Figure 1

Topography (a), photocurrent (b), and dark-current (c) maps measured on an $x=0.33$ CIGSe film with 50 mV bias voltage. Scan size: $4\times 4\mu {\mathrm{m}}^2$. (d) The photocurrent histograms at zero bias (red) and at 50 mV (blue), showing that the reversed photocurrent vanishes (on average) at $\sim 50\mathrm{mV}$.Reuse & Permissions

Figure 2

(a) Average $V_c$ as a function of the Ga content $x$. (b) The device efficiency of the reference solar cells. The error bars reflect the spread in the measurements, and the lines are guides to the eye.Reuse & Permissions

Figure 3

$0.5\times 0.5\mu {\mathrm{m}}^2$ STM topographic images measured on CIGSe films with $x=0$ (a), $x=0.33$ (c), and $x=0.75$ (e) and corresponding tunneling spectra (b), (d), and (f) measured on grain centers (red curves, “G”) and grain boundaries (blue curves, “GB”) at positions indicated in the respective images. The spectra presented in (d) exhibit a pronounced spectral shift, manifesting downwards band bending at GBs to a degree that yields type inversion, from $p$-type (on grain) to $n$-type (on the GB). The above spectra manifest also reduced DOS at around the band center at the GB with respect to the adjacent grain. In contrast, neither a spectral shift nor a reduction in the GB DOS at the chemical potential is observed for $x=0.75$ (f). Note that the plots have different scales.Reuse & Permissions

Figure 4

(a) $0.5\times 0.5\mu {\mathrm{m}}^2$ STM topographic image measured on a CIGSe film with $x=0.33$. (b) Two tunneling spectra, one measured on the grain center (red curve, G) and the other on the grain boundary (blue curve, GB), at positions indicated in the topographic image. A reduced DOS around the chemical potential is observed at the GB, but no clear signature for band shifts (band bending) is detected on this GB.Reuse & Permissions