Ecotoxicity of selected carbon-based nanomaterials

The widespread use of the nanomaterials increases the emission of nano-pollutants into the water. Carbon nanomaterials are particularly interesting. They are characterized by relatively stable structure, which makes them able to migrate and accumulate in the environment. Therefore, the aim of this study was to determine the potential toxicity at the different trophic levels of four selected carbon nanostructures: graphene oxide (GO), reduced graphene oxide (rGO), multi-walled carbon nanotubes (MWCNTs) and oxidized multi-walled carbon nanotubes (f-MWCNTs) on indicators at three trophic levels. Producers was represented by Lemna minor in growth inhibition test. The ecotoxicological effect for consumers was estimated by acute tests on Artemia franciscana, Brachionus calyciflorus and Thamnocephalus platyurus, while the acute toxicity on decomposers was studied on bacteria Escherichia coli. Results show that the short-term exposure on MWCNTs, f-MWCNTs, GO and rGO can be toxic at three trophic levels. The influence of the tested materials was much higher for the consumers, than for the producers. The lowest toxicity from all researched carbon-based nanomaterials was presented by GO. Moreover, generation of high reactive form of oxygen, mechanical damage of cell wall and membrane is one of the main toxicity mechanism; thus, the toxicity depends heavily of the dose and the shape of the nanomaterials.


Introduction
The application of nanomaterials allows to change the fundamental chemical and physics properties of conventional materials. Changing material properties is possible as their size reduced to the nanoscale, what offering new materials with unique optical, electrical and mechanical properties. Nanomaterials are increasingly used in such fields as biomedicine, electronics, mechanical engineers, photonics, energy generation and storage (Basiuk et al. 2011;Novoselov et al. 2012;Montagner et al. 2016). Among all the nanomaterials, special attention should be paid to the carbon-based nanomaterials such as the carbon nanotubes (CNTs) and graphene (Montagner et al. 2016). Currently, these nanomaterials are widely studied, and hold immerse to impact several scientific areas (Lalwani et al. 2016).
Graphene is a 2D layer sheet of sp 2 hybridization carbon atoms. Basic graphene is a polar and hydrophobic, so to improve its dispersibility in aqueous media its need to be oxidized for example via chemical oxidation (Lalwani et al. 2016). Currently, the production of graphene-based nanomaterials is not as important for the global economy as carbon nanotubes (CNTs). However, investment in graphene-based nanomaterials is expected to reach $ 400 million by 2025 (Montagner et al. 2016). Such a large production and industrial application of both graphene and CNTs nanomaterials would transfer these nanomaterials into the environment where they may have potential deleterious impact on the flora and fauna as well as microorganisms.
Current studies of the toxicological effect of carbon-based nanomaterials are mainly focused on the animal and human cells, while the studies considering their effect on the development and physiology of plants, algae, fungi and bacteria are limited (Boros and Ostafe 2020;Penga et al. 2020). In this study Lemna minor (duckweed), Artemia franciscana, Brachionus calyciflorus (rotifer), Thamnocephalus platyurus (shrimp) and Escherichia coli were used to evaluate the ecotoxic effect of four carbon nanomaterials (graphene oxide-GO, reduced graphene oxide-rGO, multi-walled carbon nanotubes-MWCNTs, functionalized multiwalled carbon nanotubes-f-MWCNTs). The test organisms selected for the study are characterized by fast growth, small size, ease in culturing and cost effectivity as well as are sensitivity to toxic substances (Gamoń et al. 2019), that make them good indicators. Khodakovskaya et al. (2011) investigated the genotoxicity of several carbon nanomaterials, such as activated carbon, graphene, SWCNTs and MWCNTs on the germination of tomato seedlings. Results of the research over leaves and roots show that graphene induced the lowest activation of stress-related LeAqp2 gene encoding tomato water-channel protein, while the highest activation stress was observed for CNTs. Tabei et al. (2019) studied the toxicity of MWCNTs and showed that it has very high level of phagocytic activity for undifferentiated HL-60 cells and a low level of cytotoxicity for differentiated HL-60 cells. MWCNTs affect the repair mechanism of DNA, showing genotoxic activity. Moreover, the molecular larval studies at aquatic levels conducted by Martinez-Paz et al. (2019) showed that MWCNTs affect the transcription of genes involved in apoptosis. On the other hand, important factors influencing the MWCNTs toxicity are also their length, diameter and structure. Stronger toxicity is revealed by the longer tubes (Shi et al. 2017).
Antifungal effects of reduced graphene oxide against Aspergillus niger, Aspergillus oryzea and Fusarium oxysporium were investigated by Sawangphruk et al. (2012) via quantifying mycelial growth inhibition. This research indicated a good antifungal activity of rGO reaching IC 50 value between 50 and 100 µg/mL against all three fungi. Antimicrobial activity of the graphene oxide nanowall and reduced graphene oxide nanowall was reported by Akhavan et al. (2013). These results showed that rGO nanowalls were more toxic against Escherichia coli and Staphylococcus aureus than the GO nanowalls because of effective charge transfer between bacteria cells and the edges of nanomaterials during direct contact. However, E. coli demonstrated lower impact than S. aureus as effect of cell wall structures. Gram-positive S. aureus doesn't include cell membrane compered to gramnegative E. coli making them less resistant (Akhavan et al. 2013;Lalwani et al. 2016). Chen et al. (2013) compared antimicrobial activity of the GO and rGO against gram-negative Xanthomonas oryzae pv. oryzae. However, these results showed grater antibacterial activity of the GO in comparison to the rGO. Graphene phytotoxicity was evaluated by Begum et al. (2011) on roots as effect of shoot growth and shape, cell death and biomass by incubating seedlings of cabbage, tomatoes, red spinach and lettuce with 500-2000 mg/L for 20 days. The physiological and morphological analysis showed that graphene significantly inhibited plant growth and biomass production. Nanomaterials toxicity was also investigated in some model marine organisms. Artemia salina was not affected by graphene concentration in range 0.675-5 mg/L (Pretti et al. 2014), while in concentration between 0 and 700 mg/L acute mortality was observed by Mesarič et al. (2015). Graphene-related materials also have an effect on the lysosomal function and metabolic activity of fish cells and the degree of graphitization is highly related to the toxicity-they are inversely proportional (Kalman et al. 2019). Analysis provided by Evariste et al. (2020) indicated that communities and organisms from the sediment are impacted more than those from the water column due to sedimentation of the CNTs.
Development of nanotechnology including carbon nanomaterial bring potential risk for environment. Especially water environment is regarded to be more influenced by all types of nanoparticles, as these compounds are washed out from other parts of the environment and directed to water from soil and air. Despite an increasing number of studies examining the toxicity of carbon nanomaterials, current knowledge is still insufficient, and it is extremely important to research their potentially hazardous effects on the ecosystem (Ghadimi et al. 2020). Therefore, in this study the ecotoxic effect of four carbon nanomaterial at different trophic levels was investigated to present the full spectrum of the results of these selected types of carbon nanomaterials. This knowledge enables to understand the mechanisms of the carbon nanomaterials influence on variety of organisms in the environment.

Carbon nanomaterials characteristics
Commercially available GO, rGO (Nano Carbon, Poland) and MWCNTs (Nanocyl NC7000™, Belgium) were used. MWCNTs were functionalized according to Kolanowska et al. (2019) by oxidative treatments with a mixture of nitric acid and sulfuric acid solution to obtain f-MWC-NTs. Nanomaterials samples were analysed by transmission electron microscope using a S/TEM TITAN 80-300 (FEI Company). The samples were prepared by dispersing the nanoparticles in ethanol (99.8%) using an ultrasound washer, and then depositing one or two drops of the dispersion on a TEM copper 200 mesh grids coated with a carbon film. The samples were air-dried at room temperature. Nanomaterials solutions for toxicity assays were prepared in test media. Each test was provided with the same concentration of nanomaterials-25, 50, 100, 150 and 200 mg/L for GO and rGO, while MWCNTs and f-MWCNTs concentrations were 125, 250, 500, 750 and 1000 mg/L.

Phytotoxicity
The phytotoxicity was evaluated using growth inhibition tests on Lemna minor, according OECD Test No. 221 (2006). L. minor plants, with at least 12 leaves, were placed in plastic Petri dishes containing 20 mL of the test solution and incubated in a thermostatic cabinet, at 24 ± 1 °C under 16 h photoperiod (16 h of light: 8 h of darkness). After 14 days of incubation, the number of leaves was counted, and the percentage of the growth inhibition was determined. Standard growth medium for macrophytes was used as a negative control and to prepare nanomaterials solutions. Tests were performed in triplicate and IC 50 values were calculated with a probit model.

Artoxkit M
The acute toxicity test-Artoxkit M kit was obtained from MicroBioTests Inc. (Belgium). It is based on the mortality of Artemia franciscana. Test was developed by the research team at the State University of Ghent (Belgium) (Van Steertegem and Persoone 1993) and is based on the ASTM Standard Guide E1440-91 (2012). Assay was performed according to the standard procedure (Artoxkit 2020). The standard seawater medium with a salinity of 35 ppt was prepared with deionized water and NaCl. It was used to prepare nanomaterials dilutions and as a negative control. The bioassay was conducted in a multi-well test plate with 24 (6 × 4) test wells. Instar II-III larvae of the brine shrimp Artemia franciscana were hatched from cysts 30 h before the start of the toxicity test. The test was based on one control and five nanomaterials concentrations, each with 3 replicates of 10 animals. The test plates were covered with a strip of parafilm and incubated at 25 °C in darkness for 24 h. IC 50 values were calculated with a probit model.

Rotoxkit F
The acute toxicity was assessed based on the mortality of Brachionus calyciflorus after 24 h exposure in accordance with ASTM Standard Guide E1440-91 (2012). Thamnotoxkit F kit was obtained from MicroBioTests Inc. (Belgium). Assay was performed according to the standard procedure (Rotoxkit 2020). The standard freshwater medium consisted of NaHCO 3 (96 mg/L), CaSO 4 ·2H 2 O (120 mg/L), MgSO 4 ·7H 2 O (123 mg/L), KCl (4 mg/L) was used to prepare nanomaterials dilutions and as a negative control. Larvae of Brachionus calyciflorus hatched from cysts after 18 h of incubation were placed to the multi-well plates with one control solution and five toxicant concentrations, each with 6 replicates of 5 animals. The test plates were covered with a strip of parafilm and incubated at 25 °C in darkness for 24 h. IC 50 values were calculated with a probit model.

Thamnotoxkit F
The acute toxicity test-Thamnotoxkit F kit was obtained from MicroBioTests, Inc. (Belgium). It is based on the mortality of Thamnocephalus platyurus. Test was developed by the research team by the research team of Prof. Dr. G. Persooneat at the State University of Ghent (Belgium) and is based on the standards of ISO 14380 (2011). Assay was performed according to the standard procedure (Thamnotoxkit 2020). The standard freshwater medium consisted of NaHCO 3 (96 mg/L), CaSO 4 ·2H 2 O (120 mg/L), MgSO 4 ·7H 2 O (123 mg/L), KCl (4 mg/L) was used to prepare nanomaterials dilutions and as a negative control. Larvae of Thamnocephalus platyurus hatched from cysts after 24 h of incubation were placed to the multi-well plates with one control and five nanomaterials concentrations, each with 3 replicates of 10 animals. The test plates were covered with a strip of parafilm and incubated at 25 °C in darkness for 24 h. IC 50 values were calculated with a probit model.

Toxi-ChromoPad™
The acute sediment toxicity test is based on the ability of nanomaterials to reduce the synthesis of an inducible enzyme-β-galactosidase-in a highly permeable mutant of the bacterium E. coli. Toxi-ChromoPad™ was obtained from Environmental Biodetection Products, Inc. (Canada). Test was performed according to the standard procedure. Bacteria-sediment mixture solution was placed to the test tubes with one control and five nanomaterials concentrations. All tubes were incubated for two hours at 37 °C. After this time test tubes were hand-mixed vigorously and a small drop from each tube was transferred onto the chromogenic pad. Then, samples were incubated overnight at 37 °C before results reading. Blue colour intensity is correlated with enzyme synthesis. Toxicity reduces enzyme synthesis and colour development. Blue colour intensity was measured based on image histogram.

Results and discussion
Nanomaterials morphology was observed using TEM and Figs. 1 and 2 present representative images of the observed samples. MWCNTs are elongated hollow nanotubes formed by curled multi-layer graphite sheets. MWCNTs functionalized by acid treatment contains carboxyl and hydroxyl groups at the walls (Sezer and Koç 2019). Both types of carbon nanotubes were characterized by an average inner diameter of 2-6 nm, outer diameter of 5-20 nm and lengths in the microns. In turn, GO is a single layer of carbon atoms (graphene), comprising carboxyl and hydroxyl groups. Reduction of GO removes some of these oxygen-containing groups and leads to the formation of rGO. Both of them occur in the water environment in the form of few-layer flakes and Fig. 2 illustrates much higher disintegration of the carbon basal plane in the rGO sample.
As it has been previously reported that even a small change in the physical-chemical characteristics of the nanomaterial can be a reason of its toxicity change (Manshian et al. 2013;Parise et al. 2014). However, the detailed mechanism of CNMs toxicity is still unclear and requires further elucidation, but three main phenomena are currently considered: (1) mechanical damage to the microorganism's cell wall and membrane; (2) reactive oxygen species (ROS) production, which can damage membranes, proteins, and DNA; (3) other physical-chemical properties related, i.e. with their functionalization or electron transport and multielectron accepting ability (Chen et al. 2019).
The ecotoxicological effect of the chosen carbon-based nanomaterials was estimated at three trophic levels in order to present the wide spectrum of these compounds influence on living organisms. For this experiment graphene oxide, reduced graphene oxide, multi-walled carbon nanotubes and their functionalized form were used. rGO and MWCNTs are hydrophobic, while GO and f-MWCNTs are hydrophilic , thus more toxic towards bacteria (Nguyen and Rodrigues 2018;Santos et al. 2012). These features should be taken into consideration while discussing their influence on living organisms. Moreover, much concentrated graphene-based nanomaterials aggregate in larger complexes, becoming less available for microorganisms and limiting their toxic activity. Their features differ; thus, it should be expected that the mechanism of their influence on various organisms will differ as well.
Lemna minor was a representative of produces for ecotoxicological tests. According to the results obtained in this experiment the increasing concertation of GO, rGO in range of 25-200 mg/L and the range of 125-1000 mg/L for MWCNTs is successfully inhibiting the producer's growth. In case of all used GO concentrations, as well as rGO at concentration 100, 150 and 200 mg/L and MWCNTs at concentration of 250, 500, 750 and 1000 mg/L the results are statistically significant. Interestingly for functionalized form of MWCNTs the increasing inhibition was observed for concentration of 125, 250 and 500 mg/L, while above this value the inhibition was decreasing for ca. 4% and 6% for 750 and 1000 mg/L, respectively. The results obtained for 125-750 mg/L were statistically significant (Fig. 3).
The 50% lethal concentration (LC 50 ) is comparable for GO and MWCNTs (543 and 431 mg/L, respectively). For rGO LC 50 is lower-144 mg/L. No observed effects concentration (NOEC) was estimated for rGO and MWCNTs at level of 50 and 125 mg/L, while for GO it was not possible. The highest value for the lowest observed effects concentration (LOEC) was also for MWCNTs, while for GO and rGO it was 25 and 100 mg/L. For functionalized MWCNTs it was impossible to estimate LC 50 as well as NOEC and LOEC values ( Table 1).
The ecotoxicological effect for consumers was estimated on the basis of the tests for Artemia franciscana, Brachionus calyciflorus and Thamnocephalus platyurus. In case of A. franciscana no inhibition effect was estimated for GO at concertation of 25 mg/L and for MWCNTs at concertation of 125 mg/L. For rGO and f-MWCNTs all researched concentrations were inhibiting test organisms. It is also important to notice that the influence of the tested substances was far higher for consumers, than for producers. It could be explained with the presence of plant cell wall preventing the cell from the carbon-based nanomaterials penetration (Khodakhowskaya et al. 2011).
In case of L. minor, the highest observed inhibition value (ca. 18%) was for MWCNTs at concertation of 1000 mg/L, while for producers only the lowers concentrations values for all tested substances gave the mortality result lower than 20% (Fig. 4). The highest used concentrations of GO and rGO caused higher mortality of A. franciscana (95% and 90%, respectively) than the highest used concentrations of MWCNTs and f-MWCNTs (90% and 66%, respectively) (Fig. 4). The values of LC 50 , NOEC and LOEC were comparable for rGO and GO. LC 50 for MWCNTs was slightly lower (568 mg/L) than for f-MWCNTs (631 mg/L). In case of NOEC and LOEC the valued obtained for MWCNTs and f-MWCNTs were twice higher (250 and 125 mg/L for NOEC and 500 and 250 mg/L for LOEC, respectively (Table 2).
B. calyciflorus appeared to be more sensitive towards the tested compounds in their lower concentrations. Only 25 mg/L GO did not caused mortality for this testing organism. For GO, rGO and MWCNTs their increasing concertation caused increasing mortality with the highest value of 98 and 95%, respectively, for GO and rGO at concentration of 200 mg/L and 75% mortality for 1000 mg/L of MWCNTs. It is important to point that the mortality for the increasing concentration of functionalized MWCNTs was the highest for 500 mg/L, but for 750 and 1000 mg/L the mortality per cent dropped from over 80% to 32 and 35%, respectively (Fig. 5). It could be explained by the fact that more oxidized carbon-based nanomaterials are more cytocompatible thus less harmful for animal cells (Sasidharan et al. 2011;Lalwani et al. 2016).
As in many animal-based studies these results also show that GO presents the lowest toxicity from all carbon-based nanomaterials used in this study (Talukdar et al. 2014;Yuan et al. 2014). It was also stated that the toxicity of these materials depends heavily on dose and the shape of the structure (Akhavan et al. 2013;Lawani et al. 2016).
It was also impossible to estimate LC 50 for f-MWCNTs. NOEC and LOEC values for this compound were 125 and 250 mg/L, respectively. LC 50 for GO, rGO and MWCNTs were 87, 65 and 712 mg/L, respectively. The value of NOEC and LOEC for rGO were lower than for GO (25 and 50 mg/L for NOEC and 50 and 100 mg/L for LOEC, respectively). The NOEC and LOEC values for MWCNTs were 125 and 250 mg/L, respectively (Table 3).
For T. platyurus only in case of rGO the lowest concertation used (25 mg/L) did not caused mortality effect for this testing organism and values of 50 and 100 mg/L have similar mortality effect (ca. 25%; Fig. 6). GO low concentrations (25 and 50 mg/L) have similar mortality effect on T. platyurus while concentration 100 mg/L and above caused increasing mortality reaching 92% for 200 mg/L. Interestingly, MWCNTs influence seems to be very strong. This substance Fig. 3 Growth inhibition of Lemna minor after 14 days of exposure on A graphene oxide (GO), B reduced graphene oxide (rGO), C multi-walled carbon nanotubes (MWCNTs) and D functionalized multiwalled carbon nanotubes (f-MWCNTs). Bars represent standard error; *a statistically significant difference with control at p < 0.05 (t-test) Table 1 The results of Lemna minor Inhibitory Concentration, 10% (IC 10 ); No Observed Effects Concentration (NOEC) and Lowest Observed Effects Concentration (LOEC) induced by graphene oxide (GO), reduced graphene oxide (rGO), multi-walled carbon nanotubes (MWCNTs) and functionalized multi-walled carbon nanotubes (f-MWCNTs) IC 10  543  144  431  -NOEC  -50  125  -LOEC  25  100 250 concentration at 125 mg/L caused 18% mortality and this rate is increasing heavily and reaching 82% for 1000 mg/L. Functionalized MWCNTs also influence T. platyurus more heavily than rGO and GO, but the concentrations between 125 and 750 mg/L did not give such a strong effect as nonfunctionalized MWCNTs (mortality at level of 10-42%). However, the mortality for the highest concentration of f-MWCNTs unexpectedly reached almost 100% (Fig. 6). The values of NOEC and LOEC for GO and rGO presented similar trend as for B. calyciflorus: 50 and 25 mg/L for GO and 100 and 50 mg/L for rGO, respectively. The LC 50 for these compounds was at the comparative level (94 and 99 mg/L, respectively). LC 50 as well as NOEC and LOEC values for f-MWCNTs were twice higher than for MWCNTs. LC 50 , NOEC and LOEC for f-MWCNTs were 894, 250 and 500 mg/L, respectively. While for MWCNTs these values were 404, 125 and 250 mg/L, respectively (Table 4).
The Toxi-ChromoPad™ tests the toxic influence of the compounds studied on decomposers-bacteria Escherichia coli producing-β-galactosidase. The increase in the toxicity causes the decrease in enzyme production and the colour development. According to the results obtained in this test it could be stated that GO and rGO in concentration from 25 to 200 mg/L are less toxic than MWCNTs and f-MWCNTs used in the concertation between 125 and 1000 mg/L. For GO/rGO the values of colour intensity are estimated at level for 4-6 a.u. (arbitrary unit), while for MWCNTs/f-MWCNTs at level of ca. 4 a.u (Fig. 7). These results let to suspect that in case of carbon nanomaterials the more complex and sharper the structure is, more harmful it is to bacterial cell wall. As it has been already studied in case of GO and rGO, the latter is found its sharper structure and its better charge transfer ability (Yang et al. 2013). Considering mechanical damages of cell membranes caused by nanoparticles, TEM images demonstrate that rGO is more irregularly shaped than GO, with sharp folds and edges. It is also possible that in case of MWCNTs and f-MWCNTs there are far more points of the direct contact for the bacterial cells which direct to their damage as it has been presented by Kurantowicz et al. (2015) who hypothesized that the presence of the oxidative functional groups on graphene-based nanomaterial can act as an attractant for bacterial cells enabling the direct Fig. 4 Mortality of Artemia franciscana after 24 h of exposure on A graphene oxide (GO), B reduced graphene oxide (rGO), C multi-walled carbon nanotubes (MWCNTs) and D functionalized multi-walled carbon nanotubes (f-MWCNTs). Bars represent standard error; *a statistically significant difference with control at p < 0.05 (t-test) contact and cell impairment as its consequence. Moreover, according to Saleemi et al. (2020) functionalized CNTs can destroy cell membrane and affect microbial activity. As illustrated in Fig. 1B, f-MWCNTs have sharp ends, with an average diameter between 5 and 20 nm, which could easily damage membrane integrity.

Toxicity mechanism
The toxicity mechanism of carbon nanomaterials dependent mostly on the type of cells on which it acts (prokaryotic or eukaryotic), as well as the physicochemical properties of the nanomaterials as shape, size, types and density of functional group and charge transfer abilities (Lalwani et al. 2016). The generation of reactive oxygen species (ROS) is mentioned as the main toxicity mechanism of nanomaterials against various types of cells inducing damage to proteins and DNA which finally leads to cell death via apoptotic or necrotic pathways (Li et al. 2012). Previous studies have also reported that directly contact of sharp edges of carbon nanomaterials (graphene, CNTs) with cells lead to loss of cell integrity as well as may cause DNA fragmentation and chromosomal aberration (Akhavan et al. 2013). Moreover, it was observed that long sheet of graphene nanomaterials can wrap around bacterial cell leading to growth inhibition (Carpio et al. 2012). The strong cytotoxicity effect may be the result of integration between graphene-based nanomaterials and phospholipids from the cell membrane, which are extract from the membrane structures (Tu et al. 2013).
Thus, it is difficult to demonstrate the evidential mechanism of toxicity of nanomaterials on cells. It is suggested that the toxic mechanism is the result of the simultaneous action of nanomaterials on different cellular structures.

Conclusion
Presented work investigates the ecotoxicity of four carbon nanomaterials at different trophic levels. Results show that the short-term exposure on MWCNTs, f-MWCNTs, GO and rGO can be toxic at all trophic levels. However, the influence of the tested materials was much higher for consumers, than for producers, what can be explained with the presence of plant cell wall preventing the cell from the carbon-based  nanomaterials penetration. The toxicity of these materials depends heavily of dose and the sharpness of the structure, because mechanical damage of cell wall and membrane is one of the main toxicity mechanism. Furthermore, GO presents the lowest toxicity from all the carbon-based nanomaterials used in this study. Considering mechanical damages, Fig. 6 Mortality of Thamnocephalus platyurus after 24 h of exposure on A graphene oxide (GO), B reduced graphene oxide (rGO), C multi-walled carbon nanotubes (MWCNTs) and D functionalized multiwalled carbon nanotubes (f-MWCNTs). Bars represent standard error; *a statistically significant difference with control at p < 0.05 (t-test)

Fig. 7
Blue colour development induced by β-galactosidase in E. coli after exposure on A graphene oxide (GO), B reduced graphene oxide (rGO), C multi-walled carbon nanotubes (MWCNTs) and D functionalized multi-walled carbon nanotubes (f-MWCNTs). a.u. arbitrary unit TEM images demonstrates that rGO and CNTs aggregates are more irregular in shaped than GO, with the sharp folds and edges. Obtained results are an important contribution to evaluate and understand hazards resulting from the carbon nanomaterials presence in the water environment. However, studies on much lower concentrations between pg/L and ug/L during long-term exposure and accumulation in real environment should be take into consideration during further studies.