Ultrafast hot carrier dynamics of ${\mathrm{ZrTe}}_5$ from time-resolved optical reflectivity

${\mathrm{ZrTe}}_5$ is a layered pentatelluride with very weak interlayer coupling, high in-plane carrier mobility, and strong electron-electron interaction. We investigate the hot carrier dynamics of ${\mathrm{ZrTe}}_5$ by ultrafast time-resolved optical reflectivity. Our results reveal a phonon-mediated across-gap recombination and an electron-phonon scattering channel in thermalization, with timescales of few and sub-ps, respectively. In addition we find a fast relaxation right after the initial photoexcitation with a time constant of $\sim 250$ fs, possibly arising from an electron-electron scattering channel. Our findings indicate that correlated many-body effects play an important role for the transient dynamics of ${\mathrm{ZrTe}}_5$.

Figure 1

Time-resolved reflectivity changes of ${\mathrm{ZrTe}}_5$. (a) Photoexcitation and relaxation processes. The initial photoexcitation time is marked as $t_1$. The BCB buildup time is denoted as $t_2$. Hot electrons from high-lying conduction band ($t_1$) lose energy and accumulate to the bottom of the low conduction band ($t_2$) through e-e scattering and e-ph scattering. The dashed line illustrates the impact excitation for e-e scattering, in which a photoexcited electron collides with an electron in the valence band. (b) Schematic of ultrafast time-resolved optical reflectivity. (c) The a-c plane of the ${\mathrm{ZrTe}}_5$ crystal sheet, formed by ${\mathrm{ZrTe}}_3$ chains along the a-axis and linked via zig-zag chains along the c-axis. The bulk crystal is constructed by stacking the ${\mathrm{ZrTe}}_5$ sheets along the b-axis (perpendicular to a-c plane) with very weak interlayer coupling. The blue and orange spheres represent Te atoms, while Zr atoms are marked as the green ones. (d) Representative reflectivity changes at 6, 135 and 200 K, using pump pulses of 3.10 and 1.55 eV with $F=18\mu \mathrm{J}/{\mathrm{cm}}^2$. The zero time delay is determined by the pink shaded curve, which is the cross-correlation of the pump and probe beams, with FWHM of 120 fs. The black arrow denotes the maximum of $|\mathrm{\Delta }R/R|$, corresponding to the BCB buildup at $t_2$.

Figure 2

Across-gap relaxation of the BCB buildup after $t_2$. (a) Comprehensive $\mathrm{\Delta }R/R$ results with pump fluence $F=18 \mu \mathrm{J}/{\mathrm{cm}}^2$ by 3.10 and 1.55 eV. (b) Relaxation time ${\tau }_r$ as a function of temperature, extracted from (a). The error bars are obtained from fits. The pink circles in (b) are the inverse of the $T$-dependent gap, confirming ${\tau }_r\propto 1/E_g$ (${\mathrm{E}}_g$ taken from Ref. [3] with authors' permission). (c) Normalized amplitude $A_r$ as a function of temperature. The gray thick curves in (b) and (c) are used to guide the eye for the slope changes at 105 and 165 K.

Figure 3

BCB buildup dynamics between $t_1$ and $t_2$. (a) Reflectivity changes at 6, 135, and 300 K between $t_1$ and $t_2$. The vertical bar presents FWHM of the pump-probe cross correlation. The red curve is a two-component exponential fit convolved with the width of the cross correlation. The red and orange dashed lines are single exponential results for decay channels via e-e (fast) scattering and e-ph (slow) scattering. The black dashed line denotes the kink position. The black arrows denote $t_1, t_{\mathrm{Kink}}$, and $t_2$. The black triangles mark the BCB buildup, the maximum of $|\mathrm{\Delta }R/R|$ at $t_2$. (b) and (c) ${\tau }_{\mathrm{slow}}$ and ${\tau }_{\mathrm{fast}}$ as a function of temperature, excited by 3.10 and 1.55 eV, respectively. The gray lines are used to guide the increase of ${\tau }_{\mathrm{slow}}$ with temperature increasing.

Figure 4

$F$-dependent reflectivity changes. (a) and (b) Representative $F$-dependent differential reflectivity at 6 K excited by 3.10 and 1.55 eV. The rising edge between $t_1$ and $t_2$ is fitted by a two-component exponential function in order to extract ${\tau }_{\mathrm{fast}}$ and ${\tau }_{\mathrm{slow}}$, while the relaxation ${\tau }_r$ is obtained by a single exponential fit after $t_2$. The insets are normalized curves, showing good agreement with each other. (c) and (d) $A_r, {\tau }_r, {\tau }_{\mathrm{fast}}$, and ${\tau }_{\mathrm{slow}}$ as a function of pump fluence. Error bars are obtained from fits.