Registration Dossier

Environmental fate & pathways

Endpoint summary

Administrative data

Description of key information

Statement literature and substance evaluation:

A large number of studies identified during the literature search applied so-called titanium dioxide as test item without any or with limited documentation on the identity of the tested substance. It is problematic to base the hazard and risk assessment of clearly characterised titanium dioxide on information generated with poorly characterised test items that are without any market relevance. Consequently, all studies conducted with self-synthesised titanium dioxide nanomaterials were rated not relevant for hazard and risk assessment purposes (in accordance with ECHA Guidance Chapter R.4: Evaluation of available information, Version 1.1, December 2011). Furthermore, all references obtained during the literature search were evaluated according to the criteria laid down in ECHA Guidance R.4 (Version 1.1, December 2011), the OECD Guidance Document 34 (2005) and the reference by Klimisch et al. (1997). The screening foresees an incremental procedure evaluating the relevance, reliability and adequacy (for further information refer to the attached statement on the literature evaluation criteria in IUCLID section 13).

The main environmental processes determining the environmental fate of nano- and microsized TiO2 separated in four categories are of different importance (see Table).

Table: Importance of environmental processes

             

Environmental process

Low

Medium

High

Chemical processes

Solubility/dissolution

micro/nano

 

 

 

Physical processes

Aggregation/Agglomeration

 

 

micro/nano

 

Sedimentation

 

 

micro/nano

 

Adsorption/desorption

Soil retention

 

 

micro/nano

 

Retention in sewage treatment plants

 

 

micro/nano

 

Biologically mediated processes

Biodegradation

Not relevant

 

 

 

Biodegradation reactions are not relevant for metals and metal compounds that are not biodegradable, including TiO2. However, the degradation of the coatings of TiO2 materials has to be considered since the coatings may alter the surface characteristics of the TiO2 material that determines the agglomeration behaviour and further fate of TiO2 materials in aqueous environments.

Titanium has very low mobility under almost all environmental conditions, mainly due to the high stability of the insoluble oxide TiO2 under all, but the most acid conditions, i.e., below pH 2 (Brookins, 1988 referenced in Salminen et al, 2015). Titanium only exists in a fully hydrated form, TiO(OH)2, in water above pH 2, and is thus transported in a colloidal state rather than as dissolved ion. The low concentrations of ‘dissolved’ Ti decrease typically further with increasing ionic strength.

Transformation dissolution data (OECD Series No. 29) indicate that nano- and microsized TiO2 is poorly soluble in environmental media. Thus, dissolution is not considered to be an important process for nano- and microsized TiO2 materials in the environment.

For the assessment of TiO2 nanomaterials, the dissolution as well as the dispersability are key factors affecting environmental toxicity according to the OECD Series on the Safety of Manufactured Nanomaterials No. 40.

In aqueous environments agglomeration and subsequent sedimentation of TiO2 materials was observed in several studies. Results indicate that the aggregation of n-TiO2 depends on several water parameters, such as ionic strength, pH, presence of relevant ions, concentration and source of natural organic matter (NOM), turbidity, alkalinity, but also on initial TiO2 concentrations and on the surface properties of the TiO2 material as well as the interaction of all these factors. While pronounced sedimentation resulting in a transport of TiO2 particles from the water to the sediment phase was observed in most natural waters, stable TiO2 nanoparticle aggregates were also present in the water column in some natural water samples. Overall, sediments appear to be the main receiving environmental compartment in the aquatic environment but the presence of TiO2 aggregates in the water phase cannot be ruled out.

Solid-solution partitioning of TiO2 is a key property that determines its environmental fate in terrestrial and aquatic systems. The Kd value can be used to provide information on potential exposure pathways and which biota are likely to be most relevant once the contaminant is released into the environment.

Different partition coefficients were derived from monitoring data for dissolved/dispersed Ti concentrations in water and corresponding sediments or suspended matter. These results correspond to steady-state conditions of Ti in the respective environmental compartment, independent of Ti speciation. High sediment/suspended matter - water distribution coefficients provide evidence that TiO2 is effectively removed from the water phase and mostly associated with particles or colloids of the sediment / suspended matter.

Dissolved/dispersed titanium concentrations in European stream waters, can go up to 16.8 µg/L but are generally below 4.5 µg/L (95th P = 4.21 µg/L). Titanium exists only in a fully hydrated form, TiO2 * n H2O, in water above pH 2, and is transported in a colloidal state rather than as dissolved ion .

Soil leaching (OECD TG 312) and soil adsorption (OECD TG 306) tests indicate a low mobility of nanosized TiO2 materials in soil and provide further evidence for the adsorption/ retention in soil. A similar lack of mobility may be assumed for microsized material based on common physico-chemical properties including solubility or the lack thereof. 

Studies on the fate of TiO2 in wastewater indicate that up to 97% of TiO2 is removed during sewage treatment. Based on monitoring data, it is expected that the entry of nanosized-TiO2 materials into the aquatic environment via STP effluents is low.

Based on the stability of the insoluble oxide TiO2 under environmental conditions, it is assumed that the dissolved Ti ion accounts to a limited extent for the uptake of nano- and microsized TiO2 materials into organisms (if any). Terrestrial and aquatic bioaccumulation data indicate an absence of a bioaccumulation and biomagnification potential of TiO2.

Additional information

Solubility of microsized TiO2

Klawonn et al. (2016) determined the water solubility of microsized TiO2 according to OECD TG 105 in ultrapure water. Under the conditions of this test (flask method, mean loading of 500.6 mg/L), a solubility equilibrium was not observed during the test. Mean background-corrected dissolved Ti concentrations of test solutions in samples (operationally defined as fraction < 2.1 nm, i.e. after filtration through a 0.2 µm membrane and centrifugal filtration) ranged from 0.024 µg Ti/L to 0.172 µg Ti/L, which corresponds to 0.04 to 0.287 µg TiO2/L.

The solubility of microsized TiO2 in environmental media is low based on results of transformation / dissolution (T/D) testing of three TiO2 materials (Klawonn et al. 2017a-c). The T/D of three microsized TiO2 materials, ranging in particle size from 120 to 165 nm and specific surface area from 7 to 13 m²/g, was tested according to OECD Series No. 29. Measured dissolved Ti concentrations (operationally defined as the dissolved Ti fraction after centrifugal filtration (~2.1 nm)) after 28 d at a loading of 1 mg/L at pH 6 and pH 8 were below the limit of detection / quantification (<0.11 / <0.34 µg Ti/L). After 28 d and at a loading of 1 mg/L, suspended Ti concentrations (operationally defined as the Ti fraction after 0.2 µm filtration) were below the LOD/LOQ (<0.11/<0.34 µg Ti/L) at pH 8, whereas at pH 6 they ranged from below the LOQ (<0.34 µg Ti/L) to 3.76 µg/L during the 28-day test. These findings indicate that microsized TiO2 materials do not dissolve to any relevant extent under regular environmental conditions. Further, measured Ti concentrations in the water phase decreased with decreasing filter size fraction (0.2 µm vs 2.1 nm), suggesting that Ti is associated with particles or colloids.

In sum, microsized TiO2 does not dissolve to any relevant extent under regular environmental conditions. Thus, dissolution is not an important process for the environmental fate and behavior of microsized TiO2 in environmental compartments.

Solubility of nanosized TiO2

The solubility of nanosized TiO2 in environmental media is low based on results of transformation/dissolution (T/D) testing of three TiO2 materials (Klawonn et al. 2017d-f). The T/D of three TiO2 materials ranging in their particle size from 19 to 23 nm and specific surface area from 50 to 85 m²/g was tested according to OECD Series No. 29. Measured dissolved Ti concentrations (operationally defined as the dissolved Ti fraction after centrifugal filtration (~2.1 nm)) after 28 d at a loading of 1 mg/L at pH 6 and pH 8 were below the limit of detection / quantification (<0.11 / <0.34 µg Ti/L). After 28 d and at a loading of 1 mg/L, suspended Ti concentrations (operationally defined as the Ti fraction after 0.2 µm filtration) were below the LOD/LOQ (<0.11/<0.34 µg Ti/L) at pH 8, whereas at pH 6 they ranged from below the LOQ (<0.34 µg Ti/L) to 5.62 µg/L during the 28-day test. These findings indicate that nanosized TiO2 materials do not dissolve to any relevant extent under regular environmental conditions. Further, measured Ti concentrations in the water phase decreased with decreasing filter size fraction (0.2 µm vs 2.1 nm), suggesting that Ti is associated with particles or colloids.

In sum, nanosized TiO2 does not dissolve to any relevant extent under regular environmental conditions. Thus, dissolution is not an important process for the environmental fate and behavior of nanosized TiO2 in environmental compartments.

Environmental concentrations

A total of 808 stream water samples were processed in the FOREGS-program to determine typical Ti stream water concentrations (Salminen et al. 2005). The FOREGS Geochemical Baseline Mapping Programme’s main objective is to provide high quality, multi-purpose homogeneous environmental geochemical baseline data for Europe. Stream water samples were filtered (< 0.45 µm) and Ti concentrations measured by ICP-MS (limit of quantification (LOQ): 0.01 µg/L). Based on the FOREGS dataset, dispersed/dissolved titanium concentrations in European stream waters may go up to 16.8 µg/L but are typically below 5 µg/L (95th P = 4.21 µg/L; EBRC et al. 2017). Titanium exists only in a fully hydrated form, TiO2 * n H2O, in water above pH 2, and is, therefore, transported in a colloidal state rather than as dissolved ion. This is further supported by thermodynamic stability data and speciation modelling, which suggests TiO2 * n H2O to consitute the dominating species within environmentally relevant pH values and within the ionic composition of stream waters as provided by the FOREGS dataset. Because of the propensity of TiO2 for hydrolysis, other titanium hydroxo complexes than TiO2 * n H2O are rare within a pH range of pH 4-10. Considering the low solubility of TiO2 as observed in T/D tests and confirmed by speciation modleling, it may be assumed that the stream water concentrations represent dispersed concentrations of TiO2 rather than dissolved concentrations.

In addition, Johnson et al. (2011) measured mean Ti concentrations in the fraction < 0.45 μm of raw sewage influent, sewage following primary settlement and in the final effluent of a sewage treatment plant serving over 200,000 people that decreased from 30.5 μg/L, to 26.7 μg/L, and to 3.2 μg/L, respectively, suggesting also that Ti is associated with particles or colloids.

Solubility - Conclusion

In sum, the micro- and nanoforms of TiO2 do not dissolve to any relevant extent under regular environmental conditions. Thus, dissolution is not important for the environmental fate and behavior of nano-and microsized TiO2 in environmental compartments.

Aggregation/agglomeration and sedimentation of microsized TiO2

Studies on the agglomeration/sedimentation of microsized TiO2 are lacking. However, titanium is relatively abundant in sediments (the 95th percentile in European sediments is 0.81 %), yet its concentration in freshwater is extremely low (95th percentile in European stream water is 4.21 µg dissolved/dispersed Ti/L, operational defined fraction < 0.45 µm) (Salminen, 2005). Even though titanium concentrations of sediments can be as high as 3 %, dissolved/dispersed Ti concentrations in the water column of European stream waters are well below 20 μg/L, indicating that sedimentation and presumably agglomeration remove TiO2 effectively from the water column.

 

Johnson et al. (2011) measured mean Ti concentrations in the fraction < 0.45 μm of raw sewage influent, sewage following primary settlement and in the final effluent of a sewage treatment plant serving over 200,000 people. They observed decreasing Ti concentrations of 30.5 μg/L, 26.7 μg/L and 3.2 μg/L from sewage influent to primary settlement to final effluent, respectively. These results suggest that TiO2 is effectively removed from the waste water and associated with sludge particles or colloids.

 

Aggregation/agglomeration and sedimentation of nanosized TiO2

Several studies indicate that agglomeration and subsequent sedimentation of nanosized TiO2 particles are of high relevance in the environment. Pronounced sedimentation was observed in a number of different waters: In natural and synthetic fresh- and saltwater as well as in Milli-Q water. Distinct sedimentation of 33.5-98.3% of the nominal applied nano-TiO2 concentration after relatively short settling periods ranging from 2 h to 50 h was observed (Brunelli et al. 2013, Doyle et al. 2014, Hildebrand et al. (2014), Li et al. 2017, Ottofuelling et al. 2011). Furthermore, a number of studies demonstrate that agglomeration of nano-TiO2 is influenced by several factors, including different water parameters, initial TiO2 concentration and surface properties of the TiO2 nanomaterials.

 

Doyle et al. (2014) observed a pronounced sedimentation of greater than 92% of the initial mass of three nano-sized TiO2 materials (UV-Titan M212, P25, Meliorum Technologies) in natural seawater after 72 h of mixing followed by a 45-min settling period.

 

Hildebrand et al. (2014) performed a sedimentation experiment and revealed that almost 90% of the TiO2 particles (P25, 28 nm) precipitated from the unstable TiO2 suspension (1 mM NaCl, pH 7) after 2 h, in which aggregated particles were characterized with a zeta potential of -10 mV and a hydrodynamic diameter of 1100 nm. They further investigated the influence of pH, fulvic acid, and ionic strength (mono-, di-, and trivalent cations) on surface charge, HDD and sedimentation of nano-TiO2 P25 particles. In TiO2 dispersions with 1 mM NaCl, the IEP of P25 particles was around 7. The zeta potential of TiO2 particles remained negative (-20 to -60 mV) in presence of fulvic acids (5, 10, 25 ppm) irrespective of the solution pH ranging from 3-10. Cations affected zeta potential and HDD of dispersed TiO2 particles. The HDD of TiO2 particles at pH 4 increased with rising cation (Na+, Ca2+ and Al3+) concentrations (250 nm at 0.001 mol/L to 1300-1650 nm at 0.100 mol/L). At pH 7, the HDD of TiO2 particles increased after the addition of Na+ (1100-2900 nm) and Ca2+ (2000-2400 nm) compared to dispersions without salts (250 nm). The HDD of TiO2 was, however, not affected by 0.001 and 0.010 mol/L Al3+ but increased at 0.100 mol/L (1400 nm). In the presence of 5 mg/L fulvic acid and cations (Na+, Ca2+ and Al3+), HDD of TiO2 particles were comparable to those observed in the absence of fulvic acid at pH 4 and pH 7.

 

Ottofuelling et al. (2011) performed colloidal stability tests with n-TiO2 (P25, primary particle size: 19.8 nm) in well-controlled synthetic waters (EPA very soft, moderately hard and very hard water), covering a wide range of water chemistries, and also in natural waters, including groundwater, lake, sea and tap water, peat bog water, and waste water inflow and outflow. In synthetic waters, sedimentation was influenced by the tested water parameters (pH, ionic strength, presence of relevant monovalent and divalent ions & presence of natural organic matter (NOM)), resulting in supernatant concentrations (< 0.2 µm) ranging from 5% to 30% of the nominal concentration after 15 h. Multi-dimensional testing on the effects of water chemistry on TiO2 aggregation revealed the dependency of TiO2 aggregation on IS and pH of the solution, the presence of monovalent and divalent ions, the presence of natural organic matter (NOM) as well as their interaction. In Milli-Q water, aggregation and subsequent sedimentation is strongly influenced by pH: At solution pH close to the isoelectric point (IEP), determined at pH 5 for the test material P25, TiO2 particles were present as large aggregates (> 2.5 µm in diameter). If the pH is below or above the IEP (tested pH range 3 to 10), aggregates were smaller (~ 300 nm).

Furthermore, Ottofuelling et al. (2011) investigated the effect of the following salts: NaCl, CaCl2, Na2SO4. At NaCl concentrations < 5 mM, the aggregation of TiO2 particles is still controlled by pH whereas concentrations > 5 mM result in an increased aggregation and a shift of the IEP to pH 6.5 at a maximum of 500 mM NaCl. The effect of 0.01 to 10 mmol/L CaCl2 (i.e. effect of divalent cations and monovalent anions) was tested. Below 0.1 mM CaCl2, TiO2 aggregation is controlled by pH whereas concentrations between 0.1 mM and 5 mM result in a stabilisation of TiO2 particles at pH ≤ 5, associated with a shift of the IEP to pH 7. However, an increased aggregation (aggregates > 1000 nm) is observed at concentrations > 5 mM CaCl2. The presence of 0.01 to 10 mmol/ Na2SO4 L (divalent anions) strongly promoted aggregation under all tested conditions. The combined effect of divalent cations and anions was tested with CaSO4 (at concentrations of 0.01 to 5 mmol/L) and an effective aggregation of TiO2 particles at CaSO4 > 0.1 mM (aggregates > 1000 nm) was observed. Only at concentrations < 0.1 mM and a pH < 5, particles aggregated less (diameters 500-1000 nm). Natural organic matter (NOM) at concentrations between 0.5 and 100 mg/L corresponding to DOC concentrations of 0.2 to 41.1 mg/L had a stabilizing effect on TiO2 particles independent of pH with HDD between 250 and 750 nm. However, an effective aggregation was observed in a solution with 1 mg DOC/L and CaCl2 at concentrations ≥ 1 mmol/L (40.0 mg Ca/L).

A pronounced aggregation and sedimentation of nanosized TiO2 particles (mean particle diameters: 750-1520 nm; remaining TiO2 in supernatant after 15 h: 4 - 32%) was further observed by Ottofuelling et al. (2011) in natural waters with typical water characteristics of lake, river and sea water. In contrast, almost no sedimentation and smaller aggregates were found in peat bog water (remaining TiO2 in supernatant: 96%; mean particle diameter: 219 nm), which deviates however substantially from typical compositions of natural waters with regard to its low ionic strength (IS) and pH, low Ca2+ and high dissolved organic carbon (DOC) concentration.

 

Li et al. (2016) studied the influence of different natural water properties on the aggregation and deposition of TiO2 nanoparticles (P25; primary particle size: 21 nm) while investigating the behaviour of the nanoparticles in 39 different natural water samples. The deposition rate constant for nano-TiO2 particles was highest in brackish water (54.3 x 10-5/min), which was in accordance with the low zeta potential, large aggregates and high aggregation rate constant observed in these samples. Deposition rate constants were low in Humus-poor lake (-3 x 10-5/min to 5.2 x 10-5/min) and humic lake samples (-10.1 x 10-5 to 12.0 x 10-5/min) corresponding to a more negative zeta potential, a smaller hydrodynamic diameter (HDD) and a slow aggregation rate of nano-TiO2 particles. Deposition rate constants in nutrient-rich lake samples were more variable (0.16 x 10-5/min to 4.9 x 10-5/min). Several water parameters, including alkalinity, pH, electrical conductivity, total nitrogen, total phosphorus and turbidity, were significantly positively correlated with different stability parameters.

 

Li et al. (2017) investigated the stability, aggregation and sedimentation of the nano-sized TiO2 material Aeroxide P25 (primary particle size: 21 nm) in Milli-Q water as well as in 6 freshwater samples of different origin. As zeta potential and HDD of P25 nanoparticles in different natural freshwater samples were only measured in the presence of surfactants (sodium dodecyl sulfate or Tergitol NP-9), a comparison to the stability and aggregation of P25 in Milli-Q water cannot be made (Zeta potential: 14.72 mV; HDD: 1250 nm in Milli-Q water). An increased sedimentation was observed in all natural water samples as 11-19 % n-TiO2 remained in solution after 24 h compared to 44.6 % in Milli-Q water.

 

Nur et al. (2015) studied the effect of the composition of different ecotoxicity test media on agglomeration and stability of nano-sized rutile TiO2 material NM 104 (particle size: 36 ± 14 nm). The HDD ranged from 1024 to 1792 nm, and critical coagulation concentrations ranged from 17.6 to 54.0 % v/v for the different test media diluted in ultra-pure water.

 

Keller et al. (2010) investigated the effect of properties of 8 different natural waters on aggregation and deposition of TiO2 nanoparticles (Evonik, 27 ± 4 nm), including seawater, ground water, lake water and wastewater collected in California (US), artificial seawater and mesocosm freshwater medium. Aggregation and deposition of TiO2 nanoparticles varied distinctly in different natural and artificial waters. Aggregation and deposition was fast in waters with a high ionic strength (IS) and low/medium total organic carbon (TOC), in particular in all seawater samples, depending further strongly on the TiO2 concentration (with aggregate sizes ~ 1 µm at 10 mg TiO2/L, and up to 2 µm at 50 and 200 mg TiO2/L). Smaller, more stable aggregates (~ 300 nm) and a slow sedimentation that remained independent from initial TiO2 concentrations were observed in waters with a low IS and medium/high TOC content (storm water, treated WWTP effluent, mesocosm freshwater).

 

Lu et al. (2015) investigated the influence of pH, ionic strength and NOM on sedimentation, zeta potential and HDD of TiO2 nanoparticles (particle size: 21 nm) in ultrapure water for 120 min. Sedimentation increased at a solution pH around the point of zero charge (PZC) of the respective TiO2 nanoparticle (pH 5.2) and with increasing ionic strength (3-100 mM NaCl). By contrast, increasing humic acid (HA) concentrations (0, 1, 5 mg/L) stabilized TiO2 nanoparticles in ultrapure water.

 

Von der Kammer et al.(2010) observed that two nanosized TiO2 materials (P25 and Hombikat UV-100) behaved similarly in presence of NOM and CaCl2. All tested NOM concentrations (1-10 mg/L) enhanced the stabilisation of TiO2 nanoparticles. The dispersion of TiO2 nanoparticles was stable at elevated CaCl2 concentrations (around 1 x 10-2mol CaCl2/L at pH 5 to 7) but not at intermediate (environmentally relevant) CaCl2 concentrations (10-5to 10-3mol/L) and a solution pH around the IEP.

 

Von der Kammer & Hofmann (2015) tested different TiO2 nanoparticles (P25, Hombikat UV100, Titan M212, Titan M262) at varying pH, ionic strength and NOM concentrations. Whereas test materials behaved slightly differently under NOM addition, an overall stabilizing effect of increasing NOM concentrations was observed starting at 0.1 mg/L DOC. Increased NaCl concentrations promoted the aggregation of P25 TiO2 particles, however, the other test materials behaved differently and remained more dependent on pH. In the presence of CaCl2 (divalent cations and monovalent counter ions), P25, Titan M212 and Titan M262 were stabilized especially at higher CaCl2 concentrations whereas the stability of Hombikat UV100 was slightly reduced. Hombikat UV100, Titan M212, Titan M262 and Evonik P25 behaved differently after the addition of NaSO4 (monovalent cations and divalent anions). For Hombikat UV100 an extended stability under high sulfate concentrations and low pH can be observed whereas the aggregation of P25, Titan M212 and Titan M262 increased at higher NaSO4 concentrations.

 

Thio et al. (2011) studied the influence of pH, HA, and ionic strength (NaCl, CaCl2) on the surface charge and HDD of nano-TiO2 particles (30 nm). The PZC in the absence of electrolytes and HA was determined at a pH of ~ 5.5. TiO2 nanoparticles were found to aggregate immediately upon their introduction to water, with the smallest aggregates slightly above 250 nm in HHD. In the presence of NaCl, the zeta potential of TiO2 nanoparticles became less positive or more negative with an increase of the IS (10, 100 mM). The critical coagulation concentration (CCC) was determined at ca. 15 mM in the absence of HA and 200 mM in the presence of HA (10 mg/L). At 1 mM CaCl2, the zeta potential of TiO2 particles was either less positive (at pH < PZC) or less negative (at pH > PZC), thus closer to 0 mV, compared to TiO2 nanoparticles in the absence of electrolytes. At 10 and 100 mM CaCl2, Ca2+ determined the surface charge of the particles, and the zeta potential of particles ranged from +18 mV to +25 mV in the pH range of 4 - 9. The CCC for CaCl2 was determined at ca. 0.1 mM in the absence of HA and 5 mM in the presence of 10 mg/L HA. At 10 mg/L HA, the zeta potential of TiO2 NPs remains negative in the pH range of 4 - 9 regardless of the IS and the type of electrolyte. Higher IS leads to a decrease in the magnitude of the zeta potentials, which is more pronounced with CaCl2 than with NaCl. Based on the IS of natural waters and not taking NOM into consideration, TiO2 would be predicted to undergo fast aggregation. The IS plays a role, but only in highly saline environments. When humic substances are present in natural waters, TiO2 particles are stabilized. Thus, NOM plays an important role in determining the stability and mobility of TiO2 NPs in natural aqueous matrices.

 

Wang et al. (2014) measured HDDs and zeta potentials of a nanosized TiO2 material (particle size: 21 nm) in reconstituted seawater for up to 240 hours. The aggregation of TiO2 nanoparticles accelerated with increasing salinity (1, 5, 30 ppt) whereas increasing DOC concentrations (0.25-10.5 mg/L) stabilized particle dispersions over time.

 

Not only the DOC concentration but also the source of HA and the size fraction of NOM seem to influence adsorption and stabilisation of nanosized TiO2 dispersions. The influence of different sources of HA and the inorganic background solution on the adsorption of HA to nanosized TiO2 material P25 (21 nm), and on the sedimentation, aggregation and stability was investigated by Erhayem et al. 2014. Adsorption constants of HA were significantly higher in NaCl solution than in phosphate buffer since phosphate may compete with HA on the surface of TiO2. Aromaticity of the different tested HAs influenced the respective adsorption to nano-TiO2, resulting in Kads values ranging from 2.54 to 10.3 with higher Kads values for soil than for estuarine HA. The concentrations of HA affected sedimentation, i.e. nano-TiO2 particles destabilised at low concentrations (10 mg/L) but stabilised at higher HA concentrations (≥ 25 mg/L). Furthermore, poly-condensed structures of HA preferentially adsorbed to nano-TiO2. In sum, structural differences of sedimentary and soil HA, as well as the HA concentration affect the stability and fate of nano-TiO2.

 

Mwaanga et al. (2014) studied the sorption of natural organic matter (NOM) to nanosized titanium dioxide P25 (particle size: <50 nm) in media differing in pH and ionic strengths. Results indicate that larger-sized fractions of NOM preferentially adsorb to TiO2 irrespective of NOM concentrations. Further, sorption of the larger-size fractions of NOM increased with decreasing pH and NOM sorption increased with increasing IS at a respective pH based on two independent analyses (absorbance and fluorescence spectrometry).

 

Adeleye et al. (2016) investigated the effect of soluble extracellular polymeric substances (sEPS) produced by freshwater (C. reinhardtii) and marine algae (D. tertiolecta) and Suwannee River natural organic matter (SRNOM) on surface properties and fate of three TiO2 nanoparticles, including Hombikat UV 100 (3.3 nm, anatase, uncoated), UV-Titan M212 (23 nm, rutile, hydrophilic alumina and glycerol coating) and UV-Titan M262 (24 nm, rutile, hydrophobic alumina and dimethicone polymer coating). Measurement of zeta potentials and CCC of nano-TiO2 particles in presence of different concentration of sEPS (0-2 mg-C/L) and SRNOM (0.25 mg-C/L) revealed that i) sEPS and SRNOM can adsorb to nano-TiO2, ii) sEPSand SRNOM may cause a charge reversal of the nano-TiO2 particles and thereby form negatively charged surface complexes with nano-TiO2 in aqueous media, iii) sEPSand SRNOM can increase the stability of nano-TiO2 by steric and electrostatic stabilisation in aqueous media. Furthermore, FTIR analyses may suggest that a chemical bond between the carboxylic group of sEPS and the surface of nano-TiO2 can form, which is considered to be long-lived. Hence, the fate of nano-TiO2 in complex environmental systems may not be accurately predicted based on the intrinsic physiochemical properties of the particles alone.

 

Brunelli et al. (2013 and 2016) identified the initial loading of n-TiO2 as a further parameter that affects agglomeration and sedimentation of nanosized TiO2. Apparently, the so-called settling rate constants of nanosized TiO2 (P25, 21 nm) in artificial freshwater (ISO 6341-1982) increased from 1 x 10-4 s-1 to 2 x 10-4 s-1 with an increase of initial TiO2 concentration from 15 to 100 mg/L but remained constant at 5-15 mg TiO2/L (9.6 – 1.0 x 10-5 s-1) (Brunelli et al. 2016). Furthermore, Brunelli et al. (2013) investigated the influence of salinity (0-35‰), ionic composition and strength, pH and DOC on agglomeration and sedimentation of nanosized TiO2 (P25, 21 nm) in artificial and natural fresh- and seawater, including also estuarine and lagoon water. The tested n-TiO2 underwent a fast agglomeration as soon as it was dispersed in both synthetic and real aqueous dispersions while sedimentation was slower. Sedimentation rates ranged from 3 x 10-6 to 2 x 10-5 s-1 and increased with increasing initial TiO2 concentration (0.01, 0.1, 1, and 10 mg/L; nominal) in all media. Settled n-TiO2 after 50 h varied from 33.5 - 52.2%, from 53 - 79% and from 94.5 - 98.3% in the 0.01/0.1 mg/L, 1 mg/L and 10 mg/L treatment groups of the different test media, indicating that the difference between the lowest and highest treatment group was greater than the variation between the different test media, i.e. fresh- and saltwater. Thus, sedimentation and agglomeration in seawater depend more on the initial loading of n-TiO2 than on ionic strength, salt and DOC content under the tested conditions.

 

The presence of a coating on nano-TiO2 particles may also influence the agglomeration through electrostatic or steric stabilisation. Two studies investigated if coatings degrade under environmentally relevant conditions or during the preparation of dispersions for standard ecotoxicological testing by ultrasonication:

 

Labille et al. (2010) examined ageing of coated TiO2 nanomaterial T-lite SF (TiO2 core with Al(OH)3 and PDMS coating) in ultrapure water under dark and simulated sunlight conditions for 7 days. Results indicate that ageing (simulated by magnetic stirring at 690 rpm) leads to a release of 90% and 30% of the PDMS coating under dark and simulated sunlight conditions, respectively, whereas the Al(OH)3 coating remained on the TiO2 core. A positive zeta potential was measured after 3 h and 18 h of ageing at sunlight and under dark conditions, pointing to a rather quick surface charge reversal. Since solution pH did not change, the positive charge is most likely due to the exposition of the Al(OH)3 layer beneath the PDMS coating known to have its isoelectric point (IEP) at pH 7 - 8 and a positive surface charge at pH 6.5 (test conditions). Whereas the PDMS coating appears to be unstable, Al(OH)3 remains inert under the tested aqueous conditions.

 

Nickel et al. (2013) sonicated dispersions of coated TiO2 materials NM 103 (20 nm, rutile, PDMS and Al2O3 coating) and NM 104 (20 nm, rutile, glycerol and Al2O3 coating) in deionised water. Sonication caused the release of 86 ± 12% and 88 ± 8% of glycerol and PDMS, whereas the Al2O3 shell was shown to remain on the TiO2 core of NM 103 and NM 104. The IEP of the TiO2 materials in dispersion, i.e. NM 104: pH 8 -9, NM 103: 8 -9.5, were in the range of the IEP of Al2O3 (pH 8 -9) but not of the TiO2 core (pH 5.5 - 7). Thus, the degradation of coatings alters the surface characteristics of TiO2 materials, resulting e.g. in a modified IEP, which further affects agglomeration.

 

Conclusions – aggregation/sedimentation

In sum, TiO2 nanoparticles aggregate in most of the tested aqueous media with aggregates being > 100 nm, often ≥ 1 µm. Thus, nanosized TiO2 particles (< 100 nm) are not expected to be present for a long time after the release in the aquatic environment as only aggregated, microsized TiO2 are found in natural waters. Since sedimentation of TiO2 aggregates was observed under a variety of conditions, the importance of sediments as receiving environmental compartment for TiO2 nanomaterials was confirmed. Results further indicate that the aggregation of n-TiO2 depends on several water parameters, such as ionic strength, pH, presence of relevant ions, concentration and source of NOM, turbidity, alkalinity, but also on initial TiO2 concentrations and the surface properties of the TiO2 material. Furthermore, it becomes apparent that the interaction between these factors is crucial in determining the fate of TiO2 nanoparticles in aqueous media.

In the majority of studies, aggregation is high near the IEP, and a solution pH above or below the IEP promotes stabilisation (Hildebrandt et al. 2014; Lu et al. 2015; Nur et al. 2015; Ottofuelling et al. 2011; Von der Kammer et al. 2010, 2015). This parameter is however influenced by a variety of different factors: First of all, the characteristics and surface properties of the TiO2 nanoparticles and the applied coatings determine the surface charge and IEP. Furthermore, the presence of environmentally relevant anions and cations, their identity and the associated ionic strength can shift the IEP, thereby affecting the aggregation of TiO2 particles. In conclusion, pH affects aggregation and sedimentation of TiO2 nanoparticles, but so do other factors and the behaviour of TiO2 nanoparticles in natural waters cannot be accurately predicted solely from pH.

Presence of mono- and divalent cations and anions significantly affects the aggregation of nano-TiO2. A destabilizing effect on TiO2 nanoparticles with increasing IS was observed in most of the studies (Doyle et al. 2014; Hildebrandt et al. 2014; Keller et al. 2010; Li et al. 2016; Li et al. 2017; Lu et al. 2015; Nur et al. 2015; Ottofuelling et al. 2011; Thio et al. 2011; Von der Kammer et al. 2010, 2015). Monovalent cations and counter ions such as NaCl also affect the stability of TiO2 nanoparticles. For the test material P25, destabilizing effects were observed in the range from 0.1 to 100 mM NaCl. A destabilizing effect was observed at concentrations > 5 mM by Ottofuelling et al. (2011) whereas other studies report an effect at lower concentrations (0.1 mM, Von der Kammer et al. 2010; 1.0 mM, Von der Kammer et al. 2015 & Hildebrand et al. 2014). Effects were not observed up to 100 mM NaCl with Titan M212 and M262 (Von der Kammer et al. 2015), whereas results of Hombikat UV-100 were contradictory (Von der Kammer et al. 2010 & Von der Kammer et al. 2015). In sum, a slight destabilizing effect due to Na and Cl depending on the TiO2 material and pH might be expected at environmentally relevant concentrations of NaCl (5th-95thpercentile of EU freshwaters: Cl: 0.01-2.04 mM, Na: 0.04-2.10 mM).

Available data for divalent cations such as Ca2+are contradictory: Hildebrand et al. (2014) observed a destabilisation of P25 at 1.0 -100 mM Ca2+whereas Von der Kammer et al. (2015) reports stabilisation of P25, UV-Titan M212 and UV-Titan M262 nanoparticles but not of Hombikat UV-100 at 1.0-10.0 mM Ca2+. Ottofuelling et al. (2011) observed stabilisation of P25 at lower concentrations (0.05-5.0 mM Ca2+) and destabilisation > 5 mM. Thus, based on available data, the aggregation of TiO2 nanoparticles at environmentally relevant calcium concentrations (5th-95thpercentile of EU freshwaters: 0.04-3.70 mM Ca2+) appears to also depend on other factors including properties of the TiO2 material itself.

Increasing concentrations of divalent anions such as sulfate result consistently in destabilisation of TiO2 dispersions, including P25, UV-Titan M212 and UV-Titan M262 but not Hombikat UV-100. Tested concentrations ranging from 0.01 to 10.0 mM (Ottofuelling et al. 2011; Von der Kammer et al. 2010; Von der Kammer et al. 2015) imply that at concentrations present in EU freshwaters (5th-95thpercentile of EU freshwaters: 0.01-1.75 mM SO42-), aggregation of the TiO2 nanoparticles except of Hombikat UV-100 is promoted. The latter may be explained with the low IEP (> pH 4) and the observation that dispersions of TiO2 nanoparticles tend to be least stable at a pH close to the respective IEP.

Regarding the effect of NOM or DOC, available studies point to a stabilisation of TiO2 nanoparticles by DOC/NOM. Tested concentrations range from 0.1 to 41.41 mg DOC/L or 1-10 mg NOM/L implying that dispersions of TiO2 nanoparticles are fairly stable at environmentally relevant concentrations (5th-95thpercentile of EU freshwaters: 0.61-23.08 mg/L DOC).

In conclusion, the aggregation and subsequent sedimentation of TiO2 nanoparticles depends on several interacting factors, including the initial TiO2 loading, the IEP of TiO2 materials, pH, hardness, ionic strength, presence of monovalent or divalent cations and anions and organic matter. Whereas several studies in a selection of representative natural waters, including groundwater, lake water, seawater, tap water, and waste water inflow and outflow, report substantial aggregation and pronounced sedimentation (Brunelli et al. 2013, Doyle et al. 2014, Li et al. 2016, Li et al. 2017, Lv et al. 2016, Ottofuelling et al. 2011, Schirmer 2014), resulting in the removal of TiO2 particles from the water column and subsequent delivery to the sediment phase, results of some studies indicate that under certain conditions, in particular in waters with higher organic matter concentrations and a low IS, such as bog water or mesocosm freshwater, smaller aggregates (~ 300 nm) remain stable in the aqueous phase and sedimentation is slow (Keller et al. 2010; Ottofuelling et al. 2011). Overall, findings of available studies indicate that sediments are the major receiving environmental compartment in the aquatic environment butTiO2 nanoparticle aggregates might be present in the water column under certain conditions.

Further multi-dimensional testing would be required to make definitive predictions on nanoparticle behaviour.

 

Solid-water partitioning

Solid-solution partitioning of TiO2 is a key property that determines its environmental fate in terrestrial and aquatic systems. The partition coefficient can be used to provide information on potential exposure pathways and which biota are likely to be most relevant once TiO2 is released into the environment. Nano- and microsized TiO2 is poorly soluble and sufficiently stable to not transform to a water-soluble form. Thus, TiO2 itself (and not a soluble form) should be assessed taking into account the particle-specific partitioning characteristics. Further, according to ECHA’s guidance on IR and CSA Appendix R.7.13-2 (July 2008), “the distribution of metals over the solid and liquid phase is not only controlled by pure adsorption/desorption mechanisms. Other processes like precipitation or encapsulation in the mineral fraction also play a role.”

 

Sediment – water partitioning

The assessment of the partitioning of TiO2 in environmental media is based on Kp values derived from monitoring data for elemental Ti concentrations in water and corresponding sediments provided by the FOREGS Geochemical Baseline Mapping Programme that aimed to provide high quality, multi-purpose homogeneous environmental geochemical baseline data for Europe. A total of 757 paired samples, i.e. samples with the same coordinates for the sampling location of stream water (filtered to < 0.45 µm) and sediment (wet sieved in the field to <0.15 mm) were processed (Salminen et al. 2005) and results correspond to steady-state conditions of Ti, independent of Ti speciation.

Sampled stream water and sediments cover a wide range of environmental conditions. Water parameters such as pH, hardness and organic carbon concentrations cover several magnitudes. Dissolved/dispersed titanium water levels range from < 0.10 to 16.80 µg Ti/L with 5th and 95th percentiles of 0.10 and 4.21 µg Ti/L, respectively. The FOREGS dataset reports titanium dioxide concentrations of the sediment. Respective TiO2-data were converted into titanium concentrations to enable comparison of sediment with dissolved/dispersed concentrations in the stream water. Sediment concentrations of Ti range from 0.01 to 2.99 % Ti with 5th and 95th percentiles of 0.14 and 0.81 % Ti, respectively. Taking into account the high quality and representativeness of the data set, the 95th percentile of 4.21 µg Ti/L can be regarded as typical background concentration for dissolved/dispersed titanium in European surface waters and the 95th percentile of 0.81 % Ti as typical background concentration of European stream sediments.

Regarding the partitioning of titanium in the water column, stream water/sediment partition coefficients range from 182,527 L/kg to 119,329,191 L/kg. Since FOREGS sampled on a grid aiming to equally represent geochemical baseline concentrations across Europe, a European median log Kd value of 6.57 is derived for sediment-water partitioning.

Roychoudhury and Starke et al. (2006) measured total and dissolved Ti concentrations of surface water and sediments from 20 sites of the stream Blesbokspruit (South Africa), and the mean log Kd for sediment is 4.61 L/kg dw (range: 3.79-5.13).

Additionally, Boncagni et al. (2009) studied the transport of TiO2 nanoparticles between stream water and stream beds in recirculating flume experiments: Fast deposition was observed for P25 nanoparticles at pH 6.0, with almost all TiO2 particles deposited after 24 h, while a stable in-stream concentration around 60% was observed at pH 10.0. When the pH was increased from 6.0 to 11.3 at the end of the first experiment, a small percentage of nanoparticles (~5%) was released, which was further enhanced when the stream flow rate was doubled (~ 12%).

Conclusion - Sediment – water partitioning

Even though Ti concentrations of sediments can be as high as 3 %, dissolved/dispersed Ti concentrations in the water column of European stream waters are well below 20 µg/L. These monitoring stream water and sediment data provide strong evidence that Ti is sparingly soluble in a wide representative range of European stream waters and is mostly associated with particles or colloids of the sediment.

 

Suspended matter – water partitioning

Vesley et al. (2001) analysed titanium concentrations of water and particulate matter from 54 rivers at 119 sites (Czech Republic), and the median log Kd for suspended matter amounts to 5.36 L/kg dw indicating that titanium in the water column is mostly associated with suspended matter.

Soil – water partitioning

In a standardized soil adsorption test according to OECD 106, Kuhlbusch et al. (2012) measured the adsorption of nanosized TiO2 (P25, 21 nm and UV Titan M262, 20 nm, coated). Adsorption was measured at one single concentration (2.5 mg TiO2, 25 mL, 5 g soil) in six different soils varying in texture (from loamy sand to silt loam), and pH from 4.78 to 6.78 (strongly acid to very-sub acid), effective CEC from 37.9 to 236 mmolc/kg and OC content from 0.93 to 3.85 % (very light humic to strongly humic). Centrifugation (2700 g, 10 min) removed soil particles unadsorbed TiO2 agglomerates larger than 177 nm so that a differentiation of adsorbed and unadsorbed TiO2 particles is not possible. Titanium concentration were below 1% and 5 % of the added material in the supernatant after centrifugation of soils mixed with UV Titan M262 and P25, respectively. Based on the solid-solution partitioning of n-TiO2, conservative coefficients of Kd < 495 L/Kg (UV Titan M262) and Kd < 95 L/Kg (P25) can be estimated. Due to the limited test design, the partition coefficients should not be applied without further confirmatory experiments.

 

Kuhlbusch et al. (2012) applied dispersions of the nano-sized TiO2 materials P25 (21 nm), PC 105 (15-25 nm) and UV Titan M262 (20 nm) to columns of three test soils varying in pH, texture and cationic exchange capacity (CEC) according to OECD 312 soil leaching tests. After 48 h of artificial rain onto the soil columns, the Ti concentration in the eluates were below the LOD (< 5 µg/L), indicating that less than 0.04% of the applied TiO2 leached out of the columns. Except for the TiO2 material UV Titan M262, the highest TiO2 concentrations were measured in the first segment (0 - 1 cm) of the soil columns and Ti concentrations did not rise in subsequent segments. Apparently, up to 14 % and 19 % of UV Titan M262 were transported to the second (3 – 4 cm) and last segment (29 - 30 cm), respectively.

 

Nickel et al. (2013) applied dispersions of nanosized TiO2 (P25, 21 nm) to columns of three test soils varying in pH, texture and CEC according to OECD 312 soil leaching tests. The Ti concentrations of eluates were below the LOD (< 0.12 µg/L), and Ti concentration did not increase in all subsequent soil segments.

 

Gogos et al. (2016) performed a non-standardized pot test with red clover and wheat, which were exposed to natural soils blended with quartz sand and spiked with food grade E171 TiO2 powder (primary particle size: 92 ± 31 nm) or with the titanium dioxide nanoparticle powder P25 (primary particle size: 29 ± 9 nm). Soil cores sampled after 12 weeks for wheat experiments showed no statistically significant differences in Ti concentrations for the different soil depths (0-5, 5-10, 10-15 cm). After watering the pots to 110% water holding capacity, leachate analysis at week 11 (wheat) showed a significant increase of the Ti concentrations for the highest E171 concentration tested (1000 mg TiO2/kg soil) compared to the control. However, it has to be considered that only 0.0001% of the applied Ti was found in the leachate. Significantly different Ti leachate concentrations were not observed between P25 treatments and controls.

Conclusion - soil – water partitioning

Results of soil leaching studies indicate a low leachability and mobility of nanosized TiO2 materials in soil and provide further evidence for the adsorption/retention in soil. A similar lack of mobility is indicated for macrosized TiO2 by the study of Gogos et al. (2016) and thus may be assumed for microsized material based on common physico-chemical properties including the poor solubility. 

 

Sewage sludge – waste water partitioning

Johnson et al. (2011) studied the fate of TiO2 particles in a sewage treatment plant (STP) serving over 200,000 people. Apparently, concentrations of Ti in the fraction < 0.45 µm differ significantly between influent and effluent. Whereas the TiO2 particles do not undergo removal during the primary settlement stage, the greatest decline (ca. 90%) followed the biological stage of activated sludge with mean Ti concentrations in the fraction < 0.45μm in raw sewage influent, following primary settlement and in final effluent of 30.5μg/L; 26.7μg/L; and 3.2μg/L, respectively.

Westerhoff et al. (2001) determined titanium concentrations in raw sewage and treated effluent of 10 wastewater treatment plants (WWTP) and showed that over 96 % of influent titanium was removed, with all WWTPs having effluent titanium concentrations of less than 25 µg/L. Furthermore, the presence of spherical titanium oxide nanoparticles (crystalline and amorphous) with diameters of 4 to 30 nm in WWTP effluents was indicated by TEM and EDS analysis.

 

Three sewage sludge samples selected from the USEPA Targeted National Sewage Sludge Survey (TNSSS) by Kim et al. (2012) showed Ti concentration ranging from 96.9 to 4510 mg/kg (on a dry weight basis). Nano- and larger particles of TiO2 (40-300 nm) were repeatedly identified by electron microscopic analysis across the sewage sludge types tested, having faceted shapes with the rutile crystal structure, and typically forming small, loosely packed aggregates.

Results from a lab-scale sequencing reactor indicate that biological STPs that operate with suspended biomass such as activated sludge have the potential to remove 97 ± 1% of n-TiO2 from wastewater (Wang et al. 2012). These results are confirmed by Kuhlbusch et al. (2012) who investigated the fate of nano-sized TiO2 (P25, 21 nm) stabilized with sodium metahexaphosphate in a laboratory STP according to OECD TG 303 A. Apparently, 4 % of the initially applied TiO2 were recovered in the overflow after 22 d and consisted mainly of particles < 0.6 µm. However, these results have to be considered carefully since the influence of the stabilizing agent on the fate of P25 in the laboratory STP was not clarified.

Pronounced sedimentation and adsorption to the solid phase of sewage was also observed in a non-standardized study by Barton et al. (2015). Mean distribution coefficients of the solid and liquid phase of primary and secondary wastewater sludge at steady state of nanosized TiO2 (P25, 21 nm) after 1 h of mixing and 30 min of sedimentation amounted to 8800 L/kg for primary and 9100 L/kg for secondary sludge. These results provide further evidence that the majority of the applied nano-TiO2 is removed from the water phase of wastewater by sedimentation within 30 min. Thus, the mean sludge – waste water distribution coefficient of 8950 L/kg was applied in the CSA.

 

Conclusion - Sewage sludge – waste water partitioning

Studies on the fate of TiO2 particles in wastewater provide evidence that up to 97% are removed during sewage treatment. Thus, the emission of TiO2 materials into the aquatic system via the effluent of STPs is low since, for example, measured Ti concentrations of the effluent of an operating STP amount to 3.2 µg Ti/L despite higher Ti concentrations of the respective influent.

Potential entry of uncoated TiO2 materials from paints into the aqueous environment

Kaegi et al. (2008) investigated the release of TiO2 from new and aged outdoor paints containing (not further characterised) pigment-grade but not nanosized TiO2 particles. The particle size of 85 - 90% of the total Ti in facade-runoff samples ranged from 20 to 300 nm without further size separation. Thus, a quantification of TiO2 particles below 100 nm is not possible.

Electron microscopy identified nanoparticles (< 100 nm) in the runoff of new and aged facades as well as in the urban runoff. These particles were identified as TiO2 by TEM-EDX analysis. All the measured Ti was traced back to the painted facade, however, there is not any evidence that the TiO2 nanoparticles were actually released from the facade. Background concentrations were not measured, and a background correction was not performed. Information on titanium particles in the rainwater was not provided. Rainwater measurements are relevant to exclude it as a potential source of particles. Copper particles were detected in all runoff samples but not in the new facade paint. It is thus unlikely that copper was released from the facade. Due to the sample preparation (centrifugation) and the density of relevant copper substances it may be assumed that copper particles are in the same size range or smaller than TiO2 particles. However, the occurrence of copper particles was not mentioned or discussed by the authors.

Kaegi et al. (2008) tried to estimate nanoparticle concentrations in urban runoff based on the particles ranging from 20 – 300 nm. Furthermore, it was assumed that 10% of the particles are smaller than 100 nm and a particle number for the fraction < 100 nm was estimated. This assumption is, however, not supported by the provided information on the particle size distribution (lognorm (150, 1.6) A PSD with such a small standard deviation would not contain particles below 100 nm.

Incidental TiO2 nanoparticles were observed by Kaegi et al. (2008) in the runoffs but the concentration, number or origin of these TiO2 nanoparticles (<100 nm) was not further clarified. Therefore, the study by Kaegi et al. (2008) is not considered reliable.

Furthermore, a comparison of the quantity of TiO2 from the new facade compared to the aged facade is difficult, as size of the aged facade as well as the water volume are unknown. A quantitative conclusion on the release of TiO2 particles from an outdoor facade is not possible.

Al-Kattan et al. (2013) exposed panels covered with paint with and without n-TiO2 to simulated weathering by sunlight and rain in climate chambers. The same paints were also studied in small-scale leaching tests to elucidate the influence of various parameters on the release such as composition of water, type of support and UV-light. During the climate chamber experiments, representing accelerated weathering compared to natural conditions, both paints released very low amounts of Ti into the leachate. Even the first leachate sample contained only 1.5 µg/L Ti, a concentration that is only slightly above the background Ti concentration in the leaching water. The total release of Ti over the 113 weathering cycles was 0.007% of the total Ti, indicating that TiO2 was strongly bound in the paint matrix. The size fractionation (< 100 nm) of Ti released from the paint was more or less constant throughout the test (about 0.5 µg/L).

In the leaching tests, concentrations in the leachate ranged from 0.5 to 14 µg/L (0.5 – 13.5 µg/L for the size fraction < 100 nm), with the highest concentrations observed after prolonged UV-exposure. Considering the volume of the leachate water and the weight of the exposed paint, measured Ti in the first leachate in the climate chamber experiment amounts to 0.22 Ti µg/g of paint. Accordingly, in the small-scale leaching experiments the released Ti amounts to 0.05–1.3 µg/g. Therefore, similar release rates were observed in the two types of experiments. It can be concluded that paints containing n-TiO2 may release a very limited amount of particles (including particles < 100nm) into the environment, at least over the timespan investigated in this work.

Conclusion - potential entry of uncoated TiO2 materials from paints into the aqueous environment

Uncoated TiO2 nanomaterials do not enter directly the aquatic system via run-off from facades at ecotoxicologically relevant concentrations.