ABSTRACT
The physical phenomena governing hyperpolarization through optical pumping of conduction electrons continue to be explored in multiple semiconductor systems. One early finding has been the asymmetry between the optically pumped nuclear magnetic resonance (OPNMR) signals when generated by different circular polarizations (i.e., light helicities). Because these resonances are asymmetric, the midpoint between the signals prepared with each of the two circular polarizations is either a positive or negative value, termed an "offset" that is representative of an optical Overhauser enhancement. Both negative offsets (in GaAs) and positive offsets (in CdTe) have been observed. The origins of these offsets in semiconductors are believed to arise from thermalized electrons; however, to the best of our knowledge, no study has systematically tested this hypothesis. To that end, we have adopted two configurations for OPNMR experiments-one in which the Poynting vector of the laser light and magnetic field are parallel, and one in which they are antiparallel, while other experimental conditions are kept the same. We find that the OPNMR signal response to a fixed helicity of light depends on the experimental configuration, and this configuration needs to be accounted for in order to properly describe the OPNMR results. Further, studying the offsets as a function of field strength shows that the optical Overhauser enhancement (the offset) increases in magnitude with field strength. Finally, by describing all angular momentum and phasing conventions unambiguously, we are able to determine that the absorptively-phased appearance of 113Cd (and 125Te) OPNMR in CdTe is a consequence of the sign of the nuclear gyromagnetic ratios for these isotopes.
ABSTRACT
Instrumentation for optically-pumped and optically-detected nuclear magnetic resonance (OPNMR and ODNMR) has been developed and implemented as a single experimental apparatus to study semiconductors such as GaAs and CdTe. These two measurement schemes use many of the same components for experiments. Here we describe, in two parts, the apparatus that can record such measurements and give examples of representative data. In Part 1, the radio-frequency probe and low-temperature cryostat are described, including single-channel and two-channel static cryogenic probes that both incorporate a modified solenoid coil that permits better optical access. In Part 2, the optical bench is described in detail, which uses a set of experiments (magneto-photoluminescence, photoluminescence excitation, detection of polarized photoluminescence) as important input for ODNMR. We are able to portray a robust design that encompasses multiple measurement modalities, along with the ability to change many experimental parameters with ease.
ABSTRACT
We have modified the model for optically-pumped NMR (OPNMR) to incorporate a revised expression for the expectation value of the z-projection of the electron spin, ãSzã and apply this model to both bulk GaAs and a new material, InP. This expression includes the photon energy dependence of the electron polarization when optically pumping direct-gap semiconductors in excess of the bandgap energy, Eg. Rather than using a fixed value arising from coefficients (the matrix elements) for the optical transitions at the k=0 bandedge, we define a new parameter, Sopt(Eph). Incorporating this revised element into the expression for ãSzã, we have simulated the photon energy dependence of the OPNMR signals from bulk semi-insulating GaAs and semi-insulating InP. In earlier work, we matched calculations of electron spin polarization (alone) to features in a plot of OPNMR signal intensity versus photon energy for optical pumping (Ramaswamy et al., 2010). By incorporating an electron spin polarization which varies with pump wavelength into the penetration depth model of OPNMR signal, we are able to model features in both III-V semiconductors. The agreement between the OPNMR data and the corresponding model demonstrates that fluctuations in the OPNMR intensity have particular sensitivity to light hole-to-conduction band transitions in bulk systems. We provide detailed plots of the theoretical predictions for optical pumping transition probabilities with circularly-polarized light for both helicities of light, broken down into illustrative plots of optical magnetoabsorption and spin polarization, shown separately for heavy-hole and light-hole transitions. These plots serve as an effective roadmap of transitions, which are helpful to other researchers investigating optical pumping effects.
ABSTRACT
Optically-pumped (69)Ga NMR (OPNMR) and optically-detected measurements of polarized photoluminescence (Hanle curves) show a characteristic feature at the light hole-to-conduction band transition in a GaAs/AlxGa1-xAs multiple quantum well sample. OPNMR data are often depicted as a "profile" of the OPNMR integrated signal intensity plotted versus optical pumping photon energy. What is notable is the inversion of the sign of the measured (69)Ga OPNMR signals when optically pumping this light hole-to-conduction band energy in OPNMR profiles at multiple external magnetic fields (B0=4.7T and 3T) for both σ(+) and σ(-) irradiation. Measurements of Hanle curves at B0=0.5T of the same sample exhibit similar phase inversion behavior of the Hanle curves at the photon energy for light hole excitation. The zero-field value of the light-hole state in the quantum well can be predicted for the quantum well structure using the positions of each of these signal-inversion features, and the spin splitting term in the equation for the transition energy yields consistent values at 3 magnetic fields for the excitonic g-factor (g(ex)). This study demonstrates the application of OPNMR and optical measurements of the photoluminescence to detect the light hole transition in semiconductors.