Dual phototransformation of the pollutants methyl orange and Cr (VI) using phthalocyanine-cobalt ferrite based magnetic nanocomposites

Abstract

Bifunctional nanocomposites based on zinc phthalocyanines and glutathione capped CoFe₂O₄ magnetic nanoparticles (GSH-CoFe₂O₄ MNPs) are applied in a binary system wherein simultaneous photooxidation of methyl orange (MO) and photoreduction of Cr (VI) are conducted. The photoactivity of two zinc Pcs with different functional moieties are compared based on their interactions with GSH-CoFe₂O₄ MNPs. Conjugation of the Pcs to the GSH-CoFe₂O₄ MNPs not only enhanced their singlet oxygen production but also their photocatalytic activity in both photooxidation and photoreduction experiments. Using electron paramagnetic resonance (EPR) spectroscopy, the Pc-MNP conjugates reported herein were found to exhibit superparamagnetic behaviour, giving the advantage of easy separation using an external magnetic field post application, an attractive attribute for heterogeneous catalysis. The catalysts reported herein are therefore good candidates as catalysts for real life water purification analyses as they facilitate the treatment of both organic and inorganic water pollutants.

1. Introduction

Phthalocyanines (Pcs) have been extensively studied in applications such as electrochemical sensors, liquid crystals, dye-sensitized solar cells, nonlinear optics and photodynamic therapy [1, 2, 3, 4, 5, 6]. Due to their excellent visible/near infrared absorption, high chemical and thermal stability and the ability of generate singlet oxygen, Pcs have also been reported as good photocatalysts for the degradation of organic dyes [7, 8]. Photocatalysis has been proposed as an effective advanced oxidation process (AOP) for the degradation of organic pollutants. Photocatalysis may completely degrade pollutants without giving rise to undesired by-products, which may be even more toxic than the parent compound. However, photocatalysis using Pcs has mainly been used for organic pollutants and this work presents the use of Pcs for photocatalysis of an inorganic pollutant, Cr(VI), in the presence of ferrites.

Ferrites are also commonly used photocatalysts, they have band gaps of approximately 2 eV which in turn enable them to absorb visible light [9, 10]. They have been used in various applications including photocatalysis, as multiferroic materials and magnetic labelling of biological systems [10, 11, 12, 13].

In this work, we link the two photocatalysts (Pcs and ferrites) for improved photocatalytic activity by a synergistic effect. Conjugates of gluthathione capped cobalt ferrite magnetic nanoparticles (GSH-CoFe₂O₄ MNPs) with 2(3),9(10),16(17),23(24)-tetrakis-(4’-(4′-6′-diaminopyrimidin-2′-ylthio))) phthalocyaninato zinc (II) (complex 1) and zinc tetraaminophenoxy phthalocyanine (complex 2) are synthesised. They are then employed for the photodegradation of methyl orange (MO) and the photoreduction of hexavalent chromium, common water pollutants.

Complexes 1 and 2 are conjugated to mixed metal MNPs which show more catalytic activity compared to the corresponding single component metal oxides [14]. Glutathione was used for functionalising the CoFe₂O₄ MNPs because in addition to it having the appropriate terminal groups for amide bond formation with the Pc complexes, glutathione has itself been reported to possess photoreduction ability for Cr (VI) [15, 16]. Apart from our recent report [17], there are no reports on linking Pcs to bimetallic MNPs. The conjugates reported herein are employed as dual photocatalysts for both photoreduction of Cr (VI) and photodegradation of MO. This is the first time that Pc-MNP conjugates are employed as dual photocatalysts for both organic and inorganic species using the same photocatalyst.

The reduction of Cr (VI) to Cr (III) is often accelerated in the presence of sacrificial donors such as 4-chlorophenol which prevent the re-oxidation of Cr (III) to Cr (VI). Cr (III) is significantly less toxic than Cr (VI) and is readily removable by alkaline precipitation [18]. In this work, in addition to studying the photodegradation of MO, we use MO as a sacrificial electron donor during the photoreduction of Cr (VI). This is because the hydroxyl radicals ($·$OH) generated during the photolysis may oxidize Cr (III) back to Cr (VI) [19]. In the presence of MO however, the Cr (III) to Cr (VI) oxidation is suppressed since MO acts as a sacrificial molecule which will react preferentially with the $·$OH.

Complex 1 has eight amine groups while complex 2 has four and therefore a comparative study on the effect of the number of amine groups on the photocatalytic behaviour will be examined. The presence of the CoFe₂O₄ MNPs in the photocatalysts gives the advantage of magnetic regeneration of the catalyst and hence reusability. In addition, the MNPs also encourage intersystem crossing (ISC) to the triplet state of the Pc, through the heavy atom effect of the metals contained in the MNPs. This enhances the singlet oxygen generating ability of the Pc [7, 8], which is fundamental for their photocatalytic efficiency.

In this work nanocomposites of Pcs and bimetallic CoFe₂O₄ MNPs will for the first time be used to facilitate both photoreduction and photodegradation, using Cr (VI) and MO as models, respectively. MO is a synthetic azo (-N=N-) dye that is commonly used in textiles and paper industries [20]. Cr (VI) on the other hand is also used in textiles, tannery industries as well as fertilizers, amongst others [21]. Both of Cr (VI) and MO are toxic pollutants, hence the need of finding means of degrading or transforming them [22, 23].

2. Experimental

2.1. Materials

Dimethylsulphoxide (DMSO), dimethylformamide (DMF), Zn phthalocyanine (ZnPc), Na₂HPO₄ and KH₂PO₄ were obtained from SAARChem. N,N-dicyclohexylcarbodiimide (DCC), 1,3-diphenylisobenzofuran (DPBF), methyl orange (MO), chromium (VI) oxide, reduced glutathione (GSH) and iron (III) chloride hexahydrate were purchased from Sigma–Aldrich. Cobalt (II) chloride hexahydrate was purchased from UniLab. All other reagents and solvents were obtained from commercial suppliers and used as received. Aqueous solutions were prepared using millipore water from Milli-Q Water Systems and adjusted to a pH of 2.75 with phosphoric acid from Minema Chemicals. 2(3),9(10), 16(17),23(24)-Tetrakis-(4’-(4′-6′-diaminopyrimidin-2′-ylthio)))phthalocyaninato zinc (II) (1) [24] and zinc tetraaminophenoxy phthalocyanine (2) [25,26] were synthesized according to literature methods.

2.2. Methods

2.2.1. Synthesis of magnetic nanoparticles (MNPs)

2.2.1.1. Synthesis of bare CoFe₂O₄ MNPs (Fig. 1A)

Fig. 1.

(A) Synthesis of bare CoFe₂O₄ MNPs and (B) functionalization of CoFe₂O₄ MNPs with GSH.

The MNPs were prepared by the conventional co-precipitation method as reported before [27] with slight modifications as follows: iron (III) chloride hexahydrate (25 mL, 0.4 M) and cobalt (II) chloride hexahydrate (25 ml, 0.2 M) were mixed together. Ammonium hydroxide (25%) was added dropwise to the solution until a pH of 10–12 was reached. The solution was then heated to 85 °C for 1 h and then cooled to room temperature. The precipitated nanoparticles were then retrieved by magnetic decantation, washed with water three times and then dried overnight under vacuum to form the bare CoFe₂O₄ MNPs.

2.2.1.2. Functionalisation of the bare CoFe₂O₄ MNPs to form GSH-CoFe₂O₄ (Fig. 1B)

The bare CoFe₂O₄ MNPs were modified by anchoring of glutathione (GSH) onto the surface of the ferrite as reported before [28] with slight modifications as follows: bare CoFe₂O₄ MNPs (0.5 g) were dispersed in a solvent mixture of water (15 mL) and methanol (5 mL) and sonicated for 15 min. Reduced glutathione was then dissolved in water (15 mL) and added to the solvent mixture drop-wise followed by sonication for a further 2 h. The resultant GSH-CoFe₂O₄ MNPs were then magnetically separated washed with methanol and dried under vacuum overnight.

2.2.2. Conjugation of complexes 1 and 2 to GSH-CoFe₂O₄ forming CoFe₂O₄-1 and CoFe₂O₄-2 (Figs. 2 and 3)

Fig. 2.

Conjugation of GSH-CoFe₂O₄ MNPs to complex 1.

Fig. 3.

Conjugation of GSH-CoFe₂O₄ MNPs to complex 2.

Conjugation of the GSH-CoFe₂O₄ MNPs with complexes 1 (Fig. 2) and 2 (Fig. 3) was done as reported in literature [29] as follows: in two separate round bottomed flasks, a mixture of GSH-CoFe₂O₄ MNPs (0.1 g) and DCC (0.02 g, 0.097 mmol) in DMF (4 mL) was added, followed by stirring at room temperature for 48 h. Complexes 1 and 2 (0.10 g, ∼ 0.088 mmol) were then added to the mixtures and stirring was continued for another 48 h. The conjugates were collected and purified with acetone, methanol, and ethanol in succession. The resulting products (CoFe₂O₄-1 and CoFe₂O₄-2) were dried under vacuum overnight.

2.2.3. Photophysical parameters

Fluorescence (Φ_F), triplet (Φ_T), and singlet oxygen (Φ_Δ) quantum yields were determined in DMSO using comparative methods described before in literatures [30, 31, 32, 33]. Unsubstituted ZnPc in DMSO was used as a standard with Φ_F = 0.20 [31], Φ_Δ = 0.67 [33] (using DPBF as a singlet oxygen quencher), and Φ_T = 0.65 [32]. The solutions for triplet state studies were de-aerated with argon for 15 min before measurements.

2.3. Instrumentation

UV–Vis absorption spectra were measured at room temperature on a Shimadzu UV-2550 spectrophotometer using a 1 cm pathlength cuvette in solution. Fluorescence emission and excitation spectra were obtained on a Varian Eclipse spectrofluorimeter using a 1 cm pathlength quartz cuvette. IR spectra were recorded on a Perkin-Elmer Spectrum 100 ATR FT-IR spectrometer.

Fluorescence decay times were measured using a time correlated single photon counting (TCSPC) setup (FluoTime 300, Picoquant GmbH). The excitation source was a diode laser (LDH-P-670 driven by PDL 800-B, 670 nm, 20 MHz repetition rate, 44 ps pulse width, Pico quant GmbH). Triplet quantum yields were determined using laser flash photolysis system. The excitation pulses were produced using a tunable laser system consisting of a Nd:YAG laser (355 nm, 135 mJ/4–6 ns) pumping an optical parametric oscillator (OPO, 30 mJ/3–5 ns) with a wavelength range of 420–2300 nm (NT-342B, Ekspla) as described before [34].

Energy dispersive X-ray spectroscopy (EDX) was done on an INCA PENTA FET coupled to the VAGA TESCAM using 20 kV accelerating voltage. Dynamic light scattering (DLS) experiments were done on a Malvern Zetasizer Nanoseries, Nano-ZS90. X-ray powder diffraction (XRD) patterns were recorded on a Bruker D8 Discover equipped with a Lynx Eye Detector, using Cu-Kα radiation (λ = 1.5405 A, nickel filter). Details have been reported before [8].

X-ray photoelectron spectroscopy (XPS) analysis was done using an AXIS Ultra DLD, with Al (monochromatic) anode equipped with a charge neutraliser, supplied by Kratos Analytical. The following parameters were used: the emission was 10 mA, the anode (HT) was 15 kV and the operating pressure below 5 × 10⁻⁹ torr. A hybrid lens was used and resolution to acquire scans was at 160 eV pass energy in slot mode. The centre used for the scans was at 520 eV with a width of 1205 eV, steps at 1 eV and dwell time at 100 ms. The high resolution scans were acquired using 80 eV pass energy in slot mode.

Electron paramagnetic resonance (EPR) measurements were carried out using a Bruker EMX Plus EPR spectrometer, model number: EMP-9.5/12B/P. EPR settings were: the microwave power was 0.632 mW, frequency 9.714 GHz, resolution 2048 points, and centre field 3460 G and 100 G for the sweep width. The position of resonance of a paramagnet is defined by the g-factor and it was calculated using Eq. (1):

$$\text{g}=\frac{h\text{υ}}{H\beta }$$ (1)

where h is the Planck's constant, υ is the frequency of the microwave, H is the magnetic field occurring, and β is the Bohr magneton [35].

Measurement of singlet oxygen quantum yields as well as photocatalytic reactions for MO and Cr (VI) were carried out using irradiation from a halogen lamp (300 W), 600 nm glass (Schott) and water were used to filter off ultra-violet and far infrared radiation, respectively, as described before [36]. Light intensities were measured with a POWER MAX 5100 (Molelectron detector incorporated) power meter and were found to be 3.8 × 10¹⁶ photons/cm² s for singlet oxygen studies.

3. Results and discussion

3.1. Synthesis and characterization of the conjugates

GSH-CoFe₂O₄ MNPs were synthesized as reported before [27, 28] with modifications (Fig. 1). Various methods have been reported on the synthesis of cobalt ferrite MNPs such as the sonochemical method, mechanochemical method, co-precipitation and the micro-emulsion procedure [37, 38, 39, 40, 41]. In this work the co-precipitation method is implemented due the fact that, relative to other known techniques, small particle sizes can be obtained at low temperatures [40].

Following activation of the carboxylic acid groups of the GSH-CoFe₂O₄ MNPs with DCC, they were linked via amide bond formation with complexes 1 and 2 forming CoFe₂O₄-1 and CoFe₂O₄-2 respectively (Figs. 2 and 3). Since Pcs have sizes of about ∼1 nm and the GSH-CoFe₂O₄ MNP have an average size of 10.08 nm (from DLS measurements below), it is unlikely for more than one MNP to be attached to the Pc, but it is possible for more than one Pc to link to the MNP. The number of Pc molecules bonded to the GSH-CoFe₂O₄ MNPs were determined following literature methods, but using absorption instead of fluorescence [42]. This involves comparing the Q band absorbance intensity of the Pc in the conjugate with that of the initial Pc before the conjugation. The ratio of the Pcs:MNPs in CoFe₂O₄-1 and CoFe₂O₄-2 were found to be 5:1 and 8:1 respectively, Table 1, showing that more molecules of complex 2 are loaded onto the MNPs than there are of complex 1.

Table 1.

Photophysicochemical data of complexes 1 and 2 and their respective conjugates (CoFe₂O₄-1 and CoFe₂O₄-2) in DMSO.

Complexes	Ratio Pc/MNPs	DLS size (nm)a	λabs (nm)	ΦF	τF (ns)	ΦT	τT (μs)	ΦΔ
1	-	-	689b	0.11b	2.53b	0.83b	351b	0.30b
CoFe2O4-1	5:1	14.10	689	0.058	2.29	0.88	355	0.65
2	-	-	686c	0.060c	2.79c	0.51c	315c	0.38c
CoFe2O4-2	8:1	16.95	686	<0.010	2.20	0.85	256	0.62

^aDLS size for GSH- CoFe₂O₄ MNPs alone = 10.08 nm.

^bvalues from reference 24.

^cvalues from reference 26.

The UV–vis spectra of complex 1 and CoFe₂O₄-1 are shown in Fig. 4A and those of complex 2 and CoFe₂O₄-2 are shown in Fig. 4B. The narrow Q bands of complexes 1 and 2 confirm their monomeric behaviour. No shifts were observed in the Q bands of both Pcs after conjugation (Table 1).

Fig. 4.

Ground state absorption spectra of (A) (i) complex 1, (ii) CoFe₂O₄-1, and (B) (i) complex 2, (ii) CoFe₂O₄-2 in DMSO.

Transmission electron microscopy (TEM) was employed so as to obtain the morphology and dispersion of the MNPs and their respective conjugates. The images (Fig. 5) reveal that the GSH-CoFe₂O₄ MNPs and CoFe₂O₄-1 have a predominantly spherical morphology. Due to the superparamagnetic nature of the MNPs however, the TEM images also show that the GSH-CoFe₂O₄ MNPs are aggregated and even more so when conjugated to the Pcs (CoFe₂O₄-1 used as an example). The aggregates appear as big clusters of predominantly spheres in the images with some rods also observed. The aggregation is possibly due to the π-π interaction between Pcs on neighbouring MNPs since Pcs are known for their π-π stacking to form aggregates [43]. The size distribution histograms show the estimates of the sizes of the spheres representing the GSH-CoFe₂O₄ MNPs, CoFe₂O₄-1 and CoFe₂O₄-2 to be 10.75 nm, 14.85 nm and 16.10 nm, respectively.

Fig. 5.

TEM images of the (A) GSH-CoFe₂O₄ MNPs and (B) CoFe₂O₄-1 and their respective histograms.

Dynamic light scattering (DLS), Fig. 6, showed an increase in size of the GSH-CoFe₂O₄ MNPs following conjugation to complexes 1 and 2, Table 1. The sizes obtained for GSH-CoFe₂O₄ MNPs, CoFe₂O₄-1 and CoFe₂O₄-2 are 10.08, 14.10 and 16.95 nm respectively, similar to those obtained from TEM. It is expected that sizes from DLS would be slightly higher than those from TEM due to the interference of the dispersant into the hydrodynamic diameter, however, DLS-numbers are close to the TEM results. As observed before [44] DLS number distribution can be close to those obtained from TEM while DLS intensity distribution usually has a large difference with TEM. It was proven that the DLS-numbers show a good approximation to the size parameters obtained by TEM while DSL-intensity does not. These errors are more evident on the results of mixed samples with different sizes, where the DLS-intensity present values may be up to five times larger than TEM.

Fig. 6.

DLS plots of (A) GSH-CoFe₂O₄, MNPs, (B) CoFe₂O₄,-1, and (C) CoFe₂O₄-2.

X-ray Diffraction (XRD) patterns of the compounds were analysed as shown in Fig. 7. The GSH-CoFe₂O₄ MNPs show characteristic peaks attributed to a cubic spinel structure with peaks at 2θ = 29^o, 36^o, 43^o, 54^o, 57^o and 63^o corresponding to hkl Miller indices of (220), (311), (400), (422), (511) and (440), respectively [45]. Complex 1 has a broadened peak at ∼2θ = 26^o which is typical for Pcs [46]. The CoFe₂O₄-1 on the other hand has peaks corresponding to both the GSH-CoFe₂O₄ MNPs and complex 1. This is an indication that the conjugate is indeed a composite of the two photocatalysts and that the GSH-CoFe₂O₄-NH₂ MNPs retain their crystallinity when conjugated to Pcs, which also retain their amorphous nature in the conjugate. Similar diffraction patterns were obtained for complex 2 and CoFe₂O₄-2 (figure not shown).

Fig. 7.

X-ray diffraction patterns of (A) GSH-CoFe₂O₄ MNPs, (B) CoFe₂O₄-1, and (C) complex 1.

Energy Dispersive X-ray Spectroscopy (EDX) was conducted for confirmation of the elemental composition of the MNPs, Pcs and conjugates (using the bare and functionalised CoFe₂O₄ MNPs, complex 2 and CoFe₂O₄-2 as examples). As shown in Fig. 8, the GSH-CoFe₂O₄ have S peak which is absent in the bare CoFe₂O₄ MNPs, an indication of efficient incorporation of glutathione on the surface of the bare MNPs. Complex 2 has the expected C, N, O and Zn peaks of a Pc and its Pc-MNP conjugate (CoFe₂O₄-2) on the other hand has peaks from the GSH-CoFe₂O₄ MNPs and an additional Zn peak from complex 2, that being an indication of pure elemental composition of the Pcs, MNPs and conjugates.

Fig. 8.

EDX spectra of (A) Bare CoFe₂O₄ MNPs, (B) GSH-CoFe₂O₄ MNPs, (C) complex 2, and (D) CoFe₂O₄-2.

Fourier-transform infrared (FT-IR) spectroscopy was used to elucidate the structural properties of the MNPs. It was also used to confirm functionalization of the MNPs with glutathione as well as efficient covalent bond formation between the MNPs and Pc complexes, Fig. 9 (using complex 1 as an example). IR bands at 3033 cm⁻¹ and 2540 cm⁻¹ representing carboxylic acid O-H stretch and the S–H stretch, respectively, were observed for glutathione (Fig. 9 A(i)). The S-H stretch diminished in the GSH-CoFe₂O₄ MNPs (Fig. 9 A(ii)) due to covalent interaction of the glutathione to the bare MNP through the thiol group. There was also an observed IR band at 1545 cm⁻¹ for glutathione (Fig. 9 A(i)), representing the amide moieties in its chain as well as bands between $\sim$ 1000 cm⁻¹ and 500 cm⁻¹, representing its C-H, C-N and C-C bonds. The bands at 444 cm⁻¹ and 592 cm⁻¹, observed for GSH-CoFe₂O₄ MNPs (Fig. 9 A(ii)) are characteristic for metal-oxygen bonds for spinel ferrite structure [47]. They represent the octahedral metal-oxygen and tetrahedral metal-oxygen stretching vibrations, respectively [48]. The amide bond and carboxylic acid O-H stretch are also maintained with bands at 1595 cm⁻¹ and 3012 cm⁻¹, respectively. The FT-IR spectrum of complex 1 (Fig. 9 B(i)) showed peaks corresponding to the N-H stretch were observed at 3185 and cm⁻¹ and 3320 cm⁻¹ and an N-H bend peak was observed at 1548 cm⁻¹, due to the primary amine. The carboxylic acid O-H stretch of the GSH-CoFe₂O₄ MNPs and the primary amine stretch of complex 1 are absent in their conjugate (CoFe₂O₄-1) instead there is an observed C-H stretch at 2951 cm⁻¹. There is also an amide peak at 1565 cm⁻¹ in addition to the characteristic peaks of the spinel MNPs, this is indication of successful amide bond formation between the MNPs and Pc complexes (complex 1 in this case).

Fig. 9.

FT-IR spectra of (A) (i) glutathione only, (ii) GSH-CoFe₂O₄ MNPs, (B) (i) complex 1 and (ii) CoFe₂O₄-1.

The EPR spectroscopy is highly sensitive to the local electronic structure, oxidation state and the vicinity of the magnetic species. EPR was used in this work to investigate the magnetic properties at high frequency due to the resonance originating from the interaction between spins and electromagnetic waves. At room temperature, the first derivative EPR spectrum (Fig. 10) of the GSH-CoFe₂O₄ MNPs shows a mixing of two lines consisting of a broad component corresponding to typical anisotropic contributions and a narrow one, which has been attributed to surface isotropic contributions [35]. This is typical for superparamagnetic resonance [35]. The position of resonance i.e. effective g-value was calculated to be 2.005 for the GSH-CoFe₂O₄ MNPs, a value that is typical for paramagnetic compounds [49]. Upon conjugation of the GSH-CoFe₂O₄ MNPs with the Pcs (using CoFe₂O₄-1 as an example), there was a slight reduction in the intensity of the spectrum, suggesting that the magnetic behaviour of the MNPs slightly decreased after conjugation. There is also an observed narrow, high intensity signal with an effective g-value of 2.002 suggesting the presence of ZnPc^•+ in the conjugate [50]. The obtained results are therefore an indication that the MNPs and Pc-MNP conjugates reported herein are indeed superparamagnetic and were magnetically retrieved for possible re-use.

Fig. 10.

EPR spectra of (A) GSH-CoFe₂O₄ MNPs and (B) CoFe₂O₄-1.

X-ray photoelectron spectroscopy (XPS) survey scans (Fig. 11A and C) shows the expected elements for both the GSH-CoFe₂O₄ MNPs and the Pc complexes while the conjugates (CoFe₂O₄-1 and CoFe₂O₄-2) have elements from their respective Pcs and GSH-CoFe₂O₄ MNPs. Although complex 1 does not have oxygen but there is an observed O1s peak. It has been reported that samples exposed to the atmosphere, whether due to adventitious contamination, oxidation or exposure to water will usually have an oxygen peak [51].

Fig. 11.

(A) Wide scan XPS for (i) GSH-CoFe₂O₄ MNPs, (ii) CoFe₂O₄-1 and (iii) complex 1, (B) High resolution XPS (N Is) spectra for (i) GSH-CoFe₂O₄ MNP and (ii) CoFe₂O₄-1, (C) Wide scan XPS for (i) GSH-CoFe₂O₄ MNPs, (ii) CoFe₂O₄-2 and (iii) complex 2, and (D) High resolution XPS (N Is) spectra for CoFe₂O₄-2.

The N 1s high resolution peak for the GSH-CoFe₂O₄ MNPs was deconvoluted to yield three components at 398.3 eV, 399.5 eV, and 401.7 eV corresponding to N-C, N-H and N-C = 0, respectively (Fig. 11B(i)). The component at high binding energy (401.7 eV) is representative of the amide bond present in the glutathione backbone. The high resolution N 1s XPS spectrum for complex 1 was deconvoluted to show two peaks [24]. The high resolution N 1s for CoFe₂O₄-1 on the other hand was deconvoluted to four components at 398.7 eV (N=C), 399.4 eV (N-C), 400.3(N-H) and 401.2 eV (N-C=O), Fig. 11B (ii). The high resolution N 1s for complex 2 was deconvoluted to show four components [17]. The high resolution N 1s for CoFe₂O₄-2 on the other hand was deconvoluted and showed five components at 397.8eV (N=C), 398.6 eV (N), 399.8 eV (N-C), 400.4 eV (N-H) and 401.2 eV (N-C=O), Fig. 11D. The intensities of the higher binding energy components (401.2 eV) are higher at 11730 cps and 14620 cps for CoFe₂O₄-1 and CoFe₂O₄-2 respectively compared to 2630 cps obtained for the GSH-CoFe₂O₄ MNPs. These results suggest that there is an increase in the number of amide bonds upon conjugation of the GSH-CoFe₂O₄ MNPs with the respective Pcs. The results also show that although complex 1 which has 8 terminal amine groups, forms less covalent bonds (less intensity of the amide bond peak) than complex 2. This could be attributed to steric hinderance due to the multiple sites of attachment and hence various ways that the complex 1 and GSH-CoFe₂O₄ MNPs can bond.

3.2. Photophysical and photochemical properties

3.2.1. Fluorescence quantum yields (Φ_F) and lifetimes (τ_F)

When Pcs are conjugated to MNPs, the latter introduces a heavy atom effect which enhances intersystem crossing (ISC) of the Pc to the triplet state and hence decreases fluorescence. This therefore means that lower Φ_F are expected for Pc-MNP conjugates compared to their respective Pcs. As seen in Table 1, complex 1 has a Φ_F of 0.11 [24] while CoFe₂O₄-1 has a Φ_F of 0.058. Complex 2 on the other hand has a Φ_F of 0.060 [26] while CoFe₂O₄-2 has a Φ_F of <0.01. These results therefore show that most of the energy of the excited photosensitizer is transferred through ISC and very little is given off through fluorescence.

A fluorescence decay curve using CoFe₂O₄-1 as an example is shown in Fig. 12A, where the curves show biexponential decays. This type of decay may occur for Pcs due to the formation of aggregates which are non-fluorescent, but which can quench the monomer [52]. For the Pc-MNP conjugates, the bi-exponential fluorescence lifetimes could be due to the orientation of the Pcs around the NPs [53]. The reported τ_F for complexes 1 and 2 are 2.53 and 2.79 ns, respectively [24, 26], while those for CoFe₂O₄-1 and CoFe₂O₄-2 are 2.29 and 2.20 ns, respectively. The shortening in τ_F corresponds to the decrease in Φ_F of the Pcs when in the presence of MNPs.

Fig. 12.

(A) Fluorescence lifetime decay and (B) triplet decay curves for CoFe₂O₄-1 in DMSO.

3.2.2. Triplet quantum yields (Φ_T) and lifetimes (τ_F)

A triplet decay curve of a Pc-MNP conjugate is shown in Fig. 12B (using CoFe₂O₄-1 as an example). The Φ_T of complex 1 is 0.83 [24] while that of CoFe₂O₄-1 was found to be 0.88, Table 1. Complex 2 has been reported to have a Φ_T of 0.51 [26] while that of CoFe₂O₄-2 was found to be 0.85. The results show that when Pcs are in the presence of MNPs, there is an increase in Φ_T due to the heavy atom effect of the CoFe₂O₄ MNPs. The greater enhancement in the Φ_T for CoFe₂O₄-2 than CoFe₂O₄-1 could be due to the larger loading of Pc onto MNPs in the former. The reported τ_T for complexes 1 and 2 are 351 and 315 μs [24, 26] while CoFe₂O₄-1 and CoFe₂O₄-2 were found to have τ_T of 355 and 256 μs, respectively. It has been reported that when Φ_T increases, τ_T will decrease [54] as observed for the CoFe₂O₄-2. The lengthening of τ_T for the Pc in CoFe₂O₄-1 could be due to it being shielded by the MNPs as observed before [17].

3.2.3. Singlet oxygen quantum yield (Φ_Δ)

In this work, the Φ_Δ values of the photocatalysts were determined using DPBF as a singlet oxygen quencher in DMSO. The decay of DPBF upon irradiation with increased time is shown in Fig. 13 wherein CoFe₂O₄-1 was used as an example. No significant changes in the intensity of the Q band were observed for the period of irradiation, that being an indication of the photostability of the Pcs and their respective conjugates. DPBF did degrade though and was monitored at 417 nm. The reported Φ_Δ for complex 1 is 0.30 [24] while CoFe₂O₄-1 was found to have a Φ_Δ of 0.65. Complex 2 on the other hand has a reported Φ_Δ of 0.38 [26] while CoFe₂O₄-2 has a Φ_Δ of 0.62, Table 1. When conjugated to single metal Fe₃O₄ MNPs, the Φ_Δ of complex 1 was found to be 0.41 [24] compared to the 0.65 found here when GSH-CoFe₂O₄ MNPs are used. Thus, the conjugation of Pcs to mixed metal MNPs (CoFe₂O₄ MNPs in this case) enhances Φ_Δ compared to when single metal MNPs are used. This is not surprising as mixed metal MNPs have been reported to show more catalytic activity compared to their corresponding single component metal oxides [14].

Fig. 13.

Photodegradation of DPBF in the presence of CoFe₂O₄-1 in DMSO with an initial concentration of DPBF = 5.9$\times$10⁻⁵ molL⁻¹ at 10 s irradiation intervals.

3.3. Photocatalysis

3.3.1. Photodegradation of MO

Photooxidation was conducted at room temperature with pH of 2.75 because it has been reported that the photodegradation of MO is enhanced at pH values in the range 2–3 [55,56]. The photocatalysts (Pcs or their conjugates) are not soluble in water, hence they were suspended in MO solution as solids. The peak at 506 nm corresponding to the azo bond decreased in intensity with increased irradiation in the presence of CoFe₂O₄-1 (used as example) while the bands below 330 nm increased in intensity, Fig. 14. The latter bands are due to the benzene rings [57, 58]. The obtained spectra therefore suggest that the benzene rings remain intact and are not degraded during the photooxidation process as observed before [57]. Since the Pc complexes used for the study are not water soluble, there is no observed Pc Q band in the degradation spectra. The color of MO was observed to also fade with irradiation time, this is because the azo bond determines the colour of the azo dyes and is very reactive usually undergoing oxidation, leading to fading of their colour [59].

Fig. 14.

Absorption spectral changes of 2.67 × 10⁻⁵ molL⁻¹ MO solution during visible light photocatalysis using CoFe₂O₄-1 in water with irradiation intervals of 2 min at pH 2.75. Insert = the structure of MO.

The photocatalysts were applied in the degradation of five concentrations of MO with a catalyst loading of 1.0 mg/mL. The initial rates and rate constants (k_obs) for the photodegradation of MO increased (while the half-life decreased) with an increase in the MO concentration (Table 2). Plots for ln (C_O/C) versus irradiation time are linear (Fig. 15) confirming pseudo first order kinetics. Both conjugates (CoFe₂O₄-1 and CoFe₂O₄-2) showed larger rates of degradation than their respective Pcs (complexes 1 and 2). GSH-CoFe₂O₄ MNPs on the other hand were found to have the poorest photocatalytic activity. This is because, even though ferrites have been reported to have photocatalytic activity, the addition of oxidants such as H₂O₂ is often required to enhance the production of reactive oxygen species (ROS) and hence photocatalytic ability. This is because the e⁻/h⁺ pairs (formed on excitation) tend to recombine quickly, thereby reducing the photocatalytic ability of the ferrites [10]. When the photocatalysts were employed in the absence of irradiation and absence of oxygen (i.e. nitrogen purged solutions), only slight UV–vis spectral changes were observed, an indication that light and molecular oxygen are prerequisites in the degradation process.

Table 2.

Rates, rate constants (k_obs) and half-lives (t_1/2) of various initial concentrations of MO using GSH-CoFe₂O₄, complexes 1 and 2 and their respective conjugates (CoFe₂O₄-1 and CoFe₂O₄-2).

[MO] × 10−5 (mol L−1)	kobs (min−1)					Rate (10−7 mol L-1 min−1)					t1/2 (min)
	GSH-CoFe2O4	1	CoFe2O4-1	2	CoFe2O4-2	GSH-CoFe2O4	1	CoFe2O4-1	2	CoFe2O4-2	GSH-CoFe2O4	1	CoFe2O4-1	2	CoFe2O4-2
3.40	0.118	0.132	0.221	0.136	0.221	40.1	44.9	75.1	46.2	75.1	5.87	3.14	3.13	5.10	3.13
2.67	0.0520	0.0606	0.106	0.0610	0.105	13.9	16.2	28.3	16.3	28.0	13.3	11.4	6.54	11.4	6.60
2.13	0.0280	0.0360	0.0629	0.0380	0.0625	5.96	7.66	13.4	8.09	13.3	24.8	19.3	11.0	18.2	11.1
1.85	0.0070	0.0100	0.0254	0.0103	0.0250	1.29	1.85	4.69	1.90	4.62	99.0	69.3	27.3	67.3	27.7
1.36	0.0010	0.0042	0.0093	0.0042	0.0094	0.130	0.570	1.26	0.570	1.27	693	165	74.5	165	73.7

Fig. 15.

Kinetics plots for degradation of MO; i. 1.85×10⁻⁵, ii. 2.13×10⁻⁵, iii. 2.67×10⁻⁵ and iv. 3.40×10⁻⁵ mol L⁻¹ using CoFe₂O₄-1 as a photocatalyst in water at pH 2.75.

3.3.2. Photoreduction of Cr (VI) to Cr (III)

UV-vis spectra for the photoreduction of Cr (VI) (Fig. 16A, using CoFe₂O₄-1 as an example) show a decrease in two absorption peaks at 252 nm and 349 nm while there was an increase in absorption at 450 nm. The spectral changes are typical of the reduction of Cr (VI) to Cr (III) [60], wherein the latter is less toxic. The absorbance of Cr (III) was recorded (Fig. 16A (Insert)) so as to confirm that its spectrum matched that of the 450 nm peak observed in the photoreduction of Cr (VI). Photoreduction studies were conducted at a pH of 2.75 as it has been reported that the reduction of Cr (VI) is enhanced at acidic pH ranges due to the fact that the net reaction in the photoreduction process consumes H⁺ [60, 61, 62].

Fig. 16.

Absorption spectral changes of 7.08 × 10⁻⁴ mol L^-1 Cr (VI) solution in the absence (A) and presence (B) of 3.40 × 10⁻⁵ molL⁻¹ MO, using CoFe₂O₄-1 in water at 2 min interval at pH 2.75. Insert in (A) is the absorption spectrum of Cr(III).

Photoreduction of Cr (VI) was not effective due to the re-oxidation of Cr (III) to Cr (VI) as explained in the introduction. In the presence of MO, substantial photoreduction results were obtained. This is possibly because, when irradiated with light, MO itself absorbs the photons and gets excited. When excited, it can then transfer electrons to Cr (VI) ions through the conduction band of the ferrites enhancing the reduction of Cr (VI) [63].

Spectral changes observed for Cr (VI) photoreduction in the presence of MO, are shown in Fig. 16B. There was an observed decrease of spectral peaks corresponding to both Cr (VI) (349 nm) and MO (506 nm). An increase in the band 250 nm is a result of the absorption by the oxidation products of MO as observed in Fig. 14. Due to the observed overlap between the Cr (III) and MO bands, increase in the spectral peak for Cr (III) could not be monitored so photoreduction was monitored based solely on decrease of the Cr (VI) peak.

For reasons provided above, kinetic studies for the photoreduction of Cr (VI) were conducted in the presence of 3.40× 10⁻⁵ mol L^-1 MO for each concentration of Cr (VI). The decrease of the peak at 349 nm was monitored and complexes 1 and 2 and their respective composites showed an increase in photoreduction efficiency (initial rates and k_obs) with increase in the concentration of Cr (VI), Table 3.

Table 3.

Rates, rate constants (k_obs) and half-lives (t_1/2) of various initial concentrations of Cr (VI) (in the presence of 3.40 $\times$10⁻⁵ molL⁻¹ MO) using GSH-CoFe₂O₄, complexes 1 and 2, CoFe₂O₄-1 and CoFe₂O₄-2 as photocatalysts.

[Cr (VI)] × 10−4 (mol L−1)	kobs (min−1)					Rate (10−6 mol L-1 min−1)					t1/2 (min)
	GSH-CoFe2O4	1	CoFe2O4-1	2	CoFe2O4-2	GSH-CoFe2O4	1	CoFe2O4-1	2	CoFe2O4-2	GSH-CoFe2O4	1	CoFe2O4-1	2	CoFe2O4-2
9.54	0.0250	0.0309	0.103	0.0305	0.102	23.9	29.5	98.2	29.1	97.3	27.7	22.4	6.70	22.7	6.79
9.03	0.0122	0.0156	0.0674	0.0165	0.0660	11.0	14.1	60.9	14.9	59.6	56.8	44.4	10.3	42.0	10.5
8.54	0.0064	0.0084	0.0450	0.0079	0.0425	5.47	7.17	38.4	6.74	36.3	108	82.5	15.4	87.7	16.3
7.89	0.0020	0.0038	0.0292	0.0035	0.0285	1.58	3.00	2.30	2.76	2.25	346	182	23.7	198	24.3
7.08	0.0009	0.0018	0.0097	0.0019	0.0092	0.637	1.27	6.86	1.35	6.51	770	385	71.4	364	75.3

Plots for ln (C_O/C) versus irradiation time for the photocatalysts are linear, Fig. 17 (using CoFe₂O₄-1 as an example), again showing pseudo-first order reaction kinetics. Complexes 1 and 2 showed lower photocatalytic activity for the reduction of Cr (VI) compared to when linked to MNPs, while GSH-CoFe₂O₄ showed the worst activity for reasons explained above.

Fig. 17.

First order kinetics plots for photoreduction of Cr (VI); i. 7.08 × 10⁻⁴, ii. 7.85 × 10⁻⁴, iii. 8.54 × 10⁻⁴, iv. 9.03 × 10⁻⁴ and v. 9.54 × 10⁻⁴ mol L⁻¹ in the presence of 3.40 × 10⁻⁵ molL⁻¹ MO using CoFe₂O₄-1 as a photocatalyst in water at pH 2.75. Study based on spectra monitored at Cr (VI) peak of 349 nm.

3.3.3. Mechanism of simultaneous photodegradation and photoreduction

It can be elucidated that photocatalysis of Pcs in the presence of CoFe₂O₄ MNPs may occur in two ways, Fig. 18 A and B. Upon excitation with visible light, an electron–hole pair is formed in the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the Pc (route 1 in Fig. 18 A) [64, 65]. In ferrites, upon exposure to visible light, an electron–hole pair is formed in the conduction band (CB) and valence band (VB) respectively (route 2 in Fig. 18 A). The electrons in the LUMO of the Pc are then injected into the conduction band of CoFe₂O₄ MNPs (route 3 in Fig. 18 A). The photoelectrons can attack molecular oxygen to produce various ROS including hydroxyl radicals ($·$OH) (route 4 in Fig. 18 A) [19]. The photoholes possess highly oxidizing ability and can directly participate in oxidative degradation of organics or oxidise water to produce ROS including $·$OH (route 5 in Fig. 18 A).

Fig. 18.

The formation of ROS by photosensitization of CoFe₂O₄ MNPs and Pcs upon excitation with visible light. HOMO = highest occupied molecular orbital, LUMO = lowest unoccupied molecular orbital, ISC= intersystem crossing. T1= Type 1 and T2= Type 2.

Simultaneous photooxidation and photoreduction of MO and Cr (VI), respectively have been reported before using catalysts such as cobalt ferrite meso–macroporous nanospheres, Fe(III)-grafted K₂Ti₆O₁₃ nanobelt and TiO₂ microspheres [63, 66, 67]. Interestingly, although the catalysts and reaction conditions differed, the mutual finding was that the photoreduction of Cr (VI) is enhanced in the presence of MO. This is because when Cr(VI) and MO are in a binary system, Cr(VI) acts as a strong oxidant and rapidly consumes photoelectrons at the conduction band of the MNPs thereby suppressing the charge recombination, MO on the other hand acts an oxidisable organic pollutant which consumes the photoholes at the valence band to further suppress the charge recombination. The suppressed electron−hole recombination then leads to enhanced photocatalytic efficiencies when compared to the single systems [68], as observed in this work and also enhanced by the presence of the Pcs.

A second possible mechanism (B) involves the intersystem crossing (ISC) of the excited Pc to the triplet state. The triplet state (³Pc) has a longer lifetime (μs) than that of the excited singlet state (¹Pc) (ns), enabling the Pc in the triplet excited state to react with molecular oxygen in two different ways, denoted T1 (Type 1) and T2 (Type 2) in Fig. 18 B. In the Type 1 reaction (T1), the Pc in the excited triplet state (³Pc) transfers an electron to molecular oxygen (³O₂) generating various ROS including hydroxyl radicals ($·$OH), peroxides (H₂O₂) and hydroxide ions (OH⁻) [68] which have the ability to readily degrade organic pollutants. The Type 2 reaction (T2) on the other hand entails the transfer of energy from the Pc in triplet excited state (³Pc) to molecular oxygen (³O₂) thereby generating singlet oxygen (¹O₂) [8] which is itself reactive to organic pollutants.

4. Conclusion

The conjugates of complexes 1 and 2 with GSH-CoFe₂O₄ MNPs (represented as CoFe₂O₄-1 and CoFe₂O₄-2 respectively), have been successfully synthesized and applied in a binary system for simultaneous photooxidation of MO and photoreduction of Cr (VI), using visible light. The results showed that conjugation of Pcs to GSH-CoFe₂O₄ MNPs results in improved singlet oxygen quantum yield, and enhanced photocatalytic abilities of the Pcs. The conjugates (CoFe₂O₄-1 and CoFe₂O₄-2) were found to be the best photocatalyst for both the photodegradation of MO and the photoreduction of Cr (VI) relative to their respective Pcs and MNPs. The results in this work have also proven that real life applications for water purification can be devised from Pc-MNP based photocatalysts as they are effective in water containing a mixture of organic and inorganic pollutants.

Declarations

Author contribution statement

Sivuyisiwe Mapukata: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Olawale L. Osifeko: Performed the experiments.

Tebello Nyokong: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This work was supported by the department of Science and Technology, Republic of South Africa and National Research Foundation through DST/NRF South Africa Research Chairs Initiative for Professor of Medicinal Chemistry and Nanotechnology (Grant number UID 62620) and Rhodes University.

Competing interest statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.