Synthesis, characterization and properties of sulfate-modified silver carbonate with enhanced visible light photocatalytic performance

Abstract

Sulfate-modified Ag₂CO₃ was successfully synthesized via a simple precipitation method. Its visible light photocatalytic performance against the removal of Orange G was found to be significantly enhanced in comparison with the one of pure Ag₂CO₃. While SO₄²⁻-Ag₂CO₃ ensured a removal efficiency of 100% of OG within 30 min, the unmodified Ag₂CO₃ exhibited a degradation threshold at hardly 60%. Likewise, the degradation rate constant in the presence of SO₄²⁻-Ag₂CO₃ photocatalyst was assessed to be twice that determined upon the involvement of pristine Ag₂CO₃. Furthermore, Total Organic Carbon (TOC) measurements evidenced the occurrence of a quasi-total mineralization of the dye pollutant upon the use of SO₄²⁻-Ag₂CO₃ photocatalyst. Scavenger experiments highlighted the dominant role of photo-induced h⁺ along with ˙O₃⁻ ozonide radicals in the OG photocatalytic oxidation mechanism. Reuse cycles revealed that the modification by SO₄²⁻ is a promising route to improve the stability of silver carbonate against photocorrosion. All these improvements could be ascribed to electronic transfer from the upper SO₄²⁻ HOMO to the lower Ag₂CO₃ conduction band.

---

1 Introduction

Great attention was recently paid to silver based photocatalysts such as Ag₂O,¹ AgX (X = Cl, I and Br),² Ag₃PO₄ (ref. 3) and Ag₂CO₃.⁴ This is due to their high visible light absorption that yields an effective photocatalytic degradation of organic pollutants from aqueous media.^5,6 Nevertheless, their photoactivity efficiency depends on the anion nature since for instance, Ag-oxosalt photocatalytic performances increased in the order Ag₃AsO₄ > Ag₃PO₄ > Ag₂CO₃ > Ag₂SeO₄ > Ag₂SO₄.⁷ However, this order seemed to be controversial in that a recent study reported that Ag₂CO₃ appeared more photoactive and photostable than Ag₃PO₄.⁸ Likewise, this order depended on the nature of the pollutant.⁸ Despite this, Ag₃PO₄, has been the most studied in photocatalysis with about 900 research papers dedicated to this compound compared to about 150 for Ag₂CO₃ in the same period.⁹ This highest interest for Ag₃PO₄ was likely due to its better thermal stability and higher photoactivity, which could be further enhanced by doping.

In this context, several studies reported Ag₃PO₄ doping with Ni²⁺ (ref. 10) or Cu²⁺ (ref. 11) cations, and SO₄²⁻ (ref. 12) or CO₃²⁻ anions.¹³ Considering anion doping, it is difficult to claim the achievement of the doping of an Ag-oxosalt by a polyatomic anion since the authors do not clearly explain how 5 ions (S⁶⁺ and O²⁻) issued for instance from SO₄²⁻ group were distributed in the crystalline matrix. A more generic term than doping could be activation or treatment of the photocatalyst. For example, a so-called doping by SO₄²⁻ could be similar to a doping by sulfur. Anyway, in the particular case of Ag₃PO₄ doping by SO₄²⁻, it was suggested that an electronic transfer occurred between SO₄²⁻ and PO₄³⁻, more precisely from atomic orbitals S 3s²3p⁴ to P 3s²3p³.¹² The photocatalyst Ag₂CO₃ crystallizes in a monoclinic structure significantly different than the cubic structure of Ag₃PO₄. In addition, it presents an optical band gap (≈2.4 eV) in the same order of magnitude as the silver phosphate and thus is photoactive under visible light irradiation.⁴ Nonetheless, in comparison with Ag₃PO₄, it was less studied as photocatalyst, probably due to the fact that, as other silver oxosalts, it exhibited drawbacks related for instance to a lower transfer rate of charge carriers¹⁴ and a higher sensitivity to photocorrosion.¹⁵ By contrast with Ag₃PO₄, fewer strategies have been explored to improve the performance of Ag₂CO₃.

To overcome these issues, several routes have been proposed such as heterojunctions between Ag₂CO₃ and another semiconductor, which can promote high separation of the photo-generated electrons–holes pairs. This approach has been extended to nanocomposites with metal nanoparticles as Ag to produce Schottky junctions. For this purpose, beneficial heterojunctions were reported for TiO₂/Ag₂CO₃,¹⁶ ZnO/Ag₂CO₃,¹⁷ Ag/Ag₂CO₃,¹⁸ Ag₂O/Ag₂CO₃,¹⁹ Ag₃PO₄/Ag₂CO₃ (ref. 20) and AgX (X = Cl, I and Br)/Ag₂CO₃.²¹ Another way²² reported the effect of varying types and ratios of solvents upon the synthesis in tunning the size and morphology of Ag₂CO₃ and consequently improving its photocatalytic activity and photostability. In another route, it was reported that the immobilization of Ag₂CO₃ particles on supports with a large specific surface area such as graphene oxide,²³ C₃N₄ (ref. 24–26) and some clay minerals like palygorskite²⁷ was also a promising way. Surprisingly, in contrast to Ag₃PO₄, anionic modification of Ag₂CO₃ was not investigated. In particular, the route of modification by SO₄²⁻ was not yet reported for improving photocatalytic properties of Ag₂CO₃.

Since both Ag-oxosalts differ in their crystallographic structure and anion nature (size, charge, composition), the response of these photocatalysts to a strategy will not necessarily be the same from one to another. The fact that SO₄²⁻ presents an ionic radius of the same order of magnitude than CO₃ (ref. 2–28) should facilitate bulk modification of Ag₂CO₃. However, the differences both in crystallographic structure between Ag₂CO₃ and Ag₃PO₄, and in charge transfer expected between the modifying anion (SO₄²⁻) and crystal lattice anion (CO₃²⁻), does not allow to think that the effects of the modification by SO₄²⁻ should be similar for Ag₂CO₃ and Ag₃PO₄.¹² These differences increase the interest of exploring Ag₂CO₃ modification with SO₄²⁻ as a route to enhance photocatalytic properties and paving a new way to improve the stability against photocorrosion for practical applications.

In this paper, the photocatalytic performances were assessed (i) by measuring the degradation rate of a model pollutant under visible light irradiation (kinetic data), (ii) by analyzing the extent of mineralization (TOC analysis), (iii) by determining the nature of photogenerated oxidizing species responsible of the dye photocatalytic degradation through scavenger experiments (resulting in hypothesis on the reaction mechanism) and (iv) by studying the photo-stability in reuse experiments.

2 Experimental

2.1 Synthesis of the photocatalysts

All reagents were of analytical grade and used without further purification. Pure Ag₂CO₃ was synthesized according to a method previously published.^4,27 For Ag₂CO₃ treatment with SO₄²⁻, the method was adapted from that already reported for sulfate-doped Ag₃PO₄.¹² In a typical procedure, 20 mL of a translucent mixture of AgNO₃ (0.272 M) and Na₂SO₄ (0.0125 M) was prepared at room temperature in distilled water and stirred during 10 min for homogenization. Then 40 mL of Na₂CO₃ (0.068 M) was added dropwise under continuous stirring. The formed precipitate (i.e. SO₄²⁻-modified Ag₂CO₃) was collected by centrifugation, washed with distilled water and ethanol three times to remove reaction byproducts mainly NaNO₃, Na₂SO₄ and traces of remaining reactants. The samples were dried at 60 °C overnight, then stored in the dark at room temperature before being used. With these synthesis conditions, the theoretical SO₄²⁻ content of the photocatalyst cannot exceed 9 at% since this was the value of the SO₄²⁻/CO₃²⁻ anion ratio used.

2.2 Characterization techniques

The phase identification was carried out by X-ray diffraction (D8 Advance Diffractometer Bruker) with Cu Kα radiation source. The scanning was performed at room temperature in the 2θ range from 5° to 70° with a step size of 0.05° for 1 s. Infrared spectroscopy analysis was performed on a pellet of KBr–photocatalyst mixture (≈1 wt%) using a PerkinElmer FTIR in the range 400–4000 cm⁻¹. The SEM microscopy observations were performed using a scanning electron microscope VEGATE-SCAN3 coupled with an EDAX analyzer for elemental analysis. UV-vis Diffuse Reflectance Spectra (DRS) were recorded with respect to PTFE as standard using a Cary 5000 UV-vis-NIR spectrophotometer for wavelengths ranging from 200 to 800 nm. The zeta potential of the synthesized Ag-oxosalt powders was measured at room temperature in an aqueous suspension using a Zetasizer apparatus from Nano ZS, Malvern instruments.

2.3 Photocatalytic tests under visible light

The photocatalytic activity was evaluated by recording the temporal variation of Orange G (OG) concentration during its degradation under visible light irradiation in presence of the photocatalyst. OG is an anionic azo dye with the formula C₁₆H₁₀N₂Na₂O₇S₂. It was selected as model pollutant because it is frequently used as dye in textile industry, and it does not undergo photolysis under the conditions of this test. The visible light irradiation was provided by 4 light color/840 lamps (13 W each) coupled with a UV cut-off filter (λ < 400 nm). Practically, 25 mg of the photocatalyst was dispersed into a quartz photo-reactor containing 25 mL of OG (10⁻⁵ M) aqueous solution. It should be noted that such conditions were the same as the ones already established by Lakbita et al.²⁷ to be optimal for the study of photocatalytic performances under visible light of Ag₂CO₃ supported palygorskite clay mineral in a similar slurry reactor. Afterwards, the suspension was stirred in the dark for 30 min to achieve adsorption–desorption equilibrium before to start the photocatalytic test by switching on the lamps. At regular time intervals, aliquots were picked up, centrifuged and OG concentration in supernatant was determined using UV-vis spectrophotometer from the intensity of the OG absorption band in the range 470–500 nm. It should be noted that a calibration curve in the Beer–Lambert linearity domain was checked beforehand in the low concentration range (10⁻⁷–10⁻⁴ M) that includes the test value (10⁻⁵ M). The experimental data reported here were the average of 3 photocatalytic tests and an error bar represented the dispersion of data.

To check if the photocatalytic degradation of OG organic dye pollutant led to its partial or total mineralization, the Total Organic Carbon (TOC) corresponding to remaining organic dye plus possibly organic byproducts was determined after the test using a TOC analyzer (TOC-L CPH/CPN, Shimadzu). For this purpose, after the test, photocatalyst dispersion was centrifuged at 12 000 rpm during 5 min and the resulting supernatant was analyzed.

To further investigate the photocatalytic mechanism that would ensure the OG photocatalytic degradation, isopropanol (IPA) was utilized to quench hydroxyl radicals (˙OH),¹² trichloromethane (TCM) served as scavenger of superoxide radical anions (˙O₂⁻),²⁹ indigo carmine (IC) used to trap the ozonide radicals (˙O₃⁻)⁸ and ethylenediamine tetra-acetic acid disodium (EDTA) was used as a hole (h⁺) scavenger.³⁰ Before adding Ag₂CO₃ or SO₄²⁻-Ag₂CO₃ powders and running photocatalysis experiment, 1 mM of each scavenger was added to the OG dye solution.³¹

3 Results and discussion

3.1 Bulk modification of Ag₂CO₃ by SO₄²⁻

As shown in Fig. 1, XRD patterns recorded on unmodified and SO₄²⁻ modified Ag₂CO₃ appeared quite similar with several diffraction peaks amongst which the ones at 2θ around 18.32° (d₁₀₀ = 4.85 Å), 18.54° (d₀₂₀ = 4.78 Å), 20.5° (d₁₁₀ = 4.32 Å), 32.74° (d 1̄ 01 = 2.73 Å), 33.81° (d₁₃₀ = 2.65 Å), 37.07° (d₂₀₀ = 2.42 Å) and 39.60° (d₀₃₁ = 2.27 Å). These peaks were indexed on the basis of the stable monoclinic Ag₂CO₃ phase (ICDD file no. 01-070-2184). This result confirmed that the Ag₂CO₃ treatment by SO₄²⁻ had no effect on the crystallographic structure of the silver carbonate that remained monoclinic while several polytypes exist.²⁷ Furthermore, the non-appearance of additional peaks in the SO₄²⁻ modified Ag₂CO₃ pattern denoted the absence of any crystallized secondary phases as byproducts from the base reactions (e.g. Ag₂SO₄) or from side decomposition reactions (e.g. Ag₂O, Ag). Regarding Ag₂SO₄, the corresponding main XRD peaks, expected at 31.13° (311) and 28.08° (040), were not present (Fig. 1). This is consistent with the fact that it is water-soluble and was therefore washed away from the final photocatalyst powder (see Section 2.1).

Fig. 1 X-ray diffraction patterns of pure Ag₂CO₃ and SO₄²⁻ modified Ag₂CO₃ (left) and zoom in the 32.0°–34.5° angular region corresponding to (1̄ 01) and (130) reflection planes of both Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ (right).

Further detailed analysis of the diffractograms interestingly revealed a slight shift of the peaks towards smaller 2θ angles. Indeed, in the 2θ region between 32.0° and 34.5°, (1̄ 01) and (130) peaks were observed respectively at 32.74° and 33.81° for pure Ag₂CO₃ while they appeared at 32.64° and 33.69° for SO₄²⁻ modified Ag₂CO₃ (Fig. 1b). The values of a, b, and c lattice parameters and β angle of unmodified- and SO₄²⁻ modified Ag₂CO₃ samples, were calculated from all the XRD peaks, using the Unit Cell software.³² They were reported in Table S1 (ESI).† They revealed an increase of the lattice parameters for SO₄²⁻-modified Ag₂CO₃ with respect to unmodified Ag₂CO₃ corresponding to a volume expansion of the unit cell of ca. 1.0%, which was certainly due to the higher ionic radius of SO₄²⁻ (0.218 nm) compared to the one of CO₃²⁻ (0.189 nm).²⁸ This finding ascertained that CO₃²⁻ anions in the crystal lattice were substituted by modifying SO₄²⁻ anions larger by 15% and hence confirmed the achievement of a bulk modification of Ag₂CO₃ by sulfate anions. This ruled out a simple surface functionalization by SO₄²⁻ anions grafted on Ag₂CO₃ particles, which should not induce an expansion of the crystal lattice.

To complete the crystallographic analysis, the two photocatalysts were characterized by FTIR spectroscopy. In addition to the presence of H₂O molecules that was revealed by absorption bands around 3200 and 1650 cm⁻¹, both FTIR spectra exhibited characteristic absorption bands of Ag₂CO₃ (Fig. 2). The two intense bands at 1442 and 1377 cm⁻¹, as well as those at 879 and 697 cm⁻¹ may be assigned to the distinctive vibrations of the planar CO₃²⁻ groups, i.e. deformation and stretching, respectively.³³ Additional bands were exclusively detected in SO₄²⁻-modified Ag₂CO₃ sample with bands at 1113, 1056, and 606 cm⁻¹ attributed to the stretching mode of S–O and O–S–O bonds in SO₄²⁻ tetrahedra, as reported for Ag₂SO₄ elsewhere.³⁴ Except the bands of CO₃²⁻ and SO₄²⁻ anions (and H₂O traces), no other absorption band was observed in the sulfate modified Ag₂CO₃ photocatalyst. This result further supported the beforehand mentioned statement issued from XRD analysis according to which on one hand a bulk modification of Ag₂CO₃ with SO₄²⁻ anions was successfully achieved rather than a sulfate anion grafting on the surface of Ag₂CO₃ particles and on the other hand the absence of secondary phases formation.

Fig. 2 FTIR spectra of pure Ag₂CO₃ (bottom) and SO₄²⁻ modified Ag₂CO₃ (top). The absorption bands were assigned according to literature data of carbonates and sulfates including Ag₂CO₃ ³³ and Ag₂SO₄.³⁴

SEM analyses showed that pure Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ samples exhibited the same microstructure with micron-size and monodisperse particles with pretty similar polyhedral shapes (Fig. 3). They appeared well crystallized, in the form of facetted rhombohedra with an average size of 0.3 μm according to the Dynamic Light Scattering (DLS) particle size distribution analysis (Fig. S1†). This microstructure was similar to that previously reported for silver carbonate synthesized by the same method.²⁷ Therefore, SO₄²⁻-modification did not change the microstructure of Ag₂CO₃ particles (Fig. 2). This should be consistent with the incorporation into the Ag₂CO₃ crystal lattice of only a very small amount of SO₄²⁻ anions. This assumption was further supported by EDS analysis showing the SO₄²⁻-Ag₂CO₃ sample was mainly composed of Ag, C, O. Very small peak, corresponding to sulfur, was detected in the energy range (Fig. 3d). Moreover, the SEM-EDS elemental mapping of the SO₄²⁻-Ag₂CO₃ powder spread on the sample holder did not reveal any evidence for S-rich regions, which additionally confirmed the absence of secondary phase such as Ag₂SO₄ (Fig. 3). Thus, the S amount in SO₄²⁻-Ag₂CO₃ sample was estimated to be around 1.8 atomic% fraction, which is way lower than the 9% allowed by the initial proportion of reagents and confirms the Ag₂CO₃ modification with SO₄²⁻.

Fig. 3 SEM micrographs of pure Ag₂CO₃ (left) and SO₄²⁻ modified Ag₂CO₃ (right), and the EDS spectra corresponding to each image.

UV-vis diffuse reflectance spectra recorded on pure Ag₂CO₃ and SO₄²⁻ modified Ag₂CO₃ samples did not reveal significative differences. For both the samples, they displayed a high and sharp absorption below 500 nm, in the visible light range, denoting as expected their wide band gap semiconductor characters (Fig. 4a). In both the cases, there was no evidence of absorption band above 500 nm that could be due to plasmonic effect induced by Ag nanoparticles as reported elsewhere.³⁵ The recorded UV-vis spectra herein were in good agreement with those previously reported for monoclinic Ag₂CO₃.^4,8,15 The absorption edges, corresponding to the intersection of the extrapolation of the linear part of the absorption curve with the wavelength axis, were determined to be around 509 nm (2.44 eV) and 533 nm (2.33 eV) for pure Ag₂CO₃ and SO₄²⁻-modified Ag₂CO₃ samples, respectively (Fig. 4a).

Fig. 4 UV-vis DRS spectra of Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ samples (a), and corresponding plot (αhν)² vs. energy hν (b). Red straight lines are extrapolations to determine optical gap (absorption edge) from (a) and the Tauc band gap energy from (b).

As both the Ag₂CO₃-based compounds are indirect semiconductors according to previously reported studies,^4,7,8 the corresponding band gap energies (E_g) determined by the Tauc method⁴ were found to be about 2.62 and 2.56 eV, for unmodified and SO₄²⁻ modified Ag₂CO₃ respectively (Fig. 4b). These E_g values were consistent with those of Ag₂CO₃ which were reported over a fairly wide range from 2.30 eV¹⁵ to 2.62 eV²⁰ depending on the synthesis method, the crystallographic structure, the microstructure and purity (Table S2 in ESI†). In this respect, as beforehand evidenced that the treatment by SO₄²⁻ did change neither the structure (Fig. 1) nor the microstructure of Ag₂CO₃, the slight decrease of E_g (≈2.3%) between both the band gap energies may result from the anionic modification effect. This SO₄²⁻ effect contrasted with that exerted by Ag₃PO₄ doping by SO₄²⁻ yielding rather to a slight increase of E_g as reported elsewhere.¹² Therefore, SO₄²⁻-modified Ag₂CO₃ photocatalyst exhibited a slightly better absorption of the photons in the visible range. This means that, as more discussed below, more photo-generated electrons (e⁻) were transferred from the valence band (VB) to the conduction band (CB), leaving more holes (h⁺) in the VB that will be capable of oxidizing more organic species in primary reactions.

Furthermore, the valence band position (E_VB) and the conduction band position (E_CB) of SO₄²⁻-Ag₂CO₃ can be computed through the Mulliken electronegativity empirical equations:¹⁵

E_VB = χ − E₀ + 0.5E_g (1)

E_CB = E_VB − E_g (2)

where χ is the semiconductor absolute electronegativity (6.023 eV for Ag₂CO₃ (ref. 8 and 15) and E₀ is the free electron energy on the hydrogen scale, otherwise the Fermi level of NHE (Normal Hydrogen Electrode) with respect of the vacuum level (ca. 4.5 eV). E_VB and E_CB are the VB and CB edge potentials respectively, and E_g is the band gap of the semiconductor. The E_VB and E_CB values hence deduced for SO₄²⁻-modified Ag₂CO₃ were +2.80 and +0.24 eV, respectively. For pure Ag₂CO₃ the values found were +2.83 and +0.21 eV respectively. These values were in agreement with literature data for Ag₂CO₃ based compounds (Table S2†).

3.2 Visible-light photocatalytic properties

The variation of OG dye concentration versus irradiation time under visible light in presence of pure Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ photocatalysts was reported in Fig. 5. It is worth noting that keeping beforehand in the dark during 30 min the dispersion of the two photocatalysts in OG aqueous solution did not reveal OG adsorption onto the surface of the two photocatalysts due to electrostatic repulsions between anionic OG dye molecules and negatively charged surfaces of the photocatalysts. This was confirmed by the zeta potentials measured to be −23 mV for pure Ag₂CO₃ and −27 mV for SO₄²⁻-Ag₂CO₃, which was consistent with that reported elsewhere (−36 mV) for the pure silver carbonate.⁷ Furthermore, the concentration of OG solution free from photocatalyst appeared not decreasing during the whole duration of its exposure to visible light. This proved that OG dye did not undergo photolysis (Fig. 5a).

Fig. 5 Variation of the relative OG concentration vs. irradiation time over Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ (a). Plots of −ln(C_t/C₀) vs. irradiation time for the two photocatalysts (b). Proportion of Total Organic Carbon (TOC) removal analyzed upon OG photocatalytic degradation under visible light irradiation during 30 min (c). The errors bars in (a) and (c) correspond to the data dispersion from 2 experiments.

In the presence of pure Ag₂CO₃, OG concentration considerably decreased during the first 20 minutes of irradiation to reach an OG degradation plateau after a removal of ca. 60% (Fig. 5a) beyond which no more dye photocatalytic degradation occurred. This striking behavior could be ascribed either to (i) surface Ag₂CO₃ poisoning by byproducts issued from photocatalytic reactions, or (ii) to photocorrosion of the photocatalyst induced by premature Ag₂CO₃ decomposition under light irradiation. The second assumption seemed to be of minor impact since the catalyst has been found still active in reuse experiments as further discussed in the Section 3.3. The observation of photocatalytic degradation threshold has already been reported upon assays involving pure Ag₂CO₃ against Methyl Orange (MO) anionic dye,^7,8 and Rhodamine B (RhB)^7,8,18 and Methylene Blue (MB)⁷ cationic dyes despite their electrostatic attraction by negatively Ag₂CO₃ surfaces. In contrast, the most salient result from the use of SO₄²⁻-Ag₂CO₃ photocatalyst was that the relative OG concentration continuously decreased during the photocatalytic test to reach remarkably a total removal of the dye pollutant within 30 min (Fig. 5a). This denoted its relative better photostability, i.e. a higher resistance to both surface poisoning and photo-corrosion issues as below evidenced in Section 3.3.

For both the pristine photocatalyst and SO₄²⁻-modified photocatalysts, OG dye photocatalytic degradation obeyed a pseudo first-order kinetics during approximately the first 15 min of visible light illumination. The rate constant assessed in this time range was about 5.1 × 10⁻² min⁻¹ for pure Ag₂CO₃ and 7.8 × 10⁻² min⁻¹ for SO₄²⁻-Ag₂CO₃ (Fig. 5b). This result denoted that the photocatalytic activity of the SO₄²⁻-modified sample was ca. 1.5 times faster than that of pure Ag₂CO₃.

The mineralization of OG, i.e. its ability to be converted into H₂O and CO₂ upon photocatalytic degradation, was also investigated. Fig. 5c depicted the total organic carbon (TOC) removal analyzed in the supernatant aqueous solution after a 30 min photocatalytic test using each photocatalyst. In presence of Ag₂CO₃, only around 52% of the TOC was removed, otherwise the proportion of OG and eventually its organic byproducts originated from photocatalytic reactions which was able to be mineralized. It should be noted that this OG amount was of the same magnitude order than that (≈60%) previously assessed by in situ UV-VIS spectrophotometry during the photocatalyst test. Nevertheless, it is worth noting that upon the use of SO₄²⁻-Ag₂CO₃, almost 94% of the TOC was removed, indicating a relatively huge mineralization of OG dye. Interestingly, it should be noted that this amount was fairly close to the 100% of OG photo-catalytically degraded assessed above by UV-VIS spectrophotometry.

3.3 Effect of SO₄²⁻ modification on Ag₂CO₃ stability

The above presented results highlighted that modification of Ag₂CO₃ by SO₄²⁻ could be an efficient route to enhance the photocatalytic activity under visible light with (i) a higher degradation rate, (ii) an almost complete decomposition of the organic pollutant and (iii) a high efficiency of mineralization of the OG dye. However, the effect of SO₄²⁻ anions on both the stability of the photocatalyst and the photocatalytic oxidation mechanism needed to be investigated for further improvement and practical applications.

To investigate the behavior of pure and sulfate-modified Ag₂CO₃ photocatalysts towards photo-corrosion issue, cycling experiments were performed. Fig. 6a reported the relative variation of OG concentration versus irradiation time under visible light recorded in presence of both the photocatalysts during three successive runs of 30 min each one. As previously mentioned, when using pure Ag₂CO₃, the dye photocatalytic degradation curve reached a plateau after 20 min of irradiation in the 1st cycle and the same behavior tended to be exhibited during the 2nd and 3rd cycles. It should be noted that as the cycles number increased, the maximal amount of OG removed decreased from approximately 60% at the end of the 1st cycle, to 45 and 35% after the 2nd and 3rd ones respectively. Accordingly, the corresponding rate constants deduced over the first 15 minutes, labelled R₁₅, also rapidly decreased from 5.1 × 10⁻² min⁻¹ for the 1st cycle to 2.6 × 10⁻² and 1.7 × 10⁻² min⁻¹ for the 2nd and 3rd ones, respectively (Fig. 6b). This steady decline of rate constants corresponded to an efficiency loss with respect to the 1st cycle of 50 and 65% for the 2nd and 3rd ones respectively. These results ascertained the poor stability of pure Ag₂CO₃ in agreement with elsewhere reported studies.^8,15 The variation of the OG half-degradation time (t_1/2) versus cycling recorded upon the use of Ag₂CO₃ (Fig. 6) showed that t_1/2 of about 14 min for the 1st cycle strongly increased to 35 min for the 2nd one while it cannot be determined after 3 cycles since the total amount of OG removed did not exceed 35% (Fig. 6c). These results further supported previous deductions (Section 3.2) regarding the low stability and/or deactivation, by being sensitive to photo-corrosion and/or poisoning issues respectively, of the unmodified Ag₂CO₃.

Fig. 6 Cycling runs for OG photodegradation under visible light irradiation over pure Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ (a). Rate constant (b) and half-photodegradation time of OG (t_1/2) versus cycling of pure Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ photocatalysts (c). X-ray diffraction patterns of unmodified Ag₂CO₃ and SO₄²⁻ modified Ag₂CO₃ after the 3rd cycle (d). The errors bars in (a)–(c) correspond to the data dispersion from 2 experiments.

By contrast, the OG concentration variation curves versus cycling, recorded in presence of SO₄²⁻-Ag₂CO₃ photocatalyst, showed a continuous and very noticeable OG photocatalytic degradation versus irradiation time under visible light regardless the cycling run number (Fig. 6a). In fact, while OG dye was completely removed after 30 min during the 1st cycle, denoting 100% of removal efficiency, the efficiency remained greater than 90% and 80% over the same period upon the 2nd and the 3rd cycles, respectively. Likewise, the corresponding R₁₅ rate constants slightly decreased from 7.8 × 10⁻² min⁻¹ after the 1st cycle to 6.3 × 10⁻² min⁻¹ after the 2nd and to 5.1 × 10⁻² min⁻¹ after the 3rd cycle, which corresponded to a thorough efficiency loss of about 35% being slighter than that (65%) previously computed for pure Ag₂CO₃ (Fig. 6b). The better stability of SO₄²⁻-Ag₂CO₃ photocatalyst was also illustrated throughout the assessment of t_1/2 values versus cycling. Fig. 6c revealed that t_1/2 remained quite of the same magnitude order around 10 min regardless the cycle run number. Furthermore, powder XRD carried out on unmodified Ag₂CO₃ sample, recovered after the 3rd cycle, revealed the appearance of a very intense peak ascribed to (111) metallic silver (Ag) (Fig. 6d). This observation denoted that bare Ag₂CO₃ underwent photo-corrosion issue. Nevertheless, it is worth noting that such Ag peak was very hardly observed in SO₄²⁻-Ag₂CO₃ after the same run of photocatalysis test. The whole of these results concorded for proving that Ag₂CO₃ modification by SO₄²⁻ likely appeared to be an effective and efficient way to improve the resistance of Ag₂CO₃ against photo-corrosion as well as poisoning issues, and thus to enhance photocatalytic activity.

3.4 Photocatalytic mechanism

In order to explore the photocatalytic mechanism of both Ag₂CO₃ and SO₄²⁻-Ag₂CO₃ photocatalysts, reactive species trapping experiments were performed to elucidate the main oxidants responsible of OG degradation under visible light irradiation, including hydroxyl radical (˙OH), superoxide radicals (˙O₂⁻), ozonide radicals (˙O₃⁻) and photoinduced holes (h⁺). The obtained results, gathered in Fig. 7(a) and (b), did not reveal significant differences in the behavior of both the investigated Ag₂CO₃ based photocatalysts towards the radicals trapping agents above-mentioned. Indeed, the addition of TCM, as ˙O₂⁻ scavenger, had no noticeable effect on OG photocatalytic degradation since the efficiency almost remained constant in both the cases like before using the trapper. This likely stood for the noninvolvement of ˙O₂⁻ in OG degradation. The use of IPA, as hydroxyl radical trapper, yielded a small decrease in the photocatalytic degradation efficiency of OG within 30 min from initially almost 62% and 100% to around 58% and 82% over unmodified and modified Ag₂CO₃ respectively. Hence, ˙OH radicals seemed to play a minor role in OG photocatalytic degradation. Nevertheless, it is noteworthy that using IC quencher of ˙O₃⁻ provoked a relatively huge decrease of the OG degradation efficiency estimated at almost 40% in both the cases. This denoted that the contribution of ˙O₃⁻ radicals in OG photocatalytic degradation may be substantial. The most salient effect was that of the addition of EDTA as h⁺ trapper which thoroughly made both the Ag₂CO₃ based photocatalysts unactive towards the OG removal. This likely ascertained that photoinduced holes (h⁺) was the major reactive species responsible of OG photocatalytic oxidation over both the investigated photocatalysts.

Fig. 7 Photocatalytic degradation of OG by Ag₂CO₃ (a) and SO₄²⁻-Ag₂CO₃ (b) under visible light irradiation in the presence of scavengers: IPA, EDTA, TCM and IC. Schematic illustration of the plausible mechanism of OG degradation in presence of SO₄²⁻-Ag₂CO₃ photocatalyst (c).

These results were further supported by the values of redox potentials of the top of valence band (VB) and the bottom of conduction band (CB) beforehand computed for both the Ag₂CO₃ based photocatalysts studied herein (Section 3.1) that were reported in Fig. 7c on which were also placed potentials of different reactive species in photocatalysis. In fact, the potential of photoelectrons on CB is too positive (0.24 eV) to be able to permit the reduction of dissolved O₂ in aqueous medium into ˙O₂⁻ species whose potential is −0.33 eV. Therefore, ˙O₂⁻ entities cannot come into play in the OG photocatalytic degradation over Ag₂CO₃ based photocatalysts, and must be excluded as also emphasized by W. Jiang et al.⁸ This deduction was in contrast to results reported by several authors^4,34 proposing the involvement of ˙O₂⁻ in the photodegradation of rhodamine B and/or Methylene Orange and/or Methylene Blue over silver carbonate although the potential of photoelectrons they estimated was 0.29 eV and 0.37 eV respectively, i.e. too positive with respect to the negative one characterizing O₂/˙O₂⁻ couple. Nevertheless, the ozone, able to be provided from dissolved oxygen upon solution illumination,⁸ present an electrode potential in neutral and acidic medium of 1.24 eV and 2.076 eV respectively so that it can be reduced by photoelectrons generated herein (0.24 eV) to yield the formation of oxidative ˙O₃⁻ radicals which assured OG photocatalytic degradation onto both the investigated photocatalysts. Furthermore, taking into account the strongly oxidative electrode potential of photo-holes (h⁺) herein estimated of about 2.80 eV, these photoinduced entities assured direct OG degradation. Likewise, h⁺ specie could also hardly oxidize H₂O to provide ˙OH radical whose the potential is close to its own (2.8 eV). This denoted that a part of h⁺ may indirectly participate through a minor amount of ˙OH radical in the OG photodegradation onto both the studied photocatalysts. In light of all these results, the OG photocatalytic degradation under visible light seemed to be achieved according to on the whole a similar mechanism illustrated in Fig. 7c onto unmodified and sulfate modified Ag₂CO₃. The plausible mechanism likely mainly implied photo-holes acting as latent oxidative entities along with to a relatively lesser extent ˙O₃⁻ reactive radicals and finally a minor participation of ˙OH hydroxyl radicals. Nonetheless, in view of the so proposed mechanism, one can ask how Ag₂CO₃ modification by SO₄²⁻ improved photostability as well as photocatalytic efficiency beforehand evidenced (Section 3.2 and 3.3)?

One beneficial key role of SO₄²⁻ anions would be their ability to contribute with 32 valence electrons to the lattice of the sulfate-modified Ag₂CO₃ photocatalyst, instead of 24 electrons for CO₃²⁻ anions, providing therefore a surplus of 8 valence electrons. As reported by Bishop et al.,³⁶ the Highest Occupied Molecular Orbital (HOMO) of SO₄²⁻ is located at around −7.5 eV, so that sulfate electrons can be transferred to lower AgCO₃ conduction band. Thus, these excess electrons provided by SO₄²⁻ together with those photogenerated in AgCO₃ upon illumination can reduce more ozone to produce more ˙O₃⁻ reactive entities which can act as additional reactive species as evidenced above by IC trapping experiments. Consequently, the OG photocatalytic degradation could speed up as proved by corresponding higher rate constant with respect to that of the pristine Ag₂CO₃ above-computed in Section 3.2.

Another possible key effect of SO₄²⁻ on the improvement of Ag₂CO₃ photostability, beforehand evidenced upon photocatalytic tests (Section 3.2) and cycling experiments (Section 3.3), may be its capacity of electrons transfer to Ag⁺ higher than that permitted by CO₃²⁻. In fact, as widely reported, primary mechanism of photocorrosion of pure Ag₂CO₃ could be attributed to the reduction of Ag⁺ cations, issued from Ag₂CO₃ solubilizing, by photoinduced e⁻, which produced metallic Ag and, over time, a progressive Ag₂CO₃ depletion into Ag cations resulting at the end in the photocatalyst degradation.^2,4,15 Considering an isolated Ag⁺–anion interaction, the fractional number of electrons transferred from the anion to Ag⁺ cation can be estimated, according to the Pearson model³⁷ (details reported in the ESI†), to be about 0.29 and 0.21 for SO₄²⁻ and CO₃²⁻ respectively. This indicated that an interaction of SO₄²⁻ with Ag⁺ cation was 38% greater than that of CO₃²⁻. Such an effect can contribute to improve the stability of the photocatalyst because Ag⁺ would be less reactive towards photoelectrons to form metal Ag, hence yielding photocorrosion issue to get mitigated.

4 Conclusion

Ag₂CO₃ modification with sulfate was achieved according to a simple precipitation method. This was confirmed by XRD that revealed a volume expansion of the monoclinic Ag₂CO₃ unit cell due to the substitution of CO₃²⁻ groups by the larger SO₄²⁻ anions. This was hence consistent with a bulk modification rather than a simple surface grafting of SO₄²⁻ ions on Ag₂CO₃ particles. However, Ag₂CO₃ modification with SO₄²⁻ did not significantly affect the Ag₂CO₃ band gap. The evaluation of the photocatalytic degradation of OG dye under visible light revealed that the sulfate modified Ag₂CO₃ was more active than bare Ag₂CO₃ leading to the achievement of a quasi-total mineralization within 30 min of irradiation. In addition, evidence for poisoning and/or photocorrosion issues observed for pure Ag₂CO₃ did not occur for the modified photocatalyst. The active species trapping experiments revealed that the OG photocatalytic degradation seemed to be assured over both the photocatalysts by photoinduced holes (h⁺) along with ˙O₃⁻ radicals, while ˙OH radicals would play a minor role. Improvements allowed by Ag₂CO₃ modification by SO₄²⁻, in terms of increase of photocatalytic rate and enhancement of Ag₂CO₃ stability, could be due to valence electrons added to photoinduced electrons, as a result of their transfer from SO₄²⁻ HOMO to Ag₂CO₃ conduction band, which may together assure the formation of much active ˙O₃⁻ radicals entities. This proposed mechanism ought to be further validated by using more sophisticated characterization technique such as ESR and PL analyses and DFT calculations. On the whole, in view of the simple synthesis route of SO₄²⁻ modified Ag₂CO₃ photocatalyst and its promising performances, in term of enhancement of photoactivity and photostability under visible light at laboratory scale, this study should be carried on by evaluating its photocatalytic properties in an outdoor solar pilot. This will increase its prospect of application in the field of tertiary treatment of wastewater.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

The support from the “Programme de soutien aux Centres d’Etudes Doctorales, Franco-Moroccan cooperation” is gratefully acknowledged. The authors further thank the Center of Analysis and Characterization of Cadi Ayyad University, Marrakesh.

References

1. Y. Wang, N. Bi, H. Zhang, W. Tian, T. Zhang, P. Wu and W. Jiang, Colloids Surf., A, 2020, 585, 124105.
2. Y. Wang, H. Liu, B. Wu, T. Zhou, J. Wang, J. Zhou, S. Li, F. Cao and G. Qin, J. Alloys Compd., 2019, 776, 948–953.
3. U. Sulaeman, F. Febiyanto, S. Yin and T. Sato, Catal. Commun., 2016, 85, 22–25.
4. H. Dong, G. Chen, J. Sun, C. Li, Y. Yu and D. Chen, Appl. Catal., B, 2013, 134–135, 46–54.
5. G. Li, Y. Wang and L. Mao, RSC Adv., 2014, 4, 53649–53661.
6. J. Li, W. Fang, C. Yu, W. Zhou, L. zhu and Y. Xie, Appl. Surf. Sci., 2015, 358, 46–56.
7. X. Jin, I. Y. Kim, Y. K. Jo, J. L. Bettis Jr, H.-J. Koo, M.-H. Whangbo and S.-J. Hwang, J. Phys. Chem. C, 2013, 117, 26509–26516.
8. W. Jiang, Y. Zeng, X. Wang, X. Yue, S. Yuan, H. Lu and B. Liang, Photochem. Photobiol., 2015, 91, 1315–1323.
9. Document search – Web of Science Core Collection, https://www.webofscience.com/wos/woscc/basic-search, (accessed November 9, 2022).
10. L. Song, Z. Chen, T. Li and S. Zhang, Mater. Chem. Phys., 2017, 186, 271–279.
11. H. El Masaoudi, I. Benabdallah, B. Jaber and M. Benaissa, Chem. Phys., 2021, 545, 111133.
12. W. Cao, Z. Gui, L. Chen, X. Zhu and Z. Qi, Appl. Catal., B, 2017, 200, 681–689.
13. J. Luo, Y. Luo, Q. Li, J. Yao, G. Duan and X. Liu, Colloids Surf., A, 2017, 535, 89–95.
14. H. Wang, J. Li, P. Huo, Y. Yan and Q. Guan, Appl. Surf. Sci., 2016, 366, 1–8.
15. G. Dai, J. Yu and G. Liu, J. Phys. Chem. C, 2012, 116, 15519–15524.
16. Y. Wang, P. Ren, C. Feng, X. Zheng, Z. Wang and D. Li, Mater. Lett., 2014, 115, 85–88.
17. Z. Xiang, J. Zhong, S. Huang, J. Li, J. Chen, T. Wang, M. Li and P. Wang, Mater. Sci. Semicond. Process., 2016, 52, 62–67.
18. N. Yu, R. Dong, J. Liu, K. Huang and B. Geng, RSC Adv., 2016, 6, 103938–103943.
19. W.-K. Jo, S. Kumar, P. Yadav and S. Tonda, Appl. Surf. Sci., 2018, 445, 555–562.
20. W. Fa, P. Wang, B. Yue, F. Yang, D. Li and Z. Zheng, Chin. J. Catal., 2015, 36, 2186–2193.
21. J. Dostanić, D. Lončarević, V. Đorđević, S. P. Ahrenkiel and J. M. Nedeljković, J. Photochem. Photobiol., A, 2017, 336, 1–7.
22. X. Yang, R. Li, Y. Wang, K. Wu, S. Chang and H. Tang, Ceram. Int., 2016, 42, 13411–13420.
23. C. Dong, K.-L. Wu, X.-W. Wei, X.-Z. Li, L. Liu, T.-H. Ding, J. Wang and Y. Ye, CrystEngComm, 2013, 16, 730–736.
24. N. Tian, H. Huang, Y. He, Y. Guo and Y. Zhang, Colloids Surf., A, 2015, 467, 188–194.
25. C. Feng, L. Tang, Y. Deng, J. Wang, J. Luo, Y. Liu, X. Ouyang, H. Yang, J. Yu and J. Wang, Adv. Funct. Mater., 2020, 30, 2001922.
26. C. Feng, Z.-P. Wu, K.-W. Huang, J. Ye and H. Zhang, Adv. Mater., 2022, 34, 2200180.
27. O. Lakbita, B. Rhouta, F. Maury, F. Senocq, M. Amjoud and L. Daoudi, Appl. Surf. Sci., 2019, 464, 205–211.
28. M. C. Simoes, K. J. Hughes, D. B. Ingham, L. Ma and M. Pourkashanian, Inorg. Chem., 2017, 56, 7566–7573.
29. A. Bouziani, M. Yahya, Y. Naciri, A. Hsini, M. A. Khan, M. Sillanpää and G. Celik, Surf. Interfaces, 2022, 34, 102328.
30. K. Perumal, S. Shanavas, T. Ahamad, A. Karthigeyan and P. Murugakoothan, J. Environ. Sci., 2023, 125, 47–60.
31. X. Bu, C. Chen, X. Zhao, Q. Huang, X. Liao, H. Fan, P. Wang, H. Hu, Y. Zhang and Z. Huang, Appl. Surf. Sci., 2022, 588, 152887.
32. T. J. B. Holland and S. A. T. Redfern, Mineral. Mag., 1997, 61, 65–77.
33. H. Zeng, Z. Yu, L. Shao, X. Li, M. Zhu, Y. Liu, X. Feng and X. Zhu, Desalination, 2020, 491, 114558.
34. E. Tomaszewicz, M. Kurzawa and L. Wachowski, J. Mater. Sci. Lett., 2002, 21, 547–549.
35. M. Rehan, A. Barhoum, T. A. Khattab, L. Gätjen and R. Wilken, Cellulose, 2019, 26, 5437–5453.
36. D. M. Bishop, M. Randič and J. R. Morton, J. Chem. Phys., 2004, 45, 1880–1885.
37. R. G. Pearson, Inorg. Chem., 1988, 27, 734–740.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ra03120a

---