850 nm VCSEL with sub quantum well and p-type δ-doping in the active layers for improved high-speed and high-temperature performance

The rapid growth of high performance computing data centers largely depend on the development of high-speed optical interconnects (OI). The current data centers have reach total data rates of up to 400 Gbit/sec based on the 8 × 50 Gbit/sec array transmitter according to the 400GBASE-SR8 specification in IEEE 802.3 cm [1]. The vertical cavity surface emitting laser (VCSEL) has long been recognized as the ideal directly modulated light source for OI because of their small size, low energy consumption and economical fabrication cost  [2], [3], [4], [5], [6]. However, due to the heat generated from laser driver integrated circuits (ICs) and high ambient temperature, the transmission speed and lifetime of VCSEL usually degrades [7], [8]. Thus, it is vital to develop a VCSEL with nearly invariant high-speed and long-life cycle under high temperature operation.

To enhance the modulation bandwidth and bit rate of VCSEL, the focus has been on optimizing the differential gain (∂g/∂n), electrical parasitic, and self-heating etc. [9], [10], [11]. Typical solution to improve the ∂g/∂n was incorporated InGaAs strained multiple quantum wells (MQWs) in 900–1100 nm VCSELs [12], [13]. But it is not feasible to use such a large indium (In) mole fraction (>30%) for the 850 nm VCSEL. Another effective approach is adding p-type doping in quantum well (QW), which could further improve the ∂g/∂n and minimize the induced gain compression at high-current [14], [15], [16], [17], [18]. Nevertheless, high-level p-type doping in QWs would lead to an increase in the free-carrier absorption loss in active layers and degrade laser’s performance. So, the position and concentration of p-type doping should be considered comprehensively. Multi-layer-oxide structures are usually used in high-speed VCSELs, which could reach upper than 50 Gbit/s data rate under non-return-to-zero on-off keying (NRZ-OOK) transmission  [19], [20]. However, small oxide aperture typically accumulates large thermal in the devices, reducing device lifetime and modulation bandwidth. Thus, for further improving the bandwidth of VCSEL, it is necessary to maintaining high ∂g/∂n and temperature stability meanwhile.

Recently, composite QWs with single or multiple functional nanolayers have been widely studied because of their unique electron transport and quantum characteristic [21], [22]. This composite QW is broadly used in high electron mobility transistor (HEMT) [23], ultra-high-speed switching devices etc [24]. The composite functional nanolayers could change the shape of the QW potential deforms and the transition energies of the sublevels, which could rise tunnel penetration effect and decreases temperature sensitivity of Hall electron concentration [23]. In 2020, Ajia et al. conducted experimental study on sub quantum well (sub QW) AlGaN/AlGaN LED. Through inserts different nanolayers in QW or quantum barrier (QB), the efficiency and radiative recombination could be enhanced in AlGaN/AlGaN LED, and it also exhibits good temperature stability [25]. This method provides a new idea in composite QW in optoelectronic devices. As far as we know, few works have been devoted to the performance of inserts nanolayers in InGaAs QW laser. Especially, utilizing inserts nanolayers in the QWs of high-speed VCSEL.

In this paper, we propose a composite MQWs with p-type δ-doping and sub QW insert layer in high-speed 850 nm VCSEL structure and compare this composite MQWs structure with the other three basic MQWs in VCSELs. These MQWs modifications are assessed in terms of their potential for reducing the thermal rollover in output performance, and for increasing the maximum intrinsic modulation bandwidth of the laser. Since thermal rollover is a primary negative factor in limiting bandwidth and output performance in the multi-oxide layer VCSEL, we used the VCSEL structure with six oxide layers in this paper to study the effect of composite MQWs on thermal rollover. The simulation dynamic/static performance of those structures are systemically studied through Crosslight PICS3D software [26]. Results show that this novel MQWs design could improve the output performance and 3-dB bandwidth of VCSEL in 25 °C and 85 °C operation. Especially, the sub QW insert layer in QW could significant delay the thermal rollover of VCSEL at high current condition. Furthermore, we studied the proposed four MQWs by molecular beam epitaxy (MBE) system. Temperature-dependent optical analyses show that the novel MQWs have strong radiative recombination rate and outstanding photoluminescence (PL) intensity, which is ideal MQWs in high-speed VCSEL. In summary, this research proposed a new method for the active region design of optoelectronic devices in special environments with tremendous advantages in monolithic laser/transistor integration.

留言 (0)

沒有登入
gif