熱載子式矽與砷化鎵次能隙紅外光偵測器與非破壞式磊晶薄膜成分研究

Abstract

本論文主要針對紅外光寬波段光偵測器進行探討，透過本實驗室設計的深溝槽孔洞結構與蕭特基金半接面，本研究的光偵測器將光轉換成熱載子後輸出電流及電壓訊號，其工作波段可以涵蓋近紅外光與中紅外光波段，並且搭配降溫量測系統進行低溫量測，可以在上述波段得到良好的光響應特性。 本研究將光偵測器分為兩種金屬與基板材料進行探討，分別為氮化鈦搭配P型矽基板與金搭配半絕緣砷化鎵基板，其中第一部分會探討氮化鈦與矽基板之元件，由於氮化鈦薄膜在不同製程條件下會具有不同的光學行為，因此我們將透過製程中的氮氣流量參數調控，製備出具有不同光學性質的氮化鈦薄膜，並透過橢圓儀及光譜儀等分析取得其光學常數後，藉由模擬軟體建立模型並最佳化元件結構的孔洞大小、週期與深度等參數，最後透過電性量測系統與降溫系統的組合，輔以鎢燈與碳化矽燈等光源進行量測分析。由於不同氮氣流量所製作出的氮化鈦薄膜具有不同的結晶型態、光學與電特性，在製程條件調整的過程中我們發現當氮氣流量增加時反射值會先大幅提升再稍微下降，而穿透值則無明顯差異，透過電性分析我們發現氮氣流量增加後金半接面的蕭特基特性會逐漸變好，最後我們選擇氮氣流量15 sccm的氮化鈦薄膜，分別在光學與電性方面皆具有最優異的表現。利用此流量之氮化鈦薄膜搭配深溝槽孔洞結構所製備出的光偵測元件，在室溫下分別於2 μm、3.25 μm、6 μm與10.6 μm四個波段我們量測到1.27×10-4 A/W、7.35×10-5 A/W、1.06×10-4 A/W、9.92×10-5 A/W的光電流響應度與0.07 V/W、0.04 V/W、0.06 V/W、0.06 V/W的光電壓響應度；而同樣針對四個波段並透過降溫系統於-100°C下，我們量測到10.08 V/W、5.60 V/W、8.69 V/W、8.12 V/W的光電壓響應度。透過降溫量測的分析結果，我們證實了隨著溫度下降蕭特基能障減少與元件電阻上升，皆會對光電流與光電壓響應特性造成影響。接下來，透過實驗室內的雷射、截波器與示波器組合，我們探討了上述元件在10.6 μm光源照射下的速度特性，藉由此系統來分析元件的開關特性，並透過電壓值的百分之十至百分之九十分別計算出上升時間與下降時間，來確認元件工作原理是否受環境熱影響。我們針對氮化鈦/P型矽元件於10.6 μm進行分析，同樣透過截波器頻率的調整在1 Hz與5 Hz的頻率下量測，我們於5 Hz下測得了0.00576秒的上升時間與0.00537秒的下降時間。由這些快速的上升與下降時間我們證實了元件並非受到熱的影響，而是中紅外光波段之入射光照射元件後，熱載子有足夠能量得以穿過蕭特基能障並被量測到。 論文中的第二部分我們採用金膜與半絕緣砷化鎵基板的搭配，製備了近紅外光波段工作的元件並分析環境溫度之影響與其速度特性，透過我們設計好的深溝槽孔洞結構，金/砷化鎵元件表現出高達922.83 V/W的優異光電壓響應度特性。在升降溫系統的量測中，我們發現該元件於-50°C有最佳的電性表現，並且當溫度升高至200°C的狀況下仍然可以工作。接著我們針對金/砷化鎵元件於1550 nm進行速度特性分析，透過截波器頻率的調整分別於500 Hz、1000 Hz、2000 Hz與3000 Hz四種頻率進行量測，並且我們在3000 Hz下測得了0.00018秒的上升時間與0.00020秒的下降時間。 論文中的第三部分將利用拉曼量測系統來分析MOCVD製程所製備的三五族InAlAs磊晶薄膜品質，透過製程溫度與製程氣體的調控準備了不同的InAlAs磊晶薄膜，分為溫度與氣體流量兩組進行探討。首先，藉由拉曼系統確認位於230 cm-1與360 cm-1波數附近的峰值，而這兩個峰值分別對應到InAs與AlAs的拉曼特徵峰，接著透過減光量測的方式，確認InAs與AlAs特徵峰值減弱的趨勢差異，並透過峰值比例計算發現在高溫與砷化氫氣體流量不足時，其製備出的薄膜會有As原子流失的狀況，並透過X射線光電子能譜的量測驗證了拉曼分析結果。藉由拉曼量測系統及減光技巧，可以透過非破壞性的檢測方式來分析薄膜成分分布性質，有助於提升薄膜成長領域的研究。

In this thesis, we investigated the broadband photodetector with deep trench thin metal (DTTM) structure in the infrared ray regimes ranging from 1550 nm to 10.6 μm. As we all known, the photodetectors convert the light from near infrared (NIR) to mid-infrared (MIR) regime into the hot carriers, which produce the photocurrent and photovoltage. In this research, we discussed two research topics. In the first research topic, we discussed the infrared (IR) photodetector which shows good optical and electrical behaviors with respect to the experimental setup temperature. In the second research topic, we determined the composition distribution of III-V InAlAs epitaxial thin films using non-destructive method with Raman spectroscopy and we verified with X-ray photoelectron spectroscopy (XPS). In the first research topic, we introduced two different kinds of design for IR photodetector which was fabricated using different metals and substrates. In the first design, we used sputtering process to deposit a titanium nitride (TiN) film on p-type silicon (p-Si) to form the Schottky junction. In this sputtering process, we had modified the nitrogen (N2) flow to deposit TiN on p-Si substrate and we repeated the same sputtering processes with different N2 flows which shows different optical and electrical behaviors at different wavelengths. Here we used the ellipsometer and spectrometer to measure the refractive index (n) and extinction coefficient (k) values. By using the three dimensional finite-difference time-domain (3D-FDTD) simulation, we optimized the hole size, array period and depth of DTTM structure. From the current-voltage (I-V) curve, we found that when the N2 flow increase in the sputtering process that will also increase the contact quality in the TiN/p-Si Schottky junction device. We demonstrated the experimental setup at different temperature. First, we performed experimental setup at ambient temperature of 25°C with different wavelength of 2 μm, 3.25 μm, 6 μm and 10.6 μm. We also obtained the photovoltage responsivity of 0.07 V/W, 0.04 V/W, 0.06 V/W and 0.06 V/W, respectively. Under the illumination of CO2 laser at 10.6 μm, we showed the rise time at 0.00576 seconds and fall time at 0.00537 seconds of TiN/p-Si Schottky junction device with respect to the chopper frequency of 5 Hz. We also demonstrated the experimental setup at lower temperature at -100°C by using liquid nitrogen and we calculated the photovoltage responsivity values of 10.08 V/W, 5.60 V/W, 8.69 V/W and 8.12 V/W with respect to different wavelength as we mentioned before. Finally, we used the experimental setup with both ambient at low temperature for TiN/p-Si Schottky junction device with the N2 flow rate of 15 sccm performed the excellent performance in both optical and electrical behaviors. In the second design, we used the same sputtering process to deposit the gold (Au) on intrinsic gallium arsenide (GaAs) substrate to form the Scohttky junction. We made the experimental setup for Au/GaAs Scohttky junction device with three different working temperatures. The experimental setup with lower temperature of -50°C shows the best performance with respect to the wavelength of 1550 nm. We presented that Au/GaAs Scohttky junction device can even work at very high temperature up to 200°C. In experimental setup with ambient temperature of 25°C, we obtained the photovoltage responsivity value of 922.83 V/W. We also presented the rise time at 0.00018 seconds and fall time at 0.00020 seconds with respect to the chopper frequency of 3000 Hz. In the second research topic, The III-V InAlAs epitaxial thin film which was deposited by Metal Organic Chemical Vapor Deposition (MOCVD) process we employed. The Raman signal for the III-V InAlAs epitaxial thin film was measured to analyze the Raman peak intensity and Raman peak position to distinguish the composition distribution of III-V InAlAs epitaxial thin film. We found the Raman peak position around 230 cm-1 and 360 cm-1, which correspond to the peak position of InAs and AlAs respectively. The ratio of InAs/AlAs were calculated from Raman peak intensity. We used the light intensity reduction method to find the difference in the Raman peak intensity ratio of InAs/AlAs. The experimental trend shows the decreasing the laser intensity will also decrease the Raman peak intensity ratio of InAs/AlAs. In the MOCVD process, the process temperature and arsine gas flow are the two different parameters which affect the quality of epitaxial thin film. Epitaxial thin film which used here were deposited by MOCVD under high process temperature and low arsine gas flow that makes an observable difference in the Raman peak intensity ratio compared to epitaxial thin film deposited by MOCVD under normal temperature and arsine gas flow. This indicates that the leakage of As atoms during the MOCVD process. The intensity ratio of Raman peak and the composition distribution properties were also verified by XPS.