinoterb acetate Di-n-propyl phthalate Diofenolan I Diofenolan II {CAS # 63837-33-2} Dioxabenzofos Dioxacarb Dioxathion Diphacinone Diphenamid Diphenyl phthalate Diphenylamine Dipropetryn Dipropyl isocinchomeronate Disulfoton Disulfoton sulfone Ditalimfos Dithiopyr Diuron Diuron Metabolite [3,4-Dichlorophenyl isocyanate] Dodemorph I Dodemorph II {CAS # 1593-77-7} Drazoxolon Edifenphos Empenthrin I Empenthrin II {CAS # 54406-48-3} Empenthrin III {CAS # 54406-48-3} Empenthrin IV {CAS # 54406-48-3} Empenthrin V {CAS # 54406-48-3} Endosulfan (alpha isomer) Endosulfan (beta isomer) Endosulfan ether Endosulfan lactone Endosulfan sulfate Endrin Endrin aldehyde Endrin ketone 13

EPN Epoxiconazole EPTC Erbon Esfenvalerate Esprocarb Etaconazole Ethalfluralin Ethidimuron Ethiofencarb Ethiolate Ethion Ethofenprox Ethofumesate Ethofumesate, 2-Keto Ethoprophos Ethoxyfen-ethyl Ethoxyquin Ethylenethiourea Etoxazole Etridiazole Etridiazole, deschloro- (5-ethoxy3-dichloromethyl-1,2,4-thiadiazole) Etrimfos Eugenol Exaltolide [15-Pentadecanolide] Famoxadon Famphur Fenamidone Fenamiphos sulfoxide Fenamiphos-sulfone Fenarimol Fenazaflor Fenazaflor metabolite Fenazaquin Fenbuconazole Fenchlorazole-ethyl Fenchlorphos Fenchlorphos-oxon Fenclorim Fenfuram Fenhexamid Fenitrothion Fenitrothion-oxon Fenobucarb Fenoprop 14

Fenoprop methyl ester Fenothiocarb Fenoxanil Fenoxaprop-ethyl Fenoxycarb Fenpiclonil Fenpropathrin Fenpropidin Fenson Fensulfothion Fensulfothion-oxon Fensulfothion-oxon -sulfone fensulfothion-sulfone Fenthion Fenthion sulfoxide Fenthion-sulfone Fenuron Fenvalerate I Fenvalerate II {CAS # 51630-58-1} Fepropimorph Fipronil Fipronil, desulfinylFipronil-sulfide Fipronil-sulfone Flamprop-isopropyl Flamprop-methyl Fluacrypyrim Fluazifop-p-butyl Fluazinam Fluazolate Flubenzimine Fluchloralin Flucythrinate I Flucythrinate II {CAS # 70124-77-5} Fludioxonil Flufenacet Flumetralin Flumiclorac-pentyl Flumioxazin Fluometuron Fluoranthene Fluorene Fluorodifen Fluoroglycofen-ethyl Fluoroimide Fluotrimazole

Fluoxastrobin cisFluquinconazole Flurenol-butyl ester Flurenol-methylester Fluridone Flurochloridone I Flurochloridone II {CAS # 61213-25-0} Flurochloridone, deschloroFluroxypyr-1-methylheptyl ester Flurprimidol Flurtamone Flusilazole Fluthiacet-methyl Flutolanil Flutriafol Fluvalinate-tau-I Fluvalinate-tau-II {CAS # 102851-06-9} Folpet Fonofos Formothion Fosthiazate I Fosthiazate II {CAS # 98886-44-3} Fuberidazole Furalaxyl Furathiocarb Furilazole Furmecyclox Halfenprox Haloxyfop-methyl Heptachlor Heptachlor epoxide isomer A Heptachlor exo-epoxide isomer B Heptenophos Hexabromobenzene Hexachlorobenzene Hexachlorophene Hexaconazole Hexazinone Hexestrol Hydroprene Imazalil Imazamethabenz-methyl I Imazamethabenz-methyl II {CAS # 81405-85-8} Imibenconazole Imibenconazole-desbenzyl

Indeno[1,2,3-cd]pyrene Indoxacarb and Dioxacarb decomposition product [Phenol, 2-(1,3-dioxolan-2-yl)-] Ioxynil Ioxynil octanoate Ipconazole Iprobenfos Iprodione Iprovalicarb I Iprovalicarb II {CAS # 140923-25-7} Irgarol Isazophos Isobenzan Isobornyl thiocyanoacetate Isocarbamide Isocarbophos Isodrin Isofenphos Isofenphos-oxon Isomethiozin Isoprocarb Isopropalin Isoprothiolane Isoproturon Isoxaben Isoxadifen-ethyl Isoxaflutole Isoxathion Jasmolin I Jasmolin II Jodfenphos Kinoprene Kresoxim-methyl Lactofen Lenacil Leptophos Leptophos oxon Lindane Linuron Malathion Malathion-o-analog MCPA methyl ester MCPA-butoxyethyl ester MCPB methyl ester m-Cresol Mecarbam

Mecoprop methyl ester Mefenacet Mefenpyr-diethyl Mefluidide Menazon Mepanipyrim Mephosfolan Mepronil Metalaxyl Metamitron Metasystox thiol Metazachlor Metconazole I Metconazole II {CAS # 125116-23-6} Methabenzthiazuron [decomposition product] Methacrifos Methamidophos Methfuroxam Methidathion Methiocarb Methiocarb sulfone Methiocarb sulfoxide Methomyl Methoprene I Methoprene II {CAS # 40596-69-8} Methoprotryne Methoxychlor Methoxychlor olefin Methyl (2-naphthoxy)acetate Methyl paraoxon Methyl parathion Methyl-1-naphthalene acetate Methyldymron Metobromuron Metolachlor Metolcarb Metominostrobin (E) Metominostrobin (Z) {CAS # 133408-50-1} Metrafenone Metribuzin Mevinphos Mirex Molinate Monalide

Monocrotophos Monolinuron Musk amberette Musk Ketone Musk Moskene Musk Tibetene (Moschustibeten) Musk xylene Myclobutanil N,N-Diethyl-m-toluamide N-1-Naphthylacetamide Naled Naphthalene Naphthalic anhydride Naproanilide Napropamide Nicotine Nitralin Nitrapyrin Nitrofen Nitrothal-isopropyl N-Methyl-N-1-naphthyl acetamide Nonachlor, cisNonachlor, transNorflurazon Norflurazon, desmethylNuarimol o,p'-DDD o,p'-DDE o,p'-DDT Octachlorostyrene o-Dianisidine o-Dichlorobenzene Ofurace Omethoate o-Phenylphenol Orbencarb ortho-Aminoazotoluene Oryzalin Oxabetrinil Oxadiazon Oxadixyl Oxamyl Oxycarboxin Oxychlordane Oxydemeton-methyl Oxyfluorfen 15

p,p'-DDD p,p'-DDE p,p'-DDM [bis(4-chlorophenyl)methane] p,p'-DDT p,p'-Dibromobenzophenone p,p'-Dicofol Paclobutrazol Paraoxon Parathion PBB 52 Tetrabrombiphenyl PBB 101 PBB 15 PBB 169 Hexabrombiphenyl PCB 101 PCB 105 PCB 110 PCB 118 PCB 126 PCB 127 PCB 131 PCB 136 PCB 138 PCB 153 PCB 169 PCB 170 PCB 180 PCB 30 PCB 31 PCB 49 PCB 77 PCB 81 p-Dichlorobenzene Pebulate Penconazole Pendimethalin Pentachloroaniline Pentachloroanisole Pentachlorobenzene Pentachloronitrobenzene Pentachlorophenol Pentanochlor Permethrin I Permethrin II {CAS # 52645-53-1} Perthane Phantolide Phenamiphos 16

Phenanthrene Phenanthrene-d10 Phenkapton Phenol Phenothiazine Phenothrin I Phenothrin II Phenoxyacetic acid Phenthoate Phorate Phorate sulfone Phorate sulfoxide Phorate-oxon Phosalone Phosfolan Phosmet Phosphamidon I Phosphamidon II {CAS # 13171-21-6} Phthalide Phthalimide Picloram methyl ester Picolinafen Picoxystrobin Pindone Piperalin Piperonyl butoxide Piperophos Pirimicarb Pirimiphos-ethyl Pirimiphos-methyl Plifenat p-Nitrotoluene Potasan Prallethrin, cisPrallethrin, trans- {CAS # 23031-36-9} Pretilachlor Probenazole Prochloraz Procymidone Prodiamine Profenofos Profenofos metabolite (4-Bromo2-chlorophenol) Profluralin Prohydrojasmon I Prohydrojasmon II {CAS # 158474-72-7}

Promecarb Promecarb artifact [5-isopropyl3-methylphenol] Prometon Prometryn Propachlor Propamocarb Propanil Propaphos Propargite Propargite metabolite [Cyclohexanol, 2-(4-tert-butylphenoxy)] Propazine Propetamphos Propham Propiconazole-I Propiconazole-II {CAS # 60207-90-1} Propisochlor Propoxur Propyzamide Prosulfocarb Prothioconazole-desthio Prothiofos Prothoate Pyracarbolid Pyraclofos Pyraflufen-ethyl Pyrazon Pyrazophos Pyrazoxyfen Pyrene Pyrethrin I Pyrethrin II Pyributicarb Pyridaben Pyridaphenthion Pyridate Pyridinitril Pyrifenox I Pyrifenox II {CAS # 88283-41-4} Pyriftalid Pyrimethanil Pyrimidifen Pyriminobac-methyl (E) Pyriminobac-methyl (Z) {CAS # 136191-64-5}

Pyriproxyfen Pyroquilon Quinalphos Quinoclamine Quinoxyfen Quintozene metabolite (pentachlorophenyl methyl sulfide) Quizalofop-ethyl Rabenzazole Resmethrin Resmethrine I Resmethrine II {CAS # 10453-86-8} Rotenone S,S,S-Tributylphosphorotrithioate Schradan Sebuthylazine Sebuthylazine-desethyl Secbumeton Silafluofen Silthiopham Simazine Simeconazole Simetryn Spirodiclofen Spiromesifen Spiroxamine I Spiroxamine II {CAS # 118134-30-8} Spiroxamine metabolite (4-tert-butylcyclohexanone) Sudan I Sudan II Sudan Red Sulfallate Sulfanilamide Sulfentrazone Sulfotep Sulfur (S8) Sulprofos Swep Tamoxifen TCMTB Tebuconazole Tebufenpyrad Tebupirimifos Tebutam Tebuthiuron

Tecnazene Tefluthrin, cisTemephos Terbacil Terbucarb Terbufos Terbufos-oxon-sulfone Terbufos-sulfone Terbumeton Terbuthylazine Terbuthylazine-desethyl Terbutryne Tetrachlorvinphos Tetraconazole Tetradifon Tetraethylpyrophosphate (TEPP) Tetrahydrophthalimide, cis-1,2,3,6Tetramethrin I Tetramethrin II {CAS # 7696-12-0} Tetrapropyl thiodiphosphate Tetrasul Thenylchlor Theobromine Thiabendazole Thiazopyr Thifluzamide Thiofanox Thiometon Thionazin Thymol Tiocarbazil I Tiocarbazil II {CAS # 36756-79-3} Tolclofos-methyl Tolfenpyrad Tolylfluanid Tolylfluanid metabolite (DMST) Tolyltriazole [1H-Benzotriazole, 4-methyl-] Tolyltriazole [1H-Benzotriazole, 5-methyl-] Tonalide Toxaphene Parlar 26 Toxaphene Parlar 50 Toxaphene Parlar 62 trans-Chlordane Transfluthrin Traseolide Triadimefon

Triadimenol Tri-allate Triamiphos Triapenthenol Triazamate Triazophos Tributyl phosphate Tributyl phosphorotrithioite Trichlamide Trichlorfon Trichloronate Triclopyr methyl ester Triclosan Triclosan-methyl Tricresylphosphate, metaTricresylphosphate, orthoTricresylphosphate, para Tricyclazole Tridemorph, 4-tridecylTridiphane Trietazine Triethylphosphate Trifenmorph Trifloxystrobin Triflumizole Trifluralin Triphenyl phosphate Tris(2-butoxyethyl) phosphate Tris(2-chloroethyl) phosphate Tris(2-ethylhexyl) posphate Triticonazole Tryclopyrbutoxyethyl Tycor (SMY 1500) Uniconizole-P Vamidothion Vernolate Vinclozolin XMC (3,4-Dimethylphenyl N-methylcarbama XMC (3,5-Dimethylphenyl N-methylcarbama Zoxamide Zoxamide decomposition product

17

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA April 18, 2006 5989-5076EN

Screening for Hazardous Chemicals in Homeland Security and Environmental Samples Using a GC/MS/ECD/FPD with a 731 Compound DRS Database Application Note

Homeland Security, Environmental

Authors
Bruce Quimby Mike Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 UASA

interpretation of the MS data, especially in samples with high matrix contamination. The combination of selective GC detectors, SIM/Scan, and deconvolution makes a very powerful hazardous chemical analysis system that shows significant progress toward the above goals.

Introduction
In recent years, there has been increasing concern over the release of hazardous chemicals through either accidental or intentional acts. Both the homeland security and environmental communities recognize the need for preparing analytical laboratories that can respond quickly to such incidents. The terms toxic industrial chemicals/toxic industrial materials (TIC/TIM) are used in homeland security to refer to hazardous chemicals, while the environmental community uses different terminology like hazardous materials. In either case, the challenge is to develop laboratory methods with the capability of identifying any hazardous chemical(s) involved in an incident and to be able to measure its concentration in collected samples. There are several significant challenges to face when developing methods for this analysis. The methods must able to: Rapidly and accurately identify the specific toxic agents involved Measure concentration correctly at high levels of agent at the epicenter (high dynamic range) Measure concentration correctly at low levels of agent at perimeters and during decontamination (low detection limits)

Abstract
Response to homeland security or environmental incidents involving hazardous chemicals requires first, the rapid and accurate identification of the chemical agent(s) involved and second, the quantitative measurement of that agent in large numbers of samples to aid in managing the response. Given the unknown nature of the analytes and the complexity of matrices that could be encountered, developing analytical methods for this analysis is challenging. The approach described in this work uses a gas chromatography/mass spectrometry (GC/MS) system with a micro-fluidic splitter added to the end of the column. The splitter divides the column effluent between the MS and either a dual-wavelength flame photometric detector (DFPD) or a micro electron-capture detector (ECD) and a single-wavelength FPD. This approach allows the simultaneous collection of MS and two channels of selective GC detector data from a single injection. This multisignal configuration provides: full-scan MS data for library searching, selective ion monitoring (SIM) data for trace analysis, ECD and FPD data for excellent selectivity and sensitivity in complex matrices. The systems use retention time locking (RTL) to produce retention times (RTs) that precisely match those in a 731 compound database of hazardous chemicals. Deconvolution Reporting Software (DRS) is used to provide fast and accurate

Be highly selective over matrix interferences (wood smoke, fuels, burning tires, etc.) to minimize both false positives and false negatives Indentify as many toxic agents as possible Handle large numbers of samples It is clear that there is no single analytical technique that can be used for detecting all possible hazardous chemicals. However, one technique that is widely applicable for the identification and measurement of broad classes of hazardous chemicals is GC/MS. GC/MS is widely used in laboratories worldwide for the analysis of thousands of different chemicals. GC/MS methods are typically developed to analyze between 10 and 100 individual compounds. A target compound is deemed to be present if the target ion and two or three qualifier ions, with specific abundance ratios, fall within a defined RT window. The identity of the target may be further confirmed by comparison of the scan at the apex of the peak with a library reference spectrum. Matrix interferences are usually minimized by optimizing a combination of the sample preparation, GC, and MS parameters. Since most methods only deal with at most a few matrix types, the ions chosen for identification purposes can be selected such that they are minimized in the matrix. With the limited number of targets addressed by the method, recalibration of response factors, RTs, and qualifier ion abundance ratios can be accomplished with the injection of a few calibration mixtures. General screening methods for very large numbers of targets in widely varying and complex matrices offer a new set of challenges for the method developer. When screening for hundreds of targets, several factors must be addressed: Use of sample preparation to reduce matrix interferences is now significantly limited because rigorous cleanup steps may unintentionally remove targets. This reduced level of cleanup can result in significantly higher levels of matrix interferences to contend with. Recalibration of response factors, RTs, and qualifier abundance ratios is difficult or impossible because of the large number of targets. The methods may be deployed in laboratories without access to standards for all of the targets.

The time required for data review of hundreds of targets in complex matrices can become unmanageably large. Even with a very large database of targets, it is possible that hazardous chemicals not in the target list could be present in a sample. Recently, several techniques have become available to help address the above set of challenges. RTL produces RTs that precisely match from instrument-to-instrument and to those in a database [1]. This eliminates the need for recalibration of the individual RTs and timed events. The introduction of reliable and inert microfluidic splitters allows for the simultaneous collection of mass spectral data and, for example, phosphorus, sulfur, and/or electron capture data [2]. The selective detector chromatograms can highlight suspect compounds even if they are not in the MS target list. They can also offer an alternative means for quantitation of target analytes. The introduction of the synchronous SIM/Scan feature allows for the simultaneous acquisition of both full scan and SIM data from the same injection [2, 3]. The scan data can be used for screening the full list of targets in the database while the SIM data looks for a high priority subset of compounds down to very low levels. One of the most significant tools developed for dealing with complex matrices is Agilents Deconvolution Reporting Software (DRS) [4]. It uses advanced computational techniques to extract the spectra of targets from those of overlapped interference peaks. It then compares the extracted spectrum with a library to determine if the target is present. Any hits are confirmed by searching against the main NIST MS reference library. This process is automated and provides significant time savings in data interpretation. Since it deals with the entire spectrum instead of just four ions, DRS can often correctly identify a target in the presence of interferences where the typical approach would fail. The use of DRS substantially reduces the number of both false positives and false negatives. This application note describes the combination of the above techniques with a database of 731 hazardous chemicals, the Agilent Hazardous Chemical DBL (HCD), to be used for screening purposes. The compounds were chosen because of their significance in environmental or food safety analysis. The reasoning is that if the materials are manufactured in significant quantities and are toxic, they would be likely to appear in an

environmental method. The pesticides are included because many exhibit toxicity. The list is comprised of: Chlorinated Dioxins and Furans: EPA 8280A, 10 compounds Polychlorinated biphenyls: EPA 8082, 19 compounds Volatiles: EPA 502/524, 60 compounds Semivolatiles: EPA 8270C Appendix IX, 140 compounds Pesticides: Agilent RTL Pesticide Database (adapted), 567 compounds Total: 796 compounds, with 65 compounds in two groups, or 731 individual compounds The names of all the compounds in the database are listed in Appendix A at the end of this note. The above list by no means contains all of the hazardous chemicals that could be encountered. However, it does screen for a large number of known hazards and with the addition of selective detection can highlight other nontarget compounds that may be of interest.

and where the fuel components are not of interest. In the examples shown below the database with hydrocarbons removed was used, since fuels were used as prototype matrices.

System Configuration
The system configurations used are shown in Figure 1A and 1B.

A
Auto-sampler Phosphorus FPD
3-Way effluent splitter with makeup

AUX EPC 3.8 psig

ECD

Column 6890N GC

5975 Inert MSD

The chromatographic conditions chosen for development of the database are general in nature and are compatible with the analysis of other types of compounds beyond those in the table. For example, laboratories with access to calibration standards for chemical warfare agents (CWA) can add CWA data to the tables and screen for them as well. The RTs for compounds in the database were collected with the column outlet pressure at 3.8 psig using a microfluidic splitter. This was done to assure that the RTs observed during sample analysis would closely match those in the database when a microfluidic splitter or QuickSwap is used. The chromatographic conditions for the database were chosen to be compatible with the method translation technique. Constant pressure mode was used in the GC inlet so that method translation can be used to precisely time scale the methods for faster operation [5]. Provided with the Agilent Hazardous Chemicals DBL are the files to run the analysis precisely threefold (3X) and sevenfold (7X) faster than the primary database (1X). Also, each of the three-speed variations of the database are provided in two forms: one with the entire set of 731 compounds and one with the 36 aromatic hydrocarbons removed. The latter is provided for use with samples known to contain fuels

Auto-sampler Dual Flame Photometric Detector Sulfur Phosphorus AUX EPC 3.8 psig 3-Way effluent splitter with makeup

Column 6890N GC

5975 Inert MSD + Performance electronics

Figure 1.

System configurations. A). GC/MS/ECD/FPD system used for 1X and 3X screening analyses. B). GC/MS/DFPD system used for 7X screening analyses.

Key components are: Fast Oven The primary 1X method only requires the 120V oven. With the 6890N 240V oven (option 002), the screening analysis method can be run precisely three times faster (14.33 min) using a 15-m HP-5MS column. If the 240V GC is further equipped with SP1 2310-0236 (puts MSD interface in back of oven under rear injection port) and
3

using the G2646-60500 oven-insert accessory, the speed can be increased to seven times faster (6.14 min) with a 5-m HP-5MS column. Note that use of the oven insert prevents use of the front inlet and detector positions. Only one detector is available for splitting. The DFPD is a good choice for this configuration, as it uses only one detector position but generates two signals. ECD The 6890N Option 231 is a ECD. The signal from the electron capture detector (ECD) is collected, stored, and processed by the MS ChemStation simultaneously with the MS data. ECDs are selective in nature and exhibit very sensitive response to halogenated compounds, with detection limits below 1 pg for polyhalogenates. They also respond to several other functional groups like nitro compounds. They do, however, also respond to some fairly low-priority compounds, like phthalate esters. The ECD data can be used in several ways. Nontarget halogenated or nitro compounds are highlighted. The presence of an electrophore at the RT of an identified compound can be used to support confirmation of identity. The response on the ECD can be used for quantitative analysis, but only after calibration with a standard, as the response factors are compound dependent and can vary significantly with compound class. Single FPD The 6890N Option 240 is a single FPD. It is used to selectively detect either sulfur or phosphorus. The detector is usually run in the phosphorus mode to highlight such compounds as organophosphorus pesticides and nerve agents. In the phosphorus mode, the detector is highly selective (>106) with a very low (~0.050 pg) detection limits for phosphorus. The ability of the FPD to uncover nontarget organophosphorus compounds like new pesticides or designer nerve agents is especially helpful. The presence of phosphorus at the RT of an identified compound can be used to support confirmation of identity. Because the response per unit weight of phosphorus is relatively consistent from compound to compound, the FPD can be used for semi-quantitative analysis in situations where no calibration standard is available for an identified analyte. Dual FPD The 6890N Option 241 is a DFPD with two optical detection channels that measures sulfur and phosphorus simultaneously. The DFPD sulfur response is also selective (>104) and sensitive (detection limits <10 pg) , although not as much as phosphorus.
4

The sulfur signal is also quadratic with respect to the amount of sulfur injected. It is often used to detect sulfur-mustard agents and for confirmation of sulfur-containing pesticides. The response per unit weight of sulfur is relatively consistent from compound to compound, but varies more than that of the phosphorus signal. Microfluidic Splitter The 6890N Option 890 (3-way splitter) or Option 889 (2-way splitter) uses diffusion-bonded plate technology combined with metal column ferrules to make an inert, easy-to-use, leak free, hightemperature column-effluent splitter. The splitter uses Auxiliary EPC for constant pressure makeup (6890N Option 301). The Auxiliary EPC makeup can be pressure programmed at the end of the run to higher pressure, while at the same time the inlet pressure is lowered to near ambient. This causes the flow in the column to reverse direction, backflushing heavy materials out the split vent of the inlet. Backflushing can greatly reduce analysis times for samples that contain high-boiling matrix components [6]. The Aux EPC also allows column changing and maintainance without venting the MSD. When the column fitting is removed from the splitter, helium from the makeup supply purges the fitting, preventing air from entering the MSD. If the column is attached to the splitter but removed from the inlet, helium flows backwards through the column and out the inlet end. Inlet maintainance or column headtrimming can be done without cooling and venting the MSD to prevent sucking air into a hot source. MSD System The 5975 inert MSD with performance turbo (G3243A) or 5973N inert MSD with performance electronics and performance turbo (G2579A), EI (electron impact ionization mode) MSD is used. These configurations provide faster full scan rates while maintaining sensitivity. The scan rates are compatible with the narrower peaks generated by fast chromatography. The performance turbo pump is required to handle the higher flows associated with the screening method. Synchronous SIM/Scan The D.02.00 (or higher) revision of the Agilent MSD ChemStation is used because it supplies the synchronous SIM/Scan feature. SIM/Scan operates by collecting SIM data every other cycle and scan data on alternate cycles throughout the entire chromatogram. The signal-to-noise performance of the collected SIM and scan data is virtually identical to that obtained with SIM-only and scan-only

methods. As with conventional SIM methods, not all 731 targets can be monitored in a single run due to the required time separation between SIM groups. In general, the acquisition of SIM data is set up to collect high-priority targets at very low levels. Examples would be the chlorinated dioxins and CWAs. DRS Software (G1716AA) Spectral deconvolution of the MS data enables identification of analytes in the presence of overlapped matrix peaks [4]. This significantly reduces chromatographic resolution requirements, which allows detection of targets in higher levels of matrix or can be used with fast chromatography to shorten analysis times. DRS uses the AMDIS deconvolution program from NIST, originally developed for trace chemical-weapons detection in complex samples. DRS presents the analyst with three distinct levels of compound identification: ChemStation, based on RT and four-ion agreement AMDIS, based on cleaned spectra full-ion matching and locked RT NIST05 search using a 163000 compound library Hazardous Chemical DBL (G1671AA) This supplies the mass spectral library, method, and DRS files for the 731 compound-screening method.

volatility from gases to large polynuclear aromatic hydrocarbons (PAHs). Splitless injections are usually incompatible with the lowest boiling volatiles due to problems with the solvent. For low matrix samples where semivolatiles are of interest, splitless injections can be used. For ambient headspace analysis [7], the conditions are listed separately at the bottom of Table 1. The liner used for ambient headspace was 1-mm id straight through (no glass wool) and Siltek coated (Restek, part number 20973-214.5). The auto injector parameters are critical in ambient headspace and are listed in Table 1. The volatiles samples run by ambient headspace were prepared as described in Reference 7. While the targets in the table cover a very broad range of boiling points, it is usually not practical to screen for all of them in one run. This is because an analysis for semivolatile compounds would be done with a solvent that would occlude the lowest boiling volatiles in the table. Conversely, a method for injecting the lowest boiling compounds would usually not be suitable for the highest boiling. The MSD solvent delays listed in Table 1 are based on isooctane as the solvent in a semivolatiles analysis. If a lower boiling solvent is used, it may be possible to reduce these delays accordingly. Some of the target compounds were found to have sufficiently high boiling points to require higher inlet and detector temperatures. These were the higher molecular weight PAHs, the polychlorinated dioxins, and the polychlorinated furans. For these compounds the inlet temperature, MS source, and transfer line were also raised to 300 C. Without this increase in temperature, the compounds would exhibit tailing and in some cases reduction in signal. The trade-off with temperature is that the performance of some thermally labile compounds is degraded at the higher temperatures. The MSD data acquisition sampling rates listed in Table 1 are for scan mode only. For volatiles analysis, the scan rate is increased one step. It is also increased one step when SIM/Scan is used. In SIM/Scan mode the SIM dwell time was set to 40 milliseconds for each ion monitored. The microfluidic splitter parameters are chosen to provide the desired flow ratio between detectors while meeting the flow requirements of the detectors used. A primary consideration is to make sure that the flow to the MSD does not exceed ~4 mL/min while collecting analyte data. It was also desired to split the effluent equally between the DFPD and MSD in the 2-way split configuration. In the 3-way configuration, the split to the ECD was reduced
5

Instrument Operating Parameters


The instrument operating parameters used (unless noted otherwise) are listed in Table 1. These are starting conditions and may have to be optimized. The split/splitless injection port was used for all work described here. It was chosen for its flexibility, allowing splitless injections for clean samples and split injections for dirty or high-concentration samples. It is also compatible with column backflushing. For all cases (except ambient headspace), the inlet liner used was the 4-mm id Siltek Cyclosplitter (Restek, part number 20706-214.1). This inlet liner was found to be of low activity, as it does not contain glass wool. Proper mixing for split injections is done by the internal liner geometry. Except as noted, split injections with a split ratio of 10:1 were used. For high matrix samples, this roughly matches the amount of matrix injected with the column capacity. If excess amounts of matrix are injected, the RTs of targets can shift. Split injection is also the easiest and most reliable way of screening samples for analytes ranging in

Table 1.

Gas Chromatograph and Mass Spectrometer Conditions Original 1X Method 3X Method 7X Method

GC Agilent Technologies 6890N 7683 Autoinjector and Tray Inlet Mode Injection type Injection volume (uL) Inlet temp ( C) Pressure, nominal (psig) RT Locking compound RT Locking time (min) Split ratio Gas saver Gas type Oven Voltage (VAC) Initial oven temp (C) Initial oven hold (min) Ramp rate (C/min) Final temp (C) Final hold (min) Total run time (min) Equilibration time (min) Column Type Agilent part number Length (m) Diameter (mm) Film thickness (um) Outlet pressure (AUX EPC, psig) FPD or DFPD Type Temperature (C) Hydrogen flow (mL/min) Air flow (mL/min) Mode: Constant makeup flow Nitrogen makeup flow (mL/min) Data rate (Hz) ECD Temperature (C) Nitrogen makeup flow (mL/min) Mode: Constant makeup flow Data rate (Hz) AUX EPC Pressure Pressure (psig) Gas type EPC Split/Splitless Constant pressure Split 1.0 250 31.17 Tripropyl phosphate 12.874 10:1 Off Helium EPC Split/Splitless Constant pressure Split 1.0 250 23.96 Tripropyl phosphate 4.291 10:1 Off Helium EPC Split/Splitless Constant pressure Split 1.0 250 8.84 Tripropyl phosphate 1.839 10:1 Off Helium

120 or 240 40 2 10 300 15 43.00 0.5

240 40 0.667 30 300 5 14.33 0.5

240 (and pillow) 40 0.286 70 300 2.143 6.14 0.5

HP 5-MS inert 19091S-433i 30 0.25 0.25 3.8

HP 5-MS 19091S-431 15 0.25 0.25 3.8

HP 5-MS Custom 5 0.25 0.25 3.8

Single, Phosphorus 250 75 100 60 5

Single, Phosphorus 250 75 100 60 10

Dual, S and P 250 75 100 60 10

30 0 60 5

300 60 10

N/A N/A N/A

3.8 Helium

3.8 Helium

3.8 Helium

Table 1.

Gas Chromatograph and Mass Spectrometer Conditions (Continued)

MSD Agilent Technologies Tune file Mode Solvent delay (min) EM voltage Low mass (amu) High mass (amu) Threshold Sampling Scans/s Quad temp (C) Source temp (C) Transfer line temp (C) Splitter Type 6890N option number Flow ratio [Deactivated fused silica tubing] MSD restrictor length (m) MSD restrictor id (mm) FPD/DFPD restrictor length (m) FPD/DFPD restrictor id (mm) ECD restrictor length (m) ECD restrictor id (mm) Ambient Headspace Inlet Mode Injection type Inlet temp ( C) Pressure, nominal (psig) RT locking compound RT locking time (min) Split ratio Gas saver Gas type Autoinjector Sample washes Sample pumps Injection volume (L) Syringe size (L) PreInj Solvent A washes PreInj Solvent B washes PostInj Solvent A washes PostInj Solvent B washes Viscosity delay (s) Plunger speed Pre-injection dwell (min) Post-injection dwell (min) Sampling depth (mm) [critical!]

5975 inert MSD Atune.U Scan 2.20 Atune voltage 35 565 0 1 5.23 150 230 280

5975 inert MSD Atune.U Scan 0.82 Atune voltage 35 565 0 1 5.23 150 230 280

5973 inert with Performance Electronics Atune.U Scan 0.40 Atune voltage 35 565 0 0 9.46 150 230 280

3 way 890 1:1:0.1 MSD:FPD:ECD 1.44 0.18 0.53 0.18 0.51 0.10

3 way 890 1:1:0.1 MSD:FPD:ECD 1.44 0.18 0.53 0.18 0.51 0.10

2 way 889 1:1 MSD:DFPD 1.44 0.18 0.53 0.18 N/A N/A

EPC Split/Splitless Constant pressure Split 200 31.17 Tripropyl phosphate 12.874 1:1 Off Helium

0 3 50 100 0 0 1 3 5 Fast 0 0 20

to 1/10th that going to the MSD and FPD because of the extreme sensitivity of the detector. The lengths and diameters of the detector restrictors were calculated using the spreadsheet calculator included with the splitter. The peak recognition windows used in the Agilent ChemStation were set to 0.2 min and in AMDIS to 12 s. these values were found to be sufficiently wide enough to compensate for some RT drift yet narrow enough to minimize the number of false positives. The minimum match factors setting in AMDIS was set to 45. This value seemed to give the least number of false positives and false negatives.

relatively non-polar volatiles in water. It is convenient for labs that need to screen samples for volatiles but do not have a dedicated headspace sampler. The conversion from liquid sampling to ambient headspace simply requires changing the inlet liner and the autosampler syringe. Figure 2 shows the chromatograms from a run using the system in Figure 1A. A mixture of 14 halogenated volatiles was spiked into water at 2 ppm. Fifty microliters of the approximately 1 mL of headspace in the vial was injected. With the exception of peaks 3 and 4, which coelute, the compounds are well separated. The ECD chromatogram is inverted for comparison with the MS total ion chromatogram from the full-scan data. All of the volatiles respond on the ECD, although the response to compounds 1, 2, and 8 is significantly lower than for the rest of the compounds. In general with an ECD the response to a compound increases dramatically with the number of halogens in the molecule. Since none of the compounds contain phosphorus, there is no response on the FPD. Figure 3 shows the DRS report for the sample. For each compound identified, the RT, Chemical Abstracts number (CAS#), and compound name are listed. A line is generated in the report if a compound is found by either the Agilent ChemStation, AMDIS, or both.

Results
Volatiles To evaluate the HCD method for volatiles analysis, headspace injection was chosen. Headspace injections are usually done with an automated heated sampler specifically designed for the purpose. Ambient headspace [7] is a variant of the technique that uses a gastight syringe in the liquid autosampler and injects the headspace from a 2-mL vial. It is unheated, and is thus limited to compounds that are volatile at room temperature. Ambient headspace works well for the analysis of

3,4 7 2 1 5 6 8 10 9 11 12

TIC: 2ppmMIX 3_Only_simscan.D\DATA.MS

14 13

TIC

ECD

FPD
2 4 6 8 10 12

Peak identities 1) 1,2-Dichloroethane 2) 1,1-Dichloropropylene 3) 1,2-Dichloropropane 4) Trichloroethylene

5) 6) 7) 8) 9)

cis-1,3-Dichloropropylene trans-1,3-Dichloropropylene 1,1,2-Trichloroethane 1,3-Dichloropropane 1,2-Dibromoethane

10) 11) 12) 13) 14)

1,1,1,2-Tetrachloroethane 1,1,2,2-Tetrachloroethane 1,2,3-Trichloropropane 3-Chloro-1,2-dibromopropane Hexachlorobutadiene

Figure 2. 8

Ambient headspace analysis of volatile organics in water, spiked at 2 ppm per component.

Figure 3.

DRS report for the analysis in Figure 2.

The report shows that a compound has been determined as present by the Agilent ChemStation if a value appears in the Agilent ChemStation Amount column. This means the identification criteria set in the DATA ANALYSIS section of the method have been met. Typically the criteria are that the target ion is present and all three qualifier ions are present in ratios that fall within the percent uncertainty values for that compound. The Agilent ChemStation Amount listed is a very rough approximation of the amount of the compound, in nanograms, reaching the MS. This is based on the response factor originally observed when the HCD table data was collected. Since valid quantitation requires recent recalibration of response factors on the specific instrument used for analysis, the numbers in this column should never be used to report concentrations of identified analytes. The error in these values can easiliy be a factor of 10 or higher. The purpose of the listed values is to give an approximate amount that can be used to guide standard preparation for quantitative calibration of the compound, if needed. The match value listed under the AMDIS column is the degree to which the extracted (deconvolved) spectrum of the peak at that RT matched the spectrum in the HCD AMDIS target library. The higher this number, the better the spectra agree. The

column R.T. Diff sec. lists the difference in seconds between the observed RT and that in the AMDIS target library. The lower this number, the better the RTs agree. The NIST column lists the reverse-match quality of the extracted spectrum compared with the NIST05 main library spectrum with the same CAS#. The entry Hit Num. is the number of the hit in the NIST search results that has the same CAS# as the identified compound. The higher the reverse-match value and the lower the hit number, the better the extracted spectrum matches with NIST05. The NIST column serves as a second opinion on the identity of the extracted spectrum. The analysis in Figure 2 is of course an easy one, but serves to demonstrate how the system works. All 14 spiked compounds were found by both the Agilent ChemStation and AMDIS. The certainty of identification is very high because: The target ion and three qualifier ions are present in appropriate ratios and at the appropriate time as determined by the Agilent ChemStation The deconvolved spectrum and the RT at which it appears closely matches the data in the AMDIS target library.
9

The extracted spectrum of the identified compound also matches the spectrum with the same CAS # in the NIST05 library. The compounds all have a significant response on the ECD, as expected from their halogen content. To challenge the system in a more realistic way, the effect of matrix and dilution of the analytes was studied. Additional samples were prepared that contained: the same 2-ppm mixture of analytes plus 100 ppm of pump gasoline; 100 ppb of analytes only; and 100 ppb of analytes plus 100 ppm of pump gasoline. Figure 4 shows the chromatograms from the 100 ppb of analytes with 100 ppm of gasoline. The complexity of the TIC chromatogram illustrates the severe matrix challenge presented by the thousand-fold excess of gasoline. In the ECD chromatogram, interference peaks are now apparent. However, with the exception of peaks 1, 2, 8, and 12, all of the analytes peaks are still visible above the matrix interferences.

Table 2 summarizes the results from the matrix and dilution experiments. In the sample that was 2 ppm of analytes with 100 ppm of gasoline, the Agilent ChemStation (column labeled Quant) found all but two of the compounds. Those two compounds had qualifier ions out of range due to interferences from the matrix. AMDIS successfully found all 14 compounds. Also, with the exception of compound 8, all of the analytes were clearly visible above the matrix responses on the ECD chromatogram. In the sample that contained 100 ppb of analytes but without gasoline, quant found 7 of the 14 analytes. Using full-scan data, the signal to noise ratio for most of the analytes at the 100-ppb level is very low. This results in difficulties with finding the qualifier ions in ratios that fall within the specified uncertainty range in the quant calibration table. AMDIS found 11 of the 14 compounds. Peak 3 was not found due to a severe overlap with the coeluting peak number 4. Peaks 9 and 13 were missed by AMDIS because the signal to noise ratio was too low.

TIC

8 1 2 5 6 7 11

12

ECD
9 10 13 14

3,4

FPD
2 4 6 8 10 12

Peak identities 1) 1,2-Dichloroethane 2) 1,1-Dichloropropylene 3) 1,2-Dichloropropane 4) Trichloroethylene

5) 6) 7) 8) 9)

cis-1,3-Dichloropropylene trans-1,3-Dichloropropylene 1,1,2-Trichloroethane 1,3-Dichloropropane 1,2-Dibromoethane

10) 11) 12) 13) 14)

1,1,1,2-Tetrachloroethane 1,1,2,2-Tetrachloroethane 1,2,3-Trichloropropane 3-Chloro-1,2-dibromopropane Hexachlorobutadiene

Figure 4.

Ambient headspace analysis of volatile organics in water. Analytes at 100 ppb plus pump gasoline at 100 ppm.

10

Table 2.

Effect of Matrix and Concentration on DRS Results 2 ppm STD only 2 ppm STD with 100 ppm gasoline Quant AMDIS (ng) (match) 2.47 7.34 5.59 7.71 4.81 93 98 64 97 98 84 3.05 3.50 96 97 95 5.32 2.41 1.85 2.40 3.54 12 99 98 98 90 89 14 0.65 7 0.07 0.12 0.37 0.21 0.40 0.21 100 ppb STD only Quant (ng) AMDIS (match) 73 89 Overlap 90 88 53 72 66 S/N 89 48 79 S/N 75 10 0.36 7 0.31 0.19 0.14 0.22 0.30 0.23 100 ppb STD with 100 ppm gasoline Quant AMDIS (ng) (match) 65 85 Overlap 82 74 Overlap Overlap 46 66 88 53 75 59 52 11

RT (min) 1.491 1.536 1.793 1.863 2.317 2.658 2.735 2.938 3.250 4.003 5.151 5.283 8.208

Compound 1,2-Dichloroethane 1,1-Dichloropropylene 1,2-Dichloropropane Trichloroethylene cis-1,3-Dichloropropylene trans-1,3-Dichloropropylene 1,1,2-Trichloroethane 1,3-Dichloropropane 1,2-Dibromoethane 1,1,1,2-Tetrachloroethane 1,1,2,2-Tetrachloroethane 1,2,3-Trichloropropane 3-Chloro-1,2-dibromopropane

Peak Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Quant (ng) 2.27 7.60 4.92 7.58 4.39 3.30 2.82 3.39 2.60 5.15 2.38 1.89 1.62 16.46 14

AMDIS (match) 97 100 95 99 98 97 99 98 91 100 99 98 93 94 14

10.435 Hexachlorobutadiene Total Found

With 100 ppm of gasoline added to the 100-ppb sample, quant again found 7 of the 14 compounds and AMDIS again found 11 of the 14. Curiously, in both cases some of the compounds missed in the absence of matrix were now found. It is possible that the presence of matrix enhances the concentration of some of the analytes in the headspace. The compounds missed in quant were again the result of low signal to noise and/or interference. In AMDIS the three missed peaks were due to severe interferences from the gasoline. As indicated above, the ECD response from 10 of the 14 compounds was still visible above the peaks due to interferences. SIM/Scan The quant data in Table 2 was generated using full scan mode. Peak 13 was missed in quant due to low signal to noise ratio. SIM/Scan mode can be

used to collect SIM data simultaneously with the scan data. The 100 ppb plus 100-ppm gasoline sample was run in SIM/Scan mode with SIM groups for each of the 14 analytes. Figure 5 compares the target and qualifier extracted ion chromatograms in both modes with the ECD response for peak 13. The signal-to-noise (peak to peak) for the target ion increases from 34 in full scan mode to 433 in SIM mode. The peaks lost in quant due to low signal-to-noise were all recovered in SIM mode. This example demonstrates the power of SIM/Scan when looking for high-priority targets at low levels. If necessary, the ECD could also be used for quantitation, as it has a high signal to noise ratio and is free from interference.

11

Ion 157 Scan Ion 75 Scan

s/n (pk-pk) = 34

Ion 155 Scan

Ion 39 Scan s/n (pk-pk) = 433

Ion 157 SIM Ion 75 SIM Ion 155 SIM

Ion 39 SIM

ECD

s/n (pk-pk) = 122

8.0

8.1

8.2

8.3

8.4

Figure 5.

Target and qualifier extracted ion chromatograms for peak 13 (3-Chloro-1,2-dibromopropane) in Figure 4. SIM, scan, and ECD data collected simultaneously.

AMDIS Figure 6 illustrates the ability of AMDIS to clean the interference ions from the spectrum of an analyte. The raw spectrum at the top of Figure 6 was taken at the apex of peak 13 in the 100 ppb plus 100-ppm gasoline sample. When searched against the NIST05 library using the NIST search program, the actual compound (3-Chloro-1,2-dibromopropane) was the 70th hit in the search results. Using manual subtraction of nearby spectra in the Agilent ChemStation data analysis program improved the quality of the spectrum so that it was now the second hit when searched in NIST. This is a tedious process, however, when dealing with a large number of analytes. The spectrum as deconvolved by AMDIS is shown in Figure 6 above the

NIST05 library spectrum. When this spectrum is searched, it is the first hit in the results. The automated deconvolution provided by AMDIS saves an enormous amount of time in the data review process. Fast Methods When a retention time locked database is constructed, the RTs are (or at least should be) collected under the highest resolution conditions expected for the application. If the database is collected under constant pressure mode, method translation can then be used to adjust the speed of the method to meet the needs of different situations.

12

100

155 44 75 55 81 97 117 132 188

50

Raw spectrum
211 227 246 261 280 300

0 157 75 38 0 30 58 60 95 90 115 120 132 150 157 75 50 39 49 0 49 50 100 30 50 39 75 70 90 110 130 157 150 170 190 61 85 93 93 105 136 119 129

100

50

Manual subtraction
188 180

211 210

230

246 240

261 270

282

300 300

100

Deconvolved spectrum
188 187 199

NIST 05 library

Figure 6.

Comparison of raw, manually subtracted, AMDIS deconvoluted, and NIST05 reference spectra for peak 13 (3-Chloro-1,2-dibromopropane) in Figure 4.

The 3X method uses RTs in its database that are simply the RTs from the 1X method divided by exactly 3. The 7X method likewise uses RTs that are 1/7 of those in the original database. The quality of RTs matching between the two new faster methods and the new divided databases is demonstrated in Figure 7. Three different mixtures containing 13 chlorinated hydrocarbons and 36 pesticides were run with the two methods. The RTs were compared to those in the two new databases. The graph at the top of Figure 7 plots the database RT on the x-axis versus the difference of the measured RT from the database on the y axis. If the RT matching were perfect, the plot would be a straight horizontal line at zero height on the y axis. The maximum deviation from the table values for the 3X method was 0.047 min. The plot

indicates that a peak recognition window of 0.1 min should be sufficient. The maximum deviation in the 7X plot at the bottom of Figure 8 is +0.032 min indicating that the same peak recognition window could be used here as well. In general the RTs in scaled methods agree very well with the predicted RTs. The conditions for the two higher-speed methods were chosen to increase speed while maintaining the same column capacity. The capacity is important for both the dynamic range of quantitative measurements and for minimizing analyte RT shifts in samples with high levels of matrix. In gas chromatography, the well-known triangle of speed, resolution, and capacity dictates that if the capacity is to be maintained and the speed is to be increased, then the resolution will decrease.

13

0.100

3X Measured vs. Table

0.050 Difference (min)

0.000

_0.050

_0.100 0.0 2.0 4.0 6.0 Table RT (min) 8.0 10.0 12.0

0.100

7X Measured vs. Table

Difference (min)

0.050

0.000 _0.050 _0.100 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Table RT (min)

Figure 7.

Difference between scaled HCD table and experimental retention times for 50 compound test set. Y axis is table value minus experimental, X axis is table RT. Top plot is 3X, bottom is 7X.

Figure 8 shows three sets of chromatograms using the HCD database at three different speeds. The sample consists of nine organophosphorus pesticides (identified in the caption to Figure 8) at 50 ppm and a matrix consisting of an equal volume mixture of gasoline, kerosene, and diesel fuel spiked at 50,000 ppm total mixture. The 1X and 3X data were collected on the three-way splitter instrument and the 7X was collected on the DFPD instrument. All nine compounds also contained sulfur as can be seen in the DFPD sulfur chromatogram at the bottom of Figure 8. Note that the sulfur tails somewhat compared to the phosphorus.

14

TIC: OP_50K_GKD1.D\DATA.MS

TIC
1 2 3 4 5 6 7 8 9

1X

P (FPD) ECD
4 8 12 16 20 24 28

TIC: OP_50K_GKD1.D\DATA.MS

TIC P (FPD) ECD


1 2 3 4 5 6 7 8 9

3X

TIC: 50_OP_50K_GKD.D\DATA.MS

TIC

7X

P (FPD) S (FPD)
1 2 3 4

Peak identities

1) 2) 3) 4)

O,O,O-triethyl phosphorothioate Thionazin Sulfotepp Phorate

5) 6) 7) 8) 9)

Dimethoate Disulfoton Methyl parathion Parathion Famphur

Figure 8.

Comparison of 1X, 3X, and 7X chromatograms. 1X and 3X were run on GC/MS/ECD/FPD system, 7X on GC/MS/DFPD.

15

Figure 9 expands the RT region of the phosphorous chromatogram containing peaks 3, 4, and 5 from Figure 8. The decrease in resolution with increasing speed is clearly evident. If only the standard target and three qualifier ion approach is used, the loss in resolution causes a significant problems. With the 1X method, all nine of the analytes are identified and eight false positives are reported. With the 3X method, all analytes are again found but now with 25 false positives. With the significantly decreased resolution of the 7X method, only seven of the nine analytes are identified and 48 false positives are reported.
3

The situation is much different when using the approach described here. Even in the worst situation, the 7X method, AMDIS finds all nine analytes with high-quality matches and only three false positives. The DRS report for the 7X analysis is shown in Table 3. To simplify the table, the 48 false positives that only appear in the quant column are not shown. The analyte compounds are shown in bold. All show close RT and high-quality spectral matches to both the AMDIS target library and to the NIST05 library.

1X

16.50

17.75

3X

5.50

5.916

7X

2.365

2.544

Peak identities

3) Sulfotepp 4) Phorate 5) Dimethoate Figure 9. Comparison of FPD phosphorus chromatograms from 1X, 3X, and 7X runs in Figure 8.

16

Table 3.

DRS Report for 7x Analysis of 50 ppm Pesticides In 50,000 ppm Gasoline/Kerosine/Diesel Matrix Agilent ChemStation amount (ng) 13.92 AMDIS match 71 69 46 64 55 91 88 90 84 92 92 91 93 RT Diff (sec.) 9.5 0.7 7.6 0.6 2.3 0.5 0.5 0.6 0.7 0.6 0.6 0.7 0.8 NIST reverse match 74 71 74 80 85 85 83 85 85 88 82 85 85 Hit number 50 1 1 3 1 1 1 1 1 1 1 1 1

RT 0.973 1.380 1.520 1.520 2.113 2.138 2.138 2.275 2.417 2.427 2.485 2.619 2.748 2.901 3.360

Cas # 98862 126681 94597 52417502 132649 90437 2131411 297972 3689245 298022 60515 298044 298000 56382 52857

Compound name Acetophenone O,O,O-triethyl phosphorothioate Safrole Benzeneacetaldehyde, ,2,5-trimethylDibenzofuran o-Phenylphenol Naphthalene, 1,4,5-trimethylThionazin Sulfotepp Phorate Dimethoate Disulfoton Methyl parathion Parathion (ethyl) Famphur (48 quant-only hits not shown)

0.35

89.2 23.31 27.34 22.7 25.12

The peak at 0.973 minutes is a reasonable spectral match to acetophenone, but the large time difference and being the 50th hit in the NIST search results suggests that this is not the compound. The peak at 1.520 min is a poor spectral match with a large time difference. The absence of a NIST reverse search and hit entry means that the listed compound was not in the top 100 hits in the NIST search. The next compound listed at 1.520 min is the top entry from the NIST search. It is quite clear that safrole is not present. The peak at 2.113 min, dibenzofuran, was not one of the analytes added to the sample. However, it probably is present in the diesel fuel matrix. Its presence is supported by both reasonably good spectral matches and close time matching with a database. The last extraneous peak at 2.138 min is also questionable. The time match is somewhat poor and the NIST reverse search suggests the identification is not correct.

All nine analytes are detected with the FPD on both the phosphorus and sulfur chromatograms. All analytes except peak 1 are detected selectively on the ECD as well. These results suggest that while the loss of resolution in going to 7X is unacceptable when using only conventional screening approaches, with the method discussed here, it is a viable option. By using the DRS report combined with the selective detector data, the number of false positives and false negatives are significantly reduced. For those situations where speed is a critical factor, for example in response to homeland security incidents, the fastest method may be the one of choice. For many laboratories, the 3X method would be an attractive choice. It has higher resolution than the 7X and higher speed than the 1X and still allows the use of two GC detectors in parallel with the MSD. It also only requires a 240V oven, not the repositioning of the MSD to the back position.

17

Conclusions
The systems described here offer several advantages when screening samples for the presence of hazardous chemicals. The advantages derive from a combination of techniques that result in both faster and more accurate screening results. Retention time locked target database of 731 hazardous chemicals for screening with MS Microfluidic splitter - using selective detection simultaneous with MS data for added confirmation, finding non-target suspect compounds, and alternate quantitation SIM/Scan - Acquire SIM data on high priority targets simultaneously with scan data. Saves time by eliminating need to run samples in both modes. DRS - automated deconvolution dramatically increases accuracy of target identification, even in the most challenging matrices. The reduction of data interpretation from hours to minutes is especially useful for response to hazardous chemical incidents. Fast chromatography using shorter columns, faster ovens, and backflushing to greatly reduce run times. This combination of techniques offers a viable solution to the hazardous chemicals challenge.

3. Chin-Kai Meng, Improving Productivity with Synchronous SIM/Scan, Agilent Technologies, publication 5989-3108EN www.agilent.com/chem 4. Philip Wylie, Michael Szelewski, Chin-Kai Meng, Christopher Sandy, Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN www.agilent.com/chem 5. B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie, Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking, Agilent Technologies, publication 5967-5820E www.agilent.com/chem 6. Michael J. Szelewski and Bruce Quimby, New Tools for Rapid Pesticide Analysis in High Matrix Samples, Agilent Technologies, publication 5989-1716EN www.agilent.com/chem 7. Michael J. Szelewski and Bruce D. Quimby, Ambient Headspace GC and GC-MSD Analysis of Nonpolar Volatiles in Water, Agilent Technologies, publication 5968-9455E www.agilent.com/chem

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

References
1. Retention time locking (RTL), Vince Giarrocco, Bruce Quimby, and Matthew Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E www.agilent.com/chem 2. Chin Kai-Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Agilent Technologies, publication 5989-3299EN www.agilent.com/chem

18

Appendix A

Lists of Compounds in Databases


Volatiles: EPA 502/524, 60 compounds
1,1,1,2-Tetrachloroethane 1,1,1-Trichloroethane 1,1,2,2-Tetrachloroethane 1,1,2-Trichloroethane 1,1-Dichloroethane 1,1-Dichloroethylene 1,1-Dichloropropylene 1,2,3-trichlorobenzene 1,2,3-Trichloropropane 1,2,4-trichlorobenzene 1,2,4-trimethylbenzene 1,2-Dibromoethane 1,2-dichlorobenzene 1,2-Dichloroethane 1,2-Dichloropropane 1,3,5-trimethylbenzene 1,3-dichlorobenzene 1,3-Dichloropropane 1,4-dichlorobenzene 2,2-Dichloropropane 2-chlorotoluene 3-Chloro-1,2-dibromopropane 4-chlorotoluene Benzene Bromobenzene Bromochloromethane Bromodichloromethane Bromoform Bromomethane Carbon Tetrachloride Chlorobenzene Chlorodibromomethane Chloroethane Chloroform Chloromethane cis-1,2-Dichloroethylene cis-1,3-Dichloropropylene Dibromomethane Dichlorodifluoromethane Ethylbenzene Hexachlorobutadiene Isopropylbenzene Methylene Chloride m-xylene Naphthalene n-butylbenzene n-propylbenzene o-Xylene p-isopropyltoluene p-xylene Styrene tert-butylbenzene Tetrachloroethylene Toluene trans-1,2-Dichloroethylene trans-1,3-Dichloropropylene Trichloroethylene Trichlorofluoromethane Vinyl chloride 4,4'-DDD 4,4'-DDE 4,4'-DDT 4-aminobiphenyl 4-bromophenyl phenyl ether 4-chloro-3-methylphenol 4-chloroaniline 4-chlorophenyl phenyl ether 4-nitroaniline 4-nitrophenol 4-nitroquinoline-1-oxide 5-nitro-o-toluidine 7,12-dimethylbenz[a]anthracene a,a-dimethylphenethylamine Acenaphthene Acenaphthylene Acetone Acetophenone Aldrin Alpha-BHC (alpha-HCH) Aniline Anthracene Aramite (total) Benz[a]anthracene Benzene Benzo[a]pyrene Benzo[b]fluoranthene Benzo[ghi]perylene Benzo[k]fluoranthene Benzyl alcohol Beta-BHC (beta-HCH) Bis(2-chloroethoxy)methane Bis(2-chloroethyl) ether Bis(2-chloroisopropyl) ether Bis(2-ethylhexyl)phthalate Butyl benzyl phthalate Chlorobenzilate Chrysene Delta-BHC (delta-HCH) Diallate (total) Dibenz[a,h]anthracene Dibenzofuran Dieldrin Diethyl phthalate Dimethoate Dimethyl phthalate Di-n-butyl phthalate

Semivolatiles: EPA 8270C Appendix IX, 140 compounds


1,2,4,5-tetrachlorobenzene 1,2,4-trichlorobenzene 1,2-dichlorobenzene 1,3,5-trinitrobenzene 1,3-dichlorobenzene 1,4-dichlorobenzene 1,4-naphthoquinone 1-naphthylamine 2,3,4,6-tetrachlorophenol 2,4,5-trichlorophenol 2,4,6-trichlorophenol 2,4-dichlorophenol 2,4-dimethylphenol 2,4-dinitrophenol 2,4-dinitrotoluene 2,6-dichlorophenol 2,6-dinitrotoluene 2-acetylaminofluorene 2-chloronaphthalene 2-chlorophenol 2-methyl-4,6-dinitrophenol 2-methylnaphthalene 2-naphthylamine 2-nitroaniline 2-nitrophenol 2-picoline 3,3'-dichlorobenzidine 3,3'-dimethylbenzidine 3-methylcholanthrene 3-nitroaniline

19

Di-n-octyl phthalate Dinoseb Diphenylamine Disulfoton Endosulfan I Endosulfan II Endosulfan sulfate Endrin Endrin aldehyde Ethyl methanesulfonate Famphur Fluoranthene Fluorene Gamma-BHC (lindane) Heptachlor Heptachlor epoxide -isomer B Hexachlorobenzene Hexachlorobutadiene Hexachlorocyclopentadiene Hexachloroethane Hexachlorophene Hexachloropropene Indeno[1,2,3-cd]pyrene Isodrin Isophorone Isosafrole Kepone m-cresol (3-methylphenol) m-dinitrobenzene Methapyrilene Methoxychlor Methyl methanesulfonate Methyl parathion Naphthalene Nitrobenzene N-nitrosodiethylamine N-nitrosodimethylamine N-nitrosodi-n-butylamine N-nitrosodi-n-propylamine N-nitrosodiphenylamine N-nitrosomethylethylamine N-nitrosomorpholine (4-nitrosomorpholine) N-nitrosopiperidine (1-nitrosopiperidine) N-nitrosopyrrolidine (1-nitrosopyrrolidine) O,O,O-triethyl phosphorothioate o-cresol (2-methylphenol) o-toluidine p-(dimethylamino)azobenzene Parathion (ethyl) p-cresol (4-methylphenol) Pentachlorobenzene Pentachloroethane Pentachloronitrobenzene Pentachlorophenol Phenacetin 20

Phenanthrene Phenol Phorate p-phenylenediamine Pronamide Pyrene Pyridine Safrole Sulfotepp Thionazin

Chlorinated Dioxins and Furans: EPA 8282, 19 compounds


2,3,7,8-Tetrachlorodibenzo-p-dioxin 1,2,3,7,8-Pentachlorodibenzo-p-dioxin 1,2,3,4,7,8-Hexachlorodibenzo-p-dioxin 1,2,3,4,6,7,8-Heptachlorodibenzo-p-dioxin Octachlorodibenzo-p-dioxin 2,3,7,8-Tetrachlorodibenzofuran 1,2,3,7,8-Pentachlorodibenzofuran 1,2,3,4,7,8-Hexachlorodibenzofuran 1,2,3,4,6,7,8-Heptachlorodibenzofuran Octachlorodibenzofuran

Polychlorinatedbiphenyls: EPA 8082, 19 compounds


2-chlorobiphenyl 2,3-dichlorobiphenyl 2,2',5-trichlorobiphenyl 2,4',5-trichlorobiphenyl 2,2',5,5'-tetrachlorobiphenyl 2,2',3,5'-tetrachlorobiphenyl 2,3',4,4'-tetrachlorobiphenyl 2,2',4,5,5'-pentachlorobiphenyl 2,2',3,4,5'-pentachlorobiphenyl 2,3,3',4',6-pentachlorobiphenyl 2,2',3,5,5',6-hexachlorobiphenyl 2,2',4,4',5,5'-hexachlorobiphenyl 2,2',3,4,5,5'-hexachlorobiphenyl 2,2',3,4,4',5'-hexachlorobiphenyl 2,2',3,4',5,5',6-heptachlorobiphenyl 2,2',3,4,4',5',6-heptachlorobiphenyl 2,2',3,4,4',5,5'-heptachlorobiphenyl 2,2',3,3',4,4',5-heptachlorobiphenyl 2,2',3,3',4,4',5,5',6-nonachlorobiphenyl

Pesticides: Agilent RTL pesticide database (adapted), 567 compounds


1,2,4-Trichlorobenzene 1,2-Dibromo-3-chloropropane 17a-Ethynylestradiol 2-(1-naphthyl)acetamide 2-(2-Butoxyethoxy)ethyl thiocyanate 2-(Octylthio)ethanol

2,3,4,5-Tetrachlorophenol 2,3,4,6-Tetrachlorophenol 2,3,5,6-Tetrachlorophenol 2,3,5-Trichlorophenol 2,3,5-Trimethacarb 2,3,5-Trimethylphenyl methyl carbamate (Trimethacarb) 2,3,7,8-Tetrachlorodibenzofuran 2,3,7,8-Tetrachlorodibenzo-p-dioxin 2,4,5-T methyl ester 2,4,5-Trichlorophenol 2,4,6-Trichlorophenol 2,4-D methyl ester 2,4-D sec-butyl ester 2,4-DB methyl ester 2,4-Dichlorophenol 2,4-Dichlorophenyl benzenesulfonate 2,4-Dimethylaniline 2,6-Dichlorobenzonitrile 2,6-Dimethylaniline 2-[3-Chlorophenoxy]propionamide 2-Ethyl-1,3-hexanediol 2-Hydroxyestradiol 2-Methylphenol 2-Phenoxypropionic acid 3,4,5-Trimethacarb 3,4-Dichloroaniline 3,5-Dichloroaniline 3-Chloroaniline 3-Hydroxycarbofuran 4,4'-Dichlorobenzophenone 4,6-Dinitro-o-cresol (DNOC) 4-Chloroaniline 4-Methylphenol 5,7-Dihydroxy-4'-methoxyisoflavone 9,10-Anthraquinone Acephate Acetochlor Acifluorfen methyl ester Alachlor Aldrin Allidochlor Ametryn Amidithion Aminocarb Amitraz Ancymidol Anilazine Aniline Atraton Atrazine Azaconazole Azamethiphos Azinphos-ethyl Azinphos-methyl Aziprotryne Azobenzene Barban

Benalaxyl Benazolin-ethyl Bendiocarb Benfluralin Benfuresate Benodanil Bentazone Bentazone methyl derivative Benthiocarb Benzo(a)pyrene Benzophenone Benzoylprop ethyl b-Estradiol BHC alpha isomer BHC beta isomer BHC delta isomer Bifenox Bifenthrin Binapacryl Bioallethrin Bioallethrin S-cyclopentenyl isomer Bioresmethrin Biphenyl Bis(2-ethylhexyl)phthalate Bisphenol A Bitertanol I Bitertanol II Bromacil Bromobutide Bromocyclen Bromophos Bromophos-ethyl Bromopropylate Bromoxynil Bromoxynil octanoic acid ester Buprofezin Butachlor Butamifos Butoxycarboxim Butralin Butyl benzyl phthalate Butylate Butylated hydroxyanisole Captafol Captan Carbaryl Carbetamide Carbofuran Carbofuran-3-keto Carbophenothion Carbosulfan Carboxin Chinomethionat Chloramben methyl ester Chloranocryl Chlorbenside Chlorbromuron

Chlorbufam Chlordecone Chlordimeform Chlorfenethol Chlorfenprop-methyl Chlorfenson Chlorfenvinphos Chlorflurecol-methyl ester Chlormefos Chlornitrofen Chlorobenzilate Chloroneb Chloropropylate Chlorothalonil Chlorotoluron Chlorpropham Chlorpyrifos Chlorpyrifos Methyl Chlorthal-dimethyl Chlorthiamid Chlorthion Chlorthiophos Chlorthiophos sulfone Chlorthiophos sulfoxide Chlozolinate cis-Chlordane Clomazone Coumaphos Crimidine Crotoxyphos Crufomate Cyanazine Cyanofenphos Cyanophos Cycloate Cycluron Cyfluthrin I Cyfluthrin II Cyfluthrin III Cyfluthrin IV Cyhalothrin I (lambda) Cymoxanil Cypermethrin I Cypermethrin II Cypermethrin III Cypermethrin IV Cyprazine Cyprofuram Cyromazine d-(cis-trans)-Phenothrin-I d-(cis-trans)-Phenothrin-II Dazomet Decachlorobiphenyl Deltamethrin Demephion

Demeton-S Demeton-S-methylsulfon Desbromo-bromobutide Desmedipham Desmetryn Dialifos Di-allate I Di-allate II Diamyl phthalate Diazinon Dibrom (naled) Dicamba Dicamba methyl ester Dicapthon Dichlofenthion Dichlofluanid Dichlone Dichlormid Dichlorophen Dichlorprop Dichlorprop methyl ester Dichlorvos Diclobutrazol Diclofop methyl Dicloran Dicrotophos Dicyclohexyl phthalate Dicyclopentadiene Dieldrin Diethatyl ethyl Diethofencarb Diethyl dithiobis(thionoformate) (EXD) Diethyl phthalate Diethylene glycol Diethylstilbestrol Difenoconazol I Difenoconazol II Diflufenican Dimefox Dimethachlor Dimethametryn Dimethipin Dimethoate Dimethylphthalate Dimethylvinphos(z) Dimetilan Di-n-butylphthalate Diniconazole Dinitramine Dinobuton Dinocap I Dinocap II Dinocap III Dinocap IV

21

Dinoseb Dinoseb acetate Dinoseb methyl ether Dinoterb Dinoterb acetate Di-n-propyl phthalate Dioxacarb Dioxathion Dioxydemeton-S-methyl Diphacinone Diphenamid Diphenylamine Dipropetryn Disulfoton Ditalimfos Dithiopyr Diuron Dodemorph I Dodemorph II Drazoxolon Edifenphos Endosulfan (alpha isomer) Endosulfan (beta isomer) Endosulfan ether Endosulfan lactone Endosulfan sulfate Endrin Endrin aldehyde Endrin ketone EPN Epoxiconazole EPTC Erbon Esfenvalerate Esprocarb Etaconazole Ethalfluralin Ethiofencarb Ethiolate Ethion Ethofumesate Ethoprophos Ethoxyquin Ethylenethiourea Etridiazole Etrimfos Famphur Fenarimol Fenazaflor Fenbuconazole Fenchlorphos Fenfuram Fenitrothion Fenobucarb Fenoprop Fenoprop methyl ester Fenoxycarb 22

Fenpropathrin Fenson Fensulfothion Fenthion Fenthion sulfoxide Fenuron Fenvalerate I Fenvalerate II Fepropimorph Flamprop-isopropyl Flamprop-methyl Fluazifop-p-butyl Flubenzimine Fluchloralin Flucythrinate I Flucythrinate II Flumetralin Fluometuron Fluorodifen Fluotrimazole Flurenol-butyl ester Flurenol-methylester Fluridone Flurochloridone I Flurochloridone II Fluroxypyr-1-methylheptyl ester Flusilazole Flutolanil Flutriafol Fluvalinate-tau-I Fluvalinate-tau-II Folpet Fonofos Formothion Fuberidazole Furalaxyl Furathiocarb Furmecyclox Heptachlor Heptachlor epoxide Heptachlor exo-epoxide isomer B Heptenophos Hexabromobenzene Hexachlorobenzene Hexachlorophene Hexaconazole Hexazinone Hexestrol Imazalil Ioxynil Iprobenfos Iprodione Isazophos Isobenzan Isobornyl thiocyanoacetate Isocarbamide Isodrin

Isofenphos Isomethiozin Isoprocarb Isopropalin Isoprothiolane Isoproturon Isoxaben Isoxathion Jodfenphos Kinoprene Lenacil Leptophos Leptophos oxon Lindane Linuron Malathion Malathion-o-analog MCPA methyl ester MCPB methyl ester m-Cresol Mecarbam Mecoprop methyl ester Mefenacet Mefluidide Menazon Mephosfolan Mepronil Metalaxyl Metamitron Metasystox thiol Metazachlor Methacrifos Methamidophos Methfuroxam Methidathion Methiocarb Methiocarb sulfone Methiocarb sulfoxide Methomyl Methoprene I Methoprene II Methoprotryne Methoxychlor Methyl paraoxon Methyl parathion Methyl-1-naphthalene acetate Methyldymron Metobromuron Metolachlor Metolcarb Metribuzin Mevinphos Mirex Molinate Monalide Monocrotophos Monolinuron

Myclobutanil N,N-Diethyl-m-toluamide N-1-Naphthylacetamide Naphthalic anhydride Napropamide Nicotine Nitralin Nitrapyrin Nitrofen Nitrothal-isopropyl N-Methyl-N-1-naphthyl acetamide Norflurazon Nuarimol o,p'-DDD o,p'-DDE o,p'-DDT Octachlorostyrene o-Dichlorobenzene Omethoate o-Phenylphenol Oryzalin Oxabetrinil Oxadiazon Oxadixyl Oxamyl Oxycarboxin Oxychlordane Oxydemeton-methyl Oxyfluorfen p,p'-DDD p,p'-DDE p,p'-DDT Paclobutrazol Paraoxon Parathion p-Dichlorobenzene Pebulate Penconazole Pendimethalin Pentachloroaniline Pentachloroanisole Pentachlorobenzene Pentachloronitrobenzene Pentachlorophenol Pentanochlor Permethrin I Permethrin II Perthane Phenamiphos Phenkapton Phenoxyacetic acid Phenthoate Phorate Phosalone Phosfolan Phosmet Phosphamidon I

Phosphamidon II Phthalide Picloram methyl ester Pindone Piperalin Piperonyl butoxide Piperophos Pirimicarb Pirimiphos-ethyl Pirimiphos-methyl Plifenat p-Nitrotoluene Pretilachlor Probenazole Prochloraz Procymidone Profenofos Profluralin Promecarb Prometon Prometryn Propachlor Propamocarb Propanil Propargite Propazine Propetamphos Propham Propiconazole-I Propiconazole-II Propoxur Propyzamide Prothiofos Prothoate Pyracarbolid Pyrazon Pyrazophos Pyrazoxyfen Pyributicarb Pyridaben Pyridaphenthion Pyridate Pyridinitril Pyrifenox I Pyrifenox II Pyrimethanil Pyroquilon Quinalphos Quinoclamine Quizalofop-ethyl Resmethrin S,S,S-Tributylphosphorotrithioate Sebuthylazine Secbumeton Simazine Simetryn Sulfotep

Sulfur (S8) Sulprofos Swep Tamoxifen TCMTB Tebuconazole Tebutam Tecnazene Temephos Terbacil Terbucarb Terbufos Terbumeton Terbuthylazine Terbutryne Tetrachlorvinphos Tetradifon Tetraethylpyrophosphate (TEPP) Tetramethrin I Tetramethrin II Tetrapropyl thiodiphosphate Tetrasul Thenylchlor Thiabendazole Thiofanox Thiometon Thionazin Tiocarbazil I Tiocarbazil II Tolclofos-methyl Tolylfluanid trans-Chlordane Triadimefon Triadimenol Tri-allate Triamiphos Triazophos Tributyl phosphate Tributyl phosphorotrithioite Trichlorfon Trichloronate Triclopyr methyl ester Tricyclazole Tridiphane Trietazine Triflumizole Trifluralin Tryclopyrbutoxyethyl Tycor (SMY 1500) Uniconizole-P Vamidothion Vernolate Vinclozolin

23

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA February 9, 2006 5989-4834EN

Analysis of Post-Harvest Fungicides and Their Metabolites in Citrus Fruits and Juices by Time-of-Flight and Ion Trap LC/MS Application

Food Safety

Authors
E. Michael Thurman and Imma Ferrer Pesticide Residue Research Group University of Almera, 04120 Almera, Spain Michael Woodman and Jerry Zweigenbaum* Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
*Agilent contact

Introduction
Thiabendazole and imazalil are widely used for post-harvest treatment of citrus fruits (Figure 1) in order to preserve the fruit during the transport process, which may take from several days to several weeks. The fungicide treatment is of the fruit skin only. The maximum residue limits (MRLs) for imazalil and thiabendazole in Europe and the United States require that these compounds be monitored before consumption of the fruit [1, 2]. Since these compounds are used ubiquitously, they are frequently detected in imported citrus fruits [3]. Thus, many commercial samples must be inspected to ensure food safety. Additionally, imazalil is easily metabolized to 1-(2,4dichlorophenyl)-2-(1H-imidazole-1-yl)-1-ethanol, which is sometimes detected in citrus fruits [4]. In the United States, the sum of imazalil and its metabolite is regulated, so a method for the parent compound should also include its metabolite [5]. While papers have reported analytical LC/MS procedures for imazalil and thiabendazole, less is known about the metabolite of imazalil, and there are, as yet, no published reports using LC/MS accurate-mass analysis [6]. This application note describes a rapid and simultaneous method using electrospray liquid chromatography/time-of-flightmass spectrometry and liquid chromatography/ion trap mass spectrometry (LC/TOFMS and LC/ITMS) for the analysis of these two important postharvest fungicides in citrus. An earlier application note reported on the analysis of carbendazim and thiabendazole in fruits and vegetables and may be of interest as well [7].

Abstract
This application note applies both time-of-flight and ion trap liquid chromatography/mass spectrometry (LC/TOFMS and LC/ITMS) to determine two important post-harvest fungicides (imazalil and thiabendazole) and a metabolite of imazalil in oranges and in their juice. Included is both a detailed rapid procedure for sample preparation of citrus fruits and a solid-phase extraction for the isolation of these fungicides and metabolite from juice. Identification of these fungicides and metabolite is carried out using both accurate mass for elemental composition and fragmentation pathways with LC/ITMS to help elucidate pathways of degradation. The results of the analysis of actual samples from the marketplace for fruit and juice are included.

1 ppm

0 ppm

5. Shake the sample vigorously for 1 minute by using a Vortex mixer at maximum speed or hand shaking. 6. Centrifuge for 3 minutes at 3700 rpm.

Treated

Untreated

7. Add 5 mL of supernate into a 15-mL tube. 8. Add 250 mg of PSA adsorbent and 750 mg of MgSO4. 9. Vortex and shake for 20 s.

Cl O H2C N NH

10. Centrifuge again for 3 minutes at 3700 rpm.


HO Cl

11. Transfer 1.0 mL into a LC/MS vial. 12. Evaporate the 1.0 mL supernate to dryness and bring back up in 8%/92% methanol/water for LC/TOFMS analysis and ion trap analysis. LC/TOFMS analysis of fruit and vegetable extracts is performed by injecting 50 L. Nonfortified samples are analyzed directly at this same point by either LC/TOFMS or LC/ITMS. Fruit-Juice Extraction

Cl
m/z 297

NH

Cl

m/z 257

Imazalil
H2 N N N

Imazalil metabolite

m/z 202

Thiabendazole Figure 1. Structure of imazalil, thiabendazole, and an imazalil metabolite, which are post-harvest fungicides used on citrus (oranges and lemons), showing the [M+H]+ ion. Top panel compares treated and untreated oranges.

Solid-phase extraction (SPE) is used for the extraction of post-harvest fungicides from fruit juices and sodas based on juices. The cartridge of choice is the Accubond C-18 SPE cartridge (part number 188-1356). 1. Prepare the 5-mL SPE cartridge by taking 5 mL of methanol to wet the cartridge followed by 5 mL of deionized water to remove the methanol for good recovery of the fungicides. 2. Allow juice to stand, equilibrate to room temperature and degas prior to SPE, especially if bubbles are present. 3. Centrifuge and filter (0.2-micron TFPE filter to remove remaining solids) 5 mL of juice prior to dilution with 5 mL of deionized water. 4. Pass the juice through the cartridge and rinse with 2 mL of deionized water. 5. Elute with 5 mL of methanol. 6. Evaporate the methanol to 200 L under nitrogen. 7. Dilute 1:3 with deionized water before injection into the LC/TOFMS or LC/ITMS.

Experimental Methods
Fruit Extraction: QuEChERS QuEChERS (quick, easy, cheap, effective, rugged, and safe) is a method that is receiving wide acceptance for rapid extraction of pesticides in food [8]. 1. Weigh 15 g of a previously homogenized sample into a 40-mL Teflon centrifuge tube. 2. Add 15 mL of acetonitrile (containing 1% acetic acid). 3. Add 6 g of anhydrous MgSO4. 4. Add 2.5 g of NaAc3H2O (sodium acetate trihydrate).

LC/TOFMS Method
LC Pumps Column Mobile Phases Gradient Instrument Nebulizer Drying gas Voltages Reference masses Resolution Mass range Flow rate Agilent 1100 binary pumps, injection volume 50 L with standard Agilent 1100 ALS ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, product number 993967-906 A = acetonitrile and B = 0.1% formic acid in water 5 minutes isocratic at 10% A, followed by a linear gradient to 100% A in 30 minutes at a flow rate of 0.6 mL/min Agilent LC/MSD TOF time-of-flight mass spectrometer with dual electrospray source, positive ESI, capillary 4000 V 40 psig 9 L/min, 300 C Fragmentor 190 V, skimmer 60 V, Oct DC1 37.5 V, OCT RF V 250 V m/z 121.0509 and 922.0098 9500500 at m/z 922.0098 501000 m/z Reference A Sprayer 2 is at a constant flow rate during the run

LC/ITMS Method
Chromatography LC Pumps Column Mobile phases Gradient Instrument Detection Identical to LC/MSD TOF methods for direct comparison of peaks Agilent 1100, injection volume 50 L ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, product number 993967-906 A = acetonitrile and B = 0.1% formic acid in water 5 minutes isocratic at 10% A, followed by a linear gradient to 100% A in 30 minutes at a flow rate of 0.6 mL/min Agilent 1100 Series LC/MSD Trap. Positive ESI, Capillary 3200 V, Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C. Capillary exit 70 V, Skimmer 60 V Trap accumulation, 50,000 counts with maximum accumulation time of 200 ms

Results and Discussion


Imazalil in Oranges Figure 1 shows two oranges, one with a treatment of the post-harvest fungicides (imazalil and thiabendazole) and the other orange without treatment. Note the intense growth of mold when no treatment of post-harvest fungicides is used. The purpose of the post-harvest fungicides is to prevent serious mold formation during transport and sale of citrus fruits. This treatment is directed at the orange peel itself, rather than the entire fruit; therefore, it poses a lesser risk to the consumer. Figure 2 shows the LC separation and chromatogram of the orange-peel extract that contains two large peaks, which were suspected to be the post-harvest fungicides, imazalil and thiabendazole. They were identified using the following

technique. The accurate mass spectrum of each peak was taken and A+2 isotopes were examined (Figures 2 and 3). For example, the peak at 17.9 minutes was suspected to be a chlorinecontaining species based on the mass spectrum (Figure 3). In fact, the presence and number of chlorine atoms in the compound can be easily attained taking into account both the relative intensity of the 37Cl/35Cl signals and the accurate mass differences between the two masses. As can be seen in Figure 3, the accurate mass of the m/z 297 peak was 297.0556 with a 37Cl isotope signal of 299.0527, with a relative intensity of about two-thirds of the main peak. The mass difference between both signals is 1.9971, which is very near to the exact mass difference between 35Cl (34.9689) and 37Cl (36.9659) (1.9970). This evidence combined with peak area shows that the unknown compound unequivocally contains chlorine atoms.
3

3.9e7 3.8e7 3.6e7 3.4e7 3.2e7 3.0e7 2.8e7 2.6e7 2.4e7 Intensity, cps 2.2e7 2.0e7 1.8e7 1.6e7 1.4e7 1.2e7 1.0e7 8.0e6 6.0e6 4.0e6 2.0e6 0.0 2 4 6 8 10 12 14 2.1 2.4 2.7 7.6 16.7

17.9

Imazalil

Thiabendazole Imazalil degradate


14.9 18.2 21.0 20.4 21.5 24.8 26.0 26.4 22.7 23.9 25.4

16 18 Time, min

20

22

24

26

28

30

Figure 2.

LC/MSD TOF total ion chromatogram (TIC) of orange peel, note the two large peaks in the chromatogram are imazalil and thiabendazole.

Furthermore, the relative abundance of the isotopic signal indicates that the chlorine isotope is present with two atoms (the natural distribution of 35 Cl: 37Cl is nearly 3:1 (75.77% 35Cl; 24.23 % 37Cl), and two chlorine atoms give the characteristic pattern shown in Figure 3. This fact makes much easier the assignment of an elemental composition for the suspected species. Using the calculator tool of the TOF software, we set, as a stringent criterion, the number of chlorine atoms to two. Using an accuracy error threshold of 3 ppm (the maximum error of the LC/MSD TOF), only one formula

was found (C14H15N2OCl2, 0.3 ppm error). This formula was searched against the The Merck Index database. The use of the Merck Index is explained in an application note [9]. It has to be taken into consideration that the additional hydrogen atom present in the measured ions due to the positive ionization mode must be subtracted from the elemental compositions provided by the calculator tool before entering the formula into the database. The Merck Index gave a unique match with the formula for imazalil.

297.0556

A (Cl35)

Imazalil

A+2 (Cl35 and Cl37)


299.0527

Intensity

Accurate mass Accurate mass spectrum spectrum

299.0501

A+4 (Cl37 and Cl37)


301.0498

288

290

292

294

296

298

300 302 m/z, amu

304

306

308

310

312

314

Figure 3.

Imazalil detection in orange peel measured accurate mass, m/z = 297.0556 theoretical mass = 297.0556, accuracy 0.3 ppm, formula is C14H15N2OCl2.

Likewise the mass spectrum was taken for the peak at 7.6 minutes in Figure 2 and the spectrum is shown in Figure 4, m/z 202.0434. The relative abundance of the isotopic signal indicates that sulfur may be present with one atom. The relative intensity of the A+2 peak is 4.2% with a mass depletion of 1.996 daltons when going from 32S (31.9721) to 34S (33.9679). The mass spectrum shown in Figure 4 has a mass difference of 1.997 daltons and a peak height of 4.7%, which is

quite close to the expected values. Thus, these data indicate the presence of 1 sulfur atom. Entering a mass of m/z 202.0434 gives the unique formula of C10H8N3S, which is the formula for a protonated ion of thiabendazole, based on the Merck Index search. The next step is to examine the fragmentation of the molecules using the LC/MS ion trap for further structural confirmation.

1.9e5 1.8e5 1.7e5 1.6e5 1.5e5 1.4e5 Intensity, counts 1.3e5 1.2e5 1.1e5 1.0e5 9.0e4 8.0e4 7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 186 188 190 192 194 196 198 200 202 m/z, amu 203.0458 204.0406 204 206 208 210 212 214 216 202.0434

Figure 4.

LC/MSD TOF mass spectrum of thiabendazole.

Fragmentation of Imazalil The next step in the discovery process was to search for characteristic fragment ions of the pesticide to confirm (or refute) its identity as imazalil. The fragmentation pathway of the large peak at 17.9 minutes was determined with LC/ITMS at MS3. The pathway is examined to see if reasonable structures could be drawn for the imazalil fragments based on the detection by LC/TOFMS. Figure 5 shows the structure for the proposed ions from the fragmentation of the unknown peak. The m/z 255 ion results from the loss of 42. MS/MS of the m/z 255 ion gives the m/z 187 and 159 ions, whose proposed structures are shown in Figure 5.

The m/z 159 ion is proposed as a doubly chlorinated tropylium ion. All the structures are consistent with the proposed identification as imazalil. Next is to examine the accurate mass spectrum from the LC/TOFMS. Here we found two fragment ions (m/z 255 and 159) with a relative abundance of respectively, 5% and 10% of that of the protonated molecule. The accurate mass of fragment 1 was 255.0084 (Table 1), and a 37Cl signal m/z of 257.0055 (data not shown). From both the relative intensity of these signals and the difference between the two masses, it can be deduced that the two chlorine atoms present in imazalil remain in this fragment.

5 6 4 2 0 100 200 300

297.0

Cl O H2C N
400 500 m/z 600 700 800 900

Cl Neutral loss of 42 H2C HO

NH

Cl

NH

Cl
m/z 255 at MS2

5000 4000 Intensity 3000 2000 1000 109.0 0 100 200 200.8 158.7

254.8

Imazalil, EE ion

EE ion fragment

MS/MS (297 255, 201,159)

Neutral loss of 68

NH

Cl
300 400 500 m/z 600 700 800 900

Cl _CO of 28

+
1.0 0.8 Intensity x104 0.6 0.4 0.2 0.0 100 200 300 400 500 m/z 600 700 800 900 158.7
m/z 159 at MS4

Double Cl trophilium Cl

Cl

m/z 187 at MS3

MS3 (255 187,159)


186.7

EE ion fragment

EE ion fragment

82.0

219.8

Figure 5.

Ion trap MS2 and MS3 to show fragmentation pathway of imazalil.

As can be seen in Table 1, the accurate mass of this fragment gave also three possible elemental compositions in the calculator tool. The first formula (C11H9N2OCl2), (0.8 ppm error) fitted with the proposed structure, involving a loss of C3H6 (accurate mass loss of 42.0469 versus 42.0465 u) in relation with the proposed parent species. Moreover, the accurate mass spectrum (relative intensity and mass differences) evidenced the presence of two chlorine atoms on fragment 2. The measured mass (m/z 158.9762) gave a unique elemental composition (C7H5Cl2), that corresponds to the formation of a doubly chlorinated tropylium fragment-ion. At this point, we combined the fragmentation information from the LC/MSD Trap, as shown in Figure 5 for the proposed structure of

imazalil, with the accurate mass information, that all show a perfect match for the detection of imazalil. These two lines of evidence (LC/TOFMS and LC/ITMS) provide fragment ions and information to confirm the identity of the proposed species based on fragmentation of the parent structure as imazalil, without the use of standards! Likewise, thiabendazole was examined by ion trap and gives a m/z 175 fragment ion, that fits with the structure of thiabendazole. The data are not shown here but were shown in a previous application note for thiabendazole [9]. We verified these determinations by standards for final confirmation.

Table 1.

LC/MSD TOF Fragment Ions of Imazalil Elemental compositions C14H15N2OCl2 C3H15N8O4Cl2 C11H19N2OSCl2 m/z calculated 297.0556 297.0588 297.0590 255.0086 255.0073 255.0060 158.9763 Error, ppm <0.1 8.3 9.0 0.8 4.3 9.4 0.6 Comments Imazalil Fragment 1 Fragment 2

m/z experimental 297.0556

255.0084

C11H9N2OCl2 C9H7N5Cl2 C8H11NO4Cl2

158.9762

C7H5Cl2

Imazalil Metabolite In the orange extract, we found a peak (at a retention time of 14.9 minutes, see Figure 2) with an ion with the same isotopic pattern as imazalil (two chlorine atoms). Taking into account that it had the same number of chlorine atoms, and it also appeared before imazalil, we considered that it could be related to imazalil, perhaps a metabolite. The accurate mass of the ion was 257.0245 with a 37 Cl signal of 259.0216 (see Figure 6). It gave a unique elemental composition in the calculator tool: C11H10N2OCl2 (0.8 ppm error, also see Table 2). For confirmation purposes, we searched for additional fragments but we did not find any characteristic fragment ion of that compound by TOF LC/TOFMS. The obtained elemental composition was then searched against two databases (The Merck Index and ChemIndex) with no positive

results. Then, we search the elemental composition against the Sigma-Aldrich commercial electronic catalogue, and we found a unique match: 1-(2,4dichloro-phenyl)-2-imidazol-1-yl-ethanol. The structure of this compound, shown in Figure 1, is compatible with the degradation of imazalil (see Figure 5 the m/z 255 ion). This suggests that the degradation product is the neutral species corresponding to the degradation of imazalil at the same site that cleaves to yield the m/z 255 fragment. A literature search on imazalils degradation products produced data and reports that agreed with our results (Imazalil-metabolite, R14832) [4]. In fact, in the United States, the sum of imazalil and imazalil metabolite is regulated, so the survey of residual imazalil metabolite is also required [4]. Finally, we confirmed the identity of the degradate by standard matching.

Table 2.

Accurate Mass and Elemental Composition of Imazalil, Imazalil Degradate, and Thiabendazole in Orange Peel Extracted with Methanol Selected ion [M+H]+ [M+H]+ [M+H]+ m/z experimental 297.0556 257.0245 202.0434 m/z calculated 297.0556 257.0243 202.0434 Error mDa +0.1 +0.2 +0.2

Compound Imazalil Imazalil Degradate Thiabendazole

Formula C14H14Cl2N2O C11H10Cl2N2O C10H7N3S

ppm 0.3 0.8 <0.1

2.4e5 2.3e5 2.2e5 2.1e5 2.0e5 1.9e5 1.8e5 1.7e5 1.6e5 1.5e5 Intensity, counts 1.4e5 1.3e5 1.2e5 1.1e5 1.0e5 9.0e4 8.0e4 7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 0.0 235 240 245 250

257.0245

C11H11N2OCl2
HO N N Cl
1 (2,4 dichloro-phenyl)-2-imidazol-1-YL ethanol (Imazalil degradate)

[M+H]+ Protonated imazalil degradate Cl

259.0216

261.1115

255

260 m/z, amu

265

270

275

280

285

Figure 6.

LC/MSD TOF spectrum of imazalil degradate.

Identification of Imazalil in Orange Juice Figures 7 and 8 show the LC/TOFMS chromatogram (see peak at 18.0 minutes) and identification of imazalil (m/z 297.0556) in a juice product made from oranges. Several samples were analyzed and only this sample contained imazalil. No identifications of thiabendazole were observed. It is assumed that the production of juice usually does not involve the squeezing of the peel. However, in some types of juice machines the entire orange is squeezed without removing the peel. In this case, the opportunity for including fungicides in the juice may occur. Note that the LC/TOFMS analysis gave good results in spite of the complex matrix and complex chromatogram of the juice extract.

3.7e7 3.6e7 3.4e7 3.2e7 3.0e7 2.8e7 2.6e7 2.4e7 Intensity, cps 2.2e7 2.0e7 1.8e7 1.6e7 1.4e7 1.2e7 1.0e7 8.0e6 6.0e6 4.0e6 2.0e6 0.0 2 2.1 3.0

3.8

25.4 3.4 16.7 12.2 17.5 24.3 23.4 6.2 11.5 4.9 14.4 14.8 15.3 13.8 21.1 23.1 18.0 19.5 20.3 21.5 22.1 28.5 27.9 26.1

24.7

10

12

14

16 Time, min

18

20

22

24

26

28

30

Figure 7.

LC/MSD TOF TIC of orange juice extract.

9.2e4 9.0e4 8.5e4 8.0e4 7.5e4 7.0e4 6.5e4 6.0e4 Intensity, counts 5.5e4 5.0e4 4.5e4 4.0e4 3.5e4 3.0e4 2.5e4 2.0e4 1.5e4 1.0e4 5000.0 0.0 260 265 270 275 280 285 290

297.0556

299.0528

298.0585

295 m/z, amu

300

305

310

315

320

325

330

Figure 8.

LC/MSD TOF mass spectrum of imazalil in orange juice made from peel.

10

Conclusions
LC/TOFMS analysis is a powerful tool for identification of fungicides in fruits and juices and their metabolites and is a new tool for environmental food chemistry. Elemental composition and structural fragmentation of standards and fragment ions are possible, especially with ion trap MS/MS confirmation. The identification of metabolites is possible, in the case of imazalil, without standards, when both LC/TOFMS and LC/ITMS are used. These concentrations are equal to or lower than the EU directives for controlled fungicides in fruits and juices.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

References
1. EU Food Directives 2002, 91/414/EEC. 2. Food Quality Protection Act, 1998 (FQPA). 3. M. Fernandez, R. Rodriguez, Y. Pico, J. Manes, (2001) J. Chromatogr. A 912, 301310. 4. N. Yoshioka, Y. Akiyama, K. Teranishi, (2004) J. Chromatogr. A, 1022, 145150. 5. http://www.epa.gov./pesticides/registration/ imazalil/ImazProductChem.pdf. 6. E. M. Thurman, Imma Ferrer, Jerry A. Zweigenbaum, Juan F. Garca-Reyes, Michael Woodman, and Amadeo R. Fernndez-Alba, (2005) Discovering Metabolites of Post-Harvest Fungicides in Citrus with LC/MS TOF and Ion Trap MS/MS, J. Chromatogr., in press. 7. Determination of Fungicides in Fruits and Vegetables by LC/MS/ESI/TOF and LC/MS Ion Trap, Agilent Technologies, publication 5989-2209EN www.agilent.com/chem 8. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, and F.J. Schenck, (2003) J. AOAC Internat., 86, 412431. 9. Identification of Unknown Pesticides in Food Using Both LC/MSD TOF and Ion Trap MSn, Agilent Technologies, publication 5989-1924EN www.agilent.com/chem

11

www.agilent.com/chem

Acknowledgments We acknowledge Amadeo Fernandez-Alba and Juan-Francisco Garcia-Reyes, Department of Hydrogeology and Analytical Chemistry, University of Almera, for help, ideas, and analytical support. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. ZORBAX is a registered trademark of E.I. duPont de Nemours Co. Inc. Agilent Technologies, Inc. 2005 Printed in the USA August 24, 2005 5989-2728EN

Analysis of Terbuthylazine in Olive Oil by LC/Ion Trap MS and LC/Time-of-Flight MS Application

Food safety

Authors
Imma Ferrer and E. Michael Thurman University of Almera, 04120 Almera, Spain

Introduction
Olive oil is one of the most-used food products in Mediterranean countries. The positive effects of olive oil on health have prompted a demand for this product world-wide. Virgin olive oil is obtained from the fruit of the olive tree (Olea Europaea) exclusively by mechanical and physical processes without any further treatment, which may alter the olive oil quality. The most extensively applied agrochemicals in olive plantations of Mediterranean countries (Greece, Spain, and Italy) are herbicides and insecticides. Although herbicides are mainly applied to soils, some residues can persist to the harvest stage, thus contaminating the olives picked up from soil. This can result in the presence of trace amounts of these pesticides in olive oil. Consequently, both the European Union and the Codex Alimentarius Commission of the Food and Agriculture Organization (FAO) of the United Nations have established maximum pesticide residue levels in olives and olive oil. Currently, various olive oil pesticide residue regulatory programs are being carried out to update and to establish new and more stringent regulations concerning the maximum residue levels in these commodities [1]. This fact has fostered the development of more powerful analytical tools in order to provide enough sensitivity and selectivity to meet these requirements in food samples such as edible oils, which have a complex matrix due to the high fat content of the extracts obtained after the sample treatment step.

Abstract
This application note describes the use of liquid chromatography/ion trap mass spectrometry (LC/ITMS) and liquid chromatography/time-of-flight mass spectrometry (LC/TOFMS) for the identification and quantitation of terbuthylazine in olive oil samples. The method includes a sample treatment step based on a preliminary liquidliquid extraction, followed by matrix solid-phase dispersion (MSPD) using an aminopropyl-bonded silica as a sorbent material. A final clean-up step is performed with Florisil using acetonitrile as an eluting solvent. The analysis by ion trap was achieved in MS/MS mode, monitoring the characteristic fragment ion at m/z 174. The identification by LC/TOFMS was accomplished with the accurate mass (and the subsequently generated empirical formula) of the protonated molecule ([M+H]+ m/z 230), along with the accurate mass of the fragment ion and the characteristic chlorine isotope cluster present in terbuthylazine. The accuracy typically obtained was routinely better than 2 ppm. The method sensitivity, linearity, precision, accuracy, matrix effects, and limit of detection were also studied; they supported the potential of LC/TOFMS and LC/ITMS for the routine quantitative analyses of terbuthylazine in olive oil.

Many multiresidue procedures employing different clean-up techniques and a variety of detection methods are reported on the determination of pesticide residues in olive oil. The most commonly used methodology is based on GC after a comprehensive clean-up step, in most cases based on liquid-liquid partitioning or gel permeation chromatography to separate the low molecular mass pesticides from the higher molecular mass fat constituents of the oil, such as triglycerides [2, 3]. The preparation of oil samples for the determination of pesticides by GC requires the complete removal of the high-molecular-mass fat from the sample to maintain the chromatographic system in working order. Recently, a multiresidue method for the determination of triazines and organophosphorous pesticides using MSPD followed by GC/MS and ion trap MS techniques was reported [4]. In this work, we further explore the capabilities of LC/ITMS and LC/TOFMS techniques for the identification of terbuthylazine, one of the most common pesticides found in olive oil.

- Another 10 mL of acetonitrile saturated with petroleum ether was added to the petroleum ether extract. The mixture was shaken again for 3 minutes and the acetonitrile phase was collected and added to the previous one. - Finally, a 7-mL aliquot of the acetonitrile extract was transferred to a 10-mL glass test tube. The extract was then carefully evaporated down to a final volume of about 2 mL. This remaining extract was transferred to a glass mortar to be subjected to matrix solid-phase dispersion. 2. The MSPD: - The extract obtained in the liquid-liquid extraction step was homogenized with 2 g of aminopropyl-bonded sorbent (Bondesil-NH2, 40-m particle size, Varian Inc.) until a fine powder was obtained. - The mixture was transferred to a commercially available minicolumn containing 2 g of Florisil (12-mL Bond-Elut Varian minicolumn, Varian Inc.). This minicolumn was connected to a vacuum system for solid phase extraction adjusting the flow to 3 mL/min. - An elution step was carried out with 2 5 mL of acetonitrile. The final extract was evaporated until near dryness, then dissolved with 1:1 acetonitrile:water. - Prior to LC/ITMS and LC/TOFMS analysis, the extract was filtered through a 0.45-m PTFE filter (Millex FG, Millipore, Milford, MA, USA). Using the MSPD method, recoveries for terbuthylazine from olive oil samples were higher than 96% with a 6% relative standard deviation (RSD) (n = 5).

Experimental Methods
Olive Oil Extraction MSPD was used for the extraction of terbuthylazine from olive oil after a preliminary liquidliquid extraction with organic solvents. 1. The liquid-liquid extraction: - An aliquot of approximately 5 g (ca. 5.5 mL) of olive oil sample was mixed with 15 mL of petroleum ether, saturated with acetonitrile, in a 100-mL separatory funnel, in which a two-step liquid-liquid extraction was performed. - A solution of 25 mL of acetonitrile, saturated with petroleum ether, was added to the olive oil mix obtained previously. The funnel was shaken vigorously for 3 minutes, and the acetonitrile phase was separated from the petroleum ether phase.

Agilent 1100 Series LC/MSD TOF Methods


LC conditions LC Pumps were Agilent 1100 binary pumps Injection volume: 50 L with standard Agilent 1100 ALS Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, p/n 993967-906 Mobile phases: A = acetonitrile and B = 0.1% formic acid in water Gradient: 5 minutes isocratic at 10% A, followed by a linear gradient to 100% A in 25 minutes at a flow rate of 0.6 mL/min MS conditions LC/MSD TOF Source: Capillary: Nebulizer: Drying gas: Gas temp: Fragmentor: Skimmer: Oct DC1: Oct RF V: Reference masses: Resolution: Mass range: Reference A sprayer 2:

Positive ESI (electrospray ionization) 4000 V 40 psig 9 L/min 300 C 190 V 60 V 37.5 V 250 V m/z 121.0509 and 922.0098 9500 500 @ m/z 922.0098 501000 m/z Constant flow during the run

Agilent 1100 Series LC/MSD Trap Methods LC/MSD Trap Methods identical to LC/MSD TOF for direct comparison of peaks. LC Pumps: HP 1100 Inject vol: 50 L Column: ZORBAX Eclipse XDB-C-8, 4.6 mm 150 mm, 5 m, p/n 993967-906 Mobile phases: A = acetonitrile (ACN) B = 0.1% formic acid in water Gradient: 10% A, isocratic, for 5 minutes followed by linear gradient to 100% A in 25 min at a flow rate of 0.6 mL/min LC/MSD Trap Positive ESI: Nebulizer: Drying gas: Capillary exit: Skimmer: Trap accumulation: Capillary 3200 V 40 psig 9 L/min, gas temp 300 C 70 V 60 V 50,000 counts with maximum accumulation time of 200 ms

Results and Discussion


Identification of Terbuthylazine by LC/ITMS and LC/TOFMS Olive oil is one of the most difficult food matrices due to the presence of numerous interferences that show up in full-scan mode. For this reason, the ion trap method was optimized in MS/MS mode by isolating the precursor ion at m/z 230, which corresponds to [M+H]+. An isolation mass window of m/z 2 and optimized fragmentation amplitude was used in order to enhance the signal-to-noise (S/N) ratio. An amplitude voltage of 0.8 V was used to

obtain good fragmentation for terbuthylazine, which fragments by losing the terbutyl group (56) forming the m/z 174 fragment ion. Figure 1 shows the analysis of an olive oil sample (spiking level 0.025 mg/kg) by ion trap MS/MS. The extracted ion chromatogram (EIC) for m/z 174, the main fragment of terbuthylazine, as well as its mass spectrum is shown. The fragmentation occurs at the CN bond via the cleavage of the terbutyl group. This represents a characteristic fragmentation for this analyte, allowing for the structural confirmation of terbuthylazine in a relative complex matrix such as olive oil.

1.25

1.00 Intensity 105

0.75

0.50

0.25

0.00 0 5 10 15 20 25 Time [min]

MSPAC25.D: EIC 174 All 1.50

173.8
1.25 Cl Intensity 105 1.00 0.75 0.50 0.25 0.00 100 N N H N N N H _56

+MS2(230.0), 23.7 min #556

210.8
200 300 400 500 600 700 800 900 m/z

Figure 1.

Ion Trap MS/MS chromatogram corresponding to the analysis of a spiked olive oil sample with terbuthylazine (0.025 mg/kg). The EIC for m/z 174 and its corresponding spectrum are shown.

LC/TOFMS analyses were optimized in terms of fragmentation. The in-source collisionally induced dissociation (CID) fragmentation is greatly enhanced at high fragmentor voltages. This provides highly valuable structural information since the accurate mass of the characteristic fragment ion can be used along with that of the molecular ion for confirmation purposes. The relative abundances for both the main fragment and the protonated molecule of terbuthylazine are summarized in Table 1 at three different voltages: 160 V (low),
Table 1. m/z ion 230 [M+H]+ 174 [M+HC4H8]+ Effect of the Fragmentor Voltage on CID Fragmentation for LC/TOFMS Relative abundance 160 V 190 V 230 V 100 5 100 30 20 100

190 V (medium), and 230 V (high). In order to obtain sufficient sensitivity for quantitative purposes (using the protonated molecule) and additional qualitative spectral information provided by the fragment ions generated by in-source fragmentation, a value of 190 V was chosen for further analyses. The accurate masses for both the main fragment and the protonated molecule of terbuthylazine in a spiked olive oil matrix are shown in Table 2.

Table 2.

LC/TOFMS Accurate Mass Measurements for Terbuthylazine and its Main Fragment Ion in Olive Oil Matrix-Matched Standard (Fragmentor Voltage 190 V). Spiking Level: 0.025 mg/Kg Ion [M+H]+ Theoretical mass 230.1167 232.1137 174.0541 176.0511 Measured mass 230.1168 232.1134 174.0542 176.0511 Error mDa 0.1 0.3 0.1 0.05 ppm 0.4 1.5 0.6 0.3

Elemental composition C9H16N535Cl C9H16N537Cl C5H9N535Cl C5H9N537Cl

[M+HC4H8]+

Along with the accurate mass of the protonated molecule and the information provided by the fragments obtained with an optimized in-source fragmentation, terbuthylazine presents another feature which enables its identification; the presence of a chlorine atom. The accurate mass pattern of the 37 Cl isotope signal confirms that the peak unequivocally contains only one chlorine atom (Figure 2). Therefore, the accurate mass of each ion as well as the presence of the chlorine signal, together with the characteristic retention time represent sufficient information to unequivocally identify and confirm this specie in such complex matrices.

230.1168 9.50e4 8.50e4 7.50e4 6.50e4 5.50e4 4.50e4 3.50e4 2.50e4 1.50e4 5000.00

Intensity, counts

174.0542 176.0511

232.1134

60 80 100 120 140 160 180 200 220 240 260 280 m/z, amu

Figure 2.

LC/TOFMS total ion chromatogram (TIC) corresponding to the analysis of a spiked olive oil sample with terbuthylazine (0.025 mg/kg). The EIC for m/z 230 and its corresponding spectrum are also shown.

Analytical Performance The analytical performance of the proposed methods was studied in order to explore the ability to carry out quantitative analyses of terbuthylazine in these complex matrices with a high content of fat. The calibration was carried out using matrixmatched standards prepared by the extraction method based on MSPD described in the experimental section. Linearity was evaluated by analyzing solutions of matrix-matched standards at seven different concentration levels in the range 0.0050.5 mg/kg. Using ion trap the quantitation was performed in MS/MS mode of the m/z 230 ion. Using LC/TOFMS the quantitation was carried out using the peak area from the EIC of the protonated molecule with a mass window of 0.1 Da. As an example, Figure 3 shows the linear calibration curve obtained by LC/TOFMS for terbuthylazine in an olive oil matrix compared to the curve obtained in pure solvent.

The limits of detection (LOD) were estimated from the injection of matrix-matched standard solutions with concentration levels giving a S/N ratio of about 3. The results obtained are shown in Table 3. The LOD obtained are remarkable since they are far below the maximum residue level regulations established for these pesticides. In this sense, LC/TOFMS analyses benefits of the use of narrow mass windows for quantitation purposes, which results in enhanced S/N ratio, thus providing lower detection limits. This fact illustrates the analytical potential of the proposed method based on MSPD and LC/TOFMS for the analyses of pesticides in complex matrices with high content of fat.

40

Pure solvent
y = 70.697x + 0.5666 R2 = 0.9977
30

Spiked oil solvent


y = 60.442x + 0.7386 R2 = 0.9977

Area 10 E+06

20

10

0 0.0 0.1 0.2 0.3 Concentration (mg/kg) 0.4 0.5 0.6

Figure 3.

Calibration plot obtained from spiked olive oil samples versus solvent based samples by LC/TOFMS.

Table 3.

Analytical Parameters for the Analysis of Terbuthylazine in Olive Oil Samples by Ion trap MS/MS and LC/TOFMS Concentration Linearity LOD RSD (%) Method range (mg/kg) (R2) (g/kg) n = 5 LC/ITMS LC/TOFMS 0.0050.5 0.0050.5 0.991 0.995 0.2 1 5.5 2

overcome using the isotopic chlorine signal or a characteristic fragment not affected by other interfering species.

Conclusion
LC/ITMS and LC/TOFMS can be used for the identification and quantitation of terbuthylazine in olive oil samples after performing several preliminary sample treatments. The analysis by ion trap was achieved in MS/MS mode, monitoring the characteristic fragment ion at m/z 174. The identification by LC/TOFMS was accomplished with the accurate mass (and the subsequently generated empirical formula) of the protonated molecule ([M+H]+ m/z 230), along with the accurate mass of the fragment ion and the characteristic chlorine isotope cluster present in terbuthylazine. The accuracy typically obtained was routinely better than 2 ppm. The method sensitivity, linearity, precision, accuracy, matrix effects, and LOD were also studied and they supported the potential of LC/TOFMS and LC/ITMS for the routine quantitative analyses of terbuthylazine in olive oil.

Analysis of Olive Oil Samples To evaluate the effectiveness of the proposed method, it was applied to the analysis of olive oil samples. An example is shown in Figure 4 where traces of terbuthylazine were found in an olive oil extract; the EIC for m/z 230 is also shown. This is a real example illustrating the usefulness of the routine accurate mass measurement capabilities of LC/TOFMS, especially when analyzing traces of pesticides in complex samples such as olive oil. In this sense, the selectivity of LC/TOFMS relies on the resolving power of the instrument on the m/z axis, enabling discrimination between the target species and an isobaric interference within 0.05 Da of the mass difference (using 230 m/z, as example). In the case of terbuthylazine, it is easily

Figure 4.

Upper: LC/TOFMS TIC of a "positive" for terbuthylazine in an olive oil sample. Lower: EIC of terbuthylazine using a 20 mDa mass window.

www.agilent.com/chem

References
1. Proposal for a Regulation of the European Parliament and of the Council on maximum residue levels of pesticides in products of plant and animal origin, COM/2003/0117/final COD2003/0052*/; www.europa.eu.int/pol/food/indexes.htm 2. Ch. Lentza-Rizos, E. J. Avramides, and F. Cherasco, (2001) J. Chromatogr. A 912: 135142. 3. J. J. Vreuls, R. J .J. Swen, V. P. Goudriaan, M. A. T. Kerkhoff, G. A. Jongenotter, and U. A. Th. Brinkman, (1996) J. Chromatogr. A 750: 275286. 4. C. Ferrer, M. J. Gmez, J. F. Garca-Reyes, I. Ferrer, E. M. Thurman, and A. R. FernndezAlba, (2005) J. Chromatogr. A 1069: 183194.

Acknowledgments
We acknowledge Amadeo Fernandez-Alba, JuanFrancisco Garca-Reyes and Carmen Ferrer from the Department of Hydrogeology and Analytical Chemistry, University of Almera, for help, ideas, and analytical support. Agilent contact: Jerry Zweigenbaum

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice.

Agilent Technologies, Inc. 2005 Printed in the USA August 23, 2005 5989-3573EN

Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection Application

Food Safety, Environmental

Authors
Chin-Kai Meng and Bruce Quimby Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
In this application note, a gas chromatography/mass spectrometry (GC/MS) system capable of providing up to four signals from a single injection is described. When a three-way micro-fluidic splitter is added to the end of the column, two additional signals from GC detectors can be acquired together with the MS data from a single injection. This multi-signal configuration provides: full-scan data for library searching, selective ion monitoring (SIM) data for trace analysis, micro-electron capture detector and flame photometric detector data for excellent selectivity and sensitivity from complex matrices. A combination of element selective detectors, SIM/Scan, and deconvolution reporting software makes a very powerful pesticide analysis system. Examples for trace-level compound quantitation/confirmation or for screening are discussed.

Chromatograph (LC) depending on the nature of the compounds. For GC amenable compounds, the traditional detectors are NPD (Nitrogen Phosphorus Detector), ECD (micro-Electron Capture Detector), and FPD (Flame Photometric Detector) for their excellent sensitivity and selectivity. However, even with dual-column confirmation analysis, these GC detectors cannot be used to verify the identity of the compounds with high confidence. Full scan mass spectral data and library searching are typically used for final compound verification. However, full-scan analysis has a worse (higher) detection limit (DL) compared to selective detectors on a GC. To improve the DL, the technique selective ion monitoring (SIM) is often used. With SIM, the MS monitors only a few characteristic ions for each target compound within the retention time (RT) range that the target elutes from the column. By monitoring only a few specific ions, the signal-to-noise ratio (S/N) improves significantly. The ions monitored are time programmed in groups corresponding to the RTs of the targets. SIM analyses with closely eluting targets require precise alignment of chromatographic RTs with the time programming of SIM groups. The Retention Time Locking (RTL) technique can be applied to eliminate the need to adjust SIM group time-windows after column maintenance or replacement. In this application note, a GC/MS system capable of providing up to four signals from a single injection is described. The benefits of the multi-signal detection include: Confirmatory information Full-scan data for library search capability

Introduction
Many laboratories in the world are analyzing pesticide residue levels in both foods and the environment to protect human health. The process usually involves homogenizing the sample, extracting the pesticides, and analyzing the target compounds with a Gas Chromatograph (GC) or a Liquid

Maximum sensitivity SIM data enables trace analysis Excellent selectivity ECD and FPD detect trace-level hetero-compounds from complex matrices

Experimental
A recent technical note describes Synchronous SIM/Scan, which takes advantages of the Performance Electronics in the 5975 inert MSD to get both SIM and full-scan signals in a single run without sacrificing performance [1]. The SIM method can be easily developed automatically using the ChemStations AutoSIM tool [2]. By simply selecting a checkbox in the method, the SIM and fullscan data can be acquired together. The trade-off is giving up some cycles per second but gaining an additional signal (full-scan data or SIM data) for the whole analysis. With properly chosen acquisition parameters, for example, increasing the scan speed, the decrease of cycles per second is usually not significant and does not affect peak quantitation or the quality of results (for example, S/N).

Figure 2.

A close-up view of the micro-fluidic three-way splitter in the 6890 GC oven.

EPC Injection MSD (SIM/Scan) Split vent ECD Splitter FPD (P)

At the end of the column, effluent flow is split three ways according to the length and diameter of the capillary tubing (restrictor) used.

Figure 1.

A schematic of the multi-signal configuration. Note: the EPC flow adds to the column flow into the splitter.

The size of the micro-fluidic plate is 1.25 inches (3.2 cm) wide and 2.5 inches (6.4 cm) tall. The device was designed to eliminate the common problems of large thermal mass, excess dead volume, and leaky connection due to oven temperature cycling etc. The splitter's flow paths and connection points are laid out and etched onto a thin, stainless steel plate using photolithography and chem-milling technologies. The plate is diffusion bonded, mounted with column connectors, and surface deactivated, resulting in an integrated and compact micro-fluidic splitter. Metal ferrules are used at the connectors that are leak-free after temperature cycling and will not absorb solvents or sample matrix, improving sensitivity for trace analysis applications. Deactivated capillary tubing between the splitter and each detector was used as a flow restrictor. Aux EPC pressure and the restrictor dimensions were determined using a spreadsheet-like calculator program to achieve the proper split ratio among all detectors. The three-way splitter can easily turn into a two-way splitter when a connector is capped. Other advantages of a splitter include backflushing [3] and quick-swapping. The Aux EPC flow can be run-time programmed to a higher pressure, while at the same time the inlet pressure is lowered to near ambient. This causes the column flow to reverse direction, back-flushing the less volatile materials out of the split vent of the inlet. The Aux EPC on the splitter also allows column changing and inlet maintenance without cooling and venting the MSD. The splitters flow paths and connection points were designed in such a way

Besides the SIM/Scan data, the ChemStation software can simultaneously acquire up to two additional GC detector signals, for example, FPD (in phosphorus- or sulfur- mode) and NPD (nitrogenphosphorus detector) signals or both P- and S- signals from a dual-wavelength FPD (DFPD). See Figure 1. Figure 1 is a schematic for multi-signal detection. At the end of the column, a three-way micro-fluidic splitter was used to split the column effluent to different detectors [3]. For this study, an FPD and a ECD were installed. Notice on the figure that an Auxiliary Electronic Pneumatics Control (Aux EPC) gas channel was connected to the splitter to maintain the pressure at the end of the column so that the split ratios/flows are kept constant throughout a run. Figure 2 shows a close-up view of the microfluidic splitter installed in the GC oven.
2

that when the column fitting is removed, the helium gas from the Aux EPC purges the fitting, preventing air from entering the splitter/MSD. See Table 1 for hardware details and settings.

Table 1. GC Inlet

Gas Chromatograph, Mass Spectrometer, and Three-Way Splitter Operating Parameters Agilent Technologies 6890 EPC Split/Splitless Splitless, 1.0 L injected (7683 ALS) 280 C ~27 psi (chlorpyrifos methyl RT locked to 16.596 min) 50.0 mL/min 0.75 min 55.3 mL/min Off Helium Siltek Cyclosplitter, 4-mm id, Restek p/n 20706-214.1

Mode Inlet temp Pressure Purge flow Purge time Total flow Gas saver Gas type Inlet liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Total run time Equilibration time Oven max temp Column Length Diameter Film thickness Mode Nominal initial flow Outlet Outlet pressure

C/min 25 3 8

Final (C) 70 150 200 280

Hold (min) 2.00 0.00 0.00 15

46.87 min (last standard elutes around 35 min) 0.5 min 325 C Agilent Technologies HP 5-ms, p/n 19091S-433 30.0 m 0.25 mm 0.25 m Constant pressure 2.5 mL/min Unspecified 3.8 psi (Aux EPC pressure to splitter)

Front detector (FPD) Phosphorus mode Hydrogen flow: Oxidizer flow: Temperature: Oxidizer gas type: Mode: Makeup flow: Makeup gas type: Lit offset: Data rate: 75.0 mL/min 100.0 mL/min 250 C Air Constant makeup flow 60.0 mL/min Nitrogen 2.00 5 Hz Sulfur mode Hydrogen flow: Oxidizer flow: 50.0 mL/min 60.0 mL/min

Table 1.

Gas Chromatograph, Mass Spectrometer, and Three-Way Splitter Operating Parameters (Continued) 300 C Constant makeup flow 60.0 mL/min Nitrogen 5 Hz MSD Transfer line heater 280 C Helium 3.80 psi 0.00 min (this value will follow oven ramp) Agilent Technologies 5975 inert MSD Atune.U Scan 3.00 min Atune voltage 45 amu 555 amu 100 2 4 2.89 150 C 230 C Agilent 6890N Option 890, when installed on the GC during factory assembly 10:10:1 MSD:FPD:ECD 1.444 m 0.18-mm id Deactivated fused silica tubing 0.532 m 0.18-mm id Deactivated fused silica tubing 0.507 m 0.10-mm id Deactivated fused silica tubing 1.53 mL/min 1.53 mL/min 0.153 mL/min 1.38 mL/min

Back detector (ECD) Temperature: Mode: Makeup flow: Makeup gas type: Date rate: Thermal AUX 2 Use: Initial temp: Pressure AUX 5 Gas type: Initial pressure: Initial time: MSD Tune file Mode Solvent delay EM voltage Low mass High mass Threshold Sampling A/D Samples Scans/s Quad temp Source temp Three-way splitter Split ratio MSD restrictor FPD restrictor ECD restrictor Flow to MSD (at 280 C) Flow to FPD (at 280 C) Flow to ECD (at 280 C) Makeup flow (at 280 C)

Software Used in this Application Note GC/MSD ChemStation G1701DA Deconvolution Reporting Software (DRS) G1716AA NIST Library G1033A AMDIS (included for free with the NIST library CD)

Results and Discussion


Figure 3 shows four signals that were simultaneously acquired from a single injection of a pesticide mixture. Due to the high sensitivity of the ECD, the split ratios for the three detectors was set to MSD:FPD:ECD = 10:10:1. This split ratio distributes the sample of a 1-L splitless injection of a 1-ppm (1000 pg/L) sample to the different detectors as labeled in Figure 3.

Scan: ~ 480 pg

SIM: ~ 480 pg

ECD: ~ 48 pg

FPD (P): ~ 480 pg

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

Figure 3.

Signals acquired simultaneously from a 1-L splitless injection of 1-ppm standard. The split ratios were MSD:FPD:ECD = 10:10:1.

Figure 4 shows the signals when the pesticide standard was diluted 100-fold in a produce matrix. The total ion chromatogram (TIC) from full scan was not shown due to the lack of sensitivity. The FPD(P) and ECD were able to detect all the pesticides spiked in this extract. For trace-level target compound analysis, the SIM signal can be used for quantitation and the GC signals used for further confirmation.
1800 1700 1500 1300 1100 900 700 500 300 5.00 10.00 15.00 20.00 25.00 30.00 35.00

SIM

Produce extract 2688, spiked

700000 600000 500000 400000 300000 5.00 1240000 1200000 1160000 1120000 1080000 1040000 1000000 5.00

FPD(P)

10.00

15.00

20.00

25.00

30.00

35.00

ECD

10.00

15.00

20.00

25.00

30.00

35.00

Figure 4.

Data of a produce extract spiked at 10 ppb. FPD and ECD were able to detect the respective standards spiked into the extract.

Another application for this multi-signal system is for screening. In screening, no target list is available for the analysis; therefore, SIM acquisition or MS/MS is not possible. Figure 5 shows three signals (no SIM) from a produce extract.

800000 700000 600000 500000 400000 300000 200000 100000 5.00 50000 45000 40000 35000 30000 25000 20000 15000 10000 5.00 110000 100000 90000 80000 70000 60000 50000 40000 30000 5.00 10.00 15.00 20.00 25.00 30.00 35.00 10.00 15.00 20.00 25.00 30.00 35.00 10.00 15.00 20.00 25.00 30.00 35.00

Produce extract 13-10927

SCAN

40.00

45.00

FPD (P)
40.00 45.00

ECD

40.00

45.00

Figure 5.

Full-scan, FPD(P), and ECD data for extract 13-10927.

The Deconvolution Reporting Software (DRS) [3, 4] found several pesticides in the TIC as shown in Figure 6.

Figure 6.

Report for extract 13-10927 generated from DRS.

The possible pesticides in the sample were benzophenone, chlorpyrifos methyl, and thiabendazole. Propoxur and metamitron were not confirmed by both AMDIS and NIST; therefore, they were most likely false positives. Due to the complexity of the sample matrix and other interferences, it is sometimes difficult to get a high library match factor from peaks in the TIC, even after background subtraction. Therefore, element selective detectors would be very useful in providing the supporting information for compound confirmation. The multi-signal system was retention time locked, therefore, from the RT and the aligned peaks from the FPD(P) and the ECD responses, chlorpyrifos methyl (C7H7Cl3NO3PS) was confirmed.

It usually takes less than 3 minutes to turn off the FPD photomultiplier, swap the P-filter with the S-filter, and turn the photomultiplier back on. After the swap, adjust the detector gas flows to optimize the response in either P- or S- mode. A new injection of the same extract was made in FPD(S) mode. The FPD(S) result is shown with previously acquired signals in Figure 7. Two major peaks were seen on the FPD(S) chromatogram. From the peak RTs, they supported the presence of chlorpyrifos methyl and thiabendazole (C10H7N3S) respectively. Note that the full-scan TIC barely showed a peak for either compound, which made it impossible for traditional data analysis to identify both compounds. The FPD(S) mode is very selective, but it is not as sensitive as the FPD(P) mode. Although the ECD is very sensitive, it is not as selective as the FPD. A combination of GC detectors, SIM/Scan, and DRS makes a very powerful pesticide analysis system.

Conclusion
The Synchronous SIM/Scan provides users with library searchable full-scan spectra as well as trace level SIM data in a single analysis. When a threeway micro-fluidic splitter is added to the end of the column, two additional signals from element selective detectors can be acquired together with the MS data from a single injection. This configuration makes it very attractive for the analysis of trace-level pesticide residues in foods or environmental samples. This multi-signal configuration provides: full-scan data for library searching, SIM data for trace analysis, ECD and FPD data for excellent selectivity and sensitivity from complex matrices. In this application note, examples of ECD signal and FPD signal (P- or S- mode) were acquired together with the SIM/Scan data from a single injection for trace-level compound quantitation/confirmation, or for screening.

800000 700000 600000 500000 400000 300000 200000 100000 5.00 23000 21000 19000 17000 15000 13000 5.00 50000 45000 40000 35000 30000 25000 20000 15000 10000 5.00 10.00 15.00 20.00 25.00 30.00 10.00 15.00 20.00 25.00 30.00 10.00 15.00 20.00 25.00 30.00

Produce extract 13-10927

SCAN
35.00 40.00 45.00

Thiabendazole

FPD (S) (new injection)

Chlorpyrifos methyl

35.00

40.00

45.00

FPD (P)
35.00 40.00 45.00

Figure 7.

Full-scan, FPD(S), and FPD(P) data for extract 13-10927.

www.agilent.com/chem

References
1. Chin-Kai Meng, Improving Productivity with Synchronous SIM/Scan, Agilent Technologies, publication 5989-3108, www.agilent.com/chem 2. Harry Prest and David W. Peterson, New Approaches to the Development of GC/MS Selected Ion Monitoring Acquisition and Quantitation Methods, Agilent Technologies, publication 5988-4188, www.agilent.com/chem 3. Michael J. Szelewski and Bruce Quimby, New Tools for Rapid Pesticide Analysis in High Matrix Samples Agilent Technologies, publication 5989-1716, www.agilent.com/chem 4. Philip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD Using Deconvolution Reporting Software Agilent Technologies, publication 5989-1157, www.agilent.com/chem

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2005 Printed in the USA July 6, 2005 5989-3299EN

Determination of Fungicides in Fruits and Vegetables by Time-of Flight and Ion Trap LC/MS Application

Food Safety

Author
Imma Ferrer and E. Michael Thurman Pesticide Residue Research Group University of Almera, 04120 Almera, Spain

Introduction
The importance of the identification and quantitation of fungicides in fruits and vegetables was described [1, 2]. Increasingly, LC/MS is being used for the analysis of many polar fungicides of the European Union (EU) Directives and is rapidly becoming an accepted technique in pesticide residue analysis for regulatory monitoring. In particular LC/MS works well on the various families of fungicides that are polar and labile. Thus, often LC/MS methods are preferred over the older gas chromatography/mass spectrometry (GC/MS) methods. For example, the use of MS and MS/MS for pesticides (including fungicides) in food was reviewed by Pico et al. [3], and there are currently about 100 published papers dealing with the analysis of pesticides in food. Of these, there are approximately 25 papers that deal with fungicides by LC/MS, the majority published in the early 2000s. Thus, LC/MS is an emerging technology for the analysis of food products. Interestingly though, none of the reported papers in this most recent review deal with the use of LC/TOFMS and accurate mass analysis for fungicides in food. Thus, there is an important need for research studies and methods development on the analysis of pesticides in food by accurate mass using LC/ TOFMS [1, 2]. Our study in this report is one of the first of its kind to examine the new Agilent LC/MSD TOF time-of-flight mass spectrometer for the analysis of fungicides in food. This topic was chosen because of the relevance of these fungicides and their significant use on apples, oranges, lemons, melons, tomatoes, broccoli, and peppers.

Abstract
This application examines the feasibility of the new instrumentation of electrospray and liquid chromatography/ time-of-flight mass spectrometry (LC/TOFMS) in conjunction with liquid chromatography/ ion trap mass spectrometry (LC/ITMS) to analyze five major fungicides in fruit extracts (apple, lemon, melon, and orange) and in salad vegetables (tomato, broccoli, and pepper). Included are the LC/TOFMS and LC/ITMS MS/MS spectra of five important fungicides (carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole), the TOF empirical formula and MS/MS fragmentation and diagnostic ion(s) for these fungicides in the matrices of various important fruits and vegetables. A detailed rapid procedure for sample preparation and extraction of these fungicides from the fruit and vegetables is given. Spectral quality at the limit of detection (LOD), linearity, and quantitation of the fungicides in pure solvent and in the fruit and vegetable extracts using TOF and ion trap are provided. Last are the results of the analysis of real samples from the marketplace for these fungicides in fruits and vegetables: apples, oranges, lemons, and melons.

Furthermore, LC/ITMS will be demonstrated as a companion technique to LC/TOFMS for the analysis of fungicides in fruits and vegetables. The companion use of LC/ITMS was described in more detail [2]. The molecular structures of the common fungicides of this study (carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole) are shown in Figure 1.

2. Take 5 mL of the supernatant into a 15-mL tube, add 250 mg of PSA adsorbent and 750 mg of MgSO4, and vortex for 20 s. Then microcentrifuge again for 3 min at 3700 rpm. Transfer 1.0 mL into an autosampler vial. Then evaporate the fruit and vegetable residues to dryness and redissolve in 8/92% methanol/ water. Standards were prepared by fortifying with a mixture of carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole at concentrations ranging from 0.01 to 0.5 mg/kg for analysis using the Agilent LC/MSD TOF and the Agilent 1100 Series LC/MSD Trap. Nonfortified samples were analyzed directly at this same point by either the LC/MSD TOF or the LC/MSD Trap by injecting 50 L. The extraction method is summarized in Figure 2.

O H N NH N Carbendazim [M+H]+ = m/z 192 OCH3

H N

S N

Thiabendazole [M+H]+ = m/z 202

N O CN Azoxystrobin [M+H]+

N O H3CO = m/z 404 O OCH3 Cl F

Sample 1 kg of fruit or vegetable. F

H3CO

OCH3 O N O H3C O N

N Cl Dimethomorph [M+H]+ = m/z 388 Triflumizole [M+H]+ = m/z 346

Combine 15 g of sample, 15 mL of acetonitrile (1% acetic acid), extract and shake for 60 s with some Vortex to mix. Add 6 g of MgSO4 and 2.5 g of NaAc3H2O and shake. Centrifuge for 3 min at 3700 rpm.

Put 5 mL of sample into a 15-mL tube, add 250 mg of PSA adsorbent and 750 mg of MgSO4, and vortex and shake for 20 s. Then microcentrifuge for 180 s at 3700 rpm.

Figure 1.

Structures of major fungicides used on fruits and vegetables, and the corresponding [M+H]+ ion. Figure 2.

Transfer 1.0 mL into LC/MS vials

Experimental Methods
Fruit and Vegetable Extraction (QuEChERS) QuEChERS is the acronym for the method of extraction, which stands for quick, easy, cheap, effective, rugged, and safe [4]. It is a method that is receiving wide acceptance for rapid extraction of pesticides in food. 1. Weigh 15 g of a previously homogenized sample into a 40-mL Teflon centrifuge tube. Add 15 mL of acetonitrile (containing 1% acetic acid), 6 g of anhydrous MgSO4 and 2.5 g of NaAc3H2O (sodium acetate trihydrate) and shake the sample vigorously for 1 min by using a Vortex mixer at maximum speed or by hand shaking. Afterwards, centrifuge for 3 min at 3700 rpm.
2

QuEChERS Extraction method for fungicides in fruits and vegetables.

Agilent LC/MSD TOF Methods LC Pumps were Agilent 1100 binary pumps, injection volume 50 L with standard Agilent 1100 ALS. Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, part number 993967-906 Mobile Phase A=acetonitrile and B=0.1% formic acid in water, gradient was 10% A for 5 min, then to 100% A in 25 min at a flow rate of 0.6 mL/min Agilent LC/MSD TOF with electrospray source Positive ESI, Capillary 4000 V

Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Fragmentor 190 V, Skimmer 60 V, Oct DC1 37.5 V, OCT RF V 250 V Reference Masses: m/z 121.0509 and 922.0098, resolution: 9500500 @ m/z 922.0098, Reference A Sprayer 2 is constant flow rate during the run Agilent 1100 Series LC/MSD Trap Methods Chromatographic methods were identical to Agilent LC/MSD TOF for direct comparison of peaks LC Pumps were Agilent 1100, injection volume 50 L Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, part number 993967-906 Mobile Phase A=acetonitrile and B=0.1% formic acid in water, gradient was 10% A for 5 min, then to 100% A in 25 min at a flow rate of 0.6 mL/min. Agilent 1100 Series LC/MSD Trap Positive electrospray ionization (ESI), Capillary 3200 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Capillary exit 70 V Trap accumulation, 50,000 counts
Intensity, cps

17.6 5.0e7 4.0e7 3.0e7 2.0e7 1.0e7 0.0 2 1.6e6 1.4e6 1.2e6 1.0e6 8.0e5 6.0e5 4.0e5 2.0e5 0.0 2 4 4 6 8 2.1 3.6 4.1 13.4 17.4 15.1 18.0 20.8 22.7 24.0 25.1

(a) 28.9

10 12 14 16 18 20 22 24 26 28 30 Time, min 4 (b)

Intensity, cps

3 5

10 12 14 16 18 20 22 24 26 28 30 Time, min

Figure 3.

(a) TIC of a lemon standard solution (0.125 ppm); (b) EICs of carbendazim (1), thiabendazole (2), dimethomorph (3), azoxystrobin (4), and triflumizole (5)

of the compounds are shown in Figure 3b. The window of extraction was clean, which is commonly a feature of accurate mass extraction of ions, where the width of the window of extraction may be narrowed to (0.02 amu) or 100 ppm. In Figure 3, the window of extraction was 0.1 amu. At concentrations as low as 0.05 mg/kg, the extracted ions (m/z 192, 202, 346, 388, and 404) still yielded clean chromatograms testifying to the importance of accurate mass and its ability to give clean extracted ion chromatograms (EICs) with a narrow mass window. This accurate mass window is reflected in the determinations of the accurate mass of the protonated molecule of each of the fungicides. Table 1 shows the elemental composition, the exact mass, and errors, in mDa and ppm, for each of the fungicides at the 0.50 mg/kg concentration. Mass accuracy was always better than 2 ppm in all the fruit and vegetable matrices, except for dimethomorph.

Results and Discussion


Fragmentation and Mass Accuracy Figure 3 shows the LC separation of five fungicides: carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole. The sample contained the fungicides at 0.125 mg/kg (ppm) in a fortified lemon extract. The extracted ions for each
Table 1.

Accurate Mass and Elemental Composition of Five Fungicides at 0.50 mg/kg in Lemon Matrix Error mDa -0.05 -0.34 +1.1 0.20 -0.35 Error ppm 0.27 1.7 2.5 0.50 1.0 3

Compound Carbendazim Thiabendazole Dimethomorph Azoxystrobin Triflumizole

Formula C9H9N3O2 C10H7N3S C21H22NO4Cl C22H17N3O5 C15H15N3OF3Cl

Selected ion [M + H]+ [M + H]


+

m/zexperimental 192.0767 202.0430 388.1321 404.1243 346.0925

m/zcalculated 192.07675 202.04334 388.13101 404.1242 346.09285

[M + H]+ [M + H]+ [M + H]+

These results show that the use of continuous calibration is effective for accurate mass across an order of magnitude concentration range in complex fruit and vegetable matrices. Figures 4a4e show the fragmentation pathway of each of the five fungicides, based on accurate mass spectra with pure standards in a methanol solution. The fragmentation ions of the LC/MSD TOF were verified with the LC/MSD Trap spectra at MS2 with identical chromatographic conditions. Beginning with carbendazim, the MS2 shows that the protonated molecule loses methanol (32 Da) to give the m/z 160 fragment ion (Figure 4a). This fragment ion of m/z 160 is also seen in the LC/MSD TOF. The fragmentation pathway for thiabendazole is shown in Figure 4b with the protonated molecule, m/z 202, fragmenting to give the m/z 175, with the neutral loss of 27 Da or HCN. The fragmentation pathway for azoxystrobin is shown in Figure 4c with the protonated molecule, m/z 404, fragmenting to m/z 372 (loss of methanol of 32 Da) at MS2, and to m/z 342 at MS3. The fragmentation pathway for dimethomorph is shown in Figure 4d with the protonated molecule, m/z 388, fragmenting to give the m/z 301 at MS2, then m/z 165 at MS3. The fragmentation pathway for triflumizole, m/z 346, is to m/z 278 at MS2, then to m/z 250 at MS3 (Figure 4e). The combination of the LC/MSD Trap and LC/MSD TOF is extremely valuable for interpretation of spectra in that both instruments work well for these compounds and give complementary information on the structure and identity.

MSD Ion Trap of carbendazim


+H+ NH O

Fragment ion at MS2


+H+ O N

N H

OCH3

Carbendazim, [M+H]+ = m/z 192

Fragment ion, m/z 160

192.0766
4.7e4 4.0e4
N N O
H2 N N
O

Int ensity, counts

N H

NH

OCH3

3.0e4

2.0e4

160.0502

1.0e4

193.0796
0.0 145 155 165 175 m/z, amu 185 195 205 215

Figure 4a. Structure and fragmentation pathways for carbendazim using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.

MSD Ion Trap of thiabendazole


H N +H+ S

Fragment ion at MS2


+H+ N S

N Thiabendazole, [M+H]+

N N = m/z 202 Fragment ion, m/z 175

1.00e5 9.00e4 8.00e4 Intensity, counts 7.00e4 6.00e4

202.0433

S 5.00e4 4.00e4 3.00e4 202.2072 2.00e4 203.0459 1.00e4 200.0 202.0 204.0404 204.0 206.0 m/z, amu 208.0 210.0 N H

Figure 4b. Structure and fragmentation pathways for thiabendazole using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF. 4

N O

N Cl O H3CO +H+ OCH3 O

Ion Trap MS/MS of dimethomorph

Fragment ion at MS2


H3CO OCH3

Ion Trap MS/MS of azoxystrobin, [M+H]+ = m/z 404

CN

Azoxystrobin, m/z 404

+H+ N H3CO H3CO O Dimethomorph, [M+H+]+ = m/z 388 O

N O +H+

Fragment ion, m/z 372, at MS2


CN

Cl Fragment ion, m/z 301

Fragment ion, m/z 372 H3CO O

H3CO

Fragment ion at MS3 Fragment ion, m/z 342, at MS3


CN N O N O +H+ O 8.0e4 7.0e4 Intensity, counts Intensity, counts 4.0e5 1 ppm accuracy 3.0e5 O 2.0e5 CN H3CO 1.0e5 372.0982 OH CN H3CO OH 405.1275 426.1061 427.1092 420 430 N N O 1 ppm accuracy N O N O OCH3 2.0e4 1.0e4 404.1246 6.0e4 5.0e4 4.0e4 3.0e4 389.1341 Cl 390.1280 O O 388.1310 O NH

O +H+ Fragment ion, m/z 165 MS3

Fragment ion, m/z 342

CH3 O CH3

391.1317 386.0 387.0 388.0 389.0 390.0 391.0 392.0 393.0 394.0 m/z, amu

373.1008 370 380 390

406.1301 400 410 m/z, amu

Figure 4d. Structure and fragmentation pathways for dimethomorph using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.

Figure 4c. Structure and fragmentation pathways for azoxystrobin using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.

Ion Trap MS/MS of triflumizole

Fragment ion at MS2

+H+ O+ O

N N N N

CF3

CF3

Cl

Triflumizole [M+H]+ = m/z 346

Cl

Fragment ion, m/z 278

Cl

Fragment ion at MS3


N O + CF3 Fragment ion, m/z 250 at MS3

broccoli, and pepper). Results showed the similarity among matrices with the LC/MSD TOF and LC/MSD Trap, given the variability in fruit and vegetable matrices. For example, Figures 5a5b show the standard curves for two representative fungicides, carbendazim and azoxystrobin. These figures demonstrate that there was little or no matrix suppression in the LC/MSD TOF system for the fruit extracts up to 0.5 ppm of parent compound for carbendazim, thiabendazole, and azoxystrobin. However, triflumizole and dimethomorph showed both enhanced signal and suppressed signal (data not shown here). The enhancement occurred for the fruit extracts of melon, orange, and lemon. The vegetable extracts of pepper showed no enhancement and some suppression for broccoli was observed. One possible explanation for these results for triflumizole has to do with the late retention time of the analyte (26 min), which means that the mobile phase is approximately 80% acetonitrile. The ESI signal is susceptible to matrix effects at these high concentrations of organic solvent [5], which indicates the importance of using matrix matched standards for unknown analysis of food extracts. Thus, best accuracy for reporting concentrations was with the standard curve made up in matrix for each fruit or vegetable. This was the procedure that was used for unknown analysis in real fruit and vegetable samples.
(a)
F

Cl 278.0558 2.5e5

Carbendazim
30 F Area 10 E+06

O+ H3C N CF3 O N

2.0e5 Intensity, counts

20

1.5e5

10

Solvent Pepper Lemon Orange Broccoli Melon

1.0e5

Cl 280.0529 279.0586 281.0556 280 290 300 310 320 m/z, amu 330

N 346.0930

0.1

0.2

5.0e4 1.0e4

0.3 0.4 0.5 Concentration (mg/kg)

0.6

347.0958 348.0912 340 350 Area 10 E+06

50

Azoxystrobin

(b)

40

Figure 4e. Structure and fragmentation pathways for triflumizole using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.

30 Solvent Pepper Broccoli Lemon Orange Melon

20

Linearity and Detection Limits


Calibration curves were established for the five fungicides using both the LC/MSD TOF and LC/MSD Trap over the analyte concentration range of interest, which was from 0.01 mg/kg to 0.5 mg/kg (ppm) in solvent and in each of the fruit and vegetable matrices (orange, lemon, melon,
6

10

0 0 0.1 0.2 0.3 0.4 0.5 Concentration (mg/kg) 0.6

Figure 5. Overlay of standard curves for the LC/MSD TOF analysis of carbendazim (5a) and azoxystrobin (5b) in five matrices.

Furthermore, Figures 5ab also show that the standard curves were linear across the range of concentration tested with correlation coefficients of 0.990 to 0.999 for all matrices tested and for all four of the fungicides (Table 2). Similar results for standard curves were seen with the LC/MSD Trap, with correlation coefficients typically of 0.9900.001.
Table 2. Correlation Coefficients for Standard Curves for Various Pesticides in Representative Fruit Matrices Orange 0.998 0.999 0.990 0.996 Melon 0.997 0.999 0.990 0.999 Lemon 0.999 0.999 0.997 0.996

For example, carbendazim is regulated at 0.10 mg/kg in tomatoes, thiabendazole is 0.50 mg/kg for tomatoes, while azoxystrobin is not regulated in tomatoes in Spain; therefore, the maximum residue limit is automatically 0.01 mg/kg. The LC/MSD TOF is capable of these LODs for each of the fungicides in the various fruit and vegetable matrices. The quantitation for real fruit extracts from actual grocery store samples are shown in Table 5 for LC/MSD TOF analysis of fungicides with LC/MSD Trap confirmation. Thus, these data show that the LC/MSD TOF is capable of accurate mass measurements of the fungicides in fruit and vegetable extracts with LODs needed for their monitoring in the EU.

Compound Carbendazim Thiabendazole Azoxystrobin Trifluomizole

Table 3 shows the limits of detection (LODs) of the fungicides for various matrices. Typically, the values of Table 3 vary from 0.001 ppm to 0.010 ppm. The European standard for pesticides with no regulatory standard is 0.010 ppm. Thus, the LC/MSD TOF is sufficiently sensitive to detect these compounds in all matrices. A similar result was achieved for the LC/MSD Trap for LODs (Table 4) using single MS. The accurate mass capabilities of the LC/MSD TOF appear capable of seeing through the complexity of the fruit and vegetable matrices without interferences to the LODs. These LOD results of the various fungicides in fruits and vegetables are capable of meeting the regulation limits of Spain and at those of the EU.
Table 3.

Table 4.

LOD in mg/kg for Fungicides by LC/MSD Trap in Three Matrices in Full-Scan Mode Orange 0.005 0.005 0.005 0.002 0.002 Melon 0.005 0.010 0.003 0.005 0.002 Lemon 0.010 0.005 0.010 0.010 0.005

Compound Carbendazim Thiabendazole Dimethomorph Azoxystrobin Triflumizole

Table 5.

Concentration (in mg/kg) of Fungicides by LC/MSD TOF and Confirmed by LC/MSD Trap in Samples of Fruits from Grocery Store Produce Orange 0.03 <LOD 0 0 0 Melon <LOD 0 0 0 0 Apple 0.1 0 0 0 0 Lemon 0.5 0.1 0 0 0

Compound LOD in mg/kg for Fungicides by LC/MSD TOF in Three Carbendazim Matrices Thiabendazole Compound Orange Melon Lemon Dimethomorph Carbendazim 0.004 0.005 0.008 Azoxystrobin Thiabendazole 0.005 0.010 0.005 Triflumizole Dimethomorph 0.005 0.002 0.008 Azoxystrobin Triflumizole 0.001 0.001 0.001 0.001 0.001 0.001

www.agilent.com/chem

Conclusions
LC/MSD TOF analysis is a powerful tool for identification of fungicides in fruits and vegetables and is a new tool for environmental food chemistry. Quantitation is easily possible over 2 orders of magnitude with accuracy better than 3 ppm, typically less than 2 ppm, and in this work was an amazing 1 ppm in ESI+ ion for most compounds! Elemental composition of fungicides and fragment ions are possible with LC/MSD TOF, and further confirmed with LC/MSD Trap. LODs of the five fungicides in fruit and vegetables were from 0.001 to 0.010 g/g. These concentrations are equal to or better than the EU directives for controlled fungicides in fruits and vegetables.

References
1. Imma Ferrer and E. Michael Thurman, (2004) "Determination of chloronicotinyl insecticides in salad vegetables by LC/MSD TOF and LC/MSD ion trap", Agilent Technologies, publication 5989-1842EN www.agilent.com/chem 2. E. Michael Thurman and Imma Ferrer, (2005) "Identification of unknown pesticides in food using both LC/MSD TOF and Ion Trap MSn", Agilent Technologies, publication 5989-1924EN www.agilent.com/chem 3. Y. Pico, C. Blasco, and G. Font, (2004) Mass Spectrometry Reviews, 23, 45-85. 4. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, and F.J. Schenck, (2003) Journal of AOAC International, 86: 412-431. 5. E.M. Thurman, I. Ferrer, and A.R. FernndezAlba, Chapter 8: LC/MS I. Basic Principles and Technical Aspects of LC/MS for Pesticide Analysis, in Chromatographic-Mass Spectrometric Food Analysis for Trace Determination of Pesticide Residues, Ed. A.R. Fernndez-Alba, Elsevier, Amsterdam, 2005.

Acknowledgements
We acknowledge Amadeo Fernndez-Alba, Department of Hydrogeology and Analytical Chemistry, University of Almera for analytical support.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem. Agilent Contact: Jerry Zweigenbaum

ZORBAX is a registered trademark of E.I. duPont de Nemours Co. Inc. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2005 Printed in the USA March 15, 2005 5989-2209EN

Identification of Unknown Pesticides in Food Using Both LC/MSD TOF and Ion Trap MSn Application

Food Safety

Author
E. Michael Thurman and Imma Ferrer Pesticide Residue Research Group University of Almera 04120 Almera Spain

searching of ions. The procedure here consists of multiple steps. First is the initial detection of a possible unknown pesticide in actual market-place vegetable extracts (tomato or another salad vegetable) using either the LC/MSD TOF or the LC/MSD ion trap. Second is the application of the "Dual Approach." This consists of accurate mass identification of an empirical formula or molecular mass (usually containing an A+2 isotope signature, such as from S or Cl), followed by MS/MS ion trap confirmation of the chlorinated and/or sulfur-containing pesticides and their major fragment ions and chlorine or sulfur signatures. Third is the database analysis of the empirical formula, which is carried out with either the Merck Index CD, containing 10,000 compounds, or the ChemINDEX CD, containing 77,000 compounds, and a laptop computer. In the examples given, the unknowns are identified as the organophosphate insecticide, malathion, in cucumbers; and the insect growth regulator, buprofezin in tomato extracts. The fourth and final step involves confirmation with authentic standards.

Agilent Contact: Jerry Zweigenbaum 2850 Centerville Road Wilmington, DE 19808-1610

Abstract
Traditionally, the screening of unknown pesticides in food was accomplished by gas chromatography/mass spectrometry (GC/MS) methods using conventional library searching routines. However, many of the new polar and thermally labile pesticides are more readily and easily analyzed using liquid chromatography/mass spectrometry (LC/MS) methods, even though no searchable libraries currently exist (with the exception of some limited user libraries). There is, therefore, a need for LC/MS methods to detect true pesticide unknowns. This application develops an identification scheme for unknown pesticides using a combination of liquid chromatography/mass selective detector (LC/MSD) with time-of-flight (TOF) and ion trap (MSn) options, and searching for the empirical formula using the accurate mass and the ChemINDEX or Merck Index databases. The approach is different from the conventional library

Introduction
The identification and quantitation of insecticides in vegetables is of great importance to individuals and health organizations around the world. The European Union (EU), has set new Directives for pesticides at low levels in vegetables in order to

meet these health concerns. For example, new laws such as the European Directive 91/414/EEC [1], or the Food Quality Protection Act (FQPA) [2] in the US have increased the standards for human health, workers, and environmental protection. They also require re-registration for older pesticides. Furthermore, the review programs have withdrawn authorizations for many of the crop protection products currently on the market, 177 compounds in the US and 320 in Europe. Moreover, it was announced in Europe that a total of 110 products will be withdrawn in the near future [1]. The revision of the review program in Europe dictates that over 450 existing active substances will be taken off the market by 2008 with about 400 pesticides remaining in use. Next, the quality standards within the new regulations include the re-assessment of the maximum residue limits (MRLs) for vegetables, and EU directives are setting different MRLs for each pesticide within each food group. Typically, the MRLs are lower than the previous ones. Furthermore, the new Directive leads to different MRLs for each EU country, which are still being decided. The EU Directive states that individual country MRLs will be maintained in the new program. Finally, banned compounds have the lowest MRLs, which is set now at 0.01 mg/kg (ppm). With the planned program to remove so many compounds from the market, it is important, even necessary, that screening for unknown pesticides may be done by LC/MS on vegetable extracts. Because it is not always possible to know which banned substances may be used, it is of vital importance to environmental food monitoring that there be a system to give fast and accurate screening of unknown substances in food and food products. Thus, there is an important need for research studies and methods development on the analysis of unknown pesticides in food by new LC/MS methods, such as the combination of accurate mass using LC/MSD TOF, and MS/MS using LC/MSD ion trap. Our study in this report is one of the first of its kind to examine the new Agilent LC/MSD TOF combined with LC/MSD ion trap, and the use of commercial databases, such as the Merck Index and the ChemINDEX database to identify unknown pesticides in food. Several advantages of the combination of LC/MSD TOF and ion trap are that accurate mass and empirical formulas may be combined with the MS/MS spectra or even MSn for spectral information. The identification of unknowns using the

LC/MSD TOF and ion trap consists of six steps, which are outlined below. 1. Analyze the vegetable extract with ion trap in full-scan mode looking for important unknown peaks or alternatively, using in-source CID on LC/MSD TOF. 2. Next check for A+2 isotope patterns to see if Cl, Br, or S is present. 3. Search Merck Index or ChemINDEX for possible unknowns using molecular mass (nominal mass weight), A+2 isotopes, and molecular formula. 4. Proceed to ion trap MS/MS with proposed ideas for structure and do MS2 or MS3. Identify ion fragments, accurate mass fragments, and possible structures. 5. Combine with LC/MSD TOF data of accurate fragment ions and protonated molecule to make tentative identification. 6. Obtain and analyze appropriate standard for final confirmation. Given are two detailed examples of this process using store-purchased salad vegetables, cucumbers and tomatoes, both of which contained "unknown white powders" that were subsequently identified by the above process for various "unknown pesticides."

Experimental Methods
Standard Vegetable Extraction [3] 1. Tomato, lettuce, pepper, and cucumber, in 2-kilogram portions, were obtained from the market place and homogenized by chopping. 2. Then 15 g of the homogenized vegetable was accurately weighed and placed into a 200-mL PTFE centrifuge tube. 3. Ethyl acetate (45 mL) and 6.5 M NaOH (1 mL) were added and blended in a high-speed blender (Polytron) for 30 s at 21,000 rpm [4]. After this time, 13 g of anhydrous Na2SO4 was added and the extraction repeated for 30 s. 4. The combined extracts were then filtered through a thin layer of 20 g of anhydrous Na2SO4. The solid was washed with 50 mL of ethyl acetate and the combined extracts were evaporated to dryness on a vacuum rotary evaporator using a water bath at 45 5 C.

5. The dried residue was dissolved by sonication in 15 mL of methanol. The extracts, which contained 1 g of sample per mL, were filtered through 0.2-m PTFE filters (Millex FG, Millipore) prior to LC/MS analysis. Rapid Vegetable Extraction 1. Select tomato or cucumber containing white powder from a commercial market place. 2. Carefully wash the vegetable skin three times with methanol to remove the white powder, recovering a total of 2 to 5 mL of wash, (depending on the size of the vegetable) into a 150-mL Pyrex beaker. 3. After mixing, transfer the methanol wash to a 5-mL syringe and filter through a Millex-FH PTFE filter and aliquot 0.3 mL. 4. Dilute with 0.6 mL of deionized water. 5. Analyze by either LC/MSD TOF or LC/MSD ion trap. LC/MSD TOF Methods Instrument: Agilent model LC/MSD TOF with electrospray source LC pumps were Agilent 1100 binary pumps, injection volume 50 L with standard Agilent ALS Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm 5 m, part number 993967-906 Mobile phases: A: ACN, B: 0.1% formic acid in water Gradient: 15% A to 100% A over 30 minutes, at 0.6 mL/min Positive ESI, capillary: 4000 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Fragmentor: 190 V, skimmer: 60 V Oct DC1: 37.5 V, OCT RF V: 250 V Reference masses: m/z 121.0509 and 922.0098, resolution: 9500 500 @ m/z 922.0098 Reference A Sprayer 2 set to constant flow rate during the run

LC/MSD Ion Trap Methods Chromatographic conditions identical to those using LC/MSD TOF for direct comparison of peaks. Instrument: LC/MSD Ion Trap Classic LC pumps were Agilent 1100 binary pumps, injection 50 L with standard Agilent ALS Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm 5 m, part number 993967-906 Mobile phases: A: ACN, B: 0.1% formic acid in water Gradient: 15% A to 100% A over 30 minutes, flow rate: 0.6 mL/min Positive ESI, Capillary 3000 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 350 C Skimmer 1: 20 V, Skimmer 2: 10 V, Capillary exit offset: 50 V, Capillary exit: 70 V Octopole 3 V, Oct RF 100 V, Octopole : 2 V, Lens 1: -3 V, Trap drive: 50, Lens 2: -50 V Target: 50,000, max. accumulation time: 200 ms, scan m/z 50 to 800, averages: 5 Rolling average: on, MS/MS m/z 2.0 isolation width, amplitude: 1.2 V, fragmentation cutoff m/z 83

Results and Discussion


Cucumber Extract Figure 1 shows the ion trap total ion chromatogram (TIC) for the rapid extract of the white powder present on a store-purchased cucumber. The largest peak in the chromatogram, deemed the Star*TIC, has a spectrum shown in Figure 1 (bottom panel) with ions at m/z 331 and 353. The spacing of m/z 22 indicates a sodium adduct. Because the analysis was carried out in positive ion, the [M+H]+ is the m/z 331, and the [M+Na]+ is the m/z 353. Furthermore, note that the m/z 331 ion also contains an A+2 ion at m/z 333, which is 8% of the m/z 331 ion. The two important A+2 ions to consider in unknown ID are chlorine and sulfur (occasionally bromine). Sulfur-34 is present in natural abundance of 4.8%; thus, the 8% peak is most likely due to two sulfur-34 atoms present in the unknown at m/z 331.
Intens. 106 2.5 2.0 1.5 1.0 0.5 0.0 0 5 TIC EIC 301 Intens. 106 0.8 0.6 0.4 0.2 0.0 98.9 170.0 100 200 285.1 300 400 500 Possible fragment ions 353.1 575.5 600 682.8 709.5 700 m/z 10 15 20 EIC 331 m/z 331 25 m/z 301 Time [min]

Intens. 106 2.0 1.5 1.0 0.5 0.0 20.5 Intens. 3000 2000 1000 0 Fragment ions 126.9 172.9 240.1 98.7 100 200 300 284.9 21.0 21.5 22.0 22.5 23.0 23.5 24.0 24.5 Time [min] TIC

MS/MS of 331 No fragmentation of 353 verifying it is a Na adduct The m/z 285, 240, and 99 ions are present 400 500 600 700 m/z

Figure 2.

Ion-Trap MS/MS and corresponding product mass spectrum for m/z 331.

TIC of cucumber skin Star*TIC

The next step is the search of the Merck Index for the molecular weight and sulfur atom content. Figure 3 shows an example search sheet on a laptop computer. Because the ion trap gives the protonated molecule (m/z 331) the search is done by subtracting a proton to give the molecular weight of 330. The molecular weight is then searched across one mass unit to 331. The molecular formula has only the S2, which is a requirement now of the database search.

331.1

1. [M+H]+ = m/z 331 2. m/z 353 = [M+Na]+ 3. S-34 ~8% = 2 S atoms

Figure 1.

Upper: cucumber skin ion-trap TIC; lower: extracted ion chromatogram (EIC) for m/z 331. The lower panel is the spectrum of the Star*TIC peak.

Figure 2 shows the LC/MS/MS of m/z 331. The ions present include: m/z 285, 240, 127, and 99. These ions will be used later for structural identification after a database search on the Merck Index finds a possible structure. These ions will then be compared with that structure.

Figure 3.

Database search for MW 330-331.

Figures 4a and 4b show the two results of the Merck Index database search. The compounds found were penicillin O and malathion, with molecular weights of 330.43 and 330.36, respectively.

Results of library search for MWt = 330-331 Molecular Formula = S2*

Monograph Number: 7170 Title: Penicillin O CAS Registry Number: 87-09-2


H2C S N H N O COOH S CH3 CH3 O H H

CAS Name: (2S,5R,6R)-3,3-Dimethyl-7-oxo-6-[[(2-propenylthio)acetyl]amino]-4-thia-1-azabicyclo[3.2.0]heptane-2carboxylic acid Additional Names: [(allylthio)methyl] penicillin; allylmercaptomethylpenicillin; allylmercaptomethylpenicillinic acid; penicillin AT Molecular Formula: C13H18N2O4S2
Molecular Weight: 330.43. Percent Composition: C 47.25%, H 5.49%, N 8.48%, O 19.37%, S 19.41% Literature References: Antibiotic produced by Penicillium chrysogenum. Biosynthesis of salts: Behrens et al., J. Biol. Chem. 175, 793 (1948); Rhodehamel, Behrens et al., US 2528175 and US 2623876 (1950, 1952, both to Lilly); Ford et al., Antibiot. & Chemother. 3, 1149 (1953); Ford, US 2647894 (1953 to Upjohn); Paleckov, Slechta, C.A. 50, 17309g (1956).

Derivative Type: 2-Chloroprocaine salt monohydrate Additional Names: Chloroprocaine penicillin O; penicillin O 2-chloroprocaine Trademarks: Depo-Cer-O-Cillin Chloroprocaine (Upjohn) Molecular Formula: C H ClN O S .H O 26 37 4 6 2 2 Molecular Weight: 619.20. Percent Composition: C 50.43%, H 6.35%, Cl 5.73%, N 9.05%, O 18.09%, S 10.36% Properties: Slender needles from hot water, mp 79-81. Practically insol in cold water. Stable in dry form at room temp. Aq suspensions are stable at room temp for 1 week, at refrigerator temps for 3 weeks. Calculated activity: 949 units/mg. Solubilities: Weiss et al., Antibiot. & Chemother. 7, 374 (1957). Melting point: mp 79-81

Figure 4a. First result for database search for MW 330-331.

Monograph Number: 5723 Title: Malathion CAS Registry Number: 121-75-5 CAS Name: [(Dimethoxyphosphinothioyl)thio]butanedioic acid diethyl ester
H3CO H3CO S P S

O O

CH3 CH3

Additional Names: diethyl mercaptosuccinate S-ester with O,O-dimethyl phosphorothioate; S-(1,2dicarbethoxyethyl) O,O-dimethyldithiophosphate; insecticide no. 4049; carbofos; malathon (obsolete); mercaptothion; phosphothion Manufacturers' Codes: ENT-17034 Trademarks: Cythion; Derbac-M (International); Malamar 50; Malaspray; Organoderm (Mundipharma); Prioderm (Coates & Cooper); Suleo-M (International) Molecular Formula: C10H19O6PS2
Molecular Weight: 330.36. Percent Composition: C 36.36%, H 5.80%, O 29.06%, P 9.38%, S 19.41% Literature References: Cholinesterase inhibitor. Prepn: Johnson et al., J. Econ. Entomol. 45, 279 (1952); Cassaday, US 2578652 (1951 to Am. Cyanamid). Purification: Usui, US 2962521 (1960 to Sumitomo). Treatment of Pediculus humanus (head lice) infestation: D. Taplin et al., J. Am. Med. Assoc. 247, 3103 (1982). Toxicity data: T. B. Gaines, Toxicol. Appl. Pharmacol. 14, 515 (1969). Review of distribution, transport and fate in the environment: M. S. Mulla et al., Residue Rev. 81, 1-159 (1981); of carcinogenic risk: IARC Monographs 30, 103-129 (1983). Properties: Deep brown to yellow liq, mp 2.9. bp 156-157. Characteristic odor. d 25 1.23. n 25 1.4985. Vapor pressure at 30: 4 10-5 mm Hg. Slightly sol in 4 D and alkylated aromatic hydrocarbons and vegetable oils. water (145 ppm). Misc with many organic solvents 0.7 including alcohols, esters, ketones, ethers, aromatic Limited soly in certain paraffin hydrocarbons. Petroleum ether is sol to about 35% in malathion. Hydrolyzed at pH >7.0 or <5.0. Stable in an aq soln buffered to pH 5.26. LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines). 50 Melting point: mp 2.9 Boiling point: bp 156-157 0.7 Index of refraction: n 25 1.4985 D Density: d 25 1.23 4 Toxicity data: LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines) 50 CAUTION: Potential symptoms of overexposure are miosis, aching eyes, blurred vision and lacrimation; eye and skin irritation; salivation; anorexia, nausea, vomiting, abdominal cramps, diarrhea, giddiness, confusion and ataxia; rhinorrhea, headache; tightening of chest, wheezing and laryngeal spasms. See NIOSH Pocket Guide to Chemical Hazards (DHHS/NIOSH 97-140, 1997) p 188. Use: Insecticide. Therap-Cat: Pediculicide. Therap-Cat-Vet: Ectoparasiticide.

Figure 4b. Second result for database search for MW 330-331. 5

The fragmentation ions shown in Figure 5 fit the ions from the MS/MS experiment of the m/z 331 (Figure 2). Therefore, malathion is a good possibility of an identification from the Database search of the Merck Index.
Intensity, cps OH S H3CO H3CO P S O C10H20O6PS2 Exact Mass: 331.0433 Error = 0.8 ppm
+

2.1 1.00e7 0.7 8.00e6 2.8 6.00e6 4.00e6 2.00e6 4.0 18.6 2.4 3.5 Star*TIC 27.5 22.5 28.6 24.9 26.0 27.8 30.5 21.1 23.3

O O

CH3 CH3 S H3CO H3CO P S Na

O O O O CH3 CH3

0.00 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Time, min

Figure 6a. LC/MS TOF analysis of a cucumber extract.

-46

HO

C10H19NaO6PS2+ Exact Mass: 353.0253 Error = 1 ppm

H3CO H3CO

S P S O

When this formula is entered into the Merck Index database the only match is malathion (Figure 7).
OH

H3CO H3CO

S P S

O C8H14O5PS2+ Exact Mass: 285.0015 Error = 0.8 ppm HO

O C7H14O4PS2+ Exact Mass: 257.0066 Error = 0 ppm O

O C6H7O3+ Exact Mass: 127.0390 Error = 0 ppm

O H

C4H3O3+ Exact Mass: 99.0077 Error = 3.0 ppm

Thus, the accurate mass gives the same formula match that is consistent with the identification by ion trap. Furthermore, Figure 6b also shows the accurate mass ions for the possible fragmentation of the unknown at masses of m/z 353.0256, 285.0017, 127.0389, and 99.0081. These accurate mass ions are the same ones measured with the ion trap in MS/MS mode of the m/z 331 at nominal mass (with the exception of the m/z 353 adduct). The masses of these ions match the formulas shown in Figure 5 including the sodium adduct of the m/z 331. Thus, the probability of this being the correct identification is quite high. The final step is the identification by standard matching. This was done by both ion trap MS/MS and by TOF; both gave this compound as the correct identification. Malathion is a banned substance for cucumbers so this identification represents an important finding. Based on the identification point scheme for unknowns, this compound receives 2.0 points for the protonated molecule, 2.0 points for the m/z 285 ion, and 1.5 points for the MS/MS transition from m/z 331 to 285 for a total of 5.5 points. Four identification points are required for banned substances in food (see reference by Stolker et al. [5] in book by Ferrer and Thurman [6]). The complementary nature of the two instruments is also shown with the next unknown example.

Figure 5.

Fragmentation possibilities for malathion.

The next step is obtaining the accurate mass with the LC/MSD TOF. Figure 6a shows the TIC and the nearly identical retention times between the two instruments for the unknown Star*TIC because of using the same HPLC column. Figure 6b shows the accurate mass for the m/z 331.0435 and only one formula match of C10H19O6PS2.

OH 1.6e5 1.5e5 1.4e5 1.3e5 1.2e5 1.1e5 1.0e5 Intensity, counts 9.0e4 8.0e4 7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 333.0404 O 127.0389 O H3CO H3CO S P S O OH 257.0066 353.0256 99.0081 HO H3CO O HO S P S O H3CO H3CO O S P S O Na O O O CH3 CH3 331.0435

Accurate Mass of 331 m/z 331.0435 Formula matches...C10H19O6PS2

S H3CO H3CO P S O

O O

CH3 CH3

285.0017

O H H3CO

60

80

100

120

140

160

180

200

220

240 m/z, amu

260

280

300

320

340

360

380

400

Figure 6b. LC/MS TOF spectrum of peak at 22.5 min (m/z 331).

Monograph Number: 5723 Title: Malathion CAS Registry Number: 121-75-5 CAS Name: [(Dimethoxyphosphinothioyl)thio]butanedioic acid diethyl ester
H3CO H3CO S P S

O O

CH3 CH3

Additional Names: diethyl mercaptosuccinate S-ester with O,O-dimethyl phosphorothioate; S-(1,2dicarbethoxyethyl) O,O-dimethyldithiophosphate; insecticide no. 4049; carbofos; malathon (obsolete); mercaptothion; phosphothion Manufacturers' Codes: ENT-17034 Trademarks: Cythion; Derbac-M (International); Malamar 50; Malaspray; Organoderm (Mundipharma); Prioderm (Coates & Cooper); Suleo-M (International) Molecular Formula: C10H19O6PS2
Molecular Weight: 330.36. Percent Composition: C 36.36%, H 5.80%, O 29.06%, P 9.38%, S 19.41% Literature References: Cholinesterase inhibitor. Prepn: Johnson et al., J. Econ. Entomol. 45, 279 (1952); Cassaday, US 2578652 (1951 to Am. Cyanamid). Purification: Usui, US 2962521 (1960 to Sumitomo). Treatment of Pediculus humanus (head lice) infestation: D. Taplin et al., J. Am. Med. Assoc. 247, 3103 (1982). Toxicity data: T. B. Gaines, Toxicol. Appl. Pharmacol. 14, 515 (1969). Review of distribution, transport and fate in the environment: M. S. Mulla et al., Residue Rev. 81, 1-159 (1981); of carcinogenic risk: IARC Monographs 30, 103-129 (1983). Properties: Deep brown to yellow liq, mp 2.9. bp 156-157. Characteristic odor. d 25 1.23. n 25 1.4985. Vapor pressure at 30: 4 10-5 mm Hg. Slightly sol in 4 D and alkylated aromatic hydrocarbons and vegetable oils. water (145 ppm). Misc with many organic solvents 0.7 including alcohols, esters, ketones, ethers, aromatic Limited soly in certain paraffin hydrocarbons. Petroleum ether is sol to about 35% in malathion. Hydrolyzed at pH >7.0 or <5.0. Stable in an aq soln buffered to pH 5.26. LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines). 50 Melting point: mp 2.9 Boiling point: bp 156-157 0.7 Index of refraction: n 25 1.4985 D Density: d 25 1.23 4 Toxicity data: LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines) 50 CAUTION: Potential symptoms of overexposure are miosis, aching eyes, blurred vision and lacrimation; eye and skin irritation; salivation; anorexia, nausea, vomiting, abdominal cramps, diarrhea, giddiness, confusion and ataxia; rhinorrhea, headache; tightening of chest, wheezing and laryngeal spasms. See NIOSH Pocket Guide to Chemical Hazards (DHHS/NIOSH 97-140, 1997) p 188. Use: Insecticide. Therap-Cat: Pediculicide. Therap-Cat-Vet: Ectoparasiticide.

Figure 7.

Database search for C10H19O6PS2. 7

Tomato Extract In this next example, a similar protocol is followed except that this time we use the LC/MSD TOF first to obtain an in-source CID spectrum. Figure 8a shows the Star*TIC for the rapid extraction of the tomato white powder.
23.9 3.0e7 2.5e7 2.0e7 Intensity, cps 14.7 1.5e7 1.0e7 2.2 5.0e6 0.0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Time, min 8.3 14.5 11.3 18.5 24.9 3.2 2.0 26.0 26.7 30.4

The accurate mass is m/z 306.1642, which results in two possible formulae after examination of the A+2 isotope pattern, which shows one sulfur atom in the molecule. Furthermore, the A+1 has an area relative to the m/z 306 of 17%, which indicates 15 carbon atoms. See Figure 8b.
4.0e5 306.1642 m/z 306.1642 C16H23N3OS C14H28N O2PS 2.4 ppm -3.0 ppm

Star*TIC
3.0e5 Intensity, counts

2.0e5

1.0e5 306.6200 0.0 280 290 300

307.1664 A+117%~C15 308.1624 A+25%S1 310 m/z, amu 320 330

Figure 8b. LC/MS TOF spectrum of peak at 23.9 min (m/z 306).

Figure 8a. LC/MS TOF analysis of a tomato extract.

The two formulae were both then searched in the Merck Index and only one formula gave a database hit. That compound was C16H23N3OS, the insect growth regulator buprofezin. Buprofezin is used extensively on white flies according to the Merck Index (Figure 9). Thus, this compound was a good candidate for a positive identification.

Monograph Number: 1486 Title: Buprofezin


N N O CH3 H3C S N C(CH3)3

CAS Registry Number: 69327-76-0

CAS Name: 2-[(1,1-Dimethylethyl)imino]tetrahydro-3-(1-methylethyl)-5-phenyl-4H-1,3,5-thiadiazin-4-one Additional Names: 2-tert-butylimino-3-isopropyl-5-phenylperhydro-1,3,5-thiadiazin-4-one Manufacturers' Codes: NNI-750; NNK-758; NN-29285; PP-618 Trademarks: Applaud (Nihon Nohyaku) Molecular Formula: C16H23N3OS
Molecular Weight: 305.45. Percent Composition: C 62.92%, H 7.59%, N 13.76%, O 5.24%, S 10.50% Literature References: Insect growth regulator which inhibits chitin synthesis. Prepn: Z. Grnecker et al., DE 2824126; K. Ikeda et al., US 4159328 (1978, 1979 both to Nihon Nohyaku); H. Kanno, Pure Appl. Chem. 59, 1027 (1987). Mode of action study: T. Asai et al., Appl. Entomol. Zool. 20, 111 (1985). Control of whiteflies and scale insects: I. Ishaaya et al., Meded. Fac. Landbouwwet., Univ. Gent 54, 1003 (1989). GC-MS determn in clementine citrus: P. Cabras et al., J. Agr. Food Chem. 46, 4255 (1998). Persistence in olives and olive oil: idem et al., Food Addit. Contam. 17, 855 (2000). Review of physical properties, activity and field trials: H. Kanno et al., Proc. Br. Crop Prot. Conf. - Pests Dis. 1981, 59-66. Properties: Crystals from isopropyl alcohol, mp 106.1. Soly at 25 (g/l): acetone 240, chloroform 520, ethanol 80, toluene 320; water 0.9 mg/l. Vapor pressure at 25: 9.4 10-6 mmHg. LD in mice, rats (mg/kg): 10000, 8740 orally. LC (48 hr) in carp: 2-10 mg/l (Kanno, 1981). 50 50 Melting point: mp 106.1 Toxicity data: LD Use: Insecticide. 50 in mice, rats (mg/kg): 10000, 8740 orally; LC 50 (48 hr) in carp: 2-10 mg/l (Kanno, 1981)

Results of Library SearchC14H28NO2PSNo hit. Molecular Formula = C16H23N3OS


Figure 9. Database search for C14H28NO2PS and C16H23N3OS. 8

Note in the spectrum the presence of the ion at m/z 201.1059 (Figure 10). LC/MSD ion trap MS/MS of the m/z 306 ion gave the m/z 201 and the further MS3 yielded the m/z 116 ion.
Intens. 107 1.0 306.1656 0.5 0.0 23.00 Intens. 6 10 3 2 1 115.9 1.0e5 201.1059 116 0.0 100 200 300 400 Formula = C9H17N2OS 500 600 m/z, amu 700 800 900 1000 0 MS3 gives 116

4.0e5

23.50

24.00

Intensity, counts

3.0e5

24.50 25.00 Time [min]

200.9 MS/MS gives 201

2.0e5

Furthermore, the expected mass for the 201.1056 fragment ion matched the value from the LC/MSD TOF (201.1059) quite closely (0.0003 u), which gave a high certainty for identification. After obtaining the buprofezin standard, the final data show a perfect match, which further shows the ability of the LC/MSD TOF and LC/MSD ion trap to identify unknowns. Buprofezin is allowed for tomatoes; therefore, it is not a banned substance. The quantitative tolerance for buprofezin may then be measured by the standard extraction and analysis for vegetable acceptance to European Union standards [1].

100 200 300 400 500 600 700 m/z

Conclusions
LC/MSD TOF and LC/MSD Trap are complementary and powerful tools for identification of pesticides in vegetables and represent a new approach for environmental food chemistry using LC/MS. The combination of these two tools, the twin mass spectrometer(s), with a pesticide database, Merck Index or ChemINDEX, works often for identification of unknown pesticides. The use of identification of fragment ions with MS2 and MS3 is also powerful when combined with accurate mass of LC/MSD TOF CID spectra. Combining TOF and Trap for unknown ID works [4, 7]! Future work should include chromatographic data to help in the identification of unknowns and to provide a simple pesticide library generated through calculation of empirical formula and isotope ratios.

Figure 10. LC/MS TOF spectrum and LC/MS/MS Ion Trap spectrum for m/z 306.

It was possible to draw reasonable chemical structures for the m/z 201 and 116 fragment ions (Figure 11).
Buprofezin m/z 306 S N N O H3C CH3 N C(CH3)3

MS2

S MS3 N N O H3C 201 m/z fragment ion C9H17N2OS Exact mass: 201.1056 CH3 C(CH3)3

H N m/z 116 C(CH3)3

Conclusion: Buprofezin Standard confirmationyes.

Figure 11. Reasonable structures of fragment ions consistent with the structure of buprofezin.

www.agilent.com/chem

References
1. EU Food Directives, 2002, 91/414/EEC 2. Food Quality Protection Act, 1998 (FQPA) 3. Aguera, A.; Lopez, S.; Fernandez-Alba, A.R.; Contreras, M.; Crespo, J.; Piedra, W., 2004, J. Chromatogr. A., 1045: 125-135. 4. Imma Ferrer and E. Michael Thurman, "Determination of Chloronicotinyl Insecticides in Salad Vegetables by LC/MS/ESI/TOF and LC/MS Ion Trap", Agilent Technologies publication 5989-1842EN, www.agilent.com/chem 5. A.A.M Stolker, E. Dijkman, W. Niesing, E.A. Hogendoorn, 2003, "Identification of residues by LC/MS/MS" In: Imma Ferrer and E. Michael Thurman, editors Liquid Chromatography/Mass Spectrometry/Mass Spectrometry and Time-ofFlight Mass Spectrometry for the analysis of emerging contaminants, (2003) American Chemical Society Symposium Volume 850. 6. Imma Ferrer and E.M. Thurman, (2003), Liquid Chromatography/Mass Spectrometry/Mass Spectrometry and Time-of-Flight Mass Spectrometry for the analysis of emerging contaminants, American Chemical Society Symposium, 850. 7. E. Michael Thurman, Imma Ferrer, A. R. Fernandez-Alba, (2005), "Matching Unknown Empirical Formulas to Chemical Structure Using LC/MS TOF Accurate Mass and Database Searching: Example of Unknown Pesticides on Tomato Skins" Journal of Chromatography, Special Issue on Mass Spectrometry, In press.
Acknowledgments Amadeo Fernandez-Alba, Department of Hydrogeology and Analytical Chemistry, University of Almera for analytical support. ChemOffice with Chemfinder, the Merck Index, and ChemIndex are products of CambridgeSoft. ZORBAX is a registered trademark of E.I. duPont de Nemours Co. Inc. Millex is a registered trademark of the Millipore Corp. Merck Index is a registered trademark of Merck. Pyrex is a registered trademark of Corning Corporation. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2005 Printed in the USA February 18, 2005 5989-1924EN

New Tools for Rapid Pesticide Analysis in High Matrix Samples Application

Food Analysis

Authors
Mike Szelewski and Bruce Quimby Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Tools like Method Translation [1] have made it straightforward to reduce analysis time by a known factor and maintain the exact relative elution order of the analytes. The use of appropriate shorter and smaller diameter columns can maintain the same resolution while achieving a much shorter analysis time. One application area where this approach has met difficulty, however, is the gas chromatography/ mass spectrometry (GC/MS) analysis of pesticides in complex matrices like food. This application requires that speed-up schemes maintain column capacity in order to handle the large matrix peaks and achieve low detection limits for analytes. Since chromatography is governed by the triangle of speed, resolution, and capacity, resolution must be sacrificed to increase speed at the same capacity. The problem is that chromatographic resolution is also needed to confirm the identity of any target analytes detected in the presence of interferences from the sample. In this note, the reduction in chromatographic resolution in faster analysis is more than adequately compensated for by use of spectral deconvolution [2] and simultaneous element-selective detection for the confirmation step. The system consists of a GC/MS with a dualwavelength flame photometric detector (DFPD) for the simultaneous collection of phosphorus, sulfur, and mass spectral data. The GC column effluent is split between the two detectors in the ratio of 2:1 in favor of the mass selective detector (MSD). The system is retentiontime locked to the Agilent pesticide library [1] scaled to threefold faster times, which contains the

Abstract
Recent developments in GC/MS hardware and software make it possible to analyze samples with high levels of matrix contamination much faster than ever before. New tools such as mass spectral deconvolution, reliable and inert effluent splitters, and column backflushing capabilities can be combined to produce large time savings. By accelerating the chromatographic run, post-run bakeout, and data interpretation steps, analysis times can be shortened by at least three-fold versus conventional methods. These tools are especially useful in analyses with high levels of matrix background, such as the inspection of the food supply for contaminants. In addition to monitoring for pesticide residues, the threat of terrorism has recently raised concerns over deliberate contamination of food with other toxic materials. This article describes a GC/MS system for the rapid screening of foodstuffs for chemical contaminants with a special emphasis on pesticides, organophosphorus, and organosulfur compounds.

Introduction
Techniques for decreasing the analysis time of gas chromatography (GC) methods have been developed in recent years.

retention times and spectra for 567 pesticides used worldwide. Samples are analyzed with MS in fullscan electron-impact ionization (EI) mode. The combination of precise retention times, elemental, and mass spectral data is used to screen for specific target compounds. The flame photometric detector (FPD) data also highlights any non-target, P- or S-containing compounds for identification by MS. The MS data is screened using the standard quantitation software based on retention time (RT), ion ratios, and spectral cross correlation. The MS data is also processed using spectral deconvolution software, which greatly reduces spectral interferences from the matrix. The deconvoluted spectra are then searched against a table of targets. Any hits are confirmed by searching against the main NIST library. This process is automated by the Agilent Deconvolution Reporting Software (DRS), also providing significant time savings in data interpretation.

The system described here uses column backflushing, a technique used to save large amounts of time with complex samples. Backflushing is done with the splitter hardware. This technique removes heavy residues from the column much faster and at lower temperatures than the conventional bakeout step at the end of the run. This reduces MS source contamination by preventing the higher levels of column bleed and heavy matrix components from entering the MSD. It also increases the column lifetime. The approach used thus reduces analysis time in three major ways: shortening the chromatographic run time; automating data interpretation; and reducing bakeout time. Other notable advantages are the ability to change columns and/or inlet liners without venting the MSD, and a reduced need for MS source cleaning. System Configuration The system configuration used is shown in Figure 1. Key components are:

Auto-sampler Dual Flame Photometric Detector


Sulfur Phosphorus

Effluent Splitter with Makeup

AUX EPC

0.814 m 0.18 mm id

Column

1.1 m 0.18 mm id

6890N GC

15 m 0.25 mm id 0.25 um HP-5MS

5973 Inert MSD

Figure 1.

System configuration.

Key Components Fast Oven With the 6890N 220V oven (option 002), the pesticide analysis method can be run precisely 3 times faster (14 min) using a 15 m HP-MS column. If the 220V GC is further equipped with SP1 2310-0236 (puts MSD interface in back of oven under rear injection port) and the G2646-60500 oven insert accessory (reduces oven volume twofold), the speed can be increased to 4.8 times faster (9 min). The cool-down time of the oven is also reduced. Dual FPD 6890N Option 241 is a single flame photometric detector with two optical detection channels, one for sulfur and one for phosphorus. The signals from the DFPD are collected, stored, and processed by the MS ChemStation simultaneously with the MS data. The FPD data can be used in several ways. Nontarget organophosphorus compounds like new pesticides or designer nerve agents are highlighted. The presence of an element at the retention time of an identified compound can be used to support confirmation of identity. The response on the FPD can be used for quantitative or semi-quantitative analysis, especially for situations where no calibration standard is available for an identified analyte. Microfluidic Splitter 6890N Option 889 uses diffusion bonded plate technology combined with metal column ferrules to make an inert, easy-to-use, leakfree, high-temperature column effluent splitter. The splitter uses Auxillary (Aux) electronic pneumatics control (EPC) for constant pressure makeup (6890N Option 301). The Aux EPC makeup can be pressure programmed at the end of the run to higher pressure, while at the same time the inlet pressure is lowered to near ambient. This causes the flow in the column to reverse direction, backflushing heavy materials out the split vent of the inlet. The Aux EPC also allows column changing and maintainance without venting the MSD. When the column fitting is removed from the splitter, helium from the makeup supply purges the fitting, preventing air from entering the MSD. If the column is attached to the splitter but removed from the inlet, helium flows backwards through the column and out the inlet end. Inlet maintainance or column headtrimming can be done without cooling and venting the MSD and air is not introduced into a hot source. MSD System The 5973N Inert with Performance Electronics and performance turbo (G2579A) EI MSD is used. This configuration provides faster full scan rates while maintaining sensitivity. The scan rates are compatible with the narrower peaks

generated by fast chromatography. The performance turbo pump is required to handle the higher flows associated with fast chromatography and backflushing. DRS Software (G1716AA) Spectral deconvolution of the MS data allows identification of analytes in the presence of overlapped matrix peaks. This significantly reduces chromatographic resolution requirements, allowing much shorter analysis times. DRS utilizes the AMDIS deconvolution program from NIST, originally developed for trace chemical weapons detection in complex samples. DRS presents the analyst with three distinct levels of compound identification: (1) ChemStation, based on retention time and four ion agreement; (2) AMDIS, based on cleaned spectra full ion matching and locked retention time; (3) NIST02 search using a >147,000 compound library. Instrument Operating Parameters The recommended instrument operating parameters are listed in Table 1. These are starting conditions and may have to be optimized. Split injection was used to match the amount of matrix to the column capacity. Citrus oils cause retention shifts if excess sample is injected. Splitless injection could be used for samples with significantly less matrix. The inlet liner was found to be of low activity, as it does not contain glass wool. Proper mixing for split injections is done by the internal liner geometry. The 6890 220V oven was needed for the ramps described in Tables 1 and 2. This oven program is necessary for the precise 3 speed increase of the RTLocked pesticide database. The 15-m HP-5ms column has the same phase ratio as the 30 m column traditionally used for the 1 method. This shorter column allows a flow rate for a 3 precisely scaled faster method. The outlet is listed as unspecified because the column connects to the splitter. The splitter pressure is operated at a constant 3.8 psig using an auxillary EPC module. The 5973 inert Performance Electronics data acquisition sampling rate was set to 1, which is faster than the typical setting of 2. Signal-to-noise is improved over previous systems at faster sampling rates. More data points allows for easier integration and better deconvolution to compensate for the loss in resolution using a shorter column. The microfluidic splitter parameters are chosen to provide the desired split ratio between detectors
3

while meeting the flow requirements of the detectors used. A primary consideration with the current system is to make sure that the flow to the MSD does not exceed ~4 mL/min while collecting analyte data. It was also desired to split the effluent 2:1 in favor of the MSD. These parameters were entered into the spreadsheet calculator (included with the splitter), which calculated the lengths and diameters of the detector restrictors
Table 1. GC Inlet Mode Inlet temp Pressure Split ratio Split flow Total flow Gas saver Gas type Inlet Liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Ramp 4 Total run time Total run time Equilibration time Oven max temp Column Length Diameter Film thickness Mode Inlet Outlet Outlet pressure Back Detector (FPD) Temperature Hydrogen flow Oxidizer flow Oxidizer gas type Mode Makeup flow Makeup gas type Flame Lit offset Photo multiplier Gas Chromatograph and Mass Spectrometer Operating Parameters Agilent Technologies 6890 EPC Split/Splitless Split, 1.0 L injected 250 C 23.84 psi 10:1 44.1 mL/min 48.1 mL/min Off Helium Siltek Cyclosplitter, 4 mm id, Restek part number 20706-214.1 220V C/min 75 9 24 50 Next C 70 150 200 280 320 Hold min 0.67 0.00 0.00 3.33 (end of pesticide ramp) 50.0 (end of oil elution)

13.96 min to elute pesticides 64.76 min to elute heavy components from citrus oils 0.5 min 325 C Agilent Technologies HP-5MS, p/n 19091S-431 15.0 m 0.25 mm 0.25 m Constant Pressure = 23.84 psi Front Unspecified 3.8 psi (aux pressure to splitter) 250 C 75.0 mL/min 100.0 mL/min Air Constant makeup flow 60.0 mL/min Nitrogen On 5.00 On

Table 1.

Gas Chromatograph and Mass Spectrometer Operating Parameters (Continued) 5 Hz Back detector On 0.0 (Off) 0 Off 0 Signal 2 Data rate: Type: Save data: Zero: Range: Fast Peaks: Attenuation: 5 Hz Front detector On 0.0 (Off) 0 Off 0

Signal 1 Data rate Type Save data Zero Range Fast Peaks Attenuation AUX Pressure 5 Description Gas type Initial pressure Initial time MSD Tune file Mode Solvent delay EM voltage Low mass High mass Threshold Sampling Scans/sec Quad temp Source temp Transfer line temp Splitter Split ratio MSD restrictor DFPD restrictor

Helium 3.80 psi 0.00 min (this value will follow oven ramp) Agilent Technologies 5973 Inert Performance Electronics Atune.U Scan 1.00 min Atune voltage 45 amu 450 amu 0 1 6.68 150 C 230 C 280 C Agilent 6890N Option 889 2:1 MSD:DFPD 1.1 m 0.18 mm id deactivated fused silica tubing 0.81 m 0.18 mm id deactivated fused silica tubing

Backflush Instrument Operating Parameters Instrument operating conditions for backflushing are shown in Table 2. These are starting parameters and will have to be optimized depending on sample matrix. Conditions listed here are only those that are different from the Table 1 conditions. Instead of baking the column at 320 C for 50 min, the heavy matrix material is backflushed through the column inlet, out the split vent. This is accomplished in 5 min at 300 C, saving 45 min of run time, preserving column life, and shortening cool down time. The column inlet pressure is reduced to 1.1 psi by using the ramped pressure feature of the EPC. At the end of the backflush time, it is ramped back to initial conditions.

Simultaneous with ramping the inlet pressure down to 1.1 psig, the Aux EPC splitter pressure is ramped up to 23 psig. This increase in pressure at the column outlet, along with the decrease in inlet pressure, backflushes the column. The backflush pressure for the Aux EPC is set to limit the flow to the MSD to < 8 mL/min. This is calculated using the Agilent Flow Calculation software program, also supplied with the splitter kit. The calculator is used to find the pressure which gives the desired flow through the MSD splitter restrictor (1.1 m 0.18 mm id) at the backflushing temperature 300 C. The result was 23.6 psig which would produce a flow of 8 mL/min, so 23 psig was used. The backflush time is determined empirically and depends on the sample matrix. The process is to try a backflush run followed by a blank run with the conventional long-hold to see if the heavy materials

are completely removed. If not, the process is repeated with a longer backflush time. As a very rough guide, start with a backflush time of five void times for the backwards flow. Using the Agilent Flow Calculation software with the inlet pressure at 23 psig, the outlet pressure at 1.1 psig, and the temperature at 300 C, the hold-up time (void time) is 0.423 min. The rough guide says that the column should be backflushed for 2.12 min. This works for removing most heavies, but 5 min is required in this case to remove them all. At the end of the backflush time, the Aux EPC is ramped back to initial conditions. The MSD is time-programmed to collect data over the time range of 1 to 13.96 min. All of the pesticides elute during this time range.

Table 2.

Backflush Gas Chromatograph and Mass Spectrometer Operating Parameters. Use Table 1 Conditions Except for These Differences C/min 75 9 24 50 Next C 70 150 200 280 300 Hold min 0.67 0.00 0.00 3.33 (end of pesticide ramp) 5.40 (end of oil backflush)

Backflush Oven Conditions Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Ramp 4 Total run time Total run time

13.96 min to elute pesticides 19.76 min to elute heavy components from citrus oils

Backflush Column Conditions Mode Ramped Pressure Press ramp psi/min Initial Ramp 1 99 Ramp 2 99 Backflush AUX 5 Conditions Press ramp psi/min Initial Ramp 1 99 Ramp 2 99 Backflush MSD Conditions Timed MS detector entries Time (min) 13.96

Next psi 23.84 1.1 23.84 Next psi 3.8 23.0 3.8

Hold min 13.96 (end of pesticide ramp) 5.57 (backflush time) 0.00 Hold min 13.96 (end of pesticide ramp) 5.61 (backflush time) 0.00

State (MS on/off) Off

Results
A mixture of 25 pesticides was run at 1, 3, and 4.8 speeds. The Total Ion Chromatograms (TICs) are shown in Figure 2. There is some loss in resolution, but the loss is limited because the shorter columns are run at flows closer to the optimum. The resolution losses are mitigated by using the faster scanning capabilities of the performance electronics together with DRS.

1 30 m

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

3 15 m

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

11.00

12.00

13.00

4.8 10 m

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

5.00

5.50

6.00

6.50

7.00

7.50

8.00

8.50

Figure 2.

TICs of pesticide test mix at three different scaled speeds. All columns have the same phase ratio.

A neat lemon oil was analyzed using the 3 speed conditions. The TIC is shown in Figure 3 with the DFPD phosphorus (P) and sulfur (S) data channels. The ChemStation software can simultaneously acquire two signals of GC data with the MSD data. The pesticides elute within the 14 min window shown, but the matrix continues to elute for an additional 50 min at 320 C.

TIC

7.441 min

FPD (P)
7.257 min

FPD (S)

10

12

14

Figure 3.

MS and DFPD data from lemon oil analyzed with 3x method.

The P and S chromatograms indicate the possible presence of numerous pesticides. The largest P peak, 7.441 min, also contains S. A PBM reverse Library search identified the peak as Methidathion (C6H11N2O4PS3) with a match quality of 45. It was also identified using a target ion and three qualifier ions. The raw apex spectrum of the P peak at 7.257 min is shown in the top of Figure 4. No match was found for a P-containing compound in the top 20 library search results. It was also not identified by the ChemStation target and qualifier ion criteria due to out-of-range ratios.

69 81 93 55 131

Raw spectrum at 7.257 min


109 121 137 159 171 187
180

206
200

227 241
220 240

256 270 283 296


260 280 300

314

329
320

40

60

80

100

120

140

160

100
58 86 76 65

131 97 159

Deconvoluted spectrum
296
171 198 226 252

329

0
86 97
50 80 110

296

329

159 131
140 170

100

Library spectrum of Mecarbam


200 230 260 290 320

Figure 4.

Top - Raw mass spectrum of peak at 7.257 min in lemon oil. Bottom - Deconvoluted spectrum of 7.257 min peak compared to NIST02 library spectrum of Mecarbam.

The DRS report is shown in Figure 5. The peak at 7.257 min is clearly identified as Mecarbam by the DRS software. The deconvoluted spectrum has a match factor of 72, compared to both the Agilent Pesticide AMDIS database and the NIST02 library. Additionally, the DRS report shows the elution time to be only 0.5 s from expected. Further confirmation is the presence of P with S barely visible. The deconvoluted spectrum for the peak at 7.257 min is shown at the bottom of Figure 4, together with the NIST library spectrum of Mecarbam. The DRS report displays the amount for compounds found by the normal ChemStation quant process. The compounds must be properly

calibrated to have a meaningful value. In this lemon oil, the ChemStation found four compounds. The AMDIS portion of DRS found two of the same compounds and an additional five compounds missed by the ChemStation. The NIST02 library search confirmed all of the compounds found by AMDIS using the NIST02 >147,000 compound library. The DRS results show good match quality at the locked retention times for seven target compounds. No single software package can produce this same confidence level in compound identification. An experienced analyst would take 14 hours to process this complex sample manually with each of the three software packages used by DRS. DRS produced this report is less than two minutes.

Figure 5.

DRS Report for lemon oil.

10

Backflushing Citrus oils contain significant amounts of material that elute after the last pesticide. This requires a 150-min hold at 320 C to elute all of the heavy material with a 1 method. The total run time for the 1 method is therefore 195 min, as shown in Table 3.
Table 3. Method Run Time Comparison 30 m 1 min 42 195 n/a 15 m 3 min 14 65 20 10 m 4.8 min 8.75 40.6 12.5

The 3 method requires a 50-min hold at 320 C, as shown at the top in Figure 6, resulting in a 65-min run time. With backflushing, all of this heavy material is removed in 5 min at 300 C, as shown in the bottom of Figure 6. This is a 9-fold reduction in analysis time compared to the 1 method. 4.8x Method Using the 220V oven, SP1 2310-0236, and oven insert accessory, the method can be scaled to 4.8 faster, as shown in Table 3. There is a practical limit to the amount of matrix that can be tolerated with the reduced resolution using a 10-m column. However, for matrices less complex than a citrus oil, the 4.8 method can be successfully used to save even more time. The Performance Electronics allows running at faster scan speeds while maintaining signal/noise ratio. Sufficient points across a peak are maintained even with faster chromatography.

Column Speed Run time Pesticides No backflush matrix With backflush matrix

Normal bakeout: 65 min run

Backflush time range

Backflush: 19 min run

10

20

30

40

50

60

Figure 6.

Top - 3 lemon oil analysis with 50 min bakeout at 320 C. Bottom - 3 lemon oil analysis with 5 min backflush at 300 C.

11

www.agilent.com/chem

Conclusions
New tools are available for the analysis of pesticides in complex matrices. These tools can be combined to significantly reduce analysis and data processing time. Fast oven - programming rates needed for faster runs Shorter column - 3 faster runs with sufficient resolution Microfluidic splitter - confidence in results using selective detection simultaneous with MS data Backflush - additional 3 reduction in run time with lower column temperatures for extended lifetime Performance Electronics - maintain signal/noise at faster sampling rates DRS - three levels of target analyte identification in less than two minutes Using the above tools, the run time for analysis of lemon oil was reduced from 195 minutes to 20 minutes (nine-fold). DRS reduced the data analysis from hours to minutes.

References
1. B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking Agilent Technologies publication 5967-5820E www.agilent.com/chem 2. P. Wylie, M. Szelewski, and C. K. Meng Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software Agilent Technologies publication 5989-1157EN www.agilent.com/chem

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA October 13, 2004 5989-1716EN

A Blind Study of Pesticide Residues in Spiked and Unspiked Fruit Extracts Using Deconvolution Reporting Software Application

Food

Author
Christopher P. Sandy Agilent Technologies, Inc. Block A, CSC Eskdale Road Winnersh Triangle Estate Wokingham, Berk RG41 5DZ United Kingdom

Abstract
The Agilent Technologies mass selective detector (MSD) coupled with deconvolution reporting software (DRS) provides additional powerful data processing capabilities to the MSD ChemStation software. Reviewing full scan gas chromatography/mass spectrometry data for the confirmation of pesticide residues can be a labor-intensive and time-consuming process requiring great skill and concentration by an experienced analyst. The DRS is able to process a complex food extract total ion chromatogram in about 1 minute, whereas an experienced analyst may take more than 30 minutes to achieve the same quality result. Extensive data shown in this report supports the high confidence level that an analyst can have in results rapidly produced by the DRS.

ion ratios. It is sometimes very difficult to confirm target compounds from high matrix background because the matrix affects the ion ratios of the target compounds or complicates the spectrum with additional ions. To be certain of the results, background subtraction and manual integration are often practiced. It is, therefore, a timeconsuming process to confirm target compounds in a dirty matrix. It can take an experienced analyst 15 to 30 minutes to review/confirm one data file. Two powerful gas chromatography/mass spectrometry (GC/MS) techniques - Retention Time Locking (RTL) and deconvolution were combined to create a quantitation and screening tool that can identify 567 pesticides and endocrine disrupters from a single run in 12 minutes. The Agilent Technologies GC/MSD-DRS provides the additional functionality to the MSD ChemStation.

Experimental
DRS Overview A detailed overview of the DRS is given in an application note 5989-1157EN [1], available for download at www.agilent.com/chem. The operating principles of the DRS appear in Figure 1.

Introduction
Typical mass spectral pesticide residue analysis requires finding target ions and meeting qualifier

2.5E7 2E7 1.5E7 1E7 .5E7 5 10 15 20 25 30 35 40

Total ion chromatogram (TIC)

Targets are identified by comparison to locked retention times (RTs) and three qualifying ion ratios, quantified using target ion area versus internal standard (ISTD) calibration table Quant results

AMDIS 32 deconvolutes component spectra and searches target MS database, locked RT used as a qualifier

Deconvoluted target spectra confirmed by AMDIS searched against NIST02 MS database

Confirmed AMDIS hits

Confirmed NIST02 hits

Combined quantitative and qualitative HTML Summary report

Figure 1.

Schematic diagram summarizing the GC/MS DRS.

The quantitation capabilities of the MSD ChemStation are combined with the deconvolution power of the industry standard AMDIS program from NIST. AMDIS is able to separate spectra of interest from dirty matrix spectra present in samples analyzed for pesticides. A third level of confidence is obtained by sending the deconvoluted spectra for library searches of the NIST02 145,000 compound library. A comprehensive report is produced in about 1 minute.

Samples Six samples of fruit extracts, supplied in 90/10 iso-octane/toluene solvent were received for analysis by GC/MS. The samples were prepared by an accredited food pesticide laboratory based in Scandinavia. Three of the samples were spiked with a number of pesticides at varying concentration levels. Although the range of concentrations of the pesticides in each sample was given, neither the actual number of pesticides spiked into each control sample nor the identities were supplied. Details of the samples appear in Table 1. The other three samples were real, unspiked extracts.

Table 1. Sample number 1 2 3 4 5 6

Sample Details for Blind Study Matrix extracted Orange Lettuce Apple Grapes Orange Apple Number of pesticides 2040 2040 2040 24 24 24 Concn range (mg/Kg) 0.020.20 0.020.20 0.010.20 0.11.0 0.25.0 0.052.0 Comments Control sample - spiked Control sample - spiked Control sample - spiked Real sample Real sample Real sample

Instrumentation
The samples were analyzed by full-scan GC/MS using the analytical conditions given in Table 2. Data processing and reporting were performed using the default settings provided with the DRS.

Table 2.

RTL GC/MS Analysis Conditions for Fruit Extract Samples Agilent 6890N 30 m x 0.25 mm id x 0.25 m HP-5MS (p/n 19091S-433) Helium 1.9 mL/min at 70 C 18 psig, constant pressure mode Method RTLocked to methyl chlorpyrifos at 16.593 min PTV, septumless head 90 C (0.3 min) - 1720 C/min - 250 C 0.2 min 30 mL/min 0 psig 60 mL/min 1.0 min 50 L 15 L Empty multibaffle 70(2)-25-150(0)-3-200(0)-8-280(10)

Gas chromatograph Column Carrier gas Flow rate Head pressure Injector type Injector temperature (C), hold time (min), and ramp rate (C/min) Vent time Vent flow Vent pressure Purge flow Purge time Syringe volume Injection volume Liner Oven program: temperature (C), hold time (min), and ramp rate (C/min) MSD MS interface MS source MS quad Detection mode EM voltage

Agilent 5973 inert 280 C 230 C 150 C EI, Scan 40550 amu ATUNE value

Results
The results for the three spiked extracts appear in Table 3 - note that the details of which pesticides were added to the spiked samples were not supplied until after the results were shown to the customer. Those pesticides confirmed by the DRS, are shown lightly shaded. The analytes, shown darkly shaded, are not present in the Agilent RTL Pesticides database. Analyte entries left unshaded were not confirmed.

Table 3.

MSD-DRS Results for Three Spiked Fruit Extract Samples Sample 2: Control-lettuce, spiked Added mg/kg 0.10 0.10 0.10 0.10 0.20 0.10 0.10 0.04 0.02 0.10 0.10 0.10 0.04 0.10 0.04 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.04 0.04 0.10 0.10 0.10 0.10 0.10 0.04 0.04 Pesticide Diphenylamine HCB Lindane (HCH-gamma) Diazinon Chlortalonil Vinclozolin Carbaryl Metalaxyl Pirimiphos-methyl Malathion Chlorpyrifos Cyprodinil Penconazole Captan Folpet** Procymidone Endosulfan-a pp-DDE Bupirimate Endosulfan-b Aclonifen Ethion Triazophos Endosulfan-sulfate Iprodione Bromopropylate Methoxychlor Phosalone Lambda-Cyhalothrin Permethrin Cypermethrin Fenvalerate Deltamethrin Added mg/kg 0.10 0.02 0.04 0.04 0.04 0.04 0.20 0.10 0.10 0.10 0.10 0.04 0.04 0.10 0.10 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.10 0.10 0.10 0.04 0.10 0.10 0.04 0.10 Sample 3: Control- apple, spiked Pesticide Mevinphos Trichlorfon Heptenophos Tecnazene HCH alpha HCH beta Dichloran Pyrimethanil Etrimphos Ethiofencarb Metribuzin Toclophos methyl Linuron Aldrin Diethofencarb Trichloronate Triadimenol Disulfoton sulfoxide Disulfoton sulfone Fluazinam Chlorbenzilate Oxadixyl Benalaxyl Dicofol Fenazaquin Pyrazophos Acrinathrin Bitertanol Cyfluthrin beta Alpha cypermethrin Added mg/kg 0.05 0.05 0.02 0.01 0.01 0.02 0.05 0.02 0.02 0.10 0.05 0.01 0.05 0.02 0.02 0.02 0.05 0.20 0.02 0.05 0.05 0.05 0.05 0.05 0.02 0.05 0.02 0.05 0.05 0.05

Sample 1: Control-orange, spiked Pesticide Methamidofos* Dichlorvos* Acephate* Omethoate Propachlor Chlorprofam Monocrotophos Dimethoate Quintozene Parathion-methyl Dichlofluanid Fenpropimorph Triadimefon Thiabendazole Tolylfluanid Mecarbam Methidation Vamidothion Imazalil Myclobutanil Kresoxim methyl Tebuconazole Phosmet Fenpropathrin Tetradifon Azinphos-methyl Fenarimol Azinpfos-ethyl Prochloraz Flucythrinate Esfenvalerate Azoxystrobin

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 * See Discussion item 1. ** See Discussion item 2.

The results for the three real extracts appear in Table 4. Those pesticides confirmed by the DRS are shown lightly shaded. The darkly-shaded analytes are not present in the Agilent RTL Pesticides database. Analyte entries left unshaded were not confirmed. Analytes with an associated concentration were confirmed as present by the customer using NPD/ECD. Lightly-shaded analytes without a concentration label were detected and confirmed by the DRS, but not by the customer.
Table 4. MSD-DRS Results for Three Real Fruit Extract Samples Sample 4: Grapes 0.68 mg/Kg Captan 0.21 mg/Kg Cyprodinil 0.27 mg/Kg Fludioxinil Diphenylamine Sample 5: Orange 2.5 mg/Kg Imazalil 0.25 mg/Kg Medidathion 3.0 mg/Kg Thiabendazole Sample 6: Apple 0.86 mg/Kg Diphenylamine 0.05 mg/Kg Chlorpyrifos 0.79 mg/Kg Thiabendazole Dimethoate Ethoxyquin Methyl parathion Endosulfan sulfate Propargite

2. Control - Lettuce spiked extract This control sample was spiked with 33 pesticides at levels ranging between 0.02 and 0.20 mg/kg. Twenty-nine pesticides were detected and confirmed by the DRS software, three were not reported since they are not present in the Agilent RTL Pesticide database and one was not detected. The one undetected analyte, (Folpet, marked with two asterisks in Table 3), was detected and confirmed if a higher sensitivity setting was used in the AMDIS deconvolution program. 3. Control - Apple spiked extract This control sample was spiked with 30 pesticides at levels ranging between 0.01 and 0.20 mg/kg. Twenty-two pesticides were detected and confirmed by the DRS software, six were not reported since they are not present in the Agilent RTL Pesticide database and two were not detected. Overall, of the 95 spiked analytes in the three control samples, 93% of the pesticides present in the Agilent RTL Pesticide database were detected and confirmed by full-scan library searching of the deconvoluted mass spectra. 4. Real Grape extract The customer had detected and confirmed three pesticide residues in the Grape extract sample Captan, Cyprodinil, and Fludioxinil. Of these three analytes, Captan was confirmed by the DRS and Cyprodinil and Fludioxinil are not entries in the Agilent RTL Pesticide database. However, DRS also confirmed an additional pesticide residue Diphenylamine, which was not reported by the customer. 5. Real Orange extract The customer had detected and confirmed three pesticide residues in the Orange extract sample Imazilil, Methidathion, and Thiabendazole. All three of these pesticides were confirmed by the DRS software and no other analytes were confirmed.

Discussion
1. Control - Orange spiked extract This control sample was spiked with 32 pesticides at levels ranging between 0.02 and 0.10 mg/kg. Twenty-six pesticides were detected and confirmed by the DRS software, two were not reported since they are not present in the Agilent RTL Pesticide database and four were not detected. The spiking was done to the raw matrix, not to a matrix extract. For the polar pesticides (methamidofos and acephate), the recovery was in the 20%30% range as confirmed by NPD/ECD. Therefore, that explains why these pesticides were not detected by DRS.

www.agilent.com/chem
6. Real Apple extract The customer had detected and confirmed three pesticide residues in the Apple extract sample Diphenylamine, Chlorpyriphos, and Thiabendazole. All three of these pesticides were confirmed by the DRS. In addition, the DRS also confirmed the presence of five additional pesticide residues Dimethoate, Ethoxyquin, Methyl Parathion, Endosulfan Sulfate, and Progargite. These five pesticides had not been reported by the customer. The extensive data shown in this report, run under totally blind conditions, shows the high degree of confidence that an analyst can have in the results produced by the DRS in minutes.

Reference
1. Philip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD Using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN, www.agilent.com/chem

Conclusions
The Agilent Technologies MSD-DRS provides additional powerful data processing capabilities to the MSD ChemStation software. Reviewing full scan GC/MS data for the confirmation of pesticide residues can be a labor-intensive and time consuming process requiring great skill and concentration by an experienced analyst. The DRS is able to process a complex food extract TIC in the order of 1 minute, whereas an experienced analyst may take more than 30 minutes to achieve the same quality result. The DRS software was proven to report the lowest number of false positives and false negatives in the shortest time period. In scan mode, the detection limit is not as low as in selected ion monitoring (SIM) mode; however, any prior knowledge of the target analytes (retention times or characteristic ions) is not required for the DRS.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA October 5, 2004 5989-1654EN

Analysis of Organochlorine and Pyrethroid Pesticides with Agilent 6820 Gas Chromatograph/Micro-Electron Capture Detector Application

Environmental and Food Analysis

Author
Chuanhong Tu Agilent Technologies Co., Ltd. (Shanghai) 412 YingLun Road Waigaoqiao Free Trade Zone Shanghai, 200131 P. R. C.

To address these problems, the ECD, developed by Agilent Technologies, uses a smaller flow cell. The ECD is optimized for capillary columns and designed for improved sensitivity. It was successfully used with an Agilent 6890 series GC with better detector sensitivity and a wider linear range. In this note, the Agilent 6820 GC with ECD was used to determine organochlorine and pyrethroid pesticides following the Chinese National Standard Method GB/T 5009.146-2003 [2]. System sensitivity and limits of detection (LOD) were examined for organochlorine and pyrethroid compounds.

Abstract
The Agilent 6820 gas chromatograph (GC) with microelectron capture detector (ECD) was used to analyze organochlorine and pyrethroid compounds. All compounds demonstrated good linearity among a wide concentration range. The sensitivity provided by ECD was much better than the requirements of the routine pesticide residue analysis.

Experimental
All experiments were performed on an Agilent 6820 GC with split/splitless inlet and ECD. Single-tapered deactivated liner (p/n 5183-4696) and Agilent green septa (p/n 5183-4759) were used. Cerity Networked Data System (NDS) software was used for instrument control, signal acquisition, and data processing. Samples were manually introduced into the GC with a 10-L micro-syringe (p/n 5182-3428). Experimental conditions are listed in Table 1. All organochlorine and pyrethroid compounds were diluted with hexane.

Introduction
The electron capture detector (ECD) is a type of detector with high sensitivity and selectivity for halogenated compounds. However, there are some drawbacks in ECD design. The ECD is inherently nonlinear, with a limited linear range. Due to the narrow linear range, sample concentration or dilution, and re-analysis have to be employed, resulting in lower productivity. In addition, in traditional ECD design, a large flow cell is necessary to be compatible with both packed and capillary columns, leading to lower detector sensitivity [1].

Table 1. Instrumental Parameters Instrument Agilent 6820 GC Software Inlet Injection volume Column Carrier Oven Cerity NDS Chemical for QA/QC Split/Splitless; 250 C; splitless mode; purge time: 0.75 min 2 L HP-1, 30 m 0.32 mm 0.25 m (p/n 19091Z-413) Nitrogen, 6.0 psi, constant head pressure mode, 1.2 mL/min (60 C) 60 C (1 min), 30 C/min to 180 C, 5 C/min to 250 C (5 min), 3 C/min to 280 C (10 min) ECD; 330 C; make-up: nitrogen, 60 mL/min

Results
ECD Sensitivity A chromatogram of organochlorine and pyrethroid pesticides on an HP-1 column is shown in Figure 1. The concentrations, for four benzene hydrochlorides (BHCs), heptachlor, aldrin, heptachlor epoxide, and six pyrethroids are 10 ppb; for p,p'-DDE, dieldrin, endrin, endosulfan I, and endosulfan II, 20 ppb; and for p,p'-DDD, endrin aldehyde, endosulfan sulfate, and p,p'-DDT, 60 ppb. Except for endrin and endosulfan II, the other 20 compounds were fully separated. Among pyrethroid pesticides, two permethrin, four cypermethrin and two fenvalerate isomers were separated.

Detector

Hz 1400

Peak identification
1200 1. 2. 3. 4. 5. 6. 7. 8. 3 600 4 5 400 2 6 7 -BHC -BHC -BHC -BHC Heptachlor Aldrin Heptachlor epoxide Endosulfan I 9. 10. 11. 12. 13. 14. 15. 16. p,p'-DDE Dieldrin Endrin Endosulfan II p,p-DDD Endrin aldehyde Endosulan sulfate p,p'-DDT 8 10 12 11 14 13 15 16

1000

800

200

10 Hz 240

12

14

16

18

20

min

18 17. 18. 19. 20. 21. 22. 17 Fenpropathrin Cyhalothrin Permethrin Cypermethrin Fenvalerate Deltamethrin

220

200

180 19 160

20

21 22

140

120 25 27.5 30 32.5 35 min

Figure 1. 2

Chromatogram of organochlorine and pyrethroid pesticides on HP-1 column.

The signal-to-noise ratios are larger than 20 for organochlorine pesticides at concentrations of 16 ppb. For 5 ppb pyrethroids, the signal-to noise ratios are larger than 10. They can be easily quantitated. Therefore, the ECD provides more than enough sensitivity to meet the requirements of quantitative analysis of pesticides residues. Linear Range and Response Factors The calibration curves of -BHC and permethrin, typical of organochlorine and pyrethroid pesticides, are shown in Figures 2 and 3, respectively. The linear range and response factors (RFs) are listed in Table 2. The RFs are the ratios of compound concentrations to peak areas. The relative standard deviations (RSDs) of RFs are less than 20%, better than the precision requirements for response factors in the contract laboratory program of the USEPA (the United States Environmental Protection Agency). The linear correlation coefficients for all compounds are better than 0.995.

pA 60000

40000

-BHC y = 119.98x + 354.05 R2 = 0.9988

20000

0 0 100 200 Concentration, ppb 300 400

Figure 2.
pA 4000

Calibration curve of -BHC, a typical organochlorine pesticide.

3000

Permethrin y = 9.0737x + 97.257 R2 = 0.9986

2000

1000

0 0 100 200 Concentration, ppb 300 400

Figure 3.

Calibration curve of permethrin, a typical pyrethroid.

www.agilent.com/chem
Table 2. Compound -BHC -BHC -BHC -BHC Heptachlor Aldrin Heptachlor epoxide Endosulfan I p,p'-DDE Dieldrin Endrin Endosulfan II p,p'-DDD p,p'-DDT Endrin aldehyde Endosulfan sulfate Fenpropathrin Cyhalothrin Permethrin Cypermethrin Fenvalerate Deltamethrin Linearity Results for Organochlorine and Pyrethroid Pesticides Average RF 0.0064 0.0159 0.0078 0.0087 0.0097 0.0078 0.0100 0.0109 0.0088 0.0123 0.0152 0.0121 0.0379 0.0175 0.0139 0.0140 0.0500 0.0225 0.1007 0.0809 0.0728 0.0461 %RSD of RF 7.5 18.1 9.4 8.8 13.2 8.0 11.7 11.7 11.0 13.0 16.3 15.8 8.7 15.8 10.6 10.0 18.0 8.6 10.2 7.2 6.9 15.0 Linear range (ppb) 1400 1400 1400 1400 1400 1400 1400 2800 2800 2800 2800 2800 62400 62400 62400 62400 1400 1400 10400 10400 10400 10400 R2 0.9983 0.9984 0.9988 0.9986 0.9987 0.9982 0.9983 0.9983 0.9948 0.9981 0.9991 0.9987 0.9983 0.9972 0.9989 0.9988 0.9951 0.9982 0.9986 0.9991 0.9990 0.9996

Conclusion
The Agilent 6820 GC/ ECD system shows good sensitivity and wide linear range for organochlorine and pyrethroid pesticides, and are much better than routine pesticide residue analysis requirements.

References
1. Channel, I., and Chang, I. L., Analysis of Organochlorine Pesticides and PCB Congeners with the Agilent 6890 Micro-ECD, Agilent Technologies, Publication (23) 5965-8556E www.agilent.com/chem 2. China National Standard Method GB/T 5009. 146-2003, Multiresidue analytical methods for organochlorine and pyrethroid pesticides for plant-originated food, August, 2003
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA August 13, 2004 5989-1333EN

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Analysis of Organophosphorus Pesticides with Agilent 6820 Gas Chromatograph/ Nitrogen Phosphorus Detector Application

Environmental and Food Analysis

Author
Chuanhong Tu Agilent Technologies Co., Ltd. (Shanghai) 412 YingLun Road Waigaoqiao Free Trade Zone Shanghai, 200131 P. R. C.

Abstract
Fifteen of the most common and important organophosphorus pesticides (OPs) were studied using the Agilent 6820 gas chromatograph (GC) equipped with the new nitrogen-phosphorus detector (NPD). The 6820 GC-NPD demonstrated good linearity for concentrations of pesticides range from 1 to 500 ng/mL (R2 >0.999) for most compounds. All OPs produced high signal-to-noise ratios for splitless injections at 10 ng/mL (ppb) concentrations with the NPD detector. Instrumental limits of detection for most OPs studied were at low or sub-ppb concentrations. This suggests the 6820 GC with an NPD is well suited for OP pesticide residue determinations in foods, water, or other samples.

food safety and the identities and residual concentrations of pesticides due to their stability, inappropriate or illegal usage. In China, organophosphorus pesticides (OPs) account for 70% of the total amount of the pesticides used [1]. Maximum residue levels (MRLs) have been set up for 27 OPs in different kinds of food, and analytical methods have been developed for their analysis [2]. The China National standard method GB/T 5009.1452003 is a method for the determination of 16 organophosphorus and 4 carbamate pesticides by GC-NPDs [3]. In this application note, the Agilent 6820 GC equipped with an NPD is employed to determine 15 of the most common OPs of concern.

Experimental
The experiments were carried out on an Agilent 6820 GC with split/splitless inlet and NPD. De-activated liners for splitless injection (p/n 5183-4696) were used to improve the inertness of the system; the septa were Agilent green septa (p/n 5183-4759). Cerity Networked Data System (NDS) for Chemical QA/QC software was used for instrument control, data collection, and data processing. The sample was introduced manually with a 10-L syringe (p/n 5182-3428). All target compounds were dissolved in acetone. The experimental conditions are listed in Table 1. All flows were set using the Veriflow-500 digital flowmeter (p/n HVF-500-2).

Introduction
Synthetic organic pesticides are widely used in modern agriculture to protect crops and improve production. The green revolutions of many countries are obtained through the application of these compounds. There are increasing concerns over

Table 1. Software Inlet

Instrumental Parameters Cerity NDS for chemical QA/QC Split/Splitless inlet 250 C Splitless 1 L 0.75 min HP-5ms, 30 m 0.32 mm 0.25 m (p/n 19091S-413) He, head pressure: 12 psi, 2.5 mL/min at 60 C. 60 C for 1 min, to 200 C at 10 C/min, to 250 C at 5 C/min, 5 min hold. NPD at 325 C with white rubidium bead (p/n G1534-60570) H2: 3 mL/min; Air: 60 mL/min; makeup N2: 10 mL/min

Inlet temperature Injection mode Injection volume Purge time Column Carrier gas Oven temperature Detector Detector gases

by 1-L injections of standards at 1, 10, 50, 100, and 500 ng/mL concentrations were linear with R2 >0.999 for most compounds. The calibration curve for dichlorvos, a typical OP, is shown in Figure 2.

Results and Discussions


The separation of 15 OPs is illustrated in Figure 1. The compounds, except for chlorpyrifos and parathion, were well separated with the HP-5ms column. Peak identifications are listed in Table 2. Calibration curves constructed from data obtained

Table 2. Retention Times of Target Pesticides Peak number Compound Retention time 1 Methamidophos 8.26 2 Dichlorvos 8.63 3 Acephate 11.21 4 Monocrotophos 14.44 5 Phorate 14.60 6 Dimethoate 15.00 7 Parathion-methyl 17.03 8 Fenitrothion 17.75 9 Malathion 18.04 10 Fenthion 18.27 11 Parathion 18.34 12 Chlorpyrifos 18.34 13 Methidathion 19.97 14 Ethion 22.52 15 Triazophos 22.92

pA 600 11, 12

400 2 1 200 3 4 5

6 7 10 8 9 13 15 14

0 10 15 20 min

Figure 1.

Chromatogram of 15 OPs at 1-ppm using the NPD.

pA 800

600

Dichlorvos y = 1.3849x + 3.3834 R2 = 0.9999

400

200

0 0 100 200 300 Concentration, ppb 400 500 600

Figure 2.

Calibration curve for dichlorvos, a typical OP.

Approximate Instrumental Limits of Detection The chromatogram for 10-ppb pesticides using the NPD is shown in Figure 3. All compounds are easily quantitated. Acephate, with the lowest response factor, provided around a 30 ratio of signal to noise. In fact, most compounds at 1 ppb show good peaks except methamidophos, acephate, and monocrotophos. The limits of detection (LODs) are much lower than the maximum residue levels (MRLs) for the OPs.
11, 12 10 9
50

pA

52

14

7 6 5

13 15

48

2
46

44

3 1
42

40

7.5

10

12.5

15

17.5

20

22.5

min

Figure 3.

Chromatogram of 10-ppb OPs using the Agilent 6820 GC/NPD.

www.agilent.com/chem
Sample Solvent Precautions In some sample extraction procedures, organic solvents, containing elements with high electronegativity, (such as methylene chloride), may be used. However, these solvents may cause baseline shifts, change the detector sensitivity, shorten lifetime of the rubidium bead, or even make the detector unusable. This is true for all types and vendors of NPD [4]. If the final solution for injection consists of methylene chloride or chloroform, it is best to change the solvent to acetone or hexane. visit our Web site at www.agilent.com/chem.

Conclusion
The Agilent 6820 GC with NPD can be used for the sensitive and selective determination of OPs. The NPD detector provides good linearity for most of these compounds in the 1500-ppb range.

References
1. Xingui Sun et al, Survey of organophosphorus pesticide residues in vegetables and fruits in Beijing, (2003) Chinese Journal of Food Hygiene, 15, No. 6, 536-538. 2. Jiming Ye et al, Introduction of maximum residues limits in China, (2000) Pesticide Science and Management. 3. China National Standard Method GB/T 5009.145-2003, Determination of organophosphorus and carbamate pesticides multiresidues in vegetable foods. 4. Kai Meng, Yeugeny Kaplun, Rich White, The New Features of the HP 6890 NitrogenPhosphorus Detector to Deal with Hostile Solvents, Agilent Technologies, publication (23) 5963-6808E. www.agilent.com/chem

For More Information


For more information on our products and services,
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA August 13, 2004 5989-1335EN

Analysis of Carbamate Pesticides Using Agilent 6820 Gas Chromatograph/Nitrogen Phosphorus Detector Application

Environmental and Food Analysis

Authors
Weijun Yao, Chuanhong Tu Agilent Technologies Co., Ltd. (Shanghai) 412 YingLun Road Waigaoqiao Free Trade Zone Shanghai, 200131 P. R. C.

because of the excellent separation capability of the capillary column. In this note, the Agilent 6820 GC/NPD is used to determine seven common carbamates including metolcarb, isoprocarb, baycarb, propoxur, carbofuran, primicarb, and carbaryl.

Experimental
All experiments were performed on an Agilent 6820 GC with NPD, deactivated splitless inlet liner (p/n 5183-4696), and Agilent advanced green septa (p/n 5183-4759). Agilent Cerity Networked Data System (NDS) is used for instrument control and data acquisition. Instrumental parameters are listed in Table 1. The standard pesticides were dissolved in acetone solution. Samples were injected manually using a 10-L syringe (p/n 5182-3428).
Table 1. Instrumental Parameters Instrument Agilent 6820 GC Software Cerity NDS for chemical QA/QC Inlet Split/Splitless inlet Inlet temperature 250 C Injection mode Splitless Injection volume 1 L Purge time 0.75 min Column HP-5ms, 30 m 0.32 mm 0.25 m (p/n 19091S-413) Carrier gas Nitrogen, constant pressure of 5 psi, 1.0 mL/min (50 C) Oven temp 50 C (1 min) Ramp 1: 20 C/min to 100 C Ramp 2: 5 C/min to 150 C (5 min) Ramp 3: 10 C/min to 200 C (10 min) Detector NPD, 325 C, white bead (p/n G1534-60570) H2, 3 mL/min Air, 60 mL/min Detector gas Makeup, N2, 10 mL/min

Abstract
The Agilent 6820 gas chromatograph/nitrogen phosphorus detector (GC/NPD) was employed to determine carbamate pesticides, following the China National Standard Method GB/T 5009.104-2003. Seven common carbamates were fully separated using an HP-5ms column, 30 m 0.32 mm 0.25 m. Each compound showed good linearity in the 101000-ppb range, with a detection limit lower than 10 ppb, which is 101000 times lower than the maximum allowable residue level.

Introduction
The nitrogen phosphorus detector (NPD) is both sensitive and selective for organic compounds containing nitrogen and phosphorus, and is widely used in the trace analysis of organonitrogen or organophosphorus pesticides. The carbamates are types of organic synthetic pesticides, widely used to prevent plant diseases. The application of these pesticides may leave residues in agricultural products that may pose potential human health risks via the food chain. The Chinese government publishes maximum residue limits for the pesticides carbofuran, primicarb, and carbaryl in food. GB/T 5009.104-2003 (originally GB14877-1994) is a method used to determine six carbamate residues based on GC/NPD using a packed column [1]. In recent years, capillary-column methods have gradually replaced packed-column methods

Results and Discussions


All compounds were separated with a HP-5ms column. The resolution of baycarb and propoxur is 1.0, although their structures are very similar. The chromatogram of these pesticides is shown in Figure 1. The peak areas are linear with concentrations for all carbamates in the range of 101000 ppb. Typical calibration curves for carbofuran and primicarb are shown in Figure 2.
pA 90

80

70

Peak identification 5. Carbofuran 1. Metolcarb 6. Primicarb 2. Isoprocarb 7. Carbaryl 3. Baycarb 4. Propoxur 1

60

50

3 4 2

5 7

40

30

20 16 18 20 22 24 26 min

Figure 1.

Chromatogram of 1-ppm carbamates using NPD detector.

pA 140

120

100

Primicarb y = 0.1277x + 0.5557 R2 = 0.9992

80

60

40

Carbofuran y = 0.0706x + 0.7961 R2 = 0.9969

20

0 0 200 400 600 Concentration, ppb 800 1000 1200

Figure 2.

Calibration curves for primicarb and carbofuran.

The chromatogram for the 10-ppb carbamate standard is shown in Figure 3. The signal to noise ratios are between 612 for the 10-ppb carbamates. The limits of detection are 46 ppb; these are 10 to 1000 times lower than the maximum allowable residue limits stipulated in China National Standard (GB14928.2-94, GB14928.7-94, and GB14971-94), which is 50-ppb primicarb and 5000-ppb carbaryl in cereals [2, 3, 4]. The NPD can easily meet these requirements for the routine analysis of carbamate pesticide residues.
pA

6
30.6

30.5

30.4

4 1 3 2

30.3

30.2

30.1

30

16

18

20

22

24

26

min

Figure 3.

Chromatogram of 10-ppb carbamates standard with 1-L injection.

Conclusion
The Agilent 6820 GC with NPD detector can be used for the sensitive and selective measurement of carbamate pesticide residues in food. The detection limit is much lower than the maximum residue limits. Good linearity was obtained in the range of 101000 ppb.

References
1. GB/T 5009.104-2003, Determination of carbamate pesticide residues in plant-originated foods. 2. GB14928.2-94 Maximun residue limits of primicarb in food. 3 GB14928.7-94 Maximun residue limits of carbofuran in rice grains. 4. GB14971-94 Maximun residue limits of carbaryl in food.
3

www.agilent.com/chem

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA August 11, 2004 5989-1334EN

Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software Application

Food Safety

Author
Philip L. Wylie, Michael J. Szelewski, and Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA Christopher P. Sandy Agilent Technologies, Inc. Block A, CSC Eskdale Road Winnersh Triangle Estate Wokingham, Berk RG41 5DZ United Kingdom

pesticides in food and environmental samples to track their distribution in the environment and to ensure a safe food supply. Current analytical methods target only a subset of the possible compounds. Whether for food or environmental samples, analyses are often complicated by the presence of co-extracted natural products. Food or tissue extracts can be exceedingly complex matrices that require several stages of sample cleanup prior to analysis [3]. Even then, it can be difficult to detect trace levels of contaminants in the presence of the remaining matrix. For efficiency, multiresidue methods (MRMs) must be used to analyze for most pesticides. Traditionally, these methods have relied upon gas chromatography (GC) with a constellation of element-selective detectors to locate pesticides in the midst of a variable matrix [4, 5, 6]. GC with mass spectral detection (GC/MS) has been widely used for confirmation of hits. Liquid chromatography (LC) has been used for those compounds that are not amenable to GC [2]. Today, more and more pesticide laboratories are relying upon LC with mass spectral detection (LC/MS) and GC/MS as their primary analytical tools [7, 8]. Still, most MRMs are target compound methods that look for a small subset of the possible pesticides. Any compound not on the target list is likely to be missed by these MRMs. Using the techniques of retention time locking (RTL) [9, 10, 11] and spectral deconvolution [12], a method has been developed to screen for 567 pesticides and suspected endocrine disrupters in a single GC/MS analysis. Spectral deconvolution

Introduction
According to The Pesticide Manual, more than 700 pesticides are currently approved for use around the world [1]. About 600 more were used in the past, but are either banned or no longer marketed. In spite of their discontinuance, some of these still persist in the environment where they may bioaccumulate in the flora and fauna. Many pesticides or their degradation products can be found at trace levels in food and beverages; in soil, water, and air; in aquatic and terrestrial flora and fauna; and in human blood, adipose tissue, and breast milk. The World Health Organization has classified pesticides into five groups based upon their acute toxicity to humans [2]. The categories range from Acutely Hazardous to those that are Unlikely to Present Acute Hazard in Normal Use. Certain pesticides are classified as persistent organic pollutants (POPs), carcinogens, teratogens, or endocrine disrupters. It is now common to analyze for

helps to identify pesticides even when they are buried under co-eluting matrix compounds. RTL helps to eliminate false positives and gives greater confidence in the results. Users can easily add compounds to the method if they wish.

Samples Vegetable extracts were obtained from Dr. Mark Lee and Stephen Siegel at The California Department of Food and Agriculture (CDFA; Sacramento, CA USA) and from Dr. J.G.J. Mol at TNO Nutrition and Food Research (Zeist, The Netherlands). Seventeen data files from the GC/MS analysis of surface water samples were also contributed by CDFA and were processed in this laboratory using the Deconvolution Reporting Software (DRS). GC/MS data files (locked to the Agilent Pesticide Method) for 17 crop extracts were supplied by NRM Laboratories, Berkshire, UK.

Experimental
Table 1 lists the instrumentation, software, and analytical parameters used by Agilent for pesticide analysis. Depending upon the desired injection volume, a PTV inlet or split/splitless inlet can be used.

Table 1.

Instrumentation and Conditions of Analysis Agilent 6890N Agilent 7683 Agilent PTV operated in the solvent vent mode Agilent 30 m 0.25 mm 0.25 m HP-5MS (p/n 19091S-433) Helium in the constant pressure mode Chlorpyrifos-methyl locked to 16.596 min (nominal column head pressure = 17.1 psi) 70 C (2 min), 25 C/min to 150 C (0 min), 3 C /min to 200 C (0 min), 8 C /min to 280 C (1015 min) Temp program: 40 C (0.25 min), 1600 C/min to 250 C (2 min); Vent time: 0.2 min; Vent flow: 200 mL/min; Vent pressure: 0.0 psi; Purge flow: 60.0 mL/min; Purge time: 2.00 min 15 L (using a 50-L syringe) Agilent 5973 inert 50550 amu 230, 150, and 280 C, respectively 4.00 min Autotune voltage

Gas chromatograph Automatic sampler Inlet Column Carrier gas RTL Oven temperature program PTV inlet parameters Injection volume Mass Selective Detector (MSD) Scan range Source, quad, transfer line temperatures Solvent delay Multiplier voltage Software GC/MSD ChemStation Deconvolution Reporting Software (DRS) Library searching software Deconvolution software MS Libraries

Agilent p/n G1701DA (Version D01.00 sp1) Agilent p/n G1716AA NIST MS Search (version 2.0) (included with NIST '02 mass spectral library, Agilent p/n G1033A) Automated Mass Spectral Deconvolution and Identification Software (AMDIS) (included with NIST '02 mass spectral library, Agilent p/n G1033A) NIST '02 mass spectral library (Agilent p/n G1033A); Agilent RTL Pesticide Library (p/n G1049A)

Results and Discussion


RTL and RTL Databases RTL is a technique developed by Agilent that allows users to match analyte retention times (RTs) on any Agilent 6890 GC, in any laboratory in the world, so long as the same nominal GC method and capillary column are used [13]. Using RTL, Agilent has developed several retention-timelocked databases for GC and GC/MS that include the locked retention time, compound name, CAS number, molecular formula, molecular weight, and mass spectrum (GC/MS databases only) for each entry [14]. The Agilent RTL Pesticide Library contains this information for almost all GC-amenable pesticides, as well as several endocrine disrupters - 567 compounds in all. For use with the DRS discussed below, this library was converted into the NIST format [15]. Separate Automated Mass Spectral Deconvolution and Identification Software (AMDIS) libraries for the RTs and compound information were created from the original RTL Pesticide Library. Users can easily augment these libraries with newer pesticides or other compounds of interest [15].

Basics of Deconvolution In GC/MS, deconvolution is a mathematical technique that separates overlapping mass spectra into cleaned spectra of the individual components. Figure 1 is a simplified illustration of this process. Here, the total ion chromatogram (TIC) and apex spectrum are shown. As is often the case, the peak is composed of multiple overlapping components and the apex spectrum is actually a composite of these constituents. A mass spectral library search would give a poor match, at best, and certainly would not identify all of the individual components that make up the composite spectrum. The deconvolution process finds ions whose individual abundances rise and fall together within the spectrum. In this case, it first corrects for the spectral skew that is inherent in quadrupole mass spectra and determines a more accurate apex RT of each chromatographic peak. As illustrated in Figure 1, deconvolution produces clean spectra for each overlapping component. These individual spectra can be library searched with a high expectation for a good match. The AMDIS that is incorporated into the Agilent DRS is supplied by the National Institute of Science and Technology (NIST) [12].

TIC and spectrum

Deconvoluted peaks and spectra

TIC Component 1 Component 2 Component 3 Matrix

Deconvolution
Interference

Target

Figure 1.

An illustration of mass spectral deconvolution process. 3

DRS Agilent's DRS results from the combination of three different GC/MS software packages: 1) the Agilent GC/MS ChemStation, 2) the NIST Mass Spectral Search Program with the NIST '02 MS Library, and 3) the AMDIS software, also from NIST. Included in the DRS, are mass spectral and locked RT libraries for 567 pesticides and suspected endocrine disrupters. Three separate, but complimentary, data analysis steps are combined into the DRS. First, the GC/MS ChemStation software performs a normal quantitative analysis for target pesticides using a target ion and up to three qualifiers. An amount is reported for all calibrated compounds that are detected. For other compounds in the database, an estimate of their concentration can be reported based upon an average pesticide response factor

(RF) that is supplied with the DRS software. The DRS then sends the data file to AMDIS, which deconvolutes the spectra and searches the Agilent RTL Pesticide Library (in AMDIS format) using the deconvoluted full spectra. A filter can be set in AMDIS, which requires the analyte's RT to fall within a user-specified time window. Because RTL is used to reproduce the RTL database RTs with high precision, this window can be quite small typically 20 seconds or less. Finally, the deconvoluted spectra for all of the targets found by AMDIS are searched against the 147,000-compound NIST mass spectral library for confirmation; for this step, there is no RT requirement. Once the appropriate method is loaded, the DRS report can be generated with a single mouse click as shown in Figure 2. The software can run automatically after each analysis or at a later time on a single file or a batch of files.

Figure 2.

ChemStation pull down menu showing options for running the DRS on single or multiple files.

Pesticides in an Herbal Mix Figure 3 shows a TIC from the extract of an herbal mix. Figure 4 shows the MSD Deconvolution Report for this sample, which is produced in html format so it can easily be emailed or copied into a spreadsheet. This sample was chosen because herbs are among the most difficult vegetable products to analyze. Their extracts contain a large number of natural products that interfere with pesticide analysis.

450000 400000 350000 Abundance 300000 250000 200000 150000 100000 50000 5.00 10.00 15.00

TIC: MOL_4A.D

20.00

25.00 Time

30.00

35.00

40.00

45.00

Figure 3.

TIC of an herbal mix.

Figure 4.

MSD Deconvolution Report generated for the herbal mix extract shown in Figure 3.

The DRS report in Figure 4 lists the RT, CAS number, and compound name for each hit. Phenanthrene-d10, listed at the bottom of the report, is the internal standard (ISTD) used by the ChemStation to estimate the quantity of each compound that it found. Since an average pesticide response factor was used for all 567 target compounds, the amounts listed in column 4 are only estimates. Experience has shown that most estimates reported using an average pesticide response factor fall within a factor of 10 of their actual values. True quantitation requires calibration with pesticide standards in the normal way, but this is not practical for all of the pesticides in the database. A laboratory would normally generate calibration curves for their target set of pesticides and use the average RF for the remaining compounds in the database. In this way, when a new compound is detected, the lab can immediately get a rough estimate of its concentration and decide if it should be added to the calibration list. Column 5 in the report shows the match factor obtained through AMDIS deconvolution and RTL Pesticide Library searching using the deconvoluted full spectra. In this case, several more targets were identified by AMDIS than were found by the ChemStation software (for example, Prometon and p,p-DDE), which is typical for complex samples. When locked RTs are available, it is a significant advantage to set a RT requirement in the AMDIS software. In this case, hits that did not fall within 10 seconds of the database RT were eliminated. Column 6 shows the RT difference (in seconds) between the compound's library RT and its actual value in the chromatogram. Figure 4 shows that the software identified two phthalates (suspected endocrine disrupters) in addition to the pesticides. Phthalates are ubiquitous in the environment and are extremely difficult to remove from the background. In this case, no attempt was made to determine if the phthalates were actually extracted from the sample or were introduced in the laboratory. The last two columns in the DRS report show the results from searching all of the AMDIS hits against the NIST 147,000-compound mass spectral library. When the NIST library search finds a compound in the top 100 matches (a user-settable value) that agrees with the AMDIS results, its match factor is listed in column seven. The hit number is shown in the last column, with 1 being the best match (highest match factor) in the NIST database. Occasionally, the NIST library search does not find the AMDIS hit among the top

100 spectral matches. In this case, the next line in the report shows the best library match for that spectrum. This is evident for fluvalinate-tau-I (Figure 4), which eluted at 34.779 min. The next line shows the best NIST library match for that spectrum - fluvalinate. In this case, no compound with the same CAS number as fluvalinate-tau-I is contained in the NIST mass spectral library. In fact, fluvalinate-tau-I is the D isomer, while fluvalinate is the DL isomer mixture. Blind Comparison Between DRS and Traditional Data Review Many comparisons have shown that the DRS is much better than conventional methods at identifying target compounds in complex samples, such as food and environmental extracts. Two such studies are described here. In the first case, 17 unspiked crop samples were analyzed by NRM Laboratories in Berkshire, UK using Agilent's RT-locked pesticide method. The data files, but not their list of pesticide hits, were sent to Agilent for analysis using the new DRS. Table 2 shows a comparison of the results from the two laboratories. Using manual data review, NRM identified 28 pesticides in the 17 samples, four of which were below their lowest calibration level. Using the same data files, the DRS identified 33 pesticides. Agilent's automated method did not identify azoxystrobin in the spring onion sample because it is not included in the RTL pesticide library. While it can be found in the NIST library, it has a molecular ion at 403 amu and method used at NRM only scanned to 400 amu. The DRS method confirmed all four pesticides that were below the NRM calibration range and found five more (terbacil, pyrimethanil, methiocarb, pyridaben, and propamocarb) that were not included in their method. The agreement between the manual and automated methods was excellent. However, the DRS looks for many more pesticides and was able to find several that were missed by the manual method. In addition, manual data review took a chemist about 7 hours for the 17 samples while the DRS finished the task in 50 minutes of unattended computer time.

Table 2.

A Comparison of the Pesticides Found in 17 Unspiked Crop Samples Using Conventional Data Review and Agilent's DRS. Pesticides that Were Found by Only One Method Are Underlined Agilent DRS results* Propyzamide Chlorthal-dimethyl p,p'-DDE Terbacil Pirimicarb Chlorthal-dimethyl Propyzamide Pyrimethanil Pirimicarb Metalaxyl Iprodione Not in DRS library Methiocarb Iprodione Procymidone Pyridaben Propamocarb Procymidone Buprofezin Endosulfan sulfate Iprodione Chlorthal-dimethyl Iprodione Dimethoate Metalaxyl Procymidone Terbuconazole Omethoate Procymidone lamda-Cyhalothrin Chloropropham Pirimicarb Pirimicarb NRM Manual Analysis** Propyzamide Chlorthal-dimethyl p,p'-DDE Not found*** Pirimicarb Chlorthal-dimethyl Propyzamide Not found*** Pirimicarb Metalaxyl Iprodione Azoxystrobin Not found*** Iprodione Procymidone Not found*** Not found*** Procymidone Buprofezin Endosulfan sulfate Iprodione Chlorthal-dimethyl Iprodione Dimethoate Metalaxyl Procymidone Terbuconazole Omethoate Procymidone lamda-Cyhalothrin Chloropropham Pirimicarb Pirimicarb

Sample Coriander

Rosemary

Spring Onion

Chives Cherry Tomato Courgette Aubergine

Flat Leaf Parsley Lambs Lettuce Cos Lettuce

Fine Endive Red Potato Fine Endive


* **

Pesticides found by re-analyzing NRM datafiles using Agilent's DRS software. Pesticides found by NRM using target compound analysis and manual verification.

*** This compound was not in the NRM target compound list.

This compound is not included in the Agilent RTL Pesticide Library or the DRS software. Found by the Agilent ChemStation but not found by AMDIS or NIST library searching after deconvolution. After careful review of this hit, omethoate was judged not to be in the sample. Compound was detected but was below the calibration range.

Analysis of Surface Water Samples: In another study, the CDFA analyzed 17 surface water extracts for pesticides. TICs for two typical samples are shown in Figure 5. The CDFA used RTL and RTL database searching but without the benefit of spectral deconvolution. The same data files were then analyzed using the DRS for comparison. Table 3 shows the results from the CDFA manual analysis of the 17 samples compared to results using the DRS. The CDFA found 38 pesticide hits in the 17 samples, some of which were for the same pesticide in multiple samples. It took a skilled analyst about 8 hours to review the results, eliminate false positives, and verify all of the hits. The DRS found 37 of the compounds seen by the CDFA and identified one CDFA hit as a false positive. In addition, 34 more pesticides were

found for a total of 71 hits in the 17 samples. The process was fully automated and took about 20 minutes of unattended computer time to process all of the data files.
Table 3. A Comparison of Results from the Analysis of 17 Surface Water Samples by GC/MS. The CDFA Used RTL and RTL Database Searching, but No Deconvolution. Agilent's DRS Was Used to Analyze the Same Data Files CDFA Number of pesticide hits Number of false positives Time required for analysis 37 1 ~ 8 hours DRS Same 37 + 34 additional 0 20 minutes

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

Figure 5.

TICs of typical surface water extracts provided by the CDFA.

Conclusions
Agilent's new DRS solution for pesticide analysis offers laboratories a number of real benefits. Ease of use: This software solution is very simple to use and takes no more skill than is needed to operate the 6890N/5973 inert GC/MS system. There is no need for the user to learn about the intricacies of deconvolution or to master a new software package. Automation: The deconvolution report can be generated automatically after each run or a batch of samples can be processed all at once. Time savings: Data review is reduced from hours to minutes. Quality: It produces results with the fewest false positives and false negatives. Reproducibility: Results are not dependent upon the skill or experience of the operator. Accuracy: Comparisons such as those discussed in this application note show that the DRS finds pesticides with greater accuracy than manual methods of data analysis. It is particularly useful for relatively complex samples where co-eluting matrix components might obscure traces of target pesticides. Comprehensive: This method screens for almost all GC-amenable pesticides as well as several suspected endocrine disrupters in a single GC/MS run. With 567 compounds in the method, it is the most comprehensive pesticidescreening tool available. Users can add more compounds to the method as needed. Produces quantitative, semi-quantitative, and qualitative results: All calibrated compounds can be quantified. The concentrations of any other compounds can be estimated using an average pesticide response factor provided with the software. Use of the DRS is not limited to pesticide analysis. Other target compound mass spectral libraries can be converted into the AMDIS format and used with this software. For example, one could use existing libraries for forensic drugs, flavors and fragrances, organic pollutants, etc. Users can even generate their own libraries and use them with the DRS. While not required, it is a big advantage to have an RTL library with locked RTs for each entry, as this will give the fewest false positives.

Acknowledgements
The authors wish to thank Dr. Mark Lee and Stephen Siegel of the California Department of Food and Agriculture, Dr. J.G.J. Mol of TNO Research, The Netherlands, and the management and staff at NRM Laboratories, UK, for their contribution of samples and data.

References
1. C.D.S Tomlin, editor. The Pesticide Manual, 13th edition, British Crop Protection Council, Surry, UK (2003). 2. http://www.who.int/pcs/docs/ Classif_Pestic_2000-02.pdf 3. J. Cook, M.P. Beckett, B. Reliford, W. Hammock, and M. Engel (1999) J. AOAC Int., 82 (6), 14191435. 4. M.A. Luke, J.E. Froberg, and H.T. Masumoto, (1975) J. Assoc. Off. Anal. Chem. 58, 1020-1026. 5. M. Luke, J. Froberg, G. Doose, and H. Masumoto, (1981) J. Assoc. Off. Anal. Chem. 64, 1187-1195. 6. B. McMahon and N. Hardin (1994) Pesticide Analytical Manual, Vol. 1, 3rd Ed., U.S. Food and Drug Administration, Washington, DC. 7. J. Fillion, R. Hindle, M. Lacroux, and J. Selwyn (1995) J. AOAC Int. 78, 1252-1266. 8. J. Fillion, F. Sauv, and J. Selwyn (2000) J AOAC Int., 83, 698-712. 9. P. Wylie and B. Quimby, A Method Used to Screen for 567 Pesticides and Suspected Endocrine Disrupters, Agilent Technologies, publication 5967-5860E www.agilent.com/chem. 10.H. Prest, P. Wylie, K. Weiner, and D. Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the 6890/5973 GC/MSD System, Agilent Technologies, publication 5968-4884E www.agilent.com/chem. 11.K. Weiner and H. Prest, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries, Agilent Technologies, publication 5968-8657E www.agilent.com/chem. 12.National Institute of Standards and Technology, AMDIS Literature and Downloads website: http://www.amdis.net/What_is_AMDIS/ AMDIS_Literature_and_Downloads/ amdis_literature_and_downloads.html.

www.agilent.com/chem
13.V. Giarocco, B. Quimby and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E www.agilent.com/chem. 14.http://www.chem.agilent.com/cag/servsup/ usersoft/main.html#RTL. 15.M. Szelewski and C.K. Meng, Building and Editing RTL Screener Databases and Libraries, Agilent Technologies, publication 5989-0916EN www.agilent.com/chem.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA May 19, 2004 5989-1157EN

Improving the Analysis of Organotin Compounds Using Retention Time Locked Methods and Retention Time Databases Application

Environmental

Authors
Frank David Research Institute for Chromatography Pres. Kennedypark 20, B-8500 Kortrijk, Belgium Pat Sandra University of Gent Krijgslaan 281 S4, B-9000 Gent, Belgium Philip L. Wylie Agilent Technologies 2850 Centerville Road Wilmington, DE 19808-1610 USA

Introduction
For many years, organometal speciation has been an important topic in environmental analysis, primarily due to increasing awareness of the toxicological effects of many organometal compounds. Within the class of organometalics, organotin compounds are probably the most widely spread in the environment due to their use as additives in polymers and in antifouling paints. Organotin compounds degrade in the environment into more polar metabolites [1]. Tributyltin, one of the most frequently used organotin additives (as tributyltinchloride or tributyltinoxide), for instance, degrades into dibutyltin and monobutyltin species. Consequently, a large diversity of organotin compounds can be detected in various environmental samples [2]. More recently, organotin contamination of diapers and printed T-shirts was reported and numerous analyses were performed on different consumer products, including all types of absorbent hygiene products. Different methods were used for the extraction and analysis of organotin compounds in environmental, food, and consumer product matrices. Since the organotin compounds with less than four alkyl groups are very polar, they cannot be analyzed directly by GC and must be derivatized into tetraalkyltin compounds prior to analysis. Initially, most methods were based on extraction with

Abstract
The analysis of organotin compounds is becoming increasingly important in both environmental analysis and in food and consumer product analysis. This application note describes a retention time locked (RTL) gas chromatography/mass spectrometry (GC/MS) method for the analysis of derivatized organotin compounds. Three retention time locked libraries are made available, corresponding to three different derivatization methods. The retention time databases allow easy peak location and identification of the target solutes based on mass spectra and retention times.

tropolone (a complexing agent) and n-hexane, followed by Grignard derivatization and determination with GC-flame photometric detection (FPD) [39]. Recently, in-situ ethylation with sodium tetraethylborate (NaBEt4) [1013] has largely replaced Grignard derivatization. At the same time, mass selective detectors (MSD) and atomic emission detectors (AED) have replaced the FPD as the preferred GC detector for organotin compounds [11,13]. A few years ago, solid phase micro extraction (SPME) in combination with capillary gas chromatography-inductively coupled plasma mass spectrometry (CGC-ICP-MS) was used for the determination of volatile and semi-volatile organometal compounds, resulting in excellent sensitivity and selectivity [14,15]. SPME was performed in the headspace or directly in the aqueous sample using a 100 mm polydimethylsiloxane (PDMS) coated fiber. Using NaBEt4, organotin compounds could be derivatized in-situ and simultaneously extracted into the PDMS phase. More recently, stir bar sorptive extraction (SBSE) using a magnetic stir bar coated with a 0.51 mm PDMS layer was developed [16]. After extraction, the solutes were thermally desorbed online to GC/MS, GC-AED or GC-ICP-MS. SBSE in combination with CGC-ICP-MS was applied for the determination of organotins in environmental samples after in-situ derivatization with NaBEt4, resulting in unsurpassed sensitivity with detection limits reaching the ppq (pg/L) level [17]. For standard applications such as the determination of organotin compounds in sediments, or soils, and in extracts or leachates of consumer products, these extremely high sensitivities are not required. For these applications, sufficient sensitivity is obtained using mass spectrometric detection. In comparison to AED or ICP-MS, where specific tin-chromatograms are obtained, the chromatograms obtained by mass spectroscopy are far more complex, even when using the selected ion monitoring (SIM) mode. Several ions per solute need to be monitored, and the derivatized sample extracts often contain many co-extracted solutes

or by-products of the derivatization reaction. Therefore, data interpretation is more demanding requiring the use of extracted ion chromatograms, retention time matching, and calculation of the relative abundances of target and qualifier ions. In this respect, the use of retention time locked methods offers several advantages. If a selected ion method is used, the switching times between groups of monitored ions are fixed and do not need to be adjusted after column maintenance or column change, since the retention times of all solutes can be relocked. Moreover, quantification databases do not need to be updated for variations in retention times. Finally, a retention time locked database can be used, allowing easy peak allocation. Solute detection and confirmation are far more reliable using the results screener option [18,19], which combines the power of spectral matching with locked retention time matching. In this application note, a GC/MS method is described for the analysis of organotin compounds in environmental, food, or consumer product extracts. Since derivatization by Grignard reaction and derivatization using NaBEt4 are both easy and convenient, three types of derivatives are considered: methyl-derivatives using methylmagnesium bromide, pentyl- derivatives using pentylmagnesium bromide (both Grignard reagents), and ethylderivatives using NaBEt4. The most important organotin compounds are listed in Table 1 together with typical ions for the mass spectra of all three derivatives. Tin has several isotopes and the mass spectra are characterized by typical isotope clusters. The relative abundances of the tin isotopes are Sn-116 (14.24%), Sn-117 (7.57%), Sn-118 (24.01%), Sn-119 (8.59%), Sn-120 (32.97%), Sn-122 (4.71%), and Sn-124 (5.98%). For the organotin compounds listed in Table 1, mass spectral libraries and retention-time-locked screener libraries were created for all three types of derivatives. After selecting the appropriate derivitization method, a library and screener database can be selected, allowing fast data interpretation. Sample extraction and clean-up are beyond the scope of this application note.

Experimental
Samples The organotin compounds listed in Table 1 were purchased from Dr Ehrenstorfer, Augsburg, Germany (http://www.analytical-standards.com). For analysis, the standards were dissolved in methanol at a 1000 ppm (1mg/mL) concentration. These solutions were further diluted, depending on the derivatization method used. For creation of the databases, approximately 10 g of compound was derivatized, resulting in a final concentration of 10 ppm. Derivatization method 1: The sample extract is concentrated to 1 mL in an apolar solvent (typically hexane) in a reaction tube. To this solution, 0.5 mL methylmagnesiumbromide Grignard reagent (1.4 M in 75/25 toluene/THF, SigmaAldrich cat no 28,223-5) is added. The solution is vortexed for 10 s and allowed to stand at room temperature for 15 min. This procedure should be performed in a fume hood, since toxic vapors evolving from the reaction and the solvents are

flammable. The reaction is stopped and the excess reagent is removed by adding 2 mL of a saturated ammoniumchloride solution in water or 2 mL 0.25 mol/L aqueous sulphuric acid. The mixture is vortexed for 10 s and the two phases are allowed to separate. The clear upper layer (apolar hexane phase) is transferred to an autosampler vial for analysis. The resulting organotin compounds are the methyl-derivatives. Derivatization method 2: The sample extract is concentrated to 1 mL in an apolar solvent (typically hexane) in a reaction tube. To this solution, 0.5 mL pentylmagnesiumbromide Grignard reagent (2 M in diethylether, Sigma-Aldrich cat no 29,099-8) is added. The remaining steps in this procedure are identical to those used in derivitization method 1. The resulting organotin compounds are the pentyl-derivatives. Derivatization method 3: The sample extract is concentrated to 1 mL in a polar solvent (typically ethanol) in a reaction tube. To this solution, 1 mL acetate buffer (82 g/L sodium acetate in water, adjusted to pH 4.5 with acetic acid) and 50 L

Table 1:

Organotin Compounds and Characteristic Ions for the Three Derivatization Products Abbreviation Derivatization 1 Methylmagnesium bromide Methyl193, 191, 165, 163 207, 205, 179, 177 179, 177, 221, 219 249, 247, 207, 205 165, 163, 151, 149 151, 149, 207, 205 193, 191, 249, 247 291, 289, 179, 177 227, 225, 223, 197 289, 287, 285, 197 351, 349, 347, 197 351, 349, 347, 197 301, 299, 219, 217 165, 163, 263, 261 263, 261, 151, 149 Derivatization 2 Pentylmagnesium bromide Pentyl179, 177, 249, 247 207, 205, 179, 177 277, 275, 165, 163 249, 247, 207, 205 319, 317, 193, 191 319, 317, 179, 177 305, 303, 179, 177 291, 289, 179, 177 339, 337, 197, 195 345, 343, 197, 195 351, 349, 347, 197 351, 349, 347, 197 357, 355, 205, 203 375, 373, 193, 191 417, 415, 375, 373 Derivatization 1 Sodium tetraethylborate Ethyl207, 205, 179, 177 207, 205, 179, 177 235, 2331, 249, 247 249, 247, 207, 205 235, 233, 179, 177 263, 261, 207, 205 291, 289, 207, 205 291, 289, 179, 177 255, 253, 197, 195 303, 301, 197, 195 351, 349, 347, 197 351, 349, 347, 197 315, 313, 233, 231 291, 289, 179, 177 375, 373, 263, 261 3

Organotin solute Reagent Derivatives Triethyltin Tetraethyltin Tripropyltin Tetrapropyltin Monobutyltin Dibutyltin Tributyltin Tetrabutyltin Monophenyltin Diphenyltin Triphenyltin Tetraphenyltin Tricyclohexyltin (Cyhexatin) Monooctyltin Dioctyltin

TET TeET TPT TePT MBT DBT TBT TeBT MPhT DPhT TPhT TePhT TCT MOT DOT

derivatization reagent are added. The derivatization reagent is prepared by dissolving 2 g NaBEt4 (Sigma-Aldrich cat no 48,148-3) in 10 mL ethanol. This solution should be freshly prepared. The sample is shaken and allowed to react for 30 min. After addition of 5 mL water, the derivatized compounds are extracted in 1 mL hexane. The mixture is vortexed for 10 s and the two phases are allowed to separate. The clear upper layer (apolar hexane phase) is transferred to an autosampler vial for analysis. The resulting organotin compounds are the ethyl-derivatives. These derivatization methods can be adapted to the type of sample analyzed. For example, derivatization method 3 is often applied to aqueous samples directly, combining in-situ derivatization and simultaneous extraction. This method is also used for sediment samples. Typically 1 g sample (dry weight) is extracted with 10 mL acetate buffer, 7 mL methanol and 10 mL hexane. Four mL of a 5% NaBEt4 solution is added while stirring. The derivatized organotin compounds are simultaneously extracted into the hexane layer.

Analytical Conditions
All analyses were performed on an Agilent 68905973N GC-MSD system. Automated splitless injection was performed using an Agilent 7683 automatic liquid sampler. The instrumental configuration and analytical conditions are summarized in Table 2. The retention time of tetrabutyltin (used as the locking standard) was locked at 16.000 min. To duplicate this method, the initial column head pressure can be set to the pressures indicated in Table 2 (nominal pressure). Then the retention time locking (RTL) calibration runs can be performed automatically (at 20%, 10%, +10% and +20% of the nominal pressure) [18]. The retention time versus head pressure curve is then calculated and stored in the method. Agilents RTL software uses this curve to set the column head pressure so that retention time of the locking standard (tetrabutyltin) is 16.000 min.

Table 2.

Instrumentation and Conditions of Analysis

Instrumentation Chromatographic system Inlet Detector Automatic sampler Liner Column Experimental conditions Inlet temperature Injection volume Injection mode Carrier gas Head pressure Oven temperature Transfer line temperature Detector 280 C 1 L Splitless, purge time: 1 min, purge flow: 50 mL/min. Helium Tetrabutyltin is retention time locked at 16.000 min (pressure around 45 kPa at 50 C, 34 cm/s at 50 C) 50 C, 1 min, 10 C/min to 300 C, 4 min. 300 C Scan (40550 amu), threshold 100, MS quad 150 C, MS source 230 C. Solvent delay: 4 min SIM mode: 50 ms dwell time per ion, ions listed in Table 3 Agilent 6890 GC Split/Splitless Agilent 5973 N MSD Agilent 7683 Splitless liner (part number 5062-3587) 30 m 0.25 mm id 0.25 m HP-5MS (Agilent part number 19091S-433)

derivatization. With this derivatization, the elution sequence of the butyltin compounds is MBT A typical chromatogram, for an organotin standard (10 C atoms) < DBT (12 C atoms) < TBT (14 C atoms) < TeBT (16 C atoms). The spectrum obtained for mixture, derivatized using method 3 (ethylderivatives with NaBEt4), is shown in Figure 1. The tributyltin (as tributylethyltin) is shown in Figure 2. The typical ion clusters, resulting from the different compounds elute according to their boiling point, tin isotopes, are clearly detected. and the elution sequence can be predicted by calculating the total number of carbon atoms after
3000000

Results and Discussion

TCT
2800000 2600000 2400000 2200000 2000000 1800000 Abundance 1600000 1400000 1200000 1000000 800000 600000 400000 200000 0 10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00 20.00 21.00 22.00 23.00 Time

TBT

TeBT

DOT MBT TPT TePT DBT MOT

Figure 1.

GC/MS chromatogram for the analysis of an organotin standard mixture after derivatization with NaBEt4 (ethyl-derivatives).
207

130000 120000 110000 100000 90000 Abundance 80000 70000 60000 50000 40000 30000 20000 10000 0 40 60 41 57 81 80 103 100 120 140 160 180 200 220 m/z 240 260 280 313 300 334 320 340 360 389 406 380 400 121 235 151 177 263 291

Figure 2.

Mass spectrum of tributyltin after derivatization with NaBEt4 (ethyl-derivative). 5

The analysis of a coastal sediment sample is shown in Figure 3. In this case, derivatization method 2 (Grignard reaction with pentylmagnesium bromide) was applied and a complex chromatogram was obtained. Using the extracted ion chromatogram at m/e 179 the butyltin compounds were easily detected (Figure 4). Tetrabutyltin, eluting at 16.000 min, was added as internal standard. In this case, pentyl- derivatives are analyzed. Therefore the elution order is reversed since the derivatization adds a C5-group for every free valency. The elution sequence is now TeBT (16 C atoms = unchanged) < TBT (17 C atoms) <DBT (18 C atoms) < MBT (19 C atoms). The mass spectrum obtained for the pentyl derivative of tributyltin is shown in Figure 5.

9500000 9000000 8500000 8000000 7500000 7000000 6500000 6000000 Abundance 5500000 5000000 4500000 4000000 3500000 3000000 2500000 2000000 1500000 1000000 500000 0 15.50 16.00 16.50 17.00 17.50 Time 18.00 18.50 19.00 19.50

Figure 3.

GC/MS chromatogram for the analysis of a coastal sediment sample after derivatization with pentylmagnesium bromide (pentyl-derivatives).

16.01
650000 600000 550000 500000 450000 400000 Abundance 350000 300000 250000 200000 150000 100000

TeBT

17.00

TBT

DBT
17.94

MBT
50000 0 15.50 16.00 16.50 17.00 17.50 Time 18.00 18.50 19.00 19.50

17.72

19.00

Figure 4.

Extracted ion chromatogram showing the presence of butyltin compounds in the coastal sediment sample extract(shown in Figure 3) after derivatization with pentylmagnesium bromide (pentyl-derivatives).

179 280000 260000 240000 220000 200000 235 180000 160000 Abundance 140000 120000 100000 80000 60000 40000 20000 0 305

121

41 69 147 99 207

254 283 327

355

405

452

477

50

100

150

200

250
m/z

300

350

400

450

Figure 5.

Mass spectrum of tributyltin after derivatization with pentylmagnesium bromide (pentyl-derivative).

Using the Agilent results screener and the appropriate screener library, the files can be screened for the presence of all compounds listed in the screener database. Figure 6 shows a typical result, with the identification of pentyltricyclohexyltin at 24.908 min. The target ions for this compound are extracted and overlaid in the top window. For easy comparison, the apex mass spectrum is displayed. Though not shown in Figure 6, the Agilent RTL Screener Software can display the library and apex spectra together for easy spectral comparison. In addition, the relative abundances of the

target ion and qualifiers are measured and compared to the library data. What distinguishes the Agilent screener methods from conventional GC/MS techniques is the comparison of a peak's locked retention time to values stored with the RTL database. In this case, the locked retention time of pentyltributyltin is within 0.002 min (0.12 s) of the database value. The Agilent results screener compares locked retention times and spectral information for fast peak allocation and more reliable identification.

320000 280000 240000 Abundance 200000 160000 120000 80000 40000 0 24.00 24.20 24.40 24.60 24.80 25.00 Time 25.20 25.40 25.60 25.80

24.91

Ion 205 Ion 203 Ion 357 Ion 287

205 200000 180000 160000 287 140000 Abundance 120000 100000 80000 121 60000 40000 20000 147 0 50 100 150 200 250 177 231 253 309 300 m/z 331 350 378 405 429 451 479 503 529 400 450 500 55 81 357

(13) Pentyltricyclohexyltin 24.908 min (-0.002) response 4581312 Ion 205.00 203.00 357.00 287.00 Exp % 100 82.90 77.20 66.70 Act % 100 83.87 80.36 68.98

Figure 6.

Screener result for the detection of tricyclohexyl tin in a sample extract after derivatization with pentylmagnesium bromide (pentyl-derivative).

For added specificity and sensitivity, SIM methods were developed for all three alkyl derivatives of the target tin compounds. Table 3 lists the SIM ions and target compounds in each group. Note that the start time for each SIM group is also listed. Normally, this timing could not be published with confidence, because of retention time differences

between instruments. However, RTL allows analysts to duplicate locked methods directly and reproduce all analyte retention times within a few thousandths of a minute. Thus, it is possible to apply this method directly, including the SIM group timing, after locking tetrabutyltin to the method-specified retention time of 16.000 minutes.

Table 3.

SIM Groups and Timing for Methyl, Pentyl, and Ethyl Derivatives of the Target Tin Compounds Listed in Table 1. The GC/MS Method Shown in Table 2 was used with the Retention Time of Tetrabutyltin Locked to 16.000 Minutes. Start time (min)

Solutes

Ions

Derivatization 1 1 2 3 4 5 6 7 8 Derivatization 2 1 2 3 4 5 6 7 8 Derivatization 3 1 2 3 4 5 6 7 8 5.00 8.50 11.40 13.00 14.80 17.00 22.00 25.00 TeET (=TET) MBT, TPT TePT, DBT MPhT, TBT MOT, TeBT DPhT, DOT TPhT, TCT TePhT 207, 205, 179, 177 235, 233, 179, 177, 249, 247 249, 247, 207, 205, 263, 261, 207, 205 255, 253, 197, 195, 291, 289 291, 289, 179, 177, 291, 289 303, 301, 197, 195, 375, 373, 263, 261 351, 349, 347, 197, 315, 313, 233, 231 351, 349, 347, 197 5.00 9.00 13.50 16.50 20.00 22.80 24.00 26.00 TeET TET, TePT TPT, TeBT TBT, DBT, MBT MPhT, MOT, DPhT DPhT, TCT DOT, TPhT TePhT 207, 205, 179, 177 179, 177, 249, 247, 207, 205 277, 275, 165, 163, 291, 289, 179, 177 305, 303, 179, 177, 319, 317, 193, 191 339, 337, 197, 195, 375, 373, 193, 191 345, 343, 275, 273, 357, 355, 205, 203 417, 415, 375, 373, 351, 349, 347, 197 351, 349, 347, 197 5.00 6.50 8.00 10.50 12.50 16.40 21.00 25.00 TET, MBT TeET MPhT, DBT, TPT MOT, TePT TBT, TeBT DPhT, DOT TCT, TPhT TePhT 193, 191, 165, 163, 151, 149 207, 205, 179, 177 227, 225, 223, 151, 149, 207, 205, 179, 177, 221, 219 165, 163, 263, 261, 249, 247, 207, 205 193, 191, 249, 247, 291, 289, 179, 177 289, 287, 285, 197, 263, 261, 151, 149 301, 299, 219, 217, 351, 349, 347, 197 351, 349, 347, 197

www.agilent.com/chem

Conclusion
A GC/MS method is presented for the analysis of organotin compounds in extracts of environmental, food, or consumer product samples. Three different derivatization methods are described. For each derivatization method, mass spectral and retention time-locked screener databases were created. By itself, RTL is a valuable tool for maintaining GC and GC/MS methods and for comparing results among different laboratories. It also allows analysts to duplicate methods exactly, including SIM group timing and peak timing in quantitative methods. When combining RTL with locked mass spectral database searching, peak identifications become far more convenient and reliable. While many compounds can have similar spectra, they usually do not have similar spectra and identical retention times. Agilents ability to reproduce retention times for a given method on any 6890 GC makes it possible to differentiate closely-related compounds and to screen for large numbers of analytes in a matter of seconds. This rapid GC/MS screening technique is now available for a wide variety of important tin compounds. The three organotin databases are available for free from the Life Sciences and Chemical Analysis portion of the Agilent web site (www.agilent.com).

8. M. Nagase and K. Hasabe, (1993) Anal. Sci., 9, 517. 9. H. H. Van de Broek, G. B. M. Hermes, and C. E. Goewie, (1988) Analyst, 113, 1237. 10. N. Flsvik and E. M. Brevik, (1999) J. High Resolut. Chromatogr., 22, 177. 11. M. Ceulemans, C. Witte, R. Lobinski, and F. C. Adams, (1994) Appl. Organomet. Chem., 8, 451. 12. Y. Morcillo and C. Porte, (1998) Trends Anal. Chem., 17, 109. 13. J. S. Lobinska, M. Ceulemans, R. Lobinski, and F. C. Adams, (1993) Anal. Chim. Acta, 278, 99. 14. L. Moens, T. De Smaele, R. Dams, P. van den Broek, and P. Sandra, (1997) Anal. Chem., 69, 1604. 15. J. Vercauteren, A. De Meester, T. De Smaele, F. Vanhaecke, L. Moens, R. Dams, and P. Sandra, (2000) J. Anal. At. Spectrom., 15, 651. 16. E. Baltussen, P. Sandra, F. David, and C. Cramers, (1999) J. Microcolumn Sep., 11, 737. 17. J. Vercauteren, C. Prz, C. Devos, P. Sandra, F. Vanhaecke and L. Moens, (2001) Anal. Chem., 73, 1509. 18. V. Giarrocco, B. Quimby and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E www.agilent.com/chem 19. K.R. Weiner and H.F. Prest, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries, Agilent Technologies, publication 5968-8657E www.agilent.com/chem

References
1. P. J. Graig, (1986) Organometallics in the environment; Principles and Reactions, 133. 2. Harino H., M. Fukushima, Y. Yamamoto, S. Kawai. N. Miyazaki, (1998) Archives of Toxicology, 35, 558. 3. A. M. Caricchia, S. Chiavarini, C. Cremisini, R. Morabito, and R. Scerbo, (1994) Anal. Chim. Acta, 286, 329. 4. K. Fent and J. Hunn, (1991) J. Environ. Sci. Technol., 25, 956. 5. J. L. Gomez-Ariza, E. Morales, and M. Ruiz-Benitez, (1992) Analyst, 117, 641. 6. H. Harino, M. Fukushima, and M. Tanaka, (1992) Anal. Chim. Acta, 264, 91. 7. M. D. Mller, (1987) Anal. Chem., 59, 617.

For More Information


For more information on our products and services, visit our web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2003 Printed in the USA April 21, 2003 5988-9256EN

Validated Method for the Determination of Phenyl Urea and Triazine Herbicides in Potable and Groundwater by LC/MS Using Selective Ion Monitoring Application

Environmental

Authors
Neil Cullum Anglian Water Services Laboratories Huntingdon UK Paul Stephens Agilent Technologies, Ltd. Bracknell UK

(a triazinone herbicide). Phenyl urea herbicides, such as isoproturon, are widely used for weed control in crops. The triazine herbicides, such as atrazine, are extensively used general weed control agents. Both phenyl urea and triazine herbicides are found in environmental samples and are detected in drinking water. The Prescribed Concentration or Value (PCV) for an individual pesticide in drinking water in the UK, as defined by the Water Supply (Water Quality) Regulations, is set at 0.1 g/L. Ideally, the method of analysis should be capable of detecting 10% to 20% of the PCV, that is 0.01 to 0.02 g/L. Standard methods for the determination of these herbicides in water matrices involve either liquidliquid extraction or, more recently, solid phase extraction (SPE) followed by high performance liquid chromatography (HPLC) with ultra violet (UV)/diode array detection for both classes of herbicide, but more typically gas chromatograph/ nitrogen phosphorus detector (GC/NPD) or gas chromatography/mass spectrometry (GC/MS) detection for the triazines. The phenyl urea herbicides are not suited to GC analysis as they are thermally labile. This application note details the extraction method (SPE) and analysis using liquid chromatography/mass spectrometry (LC/MS) as a combined suite, using a relatively small volume of sample.

Abstract
This application note describes a validated liquid chromatography/mass spectrometry method for phenyl urea and triazine herbicides in potable and groundwater using an atmospheric pressure electrospray ionisation source in both positive and negative ion mode. The performance requirements set by the Drinking Water Inspectorate for standard deviation, bias, and total error are all met for each of the 16 herbicides studied.

Introduction
This application note describes the analysis of a single analytical suite composed of five phenyl urea herbicides including eight triazine herbicides, carbetamide (a carbamate herbicide), chloridazon (a pyridazinone herbicide), and metamitron

Experimental
All analyses were performed using the Agilent 1100 series LC/MS quadrupole coupled to an Agilent 1100 series LC system consisting of a binary pump, autosampler, thermostated column compartment, and vacuum degasser. The system also had a diode array detector in-line before the mass spectrometer, as a trouble shooting tool. The quadrupole mass spectrometer was operated with an atmospheric pressure electrospray ionisation (API-ES) source in positive and negative ion modes.

MS Conditions
Ionisation mode: Drying gas flow: Nebulizer pressure: Drying gas temperature: Vcap voltage: Positive/Negative API-ES 13.0 L/min 40 psig 350 C 3000 V (positive), 2500 V (negative)

LC Conditions
Column: 40 C Flow rate; Injection volume: Mobile phase: 0.5 mL/min 25 L A = 0.001% formic acid in water B = Methanol Initial 2.0 min 15.0 min 15.1 min 22.0 min A 90% 90% 30% 90% 90% B 10% 10% 70% 10% 10% Zorbax Eclipse XDB-C8 50 mm long x 2.1 mm id, 3.5 m particles,

The selected ion monitoring (SIM) ions and fragmentor voltages listed in the SIM table parameters of Table 1 were all optimised using Flow Injection Analysis (FIA). Ten mg/L standard solutions of each herbicide were injected using scan mode 150 400 amu and the fragmentor voltage was ramped from 70 to 150 V in steps of 5 V.

Sample Preparation
SPE was performed using automated equipment and Baker SDB1 200 mg, 3 mL cartridges. The cartridges were conditioned with 5 mL of ethyl acetate, followed by 5 mL methanol, followed by 2 mL HPLC grade water. Fifty mL of sample was diluted to 200 mL using de-ionised water. One hundred mL of the diluted solution was pumped through the conditioned cartridge at 10 mL/min. Cartridges were dried for 25 minutes by forcing air through, followed by elution using ethyl acetate 1 1.5 mL and 1 1 mL. The final extracts were evaporated to dryness using a heated block set at 45 C

Gradient program:

Table 1.

SIM Table parameters, positive ion mode Compound Metamitron Chloridazon Monuron Simazine Carbetamide Cyanazine Isoproturon Chlortoluron/ Isoproturon-d6 Atrazine Propazine Terbuthylazine Trietazine Prometryn Terbutryn Time 2.50 7.50 Group 1 2 SIM ions, Quantitation, Qualification (q) 203.0, 204.0q 222.0, 224.0q 199.0, 201.0q 202.0, 204.0q 237.0, 238.0q 241.0, 243.0q 207.0, 208.0q 213.0, 215.0q 216.0, 218.0q 12.40 4 230.0, 232q Fragmentor voltage 70 100 115 70 95 130 140 130 135 130

Compound # 1 2 3 5 4 6 9 7 8 12 13 15 14 16

10.25

242.0, 243.0q

130

SIM Table parameters, negative ion mode 10 11 q Qualifier ion Diuron Linuron 2.50 1 231.0, 233.0q 247.0, 249.0q 130 115

and a gentle stream of air. The residue was redissolved in 250 L 90:10 HPLC grade water:methanol. The final make-up solution contains an internal standard at a concentration of 0.1 g/L. The internal standard used is isoproturon-d6. Isoproturon-d6 is also present in the calibration standards at the above concentration for all levels of the calibration.

450000 400000 350000 300000

MSD1 TIC, MS File (T:\AGILEN~2\922259~1\DATA\PT1\SAMPLE03.D) API-ES, Pos, SIM, Frag: 70 (TT)

15 + 16 14

9 250000 200000 150000 100000 50000 0 4 6 8 10 12 14 16 18 min 1 3 2 4 5+6 7 12.902 10 18000 16000 14000 12000 10000 8000 11 6000 4000 2000 0 4 6 8 10 12 14 16 18 min
MSD2 TIC, MS File (T:\AGILEN~2\922259~1\DATA\SAMPLE03.D) API-ES, Neg, SIM, Frag: 130 (TT)

8 12 13

Figure 1. Chromatogram for the low level standard in positive ion mode.

Figure 2. Chromatogram for the low level standard in negative ion mode.

IS 70000 60000 50000 40000 30000 20000 10000 0 4 6 8 10 12 14 16 18 min 7


MSD1 213, EIC=212.7:213.7 (T:\AGILEN~2\922259~1\DATA\PT1\SAMPLE05.D) API-ES, Pos, SIM, Frag: 70 (TT)

Figure 3. Extracted ion chromatogram from a spiked tap water sample containing chlortoluron at 0.10 g/L. The isoproturon-d6 internal standard (IS) is also shown. 3

Results
A typical chromatogram for the low level standard (each compound at 0.1 g/L) appears in Figures 1 and 2 in both positive and negative ion modes respectively. Figure 3 shows an extracted ion chromatogram from a spiked tap water sample containing chlortoluron (0.10 g/L) and the isoproturon-d6 internal standard. Validation of the method was done on 11 batches of samples. Standards were spiked at three levels: 0.01, 0.10, and 0.40 g/L. The borehole raw water was spiked at two levels: 0.01 and 0.10 g/L. The potable tap water (which was from a surface water source) was also spiked at two levels: 0.01 and 0.10 g/L . All samples were analysed in duplicate in each batch in a random order. See Table 2. The limit of detection (LOD) for each herbicide was calculated from the within-batch standard deviation of the standard spiked at 0.01 g/L. Recovery for both the groundwater and potable water samples was calculated from the 0.10 g/L spike after the subtraction of 0.01 g/L, hence the recovery value is based on 0.09 g/L. Calibration curves were produced using three calibration levels at 0.1, 0.3, and 0.5 g/L. The calibration curves were all produced using a quadratic fit and forced through the origin and referenced against an internal standard. Typical correlation values are 0.9997 or better for all the herbicides in the suite.
Table 2. Results Groundwater sample Compound Atrazine Carbetamide Chloridazon Chlortoluron Cyanazine Diuron Isoproturon Linuron Metamitron Monuron Prometryn Propazine Simazine Terbuthylazine Terbutryn Trietazine Recovery % 83.4 92.0 94.8 89.5 95.1 92.7 93.0 89.4 102.4 96.6 81.8 85.8 91.0 78.1 81.9 79.1 RSD % 5.1 7.5 5.8 3.9 4.3 4.6 2.6 4.9 2.8 4.6 4.6 5.8 4.7 8.3 4.4 4.7

Discussion
All 16 compounds could be analysed in positive ionisation mode, but diuron and linuron were analized in negative ionisation mode because some interferences were observed around the 0.01 g/L level. No interferences were observed for these compounds in negative ionisation mode and good limits of detection were obtained. The recovery values obtained for all compounds in both water matrices tested are all in agreement. It appears that there is no suppression of ionisation, which has not been the case for other methods [1] where suppression is noted by a reduction in the recovery of individual compounds in some matrices. Complete separation of all 16 compounds was not achieved by this method. Compounds 5 and 6 coeluted as well as compounds 15 and 16. However, all compounds were still quantitatively analysed as they differ in molecular weight. For instance, since compound 5 (simazine) has a molecular weight of 201 amu ( [M + H]+ = 202 m/z) and compound 6 (cyanazine) has a molecular weight of 240 amu ( [M + H]+ = 241 m/z), they were distinguished by selected ion mass spectroscopy. The same is applicable for compounds 15 and 16. Three compounds 12, 13, and 15 are isomers and all have a molecular weight of 229, but all three compounds show good chromatographic separation.

Potable water sample Recovery % 83.3 88.5 94.0 89.5 96.3 94.3 93.2 91.0 102.8 97.3 82.6 85.6 91.1 80.2 83.6 78.7 RSD % 5.6 7.8 4.6 4.6 4.7 4.5 3.3 4.6 3.1 4.8 4.4 6.2 4.8 6.8 4.6 4.8 LOD g/L 0.00146 0.00544 0.00293 0.00219 0.00352 0.00348 0.00209 0.00330 0.00257 0.00221 0.00208 0.00218 0.00179 0.00397 0.00268 0.00179

Figure 4 shows an extracted ion chromatogram, at 230 m/z ( [M + H]+ ) for compounds 12, 13, and 15. Two other compounds, 14 and 16, are also isomers with a molecular weight of 241 amu ( [M + H]+ = 242 m/z). Compounds 14 and 16 are also separated by the chromatography.
13.979 13.105 13.391 14 14.775 4 6 8 10 12
MSD1 230, EIC=229.7:230.7 (T:\AGILEN~2\922259~1\DATA\SAMPLE01.D) API-ES, Pos, SIM, Frag: 70 (TT)

800000 700000 600000 500000 400000 300000 200000 100000 0 16 18 min

Figure 4. Extracted positive ion chromatogram, 230 m/z, for compounds 12, 13, and 15.

Conclusion
The data shows that the method presented is capable of quantitative analysis for the 16 herbicides in a single analytical suite. The performance requirements set by the Drinking Water Inspectorate (DWI) for standard deviation, bias, and total error are all met for each individual herbicide. Although spiked recovery targets of 90%110% are not achieved in all cases, this can be compensated for by the application of recovery factors calculated from the performance data.

References
1. N. Cullum, P. Stephens and S.Evans, Determination of Acidic Herbicides in Groundwater and Potable Water by LC/MSD Using Selective Ion Monitoring. Agilent Technologies, publication 5988-5882EN www.agilent.com/chem

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2003 Printed in the USA April 2, 2003 5988-8595EN

Analyzing Phenyl Ureas and Carbamate Pesticides Using ESI-, APPI-, and APCI-LC/MSD Application

Environmental and Food

Author
Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
Carbamates (a class of highly effective insecticides) and phenyl ureas (a class of herbicides) were successfully analyzed by liquid chromatography/mass spectrometry using electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), and atmospheric pressure photo ionization (APPI) sources. APPI and APCI exhibited lower background signals and fewer background peaks than ESI. Phenyl ureas, in general, exhibited better signal-to-noise ratios (S/N) from APCI or APPI than from ESI. However, it was just the opposite for carbamates. The S/N was better from ESI than from APCI or APPI. A small amount of post-column acetone was needed in the APPI source as a photoionizable dopant for charge transfer to the analyte.

Gas chromatography/mass spectrometry (GC/MS) systems have traditionally been used for analyzing pesticides in environmental samples. Due to the thermally labile nature of these compounds, liquid chromatography has been the method of choice for the separation of these pesticides. Many of the pesticides within the same class exhibit similar ultraviolet (UV) spectra. With the development of atmospheric pressure ionization (API) techniques, liquid chromatography/mass spectrometry (LC/MS) systems have become the preferred method for the analysis of ground- and surface-water contaminants. The LC/MS provides better sensitivity and specificity than typical diode array detectors (DAD), even if the analytes are not fully resolved from neighboring eluants. Three different API sources were used in this study for comparison: electrospray ionization (ESI), chemical ionization (APCI) [1], and photo ionization (APPI) [2]. In ESI, the ionization process happens before the solvent evaporation process. It is, therefore, a more universal ionization for polar compounds. In APCI and APPI, the analyte is not ionized until after solvent evaporation. APCI involves a charge transfer (proton or electron) between ionized reagent gas and the analyte. APPI requires that the analytes and/or a dopant be photoionized by absorbing photons from the 10.6 eV krypton light. The dopant then transfers the charge to the analyte, which could be thought of as photon-induced chemical ionization.

Introduction
Carbamates, a class of highly effective insecticides, are widely used worldwide to protect crops from pests. Phenyl ureas, a class of herbicides, are highly potent chemicals for agricultural weed control. All are potential endocrine disrupters found in ground- and surface-water samples.

Experimental
A mixture of carbamates and urea pesticides at 10 ppm (ng/L) in acetonitrile was purchased from AccuStandard (New Haven, CT). A series of dilutions in acetonitrile were made for the linearity studies. The compounds and their structures are shown in Figures 1 and 2, respectively. All experiments were performed on two G1946D LC/MSD systems equipped with ESI on one and APCI/APPI on the other. The LC/MS conditions are listed in Table 1.

O H3C H3C N H3C Cl O Cl H3C NH O HN O O

O NH CH3 H 3C NH

O O O CH3 CH3

Aminocarb

Barban

Carbaryl

Carbofuran

NH O Cl

CH3 CH3 H3C S

CH3 O O NH CH3 H3C S N CH3 O

O NH CH3

H 3C H3C N CH3

O HN

CH3

H3C

H3C

Chlorpropham
CH3 H3C N H3C S CH3 N O O HN O

Methiocarb

Methomyl

Mexacarbate

H3C NH O O CH3 O CH3 O

CH3 O NH CH3

Cl Cl

O O NH

CH3

Oxamyl

Propham

Propoxur

Swep

Figure 1.

The 12 carbamates used in this study.

Cl H3C O N CH3 NH Cl NH

O N CH3 F F NH F

O N CH3 Cl Cl

NH N

O O CH3

H3C

CH3

CH3

Diuron

Fenuron

Flurometuron

Linuron

CH3 NH Cl O N CH3

Cl Cl NH

O N CH3 CH3 CH3 NH

O NH

Monuron

Neburon

Siduron

Figure 2.

The seven phenyl ureas used in this study.

Table 1.

LC/MS Conditions ESI APCI APPI

Column Column temperature Column flow rate Solvent A Solvent B

Zorbax Eclipse XDB-C8, 4.6 x 50 mm, 3.5 m (p/n 935967-906) 30 C 1 mL/min H2O, 0.1% acetic acid or as specified Acetonitrile, 0.1% acetic acid or as specified n/a B: 10% at 0 min, 30% at 4 min, 80% at 10 min, 80% at 16 min 2 L 12 L/min 350 C 60 V 3500 V 60 psi n/a 0.1 0.15 Off 150800 Positive 30 C 1 mL/min H2O Acetonitrile 30 C 1 mL/min H2O Methanol

Post column Solvent gradient

n/a

40 L/min acetone added as dopant

B: 30% at 0 min, 40% at 4 min, 70% at 13 min, 80% at 16 min or as specified in the figure 2 L 11 L/min 275 C 110 V 4500 V 35 psi 225 C 0.1 0.15 Off 115500 Positive 2 L 11 L/min 275 C 110V 4500V 35 psi 225 C 0.1 0.15 Off 115500 Positive

Injection volume Drying gas flow Drying gas temperature Fragmentor voltage Vcap Nebulizer pressure Vaporizer Step size Peak width (min) Time filter Scan (m/z) Polarity

Results and Discussion


Figure 3 shows the effect of adding a post-column dopant (acetone) in APPI to significantly improve the responses of the analytes. Acetone at a flow rate of 40 L/min was introduced to the column effluent before entering the nebulizer. A gain of 1000 in peak height was seen for some of the analytes. The difference in peak-height gain among the analytes may be due to their structural difference and the charge transfer efficiency with the dopant.

140000 120000 100000 80000 60000 40000 20000 0 2 4 6 8 10 12 14 min

APPI without post-column acetone

2250000 2000000 1750000 1500000 1250000 1000000 750000 500000 250000 0 2 4 6 8 10 12 14 min

APPI with post-column acetone

Figure 3.

TICs show the effect of adding post-column acetone to enhance analyte signal in APPI. Methanol was used as solvent B for both TICs.

The LC mobile phase in APPI can interfere with ionization if the solvent has more affinity for the proton than the analyte. Figure 4 shows that acetonitrile can be a problem for baseline stability and it is always best to also try methanol in APPI. Typical background spectra between 15 and 15.5 minutes from the three sources appear in

Figure 5. Because ESI is a more universal ionization source for polar compounds and ionic modifiers are used in ESI, the baseline has a lot more peaks and the noise is usually 5 to 15 times higher than APCI and APPI.

1200000 1000000 800000 600000 400000 200000 2 4 6 8 10

Mobile A: H2O Mobile B: ACN with dopant

min

1000000 800000 600000 400000 200000 0 2 4 6 8 10

Mobile A: H2O Mobile B: MeOH with dopant

min

B: 10% at 0 min, 30% at 3 min, 80% at 8 min, 80% at 12 min

Figure 4.

TICs show that methanol is a better mobile phase than acetonitrile in APPI. Acetone was used as dopant.

6
116.1 117.1 158.9 183.0 200 210.0 233.1 158.1 256.2 170.2 179.2 299.2 338.1 201.1 400 363.0 385.1 219.2 233.2 250 279.1 298.3

Figure 5.
149.1
128.1 138.2 150

136.1

150 205.1
200 188.1

171.2

200 233.2
220.2

337.1

426.1

250 250.1
260.2 259.2

259.2

485.0 522.5 275.1 550.5 288.3

280.2 288.3 279.2

300
300

600

313.2

ESI

APPI

APCI

Max: 3155

Max: 1091

Max: 20907

633.1

350
350

Baseline noise of the three sources are compared. ESI exhibits higher noise and more peaks in the baseline.
391.3
794.6

The spectra of 20 ng Diuron on column from the three sources (ESI, APCI, and APPI) appear in Figures 6, 7, and 8. The peak at mass 233 is the [M+H]+ peak. Peaks at masses 233, 235, and 237 match the isotope peak-intensity pattern for two chlorine atoms, which is additional information for compound confirmation. Although ESI gives better response compared to the other two sources, it has lower signal-to-noise ratio (S/N).

4.777

6.637

2500000

Diuron

ESI
9.134 8.458 9.201 9.453 10.427

2000000

1500000 2.266 2.340 2.230 2.687 5.095

9.607

1000000

10.676 10.538

11.445 11.624

4.667

500000

0 2 4 6 8 10 12 14 min

60 40

Diuron

235.0

80

233.0

100

10.237

Max: 439851

210.1

20 0 160 180 200

220

222.1

237.0

240

m/z

0.05% formic acid in both solvents A and B

Figure 6.

The TIC of all target compounds from ESI. The bottom half is the ESI spectrum of 20 ng Diuron on column.

13.180

300000 5.476 250000 200000 150000 1.017 1.208 1.910 3.954 100000 0.449 50000 0 2 4 6 8 10 12 14 5.977 6.148 6.346 6.580 8.115

2.300

APPI
Diuron
11.041 11.292 11.667 13.453

9.040

10.012 10.426

7.225

9.550

min

60 40 20 0
160 180

Diuron

235.1 200 220 237.1 240

80

233.1

100

Max: 66405

m/z

Figure 7.

The TIC of all target compounds from APPI. The bottom half is the APPI spectrum of 20 ng Diuron on column. APCI

400000 5.310 300000 200000 5.573 7.124 8.356 100000 0 2 4 6 8 10 12 14 min 1.008 1.170 6.224 6.471 8.009

2.202

Diuron
10.957 13.401

10.381 10.569

11.196 11.569

8.942

100 233.1 80 60 40 20 0 160 180 200 220 240 m/z 237.1 235.1

Diuron

Max: 85332

Figure 8.

The TIC of all target compounds from APCI. The bottom half is the APCI spectrum of 20 ng Diuron on column.

15.027 15.386 15.551

Figure 9 shows the spectra of Siduron (protonated ion at m/z 233) from all three sources. Protonated dimer (m/z 465) and sodiated dimer (m/z 487) of Siduron are fairly significant peaks in the ESI spectrum. However, only negligible amounts of protonated dimers (no sodiated dimers) were observed in the APPI and APCI spectra. This is mainly due to the different sequence of desolvation and ionization steps. In ESI, ionization happens before desolvation. In APPI and APCI, ionization comes after desolvation.

233.2

7 6 5 4 3 2 1 0

APPI

Max: 111072 234.2


250

300

350

400

450

500 m/z

10 8 6 4 2 0

233.2

APCI

Max: 171264 234.2


250

300

350

400

450

500 m/z

233.1

100 80

ESI
Max: 1.62867e+006 465.2

[M+H]+
60 40

234.1

466.3

[2M+H]+
255.1

487.2

[2M+Na]+

20 0

250

300

350

400

450

488.3
500 m/z

Figure 9.

Spectra of Siduron from APPI, APCI, and ESI. Significant dimer peaks are observed in ESI.

Figure 10 compares the S/N of some of the target compounds analyzed by the three sources. The figure shows that carbamates, in general, give better S/N from ESI than from APCI or APPI. However, it is just the opposite for phenyl ureas, for example, the S/N is better from APCI or APPI than from ESI. It is also worth noting that APPI shows consistently higher S/N than APCI for this study.

A calibration based on the Methomyl [M+H]+ (m/z 163) is linear over the concentration range of 202000 pg on column. Figure 11 shows the ESI calibration curve with the linear correlation coefficient of 0.9996. Four of the Methomyl peaks that were used for the linearity calibration appear in Figure 12.

ESI APPI APCI

Carbofuran

Mexacarbate

Diuron

Monuron

Carbamate Figure 10. The S/N of four analytes from the three ionization modes.

Phenyl urea

Figure 11. The calibration curve of Methomyl, by ESI in SIM mode (mass 163), shows good linearity over the concentration range of 202000 pg on column (2-L injection).1
1

The calibration curve was generated using the PE Sciex AnalystTM software.

10

20 pg on column

100 pg

200 pg

1000 pg

Figure 12. The AnalystTM software shows four of the Methomyl peaks (163 amu) that were used for generating the calibration curve in Figure 11.

11

Three different modifiers (formic acid, ammonium acetate, and acetic acid) were used in the ESI mode to compare their effectiveness. The results in Figure 13 show that, in general, the peak intensities are comparable among three modifiers. However, the elution order of the peak Mexacarbate (as well as Aminocarb) was different. Therefore, a single selected-ion-monitoring (SIM) method should not be used for different modifiers.

1000000 800000 600000 400000 200000 0 2 1000000 800000 600000 400000 200000 0 2

Mexacarbate

0.1% Formic acid

10

12

min

Mexacarbate

10 mM NH4Ac

10

12

min

1000000 800000 600000 400000 200000 0 2

Mexacarbate

0.1% Acidic acid

10

12

min

Figure 13. Different modifiers in ESI resulted in comparable peak intensities but changed the peak elution order.

12

Conclusions
Carbamates and phenyl ureas were successfully analyzed using ESI, APCI, and APPI sources. APPI and APCI have lower background signals and fewer background peaks than ESI. The results show that carbamates, in general, give better S/N from ESI than from APCI or APPI. However, it is just the opposite for phenyl ureas, for example, the S/N is better from APCI or APPI than from ESI. A small amount of post-column acetone was needed in the APPI source as a dopant. The dopant is photoionizable and is used to transfer charge to the analyte. The typical ESI quantitation limit for these compounds is 20 pg on column using SIM. Some carbamates and phenyl ureas (for example, Propoxur, Carbofuran, and Siduron) exhibited protonated and sodiated dimers in the full scan ESI spectra.

References
1. Basics of LC/MS, Agilent Technologies Primer, Publication 5988-2045EN, February 15, 2001. 2. PhotoMate Atmospheric Pressure Photoionization Source from Syagen Technology, Agilent Technologies Brochure, Publication 5988-3130EN, May 25, 2001.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

13

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA May 24, 2002 5988-6635EN

Analysis of the Active Compound in an Agricultural Fungicide Formulation by Liquid Chromatography Application

Agricultural, Speciality Chemical, Environmental, Ag Chem

Author
James D. McCurry Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
Liquid chromatography was shown to be an excellent tool for the routine analysis of the active ingredient in an agricultural fungicide formulation. No extensive sample preparation or cleanup was required to remove the active ingredient from the rest of the sample matrix. The active ingredient was easily separated and detected using conventional reverse-phase conditions with a UV/VIS detector.

This work was done on a commercially available fungicide formulation. The active ingredient in this product is 6.5 % (wt) of N,N-[1,4-piperazinediylbis (2,2,2-trichloroethylidene)] bisformamide. This is also known as triforine (CAS registry number 26644-46-2) and the structure is shown in Figure 1. The inactive ingredients in this formulation are listed as cyclohexanone, N-methyl pyrrolidone and Atlox 3406-F. The Atlox 3406-F is an agricultural dispersant that contains ionic and nonionic surfactants and mixed aromatic solvents. Electrospray ionization liquid chromatography/mass spectrometry (LC/MS) analysis has shown that the triforine can be easily separated and identified in the formulation [1].
Cl Cl * Cl NH O

Introduction
Agricultural chemical formulations usually contain an active ingredient and several inert components, such as surfactants, that are designed to enhance the efficacy of the product. Gas chromatographic analysis of these formulations cannot be performed due to the polarity or thermal instability of the active ingredient as well as the high molecular weight and polarity of the surfactants. Therefore, liquid chromatography offers the best solution for the routine analysis of the active ingredients in an agricultural formulation.
O HN Cl Cl N * Cl

*Optically active

Figure 1.

Chemical structure of triforine, the active ingredient in some commercial fungicide formulations.

Experimental
A 10% (v/v) solution of the fungicide formulation was made in acetonitrile. This solution was run on the Agilent 1100 Series LC System. This system included a vacuum degasser, a binary pump, an autoinjector, a thermostated column compartment, and a diode array UV/VIS detector. LC instrument conditions for this analysis are shown in Table 1.
Table 1. LC Analysis Conditions

Results and Discussion


Figure 2 shows the chromatogram of the fungicide formulation. The active ingredient in the formulation, triforine, elutes as two chromatographic peaks between 7.5 minutes and 7.8 minutes. The presence of two triforine peaks is due to the stereochemistry of the structure. Figure 3 shows the four triforine stereoisomers. These four configurations can be grouped into two pairs of mirror images that are diastereoisomers. The S,R and R,S configurations are mirror images that are superimposable, resulting in a meso compound that exhibits no optical activity or differences in physical properties. Therefore, because the S,R and R,S configurations are identical, they will elute as one chromatographic peak. The second pair of mirror images are the R,R and S,S configurations. These are not superimposable and are, therefore, enatiomers that will have different optical activity, but identical physical properties. Conventional reverse-phase liquid chromatography cannot separate these enantiomers, and they will co-elute as a single peak. However, these enantiomers can be separated from the meso compound by reversephase LC. This is why there are two triforine peaks, one for the meso compound and one for the enatiomers. Without pure standards of the stereoisomers, it is not possible to determine which configurations can be attributed to the observed chromatographic peaks.

Liquid chromatograph conditions Column: Mobile phase A: Mobile phase B: Mobile phase gradient: Flow rate: Injection volume: Column temperature: Detector: Signal wavelength: Signal bandwidth: Reference wavelength: Reference bandwidth: Zorbax XDB-C8,150 4.6 mm, 5 m (p/n 993967-906) 0.1% Formic acid in water Acetonitrile 30% B at 0 min, 50% B at 7 min, 95% B at 10 min 1.0 mL/min 1 mL 30 C Diode array 254 nm 10 nm 500 nm 40 nm

7.554 min
mAU

7.757 min

40

30

20

254 nm

10

0 0 2 4 6 8 10 12 14 min

Figure 2.

LC of an agricultural fungicide formulation containing the active ingredient triforine. The two peaks at 7.554 min and 7.757 min were shown to be optical isomers of triforine.

Mirror

Mirror

O HN H
(S)

O NH HN H Cl3C
(R)

O NH

CCl3

Cl3C

(R)

(S)

CCl3

N H
(R)

N CCl3 Cl3C
(S)

N H H
(R)

N CCl3 Cl3C
(S)

NH O (S,R)

HN O (R,S) O

NH

HN O

(R,R)

(S,S)

Meso structure no optical activity

Enantiomers optically active

Figure 3.

The four triforine stereoisomers arising from the two chiral carbons in the structure. These two pairs of mirror images account for the two triforine peaks observed in the chromatogram.

Conclusions
Liquid chromatography was shown to be an excellent tool for the routine analysis of the active ingredient in an agricultural fungicide formulation. No extensive sample preparation or cleanup was required to remove the active ingredient from the rest of the sample matrix. The active ingredient was easily separated and detected using conventional reverse-phase conditions with a UV/VIS detector.

References
1. McCurry, J.D., and Zavitsanos, P., Analysis of Components, Contaminants, and Impurities in Fungicide Chemical Formulations by GC/MS and LC/MS, Agilent Technologies Application Note, Publication 5988-6085EN, April 2002.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA May 14, 2002 5988-6686EN

Herbicides

Examples of Bonded-Phase Selectivity Differences

Application
Environmental
Robert Ricker

1. Bentazon 2 .Tebuthiuron 3. Simazine 4 Atrazine 5. Prometron 6. Diuron 7. Propazine 8. Propanil 9. Prometryne 10. Metolachlor

Highlights
Useful differences in band spacing can often be obtained by varying column type, keeping mobile-phase conditions constant.

For the same mobile phase, stationaryphase strength (retentivity) is in the order C8>Phenyl>CN for these sterically protected bonded phases.

Long-term stability and reproducibility are provided by ZORBAX StableBond columns.


Conditions: ZORBAX StableBond Columns (4.6 x 150 mm) (Agilent P/N: above) Mobile Phase: 35% ACN, 65% H20, 1mL/min

A wide range of pesticides can be separated with excellent peak efficiency and peak shape.

Robert Ricker is an application chemist based at Agilent Technologies, Wilmington, Delaware.

For more information on our products and services, visit our website at: www.agilent.com/chem

Copyright 2002 Agilent Technologies, Inc. All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Printed in the USA April 25, 2002 5988-6340EN

Pesticides

Analysis of Pesticides in Drinking Water

Application
Environmental
Robert Ricker

1. Desisopropylatrazine 2. Metamitron 3. Fenuron 4. Chloridazon 5. Desethylatrazine 6. Metoxuron 7. Carbetamid 8. Bromacil 9. Hexazinon 10. Simazine 11. Metribuzin 12. Desethylterbutylazine 13. Carbutilat 14. Methabenzthiazuron 15. Chlortoluron 16. Atrazine 17. Monolinuron 18. Diuron 19. Isoproturon 20. Metobromuron 21. Metazachlor 22. Buturon 23. Propazine 24. Dimefuron 25. Terbuthylazine 26. Linuron 27. Chlorbromuron 28. Chloroxuron
Courtesy of Dr. rer.nat. Claus Schlett, Gelsenwasser AG

Highlights
The 3mm-diameter ZORBAX LowVolume Columns offer significant advantages over standard 4.6 mm i.d. columns: - A 2-fold increase in detection sensitivity -- less sample required - A 50% solvent savings -- and reduced solvent-disposal costs

ZORBAX SB-C18 has a sterically protected bonded phase that permits reliable results run after run.

28 pesticides are separated with good resolution and peak shape in a single run using simple mobile phases.

Conditions: ZORBAX SB-C18 (3.0 x 250 mm) (Agilent P/N: 880975-302) Mobile Phase:! A=2mM Sodium Acetate (pH 6.5) with 5% ACN B=100% Acetonitrile!(ACN) Gradient Elution:!!2min, 10% B; 10 to 45% B in 70 min. Injection volume 25l, 0.35 mL/min, 40C, Detect. UV (245 nm)

SUMMARY A variety of pesticides have had extensive use in many countries around the world over the last twenty years. These chemicals are currently present in surface water in very low concentrations, and need to be analyzed. High-Performance Liquid Chromatography with diode-array detection is an excellent tool for analysis of these compounds. TECHNICAL DETAILS Drinking-water regulations have been developed in many locations that set limits for maximum allowable levels of pesticides. A reliable method of analysis is required to monitor these levels, preferably in a single run. HPLC using diode-array detection after solid-phase extraction can meet this need. Generally, substances can be detected in concentrations less than 0.1 mg/L (i.e., the maximum level set in the drinking-water regulation of Germany). ACKNOWLEDGMENT Agilent Technologies, Inc. wishes to thank Dr. rer.nat. Claus Schlett, Gelsenwasser AG who developed the method and provided this chromatogram.

Robert Ricker is an application chemist based at Agilent Technologies, Wilmington, Delaware.

For more information on our products and services, visit our website at: www.agilent.com/chem

Copyright 2002 Agilent Technologies, Inc. All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Printed in the USA April 25, 2002 5988-6341EN

High-Speed Separation of Pesticide Mixture

Application
Environmental
Robert Ricker

Columns with particle sizes under 5 m and columns with dimensions up to ten-times smaller than standard analytical sized columns are ideal for high-speed analyses. In this application, smaller particles and column dimensions are used to resolve six pesticides in less than two minutes. Extra-column volume is a crucial factor in chromatographic performance when using low volume columns. In this example, no modification to a modern instrument (plumbing or flow cell) is necessary.
Operating Conditions:
HPLC System: Column: Mobile Phase: Detection: Flow: Temperature: Agilent 1100 with quaternary pump ZORBAX Eclipse XDB-C18 Rapid-Resolution (3.5m) Cartridge-Column, 4.6 x 30 mm Agilent Part No. 931975-932 MeOH: water (60:40) UV 254 nm with standard flow cell (13 L) 2 mL/ min. ambient

Highlights
Reducing column length and particle size simultaneously can: - Reduce analysis time - Maintain resolution - Reduce solvent use

mAU 500 400 300 200 100 0 0

ZORBAX Eclipse XDB-C18


1. Tebuthiuron 2. Atrazine 3. Diuron 4. Propazine 5. Propranil 6. Prometryne

Higher flow rates are possible due to decreased pressure when using ZORBAX Rapid Resolution cartridge columns.

6 4

0.5

1.5

2.5

3.5

min

Robert Ricker is an application chemist based at Agilent Technologies, Wilmington, Delaware.

For more information on our products and services, visit our website at: www.agilent.com/chem

Copyright 2002 Agilent Technologies, Inc. All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Printed in the USA April 25, 2002 5988-6347EN

Analysis of Components, Contaminants, and Impurities in Fungicide Formulations by GC/MS and LC/MS Application

Agriculture, Specialty Chemical, Environmental, Ag Chem

Author
James D. McCurry Paul Zavitsanos Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
A commercially available fungicide formulation was analyzed by both gas chromatography/mass spectrometry (GC/MS) and electrospray ionization liquid chromatography/mass spectrometry (ESI-LC/MS). The GC/MS analysis provided a detailed look at the volatile components in the formulation, but did not yield any results for the active ingredient, triforine. The ESI-LC/MS provided information on the stereoisomers of triforine as well as the nonvolatile surfactants and contaminants in the formulation. This paper demonstrates the complementary nature of these two analytical techniques when trying to fully characterize a complex chemical formulation containing a broad range of components.

unexpected impurities and breakdown products that can affect product quality. However GC/MS can only provide meaningful information for compounds that are volatile, nonionic, thermally stable, and have relatively low molecular weight. Liquid chromatography is much better suited to analyzing compounds that are nonvolatile, ionic, polar, thermally labile, or have high molecular weight. This includes about 80% of all known organic compounds [1]. When coupled with a modern atmospheric pressure ionization (API) mass spectrometer, LC/MS offers a complementary tool to GC/MS in the chemical diagnostic laboratory. Commercial pest control formulations contain one or more active compounds along with a recipe of ingredients that can play an important role in the products efficacy. These inactive ingredients are often a combination of solvents and surfactants that allow for easy application and dispersal of the active ingredient onto the target substrate. For this work, an over-the-counter fungicide formulation was purchased at a local home products store. The active ingredient in this product is 6.5 % (wt) of N,N-[1,4-piperazinediylbis(2,2,2-trichloroethylidene)] bisformamide. This is also known as triforine (CAS registry number 26644-46-2), and the structure is shown in Figure 1. The inactive ingredients in this formulation are listed as cyclohexanone, N-methyl pyrrolidone, and Atlox 3406-F. The Atlox 3406-F is an agricultural dispersant that contains ionic and nonionic surfactants and mixed aromatic solvents.

Introduction
Gas chromatography/mass spectrometry (GC/MS) is an indispensable tool for solving complex problems in the chemical industry. This fast and powerful technique yields detailed information about the expected compounds in the mixture along with any

Cl Cl * Cl NH O

Table I.

GC/MS Analysis Conditions

Gas chromatograph conditions Column: Carrier gas: Flow rate: Inlet: 30 m 0.25 mm HP5-MS, 0.25 m (p/n 19091S-433) Helium at 13.00 psi 1.6 mL/min., constant flow mode Cool on-column at 50 C, oven track mode

N O HN Cl Cl Figure 1. N * Cl

*Optically active

Oven temperature program: 50 C for 3 min 10 C/min to 275 C 275 C for 4 min MS Transfer line: Injection volume: Electron multiplier: Solvent delay: Scan range: Scan threshold: A/D Samples: Scan rate 280 C 1 L 1400 V 3 min 30 to 800 m/z 50 counts 2 1.95 scans/s

Chemical structure of triforine, the active ingredient in some commercial fungicides. The nominal molecular weight is 432, and the structure contains two optically active carbons.

Mass spectrometer conditions

A complete analysis of this formulation requires GC/MS to separate and identify the volatile components and LC/MS for the surfactants and polar components. Analysis of the active ingredient, triforine, presents a separate challenge. References for triforine analysis cite gas chromatography as the method of choice when analyzing environmental residues [2]. However, the melting point is reported to be 155 oC with decomposition, indicating that gas chromatography may only be possible with on-column injection.

Electrospray Ionization Liquid Chromatography/Mass Spectrometry (ESI-LC/MS) The same fungicide sample was run on the Agilent 1100 Series LC/MSD. This system included a vacuum degasser, a binary pump, an autoinjector, a thermostatted column compartment, and the LC/MSD SL quadrupole mass spectrometer. LC/MS instrument conditions for this analysis are shown in Table 2.

Experimental
Gas Chromatography/Mass Spectrometry (GC/MS) A 1% (v/v) solution of the triforine formulation was made in acetonitrile and the GC/MS analysis was performed with an Agilent 5973 GC/MS system. The components in this system were a 6890N gas chromatograph, a 7683 autoinjector, and a 5973 mass spectrometer. A cool-on-column inlet in the Agilent 6890 GC was used to avoid decomposition of the triforine. Instrument conditions for the GC/MS analysis are listed in Table 1.

Results and Discussion


Gas Chromatography/Mass Spectrometry (GC/MS) The complex nature of this fungicide formulation is revealed when one looks at the GC/MS data. Figure 2 shows the total ion chromatogram (TIC) of the fungicide sample. The volatile components

Table 2.

LC/MS Analysis Conditions

Liquid chromatograph conditions Column: Mobile phase A: Mobile phase B: Mobile phase gradient: Flow rate: Column temperature: Injection volume: Source: Drying gas flow: Nebulizer: Drying gas temperature: Vcap: Stepsize: Peak width: Time filter: Scan range Fragmentor 150 4.6 mm Zorbax XDB-C8, 5 m (p/n 993967-906) 0.1% Formic acid in water 0.1% Formic acid in acetonitrile 30% B at 0 min; 50% B at 7 min; 95% B at 10 min 1.0 mL/min 30 C 1 L Electrospray 12 L/min 40 psig 350 C 3500 V (positive) and 3000 V (negative) 0.1 amu 0.1 min On 120 to 1200 m/z Fixed at 60 V

in the formulation are easily identified from the mass spectral data. The major solvents, cyclohexanone and N-methyl-2-pyrrolidone, dominate the chromatogram while smaller amounts of C9 aromatics, C10 aromatics, and substituted napthalenes are easily separated and identified. There were no peaks in the TIC whose spectra matched the triforine reference spectra from the Wiley mass spectral library. An extracted ion profile using the triforine base peak of 203 m/z did not produce any chromatographic peak indicating the presence of triforine. From this data, it appears that the triforine did not elute from the column into the mass spectrometer. However, a spectral average of the large hump between 18 and 20 minutes shows an isotope pattern indicating one chlorine atom (Figure 3A). Since no chlorinecontaining species other than triforine are components in the formulation, the presence of chlorine and the broad peak shape indicates triforine decomposition in the gas chromatograph. The peak at approximately 20-minute retention time also has a mass spectrum containing an isotope pattern indicating the presence of two chlorine atoms in the structure (Figure 3B). This peak could be a decomposition product or a contaminant in the formulation.

Mass spectrometer conditions

5 2 4 1 2 3 4 5 6 7 1-hexanol Cyclohexanone C9-aromatics N-methyl-2-pyrrolidone C10-aromatics Naphthalenes Triforine decomposition

6 3

10

12

14

16

18

20

Figure 2.

GC/MS TIC showing the complex volatile components in the commercial fungicide formulation.

145 158 187 152

215
140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 m/z

288

144
140 150

157
160 170 180

189
190 200 210

225
220 230 240 250

259
260 270 280 290 300 310 320 m/z

Figure 3.

(A) Average mass spectrum of broad hump between 18 and 20 minutes of TIC. Isotope patterns of the peaks at m/z 145, 158 and 187 indicate the presence of one chlorine atom. (B) Mass spectrum of the peak at 20 minutes shows the presence of two chlorine atoms in the structure.

Electrospray Ionization Liquid Chromatography/Mass Spectrometry (ESI-LC/MS) The positive ion ESI-LC/MS chromatogram is shown in Figure 4. Several major peaks are observed along with several minor components. The spectra of the three peaks eluting between 0 and 2 minutes are shown in Figure 5. Since electrospray is a soft ionization technique, these spectra do not exhibit the detailed fragmentation

needed to interpret structures for these three compounds. However, peak number 2 does have an isotopic pattern indicating the presence of two chlorine atoms in the structure. This compound could be a contaminant related to triforine production or a triforine decomposition product. Figure 6 shows the spectra of the three peaks between 10.5 and 13 minutes. These compounds are the various surfactants that make up the agricultural dispersant used in the formulation.

3 8

2 6

10

12

Figure 4.

TIC from positive ion ESI-LC/MS of fungicide formulation.

167

122

Peak 1
126 146 185 208
200 m/z

120

140

160

180

260 262

Peak 2
264
250 275

300

325

350

375

m/z

199

Peak 3
200
180 190 200

210

220

m/z

Figure 5.

Electrospray spectra from LC/MS peaks 1, 2, and 3. The spectra from peak 2 shows an isotope pattern indicating two chlorine atoms in this structure. This compound may be a contaminant in the formulation from the active ingredient triforine.

Peak 6

600

700

800

900

m/z

Peak 7

400

600

800

1000

m/z

Peak 8

400

500

600

700

800

m/z

Figure 6.

Electrospray mass spectra of LC/MS peaks 6, 7, and 8 from Figure 4. These compounds are the surfactants used in the formulation.

The spectra of LC/MS peaks 4 and 5 (Figure 7) are identical and correspond to the active ingredient, triforine. The protonated molecular ion is observed at m/z 433 along a sodium adduct at m/z 455. The multiplets for m/z 433 to 439 and m/z 455 to 461 exhibit an isotopic pattern consistent with six chlorine atoms. The ion at m/z 388 is due to a rearrangement and subsequent loss of a formamide group from the protonated molecular ion (m/z 433). This is also confirmed by the isotopic pattern indicating six chlorine atoms (m/z 388 to 396). The presence of two triforine peaks in Figure 4 can be explained by the stereochemistry of the structure. Triforine contains two optically active carbons that give rise to four stereoisomers. Figure 8 shows the four configurations that can be grouped into two pairs of mirror images that are diastereomers. The S,R and R,S configurations are mirror

images that are superimposable, resulting in a meso compound that exhibits no optical activity or differences in physical properties. Therefore, because the S,R and R,S configurations are identical, they will elute as one chromatographic peak. The second pair of mirror images are the R,R and S,S configurations. These are not superimposable and are, therefore, enatiomers that will exhibit different optical activity, but identical physical properties. Conventional reverse-phase liquid chromatography cannot separate these enantiomers, and they will co-elute as a single peak. However, these enantiomers are not mirror images of the meso compound and can be chromatographically separated from the meso compound. This is why there are two triforine peaks, one for the meso compound and one for the enatiomers. Without pure standards of the stereoisomers, it is not possible to determine which configurations can be attributed to the observed chromatographic peaks.

(M+H)+ -43

Peak 4
7.554 min

(M+H)+ (M+Na)+

390 392

Peak 5
7.757 min

388

394

433 435 437 439

400

420

440

457 459 461


460 m/z

Figure 7.

Electrospray mass spectra of peaks 4 and 5 from Figure 4. Both spectra show a protonated molecular ion at m/z 433 representing the active ingredient triforine. There is also a sodium adduct (m/z 455) of triforine observed for both peaks. A rearrangement and loss of a formamide group from the protonated molecular ions give rise to the multiplet at m/z 388 to 396. Mirror Mirror

O HN H
(S)

396

O NH HN H Cl3C
(R)

O NH

CCl3

Cl3C

(R)

(S)

CCl3

N H
(R)

N CCl3 Cl3C
(S)

N H H
(R)

N CCl3 Cl3C
(S)

NH O (S,R)

HN O (R,S) O

NH

HN O

(R,R)

(S,S)

Meso structure no optical activity Figure 8.

Enantiomers optically active

The four triforine stereoisomers arising from the two chiral carbons in the structure. These two pairs of mirror images account for the two triforine peaks observed in the chromatogram (Figure 4). 7

The fungicide formulation was also run by ESI-LC/MS in the negative ion mode. The results of this analysis are shown in Figure 9. The negative ion mass spectra for these two peaks are shown in Figure 10. For both triforine peaks, the most stable negative ion species is the chloride adduct (m/z 467). However, the spectra for the first peak contains a
7.554

deprotonated molecular ion (m/z 431) and a formate adduct (m/z 477) that is not observed in the spectra of the later eluting peak. This selective adduct formation is likely related to the stereochemstry of the triforine, but again, without pure standards, the correct configurations cannot be assigned to the chromatographic peaks.

7.757

10

12

Figure 9.

TIC from negative ion ESI-LC/MS of fungicide formulation.


_ (M+Cl)
469 471

RT = 7.554 min
(M+HCO2)
479 _

(M+TFA)
_ 545 547

467

473

477

483

(M_H)

(M+NH4CO2)
494 494 494 494

431 433 435 437

RT = 7.757 min

425

450

475

500

525

550

551 m/z

Figure 10. Negative ion electrospray mass spectra of the two triforine peaks. The spectra from peak at 7.554 minutes shows a deprotonated molecular ion (m/z 431) and a formate adduct (m/z 477) that is not seen in the later eluting peak (7.757 minutes).

549

481

Conclusions
This paper demonstrates the complimentary nature of GC/MS and LC/MS when trying to characterize a formulation that is composed of many different chemical species. The volatile compounds in the formulation can be easily separated and identified by GC/MS. In this case, polar solvents such as cyclohexanone and N-methyl-2-pyrrolidone were the major components while 1-hexanone, C9 aromatics, C10 aromatics, and substituted naphthalenes were present as minor components or contaminants. However, GC/MS did not yield any information on the active fungicidal ingredient, triforine, a hexachlorinated compound. This was most likely due to thermal decomposition during GC/MS analysis. Evidence for this was seen in a broad chromatographic hump containing chlorine-containing constituents. The nonvolatile components in this fungicide were quickly analyzed by ESI-LC/MS. This analysis yielded information on several polar contaminants, some containing chlorine, which may be by-products of triforine production or triforine breakdown products. Also observed were several surfactants that are used in agricultural products as dispersants. The LC/MS analysis did yield significant information on the triforine active ingredient, showing a distribution of stereoisomers in the formulation.

References
1. Willoughby, R., Sheehan, E, and Mitrovich, S. A., Global View of LC/MS: How to Solve your Most Challenging Problems, p. 80, Global View Publishing, 1998. 2. The Pesticide Manual 9th Edition, Worthing, C.R. and Hance R.J. eds., p. 853, The British Crop Protection Council, 1991.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA April 24, 2002 5988-6085EN

Rapid Analysis of CLP Pesticides Using High-Temperature DB-35ms and DB-XLB Columns Application

Environmental

Author
Cameron George Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, CA 95630-4714 USA

thermal stability needed to achieve optimum resolution and sensitivity in the shortest possible time. These needs are realized with Agilent Technologies J&W Brand DB-35ms (primary) and DB-XLB (confirmation) columns. The excellent selectivity of high phenyl content phases for chlorinated pesticides is well documented. However, these phases typically suffer from poor thermal stability resulting in high bleed and excessively long analysis times. DB-35ms uses arylene-phase technology to provide improved thermal stability through a stiffening of the polymer backbone. The result is increased sensitivity and an upper temperature limit of 360C. The column bleed contribution to background noise is reduced, giving a much improved signal-tonoise ratio and increased usable sensitivity compared to standard 35%-phenyl phases. The high thermal limit translates into shorter analysis times, increased column lifetime and the ability to periodically bake the column at a high temperature to remove semivolatile contaminants. DB-XLB uses a proprietary second-generation arylene technology giving it the same 360C upper temperature limit and the lowest bleed of any phase available.

Abstract
Arylene-phase column pairs (primary and confirmation) permit high oven temperature for rapid analysis of CLP chlorinated pesticides. The columns are also suitable for phenoxy acids, haloacetic acids, polychlorinated biphenyls and Environmental Protection Agency Method 508.1 pesticides.

Introduction
Accurate identification and confirmation of trace level chlorinated pesticides are difficult tasks facing environmental laboratories. The chromatographic system, including the analytical columns, must be optimized. The gas chromatography (GC) columns must possess the selectivity, inertness and

Experimental
The columns and related inlet parts are described in Table 1.

Table 1.

Columns and Related Parts


id (mm) 0.32 0.32 0.53 Film (m) 0.25 0.50 Length (m) 30 30 5 Part Number 123-3832 123-1236 5181-3398 160-2535-5

Phase/Description DB-35ms DB-XLB Quartz deactivated splitter Deactivated fused silica guard column

This is a small sampling of the many DB columns and dimensions available.

Results and Discussion


Simple window diagramming identified the exact film thickness necessary to allow DB-XLB to give optimum confirmation power when run using the primary column temperature program. Figure 1 shows the optimized primary and confirmation chromatograms for the DB-35ms/DB-XLB column pair. Because these columns are designed for enhanced thermal resistance, it is not necessary to bake them excessively upon installation to reduce bleed to acceptable levels. A simple 1 to 2 hour conditioning period is typically more than adequate. Conditioning columns overnight is a common requirement with cyanopropyl- and trifluoropropyl-containing CLP pesticide columns. This practice can result in increased column activity and decreased column life time, but is not required with DB-35ms/DB-XLB. In short, you are ready sooner after column installation. Environmental laboratories are also interested in other gas chromatograpy/electron capture detector (GC/ECD) methods with the same dual column

pair used for the chlorinated pesticides. These methods include phenoxy acid herbicides (EPA Method 8151A), haloacetic acids (EPA Method 552.2), PCBs (EPA Method 8082) and EPA Method 508.1 pesticides. This column pair provides baseline resolution of all 8151A analytes on both columns in just over 16 minutes. In addition, DB-35ms and DB-XLB provide the best confirmation and the fewest coelutions of any dual column pair commercially available for 508.1 pesticides.

Conclusions
The DB-35ms and DB-XLB column pair perform rapid, high-resolution separations of CLP pesticides. The high temperature limit and low column bleed make these columns attractive for analyses of similar semivolatile sample mixtures.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Column:

DB-35ms 30 m 0.32 mm id, 0.25 m P/N: 123-3832 DB-XLB 30 m 0.32 mm id, 0.50 m P/N: 123-1236

Carrier: Oven:

Helium at 45 cm/sec (EPC in constant flow mode) 110 C for 0.5 min 110-320 C at 15 /min 320 C for 2 min Splitless, 250 C 30 sec purge activation time 50 pg per component

Column:

Injector:

Detector: Micro ECD, 350 C Nitrogen makeup gas (column + makeup flow = 30 mL/min constant flow)

2 3 7 12 8 1 5 9 10 11 13

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Tetrachloro-m-xylene (SS) -BHC -BHC -BHC Heptachlor -BHC Aldrin Heptachlor epoxide -Chlordane -Chlordane Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Decachlorobiphenyl (SS)

SS-Surrogate Standard
15 16 21 18 19 22

DB-35ms

14

17

20

C784

10 Time (min)

12

14

16

7 8 9 10

12 13 16 15

Complete resolution and confirmation of 22 CLP Pesticides in under 16 minutes!

11 21 18 17 19 22

6 5 14

DB-XLB

20

C783

10 Time (min)

12

14

16

Figure 1. CLP Pesticides 3

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the USA December 18 2001 5988-4973EN

A Complete Solution for Chlorinated Pesticides and Herbicides Using DB-35ms and DB-XLB Columns Application

Environmental

Author
Cameron George Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

confirmation in less than 16 minutes. However, there is always the desire for a faster analysis. This note reports the result of changing to hydrogen carrier gas and increasing the oven ramp rate.

Experimental
Table 1 describes the columns and related inlet parts.
Table 1. Columns and Related Parts
id (mm) 0.32 0.32 0.53 Film (m) 0.25 0.50 Length (m) 30 30 5 Part number 123-3832 123-1236 5181-3398 160-2535-5

Abstract
DB-35ms (primary) and DB-XLB (confirmation) columns, used with inert inlet components and hydrogen carrier gas, perform CLP pesticide analyses in less than 15 minutes total cycle time. Phenoxy acids can be analyzed with the same configuration.

Phase/Description DB-35ms DB-XLB Quartz deactivated splitter Deactivated fused silica guard column

Introduction
Obtaining high quality data in the shortest possible time is a desire of all analytical testing laboratories. To achieve this goal, all aspects of the chromatographic system must be optimized. The GC columns must possess the selectivity, inertness and thermal stability needed to achieve optimum resolution and sensitivity in the least amount of time. For the analysis of CLP pesticides and phenoxy acid herbicides, these needs are met with DB-35ms (primary) and DB-XLB (confirmation) columns. In another Application Note[1], DB-35ms and DB-XLB show excellent selectivity and inertness for CLP pesticides, achieving one hundred percent

This is a small sampling of the many DB columns and dimensions available.

Results and Discussion


To reduce analysis time without a significant loss in resolution, the carrier gas was changed to hydrogen. Using hydrogen carrier gas with a linear velocity of 65 cm/sec, and increasing the oven ramp rate from 15 C/min to 25 C/min, reduces analysis time to less than 10 minutes. Considering a typical cool-down time of 4 to 5 minutes, the total instrument cycle-time is now less than 15 minutes.

Results and Discussion


Figure 1 shows the excellent resolution and confirmation available for CLP pesticides using DB-35ms/DB-XLB with hydrogen carrier gas and a properly scaled temperature program.

Baseline resolution for 22 CLP pesticides


7 2 3 12 10 13 8 9 11 56 4

DB-35ms

14 1516 18 19 17 21

22

20

Columns: DB-35ms 30 m 0.32 mm I.D., 0.25 m Part No.: 123-3832 DB-XLB 30 m 0.32 mm I.D., 0.50 m Part No.: 123-1236 Carrier: Hydrogen at 65 cm/sec (EPC in constant flow mode) Oven: 110 C for 0.5 min 110-320 C at 25 C/min 320 C for 2 min Injector: Splitless, 250 C 30 sec purge activation time 50 pg per component Detector: ECD, 350 C Nitrogen makeup gas (column + makeup flow = 30 mL/min constant flow)
10

C794

7 Time (min)

100% confirmation in under 10 minutes


12 2 7

DB-XLB
3 65 4

11 8 10 9

13 14 1516 18 19 17 20 21

22

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
10

Tetrachlorom-xylene (SS) -BHC -BHC -BHC Heptachlor -BHC Aldrin Heptachlor epoxide -Chlordane -Chlordane Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Decachlorobiphenyl

C795

7 Time (min)

Figure 1 Analysis of CLP pesticides using DB-35ms and DB-XLB columns.

DB-XLB and DB-35ms have flexibility for a range of GC/ECD methods, a result of their ideal selectivity, inertness and the robustness of low bleed phases. Phenoxy acid herbicides (EPA Method 8151A) are nicely resolved with these columns. All twenty common herbicides are resolved in slightly

over 16 minutes, as shown in Figure 2. The analysis can be optimized for faster analysis. To obtain chromatograms and analysis conditions for additional GC/ECD methods, go to Agilent's Technical Support at www.agilent.com/chem

Columns: DB-35ms 30m 0.32 mm I.D., 0.25 m Part No.: 123-3832 DB-XLB 30m 0.32 mm I.D., 0.50 m Part No.: 123-1236 Carrier: Helium at 45 cm/sec (EPC in constant flow mode) Oven: 50 C for 0.5 min 50-100 C at 25 C/min 100-320 C at 12 C/min 320 C for 2 min

Injector:

Splitless, 250 C 30 sec purge activation time 50 pg per component Detector: ECD, 350 C Nitrogen makeup gas (column + makeup flow = 30 mL/min constant flow) 1. Dalapon 2. 3,5-Dichlorobenzoic acid 3. 4-Nitrophenol 4. Methyl-2,4-dichlorophenyl-acetate (SS) 5. Dicamba 6. MCPP

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

MCPA 4,4, Dibromooctafluorobi-phenyl (IS) Dichloroprop 2,4-D Pentachlorophenol 2,4,5-T,P 2,4,5-T Chloramben Dinoseb 2,4-DB Bentazone DCPA Picloram Acifluorofen

11

DB-35ms
8 18 13 14 15 9 6 7 10 16 17 19 20

12 5 1 2 3 4

C784-2

10 Time (min)

12

14

16

11

DB-XLB
5 1 2 3 4 6 7 9 10

12

18 19 15 16 17

13 14

C785

10 Time (min)

12

14

16

Figure 2.

EPA 8151A phenoxy acid herbicides.

www.agilent.com/chem

Conclusions
DB-XLB and DB-35ms columns, when used with inert inlet components, hydrogen carrier gas and an appropriate carrier velocity, yield these benefits: Short analysis times for better productivity Excellent thermal stability with 360 C upper temperature limit Confirmation for CLP pesticides and phenoxy acid herbicide

Reference
1. Rapid Analysis of CLP Pesticides Using HighTemperature DB-35ms and DB-XLB Columns, Application Note 5988-4973EN, Nov 26, 2001.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the USA December 17, 2001 5988-4971EN

Routine Analysis of Trace Level Carbamate Pesticides in Food Using LC/MSD

Application

Food, Environmental

Author
Masahiko Takino Agilent Technologies, Inc. 3-3-11, Niitaka, Yodogawa-ku, Osaka, 532-0033 Japan

Abstract
System linearity, sensitivity, and repeatability were evaluated on nine carbamates using positive electrospray ionization. Instrument detection limits were determined to be about 3 to 10 ppb for all the analytes in broccoli extract. The results demonstrate that the liquid chromatography with mass selective detection method is a viable technique to replace the nonspecific post-column derivatization process using fluorescence detection.

including N-methyl carbamates, are also suspected endocrine disrupters. The most often used method for carbamate determination is the post-column derivatization technique using fluorescence detection. Drawbacks of the fluorescence method include false positives due to lack of specificity, required additional hardware and plumbing, and insufficient limits of detection. False positives were confirmed in samples having a positive carbaryl identification by a liquid chromatograph (LC) with a fluorescence detector [1]. The goal of this study is to develop a routine liquid chromatograph/mass selective detector (LC/MSD) assay to detect carbamates in foods at low ppb levels without the derivatization step.

Experimental
A mixture of nine analytes (see Figure 1) in methanol was used to evaluate system linearity, sensitivity, and repeatability. A spiked broccoli extract was used to study the matrix effect on system performances. The LC/MSD system used was the Agilent 1100 MSD SL Quad system.

Introduction
Pesticides in food are a significant route to human exposure. Besides toxicity, many of the pesticides,

O (CH 3)2NC C N S CH3 O

O C NHCH3
CH3 S

O C(CH3 )2 CH N OCNCH3
OCONHCH3 CH2SC 2H5

Oxamyl

Aldicarb

Ethiofencarb

OCONHCH3

OCONHCH3 O CH3 CH3


CH3

OCONHCH3

CH3 O S

CH3 CH 3

CH3 S CH3

Methiocabsulfoxide

Bendiocarb

Methiocarb

OCONHCH3

O (CH 3) N
2

CH3

CO CH3

N N CH3

N( CH3 )2

C 2H 5

CH

OCONHCH3

CH3 S O

CH3 CH3 O

Methiocabsulfone
Figure 1. Structures of the analytes used in this study.

Pirimicarb

Fenobcarb

Table 1.

LC and MS Conditions Inertsil ODS3 (150 mm 2.1 mm 5 m) Zorbax Eclips XDB C18 (150 mm 2.1 mm 5 m) A: MeCN B: 10 mM CH3COONH4 in H2O 20% A/B to 100% A in 30 min 0.2 mL/min 40 C 10 L Positive electrospray 100 to 500 Base peak 10 L/min at 350 C 50 psi 60 V 7

Results and Discussion


TIC and SIM Figure 2 shows the total ion chromatogram (TIC) of all nine analytes at 1 ppm. The base peaks for the analytes were either MH+ or MNH4+, except Aldicarb, which was a fragment from the breakage of the N-O bond. Table 2, later in this note, lists the base peaks. Figures 3 and 4 are the selected ion monitoring (SIM) chromatograms of all nine carbamate standards at 0.2 ppb. The signal-to-noise (S/N) ratios of these peaks range from 7 to 56.

LC Conditions Columns

Mobile phase

Flow rate Column temp Sample volume MS Conditions Ionization Scan range SIM ion Drying gas Nebulizer gas Fragmentor EM gain

160000

1 Oxamyl
140000

6 Pirimicarb 7 Ethiofencarb 8 Fenobcarb 9 Methiocarb


5

8,9

2 Methiocarbsulfoxide 3 Methiocarbsulfone
1

120000

4 Aldicarb 5 Bendiocarb
2

100000

80000

4
60000

40000

20000 0 0 2.5 5 7.5 10 12.5 15 17.5 20

Retention time/min

Figure 2.

TIC of carbamate standards at 1 ppm.

1
700 600 500 400 0 2.5

Oxamy

m/z = 237
5 7.5 10 12.5 15 17.5 20

1200 1000 800 600 0

2 Methiocarbsulfoxide
m/z = 242
2.5 5 7.5 10 12.5 15 17.5 20

900 700 500 0

3 Methiocarbsulfone
m/z = 275
2.5 5 7.5 10 12.5 15 17.5 20

4 Aldicarb
2600 2200 1800 0 2.5 5 7.5 10 Retention time/min 12.5 15 17.5 20

m/z = 116

Figure 3.

SIM chromatograms of four carbamate standards at 0.2 ppb. The base peak is labeled in each chromatogram.

1500 1000 500 0 0 2.5 5 7.5 10 12.5

5 Bendiocarb
m/z = 224
15 17.5 20

6000 4000 2000 0 0 2.5 5 7.5 10 12.5

6 Pirimicarb
m/z = 239
15 17.5 20

9 Methiocarb
1500 1000 500 0 0 1500 1000 500 0 0 2.5 5 7.5 10 12.5 15 17.5 20 Retention time/min 2.5 5 7.5 10 12.5 15 17.5 20

7 Ethiofencarb
m/z = 226

8 Fenobcarb
m/z = 208

Figure 4.

SIM chromatograms of five carbamate standards at 0.2 ppb. The base peak is labeled in each chromatogram.

Ionization Electrospray is a soft ionization technique that produces a large number of molecular-related ions. These ions are typically protonated molecular ions [M+H]+. In this study, both [M+H]+ and [M+NH4]+ are dominant adduct ions. The fragmentor voltage is applied to the exit of the capillary and affects the transmission and fragmentation of sample ions by the in-source collision-induced dissociation in this region. By changing the voltage of the fragmentor, various degrees of fragmentation may be achieved. With a low voltage, there is little fragmentation; with higher voltages, a molecularrelated ion is fragmented to a larger degree.

Figure 5 has five SIM chromatograms collected at different fragmentor voltages ranging from 40 V to 120 V. Using the abundance of carbamates at 40 V as reference, Figure 6 shows the changes in relative abundances of carbamates at four fragmentor voltages. For compounds that do not fragment easily, ion transmission improves at the higher fragmentor voltage because the fragmentor voltage gives these ions a push that helps them travel through the relatively high pressure region between the exit of the capillary and the skimmer. For stable compounds, base peak abundance increases with fragmentor voltage up to about 100 V. Setting the fragmentor to either 60 or 80 V will have the best results.

Fragmentor voltage
3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20

40 V

1 2 3 4

8,9

60 V

80 V

100 V

120 V

Retention time/min Figure 5 SIM chromatograms of carbamate standards at different fragmentor voltages.

140

120

100

80

1 Oxamyl
60

6 Pirimicarb 7 Ethiofencarb 8 Fenobcarb 9 Methiocarb

2 Methiocarbsulfoxide 3 Methiocarbsulfone 4 Aldicarb 5 Bendiocarb

40

20

0 40 60 80 Fragmentor voltage (V) 100 120

Figure 6

Relative abundance of carbamate standards at different fragmentor voltages. Abundances at 40 V were used as the reference values. 5

Repeatability and Linearity Figure 7 shows five repeat injections of the carbamate standards at 5 ppb. The %RSD for each analyte ranges from 1.4% to 13% (Methiocarbsulfone). The 13% variation reflects the relatively weak response of the peak at this low concentration.

Figure 8 shows some of the calibration curves of carbamate standards. The responses are linear for all the analytes over a very wide concentration range. For the range from 0.2 to 1000 ppb, the correlation coefficients of all analytes are all greater than 0.998 (see Table 2).

20000 10000 0 2.5 20000 10000 0 2.5 20000 10000 0 2.5 20000 10000 0 2.5 20000 10000 0 2.5 5 7.5 10 12.5 5 7.5 10 12.5 5 7.5 10 12.5 5 7.5 10 12.5 5 7.5 10 12.5

5 (2.2%) 1 (2.5%) 2 (4.1%) 3 (13%) 4 (1.4%)

6 (1.9%) 7 (2.1%)

8 (1.7%) 9 (1.7%)

15

17.5

20

15

17.5

20

15

17.5

20

15

17.5

20

15

17.5

20

Retention time/min

Figure 7.

Repeatability of five repeat injections of the carbamate standards at 5 ppb. The percentage number in each parenthesis is the %RSD value for the peak.

Fenobcarb, 0.2-1000 ppb correlation coefficient = 0.998


7
800000 6000000 600000

160000 120000 80000 40000

4 5 6 8 7 9
0 50 ppb 4000000

8 9 11 10 12
5 ppb Correlation: 0.99903

400000 200000 0

2 3 5 4 6
0 500 ppb

Correlation: 0.99938

2000000

Correlation: 0.99892

0 0

0.2-10 ppb

2-100 ppb

20-1000 ppb

Methiocarbsulfone, 0.2-1000 ppb correlation coefficient = 0.999


40000

300000

4
3000000

30000 200000 20000

5 8 9 11 10
Correlation: 0.99836 100000

2000000

2 3 54
6 0 500 ppb

10000

0 0

12
5 ppb

8
0 0

6
Correlation: 0.99973

1000000

Correlation: 0.99987

9
50 ppb

0.2-10 ppb

2-100 ppb

20-1000 ppb

Figure 8.

Calibration curves of carbamate standards.

Column Size Figure 9 is a comparison of 10-ppb chromatograms on a 2.1-mm column and a 4.6-mm column. The column length and particle size are the same. The responses from the 4.6-mm column are about 40% of the responses from the 2.1-mm column even though the flow rate is five times higher on the 4.6-mm column. Matrix Effect and LOD In many cases, the sample matrix interferes with the analyte responses. A SIM method collects less background signal. See Figure 10 for the SIM

chromatograms of 10-ppb carbamates in broccoli extract. The injection was 5 L on a 2.1-mm column. The percentage number in each parenthesis is the relative response of the analyte from the broccoli extract to the peak intensity from the analyte standard in methanol. Most of the analytes showed responses very similar (98 to 115%) to the standards, except Methiocarbsulfoxide, which had 71% of the response of the standard in methanol. The Limit of Detection (LOD) is defined as the sample concentration that produces a signal-to-noise (S/N) ratio of three in the broccoli extract. The LOD values of the nine carbamates are listed in Table 2.

6
35000 30000 25000 20000 15000 10000 5000 0 2.5 30000 25000 20000 15000 5 7.5 10 12.5 15 17.5 20

2.1-mm column (Flow rate: 0.2 mL/min)


5 2 1 3 7 4

8 9

4.6-mm column (Flow rate: 1.0 mL/min)


5 (36%)

6 (44.8%)

7 (30.8%)

8 (41%) 9 (34.6%)

2 (41.5%)
10000 5000 0 2.5 5 7.5 10 12.5 Retention time/min 15 17.5 20

3 (31.7%) 1 (29.4%)

4 (36.6%)

Figure 9.

Comparison of 10-ppb chromatograms on a 2.1-mm column and a 4.6-mm column. The percentage number in each parenthesis is the relative response of the analyte from the 4.6-mm column to the peak intensity from the 2.1-mm column.

2000 1000 0 2.5 3000 2000 1000 0 2.5 3000 2000 1000 0 2.5 4000 2000 0 2.5

1 (98%)

8000 4000 0

5 (98%)

7.5

10

12.5

15

17.5

20 20000 10000 0

2.5

7.5

10

12.5

15

17.5

20

2 (71%)

6 (101%)

7.5

10

12.5

15

17.5

20 10000 5000 0

2.5

7.5

10

12.5

15

17.5

20

3 (109%)

9 (121%) 7 (98%)

7.5

10

12.5

15

17.5

20 8000 4000 0

2.5

7.5

10

12.5

15

17.5

20

4 (115%)

8 (101%)

7.5 10 12.5 Retention time/min

15

17.5

20

2.5

7.5

10 12.5 15 Retention time/min

17.5

20

Figure 10. SIM chromatograms of 10-ppb carbamates in broccoli extract.

Table 2. # 1 2 3 4 5 6 7 8 9

Carbamates with the Base Peak, Correlation Coefficient and LOD in Broccoli Extract Base peak 237 [M+NH4]+ 242 [M+H]+ 275 [M+NH4]
+

Compound Oxamyl Methiocarbsulfoxide Methiocarbsulfone Aldicarb Bendiocarb Pirimicarb Ethiofencarb Fenobcarb Methiocarb

Correlation coefficient (0.21000 ppb) 0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.998 0.998

LOD (S/N = 3 in broccoli extract) 3 ppb 10 10 2 1 3 2 1 2

116 (fragment) 224 [M+H]+ 239 [M+H]


+

226 [M+H]+ 208 [M+H]+ 226 [M+H]


+

Conclusions
Several carbamates can be analyzed routinely using LC/MSD with low ppb detection limits and excellent linearity in broccoli extract. The results demonstrate that the LC/MSD method is a viable technique to replace the non-specific post-column derivatization process using fluorescence detection.

References
1. Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System, Agilent Application Note, 5980-0332, April, 2000.

For More Information


For more information on our products and services, visit our Web site at: www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the USA December 7, 2001 5988-4708EN

Identification and Quantitation of Pesticides in the Parts-per-Trillion Range Using Retention Time Locking and GC/MS Application

Environmental, Food

Author
C. Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
The typical pesticide quantitation limit for a mass spectrometer in the Scan mode is in the sub-ppm range. By using a selected ion monitoring method, a lab can lower the target compound quantitation limit to the low parts-per-billion (pg/L) range using a retention time locked gas chromatography/mass spectrometry method. By adding large volume injection capability to the method, target compounds at parts-per-trillion can be quantified. A specially developed 567-compound retention time locking pesticide mass spectral library can automatically screen an acquired samples data file for all 567 compounds in seconds. The library can also be applied for rapid screening of samples acquired in selected ion monitoring method. Using the compound library information, a selected ion monitoring method for 80 target compounds was created in less than 2 hours without running any analyses.

detection limits and are easy to operate, the data do not provide sufficient information to confirm a compounds presence with confidence. Due to the universal nature of mass spectrometric detection, a mass spectrometer (MS) provides additional information and increased confidence in the assignment of compound identity. With recent advances in GC/MS hardware and software and the decrease in cost of ownership, more and more laboratories are routinely analyzing pesticide residue samples with MS detection. To match the GC/ESD detection limits and/or to eliminate sample concentration steps, a user must lower the MS detection limit by 2 to 3 orders of magnitude. This application note, discusses the following approaches. Run the MS in single ion monitoring (SIM) mode Make large volume injections (LVIs) Use higher electron multiplier voltage (EMV) For compound identification, a specially developed 567-compound retention time locking (RTL) [1] pesticide library could perform the entire 567-compound screening in seconds using Scan data. A subset of the library could be screened in seconds from SIM data.

Introduction
Most pesticides are typically analyzed on a gas chromatograph (GC) with element-selective detectors (ESDs). Although these ESDs provide low ppb

Experimental
A pesticide standard mixture was used to compare the lowest detection limits of splitless injection and LVI under Scan and SIM modes.

System Configuration for Screening and Quantitation: 6890 GC with a programmable temperature vaporizer (PTV) [2,3] inlet 5973 Mass Selective Detector (MSD) 7683 Automatic Liquid Sampler (ALS) tray and autoinjector HP-5MS capillary column (30 m 0.25 mm 0.25 m), P/N 19091S-433 G1701BA version B.00.00 MSD ChemStation software or higher G1049A MSD RTL Pesticide Database/Library

Table 3. MS Method Parameters Solvent delay 3 min Tune file Atune.u Transfer line 280 C MS Quad 150 C MS source 230 C Threshold 150 Sample # 2 Scan range 35 to 500 amu (in Scan mode) Forty (40) SIM groups (in SIM mode)

Table 4. Pesticide Screening Parameters for the SIM Method Extraction window Qualifier mode Qualifier % Zero qualifiers Subtraction mode Screen database 0.100 minute Absolute 30 Included Average start/stop Rtlpest.SCD

Table 1. GC Method Parameters Oven Inlet 70(2)/25/150(0)/3/200(0)/8/280(10) = 41.87 min PTV

Inlet pressure 17.30 psi (locked to methyl chlorpyrifos at 16.593 min), constant pressure mode

Results and Discussion


RTL [1] was used to: 1. Expedite data comparison in overlay format

Table 2. Injection Parameters Injection mode Injection volume (syringe) Injection speed Inlet temp Solvent vent 25 L (50-L syringe, P/N 5183-0318) Inject @ 100 L/min Draw @ 300 L/min Dispense @ 4500 L/min 40(0.35)/600/320 (3)/50/200 (Hold until end) Vent time = 0.29 min Vent flow = 150 mL/min Vent pressure = 0.00 psi 60 mL/min @ 2 min Deactivated, Multi Baffled (P/N 5183-2037) Liquid CO2 60 mL/min @ 2 min Deactivated, Multi Baffled (P/N 5183-2037) None Splitless 1 L (10-L syringe, P/N 9301-0713) Fast

2. Achieve lower target compound detection limit 3. Allow rapid pesticide screening using the RTL pesticide database/library 4. Help to differentiate isomers by their retention time (RT) differences 5. Eliminate the tedious SIM method RT updating process after column maintenance

280 C

6. Simplify the editing of the SIM ion groups A mixture from the California Department of Food and Agriculture (CDFA) of 80 pesticides at 5000 pg/L each was used as the stock solution for this study. The mixture contained carbamate, organochlorine, organophosphorus, and organonitrogen pesticides. Figure 1 is an offset overlay of three total ion chromatograms (TIC) with 50, 100, and 500 pg of each of the pesticides injected. These TICs were obtained in the Scan mode from 1-L spiltless injections. For many of these pesticides the quantitation limit in the Scan mode is about 500 pg on column.

Vent

Purge Liner

Inlet cooling

240000

50 pg on column
220000

180000

100 pg on column

140000

100000

60000

500 pg on column

20000

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

Figure 1.

Total ion chromatograms from 1-L splitless injections of 80 pesticides with 50, 100, and 500 pg of each compound injected.

SIM Mode To lower the detection limit, a SIM method was created. Instead of the traditional way of making a SIM method, a user can use the information in the RTL Pesticide Database to build a SIM method

without running an analysis. Here are the steps for editing SIM ion group parameters: 1. List the MSD RTL Pesticide Database from the ChemStation (Figure 2 is a partial listing) and paste the complete listing into a spreadsheet.

Figure 2.

A partial listing of the pesticide screener database. The listing includes the compound number, compound name, target ion, expected retention time, and three qualifier ions. 3

2. In the spreadsheet, delete the rows of the compounds not needed in the method. 3. Separate target compounds into groups (see the added Group # column on Figure 3) using these criteria: One to three compounds in each group, and The RTs of the adjacent compounds in adjacent groups are at least 0.2 minute apart. For example, compounds 42 and 51 are more than 0.2 minute apart, so they are in different groups. Compounds 51 and 55 are less than 0.2 minute apart, so they are in the same group. 4. Use the average RT of the adjacent compounds in adjacent groups as the SIM group RT (see the added Group RT column on Figure 3). For example, the average retention time of compound 42 (7.91 min, in group 2) and compound 51 (8.78 min, in group 3) is 8.35 minute which is used as the starting retention time of group 3. When all the group numbers and respective starting retention times are determined, make a hardcopy of the spreadsheet for easy entry into the MS SIM/Scan Parameters in the next step. 5. Enter the target ion and qualifier ion(s) (Q1, Q2, and/or Q3) of all compounds into the respective ChemStation SIM group (Figure 4). Notice that all the information for building the SIM groups came from Figure 3.
# 24 35 42 51 55 76 82 92 98 102 103 104 111 113 117 120 122 124 129 Compound Name 2,6-Dichlorobenzonitrile Mevinphos Propham o-Phenylphenol Pentachlorobenzene Propoxur Diphenylamine Chlorpropham Ethalfluralin Bendiocarb Trifluralin Benfluralin Phorate BHC alpha isomer Hexachlorobenzene Dicloran Demeton-S Dimethoate Simazine MSD_RT 6.75 7.60 7.91 8.78 8.95 10.35 10.52 11.05 11.28 11.54 11.64 11.73 11.96 12.09 12.38 12.56 12.63 12.68 12.91 T 171 127 93 170 250 110 169 127 276 151 306 292 75 181 284 206 88 87 201 Q1 173 192 179 169 252 152 168 213 316 126 264 264 121 219 286 176 60 93 186 Group # Group RT 1 3.00 2 3 4 5 6 7.75 8.35 9.60 10.76 11.41

The number of qualifier ions used in a SIM method depends on the number of analytes of interest. For a method monitoring 20 to 30 compounds, all three qualifier ions should be used in the SIM method. As the list of target compounds grows, fewer qualifier ions should be used in the method to maintain a reasonable and comparable ion dwell time and sampling rate. In general, 10 scans (cycles) per peak are recommended for quantitation purposes. For example, if an analyte peak is 6 seconds wide, about 1.7 cycles per second should be maintained for that SIM ion group. Once the number of cycles per second is determined, the dwell time of the ions can be varied to meet that. As the dwell time is entered for each ion, the ChemStation automatically shows the number of cycles per second. In Figure 4, Group 6 has 3.03 cycles per second.

Figure 4.

A screen capture of the MSD ChemStation showing the MS and SIM parameters. The SIM parameters (group ID, group retention time, and ions) were all derived from Figure 3.

Figure 3.

A spreadsheet of target compounds separated into different SIM groups with RTs of the adjacent compounds in adjacent groups at least 0.2 minute apart. The starting retention time of each group was determined by calculating the average RT of the adjacent compounds in adjacent groups.

Figure 5 shows two chromatograms obtained from 1-L splitless injections at 50 pg/L using both Scan and SIM modes. The Scan mode has significantly higher baseline noise than the SIM mode. Some of the compounds, especially the late eluters, were not detected in the Scan mode. When the Scan method was changed to a SIM method at this concentration, the signal-to-noise ratio (S/N) increased by a factor of 100. It is worth pointing out that a SIM method does not record background ions from the sample matrix, therefore minimizing the baseline noise and improving the S/N.

21000 19000 17000 15000 13000 11000 9000 7000 5000 3000 1000 5.00 10.00 15.00 20.00 25.00 30.00 Injection volume: 1 L Concentration: 50 pg/L PTV mode: Splitless

Scan

SIM
35.00 40.00

Figure 5.

Chromatograms of 1-L splitless injections at 50 pg/L from Scan and SIM modes.

In a SIM method, the retention times of the ion groups normally need updating after column maintenance. By using RTL, a user can not only eliminate the tedious RT updating process [4] but also decrease the detection limit. With reproducible known RTs of target compounds, the start and end time of each ion group can be determined optimally. By narrowing the time windows of an ion group to monitor only one or two compounds at a time, the MS can monitor fewer ions in each window, allowing more sampling time for the target ions. Ideally, a SIM method will have the maximum number of ion groups and the minimum number of ions in each group. In this way, each ion group can get more scans per unit time resulting in better peak shape and more accurate quantitation.

LVIs To decrease the detection limit further, a user can put more sample on column using the LVI technique. The typical solvent-vent approach is to inject the sample slowly into a PTV inlet at a temperature just below the solvent boiling point and let solvent evaporate before ramping up the inlet temperature to move the compounds onto the capillary column. Figure 6 compares a 1-L splitless injection with a 25-L solvent-vent injection. Both injections resulted in 50 pg per compound on column. Note that the solvent-vent ion chromatogram is plotted upside down for ease of comparison with the splitless ion chromatogram. It is obvious from the figure that the two techniques provide very similar results. This demonstrates that the solvent-vent technique is a viable approach for sample introduction.

7000 6000 5000 4000 3000 2000 1000 0 -1000 -2000 -3000 -4000 -5000 6.00 8.00

SIM

Injection volume: 1 L Concentration: 50 pg/L PTV mode: Splitless

Injection volume: 25 L Concentration: 2 pg/L PTV mode: Solvent vent

10.00

12.00

14.00

16.00

18.00

20.00

22.00

24.00

26.00

28.00

30.00

32.00

Figure 6.

SIM results of 50 pg on column using either a 1-L splitless or a 25-L solvent-vent injection.

Higher EMV It is known that the signal increases with higher EMV on the MS. In Figure 7, the upper signal, after 10-fold magnification, is a 25-L LVI of 0.5 pg/L at tune voltage. The bottom signal is the same injection with the electron multiplier set to tune +400 V. Adding 400 V to the EMV increases

the signal by 10X, which makes the integration more accurate. However, the baseline noise also increases by 10X, so the S/N stays the same. Although increasing the EMV does help to bring small peaks over the detection threshold, it shortens the life of the multiplier. In general, the EMV should be kept at the tune voltage.

40000 35000 30000 25000 20000 15000 10000 5000 0 -5000 -10000 -15000 -20000 -25000 -30000 -35000 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00

Tune voltage
(Magnified 10X)

Injection volume: 25 L Concentration: 0.5 pg/L

Tune voltage + 400 V

Figure 7. SIM results of 12.5 pg on column using either EMV at Tune voltage or Tune +400 V.

LVIs in Combination with SIM Methods Combining LVI and SIM, Figures 8 and 9 show quantifiable peaks of three compounds at as low as 5 pg on column. In Figure 8, ion chromatograms of endosulfan sulfate and p,p-DDT at 0.2 and 500 pg/L are shown. The top chromatogram was from a 25-L solvent-vent SIM method and the bottom chromatogram was from a 1-L splitless Scan method. By using LVI and SIM, it is interesting to see that similar S/N ratios were achieved

even with a 2500-fold decrease (from 500 to 0.2 pg/L) in sample concentration. By increasing the injection volume to 100 L, samples at concentration as low as 0.05 pg/L can also be quantified as shown in Figure 9. The top portion shows the chlorthal-dimethyl extracted ion chromatograms (EIC) of mass 299 and 301 from a 100-L full Scan run. The bottom portion shows the same ions from a 100-L SIM run. The SIM method shows better peak shape and lower baseline noise.

280 250 220 190 160

Injection volume: Concentration:

25 L 0.2 pg/L (5 pg on column)

SIM

30000 24000 18000 12000 6000 26.30

Injection volume: Concentration:

1 L 500 pg/L (500 pg on column)

p,p-DDT

Scan

Endosulfan sulfate

26.40

26.50

26.60

26.70

26.80

26.90

27.00

27.10

27.20

27.30

Figure 8.

Ion chromatograms of endosulfan sulfate and p,p-DDT at 0.2 and 500 pg/L. The top chromatogram was from a 25-L solvent-vent SIM method and the bottom chromatogram was from a 1-L splitless Scan method.

1400 1200 1000 800 600 400 200 0 1400 1200 1000 800 600 400 200 0 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 20.4

Ion 301

Scan
Concentration: 0.05 pg/L

Ion 299

1000 800 600 400 200 0 1000 800 600 400 200 0 18.6 19.0 19.2

SIM
Ion 301
Concentration: 0.05 pg/L

Ion 299 *SIM group start time

*
19.4 19.6

*
19.8 20.0 20.2

Figure 9.

Ion chromatograms of 100-L chlorthal-dimethyl injected at 0.05 pg/L. The top portion was from a full Scan run and the bottom portion was from a SIM run.

Target Compound Screening Combing RTL and the G1049A MSD RTL Pesticide Database/Library, a user can screen for 567 pesticides and suspected endocrine disrupters from any Scan run [5]. A user can screen a subset of the library with improved sensitivity using a SIM method. The MSD ChemStation can generate a

567-compound screening report automatically in less than 30 seconds. Figure 10 is a report of the 0.5 pg/L sample (25 L injected in SIM mode) that lists the probable hits (marked with an x) and possible hits (marked with a ?). All target compounds at this 12.5 pg on column level were found by the software.

Figure 10. Typical report from the GC/MS pesticide screener showing probable "hits" (marked with an x) and possible hits (marked with a ?). Other information includes the library retention time followed by the RT difference in this chromatogram, the target ion, its abundance, out of range qualifier(s), and a cross correlation value with the library spectrum.

Conclusions
Using the information (compound names, retention times, and ion masses) in the RTL pesticide database, a SIM method of 80 target compounds can be created in less than 2 hours without running any analyses. The examples show that both LVI and SIM are effective techniques to decrease the quantitation limit of target compounds from sub-ppm to ppt. Any lab can decrease the quantitation limit by a factor of 100 without any hardware modification. Lowering the quantitation limit from 500 pg down to 5 pg on column can be done using a SIM method and RTL. By adding LVI to the system, target compounds in femtogram/L can be quantified.

Acknowledgement
The author would like to acknowledge Alex Chung and Mark Lee at CDFA for providing the pesticide mixture used in this study.

References
1. Vince Giarrocco, Bruce Quimby, and Matthew S. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies Application note, 5966-2469E, printed March 2000, www.agilent.com/chem 2. Bill Wilson, Philip L. Wylie, and Matthew S. Klee, Large Volume Injection for Gas Chromatography Using a PTV Inlet, Agilent Technologies Application note, 5965-7770E, printed February 2000, www.agilent.com/chem

www.agilent.com/chem
3. Philip L. Wylie, Trace Level Pesticide Analysis by GC/MS Using Large-Volume Injection, Agilent Technologies Application note, 5966-1214E, printed April 2000, www.agilent.com/chem 4. Prest, H. and P. Cormia, Retention Time Locking: Advantages in GC/MS SIM Analysis, Agilent Technologies Application Note, 5968-3797E, printed December 1999, www.agilent.com/chem 5. Harry Prest, Philip L. Wylie, Ken Weiner, and Doug Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the 6890/5973 GC/MSD System, Agilent Technologies Application note, 5968-4884E, printed December 1999, www.agilent.com/chem

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the USA November 14, 2001 5988-4392EN

GC/MS Approaches to the Analysis of Monochloropropanediol Application

Foods and Flavors

Author
Harry Prest Agilent Technologies, Inc. 1601 California Avenue Palo Alto, California 94301-1111 USA

monochloropropanediol byproduct was classified by the European Union's Scientific Committee for Food as a suspected carcinogen [1]. Although efforts were made to reduce the presence of 3-MCPD, continuing concerns about its presence lead to regulation of the allowable concentration. Recently the Association of Official Analytical Chemists (AOAC) has published a method for the extraction, separation and identification of 3-MCPD in foods and ingredients using gas chromatography with mass spectrometric detection [2]. In brief, a homogenized sample is mixed with a salt solution, then mixed with an Extrelut refill pack before being added to chromatographic column. The 3-MCPD is eluted with diethyl ether and a portion is derivatized with heptafluorobutyrylimidazole. Quantitation with GC-MS using electron impact ionization provides detection limits less than 0.01 mg/kg (which is equivalent to about 10 pg/L in the final extract at injection). This brief examines approaches to 3-MCPD as the heptafluorobutyryl-derivative described in the AOAC method using the Agilent 5973N MSD.

Abstract
The suspected carcinogen 3-chloro-1,2-propanediol (3-MCPD) is found in hydrolyzed vegetable protein, a widely used flavoring. Gas chromatography with mass spectrometric detection is a standard AOAC method. Electron impact ionization permits subpicogram measurement. Electron capture negative ionization is more selective and probably better suited to actual samples, with sensitivity of a few picograms in scan mode and less than 1 picogram in the selected ion mode.

Introduction
Hydrolyzed vegetable protein (HVP) is a widely used flavoring found in soups, sauces, and some meat products, etc. HVP is traded internationally both as solid and liquid depending upon the intended application. During the acid hydrolysis process of vegetable proteins, the hydrochloric acid agent reacts with triglycerides (Equation 1) to produce 3-chloro-1,2-propanediol (3-MCPD). This
O O R O O O O R" R' HCl OH OH OH HCl Cl OH OH

Experimental
3-MCPD liquid (Sigma Scientific, St. Louis, MO.) was diluted in dichloromethane (VWR Scientific, San Francisco, CA). An aliquot was added to a reaction vial containing 1 mL isooctane and derivatized with heptafluorobutyrylimidazole (Pierce, Rockford, IL) at 70 C for 30 minutes according to the procedure outlined in the AOAC method [2].

Equation 1. The acid hydrolysis of vegetable matter from triglycerides to glycerol, and then to 3-MCPD.

Results and Discussion


Derivatizing 3-MCPD with heptafluorobutyrylimidazole replaces the hydrogens on the diol groups with ester linkages to a perfluorinated propyl side chain. The molecular formula of the derivative is C3H5O2Cl(COC3F7)2, and Figure 1 shows the molecular structure. Figure 2 shows the electron impact (EI) ionization mass spectrum of the heptafluorobutyrylimidazole derivative of 3-MCDP from 60 to 510 m/z.
F

197 m/z

275 m/z

F F F F F O 169 m/z F

F O F F O 453 m/z 289 m/z -HCl 253 m/z Cl

Figure 1. The molecular structure and a suggested fragmentation pattern for heptafluorobutyrylimidazole derivative of 3-MCPD in electron impact ionization.

Abundance 100 90 80 70 60 50 40 30 20 10 0 100 150 200 250 300 350 400 450 500 m/z

169

69 197

253 289 100 453

Figure 2. Electron impact ionization mass spectrum of the 3-MCPD heptafluorobutyryl derivative at 70 eV for the 60 to 510 m/z mass range. The molecular ion [M]+ would be expected at 502 m/z.

Electron impact ionization produces a mass spectrum that lacks a molecular ion and has a base peak at m/z 169 from [C3F7]+ fragments. As seen from the fragmentation diagram of Figure 1, the ions at 169 m/z and 197 m/z contain no structural relevance to 3-MCPD and consequently can not be used to indicate 3-MCPD. This suggests the 453, 289, 275, and 253 m/z fragments, which contain 3-MCPD structure, be used for detection. In the AOAC collaborative study, several laboratories had difficultly detecting the 453 m/z fragment. This is not a problem for the 5973N due to the high-energy dynode arrangement and high transmission quadrupole which provide good signal for high molecular weight fragments. In fact, work with the standard showed good signal-to-noise for the 453 m/z ion in the scan mode even at only a few picograms injected. Selected ion monitoring using the four ions suggests detection at subpicogram levels is possible. In actual samples, matrix interferences may emerge and contribute to noise. However, the derivatization technique suggests applying electron capture negative ionization (ECNI) mass

spectrometry which provides more selective ionization than electron impact. Figure 3 shows the ECNI mass spectrum of the 3-MCPD derivative under standard conditions with methane buffer gas (tat is, source 150 C, methane at 2 mL/min). Unlike the EI results, the molecular ion (502 m/z) is detected, although at low relative intensity. Unfortunately like EI, the base peak at 213 m/z and next most abundant peak at 194 m/z, due to [OCOC3F7]+ and [OCOC2F7]+ fragments respectively, also contain no structural relationship to 3-MCPD. This leaves the 502, 482 and 446 m/z ions as good candidates for 3-MCPD detection and quantitation. Analysis of standards suggests that it would be possible to detect a few picograms in the scanning mode and less than one picogram in selected ion mode (SIM). Further optimization of ECNI is possible, such as a lower source temperature to take advantage of the low boiling point of the derivative, which may improve the spectrum and detection limits. The real advantage of ECNI is expected to be in typical food samples where the greater selectivity of ECNI will demonstrate a strong suppression of chemical noise and enhance method detection limits.

Abundance 100 90 80 70 60 50 40 30 20 10 0 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 m/z-->

213

194 482 179 233 446 502

Figure 3. Electron capture negative ion chemical ionization mass spectrum of derivatized 3-MCPD with methane buffer gas from 150 to 510 m/z. Note the presence of the molecular anion [M]- not seen in EI.

www.agilent.com/chem

Acknowledgments
The author is grateful to Phil Wylie and Norman Low for their contributions to the manuscript.

References
1. Reports of the Scientific Committee for Food, Food Sciences and Techniques, 36th Series, European Commission, Luxembourg, Belgium. Brereton, P., et al., Determination of 3-chloro1,2-propanediol in foods and food ingredients by gas chromatography with mass spectrometric detection: Collaborative study. Journal of the AOAC International, 2001. 84(2): p. 455-465.

2.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the U.S.A. November 8, 2001 5988-4287EN

SPE of Chlorinated Pesticides in Water

Application Brief

Condition: Add 5 mL acetone to the cartridge. Apply vacuum and discard the eluant. Repeat with 5 mL methanol then 5 mL water. Do not allow the sorbent to go dry at any point during this step. Load: Attach a sample reservoir to the top of the cartridge. Add 0.2 mL methanol to 20 mL of the water sample.1 Mix, then add to the cartridge. Apply the vacuum and discard the eluant. The flow rate should be no greater than 10 mL/min.2 Rinse: Add 3 mL water to the cartridge. Apply vacuum and discard the eluant. Leave the vacuum on for 30 seconds after all of the water has passed through the cartridge. Centrifuge the cartridge at 10001500 rpm for 5 minutes.3

Elution: Place a collection tube beneath the cartridge. Add 3 mL of acetone to the cartridge. Apply vacuum and collect the eluant. Concentrate to dryness under a stream of dry nitrogen4 (transfer to a small vial may be necessary). Dissolve the residue in 200 L acetone. Inject 2L into the GC.
1

Equipment
AccuBOND II ODS (C18) 6 mL/500 mg cartridge (P/N 188-1356) 25 mL sample reservoir (P/N 700-4007) coupling fitting (P/N 5185-5794) vacuum manifold (P/N 5185-5754, 10-port) (P/N 5185-5765, 20-port)

Volume up to 100 mL may be used. Adjust the amount of added methanol so that the final concentration is 1%. Using slower flow rates will result in slightly better recovery values. Centrifuging removes additional water which aids in sample concentration. The use of heat to aid in sample concentration may result in reduced recovery values.

Reagents
water (HPLC grade) methanol (pesticide grade) acetone (pesticide grade)

Compound
Trifluralin Chloroneb Propachlor Hexachlorobenzene (HCB) -BHC -BHC (lindane) -BHC Heptachlor Chlorothalonil -BHC Alachlor Aldrin DCPA Heptachlor epoxide o,p'-DDE Endosulfan 1 p,p'-DDE Dieldrin o,p'-DDD Chlorobenzilate Endrin p,p'-DDD o,p'-DDT Endosulfan II p,p'-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor
Water spiked at 50 ppb n=4

k
5.52 5.63 6.65 7.10 7.43 8.16 8.26 8.75 8.76 8.93 9.03 9.38 9.80 10.86 11.49 11.55 12.15 12.52 13.12 13.35 13.85 14.17 14.25 14.48 15.27 15.64 16.07 18.50

x
90 90 87 69 73 49 83 62 86 97 83 67 66 72 66 81 88 86 68 86 97 80 64 91 79 85 90 92

Std. Dev.
12 6 7 13 3 18 4 7 2 4 5 5 3 7 11 9 5 11 9 8 12 8 14 7 6 3 7 11

Pond Water Extract Containing Chlorinated Pesticides


DB-608 30 m x 0.53 mm I.D. P/N: 125-1730 Film Thickness: 0.83 m Carrier: Helium at 40 cm/sec (measured at 140C) Oven: 140C for 2 min 140-240C at 10/min 240C for 5 min 240-265C at 5/min 265C for 5 min Injector: Megabore direct, 250C Detector: ECD, 325C Nitrogen makeup gas at 30 mL/min Column: 1. Hexachlorobenzene 2. a-BHC 3. g-BHC 4. b-BHC 5. Heptachlor 6. d-BHC 7. Aldrin 8. Heptachlor epoxide 9. Endosulfan 1 10. p,p'-DDE 11. Dieldrin 12. Endrin 13. p,p'-DDD 14. Endosulfan II 15. p,p'-DDT 16. Endrin aldehyde 17. Endosulfan sulfate 18. Methoxychlor

Pond Water Extract Containing Chlorinated Pesticides


Column: DB-608 30 m x 0.53 mm I.D. P/N: 125-1730 Film Thickness: 0.83 m Carrier: Helium at 40 cm/sec (measured at 140C) Oven: 140C for 2 min 140-240C at 10/min 240C for 5 min 240-265C at 5/min 265C for 5 min Injector: Megabore direct, 250C Detector: ECD, 325C Nitrogen makeup gas at 30 mL/min 1. Trifluralin 2. Chloroneb 3. Propachlor 4. Hexachlorobenzene 5. Chlorothalonil 6. Alachlor 7. DCPA 8. o,p'-DDE 9. o,p'-DDD 10. Chlorobenzilate 11. o,p'-DDT

k = partition ratio (a measure of retention) x = % recovery To order or for technical assistance, contact your local authorized Agilent distributor, or visit our website at www.agilent.com/chem.
Agilent Technologies, Inc. 2001 Information, descriptions and specifications in this publication are subject to change without notice. All rights reserved. Reproduction, adaptation, or translation without prior written permission is prohibited, except as allowed under copyright laws. Printed in the U.S.A. October 25, 2001 5988-4057EN

The Analysis of Organophosphate Pesticides by LC/MS Application

LC-MS

Authors
Paul Zavitsanos Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA Paul Yang MOE Ontario Canada Lorna Grey MOE Ontario Canada

Additionally, LC-MS provides unequivocal identification of each pesticide, even if the pesticide was not completely resolved from neighboring eluants. Traditional UV detection cannot provide the required specificity because many of the pesticides within the same class exhibit similar UV spectra.

Sample case
A mixture of organophosphate pesticides and an internal standard were analyzed using an Agilent 1100 LC/MS with an ESI source (Table 1).

Table 1. Mixture of Organophosphate Pesticides

Abstract
Organophosphate pesticides were readily analyzed using liquid chromatography-mass spectrometry with electrospray ion source. Sensitivity and selectivity were significantly better than using a diode-array UV detector.

Elution order 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Compound Mevinphos isomer 1 Dimethoate Mevinphos isomer 2 Dichlorvos Azinphos methyl Parathion methyl Malathion Diazinon Triphenyl orthophosphate* Parathion ethyl Phorate Reldan Ronnel Terbuphos Dursban Ethion Temephos

[M+H]+ 225 230 225 221 318 264 331 305 327 292 261 322 321 289 350 385 467

Concentration g/mL 0.2 0.5 0.5 0.5 0.05 0.2 0.5 0.2 1.0 0.1 0.1 0.5 0.1 0.2 0.1 0.2 0.1

Overview
Liquid chromatography-mass spectrometry (LC-MS) is rapidly becoming a routine technique for efficient trace analysis of polar pesticides in various types of samples. In comparison to existing methodologies, such as gas chromatography-mass spectrometry (GC-MS) and ultraviolet (UV) detection, LC-MS considerably simplifies cleanup procedures, reducing both time of analysis and method development time.1 LC-MS with an electrospray ion (ESI) source avoids the thermal degradation of labile pesticides encountered with GC and eliminates the need for preliminary derivatization to increase compound volatility.

* Internal standard

Method summary
Column 2.1 mm id 5 cm long, filled with

Results
Simultaneous UV (220 nm) and MS detector outputs are compared in Figure 1. The MS plot is a composite of all the individual extracted ion chromatograms. Each was obtained at the [M+H]+ value given in Table 1, and are separated and stacked in Figure 2 for easy comparison.

3.5 m particles, C18 chemistry


20 mM ammonium acetate vs. acetonitrile

mobile phase gradient 5 % to 95 % acetonitrile in 4 minutes Hold 2 minutes


Splitless 400 L/min flow 3 L injection volume Scan data 120 to 600 m/z

SIM data as per Table 1. 95 msec dwell/ion in two groups

5 0 -5 -10 -15 -20 -25 0 2 4 6 8 Normalized UV Plot 10 12 14 min.

7000000 6000000 5000000 4000000 3000000 2000000 1000000 0 2 4 6 8 10 12 14 min.

MS selected ion mode (SIM), overlayed and normalized

Figure 1. Comparison of UV and MS chromatograms.

17 16

m/z 467 385

15 350 14 289 13 321 12 322 11 261 10 292 9 327 8 305 7 331 6 264 5 348 4 221 2 230 1
2

3 225
4 6 Minutes 8 10

Figure 2. Stacked normalized extracted ion chromatograms for compounds 1 through 17.

Figures 3 through 6 show the resulting normalized mass ion spectra for each compound included in Table 1.

Mevinphos isomer 1

193.0

[M+H]+

225.0

Dimethoate

[M+H]+

232.0

230.0

226.0

[M+H]+

192.9

222.9

Dichlorvos

[M+H]+

220.9

237.9

100

150

200

224.8

250

[M+NH4]+

239.9

242.9

[M+NH4]+

Mevinphos isomer 2

225.0

242.1

283.0
300

m/z

Figure 3. Stacked normalized ion mass spectra for compounds 1 through 4.

Diazinon

Malathion

Azinphos methyl

Parathion methyl

100

118.9 133.1 150.0 160.0

132.0

150

200

235.4

250

Figure 4. Stacked normalized ion mass spectra for compounds 5 through 8.


[M+H]+ 265.1 285.0 [M+H]+ 305.1 291.4 Background 263.9 260.9 290.1 291.5 290.4 [M+H]+ [M+H]+ 332.0 348.1 331.0 340.1 318.8 317.9

300

306.1

m/z

Triphenyl orthophosphate

[M+H]+

327.0

Parathion ethyl

[M+H]+

292.1

235.1

[M+H]+

Phorate

263.0

261.0

293.0

304.9

[M+H]+

Reldan

325.8
100 150 200 250 300

321.9 323.9
324.9

328.0
m/z

Figure 5. Stacked normalized ion mass spectra for compounds 9 through 12.

[M+H]+ 324.9 101.9

322.9 232.9 253.3 244.8 186.9 198.9 234.9 [M+H]+


150 200 250

Ronnel

320.8

Terbuphos
289.0

100

291.0

98.0

300

324.0

339.1

m/z

350.9

Dursban

[M+H]+

353.9

351.9

[M+H]+

Ethion

Temephos
[M+NH4]+ 386.9 467.0
400 425 450

325

350

375

[M+H]+

475

485.0

484.0

m/z

Figure 6. Stacked normalized ion mass spectra for compounds 13 through 17. 7

www.agilent.com Conclusions
When determining organophosphate pesticides using LC-MS with an ESI source:
All the tested organophosphate pesticides

ionized well and gave definite [M+H]+ ions


Sensitivity and selectivity are significantly

better than using diode-array UV detector


Overall chromatography and analysis is simple

and straightforward
Positive identification and quantification are

performed using integrated software For more information, contact your local Agilent sales representative, or visit www.agilent.com.

Reference
1. Elbert Hogendoorn and Peit van Zoonen, Journal of Chromatography A, 892 (2000) 435-453.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies, Inc. Printed in the USA September 13, 2001 5988-3774EN

A New Approach to the Analysis of Phthalate Esters by GC/MS Application


Gas Chromatography/Mass Spectrometry March 2001

Authors
Cameron George Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, California 95630-4714 USA Harry Prest Agilent Technologies, Inc. 1601 California Avenue Palo Alto, California 94304-1111 USA

Introduction
The widespread use and manufacture of plastics have made the phthalate esters one of the most ubiquitous classes of compounds in our everyday environment. These plasticizers increase polymer flexibility due to their function as intermolecular lubricants. Because they are additives and not reagents, they are not chemically bound in the polymer and are available to leach from the matrix. Phthalates are also components of cosmetics, detergents, building products (flooring, sheeting, films), lubricating oils, PCB substitutes, carriers in pesticide formulations and solvents. Consequently, the potential for human exposure is very high. Toxicological studies have linked some of these compounds to liver and kidney damage, and to possible testicular or reproductive-tract birth defect problems, characterizing them as endocrine disruptors. Scientists at the U.S. Centers for Disease Control have, for the first time, documented human exposure to phthalates by determinations of the monoester metabolites in human urine [1]. Their work leads to the conclusion that phthalate exposure is both higher and more common than previously suspected. Of particular concern were the significantly higher concentrations of the dibutyl phthalate metabolite in urine of women of childbearing age (20-40 years) than in other portions of the population. The presence of phthalate esters in polyvinyl chloride (PVC) toys has generated the most

Abstract
A new instrumental method for the determination of 29 phthalate esters, including six recently banned from baby toys by the European Union, using positive chemical ionization and retention-time locking is described. Positive chemical ionization provides a high degree of selective ionization for the phthalates, primarily producing spectra in which the protonated molecule (M+1) is the base peak. This provides easy discrimination among the phthalates on the basis of their molecular weight, while retentiontime locking increases confidence in the identification of the various isomers. In this approach, both pure compounds and technical mixtures are considered. Although this work focuses on the more commonly used 1,2-substituted esters, the 1,3-isomers and 1,4-isomers are also characterized. The combination of positive chemical ionization and retention-time locking makes the method rugged, durable and applicable to a wide variety of matrices.

controversy. While regulators in Greece have completely banned soft PVC toys, Austria, Denmark, Finland, France, Germany, Norway and Sweden have unilaterally banned phthalates in PVC toys for children under three years old. In December of 1999, the European Union (EU), concerned with a serious and immediate risk to children, placed an emergency ban on six of the phthalate esters in soft PVC toys and childcare products meant to be placed in the mouths of children under the age of three [2]. None of the six banned phthalates may exceed 0.1% by weight. These heightened concerns suggest the need for an improved method of detecting and characterizing phthalate esters which is applicable to a wide variety of matrices. This application note describes such an analytical method.

O R O O R' O

O R

Phthalate Structure and Mass Spectra


The three primary structures of phthalates are shown in Figure 1. Although there are three possible positions for the ester linkages, the most commonly used phthalates are based on the 1,2-benzenedicarboxylic acid structure (top). There are an infinite number of possible alkyl side chains, (R) and an infinite number of combinations of the side groups (R and R'). For example, the diisononyl phthalate consists of an array of compounds due to the isomeric branched-chain alkyl groups on both side chains. For phthalate esters with saturated alkyl side chains (without oxygen), the most intense peak in the electron impact (EI) ionization mass spectrum at 70 eV is always at m/z 149 due to the rapid formation and stability of the ion shown in Figure 2. (The only exception is R=R'=CH3 where the base peak is at m/z 163).

O R' O

O R

R' O O

Figure 1. Phthalic ester (top) or the 1,2-benzenedicarboxylic acid ester, isophthalic ester (middle) or the 1,3-benzenedicarboxylic acid ester, and terephthalic ester (bottom) or the 1,4-benzenedicarboxylic acid ester. R and R' represent alkyl side chains which may be branched and contain oxygen.

OH

Figure 2. The most abundant ion in the mass spectra of the phthalate esters with saturated alkyl side chains; m/z 149. The exception is for dimethyl phthalate where both R and R' are CH3 and so the H on the oxygen is replaced by CH3 and consequently m/z 163 becomes the base peak.

Invariably, the molecular ion is very weak or altogether absent; other fragments that provide information on the phthalate identity are also of very low abundance. As an example, consider the EI mass spectrum of dibutyl phthalate, one of the six banned by the EU, and bis(4-methyl-2-pentyl) phthalate in Figure 3. Identifying fragments have relative intensities of less than 10%. Gas chromatography provides some separation of the phthalates, but with the array of possible isomers and essentially a single identifying ion (i.e., m/z 149), distinguishing the individual phthalates of concern is difficult. More confident identification of the phthalates is possible using chemical ionization mass spectrometry in conjunction with retention-time locking (RTL).

1000000 900000 800000 700000 Abundance 600000 500000 400000 300000 200000 100000 0 40 60 80 100 120

149 O O O O

57

76

104

121
140

167
160

187
180

205
200

223 250
220 240 260

278
280

296 313
300 320

341
340

149 800000

O
700000 600000 Abundance 500000 400000 300000 200000 100000 55 0 40 60 76 80 104 100 121 120 140 167 160 189 180 m/z 208 200 233 220 240 260 251 278 280 306 300 334 320 340

O O O

Figure 3. Electron impact ionization mass spectra of di-n-butyl phthalate (upper panel) and bis(4-methyl-2-pentyl) phthalate (lower panel) from m/z 50 to 350 at 70 eV. Notice the lack of intense fragments and molecular ions. The molecular weights are 278 and 334 g/mole, respectively.

Retention-time locking allows compound retention times achieved on any one Agilent 6890 gas chromatograph (GC) to be replicated to within a few seconds on any other Agilent 6890 gas chromatograph (GC) applying the same GC method [3-5]. RTL is a powerful approach to compound identification. RTL allows the creation of compound acquisition methods and quantitation databases that can be reproduced in any laboratory, anywhere, because a compound can have a universally fixed and reproducible retention time. It is important that RTL be applied in conjunction with the appropriate detection scheme and sample reparation methods. Chemical ionization provides a more selective form of ionization than electron impact [6]. By judicious choice of the reagent gases, the degree of compound fragmentation can be controlled to a certain extent. In positive chemical ionization, methane reagent gas usually provides more fragmentation than gases of higher proton affinity such as ammonia. Less fragmentation would be helpful in identifying the phthalates. Instead of all phthalates generating a single, similar ion, positive ionization can provide phthalate ester molecular weights.

Injection Parameters Injection Mode Injection Port Temperature Pulse Pressure & Time Purge Flow & Time Gas Saver Flow & Time Oven Parameters Temperature Program 50.00C/min 15.00C/min Oven Equilibrium Time MSD Transfer Line Temp Column Parameters GC column (122-5532) Initial Flow & Mode Detector & Outlet Pressure Mass Spectrometer Parameters Tune Parameters Electron Multiplier Voltage Solvent Delay Scan Parameters Quadrupole Temperature Source Temperature Ammonia Gas Flow (MFC setting) Miscellaneous Parts Septa Liner GC column ferrule MSD interface ferrule

Pulsed Splitless 300C 25.0 psi 20.0 mL/min 20.0 mL/min

1.00 min 3.00 min 3.00 min

80C 200C 350C 0.25 min 325C

1.00 min 0.00 min 2.00 min

DB-5MS 30 m; 0.25 mm i.d.; 0.25 m film 1.2 mL/min Constant Flow MSD Vacuum

PCI Autotune (NH3) Autotune + 400V 4.00 min 194 - 550 m/z 150C 250C 0.5 mL/min (10%)

Experimental
Phthalate esters were obtained from Ultra Scientific (North Kingstown, RI), AccuStandard (New Haven, CT), and ChemServices (West Chester, PA) as neat compounds and mixtures. Dilutions were made in isooctane (Burdick and Jackson Grade, VWR Scientific). The configuration and operating parameters of the Agilent 6890Plus GC (standard 120V or faster ramping 220V), 7683 Automatic Liquid Sampler and 5973N MSD with CI option used for acquiring the data are given in the following tables. PCI reagent gas purities were 99.99% or higher.

5182-0739 5062-3587 5181-3323 5082-3508

BTO septa (400C) Deactivated 4 mm i.d. single taper 250 m Vespel 0.4 mm i.d. graphitized Vespel

Results and Discussion


As expected using methane as the reagent gas, the PCI mass spectra of the phthalates show ions corresponding to the protonated molecule [M+H]+ and adducts [M+C2H5]+ and [M+C3H5]+. Because of the relatively vigorous fragmentation produced by methane, the spectra of the dialkyl phthalate esters still resemble that produced in EI. In most cases, the fragment at m/z 149 is the base peak, however ions at m/z M+1, M+29 and M+41 are relatively intense with [M+H]+ from 10% to 30% (Figure 5). The dialkyl phthalate spectra also show a fragment corresponding to loss of one of the alkyloxy side chains to produce an ion shown in Figure 4. This is the most intense fragment for the dimethyl and diethyl phthalates and for the dibutyl and dipentyl (diamyl) phthalates, about 75% of the 149 base peak. As the length of ester alkyl chain increases, the intensity of this fragment decreases. (Apparently, in the dialkyl isophthalates, loss of the alkyl side chain not accompanied by the oxygen may be a preferred route.) Although positive chemical ionization with methane provides more information than EI on phthalate identity, the methane reagent is still rather unselective in ionization and will produce more chemical noise in the background, complicating identification in complex matrices.

OR

O
Figure 4. One of the most intense fragments in the methane PCI spectra of the phthalate esters is formed by loss of one of the alkyloxy side groups.

Applying ammonia as the reagent gas in PCI to reduce chemical noise and enhance identification of the phthalates is a more useful approach. The relatively gentle ionization produces protonation of the dialkyl phthalates, with m/z M+1 the base peak in their spectra. When combined with retentiontime locking, identification of phthalates becomes further simplified. Compare the spectra of the di-n-butyl phthalate acquired using methane versus ammonia as the reagent gas (Figure 5). The protonated molecule is the single dominant peak in the ammonia PCI mass spectrum of the di-n-butyl phthalate, and the adduct at m/z 296 ([M+NH4]+) is relatively small.

149
280000 260000 240000 220000 200000 180000 Abundance 160000 140000 120000 100000 80000 60000 40000 20000 0 80

O O O 205 177 O 205

177 123 84 105


100 120

279 189 223


200 220 240

135
140

163
160

251 263
260 280

307
300

319 333
320

180

1300000 1200000 1100000 1000000 900000 800000 Abundance 700000 600000 500000 400000 300000 200000 100000 0 80 100 120 140 160 180 200 m/z 220 240 260

279

94

108

122

136 148

166 177

194 205

222

240 252
280

296
300 320

Figure 5. PCI methane (upper panel) and ammonia (lower panel) mass spectra of di-n-butyl phthalate. The PCI methane mass spectrum shows substantial fragmentation but relative to the EI spectrum in Figure 3, high abundance for the higher m/z ions such as the protonated molecule at m/z 279. The ion at m/z 205 is generated by loss of an oxybutyl fragment; a process described in Figure 4. The PCI-ammonia mass spectrum consists almost completely of the protonated molecule.

This implies an easy method for identification. Whereas the EI spectra of the phthalates most frequently result in a base peak at m/z 149, the dialkyl phthalate PCI-ammonia spectra have base peaks at m/z = M+1. All dialkyl phthalates molecular formulas can be expressed as C8H6O4(CH2)y (CH2)x. These phthalates have (nominal) molecular masses given by M = 166 + (x + y)14, or M = 166 + (w)28, where x and y are the side chain lengths, and the second formula applies to symmetrical side chains (i.e., x = y = w). For example, di-n-butyl phthalate has x = y = 4, and therefore a (nominal) molecular mass of 278 which produces m/z 279 as the base

peak. Interestingly, the PCI-ammonia spectra of the dialkyl isophthalates and terephthalates appear to have base peaks at m/z M+18 due to [M+NH4]+. Because of the greater steric access to the ester linkages, adduct formation may be preferred. Table 1 gives the phthalate names, CAS numbers, molecular formula, nominal molecular mass, base peak in the PCI-ammonia spectrum and the RTL elution times. These retention times are "locked" relative to diphenyl phthalate, which has been chosen as the RTL locking compound and locked to elute at 9.450 min. Notice that the branched chain isomers elute prior to their straight chain forms on this column phase.

Table 1. Phthalate compound names, Chemical Abstracts Services numbers (CAS), molecular weights (M. Wt.), molecular formulas, nominal base peak in the PCI-ammonia spectrum and retention time (RT) in minutes. Retention times are locked relative to diphenyl phthalate (9.450 min). Retention time ranges are given for the isoalkyl phthalate technical mixtures. Phthalates banned by the EU are indicated by an asterix*. Benzyl benzoate is included since it is used as a surrogate in U.S. Environmental Protection Agency Method 8061. Name dimethyl phthalate dimethyl isophthalate diethyl phthalate diethyl terephthalate benzyl benzoate diisobutyl phthalate di-n-butyl phthalate* bis(2-methoxyethyl) phthalate diamyl phthalate bis(2-ethoxyethyl) phthalate butyl benzyl phthalate* diphenyl phthalate diphenyl isophthalate dicyclohexyl phthalate bis(4-methyl-2-pentyl) phthalate diisohexyl phthalates dihexyl phthalate dibenzyl phthalate hexyl-2-ethylhexyl phthalate bis(2-n-butoxyethyl) phthalate bis(2-ethylhexyl) phthalate* di-n-octyl phthalate* dioctyl isophthalate diisononyl phthalates* dinonyl phthalate diisodecyl phthalates* didecyl phthalate diundecyl phthalate didodecyl phthalate ditridecyl phthalate CAS 131-11-3 1459-93-4 84-66-2 636-09-9 120-51-4 84-69-5 84-74-2 117-82-8 131-18-0 605-54-9 85-68-7 84-62-8 744-45-6 84-61-7 146-50-9 146-50-9 84-75-3 523-31-9 75673-16-4 117-83-9 117-81-7 117-84-0 137-89-3 28553-12-0 84-76-4 26761-40-0 84-77-5 3648-20-2 2432-90-8 119-06-2 M. Wt. 194 194 222 222 212 278 278 282 306 310 312 318 318 330 334 334 334 346 362 366 390 390 390 418 418 446 446 474 502 530 Molecular Formula C8H4O4(CH3)2 C8H4O4(CH3)2 C8H4O4(C2H5)2 C8H4O4(C2H5)2 C14H12O2 C8H4O4(C4H9)2 C8H4O4(C4H9)2 C8H4O4(C2H4OCH3)2 C8H4O4(C5H11)2 C8H4O4(C2H4OC2H5)2 C8H4O4(C4H9)(CH2C6H5) C8H4O4(C6H5)2 C8H4O4(C6H5)2 C8H4O4(C6H11)2 C8H4O4(CH3C5H10)2 C8H4O4(C6H13)2 C8H4O4(C6H13)2 C8H4O4(CH2C6H5)2 C8H4O4(C2H5C6H12)(C6H13) C8H4O4(C2H4OC4H9)2 C8H4O4(C2H5C6H12)2 C8H4O4(C8H17)2 C8H4O4(C8H17)2 C8H4O4(CH3C8H17)2 C8H4O4(C9H19)2 C8H4O4(CH3C9H18)2 C8H4O4(C10H21)2 C8H4O4(C11H23)2 C8H4O4(C12H25)2 C8H4O4(C13H27)2 Base Peak 195 212 223 240 230 279 279 283 307 311 313 319 319 331 335 335 335 347 363 367 391 391 408 419 419 447 447 475 503 531 RT (min) 4.32 4.54 4.81 5.06 5.62 5.95 6.40 6.57 6.94 7.13 8.42 9.45 10.30 9.32 6.93 7.55 - 8.28 8.34 10.51 8.84 8.98 9.32 10.28 10.84 9.40 - 11.10 11.19 10.16 - 11.86 12.05 12.87 13.65 12.01 - 13.81

Technical formulations of the isoalkyl phthalates tended to contain substantial amounts of the straight chain isomer, which may convolute quantitation as well as peaks that may be construed as originating from nonequivalent side chains i.e., x y in equation 1). These impurities can be detected as M14 around the mass of the nominal isomer. For example, technical grade diisononyl phthalate contains compounds that generate ions at m/z 391 (minor), 405, 433, and 447 in addition to the nominal diisononyl phthalate compound at m/z 419. The gentle ionization of ammonia reagent gas, the elution times and the study of the

spectra of other pure isomers, such as the dinonyl phthalate, suggest that these fragments are not formed by the PCI process but are due to these different alkyl side chain impurities (Figure 6). To demonstrate the utility of the PCI-ammonia compared to conventional EI analysis, consider the chromatograms presented in Figure 7. The EI spectra of the phthalates produce m/z 149 as the base peak for all the phthalates present; distinguishing ions are minor constituents (<10% relative intensity), making identification complicated. However, by examining the appropriate PCI-ammonia ions, the various phthalates are easily distinguished.

m/z 419

Abundance

m/z 433

m/z 405

m/z 447

9.40

9.60

9.80

10.00

10.20

10.40 Time

10.60

10.80

11.00

11.20

11.40

Figure 6. PCI-Ammonia extracted ion chromatogram of technical diisononyl phthalate. The diisononyl appears as the major component at m/z 419 while ions at m/z 405, 433, and 447 indicate alkyl chains shorter by one CH2 unit and longer by one and two CH2 units respectively.

EI TIC
Abundance 9.50

10.00

10.50

11.00

11.50

12.00

12.50

13.00

13.50

14.00

Abundance

m/z 149

9.50

10.00

10.50

11.00

11.50 di-n-nonyl

12.00

12.50

13.00

13.50

14.00

di-n-undecyl di-n-dodecyl

di-n-decyl

diisodecyl Abundance

diisononyl

ditridecyl

9.50

10.00

10.50

11.00

11.50

12.00 Time

12.50

13.00

13.50

14.00

Figure 7. Chromatograms of dinonyl, diisononyl, didecyl, diisodecyl, diundecyl, didodecyl, ditridecyl phthalate esters in EI (upper panel), EI as an extracted ion chromatogram at m/z 149 (middle panel), and PCI-extracted ion chromatogram with ions selected for the individual phthalate classes as given in Table 1. The EI information is insufficient to identify coeluting phthalates. For example, the dinonyl and diundecyl phthalates are "buried under" the signals from the isodecyl and ditridecyl phthalates.

Conclusions
Applying GC - electron impact (EI) mass spectrometry to the determination of phthalates requires full chromatographic separation. The EI spectra of the phthalates are distinguished only by ions of very low intensity. In EI, the phthalates produce a single common ion (m/z 149) as the most intense spectral peak, regardless of the alkyl side chain substitution. Applying tandem mass spectrometry (i.e., EI/MS/MS) gains nothing, because there is a common parent ion, and therefore any daughter ions would also be non-unique. However, the combination of positive chemical ionization with retention-time locking allows even complex mixtures of phthalates to be characterized. Ammonia reagent gas produces the protonated molecule as the base peak, which immediately allows the phthalates to be distinguished on the basis of their substitution. PCI is also an advantage in complex matrices, where the non selective ionization of EI produces a high chemical background. This method should therefore be suitable for use in phthalate determinations in environmental media, plastics, cosmetics and many other matrices. Locking the retention time enhances confidence in the characterization of the various phthalate isomers on the basis of their definitive retention time. This is especially helpful for determinations using selected ion monitoring (SIM), since SIM groups need not be edited after column maintenance [4]. The data in Table 1 facilitate the development of a SIM method. The extension of the method to phthalates which elute at higher temperatures ( >350C) is also easily accomplished.

References
1. Blount, B.C., et al., Levels of seven urinary metabolites in a human reference population. Environmental Health Perspectives, 2000. 108 (10): p. 979-982. 2. Official Journal of the European Communities, Decision 198/815/EC. 1999, European Commission; European Union Scientific Committee on Toxicology, Ecotoxicology, and the Environment. 3. Giarrocco, V., B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications. 1998, Publication number (23) 5966-2469E, Agilent Technologies. 4. Prest, H. and P. Cormia, Retention Time Locking: Advantages in GC/MS SIM Analysis. 1999, Publication number (23) 5967-3797E, Agilent Technologies. 5. Prest, H. and K. Weiner, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries. 1999, Publication number (23) 5968-8657E, Agilent Technologies. 6. Harrison, A.G., Chemical Ionization Mass Spectrometry. Second ed. 1992: CRC Press.

For more information


For more information on our products and services, you can visit our site on the World Wide Web at: http://www.agilent.com/chem.

Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies All rights reserved. Reproduction and adaptation are prohibited. Printed in the U.S.A. March 7, 2001 5988-2244EN

Development of an LC/MS Method for the Analysis of Rodenticides Application Note


Environmental
Friedrich Mandel and Juergen Wendt Agilent Technologies Raffaele Vistocco and Christian Bachmann Chromatography Department, A.P.P.A. Bolzano (Italy)

Introduction
Rodents such as rats and mice must be controlled because they destroy food supplies and serve as vector hosts for human diseases such as hantavirus. Although individual animals or small groups can be removed by trapping, rodenticides are frequently used in rodent control. Most rodenticides are also toxic to humans and domesticated animals such as dogs. In this work, an LC/MS method was developed to monitor several coumarin rodenticides in complex matrices. Anticoagulant poisons work by interfering with the blood clotting mechanism. After repeated ingestion of relatively small doses, a lethal dose accumulates and causes death due to internal bleeding. Anticoagulant poisoning can also cause spontaneous bleeding from the nose, gums and the gastrointestinal and urinary tracts.

Development of an LC/MS Method for the Analysis of Rodenticides

Agilent Technologies

Experimental
All experiments were done on an Agilent 1100 Series LC/MSD system that was comprised of a binary pump, vacuum degasser, autosampler, thermostatted column compartment with column-switching valve, diode-array detector, and an LC/MSD. The LC/MSD was used with either the electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) source. Complete system control and data evaluation was done on the Agilent ChemStation for LC/MS. Reagent grade chemicals and HPLC grade solvents were used for preparing mobile phases and standards. The compounds used for this study are all classified as coumarin rodenticides due to the common coumarin moiety in the structures (Figure 1). Method development included assessing ionization technique (ESI and APCI), ionization polarity and fragmentor voltage to determine the best conditions for the three analytes. In order to test the performance of the method, spiked sausage extract and dog stomach extract were analyzed. The samples were spiked at either 0.5 or 5 ppm, then subjected to a general extraction and sample cleanup.1 The spiked samples were extracted in acetone for four hours. The acetone was then evaporated to a volume of approximately 50 ml and extracted with 100 ml of methylene chloride. The acetone layer was dried with anhydrous sodium sulfate, filtered, and evaporated to dryness. After redissolving in 2 ml of acetone, the extract was applied to a column containing 15 g silica gel and 1 g charcoal. After elution with methylene chloride/benzene/acetone (10/2/2 v/v), the eluate was evaporated to dryness and redissolved in 5 ml of methanol. The sample extracts were then analyzed by ESI-LC/MS.
O O CH OH H2C OH O C CH3 O O O

OH

Difenacoum C31H24O3 mass 444.173


O O

Coumatetralyl C19H16O3 mass 292.11


Figure 1. Coumarin rodenticides.

Coumafuryl C17H14O5 mass 298.084

Results and Discussion


Both coumafuryl and coumatetralyl showed poor response in positive mode ESI compared to negative mode ESI (Figure 2). The positive mode response for these compounds was better in APCI than ESI, but the negative mode APCI response for coumatetralyl was weak compared to ESI. Based on these results, ESI negative mode was selected for the optimized method.

Development of an LC/MS Method for the Analysis of Rodenticides

Agilent Technologies

10000 0

Positive m/z 299

ANALYSIS METHOD Chromatographic Conditions Column: 150 2.1 mm Zorbax XDB-C18, 5 m (p/n 993700-902) Mobile phase: A = 2 mM ammonium acetate in water B = methanol Gradient: start with 30% B at 2 min 50% B at 4 min 100% B Flow rate: 0.4 ml/min from 0 to 2 min, then 0.5 ml/min Column temp: 50C Injection vol: 2 l Diode-array detector: signal: 280, 16 nm; reference: 550, 100 nm

Negative m/z 297 40000 0 Positive m/z 293 1000 0 10000 0 Positive m/z 445 20000 0 20000 0 0 1 2 3 4 5 6 7 min Negative m/z 443 Negative m/z 291

coumafuryl

coumatetralyl

difenacoum

ESI-MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: Stepsize: Peakwidth: Time filter: Scan: SIM ions (negative mode):

Fragmentor:

ESI 10 l/min 40 psig 350C 2500 V (positive and negative) 0.1 0.2 min On m/z 150500 296, 297, 298 Coumafuryl 290, 291, 292 Coumatetralyl 442, 443, 444 Difenacoum variable 180 V (50275) 100 V (280) 120 V (400)

Figure 2. Comparison of positive and negative ionization modes for electrospray LC/MS.

Typically, positive ions will fragment more easily than negative ions, so the fragmentor voltage was optimized for both modes. For example, difenacoum showed extensive fragmentation at 160 V in positive ion mode but no fragmentation at the same voltage in negative ion mode. Optimized fragmentor voltages were used for each ion in the final method. For samples extracted from a complex matrix, the deprotonated molecule as well as the ions 1 m/z above and below can be monitored to check for coeluting artifacts. In the final method, this strategy was employed so a set of three ions was used for each analyte. The ions for each analyte were time-programmed as a SIM ion group in order to maximize the dwell time and thus maintain maximum sensitivity. Figure 3 shows the results for the

spiked sausage extract. For this analysis, data was collected in scan mode because the coumafuryl and difenacoum were present in sufficient concentration. The mass spectra from the spiked analytes show good quality at the 5 ppm level. The dog stomach extract was spiked at a lower level (0.5 ppm) so SIM analysis was done for this sample (Figure 4). Only the spiked analyte, coumafuryl, shows a significant signal. This work demonstrates that anticoagulant rodenticides can be detected by APCI and ESI in both positive and negative ion modes. Negative mode ESI-LC/MS was found to be the best choice for the three target compounds. This LC/MS method was capable of detecting the rodenticides in complex matrices.

Development of an LC/MS Method for the Analysis of Rodenticides

Agilent Technologies

coumafuryl
250000 150000 50000 m/z 297

coumafuryl
60000 m/z 297

2.833

2.882

40000 443.1 20000 0

coumafuryl

Max: 153123 311.2 250000 150000 265.1 343.2 50000 m/z 291

60000 40000 20000 1 2 3 4 5 minutes 6 200 300 400 m/z 0

m/z 291

difenacoum

difenacoum
297.0 60000 Max: 60723 5.374 40000 20000 0 211.1 0 1 2 3 4 5 6 min m/z 443

250000 150000 50000

m/z 443

3 4 5 minutes

200 300 400 m/z

Figure 4. Extracted ion chromatograms from the analysis of dog stomach extract spiked with 0.5 ppm of coumafuryl. Data was collected in negative ion using SIM mode.

Figure 3. Extracted ion chromatograms (left) and mass spectra (right) from the analysis of sausage extract spiked with 5 ppm of coumafuryl and difenacoum. Data was collected in negative ion using scan mode.

References
1. Faucannet, V., Pouliquen, H., and Pinault, L., Journal of Analytical Toxicolocy, 27, 1997.

For more information on our products and services, you can visit our site on the World Wide Web at: http://www.agilent.com/chem. Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies All rights reserved. Reproduction and adaptation is prohibited. Printed in the USA June 2000 (23) 5968-0561E

Authors
Friedrich Mandel and Juergen Wendt are application chemists at Agilent Technologies in Waldbronn, Germany. Raffaele Vistocco and Christian Bachmann are in the Chromatography Department of A.P.P.A. Bolzano in Italy.

Fast Screening of Pesticides and Endocrine Disrupters Using the Agilent 6890/5973N GC/MSD System, Part II

Application
Gas Chromatography May 2000
for 567 of the most common pesticides and endocrine disrupters of concern worldwide. A GC/MSD method was developed based on the standard 42-min method1 to screen for all 567 of the most common analytes. A specific combination of column stationary phase, carrier gas flow rate, and oven temperature programming is required to lock all the compounds to an expected retention timetable2. Compound identification based only on spectral searching alone is difficult when analyzing extracts containing significant sample matrix content because of overlapping peaks and noisy baselines. The new screening tool, integrated within Agilent's ChemStation for MSD, searches for all 567 compounds. It first checks and integrates four characteristic ions within the expected time window and then prints a report showing "hits" and "possible hits" (ratios of characteristic ions that do not match the expected values in the library within specified limits). In Part I of the MSD fast screening application brief 3 , a 10 m 0.1 mm 0.1 m Agilent HP-5 column was used to increase analysis speed up to fourfold. In this application brief, a 15 m 0.25 mm 0.25 m Agilent HP-5MS column was used. The faster methods were scaled exactly as predicted by using a combination of Agilent's method translation (MTL) and RTL software. Because scaling was exact, these faster methods can be used with precisely-scaled pesticide libraries, making the screening process even more powerful and adaptable to individual needs.

Authors C. Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, Delaware 19808-1610 USA Michael J. Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract Agilent Technologies' new, fast GC/MSD method can significantly speed up the screening of pesticides. Agilent's GC Method Translation software (available free from the Agilent Technologies Web site, http://www.chem.agilent.com/cag/ servsup/usersoft/main.html#mxlator) was used in developing the new method based on the standard 42-min method. A 15 m 0.25 mm 0.25 mm Agilent HP-5MS column was used to increase analysis speed up to fourfold. The time savings were implemented in increments (down to 10.5 minutes) to verify the predictability of scaling and the effect of scaling on the signal-to-noise ratio. Key Words
RTL, pesticide, environmental, screening, fast GC, method translation, 5973, 6890, MTL

Introduction
Analysts want faster analyses to improve laboratory productivity. Often, when speeding up GC methods, an analyst will trade resolution for increased analysis speed. This loss of resolution can complicate peak identification, even with a mass selective detector (MSD). Agilent Technologies has developed new techniques to solve the peak identification problem based on Agilent's retention time locking (RTL) and a new mass spectral library that contains the locked retention times and characteristic ions

Experimental
The GC method translation software tool was used to find operating conditions for the faster methods. Figure 1 is a screen capture of MTL software data entry showing the original conditions and the new chromatographic conditions for a fourfold speed gain. The column flow rate, which is helpful to avoid exceeding MSD pumping capacity 4, also is found in the table. In this study, a turbo pump was used, which could handle the 3.8 mL/min carrier flow. The program also determined the required column head pressure and corresponding oven ramp. The Agilent 6890 GC fast oven option (220/240V in the U.S.) was required for the faster oven ramp used in this study. General chromatographic conditions are listed in table 1. The standard used was a mixture of 26 pesticides at 10 ppm. A 15 m 0.25 mm 0.25 m Agilent HP-5MS column (part number 19091S-431) was used. The head pressure determined by the method translation software (18 psi) was used as the starting point for retention time locking. The column head pressure required to lock retention times of the compounds to the library (the original retention time divided by 4) was determined using the automated RTL process integrated within the Agilent ChemStation for MSD.

Figure 1. Screen capture showing the method translation (MTL) software data entry used in a 4X speed gain translation.

This process (first translate the method then lock the retention times) was repeated for the 2.5X time reductions. Figure 2 shows the results of the shortened analysis times. The three chromatograms look extremely similar, except that the time axis is scaled proportionally. Because MTL followed by RTL scales methods very precisely, scaled screening libraries for corresponding time reductions can be obtained by dividing the retention times in the library by the speed gain (which does not have to be an integer). Using the same injection method (1-L splitless), the peak heights of the faster runs were twice those from the original

Table 1 Chromatographic Conditions

Speed GC Column Injection mode Column head pressure Column flow (mL/min) Inlet control mode Carrier gas Injector Temp. Oven Temp. Ramp 1 Ramp 2 Ramp 3 Oven equilibration Injection volume Liner MS Conditions (Turbo pump) Solvent delay Tune file Low mass High mass Threshold Sampling Scans/sec Quad Temp. Source Temp. Transfer line Temp. Acquisition mode

Onefold 110 V 30 m 0.25 mm 0.25 mm HP-5MS (P/N 19091S-433) Splitless 18.0 psi 1.9 Constant pressure Helium 250 C 70 (2 min) 25 C/min 150 (0 min) 3 C/min 200 (0 min) 8 C/min 280 (10 min) 2 min 1 mL 5183-4647

Two and a half fold 220/240 V 15 m 0.25 mm 0.25 mm HP-5MS (P/N 19091S-431) Splitless 5.74 psi 1.49 Constant pressure Helium 250 C 70 (0.8 min) 62.5 150 (0 min) 7.5 200 (0 min) 20 280 (4 min) 2 min 1 mL 5183-4647

Fourfold

18.0 psi 3.8

70 (0.5 min) 100 150 (0 min) 12 200 (0 min) 32 280 (2.5 min)

3 min Atune.u 35 amu 500 amu 150 2 3.15 150 C 230 C 280 C Scan (EI)

1.44 min Atune.u 35 amu 450 amu 250 2 3.50 150 C 230 C 280 C Scan (EI)

0.9 min

1 6.54

analysis. A faster oven ramp and the shorter column made the peaks narrower and higher, so an improvement in the signal-to-noise ratio is realized with the faster methods.

Abundance 7000000 6000000 5000000 4000000

TIC: RTLDEMO.D

1X

Conclusion
The highly accurate and reproducible pressure and temperature control of the Agilent 6890 GC allows precise scaling of the standard 42-min GC/MSD pesticide method. Run time was shortened to 10.5 minutes using a fast oven ramp rate and a 15-meter, 250-micron column. The combination of MTL and RTL facilitated scaling and yielded exact scaling. RTL libraries can be scaled accurately to correspond to the faster analyses.

3000000 2000000 1000000 0 Time 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00

Agilent HP-5MS, 30 m 0.25 mm 0.25 m


Abundance 18000000 16000000

TIC: 15MMIX1A.D

References 1. B. D. Quimby, L.M. Blumberg, M. S. Klee, and P. L. Wylie, "Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking," Application Note 228-401, Agilent publication number 5967-5820E, May 1998. 2. H. Prest, P. L. Wylie, K. Weiner, and D. Agnew, "Efficient Screening for Pesticides and Endocrine Disrupters Using the HP 6890/ 5973 GC/MSD System," Agilent publication number 5968-4884E, April 1999.
3. C. K. Meng and M. Szelewski, "Fast Screening of Pesticide and Endocrine Disrupters Using the Agilent 6890/5973N GC/MSD System", Agilent publication number 5968-9220, January 2000. 4. H. Prest, "GC Column Selection and Pumping Considerations for Electron and Chemical Ionization MSD operation," Agilent publication number 5968-7958E, November 1999.

14000000 12000000 10000000 8000000 6000000 4000000 2000000 Time 0 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 16.00

2.5X

Abundance 18000000 TIC: 4X-MIX1A.D 16000000 14000000 12000000 10000000 8000000 6000000 4000000 2000000 0 Time 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00

4X

Agilent HP-5MS, 15 m 0.25 mm 0.25 m


Figure 2. The TICs of the 2.5X and 4X speedups. The standard analysis (1X) was 42 minutes long.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies, Inc. Printed in the USA 5/00 5980-1057E

Trace Level Pesticide Analysis by GC/MS Using Large-Volume Injection

Application
Gas Chromatography September 1997

Author
Philip L. Wylie Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

were injected into a single PTV liner. This application note includes recommendations for doing LVI using the PTV/6890/5973 GC/MSD system.

Introduction
More than 700 pesticides are registered for use in the world1 , and many more continue to persist in the environment, even though they are no longer being applied. For the protection of human health and the environment, pesticide residues are routinely monitored in food, water, soil, and tissue samples. "Acceptable" residue limits have been set for various foods and environmental samples by agencies such as the United States Environmental Protection Agency (U.S. EPA), the Codex Alimentarius Commission2 , and many other governmental organizations around the world. A great many methods have been developed to screen for pesticides in food3-7 and the environment8-10 to ensure that risks associated with pesticide use are minimized.

Abstract
Large-volume injection (LVI) using the Agilent programmable temperature vaporizing (PTV) inlet can improve gas chromatography system detection limits by one to two orders of magnitude over standard methods that call for 1- or 2- L injections. An Agilent 6890 Series gas chromatograph (GC), configured with a PTV inlet, a 6890 Series automatic liquid sampler (ALS), and an Agilent 5973 mass selective detector (MSD), was used for the analysis of pesticides in standards and several food extracts. By making 100- L injections, several pesticides could be identified by scanning gas chromatography/mass spectrometry (GC/MS) at the 100 ppt (100 ng/L) level. The PTV inlet tolerated dirty food extracts very well; more than 1,500 L of such samples

Recently, concern has increased that certain pesticides and other synthetic chemicals may be acting as pseudo hormones which disrupt the normal function of the endocrine system in wildlife and humans. Birth defects, behavioral changes, breast cancer, lowered sperm counts, and reduced intelligence are among the many disorders that have been blamed on these "endocrine disrupting" compounds, though much research must be done to verify these assertions. In 1996, Colborn, Domanoski, and Myers11 brought these issues into the public spotlight with the publication of their book Our Stolen Future. Recently, the United States Congress passed legislation calling for increased testing of suspected endocrine disrupters and monitoring their levels in food12 and water13 supplies. Because the endocrine system can be exquisitely sensitive to extremely low hormone concentrations, there is a need to measure concentrations of suspected endocrine disrupters (many of which are pesticides) at very low levels. Initiatives such as the Pesticide Data Program, developed by the United States Department of Agriculture14 , seek to

determine the lowest measurable pesticide levels in various foods to develop a total exposure model. Clearly, there is pressure to push pesticide detection limits to even lower levels than are routinely achieved today. Most residue measurements are made by gas chromatography using a variety of element-selective or mass spectral detectors (GC/MS). Therefore, to achieve lower detection limits, it is necessary to improve the detection limits of these GC methods. In GC, there are primarily four ways to improve method detection limits: 1) increase the concentration of analytes in a sample, usually by reducing the volume of an extract; 2) increase the sensitivity of the detector; 3) increase the selectivity of the detector to reduce chemical background "noise" or 4) increase the volume of sample injected. Because GC/ MS can be highly selective and extremely sensitive, it is often the method of choice for pesticide analysis and/or confirmation. However, for the reasons discussed above, there are occasions when even greater sensitivity is required. This application note describes a method for increasing GC/MS system detection limits by making large-volume injections (LVI) using Agilent's new programmable temperature vaporizing (PTV) inlet. Because this LVI technique is detector-independent, it is applicable to other GC configurations that may be used for pesticide residue analysis.

I, p,p'-DDE, propargite, iprodione, methoxychlor, and fenvalerate (mix of isomers I and II) to individual 20mL vials and diluting with 10.0 mL of acetone. Permethrin was obtained as a mixture of permethrin I and permethrin II comprising 32 percent and 27 percent of the sample, respectively, so 16.95 mg of this mixture was diluted with 10 mL of acetone giving a solution in which the combined permethrins represented 1 mg/mL. A stock mixture was prepared by adding 4 mL of the permethrin and fenvalerate solutions and 1 mL of each of the other stock solutions to a 100-mL volumetric flask and diluting to volume with acetone. The resultant solution contained 40 ng/ L each of the combined permethrin and fen-

valerate isomers and 10 ng/ L each of the other 12. This sample was diluted further with acetone to prepare standards that were analyzed by LVI. All these pesticides were obtained in neat form from Chem Service (West Chester, PA USA).

Extracts
Fruit and vegetable extracts were obtained from the Florida Department of Agriculture and Consumer Services (Tallahassee, FL USA). Commodities were extracted using a version of the Luke procedure15-17 that gave a final sample representing 1.75 g of the commodity per mL of extract.

Table 1.

Instrumentation and Conditions Used for Pesticide Samples


6890 Series GC 6890 Series ALS 5973 Series MSD PTV with CO2 cooling HP Vectra XU 6/200 G1701AA Version A.03.00 running Microsoft Windows 95 30 m x 0.25 mm x 0.25 m Agilent HP-5MS

GC/MS System Gas chromatograph Automatic liquid sampler Mass spectral detector Programmable temperature vaporizing inlet Computer for data acquisition and analysis Software Column Instrumental Conditions GC Parameters Carrier gas Inlet liner Syringe size Injection volume Injection delay Inlet temperature program Vent flow Purge flow to split vent Column head pressure Oven temperature program MSD Parameters Acquisition mode Temperatures

Experimental
Pesticide Standard Solution
Stock solutions of 14 pesticides were prepared at 1 mg/mL by adding 10 mg each of trifluralin, hexachlorobenzene, pentachloronitrobenzene, dichloran, chlorothalonil, chlorpyrifosmethyl, chlorpyrifos, endosulfan

Helium Prototype deactivated borosilicate with fritted glass on interior walls (part no. 5183-2041) 50 L 100 L (Inject 10 L 10 times) 12 sec 40 C (4.2 min), 200 C/min to 320 C (2 min) 400 mL/min Vent pressure 0.0 psi for 4.00 min 50.0 mL/min at 6.50 min 0 psi (4 min) then 17.3 psi (constant pressure) 50 C (6.13 min), 30 C/min to 150 C (2 min), 3 C/min to 205 C (0 min), 10 C/min to 250 C (20 min) Scan (35-550 amu) Transfer line = 280 C, MS quad = 150 C, MS source = 230 C

Instrumentation
Table 1 lists the instrumentation and chromatographic conditions used for LVI and GC/MS analysis of pesticide samples.

manual injections are usually impractical and good precision may be hard to achieve. The 6890 Series ALS is designed to make one or more injections of up to 25 L into the PTV inlet. After the desired number of injections has been made, the inlet is heated and the chromatography begins. Though the system controls allow up to 99 injections, a reasonable upper limit is about 10, making 250 L the typical injection volume limit for this system. For even larger injections, the controlled speed injector19 should be used. For all of the analyses described below, 100 L were injected by making 10 sequential injections of 10 L each.

How the PTV Works in the Solvent Vent Mode


Figure 1 shows a diagram of the PTV inlet. For large-volume injections, three steps are required. These are: 1) injection and solvent elimination; 2) splitless sample transfer to the GC column; and 3) chromatographic separation and, if desired, a simultaneous inlet bake-out step. The steps are described more completely below.

Brief PTV Tutorial


Before focusing on the PTV/GC/ MS analysis of pesticides, it is important to understand how the PTV inlet operates in the solvent vent mode for large-volume injections.

The PTV Inlet


The PTV inlet has the same basic functions as the split/ splitless inlet except that it is temperature programmable from -60 C (using CO2 cooling) or -160 C (using liquid N2 cooling) to 450 C at rates up to 720 C/min. However, the PTV's design has been optimized for its main uses-LVI and cold split/splitless injection. Although hot split and splitless injections may be made with or without a pressure pulse, care must be taken not to exceed the small internal volume of the PTV inlet. In practice, it is best to choose the Agilent split/splitless inlet for hot injections and the PTV inlet for LVI and cold split/ splitless techniques. Most GC pesticide methods call for injecting 1-2 L; splitless injection is used because it is compatible with dirty extracts of food, soil, or water. Pulsed splitless injection allows one to make injections of up to 5 L using standard equipment18. Enormous gains in system sensitivity can be realized by using the PTV inlet in the "solvent vent" mode, which is compatible with injections of 5-1,000 L. These large injections may be made manually or automatically using either a standard 6890 Series ALS in the multiple injection mode or by using a controlled speed injector available from Gerstel19. Because the injection process may take several minutes,

Injection and Solvent Elimination (Step 1)


During injection, the column head pressure is set to 0 psi to eliminate or, in the case of GC/MS, reduce the flow through the column. When mass spectral detection is used, there is still

Septumless Sampling Head Carrier Gas Line Coolant Liner Seal Heating Coil

Split/Splitless Solenoid Proportional Valve

Glass Inlet Liner

Capillary Column
Figure 1. The PTV inlet shown with the septumless head. The inlet is also available with a septum head that may be equipped with a standard septum or a Merlin Microseal. (Figure reproduced with permission of Gerstel GMBH.)

some flow because the column outlet is under vacuum. At the same time, a steady stream of carrier gas passes through the inlet and out through the split vent. This flow is typically between 100 and 500 mL/min. The sample is injected into the cool liner where it remains as a liquid, dispersed over the liner walls or any packing material that may be in the liner. The steady flow of carrier gas through the liner causes the solvent (and any volatile fraction of the sample) to evaporate and be swept with the carrier gas out through the split vent. This is analogous to "blowing down" a sample with a stream of inert gas, except that this takes place inside the PTV inlet. When most of the solvent has evaporated, the next injection is made and the evaporation process repeats, accumulating more sample in the inlet. To recover an analyte completely, its boiling point should be at least 100 C greater than that of the solvent; most pesticides fall into this category.

Normal

The timing of these multiple injections can be important. If the sample is introduced too rapidly, the liner may become flooded and liquid will be forced out through the split vent. Chromatographically, this shows up as reduced area counts for all analytes (see figure 2A). If there is too much time between injections, all of the solvent may evaporate and more of the volatile analyte fraction may be lost too. This results in poor recovery of volatiles but 100 percent recovery of the less volatile compounds (see figure 2B). Set-points such as inlet temperature, vent flow, and injection delay times can affect recovery of volatiles. Note that for 100 percent recovery, an analyte should have a boiling point at least 100 C greater than the solvent. One can adjust the delay between injections by entering the desired value in the ChemStation software. Some experimentation is usually necessary when setting this delay for a new method. It will be dependent upon such factors as the solvent type, injection volume, vent flow, and inlet temperature.

a splitless injection, except that instead of flash vaporization, the sample is transferred as the inlet temperature is programmed up. For the most gentle treatment of labile analytes, slow ramp rates may be used. This allows analytes to be flushed into the column at the minimum temperature needed for volatilization. When sample decomposition is not a problem, the inlet may be heated as fast as 720 C/min.

Chromatographic Separation (Step 3)


During sample transfer, the oven temperature is usually held between 30 C below and 20 C above the solvent's atmospheric boiling point, depending on whether the solvent effect is needed to focus the more volatile fraction of the analytes. Again, some experimentation is necessary to optimize peak shapes. After the sample has been transferred in step 2, the oven temperature is programmed up and chromatography begins. After the inlet has reached its maximum temperature and sufficient time has elapsed to transfer the sample to the column, a purge flow of 30-50 mL/min is restored to the split vent. If desired, one can set a very large split flow for a few minutes and bake out the inlet at a higher temperature to remove nonvolatile impurities. To conserve carrier gas, gas saver should be turned on at the end of this bake-out step.

Splitless Sample Transfer to the GC Column (Step 2)


Inlet Flooding
Once the desired number of injections has been made, the column head pressure is restored and the vent flow is tur ned off. At this point, the inlet temperature is programmed up to a value that is sufficient to transfer all of the desired analytes to the GC column. This step is similar to

A Sample is injected too rapidly

Normal

Volatiles Lost

B Solvent evaporates completely between injections

Figure 2. Chromatograms A and B illustrate the result of poor timing of multiple injections.

Entering PTV Inlet Parameters into the Agilent ChemStation


When preparing the PTV portion of a GC method, one should first decide on the sample size and how many injections are required. In this work, ten 10- L injections were made for a total of 100 L. When entering parameters into the ChemStation screen, the Injector icon is first selected (figure 3) under the "GC edit parameters" menu. Next, the Configure button is pressed to enter the syringe size and enable multiple injections. From the main injector screen, the injection volume (10 L) and number of injections are entered10 . For this work, a 12-second delay was chosen between injections to allow for solvent evaporation. The estimated total injection time is listed on the Inlets screen (figure 4). This is helpful when setting the inlet and oven parameters. First, the vent flow rate (400 mL/min for these analyses) is chosen, which sets the vent pressure to 0 psi until the injection sequence is done and solvent from the last injection has largely evaporated (4.00 min in figure 4). This is done by entering these values in the following fields: Vent Flow 400 mL/min Vent pressure 0.0 psi until 4.00 min Next, the purge flow and elapsed time are set by entering values in the following field: Purge Flow to Split Vent 50.0 mL/min @ 6.50 min Note that as an aid in setting up the method, the "estimated total injection time" is shown just above the previous data entry fields.

In this example, the normal column head pressure was restored and the vent flow was turned off at 4.00 min. This prepares the inlet for the splitless transfer of the sample to the column. The vent flow remained off until it was set to 50 mL/min at 6.5 min. Thus, there is a 2.5-min period for inlet temperature

programming and splitless sample transfer to the column. In this example, the inlet was held at 40 o C for 4.2 min, enough time to make 10 injections, turn off the purge flow, and restore the column head pressure; the PTV was then programmed to 320 o C at 200 o C/min (figure 4).

Figure 3. The injector screen from Agilent GC and GC/MS ChemStation software showing the setpoints available for multiple injections. To configure the sampler for multiple injections, set the syringe size, and choose slow injection, click on the Configure button.

Figure 4. The inlets screen from Agilent GC and GC/MS ChemStation software showing the setpoints available for operation of the PTV inlet in the solvent vent mode.

Although not done for these analyses, the inlet could be baked out by setting the "purge flow to split vent" to a large value (perhaps 500 mL/min) at the end of the splitless time (6.50 min) and at the same time, program the inlet to a higher temperature. After the bake-out period, the inlet temperature is programmed downward and gas saver is turned on. Normally, the GC oven is held at its starting temperature until the splitless injection is complete (6.50 min in this case) at which time oven temperature programming is begun. For this work, the oven temperature program was begun at 6.13 min so that the pesticide retention times would match a retention time data base that was in use. Figure 5 diagrams the PTV and GC oven setpoints used for this work.

PTV Inlet Liner Considerations


The correct liner choice is critical to the success of any pesticide analysis by PTV injection. The liner must be thoroughly deactivated or many labile pesticides may decompose or adsorb in the inlet. In general, any liner containing glass wool will be unsatisfactory for the analysis of labile pesticides, whether or not the glass wool is deactivated. At this time, two PTV liners are suggested for pesticide analysis: Part no. 5183-2037 is a deactivated, open multibaffled liner with no internal packing that may be used for single or multiple injections of 5 L or less. This liner gives very good recovery for pesticides, even extremely difficult ones such as acephate and methamidophos.

Part no. 5183-2041 is a deactivated liner with an internal coating of sintered glass to give it more surface area and is, therefore, suitable for single or multiple 25- L injections. This liner gives better than 70 percent recovery for most pesticides, although tests have shown that acephate and methamidophos cannot be analyzed using this liner, and that recoveries of guthion are often less than 50 percent. A prototype version of this liner was used for all of the work described in this application note.

Multiple injections

12 sec injection delay

PTV purge flow 400 mL/min 4.00 min 0 psi 4.00 min 40 C 4.20 min 50 C 6.13 min

Column head pressure PTV temperature Oven temperature

Figure 5. Illustration of the GC and sampler setpoints used for 100- L injections of pesticide samples. Note that normally, the GC oven hold period would have been at least 6.5 min for this method. A value of 6.13 min pesticide retention times to a data base.

}
n 0 mL/min 280 C 200 C/min 50 mL/min 6.50 min 30 C/min
6

Results and Discussion


When compared to a typical 2-L splitless injection, 100- L PTV injections can often result in a 50-fold improvement in system detection limits. Selective detectors such as the MSD can help the analyst to realize the full measure of this sensitivity improvement by excluding background that may be introduced from solvent impurities, vial cap extract, and indigenous compounds coextracted with the analytes. In this application, it was possible to see most of the pesticides in the 14-component mixture at 100 ppt in the scan mode (400 ppt for the isomer mixes of permethrin and fenvalerate). Figure 6 shows extracted ion chromatograms for trifluralin and hexachlorobenzene (HCB) at 100 ppt. Library searching gave a match quality of 93 for the HCB peak. Fenvalerate isomers I and II were found in the solution in a ratio of about 78:22. Figure 7 shows extracted ion chromatograms for fenvalerate I at a concentration of 311 ppt.

Trifluralin (100 ppt) m/z 306

m/z 264

A Extracted ion current chromatograms of trifluralin

Hexachlorobenzene (100 ppt) Extracted ions 284, 286, and 282 Match quality = 93

B Extracted ion current chromatogram of HCB with its mass spectrum and library match

Figure 6. Scanning GC/MS results for a pesticide standard containing Trifluralin and Hexachlorobenzene at 100 ppt. (Ten 10- L injections were made using the PTV inlet.)

Fenvalerate I (311 ppt) m/z 167

m/z 125

m/z 225

Figure 7. Extracted ion current chromatograms of Fenvalerate I at a concentration of 311 ppt in a pesticide standard. (Ten 10- L injections were made using the PTV inlet.)

Analysis of a bell pepper extract revealed several pesticide residues. As seen in figure 8, chlorpyrifos and the endosulfans were easily detected. The Florida Department of Agriculture determined the concentration of chlorpyrifos, alpha-endosulfan, betaendosulfan, and endosulfansulfate to be 0.210, 0.011, 0.018, and 0.013 ppm, respectively. It is important to note that these compounds could be detected with very high selectivity by extracting high mass ions that are characteristic of these pesticides but not of the matrix. Using LVI, there is ample signal from these less abundant ions for good quantitation. With normal injection volumes, selectivity may have to be compromised and the most abundant ions extracted in a pesticide spectrum to gain sensitivity. Phosmet, captan, and propoxur were all easily detected in a pear sample. The total ion current chromatogram (TIC) is shown in figure 9 along with spectrum obtained for captan juxtaposed with the library spectrum. Figure 10 shows the propoxur peak along with 2,4,6-tribromoanisole and 2,4,6-tribromophenol, two other compounds that were surprising to find in a pear sample. Though the origin of these brominated compounds is not known, a recent paper by Hoffmann and Sponholz 20 suggests that tribromophenol is used to treat storage palettes for the prevention of fire and mold growth, and that the anisole is formed from the phenol microbiologically. Perhaps these pears were shipped in containers that had been similarly treated.

Figure 8. GC/MS Analysis of a bell pepper extract. (Ten 10- L injections were made using the PTV inlet.) Using LVI, there was sufficient signal to use high mass ions with smaller abundances to achieve greater selectivity.

Figure 9. TIC of a pear extract resulting from a 100- L Injection (10 x 10 L). Captan was easily detected, and its spectrum gave a library match quality of 96.

Figure 10. TIC of a pear extract resulting from a 100- L Injection (10 x 10 L). Propoxur and two brominated phenolics were easily identified.

A single sintered glass coated liner of the type described above (part no. 5183-2041) was used for about ten 50- and ten 100- L injections (ca. 1,500 L total) of vegetable extracts before it was replaced. All of the extracts were rather dirty, and an inlet bake-out step was not used. Although the liner looked somewhat discolored for about 2 cm where injections were made, it still performed well at the time it was replaced.

Conclusion
Using the PTV inlet in the solvent vent mode, it is relatively simple to increase system detection limits by one or two orders of magnitude. When combined with the Agilent 6890 Series automatic liquid sampler,

multiple injections of up to 25 L each into the inlet can be made, allowing the solvent to vent while pesticides and other less volatile analytes accumulate. After the desired sample volume has been introduced (typically 5-250 L), the solvent vent is closed and the sample is transferred to the column in a temperature-programmed splitless injection. By making 100- L injections into a PTV-equipped Agilent 6890 Series GC coupled to the Agilent 5973 MSD, it was possible to see several pesticides at the 100 ng/L level (100 ppt) in the scan mode. With such low detection limits, less abundant ions can be used to identify and quantitate pesticides at low ppb levels, thereby gaining in selectivity as well.

When performing LVI, there are several parameters to adjust and some method development time is usually required. However, the method described herein worked well and can be duplicated for the PTV/GC/MS analysis of pesticides in food.

Acknowledgment
The author wishes to thank Ms. Joanne Cook of the Florida Department of Agriculture and Consumer Services for supplying the food extracts used in these experiments and Dr. Bill Wilson (Agilent Technologies) for supplying liner deactivation test results.

References
1. Tomlin, Clive, ed (1994), The Pesticide Manual, Tenth Edition, British Crop Protection Council, Surry, UK. 2. Miller, R. W., This is Codex Alimentarius, Secretariat of the Joint FAO/WHO Food Standards Programme, Food and Agriculture Organization of the United Nations, Rome. 3. McMahon, B. M. and Hardin, N. F. eds. (1994), Pesticide Analytical Manual, Vol I, Third Edition, U.S. Food and Drug Administration, Washington, DC. 4. Lee, S. M., Papathakis, M. L., Feng, H.-M. C., Hunter, G. G., and Carr, J. E. (1991), Fresenius' A. Anal Chem 339, 376-383. 5 Fillion, J., Hindle, R., Lacroix, M., and Selwyn, J. (1995), J AOAC Int 78, 1252-1266.

7. Luke, M. A., Froberg, J. E., Doose, G. M., Masumoto, H. T. (1981), J Assoc Off Anal Chem 64, 1187-1195. 8. Stan, H. J., ed. (1995), Analysis of Pesticides in Ground and Surface Water II, SpringerVerlag, Berlin, Germany. 9. Wagner, R. E., Kotas, W., and Yogis, G. A., eds. (1994), Guide to Environmental Analytical Methods, 2nd Edition Genium, Schenectady, NY. 10. U.S. Environmental Protection Agency, Test Methods for Evaluating Solid Waste, SW-846, Draft Method 8085: Pesticides by GC/AED. 11. Colborn, T., Dumanoski, D., and Myers, J. P. (1996), Our Stolen Future, Penguin, New York, NY. 12. Food Quality Protection Act of 1996, Public Law 104-170, Congressional Record pp. H8127-H8141. 13. Safe Drinking Water Act Amendments of 1996, Public Law 104-182, Congressional Record pp. H9680-H9700.

14. Pesticide Data Program Annual Summary Calendar Year 1994, U.S. Department of Agriculture, Agricultural Marketing Service, Washington, DC. 15. Luke, M. A., Froberg, J. E., Doose, G. M., and Masumoto, H. T. (1981), J Assoc Off Anal Chem 64, 1187-1195. 16. Luke, M. A., and Doose, G. M. (1983), Bull Environ Contamin Toxicol 30, 110-116. 17. Sawyer, L. D. (1985), J Assoc Off Anal Chem 68, 64-71. 18. Wylie, P. L., Phillips, R. J., Klein, K. J., Thompson, M. Q., and Hermann, B. W. (1991), J High Resol Chromatog 14, 649-655. 19. The controlled speed injector is available from Gerstel US, 1510 Caton Center Dr., Baltimore, MD 21227 USA. 20. Hoffmann, A. and Sponholz, W. (March 1997), American Laboratory, 22-23.

6. Working Group on Development and Improvement of Residueanalytical Methods (1996), Analytical Methods for Pesticide Residues in Food-stuffs, General Inspectorate for Health Protection, Ministry of Public Health, Welfare & Sport, The Netherlands.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Microsoft is a U.S. registered trademark and Windows is a U.S. trademark of Microsoft Corporation. HP is a registered trademark of Hewlett-Packard Company. Copyright 2000 Agilent Technologies, Inc. Printed in the USA 4/2000 5966-1214E

Fast Dual-Column GC/ECD Analysis of Chlorinated PesticidesEPA Methods 608 and 8080
Application Note 228-305

Author
Imogene L. Chang, Ph.D. Agilent Technologies Wilmington, DE 19808-1610

Experimental
EPA Method 608 and 8080 targeted pesticides were separated using 30 m x 0.53 mm x 1.0 m HP-35 and HP PAS-1701 columns (part no. 19095G123 and 19094U-023, respectively). Analyses were performed on an HP 5890 Series II GC with EPC, dual split/splitless inlets, and dual ECDs. An Agilent 7673 automatic liquid sampler was used to process the simultaneous splitless injections. A deactivated single-tapered glass liner with a small plug of glass wool (part no. 5181-3316) and a Merlin Table 1. Experimental Conditions
Instrument Requirement Gas Chromatograph Injection Ports Column Detector Sample Introduction Data Collection Experimental Conditions Injection Carrier gas

Microseal septum (part no. 51818816) were used with each split/ splitless inlet. Instrumentation and GC conditions are listed in Table 1. A test mix containing 18 pesticides (50 ppb per component) and two surrogates was prepared from the dilution of certified standard mixes with pesticide-grade hexane (Burdick & Jackson). Pesticides in the test mix are listed in Table 2.

Abstract
Dual-column analysis with HP-35 and HP PAS-1701 columns was used to analyze chlorinated pesticides targeted in EPA Methods 608 and 8080 for wastewater and solid wastes. GC parameters were optimized using the Agilent 5890 Series II gas chromatograph (GC) with electronic pressure control (EPC), a dual injector, and a dual electron capture detector (ECD) system. The analysis of 18 pesticides was completed in 12 minutes.

Agilent Technologies 5890 Series II with EPC Dual split/splitless inlets HP-35, 30 m x 0.53 mm x 1.0 m (Part no. 19095G-123) HP PAS-1701, 30 m x 0.53 mm x 1.0 m (Part no. 19095S-123) Dual ECD 7673 automatic sampler with dual injectors 3365 ChemStation and HP Vectra 486/33T PC

Introduction
Currently, many testing laboratories use dual-column/dual-ECD GC systems to analyze the chlorinated pesticides specified in EPA Methods 608 and 80801,2. For this application, EPC was used with an HP-35 column (35% phenyl, 65% methyl polysiloxane phase) as the primary column and the HP PAS-1701 column for confirmation. The unique selectivity of the HP-35 column for this set of chlorinated pesticides permitted focus on the optimization of oven temperature for the HP PAS-1701 column. Individual EPC ports for each injector permitted individual regulation of column flow for both the HP-35 and the HP PAS-1701.

Splitless 1 l, purge delay, 0.75 min, inlet temperature of 250C (A) HP-35, pressure program: 8.6 psi (1 min) at 0.5 psi/min to 12 psi and at 3.0 psi/min to 25 psi (0 min) (B) HP-1701, helium, 10 ml/min constant flow 160C (1 min) to 280C at 10C/min and to 300C (2 min) at 25C/min ECD (300C), 120 ml/min N2 makeup, 6 ml/min anode purge

Oven Detector

Results and Discussion


In a dual-column/dual-ECD system, samples introduced in a single injection can be split between two columns using a Y-connector and detected by different ECDs. However, when using a Y-connector without EPC, the split sample flow to each column cannot be optimized, and equal and consistent sample splits cannot be presumed. The only variable that can be optimized, in dual-column ECD analysis using a Y-connector is the oven temperature program, which can be optimally balanced for the two dissimilar columns. Using dual-column GC/ECD without EPC, it would typically require 45 to 60 minutes to obtain baseline separations for EPA Method 608 and 8080 targeted pesticides (see Figure 1). A typical run from an environmental testing laboratory for a test mix containing 18 targeted pesticides and two surrogates is shown in Figure 1. A

Table 2. Chlorinated Pesticides.


Peak No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Pesticides Tatrachloro-m-xylene (SS1) alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Decachlorobiphenyl (SS2)

Yconnector was used to split samples for both columns, DB-608 and DB-1701, and good baseline separations were obtained for most analytes. This dual-column run was completed in 45 to 53 minutes using the following oven temperature program: 150C (1 minute) to 260C (18.34 minute) at 3C/minute, then to 275C (5 minutes) at 25C/minute. Clearly this oven temperature program was optimized to separate critical pairs, such as DDE/dieldrin, DDD/endosulfan II, endosulfan sulfate/mehtoxychlor, and methosychlor/endrin ketone for both columns. Figure 2 shows chromatograms of the same pesticide test mix using the HP-35 and HP PAS-1701 columns and EPC. The oven program, 160C (1 minute) to 280C at 10C/minute and to 300C (2 minutes) at 25C/minute, was optimized to separate the critical pairs, endosulfan

DB-608
1 2 3 4 5 6 7 18 8 9 10 11 12 13 15 16 17 19 20

14

12

16

20

24

28

32

36

40

44

48

52

56 min

DB-1701
1 2 3 5 7 6 8 9 10 11 12 14 13 32 15 16 17 18 20 19 4

12

16

20

24

28

36

40

44

48

52

56 min

Figure 1. Typical chromatograms of a pesticides standard mix using DB-608 and DB-1701 columns under GC conditions used in environmental testing laboratories. (See Table 2 for peak identification.)

HP-35 30 m x 0.53 mm x 1m

1 3 20 2 6 4 5 7 8 9 10 11 14 16 17 13 12 15 19 18

00 2 1 4 6 8 10 20 2 3 5 4 7 6 8 10 11 9 14 13 16 18 17 15 0 0 2 4 6 8 10 12 min 12 min

HP PAS-1701 30 m x 0.53 mm x 1m

12

19

Figure 2. Chromatograms of a pesticides standard mix using HP-35 and HP PAS-1701 columns under the GC conditions listed in Table 1. (See Table 2 for peak identification.) II/DDT and methoxychlor/endosulfan sulfate, for the HP PAS-1701 column. In this run, EPC provided a constant 10 ml/minute helium flow to the HP PAS-1701 column throughout the entire run. For the HP-35 column, the following pressure program was used: 8.6 psi (hold 1 minute) at 0.5 psi/minute to 12 psi and at 3.0 psi/minute to 25 psi (hold for constant flow for the remaineder of the run). This pressure program actually provided a 10 ml/minute constant flow to elute most of the pesticides and an increased flow (up to 20 ml/minute) near the end of the run to elute the last analyte, surrogate decachlorobiphenyl and other high-boiling materials from the column. GC parameters optimized for dualcolumn/dual-injector/dual-ECD analysis of chlorinated pesticides reduced analysis time to less than
Copyright , 1995, 2000 Agilent Technologies Printed in USA 04/00 Printed on recycled paper Publication Number 5963-6734E

12 minutes. In addition to speed, all EPA Methods 608 and 8080 targeted pesticides and surrogates were well resolved with good sharp peaks for accurate quantitation.

Acknowledgement
The author wishes to thank Ms. Joann Faulkner and her colleagues at the Pace Laboratory in Petaluma, California, for providing chromatograms and pesticide standards.

Conclusion
The use of EPC permitted individual column flow control to each ECD. The unique selectivity of the HP-35 column for chlorinated pesticides permitted focus on the optimization of oven temperature for the HP PAS1701 column. Run time was 11.5 minutes with good baseline separations for all 20 target pesticides and surrogates. The result was a reduction in sample turnaround time from 54 to 11.5 minutes for a 400% increase in productivity. This is more than a twofold improvement in productivity when compared with conventional methods currently used at many environmental testing laboratories with DB-608 and DB-1701 columns.

References
1. USEPA SW-846 Test methods for Evaluating Solid Wastes, Methods 8000 and 8080, September 1986. 2. USEPA Methods for Organic Chemical Analysis of Municipal and Industrial Wastewater, Method 608, 1982. 3. I. L. Chang, The Analysis of Chlorinated Pesticides and PCBs Using the HP-608 Capillary Column, Agilent Application Note 228-236, Publication No. 5091-7567E.

Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System Application Note
Environmental
Jeff Keever, Research Triangle Park Ken Imatani and Paul Goodley, Agilent Technologies

Introduction
Pesticides in food and beverages can be a significant route to human exposure. Extracts of composite food samples typically contain many compounds, which produce interfering compounds at the retention time of interest. This note describes the application of ion trap LC/MS/MS to determine pesticide and herbicide contamination in food samples. The goal of this method was to develop a confirmaThe specific determination of the carbamate pesticide carbaryl in composite food extracts that represent typical meals have previously been investigated using a fluorescence technique.13 tion assay to detect carbaryl in foods at low ng/g levels. Absolute confirmation, identification and quantification of the low levels of carbaryl can be acheived with the LC/MSD Trap primarily due to the high sensitivity and specificity of MS/MS with the ion trap. However, the fluorescence method has drawbacks related to: (1) time consuming post-column derivatization, (2) false positives due to lack of specificity, (3) long sample preparation, (4) insufficient limits of detection, and (5) the lack of confirmatory information from the analytical results.

Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System

Agilent Technologies

Experimental
Pesticides in food and beverages are considered a significant route for human exposure. To better assess exposure from dietary intake, whole food samples were collected over a 24-hour period, homogenized using a 5-gallon blender, and then analyzed as a composite sample. Aliquots of the homogenate were extracted using a nine step scheme: (1) acid precipitation using zinc acetate, (2) drying with potassium oxalate and (3) sodium sulfate, (4) soxhlet extraction, (5) solvent exchange, (6) liquid-liquid extraction followed by (7) GPC cleanup with a (8) final solvent exchange and (9) concentration into acetonitrile. Table 1 lists the HPLC method used for the determination of carbaryl in food extracts.
Table 1. LC Conditions. LC column: Mobile phase: 4.6 X 250 mm Zorbax XDB-C8 A = 95% H20: 5% ACN, 0.1M NH4Ac, 0.1% CH3COOH B = 40% H20: 60% ACN, 0.1 NH4Ac, 0.1% CH3COOH

the detection of non-target compounds in the food extract. The selectivity offered by MS/MS enabled the detection of the carbaryl product ion at m/z 145 (Figure 2) generated from the [M+H]+ ion of carbaryl. The subsequent quantitative analysis was performed using this product ion at m/z 145. The limit of detection in this food matrix was found to be 1 ng/g and the limit of quantitation (LOQ) used for the determination in food was 10 ng/g, exhibiting a signal-to-noise of 40:1 in the composite food matrix ( Figure 3). A calibration based on the m/z 145 product ion was linear (linear correlation coefficient of 0.994) over the concentration range of 11000 ng/g (Figure 4). When the Limit of Quantitation of the carbaryl for both techniques were compared, the LC/MS/MS was shown to be 12 times more sensitive than the HPLC fluoresence technique. The LOQ for LC/MS/MS was found to be 110 ng/g while the LOQ for the fluoresence was 120 ng/g.

1.50

Gradient program: 20% B for 10 minutes; ramp up to 100% B over 30 minutes; hold for a further 11 minutes at 100% B Flowrate: 1 mL/min

1.25

Intensity 10 8

1.00

LC/MS/MS was performed using an Agilent LC/MSD Trap mass spectrometer in the MS/MS full scan mode. Carbaryl was detected using positive ion electrospray, in which MS/MS was used to provide additional fragmentation information needed to confirm the presence of carbaryl.

0.75

0.50

Results and Discussion


Extracts of composite food samples typically contain many compounds (Figure 1) which produce interfering compounds at the retention time of interest. This LC/MS/MS positive ion electrospray method for carbaryl proved 2030 times more sensitive compared to data acquired from a scanning instrument, such as a single quadrupole, enabling

0.25

0.00 0 10 20 30 40

Time
Figure 1. HPLC/ion trap MS full scan determination of carbaryl (retention time of carbaryl shown by arrow) in a composite food sample.

Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System

Agilent Technologies

145 m/z
100

145

+H O O C N H CH3

145 m/z
8000

S/N = 40:1

80
Carbaryl

Abundance

6000 60

Intensity
225 250 275

+ O

4000

40

20

2000

[M + H] + 202
0 100 125 150 175 200 0 0 10 20 30 40

m/z
Figure 2. HPLC/MS/MS spectrum of carbaryl from Figure 1. The product ion spectrum was generated by mass selecting the [M+H]+ ion at m/z 202 for collision induced decomposition using an end cap fragmentation voltage of 0.9V for 20 ms.

Time
Figure 3. Ion trace of the carbaryl product ion at m/z 145 at the limit of quantitation LOQ (10 ng/g) in food.

r 2 = .994 Response 107


0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 5 6 7 8 9 10

The LC/MS/MS ion trap method confirmed that there were false positives in samples having a positive carbaryl identification by HPLC fluorescence (5 g/g of carbaryl). The coeluting interference that also underwent post column derivatization, was a different molecular weight (m/z 213 vs 201 for carbaryl) and showed a different product ion spectrum (Figure 5), compared to carbaryl (Figure 2). Based on the MS/MS spectrum of the interference, a possible structure could be assigned as shown in Figure 5.

Concentration (ng/g)[100]
Figure 4. Calibration curve based on the m/z 145 product ion for carbaryl over the concentration range of 11000 ng/g.

Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System

Agilent Technologies

References
O C2 H4 N O
+

214
NH C2 H5

100
N H 86 H 145 O mw 213 C O

80

1. Forest, D. L., U. S. Environmental Protection Agency, Method 531.1. Measurement of N-Methylcarbamoyloximes and N-Methylcarbamates in Water by Direct Aqueous Injection HPLC with Post Column Derivatization, Revision 1.0 (1985). 2. Engels, T., National Pesticide Survey Method 5, Revision 2.0 (1987).

Abundance

60

+ N H H

3. Graves, R. L., Method 531.1, Revision 3.0 (1989).


O N H C2 H4 N O+

86
40

Acknowledgements
This work was supported by EPA contract No. 68-C5-0011.

20

145

Author
0 60 80 100 120 140 160 180 200 220

m/z
Figure 5. HPLC/MS/MS spectrum of the coeluting interference that produced a false positives in the HPLC fluorescence method.

Dr. Jeff Keever is a research associate at Research Triangle Institute, Research Triangle Park, NC. Ken Imatani is a product manager and Paul Goodley is a senior applications chemist at Agilent Technologies, Palo Alto, CA.

Conclusions
This application has shown that the analysis of food samples for pesticides/herbicides using ion trap MS/MS detection can be accomplished successfully. This method allows for the detection, confirmation and quantitation of the carbamate pesticide carbaryl in extracts of whole food samples down to 1 ng/g levels.

For more information on our products and services, you can visit our site on the World Wide Web at: http://www.agilent.com/chem. Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies All rights reserved. Reproduction and adaptation is prohibited. Printed in the USA April 2000 (23) 5980-0332E

A Method Used to Screen for 567 Pesticides and Suspected Endocrine Disrupters

Application
Gas Chromatography May 1998

Authors
Philip L. Wylie and Bruce D. Quimby Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Abstract
A gas chromatographic (GC) method has been developed that can be used to screen for 567 pesticides and suspected endocrine disrupters. In principle, it can be used to screen for any GC-amenable pesticide, metabolite, or endocrine disrupter. The method relies on a technique called retention time locking (RTL). RTL is a procedure that allows the chromatographer to reproduce analyte retention times independent of GC system, column length, or detector so long as columns with the same stationary phase, nominal phase ratio, and diameter are used. Because RTL increases retention time precision and predictability, raw retention times can be used as a more reliable indicator of compound identity. The chromatographer first locks the GC method so that all retention times match those listed in a 567-compound pesticide and

endocrine disrupter retention time table. After analyzing a sample by GC with atomic emission detection (GC-AED), the analyst enters a peaks retention time and known elemental content (presence or absence of heteroatoms) into a dialog box. If elementselective detectors are used, detector response can be entered in addition to or in place of GC-AED data. The software then searches the pesticide table for those compounds that elute at the correct retention time and have the right elemental content or detector response. Most often, the software finds just one compound that meets these criteria, and rarely does it find more than three. Confirmation is performed by GC with mass spectral detection (GC-MS) or by calculation of elemental ratios using GC-AED data. With retention time locking, pesticides have the same retention time on all GC systems; this makes GC-MS confirmation much easier because the analytes retention time is already known.

Introduction
The Pesticide Manual1 lists 759 compounds and biological agents that are used currently as active ingredients in various pesticide formulations. Many compounds, though no longer used, still persist in the environment. For the protection of human health and the environment, acceptable limits in food and water have been set by governmental bureaus such as the United States Environmental Protection Agency (USEPA) and the Codex Alimentarius Commission.2 Numerous methods have been developed to screen for pesticide contamination in food37 and the environment810 to ensure that these standards are met. Certain pesticides and other synthetic chemicals have been suspected of behaving as pseudo hormones, disrupting normal functions of the endocrine system in wildlife and humans. Maladies such as birth defects, behavioral changes, breast cancer, lowered sperm counts, and reduced intelligence have been blamed on exposure to endocrine disrupters.11 The 1996 publication of Our Stolen Future, a book by Colborn,

Key Words
Pesticides, endocrine disrupters, gas chromatography, retention time locking, RTL

Dumanoski, and Myers,11 brought these concerns to the attention of the public. Recently passed legislation in the U.S. calls for more testing of suspected endocrine disrupters and monitoring of them in food12 and water13 supplies. To facilitate more research into the endocrine disrupter issue, methods are needed to detect suspected compounds at trace levels. Because so many pesticides are in use, it is usually impractical to screen for large numbers of them individually and, therefore, multiresidue methods are preferred. Most laboratories that analyze for pesticides in food or the environment screen for only a few dozen compounds because it is often very difficult to screen for more. Recently however, methods have been developed using gas chromatography with mass spectral detection (GC-MS), that can screen for more than 2005 or even 3006 pesticide residues. Still, there is no universal method to analyze for all GC-amenable pesticides. While GC-MS methods are gaining in popularity, there are still some limitations. When methods employ selected ion monitoring (SIM) or tandem mass spectrometry (MS-MS), method development is more tedious and any shift in GC retention times requires that individual analyte retention time windows be shifted accordingly. These methods are only capable of detecting compounds on the target list; there are still hundreds of pesticides, metabolites, and suspected endocrine disrupters that could be missed. On the other hand, methods based on scanning GC-MS alone may require more sample cleanup to avoid interferences from co-extracted indigenous compounds. Typically, these methods do not screen for many pesticide metabo-

lites, endocrine disrupters, or other environmental contaminants. A method that could be used to screen for endocrine disrupters and almost all of the volatile pesticides and metabolites would offer a better means of monitoring the food supply and the environment. This paper describes a universal method that, in principle, could be used to screen for any pesticide, metabolite, or endocrine disrupter that can elute from a gas chromatograph. The screening procedure relies on a new gas chromatographic technique called retention time locking (RTL)1416 with database searching based on retention time and elemental content or detector response. This technique is used to narrow an analytes identity to a few possibilities. Confirmation is performed by GC-MS or by calculation of a compounds elemental ratio using GC with atomic emission detection (GC-AED).

intended for analysis by halogenselective detectors were also subjected to floracil SPE.

Pesticide Retention Time Table


The table containing GC and GC-MS retention times for 567 pesticides, metabolites, and suspected endocrine disrupters was obtained from Agilent Technologies, Wilmington, DE, USA (G2081AA).

Instrumentation
Table 1 lists the instrumentation and chromatographic conditions used for GC-AED screening and GC-MS confirmation.

Software for Method Translation


Software for use in translating the normal GC method to one that runs three times faster was obtained from Agilent Technologies,Wilmington, DE, USA.17

Experimental
Standards and Extracts
Pesticide standards used to develop the retention time table were obtained from Chem Service (West Chester, PA, USA), Promochem Ltd (Welwyn Garden City, Hertfordshire, England), Dr. Ehrenstorfer (Augsburg, Germany), Hayashi Pure Chemical Industries, Ltd (Osaka, Japan), Wako Pure Chemical Industries, Ltd (Osaka, Japan), and GL Sciences Inc (Tokyo, Japan). Fruit and vegetable extracts were obtained from the Florida Department of Agriculture and Consumer Services (Tallahassee, FL, USA). Samples were extracted with acetonitrile followed by solid-phase extraction (SPE) using a C-18 cartridge. Extracts

Results and Discussion


Retention Time Locking
Key to the development of this method is a new concept in gas chromatography called retention time locking (RTL).1416 Agilent RTL software allows the chromatographer to match analyte retention times from run to run, independent of the GC system, detector, or manufacturing variations in column dimensions. The only requirement is that the columns used have the same stationary phase and the same nominal diameter and phase ratio. For example, with RTL it is possible to match analyte retention times on a GC-AED and a GC-MS even though the MS operates under vacuum and the AED operates at 1.5 psi above ambient pressure. The

procedure also compensates for differences in GC column length resulting from variations in manufacturing or from column cutting required during routine maintenance. RTL is accomplished by adjusting the GC column head pressure until a given analyte, such as an internal standard, has the desired retention time. When this is done, all other analytes in the chromatogram will have the correct retention times as well. Software has been developed that can be used to determine the column head pressure that will lock the retention times correctly after one or two scouting runs. With RTL, it is possible to measure pesticide retention times using a given GC method, and then reproduce those retention times in subsequent runs on the same or different instruments. With this increased retention time precision and predictability, retention times become a far more useful indicator of analyte identity. For many years, relative retention times3,6 or retention indices18,19 have been used to identify compounds. These techniques were developed to compensate for the fact that retention times were not predictable from day to day, column to column, or instrument to instrument. With the increased retention time precision of the Agilent 6890 GC and RTL, it seemed that raw retention times could be used for compound identification instead of retention indices. The chromatographer could simply scan a table of pesticide retention times, eliminating all possibilities but those with close elution times under the same locked GC conditions.

Table 1.

Instrumentation and Conditions of Analysis


6890 6890 Series automatic sampler G2350A atomic emission detector G2360AA GC-AED software running on Microsoft Windows 3.11 30 m 0.25 mm 0.25 m HP-5MS (part no. 19091S-433) Split/splitless, 250 C or 260 C Single-tapered deactivated (part no. 5181-3316) with 2-cm deactivated glass wool plug centered ~3 cm from the top 35 L splitless when running method at 3 speed; 23 L splitless at 1 speed 87.5 psi constant pressure for method at 3 speed; 27.6 psi constant pressure for 1 speed 60 psi (2.01 min), 10 psi/min to 27.9 psi (hold) 70 C (2 min), 25 C/min to 150 C (0 min), 3 C/min to 200 C (0 min), 8 C/min to 280 C (10 min) 290 C 320 C Group 1: Cl 479, Br 478 Group 2: C 193, S 181, N 174 Group 3: P 178 Group 4: F 690 (optional) 6890 6890 Series automatic sampler 5973 MSD G1701AA Version A.03.00 running on Microsoft Windows 95 30 m 0.25 mm 0.25 m HP-5MS (part no. 19091S-433) Split/splitless, 250 C Single-tapered deactivated with small amount of glass wool at the bottom (part no. 5062-3587) 2 L 15.5 psi (constant pressure) Same as GC-AED Scan (35550 amu) 200 rel 3.20 min 150 2.86 Transfer line = 280 C, MS quad = 150 C, MS source = 230 C

Agilent GC-AED System Gas chromatograph Automatic sampler Atomic emission detector Software Column GC inlet Inlet liner Injection volumes Inlet pressure (splitless)* Inlet pressure program (pulsed splitless)* Oven temperature program AED transfer line temperature AED cavity temperature AED elements and wavelengths (nm)

Computer for data acquisition and analysis HP Vectra XM Series 4 5/150

Agilent GC-MS System Gas chromatograph Automatic sampler Mass selective detector Software Column Inlet Inlet liner Injection volume Inlet pressure* Oven temperature program MSD parameters Acquisition mode EM voltage Solvent delay Threshold Scans/sec Temperatures

Computer for data acquisition and analysis HP Vectra XU 6/200

*The column head pressures shown are typical values. Exact values were determined as part of the retention time locking procedure.

Pesticides almost always contain heteroatoms and often have several in a single molecule. The most frequently encountered heteroatoms are O, P, S, N, Cl, Br, and F. GC with atomic emission detection (GC-AED) has been shown to be a useful tool for pesticide screening because it is selective for all of the elements found in these compounds.2022 Thus, GC-AED screening provides valuable information about the elemental content of an unknown molecule. By including this elemental information along with the retention time, it should be possible to narrow pesticide hits to just a few possibilities. To implement this screening procedure, a table of pesticide and endocrine disrupters retention times had to be created using a suitable method under locked conditions.

tested to see if it could still be locked with a retention gap installed. The column chosen for the method was a 30 m 0.25 mm 0.25 m HP-5MS because the same column could be used with any GC-detector combination. In particular, this column was chosen for its low bleed at high temperatures and because its optimum column flow is compatible with GC-MS. The 5% phenyl methyl silicone phase in this column has been widely used for pesticides. Method translation software17,25,26 can be used to increase the speed of a method while retaining the same relative retention times. This can be done by translating the method to a column having the same phase ratio but a smaller id or by increasing the flow rate and oven temperature program while using the same column. The final goal was to design a method that could run at three times the normal speed on the 30-m 0.25-mm 0.25- m HP-5MS column or be translated to a 100- m id column. After several weeks of method development, the GC oven temperature program shown in figure 1a was chosen because it met all of the development criteria. Chlorpyrifos-methyl (C7H7Cl3NO3PS) was chosen as the locking standard. It is an ideal choice because chlorpyrifos-methyl elutes near the middle of the chromatogram (16.596 minutes), has good peak shape, and can be seen by most element-selective detectors. Because GC-AED requires three runs to generate element-selective chromatograms for C, Br, Cl, N, S, and P, the method was translated to run three times faster using software for method translation.17,25,26 The faster oven temperature program used by this method requires 6890 GC systems that are configured for fast oven tem-

perature ramping. The method translation software can be used to speed up the method by any desired factor; even 120-V 6890 GCs can run the method two times faster. However, the original method must be used for GC-MS because of the restriction in flow rates into the MSD. Figure 1b lists the threefold (3  faster GC method.

Pesticide Retention Time Table


Once developed, this method was employed to create a table of locked retention times for the 567 pesticides, metabolites, and suspected endocrine disrupters. Increasing international food trade requires the analysis of pesticides that may be used in the supplying country but not in the recipient country. The goal was to create a table that included pesticides used around the world so pesticide standards were obtained from sources in Europe, Japan, and the USA. A list of suspected endocrine disrupters was compiled from various lists published on the World Wide Web.2731 Many of these compounds are, in fact, pesticides. Most of the GC-amenable endocrine disrupters were analyzed and their retention times appear in the table. However, the 209 polychlorinated biphenyl congeners were not included because their inclusion might actually complicate the identification of organochlorine pesticides. Standards, diluted to 10 ppm in acetone, were first analyzed by GC-MS using the oven temperature program shown in figure 1a and instrumental conditions listed in table 1. Compound identities were verified by matching their spectra to library entries,32 by comparison with a published spectral compendium,33 or by matching spectra to a list of charac-

GC Method for Pesticide Screening


First, a GC method was needed that could elute hundreds of pesticides and endocrine disrupters in a reasonable time with adequate separation. However, the goal was not to separate every possible analyte in a single GC run. Because the intention was to build a table of locked retention times using this method, it had to reproduce these retention times under a variety of conditions. For example, the method needed to accommodate a variety of injection techniques including splitless, pulsed splitless,23,24 cold splitless using a PTV inlet, and oncolumn injection which is occasionally used for the more labile pesticides. The method also needed to perform well with samples dissolved in common solvents such as acetone and methylene chloride. Because a retention gap (or guard column) is sometimes added to protect the analytical column, the method had to be

teristic ions.6 When reference spectral information was not available, the pesticides were verified by spectral interpretation. Samples were then analyzed on two different 6890 GC-FID instruments under the same locked conditions (chlorpyrifosmethyl retention time = 16.596 minutes). The GC-MS retention time and the average of the two GC-FID retention times were tabulated for each compound along with its molecular formula, molecular weight, and CAS number. In addition to these fields, there are four user-definable columns in table 2 that can be used to add such things as mass spectral information, internal catalog numbers, or comments. Table 2 lists a small portion of the database. It must be noted that all retention time values were created using constant column head pressure. This is because GC-MS retention times are very close to those obtained with other detectors when constant pressure is used. In this mode, GC-MS and GC-FID retention times match within 0.1 minute except for three compounds that elute at the very end of the chromatogram. Even in this case, the differences are no more than 0.2 minute. The discrepancy between GC-MS and GC-FID retention times is larger in the constant flow mode.

280 C 10 min 200 C 0 min 150 C 0 min 70 C 2 min b) 3 Speed 150 C 0 min 70 C 0.67 min 75 C/min 9 C/min 25 C/min 200 C 0 min 3 C/min 280 C 3.3 min 24 C/min 8 C/min

a) Normal Speed

Figure 1. a) GC oven temperature program for the Agilent pesticide method at normal speed. When using this method, chlorpyrifos-methyl must be locked to 16.596 minutes. This method is used by GC-MS and can be used by any other GC system. b) GC oven temperature program for the Agilent pesticide method translated to run three times faster. This method may be used with 6890 GCs configured with any detector except an MSD so long as the GC is configured for fast oven temperature ramping. Chlorpyrifos-methyl must be locked to 5.532 minutes.

Table 2.

Small Portion of the Pesticide and Endocrine Disrupter Retention Time Table That Contains 567 Entries. The retention times shown here are for the pesticide method run at normal speed as shown in figure 1a. Chlorpyrifos-methyl was locked to 16.596 minutes ( 0.015 minute for the collection of the tabulated retention time values. The table includes four additional columns for user-defined information.
Name Acetochlor Fuberidazole Methyl parathion Chlorpyrifos methyl Vinclozolin Plifenat Terbucarb Chloranocryl Heptachlor Carbaryl CAS No. 34256-82-1 3878-19-1 298-00-0 5598-13-0 50471-44-8 21757-82-4 001918-11-2 2164-09-2 76-44-8 63-25-2 Molecular Formula C:14,H:20,Cl:1,N:1,O:2, C:12,H:8,N:2,O:1, C:8,H:10,N:1,O:5,P:1,S:1, C:7,H:7,Cl:3,N:1,O:3,P:1,S:1, C:12,H:9,Cl:2,N:1,O:3, C:10,H:7,Cl:5,O:2, C:17,H:27,N:1,O:2, C:10,H:9,Cl:2,N:1,O:1, C:12,H:15,N:1,O:4, C:10,H:5,Cl:7, C:12,H:11,N:1,O:2, MW 269.77 196.21 263.20 322.53 286.11 336.43 277.41 230.09 237.26 373.32 201.22 MSD RT 16.542 16.549 16.594 16.593 16.630 16.641 16.686 16.736 16.741 16.796 16.806

FID RT 16.542 16.549 16.583 16.596 16.637 16.650 16.689 16.730 16.752 16.773 16.800

Pesticide Screening Method


Figure 2 diagrams the pesticide screening method. First, RTL was used to match GC-AED and GC-MS analyte retention times to those listed in the pesticide table. Software for RTL1416 was used to determine the

3-Hydroxycarbofuran 16655-82-6

column head pressure needed to produce a retention time of 16.596 minutes for chlorpyrifos-methyl. When analyzing samples by GC-AED, the method was usually run at 3 speed and chlorpyrifos-methyl was locked to 5.532 minutes. Figure 3 shows the RTL software screen that is used to develop the retention time calibration. To accomplish this for the pesticide method, one should install the 30 m 0.25 mm 0.25 m HP-5MS column (part no. 19091S-433) and set the column head pressure to one of the appropriate nominal values as shown below, making sure to use the constant pressure mode. 26 psi for atmospheric pressure detectors run at normal speed (eg, NPD, FPD) 16 psi for GC-MSD operated at normal speed 27.5 psi for GC-AED operated at normal speed 88 psi for GC-AED operated at 3 speed

Run GC/AED element-selective chromatograms Use retention time locking so GC/AED, GC/MS, and database have same RTs GC/MSD confirmation Confirmation using GC/AED element ratioing Possible compounds Second column confirmation Done -pesticide identified

Perform pesticide database search based on RT and elemental content

Figure 2. Diagram of the screening method that uses retention time locking and retention time table searching to identify pesticides and suspected endocrine disrupters.

To prepare a calibration table similar to the one shown in figure 3, the chromatographer must make five analyses of chlorpyrifos-methyl at the following column head pressures: the nominal pressure, the nominal pressure + 20%, the nominal pressure + 10%, the nominal pressure 10%, and the nominal pressure 20%. Because of the first run affect, it is usually wise to make one or two blank runs before performing the five calibration runs. The five pressures and the chlorpyrifos-methyl retention times are entered into the table provided by the RTL software. This calibration table stays with the method and can be used to lock, or re-lock, the GC

Figure 3. RTL software screen showing typical retention time locking calibration data for the pesticide method run at normal speed using a GC detector that operates at atmospheric pressure.

method as long as that method is used. That is, the five calibration runs only need to be made once for a given method. The software screen for locking the GC method is shown in figure 4. To lock the method, one enters the retention time of chlorpyrifos-methyl and clicks on the Calc new pressure button. The RTL software calculates the pressure needed to lock the chlorpyrifos-methyl peak at the desired retention time. By clicking on the Update current 6890 Method button, this value is entered automatically into the method. One can use Agilents software for method translation17 to convert the method to other speeds (eg, 1.9 ) and determine the nominal column head pressure required. If this is done, the pesticide table must be exported to a spreadsheet program where the analyte retention times can be divided by the appropriate factor (1.9 in this case). This new table can then be imported back into the ChemStation for use with the new method. After locking the method to the table, GC-AED element-selective chromatograms were obtained for C, Cl, Br, N, S, P, and sometimes F. From the GC-AED chromatograms, it was usually possible to determine which heteroatoms were present or absent in the suspected pesticide peak. RTL software was then used to search the database by retention time and elemental content. Figure 5 shows the RTL software screen used for retention time table searching. One can enter the elements known to be present or not present in the GC-AED peak of interest. Up to six other element-selective detectors can be configured for use in the search algorithm. When the presence or absence of a heteroatom is uncertain,

Figure 4. RTL software screen used to calculate the column head pressure needed to lock or re-lock a method. In this case, the chlorpyrifos-methyl retention time was 16.581 minutes and the pressure needed to re-lock the method was calculated to be 26.33 psi. By clicking on the Update current 6890 Method, button, the new pressure is entered automatically into the GC method.

Figure 5. RTL software screen used to search a retention time table on the basis of retention time and known elemental content. In this case, the software will search the Agilent pesticide table at 16.638 0.1 minutes for compounds that contain N, P, and S but do not contain Br or Cl. If element-selective detectors (such as the NPD) are used, this information can be provided to the search routine. Up to six different element-selective detectors can be configured as shown for NPD, FPD (P), FPD (S), and ELCD. nothing is added to the search routine for that element. One must choose a search time window wide enough to include the correct analyte, but narrow enough to eliminate as many extraneous hits as possible. Experience has shown that the normal speed method requires a search window of 0.2 to 0.3 minute. The 3 speed method can use a search window of 0.1 minute. If the heteroatom content is known for a peak, retention time table searching

with these search windows most often finds just one pesticide and rarely finds more than three possibilities. Confirmation is usually done by GC-MS under locked conditions so that all GC-MS retention times match the values listed in the pesticide retention time table. This was found to be of enormous benefit. Prior to GC-MS confirmation, the analyst already knows which pesticides to look for and their expected retention times. Alternatively, when there is adequate signal to quantitate the analyte in multiple AED element-selective chromatograms, it is often possible to confirm a pesticides identity simply by calculating its heteroatom ratio. GC-AED software for element ratioing facilitates this procedure.

easily confirmed at the expected retention time. In addition, the pesticides trichlorophenol, chlorothalonil, propoxur, and prochloraz were identified. Searching the Cl peak at about 6 minutes gave no pesticide hits. However, GC-MS suggested the presence of a trichloronaphthalene isomer at the corresponding retention time in the GC-MS chromatogram (about 12 minutes because the GC-MS was operated at normal speed). Though not a pesticide, trichloronaphthalene is considered to be a hazardous compound that should not be in food. The same green onion sample was then analyzed by the newer model GC-AED system (6890/ G2350A) at 3 speed (figure 8). Several more pesticides were identified by searching the pesticide/ endocrine disrupter table using a 0.1-minute retention time window. Table 3 lists the pesticide hits that were obtained for each retention time search using the available GC-AED data. Sulfur was not included in any of the searches

because onion extracts have such a high sulfur background. Confirmation by GC-MS was much easier because the GC-MS retention time for each pesticide hit was printed out with the RT search report. Thus, the retention times and probable identities of each pesticide were already known before the GC-MS analysis was run. As is shown in figure 7 for folpet, one can simply extract the ions characteristic for each pesticide hit and look in the extracted ion chromatogram at the expected retention time.

Quantitative Analysis
The Agilent pesticide screening method is a qualitative tool to identify any of the 567 pesticides and endocrine disrupters listed in the retention time table. This, of course, is the first step in any pesticide screening method. Quantitative analysis can be performed in one of two ways.

Analysis of a Green Onion Extract


Numerous samples of fruit and vegetable extracts have been analyzed using this methodology. The results for a green onion extract illustrate the versatility and potential of this method. Green onion extracts are usually very dirty and contain a large number of co-extracted sulfur compounds that can obscure sulfur-containing pesticides. The onion chromatograms shown in figure 6 were run under locked conditions at 2 speed in Tallahassee, Florida, by the Department of Food and Agriculture using a 5890 SERIES II/5921A GC-AED system. Retention time searching indicated that folpet was present in the sample, but it could not be confirmed at the time. The same sample was sent to the Agilent Technologies Little Falls Site in Wilmington, DE, where it was analyzed by scanning GC-MS using an 6890/5973 system. As shown in figure 7, folpet was

40 2,4,5-Trichlorophenol 30 20 10 35 25 Prochloraz 15 4 6 8 10 12 14 16 18 Propoxur Chlorothalonil Trichloronaphthalene Folpet

Cl

N
20

Figure 6. Cl- and N-selective chromatograms of a green onion extract from an 5890/5921A GC-AED system. The analysis was performed at 2 speed under locked conditions in Tallahassee, Florida, by the Department of Agriculture and Consumer Services. In addition to folpet, trichlorophenol, propoxur, and prochloraz were identified by retention time table searching and confirmed by GC-MS at their expected retention times. There were no hits for the Cl peak at about 6 minutes, which was identified by GC-MS as a trichloronaphthalene isomer.

The traditional method is to inject standards into the GC, GC-AED, or GC-MS system to determine response factors from which quantitative results are calculated by the ChemStation software. However, because the GC-AED elemental response is almost independent of molecular structure, compound-independent calibration (CIC) can be used to quantitate all of the pesticides and endocrine disrupters that are found. For example, one could spike chlorpyrifos-methyl (C7H7Cl3NO3PS) at a known concentration into each pesticide extract and obtain elementspecific calibration curves for Cl, N, P, and S. These curves could then be used to calibrate for any other compound containing one or more of these elements. Because the GC-AED is quite stable, external standard CIC often works just as well. The GC-AED software facilitates CIC. Unfortunately, this procedure determines the amount of a compound that reaches the AED and does not compensate for losses due to decomposition or adsorption in the inlet or column.

Green Onion
Folpet

4.00

8.00

12.00

16.00

20.00

24.00

28.00

M/Z 260, 294, 297 Folpet (21.637 min) Pesticide table RT = 21.594 min

21.20

21.60

22.00

22.40

22.80

23.20

Figure 7. Confirmation of folpet in a green onion extract. The tabulated GC-MS retention time is 21.594 minutes, and folpet was detected in this sample at 21.637 minutes by simply extracting its characteristic ions.

N 174
6 1

1. Dichlorvos 2. 2,4,5-Trichlorophenol 3. Propoxur 4. Trichloronaphthalene 5. Chlorothalonil 6. Chlorpyrifos-methyl 7. Folpet 8. Mirex 9. Prochloraz

Conclusions
Most screening procedures in use today are capable of finding only a fraction of the pesticides that are registered around the world. This new method has the capability of screening for virtually any volatile pesticide, metabolite, or endocrine disrupter. Although confirmation is usually required, GC-MS analysis is made much easier and more reliable because the pesticides retention time and probable identity are already known.

P 178
4 5

Cl 479
7 8 9 12 14

10

Figure 8. Element-selective chromatograms obtained for the same green onion extract shown in figure 6. These chromatograms were obtained at 3 speed using an 6890/G2350A GC-AED system.

While GC-AED is an ideal tool for element-selective pesticide screening,2022 many laboratories rely on a combination of other selective detectors. It is still possible to apply this method if each GC system runs the Agilent pesticide method under the same locked conditions. Any combination of GC-AED and/or elementselective detector response data can be entered into the RTL searching software. When combined with RTL and retention time searching, GC-AED and GC-MS provide the most comprehensive and reliable screening method available for pesticides, metabolites, and suspected endocrine disrupters. Unlike most target compound methods in use today, this procedure has a good chance of finding and identifying unexpected or unknown pesticides, even in complex food extracts. RTL software makes it easy to add more compounds to the method, simply by determining their retention times under the same locked conditions. Retention time locking with database searching could easily be applied to similar types of analyses. For example, one might use the procedure to identify polychlorinated biphenyls, polynuclear aromatics, drugs of abuse, or flavor and fragrance compounds.

Table 3.

Green onion pesticide "hits" obtained by searching the 567-compound pesticide/endocrine disrupter RT table using a 0.1-minute RT window and elementselective GC-AED data. Compounds confirmed by GC-MS are shown for comparison. The GC/MS (figure 7) and GC-AED (figure 8) chromatograms were obtained at normal and 3X speeds, respectively. Sulfur peaks were not used to narrow the search because of the high background of sulfur-containing compounds in the onion extract.
RT Search Hits Dichlorvos 2,4,6-Trichlorophenol 2,4,5-Trichlorophenol Fenobucarb Propoxur 4,6-Dinitro-o-cresol No pesticide hits Terbacil Chlorothalonil Chlorpyrifos-methyl Folpet Chlorbenside Mirex Prochloraz Confirmed by GC-MS Dichlorvos 2,4,5-Trichlorophenol Propoxur

GC-AED RT 1.933 2.281 3.440

3.854 4.955 5.538 7.232 9.965 10.588

Trichloronaphthalene isomer Chlorothalonil Chlorpyrifos-methyl Folpet Mirex Prochloraz

beta testing the method; Matthew Klee and Leonid Blumberg for many useful discussions; and James Green and Takeshi Otsuka for their help in developing the retention time table.

5. J. Fillion, R. Hindle, M. Lacroix, and J. Selwyn, J AOAC Int 78, 12521266, 1995. 6. Working Group on Development and Improvement of Residue-analytical Methods, Analytical Methods for Pesticide Residues in Foodstuffs, General Inspectorate for Health Protection, Ministry of Public Health, Welfare and Sport, The Netherlands, 1996. 7. M. A. Luke, J. E. Froberg, G. M. Doose, and H. T. Masumoto, J Assoc Off Anal Chem 64, 11871195, 1981. 8. H.-J. Stan (Ed.), Analysis of Pesticides in Ground and Surface Water II, Springer-Verlag, Berlin, Germany, 1995. 9. R. E. Wagner, W. Kotas, and G. A. Yogis (Eds.), Guide to Environmental Analytical Methods, 2nd Edition Genium, Schenectady, NY, 1994. 10. U.S. Environmental Protection Agency, Test Methods for Evaluating Solid Waste, SW-846, Draft Method 8085: Pesticides by GCAED.

References
1. C. Tomlin (Ed.), The Pesticide Manual, Eleventh Edition, British Crop Protection Council, Farnham, Surry, UK, 1997. 2. R. W. Miller, This is Codex Alimentarius, Secretariat of the Joint FAO/WHO Food Standards Programme, Food and Agriculture Organization of the United Nations, Rome. 3. B. M. McMahon and N. F. Hardin (Eds.), Pesticide Analytical Manual, Vol. I, Third Edition, U.S. Food and Drug Administration, Washington, D.C., 1994. 4. S. M. Lee, M. L. Papathakis, H.-M. C. Feng, G. G. Hunter, and J. E. Carr, Fresenius J Anal Chem 339, 376383, 1991.

Acknowledgments
The authors wish to thank the following people for their contributions to the development of this method: Joanne Cook and Marc Engel (Florida Department of Agriculture and Consumer Services) for supplying a green onion chromatogram, for contributing fruit and vegetable extracts, and for

10

11. T. Colborn, D. Dumanoski, and J. P. Myers, Our Stolen Future Penguin, New York, NY, 1996. 12. Food Quality Protection Act of 1996, Public Law 104-170, Congressional Record, pp. H8127H8141. 13. Safe Drinking Water Act Amendments of 1996, Public Law 104-182, Congressional Record pp. H9680H9700. 14. V. Giarrocco, B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications, Application Note 228-392, Publication (23) 5966-2469E, December 1997. 15. P. L. Wylie and B. D. Quimby, Abstracts of the 1st European Pesticide Residue Workshop, Alkmaar, The Netherlands,| Paper # O-023, June 1012, 1996. 16. M. Klee and T. Sullivan, Environ. Testing and Analysis, pp. 1617 and 4445, March/April 1998. 17. Hewlett Packard Company, Software for Method Translation, Available on the World Wide Web at: http://www.chem. agilent.com/cag/servsup/ usersoft/main.html. 18. M. L. Lee, F. J. Yang, and K. D. Bartle, Open Tubular Column Gas Chromatography, Theory and Practice, Wiley, New York, NY, pp. 217218, 1984. 19. C. Bicchi, A. DAmato, and A. Binello, J High Resol Chromatogr 19, 8084, 1996. 20. S. M. Lee and P. L. Wylie, J Agric Food Chem 39, 21922199, 1991. 21. P. L. Wylie and R. Oguchi, J Chrom 517, 131142, 1990. 22. N. L. Olson, R. Carrell, R. K. Cummings, and R. Rieck, LC-GC 12, 142154, 1994. 23. P. L. Wylie and K. Uchiyama, J AOAC Int 79, 571577, 1996.

24. P. L. Wylie, K. L. Klein, M. Q. Thompson, and B. W. Hermann, J High Resol Chromatog 15, 571577, 1992. 25. W. D. Snyder and L. M. Blumberg, Proceedings of the Fourteenth International Symposium on Capillary Chromatography, Baltimore, MD, pp. 2838, May 2529, 1992. 26. B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie, Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking, Application Note 228401, Publication 5967-5820E, April 1998. 27. http://www.nijs.go.jp/hse/ environ/edsubs/substancesnew. html. 28. http://www.wwfcanada.org/ hormone-disruptors. 29. http://www.epa.gov/opptintr/ opptendo/index.htm. 30. http://www.osf-facts.org/ basics.chemlist.html. 31. http://www.mst.dk/liste.htm. 32. Hewlett-Packard Company. Mass Spectral Libraries: HP Pest library, Wiley 275 library, NBS75K Library, Palo Alto, CA. 33. Pesticide Reference Spectra, Vol. 14, Spectral Service Gmbh, Kln, Germany and Riedel-de Han AG, Seelze, Germany, 1992.

11

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Microsoft and Windows NT are U.S. registered trademarks. HP is a registered trademark of Hewlett-Packard Company. Copyright 2000 Agilent Technologies, Inc. Printed in the USA 3/2000 5967-5860E

The Analysis of Chlorinated Pesticides and PCBs Using the HP-608 Capillary Column
Application Note 228-236

Authors
Imogene L. Chang, PhD Winfred J. Sanders, PhD

Abstract
Chlorinated pesticides and PCBs targeted in EPA Methods 608, 8080, 8081, and CLP pesticides for wastewater and solid wastes are analyzed under optimum conditions at a constant flow of 2.4 ml/min. The merits of splitless and on-column injection techniques using the Agilent 5890 Series II GC with electronic pressure control (EPC) are compared. Key Words: chlorinated pesticides, PCBs, on-column injection, splitless injection, HP-608 capillary column, EPA 608, EPA 8080/8081, CLP pesticides, electronic pressure control.

These EPA Methods allow laboratories to substitute columns of their choice provided that performance data such as chromatographic resolution, analyte breakdown, and MDLs (minimum detectable levels) are equal to or better than those provided with the EPA methods. The HP-608 is a wide bore (530 m-id) capillary column specially designed for the analysis of organic pesticides. GC/ECD separations of chlorinated pesticides and PCBs were done using the HP-608 column with both on-column and splitless inlet sample introductions. In both cases, the HP-608 provided superior chromatographic resolution, excellent reproducibility, and minimal analyte breakdown for the analysis of pesticides and PCBs. Table 1. Experimental Conditions
Instrument Requirements Gas chromatograph: Injection ports: Column: Detector: Sample introduction: Data collection: Experimental Conditions Column: Carrier gas: Oven:

Experimental
A 30 m x 530 m x 0.5 m HP-608 column (part no. 19095S-023) was used under constant carrier gas flow using the 5890 Series II GC with EPC equipped with a split/splitless inlet and a cool on-column inlet. Equipment included the 7673 automatic sampler with tray and the electron capture detector (ECD). Samples were introduced in both the on-column and splitless modes. The MerlinTM Microseal septum (part no. 5181-8816) was used in the split/splitless inlet to replace the conventional inlet septum. A deactivated tapered glass liner (part no. 51813316) was used for all splitless injection runs. GC conditions were controlled using the HP 3365

Introduction
Chlorinated pesticides and PCBs have been banned in the U.S. for several years. However, because of their persistence in the environment, EPA methods 8080/8081 and CLP pesticides target 16 to 20 chlorinated organic pesticides in the evaluation of solid waste. This includes pesticides, their degradation products, technical grades of chlordane, toxaphene, and PCBs in solid waste.1,2 EPA Method 608 targets similar pesticides in industrial and wastewater discharges.3 EPA Methods 608 and 8080 prescribe packed-column analysis, whereas Methods 8081 and CLP pesticides prescribe capillary column analysis.

Agilent 5890 Series II with EPC Split/splitless inlet with temperature and pressure programmable features On-column inlet with temperature and pressure programmable features HP-608, 30 m x 530 m x 0.5 m (Part number 19095S-023) ECD 7673 splitless fast injection On-column injection 3365 ChemStation and HP Vectra 486/133T HP-608, 30 m x 530 m x 0.5 m (Part number 19095S-023) He, 20 cm/sec, 2.2 psi at 80C with EPC under constant flow of 2.4 ml/min First ramp: 80C (hold 1 min) to 190C at 30C/min Second ramp: 190C to 280C (hold 1 min) at 6C/min Third ramp: 280C to 300C (hold 2 min) at 20C/min Splitless: 1 l, inlet temperature of 250C On-column: 1 l oven track for inlet temperature program ECD (330C), 65 ml/min N2 makeup, 6 ml/min anode purge Pesticides and PCB standard solutions in isooctane

Injection: Detector: Sample:

ChemStation. Data was managed with a HP Vectra PC (486/33T). Instrument parameters and experimental conditions are listed in Table 1. Pesticide solutions containing 16 to 22 components were prepared from the dilution of certified standards (part no. 8500-5873 and 8500-5876, mixes A and B: level 2) with isooctane (pesticide residue grade from Burdick & Jackson). Pesticide standards (part no. 5062-3589), including four vials of 16 EPA-608 pesticides and two vials of two component inlet check solutions (endrin/DDT concentrations are 50 ppb/100 ppb), were used without further dilution. These pesticide compounds are listed in Table 2.

Figure 1.
1.6e5 1.4e5 1.2e5 1.0e5 8.0e4 6.0e4 4.0e4 2.0e4 0 6 1.6e5 1.4e5 1.2e5 1.0e5 8.0e4 6.0e4 4.0e4 2.0e4 0 6

Chromatograms of the 16 chlorinated pesticides under optimum GC conditions, 100 pg of each pesticide injected. Peak identification in Table 2.
1 2 5 6 4 3 7 10 11 12 15 14 17 18 13 16

Figure 1A. Splitless Injection

a 8 1 10 12 14 16 18 20 min

Figure 1B. On-Column Injection


2 5 6 4 7 11 10 12 13 3 16

15 14 17 18

a 8 10 12 14 16 18 20 min

Table 2. Chlorinated Pesticides


Peak No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SS1 SS2 EPA-608 alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Compound Name EPA-8080/8081 alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Chlordane-gamma Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor EPA-CLP Pesticides alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Chlordane-gamma Chlordane-alpha Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Tetrachloro-m-xylene Decachlorobiphenyl

Results and Discussion


Splitless Analysis
Figure 1A shows the analysis of a standard solution containing the 16 EPA-608 targeted pesticides at a constant column flow of 2.4 ml/minute. One microliter of sample (100 pg of each component) was introduced in splitless mode at 250C under the conditions4 listed in Table 1. All 16 components were well resolved in sharp symmetric peaks, and the analysis was completed in less than 17 minutes. The 30-m HP-608 (530 m id) column possesses sufficient efficiency to completely resolve the complex pesticides mix, including chlorinated compounds with similar or isomeric structures. The absence of coeluting peaks on the HP-608 column permitted fast and accurate identification and quantitation.

Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate a-Degradation product

Low-Temperature On-Column Analysis


Figure 1B shows the same pesticides standard mix using the cool on-column injection technique. On-column injection of 1 l of sample at 80C resulted in little sample degradation, minimal byproducts, and good sensitivity (see Table 3). Common to both Figures 1A and 1B is the absence of tailing peaks, including the endrin aldehyde peak (peak 17), indicating the HP-608 column surface is very inert.

Table 3. Reproducibility of Pesticide Analysis


Retention Times, min Pesticides alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate a, Degradation product alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate a, Degradation product Mean 8.423 9.225 9.352 9.984 10.181 10.760 13.036 13.623 13.838 14.814 15.135 15.311 15.975 16.208 16.570 18.690 Std Dev 0.004 0.004 0.004 0.004 0.005 0.004 0.003 0.004 0.004 0.004 0.004 0.004 0.004 0.003 0.004 0.003 0.003 % RSD 0.047 0.046 0.046 0.042 0.044 0.039 0.028 0.031 0.026 0.027 0.025 0.024 0.025 0.021 0.022 0.021 0.017 Mean 431643 393514 208287 310294 390027 359246 359586 321622 341930 336042 268560 254389 297580 259369 205588 281397 3416 A. On-column injection (100 pg each component) 7497 6496 3428 5430 7428 6996 5740 5478 7070 4832 5298 3017 4326 3881 1876 4143 97 1.74 1.65 1.65 1.75 1.90 1.95 1.60 1.70 2.07 1.44 1.97 1.19 1.45 1.50 0.91 1.47 2.83 Area Counts Std Dev % RSD

Heptachlor epoxide 12.385

Reproducibility
Reproducibility for the analysis of chlorinated pesticides using HP-608 columns with the HP GC/ECD system was excellent (see Table 3). The RSD (relative standard deviation) in absolute area counts for all 16 EPA targeted pesticides was less than 2% for on-column runs (two sets of six replicate injections). Similarly, the peak area counts reproducibility for all splitless injection runs (three sets of six replicate injections) was in the 1% to 2% RSD range using the same standard sample. The standard deviation of retention times was within 0.0030.005 minutes and 0.002 minutes for on-column and splitless runs, respectively. In comparison, the standard deviation of retention times for EPA Method 8081 analysis (Table 10, reference 1) using wide-bore capillary columns ranged from 0.007 minutes to 0.013 minutes for the same set of pesticides. This clearly demonstrates that chromatographic reproducibility obtained using the HP-608 capillary column is better than that obtained using the capillary columns stipulated in EPA Method 8081.

B. Splitless injection (100 pg each component) 8.351 9.146 9.273 9.898 10.097 10.671 12.938 13.527 13.735 14.710 15.034 15.207 15.874 16.103 16.467 18.584 0.002 0.002 0.002 0.002 0.001 0.002 0.001 0.002 0.001 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.020 0.020 0.018 0.018 0.013 0.015 0.011 0.014 0.011 0.014 0.013 0.013 0.015 0.012 0.010 0.013 0.012 376446 317405 165105 207924 301779 308689 289985 253489 313249 209054 160235 168113 228810 168810 148655 190284 21513 7222 6592 3129 4637 6113 6422 6216 5496 6102 3925 3104 3094 4868 2129 3687 3003 1747 1.92 2.08 1.90 2.23 2.03 2.08 2.14 2.17 1.95 1.88 1.94 1.84 2.13 1.26 2.48 1.58 8.12

Heptachlor epoxide 12.289

Comparison of Sample Introduction Techniques


For all on-column injection runs, degradation was negligible due to the low initial column temperature (80C) and the direct introduction of a liquid sample plug into an inert column. As a result, inlet-related sample discrimination, alteration, and degradation were eliminated, while the advantages of solvent focusing and stationary phase focusing were maximized. Routine analysis of the inlet check solution (specified by the EPA methods) showed that the average degradation was less than 3% for endrin and 1% for DDT. As demonstrated by the clean baseline in Figure 1A, little sample degradation occurred at an inlet temperature of 250C. However, a small endrin ketone peak (RT of 18.69 minutes) appeared on the chromatograms from the GC runs with both on-column and splitless injection shown in Figures 1A and 1B. A closer look (Table 3), shows that the area counts for endrin ketone (peak a, a byproduct of endrin degradation) measured 5 times larger in the splitless runs than for the on-column runs (average absolute area counts of 3,400 versus 21,000). The GC runs of the inlet check standard (after 200 repeated splitless injections), showed a 7% endrin degradation and 10% DDT degradation. These values were well below the EPA requirement of 15% degradation for both endrin and DDT. Use of the MerlinTM Microseal5 and the deactivated glass liner also contributed directly to the low degradation rate in the splitless mode. The Microseal is designed to provide a good inlet seal without using a conventional septum. By eliminating the introduction of particulates into the inlet liner from conventional septum, useful life for the inlet liner is extended, down time (to change a liner and a conventional septum) is reduced, and laboratory throughput is increased. The use of splitless injection technique may also prevent interference from extraneous and high boiling

Figure 2.
2.2e4 2.0e4

Chromatograms of isooctane under optimum GC conditions, 1 l injected. (b,k=solvent contaminants)

Figure 2A. Splitless Injection


1.8e4 1.6e4 1.4e4 1.2e4 1.0e4 8000 6000 4000 6 2.2e4 2.0e4 8 10 12 14 16 18 20 min b b

Figure 2B. On-Column Injection


1.8e4 1.6e4 1.4e4 1.2e4 1.0e4 k 8000 6000 4000 6 8 10 12 14 16 18 20 min b b b b

materials in dirty samples. This is demonstrated in Figures 2A and 2B. Figure 2 shows the analysis of isooctane solvent (pesticide-residue grade) using both splitless (Figure 2A) and on-column injection (Figure 2B). The late-eluting peak (peak k) , at 16.69 minutes retention time in the on-column run, does not appear in the chromato-gram of the splitless run (Figure 2A). This peak, possibly a high boiling contaminant in isooctane, appears again in Figure 3B. Figures 3A and 3B show analyses of a 10-ppb pesticide standard using splitless injection and on-column injection, respectively. The peak (peak k) eluting just before endosulfan sulfate

(peak 18) may cause a higher value for the determination of trace endosulfan sulfate in the sample. Both area counts and peak heights for the splitless runs were smaller than those for the on-column injection runs (see Table 3). For example, the average counts of lindane from the splitless runs were approximately 80% of those from the on-column injections (Table 3). Therefore, oncolumn injection is a good choice for clean samples and trace analyses demanding high sensitivity and low detection limits (large area counts).

Analysis of PCBs and EPA Methods 8080, 8081, and CLP Pesticides
For wastewater and solid waste samples, the EPA recommends splitless injection for the determination of pesticides and PCBs. Using splitless injection under optimum 5890 Series II GC conditions, all 17 pesticides targeted by EPA Method 8080B are resolved as shown in Figure 4. Among the 20 components targeted by EPA Methods 8081 and CLP pesticides, all but alpha-chlordane and endosulfan I (they are partially separated) are well resolved by the HP-608 column (Figure 5). Since the HP-608 column can effectively separate the complex mix of these pesticides, it is a good column choice for the determination of PCBs and multiple-peak response pesticides such as chlordane and toxaphene. Figure 6 shows a comparison of chromatograms for technical grade chlordane and toxaphene, while Figure 7 is a comparison of chromatograms for seven PCBs, all analyzed under the same GC conditions using the HP-608 capillary column.

Figure 3.
2.2e4 2.0e4

Chromatograms of dilute pesticides mix under optimum GC conditions; 10 pg of each pesticide injected. (Peak ID, see Table 2)

Figure 3A. 1 Splitless Injection


1.8e4 1 1.6e4 1.4e4 1.2e4 3 1.0e4 8000 a 6000 4000 6 2.2e4 1.0e4 1.2e4 1.6e4 1.4e4 1.2e4 1.0e4 8000 6000 4000 6 8 10 12 14 16 18 20 min k a 3 8 1 2 5 4 6 7 10 11 14 12 13 15 16 18 17 10 12 14 16 18 20 min 4 2 5 6 7 10 11 12 13 15 14 16 18 17

Figure 3B. 1 On-Column Injection

Figure 4.

Chromatograms of the EPA-Method 8080 pesticides under optimum GC conditions. Splitless injection of 100200 pg per component. (Peak ID, see Table 2)
1 2 5 6 7 10 4 3 13 12 14 15 18 16 17 19

1.2e5

11

a 0 6 8 10 12 14 16 18 20 min

Conclusion
Under optimal conditions, the HP-608 column separates 16 EPA-608 pesticides in 17 minutes and 20 EPA-CLP pesticides (and EPA-8081 pesticides) in 19 minutes (22 minutes including the surrogate, decachlorobiphenyl). Both splitless and on-column injections yield little sample degradation and provide excellent reproducibility of retention times and area responses. On-column injection is more suitable for clean samples and trace analysis, while splitless injection is better used for wastewater and waste samples.

Figure 5.

Chromatogram of pesticides targeted in EPA-method 8081 and CLP pesticides under optimum GC conditions. Splitless injection of 50100 pg per component. (Peak ID, see Table 2)
SS1 11 19 SS2

1.2e5

1.0e5 1 8.0e4 12 10 5 6.0e4 4 4.0e4 3 6 7 9 8 13 15 20 14 16 17 18

2.0e4

0 5 10 15 20 min

Figure 6.

Chromatogram of technical grade toxaphene and chlordane under optimum GC conditions. Splitless injection of 1 l 2.5 ppm mix

2.0e5

Chlordane

1.5e5

1.0e5

6.0e5

0 5 1.4e5 1.2e5 1.0e5 8.0e5 6.0e5 4.0e5 2.0e5 0 5 10 15 20 25 min Toxaphene 10 15 20 25 min

Figure 7.
1.4e5

A comparison of seven PCBs under optimum GC conditions. Splitless injection of 1 l 2.5 ppm each
1.6e5 Aroclor 1016 Aroclor 1248

0 6 2.0e5 Aroclor 1221 8 10 12 14 16 18 20 min

0 6 1.6e5 Aroclor 1254 8 10 12 14 16 18 20 min

0 6 70000 Aroclor 1232 8 10 12 14 16 18 20 min

0 6 1.6e5 Aroclor 1260 8 10 12 14 16 18 20 min

0 6 1.6e5 Aroclor 1242 8 10 12 14 16 18 20 min

0 6 8 10 12 14 16 18 20 min

0 6 8 10 12 14 16 18 20 min

Acknowledgment
The authors wish to thank Dr. D. Pautler for his many helpful discussions.

References
1. EPA Method 8080B and 8081, Test Methods for Evaluating Solid Waste, SW-846, Revision 2, Nov. 1990. 2. U.S. EPA Contract Laboratory Program Statement of Work for Organics Analysis Document Number OLM01.0, 1990. 3. EPA Method 608, Methods for Organic Chemical Analysis of Municipal and Industrial Wastewater, PB82-201798, 1982. 4. I. L. Chang and W. J. Sanders, Method Development for EPA608 Analysis Using a HP-608 Capillary Column, HewlettPackard Application Note 1993 (in preparation). 5. Introducing the Merlin MicrosealTM Septum, Pub. No. 5091-3197EUS, 1991.

www.agilent.com

Copyright 1993, 2000 Agilent Technologies Printed in USA 3/00 5091-7567E

Fast Screening of Pesticide and Endocrine Disrupters Using the Agilent 6890/5973N GC/MSD System, Part I

Application
Gas Chromatography January 2000 retention times and characteristic ions for 567 of the most common pesticides and endocrine disrupters of concern worldwide. A GC/MSD method was developed based on the standard 42-min method1 to screen for all 567 of the most common analytes. A specific combination of column stationary phase, carrier gas flow rate, and oven temperature programming is required to lock all the compounds to an expected retention timetable2. Compound identification based only on spectral searching alone is difficult when analyzing extracts containing significant sample matrix content because of overlapping peaks and noisy baselines. The new screening tool, integrated within Agilents ChemStation for MSD, searches for all 567 compounds by first checking and integrating four characteristic ions within the expected time window, and second by printing out a report showing hits and possible hits (ratios of characteristic ions that do not match the expected values in the library within specified limits). In one application, the analysis time of the standard pesticide method was reduced by one half, two-thirds, and three-fourths. The faster methods were scaled exactly as predicted by using a combination of Agilents method translation (MTL) and RTL software. Because scaling was exact, these faster methods can be used with precisely-scaled pesticide libraries, making the screening process even more powerful and adaptable to individual needs.

Authors C. Kai Meng Agilent Technologies 2850 Centerville Road Wilmington, Delaware 19808-1610 USA Michael Szelewski Agilent Technologies 2850 Centerville Road Wilmington, Delaware 19808-1610 USA

Abstract Agilent Technologies new, fast GC/MSD method can significantly speed up the screening of pesticides. Agilents GC method translation software (available free from the Agilent Technologies Web site, http://www. chem. agilent.com/cag/ servsup/usersoft/main.html#mxlator) was used in developing the new method based on the standard 42-min method. A 10 m x 0.1 mm x 0.1 m HP-5 column was used to increase analysis speed up to fourfold. The time savings were implemented in increments (down to 10.5 minutes) to verify the predictability of scaling and the effect of scaling on the signal-tonoise ratio. Key Words RTL, pesticide, environmental, screening, fast GC, method translation, 5973, 6890, MTL Introduction Analysts want faster analyses to improve laboratory productivity. Often, when speeding up GC methods, an analyst will trade resolution for increased analysis speed. This loss of resolution can complicate peak identification, even with a mass selective detector (MSD). Agilent Technologies has developed new techniques to solve the peak identification problem based on Agilents retention time locking (RTL) software and a new mass spectral library that contains the locked

Experimental The GC method translation software tool was used to find operating conditions for the faster methods. Figure 1 is a screen capture of MTL software data entry showing the original conditions and the new chromatographic conditions for a twofold speed gain. The column flow rate, which is helpful to avoid exceeding MSD pumping capacity3, is also found in the table. A 16:1 split ratio was suggested in the table as a proportional scaling from the original column to the smaller i.d. column with corresponding lower capacity. The program also determined the required column head pressure and corresponding oven ramp. The Agilent 6890 GC fast oven option (220/240V in the U.S.) was required for the faster oven ramp used in this study.

Figure 1. Screen capture showing the method translation (MTL) software data entry used in a twofold speed gain translation.

General chromatographic conditions are listed in table 1. The standard used was a mixture of 26 pesticides at 10 ppm. A 10 m x 0.1 mm x 0.1 m HP-5 column (part number 19091J141) was used. The head pressure determined by the method translation software (30.72 psi) was used as the starting point for retention time locking. The column head pressure required to lock retention times of the compounds to the library (the original retention time divided by 2) was determined using the automated RTL process integrated within the Agilent ChemStation for MSD. This process (first translate the method then lock the retention times) was repeated for the threefold and fourfold time reductions.

Table 1. Chromatographic Conditions Speed GC Column Injection mode Column head pressure Column flow (mL/min) Inlet control mode Carrier gas Injector temperature Oven temperature Ramp 1 Ramp 2 Ramp 3 Oven equilibration Injection volume Liner MS Conditions Solvent delay Tune file Low mass High mass Threshold Sampling Scans/sec Quad temperature Source temperature Transfer line temperature Acquisition mode Onefold (1X) 110 V 30 m x 0.25 mm x 0.25 m HP5-MS (P/N 19091S-433) Splitless 18.0 psi 1.5 Constant pressure Helium 250 C 70 (2 min) 25 C/min 150 (0 min) 3 C/min 200 (0 min) 8 C/min 280 (10 min) 2 min 1 L 5183-4647 Twofold (2X) Threefold (3X) 220/240 V 10 m x 0.1 mm x 0.1 m HP-5 (P/N 19091J-141) 16:1 split 36.55 psi 63.17 psi 0.4 0.8 Constant pressure Helium 250 C 70 (1 min) 70 (0.67 min) 50 75 150 (0 min) 150 (0 min) 6 9 200 (0 min) 200 (0 min) 16 24 280 (5 min) 280 (3.33 min) 2 min 1 L 5183-4647 Fourfold (4X)

90.0 psi 1.5

70 (0.5 min) 100 150 (0 min) 12 200 (0 min) 32 280 (2.5 min)

3 min Atune.u 35 amu 500 amu 150 2 3.15 150 C 230 C 280 C Scan (EI)

1.8 min Atune.u 35 amu 450 amu 250 2 3.50 150 C 230 C 280 C Scan (EI)

1.2 min

0.9 min

1 6.54

1 6.54

Figure 2 shows the results of the shortened analysis times. The three chromatograms look extremely similar, except that the time axis is scaled proportionally. Because MTL followed by RTL scales methods very precisely, scaled screening libraries for corresponding time reductions can be obtained by dividing the retention times in the library by the speed gain (which does not have to be an integer). The peak heights from all the methods are very similar. Although the sample was split 16:1 for the smaller column, the small column i.d. and faster oven ramp combination made the peaks narrower and higher, so there was minimal loss in the signal to noise ratio. Conclusion The highly accurate and reproducible pressure and temperature control of the Agilent 6890 GC allows precise scaling of the standard 42-min GC/MSD pesticide method. Run time was shortened to 10.5 minutes using a fast oven ramp rate and a 10-meter 100-micron column. The combination of MTL and RTL facilitated scaling and yielded exact scaling. RTL libraries can accurately be scaled to correspond to the faster analyses. References 1. B. D. Quimby, L.M. Blumberg, M. S. Klee, and P. L. Wylie, Precise TimeScaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking, Application Note 228-401, Agilent publication number 5967-5820E, May 1998. 2. H. Prest, P. L. Wylie, K. Weiner, and D. Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the HP 6890/ 5973 GC/MSD System, Agilent publication number 5968-4884E, April 1999. 3. H. Prest, GC Column Selection and Pumping Considerations for Electron and Chemical Ionization MSD operation, Agilent publication Number 5968-7958E, November 1999.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies Printed in USA 2/00 5968-9220E

Abundance 2,400,000 2,200,000 2,000,000 1,800,000 1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 Time 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00

2X Speed

18 min
Abundance 6,000,000 5,500,000 5,000,000 4,500,000 4,000,000 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 Time 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00

3X Speed

Abundance 6,000,000 5,500,000 5,000,000 4,500,000 4,000,000 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 Time 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

12 min

4X Speed

9.00

10.00

9 min
Figure 2. Three TICs of the 2X, 3X, and 4X speedups. The standard analysis (1X) was 42 minutes long. The two vertical lines on the figure are used as references to show the similarity of the TICs.

Analysis of Glyphosate in Water with Postcolumn Derivatization using HPLC


Rainer Schuster Food

Abstract The HPLC method presented here was used for the direct analysis of glyphosate in water with postcolumn derivatization. HPLC method performance Limit of detection 1ppb Repeatability of RT over 10 runs <0.8 % of areas over 10 runs <2.2 %

Conditions
Column 150 4 mm cation exchange, K+ form from Pickering, 8 m Mobile phase A = 5 mM KH2PO4, pH = 2.0, B = 5 mM KOH Flow rate 0.4 ml/min Gradient at 15 min 0% B; at 17 min 100% B Column compartment 55 C Injection vol 50 l standard Fluorescence detector Excitation wavelength: 230 nm or 330 nm, Emission wavelength: 425 nm Slit width excitation: 2 mm (25 nm) Slit width emission 1: 4 mm (50 nm) Slit width emission 2: 4 mm (50 nm) Photomultiplier gain: 2 Cut-off filter: 370 nm Lamp: 55 Hz (always on) Response time: 4 s Derivatization reagent pump flow rate for hydrolization agent: 0.3 ml/min (OCl-) flow rate for derivatization agent: 0.3 ml/min (OPA). Sample preparation None

Norm. 7.5 7

Glyphosate AMPA

6.5 6 5.5 5 2.5 5 7.5 10 12.5 Time [min] 15 17.5 20

Figure 1 EIC traces from amine standards

Agilent Technologies
Innovating the HP Way

Equipment
Agilent 1100 Series degasser quaternary pump autosampler Pickering post-column derivatization system fluorescence detector, Agilent ChemStation + software

Rainer Schuster is application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Copyright 1997 Agilent Technologies Released 09/97 Publication Number 5966-0743E

Agilent Technologies
Innovating the HP Way

Analysis of Pesticides in Salad Samples and Spices using HPLC


Rainer Schuster Food

Abstract The following compound classes of pesticides have been analyzed: triazines, phenylureaherbicides, methabenzthiazuron, diquat, paraquat, and mercaptobenzothiazol. Carbamates and glyphosate also have been analyzed but with different equipment. In most countries, growing concern about the residues of pesticides in food products is evident. Therefore, regulations limiting the concentration of pesticides in foodstuffs have been introduced to protect consumers from contaminated food products. Several methods are used to control these limits. HPLC is recommended for the analysis of low volatile compounds and for compounds that are unstable when heated.

Conditions
mAU 120
80

3 different salad samples

Vinclozolin

Folpet 40 0 10 15
20 25 Time [min] 30

Carbendazim*

35

40

Figure 1 Analysis of pesticide residues in three different salad samples * Carbendazim has a low recovery rate of only approximately 40 %

Column 100 3 mm Hypersil BDS, 3 m Mobile phase water/ACN (95:5) Gradient at 10 min 25% ACN at 26 min 42% ACN; at 34 min 60% ACN Flushing time 10 min at 100% ACN Post time 6 min Flow rate 0.5 ml/min Oven temp 42 C Injection vol 310 l Detector UV-DAD detection wavelengths 214/15 nm, 230/20 nm, and 245/20 nm reference wavelength 400/80 nm Sample preparation Salad was homogenized and then extracted with liquid/liquid extraction. The extract was cleaned with gel permeation chromatography using cyclohexane/ethyl acetate. Spices were prepared according to Specht 22 with gel permeation chromatography.

Agilent Technologies
Innovating the HP Way

Sample preparation Sample preparation and enrichment depend strongly on the matrix. Drinking water samples, for example, must be extracted using solidphase extraction, whereas vegetables are extracted with liquid/liquid extraction after homogenization, followed by additional cleaning and sample enrichment. Chromatographic conditions The HPLC method presented here was used for the analysis of pesticides in salad samples and spices.1, 2 HPLC method performance Limit of detection 0.01 g/l Repeatability of RT over 10 runs <0.2 % of areas over 10 runs <1 %

Equipment
Agilent 1100 Series degasser quaternary pump autosampler thermostatted column compartment diode array detector, Agilent ChemStation + software

mAU 100
80 60

Vinclozolin Procymidon Nitro compounds Chlorpyripho-ethyl Paprika (Spain)

40 20 0 10 20 30 Time [min] 40

Paprika (Turkey)

50

Figure 2 Analysis of pesticide residues in two paprika samples

Rainer Schuster is application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Copyright 1997 Agilent Technologies Released 09/97 Publication Number 5966-0742E

Agilent Technologies
Innovating the HP Way

Pesticides & Residues Antibacterial Drug Residues

> Return to Table of Contents > Search entire document

Detection, Confirmation, and Quantification of Chloramphenicol in Honey, Shrimp and Chicken Using the Agilent 6410 LC/MS Triple Quadrupole
Application
Food Safety

Authors
Yanyan Fang Agilent Technologies (Shanghai), Co. 412 Yin Lun Road 200131 China Jerry Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilimington DE 19808 USA Zhuwei Wang Agilent Technologies (China) Beijing

meet the EU identification points needed for confirmation. This study is a valuable indicator of the ability of the QQQ for routine quantitative trace analysis of chloramphenicol in honey, shrimp, and chicken.

Introduction
Chloramphenicol (CAP) is a broad-spectrum antibiotic. It was concluded that human exposure to CAP can cause aplastic anemia [1]. Chloramphenicol and other bacterial inhibitors have arguably been the biggest issue facing international seafood trade over the past year. Because chloramphenicol has displayed significant toxicological effects on humans, it has been banned from foods in the European community, Japan, and the United States at levels of 0.3 ppb. LC/MS has been demonstrated for this analysis by the U.S. Food and Drug Administration[2-4] and others[5]. In this application, the Agilent G6410AA QQQ is used. This method employs negative ion mode with electrospray ionization. An internal standard (IS), CAP-d5, is added at the beginning of the extraction. The use of this IS self-corrects for any extraction variability from sample to sample and response variability caused by the matrix. With the use of this IS, 50 parts per trillion (ppt) CAP levels can be reliably quantified. A solid phase extraction (SPE) procedure is used along with a mobile phase of only methanol and water without salt buffers, which should help minimize MS maintenance.

Abstract
This method was developed using the Agilent G6410AA Triple Quadrupole Mass Spectrometer (QQQ) for chloramphenicol in honey, shrimp, and chicken. The sensitivity obtained exceeds the minimum required performance level (MRPL) established by the European Union regulation for food monitoring programs. Using a deuterated internal standard and one simple sample solid phase extraction (SPE) procedure can provide a limit of detection at 10 ppt in sample matrix. The analytical performance of the method was evaluated for three different matrixes and the results show little or no matrix effects. Linearity of response over 2 orders of magnitude was demonstrated (r > 0.99). In addition, good reproducibility of the two required product ion ratios was obtained to

Experimental
Reagents and Materials Agilent AccuBond SPE ENV PS DVB Cartridges (P/N 188-3060) Ethyl acetate from Burdick and Jackson (Morristown, NJ) Methanol HPLC-Grade from Burdick and Jackson Water (18 MW) from Milli-Q Synthesis System Chloramphenicol (CAP) from Aldrich Chemical Co. (Milwaulkee, WI) Deuterated (d5) CAP internal standard from Cambridge Isotope Laboratories (CIL, Andover, MA, U.S.) Syringe filter (0.2 m, PTFE) from Agilent (P/N 5185-5843) Overview of Method Internal Standard Preparation 1. A 100-g/mL (100 ppm) stock standard CAP-d5 solution in methanol (MeOH) is purchased from Cambridge Isotope Laboratories, Inc. (Lot SCCE-005) 2. A 1:100 dilution in MeOH of the stock standard gives an intermediate standard concentration of 1 g/mL (1 ppm) or 1000 ng/mL CAP-d5 3. A 1:100 dilution in MeOH gives a diluent solution (This diluent solution is used to prepare the samples) concentration of 10 ppb. 4. Every 1-g sample is fortified with 25 L of CAP-d5 diluent solution for a 0.25 ppb IS (internal standard) concentration Standard Solution Preparation 1. A 100-g/mL stock standard CAP solution in methanol (MeOH) is prepared by weighing 5.0 mg CAP std into 50 mL methanol. 2. A 1:100 dilution with methanol of the stock standard gives an intermediate standard concentration of 1 g/mL (1 ppm) or 1000 ng/mL CAP 3. Add 25 L CAP-d5 diluent solution into each vial. 4. Prepare standard solutions in these vials: 1 ppb, 0.2 ppb, 0.1 ppb, 0.02 ppb, and 0.01 ppb, with IS at 0.25 ppb level.

Sample Preparation All SPE cartridges are conditioned with 2 mL of water before use. 1. Honey, 1 g of sample is diluted to 5 mL with water and 25 L 10 ppb IS is added. The solution is loaded onto the SPE cartridge and allowed to stand for 5 min. Elution is performed with 10 mL ethyl acetate. The eluate is collected and the solvent is evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL methanol and put in an ultrasonic bath for 1 min. The solution is filtered, using a syringe filter, before injection. No additional clean-up of the sample solution is performed. 2. Shrimp, 1 g of shrimp is defrosted and mixed in a blender. To the 1 g of the mixed shrimp, 3 mL of water and 25 L 10 ppb IS is added. The portion is centrifuged for 5 min (8,000 rpm). The supernatant is loaded on the cartridge and allowed to stand for 5 min. Elution is performed with 5 mL ethyl acetate. The eluate is collected and the solvent evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL methanol and put in an ultrasonic bath for 1 min; the solution is filtered before injection. 3. Chicken, 1 g of chicken is defrosted and mixed in a blender. To the 1 g of the mixed chicken, 3 mL of water and 25 L 10 ppb IS is added. The portion is centrifuged for 5 min (8,000 rpm). The supernatant is loaded on the cartridge and allowed to stand for 5 min. Elution is performed with 5 mL ethyl acetate. The eluate is collected and the solvent evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL methanol and put in an ultrasonic bath for 1 min.; the solution is filtered before injection. LC/MS conditions The LC system was the Agilent 1200-SL binary pump with the ALS-SL autosampler. The MS was an Agilent 6410 LC/MS triple quadrupole mass spectrometer. See Table 1 for conditions.

Table1. HPLC Column

LC/MS Conditions ZORBAX SB-C18, 2.1 50 mm, 1.8 m (p/n 827700-902) 0.4 mL/min A: water B: methanol 0-5 min, 30~70% B 5-6 min, 70~100% B 8 min, 100% B 4 min 45 C 5 L ESI Negative 350 C 10 L/min 45 psi 3500 V 100 V 10 V for m/z 257(qualifier ion) 15 V for m/z 152 (quantitation ion)

Table 2.

Structure and Fragment Ions of CAP and CAP-d5 (* indicates deuterated positions for the CAP-d5 IS)

Chloramphenicol structure
OH

Flow rate Mobile phase Gradient

*
O2N

* * *
m/z 257 152 262 157

H N O

CI CI

OH

Post time Temperature Injection MS Source Settings Source Ion polarity Drying gas temperature Drying gas flow rate Nebulizer Vcap Fragmentor Collision energy

CAP CAP-d5

Qualifier ion Quantitiation ion

Optimization of MS Condition Figure 1 shows the results of varying the Vcap. For this analyte there was little effect from varying this parameter. Only a slight increase in signal is observed at 3,500 V, and this voltage was used. The fragmentor was varied from 90 V to 160 V. Above 120 V, fragment ions are observed and the precursor ion signal drops significantly. At 160 V on the fragmentor almost no m/z 321 is observed. This results show that 100 V on the fragmentor provided the highest precursor ion signal. Finally, using a product ion scan of the precursor, m/z 321, the collision energy (CE) was varied from 2 V, 5 V, 8 V, 10 V, 15 V, 18 V to 40 V.
Response 4.00E+06 3.00E+06 2.00E+06 1.00E+06 0.00E+00 4500 4000 3500 3000 2500 Vcap

Results and Discussion


Spectral Quality and Sensitivity of Standard Table 2 lists the structure of the CAP and the fragment ions used for quantitation and confirmation as described by the identification point system.[6] To obtain the most sensitivity, only two or three parameters need to be optimized on this instrument. They are the fragmentor, to provide highest transmission of the precursor ion, the collision energy, to maximize signal for the quantitation and qualifier ion, and possibly the Vcap (electrospray voltage), to maximize the number of ions generated.

Figure 1.

Plot of Vcap voltage vs. response of precursor ion at m/z 321.

Comparison of extracted ion chromatograms of the quantitation and qualifier ions showed that response maximized at 10 V for m/z 257 and at 15 V for m/z 152. The product ion spectra for these two collision energy experiments are shown in Figure 2 and Figure 3. As shown in Table 3, the same CE were used for the deuterated internal standard.
104 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 100 120 140 160 180 200 220 240 260 Abundance vs. Mass-to-Charge (m/z) 280 300 320 340

321.0 152.1

257.1

194.0

176.0 249.0 127.0 164.3 206.9 219.1 237.0

Figure 2.
x104 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Product ion spectrum of m/z 321 at 10 V collision energy.


152.1

176.0

194.1 121.0
120

257.1 207.1 237.2 248.7


220 240 260 280 300

105.8
100

136.2
140

158.9 160 180 200

320.8
320 340

Abundance vs. Mass-to-Charge (m/z)

Figure 3.

Product ion spectrum of m/z 321 at 15 V collision energy.

Table 3.

MRM Mode Parameters Transition 321257 321152 Dwell time (ms) 200 200 200 200 Fragmentor Voltage (V) 100 100 100 100 Collision) Energy (V) 10 15 10 15 MS2 resolution Unit Unit Unit Unit

Compound CAP

CAP-d5

326262 326157

Repeatability Using honey matrix spiked at 0.1 ppb level as an example, the repeatability was tested by running the extract 15 times. Table 4 shows the area of the qualifier and quantitation ions in both the analyte and the IS. On average the areas of each ion vary about 8% and the ratios 5%, well within the 20% required for ratios 50% and above. Masshunter quantitation software tabulates these results and gives a graphic representation as shown in Figure 4.

Table 4.

Integrated Areas of the Quantitation Ion and Qualifier Ion and Their Associated Internal Standard Ion Chloramphenicol Quantitative ion (321152) 350 346 346 313 301 313 320 326 317 290 300 281 303 290 261 8.11% Qualifier ion (321257) 165 157 5 164 154 168 160 145 141 135 138 136 143 140 131 8.30% Ratio 47.1 45.2 44.6 52.3 49.5 53.6 50.1 44.5 44.5 46.6 46.2 48.4 47.3 48.3 50.3 5.91% d5-chloramphenicol Quantitative ion (326157) 262 258 259 267 261 253 228 225 241 226 253 240 220 214 217 7.67% Qualifier ion (326262) 121 114 118 127 121 124 111 113 117 107 90 90 101 107 101 9.99% Ratio 50.4 55.3 49.4 47.6 46.4 49.0 48.6 50.4 48.6 47.1 45.7 47.6 45.9 49.8 46.6 4.83%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 RSD

- MRM (321.0 -> 152.0) x102 Abundance 5 4 3 2 1 0 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8

321 -> 152.0 , 257.0

1.849
Abundance

x102 5 4 3

Ratio = 46.8

2 1 0 1 1.2 1.4 1.6 1.8 2 2.2

2.4

2.6

2.8

Acquisition Time (min) - MRM (326.0 -> 157.0) x101 Abundance 7 6 5 4 3 2 1 0 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 326 -> 157.0 , 262.0

Acquisition Time (min)

1.818
Abundance

x101 7 6 5 4

Ratio = 50.3

3 2 1 0 1 1.2 1.4 1.6 1.8 2

2.2

2.4

2.6

2.8

Acquisition Time (min)

Acquisition Time (min)

Figure 4.

Panels A and B show the CAP and IS peak for the quantitation transition. Panels C and D are the graphic representation of quantitation ion and qualifier ion ratio as shown by MassHunter software.

Linearity The linearity of the method was determined for CAP in solvent and each of the matrices. This was done from 10 ppt to 1 ppb, well below the minimum required performance level (MRPL) and above that concentration. Figures 5 through 8 show the graphic representation of those results. Each was well above an r2 value of 0.99.
Relative Responses 8 7 6 5 4 3 2 1 0 -0.2 0.2 0.6 1 1.4 1.8 2.2 2.6 3 3.4 3.8 4.2 Relative Concentration

Relative Responses 6 5 4 3 2 1 0 -0.2 0.2 0.6 1 1.4 1.8 2.2 2.6 3 3.4 3.8 4.2 Relative Concentration

y = 1.9565 * x + 0.0669 R^2 = 0.99967712

y = 1.3871 * x + 0.3325 R^2 = 0.99854947

Figure 5. 6

Linearity of CAP in solvent from 10 ppt to 1 ppb.

Figure 6.

Linearity of CAP in honey from 10 ppt to 1 ppb.

Relative Responses 8 7 6 5 4 3 2 1 0 -0.2 0.2

Relative Responses 7 6 5 4 3 2 1 0

y = 1.7891 * x + 0.2085 R^2 = 0.99890473

y = 1.7027 * x + 0.0150 R^2 = 0.99985348

0.6

1.4

1.8

2.2

2.6

3 3.4 3.8 4.2 Relative Concentration

-0.2

0.2

0.6

1.4

1.8

2.2

2.6

3 3.4 3.8 4.2 Relative Concentration

Figure 7.

Linearity of CAP in shrimp from 10 ppt to 1 ppb.

Figure 8.

Linearity of CAP in chicken from 10 ppt to 1 ppb.

Sensitivity The sensitivity of CAP standard in solvent is observed at 10 ppt with an injection volume of 5 L. The MRM chromatogram is shown in Figure 9. Although this demonstrates the sensitivity of the instrument, it is also important to determine the sensitivity in real sample matrix. This is shown in Figure 10 with a spike concentration of CAP at 10 ppt with a 5-L injection. Not only is the analyte detectable, but the ratio of the qualifier ion is within the specified tolerance so confirmation can be obtained.
x10 2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
- TIC MRM (****) CAP 10 ppt

Recovery Recovery was determined by spiking CAP into three samples of matrix and extracting using the specified SPE. Table 5 shows both the repeatability of extraction and analysis and the mean recovery. Using the internal standard spiked before extraction, recovery is automatically compensated. Thus accuracy of the quantification is very good using this methodology. The recovery results show the overall effectiveness of the method.

1.724

2.6

2.8

3.2

3.4

3.6

3.8

Abundance vs. Acquisition Time (min)

Figure 9. Table 5.

MRM chromatogram of 10 ppt CAP in solvent with injection volume of 5 L.

Recovery of CAP at 0.1 ppb Where Three Sample Aliquots of Each Matrix Were Spiked and Determined Honey Shrimp Chicken (n=3) (n=3) (n=3) RSD (%) 6.29 3.93 3.29 Recovery (%) 89.5 85.4 86.4

Conclusions
The method described herein for the analysis of CAP in three important matrices has been shown to be highly effective and meet the criteria for

quantitation and confirmation well below the required 0.3 ppb MRPL. Optimization of the method was simple, as few parameters in the mass spectrometer need adjustment. In addition, the requirements for a validated method have been shown. These include sensitivity, repeatability, linearity, and recovery. The Agilent 6410 LC/MS triple quadrupole instrument has been shown to be a highly effective instrument for the analysis of chloramphenicol.

www.agilent.com/chem
10 1 1.0 0.9 0.8 0.7 Abundance Abundance 0.6
- MRM (321 152.0) CAP in Honey.d

10 1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 _ 0.1

321 152.0, 257.0

1.751

Ratio=57.8

0.5 0.4

0.3 0.2 0.1 0.0 _ 0.1 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 Acquisition Time (min)

1.2

1.4 1.6 1.8 2 2.2 2.4 2.6 Acquisition Time (min)

Figure 10. MRM chromatogram of quantitation ion and ratio of qualifier ion for 10 ppt CAP in honey.

References
1. Chemical Safety Information from Intergovernmental Organisations (IPCS-INCHEM), Web page http://www.inchem.org/documents/ jecfa/jecmono/v33je03.htm 2. S. Turnipseed, et al. (2002) Confirmation of Multiple Phenicol Residues in Honey by Electrospray LC/MS, Laboratory Information Bulletin (4281) U.S. Food and Drug Administration. 3. A. Pfenning, et al. (2002) Confirmation of Multiple Phenicol Residues in Shrimp by Electrospray LC/MS, Laboratory Information Bulletin (4284) Food and Drug Administration. 4. B. K. Neuhaus, et al. (2002) LC/MS/MS Analysis of Chloramphenicol in Shrimp, Laboratory Information Bulletin (4290) Food & Drug Administration 5. P. Mottier, V. Parisod, E. Gremaud, P. A. Guy, and R. H. Stadler. Determination of the antibiotic chloramphenicol in meat and seafood products by liquid chromatography - electrospray ionization tandem mass spectrometry. Journal of Chromatography A 2003, 994, (1-2), 75-84. 6. Commission, E., 2002/657/EC: Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results (Text with EEA relevance) (notified under document number C(2002) (3044) 2002.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

The information contained in this publication is intended for research use only and is not to be followed as a diagnostic procedure. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA August 21, 2007 5989-5975EN

Analysis of Nitrofuran Metabolites in Tilapia Using Agilent 6410 Triple Quadrupole Application

Food Safety

Authors
Feng Liang, Peibin Hu, and Ping Li Agilent Technologies, Inc. Beijing, China

analytical detection of the metabolites and not the parent compounds are required in samples of animal origin. The criteria for detection and confirmation of veterinary drugs in animal and animal products established by the European Union (EU) [2] has been accepted in much of the world. This criteria mandates a separation technique combined with a spectrometric technique. For banned substances such as the nitrofurans, no maximum residue limit (MRL) could be set. Therefore a minimum required performance level (MRPL) was set at 1 g/kg for each metabolite [3]. Only LC/MS could meet these criteria, and very good methods have been reported [46]. However, the most widely accepted methodology employs triple quadrupole tandem mass spectrometers. This is the first report showing analysis of these metabolites using the new Agilent triple quadrupole LC/MS system.

Abstract
The metabolites of nitrofuran antibiotics banned in meat and meat products are analyzed by LC/MS/MS with the new Agilent 6410 triple quadrupole. The method is shown to be highly sensitive, to 0.01 ppb (10 ppt), for each of the four analytes. Calibration from 0.1 ppb to 10 ppb is presented with all criteria for confirmation as set by the European Union decisions for analytical method performance. Extracts of tilapia are used to show the performance of the LC/MS/MS method for aquaculture samples.

Introduction
Nitrofurans are inexpensive antibiotics used for Gram positive and Gram negative bacteria. They have been used to treat gastrointestinal and dermatological infections in farm animals and fish. In addition, they have been used to treat bacteria in bees. Because both parent compounds and their metabolites are suspect carcinogens, they have been banned around the world. The Rapid Alert System for Food and Feed Annual report for 2005 [1] shows that these compounds continue to be detected in food samples and remains a major concern for food safety. The four compoundsfurazolidone, furaltadone, nitrofurantoin, and nitrofurazonehave been found to metabolize rapidly, and the metabolites bind to muscle tissue. Thus the

Experimental
Chemicals Derivatized standards of nitrofuran metabolites and all chemicals for sample preparation were received from a food manufacturing company. Acetonitrile was HPLC grade from Merck (Darmstadt, Germany). Formic acid was reagent grade from Merck (Darmstadt, Germany). Sample Preparation The accepted procedure for sample preparation was followed. To 2 g of tilapia was added 15 mL 0.125 M HCl and the mixture homogenized.

To this solution, a 50-L solution of 2-nitrobenzaldehyde (50 mM in DMSO) was added and shaken. The solution was then incubated at 37 C for 16 hours. This was followed by neutralization to pH ~7 with NaOH and K2HPO4. The neutral derivatized sample was then extracted with ethyl acetate, concentrated to dryness, and reconstituted in 100 L of initial LC mobile phase. Standards of the four metabolites were spiked into 0.125 M HCl, derivatized, and extracted for calibration using the same procedure as was used for the samples of tilapia. LC/MS/MS Method
LC Conditions Instrument: Column: Column temp.: Mobile phase: Gradient: Agilent 1100 LC C18, 2.1 mm 150 mm, 3 m 40 C A = 0.1% formic acid in water B = acetonitrile 22% B at 0 min 99% B at 6 min 99% B at 9 min 0.3 mL/min 50 nL

Quantitation Quantitative analysis was done with the first transition listed in the MRM parameter table. The second transition was used as a qualifier ion for confirmation as per the confirmation criteria. Quantitative results were performed with the new MassHunter quantitative analysis software.

Results and Discussion


The instrument sensitivity is an important performance parameter for this analysis when considering the derivatization and extraction needed to meet the required detection limit of 1 ppb for each metabolite, aminohydantoin (AH), 3-amino-5-morpholinomethyl-2-oxazolidinone (AMOZ), 3-amino2-oxazolidinone (AOZ), and semicarbazide (SC). To demonstrate this performance, a standard of each 2-nitrobenzaldehyde (2-NBA) derivatized metabolite is shown at 0.01 ppb (10 ppt) in Figure 1. The structure of each derivatized metabolite is given and each is shown with a signal-to-noise ratio of greater than 3:1. Another indicator of performance is linearity. Calibration curves from this concentration (10 ppt) to 10 ppb are displayed in Figure 2 showing the linearity for each compound. Treatment of fish with nitrofurans is a continual problem for food safety and import into EU member countries. To demonstrate the capability of the Agilent triple quadrupole LC/MS, tilapia samples were spiked with the four metabolites, hydrolyzed, derivatized, and extracted. An analysis of a tilapia extract at 500 ppt is shown in Figure 3 and demonstrates the signal obtained at half the MRPL. In addition to meeting the sensitiv-

Flow rate: Injection volume: MS Conditions Instrument: Ionization mode: Drying gas flow: Nebulizer: Drying gas temp.: Vcap:

Agilent 6410 LC/MS Triple Quadrupole Positive ESI 10 L/min 35 psig 350 C 5000 V

MRM Mode Parameters Dwell time (ms) 60 60 60 60 60 60 60 60 Fragmentor voltage (V) 100 100 100 100 100 100 100 100 Collision energy (V) 5 5 5 5 5 5 5 5 MS2 resolution Unit Unit Unit Unit Unit Unit Unit Unit

Compound AMOZ

Transition 335.1 & 291.4 335.1 & 262.4

SC

209.1 & 192.3 209.1 & 166.3

AH

249.1 & 134.2 249.1 & 104.2

AOZ

236.0 & 134.1 236.0 & 104.1

ity requirement, the analysis must also meet the confirmation criteria, including both chromatographic retention time match with the standards and measuring a qualifying ion with a relative intensity ratio within a specified tolerance of the quantitation ion. This tolerance is set by the ratio obtained when analyzing standards and increasing as that ratio decreases. This tolerance ranges from 20% for ions with relative ratio intensities above 0.5 and to 50% for ratios below 0.1. Table 1 shows tilapia samples spiked with the metabolites, derivatized and extracted. The spikes

were used as the calibrants, so the final concentration is obtained from the curve. The table is produced as the batch using the MassHunter software results with outliers highlighted in blue (low) and red (high). The table shows that in the blank a peak is found within the tolerance set for the retention time of AMOZ but the qualifier ratio is low. For AOZ and SC, retention times for suspect peaks are below the specified retention in the same blank. For AH, the 0.5 ppb spike, the qualifier ion ratio is outside the 35% tolerance limit set for this ion (again low).

x10 2 0.9 0.7 0.5

+ MRM (335.1 291.4) 10ppt_1.d Smooth (1)

12

23

34 O N O O N
+ _

O O O N N N+ O
_

x10 1 + MRM (209.1 192.3) 10ppt_1.d Smooth (1) 7 6 5 x10 1 6.0 + MRM (249.1 134.2) 10ppt_1.d Smooth (1) 5.6 5.2 4.8

12

23

34

N NH

O NH2

12

23

34

O O N
+

N N O O
_

x10 + MRM (236.0 134.3) 10ppt_1.d Smooth (1) 1.2 1.0 D 0.8 0.6
2

12

23

34

N N

O NH

O 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2

Abundance vs. acquisition time (min)

Figure 1.

The MRM quant ion chromatogram for each derivatized metabole at 10 ppt of A) 2-NBA AMOZ, B) 2-NBA SC, C) 2-NBA AH, and D) 2-NBA-AOZ

x105 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

y = 49247.4270 * - 1850.8024 R 2 = 0.99989213

Response

R 2 > 0.999

Response

AMOZ

x105 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 _0.1

y = 11978.9741 * - 1213.1580 R 2 = 0.99852795

SEM
R 2 > 0.998

0 x104 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 _0.5

3 4 5 6 7 Concentration (ng/mL)

10

3 4 5 6 7 Concentration (ng/mL)

10

y = 5697.5004 * - 565.5495 R 2 = 0.99879989

Response

R 2 > 0.998

Response

AHD

x105 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

y = 45077.4032 * + 4420.7065 R 2 = 0.99523855

AOZ
R 2 > 0.995

3 4 5 6 7 Concentration (ng/mL)

10

3 4 5 6 7 Concentration (ng/mL)

10

Figure 2.

Calibration curves of nitrofuran metabolites linear range from 10 ppt to 10 ppb.

x10 3

+ TIC MRM (** **) 0619_500ppt_spike_MRM_50uL_4seg_4.d Smooth (1)

12

23

34

AOZ AMOZ
1.0

SC AH

Abundance vs. acquisition time (min)

Figure 3.

Spiked tilapia sample extract at 500 ppt each metabolite.

Table 1.

Analysis of Talapia Spikes Self-Calibrated. Note Qualifier Ratios and Retention Times Reported.

AH
Name Nitrofuran Blank Nitrofuran 0.5 ppb Nitrofuran 1 ppb Nitrofuran 3 ppb Nitrofuran 5 ppb RT 8.66 8.66 8.67 8.66 Final Conc. 0.00 0.54 1.07 3.42 5.71

Qualifier
Ratio 6.07 15.83 13.78 14.24 RT 2.72 2.62 2.64 2.65 2.66

AMOZ
Final Conc. 0.02 0.51 0.92 3.41 4.66

Qualifier
Ratio 8.70 17.44 20.42 18.67 19.28 RT 8.71 9.19 9.19 9.20 9.19

AOZ
Final Conc. 0.00 0.37 1.06 3.18 4.89

Qualifier
Ratio 7.99 10.19 9.51 8.61 RT 7.94 8.44 8.44 8.44 8.44

SC
Final Conc. 0.00 0.50 1.05 2.89 5.05

Qualifier
Ratio 65.32 70.85 67.47 69.06

Conclusions
This work shows the high performance of the new Agilent 6410 LC/MS triple quadrupole system for the sensitive analysis of the nitrofuran metabolites in fish samples. The system readily meets the performance requirements and provides advanced quantitation software for calculating and reporting all confirmation parameters specified by the European Commission decision.

4. B. Wst, C. Sauber, and H. J. A. van Rhijn, Quantitation of Nitrofuran Metabolites in Shrimp and Poultry by LC/MS/MS Using the Agilent LC/MSD Trap XCT, Agilent Technologies publication, 5989-0738EN: March 25, 2004. 5. M. Takino, Determination of the Metabolites of Antibacterial Drugs in Chicken Tissue by Liquid Chromatography Electrospray Ion ization Mass Spectrometry (LC-ESI-MS), Agilent Technologies publication, 5988-8903EN: March 19, 2003. 6. F. Mandel, B. Wst, and C. Sauber, High Resolution Quantitative Analysis of Nitrofuran Derivatives in Poultry and Shrimp Using a New oa-ESI-TOF, Agilent Technologies publication, 5989-1302EN: July 9, 2004.

References
1. The Rapid Alert System for Food and Feed (RASFF) Annual Report 2005, p 29, http://ec.europa.eu/food/food/rapidalert/ index_en.htm. 2. E. Commission, 2002/657/EC: Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results (Text with EEA relevance) (notified under document number C [2002] 3044) 2002. 3. E. Commission, 2003/181/EC: Commission Decision of 13 March 2003 amending Decision 2002/657/EC as regards the setting of minimum required performance limits (MRPLs) for certain residues in food of animal origin (Text with EEA relevance) (notified under document number C[2003] 764) 2003.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA October 31, 2006 5989-5808EN

Determining Malachite Green and Leucomalachite Green in Food by LC/MS/MS Application

Food Safety

Authors
Feng Liang, Peibin Hu, and Ping Li Agilent Technologies Beijing, China

have already banned MG in fishery. Due to its low cost and antifungal effectiveness, MG is still being used illegally as indicated in the European Rapid Alert System for Food and Feed.2 HPLC with UV detection has been used to analyze MG and LMG. Figure 1 shows the structure of the two compounds. Loss of conjugation by reduction changes the chromaphore of LGM significantly. To obtain the sum of both, the method employs postcolumn oxidation with PbO2 to convert LMG to MG, thus providing a sum of both comounds.3 Most recently, LC/MS has been used to both meet the EU confirmation criteria and provide quantitative results for both compounds without the need for post-column oxidation. In this application, a simple and sensitive method for simultaneously determining MG and LMG is presented.4, 5 The LC/MS/MS methods LOQ is 0.01 g/Kg, which easily meets the import requirement set by Japan or the EU.6

Abstract
This application note demonstrates a complete method to rapidly and precisely determine residue levels of malachite green and leucomalachite green in fish with the new Agilent 6410 LC/MS triple quadrupole system. Using positive mode electrospray ionization (ESI+) and multiple reaction monitoring (MRM), qualification and quantification were accomplished without the traditional tedious PbO2 oxidation process. The LC/MS/MS methods LOQ is 0.01 g/Kg, which easily meets the import requirement of 2 g/Kg set by Japan and the EU.

Introduction
Malachite green (MG) is a metallic-looking crystal. It dissolves in water easily as a blue-green solution. It is a toxic chemical primarily used as a dye and has been found very effective in treating parasites, fungal infections, and bacterial infections in fish and fish eggs.1 On uptake, MG is rapidly reduced into leucomalachite green (LMG) and deposited in the fatty tissue of the fish with little MG remaining. MG can cause significant health risk for humans who eat contaminated fish. For example, it can cause liver tumor formation and is suspected of carcinogenesis.1 The United States, Japan, China, the European Union, and many other countries

Experimental
Reagents MG LMG Sigma-Aldrich, CAS 569-64-2, USA Dr. Ehreastorfer's lab, D-86199, 99% pure, Augsburg, Germany CAS 75-05-8; Burdick & Jackson; Morristown, New Jersey, USA Merck, Germany CAS 631-61-8, Acros Organics, Morris Plains, New Jersey, USA

Acetonitrile

Acetic acid Ammonium acetate

H C N C N

N +

Cl

Malachite green
Figure 1.

Leucomalachite green

Molecular structure of malachite green and leucomalachite green.

Calibration Solutions A stock standard solution of MG and LMG in acetonitrile was prepared at 100 g/mL and stored at %18 oC, avoiding light. The stock solution was diluted in 50:50 acetonitrile:water to make the calibration solutions+10, 50, 100, 500, 1000, 5000, and 10,000 fg/L. Sample Preparation To 5 g tilapia tissue was added 1 mL (0.25 mg/mL) hydroxylamine, 2 mL 1 M toluene sulfonic acid, 2 mL of 0.1 M ammonium acetate buffer (pH 4.5), and 40 mL acetonitrile. The mixture was then homogenized for 2 min. The supernatant was decanted, and to the precipitate was added 20 mL acetonitrile. This was filtered and added to the supernatant. To the combined acetonitrile extracts, 35 mL water and 30 mL methylene chloride were added. The solution was shaken and the methylene chloride layer collected. A second extract of 20 mL methylene chloride was made, and this layer added to the first extract. The methylene chloride was taken to dryness with a gentle stream of nitrogen and the extract reconstituted in 100 L of acetonitrile

Instrumentation
LC Column Column temp. Mobile phase 1100 LC C18, 2.1 x 150 mm, 5 m 40 oC A % 10 mmol/L ammonium acetate (adjust to pH 4.5 with acetic acid) B % acetonitrile 0.3 mL/min Time %B 0 30 1 50 2 95 8 95 8.01 30 13 30 10 L

Column flow Gradient

Injection vol. MS

Agilent 6410 LC/MS Triple Quadrupole Ionization ESI(+) Capillary 4000 V Nebulizer P. 35 psi Drying gas 11 L/min Gas temp. 350 oC Skimmer 15 V OctDc1 (Skim2) 45 V Oct RF 500 V Q1 resolution Unit Q3 resolution Unit Collision gas Nitrogen The MRM parameters are listed in Table 1.

Table 1. Time 0 7

MRM Method Parameters Compound MG LMG Precursor 329.3 329.3 331.3 331.3 Product 313.3 208.2 316.3 239.2 Dwell (ms) 40 40 40 40 Fragmentor (V) 100 100 100 100 Collision Energy (V) 40 40 30 30

Results and Discussion


To obtain the most sensitive results, optimization of certain fragmentor voltages is important. Figure 2 shows the EICs of both target compounds at fragmentor values of 70 V, 90 V, and 100 V. The results show that the three different fragmentor values have little effect on the intensity of [M+H]+ ions. Thus, 100 V was chosen for this study. In addition, an optimal collision energy for the MS/MS must be set. Figure 3 shows the MS/MS spectra from three different collisional voltages,

(a) 20 V, (b) 30 V, and (c) 40 V. Due to their structural differences, the voltage required for optimum fragmentation of each compound is different. For MG, the optimum fragmentation was observed at 40 V. The ion m/z 313 was due to the neutral loss of methane. The ion at m/z 208 was due to the neutral loss of N,N-dimethylaniline. For LMG, the optimum fragmentation was observed at 30 V. The ion at m/z 316 was due to the loss of a methyl radical. The ion at m/z 239 resulted from a subsequent loss of a benzene radical or, more likely, the rearrangement and neutral loss of toluene.

x107 + EIC(329.4, 331.4 m/z) Scan optimizing FRG70_3.d 5 3 1


+ EIC(329.4, 331.4 m/z) Scan optimizing FRG90_4.d

70 V

x107 5 3 1 x107 5 3 1

90 V

+ EIC(329.4, 331.4 m/z) Scan optimizing FRG100_5.d

100 V
0 1 2 3 4 5 6 7 8 9 10 11 12

Abundance vs. acquisition time (min)

Figure 2.

EICs of malachite green and leucomalachite green at fragmentor values of 70 V, 90 V, and 100 V.

329.3 x105 1.2 1.0 0.8 0.6 0.4 0.2 0.0 193.1 208.3 236.9 268.4 284.3 313.4 331.8
+ Product Ion (5.499-5.633 min, 17 scans) (329.3, 331.4 **) optimizing MS2_FRG100_CE20_2.d

x105 + Product Ion (8.349-8.466 min, 15 scans) (331.4, 329.3 **) optimizing MS2_FRG100_CE20_2.d 239.8 2.0 1.6 1.2 0.8 0.4 0.0 40 60 80 100 120.8 134.5 120 140 165.6 160 195.8 180 209.8 223.9 200 220 240 272.8 286.6 301.8 260 280 300 316.7

320

340

Abundance vs. mass-to-charge (m/z)

Figure 3a. MS/MS spectra of MG and LMG at collisional voltage of 20 V.


x104 7 6 5 4 3 2 1 0 134.3 165.1 192.8 208.2 217.4 237.2 251.4 270.3 285.3298.9 313.4
+ Product Ion (5.457-5.591 min, 17 scans) (331.4, 329.3 **) optimizing MS2_FRG100_CE30_3.d

329.3

+ Product Ion (8.349-8.457 min, 14 scans) (331.4, 329.3 **) optimizing MS2_FRG100_CE30_3.d

239.8

x105 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 40 60 80 100 120

315.8 134.5 140 209.8 223.9 165.8 180.6 194.7 160 180 200 220 240 Abundance vs. mass-to-charge (m/z) 260 272.7 286.8 280 301.9 300 320 331.8 340

Figure 3b. MS/MS spectra of MG and LMG at collisional voltage of 30 V.

x104 4.0 3.0 2.0 1.0 0.0

+ Product Ion (5.474-5.591 min, 15 scans) (331.4, 329.3 **) optimizing MS2_FRG100_CE40_4.d

313.3

208.2 165.3 241.4 284.2 299.4 329.4

91.5

117.9 134.1

192.1

221.0

270.3

+ Product Ion (8.340-8.499 min, 20 scans) (329.3, 331.4 **) optimizing MS2_FRG100_CE40_4.d

x105 2.5 2.0 1.5 1.0 0.5 0.0 40 60 80 91.6 100 120.6 134.5 120 140 165.8 180.7 194.7 208.7 223.8

239.8

315.8 257.9 272.7 286.7 260 280 300 320 340

160 180 200 220 240 Abundance vs. mass-to-charge (m/z)

Figure 3c MS/MS spectra of MG and LMG at collisional voltage of 40 V.

Figure 4 shows the calibration curves for both MG (4a) and LMG (4b). Calibration solution concentrations were from 10 to 10,000 fg/L. The linear calibration range is 100 to 100,000 fg on column for both compounds. The R2 for both compounds was > 0.999 (origin ignored and no weighting). To demonstrate the sensitivity of the instrument,
x105 2.6 2.4 2.2 2.0 1.8 1.6 Response 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 5 6 7 Concentration (ng/mL)

Figure 5 shows MS/MS spectra of a blank sample extract (5a) and sample extract spiked with 10 ppt of each compound (5b). A sample of tilapia spiked at 100 ppt MG and LMG before extraction was made to demonstrate method performance. The MRM results after extraction and cleanup are shown in Figure 6. The recover-

Malachite green - 7 levels, 7 levels used, 14 points, 14 points used, 0 QCs y = 23363.3374 * - 1766.9951 R 2 = 0.99946103

R 2 = 0.9995

10

11

12

Figure 4a. Calibration curve of malachite green, linear range: 10 ppt to 10 ppb.

x10 6 1.0 0.9 0.8 0.7 Response 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 1 2 3 4 5 6 7 Concentration (ng/mL) 8 9 10 11 12
Leucomalachite green - 7 Levels, 7 Levels Used, 14 Points, 14 Points Used, 0 QCs y = 93199.4712 * - 7543.3588 R 2 = 0.99942595

R 2 = 0.9994

Figure 4b. Calibration curve of leucomalachite green, linear range: 10 ppt to 10 ppb.
8.433 x101 2.8 2.4 2.0 1.6 1.2 0.8 0.4 0.0 0 1 2 3 4 5 6 7 Abundance vs. acquisition time (min) 8 9 10 11 12
+ MRM MRM (331.3 239.2) malachite green_200606121.d

12

Figure 5a. MG and LMG MRM of a blank sample.


x102 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 5 6 7 Abundance vs. acquisition time (min) 8 9 10 11 12 5.440

+ MRM MRM (329.3 313.3) malac hite green_200606121.d

12

ppt spiked sample.


Figure 5b. MG and LMG MRM of a 10-ppt spiked sample.

x102 3.2 2.8 2.4 2.0 1.6 1.2 0.8 0.4 0.0 0

+ MRM MRM (331.3 239.2) Spike_100 ppt_1.d

12

8.315

5 6 7 Abundance vs. acquisition time (min)

10

11

12

Figure 6. MRM result of talapia extract spiked with 100-ppt MG and LMG.

ies for MG were 48% and 23% for LMG. A mixture of MG and LMG at 100 fg/L in 50:50 acetonitrile: ammonium acetate was used for the repeatability study for instrument performance. The RSD from eight injections for MG was 3.52% (S/N > 20). The RSD from eight injections for LMG was 2.25% (S/N > 40).

4. M. D. Hernando, M. Mezcua, J. M. SuarezBarcena, and A. R. Fernandez-Alba, Liquid chromatography with time-of-flight mass spectrometry for simultaneous determination of chemotherapeutant residues in salmon. Analytica Chimica Acta 2006, 562, (2), 176%184. 5. K.-C. Lee, J.-L. Wu, and Z. Cai, Determination of malachite green and leucomalachite green in edible goldfish muscle by liquid chromatography-ion trap mass spectrometry. Journal of Chromatography B 2006, In Press, Corrected Proof. 6. 2004/25/EC: Commission Decision of 22 December 2003 amending Decision 2002/657/EC as regards the setting of minimum required performance limits (MRPLs) for certain residues in food of animal origin (Text with EEA relevance) (notified under document number C [2003] 4961). 2003.

Conclusions
This application note demonstrates a complete method to rapidly and precisely determine residue levels of malachite green and leuco-malachite green in fish. Using positive mode electrospray ionization (ESI+) and multiple reaction monitoring (MRM) technique, the LC/MS/MS method shows detection limit of 10 ppt, which easily meets the import requirement set by Japan or EU.

References
1. S. Srivastava, R. Sinha, and D. Roy, Toxicological effects of malachite green. Aquatic Toxicology 2004, 66, (3), 319%329. 2. The Rapid Alert System for Food and Feed (RASFF) Annual Report 2005. 2005, 29. 3. C. A. Hajee and N. Haagsma, Simultaneous determination of malachite green and its metabolite leucomalachite green in eel plasma using post-column oxidation. Journal of Chromatography B Biomed Appl. 1995, 669, (2), 219%227.

Acknowledgement
The authors would like to thank Dr. Yanqin Liu for providing the standard solutions and sample extracts.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA October 25, 2006 5989-5807EN

Quantitation of Nitrofuran Metabolites in Shrimp and Poultry by LC/MS/MS Using the Agilent LC/MSD Trap XCT Application

Food

Authors
Bernhard Wst, Christian Sauber Agilent Technologies GmbH, Hewlett-Packardstr. 8, 76337 Waldbronn Germany Hans (J.) A. van Rhijn RIKILT -Institute of Food Safety Bornsesteeg 45,6700 AE Wageningen the Netherlands

shrimp, and poultry. In 1995 the four drugs Furazolidone, Furaltadone, Nitrofurantoin, and Nitrofurazone were banned by the European Union (EU) for their usage in food-producing animals. Due to their rapid metabolism nitrofuran parent substances are not suitable for monitoring and typically their metabolites are analyzed. A liquid chromatography /mass spectrometry/mass spectrometry (LC/MS/MS) method was developed for the sensitive, qualitative, and quantitative analysis of four derivatized nitrofuran metabolites found in shrimp and poultry. See Figure 1.

Abstract
An LC/MS/MS method was developed for the qualitative and quantitative measurement of nitrofuran metabolites in chicken and shrimp using the Agilent LC/MSD XCT Ion Trap. The limit of quantitation for all four nitrofurans investigated easily met the specified European Union Minimum Required Performance Level of 1 g/kg and ranged from 0.125 g/kg to 0.5 g/kg.

Experimental
All liquid chromatography/mass spectrometry (LC/MS) experiments were performed using an Agilent 1100 Series LC system coupled to a mass selective detector (MSD) Ion Trap XCT mass spectrometer. The Ion Trap was operated with an orthogonal electrospray source in positive ion mode using multiple reaction monitoring (MRM). A gradient method was used for chromatography. Deuterated NBA-AMOZ was used as the internal standard (ISTD) for NBA-AMOZ and deuterated NBA-AOZ was used as the ISTD for the other metabolites.

Introduction
Nitrofuran antibiotics are widely used for the treatment of infectious diseases in cattle, pigs,

O O

O O O N N O H2N N O O N

O O N N O

N+

Furazolidone
O O
_

AOZ
O O N N

NBA-AOZ

N+

O N N O O N

H2N

N N

O N

O N

Furaltadone
_

AMOZ
O O N O

NBA-AMOZ
O H N O

O O

O O N N O NH H2N N O

N+

NH

Nitrofurantoin
O O
_

AHD
O O N O H N NH2

NBA-AHD

N+

O N

H N NH2

H2N

H N NH2

Nitrofuranzone

SEM

NBA-SEM

Drugs Figure 1.

Metabolites

Derivatized metabolites

Chemical structures of nitrofurans, nitrofuran metabolites, and corresponding derivatives.

Sample Preparation One (1) g of homogenized tissue from shrimp or poultry was mixed with 5 mL of a 0.2 M hydrochloric acid and 50 L of 2-nitrobenzaldehyde (2-NBA, 100 mM in methanol) and incubated overnight at 37 C. This is a protocol from the State Institute for Quality Control of Agricultural Products (RIKILT, The Netherlands). Using this protocol, tissue-bound residues of the nitrofurans with an intact side chain are released through acidic hydrolysis of the imine bond. The free amino groups of the corresponding metabolites are derivatized with 2-NBA to form an aromatic imine bond.

The sample was then neutralized with 500 L of a 0.3 M Na3PO4 solution in water and adjusted to pH7 with 2 M NaOH solution. After the addition of 4-mL ethyl acetate, the sample was centrifuged and the organic layer was transferred to a clean tube. The sample was further extracted with a 4-mL aliquot of ethyl acetate, centrifuged, and the organic layer added to the first extract. After evaporation to near dryness, the sample was reconstituted in 500-L solvent (50-mL acetonitrile, 80-mL water and 0.5-mL acetic acid) for subsequent LC/MS analysis. Calibration standards were made by spiking known concentrations of all four underivatized analytes into shrimp and poultry prior to sample preparation.

Chemicals
Methanol Hydrochloric acid, HCL 32% 2-Nitrobenzaldehyde (2-NBA), C7H5NO3 Tri-sodiumphosphate-dodecahydrate, Na3PO4 12(H2O) Sodium hydroxide, NaOH Ethyl acetate, CH3COOC2H5 Acetic acid, CH3COOH Acetonitrile, CH3CN Methanol-d, 99.5% D (Biosolve, 13683502) (Merck, 100319) (Sigma, N6001) (Merck, 106578) (Merck,106498) (Biosolve 05402602) (Merck 10063) (Biosolve, 01203502) (Aldrich, 15.193-9)

LC/MS/MS Method Details


HPLC: Flow rate: Column: Mobile phases: Agilent 1100 0.4 mL/min Zorbax XDB-C18, 2.1 mm 150 mm, 3.5 m A: Water + 0.1% acetic acid B: Acetonitrile + 0.1% acetic acid Gradient: 014 min: 10% A - 45% A; 1416 min: 45% A - 90% A 50 L out of 500 L

Injection: MS 1100 LCMSD XCT Ion Trap Ionization mode: Nebulizer pressure: Drying gas flow: Drying gas temperature: Skimmer: Capillary exit: Trap drive: ICC: MRM mode 4 Segments: Segment 1: Segment 2: Segment 3: Segment 4:

Positive ESI 45 psi 12 L/min 350 C 20 V 55 V 55 On

02.2 min, 2.27.4 min, 7.410.6 min, 10.613.9 min Divert valve: to waste, no spectra MS/MS of m/z 335, Isolation width: 2.0, Cut-off: 140; MS/MS of m/z 340, Isolation width: 2.0, Cut-off: 140; MS/MS of m/z 209, Isolation width: 2.0, Cut-off: 120; MS/MS of m/z 249, Isolation width: 2.0, Cut-off: 100; MS/MS of m/z 236, Isolation width: 2.0, Cut-off: 100; MS/MS of m/z 240, Isolation width: 2.0, Cut-off: 100;

Amplitude: 1.16 Amplitude: 1.16 Amplitude: 1.28 Amplitude: 1.25 Amplitude: 1.25 Amplitude: 1.25

Quantitation NBA-AMOZ: NBA-dAMOZ: NBA-SEM: NBA-AHD: NBA-AOZ: NBA-dAOZ: Maximum accumulation time: Smart target: Scan: EIC of 261 + 291 (MS/MS of 335), EIC of 266 + 296 (MS/MS of 340), EIC of 166 + 192 (MS/MS of 209), EIC of 134 (MS/MS of 249), EIC of 134 (MS/MS of 236), EIC of 134 (MS/MS of 240), 150 ms 100.000 100350 Ret. Time: 4.5 min Ret. Time: 4.5 min Ret. Time: 9.9 min Ret. Time: 10.0 min Ret. Time: 10.8 min Ret. Time: 10.8 min

Results and Discussion


Very low limits of detection (LOD) are required for nitrofuran metabolites and the derivatization method increased the ionization efficiency, as well as improving the chromatographic separation. A liquid-liquid extraction procedure was used which resulted in a relatively high concentration factor to further improve LOD. The ion trap mass spectrometer was operated in MRM mode. In this mode, only precursor ions are chosen and full-scan MS/MS-spectra of the corresponding analytes are acquired. These full scanMS/MS spectra are then used for identification by comparing them with MS/MS-spectra stored in a library. No further qualifier ion has to be monitored. See Figure 2.

Intensity x104 3 2 1 0 x105 0.8 0.4 0.0 x104 2.0 1.0 0.0 x104 2.0 1.0 0.0 6000 4000 2000 0 x104 1.2 0.8 0.4 0.0 2 4 6 8 10 12 Time [min]

Intensity

2A

NBA-AMOZ

x104 3 2 1 0
x104 6 4 2 0

2B

NBA-AMOZ
192.9

290.9

261.9

317.8 333.9

NBA-d5AMOZ

NBA-d5AMOZ
264.9

295.9

x104 3.0

NBA-AOZ

2.0 1.0 0.0 x104 2.0


119.9

164.7 133.8

217.7 235.8 206.7

NBA-AOZ

148.8

171.8

190.7

NBA-d4AOZ

133.8

1.0 0.0 8000 6000 4000 2000 0 8000 6000 4000 2000 0 100 125
122.9 158.9 176.8 148.8 166.7 239.9 204.8 222.8

NBA-d4AOZ

133.8 120.8
138.7

162.7 230.7 170.7 188.7 200.7212.8 248.8

NBA-AHD

144.8

NBA-AHD

165.7

NBA-SEM

NBA-SEM
135.8 148.7 190.8 207.8 217.8 228.7

150

175

200

225

250

275

300

325

m/z

Figure 2.

Representative chromatograms (2A) and MS/MS spectra (2B) for all analytes plus ISTDs (1 g/kg).

Quantitation is performed by selecting one or more product ions to create extracted ion chromatograms for each analyte and ISTD. The product ions used for quantitation were selected for best signal-tonoise (S/N) ratio post-acquisition. See Figures 3 and 4.
4000

3000

Intensity

2000

1000

0 2.5

3.0

3.5

4.0

4.5 5.0 Time [min]

5.5

6.0

6.5

7.0

Figure 3.

Limit of quantitation (LOQ) for NBA-AMOZ, 0.125 g/kg in shrimp matrix.

1.0

1.0

0.8

0.8

ISTD Response x106

Relative response

0.6

0.6

0.4

0.4

0.3

0.3

0.0 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 Relative concentratiion

0.0

Figure 4.

Calibration curve for NBA-AMOZ, 0.125 g/kg 2 g/kg poultry matrix, three replicates.

NBA-AMOZ and NBA-SEM were quantified using the sum of two product ions, while NBA-AHD and NBA-AOZ were quantified using one product ion. The European Union has set a Minimum Required Performance Level (MRPL) of 1 g/kg for nitrofuran metabolites. These detection limits are easily reached using this method with LOQs ranging from 0.125 g/kg for NBA-AMOZ to 0.25 g/kg for NBA-AOZ and NBA-SEM, and 0.5 g/kg for NBA-ADH. See Figure 5.
Intensity x104 4 3 2 1 0

NBA-AMOZ

x104 6 4 2 0 x104 1.50 1.00 0.50 0.00 x104 1.2 0.8 0.4 0.0 6000 5000 4000 3000 2000 1000 0 x104 4 3 2 1 0 2 4

NBA-d5 AMOZ

NBA-AOZ

NBA-d4 AOZ

NBA-AHD

NBA-SEM

10

12

Time [min]

Figure 5.

Representative chromatogram of a positive shrimp sample at a level of 0.25 g/kg.

Linearity of the method was evaluated up to twice the MRPL (2 g/kg) and showed a linear weighted regression (1/) with coefficients of correlation of 0.99 or better. Intraday relative standard deviations (RSDs) were below 10% for all analytes at all concentrations. See Table 1.
Table 1. Method Reproducibility and Accuracy for the Four Target Derivatized Metabolites NBA-AOZ SD % 2.29 2.52 3.35 3.01 2.13 n=6 Accuracy % (average) 98.94 101.76 100.58 101.90 96.82 NBA-AMOZ Accuracy % SD % (average) 4.34 101.10 5.87 95.55 6.21 105.19 5.46 99.11 5.77 99.05 NBA-AHD SD % 6.05 4.59 6.53 6.97 Accuracy % (average) 98.31 103.87 100.22 97.60

NBA-SEM Standard Accuracy % (g/kg) SD % (average) 0.125 3.77 98.81 0.25 2.26 102.72 0.5 3.40 100.72 1 3.56 96.62 2 2.94 101.12 All calibration curves linear weighted 1/x

Conclusions
An LC/MS/MS method was developed for the qualitative and quantitative measurement of nitrofuran metabolites in chicken and shrimp using the Agilent XCT Ion Trap. The LOQ for all four nitrofurans investigated easily met the specified EU MRPL of 1 g/kg and ranged from 0.125 g/kg to 0.5 g/kg.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA March 25, 2004 5989-0738EN

The Analysis of Fluoroquinolones in Beef Kidney Using HPLC Electrospray Mass Spectrometry Application

Food

Authors
Ralph Hindle Access Analytical Labs #3, 2616 - 16 Street N.E. Calgary, Alberta T2E 7J8 Canada Agilent contact: Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Introduction
Fluoroquinolones are synthetic antibacterial compounds derived from nalidixic acid, and are useful to treat animal infections that are resistant to other antibacterial agents. They have a broad spectrum of activity, acting against both gram-positive and gram-negative bacteria. The maximum residue limit (MRL) for enrofloxacin (as the sum of enrofloxacin and ciprofloxacin) was entered into Annex 1 of Council Regulation (EEC) No. 2377/90 for kidney at 200 g/kg in bovine and ovine species, and 300 g/kg for porcine, poultry, and rabbits. For all other food producing species, the MRL is 200 g/kg in kidney [1]. There are a number of methods describing the analysis of fluoroquinolones in various tissues, with HPLC coupled with fluorescence and mass spectrometric detection being very popular. Most methods involve extraction into acidic or basic organic solvents, followed by some type of cleanup, most notably solid phase extraction (SPE). The Canadian Food Inspection Agency extracts animal tissue with acidic ethanol, followed by strong cation exchange SPE cleanup, and HPLC fluorescence analysis [2]. Chen and Schneider [3] described a screening method for enrofloxacin in chicken, where extracts were detected by fluorescence without cleanup, following extraction and centrifugation.

Abstract
A fast and simple screening method was validated for the analysis of three fluoroquinolone antibiotics in beef kidney. Samples were extracted with acidified methanol, centrifuged, diluted with water, and filtered. The diluted extract was analyzed directly by HPLC mass spectrometry using electrospray ionization in positive ion mode. Using an internal standard, mean recoveries were 73%96% at spiking levels of 33 g/kg (ppb), with statistically derived detection limits of 819 g/kg. This is below the European Union maximum residue limit of 200 g/kg for enrofloxacin and ciprofloxacin in bovine kidney. The method is evaluated relative to the requirements of the European Commission Decision 2002/657/EC for use as a confirmatory method.

European Community Commission Decision 2002/657/EC allows the use of HPLC coupled with fluorescence detection [4] for substances in Group B of Annex I to Directive 96/23/EC. Quinolones and other veterinary drugs fall into Group B, where three identification points are required for confirmation by Selected Ion Monitoring (SIM) using mass spectrometry (MS). With low resolution HPLC/MS, one point can be earned for each ion detected, provided that the ion ratios meet relative intensity criteria. Additional requirements of Directive 2002/657/EC, based on spiking levels of 33 g/kg carried out in this study, are as follows: The internal standard (IS) shall be added to the test portion at the beginning of the extraction procedure. In order to allow the use of data corrected for mean recovery, the range of recoveries allowed are 20% to +10%. The reproducibility of coefficient variation (CV) (%) is expected to be about one-half to twothirds of the 100 g/kg CV, which is 23%, at a concentration of half the permitted limit. For liquid chromatography/mass spectrometry (LC/MS) procedures, the minimum acceptable retention time (RT) for the analyte under examination is twice the RT corresponding to the void volume of the column. The ratio of the chromatographic RT of the analyte to that of the IS, that is, the relative RT of the analyte, shall correspond to that of the calibration solution at a tolerance of 2.5% for LC. The molecular ion shall preferably be one of the selected diagnostic ions. The maximum permitted tolerances for relative ion intensities shall meet the criteria in the Annex, (in this case, either 25% or 30%), as reproduced in Table 6.

Acidified methanol was prepared by adding 100 L of 98% formic acid to 100 mL of methanol. Acidified deionized water was prepared by adding 100 L of 98% formic acid to 100 mL deionized water. Ultra-Turrax T25 homogenizer, 50-mL polypropylene centrifuge tubes, and 13-mm polyvinylidene fluoride (PVDF) syringe filters (0.2 m), were purchased from VWR Scientific. All fluoroquinolones, including the IS, were provided as a gift from the Canadian Food Inspection Agency, Calgary, Alberta, Canada, as stock solutions of 100 ng/L (ppm) in 1% acetic acid in methanol. Solutions were stored at 4 C. Standard solutions at different concentrations were prepared for spiking by dilution with acidified methanol solution. The analytes ciprofloxacin, enrofloxacin, and sarafloxacin were chosen as targets since these compounds are included in the Canadian Food Inspection Agencys proficiency check samples. The spiking standard for these compounds (1 ng/L) was prepared by diluting 100 L of each the stock solutions to a 10-mL volumetric flask, and made to volume with acidified deionized water. A separate IS solution at 1 ng/L was prepared the same way, except that it only contained norfloxacin and danofloxacin. Sample Preparation 1. For beef kidney, 3 g samples were weighed directly into 50-mL polypropylene centrifuge tubes. 2. For spiked samples, 100 L of the 1-ng/L (100 ng) spiking solution was added, resulting in fortification levels of 33 g/kg. Samples were allowed to stand for 1 hour before subsequent extraction. 3. For the sample blank, 100 L of acidified methanol solution was added. 4. For all spiked samples, 100 L of the 1-ng/L (100 ng) IS solution was added just prior to extraction. Norfloxacin was included in this solution at the same level, to be used as an alternate IS, if required due to potential interferences for danofloxacin. 5. The samples were homogenized for 2 min with 15 mL of acidified methanol using the Ultra-Turrax homogenizer. 6. The samples were then centrifuged for 10 min, and the supernatant decanted into a clean test tube.

Experimental
Chemicals and Materials HPLC-grade methanol and acetonitrile were purchased from Caledon Labs (Georgetown, Ontario). Formic acid, min. 98%, was purchased from EM Science. Acidified methanol solution: 30% methanol in pH 3 deionized water (100 L of formic acid per 100 mL of water).

7. The extract was diluted with acidified deionized water 1 in 4 (250 L of extract + 750 L of water), filtered through a 0.2-m PVDF filter into an autosampler vial, and analyzed directly by LC/MS. By adding an accurately known amount of IS to the initial sample before extraction, there is no need to measure the final volume of the extracts, nor the aliquot to be diluted. The IS calculations, performed by the ChemStation, measure the relative amounts of the analytes and IS. This corrects for any concentration or dilution effects in the samples. Standard Preparation A 5-point calibration curve was used for the determination of each of the three target compounds, and a 1-point curve was used for norfloxacin, the alternate IS. Table 1 gives the volumes of the IS and target solutions added (1 ng/L each) to each of five test tubes. The standards were prepared by adding 250 L of the blank extract and 750 L of acidified deionized water to the tubes containing the analytes, after which the solutions were filtered through 0.2-m PVDF filters. The final solution of each standard contained 5 ng of IS per mL of diluted extract, or 5 pg/L. With 50 L injected, this results in 250 pg injected. The amount of target analyte in each of the five solutions varies to produce the calibration curves, as shown in Table 1. The correlation coefficient (R ) for the target analytes ranged from 0.9987 to 0.9992, as shown in Table 4. Preparation of the standards in this fashion will compensate for any ion suppression or enhancement that may occur, due to the presence of co-eluting material at the MS source, which may not otherwise occur if pure solvents alone are used.
Table 1. Preparation of Analytical Standards (50-L Injections into LC/MSD) Target volume added (L) 1 2 5 10 20 Target amount injected (pg) 50 100 250 500 1,000
2

LC/MS Conditions The HPLC system was made up of an Agilent Technologies 1100 series solvent degasser, binary pump, autosampler, column oven, diode array detector (DAD), and quadrupole mass selective detector (MSD) (Table 2).
Table 2. HPLC Column Zorbax Eclipse XDB-C8, 150 mm 4.6 mm, 5 m (P/N 993967-906) 0.1% Formic acid in water 0.1% Formic acid in acetonitrile t0 = 20% B t1 = 20% B t8 = 90% B t15 = 90% B Post time = 2.0 min 0.4 mL/min 50 L 30 C LC/MSD Conditions

Solvent A Solvent B Gradient

Flow rate Injection volume Column temp MSD Source Ion dwell time Fragmentation Drying gas flow Nebulizer pressure

Electrospray Ionization (ESI) (positive ion mode) 14 ions at 40 ms each Varies by ion, see Table 3 12 L/min 30 psi

Drying gas temperature 350 C Capillary voltage 4000 V

Table 3.

Fragmentor Voltages for Acquired Ions in SIM (single acquisition group) Ion 320 302 276 332 314 288 358 340 360 342 316 386 368 342 Fragmentor (V) 120 200 200 120 200 200 120 220 120 220 220 120 220 220 3

Compound Norfloxacin (IS)

Ciprofloxacin

Standard 1 2 3 4 5

IS Volume added (L) 5 5 5 5 5

IS Amount injected (pg) 250 250 250 250 250

Danofloxacin (IS) Enrofloxacin

Sarafloxacin

All ions were included in a single acquisition group, which started at injection (time = 0). An alternative approach would be to set the group start time to a value around half a minute before the elution of the first compound, as this will keep the eluant stream diverted to waste as long as possible. This will reduce the amount of co-extracted material being introduced into the source, reducing contamination. Another alternative is to add an additional timeprogrammed acquisition group to the method, and only include the ions for compounds eluting within the group times. This will take on more significance as the overall number of compounds in a method increases, and with three ions per compound required for identity confirmation. Fragmentor voltages were chosen that maximized the response for each selected ion. For each fluoroquinolone, a value of 120 V produced only the protonated parent ion, while higher voltages were required to induce fragmentation to confirmatory ions. The ions monitored corresponded to the neutral losses of water and carbon dioxide in each case. Note that although mass 342 is acquired for both enrofloxacin and sarafloxacin, it is only added to the MSD acquisition table once.

Chromatography All compounds eluted between 5 and 9 minutes, however the total run time was set to 15 minutes with 90% organic solvent to allow co-extractives to elute from the column. Otherwise, their eventual elution could interfere with subsequent injections. This is more of a potential problem when methods with abbreviated cleanups, such as dilution-only, are used. The following figures compare the blank beef kidney sample to a sample fortified at 33 g/kg. In each case, the selected ions are the protonated forms of the parent ion, as well as the protonated ions resulting from the loss of H2O (M-18) and CO2 (M-44). The qualifier ion for danofloxacin, the compound used as the IS for this study, is mass 340. The matrix causes an interference at mass 340. The interference is shown as a small peak in the beef kidney blank as shown in Figure 1. Since a diagnostic qualifier ion is not required for the IS calculations, it had no impact on the results. It does, however, indicate that there is elution of co-extractive material in the samples, and that without further cleanup, ion suppression may result from its presence. All standards were prepared in blank beef kidney extract in order to compensate for these potential effects.

Norfloxacin (IS) 320 = [M + H]+ 302 = [M - H2O + H]+ 276 = [M - CO2 + H]+

12000 8000 4000 0 12000 8000 4000 0 12000 8000 4000 0 5.5

5.530 - NOR 320

5.531- NORq1

302

5.532 - NORq2
6 6.5 7 7.5 8 8.5 9

276
min

Beef kidney blank

12000 8000 4000 0 12000 8000 4000 0 12000 8000 4000 0

5.484 - NOR 320

5.485 - NORq1

302

5.482 - NORq2
5.5 6 6.5 7 7.5 8 8.5 9

276
min

Figure 1.

Comparative extracted ion chromatograms for fluoroquinolones spiked into beef kidney.

Ciprofloxacin spike 332 = [M + H]+ 314 = [M H2O + H]+ 288 = [M CO2 + H]+

6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5

5.854 - CIPRO 332

5.850 - CIPROq1

314

5.846 - CIPROq2
6 6.5 7 7.5 8 8.5 9

288
min

Beef kidney blank

6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5 6 6.5 7 7.5 8 8.5 9

332

314

288
min

Danofloxacin (IS) 358 = [M + H]+ 340 = [M - H2O + H]+

8000 6000 4000 2000 0 8000 6000 4000 2000 0 5.5 6

6.009 - DANO 358 6.012 - DANOq1 340


6.5 7 7.5 8 8.5 9 min

Beef kidney blank

8000 6000 4000 2000 0 8000 6000 4000 2000 0 5.5 6

358

6.04 - DANOq1
6.5 7 7.5 8 8.5 9

340
min

Figure 1.

Comparative extracted ion chromatograms for fluoroquinolones spiked into beef kidney (Continued).

Enrofloxacin spike 360 = [M + H]+ 342 = [M H2O + H]+ 316 = [M CO2 + H]+

6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5 6 6.5

6.910 - ENRO 360

6.914 - ENROq1

342

6.917 - ENROq2
7 7.5 8 8.5 9

316
min

Beef kidney blank

6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5 6 6.5 7 7.5 8 8.5 9

360

342

316
min

8000 4000

8.604 - SARA 386 8.601- SARAq1 368 8.594 - SARAq2 342


5.5 6 6.5 7 7.5 8 8.5 9 min

Sarafloxacin spike 386 = [M + H]+ 368 = [M H2O + H]+ 342 = [M CO2 + H]+

0 8000 2000 0 8000 4000 0

8000 4000

386

Beef kidney blank

0 8000 4000 0 8000 4000 0

368

342

Figure 1.

Comparative extracted ion chromatograms for fluoroquinolones spiked into beef kidney (Continued).

Sarafloxacin elutes from the column in the same region as a number of other co-extractives, making identification and quantitation more difficult. However, as shown in Table 7, the qualifier ions still meet the identification criteria for relative responses of the qualifiers, and so further cleanup of the samples may not be necessary. The effect of these co-extractives will also be reduced at higher incurred residue levels, closer to those permitted by the European Union MRL. Recoveries In order to allow results to be corrected for recoveries, where the determined incurred levels are divided by the percent recovered from certified reference materials or spiked samples, Table 2 of the Annex requires that the recoveries for analytes at levels greater than 10 g/kg be within the range of 80% to 110%. Table 4 shows that recoveries for ciprofloxacin and enrofloxacin meet this requirement, with 96.3% and 86.0%, respectively. However, sarafloxacin fails the requirement, with only 72.6%

mean recovery. With a CV of only 8% for this compound, it looks as though the method may still produce acceptable results for screening purposes, but some additional work may be required to produce higher recoveries. Since the work presented here involves spiked samples only, recoverycorrection calculations do not apply. Norfloxacin was added along with danofloxacin as an additional IS. However, examination of the blank beef kidney used in this study shows norfloxacin to be present as an incurred residue, at a concentration approximately one half of the spiking level. Assuming a linear response through the origin, this would mean that norfloxacin was detected at approximately 1520 g/kg, which is about 10% of the permitted level for enrofloxacin in bovine kidney. Recoveries for norfloxacin are included in Table 4, even though they were calculated with a single point calibration, and not corrected for incurred residues. However, there is some compensation for this since the standards used for calibration were prepared by addition of the targets to the blank extracts.

Table 4.

Recoveries of Fluoroquinolones from Beef Kidney Amount recovered (ng) Ciprofloxacin Enrofloxacin 96.3 84.9 94.0 85.6 89.6 83.8 95.4 86.2 93.8 85.4 109.3 87.9 101.3 83.3 90.8 91.3 100.0 100.0 96.3 86.0 6.3 2.5 19.0 7.6 63.4 25.4 6.6 3.0 96.3 86.0 0.9987 0.9992 3.00 3.00

Description Kidney spike 1 Kidney spike 2 Kidney spike 3 Kidney spike 4 Kidney spike 5 Kidney spike 6 Kidney spike 7 Kidney spike 8 Amount spiked (ng) Mean SD (Precision) ng MDL (SD t-stat) ng LOQ (SD 10) ng CV (SD/Mean) % Accuracy (%) Linearity (R2) t-stat (N = 8)

Norfloxacin 111.8 93.1 88.0 98.9 82.2 143.0 102.6 110.6 100.0 103.8 18.9 56.7 189.1 18.2 103.8 0.9895 3.00

Sarafloxacin 68.5 64.1 77.6 75.2 82.1 72.9 73.0 67.7 100.0 72.6 5.8 17.4 58.1 8.0 72.6 0.9987 3.00

Compound Identification For chromatographic separation, Section 2.3.3.1 of the Annex to 2002/657/EC requires that the minimum acceptable RT for the analyte under investigation be at least twice the RT corresponding to the void volume of the column (k'=1). The first compound to elute under these conditions is norfloxacin, with a k' of 2.6, therefore this condition is easily met. The second condition is that the ratio of the RT of the analyte to that of the IS, that is the relative RT, shall correspond to that of the calibration solution at a tolerance of 2.5% for LC. Table 5 shows the RT times of each analyte in the spiked samples, compared to those of the standards, and that they are well within the allowable tolerance.
Table 5. Relative RTs of Analytes in Samples, Compared to Standards Average RRT in standards (N = 15) 0.922 0.975 1.150 1.439 CV (%) RRT in standards (N = 15) 0.12% 0.05% 0.16% 0.47% RRT in samples, relative to standards (N = 8) 99.8%100.1% 99.9%100.1% 99.8%100.2% 99.5%100.3%

Compound Norfloxacin Ciprofloxacin Enrofloxacin Sarafloxacin

Compound Confirmation Section 2.3.3.2 of the Annex to 2002/657/EC gives the maximum permitted tolerances for relative ion intensities, which is reproduced in Table 6.

Table 6.

Maximum Permitted Tolerances for Relative Ion Intensities Using a Range of Mass Spectrometric Techniques GC/MS(EI) (relative) 10% 15% 20% 50% GC/MS(CI), GC/MSn, LC/MS, LC/MSn (relative) 20% 25% 30% 50%

Relative intensity (% of base peak) >50% >20% to 50% >10% to 20% 10%
Note MSn equals MS/MS if n = 2

Table 7 shows the relative intensities for each of the qualifier ions for the three target compounds, as well as norfloxacin and danofloxacin (one ion). As expected, norfloxacin meets the criteria in each of the eight spiked samples, even though it had incurred residues. The presence of additional norfloxacin should not negatively affect this qualitative aspect of performance, and it does not. Danofloxacin, however, showed an interference for the single qualifier ion monitored, and so the relative amount of this signal would be expected to vary to a larger degree, depending upon the exact amount of blank extract used in preparing the

sample dilutions and standards. As previously mentioned, the standards are prepared by accurately measuring the relative amounts of target and IS compounds into a tube or vial, followed by addition of blank kidney extract and water. The exact proportions of extract and water do not have to be known, since the IS calculations uses amount and response ratios, rather than absolute amount and response, in determining concentrations in unknowns. An accurate measurement of extract and water volumes can, however, reduce interference variability.

Table 7.

Relative Intensities of Qualifier Ions for Fluoroquinolones in Beef Kidney, Compared to Permitted Tolerances Relative intensities (%) of qualifier ions Ciprofloxacin Danofloxacin Enrofloxacin Q1 = 314 Q2 = 288 Q1 = 340 Q1 = 342 Q2 = 316 47 17 64 44 28 48 17 58 46 24 46 20 63 44 28 46 17 60 43 30 45 21 65 44 26 39 17 72 45 29 41 18 65 46 29 42 19 62 39 24 44 20 86 43 26 3 2 22 2 2 25 30 20 25 25 33 14 69 32 19

Sample Spike 1 Spike 2 Spike 3 Spike 4 Spike 5 Spike 6 Spike 7 Spike 8 Average for Stds Std Dev for Stds Tolerance(Table 7) Lower Allowable (calculated) Upper Allowable (calculated)

Norfloxacin Q1 = 302 Q2 = 276 49 15 45 15 48 16 42 15 49 17 50 19 49 17 47 17 49 20 2 1 25 30 37 14

Sarafloxacin Q1 = 368 Q2 = 342 50 15 41 17 45 14 43 13 45 11 43 12 46 13 46 12 47 15 6 1 25 30 35 11

62

26

55

25

103

53

32

59

20

www.agilent.com/chem

Conclusion
A fast and sensitive single quadrupole LC/ESI/MS method was validated for the detection of three fluoroquinolone antibiotics (ciprofloxacin, enrofloxacin, and sarafloxacin) in beef kidney. The detection limits ranged from 8 to 19 g/kg (ppb), with direct analysis of sample extracts after dilution with water. All qualitative requirements were met with respect to the Annex to EU Directive 2002/657/EC for spiked samples, and recoveries of two of the three compounds met the quantitative requirements. Recovery of sarafloxacin was slightly lower than the level required to allow correction for recoveries in reported results.

References
1. The European Agency for the Evaluation of Medicinal Products, Veterinary Medicines and Inspections, EMEA/MRL/820/02-FINAL, January 2002. 2. Determination of Fluoroquinolones in Bovine, Porcine and Avian Tissues by Liquid Chromatography with Fluorescence Detection, FQL-SP04, Canadian Food Inspection Agency, Saskatoon, Saskatchewan, Canada; 2001/03. 3. Chen, G., Schneider, M. J., (2003) A Rapid Spectrofluorometric Screening Method for Enrofloxacin in Chicken Muscle. J. Agric. Food Chem., 51(11), 3249-3253. 4. Annex of Commission Decision 2002/657/EC, Commission Decision of 12 August 2002, implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results, Official Journal of the European Communities, 17.8.2002, L 221/8-36, Table 5, Footnote 4.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA February 10, 2004 5989-0596EN

A Validated Atmospheric Pressure Chemical Ionization Method for Analyzing Sulfonamides in Pork Muscle Application

Food

Author
Ralph Hindle Access Analytical Labs #3, 2616 - 16 Street N.E. Calgary, AB. T2E 7J8 Canada Agilent contact: Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA

Introduction
Meat, edible organs, animal feed, and animal waste may contain antibiotics, growth hormones, and other chemicals that can enter the food supply. These compounds are added to maintain animal health, to increase animal growth rate, and to reduce stress. Human exposure can result from eating contaminated meat, or contacting runoff and leaching from manure and compost. Health specialists warn that there may be reduced options for effectively treating disease with antibiotics, such as penicillin and sulfa drugs, since antibioticresistant strains of bacteria may develop from the low-level exposure. Sulfonamides are broad-spectrum antimicrobials used in both humans and animals. The maximum residue limit (MRL) in Canada for sulfonamides in meat is 100 ppb (ng/g), and 10 ppb in milk, while the MRL in the European Union is 100 ppb for both of these matrices. The Canadian Food Inspection Agency method for sulfonamides in meat tissue calls for extraction in ethyl acetate, partitioning with glycine buffer, followed by a pH-adjusted back extraction into methylene chloride [1]. Extracts are evaporated, reconstituted, then separated by thin layer chromatography (TLC), derivatized, and quantitated by densitometry. Alberta Agriculture has improved the quantitative and qualitative aspects by using liquid chromatography/mass spectrometry (LC/MS) with atmospheric pressure chemical ionization (APCI) for the final analysis [2]. There are a number of

Abstract
This application note presents a simple method for the analysis of sulfonamide antibiotics in pork muscle. Samples were extracted with acidified methanol, centrifuged, and a portion of the extract was diluted with water. This dilution was analyzed directly by HPLC mass spectrometry using chemical ionization, with all compounds eluting in less than 5 minutes. Using an internal standard, recoveries for seven sulfonamides ranged from 84%118% at a spiking level of 50 ppb (ng/g). The statistically derived detection limit was 1025 ppb. A comparison was made to the cleaned extracts using solid phase extraction, as well as a comparison of mass selective detector settings for both screening (maximum sensitivity) and confirmation (greater fragmentation). The enhanced sensitivity of the Agilent quadrupole mass selective detector allows this dilution cleanup technique to be used in labs where high throughput is required.

extraction steps in the Alberta method, and a faster method would greatly benefit laboratories monitoring the food supply for residues. The goal of this method was to reliably quantitate the sulfa drugs at one-half of the regulatory limit or lower, with minimal sample preparation, and a maximum injection cycle time of 10 minutes. Maximum sensitivity is generally obtained by forming as many parent ions [M+H]+ as possible and minimizing fragmentation. Due to the operational complexity of triple quadrupole instruments, it is also desirable to confirm positive findings on a single quadrupole. This could be achieved by using collision induced dissociation (CID) to enhance fragment ions characteristic of the compounds.

Sample Preparation 1. For pork muscle, 3 g samples were weighed directly into 50-mL polypropylene centrifuge tubes. 2. The samples were homogenized for 3 minutes with 10 mL acidified methanol using the UltraTurrax homogenizer. 3. The samples were then centrifuged for 10 minutes, and the supernatant decanted into a clean test tube. 4. The samples were then re-extracted with a further 10 mL acidified methanol, and centrifuged again. 5. The supernatants were combined, and 1 mL IS (2 mg) was added to the combined extract. 6. The extract was diluted with de-ionized water 1 in 4 (250 L extract + 750 L water), filtered through a 0.2 m PVDF filter into an autosampler vial, and analyzed directly by LC/MS. By adding an accurately known amount of IS to the combined extracts, there is no need to measure the final volume of the extract. The IS calculations performed by the ChemStation measure the relative amounts of the analytes and IS. This corrects for any concentration or dilution effects in the samples. Sample extracts were also taken through SPE cleanup cartridges in order to compare with the dilution-only extracts. The 60-mg Oasis HLB cartridges are prewashed by eluting 1.5 mL acidified methanol, followed by 1.5 mL de-ionized water. The 1 mL extract was diluted to 10 mL with de-ionized water, eluted through the cartridges, and the eluant was discarded. The sulfa drugs were then eluted with 1.5 mL acidified methanol. This eluant was evaporated to near dryness under nitrogen. Samples were reconstituted in 1 mL of 25% methanol in water, filtered, and analyzed by LC/MS. A further comparison was done by evaporating 1 mL methanol extract to near dryness, and reconstituting it in 1 mL of 25% methanol in water without the SPE cleanup. This gave the sample extract the proper solvent composition for HPLC analysis, but without the dilution step to negatively affect the detection limits (DL) of the compounds.

Experimental
Chemicals and Materials All sulfonamide standards were purchased from Sigma Aldrich Canada, with a minimum purity of 99%. Stock solutions were prepared at 2 mg/mL in acetone, with the exceptions of sulfadiazine and the sodium salt of sulfaquinoxaline. Three mL of 0.2N NaOH was added in order to completely dissolve these compounds. Standard solutions at different concentrations were prepared for spiking and quantitation by diluting with de-ionized water. Internal standard (IS): sulfachloropyridazine (SCPD) at 2 mg/mL in de-ionized water. HPLC-grade methanol and acetonitrile were purchased from Caledon Labs (Georgetown, Ontario). Formic acid (min. 98%), was purchased from EM Science. Acidified methanol was prepared by adding about 100 L of 98% formic acid to 100-mL methanol. Ultra-Turrax T8 homogenizer with 8-mm diameter dispersing element, 50-mL polypropylene centrifuge tubes, and 13-mm polyvinylidene fluoride (PVDF) syringe filters (0.2 m), were purchased from VWR Scientific. Oasis HLB (3 cc, 60 mg) solid phase extraction (SPE) cartridges were purchased from Waters.

LC/MS Conditions
The LC/MS system was made up of Agilent Technologies 1100 Series solvent degasser, binary pump, autosampler, column oven, diode array detector, and quadrupole mass selective detector (MSD) (Table 1).

Compound Identification and Confirmation


In general, the goal of a monitoring method for target analytes is to separate the compounds from potential interferences and maximize sensitivity on the instrument. Using mass spectrometry (MS), maximum sensitivity is achieved by the production of a single ion, for example, the protonated parent

ion [M+H]+ in LC electrospray ionization (ESI) or APCI in target ion mode. However, once a positive is detected, a confirmation must be made as to whether the suspect peak is actually the target analyte, or simply a co-eluting compound that produces the same ion. There are a number of ways to perform the confirmation: re-extract the sample with a different solvent system; further clean up the sample to a higher final concentration, to allow detection of additional confirmation ions or analysis in scan mode; derivatize and analyze by gas chromatography/mass spectrometry (GC/MS); or re-analyze the extract on a triple quadrupole LC/MS/MS. All of these techniques are useful, but the drawback is the additional time and expense involved, especially with LC/MS/MS.

Table 1. HPLC Column Solvent A Solvent B Gradient

LC/MSD Conditions Zorbax Eclipse XDB-C8, 150 mm 4.6 mm, 5 m (p/n 993967-906) 0.1% Formic acid in water 0.1% Formic acid in acetonitrile t0 = 20% B t1 = 20% B t3 = 90% B t6.5 = 90% B Post time = 1.5 min 1.0 mL/min 50 L 30 C APCI (positive ion mode) 8 Ions at 63 ms each 70 V 6.0 L/min 60 psi 350 C 400 C 3000 V 4 A

Flow rate Injection volume Column temp MSD Source Ion dwell time Fragmentor Drying gas Nebulizer pressure Drying gas temperature Vaporizer temperature Capillary voltage Corona current

The Agilent 1100 MSD has the capability of acquiring up to four separate MS signals during the same run, where each signal can be made up of a number of selected ions (SIM) or a full scan spectrum. For example, Signal 1, with a low fragmentor voltage to maximize parent ion response, can include each of the [M+H]+ ions in the target list, while Signal 2, at higher fragmentor voltages can acquire the confirmatory fragment ions. For analytes expected at higher concentrations, Signal 1 could acquire in SIM mode for quantitation, while Signal 2 could be set for scan mode for identification. Figure 1 demonstrates the former example, with the Fragmentor set to 70 V for Signal 1 (MSD1), and 200 V for Signal 2 (MSD2).

MSD1 250, EIC=249.7:250.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 70, "Quantitation"

8000 4000 0 2.5 3.0

3.365 - SPY

Simultaneous 2-signal aquisition; Fragmentor at 70 V or 200 V

3.5

4.0

4.5

5.0

5.5

min

MSD1 285, EIC=284.7:285.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 70, "Quantitation"

8000 4000 0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 min

4.344 - SCPD (IS)

MSD2 108, EIC=107.7:108.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 200, "Confirmation"

3000 2000 1000 2.5 3.0 3.5 4.0 4.5 5.0 5.5 min

MSD2 156, EIC=155.7:156.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 200, "Confirmation"

700 500 300 2.5 3.0 3.5 4.0 4.5 5.0 5.5 min

Figure 1.

Dual MSD acquisition signals (Masses 108 and 156 are class-specific fragments for sulfonamides).

Table 2 shows the mass spectra for the sulfonamides using various fragmentor voltages. Masses 108 and 156 are class-specific fragments for sulfonamides (H2N+=[C6H4]=O and H2N+=[C6H4]=SO2, respectively), and, as such, are very useful diagnostic ions, when acquired along with the protonated molecular ion.

Table 2.
Compound

APCI Spectra of Sulfonamides, Using Various Fragmentor Voltages


Fragmentor 70 V
*MSD1 SPC, time=2.859:3.513 of SFCISCAN\SULFA006.D APCI, Pos, Scan, Frag: 70

Fragmentor 160 V
*MSD1 SPC, time=2.891:3.337 of SFCISCAN\SULFA008.D APCI, Pos, Scan, Frag: 160

Fragmentor 200 V
*MSD1 SPC, time=2.907:3.258 of SFCISCAN\SULFA009.D APCI, Pos, Scan, Frag: 200

156.1

256.1

100

101.1 108.2

0
100 200 300

m/z

0 100 200 300

107.5

m/z

0 100 150 200 250

256.1 m/z

20

20

157.1

257.1

20

*MSD1 SPC, time=2.971:3.369 of SFCISCAN\SULFA010.D APCI, Pos, Scan, Frag: 70

*MSD1 SPC, time=2.986:3.369 of SFCISCAN\SULFA012.D APCI, Pos, Scan, Frag: 160

*MSD1 SPC, time=2.987:3.242 of SFCISCAN\SULFA013.D APCI, Pos, Scan, Frag: 200

251.1

251.1

252.1

158.1

252.1

0
100 200

300 m/z

0 100 200

185.1

20

20 300 m/z 0

100

150

200

m/z

*MSD1 SPC, time=3.212:3.467 of SFCISCAN\SULFA002.D APCI, Pos, Scan, Frag: 70

*MSD1 SPC, time=3.210:3.608 of SFCISCAN\SULFA004.D APCI, Pos, Scan, Frag: 160

*MSD1 SPC, time=3.210:3.529 of SFCISCAN\SULFA005.D APCI, Pos, Scan, Frag: 200

250.1

108.1

156.1

120.1

101.2

251.1

20 0

108.1

184.1

251.1

20 m/z 0

20 m/z 0 100

157.1

183.4

40

40

40

107.4

Sulfapyridine (SPY) C11H11N3SO2 MW = 249 RT = 3.33 min

80 60

250.1

80 60

80 60 156.1

184.1

100

Max: 35037

100

Max: 16213

100

Max: 2943

100

150

200

250

100

200

300

150

200

m/z

*MSD1 SPC, time=3.626:3.865 of SFCISCAN\SULFA002.D APCI, Pos, Scan, Frag: 70

*MSD1 SPC, time=3.608:3.991 of SFCISCAN\SULFA004.D APCI, Pos, Scan, Frag: 160

*MSD1 SPC, time=3.672:3.927 of SFCISCAN\SULFA005.D APCI, Pos, Scan, Frag: 200

265.1

265.1

156.1 172.1 184.1 199.1

101.2 117.1

110.2

250.1

266.2

20 0

20
300 m/z

155.3 172.1

20
300 m/z

100

200

100

200

109.4 120.1

266.1

265.1

Sulfamerazine (SMR) C11H12N4SO2 MW = 264 RT = 3.78 min

80 60 40

80 60 40

80 60 40
108.2

110.2

100

Max: 20181

100

Max: 9216

100

Max: 4083

100

200

300

m/z

250.1

251.1

20

158.0

185.1

40

107.5 109.1 120.1

40

40

107.5 108.2

156.1

Sulfadiazine (SDZ) C10H10N4SO2 MW = 250 RT = 3.09 min

60

60

156.1

80

80

80 60

108.1

100

Max: 66929

100

Max: 23650

100

107.5 120.1

40
257.1

40

40

156.1

Sulfathiazole (STZ) C9H9N3S2O2 MW = 255 RT = 3.05 min

80 60

80 60

256.1

Max: 151737

100

Max: 69275

100 80 101.1 60

108.1

Max: 32290

Max: 10375

Table 2.
Compound

APCI Spectra of Sulfonamides, Using Various Fragmentor Voltages (Continued)


Fragmentor 70 V
*MSD1 SPC, time=3.993:4.168 of SFCISCAN\SULFA002.D APCI, Pos, Scan, Frag: 70

Fragmentor 160 V
*MSD1 SPC, time=3.991:4.262 of SFCISCAN\SULFA004.D APCI, Pos, Scan, Frag: 160

Fragmentor 200 V
*MSD1 SPC, time=4.007:4.246 of SFCISCAN\SULFA005.D APCI, Pos, Scan, Frag: 200

100

Max: 57189
279.1

100 80 60 40

Max: 35356
279.1

100 80 60
108.1 123.5 124.2

Max: 11707
279.1 156.1 186.1 213.2 280.1 300 m/z 200

280.1

124.2

20 0
100 200

20
m/z

186.1

280.1

Sulfamethazine (SMZ) 60 C12H14N4SO2 MW = 278 40 RT = 4.06 min

80

40 20
m/z

300

100

200

300

100

*MSD1 SPC, time=4.184:4.423 of SFCISCAN\SULFA002.D APCI, Pos, Scan, Frag: 70

*MSD1 SPC, time=4.262:4.453 of SFCISCAN\SULFA004.D APCI, Pos, Scan, Frag: 160

*MSD1 SPC, time=4.262:4.438 of SFCISCAN\SULFA005.D APCI, Pos, Scan, Frag: 200

100
285.0

Max: 15461

100
156.0

Max: 6222
285.0

111.2 123.3 124.2 135.3 138.1

40 20
m/z

163.1 179.1

107.4 120.1 130.1

40 20 0

40 20

265.1 284.3 287.0

101.2 115.2

100

200

300

m/z

100

200

300

184.2

Sulfachloropyridazine (SCPD); istd C10H9N4SO2Cl MW = 284 RT = 4.31 min

80 60

80 60
287.1

80 60

108.1

100

Max: 3906

156.0

100

150

200

250

m/z

*MSD1 SPC, time=4.470:4.725 of SFCISCAN\SULFA014.D APCI, Pos, Scan, Frag: 70

*MSD1 SPC, time=4.389:4.756 of SFCISCAN\SULFA016.D APCI, Pos, Scan, Frag: 160

*MSD1 SPC, time=4.469:4.772 of SFCISCAN\SULFA017.D APCI, Pos, Scan, Frag: 200

301.1

301.1

100 80

80 60
302.1 302.1

80 60 40 20
m/z

40 20 0
100 146.1

40
108.2 129.1

107.5 120.1 129.1 145.3

235.1

20
m/z

200

300

100

200

300

288.4 302.1

Sulfaquinoxaline (SQ) C14H12N4SO2 MW = 300 RT = 4.51 min

156.1

301.1

60

156.1

Max: 148192

100

Max: 63906

100

108.2

Max: 16120

100

200

300

m/z

*MSD1 SPC, time=4.485:4.740 of SFCISCAN\SULFA018.D APCI, Pos, Scan, Frag: 70

*MSD1 SPC, time=4.454:4.900 of SFCISCAN\SULFA020.D APCI, Pos, Scan, Frag: 160

*MSD1 SPC, time=4.485:4.900 of SFCISCAN\SULFA021.D APCI, Pos, Scan, Frag: 200

311.1

311.1

156.1

100

Max: 467058

100 80 60
156.1

Max: 261189

100 80 60

311.1 218.1 245.2

Max: 80490

312.1

312.1

156.1

20 0
100

20
m/z

20
m/z

200

300

100

200

300

157.1

312.1

Sulfadimethoxine (SDMX) C14H12N4SO2 MW = 310 RT = 4.54 min

80 60 40

108.1

40

40

155.3

100

200

300

m/z

Chromatography
While complete separation of target compounds is not always necessary when using mass spectral detection, it is, however, essential when common ions are present. For example, the protonated molecular ion of SPY is 250 mass units. Due to the naturally-occurring C13 isotope, ions 251 coexist with the parent ions 250. Separating SPY from SDZ (m/z = 251) was, therefore, important when trying to optimize the chromatographic conditions, and was achieved as shown in Figure 2. While this results in a slightly longer chromatographic run than would otherwise be necessary, there is more consistent integration of the peaks during data

analysis; the chromatogram is easier to interpret; and the amount of SDZ is not underestimated due to co-elution of SPY in the standard mix. A recently published application shows four sulfonamides were analyzed with an injection cycle time of 1.1 minutes, using a 2-position 10-port valve, two analytical columns in parallel, and a second binary pump [3]. Since most labs do not have such high sample volume requirements, the method described in this application note was developed using more conventional techniques, without the additional hardware costs. Conditions were set up to provide good chromatographic separation in a relatively short time of 6 minutes (total cycle time was 10 minutes).

MSD1 256, EIC=255.7:256.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

3.051 - STZ
1
2

min

MSD1 251, EIC=250.7:251.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

3.084 - SDZ
1
2

min

MSD1 250, EIC=249.7:250.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

3.332 - SPY
1
2

min

MSD1 265, EIC=264.7:265.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

3.783 - SMR
1
2

min

MSD1 279, EIC=278.7:279.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

4.056 - SMZ
4 5 min

MSD1 285, EIC=284.7:285.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

4.314 - SCPD (IS)


4 5 min

MSD1 301, EIC=300.7:301.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

4.514 - SQ
4 5 min

MSD1 311, EIC=310.7:311.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70

4.537 - SDMX
4 5 min

Figure 2.

Sulfonamide standard mix, 500 pg each (SIM).

Sample Cleanup
The total ion chromatograms (TIC) in Figure 3 show that there is considerable matrix background from the samples. A simple solvent exchange was performed, where 1 mL of extract was evaporated under nitrogen, and reconstituted in 25% methanol in water. One of the problems with solvent exchange only is the amount of matrix material that is injected onto the HPLC column. Peak shape can be negatively affected by overloading, and eventually the performance of the column will deteriorate. All of this matrix material is also introduced into the MSD. Frequent cleaning and maintenance may be required for the MSD, further reducing productivity.

In order to develop a high-throughput method, keep the number of required steps to a minimum. The Agilent liquid chromatography/mass selective detector (LC/MSD) has enough sensitivity to allow simple dilution of the extracts with water to act as a cleanup technique. This eliminates the need for costly SPE cartridges and analyst time to further prepare the samples. Minimal sample handling can also improve recoveries, since losses are possible at each step. The third chromatogram in Figure 3 shows how the use of SPE cleanup techniques can remove the majority of co-extracted materials, allowing for a more concentrated final extract and ultimately lower DLs. This also results in a simpler chromatogram for integration and interpretation.

MSD1 TIC, MS File (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70 200000 150000 100000 50000 0 1 2 3 4 5 6 min

Spiked pork extract - solvent exchange to 25% MeOH in water

MSD1 TIC, MS File (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70 200000 150000 100000 50000 0 1 2 3 4 5 6 min

Spiked pork extract - Diluted 1 in 4 with water

MSD1 TIC, MS File (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70 200000 150000 100000 50000 0 1 2 3 4 5 6 min

Spiked pork extract - Oasis HLB cleanup

Figure 3.

TIC comparisons of various cleanup techniques.

However, where the goal of a method is to screen large numbers of samples to find potential violations of MRLs, a simple dilution technique may be preferred. Dilution could offer enough cleanup for good chromatographic separation, while remaining concentrated enough to meet DL requirements. The second chromatogram in Figure 3 shows a much improved baseline. Figures 4 through 6 show the same analyses with all the target ions in SIM mode.

MSD1 256, EIC=255.7:256.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

3 .0 5 4 - S T Z 3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 251, EIC=250.7:251.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

3.094 - SDZ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 250, EIC=249.7:250.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

3.340 - SPY
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 265, EIC=264.7:265.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

3.794 - SMR
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 279, EIC=278.7:279.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

3.180
3.0 3.5 4.0

4.083 - SMZ
4.5

4.760
5.0 5.5 min

MSD1 285, EIC=284.7:285.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

4.345 - SCPD (IS)


3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 301, EIC=300.7:301.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

4.549 - SQ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 311, EIC=310.7:311.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70

4.572 - SDMX
3.0 3.5 4.0 4.5 5.0 5.5 min

Figure 4.

Solvent exchange only (SIM)

MSD1 256, EIC=255.7:256.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

3.044 - STZ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 251, EIC=250.7:251.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

3.078 - SDZ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 250, EIC=249.7:250.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

3.324 - SPY
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 265, EIC=264.7:265.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

3.782 - SMR
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 279, EIC=278.7:279.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

3.139
3.0 3.5 4.0

4.056 - SMZ
4.5 5.0 5.5 min

MSD1 285, EIC=284.7:285.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

4.316 - SCPD (IS)


3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 301, EIC=300.7:301.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

4.517 - SQ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 311, EIC=310.7:311.7 (SF030816\SULFA011.D) APCI, Pos, SIM, Frag: 70

4.541 - SDMX

3.0

3.5

4.0

4.5

5.0

5.5

min

Figure 5.

Diluted 1 in 4 with water.

10

MSD1 256, EIC=255.7:256.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

3.042 - STZ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 251, EIC=250.7:251.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

3.078 - SDZ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 250, EIC=249.7:250.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

3.322 - SPY
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 265, EIC=264.7:265.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

3.777 - SMR
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 279, EIC=278.7:279.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

3.166
3.0 3.5 4.0

4.077 - SMZ
4.5 5.0 5.5 min

MSD1 285, EIC=284.7:285.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

4.343 - SCPD (IS)


3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 301, EIC=300.7:301.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

4.546 - SQ
3.0 3.5 4.0 4.5 5.0 5.5 min

MSD1 311, EIC=310.7:311.7 (SF030817\SULFA010.D) APCI, Pos, SIM, Frag: 70

4.569 - SDMX
3.0 3.5 4.0 4.5 5.0 5.5 min

Figure 6.

After HLB cleanup (SIM).

Results and Discussion


The recoveries obtained for seven samples spiked at a level of 50 ppb (150 ng of each sulfonamide in 3 g sample) appear in the following tables. The spiking solutions were added before homogenization and allowed to stand for at least 30 minutes before extraction. SMR (sulfamerazine) was added separately at 300 ng per sample before homogenization, and could be used as a surrogate. Results in Table 3 were obtained by simply diluting the extracts 4-fold with water (recovery 84%118%), while results in Table 4 are from extracts taken through SPE cleanup (recovery 79%104%).

In both cases, a five-point IS calibration with SCPD was used, with 20 to 200 pg of each target compound injected, plus 2,000 pg SCPD. The five standards were injected both before and after the set of seven spikes, and the curves were created by using the average responses of the two sets of standards. Peak height was used to measure response, as there was less variability compared to peak area, due to the noticeable tailing of these compounds. The linearity results (R2) are tabulated in Tables 3 and 4.

11

Table 3.

Recoveries of Sulfonamides by Diluting Extracts 1 in 4 with Water Amount recovered (ng) STZ 167 168 160 158 151 147 144 150 156 9 29 94 6 104 0.9997 3.14 SDZ 172 197 183 189 169 161 72 150 178 13 40 126 7 118 0.9996 3.14 SPY 164 68 158 167 154 144 141 150 157 11 34 108 7 104 0.9997 3.14 SMR 317 343 315 336 295 322 272 300 314 24 77 245 8 105 0.9972 3.14 SMZ 151 164 157 156 169 143 151 150 156 9 28 88 6 104 0.9996 3.14 SCPD(IS) 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 100 1.0000 3.14 SQ 148 169 133 138 133 120 124 150 138 17 53 167 12 92 0.9984 3.14 SDMX 130 137 121 129 129 112 125 150 126 8 26 82 7 84 0.9992 3.14

Description Pork spike 1 Pork spike 2 Pork spike 3 Pork spike 4 Pork spike 5 Pork spike 6 Pork spike 7 Amount spiked (ng) Mean SD (Precision) MDL (SD t-stat) ng LOQ (SD 10) ng RSD (SD 100/Mean) Accuracy (%) Linearity (R ) t-stat (N=7)
2

Table 4.

Recoveries of Sulfonamides Using Oasis HLB Cleanup Cartridges Amount recovered (ng) STZ 161 154 149 145 151 136 148 150 149 8 24 76 5 99 0.9994 3.14 SDZ 157 156 158 152 162 147 161 150 156 5 17 53 3 104 0.9994 3.14 SPY 132 132 124 122 127 127 128 150 127 4 11 36 3 85 0.9997 3.14 SMR 273 293 267 279 294 274 275 300 279 10 33 104 4 93 0.9979 3.14 SMZ 149 157 155 144 149 136 155 150 149 7 23 73 5 100 0.9998 3.14 SCPD(IS) 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 100 1.0000 3.14 SQ 139 153 132 119 127 116 124 150 130 13 40 128 10 87 0.9989 3.14 SDMX 126 131 113 111 121 108 116 150 118 8 26 82 7 79 0.9989 3.14

Description Pork spike 1 Pork spike 2 Pork spike 3 Pork spike 4 Pork spike 5 Pork spike 6 Pork spike 7 Amount spiked (ng) Mean SD (Precision) MDL (SD t-stat) ng LOQ (SD 10) ng RSD (SD 100/Mean) Accuracy (%) Linearity (R ) t-stat (N=7)
2

12

Table 5 summarizes the comparison of recoveries when diluted with water versus using Oasis HLB cartridge cleanup. Generally there is a greater difference in recoveries for the early eluting compounds, as one might expect. Since the samples are loaded onto the cartridge with a mostly aqueous phase (10% methanol in water), the water-soluble matrix components would tend to pass through the cartridge to waste. Because these early eluting compounds were removed prior to injection on the HPLC column, the chromatograms are cleaner with more reproducible chromatography, as shown by the smaller standard deviations in recoveries. The results from the HLB cleanup exhibited smaller standard deviations and lower minimum detection levels (MDLs).

when SPE cleanup is used. Instrumental conditions allow injection cycle-time of 10 minutes using typical columns and conditions for most labs.

References
1. TLC-Densitometric Procedure for Sulfonamide Residues in Animal Tissue, SUL-SP08, Canadian Food Inspection Agency, Saskatoon, Saskatchewan, Canada; 2001/04. 2. Sulfonamides in Tissue by LC/MS, Alberta Agriculture, Edmonton, Alberta, Canada, Standard Operating Procedure TX-0278-01. 3. Mark Stahl, High-throughput analysis with the Agilent 1100 Series high-throughput LC/MS system, Agilent Technologies, publication 5988-9638EN. www.agilent.com/chem

Conclusion
A fast and sensitive single quadrupole LC/APCI/MS method was developed and validated for detection of sulfonamide residues in pork. The DL ranged from 10 to 25 ng/g of tissue when analyzed by simple dilution of the extracts, and 4 to 13 ng/g

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

Table 5.

Comparison of Recoveries Obtained by Dilution vs Oasis HLB Cleanup STZ 104 9.4 29 99 7.6 24 SDZ 118 12.6 40 104 5.3 17 SPY 104 10.8 34 85 3.6 11 SMR 105 24.5 77 93 10.4 33 SMZ 104 8.8 28 100 7.3 23 SCPD(IS) 100 100 SQ 92 16.7 53 87 12.8 40 SDMX 84 8.2 26 79 8.2 26

Description Accuracy % (1 in 4 dilution) SD (Precision) MDL (ng) Accuracy % (HLB cleanup) SD (Precision) MDL (ng)

13

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2003 Printed in the USA October 20, 2003 5989-0182EN

Detection, Confirmation, and Quantification of Chloramphenicol in Honey and Shrimp at Regulatory Levels Using Quadrupole and Ion Trap LC/MS Application

Foods, Environmental

Authors
Gerd Vanhoenacker, Frank David, Pat Sandra Research Institute for Chromatography Kennedypark 20, B-8500 Kortrijk, Belgium Non-Author Contact Jerry Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road, Wilmington DE 19808-1610 USA e-mail: Jerry_Zweigenbaum@Agilent.com

Abstract
Methodology capable of meeting regulatory requirements has been developed for the determination of chloramphenicol in honey and shrimp. Samples of the two foodstuffs are extracted with Isolute HN-M cartridges and analyzed with both the Agilent 1100 LC/MSD Trap (SL) and the Agilent 1100 LC/MSD (SL) quadrupole with negative mode electrospray ionization. Using deuterated internal standard and one simple sample extraction procedure, both instruments provide a limit of detection at or below 0.1 ppb in both shrimp and honey. Detection limits are lower using the ion trap for shrimp because of less matrix interference. The Agilent 1100 LC/MSD gives quantitative results and the Agilent 1100 LC/MSD Trap gives full spectrum confirmation.

and shrimp. Because it has displayed significant toxicological effects on humans, it has been banned from foods in the European community and the United States at levels greater than 0.1 ppb. Analytical methods used to determine this limit must achieve both the required sensitivity and maintain sufficient selectively. LC/MS has been demonstrated by the US Food and Drug Administration for these analysis [1-3]. In addition, the Commission of European Communities has issued guidelines stipulating that for mass spectral detection, a molecular ion (or quasimolecular ion) and at least two fragment ions are needed for positive confirmation [4]. For quantitative analysis the Agilent 1100 LC/MSD provides excellent results and can give some confirmation information. The Agilent 1100 LC/MSD Trap gives excellent full spectrum confirmation at the regulated concentration.

Experimental
Reagents and Materials ISOLUTE HM-N cartridges from IST (Hengoed, UK, Part-nr. 800-1300-FM) Ethyl acetate from Vel (Merck Eurolab, Leuven, Belgium) Methanol HPLC-grade from Merck (LiChrosolv, Darmstadt, Germany) Deuterated (d5) CAP internal standard from Cambridge Isotope Laboratories (CIL, Andover, MA, USA) Syringe filters (0.2 m, PTFE) from Alltech Associates Inc. (Lokeren, Belgium)

Introduction
Chloramphenicol is a broad range antibiotic that has found its way into foodstuffs such as honey

Sample Preparation For honey, 5 g of sample is diluted to 20 mL with water and 5 L of 1 ng/L internal standard (IS) is added. The solution is loaded on the cartridge and allowed to stand for 5 minutes. Elution is performed with 50 mL ethyl acetate. The eluate is collected and the solvent is evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL water/methanol (9/1, v/v) and put in an ultrasonic bath for 1 minute. The solution is filtered, using a syringe filter, before injection. No additional clean-up of the sample solution is performed. For shrimp, a portion of at least 10 g of frozen shrimp is defrosted and mixed in a blender. To 10 g of the mixed shrimp, 30 mL of water and 10 L of

1 ng/L IS is added. This portion is centrifuged for 10 minutes (2000 rpm). A 20-mL portion of the supernatant is loaded on the cartridge and allowed to stand for 5 minutes. Elution is performed with 50 mL ethyl acetate. The eluate is collected and the solvent evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL water/methanol (9/1, v/v) and put in an ultrasonic bath for 1 minute. The solution is filtered before injection. LC/MS Conditions The LC/MS systems were the Agilent 1100 LC/MSD quadrupole mass spectrometer and the Agilent 1100 LC/MSD Trap. Both were equipped with Agilent 1100 binary pumps and 1100 well plate autosamplers. See Table 1.

Table 1. HPLC Column Flow-rate

LC/MS Conditions

Eclipse XDB C18, 4.6 mm 150 mm, 5 m (p/n 993967.902) 0.9 mL/min 10 mM ammonium acetate in water (solvent A) Methanol/acetonitrile 1/9 (solvent B) both from Merck (LiChrosolv, Darmstadt, Germany) 01 min 18 min 88.5 min 8.512 min Post time 30% B 30%70% B 70%100% B 100% B 4 min at 30% B

Mobile phase

Gradient

Injection Injection solvent Column temperature MSD source settings Source Ion polarity Drying gas temperature Drying gas flow-rate Nebulizer pressure Vcap Quadrupole MSD MSD acquisition on Fragmentor SIM settings

100 L with needle wash (methanol) Water/methanol (9/1 v/v) for both standards and samples 30 C ESI Negative 340 C 11 L/min 50 psig 3500 V Between 3 and 7.5 min 160 V m/z 257, 321, 323 (CAP) m/z 262, 326, 328 (CAP-d5)

Table 1. Trap MSD

LC/MS Conditions (continued)

MSD acquisition on Target mass (SPS) Trap parameters Max. accumulation time ICC target Scan range Averaging

Between 3 and 7.5 min 323 m/z

300 ms 30,000 160340 2

Fragmentation parameters (MS/MS) Smart Frag Isolation mass Isolation width Fragmentation amplitude Fragmentation cutoff On, 30%200% (default) m/z 325.0 10.0 m/z 1.0 V m/z 88

Results and Discussion


Spectral Quality and Sensitivity of Standards For analysis with the quadrupole LC/MSD, selected ion monitoring (SIM) was used to obtain the required sensitivity. Table 2 shows the structure, fragment ions and identity of CAP and CAP-d5. Figure 1 shows the analysis of a standard mixture containing 2.5 pg/L CAP and 5 pg/L CAP-d5. By applying a fragmentor voltage of 160 V, fragment ions at m/z 257 and 262 are detected for confirmation purposes. Lowering the fragmentor voltage to optimize for the m/z 321 and m/z 326 and monitoring those ion alone would obtain greater sensitivity. However, the confirmation of the fragment ions would be lost. For screening analysis without confirmation this would be acceptable and provide a much lower limit of detection (LOD).

Table 2

Structure and Fragment Ions and Identity of CAP and CAP-d5 (* Indicates Deuterated Positions for the CAP-d5 IS)

Chloramphenicol structure OH * * O N+ * O* * H N Cl OH O Cl

m/z CAP 257 249 194 176 262 254 199 180

Identity [M-H-HCOCl] [M-H-2HCl] [M-H-NH2CoCl2H] [M-H-NH2CoCl2H-H2O] [M-H-HCOCl] [M-H2HCl] [M-H-NH2COCl2H] [M-H-NH2COCl2H-HDO]

CAP-d5

7000

CAP (2.5 pg/L)

321 m/z

4000

323 m/z

257 m/z
1000

14000

CAP-d5 (5 pg/L)

326 m/z

8000

328 m/z

2000

262 m/z

Time (min)

Figure 1.

Analysis of a standard solution containing 2.5 ppb of CAP and 5 ppb of CAP-d5 (IS) on the quadrupole MSD. The extracted ion chromatogram for the corresponding ions are shown.

Using the LC/MSD Trap in MS/MS mode both the needed sensitivity (through reduction in chemical noise) and selectivity (for confirmation) is obtained. The compound shows a clear and reproducible fragmentation pattern. An example of the analysis of the standard mixture together with the corresponding MS/MS spectra is shown in Figure 2. Optimizing the fragmentation energy [turning off Smart Frag] and fragmentation cutoff in the ion trap will increase sensitivity even further than shown here. Using an isolation width of 10 m/z allows inclusion of the chlorine isotopes in the resulting full scan mass spectra of the analyte and the Cl35 isotope of the internal standard. Contact Agilent for more details on these and other ion trap settings.

CAP: EIC 176, 194, 249, 257 m/z (2.5 pg/L)


1000

2000

CAP-d5: EIC 180, 199, 254, 262 (5 pg/L)

1000

0 4

Time (min)

2.5 pg/L CAP

257

194

5 pg/L d5-CAP6 262 249

176
160 240

0.2 pg/L CAP (LOD)

257 180
160

199
240

254
320 m/z

176

194 249

160

240

320

m/z

Figure 2.

Analysis of a standard solution containing 2.5 pg/L CAP and 5 pg/L CAP-d5 (IS) on the LC/MSD Trap together with the corresponding MS/MS spectra and the MS/MS spectrum resulting from an analysis of a standard solution containing 0.2 pg/L CAP.

Method Performance Standard solutions of CAP containing 5 pg/L of CAP-d5 were injected six consecutive times to test repeatability of injection on the mass selective detector (MSD) quadrupole instrument. This was done at two concentration levels. Each time, the response of CAP relative to CAP-d5 was recorded. For a solution containing 0.5 pg/L CAP the relative standard deviations (RSDs) on the relative response were 5.05%. This 0.5-pg/L level would correspond to a sample containing approximately 0.1 ppb CAP with the five-fold concentration step. When a solution containing 5 pg/L CAP was analyzed, RSDs on the relative response were 1.28% for the quadrupole.

A calibration line was constructed by injecting standard solutions of CAP with a concentration of 0 to 25 pg/L with 5 pg/L of the IS added to each solution. One injection was performed per concentration. The quadrupole showed a linear response for CAP in this concentration range. Calibration curves and correlation coefficients are shown in Figure 3. The LOD with this method was determined to be ca. 0.2 pg/L in a standard solution for both mass spectrometers. With the 100-L injection used, this corresponds with 20 pg on-column.

8.E+05

CAP (without IS)

R2 = 0.9994

CAP/CAP d5 (with IS) R2 = 1

4 6.E+05 3 4.E+05 2 2.E+05

0.E+00 0 5 10 15 20 25 30 Concentration CAP (ppb)

0 0 5 10 15 20 25 30 Concentration CAP (ppb)

Figure 3.

Calibration graphs for standard solutions of CAP on the quadrupole with and without CAP-d5 (IS).

Extraction Recovery and Repeatability of Extraction The extraction procedure was evaluated on repeatability and linearity with the quadrupole instrument. Blank honey was spiked with 1 ppb CAP and 1 ppb CAP-d5. The extraction procedure was carried out six times and the recovery was calculated. The recovery for CAP varied from 85.31% to 94.94% and the mean recovery was 90.60%. The RSD on the recovery was 4.34% for CAP and 3.39% when the IS was taken into account. An analysis of blank honey spiked only with the IS is shown in Figure 4 run on both instruments. With the quadrupole, LC/MSD matrix interferences are present but chromatographically separated from the CAP signal. The ion trap results show that no matrix interference is present in the isolation window from m/z 318 to 328. The data suggest that other endogenous compounds in honey produce fragments at the same m/z as CAP. This supports an even lower detection limit for this matrix if a screening analysis were conducted with a lower fragmentor voltage monitoring only the m/z 321.

60000

TIC
30000

15000

EIC 257, 321, 323 m/z (CAP)

QUADRUPOLE

7500

12000

EIC 262, 326, 328 m/z (CAP-d5)


6000

0 4 5 Time (min) 15000 6 7

EIC 176, 194, 249, 257 m/z (CAP)

0 15000

TRAP

EIC 180, 199, 254, 262 m/z (CAP-d5)

0 4 5 Time (min) 6 7

Figure 4.

Analysis of a blank honey sample containing 1 ppb CAP-d5.

A calibration curve was constructed with blank honey samples spiked with 0, 0.1, 0.2, 0.5, 1.0, and 2.0 ppb CAP. The samples also contained 1 ppb of the IS. The correlation coefficients were 0.9997 and 0.9998 without and with correction with the IS, respectively. The slope for the calibration curve constructed with these extracts for CAP with correction with the IS was 0.1822. This is in good agreement with the slope obtained with the standard solutions, which is 0.1758 (see Figure 3).

Spectra on the trap were similar for standard solutions and real samples. An example of an MS/MS spectrum of an extract of a honey sample spiked with 0.5 ppb CAP and 1 ppb CAP-d5 is shown in Figure 5. Since the analyte and the IS coelute, a mixed spectrum is obtained. This could be avoided by using a smaller isolation width and the multiple reaction monitoring (MRM) function of the ion trap. Note that the chlorine isotope for Cl35Cl37 is not observed for the deuterated internal standard because its precursor ion is at the edge of the isolation width and thus not trapped.

262

257 194 176 180


160

249 199
240

254
320 m/z

Figure 5.

Ion trap MS/MS spectrum from analysis of a honey sample spiked with 0.5 ppb CAP and 1 ppb CAP-d5.

Analysis of Honey The extraction procedure and LC/MS methods were applied to the analysis of honey samples that were known to contain CAP. Sample results obtained with the quadrupole and trap MSD were compared (Figure 6).

80000 40000 0 14000

TIC

QUADRUPOLE

EIC 257, 321, 323 m/z (CAP)


8000 2000 15000

EIC 262, 326, 328 m/z (CAP-d5)


7500 0 4 5

Time (min)

EIC 176, 194, 249, 257 m/z (CAP)


15000

TRAP

0 15000

EIC 180, 199, 254, 262 m/z (CAP-d5)

0 4 5

Time (min)

Figure 6.

Analysis of a honey sample containing 0.5 ppb CAP and 1 ppb CAP-d5.

The LOD for the honey samples varies between detectors. For the quadrupole, it is found to be 0.5 pg/L in the analytical solution. This corresponds with 50 pg on-column. Taking into account the sample preparation with a five-fold concentration, samples containing 0.1 ppb CAP can be detected. It is obvious that the sample matrix interferes with the sensitivity (Figures 4 and 6). Due to the increased selectivity using MS/MS in the trap, the LOD with this MS is similar for honey samples as for the standard solutions and is ca. 0.2 pg/L in the analytical solution. This is equivalent to 0.04 ppb CAP in the sample because of the five-fold concentration step.

Analysis of Shrimp The same sample preparation method was applied to the analysis of shrimp. The total volume of shrimp and water added was about 40 mL. Taking 20 mL of the 10 g shrimp aliquot for the Isolute sample preparation and reconstituting the dried extract in 1 mL produced a five-fold concentration as with the honey. This sample preparation shows less matrix interference with the analysis compared to honey samples. An example of an analysis of shrimp is shown in Figure 7. Due to the reduced matrix effect, the LOD with the quadrupole is lowered to nearly the same level as for the trap (0.05 ppb in the sample with the five-fold concentration). A concentration of 0.35 ppb was recovered in the shrimp sample by both the quadrupole and the trap MSD. Extraction recovery was approximately 85%.

30000

TIC
15000

7000

QUADRUPOLE

EIC 257, 321, 323 m/z (CAP)


4000

16000

EIC 262, 326, 328 m/z (CAP-d5)


8000

Time (min)

10000

EIC 176, 194, 249, 257 m/z (CAP)

TRAP

0 10000

EIC 180, 199, 254, 262 m/z (CAP-d5)

0 4 5

Time (min)

Figure 7.

Analysis of a shrimp containing 0.35 ppb CAP and 1 ppb CAP-d5.

www.agilent.com/chem

Conclusion
Honey and shrimp samples were successfully analyzed for CAP with both the quadrupole and trap MSD. A simple liquid-liquid extraction procedure using ISOLUTE HM-N cartridges was found to perform excellently in view of recovery and repeatability. The LC method used a standard 4.6-mm id column and produced the required sensitivity on both instruments. The LC/MSD quadrupole instrument produced excellent linearity and demonstrated its quantitative ability. The LC/MSD Trap showed the needed sensitivity with excellent full scan capability below the regulated limit in both sample matrices. The use of a broad isolation window for full scan spectra using the ion trap produced more transition ions than required for confirmation.

References
1 S. Turnipseed, et al. (2002) Confirmation of Multiple Phenicol Residues in Honey by Electrospray LC/MS, Laboratory Information Bulletin (4281) U.S. Food and Drug Administration. 2 A. Pfenning, et al. (2002) Confirmation of Multiple Phenicol Residues in Shrimp by Electrospray LC/MS, Laboratory Information Bulletin (4284) Food and Drug Administration. 3 B. K. Neuhaus, et al. (2002) LC/MS/MS Analysis of Chloramphenicol in Shrimp, Laboratory Information Bulletin (4290) Food & Drug Administration. 4 D. Byrne, (2002) Performance Criteria, other Requirements and Procedures for Analytical Methods. Official Journal of European Communities L221, 1417.

For More Information


For more information on our products and services visit our web site at www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2003 Printed in the USA August 6, 2003 5988-9920EN

Determination of Chloramphenicol in Fish Meat by Liquid Chromatograph-Atmospheric Pressure Photo Ionization-Mass Spectrometry (LC-APPI-MS) Application

Foods, Environmental

Author
Masahiko Takino and Shigeki Daishima Yokogawa Analytical Systems Inc 2-11-13, Nakacho, Musashino Tokyo, 180-8453 Japan

Abstract
A liquid chromatography-atmospheric pressure photoionization-mass spectrometry method was developed for the determination of chloramphenicol antibiotics in fish meats. For the optimization of APPI, several ion source parameters were examined. Using the optimized parameters, simple mass spectra and a strong signal cor responding to [M-H] was observed. The samples were extracted with ethylacetate and evaporated to dryness followed by a clean-up step using liquid-liquid distribution by acetonitrile and n-hexane. Mean recoveries of chloramphenicol from young yellowtail meat and flatfish meat spiked at 0.12 ng/g were 89.3%102.5% and 87.4%94.8%, respectively. The limit of detection (signal-to-noise = 3) of the young yellowtail meat and the flatfish meat were 0.27 and 0.10 ng/g.

therapy can result in a well-understood and irreversible type bone marrow depression called aplasia or hypolasia. This, in turn, can lead to aplastic anemia and although uncommon, it is often fatal. Because of these health concerns, a joint Food and Agriculture Organization/World Health Organization (FAO/WHO) Expert Committee on Food Additives has proclaimed that CAP residues in the human food supply are unacceptable [1]. The use of CAP in food products has been banned in EU and U.S.A. However, CAPs broad-spectrum activity, ready availability, and low cost attract its use by some third world countries. Admittedly, whenever CAP is accessible, indiscriminate and illegal use potentially exists. In fact, the presence of CAP has been detected in shrimp imported from China and Vietnam that was intended for human consumption. Liquid chromatography/mass spectrometry (LC/MS) methods are very useful in analyzing CAP in food because of the high selectivity and sensitivity of MS detection [2-7]. Atmospheric pressure ionization (API) interfaces, represented by atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI), are commonly used in LC/MS. Atmospheric pressure photoionization (APPI) is a new ionization technique for LC/MS [8, 9]. The APPI source is based on a high-fluence gas discharge lamp that generates vacuum-ultraviolet (VUV) photons of 10 and 10.6 eV energy. The energy of this discharge lamp is normally greater

Introduction
Chloramphenicol (CAP) is a broad-spectrum antibiotic, that exhibits activity against a variety of aerobic and anaerobic microorganisms. Its action works through interference with or inhibition of protein synthesis. However, weeks or months of CAP

than a first ionization potential (IP) of an analyte because many organic compounds have IPs in the range of 710 eV. On the other hand, the IPs of the most common LC solvents, which are used as a mobile phase, have higher values (water, IP = 12.6 eV; methanol, IP = 10.8 eV; acetonitrile, IP = 12.2 eV). This provides ionization of many analytes with lower IPs without interference from the mobile phase. To our knowledge, APPI has not yet been applied to residual analysis in food. This application note describes how parameters affect the ionization efficiency of APPI for the analysis of CAP. In addition, the suitability of LC/MS and liquid-liquid extraction using the APPI technique is evaluated for the determination of CAP in fish meat.

step was repeated with another 1 mL of n-hexane. Finally, the acetonitrile phase was evaporated to dryness under a stream of dry nitrogen using a heating block at 50 C, redissolved in 5 mL of a 10% acetonitrile in 10 mM ammonium acetate water solution, and filtered through a 0.22 m nylon centrifuge filter. The samples were spiked with 0.1100 ng/mL of CAP after the homogenation step to generate a calibration by LC/APPI-MS selected ion monitoring (SIM). LC/MS An Agilent 1100 series LC, consisting of a vacuum solvent degassing unit, a binary high-pressure gradient pump, a standard automatic sample injector, and a column thermostat, was used for the separation. An 1100 series diode array detector (DAD) was connected in line with an 1100 MSD for detection and confirmation. See Table 1. The separation was performed on a 150 3 mm id column packed with 5 m Zorbax Eclipse XDB C18 (Agilent Technologies, Palo Alto, USA). A 15-min linear solvent gradient was used for elution with the mobile phase. Quantitative analysis was carried out using SIM of m/z 321 with a dwell time of 500 msec. The following six parameters were optimized using
Table 1. LC: Column: Solvent A: Solvent B: Dopant: Gradient: Column temp: Sample volume: Flow rate: MS: Ionization: Scan range: SIM ion: Drying gas: Nebulizer gas: Fragmentor: Capillary: Vaporizer temp: Instrument Parameters 1100 series LC Zorbax Eclipse XDB C18 (150 mm 3 mm, 5 m) Water with 10 mM ammonium acetate Methanol Acetone at 0.05 mL/min 90/10 A/B 15 min to 70/30 A/B 40 C 20 L 0.5 mL/min 1100 MSD, SL APPI (Negative) m/z 100400 for optimization m/z 321; (M-H)
_

Experimental
Chemicals and Solvents CAP was purchased from Sigma-Aldrich Japan (Tokyo, Japan). The purity of this compound was greater than 99%. Stock solutions at 1 mg/mL were prepared in methanol, stored in the dark at 4 C, and diluted to the desired concentrations prior to use. Ammonium acetate, pesticide-grade ethyl acetate, anhydrous sodium sulfate, acetonitrile, HPLC-grade methanol and n-hexane were obtained from Wako Chemical (Osaka, Japan). Water was purified with a Milli-Q system (Millipore, Tokyo, Japan). A nylon-type 0.22 m centrifuge filter was obtained from Toyo Soda (Tokyo, Japan). Sample Preparation The samples analyzed (young yellowtail and flatfish) were obtained from a local market. To a centrifuge tube, 5 g fish meat and 5 g anhydrous sodium sulfate were weighed and 10 mL ethyl acetate was added. The mixture was homogenized for 20 s with an Ultra-Turrax TP 18/10 (Janke & Kunkel KG, Staufen, Germany). After centrifugation for 5 min at 6000 rpm, the supernatant was removed and transferred to a round flask. The extraction step was repeated twice, each with 10 mL ethyl acetate. The combined ethyl acetate extract was then evaporated in a rotary evaporator at 40 C under vacuum. One mL acetonitrile and 1 mL n-hexane was added to the residue, transferred into a graduated glass stopper reagent bottle, and shaken. The n-hexane phase was discarded. The

Nitrogen, 7 L/min at 350 C Nitrogen, 50 psi 120 V 3500 V 350 C

the analytical column with CAP at 100 ng/mL: the voltages for in-source-fragmentation (the fragmentor voltage), the capillary voltage (Vcap), the drying gas flow rate, the nebulizer pressure, the mobile phase composition, and the mobile phase flow rate. The ion lens voltages in the MS were automatically optimized using a Calibrant Delivery System and the AutoTune program. Negative ion mass spectra were acquired over the scan range m/z 100400 using a step size of 0.1 amu and a scan rate of 2 s per scan for the optimization of fragmentor voltage.

was found that modification of drying gas flow rate and nebulizer gas pressure did not drastically improve the sensitivity of CAP. In addition, the fragmentor voltage was included in optimization because of its compound dependence and its significant effect on the mass spectral response. Effect of Capillary Voltage The capillary voltage is applied to the inlet of the capillary and influences the transmission efficiency of the ions through the capillary sampling orifice. To establish the optimum capillary voltage, this parameter was varied from 1000 to 4000 V. As shown in Figure 1, 1500 V was found optimum. A tremendous effect of this parameter on the intensity of CAP was observed in the case where acetone was not used as the dopant. On the other hand, when acetone was introduced into the APPI source as the dopant, the maximum intensity of the ion was found at 3500 V. The intensity found at 3500 V with the dopant was higher than the maximum intensity without the dopant. Based on the above results, the capillary voltage was set at 3500 V with acetone.

Results and Discussion


Optimization of the APPI Parameters To optimize the APPI conditions, parameters that influence the ionization efficiency were investigated. The drying gas flow, the nebulizer gas pressure, the vaporizer temperature, the capillary voltage, and the mobile phase composition were evaluated under the chromatographic conditions mentioned in the Experimental section by SIM mode using the m/z 321 ion as the target ion. It

3000000

2500000

2000000 Peak intensity

1500000

Without acetone as the dopant With acetone as the dopant


1000000

500000

0 1000

1500

2000

2500 Capillary voltage [V]

3000

3500

4000

Figure 1.

The effect of the capillary voltage on the peak intensity of CAP concentration : 1 ng/mL. For the other conditions, see Experimental section.

Effect of Vaporizer Temperature In APPI, the vaporizer temperature plays a key role for the complete evaporation of CAP because ionization occurs in the vapor state like APCI. Thus, in the case of using linear gradient elution, this temperature must be kept sufficiently high so that the change of mobile phase composition does not influence the ion intensity of CAP. Under high temperature, however, the risk of thermal degradation occurs. In this study, the vaporizer temperature was modified between 250 and 450 C to optimize the intensity and the S/N ratio. The highest temperature for a maximum intensity and S/N ratio of CAP was observed at 350 C. The intensity of CAP decreased as the vaporizer temperature was increased over 400 C. In addition, intense fragmentation was observed in the mass spectrum at 400 C. Therefore, the decrease in intensity above 400 C seems to be a result of the thermal degradation. Based on the above results, the vaporizer temperature was set at 350 C.

Optimization of Fragmentor Voltage The fragmentor voltage is applied to the exit of the capillary and affects the transmission and fragmentation of sample ions between the exit of the capillary and the skimmer at relatively high pressure (3 torr). In general, the higher the fragmentor voltage (which helps the transfer of ions), the more fragmentation will occur. To establish the optimum fragmentor voltage for the analysis of CAP, the intensity of this compound versus the fragmentor voltage was studied in the range from 80 to 200 V. As shown in Figure 2, the optimum fragmentor voltage was found at 120 V, whereas at higher values a significant intensity reduction was observed. Further, the best S/N ratio was also observed at 120 V. The mass spectra of CAP at optimal and higher fragmentor voltages are shown in Figure 3. The deprotonated molecule (M-H) was the predominant ion at 120 V, and this included isotopic ions (m/z 321, Cl35 Cl35; m/z 323, Cl35 Cl37; m/z 325, Cl37 Cl37) because CAP includes two

2500000

2000000

1500000 Peak intensity 1000000 500000

0 80 100 120 140 Fragmentor voltage [V] 160 180 200

Figure. 2. The effect of the fragmentor voltage on the peak intensity of CAP concentration : 1 ng/mL. For the other conditions, see Experimental section.

chlorines. A higher fragmentor voltage (180 V) generated structurally relevant fragment ions. The m/z 152 fragment ion gives the greatest intensity and might be produced by the cleavage of the carbon-carbon bond on the alkyl branch as shown in Figure 3. Other fragment ions are observed at m/z 121 and 257. The m/z 121 may be the nitrophenyl fragment. The m/z 257 fragment might be explained by a charge migration hydrogen shift with a concerted loss of HCl and CO. These observed fragment ions in the APPI source corresponded with the fragment ions in an ESI source and an APCI source. Based on the above results, the fragmentor voltage was set to 120 V.

Optimization of the Chromatographic Conditions The separation of CAP from sample matrix peaks was optimized using acetonitrile, methanol, water, and ammonium acetate. The combination of methanol and ammonium acetate was found optimum for the separation of CAP. When methanol was replaced with acetonitrile, a significant signal intensity and S/N decrease was observed. This result indicates that methanol may be a source of electrons for the hydrogen abstraction from CAP. Therefore, methanol and 10 mM ammonium acetate was used as the mobile phase in this study. The flow rate was set at 0.5 mL/min considering the size of the used column.

321

600000

(A). Low fragmentor voltage (120 V)

400000

200000

100000 100 150 200 250 300 350 400

(B). High fragmentor voltage (180 V)


250000 152 200000

321

121 O2N

H CH OH

NH

CO

CHCl2

150000

CHCH2OH

100000

152
257 121

50000

0 100 150 200 250 300 350 400

Figure 3.

The mass spectra of CAP at two different fragmentor voltages.

Linearity, Detection Limit and Precision of LC/APPI-MS System The analytical performance characteristics of the optimized LC/APPI-MS were first determined on standard solutions of CAP in pure solvent. See Figure 4. In order to achieve optimum sensitivity, all experiments were carried out under SIM mode using the mass corresponding to the [M-H] ions for CAP. To test the linearity of the calibration curves, various concentrations of CAP ranging from 0.1 to 100 ng/mL were analyzed. The calibration curves of APPI showed good linearity with correlation coefficients (r2) = 0.9998. The repeatability of APPI for a standard solution was calculated on the basis of five replicates at 0.5 ng/mL in the same day. The limit of detection (LOD) was calculated by using a S/N ratio of 3 at 0.1 ng/mL. The SIM chromatogram of CAP with APPI is shown in Figure 4 (the S/N ratio of this chromatogram was 4.2); LOD and RSD were 0.07 ng/mL and 2.1%.

900

700

500

300

10 Retention time [min]

15

20

Figure 4.

SIM chromatogram of CAP in pure solvent at 0.1 ng/mL with APPI.

APPI Method Evaluation To evaluate recoveries, the proposed method was applied to the analysis of spiked CAP-free samples of young yellowtail and flatfish meat. Eighteen samples of two different fish were each spiked with CAP and each sample was spiked at three levels. The spiking levels ranged from 0.1 to 2 ng/g. Typical chromatograms from the fish meat extracts spiked at 1 ng/g and 0.1 ng/g are shown in Figure 5.

20000 15000 10000 5000 0 16000 12000 8000

(A)

10

15

20

(B)

0 12000 8000 4000

10

15

20

(C)

0 10000

10

15

20

(D)

6000

2000 0 5 10 Retention time [min] 15 20

Figure 5.

SIM chromatograms of A) Young yellowtail meat, B) Spiked young yellowtail meat at 1 ng/g CAP, C) a flatfish meat, and D) a spiked flatfish meat at 0.1 ng/g CAP.

Data from 18 spiked samples led to recoveries and RSD are summarized in Table 2.

Table 2

Recovery of CAP for Spiked Fish Meat Recovery [RSD (%)]* Young yellowtail Flatfish 89.3 5.1 102.5 4.9 96.1 4.3 87.4 6.1 94.8 6.7 91.8 4.9

Spiking levels (ng/g) 0.1 0.5 2.0

*Three spiked samples at the same amount were analyzed.

Mean recoveries ranged from 87.4% to 102.5% with RSD of 4.3% to 6.7%. The LODs of CAP in fish meats were determined by the signal corresponding to three times the background noise on SIM chromatogram of spiked sample at 1 ng/g and 0.1 ng/g and shown in Table 3. The intraday precision (repeatability) was estimated by injecting the same spiked fish meat extract at 0.1 ng/g five times during a working day. The interday precision (reproducibility) was evaluated by analyzing the same sample over 5 working days. The repeatability and reproducibility for CAP in fish meats were 4.8%, 9.4% and 2.1%, 7.3%, respectively. These results indicate that this LC/APPI-MS method is suitable for the analysis of residues of CAP in fish meats.

Table 3.

LODs, Repeatability, and Reproducibility of CAP in Standard Solution Using APPI LODs* Repeatability** Reproducibility*** Fish meats (ng/g) (RSD, %) (RSD, %) Young yellowtail 0.27 4.8 9.4 Flatfish 0.10 2.1 7.3
*Detection limit is LOD defined as S/N = 3 at 0.1 ng/mL. **Repeatability was calculated on the basis of five replicates at 0.5 ng/mL within 1 day. ***Reproducibility was calculated by analyzing one fish meat spiked at 0.1 ng mL1 per day during 5 days.

Conclusion
APPI is an ideal ionization technique because of high sensitivity and high selectivity for the determination of CAP in fish meats. An important advantage of using APPI for CAP content of fish meats is that sample matrix did not significantly affect ion intensity of CAP. The data presented here demonstrate that this method is convenient for routine analysis of CAP residues in fish meats at trace levels, as excellent recoveries and precision for different samples were obtained.

References
1. A. H. Allen J. AOAC Int. 1985, 68, 990. 2. T. L. Li; Y.J. Chung-Wang; Y. C. Shih J. Food Science 2001, 67, 21. 3. B. Roudaut J. Liq. Chrom. & Rel. Technol. 1994, 19, 1097. 4. C. N. Kenyon; A. Melera; F. Mrmi J. Anal. Toxicol. 1981, 5, 216. 5. V. Hormazabal; M. Yndestad J. Liq. Chrom. & Rel. Technol. 2001, 24, 2477. 6. K. Richard; V. Kruft; H. Sommer LaborParaxis 2000, 24, 91. 7. D. G. Kennedy; R. J. McCracken; A. Cannavan; S. A. Hewitt J. Chromatogr. A. 1998, 812, 77. 8. D. B. Robb; T. R. Covey; A. P. Bruins Anal. Chem. 2000, 72, 3653. 9. J. A. Syage; M. D. Evans; K. A. Hanold Amer. Lab. 2000, 32, 24.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2003 Printed in the USA April 9, 2003 5988-8999EN

Determination of the Metabolites of Nitrofuran Antibacterial Drugs in Chicken Tissue by Liquid Chromatograph-Electrospray Ionization-Mass Spectrometry (LC-ESI-MS) Application

Food, Environmental

Author
Masahiko Takino Yokogawa Analytical Systems Inc 2-11-13, Nakacho, Musashino, Tokyo, 180-8453 Japan

additives to prevent bacterial enteritis by Escherichia coli and Salmonella in cattle, fish, swine, and poultry. The occurrence of furazolidone residue in edible tissue is a major human health concern. Effective June 1995, these drugs were banned from use in food animal production in the European Union (EU) because of concerns about their carcinogenicity and mutagenicity (Commission Regulation 1442/95). Nitrofuran antibacterial drugs are characterized by their rapid metabolism, with in vivo half-lives of less than a few hours. Therefore, the detection of parent drugs in animal tissue is not practical. Studies using radioactive-labeled furazolidone have shown that protein-bound metabolites are formed in tissues [1-3]. The tissue-bound metabolites are detectable for several weeks after administration. Hence, the analysis of nitrofuran drugs is based on the detection of the tissue-bound metabolites of the parent drugs. These tissue-bound metabolites are very small molecules which are not UV absorbing, and they elute too quickly out of a column. To induce UV absorption in the molecule and to be reasonably retained on a column, they are derivatized. It is possible to release these metabolites from the proteins under moderately acidic conditions and derivatize the metabolites with 2-nitrobenzaldehyde (2-NBA) to produce 2-NBA-derivatives for liquid chromatography (LC), UV detection, and mass spectrometry (MS) confirmation. The goal of this study is to develop a routine analytical method to simultaneously detect the target nitrofuran metabolites. Because no maximum residue limit (MRL) has

Abstract
A liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS) method was developed for the simultaneous determination of the metabolites of four nitrofuran antibacterial drugs in chicken tissues: furazolidone, furaltadone, nitrofurazone, and nitrofurantoin. Sample clean-up and analyte enrichment were performed by liquid-liquid extraction with ethyl acetate followed by solvent washing, hydrolysis of the protein-bound drug metabolites, and derivatization with 2-nitrobenzaldehyde (2-NBA). ESI parameters were optimized, and the chromatographic separation of all metabolites was examined. Each metabolite produced a simple mass spectrum containing a strong signal corresponding to [M+H]+. Metabolite calibration curves, in the 0.25 to 1 ng/mL range, exhibited correlation coefficients greater than 0.999. The limit of detection (LOD) for each analyte ranged from 0.02 to 0.06 ng/mL.

Introduction
The four drugs shown in Figure 1, furazolidone, furaltadone, nitrofurazone, and nitrofurantoin, belong to the group of nitrofuran antibacterial drugs. These drugs have been widely used as feed

Drugs
O N N2O O N O H2N

Metabolites
O N O

Furazolidone

3-amino-2-oxazolidinone (AOZ)

O N N2O O N O H2N N

O O

N
Furaltadone

5-morpholinomethyl-3-amino-oxazolidinone (AMOZ)

O N N2O O N H
Nitrofurazone

O H2N NH2 N H NH2

Semicarbazide (SEM)

O N N2O O N NH H2N N

O NH

O
Nitrofurantoin

O
1-aminohydantoin (AHD)

Figure 1.

Structure of the nitrofuran antibacterial drugs and their metabolites .

been set by any regulatory agency, the goal of the analytical method was to estimate the lowest possible detection limit.

Experimental
Chemicals and Solvents Three metabolites: 3-amino-2-oxazolidinone (AOZ), semicarbazide (SEM), and 1-aminohydantoin (AHD) were purchased from Sigma Aldrich Japan (Tokyo, Japan). The purity of these compounds was greater than 99%. The 2-NBA derivatives of these metabolites were prepared by the Livestock

Department in Thailand (Palm Thani, Thailand) using the procedure described by Leitner [4]. Stock solutions of these three 2-NBA derivatives were prepared in methanol at 1000 ng/mL and stored in the dark at 4 C. The stock solution was diluted to the desired concentration just prior to its use for the optimization of ESI parameters. Acetonitrile, ethyl acetate, formic acid, and dimethyl sulfoxide (DMSO) were supplied by Wako Chemical (Osaka, Japan). Hydrochloric acid and 2-NBA were purchased from Tokyo Kasei Kogyo (Tokyo, Japan). Water was purified with a Milli-Q system (Millipore, Tokyo, Japan).

Sample Preparation Sample preparation procedures included solvent wash and acid extraction by homogenization and derivatization with 2-NBA. Chicken muscle and liver were prepared by the Livestock Department in Thailand. Calibration curves for the four nitrofuran metabolites (from the Livestock Department) were constructed in the range 0.25 to 1.0 ng/mL. The derivatization and sample preparation procedures used by the Livestock Department are the following: 1. The four metabolite solutions in water, 12.5, 25.0, 37.5, and 50 L at 100 ng/mL, were transferred to separate 40 mL glass vials with screw caps. 2. A solution of 10 mL HCl (125 mM in water) and 200 L 2-NBA (50 mM in DMSO) were added to each vial. 3. The reaction mixtures were kept in a water bath at 37 C for 16 hours. 4. The solutions were cooled to room temperature. 5. The pH was adjusted to about 7.4 by adding 0.1 M aqueous KHPO4 or 0.8 M aqueous NaOH. 6. A 5-mL measure of ethyl acetate was added to each reaction mixture, and shaken for 2 min. 7. Each ethyl acetate phase was transferred to a separate glass vial and evaporated under a stream of nitrogen. 8. Finally, each residue was reconstituted in 5 mL of 1:1 methanol:water (V/V). The calibration curve was based on the metabolite concentration in clean solvent and derivatization using 2-NBA. Previous studies done by the Livestock Department showed that recovery of all metabolites from chicken extracts was above 80%. Therefore, the amounts of metabolite in chicken extract can be calculated by comparing the responses of 2-NBA derivatives from the samples against the calibration curve.

Instrument and Experimental Conditions An Agilent 1100 series LC, with a solvent degassing unit, a binary high-pressure gradient pump, an automatic sample injector, and a column thermostat, was used for separation. An 1100 series diode array detector (DAD) was connected in line with an 1100 MSD for detection and confirmation. The column and MS conditions are described in Table 1.
Table 1. LC: Column: Solvent A: Solvent B: Gradient: Column temp Sample volume Flow rate: Instrument Parameters Agilent 1100 series Inertsil ODS3, 150 mm 2.1 mm, 5 m (GL Science, Tokyo, Japan) Acetonitrile Aqueous 0.5% formic acid 20/80 A/B to 70/30 A/B in 20 min 20 C 30 L 200 L/min

MS: Ionization: Scan range: SIM ion: Drying gas: Nebulizer gas: Fragmentor: Vcap

Agilent 1100 MSD, SL ESI (Positive) 100500 m/z for optimization Base peak for quantitation Nitrogen, 10 L/min at 350 C Nitrogen, 50 psi 120 or 140 V 2000 V

Quantitative analysis was carried out using selective ion monitoring (SIM) of the base peak ions according to the program shown in Table 2. To confirm the presence of the target analytes in chicken extract, the sodium adduct ions (qualifier ions) of all target analytes were also monitored.

Table 2. Group 1 2 3

SIM Program Time window min 06 612.5 12.514 Analyte(s) 2-NBA-AMOZ 2-NBA-SEM and 2-NBA-AHD 2-NBA-AOZ Target ion 335 209 and 249 236 Qualifier ion 357 231 and 271 258 Dwell time msec 500 250 and 250 500 Fragmentor voltage, V 140 120 and 140 140

System Optimization Positive ion mass spectra were acquired over the scan range m/z 100500 using a step size of 0.1 amu and a scan rate of 2 seconds per scan for the optimization of fragmentor voltage. Ion lens voltages in the MS were automatically optimized using a Calibrant Delivery System and the AutoTune program. Using the analytical column and three 2-NBA derivatives (AOZ, SEM, and AHD) at 100 ng/mL, instrument performance was optimized by adjusting the four major ESI parameters: the capillary voltage, fragmentor voltage, the nebulizer gas pressure, and the drying gas flow rate. However, significant variation in the intensity of analytes was not observed when the drying gas flow rate and nebulizer gas pressure were varied from 4 L/min to 13 L/min and 20 psi to 60 psi, respectively. Capillary and fragmentor voltages applied to the inlet and exit end of the capillary affected the ion

transmission significantly. Fragmentor voltage also affected the fragmentation of sample ions. In general, higher fragmentor voltage helps the transmission of ions through the relatively high-pressure region between the exit of the capillary and the entrance of the skimmer. High fragmentor voltage can cause fragmentation to occur which provides structural information of the ion. For compounds that do not fragment easily, higher fragmentor voltage often results in better ion transmission. Optimal fragmentor voltage is compound dependent. Evaluation of the fragmentor voltages for the three 2-NBA-metabolites was done under the same chromatographic conditions as the analysis. Mass spectra of three 2-NBA-metabolites are shown in Figure 2. Each mass spectrum exhibited [M+H]+ as the base peak. Adducts ions [M+NH4]+ and [M+Na]+ were observed at lower fragmentor voltage (120 V) and some fragment ions (m/z=178 and 192) were observed at higher fragmentor voltage (180 V). Interestingly, the [M+NH4]+ ion was not observed at

Fragmentor = 120 V
800000

Fragmentor = 180 V
209(M+H)+ 192
200

600000

231(M+Na)+

400000 200000 0

400000 200000 100000 100 500000 400000 300000 200000 100000 0 100 150 150

200

250

300

350 400000

100

150

166

231(M+Na)+
250

600000

209(M+H)+

2-NBA-SEM

2-NBA-SEM

300

350

249(M+H)+ 266(M+NH4)+

249(M+H) 178
200 250

271(M+Na)+

2-NBA-AHD

300000 200000 100000 0

2-NBA-AHD

200

250

300

350

100

150

271(M+Na)+
300

350

236(M+H)+

40000

258(M+Na)+

40000 20000

20000

0 100 150 200 250 300 350

0 100 150 200 250 300 350

Figure 2.

Mass spectra of 2-NBA-SEM, 2-NBA-AHD, and 2-NBA-AOZ from two ESI fragmentor voltages.

192

258(M+Na)+

2-NBA-AOZ

253(M+NH4)+

60000

2-NBA-AOZ

236(M+H)+

180 V fragmentor voltage due to its stability. As seen in Figure 3, in order to ensure the best sensitivity, the fragmentor voltage for 2-NBA-SEM was set to 120 V and that of 2-NBA-AHD and 2-NBA-AOZ was set to 140 V for the analysis. Although 2-NBA-AMOZ was not examined, fragmentor voltage of this compound was set to 140 V because of its structural similarity to 2-NBA-AOZ. For the capillary voltage varied between 1500 and 4500 V, the optimal voltage was found to be 2000 V for all three metabolites.

Linearity, Detection Limits, and Precision In order to achieve optimal sensitivity, all quantitation experiments were carried out under SIM conditions, and the [M+H]+ ions were monitored for all 2-NBA-metabolites. To evaluate the linearity of the calibration curves, various metabolite solutions ranging from 0.25 ng/mL to 1 ng/mL were derivatized and then analyzed. As shown in Table 3, the linearity was very good for all 2-NBA-metabolites with correlation coefficients (r2) greater than 0.999.

1000000

800000

700000

600000 Peak intensity

500000

400000

2-NBA-SEM
300000

2-NBA-AOZ 2-NBA-AHD

200000

100000

0 100 120 140 Fragmentor voltage/V 160 180 200

Figure 3.

Effect of fragmentor voltage on peak intensity. Mobile phase, 20% acetonitrile/80% water 0.1% formic acid; Analyte concentration, 100 ng/mL.

The LOD for all 2-NBA-metabolites was estimated by extrapolating to a signal-to-noise ratio (S/N) of 3 using the signal from the standard solution at 0.25 ng/mL. These SIM chromatograms are shown in Figure 4. The LODs of the metabolites were in the range of 0.02 ng/mL to 0.06 ng/mL. These LODs were lower than those of the LC/MS/MS method developed by Leitner [4]. The intraday

instrument precision (repeatability) was determined by injecting aqueous standard solutions containing all of the 2-NBA-metabolites at 0.5 ng/mL five times during a working day. The interday instrument precision (reproducibility) was evaluated by analyzing the same sample three times over 3 working days. The precision for all analytes ranged from 3.1% to 8.2%, as seen in Table 3.

6000 6000 5500

2-NBA-AMOZ
5000

5500

2-NBA-AHD

5000 4500 4500 4000 4 8000

m/z = 357
6 8 10 12 14 60000 4 6 8

m/z = 271
10 12 14

7000

2-NBA-SEM
6000

56000

2-NBA-AOZ

52000 5000

m/z = 258

4000

m/z = 231
4 6 8 10 12 14

48000 4 6 8 10 12 14

Retention time(min)

Retention time(min)

Figure 4.

SIM chromatograms of aqueous 2-NBA nitrofuran metabolites solution at 0.25 ng/mL.

Table 3.

Linearity, LOD, and Instrument Precision of Metabolites in Aqueous Solutions Instrument precision (%RSD) Repeatability** Reproducibility*** 5.0 7.3 4.7 8.1 4.9 7.9 3.1 8.2

Metabolites AMOZ SEM AHD AOZ

r2 0.9999 0.9998 0.9989 0.9997

LOD* (ng/mL) 0.04 0.02 0.06 0.06

*Detection limit is LOD defined as S/N = 3 for standard solution at 0.25 ng/mL **Repeatability was calculated based on five replicates at 0.5 ng/mL within 1 day ***Reproducibility was calculated based on once per day for 3 days at 0.5 ng/mL

Evaluation of Chromatographic Separation Several reverse-phase columns were evaluated for HPLC performance. In terms of minimizing the inherent matrix suppression effects on the ESI process, Inertsil ODS3 column provided the best separation between analytes and the majority of the matrix components with the given mobile phase. Further, the linear solvent gradient gave the best compromise between short analysis time and sufficient matrix and analytes separation. Figure 5 shows individual SIM chromatograms for the four metabolite derivatives in spiked chicken muscle at 0.2 ng/g. No interference peaks were observed for 2-NBA-AMOZ and 2-NBA-AOZ, but it was difficult to separate 2-NBA-SEM and 2-NBA-AHD from the interfering matrix peaks. However, these peaks could still be identified by comparison with the blank sample, and the analyte amounts could then be calculated.

25000 15000 5000 0 30000 2

2-NBA-AMOZ

m/z = 335
4 6 8 10 12 14 16

2-NBA-SEM
20000 10000 0 0 40000 30000 20000 10000 0 0 50000 40000 30000 20000 10000 0 0 2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16 2 4 6 8 10 12

m/z = 209
14 16

2-NBA-AHD

m/z = 249

2-NBA-AOZ
m/z = 236

Retention time(min)

Figure 5.

SIM chromatograms of a spiked chicken muscle tissue sample containing 0.2 ng/g of each of the four 2-NBA nitrofuran metabolites.

Application of the Method to Chicken Liver Samples It has been reported that AOZ concentrations in liver tissue are several times higher than in muscle tissue [1,2]. This indicates that detection of nitrofuran metabolites in liver would be possible over an even longer period of time. Since the nature of the liver matrix is considered to be different from muscle and more difficult to separate target compounds from the interfering matrix, the developed LC/MS muscle method was also tested for the applicability to liver matrix. Figure 6 shows individual SIM chromatograms of the four metabolite derivatives in spiked chicken liver tissue. AMOZ, SEM, and AOZ derivatives were identified unambiguously and quantified down to 0.2 ppb. However, the AHD derivative overlapped with the matrix component and was difficult to quantify.

35000 25000 15000 5000 0 30000 20000 10000 0 0 25000 20000 15000 10000 5000 0 0 70000 50000 30000 10000 0

2-NBA-AMOZ (0.84 ppb)

m/z = 335
2 4 6 8 10 12 14 16

2-NBA-SEM (1.21 ppb)


m/z = 209
2 4 6 8 10 12 14 16

2-NBA-AHD (0.33 ppb)


m/z = 249
2 4 6 8 10 12 14 16

2-NBA-AOZ (0.72 ppb)

m/z = 236
12 14 16

10

Retention time (min)

Figure 6.

SIM chromatograms of a spiked chicken liver tissue sample containing 0.2 ng/g of each of the four 2-NBA nitrofuran metabolites.

Conclusion
The development of a routine and sensitive LC/MS method allows for the simultaneous detection of four nitrofuran metabolite derivatives. The detection limit of each analyte ranges from 0.05 to 0.2 ng/g in chicken muscle and liver tissues.

References
1. L.A.P. Hoogenboom, M. van Kammen, M.C.J. Berghmans, J.H. Koeman and H.A. Kuiper, Food Chem. Toxicol. 1991; 28: 321. 2. L.A.P. Hoogenboom, M.C.J. Berghmans, T.H.G. Polman, R. Paker and I.C. Shaw, Food Addit. Contam. 1992; 9: 623. 3. D.W. Gottschall and R. Wang, J. Agric. Food. Chem. 1995; 43: 2520. 4. A. Leitner, P. Zollner, W. Lindner, J. Chromatogr. A 2001; 939: 49.

Acknowledgement
The author would like to acknowledge the Livestock Department staff in Thailand for providing the four 2-NBA-metabolite standards and chicken muscle and liver extracts.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.

www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2003 Printed in the USA March 19, 2003 5988-8903EN

HPLC Analysis of Antibacterial Drugs with Penicillin-Like Structure Application

Drug Development
Udo Huber, Adebayo O. Onigbinde

Penicillins can be isolated from the culture medium of certain fungi-producing natural penicillin, such as Penicillium notatumor and P. chrysogenum. Other penicillins can be synthesized semisynthetically or by precursor-indicated biosynthesis. Total synthesis would not be economical. Penicillin inhibits the polymerization of murin, which is responsible for the stability of the bacteria's cell wall. Because many antibacterials are toxic, various countries regulate the level of antibacterial residues in agricultural, veterinary, dairy, and meat-based food products. Figure 1 shows the HPLC separation of four common antibacterial drugs with pencillin-like structure (amoxicillin, ampicillin, penicillin G, and penicillin V) on an SB-C18 reversed phase column.

Highlights
There is excellent resolution of penicillin analogs without ion pairing agent. There is rapid resolution of the penicillin analogs on the SB-C18 column at low pH and buffer concentration. Penicillins are eluted from the column with good and narrow peak shape. Extreme stability of sterically protected SB-C18 bonded phases allows for excellent separation at low pH. The SB-C18 column provides excellent peak shape and selectivity for antibacterial drugs. The HPLC method shows an easy but reliable and precise analysis of the antibacterial drugs. The values for limit of detection (LOD), precision of retention time (RT) and area show the good performance of the HPLC analysis.

1000 Absorbance (mAU) 800 600 400 200 0 0 2 4 6 Time (min) 8 10 12

1 2 3 4 4

Amoxicillin Ampicillin Penicillin G Penicillin V

Figure 1. Separation of four penicillin analogs.

Experimental Conditions
Equipment: Agilent 1100 Series HPLC; UV Detector: Variable wavelength detector, 204 nm, standard cell; Column: Zorbax SB-C18, 3. 5 m, 4.6 75 mm (part number 866953-902); Mobile phase: A = 0.025 M KH2PO4 in water (pH = 3), B = acetonitrile; Injection volume: 5 L; Temp: 40 C; Flow rate: 1.0 mL/min; Gradient: at 5% B to 60% in 10 min; Column wash: 60% B to 5% B in 2 min; Stop time: 12 min; Post time: 5 min

www.agilent.com/chem
Table 1. HPLC Method Performance of Antibacterial Drugs with Penicillin-Like Structure LOD for S/N = 2 (mg/L)* 1.0 1.0 1.0 1.0 Precision of RT (RSD of 10 runs) (100 mg/L)* 0.32 0.32 0.32 0.25 Precision of area (RSD of 10 runs) (100 mg/L)* 0.54 0.55 0.49 0.48

Compound Ampicillin Amoxicillin Penicillin G Penicillin V


*Injection volume: 5 L

Udo Huber is an application chemist based at Agilent Technologies, Waldbronn, Germany. Adebayo Onigbinde is an applications chemist based at Agilent Technologies, Wilmington, Delaware, USA.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA September 10, 2002 5988-7629EN

HPLC Separation of Antibacterial Drugs with Tetracycline Structure Application

Drug Development
Udo Huber, Adebayo O. Onigbinde

Tetracyclines occur naturally in some streptomyces species. Besides being used in human and veterinary medicine, they are fed as nutritional antibiotics in pig and poultry farming. Because of their long half-life and resistance, there is a high restriction on their usage in some European countries, such as Germany. Figure 1 shows the HPLC separation of three common tetracycline analogs on a Zorbax SB-C18 reversed phase column. This application demonstrates separation without ion pairing and the use of an alternative mobile phase to TFA in separating antibacterial drugs.
2
120 100 Absorbance (mAU) 80 60 40 20 0 0 2 4 Time (min) 6 8 10

Highlights
The SB-C18 column provides excellent peak shape and selectivity for basic antibacterial drugs. The SB-C18 column shows excellent stability at low pH. The SB-C18 column shows excellent and rapid resolution of antibiotics at low pH and buffer concentration. The HPLC method shows an easy but reliable and precise analysis of the antibacterial drugs. The values for limit of detection (LOD), precision of retention time (RT), and area show the good performance of the HPLC analysis.

1 3

1 Minocycline 2 Tetracycline 3 Doxycycline

Figure 1. Separation of three antibacterial drugs with tetracycline structure.

Experimental Conditions
Equipment: Agilent 1100 Series HPLC; UV Detector: Variable wavelength detector, 350 nm, standard cell; Column: Zorbax SB-C18, 3. 5 m, 4.6 75 mm (part number 866953-902); Mobile phase: A = 0.025 M KH2PO4 in water (pH = 3), B = acetonitrile; Injection volume: 5 L; Temp: 25 C; Flow rate: 1.0 mL/min; Gradient: at 5% B to 60% B in 10 min; Column wash: 60% B to 5% B in 2 min

www.agilent.com/chem
Table 1. HPLC Method Performance of Antibacterial Drugs with Tetracycline Structure LOD for S/N = 2 (mg/L)* 0.1 0.1 0.1 Precision of RT (RSD of 10 runs) (100 mg/L)* 0.06 0.05 0.04 Precision of area (RSD of 10 runs) (100 mg/L)* 0.14 0.13 0.21

Compound Minocycline Tetracycline Doxycycline


*Injection volume: 5 L

Udo Huber is an application chemist based at Agilent Technologies, Waldbronn, Germany. Adebayo Onigbinde is an applications chemist based at Agilent Technologies, Wilmington, Delaware, USA.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA September 10, 2002 5988-7630EN

Analysis of Residual Synthetic Antibacterials in Meat by HPLC Application

Food
Hiroki Kumagai, Adebayo Onigbinde

Many domestic cattle receive various antibacterials in their feed for the prevention and control of disease caused by fungi and bacteria. Residues of antibacterials are found in food made from the meat of these animals. Since many antibiotics are toxic, many countries regulate acceptable residue levels of compounds allowable in agricultural and animal products. Many alkyl-C18 columns tail with basic compounds and have a shorter life time at low pH. Purospher column separated basic antibacterials with good resolution, peak shape,and efficiency.
Absorbance [mAU] 25 20 15 10 5 0 0 Absorbance [mAU] 25 20 15 10 5 0 0 5

Highlights
Separation of 10 antibacterials in meat at low pH Excellent and rapid resolution of antibacterials at low sample concentration Elution of antibacterials from the column with good peak shape and narrow peak width Separation of low level amounts of a wide range of pharmaceutical compounds with differing structures in a single analysis by Purospher column

3 8 9 10
1 2 3 4 5 Time (min) SMr PYM TCP SDD FZD 6 7 8 9 10 SMMX DFZ SDMX SQX OXA

at 224 nm
1
10 15

4 5

67

20

25

30

35

at 360 nm

DFZ FZD

10

15

20

25

30

35

Time (min)

Figure 1. Chromatogram of standard solution, 2 g/mL each analyte.


Absorbance [mAU] 4 3 2 1 0 -1 0 Absorbance [mAU] 4 3 2 1 0 -1 0 5 10

at 224 nm SDD

15

20

25

30

35

Time (min)

Analyzed Compounds Sulfamerazine (SMR) Sulfadimidine (SDD) Sulfamonomethoxine (SMMX) Sulfadimethoxine (SDMX Sulfaquinoxaline (SQX) Pyrimethamine (PYM) Thiamphenicol (TPC) Furazolidone (FZD) Difurazone (DFZ) Oxolinic acid (OXA) Sample: Extracts from bovine muscle Sample preparation: According to the official procedure of the Japanese food sanitation law.

at 360 nm

10

15

20

25

30

35

Time (min)

Figure 1. Chromatogram of extract of bovine muscle.

Instrument: Agilent 1100 Series HPLC; Column: 250 mm 4 mm id, RP-18 Purospher, 5 m, Part no. 79925PU-584; Mobile phase: A = 0.7 % Phosphoric acid, B = CH3CN; Gradient: 0.0 min 5% B; 10.0 min 65% B; 40.0 min 65% B; 45.0 min 65% B; Post Time 7.0 min 5% B; Flow rate: 1.0 mL/min; Temperature: 40 C; Injection volume: 20 L; Diode array detector: A338/10 nm, reference wavelength off; B264/8 nm, reference wavelength off; C360/8 nm, reference wavelength off

www.agilent.com/chem

Hiroki Kumagai is the PHS Support manager based at Yokogawa Analytical Systems Inc., Tokyo, Japan. Adebayo Onigbinde is an application chemist based at Agilent Technologies, Wilmington, Delaware, USA.

For More Information


For more information on our products and services, visit our Web site at www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA June 18, 2002 5988-7135EN

Analysis of tetracyclines by HPLC

Rainer Schuster Food

Abstract Tetracyclines are used worldwide as oral or parenteral medication in the form of additives in animal feed. In food-producing animals, these drugs exhibit a high degree of activity toward a wide range of bacteria.1, 2 Sample preparation After homogenization or mincing and addition of mineral acids to dissociate tetracyclines from proteins, the samples were extracted using liquid/liquid extraction followed by degreasing and/or deproteinization, purification, and concentration.3 Chromatographic conditions The HPLC method presented here for the analysis of meat is based on reversed-phase chromatography and UV-visible diode-array detection. UV spectra were evaluated as an additional identification Conditions tool.
Absorbance [mAU] 1 2 3 4 5 6 7 Oxytetracycline Tetracycline epi-Tetracycline Demeclocycline epi- Demeclocycline Chlortetracycline Doxycycline

6 5 4 3 2 1 0 -1 2 4 6 1

2 4 6

Column: 100 4 mm Hypersil BDS, 3 m Mobile phase: A = water, pH = 2.1 with sulfuric acid B = ACN Gradient: start with 15 % B at 10min 60% B Flow rate: 0.5 ml/min Column compartment: 25 C Detector: UV-DAD detection wavelength 355 nm/20 nm, reference wavelength 600/100 nm

5 10 ng each

Sample preparation
1. 1 g sample was mixed with citric acid (100 mg). 2. add 1 ml nitric acid (30 %) or 0.1 m oxalic acid 3. add 4 ml methanol 5 min in the ultrasonic bath 4. add water up to 10 ml total volume 5. centrifuge 6. inject

1 ng each 8 10 12 14 Time [min]

Figure 1 Analysis of tetracyclines by HPLC

Agilent Technologies
Innovating the HP Way

Area Area = 2.59596374*Amt -0.0821516 25 20 15 10 5 0 0 5 Amount [ng/ul] 3 2 Correlation: 0.99996 Rel. Res%(3): 2.461 1

Equipment
Agilent 1100 Series vacuum degasser quaternary pump autosampler thermostatted column compartment diode array detector, Agilent ChemStation + software

HPLC method performance


Limit of detection for UV-DAD 100 ppb Repeatability of RT over 10 runs <0.2 % of areas over 10 runs <2 %

Figure 2 Linearity for oxytetracycline 1-10 ng


Absorbance [mAU] 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 200 250 300 350 400 Wavelength [nm] 1 ng Tetracycline Library Tetracycline

References
1. H. Malisch et al., Determination of residues of chemotherapeutic and antiparasitic drugs in food stuffs of anomaly origin with HPLC and UV-Vis diodearray detection J. Liq. Chromatogr., 1988, 11 (13), 28012827.14. 2. M.H. Thomas, J. Assoc. Off. Anal.; 1989, 72 (4) 564. 3. Farrington et. al., Food Additives and Contaminants, 1991, Vol. 8, No. 1, 55-64.
Copyright 1997 Agilent Technologies Released 09/97 Publication Number 5966-1619E

Figure 3 Analysis of tetracyclines at 100 ppb by HPLC


Rainer Schuster is application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem

Agilent Technologies
Innovating the HP Way

Contaminants Acrylamides

> Return to Table of Contents > Search entire document

Gas Chromatography/Mass Spectrometry Approaches to the Analysis of Acrylamide in Foods Application

Food Safety

Author
Bernhard Rothweiler Agilent Technologies Deutschland GmbH Hewlett-Packard Strasse 8 76337 Waldbronn Germany Eberhardt Kuhn Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, CA USA Harry Prest Agilent Technologies, Inc. 5301 Stevens Creek Blvd. Santa Clara, CA USA

Introduction
The discovery announced in April 2002 by scientists at Swedens National Food Administration of acrylamide (2-propenamide) in fried and baked foods at levels many times that allowed in water suggested a much higher exposure than previously estimated [1-3]. Acrylamide (Figure 1), a known neurotoxin, is considered a probable human carcinogen. The World Health Organization considers 0.5 g/L the maximum level for acrylamide in water. However, foods such as french fries, baked potato chips, crisp breads, and other common cooked foods, were found to contain acrylamide between 100 and 1000 g/kg. Acrylamide was not found in the raw foodstuffs and cooking by boiling produced no detectable levels. Recent work has suggested that acrylamide forms via the Maillard reaction, which occurs when amino acids and sugars (for example, asparagine and sucrose) are heated together [4]. The concern over these relatively high concentrations has led to studies of the occurrence of acrylamide in a wide variety of foods.
H H H O H2N

Abstract
Discovery of acrylamide in cooked foods has required an examination of foods for potential exposure. A classic approach employs extracting acrylamide from the food with water and converting the acrylamide to brominated derivatives. These derivatives are described here in terms of their spectra and response in electron impact and positive chemical ionization. Additionally, a more