二維氮化鎵電子系統與氮化銦之磁阻研究

Abstract

It is well known that III-nitride semiconductors have many applications in optoelectronic devices. We hope that a possible integration of magnetic storage or magnetic sensors with optoelectronic devices can be realized on the same material system. Therefore we studied the electrical properties of an InN film in the presence of magnetic fields through magneto-resistivity. In addition, a high electron mobility transistor (HEMT) structure not only has many applications such as cellphones, radar, satellite communication, but also is a suitable system to study low-dimensional physics such as the magnetic-field-induced transition from an insulator (I) to a quantum Hall (QH) state observed in a two-dimensional electron system (2DES). A comprehensive understanding of physics of HEMTs is helpful to improve the performance of HEMT based devices. Thus we studied the I-QH transition in low magnetic fields in an AlGaN/GaN HEMT structure. At first, we characterized the structure of the InN film by the high-resolution X-ray diffraction (XRD) analysis (Bede D1 system) and Transmission electron microscopy (TEM) (JEOL JEM-2000FX). The XRD q/2q-scan pattern of our InN film shows three obviously sharp peaks corresponding to Al2O3(0006), GaN(0002), and InN(0002), respectively. No obvious off-axis crystal plane and metallic In(101) peak can be found in the pattern, suggesting that our InN film is of high purity and has an epitaxial relation of InN(0002)//GaN(0002)//Al2O3(0006). The TEM image of the InN film indicates that both the InN and the GaN layer were grown continuously from the interface without any voids and that most dislocations in the InN film originate in the GaN buffer layer and gradually disappear toward to the surface. No obvious In droplet is found in the image, consistent with the XRD results. The corresponding selected-area electron diffraction (SAED) pattern taken at the interface of InN and GaN only shows clear InN and GaN diffraction spots in the pattern, suggesting the epitaxial relationship between InN(0001) and GaN(0001) and the thin film with high purity and single wurtzite phase even at the interface. Although the XRD and TEM analyses have no clue to the appearance of metallic indium droplets or a thin indium layer in the InN film, minute metallic indium in the sample is still possible.

In the study of the InN film, we observed huge positive magnetoresistance ( %) at low temperatures as a very low magnetic field is applied (B~0.15 T). The huge PMR shows a strong dependence on temperature and dies out as the temperature is increased above 4 K. From the temperature and magnetic-field dependence of the resistance, we inferred that the huge PMR may be related to superconductivity. However, we can not make sure whether superconductivity is an intrinsic property of our InN film because of the possible existence of minute metallic indium (type I superconductivity). Magnetoresistivity measurements performed on the AlGaN/GaN heterostructure at various temperatures ranging between 0.27 K and 2.2 K. An approximately T -independent point is observed in for which ~ . We can see that increases with decreasing T at B ＜ while it decreases with decreasing T at B ＞ , consistent with a direct I-QH transition point at low B . When the perpendicular magnetic field is larger than , the SdH oscillations begin to appear, inconsistent with Huckestein’s argument that should coincide with . In addition, his argument also implies that the quantum mobility must be approximately equal to electron mobility μ, where quantum mobility , is the quantum lifetime corresponding to Landau level broadening, and is the effective mass. We can inspect this implication by comparison among various mobilities. The magnetoresistivity data can be used to obtain different mobilities from the following ways: 1. m0 = 1/(ner0) from the zero-field resistivity r0. 2. mR = rxy/(rxxB) at B =Bcr [4]. 3. mcr = 1/Bcr from the approximately T-independent point in the I-QH transition. 4. me = 1/B, where B is the magnetic field at which ρxx ~ρxy. 5. The renormalized mobility m΄ is obtained by fitting . 6. The quantum mobility mq is determined from SdH oscillations. We found that the quantum mobility is much lower than other mobilities, inconsistent with Huckestein’s arguments. The same results are found in other two samples. For the whole temperature range (0.27 K T 80 K), the temperature dependence of the renormalized mobility m΄ obtained from the information in low magnetic fields is approximately the same as that of m0 determined from zero-field information. The result implies that the strength of the weak localization effect is smaller than that of Electron-electron-interaction (EEI) effects at zero field. By the huge PMR observed in our non-magnetic InN film, magnetic sensing and recording devices, which are not susceptible to ferromagnetic noise, can be achieved. Furthermore, an advantage of InN is that InN can be grown on Si which is fully compatible with the silicon CMOS technology. Most importantly, it is deserved to mention that III-nitride semiconductors are promising material for optoelectronic devices such as LED and solar cells, therefore the experimental results inspire the integration of magnetic sensing and recording devices with optoelectronic devices using the same nitride material system. For the AlGaN/GaN heterostructure, we conclude that the corrections to Huckestein’s arguments are necessary to understand the direct I-QH transitions because and . In addition, the coincidence of m0 and m΄ implies that the zero-field mobility may be dominated by EEI in our study. The results together with the observed logarithmic T dependence of the Hall slope and the linear T dependence of the renormalized mobility, the characteristic of EEI effects, are the experimental evidence of EEI effects in our AlGaN/GaN HEMT structure.