Effect of preparation of iron-infiltrated activated carbon catalysts on nitrogen oxide conversion at low temperature

https://doi.org/10.1016/j.apcatb.2014.05.010Get rights and content

Highlights

  • Effect of preparation of iron-infiltrated activated carbon catalysts on nitrogen oxide conversion at low temperature.

  • We compare the preparation of catalysts by chemical vapor infiltration and by the incipient wetness method.

  • Crystalline iron oxide is only detected in catalysts prepared by chemical vapor infiltration.

  • The co-deposition of silica and iron keeps the pores well accessible during the infiltration and catalysis.

  • Fe/SiO2-co-deposited activated carbon exhibits the best NOx conversion performance at 425 K.

Abstract

Nitrogen oxides are toxic and their concentration in human workspace should be reduced to a minimum level. Among the possible catalyst materials activated carbon based catalysts are a cheap and non-toxic alternative of high availability. In this paper we investigate two different methods for the preparation of iron-infiltrated activated carbon catalysts: chemical vapor infiltration (CVI) and the incipient wetness method (IWM). The effects of the preparation method on the structure and catalytical performance are compared with the effects of infiltration load and co-deposition of silicon dioxide. The study elucidates profound differences in the nitrogen dioxide adsorption and catalytic nitrogen oxide decomposition, depending on the catalyst preparation technique. Samples prepared by chemical vapor infiltration exhibit well dispersed iron/iron oxide particles all over the sample cross section. Crystalline iron oxide is only detected in the samples prepared via the gas phase and not in samples prepared by IWM. The nitrogen dioxide adsorption is notably enhanced in samples with a large accessible micropore volume. All samples containing iron catalyze the conversion of nitrogen oxides into nitrous oxide and carbon monoxide, but especially the co-deposition of silica enhances the nitric oxide conversion into less harmful species. The iron/silica-co-deposited activated carbon catalyst prepared via incipient wetness method exhibits the best catalytical performance of all investigated catalysts at 425 K.

Introduction

More than 90% of nitrogen oxide (NOx) emissions from fossil fuel power units are nitric oxide, but in the atmosphere nitric oxide is quickly oxidized to nitrogen dioxide [1]. Nitrogen dioxide however, is even more harmful than nitric oxide [2] and hence the removal of any NOx in human workspace is of strong interest. Catalytic removal of nitrogen oxides by existing air filters would be a great progress. Since common filter materials already apply activated carbons, a modification of these seems favorable.

Carbonaceous materials like activated carbons (AC) have a long history in the technical nitrogen oxide decomposition [3], [4]. Several mechanisms for the conversion reaction have been proposed [5], [6], [7], [8], [9] and a good overview is provided by Gao et al. [10]. Unsaturated carbon surface atoms or highly basic edge sites typically serve as reactive sites [11]. Even if a small quantity of carbon may be consumed during the reaction, the application of solid carbon as a reducing agent is probably less harmful than the possible spill of ammonia or hydrocarbons, especially in human workspace [1], [12]. Furthermore, the working temperature of activated carbon based catalysts can be considerably lower than required in the efficient selective catalytic reaction (SCR) with ammonia [13], [14].

Bashkova and Bandosz investigated the adsorption of nitrogen dioxide on iron-containing polymer-based porous carbons [15]. They found that the retention of nitrogen dioxide at room temperature is a function of the pore volume, of the degree of the iron-containing particle dispersion and of the sort of predominant iron-containing species. Additionally, the NO2 retention is highly influenced by the chemical composition of the adsorbent and by the presence of water, possibly involving a variety of surface complexes of present carbonaceous or inorganic phases [2], [6], [10], [16].

At room temperature a significant amount of the supplied nitrogen dioxide is adsorbed by activated carbon and reduced to nitric oxide [17]. The observed increasing acidity of the sorbent is possibly caused by the formation of nitric or nitrous acid and this effect even further increases in presence of water.

Zhang et al. studied the reaction of nitrogen dioxide with activated carbons between room temperature and 423 K [7]. They suggest that nitrogen dioxide would initially oxidize the carbon surface and is simultaneously reduced to nitric oxide, which either remains chemisorbed at the carbon surface or is released to the gas phase. However, as the carbon surface becomes progressively oxidized, the rate of nitrogen dioxide conversion is significantly decreasing.

The deposition of various transition metals has been evaluated in its activity for NOx conversion [18], [19] and iron has been identified to be one of the most promising candidates [15], [20], [21]. The adsorption of nitrogen dioxide can be closely related of the content of metal in the carbon material, as observed for carbonaceous adsorbents containing silver nanoparticles [22]. The adsorbed nitrogen dioxide is further converted to nitric oxide, nitrous oxide, oxygen or is retained on the carbon surface by interaction with the metallic phases, e.g. by formation of chelate complexes.

In addition to the catalytic effect of the metal and carbonaceous phases, the effective size of the micropores in activated carbons seems also to have an important influence on the adsorption [10] as well as on the catalytic properties of the material [23]. However, much less research has been directed on the effect of the preparation method itself, although such a comparison can uncover profound differences in the catalytic reaction [24].

In the present work we compare for the first time the catalytic activity of four different inexpensive activated carbon based catalysts, prepared by deposition of iron or iron oxide clusters and silica inside the matrix. Two different preparation methods are applied for direct comparison: chemical vapor infiltration (CVI) and the incipient wetness method (IWM). The samples from either method are likely to differ structurally and chemically and probably differ in their catalytic activity [24]. The chosen chemical vapor infiltration process comprises two main steps: (1) homogeneous infiltration of the porous sample matrix and (2) subsequent precursor decomposition. The homogeneous sample infiltration from an alternatively applicable continuous precursor flow [25], [26], [27] is probably more difficult to control in such complex pore systems and is likely to decrease the deposition rate and precursor efficiency [28].

In addition to differences caused by the preparation method, effects of the extent of the infiltration and of the co-deposition of silicon dioxide are also evaluated. The latter may possibly prevent the sintering of adjacent iron oxide particles and stabilize the catalyst surface [29].

The infiltration homogeneity is investigated by (high resolution) transmission electron microscopy (TEM/HRTEM) from thin catalyst lamellas that have been extracted using a focused gallium ion beam (FIB). Surface structure changes during the synthesis and catalysis process are analyzed by scanning electron microscopy (SEM), X-ray diffraction (XRD) and nitrogen adsorption. The catalytic activity of the modified activated carbons is investigated by nitrogen dioxide adsorption in a thermal balance system and in a recycle flow reactor applying 0.9 mol-% NO2 at 425 K and 100 kPa for 120 min. The experimental temperature is chosen above room temperature, in order to obtain reasonable conversion yields within a few hours. The catalytic effects of the samples are compared to those of iron oxide powder and of a commercial platinum reference catalyst.

Section snippets

Catalyst preparation

Two different activated carbon materials are used as substrate materials: spherical activated carbon particles (Rütgers/CarboTech R1407, 1604 m2/g surface area, 0.67 cm3/g pore volume, ca. 0.5 mm particle diameter) and a commercial spherical silica adsorbent containing 14 wt.-% of activated carbon (BASF EnviSorb B+, 762 m2/g surface area, 0.69 cm3/g pore volume, 3–5 mm particle diameter). The applied artificially activated carbon obtains no considerable amount of impurities, which has been

Catalyst composition and structure

The sample preparation by chemical vapor infiltration as compared to the incipient wetness method can produce very different results in the structure of the catalyst and largely affect its catalytic performance [24]. One main concern is the homogeneity of the infiltration throughout the sample and the dispersion of the deposited, catalytically active particles. Fig. 2 shows a SEM micrograph of the cross section of the iron-loaded gas phase infiltrated sample (Fe–AC–CVI). For this investigation,

Summary and conclusions

In the current paper we investigate effects of the sample preparation of iron-infiltrated activated carbon catalysts on the nitrogen oxides adsorption and decomposition at low temperature. Two different preparation methods are chosen: the incipient wetness method and the chemical vapor infiltration. Additionally, the infiltration of a larger iron load and the co-deposition of silica in the sample matrix are evaluated.

The applied chemical vapor infiltration comprises two main steps: (1)

Acknowledgments

The authors thank Mr. A. Hein, Mr. S. Suleiman, Mr. S. Lorenz, Mr. A Görnt and Mr. F. Sen for technical support. The authors would like to thank for financial support from the German Federal Ministry of Economics and Technology (Grant no. 15751N) within the agenda for the promotion of industrial cooperative research and development (IGF) based on a decision of the German Bundestag. The access was opened by the IUTA e. V., Duisburg, and organized by the (IGF-Project No. 15751N).

References (71)

  • J.K. Neathery et al.

    Carbon

    (1997)
  • R. Pietrzak et al.

    Carbon

    (2007)
  • M. Jeguirim et al.

    J. Anal. Appl. Pyrol.

    (2004)
  • N. Shirahama et al.

    Carbon

    (2002)
  • X. Gao et al.

    Fuel Process. Technol.

    (2011)
  • M.J. Illán-Gómez et al.

    Appl. Catal. B: Environ.

    (1999)
  • K. Skalska et al.

    Sci. Total Environ.

    (2010)
  • G. Busca et al.

    Appl. Catal. B: Environ.

    (1998)
  • H. Muckenhuber et al.

    Carbon

    (2006)
  • W.J. Zhang et al.

    Appl. Catal. B: Environ.

    (2008)
  • A.E. Aksoylu et al.

    Appl. Catal. A: Gen.

    (2003)
  • Z. Hu et al.

    Carbon

    (2001)
  • W. Zhang et al.

    Carbon

    (2001)
  • B. Curdts et al.

    Phys. Proc.

    (2013)
  • S. Stegenga et al.

    Appl. Catal. B: Environ.

    (1993)
  • I. Aarna et al.

    Fuel

    (1997)
  • W.M. Shchedrin et al.

    J. Chem. Thermodyn.

    (1978)
  • I. Menendez et al.

    Carbon

    (2001)
  • M. Gurrath et al.

    Carbon

    (2000)
  • S. Bashkova et al.

    J Colloid Interface Sci

    (2009)
  • M. Iwamoto et al.

    Catal. Today

    (1991)
  • R. Pietrzak et al.

    J. Hazard. Mater.

    (2008)
  • L.K. Chan et al.

    Combust. Flame

    (1983)
  • T. Furusawa et al.

    Fuel

    (1985)
  • J.A. Moulijn et al.

    Carbon

    (1995)
  • S.A. Carabineiro et al.

    Catal. Today

    (2008)
  • P.G. Gray

    Gas Separat. Purif.

    (1993)
  • G.L. Bertrand et al.

    Solid State Ionics

    (2000)
  • G.T. Linteris et al.

    Proc. Combust. Inst.

    (2000)
  • K. Kuwana et al.

    Carbon

    (2005)
  • L.M. Dyagileva et al.

    J. Organometal. Chem.

    (1979)
  • K. Kuwana et al.

    Proc. Combust. Inst.

    (2007)
  • T. Hirasawa et al.

    Combust. Flame

    (2004)
  • M.S. Shah

    J. Chem. Soc.

    (1929)
  • M.S. Shah

    J. Chem. Soc.

    (1929)
  • Cited by (11)

    • Reaction of activated carbon zerovalent iron with pentachlorophenol under anaerobic conditions

      2021, Journal of Cleaner Production
      Citation Excerpt :

      Nevertheless, the iron oxide species and activated carbon do not have a clear diffraction peak. Indeed, activated carbon is composed of amorphous carbon with a broad reflection peak at 2θ = 44.7° (Busch et al., 2014). As for Fe oxide, the XRD data cannot reflect all iron species information due to a less content and amorphous species.

    • Reconciling results of MOCVD of a CNT composite with equilibrium thermodynamics

      2016, Journal of Crystal Growth
      Citation Excerpt :

      We have investigated a modified procedure to account for the observed simultaneous deposition of carbon nanotubes (CNTs), metallic iron, and magnetite (Fe3O4) from a single precursor, a metalorganic complex of Fe. It is to be noted in the context that carbon composites of iron oxide, especially with CNTs, are of practical interest and have been prepared and tested for application in supercapacitors, batteries, and adsorbents [5–9]. For example, Zhao et al. have used an electrode of Fe/Fe3O4/carbon as anode material for Li-ion battery [10].

    View all citing articles on Scopus
    View full text