Open Access
Add to BibSonomy
Abstract
GaAs-based oxide-confined vertical-cavity surface-emitting lasers (VCSELs) exhibit relatively low resistance against reliability-related damage. In order to gain a deeper understanding of the degradation and failure mechanism in oxide-confined VCSELs caused by electrostatic discharge (ESD)-induced defect proliferation, we investigated the effects of ESD stress on the degradation of optical-electrical characteristics and the evolution of defects in VCSELs under human body model test condition. The degradation threshold values for forward and reverse ESD pulse amplitudes were estimated to be 200 V and -50 V, respectively. Notably, VCSELs demonstrated greater sensitivity to reverse bias ESD compared to forward bias ESD. Analysis of optical emission and microstructure provided evidence that the device failure is attributed to an increase in ESD current density, leading to the multiplication of dark line defects (DLDs) originating from the edge of the device's oxide aperture. The formation of defects occurred suddenly in discrete events within small regions, rather than progressing gradually and uniformly. These defects propagated and led to damage across the entire active region. We believe that our results would be meaningful for improving the reliability of VCSEL in the future.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Since VCSELs have many unique features, they have solidified their status as a major active device in microwave photonic links, short-distance optical interconnections, laser radar and chip-scale atomic clocks [1–5]. Meanwhile, gigabit Ethernet and fiber channel applications are driving the current VCSEL market [6–10].
However, one of the challenging issues in VCSEL reliability is their sensitivity to ESD [11,12]. There is still a critical issue to restrict the further expansion of VCSEL application. Due to the active region of VCSEL is much smaller than that of edge-emitting laser, the sensitivity to ESD is notably heightened [13,14]. Even the unintentional touch by a finger could result in a degradation of the device's performance. ESD events are difficult to prevent and monitor. This concern becomes even more pronounced for high-speed GaAs-based oxide-confined VCSELs, as the pursuit of higher modulation frequencies necessitates the reduction of device dimensions [15].
ESD events occur because of a charge imbalance between the electronic device and another object. As one of the electrical overstress failure modes, it is characterized by limited energy flow and short duration. The real challenge of ESD damage stems from the fact that the performance degradation of devices may not immediately be revealed, potentially resulting in significant costs. As the human body is a primary and highly damaging source of electrostatic discharge, the most common measurement of ESD sensitivity is performed using the human body model (HBM). As a result, the HBM has become the most widely used model for ESD testing of semiconductor lasers [15–17].
Although the increasing vulnerability of VCSELs in respect to ESD damage and the associated potential costs cannot be estimated, it is surprising that only a limited amount of research has been conducted on the impact of ESD events on VCSEL performance. While it has been suggested that ESD damage can have a fatal impact on the optical properties of VCSELs [12,18], a detailed analysis of the underlying mechanisms to comprehend the influence of ESD-induced defect evolution on the optical-electrical characteristics of VCSELs has not yet been established. Thus, the purpose of this investigation is to analyze the degradation in the electrical and optical characteristics of GaAs-based oxide-confined VCSELs by subjecting them to ESD pulse voltages, to demonstrate the location of defects caused by ESD, as well as to track the evolution of these defects with varying amplitude of ESD pulses. The outcomes of this study will significantly enhance our understanding of the mechanisms leading to device degradation and failure caused by ESD events.
Section 2 provides a concise overview of the tested devices and the ESD test setup. Section 3 focuses on presenting the alterations in electro-optical characteristics resulting from forward and reverse bias ESD, with particular emphasis on the degradation phase. Furthermore, Section 3 delves into the discussion of the degradation process and the mechanisms leading to failure. Finally, Section 4 summarizes the conclusions drawn from this study.
2. Description of device and test method
In this study, the GaAs-based oxide-confined VCSEL is designed for 850 nm operation, the outlook and schematic cross-section of a VCSEL are shown in Fig. 1(a) and (b), respectively. The VCSEL consisted of p-type distributed Bragg reflectors (DBR), n-type DBR and λ cavity, the λ cavity sandwiched by 18.5 pairs p-type doped the top-DBR and 40.5 pairs n-type bottom-DBR mirror layers. The λ-cavity is composed of two Al0.3Ga0.7As spacers. The oxide layer, a 33 nm thick Al2O3, is positioned above the spacers. The active region is composed of three 10 nm GaAs quantum-well layers with 8 nm Al0.3Ga0.7As barriers between the wells. The epitaxial material is grown by metal-organic chemical vapor deposition (MOCVD) on an n-type GaAs substrate. Each device is mounted on a TO-46 header with a glass cap. All test VCSELs are provided by CGPhoton Incorporation. The cap is removed when observing the surface of the device and electroluminescence spectrum. All the tests and device characterizations are performed at room temperature.
Fig. 1. (a) Outlook of the VCSEL under microscope test. (b) Schematic cross-section of the device. (c) Circuit diagram of the HBM model. (d) Typical waveform of an electrostatic generator’s output current.
The pulse stress was applied to the devices using a custom ESD system, and the topology for applied stress is shown in Fig. 1(c). In the test, a high voltage from the DC power supply was applied. When the relay was turned on, a charged 150 pF capacitor discharged its energy to the laser through a resistor. The pulse duration was approximately 100 ns. The 150 pF capacitor simulates the charge stored on an average human body, while the resistor simulates the resistance of the human body and skin. The voltage step was set to 200 V for forward bias ESD pulse accumulation and -100 V for reverse bias ESD pulse accumulation, respectively. During the test, the possibility of excessive electrostatic discharge stress causing an open circuit in the device was eliminated. Figure 1(d) displays the typical waveform of the output current from an electrostatic generator during the ESD pulse.
The electrical parameters were measured using an Agilent B1500A semiconductor parameter analyzer. The optical output power was measured with a PM-100D photoelectric detector. Electroluminescence spectrum characterization was performed using an Andor SR-303i-A spectrum analyzer to observe changes in the device’s optical mode behavior. The output light from each TO-46 mounted device was directly coupled to the optical spectrum analyzer via multi-mode fiber. The emission spectra were measured at room temperature with a wavelength resolution of 0.05 nm. Low-frequency noise characterization was conducted by an SR785 dynamic signal analyzer, with a filter and amplifier units provided by Pro-plus 9812B. Emission microscopy imaging (EMMI) was carried out using a Phemos-1000 photo emission microscope (PEM). The microstructures were analyzed using a Helios G4 CX model scanning electron microscope (SEM) dual-beam system.
3. Results and discussion
3.1 Effect of the ESD on light-current-voltage
Figure 2(a) shows the typical degradation of the optical output power-current (L-I) characteristics of the devices during the forward bias ESD pulse accumulation test. As is evident, the optical output power decreased faster and faster with increasing ESD pulse amplitude. The slope efficiency and threshold current of the devices did not exhibit noticeable changes until an applied pulse amplitude of 200 V. Beyond this threshold, a decline commenced at the application of a 400 V pulse amplitude, resulting in a 10% reduction in slope efficiency compared to the fresh device. The threshold current increased by 14.1%, and the peak optical output power decreased by 9.3%. These observations suggest that the ESD damage threshold of the device lies around 200 V for forward bias ESD. Figure 2(b) shows the typical degradation of the L-I characteristics of the devices during the reverse bias ESD pulse accumulation test. We observed a similar trend to that of the forward bias ESD for all parameters. The results demonstrate minimal changes in the L-I characteristics when the reverse bias ESD voltage amplitude up to -50 V. However, when the reverse bias ESD amplitude is increased to -150 V, the optical output power starts to decrease rapidly, and the slope efficiency drops by 51%. As the reverse bias ESD amplitude up to -250 V, the slope efficiency decreased to 25.5% of the initial value. Additionally, the threshold current increases from 0.71 mA to 6.47 mA under this reverse bias ESD amplitude. In order to provide a clear representation of the variation in electrical-optical performance parameters, we have presented the values in Table 1, where EQE is abbreviation for the external quantum efficiency.
Table 1. Electrical-optical performance parameters of the VCSELs by the ESD pulses
Figure 3(a) displays the measured current-voltage (I-V) curves obtained from the forward bias ESD pulse accumulation test. Prior to the ESD test, a reverse bias breakdown voltage of approximately 9.72 V is observed. As the ESD pulse amplitude increases, the reverse bias breakdown voltage decreases and the reverse leakage current gradually increases. Notably, a significant increase in reverse leakage current is observed when the ESD pulse amplitude up to 600 V. However, no significant changes in the forward bias current are observed at an ESD pulse amplitude of 400 V. Upon increasing the ESD pulse amplitude to 600 V, the forward bias current increases for bias voltages below 1.5 V. The forward bias and reverse bias currents exhibit symmetry with respect to the 0 V axis for operating voltages from -0.5 V to 0.5 V. This symmetry suggests that the dominant component of the total operating current at low bias voltages is the resistive leakage current. Furthermore, the total operating current increases as the ESD pulse amplitude is elevated. These trends are consistently observed in both forward and reverse bias ESD tests. Figure 3(b) presents the measured I-V curves obtained from the reverse bias ESD pulse accumulation test. The reverse current only increases by 80 pA at an operating voltage of -4 V when the ESD voltage is adjusted from 0 V to -150 V. However, a drastic increase to 749 pA is witnessed following the application of a -250 V pulse. Similar to the forward bias ESD, the resistive leakage current component dominates the total operating current at low bias voltages in the reverse bias ESD scenario. Moreover, the total operating current rises with an increase in the ESD pulse amplitude, which is consistent with previous observations.
In order to determine the minimum ESD pulse amplitude that causes degradation of the device performance, we monitored the current at a given reverse bias voltage and the emitted optical power at a given bias current after each pulse. The bias parameters were set to remain well below the reverse bias breakdown voltage and well above the laser threshold currents to prevent degradation. The development of optical power at a given laser current of 8 mA and the reverse bias current at a fixed reverse bias voltage of -6 V during forward and reverse ESD step stress tests of the VCSELs are shown in Fig. 4. In Fig. 4(a), we observe different thresholds at which optical and electrical properties start to deteriorate during forward bias ESD stress tests, rather than a same ESD pulse amplitude degradation threshold value. The optical power starts to decrease after the application of a 200 V pulse, while the reverse bias current remains stable up to 400 V and increases sharply after applying a 600 V pulse. Similar behavior is observed during reverse bias ESD stress tests in Fig. 4(b). The optical power starts to decrease significantly, and the reverse bias current remains stable after the application of a -150 V pulse, but the reverse bias current increases sharply when the reverse ESD pulse amplitude reaches -250 V. Therefore, we estimate that the forward ESD pulse amplitude degradation threshold value is 200 V, and the reverse ESD pulse amplitude degradation threshold value is -50 V.
3.2 Effect of ESD on optical spectrum
The far-field pattern is an important factor of VCSELs when considering them as a laser source. The 2D far-field profiles and relative intensity distribution at a distance of 40 mm from the VCSELs, under various ESD pulse amplitudes, are shown in Fig. 5, with the driving current set at 8 mA. It is evident that the far-field distribution shows minimal degradation until the forward bias ESD pulse amplitude reaches up to 400 V. A similar observation can be made for the reverse bias ESD pulse amplitude, which remains stable up to -50 V. However, the degradation process accelerates when the forward bias ESD pulse amplitude goes beyond 600 V, or in the case of the reverse bias ESD pulse amplitude exceeding -150 V. And the fundamental mode having its intensity confined to the central region of quantum wells that is significantly suppressed. This phenomenon is likely attributed to both the local depletion of carriers in the central quantum wells due to the stimulated emission process and enhanced absorption in the central region by free carriers with degraded mobility according to Drude mode [19].
Fig. 4. Optical power at given bias current and reverse bias current at a fixed bias voltage during: (a) the forward bias ESD step stress and (b) the reverse bias ESD step stress testing of the devices.
Fig. 5. 2D far-field profiles and relative intensity distribution of the VCSELs during the forward bias ESD and the reverse bias ESD.
To better identify early signs of deterioration in the optical properties and gain valuable insights into variations in the performance of the VCSEL's active layer, we conducted an investigation into the stimulated emission spectrum of the laser. This spectrum is presented in Fig. 6, with the laser current set at 8 mA. Specifically, Fig. 6(a) illustrates the optical emission spectra for the forward bias ESD conditions.
Initially, the optical emission spectrum of the pristine VCSEL exhibited two distinct peaks. However, after subjecting the VCSEL to ESD pulses with amplitudes of 200 V and 400 V, the optical emission spectrum revealed three peaks. Notably, when exposed to an ESD pulse amplitude of 600 V, the optical emission spectrum displayed only a single peak. Additionally, it was observed that the peak wavelength following the ESD pulse was shorter compared to that of the pristine VCSEL. Conversely, in cases where the ESD pulse amplitude reached up to 600 V, a red shift in the peak wavelength was evident. This observation implies that ESD-induced degradation primarily occurs along the active area adjacent to the oxidized region. Consequently, there is a reduction in the effective active area, leading to a decrease in the effective refractive index and subsequently causing a blue shift in the lasing spectrum. Notably, when subjected to ESD levels as high as 600 V, the VCSELs experienced substantial resistivity degradation, resulting in excess Joule heating. This phenomenon led to both an increase in the refractive index and an expansion of the cavity length, ultimately causing the observed red shift in wavelength. As a corollary, the decrease in the number of lasing modes can also be attributed to these same factors [19,20].
To investigate the spatial mode behavior when subjected to reverse bias ESD pulses, we conducted optical spectrum measurements on the devices following the same procedure employed in the forward bias ESD pulse accumulation test. As depicted in Fig. 6(b), the optical spectrum of the pristine VCSEL exhibited a multi-mode behavior characterized by two peaks. In contrast, the spectrum from the ESD-damaged device displayed only a single peak for reverse ESD pulse amplitudes exceeding -150 V. It's worth noting that the peak wavelength following the ESD pulse at -50 V was shorter than that of the pristine VCSEL. Similarly, a red shift in the peak wavelength was evident when the ESD pulse amplitude reached up to -150 V. These findings can be explained in a manner consistent with the explanation provided in Fig. 6(a).
3.3 Effect of ESD on low-frequency noise
To analyze the possible noise sources and defect evolution within the devices under ESD stress, low-frequency noise analysis was applied. The change in the current noise power spectral density of the devices before and after ESD was examined, with a bias current equals to 0.1 mA. As shown in Fig. 7, the low-frequency noise is the purely 1/f noise characteristics within the frequency range of 1-1000 Hz. Furthermore, the current noise power spectral density increases with higher ESD pulse amplitudes, signifying a degradation in the performance of the 1/f noise. According to the theory of Reimbold, as expressed below [21,22]:
the higher SI, the higher the defect density (Nt) in the device, it is indicated that there were more defects created in internal of the devices. The case was closely related to the reliability of the devices.Fig. 7. The current noise power spectral density curve versus frequency of the VCSELs. (a) the forward bias ESD. (b) the reverse bias ESD.
The relationship between the power spectral density of the current noise (SI) and the bias current level is depicted in Fig. 8. The ambient temperature is 295 K, and the noise frequency is set at 10 Hz. The curves can be categorized into three regions. As illustrated in Fig. 8(a) for the forward bias ESD. In the high current region (> 5 × 10−4 A), SI and I exhibit the following relationship: $S_I\propto I^{n}$ (n≈2). Here, the 1/f noise predominantly arises from fluctuations in the series resistance. The noise in this region increases with the ESD pulse amplitude increase, indicating a degradation in contact properties. In the low current region (I < 10−6 A), SI is proportional to I, with the noise originated from the junction area. When the ESD pulse amplitude is up to 600 V, it can be inferred that some defects in the junction area are rapidly generated or amplify. In the median current region (5 × 10−4 A > I > 10−6 A), the noise curves exhibit a transitional region that changes gradually with an increase in bias current. While the noise in this region rises with the ESD pulse amplitude, indicating increasing defects. The findings are displayed in Fig. 8(b) for the reverse bias ESD. Similarly, for the reverse bias ESD, the results in Fig. 8(b) show that SI is approximately proportional to I2 in the high-current region. The noise in this region increases gradually with the ESD pulse amplitude. In the low current region, $S_I\propto I$, and the noise rapidly increases with the ESD pulse amplitude, suggesting a rapid proliferation of defects in the junction area. In the median current region, the noise curves exhibit a transition region, with noise increasing alongside the ESD pulse amplitude.
Fig. 8. The current noise power spectral density versus the applied forward bias current for the VCSELs with the ESD pulse amplitude.
The plot in Fig. 9 illustrates the relative change in the current noise power spectral density, ΔSI/SI0, concerning both forward bias and reverse bias ESD pulse amplitude. In this context, ΔSI =SI-SI0, where SI0 represents the current noise power spectral density before the ESD stress, SI represents the current noise power spectral density after ESD stress. As depicted in Fig. 9, there is an increase in ΔSI/SI0 corresponding to higher ESD pulse amplitudes. Notably, when the forward bias ESD pulse amplitude reaches 800 V, the relative change in current noise power spectral density escalates by a factor of 104 compared to before the ESD at a frequency of 10 Hz. Similarly, with a reverse bias ESD pulse amplitude of -250 V, the relative change in current noise power spectral density also undergoes a 104-fold increase relative to before the ESD. These observations suggest that the VCSELs exhibited greater sensitivity to reverse bias ESD in comparison to forward bias ESD. Moreover, the degradation resulting from reverse bias ESD appears to be more pronounced.
Fig. 9. The curves of the relative change in current noise power spectral density versus the ESD pulse amplitude for the VCSELs.
3.4 EMMI analysis of ESD induced defects
EMMI images of the VCSELs before and after ESD are shown in Fig. 10. The figure displays the typical emission profile of the device after the ESD, revealing distinct bright red spots at the edge of the light output aperture. In this study, the reverse microampere-level current method was employed to detect characteristic changes in the optical emission as a potentially fingerprint of this failure mode. The operating current was set at -1µA. The provided microscopic images were captured immediately after the ESD stressing.
For comparison, Fig. 10(a) displays EMMI images before the application of ESD, where only a few faint glowing spots are visible. In contrast, the other images in Fig. 10 clearly shown that prominent bright spots after the application of ESD pulse, and the size of these bright spots noticeably increase with the ESD pulse amplitude increase.
A distinct bright spot emerges following the application of a 200 V ESD pulse, while densely distributed small bright spots surround two main bright spots after an ESD pulse amplitude of 400 V. Upon increasing the ESD pulse amplitude to 600 V, bright speckles appear along the edge. This escalation in the amplitude of the ESD pulse exacerbates ESD-induced degradation, causing significant damage across the entire active area, including the edge region, at this voltage. In these areas, the number of non-radiative centers outweighs the number of radiative ones. Under reverse bias conditions, the VCSEL emission intensity hinges on the intrinsic shunt resistance. However, these spots introduce additional leakage pathways, leading to localized reduction in reverse device breakdown. As a consequence, these spots exhibit heightened electrical and optical activities. A similar phenomenon occurs in the case of reverse bias ESD. The internal crystal structure of VCSEL sustains damage from ESD, creating increased defects. These defects act as non-radiative recombination centers, absorbing carriers and releasing energy. As a result, the bandgap contracts as the junction heats up. This process can be visualized as a reduction in barrier height with increasing junction temperature. Initially, this leads to a rise in the threshold carrier density with temperature. At sufficiently high temperatures, the hetero-structure becomes inadequate in confining the carriers, causing leakage over the hetero-structure to dominate.
3.5 Microstructure analysis of ESD induced defects
To uncover microstructure changes in the dark area of the device after ESD damage, and gain a deeper understanding of the defects responsible for the observed EMMI, a dual-beam system utilizing focused ion beam (FIB) was employed. This system was used to cut and mill the dark area, creating cross-sectional SEM devices from the failed device.
The electroluminescence images of the fresh and failed devices are presented in Fig. 11(a) and 11(b), respectively. The forward bias voltage of the device was set at 1.3 V, which is below laser threshold. The emission intensity of the failed device was obviously lower than that the fresh device. Additionally, the emission intensity distribution of the failed device was uneven, resulting in the appearance of dark areas in the emission. Consequently, the optical activity is reduced. The reverse bias EMMI image of the failed device is depicted in Fig. 11(c). The emission pattern reveals activity across the entire emitter, with four main spots at the edge of the laser aperture under forward bias conditions, which corresponds to the device shows a lower light output in the edge of the laser aperture, indicating the presence of defects in the active region rather than within DBRs of this device in this region, these defects are DLDs are presumed [23].
The colorized close-up of the insert corresponds to the red dots areas in Fig. 11(c), and the solid green line shows the precise position of FIB incision. The incision image of the FIB is shown in Fig. 11(d), corresponding to the SEM image of the FIB cross section shown in Fig. 11(e). The cross section of the FIB covers the entire failure region. The SEM image reveals that some areas are noticeably blurred at the edge of the oxide aperture and the quantum well hetero-junction in the active region of the device. The defects seem to originate from the active layers beneath the edge of the oxide layer. Figure 11(f) shows a magnified image of the region of interest (the red box) in Fig. 11(e). The result shows that evident traces of dislocation at the boundary of the oxide aperture, and the hetero-junction interface appears blurred, with a large number of high-density dislocation defects generated and spreading to the surrounding areas. There are also traces of dislocation in the active region below the boundary of the oxide aperture, which can be observed to extend to the bottom DBR. Generally, during electrically accelerated aging or ESD damage, dislocation defects generate in the active region of VCSEL [24,25], leading to the spread of these dislocation defects to the surrounding area under electrical stress. This confirms that the expansion of red dark spots in the EMMI results is caused by the formation of defects.
Combining the analysis of the EMMI and the microstructure, the damaged region was consistently located near the oxide layer in all ESD tests. The most severe damage was always manifested at the edge of the oxide aperture, and the interface between oxide layer and the surrounding mirror layers was decorated with defects. Generally, ESD damage arises from either by high current densities or high electric fields. The failure mode typically occurs in thin conductive films [26]. As depicted in Fig. 12(a), electrical stress accumulates rapidly during the ESD process. In the vicinity of the oxide aperture edge, the aperture geometry of the VCSEL causes an elevated current density compared to the center of the oxide aperture. Given the highly conductive nature of the aperture region, it is presumed that the ESD induces localized joule heating, resulting in a rapid increase in junction temperature and the formation of a steep temperature gradient within the device. The lattice mismatch, coupled with the difference in thermal expansion coefficients between the oxide and the semiconductor, leads to strain that generates structural defects near the oxide layer. These defects proliferate and propagate, culminating in an expanded region of DLDs (defect-related luminance distributions). This sequence explains why VCSELs experience the proliferation of defects, ultimately leading to device failure. Additionally, these defects initially formed close to the sides of the metal lines bordering the oxide aperture, which was shown in the red dashed box in Fig. 12(b). This observation further underscores that regions with higher current density are more susceptible to ESD damage.
Fig. 12. A schematic illustration of defects provide a pathway for leakage current at the oxide aperture edge.
Furthermore, the localized nature of these defect sites, as shown in Fig. 11(c), suggests discrete failure sites. By observing the changes in reverse emission intensity with varying ESD pulse amplitudes, as depicted in Fig. 10, it becomes easier to see that defect formation around the edge of the oxide aperture occurs in sudden, discrete events within small regions, rather than gradually and uniformly. This situation may be caused by the high current density, which accelerates the propagation of defects at the intrinsic defects within the internal devices.
4. Conclusion
In the current study, we focused on the degradation and failure mechanism of GaAs-based oxide VCSEL caused by the proliferation of ESD-induced defects. The degradation threshold values for forward and reverse ESD pulse amplitudes are estimated to be 200 V and -50 V, respectively. The degradation of optical-electrical properties exhibited an accelerated decline as the ESD pulse amplitude increased. Notably, the intensity of the fundamental mode, which is confined to the central region of the quantum wells, was significantly suppressed. The 1/f noise was found to originate primarily from the active region. Interestingly, the VCSELs more sensitive to reverse bias ESD than forward bias ESD, resulting in faster generation and accumulation of defects. This was evident through far-field patterns, EMMI images, and SEM images. These findings collectively indicate that the failure of the devices can be attributed to the elevated ESD current density, leading to the multiplication of DLDs from the edges of the device's oxide aperture. The formation of defects occurred suddenly in discrete events within localized regions and subsequently spread, ultimately causing damage throughout the entire active region. It is hypothesized that defects formed in the active layers propagate at an accelerated rate, contributing to a more rapid deterioration of the device.
Funding
Natural Science Foundation of Jilin Province (20230101352JC); The Fundament Application fundamentals of Guangzhou (202201011150); State Key Lab of Digital Manufacturing Equipment and Technology (JAS212800040); Guangzhou Municipal Science and Technology Project (2023A04J2041).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. M. Janczak, R. P. Sarzała, M. Dems, et al., “Threshold performance of pulse-operating quantum-cascade vertical-cavity surface-emitting lasers,” Opt. Express 30(25), 45054–11365 (2022). [CrossRef]
2. P. Fu, P. N. Ni, B. Wu, et al., “Metasurface Enabled On-Chip Generation and Manipulation of Vector Beams from Vertical Cavity Surface-Emitting Lasers,” Adv. Mater. 35(12), 2204286 (2023). [CrossRef]
3. G. Larisch, S. Tian, and D. Bimberg, “Optimization of VCSEL photon lifetime for minimum energy consumption at varying bit rates,” Opt. Express 28(13), 18931–18937 (2020). [CrossRef]
4. Y. Zhou, Y. Jia, X. Zhang, et al., “Large-aperture single-mode 795 nm VCSEL for chip-scale nuclear magnetic resonance gyroscope with an output power of 4.1 mW at 80 °C,” Opt. Express 30(6), 8991–8999 (2022). [CrossRef]
5. G. Pan, Y. Sun, M. Xun, et al., “Extrinsic Parasitics and Design Considerations on Modulation Bandwidth of 850-nm Vertical Cavity Surface Emitting Lasers,” IEEE Trans. Electron Devices 69(9), 4992–4997 (2022). [CrossRef]
6. L. Chorchos, N. N. Ledentsov, J. R. Kropp, et al., “Energy Efficient 850 nm VCSEL Based Optical Transmitter and Receiver Link Capable of 80 Gbit/s NRZ Multi-Mode Fiber Data Transmission,” J. Lightwave Technol. 38(7), 1747–1752 (2020). [CrossRef]
7. X. Zhang, G. Zhang, Y. Wen, et al., “50 Gbit/s PAM4 Data Transmission Over 500-m Single-Mode Fiber Using an 880-nm Single Mode VCSEL,” IEEE Photonics Technol. Lett. 34(22), 1210–1213 (2022). [CrossRef]
8. K. Li, X. Chen, J. E. Hurley, et al., “High data rate few-mode transmission over graded-index single-mode fiber using 850 nm single-mode VCSEL,” Opt. Express 27(15), 21395–21404 (2019). [CrossRef]
9. M. Wasiak, P. Śpiewak, N. Haghighi, et al., “Numerical model for small-signal modulation response in vertical-cavity surface-emitting lasers,” J. Phys. D: Appl. Phys. 53(34), 345101 (2020). [CrossRef]
10. A. Liu, P. Wolf, J. A. Lott, et al., “Vertical-cavity surface-emitting lasers for data communication and sensing,” Photonics Res. 7(2), 121–136 (2019). [CrossRef]
11. J. Krueger, R. Sabharwal, S. McHugo, et al., “Studies of ESD-related failure patterns of Agilent oxide VCSELs,” Proc. Vertical-Cavity Surface-Emitting Lasers VII. SPIE 4994, 162–172 (2003). [CrossRef]
12. H. Neitzert, A. Piccirillo, and B. Gobbi, “Sensitivity of proton implanted VCSELs to electrostatic discharge pulses,” IEEE J. Select. Topics Quantum Electron. 7(2), 231–241 (2001). [CrossRef]
13. H. Jia-Sheng, T. Olson, and E. Isip, “Human-Body-Model Electrostatic-Discharge and Electrical-Overstress Studies of Buried-Heterostructure Semiconductor Lasers,” IEEE Trans. Device Mater. Relib. 7(3), 453–461 (2007). [CrossRef]
14. T. Kim, T. Kim, and S. Kim, “Degradation behavior of 850 nm AlGaAs/GaAs oxide VCSELs suffered from electrostatic discharge,” ETRI journal 30(6), 833–843 (2008). [CrossRef]
15. H. Kim, W. Jeong, H. Shin, et al., “Reliability in the oxide vertical-cavity surface-emitting lasers exposed to electrostatic discharge,” Opt. Express 14(25), 12432–12438 (2006). [CrossRef]
16. H. Ichikawa, S. Matsukawa, K. Hamada, et al., “Failure Analysis of InP-Based Edge-Emitting Buried Heterostructure Laser Diodes Degraded by Forward-Biased Electrostatic Discharge Tests,” Jpn. J. Appl. Phys. 48(5), 052102 (2009). [CrossRef]
17. H. C. Neitzert, “ESD sensitivity of AlGaAs and InGaAsP based Fabry-Perot laser diodes,” Microelectron. Reliab. 50(9-11), 1563–1567 (2010). [CrossRef]
18. S.-H. Lee, S.-H. Pyun, H. D. Kim, et al., “Prediction and evaluation of infant failures of oxide VCSELs,” J. Korean Phys. Soc. 65(11), 1848–1852 (2014). [CrossRef]
19. X. Shan, B. Li, F. Zhao, et al., “Radiation effects of heavy ions on the static and dynamic characteristics of 850 nm high-speed vertical cavity surface emitting lasers,” J. Lumin. 237, 118136 (2021). [CrossRef]
20. A. Kalavagunta, C. Bo, M. A. Neifeld, et al., “Effects of 2 MeV proton irradiation on operating wavelength and leakage current of vertical cavity surface emitting lasers,” IEEE Trans. Nucl. Sci. 50(6), 1982–1990 (2003). [CrossRef]
21. J. Hu, L. Du, Y. Zhuang, et al., “Noise as a representation for reliability of light emitting diode,” Acta Phys. Sin. 55(3), 1384–1389 (2006).
22. C. Tzung-Te, F. Han-Kuei, D. Chun-Fan, et al., “Low-Frequency Noise Analysis of Electrostatic Discharge Tolerance of InGaN Light-Emitting Diodes,” IEEE Trans. Electron Devices 60(11), 3794–3798 (2013). [CrossRef]
23. R. Fabbro, T. Haber, G. Fasching, et al., “Defect localization in high-power vertical cavity surface emitting laser arrays by means of reverse biased emission microscopy,” Meas. Sci. Technol. 32(9), 095406 (2021). [CrossRef]
24. J. Zhang, W. Liao, X. Wang, et al., “Degradation Characteristics and Mechanism of High Speed 850 nm Vertical-Cavity Surface-Emitting Laser during Accelerated Aging,” Photonics 9(11), 801 (2022). [CrossRef]
25. A. W. Bushmaker, Z. Lingley, M. Brodie, et al., “Optical Beam Induced Current and Time Resolved Electro-Luminescence in Vertical Cavity Surface Emitting Lasers During Accelerated Aging,” IEEE Photonics J. 11(5), 1–11 (2019). [CrossRef]
26. M. Vanzi, G. Mura, G. Marcello, et al., “ESD tests on 850 nm GaAs-based VCSELs,” Microelectron. Reliab. 64(5), 617–622 (2016). [CrossRef]
