Harmonizing nanomaterial exposure methodologies in ecotoxicology: the effects of two innovative nanoclays in the freshwater microalgae Raphidocelis subcapitata

Abstract Layered double hydroxides (LDHs) are innovative nanomaterials (NMs) with a typical nanoclay structure (height <40 nm) consisting of layers of metallic cations and hydroxides stabilized by anions and water molecules. Upon specific triggers, anions can exchange by others in the surrounding environment. Due to this stimuli-responsive behavior, LDHs are used as carriers of active ingredients in the industrial or pharmaceutical sectors. Available technical guidelines to evaluate the ecotoxicity of conventional substances do not account for the specificities of NMs, leading to inaccuracies and uncertainty. The present study aimed to assess two different exposure methodologies (serial dilutions of the stock dispersion vs. direct addition of NM powder to each concentration) on the ecotoxicological profile of different powder grain sizes of Zn–Al LDH-NO3 and Cu–Al LDH-NO3 (bulk, <25, 25–63, 63–125, 125–250, and >250 µm) in the growth of the freshwater microalgae Raphidocelis subcapitata. Results revealed that the serial dilutions methodology was preferable for Zn–Al LDH-NO3, whereas for Cu–Al LDH-NO3 both methodologies were suitable. Thus, the serial dilutions methodology was selected to assess the ecotoxicity of different grain sizes for both LDHs. All Zn–Al LDH-NO3 grain sizes yielded similar toxicity, while Cu–Al LDH-NO3 powders with smaller grain sizes caused a higher effect on microalgae growth; thus, grain size separation might be advantageous for future applications of Cu–Al LDH-NO3s. Considering the differences between exposure methodologies for the Zn–Al LDH-NO3, further research involving other NMs and species must be carried out to achieve harmonization and validation for inter-laboratory comparison.


Introduction
Nanomaterials (NMs) are designed to exhibit unique, desirable physicochemical characteristics for a broad set of applications.Therefore, some attention has to be devoted to the approaches and methodologies employed in their risk characterization phase, as current standardized test methodologies or guidelines are designed for conventional chemicals.These approaches do not fully account for the unique and inherent physicochemical characteristics of NMs (Hund-Rinke et al. 2016;Khan et al. 2017;Skjolding et al. 2016).A significant number of reviews highlight the necessity of developing nano-specific test methodologies (e.g.Handy et al. 2008Handy et al. , 2012;;Hartmann et al. 2017;Hund-Rinke et al. 2016) and the Organisation for Economic Co-operation and Development (OECD) is following this path as one of the institutions with the critical role in the development and harmonization of test guidelines (TG) (Petersen et al. 2015;Rasmussen et al. 2019).In 2020, the OECD published a guideline document (GD number 317) addressing nano-specific testing issues and recommendations on aquatic and sediment ecotoxicological bioassays (OECD 2020).Some of these include: (i) a mandatory NM physicochemical characterization in relevant test media and as-produced material, (ii) appropriate exposure and dispersion methodologies to ensure homogeneous NM dispersion, and (iii) alterations to already published and widely used OECD TG for environmental hazard assessment (e.g.OECD TG 201 and OECD TG 211), to account for possible outcome scenarios obtained from the NM exposure behavior (OECD 2020).The GD 317 poses a promising starting point for harmonizing test methodologies.However, like other comprehensive GDs, it may not fit all NMs due to the constant development of innovative materials with a broad range of properties and may require updates or adaptations.The OECD TG 201 -Freshwater Alga and Cyanobacteria, Growth Inhibition Test (OECD 2011) is one of the most commonly used TGs in hazard assessment studies of chemical substances, as microalgae act as primary producers and carbon dioxide fixators, playing a central role in the dynamics of freshwater ecosystems and geochemical cycles.Being considered one of the most challenging protocols for NMs testing, it is one of the targeted TGs for test harmonization in the NanoHarmony project (https://nanoharmony.eu/ ), along with the OECD TGs 202 (Daphnia sp.acute immobilization test, OECD 2004) and OECD 203 (fish acute toxicity test, OECD 2019).Most challenges in conducting ecotoxicological bioassays are related to the instability and poorly understood dynamics of some NMs in test media, presenting low reproducibility in results, thus making their hazard evaluation very challenging (Johnston et al. 2020;Skjolding et al. 2016).Besides, some artifacts may arise while testing NMs, like impurities, dissolution, reactive oxygen species (ROS) formation during dispersion, interaction with test reagents, biological matrix or measuring procedure, inducing some constraints while measuring endpoints, such as biomass in the case of microalgae, or identifying their internalization in organisms' tissues (K€ uhnel and Nickel 2014;Petersen et al. 2015).These events must be considered upon microalgae testing with different NMs to derive accurate, relevant, reproducible, and reliable experimental data.Also, the test methodologies must tackle such specific NM-related test issues during microalgae bioassays that are reported and well described in the literature to enable inter-laboratory experimental reproduction and data comparison.Ultimately, this will boost confidence in NM testing and lead international organizations, like the OECD, to update microalgae testing methods to account for such specificities (Hartmann et al. 2013(Hartmann et al. , 2017;;Hund-Rinke et al. 2016).
Considering the above, this study aimed to fill the gap in harmonizing exposure methodologies to derive accurate toxicological profiles of NMs with characteristics that make them challenging in testing.For that, two commercially available engineered nanoclays, known as layered double hydroxides (LDHs), stabilized with a NO 3 -anion, Zn-Al LDH-NO 3 , and Cu-Al LDH-NO 3 were tested in the green microalgae Raphidocelis subcapitata.LDHs are nanoclays with 20-40 nm in height that can be found individualized or slightly aggregated when dispersed in aqueous media (Martins et al. 2017(Martins et al. , 2022;;Santana et al. 2022).However, manufacturers use a spray dryer to obtain a fine powder that amalgamates the LDHs and generates micro-sized aggregates/grains (Gomes et al. 2020).Although LDHs tend to recover the individual state when dispersed in aqueous media easily, it is unclear if the different grain sizes may interfere with the stability and ecotoxicity of both selected LDHs.Thus, we took advantage of the availability of different powder grain sizes of LDHs (bulk, <25, 25-63, 63-125, 125-250, and >250 mm) and tested them in R. subcapitata, using the best methodology provided previously.These NMs exhibit peculiar characteristics, such as good biocompatibility, high chemical stability, and pH-dependent solubility (Mishra et al. 2018).The functional and compositional flexibility (e.g.numerous metal-anion combinations) and the anionic exchangeability pose as the two primary advantages of LDHs (Kuthati et al. 2015;Mishra et al. 2018).Due to their characteristics, these nanoclays are very promising NMs for numerous applications, namely antifouling or anticorrosion coatings, wastewater treatment, agrochemical, pharmaceutical or cosmetic products, flame-retardants, textiles, among others (As ¸c¸ı 2017; Avelelas et al. 2017;Barik et al. 2017;Eisentraeger et al. 2003;Lead et al. 2018;Lead and Wilkinson 2006;R. Liu et al. 2011;Maurer-Jones et al. 2013;Sayre et al. 2017;Singh et al. 2019;Wilhelm et al. 2021).Additionally, it has been reported that this class of NMs exhibits low marine ecotoxicity (Avelelas et al. 2017;Ding et al. 2018;Gutner-Hoch et al. 2018, 2019;Koba-Ucun et al. 2021;Martins et al. 2017Martins et al. , 2022;;Santana et al. 2022) making them suitable for eco-friendly applications.Nevertheless, the ecotoxicity of the versatile and widely produced Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 in freshwater photosynthetic species is still unknown.As far as authors are concerned, there is only a single study on the freshwater ecotoxicity of such LDHs, explicitly focusing on the early life stages of the fish Danio rerio upon exposure to Zn-Al LDH-NO 3 (Carneiro et al. 2023).
The information generated in this work will be of utmost importance for the compliance of the NMs exposure methodologies for microalgae testing.Additionally, the outputs of the present study will be critical for a future environmental hazard assessment of this NM in the freshwater compartment, as well as for the NMs' manufacturers and industry stakeholders aiming at an eco-friendlier NMs design.

Test nanomaterials
Two different LDHs were tested, (i) zinc-aluminum LDH (Zn-Al LDH-NO 3 ) and (ii) copper-aluminum LDH (Cu-Al LDH-NO 3 ).Both LDHs were synthesized as described by Gomes et al. (2020) by the Smallmatek, Lda.(Aveiro, Portugal).Briefly, LDHs were manufactured by co-precipitation of the respective hydroxide metallic salts in a solution oversaturated with sodium nitrate (NaNO 3 ), being the pH value kept constant (pH ¼ 10 ± 0.5) with the addition of sodium hydroxide (NaOH) (Gomes et al. 2020).A slurry was obtained and washed in deionized water and then dried using an industrial spray dryer to ensure the formation of fine powders of each bulk material (Gomes et al. 2020).During this drying step, LDHs are amalgamated; thus, the bulk material corresponds to heterogeneous powder with an extensive range of micro-sized aggregates/grains that are expected to return to the original individualized state after dispersion in water.Thus, these powders were separated into five different grain size groups (<25, 25-63, 63-125, 125-250, and >250 mm) using a vibratory sieve shaker (Retsch, Haan, Germany).
Full details regarding the morphological, chemical, and textural characterization of Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 are available in Martins et al. (2017) and Santana et al. (2022), respectively.These studies report a usual hexagonal morphology with less than 40 nm in height for both LDHs, with an average size distribution between 300 nm (width) and 600 nm (length) for Zn-Al LDH-NO 3 (Avelelas et al. 2017) and 1.31 mm (width) and 1.58 lm (length) for Cu-Al LDH-NO 3 (Santana et al. 2022).

Microalgae growth inhibition testing
2.2.1.Freshwater microalgae Raphidocelis subcapitata cultures Cultures of R. subcapitata were maintained at 20 C ± 1 C under continuous PAR light (photosynthetic active radiation).Inoculates were prepared 3-5 days before the bioassays to obtain cells in the exponential growth phase, being the cultures renewed after seven days by inoculation in a new culture medium (OECD 2011).Woods Hole MBL culture medium (pH ¼ 7.2) was prepared according to Stein et al. (1973) and used for algae culturing and bioassays.All the materials were previously sterilized in an autoclave at 121 C and 1 bar for 20 min.

Growth inhibition tests
Algae growth inhibition tests with R. subcapitata were performed following the OECD TG 201 adapted by Eisentraeger et al. (2003), with different exposure conditions based on two different exposure methodologies.Exponential growth cells were collected from the algae culture, and their absorbance was measured in a microplate reader at 440 nm, being the concentration calculated using the following equation: where C represents the algae concentration in cells/mL and ABS is the absorbance recorded at 440 nm.The test was assembled in 24-well microplates for 72 hours.Each treatment's replicate was obtained by incubating 100 mL of a R. subcapitata cell suspension (5 Â 10 5 cells/mL) with 900 mL of LDH test dispersion in a well.Blanks for the negative control and the tested concentrations (culture medium and dispersed LDHs with no algae) were also prepared.This provided information on potential artifacts in the spectrophotometry measurements created by the NMs' presence.Additionally, falcon tubes (50 mL) with the same suspensions were added to the experimental setup to allow the measurement of pH and conductivity parameters (using the respective probes) and to assess the suspensions' behavior (e.g.sedimentation) over time by dynamic light scattering (DLS).The microplates were incubated under recommended TG conditions (temperature: 23 C ± 1 C; continuous PAR light; constant agitation) for 72 hours.Every 24 hours, absorbance measurements were recorded in a microplate reader at 440 nm, and the concentration of algae was calculated using Eq.(1).The average specific growth rate for each concentration was calculated as a logarithmic increase in the algal biomass using the following equation: where l iÀj represents the average specific growth rate from time i to time j (day À1 ); Ti is the initial exposure time (days, i.e. day 0 marks the initial exposure); Tj is the end time of exposure (in days, day 1, 2, or 3 represent the 24, 48, and 96 h exposure time); Ci and Cj the cellular concentration (cells/mL) at time i and time j, respectively.Moreover, the inhibition percentage in average specific growth rate was also accessed by using the following equation: where %Ir is the inhibition percentage in average specific growth rate; l C is the mean value for the average specific growth rate in the negative control (day À1 ), and l T is the mean value for each treatment's average specific growth rate (day À1 ).

Exposure methodologies to obtain NMs concentrations range
The first objective of this study was to determine the best methodology to test LDHs as challenging test NMs.Two exposure methodologies were tested based on the recommendations from the OECD GD 317.One was the serial dilutions from the stock dispersion, previously prepared in Woods Hole MBL medium, to build a range of concentrations (serial dilutions methodology).In the other methodology, each concentration was made individually by directly adding the required NM powder weight to MBL media to achieve a specific concentration (direct addition methodology) (OECD 2020).For each NM type and grain size, all suspensions were achieved by ultrasonic bath sonication (Selecta; 40 kHz) for one hour to guarantee maximum particle dispersion in Woods Hole MBL medium.
In order to evaluate data reproducibility, two independent growth algae inhibition tests were conducted for each exposure methodology, with two replicates per treatment and four for the negative control, following the protocol described in Section 2.2.2.
Tests were only performed with the bulk powder for the Zn-Al LDH-NO 3 and <25 mm grain size for the Cu-Al LDH-NO 3 , as according to the information provided by the manufacturer, the behavior of these sizes of Zn-Al and Cu-Al LDH-NO 3 s in ultrapure water is similar and can be used as a starting point for testing the media contamination procedure.Tested concentrations were 10, 100, and 200 mg Zn-Al LDH-NO 3 /L and 1.75, 3.5, and 7 mg Cu-Al LDH-NO 3 /L (determined in previous rangefinding tests).

Grain size-dependent toxicity tests
The second objective of this study was to unravel the possible grain size toxicity effect of the LDH powders used by applying the best exposure methodology (from Section 2.3).The serial dilutions methodology was selected based on the results gained from the exposure methodology tests.A growth inhibition test was performed as previously described (see Section 2.2.2) for each grain size (i.e.<25, 25-63, 63-125, 125-250, and >250 mm) and bulk LDH powder for both LDHs.The test concentrations used were: 0, 60, 90, 120, 150, 180, and 210 mg/L and 0, 1, 2, 3, 5, 7, and 9 mg/L for Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 , respectively, with four replicates per treatment and eight for the negative control.Also, blanks for the negative and the tested concentrations (culture medium and dispersed LDHs with no algae) were prepared for artifact prevention.The average specific growth rate in each treatment, for each grain size from each LDH, was calculated every 24 hours.

Stability of LDH NMs in MBL media
The stability study of both LDHs was used to define the best methodology for ecotoxicity data reproducibility and understand the grain size-dependent effects.

Physicochemical characterization of LDHs
DLS and zeta-potential (fP) measurements were performed in the Woods Hole MBL culture medium used for the microalgae growth inhibition tests using a Zetasizer Nano-ZS (Malvern Analytical, Malvern, UK).One concentration (10 mg/L), for each LDH and grain size fraction, was prepared and submitted to sonication in an ultrasonic bath (Selecta; 40 kHz) for one hour (see Section 2.1), prior to the measurements.All measurements were performed in triplicate, at constant room temperature, at fivetime points (0, 6, 24, 48, and 72 h).The pH and conductivity were measured simultaneously using pH and conductivity probes (WTW, Weilheim, Germany).

Nitrate release from LDHs
Nitrates can be released from interlamellar galleries of LHDs to the exposure media and be responsible partly for the effects.Therefore, nitrates release was measured in MBL media using the method 8039 for HACH DR/2000 field spectrophotometer (Hach Company, Colorado, CA), where nitrate species quantification is based on cadmium reduction.Dispersions for both LDHs and respective grain sizes were prepared in MBL as previously described (i.e.ultrasonication bath for one hour), and the test concentrations were achieved by serial diluting from a stock solution.The measurements were performed at 0 (immediately after ultrasonication) and 72 hours (end of the algae growth inhibition test), with only one replicate.Between measurements, the prepared dispersions were maintained under the same exposure conditions used for the algae growth inhibition tests to mimic the test scenario.Three concentrations were selected for both LDH types and grain sizes: 60, 120, and 210 mg/L for Zn-Al LDH-NO 3 , and 1, 5, and 9 mg/L for Cu-Al LDH-NO 3 , which correspond to the lowest, intermediate, and highest concentrations tested for the algae growth inhibition tests with each type of LDH and grain size.
2.5.3.Dissolution of Zn, Cu, and Al from LDH nanomaterials Concentrations of the particulate and dissolved forms of Zn, Al, and Cu were measured by inductively coupled plasma mass spectrometry (ICP-MS) for all grain sizes of both LDHs.For this, the lowest, intermediate, and highest tested concentrations (Zn-Al LDH-NO 3 : 60, 120, and 210 mg/L; Cu-Al LDH-NO 3 : 1, 5, and 9 mg/L) used in the algae growth inhibition tests were selected, and measurements were performed at time 0 h (immediately after ultrasonication) and 72 h (end of the growth inhibition test) (n ¼ 1).A stock dispersion of each grain size, from both LDHs, was prepared in Woods Hole MBL medium as previously described (see Section 2.1), being the remaining test concentrations achieved by serial dilution from the mentioned stock dispersion.These samples were then immediately centrifuged at 15 000 Â g, for five minutes, to enable the separation between the particulate fraction (present in the pellet) and the dissolved fraction (present in the supernatant).Both fraction samples were acidified to a final concentration of 2% HNO 3 (v/v).Between sampling times (0 and 72 h), test dispersions were maintained under the same exposure conditions as the growth inhibition tests (but without algae).Finally, samples were readily stored at 4 C until ICP-MS analysis.All analyses were conducted at the Central Laboratory of Analysis of the University of Aveiro, Aveiro, Portugal (certified laboratory).The detection limit for the targeted chemical elements was 0.05 mg/L.

Statistical analysis
The following null hypotheses were addressed to fulfill the aims of this study: H 0 1: the specific growth rate of R. subcapitata upon Zn-Al or Cu-Al LDHs exposure is similar regardless of the adopted OECD guidance recommended for the exposure methodology.
H 0 2: the specific growth rate of R. subcapitata upon LDHs exposure is similar regardless of the LDH chemical nature (Zn-Al vs. Cu-Al) and grain size (bulk form vs. each grain size fraction).
To test both hypotheses, initially, ecotoxicity data were analyzed for normality and homogeneity of variances using the Shapiro-Wilk and Levene's tests, respectively.When data followed a normal distribution and homoscedasticity, a one-way analysis of variance (ANOVA) was performed to assess differences between treatments (p 0.05), followed by a post hoc multiple comparisons test using Dunnett's method to compare each treatment with the control, whenever differences were attained.When data failed normality distribution and homoscedastic, a Kruskal-Wallis ANOVA on ranks was performed, followed by a post hoc multiple comparisons Dunn's method when significant differences were obtained (p 0.05).Moreover, median effective concentrations (EC 50 ) values for all bioassays' average specific growth rates were calculated using a non-linear regression three-parameter (3P) logistic model.Statistical treatment was performed on SigmaPlot version 12.0 (Systat Software Inc., San Jose, CA).Additionally, the slopes obtained from the non-linear regression and the EC 50 values from the exposure methodology-related assays and the grain size-dependent toxicity assays were compared by a generalized likelihood ratio test (Chi-square (v 2 ) test).

Physicochemical characterization
Table 1S and 2S show the results of the tested NMs' behavior in Woods Hole MBL medium over time, including pH, conductivity, fP, and hydrodynamic size (as average peak size) data for each grain size class and time point (0, 6, 24, 48, and 72 h) for Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 dispersions, respectively.Conductivity and pH showed low variation (less than 1%) for the different sizes over 72 h for Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 dispersions (Table 1S and 2S).Tables 1 and 2 report the dispersion's stability over time, based on fP and hydrodynamic size data variation between the first and last time points (i.e.0 and 72 h) for Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 , respectively.
Zeta-potential results showed a general pattern of increase in time from 0 to 72 h for all grain sizes for both LDHs.However, they are still classified as slightly unstable colloidal dispersion according to Kumar and Dixit (2017) (Table S1 and 2S).
The hydrodynamic size results of the Zn-Al LDH-NO 3 s were also very variable for all grain sizes over time, as no clear grain size-dependent or timedependent change was observed (Table 1).For this engineered nanoclay, the grain size fractions that produced stable dispersions were the bulk powder, 125-250 and >250 mm, showing variations of 7.85%, 16.65%, and À11.70%, respectively.The hydrodynamic size of Cu-Al LDH-NO 3 showed a general tendency to increase across all different grain sizes over time; however, no grain-size related change was observed (Table 2).The 25-63 mm dispersion of Cu-Al LDH-NO 3 is deemed stable For both parameters, the % variation between timepoints (0-72 h) was calculated.For both parameters, the % variation between timepoints (0-72 h) was calculated.
(-0.05% of variation), as well as the dispersions of 125-250 mm (-16.53%) and >250 mm (-15.27%)grain size fractions (Table 2).However, and differently from the Zn-Al LDH-NO 3 , the results show that all grain size fractions produced stable dispersions, depending on the evaluated parameter.
Nonetheless, looking at the variation percentages in time of both parameters (fP or hydrodynamic size), all dispersions for both types of LDH and regardless of grain size, are considered stable considering the variation percentage of the least variable parameter.For example, the 25-63 mm grain size fraction of the Zn-Al LDH-NO 3 is considered unstable considering the hydrodynamic size variation over time; however, the same dispersion can be considered stable if the fP variation is set as a parameter with minor variability parameter (42.79% and À8.84%, respectively; Table 1).However, this scenario does not apply to the <25 mm grain size dispersion of the Zn-Al LDH-NO 3 as the variation on both parameters is higher than 20% (-37.33% and 72.34% of variation for the fP and hydrodynamic size, respectively; Table 1).

Nitrate release measurement
Nitrate concentrations were measured to assess the anionic exchange release from all Zn-Al and Cu-Al LDH-NO 3 grain size fractions and are presented in Tables 3 and 4, respectively.
For the Zn-Al LDH-NO 3 , the nitrate release was LDH concentration-dependent, and only minor variations were observed from 0 h to 72 h (Table 3).For the Cu-Al LDH-NO 3 , the concentration dependency was not so marked, as only a very low variation in nitrate concentrations was verified from 1 mg/L to 9 mg/L, and, in some cases, at time 0 h nitrate concentrations were similar to the negative control.However, as previously mentioned, the dispersions' stability evaluation based on nitrate release was not performed.The nitrate release occurred rapidly in Woods Hole MBL medium (see Tables 3 and 4), and a possible anionic-exchange plateau may be rapidly reached; hence, no variation between this study's time points would be observed.5 and 6, respectively.The dissolved fraction may still contain small-sized nanoparticles because no filter was used, and the centrifugation does not guarantee the removal of all particles.This dissolved fraction may include ions or metallic complexes formed while dissolved in the water media.
Regarding the Zn-Al LDH-NO 3 , both Zn and Al concentrations (particulate and dissolved) showed  Measurements were performed at 0 and 72 hours (n ¼ 1).
an overall concentration-dependent increase, being more pronounced for the particulate fraction.An opposite tendency was observed between particulate and dissolved forms for both Zn and Al concentrations when considering time.An overall decrease for the particulate form was observed for both elements, while an increase was observed for the dissolved form, with minor exceptions (e.g.Zn concentration for the particulate form increases between time points for the two highest test concentrations in the >250 mm grain size).Moreover, the concentration values for elements and respective forms were unrelated to grain size.Cu-Al LDH-NO 3 showed similar behavior with increased particulate and dissolved forms in a concentration-dependent manner.Identically to Zn-Al LDH-NO 3 , the particulate form of Cu and Al concentrations decreased over time, whereas the concentration of both Cu and Al dissolved forms increased.Moreover, a grain size-dependent decrease was observed for the Cu dissolved form.The Cu's dissolved form concentration was higher in the <25 mm grain size and decreased with the increasing grain size, with the lowest value obtained for the larger grain size.The values obtained for both LDH types and their respective grain size fractions showed an overall good stability status for all dispersions in MBL (variations of ±20%), with few minor exceptions.

Exposure methodology test
In the exposure methodology approach, algae growth inhibition tests assessed differences between exposure methodologies for serial dilutions from the stock dispersion (in Woods Hole MBL) and the direct addition of NM powder to achieve each required concentration.Tests were done twice in order to evaluate their reproducibility.Regarding the Zn-Al LDH-NO 3 , both exposure methodologies exhibit similar toxicological profiles, with a concentration-dependent decrease in the specific growth rate of R. subcapitata.For the first independent test (Figure 1(A)), significant differences (p < 0.05) were found in the two highest concentrations, despite the exposure methodology, but not for the 10 mg/L concentration.Moreover, a significant difference between exposure methodologies was found at the 100 mg/L test concentration.In the second independent test (Figure 1(B)), significant differences (p < 0.05) were also found for all test concentrations, regardless of the exposure methodology, except for the 10 mg/L treatment.
Additionally, when comparing methodologies for the same test concentration, significant differences were observed for the highest test concentrations (i.e.200 mg/L).The calculated EC 50 values for each exposure methodology with Zn-Al LDH-NO 3 are summarized in Table 7.For the direct addition of NM powder to each test concentration methodology, EC 50 values are statistically different between trials (v 2 ¼ 4.60, p < 0.05), whereas for the serial dilution methodology, no significant difference (v 2 ¼ 2.95, p > 0.05) was observed (Table 7).
The Cu-Al LDH-NO 3 exposure to R. subcapitata derived a dose-response for the specific growth of the microalgae in both first and second independent assays (Figure 2(A,B), respectively).Regardless of the exposure methodology, significant differences were noted for all concentrations in both independent assays.When comparing exposure methodologies within the same test concentration, only statistical differences were found for the 1.75 and 7 mg/L concentrations in the second independent test (Figure 2(B)).The calculated EC 50 values are summarized in Table 8, which showed no significant differences between exposure methodologies (v 2 ¼ 0.46, p > 0.05 for the direct addition methodology, and v 2 ¼ 0, p > 0.05 for the serial dilutions methodology).

Grain size-depended toxicity tests
The results for the average specific growth of R. subcapitata exposed to Zn-Al LDH-NO 3 are shown in Figure Figure 1.Average specific growth rates for Raphidocelis subcapitata when exposed to Zn-Al LDH-NO 3 (bulk form) using two different exposure methodologies: (A) 1st independent assay and (B) 2nd independent assay.Ã Evidence significant differences (p < 0.05) in the average growth rate when compared with the respective control for the exposure methodology; bars represent the mean ± standard error.The letter 'a' represents statistical differences when comparing data for the same concentration between methodologies (p < 0.05).9.A dose-response curve was derived for each grain size fraction, and significant differences (p < 0.05) were noted for all tested concentrations compared to the respective controls (Figure 3).The EC 50 values for the average specific growth rate derived for each grain size were not different from each other (v 2

and Table
(1) < 3.841, p > 0.05) (Table 8).The average specific growth rates of R. subcapitata exposed to Cu-Al LDH-NO 3 are shown in Figure 4 and Table 10.A dose-response curve for this nanoclay was achieved for the bulk powder, <25, 25-63, and 63-125 mm grain size fractions (Figure 4(A-D)).Moreover, calculated EC 50 values for the average specific growth and growth inhibition showed a noticeable grain-size dependent increase, with significant differences being found between the bulk powder and both <25 mm (v 2 ¼ 21.79, p < 0.05) and 63-125 mm (v 2 ¼ 44.77, p < 0.05) grain size fractions (Table 4S).Considering that a doseresponse curve could not be derived for the 125-250 and >200 mm grain size fractions, it is also concluded that they differ from the bulk material.
Table 7. EC 50 values with respective standard deviation (SD) expressed as (i) total mass of Zn-Al LDH (mg Zn-Al LDH/L) and as (ii) Zn mass (mg Zn/L), calculated for the 1st and 2nd independent algae growth inhibition tests with Raphidocelis subcapitata, for both exposure methodologies, using Zn-Al LDH-NO 3 (bulk form).(v 2 (1) > 3.841, p < 0.05) between EC 50 values of each independent test for the same methodology.
Figure 2. Average specific growth rates for Raphidocelis subcapitata when exposed to Cu-Al LDH-NO 3 (<25 mm grain size) using two different exposure methodologies: (A) from the 1st independent assay and (B) from the 2nd independent assay.Ã Evidence significant differences (p < 0.05) in the average growth rate when compared with the respective control for the exposure methodology; bars represent the mean ± standard error.The letter 'a' represents statistical differences when comparing data for the same concentration between methodologies (p < 0.05).
Table 8.EC 50 values with respective standard deviation (SD) expressed as (i) total mass of Cu-Al LDH (mg Cu-Al LDH/L) and as (ii) Cu mass (mg Cu/L), calculated for the 1st and 2nd independent algae growth inhibition tests with Raphidocelis subcapitata, for both exposure methodologies, using Cu-Al LDH-NO 3 (<25 mm grain size).

Zn-Al LDH and Cu-Al LDH behavior in MBL media
The OECD GD 317 (OECD 2020) refers that a NM dispersion is considered stable if 'the mass concentration does not derivate more than 20% from the initial value due to sedimentation within a relevant time scale'.However, in this study, the mass concentration variation over time was not assessed, as LDH NMs cannot be quantified in terms of mass, as their metallic constituents are known with some uncertainty and NO 3 and water are difficult to predict.Nonetheless, the same OECD document states that 'other manufactured NM exposure metrics (e.g. the extent of agglomeration, dissolution, or change in other manufactured NM characteristics, fP) in addition to total concentration may also be evaluated relative to the ±20% variability target, where practicable'.Accordingly, the fP and hydrodynamic size, nitrate release, Zn, Cu, and Al dissolution (and as particles) were quantified.Cu-Al LDH-NO 3 and Zn-Al LDH-NO 3 dispersions behaved in time accordingly to the described in the literature for other media.Gomes et al. (2020) reported high stability of Zn-Al LDH-NO 3 dispersed in distilled water in a wide pH range (pH 3-12), with only $20% of the LDH being dissolved during one  Values in parentheses represent the 95% confidence intervals.
month in neutral (or close to neutral) pHs.Such partial dissolution of both LDHs leads to an increase in the concentration of Zn, Al, and Cu elements (Tables 1 and 2), as well as nitrate anions, in the media, as they are released from the LDH structural matrix upon dissolution (Galvão et al. 2016).This input of different metallic and nonmetallic elements may lead to the formation of different chemical complexes in the test media (Iftekhar et al. 2018), promoting a shift from the initial media's pH value (Woods Whole MBL, pH ¼ 7.2) to the ones measured immediately after the dispersion process (Table 1S and 2S).
In the present study, DLS and fP values measured in all grain size fractions of both LDHs were variable (Table 1S and 2S).This might be explained based on the LDHs' trademark dual-layered geometry.DLS is considered the most used tool for the size and fP characterization of NMs.However, it has already been shown to be unsuitable for nonspherical geometries (Levin et al. 2016(Levin et al. , 2017)), as many nano-specific behaviors (e.g.rotational diffusion) are not taken into account during the measurement (Khlebtsov and Khlebtsov 2011;T. Liu and Xiao 2012).LDHs powders possess an intrinsic ability to readily disperse and individualize themselves in ultra-pure water (Choy et al. 2000;Yan and Chen 2013); therefore, a similar hydrodynamic size should be expected for all grain sizes of each LDH type in equivalent time points, disregarding the possible DLS unsuitability to characterize non-spherical NMs.However, this was not the case in the present study; even though, the differences are softened when considering the average hydrodynamic size obtained from all time points measurements in >1.17 >1.17 >250 mm 5 7 >9 >9 >1.17 >1.17 Values in parentheses represent the 95% confidence intervals.
MBL, for all grain size fractions of both LDHs.Still, the differences found can be explained by the natural variance that may derive from: (i) the LDHs' dynamics in high ionic strength test media, as the continuous anionic exchange might affect the NMs dynamic and behavior or (ii) from the LDHs drying process, which depending on its duration might affect the LDHs behavior dynamics (Galvão et al. 2016;Xu et al. 2006).Despite the previously mentioned variations, the fP measurements for Zn-Al and Cu-Al LDH-NO 3 indicate an overall tendency to increase over time; however, a grain size-dependent influence is unclear.The values obtained for all Zn-Al and Cu-Al grain sizes are negative, ranging between À5 and À30 mV, regardless of the time point.This finding indicates that both LDHs have incipient colloidal stability, which leads to some extent to aggregation and sedimentation events (Kumar and Dixit 2017;Riddick 1966).Such behavior was observed in our study, which was expected due to the high ionic strength of the test media, leading LDHs to aggregate, as reported by Pavlovic et al. (2017).
The release of nitrate anions from the interlamellar region was also assessed for all grain sizes of both LDHs (Tables 3 and 4, for Zn-Al and Cu-Al LDH-NO 3 , respectively).An overall expected concentration-dependent increase in the nitrate concentration was observed for both LDHs, with the nitrate anion content available for anionic exchange increasing with the increasing LDH concentrations.Such fast anionic exchange of nitrates observed in high ionic strength media was also reported by Kotlar et al. (2020).However, the lowest Cu-Al LDH-NO 3 concentration (1 mg/L) yielded nitrate concentrations equivalent to the ones observed in control (Woods Hole MBL) for both time points of all grain sizes.This suggests that such low LDH concentrations lead to a lower release of nitrates to the test media, being the final nitrate concentration similar to the initial values.
Moreover, a subtle increase in the nitrate concentration between time points for all grain size fractions was also noted.This minor increase between time points is consequently related to the prompt and fast release of the intercalated nitrate anion, which occurs within the first minutes after addition to an aqueous ionic solution, as reported in a study conducted by Tedim et al. (2012) using a Zn-Al LDH-NO 3 intercalated with nitrate anions.A similar fast release of intercalated nitrate anions has also been reported in soils (Kotlar et al. 2020).
The study of Zn-Al LDH-NO 3 behavior showed high concentrations of particulate and dissolved Zn compared to Al concentrations, which can be justified by synthesizing this LDH using a 3 Zn 2þ :1 Al 3þ ratio.

Exposure methodology test
The OECD GD 317 reports that a serial dilution from a stock dispersion methodology is considered more appropriate in cases where good stability of the NM is a test media is observed (OECD 2020).Moreover, these findings suggest that further studies with Zn-Al LDH-NO 3 using the serial dilutions methodology should have good reproducibility, regardless of the number of independent assays, whereas, for the Cu-Al LDH-NO 3 , either methodology may produce similar related results.Ultimately, the results obtained in the grain-size dependent toxicity assays (see Section 3.3) for both tested grain sizes (i.e.bulk powder and <25 mm grain size for the Zn-Al and Cu-Al LDH-NO 3 s, respectively) revealed similar EC 50 values to the ones obtained in this part of the study, hence, emphasizing the future tests' good reproducibility involving these particular LDH types.
When looking at the zinc-containing LDHs, the results obtained in our study present comparatively lower toxicity of Zn-Al LDH-NO 3 exposure to R. subcapitata.However, some difficulties were found when comparing results among studies due to the differences in units expressing ecotoxicological endpoints.Koba-Ucun et al. ( 2021) reported a 72-h EC 50 lower than 10 mg/L for the average growth of R. subcapitata exposed to Zn-Fe LDH, a commonly used catalyst.As the authors do not report the EC 50 value based on Zn mass (of the Zn-Fe LDH), the previous comparison was performed considering the total LDH mass concentration, although the ratio of Zn in the NMs from both studies is similar.Furthermore, compared to other zinc or aluminum forms (e.g.nano-forms, such as ZnO, or ionic forms) described in the literature, Zn-Al LDH is also less toxic.Samei et al. (2019) reported that different shapes and sizes of ZnO nanoparticles (small and large spherical and small and large rods) inhibited the growth rate of R. subcapitata at concentrations lower than 0.64 mg Zn/L.Andreani et al. (2020) reported a 72-h EC 50 of 3.90 mg Zn/L (converted to mg Zn/L for a more relevant and convenient comparison) for R. subcapitata average growth rate exposed to ZnO nanoparticles.Aruoja et al. (2009) also reported 72-h EC 50 values of 0.042 mg Zn/L after exposing R. subcapitata to ZnSO 4 and nano-ZnO.
Similar to what was found for Zn-Al LDH, few studies focused on the effects of Cu-Al LDHs or chemically similar LDHs.Ding et al. (2018) reported a 72-h EC 50 of 10.00 ± 2.28 mg/L when exposing the freshwater microalgae Scenedesmus quadricauda to Cu-Mg-Fe LDH.However, as the authors do not express the EC 50 value in mg Cu/L or give the Cu-Mg-Fe LDH's elemental ratio, no reliable toxicity comparison can be drawn.Comparatively to other Cu forms (e.g.nano-CuO, or CuSO 4 ), the Cu-Al LDH used in the present study can be either more or less toxic to freshwater microalgae.Alho et al. (2020) reported a 72-h EC 50 of 0.64 mg Cu/L on the growth inhibition of R. subcapitata upon exposure to CuO NPs (<50 nm) in the same range of magnitude of the Cu-Al LDH-NO 3 (EC 50 ffi 0.41 mg Cu/L).Furthermore, Aruoja et al. (2009) reported an EC 50 of 0.71 mg Cu/L following nano-CuO exposure to R. subcapitata.Additionally, the toxicity of the CuSO 4 (as metal salt control) was also assessed for R. subcapitata with an EC 50 of 0.02 mg Cu/L.This indicates that compared to other nano-forms, the present study's Cu-Al LDH-NO 3 yields comparable Cu based toxicity when exposed to freshwater microalgae, however less toxic than the salt Cu forms like CuSO 4 .

Grain size-dependent toxicity
Smaller NMs/NPs are usually related to higher toxicity due to a higher surface area/volume ratio, leading to higher bioreactivity or dissolution (e.g.Cl ement et al. 2013;Hu et al. 2018;Ivask et al. 2014;Karlsson et al. 2009;Lopes et al. 2014;Pan et al. 2012;Silva et al. 2014;Xiong et al. 2013;Zhao and Wang 2012).
In the present study, although the two LDH-NO 3 types (i.e.Zn-Al and Cu-Al) are the same size, the manufacturer supplied them as they are commercialized, i.e. with different LDH grain size powders.The Zn-Al LDH-NO 3 microalgae test results suggest no differences between the bulk and other grain sizes from an ecotoxicological point of view, as no statistical differences were encountered between EC 50 values.Unlike other NMs, LDHs are reported to rapidly disperse and individualize upon aqueous solutions, acquiring a very similar hydrodynamic size, regardless of the initial powder grain size (Choy et al. 2000;Sun and Dey 2015;Yan and Chen 2013).The present study's DLS results for Zn-Al LDH-NO 3 partially corroborate such behavior, as no clear trend or statistical differences were observed between grain size fractions.
Differently from Zn-Al LDH-NO 3 , toxicological differences are observed between the bulk and the four grain size fractions of the Cu-Al LDH-NO 3 .The EC 50 obtained for the bulk and <25 mm grain size (5.11 and 3.27 mg Cu-Al LDH-NO 3 /L, respectively) suggests that the <25 mm grain size is more toxic to the tested microalgae.The different toxicity between grain sizes might be due to the difference in Cu dissolution, with higher dissolved Cu concentrations being detected in the <25 mm test dispersions, considered an extremely toxic metal to R. subcapitata, even at low concentrations (Al-Hasawi et al. 2020).The fractions 63-125, 125-250, and >250 mm were less toxic than the bulk counterpart (mixture of grain sizes), possibly related to the lower Cu dissolution on those dispersions and not to the hydrodynamic size of the dispersed Cu-Al LDH-NO 3 (since the hydrodynamic size is very similar between grain size fractions).As the grain size increases, the dissolved Cu concentration and growth inhibition effects tend to decrease.For the same microalgae species, Aruoja et al. (2009) reported a 72 h-EC 50 for Zn and Cu of 0.042 and 0.020 mg/L, respectively, lower than the ICP-MS detection limit used in our study (0.050 mg/L).Despite these EC 50 values being lower than the concentrations measured in the dissolved fractions of both LDHs, the concentrations measured in the present work do not reflect the bioavailable fraction of the microalgae.In the dissolved fraction, Zn, Cu, and Al might be complexed with dissolved species present in the media, thus being less available for uptake by R. subcapitata.It should be noted, however, that the dissolved fraction includes the bioavailable fraction of the metallic elements and metallic complexes.As the bioavailable fraction was not assessed, we cannot conclude specifically this difference in toxicity.
The LDHs' toxicity to R. subcapitata reported in our study results from the combined effects of: (i) the exposure to the metallic elements (majorly from bioavailable dissolved fraction), present in both LDH types (Zn, Al, and Cu), which may induce toxicity to algae by generating ROS (Foyer and Shigeoka 2011;Xia et al. 2015), impairing the photosynthetic system (Gunawan et al. 2013), or by increasing the lipid peroxidation of the algae's cell membrane (Ozkaleli and Erdem 2018);(ii) the interactions between LDHs and algae cells, leading to mechanical effects inducing agglomeration and sedimentation, shading effects with reduced available light, or by adsorption of the LDHs to the algae's cell wall (D eniel et al. 2019;Figueiredo et al. 2019;Hartmann et al. 2010;Hartmann et al. 2013;Schwab et al. 2011).
Several alterations to test methodologies addressing the previously mentioned interactions are starting to be reported in the literature, leading to less variability in the results, hence increasing the relevancy and reliability of the toxicological results (Hartmann et al. 2013;Hund-Rinke et al. 2016).Moreover, the nitrate anion concentration in test media probably does not explain or affect the toxicological profile of both LDHs (regardless of the tested grain size or tested concentration) since Jeanfils et al. (1993) reported a slight decrease in the growth of Chlorella vulgaris only upon exposure to nitrate concentrations above 310 mg/L.
Finally, Zn-Al LDH-NO 3 proved to be less toxic than Cu-Al LDH-NO 3 , considering all EC 50 values obtained.The ecotoxicological difference between LDHs results from their metallic composition.According to a study published by Al-Hasawi et al. (2020), Cu yields higher toxicity to R. subcapitata than Zn and Al, the latter being the least toxic.Moreover, similar results have been previously reported by Ouyang et al. (2012) for C. vulgaris.

Conclusions
The exposure methodology for reproducible microalgae testing depends on the NM tested since different reproducibility of ecotoxicity results were obtained using two similar LDH-NO 3 .Notwithstanding, the serial dilutions exposure methodology proved suitable for both Zn-Al LDH-NO 3 and Cu-Al LDH-NO 3 .Besides the NM size, powder grain size is an interesting feature to investigate regarding commercialized NM provided as powders.No grain size-dependent toxicity was observed for the Zn-Al LDH-NO 3 , while for the Cu-Al LDH-NO 3 , the three grain size fractions are of interest when considering possible future applications due to their lower toxicological profile comparatively to the bulk counterpart (with mixed grain sizes).Results like these are crucial to be reported to manufacturers, leading to a selection of particular grain sizes for specific applications, where less/no environmental hazard profiles could be observed.
Furthermore, Zn-Al LDH-NO 3 revealed lower toxicity to R. subcapitata than Cu-Al LDH-NO 3 ; therefore, in a specific application where both LDHs could be employed, Zn-Al LDH-NO 3 would be advised as the NM to be incorporated.Future studies should also consider using other state-of-the-art techniques for the LDHs' physicochemical characterization (e.g.particle size distribution by electric birefringence-based methods) over commonly used ones like DLS, which do not account for inherent specificities (e.g.NM geometry, or shape).Moreover, addressing the LDHs' behavior (by sedimentation/agglomeration preliminary testing to ensure constant NM test concentrations) is a major advantage for deriving a testing strategy for accurate exposure and derivation of effects.Further ecotoxicological testing with other ecologically relevant freshwater species and from different trophic levels should be performed to ensure a full hazard assessment, which is critical for regulatory purposes.
Ultimately, this study contributes toward a most needed NM test methodology harmonization, enabling a realistic inter-laboratory data comparison while granting environmental relevance to the study.
3.1.3.Dissolution of Zn, Cu, and Al from LDH nanomaterials The concentrations of particulate and dissolved Zn and Al, and Cu and Al measured in Zn-Al and Cu-Al LDH-NO 3 dispersions are displayed in Tables

Table 1 .
Zeta-potential (fP)and hydrodynamic size as indicators of dispersion stability for each Zn-Al LDH-NO 3 grain size dispersion (10 mg/L).

Table 2 .
Zeta-potential (fP)and hydrodynamic size as indicators of dispersion stability for each Cu-Al LDH-NO 3 grain size dispersion (10 mg/L).

Table 3 .
Nitrate concentration released (mg NO 3 -/L) in time for all Zn-Al LDH-NO 3 grain size fractions at 60, 120, and 210 mg/L nominal concentrations.

Table 4 .
Nitrate concentration released (mg NO 3 -/L) in time for all Cu-Al LDH-NO 3 grain size fractions at 1, 5, and 9 mg/L nominal concentrations.

Table 5 .
Measured concentrations of particulate and dissolved Zn and Al, at 0 h and 72 h in MBL media for all grain size fractions of Zn-Al LDH-NO 3 at concentrations of 60, 120, and 210 mg/L (n ¼ 1).

Table 6 .
Measured concentrations of particulate and dissolved Cu and Al, at 0 h and 72 h in MBL media for all grain size fractions of Cu-Al LDH-NO 3 at concentrations of 1, 5, and 9 mg/L (n ¼ 1).
< 3.841, p > 0.05) between EC 50 values of each independent test for the same methodology.

Table 9 .
No observed effect concentration (NOEC), lowest observed effect concentration (LOEC), and EC 50 values for the specific growth rate of Raphidocelis subcapitata exposed to Zn-Al LDH-NO 3, for all grain size fractions.