1.State Key Laboratory of Luminescence and Applications, Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130033, China 2.Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100049, China
Abstract: The vertical external cavity surface emitting laser (VECSEL) is one of the hottest research fields of semiconductor lasers, due to its high power and good beam quality. However, there are few reports about how to systematically design the active region of VECSEL. In this paper, the gain design of quantum wells, which are the most important region within the VECSEL, is carried out. To achieve low power consumption under high temperature condition, epitaxial structure of the VECSEL is optimized by using the commercial software PICS3D. Firstly, the relationship between the structure of quantum well and the gain is simulated by the k·p method. Then, the gain spectra of quantum wells at different carrier densities and temperatures are compared with each other, and the optimal composition and thickness of quantum well are thus determined. The temperature drift coefficient is 0.36 nm/K, obtained by simulating the drift of the gain peak wavelength at the working temperature. Finally, the gain spectra of quantum wells with five different barriers are compared with each other. The slight blue shift of the gain peak in the quantum well with five different barriers accommodates the different emission thermal drifts of the quantum well at high temperature operation. With the GaAsP barriers on both sides of quantum well the gain characteristics of quantum wells can be improved efficiently. The designed structure is deposited by the MOCVD system. According to the reflection spectrum of the gain chip, measured by ellipsometer, the stop-band over 100 nm is centered at the about 970 nm wavelength, confirming accurate growth of the VECSEL. The 808 nm pump laser is focused on the surface of VECSEL chip at an incident angle from 30° to 50°. The VECSEL light-light characteristics are tested under the output coupling mirror with different reflectivity. The output power of VECSEL with a 97.7% reflectance output coupling mirror reaches 9.82 W at the pumping power of 35 W, without saturating the power curve. By using the external mirrors with different reflectivity, there appears the wavelength shift with the pumping power changing from 0.216 nm/W to 0.16 nm/W. Thus, the internal heating effects are different for VECSEL with different mirrors. The divergence angles at two orthogonal directions are 9.2° and 9.0°, respectively. And the circle profile of optical field shows good symmetry. Keywords:optically-pumped vertical external cavity surface emitting semiconductor laser/ quantum well/ gain chip/ high power
其中q是自由电子电荷; n是折射率; m0为有效载流子质量; c为光速; ε0为真空介电常数; h为普朗克常数. InGaAs量子阱是980 nm波段最常使用的量子阱结构, 这是因为InGaAs量子阱带来的压应变效应可以使得价带的轻重空穴带有效分离, 实现低的透明载流子密度和高的材料增益. 因此, 本研究VECSEL增益区采用InGaAs量子阱, 并首先开展量子阱的组分及厚度设计. 在模拟中采用Al0.06Ga0.94As作为势垒层材料. 使用PICS3 D软件中基于上述增益计算方法的理论模型, 对不同组分及厚度InGaAs量子阱的能带波长进行筛选. 图2(a)为计算的InGaAs量子阱能级分立后发光波长分别在970, 975以及980 nm的量子阱In组分、量子阱厚度关系. 可以看出, 这3个出光波长所对应的3条量子阱组分与厚度关系的变化趋势相一致. 对于同一种发光波长, 当量子阱厚度增加时, 需要降低In组分的含量以保持发光波长不变. 当量子阱的厚度小于6 nm时, In组分的含量随量子阱厚度的增加快速降低; 而当量子阱厚度大于6 nm后, 该趋势较为平缓. 这是因为量子阱厚度较大时, 材料的量子效应弱化, 能级分立效应不明显, 导致能带的变化主要依靠In组分来调节. 对比3种不同波长所对应的量子阱In组分和厚度曲线, 可以看出, 对于相同厚度的量子阱, 增加量子阱In组分可以得到更长的出光波长. 图 2 (a) 室温下InGaAs量子阱的发光波长为970, 975, 980 nm时, 量子阱中In组分与厚度的关系; (b) 增益峰值在980 nm InGaAs量子阱增益光谱对比; (c) 对应980 nm波长的InGaAs量子阱的价带结构(HH1, 第一重空穴; LH1, 第一轻空穴); Figure2. (a) Relationships between the In content and thickness of quantum wells when its emitting wavelength is 970, 975, 980 nm; (b) the gain spectra of different quantum wells with the same gain peak wavelength of 980 nm; (c) the valence subband structures of InGaAs QWs corresponding to a wavelength of 980 nm (HH1, the first heavy hole subband; LH1, the first light hole subband.).
图2(b)为出光波长在980 nm的InGaAs量子阱增益谱曲线. 可以看出, 不同组分及厚度的InGaAs量子阱增益谱差异非常明显. 随着量子阱厚度的增加, 量子阱的增益谱峰值不断降低. 同时, 对于厚度超过9 nm的量子阱结构, 其增益谱左侧出现一个次级的增益峰. 这是因为厚度增加, 量子阱能级分立效应变弱, 第一子能级和第二子能级分立不足, 导致两个子能级同时发光. 而第二子能级对应的能带宽度要大于第一子能级, 因此, 出现的次级增益峰所对应的波长比主增益峰要短. 总体看来, 使用窄的量子阱厚度, 可以实现高的增益峰值, 然而增益谱的宽度也在不断缩小. 并且, 量子阱厚度由4 nm增加至5 nm时, 增益峰值相差不大, 但是量子阱的增益谱宽度明显增加. 增益谱宽度的增加, 伴随着纵模限制因子的增加, 将进一步地限制光场和载流子, 防止其泄露[21]. 此外, 当量子阱厚度从5 nm继续增加至6 nm时, 增益谱宽度没有特别明显的变化, 但是增益峰强度却在降低. 因此, 从增益谱看来, 采用5 nm的量子阱厚度是比较理想的选择. 量子阱的组成和厚度在保持增益峰值为980 nm的情况下被进行了调整, 以优化增益特性. 在图2(c)中, 自上往下依次是第一重空穴带(HH1)、第一轻空穴带(LH1). 选取5, 8, 12 nm厚的不同的In组成的InGaAs量子阱系列的计算价带子带被绘成平面内波长kt的函数. 随着QWs中In含量的增加, 压缩应变增大, 导致价带的态密度降低. 同时, 第一重空穴HH1的曲率随着In组成的增加而增大, 使得价带和导带的态密度更加匹配. 这些效应有望导致更高的差分增益和更低的透明载流子密度. 这更加确定了量子阱的厚度为5 nm较为理想的推论. 量子阱的增益谱可以看出量子阱工作时可提供的各波长增益情况, 而量子阱的峰值增益随载流子浓度的变化可以看出量子阱实际工作时的增益输出能力. 图3(a)为模拟的InGaAs量子阱峰值增益随量子阱内部载流子浓度的变化关系, 增益峰值波长在980 nm. 可以看出, 不同厚度的量子阱结构随载流子浓度的变化曲线具有明显差异. 在图3(a)中, 当量子阱的厚度从4 nm增加到12 nm时, 载流子密度明显降低. 这是因为为了保持设计的波长不变, 量子阱中的In成分相应地降低, 此时量子阱材料带隙的增加导致价带顶能态密度的减小, 进而导致载流子流密度减小[22]. 此外, 量子阱厚度大于9 nm时, 很低的载流子浓度就可以产生光增益, 然而随着载流子浓度的增加, 量子阱的增益曲线快速饱和. 随着量子阱厚度不断减小, 增益曲线的饱和值不断增加, 同时, 量子阱的微分增益dg/dn不断增加. 这说明随着量子阱厚度的降低, 量子阱在高的抽运功率下可以提供的增益更高, 并且单位载流子产生的光增益也更高. 从图3(a)可以看出, 5—6 nm的量子阱厚度可以在很大的载流子注入浓度范围下获取高的增益. 图 3 (a) InGaAs量子阱的峰值增益随载流子浓度的变化关系(增益谱峰值波长位于980 nm); (b) 不同InGaAs量子阱的材料增益随工作温度变化 Figure3. (a) The change of gain peak with the carrier density within quantum wells when the gain peak wavelength is 980 nm; (b) the change of material gain with the operating temperature.
图3(b)为不同量子阱的峰值增益随工作温度的变化曲线. 随着工作温度的增加, 量子阱的峰值增益不断降低. 可以看出, 量子阱的厚度越小, 其峰值增益随工作温度的变化越明显, 也就是说, 在高温下量子阱的增益衰减就越明显. 这是因为随着工作温度的增加, 薄的量子阱厚度更容易产生注入载流子的泄露, 从而降低载流子利用效率. 然而从图3(b)的曲线可以看出, 即使薄的量子阱厚度具有更快的增益衰减速度, 在相同的工作温度下, 薄的量子阱仍然能提供更高的光增益. 根据上述对量子阱增益谱及增益-载流子浓度关系的分析, 可以确定量子阱的最佳厚度在5 nm左右. 上述设计是量子阱增益峰值位于980 nm的设计结果, 在实际器件工作时, 由于自产热效应, VECSEL的增益区往往工作在较高的工作温度. 为了保证工作时增益峰值波长与激光器的实际出光波长实现较好的匹配, 对量子阱的增益谱及增益峰值波长随工作温度的漂移情况进行了模拟, 如图4所示. 图 4 (a) 5 nm厚度的InGaAs量子阱的增益光谱随工作温度的变化; (b) 增益峰值波长随工作温度的变化 Figure4. (a) The gain spectra and (b) the gain peak wavelength of 5 nm InGaAs quantum well at different opera-ting temperatures.