Band structure dynamics in indium wires

One-dimensional indium wires grown on Si(111) substrates, which are metallic at high temperatures, become insulating below $\sim 100$ K due to the formation of a charge density wave (CDW). The physics of this transition is not conventional and involves a multiband Peierls instability with strong interband coupling. This CDW ground state is readily destroyed with femtosecond laser pulses resulting in a light-induced insulator-to-metal phase transition. The current understanding of this transition remains incomplete, requiring measurements of the transient electronic structure to complement previous investigations of the lattice dynamics. Time- and angle-resolved photoemission spectroscopy with extreme ultraviolet radiation is applied to this end. We find that the transition from the insulating to the metallic band structure occurs within $\sim 660$ fs, which is a fraction of the amplitude mode period. The long lifetime of the transient state ($>100$ ps) is attributed to trapping in a metastable state in accordance with previous work.

Figure 1

Metal-to-insulator transition in one-dimensional indium wires on Si(111): Sketch of the atomic structure above (a) and below (b) $T_{\text{C}}$. Sketch of the electronic structure above (c) and below (d) $T_{\text{C}}$ based on density functional theory calculations from [23]. (e) Sketch of the potential energy surface below (black) and above (orange) $T_{\text{C}}$. $Q$ is the generalized reaction coordinate of the ($4\times 1$) to ($8\times 2$) transition as defined in [18]. The vertical blue arrow indicates the light-induced phase transition. (f) Difference between panels (c) and (d) simulating the pump-probe signal for the light-induced phase transition.

Figure 2

Light-induced insulator-to-metal transition viewed by tr-ARPES at $T=40$ K. (a) Equilibrium photocurrent. (b) Pump-induced changes of the photocurrent at $t=0.1$ ps, $t=1.1$ ps, and $t=3.0$ ps for a pump photon energy of 1.0 eV and a pump fluence of 5.6 $\mathrm{mJ}/{\mathrm{cm}}^2$. The black boxes (centered at $\pm 0.15$ and $+0.35$ eV) indicate the area of integration for the pump-probe traces in Fig. 3.

Figure 3

Energy dependence of the pump-probe signal at $T=40$ K. (a) Pump-probe traces (thin lines) obtained by integrating the pump-induced changes of the photocurrent over the areas indicated by black boxes in Fig. 2. Thick lines are exponential fits. (b) Rise and (c) decay time as a function of energy. The values for the room temperature (RT) measurement from Fig. 4 are included in (b) and (c) for direct comparison. Continuous lines in (b) and (c) are guides to the eye.

Figure 4

Dynamics at room temperature. (a) Equilibrium photocurrent. (b) Simulated pump-probe signal assuming a hot electronic distribution for the transient state. (c) Pump-induced changes of the photocurrent at $t=0.1$ ps and $t=1.1$ ps for a pump photon energy of 1.55 eV and a pump fluence of 2 ${\mathrm{mJ}/\mathrm{cm}}^2$. The black boxes indicate the area of integration for the pump-probe traces in panel (d). Thick lines in (d) are exponential fits. The pump-probe trace at 0.1 eV shows coherent oscillations at $1.8\pm 0.1$ THz. (e) Electronic temperature as a function of pump-probe delay together with exponential fit.