Effect of Adding Chelating Ligands on the Catalytic Performance of Rh-Promoted MoS₂ in the Hydrodesulfurization of Dibenzothiophene

Abstract

Hydrodesulfurization (HDS) is a widely used process currently employed in petroleum refineries to eliminate organosulfur compounds in fuels. The current hydrotreating process struggles to remove organosulfur compounds with a steric hindrance due to the electronic nature of the current catalysts employed. In this work, the effects of adding chelating ligands such as ethylenediaminetetraacetic acid (EDTA), citric acid (CA) and acetic acid (AA) to rhodium (Rh) and active molybdenum (Mo) species for dibenzothiophene (DBT) HDS catalytic activity was evaluated. HDS activities followed the order of RhMo/ɣ-Al₂O₃ (88%) > RhMo-AA/ɣ-Al₂O₃ (73%) > RhMo-CA/ɣ-Al₂O₃ (72%) > RhMo-EDTA/ɣ-Al₂O₃ (68%). The observed trend was attributed to the different chelating ligands with varying electronic properties, thus influencing the metal–support interaction and the favorable reduction of the Mo species. RhMo/ɣ-Al₂O₃ offered the highest HDS activity due to its (i) lower metal–support interaction energy, as observed from the RhMo/ɣ-Al₂O₃ band gap of 3.779 eV and the slight shift toward the lower BE of Mo 3d, (ii) increased Mo-O-Mo species (N_Mo-O-Mo ~1.975) and (iii) better sulfidation of Rh and MoO in RhMo/ɣ-Al₂O₃ compared to the chelated catalysts. The obtained data provides that HDS catalytic activity was mainly driven by the structural nature of the RhMo-based catalyst, which influences the formation of more active sites that can enhance the HDS activity.

1. Introduction

Deleterious refractory organosulfur compounds in fuel oils have contributed to SOx emissions [1,2]. Therefore, it has become very important to remove these compounds due to the introduction of the strict environmental regulations of the Euro V limits of 10 ppmS in diesel fuels [1,2].

Hydrodesulfurization (HDS) is the most commonly used technology to produce clean fuels by employing hydrotreating catalysts, mainly from Co(Ni)/Mo oxides supported on alumina [3,4,5,6,7,8,9,10,11]. The production of ultra-low fuels using the HDS process at present requires extreme and expensive operating conditions, viz., high temperatures, hydrogen and highly active catalysts.

However, other methods of fuel desulfurization have been reported, and these techniques are bio-desulfurization (BDS), oxidative-desulfurization (ODS), adsorptive and extractive desulfurization (ADS and EDS) [1,2]. Bio-desulfurization (BDS), which involves the use of sulfur-consuming bacteria to desulfurize fuels, is nonetheless limited in meeting very deep desulfurization, as 50–200 ppmS have been reported [1]. The ODS process of recalcitrant sulfur-containing compounds such as dibenzothiophene have been reported to produce nonessential side products during oxidation [2]. Adsorptive and extractive desulfurization (ADS and EDS) is a process which employs the use of use of solid sorbents and extraction solvents for removing the recalcitrant sulfur-containing compounds, respectively. Adsorbents have been limited in organosulfur compound selectivity, as well as low adsorption capacity, while the organosulfur extraction solvents such as dimethylsulfoxide (DMSO) and N,N-dimethylformamide (DMF) have high boiling points, and hence the solvent recovery may be impossible [1,2]. The use of ODS, BDS, ADS and EDS desulfurization techniques are currently limited, as they may pose problems for larger scale applications [1,2]. As a result, more attention is required to improve the HDS system by redesigning the current HDS catalysts’ inadequacies and by introducing chelating ligands and precious metals (PGMs), which could drive HDS under mild conditions [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26].

Several transition metal sulfides have been reported as possible candidates for the HDS catalysts [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]. Rhodium-based catalysts have exhibited promising properties, hence showing great potential as an HDS catalyst. A few HDS studies of dibenzothiophene (DBT) over Rh-based catalysts have been reported. Lee et al. [27] reported that the RhCs/Al₂O₃ catalyst was more active than a conventional CoMo/Al₂O₃ catalyst for hydrotreating dibenzothiophene. RhCs/Al₂O₃-catalyzed DBT hydrodesulfurization was mainly controlled by the DDS (direct desulfurization) mechanism. Similarly, a synergetic effect was reported with the RhMo/ɣ-Al₂O₃ catalysts, suggesting that Rh and Mo interact when employed for DBT HDS [28]. While there are reports on the use of RhMo-based catalysts [27,28], studies with regards to the influence of chelating ligands, especially acetic acid (AA), on hydrotreating activity is lacking. Chelating ligands are molecules with two or more donor atoms available to bind a metal cation, and they have been reported to improve hydrotreating activity [15,17,29,30,31].

The scientific novelty of the research is to arrive at a fundamental understanding of the nature of the sulfur tolerance of the supported Rh-Mo catalysts (chelated and unchelated RhMoS/ɣ-Al₂O₃), and to offer clarification of the synergetic effect of chelating ligands such as ethylenediaminetetraacetic acid (EDTA), citric acid (CA) and acetic acid (AA) on the individual metal, the Rh and the Mo components in the catalytic hydrodesulfurization of DBT. The originality of the research was in the design and the application of the nanostructured RhMo catalysts in the presence of the chelating ligands so to ensure the uniform composition of the catalytic metal species for ease of physico-chemical characterization and the fundamental understanding of the structure-activity relationships. In this study, the HDS catalytic activity of the RhMo-based catalysts were carried out on dibenzothiophene (DBT). The as-synthesized catalysts were characterized using powder X-ray diffraction (PXRD), ultraviolet-visible diffuse reflectance spectroscopy (UV-vis DRS), Fourier transform infrared spectroscopy (FT-IR), scanning electron microscope (SEM), energy-dispersive X-ray spectroscopy (EDX), X-ray photoelectron spectroscopy (XPS), transmission electron microscopes (TEM) and thermogravimetric analysis (TGA)—differential scanning calorimetry (DSC) (TGA-DSC) to determine the catalysts’ bulk chemical compositions, morphologies and thermal stability.

The main findings are as follows:

(1) The prepared catalysts presented band gaps of 3.779 eV (RhMo/ɣ-Al₂O₃), 4.341 eV (RhMo-EDTA/ɣ-Al₂O₃), 4.394 eV (RhMo-AA/ɣ-Al₂O₃) and 4.478 eV (RhMo-CA/ɣ-Al₂O₃), respectively.

(2) The introduction of different chelating ligands increases the metal–support interaction, which prevents the formation of easily reduced Mo species.

(3) The HDS activity decreased in the following order: RhMo/ɣ-Al₂O₃ (88%) > RhMo-AA/ɣ-Al₂O₃ (73%) > RhMo-CA/ɣ-Al₂O₃ (72%) > RhMo-EDTA/ɣ-Al₂O₃ (68%).

(4) The observed catalytic results were ascribed to the introduction of different ligands, thus increasing the metal–support interaction and increasing the e-charge transfer from the valance band Rh 4d orbital to the conduction band of the Mo species. This led to the excessive weakening of the Mo-S bond by inhibiting the absorption of sulfur (S) compound (DBT) on the active sites, hence leading to a reduced activity.

2. Results and Discussion

2.1. UV-Vis Spectroscopy

UV-vis spectroscopy was applied to study RhMo/γ-Al₂O₃, RhMo-EDTA/γ-Al₂O₃ and RhMo-CA/γ-Al₂O₃, and are shown in Figure 1. As shown in Figure 1, the UV-vis spectra recorded for RhMo/γ-Al₂O₃ and RhMo(x)/γ-Al₂O₃ (x = EDTA, CA and AA) exhibited a broad absorption band at 210–290 nm and were assigned to the O²⁻→Mo⁶⁺ ligand–metal charge transfer transitions in an octahedral environment [32]. A weak band observed in the region of 340–360 nm with RhMo/γ-Al₂O₃ was associated with Rh(III) oxides, and another weak broad band at a visible region displayed at 452 nm was due to the presence of Rh(III) in oxide form [33]. A weak absorbance at 530 nm provided strong evidence of well-dispersed octahedral Rh oxide species [34,35], which are known to be easily reduced and sulfided [36,37,38]. A band around 550–680 nm on RhMo/γ-Al₂O₃ was due to Rh³⁺ interacting with γ-Al₂O₃ support to form a rhodium aluminate complex [32]. With the addition of the ligands, the absorbance of Rh³⁺ shifts, as do the weak broad bands around 430 nm, 410–470 nm and 420 nm associated with the metal to ligand transfer, and Rh-O species for RhMo-AA/γ-Al₂O₃, RhMo-EDTA/γ-Al₂O₃ and RhMo-CA/γ-Al₂O₃ were observed. The observed shifts to the lower wavelength suggested that the Rh–ligand complexes inhibit the formation of the Rh-γ-Al₂O₃ phase [32,39]. The shift to a lower wavelength also indicated a decreased agglomeration of the Mo species [32,40,41].

2.2. Band Gaps of RhMo Catalysts

The E_g value obtained from the Tauc and Davis–Mott Equation (1) demonstrates the dispersion of the Mo species. The band gaps for RhMo/γ-Al₂O₃, RhMo-EDTA/γ-Al₂O₃, RhMo-CA/γ-Al₂O₃ and RhMo-AA/Al₂O₃ were determined and are displayed in Figure S1. According to the literature, the higher the E_g value, the more improved the dispersion of the Mo species [42]. The obtained band gap for RhMo/Al₂O₃ was 3.779 eV (Figure S1), the E_g value obtained for RhMo-EDTA/γ-Al₂O₃ was 4.341 eV, the E_g value RhMo-AA/γ-Al₂O₃ was E_g = 4.394 eV and the E_g value for RhMo-CA/γ-Al₂O₃ was 4.478 eV. The RhMo-CA/γ-Al₂O₃ catalyst exhibited the highest E_g value, which implies a decrease in the average particle size and an increased charge transfer [41,42,43,44].

$${(\mathsf{\alpha }\mathrm{h}\mathsf{\nu })}^{\frac{1}{\mathrm{n}}}=\mathrm{A} \left(\mathrm{h}\mathsf{\nu }-{\mathrm{E}}_{\mathrm{g}}\right)$$

(1)

where α is the absorption coefficient, hν is the incident photon energy, A is the proportionality constant, E_g is the optical band gap energy and n represents the nature of the electronic transition (n = 1/2 for direct transition).

The bridging Mo-O-Mo bonds, which determine the degree of polymerization/aggregation of Mo(VI), were determined from the E_g values by using the formula (N_Mo-O-Mo = 11.8 − 26E_g) presented by Tian et al. (2010) [42]. It was established that the higher E_g values of the catalysts corresponded to the lower average number (of covalent bridging of the central Mo⁶⁺ cation) nearest to the Mo⁶⁺ neighbours (

Table 1. Band gap energy and average number of nearest Mo⁶⁺ neighbors (N_Mo-O-Mo) in deposited clusters, as determined from UV spectra of the oxide RhMo catalysts.

Catalyst	Eg Values	NMo-O-Mo
RhMo/γ-Al2O3	3.779	1.975
RhMo-AA/γ-Al2O3	4.341	0.5134
RhMo-EDTA/γ-Al2O3	4.394	0.3756
RhMo-CA/γ-Al2O3	4.478	0.1572

N_Mo-O-Mo = 11.8–2.6E_g.

), thus confirming the Mo(VI) cation structural variations in the catalysts. RhMo/γ-Al₂O₃ was reported to offer more polymeric/aggregated Mo species compared to the chelated catalysts (RhMo-AA/γ-Al₂O₃, RhMo-EDTA/γ-Al₂O₃ and RhMo-CA/γ-Al₂O₃).

2.3. Fourier Transform Infrared Spectroscopy (FT-IR)

FT-IR bands at 3600–2800 cm⁻¹ correspond to the OH (from H₂O) stretching. At the region between 550 cm⁻¹, the corresponding band was assigned to the Mo-O-Mo bridge stretching, while the bands that were found, and the Mo=O stretching, were located at 950 cm⁻¹ [45]. RhMo-EDTA/Al₂O₃ showed an absorption band at 978 cm⁻¹, which could be assigned to the Mo-N band [35,45]. At 1588 cm⁻¹, the absorption band could be assigned to -COO⁻ vibrations with H₂O [46] (Figure S2).

2.4. Energy Dispersion Spectroscopy (EDX)

The EDS of the sulfided RhMo/Al₂O₃, RhMo-EDTA/Al₂O₃, RhMo-AA/Al₂O₃ and RhMo-CA/Al₂O₃ catalysts confirmed that the catalysts are made up of Rh, Mo, O, S, C and Al (Figure S3). The peaks at ~2.37, ~2.7, ~2.81 and ~3.28 KeV corresponded to the theoretical L_α, K_α and K_β of Rh, respectively. The peaks at ~2.3 and 2.81 KeV corresponded to the theoretical L_α, K_α and K_β of Mo, and the O peak was obtained at ~0.5 KeV and the S peak at ~2.3 KeV. The additional peak at ~1.5 KeV corresponded to Al from the support, and the presence of carbon (~0.28 KeV) was due to the carbon tape that was used for the sample analysis [47].

Table 2. Qualitative atomic percentage of Rh, C, O, S and Mo for sulfided HDS catalysts.

Catalysts	Atomic Percentage (wt. %)						S/Mo	C/Mo
	C K	O K	Al K	S K	Rh L	Mo L
RhMo/ɣ-Al2O3	7.97	63.24	26.47	0.68	0.17	1.46	0.47	5.46
RhMo-EDTA/ɣ-Al2O3	10.20	57.52	30.03	1.48	0.44	0.32	4.63	31.88
RhMo-AA/ɣ-Al2O3	9.38	62.61	25.63	1.35	0.19	0.73	1.85	12.85
RhMo-CA/ɣ-Al2O3	12.07	57.56	26.88	1.60	0.94	0.90	1.78	13.41

illustrates the qualitative atomic percentage of the present elements for each HDS catalyst. The S/Mo atomic ratio for the catalysts are around (0.47–4.63) and the amounts of carbon (5.46 ≤ C/Mo ≤ 31.88) are found in all of the MoS₂ catalysts. The main source of carbon is most probably the heptane solvent [48,49].

2.5. X-ray Diffraction (XRD)

The XRD analysis was performed to identify the diffraction phases and dispersion of the synthesized RhMo/γ-Al₂O₃ and RhMo-x/γ-Al₂O₃ (x = EDTA, acetic acid (AA), citric acid (CA)) catalysts. Figure 2 showed that all of the RhMo catalysts in the oxide phase had similar diffraction patterns at 2θ = 19.6°, 32.0°, 37.6°, 39.5°, 45.5°, 60.9° and 67.0°, and were assigned to (220), (311), (222), (400), (511) and (440), characteristic of the γ-Al₂O₃ face-centered cubic phase, respectively. For the RhMo/ɣ-Al₂O₃ catalyst, more patterns were observed at 2θ = 12.1°, 18.5° and 28.5°, which were due to the orthorhombic MoO₃ crystalline phase [50], and the diffraction pattern at 49.0° could be ascribed to the monoclinic crystalline phase of MoO₃. The pattern observed at 2θ = 34.1° could be attributed to Rh₂O₃ phase, and diffraction patterns at 56.0° and 57.8° were also observed. All of the chelated catalysts showed the characteristic reflections of alumina supports and very weak reflection peaks for RhMo-CA/ɣ-Al₂O₃, which indicated that the addition of the chelating agent could promote the redispersion of the bulk MoO₃ [40,51,52]. The broadness and the amorphous nature of the diffraction pattern observed in the chelated catalysts indicated the absence of crystalline MoO₃ (RhMo-CA/ɣ-Al₂O₃). Sulfided RhMo/ɣ-Al₂O₃ presented a hexagonal MoS₂ phase at 14.8°, 29.5°, 33.3°, 38.5° and 60.4°. According to the obtained result, it was shown that the addition of the chelating ligands resulted in the better dispersion of molybdenum oxide [53]. Additional peaks attributed to the rhodium sulfide phase (Rh₂S₃ and/or Rh₃S₄), with characteristic peaks at 2Ɵ = 36–42, were detected on all the sulfided RhMo-x/ɣ-Al₂O₃, with RhMo/ɣ-Al₂O₃ exhibiting more characteristic peaks compared to RhMo-x/ɣ-Al₂O₃. The broad diffraction peaks of the sulfided catalysts (Figure 2b–d) compared to RhMo/ɣ-Al₂O₃ (Figure 2a) showed bulk and relatively smaller crystallite sizes. The RhMo catalysts’ crystallite sizes decreased in the order of RhMo/ɣ-Al₂O₃ (5.903 nm) > RhMo-CA/ɣ-Al₂O₃ (5.809 nm) > RhMo-EDTA/ɣ-Al₂O₃ (5.770 nm) > RhMo-AA/ɣ-Al₂O₃ (5.750 nm).

2.6. XPS Analysis

The XPS survey spectrum of sulfided RhMo/ɣ-Al₂O₃ and RhMo-EDTA/ɣ-Al₂O₃ catalysts with the detected species, viz., S 2p, Rh 3d and Mo 3d, are presented in Figure 3 and Figure 4, respectively. The survey scan spectrum, shown in Figure 3a and Figure 4a, demonstrated the presence of the key elements, O 1s, S 2p, C 1s, Al 2p and 2s, Rh 3d and Mo 3d in the catalysts. The highly resolved measurements of these individual elements of O 1s, C 1s, Al 2p, and Al 2s are demonstrated in Figures S3a–d and S4a–d for the RhMo/ɣ-Al₂O₃ and RhMo-EDTA/ɣ-Al₂O₃ catalysts. The binding energies of these elements are shown in

Table 3. Binding energies determined in XPS experiments for RhMo/Al₂O₃ and RhMo-EDTA/ɣ-Al₂O₃.

Elements (eV)	RhMo/ɣ-Al2O3	RhMo-EDTA/ɣ-Al2O3
C 1s	289.5	285.5
O 1s	530.0	529.0
Mo 3d	226.2; 230.0; 233.2	226.0; 230.0
Rh 3d	306.5; 310.6	305.0; 310.1
S 2p	160.1; 167.0	160.5; 166.1
Al 2p	72.4	72.0
Al 2s	117.0	117.0

, and these energies signify the presence of the elements on the catalysts. The XPS result of the Rh 3d peaks for RhMo/ɣ-Al₂O₃ showed doublets at the binding energies of 306.5 and 310.6 eV, and for RhMo-EDTA/ɣ-Al₂O₃, the Rh 3d binding energies were obtained at 305.0 and 310.1 eV, respectively [54,55]. Mo 3d showed three characteristic peaks observed at 226, 230.0, and 233.2 eV for RhMo/ɣ-Al₂O₃ and two visible peaks at 226 and 230.0 eV for RhMo-EDTA/ɣ-Al₂O₃, respectively.

To investigate the different phases within the samples for Rh 3d and Mo 3d for the sulfided RhMo/ɣ-Al₂O₃ and RhMo-EDTA/ɣ-Al₂O₃, the spectra were carefully deconvoluted and the obtained results are presented in Figure 5a–d. The rhodium oxide (Rh₂O₃) showed characteristic peaks between 307–310 for Rh 3d_5/2, and for Rh 3d_3/2 showed characteristic peaks at a region between 312–315 eV [52,56], and the corresponding results are presented in Figure 5a,c. The Rh₂S₃ phase showed a doublet at the binding energies of 307–309.2 and 312.7–314 eV in the sulfided RhMo/ɣ-Al₂O₃ and RhMo-EDTA/ɣ-Al₂O₃ samples (Table S2), corresponding to the Rh 3d_5/2 and Rh 3d_3/2 states for rhodium sulfide, respectively [54,55,56]. The Mo 3d was comprised of three main peaks with oxidation states of +V (oxide), +V (oxysulfide) and +IV (sulfide) [57,58]. Figure 5b,d shows the deconvolution of the Mo 3d spectra. In the case of Mo, its deconvolution consisted of Mo⁴⁺ (228.4–229.1 eV, sulfide MoS₂), Mo⁵⁺ (229.7–230.5 eV, oxysulfide MoS_xO_y) and Mo⁶⁺ (232.1–232.7 eV, oxide MoO₃) [59].

The doublet at a binding energy (BE) of (±) 229 eV and (±) 232.1 0.1 eV was attributed to Mo 3d_5/2 and Mo 3d_3/2 levels of MoS₂ (Mo⁴⁺), and the two contributions observed at (±) 230.0 eV and (±) 235.3 0.1 eV were assigned to Mo 3d_5/2 and Mo 3d_3/2 of Mo oxysulfide (MoO_xS_y, Mo⁵⁺) [59], while the binding energy of the Mo 3d_5/2 component located at (±) 232.2 0.1 eV was assigned to the Mo⁶⁺ (MoO_x) species [60], and for the Mo 3d_3/2 energy level, the binding energy was (±) 236.0 eV, respectively [61]. The peak presented at 226.3 and 226.1 eV was ascribed to the S 2s level of sulfur (Figure 5e,f). The analysis results, including the detailed binding energies and the sulfidation degree of the Mo species obtained by the deconvolution, are shown in Table S1, and the different phase compositions of the catalysts are calculated from the area of the deconvoluted peaks. Figure 5e,f displays a contribution at 162.1 eV at the lower binding energy of the S 2p peak originating from the S²⁻ precursor [62], and it was clearly visible in both catalysts. The S 2p region suggests the existence of S²⁻, S₂²⁻ and SO₄²⁻ species [63,64]. The peaks at 162.3 and 163.5 eV are assigned to the S²⁻ in the 2p_3/2 and 2p_1/2 levels, respectively, in MoS₂ [40,51,63,65], and the characteristic peak in S 2p at 166.7 eV was attributed to SO₄²⁻ [66]. The sulfidation degree of the Mo species was calculated by the following Formula (2):

$$\left[{\mathrm{Mo}}^{4+}\right](\%)=\frac{{\mathrm{A}}_{{\mathrm{Mo}}^{4+}}}{{\mathrm{A}}_{{\mathrm{Mo}}^{4+}}+{\mathrm{A}}_{{\mathrm{Mo}}^{5+}}+{\mathrm{A}}_{{\mathrm{Mo}}^{6+}}}\times 100\%$$

(2)

where [Mo⁴⁺] is the sulfidation degree of the samples, and A_Mo⁴⁺, A_Mo⁵⁺ and A_Mo⁶⁺ are the areas of the peaks which are assigned to the Mo⁴⁺, Mo⁵⁺ and Mo⁶⁺ species, respectively [67,68].

A catalyst with a higher sulfidation degree (MoS₂) would suggest that there was a lower metal–support interaction with the active metal, bringing about easier catalyst reduction and sulfidation [67]. The slight shift toward the lower BE of Mo 3d in the RhMo/ɣ-Al₂O₃ catalyst (Table S1) could be attributed to a weaker metal–support interaction (caused by electron effects of the defects at the surface on the alumina support), which enhances the HDS catalytic activity [68,69].

2.7. Transmission Electron Microscopy (TEM)

Information on active MoS₂ crystallite dispersion and sizes in the sulfided RhMo/ɣ-Al₂O₃ and RhMo-x/ɣ-Al₂O₃ (where x = AA, EDTA, CA) was obtained by means of TEM measurement. Figure 6a–d shows the distribution of MoS₂ crystallites in the sulfided catalysts with and without the chelating agents, and their statistical distribution results for the length of the MoS₂ slab. The addition of the chelating agents influenced the particle size distribution. The RhMo-CA/ɣ-Al₂O₃ had the lowest average diameter (1.86 nm) and the RhMo/ɣ-Al₂O₃ resulted in the highest average diameter (4.72 nm). The decrease of the MoS₂ slabs due to the chelating ligand’s decomposition improved the dispersion of the active phase [70,71]. The average slab length in diameter observed for the catalysts are shown in

Table 4. Average length of MoS₂ crystallites in RhMo/Al₂O₃ and RhMo-x/ɣ-Al₂O₃ (x = EDTA, AA, CA).

Catalysts	Average Diameter ± SD (nm)
RhMo/ɣ-Al2O3	4.4 (±1.38)
RhMo-EDTA/ɣ-Al2O3	4.1 (±1.220)
RhMo-AA/ɣ-Al2O3	3.3 (±0.757)
RhMo-CA/ɣ-Al2O3	1.6 (±0.860)

, and the average slab length of the MoS₂ slab with the highest frequency was distributed between 2.5–6.0 nm. Along with these crystallites, there were a few regions with big agglomerations of molybdenum sulfide, as confirmed by the strings of highly stacked crystallites.

The results for the MoS₂ slab average diameter in Table 4 indicated that the MoS₂ dispersion obtained for the RhMo catalysts decreased in the order of RhMo-CA/ɣ-Al₂O₃ > RhMo-AA/ɣ-Al₂O₃ > RhMo/ɣ-Al₂O₃ > RhMo-EDTA/ɣ-Al₂O₃. The higher dispersion for the chelated catalysts was due to the complexation of metal-chelating ligand, reducing the metal–support interaction and leading to the delay of the sulfidation of the metals. The high MoS₂ dispersion could facilitate the generation of more active sites [72,73,74].

2.8. Scanning Electron Microscopy (SEM)

The SEM images of RhMoO_x are shown in Figure 7a,c,e,g for RhMo/ɣ-Al₂O₃ and RhMo-x/ɣ-Al₂O₃ (x = EDTA, AA, CA). The images showed that particles are closely spherical in shape with an average uniform distribution. All the samples had an average particle size (82.5–102.6 μm). Figure 7b,d,f,h represents the sulfided RhMo/ɣ-Al₂O₃ and RhMo-x/ɣ-Al₂O₃ (x = EDTA, AA, CA) catalysts, and the zoomed images show that all the chelated catalysts highly agglomerated with spherical-like materials with fluffy-like particles, which could indicate that the catalysts are porous in nature. The particle distributions for the sulfided catalysts could not be measured due to the high agglomeration of the particles.

2.9. Stability of Catalysts—TGA and DSC Thermal Analyses

TGA-DSC is a technique used for thermal analysis to characterize materials by measuring their change in mass as a function of temperature. It is coupled with DSC to provide complementary information such as measuring the heat flow as a function of time and temperature at a controlled environment.

RhMo/ɣ-Al₂O₃: The first weight loss (2.5%) for RhMo/ɣ-Al₂O₃ occurred in the range of 50–150 °C due to the desorption of the physically adsorbed water from the surface of the catalyst, and this was accompanied by a broad exothermic peak in the range of 120–210 °C (Figure 8a). A second weight loss of 3% was observed between 200–580 °C, and it was reflected by a very weak endothermic peak between 360–610 °C, associated with the decomposition of nitrate radical, hexaammonium molybdate and dihydroxylation [75]. An endothermic peak at 810 °C was observed, and it was attributed to the formation of a stable MoO₃ phase.

RhMo-EDTA/ɣ-Al₂O₃: The first weight loss (5.2%) in Figure 8b was below 100 °C and was mainly due to water desorption. From 160–800 °C, there was a gradual weight loss of 3% and not many events were happening in those stages. This weight loss was associated with the loss of the complex, and a partial dehydration–decomposition of the Rh and Mo species of RhMo-EDTA/ɣ-Al₂O₃. The first and second weight losses were accompanied by a broad exothermic peak at a maximum of 180 °C, corresponding to the decomposition of the complex. A broad exothermic peak between 280–670 °C was observed, and was due to the decomposition of EDTA and the further combustion of the residual organic matrix [76], and to the total transformation of the partially decomposed Rh and Mo precursor species into the catalyst oxidic precursor [77].

RhMo-AA/ɣ-Al₂O₃: The results obtained for RhMo-AA/ɣ-Al₂O₃ (Figure 8c) showed a weight loss (3.6%) taking place in the range of 50–150 °C followed by an exothermic peak 90 °C, which was mainly due to water desorption. The subsequent weight loss of 2.8% between 180–410 °C was attributed to the dehydration–decomposition of precursor species and the partial dehydroxylation of alumina [77]. A second exothermic peak occurred at a maximum of 280 °C, which corresponded to the decomposition of the complex (metal–AA) and the total decomposition and partial dehydration–decomposition of Rh and Mo precursor species. A third weight loss (1.8%) between 410–850 °C was due to the formation of the monometallic oxidic precursor. The DCS curve displayed an endothermic peak at a maximum of 580 °C.

RhMo-CA/ɣ-Al₂O₃: The results obtained for RhMo-CA/Al₂O₃ (Figure 8d) showed a weight loss of 5.2% below 150 °C, followed by an endothermic peak of 80 °C, was mainly due to H₂O removal. The second weight loss of 2.8% between 180–400 °C was attributed to a decomposition and combustion of the precursor species, and the complete breakdown of citric acid [46]. A broad and weak exothermic peak occurred at a maximum of 400 °C, which corresponded to the decomposition of the remaining complex (metal–CA) and the total decomposition and partial dehydration–decomposition of Rh and Mo precursor species. Above 400 °C, not much loss of weight loss was observed, and this indicated the formation of stable metallic oxidic precursors.

2.10. Catalytic Activity

The conversion of dibenzothiophene (DBT) was used to estimate the catalytic activity in HDS (Equation (1)). RhMo/ɣ-Al₂O₃ (88%) had the highest catalytic activity, and the activity for the chelated catalysts followed this order: RhMo-AA/ɣ-Al₂O₃ (73%) > RhMo-CA/ɣ-Al₂O₃ (72%) > RhMo-EDTA/ɣ-Al₂O₃ (68%) (

Table 5. Catalytic performances of RhMo/ɣ-Al₂O₃ and RhMo-x/ɣ-Al₂O₃ (x = EDTA, AA, CA) in hydrotreating of DBT as simulated fuel.

Catalysts	Crystallite Sizes (nm)	Eg Values	HDS (%)	BP(%)	PhCh(%)	HYD/DDS Ratio	TOF (h−1) a
RhMo/ɣ-Al2O3	5.903	3.779	88	65	13	0.20	51
RhMo-EDTA/ɣ-Al2O3	5.770	4.394	68	16	1	0.06	60
RhMo-AA/ɣ-Al2O3	5.750	4.341	73	65	3	0.05	79
RhMo-CA/ɣ-Al2O3	5.809	4.478	72	36	2	0.06	223

Catalyst (molybdenum content) employed = 0.1 g (4.119 × 10⁻⁵ moles). Hydrodesulfurization (HDS) time = 6 h; reaction temperature = 300 °C; reaction pressure = 40 bar. Phenylcyclohexane (PhCh) or biphenyl (BP) ^a TOF, h⁻¹: (turnover frequency).

). The observed catalytic results were ascribed to the introduction of the different ligands, which increased the metal–support interaction and increased the e-charge transfer (energy band gap) from the valance band Rh 4d orbital to the conduction band of the Mo species. The absence led to a weaker Mo-S bond strength, a higher concentration of CUS and a higher HDS activity [60]. Crystallite sizes were also observed to influence the catalytic activity, as RhMo/ɣ-Al₂O₃ (with crystallite size of 5.903 nm) presented the highest activity, and this may be due to the formation of bigger MoS₂ crystals when compared to others. A combined electron donating effect of the chelates, and the crystallite sizes of MoS₂, may have influenced the chelated catalyst activity.

The chelated RhMo/ɣ-Al₂O₃ catalysts resulted in slightly lower catalytic activity due to the formation of rhodium-chelating ligands and a molybdenum–chelate complex. For all the catalysts, DBT converted mainly via the DDS pathway (Table 5). There was not much difference obtained in terms of selectivity when comparing the HYD/DDS selectivity ratio. RhMo/ɣ-Al₂O₃ showed a slightly higher HYD/DDS ratio of ~0.20 when compared to the chelated catalysts (see GC chromatogram, Figures S5–S7). The addition of the chelating ligands showed a slight difference, and therefore we can conclude that the addition of the chelating ligand on the catalysts did not have much influence on the selectivity (Table 5).

The HDS selectivity correlated linearly with the slab length of the MoS₂ phase (TEM), the longer slab length indicated a high ratio of edge/corner and better HDS selectivity, with RhMo/ɣ-Al₂O₃ presenting the longest slab length [64,77], and the edge sites only catalysed the HDS reaction [78]. The values presented by the current RhMo catalysts (

Table 6. Comparison of catalyst performance with literature reports in DBT hydrodesulfurization.

Catalysts	Model Compound	Reaction Temperature (°C)	HDS (%)	Reaction Pressure (Bar)	Reference
RhMo/ɣ-Al2O3	DBT	300	88	40	This work
RhMo-EDTA/ɣ-Al2O3	DBT	300	68	40	This work
RhMo-AA/ɣ-Al2O3	DBT	300	73	40	This work
RhMo-CA/ɣ-Al2O3	DBT	300	72	40	This work
Ni2P	DBT	340	35	40	[79]
Ni2P	TH-DBT	340	50	40	[79]
NiMoP/γ-Al2O3	DBT	<320	22–90	<25	[80]
RuxMoNi	DBT	320	24–92	54.5	[81]
NiMo	DBT	320	62	54.5	[81]
NiMo/TiO2-6	DBT	300	90	20	[82]
NiMo/MCM-41-Na	DBT	300	>95	50	[83]
Fe-Zn/TiO2-Al2O3	DBT	380	>98	40	[84]
RhMo/ɣ-Al2O3	DBT	310	84	50	[28]

Dibenzothiophene (DBT); TH-DBT = 1,2,3,4-tetrahydro-dibenzothiophene (TH-DBT).

) exhibit certain benefits and compare well with the other catalysts reported in the literature [79,80,81,82,83,84]. We concluded that the HDS % conversion and desulfurization route was influenced by the catalyst composition, the electronic properties and the HDS reaction conditions.

2.11. Proposed Mechanism

The large energy gap (E_g) values of the chelated ligands RhMo-AA/ɣ-Al₂O₃ (4.341 eV), RhMo-EDTA/ɣ-Al₂O₃ (4.394 eV) and RhMo-CA/ɣ-Al₂O₃ (4.478 eV), supported the increased charge transfer of Rh, chelates and Mo species catalysts when compared to RhMo/ɣ-Al₂O₃ (3.779 eV). However, according to Figure 5a–d, the higher amount of Mo-S/RhMo-S phases are formed in the absence of a chelating ligand, owing to the electron transfer between the Rh and Mo-phase. Furthermore, the BE of O 1s increases upon chelation, confirming less neutralization of the surface O-H in ɣ-Al₂O₃ (O 1s 531.08 eV). The O 1s of RhMo/Al₂O₃ is 529.139 eV, and for RhMo-EDTA/Al₂O₃ is 529.280 eV (Figure 9).

Thus, this implies that ligand presence decreases the neutralization of the surface Brønsted acid site OH^δ+ of ɣ-Al₂O₃ (an observed increase in the oxygen binding energies), and this prevents the formation of Mo-S/RhMo-S bonds, but more Rh-S bonds are formed due to the higher charge transfer between Rh and the chelating ligands (see Figure 5c,d). Generally, the free electrons promoted catalytic activities by donating electrons to the conduction band of the Mo species, thus promoting Mo-S/RhMo-S bond cleavage to form more coordinatively unsaturated sites (CUS) (Scheme 1). In a typical HDS reaction, the Rh₂O₃ phase is reduced to metallic rhodium (BE of Rh 3d_5/2 at 307.0–307.1 eV, predominately observed in RhMo/ɣ-Al₂O₃ [55]) and Rh₂O₃-x(RhO), with unpaired free electrons in the Rh 4d orbital and oxygen vacancies transferred to the conduction band of Mo species, thereby promoting the cleavage of Mo-S/RhMo-S to form the CUS. The increased charge transfer (energy gap (E_g) values) in chelated RhMo/ɣ-Al₂O₃ between Rh and Mo species led to the excessive weakening of the Mo-S bond, preventing S-compound absorption on the active sites, thereby leading to reduced activity [84].

3. Experimental Section

3.1. Materials

All chemicals used were obtained from Merck/Sigma-Aldrich, South Africa. These include rhodium(III) chloride (98%), ammonium heptamolybdate (99%), ethylenediaminetetraacetic acid (EDTA, 97%), citric acid monohydrate (CA, 99.5%), acetic acid (AA, 99%), heptane, dibenzothiophene (98%) and gamma alumina support (ɣ-Al₂O₃).

3.2. Synthesis of RhMo Catalysts Prepared with Ethylenediaminetetraacetic Acid (EDTA), Citric Acid (CA) and Acetic Acid (AA)

Rh(x)Mo(y) catalysts were prepared by wet impregnations of the precursor salts [28]. Unchelated catalyst: Rh from RhCl₃ (0.0421 g, 2 × 10⁻⁴ mol) and Mo from (NH₄)₆Mo₇O₂₄·4H₂O (0.496 g, 4 × 10⁻⁴ mol) were added in 30 mL deionized water to obtain the desired metal content ratio (Rh/Rh + Mo) 0.3, and the pH adjusted to pH = 9. The solution was added to γ-alumina (1 g), and the resulting mixture was transferred to a 100 mL Teflon-lined stainless steel autoclave, and hydrothermally treated at 453 K for 4 h. The resulting solid was filtered, dried at 393 K for 12 h and calcined at 773 K for 4 h to obtain the RhMo oxide on γ-alumina.

Chelated catalysts: Generally, for the synthesis of chelated catalyst, molar ratios 1:2 Rh to chelates (EDTA, CA and AA) was employed. RhMo-AA/γ-Al₂O₃: A mixture of RhCl₃ (0.0421 g, 2 × 10⁻⁴ mol), (NH₄)₆Mo₇O₂₄·4H₂O (0.496 g, 4 × 10⁻⁴ mol) and AA (0.0241 g, 4 × 10⁻⁴ mol) was dissolved in 20 mL H₂O solution. RhMo-CA/γ-Al₂O₃: A mixture of RhCl₃ (0.0421 g, 2 × 10⁻⁴ mol), (NH₄)₆Mo₇O₂₄·4H₂O (0.496 g, 4 × 10⁻⁴ mol) and AA (0.0770 g, 4 × 10⁻⁴ mol) was dissolved in 20 mL H₂O solution. RhMo-EDTA/γ-Al₂O₃: A mixture of RhCl₃ (0.0421 g, 2 × 10⁻⁴ mol), (NH₄)₆Mo₇O₂₄·4H₂O (0.496 g, 4 × 10⁻⁴ mol) and EDTA (0.0745 g, 2 × 10⁻⁴ mol) was dissolved in 20 mL H₂O solution. In all of the solution mixtures, the molar ratio with Rh/Rh + Mo molar ratio of 0.3 was added to γ-Al₂O₃ (1 g, calcined at 500 °C) and the mixture was stirred for 4 h and pH adjusted to 9. The resulting solid was dried at 120 °C overnight to obtain RhMo-AA/γ-Al₂O₃, RhMo-CA/γ-Al₂O₃ and RhMo-EDTA/γ-Al₂O₃. For the chelated catalysts, RhMo-AA/γ-Al₂O₃, RhMo-CA/γ-Al₂O₃ and RhMo-EDTA/γ-Al₂O₃ were only treated at 120 °C to preserve chelating ligands (CA, AA and EDTA) in catalyst until the activation stage [46,50].

3.3. Catalyst Characterization

Ultraviolet-visible diffuse reflectance spectroscopy (DRS) and band gap energies of the catalysts were processed from a Shimadzu UV-vis DRS spectrophotometer UV-3100 UV-vis spectrophotometer from a wavelength range from 200 to 800 nm.

FT-IR spectroscopy of the catalysts was acquired using a Bruker Tensor 27 platinum ATR-FTIR spectrometer (wavelength range from 4000 to 400 cm⁻¹).

Thermogravimetric analysis (TGA-DSC) was measured using a Perkin Elmer STA 6000 at a heating range of 55 to 900 °C at 20 °C/min with N₂ flow of 30 mL min⁻¹.

X-ray powder diffraction (XRD) analysis was carried out on a Bruker D2 powder X-ray diffractometer using Cu radiation with a LynxEye detector with a scan range of 5 to 80° 2 theta.

Milled samples (RhMo/γ-Al₂O₃, RhMo-AA/γ-Al₂O₃, RhMo-CA/γ-Al₂O₃ and RhMo-EDTA/γ-Al₂O₃) were gold coated and imaged for morphological evaluation using a JOEL 7001f scanning electron microscope (SEM). JEOL JEM-2010 transmission electron microscope (TEM) operated at 200 kV was employed for TEM imaging.

X-ray photoelectron spectrometer (XPS) was performed on a Kratos Axis Ultra X-ray photoelectron spectrometer equipped with a monochromatic Al K_α source (1486.6 eV).

3.4. Catalyst Sulfidation and Hydrodesulfurization Measurements

The sulfidation and HDS tests were carried in a 2 L Parr pressure reactor 4842 (350 bar, max tem = 425 °C). For catalyst sulfidation: HDS catalysts (2 g, mol/mL ratio of Mo:DBT (~1:100)) were pressurized to 4.0 MPa (40 bar) in a 100 mL heptane solution containing 10 wt.% of CS₂ (sulfiding agent) under hydrogen flow (40 mL/min) for 4 h and at 573 K to ensure complete sulfidation.

After sulfidation, the reactor was cooled down to room temperature for the dibenzothiophene HDS test. At this temperature, the liquid feed was switched to dibenzothiophene solution (0.22 g, 1.194 × 10⁻⁵ mol/mL). The temperature was adjusted to 573 K under H₂ pressure of 4.0 MPa (40 bar) and maintained for 6 h. The dibenzothiophene content was measured with an Agilent 6890 gas chromatograph equipped with a FID detector and a 30 m × 0.25 mm × 0.25 μm capillary column (ZB-5MSi, 5% Phenyl column). HDS catalytic activity was estimated using (3):

$${\mathrm{X}}_{\mathrm{HDS}}(\%)=\frac{{\mathrm{C}}_{\mathrm{DBT}}^0-{\mathrm{C}}_{\mathrm{DBT}}}{{\mathrm{C}}_{\mathrm{DBT}}^0}\times 100$$

(3)

where ${\mathrm{C}}_{\mathrm{DBT}}^0$ is the DBT content in the feedstock (wt.%) and C_DBT is the DBT content in the products (wt.%) [29]. Therefore, the catalytic selectivity ratio between the hydrogenation (HYD) and direct desulfurization (DDS) is estimated from Equation (4):

$${\mathrm{S}}_{\mathrm{HYD}}/{\mathrm{S}}_{\mathrm{DDS}} =\frac{{\mathrm{C}}_{\mathrm{x}}}{{\mathrm{C}}_{\mathrm{DBT}}^0-{\mathrm{C}}_{\mathrm{DBT}}}$$

(4)

where C_x is the content of phenylcyclohexane (PhCh) or biphenyl (BP) [30].

4. Conclusions

In the present work, a series of new catalysts (RhMo/Al₂O₃, RhMo-x/Al₂O₃ (where x = EDTA, AA, CA) were successfully synthesized and characterized. The UV-vis analysis confirmed the presence of octahedral molybdate species between 320–360 nm for RhMo/Al₂O₃, and the shift to a lower wavelength in the visible part of 220 nm for the chelated catalysts was observed and indicated the formation of less polymerized molybdate species and heteropoplymolybdates. Band gaps of 3.779 eV (RhMo/ɣ-Al₂O₃), 4.341 eV (RhM-EDTA/ɣ-Al₂O₃), 4.394 eV (RhMo-AA/ɣ-Al₂O₃) and 4.478 eV (RhMo-CA/ɣ-Al₂O₃) were obtained. The TEM imaging confirmed that the materials had fringe-like morphologies, deemed as MoS₂ slabs. The chelated catalysts showed a greater dispersion when compared with the unchelated catalysts, and this was confirmed by XRD analysis by the absence of crystalline peaks for the chelated catalysts. RhMo/Al₂O₃ resulted in higher catalytic activity when compared with the chelated catalysts, and this was confirmed by XPS showing more MoS₂ phases of the RhMo/Al₂O₃ catalyst to be (63%), and it was also confirmed by HDS activity where RhMo/Al₂O₃ exhibited the highest DBT conversion of (88%). The addition of the chelating ligands (EDTA, AA and CA) resulted in lower HDS activity. Since this is a new catalyst, a lot of parameters must be investigated, such as the sulfidation temperature for Rh, the crystallite size effects, the molar ratio of Rh:chelating ligand to be used, and the type of the chelating ligand to be used. This would help to understand the catalyst and how the activity could be enhanced.