Open Access
Add to BibSonomyAbstract
This study of the optoelectronic properties of blue light-emitting diodes under direct current stress. It is found that the electroluminescence intensity increases with duration of stress, and the efficiency droop curves illustrated that the peak-efficiency and the peak-efficiency-current increases and decreases, respectively. We hypothesize that these behaviors mainly result from the increased internal quantum efficiency.
© 2014 Optical Society of America
1. Introduction
High-efficiency InGaN-based light-emitting diode (LED) has been used in solid-state illumination and display products [1]. The lifetime test of an LED is important for ensuring its reliability. When the LED is stressed under fixed current and ambient temperature, it is easy to observe a meaningful enhancement of the electroluminescence (EL) spectra which has been reported in previous literature [2]. However, the physical origin of the enhancement phenomenon remains unclear. Another important issue of InGaN-based LEDs is “efficiency droop”, i.e., the efficiency of conventional LEDs reaches its peak at low current density and monotonically drops with increasing the drive current. Since the enhanced EL is inherently an efficiency issue, hence, the physics origin of these phenomena should be highly related. The efficiency droop is mainly caused by the non-radiative recombination processes. Hypothesized mechanisms include electron leakage [3], poor hole injection efficiency [4], electron overflow [5], carrier delocalization [6,7], and Auger non-radiative recombination [8]. Both electron leakage and poor hole injection efficiency have strong relationship with the quantum confined Stark effect (QCSE) induced by the piezoelectric polarization [9,10]. These leakage electrons and not injected holes may recombine outside the active layers through the non-radiative process.
This study analyzed optoelectronic properties in direct-current stressed LED samples. Experiments showed that the EL intensity and the reverse leakage current increase simultaneously. In previous reports of Ref. 2, the Mg-dopant activation of p-type GaN increased both hole injection efficiency and conductivity, which respectively caused the EL enhancement and the reverse leakage current increase. However, in our experiments, the identical capacitance-voltage (C-V) curves obtained imply that the increase in carrier concentration is not significant (not shown here). The increased thermal activation energy indicates a change in the active region after application of the stress. Thus, we hypothesize that the EL enhancement of our sample results from the increased internal quantum efficiency (IQE) which is mainly attributable to the annealing effect. The annealing effect reduces the defect density, especially in active regions. Since the non-radiative recombination channels are impeded, free electrons are more easily injected into the radiative states and then overflow the potential minimum of the radiative states (accumulation effect). The linear dependency trend in the integrated EL intensity versus current density curves further support our inferences. Finally, experiments were performed under different stressing conditions to investigate the mechanism of EL enhancement.
2. Experiment
The 1 mm2 InGaN-based LED sample used in the experiment was grown on a (0001) sapphire substrate. The LED structure comprised of a 3 um undoped GaN layer, a 2 um highly conductive n-type GaN layer, an InGaN/GaN (~3 nm/10 nm) multiple quantum wells (MQWs) active layer, and a 200 nm p-type GaN layer (Mg doping concentration about 1.0x1017 cm−3). The LED sample was soldered on the Cu-Al base and stressed under fixed direct current in an isothermal chamber. The optoelectronic characteristics were measured before and during the stressing period. The EL measurement was performed with a Keithley 2430 power supply with a ðxed pulse width of 20 ms to provide a high signal-to-noise ratio, and the optical measurement was performed with an integrating sphere (IS;CAS 140B). The current-voltage (I-V) characteristics were measured with a Keithley 236 source measure unit.
3. Results and discussion
The LED sample was stressed under fixed direct current of 100 A/cm2 in an isothermal chamber (25 °C) for up to 1574 hours. Before and after different stressing time the sample was measured at different current (0.1, 0.3, 0.5, 5 and 50 A/cm2) at room temperature. Figure 1(a) shows the time-dependent EL intensities, which were normalized to that of the unstressed sample for each current level before and after application of stress. The EL intensity approached saturation during the first 100 hours. The enhancement at low current density exceeded that at high current density. It is because the accumulation effect is particularly evident at low injection current, hence, the degree of the enhancement is diminishes at the higher injection current. Figure 1(b) compares the EL spectra at low and high current density. The main peaks of the EL spectra were attributable to the band-edge transitions, and no defect signals were observed. Notably, although the EL intensity substantially increased, the peak position and line width were unaltered for each current level, which indicates that the radiative recombination channel remain the same after the stress.
Fig. 1 (a) Normalized EL intensity and (b) EL spectra of the sample measured after different stressing durations.
Figure 2(a) shows the efficiency droop behavior, which was analyzed to identify the enhancement mechanism. The deviation in the efficiency droop curve was large at low current density but small at high current density as the enhancement variation of EL intensity. These behaviors were not attributable to improved extraction efficiency, which showed similar enhancement at all current level. Therefore, these behaviors should be the result from the improved IQE. According to the previous report [2] and the unchanged radiative recombination channels in this study, we infer that the enhanced IQE should be resulted from the dopant activation of p-type GaN and the annealing effect of the active region, respectively. However, the identical C-V curve indicates that the EL enhancement caused by the annealing effect is larger than that caused by dopant activation. The annealing effect reduces defect density, hinders the non-radiative recombination channels, and then enhances electron injection into the radiative states. The electron accumulation process in the potential minima of radiative state is particularly evident in the low current region, which causes the non-uniform improvement in IQE. As stressing time increases, the defect density further decreases, which causes a peak-efficiency-current shift toward to the low drive current and the peak-efficiency increases as shown in the Fig. 2(b). The IQE can be described by the following equation:
where the confections A, B, C are non-radiative, radiative, and Auger non-radiative recombinations, respectively, and where n is the carrier density. Because the peak-efficiency-current is smaller than 3 A/cm2. In this low current region (<3 A/cm2), the Auger recombination can be ignored. The reduced defect density in active region causes a reduced A coefficient, which could yield an increase peak efficiency, a decreased peak-efficiency-current, and a enhanced IQE.Fig. 2 (a) Wall plug efficiency (WPE) as a function of the current density. (b) Peak-efficiency and peak-efficiency current as a function of stressing time.
Figure 3 compares the I-V characteristics of the sample. The figure shows that the reverse leakage current increases as stressing time increases. It is well known that the threading dislocations (TDs) are the major leakage current paths [5,11], and the In-rich regions enhance the wave function overlap between the electron-hole pairs, which prevent carriers from being trapped in the non-radiative pathways inside TDs [7]. Thus, we believe the damage of LEDs occurred mainly in the position of TDs and most of the increased leakage current are passed through the TDs, simultaneously, the reduced defect density at the In-rich regions which causes most carriers at the In-rich regions to recombine radiatively. Under forward bias, most carrier injection passes through the p-n junction and In-rich regions, which has lower resistance. However, at reverse bias, most carrier injection passes through the TDs. Thus, the EL enhancement and reverse leakage current increase can occur simultaneously. The literature shows that TD density has a minor role in the efficiency droop behavior, which further supports our inference [5]. We hypothesize that most damage occurs near the TDs, and the annealing effect is concentrated in the active (In-rich) region. The almost unchanged forward I-V and C-V (not shown here) curves also imply the p-n junction is unchanged.
The emission properties during the stressing process were further clarified by constructing L-I curves at room temperature and constructed the Arrhenius plots of the integrated photoluminescence (PL) intensity range from 10 to 300 K. Figure 4(a) shows that the L-I curves were characterized by
where L is the integrated EL intensity (W), I is the current density (A/cm2), P is a constant, and m is an exponent parameter [12]. At low current injection (I<1 A/cm2), exponent m significantly decreases and approaches 1 as stressing time increases. The value of P0 is 0.0129, P6 is 0.0134, and P74 is 0.014. At high current density (I>20 A/cm2), exponent m approaches 0.7, which approximates the ideal value for Auger processes (m = 2/3). The value of P0 is 0.0335, P6 is 0.033, and P74 is 0.0328. At medium current density (1<I<20 A/cm2), exponent m increases to approximately 0.87, which implies that the emission is only partially limited by the Auger recombination. The value of P0 is 0.0165, P6 is 0.0174, and P74 is 0.0182. The linear dependency (m~1) of the low current injection indicates that most of the injected carriers recombine radiatively. That is the Shockley-Read-Hall (SRH) processes occur via non-radiative recombination centers and surface/interface recombination are irrelevant for the stressed sample. For the Arrhenius plots shown in Fig. 4(b), a pulse diode laser operating at 405 nm with an average power of 1 mW was used as the excitation source. Because the 405 nm pulsed laser could only excite the InGaN quantum wells, the variation in optical properties had a strong relationship with the active regions. The PL signal was recorded with an SPEX-1403 double-grating monochromator and detected by a PMT detector. The temperature PL intensity (IPL(T)) can be described by the following equation:where Ea is the thermal activation energy, coefficient C is the strength of quenching processes, k is the Boltzman constant, and T is the temperature [13,14]. The different Ea value reflects that the temperature dependency of the captured cross section to the non-radiative centers varies according to the degree of localization. A larger Ea of the stressed sample indicates higher activation energy for the transfer of confined carriers to non-radiative recombination center.Fig. 4 (a) Integrated EL intensity as a function of current density after varying durations of stress. (b) Logarithmic value of the integrated PL intensity as a function of temperature.
In order to understand the mechanism, we have derived the LEDs at three different stressing conditions: 350 mA at 25°C (condition A), 700 mA at 55°C (condition B), and 1 A at 25°C (condition C). The junction temperatures of the LED chips were estimated using the following equation:
where Tj is the junction temperature, Rth is the thermal resistance, IV is the injection power, and Tb is the temperature of Al base (measured by the thermocouple). The Rth values for the samples were measured using a transient thermal tester (T3Ster, MicReD Ltd.) Additional details of the measurement procedure are given in the literature [15]. Figure 5(a) shows that the Rth (IV; Tb) of conditions A, B, and C are 8 k/W (IV = 1.18 W; Tb = 33°C), 9.5 k/W (IV = 2.8 W; Tb = 71.2°C), and 11.5 k/W (IV = 4.58 W; Tb = 46.4°C), respectively. Hence, the respective junction temperatures are 42.4°C, 97.8°C, and 99.1°C. These estimated junction temperatures are average value which may lower than the local temperature caused by the current crowding effect. Because the annealing effect is attributable to the high temperature. Figure 5(b) shows that the sample under condition B or C had higher EL enhancement compared to condition A. The EL enhancement under condition B and C are similar due to the similar junction temperature. Besides, the high electric current induces the defect generation at the same time. The competing effects of annealing and defect generation are apparent during the stressing process. In the early stage, the defect generation is not dominant. However, after 800 hours, further defect generation in the active region causes a gradual decrease in EL intensity. Under condition A, the decrease in EL intensity is not evident due to the low electric current. These experimental results explain the variation in EL intensity during the stressing process. As reported previously, further increases in injection current or junction temperature may cause the defect generation to dominate and result in an EL intensity decrease in the early stage [2].Fig. 5 (a) Cumulative structure functions of different injection current densities. (b) Normalized EL intensity as a function of the stressing time under different stressing conditions.
4. Conclusion
In conclusion, this study showed that enhancement of EL intensity is highly related to the increased IQE due to the annealing effect, which mainly reduce defect density in the active region. The annealing effect and defect generation are competing with each other during the stressing process. The increased reverse leakage current passes through the TDs. Besides, the linear dependent L-I curves and the increased thermal activation energy also support our inferences.
References and links
1. G. Harbers, S. J. Bierhuizen, and M. R. Krames, “Performance of high power light emitting diodes in display illumination applications,” J. Display Technol. 3(2), 98–109 (2007). [CrossRef]
2. L. Liu, M. Ling, J. Yang, W. Xiong, W. Jia, and G. Wang, “Efficiency degradation behaviors of current/thermal co-stressed GaN-based blue light emitting diodes with vertical-structure,” J. Appl. Phys. 111(9), 093110 (2012). [CrossRef]
3. L.-B. Chang, M.-J. Lai, R.-M. Lin, and C.-H. Huang, “Effect of electron leakage on efficiency droop in wide-well InGaN-based light-emitting diodes,” Appl. Phys. Express 4(1), 012106 (2011). [CrossRef]
4. Y.-K. Kuo, J.-Y. Chang, and M.-C. Tsai, “Enhancement in hole-injection efficiency of blue InGaN light-emitting diodes from reduced polarization by some specific designs for the electron blocking layer,” Opt. Lett. 35(19), 3285–3287 (2010). [CrossRef] [PubMed]
5. Y. Yang, X. A. Cao, and C. H. Yan, “Rapid effciency roll-off in high-quality green light-emitting diodes on freestanding GaN substrates,” Appl. Phys. Lett. 94(4), 041117 (2009). [CrossRef]
6. J. Wang, L. Wang, W. Zhao, Z. Hao, and Y. Luo, “Understanding efficiency droop effect in InGaN/GaN multiple-quantum-well blue light-emitting diodes with different degree of carrier localization,” Appl. Phys. Lett. 97(20), 201112 (2010). [CrossRef]
7. S.-F. Yu, R.-M. Lin, S.-J. Chang, and F.-C. Chu, “Efficiency droop characteristics in InGaN-based near ultraviolet-to-blue light-emitting diodes,” Appl. Phys. Express 5(2), 022102 (2012). [CrossRef]
8. Y. C. Shen, G. O. Mueller, S. Watanabe, N. F. Gardner, A. Munkholm, and M. R. Krames, “Auger recombination in InGaN measured by photoluminescence,” Appl. Phys. Lett. 91(14), 141101 (2007). [CrossRef]
9. J. Xu, M. F. Schubert, A. N. Noemaun, D. Zhu, J. K. Kim, E. F. Schubert, M. H. Kim, H. J. Chung, S. Yoon, C. Sone, and Y. Park, “Reduction in efficiency droop, forward voltage, ideality factor, and wavelength shift in polarization-matched GaInN/GaInN multi-quantum-well light-emitting diodes,” Appl. Phys. Lett. 94(1), 011113 (2009). [CrossRef]
10. J.-Y. Chang and Y.-K. Kuo, “Influence of polarization-matched AlGaInN barriers in blue InGaN light-emitting diodes,” Opt. Lett. 37(9), 1574–1576 (2012). [CrossRef] [PubMed]
11. X. A. Cao, J. M. Teetsov, M. P. D’Evelyn, D. W. Merfeld, and C. H. Yan, “Electrical characteristics of InGaN/GaN light-emitting diodes grown on GaN and sapphire substrates,” Appl. Phys. Lett. 85(1), 7–9 (2004). [CrossRef]
12. W.-H. Chang, A. T. Chou, W. Y. Chen, H. S. Chang, T. M. Hsu, Z. Pei, P. S. Chen, S. W. Lee, L. S. Lai, S. C. Lu, and M.-J. Tsai, “Room-temperature electroluminescence at 1.3 and 1.55μm from Ge/Si self-assembled quantum dots,” Appl. Phys. Lett. 83(14), 2958–2960 (2003). [CrossRef]
13. J. I. Pankove, Optical Processes in Semiconductors (Prentice-Hall, 1991), Chap. 7.
14. M. Leroux, N. Grandjean, B. Beaumont, G. Nataf, F. Semond, J. Massies, and P. Gibart, “Temperature quenching of photoluminescence intensities in undoped and doped GaN,” J. Appl. Phys. 86(7), 3721 (1999). [CrossRef]
15. C.-P. Wang, T.-T. Chen, H.-K. Fu, T.-L. Chang, and P.-T. Chou, “Transient analysis of partial thermal characteristics of multistructure power LEDs,” IEEE Trans. Electron. Dev. 60(5), 1668–1672 (2013). [CrossRef]
