The Effects of Oxyanion Adsorption on Reactive Oxygen Species Generation by Titanium Dioxide

Abstract

The growing use of nano titanium dioxide (TiO₂) in consumer and agricultural products has accelerated its introduction into terrestrial environments, where its impact has not been documented extensively. TiO₂ toxicity arises primarily from its ability to photochemically generate reactive oxygen species (ROS), including hydrogen peroxide (H₂O₂). While common ligands in soil porewaters can either hinder or enhance the degradation of organic contaminants by TiO₂, their effects on ROS production by TiO₂ have not been understood clearly. The objective of this study was to assess the effect of phosphate (P) and nitrate on UV-irradiated anatase, nano-TiO₂. Accordingly, H₂O₂-generation kinetics experiments were conducted in UV-irradiated TiO₂ under environmentally relevant concentrations of the ligands (0, 50, 100, and 250 μM) and pH values (4.00 ± 0.02 and 8.00 ± 0.02) from 0–100 min. Under all conditions, H₂O₂ grew logarithmically and reached between 5.38 and 22.98 μM after 100 min. At pH 4.00 ± 0.02, H₂O₂ production was suppressed by P but not by nitrate. Conversely, at pH 8.00 ± 0.02, nitrate did not affect H₂O₂ concentration while P increased it. Non-specific, minimal adsorption of nitrate prevented interference with the photoreactivity of TiO₂. Due to the pH-dependent behavior of suspended TiO₂ and H₂O₂ degradation rates, specific adsorption of P on TiO₂ impeded its ability to produce H₂O₂ photochemically at pH 4.00 ± 0.02 but amplified it at pH 8.00 ± 0.02.

Introduction

Engineered nanomaterials (ENM) are utilized in various consumer products, including cosmetics, plastics, fuel additives, textiles, paints, coatings, dietary supplements, medical applications, etc. (Keller & Lazareva 2013; Manke et al. 2013; Sajid et al. 2015). During their production, use, and disposal each year, ~51,600 metric tons of ENMs are released into the soil globally (Keller & Lazareva 2013). Within terrestrial environments, ENMs have both beneficial and detrimental impacts (Gardea-Torresdey et al. 2014; Sajid et al. 2015). The positive effects of ENMs primarily involve increased yields of various crops and enhanced degradation of toxic chemicals (Ravichandran et al. 2010; Gardea-Torresdey et al. 2014). Conversely, the risks of ENMs include cyto- and genotoxicity to humans, animals, and plants as well as decreased crop biomass (Ghosh et al. 2010; Manke et al. 2013; Fu et al. 2014; Gardea-Torresdey et al. 2014; Sajid et al. 2015). These impacts vary widely depending on the type and concentration of the ENM and the type of organism exposed (Manke et al. 2013).

Titanium dioxide is one of the most common ENMs and is frequently released into the environment through cosmetic, agricultural, medicinal, and food-additive applications (Weir et al. 2012; Fu et al. 2014; Gondikas et al. 2014; Sajid et al. 2015). Unlike naturally occurring TiO₂ particles, TiO₂ environmental nanoparticles (ENPs) are manufactured to be homogeneous in size, shape, and structure. They also have smaller diameters and greater reactive surface area than natural TiO₂ (Bernhardt et al. 2010). The homogeneity, purity, and large reactive surface area make engineered TiO₂ NPs more reactive. Smaller particles become more photoreactive because of an increase in band-gap energies (Kavan et al. 1993). Their release into soil and water environments, in particular, is extensive because they are utilized in various agricultural applications, including soil remediation nano-technologies and nano-enabled pesticides, fertilizers, soil additives, and growth regulators (Gardea-Torresdey et al. 2014). Additionally, TiO₂ is added to the soil through the application of biosolids, which tend to have high concentrations of TiO₂ (10–70 mg kg^–1) (Keller & Lazareva 2013; Gardea-Torresdey et al. 2014). Consequently, soil TiO₂ concentrations can be significant. For instance, an exposure model predicted that nano-TiO₂ concentrations in Swiss soils fall between 0.4 and 4.8 μg kg^–1, depending on the extent of the emissions (Mueller et al. 2009). In the USA, soil nano-TiO₂ increases by 0.43–2.13 μg kg^–1 y^–1 across all soil types, and by 34.5–170 μg kg^–1 y^–1 in sludge-treated soils (Gottschalk et al. 2009).

Elevated TiO₂ concentrations in soils and porewater can lead to adverse effects. Past studies have reported that plant exposure to TiO₂ altered the chlorophyll content and antioxidative enzyme activities and led to Ti accumulation in fruit, roots, stems, and leaves (Servin et al. 2013; Gardea-Torresdey et al. 2014). Additionally, TiO₂ can negatively affect wheat growth and soil enzyme activity and lead to phytotoxicity and DNA fragmentation (Sajid et al. 2015).

Furthermore, TiO₂ toxicity may also detrimentally affect aquatic and terrestrial organisms, including arthropods, nematodes, and earthworms. The toxicity of anatase TiO₂ arises from its ability to generate reactive oxygen species when exposed to UV light (Ma et al. 2012). Reactive oxygen species cause oxidative stress to various soil organisms. While TiO₂ exposure can increase the feeding rates and catalase activities of the terrestrial arthropod Porcellio scaber (Drobne et al. 2009), it can also destabilize the cell membranes in these arthropods’ isolated digestive glands (Valant et al. 2009). Moreover, the fertility and survival of nematode Caenorhabditis elegans was compromised by exposure to nano-TiO₂ (Wang et al. 2009). In fact, the degree of those effects were inversely correlated with the size of the TiO₂ particle (Roh et al. 2010). Nano-TiO₂ with a diameter <30 nm can enter plant cell-wall pores and be taken up into the roots. In roots, TiO₂ can interrupt water flow and hydraulic potential. It can also be translocated from the roots into the leaves, where it influences pigments (i.e. chlorophyll, anthocyanin, and carotenoid) (Tan et al. 2018). Anatase TiO₂ accelerates the nitrogen metabolism of spinach by increasing the activities of nitrate reductase, glutamate dehydrogenase, glutamine synthase, and glutamic–pyruvic transaminase and by promoting the absorption of inorganic nitrogen and its subsequent conversion to organic forms (Yang et al. 2006). Finally, TiO₂ can bio-accumulate in earthworms and lead to oxidative stress, DNA and mitochondrial damage, and inhibition of cellulase (Hu et al. 2010).

Many of these risks arise because TiO₂ generates reactive oxygen species (ROS) under exposure to UV-light (Skocaj et al. 2011; Fu et al. 2014). Specifically, UV-irradiated TiO₂ produces H₂O₂ in aqueous solutions (Pappas & Fischer 1975; Konaka et al. 1999). When exposed to UV-radiation, TiO₂ photochemically generates e^–, which contributes to the reduction of O₂ into H₂O₂, as seen below, in Eq. 1 (Burek et al. 2019):

(1) $O_2\stackrel{+e^{-}}{\to }·O_2^{-}\stackrel{+H^{+}}{\to }·{\mathrm{HO}}_2\stackrel{+e^{-}}{\to }·{\mathrm{HO}}_2^{-}\stackrel{+H^{+}}{\to }{H}_2{O}_2$

Subsequently, H₂O₂ can either be re-oxidized to O₂ or be further reduced to water (Burek et al. 2019).

These adverse effects have led lawmakers throughout the world to take different regulatory actions (Amenta et al. 2015; Gupta & Xie 2018). The World Health Organization has classified the carcinogenicity of TiO₂ into group 2B, deeming it “possibly carcinogenic to humans” (World Health Organization 2019). Furthermore, the European Chemicals Agency is seriously considering classifying TiO₂ as a class 1B carcinogen, deeming it “presumed to have carcinogenic potential for humans” (ANSES 2016; European Chemicals Agency 2017). The US Food and Drug Administration mandates that TiO₂ makes 1.0% or less of food by weight (U.S. Food and Drug Administration 2019) while the US National Institute for Occupational Safety and Health has recommended that TiO₂ be classified as a potential occupational carcinogen and that exposure be limited to 0.2 mg m^–3 (National Institute for Occupational Safety and Health 2011).

Moreover, photocatalysis of TiO₂ may be influenced by common, nearly ubiquitous environmental oxyanions in soil porewaters, including nitrate and phosphate). While the TiO₂ photocatalytic reactions have been studied extensively in basic research, how environmentally relevant conditions (e.g. ligand interaction) affect its photoreactivity is not well understood. Researchers previously reported that at pH 7.9, large concentrations (5–25 mmol L^–1) of a common ligand, phosphate (P), decreased hydroxyl radical (∙OH) generation by TiO₂ in freshwaters while nitrate decreased∙OH generation to a lesser extent ( Budarz et al. 2017). Others have concluded that P adsorption promoted the photocatalytic degradation of organic compounds bound weakly to TiO₂ while also inhibiting degradation of strongly sorbed organic compounds (Zhao et al. 2008). Phosphorus adsorption accelerated organic contaminant degradation at pH 6.2 in another study (Long et al. 2017). Furthermore, the presence of large concentrations of P (≥200 mM) under above neutral pH conditions boosted H₂O₂ production by TiO₂ (Xiong et al. 2018; Burek et al. 2019).

The purpose of the present study was to investigate the effect of common ligands (e.g. nitrate and phosphate) on the generation of H₂O₂ by UV-irradiated TiO₂. The results should have implications for better assessing the toxicity of TiO₂ in the aquatic and terrestrial environment. The consideration of lower, environmentally relevant nitrate and phosphate concentrations (i.e. 50 μM, 100 μM, 250 μM) and various pH conditions (4.00 ± 0.02 and 8.00 ± 0.02) differentiate the present study from previous ones. The hypothesis was that specific adsorption of P will suppress H₂O₂ production by UV-irradiated TiO₂ while the non-specific adsorption of nitrate will not interfere with the photoreactivity of TiO₂.

Materials and Methods

Materials

All reagents were American Chemical Society (ACS) grade and all solutions were prepared using ultrapure water.

Nanopowder TiO₂ (Nanostructured and Amorphous Materials, Inc., Katy, Texas, USA) and anatase 99+% with a Brunauer, Emmett, and Teller (BET) surface area of 60 m² g^–1 (Nanostructured & Amorphous Materials, Inc., Texas, USA) were used in this study. Powder X-ray diffraction (XRD) was performed on TiO₂ using a Siemens-Bruker D5000 XRD System (Bruker Corporation, Billerica, Massachusetts, USA). The voltage was adjusted to 40.0 kV and the current to 30.0 mA. The XRD angle ranged from 10 to 90°2θ, in accordance with previous studies (Sakthivel et al. 2003; Djerad et al. 2004; Anandan et al. 2008; Kathiravan & Renganathan 2009; Ravichandran et al. 2010). Furthermore, the scan speed was set at 1.2°2θ min^–1 (Djerad et al. 2004; Ravichandran et al. 2010). The 2θ position and intensity ratio of each peak on the collected spectra were compared to an anatase TiO₂ reference pattern from the International Center for Diffraction Data (ICDD 1971).

Scanning electron microscopy (SEM) analysis was conducted on TiO₂, using a Hitachi S-4700 High Resolution SEM. Because of its low conductivity, the sample was coated with gold-palladium (Au-Pd) before analysis. The conditions of the analysis included an accelerating voltage of 15.0 kV, an emission current of 10 μA, and magnification of 90,032. After the image was captured, the diameters of ~20 individual TiO₂ particles were measured and averaged.

The zeta potential and particle size of the TiO₂ was determined using a Malvern Zetasizer (Malvern Instruments, Malvern, Worcestershire, UK), using laser Doppler velocimetry and dynamic light scattering (DLS) techniques, respectively. TiO₂ samples were prepared with suspension densities of 0.1 g L^–1 and background solution of 0.001 M NaCl and in the presence or absence of 10 μM or 25 μM NO₃ ^– or P, using sodium nitrate (NaNO₃) and sodium phosphate (Na₂HPO₄), respectively. The samples were shaken on an end-over shaker at 50 rpm for 48 h and adjusted to various pH conditions, using 0.01 M HCl and NaOH. The refractive index of 2.488 was used (Zhang et al. 2011). Prior to the measurements, the zetasizer was calibrated with a polystyrene latex standard.

Methods

Adsorption of Phosphate and Nitrate by TiO₂

TiO₂ suspensions (suspension density: 1 g L^–1) were prepared in 0.001 M NaCl (Fisher Scientific International, Inc., Fair Lawn, New Jersey, USA) and immediately sonicated for 30 s. The suspensions were kept in the dark at ambient temperature on a stir plate at 300 rpm for 24 h. The pH values of the solutions were adjusted to 4.00 ± 0.02 or 8.00 ± 0.02, using 0.01 M HCl (Fisher Scientific International, Inc., Fair Lawn, New Jersey, USA) and NaOH (Fisher Scientific International, Inc., Fair Lawn, New Jersey, USA). After the pH had stabilized, Na₂HPO₄ (Fisher Scientific International, Inc., Fair Lawn, New Jersey, USA) also prepared in 0.001 M NaCl and adjusted to pH 4.00 ± 0.02 or 8.00 ± 0.02, was added to the treatment solutions to bring the final P concentrations of the eight samples to 0, 25, 50, 75, 100, 150, 200, and 250 μM, respectively. The solutions were mixed on a reciprocal shaker at 80 rpm, and aliquots of 10 mL were removed at 48 h and passed through a 0.2 μm polyvinylidene fluoride (PVDF) syringe filter. Preliminary centrifugation experiments of filtrate showed that quantifiable nanoparticles did not pass through the filter. Solution P concentrations were measured colorimetrically using the molybdenum blue method (Murphy & Riley 1962). The same procedure was carried out to determine nitrate adsorption, except NaNO₃ (Fisher Scientific International, Inc., Fair Lawn, New Jersey, USA) was added to bring the concentration of the samples to 0, 50, 100, 150, 200, and 250 μM, respectively. Nitrate concentrations were determined using UV spectrometry at 220 nm (Patey et al. 2008).

UV-irradiation of TiO₂ and H₂O₂ Determination

TiO₂ solutions (suspension density: 1 g L^–1) were prepared in 0.001 M NaCl and 10% 2-propanol (VWR, Radnor, Pennsylvania, USA) and adjusted to pH 4.00 ± 0.02 or 8.00 ± 0.02, as described above. The addition of 2-propanol optimized the analytical determination of H₂O₂ by providing additional e^– for the reduction of O₂ to H₂O₂ and suppressing H₂O₂ degradation (Burek et al. 2019). The reaction vessels, 250-mL beakers wrapped in aluminum foil, were filled with 100 mL of TiO₂ solution and placed on a magnetic stir plate set to 350 rpm. A 60 W UV-A light bulb (Feit Electric Company, Pico Rivera, California, USA) (λ: 320–400 nm) was positioned 5.5 cm above the solution surface, which had an area of 36.32 cm². Aliquots of 3 mL were removed from the samples at 0, 1, 5, 10, 15, 20, 30, 40, 60, 80, and 100 min and filtered through the 0.2 μm PVDF syringes. Immediately following filtration, the process of H₂O₂ determination was carried out on the aliquots, as described below.

Hydrogen peroxide was determined fluorometrically using homovanillic acid (HVA) (Acros Organics, New Jersey, USA) as a substrate (Guilbault et al. 1967, 1968; Staniek & Nohl 1999; Khosravi et al. 2013; Paital 2014). Solutions, prepared in ultrapure water, included 0.025 M tris hydroxymethyl aminomethane (tris) buffer (Macron Fine Chemicals, Center Valley, Pennsylvania, USA) , adjusted to pH 7.5; 10 U (Universal Enzyme unit in μmol/min is defined as the amount of the enzyme that catalyzes the conversion of 1 μM of substrate per min under the specified conditions) horseradish peroxidase (HRP) (Alfa Aesar, Ward Hill, Massachusetts, USA), and 125 μM HVA. To determine H₂O₂, solutions were combined in dark 15-mL vessels in the following order: 6.6. mL tris buffer, 0.9 mL HRP, 0.6 mL HVA, and 0.9 mL irradiation aliquot (Khosravi et al. 2013). Solutions were gently shaken and left to incubate for 2 h (Paital 2014). Then, each sample was transferred to a quartz fluorescence cuvette. The sample was illuminated with an excitation wavelength of 315 nm (Ci & Wang 1991; Paital 2014). Fluorescence emissions at 425 nm were measured using an Ocean Optics UV-Vis spectrophotometer (Ocean Optics, Inc., Largo, Florida, USA).

Prior to measuring the H₂O₂ concentration of irradiated samples, H₂O₂ polynomial standard curves were prepared, using 0–15 μM H₂O₂ standards. All H₂O₂ standards were made in 0.001 M NaCl and 10% 2-propanpol. The effect of background oxyanions on H₂O₂ measurements was tested by making additional standard curves using 50, 100, and 250 μM Na₂HPO₄ and 50, 100, and 250 μM NaNO₃ background solutions.

Kinetic and Statistical Analyses

To analyze the rate of H₂O₂ production under the different pH and oxyanion conditions, time vs H₂O₂ data for each treatment were substituted into the logarithmic regression curve (i.e. y = a ∙ ln(x) + b) and the statistical fitting parameters were determined.. At a given time x, H₂O₂ increased at a rate of a∙x ^–1 μM min^–1. Coefficients of determination (R² values) were calculated for each regression curve. The H₂O₂ data were analyzed using two sample t-tests to determine if the population means of H₂O₂ varied significantly at 100 min based on treatment effect (i.e. the presence or absence of nitrate or phosphate).

Results

TiO₂ Characterization

Various analyses were utilized to characterize TiO₂. The particle size of TiO₂ (0.1 g L^–1) in 0.001 M NaCl, measured by DLS, was 1040.67 ± 107.30 nm at pH 4.00 ± 0.05 and 808.57 ± 76.15 nm at pH 8.00 ± 0.05 (Fig. 1a). In 10 μM P, the particle size of TiO₂ was 2071.67 ± 186.92 nm at pH 4.00 ± 0.05 and 764.73 ± 7.81 nm at pH 8.00 ± 0.05. In 25 μM P, the particle size of TiO₂ was 2141.67 ± 189.87 nm at pH 4.00 ± 0.05 and 498.00 ± 28.52 nm at pH 8.00 ± 0.05.

Fig. 1. Characterization of TiO₂: ( a) particle size diameter (nm) of TiO₂ (0.1 g L^–1) in 0.001 M NaCl, as affected by pH and phosphate concentration; (b) X-ray diffraction pattern of anatase TiO₂ powder, with d-spacing values (Å) displayed over major peaks; (c) SEM image of anatase TiO₂ powder; and (d) zeta potential (mV) of TiO₂ (0.1 g/L) in 0.001 M NaCl in the presence and absence of phosphate and nitrate, as affected by pH (n = 3).

XRD patterns of the TiO₂ are shown in Fig. 1b. The XRD reflections (12.65, 18.47, 24.02, 26.94, 27.53, 31.34, 70.29, 35.14, 37.043, and 41.34°2θ) confirmed the presence of anatase. A SEM image of anatase TiO₂ powder revealed that the average individual particle size was 35.89 ± 4.18 nm (n = 19) (Fig. 1c).

Zeta Potential of TiO₂ in the Presence and Absence of Oxyanions

From approximately pH 3 to 9.5, the zeta potential of TiO₂ (0.1 g L^–1) in 0.001 M NaCl followed very similar trends in the presence of 10 μM and 25 μM nitrate and the absence of any oxyanion ligands (Fig. 1d). For instance, the zeta potential was 11.42 ± 1.34 mV at pH 4.58 in the control, 11.67 ± 0.42 mV at pH 4.55 in 10 μM nitrate, and 13.73 ± 0.91 mV at pH 4.69 in 25 μM nitrate. The isoelectric point (IEP) occurred at around pH 5.5 in all three samples.

In the presence of 10 μM and 25 μM P, however, the zeta potential of TiO₂ deviated greatly and was much smaller compared to the other conditions from approximately pH 3 to 7. For instance, the zeta potential was –20.43 ± 1.07 mV at pH 4.76 in 10 μM P and –19.97 ± 0.58 mV at pH 4.40 in 25 μM P. In 10 μM P, the IEP of TiO₂ shifted to ~pH 3.25 while, in 25 μM P, the isoelectric point occurred at a pH lower than that of the most acidic sample. Nevertheless, at ~pH 7, the zeta potential of TiO₂ in both P concentrations became similar. At pH 9.15 ± 0.23, the zeta potential of TiO₂ in all conditions averaged –37.34 ± 0.89 mV.

Adsorption of Phosphate and Nitrate at the TiO₂–Water Interface

Adsorption of P by TiO₂ was greater at pH 4.00 ± 0.02, than at pH 8.00 ± 0.02. This was anticipated, as the positively charged TiO₂ attracts negatively charged P (Connor & McQuillan 1999). At pH 8.00 ± 0.02, the surface of TiO₂ is negatively charged, weakening this attraction. The trend line of C _eq vs adsorbed q for P (Fig. 2A) followed y = 0.261 ln(x) + 2.322 (R² = 0.9647) at pH 4.00 ± 0.02 and y = 0.110 ln(x – 0.015) + 1.331 (R² = 0.9950) at pH 8.00 ± 0.02.

Fig. 2. Adsorption of (a) phosphate and (b) nitrate by TiO₂ (1 g L^–1) in 0.001 M NaCl, as affected by pH (n = 3).

Conversely, adsorption of nitrate by TiO₂ was low at pH 4.00 ± 0.02 and did not occur at pH 8.00 ± 0.02. At pH 4.00 ± 0.02, the trend line of C _eq vs adsorbed q for nitrate followed y = 9.097 x ² – 0.00601 x + 0.00165 (R² = 0.9043) (Fig. 2b).

H₂O₂ Generation by UV-Irradiated TiO₂

UV-irradiated TiO₂ generated H₂O₂ to different degrees, depending on pH and the type and concentration of oxyanion ligand (Fig. 3; Table 1). In the control at pH 4.00 ± 0.02, H₂O₂ increased at a rate of 4.1494 x ^–1 μM min^–1 and reached 18.61 μM at 100 min. The addition of P suppressed this production. At pH 4.00 ± 0.02, the rates of H₂O₂ generation were 1.76 x ^–1 μM min^–1 in the presence of 50 μM P, 0.84 x ^–1 μM min^–1 in the presence of 100 μM P, and 1.99 x ^–1 μM min^–1 in the presence of 250 μM P. H₂O₂ reached 8.70 μM, 5.38 μM, and 9.85 μM at 100 min in 50 μM P, 100 μM P, and 250 μM P, respectively. After 100 min of irradiation, H₂O₂ was significantly (p < 0.01) less in the presence of P compared to its absence.

Fig. 3. H₂O₂ generation by UV-irradiated nano TiO₂ over the course of 100 min, as affected by various concentrations of (a) phosphate at pH 4, (b) nitrate at pH 4, (c) phosphate at pH 8, and (d) nitrate at pH 8 (n = 2)

Table 1. H₂O₂ generation rates, where x = time, in min, by UV-irradiated nano TiO₂, as affected by pH and oxyanion type and concentration. Solutions were prepared in 0.001 M NaCl and 10% 2-propanol, in addition to the oxyanion solutions below, and then adjusted to the specified pH

Compared to P, the addition of nitrate did not cause the same inference in photocatalysis. At pH 4.00 ± 0.02, the rates of H₂O₂ generation were 4.15 x ^–1 μM min^–1 in the presence of 50 μM nitrate, 4.19 x ^–1 μM min^–1 in the presence of 100 μM nitrate, and 4.76 x ^–1 μM min^–1 in the presence of 250 μM nitrate. At 100 min, H₂O₂ reached 22.98, 19.91, and 22.44 μM in 50 μM nitrate, 100 μM nitrate, and 250 μM nitrate, respectively. After 100 min of irradiation, H₂O₂ was not significantly different in the presence of nitrate compared to its absence.

At pH 8.00 ± 0.02, H₂O₂ concentration increased at markedly lower rates. In the control, H₂O₂ yield followed 0.99 x ^–1 μM min^–1 and hit 5.79 μM at 100 min. In the presence of nitrate, H₂O₂ similarly reached 6.29, 5.58, and 5.53 μM, at 100 min in 50 μM nitrate, 100 μM nitrate, and 250 μM nitrate, respectively. The production follows the trends of 1.1430 x ^–1 μM min^–1 in 50 μM nitrate, 1.1081 x ^–1 μM min^–1 in 100 μM nitrate, and 0.97 x ^–1 μM min^–1 in 250 μM nitrate. After 100 min of irradiation, H₂O₂ was not significantly different in the presence of nitrate compared to its absence.

In the presence of P at pH 8.00 ± 0.02, H₂O₂ grew at rates of 2.02 x ^–1 μM min^–1, 2.08 x ^–1 μM min^–1, and 1.51 x ^–1 μM min^–1 and reached 10.04, 10.48, and 8.11 μM at 100 min in 50 μM P, 100 μM P, and 250 μM P, respectively. After 100 min of irradiation, H₂O₂ was significantly (p < 0.01) greater in the presence of P compared to its absence.

Discussion

Physicochemical Properties of TiO₂

Physicochemical properties of TiO₂ were studied using XRD, SEM, and zeta potential analyses. No differences were observed when the measured XRD pattern was compared to the anatase TiO₂ reference from the ICDD database (ICDD 1971), confirming that the TiO₂ used in the present study was in fact anatase (ICDD 1971). SEM analysis revealed that the diameter of the particles of the TiO₂ powder was larger at 35.89 ± 4.18 nm (n = 19) than the size advertised by the vender (i.e. 10 nm). Furthermore, the SEM image illustrated the aggregation of the dry powder into large clusters, which is consistent with previous characterizations (French et al. 2009; Yin et al. 2012).

The zeta potential values of the control aligned with those reported previously for TiO₂ under similar conditions (Nelson et al. 2000; Dutta et al. 2004; Kataoka et al. 2004; Kim & Choi 2011; Wan et al. 2016). Below pH 5.5, TiO₂ became increasingly positively charged as the surface became more and more protonated (i.e. TiOH₂ ⁺) (Kataoka et al. 2004). The reverse was true above pH 5.5, as the surface became deprotonated (i.e. TiO^–) (Kataoka et al. 2004). Moreover, in the presence of P, the zeta potential and IEP of TiO₂ shifted significantly downward due to the specific adsorption of P (Nelson et al. 2000; Wan et al. 2016).

The Effect of Nitrate on H₂O₂ Generation by UV-Irradiated TiO₂

H₂O₂ generation by UV-irradiated TiO₂ was similar between the control and nitrate treatments at both pH 4.00 ± 0.02 and 8.00 ± 0.02, while adsorption of nitrate by TiO₂ was minimal to zero. Nitrate is considered an indifferent ligand because it does not specifically interact with TiO₂ (Budarz et al. 2017). This lack of specific interaction between nitrate and TiO₂ was supported by zeta potential measurements, in which differences were not discernible between the control TiO₂ and the TiO₂ in 10 and 25 μM nitrate from pH 3 to 9.5. Consequently, the presence of nitrate did not effectively hinder or interfere with the ability of TiO₂ to produce H₂O₂ photochemically.

The Effect of Phosphate on H₂O₂ Generation by UV-Irradiated TiO₂

Phosphate strongly adsorbs to TiO₂, including the anatase form used in the present study, through inner-sphere monodentate and bidentate complexes (Hadjiivanov et al. 1989; Connor & McQuillan 1999; Zhao et al. 2008; Alshameri et al. 2014). Past infrared spectroscopic studies revealed that multinuclear complexation of P at the TiO₂ surfaces dominates (Connor & McQuillan 1999; Gong 2001; Ronson & McQuillan 2002).

Throughout the irradiation, adsorption of P by TiO₂ remained relatively constant under both acidic and alkaline conditions. Within the eleven aliquots taken from 0 to 100 min, aqueous P averaged 0.54 ± 0.46 μM at pH 4 in the 50 μM treatment, 22.23 ± 0.91 μM at pH 4 in the 100 μM treatment, 217.63 ± 2.15 μM at pH 4 in the 250 μM treatment, 23.74 ± 1.97 μM at pH 8 in the 50 μM treatment, 70.52 ± 0.73 μM at pH 8 in the 100 μM treatment, and 224.13 ± 5.03 μM at pH 8 in the 250 μM treatment. Across the treatments, aqueous P varied, on average, 1.88 μM P during the 100 min irradiations.

The effects of P adsorption by TiO₂ photoreactivity were more complex than nitrate and varied with pH. On one hand, much adsorption of P by TiO₂ at pH 4.00 ± 0.02 corresponded to suppressed H₂O₂ generation. After 100 min of UV-irradiation, H₂O₂ was more than three times less in the presence of 100 μM P than in the absence of P. On the other hand, comparatively less adsorption of P by TiO₂ at pH 8.00 ± 0.02 corresponded to enhanced H₂O₂ production. After 100 min of UV-irradiation, H₂O₂ was nearly two times greater in the presence of 100 μM P than in the absence of P. These results can be explained by the contrasting behavior of suspended TiO₂ and H₂O₂ degradation under the different pH conditions.

Adsorbed P caused a shift in the behavior of suspended TiO₂. At pH 4.00 ± 0.05, the particle size of TiO₂ was approximately two times larger in the presence of 10 μM and 25 μM P than in the control. This is consistent with past research, which demonstrated that the adsorption of P (>50 μM) by TiO₂ led to its aggregation at pH ~3, which is lower than the IEP of the pure solid, 6.3 (Wan et al. 2016). Conversely, at pH 8.00 ± 0.05, the particle size of TiO₂ in the presence of 10 μM and 25 μM P was smaller than in the control. Past researchers have similarly observed that the adsorption of P (>50 μM) by TiO₂ under neutral to alkaline conditions (i.e. pH 7.0 and 7.9) facilitates the dispersion (Wan et al. 2016; Budarz et al. 2017).

In the present work, specific adsorption of P at pH 4.00 ± 0.02 caused the TiO₂ to aggregate and, thus, reduced its ability to generate H₂O₂ photochemically under all three concentrations (Connor & McQuillan 1999; Fujishima et al. 2008; Budarz et al. 2017). Inner-core particles within TiO₂ aggregates will not be as accessible as outer-core particles. The degree to which H₂O₂ production was suppressed correlated positively with the magnitude of P adsorption (100 μM > 50 μM > 250 μM) during irradiation. Previous studies confirmed that the aggregation of nano-TiO₂ reduces its photoreactivity, due to the reduction of reactive surface area and photon penetration (Maira et al. 2000; Hotze et al. 2010; Jassby et al. 2012). Thus, the presence of P reduced the production of H₂O₂ by UV-irradiated TiO₂ at pH 4.00 ± 0.02.

Conversely, at pH 8.00 ± 0.02, greater H₂O₂ was produced in the presence of P than in its absence. Interestingly, the extent to which H₂O₂ production was enhanced correlated with the degree of P adsorption (50 μM ~ 100 μM > 250 μM) during irradiation. The degradation of H₂O₂ at pH 8.00 ± 0.02 was approximately three times greater than at pH 4.00 ± 0.02 (Burek et al. 2019). However, under alkaline conditions, the adsorption of P by TiO₂ suppressed the adsorption of H₂O₂ onto TiO₂ and, thus, hindered its photocatalytic degradation, enabling greater aqueous H₂O₂ accumulation in the presence of P than in its absence. This phenomenon has been documented by several researchers (Moon et al. 2014; Xiong et al. 2018). Moreover, the dispersion of TiO₂ caused by P adsorption under alkaline conditions optimized its photoreactivity. Consequently, the presence of P increased aqueous H₂O₂ at pH 8.00 ± 0.02.

Environmental Implications and Uncertainties

The presence of ≥50 μM P at pH 4.00 ± 0.02 facilitated the aggregation of TiO₂, reducing its photoreactivity and ability to produce H₂O₂. Because P is nearly ubiquitous within porewaters and surface waters throughout the world, TiO₂ toxicity in terrestrial environments, under acidic conditions, may be of less concern than originally thought. However, water-soluble P varies widely from <1 to 1000 μM (Fried & Broeshart 1967; Sharpley & Smith 1989), so the extent of this suppression by P in soil environments likely varies widely. At the same time, P enhanced H₂O₂ generation at pH 8.00 ± 0.02 because P adsorption reduced the degradation of H₂O₂.

Additionally, dissolved nitrate was up to 10³ times more prevalent in soil pore waters than P (Fried & Broeshart 1967) and did not effectively suppress or increase H₂O₂ production. Lastly, the effects of other compounds (e.g. other anions, organic compounds, colloids) commonly found in the soil matrix and their complex interactions on the photochemistry of TiO₂ are not yet understood and should be evaluated further. Thus, many uncertainties remain regarding the extent of ROS generation by TiO₂ in terrestrial and aquatic environments.

Conclusions

The ability of nano-TiO₂ to generate reactive oxygen species under UV light exposure raises concerns about its potentially adverse effects in soil, where agricultural applications and the disposal of consumer products have enabled its accumulation. In the present study, the production of H₂O₂ by UV-irradiated TiO₂ was monitored in the presence and absence of common environmental oxyanion ligands (i.e. nitrate and P) of different concentrations under acidic and alkaline conditions. At pH 4.00 ± 0.02, aggregation, caused by the specific adsorption of P onto TiO₂, suppressed the generation of H₂O₂. Conversely, adsorption of P at pH 8.00 ± 0.02 reduced H₂O₂ degradation, enabling higher H₂O₂ accumulation in solution. Nitrate, which does not specifically interact with TiO₂, did not affect the photoreactivity of TiO₂ at pH 4.00 ± 0.02 or 8.00 ± 0.02. The present work provided evidence that P positively or negatively influences TiO₂ photoreactivity at concentrations that are environmentally relevant in soil porewaters, depending on pH, while these same small concentrations of nitrate do not influence the photoreactivity of TiO₂.