Photocatalytic Decomposition of an Azo Dye Using Transition-Metal-Doped Tungsten and Molybdenum Carbides.

The preparation, characterization, and photocatalytic application of tungsten or molybdenum carbides (Ni-WC, 1, Co-WC, 2, Ni-MoC, 3, Co-MoC, 4, NiCo-WC, 5, NiCo-MoC, 6, NiFe-WC, 7, and NiFe-MoC, 8) doped with transition metals (Fe, Co, and Ni) are reported. These transition-metal carbide (TMC) particles show that the submicrometer globular particles agglomerated to form larger particles, with smaller crystallites present on the surface of the large particles. These crystallite sizes range between 4 and 34 nm (as calculated from X-ray diffraction data) depending on the metal dopant and type of carbide. Oxidation of the metal carbides is evident from the two sets of photoelectron lines present in the X-ray photoelectron spectroscopy (XPS) of the W 4f area. The Mo 3d spectra reveal four sets of photoelectron lines associated with oxidized MoO2 and MoO3 as well as Mo2+ and Mo3+ associated with MoC1-x . The XPS of the dopant metals Ni, Co, and Fe also show partial oxidation. The photocatalytic decomposition of Congo red (an azo dye) is used as a model reaction to determine the photocatalytic activities of the transition-metal carbides, which is related to the TMCs' optical band gap energies.

Introduction

Transition-metal carbides (TMCs), such as molybdenum and tungsten carbides, exhibit remarkably different chemical and physical properties compared to the parent metal or metal oxides (from which they are prepared) on account of the incorporation of the metal–carbon bond. These TMCs have shown potential as a more economical alternative catalyst for an assortment of reactions routinely catalyzed by noble metals. These catalysis reactions include hydrogenation, biomass conversion, water electrolysis, water gas shift reactions, alcohol electrooxidation, and the removal of contaminants (such as nitrogen and sulfur) through hydrotreating.[1] The catalytic properties of TMCs can be enhanced by incorporating a second metal, e.g., Co/MoC and Ni/MoC, due to an improved surface structure.[2] When β-Mo2C is doped with a small amount of Fe, Co, or Ni, the activity and stability for steam reforming of methanol are improved.[3] It was also reported that the addition of Co to Mo2C not only improves the activity and selectivity of the Co-doped Mo2C catalyst toward CO2 reduction but also improves its durability.[4] Transition metals such as Ni, Co, and Fe are often selected as doping agents because of their ability to combine electrical and optical properties into a single material. They provide a substitute level nearly above the conduction band, which improves the absorption of visible light and utilization, and decrease the electron–hole recombination rate,[2] which is useful during photocatalysis. Recently, metal carbides have also emerged as a promising photocatalyst,[5] Co-doped MoC has been reported to photocatalytically degrade Maxilon Blue GRL 300 basic dye.[2] Since many industries such as textile, plastic, printing, photographic, paper-pulp, paint, and leather factories discard wastes,[6−11] such as hazardous chemicals and synthetic dyes (including highly toxic azo dyes), into rivers and streams, pollution of water has increased dramatically. This then affects the aquatic environment. Water pollution blocks the sunlight from penetrating the water, causing algae to grow and endangering water life. Azo dyes are colored organic compounds having two nitrogen atoms linked to each other (−N=N−). Congo red (see Figure for the chemical structure) is a synthetic dye that belongs to this azo group, with two azo chromophores. It is a highly toxic dye that is very difficult to degrade because of its stable aromatic structure.[12] Congo red is no longer used for dye purposes due to its carcinogenic nature,[13] but it is still used for staining to detect amyloids under a microscope.[14]

Figure 1

Molecular structure of Congo red.

There are many methods to treat wastewater, namely, coagulation, electrocoagulation, chlorination, ozonation, flotation, chemical oxidation, filtration, membrane separation, adsorption, and ultrafiltration.[15−23]

However, photocatalytic degradation is considered a promising and energy-saving technology to remove high concentrations of biodegradable pollutants. The dye molecule decomposes when it interacts with a photocatalyst in the presence of ultraviolet/visible light from either solar energy or artificial light sources. During the interaction between the dye and the photocatalyst, charge separation is generated by a charge-transfer process. This leads to the formation of reactive oxygen species necessary for the oxidation and subsequent mineralization of the organic contaminants.[24]

Modified TMCs could be a good choice for practical applications in photocatalytic processes because of their efficiency to degrade organic pollutants in wastewater. The present report discusses the synthesis and characterization (X-ray diffraction (XRD), scanning electron microscopy (SEM), time of flight secondary mass spectroscopy (ToF-SIMS), and X-ray photoelectron spectroscopy (XPS)) of tungsten and molybdenum carbides doped with transition metals (Fe, Co, and Ni). These TMC catalysts are tested for the photocatalytic decomposition of the pollutant Congo red, which is an example of an azo dye. Simulated sunlight is used as the irradiation source and the photocatalytic activity of the TMCs is related to the type of carbide, metal dopant, and the particles’ optical band gap energy.

Experimental Section

Materials

Nickel(II) nitrate hexahydrate [Ni(NO3)2·6H2O], cobalt(II) nitrate hexahydrate [Co(NO3)2·6H2O], iron(II) nitrate hexahydrate [Fe(NO3)2·6H2O], ammonium metatungstate hydrate [(NH4)6H2W12O40·xH2O], ammonium molybdate tetrahydrate [(NH4)6Mo7O24·4H2O], multiwalled carbon nanotubes, MWCNTs (>98% carbon basis, O.D. × L6–13 nm × 2.5–20 μm), and Congo red (indicator grade) were used in this study. All these analytical grade chemicals were purchased from Merck (Pty) Ltd. and used without further purification.

Synthesis of Transition-Metal Carbide Nanoparticles

A solid-state reaction (carburization method) was used to synthesize transition-metal carbides. The process involves reducing the desired metal precursors using a carbon-based source (MWCNTs) as a reducing agent at high temperatures.

Synthesis of Ni-WC, 1

A combination of Ni(NO3)2·6H2O (0.049 g; 0.19 mmol; ∼15 equiv), (NH4)6H2W12O40·xH2O (0.038 g; 0.013 mmol; 1 equiv), and MWCNTs (0.013 g; 0.07 mmol; ∼5 equiv) was mixed homogeneously and ground using a pestle and mortar resulting in a fine powder. The powder was then placed in a furnace for three hours and heated at 950 °C in an H2–N2 mixture atmosphere. While the furnace cooled, the gas mixture was replaced by argon until room temperature was attained. This resulted in the isolation of Ni-WC as a gray powder.

Synthesis of Co-WC, 2

Co-WC was prepared by the same procedure used for Ni-WC except Ni(NO3)2·6H2O was replaced with 0.049 g (0.17 mmol; ∼13 equiv) cobalt(II) nitrate hexahydrate [Co(NO3)2·6H2O]. This resulted in the isolation of Co-WC as a dark gray powder.

Synthesis of Ni-MoC, 3

Ni-MoC was prepared by the same procedure used for Ni-WC except a combination of Ni(NO3)2·6H2O (0.049 g; 0.19 mmol; 10 equiv), (NH4)6Mo7O24·4H2O (0.038 g; 0.019 mmol; 1 equiv), and MWCNTs 0.013 g (0.07 mmol; ∼3.7 equiv) was used. This resulted in the isolation of Ni-MoC as a gray powder.

Synthesis of Co-MoC, 4

Co-MoC was prepared by the same procedure used for Ni-WC except a combination of Co(NO3)2·6H2O (0.049 g; 0.17 mmol; ∼9 equiv), (NH4)6Mo7O24·4H2O (0.038 g; 0.019 mmol; 1 equiv), and MWCNTs 0.013 g (0.07 mmol; ∼3.7 equiv) was used. This resulted in the isolation of Ni-MoC as a gray powder.

Synthesis of NiCo-WC, 5

A combination of Ni(NO3)2·6H2O (0.062 g; 0.24 mmol; ∼19 equiv), Co(NO3)2·6H2O (0.028 g; 0.096 mmol; ∼7.5 equiv), (NH4)6H2W12O40·xH2O (0.038 g; 0.013 mmol; 1 equiv), and MWCNTs (0.013 g; 0.07 mmol; ∼5 equiv) was mixed homogeneously with a pestle and mortar. The mixture was transferred to a tube furnace and heated at 950 °C for three hours in an H2–N2 mixture atmosphere. While the furnace was cooled, the gas mixture was replaced by argon until room temperature was achieved. This resulted in the isolation of NiCo-WC as a black powder.

Synthesis of NiCo-MoC, 6

NiCo-MoC was prepared by the same procedure used for NiCo-WC except a combination of Ni(NO3)2·6H2O (0.062 g; 0.24 mmol; ∼19 equiv), Co(NO3)2·6H2O (0.049 g; 0.17 mmol; ∼9 equiv), (NH4)6Mo7O24·4H2O (0.038 g; 0.019 mmol; 1 equiv), and MWCNTs (0.013 g; 0.07 mmol; ∼3.7 equiv) was used.

This resulted in the isolation of NiCo-MoC as a light gray powder.

Synthesis NiFe-WC, 7

NiFe-WC was prepared by the same procedure used for NiCo-WC except a combination of Ni(NO3)2·6H2O (0.062 g; 0.24 mmol; ∼19 equiv), Fe(NO3)2·6H2O (0.028 g; 0.07 mmol; ∼5 equiv), (NH4)6H2W12O40·xH2O (0.038 g; 0.013 mmol; 1 equiv), and MWCNTs (0.013 g; 0.07 mmol; ∼5 equiv) was used.

This resulted in the isolation of NiFe-WC as a black powder.

Synthesis of NiFe-MoC, 8

NiFe-MoC was prepared by the same procedure used for NiCo-WC except a combination of Ni(NO3)2·6H2O (0.062 g; 0.24 mmol; ∼19 equiv), Fe(NO3)2·6H2O (0.028 g; 0.07 mmol; ∼5 equiv), (NH4)6Mo7O24·4H2O (0.038 g; 0.013 mmol; 1 equiv), and MWCNTs (0.013 g; 0.07 mmol; ∼5 equiv) was used.

This resulted in the isolation of NiFe-MoC as a black powder.

Characterization

The crystal structures of various TMCs were analyzed by a PAN analytical X’pert PRO X-ray diffractometer (XRD) using Cu Kα radiation in the range of 25–80°. The average crystallite size (Dhkl) was estimated using the Debye–Scherrer equationwhere k is the crystallite shape coefficient ∼0.9, λ is the wavelength of the radiation, β is the full width at half-maximum (FWHM), and θ is the Braggs angle of diffraction.

Surface particle morphology and elemental composition of TMCs (1–8) were analyzed using field emission scanning electron microscopy (FE-SEM) coupled with energy-dispersive X-ray spectroscopy (EDS) respectively. The SEM images were captured using a Shimadzu Superscan ZU SXX-550 electron microscope. The electron beam energies were in the range of 5 keV.

Surface characterization of these TMCs was carried out with a PHI 5000 Versaprobe-Scanning XPS. A monochromatic Al Kα radiation with hv = 1486.6 eV was used, which was generated by a 25 W, 15 kV electron beam. A low-energy neutralizer electron gun was used to minimize the charging of the samples. For high-resolution spectra, the hemispherical analyzer pass energy was maintained at 93.90 eV with a 0.1 eV step. The resolution of the PHI 5000 Versaprobe system was FWHM = 0.53 eV at the pass energy of 23.5 eV and FWHM = 1.44 eV at the pass energy of 93.9 eV. The X-ray beam size used for the XPS measurements was 10 μm. The pressure during acquisition was less than 1 × 10–8 Torr. All of the absolute binding energies of the photoelectron spectra were corrected with a C 1 s signal at 284.8 eV (the lowest binding energy of the simulated adventitious C 1 s photoelectron line).[30] The XPS data was analyzed by a Multipak version 8.2c computer software,[31] and applying Gaussian–Lorentz fits (the Gaussian/Lorentz ratios were always >95%). The neat samples were held in place on the sample holder by means of carbon tape, and the samples were sputtered with different Ar+ ion beam energies in the range of 0.5–4.0 keV.

ToF-SIMS is a surface characterization technique used to investigate the composition of a sample. This is achieved by sputtering the surface with an ion beam and analyzing the collected secondary ions. ToF-SIMS measurements were conducted on a PHI TRIFT V nanoTOF. The UV–vis absorption spectra of the TMCs, for determining their band gap energy, were recorded using a Perkin Elmer Lambda 950 UV–vis at room temperature in the range of 300–800 nm.

The optical band gap energy is an important parameter of photocatalytic performance. The optical band gap energies of the transition-metal carbides were calculated from the % reflectance vs wavelength graphs using Tauc’s equation (eq )where α is an absorption coefficient, hν is the photon’s energy, A is the proportional constant, Eg is the band gap energy, and n is the nature of the sample transition. The value of n for allowed direct, allowed indirect, forbidden direct, and forbidden indirect transitions are 0.5, 2, 3/2, and 3, respectively.[25] According to Kubelka–Munk,[26] α is proportional to the measured reflectance (R) and can be expressed byThis then allows the construction of a Tauc plot, which shows a relationship between the absorption coefficient and the optical band gap. The optical band gap of the TMCs was determined by extrapolating the linear portion of the curves (see the Supporting Information for the graphs).

Photocatalytic Application (Congo Red Dye Degradation)

The photocatalytic activity of the transition-metal-doped TMC particles was studied for the photocatalytic degradation of Congo red dye (Figure ) under simulated sunlight irradiation using a 350 W metal halide lamp in the open air at room temperature. Fifty milligrams of the catalyst was placed in a 250 cm3 beaker containing 100 cm3 of dye solution (20 mg dm–3, pH = 7.0), which were stirred magnetically. Five cubic centimeter samples were taken at random time intervals and the change in absorption of the Congo red was measured using a Shimadzu CPS-240A UV–vis spectrophotometer, corresponding to λmax of dye = 495 nm. Dark adsorption (in the presence of a catalyst but no light) and photolysis (exposure to light but no catalyst) degradation reactions were carried out to differentiate between the adsorption and photocatalytic degradation. Adsorption experiments were performed in the dark, while photocatalytic tests were performed with light irradiation.

Results and Discussion

Synthesis

Various transition-metal-doped tungsten and molybdenum carbides (Ni-WC, 1, Co-WC, 2, Ni-MoC, 3, Co-MoC, 4, NiCo-WC, 5, NiCo-MoC, 6, NiFe-WC, 7, and NiFe-MoC, 8) were prepared by carbothermal reduction. This process involves the reduction of the desired combination of metal oxides at a high temperature (950 °C) using MWCNTs as the reducing agent in an H2–N2 atmosphere.

Structural Analysis

The crystal phase, purity, and crystallite size of the TMCs, 1–8, were analyzed by XRD and are displayed in Figure . Acceptable matches were observed for the bimetallic catalysts. The Co-WC (1) pattern correlated to the Co3W9C4 (ICDD 01–072–1362) and Co6W6C (ICDD 00-023-0939), the Ni-WC (2) pattern to the Ni2W4C (ICDD 00-020-0796), and Ni-MoC to Ni6Mo6C (ICDD no. 03-065-4436) and MoC (ICDD no. 20-0748), while Co-MoC corresponded to Co3Mo3C (ICDD no. 03-065-7128) and β-Mo2C (ICDD no. 45–1014), the Fe-WC to Fe6W6C (ICDD no. 01–089–2616) and W2C (ICDD no. 01-089-2371), and Fe-MoC to Fe2Mo4C (ICDD no. 01-089-4884) and Fe7Mo3 (ICDD no. 00-045-1230). The trimetallic carbides (NiCo-WC, 5, NiCo-MoC, 6, NiFe-WC, 7, and NiFe-MoC, 8) matched well with their corresponding bimetallic carbides.

Figure 2

XRD patterns of (a) Co-WC (2), NiCo-WC (5), and Ni-WC (1); (b) Co-MoC (4), NiCo-MoC (6), and Ni-MoC (3); (c) NiFe-WC (7) and Ni-WC (1); and (d) NiFe-MoC (8) and Ni-MoC (3). The XRD peaks in the trimetallic carbides are assigned to the characteristic peaks of both the bimetallic carbides. The peaks marked with an asterisk (*) are not from impurities but from tungsten oxide (WO3).

From the crystallite sizes of 1–8 presented in Table , the Co-doped TMCs displayed smaller crystallite sizes for the bimetallic TMCs (2 and 4vs1 and 3) and trimetallic TMCs (5 and 7vs6 and 8), while the WC trimetallic TMCs (5 and 6vs7 and 8) displayed smaller crystallites compared to MoC.

Table 1

Average Crystallite Sizes Calculated for 1–8 from the XRD Data Using (eq )a

no.	sample	average crystallite size (nm)	% decolorization of Congo red after 25 min	optical band gap energy (eV)
1	Ni-WC	34	86	2.77
2	Co-WC	4	66	2.50
3	Ni-MoC	25	94	2.00
4	Co-MoC	22	72	1.83
5	NiCo-WC	7	97	2.3
6	NiCo-MoC	15	72	2.34
7	NiFe-WC	22	44	1.97
8	NiFe-MoC	32	41	1.66

The % decolorization of Congo red after 25 min of photocatalytic reaction time in the presence of the TMC catalyst. The optical band gap energy was determined using the Tauc plots of 1–8.

Morphology and Chemical Composition Analysis

Figure shows the SEM images with a low magnification of 1–8 (the high-magnification SEM and EDS data are presented in the Supporting Information). Apart from 8, the TMCs (1–7) show agglomeration of submicrometer globular particles into larger particles. Small crystallites can be observed on the surface of the large particles in the high-magnification SEM images. This type of morphology is consistent with other transition-metal-doped tungsten and molybdenum carbides reported by Regmi et al.[27] EDS confirms the presence of the desired elements in the TMCs 1–8 (see the Supporting Information for the EDS data).

Figure 3

SEM images of 1–8.

Surface Composition and Chemical State Analysis (XPS)

XPS is a convenient technique for identifying the elements present in a sample, the oxidation state of the elements, and even the chemical environment that surrounds the element.

The measured binding energies for the main photoelectron lines were charged correctly against the simulated adventitious carbon set at 284.8 eV.[28] Simulated curve fitting of the C 1s area of the TMCs (1, 2, and 5–8) enabled the location of the carbide photoelectron line at ca. 283.5 eV (see Table and Figure ), which correlates well with the reported binding energy range of 283.3–283.5 eV for WC1–.[29−31]

Table 2

Binding Energies (BE Measured in eV) of the Main Photoelectron Lines and If Applicable the Satellite Structures of the W 4f7/2, Mo 3d5/2, C 1s, O 1s, Ni 2p3/2, Co 2p3/2, and Fe 2p3/2 of 1, 2, and 5–8

	W 4f7/2		Mo 3d5/2				C 1s	O 1s	Ni 2p3/2			Co 2p3/2			Fe 2p3/2
	WC	WO2/3	Mo2+	Mo3+	Mo4+	Mo6+	carbide	oxide	Ni0	Ni2+	Nisat	Co0	Co2+/3+	Cosat	Fe0	Fe2+	Fesat
1: Ni-WC	31.4	35.6					283.8	530.8	852.4	855.4	860.2
2: Co-WC	31.6	35.5					283.8	530.9				778.7	780.9	786.3
5: NiCo-WC	31.5	33.6					283.7	529.9	853.2	855.7	860.5	778.6	780.8	786.3
6: NiCo-MoC			228.5	229.7	231.1	232.8	283.6	530.9	853.4	855.2	859.1	778.8	781.4	786.3
7: NiFe-WC	31.6	35.9					283.1	530.7	852.9	854.8	859.0				706.9	709.0	712.2
8: NiFe-MoC			229.0	230.3	231.6	232.5	283.5	530.9	852.9	855.1	859.0				707.1	709.5	712.8

Figure 4

High-resolution XPS scans of the C 1s (of 7), Mo 3d (of 8), W 4f (of 7), Ni 2p (of 7), Co 2p (of 2), and Fe 2p (of 8) areas showing the simulated components. The red line indicates the accumulative simulation of all of the fitted components.

Figure compiles as representative examples the C 1s, W 4f, Ni 2p, and Fe 2p high-resolution spectra of 7 as well as the Co 2p area of 2 and the Mo 3d area of 8 (the XPS of all of the metals of 1, 2, and 5–8 are presented in the Supporting Information), while the data extracted from the XPS are reported in Tables and 3.

Table 3

Atomic Ratios (as Estimated from the XPS Data) between Different Metals of the Prepared TMCs 1–8

	W				Ni			Co			Fe			Mo
	Wtot	W0	WC	WO	Nitot	Ni0	NiO	Cotot	Co0	CoO	Fetot	Fe0	FeO	Motot	MoC	MoO
1: Ni-WC	1.0		0.82	0.18	1.39	0.80	0.59
2: Co-WC	1.0	0.10	0.53	0.37				1.36	0.77	0.59
5: NiCo-WC	1.0		0.57	0.43	1.11	0.59	0.51	1.74	1.24	0.50
6: NiCo-MoC					1.04	0.53	0.51	1.60	1.19	0.41				1.0	0.41	0.59
7: NiFe-WC	1.0		0.88	0.12	1.05	0.69	0.36				0.86	0.49	0.37
8: NiFe-MoC					0.60	0.40	0.20				1.04	0.87	0.16	1.0	0.70	0.30

Two sets of photoelectron lines were observed for the W 4f spectra of all of the W-containing TMCs (1, 5, and 7). The 4f7/2-4f5/2 doublets displayed a spin–orbit splitting ΔBE ≈ 2.2 eV. The photoelectron lines at ca. 31.5 and 35.2 eV were assigned to the W 4f7/2 line of WC and WO3, respectively. These binding energy assigned to the WC photoelectron lines were in accordance with the range values of 31.5–31.8 eV reported for the WC1– phase in the literature.[29−31] Partial surface oxidation was revealed by the presence of a 4f7/2-4f5/2 doublet between 35 and 38 eV corresponding to WO3.[32]

Molybdenum is present in a mixed-valence state for the MoC TMCs (6 and 8). Four different valence states, namely, 2+, 3+, 4+, and 6+, were detected in the XPS Mo 3d spectra (see Figure and Table ). The binding energy of the Mo 2d5/2 photoelectron lines of the higher-valence states Mo4+ and Mo6+ present at 231.4 and 232.6 eV (with a ca. ΔBE ≈ 3.1 eV), respectively, were assigned to the partial oxidation (MoO2 and MoO3) of the MoC1– surface.[33−35] The lower valence states Mo2+ and Mo3+, which are associated with the carbide, revealed Mo 2d5/2 photoelectron lines at ca. 228.7 and 230.0 eV, respectively, in correlation with that reported for MoC1– in the literature.[33,34,36,37]

Ni is present in two different valence states (0 and 2+) in 1 and 5–8, as evident from its Ni 2p XPS (see Figure and Table ). The Ni 2p3/2 simulated photoelectron lines at ca. 852.9 eV are characteristic of metallic Ni0. The lines at ca. 855.2 eV and its associated satellite structures situated at 859.6 eV are assigned to Ni2+, indicating that Ni was partly oxidized.[38] The degree of oxidation of Ni was found to be ca. 7% higher when doped in Mo than when doped in W (6 and 8vs5 and 7). Also, when the codopant was Co (5 and 6), a higher degree of oxidation occurred for Ni as compared to when Fe was used as a codopant (7 and 8).

The Co 2p envelope (of 2, 5, and 6) was deconvoluted to fit three sets of Co 2p3/2–Co 2p1/2 doublets. The two simulated photoelectron lines at ca. 778.7 and 793.7 eV corresponded to Co 2p3/2 and Co 2p1/2, respectively, of metallic Co0. The Co 2p3/2 photoelectron line positions at ca. 781.0 and 786.3 eV were assigned to the main and satellite structure, respectively, of Co2+ and/or Co3+ associated with oxidized cobalt.[39]

Mixed valencies were also observed for the iron in 7 and 8. According to the simulated fitting of the photoelectron lines, ca. 35% of the iron was present as metallic Fe0. These Fe0 lines presented as sharp well-defined peaks at ca. 707 eV with a full width at half-maximum (FWHM) of ca. 1.2 eV. Further deconvolution of the Fe 2p envelope indicated the main photoelectron line (ca. 709.3 eV) and satellite structures (712.5 eV) for Fe2+, which is associated with FeO.[40]

Although not normally used as a quantitative technique, XPS accurately detected the ratios between atomic %,[41−48] thus giving a good indication (estimation) of the composition. A summary of the atomic ratios between the metals as well as the ratios between different compounds (e.g., WC and WO) is given in Table . Using these atomic ratios obtained between the metals, as well as the relative % of metal species (as defined by their oxidation state and binding energy position) from the XPS, it is possible to estimate the stoichiometric compositions of 1–8:

1: Ni0.8(NiO)0.6(WC)0.8(WO3)0.2

2: Co0.6(CoO)0.8W0.1(WC)0.5(WO3)0.4

5: Ni0.6(NiO)0.5 Co0.5(CoO)1.2(WC)0.6(WO2)0.4

6: Ni0.5(NiO)0.5 Co0.4(CoO)1.2(MoC1–)0.4(MoO2/3)0.6

7: Ni0.7(NiO)0.4Fe0.4(FeO)0.5(WC)0.9(WO2)0.1

8: Ni0.4(NiO)0.2Fe0.2(FeO)0.9(MoC1–)0.7(MoO2/3)0.3

Time-of-Flight Secondary Ion Mass Spectrometry Analysis

The mass spectra of both the positive and negative secondary ions were recorded for the trimetallic carbides 5–8.

The false color overlay ToF-SIMS images (both positive and negative modes) of the different elements of the trimetallic carbides 5–8 are shown in Figure . The emitted secondary ions captured by the detector are within the area of 100 × 100 μm2. An image with high intensity means high ionic concentration, while a low color intensity means low ionic concentration. From the ToF-SIMS images in Figure , the ions of the dopants are uniformly distributed in the metal carbide particles, indicating a homogeneous scattering of all of the different transition metals, with no agglomeration.

Figure 5

ToF-SIMS images collected from (left column) the positive-ion mode and (right column) the negative-ion mode of 5–8.

Photocatalytic Decomposition of Congo Red

The photocatalytic activity of TMCs (1–8) was compared using the model decomposition of the Congo red under simulated sunlight irradiation. Congo red is a toxic azo dye that is very difficult to decompose because of its stable aromatic structure.[12]

For a comparison of the removal efficiency of various TMCs (1–8), Congo red dye decomposition was executed in the dark (dark absorption, catalyst but no light), photolysis (no catalyst but in light), and photocatalytic conditions (in the presence of both catalyst and light). The photocatalytic decomposition process was performed under the same conditions for all of the TMCs: initial Congo red concentration = 20 mg dm–3, catalyst loading = 50 mg, solution pH = 7.0, irradiation = 350 W metal halide lamp, and λmax of Congo red = 495 nm.

Figure shows the UV–vis absorption spectra of the Congo red solution over time during photolysis (in the presence of light and no catalyst) and (b) dark absorption (in the presence of a catalyst but in the absence of light) and photocatalytic decomposition (in the presence of light and a catalyst). Ni-WC (1) is presented as a representative example (graphs of 2–8 are presented in the Supporting Information). Negligible degradation was observed in both the dark absorption and photolysis, as can be seen from the lack of decrease in the peak intensity in Figure a,b, indicating that both light and photocatalyst are needed for the effective decomposition of dye molecules in solution. The decomposition of the dye’s molecules in Figure c is confirmed by the decrease in the intensity of the bands in relation to time. The absorption peaks gradually decrease without a change in the wavelength of λmax before disappearing. This is in agreement with a report by Dantas et al., who analyzed the photocatalytic degradation of the Maxilon Blue GRL 300 textile dye in the presence of the Ni-Mo2C catalyst at different time intervals. The absorbance results show a decrease in the peak intensity over time, implying a degradation of dye molecules in solution.[2]

Figure 6

UV–vis absorption spectra of the Congo red solution during (a) photolysis, (b) dark absorption (with Ni/WC, 1), and (c) photocatalytic decomposition (time 0–45 min) with Ni-WC as an example (graphs of 2–8 are presented in the Supporting Information). (c) Photocatalytic decolorization graph (of 1) showing concentration (C/C0) vs time measured at λ = 495 nm. (d) The kinetics of photocatalytic degradation of Congo red under different conditions: in the dark with catalyst (green square), without catalyst under irradiation (blue dot), and with catalyst under irradiation (magenta dot).

The photocatalytic decolorization is calculated from the dye’s concentration (C/C0) over time (see Figure d and graphs in the Supporting Information). C0 is the initial concentration of the dye solution and C is the concentration at time t (min).

The photocatalytic decolorization fraction (C/C0) of the TMCs decrease drastically compared to dark absorption and photolysis, affirming that visible light irradiation and a photocatalyst in this case TMCs are needed for the decomposition of Congo red.

To compare the activity of the photocatalytic properties of 1–8, the % decolorization of Congo red was determined after 25 min reaction time (see Table and Figure ). NiCo-WC (97.1%) displayed the highest decolorization % after 25 min, closely followed by Ni-MoC, with 94% decolorization after 25 min. This was much better when compared to the that of reported for Mo2C by Dantas et al., who showed the maximum decolorization of 90.5% for Congo red after 1 h. From the data in Figure , it could be concluded that the Ni-doped carbide derivatives (NiCo-WC, Ni-MoC, and Ni-WC) were the most active photocatalysts. While the Co- and Fe-doped carbides result in the least decolorization after 25 min.

Figure 7

Left: % Decolorization after 25 min of photocatalytic reaction time. Right: Comparison between the optical band gap energy of the TMCs and the % decolorization of Congo Red after 25 min photocatalytic reaction in the presence of TMCs.

It is also reported that when Mo2C is doped with Co the % decolorization decreased from 86.7% for neat Mo2C to 83.7% for the 10% Co-doped Mo2C at a pH of 9. This affirms our results that doping with Co results in a decrease in photocatalytic activity.

A rough general trend favors an increase in band gap energy (Table ), which is associated with an increase in the decolorization activity. For materials having large band energies, the recombination rate of the electron–hole pair is reduced, causing the pair to have a longer time to interact with surface molecules, in turn resulting in better photocatalytic activity (faster decolorization).[49,50]

The graph comparing the optical band gap energy of the TMCs against the % decolorization of Congo red after 25 min of photocatalytic reaction in the presence of the TMCs is shown in Figure . The data points on this graph were grouped into three sets: TMCs containing WC (1, 2, and 7), TMCs containing MoC (3, 4, and 8), and TMCs doped with both Ni and Co (5 and 6). For the WC (1: BGE = 2.77, 86% decolorization; 2: BGE = 2.5, 66% decolorization; and 7: BGE = 1.97, 44% decolorization) and MoC (3: BGE = 2.00, 94% decolorization; 4: BGE = 1.83, 72% decolorization; and 8: BGE = 1.66, 41% decolorization) containing TMCs, the % decolorization of the Congo red increases as the band gap energy increases during the photocatalytic reaction and correspondingly the photocatalytic activity of the TMC. This increase in the photocatalytic activity associated with an increase in the band gap energy is more pronounced for the WC-containing TMCs than for the MoC containing TMCs (as can be seen by the slope of the graphs). For the NiCo-containing TMCs (5: BGE = 2.3, 97% decolorization; and 6: BGE = 2.34, 72% decolorization), change in band gap energy seems to have a negligible influence on the photocatalytic activity of the TMCs.

No relationship could be established between the amount of oxidation of the carbides or the doping metal or the amount of doping metals added.

Conclusions

Various transition-metal carbides (Ni-WC, 1, Co-WC, 2, Ni-MoC, 3, Co-MoC, 4, NiCo-WC, 5, NiCo-MoC, 6, NiFe-WC, 7, and NiFe-MoC, 8) were prepared by the carbothermal reduction of the parent metal oxides. It was found that the crystallite sizes were dependent on the dopant and metal carbide. The Co-doped TMCs displayed smaller crystallite sizes for the bimetallic TMCs (2 and 4vs1 and 3) and trimetallic TMCs (5 and 7vs6 and 8), while the WC trimetallic TMCs (5 and 6vs7 and 8) displayed smaller crystallites compared to MoC. From the SEM images, the TMCs (1–7) showed agglomeration of submicrometer globular particles into larger particles, the 4–34 nm crystallites (as determined by XRD) can be observed on the surface of the large particles.

Partial oxidation of all of the metals in the TMCs 1–8 occurred as was evident from the XPS, which showed the presence of the metal oxides along with the carbide and metallic photoelectron lines. Additionally, it was found that the ions detected by ToF-SIMS were not distributed evenly. Despite the oxidation that occurred and the uneven distribution of the ion in the TMCs, good photocatalytic activity was obtained for the decomposition of Congo red. This photocatalytic activity is highly dependent on the type of carbide, metal dopant, and the band gap energy of the material. WC-containing TMCs with higher band gap energies (Ni-WC, 1, band gap energy = 2.77 eV; NiCo-WC, 5, band gap energy = 2.30 eV) gave the best overall decolorization after 25 min (Ni-WC, 1, % decolorization after 25 min = 86%; NiCo-WC, 5, % decolorization after 25 min = 97%) and accordingly the best photocatalytic activity. However, Ni-MoC, 3, displayed a high band gap energy of 2.00 eV and resulted in very good photocatalytic activity with 94% decolorization after 25 min. The Fe-doped carbides NiFe-WC, 7, and NiFe-MoC, 8, displayed a poor photocatalytic activity of 44 and a decolorization of 41% after 25 min.