量子阱
异质结
偶极子
凝聚态物理
密度泛函理论
极化(电化学)
带偏移量
材料科学
宽禁带半导体
电子能带结构
电场
分子物理学
化学
物理
光学
带隙
计算化学
激光器
量子力学
价带
物理化学
作者
Paweł Strąk,Paweł Kempisty,Maria Ptasinska,Stanisław Krukowski
摘要
A critical comparison of three polarization based approaches with the fields in AlN/GaN multiple quantum wells (MQWs) systems proved that they give identical results. The direct density functional theory (DFT) results, i.e., the fields, are in qualitative agreement with data obtained within the polarization theory. The results of DFT calculations of an AlN/GaN MQW system were used in the projection method to obtain a spatial distribution of the bands in the structure with atomic resolution. In parallel, the plane averaged and c-smoothed potential profiles obtained from the solution of the Poisson equation were used to determine the electric field in the multiquantum well structures and the magnitude of dipole layers at the AlN/GaN heterostructures. The dipole layers cause potential jumps of about 2.4 V that seriously affects the band offsets. The presence of the dipole layer is in good agreement with the potential measurements by electron holography. It was shown that the wells of the width up to 4 Ga layers behave as potential minima, but the wider layers behave as standard quantum wells. The barriers up to 3 Al layers do not localize the carriers. It is shown that the Quantum Confined Stark Effect causes a huge decrease of their energies and oscillator strengths of the optical transitions, especially for wider structures. For wider wells, the strengths fall much faster for perpendicular polarization which indicates the important role of the anisotropic band offsets. A direct simulation shows that the band offset for the valence band crystal field split off hole states, i.e., pz states are different from heavy and light hole (i.e., p⊥=px⊗py) states being equal to valence band offset (VBO)⊥=0.85 eV and rough estimate of VBOII≅0.5 eV, respectively. These values are in good agreement with the recently reported measurement of AlN/GaN offsets.
科研通智能强力驱动
Strongly Powered by AbleSci AI