|M.Sc Student||Ezra Uri|
|Subject||Intersubband Electro and Photo-induced Absorption in|
InGaAs/InP Quantum Wells
|Department||Department of Physics||Supervisors||Professor David Gershoni|
|Professor Emeritus Eitan Ehrenfreund|
We report on investigations of In0.53Ga0.47As/InP supperlattices using intersubband optical spectroscopy. We measured intersubband absorption, photo- and electro-induced absorption, due to optical transitions between the conduction minibands. The samples consist of 60Å InGaAs well and InP barrier of various widths in each period. The barriers, which separate between adjacent wells determine the overlap between the wave functions of carriers within each quantum well, and thereby the energy width of the superlattice's minibands. The narrower the barriers are, the wider the minibands are. The energy width of a miniband () is related to the lifetime of a carrier within this miniband (τ) by τ = /. The tunneling probability of an electron through the superlattice structure is calculated as a function of its energy. This probability is, normally very small, but when the incident electron energy coincides with a miniband energy, there is a resonant amplification in the tunneling probability. The lifetime is then evaluated from the full energy width at half the maximum of the resonant tunneling peak. In our samples the first electronic level (e1) is quite deep, while the second level (e2) is quite shallow. Consequently, the width of the first superlattice miniband is rather narrow, while that of the second one is quite wide. Since intersubband transitions are vertical in momentum space, their energy is extremely sensitive to the momentum component along the superlattice axis of the e1 electron. When external electric field is applied to a superlattice, its energy bands ‘bend’ and change slightly. In addition there is a flow of carriers through the structure and their steady state population undergoes redistribution. Under these conditions τ represents the dwell time that a carrier spends in its miniband.
The density of electrons in the first miniband (n1) can be measured by the intersubband absorption (e1-e2). Using electro-or photo-induced absorption the induced changes in this density can be extracted. By correlating the measured density with the measured current, we tried to get an estimate for the lifetime τ. In the most straightforward case, the dwell time is deduced from the current density () by: τ=n1/. We measured electron density of ~ 1010 cm-2 per period due to unintentional doping. However, two orders of magnitude smaller variations were measured in the electro- induced absorption. This means that only a small fraction of the resident electron population is influenced by the applied bias.
We found that in the differential electro induced intersubband absorption spectra there are both positive and negative spectral regions. This means that under bias, electrons of different momentum redistribute differently.
These differences are qualitatively explained in terms of “effective mass filtering”. From the second derivative of the calculated e1 dispersion we get the electron effective mass and thereby its mobility. In the Brillouin zone center (kz=0) the electrons have positive relatively small in absolute value, effective mass. In the zone edge (kz= /lz , where lz is the superlattice period) they have negative relatively small in absolute value effective mass. In between, the effective mass changes sign and its absolute value is rather large. Therefore, zone edge and zone center electrons are rather mobile, but they move in opposite directions when subjected to an applied electric field. In contrary, mid zone electrons are very heavy and immobile.
We outline a more sophisticated model, which takes into account the Poisson and the continuity equations in order to quantitatively account for our experimental observations.