Open Access
Add to BibSonomy
Abstract
The fabrication processes of high-speed oxide-confined single-mode (SM)-vertical-cavity surface-emitting lasers (VCSELs) are complex, costly, and often held back by reliability and yield issues, which substantially set back the high-volume processing and mass commercialization of SM-VCSELs in datacom or other applications. In this article, we report the effects of Al2O3 passivation films deposited by atomic layer deposition (ALD) on the mesa sidewalls of high-speed 850-nm SM-VCSELs. The ALD-deposited film alleviates the trapping of carriers by sidewall defects and is an effective way to improve the performance of SM-VCSELs. The ALD-passivated SM-VCSELs showed statistically significant static performance improvements and reached a believed to be record-breaking SM-modulation bandwidth of 29.1 GHz. We also propose an improved microwave small-signal equivalent circuit model for SM-VCSELs that accounts for the losses attributed to the mesa sidewalls. These findings demonstrate that ALD passivation can mitigate processing-induced surface damage, enhance the performance of SM-VCSELs, and enable mass production of high-quality SM-VCSELs for mid- to long-reach optical interconnects.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
850-nm multi-mode (MM)-vertical-cavity surface-emitting lasers (VCSELs) are the critical constituents of optical interconnects (OIs) in modern networking systems [1,2], whereas the next-generation OIs require a principal emphasis on light sources with narrow spectral widths. Few-mode (FM)-VCSELs and single-mode (SM)-VCSELs offer improved beam profile qualities [3], and significantly superior performances in mid- to long-reach OIs [4–6]. Meanwhile, the adoption of advanced modulation formats formulates a new set of requirements for next-generation light sources. The IEEE Std 802.3bs-2017 standard set forth the 200G and 400G Ethernet standards by implementing a combination of pulse amplitude modulation (PAM)-4 encoding and multiple parallel lanes [7]. The PAM-N signaling put extra requirements on spectral rms bandwidths and linearity of the light sources. FM-VCSELs and SM-VCSELs have significantly narrower spectral linewidths and allow significantly enhanced signal transmission quality over long fibers due to reductions in modal and chromatic dispersions [2,6].
High-speed SM-VCSELs have the innate characteristics of single wavelength emission, small spectral rms bandwidth, and have great significance for future high-speed OI applications. And are crucially important for mid- to long-reach OIs in MMF and SM fiber (SMF) links because of their minimal differential mode delays, reduced dispersion effects, and power losses for longer distance transmission [2,5,6]. However, the fabrication processes of high-speed oxide-confined SM-VCSELs are complex, costly, and often held back by reliability and yield issues, which substantially set back the high-volume processing and mass commercialization of SM-VCSELs in datacom or other applications.
Many techniques have been introduced for the realization of SM-VCSELs including oxide aperture [8,9], proton implantation [10], surface relief [11], photonic crystal [12,13], impurity-induced disordering [14–16], high-index contrast grating [17], metal ring aperture [18,19], anti-phase coatings [16] or integrated mode-selective filter [20–22]. Nevertheless, the latter techniques require complex fabrication processes that are not commercially feasible for high-volume manufacturing or are unfavorable for high-speed applications. Oxide-confined SM-VCSELs have been previously demonstrated to support long-distance and high-speed optical fiber transmissions [6,23]. However, VCSELs with small oxide aperture diameters typically suffer from higher differential resistances and thermal impedances [24], which lead to increased junction temperatures due to severe Joule heating. This causes redshifts of the quantum well (QW) gain peaks, increased threshold currents (Ith), and reduced rollover currents (Iro), fundamentally limiting the devices’ performance at higher current densities and temperatures. Meanwhile, defects on mesa sidewalls may cause surface recombinations and undesirable current leakage paths that deteriorate device performance, increase Ith, reduce modulation bandwidth, and create premature failures and reliability issues [25,26].
VCSELs have considerably high perimeter-to-area ratios because of their multiple distributed Bragg reflector (DBR) pairs and mesa/pillar structures. The high perimeter-to-area ratios may lead to increased surface recombination rates at the mesa sidewalls. Studies have revealed that significant temperature-dependent carrier leakage in the active region and non-radiative recombination losses are much higher in smaller aperture VCSEL devices [27–29]. Additionally, the oxidized regions are susceptible to moisture in the air, which may result in reliability issues such as semiconductor cracking, dark line defects (DLDs) and aperture surface degradation over prolonged operation [30]. Sidewall passivation processes with dielectric films may effectively alleviate these issues and improve the performance of VCSELs [31–33]. Several sidewall passivation techniques for VCSELs have been explored using SiO2 [relative permittivity ($\varepsilon _\mathrm {r}$) = 3.9], SiNx ($\varepsilon _\mathrm {r}$ = 7.5) [31], polyimide ($\varepsilon _\mathrm {r}$ = 3.4) [32,34], SiON ($\varepsilon _\mathrm {r}$ = 7.9) [35], and benzocyclobutene (BCB) ($\varepsilon _\mathrm {r}$ = 2.65) [33,34,36]. Low $\varepsilon _\mathrm {r}$ materials such as BCB are commonly used as passivation films because they are capable of minimizing parasitic effects in VCSELs [33,36].
Due to various challenges in the growth of III-V materials, many optoelectronic devices suffer from interface defect density, current leakage and instability issues. Various surface treatments have shown superior results in suppressing Fermi-level pinnings, and current leakages [37]. The smaller device footprints of the SM-VCSELs have particularly high perimeter-to-area ratios that are especially hampered by these surface non-radiative recombination losses and reliability issues. These have been considered as agonizing issues faced by the VCSEL industry. There are two main methods employed by the industry for the past several decades to confine the current flows and transverse modes: ion implantation and oxidation [10,38,39]. The oxidation approach requires plasma etching to expose the oxidizable high-Al layers. Therefore, the mesa sidewalls could be constantly exposed to moisture and other particles if not adequately passivated. The sidewall passivations could block moisture from reaching the oxide layers and active regions, inhibiting the formation of DLDs and prolonging the devices’ lifetime [30].
Atomic layer deposition (ALD) is an advanced thin-film deposition process that offers precise coating thickness control, good step coverage, low defect densities, and high-quality layer deposition on structures with high aspect ratios [40,41]. ALD-grown passivation films (e.g., Al2O3, SiNx) have also been demonstrated to rectify these issues and enhance the electrical and optical performance of many optoelectronic devices [37,40,42–44]. The investigation of applying the ALD-grown sidewall passivation films on VCSELs was not conducted until 2021 [41]. The ALD passivation process is highly suitable for defect passivation on the mesa sidewalls [40,44,45], and researchers have reported the applications of ALD-grown passivation films on various optoelectronic devices, such as light-emitting diodes [41,44–47], solar cells [48], photodetectors [49], phototransistors [50], and lasers [41,51]. ALD-deposited Al2O3 films have the advantages of excellent adhesions and good isolations. And most importantly, they are highly compatible with the oxidized edges of the high-Al DBR layers on the sidewalls and have excellent film qualities as deposited [52].
This article is the first to report the application and effects of ALD-grown Al2O3 sidewall passivation treatment on SM-VCSELs (ALD-VCSELs). Using static characteristics, we statistically compare their characteristics against conventional oxide-confined SM-VCSELs (C-VCSELs). Furthermore, we conduct an in-depth analysis of the static and microwave performance of the C-VCSELs and ALD-VCSELs and use microwave small-signal modeling to study the enhancement effects of the ALD-grown Al2O3 sidewall passivation films on SM-VCSELs. Lastly, we propose an improved microwave small-signal equivalent circuit model for SM-VCSELs that accounts for the losses attributed to the mesa sidewalls. The results demonstrate that the ALD-passivation process could mitigate various sidewall damage issues, particularly hampering SM-VCSELs, and enable mass production of high-quality SM-VCSELs for mid- to long-reach OIs.
2. Method
The VCSELs investigated in this work had an epitaxial layer structure design for modulation bandwidth in the range of 25–30 GHz and were optimized for next-generation short-reach 850-nm optical interconnects. The bottom and top DBR mirrors comprised 33 and 20 periods, respectively, and the active region contained strained InGaAs QWs for increased differential gain [53]. The top DBR mirror possessed a multioxide layer design composed of Al0.98Ga0.02As and Al0.96Ga0.04As layers to reduce oxide capacitance [2,54,55]. This was followed by the growth of a highly doped p-type GaAs cap layer for p-contact metallization. The VCSEL microcavities were additionally designed for optimal optical confinement, reduced threshold gain, and narrow linewidth [56,57].
The cross-sectional schematic of the ALD-VCSELs is depicted in Fig. 1. We deposited 10 nm of Al2O3 on the mesa sidewalls of the ALD-VCSELs, whereas all other fabrication steps were kept unchanged from the C-VCSELs [36]. The fabrication process began with the deposition of p-type Ti/Pt/Au contacts, followed by mesa etching with a SAMCO RIE-101iPH inductively coupled plasma system for the formation of mesa structures that were 14 µm in diameter. The devices were then placed in a tube furnace with flowing gaseous streams of N2 and H2O for selective oxidation of the high-Al layers to form 2.5-µm oxide apertures [38]. Next, AuGe/Ni/Au layers were evaporated and annealed to form n-type ohmic contacts. The sample was then diced in half, and one-half underwent the Al2O3 ALD passivation process, while the other half did not. Approximately 10 nm of Al2O3 (100 cycles) was deposited on the ALD-VCSELs using a PICOSUN R-200 ALD system before spin coating with BCB resin. The subsequent steps included via-hole etching and the deposition of Ti/Au interconnects and coplanar waveguides for microwave probing.
3. Results
The VCSEL devices were placed on a temperature-controlled probe station and analyzed with a ground-signal probe. Microwave signals were generated from a network analyzer and a bit pattern generator (BPG), and DC signals were generated from a source meter. Both were provided through a bias tee to the VCSELs for static and microwave characterization. A Thorlabs FDS1010 Si photodiode was used to quantify the total optical output power of the devices, and the optical outputs were coupled with a lensed MMF for the optical spectra, microwave, and bit-error-rate (BER) measurements.
3.1 Static characteristics
The light-current-voltage (L-I-V) of the C-VCSEL and ALD-VCSEL chips used in the following characterization are illustrated in Fig. 2. Their similar Iro indicates that they have almost identical oxide aperture sizes [58]. While both devices exhibit similar Iro, the ALD treatment reduced Iro from 0.79 to 0.68 mA. The output light of the VCSELs was coupled into an MMF and fed into an Advantest Q8384 optical spectrum analyzer for spectral measurements. Both C-VCSEL and ALD-VCSEL exhibit highly SM spectra as presented in Fig. 3.
Fig. 2. L-I-V curves of select C-VCSEL (green lines) and ALD-VCSEL (red lines) devices. The solid lines denote the L-I curves, and the dash-dotted lines denote the V-I curves.
3.2 Statistical analysis
Yield, consistency, and reliability are what trouble foundries in the development of new processes and devices. The shrinkage of the devices for transverse mode control would introduce device uniformity issues which impact yields and profit margins. Besides, smaller devices would substantially increase current densities and put heavy burdens on the devices’ lifetimes. SM-VCSELs are usually having yield and reliability issues that prohibit the mass adoption of these devices in datacom applications. In this section, we are trying to discuss the changes in the devices’ static characteristics and the improvements in the ALD-passivated SM-VCSELs with statistical evidence.
We compared two groups (80 chips total) of the C-VCSELs and ALD-VCSELs with identical mesa and oxide aperture sizes. The box plot distributions of Ith and Iro for both groups are illustrated in Fig. 4. We selected one device from each group that had static characteristics (Ith and Iro) that were closest to their respective means to allow for a fair analysis of the effects of the ALD passivation process on SM-VCSELs.
Figure 4(a) shows that both groups have similar Iro indicating that the devices have almost identical oxide aperture sizes as previously mentioned. Furthermore, it rules out that the enhancements in the high-speed performance of ALD-VCSELs are not potentially the results of smaller laser microcavities. At the same time, the reduction in Ith indicates the enhancement effects of the ALD-grown Al2O3 passivation process on SM-VCSELs.
As indicated in the box plot in Fig. 4(b), the ALD-VCSELs had a mean Ith of 0.65 mA, which is on average 18% lower than C-VCSELs (0.79 mA). The lower quartile, the median, and the higher quartile show an average of 14% decrease from 0.7 to 0.6 mA, an average of 12% decrease from 0.8 to 0.7 mA, and an average of 22% decrease from 0.9 to 0.7 mA, respectively. The interquartile range of Ith decreased from 0.2 to 0.1 mA after the ALD passivation process. Hence, ALD-VCSELs display a statistically significant reduction in Ith, indicating that the ALD Al2O3 passivation process may have effectively reduced the non-radiative recombination rate and resulted in lower lasing thresholds.
The statistical analysis demonstrates that the ALD-passivation process mitigates some of the processing-induced surface damages in the fabrication processes and gives an overall uplift to the ALD-VCSEL group of devices. In the next subsection, we shall discuss the changes in microwave performances with the ALD-passivation film.
3.3 Microwave characteristics
To understand the effects of the thin ALD-grown Al2O3 passivation film on high-speed VCSELs, we measured and investigated the microwave characteristics of the C-VCSEL and ALD-VCSEL. Commonly used techniques to improve the VCSELs’ electrical-to-optical microwave responses, $S_{21,VCSEL} {(f)}$, include reducing device footprints [2,58] or lowering the pad parasitics [2,32,33]. Although a smaller aperture size can provide highly SM emission, low threshold current density, larger modulation current efficiency factor [59], and larger modulation bandwidth potentials. It also introduces issues such as Joule heating, current crowding, and thermal lensing [3,60]. Moreover, high perimeter-to-area ratios and smaller oxide apertures contribute to increased surface recombination losses and negatively impact device performance [28,29].
The microwave characteristics of our devices were measured using a 50-GHz Keysight N5225A vector network analyzer. A lensed MMF coupled the optical outputs into a 30-GHz Thorlabs DXM30BF photodetector and sent them to the vector network analyzer as electrical signals. Because the microwave response of the photodetector, $H_{PD}{(f)}$, has a limited 3-dB bandwidth of 30 GHz, the photodetector response was de-embedded from the measured microwave response, $S_{21,measured}{(f)}$, to extract the correct $S_{21,VCSEL} {(f)}$ as described in Eq. (1):
The de-embedded $S_{21,VCSEL} {(f)}$ is a three-pole transfer function that comprises two superimposed microwave responses: the two-pole laser intrinsic optical transfer function, $S_{21,int} {(f)}$, and the single-pole electrical parasitic transfer function, $H_{par} {(f)}$.
Figure 5 shows the $S_{21,VCSEL} {(f)}$ of the C-VCSEL and ALD-VCSEL when biased at 2, 3, 4, and 5 mA. While they have similar bandwidths at lower drive currents of 2–3 mA, the ALD-VCSEL experience smaller magnitude drops at higher frequencies from 4 mA onwards and thus exhibit distinguishably faster modulation responses over the C-VCSEL at these bias conditions. Therefore the maximum optical microwave response ($f_{3dB,max}$) of the ALD-VCSEL increases from 27.7 to 29.1 GHz. To the best of our knowledge, the 29.1-GHz SM bandwidth of the ALD-VCSEL reported in this article is the fastest optical modulation bandwidth for 850-nm oxide-confined SM-VCSELs to date [61,62]. The previous record was the 26-GHz SM-VCSEL reported by Ledentsov et al. in 2019 [61].
A dual-channel 56-Gb/s SHF 12103A BPG and an SHF 11104A error analyzer installed in an SHF 10001A bit-error-rate tester (BERT) mainframe unit were used to evaluate the BER performance of the optical links. The C-VCSEL and ALD-VCSEL were modulated by 27-1 pseudo-random binary sequence (PRBS) non-return-to-zero (NRZ) signals generated by the BPG, and their optical outputs were collected by a 22-GHz New Focus 1484-A-50 photoreceiver. The converted electrical signals were then sent to the BERT through a continuously variable neutral density filter to test the BER performance of the PRBS NRZ signals at different received optical levels. A 70-GHz Agilent 86118A module installed on an Agilent Infiniium DCA-J 86100C sampling scope was used for eye-diagram characterization.
The BER measurements of the C-VCSEL and ALD-VCSEL were performed at 4 mA while varying the optical powers. The highest error-free (BER $< 10^{-12}$) data transmission rate attained by the ALD-VCSEL surpassed that of the C-VCSEL and reached 48 Gb/s. Figure 6(a) exhibits the BER testing results, and Fig. 6(b) and Fig. 6(c) show the measured eye diagrams of the C-VCSEL and ALD-VCSEL at their maximum error-free data rates (C-VCSEL: 44 Gb/s; ALD-VCSEL: 48 Gb/s).
Fig. 6. (a) Measured BER versus received optical power of the C-VCSEL and ALD-VCSEL. The eye diagrams were captured at the highest error-free PRBS NRZ data rates for (b) C-VCSEL (44 Gb/s) and (c) ALD-VCSEL (48 Gb/s).
4. Discussion
4.1 Effects of the ALD-grown sidewall passivation films
The 10-nm ALD-grown Al2O3 films with excellent sidewall coverage encapsulate the exposed oxidized high-Al layers and active regions after the plasma etching processes. ALD-grown Al2O3 films are typically high-quality and have low defect and pinhole densities [40,43], and they possess excellent thermal conductivities (Al2O3: 35 W/mK; BCB: 0.29 W/mK) [36,63]. Many have reported that ALD-grown passivation films may repair sidewall damages, reduce surface recombination, and essentially lower non-radiative recombination rates in optoelectronic devices [64].
Like other light-emitting devices, two major types of recombinations occur in the active region of VCSELs: radiative and non-radiative recombinations. Unlike radiative recombination, which converts a large proportion of the injected carriers into photons, defects or dislocations in the crystal structure and the semiconductor surfaces lead to non-radiative recombination and heating of the device internally and at the surface. These unwanted non-radiative recombinations may increase lasing threshold and Ith [28,65]. The reduced Ith for the ALD-VCSELs suggests reduced non-radiative recombination rates in the ALD-passivated devices. After the simulated emissions kick in, the optical emission would increase significantly. Therefore, the enhancement on optical powers provided by the ALD-passivated sidewalls would not be that prominent, resulting in comparable L-I curves between the two devices.
For a given bias current on lasers, smaller Ith, shrunken optical volumes, and larger differential gains increase the resonance frequencies at given currents [66]. And allow higher modulation bandwidths and data rates. Therefore, as shown in Fig. 2 and Fig. 4, the decreases in Ith suggest ALD-VCSELs may offer improved high-speed modulation bandwidths and bit rates. $S_{21,VCSEL} {(f)}$ bandwidths of both devices are nearly almost identical at lower bias currents (2–3 mA), but at higher bias currents (4–5 mA), the ALD-VCSEL had shown noticeable performance edges over the C-VCSEL. These support our conjecture that the parasitic effects caused by surface recombination losses and current leakages in C-VCSELs may dampen the microwave responses at higher frequencies.
The intrinsic modulation bandwidths of lasers are majorly limited by the photons in the optical cavities, including photon cavity lifetime, internal quantum efficiency and etc. On the other hand, the RC parasitics of the defects on the sidewall could cause early RC rolloffs, another main factor that may cause reductions in the overall optical modulation bandwidths. The ALD-deposited film can reduce the non-radiative recombination losses and improve the efficiencies of lasers. Additionally, this also reduces the RC parasitic component in the small-signal model that limits the frequency responses of VCSELs. Therefore, we firstly introduced a modified small-signal model to model the sidewall effects in the small-signal perspective. In the following subsections, we are going to deduce the parasitic sidewall components that account for the losses attributed to the mesa sidewalls and investigate the microwave enhancement effect of the ALD-grown film from a small-signal analysis perspective.
4.2 Proposed small-signal equivalent circuit model considering the sidewall losses
To investigate the microwave responses of the VCSELs, we built a small-signal equivalent circuit model for SM-VCSELs, as illustrated in Fig. 7, and extracted their small-signal equivalent circuit parameters [67–69]. The circuit elements are labeled on the schematic cross-sectional view. Here, we propose a modified small-signal model with additional parasitics ($R_{\textrm{w}}$ and $C_{\textrm{w}}$) to model the effect of the time constant attributed to the surface recombination. Other parasitic parameters include $C_{\textrm{p}}$ representing the capacitance between the contacts, and $R_{\textrm{p}}$ representing the impedance loss. They are heavily influenced by the layout design and can be regarded as non-current dependent variables. $R_{\textrm{m}}$ models the contact resistance and the mirror resistance of the top and bottom DBR mirrors. The mesa capacitance, $C_{\textrm{m}}$, is the combined capacitance of the oxide capacitance, $C_{\textrm{ox}}$, and the intrinsic region capacitance, $C_{\textrm{int}}$. $R_{\textrm{m}}$ and $C_{\textrm{m}}$ are majorly influenced by injected currents. The abovementioned parasitic elements cause parasitic RC roll-offs and lessen the microwave responses at higher frequencies. The active region parameters include the junction resistance, $R_{\textrm{j}}$, and the junction capacitance, $C_{\textrm{j}}$.
Fig. 7. Schematic cross-sectional view of a VCSEL with its small-signal equivalent circuit components labeled (not-to-scale drawing).
The values of $C_{\textrm{p}}$ are usually on the order of fF and $R_{\textrm{p}}$ are usually within dozens of $\Omega$. They are heavily influenced by $\varepsilon _\mathrm {r}$ and the dimensions of the insulating passivation films. They exhibited almost no changes at different current injection levels and are therefore minimally current-dependent. The mesa capacitance, $C_{\textrm{m}}$, is also a non-current-dependent component that relates to the dimensions of the mesas and the oxidized regions.
4.3 Extraction of the small-signal model parameters
Compared to larger-size VCSELs, the smaller mesa diameters and oxide apertures of SM-VCSELs have much higher perimeter-to-area ratios, introducing non-negligible sidewall losses for the input signals to the active region of the SM-VCSELs. This article proposes two additional RC components, $R_{\textrm{w}}$ and $C_{\textrm{w}}$, to the SM-VCSEL small-signal equivalent circuit to model the microwave losses due to the non-radiative recombinations or current leakages. We envision that the ALD-grown films applied to ALD-VCSELs could provide a better field insulation layer and significantly passivate the undesired surface recombination centers to reduce non-radiative recombination losses.
With the equivalent circuit model shown in Fig. 7, a data fitting between the experimental and small-signal models of the C-VCSEL and ALD-VCSEL is performed. The reflection S11,VCSEL microwave responses of the C-VCSEL and ALD-VCSEL are characterized at 2, 3, 4, and 5 mA. The characterization and fitting plots are illustrated using Smith Charts in Fig. 8. The calculated and extracted small-signal equivalent circuit parameters of C-VCSEL and ALD-VCSEL are shown in Table 1.
Fig. 8. S11,VCSEL reflection microwave responses of the (a) C-VCSEL and (b) ALD-VCSEL at 2, 3, 4, and 5 mA. The solid lines denote the measured microwave responses, and the dotted lines denote the microwave small-signal equivalent circuit model fitting curves.
Table 1. The extracted microwave small-signal equivalent circuit parameters of C-VCSEL and ALD-VCSEL
According to the fitting results, both the devices share identical Cp, Rp, Cm and Cj values. The parameters that differ ALD-VCSEL from C-VCSEL are the parasitic elements that correlate to the sidewall losses, Cw and Rw. The ALD process decreases Cw from 90 to 10 fF and increases Rw from 0.2 to 0.8 M$\Omega$. These results demonstrated that the ALD-grown passivation films lower Cw and increase Rw. A more prominent capacitive element Cw or a smaller resistive element Rw may limit bandwidths as it shunts a fraction of the input signals and reduces the current injected into the active region.
The extracted sidewall parasitic components, Cw and Rw, are 90 fF and 0.2 M$\Omega$ for the C-VCSEL and 10 fF and 0.8 M$\Omega$ for ALD-VCSEL. A quick simulation with the circuit model with the extracted Cw and Rw sees that these sidewall effects become the bandwidth bottleneck at the microwave responses from 23 GHz onwards. Therefore, the ALD-passivated sidewalls alleviate these issues and enhance the modulation bandwidths and the data rates.
In this way, the ALD-grown Al2O3 passivation films mitigate the non-radiative recombination losses, extend the RC parasitic roll-off frequencies, and ultimately increase the modulation bandwidths of ALD-VCSELs. Therefore, the sidewall parasitic RC components, Cw and Rw, essentially act as an extra one-pole low-pass filter that cuts off the higher frequency responses of VCSELs, and limits their modulation bandwidths and bit rates.
5. Conclusion
This article demonstrates that applying ALD passivation on the mesa sidewalls of SM-VCSELs enhances the device characteristics and thoroughly compares the static, microwave, and signal transmission characteristics of the C-VCSELs and ALD-VCSELs. The experimental results show that the ALD-VCSELs have a statistically significant lower average Ith than the C-VCSELs, which we attribute to the ALD passivation significantly reducing the non-radiative recombination losses and shunt leakages on the mesa sidewalls. A modified microwave small-signal equivalent circuit model is firstly proposed to explain the effect of microwave losses due to surface recombination. The microwave performance $f_{3dB,max}$ of the ALD-VCSEL improves from 27.7 to 29.1 GHz, and the maximum error-free PRBS NRZ modulation bandwidth improves from 44 to 48 Gb/s. These findings demonstrate that the ALD-grown Al2O3 passivation films enhance the static and microwave characteristics of oxide-confined SM-VCSELs. These results suggest that the use of ALD passivation is desirable for high-performance and robust long-reach SM OIs based on SM-VCSELs with enhanced bandwidth and data rates. Lastly, the ALD passivation process has demonstrated a solution to the long-standing issues concerning the mass production of reliable, high-performance SM-VCSELs that may enable the commercial mass adoption of SM-VCSEL-based OIs in mid- to long-reach datacom applications.
Funding
Ministry of Science and Technology, Taiwan (110-2224-E-992-001, 110-2622-8-002-021, 110-2622-8-A49-008-SB, 111-2119-M-002-008, 111-2119-M-002-009, 111-2221-E-002-051-MY3, 111-2622-8-002-001); National Taiwan University (109L7819, 110L2033-49); NTUS Innovation Cooperation (11112071002).
Acknowledgments
The authors appreciate Ting-Yu Huang, Yu-Jie Lo, Yun-Cheng Yang, Guei-Ting Hsu and Kevin Tsao for their constructive criticisms of the manuscript and experimental assistance. We thank the anonymous reviewers for their careful reading of our manuscript and their many insightful comments and suggestions. The authors also value the support from all members of our laboratories, the Integrated Optoelectronics Device (IOED) Group at National Taiwan University and the High-Speed Integrated Circuits (HSIC) Group at the University of Illinois at Urbana-Champaign, who contributed to our researches on high-speed VCSELs and other optoelectronic devices. We appreciate Ms. C.-Y. Chien and Ms. S.-J. Ji of NSTC (National Taiwan University) for assistance in FIB and SEM experiments. We are also appreciative of the Semiconductor Manufacturing Lab of the Consortia of Key Technologies, and the Nano-Electro-Mechanical-System Research Center, National Taiwan University for experimental support.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this article may be obtained from the authors upon reasonable request.
References
1. A. Liu, P. Wolf, J. A. Lott, and D. Bimberg, “Vertical-cavity surface-emitting lasers for data communication and sensing,” Photonics Res. 7(2), 121–136 (2019). [CrossRef]
2. H.-T. Cheng, Y.-C. Yang, T.-H. Liu, and C.-H. Wu, “Recent advances in 850 nm VCSELs for high-speed interconnects,” Photonics 9(2), 107 (2022). [CrossRef]
3. C.-Y. Peng, K. Tsao, H.-T. Cheng, M. Feng, and C.-H. Wu, “Investigation of the current influence on near-field and far-field beam patterns for an oxide-confined vertical-cavity surface-emitting laser,” Opt. Express 28(21), 30748–30759 (2020). [CrossRef]
4. T.-Y. Huang, J. Qiu, C.-H. Wu, H.-T. Cheng, M. Feng, H.-C. Kuo, and C.-H. Wu, “A NRZ-OOK modulated 850-nm VCSEL with 54 Gb/s error-free data transmission,” in Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference (Optica Publishing Group, 2019), paper cb_p_22.
5. H.-Y. Kao, C.-T. Tsai, S.-F. Leong, C.-Y. Peng, Y.-C. Chi, H.-Y. Wang, H.-C. Kuo, C.-H. Wu, W.-H. Cheng, and G.-R. Lin, “Single-mode VCSEL for pre-emphasis PAM-4 transmission up to 64 Gbit/s over 100-300 m in OM4 MMF,” Photonics Res. 6(7), 666–673 (2018). [CrossRef]
6. M.-J. Li, K. Li, X. Chen, S. K. Mishra, A. A. Juarez, J. E. Hurley, J. S. Stone, C.-H. Wang, H.-T. Cheng, C.-H. Wu, H.-C. Kuo, C.-T. Tsai, and G.-R. Lin, “Single-mode VCSEL transmission for short reach communications,” J. Lightwave Technol. 39(4), 868–880 (2021). [CrossRef]
7. Institute of Electrical and Electronics Engineers, “IEEE Standard for Ethernet - Amendment 10: Media Access Control Parameters, Physical Layers, and Management Parameters for 200 Gb/s and 400 Gb/s Operation,” IEEE Std 802.3bs-2017 (Amendment to IEEE 802.3-2015 as amended by IEEE’s 802.3bw-2015, 802.3by-2016, 802.3bq-2016, 802.3bp-2016, 802.3br-2016, 802.3bn-2016, 802.3bz-2016, 802.3bu-2016, 802.3bv-2017, and IEEE 802.3-2015/Cor1-2017) pp. 1–372 (2017).
8. C. Jung, R. Jager, M. Grabherr, P. Schnitzer, R. Michalzik, B. Weigl, S. Muller, and K. J. Ebeling, “4.8 mW singlemode oxide confined top-surface emitting vertical-cavity laser diodes,” Electron. Lett. 33(21), 1790–1791 (1997). [CrossRef]
9. J. M. Dallesasse and N. Holonyak, “Oxidation of Al-bearing III-V materials: A review of key progress,” J. Appl. Phys. 113(5), 051101 (2013). [CrossRef]
10. M. Orenstein, N. G. Stoffel, A. C. Von Lehmen, J. P. Harbison, and L. T. Florez, “Efficient continuous wave operation of vertical-cavity semiconductor lasers using buried-compensation layers to optimize current flow,” Appl. Phys. Lett. 59(1), 31–33 (1991). [CrossRef]
11. H. Unold, M. Grabherr, F. Eberhard, F. Mederer, R. Jager, M. Riedl, and K. J. Ebeling, “Increased-area oxidised single-fundamental mode VCSEL with self-aligned shallow etched surface relief,” Electron. Lett. 35(16), 1340–1341 (1999). [CrossRef]
12. D.-S. Song, S.-H. Kim, H.-G. Park, C.-K. Kim, and Y.-H. Lee, “Single-fundamental-mode photonic-crystal vertical-cavity surface-emitting lasers,” Appl. Phys. Lett. 80(21), 3901–3903 (2002). [CrossRef]
13. C.-Y. Peng, H.-T. Cheng, Y.-H. Hong, W.-C. Hsu, F.-H. Hsiao, T.-C. Lu, S.-W. Chang, S.-C. Chen, C.-H. Wu, and H.-C. Kuo, “Performance analyses of photonic-crystal surface-emitting laser: Toward high-speed optical communication,” Nanoscale Res. Lett. 17(1), 90 (2022). [CrossRef]
14. Y. Yang, T. Dziura, S. Wang, G. Du, and S. Wang, “Single-mode operation of mushroom structure surface emitting lasers,” IEEE Photonics Technol. Lett. 3(1), 9–11 (1991). [CrossRef]
15. C.-Y. Peng, H.-T. Cheng, H.-C. Kuo, and C.-H. Wu, “Design and optimization of VCSELs for up to 40-Gb/s error-free transmission through impurity-induced disordering,” IEEE Trans. Electron Devices 67(3), 1041–1046 (2020). [CrossRef]
16. J. M. Dallesasse, P. Su, K. P. Pikul, L. Espenhahn, and M. Kraman, “(Invited) Achieving high-power single-mode operation in vertical-cavity surface-emitting lasers via scalable, higher-order mode suppression techniques,” ECS Trans. 109(5), 15–26 (2022). [CrossRef]
17. K.-B. Hong, T.-C. Chang, F. Hjort, N. Lindvall, W.-H. Hsieh, W.-H. Huang, P.-H. Tsai, T. Czyszanowski, Å. Haglund, and T.-C. Lu, “Monolithic high-index contrast grating mirror for a GaN-based vertical-cavity surface-emitting laser,” Photonics Res. 9(11), 2214–2221 (2021). [CrossRef]
18. S.-Y. Min, H.-T. Cheng, J.-S. Pan, W.-H. Lin, and C.-H. Wu, “Oxide-confined VCSEL with metal apertures for high-speed 850nm transmission,” in Opto-Electronics and Communications Conference (2020), paper T4-1.2.
19. C.-H. Wu, S.-Y. Min, and H.-T. Cheng, “Vertical-cavity surface emitting laser for emitting a single mode laser beam,” (2022), US Patent App. 16/932, 839.
20. M. Feng and X. Yu, “Single mode VCSELs with low threshold and high-speed operation,” (2020), US Patent App. 16/600, 030.
21. H.-L. Wang, J. Qiu, X. Yu, W. Fu, and M. Feng, “The modal effect of VCSELs on transmitting data rate over distance in OM4 fiber,” IEEE J. Quantum Electron. 56(6), 8000106 (2020). [CrossRef]
22. J. Qiu, D. Wu, H.-L. Wang, M. Feng, and X. Yu, “Advanced single-mode 850 nm VCSELs for record NRZ and PAM4 data rate on SMF-28 fiber up to 1 km,” in Optical Fiber Communication Conference (Optical Society of America, 2021), paper Tu5C.2.
23. J. Qiu, X. Yu, and M. Feng, “85°C operation of single-mode 850 nm VCSELs for high speed error-free transmission up to 1 km in OM4 fiber,” in Optical Fiber Communications Conference and Exhibition (2019), paper W3A.4.
24. B. Hawkins, R. Hawthorne, J. Guenter, J. Tatum, and J. Biard, “Reliability of various size oxide aperture VCSELs,” in 52nd Electronic Components and Technology Conference (Cat. No.02CH37345) (IEEE, 2002), pp. 540–550.
25. S. A. McHugo, A. Krishnan, J. J. Krueger, Y. Luo, N. Tan, T. Osentowski, S. Xie, M. S. Mayonte, R. W. Herrick, and Q. Deng, “Characterization of failure mechanisms for oxide VCSELs,” Proc. SPIE 4994, 55–66 (2003). [CrossRef]
26. R. W. Herrick, Reliability and Degradation of Vertical-Cavity Surface-Emitting Lasers (Springer, 2013), pp. 147–205.
27. J. Yan, J. Wang, C. Tang, X. Liu, G. Zhang, and Y. He, “An electrooptothermal-coupled circuit-level model for VCSELs under pulsed condition,” IEEE Trans. Ind. Electron. 66(2), 1315–1324 (2019). [CrossRef]
28. J. Shin and Y. H. Lee, “Determination of nonradiative recombination coefficients of vertical-cavity surface-emitting lasers from lateral spontaneous emission,” Appl. Phys. Lett. 67(3), 314–316 (1995). [CrossRef]
29. J. H. Shin, H. E. Shin, and Y. H. Lee, “Effect of carrier diffusion in oxidized vertical-cavity surface-emitting lasers determined from lateral spontaneous emission,” Appl. Phys. Lett. 70(20), 2652–2654 (1997). [CrossRef]
30. Y.-q. Zhang, Z.-y. Zuo, Q. Kan, and J. Zhao, “Common failure modes and mechanisms in oxide vertical cavity surface emitting lasers,” Chin. Opt. 15(2), 187–209 (2022). [CrossRef]
31. D. Young, A. Kapila, J. Scott, V. Malhotra, and L. Coldren, “Reduced threshold vertical-cavity surface-emitting lasers,” Electron. Lett. 30(3), 233–235 (1994). [CrossRef]
32. K. L. Lear, V. Hietala, H. Hou, J. Banas, B. E. Hammons, J. Zolper, and S. Kilcoyne, “Small and large signal modulation of 850 nm oxide-confined verticai-cavity surface-emitting lasers,” CLEO ’97., Summ. Pap. Present. at Conf. on Lasers Electro-Optics 11, 193–194 (1997). [CrossRef]
33. T. Tanigawa, T. Onishi, S. Nagai, and T. Ueda, “High-speed 850nm AlGaAs/GaAs Vertical Cavity Surface Emitting Laser with low parasitic capacitance fabricated using BCB planarization technique,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science and Photonic Applications Systems Technologies (Optical Society of America, 2005), paper CWI3.
34. A. Al-Omari and K. Lear, “Dielectric characteristics of spin-coated dielectric films using on-wafer parallel-plate capacitors at microwave frequencies,” IEEE Trans. Dielectr. Electr. Insul. 12, 1151–1161 (2005). [CrossRef]
35. J. S. Pan, C. J. Wu, I. H. Wu, and K. F. Tseng, “Method for fabricating oxide-confined vertical-cavity surface-emitting laser,” (2014), US Patent 8, 679, 873.
36. The Dow Chemical Company, “Advanced packaging polymers product selection guide,” (2014).
37. L. Zhou, B. Bo, X. Yan, C. Wang, Y. Chi, and X. Yang, “Brief review of surface passivation on III-V semiconductor,” Crystals 8(5), 226 (2018). [CrossRef]
38. J. M. Dallesasse, N. Holonyak, A. R. Sugg, T. A. Richard, and N. El-Zein, “Hydrolyzation oxidation of AlxGa1-xAs-AlAs-GaAs quantum well heterostructures and superlattices,” Appl. Phys. Lett. 57(26), 2844–2846 (1990). [CrossRef]
39. D. L. Huffaker, D. G. Deppe, K. Kumar, and T. J. Rogers, “Native-oxide defined ring contact for low threshold vertical-cavity lasers,” Appl. Phys. Lett. 65(1), 97–99 (1994). [CrossRef]
40. S. M. George, “Atomic layer deposition: An overview,” Chem. Rev. 110(1), 111–131 (2010). [CrossRef]
41. Y.-W. Yeh, S.-H. Lin, T.-C. Hsu, S. Lai, P.-T. Lee, S.-Y. Lien, D.-S. Wuu, G. Li, Z. Chen, T. Wu, and H.-C. Kuo, “Advanced atomic layer deposition technologies for Micro-LEDs and VCSELs,” Nanoscale Res. Lett. 16(1), 164 (2021). [CrossRef]
42. M. L. Huang, Y. C. Chang, C. H. Chang, Y. J. Lee, P. Chang, J. Kwo, T. B. Wu, and M. Hong, “Surface passivation of III-V compound semiconductors using atomic-layer-deposition-grown Al2O3,” Appl. Phys. Lett. 87(25), 252104 (2005). [CrossRef]
43. X. Luo, Y. Rahbarihagh, J. C. M. Hwang, H. Liu, Y. Du, and P. D. Ye, “Temporal and thermal stability of Al2O3-passivated phosphorene MOSFETs,” IEEE Electron Device Lett. 35(12), 1314–1316 (2014). [CrossRef]
44. M. S. Wong, J. A. Kearns, C. Lee, J. M. Smith, C. Lynsky, G. Lheureux, H. Choi, J. Kim, C. Kim, S. Nakamura, J. S. Speck, and S. P. DenBaars, “Improved performance of AlGaInP red micro-light-emitting diodes with sidewall treatments,” Opt. Express 28(4), 5787–5793 (2020). [CrossRef]
45. B. Lu, Y. Wang, B.-R. Hyun, H.-C. Kuo, and Z. Liu, “Color difference and thermal stability of flexible transparent InGaN/GaN multiple quantum wells mini-LED arrays,” IEEE Electron Device Lett. 41(7), 1040–1043 (2020) [CrossRef] .
46. A.-C. Liu, K. J. Singh, Y.-M. Huang, T. Ahmed, F.-J. Liou, Y.-H. Liou, C.-C. Ting, C.-C. Lin, Y. Li, S. Samukawa, and H.-C. Kuo, “Increase in the efficiency of III-Nitride Micro-LEDs: Atomic-layer deposition and etching,” IEEE Nanotechnology Mag. 15(3), 18–34 (2021). [CrossRef]
47. S. Lai, W. Lin, J. Chen, T. Lu, S. Liu, Y. Lin, Y. Lu, Y. Lin, Z. Chen, H.-C. Kuo, W. Guo, and T. Wu, “The impacts of sidewall passivation via atomic layer deposition on GaN-based flip-chip blue mini-LEDs,” J. Phys. D: Appl. Phys. 55(37), 374001 (2022). [CrossRef]
48. B. Vermang, F. Werner, W. Stals, A. Lorenz, A. Rothschild, J. John, J. Poortmans, R. Mertens, R. Gortzen, P. Poodt, F. Roozeboom, and J. Schmidt, “Spatially-separated atomic layer deposition of Al2O3, a new option for high-throughput Si solar cell passivation,” in 37th Photovoltaic Specialists Conference (IEEE, 2011), pp. 001144–001149.
49. O. Salihoglu, A. Muti, K. Kutluer, T. Tansel, R. Turan, C. Kocabas, and A. Aydinli, “Passivation of type II InAs/GaSb superlattice photodetectors with atomic layer deposited Al2O3,” Proc. SPIE 8353, 83530Z–83530Z-7 (2012). [CrossRef]
50. M.-S. Park, M. Razaei, K. Barnhart, C. L. Tan, and H. Mohseni, “Surface passivation and aging of InGaAs/InP heterojunction phototransistors,” J. Appl. Phys. 121(23), 233105 (2017). [CrossRef]
51. H. Cha, J. Lee, L. R. Jordan, S. H. Lee, S.-H. Oh, H. J. Kim, J. Park, S. Hong, and H. Jeon, “Surface passivation of a photonic crystal band-edge laser by atomic layer deposition of SiO2 and its application for biosensing,” Nanoscale 7(8), 3565–3571 (2015). [CrossRef]
52. N. Batra, J. Gope Vandana, J. Panigrahi, R. Singh, and P. K. Singh, “Influence of deposition temperature of thermal ALD deposited Al2O3 films on silicon surface passivation,” AIP Adv. 5(6), 067113 (2015). [CrossRef]
53. C.-C. Shen, T.-C. Hsu, Y.-W. Yeh, C.-Y. Kang, Y.-T. Lu, H.-W. Lin, H.-Y. Tseng, Y.-T. Chen, C.-Y. Chen, C.-C. Lin, C.-H. Wu, P.-T. Lee, Y. Sheng, C.-H. Chiu, and H.-C. Kuo, “Design, modeling, and fabrication of high-speed VCSEL with data rate up to 50 Gb/s,” Nanoscale Res. Lett. 14(1), 276 (2019). [CrossRef]
54. N. Nishiyama, M. Arai, S. Shinada, K. Suzuki, F. Koyama, and K. Iga, “Multi-oxide layer structure for single-mode operation in vertical-cavity surface-emitting lasers,” IEEE Photonics Technol. Lett. 12(6), 606–608 (2000). [CrossRef]
55. M. Azuchi, N. Jikutani, M. Arai, T. Kondo, and F. Koyama, “Multioxide layer vertical-cavity surface-emitting lasers with improved modulation bandwidth,” CLEO/Pacific Rim 2003. The 5th Pac. Rim Conf. on Lasers Electro-Optics (IEEE Cat. No.03TH8671) 1, 163 (2003). [CrossRef]
56. S. Corzine, R. Geels, J. Scott, R.-H. Yan, and L. Coldren, “Design of Fabry-Perot surface-emitting lasers with a periodic gain structure,” IEEE J. Quantum Electron. 25(6), 1513–1524 (1989). [CrossRef]
57. Y. Huang, Z. Pan, and R. Wu, “Analysis of the optical confinement factor in semiconductor lasers,” J. Appl. Phys. 79(8), 3827–3830 (1996). [CrossRef]
58. F. A. I. Chaqmaqchee and J. A. Lott, “Impact of oxide aperture diameter on optical output power, spectral emission, and bandwidth for 980 nm VCSELs,” OSA Continuum 3(9), 2602–2613 (2020). [CrossRef]
59. P. Moser, J. A. Lott, and D. Bimberg, “Energy efficiency of directly modulated oxide-confined high bit rate 850-nm VCSELs for optical interconnects,” IEEE J. Sel. Top. Quantum Electron. 19(4), 1702212 (2013). [CrossRef]
60. V. P. Kalosha, V. A. Shchukin, N. N. LedentsovJr., J.-R. Kropp, and N. N. Ledentsov, “Robustness versus thermal effects of single-mode operation of vertical-cavity surface-emitting lasers with engineered leakage of high-order transverse optical modes,” Proc. SPIE 10122, 101220K (2017). [CrossRef]
61. N. Ledentsov, L. Chorchos, M. Agustin, N. N. Ledentsov, and J. P. Turkiewicz, “850 nm single-mode VCSEL for error-free 60 Gbit/s OOK operation and transmission through 800 m of multi-mode fiber,” in Optical Fiber Communications Conference and Exhibition (OSA, 2019), paper Th4B.6.
62. D. Wu, X. Yu, H. Wu, W. Fu, and M. Feng, “Single-mode 850nm VCSELs demonstrate 96 Gb/s PAM4 OM4 fiber link for extended reach to 1km,” in Optical Fiber Communication Conference (Optica Publishing Group, 2022), paper W2A.7.
63. J. Rösler, M. Bäker, and H. Harders, Mechanical Behaviour of Engineering Materials Metals, Ceramics, Polymers, and Composites (Springer-Verlag, 2007).
64. M. S. Wong, D. Hwang, A. I. Alhassan, C. Lee, R. Ley, S. Nakamura, and S. P. DenBaars, “High efficiency of III-nitride micro-light-emitting diodes by sidewall passivation using atomic layer deposition,” Opt. Express 26(16), 21324–21331 (2018). [CrossRef]
65. Y. Lee, J. Jewell, B. Tell, K. Brown-Goebeler, A. Scherer, J. Harbison, and L. Florez, “Effects of etch depth and ion implantation on surface emitting microlasers,” Electron. Lett. 26(4), 225–227 (1990). [CrossRef]
66. B. Zhao, T. R. Chen, Y. H. Zhuang, A. Yariv, J. E. Ungar, and S. Oh, “High speed operation of very low threshold strained InGaAs/GaAs double quantum well lasers,” Appl. Phys. Lett. 60(11), 1295–1297 (1992). [CrossRef]
67. Y.-C. Chang and L. A. Coldren, “Efficient, high-data-rate, tapered oxide-aperture vertical-cavity surface-emitting lasers,” IEEE J. Sel. Top. Quantum Electron. 15(3), 704–715 (2009). [CrossRef]
68. C. H. Wu, F. Tan, M. K. Wu, M. Feng, and N. Holonyak, “The effect of microcavity laser recombination lifetime on microwave bandwidth and eye-diagram signal integrity,” J. Appl. Phys. 109(5), 053112 (2011). [CrossRef]
69. A. Grabowski, J. Gustavsson, Z. S. He, and A. Larsson, “Large-signal equivalent circuit for datacom VCSELs,” J. Lightwave Technol. 39(10), 3225–3236 (2021). [CrossRef]
