Anthracene Aquacracking Using NiMo/SiO2 Catalysts in Supercritical Water Conditions

Anthracene aquacracking using NiMo/SiO₂ catalysts in supercritical water conditions

T.R. Reina ¹, P.Yeletsky ², J.M. Bermúdez ¹, P. Arcelus-Arrillaga ¹, V.A. Yakovlev ^2*, M.Millan ^1*

¹Department of Chemical Engineering, Imperial College London London SW7 2AZ, UK

² Boreskov Institute of Catalysis, Lavrentieva Ave. 5, Novosibirsk, Russian Federation

 

*Corresponding authors: marcos.millan@imperial.ac.uk, yakovlev@catalysis.ru

Introduction

The current status of world oil reserves is a contentious matter, but it is widely accepted that conventional resources are dwindling and their reserves are less easily accessible [1]. Therefore, the production of heavy crude oil (HCO), which is the remnant of conventional oil has become more relevant and will remain so in the foreseeable future [2]. In this sense, there is a need for more efficient refining processes to transform HCO into lighter fuels. Conventional processes for increasing the value of heavy oil fractions aim to increase the H/C ratio of fuel, generating lighter fractions. However, this implies either rejecting a large amount of the carbon in the feed as in thermal and catalytic cracking processes, or using high pressure hydrogen, an expensive gas, in hydrocracking processes [3].

As an alternative, a fairly new approach to upgrade heavy oils consists in the catalytic cracking in supercritical water (SCW) conditions using metal based catalysts. This method takes advantage of the properties of SCW (T=374˚C, P=220 bar and d=0.32 g/ml and above), in which water inverses its properties as a solvent and became a non-polar solvent [4]. Under these conditions, aromatic rings are well dissolved and dispersed in the reaction media favoring the contact between reactants and catalysts particles.

On the other hand, the chemistry of heavy oils is rather complex making difficult to obtain deep comprehension of the process. Studies with model compounds representing chemical structures found in heavy oils give relevant information about the reactivity in the medium. In addition, the use of model compounds facilitates to understand the role of the catalysts and to identify the most likely reaction pathways.

In this scenario, the aim of this work is to apply a series of NiMo/SiO₂ catalysts with different Ni/Mo ratio in the anthracene upgrading in SWC conditions. A complete physico-chemical characterization of the fresh and spend catalysts in order to correlate the catalysts' features with catalytic behavior is also a subject of this study.

Experimental

The catalysts were prepared by sol-gel method and partially reduced in H₂ flow at 550 ^oC for 1 h followed by passivation with ethanol. Anthracene aquacraking experiments were performed in a stainless steel batch reactor with a volume of 18 ml. The operation of the reactor has been described elsewhere [3]. The reactor was filled with 0.2 g of anthracene 98% (Sigma-Aldrich), 0.2 g of catalysts and the amount of water required providing 230 bar pressure at 425 ºC. 

Gas products formed were analyzed in a Perkin Elmer Clarus GC with TCD detector. Liquid products were separated from the remaining water, extracted with a mixture of chloroform/methanol 4:1 and then filtered. This liquid fraction was liquids in a Perkin Elmer Clarus GC with FID detector and in a Varian Star 3400/Saturn 2000 GC/MS to identify the main products obtained. The spent catalysts were recovered, dried and prepared for post-reaction studies.

All the fresh samples have been fully characterized by means of XRD, TEM, XPS, S_BET, H₂-TPR. Some of these techniques were applied also to study the spent catalysts.

 

Results and Discussion

The nominal compositions of the catalysts (if they were completely reduced to metallic Ni and Mo) together with their textural properties and the anthracene conversions are presented in Table 1.

Table 1.  Chemical composition (wt.%), textural properties of the prepared catalysts (if they were completely reduced to metallic Ni and Mo) and anthracene conversions

 

Sample	Ni (wt.%)	Mo (wt.%)	SiO2 (wt.%)	SBET (m2·g-1)	VPore (cm3·g-1)	Dpore (nm)	Anth. Conversion (%)
404020	40	40	20	87	0.17	8.1	46.7
305020	30	50	20	74	0.24	13.0	28.2
206020	20	60	20	68	0.25	14.9	35.0
107020	10	70	20	44	0.12	10.7	17.3

All the samples are mesoporous materials with specific surface area and porosity governed by the Ni/Mo ratio. In particular, the S_BET decreases when molybdenum loading increases. Concerning the catalytic performance, all the solids are active in the anthracene upgrading and their activity can be linked to the catalysts composition. More precisely, the presence of Ni seems to be crucial to achieve high conversions since the most active catalysts is the one with the highest Ni content and the least active sample the one with the lowest Ni loading.

Assuming that the upgrading process in supercritical water media may begin with a partial oxidation of the rings leading to C-O bonds that are easier to crack than C-C double bonds, the redox properties of the catalysts must be a relevant factor. Indeed, some interesting results can be extracted from the H₂-TPR (Figure 1)

                                                                                                                                                              

 Figure 1. H₂-TPR profiles of the NiMo/SiO₂ catalysts

As seen in Figure 1 all the catalysts present a complex profile accounting for the simultaneous reduction of NiO to Ni and several reduction steps of MoO₃.

Moreover this reduction profile is influenced by the Ni and specially Mo particles size. XRD data (not shown here) revealed that 404020 and 305020 catalysts are composed by small Ni-Mo particles forming a solid solution while for the 206020 and 107020 some big molybdenum dioxide particles (ca. 30 nm) were observed. The overall result is that only 404020 and 305020 samples achieved complete reduction in the TPR experiment indicating that these solids present better redox properties than the others. Furthermore, it seems very clear that Ni assists Mo reduction shifting the reduction zones to lowers temperatures. The later also may indicate an intimate Ni-Mo contact resulting in a positive effect in terms of reducibility.

Very interestingly Ni/Mo ratio also influences the liquid and gas products distribution. Figure 2 shows liquids products distribution. Flourenone and Xanthene are the dominant products for high Ni loadings in the catalysts composition while some different products as for example Anthraquinone and Xantone appear for the l07020 sample. In any case a diversity of liquids products including oxidation products, ring opening molecules and hydrogenation products were obtained.  

 

Figure 2. Liquids products distribution. Labels: 1 (404020); 2 (305020); 3 (206020); 4 (107020)

Regarding the gaseous products, very interesting results were obtained (Figure 3). Irrespectively of the catalysts composition, H₂ was the most abundant gas produced in the reaction. Together with H₂ some CO, CO₂ and CH₄ were produced. Again, as observed for the liquid fraction, the gas product distribution is controlled by the catalysts composition. In fact, some parallel reactions taking place simultaneously may account for the obtained gases. For instance, the water gas shift reaction and the CO and CO₂ methanation are plausible since Ni is a typical active phase for both processes. Furthermore, partial oxidation of anthracene and cracking reactions may happen according to the obtained liquids products.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

                                                                                                                                    Figure 3 Gas products distribution. Labels: 1 (404020); 2 (305020); 3 (206020); 4 (107020)

 

 

Finally, the post reaction characterization (not included in the abstract for sake of briefness) indicates a certain degree of damage in the catalystsx structure. This observation opens up a new challenge in terms of developing robust catalysts to be employed in upgrading reactions in SWC conditions.

 

 

Conclusions

 

A series of NiMo/SiO₂ catalysts has been successfully applied in the anthracene aquacraking in SWC conditions. The samples present different structural and redox properties depending on the Ni/Mo ratio. The best performances are obtained when small Ni-Mo particles are in close contact and no segregation of MoO₂ is attained.

A variety of liquids products was obtained including ring opening, oxygenated and hydrogenated molecules confirming the successful upgrading of anthracene. Additionally, a valuable gas stream (rich in H₂ with some CH₄) resulted from the process making it more interesting in terms of energy efficiency. Both liquids and gaseous products distribution are controlled by the Ni/Mo ratio that seems to be the key parameter in the catalystsx design in order to maximize the overall performance.

 

Acknowledgments

This research was performed under the UNIHEAT project. The authors wish to acknowledge the Skolkovo Foundation and BP for financial support.

References

(1)      Owen, N. A., Inderwildi, O. R. & King, D. A. The status of conventional world oil reserves—Hype or cause for concern? Energy Policy. (2010) 38 (8), 4743-4749.

(2)     Hart Energy Research Group. Heavy Crude Oil: A Global Analysis and Outlook to 2035. (2011) Hart Energy.

(3)     J.L. Pinilla, P. Arcelus-Arrillaga, H. Puron, M. Millan, Appl. Catal. A 459 (2013) 17-25

(4)     H. Weingärtner, E.U. Franck, Angew. Chemie Int. Ed. 44 (2005), 2672–2692.