高功率半導體雷射之功率效率與熱議題研究

Abstract

本論文探討半導體雷射元件熱模擬、晶粒封裝製程技術、功率轉換效率計算的研究。在元件熱模擬上，我們使用了四種不同的模型對半導體雷射與散熱基座做模擬，模型分別為主動層熱源、電流分佈修正、溫度參數修正、界面熱阻的修正。模擬結果發現，未考慮後三項修正的主動層熱源模型與考慮修正模型的溫度差，在P side up 封裝超過100℃，在P side down封裝近80℃，這說明只單純使用主動層熱源計算熱模擬，在高電流與雷射內部溫度較高的情況下會有很大的誤差。因此在半導體雷射元件在高電流及內部溫度極高的狀況下，不能忽略電流分佈、溫度參數的修正。 在晶粒封裝製程技術上，半導體雷射封裝大多使用AuSn的硬式焊料。此焊料需要良好的共晶溫度與下壓力才能為雷射元件提供最好的散熱。在封裝品質的檢測上，一般都使用推離元件，觀察表面成份來確定封裝條件品質好壞，或者進行長時間的光性燒測實驗。然而這樣的方法除了會損壞元件之外，需要大量的時間分析。因此暫態熱電阻量測提供了省時、且非破壞性的量測。 本論文中我們結合了封裝條件與暫態熱電阻量測完成了一種非破壞性檢測封裝品質優劣的方法。暫態熱電阻提供了我們得知封裝焊料界面熱阻大小，並且能快速地得知每個封裝條件的界面熱阻值。我們使用最佳的封裝條件，讓封裝前熱飽和功率僅135mW的半導體雷射，得到了近瓦級操作的改善。封裝後雷射元件在1000mA 連續操作下尚未熱飽和，出光功率約830mW。 在功率轉換效率的研究上，我們提供了最佳效率共振腔的分析與功率餅圖的分析。我們證明了使用電流阻擋層的高功率大面積半導體雷射，提升了元件的功率轉換效率最大值。另外我們從功率餅圖上分析得知，高功率半導體雷射在臨界電流條件差異性不大時，增加微分量子效率對於功率轉換效率的改善是最佳的首選。

This thesis discusses thermal simulations of semiconductor laser devices, techniques of die bonding process, and studies of power conversion efficiency. In thermal simulations of semiconductor laser devices, we utilize four different models to simulate thermal distribution between semiconductor lasers and submount. These four models are successively considering heat sources in active region, calibration of current diffusion and joule heat, calibration of laser characteristic temperature, and calibration of interface thermal resistance between chip and submount. In our simulation results, we identify there are temperature differences over 100℃ in P-side-up packaging, and 80℃ in P-side-down packaging if we merely consider heat sources in active region other than considering these four factors. This result illustrates in thermal simulation, simply evaluating heat sources in active region will cause significant inaccuracy. Therefore, we can’t ignore other three factors whenever the semiconductor laser devices are operated at high injection current or their internal temperatures are extremely high. In techniques of die bonding process, semiconductor lasers packaging are typically utilizing hard solder of AuSn. This kind of hard solder requires adequate bonding temperature and bonding force to provide best condition of heat dissipation in laser devices. Normally we use shear test to confirm packaging condition is good or bad by observing surface quality. However, such test will cause damage to the devices. On the other hand, if we want to confirm packaging condition via non-destructive test, we need long-period constant light output aging measurement. Owing to these drawbacks, we combine packaging condition and transient thermal resistance measurement to implement a non- destructive test confirming packaging quality. Transient thermal resistance provides the values of interface thermal resistance between chip and solder and rapidly tells us each interface thermal resistance corresponding to its packaging condition. The thermal rollover power of semiconductor laser is merely 135mW before packaging. By employing the optimized packaging condition, we get nearly 1W thermal rollover power improvement. The output power of semiconductor laser device is about 830mW operated under 1000mA continuously after packaging. In studies of power conversion efficiency, we provide analyses of optimum cavity length for high power efficiency and power pie chart. We verify that implementing current blocking region design in high power broad area semiconductor lasers will enhance the maximum of power conversion efficiency of devices. Furthermore, we perceive whenever there are only minor differences in threshold current condition between high-power semiconductor lasers, increasing differential quantum efficiency is the best way to improve power conversion efficiency from the analysis of power pie chart.