High Surface Area Ceria for CO Oxidation Prepared from Cerium t-Butoxide by Combined Sol–Gel and Solvothermal Processing
- 2k Downloads
CeO2 was synthesized by combined sol–gel and solvothermal processing of gels obtained from acetaldoximate-modified cerium(IV) t-butoxide in the presence of the non-ionic surfactant Pluronic F127. The use of cerium(IV) t-butoxide as precursor contrasts very favorably with the often used ceric ammonium nitrate and results in more reliable and tailorable properties of the final materials. The kind of post-synthesis treatment of the gels and the composition of the precursor mixture proved to be crucial for obtaining high surface area ceria with a high Ce3+ proportion. Calcination in air or under nitrogen was compared with solvothermal treatment in ethanol or water and a combination of solvothermal treatment and calcination. The obtained materials are composed of 3.5–5.5 nm ceria nanoparticles. The highest specific surface area of 277 m2/g was obtained after solvothermal treatment, and 180 m2/g when solvothermal treatment was followed by calcination in air to remove residual organic groups. The highest Ce3+ proportion was 18 % after solvothermal treatment in ethanol and additional calcination in air. CO oxidation on selected samples indicated that activity scaled with surface area and thus was largest for samples solvothermally treated in ethanol. The reaction rate of the best sample was about 75-times larger than that of commercial ceria.
KeywordsCeria Cerium butoxide Sol–gel processing Solvothermal treatment CO oxidation
Ceria (CeO2) has received attention in environmental catalysis because of its high oxygen storage-release capacity associated with the Ce4+/Ce3+ redox cycle . It was suggested that nano-scaled ceria with abundant oxygen vacancies would further enhance the catalytic activity [2, 3].
Various methods have been employed to synthesize CeO2, including the sol–gel method . The latter has many advantages because the materials composition and texture can be influenced by the proper choice of precursors and processing conditions. Inorganic cerium salts, such as Ce(NO3)3·6H2O [5, 6, 7, 8, 9] CeCl3·7H2O  or ceric ammonium nitrate (NH4)2[Ce(NO3)6] (CAN) , were mainly used as precursors for sol–gel processing of ceria (including the citrate-gel route) because they are commercially available. Qi et al.  found that the choice of the precursor [Ce(NO3)3 and (NH4)2Ce(NO3)6] influenced the structure, surface state, reducibility and CO oxidation activity of Cu doped CeO2 materials.
Cerium alkoxides, Ce(OR)3 or Ce(OR)4, are less often employed as precursors because they are not easily available commercially and more reactive to humidity. Nevertheless, metal alkoxides have often advantages because the materials textures and properties can be easily modified by varying the reaction parameters. Cerium alkoxides were used, for example, to prepare ceria thin films [13, 14] or bimetallic oxides [15, 16, 17].
Apart from the precursor, the post-synthesis treatment of the gels is also an important parameter influencing the materials properties. Gels obtained by sol–gel processing are often heat-treated in air, sometimes also in N2 or a combination of both, to remove residual organic groups. Recently, a combined sol–gel and solvothermal process was applied for the preparation of titania which possessed a special microstructure with high surface area [18, 19].
In the work reported here, the influence of different precursors (CAN and cerium(IV) t-butoxide (Ce(OtBu)4), CeB) and of different post-synthesis treatment of the gels on the morphology and microstructure of ceria was investigated. Furthermore, the catalytic activity of various CeO2 samples for CO oxidation was tested, with the specific surface area (SBET) and the Ce3+ (surface) concentration being very important parameters.
All solvents were dried by standard methods, and all experiments involving metal alkoxides were carried out under moisture- and oxygen-free argon using standard Schlenk or glove box techniques. (NH4)2[Ce(NO3)6] was obtained from Alfa-Aesar and used as received. Commercial CeO2 powder from Aldrich was used for comparison.
2.1 Synthesis of Cerium t-Butoxide
Cerium(IV) t-butoxide (Ce(OtBu)4, CeB) was synthesized according to the literature from (NH4)2[Ce(NO3)6] (CAN) and sodium t-butoxide in 1,2-dimethoxyethane [22, 23]. The oily residue obtained after removal of the solvent was used directly for sol–gel processing without further purification. X-ray photoelectron spectra (XPS) proved that the sample contained no residual sodium from the preparation.
2.2 CeO2 Synthesis
Holding at 500 °C in air for 2 h in an Al2O3 crucible; heating rate 2°/min (AC).
Solvothermal treatment (ST) at 200 °C for 6 h (ST) in a 60 mL autoclave with 30 mL solvent (gel prepared from 5 mmol of CeB). Two different solvents were used, viz. EtOH (STE) and H2O (STH).
- ST, followed by calcination as described for AC (ST-AC) (Scheme 1).
Thermogravimetric analysis (TGA) was performed on a Netzsch Iris TG 209 C in a platinum crucible in synthetic air with a heating rate of 10 °C/min. Infrared spectra (IR) were recorded on a Bruker Tensor 27 working in ATR Micro Focusing MVP-QL with a ZnSe crystal, using OPUS software version 4.0 for analysis.
X-Ray powder diffraction (XRD) measurements were performed on a Philips X’Pert diffractometer using Cu-Kα radiation (λ = 1.5406 Å). Scanning electron micrographs (SEM) were obtained on a Quanta 200 (FEI) equipped with a Genesis (EDAX) energy dispersive spectrometer, with a voltage of 7.5 kV. XPS were measured on a Specs XPS system (XR 50 Mg/Al-Dual Anode, PHOIBOS 150 hemispherical analyzer).
High resolution transmission electron micrographs (HRTEM) were recorded on a TECNAI F20 operated at 200 kV. Before the measurements, the samples were ultrasonically dispersed in EtOH for 30 min, and then deposited on copper grids covered with carbon films.
Nitrogen sorption measurements were performed on an ASAP 2020 (Micromeritics). The samples were degassed in vacuum at room temperature for at least 5 h prior to measurement. The total surface area was calculated according to Brunauer, Emmett and Teller (BET), and the pore size distribution (from the desorption branch) according to Barrett, Joyner and Halenda (BJH).
2.4 Catalytic Activity
The CO oxidation reaction was performed in a continuous-flow fixed-bed quartz reactor under atmospheric pressure. An amount of 20 mg of each sample was loaded into the reactor and pretreated with synthetic air (30 mL/min) at 200 °C for 40 min (heating rate 10 °C/min). Then the sample was cooled to 30 °C in flowing synthetic air, and a mixture of 5 vol% CO, 10 vol% O2 and 85 vol% He (total flow 50 mL/min) was introduced. The system was then heated to 650 °C with a ramping rate of 5 °C/min. The concentrations of CO and CO2 in the outlet streams were monitored by gas chromatography with a HP-PLOT Q column and a flame ionization detector.
CO temperature-programmed reaction (CO-TPR) over ca. 20 mg samples were also performed in a continuous-flow fixed-bed quartz reactor under atmospheric pressure. The catalyst samples were pre-treated with synthetic air at 200 °C for 40 min (heating step 10 °C/min) at a flow rate of 30 mL/min. After cooling to room temperature the samples were exposed to a mixture of 5 vol% CO and 95 vol% He (total flow 50 mL/min) at room temperature. Then the system was ramped up to 900 °C at a heating rate of 10 °C/min. The gas stream was analysed by an online quadrupole mass spectrometer (QMS) (Prisma Plus QMG 220, Pfeiffer Vacuum) equipped with a Faraday detector.
3 Results and Discussion
3.1 Gel Preparation
Sols were prepared from CeB (or CAN), AO, F127 and 1,2-dimethoxyethane (DME) in a 1:2:0.005:40 ratio. In some reactions no F127 was added to check the influence of the surfactant. Gelation was induced by exposure to ambient humidity at room temperature. The obtained solid gels were then treated by different methods, viz. calcination in air (AC), ST in EtOH (STE) or H2O (STH), or ST, followed by calcination (ST-AC) (see Sect. 2.2).
When (NH4)2[Ce(NO3)6] (CAN) was used, the onset temperature was about 230 °C and the gel decomposed completely below 300 °C with a weight loss of about 72 %. This indicates that part of the material was lost during the TGA measurements due to deflagration. It should be pointed out that samples obtained from CAN sometimes exploded upon calcination. The solid obtained from CAN after sol–gel processing was water-soluble; thus no network had been formed with ambient moisture. The exothermic decomposition of CAN and the solubility of the obtained gel render CAN unappealing as precursor. For this reason CAN-derived samples will no longer be considered in the following.
In contrast, when CeB was used as a cerium precursor, the onset temperature in TGA was about 495 °C, and decomposition was complete at about 600 °C, with 22 % weight loss. The X-ray diffractogram after calcination (sample CeB/AO/F127-AC) showed that the sample was phase-pure ceria (PDF-No. 34-0394) with a particle size of 9.2 nm, as calculated by the Scherrer equation based on the strongest reflection at 28.6°.
We also tested whether thermolysis at 500 °C in N2 for 3 h, followed by calcination in air for 2 h would result in significantly different materials properties. The particle size was only slightly smaller (8.6 nm) and the specific surface area (see below) slightly higher. For this reason, only calcination in air was used in the following.
3.2 Influence of Different Post-synthesis Treatments
BET results for CeB:AO = 1:2 samples with different post-synthesis treatment
3.3 Influence of the Sol Composition
CeB: only CeB in DME in the ratio CeB:DME = 1:40.
CeB/AO: CeB and AO in DME in the ratio of CeB:AO:DME = 1:2:40.
CeB/F127: CeB and F127 in DME in the ratio CeB:F127:DME = 1:0.005:40.
CeB/AO/F127: CeB, AO and F127 in DME in the ratio CeB:AO:F127:DME = 1:2:0.005:40.
BET and XRD results summary of samples prepared by ST and ST-AC treatment
ST-AC500 °C series
3.4 Calcination After Solvothermal Treatment
The changes of CeB/AO/F127-STH were similar to that of the STE samples (Fig. 5, right). Significant differences between the STE-AC and the STH-AC samples were, however, that the new band at 990 cm−1 of the STH-AC sample became much stronger during calcination than that of the STE-AC sample, while the original bands at 924 and 852 cm−1 and other small bands became weaker.
The experiment indicated that calcination in air at 500 °C removes most of the residual organic groups without causing excessive sintering of the materials. For the following experiments, the samples were therefore treated at 500 °C.
BET measurements after calcination (Table 2) showed that the surface area and average pore size of the STH samples were not significantly changed. In contrast, the surface area of the STE-AC samples was significantly reduced compared with the corresponding STE samples while the average pore sizes only changed slightly. For example, CeB/AO/F127-STE had a surface area of 277 m2/g and an average pore size of 3.4 nm, but the corresponding values of CeB/AO/F127-STE-AC changed to 88.9 m2/g and 3.9 nm, respectively. The change of the surface area of the STE samples can possibly be attributed to the higher proportion of organic constituents present after ST and their removal upon AC. The surface area of the STE samples after calcination is in each case significantly lower than that of the STH samples, while the average pore diameters are in the same range.
The XRD pattern after calcination (Table 2) showed that all samples were nanocrystalline CeO2 with similar particle sizes in the range 3.5–5.5 nm (calculated by the Scherrer equation based on the strongest peak at 28.6°). Interestingly, this was only half the size than that of the particles obtained by calcination only, i.e. without ST. For comparison, we also analyzed the particle sizes of two samples after ST prior to calcination. The average particle size of CeB/AO/F127-STE was <3 nm (4.0 nm after calcination) and that of CeB/AO/F127-STH 5.1 nm (5.3 nm after calcination). The crystallite size of the final material is thus determined by the ST and is not significantly changed when the residual surface groups are removed upon subsequent calcination.
A possible explanation for the different particle sizes (and surface areas) between the AC and STE-AC or STH-AC series is that dissolution-deposition equilibria under solvothermal conditions determine the particle size. Similar results were found for ceria solvothermally treated in different alcohols . The obtained particles are capped by residual (organic) groups which are then removed upon calcination. Only crystallite sizes can be determined by XRD, but no information is obtained about agglomeration/aggregation of the crystallites. The lower BET surface area in the STE series (see Table 2) may thus be due to a more pronounced agglomeration/aggregation of the nanocrystals. In contrast, when the initially obtained gels were calcined without prior ST, removal of the organic groups and mass transport coincide and the formation of larger particles is thus favored.
3.5 Catalytic Properties
Properties of different ceria samples (20 mg) for CO oxidation
T10 % (°C)e
T90 % (°C)f
R250 °C (mol/s m2)g
r250 °C (mol/s g)h
2.32 × 10−8
6.42 × 10−6
2.52 × 10−8
2.24 × 10−6
2.75 × 10−9
3.64 × 10−7
5.76 × 10−8
8.63 × 10−8
For a better comparison of the catalytic activity of the different samples, the normalized reaction rates at 250 °C per gram (r250 °C) and per unit surface area (R250 °C) are also shown in Table 3. The sequence of r250 °C is CeB/AO/F127-STE > CeB/AO/F127-STE-AC > CeB/AO/F127-STH-AC > commercial CeO2. The reaction rate at 250 °C (r250 °C) of the best sample CeB/AO/F127-STE is about 75-times larger than that of commercial CeO2, that at 300 °C about 125-times larger.
The CO oxidation activity of the STE and STE-AC catalysts correlates with the specific surface area (Table 3). About three times higher surface area of STE produces about three times higher rates. This can be seen from comparing the reaction rate per surface area, i.e. the amount of CO converted on a unit surface area of the catalyst at 250 °C (R250 °C), which are basically identical. Accordingly, when the specific activities of the two catalysts CeB/AO/F127-STE and CeB/AO/F127-STE-AC are compared based on weight, the rates r250 °C (reaction rate of CO oxidation at 250 °C/g) show an about threefold difference. Commercial ceria roughly fits into this sequence; the specific surface area is about 180-times smaller than that of CeB/AO/F127-STE, which reduces r250 °C by a factor of 75.
The third sample CeB/AO/F127-STH-AC exhibits a specific surface area somewhat larger than CeB/AO/F127-STE-AC but shows lower rates (both R250 °C and r250 °C) by a factor of ten. This must be due to other parameters besides the surface area. In the literature different effects have been made responsible for activity differences of ceria. For example, the crystallite size of CeO2 is known to be a crucial factor for CO oxidation because size may affect the abundance of oxygen vacancies and thus of associated Ce3+ ions. Smaller crystallite sizes induce a higher fraction of oxygen vacancies due to lower barrier of vacancy formation [27, 28]. One of the proposed mechanisms for CO oxidation on CeO2 is oxidation by lattice oxygen of CeO2 with subsequent creation of oxygen vacancies (and Ce3+) which in turn induces formation of active oxygen species [29, 30].
Nevertheless, a correlation between the Ce3+ proportion and the specific catalytic activity was not observed. It may explain the difference between the specific catalytic activity of the STH-AC and STE-AC samples. On the other hand, however, STE and STH-AC have a comparable Ce3+ proportion but STE has one magnitude higher catalytic activity.
As discussed above, CO oxidation by ceria depends on the oxygen vacancies (Ce3+), but the surface OH can also contribute to the reaction. The same catalytic activity of STE and STE-AC may be due to the compensation of the Ce3+ proportion (which is lower for STE) and the influence of highly reactive OH groups resulting from the residual organic species in the STE sample. STE-AC and STH-AC do not contain such OH groups, as the organic species were largely removed during the calcination. Therefore, despite the missing OH groups, the specific activity exhibited by STE-AC is the result of high Ce3+ content. A possible explanation for the different specific catalytic activity may thus be the different CeO2 surface states formed in the different post-synthesis treatments, and the reductive ability of EtOH might be a key factor.
We have shown in this work that the use of cerium alkoxides as sol–gel precursor for ceria has clear advantages compared with (NH4)2[Ce(NO3)6]. Gels were prepared from acetaldoximate-substituted cerium t-butoxide in the presence of the surfactant F127. The latter served to create interparticle porosity after calcination. However, the post-synthesis treatment of the obtained gels had a larger influence on the resulting specific surface area than the composition of the starting mixture. ST in either ethanol or water resulted in materials with distinctly higher specific surface areas and smaller crystallite size than by calcination alone. Specific surface areas of up to 277 m2/g were obtained which is unprecedented for ceria (commercial ceria ranges around 2 m2/g). Post-synthesis calcination of the gels allows removing of the residual organic groups. Both the solvent used for ST and the calcination process can influence the Ce3+/Ce4+ ratio. The highest Ce3+ proportion was 18 % for CeB/AO/F127-STE-AC.
Tests of CO oxidation on selected samples indicated that a high catalytic activity is related to a high surface area and the surface groups created by the post-synthesis treatment. With regard to the latter, ST of the samples in ethanol proved to be particularly efficient. The sample with highest surface area (277 m2/g) from ST with ethanol outperformed commercial ceria by a factor of 75.
This project was funded by the Austrian Science Funds (FWF) in the framework of the Doctoral School “Building Solids for Function” (Project W1243) and SFB “Functional Oxide Surfaces and Interfaces” (Project F45). The authors thank Elisabeth Eitenberger for SEM measurements, Christina Artner for the XRD measurements, and Stefan Löffler and Michael Stöger-Pollach for HRTEM characterization at the USTEM of TU Wien.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.