Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Impact of oxide aperture diameter on optical output power, spectral emission, and bandwidth for 980 nm VCSELs

Open Access Open Access

Abstract

We produce experimental 980 nm vertical cavity surface emitting lasers (VCSELs) with a wide range of oxide aperture diameters (ø) from ∼2.5 to 15 µm on wafers designed to minimize the epitaxial growth and VCSEL design complexity. The structures are grown in batches of 12, 3-inch diameter wafers in a production metal-organic vapor phase epitaxy machine. We characterize the top emitting VCSELs at room temperature (∼25 °C) – grouped into unit cells with 16 rows and 15 columns—using an automated (university-built) wafer mapping system, resulting in two-dimensional colorized maps of several performance attributes of interest including optical output power, threshold current, and maximum power conversion efficiency. By etching part of the topmost layer of the upper distributed Bragg reflector to decrease the VCSEL optical cavity photon lifetime, we boost the small signal modulation bandwidth (f3dB). The room temperature maximum f3dB is ∼30 GHz for VCSELs with ø ∼3 µm and ∼20 GHz for VCSELs with ø ∼15 µm.

© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

1. Introduction

Vertical cavity surface emitting lasers (VCSELs) are a key component for optical interconnects and data communication systems [1,2]. New applications of VCSELs have rapidly appeared in various consumer applications [36] and as optical sources for a plethora of emerging 5G (fifth generation) mobile optical wireless devices. Modern VCSELs have numerous advantages compared to edge emitting lasers for short-reach data communications such as excellent high-speed properties, lower energy consumption, wafer scale testing, low fabrication costs, and high reliability. Commercial VCSELs are the standard light source for use with multiple-mode optical fiber (MMF) to form short-reach (up to ∼300 m) inter-rack optical links in data centers and supercomputers [7,8]. Low threshold (below 1 mA), oxide confined, 980 nm GaInAs quantum well (QW) VCSELs with AlGaAs distributed Bragg reflector (DBR) mirrors with small-signal modulation bandwidths (f3dB) ∼30 GHz and corresponding error free (bit error ratio ≤ 1 × 10−12) data transmission rates of 55 Gb/s have been measured for VCSELs with oxide aperture diameters (ø) of 7 µm [9]. The improvements on the performance of VCSELs at wavelength of 850 nm have enabled f3dB exceeding 30 GHz [10]. Room temperature 980 nm VCSELs have been demonstrated with f3dB of 26.6 to 18.6 GHz for ø ∼13.5 and 33.5 µm, respectively [11], and f3dB ∼30 GHz for 4 µm VCSELs with a hybrid binary (AlAs/GaAs) and ternary (AlGaAs/GaAs) bottom DBR design and an 18.5-period top (ternary) coupling DBR [12]. For conventional single cavity 850, 980, and 1060 nm VCSELs via 2-level pulse amplitude modulation (PAM-2) the highest reported (error free) data transmission rate across multiple mode optical fiber at these three wavelengths is ≥ 60 Gb/s [13].

Various approaches have been considered to increase the bandwidth and thus the bit rate of VCSELs including the use of optical injection locking [14], coupled cavities [15], and integration with an electro-optic modulator [16]. In addition, optical feedback has been employed to enhance the small-signal modulation bandwidth and to increase the speed of VCSELs [1719]. Another technique to enhance f3dB (at the expense of reducing the damping in the |S21| frequency response curves – thus increasing the |S21| peaking before reaching f3dB) is the etching of the topmost DBR layer to adjust the optical cavity photon lifetime [2022]. An analogous (well known) approach to reduce the cavity photon lifetime (and increase the optical output power for a given ø) is to reduce the number of top coupling DBR periods [23]. We note the possibility of adding dielectric capping layers to a VCSEL – such as SiN [24] – to detune the VCSEL resonance and thus to tweak (to reduce or to increase) the cavity photon lifetime and concurrently enhance or inhibit the maximum f3dB [25].

While 850 nm is the standard wavelength for OM1-OM4 MMF for short-reach VCSEL-based data transmission, the new OM5 optical fiber is designed specifically for optimal data transmission at 850 nm and 940 nm. There is great interest in wavelength division multiplexing, where data is sent (for example) at 850, 880, 910, and 940 nm across OM5 MMF. The wavelength division multiplexing (WDM) concept may be extended to include 980 nm and longer wavelengths including for example up to 1060 nm [13,26], 1090 nm [27], and 1530 nm [28] – especially (as we anticipate) when newer multimode and/or few mode or single mode optical fibers designed specifically to support these wavelengths are introduced. There is massive interest as well in increasing data rates for VCSEL-based links via advanced modulation schemes including 4-level pulse amplitude modulation (PAM-4) [2732] and advanced error correction techniques, including the fresh report of 168 Gb/s via PAM-4 with an 850 nm VCSEL with a bandwidth of 28 GHz [33]. We note the 980 nm VCSELs we report herein – designed with epitaxial simplicity in mind to minimize production costs – have similar (maximum) bandwidths (∼28-30 GHz).

Our VCSEL structure is designed for top (epitaxial) surface emission and grown on an (n+) GaAs substrate using a metalorganic vapor-phase epitaxy (MOVPE) growth reactor. Figure 1(a) shows the schematic cross section of the fabricated VCSEL after standard processing which includes ultraviolet (UV) contact lithography, selective thermal wet oxidation, dry (Cl2+BCl3 plasma) etching, and ohmic metal contact deposition. The 980 nm VCSEL structure consists of a 37-period n-doped (with silicon) bottom DBR and a 20.5-period p-doped (with carbon) top DBR. The two DBRs use Al0.9Ga0.1As/GaAs epitaxial layers with graded interfaces. We employ (approximately) linear compositional grading (from x = 0.0 to 0.9 and vice versa) and an approximately uniform doping increasing in density when moving away from the optical cavity active region from ∼2 × 1018 cm-3 to 4 × 1018 cm-3 for the (n) DBR and ∼2 × 1018 cm-3 to ∼6 × 1018 cm-3 for the (p) DBR to reduce electrical resistance and optical absorption. The 0.5λ optically thick cavity contains five strained Ga0.77In0.23As quantum wells surrounded by GaAs0.86P0.14 barrier layers and half of two 20 nm thick (as grown) Al0.98Ga0.02As oxide aperture layers (that are selectively thermally oxidized in water vapor at 420 °C during the device fabrication). The resultant Al0.98Ga0.02As-Al0.98Ga0.02Oz layers reside at nodes of the optical field intensity (on resonance) at the interfaces between the optical cavity and the top and bottom DBRs. A GaAs ohmic contact and buffer layer of ∼1.586 µm thickness lies just below the bottom (n) DBR. We isolate the n-doped GaAs substrate and the doped GaAs buffer layer with an undoped Al0.90Ga0.10As layer. The static results of light-current-voltage (LIV) wafer mapping at room temperature (RT ∼25 °C) of the VCSEL wafer piece (for unit cell 33 only) is given in Fig. 1(b). The VCSELs have two mesas and are arranged in repeating unit cells with 16 rows × 15 columns, thus there are 240 VCSELs in each unit cell. The VCSEL to VCSEL spacing is 600 µm in both the x and y directions. The rows of VCSELs in each unit cell are labeled row 0 to row F, while the columns are labeled column 0 to column E. We characterize the VCSELs in columns D and E in unit cell 33, located inside the lime colored rectangle in Fig. 1(b).

 figure: Fig. 1.

Fig. 1. (a) Schematic cross-section (not to scale) of a top emitting 980 nm VCSEL utilizing an GaInAs/GaAsP active region sandwiched between top and bottom AlGaAs DBR mirrors and two oxide aperture layers; and (b) part of a colorized LIV mapping of the maximum optical output power of a processed VCSEL quarter wafer piece, where the center region marked by thick horozontal white lines is a unit cell with 16 rows (rows 0 to F) ×15 columns (columns 0 to E). The black rectangles are either off the VCSEL quarter wafer piece (you can see the curved edge on the lower left) or nonfunctioning VCSELs.

Download Full Size | PDF

In Fig. 2 we illustrate the experimental test setup employed to measure the optical emission spectra and the small-signal modulation frequency response of our 980 nm VCSELs at room temperature. The optical output from a VCSEL is coupled into a cleaved-end multiple-mode optical fiber (MMF) - an OM1 MMF with a 62.5 µm diameter core, and then into an optical spectrum analyzer (OSA) Ando model AQ6317C to analyze the optical emission spectra over ∼780 to 1180 nm (or over a narrower range as needed). To perform the small signal modulation analysis the VCSEL is connected to a Hewlett-Packard vector network analyzer (VNA) model 8722C with a 50 MHz to 40 GHz frequency range. A Keithley model 2400-LV current source is connected to the VNA – so the VNA provides both the DC (the current bias) and AC (the modulated sinusoidal small signal) to the VCSEL using on wafer probing via a GSG (ground-signal-ground) co-planar probe head. The VCSEL’s optical output is coupled into a 25 GHz bandwidth New Focus photodetector model 1434-50 via the OM1 MMF patch cord. The VNA, OSA, and Keithley current source operate under computer control. The VCSEL wafer is placed on a probe station platen (a flat metal surface with small vacuum holes) and held at room temperature. We control all measurements via a proprietary LabVIEW program running on a Microsoft Windows-based personal tower computer.

 figure: Fig. 2.

Fig. 2. Experimental test setup for the emission spectra and the small signal modulation frequency response (2-port S21 parameter) measurements for the 980 nm VCSELs.

Download Full Size | PDF

2. Results and discussion

We present our room temperature static characterization of two columns of VCSELs in Fig. 3. The measured LIV curves for a column of VCSELs with ø ∼2.5 to 15 µm (from row 1 column E to row F column E; without rows 0 and D) are shown in Fig. 3(a). We estimate the oxide aperture diameters based on: 1) our oxidation tests on several test pieces prior to VCSEL fabrication (we study oxidation length versus oxidation time via scanning electron microscopy (SEM) cross section images); 2) the knowledge that the top mesa diameters increase in steps of 0.5 and 1.0 µm in our VCSEL columns; and 3) the nature of the emission spectra at various bias currents based on our historical experience. As a final sanity check we use the simplified model in [34] which yields an estimate of the oxide aperture diameter from the difference in the peak wavelengths of the fundamental emission mode (LP01) and the first higher order emission mode (LP10). From this rough model we find our estimated ϕ are within plus or minus 1 µm of the calculated values (and our VCSELs in Rows 5 to A are within 0.5 µm).

 figure: Fig. 3.

Fig. 3. Room temperature static characteristics of the VCSELs with a top mesa diameter from 18.5 to 31 µm and corresponding bottom mesa diameters of 78.5 to 91 µm (column E) and 58.5 to 71 µm (column D): (a) LIV from row 1 to row F in column E; (b) LIV from row 1 to row F in column D.

Download Full Size | PDF

Because we use two oxide apertures centered at optical field intensity nodes (the first nodes on both sides of the optical cavity), and due to thermal lensing we are not able to precisely determine the oxide aperture diameters. It is well known, however that ∼850-980 nm VCSELs achieve single mode emission or quasi-single mode emission when their double oxide aperture diameters are about 3 µm or less. Our VCSELs in column E have a variable top mesa diameter from 18.5 to 31 µm, where the diameter of the top mesa is increased from 18.5 to 20 µm in 0.5 µm steps (row 0 to row 3) and increases from 20 to 31 µm in 1 µm steps (row 4 to row F) with a fixed bottom mesa diameter equal to the top mesa diameter + 60 µm (thus from 78.5 to 91 µm). The LI rollover optical output powers vary from 0.6 to 11.3 mW at threshold currents of Ith = 0.12 to 0.99 mA. In Fig. 3(b), the measured static LIV characteristics are also measured for VCSELs with ø ∼2.5 to 15 µm (from row 1 in column D to row F in column D), where the column D VCSELs have a top mesa diameter between 18.5 and 31 µm with a fixed bottom mesa diameter equal to the top mesa diameter + 40 µm. The rollover optical output powers vary from 0.45 to 10.2 mW at about the same threshold currents as for the adjacent VCSELs in column E.

The threshold current and the threshold current density as a function of the oxide aperture diameter for our VCSELs in column D and column E are shown in Fig. 4(a). In both columns the threshold current is nearly invariant between 2.5 to 4 µm and then increases as a result of increasing ø from ∼5 µm to 15 µm, while the threshold current density sharply increases when moving from 4 to 2.5 µm (as expected) and is roughly constant (within ∼0.5 to 0.8 kA/cm2) between ø ∼5 and 15 µm. For the ø ∼3 µm VCSEL the threshold current of 0.12 mA corresponds to a threshold current density of 1.68 kA/cm2 in both columns D and E. In Fig. 4(b), the optical output power and the maximum power conversion efficiency (PCE) as a function of ø are presented for the VCSELs in columns D and E. The output power at ø ∼15 µm exceeds 11 mW in column E and 10 mW in column D (where the bottom mesa area is smaller), while the maximum PCE is around 26% for all VCSELs in columns D and E (except for the smallest VCSELs with ø < 4 µm where the maximum PCE decreases due likely to increasing internal losses such as optical scattering losses).

 figure: Fig. 4.

Fig. 4. Room temperature static characteristics of the VCSELs: (a) threshold current and threshold current density as a function of oxide aperture diameter; and (b) maximum optical output power and maximum power conversion efficiency versus oxide aperture diameter from ø ∼2.5 to 15 µm.

Download Full Size | PDF

We perform spectral measurements of the VCSELs at different bias currents using a cleaved end OM1 MMF. We lightly place (butt couple) the optical fiber on each VCSEL to collect the output light and then analyze the emission data (power in dBm versus wavelength) measured via our optical spectrum analyzer. In Fig. 5, the peak emission wavelength (of the fundamental LP01 mode) as a function of operating current is plotted for a sampling of the VCSELs in columns D and E at ∼25 °C.

 figure: Fig. 5.

Fig. 5. Room temperature peak emission wavelength (for the fundamental LP01 optical mode) as a function of bias current for 980 nm VCSELs with oxide aperture diameters from ø ∼2.5 to 12 µm. The solid shapes are column D results while the open shapes are column E results.

Download Full Size | PDF

We next plot example optical emission spectra for our VCSELs in Fig. 6. The optical emission spectra measurements reveal that the root mean square (RMS) spectral width for a 3 µm VCSEL is about 0.2 nm. The spectral width for a 15 µm VCSEL exceeds 1.5 nm. The narrow spectral width of the single mode (SM) VCSEL reduces the chromatic dispersion and typically allows data transmission over longer distances [35,36]. In Fig. 6(b), the LP01 and LP10 peak emission wavelengths are about 979.2 and 977.6 nm, respectively for the ø ∼3 µm VCSEL (row 2, column E). The LP01 and LP10 peak emission wavelengths are about 979.2 and 979.0 nm, respectively for the ø ∼15 µm VCSEL (row F, column E). A small oxide aperture diameter (e.g. ø ≤ 5 µm) lowers the threshold current but increases the differential resistance and self-heating [37]. The number of transverse modes of multiple-mode VCSELs naturally increases as the oxide aperture diameter increases [38].

 figure: Fig. 6.

Fig. 6. Room temperature emission spectra for VCSELs with: (a) ø ∼3 µm (row 2, column D; blue curve) and ø ∼11 µm (row B, column D; red curve) at bias currents of 1.9 and 2 mA, respectively; and (b) ø ∼3 µm (row 2, column E; blue curve) and ø ∼15 µm (row F, column E; red curve) at a bias current of 2 mA.

Download Full Size | PDF

In Fig. 7 (top) we plot the room temperature |S21| frequency response for a VCSEL with ø ∼3 µm at forward CW bias currents I = 0.5, 1, 2, and 2.5 mA. We show the raw (smoothed and corrected) data and curve fit for 0.5 mA. For ease of viewing we show only the curve fits for the other bias currents without the raw data. We fit the |S21| data using the curve fitting equation described in [39]. The peaking of the |S21| response is about 8 dB at I = 0.5 mA reflecting a low damping (γ = 21 ns from our curve fitting) despite the large number (20.5 periods) of top DBR periods. This is because during VCSEL processing we remove over 50% of the top DBR layer (after placing the p-metal contacts), which reduces the cavity photon lifetime and increases the |S21| peaking (as explained in [20]). However, the |S21| peaking decreases to ∼1 dB when I = 2.5 mA (where γ = 134 ns from our curve fitting). The maximum small-signal modulation bandwidth (f3dB) clearly exceeds 25 GHz (for example when I = 2 mA). In Fig. 7 (bottom) we plot the room temperature |S21| frequency response for a VCSEL with ø ∼15 µm at forward CW bias currents I = 3 (where γ = 19 ns), 5, 11, and 25 mA (where γ = 65 ns). We show the raw (smoothed and corrected) data and curve fit for 11 mA, and for ease of viewing we show only the curve fits for the other bias currents. The |S21| peaking remains about 4 dB at all bias currents – reflecting the relatively large ø of this VCSEL. The maximum f3dB is about 23 GHz (when I = 25 mA).

 figure: Fig. 7.

Fig. 7. Small signal modulation response (|S21|) versus frequency at different CW bias currents for: (top) a VCSEL with ø ∼3 µm; and (bottom) a VCSEL with ø ∼15 µm. The black curves are the raw |S21| data smoothed and corrected for the photodiode response. The red curves are curve fits to the corrected |S21| data. For ease of viewing we show only the raw data for one VCSEL in each graph.

Download Full Size | PDF

In Fig. 8, we plot the room temperature small-signal modulation bandwidth (f3dB) versus oxide aperture diameter (for ø ∼3 to 15 µm in column E) for our 980 nm VCSELs at CW bias current from 3 to 13 mA in 2 mA steps. The maximum small-signal modulation bandwidth is ∼30 GHz for ø ∼3 µm at a bias current of 2.5 mA. The maximum bandwidth is reduced with increasing oxide aperture diameter when moving from ø ∼3 to ø ∼15 µm. We generally wish to operate VCSELs at a CW bias current density of J ≤ 10 kA/cm2 to avoid oxide VCSEL reliability problems [40]. This maximum J restriction limits the practical operating CW bias currents of our VCSELs to I ∼ 0.7, 6.4, and 17.6 mA for our ø ∼3, 9, and 15 µm VCSELs, respectively. Ideally, we achieve a high bandwidth (and simultaneously reasonably high optical output power) at the lowest possible bias current.

 figure: Fig. 8.

Fig. 8. The -3 dB small-signal modulation bandwidth (f3dB) for CW bias in the range I = 3 to 13 mA as a function of the oxide aperture diameter for 980 nm VCSELs of column E at 25 °C. Linear line fits are overlain on the measurements.

Download Full Size | PDF

3. Summary

We characterize top emitting 980 nm vertical cavity surface emitting lasers (VCSELs) with a simplified epitaxial structure targeted for data communication applications. Our wafer piece characterization proceeds first via a university-built static LIV wafer mapping system. We investigate and compare the performance of VCSELs with a wide range of oxide aperture dimeters from ø ∼2.5 to 15 µm within top mesa diameters from 18.5 to 31 µm and with bottom mesa diameters equal to 58.5 to 71 µm (column D) and 78.5 to 91 µm (column E). For our VCSELs with ø ∼3 µm we achieve a maximum -3 dB modulation bandwidth of ∼30 GHz, which is limited to ∼12 GHz if we wish to keep the operating current density J below 10 kA/cm2. For our VCSELs with ø ∼15 µm, however we achieve a bandwidth of ∼20 GHz for J ∼10 kA/cm2 with a moderate |S21| peaking of ∼4 dB.

Funding

MARHABA Erasmus Mundus Lot 3 (2014-0653); German Research Foundation (Collaborative Research Center 787).

Acknowledgments

FC highly appreciates the support of the TU Berlin team (Arbeitsgruppe (AG) Lott) during this work. FC also acknowledges the support of Koya University during the period of her visit in Germany.

Disclosures

The authors declare no conflicts of interest.

References

1. P. Westbergh, J. Gustavsson, Å Haglund, M. Skold, A. Joel, and A. Larsson, “High speed, low-current-density 850 nm VCSELs,” IEEE J. Sel. Top. Quantum Electron. 15(3), 694–703 (2009). [CrossRef]  

2. R. Safaisini, J. R. Joseph, and K. L. Lear, “Scalable high-CW-power high-speed 980-nm VCSEL arrays,” IEEE J. Quantum Electron. 46(11), 1590–1596 (2010). [CrossRef]  

3. F. Koyama, “Recent Advances of VCSEL Photonics,” J. Lightwave Technol. 24(12), 4502–4513 (2006). [CrossRef]  

4. K. Iga, “Vertical-cavity surface-emitting laser: its conception and evolution,” J. Appl. Phys. 47(1), 1–10 (2008). [CrossRef]  

5. F. A. I. Chaqmaqchee, “Optically and electrically pumped of Ga0.65In0.35N0.02As0.98/GaAs vertical-cavity surface-emitting lasers (VCSELs),” Arab. J. Sci. Eng. 39(7), 5785–5790 (2014). [CrossRef]  

6. F. A. I. Chaqmaqchee, “Performance characteristics of conventional vertical cavity surface emitting lasers VCSELs at 1300 nm,” ZANCO J. Pure Appl. Sci. 31(2), 14–18 (2019). [CrossRef]  

7. M. A. Taubenblatt, “Optical interconnects for high-performance computing,’,” J. Lightwave Technol. 30(4), 448–457 (2012). [CrossRef]  

8. R. Safaisini, K. Szczerba, P. Westbergh, E. Haglund, B. Kögel, J. S. Gustavsson, M. Karlsson, P. Andrekson, and A. Larsson, “High-speed 850-nm quasi-single mode VCSELs for extended reach optical interconnects,” J. Opt. Commun. Netw. 5(7), 686–695 (2013). [CrossRef]  

9. N. Haghighi, J. Lavrencik, S. E. Ralph, and J. A. Lott, “55 Gbps error free data transmission with 980 nm VCSELs across 100 m of multiple-mode optical fiber,” paper TuE3-3, Proceedings 24th OptoElectronic and Communications Conference (OECC), Fukuoka, Japan (07–11 July 2019).

10. E. Haglund, P. Westbergh, J. S. Gustavsson, E. P. Haglund, A. Larsson, M. Geen, and A. Joel, “30 GHz bandwidth 850 nm VCSEL with sub-100 fJ/bit energy dissipation at 25-50 Gbit/s,” Electron. Lett. 51(14), 1096–1098 (2015). [CrossRef]  

11. N. Haghighi, P. Moser, and J. A. Lott, “Power, bandwidth, and efficiency of single VCSELs and small VCSEL arrays,” IEEE J. Sel. Top. Quantum Electron. 25(6), 1–15 (2019). [CrossRef]  

12. M. Gębski, P.-S. Wong, M. Riaziat, and J. A. Lott, “30 GHz bandwidth temperature stable 980 nm VCSELs with AlAs/GaAs bottom DBRs for optical data communication,” J. Phys. Photonics 2(3), 035008 (2020). [CrossRef]  

13. J. Lavrencik, E. Simpanen, N. Haghighi, S. Varughese, J. S. Gustavsson, E. Haglund, W. V. Sorin, S. Mathai, M. Tan, J. A. Lott, A. Larsson, and S. E. Ralph, “Error-free 850 nm to 1060 nm VCSEL links: feasibility of 400Gbps and 800 Gbps 8λ-SWDM,” Proceedings 45th European Conference on Optical Communication (ECOC), Dublin, Ireland, P84, (22–26 September 2019).

14. S. H. Lee, D. Parekh, T. Shindo, W. Yang, P. Guo, D. Takahashi, N. Nishiyama, C. J. Chang-Hasnain, and S. Arai, “Bandwidth enhancement of injection-locked distributed reflector lasers with wire like active regions,” Opt. Express 18(16), 16370–16378 (2010). [CrossRef]  

15. C. Chen and K. D. Choquette, “Analog and digital functionalities of composite-resonator vertical-cavity lasers,” J. Lightwave Technol. 28(7), 1003–1010 (2010). [CrossRef]  

16. L. Marigo-Lombart, A. Rumeau, C. Viallon, A. Arnoult, S. Calvez, A. Monmayrant, O. Gauthier-Lafaye, R. Rosales, J. A. Lott, H. Thienpont, K. Panajotov, and G. Almuneau, “High frequency operation of an integrated electro-absorption modulator onto a VCSEL,” J. Phys. Photonics 1(2), 02LT01 (2019). [CrossRef]  

17. T. Ohta, T. Uchida, T. Kondo, and F. Koyama, “Enhanced modulation bandwidth of surface-emitting laser with external optical feedback,” IEICE Electron. Express 1(13), 368–372 (2004). [CrossRef]  

18. G. Morthier, R. Schard, and O. Kjebon, “Extended modulation bandwidth of DBR and external cavity lasers by utilizing a cavity resonance for equalization,” IEEE J. Quantum Electron. 36(12), 1468–1475 (2000). [CrossRef]  

19. E. Heidari, H. Dalir, M. Ahmed, V. J. Sorger, and R. T. Chen, “Hexagonal transverse coupled cavity VCSEL redefining the high-speed lasers,” arXiv:2008.01346v1 (04 August 2020).

20. P. Westbergh, J. Gustavsson, B. Kögel, Å Haglund, and A. Larsson, “Impact of photon lifetime on high-speed VCSEL performance,” IEEE J. Sel. Top. Quantum Electron. 17(6), 1603–1613 (2011). [CrossRef]  

21. D. Ellafi, V. Iakovlev, A. Sirbu, G. Suruceanu, Z. Micković, A. Caliman, A. Mereuta, and E. Kapon, “Control of cavity lifetime of 1.5 µm wafer-fused VCSELs by digital mirror trimming,” Opt. Express 22(26), 32180–32187 (2014). [CrossRef]  

22. E. P. Haglund, P. Westbergh, J. S. Gustavsson, and A. Larsson, “Impact of damping on large signal VCSEL dynamics,” J. Lightwave Technol. 33(4), 795–801 (2015). [CrossRef]  

23. G. M. Yang, M. H. MacDougal, V. Pudikov, and P. D. Dapkus, “Influence of mirror reflectivity on laser performance of very-low-threshold vertical-cavity surface-emitting lasers,” IEEE Photonics Technol. Lett. 7(11), 1228–1230 (1995). [CrossRef]  

24. M. Volwahsen, “The Influence of the mirror reflectivity on the temperature performance of high speed VCSELs,” Thesis, Technical University Berlin (07 August 2014).

25. G. Larisch, P. Moser, J. A. Lott, and D. Bimberg, “Impact of photon lifetime on the temperature stability of 50 Gb/s 980 nm VCSELs,” IEEE Photonics Technol. Lett. 28(21), 2327–2330 (2016). [CrossRef]  

26. E. Simpanen, J. S. Gustavsson, A. Larsson, M. Karlsson, W. V. Sorin, S. Mathai, M. R. Tan, and S. R. Bickham, “2019 nm single-mode VCSEL and single-mode fiber links for long-reach optical interconnects,” J. Lightwave Technol. 37(13), 2963–2969 (2019). [CrossRef]  

27. B. Wang, W. V. Sorin, P. Rosenberg, L. Kiyama, S. Mathai, and M. R. T. Tan, “4×112 Gbps/Fiber CWDM VCSEL arrays for co-packaged interconnects,” J. Lightwave Technol. 38(13), 3439–3444 (2020). [CrossRef]  

28. F. Karinou, C. Prodaniuc, N. Stojanovic, M. Ortsiefer, A. Daly, R. Hohenleitner, B. Kögel, and C. Neumeyr, “Directly PAM-4 modulated 1530-nm VCSEL enabling 56 Gb/s/λ data-center interconnects,” IEEE Photonics Technol. Lett. 27(17), 1872–1875 (2015). [CrossRef]  

29. S. M. R. Motaghian Nezam, I. Lyubomirsky, H. Daghighian, C. Kocot, T. Gray, J. Tatum, A. Amezcua-Correa, M. Bigot-Astruc, D. Molin, F. Achten, and P. Sillard, “180 Gbps PAM4 VCSEL transmission over 300 m wideband OM4 fibre,” Proceedings Optical Fiber Communication Conference (OFC), Anaheim, CA USA (20-24 March 2016).

30. J. Lavrencik, S. Varughese, V. A. Thomas, G. Landry, Y. Sun, R. Shubochkin, K. Balemarthy, J. Tatum, and S. E. Ralph, “14λ x 100Gbps VCSEL PAM-4 transmission over 105 m of wide band multimode fiber,” Proceedings Optical Fiber Communication Conference (OFC), Los Angeles, CA USA (19-23 March 2017).

31. J. Lavrencik, S. Varughese, V. A. Thomas, J. S. Gustavsson, E. Haglund, A. Larsson, and S. E. Ralph, “102Gbps PAM-2 over 50m OM5 Fiber using 850nm Multimode VCSELs,” 1–2, Proceedings IEEE Photonics Conference (IPC), San Antonio, TX USA (29 September-03 October 2019).

32. T. Aoki, R. Kubota, S. Yoshimoto, M. Yanagisawa, T. Ishizuka, and H. Shoji, “High-speed VCSEL arrays for 400 Gbit/s data center interconnects,” SEI Tech. Rev. 90, 71–75 (2020).

33. J. Lavrencik, S. Varughese, N. Ledentsov, Ł. Chorchos, N. N. Ledentsov, and S. E. Ralph, “168Gbps PAM-4 Multimode Fiber Transmission through 50 m using 28 GHz 850 nm Multimode VCSELs,” 1–3, W1D.3, Proceedings Optical Fiber Communication Conference (OFC), San Diego, CA USA (28 March-01 April 2020).

34. C. H. Wu, F. Tan, M. K. Wu, M. Feng, and N. Holonyak, “The effect of microcavity recombination lifetime on microwave bandwidth and eye-diagram signal integrity,” J. Appl. Phys. 109(5), 053112 (2011). [CrossRef]  

35. R. Safaisini, E. Haglund, P. Westbergh, J. S. Gustavsson, and A. Larsson, “20 Gb/s data transmission over 2 km multimode fibre using an 850 nm mode filter VCSEL,” Electron. Lett. 50(1), 40–42 (2014). [CrossRef]  

36. H.-Y. Kao, C.-T. Tsai, S.-F. Leong, C.-Y. Peng, Y.-C. Chi, J. J. Huang, H.-C. Kuo, T. Shih, J.-J. Jou, W.-H. Cheng, C.-H. Wu, and G.-R. Lin, “Comparison of single-/few-/multi-mode 850 nm VCSELs for optical OFDM transmission,” Opt. Express 25(14), 16347–16363 (2017). [CrossRef]  

37. B. Weigl, M. Grabherr, C. Jung, R. Jäger, G. Reiner, R. Michalzik, D. Sowada, and K. J. Ebeling, “High-performance oxide-confined GaAs VCSELs,” IEEE J. Sel. Top. Quantum Electron. 3(2), 409–415 (1997). [CrossRef]  

38. Y. Satuby and M. Orenstein, “Mode-coupling effects on the small signal modulation of multi transverse-mode vertical-cavity semiconductor lasers,” IEEE J. Quantum Electron. 35(6), 944–954 (1999). [CrossRef]  

39. N. Haghighi, P. Moser, and J. A. Lott, “40 Gbps with electrically parallel triple and septuple 980 nm VCSEL arrays,” J. Lightwave Technol. 38(13), 3387–3394 (2020). [CrossRef]  

40. B. M. Hawkins, R. A. Hawthorne, J. K. Guenter, J. A. Tatum, and J. R. Biard, “Reliability of various size oxide aperture VCSELs,” 540–550, Proceedings IEEE 52nd Electronic Components and Technology Conference, San Diego, CA USA (May 2002).

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (8)

Fig. 1.
Fig. 1. (a) Schematic cross-section (not to scale) of a top emitting 980 nm VCSEL utilizing an GaInAs/GaAsP active region sandwiched between top and bottom AlGaAs DBR mirrors and two oxide aperture layers; and (b) part of a colorized LIV mapping of the maximum optical output power of a processed VCSEL quarter wafer piece, where the center region marked by thick horozontal white lines is a unit cell with 16 rows (rows 0 to F) ×15 columns (columns 0 to E). The black rectangles are either off the VCSEL quarter wafer piece (you can see the curved edge on the lower left) or nonfunctioning VCSELs.
Fig. 2.
Fig. 2. Experimental test setup for the emission spectra and the small signal modulation frequency response (2-port S21 parameter) measurements for the 980 nm VCSELs.
Fig. 3.
Fig. 3. Room temperature static characteristics of the VCSELs with a top mesa diameter from 18.5 to 31 µm and corresponding bottom mesa diameters of 78.5 to 91 µm (column E) and 58.5 to 71 µm (column D): (a) LIV from row 1 to row F in column E; (b) LIV from row 1 to row F in column D.
Fig. 4.
Fig. 4. Room temperature static characteristics of the VCSELs: (a) threshold current and threshold current density as a function of oxide aperture diameter; and (b) maximum optical output power and maximum power conversion efficiency versus oxide aperture diameter from ø ∼2.5 to 15 µm.
Fig. 5.
Fig. 5. Room temperature peak emission wavelength (for the fundamental LP01 optical mode) as a function of bias current for 980 nm VCSELs with oxide aperture diameters from ø ∼2.5 to 12 µm. The solid shapes are column D results while the open shapes are column E results.
Fig. 6.
Fig. 6. Room temperature emission spectra for VCSELs with: (a) ø ∼3 µm (row 2, column D; blue curve) and ø ∼11 µm (row B, column D; red curve) at bias currents of 1.9 and 2 mA, respectively; and (b) ø ∼3 µm (row 2, column E; blue curve) and ø ∼15 µm (row F, column E; red curve) at a bias current of 2 mA.
Fig. 7.
Fig. 7. Small signal modulation response (|S21|) versus frequency at different CW bias currents for: (top) a VCSEL with ø ∼3 µm; and (bottom) a VCSEL with ø ∼15 µm. The black curves are the raw |S21| data smoothed and corrected for the photodiode response. The red curves are curve fits to the corrected |S21| data. For ease of viewing we show only the raw data for one VCSEL in each graph.
Fig. 8.
Fig. 8. The -3 dB small-signal modulation bandwidth (f3dB) for CW bias in the range I = 3 to 13 mA as a function of the oxide aperture diameter for 980 nm VCSELs of column E at 25 °C. Linear line fits are overlain on the measurements.
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.