Environ. Sci. Technol. 2006, 40, 797-802
NMR Investigation of the Behavior of an Organothiophosphate Pesticide, Chlorpyrifos, Sorbed on Montmorillonite Clays MARK R. SEGER AND GARY E. MACIEL* Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523
Phosphorus-31 nuclear magnetic resonance (31P NMR) was used to explore the decomposition of chlorpyrifos (an organothiophosphate pesticide) sorbed at high concentration (typically 2-10 wt %) on partially hydrated montmorillonite clays in four different cation-exchanged forms (Ca2+, Cu2+, Zn2+, and Al3+). Solid-state 31P NMR (using magic-angle spinning and cross polarization or direct polarization) and liquid-solution 31P NMR of DMSO and acetone extracts indicate that chlorpyrifos is initially physisorbed, appearing by solid-state 31P NMR to exhibit significant motion on the molecular level, which results in almost liquidlike solidstate spectra. Over periods ranging from hours to years, the signals due to unreacted chlorpyrifos diminish and are replaced by new 31P NMR peaks resulting from hydrolysis, isomerization, mineralization, and oxidation reactions. The 31P NMR signal characteristics indicate that these decomposition products are much more tightly bound to the clay than is chlorpyrifos. Partially hydrated Cu(II)- and Al-montmorillonites most effectively catalyzed chlorpyrifos decomposition (but with different product distributions); Ca-montmorillonites (and, as previously shown, kaolin) were least effective. Solid-state13C and 27Al NMR spectra were less useful for following the decomposition of chlorpyrifos than those obtained by 31P NMR. Pesticide loading levels (1-10% w/w) that are very much higher than those typically found in the environment were used to facilitate 31P NMR detection of less-than-dominant decomposition species.
Introduction While studies detailing the disappearance of many commercial organothiophosphate pesticides have been numerous (1, 2), the emphasis in studies of the detailed chemistry of pesticide decomposition has been on the aquatic environment. Relatively little is known about the interaction of pesticides and their decomposition residues with soil components. This paper describes the use of solid-state NMR to examine nonphotochemical/nonbiological behavior of an organothiophosphate pesticide, chlorpyrifos (I, ArOP(S)(OEt)2, where Ar ) 3,5,6,-trichloro-2 pyridyl) when sorbed on certain montmorillonite clay minerals. A previous report examined by solid-state NMR the behavior of chloypyrifos sorbed on a variety of other soil-derived materials (2b). NMR should be well suited for the analysis of chemical and physical interactions and transformations in complex * Corresponding author phone: 970-491-6480; fax: 970-491-1801; e-mail:
[email protected]. 10.1021/es051704h CCC: $33.50 Published on Web 12/15/2005
2006 American Chemical Society
systems such as pesticides sorbed on soils, which are a complicated mixture of organic compounds (e.g., humic materials) and minerals (e.g., silica, silicates, clays) that are reported to be catalytically active, including in the decomposition of organophosphate pesticides (3-5). Organophosphate pesticides contain the 31P nuclide (100% natural abundance) and 31P NMR chemical shifts have large sensitivity to structural variation (6, 7). Since NMR is relatively insensitive and requires relatively large sample concentrations for detection, this study does not attempt to simulate actual concentrations found in the environment; instead this study focuses on characterizing the chemistry of interactions and decompositions of chlorpyrifos (6-13), which should carry over from one concentration to another. Scheme 1 of ref 2b summarizes likely or possible pesticide decomposition models. Hydrolysis to products of the type II, HOP(S)(OEt)2 h HS(O)(OEt)2, and/or III, ArOP(S)(OH)OEt h ArOP(O)(SH)OEt, are reported to predominate initial decomposition in the environment (9-14). Aqueous base hydrolysis of chlorpyrifos is reported to remove the aryl moiety preferentially, whereas neutral or acidic water tends to cleave the alkoxide moiety (11). Full hydrolysis to an unsubstituted aqueous phosphate ion or incorporation into the clay structure (“mineralization”) may be the pesticide’s ultimate fate. Isomerization is known to occur during synthesis or storage at elevated temperatures, but is usually not considered a major decomposition process in the environment. No literature reports were found for isomerization to the S-aryl analogue (IV), ArSP(O)(OEt)2, but isomerization to the S-alkyl isomer (V), ArOP(O)(SEt)(OEt), is known (6). Oxidation to the oxon form (VI), ArOP(O)(OEt)2, is of special concern, since this form has much greater mammalian toxicity than does chlorpyrifos (7, 8). S-alkyl isomerization (to V) also greatly increases mammalian toxicity (6).
Experimental Section Materials. Calcium montmorillonite, denoted STx-1 by the supplier, Source Clay Mineral Repository (University of Missouri-Columbia): traces of quartz, silica and carbonate (IR analysis); 0.57% Fe, 0.011% P by weight (elemental analysis). Dehydration of the “as received” material at 100 °C and 3 × 10-3 Torr to constant weight indicates 2.1% water content by weight, as received from the supplier. The Zn(II)-, Al(III)-, and Cu(II)-exchanged montmorillonites were prepared by several washes of “as received” Ca-montmorillonite with 1 M aqueous solutions of the corresponding metal chloride, followed by multiple washes with water until the wash solution did not precipitate AgCl(s) when added to 1 M AgNO3(aq). The cation-exchanged montmorillonites were dried at 100 °C and 3 × 10-3 Torr to constant weight. Weighed samples of “as received” STx-1, when placed in a humidifying chamber at room temperature, sorbed water until constant weights were obtained (after about 3 months): 19.5% water (w/w). Chlorpyrifos. Analytical grade chlorpyrifos, I (IUPAC name O,O-diethyl O-3,5,6-trichloro-2-pyridyl phosphorothioate, CAS Registry 2921-88-2) was donated by DowElanco of Indianapolis, IN; Mp 42-43 °C, lit. 42-43.5 °C (15), and was characterized as described previously (2). Chlorpyrifos-Loaded Clays. Two methods were used to sorb water onto a weighed amount (typically 2 g) of the clay minerals, as described previously (2). NMR Spectroscopy. Most of the liquid-solution 1H, 13C, and 31P NMR spectra were collected using a modified Bruker AM-500 NMR spectrometer and a Chemagnetics Infinity600 NMR spectrometer. Solid-state 31P NMR spectra were VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
797
TABLE 1. Some Possible Chlorpyrifos Decomposition Products and Relevant 13P Chemical Shiftsa 31P
species
chemical shifts (ppm)b
I chlorpyrifos II III IV V VI oxon VIIc VIII IX
60.8 55.2 (as K+ salt) est.: ∼50 20 to 24 23 to 31 -6 to -9 est.: ∼30 32 to 34 -2 to -4
31P
species
chemical shifts (ppm)b
Xc XI XIIc XIIIc XIVc XV XVI XVIIc XVIIIc XIX
est.: ∼0 -3 to 6 est.: ∼25 est.: ∼25 est.: ∼ -7 -9 to 0 25 est.: ∼30 est.: ∼25 53 to 56
a Taken from ref 2b. b Relative to 85% H PO (aq) by substitution. 3 4 Compound not typically considered to be a likely decomposition product, or stable, in the environment.
c
obtained at 60.7 and 80.9 MHz on 400-700 mg samples, as described previously (2). 27Al MAS spectra were obtained at 156.3 MHz, using a Bruker AM-600 NMR spectrometer, as described previously (2).
Results 31P
A previous paper presented (and to a much lesser extent, 13C) NMR results on chlorpyrifos sorbed on a whole soil, a humic acid, a kaolinite clay, and a Ca-montmorillonite clay (2b). The results were interpreted in terms of the 31P chemical shifts given in Table 1 for several phosphorus-containing species that one might expect to observe in the decomposition of sorbed chlorpyrifos. Some of the entries in Table 1 (marked in the table by footnote “c”) are typically not considered to be likely decomposition products, and/or are considered very unstable, in the environment; the consideration of such species was not pursued vigorously in the study reported here, which focuses on chlorpyrifos sorbed on various montmorillonite clays (i.e., Ca2+-montmorillonite, Zn2+montmorillonite, Al3+-montmorillonite, and Cu2+-montmorillonite). In the study reported here, solid-state NMR spectra were obtained, mainly with 31P, with magic-angle spinning (MAS, for averaging chemical shift anisotropy), using either cross polarization (CP) or direct polarization (DP). Some details on the techniques on relevant relaxation data and on related issues of quantitation have been presented elsewhere (2). Figure 1 shows 31P DP-MAS and CP-MAS spectra obtained on a sample of chlorpyrifos (9.6% w/w) sorbed on partially hydrated Ca-montmorillonite (STx-1 with 5.0% water by weight) after 1 day and 3.7 years. No decomposition products are apparent after 1 day storage at room temperature (in the dark). However, it is clear from Figure 1C and D that after 3.7 years of storage some chlorpyrifos decomposition has occurred. The tallest peak (61 ppm) seen in the DP-MAS spectrum (Figure 1C), not seen by CP-MAS (Figure 1D), is due to the highly mobile, adsorbed chlorpyrifos phase noted previously with chlorpyrifos/kaolinite (2). This apparently physisorbed-chlorpyrifos peak contributes less than half of the total integrated intensity of the DP-MAS spectrum; thus more than half of the chlorpyrifos (not lost through volatilization) has decomposed to at least three different phosphorus-containing species. In addition to a small amount of aryl-hydrolyzed product (II), represented by the peak around 55 ppm (mostly easily seen by CP-MAS), peaks are seen at 30, 20, and 4 ppm, and a broad intensity is noted between zero and -10 ppm. The peak at 30 ppm is probably due to one of the S,O-exchanged isomers (V); the other S,Oexchanged isomer (IV) may be responsible for the peak at 20 ppm in the 31P MAS spectra after 3.7 years. 798
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006
FIGURE 1. 60.7 MHz 31P DP NMR spectra of chlorpyrifos (9.6% w/w) adsorbed on partially hydrated Ca-montmorillonite (5.0% by wt. water), with 4.5 kHz MAS: (A) DP-MAS spectrum after 1 day; (B) CP-MAS spectrum after 1 day; (C) DP-MAS, spectrum after 3.7 years; (D) CP-MAS, spectrum after 3.7 years. When the sample represented in Figure 1C and D was extracted by acetone-d6 for 2 h at room temperature, only three peaks were seen (at 61, 27, and 24 ppm) in the extract (Figure SI-1C of ref 2b). DMSO-d6 extraction gave the same results. The 61 ppm peak is due to unreacted chlorpyrifos, while the latter two are thought to be due to V and IV, the S,O-exchanged isomers. It appears that the latter two peaks are shifted to 30 and 20 ppm when chlorpyrifos is adsorbed on Ca-montmorillonite (5.0% w/w water content). From DPMAS and CP-MAS 31P experiments (Figure SI-1) on the phosphorus-containing species remaining adsorbed on Camontmorillonite after acetone-d6 extraction (and 10 minutes drying in air), it is clear that much of the phosphoruscontaining material was not extracted. About 80% of the unreacted chlorpyrifos (61 ppm) was extracted, but only about 50% and
