Rice Husk Derived Porous Silica as Support for Pd and CeO2 for Low Temperature Catalytic Methane Combustion

1 Leibniz Institute for Catalysis at the University of Rostock, Albert-Einstein-Str. 29a, D-18059 Rostock, Germany; liudongjing19@163.com; dominik.seeburg@catalysis.de; Stefanie.kreft@catalysis.de; Sebastian.wohlrab@catalysis.de 2 School of Energy and Power Engineering, Jiangsu University, Xuefu Str. 301, 212013 Zhenjiang, China; liudongjing19@163.com 3 Deutsches Biomasseforschungszentrum gemeinnützige GmbH, Torgauer Str. 116, 04347 Leipzig; rene.bindig@dbfz.de; ingo.hartmann@dbfz.de 4 Universität Leipzig, Institute of Chemical Technology, Linnéstr. 3, 04103 Leipzig, Germany; denise.schneider@uni-leipzig.de; dirk.enke@uni-leipzig.de


Introduction
Methane possesses a global warming potential 15 to 34 times higher than that of carbon dioxide (CO 2 ) [1].Hence, it is of tremendous importance to reduce the release of methane to the atmosphere.However, the sources of methane in the air are of different origin.Natural sources of about 37%, mainly from wetlands, but also from agricultural sources of about 45% have been identified as main contributors to the global methane emission [2].As early as 1983, Seiler et al. [3] estimated that 35 to 59 Tg/year of methane could be emitted from rice paddies worldwide.Nowadays, it is assumed that rice paddy fields add about 10% to the global methane release which is about 600 Tg/year [2].In this contribution, we demonstrate how a waste product of rice production-rice husks-could serve as a starting material for a catalyst which can convert methane to the lower potent molecule for global warming-carbon dioxide-by total oxidation.In this way, methane emissions from energy applications, e.g., from gas turbines of power plants, vehicles or ships, could be reduced which currently seriously contribute to the global methane release.
Rice ranks at the second place of the most consumed foodstuffs worldwide, and rice husks are byproducts of rice production [4].Most of the rice husks are disposed of into landfills or improperly burned, which can cause severe environmental impacts in air, water, and soil [5].However, the ash of burned rice husks obtained from controlled conditions has the potential of further applicability in several materials related fields, like energy storage, drug delivery, adsorption or simply as source for other materials [6].Due to their silica content of up to 20 wt%, rice husks can be also seen as low-cost raw materials for the production of porous silica.In this way, silica xerogels with low alkali content have been prepared by acid pretreatment [7] or pure silica powders were obtained from (NH 4 ) 2 SiF 6 intermediate precipitation using NH 3 solution [8].Biogenic silica can also be obtained by combustion of the rice husks, leaving behind a siliceous ash.High-purity silica can be produced by a leaching procedure for the removal of unwanted inorganic matter prior to combustion [9,10] and exhibits pores mainly in the meso-and macroporous region.In previous investigations Enke and coworkers also used rice husk ash for transformation into microporous MFI-type zeolite structures [11], but also for the template assisted generation of mesoporous MCM-41 and MCM-48 [12].Adam et al. [13] commented the progress of the utilization of rice husk silica as a catalyst, especially offered valuable information on the development of the techniques concerning immobilization of transition metals and organic moieties onto rice husk derived silica.
Controlled porous glass (CPG) was successfully employed by Hoffmann et al. for the preparation of novel methane combustion catalysts [14].In detail, CeO 2 was separated from Pd onto the inner surface of CPG [15] which provided a higher accessibility for oxygen to the surface of the metal oxide and a higher reactivity therewith.Although the applicability of Pd supported on metal oxides as suitable methane combustion catalysts is well established [16][17][18][19][20][21][22][23][24], the surface decoration of both components onto CPG surfaces unexpectedly delivered catalysts with an even better performance.However, the applicability of such systems is limited to the low availability of the CPG being the most expensive component in this catalyst system until now.In general, the production of CPG is very energy-and time-consuming.The first step of the production of CPG is to melt the glass from the individual components at temperatures >1000 • C.After shaping, a precise tempering treatment has to be carried out.Finally, the CPG is obtained after time consuming acidic and alkaline leaching.In addition, functionalizing Pd nanoparticles in next-generation matrix supports, such as pH-responsive polymers, is another feasible way to obtain high-performance Pd-based catalysts [25].
In contrast to the preparation of CPG, the production of biogenic silica can be regarded as economically advantageous, as its production is based on cheap agricultural waste products and the price of raw materials is therefore comparatively low.The preparation comprises only two steps: firstly, the leaching to remove inorganic impurities as well as to hydrolyze the organic matter and secondly, the combustion.Due to the heating value of 13-16 MJ/kg [26], the combustion step can be exploited for the co-generation of heat, electricity or steam.Moreover, the waste problem is solved (see above).
Hence, this study uses sustainable and low-cost agricultural waste products, namely rice husks, to generate the porous catalyst support for the metal-metal oxide couple.The synthesized catalysts are named as xPd/CeO 2 /RHS, x denotes the mass percentage of Pd in the catalysts, RHS represents the rice husk derived support.The 0.2Pd/CeO 2 /RHS, 0.6Pd/CeO 2 /RHS and 1.0Pd/CeO 2 /RHS catalysts with Pd contents of 0.2, 0.6 and 1.0 wt% respectively are obtained by a facile impregnation/calcination approach.Benchmark catalyst 0.6Pd/CeO 2 /PG151 (PG151 is a typical CPG) is prepared as well.The samples are characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), nitrogen adsorption-desorption, X-ray diffraction (XRD), hydrogen temperature programmed reduction (H 2 -TPR) techniques.The RHS-supported decoupled Pd-CeO 2 catalysts and benchmark catalyst 0.6Pd/CeO 2 /PG151 were then employed for low-temperature methane combustion in a horizontal fixed-bed microreactor under methane lean conditions (1000 ppm) in a simulated dry and wet exhaust gas.

Results and Discussion
The obtained RHS is characterized by its high silicon content of Si ~98.3 wt% (oxygen neglected).Schliermann et al. [27] investigated the effect of leaching on the metal content in the rice husks and were able to show that a significant reduction of the metal by leaching with citric acid is obtained.For example, the K content was reduced from 2650 mg/kg below the detection limit.For the RHS, impurities of Ca, P, K, Al, S, Mg, Na, Cu, Zn, Mn, Ti and Cr (for details see Table S1) were reduced to a minimum amount due to the applied synthesis method.
SEM indicated that the RHS had nonuniform particle sizes and shapes (Figure 1a).Moreover, irregular shapes of the porous surface structures were observed (Figure 1b).This material served as scaffold for incorporated Pd-CeO 2 catalysts represented by 0.6Pd/CeO 2 /RHS possessing the same composition as the benchmark catalyst 0.6Pd/CeO 2 /PG151 [14].Figure 1c shows a TEM image of 0.6Pd/CeO 2 /RHS.Small particles, sometimes aggregated, can be assigned to palladium species (grey in the dark field image).The majority of them are irregularly distributed on the surface of the porous silica.Very few palladium grains are located on CeO 2 (brighter in the dark field image) which is further verified by the EDX analysis in Figure S1.Single ceria particles were identified on the surface of RHS which were found stable and isolated in the spent catalyst (Figure S2). Figure S3 additionally shows an overview of a 0.6Pd/CeO 2 /RHS granule with significant parts of CeO 2 and Pd particles sparsely located on the RHS surface.Figure 1d displays the Selected Area Electron Diffraction (SAED) analysis of 0.6Pd/CeO 2 /RHS.The diffraction spots with distance of 0.313 and 0.166 nm are attributed to the reflection of CeO 2 (111) and CeO 2 (311) lattice place, respectively; while the diffraction spots with distance of 0.199 nm are assigned to the Pd (200) lattice plane.

Results and Discussion
The obtained RHS is characterized by its high silicon content of Si ~ 98.3 wt% (oxygen neglected).Schliermann et al. [27] investigated the effect of leaching on the metal content in the rice husks and were able to show that a significant reduction of the metal by leaching with citric acid is obtained.For example, the K content was reduced from 2650 mg/kg below the detection limit.For the RHS, impurities of Ca, P, K, Al, S, Mg, Na, Cu, Zn, Mn, Ti and Cr (for details see Table S1) were reduced to a minimum amount due to the applied synthesis method.
SEM indicated that the RHS had nonuniform particle sizes and shapes (Figure 1a).Moreover, irregular shapes of the porous surface structures were observed (Figure 1b).This material served as scaffold for incorporated Pd-CeO2 catalysts represented by 0.6Pd/CeO2/RHS possessing the same composition as the benchmark catalyst 0.6Pd/CeO2/PG151 [14].Figure 1c shows a TEM image of 0.6Pd/CeO2/RHS.Small particles, sometimes aggregated, can be assigned to palladium species (grey in the dark field image).The majority of them are irregularly distributed on the surface of the porous silica.Very few palladium grains are located on CeO2 (brighter in the dark field image) which is further verified by the EDX analysis in Figure S1.Single ceria particles were identified on the surface of RHS which were found stable and isolated in the spent catalyst (Figure S2). Figure S3 additionally shows an overview of a 0.6Pd/CeO2/RHS granule with significant parts of CeO2 and Pd particles sparsely located on the RHS surface.Figure 1d displays the Selected Area Electron Diffraction (SAED) analysis of 0.6Pd/CeO2/RHS.The diffraction spots with distance of 0.313 and 0.166 nm are attributed to the reflection of CeO2 (111) and CeO2 (311) lattice place, respectively; while the diffraction spots with distance of 0.199 nm are assigned to the Pd (200) lattice plane.
As determined by nitrogen physisorption experiments, RHS possesses a Brunauer-Emmett-Teller (BET) surface area of about 217 m 2 /g.The BET surface area of 0.6Pd/CeO2/RHS (representing the RHSsupported catalysts) is decreased to about 50% compared to RHS due to the loading mainly with CeO2.However, it is larger than that of the unsupported 0.6Pd/CeO2 (see Table 1 and Figure 2) as well as the benchmark catalyst 0.6Pd/CeO2/PG151 (82 m 2 /g [14]), respectively.Table 1.BET-specific surface area (SBET) calculated according to Ref. [28], average pore volume (Vt) as well as element analysis data.As determined by nitrogen physisorption experiments, RHS possesses a Brunauer-Emmett-Teller (BET) surface area of about 217 m 2 /g.The BET surface area of 0.6Pd/CeO 2 /RHS (representing the RHS-supported catalysts) is decreased to about 50% compared to RHS due to the loading mainly with CeO 2 .However, it is larger than that of the unsupported 0.6Pd/CeO 2 (see Table 1 and Figure 2) as well as the benchmark catalyst 0.6Pd/CeO 2 /PG151 (82 m 2 /g [14]), respectively.The nitrogen uptake of RHS during sorption experiments is quite small at a low relative pressure and increases slowly with increasing relative pressure as it can be seen in the isotherms in Figure 2a.This implies the presence of a few micropores in the pore structure of RHS.Hysteresis loops are detected among all the three samples, indicating that the main pores are located in the lower mesopore region but additionally with significant parts in the higher mesopore or macropore region.The two catalysts show different adsorption behaviour, while 0.6Pd/CeO2/RHS possesses features of the pure RHS as well as of the 0.6Pd/CeO2.The derivation of the pore size distribution according to the BJH-method (Barrett-Joyner-Halenda) [29] is shown in Figure 2b.0.6Pd/CeO2 is characterized by pores with mean diameter of ~12 nm while the pure RHS has the highest amount of pores around 4 nm, but also a significant portion of pores in a broad mesopore range.0.6Pd/CeO2/RHS possesses a bimodal pore size distribution reflecting the presence of both components, respectively.This suggests that a proportion of the smaller mesopores of the RHS have either been filled or blocked by CeO2 particles.The nitrogen uptake of RHS during sorption experiments is quite small at a low relative pressure and increases slowly with increasing relative pressure as it can be seen in the isotherms in Figure 2a.This implies the presence of a few micropores in the pore structure of RHS.Hysteresis loops are detected among all the three samples, indicating that the main pores are located in the lower mesopore region but additionally with significant parts in the higher mesopore or macropore region.The two catalysts show different adsorption behaviour, while 0.6Pd/CeO 2 /RHS possesses features of the pure RHS as well as of the 0.6Pd/CeO 2 .The derivation of the pore size distribution according to the BJH-method (Barrett-Joyner-Halenda) [29] is shown in Figure 2b.0.6Pd/CeO 2 is characterized by pores with mean diameter of ~12 nm while the pure RHS has the highest amount of pores around 4 nm, but also a significant portion of pores in a broad mesopore range.0.6Pd/CeO 2 /RHS possesses a bimodal pore size distribution reflecting the presence of both components, respectively.This suggests that a proportion of the smaller mesopores of the RHS have either been filled or blocked by CeO 2 particles.
H 2 -TPR tests were carried out to derive insights into the reducibility of RHS-supported catalysts and they were compared with the benchmark catalyst 0.6Pd/CeO 2 /PG151 from Ref. [14] (Figure 4).The intense peak of 1.0Pd/CeO 2 /RHS at the lowest temperature of ~30 • C is attributed to the reduction of the PdO x species located on the RHS surface, which corresponds well with the main peak for Pd/SiO 2 reported by Muto et al. [30] who claimed that the PdO x species supported on SiO 2 were reduced at low temperatures with maximum hydrogen consumption at ~30 • C. The main reduction temperature of unsupported catalyst Pd/CeO 2 would be slightly higher as reported by Leitenburg et al. [31].The catalyst 0.6Pd/CeO 2 /RHS shows temperatures for Pd-O reducibility in the same range as 0.6Pd/CeO 2 /PG151 but with slightly reduced H 2 consumption.The broad signal at maximum of ~485 • C can be assigned to the reduction of surface CeO 2 [32,33] indicating the successful preparation of the decoupled Pd species from CeO 2 within the sample in accordance to our previous work [14].The peaks at high temperature of ~730 • C or >800 • C, obtained for the RHS supported catalysts are related to the reduction of bulk CeO 2 [33].H2-TPR tests were carried out to derive insights into the reducibility of RHS-supported catalysts and they were compared with the benchmark catalyst 0.6Pd/CeO2/PG151 from Ref. [14] (Figure 4).The intense peak of 1.0Pd/CeO2/RHS at the lowest temperature of ~30 °C is attributed to the reduction of the PdOx species located on the RHS surface, which corresponds well with the main peak for Pd/SiO2 reported by Muto et al. [30] who claimed that the PdOx species supported on SiO2 were reduced at low temperatures with maximum hydrogen consumption at ~30 °C.The main reduction temperature of unsupported catalyst Pd/CeO2 would be slightly higher as reported by Leitenburg et al. [31].The catalyst 0.6Pd/CeO2/RHS shows temperatures for Pd-O reducibility in the same range as 0.6Pd/CeO2/PG151 but with slightly reduced H2 consumption.The broad signal at maximum of ~485 °C can be assigned to the reduction of surface CeO2 [32] - [33] indicating the successful preparation of the decoupled Pd species from CeO2 within the sample in accordance to our previous work [14].The peaks at high temperature of ~730 °C or >800 °C, obtained for the RHS supported catalysts are related to the reduction of bulk CeO2 [33].Catalytic methane combustion tests were conducted to study the influence of RHS support as well as Pd content on methane oxidation activity in the temperature range of 150-500 °C (Figure 5).0.6Pd/CeO2/RHS completely converts methane below 500 °C in dry feed gas, whereas unsupported  H2-TPR tests were carried out to derive insights into the reducibility of RHS-supported catalysts and they were compared with the benchmark catalyst 0.6Pd/CeO2/PG151 from Ref. [14] (Figure 4).The intense peak of 1.0Pd/CeO2/RHS at the lowest temperature of ~30 °C is attributed to the reduction of the PdOx species located on the RHS surface, which corresponds well with the main peak for Pd/SiO2 reported by Muto et al. [30] who claimed that the PdOx species supported on SiO2 were reduced at low temperatures with maximum hydrogen consumption at ~30 °C.The main reduction temperature of unsupported catalyst Pd/CeO2 would be slightly higher as reported by Leitenburg et al. [31].The catalyst 0.6Pd/CeO2/RHS shows temperatures for Pd-O reducibility in the same range as 0.6Pd/CeO2/PG151 but with slightly reduced H2 consumption.The broad signal at maximum of ~485 °C can be assigned to the reduction of surface CeO2 [32] - [33] indicating the successful preparation of the decoupled Pd species from CeO2 within the sample in accordance to our previous work [14].The peaks at high temperature of ~730 °C or >800 °C, obtained for the RHS supported catalysts are related to the reduction of bulk CeO2 [33].Catalytic methane combustion tests were conducted to study the influence of RHS support as well as Pd content on methane oxidation activity in the temperature range of 150-500 °C (Figure 5).0.6Pd/CeO2/RHS completely converts methane below 500 °C in dry feed gas, whereas unsupported 0.6Pd/CeO2 can only reach ~80% methane conversion at 500 °C.Moreover, the ignition temperature of 0.6Pd/CeO2/RHS is also lower than that of 0.6Pd/CeO2.The activity is comparable to our previous results using more expensive CPG as support used for 0.6Pd/CeO2/PG151 catalyst [14].The Pd Catalytic methane combustion tests were conducted to study the influence of RHS support as well as Pd content on methane oxidation activity in the temperature range of 150-500 • C (Figure 5).0.6Pd/CeO 2 /RHS completely converts methane below 500 • C in dry feed gas, whereas unsupported 0.6Pd/CeO 2 can only reach ~80% methane conversion at 500 • C.Moreover, the ignition temperature of 0.6Pd/CeO 2 /RHS is also lower than that of 0.6Pd/CeO 2 .The activity is comparable to our previous results using more expensive CPG as support used for 0.6Pd/CeO 2 /PG151 catalyst [14].The Pd content also has an influence on the methane catalytic activity as it was implied by the TPR results.1.0Pd/CeO 2 /RHS represents a higher catalytic activity with T 90 = 327 • C in dry feed gas.
As it was reported for Pd-Pt/Al 2 O 3 catalysts [34], the required combustion temperature in presence of water is significantly higher compared to dry conditions.Also in our case, with presence of water in the feed gas (10.5 vol%-H 2 O), the catalytic performance of xPd/CeO 2 /RHS decreased (Figure 6).Again, the activity of 0.6Pd/CeO 2 /RHS is comparable to 0.6Pd/CeO 2 /PG151.The activity was tremendously increased with a higher Pd loading as recorded over 1.0Pd/CeO 2 /RHS.content also has an influence on the methane catalytic activity as it was implied by the TPR results.1.0Pd/CeO2/RHS represents a higher catalytic activity with T90 = 327 °C in dry feed gas.As it was reported for Pd-Pt/Al2O3 catalysts [34], the required combustion temperature in presence of water is significantly higher compared to dry conditions.Also in our case, with presence of water in the feed gas (10.5 vol%-H2O), the catalytic performance of xPd/CeO2/RHS decreased (Figure 6).Again, the activity of 0.6Pd/CeO2/RHS is comparable to 0.6Pd/CeO2/PG151.The activity was tremendously increased with a higher Pd loading as recorded over 1.0Pd/CeO2/RHS.As it was reported for Pd-Pt/Al2O3 catalysts [34], the required combustion temperature in presence of water is significantly higher compared to dry conditions.Also in our case, with presence of water in the feed gas (10.5 vol%-H2O), the catalytic performance of xPd/CeO2/RHS decreased (Figure 6).Again, the activity of 0.6Pd/CeO2/RHS is comparable to 0.6Pd/CeO2/PG151.The activity was tremendously increased with a higher Pd loading as recorded over 1.0Pd/CeO2/RHS.The reasons for the lower activity might be the competitive adsorption of water and methane on catalyst active sites or the formation of metal hydroxides.However, unsupported 0.6Pd/CeO 2 was completely deactivated, suggesting that RHS is an efficient support for improving the activity of Pd-CeO 2 catalyst both in dry and in wet feed gas.

Catalyst Preparation
For the production of biogenic silica as catalyst support, rice husks originated from Cambodia were used.The rice husks were leached with a 1 wt% citric acid solution for 2 h at 50 • C in an immersion bath.Then the solution was filtered and rinsed with tap water with subsequent thermal drying.
The resulting leached rice husks were burned in laboratory furnace under ambient air atmosphere at 650 • C for 2 h and then cooled down to room temperature in ambient air.The obtained ash was designated as rice husk silica (RHS).
The catalyst preparation approach was described in our previous paper [14] and the catalyst described there as 0.6Pd/CeO 2 /PG151 (CPG with d pore = 151 nm) was used as benchmark.In brief, RHS was initially impregnated with molten Ce(NO 3 ) 3 •6H 2 O at 235 • C and then calcined at 450 • C with a heating rate of 5 K/min for 2 h.The required amount of Ce(NO 3 ) 3 •6H 2 O was calculated to completely fill the pores considering an approx.density of 2.38 g/cm 3 .Hereafter, as-obtained CeO 2 -containing porous silica (CeO 2 /RHS) granules were impregnated with a Pd(NO 3 ) 2 •2H 2 O solution via incipient wetness method to yield catalysts containing 0.2, 0.6, 1.0 wt% Pd.Finally, the catalyst precursors were dried in air at 60 • C overnight and calcined at 500 • C with heating rate of 5 K/min for 1 h.The as-prepared catalysts are denoted as xPd/CeO 2 /RHS, x represents the Pd loading values in wt%, RHS represents the porous silica obtained from rice husk.
For comparison, a 0.6Pd/CeO 2 reference catalyst (0.6 wt% Pd related to CeO 2 ) was synthesized without the porous silica support according to the following protocol.Ce(NO 3 ) 3 •6H 2 O was calcined in air at 450 • C with heating rate of 5 K/min for 2 h to yield pure CeO 2 .The resulting cerium oxide was also impregnated with a Pd(NO 3 ) 2 •2H 2 O solution via the incipient wetness impregnation method to yield 0.6 wt% Pd in the 0.6Pd/CeO 2 catalyst and the subsequent drying and calcination protocol was the same as described for xPd/CeO 2 /RHS catalysts.

Catalyst Characterization
Nitrogen sorption measurements were done by using a NOVA 4200e instrument from Quantachrome (Odelzhausen, Germany).XRD (X-ray diffraction) of the support and catalysts were recorded with an X'Pert PRO reflexion diffractometer from PANalytical (Almelo, The Netherlands).The surface microstructures of the porous silica are imaged by SEM (scanning electron microscopy) using a Merlin VP compact instrument (Carl Zeiss Microscopy, Jena, Germany).TEM (transmission electron microscopy) measurements were performed on a JEM-ARM200F (JEOL, Tokyo, Japan) with aberration-correction by a CESCOR (CEOS) for the scanning transmission (STEM) applications.A Varian 715-ES ICP-OES (inductively coupled plasma-optical emission spectrometer, Palo Alto, CA, USA) was used for the determination of the elemental composition of the catalysts.For this purpose a solution containing 8 mL of aqua regia and 2 mL hydrofluoric acid was prepared and the measurements were performed after complete dissolution of the catalysts within this mixture.The hydrogen temperature programmed reduction (H 2 -TPR) experiments were carried out with an AC 2920 equipped with a CryColler-unit (Mircomertics, Waltham, MA, USA).The experiments were performed after the following procedure: 80 up to 100 mg of each sample were preheated to 500 • C for 30 min in synthetic air (50 mL/min, 20 • C/min).After cooling down to −50 • C the measurement was started.Therefore each sample was heated up to 800 • C in a mixture of 5% H 2 in Ar (20 mL/min, 5 • C/min).At 800 • C the reduction was performed for 30 min.The hydrogen consumption during the whole experiment was determined by an online thermal conductivity detector.

Catalyst Testing of Methane Combustion Ability
Catalytic methane combustion tests were conducted in a horizontal fixed-bed plug-flow quartz tubular reactor (inner diameter of 8 mm) at ambient pressure.0.2 g of catalysts were placed in the isothermal reactor zone and fixed by quartz wool located downstream below the catalyst bed.The determination of the catalytic properties under dry and wet reaction conditions were performed with the methodologies as described in ref. [24].

Conclusions
Porous silica prepared from rice husks (RHS) is an effective catalyst support for separating the Pd-CeO 2 system yielding a highly reactive catalyst for complete catalytic methane oxidation at low temperatures even under wet conditions.In contrast to expensive controlled porous glass made from phase separated borosilicate glass after leaching, the RHS appears to be a promising catalyst support for real applications as it is economically viable, environmentally friendly, sustainable and available in large amounts.Potentially, RHS from rice production could help to reduce methane release from industrial processes to prevent our ecosystem from further damages.

Table 1 .
[28]specific surface area (S BET ) calculated according to Ref.[28], average pore volume (V t ) as well as element analysis data.