Flexophotovoltaic Effect and Above-Band-Gap Photovoltage Induced by Strain Gradients in Halide Perovskites

We have measured the flexophotovoltaic effect of single crystals of halide perovskites ${\mathrm{MAPbBr}}_3$ and ${\mathrm{MAPbI}}_3$, as well as the benchmark oxide perovskite ${\mathrm{SrTiO}}_3$. For halide perovskites, the flexophotovoltaic effect is found to be orders of magnitude larger than for ${\mathrm{SrTiO}}_3$, and indeed large enough to induce photovoltages bigger than the band gap. Moreover, we find that in ${\mathrm{MAPbI}}_3$ the flexophotovoltaic effect is additional to a native bulk photovoltaic response that is switchable and ferroelectric-like. The results suggest that strain gradient engineering can be a powerful tool to modify the photovoltaic output even in already well-established photovoltaic materials.

Figure 1

Schematic configuration of flexophotovoltaic experiment and results for STO and MAPB. (a) Single crystals with identical top and bottom electrodes are clamped in the thickness direction [001] and bent vertically (see Supplemental Material, Fig. S2 [17]). The lower left shows the relation between coordinate axis and crystal direction. The crystals were bent around the [010] axis and illuminated laterally along the bending axis. A polarizing filter was introduced after the measurements in order to verify the bulk nature of the photoresponse. All measurements were performed under 1000 LUX of illuminance. (b) and (f) Photocurrent density as a function of applied strain gradient. (c) and (g) Current as a function of voltage under bending and illumination. (d) and (h) Open-circuit photovoltage (voltage for which there is no current) and closed-circuit photocurrent (current at 0 V) as a function of strain gradient, as extracted from 1(c) and 1(g) (example procedure provided in Supplemental Material [17], S3). (e) and (i) Photovoltage and photocurrent as a function of the polarization of the incident light with respect to the [001] direction, which is perpendicular to the elecrodes and parallel to the flexoelectric polarization. The observed $\pi$-periodic response is consistent with a bulk photovoltaic effect.

Figure 2

Photovoltaic output of MAPI as a function of strain gradient (a),(b) and poling field (c),(d). The results show a native bulk photovoltaic effect that can be modified both by strain gradients and, hysteretically, by electric fields.

Figure 3

Above-band-gap photovoltage induced by strain gradients. (a) Scheme of the experiment, whereby a large local curvature is induced by pushing with the tip of an atomic force microscope, and the current vs voltage is measured under illumination for various indentation forces. (b) Microcrystal used for the measurement. The thickness of the crystal is $\sim 1\mathrm{\mu }\mathrm{m}$ and the bottom electrode is ITO. We observed no mechanical damage to the sample even after applying a maximal indentation force of $20\mathrm{\mu }\mathrm{N}$ (see comparison in Supplemental Material [17], Fig. S6-2. (c) Current vs voltage measured under an indentation force of $10\mathrm{\mu }\mathrm{N}$, showing a $\sim 4\mathrm{V}$ open-circuit photovoltage (d) open-circuit photovoltage as a function of indentation force (extracted from Fig. S7 [17]), showing the correlation between the two and (inset) photovoltage as a function of linear light-polarization angle, showing the same sinusoidal dependence as in the bulk crystals.