MBE成長低溫InGaAs/GaAs超晶格p-i-n結構之電性與缺陷量測分析

Abstract

結合低溫砷化鎵和InGaAs/GaAs超晶格結構在光電應用上的優越性,並 承接本實驗室過去對低溫砷化鎵材料n-i-n及p-i-n結構的研究,我們以分 子束磊晶系統成長低溫InGaAs/GaAs超晶格p-i-n結構,進一步對低溫砷化 鎵材料做研究,並同時成長正常溫度下成長的InGaAs/GaAs超晶格pi-n結構 作為研究上的參考比較之用,經由完整的電性與缺陷量測分析,我們得致如 下初步的成果. 在對漏電流的研究上,我們發現該低溫樣品在反向偏壓 下的漏電流明顯較其他參考樣品大上一個數量級,該現象在以往對低溫砷 化鎵p-i-n結構的研究中亦有發現,此現象乃是因為低溫砷化鎵材料高缺陷 濃度所導致的,量測不同溫度下反向偏壓的漏電流,我們求得一活化能 為0.65eV.在低溫77K量測個樣品正向偏壓下的電流並量測低溫範圍下該低 溫樣品各個溫度正向偏壓下的電流,發現低溫樣品的起始電壓(turn-on voltage)明顯較其他參考樣品來的大,且越低溫下越明顯,此應該為低溫砷 化鎵材料的特性,但尚未找到合理的模型來解釋其物理理機制. DLTS在 量測中所得到的缺陷,我們亦經由導納頻譜量測中找到相對應的缺陷,在阿 瑞尼尼斯圖上確定為同一個缺陷.我們在低溫樣品主要發現三個缺陷,分別 命名為91H,91E1及91E2. 91H的活化能與捕捉截面積分別為0.71eV 與1.7x10^-11cm^2,該缺陷在其他參考樣品中也有發現,故推論此缺陷的形 成機制為InGaAs/GaAs晶格不匹配所導致,在已發現的缺陷中, Y. Ashizawa等人對n-InGaAs/n-GaAs的p-i-n結構所發現的H2為同一個缺陷. 91E2的活化能與捕捉截面積分別為0.71eV與1.5x10^-15cm^2,經由該缺陷 在阿瑞尼圖上的位置和quenching effect,判定和EL2應該為同一個缺陷. 91E1的活化能與捕捉截面積分別為0.46eV與1.0x10^-16cm^2,該缺陷只有 在低溫樣品中有出現,在其他參考樣品並中沒有出現,極有可能為低溫砷化 鎵超晶格結構的特性缺陷,他在阿瑞尼斯出現的位置相當接近一般低溫砷 化鎵材料中0.66eV特性缺陷出現的範圍,但在能階深度上有一定的差異,無 法肯定為同一個缺陷. 最後我們在對超晶格樣品電容對頻率的量測中 發現一些不尋常的現象.在高溫或高反向偏壓低頻下會出現負電容的情形, 概括而言,負電容乃是因為量測時暫態電流所導致.在低溫樣品中並未出現 負電容的情形,但電容在低頻時出現隨頻率上升的情形,求該段曲線的反曲 點,得到一活化能正好對應到91E1,故電容隨頻率上升的情形極有可能為91 E1所導致,且該缺陷決定該低溫砷化鎵超晶格結構暫態電流的產生. To combine the photo-electro properties of the low temperature(LT) GaAs andthe superlattice structure, we used the molecular beam epitaxy system to growthe LT InGaAs/GaAs superlattice p-i-n structure. For comparison, the samples of the smae struectures were grown at normal temperature. We continued the research of the LT GaAs n-i-n and p-i-n structures to the electrical and deep-level measurements on the LT InGaAs/GaAs superlattice sample. The results we obtained are stated below. In the current-voltage(I-V) measurement, we observed that the leakage current of the LT sample was more than one order of magnitude higher than thatof the reference samples. A smimilar result was observed in our research of theLT GaAs p-i-n structure. It was due to the high concentration of defects in LT GaAs. Besides, at low temperarure, the LT sample had much higher turn-on voltage than the reference samples. Yet we have not found the suitable modelto explain the mechanism. Three deep levels were observed in DLTS as well as admittance spectroscopy for the LT sample. We labeled them as 91H, 91E1, and 91E2. The activation energy and corss section of 91H are 0.71eV and 1.7x10^-11cm^2.Similar deep levels as 91H were also observed in the reference samples. Therefore, this deep level was not created by the LT growth and was speculated to be the result of the lattice mismatch between InGaAs and GaAs. The activation energy and cross section of 91E2 are 0.71eV and 1.5x10^-15cm^2. This deep level was speculated to be EL2 because of its position in the Arreheniusplot and quenching effect. The activation energy and cross section of 91E1 are 0.46eV and 1.0x10^-16cm^2. It was not observed in the reference samples, suggesting that it is the uniquedeep level in the LT InGaAs/GaAs superlattice layers. The position of this deep level in the Arrehenius plot is in the vicinity of the dominating trap, 0.66eV,commonly observed in LT GaAs. This deep level may play a similar dominatingrole as the 0.66eV trap does in LT GaAs. Finally we observed unusual negative capacitance under high reverse bias or high temperature in our referencer samples by admittance spectroscopy. From theliterate, the negative capacitance is caused by the transient current. In the LT sample, there was no negatice capacitance but the capacitance increased withfrequency in the low frequency range. An activation energy obtained by takingthe inflexion points of the capacitance-frequency(C-F) curves in the low frequency range was found to correspond to 91E1. So, this deep level, 91E1, had a direct influence on the transient current of the LT InGaAs/GaAs superlattice sample.