Investigation of the Unusual Behavior of Metolachlor under Chemical

Sep 18, 2011 - system coupled to a “Saturn 240MS” ion trap mass spectro- meter (Varian, Les ..... (3) Kale, V. M.; Miranda, S. R.; Wilbanks, M. S...
0 downloads 0 Views 846KB Size
LETTER pubs.acs.org/ac

Investigation of the Unusual Behavior of Metolachlor under Chemical Ionization in a Hybrid 3D Ion Trap Mass Spectrometer Pierre Henri Goulden, Sarah Coffinet, Christophe Genty, Sophie Bourcier, Michel Sablier, and Stephane Bouchonnet* Laboratoire des Mecanismes Reactionnels, Ecole Polytechnique, Route de Saclay, 91128 Palaiseau Cedex, France ABSTRACT: This article describes the strange behavior of the widely used herbicide metolachlor under chemical ionization conditions in a hybrid source ion trap mass spectrometer in gas chromatography/mass spectrometry (GC/MS) coupling. With the use of ammonia as the reagent gas, metolachlor provides a chlorinated ion at m/z 295/297, almost as abundant as the protonated molecule at m/z 284/286, which cannot be isolated to perform tandem mass spectrometry (MSn) experiments. Curiously, this ion at m/z = M + 12 is not observed for the herbicides acetochlor and alachlor, which present very similar chemical structures. The chemical structure of the m/z 295/297 ions and the explanation of the observed phenomenon based on the metastable behavior of these ions were elucidated on the basis of experiments including isotopic labeling and modifications of the operating conditions of the ion trap mass spectrometer. This work allows one to give new recommendations for an optimized use of hybrid source ion trap mass spectrometers.

A

cetochlor, alachlor, and metolachlor are widely used herbicides frequently detected in river waters.1,2 Known as potential endocrine disruptors, their toxicity for human has been clearly established.3 5The characterization of their ozonation and photolysis byproducts has been the subject of recent investigation since some of them could also been responsible for endocrine disruption.6,7 The structural elucidation of the byproducts is usually performed using liquid chromatography tandem mass spectrometry (LC MS/MS) and gas chromatography/tandem mass spectrometry (GC/ MSn) analysis for polar and little or not polar compounds, respectively. In GC/MS, the use of chemical ionization (CI) is expected to provide abundant MH+ ions allowing direct access to the molecular weight of the analyte while electron ionization (EI) provides abundant fragment ions but only trace amounts of M+. ions for the herbicides of interest. We have been recently puzzled by the strange behavior of metolachlor which, unlike the other mentioned pesticides, provides a very abundant monochlorinated ion at m/z 295/297 under ammonia positive chemical ionization using a hybrid ion trap mass spectrometer in GC/MS coupling. In the hybrid mode of operation, reagent ions are formed in an external ion source and introduced into the ion trap where they may react with the molecules directly eluting from the analytical column into the trap. In comparison with the more “classic” ion trap operating modes using either internal ionization, either external ionization, hybrid ionization allows selecting a reagent ion among a plasma of reagent ions for a better control of internal energy deposition.8 r 2011 American Chemical Society

’ EXPERIMENTAL SECTION All analyses were carried out on a “450GC” gas chromatograph system coupled to a “Saturn 240MS” ion trap mass spectrometer (Varian, Les Ulis, France). All experiments were performed automatically injecting 1.0 μL of sample in the splitless mode. High purity (99.999%) helium was used as the carrier gas at a constant flow of 1.4 mL min 1 held by electronic pressure control; it was also used as the damping gas in the hybrid mode (see below). The manifold, ion trap electrodes, and transfer line temperatures were set at 120, 220, and 300 °C, respectively. Spectra were recorded using the automatic gain control (AGC) function with a target value of 5000. The emission current was set to 20 μA, and the multiplier voltage to 1850 V. Full scan mass spectra were recorded on a scanning range from m/z 50 to m/z 650. The isolation window was set at 3 m/z for MS/MS experiments. Chemical ionization (CI) experiments were carried out using methanol and ammonia as reagent gases. In the internal ionization mode, CI was performed with the whole plasma of ions, i.e., m/z 33 and m/z 47 for MeOH, m/z 18, and m/z 35 for NH3. In the hybrid operating mode, ions m/z 33 were isolated for MeOH using an ejection amplitude of 15 V, with reagent low mass and high mass values of 25 and 45, respectively. For ammonia, m/z 18 ions were isolated using an ejection amplitude of 20 V, with reagent low mass and high mass values of 15 and 22, respectively. The reaction time was set to 100 ms in both cases. Received: June 20, 2011 Accepted: September 18, 2011 Published: September 18, 2011 7587

dx.doi.org/10.1021/ac201563w | Anal. Chem. 2011, 83, 7587–7590

Analytical Chemistry

LETTER

Figure 1. Chemical structures of acetochlor, alachlor, metolachlor, as well as their ammonia CI mass spectra recorded using a hybrid ion trap mass spectrometer.

Figure 2. Fragmentation pathways of protonated herbicides: acetochlor (up), alachlor (middle), and metolachlor (bottom).

’ RESULTS AND DISCUSSION The chemical structures of acetochlor, alachlor, and metolachlor as well as their ammonia hybrid-CI (HCI) mass spectra are

displayed Figure 1. No other ions than those displayed in Figure 1 were observed in the whole mass spectra (from m/z 50 to m/z 650). MH+ pseudo molecular ions are displayed in each herbicide mass spectrum. The acetochlor and alachlor NH3 HCI mass spectra are dominated by fragment ions at m/z 212/214 and 226/228, respectively. Protonated metolachlor also dissociates to provide a single fragment ion at m/z 252/254 attributed to the loss of a neutral fragment of 32 amu. The mechanisms assumed to explain the fragmentation mechanisms of protonated herbicides are displayed in Figure 2. Compared to those of acetochlor and alachlor, the ammonia hybrid-CI (HCI) mass spectrum of metolachlor shows a noticeable difference with the appearance of peaks at m/z 295/297 for which the shape of the isotopic distribution around m/z 295 and m/z 297 ions does not perfectly match the one expected for a monochlorinated ion. This distortion, added to the fact that these ions appear to be slightly “decalibrated” when the other ones perfectly match their expected m/z values in the mass spectrum, suggests that the corresponding ions are metastable. Furthermore, the ions at m/z 295 and m/z 297 from metolachlor disappear when attempting to isolate them for performing MSn experiments. We tried to isolate them with no activation energy reducing the activation time in the ion trap to 1 ms (the default value is 20 ms) as well as beginning the m/z scanning at m/z 280 to reduce the ion storage duration. In all cases, no ion at m/z 295 nor m/z 297 could be isolated, thus confirming their metastability. Artifact ions (also referred as “ghost” ions) have been reported by ion trap specialists in the early 1990s.9 Their formation has 7588

dx.doi.org/10.1021/ac201563w |Anal. Chem. 2011, 83, 7587–7590

Analytical Chemistry

LETTER

been rationalized in terms of nonlinear resonance effects mainly due to hexapolar and octopolar field contributions in quadrupolar ion traps having a stretched design, like the one used in this study. Nonlinear resonances increase the internal energy of ions when the later are stored in the Mathieu stability diagram on locations corresponding to βz = 2/3 (hexapolar field), βz = 1/2, or βz + βr = 1 (octopolar field).10 This increase in internal energy induces ion dissociation and/or ion neutral reactions, particularly when performing chemical ionization because of the long storage times and of the high reagent gas pressures generaly used in CI. In chemical ionization, this phenomenon can play a role on reagent ions storage as well as on the behavior of analyte ions. In the ammonia CI experiments carried out in this work, NH4+ ions were selectively stored at qz = 0.75 (low mass cutoff = 15, az = 0), i.e., kept away from the qz values of 0.65 and 0.80 where nonlinear resonance effects are significantly observed according to Eades et al.11 Consequently, activation of m/z 18 ions by nonlinear resonance effects seems very unlikely here. The behavior of

pseudomolecular ions will be discussed below in the paragraph devoted to metolachlor-d6 analysis. Signal loss during the isolation of precursor ions in collision induced dissociation (CID) experiments has also been interpreted in terms of nonlinear resonance effects. Ion ejection occurs when isolation is performed at a qz value near 0.65 or 0.80; this phenomenon is generally referred to as the “black hole” or “black canyon”.12 In our MS/MS experiments, precursor ions are isolated at qz = 0.30 so that a black hole cannot be the cause of their disappearing. The same experiments performed using methanol (isolation of the m/z 33 ion) and methane (isolation of the m/z 17 ion) as reagents for HCI provided mass spectra displaying only the pseudomolecular ion MH+ and the daughter ions at m/z 252/254, without any m/z 295/297 ion in the case of metolachlor. This result shows that the internal energy deposition in the pseudomolecular ion during proton transfer plays a critical role in the formation of m/z 295/297 ions as expected if one considers the relative proton affinity values of the chemical ionization reagents, CH4, CH3OH, and NH3, respectively. When ammonia CI experiments are performed on metolachlor using an ion trap operating in the internal mode, no ions are observed at m/z 295/297. In the internal mode, helium is used both as a vector gas and damping gas to improve ion storage; it is eluted from the chromatograph at a flow rate of 1.4 mL min 1. In the hybrid mode, helium is eluted from the capillary column at the same flow rate but additional helium is introduced at 2.5 mL min 1 (default value) through another entrance to serve as damping gas. Consequently, one can expect a higher partial pressure of helium in the hybrid mode compared to the internal one. In the scan mode, helium ensures thermalization of ions: multiple collisions with helium atoms reduce the internal energy of ions. The greater the helium pressure, the greater the efficiency of thermalization. Different damping gas flow rate values were tested between 0.5 and 3.0 mL min 1. The ratios of

Figure 3. m/z 295/m/z 284 ratio as a function of the damping gas flow rate.

Figure 4. Mechanism of formation and dissociation of the m/z 295/297 ion. 7589

dx.doi.org/10.1021/ac201563w |Anal. Chem. 2011, 83, 7587–7590

Analytical Chemistry m/z 295/297 ion abundances vs m/z 284/286 ions abundances are plotted as a function of the damping gas flow rate on Figure 3. Experimental data confirms that the relative abundance of the 295/297 ions significantly increases when increasing the helium pressure in the ion trap, confirming that the formation of the later implies efficient thermalization of ions. The ratio of the m/z 252/ 254 ions abundance vs the total ionic current (TIC) remains constant whatever the helium pressure in the trap. The TIC remains constant for damping gas flow rates ranging from 0.5 to 2.0 mL min 1 and falls down by a factor 100 at 3.0 mL min 1, likely because such a pressure in the trap makes ion ejection difficult. Additional experiments were carried out with the aim to study the formation of m/z 35 ions according to the reaction NH4+ + NH3 f (NH3)2H+ after isolation of NH4+ ions. The abundance of m/z 35 ions remained negligible in comparaison with that of m/z 18 ions whatever the helium pressure in the formentioned range. Ions at m/z 295 and m/z 297 are shifted by 7 Th when analyzing metolachlor-d6 (hydrogen atoms carried by the three carbon atoms between the amide and ether functions are replaced by deuterium atoms) in the same conditions. The corresponding ions at m/z 302 and 304 are also slightly “decalibrated” and could not be isolated for multiple stage MS experiments. This shift of 7 Th when the deuterated compound contains only six deuterium atoms implies an ion molecule reaction in the ion trap. The formation of the ion at m/z 295/ 297 has been rationalized by the mechanism proposed in Figure 4. Even if the proton affinity of the oxygen atom of the ether function is likely lower than that of the nitrogen atom too encumbered to favor such protonation, protonation of the former can occur in a large way considering the steric configuration of metolachlor.13 The mechanism proposed involves nucleophilic attack of the nonbonding electrons of the nitrogen atom of a neutral molecule of metolachlor onto the aromatic ring of the MH+ ion with simultaneous elimination of methanol, leading to the ion A on Figure 4. Formation of A involves the formation of a stable five center ring while the same mechanism with alachlor or acetochlor would have led to a much less stable four center ring due to a closer position of the ether function on their skeletons. This can explain why no ions of a greater m/z ratio than the MH+ ion are observed for alachlor and acetochlor. Formation of the m/z 295/297 ions from A is assumed to result from two α,β concerted eliminations according to mechanisms classically observed in the gas phase.14 In Figure 4, the first dissociation pathway leads to the elimination of CH3 CHd CH OCH3, the second one eliminates (Met)(Et)C6H3CH2Cl. One can expect that the order of these two mechanisms could be inverted in relation to the lability of the neutral fragments expelled. Since neither ions A nor ions B could be detected, even reducing the m/z scanning range of the mass spectrometer to reduce the time between their formation and their detection, this order could not be confirmed experimentally. During the loss of CH3 CHdCH OCH3 (CD3 CDdCD OCH3 in the case of deuterated metolachlor), a hydrogen atom carried by the carbon α of the ether function is proposed to be transferred onto the nitrogen atom, thus rationalizing the shift of +7 Th for ion C when analyzing metolachlor-d6. The metastability of ion C can be explained by the very easy elimination of HNdCdO from it. This elimination involves a very weak energy of activation since it does not imply any particular transition state. Furthermore, it provides a fully conjugated and thus stable carbocation. This carbocation, at m/z 252/254 in the mass spectrum, cannot be differentiated

LETTER

from the one resulting from methanol elimination from MH+ (see Figure 2). Results on metolachlor-d6, and particularly the shift of +7 Th for unexpected ions in comparison to the unlabeled compound, constitute the best argument to attribute the formation of these ions to an ion molecule reaction. As a matter of fact, assuming that the molecular ions of metolachlor and metolachlor-d6 are stored in very close location to the Mathieu stability diagram for a given value of qz, nonlinear resonance effects, if they had occurred in the same way for both compounds, would have likely provided artifact ions separated by 6 rather than 7 Th.

’ CONCLUSIONS This study allowed one to establish the formation mechanism of an unexpected metastable m/z 295/297 ion from metolachlor under NH3 CI using an ion trap mass spectrometer in the hybrid mode, never reported before whatever the conditions of analysis. It showed how this formation is directly correlated to operating conditions favoring ion thermalization. The detection of such kinds of ions constitutes an important problem for structural investigation because it may induce mistakes when attempting to determine the molecular weight of an unknown analyte using chemical ionization. To prevent their formation, attention should be paid to avoid the use a reagent with a too high proton affinity and the ion trap should be operated with a low damping gas pressure. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ REFERENCES (1) Byer, J. D.; Struger, J.; Sverko, E.; Klawunn, P.; Todd, A. Chemosphere 2011, 82, 1155–1160. (2) Whitall, D.; Hively, W. D.; Leight, A. K.; Hapeman, C. J.; McConnell, L. L.; Fisher, T.; Rice, C. P.; Codling, E.; McCarty, G. W.; Sadeghi, A. M.; Gustafson, A.; Bialek, K. Sci. Total Environ. 2010, 408, 2096–2108. (3) Kale, V. M.; Miranda, S. R.; Wilbanks, M. S.; Meyer, S. A. J. Biochem. Mol. Toxicol. 2008, 22, 41–50. (4) Krajcsi, P.; Oosterhuis, B.; Vukman, K.; Vagi, E.; Glavinas, H.; Jablonkai, I. Toxicology 2008, 248, 45–51. (5) Richardson, S. D. Anal. Chem. 2009, 81, 4645–4677. (6) Roberts, A. L.; Hladik, M. L.; Bouwer, E. J. Water Res. 2008, 42, 4905–4914. (7) Bouchonnet, S.; Kinani, S.; Souissi, Y.; Bourcier, S.; Sablier, M.; Roche, P.; Boireau, V.; Ingrand, V. Rapid Commun. Mass Spectrom. 2011, 25, 93–103. (8) Aguera, A.; Mezcua, M.; Mocholi, F.; Vargas-Berenguel, A.; Fernandez-Alba, A. R. J. Chromatogr., A 2006, 1133, 287–292. (9) March, R. E.; Todd, J. F. J. Quadrupole Ion Trap Mass Spectrometry, 2nd ed.; Wiley-Interscience: Hoboken, NJ, 2005. (10) Mo, W. J.; Langford, M. L.; Todd, J. F. J. Rapid Commun. Mass Spectrom. 1995, 9, 107–113. (11) Eades, D. M.; Johnson, J. V.; Yost, R. A. J. Am. Soc. Mass Spectrom. 1993, 4, 917–929. (12) Eades, D. M.; Yost, R. A. Rapid Commun. Mass Spectrom. 1992, 6, 573–578. (13) Bouchoux, G. Mass Spectrom. Rev. 2007, 26, 775–835. (14) McLafferty, F. W. Interpretation of Mass Spectra, 3rd ed.; University Science Books: Mill Valley, CA, 1980.

7590

dx.doi.org/10.1021/ac201563w |Anal. Chem. 2011, 83, 7587–7590