Electron and hole $g$ factors in InAs/InAlGaAs self-assembled quantum dots emitting at telecom wavelengths

We extend the range of quantum dot (QD) emission energies where electron and hole $g$ factors have been measured to the practically important telecom range. The spin dynamics in ${\text{InAs/In}}_{0.53}{\mathrm{Al}}_{0.24}{\mathrm{Ga}}_{0.23}\mathrm{As}$ self-assembled QDs with emission wavelengths at about 1.6 $\mu \mathrm{m}$ grown on InP substrate is investigated by pump-probe Faraday rotation spectroscopy in a magnetic field. Pronounced oscillations on two different frequencies, corresponding to the QD electron and hole spin precessions about the field, are observed from which the corresponding $g$ factors are determined. The electron $g$ factor of about $-1.9$ has the largest negative value so far measured for III-V QDs by optical methods. This value, as well as the $g$ factors reported for other III-V QDs, differ from those expected for bulk semiconductors at the same emission energies, and this difference increases significantly for decreasing energies.

Figure 1

(a) Atomic force microscopy image of an InAs QD layer deposited on an ${\mathrm{In}}_{0.53}{\mathrm{Al}}_{0.24}{\mathrm{Ga}}_{0.23}\mathrm{As}$ barrier. (b) Photoluminescence spectrum of the studied InAs/${\mathrm{In}}_{0.53}{\mathrm{Al}}_{0.24}{\mathrm{Ga}}_{0.23}\mathrm{As}$ QDs at $T=10$ K. (c) Dynamics of the QDs relative transmission measured with a laser energy of 0.79 eV and spectral width of 10 meV at $T=12$ K. The red line shows the exponential fit to the experimental data.

Figure 2

(a) Dynamics of the ellipticity signal at different magnetic fields. The red dotted lines show fits to the experimental data. The form of the fit is given by the sum of two damped oscillating functions. The curves are shifted vertically for clarity. The inset shows a FFT spectrum of the ellipticity signal at $B=1$ T. (b), (c) Magnetic field dependencies of the oscillation frequencies (b) and decay rates (c) for the fast (solid squares) and slow (open circles) oscillations in the spin dynamics. The laser photon energy is set to 0.79 eV.

Figure 3

(a) Dependence of the oscillation amplitudes on the square root of the pump power for the fast (solid squares) and slow (open circles) oscillations in the observed spin precessions. Broadband $\sim 30$ meV pump and probe beams without a pulse shaper are used. (b) Dependence of the $g$ factor moduli on the angle between the sample surface and the magnetic field for the fast (solid squares, left axis) and slow (open circles, right axis) oscillations. The zero angle corresponds to the Voigt geometry. (a), (b) The dashed lines are guides to the eye. $B=1$ T, the laser photon energy is 0.79 eV, and $T=7$ K.

Figure 4

Dependence of $g$ factor moduli (a) and oscillation amplitudes (b) on the laser energy for the electron (solid squares) and hole (open circles) spin precession signal. The solid line shows the PL spectrum (right axis). (a), (b) The dashed lines are guides to the eye. $B=1$ T, $T=7$ K.

Figure 5

Dependence of the electron $g$ factor on the QD emission energy. The solid and open symbols correspond to the transverse and longitudinal $g$ factors, respectively. The original data from the present work are shown together with the data from Refs. [3, 5, 12, 13, 26, 46]. The solid line shows the dependence for bulk semiconductors calculated according to the Roth-Lax-Zwerdling relation (1). The dotted and dashed lines show the dependencies calculated for spherical QDs and QWs using the model of Ref. [28]. The table shows the $g$ factor values determined experimentally and calculated using different approaches at $E=0.79$ eV.