Straightforward fitting of the raw terahertz conductivity spectra by the Drude-Smith model, which was abundantly
used in the literature, did not lead to a significant advance in an in-depth understanding of these phenomena.
This is mainly because of the depolarization fields which build up in any inhomogeneous system. On the one hand,
these fields reflect the sample morphology and their understanding in each particular system may provide new information about the nanostructure connectivity. On the other hand, the effect of unknown depolarization fields can hide or distort fingerprints of the nanoscopic transport.
Recently, we devoted a systematic effort to describe the depolarization fields in photoconductive nanostructures
and to disentangle their effect from that of the local carrier response function .
We proposed a general relation between the effective photoconductivity (measured in the THz experiment) and the microscopic response function. It is based on the general Bergman representation of effective medium theory for a two-component system with a single dominant depolarization factor. We have shown that our model describes both percolated and non-percolated samples and structures with complex percolation pathways .
We solved the wave equation in an inhomogeneous sample and demonstrated that the transient sheet photoconductivity is directly measured in THz experiments for any spatial profile of the photoconductivity and any thin film sample morphology [1,3].
The microscopic response function can be determined by Monte-Carlo calculations of the motion of a single carrier inside nanoparticles .
Analysis of experimental data within this framework allows us to uncover the nature of charge carrier
transport at nanoscale and to assess the sample morphology in quite arbitrary nanostructured systems.
Based on this approach we proposed microscopic models of the response of partly or completely localized
charge carriers in a number of systems including various TiO2 nano- and microparticle networks [2,5],
in Sb-doped SnO2 nanoparticles [6,7], and in various nanocrystalline silicon systems [3,8],
For example, in  we investigated a regular array of vertically aligned InP nanowires (Fig. 1).
In this structure, photonic waveguiding effects inside nanowires and interferences of the excitation
optical beam should have been properly taken into account to describe the sample photoexcitation.
We were able to determine high transversal electron mobility in nanowires which, in comparison
with the longitudinal mobility, indicates the presence of stacking faults in the growth direction (Figure 1).
(a) Sketch of the preparation of the vertically aligned InP nanowires and scheme of the investigated structure.
(b) Distribution of the excitation density in the nanowires for two different excitation wavelength λexc.
(c) The transversal mobility (μxx, μyy) is considerably higher than the longitudinal
one (μzz) which is limited by the stacking faults along the growth direction.
 P. Kuzel and H. Nemec, J. Phys. D: Appl. Phys. 47, 374005 (2014).
 H. Nemec et al., IEEE Trans. Terahertz Sci. Technol. 3, 302 (2013).
 V. Zajac et al., New J. Phys. 16, 093013 (2014).
 H. Nemec et al., Phys. Rev. B 79, 115309 (2009).
 S. A. Jensen et al., J. Phys. Chem. C 118, 1191 (2014).
 K. Peters et al., Chem. Mater. 27, 1090 (2015).
 K. V. Skoromets et al., J. Phys. Chem. C 119, 19485 (2015).
 H. H. Němec et al., Phys. Rev. B 91, 195443 (2015).
 C. S. Ponseca et al., Phys. Rev. B 90, 085405 (2014).