Strongly enhanced oxygen ion transport through samarium-doped CeO2 nanopillars in nanocomposite films

Abstract

Enhancement of oxygen ion conductivity in oxides is important for low-temperature (<500 °C) operation of solid oxide fuel cells, sensors and other ionotronic devices. While huge ion conductivity has been demonstrated in planar heterostructure films, there has been considerable debate over the origin of the conductivity enhancement, in part because of the difficulties of probing buried ion transport channels. Here we create a practical geometry for device miniaturization, consisting of highly crystalline micrometre-thick vertical nanocolumns of Sm-doped CeO2 embedded in supporting matrices of SrTiO3. The ionic conductivity is higher by one order of magnitude than plain Sm-doped CeO2 films. By using scanning probe microscopy, we show that the fast ion-conducting channels are not exclusively restricted to the interface but also are localized at the Sm-doped CeO2 nanopillars. This work offers a pathway to realize spatially localized fast ion transport in oxides of micrometre thickness.

Figure 1. Crystal structure of the nanoscaffold SDC–STO film.

(a) ‘Nano-comb'-like spontaneous phase ordering in cross-sectional view of the nanoscaffold SDC–STO film, as revealed by cross-sectional TEM image. Scale bar, 100 nm. (b) High-resolution TEM image of vertical SDC–STO interfaces in cross-sectional view. Scale bar, 10 nm. (c) Out-of-plane epitaxial relationship investigation in θ–2θ scan using X-ray diffraction. (d) Reciprocal space maps about the STO substrate for a plain SDC film (left) and a SDC–STO nanoscaffold film (right).

Figure 2. Transport properties of nanoscaffold SDC–STO films.

(a) Frequency dependence of the real part of ac conductivity σ_ac′ measured in the nanoscaffold SDC–STO film with variation of temperature. The ionic conductivity σ is defined from the plateaus of σ_ac′, as indicated by solid circles. The error bars in the solid circles represent the possible spread in plateau positions. (b) Temperature dependence of σ for the nanoscaffold SDC–STO film. For comparison, we include those for the plain SDC, YSZ and STO films that we grew. The error bars of the σ values represent the small variation of the plateaus of σ_ac′. We also include the σ values for GDC, Sm_0.075Nd_0.075Ce_0.85O_2−δ(SNDC), GDC–YSZ nanocomposites and SDC–YSZ multilayers reported in the literature.

Figure 3. Nanoscale mapping of electrochemical oxygen redox process at room temperature.

(a) Topographic image of the nanoscaffold SDC–STO film surface. The bright circular regions of ∼30 nm diameter are the SDC columns intersecting the film surface. The image size is 300 × 300 nm². Scale bar, 60 nm. (b) The bias waveform used for ESM measurement. (c) Spatial map of the ESM resonance frequency analysed using a simple harmonic oscillator model fitting, measured at the dc bias of −7 V. The open triangles in a,c show that the probing regions are identical. (d) ESM hysteresis loops with the maximum voltage of 10 V, averaged over one particular SDC column and its surrounding STO region, represented by the open box in h. (e–h) Spatial maps of ESM hysteresis loop area as a function of the peak bias. (i) ESM loop areas as a function of the peak bias averaged over the 30 points of the SDC and STO regions.

Figure 4. Spatially resolved mapping of ion-conducting channels.

(a) Topographic image of the nanoscaffold SDC–STO film. The image size is 250 × 125 nm². Scale bar, 50 nm. (b) The bias waveform used for the FORC-IV measurement and time-dependent current response averaged over the 40 × 20 points. Spatial maps of (c) current response and (d) relative FORC-IV loop area measured at −8 V. The boxed regions in a,c show that the probing regions are identical. Current versus tip bias curves at (e) the SDC nanopillars core and (f) the STO region.