Glycerol is a natural byproduct of the manufacture of soaps from the hydrolysis of animal fats and vegetable oils. Due to the incessant increment of the worldwide production of biodiesel, the production of glycerol, a by-product of the biodiesel industry, has notably increased causing as consequence a drop in its price. In this way, glycerol has been turned into an interesting starting raw material for others chemical products, where the double dehydration of glycerol to acrolein is one of proposed routes for glycerol valorisation.
Because of the crude glycerine is found diluted in water, the catalysts must be active and resistance to the presence of water avoiding a separation step of the glycerine from the water and therefore reducing the price of production of acrolein. Various solid acid catalysts have been tested in the dehydration of glycerol, including nio-bium oxide [1-3], phosphoric acid [4, 5] and zirconium niobium mixed oxides [6]. These catalysts reach the full conversion of glycerol and acrolein selectivities as high as 75mol% [3]. However, the main drawback of these cata-lysts is the deactivation with time on stream due to mainly to adsorption of high boiling point compounds. The characteristic features of the catalysts affecting to the catalytic activity are related to the amount of acid centers and the type of acid sites. The catalytic activity is related to Nb – OH groups and polymeric structures exposing Nb=O species [3, 6].
The main goal of this work is to evaluated how the preparation of the catalysts affect to the catalytic activity of supported niobia-phosporous on a mesoporous silica SBA-15 doped with zirconium. Doping with zirconium improves the acidity of the support and the stability of the catalysts as it has been shown previously [7].
2 Experimental
For the synthesis of mesoporous silica SBA-15 doped with zirconium (Si/Zr molar ratio of 5), 8 g of EO20-PO70-EO20 (Pluronic 123 from Aldrich) polymer was dissolved in 288 mL of a 1.7 M HCl aqueous solution and stirred at 313 K for 1 day. Then, 14.6 mL of tetraethoxysilane and 5.8 mL of zirconium n-propoxide were added, under vigorous stirring. The resulting solution was aged at room temperature for 3 days. The final product was fil-tered, washed with water and dried at 333 K. The solid was calcined at 823 K for 6 h. The support was labelled as SiZr
This solid was impregnated with niobium oxalate to achieve a 8wt% of Nb2O5 by using the incipient wetness impregnation. The catalyst was dried and calcined at 673 K during 4h (heating rate 2 K/min) and labelled SiZr_8Nb2O5. The niobium-phosphorous catalysts were prepared following to two different strategies. In the first one, the support was impregnated with a solution of niobium oxalate (8wt%Nb2O5) and then dried over night at 333 K. Following, the phosphorous was added by using H3PO4 (85wt%) wiith Nb/P molar ratio of 0.2. Finally, the catalyst was dried and calcined at different temperatures (523, 673 and 973 K) during 4 hours. The catalysts were labelled 0.2NbP-x where x is the temperature of calcination. In the second strategy, the support was impregnated with a solution containing niobium oxale and phosphoric acid with the Nb/P molar ratio of 0.2. This catalyst was calcined at 673 K for 4 hours and labelled 0.2NbP-400-I.
The support and catalysts were characterized by means of nitrogen adsorption, elemental analysis, XDR, XPS, TPD-NH3 and adsorption of pyridine coupled to FTIR.
The dehydration of glycerol was performed, at atmospheric pressure, in a fixed-bed continuous-flow stainless steel reactor (9.1mm in diameter, and 230mm in length), operated in the down-flow mode Catalysts were pre-treated in situ at atmospheric pressure under a nitrogen flow of 30 mLmin−1 at 598 K for 30 min. The glycerol solu-tion (10 wt.% in water) was supplied by means of HPLC pump at 0.05 mLmin−1 feed rate in a N2 flow (30mLmin−1). The evolution of catalytic tests was studied by collecting liquid samples after 2, 5 and 8 h in a vial cooled in an isopropanol–liquid N2 trap. These vials were sealed for posterior analysis by gas chromatography. The products measured were acrolein, acetaldehyde, hydroxyacetone and allyl alcohol.
3 Results and discussion
The support and catalysts were studied by XRD to confirm the presence of the mesoporous structure which was mantained after the impregnation and calcination of the active phases. Moreove, it was not detected neither niobium phase nor phosphate phaser at higher 2.The SiZr catalyst show a high glycerol conversion even higher than the SiZr_8Nb2O5 within the 8 hours of time on stream (TOS) although the selectivity of the SiZr_8Nb2O5 to-wards acrolein is higher (Table 1). This issue is related to higher Brönsted acid sites present in the SiZr_8Nb2O5 since the SiZr support only shows Lewis acid sites. When phosphorus is added to the SiZr_8Nb2O5 catalyst (Table 1), the catalytic performance is improved both in terms of glycerol conversion and acrolein selectivity. It is observed an increase of Brönsted acid sites, probably the presence of P – OH speccies and a decrease of Lewis acid sites compared to the SiZr_8Nb2O5 catalyst. This improvement is only shows when the catalyst is calcined at 673 K (Figure 1) higher and lower calcination temperatures had not beneficial effect on the catalytic activity.
It was also compared the catalytic activity when the impregnation with phosphorous and niobium was carried out in one step (Figure 2). In this case, the catalytic performance is drastically improved since the acrolein yield is increased in a 20mol%. This fact point out the presence of an active phase, no detected by XRD analysis, formed by interaction of niobium and phosphorous.
4 Conclusions
The catalytic performance of a Nb-P catalyst supported on a mesoporous silica SBA-15 doped with zirconium is ameliorated when the impregnation of the two active phases is carried out in one step reaching an acrolein yield of 46mol% and a glycerol conversion of 85mol%