原子层沉积
等离子体增强化学气相沉积
材料科学
电容
电介质
电容器
光电子学
化学气相沉积
高-κ电介质
氮化硅
氮化物
分析化学(期刊)
薄膜
硅
纳米技术
图层(电子)
电气工程
电压
化学
电极
物理化学
工程类
色谱法
作者
Jiro Yota,Hong Shen,Ravi Ramanathan
出处
期刊:Journal of vacuum science & technology
[American Vacuum Society]
日期:2012-12-06
卷期号:31 (1)
被引量:99
摘要
Characterization was performed on the application of atomic layer deposition (ALD) of hafnium dioxide (HfO2) and aluminum oxide (Al2O3), and plasma-enhanced chemical vapor deposition (PECVD) of silicon nitride (Si3N4) as metal–insulator–metal (MIM) capacitor dielectric for GaAs heterojunction bipolar transistor (HBT) technology. The results show that the MIM capacitor with 62 nm of ALD HfO2 resulted in the highest capacitance density (2.67 fF/μm2), followed by capacitor with 59 nm of ALD Al2O3 (1.55 fF/μm2) and 63 nm of PECVD Si3N4 (0.92 fF/μm2). The breakdown voltage of the PECVD Si3N4 was measured to be 73 V, as compared to 34 V for ALD HfO2 and 41 V for Al2O3. The capacitor with Si3N4 dielectric was observed to have lower leakage current than both with Al2O3 and HfO2. As the temperature was increased from 25 to 150 °C, the breakdown voltage decreased and the leakage current increased for all three films, while the capacitance increased for the Al2O3 and HfO2. Additionally, the capacitance of the ALD Al2O3 and HfO2 films was observed to change, when the applied voltage was varied from −5 to +5 V, while no significant change was observed on the capacitance of the PECVD Si3N4. Furhermore, no significant change in capacitance was seen for these silicon nitride, aluminum oxide, and hafnium dioxide films, as the frequency was increased from 1 kHz to 1 MHz. These results show that the ALD films of Al2O3 and HfO2 have good electrical characteristics and can be used to fabricate high density capacitor. As a result, these ALD Al2O3 and HfO2 films, in addition to PECVD Si3N4, are suitable as MIM capacitor dielectric for GaAs HBT technology, depending on the specific electrical characteristics requirements and application of the GaAs devices.
科研通智能强力驱动
Strongly Powered by AbleSci AI