Quaternary alkylammonium disinfectant concentrations in soils rise exponentially after long-term wastewater irrigation

Quaternary alkylammonium compounds (QAACs) are used as disinfectants and surfactants worldwide, with their usage currently increasing as a result of the COVID-19 pandemic. QAACs are released into the environment with manure, sewage sludge and wastewater. The fate of QAACs in soils is poorly understood, although QAACs are inflicted in the selection of antibiotic-resistant bacteria. We studied the temporal accumulation of QAACs in soils of the Mezquital Valley that have been irrigated with Mexico City wastewater from 0 to 88 years. Concentrations of 16 QAACs, including alkyltrimethylammonium compounds (ATMACs), dialkyldimethylammonium compounds (DADMACs) and benzylalkyldimethylethylammonium compounds (BACs), were determined using HPLC-MS/MS after ultrasonic extraction. The most abundant QAAC-homologues in the soils were BACs > ATMACs > DADMACs. The concentrations of QAACs increased linearly and slowly during the first years of irrigation (∑QAAC: 2–23 µg kg−1), but after 40 years of wastewater irrigation we observed an exponential increase in QAAC concentrations (up to 155 µg kg−1). QAACs accumulate in soils of the Mezquital Valley during long-term wastewater irrigation. In contrast to pharmaceuticals, no apparent ‘steady state’ concentration is reached after decades of wastewater irrigation.


Introduction
Quaternary alkylammonium compounds (QAACs), such as alkyltrimethylammonium compounds (ATMACs), dialkyldimethylammonium compounds (DADMACs) and benzylalkyldimethylethylammonium compounds (BACs), are used worldwide as surfactants in household chemicals, the chemical and food industry and the agricultural sector (Mulder et al 2018). They consist of a permanently positively charged nitrogen atom, which is the hydrophilic part of the molecule, and four covalently bound organic rests, such as methyl groups, benzyl group or alkyl-chains of variable length that form the hydrophobic part. Thus, they are amphiphilic compounds. The critical micelle concentration best describes their partitioning in a soil-aqueous system (table 1). Ismail et al (2010) described that CMC represents the hydrophobic as well as the ionic properties of QAAC.
The large scale production and use of QAACs started in the first half of the 20th century. QAACs were first described in the literature in 1928 as 'saure seifen' (acidic soaps) by Hartmann and Kägi (1928). Domagk (1935) described the use of 'Zephirol' in 1935, which is BAC-C12, as a new class of disinfectants. In the following years BAC-C12 was widely used in the medical sector (Frankl 1941). Botwright (1946) advocated the use of these tasteless, for humans nontoxic compounds as cost-effective germicides for the food industry. By the end of the 1940, QAACs were used in the whole food industry (quick-freezing, canning, baking, butcher-shops), at food-markets, in eating and drinking establishments, medicine, veterinary medicine and more (However, as early as 1961, first signs of resistance development of microorganism were observed (Lee and Fialkow 1961)).
Nevertheless, even today QAACs belong to the class of so-called high production volume chemicals (HPVCs) (OECD 2004), because of their antimicrobial and surfactant properties. During the COVID-19 pandemic hygiene is one of the key measures for infection control and prevention. Consequently, the use of disinfectants has risen substantially, even though exact data on disinfectant consumption are difficult to obtain. Hora et al (2020) found that in the U.S the consumption of disinfection wipes of one manufacturer in May 2020 was 146% that of the year before. Almost half of all disinfectants (600 of 1285) listed by the German Association of Applied Hygiene (Association for Applied Hygene 2020) contain QAACs as active ingredients. It should be expected that the amount of QAACs released into the environment is increasing as a consequence of the COVID-19 pandemic. After their use, QAAC are introduced into the environment with wastewater, sewage sludge and manure (Mulder et al 2018). Data on the environmental concentrations and fate of QAACs are scarce and predominantly available for wastewater and sewage sludge (Merino et al 2003, Clara et al 2007, Ruan et al 2014, followed by data on river water and river sediments (Ferrer andFurlong 2002, Lara-Martín et al 2010). Information about their concentrations in soils is rarely available (Gerike et al 1994).
Studies have shown that microorganism can degrade QAACs under aerobic conditions with halflives in the range of hours (Games et al 1982, Zhang et al 2015 in aerated laboratory systems. Regarding their effects in the environment, the main concern is that their presence promotes the selection and spread of antibiotic-resistant bacteria and resistance genes (Gaze et al 2005, Tandukar et al 2013, Mulder et al 2018. The ecotoxicologial effects of QAACs in soils and sediments are strongly reduced by their sorption. Previous work has shown, that QAACS can sorb in interlayer spaces of smectite soil clay minerals (Zanini et al 2013), which reduces their toxicological effects (and hence their bio-accessibility) at least under controlled laboratory conditions (Heyde et al 2020b). By reducing the bio-accessible concentration of molecules, sorption is also known to slow down the dissipation and degradation of organic compounds, thus increasing their persistence in strongly boundforms (Rosendahl et al 2012). Sorption occurs via cation exchange (due to the positively charged N atom) and via hydrophobic interactions of the alkyl chains with organic matter (Ismail et al 2010, Ren et al 2011, Mulder et al 2018. In semiarid and arid areas of the world, municipal wastewater is reused for irrigation of agricultural land (Asano 1998). The reuse of wastewater for irrigation is regarded as key for relieving regional water shortages, increasing food production and improving income and livelihoods of people (e.g. Bahri et al 2009, Drechsel et al 2015. Wastewater irrigation also has positive effects on the soil concentration of nutrients, the content of soil organic matter and the biomass as well as activity of soil microorganisms (Siebe 1998, Friedel et al 2000, Dalkmann et al 2012. Disadvantages of wastewater irrigation are the input of pollutants and the spread of resistance genes due to the input of pathogenic bacteria (Igbinosa et al 2011, Munir and Xagoraraki 2011, Dalkmann et al 2012. One of the largest agriculturally used areas under wastewater irrigation in the world (Siebe and Cifuentes 1995), the Mezquital Valley, covers an area of 900 km 2 and is located 80 km north of Mexico City (figure 1). Wastewater irrigation in this area started in 1912 and the irrigated area has been growing over the past 100 years due to the expansion of Mexico City and is still expanding (Broszat et al 2014). Until 2017 most of the irrigated 80 000 ha land received untreated wastewater for irrigation; since 2018 approximately 60% of the wastewater is treated prior to irrigation (pers. Comm. Ing. Cruz Atienza, director of the wastewater treatment plant).The wastewater is pumped and diverted frequently until it reaches the fields (Dalkmann et al 2012). While in the pipe system, the water is anaerobic until it reaches the fields where it is aerated. Main characteristics and metal concentrations of the water are provided in table SI1 (available online at stacks.iop.org/ERL/16/064002/mmedia). The concentrations of trace metals in the wastewater canals are high but below the Mexican wastewater guidelines. The concentrations of Al, As, Pb and methylmercury in the Tula River, the river into which the wastewater is partially discharged, exceeded the surface water guidelines (Guédron et al 2014). Few trace metal concentrations (0.024 mg l −1 Cd, 0.510 mg l −1 Pb) in the irrigation canal were exceeding the surface water guideline concentrations (Ramirez-Fuentes et al 2002).
Together with the wastewater, all kinds of chemicals that are used in households are introduced into the fields and soils. Siebe Grabach (1994) and Siebe and Cifuentes (1995) documented the accumulation of heavy metals in soils of the Mezquital Valley with increasing time under irrigation by sampling a chronosequence of soils for which irrigation started in different years in the course of the expansion of the irrigated area. The concentrations of heavy metals in the soils increased linearly with increasing time of irrigation, because these metals are continuously introduced and sorbed to clay, iron oxides and organic matter. Later studies showed that the soils are also enriched with pharmaceuticals and endocrine disruptors over time of irrigation (Durán-Alvarez et al 2009, Gibson et al 2010). Dalkmann et al (2012) revealed that concentrations of pharmaceutical residues reach an apparent steady state over time, which can be theoretically explained as an equilibrium reached between constant input and first order degradation (dissipation) of the organic compounds. QAAC concentrations in wastewater are known to be quite high, often an order of magnitude higher than commonly observed pharmaceuticals. For example, treatment plant influents in the range of 2-6000 µg l −1 were reported for Austria (Clara et al 2007, Martínez-Carballo et al 2007. Effluent concentrations between 14 and 78 µg l −1 are reported for Northern Europe by Kaj et al (2014).
Therefore, as was observed in the Mezquital Valley soils for pharmaceuticals, a similar pattern enrichment and attainment of apparent steady-state could also be expected for QAACs, which are also biodegradable under aerobic conditions (Tandukar et al 2013). We therefore hypothesized that (a) soil concentrations of QAACs increase with the duration of untreated wastewater irrigation and (b) the soil concentrations of QAAC will reach a steady state with time, analogously to the documented enrichment of pharmaceuticals in the Mezquital Valley. For testing these hypothesis we sampled an irrigation chronosequence of soils in 2018 and quantified concentrations of BACs, ATMACs and DADMACS. Results were confirmed and compared by extracting samples that were already analyzed for concentrations of pharmaceuticals by Dalkmann et al (2012). Finally, the accumulation of QAACs was compared to the accumulation of heavy metals with increasing time of irrigation.

Study area description
The Mezquital Valley is mostly located in the northeastern part of the Transmexican volcanic belt and extends into the Sierra Madre Oriental sedimentary system in its northern part. The valley bottom consists of extended piedmonts of relatively compacted and partly CaCO 3 -cemented volcanoclastic sediments of Tertiary age which grade into fluvial and lacustrine Quaternary deposits in its lower parts. Climate is temperate (mean annual temperatures between 16 • C and 18 • C) semi-arid in the South with annual mean precipitations ranging from up to 700 mm in the southern parts to less than 435 mm in the North concentrated in the summer months. Dominant soils are rendzic Leptosols, haplic, calcaric and vertic Phaeozems and pellic Vertisols. Irrigated areas are mostly cultivated with lucerne-maize rotations (3 years of lucerne followed by 2 years of maize combined with rye grass or different kinds of vegetables). Non-irrigated areas are nowadays scarce and used to cultivate dominantly rain-fed maize or agave species (Siebe Grabach 1994).

Soil samples
The first set of samples was taken in 2011 for the Dalkmann et al (2012) study from 0 to 30 cm depth from nine different fields in the Mezquital Valley representing eight periods of irrigation with untreated wastewater (0, 1, 3, 6, 8, 13.5, 35, 85 years). These samples were taken as composite samples with a soil auger (∼1 kg), collected in polyethylene bags and stored at −21 • C. The second set of samples was collected in 2018 as composite sample from 0 to 20 cm depth from ten fields representing nine irrigation durations (0, 2, 8, 10, 18, 20, 34, 70, 88 years). Sampling in 2018 was performed with a shovel (∼300-500 g) and the soil samples were wrapped in alumina foil and stored at −21 • C in polyethylene bags.
For all analyzes, except for pH measurements, the samples of 2018 and 2011 were freeze-dried (Beta 1 8 LSCplus, CHRIST, Osterode am Harz, Germany) in the lab for 5 d, sieved < 2 mm (Retsch®, Haan, Germany) and stored in high density polyethylene (HDPE) containers.

Characterization of soil samples
Soil pH-values were determined with a pH electrode (SenTix® 940-3 WTW, Weinheim, Germany) and a pH meter (handylab 2, SCHOTT Insturments, Mainz, Germany) in 0.01 M CaCl 2 . Five grams of air-dried (40 • C until constant weight) soil was mixed with 12.5 ml CaCl 2 solution and shaken at 80 rpm in a horizontal shaker (KS 10, Edmund Buehler GmbH, Bodelshausen, Germany) for 30 min. After shaking, suspensions were allowed to settle for 30 min and pH was determined in the supernatant.
Carbonates were determined volumetrically as CO 2 released from CaCO 3 with a Scheibler apparatus (Blume et al 1996, DIN EN ISO 1069. Total carbon and total nitrogen content of the samples were determined from 5 mg soil using a CN analyzer with a thermal conductivity detector (Unicube®, Elementar, Langenselbold, Germany).
The organic carbon (C org ) content was calculated by subtracting the content of carbonate-C from the total organic carbon content (Nelson and Sommers 2001).
For soil texture analysis all particles ⩾20 µm were separated with a wet sieving process and quantified gravimetrically, and particles smaller than 20 µm were separated by sedimentation according to Köhn (1928) and quantified also gravimetrically. Prior to soil texture analysis, carbonates were destroyed with HCl (10%) and soil organic matter was destroyed with H 2 O 2 (30%). Tetrasodium pyrophosphate (0.4 mol l −1 ) was added as dispersion agent to prevent flocculation of soil particles (Jensen et al 2017).
Soil trace element contents were determined by inductively coupled plasma coupled to an optical emission spectrometer (Agilent 720ES) after microwave assisted aqua regia extraction of the pseudo-total metal content based on U.S. EPA. Method 3051A (2007) modified by Öztan and Düring (2012).

QAAC analyzes
The soils were analyzed for ATMACs (chain length C-8 to C-16), BACs (C-8 to C-18) and DADMACs (C-8 to C-16). Firstly, the soils were repeatedly extracted with acidified acetonitrile (99.9% AcN/0.1% HCl v/v) and ultrasonic irradiation according to Heyde et al (2020a). The extracts were cleaned with solid phase extraction (Chromabond CN, Machery Nagel, Dueren, Germany). The analytes were separated and quantified with high performance liquid chromatography (Waters™ alliance 2690, Eschborn, Germany) and a Waters XSelect CSH Phenyl-Hexyl-Column (130 Å, 150 mm length, 2.1 mm ID, 3.5 µm particle size) with a column guard of the same material coupled to a mass spectrometer (Micromass Quattro Micro, Waters, Eschborn, Germany) following the method described elsewhere by Heyde et al (2020a).
A list of QAACs analyzed along with available compound parameters is provide in table 1.

Data evaluation
Peak integration of the QAAC analyzes was performed with Waters MassLynx 4.0, data evaluation and illustration with Systat Sigma Plot 12.0. RStudio 4.0.2 was used to perform correlation analysis i.e. the correlation matrix (tables SI2/SI3) and the correlograms (figures 2 and SI1). In order to analyze the linear correlation between irrigation time, soil properties and QAAC concentration, the Pearson correlation coefficient (r) was used. We tested the correlation of soil properties and QAAC irrigation time additionally with Spearman rank correlation to confirm the results (tables SI4, SI5).
Systat Sigma Plot 12.0 was also used to calculate regressions between (log) QAAC-concentrations (y) and duration of irrigation (x). The same was performed for the concentrations of the trace metals, pH, CaCO 3 , C org , Clay, and N total (y) and the duration of irrigation (x). We judged the increase of concentrations (values) with increasing duration of irrigation as significant, if the confidence intervals of the slope parameter of the regression did not overlap with zero. In this case the H 0 -hypotheses was: slope parameter b is 0, which means the slope did not differ significantly from 0. The level of significance was set to α = 0.05. Hence, the increase of concentrations with duration of irrigation was significant if p < 0.05.
Multiple Regression was performed to test which variable (pH, irrigation time, clay content) explains most of the observed increase in the sum of QAAC concentrations (p < 0.05).

Accumulation of QAACs in wastewater-irrigated soils
The smallest total concentrations of QAACs were found in the samples of 2011 after 0 and 3 years of wastewater irrigation. Across both sample sets the most abundant QAACs belonged to the group of BACs (2018; 2.3-131.2 mg kg −1 ), followed by ATMACs (2018; 0.1-12.7 mg kg −1 ) and DADMACs (2018; 0-11.3 mg kg −1 ) (table 2). The prevalence of BACs was expected since they are widely used in households and four of six BAC homologues are HPVCs, whereas only two of the five analyzed ATMACs and DADMACs are HPVCs (Mulder et al 2018). Our results were in the same concentration range as data of Li et al (2018), who also found DADMACs in lowest concetrations among the three group of QAACs in sewage-impacted estuarine sediments.
In every single sample BAC-C12 was detected, which is one of the most commonly used QAACs in disinfectant agents (Wagner 2017). We also found BAC-C14 in all samples, with the exception of one sample of the 2011 sampling year. Both homologues were found at highest concentrations among all QAACs. The greatest concentration of each sum of the three homologue groups (ATMAC, BAC, DAD-MAC) was in the soils that were irrigated for 88 years. BAC-C12 was found in highest concentrations among all homologues in soil that was irrigated 88 years (83.6 µg kg −1 ). The greatest increase of QAACs from 0 to 88 years of irrigation was observed for BAC-C12 (+4100%), BAC-C14 (+5500%) and ATMAC-C16 (+3500%). Concentrations of ATMAC-C8 and C10 (+0%) remained at a similarly low level.
In general, the samples taken in 2018 contained higher QAAC concentrations compared to those taken in 2011. This could be caused by aging effects of the soils and a loss due to microbial activity during storage, since QAACs are biodegradable by microorganisms (Nishihara et al 2000), especially by Pseudomonas spec. which can use QAACs as carbon source (Dean-Raymond and Alexander 1977, Geftic et al 1979, Kaech and Egli 2001, Takenaka et al 2007, Liffourrena et al 2008. Samples taken in 2011 were stored for 7 years longer before analysis than samples from 2018. Since we cannot rule out the suspected aging effects, we focus the discussion mainly on the samples of 2018. The duration of irrigation was strongly and positively correlated to ∑ ATMACs (r = 0.89), ∑ BACs (r = 0.89) and ∑ QAACs (r = 0.88) (figure 2 and tables SI2, SI3), while for ∑ DADMACs the correlation was weaker, with r = 0.60. This was supported by the data of the 2011 samples ( figure  SI1 and tables SI2, SI3). One possible reason is, as mentioned before, the lower production volumes of DADMACs (OECD 2004) resulting in lower discharges through the wastewater into soils. Further, DADMACs, especially those with alkyl chains longer Plotting the logarithmic concentrations of ∑ ATMAC and ∑ BAC versus the irrigation time resulted in highly significant linear increases (figures 3(a) and (c)), which illustrates an exponential increase of these concentrations. This pattern was also observed in soils sampled in 2011 for all three classes of QAACs, including DADMACs (figures SI 3(a), (c), (e)).  When narrowed down to the first 40 years of irrigation, a closer look at the temporal trend of ∑ ATMAC and ∑ BAC concentrations during the first 40 years of irrigation revealed a linear increase with the duration of irrigation (figures 3(b), (d) and SI 3(b), (d)). This suggested a biphasic accumulation in soil with time: an initial phase with linear accumulations, as typically observed for metals (Siebe Grabach 1994, Siebe andCifuentes 1995), followed by an exponential rise of concentrations for time periods >40 years. This observation is in contrast to the behavior of pharmaceuticals in those soils, for which an apparent 'steady state' is reached over time (Dalkmann et al 2012). We suppose the biphasic accumulation of BACs and ATMACs in soil during long-term irrigation with wastewater was caused either by (a) massive QAAC-inputs during the first half of the 20th century that ceased towards the end of the 20th century, or by (b) an increasing strength of QAAC sorption to the soils, slowing down their biodegradation in soils with a very long history of wastewater irrigation. Indeed, QAACs were intensively used following their invention as disinfectants in 1935 (Domagk 1935) as described in the introduction. Speculatively, the upcoming concerns about resistance could then have led to a drop of consumption, with subsequently lower QAAC concentrations in the wastewater and eventually in the soils, which were first irrigated during the 2nd half of the 20th century or later. However, since BACs and ATMACs are biodegradable, their accumulation requires an efficient sequestration of these compound caused by an increasingly strong sorption.
Interestingly, Zanini et al (2013) described a shift in sorption processes with increasing concentration of BACs. They found that the sorption of BAC occurred at low concentrations to soil minerals and at higher concentration to humic acids. At low concentrations QAACs act as polar as molecules and behave like other (metal) cations, but at high concentrations they form micelles and act like a nonpolar compound, which changes their behavior in solutions (Mulder et al 2020). Also the sorption to soil minerals occurred in two steps (Zanini et al 2013). At low concentrations, BACs were sorbed on the surface and in the interlayer of clay minerals and with rising concentrations the sorption occurred in bilayers due to hydrophobic-hydrophobic interaction at the clay surface. However, it is unclear whether sorption to organic matter and in bi-layers would stabilize BACs against biodegradation. Sorption in the interlayer space of clay minerals is known to protect organic molecules against biodegradation (Amelung et al 2018). As we and others have shown in previous work, QAACs widen the interlayer spaces of clay minerals (Paiva et al 2008, Heyde et al 2020b. In those interlayer spaces they can form bilayers and paraffinlike structures (Zhu et al 2003), which are poorly accessibly for enzymes and microorganisms. Possibly, such widening of interlayer spaces and the efficient diffusion of QAACs into these spaces requires the build-up of certain minimum QAAC concentrations and time, which are only given after long-term irrigation.
Since there is no ecotoxicological information available regarding no-effects concentrations of QAACs in soils, we cannot judge potential risks associated with the accumulation of the compounds in soils.

Change of soil properties with increasing duration of irrigation
The pH-values decreased the longer soils were irrigated (up to 85 years and 88 years, table 3). The duration of irrigation and pH-value were correlated negatively in the samples of 2018 (figure 2), which was confirmed by the data of samples collected in 2011 ( figure SI1). The range of pH was only from 7.1 to 7.9 in both sample sets, but the negative correlation corresponds to findings of Xu et al (2010), who showed that the pH-value of wastewater irrigated soils decreased in the upper 20 cm over time of irrigation. Siebe et al (2016) also observed this decrease in the range of neutral to slightly alkaline values due to wastewater irrigation in the Mezquital Valley. Hernandez-Martínez et al (2018) proposed that proton production by nitrification is accountable for the slightly changing pH values over time.
Apart from the soil that was irrigated for 34 years, all soils that were sampled in 2018 had a carbonate content below 1%. Among the samples from 2011, only one soil irrigated during 1 year had a carbonate content of 1.3%. Neither for the samples collected in 2011 nor for the ones collected in 2018, could we observe a significant correlation between the irrigation time and the CaCO 3 content in the soils (figures 2 and SI1). These results are similar to those of Adrover et al (2012), who found no significant change in soil carbonate content after wastewater irrigation for 30 years.
C total and N total were strongly correlated with each other with correlation coefficients of 0.93 (2018) and 0.97 (2011), but only weakly correlated with irrigation time in both sample sets. N total ranged from 0.1% to 0.2% in 2018 and 2011 and C total ranged from 0.9% to 3.0% in 2018 and up to 2.4% in 2011 (table 3 and figure SI2). Considering that C org ranged from 0.3% to 2.1% in 2011 samples, the results are in good agreement with the results from Dalkmann et al (2014), who obtained the same concentrations of N total and slightly lower C org concentrations. The highest amounts of organic carbon with 3% and 2.4% were detected for a soil that was irrigated more than 34 years (2018) and one irrigated for 85 years (2011). An increasing C org content with irrigation time was expected, since it is known that wastewater irrigation can drive the build-up of soil organic matter content (Friedel et al 2000). Sánchez-González et al (2017) observed that organic carbon stocks increase 1.5 fold after 40 years of irrigation and reach thereafter a new steady state in the Mezquital Valley.
The determined clay contents ranged from 36% to 57% in samples form 2018 and from 33% to 55% in 2011. An exception was the field that was irrigated for 0 years (table 3), where the soil had a clay content of 71%. Dalkmann et al (2014) described a clay content between 39% and 54% for soils in that area, which is close to our findings. Apparently, the clay content was independent of the irrigation duration, since the amounts of fine material introduced into the soils with wastewater were too small to increase the bulk mass of clay-sized material significantly. This is also in agreement with a study by Mamedov et al (2001), who found no significant difference in clay content between wastewater and freshwater-irrigated soils.

Accumulation of trace metals with increasing duration of irrigation
The regression between soil concentrations and irrigation duration showed that concentrations of Cd, Cr, Cu, Ni, P, Pb and Zn increased significantly with time in both sets of samples (table 4). The concentrations of Cu, Mn, Zn, Cd, Cr, Co and Pb (table 4) were similar to those that were found by Cajuste et al (1991) in wastewater-irrigated soils in Mezquital Valley, but our Ni-concentrations were about four times higher (17.6-35.4 mg kg −1 ). Pb contents in the soils that were irrigated for 70 or more years (⩾35.1 mg kg −1 ) were elevated. This coincided with the fact that the Pb-concentration in wastewater were increased. The lowest Cr-content that we found was 42.6 mg kg −1 (table 4) and background concentration of rainfed soils were around 50 mg kg −1 . These comparably large background concentrations of Cr were most likely inherited from the volcanic rocks in the Mezquital Valley area, which form the parent material for soil formation.
Strong positive correlations between duration of irrigation and the contents of Cd, Cr, Cu, Ni, P, Pb, S and Zn occurred for the samples of 2018 (figure 2). These correlations were confirmed by the data of the samples from 2011. This leads us to conclude, that metals were accumulated in the soils due to the irrigation with wastewater. Similar accumulation of Cu, Cr, Pb, Zn and Cd in wastewater irrigated soils of the Mezquital Valley was reported previously by Siebe and Cifuentes (1995). Hence, the temporal change of metal contents with increasing irrigation time underlines the suitability of our set of chronosequence samples for testing the potential accumulation of QAACs in wastewaterirrigated soils.
In summary, wastewater irrigation over a time period of 88 years leads to an accumulation of QAACs in soils. Contradictory to our second hypotheses, QAACs behaved differently to what we expected from pharmaceuticals, as their concentrations did not reach a steady state after 88 years of irrigation.

Conclusion
This study provides first evidence that wastewater irrigation leads to an increase of QAACs concentrations in the soil of the Mezquital Valley, Mexico. In fact, our study is to our knowledge is the first to report QAAC concentrations determined for an extended set of (unspiked) soil samples. We show that the concentrations of QAACs exceed the concentrations of other antibiotics and pharmaceuticals often by an order of magnitude, highlighting the need to keeping this group of substances on the radar when considering mechanisms of antibiotic resistance gene evolution. The temporal trend of QAAC accumulation is opposite to the trend observed for pharmaceuticals. No 'steady state' concentration caused by a balance between inputs and dissipation is reached after 88 years of irrigation. Instead, QAAC concentrations in long-term irrigated soils rise exponentially, potentially due to massive inputs of QAACs into the soils in the first half of the 20th century or due to a strong sorption and effective sequestration (e.g. in clay mineral interlayer spaces).
The intensive use of disinfectants during the COVID-19 pandemic will likely leave a footprint of QAACs in wastewater-irrigated soils.
The fate of QAACs in soils is different to the fate of metal cations and pharmaceuticals. Future work should address the processes and kinetics of QAAC sorption and sequestration in soils with a focus on sorption and sequestration in clay mineral interlayers.

Data Availability
All data that support the findings of this study are included within the article (and any supplementary files