Preparation , characterization of Mo catalysts supported on Ni-containing calcium deficient hydroxyapatite and reactivity for the thiophene HDS reaction

Ni-containing Calcium Hydroxyapatite (NiCaHAp; 3.31 wt.% Ni) was synthesized by coprecipitation and used as catalyst support. Molybdenum was supported on NiCaHAp by impregnation using ammonium heptamolybdate. The prepared catalysts Mo(x)/NiCaHAp (x: 2 to 8 wt % in Mo) were characterized by elemental analysis, XRD, FT-IR, N2 adsorptiondesorption and TEM-EDX. The catalysts were sulfided in-situ at 673 K under flowing H2S/H2 (15 Vol.% H2S) and tested in hydrodesulfurization (HDS) of thiophene at 673 K. The main XRD peaks of hydroxyapatite CaHAp phase were observed in all samples and a peak due probably to crystalline MoO3 phase was also identified from the results. However, no crystalline phase of NiO was found for the catalysts, which showed its Ni species were highly dispersed. The sulfided catalysts Mo(x)/NiCaHAp presented are active in HDS of thiophene, despite the presence of some large MoO3 crystallites and incomplete sulfidation. This activity may be due to interaction of NiO and MoO3 on CaHAp resulting in the formation of Ni-Mo-S phase under flowing H2S/H2. When the molybdenum content increased the HDS activity increasead slightly, which was caused by the agglomeration of MoO3. The Mo(8)/NiCaHAp catalyst is about two times less active for thiophene HDS than the commercial NiMoP/Al2O3.


INTRODUCTION
It is currently known that a great part of natural phosphates adopts a crystalline structure similar to that of synthetic hydroxyapatite Ca 10 (PO 4 ) 6 (OH) 2 [1].The composition of natural phosphates is very complexe.Indeed, the apatitic lattice allows various anionic and cationic substitutions.
Despite their low cost and their availability in the market, the use and application of apatite in catalysis is still a subject of relatively scarce investigations.For example, calcium hydroxyapatite containing transition metals were found to be very active in the methane dry reforming [2], the hydrogenolyse of S-containing substrates [3], the propane oxidative dehydrogenation [4] and the epoxidation of olefin [5] Within our research team, we have been looking for some new catalytic applications for apatitic compounds with a special interest for hydroxyapatite.Most of these studies are focused on the hydrodesulfurization (HDS) of petroleum distillate fractions for the reduction of atmospheric pollution due to the emission of hazardous compounds such as SO 2 .
Previously, we have shown that hydroxyapatite containing small amounts of carbonate species, with a Ca/P molar ratio around 2, can be used as a support for hydrodesulfurization sulfide catalysts since it exhibits the required textural properties and a good thermal stability [6].
In the present work, we studied the use of nickelcontaining calcium deficient hydroxyapatite Ni-CaHAp (3.31 wt% Ni) prepared by coprecipitation at pH basic and calcined at 773 K, as a support for molybdenum catalysts during the hydrodesulfurization of thiophene.Molydenum supported catalysts were prepared by impregnation using ammonium heptamolybdate (NH 4 ) 6 Mo 7 O 24 .4H 2 O.
The prepared catalysts were characterized using several experimental techniques such as X-ray diffraction, FT-IR spectroscopy, nitrogen adsorption at 77 K and transmission electron microscopy combined with energydispersive X-ray analysis (TEM-EDX).Meanwhile, the hydrodesulfurization catalytic activity was assessed at atmospheric pressure using thiophene as a probe molecule.

Synthesis of Ni-containing CaHAp
The nickel-containing calcium hydroxyapatite (noted as NiCaHAp) precursor was synthesized using the following precipitation method: an ammoniacal solution of (NH 4 ) 2 HPO 4 (m = 11.75 g dissolved in 1000 ml of distilled water + 100 ml of 25% aqueous ammoniac) was poured into a second solution (100 ml) of calcium nitrate (m = 18.51 g) and nickel nitrate (m = 5.75 g) while stirring at 353 K.The pH was maintained at 9 during mixing.After the dropwise addition (3 h), the mixture was kept at 353 K for 1 h.After this time, the yellow solid was separated by filtration, washed with distilled hot water, dried overnight at 393 K and calcined at 773 K.The Ca/P atomic ratio of Ni-CaHAp obtained was 1.51 indicating that the material is calcium deficient hydroxyapatite with a chemical formulae: Ni(3.31 wt%)-Ca 9.1 (HPO 4 ) 0.9 (PO 4 ) 5.1 (OH) 1.1

Synthesis of Mo(x)/NiCaHAp
Molybdenum was loaded on the NiCaHAp by pore volume impregnation.Before impregnation, 2 g of the Ni-CaHAp was calcined under dry air flow, at 773 K, for 2 h, and then added to 1 cm 3 of an aqueous solution of ammonium MATEC Web of Conferences heptamolybdate with the required concentration.The obtained slurry was left at room temperature for about one hour, and then evaporated at 333 K.All the prepared catalysts were dried overnight at 393 K in an oven, and finally calcined at 773 K for 2 h under ambient atmosphere.All catalyst will be referred to as Mo(x)/Ni-CaHAp, were x (expressed in wt.%) represents the weight percentage of Mo in the catalyst.The prepared catalysts contain loads varying from 2 to 8 wt% in molybdenum.

Characterization techniques
XRD patterns were obtained using a PW 1716 diffractometer, which utilizes Ni-filtered Cu (K α ) radiation (λ: 1.5405 Å).Samples were run as fine powders mounted on glass slides.Fourier transform infrared spectroscopy (FT-IR) studies were carried out on the Vertex 70 spectrometer in the range 400-4000 cm −1 .The transmission technique was applied and the samples were prepared as standard KBr pellets (2%).
The textural properties of all the catalysts were carried out in an ASAP 2000 Micrometrics instrument by means of nitrogen adsorption and desorption at 77 K. Before analysis, all the samples were evacuated at 573 K under vacuum.
The transmission electron microscopy (TEM-EDX) analyses were performed in a Tecnai G 2 -FEI operated at 120 KV.TEM specimens were prepared by depositing a few drops of catalysts dispersed in acetone on a carboncoated copper grid.

HDS catalytic measurements
The thiophene hydrodesulfurization reaction was carried out in a fixed bed flow reactor.Before the reaction, the oxide precursor (50 mg) was pretreated at 773 K under ambient atmosphere, and then in-situ sulfided at 673 K under flowing H 2 S/H 2 (15 Vol.% H 2 S) for 2 h 30 min at atmospheric pressure.Finally, the feed (thiophene 7.5 mol%, H 2 S 2 mol%, and H 2 90.5 mol%) was supplied to the reactor and the HDS reaction was monitored at 673 K under atmospheric pressure.In these conditions, the total conversion of thiophene was kept below 15% to operate in differential regime.Thus, the catalytic activity could be estimated from the pseudo-first order rate constant k HDS : (1) Where F is the thiophene molar flow (mol/h), W is the weight of the sulfide catalyst (g), X is the total conversion of thiophene, and C the concentration of thiophene in the feed (mol/l).
The hydrodesulfurization reaction products were separated on a gas chromatograph (GC) using an OV1 column and analyzed with a thermal conductivity detector (TCD).Besides un-reacted thiophene, the reaction products are, H 2 S and C4 (1-butene, trans-2-butene, cis-2-butane and butane).Thiophene conversion data were collected at 60 min intervals during 3 hours, and the values reported here are those obtained when the steady state is reached.

BET surface area and elemental analysis
Table 1 summarizes the physicochemical parameters of the catalysts.Chemical analysis revealed a Ca/P molar ratio ranging from 1.51 to 1.49, compatible with calciumdeficient hydroxyapatite [7].
Figure 1 shows the N 2 adsorption-desorption isotherms of the NiCaHAp and Mo(x)/NiCaHAp catalysts.All samples showed a type IV isotherm with a type H2 hysteresis loop, indicating the presence of well-developed mesoporosity in all the samples.
The BET surface area for NiCaHAp was 114 m 2 /g.For the supported catalysts, an increase in molybdenum loading results in a decrease in the surface area probably due to pore blockage, since a similar pattern is observed for pore volume measurements (Table 1).There was a 04016-p.2minimal increase of the pore diameter with an increase in the molybdenum loading.

REMCES XII
It should be noted that the changes in pore volume, pore diameter and BET surface area decreased abruptly when the Mo exceeded 5wt.%, which is attributed to the agglomeration of molybdenum oxide.

Powder X-ray diffraction
The powder XRD patterns of NiCaHAp and Mo(x)/ NiCaHAp catalysts in the angular range 2θ of 10-70 • are presented in Fig. 2. The NiCaHAp sample containing 3.31 wt.% Ni showed only diffraction lines belonging to calcium-hydroxyapatite CaHAp (JCPDS 09-0432), suggesting that the added nickel is amorphous and well dispersed on its surface.The diffraction peaks of CaHAp were clearly observed with all the Mo(x) /NiCaHAp catalysts.The diffraction peaks of CaHAp were weak, which was probably due to the small size of the particles [8].Besides the peaks characteristic of calcium hydroxyapatite, we note the presence of a diffraction peak around 2θ ≈ 29 • for high Mo loadings (8 wt.% Mo).This peak can be probably attributed to the presence of bulk MoO 3 [9,10].Formation of a small amount of NiMoO 4 (the strongest peak being at 2θ ≈ 29 • ) is not excluded [11,12].
Meanwhile, we note the presence of a vibration band at 814 cm −1 , which would correspond to the oxygen atom vibrations in Mo-O-Mo in which Mo represents hexavalent molybdenum [14].

Transmission electron microscopy
The micrograph of Mo(5)/NiCaHAp is shown in Fig. 4. Dark particles of metal oxide with a size distribution of 20-80 nm existed on the surface of CaHAp.EDX data show a Ni/Mo ratio of 1.07 which correlates with the value from elemental analysis (Table 1).

Catalytic tests
The thiophene hydrodesulfurization (HDS) model reaction has been used to establish the performance of the sulfided Mo(x)/NiCaHAp catalysts, k HDS values are plotted as a fonction of time on-stream (TOS) in Fig. 5.As can be seen, the sulfided Mo(x)/NiCaHAp catalysts, which initially display a high thiophene HDS activity, undergo a strong deactivation during the first hour.This behaviour is attributed to a loss of sulfur atoms from the active NiMoS phase during reaction with thiophene [15].The Mo(x)/NiCaHAp catalysts show a much higher HDS activity than the sulfided NiCaHAp (k HDS = 0.78 l/hg; 3.31 wt% Ni) and Mo/CaHAp (k HDS = 1.64 l/hg; 2.1 wt% Mo) catalysts, which is attributed to the promoting effect of Ni 2+ ions in the mixed NiMo sulfided materials.It is worth noting that the activity of sulfided Mo(x)/NiCaHAp catalysts slightly increases with Mo loading, because the higher Mo loading will result in more and larger MoO 3 particles and consequently the degree of sulfidation will be lower (Table 2).Among the catalysts tested, the Mo(5)/NiCaHAp catalyst showed better sulfidation than other catalysts.This may be due to a well dispersion of MoO 3 particles on NiCaHAp (Table 2).

CONCLUSION
The agglomeration of MoO 3 and the incomplete sulfidation may explain the relatively low activity of the Mo(x)/NiCaHAp catalysts in hydrodesulfurization of thiophene compared to the commercial NiMoP/Al 2 O 3 in the same experimental conditions.

Table 2 .
Rate constants k HDS and degree of sulfidation S/(Ni + Mo) of catalysts.