reverse microemulsion synthesis of nanostructured complex oxides for catalytic combustion

3
© 2000 Macmillan Magazines Ltd NATURE | VOL 403 | 6 JANUARY 2000 | www.nature.com 65 letters to nature ................................................................. Reverse microemulsion synthesis of nanostructured complex oxides for catalytic combustion Andrey J. Zarur & Jackie Y. Ying Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139-4307, USA .............................................................................................................................................. Catalysts play an important role in many industrial processes, but their use in high-temperature applications—such as energy gen- eration through natural gas combustion, steam reforming and the partial oxidation of hydrocarbons to produce feedstock chemicals— is problematic. The need for catalytic materials that remain stable and active over long periods at high operation temperatures, often in the presence of deactivating or even poisoning compounds, presents a challenge. For example, catalytic methane combustion, which generates power with reduced greenhouse-gas and nitrogen- oxide emissions 1–3 , is limited by the availability of catalysts that are sufficiently active at low temperatures for start-up and are then able to sustain activity and mechanical integrity at flame temperatures as high as 1,300 8C. Here we use sol–gel processing in reverse microemulsions to produce discrete barium hexa- aluminate nanoparticles that display excellent methane combustion activity, owing to their high surface area, high thermal stability and the ultrahigh dispersion of cerium oxide on the their surfaces. Our synthesis method provides a general route to the production of a wide range of thermally stable nanostructured composite materials with large surface-to-volume ratios 4–6 and an ultrahigh component dispersion that gives rise to synergistic chemical and electronic effects 7,8 , thus paving the way to the development of catalysts suitable for high-temperature industrial applications. Combustion catalysts can be gainfully employed to stabilize ‘lean’ flames (with low fuel to air ratios) at temperatures significantly lower than those required in non-catalytic combustion, thereby minimizing the production and emission of nitrogen oxide (NO x ) species 1–3 that contribute to acid rain and smog formation. A viable catalyst system needs to be active at low temperatures for start-up and transient periods, and to sustain activity and mechanical integrity at flame temperatures as high as 1,300 8C. In the case of methane, light-off (defined as 10% conversion of the fuel stream) should ideally be achieved at temperatures of about 400 8C. Catalyst systems based on noble metal oxides, such as PdO, have been studied for catalytic combustion applications 9 , but they generally suffer from deactivation at 700–800 8C owing to loss of active species and sintering. Catalysts that use oxides of non-noble metals 10–12 , especially in the form of complex oxides such as aluminates and perovskites, seem more promising, with systems based on barium hexaaluminate (BHA) having received particular attention 13 . BHA, which has a complex crystal structure giving rise to highly aniso- tropic crystal growth that is usually suppressed along the c-axis 14 , has been synthesized by sol–gel processing. However, this material has to be heated to 1,300 8C to achieve crystallization, retains a surface area of only 15 m 2 g -1 after calcination 15 and requires about 700 8C for light-off of a stream of 1-vol% methane (CH 4 ) in air 16 . Precursor Solution Ba(OC 3 H 7 ) 2 + Al(OC 3 H 7 ) 3 in Iso-octane Reverse Microemulsion Precursor Addition Hydrolysis Aging Condensation Particle Recovery Freeze Drying Rotoevaporation Filtration Drying Surfactant and Solvent Removal Calcination and Crystallization Figure 1 Diagram of the reverse-microemulsion-mediated sol–gel processing of barium hexaaluminate (BHA) nanoparticles. The barium and aluminium alkoxide precursors were dissolved in iso-octane. A reverse microemulsion was used as hydrolysing medium for the alkoxide precursors. The hydrolysed materials were aged within the reverse micelles to promote condensation. The nanoparticles were recovered from the reverse microemul- sion preferentially via freeze drying. The surfactants and solvents were removed by supercritical drying and calcination. The materials were then heat treated at high temperatures to promote crystallization. Figure 2 Characterization of uncalcined BHA materials. Transmission electron micrographs: a, gel-like BHA recovered from a reverse microemulsion with 1-wt% H 2 O; b, discrete BHA nanoparticles formed in a reverse microemulsion with 15-wt% H 2 O; c, a mixture of BHA filament-like and spherical particles synthesized in a reverse microemulsion with 30-wt% H 2 O.

Upload: jackie-y

Post on 23-Jul-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

© 2000 Macmillan Magazines LtdNATURE | VOL 403 | 6 JANUARY 2000 | www.nature.com 65

letters to nature

.................................................................Reverse microemulsion synthesisof nanostructured complexoxides for catalytic combustionAndrey J. Zarur & Jackie Y. Ying

Department of Chemical Engineering, Massachusetts Institute of Technology,

Cambridge, Massachusetts 02139-4307, USA

..............................................................................................................................................

Catalysts play an important role in many industrial processes, buttheir use in high-temperature applicationsÐsuch as energy gen-eration through natural gas combustion, steam reforming and thepartial oxidation of hydrocarbons to produce feedstock chemicalsÐis problematic. The need for catalytic materials that remain stableand active over long periods at high operation temperatures, oftenin the presence of deactivating or even poisoning compounds,presents a challenge. For example, catalytic methane combustion,which generates power with reduced greenhouse-gas and nitrogen-oxide emissions1±3, is limited by the availability of catalysts thatare suf®ciently active at low temperatures for start-up and arethen able to sustain activity and mechanical integrity at ¯ametemperatures as high as 1,300 8C. Here we use sol±gel processingin reverse microemulsions to produce discrete barium hexa-aluminate nanoparticles that display excellent methane combustionactivity, owing to their high surface area, high thermal stability andthe ultrahigh dispersion of cerium oxide on the their surfaces. Oursynthesis method provides a general route to the production of awide range of thermally stable nanostructured composite materialswith large surface-to-volume ratios4±6 and an ultrahigh componentdispersion that gives rise to synergistic chemical and electronic

effects7,8, thus paving the way to the development of catalystssuitable for high-temperature industrial applications.

Combustion catalysts can be gainfully employed to stabilize `lean'¯ames (with low fuel to air ratios) at temperatures signi®cantlylower than those required in non-catalytic combustion, therebyminimizing the production and emission of nitrogen oxide (NOx)species1±3 that contribute to acid rain and smog formation. A viablecatalyst system needs to be active at low temperatures for start-upand transient periods, and to sustain activity and mechanicalintegrity at ¯ame temperatures as high as 1,300 8C. In the case ofmethane, light-off (de®ned as 10% conversion of the fuel stream)should ideally be achieved at temperatures of about 400 8C.

Catalyst systems based on noble metal oxides, such as PdO, havebeen studied for catalytic combustion applications9, but they generallysuffer from deactivation at 700±800 8C owing to loss of active speciesand sintering. Catalysts that use oxides of non-noble metals10±12,especially in the form of complex oxides such as aluminates andperovskites, seem more promising, with systems based on bariumhexaaluminate (BHA) having received particular attention13. BHA,which has a complex crystal structure giving rise to highly aniso-tropic crystal growth that is usually suppressed along the c-axis14,has been synthesized by sol±gel processing. However, this materialhas to be heated to 1,300 8C to achieve crystallization, retains asurface area of only 15 m2 g-1 after calcination15 and requires about700 8C for light-off of a stream of 1-vol% methane (CH4) in air16.

Precursor SolutionBa(OC3H7)2 +Al(OC3H7)3 inIso-octane

ReverseMicroemulsion

Precursor AdditionHydrolysis

AgingCondensation

Particle Recovery• Freeze Drying

• Rotoevaporation• Filtration

DryingSurfactant and

Solvent Removal

Calcination andCrystallization

Figure 1 Diagram of the reverse-microemulsion-mediated sol±gel processing of barium

hexaaluminate (BHA) nanoparticles. The barium and aluminium alkoxide precursors were

dissolved in iso-octane. A reverse microemulsion was used as hydrolysing medium for the

alkoxide precursors. The hydrolysed materials were aged within the reverse micelles to

promote condensation. The nanoparticles were recovered from the reverse microemul-

sion preferentially via freeze drying. The surfactants and solvents were removed by

supercritical drying and calcination. The materials were then heat treated at high

temperatures to promote crystallization.

Figure 2 Characterization of uncalcined BHA materials. Transmission electron

micrographs: a, gel-like BHA recovered from a reverse microemulsion with 1-wt% H2O;

b, discrete BHA nanoparticles formed in a reverse microemulsion with 15-wt% H2O;

c, a mixture of BHA ®lament-like and spherical particles synthesized in a reverse

microemulsion with 30-wt% H2O.

© 2000 Macmillan Magazines Ltd

BHA is dif®cult to synthesize in a nanocrystalline form, due inpart to the different reactivity of the barium and aluminiumprecursors. The compositional heterogeneity in the as-synthesizedmaterial requires crystallization via solid-state reaction at very hightemperatures, which induces loss of surface area associated withgrain growth. To achieve BHA nanocrystals at relatively low tem-peratures, we developed a reverse-microemulsion-mediated syn-thesis (see Fig. 1 and Methods). In this technique, nanometre-sizedaqueous micelles dispersed in an oil phase are used as nanoreactorsfor controlled hydrolysis and condensation of Ba and Al alkoxides.Unlike conventional sol±gel processing, the reaction rate in thisapproach is controlled by the diffusion of precursors from the oilphase to the aqueous domains, instead of the hydrolysis of one ofthe precursors. Despite the different hydrolysis rates of Ba and Alalkoxides, chemical homogeneity can be attained with the media-tion of the reverse microemulsion since the alkoxides have similardiffusivities in the oil phase.

To tailor the particle morphology, structure, and surface area ofBHA nanoparticles, synthesis parameters such as the compositionof the reverse microemulsion, water to alkoxide ratio, aging time,powder recovery and drying techniques were carefully controlled.The composition of the reverse microemulsion governs the shapeand size of the reverse micellar domains, as well as the phasestability. This is re¯ected by the morphology of the powders

obtained from the reverse-microemulsion-mediated synthesis.Reverse microemulsions with a water content of ,5 wt% under-went phase separation shortly after the addition of the precursorsolution. The powders produced could not be stably suspended inthe resulting mixture, and were agglomerated to form a gel-likestructure (Fig. 2a). The ultra®ne pores of this material were easilycollapsed during calcination, yielding coarse-grained materials withsurface areas ,20 m2 g-1, similar to those obtained from conven-tional sol±gel processing.

With an intermediate water content of 5±20 wt%, discretespherical particles of 3±10 nm were obtained from the reverse-microemulsion-mediated synthesis (Fig. 2b). These nanoparticlescould be crystallized directly to the desired BHA phase at a relativelylow temperature of 1,050 8C, and sustained surface areas in excess of100 m2 g-1 after calcination to 1,300 8C. The excellent thermalstability of these nanoparticles might be attributed to the successfulseparation of the crystallization and grain-growth processes. Con-ventional materials undergo BHA crystallization between 1,250 and1,350 8C where a strong driving force for grain growth exists.Reverse-microemulsion-derived BHA nanoparticles possess super-ior chemical homogeneity, so that crystallization could be achievedat lower temperatures. Upon crystallization, grain growth at highertemperatures was suppressed, owing to the anisotropic crystalgrowth of BHA.

For systems with .20 wt% water, signi®cant percolation of themicellar domains might occur, leading to a mixture of liquid-crystallinephases. The materials recovered from these systems showedspherical and elongated particle morphologies (Fig. 2c). Thesecould be crystallized at 1,100±1,150 8C, and possessed a moderatelyhigh surface area of 40±60 m2 g-1 after calcination at 1,300 8C.

The water to alkoxide ratio is an important parameter forcontrolling the relative rates of hydrolysis and condensationreactions17. In general, higher levels of excess water lead to smallerparticle size and increased surface area after calcination. At a waterto alkoxide ratio of 1±10 times the stoichiometric value, particlegrowth was signi®cant, as re¯ected by the change in appearance ofthe reverse microemulsion from transparent to opaque white. Aftera short aging period, the medium underwent phase separation,resulting in highly agglomerated materials. At a water to alkoxideratio >100 times the stoichiometric value, the reverse microemul-sion remained transparent or translucent throughout the agingprocess; uniform, discrete particles were recovered upon drying.After calcination to 1,300 8C, these nanoparticles retained surfaceareas of 40±160 m2 g-1, depending on the reverse-microemulsioncomposition and the recovery technique used.

Particle growth may be minimized by short aging periods.However, appropriate aging was essential to ensure completion ofthe condensation reaction. Condensation of the surface hydroxylgroups led to particles that were less susceptible to agglomeration insubsequent powder processing. We determined that the optimalaging period for BHA systems was 24±48 h.

Conventional techniques (such as ®ltration) were ineffective atrecovering the materials prepared by reverse-microemulsion-mediated processing. Freeze drying was found to allow successfulrecovery of BHA nanoparticles with high yields while preservingtheir discrete, nanometre-sized particle morphology. To remove theresidual surfactants and volatiles in the recovered particles, sampleswere subjected to drying in an oven or under supercritical condi-tions. Supercritical drying prevented particle agglomeration andgrowth during the removal of organics. Materials recovered byfreeze drying and puri®ed by supercritical drying retained surfaceareas as high as 160 m2 g-1 after calcination at 1,300 8C. Crystal-lization was achieved by 1,050 8C, and grain growth was minimalafter further heat treatment. The ®nal BHA particle size was only30 nm, as shown in Fig. 3a.

Our nanocrystalline BHA materials showed excellent catalyticcombustion activity. Light-off of a stream of 1-vol.% CH4 in air at a

letters to nature

66 NATURE | VOL 403 | 6 JANUARY 2000 | www.nature.com

Figure 3 Characterization of calcined BHA nanoparticles. a, Transmission electron

micrograph and electron diffraction pattern (inset) of reverse-microemulsion-derived BHA

nanoparticles after calcination at 1,300 8C. b, High-resolution transmission electron

micrograph of BHA nanoparticles with 10-wt% nanocrystalline CeO2 surface coating after

calcination at 800 8C.

© 2000 Macmillan Magazines Ltd

high space velocity of 60,000 h-1 was observed at 590 8C fornanocrystalline BHA (Fig. 4b), compared to 710 8C required forconventional BHA (Fig. 4a). The nanocrystalline BHA catalystprovided full CH4 conversion over a wide temperature range of710±1,300 8C, with no signs of deactivation at the elevated reactiontemperatures.

Introduction of surface-active deposits onto the BHA nanopar-ticles further increased their catalytic activity. As ceria does not reactwith baria or alumina under our synthesis conditions, we were ableto deposit up to 25-wt% ceria onto BHA nanoparticles withoutforming a separate ceria domain. The ceria deposits on BHApreserved a nanocrystalline morphology even after calcination at1,300 8C, whereas unsupported ceria would have undergone severesintering and grain growth beyond 100 nm by 700 8C. High-resolutiontransmission electron microscopy (HR-TEM) illustrated that ceria(CeO2) nanocrystals were achieved with ultrahigh dispersion on thesurface of BHA nanoparticles using our synthesis approach(Fig. 3b). The ceria grain size (,6 nm) observed in HR-TEM is ingood agreement to that calculated from peak-broadening analysis ofthe X-ray diffraction (XRD) pattern. With the use of the stable BHAnanoparticles as support materials, a ®ne grain size of 18 nm waspreserved for ceria after extended calcination at 1,100 8C in thepresence of 15-vol.% H2O.

The CeO2±BHA nanoparticles allowed light-off of a stream of1-vol.% CH4 in air at a low temperature of ,400 8C (Fig. 4c),compared to 540 8C and 690 8C required for Mn- and Co-substituted BHA systems16. The apparent activation energy ofthe CeO2±BHA nanoparticles for combustion of ,4-vol.% CH4

in air was found to be ,145 kJ mol-1. In comparison, theconventional 0.5-wt% PdO/Al2O3 had an activation energyonly slightly lower (116 kJ mol-1). The CeO2±BHA catalystsattained full CH4 conversion by 600 8C, and no hysteresis wasobserved in the catalyst performance during heating±cooling±restarting cycles. We believe that the excellent low-temperaturecatalytic activity of this system arises in part from the ultrahighCeO2 dispersion on BHA nanoparticles. CeO2-based materialshave not been examined widely as combustion catalysts, sinceother metal-oxide systems (such as Co2O3 and Mn3O4) possesshigher speci®c activities18. However, the use of the transition-metal-oxide catalysts has been limited to ,1,000 8C, abovewhich severe grain growth and sintering occur. In contrast,our CeO2±BHA system not only achieved light-off at lowtemperatures, but also preserved full methane conversion activ-ity at temperatures in excess of 1,100 8C and in the presence ofH2S and water vapour. This combination of low-temperaturecatalytic activity, high-temperature thermal stability, and poi-soning resistance render our catalysts interesting for potential

practical applications in ultralean catalytic combustion ofmethane. Provided the composite oxides do not react witheach other under our synthesis conditions, the nanocompositedesign and the ultrahigh component dispersion achieved by ourreverse-microemulsion-mediated synthesis can be broadlyapplied to the development of other complex-oxide catalystswith high surface reactivity and thermal stability for a variety ofcatalytic applications. M

MethodsThe reverse microemulsions of water in oil (for example, iso-octane) were prepared with amixture of surfactants, such as polyethylene oxide adducts and linear alcohols. Typicalmicroemulsion compositions consist of 1±40 wt% water, 5±30 wt% surfactants, and 60±90 wt% hydrocarbons. Ba(OC3H7)2 and Al(OC3H7)3 were dissolved in iso-octane to formthe precursor solution, which was then slowly added (5±10 ml min-1) to the reversemicroemulsion at room temperature. The water to alkoxide ratios were varied from 1 to1,000 times the stoichiometric value (corresponding to 19 moles of water per mole ofBaO×6Al2O3 derived). The hydrolysed mixture was then allowed to age for 1±72 h.

The BHA nanoparticles could be modi®ed through the introduction of surface depositsof rare-earth and transition-metal oxides. Nitrate or acetate salts of the desired cationswere introduced to the BHA nanoparticles during their aging in the reverse microemul-sion medium 12±24 h after hydrolysis had taken place. This approach allowed thedeposition of up to 25 wt% of various metal oxides, including those of cerium, cobalt,manganese, and lanthanum.

The materials generated in the reverse microemulsion could be recovered either by®ltration, freeze drying, or phase separation (by cooling of the reverse microemulsion).The nanometre-sized particles were dif®cult to collect with high yields by ®ltration orcentrifugation. Roto-evaporation resulted in heavy particle agglomeration due to the hightemperatures needed to remove the surfactants. In contrast, freeze drying consisted ofspraying the reverse microemulsion into a Dewar containing liquid nitrogen. The frozensystem was then placed under vacuum, and heated slowly to remove iso-octane, water, andthe lighter fractions of surfactants by sublimation. After recovery, the powders weresubjected to oven or supercritical drying to remove the residual surfactants and volatiles.The materials were then calcined to 800 8C or 1,300 8C for the modi®ed and pure BHAnanoparticles, respectively, before catalytic studies. These materials were shown byelemental analysis to be free of surfactants and volatiles (with ,0.1 wt% carbon content).

Catalytic studies were conducted in a ¯ow reactor under isothermal conditions.Methane and air ¯ows were controlled by independent mass ¯ow controllers (MKS) tovary the CH4 to O2 ratio. The ef¯uent stream was analysed using a Hewlett-Packard gaschromatograph with a mass selective detector.

Received 4 May; accepted 21 October 1999.

1. Pfefferle, L. D. & Pfefferle, W. C. Catalysis in combustion. Catal. Rev. Sci. Eng. 29, 219±267 (1987).

2. Linan, A. & Williams, F. A. Fundamental Aspects of Combustion (Oxford, New York, 1993).

3. Lunsford, J. H. The role of surface generated gas-phase radicals in catalysis. Langmuir 5, 12±16 (1989).

4. Karch, J., Birringer, R. & Gleiter, H. Ceramics ductile at low temperature. Nature 330, 556±558 (1987).

5. Gleiter, H. Nanocrystalline materials. Prog. Mater. Sci. 33, 223±315 (1989).

6. Siegel, R. W. Cluster-assembled nanophase materials. Annu. Rev. Mater. Sci. 21, 559±578 (1991).

7. Ying, J. Y. & TschoÈpe, A. Synthesis and characterization of non-stoichiometric nanocrystalline cerium

oxide-based catalysts. Chem. Eng. J. 64, 225±237 (1996).

8. Ying, J. Y., TschoÈpe, A. & Levin, D. Synergistic effects and catalytic properties tailored by nanostructure

processing. Nanostruct. Mater. 6, 237±246 (1995).

9. McCarty, J. G. Kinetics of PdO combustion catalysis. Catal. Today 26, 283±293 (1995).

10. Trimm, D. L. Materials selection and design of high temperature catalytic combustion units. Catal.

Today 26, 231±238 (1995).

11. Berg, M. & JaÈraÊs, S. G. Catalytic combustion of methane over magnesium oxide. Appl. Catal. A 114,

227±241 (1994).

12. Matsuura, I., Hashimoto, Y., Takayasu, O., Nitta, K. & Yoshida, Y. Heat-stable ultra®ne single-crystal

magnesium oxide and its character as a support material for high-temperature combustion catalysts.

Appl. Catal. 74, 273±280 (1991).

13. Groppi, G., Belloto, M., Cristiani, C., Forzatti, P. & Villa, P. L. Preparation and characterization of

hexaaluminate-based materials for catalytic combustion. Appl. Catal. A 104, 101±108 (1993).

14. Park, J.-G. & Cormack, A. N. Crystal/defect structures and phase stability in Ba hexaaluminates.

J. Solid State Chem. 121, 278±290 (1996).

15. Machida, M., Eguchi, K. & Arai, H. Preparation and characterization of large surface area BHA. Bull.

Chem. Soc. Jpn 61, 3659±3665 (1988).

16. Machida, M., Eguchi, K. & Arai, H. Catalytic properties of BaMAl11O19-a for high-temperature

catalytic combustion. J. Catal. 120, 377±386 (1989).

17. Scherer, G. W. & Brinker, C. J. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing

(Academic, Boston, 1990).

18. Zwinkels, M. F. M., JaÈraÊs, S. G., Menon, P. G. & Grif®n, T. A. Catalytic materials for high-temperature

combustion. Catal. Rev. Sci. Eng. 25, 319±358 (1993).

Acknowledgements

We thank H. H. Hwu for technical assistance and M. Frongillo for assistance with the high-resolution electron microscopy studies. This work was supported by the David and LucilePackard Foundation.

Correspondence and requests for materials should be addressed to J.Y.Y.(e-mail: [email protected]).

letters to nature

NATURE | VOL 403 | 6 JANUARY 2000 | www.nature.com 67

0

10

20

30

40

50

60

70

80

90

100

400 500 600 700 800

Temperature (°C)

Con

vers

ion

(%)

c b a

Figure 4 Catalytic methane combustion experiments. Catalytic methane oxidation

activity: a, sol±gel-derived conventional BHA; b, reverse-microemulsion-derived BHA

nanoparticles; c, reverse-microemulsion-derived CeO2±BHA nanocomposite. The

reactant stream consisted of 1-vol.% CH4 in air (space velocity, 60,000 h-1).