Partial Oxidation of n-Pentane over Vanadium Phosphorus Oxide supported on Hydroxyapatites

The selective oxidation of n-pentane to value-added products, maleic anhydride or phthallic anhydride by vanadium phosphorus oxide loaded on hydroxyapatites as catalysts and oxygen as oxidant was investigated. Hydroxyapatite (HAp) and cobalthydroxyapatite (Co-HAp) were prepared by the co-precipitation method and VPO with varying weight percentages (2.5–15.0 %) were loaded on the hydroxyapatite supports by the wet impregnation technique. The catalyst materials were characterized by surface area measurements, elemental analysis, powder X-ray diffraction (XRD), infrared spectroscopy (IR) and temperatureprogrammed reduction (TPR). VPO is present in two phases, viz. (VO)2P2O7 and VOPO4. With increase in the VPO loading on the hydroxyapatites, the (VO)2P2O7 phase also increased. From catalytic results, a conversion of 75 % of n-pentane and selectivity towards maleic anhydride, about 50 % and phthalic anhydride, about 25 %, were consistently achieved with loadings of 5.0 and 7.5 wt. % VPO at 360 °C for GHSVs of 1900 and 2300 h. Under optimum conditions, product yields of up to 40 % maleic anhydride and 20 % phthalic anhydride were obtained. It is proposed that the products formed through the diene intermediate.


Introduction
Maleic anhydride (MA) and phthalic anhydride (PA) are precursors used in the manufacture of synthetic fibres, plastics, paints and adhesives 1 .The selective and controlled oxidation of hydrocarbons using heterogeneous catalysts to produce these compounds, occupy an important place in the chemical industry 2,3 .Vanadium phosphorus oxide (VPO) was found to be efficient in converting n-pentane to maleic anhydride and phthalic anhydride 4,5 .This catalytic reaction is promoted by the existence of both the vanadyl pyrophosphate (VPP) as well as of VOPO 4 phases in VPO 5,6 .Supported VPO catalysts showed an enhancement in the selectivity towards maleic anhydride and phthalic anhydride.Xiao et al. 7 used fumed SiO 2 -supported VPO catalysts, prepared by deposition-precipitation, for n-butane activation.A conversion of 33-40 % of n-butane and maleic anhydride selectivity of 65-87 mol % were obtained over the temperatures tested.In another study, Nie et al. 8 compared the efficiency of VPO supported on Al-MCM-41 with a large-pore silica-supported VPO catalyst in the activation of n-butane.The VPO supported on Al-MCM-41 showed an increase in maleic anhydride (MA) selectivity with a slight decrease in n-butane conversion, attributed to the interaction between the VPO component and the Al-MCM-41 support.Recently, VPO supported on alumina was used for propane ammoxidation and was hyped as a potentially interesting catalytic system for the industrial synthesis of acrylonitrile 9 .Michalakos et al. 10 studied the variation in the selectivity for the selective oxidation n-pentane on vanadium oxide supported on Al 2 O 3 and SiO 2 .The loadings for the Al 2 O 3 samples were 8.2 and 23.4 wt.% V 2 O 5 and for the SiO2 samples, 1.0 and 10.0 wt.% V 2 O 5 .The Al 2 O 3 -based catalysts were more active than the supported SiO 2 catalysts, but for catalysts of the same support, the activity per mole depended slightly on the loading.At low pentane conversions, the 1 wt.% V 2 O 5 /SiO 2 produced pentenes and 1,4-pentadiene, whereas the 10 wt.% V 2 O 5 /SiO 2 produced mainly carbon oxides.The Al 2 O 3 -based catalysts produced dehydrogenated products.Here, MA was produced with a selectivity of about 80 % at low conversion on the 23.4 wt.% V 2 O 5 loaded catalyst, but was a minor product on 8.2 wt.% loaded catalyst.Small amounts of PA were observed on the Al 2 O 3 -based catalysts, and none was observed on the supported SiO 2 catalysts.This study was corroborated by a separate study conducted by Klose et al. 11 .
Activation of hydrocarbons depend, to a large extent, on the acid-base properties of the material or on the isolated cations capable of activating C-H bonds 12,13 .Hydroxyapatites (HAp) are bi-functional materials containing acidic and basic sites in the crystal lattice.HAps are useful as materials in bioceramics, adsorbents, catalysts and catalyst support materials.Its hexagonal structure is made up from columns of calcium ions and oxygen atoms which are located parallel to its axis [14][15][16] .Three oxygen atoms from each PO 4 tetrahedron are shared by one column and the fourth oxygen atom is attached to the neighbouring column.The lattice of HAp is structured to allow substitution of the calcium ions with other cations by ion exchange and ion adsorption or a combination of these methods to improve its catalytic performance 15,[17][18][19] .Co-HAp offers high stability and the OH group within the phosphate framework leads to the formation of active oxygen species which are essential for ODH reactions 14,20,21 .In another study using Co-HAp as the catalyst, Opre et al. 22 showed excellent results for the epoxidation of styrene with in dimethylformamide.Hydroxyapatites are efficient as supports, in particular its basic properties are useful for the Knoevenagel reaction in heterogenous media without solvent 23 , the water gas shift reaction with gold and ruthenium as the active components 24  hydroxyapatite 25 .In our previous study 26 , we reported the activation of n-pentane using V 2 O 5 supported on hydroxyapatite as catalysts.V 2 O 5 showed excellent redox capabilities.Selectivity towards the anhydrides (MA 40 % and PA 25 %) was obtained with the 5.0 wt.% V 2 O 5 at 360 °C at a conversion of about 60 %.At higher weight loadings, although the conversion increased, the selectivity towards the desired products decreased.
In this paper, we report the preparation and characterization of VPO supported on calcium and cobalt hydroxyapatites and the effect of the VPO loading and support interaction, together with different phases present on the catalyst, in the partial oxidation of n-pentane.

Preparation of Catalysts
The catalytic materials were prepared using methods published previously, with modifications 26,27 .Concentrated ammonia (BDH, Poole, England) was added to a 60 mL solution of Ca(NO 3 ) 2 •4H 2 O (6.67 × 10 -2 mol) (Merck KGa, Darmstadt, Germany) to increase its pH to 11.The calcium solution was subsequently diluted to twice its volume with distilled water.A 100 mL solution of (NH 4 ) 2 HPO 4 (4.00 × 10 -2 mol) (Merck KGa, Darmstadt, Germany) was also adjusted to pH 11 using ammonia, diluted to 160 mL.The phosphate solution was added drop-wise to the calcium solution over a period of 30 min with continuous stirring at room temperature.A white, gelatinous precipitate that formed was boiled for 10 min with stirring.Thereafter, the precipitate was filtered under vacuum and washed thoroughly with deionised water.The solid was left to dry overnight in an oven set at 100 °C, followed by calcination under the flow of air at 500 °C.In the preparation of the cobalt hydroxyapatite, Co(NO 3 ) 2 •6H 2 O (Merck KGa, Darmstadt, Germany) was used.The method of preparation of VPO has been extensively reported in literature 3,6,8,9,28 .In a modified method for preparing VPO, a mixture containing V 2 O 5 (5.00 g) (BDH, Poole, England) and H 3 PO 4 (30 mL, 85 wt.%) (BDH, Poole, England) was added to 100 mL of deionized water and refluxed for 24 h.A yellow solid that formed was filtered, washed thoroughly with deionised water, followed by acetone (BDH, Poole, England) and air-dried in an oven at 110 °C.The resulting solid was then refluxed with isobutanol (50 mL) (Merck KGa, Darmstadt, Germany) for 20 h.The solid was filtered, dried in an oven set at 110 °C for 16 h and calcined through flowing nitrogen in a furnace set at 500 °C.Finally, in the preparation of the supported catalysts, weight percentages of 2.5, 5.0, 7.5, 10.0 and 15.0 VPO were loaded on the HAp and Co-HAp supports by wet-impregnation 29,30 .

Characterization of Catalysts
The surface area of the catalysts was determined by the BET method 30 .Catalysts were pretreated by degassing under N 2 flow overnight at 250 °C using a Micrometrics Flow Prep 060 (Micrometrics, USA).The degassed samples were analyzed in the Micrometrics Gemini 2360 (Micrometrics, USA), fully automatic, multi-point BET surface area analyzer under liquid N 2 .The diffraction patterns of the catalysts were recorded in the 2q range of 2-80 °using a Phillips PW 1830 X-ray diffractometer system, equipped with Co K a radiation (l = 1.7889).The experimental conditions corresponded to the step size of 0.040 °(2q) (number of steps = 1951), and a dwell time of 1 s.The concentrations of Ca, Co, V and P in the catalysts were measured in aqueous HNO 3 by inductively coupled plasma (ICP) using a Perkin Elmer Optical Emission Spectrometer Optima 5300 DV instrument.Standards, 500 ppm Ca, Co, V and P, were purchased from Fluka.
FT-IR spectra of the samples were recorded with a Nicolet Impact 420 spectrophotometer in the mid IR (400-4000 cm -1 ) region using the KBr disc technique (0.5 % dilution).TPR investigations were carried out with a custom made reactor system connected to a thermal conductivity unit.Hydrogen (5 %) in argon was used for the analysis.For heating, a calibrated furnace was used and a temperature control unit monitored the temperature to within 2 °C.The average vanadium oxidation state was determined by means of a redox titration 28,29 .

Measurement of Catalytic Reactions
Catalytic testing was carried out using a continuous, fixed bed tubular stainless steel reactor (i.d.= 10 mm, l = 300 mm) in down flow mode.The organic feedstock, n-pentane (Saarchem, South Africa) was injected into the reactor lines by means of a high precision isocratic pump (LabAlliance Series II) and vapourized on-line.The vapour was carried by the oxidant, an oxygenhelium mixture (20 % O 2 , balance He) (Afrox, BOC Special Products).Preliminary testing of VPO and the support materials, HAp and Co-HAp were initially conducted.After these preliminary runs, the VPO-supported catalysts were tested.In all testing, the catalyst bed (2 mL) was located at the centre of the reactor, with carborundum (24 grit, Promark Chemicals, South Africa) filling the spaces on either side of the bed and ends plugged with glass wool.The reactor was heated in an electric furnace with K-type thermocouples monitoring the temperature of the furnace and the catalyst bed.The product stream was analysed by Gas Chromatograph (Varian Star 3400) equipped with a FID module and GC-MS (Finnigan MAT GCQ) using capillary columns (J&W HP5-MS, 250 µm diameter).The carbon oxides and light organic compounds were intermittently monitored by a TCD Gas Chromatograph (Buck Scientific, SRI Instruments, USA) using a packed column (6 silica gel/6 molecular sieve).Water was quantified using a Karl Fischer titrator.

Catalyst Characterization
The surface areas and the elemental composition of all the catalysts used in the present studies are displayed in Table 1.The BET surface area of HAp of 48 m 2 g -1 agrees with that reported in literature 24 .There is a decrease in the surface area of the materials from 40 to 15 m 2 g -1 with the increase in VPO loading.It is possible that the narrow pores of the support are filled with the active component making it inaccessible to nitrogen molecules, leading to a decrease in the surface area.The total metal content was also determined for each catalyst to investigate its impact on the activity and selectivity of the catalysts.From elemental analysis, the Ca/P ratio for HAp is 1.67 (Table 1) which indicates that stoichiometric hydroxyapatite was prepared.As the loading of VPO on the supports increased, the M/P (M = Ca and Co) ratios decreased due to the contribution of phosphorus from the hydroxyapatite and VPO.
The XRD patterns of HAp (Fig. 1) confirm that the phases belong to stoichiometric hydroxyapatite with the formula, Ca 10 (PO 4 ) 6 (OH) 2 and they match well with the phases reported by Parhi et al. 32 .The patterns for supported catalysts showed the presence of (VO) 2 P 2 O 7 (l) peaks at 2q angles of 18.6 °, 23.1 °, 28.5 °, 29.9 °, 33.7 °and 36.9 °whereas a mixture of the VOHPO 4 •½ H 2 O and VOPO 4 •2 H 2 O (u) phases displayed peaks at 2q angles of 13.2 °, 20.3 °, 21.6 °, 28.9 °, 29.9 °, 41.1 °, 46.3 °, 59.5 °and 63.8 °.These agreed with patterns reported in literature 33 .For the Co-HAp and Co-HAp-supported catalysts (Fig. 2), the phases confirms the presence of hydroxyapatite although it was found to be non-stoichiometric.This was also confirmed by the Co/P ratio obtained through ICP-OES which was 1.82 instead of 1.67.The average vanadium number for the catalysts ranged from 4.11-4.46,confirming the presence of the (VO) 2 P 2 O 7 phase.
Fig. 3A and Fig. 3B displays the infrared spectra of HAp, Co-HAp and the supported catalysts.The formation of the hydroxyapatite structure is confirmed.
The signal at 3500 cm -1 is the result of the stretching vibration of the OH -ions at the lattice sites of the HAp crystals 27 .This signal was absent in the spectra of supported samples.A very broad absorption band was evident at 3400 cm -1 which suggests the presence of surface water molecules.Bands observed at 1380 cm -1 together with the signal at 1600 cm -1 are attributed to CO 3 [2][3][4][5][6][7][8][9][10][11][12][13][14][15]16,26,33 . Other tyical bands were observed at 1090 cm -1 and 1040 cm -1 due to (n as (P-O)) and 960 cm -1 (n s (P-O)).For the supported catalysts, there were additional peaks due to the stretching modes of pyrophosphate, P 2 O 7 4-(1160, 1120, 755 cm -1 ).From TPR analysis, HAp and Co-HAp were not reduced at the temperatures employed 34 , whereas for the supported VPO catalysts, reduction occurs in the 500 °C-800 °C range 6 .The TPR profiles for the 5VPO/HAp and 5VPO/Co-HAp, Fig. 4a and Fig. 4b respectively, shows a main peak around 550 °C, attributed to the removal of lattice oxygen related to V 4+ species in the (VO) 2 P 2 O 7 phase 35 .
The shoulders at 550 and 625 °C for the 15VPO/HAp and 15VPO/Co-HAp, respectively, is attributed to the removal of lattice oxygen related to V 5+ species (Fig. 4c, d) [35][36][37]    to the agglomerated VPO or the small amounts of vanadium orthophosphates present on the surface of the catalysts 38 .The reducibility of the VPO phase, in general, increased with an increase in vanadium loading.

Catalytic Testing
The partial oxidation of n-pentane in the gas phase with molecular oxygen was investigated with different loadings of VPO supported on HAp and Co-HAp using a fixed-bed continuous flow stainless steel reactor.The experiments were carried out at different flow rates and over a temperature range of 240-440 °C.Steady state for the reactions was established after 3 h.Blank runs conducted using the reactor filled with carborundum (24 grit) showed about 1.00-2.50% conversion of n-pentane producing mainly, carbon oxides (carbon monoxide and carbon dioxide).The activity increased to between 4.5 and 10.0 % with the use of hydroxyapatites as catalysts.The product stream consisted of lower hydrocarbons, oxygenated compounds, carbon oxides and very small amounts 1,3-pentadiene, furan and furfural.For the supported catalysts, the major products obtained were phthalic anhydride, maleic anhydride, carbon oxides, benzoic acid, furan, 2H-pyran-2-one and acetic acid.The minor compounds which are termed as 'others' included the C 2 -C 4 hydrocarbons, ethyl acetate, propanoic acid, 1,3-pentadiene, 2-cyclopenten-1-one, furfural, furaldehyde and 1,4-dioxane-2,3-diol.
To investigate the effect of GHSV, the 5VPO/HAp and 5VPO/Co-HAp catalysts were screened over the GHSV ranges of 1500 to 2700 h -1 at a temperature of 360 °C.Fig. 5 shows the typical profiles of the n-pentane conversion and the selectivity towards maleic anhydride and phthalic anhydride as a function of GHSVs.With increase in GHSV the conversion of n-pentane decreased due to the lower residence time of n-pentane on the catalyst.The best results were obtained at GHSVs of 1900 h -1 and 2300 h -1 .Above GHSV values of 2300 h -1 , the selectivity towards RESEARCH ARTICLE S. Singh, 4 S. Afr.J. Chem., 2016, 69, 1-8, <http://journals.sabinet.co.za/sajchem/>.phthalic anhydride (PA) and maleic anhydride (MA) decreased with selectivity towards carbon oxides increasing significantly.All catalytic runs were done in triplicate with an error of ±2.00 % and carbon balances ranged from 97 to 102 %.
Reactions in this study were carried out at six temperatures ranging from 240 to 440 °C.All catalysts showed an increase in activity with an increase in reaction temperature.Fig. 6a shows the conversion and selectivity as a function of temperature using the 5.0 wt.% VPO catalysts.
For the VPO/HAp catalyst, the conversion of the substrate increased gradually from 30 % at 240 °C to 69 % at 360 °C.There was no significant increase in the conversion above 360 °C.For the VPO/Co-HAp, although the maximum conversion was obtained at 440 °C, the selectivity to maleic anhydride was at its highest (50 %) at 360 °C for a conversion of 64 %.At higher temperatures, the MA and PA production decreased while the formation of carbon monoxide and carbon dioxide increased.Similar patterns of activity and selectivity were displayed by the 7.5 wt.% VPO-supported catalysts at the same flow rate (Fig. 6b).The effect of reaction temperature on n-pentane conversion and the product yields over 'non-equilibrated' and 'equilibrated' VPO catalysts showed that the yield in maleic anhydride increases from 280 to 400 °C, leveling at 370 °C, while the phthalic anhydride yield reaches a maximum at 350 °C and then decreases with an increase in the formation of the carbon oxides 39 .Similar trends were observed by Bignardi et al. 40 , who showed that the selectivity towards maleic anhydride increased with an increase in the reaction temperature but the selectivity towards phthalic anhydride decreased.
Weight percentages of 2.5, 5.0, 7.5, 10.0 and 15.0 % VPO were loaded on the hydroxyapatite supports.A marginal difference in the activity of HAp and Co-HAp as supports for VPO was observed.This is due to the easily reducible nature of VPO/Co-Hap which is evident in the TPR profile.The best performing catalysts were the 5.0 and 7.5 wt.% VPO loadings for a reaction temperature of 360 °C with GHSV values of 1900 and 2300 h -1 as reflected in Fig. 7.For the 5VPO/HAp catalyst, selectivity towards maleic anhydride and phthalic anhydride was 49 and 11 %, respectively, for a 68 % conversion of n-pentane with the GHSV fixed at 1900 h -1 .The selectivity towards these products improved marginally when the GHSV was set at 2300 h -1 .Using the 5VPO/Co-HAp catalyst, the selectivity towards maleic anhydride and phthalic anhydride was 52 and 22 %, respectively, for a 64 % conversion of n-pentane.However, there was a RESEARCH ARTICLE S. Singh, 5 S. Afr.J. Chem., 2016, 69, 1-8, <http://journals.sabinet.co.za/sajchem/>.slight decrease in the amounts of maleic anhydride and phthalic anhydride when the GHSV was set at 2300 h -1 .There was a significant increase in the activity in terms of conversion of n-pentane, but a slight decrease in the formation of maleic anhydride and phthalic anhydride with an increase in the VPO loadings on both the supports.This could be attributed to the non-crystalline nature of the catalyst surface due to clustering of the VPO at higher loadings leading to a decrease in V +4 sites.Lower yields of maleic anhydride and phthalic anhydide at higher VPO loadings could also be the result of an increase in the total metal content of the catalyst.
At higher VPO loadings, both the catalysts showed the similar pattern in conversion but different patterns in the selectivity.The similar trend in conversion is due to a similar reduction pattern observed in TPR and the difference in the selectivity is related to the Ca/P and Co/P ratios.Also, phosphorus stabilizes the active +4 valence state of vanadium and influences its oxidation properties 3,5,9 .In the supported VPO catalysts, phosphorus is present in the hydroxyapatite and VPO and there is a substantial increase in the amount of phosphorus at higher VPO loadings.

Proposed Mechanistic Pathway for the Formation of 1,3-Pentadiene
In the mechanism for the formation of MA and PA from n-pentane, the dimeric active site in VPP undergoes possibly a one-or two-electron oxidation by the adsorbed dioxygen or atomic oxygen and is attached to the hydroxyapatite support through the formation of a metal-oxygen bond.The presence of 1,3-pentadiene in the product stream together with ratios of the anhydride production suggests that both MA and PA forms via a common intermediate, a diene.
The initial activation of n-pentane occurs by hydrogen abstraction from a methylene group forming a hydroperoxy moiety bound to the surface and an alkoxy group on the adjacent vanadyl group.The hydroperoxy group then abstracts a hydrogen atom from a-position in the alkoxy group to produce the surface bound species, releasing a water molecule.The surface bound moiety gets protonated forming a carbocation species and the catalyst is regenerated.The loss of a proton and water molecule leads to the formation of pentadiene (Scheme 1).The pentadiene is oxidized to MA, or transformed to cyclopentadiene RESEARCH ARTICLE S. Singh, 6 S. Afr.J. Chem., 2016, 69, 1-8, <http://journals.sabinet.co.za/sajchem/>.

Scheme 1
Mechanism for the formation of the carbocation and diene.which dimerizes to a cyclic template and is finally oxidized to PA [41][42][43][44] .

Conclusion
Stoichiometric HAp and non-stoichiometric Co-HAp were successfully synthesized.Blank reactions using only HAp and Co-HAp resulted in conversions of the feed up to about 10 %, which indicate that the support itself has some activity.Total catalytic activity obtained was the resultant of the activity of the active component, the support and the active component and support interactions.Hydroxyapatite proved to be an efficient support for VPO in this study.Change in the morphology of the materials was negligible, even though the catalysts were subjected to temperatures up to 440 °C for a period of 160 hours.
The phases responsible for the activity during hydrocarbon oxidation is the vanadyl pyrophosphate, ((VO) 2 P 2 O 7 ) phase.There are also contributions from active sites within crystalline vanadyl(V) orthophosphate and the V V /V IV dimeric species of vanadyl phosphate.From XRD, TPR and oxidation state determination, the +4 oxidation state of V in (VO) 2 P 2 O 7 was more influential in the performance of the catalysts.
Optimum yields of up to 50 % MA and 24 % PA were obtained during the activation of n-pentane.In addition to the formation of the anhydrides, a number of parallel reactions took place resulting in the formation of compounds such as 1,3-pentadiene, 2-cyclopenten-1-one, furfural, furaldehyde, 1,4-dioxane-2,3-diol and 2,5-hexadione.

Figure 7
Figure 7 Conversion and selectivity as a function of VPO loading at 360 °C; GHSV fixed at (a) 1900 h -1 and (b) 2300 h -1 .(All data points in triplicate with an error of ±2.00 %.).

Figure 5
Figure 5 Conversion and selectivity as a function of GHSV over the 5.0 wt.% VPO catalyst at 360 °C (All data points in triplicate with an error of ±2.00 %).
and the partial oxidation of methane over nickel-added strontium phosphate incorporating strontium

Table 1
. Small peaks were observed between 400 and 450 °C for both the catalysts due BET surface areas and bulk composition of the catalysts determined by ICP-OES for various VPO/HAp and VPO/Co-HAp catalysts.
a Obtained in duplicate with a standard deviation of ±1.00.b From ICP (obtained in triplicate with a standard deviation of ±0.02).c From ICP (obtained in triplicate with a standard deviation of ±0.005).