Improving the photocatalytic degradation of EDTMP: Effect of doped NPs (Na, Y, and K) into the lattice of modified Au/TiO2 nano-catalysts

Improving the photocatalytic degradation of EDTMP: Effect of doped NPs (Na, Y, and K) into the lattice of modified Au/TiO 2 nano-catalysts Ramona Riedel a,* , Julia Schowarte b , Laura Semisch a , Miriam Gonz´ alez-Casta˜ no b,c , Svetlana Ivanova c , Harvey Arellano-García b , Marion Martienssen a a Chair of Biotechnology of Water Treatment Brandenburg, Institute of Environmental Technology, BTU Cottbus-Senftenberg, Cottbus, Germany b Department of Process and Plant Technology, Brandenburg University of Technology (BTU) Cottbus-Senftenberg, Cottbus, Germany c Department of Inorganic Chemistry and Materials Sciences Institute, University of Seville-CSIC, Seville, Spain ARTICLE INFO Keywords: EDTMP Nanoparticles Phosphonate TiO 2 Yttrium ABSTRACT This study presents the photocatalytic degradation of the aminophosphonate ethylenediaminetetra(methylenephosphonic acid) (EDTMP) with a range of different doped nanoparticles (NP). The photocatalysts were based on TiO 2 benchmark P25 and gold (Au) doped either with sodium (Na), potassium (K) or yttrium (Y). The synthesized photocatalysts were characterized via TEM, XRF, XRD, UV-DRS (band gap estimation) and N 2 -physisorption. Photocatalytic pre-screening at pH values of 3, 7 and 10 indicated highest o-PO 4 release of EDTMP at pH 7 and 10 for NP either doped with K or Y. The results of LC/MS analysis showed that the NPs doped with 5 % Y (Au2/Y5/P25) resulted in the fastest degradation of EDTMP. The target compound was completely degraded within 60 min, 4 times faster than photochemical treatment of unadulterated EDTMP. Importantly, also the transformation products were accelerated by the photocatalytic treatment with Au2/P25 either doped with 5 % Y or 10 % K. The results of scavenger experiments indicated that the enhanced photocatalytic degradation of EDTMP is primarily attributable to the presence of hydroxyl radicals in the bulk and to a lesser extent to •O 2 − and electron-holes (h + ) at the surface of the catalysts. The study demonstrates that the catalytic efficiency of TiO 2 nanocomposites is significantly influenced by the choice of dopants, which affect particle size, band gap, and photocatalytic activity. Yttrium at low concentrations (i.e., 5 wt% Y) doping emerged as particularly effective, enhancing both the visible light absorption and h + separation, leading to superior photocatalytic performance in the degradation of EDTMP. The Au content also plays a crucial role in enhancing the photocatalytic efficiency. However, the combination of Au and Na doping was found to be less effective for this photocatalysis in aqueous media, potentially due to larger particle sizes and insufficient dopant contents. In conclusion, the findings emphasise the necessity of optimising both the selection of dopants and the design of catalysts in order to enhance photocatalytic applications. 1. Introduction Phosphonates are chelating agents with an unique performance spectrum in aqueous solution. Their capacity to form sub-stoichiometric complexes with dissolved polyvalent metal ions, including calcium and magnesium, renders them a preferred choice over several other chelating agents. They are highly effective in preventing the precipitation of poorly soluble alkaline earth metals, improve the dispersion of solids, and greatly reduce recontamination of fabrics. Some phosphonate structures have corrosion inhibiting properties [1]. Others, such as ethylenediaminetetra(methylenephosphonic acid) (EDTMP), are important anti-scaling and bleach stabilising additives in detergents [2]. Phosphonates are easy to formulate and, as mentioned above, provide noticeable benefits to the user even at very low application concentrations. This is the main reason for their increased use in recent years. It is therefore not surprising that their release into the aquatic environment has also increased [3–5]. Phosphonates, such as EDTMP, are increasingly utilised and consumed on an annual basis. Nevertheless, it is challenging to ascertain the precise quantity of EDTMP employed in Europe and worldwide. However, based on the latest data of the European Chemicals Agency (ECHA), the total annual production of EDTMP averages 1,000 and * Corresponding author. E-mail address: [email protected] (R. Riedel). Contents lists available at ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej https://doi.org/10.1016/j.cej.2025.160109 Received 19 October 2024; Received in revised form 8 January 2025; Accepted 30 January 2025 Chemical Engineering Journal 506 (2025) 160109 Available online 31 January 2025 1385-8947/© 2025 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). 10,000 tons per year [2]. It is estimated that the number of products containing phosphonates such as EDTMP sold in Europe exceeds 2,700 [6]. Reliable data of the production of phosphonate on the American and Asian markets is also a challenging endeavour. Some useful data can be found elsewhere [7,8]. Rott et al. [7] reported a worldwide consumption of phosphonates of 94,000 tons per year in 2012 with a continuously increasing trend. Rott et al. [7] found that 20,900 tons of phosphonate from detergents of laundry and dishwasher enter municipal wastewater treatment plants (WWTP) in Europe every year. A total of 1.2 to 1.5 g phosphonate kg −1 dry substance (80 % to 95 %) is removed by adsorption to the sewage sludge. Approximately 5 % up to 20 % is discharged into the receiving water. In a further study [9], the distribution of phosphonates in 42 sediment samples was investigated, as well as the influence of wastewater effluents in the federal state of Lower Saxony in Germany. The study demonstrated that the effluent of the WWTP did indeed affect the phosphonate content of the sediment samples. It was observed that organoand aminophosphonates demonstrated elevated levels of enrichment in water bodies that had been subjected to anthropogenic influences. Over the past decade, much research has been published on the removal of phosphonates from aquatic solutions. The majority of studies have concentrated on the utilisation of advanced oxidation processes (AOP), with a particular emphasis on photochemical treatments [5,10–17]. These are frequently employed in conjunction with additives to facilitate enhanced oxidation. One of the most fundamental studies on the photochemical degradation of aminophosphonates was published by Lesueur et al. [17]. Their study focused on the influence of different pH values and the addition of Fe 3+ during the photolysis of aminotris (methylenephosphonic acid) (ATMP), EDTMP, hexamethylenediaminetetra(methylenephosphonic acid) (HDTMP) and diethylenetriaminepenta(methylenephosphonic acid) (DTPMP). The main results of this study were the identification of amino(methylenephosphonic acid) (AMPA) as the major transformation product of all the aminophosphonates tested and the determination of their half-lives. However, the latter was based only on ortho-phosphate (o-PO 4 3− ) releases, which did not accurately reflect the degradation of the parent compounds. Recently, Kuhn et al. [2] reported the complete degradation pathway of EDTMP monitored by liquid chromatography coupled mass spectrometry (LC/MS) analysis. They showed that in addition to AMPA, iminodi(methylenephosphonic acid) (IDMP) and ethylamino(bismethylphosphonic acid) (EABMP) were also released. Similar results, i.e., the release of the same transformation products, were obtained for the ultraviolet (UV) degradation of ATMP, HDTMP and DTPMP when LC/ MS analyses were applied by other scientists [1,5,18]. To date, several photochemical studies have investigated the enhanced degradation of phosphonates either by the addition of metals or other additives that predominantly accelerate the degradation kinetics of the parent compound [1,16–20]. In most cases, the release of oPO 4 3− was used to monitor the enhanced degradation of the parent compound. Only a few scientists have also been able to directly monitor the degradation of the parent compound and the release of transformation products [1,5]. However, the use of metals, permonosulphate (PMS) or persulphate (PS) significantly accelerated the degradation kinetics of the phosphonates studied, and partly also their transformation products [10,13,20–22]. The reported degradation mechanisms were dependent on direct electron transfer between metal and parent compound, or on the formation of reactive oxygen species (ROS). In particular, for applications using PS and PMS, the formation of sulphate radicals as the main ROS has been reported [11,22]. Zhu et al. [21] combined PMS with hydroxylamine and Cu(II) and found that this combination degraded HEDP 3.4 times faster than PMS/Cu(II). Interestingly, Zhu et al. [23] recently showed that bromide (Br − ) further enhanced the degradation of HEDP only in the presence of PMS. Nevertheless, such application may not be practical for wastewater treatment. Advanced degradation of both the parent compound and the transformation products with accelerated kinetics is the main objective of photodegradation studies. This can be achieved, for example, using photocatalysts. Common to all photocatalysts is the formation of electron-hole pairs (e − /h + ) when the absorbed photon energy is equal to or greater than the so-called band gap energy between the valence band and the conduction band. In such cases, the excited electron pair is transferred to the conduction band [24]. According to Chen et al. [25], O 2 or H 2 O and/or OH – absorbed at the surface of the photocatalyst then prevents the recombination of the electrons and, thus, forms either superoxide radicals (•O 2 ) or hydroxyl radicals (•OH). The degradation mechanism for photodegradation is described in more detail by Lanjwani et al. [26]. Several studies highlight the application of TiO 2 or TiO 2 -doped composites as photocatalysts for the degradation of glyphosate [25,27–30]. Only few studies, however, report the use of photocatalysts for the degradation of other phosphonates [31,32]. Very often, large band gaps, i.e., above 3.0 eV, are reported for TiO 2 - based catalysts, whether they are microor nanoscale. Large band gaps, on the one hand, prevent the absorption of visible light (i.e., sunlight), and on the other hand, promote the rapid recombination of the generated electron-hole pairs [33]. Therefore, doping can easily reduce the band gap and improve the performance of the catalyst. Several semiconductors and/or noble metals can be used for doping, depending on the application of the catalysts. Recently, yttrium (Y) has been highlighted as a very promising dopant on the surface of nano-hexagonal sheets of magnesium oxide for the treatment of industrial wastewater containing the antibiotic ciprofloxacin and methylene blue dye [34]. Mukhtar et al. [35] also highlighted the synthesis of Y-doped nanoparticles (NPs) with excellent redox properties to photodegrade hazardous dyes via sunlight. Similar beneficial effects such as increased photodegradation activity, reduction of band gap, increased photosensitivity, and/or increased surface area have been reported by many others [36–39]. Besides Y-doping, also the modification of the TiO 2 surface with gold (Au) can further enhance both the charge separation and solar-driven photocatalysis for removing organic pollutants in the aquatic environments [40–43]. Mu et al. [43] reported modification of Au and copper (Cu) on TiO 2 resulting in excellent carrier separation and better photocatalytic performance. Thus, photocatalysis based on TiO 2 such as the benchmark P25 modified with selected noble metals might also be very promising for enhanced removal of commercially available phosphonates such as EDTMP. The objective of the presented study was to investigate the photocatalytic performance of several NPs based on Au/TiO 2 , either loaded with Y, sodium (Na) or potassium (K), in the degradation of EDTMP. The modified NPs were characterised by TEM and for their chemical composition by inductively coupled plasma optical emission spectroscopy (ICP-OES), X-ray diffraction (XRD), and UV–DRS spectra were captured. The performance of all synthesised NPs was tested for their oPO 4 3− release of UV-treated EDTMP at three different pH values, namely 3.0, 7.0 and 10.0. The NPs that performed best were selected for more detailed studies including LC/MS analyses. The degradation pathway and kinetics were determined. The optimal dosage of the most effective NP in comparison with P25 was determined. Finally, scavenger tests were carried out to determine the major ROS responsible for the degradation of EDTMP. 2. Material and methods 2.1. Chemicals and reagents The phosphonates EDTMP and EABMP were provided by “Zschimmer & Schwarz Mohsdorf” (Burgst¨ adt, Germany). AMPA and IDMP and the relative internal standard glyphosate for LC/MS analysis were purchased from Sigma Aldrich (Steinheim, Germany). All standards were of analytical grade or better with purity >99 %. Ultra-pure water (LC/MS grade) was in-house generated (Adrona Sia Crystal EX, Lithuania). R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 2 Acetonitrile of LC/MS grade was purchased from VWR (Dresden, Germany), ammonium acetate and ammonium formate were of analytical grade and were purchased from VWR (Leuven, Belgium). P25 utilized as photocatalyst was purchased from Sigma-Aldrich (Steinheim, Germany). The purity of P25 was ≥99.7 %, the particle size was below 25 nm and the crystal phase was anatase. Dopants hydrogen tetrachloroaurate(III) trihydrate (HAuCl 4 ⋅3H 2 O) with purity >99,5 % was purchased from Roth (Karlsruhe, Germany), Yttrium(III) nitrate hexahydrate (Y(NO 3 ) 3 ⋅6H 2 O) with purity >99.9 % from Thermo Fisher Scientific (Kandel, Germany), KOH from Riedel de Haen (Seelze, Germany), and Na 2 CO 3 was purchased from Fluka (Neu-Ulm, Germany). Methanol (MeOH) was purchased from Merck (Darmstadt, Germany), ascorbic acid (AsA) from VWR (Leuven, Belgium), potassium iodide (KI) from BLDpharm (Shanghai, China) and isopropyl alcohol (i-PrOH) from VWR (Fontenay-sous-Bois, France). Ethanol (VWR, Leuven, Belgium; 99.5 %) and 25 % ammonia solution (Merk, Darmstadt, Germany) were used for synthesis. 2.2. Synthesis of the TiO 2 -supported catalysts P25 nano-powder was employed as support for all prepared catalysts. Prior the metals incorporation, 5.0 g P25 was dissolved in 100 mL distilled water and stirred during 1 h at 500 rpm at room temperature (RT). The solid was separated via vacuum filtration with a qualitative filter paper grade 410 with a pore size of 2 µm (VWR, Strasbourg, France), rinsed three times with 10 mL EtOH and dried overnight at 120 ◦C. Through filter cake formation from coagulating nanoparticles during filtration most material was retained in the filter surface. Finally, P25 was calcined for 4 h at 350 ◦C using a 10 ◦C min −1 linear heating ramp. Y, Na and K dopants were wetness impregnated using Y(NO 3 ) 3 ⋅6H 2 O, KOH and Na 2 CO 3 as corresponding precursors. Thus, the adequate precursor amounts were dissolved, dried and calcined at 450 ◦C for 5 h (heating ramp 10 ◦C min −1 ). For each dopant, the intended nominal contents were 2, 5 and 10 wt%. Over the resulting solids, Au catalysts were synthetized using the direct anionic exchange method assisted by ammonia described in [44]. Shortly, a HAuCl 4 solution was prepared, i.e., 0.08 g gold percussor dissolved in 20 mL distilled water for 30 min at 300 rpm at RT. Afterwards 80 mL of distilled water was added and the solution was heated up to 70 ◦C for 2 h under constant stirring at 300 rpm. Subsequently, 2.0 g of support (Y/P25 or K/P25 or Na/P25) was added to the suspension and homogenised for 1 h at 70 ◦C at 600 rpm. Afterwards, 25 % ammonia solution was added dropwise to adjust the pH at value 7 under constant stirring. The resulting suspension was vacuum filtered as above described, dried at 120 ◦C overnight, and calcined at 300 ◦C for 4 h (heating ramp 10 ◦C min −1 ). The nominal Au content was fixed for all catalysts at 2 wt%. 2.3. Characterization techniques The surface properties of the photocatalysts were characterized by porosimeter with the help of a Nitrogen adsorption–desorption isotherm. The experiments were carried out at liquid nitrogen temperature on a Micromeritics Tristar II (Iberfluid Instruments, Barcelona, Spain) equipment. Prior the analysis, the samples were degassed for 2 h at 250 ◦C in vacuum. The chemical composition of the samples was analysed by ICP-OES using an ULTIMA 2 Spectro ICP spectrometer (HORIBA Scientific, Longjumeau, France). XRD analysis was conducted on a X’Pert Pro PANalytical instrument (Malvern, Madrid, Spain). Diffractograms were recorded using Cu Kα radiation (40 mA, 45 kV) over a 2 θ-range of 20–80◦and a position-sensitive detector using a step size of 0.05◦and a step time of 300 s. The UV-DRS spectroscopy was carried out on a Varian spectroscopy model Cary 100 (Palo Alto, California, USA), equipped with an integrating sphere using BaSO 4 as a reference. The spectra were obtained in diffuse reflectance mode and the band gap (BG) energies were estimated through the Kubelka-Munk approximation. TEM micrographs were performed in a PHILIPS CM-200 (Amsterdam, Netherland) equipped with an EDS analyser with a minimum spot size of 15 nm and a maximum resolution of 2.8 Å; the electron energy was 160 kV. The average particle size was estimated according to equation (1) where D i and v i account for the geometric diameter and the number of particles, respectively. More than 300 particles were measured for ensuring representative analysis. Di=∑n 1D3 ivi ∑n 1D2 ivi (1) 2.4. Performance of pre-screening photocatalysis experiments Pre-screenings of photocatalysed experiments were carried out with 100 mg L −1 EDTMP dissolved in ultra-pure water at pH 3, 7 and 10 for 180 min each. For photocatalysis, 100 mg L −1 of catalyst material was added to the 100 mg L −1 EDTMP solution and incubated for 30 min in the dark at room temperature before running the UV treatment. The following materials were tested: 2 wt% Au on P25 (Au2/P25); 2 wt% Au, 2 wt% Na on P25 (Au2/Na2/P25); 2 wt% Au, 5 wt% Na on P25 (Au2/ Na5/P25); 2 wt% Au, 10 wt% Na on P25 (Au2/Na10/P25); 2 wt% Au, 2 wt% Y on P25 (Au2/Y2/P25); 2 wt% Au, 5 wt% Y on P25 (Au2/Y5/ P25); 2 wt% Au, 10 wt% Y on P25 (Au2/Y10/P25); 2 wt% Au, 2 wt% K on P25 (Au2/K2/P25); 2 wt% Au, 5 wt% K on P25 (Au2/K5/P25) and 2 wt% Au, 10 wt% K on P25 (Au2/K10/P25). EDTMP without catalyst addition was tested as reference. All pre-screening experiments were performed for 180 min at room temperature in triplicates with system configuration 1 shown elsewhere [1]. The UV lamp (150 W medium pressure mercury lamp, TQ 150, Heraeus Noblelight, Hanau, Germany), which has a wavelength range from 190 to 600 nm and an emission maximum at 366 nm, was turned on 10 min before each experiment to ensure comparable conditions. The maximum emission of the lamp was at 366 nm with 3.3 ×10 −5 ±0.2 ×10 −5 E s −1 incident photon flux, which corresponds to a light intensity of 1.1 mW cm −2 . All UV experiments were conducted in an open quartz-glass vessel (320 mL) with outer cooling jacket, as described by Kuhn et al. [18]. A magnetic stirrer continuously mixed the solution, and the pH was adjusted with HCl or NaOH before UV treatment. The reactor temperature was maintained between 20 and 22 ◦C throughout the experiment. Liquid samples were taken for determining the o-PO 4 3− release photometrically (see 2.9 Analytics). 2.5. Long-term photolysis experiments For 300 min UV irradiation experiments were carried out with Au2/ Y2/P25, Au2/Y5/P25, Au2/K2/P25, Au2/K10/P25 and P25. The treatment was carried out with 100 mg L −1 EDTMP and 100 mg L −1 addition of catalysts with prior incubation for 30 min in the dark at room temperature. All photocatalysis experiments were carried out at pH 7. Liquid samples were collected at the starting point, after 10 min, 30 min and then in 30 min intervals until 300 min. Samples were analysed for the standard phosphonates EDTMP, IDMP, EABMP, AMPA via LC/MS and for o-PO 4 3− photometrically. All experiments were performed as triplicates. 2.6. Photocatalysis for determining the reaction rate constants The catalysts Au2/Y2/P25, Au2/Y5/P25, Au2/K2/P25 and P25 were utilized for kinetic experiments. Again, 100 mg L −1 EDTMP was mixed with 100 mg L −1 of catalysts and incubated for 30 min in the dark at room temperature prior to photolysis treatment. All photocatalysis experiments were carried out at pH 7 and samples were collected at the starting point and then in 10 min intervals until 180 min. Samples were analysed for the standard phosphonates EDTMP, IDMP, EABMP, AMPA via LC/MS and for o-PO 4 3− photometrically. All experiments were R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 3 performed as triplicates. 2.7. Photocatalysis for determining the optimal dosage rate of Au2/Y5/ P25 The different dosages of Au2/Y5/P25 were utilized to determine the optimal photolysis condition for the catalysts. Prior to all experiments, 100 mg L −1 EDTMP was mixed with either 100 mg L −1 , 50 mg L −1 or 25 mg L −1 of Au2/Y5/P25 and incubated for 30 min in the dark at room temperature. All photocatalysis experiments were carried out at pH 7 and samples were collected at the starting point, after 10 min, 30 min and then in 30 min intervals until 300 min. Samples were analysed for the standard phosphonates EDTMP, IDMP, EABMP, AMPA via LC/MS and for o-PO 4 3− photometrically. All experiments were performed as triplicates. 2.8. Scavenger tests Scavenger experiments with MeOH, i-PrOH, KI and AsA for the scavenging of reactive species h + and •OH [45], •OH (bulk) [46], •OH & h + [47], •O 2 − [48] were conducted with the addition of scavenger before the UV treatment. The pH of the EDTMP solutions was adjusted to pH 7 before starting the experiments with the addition of 1 M NaOH. Prior to the UV treatment, 100 mg L −1 catalyst was added to 100 mg L −1 EDTMP solution and kept stirring for 0.5 h in the dark to ensure adsorption equilibrium. Subsequent, the scavenger either 0.3 mol MeOH, 0.3 mol iPrOH, 10 mmol KI or 0.5 mmol AsA was added, and the UV treatment was started. Samples were taken at the starting point and, 10 min, 20 min, 30 min, and then at 30 min intervals until a total of 180 min. All tests were performed in triplicates at 22 ◦C in the treatment solution. Liquid samples were taken for determining the o-PO 4 3− release. All scavenger experiments were conducted as triplicates. 2.9. Analytics Prior to the LC/MS analyses, all samples were centrifuged at 17,000 rpm and filtrated using a cellulose nitrate filter with a pore size of 0.2 µm polyvinylidene fluoride syringe filter (Roth, Karlsruhe, Germany). European standard procedure EN ISO 6878:2004 [49] was used for determining o-PO 4 3− measured with a Shimadzu UV-2450 spectrophotometer (Tokyo, Japan). For LC/MS analyses, all samples were mixed with (1:1 v/v) acetonitrile. The phosphonates EDTMP, EABMP, IDMP and AMPA were determined by liquid chromatography-electro spray ionization-mass spectrometry (LC-ESI-MS) using a Finnigan MAT LC/MS (LC spectral system P4000, LCQ MS Detector, autosampler AS 3000). The separation of the samples was carried out on a SeQuant ZIC-Hilic column (150 x 2.1 mm, 3.5 µm/100 Å; Merck, Darmstadt, Germany). The column temperature was 35 ◦C and the flow was 0.2 mL/min. The mobile phase consisted of a gradient mixture of water (solvent A) and acetonitrile (solvent B). The gradient started with 10 % solvent A and was held isocratic for 5 min. The percentage of solvent A was increased linearly to 60 % for 10 min and was held for 5 min. Afterwards, solvent A was linearly increased to 90 % for 3 min and was held for another 5 min. The settings for the detector were negative ionisation with 3.0 kV and spray capillary temperature of 220 ◦C. For quantitative determination of the parent compound and known breakdown products the selected-ion monitoring (SIM) with the following mass-to-charge (m/z) ratios were used: EDTMP 435, EABMP 232, IDMP 204 and AMPA 111. For qualitative determination of unknown breakdown products full scan mode was used with a m/z range from 90 to 650. Glyphosate was used as an internal standard. 3. Results 3.1. Characterisation of nano particles The chemical composition obtained for the samples according to XRF as well as the BG energies are presented below (Table 1). With respect to the intended concentrations, lower metal and dopant contents were noticed most likely due to the synthesis method. The impregnated dopants were partially lost during the Au incorporation. The particularly accentuated loss discerned in the K-doped systems could suggest weaker K-TiO 2 interactions. The results of the determined BG energies indicate the lowest BG for catalyst Au2/P25 and the highest for the pure anataseP25, 2.6 eV and 3.4 eV respectively. The further doting of Au2/P25 with Y, K and Na resulted in higher BG energies from 2.7 eV for Au2/K2/P25 up to 3.5 eV for Au2/Y2/P25. The textural properties of the prepared catalysts were evaluated by N 2 -physisorption. All samples exhibited relatively low surface areas, which is characteristic for the P25 support (P25: 50 m 2 g −1 , Au2/P25: 49 m 2 g −1 , Y-doped: 48 m 2 g −1 , K-doped: 47 m 2 g −1 , Na-doped: 43 m 2 g −1 ). The presence of all dopants resulted in higher pore volumes (V pore ; P25: 0.39 cm 3 g −1 , Au2/P25: 0.29 cm 3 g −1 , Y-doped: 0.41 cm 3 g −1 , Kdoped: 0.41 cm 3 g −1 , Na-doped: 0.41 cm 3 g −1 ) and pore diameters (D pore ; P25: 31 nm, Au2/P25: 18 nm, Y-doped: 23 nm, K-doped: 30 nm, Na-doped: 30 nm). The observed values underline the mesoporous character of the synthesized samples. The structural features of the synthetized catalysts were analysed by XRD (Fig. 1). P25 support exhibited the characteristic diffraction lines of anatase (JCPDS 73-1764) and rutile (JCPDS 78-1510) polymorphs which also were identified in all samples. Both phases were visible in all samples as compared to [50]. The low content achieved in the catalysts is responsible for the absence of distinct peaks associated with the incorporated promoters or Au metal. With regard to the variation in dopant concentration for each dopant, no discernible changes were observed in their XRD diffractograms. However, for Au2/K10/P25, a shoulder peak around 29◦emerged, which was a consequence of the higher quantity of K-precursor. The catalyst surface became saturated with the K-precursor, resulting in the formation of non-bound K-precursor on the catalyst surface. This was subsequently oxidised after calcination to form K 2 O. The UV-DRS spectra of all catalysts (Au2/P25 and their corresponding Y, K and Na-doped variations) exhibited comparable patterns (Fig. 2). They indicated a pronounced absorption cut-off at approximately 400 nm in the lower visible region, accompanied by a secondary, broader, and less intense absorption peak at around 550 nm in the visible region. This second absorption peak correlates with the purple appearance of the gold-doped catalysts, which is consistent with the findings of Melvin et al. [51]. A slight blue shift (towards shorter wavelengths) was observed for the K and Y doped samples, both in the onset of the 400-nm-cut-off and in the broad 550-nm-peak. A significant increase of absorption in the visible region was determined for Au2/K2/P25, Au2/K10/P25, Au2/ Table 1 Chemical composition and estimated band gaps for the catalyst series. Catalyst Au [wt.%] TiO 2 (P25) [wt.%] M [wt.%] BG [eV] Au2/P25 1.75 98.25 −2.6 Au2/Y2/P25 1.41 97.28 0.98 3.5 Au2/Y5/P25 1.41 94.91 3.55 3.2 Au2/Y10/P25 0.99 93.22 5.45 3.4 Au2/K2/P25 1.59 98.06 0.08 2.7 Au2/K5/P25 1.86 97.70 0.02 3.0 Au2/K10/P25 1.89 97.55 0.15 3.2 Au2/Na2/P25 0.87 98.56 0.17 3.3 Au2/Na5/P25 0.35 98.77 0.64 3.2 Au2/Na10/P25 0.40 98.18 0.94 3.3 R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 4 Y2/P25 and Au2/Y5/P25 in comparison to the reference Au2/P25. TEM was performed on selected samples along with their corresponding particle size distributions to gain better insight into the effect of the different promoters on the size of the Au NPs (Fig. 3). A mixture of smaller spherical anatase crystallites and larger angular rutile crystallites can be observed [52], a finding that has also been confirmed by XRD analysis. All samples showed well-dispersed, round and small gold NPs, which could be identified from their dark contrast similar as recently described [53]. A comparable average particle size distribution was estimated for NP Au2/P25, Au2/K2/P25 and Au2/Y2/P25, with average values of 4.5 nm, 5.1 nm and 4.7 nm, respectively (Fig. 4). In the case of the Au2/ Na2/P25, slightly larger particle sizes were obtained (7.4 nm). 3.2. Pre-screening with different doped Au2/P25 NPs for photocatalysis of EDTMP Pre-screening experiments were carried out to determine best performance conditions of the synthesized catalysts with different dopants. The release of o-PO 4 3− was used as initial indicator for successful photocatalytic degradation of EDTMP. All catalysts, applied at same treatment conditions, were tested at the pH values 3, 7 and 10. As reference, the release of o-PO 4 3− of EDTMP without catalyst was also measured (Fig. 5). Overall, the lowest o-PO 4 3− releases was found at acid pH for all photocatalytic treatments of EDTMP, regardless of whether catalysts were used or not. In particular, the reference treatment without catalyst resulted in release of 24.8 ±1.8 mg L −1 o-PO 4 3− corresponding to a relative release of 28.5 ±2.1 % after 180 min. No significant increase or decrease of the o-PO 4 3− release was measured for most of the tested catalysts. For the group of catalysts doped with Y and K, only slight increases were measured. In contrast, increased release was observed for almost all treatment conditions at pH value 7 except for NP P25 without metal doping (21.7 ±1.9 mg L −1 o-PO 4 3− corresponding to 25.0 ±3.0 %). In comparison, catalyst Au2/P25 resulted in release of 52.3 ±1.9 mg L −1 o-PO 4 3− (corresponding to 60.0 ±2.2 %) which underlines the importance of Au as key dopant (Table S1). The highest o-PO 4 3− releases were determined for the catalysts Au2/Y5/P25, Au2/K10/P25 and Au2/K2/P25 resulting in 65.7 ±3.7 mg L −1 , 63.1 ±1.8 mg L −1 and 60.8 ±2.5 mg L −1 corresponding to 75.4 ±4.2 %, 72.5 ±2.1 % and 69.8 ±2.9 % respectively. UV treatments at pH value 10 resulted predominantly for Au2/P25, Au2/Na2/P25, Au2/Na5/P25 and Au2/K5/P25 in further increases of the o-PO 4 3− release. For all other catalysts decreased release was determined. In summary, the best treatment condition for the photocatalysis of 100 mg L −1 EDTMP occurred for the catalyst group doped with Y and K at pH 7. For Na doped catalysts, the best performance was determined at pH value 10. However, the o-PO 4 3− releases were not as high for the Nadoped catalysts as compared to the Y and K dopants at pH 7. Therefore, only Au2/Y2/P25, Au2/Y5/P25, Au2/K2/P25 and Au2/K10/P25 were selected for further performance studies. 3.3. UV treatment of EDTMP with selected Au/P25 doped with Y and K The selected nano photocatalysts, Au2/Y2/P25, Au2/Y5/P25, Au2/ Fig. 1. XRD diffractograms obtained for the Au catalyst series. A) K-doped, B) Na-doped and C) Y-doped systems. Fig. 2. UV-DRS spectra obtained for the Au catalysts series. A) K-doped; B) Nadoped and C) Y-doped systems. R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 5 K2/P25 and Au2/K10/P25, were subjected to further investigations with regard to their influence on the degradation mechanism of EDTMP during photocatalytic treatment. The benchmark P25 (undoped) was also investigated as a reference nano catalyst. Due to the result of the pre-screening, all UV treatments were carried out at the pH 7 to achieve the highest degradation rates of EDTMP. The results of the LC/MS analyses confirmed rapid degradation of EDTMP during the initial 60 min with all nano catalysts compared to the reference UV treatment without catalyst. The latter required almost three-fold longer degradation time. The results further indicated that both catalysts Au2/Y5/P25 and Au2/K10/P25 seemed to degrade EDTMP fastest as compared to the nano catalysts Au2/Y2/P25, Au2/K2/ P25 and P25 alone (Fig. 6A). We further determined the release of major breakdown products, i.e., IDMP, EABMP and AMPA for all photocatalytic treatments (Fig. 6B-D). Like the reference treatment, we found IDMP as major breakdown Fig. 3. TEM micrographs obtained for the Au catalysts series. A) Au2/P25, B) Au2/K2/P25, C) Au2/Na2/P25 and D) Au2/Y2/P25. Fig. 4. Histograms of particle size distribution for the Au catalysts series. A) Au2/P25, B) Au2/K2/P25, C) Au2/Na2/P25 and D) Au2/Y2/P25. R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 6 Fig. 5. Release of o-PO 4 3− of pre-screening tests with different NPs during photocatalysis of 100 mg L −1 EDTMP with different pH values. All catalysts doped with Na, Y or K are based on Au2 and P25. The refence experiment was run without catalysts. For the sake of simplicity, P25 has been omitted and Au2 is replaced by Au from the legend in this figure. Fig. 6. Influence of selected nano catalysts on the UV treatment of 100 mg L −1 EDTMP at pH value 7. A) Degradation of the target compound EDTMP. B) Release of the transformation product IDMP. C) Release of the transformation product EABMP. D) Release of the transformation product AMPA. All catalysts doped with Y or K are based on Au2 and P25. For the sake of simplicity, P25 has been omitted and Au2 is replaced by Au from the legend in this figure. R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 7 product for all treatments with nano catalysts (Fig. 6B). The highest IDMP release within the first 30 min of treatment was observed with the use of Au2/K10/P25, followed by Au2/Y5/P25. For the catalysts Au2/ Y2/P25, Au2/K2/P25 and P25, the release of IDMP was always below of the reference treatment without catalyst. In comparison, similar release of EABMP was found as compared to the reference for photocatalytic treatment with Au2/Y2/P25, Au2/K10/P25 and P25, while very low quantities of EABMP were determined for Au2/K2/P25 and Au2/Y5/ P25 (Fig. 6C). Regarding the release of AMPA, we determined highest release for the treatment with nano catalysts Au2/Y5/P25 and Au2/ K10/P25 and the lowest release for Au2/Y2/P25 and Au2/K2/P25 (Fig. 6D). We also measured the o-PO 4 3− release throughout the complete photocatalytic treatments and determined the highest o-PO 4 3− release for Au2/Y5/P25 resulting in 80.2 ±2.1 mg L −1 corresponding to 92.1 ± 7.5 % (Table 2). In comparison, the o-PO 4 3− release after 300 min UV irradiation without photocatalyst resulted in 41.3 ±0.4 mg L −1 o-PO 4 3− corresponding to 47.7 ±1.3 % respectively. Hence, the o-PO 4 3− release almost doubled for the nano catalyst Au2/Y5/P25. Based on the results of the LC/MS analyses, we calculated the balance gaps for phosphorus (P), carbon (C), and nitrogen (N) (Table 2). The P balances only showed a P gap of 15.9 ±1.0 % for Au2/K2/P25, while there was no gap for the other UV treatments. In contrast, the nano catalysts Au2/Y5/P25, Au2/ K2/P25 and Au2/K10/P25 showed the highest relative gaps in the C and N balances. 3.4. Influence of UV treatment on kinetics using selected NPs −Au2/Y/ P25 and Au2/K/P25 Additional photocatalytic treatments were carried out with the nano catalysts Au2/Y2/P25, Au2/Y5/P25, Au2/K2/P25 and P25 for determining the kinetic parameters. All treatments were performed for only 60 min and the time interval between measurements was set to 10 min to increase the resolution. Our results showed that both nano catalysts Au2/Y2/P25 and Au2/Y5/P25 resulted in fastest degradation of EDTMP while P25 resulted in the slowest degradation (Fig. 7A). Interestingly, the highest release of IDMP was achieved by Au2/Y2/P25 (Fig. 7B). For Au2/Y5/P25 and Au2/K2/P25, the IDMP release was almost similar. For P25, no IDMP release was detectable within the initial 30 min of UV treatment. In contrast to the high IDMP releases mediated by Au2/Y2/ P25, no EABMP was detectable within the first 20 min for Au2/Y2/P25 (Fig. 7C). Compared to all tested nano catalysts, we determined the highest EABMP release for P25. Overall, the measured AMPA release were always lowest for all tested nano catalysts (Fig. 7D). We determined pseudo first-order degradation rate constants (k UV ) and the corresponding half-lives for all four UV treatments of EDTMP with different nano photocatalysts (Table 3). The highest k UV and the lowest half-life for the UV treatment were found with nano photocatalyst Au2/Y5/P25, averaging 3.97 •10 −3 s −1 and 2.91 ±0.1 min, respectively. In contrast, we determined the lowest k UV and, thereby, the highest half-life for the UV treatment with P25. The k UV and corresponding half-life averaged 9.82 10 −4 s −1 and 11.79 ±0.7 min respectively, which was about four times higher as compared to the treatment with Au2/Y5/P25. 3.5. Influence of dosage of Au2/Y5/P25 during UV treatment of EDTMP Since Au2/Y5/P25 performed best as a photocatalyst for EDTMP, we tested its optimal dosage by performing three UV treatments with different dosages and analysed all samples by LC/MS. The results clearly demonstrated the influence of dosage on the degradation kinetics of EDTMP (Fig. 8). The highest dosage of 100 mg L −1 Au2/Y5/P25 resulted in the fastest degradation of EDTMP. Interestingly, the lowest dose of 25 mg L −1 Au2/Y5/P25 appeared to have a similar degradation kinetic of EDTMP only during the first 10 min of UV treatment and then almost no further EDTMP degradation. 3.6. Determination of reactive species by scavengers Finally, scavenger experiments were carried out to better understand the underlying mechanism by ROS during the photocatalysis of EDTMP with Au2/Y5/P25. Four different scavengers were tested and showed different suppression of the o-PO 4 3− release during the treatment (Fig. 9). The highest o-PO 4 3− suppression was observed for the scavenger i-PrOH, which resulted in an overall reduction of more than 21 mg L −1 o-PO 4 3− compared to the UV treatment without scavenger addition (Table 4). The scavenger i-PrOH usually scavenges •OH from the bulk. Interestingly, the scavenger KI, which is normally used to scavenge OH and electron holes (h + ) at the catalyst surface, resulted in enhanced o-PO 4 3− release rather than suppression. More than 47.6 mg L −1 o-PO 4 3− was released with KI. Based on the results of the scavenger experiments the following degradation mechanism for the photocatalysts Au2/Y5/P2 is proposed (Fig. 10). 4. Discussion 4.1. The influence of the dopants Au, Na, Y, and K on the photocatalytic activity of P25 towards EDTMP degradation In the past year, several publications highlighted the photochemical degradation of phosphonates. However, only a few scientists report the degradation of phosphonates based on photocatalysts while the majority report the use of phosphonates to stabilise photocatalytic NPs and/or metal–organic frameworks [53,54]. One reason may be the fact that the design and synthesis of specific photocatalysts with such a specific target compound represent a significant challenge, particularly when the catalyst should also be applicable in the visible light range to reduce the input energy. The presented results of the pre-screening have unambiguously demonstrated that P25 alone did not seem to facilitate the photocatalytic degradation of EDTMP. This may be attributed to an insufficient oxidation of EDTMP or an underestimation of released o-PO 4 3− . Bachinger and Kickelbick [53] showed that the organic majority of organophosphonates is photocatalytic degraded while the phosphates remained at the TiO 2 NP surface. It may be possible that comparable effects occurred during the photocatalytic treatment of EDTMP with P25. This could explain the consistently low o-PO 4 3− release observed compared with the reference experiment, particularly at higher pH values. It may also be assumed that the photocatalyst was consumed systematically over the treatment time. In consequence, the available quantity of P25 for the photocatalytic degradation of EDTMP may be not consistently constant. Such poisoning of the photocatalyst is an unfavourable reaction that can be mitigated by doping the surface with other semiconductors. The Table 2 Release of o-PO 4 3− and mass balance of photocatalytic treatment of 100 mg L −1 EDTMP (pH 7). o-PO 4 3− release o-PO 4 3− release P gap C gap N gap [mg L −1 ] [%] [%] [%] [%] Reference 41.3 ±0.4 47.7 ±1.3 0 54.7 ± 1.5 21.7 ± 0.6 Au2/Y2/ P25 76.5 ±1.0 88.1 ±3.6 0 74.6 ± 0.5 75.3 ± 2.1 Au2/Y5/ P25 80.2 ±2.1 92.1 ±7.5 0 92.1 ± 0.7 88.2 ± 1.4 Au2/K2/ P25 69.2 ±3.2 79.6 ±1.4 15.9 ± 1.0 96.1 ± 0.8 95.5 ± 0.9 Au2/K10/ P25 76.5 ±1.2 88.0 ±4.1 0 82.8 ± 1.7 87.1 ± 2.7 P25 57.2 ±1.0 65.9 ±3.7 0 69.4 ± 2.0 45.0 ± 1.9 R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 8 doping affects the reduction of the surface area available for the undesired deposition of o-PO 4 3− . Consequently, a greater quantity of catalyst is available for the entire treatment period. In the presented study, the combination of P25 as TiO 2 -support with the semiconductor Au was found to enhance the photodegrading effect on EDTMP, particularly at a pH value of 7. It may be assumed that Au protected the surface of P25 from absorption of free o-PO 4 3− . However, it is also well documented that Au doped NPs are employed in a multitude of applications due to their exceptional catalytic properties. In more detail, Kamat [55] demonstrated that the catalytic activity of semiconductors is enhanced when precious metals are applied to them, due to the reduction of the bandgap energy of TiO 2 by Au. The reduction results in a more robust interaction between Au and TiO 2 , thereby markedly enhancing the photocatalytic activity. This is due to the enhanced charge separation between the excited electron and the hole. The electrons are rapidly transported from TiO 2 to Au and are captured. This inhibits the recombination of electrons and h + on the TiO 2 surface, particularly at defect sites, as recently demonstrated [56]. As a result, the overall system energy dynamics and photocatalytic activity is improved [57]. Furthermore, the formation of new electronic structures results in a notable reduction in the band gap [56]. In addition, the surface plasmon resonance (SPR) by Au result in enhanced absorption and scattering of light, particularly within the visible spectrum. The broad absorption at 550 nm in the UV-DRS spectra is ascribed to the SPR effect of Au NPs that have been deposited on the TiO 2 surface [42]. In summary, the reduction in band gap observed in the NP Au/P25 (2.6 eV) in the presented study is attributed to the introduction of defect states in TiO 2 and the SPR effect of Au. These factors enhance visible light absorption and promote efficient charge separation, leading to an improvement in photocatalytic performance for deposition of EDTMP. In the presented study, the effect of further doping with alkaline metals either by Na, Y or K was also investigated. Doping with alkaline metals such as Na was recently described to improve the crystallinity and stabilise the anatase phase [58]. Therefore, enhanced photocatalytic degradation of organic compounds was reported [59]. Y dopants, for example, are commonly reported being used for water–gas shift reactions and/or for CO 2 methanation. Successful doping with Y for photocatalytic application was also recently reported [60]. Also, Kdoped on TiO 2 has been tested for, for example, air pollutant treatments [61]. It is important to note, however, that in the bimetallic-doped TiO 2 system (in which Au is either doped with Na, Y or K on TiO 2 ), the photocatalytic activity under visible light is enhanced due to a narrower band gap, which allows for more efficient visible light absorption. Furthermore, the SPR effect of AuNPs is also of significance, whereby AuNPs absorb visible light, generate charge carriers and inject electrons into the Na-, Yor K-doped TiO 2 [42]. In the case of bimetallic-doped TiO 2 , Au exhibits a distinct behaviour. Rather than accepting electrons from TiO 2 as observed in Au/TiO 2 , it acts as an electron donor, transferring electrons to the conduction band of Na-, Yor K-doped TiO 2 [56]. This results in increased photocatalytic activity due to the synergistic effects of both processes. The results presented in the pre-screening of this study clearly indicate that the introduction of further doping resulted in an increase in the band gaps, which suggests that these dopants are effective in suppressing the defect energy levels that are characteristic of Au/TiO 2 systems. The three dopants, Na, Y and K, influenced the degradation of EDTMP distinctly, whereas the Y-doped catalyst showed in most cases the most Fig. 7. Influence selected nano particles on the degradation kinetics of UV treated 100 mg L −1 EDTMP at pH 7. A) Degradation of the target compound EDTMP. B) Release of the transformation product IDMP. C) Release of the transformation product EABMP. D) Release of the transformation product AMPA. All catalysts doped with Y or K are based on Au2 and P25. For the sake of simplicity, P25 has been omitted and Au2 is replaced by Au from the legend in this figure. Table 3 Degradation rate constants and half-life of EDTMP. Catalyst k uv [s −1 ] Half-life [min] RSS* Au2/Y2/P25 2.92 E-3 ±2.0 E-4 3.97 ±0.3 8.4 ±5.7 Au2/Y5/P25 3.97 E-3 ±1.9 E-4 2.91 ±0.1 54.4 ±34.0 Au2/K2/P25 1.94 E-3 ±2.2 E-4 6.02 ±0.7 60.7 ±31.8 P25 9.82 E-4 ±5.8 E-5 11.79 ±0.7 54.1 ±26.6 *RSS −Residual sum of squares. R. Riedel et al. Chemical Engineering Journal 506 (2025) 160109 9