摻入氮與銻含量對砷化銦鎵 / 砷化鎵量子井光性影響之研究

Abstract

早期中長程光纖通訊元件光源材料大多採用InGaAsP ，但此材料對熱敏感度非常大，操作時必須有散熱裝置來維持穩定性而造成封裝製程成本增加。而新穎的InGaAsN與AlGaInAs 兩種系統被發現具有更多的優勢，吸引更多學術研究單位投注在這兩大材料上的研發，尤其InGaAsNSb材料系統 以具有相當大的band offset ratio(△Ec: △Ev = 7：3或6：4)取勝，對熱的敏感度低和電子溢流的情況有很大的改善，其電子與重電洞有效質量特性也很適合發展成光圓點、體積小的面射型雷射(VCSEL)之優勢，再加上其基板為價格便宜的GaAs更能夠降低成本，發展前景相當看好。 本研究論文的主要內容為InGaAsNSb的量子井的特性；對半導體而言，能帶圖是非常重要的特性，這意謂著侷限載子的位勢，因此能帶圖對理論計算上是非常有用的。藉由變溫的PL量測，探討氮與銻含量多寡對InGaAs/GaAs量子井光性的影響。我們發現摻1%的氮在In0.33Ga0.67As 量子井中，於不同的溫度量測下發光波長有紅位移(red shift)的現象且PL強度增強；因此，可以判定氮摻入量子井後，樣品結構確實能減少部分壓縮應力造成量子能階下降而形成紅位移。隨著銻的加入至InGaAsN量子井中，在300 K時樣品的發光波長並沒有太大的變化，當溫度降至40 K後，InGaAsNSb量子井樣品的光譜中有一發光波長約在1300 nm，其發光強度遠勝於波長約在1200 nm且有較窄的半高寬，這個光譜特徵來自於含銻的量子井所影響。此外，InGaAsNSb能階半高全寬遠比另外兩個尖峰窄，且其發光波長受溫度變化較無缺陷能階顯著，因此推估為InGaAsNSb量子井能階因隨著銻的含量增加所造成；光譜中量子能階發光尖峰在隨著銻增加後，發光強度驟然降低，我們推估為銻固然能有效的將量子井發光波長紅移至1300 nm，且隨著銻的增加超過30%造成量子井受到壓縮應力增大，使得量子井鬆弛造成界面形成許多的缺陷而降低發光效率。

InGaAsP was used to be adopted in the long-term fiber optical communication, but this material is highly sensitive to heat, which requires coolers to maintain its stability and thus increases the cost of packaging. Two new systems, InGaAsN and AlGaInAs, are found to have more advantages and appeal to more academic attention. The InGaAsNSb system, with a large amount of band offset ratio (△Ec: △Ev = 7：3 or 6：4), is particularly dominant, based on its low-sensitivity to heat and improvement in electronic current overflow. Moreover, the properties of its electronic and heavy-hole effective mass are appropriate for the development of VCSEL. In addition, the cheaper GaAs substrate will surely lower the cost and bring a promising future. This thesis aims to explore the optical properties of InGaAsNSb quantum wells. For semi-conductor, energy band diagram is a vital property, which limits the location of carriers and is helpful in calculation. Through the measurement of PL temperature, we intend to investigate the effect of N and Sb on optical properties of InGaAs/GaAs quantum wells. The result shows that adding 1% of N in In0.33Ga0.67As quantum wells will increase the strength of PL in different temperature and that the wavelengths have red shift phenomenon. As a result, we can infer that the input of N enables the sample structures to partially reduce the compressive strain and drops the optical properties so as to form red shift. With the input of Sb into the quantum wells, the wavelengths of the sample have no distinctive changes at 300K. When the temperature drops to 40K, the spectrum in the sample of InGaAsNSb quantum wells generates a wavelength at about 1300 nm, much stronger than 1200nm and with narrower FWHM (Full Width at Half Maximum), which proves the influence of Sb. Moreover, in the InGaAsNSb energy state, its FWHM is narrower than the other two peaks and the temperature influence on its wavelengths is more distinctive than that of no defect states. Therefore, we may infer that with the increase of Sb, the wavelengths of the sample have no obvious change at 300K. However, when the temperature drops to 40K, adding more than 30% of Sb will greatly lower the strength of the wavelengths in the spectrum of InGaAsNSb quantum wells sample. As a result, we conclude that although Sb can effectively red-shift the wavelengths of quantum wells to 1300nm, more than 30% of Sb will increase the compressive strain and make the quantum wells slack so as to cause many defects on the interface and lower the performance of quantum wells.