Article pubs.acs.org/JAFC
Comprehensive Screening Study of Pesticide Degradation via Oxidation and Hydrolysis Evelyn Chamberlain,†,‡ Honglan Shi,‡,# Tongwen Wang,‡,# Yinfa Ma,‡,# Alice Fulmer,§ and Craig Adams*,#,⊗ †
Department of Civil, Architectural and Environmental Engineering, ‡Environmental Research Center, and Department of Chemistry, Missouri University of Science and Technology, Rolla, Missouri 66044, United States § Water Research Foundation, 6666 W Quincy Avenue, Denver, Colorado 80235-3098, United States ⊗ Department of Civil, Environmental and Architectural Engineering, University of Kansas, Lawrence, Kansas 66044, United States #
ABSTRACT: This comprehensive study focused on the reactivity of a set of 62 pesticides via oxidization by free chlorine, monochloramine, chlorine dioxide, hydrogen peroxide, ozone, and permanganate; photodegradation with UV254; and hydrolysis at pH 2, 7, and 12. Samples were analyzed using direct injection liquid chromatography−mass spectrometry detection or gas chromatography−electron capture detection after liquid−liquid extraction. Many pesticides were reactive via hydrolysis and/or chlorination and ozonation mechanisms under typical drinking water treatment conditions, with less reactivity exhibited on average for chlorine dioxide, monochloramine, hydrogen peroxide, and UV254. The pyrazole and organophosphorous pesticides were most reactive in general, whereas carbamates and others were less reactive. The screening study provides guidance for the pesticide/oxidation systems that are most likely to lead to degradates in water treatment and the environment. KEYWORDS: pesticide degradation, oxidation, photodegradation, hydrolysis, drinking water
■
INTRODUCTION The widespread use of pesticides in conventional farming practices and modern civilization has led to the increased occurrence of pesticide residues in drinking water supplies. Whereas pesticide applications are most often cited as agricultural, urban usage for private home pest control, golf course maintenance, and municipal pest management also contributes significantly to pesticide pollution in water supplies. Studies over the past few decades have documented the occurrence of pesticides and their degradates in a wide variety of water supplies. Drain-field and stormwater runoff, both urban and agricultural,1,2 surface water,3−5 groundwater,6−10 and finished drinking water11 are the most important for drinking water supplies due to their potential or direct purpose as a potable water source. Studies also indicate that pesticide pollution can have indirect impacts on drinking water through air,12 rainfall,13 and snowfall.14 Degradation of these pesticides may occur in the environment by hydrolysis and photolysis and additionally via oxidation (and reduction) in water and wastewater treatment. Various treatment processes are utilized to remove chemical and biological contaminants from potable water. Studies to date have focused on the removal and/or reactivity of a single or small set of pesticides with a limited number of treatment processes. These studies indicate that many different treatments show the potential for inducing degradate formation through chemical oxidation with disinfectants, photolysis with ultraviolet, hydrolysis at high pH, biodegradation, and ozone.15−23 The toxicity of degradates can potentially be greater than the parent with respect to ecotoxicity24,25 and for humans (e.g., for oxon degradates of organoposphate pesticides26). © 2011 American Chemical Society
The purpose of this study was to identify combinations of 62 pesticides and 6 oxidants (i.e., free chlorine, monochloramine, chlorine dioxide, permanganate, ozone, and hydrogen peroxide), UV, and pH that are most reactive, leading to pesticide degradate formation under typical disinfection exposures (i.e., concentration (C) times time (t), or “CT”) and environmental conditions. This information is important to help focus future work on degradate identification, kinetic modeling, occurrence, and toxicity.
■
EXPERIMENTAL PROCEDURES
Chemicals. All treatment chemicals (sodium hypochlorite, ammonium chloride, etc.) and solvents (e.g., hexane) were of at least reagent grade and were obtained from Fisher Scientific (Fair Lawn, NJ). The pesticides investigated in this study were purchased from Sigma-Aldrich (St. Louis, MO). Pesticide stock solutions were freshly prepared and stored in amber bottles to prevent possible lightinduced decomposition. Laboratory water (Milli-Q, or MQ) with resistivity >18.2 MΩ·cm was produced by a water purification system (model Simplicity 185, Millipore Co., Bedford, MA) to prepare working solutions. Experimental Methods. Experimental Design. The experiments were divided into two main groups, GC compounds and LC compounds, depending on which analytical method was most appropriate. For GC compounds, samples were prepared from five standard mixes (G1−G5) as listed in Table 1. Stock solutions of mixes G1, G2, and G4 were purchased from Sigma-Aldrich at individual pesticide concentrations of 500, 500, and 1000 mg/L, respectively. Received: Revised: Accepted: Published: 354
August 22, 2011 December 2, 2011 December 5, 2011 December 5, 2011 dx.doi.org/10.1021/jf2033158 | J. Agric.Food Chem. 2012, 60, 354−363
Journal of Agricultural and Food Chemistry
Article
Table 1. Hexane LLE Percent Recoveries, Standard Deviation of Recoveries, and MDLs for 36 Compounds Analyzed by GC-ECD compound
mix
mean recovery (%)
acetochlor alachlor aldrin chloradane (α) chloradane (γ) chlorobenzilate chloroneb chlorothalonil chlorpyriphos dacthal diazinon dieldrin dimethenamid endrin etridiazole fonofos heptachlor heptachlor epoxide hexachlorbenzene hexachloropentadiene isoxaflutole lindane linuron malathion methoxychlor metolachlor metribuzin nonachlor (cis) nonachlor (trans) parathion pendimethalin permethrin propachlor propanil terbufos triflualin
G3 G1 G1 G2 G2 G4 G4 G4 G5 G4 G3 G1 G3 G2 G4 G3 G2 G2 G2 G2 G3 G1 G3 G5 G2 G3 G5 G1 G1 G5 G5 G4 G4 G3 G5 G4
100.6 102.0 107.8 98.9 98.5 103.6 99.4 121.6 102.1 105.1 99.1 104.0 105.4 100.2 104.0 99.1 94.5 95.7 132.0 97.3 94.7 105.8 115.6 101.4 85.5 99.8 108.2 93.1 104.9 104.1 97.5 84.6 100.4 125.9 94.7 112.3
recovery RSD (%)
MDL (ng/L)
10.6 22.8 26.5 9.8 17.4 11.2 7.4 10.7 10.1 4.6 7.2 13.3 9.7 9.9 3.5 9.2 12.9 11.3 31.6 18.0 9.7 15.4 45.6 22.4 12.4 8.6 27.5 26.7 20.5 17.9 11.1 17.2 4.3 42.0 10.7 11.8
97.4 57.9 91.8 82.5 80.4 96.8 76.3 37.0 99.0 57.9 117.6 78.8 12.5 71.5 37.1 85.4 39.0 69.8 129.3 94.7 82.5 69.5 97.0 100.7 96.1 120.6 115.4 81.8 65.4 138.8 84.3 63.9 53.0 97.3 110.5 67.1
Table 2. MDLs for 26 Compounds Analyzed by LC-MS compound
mix
MDL (ng/L)
2,4,5-TP 2,4-D 3-hydroxycarbofuran aldicarb aldicarb sulfone aldicarb sulfoxide atrazine bentazon carbaryl carbofuran cyanazine dicamba diuron endothall EPTC fipronil fluometuron methiocarb methomyl molinate oxamyl picloram prometon propazine propoxur simazine
L3 L3 L2 L2 L2 L2 L1 L3 L2 L2 L1 L3 L1 L3 L1 L4 L1 L2 L2 L1 L2 L3 L1 L1 L2 L1
141.8 216.3 203.7 187.3 351.7 164.5 20.9 230.1 344.8 398.4 64.9 98.5 106.9 100.3 80.2 492.6 75.5 367.8 197.9 77.8 262.9 247.6 79.9 16.3 301.7 50.8
analytes. From 7 to 10 samples were taken over various time intervals, depending on the rate of reaction for a given oxidant or condition. For GC compounds, a 35 mL sample was withdrawn after a predetermined time interval (chosen to achieve an oxidant exposure representative of typical drinking water treatment conditions) and then extracted using the hexane LLE method (quenching any further reaction). All processed samples were stored in the dark at −20 °C and injected within 36 h. Oxidant concentrations, pH, and/or UV energy were periodically monitored to allow determination of exposures. Eight common oxidation/disinfection processes were evaluated in this study, including oxidation with free chlorine (FC), monochloramine (MCA), chlorine dioxide (ClO2), ozone (O3), permanganate (MnO4−), ultraviolet light (UV), hydrogen peroxide (H2O2), and hydrolysis at pH 2, 7, and 12 (HYD2, HYD7, and HYD12, respectively). The methods used for oxidant stock preparation are described below. Free Chlorine. FC stock solutions were prepared by dilution from a 5% sodium hypochlorite solution (Fisher Scientific). FC was determined by the difference between total chlorine concentration (as determined with Hach DPD method 8167 using Accuvacs; Loveland, CO) and the monochloramine concentrations (determined with HACH Nitrogen, Free Ammonia, and Chloramine (Mono) Indophenol Method 10200). The typical CT value for 3-log (99.9%) removal of Giardia Cysts using FC at pH 6−9 ranges from 24 to 146 mg·min/L at temperatures from 20 to 25 °C.27 A typical range for the FC concentration used during drinking water disinfection is 0.2−2 mg/L.28 The CT values used for this study ranged from 107 to 173 mg·min/L using FC concentrations of 2−5 mg/L as Cl2. Monochloramine. MCA stock solutions were prepared from sodium hypochlorite and ammonium chloride at a slight excess molar ratio of 1.05:1 (ammonium/hypochlorite) at pH >9 (as described in ref 29). Stable, high-concentration substock solutions of NaOCl and NH4Cl for MCA stock preparation were prepared and stored in the dark. During stock MCA creation, the pH was controlled to maintain a pH of 9.0 or greater by utilizing phosphate-buffered water, monitoring pH, and adding sodium hydroxide as needed.
Stock solutions of mixes G3 and G5 were prepared with concentrations of each pesticide at 191 and 179 mg/L, respectively. For LC compounds, samples were prepared from four standard mixes (L1−L4), as listed in Table 2. Stock solutions of mix L2 were purchased from Sigma-Aldrich at individual pesticide concentrations of 100 mg/L. Stock solutions of mixes L1, L3, and L4 were prepared with concentrations of each pesticide at 52 mg/L, 100 mg/L, and 77.8 mg/L, respectively. Experiments were conducted in 5−10 mM sodium phosphate buffered laboratory water at 23 ± 1 °C at pH 6.6 and 8.6 in 250 mL amber glass chemical reactors. A shaker table was used for mixing at 200 rpm. GC compound experiments were initiated by adding a stock pesticide mix solution to an initial concentration of 1.5−3 μg/L of each study compound, mixing, and then removing a 35 mL sample for initial concentration determination. For LC compound experiments, an initial concentration of 25 μg/L and a 1 mL sample volume were used. Next, the remaining solution was spiked with an oxidant, subjected to hydrolysis by pH adjustment, or subjected to UV photolysis at 254 nm. For LC compounds, 1 mL samples were periodically taken from the reactor and immediately injected into the LC-MS for immediate analysis. The injection time was recorded as the quench time due to immediate chromatographic separation of the oxidant from the 355
dx.doi.org/10.1021/jf2033158 | J. Agric.Food Chem. 2012, 60, 354−363
Journal of Agricultural and Food Chemistry
Article
The presence of a slight excess of ammonia was confirmed in the MCA stock solution using an ammonia probe (model Orion 9512, ThermoElectron Corp., Waltham, MA) to ensure that the MCA stock solution contained no FC. Additionally, absorbance scans of the MCA stock solution at pH 9.0 were also prepared from 200 to 400 nm to confirm the absence of FC. During experiments with MCA, the pH was controlled to maintain a pH of either 6.6 or 8.6 by utilizing phosphatebuffered water and monitoring pH. MCA concentrations during experiments were determined using Hach Nitrogen, Free Ammonia, and Chloramine (Mono) Indophenol Method 10200 using chemicals obtained from the Hach Co. (Loveland, CO). The typical CT value for 3-log (99.9%) removal of Giardia Cysts using chloramines at pH 6−9 ranges from 750 mg·min/L at 25 °C to 1100 mg·min/L at 20 °C.27 A typical range of concentrations used during drinking water disinfection is 1−4 mg/L.28 The CT values used for this study ranged from 1287 to 1430 mg·min/L using MCA concentrations of 9−14 mg/L as NH2Cl. Permanganate (MnO4−). Permanganate stock solutions were prepared by dissolving potassium permanganate solid in laboratory water. The permanganate concentration was determined using Hach DPD Method 8167 using Accuvac ampules. Permanganate is not commonly used as a disinfectant, but is used to control taste and odor compounds or to oxidize reduced manganese at concentrations in the range of 0.2−20 mg/L. Permanganate has also been shown to control Asiatic clams and zebra mussels at concentrations of 1.1−4.8 and 0.5−2.5 mg/L, respectively.28 The CT values used for this study ranged from 134 to 164 mg·min/L using permanganate concentrations of 3−5 mg/L as MnO4−. Chlorine Dioxide. Gaseous ClO2 was produced using a CDG Bench Scale ClO2 Generator (CDG, Bethlehem, PA). The gaseous chlorine dioxide stream was bubbled into phosphate-buffered laboratory water. The ClO2 concentration was measured by absorbance at 359 nm with a UV−vis spectrophotometer (Cary 50 Conc., Varian). The typical CT value for 3-log (99.9%) removal of Giardia cysts using ClO2 at pH 6−9 ranges from 11 to 15 mg·min/L at a temperature of 20−25 °C.27 A typical range of concentrations used during drinking water disinfection is 0.07−2 mg/L.28 The CT values used for this study ranged from 38 to 73 mg·min/L using ClO2 concentrations of 2−3 mg/L as ClO2. Hydrogen Peroxide. H2O2 stock solutions were prepared by dilution of a 30% H2O2. H2O2 is not typically utilized independently for drinking water disinfection, but rather in conjunction with UV or ozone as an advanced oxidation process. The CT values used for H2O2 in this study ranged from 933 to 1100 mg·min/L using H2O2 concentrations of 100 mg/L as H2O2. Ozone. Gaseous O3 was produced from compressed oxygen using a corona discharge ozone generator (model GLS-1, PCI-WEDECO Environmental Technologies, West Caldwell, NJ). The gaseous O3 stream was bubbled into phosphate-buffered laboratory water to create the ozone stock solution. O3 concentration was measured using UV absorbance at 260 nm. The O3 stock was then spiked into the samples and allowed to naturally degrade while being orbitally mixed at 200 rpm for 2 h. O3 dosing was conducted in duplicate with one sample being used to monitor the O3 concentration decay for calculation of CT (concentration × time) and the other used for subsequent chemical analysis. Samples were extracted as previously described. The typical CT value for 3-log (99.9%) removal of Giardia cysts using O3 at pH 6−9 ranges from 0.48 to 0.72 mg·min/L at a temperature of 20−25 °C.27 A typical range of O3 concentration used during drinking water disinfection is 0.995, indicating good linearity over the range of concentrations investigated. MDLs were determined via standard methods by injecting the seven separate samples, at an estimated 3 times the anticipated MDL, and multiplying the standard deviation of the concentration by 3.14. For experiments, the average %RSD between duplicate experiments was 357
dx.doi.org/10.1021/jf2033158 | J. Agric.Food Chem. 2012, 60, 354−363
Journal of Agricultural and Food Chemistry
Article
Figure 2. Chromatograms for LC-MS analysis of study compounds including internal standards (PCNB and TCX) and surrogate (DCB). Pesticide compounds were at 30 μg/L. 2.6%. The MDL for direct injection LC-MS for each compound ranged from 16.2 to 492.6 ng/L (Table 2). Quenching. For LC compounds, injection into the column provided immediate quenching for oxidation experiments (with free chlorine, monochloramine, chlorine dioxide, hydrogen peroxide, and permanganate). Ozone was allowed to completely decay while tracking ozone concentration in order to calculate exposure (mg·L/min). For GC compounds, extraction of compounds into the hexane phase served as the quenching step for oxidation experiments with FC, MCA, ClO2, H2O2, and MnO4−. Separate experiments demonstrated negligible oxidant reactivity with the pesticides, internal standard, and surrogate compounds in the hexane phase.
which may include excess lime softening (e.g., at pH 11.5), free chlorine, and ozone treatments. Specifically, for ozone, 29 and 23% of the pesticides were highly reactive with ozone at pH 6.6 and 8.6, respectively, whereas 23 and 21% were highly reactive with free chlorine at pH 6.6 and 8.6, respectively (Table 3). Of the 62 compounds studied, 90, 82−87, 92−97, and 97−98% had low reactivity with MCA, ClO2, H2O2, and UV254, respectively (Table 3). Approximately half of the pesticides highly hydrolyzed at pH 12, a pH just slightly greater than the 11.3 commonly used for excess lime softening. More study of the extent of hydrolysis of these compounds in water treatment plants is warranted. Just 16 and 11% were highly hydrolyzed (unstable) at pH 2 and 7, respectively. Over all the reactants, the classes of compounds having the highest percentage of highly reactive pesticides were (Table 3) pyrazole > organophosphorous > (triazinone, benzothiadiazole) > carbamate > others. With FC, slightly higher reactivity was observed at lower pH consistent with the stronger oxidation power of hypochlorous acid as compared with hypochlorite (although speciation of the pesticide affects reactions rates as well). All of the organophosphorus compounds were highly reactive with FC, with aldrin, fipronil, hexachloropentadiene, and selected carbamates also showing high reactivity (Table 3). Only two carbamates, aldicarb and methiocarb, and two organophosphorus pesticides, fonofos and terbofos, were highly reactive with MCA. All other pesticides were recalcitrant to MCA except hexachloropentadiene and heptachlor. ClO2 was only highly reactive with fipronil and metribuzin at higher pH and aldicarb, methiocarb, fonofos, and turbofos at both pH levels (Table 3). ClO2 was recalcitrant with all other
■
RESULTS AND DISCUSSION Overall Trends of Treatments. The results of the screening experiments showed a wide range of reactivities for pesticides by oxidants, hydrolysis, and UV254 under drinking water treatment conditions. Table 3 lists the compounds studied by class and their reactivity with each oxidant and/or condition, grouped with designations of H (“high”) for >50% removal, M (“moderate”) for between 20 and 50% removal, and L (“low”) for 50% removal under typical conditions), the order of reactivity of the study compounds was (Table 3) as follows: (hydrolysis)high pH > O3 > FC > (hydrolysis)low pH > ClO2 > MnO4− > MCA > (H2O2 = UV254). Combining compounds with either high or medium reactivity (i.e., >20% removal), the order of reactivity was (Table 3) (hydrolysis)high pH > O3 > (hydrolysis)low pH > FC > ClO2 > MnO4− > MCA > H2O2 > UV254. Thus, these results suggest conditions that may lead to greater parent removal, and the most significant degradate formation, 358
dx.doi.org/10.1021/jf2033158 | J. Agric.Food Chem. 2012, 60, 354−363
CAS Registry No.
93-72-1 94-75-7 1918-00-9 62059-43-2 1918-02-1 34256-82-1 15972-60-8 87674-68-8 51218-45-2 1918-16-7 709-98-8 25057-89-0 116-06-3 1646-88-4 1646-87-3 63-25-2 1563-66-2 759-94-4 16655-82-6 2032-65-7 16752-77-5 2212-67-1 23135-22-0 114-26-1 40487-42-1 1582-09-8 141112-29-0 309-00-2 5103-71-9 5103-74-2 2675-77-6 1897-45-6 60-57-1 72-20-8 76-44-8 1024-57-3 118-74-1 77-47-4 58-89-9 72-43-5 5103-73-1
compound
2,4,5-TP (Silvex) 2,4-D dicamba endothal picloram acetochlor alachlor dimethenamid metolachlor propachlor propanil bentazone aldicarb aldicarb sulfone aldicarb sulfoxide carbaryl carbofuran EPTC 3 - hydroxycarbofuran methiocarb methomyl molinate oxamyl propoxur pendimethalin trifluralin isoxaflutole aldrin chloradane (α) chloradane (γ) chloroneb chlorothalonil dieldrin endrin heptachlor heptachlor epoxide hexachlorbenzene hexachloropentadiene lindane methoxychlor nonachlor (cis)
acid acid acid acid acid amide amide amide amide amide amide benzothiadiazole carbamate carbamate carbamate carbamate carbamate Carbamate carbamate carbamate carbamate carbamate carbamate carbamate dinitroaniline dinitroaniline isoxazole organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride organochloride
class LC LC LC LC LC GC GC GC GC GC GC LC LC LC LC LC LC LC LC LC LC LC LC LC GC GC GC GC GC GC GC GC GC GC GC GC GC GC GC GC GC
method L L L L L L L M L L L L H L H L L H L H H H M L L L L H L L L L L L L L L M L L L
pH 6.6
FC
L L L L L L L L L L L L H L L L L H L H H H L L L L L M L L L L L L L L L H L L L
pH 8.6 L L L L L L L L L L L L H L L L L L L H L L L L L L L L L L L L L L M L L M L L L
pH 6.6
MCA
L L L L L L L L L L L L H L L L L L L H L L L L L L L L L L L L L L M L L M L L L
pH 8.6 L L L L L L L L L L L H H L L L L L L H L L L L L L L L L L L L L L L L L M L L L
pH 6.6
ClO2
L L L L L L L L L L L H H L L L L L L H L L L L L L L L L L L L L L L L L M L L L
pH 8.6 L L L L L L L L L L L L L L H L L L L M L L L L L L L H L L L L L L L L L H L L L
pH 6.6 L L L L L L L L L L L L M L H L L L L M L L L L L L L H L L L L L L L L L H L L L
pH 8.6
MnO4−
Table 3. Screening Study Results, Where H = >50% Removal, M = 20−50% Removal, and L =
