3.1 XRD and Raman shift

XRD patterns of MoOS, Co-MoOS, Nb-MoOS, 5Nb-10Co-MoOS, 10Nb-10Co-MoOS, and 15Nb-10Co-MoOS are shown in Fig. 1a. It can be seen that all the as-prepared samples exhibit three poor diffraction peaks at 2θ = 14.1o, 32.7o, and 39.5o which can be indexed to the (002), (100), and (103) it indicate the diffraction planes of hexagonal phase of MoS2 (JCPDS#37–1492). The characteristic peaks of all the as-prepared samples were broad and had low intensity which implies poor crystallinity and a highly disordered structure for each sample. Previous report indicates that addition of metal ions can help to improve the crystallinity of the MoS2 structure [24]. However, this effect was not observed for our case due to possible little contents of Nb and Co. As compared to hexagonal MoS2, the doped samples showed peak shift to higher angle. The small peak shift to a higher angle can attributed to the substitution of Mo4+ with Nb5+ and Co2+ [19]. It is also noted that no characteristic diffraction peaks detected related to Nb and Co compounds in the XRD patterns, indicating that Nb and Co were successfully doped. The as-prepared samples are further examined by Raman spectra, as shown in Fig. 1b. As we can see, the Raman spectra of all the as-prepared samples were identical to amorphous molybdenum oxy-sulfide, indicating that no secondary phases were formed, as reported in the previous literatures [25, 26]. The vibrational bands appearing at 191, 290, 343, and 431 cm− 1 were characteristic of MoS2, which refers to the Mo-S vibrational mode in MoS2 [27, 28]. The band detected at 855 and 951 cm− 1 was related to Mo-O vibrational mode. In brief, the 855 cm− 1 band is attributed to the stretching mode of doubly coordinated oxygen (Mo-O-Mo) in three dimensional arrangement and the band at 951 cm− 1 can be assigned to the vibration of terminal Mo = O bonds [29].

Fig. 1

a XRD patterns and b Raman shifts of MoOS, Nb-MoOS, Co-MoOS, and Nb/Co codoped MoOS photocatalysts with different Nb contents

3.2 Scanning electron microscopy (SEM) analysis

Surface morphologies of the as-prepared MoOS, Co-MoOS, Nb-MoOS, and 10Nb-10Co-MoOS photocatalyst are investigated by SEM as shown in Fig. 2. As can be seen from Fig. 2a, the pure MoOS showed plate-like irregular shape morphology. Figure 2b clearly shows that Nb-MoOS is composed of many aggregated particles. Figure 2c reveals the morphology of Co-MoOS, which contains particles with a non-uniform size and little aggregation. The morphology of 10Nb-10Co-MoOS clearly revealed that the sample had an irregular particle shape with slight agglomeration, as displayed in Fig. 2d. In general, the addition of Nb and Co does not have a significant change in the morphology of the as-prepared samples.

Fig. 2

SEM images of a MoOS, b Nb-MoOS, c Co-MoOS, and d 10Nb-10Co-MoOS

3.3 Transmission electron microscopy (TEM) analysis

The microstructure of 10Nb-10Co-MoOS photocatalyst was analyzed using transmission electron microscope. The TEM image shown in Fig. 3a revealed that the material was aggregate of particles. Figure 3b displays the typically selected area electron diffraction (SAED) pattern of 10Nb-10Co-MoOS with a diffuse ring pattern, which is in agreement with the XRD results for the amorphous nature of the as-prepared material. As can be seen from Fig. 3c, there is no well-defined lattice fringe for the as-prepared sample. The absence of lattice fringes in the HRTEM image indicated that the sample was amorphous [30]. The chemical composition and atomic percentage distribution of elements in the as-prepared material were examined with energy-dispersive X-ray spectroscopy (EDX) measurement, as shown in Fig. 3d. The spectrum showed the presence of all the elements i.e. Nb, Co, Mo, O, and S, except a strong peak found at around 8 keV was belonged to copper from sample holder. The presence of Nb and Co in codoped MoOS confirms the successful codoping of Nb and Co in MoOS. The weight and atomic percentage distribution of elements are tabulated and are shown in the inset of Fig. 3d.

Fig. 3

a TEM image, b selected area electron diffraction (SAED) patterns, c HRTEM lattice image, and d the EDX spectrum of 10Nb-10Co-MoOS photocatalyst

3.4 Electrochemical impedance spectroscopy (EIS), photocurrent response, photoluminescence (PL), and diffuse reflectance spectra (DRS) analysis

Electrochemical impedance spectroscopy (EIS) measurement was performed to evaluate the charge transfer resistance and separation efficiency between photo-induced charge carriers because the charge separation efficiency of photogenerated electron-hole pairs is an important factor for photocatalytic activity. The smaller the arc radius in the Nyquist plot exhibited that an effective separation of photo-produced electron-hole pairs, and the larger the arc radius the higher the recombination rate of photo-produced charge carriers [23]. Figure 4a shows the Nyquist plots of MoOS, Co-MoOS, Nb-MoOS, and 10Nb-10Co-MoOS. As we can see, the arc radius of codoped samples was smaller than the pure and mono-doped samples, which indicates that 10Nb-10Co-MoOS photocatalyst has smaller charge transfer resistance and effectively suppresses recombination rate of photogenerated electron-hole pairs. On the other hand, the arc radius of Nb-MoOS was larger than other as-prepared samples, which indicates that Nb-MoOS has a higher charge transfer-resistant and faster recombination rate of charge carriers. From these results, it is clear that codoped samples improve the charge transfer and reduce the recombination rate of electron-hole pairs. The photocurrent responses of MoOS, Co-MoOS, Nb-MoOS, and 10Nb-10Co-MoOS were analyzed under visible light irradiation and the results are shown in Fig. 4b. As we can see, the photocurrent density of 10Nb-10Co-MoOS was higher than other tested samples and the lowest photocurrent density was observed for Nb-MoOS. From these results, it can be concluded that the sample with 10 wt% Co/10 wt% Nb codoped achieve the highest charge transfer and charge separation capacity, which is important for photocatalytic activity. The PL spectra are useful to evaluate the recombination rate of charge carriers and to understand the destiny of electron-hole pairs in the semiconductor. As illustrated in Fig. S1, the highest PL intensity was observed for Nb-MoOS, which indicates that it has a high recombination rate of charge carriers. The lowest PL value was exhibited by 10Nb-10Co-MoOS, which implies that the recombination rate of electron-hole pairs was very low. The absorption band of photocatalyst will change when extrinsic dopants go into the lattice. Therefore, the as-prepared catalysts were examined by UV-Vis diffuse reflectance spectroscopy (DRS) to estimate their bandgap energy. Figure 4c showed the UV-Vis absorption spectra of all the samples. It is perceivable that the codoped samples gradually shift the absorption edge towards the longer wavelength region with the absorption edge extended to the NIR region. The energy bandgaps of the as-prepared photocatalysts were estimated from the intercept of the tangents to the plot (αhv)2 vs. photon energy from Tauc relation [31]. As shown in Fig. 4d, the bandgap energy value of pure MoOS was 1.71 eV, whereas the bandgaps of codoped samples were between 1.44 and 1.50 eV, indicating that the simultaneous codoping of Niobium and Cobalt into the crystal lattice of MoOS with a small red shift of the absorption.

Fig. 4

a Electrochemical impedance spectroscopy (EIS), b photocurrent response c diffuse reflectance spectra (DRS) and d the (αhν)2 versus hν plots of MoOS, Nb-MoOS, Co-MoOS, and Nb/Co codoped MoOS with different contents of Nb

3.5 X-ray photoelectron spectroscopy (XPS) analysis

XPS analysis was conducted to determine the elemental composition and chemical states of 10Nb-10Co-MoOS. The full scan spectrum of 10Nb-10Co-MoOS shown in Fig. 5a, sample contained mainly Nb 3 d, Co 2p, Mo 3 d, O 1 s, S 2p, and C 1s. The presence of carbon elements came from the contamination of the sample that exposed to the atmosphere or residual carbon. Figure 5b shows the high-resolution XPS spectrum of Molybdenum and sulfur, which can be resolved into three peaks with the binding energy of 226.5, 230.2, and 233.1 eV. The binding energy located at 230.2 and 233.1 eV attributed to the doublet Mo 3d5/2 and Mo3d3/2, respectively, to confirm the existence of Mo in a 4 + oxidation state [32]. The binding energy at 226.5 eV was assigned to S 2 s, which strongly indicates the presence of MoS2 [33]. As displayed in Fig. 5c, the binding energies located at 207.1 and 209.7 eV correspond to Nb 3d5/2 and Nb 3d3/2 doublets, respectively, which indicates that Nb presents in 5 + oxidation state [34]. The Co 2p3/2 and Co 2p1/2 peaks were found at binding energies of 782 eV and 794.4 eV, respectively, as displayed in Fig. 5d. This results revealed that Co was exist in a 2 + oxidation state [35], but the peak intensities of Co were too low, compared to other peaks, due to the mass loss of cobalt during preparation and washing. The S 2p spectrum shown in Fig. 5e exhibit two spin-orbit splits S 2p3/2 and S 2p1/2 at binding energies of 161.3 and 162.2 eV, respectively, which correspond to S− 2 [36, 37]. The binding energies positioned at 528.8, 530.3, and 531.4 eV shown in Fig. 5f associated with O− 2 species and were assigned to lattice oxygen (OL) bound to metals, oxygen vacancies or defects (OV), and adsorbed oxygen species (Oads.), respectively [38]. From the XPS peak area analysis for 10Nb-10Co-MoOS photocatalyst, the surface atomic composition (percentage) of Mo 3 d, Nb 3 d, Co 2p, S 2p, and O 1 s were 29.89%, 4.49%, 1.57%, 17.23%, and 46.82%, respectively, which is consistent to the value obtained from EDX analysis.

Fig. 5

XPS spectra of a a full scan, b Mo 3 d, c Nb 3 d, d Co 2p, e S 2p, and d O 1 s for 10Nb-10Co-MoOS

3.6 Photocatalytic activity

The photocatalytic performance of the prepared photocatalysts was examined by the degradation of MO, RhB, and MB in aqueous solution under visible light irradiation. For all the samples before visible light irradiation, the dye solution along with catalyst was stirred magnetically under dark for 30 min to achieve adsorption-desorption equilibrium between the catalyst and the dye solution. The degradation absorption spectra of MO were displayed in Fig. 6a. It can be seen that with prolongation of the irradiation time, the absorption peak of MO located at a wavelength of 464 nm and the orange color of the solution decreases gradually and disappear after 150 min, indicating that the molecular structure of the dye destroyed. Figure 6b displays a comparison of photocatalytic activities of MoOS, Co-MoOS, Nb-MoOS, 5Nb-10Co-MoOS, 10Nb-10Co-MoOS, and 15Nb-10Co-MoOS on the photodegradation of MO in aqueous solution under visible light. As is evident from the degradation curve, the degradation of MO in the absence of visible light (dark) almost negligible, indicating that the MO dye does not degrade without light irradiation. The codoped photocatalyst show enhanced photocatalytic activity compared to MoOS, Co-MoOS, and Nb-MoOS. The 10Nb-10Co-MoOS sample showed an optimal photocatalytic activity with 94.4 % degradation of MO in 150 min. Figure 6c shows the characteristic absorption spectra of RhB at different intervals of time in the photodegradation process over 10Nb-10Co-MoOS. It can be seen that after 20 min irradiation of visible light, the absorption peak of RhB located at 554 nm completely vanished, it indicates that the dye molecule has been completely degraded. Figure 6d illustrates the time-dependent photodegradation of RhB over the as-prepared samples, including the adsorption of RhB onto the samples in the dark (without light) for 30 min. All photocatalysts showed appreciable adsorption towards the dye during the dark period. The phenomenon of strong adsorption in the dark is beneficial for the photodegradation of dyes. After 20 min of photocatalytic reaction, the degradation efficiency of MoOS, Co-MoOS, Nb-MoOS, 5Nb-10Co-MoOS, 10Nb-10Co-MoOS, and 15Nb-10Co-MoOS were 65.5 %, 93.4 %, 78.3 %, 95.3 %, 97.1 %, and 96.8 %, respectively. These results showed that the codoped photocatalysts exhibited an enhanced photodegradation activity. By varying the Nb content while keeping a constant 10 % Co, the optimal photocatalyst was observed for Nb/Co-MoOS at 10 % Nb, which degrade 97.1 % of RhB within 20 min irradiation time. It has been observed that when the content of Nb is higher than 10%, the photodegradation efficiency of the catalyst decreases. The reason why the photodegradation activity decreased may be caused by surface coverage by excess Nb to reduce the visible light absorption and enhance the recombination of photogenerated electron-hole pairs [39]. The photodegradation of MB as a function of time under visible light irradiation is showed in Fig. 6e. As seen from the figure the absorption peak intensity decreased quickly and was completely eliminated after 40 min of exposure to visible light. The photodegradation curves of MB for all as-prepared samples are displayed in Fig. 6f. The concentration of MB did not significantly vary without visible light irradiation (dark), indicating that the degradation of MB is negligible at the condition without light illumination. Compare to pure and single metal-doped samples, the codoped catalysts exhibited excellent photocatalytic activity. Among all, the photocatalyst with 10 wt% Nb showed photodegradation efficiency of 99.7 % in 40 min. Generally, the photodegradation efficiency of MoOS, Co-MoOS, Nb-MoOS, 5Nb-10Co-MoOS, 10Nb-10Co-MoOS, and 15Nb-10Co-MoOS were 90 %, 89.3 %, 73.9 %, 95.3 %, 99.7 %, and 94.4 %, respectively. For testing at all three dyes, the photocatalytic performance increased with an increase in Nb content from 5 to 10 wt%. The possible reasons can be more active sites available on the surface, the light-harvesting ability, and the improved charge transfer capacity of the catalyst. When the Nb amount further increased to 15 wt%, the photocatalytic activity of the catalyst decreased to some extent due to more Nb content was found on the surface, which led to particle agglomeration, the coverage of accessible active sites, and also become recombination center for electron-hole pairs [40]. Generally, for the photodegradation of all dyes, 10Nb-10Co-MoOS are the optimal photocatalyst. From the PL, EIS, and photocurrent response results, 10Nb-10Co-MoOS exhibits a better separation efficiency of photo-induced charge carriers than other as-prepared materials, that is the reason why 10Nb-10Co-MoOS shows better photocatalytic activity.

Fig. 6

a Absorption spectra of MO photodegradation, b photodegradation curves of MO, c absorption spectra of RhB photodegradation, d photodegradation curves of RhB, e absorption spectra of MB photodegradation, and f photodegradation curves of MB over the as-prepared catalysts under visible light

In order to quantitatively understand the reaction kinetics of MO, RhB, and MB photodegradation over different as-prepared samples, kinetic curves were plotted. Fig. S2a-c illustrates that the photodegradation of MO, RhB, and MB follows a pseudo-first-order kinetics with the formula: ln(Ct/Co) = -kt [49], where Co is initial concentration, Ct is the concentration at time t, k is the reaction rate constant, and t is time. The reaction rate constant value of MO, RhB, and MB removal in the photodegradation process over 10Nb-10Co-MoOS was 0.016 min− 1, 0.066 min− 1, and 0.096 min− 1, respectively. It is about 2.58, 3.47, and 2.82 times higher than that of pure MoOS, respectively. Generally, the rate constants for all as-prepared samples are given in the Table 1.

Table 1 Photodegradation rate constants for different dyes over Nb/Co-MoOS

To confirm the photodegradation of MO, RhB, and MB dye over 10Nb-10Co-MoOS, HPLC analysis was carried out over time for each dye and the results are displayed in Fig. 7a-c. The intense peaks located at retention times of 3.1, 2.3, and 1.7 min in each figure are assigned to MO, RhB, and MB, respectively. After exposing the dye solutions to visible light, the peak intensity of each dye diminished gradually over time and finally disappeared without showing other intermediate products, indicating that the dyes were completely degraded. During the degradation of MO dye, the HPLC peak was shifted towards the higher retention time, indicating that cleavage of MO dye occurred.

Fig. 7

HPLC chromatogram for a MO, b RhB, and c MB dyes degradation

The total organic carbon (TOC) removal efficiencies for three common organic dyes MB, RhB, and MO over the 10Nb-10Co-MoOS photocatalyst was examined and given below in Table 2. As can be seen from Table 3, MB shows the highest TOC removal efficiency in 40 min; this indicates that MB undergoes fast photocatalytic mineralization whereas, the RhB dye shows moderate TOC removal compared to MB dye within 20 min because Rhodamine B is stable aromatic compound [50, 51]. The MO dye shows the lowest TOC removal of 58 % in 150 min, this is due to the azo bond (-N = N-) and the molecular structure are more stable [51]. Generally, these TOC results show that 10Nb-10Co-MoOS photocatalyst promising for photodegradation and mineralization of organic pollutants.

Table 2 Total organic carbon (TOC) removal efficiency for the three dyes (MB, RhB, and MO) dyes degraded over 10Nb-10Co-MoOS photocatalystTable 3 Comparison of photodegradation efficiencies of different dyes using 10Nb-10Co-MoOS with other photocatalysts reported in the literatures3.7 Reusability test

From the practical application point of view, reusability and stability of the catalyst are very essential. So that, the reusability test was investigated using the best-performed catalyst 10Nb-10Co-MoOS for the photodegradation of MO and RhB dyes for three consecutive runs and the results are displayed in Fig. 8a and b. As can be seen, no significant decrease in the photodegradation efficiency for both dyes was observed after three runs, indicating that the catalyst was stable. The slight loss in efficiency related to deactivation of the catalyst surface and loss of mass of catalyst during the washing and drying process. In addition, Fig. 8c shows that the XRD patterns of 10Nb-10Co-MoOS remain unchanged before and after the three runs of the photodegradation reaction, indicating that 10Nb-10Co-MoOS is rather photo-stable.

Fig. 8

Stability studies of 10Nb-10Co-MoOS on the photocatalytic degradation of a MO and b RhB. c XRD patterns of fresh and reused catalysts

3.8 Active species trapping experiment and reaction mechanism

The active species trapping experiments were conducted for photodegradation of MO, RhB, and MB dyes using the scavenger’s p-BQ, IPA, and EDTA to capture O2.−,.OH, and h+, respectively. As displayed in Fig. 9a the photodegradation of MO was significantly inhibited after the addition of p-BQ. Whereas, when IPA and EDTA were added in the reaction system, the photodegradation of MO was improved because of the inhibition of the recombination rate of electron-hole pairs and the enhanced formation of O2.−. These results exhibited that O2.− plays a vital role in the degradation of MO. For the degradation of RhB, the introduction of both IPA and p-BQ in the reaction system reduces the activity of the catalyst but the addition of EDTA has little effect on the degradation of RhB, as shown in Fig. 9b, implying that.OH and O2.− radicals are the main active species in the photodegradation of RhB. On the other hand, it is found from Fig. 9c that the photodegradation of MB reduced by adding all these scavengers indicates that the photodegradation reactions are related to O2.−,.OH, and h+. In specific, O2.−was playing a vital role in the degradation of MB. Therefore, 10Nb-10Co-MoOS under visible light irradiation can produce three kinds of active species for dye degradation. Basically, a photocatalyst can only be mentioned to be good if it can simultaneously to degrade the three kinds of MO, RhB, and MB model dyes. The Mott-Schottky plot displayed in Fig. S3 indicates that the photocatalyst 10Nb-10Co-MoOS was n-type semiconductor and the flat band potential of this catalyst estimated from the intercept on the X-axis was about − 0.259 V at pH = 7. Based on the formula: ENHE= EAg/AgCl +EƟAg/AgCl (EƟAg/AgCl = 0.197 V) [52], the conduction band potential obtained for this photocatalyst is estimated to be −0.062 V versus NHE, whereas the valence band potential is 1.38 V versus NHE, based on the bandgap data from the UV-Vis DRS spectra.

Fig. 9

Active species capturing experiments for a MO, b RhB, and c MB over 10Nb-10Co-MoOS photocatalyst under visible light irradiation

From the experimental results obtained above, we propose the reaction mechanism as shown in Fig. 10. When the photocatalystis illuminated with visible light, the electrons are excited from the valence band (VB) to the conduction band (CB) of 10Nb-10Co-MoOS and holes are left in the valence band (Eq. 2). Then the photogenerated electrons in the CB of 10Nb-10Co-MoOS react with an oxygen molecule in water to produce superoxide radicals because CB potential of 10Nb-10Co-MoOS (−0.062 V) is more negative than that of E(O2/O2˙−) which is −0.046 V versus NHE (Eq. 3) [53]. In addition, some of the superoxide radicals further react with protons to produce hydroxyl radicals (Eq. 4). Since the valence band potential of 10Nb-10Co-MoOS (1.38 V vs. NHE) is less positive than OH−/˙OH (+ 2.40 V vs. NHE) [54], so hole cannot convert water molecules into hydroxyl radicals. On the other hand, the holes remained on the valence band and the produced radicals can be used to oxidize the organic dyes because they have strong oxidative abilities (Eq. 5). The introduction of dopant ions it may involve the reduction of the bandgap energy to harvest visible light. The proposed photodegradation reaction mechanisms of the dyes are stated below:

$$\:10Nb - 10Co - MoOS + hv \to \:10Nb - 10Co - MoOS\left( + h^ } \right)$$

(2)

$$\:10Nb - 10Co - MoOS\left( } \right) + O_ \to \:10Nb - 10Co - MoOS + O_^}$$

(3)

$$\:O_^} + 2H^ \to \:2\mathop \limits^ OH$$

(4)

$$\:O_^} \:/ \bullet \:OH\:/h^ + Dyes \to \:\:Degradation\:\:\:products$$

(5)

.

Fig. 10

Schematic representation of the reaction mechanism proposed for photodegradation of dyes over 10Nb-10Co-MoOS catalyst under visible light irradiation