Elsevier

Chemical Engineering Journal

Volumes 176–177, 1 December 2011, Pages 172-177
Chemical Engineering Journal

Continuous H2O2 direct synthesis over PdAu catalysts

https://doi.org/10.1016/j.cej.2011.05.073Get rights and content

Abstract

H2O2 direct synthesis over two bimetallic, palladium-gold catalysts based on sulfated ceria (PdAu-CeS) and on sulfated zirconia (PdAu-ZS) has been studied in a continuous, trickle-bed reactor, at −10 °C and 10 bar. Three different liquid flow rates and seven different gas flow rates were used. The combined effect of liquid and gas flow rates, with two different H2/O2 ratio (2/18 and 4/16) in the gas feed, has been studied, by independent variations. The highest H2/O2 ratio always increases production rate and selectivity. In the case of 4/16 ratio, PdAu-CeS catalyst shows an enhancement in selectivity only, while the increase in H2O2 production rate is not significant compared to the 2/18 ratio case. PdAu-ZS catalyst shows a remarkable gain in both selectivity and production rate when the 4/16 ratio in gas feeding is used, outperforming the PdAu-CeS catalyst. The enhancement of the catalytic activity is more pronounced in the case of PdAu-ZS, compared to PdAu-CeS, when the highest H2/O2 ratio is used. A selectivity up to 90% has been measured on PdAu-ZS with the higher H2/O2 ratio and larger liquid flow rates. The maximum production rate measured is 6 μmol/min (i.e. a productivity 0.18 mol H2O2 gPd−1 h−1) with 1 ml/min liquid flow rate and 2.7 ml/min gas flow rate again on PdAu-ZS.

Highlights

► A trickle bed reactor was successfully developed for H2O2 direct synthesis. ► PdAu-ZS catalyst shows better activity compared to PdAu-CeS catalyst. ► H2/O2 ratio in the gas has an important role on catalyst activity. ► Selectivity of 90% was achieved with PdAu-ZS catalyst.

Introduction

The direct synthesis of hydrogen peroxide by heterogeneous catalysis is emerging as a relevant alternative to the anthraquinone autoxidation (AO) process, in view of an integration with other industrial applications, such as paper bleaching, waste water treatment, textile manufacturing, and caprolactam and propylene oxide synthesis. The in situ production of H2O2 would reduce or eliminate the costs and hazards of transport and handling related to concentrated solutions, moreover the capital investment and operating costs are expected to be lower than those for the well known H2O2 manufacturing process (autoxidation of anthraquinone). The numerous steps included in the AO process as extraction, concentration and purification of the working solution, that embody the hydrogen peroxide, can be avoided. Direct synthesis will therefore result in a simpler process and plants. The direct reaction of hydrogen and oxygen to form hydrogen peroxide is in principle the best atom-efficient method for producing H2O2, but still insufficiently developed. The reaction is thermodynamically favored, but as shown in Fig. 1, the reaction of H2O2 hydrogenation and water formation can affect selectivity and production rate of hydrogen peroxide formation. One of the main problems concerning direct synthesis lies in the fact that the catalyst promotes both hydrogen peroxide formation and the unwanted reactions (Fig. 1).

Both academia and industry have been extensively investigating direct synthesis, with a large number of studies being continuosly published [1], [2], [3], [4], [5], [6], [7], [8]. However, the main effort in the direct synthesis is still focused on the synthesis of new catalysts [2], [9], [10], [11], [12], [13] that can reduce or avoid the parallel and consecutives reaction involved in the reaction to form H2O2 (Fig. 1). Investigations on reactor design and operation conditions are increasing in recent years [1], [4], [14], [15], [16], [17], [18], but the industrial breakthrough is still far away. The most important issue in hydrogen peroxide direct synthesis concerns the safety this is one of the reasons why microreactors and membranes reactors are emerging as important alternatives [15], [18].

Microreactors are expected to improve safety at high concentration of hydrogen and different H2/O2 ratios, even inside the presumed flammability limit [18] measured in standard, low surface-to-volume vessels. Channel width in microreactors can be made smaller than the quenching distance of hydrogen/oxygen flames, which are thus expected to self extinguish. The same issue can be easily solved with a packed bed [18]. In addition, the bed can be formed by inert particles (e.g. SiO2) or catalyst powder. The possibility of using microreactors is further limited by the low conversion of the reagents inside the reactor (e.g. when H2 conversion is around 35% selectivity is around 30% and when H2 conversion is around 2% selectivity is 100%) [19], [20] and their cost, still high to be scaled up for industrial purpose [17], [18], [19], [20], [21]. The use of Membrane reactors is an alternative for direct synthesis [15], [16], [22]. The catalyst is deposited into the pores of a fine-porous surface layer, while O2 and H2 are kept separated by a membrane [15]. It is reported that the best configuration is dissolving H2 in the liquid phase while O2 is supplied in the coarse–porous side [16]. Tubular membranes can be used in continuous reactors [15], [16]. To avoid explosions membrane reactors are perhaps the best configuration, but the drawback of mass transfer limitation of the gaseous reagents in the catalytic layer is still an issue (e.g. in the continuous H2O2 direct synthesis performed with catalytic membranes, the laminar flow conditions in the empty membrane tube allows H2 mass transfer from the centre to the walls, where the catalytic material is deposited, only by molecular diffusion) [16].

The effects of the reaction conditions were recently investigated by Piccinini et al. [14] and by Moreno et al. [4]. These contributions provide a new interpretation of the results obtained in the direct synthesis. Moreover, our previous work [1] demonstrates how studying the operation conditions allows to find the suitable combination between catalyst and reaction conditions to increase H2O2 selectivity and production rate.

Selectivity in most of the cases is enhanced by adding some stabilizers, like alogens (NaBr, HBr) or acids (H2SO4) [9], [10], [12], [20], [23], to the reaction medium (e.g. methanol, water, ethanol), to support H2O2 production by suppressing water formation and H2O2 decomposition. We aim at finding suitable operation conditions in a TBR allowing to produce H2O2 in situ for other processes (e.g. peroxycarboxylic acid) without using any stabilizer or additive. At these conditions, the highest productivity achieved is 0.18 mol H2O2 gPd−1 h−1, comparable to catalytic membrane results (continuous operations, TiO2 membrane, active phase Palladium, at 30–50 bar, water as a solvent, 0.03 mol/l H2SO4, 15 mg/l NaBr) [16].

Two drawbacks of the TBR are that (i) the reaction mixture must fall outside the flammability limits and (ii) the low contact time between the small catalytic bed and the liquid phase leads to low conversion of the reagents. The production rate in the TBR is still low for industrial applications: and the increase of it is an open issue to be addressed with different catalysts and different feeding policies in the reactor.

Here we investigate the opportunities of engineering the reaction by implementing continuous operation with a trickle bed reactor (TBR). A common industrial reactor permits to be easily scalable for the purposes (i.e. H2O2 synthesis) and integrated to already existing processes.

This work is based on previous studies conducted by Menegazzo et al., who synthesized a number of promising catalysts, based on Pd [12], Pd-Au [24] and Pd-Pt [9] and by Biasi et al. [1], who described a method to enhance selectivity in hydrogen peroxide direct synthesis. Two promising catalyst based on palladium and gold, one supported on sulfated zirconia (ZS) and the other one on sulfated ceria (CeS) are compared with an in depth analysis of operation conditions. In addition, gas and liquid residence times are independently manipulated to affect the extent of the reaction and gas–liquid mass transfer effectiveness (a compensation for small reagents solubility, even lower if operating in the non-explosive region). Two H2/O2/CO2 gas mixtures are used, namely 2/18/80% and 4/16/80% mol and the advantages of using the richest H2 mixture are explained. Furthermore, a significant improvement in hydrogen peroxide direct synthesis, compared to semi-batch reactor, is reported.

Section snippets

Materials

ZrOCl2 (Sigma–Aldrich), (NH4)2Ce(NO3)6 (Sigma–Aldrich), (NH4)2SO4 (Merck), were used for catalyst synthesis as received. Methanol for HPLC was used as reaction medium (J.T. BAKER 99.99%), H2O2 30% w/w (Merck) for decomposition tests, while potassium iodide (Sigma–Aldrich), Hydranal-Composite 2 (Fluka), dry methanol for KFT (Fluka), Acetic Acid (Sigma–Aldrich), Sodium Thiosulfate penta-hydrate 99.5% (Sigma–Aldrich), Starch (Sigma–Aldrich), Potassium dichromate (Riedel de Haën) were used for

Results and discussion

The results of the experiments to produce H2O2 from its elements in a TBR are shown in Fig. 2 (above) and Fig. 3 (below), Fig. 4 (above) and Fig. 5 (below), with both PdAu-CeS (sulfated ceria) and PdAu-ZS (sulfated zirconia) catalysts, reporting either H2O2 production rate or selectivity, as previously defined. Different gas and liquid flow rates as well as different H2/O2 ratio in the gas phase were investigated.

The two catalysts were already developed and tested in a semibatch reactor [12],

Conclusions

A continuous reactor for hydrogen peroxide synthesis is successfully used to enhance catalyst activity. Bimetallic catalysts confirm to be promising for H2O2 direct synthesis, best results being obtained with a PdAu-ZS catalyst. The H2/O2 ratio in the gas phase has different effects on different catalysts: the catalyst based on PdAu-ZS shows better activity with a 4% H2 compared to the same experiments carried out with 2% H2. PdAu-CeS catalyst provides better selectivity with 2% H2 with a

Acknowledgments

Pierdomenico Biasi gratefully acknowledges the PCC (Process Chemistry Centre), Åbo Akademi, Finland and the Johan Gadolinian scholarship for financial support.

References (35)

  • E. Ghedini et al.

    J. Catal.

    (2010)
  • G. Bernardotto et al.

    Appl. Catal., A

    (2009)
  • R. Burch et al.

    Appl. Catal., B

    (2003)
  • F. Menegazzo et al.

    J. Catal.

    (2008)
  • A. Pashkova et al.

    Chem. Eng. J.

    (2008)
  • A. Pashkova et al.

    Chem. Eng. J.

    (2010)
  • T. Inoue et al.

    Chem. Eng. J.

    (2010)
  • Y. Voloshin et al.

    Catal. Today

    (2007)
  • F. Menegazzo et al.

    J. Catal.

    (2009)
  • V.R. Choudhary et al.

    Appl. Catal., A

    (2009)
  • J.H. Lunsford

    J. Catal.

    (2003)
  • S. Melada et al.

    J. Catal.

    (2006)
  • V.R. Choudhary et al.

    Appl. Catal., A

    (2007)
  • C. Burato et al.

    Appl. Catal., A

    (2009)
  • P. Biasi et al.

    Ind. Eng. Chem. Res.

    (2010)
  • P. Li et al.

    Phys. Chem. Chem. Phys.

    (2010)
  • T. Moreno et al.

    Green Chem.

    (2010)
  • Cited by (0)

    View full text