Iron Oxide/Polymer-Based Nanocomposite Material for Hydrogen Sulfide Adsorption Applications

Download Iron Oxide/Polymer-Based Nanocomposite Material for Hydrogen Sulfide Adsorption Applications

Post on 06-Apr-2017




1 download

Embed Size (px)


<ul><li><p>Iron Oxide/Polymer-Based NanocompositeMaterial for Hydrogen Sulfide AdsorptionApplications</p><p>The processing of iron oxide nanoparticles derived from spray flame synthesis forspecific adsorption applications is described. After the as-prepared particlesproved the ability for H2S removal in pure gas treatment, two different nanopar-ticle-based composite materials were prepared. While impregnation of activatedcarbon with the as-prepared nanoparticles showed the expected increase in H2Sadsorption capacities, a significant enhancement in desulfurization performancewas observed for a novel iron oxide nanoparticle composite material. H2S adsorp-tion was tested in fixed-bed breakthrough curve measurements. The H2S removalefficiency of the novel material under ambient conditions indicates highly promis-ing properties for potential use in industrial and air pollution control applications.</p><p>Keywords: Adsorption, Hydrogen sulfide removal, Iron oxide, Nanocomposite,Nanoparticles</p><p>Received: May 21, 2014; revised: July 18, 2014; accepted: August 21, 2014</p><p>DOI: 10.1002/ceat.201400303</p><p>1 Introduction</p><p>The reduction of sour emissions is an environmental issue ofspecial public interest in todays air pollution control as well asin certain sectors of energy technology and chemical engineer-ing, e.g., fuel cells, oil refineries, or biogas upgrader [1, 2]. Inaddition, tightened environmental regulations force manufac-turers to meet lower threshold values in respect of pollutantsrelease. While the desulfurization of fossil fuels has achieved ahigh standard, fuel-cell applications and chemical synthesisbased on heterogeneous catalysis often require a deep desulfur-ization of the respective feedstock. Especially gaseous hydrogensulfide generates serious problems due to its corrosive andtoxic properties. Even minor concentrations cause damage topipelines and energy conversion sites or lead to poisoningeffects on industrial catalysts.Gas adsorption is a preferred technology in modern gas</p><p>phase purifying and separation processes. To meet low ppm-</p><p>specifications in product and/or exhaust gases, two differentroutes for adsorptive desulfurization are commonly applied.Physisorption on regenerable adsorbent materials like zeolitesor silica gels is widely used in cyclic working multistage plants[35]. A further discussion on state-of-the-art adsorbents isfeatured in [6]. Chemisorptive materials based on impregnatedactivated carbon or metal oxide-based scavengers [7] providemuch more adsorption capacity due to the chemical conversionof the sulfurous compounds. However, they are non- or poorlyregenerable and have to be disposed after use. Furthermore,impregnation can only be performed in low concentration andis attended by a significant decrease of surface area and porosi-ty or even pore blocking [812]. Hence, there is a great interestin optimizing existing or developing new adsorbents exceedingthe capacities of present adsorption equipment.A promising approach to meet these future needs is to make</p><p>use of nanoscale materials with their high surface-to-volumeratio and their mesoscopic properties. Their elevated reactivity[13] and the dominance of surface interactions [14] makenanoparticles ideal for adsorption applications. Particularlytransition metal oxides combine these attributes due to theirchemical potential towards polar gases [15, 16]. Iron oxides aswidely available materials meet many of the desired propertiesand consequently were chosen as the object of investigation[17]. The basic approach presented in this work is based on the</p><p> 2014 WILEY-VCH Verlag GmbH &amp; Co. KGaA, Weinheim Chem. Eng. Technol. 2014, 37, No. 11, 19381944</p><p>Oliver Blatt1</p><p>Martin Helmich2</p><p>Bastian Steuten2</p><p>Sebastian Hardt1</p><p>Dieter Bathen2,3</p><p>Hartmut Wiggers1,4</p><p>1University of Duisburg-Essen,Institute for Combustion andGas Dynamics Reactive Fluids,IVG, Duisburg, Germany.</p><p>2University of Duisburg-Essen,Chair of Thermal SeparationTechnology, Duisburg,Germany.</p><p>3Institute of Energy andEnvironmental TechnologyIUTA, Duisburg, Germany.</p><p>4University of Duisburg-Essen,Center for NanointegrationDuisburg-Essen, CENIDE,Germany.</p><p>Supporting Informationavailable online</p><p>Correspondence: Martin Helmich (, Uni-versity of Duisburg-Essen, Chair of Thermal Separation Technology,Lotharstrae 1, 47057 Duisburg, Germany.</p><p>1938 Research Article</p></li><li><p>utilization of iron oxide nanoparticles in combination with thelarge surface area of an activated carbon to achieve a micropo-rous material with huge amount of active adsorption sites forsulfurous species. Although the nanoparticles are expected toprovide a high chemical activity, they require a supporting,porous host material for immobilization and accessibility.The activity of iron oxide nanoparticles from spray flame</p><p>synthesis [18] was measured directly with respect to H2S re-moval using a fixed-bed reactor. For a first proof-of-principlewith respect to their adsorption properties, these nanoparticleswere combined with high-surface-area activated carbon. Asdescribed in literature [1921], carbon was impregnated by adispersion of the pristine nanoparticles and the adsorptioncapacity was measured after removal of the dispersing solvent.An iron oxide/carbon nanocomposite material was synthe-</p><p>sized from an iron oxide/monomer dispersion with subsequentpolymerization and pyrolysis to facilitate a direct embedding ofthe nanoparticles into the carbon framework. These new mate-rials were tested for removal of H2S under ambient conditionsand the results were compared to materials synthesized by theimpregnation method.</p><p>2 Experimental</p><p>2.1 Materials and Reagents</p><p>Iron oxide (FexOy) nanoparticles from spray flame synthesis(details see below) served as raw material for further process-ing. A standard activated carbon, namely Carbotech C40/4supplied by Carbotech AC GmbH (Germany), was taken forimpregnation experiments with FexOy nanoparticles and forcomparison measurements. Gases for adsorption characteriza-tion were supplied by Air Liquide. Nitrogen (99.99%) was usedas purge gas; adsorption of hydrogen sulfide (99.5%) from car-rier gas methane (99.95%) was tested in fixed-bed break-through experiments. Pure hydrogen sulfide was employed fortreatment of as-prepared iron oxide nanoparticles. All gasesmentioned had the highest available grade and were appliedwithout further purification. Certified test gases were used forgas chromatograph calibration.</p><p>2.2 Characterization</p><p>The synthesized nanoparticles and subsequent products werestudied by transmission electron microscopy (TEM) on a Phil-ips CM12 TEM. An Oxford EDX system was used for energydispersive X-ray spectroscopy (EDX) measurements of the im-aged materials. An X-ray diffractometer Xpert PRO (PANalyti-cal) was employed for X-ray diffraction analysis (XRD) of pris-tine nanoparticles and of the composites as well as thematerials after adsorption of H2S. Additionally, the impregna-tion products were investigated with a LEO 1530 scanning elec-tron microscope (SEM). Infrared spectroscopy was carried outon a Bruker IFS66v/S infrared Fourier spectrometer (FT-IR).Thermogravimetric analysis (TGA) of polymeric materials wasperformed on the thermal analysis apparatus NETZSCH STA449 F3 Jupiter. In addition, certain materials were probed by</p><p>elemental analysis (EA) using atomic absorption spectroscopy(AAS), which was done according to DIN EN ISO/IEC17025:2005. The porosity was analyzed for each material bymeasuring the adsorption isotherm of nitrogen at 196 C. Alladsorption measurements were performed using the volumet-ric sorption apparatus BELSORP-max (BEL Japan, Inc.). Priorto all porosity analyses, samples were pretreated at high tem-perature, i.e., 200 C for at least 4 h, and low pressure at a vacu-um level below 103 Pa to remove pre-adsorbed humidity. Porevolume and BET surface were subsequently derived from thenitrogen isotherms according to DIN ISO 9277 [22]. Further-more, pore size distributions were calculated with the analysissoftware BELMaster under the assumptions of slit-shaped, car-bonaceous pores using the Non-Local Density FunctionalTheory (NLDFT) [23].H2S treatment and breakthrough curve measurements were</p><p>carried out in a fixed-bed adsorption system. A schematic illus-tration of the experimental setup is presented in Fig. 1.</p><p>The setup consists of a stainless-steel adsorption column,two thermal mass flow controllers (Bronkhorst Hi Tech), and am-gas chromatograph (Varian) for gas analysis. The instrumentis equipped with pressure and temperature sensors for processmonitoring. All devices are connected with stainless-steel tub-ing. The whole setup was temperature-controlled at 25 C. Allmedium-contacted devices like column, valves, and tubing arepassivized by an amorphous silicon coating (Restek, Silco-Nert 2000) to avoid adsorption of hydrogen sulfide on metal-lic surfaces. Data logging and processing was done withNational Instruments hardware and the software LabVIEW.For pretreatment, the adsorbent materials were heated for at</p><p>least 12 h in a drying oven, standard and nanoparticle-impreg-nated activated carbon at 200 C and nanoparticles and com-posite materials at 300 C. Prior to breakthrough curve meas-urements, adsorbent materials were packed into the column.Subsequently, the bed was purged with nitrogen (&lt; 1 ppmvH2O) at the standard adsorption temperature of 25 C. In anext step, the bed was purged with methane for at least 10minto avoid co-adsorption of methane and traces of impuritiesduring measurement and to assure equal conditions for allexperiments. Afterwards, the column was locked up and thegas stream was passed over a bypass to adjust hydrogen sulfide</p><p>Chem. Eng. Technol. 2014, 37, No. 11, 19381944 2014 WILEY-VCH Verlag GmbH &amp; Co. KGaA, Weinheim</p><p>Figure 1. Schematic illustration of the experimental setup forH2S-treatment and measurement of breakthrough curves.</p><p>Research Article</p></li><li><p>concentration. Methane and hydrogen sulfide were mixed atthe desired ratio using two mass flow controllers to reach lowconcentrations similar to practical cases. The gas mixture wasanalyzed inline with a gas chromatograph via a connected sam-pling line. After achieving stationary conditions, valves wereswitched and the gas stream was passed through the column.All adsorption experiments were carried out at 1.3 kPa and25 C.For breakthrough measurements of the composite material</p><p>and for the H2S treatment experiments, the adsorbent was en-capsulated between two PTFE disks with non-adsorbable wo-ven fabric to assure immobilization of the material. The PTFEdisks were adapted into the stainless-steel column.</p><p>3 Results</p><p>3.1 Nanoparticle Synthesis</p><p>Iron oxide nanoparticles were synthesized by spray flame syn-thesis from a 0.3mole solution of iron(III) acetylacetonate(Fe(III)acac) in toluene. Burner systems similar to the one usedhere are particularly designed for the formation of metal oxidenanoparticles [18]. The complete experimental setup consistsof a burner mounted in a reaction chamber, a liquid precursordelivery system, and a cooled filter apparatus for collecting theparticles; see Fig. 2. The precursor solution is injected into thereaction chamber through a hollow needle and atomized by acoaxial oxygen jet at 9 LSTPmin</p><p>1 in order to form a spray. Thespray is initially ignited by a premixed methane/oxygen pilotflame at 1.5 LSTPmin1CH4 /3 LSTPmin</p><p>1O2</p><p>surrounding the spray.</p><p>Nanoparticle formation and growth takes place at decreasingtemperature downstream the flame [17] and is quenched by in-jecting N2 as cooling gas at 300 LSTPmin</p><p>1. The pristine par-ticles are separated from the gas stream and collected on a filtermembrane.ATEM image of the pristine iron oxide nanoparticles is pre-</p><p>sented in Fig. 3 a. The material exhibits mostly spherical,slightly agglomerated particles with a count median diameterof 7.5 1.9 nm, calculated from a histogram of the TEM char-acterization.</p><p>Rietveld refinement of the XRD measurements using thesoftware MAUD [24, 25] yields a mean crystallite size of6 1 nm, which is in good agreement with the TEM results.The diffraction pattern indicates that the product consists pre-dominantly of g-Fe2O3 (maghemite) and additional X-ray pho-toelectron spectroscopy (XPS) measurements are supportingthis conclusion.The FT-IR spectrum of the pristine product (see Supporting</p><p>Information) mainly shows very strong and broad absorptionbands around 3260 cm1 and 1560 cm1, indicating the prevail-ing presence of adsorbed H2O and OH groups on the particlessurface. Absorption bands in the range of 13001500 cm1 areexpected to originate from adsorbed CO2 and carbonate, re-spectively [26]. The typical FeO vibrations between 400 and800 cm1 are not visible due to detector limitations.To investigate whether the pristine material shows activity</p><p>towards the adsorption of H2S, the freshly prepared powderwas treated with pure H2S gas at room temperature. ThroughH2S treatment, the particles undergo a significant change inmorphology; see Fig. 3 b. Formerly separated, spherical par-ticles have built amorphous, aggregated structures togetherwith the emergence of rod-shaped crystals, which must belongto iron oxide/hydrogen sulfide reaction products [27]. EDXprobing and elemental analysis (not shown here) denote a highconcentration of sulfur in the sample indicating a chemical re-action between iron oxide and H2S [28] such as:</p><p>Fe2O3 3 H2S fi Fe2S3 3 H2O (1</p><p>2 Fe2O3 H2S fi 4 FeOH2 SO2 (2</p><p>XRD analysis clearly indicates massive structural changesafter hydrogen sulfide admission flow compared to the as-pre-pared material (see Supporting Information). Unfortunately,identification of the resulting species was not possible due tothe high complexity of the diffraction pattern most probably</p><p> 2014 WILEY-VCH Verlag GmbH &amp; Co. KGaA, Weinheim Chem. Eng. Technol. 2014, 37, No. 11, 19381944</p><p>Figure 2. Scheme of the reactor setup for spray flame synthesisof iron oxide nanoparticles.</p><p>Figure 3. TEM images of (a) as-prepared iron oxide nanoparti-cles; (b) after treatment with H2S.</p><p>1940 Research Article</p></li><li><p>originating from several iron-sulfur compounds. In accordancewith the results mentioned before, the FT-IR spectrum of theH2S-treated material also exhibits a distinct change and indi-cates iron sulfide reaction products. The almost completelyoverlaid bands around 2930 and 2860 cm1 plus the strong ab-sorption band at 1130 cm1 and the weak ones around 890 and730 cm1 most likely belong to iron(II) sulfide [29]. Additional-ly, the signals of the OH-groups are broadened and shiftedfrom 3260 cm1 to 3200 cm1, also indicating a crucial changeof the particle surface situation. These findings are similar toresults obtained by Cummins et al. [30]. They report a trans-formation of iron oxide to iron oxide/sulfide species duringsulfurization with H2S. This process occurs on the basis of fourdiffusion processes with hydrogen molecules and sulfur, oxy-gen, and iron atoms being involved, creating different iron/oxygen/sulfur...</p></li></ul>


View more >