Supercharged ceria quantum dots with exceptionally high oxygen buffer action

Abstract

Herein we present the exceptionally high oxygen buffer action of ceria quantum dots of size 3.7 nm. Ceria nanoparticles were obtained via a facile ammonia precipitation route, and were characterized using TEM, XRD, BET surface area analysis, XPS analysis and TPR study using H₂ as the probe molecule. A supercharging effect that leads to low temperature oxygen release is observed in these ultra-fine ceria particles. Existence of under stoichiometry in the ceria nanostructures with practically no oxygen vacancies has been substantiated with experimental proof.

---

1. Introduction

Cerium oxide is a well acclaimed candidate among the rare earth oxides. Marked by its fine properties, it has great potential for environmental and energy-related applications. Some of them include catalysis ,^1–4 sensors,⁵ fuel cells⁶ etc. A few years back, a non-traditional application of ceria in the biomedical field was also highlighted.^7,8 Behind the versatile usage of ceria lies one remarkable feature, namely, its exceptional redox property, which enables it to perform as an excellent oxygen storage material.^9,10 The work by Namai et al. deserves attention in this context, which cites defect centre mobility even at room temperature due to hopping of the lattice oxygen.¹¹ Experimental studies on oxygen storage capacity (OSC) of ceria indicate that the source of released oxygen varies with the size of the crystallites.

Temperature programmed reduction (TPR) using probe molecules is a commonly used analytical tool for the measurement of OSC in metal oxides, where the amount of reducing agent consumed gives the measure of available oxygen in the system. For smaller ceria crystallites, apart from the peak corresponding to oxygen release at 800 °C signifying bulk lattice oxygen release, another peak has been as observed in the reduction profile at 580 °C, which has been assigned to the reduction of surface oxygen that is quite different from that of the bulk.¹² In an exhaustive report on the size dependent oxygen buffer action of nanoceria, Xu et al. observed that particles of size less than 10 nm (especially <5 nm) show oxygen release at a further low temperature (∼400 °C) also, which the authors attributed to superoxo ions (O₂⁻) using electron paramagnetic resonance (EPR) spectral methods.¹³ More clarity to this observation has been imparted by Kullgren et al. who made a theoretical study on the oxygen chemistry responsible for the increased OSC in smaller ceria particles (diameter d < 5 nm).¹⁴ According to them, the conventional OSC mechanism based on the explicit formation of oxygen vacancies arising from the presence of Ce³⁺ ions is inadequate to explain this oxygen release at low temperature. This phenomenon was attributed to the structure of the nanoparticle itself, which was also hinted at by the previous researchers. They proposed that very small octahedral/truncated octahedral ceria nanoparticles will have under stoichiometric surfaces without extrinsic oxygen vacancy, i.e., the surface will be oxygen deficient without containing specific vacancy sites, and can be reduced at the above mentioned lower temperature range. Usually, the term under stoichiometry refers to a lower amount of metal present in the system than that predicted by the stoichiometry. However, Kullgren et al. used the term with respect to oxygen stoichiometry; for systems with Ce_xO_2(x−L) stoichiometry, ‘L’ representing excess Ce atoms in the ceria polyhedra. Such under stoichiometric systems are expected to display a supercharged OSC, which will release oxygen at lower temperatures (∼400 °C).

It has been established that the synthesis route has a profound influence on the oxygen storage capacity of cerium oxide. The present report deals with the exceptionally high oxygen storage capacity of supercharged nanoceria structures of size 3.7 nm achieved via a facile ammonia precipitation route, first suggested by our own group.¹⁵ The oxygen storage capacity of the prepared system is the highest value reported so far, when matched with systems of comparable size. The low temperature oxygen release, a highly recommended quality for a low temperature oxidation catalyst, is also commendably high for our material. We also confirm that very small under stoichiometric ceria nanoparticles without evident oxygen vacancies have unusual oxygen storage, which needs less activation, and can be reduced at lower temperatures. A nil oxygen vacancy concept even in the presence of Ce³⁺ ions is generally difficult to digest and approve. Here we provide concrete proof for this observation with the help of X-ray Photoelectron Spectroscopy (XPS).

2. Experimental

CeO₂ nanoparticles were prepared by a process that involves a homogeneous precipitation from a cerium nitrate (Ce(NO₃)₃·6H₂O) solution with dilute aqueous ammonia. All chemicals were purchased from Aldrich (Germany), purity 99.99%. A requisite quantity of the precursor was dissolved in double distilled water followed by the dropwise addition of the solution to aqueous ammonia over a period of time. The obtained light yellow slurry was washed thoroughly with deionised water, followed by ethanol. The collected precipitate was oven-dried at 383 K for 12 h, crushed using an agate mortar, and calcined at 500 °C to obtain the oxide.

The micro structure and morphology of the samples were examined by transmission electron microscopy (TEM) images with a Philips CM 200 transmission electron microscope operating at 20–200 kV range. Sample grids were prepared by sonicating powdered samples in ethanol for 20 min and evaporating one drop of the suspension onto a carbon-coated, holey film supported on a 3 mm, 300-mesh copper grid. Powder X-ray diffraction (XRD) patterns of the samples were collected on a Rigaku D/MAX-diffractometer equipped with a rotating anode using CuKα radiation. N₂ adsorption studies for the measurement of surface area were done using a Micromeritics Gemini Surface Area analyser through a static adsorption procedure at 77 K. Samples were degassed at 150 °C in a vacuum below 10⁻³ Torr for 16 h prior to the measurements. DR UV-vis spectra were taken in the range 200–800 nm with BaSO₄ as the reference using a Jasco V-550 spectrometer. XPS measurements were carried out on a VG Micro Tech ESCA 300° instrument at a pressure of >1 × 10⁻⁹ Torr (pass energy of 50 eV, electron take off angle 60° and the overall resolution was ∼0.1 eV). Temperature-programmed reduction studies were performed using a Micromeritics Pulse Chemisorb-2705, fitted with a thermal conductivity detector (TCD). The samples (300 mg) were activated/surface-cleaned in helium at 200 °C for 30 minutes, then cooled to ambient temperature. The reactive gas composition was 10% balance N₂, its consumption was measured while heating the sample up to 800 °C at a rate of 10 °C min⁻¹.

3. Results and discussion

3.1. Preliminary characterisations

The size and morphology of the product were examined by transmission electron microscopy images (Fig. 1), which revealed particles of size <10 nm in the sample. For smaller particles, in the size range of 3–10 nm, it has been reported that the particle shapes are predominantly truncated octahedral structures defined by {100} and {111} facets.¹⁶ This was also validated from the high resolution TEM image presented in the same figure, which shows particles of size 3.7 nm. Comparatively high agglomeration was also seen, which is naturally expected based on the ultra-small size of the particles.

Fig. 1 TEM images (a) and (b).

Fig. 2 shows the powder XRD pattern of the prepared nanoceria. Broadening of the reflections in the diffractogram distinctly indicates the formation of nanocrystals. The material is pure fluorite cubic in structure (JCPDS card no. 34-0394, space group Fm3m), and there was no sign of crystalline Ce₂O₃ in the system. The mean crystallite size (D) of the ceria system was determined from the broadening full width at half maximum of the XRD peaks, through Scherrer’s formula. Notably, the particle dimension (D = 3.5 nm) agrees with the TEM outcome. The lattice parameter value (a = d₁₁₁√3; d = inter planar distance) is 0.5501 nm, and the lattice expansion for the system approached ∼2%. These values are in agreement with the reports by Hailstone et al., who also proposed a relation for ceria particles of lower dimensions, a = a_bulk + 0.036/D, where a_bulk is the lattice parameter corresponding to bulk cerium oxide.¹⁷ Our system seems to obey this proposed relation. N₂ adsorption studies using the Brunauer–Emmett–Teller (BET) method yielded a significantly high surface area for the material, 128.24 m² g⁻¹.

Fig. 2 XRD pattern of the system.

Fig. 3 illustrates the UV-vis absorption spectra of the prepared material and that of bulk ceria (Sigma-Aldrich, 99.9%). It can be seen that the absorption edge of nano CeO₂ is blue-shifted when compared to that of the bulk material. Band gap energy values calculated from the corresponding Kubelka–Munk plots were 3.00 eV and 2.8 eV respectively, for the present sample and the bulk counterpart. Since the typical size of the particles is less compared to twice the value of the exciton Bohr radius of the material (∼7–8 nm), the optical band gap shift should be classified to the strong confinement regime, which is typical of quantum dots.

Fig. 3 UV absorbance spectra of the prepared and bulk ceria.

3.2. Redox chemistry and oxygen storage capacity of the system

XPS is an identified and conventional tool for the prediction of the different oxidation states of metal ions. The Ce 3d and O 1s XPS patterns of the prepared ceria sample are provided in Fig. 4. Curve fitting enabled the identification of ten individual peaks in the 3d XPS spectra of our ceria sample. Six peaks correspond to the Ce⁴⁺ entity, and four signify the Ce³⁺ species. Two sets of spin–orbital multiplets, corresponding to the 3d_3/2 and 3d_5/2 contributions were labelled as u and v, respectively. Peaks u′′′ and v′′′ originate from the Ce 3d⁹ O 2p⁶ Ce 4f⁰ final states of Ce⁴⁺. The additional states of Ce⁴⁺, u, v, u′′ and v′′ result from a mixture of the Ce 3d⁹ O 2p⁵ Ce 4f¹ and Ce 3d⁹ O 2p⁴ Ce 4f² final states. The contribution of Ce³⁺ to the Ce 3d spectrum consists of two doublet pairs: u′, v′ and uo, vo. These doublets correspond to a mixture of the Ce 3d⁹ O 2p⁵ Ce 4f² and Ce 3d⁹ O 2p⁶ Ce 4f¹ final states. As the sum of the peak area corresponding to the ions is proportional to the concentration of ions, the concentration of the two oxidation states in the oxide sample can be obtained. This method directly yields the ratio of intensities of the Ce³⁺ and Ce⁴⁺ ions from the area under their respective peaks as reported by earlier workers.^18,19 The Ce³⁺ percentage was 28.7 in the prepared nanoceria. It was observed that the percentage of 3+ ions levelled off to ∼29% when the particle size decreased below 10 nm.¹³ Our result also supports this conclusion. The direct O/Ce ratio (the ‘x’ parameter in CeO_x, which should be 2 for the stoichiometric cubic ceria lattice) derived from the XPS data was 1.58, which was consistent with the corresponding ratio derived from EDAX measurements. Thus, both tools identified an under stoichiometric cubic fluorite ceria.¹⁴ Hence, it can be seen that though the XRD study ruled out the possibility of Ce₂O₃, where ceria existed in the Ce³⁺ state, the presence of the trimeric ion has been shown by X-ray photoelectron spectroscopy. This can be accounted for by the reduced size, which makes the system oxygen deficient. Note that even if amorphous Ce₂O₃ exists in the sample, the ‘x’ parameter should be 1.74 as per the respective stoichiometry calculation.²⁰ In any case, the value is <2, and the under stoichiometry in cubic fluorite ceria is quite evident.

Fig. 4 XPS spectra of ceria after curve fitting. Ce 3d spectrum (a) and O 1s spectrum (b).

Useful data regarding the oxygen chemistry of the system can be derived from the O 1s XPS spectrum. It can be seen that the total oxygen is made of three components, the relative percentage of each being provided in parenthesis. The peak at lower binding energy (529.8 eV) accounts for the lattice oxygen/structural oxygen (8.34%), while the one at 532.1 signifies dissociatively adsorbed surface oxygen in the form of OH ions (70.45%).²¹ These two XPS peaks are those usually detected in nanoceria systems. The additional XPS peak seen at 531 eV accounts for the supercharged oxygen (O₂⁻) at the surface (21.21%) due to the particle structure.¹³ This specific peak signifies the low temperature oxygen storage (or better oxygen releasing) capacity of the structure. Wang et al. noticed a peak at 530.4 eV also, and assigned it to the oxygen present near oxygen vacant sites in the ceria matrix.²¹ Usually, the presence of Ce³⁺ is supposed to be accompanied by oxygen vacancies in the nanoceria lattice. The absence of this peak hints that oxygen vacancies do not exist in the system, which matches the situation as suggested by Kullgren et al.¹⁴

The temperature programmed reduction plots shown in Fig. 5 compare the OSC of the system of our concern and another reference system. Since the supercharging effect was observed below 10 nm; more specifically in particles of size <5 nm,¹³ ceria particles of a higher dimension, size 11 nm (surface area 40 m² g⁻¹), were chosen as the reference. A comparison of the two TPR profiles hints at size dependent oxygen buffer action, which is in accordance with previous reports. Both the systems show signals at ∼520 °C and at ∼770 °C. The peak ‘c’ corresponds to bulk oxygen release, while ‘b’ represents surface oxygen (oxygen on the reduced bigger cerium ion, Ce³⁺, which requires lesser activation for removal, and hence lower temperature) liberation. The one centred at ∼375 °C (peak depicted as ‘a’), appearing for the 3.7 nm sized particles is of interest for discussion of the topic. This is the typical peak corresponding to oxygen adsorbed on nanoceria due to supercharging, i.e., the chemisorption of oxygen due to the structure of under stoichiometric small truncated octahedral species, leading to O₂⁻. The XPS spectra of the reference system is displayed in Fig. 6, wherein the notable absence of a peak at 531 eV is seen when compared to smaller particles, indicating the absence of a supercharging effect in the system, as expected. The relative percentages of lattice oxygen (corresponding to 529.8 eV) and surface oxygen (corresponding to 532.1 eV) were 48.9 and 51.1% respectively.

Fig. 5 TPR profiles of ceria nanoparticles of different sizes.

Fig. 6 XPS spectra of 11 nm ceria particles. The Ce 3d spectrum (a) and the O 1s spectrum (b).

The dramatically high OSC of the smaller particles is quite evident from the TPR profile, which may be attributed to oxygen adsorption due to a supercharging effect. It was observed that the OSC leveled off to ∼717 μmol g⁻¹, when the size was above 5 nm (Xu et al., 21.5 μmol/30 g), which is in good agreement with our reports. The general trend was an increase in OSC as the particle size decreased, and a size reduction to 4.4 nm enhanced the parameter to 1223.3 μmol g⁻¹ as per their observation. This result can also be extended to our system where the amount was 1550.3 μmol g⁻¹ for 3.7 nm; the highest value reported in the literature, for systems of almost comparable sizes. The low temperature oxygen storage shows a dramatic increase in our system. Relevant data in this regard are provided in Table 1. This in turn can be attributed to two factors; (i) the increase in surface area per mole of the system and (ii) the structure of the particle itself that facilitates superoxo decoration on the surface of these ultra-small polyhedra; the latter contributing more towards the oxygen storage at lower temperature region. Rather than becoming stoichiometric by absorbing oxygen from the surroundings, the particles stabilize themselves through oxygen chemisorption forming superoxo (O₂⁻) ions. Electron transfer takes place from the Ce³⁺ ions to the adsorbed O₂ molecules. In the case of the truncated octahedral particles as discussed here, superoxide formation is predicted on corners, ridges, and on the Ce terminated {100} facets. These supercharged particles are found to be more energetically stable than their fully oxidized counterparts in an oxygen containing atmosphere. These observations authenticate the commendable low temperature oxygen release by the present system.

Table 1 Oxygen storage data of ceria nanoparticles

Particle size of ceria system	Oxygen release calculated from TPR data (μmol g^−1) at various peak positions
	375 °C	525 °C	775 °C
11 nm	—	370.4	346.6
3.7 nm	1005.9	288.4	256.0

---

4. Conclusions

A simple ammonia precipitation route yielded ultra-small cubic ceria nanoparticles which showed the highest oxygen storage activity ever reported, when matched with particles of comparable size. Supercharging by oxygen has been observed in the system and is attributed to the structure of the ultra-small particle itself, rather than oxygen vacancy creation. The existence of no oxygen vacancies in these under stoichiometric particles has been proven experimentally.

Notes and references

1. D. Valechha, S. Lokhande, M. Klementova, J. Subrt, S. Rayalu and N. Labhsetwar, J. Mater. Chem., 2011, 21, 3718.
2. X. Lin, Y. Zhang, Z. Wang, R. Wang, J. Zhou and K. Cen, Energy Convers. Manage., 2014, 84, 664.
3. C. S. Pan, D. S. Zhang and L. Y. Shi, J. Solid State Chem., 2008, 181, 1298.
4. A. Asati, S. Santra, C. Kaittanis, S. Nath and J. M. Perez, Angew. Chem., 2009, 48, 2308.
5. D. A. Andersson, S. I. Simak, N. V. Skorodumova, I. A. Abrikosov and B. Johansson, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 3518.
6. N. G. Hurtado, I. Atribak, A. B. Lopez and A. G. García, J. Mol. Catal. A: Chem., 2010, 323, 52.
7. C. W. Younce, K. K. Wang and P. E. Kolattukudy, Cardiovasc. Res., 2010, 87, 665.
8. S. Das, J. M. Dowding, K. E. Klump, J. F. McGinnis, W. Self and S. Seal, Nanomedicine, 2013, 8, 1483.
9. F. Esch, S. Fabris, L. Zhou, T. Montini, C. Africh, P. Fornasiero, G. Comelli and R. Rosei, Science, 2005, 309, 752.
10. Catalysis by Ceria and Related Materials, ed. A. Trovarelli, Imperial College Press, London, 2002, p. 15.
11. Y. Namai, K. Fukui and Y. Iwasawa, J. Phys. Chem. B, 2003, 107, 11666.
12. A. Tschope, D. Staadt, R. Birringer and J. Y. Ying, Nanostruct. Mater., 1997, 9, 423.
13. J. Xu, J. Harmer, G. Li, T. Chapman, P. Collier, S. Longworth and S. C. Tsang, Chem. Commun., 2010, 46, 1887.
14. J. Kullgren, K. Hermansson and P. Broqvist, J. Phys. Chem. Lett., 2013, 4, 604.
15. N. K. Renuka, J. Alloys Compd., 2012, 513, 230.
16. Z. L. Wang and X. Feng, J. Phys. Chem. B, 2003, 107, 13563.
17. R. K. Hailstone, A. G. D. Francesco, J. G. Leong, T. D. Allston and K. J. Reed, J. Phys. Chem. C, 2009, 113, 15155.
18. J. Z. Shyu, K. Otto, W. L. Watkins, G. W. Graham, R. K. Belitz and H. S. Ganghi, J. Catal., 1988, 114, 23.
19. F. Zhang, P. Wang, J. Koberstein, S. Khalid and S. W. Chan, Surf. Sci., 2004, 563, 74.
20. P. Patsalas and S. Logothetidis, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 68, 035104.
21. X. Wang, Z. Jiang, B. Zheng, Z. Xie and L. Zheng, CrystEngComm, 2012, 14, 7579.

---