A hidden artefact: how surfactants can distort the results of springtail reproduction tests

Abstract

Soils are exposed to multiple substance groups, including surfactants, which directly enter soils when they are used as additives in firefighting liquids or pesticide mixtures due to their surface tension–lowering properties. The impact of chemicals on soil health is often tested with the springtail reproduction test. We tested the effects of the trisiloxane Break-Thru® S 301 on the reproduction of Folsomia candida in three soils according to Organisation for Economic Co-operaton and Development (OECD) guideline 232. Juveniles were extracted either by heat or flotation. In the latter method, recommended by OECD 232, test soil is flooded with water and stirred so that springtails float and swim on the water surface. Additionally, we tested the impact of Break-Thru S 301 on other life-history endpoints linked with reproduction, namely, reproduction investment and hatching success. We found a significant decline of recovered springtails at soil concentrations of Break-Thru S 301 down to 2 mg/kg in sandy soils when using flotation. However, using heat extraction, no effects were found at the same concentrations. Also, reproduction investment and hatching success did not indicate any toxicity of Break-Thru S 301 to springtails at all. In conclusion, Break-Thru S 301 reduced the water surface tension in the flotation process so that springtail juveniles just sank and disappeared from the water surface. This artefact potentially can occur for all surfactants tested this way. We propose testing surfactant impact on springtail flotation by adding a few drops of surfactant and observing springtail sinking behavior before testing toxicity. Alternatively, heat extraction or surfactant controls can be applied. Most importantly, these options should be mentioned in the respective guidelines, which are highly relevant for chemical risk assessment.

Introduction

Soil contamination is an increasing problem threatening soil function and biodiversity (Morgado et al., 2017), because there are a large number of different compound classes being released into soil. One substance group of increasing concern are surfactants, which are mainly used to reduce the surface tension of liquids for different purposes. Currently, firefighting foams are intensively discussed because they contain the highly persistent, (eco-)toxicologically very problematic groups of per- and polyfluoroalkyl substances (Brusseau et al., 2020), including perfluorooctane sulfonate (PFOS; Princz et al., 2018), but there already exist less persistent substitutes that are ecotoxicologically tested (Graetz et al., 2021; Kuperman et al., 2023). Other well-known representatives for surfactants originating from wastewater are nonylphenol ethoxylate (Domene et al., 2009) and linear alkylbenzene sulfonate (LAS, Jensen & Sverdrup, 2002; Krogh et al., 2007; see also Table 1).

The global market size of surfactants is estimated to grow from 47 billion USD in 2024 to 70 billion USD in 2032 (Fortune Business Insights, 2025). This reflects a global production volume of approximately 20 million tons (Mordor Intelligence, 2024). The biggest applications by far are home car and industrial and institutional cleaning, with together approximately 75% of the total share (Fortune Business Insights, 2025), but a market value of 1.85 billion USD is projected for agricultural surfactants in 2024 (Statista, 2023). In agriculture, surfactants are directly applied in the field. Worldwide agriculture is facing two challenges: on the one side, provision of more food for a growing world population with changing dietary habits requires a further rise in productivity—on the other side, arable land is declining due to climate change (Popp et al., 2013). One way to meet this is by increasing the efficiency of pesticides through the use of additives such as surfactants, which can change the physical-chemical properties of the spray-tank mixture (Janků et al., 2012). These surfactants were established in agricultural practice more than a century ago. They reduce the surface tension of the pesticide solution, which consequently affects either the application characteristics or the biological activity of the active ingredient, for example, by better adsorption (Green & Foy, 2004).

Currently, a new generation of surfactants have been developed and come into use, such as lipid-based nanoparticles (Gomes et al., 2024), nano-emulsions (Gomes et al., 2023), organic nano-vesicles (Gavina et al., 2016), and siloxanes (Adetunji et al., 2021; Wernecke et al., 2022). The latter were shown to be nontoxic to nontarget species of honey bees, but at the same time, to increase the toxicity of pesticides, probably due to an increased uptake into bee tissue (Wernecke et al., 2022). An increase of pesticide toxicity by a factor of 6 and 3,000 by a mineral oil–based surfactant, which was also toxic to some soil organisms, was shown for the reproduction of the whiteworm Enchytraeus crypticus and the springtail Proisotoma minuta (de Santo et al., 2019). This underlines the potential of surfactants to cause synergistic effects between different contaminants, resulting in an increasing toxicity towards organisms. Additionally, even if surfactants are considered to be mainly nontoxic, the development of new substance groups requires a regular update and screening for unforeseen effects (Gomes et al., 2023, 2024). This is crucial because their use as spray adjuvants in pesticide mixtures is not regulated in most of the European Union (EU) and the North American market (Wernecke et al., 2022).

In soil ecotoxicology, potential negative effects to soil organisms are assessed by a set of different assays and taxonomic groups. Within soil invertebrates, one of the most commonly tested taxonomic groups are springtails (Arthropoda, Collembola). This taxonomic group is highly relevant for soil health and nutrient cycling and therefore widely tested for the impact of chemical stress towards soil biota (Filser, 2002; Filser et al., 2014) This is also the reason why many different standardized guidelines for testing the effect of chemicals on springtails were implemented. One of them is guideline 232 from the Organisation for Economic Co-operation and Deveopment (OECD; 2016). This guideline is the most often used test system for investigating chemical stress to collembolans in research and also part of the EU directive for Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH, European Chemicals Agency, 2024). Other examples are the International Organition for Standardization (ISO) guideline 11267 (ISO, 2023) or EPS 1/RM/47 from the Canadian Government, namely, Environment and Climate Change Canada (ECCC, 2017).

The basic endpoint in these tests is reproduction, that is, the number of juveniles in the treated soils compared with a control soil, which is a more sensitive endpoint than survival and growth (Crouau & Moïa, 2006; Fischer et al., 2021; Xu et al., 2009). The central process to extract the juvenile collembolans is the flotation method, in which collembolans float to the water surface where they can be counted easily. This process is very time- and cost-efficient. Another method, heat extraction, was widely used about 20 years ago (e.g., Holmstrup et al., 2001; Jensen et al., 2007). For both methods, numerous studies exist on the effects of surfactants on springtails; however, recent studies always evaluate by flotation (Table 1).

The toxicity or efficacy of a substance can be highly dependent on soil properties (Chiapusio et al., 2007; Neves et al., 2019). In case of surfactants, it is expected that highly adsorbing soils with higher clay and organic contents tend to reduce their efficacy by adsorbing them (Domene et al., 2009; Princz et al., 2018). For surfactants, this effect is mainly dominated by the clay fraction (Lee et al., 2005). However, in a few cases, clay minerals can also increase the toxicity of certain substances (Gupta et al., 2016; Fischer et al., 2022). Therefore, it makes sense to test a set of soils with differing properties to account for potential interactions between the toxicant and the soil matrix.

In this study, we aimed to investigate the effect of a surfactant towards springtails, namely, the wetting agent Break-Thru® S 301 (BT), a trisiloxane by the manufacturer Alzchem (Trostberg, Germany), which is labelled as a polyether-polymethylsiloxane copolymer. To our knowledge, there exist no data on the toxicity of BT on soil organisms, but it has previously been tested on honey bees in a study by Wernecke et al. (2022), where no intrinsic toxicity of BT but an increase of pesticide toxicity by the addition of BT was found. However, the single substance tests with OECD guideline 232 in our study resulted in unexpected low numbers of recovered springtail juveniles in the soils treated with BT, because it was considered to be rather nontoxic, as shown by (Wernecke et al., 2022). Therefore, we aimed to further investigate whether this effect was due to the intrinsic toxicity of BT or rather to an artefact in the flotation process, which was caused by the reduction in water surface tension through BT and which was, to our knowledge, not yet mentioned in the literature. Because this artefact can be avoided by using heat extraction instead of flotation, we compared both methods with regard to their efficacy in extracting juveniles. To obtain robust results, we performed tests in three different soils, assuming that soil properties might additionally interfere with the toxicity or the artefact caused by the surfactant.

We show the difficulties in testing highly surface-active substances like trisiloxanes with flotation and give an overview on the current state of surfactants tested on collembolans by this method to underline the potentially wide distribution of the detected artefact (see Table 1).

Methods

Test substance, test soils, and spiking

The test substance, Break-Thru S 301 (CAS (Chemical abstract service): 134180-76-0) from Alzchem (Trostberg, Germany), was received from the distributer BayWa AG. It is defined as a pH neutral, 100% nonionic polyether trisiloxane with a density of 1.0–1.1 g/ml (Alzchem Group, 2024). In our own experiments, we assessed a density of 1.005 g/ml by weighing 10 ml BT three times on a fine balance (d = 0.1 mg, Sartorius, ENTRIS124I-1S, Göttingen, Germany). Acetone for soil spiking was received from Sigma-Aldrich (St. Louis, USA).

Soils from the Refesol system (Fraunhofer IME, Schmallenberg, Germany) were used in all tests (Table 2), namely, a sandy (S), a sandy-organic (SO), and a loamy soil (L). The soils were stored air-dried in the dark at 17 °C. The described experiments originate from two projects with different test setups; therefore, two different methods for spiking the soil with BT were used. For most full dose-response tests and the comparison between extraction methods, the soils were exposed to BT in screw-top jars using acetone as a carrier substance. Break-Thru S 301 was first dissolved in acetone and then pipetted onto a part of the dry soil to minimize the negative effects by contact with acetone on the microbial soil community in the dry soil. This method is widely used, and a ratio of 1 part of spiked to 3 parts of clean soil has proven suitable in many studies (e.g., Castro-Ferreira et al., 2012; Droge et al., 2006; Paumen et al., 2008). A ratio of 0.2 ml acetone per g dry soil was applied, which also corresponds approximately to the ratio in other studies using this methodology (e.g., Kobetičová et al., 2011; Paumen et al., 2008; Sverdrup et al., 2002; Smith et al., 2006). Immediately after spiking, the soil was stirred thoroughly and again 20 min later. The acetone was allowed to evaporate overnight under a fume hood. The following day, the remaining three quarters of the dry soil was added and stirred thoroughly. The soil was then adjusted to 50% of the maximum water holding capacity (WHC_max) with ultrapure Milli-Q water (18.2 Ω, Direct-Q 3 UV, MerckKGaA, Darmstadt, Germany), stirred again thoroughly, and stored in the dark at 17 °C for 7 days with the lid almost completely closed until the start of the test. At test start, the respective water loss was replenished to readjust to 50% WHC_max. In these tests, solvent controls were prepared for all soils in the same way as described above, and all results of the soils exposed to BT were related to the solvent control in the analysis.

For the life-history parameters, BT was dissolved in demineralized water, then it was mixed with the S soil in a ratio to reach 50% of WHC_max. Then the soil was stirred until it was homogenized. All nominal test concentrations are defined in online supplementary material Table S1. The soil concentration of BT could not be measured due to analytical restrictions.

Reproduction tests

For all tests, approximately 2–4-month-old Folsomia candida were reared at 15 °C in the dark on plaster of Paris/charcoal plates (8:1 weight ratio) and regularly fed with dry yeast. For starting synchronization, 30 adults were transferred to another plaster plate and kept for 3 days at room temperature in the dark to stimulate egg production. After 3 days, adults were removed and the juveniles usually hatched after 10 days.

The reproduction tests for collembolans were carried out according to OECD guideline 232 (OECD 2016) with two extraction methods: the flotation method and the heat extraction method. Besides the extraction procedure, all other steps were conducted in accordance with the guideline. In detail, 30 ± 0.3 g of moist soil was weighed into 100 ml snap lid glass jars (VWR, Radnor, USA), 10 juvenile collembolans of the species F. candida aged 10–12 days and a spatula tip of dry yeast were added, and the tightly closed jars were incubated at 20 °C and a 16:12-hr light:dark cycle in an incubator for 28 days. The vessels were aerated twice a week for at least 10 min and dry yeast was added. To verify the results, a full dose-response test was conducted at least two times for each test soil. In case of the S soil, five tests were run in total due to highly diverging results and its use in two different experiments.

For the flotation method, the soil was rinsed after 28 days with 100 ml tap water in plastic cups, a splash of ink was added (fountain pen ink 4001, Pelikan, Schindellegi, Switzerland), and the water surfaces were photographed. The number of juvenile and adult animals was determined manually from the digital photos using the ImageJ program (Schneider et al., 2017); sample identity was unknown to the counting person (blinding). According to the OECD guideline, a test was considered valid if an average of 80% of the adults and an average of at least 100 juveniles could be found in the control.

In the extraction method comparison test, flotation and heat extraction were used. Folsomia candida were exposed in the three test soils S, SO, and L at concentrations that were close to their previously assessed half-maximal effect concentrations (EC50s) of BT, that is, 30, 10, and 500 mg/kg, respectively (see also online supplementary material Table S1). For S soil, the value of 30 was calculated as the geometric mean of the EC50 values that were available at the time of the test start. Additionally, they were exposed to concentrations of these EC50 values divided by and multiplied with 5, respectively. Three replicates of each treatment were extracted by flotation and three by heat in a Macfadyen extractor.

In the heat extraction method, after 28 days at the end of the test, the soil was poured into a metal cylinder (6 cm diameter) that was sealed at the bottom with a coarse metal sieve (2 mm mesh size) and gauze bandage fabric so that the soil was retained but springtails could pass through. The metal cylinder was placed in a MacFadyen Extractor (ecoTech Umwelt-Meßgeräte GmbH, Bonn, Germany) a funnel system leading into a collecting vessel filled with 10 ml ethylene glycol. The extraction process started at 25 °C with an increase of 5 °C every 12 hr until the final temperature of 40 °C was reached (McKee et al., 2018). The funnel system was closed with a lid to preserve the soil moisture as much as possible during the extraction process. The heat gradient in the extractor forced the springtails into the collecting vessel where they were collected and preserved. To avoid the springtails floating on the surface and help preservation of the animals, 10 ml pure ethanol was added, so that the springtails sank down to the collecting vessel. The number of juveniles and adults collected was counted by hand under binoculars. To reduce the time effort, springtails were pipetted onto a glass Petri dish and the storage liquid was removed by a plastic pipette with a fine gauze wrapped around its tip. As a consequence, the springtails stuck immobile on the Petri dish ground and could be easily counted, supported by a white grid on black ground fixed under the Petri dish (∼20 min per sample).

Life history tests

In the life-history test, a reproduction test was also conducted as reference, where a reduced version of OECD 232 was used (Filser et al., 2014). The difference is that 5 animals are introduced instead of 10 and 10 g wet soil instead of 30 g wet soil is used. Otherwise, the procedure is the same. Test chamber conditions were 20 °C and constant darkness. After flotation, the adults were transferred to clean culturing plates: Petri dishes with plaster mixed with charcoal. Five days later, the eggs were spread with a wet brush and photographed with an AmScope 3 MP highspeed microscope digital camera on an Olympus SZH10 microscope. Ten days later, the eggs were photographed again to check hatching success based on the remaining eggs. ImageJ was used for digital measurements (Schneider et al., 2017). The measurements of reproduction parameters was based on Szabó et al. (2020). The eggs were counted on the digital pictures manually. Then, random numbers were generated to choose 10 eggs per Petri dish to measure size. The egg volume was calculated with the equation of the prolate spheroid, (V = 4/3 π × a × b2, where “a” is the longer diameter and “b” is the shorter diameter [Satterly, 1960]). The reproduction investment was calculated by multiplying the mean egg volume in a Petri dish with the egg number per adults. The hatching success was recorded by counting the remaining eggs 10 days after photographing them and calculated the ratio of remaining eggs per total number of eggs subtracted from 1 (1–remaining eggs/all eggs).

Statistics

All statistical analyses were conducted with the free software R (Ver. 4.3.3) and RStudio (Ver. 2023.12.1, R Core Team, 2025). Dose-response models were calculated using the DRC-package (Ritz et al., 2015). Significant differences between the respective treatments were assessed by a generalized linear model. The egg volume was analyzed by a linear mixed effect model with the Petri dish ID as random factor (aka random subject). The assumptions of the linear model were checked visually with a QQ-plot and a residual plot.

Results

All effects reported here refer to the respective solvent control having been spiked either by acetone or water (see also online supplementary material Table S1). The full dose-response 28-day reproduction tests were carried out in the three test soils S, SO, and L. These tests were designed based on previous range finding experiments and therefore in different concentration ranges (see online supplementary material Table S1). In the first test, all three soils were tested at the same time (Figure 1). In S and SO soil, an impact on recovered juveniles was found at much lower concentrations than in L soil. In both S and SO soil, reproduction dropped significantly within the concentration range of 0.75–12 mg BT/kg (dry soil—this applies to all subsequent soil concentrations). In S soil, it dropped from 76.7% to 40.1% and in SO soil, from 88.7% to 44.9% of the control. In contrast, in L soil, reproduction was reduced from 78.1% to 0% of the control in a concentration range from 100–3,690 mg BT/kg. Calculated EC50 values for reproduction in S, SO, and L soil were 3.8 ± 1.6, 10.4 ± 2.5, and 325 ± 113 mg/kg, respectively. For all other EC50 values, see online supplementary material Table S1.

In S and SO soil, tests were repeated. In S soil, the outcomes between the different tests were highly diverging: three tests with acetone as carrier ended up with EC50 values between 1.5–110 mg BT/kg. There was no pattern with regard to the respective solvent, that is, in the two tests with water, the toxicity was second and third highest of the five tests (Figure 2). In contrast, in SO soil, the relative number of recovered juveniles was similar in both tests, in the range from 0.75–12 mg BT/kg. In the second test, the relative number of juveniles compared with the control was constantly at approximately 40%, between 10–300 mg BT/kg.

In the life-history test, the recovered juvenile number dropped significantly from 3.7 mg BT/kg onwards (Figure 3A). However, there was barely any effect on the life-history parameters. The egg size was not affected (see online supplementary material Figure S1). The egg number significantly decreased in concentration 3.7 mg/kg (see online supplementary material Figure S1), but this effect was not significant if the whole reproduction investment was tested (Figure 3C). The hatching success increased in almost all treatments by 10%–20% compared with the control (Figure 3B).

In the comparison between flotation and heat extraction, the number of extracted juveniles was in the range of 800–1,100 juveniles in all controls, and there was no significant difference between the extraction methods. However, in all soils, there were clear differences in the number of juveniles between the extraction methods for the BT treatments. In S soil, the BT treatments extracted by heat ranged from 1,034 ± 33–1,106 ± 120 juveniles, whereas the BT treatments extracted by flotation were significantly lower than the control, with 539 ± 54 (6 mg/kg, p < 0.001) to 375 ± 9 juveniles (30 mg/kg, p < 0.001, Figure 4A). In SO soil, the average level of juveniles was slightly higher, but the overall pattern was the same. When extracted by heat, the BT treatments rendered between 972 ± 137–1,318 ± 226 juveniles, whereas the flotation-extracted BT treatments were all significantly lower than the control and between 731 ± 50 (2 mg/kg, p = 0.014) and 514 ± 25 juveniles (50 mg/kg, p < 0.001, Figure 4B). In L soil, the difference between heat extraction and flotation was much smaller than in the other soils, and significant differences to the control did not occur for both extraction methods until 500 mg BT/kg. Only at the highest concentration of 2,500 mg BT/kg was the number of juveniles significantly reduced in both extraction methods: by 43% from 958 ± 87–542 ± 26 juveniles in heat extraction (p < 0.001) and by 54% in the flotation experiment (819 ± 100–373 ± 90 juveniles, p = 0.017, Figure 4C).

Discussion

The results indicate that extraction of juvenile F. candida by flotation can cause a massive artefact when water surface tension is reduced by surfactants that are released from soil into the flotation liquid. Break-Thru S 301 and its predecessor Break-Thru S 240 can reduce the water-air surface tension from 72 to 22.4 and 22.6 mN * m⁻¹ at a concentration of 0.05% (Alzchem Group, 2024), which is more efficient than the widespread surfactant sodium dodecyl sulfate (MacLeod & Radke, 1993). The extent of this effect is comparable to other trisiloxanes (Silva et al., 2023). The reduction in surface tension by BT visibly led to an agglomeration and cluster formation of the animals, which also aggravated the counting process (Figure 5). The number of recovered juveniles was significantly reduced by flotation when BT was involved, probably due to sinking of juveniles, but not by heat extraction (Figure 4). The impact of BT on the flotation setup was also directly tested by adding 1–2 drops of BT (1 drop BT on 100 ml water equals approximately 0.05% [w/w] BT) to a vessel containing rinsed soil, which immediately led to a directly observable sinking of juveniles and, consequently, a reduced number of countable juveniles on the water surface (see online supplementary material Figure S2). This should be taken into account by all test guidelines for springtail reproduction tests recommending flotation as the extraction method, for example, OECD 232, ISO 11267, and EPS 1/RM/47 from ECCC (OECD, 2016; ECCC 2017; ISO, 2023) and alternative methods, for example, testing the impact of the surfactant on the flotation system in advance or a heat extraction for the verification of results gained by flotation when testing surfactants, should be provided.

The reproduction tests using heat extraction and the life-history test suggest that the trisiloxane BT is rather nontoxic to F. candida. From the heat extraction experiments, which were not designed as full dose-response tests, no-observable-effect-concentrations (NOECs) of 150, 50, and 500 can be derived for the test soils S, SO, and L, respectively (Figure 4, see online supplementary material Table S1). In the life-history test, the NOEC with the flotation method was 1.2 mg/kg for the juvenile number. However, in the life-history parameters, only one lower concentration produced fewer eggs (3.7 mg/kg) than the control, but this effect disappeared when the full reproduction investment was tested. This shows that it is better to use full reproduction investment instead of the egg number or the egg size. The reason for this is that F. candida has very flexible reproduction in both egg size and egg number (Tully & Ferrière, 2008); therefore, the reactions of the individual parameters are less reliable than the combined investment. An increase in hatching success was observed in F. candida in this study. In a 28-day reproduction test with the enchytraeid Enchytraeus crypticus and the same test soils S, SO, and L, EC50 values of 663 ± 22, 1157 ± 173, and 4,007 ± 373 mg BT/kg, respectively, were assessed (see online supplementary material Figure S3). This indicates, on the one side, that BT is also rather nontoxic to other soil invertebrates, and, on the other side, that the loamy soil tends to remarkably reduce the toxicity of BT.

The experiments showed high impact of the test soils on BT’s ability to lower surface tension. Clays can bind surfactants by hydrophilic interactions between the polar ethoxylate chains of the surfactant, which are also present in BT, and the hydroxyl groups on the clay mineral surface or hydrated cations on the siloxane basal plane (Krogh et al., 2003; Ishiguro & Koopal, 2016). Clays are generally more dominant in determining the adsorption of nonionic surfactants in soil than organic matter; there is a linear relation between mineral matter fraction and nonionic surfactant adsorption unless the organic matter content in soil is fairly high (Lee et al., 2005). This is in line with the larger number of recovered springtail juveniles in the BT treatments of the L soil, which might be due to reduced toxicity as well as reduced surface activity caused by the clay mineral adsorption. It can also explain the reduced toxicity towards E. crypticus, which was lower in SO soil than the low-adsorbing S soil but clearly lowest in L soil (see online supplementary material Figure S3). However, the variability of springtail recovery was highest in S soil, with EC50 values varying by a factor of up to 73, whereas in SO and L soil, they were nearly equal or varied by a factor of 2, respectively (Figure 2). As an explanation, in the low adsorbing S soil, other factors varying between tests may affect surfactant adsorption much more severely, for example, yeast residuals or mold formation. In contrast, lower EC50 values for Na-LAS were found in loamy and clayey soils than in a sandy soil (Holmstrup et al., 2001); however, in retrospect, these results were doubted by the authors themselves (Krogh et al., 2007).

The results underline that flotation bears the risk to overestimate toxicity by the described artefact, which in addition, can vary in its extent by differing soil properties. This problem is very likely also to occur for other surfactants, which are tested for springtails using flotation. Numerous studies exist on the effects of different types of surfactants and their impact on springtail reproduction (Table 1). In the early 2000s, there were many studies on the toxicity of the widespread surfactant LAS, which were all evaluated using heat extraction. They found comparably low toxicity of the genus Folsomia towards springtails, that is, EC50 values for F. fimetaria ranging from 424–803 mg LAS/kg (Holmstrup et al., 2001; Holmstrup & Krogh, 2001; Jensen & Sverdrup, 2002) , an EC10 value of 205 mg LAS/kg (Krogh et al., 2007), and an EC50 value of 1,437 mg LAS/kg for F. candida (Gejlsbjerg et al., 2001). Due to heat extraction, these results can be considered as not biased by the extraction procedure.

In contrast, most newer studies used the more time-efficient extraction by flotation. The tested surfactants vary widely in their recorded toxicity to the springtail F. candida. Some of the studies report low to very low EC50 values or a high sensitivity of F. candida to a surfactant compared with other organisms. Mixed micelles of LAS and ether sulfate-based surfactants with varying ethylene oxide units resulted in very low EC50 values of approximately 0.8 mg/kg in most cases, which was clearly below the predicted environmental concentration for LAS of 1.4 mg/kg in sludge-amended soil (Fernandes et al., 2020). Melatonin-loaded lipid surfactant submicron particles had an EC50 value of 44 mg/kg for the reproduction of F. candida but an even lower EC50 value of 15 mg/kg for the reproduction of E. crypticus (Gomes et al., 2024). In another study, the toxicity of a nano-emulsion material (lecithin, sunflower oil, borate buffer) and its dispersant (borate buffer) produced EC50 values for the reproduction of F. candida of < 50 and < 100 mg/kg, respectively (Gomes et al., 2023). In both cases, 50 and 100 mg/kg were the lowest test concentrations and caused a reproduction close to 0% of the control, that is, the actual EC50 values could be even much lower. For the reproduction of E. crypticus, EC50 values of 75 and 88 mg/kg were assessed. In a study on perfluorooctanoic acid (PFOA), a NOEC of 50 mg/kg and an EC50 value of < 100 mg/kg were found for the reproduction of F. candida; in comparison to plants, algae, and other soil invertebrates, the acute and chronic toxicity of PFOA to F. candida was one of the highest (Kwak et al., 2020). For the chemically similar substance PFOS, EC50 values for the reproduction of F. candida of 94 and 233 mg/kg in coarse and fine soil (higher clay content), respectively, were assessed, (Princz et al., 2018). These values were lower for the reproduction of the oribatid mite Oppia nitens, which was extracted by heat, with 23 and 94 mg/kg. This study supports the previously described observation of an increased adsorption of surfactants by clay minerals in soil.

Other firefighting foams or additives appeared to be much less toxic. Kuperman et al., (2023) investigated the effect of 8 firefighting foam products on the reproduction of F. candida, with EC50 values ranging from 493–5,451 mg/kg. For the reproduction of E. crypticus, which was tested as well, EC50 values of 465–10,512 mg/kg were in a similar range. Even if not directly comparable due to other metrics (percentage of solution instead of milligrams per kilogram) and missing data for recalculation, other firefighting additives can be considered to be rather nontoxic in the same manner (Graetz et al., 2021; Anderson & Prosser, 2023). This underlines the assumption that surfactants in general pose a rather low risk to soil organisms.

These high values are in line with the previously reported low toxicity of LAS to Folsomia sp., which were gained using heat extraction and therefore are not distorted by the surface-tension effects reported in this study. In a species sensitivity distribution on LAS, F. candida is ranked among the least sensitive species and F. fimetaria in the midrange of the distribution (Jensen et al., 2007). Therefore, if Folsomia sp. appear to be much more sensitive to a surfactant than other species, it might be a good indicator for verifying the springtail reproduction test outcome by another test using heat extraction. Another good indicator for an artefact may be other endpoints than reproduction, such as presented in our life-history experiment, when they do not indicate any toxicity at all. This does not mean that other surfactants could not be especially toxic to springtails, but observing these patterns may help to assess the necessity of a verification test. Furthermore, it should be noted that the described artefact may be especially relevant for newly developed surfactants of explicitly high efficiency in lowering water surface tension. It is possible that older tests on surfactants with less efficient surfactants using flotation are not affected by this artefact. Nevertheless, it underlines the necessity in adopting springtail reproduction tests for future risk assessment of potential surfactant formulations. This is all the more important when considering mixture toxicity, because surfactants might increase the bioavailability of the active ingredients.

The results of this study are highly relevant for regulatory purposes and soil ecotoxicology, because in both fields, the effects of surfactants to soil organisms are likely to be tested with the OECD guideline 232 or ISO 11267 and the recommended flotation method.

Conclusion

The observed and projected increasing production amounts of surfactants, also in the field of agriculture, require a robust and reliable risk assessment for this substance group in the soil environment to avoid harm against soil organisms. In this study, we unraveled an artefact that can occur when surfactants are tested on springtails in standard test procedures. Overall, the presented data and literature indicate that results of testing surfactant toxicity by the common springtail reproduction test using flotation can be distorted by reduced surface tension. This was shown by the comparison between flotation and heat extraction on the one side and by other endpoints relevant for the reproduction of springtails on the other. Furthermore, the extent of the distortion can greatly depend on the test soil properties, because clay minerals tend to strongly adsorb surfactants. We recommend testing the direct impact of the surfactant on the flotation system as a quick and easy first approach and a comparison of results using flotation with other endpoints and/or species. In case of doubt, verification using heat extraction should be conducted. To minimize the effort, one negative control and one test concentration, for example, at the EC50 derived from the flotation test, might be sufficient. We suggest this possible artefact to be mentioned in future versions of widespread test guidelines such as OECD 232, ISO 11267, or EPS 1/RM/47 from ECCC (OECD, 2016; ECCC 2017; ISO, 2023) and propose the control mechanisms and alternatives discussed herein.

Supplementary material

Supplementary material is available online at Environmental Toxicology and Chemistry.

Data availability

All raw data are available from the corresponding author and the open data repository platform PANGAEA.

Author contributions

Jonas Fischer (Conceptualization, Formal analysis, Investigation, Methodology, Visualization), Borbála Szabó (Conceptualization, Investigation, Methodology), Leonid Manikhin (Investigation), and Juliane Filser (Supervision)

Funding

Jo.Fi. was funded by the German Federal Environment Agency within the MIXTOX project (FKZ: 3720 73 201 0). B.S. was funded by the Humboldt Research Fellowship for Postdocs of the Alexander von Humboldt Foundation.

Conflicts of interest

None declared.

References