Temperature and frequency characteristics of the interfacial capacitance in thin-film barium–strontium–titanate capacitors

Abstract

Thin film Au/Ba0.5Sr0.5TiO3/SrRuO3 capacitor structures, with a thickness of dielectric varying between ∼70 and ∼950 nm, were deposited on {001} MgO single-crystal substrates using pulsed laser deposition. Low-field dielectric measurements were performed as a function of temperature and frequency. At all temperatures and frequencies, the dielectric response as a function of thickness was found to adhere reasonably well to the so-called “series capacitor model,” from which nominal “bulk” and “interfacial” capacitance components could be extracted. The bulk component showed weak frequency dependence but strong temperature dependence, with a peak in permittivity and dielectric loss around 250 K and 150 K, respectively. Well above 250 K, reasonable Curie–Weiss behavior was evident. Overall, the extracted bulk component behaved much as would be expected in real bulk ceramics or single crystals lending confidence as to the general applicability of the series capacitor model. The functional behavior of the extracted interfacial capacitance was rationalized as being due to a combination of a thermally independent background, and thermally activated space charge. The activation energy of the space charge (∼0.6 eV) is commensurate with the detrapping of electrons from shallow level traps associated with oxygen vacancies. Importantly, since this component acts in series with the bulk component, the functional analysis performed here implies that the oxygen vacancies lie in a plane parallel to the electrodes.

Thickness dependence of intrinsic dielectric response and apparent interfacial capacitance in ferroelectric thin films

“…Experimentally, it was revealed that the inverse of the measured capacitance varies linearly with the film thickness. 1,2,[9][10][11][12][13][14][15] This observation led to a widely accepted interpretation of the experimental data in terms of a "series capacitor model." 9 According to such model, the measured capacitance density c t can be described by the relation…”

Response to “Comment on ‘Application of the interface capacitance model to thin-film relaxors and ferroelectrics’” [Appl. Phys. Lett. 89, 196101 (2006)]

“…Respectively, both the interface capacitance and the size-driven changes in the body of the film contribute to the thickness variation of the measured permittivity, as it has been emphasized in the paper. Although the mentioned almost linear dependence d / Јϰ d can be observed in some thin-film heterostructures 5 ͑and not in others 6 ͒, it can hardly be used for determining the interface parameters.…”

“…Thus the often observed fair linear fit of the dependence of d / Ј on thickness d ͑usually four to maximum ten experimental points͒, with the temperature dependent fitting parameters, cannot be unhesitatingly explained as originating solely from the interface capacitance. Moreover, the accompanying thickness dependence of the temperature of the dielectric maximum 5 indicates the changes in the phase diagram of thin film. Respectively, both the interface capacitance and the size-driven changes in the body of the film contribute to the thickness variation of the measured permittivity, as it has been emphasized in the paper.…”

Comments on “Application of the interface capacitance model to thin film relaxors and ferroelectrics” [Appl. Phys. Lett. 88, 262904 (2006)]

“…͑1͒ in temperature space-adherence has been observed at all temperatures at which it has been applied. 4 In addition, Tyunina and Levoska argue that the apparent straight line fit to data in the paraelectric state, in their CurieWeiss plots, demonstrates that there cannot be temperature dependence in any "series capacitance component." Figure 1͑a͒ below is a plot on a similar scale to those in Fig.…”

“…On this scale, the Curie-Weiss plot looks to be linear from 340 to 400 K. However, plots of the form of Eq. ͑1͒ were previously performed on this system at a series of different temperatures 4 and showed that the value of K was actually temperature dependent. There is thus an apparent conflict.…”