- Email: [email protected]
Performance of vanadium based catalyst in a membrane contactor for the benzene hydroxylation to phenol
Applied Catalysis A: General 417–418 (2012) 87–92
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Performance of vanadium based catalyst in a membrane contactor for the benzene hydroxylation to phenol Raffaele Molinari ∗ , Cristina Lavorato, Teresa Poerio Department of Chemical Engineering and Materials, - UdR INCA, University of Calabria, Via P. Bucci, 44/A, I-87036 Rende (CS), Italy
a r t i c l e
i n f o
Article history: Received 1 August 2011 Received in revised form 2 December 2011 Accepted 16 December 2011 Available online 27 December 2011 Keywords: Direct hydroxylation of benzene Phenol production Membrane reactor Vanadium catalyst
a b s t r a c t Phenol production through the direct hydroxylation of benzene with hydrogen peroxide, as oxidant, using a vanadium catalyst in a membrane reactor has been studied. The reaction was carried out in mild condition in a liquid–liquid biphasic system separated by a polypropylene membrane. This system showed high selectivity to phenol, minimizing its over-oxidation to over-oxygenated by-products. The influence of various reaction parameters such as the addition of hydrogen peroxide mode, catalyst and sulphuric acid amounts, temperature and reducing agent effects were investigated. The vanadium catalyst avoided tar formation in all the investigated experimental conditions compared to the previous system where an iron-based catalyst was used. Use of (C5 H8 O2 )2 VO as catalyst, 18 mmol of hydrogen peroxide as oxidant pumped for 4 h in the aqueous phase with the step-by-step feeding mode gave the best system performance in terms of yield (63.2%), selectivity to phenol (97.0%), and extraction quotient (76.4%). © 2011 Elsevier B.V. All rights reserved.
1. Introduction Phenol is an important intermediate for the synthesis of petrochemicals, agrochemicals, and plastics. The current worldwide capacity for phenol production is nearly 7 million metric tonnes per year [1]. Today, almost 95% of the worldwide phenol production is based on the “cumene process”. However, this process has some disadvantages: high and damaging ecological impact, an explosive intermediate to manage (cumene hydroperoxide), large amount of acetone as by-product and a multistep character, making it difficult to achieve high phenol yield compared to benzene [2]. The search for new routes for phenol production based on the one-step direct benzene oxidation became more intensive in the last decade but this reaction is little selective because phenol is more reactive than benzene and by-products occur [3–12]. To reduce the formation of by-products the membrane reactors (MRs) and, in particular, membrane contactors, can be employed. A membrane contactor permits the combination of membrane separation and catalytic reaction in one device [13–15]. The main advantage in using a membrane in this system is the separation of phenol from the reaction mixture permitting to obtain improvements of yield and selectivity limiting side catalytic reactions. In our research group a flat-sheet membrane contactor was used for the one-step oxidation of benzene to phenol [16–19] reducing by-products formation by employing a liquid–liquid membrane
∗ Corresponding author. Tel.: +39 0984 496699; fax: +39 0984 496655. E-mail address: [email protected] (R. Molinari). 0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2011.12.031
contactor with various iron catalysts and hydrogen peroxide as oxidant. In this system phenol that did not cross the membrane rapidly remained in the reacting ambient and reacted further generating over-oxidation products such as 1-4 benzoquinone, biphenyl as trace and tars (black solid). To reduce tar formation and enhance phenol recovery in the organic phase, different aspects were studied in our previous work [20]. Despite the improved results, the black solid formation was not avoided completely. The need to avoid tar formation can be ascribed to several reasons: (i) concentration polarization and fouling phenomena that decrease membrane performance, particularly in view of the continuous operating system; (ii) bad use of a reactant (benzene); (iii) environmental impact (tar to discharge). Then the choice of the catalytic system will be a compromise between high conversion to phenol and points (i)–(iii). To avoid tar formation, catalysts different from the iron-based ones should be used considering that heavy metal oxides are generally more active than noble metal catalysts [21]. Taking into account the important role played by vanadium complexes in oxidative reactions on biological systems for the detoxification of organisms, it was interesting to investigate the performance of a vanadium catalyst in the oxidative hydroxylation of benzene, replacing the iron catalysts previously investigated. Vanadium compounds successfully oxidize benzene in most of the reactions reported in literature using H2 O2 or oxygen as oxidants. For example, Ishida et al. [22] obtained good turnover numbers (TNs) of phenol production using a supported vanadium catalyst in mild conditions. High reaction yields were claimed by Barnhard and Hughes using vanadium complexes as catalyst and hydroquinones as reducing agents [23]. However, the concentration of
88
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
required hydroquinones was comparable or even higher than that of the substrate (benzene) and, consequently, yields with respect to the reductant were low [24]. Another reducing agent used in the benzene oxidation to phenol is ascorbic acid. The use of ascorbate in aqueous solution as reducing agent was first reported by Udenfriend et al. [25]. In particular, Tanarungsun et al. reported that ascorbic acid (AA), a water-soluble vitamin which reacts with H2 O2 [21], helped to increase benzene conversion by enhancing the decomposition of H2 O2 to hydroxyl radicals [27]. In this paper, vanadium as catalyst to oxidize benzene in a twophase membrane reactor, using hydrogen peroxide as oxidant, to try to avoid tar formation has been studied. To enhance phenol production and recovery in the organic phase, different aspects have been investigated: (i) hydrogen peroxide addition mode; (ii) amount of vanadyl acetylacetonate as catalyst; (iii) influence of pH; (iv) influence of temperature and (v) effects of reducing agents. 2. Experimental 2.1. Materials Benzene (C6 H6 , 99.8% purity) from Carlo Erba Reagenti was used both as substrate and as organic phase. Phenol (C6 H5 OH, 99.99% purity), benzoquinone (C6 H4 O2 , 99.9% purity), biphenyl (C12 H10 , 99.99% purity) from Sigma–Aldrich were used for analytical calibrations. Sulphuric acid (H2 SO4 , 96% purity, d = 1.84 g mL−1 ), from Carlo Erba Reagenti, was used to achieve acidic pH in the aqueous phase. Vanadyl acetyl-acetonate ((C5 H8 O2 )2 VO, 99.99% purity), from Sigma–Aldrich, was employed as catalyst. Hydrogen peroxide (H2 O2 , 30%, w/w solution in water) from Sigma and oxygen (4.8 purity) from Pirossigeno, were used as oxidants. Zinc powder particle size lower than 150 m (Zn, MW 65.39 g mol−1 , purity 99%) from Riedel-deHaën, and ascorbic acid (C6 H8 O6 , MW = 176.13 g mol−1 , 99.5% purity) from Fluka Garantie, were used as reducing agents. Sodium thiosulfate (Na2 S2 O3 , MW = 158.11 g mol−1 , 99.9% purity) and potassium iodide (KI, MW = 166 g mol−1 , 99% purity) from Sigma Aldrich, were used for hydrogen peroxide titration with the iodometric method. Ammonium molybdate ((NH4 )6 Mo7 O24 , MW = 1235.86 g mol−1 , 99% purity) from Sigma Aldrich was used as a catalyst in the iodometric method. Starch (C12 H22 O11 , MW = 342.30 g mol−1 ), was used as an indicator in the iodometric method. Di-etiletere (C4 H10 O, MW = 74.124 g mol−1 , 99.8% purity) from Carlo Erba was used for the extraction of the aqueous solution. Hydrophobic microfiltration polypropylene porous membrane (Accurel, manufactured by Membrana, thickness 142 m; pore size 0.2 m; porosity 70%) was used in the membrane contactor. Ultrapure water was obtained from Milli-Q equipment by Millipore. 2.2. Apparatus and methods The experimental tests were conducted in a biphasic membrane contactor, immersed in a thermostatic bath. It was constituted by two compartment cells containing two phases (each one with a volume of 140 mL) separated by a flat sheet polypropylene membrane, with an exposed membrane surface area of 28.3 cm2 [17]. The two phases consist of an aqueous phase, containing the dissolved catalyst and hydrogen peroxide (oxidant), and an organic phase containing initially only benzene (reactant and solvent). Despite the use of a homogeneous catalytic system we did not immobilize
the catalyst but it was confined in the aqueous phase by means of the membrane thus avoiding the problems of the heterogeneous catalysts. This reaction (process), although can became continuous, was a batch-type only for testing the catalytic performance. The identification of the oxidation products was performed analysing the solutions by means of a gas chromatograph mass spectrometer (GC–MS QP2010S) from Shimadzu using an equity 5 column. Organics in the aqueous phases were analyzed after quantitative extraction with diethyl ether at the end of each catalytic run (270 min). The amount of hydrogen peroxide was determined at the end of the reaction using the iodometric method. Iodometry is one of the most important redox titration methods. Iodine reacts directly, fast and quantitatively with many organic and inorganic substances. Thanks to its relatively low pH independent redox potential, and reversibility of the iodine/iodide reaction, iodometry can be used both to determine amount of reducing agents (by direct titration with iodine) and of oxidizing agents (by titration of iodine with thiosulfate). In all cases the same simple and reliable method of end point detection, based on blue starch complex, can be used. A pH meter (WTW Inolab Terminal Level 3) with a glass pH electrode SenTix 81 (WTW) was used for pH measurements. The results of experimental tests have been elaborated using the following parameters: • Yield% = (Ph/mmol of hydrogen peroxide reacted) × 100; where Ph is the sum of the mol number of phenol in both phases. • Extraction quotient percentage, QE % = (norg /ntot ) × 100; where norg is the mol number of phenol in the organic phase and ntot is the sum of the mol number of phenol in the organic and aqueous phases. We considered this parameter and not the “distribution coefficient” because the system studied in this work did not reach the equilibrium condition. • Phenol selectivity (%) = [Ph/(Bq + Ph + Biph)] × 100; where Ph, Bq and Biph are the mol numbers of phenol, benzoquinone and biphenyl respectively detected in the organic phase. • Conversion of hydrogen peroxide (%) = (mmol of hydrogen peroxide consumed/mmol hydrogen peroxide initial) × 100. The amount of hydrogen peroxide reacted was calculated by subtracting the initial 18 mmol of oxidant to those measured at the end of the reaction using the iodometric method. • Overall produced phenol (OPP) = sum of the phenol amount (mg) in both phases. • Overall turnover number (OTN) = mmol oxidation products/mmol catalyst. • Phenol turnover number (PTN) = mmol phenol/mmol catalyst. • Flux (Jorg Ph ), variation of mmol of phenol in the organic phase per unit membrane area and time (mmol h−1 m−2 ). 3. Results and discussion 3.1. Feeding methods of hydrogen peroxide From our previous work [16] it was observed that the feeding method of H2 O2 is crucial for maximizing the amount of this reagent converted in the benzene oxidation to phenol. Thus, also for the new vanadium based catalyst the addition mode of the oxidant was studied. A constant amount of hydrogen peroxide (18 mmol) was added in two ways: (i) one-step mode, that means 18 mmol of oxidant added at the start of the experimental run; (ii) step-bystep mode, that means adding the H2 O2 slowly in a time of 4 h at intervals of 5 min (0.37 mmol per step). The results, reported in Table 1, show that the one-step addition mode of the oxidant gave a very low conversion of hydrogen peroxide (6.6%), low OTN and PTN and a significant low amount
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
89
Table 1 Yield (%), overall produced phenol (OPP), selectivity (%), conversion of H2 O2 (%), overall turnover number (OTN), phenol turnover number (PTN) and extraction percentage (Q%) by changing the hydrogen peroxide addition mode. H2 O2 feeding mode
Yield (%)
OPP (mg)
Selectivity (%)
Conversion of H2 O2 (%)
OTN
PTN
Q%
One step Step by step
0.29 8.2
8.0 136.5
88.6 99.4
6.6 99.6
0.3 3.7
0.2 3.6
60.8 37.1
Operative conditions: H2 O2 = 18 mmol, T = 35 ◦ C, catalyst = 0.4 mmol, pH 2.5, run time = 270 min.
Fig. 1. Phenol concentration in the organic phase vs. time by changing the hydrogen peroxide addition mode. Operative conditions: H2 O2 = 18 mmol, T = 35 ◦ C, catalyst = 0.4 mmol, pH 2.5.
In the step-by-step feeding mode, the amount of oxidant, fed every 5 min, is equal to 0.37 mmol thus, in these conditions, the catalyst has the ability to regenerate the oxidation state +4 because the competition between benzene and H2 O2 for • OH is reduced. Furthermore, the high value of selectivity to phenol reported in Table 1, using the step-by-step mode, is obtained thanks to the phenol extraction from the aqueous phase-membrane interface to the organic phase thus reducing its further contact with the catalyst which is soluble in the aqueous phase. Phenol extraction is promoted by the acidic pH of the aqueous phase because its nonionized form is more soluble in the organic phase. In the one-step mode the loss of selectivity was caused by the enhanced reactivity of the system (high • OH concentration) that oxidized the phenol produced further before its transport in the organic phase. 3.2. Amount of the catalyst-vanadyl acetylacetonate
of overall produced phenol (8.0 mg) compared to the step-by-step feeding mode. Observing the trend of the reaction course in the time (Fig. 1) it can be seen a very low value of maximum phenol concentration in the organic phase which remains constant at 120 min in the case of one-step mode. For the step-by-step feeding mode the concentration increases with an exponential trend reaching a significant value (500 mg/L) at 270 min and an overall phenol production of 136.5 mg (Table 1). The different trend can be explained considering the stoichiometry of the overall redox reaction, below reported, which takes place in the aqueous phase:
Some tests with different amounts of the catalyst vanadylacetylacetonate were carried out (Fig. 3) to study the performance of the overall system. The results show that increasing the amount of catalyst increases the concentration of produced phenol. The low concentration, obtained with 0.2 mmol of catalyst, is probably caused by its deactivation. Indeed, its amount is less than that of the oxidant added in each step every 5 min (0.37 mmol) which is almost double of the amount of catalyst present in the reaction ambient. In this condition the hydroxide radical can react with itself or, more likely, with H2 O2 to form the peroxyl radical which reacts with the vanadium(IV) causing its deactivation as shown in the following simplified reaction scheme:
H2 O2 + e− + H+ → H2 O + OH•
• OH
VO2+ + H2 O → VO2 + + 2H+ + e−
HO2 · + V4+ → V5+ + HO2 −
H2 O2 + VO2+ → OH• + VO2 + + H+ When the 18 mmol of hydrogen peroxide is present in the beginning of the reaction, the catalyst amount (0.4 mmol) is less than the oxidant. This high ratio mmol oxidant/mmol catalyst causes the transition of all the vanadium present in solution from the oxidation state of +4 to +5 (Fig. 2) inhibiting the catalyst regeneration. Indeed, the high concentration of generated • OH gives a higher probability to react with itself or with H2 O2 (see Section 3.2) instead of reacting with benzene and regenerating the catalyst.
Fig. 2. Scheme of catalyst regeneration caused by the conversion of benzene to phenol.
+ H2 O2 → HO2 · + H2 O
The selectivity has a maximum value for the amount of catalyst equal to 0.4 mmol. Indeed, at higher concentrations of catalyst a reduction of selectivity is observed (Fig. 4). This behavior is caused by the increased reactivity of the system which gives further oxidation of phenol with formation of greater amount of by-products. Concerning OTN and PTN the best results were obtained using an amount of catalyst of 0.4 mmol too (Fig. 5). Using a double
Fig. 3. Phenol concentration in the organic phase vs. the time at various amount of catalyst. Operative condition: H2 O2 = 18 mmol added step by step, T = 35 ◦ C, pH 2.5.
90
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
Fig. 4. Comparison of selectivity% using different amount of catalyst. Operative condition: H2 O2 = 18 mmol added step by step, T = 35 ◦ C, pH 2.5.
Fig. 6. Phenol concentration in the organic phase vs. time at 35, 45, 55 ◦ C. Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8.
peroxide, therefore it allows to be used as oxidant for the reaction [26]. The high value of extraction quotient (Q%) can also be attributed to the pH value. Indeed, at pH 1.4 the phenol is not dissociated in the phenolate anion and therefore it is transported in the organic phase across the membrane. A low pH allows to increase phenol yield and extraction but decreases phenol production as it is evidenced in Table 2 from the values of PTN and OPP, thus a pH 1.8 was chosen in the successive tests.
3.4. Influence of temperature
Fig. 5. Overall turnover number (OTN) and phenol turnover number (PTN) using different amount of catalyst. Operative conditions: H2 O2 = 18 mmol added step by step, T = 35 ◦ C, pH 2.5.
quantity of catalyst (0.8 mmol) the amount of produced phenol was not doubled but it was lowered because of the higher reactivity which increased the amount of phenol converted to by-products. Thus, the amount of catalyst giving the best performance, both in terms of selectivity and oxidizing power, is 0.4 mmol which means a concentration of 2.85 mM in the aqueous solution. 3.3. Influence of pH The pH influences the catalyst reactivity, thus some tests were carried out at different pH by adding in solution three different amount of sulphuric acid. The results, reported in Table 2, evidence a yield% and a phenol recovery Q% of 63.2 and 76.4, respectively, at pH 1.4. The high value of yield% (63.2) is attributed to the stability of the hydrogen peroxide in the more acidic media. Indeed, different studies report that a very acidic environment inhibits the decomposition of hydrogen Table 2 Yield (%), selectivity (%), conversion of H2 O2 (%), OTN, PTN, extraction (Q%) and OPP by changing the pH values.
The effect of temperature was studied by means of catalytic tests conducted at three different temperatures (35, 45 and 55 ◦ C). In Fig. 6, it can be observed that phenol concentration in the organic phase in the time at temperatures of 35 and 45 ◦ C are practically the same while a different trend at 55 ◦ C is observed. Indeed at 55 ◦ C, phenol recovery in the organic phase reaches a maximum after 150 min from the beginning of the reaction and then probably a deactivation of the catalyst in the aqueous phase happens. The deactivation was observed clearly during the reaction considering that colors of the different oxidation states of vanadium are: +2 (lilac), +3 (green), +4 (blue) and +5 (yellow). When the reaction was conducted at 55 ◦ C the color of the aqueous phase became yellow. In Table 3 the result of tests, conducted at the three temperatures, show that the values of OTN and PTN decrease with increasing temperature, although, the phenol diffusion from the aqueous phase to the organic phase increases at 45 ◦ C (highest Q% value). The selectivity increases with increasing temperature, while the conversion of H2 O2 decreases significantly at 55 ◦ C only. This performance is determined by the deactivation of the catalyst which is no longer able to catalyze the reaction and consume H2 O2 . From these data a temperature of 35 ◦ C can be chosen as the best performance in terms of conversion of H2 O2 , OTN and PTN.
Table 3 Yield (%), conversion of H2 O2 (%), selectivity (%), OTN, PTN and Q% by changing temperature.
pH
Yield%
Selectivity (%)
Conversion of H2 O2 (%)
OTN
PTN
Q%
OPP
T
Yield (%)
Conversion of H2 O2 (%)
Selectivity (%)
OTN
PTN
Q%
1.4 1.8 2.5
63.2 15.4 8.2
97.0 97.0 99.4
18.3 99.6 98.4
5.3 7.0 3.7
5.2 6.9 3.6
76.4 53.8 37.1
196.1 260.3 136.5
35 ◦ C 45 ◦ C 55 ◦ C
8.2 7.7 3.5
99.6 99.6 21.4
96.9 97.2 98.3
7.0 5.8 3.0
6.9 5.7 3.0
53.8 61.6 53.9
Operative conditions: time = 270 min.
H2 O2 = 18 mmol,
T = 35 ◦ C,
catalyst = 0.4 mmol,
run
Operative conditions: time = 270 min.
H2 O2 = 18 mmol,
catalyst = 0.4 mmol,
pH
1.8,
run
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
Fig. 7. Phenol concentration in the organic phase vs. time adding ascorbic acid (0.4 mmol). Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C.
91
Fig. 8. Phenol concentration in the organic phase vs. time adding different amounts of ascorbic acid. Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C.
3.5. Effects of reducing agents Following the literature reported in the introduction, some reducing agents have been studied to evaluate their performance in benzene hydroxylation by using the vanadium catalyst. Two known reducing agents such as ascorbic acid and zinc have been chosen. Ascorbic acid (AA) is a potent antioxidant in some in vitro models and prevents oxidation of a variety of biomolecules [21,25], thus it could prevent further oxidation of phenol The first step in the oxidation of AA reports the formation of dehydroascorbic acid (DHA) with the loss of hydrogen from carbons 2 and 3 through a radical intermediate. This reversible reaction is thought to be primarily responsible for the antioxidant properties attributed to AA. Zinc is a good reducing agent having relatively small ionization energies and low electro-negativities. These reducing agents were chosen because their reduction potential is less than that of vanadium dioxide thus they should be able to give electrons to vanadium catalyst and promote its regeneration by conversion of the oxidation state from +5 to +4: VO2 + + 2H+ + e− VO2+ + H2 O
E ◦ = +1.00 V
Indeed, the values of reduction potential of ascorbic acid and zinc are +0.28 and −0.76 V, respectively. Because the reducing power of ascorbic acid is such that it can reduce oxygen from air to hydrogen peroxide, preliminary tests without adding hydrogen peroxide were performed. The results showed a very low phenol concentration (Fig. 7) thus, in the experimental condition investigated, only a small amount of hydrogen peroxide formed from ascorbic acid can be hypothesized. This amount is negligible compared to the amount of H2 O2 added in the aqueous phase at the beginning of the reaction. Results of the catalytic tests carried out by adding the 18 mmol of H2 O2 with different amounts of ascorbic acid and zinc are reported in Figs. 8 and 9, respectively, and Table 4. The results indicate that the reducing agents have negative effects on the reaction. Indeed, considering the catalytic tests carried out using ascorbic acid, it is possible to hypothesize that it acts as pro-oxidant rather than as a reductant thus decreasing the selectivity of the reaction in all tests.
Fig. 9. Phenol concentration in the organic phase vs. time adding different amounts of zinc. Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C.
It has antioxidant activity when it reduces oxidizing substances, such as hydrogen peroxide, but it can also reduce metal ions which, in presence of H2 O2 leads to the generation of free oxidizing radical through the Fenton reaction: 2Mn+1 + Ascorbate → 2Mn+ + Dehydroascorbate
2Mn+ + 2H2 O2 → 2Mn+1 + 2OH• + 2OH− The metal ion in these reactions can be reduced, oxidized, and then re-reduced again, in a process called redox cycling that can generate reactive hydroxide radical reducing the selectivity of the reaction. As a result, the selectivity decreases from 97%, without ascorbic acid, to 89.3% by increasing the amount of ascorbic acid till 4.5 mmol (see Table 4). According to Tanarungsun et al. [27] too much ascorbic acid gave no yield improvement because of lower phenol selectivity. Using the zinc as reducing agent, although the results are little better than those with ascorbic acid, they show a not improved performance of the catalytic system (see Table 4).
Table 4 Yield%, conversion of H2 O2 (%), selectivity (%), OTN, PTN and extraction (Q%) with different amount of zinc and ascorbic acid as reducing agents. Reducing agent and amounts:
Yield%
Conversion of H2 O2 (%)
Selectivity (%)
OTN
PTN
Q%
Without reducing agent Zinc, 0.4 mmol Zinc, 1.2 mmol Ascorbic acid, 0.4 mmol Ascorbic acid, 0.8 mmol Ascorbic acid, 4.5 mmol
15.4 13.2 20.7 5.5 8.4 5.5
99.6 98.6 65.0 60.3 49.4 72.7
97.0 99.4 99.3 93.6 91.3 89.3
7.0 5.9 6.1 3.0 3.9 4.5
6.9 5.8 6.1 2.5 3.2 3.6
53.8 42.5 67.8 40.3 42.3 50.5
Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C, run time = 270 min.
92
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
Table 5 Comparison of OPP, selectivity and amount of black solid for the different used catalysts. Type of catalysta
OPP (mg)
Selectivity (%)
Amount of black solid (g)
Black solid/OPP (g/g)
Iron(0)–Fe (*) Iron(II) –FeSO4 (*) Vanadium(IV) – (C5 H8 O2 )2 VO (**)
324 1016 260
94 96 97
0.13 0.35 0
0.403 0.344 0
(*) pH 2.8 using acetic acid, (**) pH 1.8 using sulphuric acid. a Operative conditions: T = 35 ◦ C, H2 O2 = 18 mmol, catalyst = 0.4 mmol, run time = 270 min.
3.6. Comparison of the vanadium catalyst with iron-based catalysts In recent years our research was addressed to identify a catalytic system able to avoid the black solid formation in the aqueous phase. The comparison between the best results obtained using the vanadium catalyst and the different types of iron catalysts, studied in the previous work [20], is reported (Table 5). It can be observed that the iron catalyst is more active in terms of overall phenol production. In particular, the iron(II) catalyst, with 1016 mg of produced phenol and a selectivity of 96%, could be the “chosen catalyst” but it produces a large amount of tar (0.35 g, which means 0.344 g tar/g OPP). Tar has a particulate form and it must be avoided as much is possible in membrane processes because of polarization and fouling phenomena that reduce the performance of the system (e.g. phenol transport through the membrane in the organic phase). Using the vanadium(IV) catalyst the black solid formation was not observed in the experimental conditions investigated, thus it seems more interesting despite the overall phenol production is a quarter compared to iron(II). The vanadium-based catalyst can offer the possibility to carry out the hydroxylation reaction in a continuous membrane reactor with less problems and with a better use of a reactant (benzene). 4. Conclusion The study, using a vanadium-based catalyst, in a membrane reactor, for phenol production through the direct hydroxylation of benzene with hydrogen peroxide, gave interesting results for application purposes. This system showed high selectivity to phenol, minimizing its further oxidation to over-oxygenated by-products. The influence of various reaction parameters such as the addition of hydrogen peroxide mode, catalyst and sulphuric acid amounts, temperature and reducing agent effects were investigated. The vanadium catalyst avoided tar formation in all the experimental conditions investigated. The best system performance, determined
as yield (63.2%), selectivity to phenol (97.0%), and extraction quotient (76.4%), was obtained using (C5 H8 O2 )2 VO as catalyst, without reducing agent, and adding 18 mmol of hydrogen peroxide as oxidant, pumped step by step in the aqueous phase in 4 h. This finding should facilitate the use of a continuous membrane contactor to run the catalytic system. References [1] Y. Liu, K. Murata, M. Inaba, J. Mol. Catal. A 256 (2006) 247–255. [2] S. Niwa, M. Eswaramoorthy, J. Nair, A. Raj, N. Itoh, H. Shoji, T. Namba, F. Mizuka, Science 295 (2002) 105–107. [3] G. Bellussi, C. Perego, CATTECH 4 (2000) 4–16. [4] G.I. Panov, CATTECH 4 (2000) 18–32. [5] D. Bianchi, L. Balducci, R. Bortolo, R. D’Aloisio, M. Ricci, G. Span, R. Tassinari, C. Tonini, R. Ungarelli, Adv. Synth. Catal. 349 (2007) 979–986. [6] R. Molinari, T. Poerio, Asia-Pac. J. Chem. Eng. 5 (2010) 191–206. [7] N. Itoh, S. Niwa, F. Mizukami, T. Lnoue, A. Igarashi, T. Namba, Catal. Commun. 4 (2003) 243–246. [8] W. Laufer, W.F. Helderich, Chem. Commun. (2002) 1684–1685. [9] D. Bianchi, R. Bortolo, R. Tassinari, M. Ricci, R. Vignola, Angew. Chem. 39 (2000) 4321–4323. [10] R. Cai, S. Song, B. Ji, W. Yang, G. Sun, Q. Xin, Catal. Today 104 (2005) 200–204. [11] X.B. Wang, Y. Guo, X.F. Zhang, H. Liu, J. Wang, K.L. Yeung, Catal. Today 156 (2010) 288–294. [12] X.B. Wang, X.F. Zhang, H. Liu, K.L. Yeung, J.Q. Wang, Chem. Eng. J. 156 (2010) 562–570. [13] J. Coronas, J. Santamaria, Catal. Today 51 (1999) 377–389. [14] J.N. Armor, J. Membr. Sci. 147 (1998) 217–233. [15] G. Saracco, V. Specchia, Catal. Rev. Sci. Eng. 36 (1994) 305–384. [16] R. Molinari, T. Poerio, P. Argurio, Catal. Today 118 (2006) 52–56. [17] R. Molinari, T. Poerio, P. Argurio, Desalination 200 (2006) 673–675. [18] R. Molinari, T. Poerio, P. Argurio, Italy Patent no. CZ2006A000029 (2006). [19] R. Molinari, T. Poerio, Appl. Catal. A 358 (2009) 119–128. [20] R. Molinari, T. Poerio, Appl. Catal. A 393 (2011) 340–347. [21] E. Battistel, R. Tassinari, M. Fornaroli, L. Bonoldi, J. Mol. Catal. A 202 (2003) 107–115. [22] M.-a. Ishida, Y. Masumoto, R. Hamada, S. Nishiyama, S. Tsuruya, M. Masai, J. Chem. Soc. Perkin Trans. 2 (1999) 847–853. [23] T.M. Barnhard, A.W. Hughes, US Patent 5,912,391 (1999). [24] R.S.G. Ferreira, P.G.P. de Oliveira, F.B. Noronha, Appl. Catal. B 50 (2004) 243–249. [25] S. Udenfriend, C.T. Clark, J. Axelrod, B.B. Brodie, J. Biol. Chem. 206 (1954) 731–739. [26] D. Mantzavinos, Trans. IChemE 81 (2003) 99–106. [27] G. Tanarungsun, W. Kiatkittipong, S. Assabumrungrat, H. Yamada, T. Tagawa, P. Praserthdam, J. Ind. Eng. Chem. 13 (2007) 444–451.
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Performance of vanadium based catalyst in a membrane contactor for the benzene hydroxylation to phenol Raffaele Molinari ∗ , Cristina Lavorato, Teresa Poerio Department of Chemical Engineering and Materials, - UdR INCA, University of Calabria, Via P. Bucci, 44/A, I-87036 Rende (CS), Italy
a r t i c l e
i n f o
Article history: Received 1 August 2011 Received in revised form 2 December 2011 Accepted 16 December 2011 Available online 27 December 2011 Keywords: Direct hydroxylation of benzene Phenol production Membrane reactor Vanadium catalyst
a b s t r a c t Phenol production through the direct hydroxylation of benzene with hydrogen peroxide, as oxidant, using a vanadium catalyst in a membrane reactor has been studied. The reaction was carried out in mild condition in a liquid–liquid biphasic system separated by a polypropylene membrane. This system showed high selectivity to phenol, minimizing its over-oxidation to over-oxygenated by-products. The influence of various reaction parameters such as the addition of hydrogen peroxide mode, catalyst and sulphuric acid amounts, temperature and reducing agent effects were investigated. The vanadium catalyst avoided tar formation in all the investigated experimental conditions compared to the previous system where an iron-based catalyst was used. Use of (C5 H8 O2 )2 VO as catalyst, 18 mmol of hydrogen peroxide as oxidant pumped for 4 h in the aqueous phase with the step-by-step feeding mode gave the best system performance in terms of yield (63.2%), selectivity to phenol (97.0%), and extraction quotient (76.4%). © 2011 Elsevier B.V. All rights reserved.
1. Introduction Phenol is an important intermediate for the synthesis of petrochemicals, agrochemicals, and plastics. The current worldwide capacity for phenol production is nearly 7 million metric tonnes per year [1]. Today, almost 95% of the worldwide phenol production is based on the “cumene process”. However, this process has some disadvantages: high and damaging ecological impact, an explosive intermediate to manage (cumene hydroperoxide), large amount of acetone as by-product and a multistep character, making it difficult to achieve high phenol yield compared to benzene [2]. The search for new routes for phenol production based on the one-step direct benzene oxidation became more intensive in the last decade but this reaction is little selective because phenol is more reactive than benzene and by-products occur [3–12]. To reduce the formation of by-products the membrane reactors (MRs) and, in particular, membrane contactors, can be employed. A membrane contactor permits the combination of membrane separation and catalytic reaction in one device [13–15]. The main advantage in using a membrane in this system is the separation of phenol from the reaction mixture permitting to obtain improvements of yield and selectivity limiting side catalytic reactions. In our research group a flat-sheet membrane contactor was used for the one-step oxidation of benzene to phenol [16–19] reducing by-products formation by employing a liquid–liquid membrane
∗ Corresponding author. Tel.: +39 0984 496699; fax: +39 0984 496655. E-mail address: [email protected] (R. Molinari). 0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2011.12.031
contactor with various iron catalysts and hydrogen peroxide as oxidant. In this system phenol that did not cross the membrane rapidly remained in the reacting ambient and reacted further generating over-oxidation products such as 1-4 benzoquinone, biphenyl as trace and tars (black solid). To reduce tar formation and enhance phenol recovery in the organic phase, different aspects were studied in our previous work [20]. Despite the improved results, the black solid formation was not avoided completely. The need to avoid tar formation can be ascribed to several reasons: (i) concentration polarization and fouling phenomena that decrease membrane performance, particularly in view of the continuous operating system; (ii) bad use of a reactant (benzene); (iii) environmental impact (tar to discharge). Then the choice of the catalytic system will be a compromise between high conversion to phenol and points (i)–(iii). To avoid tar formation, catalysts different from the iron-based ones should be used considering that heavy metal oxides are generally more active than noble metal catalysts [21]. Taking into account the important role played by vanadium complexes in oxidative reactions on biological systems for the detoxification of organisms, it was interesting to investigate the performance of a vanadium catalyst in the oxidative hydroxylation of benzene, replacing the iron catalysts previously investigated. Vanadium compounds successfully oxidize benzene in most of the reactions reported in literature using H2 O2 or oxygen as oxidants. For example, Ishida et al. [22] obtained good turnover numbers (TNs) of phenol production using a supported vanadium catalyst in mild conditions. High reaction yields were claimed by Barnhard and Hughes using vanadium complexes as catalyst and hydroquinones as reducing agents [23]. However, the concentration of
88
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
required hydroquinones was comparable or even higher than that of the substrate (benzene) and, consequently, yields with respect to the reductant were low [24]. Another reducing agent used in the benzene oxidation to phenol is ascorbic acid. The use of ascorbate in aqueous solution as reducing agent was first reported by Udenfriend et al. [25]. In particular, Tanarungsun et al. reported that ascorbic acid (AA), a water-soluble vitamin which reacts with H2 O2 [21], helped to increase benzene conversion by enhancing the decomposition of H2 O2 to hydroxyl radicals [27]. In this paper, vanadium as catalyst to oxidize benzene in a twophase membrane reactor, using hydrogen peroxide as oxidant, to try to avoid tar formation has been studied. To enhance phenol production and recovery in the organic phase, different aspects have been investigated: (i) hydrogen peroxide addition mode; (ii) amount of vanadyl acetylacetonate as catalyst; (iii) influence of pH; (iv) influence of temperature and (v) effects of reducing agents. 2. Experimental 2.1. Materials Benzene (C6 H6 , 99.8% purity) from Carlo Erba Reagenti was used both as substrate and as organic phase. Phenol (C6 H5 OH, 99.99% purity), benzoquinone (C6 H4 O2 , 99.9% purity), biphenyl (C12 H10 , 99.99% purity) from Sigma–Aldrich were used for analytical calibrations. Sulphuric acid (H2 SO4 , 96% purity, d = 1.84 g mL−1 ), from Carlo Erba Reagenti, was used to achieve acidic pH in the aqueous phase. Vanadyl acetyl-acetonate ((C5 H8 O2 )2 VO, 99.99% purity), from Sigma–Aldrich, was employed as catalyst. Hydrogen peroxide (H2 O2 , 30%, w/w solution in water) from Sigma and oxygen (4.8 purity) from Pirossigeno, were used as oxidants. Zinc powder particle size lower than 150 m (Zn, MW 65.39 g mol−1 , purity 99%) from Riedel-deHaën, and ascorbic acid (C6 H8 O6 , MW = 176.13 g mol−1 , 99.5% purity) from Fluka Garantie, were used as reducing agents. Sodium thiosulfate (Na2 S2 O3 , MW = 158.11 g mol−1 , 99.9% purity) and potassium iodide (KI, MW = 166 g mol−1 , 99% purity) from Sigma Aldrich, were used for hydrogen peroxide titration with the iodometric method. Ammonium molybdate ((NH4 )6 Mo7 O24 , MW = 1235.86 g mol−1 , 99% purity) from Sigma Aldrich was used as a catalyst in the iodometric method. Starch (C12 H22 O11 , MW = 342.30 g mol−1 ), was used as an indicator in the iodometric method. Di-etiletere (C4 H10 O, MW = 74.124 g mol−1 , 99.8% purity) from Carlo Erba was used for the extraction of the aqueous solution. Hydrophobic microfiltration polypropylene porous membrane (Accurel, manufactured by Membrana, thickness 142 m; pore size 0.2 m; porosity 70%) was used in the membrane contactor. Ultrapure water was obtained from Milli-Q equipment by Millipore. 2.2. Apparatus and methods The experimental tests were conducted in a biphasic membrane contactor, immersed in a thermostatic bath. It was constituted by two compartment cells containing two phases (each one with a volume of 140 mL) separated by a flat sheet polypropylene membrane, with an exposed membrane surface area of 28.3 cm2 [17]. The two phases consist of an aqueous phase, containing the dissolved catalyst and hydrogen peroxide (oxidant), and an organic phase containing initially only benzene (reactant and solvent). Despite the use of a homogeneous catalytic system we did not immobilize
the catalyst but it was confined in the aqueous phase by means of the membrane thus avoiding the problems of the heterogeneous catalysts. This reaction (process), although can became continuous, was a batch-type only for testing the catalytic performance. The identification of the oxidation products was performed analysing the solutions by means of a gas chromatograph mass spectrometer (GC–MS QP2010S) from Shimadzu using an equity 5 column. Organics in the aqueous phases were analyzed after quantitative extraction with diethyl ether at the end of each catalytic run (270 min). The amount of hydrogen peroxide was determined at the end of the reaction using the iodometric method. Iodometry is one of the most important redox titration methods. Iodine reacts directly, fast and quantitatively with many organic and inorganic substances. Thanks to its relatively low pH independent redox potential, and reversibility of the iodine/iodide reaction, iodometry can be used both to determine amount of reducing agents (by direct titration with iodine) and of oxidizing agents (by titration of iodine with thiosulfate). In all cases the same simple and reliable method of end point detection, based on blue starch complex, can be used. A pH meter (WTW Inolab Terminal Level 3) with a glass pH electrode SenTix 81 (WTW) was used for pH measurements. The results of experimental tests have been elaborated using the following parameters: • Yield% = (Ph/mmol of hydrogen peroxide reacted) × 100; where Ph is the sum of the mol number of phenol in both phases. • Extraction quotient percentage, QE % = (norg /ntot ) × 100; where norg is the mol number of phenol in the organic phase and ntot is the sum of the mol number of phenol in the organic and aqueous phases. We considered this parameter and not the “distribution coefficient” because the system studied in this work did not reach the equilibrium condition. • Phenol selectivity (%) = [Ph/(Bq + Ph + Biph)] × 100; where Ph, Bq and Biph are the mol numbers of phenol, benzoquinone and biphenyl respectively detected in the organic phase. • Conversion of hydrogen peroxide (%) = (mmol of hydrogen peroxide consumed/mmol hydrogen peroxide initial) × 100. The amount of hydrogen peroxide reacted was calculated by subtracting the initial 18 mmol of oxidant to those measured at the end of the reaction using the iodometric method. • Overall produced phenol (OPP) = sum of the phenol amount (mg) in both phases. • Overall turnover number (OTN) = mmol oxidation products/mmol catalyst. • Phenol turnover number (PTN) = mmol phenol/mmol catalyst. • Flux (Jorg Ph ), variation of mmol of phenol in the organic phase per unit membrane area and time (mmol h−1 m−2 ). 3. Results and discussion 3.1. Feeding methods of hydrogen peroxide From our previous work [16] it was observed that the feeding method of H2 O2 is crucial for maximizing the amount of this reagent converted in the benzene oxidation to phenol. Thus, also for the new vanadium based catalyst the addition mode of the oxidant was studied. A constant amount of hydrogen peroxide (18 mmol) was added in two ways: (i) one-step mode, that means 18 mmol of oxidant added at the start of the experimental run; (ii) step-bystep mode, that means adding the H2 O2 slowly in a time of 4 h at intervals of 5 min (0.37 mmol per step). The results, reported in Table 1, show that the one-step addition mode of the oxidant gave a very low conversion of hydrogen peroxide (6.6%), low OTN and PTN and a significant low amount
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
89
Table 1 Yield (%), overall produced phenol (OPP), selectivity (%), conversion of H2 O2 (%), overall turnover number (OTN), phenol turnover number (PTN) and extraction percentage (Q%) by changing the hydrogen peroxide addition mode. H2 O2 feeding mode
Yield (%)
OPP (mg)
Selectivity (%)
Conversion of H2 O2 (%)
OTN
PTN
Q%
One step Step by step
0.29 8.2
8.0 136.5
88.6 99.4
6.6 99.6
0.3 3.7
0.2 3.6
60.8 37.1
Operative conditions: H2 O2 = 18 mmol, T = 35 ◦ C, catalyst = 0.4 mmol, pH 2.5, run time = 270 min.
Fig. 1. Phenol concentration in the organic phase vs. time by changing the hydrogen peroxide addition mode. Operative conditions: H2 O2 = 18 mmol, T = 35 ◦ C, catalyst = 0.4 mmol, pH 2.5.
In the step-by-step feeding mode, the amount of oxidant, fed every 5 min, is equal to 0.37 mmol thus, in these conditions, the catalyst has the ability to regenerate the oxidation state +4 because the competition between benzene and H2 O2 for • OH is reduced. Furthermore, the high value of selectivity to phenol reported in Table 1, using the step-by-step mode, is obtained thanks to the phenol extraction from the aqueous phase-membrane interface to the organic phase thus reducing its further contact with the catalyst which is soluble in the aqueous phase. Phenol extraction is promoted by the acidic pH of the aqueous phase because its nonionized form is more soluble in the organic phase. In the one-step mode the loss of selectivity was caused by the enhanced reactivity of the system (high • OH concentration) that oxidized the phenol produced further before its transport in the organic phase. 3.2. Amount of the catalyst-vanadyl acetylacetonate
of overall produced phenol (8.0 mg) compared to the step-by-step feeding mode. Observing the trend of the reaction course in the time (Fig. 1) it can be seen a very low value of maximum phenol concentration in the organic phase which remains constant at 120 min in the case of one-step mode. For the step-by-step feeding mode the concentration increases with an exponential trend reaching a significant value (500 mg/L) at 270 min and an overall phenol production of 136.5 mg (Table 1). The different trend can be explained considering the stoichiometry of the overall redox reaction, below reported, which takes place in the aqueous phase:
Some tests with different amounts of the catalyst vanadylacetylacetonate were carried out (Fig. 3) to study the performance of the overall system. The results show that increasing the amount of catalyst increases the concentration of produced phenol. The low concentration, obtained with 0.2 mmol of catalyst, is probably caused by its deactivation. Indeed, its amount is less than that of the oxidant added in each step every 5 min (0.37 mmol) which is almost double of the amount of catalyst present in the reaction ambient. In this condition the hydroxide radical can react with itself or, more likely, with H2 O2 to form the peroxyl radical which reacts with the vanadium(IV) causing its deactivation as shown in the following simplified reaction scheme:
H2 O2 + e− + H+ → H2 O + OH•
• OH
VO2+ + H2 O → VO2 + + 2H+ + e−
HO2 · + V4+ → V5+ + HO2 −
H2 O2 + VO2+ → OH• + VO2 + + H+ When the 18 mmol of hydrogen peroxide is present in the beginning of the reaction, the catalyst amount (0.4 mmol) is less than the oxidant. This high ratio mmol oxidant/mmol catalyst causes the transition of all the vanadium present in solution from the oxidation state of +4 to +5 (Fig. 2) inhibiting the catalyst regeneration. Indeed, the high concentration of generated • OH gives a higher probability to react with itself or with H2 O2 (see Section 3.2) instead of reacting with benzene and regenerating the catalyst.
Fig. 2. Scheme of catalyst regeneration caused by the conversion of benzene to phenol.
+ H2 O2 → HO2 · + H2 O
The selectivity has a maximum value for the amount of catalyst equal to 0.4 mmol. Indeed, at higher concentrations of catalyst a reduction of selectivity is observed (Fig. 4). This behavior is caused by the increased reactivity of the system which gives further oxidation of phenol with formation of greater amount of by-products. Concerning OTN and PTN the best results were obtained using an amount of catalyst of 0.4 mmol too (Fig. 5). Using a double
Fig. 3. Phenol concentration in the organic phase vs. the time at various amount of catalyst. Operative condition: H2 O2 = 18 mmol added step by step, T = 35 ◦ C, pH 2.5.
90
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
Fig. 4. Comparison of selectivity% using different amount of catalyst. Operative condition: H2 O2 = 18 mmol added step by step, T = 35 ◦ C, pH 2.5.
Fig. 6. Phenol concentration in the organic phase vs. time at 35, 45, 55 ◦ C. Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8.
peroxide, therefore it allows to be used as oxidant for the reaction [26]. The high value of extraction quotient (Q%) can also be attributed to the pH value. Indeed, at pH 1.4 the phenol is not dissociated in the phenolate anion and therefore it is transported in the organic phase across the membrane. A low pH allows to increase phenol yield and extraction but decreases phenol production as it is evidenced in Table 2 from the values of PTN and OPP, thus a pH 1.8 was chosen in the successive tests.
3.4. Influence of temperature
Fig. 5. Overall turnover number (OTN) and phenol turnover number (PTN) using different amount of catalyst. Operative conditions: H2 O2 = 18 mmol added step by step, T = 35 ◦ C, pH 2.5.
quantity of catalyst (0.8 mmol) the amount of produced phenol was not doubled but it was lowered because of the higher reactivity which increased the amount of phenol converted to by-products. Thus, the amount of catalyst giving the best performance, both in terms of selectivity and oxidizing power, is 0.4 mmol which means a concentration of 2.85 mM in the aqueous solution. 3.3. Influence of pH The pH influences the catalyst reactivity, thus some tests were carried out at different pH by adding in solution three different amount of sulphuric acid. The results, reported in Table 2, evidence a yield% and a phenol recovery Q% of 63.2 and 76.4, respectively, at pH 1.4. The high value of yield% (63.2) is attributed to the stability of the hydrogen peroxide in the more acidic media. Indeed, different studies report that a very acidic environment inhibits the decomposition of hydrogen Table 2 Yield (%), selectivity (%), conversion of H2 O2 (%), OTN, PTN, extraction (Q%) and OPP by changing the pH values.
The effect of temperature was studied by means of catalytic tests conducted at three different temperatures (35, 45 and 55 ◦ C). In Fig. 6, it can be observed that phenol concentration in the organic phase in the time at temperatures of 35 and 45 ◦ C are practically the same while a different trend at 55 ◦ C is observed. Indeed at 55 ◦ C, phenol recovery in the organic phase reaches a maximum after 150 min from the beginning of the reaction and then probably a deactivation of the catalyst in the aqueous phase happens. The deactivation was observed clearly during the reaction considering that colors of the different oxidation states of vanadium are: +2 (lilac), +3 (green), +4 (blue) and +5 (yellow). When the reaction was conducted at 55 ◦ C the color of the aqueous phase became yellow. In Table 3 the result of tests, conducted at the three temperatures, show that the values of OTN and PTN decrease with increasing temperature, although, the phenol diffusion from the aqueous phase to the organic phase increases at 45 ◦ C (highest Q% value). The selectivity increases with increasing temperature, while the conversion of H2 O2 decreases significantly at 55 ◦ C only. This performance is determined by the deactivation of the catalyst which is no longer able to catalyze the reaction and consume H2 O2 . From these data a temperature of 35 ◦ C can be chosen as the best performance in terms of conversion of H2 O2 , OTN and PTN.
Table 3 Yield (%), conversion of H2 O2 (%), selectivity (%), OTN, PTN and Q% by changing temperature.
pH
Yield%
Selectivity (%)
Conversion of H2 O2 (%)
OTN
PTN
Q%
OPP
T
Yield (%)
Conversion of H2 O2 (%)
Selectivity (%)
OTN
PTN
Q%
1.4 1.8 2.5
63.2 15.4 8.2
97.0 97.0 99.4
18.3 99.6 98.4
5.3 7.0 3.7
5.2 6.9 3.6
76.4 53.8 37.1
196.1 260.3 136.5
35 ◦ C 45 ◦ C 55 ◦ C
8.2 7.7 3.5
99.6 99.6 21.4
96.9 97.2 98.3
7.0 5.8 3.0
6.9 5.7 3.0
53.8 61.6 53.9
Operative conditions: time = 270 min.
H2 O2 = 18 mmol,
T = 35 ◦ C,
catalyst = 0.4 mmol,
run
Operative conditions: time = 270 min.
H2 O2 = 18 mmol,
catalyst = 0.4 mmol,
pH
1.8,
run
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
Fig. 7. Phenol concentration in the organic phase vs. time adding ascorbic acid (0.4 mmol). Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C.
91
Fig. 8. Phenol concentration in the organic phase vs. time adding different amounts of ascorbic acid. Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C.
3.5. Effects of reducing agents Following the literature reported in the introduction, some reducing agents have been studied to evaluate their performance in benzene hydroxylation by using the vanadium catalyst. Two known reducing agents such as ascorbic acid and zinc have been chosen. Ascorbic acid (AA) is a potent antioxidant in some in vitro models and prevents oxidation of a variety of biomolecules [21,25], thus it could prevent further oxidation of phenol The first step in the oxidation of AA reports the formation of dehydroascorbic acid (DHA) with the loss of hydrogen from carbons 2 and 3 through a radical intermediate. This reversible reaction is thought to be primarily responsible for the antioxidant properties attributed to AA. Zinc is a good reducing agent having relatively small ionization energies and low electro-negativities. These reducing agents were chosen because their reduction potential is less than that of vanadium dioxide thus they should be able to give electrons to vanadium catalyst and promote its regeneration by conversion of the oxidation state from +5 to +4: VO2 + + 2H+ + e− VO2+ + H2 O
E ◦ = +1.00 V
Indeed, the values of reduction potential of ascorbic acid and zinc are +0.28 and −0.76 V, respectively. Because the reducing power of ascorbic acid is such that it can reduce oxygen from air to hydrogen peroxide, preliminary tests without adding hydrogen peroxide were performed. The results showed a very low phenol concentration (Fig. 7) thus, in the experimental condition investigated, only a small amount of hydrogen peroxide formed from ascorbic acid can be hypothesized. This amount is negligible compared to the amount of H2 O2 added in the aqueous phase at the beginning of the reaction. Results of the catalytic tests carried out by adding the 18 mmol of H2 O2 with different amounts of ascorbic acid and zinc are reported in Figs. 8 and 9, respectively, and Table 4. The results indicate that the reducing agents have negative effects on the reaction. Indeed, considering the catalytic tests carried out using ascorbic acid, it is possible to hypothesize that it acts as pro-oxidant rather than as a reductant thus decreasing the selectivity of the reaction in all tests.
Fig. 9. Phenol concentration in the organic phase vs. time adding different amounts of zinc. Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C.
It has antioxidant activity when it reduces oxidizing substances, such as hydrogen peroxide, but it can also reduce metal ions which, in presence of H2 O2 leads to the generation of free oxidizing radical through the Fenton reaction: 2Mn+1 + Ascorbate → 2Mn+ + Dehydroascorbate
2Mn+ + 2H2 O2 → 2Mn+1 + 2OH• + 2OH− The metal ion in these reactions can be reduced, oxidized, and then re-reduced again, in a process called redox cycling that can generate reactive hydroxide radical reducing the selectivity of the reaction. As a result, the selectivity decreases from 97%, without ascorbic acid, to 89.3% by increasing the amount of ascorbic acid till 4.5 mmol (see Table 4). According to Tanarungsun et al. [27] too much ascorbic acid gave no yield improvement because of lower phenol selectivity. Using the zinc as reducing agent, although the results are little better than those with ascorbic acid, they show a not improved performance of the catalytic system (see Table 4).
Table 4 Yield%, conversion of H2 O2 (%), selectivity (%), OTN, PTN and extraction (Q%) with different amount of zinc and ascorbic acid as reducing agents. Reducing agent and amounts:
Yield%
Conversion of H2 O2 (%)
Selectivity (%)
OTN
PTN
Q%
Without reducing agent Zinc, 0.4 mmol Zinc, 1.2 mmol Ascorbic acid, 0.4 mmol Ascorbic acid, 0.8 mmol Ascorbic acid, 4.5 mmol
15.4 13.2 20.7 5.5 8.4 5.5
99.6 98.6 65.0 60.3 49.4 72.7
97.0 99.4 99.3 93.6 91.3 89.3
7.0 5.9 6.1 3.0 3.9 4.5
6.9 5.8 6.1 2.5 3.2 3.6
53.8 42.5 67.8 40.3 42.3 50.5
Operative conditions: H2 O2 = 18 mmol, catalyst = 0.4 mmol, pH 1.8, T = 35 ◦ C, run time = 270 min.
92
R. Molinari et al. / Applied Catalysis A: General 417–418 (2012) 87–92
Table 5 Comparison of OPP, selectivity and amount of black solid for the different used catalysts. Type of catalysta
OPP (mg)
Selectivity (%)
Amount of black solid (g)
Black solid/OPP (g/g)
Iron(0)–Fe (*) Iron(II) –FeSO4 (*) Vanadium(IV) – (C5 H8 O2 )2 VO (**)
324 1016 260
94 96 97
0.13 0.35 0
0.403 0.344 0
(*) pH 2.8 using acetic acid, (**) pH 1.8 using sulphuric acid. a Operative conditions: T = 35 ◦ C, H2 O2 = 18 mmol, catalyst = 0.4 mmol, run time = 270 min.
3.6. Comparison of the vanadium catalyst with iron-based catalysts In recent years our research was addressed to identify a catalytic system able to avoid the black solid formation in the aqueous phase. The comparison between the best results obtained using the vanadium catalyst and the different types of iron catalysts, studied in the previous work [20], is reported (Table 5). It can be observed that the iron catalyst is more active in terms of overall phenol production. In particular, the iron(II) catalyst, with 1016 mg of produced phenol and a selectivity of 96%, could be the “chosen catalyst” but it produces a large amount of tar (0.35 g, which means 0.344 g tar/g OPP). Tar has a particulate form and it must be avoided as much is possible in membrane processes because of polarization and fouling phenomena that reduce the performance of the system (e.g. phenol transport through the membrane in the organic phase). Using the vanadium(IV) catalyst the black solid formation was not observed in the experimental conditions investigated, thus it seems more interesting despite the overall phenol production is a quarter compared to iron(II). The vanadium-based catalyst can offer the possibility to carry out the hydroxylation reaction in a continuous membrane reactor with less problems and with a better use of a reactant (benzene). 4. Conclusion The study, using a vanadium-based catalyst, in a membrane reactor, for phenol production through the direct hydroxylation of benzene with hydrogen peroxide, gave interesting results for application purposes. This system showed high selectivity to phenol, minimizing its further oxidation to over-oxygenated by-products. The influence of various reaction parameters such as the addition of hydrogen peroxide mode, catalyst and sulphuric acid amounts, temperature and reducing agent effects were investigated. The vanadium catalyst avoided tar formation in all the experimental conditions investigated. The best system performance, determined
as yield (63.2%), selectivity to phenol (97.0%), and extraction quotient (76.4%), was obtained using (C5 H8 O2 )2 VO as catalyst, without reducing agent, and adding 18 mmol of hydrogen peroxide as oxidant, pumped step by step in the aqueous phase in 4 h. This finding should facilitate the use of a continuous membrane contactor to run the catalytic system. References [1] Y. Liu, K. Murata, M. Inaba, J. Mol. Catal. A 256 (2006) 247–255. [2] S. Niwa, M. Eswaramoorthy, J. Nair, A. Raj, N. Itoh, H. Shoji, T. Namba, F. Mizuka, Science 295 (2002) 105–107. [3] G. Bellussi, C. Perego, CATTECH 4 (2000) 4–16. [4] G.I. Panov, CATTECH 4 (2000) 18–32. [5] D. Bianchi, L. Balducci, R. Bortolo, R. D’Aloisio, M. Ricci, G. Span, R. Tassinari, C. Tonini, R. Ungarelli, Adv. Synth. Catal. 349 (2007) 979–986. [6] R. Molinari, T. Poerio, Asia-Pac. J. Chem. Eng. 5 (2010) 191–206. [7] N. Itoh, S. Niwa, F. Mizukami, T. Lnoue, A. Igarashi, T. Namba, Catal. Commun. 4 (2003) 243–246. [8] W. Laufer, W.F. Helderich, Chem. Commun. (2002) 1684–1685. [9] D. Bianchi, R. Bortolo, R. Tassinari, M. Ricci, R. Vignola, Angew. Chem. 39 (2000) 4321–4323. [10] R. Cai, S. Song, B. Ji, W. Yang, G. Sun, Q. Xin, Catal. Today 104 (2005) 200–204. [11] X.B. Wang, Y. Guo, X.F. Zhang, H. Liu, J. Wang, K.L. Yeung, Catal. Today 156 (2010) 288–294. [12] X.B. Wang, X.F. Zhang, H. Liu, K.L. Yeung, J.Q. Wang, Chem. Eng. J. 156 (2010) 562–570. [13] J. Coronas, J. Santamaria, Catal. Today 51 (1999) 377–389. [14] J.N. Armor, J. Membr. Sci. 147 (1998) 217–233. [15] G. Saracco, V. Specchia, Catal. Rev. Sci. Eng. 36 (1994) 305–384. [16] R. Molinari, T. Poerio, P. Argurio, Catal. Today 118 (2006) 52–56. [17] R. Molinari, T. Poerio, P. Argurio, Desalination 200 (2006) 673–675. [18] R. Molinari, T. Poerio, P. Argurio, Italy Patent no. CZ2006A000029 (2006). [19] R. Molinari, T. Poerio, Appl. Catal. A 358 (2009) 119–128. [20] R. Molinari, T. Poerio, Appl. Catal. A 393 (2011) 340–347. [21] E. Battistel, R. Tassinari, M. Fornaroli, L. Bonoldi, J. Mol. Catal. A 202 (2003) 107–115. [22] M.-a. Ishida, Y. Masumoto, R. Hamada, S. Nishiyama, S. Tsuruya, M. Masai, J. Chem. Soc. Perkin Trans. 2 (1999) 847–853. [23] T.M. Barnhard, A.W. Hughes, US Patent 5,912,391 (1999). [24] R.S.G. Ferreira, P.G.P. de Oliveira, F.B. Noronha, Appl. Catal. B 50 (2004) 243–249. [25] S. Udenfriend, C.T. Clark, J. Axelrod, B.B. Brodie, J. Biol. Chem. 206 (1954) 731–739. [26] D. Mantzavinos, Trans. IChemE 81 (2003) 99–106. [27] G. Tanarungsun, W. Kiatkittipong, S. Assabumrungrat, H. Yamada, T. Tagawa, P. Praserthdam, J. Ind. Eng. Chem. 13 (2007) 444–451.