Effect of Doping TiO2 NPs with Lanthanides (La, Ce and Eu) on the Adsorption and Photodegradation of Cyanide—A Comparative Study

Free cyanide is a highly dangerous compound for health and the environment, so treatment of cyanide-contaminated water is extremely important. In the present study, TiO2, La/TiO2, Ce/TiO2, and Eu/TiO2 nanoparticles were synthesized to assess their ability to remove free cyanide from aqueous solutions. Nanoparticles synthesized through the sol–gel method were characterized by X-ray powder diffractometry (XRD), scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), Fourier-transformed infrared spectroscopy (FTIR), diffuse reflectance spectroscopy (DRS), and specific surface area (SSA). Langmuir and Freundlich isotherm models were utilized to fit the adsorption equilibrium experimental data, and pseudo-first-order, pseudo-second-order, and intraparticle diffusion models were used to fit the adsorption kinetics experimental data. Cyanide photodegradation and the effect of reactive oxygen species (ROS) on the photocatalytic process were investigated under simulated solar light. Finally, reuse of the nanoparticles in five consecutive treatment cycles was determined. The results showed that La/TiO2 has the highest percentage of cyanide removal (98%), followed by Ce/TiO2 (92%), Eu/TiO2 (90%), and TiO2 (88%). From these results, it is suggested that La, Ce, and Eu dopants can improve the properties of TiO2 as well as its ability to remove cyanide species from aqueous solutions.


Introduction
Cyanides are highly toxic chemical compounds that can cause severe harm to health and the environment [1]. The most common cyanide compounds in the environment are present both in their free form, which comprises the cyanide ion itself (CN − ) and hydrogen cyanide (HCN), and as water-soluble inorganic salts, including sodium cyanide (NaCN) and potassium cyanide (KCN) [2]. Some cyanide compounds are naturally produced by microorganisms, although they can also be found in plants and some foods, as well as in low concentrations in soil and water. On the other hand, generation of effluents by industries is the main source of release into the environment of cyanide complexes, which are formed by union of cyanide ions with metal ions [3].
All forms of cyanide can be lethal at high levels of exposure; however, free cyanide (HCN + CN − ) is the deadliest form of all [4]. This form of cyanide in a solution can be removed and/or transformed to fewer toxic forms through various physicochemical methods, such as acidification, photocatalysis, adsorption, coagulation, reverse osmosis, precipitation, filtration, and AVR (acidification, volatilization, and reneutralization) process [5,6]. Among these methods, adsorption and photocatalysis have shown to be two promising methodologies due to their high efficiency, simplicity of design, ease of handling, and low operating cost [7,8]. The adsorption process can be carried out by different mechanisms,

Characterization of the Nanoparticles
Characterization of the nanoparticles was performed using the methodology described in our previous study [36]. For X-ray diffraction (XRD) measurements, a Bruker-AXS D8-Discover diffractometer (Bruker AXS, Karlsruhe, Germany) was used. For specific surface area (SSA) determination, ChemiSorb 2720 equipment (Micromeritics, Norcross, GA, USA) was used. Micrographs of the synthesized samples were obtained by field effect scanning electron microscopy (SEM) on a Zeiss Gemini ULTRA plus electron microscope (Carl Zeiss AG, Ober-kochen, Germany). A Phenom ProX (Phenom World BV, Eindhoven, The Netherlands) was used to achieve energy-dispersive X-ray (EDS) spectra. In the present study, quantification of the elements in the synthesized nanoparticles was also performed by inductively coupled plasma (ICP) optical emission spectrometry (OES) in an Optima 8000 ICP-OES Spectrometer (PerkinElmer, Inc., Waltham, MA, USA). FTIR spectra were recorded on a PerkinElmer GX2000-FTIR Spectrometer (PerkinElmer, Inc., Waltham, MA, USA). The diffuse UV-vis reflectance spectra (DRS) were obtained using a Phenom DesNicolet Evolution 201/220 Thermo UV-vis spectrophotometer (ThermoFisher, Waltham, MA, USA). To simulate solar light and evaluate the photoactivity of the nanoparticles, a solar box AT-LAS, SUNTEST CPS+ (Atlas Material Testing Technology, Mount Prospect, IL, USA) was used. Finally, the amount of cyanide remaining in the solutions was quantified using a Jenway 7350 spectrophotometer (Cole-Parmer, Staffordshire, UK). IBM SPSS (version 25.0; statistic software for Windows; IBM Corp.; Armonk, NY, USA, 2017) was used for the ANOVA analysis. The crystalline phases were recognized using the ICDD database (International Center for Diffraction Data, version 2018). Finally, the Chemisoft TPx system (version 1.03; data analysis software; Micromeritics, Norcross, GA, USA, 2011) was used to calculate the SSA by the single-point method using the Brunauer-Emmet-Teller (BET) equation.

Adsorption Studies
Cyanide adsorption experiments from KCN aqueous solutions were designed to evaluate the following aspects: (a) the effect of pH on cyanide adsorption, (b) maximum cyanide adsorption capacity, (c) adsorption thermodynamics and (d) the kinetic behavior of cyanide adsorption. The data obtained from the adsorption experiments were fitted to isothermal and kinetic models using the least squares nonlinear regression method [37]. Experiments were performed with all nanoparticles using the methodology described in our previous study [36]. Cyanide species adsorption experiments were performed using a batch method at room temperature and keeping the pH of the solutions at 7.0 ± 0.1 by adding 0.1 M solutions of hydrochloric acid or sodium hydroxide. The amount of catalyst used in all the experiments was 0.2 g L −1 . The maximum cyanide adsorption capacity was investigated by varying the concentration of 500 mL of KCN solution from 0.20 to 40 mg L −1 . The adsorption thermodynamics and kinetic behavior of cyanide adsorption were investigated using 500 mL of water containing 20 mg L −1 of KCN [38]. The alkaline picrate analytical method was used to quantify total cyanide in aqueous solutions [39][40][41]. To accomplish this, first, the alkaline picrate solution (PAS) was prepared, consisting of 1 g of picric acid ((O 2 N) 3 C 6 H 2 OH) and 5 g of sodium carbonate (Na 2 CO 3 ) dissolved in 200 mL of HPLC water. A volume of 4 mL of this solution was added to 1 mL of cyanide or blank solution (HPLC water) contained in a test tube. This tube was incubated for 5 min in a 95 • C water bath. After this time, the absorbance at 490 nm was measured using a UV-vis spectrophotometer. The concentration of the solutions was determined based on the previously prepared calibration curve (R 2 = 0.9996) according to the Lambert-Beer Law.
The tests were carried out in triplicate and the results were expressed as the average of Nanomaterials 2023, 13, 1068 5 of 33 three repetitions [38]. The amount of cyanide adsorbed (q e ) on the nanoparticles expressed in mg g −1 was estimated using the following equation [42]: where C 0 and C e are expressed in mg L −1 and correspond to the initial and equilibrium concentration, respectively. The mass (w) of the adsorbent is expressed in grams (g), and the volume (v) of the solution is expressed in liters (L). The equilibrium total cyanide adsorption was assessed based on the Langmuir and Freundlich isotherm models. The expression of the Langmuir isotherm model can be represented using the following equation [43]: C e q e = 1 K L q max + C e q max (2) where q max is expressed in mg g −1 and corresponds to the maximum monolayer adsorption, K L is expressed in L mg −1 and corresponds to the equilibrium Langmuir constant related to the adsorption energy, and C e is expressed in mg L −1 and corresponds to the concentration of solute at equilibrium. Additionally, the R L separation factor values, which offer an idea of the adsorption characteristics, can be represented using the following equation [43]: The suitability of the adsorption treatment based on the R L value is given as follows: 0 < R L < 1; it denotes suitable adsorption, R L > 1 unsuitable adsorption, R L = 0 irreversible adsorption, and R L = 1 linear adsorption [43].
Moreover, the Freundlich isotherm model can be represented by the following equation [43]: q e = K F C 1 /n e (4) where K F is expressed in L mg −1 and corresponds to the Freundlich constant, which specifies the adsorption affinity of the adsorbents, and 1/n is another constant that corresponds to the adsorption intensity. For favorable adsorption, the value of the constant n should be in the range of 1 to 10 [44]. For the thermodynamic studies, the experimental data were fitted according to the parameters of the thermodynamic laws described by Gibbs free energy (∆G 0 , kJ mol −1 ), enthalpy (∆H 0 , kJ mol −1 ), and entropy (∆S 0 , kJ mol −1 K −1 ) conventionally used, represented by the following equation [45] ∆G 0 = −RTln k C (5) The relationship between ∆G 0 , ∆H 0 , and ∆S 0 , is obtained by the well-known van 't Hoff equation [45] ln where k L (L mg −1 ) is the Langmuir constant and could be achieved as a dimensionless parameter. T is the absolute temperature (K), and R is the universal gas constant (8.314 J mol −1 K −1 ). k C is achieved as a dimensionless parameter by multiplying k L by a molecular weight of an adsorbate (M w , g mol −1 ) and then by factors 1000 and 55.5, which is the number of moles of pure water contained in a liter, described by the following equation [46].
The absorption kinetics were estimated by applying pseudo-first-order and pseudosecond-order models, as well as intraparticle diffusion, external-film diffusion, and internalpore diffusion models. The pseudo-first-order kinetic model is represented by the following equation [44]: ln q e − q t = ln q e − k 1 t where k 1 is the rate constant (min −1 ) and q e and q t are the cyanide adsorbed per unit weight (mg g −1 ) at equilibrium and at any time t, respectively. The pseudo-second-order kinetic is represented by the following equation [44]: where k 2 is the pseudo-second-order rate constant (g mg −1 min −1 ). Finally, to obtain an appropriate understanding of the cyanide adsorption mechanism on the nanoparticles surface, the rate-limiting step in the adsorption process was also estimated. The intraparticle diffusion model supposes that intraparticle diffusion is generally the rate-controlling step in well-mixed solutions. The intraparticle-diffusion model is represented by the following equation [44]: where k 3 (mg g −1 min −1/2 ) is the intraparticle diffusion rate constant and A (mg g −1 ) is a constant indicating the width of the boundary layer; that is, the larger the value of A, the greater the boundary layer effect. When the q t plot against the square root of time shows multilinearity, it means that the diffusion occurs in several steps during the process.
The internal pore diffusion model was also utilized in this study to explain the kinetic adsorption data. When the adsorption rate is controlled by particle diffusion, the adsorption rate is represented using the following equation [44]: On the other hand, regarding external-film-diffusion controls, the adsorption rate is expressed by the following equation [44]: where q e and q t represent the amount of solute that is taken up by the adsorbent phase at equilibrium and at a specific time t (mg g −1 ), respectively. The ion concentration in the solution and adsorbent are donated by C s (mg L −1 ) and C z (mg kg −1 ), respectively. The variable t refers to the contact time (min), while r represents the average radius of the adsorbent particles (1 × 10 −7 m). The film thickness around the adsorbent particles is denoted by h and is assumed to be 10 −6 m in poorly stirred solutions. Finally, the diffusion coefficient in the adsorbent phase is referred to as D p (m 2 min −1 ), while D f (m 2 min −1 ) represents the diffusion in the film phase surrounding the adsorbent particles.

Photodegradation Studies
Heterogeneous photocatalysis experiments were performed using the methodology described in our previous study [36]. These experiments were performed in an air-cooled solar box. The batch method was used, keeping the pH of the solution at 7.0 ± 0.1 by adding 0.1 M solutions of hydrochloric acid or sodium hydroxide. Typically, 0.2 g L −1 of nanoparticles were magnetically stirred in 500 mL of water containing 20 mg L −1 of KCN [38]. The photodegradation rate of cyanide species in the heterogeneous photocatalytic systems exposed to simulated solar light for the TO, La/TO, Ce/TO, and Eu/TO nanoparticles was followed by the Langmuir-Hinshelwood equation [47], which can be represented by the following equation [48]: where k (min −1 ) is the actual rate constant, K is the adsorption constant of the substrate on the nanoparticles, C 0 (mg L −1 ) represents the initial concentration of the substrate, and C t (mg L −1 ) represents the concentration at a specific time t (min), and k app (min −1 ) is the apparent rate constant. Plotting ln(C 0 /C t ) against time t provides the apparent rate constant (k app ) for substrate degradation from the slope of the curve fit line, and the intercept is equal to zero.
To determine the influence of several reactive oxygen species (ROS) on the reaction system, the radical quenching experiment was performed using isopropanol (i-POH), ethanol (EtOH), and p-benzoquinone (p-BQ). These radical quenchers inhibit h + , • OH, and • O 2 − , respectively. The experiments were carried out following the methodology described for the heterogeneous photocatalysis experiments but each time incorporating the radical quenchers (0.5 mM) in the reaction system [49].

Reuse of Nanoparticles
Finally, in order to verify the reusability of the nanoparticles in cyanide photodegradation, a recycling experiment was designed using the methodology described in our previous study [36]. The recycling experiment was performed for five consecutive cycles. At the end of each treatment cycle, the suspensions were precipitated by leaving them to stand for at least 1 h. After this time, the supernatant liquid was removed and the nanoparticles were carefully washed three times with HPLC water, avoiding loss of solid material. In each treatment cycle, 100 mL of fresh KCN solution (20 mg L −1 ) was used. The amount of nanoparticles used in this experiment was 0.2 g L −1 [38].  1 2), respectively. The TiO 2 crystal structure was indexed to tetragonal phase with unit cell parameters a = b = 3.79 Å and c = 9.51 Å, α = β = γ = 90 • , unit cell volume = 136. 30  The crystal sizes of the TO, La/TO, Ce/TO, and Eu/TO nanoparticles were calculated considering the most intense diffraction peak. The Scherrer equation (Equation (14)) was used for the respective calculation [54] A = Kλ β cosθ (14) where A is the size of the crystals expressed in nm, λ = 0.15406 nm and K = 0.89 represent the wavelength of the X-ray beam and the shape factor, respectively, θ represents the Bragg angle, while β represents the full width at half peak height maximum (FWHM) of the X-ray diffraction peak. On the other hand, the specific surface area (SSA) was evaluated by the single-point BET (Brunauer-Emmet-Teller) method. The SSA of the pristine TiO 2 (TO) was measured as 88 m 2 g −1 , which was slightly lower than that of the La/TO, Ce/TO, and Eu/TO nanoparticles, with values of 126, 104, and 96 m 2 g −1 , respectively. From these results, it is suggested that elements lanthanum (La), cerium (Ce), and europium (Eu) have a specific effect on inhibition of growth of TiO 2 crystallites, stabilization of this oxide in its anatase phase, as well as increasing specific surface area (SSA).

SEM and EDS Analysis
Figure 3a-d displays the SEM photomicrographs of TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. In these figures, it is evident that all the nanoparticles appear almost spherical and are highly agglomerated. These results are in agreement with those informed by other authors [60]. Doped nanoparticles are smaller than pristine TO nanoparticles. The average size of the TO nanoparticles was 32 nm, in contrast to La/TO, Ce/TO, and Eu/TO nanoparticles, which had mean sizes of 26, 27, and 28 nm, respectively. The size of the synthesized nanoparticles was measured using ImageJ2, which is public domain software for scientific image processing and analysis [61,62]. The sizes of the nanoparticles are within the range of values reported by other authors [63]. The FTIR spectrum of TiO 2 reveals the existence of a broad band in the region of 540-900 cm −1 , which is assigned to Ti-O and Ti-O-Ti stretching vibrations [55,56]. After doping, the main change was identified in this region, presenting new unresolved bands around the 500-600 cm −1 region that could be attributed to Ln-O stretching modes of vibrations. The results obtained in this analysis agree with those reported in the literature [57,58].
On the other hand, the spectra in Figure 2 show that there is a tendency to increase the vibrational frequencies as the ionic radius of the doping elements decreases. Evidence from the literature indicates that Ln-O single-bond stretching frequencies and Ln-O single-bond stretching force constants show a linear correlation with inverse ionic radii [59].

SEM and EDS Analysis
Figure 3a-d displays the SEM photomicrographs of TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. In these figures, it is evident that all the nanoparticles appear almost spherical and are highly agglomerated. These results are in agreement with those informed by other authors [60]. Doped nanoparticles are smaller than pristine TO nanoparticles. The average size of the TO nanoparticles was 32 nm, in contrast to La/TO, Ce/TO, and Eu/TO nanoparticles, which had mean sizes of 26, 27, and 28 nm, respectively. The size of the synthesized nanoparticles was measured using ImageJ2, which is public domain software for scientific image processing and analysis [61,62]. The sizes of the nanoparticles are within the range of values reported by other authors [63].
The SEM images presented in this study support the results of the XRD analysis ( Figure 1), demonstrating that doping elements lanthanum, cerium, and europium are effective in preventing crystallite growth and stabilizing the structure of the TiO 2 semiconductor.
Likewise, Figure 3a-d displays the EDS spectra of TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. These spectra support the results of the FTIR analysis ( Figure 2), thus confirming incorporation of lanthanide elements La (1.32 wt%), Ce (1.34 wt%), and Eu (1.28 wt%) on the TiO 2 structure.

Optical and Photoelectric Properties
The optical absorption capacity of the synthesized nanoparticles was determined by UV-visible (UV-vis) DRS at room temperature in the range of 200 to 700 nm. Figure 4a shows the DRS UV-vis spectrum of TO, La/TO, Ce/TO, and Eu/TO nanoparticles. Comparatively, the absorption behavior in the visible light spectrum (around 400 nm) was slightly enhanced for La/TO, Ce/TO, and Eu/TO. Due to this bathochromic shift, it is suggested that, compared to TO, doped nanoparticles have improved response to visible light.
The Tauc formula was used to estimate the direct bandgap energy (E g ) for the nanoparticles using the following equation [64]: (15) where α is the absorption coefficient, hv is photon energy, computed by hv = 1240/λ, A is a constant that is correlated with the absorption brink width parameter, and n is the optical transition phase. The value of n depends on the semiconductor transitions, that is, n = 1 for direct transitions and n = 4 for indirect transitions [64].

SEM and EDS Analysis
Figure 3a-d displays the SEM photomicrographs of TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. In these figures, it is evident that all the nanoparticles appear almost spherical and are highly agglomerated. These results are in agreement with those informed by other authors [60]. Doped nanoparticles are smaller than pristine TO nanoparticles. The average size of the TO nanoparticles was 32 nm, in contrast to La/TO, Ce/TO, and Eu/TO nanoparticles, which had mean sizes of 26, 27, and 28 nm, respectively. The size of the synthesized nanoparticles was measured using ImageJ2, which is public domain software for scientific image processing and analysis [61,62]. The sizes of the nanoparticles are within the range of values reported by other authors [63].   The SEM images presented in this study support the results of the XRD analysis (Figure 1), demonstrating that doping elements lanthanum, cerium, and europium are effective in preventing crystallite growth and stabilizing the structure of the TiO2 semiconductor.
Likewise, Figure 3a-d displays the EDS spectra of TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. These spectra support the results of the FTIR analysis (Figure 2), thus confirming incorporation of lanthanide elements La (1.32 wt%), Ce (1.34 wt%), and Eu (1.28 wt%) on the TiO2 structure.

Optical and Photoelectric Properties
The optical absorption capacity of the synthesized nanoparticles was determined by UV-visible (UV-vis) DRS at room temperature in the range of 200 to 700 nm. Figure 4a shows the DRS UV-vis spectrum of TO, La/TO, Ce/TO, and Eu/TO nanoparticles. Comparatively, the absorption behavior in the visible light spectrum (around 400 nm) was slightly enhanced for La/TO, Ce/TO, and Eu/TO. Due to this bathochromic shift, it is suggested that, compared to TO, doped nanoparticles have improved response to visible light.
The Tauc formula was used to estimate the direct bandgap energy (Eg) for the nanoparticles using the following equation [64]: αhv = A hv − E (15) where α is the absorption coefficient, hv is photon energy, computed by hv = 1240/λ, A is a constant that is correlated with the absorption brink width parameter, and n is the optical transition phase. The value of n depends on the semiconductor transitions, that is, n = 1 for direct transitions and n = 4 for indirect transitions [64].  The relationship between (αhv) 2 and photon energy (hv) shown in Figure 4b enabled estimating the direct bandgap energy (Eg) [65]. Based on the data presented in this figure, the direct Eg values achieved from the intersections of the straight line with the energy axis [66] were 3.20, 3.10, 3.14, and 3.16 eV for TO, La/TO, Ce/TO, and Eu/TO, respectively.

Effect of Nanoparticles Composition
The effect of the composition of TO, La/TO, Ce/TO, and Eu/TO nanoparticles on their total cyanide removal capacity qe (mg g −1 ) was evaluated in the absence of light (30 min) The relationship between (αhv) 2 and photon energy (hv) shown in Figure 4b enabled estimating the direct bandgap energy (E g ) [65]. Based on the data presented in this figure, the direct E g values achieved from the intersections of the straight line with the energy axis [66] were 3.20, 3.10, 3.14, and 3.16 eV for TO, La/TO, Ce/TO, and Eu/TO, respectively.

Effect of Nanoparticles Composition
The effect of the composition of TO, La/TO, Ce/TO, and Eu/TO nanoparticles on their total cyanide removal capacity q e (mg g −1 ) was evaluated in the absence of light (30 min) and then under simulated solar light. The results of this experiment are shown in Figure 5. For all the nanoparticles, it was found that, after 90 min of radiation exposure and under the test conditions, it is possible to achieve maximum removal. This time was denoted as the equilibrium time t e (min). The findings of the analysis of variance (ANOVA) carried out at te (min) are summarized in Table 1. The values in the same column that have been assigned different letters (a-g) are significantly different from each other, with a high level of statistical significance (p < 0.01). Based on these results, it is suggested that, for each dopant, the best composition was 1.0 wt.%; therefore, the La/TO (1%), Ce/TO (1%), and Eu/TO (1%) nanoparticles were selected to perform cyanide removal tests and evaluate the effectiveness of doping in relation to non-doped TiO2.  The findings of the analysis of variance (ANOVA) carried out at t e (min) are summarized in Table 1. The values in the same column that have been assigned different letters (a-g) are significantly different from each other, with a high level of statistical significance (p < 0.01). Based on these results, it is suggested that, for each dopant, the best composition was 1.0 wt.%; therefore, the La/TO (1%), Ce/TO (1%), and Eu/TO (1%) nanoparticles were selected to perform cyanide removal tests and evaluate the effectiveness of doping in relation to non-doped TiO 2 . Different letters (a-g) indicate different groups with a high level of statistical significance (p < 0.01). The means for the groups in the homogeneous subsets are displayed. * Use the sample size of the harmonic mean = 3.0. Figure 6 shows the results of the cyanide adsorption experiment as a function of the pH of the solutions. Total cyanide adsorption capacity (HCN + CN − ) remained unchanged in the pH range 9 to 12 for pristine TO (pH PZC = 6.9) and for doped nanoparticles (pH PZC = 7.1). The point of zero charge (PZC) corresponds to the pH value at which the net surface charge of the solid is equal to zero under certain conditions of pressure, temperature, and composition of the aqueous solution [67]. The point of zero charge (PZC) for all nanoparticles was determined at room temperature (21 ± 2 • C) using the pH drift procedure (∆pH = pH f − pH i = 0). The experiment was carried out in 50 mL tubes, in which 25 mL of a 0.1 M NaCl solution and 0.1 g of solid sample were placed. The pH values of the solutions contained in the tubes were adjusted to values between 3 and 12 using 0.1 M solutions of HCl or NaOH. These initial pH values were designated as pH i . The tubes were shaken for 24 h at 230 rpm. After this time, the final pH of the supernatant liquid in each tube was measured and designated as pH f . The PZC was determined from the graph of ∆pH (∆pH = pH f -pH i ) vs. pH i . The procedure was repeated for all nanoparticle samples using 0.01 and 0.05 M NaCl solutions. All experiments were carried out in triplicate, and, for each sample, the average pH PZC value was reported [68]. nanoparticles was determined at room temperature (21 ± 2 °C) using the pH drift procedure (ΔpH = pHf − pHi = 0). The experiment was carried out in 50 mL tubes, in which 25 mL of a 0.1 M NaCl solution and 0.1 g of solid sample were placed. The pH values of the solutions contained in the tubes were adjusted to values between 3 and 12 using 0.1 M solutions of HCl or NaOH. These initial pH values were designated as pHi. The tubes were shaken for 24 h at 230 rpm. After this time, the final pH of the supernatant liquid in each tube was measured and designated as pHf. The PZC was determined from the graph of ΔpH (ΔpH = pHf - pHi) vs. pHi. The procedure was repeated for all nanoparticle samples using 0.01 and 0.05 M NaCl solutions. All experiments were carried out in triplicate, and, for each sample, the average pHPZC value was reported [68]. In this study, the pHPZC of all nanoparticles was on average 7.0. This implies that, at pH > pHPZC = 7.0, the surface of the nanoparticles has a negative charge and, at pH < pHPZC = 7.0, it has a positive charge. The ability of ions to be adsorbed on a surface from a solution is greatly influenced by the pH of the solution. This is because pH affects both the surface charge of the ion adsorbent and the extent to which the ions ionize and exist in different forms. Therefore, to better understand the adsorption mechanisms of cyanide species on the surface of nanoparticles, it is important to consider the pH-dependent distribution of cyanide species in solution (Equation (16)) [7].

Effect of pH on Cyanide Adsorption
Inorganic cyanides, such as KCN and NaCN, are weak acids and hydrolyze to form hydrocyanic acid (HCN), which has a pKa value of 9.4. At pH < 9.4, HCN is the predominant species in the aqueous solution, while, at pH > 9.4, HCN dissociates into H + and CN − ions [69]. CN − ions have nucleophilic characteristics and experience a strong electrostatic attraction in solutions at pH > 9.0, which is why several authors have informed higher cyanide adsorption at pH values between 9 and 11 [42]. However, other authors have also informed that inorganic cyanides can easily dissociate to form HCN and CN − species at a neutral pH of 7.0 [70]. In fact, several investigations related to cyanide adsorption have been carried out at pH values around 7.0 (pH < pKa), suggesting that the high adsorption of HCN species is based on interactions with the existing Lewis acids sites on the adsorbent surface [71,72]. Consequently, the adsorption experiments were performed at a pH In this study, the pH PZC of all nanoparticles was on average 7.0. This implies that, at pH > pH PZC = 7.0, the surface of the nanoparticles has a negative charge and, at pH < pH PZC = 7.0, it has a positive charge.
The ability of ions to be adsorbed on a surface from a solution is greatly influenced by the pH of the solution. This is because pH affects both the surface charge of the ion adsorbent and the extent to which the ions ionize and exist in different forms. Therefore, to better understand the adsorption mechanisms of cyanide species on the surface of nanoparticles, it is important to consider the pH-dependent distribution of cyanide species in solution (Equation (16)) [7].
Inorganic cyanides, such as KCN and NaCN, are weak acids and hydrolyze to form hydrocyanic acid (HCN), which has a pKa value of 9.4. At pH < 9.4, HCN is the predominant species in the aqueous solution, while, at pH > 9.4, HCN dissociates into H + and CN − ions [69]. CN − ions have nucleophilic characteristics and experience a strong electrostatic attraction in solutions at pH > 9.0, which is why several authors have informed higher cyanide adsorption at pH values between 9 and 11 [42]. However, other authors have also informed that inorganic cyanides can easily dissociate to form HCN and CN − species at a neutral pH of 7.0 [70]. In fact, several investigations related to cyanide adsorption have been carried out at pH values around 7.0 (pH < pKa), suggesting that the high adsorption of HCN species is based on interactions with the existing Lewis acids sites on the adsorbent surface [71,72]. Consequently, the adsorption experiments were performed at a pH of 7.0.

Maximum Cyanide Adsorption Capacity
In this study, the Langmuir and Freundlich isotherms were examined as equilibrium models that rely on the initial concentration. The Langmuir isotherm represents a credible theoretical framework for the process of adsorption occurring on a uniform and completely homogeneous surface, whereby a finite quantity of identical and specific sites are available for adsorption. In this model, there is minimal interaction among the molecules. The Freundlich equation is a non-theoretical formula that does not assume uniformity in the energy of surface sites, allows for unlimited adsorption capacity, and corresponds to an exponential distribution of active sites that reflects a heterogeneous surface. Figure 7 shows the cyanide adsorption isotherms of TO, La/TO, Ce/TO, and Eu/TO nanoparticles. This figure clearly shows that, for all nanoparticles, the Langmuir model is a more suitable description of the adsorption process compared to the Freundlich model. In this study, the Langmuir and Freundlich isotherms were examined as equilibrium models that rely on the initial concentration. The Langmuir isotherm represents a credible theoretical framework for the process of adsorption occurring on a uniform and completely homogeneous surface, whereby a finite quantity of identical and specific sites are available for adsorption. In this model, there is minimal interaction among the molecules. The Freundlich equation is a non-theoretical formula that does not assume uniformity in the energy of surface sites, allows for unlimited adsorption capacity, and corresponds to an exponential distribution of active sites that reflects a heterogeneous surface. Figure 7 shows the cyanide adsorption isotherms of TO, La/TO, Ce/TO, and Eu/TO nanoparticles. This figure clearly shows that, for all nanoparticles, the Langmuir model is a more suitable description of the adsorption process compared to the Freundlich model. The values estimated at different temperatures for the constants of the Langmuir and Freundlich models are given in Table 2. In this table, it is evident that the values of the RL separation factor or equilibrium parameter were in the range of 0 to 1, while the values of the coefficient n, which represents the intensity of adsorption, were in the range of 1 to 10. Consequently, it is suggested that adsorption of cyanide species on the surface of all materials was satisfactory. The values estimated at different temperatures for the constants of the Langmuir and Freundlich models are given in Table 2. In this table, it is evident that the values of the R L separation factor or equilibrium parameter were in the range of 0 to 1, while the values of the coefficient n, which represents the intensity of adsorption, were in the range of 1 to 10. Consequently, it is suggested that adsorption of cyanide species on the surface of all materials was satisfactory.

Adsorption Thermodynamics
The thermodynamic parameters provide information about the spontaneity and possibility of a process. To determine these parameters, namely Gibbs free energy change (∆G • ), enthalpy change (∆H • ), and surface entropy change (∆S • ), the equilibrium constant was measured at different temperatures, as shown in Figure 8.

Adsorption Thermodynamics
The thermodynamic parameters provide information about the spontaneity and possibility of a process. To determine these parameters, namely Gibbs free energy change (∆G°), enthalpy change (∆H°), and surface entropy change (∆S°), the equilibrium constant was measured at different temperatures, as shown in Figure 8. The results of the thermodynamic parameters are shown in Table 3. ∆G° indicates the degree of spontaneity of the process; negative values reflect higher adsorption favorability. Likewise, negative ΔH° values reflect that the process is exothermic, while negative ΔS° values indicate a decrease in randomness at the solution-solid interface during adsorption. The results of the thermodynamic parameters are shown in Table 3. ∆G • indicates the degree of spontaneity of the process; negative values reflect higher adsorption favorability. Likewise, negative ∆H • values reflect that the process is exothermic, while negative ∆S • values indicate a decrease in randomness at the solution-solid interface during adsorption.

Kinetic Behavior of Cyanide Adsorption
To design and evaluate adsorbents for adsorption, it is important to determine the rate of the time-dependent process. Two models, the pseudo-first-order (Lagergren) and pseudo-second-order (Ho) models, were used to describe the adsorption kinetics, which are commonly used as simplified models. Both models, as shown in Figure 9, exhibit a rapid initial adsorption stage followed by a plateau stage. Table 4 indicates that the correlation coefficient of the pseudo-second-order model is higher than that of the pseudo-first-order model, suggesting a chemisorption process, according to previous literature [9]. To design and evaluate adsorbents for adsorption, it is important to determine the rate of the time-dependent process. Two models, the pseudo-first-order (Lagergren) and pseudo-second-order (Ho) models, were used to describe the adsorption kinetics, which are commonly used as simplified models. Both models, as shown in Figure 9, exhibit a rapid initial adsorption stage followed by a plateau stage. Table 4 indicates that the correlation coefficient of the pseudo-second-order model is higher than that of the pseudo-firstorder model, suggesting a chemisorption process, according to previous literature [9].  On the other hand, the intraparticle diffusion model was utilized to explain the adsorption rate, which is dependent on the rate at which the cyanide species transfer from the aqueous solution to the adsorption sites on the nanoparticles. Figure 10 shows the variation in the q t (mg g −1 ) curves as a function of time (t 1/2 ) for the TO, La/TO, Ce/TO, and Eu/ZTO nanoparticles. Table 4 shows the total cyanide adsorption kinetic parameters estimated in this study for all nanoparticles.

Kinetics of Cyanide Photodegradation
TiO 2 is a widely used photocatalyst to efficiently degrade organic compounds due to its strong oxidizing capacity, which is generated when subjected to the action of light.
In this study, the photoactivity of pristine TiO 2 nanoparticles and La/TiO 2 , Ce/TiO 2 , and Eu/TiO 2 nanoparticles was tested through model cyanide photodegradation reaction in aqueous solutions under simulated solar radiation. On the other hand, the intraparticle diffusion model was utilized to explain the adsorption rate, which is dependent on the rate at which the cyanide species transfer from the aqueous solution to the adsorption sites on the nanoparticles. Figure 10 shows the variation in the qt (mg g −1 ) curves as a function of time (t 1/2 ) for the TO, La/TO, Ce/TO, and Eu/ZTO nanoparticles.  Table 4 shows the total cyanide adsorption kinetic parameters estimated in this study for all nanoparticles.  The Langmuir-Hinshelwood equation showed a linear correlation between ln(C 0 /C t ) and t, confirming that the photocatalytic degradation reaction proceeds via a pseudofirst-order reaction. The calculated apparent rate constants (k app ) were calculated to be 0.016, 0.030, 0.018, and 0.017 min −1 for TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. These results are in agreement with those reported by other authors [48,73]. The results obtained in the photocatalytic degradation test are shown in Figure 11. From this figure, the maximum percentage of cyanide degradation for all nanoparticles is reached around the first 90 min, after which photodegradation becomes almost constant. Evidence from the literature suggests that there is a limit to the efficacy of photocatalysis for complete degradation of some pollutants [74]. In this study, maximum efficiency was reached by La/TO (96.8%), followed by Ce/TO (88.9%), Eu/TO (86.5%), and TO (83.9%). and t, confirming that the photocatalytic degradation reaction proceeds via a pseudo-firstorder reaction. The calculated apparent rate constants (kapp) were calculated to be 0.016, 0.030, 0.018, and 0.017 min −1 for TO, La/TO, Ce/TO, and Eu/TO nanoparticles, respectively. These results are in agreement with those reported by other authors [48,73]. The results obtained in the photocatalytic degradation test are shown in Figure 11. From this figure, the maximum percentage of cyanide degradation for all nanoparticles is reached around the first 90 min, after which photodegradation becomes almost constant. Evidence from the literature suggests that there is a limit to the efficacy of photocatalysis for complete degradation of some pollutants [74]. In this study, maximum efficiency was reached by La/TO (96.8%), followed by Ce/TO (88.9%), Eu/TO (86.5%), and TO (83.9%).

Effect of the Photogenerated Radicals
It is widely known that photogenerated electrons (e -), photogenerated holes (h+), hydroxyl radicals ( • OH), and superoxide radicals ( • O2 -) are the main reactive oxygen species (ROS) in photocatalytic processes. Therefore, to establish the impact of various ROS on the reaction system, a radical quenching experiment was performed incorporating different radical quenchers [75]. Figure 12 shows the results of this experiment. From this figure, it is evident that, for all nanoparticles, the efficiency of photocatalytic degradation of total cyanide decreased with introduction of isopropanol (i-PrOH) and ethanol (EtOH), which are quenchers of h + and • OH radicals, respectively. On average, the nanoparticles decreased their cyanide removal efficiency by 9.0% and 5.0% when introducing ethanol and isopropanol, respectively. On the other hand, introduction of p-Benzoquinone (p-BQ) Figure 11. Photocatalytic cyanide degradation by nanoparticles.

Effect of the Photogenerated Radicals
It is widely known that photogenerated electrons (e − ), photogenerated holes (h+), hydroxyl radicals ( • OH), and superoxide radicals ( • O 2 − ) are the main reactive oxygen species (ROS) in photocatalytic processes. Therefore, to establish the impact of various ROS on the reaction system, a radical quenching experiment was performed incorporating different radical quenchers [75]. Figure 12 shows the results of this experiment. From this figure, it is evident that, for all nanoparticles, the efficiency of photocatalytic degradation of total cyanide decreased with introduction of isopropanol (i-PrOH) and ethanol (EtOH), which are quenchers of h + and • OH radicals, respectively. On average, the nanoparticles decreased their cyanide removal efficiency by 9.0% and 5.0% when introducing ethanol and isopropanol, respectively. On the other hand, introduction of p-Benzoquinone (p-BQ) as a quencher of radials • O 2 − did not affect the efficiency of the nanoparticles for cyanide removal.

Total Efficiency and Reuse of Nanoparticles
The percentage of cyanide adsorbed and photodegraded by the synthesized nanoparticles is shown comparatively in Figure 13. In this figure, for all the nanoparticles, the photocatalysis process was more efficient for removal of cyanide species than the adsorption process. From this figure, it can be inferred that, under the conditions tested in this study, the highest adsorption and photodegradation capacity of cyanide species was achieved by La/TO nanoparticles, followed by Ce/TO, Eu/TO, and finally by TO nanoparticles.
Finally, knowing that the stability and recyclability of materials with adsorbent and photocatalytic applications are considered crucial aspects for their large-scale application, in this study, reuse experiments were performed for five consecutive cyanide removal cycles. Figure 14 shows the removal efficiency of the synthesized nanoparticles during the five cycles. as a quencher of radials ˙O2 -did not affect the efficiency of the nanoparticles for cyanide removal.

Total Efficiency and Reuse of Nanoparticles
The percentage of cyanide adsorbed and photodegraded by the synthesized nanoparticles is shown comparatively in Figure 13. In this figure, for all the nanoparticles, the photocatalysis process was more efficient for removal of cyanide species than the adsorption process. From this figure, it can be inferred that, under the conditions tested in this study, the highest adsorption and photodegradation capacity of cyanide species was achieved by La/TO nanoparticles, followed by Ce/TO, Eu/TO, and finally by TO nanoparticles.
Finally, knowing that the stability and recyclability of materials with adsorbent and photocatalytic applications are considered crucial aspects for their large-scale application, in this study, reuse experiments were performed for five consecutive cyanide removal cycles. Figure 14 shows the removal efficiency of the synthesized nanoparticles during the five cycles.

XRD and SSA Analysis
The results depicted in Figure 1 indicate that doping with lanthanide (Ln) can impact the crystal structure of pure TiO2 (anatase), causing slight broadening and reduction in intensity of its characteristic diffraction peaks. This suggests that the adsorbent and photocatalytic properties of this semiconductor oxide can be affected by the presence of Ln dopants [76]. According to the literature, it is very difficult for lanthanide ions to substitute Ti 4+ ions in the TiO2 crystal lattice. This is primarily due to the significant difference in ionic radii between Ti 4+ (0.068 nm) and the lanthanide ions, such as La 3+ (0.116 nm), Ce 3+ (0.114 nm), Ce 4+ (0.101 nm), and Eu 3+ (0.107 nm). Therefore, La 3+ , Ce 3+/4+ , and Eu 3+ Figure 14. Nanoparticle reuse experiment.

XRD and SSA Analysis
The results depicted in Figure 1 indicate that doping with lanthanide (Ln) can impact the crystal structure of pure TiO 2 (anatase), causing slight broadening and reduction in intensity of its characteristic diffraction peaks. This suggests that the adsorbent and photocatalytic properties of this semiconductor oxide can be affected by the presence of Ln dopants [76]. According to the literature, it is very difficult for lanthanide ions to substitute Ti 4+ ions in the TiO 2 crystal lattice. This is primarily due to the significant difference in ionic radii between Ti 4+ (0.068 nm) and the lanthanide ions, such as La 3+ (0.116 nm), Ce 3+ (0.114 nm), Ce 4+ (0.101 nm), and Eu 3+ (0.107 nm). Therefore, La 3+ , Ce 3+/4+ , and Eu 3+ could rather be uniformly dispersed on the TiO 2 surface in the form of La 2 O 3 , Ce 2 O 3 /CeO 2 , and Eu 2 O 3 particles, respectively [77]. It should be noted that, even though the doping elements were present in low concentration (1 wt%) and were evenly dispersed on the TiO 2 surface, Figure 1 does not exhibit identifiable diffraction peaks of the corresponding lanthanide oxides [78]. Nevertheless, the reduction in crystallite size may be attributed to the existence of La-O-Ti, Ce-O-Ti, or Eu-O-Ti bonds on the surface of TiO 2 . These findings are consistent with those reported by other researchers, who have demonstrated a decrease in crystallite size caused by formation of Ln-O-Ti bonds [79].
The results of the BET analysis would corroborate this appreciation since the doped nanoparticles (La/TO, Ce/TO, and Eu/TO) show higher specific surface area related to pristine TiO 2 . From the results of the BET analysis, it is suggested that La/TO, Ce/TO, and Eu/TO nanoparticles could have better cyanide species removal capacity than TO due to the greater accessibility of active sites on the increased surface area. Furthermore, the high surface area of the doped nanoparticles could facilitate diffusion of cyanide species and photogenerated ROS. In this way, photodegradation of cyanide could be improved by allowing more energetic photons to be adsorbed on the surface of the photocatalytic nanoparticles. Finally, the high surface area also facilitates contact between the nanoparticles and the cyanide species present in the aqueous medium, which benefits the subsequent photocatalytic reaction [48].

FTIR Analysis
The FTIR spectra shown in Figure 2b-d allowed to demonstrate that, due to doping, the broad band of the TiO 2 semiconductor, shown in the region of 540-900 cm −1 in Figure 2a, changes drastically. According to the literature, this change can be attributed to the presence of Ln-O stretching modes of vibrations [24]. Unresolved bands suggest the existence of uniform stresses in the lattice, which is generally associated with defects in the material structure. These defects, probably derived from doping, can cause shortening of the Ti-O bond, promoting vibrational mode changes observed in doped nanoparticles relative to pristine TiO 2 [80]. The results obtained in this analysis agree with those reported in the literature [57,58]. Figure 3 shows a comparison of the SEM micrographs of the TO, La/TO, Ce/TO, and Eu/TO nanoparticles. As evidenced in this figure, lanthanide elements La, Ce, and Eu, by acting as TiO 2 dopants, could promote reduction in size of semiconductor crystallites. This is probably due to the fact that dispersion of dopant ions La 3+ , Ce 3+/4+ , and Eu 3+ on the TiO 2 surface could restrict direct contact between neighboring crystallites of this semiconductor, thus inhibiting their growth [34]. According to the literature, restriction of direct contact between neighboring crystallites is due to generation of large defects in the crystal lattice of the material. These defects can change the interatomic distance and, therefore, affect the stability of the bonds between TiO 2 atoms. As a result, TiO 2 microcrystals with different sizes and random orientations can be formed [81]. The results of this study agree with those reported by other authors, who have shown that doping enables stabilization of small particles [76]. Regarding EDS analysis, the results obtained allowed us to confirm the presence of La (1.32 wt%), Ce (1.34 wt%), and Eu (1.28 wt%) on the TiO 2 structure. These results support the results obtained in the XRD and FTIR assays. Figure 4a displays the UV-vis absorption spectra of TO, La/TO, Ce/TO, and Eu/TO nanoparticles. In this study, the absorption threshold of the pristine TiO 2 (TO) semiconductor was around 350 nm. The strong absorption band observed in this region for TO would be associated with electronic transition from the valence band (VB) dominated by the 2p orbital of O to the conduction band (CB) dominated by the 3d orbital of Ti [35]. In contrast to the TO spectrum, the spectra of the nanoparticles that have been doped exhibit a red shift. This phenomenon has been described in the literature as being caused by a charge transfer transition between the f-electrons of the lanthanide ions and either the conduction band or valence band of TiO 2 [82]. As can be seen in Figure 4a, the shift of the absorption edge towards a slightly longer wavelength depends on the lanthanide ions used as dopants; however, it is evident that doping with these ions could improve the response of TO to visible light.

Optical and Photoelectric Properties
On the other hand, to estimate the bandgap energy (E g ) of the synthesized nanoparticles, the (UV-vis) DRS method was used, and the absorption data were fitted to equations for direct bandgap transitions. Figure 4b shows a reduction in the E g value for the doped nanoparticles relative to the value calculated for the pristine TiO 2 . According to the literature, the reduction in the value of E g is due to formation of impure energy levels (impure orbitals) between the VB and the CB of the doped semiconductor [83]. Furthermore, due to the presence of these impure orbitals, less energy is required to achieve the transfer of electrons from the valence band to the impure orbital and/or from the latter to the conduction band. From these results, it is suggested that doping of the TiO 2 semiconductor with lanthanide elements La, Ce, and Eu allows to reduce the bandgap energy and, consequently, to optimize the capacity of TiO 2 to absorb wavelengths with less energy (including solar radiation), as well as reduce recombination of photogenerated charges and modify the adsorption capacity of the surface of this semiconductor [13].

Effect of Nanoparticles Composition
In this study, a preliminary experiment was developed in order to evaluate the best concentration of elements La, Ce, and Eu as TiO 2 dopants for effective removal of total cyanide from aqueous systems. Figure 5 shows that the total cyanide removal capacity of the doped nanoparticles increases when the percentage by weight of the dopant in the semiconductor increases from 0.5 to 1%. However, the total cyanide removal capacity of the doped nanoparticles decreases when the percentage by weight of the dopant in the semiconductor increases from 1 to 2%. According to the literature, the reason for this is that the presence of high concentrations of doping elements can lead to an increase in the number of oxygen vacancies in the material, resulting in an increase in the number of recombination centers for photoinduced charges [84]. Similarly, when the dopant particles agglomerate and form clusters, they can obstruct the active sites on the surface of the semiconductor, leading to a reduction in its photocatalytic efficiency. This has been reported in the literature as well [85]. These findings align with previous studies conducted by other researchers, who have shown that incorporating lanthanide elements at levels of 1-2 wt% can effectively enhance photoactivity of semiconductors across a broad range of the electromagnetic spectrum [78].

Effect of pH on Cyanide Adsorption
Due to the protonation-deprotonation balance of the surface hydroxyl groups, the surface of titanium oxide (TiO 2 ) has a pH-dependent charge, so this oxide presents a positive charge under acidic conditions and a negative charge under basic conditions [67]. In the middle of these two regions is the point of zero charge (PZC), that is, the pH where the total charge on the TiO 2 surface is zero. As reported in the literature, adsorption of ionic doping species, including lanthanides La, Ce, and Eu, could change the surface charge density and position of PZC of TiO 2 [86,87].
Several studies have reported that TiO 2 behaves as a weak Brönsted acid and that, when hydrated, it can form surface hydroxyl groups (Ti-OH) [88]. The hydroxyl groups present can participate in chemical reactions involving association and dissociation of protons, thus generating a pH-dependent surface charge: TiOH + H + ↔ TiOH + 2 at pH < pH PZC (17) and TiOH ↔ TiO − + H + + H 2 O at pH > pH PZC (18) where the positive, neutral, and negative surface hydroxyl groups are represented by TiOH 2 + , TiOH, and TiO − , respectively [87]. From Figure 6, it can be inferred that, as the pH of the solution enhances, the amount of total cyanide adsorbed also increases until reaching an adsorption maximum at pH = 9. Above this value, cyanide adsorption remains constant for all nanoparticles. In this figure, it is also observed that the doped nanoparticles have a higher total cyanide removal capacity compared to pristine TiO 2 . This is probably because doping with lanthanum (La), cerium (Ce), and europium (Eu) enables greater total number of available active sites on the surface of TiO 2 nanoparticles, thus driving favorable kinetics [34].
According to the literature, the pH level of a solution has an impact on the surface charge of the adsorbents and the ionization and speciation of the adsorbate. Pristine and doped TiO 2 nanoparticles have a pH PZC of 6.9 and 7.1, respectively. This implies that, at higher pH levels, there is a preference for an increase in negatively charged groups, while, at lower pH levels, there is a preference for an increase in positively charged groups. Regarding cyanide (pKa = 9.4), at pH values < pKa, this molecule in solution associates in its molecular form (HCN), while, at pH > pKa, it dissociates mainly in its ionic form (CN − + H + ).
Evidence from several studies indicates that, at pH < pH PZC , the specific adsorption of HCN species on the OH groups present on the surface of hydrated nanoparticles is driven through formation of N-H polar covalent bonds, as indicated below [69] TiOH + HCN ↔ TiOH · · · NCH (19) or For pH < 7.0, the surface of the nanoparticles exhibits a positive charge and the degree of HCN dissociation is negligible. Therefore, electrostatic forces of attraction between a charged surface and a neutral molecule are unlikely to develop, especially at very low pH values. At pH > 7.0, the negative charge on the surface of the nanoparticles increases, but the dissociation of HCN into CN − ions is negligible until pH = pKa = 9.4 is reached. However, in the pH range between 7.0 and 9.4, an increase in the forces of electrostatic attraction is expected and, therefore, an increase in the adsorption of HCN species on the negative surface of the nanoparticles. At pH > pKa = 9.4, the proportion of CN − species in the solution increases, but, in parallel, the surface of the nanoparticles becomes more negative since pH > pH PZC . This generates an augment in the repulsive forces between the negative surface sites of the nanoparticles and the negative CN − ions, which causes a reduction in the adsorption capacity of these ions at very high pH values, as shown in Figure 6.
The strong dependence between the pH of the solution and the adsorption capacity of the nanoparticles is a clear indication that, in this study, the total cyanide adsorption mechanism was mainly driven by electrostatic interactions, although other types of interactions cannot be excluded (van der Waals and/or specific interactions). In fact, it is suggested that adsorption of CN − species on La/TO, Ce/TO, and Eu/TO nanoparticles could occur predominantly by chemical adsorption instead of physical adsorption (outer sphere complex), although the presence of physical adsorption is not ruled out completely. The results of this study are consistent with those reported in the literature [44]. When CN − is in contact with active cationic sites on the adsorbent surface, it forms a bond with them. According to the literature, lanthanum, cerium, and europium ions show strong electron withdrawal effect [25,33], so the presence of these lanthanides on the surface of TiO 2 nanoparticles contributes to generation of Lewis acids sites [28,29,89]. It should be mentioned that these active cationic sites could certainly improve cyanide adsorption on the TiO 2 surface [34], in addition to providing this semiconductor with catalytic stability in an aqueous reaction system [90]. Figure 7 shows the adsorption isotherms obtained for the synthesized nanoparticles. This figure displays that, at low concentrations of cyanide species (HCN + CN − ), the adsorption rate increases; instead, at high concentrations of these species, the adsorption rate reaches a maximum where it stabilizes. One possible reason for this phenomenon is the excessive presence of cyanide compounds that compete for the available active sites on the surface of the adsorbent nanoparticles. Evidence from these results suggests that the initial concentration of cyanide species in solution could generate a significant driving force to allow these species to migrate from the liquid phase to the nanoparticle surface [91].

Maximum Cyanide Adsorption Capacity
The adsorption data obtained in this study for TO, La/TO, Ce/TO, and Eu/TO nanoparticles were fitted to the Langmuir and Freundlich isotherm models. Table 2 describes the parameters corresponding to adjustment of the experimental data. As shown in this table, given the values of the correlation coefficient (R 2 ), it is concluded that the Langmuir model is the most suitable to describe the equilibrium adsorption behavior. Consequently, it is suggested that monolayer adsorption of cyanide species on the synthesized nanoparticles proceeds as a phenomenon of electrostatic attraction. This attraction takes place in regions with homogeneous surfaces, where the strongest binding sites are initially filled, and, as the saturation level increases, the binding strength decreases [44]. Finally, the values of the Langmuir and Freundlich constants (R L and n) shown in Table 2 confirm the favorable adsorption of cyanide species on the nanoparticles synthesized in this study.

Adsorption Thermodynamics
As mentioned above, thermodynamic parameters offer the most reliable indications for effective implementation of a process in practice. Table 3 shows the thermodynamic parameters obtained in this study for removal of cyanide species. Negative ∆G • values suggest the feasibility and thermodynamic spontaneity of total cyanide adsorption on TO, La/TO, Ce/TO, and Eu/TO nanoparticles. Furthermore, the decrease in negative values of ∆G • with increasing temperature reveals an increase in the effectiveness of the adsorption process at high temperatures. On the other hand, the positive values of ∆H • and ∆S • shown in Table 3 suggest that adsorption of cyanide species on the nanoparticles occurred as an endothermic process. Finally, the positive ∆S • also suggests an increase in randomness at the solution/solid interface, with several structural alterations in the active sites of the nanoparticles during cyanide fixation. These results are in agreement with those informed by other authors [43].

Kinetic Behavior of Cyanide Adsorption
The kinetic models of adsorption allow to determine the contact time necessary for the whole adsorption of the chemical species. Figure 9 shows the fit of the experimental data obtained in this investigation with the pseudo-first-order and pseudo-second-order kinetic models. In this figure, for all nanoparticles (TO, La/TO, Ce/TO, and Eu/TO), the concentration of cyanide species decreases very fast at the beginning of the process and tends to be constant after~90 min. The rapid adsorption that occurs in the initial stage could be due to the high concentration gradient, as well as the presence of vacant adsorption sites. Table 4 shows the kinetic parameters of adsorption obtained for all the nanoparticles. From the values shown in this table for the correlation coefficient (R 2 ), it can be concluded that the experimental data fit the pseudo-second-order model more than the pseudo-first-order model, suggesting chemical adsorption of cyanide species on the surface of nanoparticles [43].
On the other hand, Figure 10 shows the fit of the experimental data to the intraparticle diffusion model. From this figure, the adsorption of total cyanide on the nanoparticles occurs conceptually in two stages (linear regions), after which intrinsic adsorption occurs, either by chemical binding on the active sites on the nanoparticles or by physical processes. The initial stage of fast speed that is observed in Figure 10 could be described as a process of diffusion of cyanide particles through the stationary film that surrounds each adsorbent nanoparticle; therefore, it corresponds to the mass transfer from the aqueous solution to the surface of the adsorbent. In contrast, the slow-rate second stage corresponds to intraparticle mass transfer and describes the process of diffusion of cyanide particles through the pores of adsorbent nanoparticles. The linear regression analysis for the diffusion kinetic models is displayed in Table 4. From the relatively high values of A reported in this table, it is suggested that surface adsorption could be the rate-limiting step for all synthesized nanoparticles [44].

Kinetics of Cyanide Photodegradation
Regarding the photocatalytic behavior of TO, La/TO, Ce/TO, and Eu/TO, in this study, it was possible to demonstrate that these nanoparticles achieved effective removal of total cyanide in the aqueous solution. Possibly, this occurred because the doping of TiO 2 with lanthanide elements La, Ce, and Eu not only provided greater quantity of active centers for adsorption of cyanide species but also the presence of these lanthanide elements on the surface of TiO 2 promoted decrease in bandgap energy, which definitely enhanced the photoactivity of the doped nanoparticles. It is widely known that magnitude of bandgap energy is fundamental in photoactivity of semiconductors since it determines the recombination rate of the electron/hole pairs (e − /h + ). Moreover, the recombination rate may depend on other factors, such as carrier concentration, carrier mobility, and semiconductor structure [81]. Therefore, the evidence from this study suggests that the presence of La, Ce, and Eu ions on the TiO 2 surface enhanced the photoactivity of this semiconductor, probably by modifying the recombination rate of the pair (e − /h + ). In addition, in this study, it was shown that the nanoparticles doped with La/TO, Ce/TO, and Eu/TO had lower bandgap energy than TO, so it is suggested that, given the smaller separation between the valence bands (VB) and conduction (CB), doped nanoparticles could easily transfer photoinduced electrons from the bulk to the surface and be more active than TO nanoparticles under simulated solar light [92].
The photocatalytic route involving the doped nanoparticles (La/TO, Ce/TO, and Eu/TO) begins with electronic excitation under simulated solar light. The photoexcited electrons (e − ) are transferred from the VB (e − VB ) to the CB (e − CB ), leaving holes (h + ) in the VB (h + VB ) of the photocatalyst (Equation (21)). Due to this electronic transfer, electron/hole pairs (e − /h + ) are photogenerated that migrate to the surface of the photocatalyst to react directly with the adsorbed species, such as H 2 O ((H 2 O) ads ), OH − (OH − ads ), O 2 ((O 2 ) ads ), and other molecules. These (e − /h + ) pairs can also recombine immediately after their formation (Equation (22)). On the other hand, the photogenerated holes (h + ) in the VB of TiO 2 promote oxidation of both adsorbed water molecules and hydroxyl ions to produce highly reactive hydroxyl radicals (Equations (23) and (24)). Furthermore, the holes (h + ) can migrate towards the surface of the photocatalyst to create more reactive radicals and oxidize the molecules adsorbed on the surface. Lanthanide ions (Ln n+ ) have empty 5d orbitals, so they can trap photoexcited electrons in the CB of TiO 2 (Equation (25)); however, these electrons are very unstable, so they are quickly transferred to oxygenated molecules adsorbed on the surface of the photocatalyst. Due to this transfer, radical superoxide anions ( • O 2 − ) and hydroxyl radicals ( • OH) are generated through a series of sequential reactions (Equations (26)- (29)). The following reactions suggest the likely pathway for ROS photogeneration on the surface of Ln (n+) /TiO 2 (Ln (n+) = La (3+) , Ce (3+/4+) , or Eu (3+) ) [73,89]: Cyanide photodegradation occurs at the surface of the photocatalyst as a complex process involving transfer of multiple electrons and protons from the catalyst to cyanide species. Therefore, one of the key steps in photocatalytic degradation of cyanide is adsorption of cyanide ions on the surface of the photocatalyst [93].
Photocatalytic degradation of cyanide involves various reactive oxygen species (ROS), such as hydroxyl radicals ( • OH), superoxide radicals ( • O 2 − ), and singlet oxygen ( 1 O 2 ), among others. As is well-known, the hydroxyl radical ( • OH) is the most reactive species and can oxidize a wide range of contaminants, including cyanide. Oxidation of cyanide by hydroxyl radicals ( • OH) can proceed through several pathways. In the biocatalytic pathway, the hydroxyl radical attacks the carbon atom of the cyanide ion to form formamide (NH 2 CHO), which can then react with the hydroxyl radicals to form formic acid (HCOOH) and ammonia (NH 3 ) [94].
In the photocatalytic pathway, the hydroxyl radical attacks the nitrogen atom of the cyanide ion to form cyanate (CNO − ) (Equation (30)), which can be oxidized to produce nitrogen gas, carbon dioxide gas, nitrate, or nitrite, and bicarbonate (Equations (31)-(34)) [95][96][97][98]. It is worth mentioning that hydrolysis of cyanate to bicarbonate and ammonium ions (Equation (35)) is more favored under acidic conditions (pH < 7) [99]. On the other hand, direct oxidation of cyanide can also occur if the molecule reacts directly with the photogenerated holes [79].
In addition to oxidation of cyanide by hydroxyl radicals, electrons in the conduction band of the photocatalyst may also participate in the degradation process. Electrons can reduce oxidized species (CNO − , CO 2 , N 2 ) to less harmful and more stable forms, such as formate (HCOO − ) or bicarbonate (HCO 3 − ). The reduction process is driven by transfer of electrons from the photocatalyst to the oxidized species [94].
In summary, photocatalytic cyanide degradation is a process that can reduce cyanide toxicity in wastewater by generating reactive species through light absorption by a photocatalyst. The photodegradation mechanism of cyanide in an aqueous solution involves a series of reactions that depend on the type of photocatalyst used and the reaction conditions, including the initial cyanide concentration, light intensity, and the presence of other contaminants in the solution [91]. Although there are many factors that can affect the effectiveness of the cyanide photocatalysis process, the process has great potential as a treatment strategy for cyanide-contaminated wastewater.

Effect of the Photogenerated Radicals
As shown in Figure 12, the efficiency of nanoparticles for photodegradation of cyanide species decreased with incorporation of i-POH and EtOH. This probably occurs due to quenching of h + and • OH, suggesting that these play a crucial role in photocatalytic degradation of cyanide species. However, Figure 12 did not show a significant effect on efficiency of photodegradation using p-BQ, probably because the presence of the radical • O 2 − is not essential for photocatalytic degradation of cyanide species under the conditions evaluated in this study. These results are in agreement with those informed by other authors [100].

Total Efficiency and Reuse of Nanoparticles
In this study, it was shown that photodegradation of cyanide species is more efficient than adsorption of these species under the test conditions ( Figure 13). As is known, adsorption is a process that can be affected by a series of parameters, such as sorbent properties, sorbate properties, solution conditions, among others. In fact, in this study, it was shown that the pH of the solution had an important effect on adsorption capacity of cyanide species on the surface of nanoparticles. Thus, Figure 6 shows that the maximum adsorption capacity of cyanide species occurs at pH values > 9. However, since the adsorption tests were performed at pH = 7, it is possible that the relatively low adsorption percentages shown in Figure 13 were due to the limited electrostatic attraction between cyanide species (pKa = 9.4) and nanoparticle surface (pH ZPC = 7). Although adsorption of cyanide species was limited by the pH of the solution, in this study, it was also possible to demonstrate that incorporation of lanthanide elements, particularly lanthanum, in the TiO 2 structure is a good alternative to improve adsorption of these species on the semiconductor surface. Possibly, this was because incorporation of lanthanide elements as dopants enables generation of Lewis acid sites on the TiO 2 surface, which undoubtedly contributes to improving the adsorption capacity of a semiconductor.
Likewise, in this study, it was demonstrated that the combination of the adsorption and photocatalysis processes enables improvement in the efficiency of removal of cyanide species from aqueous solutions. This is probably because the cyanide species that first adsorbed and accumulated on the nanoparticle surface at the beginning of photocatalytic degradation are the ones that first degraded under simulated solar light. Consequently, the constant migration and successive photocatalytic oxidation on the surface of the nanoparticles certainly contribute to improving the removal efficiency at the solid-liquid interface. This is due to generation of a concentration gradient, which acts as the main driving force of the removal process of cyanide species from an aqueous solution. From these results, it is suggested that, in this study, removal of cyanide species is due to the cooperative effect between both adsorption and photocatalysis processes. Thus, the synthesized nanoparticles are effective for removal of cyanide from aqueous systems because they have coupled "adsorption-photodegradation" performances [88].
On the other hand, it is well-known that both the useful life of a material and its potential applications are closely related to its structural and chemical stability. In this study, an experiment was designed to evaluate reuse of synthesized nanoparticles in order to estimate their effectiveness after five cycles of consecutive use. The results of this experiment are shown in Figure 14. From this figure, it could be inferred that percentage of cyanide removal decreases with each treatment cycle. However, after five consecutive cycles, loss of cyanide removal capacity of the nanoparticles did not exceed 20% on average. In fact, in this study, it was found that La/TO nanoparticles had a lower loss of effectiveness (15.3%) at the end of the fifth reuse cycle compared to Ce/TO (16.6%) and Eu/TO (17.8%) nanoparticles. The loss in effectiveness was probably due to chemical adsorption of cyanide species on the surface of the nanoparticles since the possible formation of covalent bonds and complexes could decrease the availability of active sites on the surface. At the end of the fifth cycle, the chemical composition of the nanoparticles was verified by ICP analysis, confirming that there was no lanthanide leakage into the solution. Consequently, the synthesized nanoparticles (TO, La/TO, Ce/TO, and Eu/TO) are stable and maintain adequate activity until the fifth treatment cycle, being able to effectively remove cyanide species in aqueous solution.
Finally, Table 5 compares the maximum adsorption capacity (mg g −1 ) of the present nanoparticles and some adsorbents used for cyanide removal from aqueous solutions.  Likewise, Table 6 compares the photodegradation efficiency (%) of the present nanoparticles and some photocatalysts used for cyanide removal from aqueous solutions. As can be seen in Tables 5 and 6, among the nanoparticles synthesized in this study, La/TiO 2 showed the highest capacity for adsorption and photodegradation of cyanide species in aqueous solutions. However, compared to other materials based on metal oxides (ZnO, NiO, Fe 2 O 3 ) reported in the literature, La/TiO 2 (q max = 54.96 mg g −1 ) has a lower adsorption capacity for cyanide species under the tested conditions, although it turned out to be a better adsorbent material for cyanide species than certain zeolitic materials and biochar. On the other hand, La/TiO 2 also proved to be more efficient for cyanide photodegradation than the other nanoparticles synthesized in this study. Likewise, La/TiO 2 proved to be the same and even more efficient for cyanide photodegradation than other TiO 2 -based photocatalytic materials recently reported in the literature, reaching an efficiency of 97% after 90 min of reaction, with a catalyst load of 0.2 g L −1 and a cyanide concentration in the solution of 20 mg L −1 .
Based on the adsorption and photodegradation results obtained in this study and comparing them with those reported in the literature cited above, it is suggested that La/TO, Ce/TO, Eu/TO, and TO nanoparticles, in that order, are effective for removal of cyanide species from aqueous solutions. Furthermore, the evidence from this study suggests that the removal of cyanide species was governed by a combination of electrostatic interactions, complex formation by coordinate covalent bonds, and photo-oxidation on the surface of the synthesized nanoparticles when the reaction systems were exposed to simulated solar light.

Conclusions
The evidence from this study indicates that the adsorption and photodegradation capacity of the nanoparticles synthesized in the present study (TiO 2 , La/TiO 2 , Ce/TiO 2 , and Eu/TiO 2 ) were influenced by several factors. Some of these were dependent on material properties (chemical composition, bandgap energy, crystal structure, morphology, particle size, specific surface area, and pH PZC ), as well as the operating conditions (initial concentration of adsorbate, contact time, system temperature, and pH of the solution) of the removal processes used. These results are in agreement with those reported by other authors [120]. Investigation of these parameters enabled us to obtain valuable information on the mechanisms of adsorption and photocatalysis, which was fundamental to determining the best operational conditions for efficient removal of cyanide species from aqueous solutions at neutral pH. Likewise, synergistic coupling of the adsorption and photocatalysis processes enabled significant improvement in the capacity of the TiO 2 semiconductor for effective removal of cyanide species from aqueous solutions.
Our results showed that doping of the TiO 2 semiconductor with lanthanide elements La, Ce, and Eu represents a promising alternative to improve the adsorption capacity of the TiO 2 semiconductor and extend its photoresponse to the visible light region. Among the doped nanoparticles, La/TO was more efficient than Ce/TO and Eu/TO for total cyanide removal, probably because the La/TO nanoparticles had a higher specific surface area (126 m 2 g −1 ) in comparison with the specific surface area of Ce/TO (104 m 2 g −1 ) and Eu/TO (96 m 2 g −1 ). Regarding cyanide photodegradation, it was also influenced by type of doping lanthanide ion. La/TO nanoparticles were slightly more effective under simulated solar light than Ce/TO and Eu/TO nanoparticles, possibly because the La/TO bandgap energy (E g = 3.10 eV) was lower than that of Ce/TO (E g = 3.14 eV) and Eu/TO (E g = 3.16 eV), which is essential for absorption of light and generation of pairs of electrons and voids for photodegradation of pollutants. On the other hand, reuse of doped nanoparticles was also affected by type of lanthanide element. In this study, it was found that La/TO nanoparticles had lower loss of effectiveness (15.3%) at the end of the fifth reuse cycle compared to Ce/TO (16.6%) and Eu/TO (17.8%) nanoparticles, probably due to the smaller particle size of La/TO compared to Ce/TO and Eu/TO, which could provide greater structural stability to the photocatalyst in aqueous reaction systems. Therefore, the evidence from this comparative study suggests that each doping element has an intrinsic effect on the properties of TiO 2 , making it essential to understand these effects in the design of effective materials for wastewater treatment.
In summary, this study provides information on the potential capacity of doping TiO 2 nanoparticles with lanthanides (La, Ce, and Eu) as an innovative and effective approach to improve the adsorption and photocatalysis properties of a semiconductor. In addition, it highlights the synergistic effect of combining both techniques, adsorption and photocatalysis, to achieve complete and efficient removal of cyanide in wastewater, which is valuable for development of new technologies that contribute to treatment and recovery of hydric resources. Finally, it supports use of solar energy as a lighting source during removal of cyanide in aqueous systems, contributing to generation of beneficial effects in terms of environmental impact, energy efficiency, and remediation costs, thus promoting sustainable practices.