Enhancement of photocatalytic activity of g-C3N4 under solar light by Nd3+ doping and HPA incorporation and its application in the degradation of ceftriaxone sodium

ABSTRACT Nd3+-doped graphitic carbon nitride nanosheets incorporated with heteropoly phosphotungstic acid (Nd3+-g-C3N4-HPA) were synthesised, and their optical response, band structure and charge separation efficiency were analysed. UV–vis diffuse reflectance spectral studies revealed that Nd3+ doping and HPA incorporation led to an increase in absorption intensity, which enhances the range of visible light absorption. Improved separation efficiency and reduction in the recombination rate of photogenerated electrons and holes are supported by photoluminescence studies. This novel photocatalyst was applied for the removal of the antibiotic, ceftriaxone sodium from water and attained complete degradation within 75 min of sunlight irradiation, demonstrating that the Nd3+-g-C3N4-HPA photocatalyst would be beneficial for eliminating persistent organic pollutants like dyes and pharmaceutical compounds from wastewater. Investigations revealed that superoxide anion radicals and holes play a significant role in the photocatalytic degradation of pollutants. Furthermore, a possible degradation pathway of ceftriaxone sodium was proposed using mass spectral analysis.


Introduction
Antibiotics have been extensively used for both human treatment and animal husbandry over the past few decades.The excessive use of antibiotics resulted in incipient organic contaminants with fetotoxic and mutagenic properties [1][2][3][4][5].Ceftriaxone sodium, an antimicrobial drug with broad-spectrum bactericidal capability [6], has been widely used for treating respiratory tract infection (RTI), urinary tract infection and gonorrhoea.However, its excessive intake causes toxic effects such as abdominal pain, a decrease in the prothrombin time, renal dysfunction and allergic dermatitis [7].Due to the irrational use and inadequate treatment, a large amount of ceftriaxone sodium antibiotics has been released into the environment.Hence, its presence in the aquatic and earthbound environment is increasing in an alarming rate, causing potential threat to the aquatic ecosystems, microbial population and health of human being.Many researchers are focusing on the effective eradication of these types of antibiotic waste materials [3,4].It is presently grounded that conventional removing technologies such as adsorption, reverse osmosis (RO), biological treatment [8] and chemical oxidation [9] are improper for the total expulsion of toxic substances like pharmaceuticals from water.In these decades, advanced oxidation processes like photocatalysis [10] and piezocatalysis [11,12] have been considered as feasible, advantageous and cost-effective techniques, applied broadly for the removal of organic pollutants from water.In this study, we focus on solar light active photocatalysis to debase antibiotics delivered into the environment under mild reaction conditions.
Recently, metal-free semiconductor, graphitic carbon nitride (g-C 3 N 4 ) with a band gap energy of 2.7 eV, has been considered as one of the most potential photocatalysts for organic pollutants degradation, solar energy conversion, water splitting and hydrogen production due to its distinctive characteristics like chemical stability, facile low-cost synthesis, non-toxicity and unique two-dimensional (2D) layered structure [13].Graphitic carbon nitride provides active sites for photocatalysis; at the molecular level, the highest occupied molecular orbital (HOMO) in the g-C 3 N 4 frameworks is largely made up of N2 orbitals, whereas the lowest unoccupied molecular orbital (LUMO) is made up of N2, N3, C1 and C2 orbitals.As a result, in the photocatalytic process, N atoms supply the oxidation and reduction of active sites, whereas C atoms tend to offer the reduction of active sites [14].However, its photocatalytic degradation capacity is seriously confined, fundamentally because of the limited light absorption capacity, high recombination paces of charge carriers, low charge mobility and low surface area.Various tactics like metal/non-metal doping [15,16], surface modification [17,18], nanostructure synthesis [19] and construction of heterojunction [20] have been applied to resolve these limitations.Among these modification methods, doping is regarded as a highly effective strategy and widely utilised for the enhancement of photocatalytic activity of graphitic carbon nitride in recent years.Doping plays a crucial role in broadening the visible light response, increasing the charge transfer mobility and generating more active sites [21][22][23].Although extensive research has been carried out on metal doping in the g-C 3 N 4 photocatalyst, few works are available on lanthanide metal doping in g-C 3 N 4 .Studies revealed that lanthanide metal ion (La, Eu, Ce and Sm)-doped g-C 3 N 4 shows better photocatalytic activity than pristine g-C 3 N 4 [24][25][26], as they could significantly increase the life of photo-induced electrons and holes as well as photoabsorption of g-C 3 N 4 due to its rich energy levels and 4 f electronic transition properties.Lanthanide metal doping also improves the BET surface area and photocurrent density of graphitic carbon nitride and is found to be effective in the photocatalytic removal of organic pollutants [25,26].However, studies on the synthesis and application of Nd 3+ -doped g-C 3 N 4 for the removal of pharmaceuticals are hardly available [27].
The enhancement of the photocatalytic activity of g-C 3 N 4 via surface modification of g-C 3 N 4 by incorporating photosensitive molecules [28,29] is another novel strategy explored by researchers.Polyoxometalates (POMs) are photosensitive molecules with unique optoelectronic properties and high reduction potential, which serve as electron pools as they have great ability to relocate electrons from semiconductors to other substrates [30].Thus, they act as mediators to regulate the dynamics of photogenerated electrons and increase the lifetime of electron hole pairs.Surface modification of g-C 3 N 4 with POMs like heteropoly phosphotungstic acid (HPA) significantly curbed the recombination of photoinduced electrons and holes [31] and was found to be effective in the photooxidation of benzyl alcohol as well as the oxidative desulphurisation process [32].
Although lanthanide-doped g-C 3 N 4 as well as HPA-incorporated g-C 3 N 4 and their photocatalytic activities have been reported, the synergistic effect of lanthanide doping and HPA incorporation on the photocatalytic activity of g-C 3 N 4 has not yet been reported to the best of our knowledge.Hence, in this study, we have made an attempt to exploit the advantages of Nd 3+ doping and HPA integration on solar light-assisted catalytic activity of g-C 3 N 4 .The study reveals that the presence of neodymium ions and HPA has a significant influence on structural and optical properties of g-C 3 N 4 , which results in the enhancement of visible light absorption and reduction in the rate of the electron-hole recombination.This novel photocatalyst was applied for the photocatalytic removal process of ceftriaxone sodium in water under sunlight, and the study sheds light on an environmentally sustainable process for the removal of antibiotic waste materials from the aquatic system.Furthermore, the intermediate products of ceftriaxone sodium generated during the photodegradation process were analysed and the degradation pathway was observed by liquid chromatography-mass spectrometry (LC-MS).

Synthesis of photocatalysts
The pristine g-C 3 N 4 photocatalyst was prepared by the thermal condensation method based on a previous report [26].To synthesise Nd 3+ -doped g-C 3 N 4 , 10 g of urea was weighed and dissolved in deionised water and a certain amount of 10 −5 M solution of Nd (NO 3 ) 3 was added with thorough stirring.The mixtures were dried at 80°C, were transferred into a silica crucible, covered with a lid and subjected to calcination for 2 h at 550°C in a muffle furnace.The light-yellow coloured powder was obtained and designated as Nd 3+ -g-C 3 N 4.
To synthesise a heteropoly phosphotungstic acid-incorporated neodymium-doped g-C 3 N 4 (Nd 3+ -g-C 3 N 4 -HPA) system, 100 mg of neodymium-doped g-C 3 N 4 sample was mixed with 0.05 ml of 10 −5 M HPA solution in water.10 ml of distilled water was added to this mixture and stirred for 30 min.It is then dried at 80°C.

Characterisation
A Perkin Elmer Spectrum 400 FT-IR/FT-FIR Spectrometer was used for IR spectral analysis in the wavelength range of 4000-400 cm −1 .The crystal information of the as-prepared samples was observed by X-ray diffraction (XRD) analysis using a Rigaku Miniflex 600 diffractometer.TEM images were obtained by Jeol/JEM-2100 transmission electron microscopy.A BET analyser, Tristar II, Micrometrics, USA, was used to determine the surface area and porous properties of the samples.The X-ray photoelectron spectra (XPS) were obtained using a Kratos Axis Ultra X-ray photoelectron spectrometer, with Al Kα radiation as the exciting source.The UV-vis DRS spectra of samples were recorded using a Shimadzu UV-2600 spectrophotometer.Photoluminescence (PL) studies were carried out using a Horiba Fluorolog Fluorescence Spectrometer at an excitation wavelength of 380 nm.

Photocatalytic experiments
Photocatalytic activity of the prepared samples was examined using a Heber solar simulator (Heber Scientific, model no: HMV-88123).It comprises a tungsten halogen lamp and a mercury vapour lamp along with A.M 1.5 G filter.The photocatalytic activity of the synthesised g-C 3 N 4 photocatalyst was evaluated by the photocatalytic degradation of methylene blue (MB).For a typical photocatalytic experiment, 0.01 g of synthesised g-C 3 N 4 was suspended in 20 ml of 30 ppm MB dye solution.The resultant suspension was stirred in the dark for 30 min.To study the photocatalytic degradation under sunlight, MBg-C 3 N 4 suspension was kept under solar irradiation at room temperature for 120 min.The samples were withdrawn at different time intervals and centrifuged to separate the photocatalyst.The absorbance of MB was measured at 660 nm using a Double Beam UV-vis Spectrophotometer (Systronics 2203).The degradation of MB solution (20 ml of 30 ppm) was carried out by 120 min of solar irradiation without any catalyst.There was no significant degradation of MB in the absence of the catalyst or solar irradiation alone.The rate of degradation was evaluated using Eq.(1), where C 0 is the initial concentration of MB and C t is its concentration after different intervals of time.The active species involved and their role in photocatalysis were examined using scavengers.Isopropyl alcohol (IPA), ascorbic acid (AA) and triethanolamine (TEOA) were used as scavengers for hydroxyl-free radicals (OH), superoxide anion radicals ( • O 2

▬
) and holes (h + ), respectively.The experimental procedure involves the addition of 1 mM of scavengers to 0.01 g of photocatalyst in MB solution (20 ml of 30 ppm MB).The mixture was exposed to solar light irradiation, and the degradation rate of MB was monitored at 660 nm in a UV-vis spectrophotometer.
To investigate the photocatalytic degradation of ceftriaxone sodium under solar irradiation, an aqueous solution of 30 ml of 50 ppm ceftriaxone sodium was treated with the synthesised catalysts.The resultant suspension was stirred in the dark for 30 min to attain adsorption-desorption equilibrium and then kept under solar irradiation.The samples were withdrawn at different time intervals and centrifuged to separate the photocatalysts, and the degradation rate of ceftriaxone sodium was monitored at 238 nm in a UV-vis spectrophotometer (Systronics 2203).Photocatalytic degradation of ceftriaxone sodium (30 ml of 50 ppm), without any catalyst for the same irradiation time, was also performed.Furthermore, to ensure the complete degradation of ceftriaxone sodium and to understand the reaction intermediates, samples were taken at different time intervals (0, 25 and 75 min) from the reaction vessel and analysed by liquid chromatography mass spectrometry (LC-MS) in the positive ionisation mode.The solvents used for the LC-MS analysis were 5 mmol ammonium formate and 0.1% formic acid (gradient A) and acetonitrile (gradient B).

FT-IR analysis
Structural information of the synthesised samples, g-C 3 N 4 , Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA, was analysed by FTIR spectroscopy (Figure 1).The broad absorption band at 3200 cm −1 is ascribed to the stretching vibration of N-H groups, associated with the uncondensed amino group [31,33].A series of peaks in the range of 1200-1700 cm −1 are typical stretching modes of aromatic CN heterocycles.The peaks at 1238, 1316 and 1407 cm −1 are recognised as the stretching modes of C-N heterocycles and confirm the existence of C-N-C networks in g-C 3 N 4 [24].The absorption bands at 1567 and 1632 cm −1 are indexed as the stretching vibration of C = N.The sharp peaks at 890 cm −1 and 810 cm −1 are assigned as the bending modes of N-containing heterocycles and bending vibration of the s-triazine ring modes, respectively [34].All these bands are observed in Nd 3+ -g-C 3 N 4 , but some are slightly shifted.The characteristic peak at 3200 cm −1 corresponding to N-H vibration in g-C 3 N 4 is shifted to 3186 cm −1 in Nd 3+ -g-C 3 N 4 .All the characteristic vibrational peaks of g-C 3 N 4 befall at almost similar wavenumbers in doped samples, which indicates that there is no significant change in the structure of g-C 3 N 4 on Nd 3+ doping.The incorporation of HPA into Nd 3+ -g-C 3 N 4 causes broadening of the peak at 890 cm −1 due to the interaction of W-O-W linkage with N-containing heterocycles in g-C 3 N 4 [35].The peak at 967 cm −1 in Nd 3+ -g-C 3 N 4 -HPA is associated with the terminal bands of W = O in the exterior WO 6 octahedron of HPA molecules.An additional peak at 1070 cm −1 is also observed in the IR spectrum of Nd 3+ -g-C 3 N 4 -HPA, which corresponds to the stretching frequency of P-O in the central PO 4 tetrahedron of HPA molecules.The appearance of the IR bands at 890 cm −1 , 967 cm −1 and 1070 cm −1 in Nd 3+ -g-C 3 N 4 -HPA confirms that the primary Keggin structure of heteropoly acid is retained even after incorporating with graphitic carbon nitride [35][36][37].All these peaks clearly indicate that HPA is successfully incorporated into Nd 3+-g-C 3 N 4 .

XRD analysis
XRD analysis was performed to investigate the phase structure of the synthesised samples, and the diffractograms show two characteristic peaks (Figure 2).The low-angle diffraction peak at 12.9° in g-C 3 N 4 is attributed to the in-planar repeated tri-s-triazine (s-heptazine) units, corresponding to the (100) plane (JCPDS 87-1526).The peak at 27.4° is associated with the interlayer graphitic piling of conjugated aromatic CN units, corresponding to the (002) plane with an interlayer distance of 0.323 nm [38].The increase in the intensities of the peaks is observed in the XRD patterns of Nd 3+ -doped and HPAincorporated samples.However, characteristic diffraction peaks of neodymium and HPA were not found in the X-ray diffractograms of Nd 3+ -doped and HPA-incorporated samples, which may be due to the low Nd 3+ and HPA content.The crystallite size of the prepared g-C 3 N 4 samples was calculated using the Debye-Scherrer'equation, DXRD = kλ/ βcos θ, where 'D' is the crystallite size, 'K' is a constant with a value of 0.9, 'λ' is the wavelength of X-ray used (0.154 nm), 'β' is the full width at half maximum intensity (FWHM) (in radians) of XRD diffraction lines and 'θ' is the half of diffraction angle 2θ [36].The results are given in Table 1, which shows that the crystallite sizes of the particles are around 5 nm, and there is no considerable change in the size of g-C 3 N 4 due to neodymium doping and HPA incorporation.

TEM analysis
Morphological features of the prepared photocatalysts were analysed using transmission electron microscopy (TEM).The TEM image of pristine g-C 3 N 4 (Figure 3(a)) shows freestanding nanosheets with a layered structure.The edges of the sheets tend to be ragged in order to minimise their surface area.Moreover, the selected area electron diffraction (SAED) pattern (Figure 3(b)) shows a faint but full diffraction ring, which is indexed to the characteristic (002) plane [39].The Nd 3+ -doped sample exhibited a similar morphology to pristine g-C 3 N 4 .Nd 3+ ions are not detected in the TEM images of Nd 3+ -g-C 3 N 4 (Figure 3(c)) as it is finely dispersed.SAED patterns of Nd 3+ -g-C 3 N 4 (Figure 3(d)) and Nd 3+ -g-C 3 N 4 -HPA (Figure 3(f)) show a bright and full diffraction ring, which indicates the improved nanocrystalline nature of the Nd 3+ -doped and HPA-incorporated samples compared to pristine g-C 3 N 4 .TEM images (Figure 3(e)) of Nd 3+ -g-C 3 N 4 -HPA display an aggregated surface morphology with a layered structure and may be due to the incorporation of HPA on the surface of graphitic carbon nitride.

BET analysis
Figure 4(a) presents the N 2 adsorption-desorption isotherms of g-C 3 N 4 , Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA.All the three samples present a type IV isotherm with a H3 hysteresis loop (Figure 4(a)), suggesting the presence of mesopores [40].The BET surface areas (S BET ) of g-C 3 N 4 , Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA are obtained to be 91.2899m 2 /g, 121.7652 m 2 /g and 82.2705 m 2 /g, respectively.It was found that the surface area of g-C 3 N 4 is significantly increased upon Nd 3+ doping, which is in good agreement with previous reports [25,36].However, addition of HPA resulted in the reduction of the surface area, as HPA molecules cover the surface of Nd 3+ -doped g-C 3 N 4 .An average pore diameter of 22.8 nm for Nd 3+ -doped g-C 3 N 4 -HPA has been estimated from the BJH pore size distribution (Figure 4(b)).The BJH desorption cumulative surface area and volume of pores of samples are given in Table 2.

XPS analysis
The chemical composition and oxidation state of the synthesised Nd 3+ -g-C 3 N 4 -HPA catalyst were investigated using X-ray photoelectron spectroscopy.The XPS survey spectrum (Figure 5(a)) of Nd 3+ -g-C 3 N 4 -HPA indicates that the catalyst is composed of graphitic carbon nitride, neodymium ion and HPA.The spectrum shows peaks at 284.75 eV and 395.86 eV corresponding to C1s and N1s, respectively [41,42].Furthermore, Nd (3d) peaks are frequently observed in the 980-1010 eV range and the

Diffuse reflectance spectra
The visible light absorption properties of the as-prepared g-C 3 N 4 , Nd 3+ -g-C 3 N 4 , and Nd 3+ -g-C 3 N 4 -HPA samples were studied by UV-vis diffuse reflectance spectroscopy and are shown in Figure 6.The π→ π* and n→ π* electronic transitions between N2p and C2p orbitals in graphitic carbon nitride are responsible for the absorption peak detected in the wavelength range 200 to 340 nm in spectra [46].A small broadband in the region of 350 to 390 nm was observed due to the electrons whose energy is in between valence and conduction band energies in g-C 3 N 4 [26].
The absorption intensity is found to be highest for Nd 3+ -g-C 3 N 4 -HPA compared to Nd 3+ -g-C 3 N 4 and g-C 3 N 4 .The band gap of g-C 3 N 4 is obtained to be 2.86 eV, which is in good agreement with the literature with respect to the absorption edge located around 433 nm [47].It is clear from Figure 6 that the absorption edges of Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA were found to be red shifted to 446 and 450.9 nm, respectively.Band gap energies of the photocatalysts were determined using the equation: Eg = 1239.8/λ,where 'Eg' is the band gap (eV) and 'λ' (nm) is the wavelength of the absorption edge.The nature of the band gap was obtained using the power expression, (αhυ) n = β(hυ-E g ), where β is a constant called the band tailing parameter, E g is the energy of the optical band gap and n is the power factor of the transition mode(n = ½ for indirect transition or n = 2 for direct transition), which is dependent upon the nature of the material, whether it is crystalline or amorphous [28].Figure 7 depicts the plots of (αhν) 2 versus photon energy (hν), and the bandgap energies obtained are close to the onset energy detected in the absorption spectra, confirming that the band gap is attributed to the direct transition (Table 3).The band gap values given in Table 3 clearly indicate that the band gap of g-C 3 N 4 was decreased upon Nd 3+ doping.The band gap reduction was mainly attributed to the formation of an intermediate level between the valance band (populated by N2p orbitals) and the conduction band (formed by C2p orbitals) by the neodymium 4 f orbital [33].This intermediate level helps to trap the photoexcited electrons and reduce the rate of electron-hole recombination, thereby increasing the photocatalytic activity of graphitic carbon nitride.However, no significant effect on the band gap of graphitic carbon nitride was observed upon the integration of HPA molecules.

Photoluminescence spectra
The charge carrier separation, migration and transfer of electrons and holes in the synthesised semiconductor photocatalysts were investigated using room temperature photoluminescence spectra.The PL emission spectra of g-C 3 N 4 , Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA samples recorded at an excitation wavelength of 380 nm are shown in Figure 8.All samples exhibit a strong and broad PL emission peak ranging from 400 to 480 nm, which is ascribed to the band gap emission [26].The intensity of this emission peak was decreased in Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA compared to undoped g-C 3 N 4 .The PL spectrum of Nd 3+ -g-C 3 N 4 -HPA displays lowest intensity, signifying highest separation efficiency of the electrons and holes, leading to maximum photocatalytic activity.This confirms that doping of Nd 3+ ions and incorporation of HPA molecules enable g-C 3 N 4 to achieve a considerable reduction in the recombination rate of photogenerated electron and holes.

Photocatalytic activity
To examine the solar photocatalytic activity of synthesised g-C 3 N 4 samples, methylene blue dye was used as a reference pollutant and was irradiated under solar light.Prior to activity studies, the suspension (pollutant + catalyst) was stirred in the dark for 30 min to  attain adsorption-desorption equilibrium and it was observed that the adsorption capacities of all samples are almost the same.Figure 9 clearly demonstrates that the photocatalytic efficiency of graphitic carbon nitride was significantly enhanced upon Nd 3+ doping.Nd 3+ ions, acting as Lewis acid, could act as an effective electron scavenger to trap the conduction band electrons of g-C 3 N 4. The optimum Nd 3+ content in g-C 3 N 4 for attaining maximum photocatalytic activity is found to be 0.25 × 10 −3 wt % and gives about more than 80% degradation of MB upon 90 min of solar light irradiation (Fig. S1).In Nd 3+-doped g-C 3 N 4 , the neodymium 4 f level plays a significant role in the interfacial charge transfer and the elimination of electron-hole recombination, which would be helpful for the high photocatalytic activity.The highest photocatalytic activity of Nd 3+ -g-C 3 N 4 under sunlight is mainly ascribed to red shift of absorption spectra and the higher absorption in the visible region, as evident from diffuse reflectance spectra.The increase in the surface area and pore volume upon neodymium doping also contributes to the enhancement of its photocatalytic activity.Furthermore, when HPA is incorporated with Nd 3+ -doped g-C 3 N 4 , the photocatalytic activity is found to be significantly increased.HPA could appreciably prevent the recombination of electron-hole pairs and was in accordance with PL intensity.Surface-adsorbed HPA could act as an electron pool to which electrons trapped in the neodymium ion intermediate level formed between N2p orbitals and C2p orbitals were subsequently transferred and made available for the reaction with adsorbed MB.Nd 3+ -g-C 3 N 4 -HPA could degrade about more than 90% of MB within 75 min of solar irradiation.The enhanced photocatalytic performance of Nd 3+ -g-C 3 N 4 -HPA compared to that of Nd 3+ -doped g-C 3 N 4 and g-C 3 N 4 is mainly due to the synergistic effect of decreased bandgap energy by Nd 3+ doping and increased lifetime of photogenerated electron-hole pairs by HPA molecules.UV-visible absorption spectra of the time-dependent photodegradation of MB (Figure 10) show that high-intensity peaks at 665 nm correspond to the auxochrome group of MB.The intensity of this peak is found to be gradually decreased upon solar light illumination, which indicates the decolouration and degradation of MB [48].Photographs of degradation of MB by g-C 3 N 4, Nd 3+ -g-C 3 N 4 and Nd 3+ -g-C 3 N 4 -HPA are given in Figure 11.

Effect of pH and pH zero-point charge (pHzpc)
The pH of the pollutant causes a significant impact on the photocatalytic degradation process and is one of the most critical factors influencing the adsorption of pollutant on the catalyst.In this study, the effect of pH was studied by adjusting initial pH of the MB  dye solution to a range of pH 5 to 11, using dil.solution of 0.2 M HCl and NaOH.When the pH was adjusted to the acidic range, the photoactivity of Nd 3+ -g-C 3 N 4 -HPA was found to be decreased, whereas in basic pH, the photoactivity of Nd 3+ -g-C 3 N 4 -HPA was increased (Figure 12).Furthermore, the pH zero-point charge (pHzpc) value of the Nd 3+ -g-C 3 N 4 -HPA catalyst was determined to be 5.3 using the immersion technique (Figure 13) [49].When  the pH of the solution is less than pHzpc, the Nd 3+ -g-C 3 N 4 -HPA has a positive surface charge but becomes negative above this threshold.At lower pH (pH < pH ZPC ), the surface of the catalyst gets positively charged and tends to oppose the adsorption of cationic MB molecules and decreases the photoactivity of Nd 3+ -g-C 3 N 4 -HPA.However, at higher pH (pH > pH ZPC ), the surface tends to acquire negative charge, resulting in an increased adsorption of MB due to increasing electrostatic attraction between dye and the catalyst to enhance the activity [50].Also, it is reported that in alkaline medium, the generation of reactive species such as superoxide radicals is considerably increased, resulting in a faster reaction rate [51].

Scavenging studies
The active species or radicals generated during the photocatalytic degradation of pollutants were identified by the hole and free-radical scavenging experiment.To understand the reaction mechanism in detail, active species-capturing experiment was carried out by the degradation of MB dye solution.Isopropyl alcohol (IPA), ascorbic acid (AA) and triethanolamine (TEOA) were used as scavenging reagents for hydroxyl radicals, superoxide anion radicals and holes, respectively, and are added to the reaction medium (20 ml  of 30 ppm MB + catalyst) separately.A slight reduction in the rate photodegradation of MB was observed after the addition of IPA and implies that hydroxyl radicals are not the major active species in the present photocatalytic systems.Addition of TEOA causes a sharp decrease in the degradation rate of MB, indicating that holes are mainly responsible for the MB degradation (Figure 14).The potential of the CB and VB of g-C 3 N 4 was calculated to be −1.2V and +1.66 V, respectively, using the empirical formula: E CB = X -Ec -1/2Eg and E VB = E CB + Eg, where X is the absolute electronegativity of the atom semiconductor, Ec is the energy of free electrons of the hydrogen scale (4.5 eV) and Eg is the band gap of the semiconductor [52].Thus, the reduction potential of CB electrons in g-C 3 N 4 was more negative than the redox potential of O  , whereas the VB holes in g-C 3 N 4 were not positive enough to produce • OH [53].Therefore, the experimental results are in accordance with the fact that both superoxide anions and holes are the major active species in the photocatalytic degradation process and fewer hydroxy radicals are involved in the degradation.

Proposed mechanism of pollutant degradation in presence of Nd 3+ -g-C 3 N 4 -HPA
The photocatalytic degradation mechanism (Figure 15) was proposed to explain the improvement of photocatalytic performance of graphitic carbon nitride after the incorporation of Nd 3+ ions and HPA molecules.When the photocatalytic medium (photocatalyst + pollutant) is exposed to sunlight, electrons are excited and transferred from the valence band (VB) to the conduction band (CB) of g-C 3 N 4 and produce photoexcited electrons and holes [Eq.( 2)].Rapid recombination of photoexcited electrons and holes occurs in pristine g-C 3 N 4 and shows reduced activity.But in the Nd 3+ -g-C 3 N 4 -HPA system, Nd 3+ ions with a vacant 4 f orbital trap photogenerated electrons and are transferred into the surface-adsorbed HPA molecules, whichact as an electron pool [34].These electrons also reduce Nd 3+ to Nd 2+ [Eq.( 3)].Moreover, the lanthanides have the capacity to stock oxygen and neodymium can release oxygen to the surrounding system, whenever the concentration of oxygen in the surrounding is low [26,54].Utilising this oxygen, Nd 2+ could be oxidised back to Nd 3+ along with the formation of superoxide anion radicals [Eq.( 4)].The photogenerated holes react with water molecules to form protons and hydroxy radicals [Eq.( 5)].Surface-adsorbed OH − ions may react with generated holes and form hydroxy radicals [Eq.( 6)].The photogenerated electrons on reaction with the dissolved oxygen (O 2 ) produce • O 2 ▬ radicals [Eq.(7)].The overall reaction can be summarised as follows. .

Degradation pathway of antibiotic ceftriaxone sodium by the Nd 3+ -g-C 3 N 4 -HPA photocatalyst
Degradation of the antibiotic ceftriaxone sodium was analysed with the synthesised samples, under solar light, and the results are shown in Figure 16.Complete degradation of ceftriaxone sodium was achieved after 75 min of solar light irradiation in the presence of the Nd 3+ -g-C 3 N 4 -HPA photocatalyst.The HPLC chromatograms of ceftriaxone sodium before and after irradiation are given in Figure 17.Characteristic peak A in the HPLC chromatogram of ceftriaxone sodium before irradiation was found to be absent after 75 min of irradiation of solar light, which indicates that ceftriaxone sodium was completely degraded into CO 2 , H 2 O and other molecules [55].To examine the various intermediates generated and possible reaction pathways, samples were withdrawn at 0, 25 and 75 min of solar irradiation and mass spectra of these samples were recorded with the LC-MS system.The plausible routes of the degradation pathway of ceftriaxone sodium are sketched in Figure 18.A protonated ion peak at m/z 599 (A) due to the loss of 3.5 H 2 O molecules from the ceftriaxone sodium (Molecular formula; C 18 H 16 N 8 Na 2 O 7 S 3 •3.5H 2 O, Molecular weight; 661.59) was obtained, and this parent ion peak at m/z 599 (A) was found to be decreased with the formation of some additional peaks, viz., m/z 577.1 (C), m/z 555.1 (F), m/z 414.1 (B) and m/z 396.15 (D), which indicates that the various intermediates are formed during the photocatalysis of ceftriaxone sodium with the irradiation of solar light.The product at m/z 577 was due to the loss of the sodium atom from the carboxylate group, and the product at m/z 555.1 was due to the removal of another sodium atom from the triazine ring of ceftriaxone sodium [56].Intermediate product D (m/z 396.15) could be further degraded to G (m/z 100.05),CO 2 and H 2 O. Simultaneously, intermediate product B (m/z 414.1) could be degraded to E (m/z 181), CO 2 and H 2 O as reported in the literature [55,57].The most intense peak at m/z 555.15 in initial reaction medium was found to be absent upon photoirradiation in the presence of the catalyst.The study revealed that the species at m/z 555.15  The study demonstrates that pathway I and corresponding intermediates are similar to the previous reports on electrocatalytic and photocatalytic degradation of ceftriaxone sodium [55][56][57][58][59].However, the pathway II and corresponding fragments were not observed in their studies.

Recycling capacity of the catalyst
For assessing the stability and reusability, the catalyst, Nd 3+ -g-C 3 N 4 -HPA, was used for the degradation of different solutions of MB in three consecutive experiments by maintaining the same reaction environment (0.01 g of catalyst + 20 ml of 30 ppm MB, 75 min of solar irradiation).After each experiment, the catalyst was recovered by centrifugation, and prior to next experiment, it was dried in an air oven and powdered well.The photocatalysts retain their high catalytic activity after being recycled six times as shown in Figure 19.FT-IR spectra show that the spectrum of the recycled photocatalyst after photodegradation reactions was the same as that of the fresh photocatalyst (Figure 20).The results indicate that incorporation of Nd 3+ ions and HPA molecules with g-C 3 N 4 has increased its chemical stability and reusability.This novel photocatalyst, Nd 3+ -g-C 3 N 4 -HPA, is highly effective in degrading the antibiotic drug, ceftriaxone sodium under solar light, and it can be a potential catalyst for the degradation of organic pollutants, which are persistent in the environment.

Conclusions
In this work, we successfully synthesised a novel photocatalyst, Nd 3+ -doped graphitic carbon nitride, incorporated with heteropolyphosphotungstic acid by a facile thermal polymerisation method.UV-vis diffuse reflectance spectral analysis confirms that Nd 3+ doping and HPA incorporation led to an increase in absorption intensity, which enhances a wide range of visible light absorption.Photoluminescence studies revealed the improved separation efficiency and reduction in the recombination rate of photogenerated electrons and holes.Photocatalytic efficiency of the synthesised catalysts was assessed by the degradation of MB under sunlight, and it was found that Nd 3+ doping and HPA addition cause a significant enhancement in the activity.This photocatalyst was found to be efficient in degrading antibiotic drug, ceftriaxone sodium, and achieved complete degradation within 75 min under solar light irradiation.Also, the intermediates that are generated during the degradation process were identified, which showed the possible degradation pathways of ceftriaxone sodium.Superoxide anion radicals and holes play a significant role in the photocatalytic degradation of pollutants.This study suggests that the synthesised catalyst can be used to remove persistent organic pollutants such as dyes and pharmaceutical compounds from surface water.

Nd3d 3 /
2 core level has a peak at 1001.9 eV.The peak at 1008.95 eV in the spectrum (Figure5(b)) indicates that neodymium ions in the Nd 3+ -g-C 3 N 4 -HPA catalyst bear a +3-oxidation state[43,44].Low-intensity peaks at 35.9 eV and 134.4 eV (Figure5(c) and (d)) correspond to the 4 f level of tungsten (W4f 7/2 ) and 2p level of phosphorous in heteropoly phospotungstic acid, respectively[45].The low intensities of these peaks are due to the very low content of HPA molecules in the Nd 3+ -g-C 3 N 4 -HPA catalyst.Thus, the XPS spectrum confirms that the neodymium ion and HPA are successfully incorporated into graphitic carbon nitride.

Figure 5 .
Figure 5. (a) XPS Survey spectrum of Nd 3+ -g-C 3 N 4 -HPA and (b, c and d) enlarged views of the XPS survey spectrum of Nd 3+ -g-C 3 N 4 -HPA in different binding energy (eV) ranges.

Figure 10 .
Figure 10.UV-visible absorption spectra of the time-dependent photodegradation of methylene blue in the presence of Nd 3+ -g-C 3 N 4 -HPA.

Figure 11 .
Figure 11.Photographs of methylene blue, before and after irradiated with solar light in the presence of different catalysts.

Figure 12 .
Figure 12.Effect of pH on the photocatalytic activity of Nd 3+ -g-C 3 N 4 -HPA under various pH conditions.

Figure 14 .
Figure 14.Photocatalytic degradation of MB in the presence of Nd 3+ -g-C 3 N 4 -HPA and different radical scavengers.

Figure 17 .
Figure 17.HPLC (254 nm) chromatograms of 50 ppm ceftriaxone sodium before and after solar light irradiation in the presence of Nd 3+ -g-C 3 N 4 -HPA.

Figure 19 .Figure 20 .
Figure 19.Cyclic runs of Nd 3+ -g-C 3 N 4 -HPA for the degradation of MB under solar irradiation.
follows two degradation pathways; Pathway I: The fracture of the β-lactam ring leads to the formation of two intermediated products H (m/z 243.15) and L (m/z 261).Intermediate product H was fragmented into G (m/z 100.05) with the C-C bond rupture, whereas intermediate L was fragmented into species N (m/z 159) with C-S bond rupture.Pathway II: The second pathway is through the direct cleavage of C-S and C-N bonds in ceftriaxone sodium, which leads to the formation of fragment M (m/z 225.15) and gets converted into O (m/z 118) and P (m/z 74) through β-lactam ring rupture of M. Finally, all the intermediate products are degraded to CO 2 , H 2 O and other small, nontoxic molecules.

Table 1 .
Crystallite sizes of the synthesised samples.

Table 3 .
Band gap values of synthesised samples. 2