Open Access
Add to BibSonomy
Abstract
With the rapid development of coherent optical communication systems, lasers are required to have higher power and narrower linewidths. In this paper, a process method using buried heterojunction is proposed. By comparing the optoelectronic properties of lasers with different mesa widths and different cavity lengths, the laser with a 1000 µm cavity length and a mesa width of 2.4 µm can reach 133.7 mW at 25°C at 300 mA, and a side mode suppression ratio (SMSR) greater than 50 dB. This laser also exhibits low relative intensity noise (${\lt}-\! {150}\;{\rm dB/Hz}$ at 260 mA) and narrow linewidth (${\lt}{200}\;{\rm kHz}$).
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. INTRODUCTION
Wavelength division multiplexing (WDM) is a growing market demand for increasing the bandwidth of coherent optical communication systems. Using WDM technology, multiple wavelengths (usually in the 1550 nm C-band) can simultaneously multiplex signals through fiber [1,2]. Due to the advantages of low attenuation, C-band 1550 nm lasers are widely used for long-distance optical links. To distribute the signal over long distances (${\gt}\!{100}\;{\rm km}$) from one node to a large number of households, an adequate power budget, narrow spectral linewidth, and low relative intensity noise (RIN) are required. Compared with solid and optical fiber narrow linewidth lasers, semiconductor narrow linewidth lasers have the advantages of small size, high efficiency, a simple driving circuit, and low cost. They are especially suitable for coherent optical communication systems.
At present, the methods of compressing the linewidth of semiconductor lasers are mainly divided into the internal cavity optical feedback method and the external cavity optical feedback method. The internal cavity optical feedback technology includes distributed feedback (DFB) technology, distributed Bragg reflection technology, and coupled cavity technology. External cavity optical feedback technology includes grating feedback technology, waveguide feedback technology, and Fabry–Perot (FP) filter negative feedback technology. Among them, many teams have achieved results in the internal cavity optical feedback method. In 2011, the French III-V laboratory obtained a DFB structure designed with a large optical cavity. The side mode suppression ratio (SMSR) is greater than 50 dB, and the Lorentz linewidth is 200 kHz [3]. The Tamper University of Technology in Finland uses third-order surface grating combined with the ridge waveguide structure technology to improve the SMSR of the device and achieve a laser output of less than 250 kHz [4]. The research group of Professor He Jianjun of Zhejiang University used deep etching technology to study the influence of half-wave coupling grooves and quarter-wave coupling on the output signal. They proposed a theoretical method for calculating half-wave coupling groove with loss. The laser output with 80 kHz linewidth and 38 dB SMSR was obtained at 1550 nm band [5]. The external cavity optical feedback method has a lower linewidth. In 2016, the Australian National University research group used a gain chip with oblique light output and a blazed grating to form a Littrow external cavity laser (ECL), which achieved an output power of 300 mW and a Lorentz linewidth of 4.2 kHz at 1080 nm band [6]. In 2017, the research group of NTT Corporation in Japan conducted research on the narrow linewidth laser with an external waveguide, using the ${{\rm SiO}_2}$ planar optical waveguide (PLC) chip as the external cavity feedback optical path and compressing the linewidth of the DFB chip integrated with semiconductor optical amplifier (SOA). Using a waveguide with a length of 129 mm, a limited linewidth of less than 10 kHz was finally obtained by optimizing the coupling efficiency, and the output band was 1550 nm [7]. In 2018, the NTT completed the optimization of the coupling system, using an FP filter with a resolution of 160 MHz to compress the DFB laser signal with a linewidth of about 13 MHz to 3 kHz [8]. However, the external cavity optical feedback laser has a high cost, large size, and poor stability. It is not suitable for mass production. In the internal cavity optical feedback method, the process flow of the discrete mode laser structure and coupled cavity structure is complex, and the process conditions are harsh, and it is also not suitable for mass production.
In this paper, we propose a high-power 130 mW at 300 mA laser with a buried heterojunction structure with a RIN value below ${-}{150}\;{\rm dB/Hz}$ and a linewidth below 200 kHz. We show the power variation for different cavity lengths and mesa widths. Tunability can be obtained simply by changing the bias current and chip temperature. We also show the effect of bias current on linewidth and RIN.
2. EXPERIMENT
Active regions consisting of multiple quantum wells (MQWs) and separate confinement heterostructure (SCH) layers are grown using metal–organic chemical vapor deposition (MOCVD). The MQW region consists of three 5-nm-thick InGaAsP quantum wells and four 8-nm-thick InGaAsP barriers, and the photoluminescence (PL) of the MQW is in the range of ${1550}\;{\rm nm}\;{\pm}\;{5}\;{\rm nm}$. To form a single-mode laser, a $p$-grating layer with a thickness of 10 nm was grown.
The traditional ridge waveguide structure is mainly vertical for the confinement of carriers and photons. In the horizontal direction, the carriers can diffuse freely and even recombine before contributing to the gain. The buried heterostructure is used to confine the horizontal current, carriers, and photons through the reverse PN junction, which makes the optical field distribution more concentrated. The specific process flow is as follows. The active area is wet-etched to form the mesa structure. The mesa upper stage width is designed with three widths: A, 1.4 µm; B, 1.9 µm; and C, 2.4 µm. Subsequently, $p \!-\! {\rm InP}$ and $n\! -\! {\rm InP}$ barrier layers are grown, followed by $p\! -\! {\rm InP}$ and $p \!-\! {\rm InGaAs}$ layers to form a buried heterostructure (BH). Finally, Ti/Pt/Au was deposited on the $p \!-\! {\rm InGaAs}$ contact layer and annealed at 420°C to form $p$-metal. Figure 1(a) shows the appearance of a chip with a length of 1000 nm. The number ABC starts with three mesa-width structures, and Fig. 1(b) shows the mesa structure.
The influencing factors of the linewidth of the semiconductor laser can be determined by the modified Schawlow–Townes equation [9]:
Fig. 2. (a) PI curves of lasers with different mesa widths with a cavity length of 1000 µm. (b) PI curves of lasers with different cavity lengths with a mesa width of 2.4 µm.
These bars are coated with anti-reflective (AR) and high-reflective (HR) coatings, tested, screened, and cut into chips. Then we tested the bar to understand the chip threshold, power, oblique efficiency, and other characteristics. The chip is made into chip on carrier (COC) to test the change of chip wavelength with temperature and current, and the wavelength drift coefficient with temperature and the wavelength drift coefficient with current are obtained; parameters such as linewidth, RIN, and divergence angle can also be obtained by testing COC.
3. RESULTS AND DISCUSSION
After testing the bar, the power-current (PI) curves of three mesa widths are shown in Fig. 2(a). It can be found that, with the increase of the chip mesa width, the threshold will rise, and the slope efficiency will also increase. At 25°C and 300 mA, the chip with a mesa width of 2.4 µm can achieve a maximum power of 133.7 mW. According to the research of Henry et al. [9], when the structure is certain, the mesa width and the threshold have a certain linear relationship, which is roughly in line with our experimental results. This is because the composite term at the threshold is mainly spontaneous coincidence, which is determined by Eq. (2), and can also be shown as
where ${\eta _i}$ is the internal quantum efficiency. Figure 2(b) shows the PI curves corresponding to different cavity lengths at 25°C, 300 mA and the mesa width of 2.4 µm. It can be found that, with the increase of the cavity length, the threshold value will increase, the slope efficiency will decrease, and the saturation current point will increase. At 300 mA, the power of the 1000 µm cavity length is the largest. Combined with the practical application, the chip with a 1000 µm cavity length and a mesa width of 2.4 µm has the best electrical characteristics, so the subsequent test projects are all using this chip.Figure 3(a) shows the change of wavelength with current. When the current increases from 100 mA to 500 mA at 25°C, the wavelength increases from 1539 nm to 1540.7 nm, and the SMSR is greater than 50 dB. By fitting the peak wavelength, the wavelength drift coefficient with current is 0.005 nm/mA. Figure 3(b) shows the change of wavelength with temperature. At 300 mA, when the temperature rises from 25°C to 70°C, the wavelength rises from 1539.7 nm to 1544.1 nm, and the SMSR is greater than 47 dB. By fitting the peak wavelength, the wavelength drift coefficient with temperature is 0.09 nm/°C. It shows that wavelength tunability can be obtained simply by changing the bias current and die temperature. Figure 4 shows the optical far-field characteristics of the laser. The full width at half-maximum divergence angle at room temperature of 500 mA is ${20} \times {30}^\circ$.
Fig. 3. (a) Wavelength changes with current at 25°C. (b) Wavelength changes with temperature at 300 mA.
Fig. 4. Horizontal divergence angle of the laser far-field curve at room temperature 500 mA is 20.63°; vertical divergence angle is 30.57°.
Laser phase noise power spectral density (PSD) ${S_p}(f\;)$ is a term used to quantify the laser phase noise power distribution over frequency and can be equivalently described by the laser frequency noise PSD ${S_F}(f\;)$. It is common that tunable laser manufacturers measure ${\rm SF}(f\;)$ to ensure the laser frequency pureness, with the understanding that ${S_F}(f\;)$ and ${S_p}(f\;)$ are related by Eq. (3),
where the unit of ${S_p}(f\;)$ is ${{\rm Hz}^{- 1}}$, and the unit of ${S_F}(f\;)$ is ${{\rm Hz}^2}/{\rm Hz}$. We tested the chip laser frequency noise PSD(${S_F}(f\;)$) with a mesa width of 2.4 µm and a cavity length of 1000 µm, as shown in Fig. 5(a). There are three main frequency components in the graph: (1) the region of white noise above ${\sim}{10}\;{\rm MHz}$; (2) the region of non-flat $1/f$ noise (sometimes called “flicker noise”) between 10 kHz and 10 MHz [10]; (3) the multiple interference tones spread across the frequency range below 80 MHz. The multiple interfering tones are caused by switching power supplies and various circuits in a coherent transceiver module. The white frequency noise PSD originates from spontaneous emission noise and is directly related to the conventional term of “laser linewidth” as shown in the following equation: linewidth = white frequency noise ${\rm PSD}\;*\;\pi$. In other words, a narrower laser linewidth corresponds to a lower white frequency noise level. We characterize the linewidth of the laser by taking the average value of the relatively stable 70–80 MHz white frequency noise PSD, as shown in the inset of Fig. 5(a). The change of the linewidth with the bias current is shown in Fig. 5(b), and it can be found that the minimum linewidth is 194 kHz under the room temperature of 260 mA.We also measured the RIN of the laser at different bias currents. In testing RIN, we set resolution bandwidth (RBW) to 3 MHz and video bandwidth (VBW) to 300 KHz. The first peak value after stabilization was used as the test result of RIN value. Figure 6 shows the RIN values for bias currents from 60 mA to 460 mA at 25°C. At 60 mA, the RIN curve peaks around 3.8 GHz, which corresponds to the resonant frequency. At higher bias currents, the resonant frequency increases. At high bias currents, the RIN relaxation oscillation peak is also greatly suppressed. From 60 mA to 460 mA, the relaxation peak level dropped from ${-}{134.7}\;{\rm dB/Hz}$ to ${-}{158.5}\;{\rm dB/Hz}$ and moved from 3.82 GHz to 11.4 GHz. As the bias current increases, the RIN relaxation oscillation peak decreases slowly after 260 mA to ${-}{155}\;{\rm dB/Hz}$. This is because RIN mainly depends on the photon density in the cavity [11]; while the optical power approaches saturation as the current increases, the photon density also varies slowly.
4. CONCLUSION
Our proposed DFB laser with a MQW buried heterojunction structure has a cavity length of 1000 µm and a mesa width of 2.4 µm. It has high power, low RIN, and narrow linewidth. The manufacturing method is low in cost, simple in process, and capable of mass production. It can be used in coherent optical communication systems, and it has power greater than 130 mW at 300 mA and 25°C. Wavelength tunability can be obtained simply by changing the bias current and chip temperature. At 25°C and 260 mA, the linewidth is less than 200 kHz, and the RIN is lower than ${-}{155}\;{\rm dB/Hz}$.
Funding
International Science and Technology Cooperation Key Research and Development Program of Science and Technology Agency in Hubei Province (2021EHB018); Optics Valley Science and Technology Innovation Corridor (2021BGE005).
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. B. A. Saleh and M. C. Teich, “Fiber-optic communications,” in Fundamentals of Photonics, 1st ed. (Wiley, 1991), pp. 874–917.
2. R. T. Sahara, R. A. Salvatore, A. Hohl-Abichedid, and H. Lu, “Isolator-free transmission at 2.5 Gbits/s over 100 km of single-mode fiber by a 1.55-/spl mu/m, AlGaInAs strained-multi-quantum-well, directly modulated distributed-feedback laser diode,” IEEE J. Quantum Electron. 38, 620–625 (2002). [CrossRef]
3. C. Cayron, M. Tran, Y. Robert, M. Lecomte, M. Calligaro, O. Parillaud, M. Garcia, and M. Krakowski, “High power distributed feedback and Fabry-Perot al-free laser diodes at 780nm for rubidium pumping,” Proc. SPIE 7953, 79530A (2011). [CrossRef]
4. H. Virtanen, A. T. Aho, J. Viheriälä, V. M. Korpijärvi, T. Uusitalo, M. Koskinen, M. Dumitrescu, and M. Guina, “Spectral characteristics of narrow-linewidth high-power 1180 nm DBR laser with surface gratings,” IEEE Photon. Technol. Lett. 29, 114–117 (2017). [CrossRef]
5. Y. Wang, Y. Yang, S. Zhang, L. Wang, and J. He, “Narrow linewidth single-mode slotted Fabry–Pérot laser using deep etched trenches,” IEEE Photon. Technol. Lett. 24, 1233–1235 (2012). [CrossRef]
6. D. K. Shin, B. M. Henson, R. I. Khakimov, J. A. Ross, C. J. Dedman, S. S. Hodgman, K. G. H. Baldwin, and A. G. Truscott, “Widely tunable, narrow linewidth external-cavity gain chip laser for spectroscopy between 1.0–1.1 µm,” Opt. Express 24, 27403–27414 (2016). [CrossRef]
7. H. Ishii, N. Fujiwara, and K. Watanabe, “Narrow linewidth tunable DFB laser array integrated with optical feedback planar lightwave circuit (PLC),” IEEE J. Sel. Top. Quantum Electron. 23, 1501007 (2017). [CrossRef]
8. K. Aoyama, N. Yokota, and H. Yasaka, “3-kHz spectral linewidth laser assembly with coherent optical negative feedback,” IEEE Photon. Technol. Lett. 30, 277–280 (2018). [CrossRef]
9. C. Henry, R. Logan, and F. Merritt, “Single mode operation of buried heterostructure lasers by loss stabilization,” IEEE J. Quantum Electron. 17, 2196–2204 (1981). [CrossRef]
10. R. Zhang, K. Kuzmin, W.-J. Jiang, G. Giaretta, T. Tomimoto, Y. Weng, G.-K. Chang, and W. Way, “Impact of laser flicker noise and linewidth on 64 to 96 Gbaud/DP-nQAM metro coherent optical links,” Opt. Lett. 45, 1220–1223 (2020). [CrossRef]
11. M. Faugeron, M. Tran, O. Parillaud, M. Chtioui, Y. Robert, E. Vinet, A. Enard, J. Jacquet, and F. Van Dijk, “High-power tunable dilute mode DFB laser with low RIN and narrow linewidth,” IEEE Photon. Technol. Lett. 25, 7–10 (2013). [CrossRef]
