Formation of trifluoroacetic acid from common trifluoromethyl pesticides in agricultural soils

Abstract

The short‐chain per‐ and polyfluoroalkyl substances trifluoroacetic acid (TFA) has been recognized as a degradation product of only three pesticides, but many other trifluoromethyl pesticides (C–CF₃ pesticides) may potentially produce TFA based on their molecular structure. We determined TFA formation from seven C–CF₃ pesticides with high sales rates (trifluralin, fluazifop‐P‐butyl, fluazinam, tau‐fluvalinate, diflufenican, fluopyram, and mefentrifluconazole) over 52 weeks in a laboratory experiment with three agricultural soils. TFA was formed from all seven pesticides, but TFA formation varied both among compounds and among soils. Fluopyram had the highest TFA formation, at 10.7% of the pesticide amount added in one soil, but lower formation in the two other soils (3.3% and 7.9%). Fluazinam also had relatively high TFA formation (5.3%–6.5%), but unlike fluopyram, only minor differences were seen between soils. TFA formation was lower for the other pesticides with 2.4%–5.2% for diflufenican, 3.0%–5.3% for fluazifop‐P‐butyl, 1.1%–2.2% for trifluralin, 0.8%–3.3% for mefentrifluconazole, and 0.4%–1.0% for tau‐fluvalinate. The measured TFA formation was used to estimate how spraying with C–CF₃ pesticides may increase the annual average TFA concentrations in net precipitation infiltrating through arable topsoil (vadose zone water). The estimated TFA concentrations depended on crop type and growth stage and showed that the yearly TFA contribution from some C–CF₃ pesticides may equal the atmospheric contribution in infiltrating soil water below sprayed fields.

Core Ideas

• Seven C–CF₃ pesticides were tested for the formation of trifluoroacetic acid (TFA) in agricultural soil microcosms.

• TFA was formed from all tested pesticides.

• Conversion to TFA is likely to continue beyond 1 year.

• TFA derived from C–CF₃ pesticides may equal atmospheric input in infiltrating water below treated fields.

Plain Language Summary

Up to now, trifluoroacetic acid (TFA) was thought to come from the breakdown of only a few trifluoromethyl pesticides. However, based on the molecular structure of other similar pesticides, it seemed likely that many more could also produce TFA. Since the use of these pesticides has been rising steeply in recent years, we studied how seven common trifluoromethyl pesticides break down to TFA in agricultural soil. All seven produced measurable amounts of TFA. Simple estimates showed that the average TFA concentrations resulting from pesticide application may be at the same level as TFA from rain in newly formed groundwater below fields.

Abbreviations

FOCUS

Forum for the Coordination of Pesticide Fate Models and Their Use

FOEN

Federal Office for Environment

PE

polyethylene

SOM

soil organic matter

TFA

trifluoroacetic acid

UPLC–MS/MS

ultra‐performance liquid chromatography coupled to tandem mass spectrometry

1. INTRODUCTION

Trifluoroacetic acid (TFA) in precipitation represents a main source of TFA in the environment (Freeling et al., 2020). However, recent monitoring data indicate that local, agricultural sources contribute substantially to the TFA load in groundwater beneath arable land (Albers & Sültenfuss, 2024; Federal Office for Environment [FOEN], 2024). TFA is a short‐chain per‐ and polyfluoroalkyl substance that is chemically stable and nonbiodegradable in soil and groundwater (Alexandrino et al., 2018; Scheurer et al., 2017). In addition, TFA is highly mobile in most soils (Richey et al., 1997), resulting in a high potential for leaching to aquifers. A recent review concluded that “TFA meets the criteria of a planetary boundary threat for novel entities because of increasing planetary‐scale exposure, where potential irreversible disruptive impacts on vital earth system processes could occur” (Arp et al., 2024). Consistent with these concerns, German government agencies have classified TFA as a reproductive toxicant (Category 1B) and as “very persistent and very mobile” (Umwelt Bundesamt, 2025).

The pesticides flufenacet (European Food Safety Authority [EFSA], 2017, 2024), and with less certainty flurtamone (European Food Safety Authority [EFSA], 2017) and haloxyfop (European Food Safety Authority [EFSA], 2009), have been recognized as pesticide sources of TFA in soil. Theoretically, TFA may also be formed through the biodegradation of other pesticides containing one or more trifluoromethyl groups linked to a carbon atom (C–CF₃ pesticides), via a speculative pathway in which the parent compound is progressively degraded until TFA remains as the final persistent residue. C–CF₃ pesticides could contribute substantially to TFA in groundwater and surface waters. The relevance of this potential source is underscored by the increasing use of C–CF₃ pesticides in recent years, as illustrated by Danish pesticide sales statistics (Figure S1) and corresponding estimates of TFA equivalents (Figure 1; Figure S2). Comparable theoretical TFA loads from C–CF₃ pesticides have been reported for many regions of Europe and the United States (Joerss et al., 2024).

FIGURE 1.

Annual sales rates of C–CF₃ pesticides in Denmark expressed as potential trifluoroacetic acid (TFA) equivalents (100% conversion to TFA of all applied C–CF₃ pesticides) per km² of agricultural land.

It is not evident, however, that the presence of a trifluoromethyl group linked to a carbon atom necessarily results in degradation to TFA. Some aromatic C–CF₃ compounds may instead undergo mineralization to free fluoride. This pathway has been demonstrated for the C–CF₃ solvent benzotrifluoride (Wackett, 2022; Yano et al., 2015), in which a dihydrodiol is formed following attack by bacterial toluene dioxygenase, after which the molecule rearranges with release of one fluoride ion and subsequently reacts with water, releasing the remaining two fluoride ions. In contrast, microbial defluorination of the pharmaceutical C–CF₃ compound fluoxetine has been shown to result in concurrent accumulation of TFA, indicating at least two degradation pathways (Kahn & Murphy, 2021).

Given the existence of two reaction pathways, the central research questions of this study were as follows: (i) Is degradation to TFA a common pathway for C–CF₃ pesticides in arable soil? (ii) If so, are the amounts of TFA formed sufficient to affect groundwater quality beneath fields sprayed with C–CF₃ pesticides? To address these questions, we selected seven C–CF₃ pesticides with high historical and/or current sales rates (Figure S1): fluopyram (fungicide), fluazinam (fungicide), diflufenican (herbicide), fluazifop‐P‐butyl (herbicide), trifluralin (herbicide), mefentrifluconazole (fungicide), and tau‐fluvalinate (insecticide). These pesticides were tested for TFA formation in small soil microcosms incubated for up to 1 year. The resulting TFA formation data were subsequently used to estimate the potential contribution of C–CF₃ pesticides to annual average TFA concentrations in net precipitation infiltrating from arable topsoil (vadose zone water) under selected agricultural scenarios.

2. MATERIALS AND METHODS

2.1. Reagents

Analytical‐grade fluazinam, trifluralin, fluopyram, tau‐fluvalinate, diflufenican, fluazifop‐P‐butyl, fluazifop‐P, and TFA were obtained from Sigma‐Aldrich/Merck. Mefentrifluconazole and mefentrifluconazole‐d4 (internal standard) were obtained from LCG Standards. Sodium trifluoroacetate‐¹³C₂ internal standard was obtained from LGC Standards (product code TRC‐S673752). Ultrapure water was produced using a Milli‐Q purification system (18.2 MΩ cm; total organic carbon < 5 ppb).

2.2. Soils

Soil was sampled from the upper 15 cm of three agricultural fields at Svanholm (Skibby, Denmark). The fields had been organically managed for at least 30 years and were selected to avoid interference from pesticides potentially accumulated in conventional fields (Riedo et al., 2022). Field Sv1 had been under crop rotation for at least 30 years and was characterized by sandy soil, 3.8% soil organic matter (SOM) determined by loss on ignition at 550°C for 3 h, pH 7.3 measured in a 1:2.5 soil–water slurry with ultrapure water, and pH 6.7 determined in a 1:2.5 soil–CaCl₂ solution (10 mM) slurry. Field Sv3 had also been under crop rotation for at least 30 years but consisted of clayey soil with 4.5% SOM, pH 5.7 in water, and pH 5.2 in CaCl₂ solution. Field Sv6 had been under permanent grass for the preceding 15 years and consisted of sandy soil, 3.7% SOM, pH 6.7 in water, and pH 6.2 in CaCl₂ solution. Soil samples were collected on April 26, 2023, as composite samples consisting of 20 subsamples (100 g each) taken from 0‐ to 15‐cm depth within an area of approximately 20 m². The samples were stored in polyethylene (PE) bags at 5°C for 12 days. Before use, the soils were sieved to 2 mm and preincubated in loosely closed PE bags at 20°C for 2 days.

Core Ideas

• Seven C–CF₃ pesticides were tested for the formation of trifluoroacetic acid (TFA) in agricultural soil microcosms.

• TFA was formed from all tested pesticides.

• Conversion to TFA is likely to continue beyond 1 year.

• TFA derived from C–CF₃ pesticides may equal atmospheric input in infiltrating water below treated fields.

2.3. Microcosm experiment

To assess TFA formation from the pesticides, soil was spiked with the seven C–CF₃ pesticides individually in test tube microcosms and incubated for up to 1 year under oxic conditions. The microcosms were placed in sealed boxes with a water‐saturated atmosphere to minimize water loss. Microcosms without added pesticides served as negative controls to assess background TFA levels and potential TFA contamination during sample handling and analysis. All parent compounds were added at a final concentration of 10 µmol/kg dry weight. Concentrations of the parent compounds were determined in separate test tubes at the start and end of the experiment. Soil water content was determined gravimetrically by drying approximately 20 g overnight at 105°C. Quartz sand was acid‐washed in 6 M HCl, rinsed with ultrapure water to neutral pH, and baked at 550°C. For preparation of the microcosms, 0.5 g of baked sand (±0.05 g) was added to each of 576 12‐mL Pyrex test tubes. Subsequently, 100 µL of pesticide solution (250 µM in methanol) was added to each tube and evenly distributed over the sand. The test tubes were left overnight in a fume hood to evaporate the methanol solvent. The pesticide solutions were analyzed for TFA content. Control test tubes received 100 µL of methanol without pesticide. Field‐moist soil corresponding to 2.50 ± 0.05 g dry weight was then added to each test tube, with one soil type combined with one pesticide at a time. Soil and spiked sand were mixed using a metal spatula until the sand was evenly distributed; the same spatula was used for all test tubes containing the same combination of soil and pesticide. The open test tube microcosms were weighed and placed in plastic boxes containing wet filter paper, with each box containing only a single pesticide–soil combination. The boxes were incubated in the dark at 20°C. After 0, 4, 10, 24, 38, and 52 weeks of incubation, triplicate test tubes were removed for each treatment, sealed with screw caps fitted with homemade aluminum liners, and stored at −18°C until extraction of TFA or parent compounds. The remaining test tubes were weighed, and ultrapure water was added as needed to compensate for evaporation.

2.4. TFA and fluoride concentrations

TFA was extracted from the test tube microcosms by adding 5 mL of 10 mM KCl. The test tubes were sonicated for 15 min, placed on a shaker table overnight, and subsequently centrifuged at 1500 g for 15 min. The supernatant was filtered through 0.22‐µm polyethersulfone syringe filters, diluted 10‐fold with ultrapure water, spiked with TFA‐¹³C₂ as an internal standard, and analyzed in random order by ultra‐performance liquid chromatography coupled to tandem mass spectrometry (UPLC–MS/MS) as described previously (Albers & Sültenfuss, 2024). Briefly, TFA was analyzed using electrospray ionization in negative mode with an Atlantis Premier BEH C18 AX FIT column (150 mm × 2.1 mm, 1.7 µm, Waters) operated at 50°C and a flow rate of 0.25 mL/min, using a gradient of acetonitrile (A) and 2 mM ammonium acetate (B). The ion‐trace m/z 113 → 69 was used for quantification. All caps, liners, filters, and syringes were tested for potential TFA release to minimize contamination during sample handling. Method precision was verified by analysis of three samples at the commercial laboratory Eurofins. The limit of quantification was 0.03 µg/L. TFA formation from pesticides was calculated as the difference between pesticide‐spiked microcosms and negative controls. Fluoride concentrations in extracts from the final sampling point were determined by ion chromatography (Metrohm 819 IC equipped with a Metrosep A 150/4.0 column). The analyzed extracts included control microcosms without pesticide and control microcosms spiked with TFA (0.5 µmol/kg dry weight) and incubated for 52 weeks. For each soil, calibration curves were established by spiking with 0, 17, 50, 170, or 500 µg/kg fluoride, followed by extraction with KCl as described above.

2.5. Concentrations of parent compounds

Stock solutions of 1000 mg/L of fluopyram, fluazinam, diflufenican, trifluralin, fluazifop‐P‐butyl, fluazifop‐P, tau‐fluvalinate, and mefentrifluconazole were prepared in methanol for external calibration and stored at −20°C. Soil microcosm test tubes were extracted by adding 6 mL of acetonitrile containing 1% formic acid. The test tubes were sonicated for 15 min, placed on a shaker table overnight, and subsequently centrifuged at 1500 g for 15 min. The supernatant was filtered through 0.22‐µm polytetrafluoroethene syringe filters, diluted 20‐fold in ultrapure water:methanol (5:4, v/v), and spiked with mefentrifluconazole‐d4 as deuterated internal standard. Samples were analyzed using UPLC–MS/MS using an Acquity Classic UPLC system coupled to a Xevo TQ‐S micro mass spectrometer (Waters) using an Acquity UPLC HSS C18 column (2.1 mm × 100 mm, 1.8 µm particles). The column oven was set to 30°C. The mobile phase consisted of (A) acetonitrile with 0.1% formic acid and (B) ultrapure water with 0.1% formic acid. A 13‐min gradient was applied at a flow rate of 0.3 mL/min: 0–0.5 min 15% A; 0.5–6 min 0%→95% A; 6–10 min 95% A; 10–10.5 min 15%→0% A; 10.5–13 min 15% A. The injection volume was 5 µL. The mass spectrometer was operated with electrospray ionization in positive mode, except for fluazifop‐P, which was analyzed in negative mode. Ionization parameters were as follows: desolvation temperature, 500°C; desolvation gas, 1000 L/h; cone gas, 30 L/h; capillary voltage, 0.3 V; and source temperature, 150°C. Trifluralin could not be detected using electrospray ionization under these conditions.

Parent compounds were analyzed using both a quantifier ion and a qualifier ion. Cone voltage, collision energies, quantifier ions, and qualifier ions are presented in Table S2. Limits of quantification are presented in Table S3. Only data for which the qualifier to quantifier ion ratio in soil samples deviated by <35% from that of the external standards were accepted. The method was linear over the concentration range 0.3–90 µg/L when 5 µL was injected, and the calibration standards were prepared in blank extracts from the Sv6 soil. Tau‐fluvalinate produced a comparatively low signal and showed linearity only in the range of 3–60 µg/L. Tau‐fluvalinate concentrations were therefore determined with lower accuracy than those of other parent compounds. Method repeatability was 1.6%–5.4% relative deviation, as determined from diluted soil extracts spiked with parent compounds at concentrations corresponding to 10% of the amount added in the TFA formation incubations. Recoveries ranged from 89% to 108% and were determined as the mean of triplicate soil samples spiked with 10% of the parent compound concentration used in the TFA formation incubations. The only exception was fluazifop‐P‐butyl, for which the recovery was 70% due to hydrolysis to fluazifop‐P. Accordingly, in the TFA‐formation experiment with fluazifop‐P‐butyl, the remaining parent compound was therefore quantified as the sum of fluazifop‐P‐butyl and fluazifop‐P.

2.6. Leaching estimates

Examples of TFA leaching from arable topsoil were estimated for selected C–CF₃ pesticide products corresponding to each of the C–CF₃ pesticides used in the TFA microcosm experiment. The selected products represent the highest annual field dose that is currently approved or has previously been approved in Denmark. An exception is fluazinam, for which estimates were based on a single application, although this compound may be used up to 10 times at different crop growth stages. Product‐specific application doses and approved growth stages (BBCH stages; Forum for the Co‐ordination of Pesticide Fate Models and Their Use [FOCUS], 2014) were obtained from manufacturer user instructions retrieved from the Danish Environmental Protection Agency's Bekæmpelsesmiddeldatabasen (BMD) database (https://mst.dk/erhverv/sikker‐kemi/database‐for‐bekaempelsesmidler/bmd). The following products were included: Propulse SE 250 (fluopyram), ProTector (fluopyram), Diflanil 500 SC (diflufenican), Balaya (mefentrifluconazole), Lenvyor (mefentrifluconazole), Himalaya Pro (mefentrifluconazole), Evure Neo (tau‐fluvalinate), Mavrik 2F (tau‐fluvalinate), Frowncide (fluazinam), Shirlan Ultra (fluazinam), Ohayo (fluazinam), and LFS‐Fluazifop‐P‐butyl (fluazifop‐P‐butyl), for use in cereals, potatoes, or maize. For trifluralin, no user instructions were available in the BMD database; therefore, official recommendations for the legacy product Linulan were used (Ravn, 1979).

The average TFA concentration in net precipitation infiltrating from arable topsoil (C_a ) during the first year after spraying was estimated from the following equation:

$$C_a=D\times F_{\mathrm{ground}}\times F_{\mathrm{TFA}}\times \left(M_{\mathrm{TFA}}/M_{\mathrm{parent}}\right)\times n/Q_{10}/P_N,$$ (1)

where D is the dose of active compound (µg/m²); F _ground is the fraction of pesticide sprayed directly on the ground, that is, the fraction not intercepted by plants at the relevant growth stage according to the Forum for the Coordination of Pesticide Fate Models and Their Use (FOCUS) guideline (Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014); F _TFA is the fraction of the pesticide converted to TFA after 1 year in the microcosm experiment (geometric mean; Table 1); M _TFA and M _parent are the molecular masses of TFA and the parent compound, respectively; n is the number of TFA groups per pesticide molecule (1 or 2); Q ₁₀ is a factor of 2.58 to translate TFA formation determined at 20°C–10°C (Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014), which is more realistic for average in situ temperatures; and P_N is a net precipitation of 300 L/m², corresponding to the Hamburg climate scenario often used in pesticide approval for northern Europe (Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014).

TABLE 1.

TFA formation from degradation of C–CF₃ pesticides as a percentage of the parent compound's potential TFA equivalents in three organically managed, agricultural soils (Sv1, Sv3, and Sv6) determined after 52 weeks of incubation. For each compound, the transformation is shown as mean ± 1 standard deviation (n = 3) and as overall geometric mean.

Compound	Sv1 (%)	Sv3 (%)	Sv6 (%)	Geometric mean (%)
Fluopyram	7.9 ± 0.3	3.3 ± 0.2	10.7 ± 0.5	6.5
Fluazinam	6.5 ± 0.2	5.3 ± 0.2	6.0 ± 0.4	5.9
Diflufenican	2.9 ± 0.8	2.4 ± 0.1	5.2 ± 1.1	3.3
Fluazifop‐P‐butyl	3.0 ± 0.0	3.1 ± 0.0	5.3 ± 0.0	3.7
Trifluralin	1.7 ± 0.0	1.1 ± 0.1	2.2 ± 0.1	1.6
Mefentrifluconazole	2.6 ± 0.2	0.8 ± 0.0	3.3 ± 0.1	1.9
tau‐Fluvalinate	1.0 ± 0.1	0.4 ± 0.1	1.0 ± 0.1	0.7

Abbreviation: TFA, trifluoroacetic acid.

3. RESULTS

3.1. TFA formation

TFA formation was tested for seven C–CF₃ pesticides in small soil microcosms. Figure 1 shows the TFA formation over time, and Table 1 summarizes TFA formation at the end of the experiment after 1 year of incubation. TFA formation varied both among compounds and among soils. Fluopyram showed the highest TFA formation, with 10.7 ± 0.4% of the added fluopyram converted to TFA in Sv6 soil, lower TFA formation in Sv1 (7.9 ± 0.3%), and considerably lower formation in Sv3 (3.3 ± 0.2%). Fluazinam also showed high conversion to TFA, but in contrast to fluopyram, only minor differences were observed between the soils. Fluopyram and fluazinam each contain two C–CF₃ moieties per molecule; expressed on a molar basis (moles of TFA per mole of fluopyram or fluazinam), TFA formation from these compounds would therefore be twice the percentages reported in Table 1.

TFA formation was lower for the remaining compounds, with 2.4%–5.2% for diflufenican, 3.0%–5.3% for fluazifop‐P‐butyl, 1.1%–2.2% for trifluralin, 0.8%–3.3% for mefentrifluconazole, and 0.4%–1.0% for tau‐fluvalinate (Table 1). Although TFA formation for tau‐fluvalinate was low, it was consistently above background levels (Figure S3). For most compounds, TFA formation was highest in Sv6 (sandy soil with permanent grass) and lowest in Sv3 (clay soil with crop rotation). Differences in the size and composition of microbial communities among soils may contribute to this variability, but no further mechanistic explanation could be identified. For most incubations, TFA formation was close to linear and did not reach a plateau during the experimental period (Figure 2). Fluazinam, however, showed clearly decreasing TFA formation rates.

FIGURE 2.

Trifluoroacetic acid (TFA) formation from the degradation of C–CF₃ pesticides in soil from three organically managed agricultural fields (Sv1, Sv3, and Sv6, n = 3 for each soil). Pesticides were added at a concentration of 10 µmol/kg dry weight.

3.2. Parent compounds

We quantified six of the parent compounds with electrospray UPLC‐MS. Mefentrifluconazole and fluopyram were the most persistent, with large fractions remaining in the soils after 52 weeks of incubation (Table 2). Diflufenican and tau‐fluvalinate showed intermediate persistence, whereas fluazinam and fluazifop‐P‐butyl+fluazifop had largely disappeared. Fluazifop‐P‐butyl showed low recovery in the time‐zero samples because the butyl group was rapidly hydrolyzed in soil, yielding fluazifop as the primary reaction product. Residual concentrations were therefore calculated as the sum of fluazifop‐P and fluazifop‐P‐butyl in both the initial and final samples. Trifluralin could not be quantified with electrospray UPLC‐MS/MS.

TABLE 2.

Remaining parent compound (C/C ₀) in three organically managed, agricultural soils (Sv1, Sv3, and Sv6) after 52 weeks of incubation (mean ± 1 SD, n = 3).

Compound	Sv1 (%)	Sv3 (%)	Sv6 (%)
Fluopyram	61 ± 13	75 ± 11	70 ± 2
Fluazinam	2 ± 0	6 ± 1	3 ± 0
Diflufenican	26 ± 5	38 ± 7	28 ± 6
Fluazifop‐P+fluazifop‐P‐butyl	<0.3	1 ± 0	<0.3
Mefentrifluconazol	81 ± 1	87 ± 1	80 ± 1
tau‐Fluvalinate	12 ± 2	20 ± 2	13 ± 2

3.3. Fluoride

At the last sampling point, we measured fluoride in the extracts. The pesticides were applied at low, semi‐realistic concentrations to minimize concentration bias; consequently, any fluoride released from pesticide mineralization was expected to be small relative to background fluoride levels in the soils. Given background variability in the controls, a net increase of at least 0.15 mg/kg fluoride was required to reliably demonstrate formation of free fluoride. A fluoride concentration of 0.15 mg/kg corresponds to mineralization of 26% of the C–CF₃ groups at 10 µmol/kg for pesticides with one C–CF₃ group and 13% for pesticides with two groups. However, no clear difference was observed between incubations with and without pesticide (Figure S4), and none of the incubations showed fluoride increases of more than 0.15 mg/kg. This suggests that the marked discrepancy between TFA formation and parent‐compound removal was generally not attributable to mineralization to fluoride, although we cannot exclude that a fraction of the pesticides (<26% and <13%, respectively) was mineralized.

3.4. Estimates of leaching from sprayed fields

The measured TFA formation was used to estimate theoretical TFA leaching from arable topsoil during the first year after pesticide application (Table 3). TFA is especially suited for such simple estimates as sorption and degradation are negligible. The estimates provide an indication of the concentrations that may be expected for pesticide‐derived TFA in infiltrating water; however, they are associated with substantial uncertainty and should not be interpreted as risk assessments. In addition, field application rates may differ among other products and across countries.

TABLE 3.

Estimated contribution of spraying with C–CF₃ pesticides to the annual TFA average concentration in water infiltrating from treated fields.

Crop	Compound	Yearly dose (g/ha)	BBCH intervals and estimated TFA concentration (µg/L)
Cereals			00–12	13–19	20–29	30–39	40–69
	Fluopyram	125	–	–	–	0.12	0.06
	Diflufenican	125 a	0.08	0.15	–	–	–
	Mefentrifluconazole	100	–	–	–	0.01	–
	Trifluralin	820	0.58	–	–	–	–
	tau‐Fluvalinate	96	0.02	0.02	0.02	0.00	0.00
Potato			00–09	10–19	20–39	40–89	90–99
	Fluopyram	112.5	–	–	–	0.08	–
	Fluazinam	200 b	–	–	0.30	0.11	0.37
	Fluazifop‐P‐butyl c	375	0.53	0.45	0.21	–	–
	tau‐Fluvalinate	96	0.02	0.02	0.01	0.00	0.01
Maize			00–09	10–19	20–39	40–89	90–99
	Fluopyram	125	–	–	0.30	–	–

^Note: Estimates are based on application at the maximum approved yearly dose within approved growth stages (BBCH intervals), plant interception according to the FOCUS guideline (Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014), mean transformation to TFA according to Table 1, an average temperature of 10°C, and 300 mm net precipitation.

Abbreviations: FOCUS, Forum for the Coordination of Pesticide Fate Models and Their Use; TFA, trifluoroacetic acid.

^a Dose is only 60 g/ha when applied in BBCH 00–12.

^b Single application; may be applied up to 10 times per year.

^c Early use against grasses but no specific BBCH intervals in the user instructions.

Estimates of annual average TFA concentrations in infiltrated net precipitation varied among crops and timing. Concentrations were generally higher for early growth stages, reflecting lower plant pesticide interception compared with later stages. Mefentrifluconazole and tau‐fluvalinate yielded low estimated TFA concentrations, primarily due to low conversion to TFA in the microcosm experiment (Figure 2). Diflufenican produced intermediate TFA estimates, whereas fluopyram, fluazinam, trifluralin, and fluazifop‐P‐butyl all yielded higher estimates across several scenarios. It should be noted that the fluazinam estimates are based on a single application; however, for the simulated product, potato treatments may be repeated up to 10 times per year. Repeated fluazinam applications would therefore result in substantially higher TFA estimates than those indicated in Table 3.

4. DISCUSSION

Our study examined the extent to which pesticides containing a C–CF₃ group are converted to TFA in agricultural topsoil and whether the resulting TFA is sufficient to alter concentrations in net precipitation infiltrating to the upper groundwater. TFA was formed from all seven C–CF₃ pesticides in all three soils; however, only a small and variable fraction of the added parent compound was converted to TFA (0.4%–10.7%, Table 1). In all cases, TFA formation was substantially lower than the up to 81% TFA formation reported for flufenacet (European Food Safety Authority [EFSA], 2024). Formation rates were also lower than the assumed average formation of 30% used in a theoretical assessment of pesticide‐derived TFA loads (Joerss et al., 2024), although this difference may be less pronounced when considering that TFA formation in the present study appears to continue at near‐constant rates beyond the first year.

Most TFA formation curves were approximately linear (Figure 2). Under laboratory incubation for up to 1 year, however, reaction rates would normally be expected to decrease over time. This is because soil is not mixed by tillage, no fresh organic matter is supplied by crops, and in the absence of precipitation‐driven water flow, diffusion gradients are not disrupted, all of which may constrain degradation processes. For mefentrifluconazole and fluopyram, the near‐linearity is likely explained by the large fraction of parent compound remaining after 1 year (80%–87% and 61%–75%, respectively; Table 2). In contrast, for pesticides with little remaining parent compound, TFA was presumably formed from a pool of C–CF₃‐containing intermediates that had accumulated in the soil. The most striking examples were fluazinam with 94%–98% removal (Table 2) but only 6.3%–6.5% conversion to TFA (Table 1), and fluazifop‐P/fluazifop‐P‐butyl with 99%–100% removal but only 3.1%–5.3% conversion to TFA. Fluazinam is stable at low pH but becomes increasingly unstable as pH increases, with an aqueous hydrolysis half‐life (DT₅₀) of only 8.7 days at pH 7 (European Food Safety Authority [EFSA], 2008); its near‐complete disappearance after 1 year was therefore expected. Fluazinam hydrolysis may have proceeded fastest in the Sv1 soil, which had the highest pH (6.7–7.3). Nevertheless, both TFA formation rates (Figure 2) and the remaining fluazinam after 1 year (Table 2) were similar across all three soils despite their different pH values. Fluazinam has at least five reported C–CF₃‐containing metabolites formed during degradation in soil (Table S4). Fluazifop‐P‐butyl is rapidly hydrolyzed in soil to fluazifop‐P, which itself is prone to aqueous hydrolysis with a DT₅₀ of 15 days at pH 7 (Pesticide Properties Database). Fluazifop‐P‐butyl/fluazifop‐P has at least three reported C–CF₃ degradation products in soil (Table S4). C–CF₃‐containing degradation products have also been reported for the other parent compounds, except for mefentrifluconazole (Table S4).

It is tempting to extrapolate the observed linear formation rates to estimate the time until complete conversion to TFA (e.g., 9–30 years for fluopyram and 100–250 years for tau‐fluvalinate). Such extrapolations are, however, unrealistic for several reasons: (1) Reaction rates tend to slow down over time as reactants are consumed, so TFA formation decreases instead of remaining linear. (2) Although no formation of free fluoride was detected during the first year, we cannot rule out that some fluoride may be formed instead of TFA. (3) Most importantly, parent compounds or their C–CF₃‐containing degradation products are subject to aging in soil (Alexander, 1995), whereby an increasing fraction becomes inert and will no longer contribute to TFA formation.

The measured TFA formation (Table 1) was used to estimate TFA leaching from sprayed fields during 1 year after spraying. Fluopyram, fluazinam, fluazifop‐P‐butyl, and trifluralin all gave high estimates of 0.1–0.5 µg/L for some scenarios. These high annual TFA concentrations primarily reflect the high persistence of TFA, in contrast to many other pesticide transformation products that may undergo further degradation before leaving arable topsoil.

The test‐tube microcosms were incubated at 20°C in conformity with the FOCUS guideline (Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014). This temperature is unrealistically high for temperate climates. When the temperature is reduced from 20°C to 10°C, the reaction rate of pesticide compounds is typically reduced by a factor of 2.58 (Q ₁₀; Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014). Since TFA formations in the laboratory experiment were approximately linear, we could simply divide the TFA formation percentages by 2.58 to approximate field‐relevant temperature. For fluazinam, which showed nonlinear curves, this adjustment likely resulted in a slight underestimation of TFA formation. The average Q ₁₀ value of 2.58 is itself uncertain and has previously been suggested to be lower (2.2; Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2006), which would increase the leaching estimates. Moreover, the temperature response may be smaller if parameters other than temperature limit degradation. Degradation of the hydrophobic, sorbing pesticides mefentrifluconazole, diflufenican, and tau‐fluvalinate might, for example, be governed by bioavailability constraints rather than by temperature. Sorption effects, however, were probably higher in the laboratory experiment than in conventional, arable soils because the SOM concentration of the organically managed test soils (3.7%–4.5%) was relatively high for arable soils. Fluazinam's dissociation constant (pK_a ) is 7.22, which indicates that polarity may have differed in the higher pH Sv1 compared with the other soils; however, the pH dependence of fluazinam hydrolysis discussed is likely to outweigh any pH‐related effects on sorption. The remaining parent compounds either do not dissociate (Pesticides Properties Database) or have pK_a far outside the range of normal soil pH, rendering pH effects on sorption negligible.

The leaching estimates are based only on the fraction of the applied field dose that is deposited directly on the soil surface during spraying, as defined by the FOCUS guideline (Forum for the Coordination of Pesticide Fate Models and Their Use [FOCUS], 2014), which in some cases may be as low as 10%. Estimated TFA leaching would increase considerably if larger proportions of the parent compounds or their transformation products entered the soil. The most obvious source of underestimation is thus wash‐off by rainfall after application, but additional routes are likely to contribute. For systemic pesticides such as fluopyram, the parent compound and/or its metabolites are expected to translocate to all parts of the plant, implying that decaying stumps and roots may constitute an additional soil input. The uptake of C–CF₃ herbicides by weeds may likewise end up in soil when treated weeds senesce. Potatoes are a special case, as haulm desiccation prior to harvest may transfer TFA, C–CF₃ pesticides, or their C–CF₃‐containing metabolites from above‐ground biomass to the soil. Collectively, these processes are likely to contribute to TFA formation and leaching, although their relative importance is unknown. Conversely, some TFA may disappear from the soil environment through crop uptake of TFA dissolved in soil water (Zhang et al., 2019), followed either by removal from the field at harvest or by partial return to the soil via decaying plant residues. These uncertainties are not specific to the present study but are inherent to most model simulations used in pesticide approval procedures.

Based on the almost linear TFA formation curves, it is evident that substantially more TFA would have formed if incubation had been extended beyond the first year, either from non‐degraded parent compounds or from C–CF₃‐containing intermediates accumulated in the soil. Such delayed TFA formation and leaching have previously been demonstrated in lysimeter experiments with haloxyfop, where TFA leached at comparable concentrations during the first and second year after application (European Food Safety Authority [EFSA], 2009). Continued TFA formation beyond the first year implies that repeated annual spraying could result in higher average TFA concentrations than those estimated here. Leached TFA concentrations obviously further increase where multiple C–CF₃ pesticides are used and/or where annual application rates are higher.

Measured TFA concentrations in groundwater infiltrated after 2010 ranged from 0.2 to 0.5 µg/L (mean: 0.36, n = 12) beneath Danish forests and from 0.2 to 1.6 µg/L (mean: 0.67, n = 41) under Danish arable land (Albers & Sültenfuss, 2024). A similar pattern has been reported for Switzerland, where TFA concentrations in monitoring wells in arable land (mean: 1.2 µg/L, up to 5 µg/L) were at least twice those observed in montane areas (<0.6 µg/L), an effect presumably caused by the use of agricultural pesticides (Federal Office for Environment [FOEN], 2024).

5. CONCLUSION

All of the tested C–CF₃ pesticides degraded to TFA in agricultural topsoil to some extent. Estimates of the annual TFA concentration in net precipitation infiltrating after application indicate that, for several cultivation scenarios, average annual concentrations are likely to exceed the European groundwater threshold of 0.1 µg/L for pesticide metabolites. Our study therefore supports that increased TFA concentrations found in groundwater under arable land can be caused by C–CF₃ pesticides, with pesticide‐derived TFA inputs that may be equal to the atmospheric TFA contributions. Our results could be substantiated by field experiments, which would reduce uncertainties associated with small‐scale lab experiments and model‐based leaching estimates, especially those related to crop interception, plant uptake of TFA from soil water, and temperature effects. Such field tests should preferably be conducted on fields without prior application of C–CF₃ pesticides, allowing atmospheric inputs to be directly compared with TFA derived from single spray applications without interference from older C–CF₃ pesticides or C–CF₃ metabolites accumulated in the soil. Finally, the observation that all seven tested C–CF₃ pesticides formed TFA suggests that other commonly used C–CF₃ pesticides should also be evaluated for their potential to generate TFA in soil.

AUTHOR CONTRIBUTIONS

Anders R. Johnsen: Conceptualization; methodology; writing—original draft; writing—review and editing. Trine Henriksen: Investigation; methodology; writing—review and editing. Christian N. Albers: Conceptualization; investigation; methodology; writing—review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Additional supplemental material contains data on annual total sales of C–CF₃ pesticides in Denmark, molecular structure of the selected C–CF₃ pesticides, temporal changes in farmland area in Denmark, degradation of C–CF₃ pesticides to TFA including data for negative control microcosms, mass‐spectrometric parameters for quantification of TFA parent compounds, and a list of known C–CF₃ pesticide degradation products from the selected C–CF₃ pesticides.

Johnsen, A. R. , Henriksen, T. , & Albers, C. N. (2026). Formation of trifluoroacetic acid from common trifluoromethyl pesticides in agricultural soils. Journal of Environmental Quality, 55, e70160. 10.1002/jeq2.70160