Structural dependence of terahertz radiation from multiferroic $\mathrm{Bi}\mathrm{Fe}{\mathrm{O}}_3$ thin films

We have investigated the structural dependence of terahertz-radiation emission from the multiferroic $\mathrm{Bi}\mathrm{Fe}{\mathrm{O}}_3$ thin films on ${\left(\mathrm{La}\mathrm{Al}{\mathrm{O}}_3\right)}_{0.3}{\left({\mathrm{Sr}}_2\mathrm{Al}\mathrm{Ta}{\mathrm{O}}_6\right)}_{0.7}$ (0 0 1) substrate. The coherently strained films with thickness $\left(t\right)⩽75\mathrm{nm}$ relax partially to bulklike rhombohedral phase at $t⩾110\mathrm{nm}$ via a coexistence of these two phases in the thickness $80\mathrm{nm}⩽t⩽110\mathrm{nm}$. The strained films $(t⩽75)$ and the relaxed films $(t⩾110\mathrm{nm})$ exhibit similar terahertz-emission efficiency in the absence of electric-field bias, and this efficiency increases for the intermediate thickness films having coexisting phases. These data combined with the electric-field $\left(E\right)$ dependence of the terahertz-emission amplitude $\left(E_{\mathrm{THz}}\right)$ depict that the terahertz-emission efficiency in zero bias is nearly independent of the lattice, while the ferroelectric properties (revealed by $E_{\mathrm{THz}}\text{−}E$ hysteresis loops) are characteristics of both the lattice and the leakage current.

Figure 1

(Color online) (a) Peak amplitude $\left(E_{\mathrm{THz}}\right)$ of the time-domain wave form in zero-bias electric field obtained after poling in $+200\mathrm{kV}∕\mathrm{cm}$ for various thin films. The inset figure shows a typical terahertz time-domain wave form.

Figure 2

(a) Peak terahertz amplitude as a function of structural phase fraction (fully strained/relaxed). (b) XRD pattern around (001) peak for all the films; the dashed vertical lines indicate the BFO peaks [peaks at lower and higher $2\theta$ angles represent the strained phase and the (partially) relaxed phase, respectively]. (c) The left panel shows peak terahertz amplitude versus the structural phase ratio (small phase to large phase), while the right panel shows respective leakage current of various films; the solid line is a guide for the eye. It should be noted that nearly the same leakage current for films-A, B, G, and H might not enable us to clearly distinguish their data points.

Figure 3

The electric field $\left(E\right)$ versus peak terahertz-amplitude $\left(E_{\mathrm{THz}}\right)$ hysteresis loops (a) for film-A, film-B, and film-H, and (b) for film-D and film-E. The sample was first poled in an electric field of $-200\mathrm{kV}∕\mathrm{cm}$, and then the data were collected in the electric-field sequence of $0\mathrm{kV}∕\mathrm{cm}\to 200\mathrm{kV}∕\mathrm{cm}\to 0\mathrm{kV}∕\mathrm{cm}\to -200\mathrm{kV}∕\mathrm{cm}\to 0\mathrm{kV}∕\mathrm{cm}$.

Figure 4

The electric field $\left(E\right)$ versus peak terahertz-amplitude $\left(E_{\mathrm{THz}}\right)$ hysteresis loops for film-H and film-B. A single dashed line in the hysteresis loop of film-H indicate one slope while transformation from $-E_{\mathrm{THz}}$ to $+E_{\mathrm{THz}}$, while two dashed lines in the hysteresis loop of film-B indicate that $-E_{\mathrm{THz}}$ transforms to $+E_{\mathrm{THz}}$ in two steps. The arrows indicate the change of slope.