iStock.com/NeoLeo The emergence of 3D sensing in the consumer, industrial, and automotive markets has led to a second significant boom for VCSELs, following their initial use in high-bandwidth data communications. Today, VCSELs can be found in an increasingly broad range of applications, including wearables, medical, security, augmented and virtual reality, drones, logistics, robotics, industrial safety, passenger monitoring, gesture recognition, and lidar. VCSEL-based imaging has evolved beyond stereo vision to include structured light and time of flight, along with hybrid utilizations. Stereo vision involves using two cameras with a known separation between them to capture images from different angles and construct a 3D image. The structured-light approach projects a known pattern of light dots, which converts the observed distortion into information about the third dimension of the illuminated object. Time of flight relies on measuring the round-trip transit time or phase shift of a laser pulse. VCSEL technology is used in many applications in the industrial, consumer, medical, and automotive markets. Courtesy of ams OSRAM. Each method uses VCSEL illumination over a wide range of operating conditions, including peak optical power and pulse parameters. Optimizing laser output power and efficiency, as well as beam divergence, is essential in terms of VCSEL performance. In addition, the die design and integration of high-efficiency VCSELs into user-specific modules will also be key to the lasers’ future success in enabling a wide variety of applications. VCSELs are frequently combined into modules that include optics and/or drivers to create the desired illumination profiles. Innovative methods in VCSEL design and integration can improve packaging to enhance both device footprint and laser performance. Boosting power output While maximizing laser efficiency has always been a goal for VCSEL developers, 3D sensing has been pushing against the limits of peak power density available in VCSEL sources. Traditionally, to reach the power levels required for 3D sensing, a larger array of VCSEL apertures was required. Thermal crosstalk between individual apertures limited the VCSEL’s peak efficiency and power density, leading to a trade-off between aperture pitch and quantity in the array. A standard 1-sq-mm VCSEL die can produce a usable peak power of 3 W for a relatively long pulse width (>1 ms) up to 80 °C. Although a larger pitch and aperture quantity can improve the total power output at elevated temperatures, the improvement comes at the cost of a larger chip and ultimately leads to higher production costs, without adding significant improvements to the VCSEL power density. Trade-offs need to be made when designing die size to prevent thermal degradation. While VCSELs are primarily used in 3D sensing in automotive and mobile applications, they are also enabling both similar and complementary roles in the industrial and medical sectors. Robotics and logistics require advanced automation, involving proximity sensing and object measurement methods that utilize VCSELs to improve signal integrity. The narrow linewidth helps to reject unwanted ambient noise, with the assistance of a bandpass filter. The filter targets narrow and specific absorption wavelengths to detect elements and their properties. Narrow linewidth is beneficial in both industrial and medical uses, as it can measure energy transfer with high control, enabling atomic sensing, light therapy, and industrial heating functions with its high degree of spectrum selectivity. The technical approaches to VCSEL-based imaging have diversified to include stereo vision, structured light, and time-of-flight methods. EEL: edge-emitting laser; IRED: infrared LED. Courtesy of ams OSRAM. Groundbreaking performance improvements require reinventing the VCSEL design and epitaxial growth process via the development of multijunction VCSELs. Utilizing multiple active regions separated by tunnel junctions increases the lasing gain and reduces current requirements inside the VCSEL cavity. Optimization in the multijunction structure designs results in significant boosts in both efficiency and output power density, powering high-speed and other demanding applications. A multijunction VCSEL incorporates highly doped tunnel junctions sandwiched between separated multiple quantum well light-emission regions. By optimally controlling the tunnel junction doping, current can pass through the junctions with low resistance. Light is generated by carrier recombination within each multiple quantum well, thus effectively expanding the gain region inside the laser cavity. The tunnel junctions must be positioned at the internal electric field minimum to reduce unwanted absorption inside the cavity. The required voltage must overcome the forward bias for each multiple quantum well and the tunnel junctions’ reverse bias to enable current flow through the laser. After reaching the turn-on voltage, the slope of voltage over current is nearly the same for all three VCSEL types, indicating that the increasing junctions add little to no additional series resistance. Results show record efficiency measurements of 60% for three junctions and 61.5% for five junctions. Furthermore, high efficiency (>50%) is achievable for a broader range of driving conditions. VCSELs with multiple junctions exhibit an increase in slope efficiency and require significantly less current to obtain target power criteria. Reducing forward current requirements diminishes unwanted internal heating in electrical resistance as a result and improves VCSEL performance over temperature. The improvement in slope efficiency also has a significant impact on pulse rise time for high-speed applications. VCSELs with multiple junctions can achieve a faster rise time (dI/dt) to reach a target peak power, since the current required to overcome the parasitic inductance is significantly reduced. The slope efficiency is approximately proportional to the number of junctions, so that higher current and power densities are possible in a narrower pulse width. Multijunction VCSEL technology is also beneficial in the development of long-distance automotive lidar because VCSELs with a high quantity of junctions can improve both pulse rise time and output power density. The development of five-junction 905-nm VCSELs demonstrated that the multijunction technology could be incorporated into VCSELs under various operating wavelengths. Experimental results with five-junction 905-nm VCSELs have shown peak power densities >2 kW/mm2 at 105 °C and a decrease of An illustration of a multijunction VCSEL. P-DBR/N-DBR: P- and N-type distributed Bragg reflectors; MQW: multiple quantum well; TJ: tunnel junction. Courtesy of ams OSRAM. The development of an illumination module requires more than a high-performing VCSEL die. The light emitted from the VCSEL must be shaped into a profile that properly illuminates the target field of illumination. The first generation of VCSELs integrated with surface-mount device (SMD) packaging involved the incorporation of beam- shaping optics into the module. Next-generation packaging incorporates a driver chip, which both enables miniaturization and reduces parasitic inductance for high-speed and time-of-flight applications. Future generations of VCSEL integration are targeting improvements in enhanced field of illumination, miniaturized packaging, and faster rise times. Surface-mount device (SMD) packaging for VCSELs can incorporate beam-shaping optics (a, b) and improved electrical integration (c) designed for various applications. Courtesy of ams OSRAM. VCSEL dies can be designed to improve packaging requirements and capabilities available to the laser integrator. One method for improving performance is by using segmented and addressable arrays. Traditional VCSEL dies are binary light sources, and thus, multiple laser illumination sources would require an appropriate quantity of individual laser dies. Alternatively, the VCSEL die can be segmented and paired with appropriate optics to sequentially illuminate various segments in the field of view to compensate for the limited size of detector arrays and reduce glare from high- reflectivity objects in detected regions. Segmenting the VCSEL die into addressable units enables the VCSELs to make field-of-illumination adjustments to improve illumination and sensing control. If each segment can target a subunit in the field of illumination, each segment can be adjusted to control signal intensity, improve module efficiency, and reduce unwanted glare from the illumination target. A segmented VCSEL with a custom aperture array can be combined with projection optics and diffractive optical elements to build a flexible dot pattern for advanced structured-light algorithms. VCSELs can also incorporate segments for various functionalities to combine multiple sensing features into a single package. As an example, low-power and high-power segments can be incorporated together on the same die to combine proximity sensing and a flood illumination, respectively. More can be placed on a chip The next level of integration is the development of flip-chip VCSELs that can be bump bonded onto a silicon substrate. This design reduces the footprint compared to top-emitting VCSELs that require additional space for wire bonding. Flip-chip VCSELs can be bonded in closer proximity to or directly on top of a laser driver chip to further reduce both package size and parasitic inductance. Furthermore, optical components can be integrated into the back-emitting side of the chip to further reduce the module’s size. Illustrations of a VCSEL designed for binary operation (a), 1D line addressability (b), and segmentation that incorporates proximity and 2D flood illumination (c). Courtesy of ams OSRAM. To achieve a desired field of illumination, traditional top-emitting VCSELs require an external optical diffuser to shape the VCSEL output beam. This diffuser must be positioned at a minimum distance away from the VCSEL apertures to function properly. This solution needs to be robust in harsh environments to prevent eye-safety hazards when the diffuser is cracked or detached. Alternatively, an optical diffuser pattern can be deposited or etched onto a flip-chip VCSEL by using the gallium arsenide (GaAs) VCSEL substrate to control the spacing between VCSEL apertures and optical diffusers. Applying optics directly onto a flip-chip VCSEL die to expand and shape the output beam for wide illumination ensures that the source is eye safe without requiring a robust package solution. Top-emitting VCSELs require external optics for efficient beam shaping (left). Wafer-level optics can be built directly on flip-chip VCSELs for beam shaping (right). Courtesy of ams OSRAM. The design incorporated both cathode and anode contacts with bump bonds on the original top surface of the wafer, along with solder bumps for a subsequent solder reflow attachment to a submount. The VCSEL is designed to emit light down into the substrate side of the GaAs wafer, so that the die is flipped to achieve upward emission. Unless the substrate is removed from the wafer, the operating VCSEL wavelength must be longer than 900 nm to avoid excessive absorption from the GaAs substrate. Flip-chip VCSELs exhibit improved power density due to their unique design. Traditional top-emitting VCSELs are limited to 20-μm-diameter apertures due to the difficulties in spreading current across the open aperture. In contrast, the apertures in flip-chip VCSELs can be fully metalized, significantly improving uniform current spreading across larger-size apertures, with minimal performance degradation in apertures as large as 100 μm in diameter. The emergence of 3D sensing technology in the consumer, industrial, and automotive markets, as well as increasing demand for applications requiring this technology, has led to a second significant expansion of the market for VCSELs. The preferred performance characteristics of the devices for these applications is differentiated for each methodology, which has driven VCSEL technology innovation in many directions. The investment in multijunction VCSEL technology enables continued improvement in efficiency, peak powers, and pulse rise time at multiple operating wavelengths. With the advent of die segmentation, the VCSEL can also be designed for flexible and multifunctional laser illumination. Requests to increase pulse performance and enable further miniaturization have led to the emergence of flip-chip VCSEL technology. All of these known technological advancements can be integrated into high-performing, customized, next-generation VCSEL modules and are expected to be available for commercial applications in the next few years due to strong economic demand within the consumer, industrial, and automotive markets. Meet the author Kevin Kruse, Ph.D., is senior applications engineer at ams OSRAM. He assists customers on design and manufacturing, researches technical trends in the optoelectronics market, and presents technical trainings and conference proceedings. Kruse has worked at ams OSRAM for over five years, after obtaining his doctorate in electrical engineering from Michigan Technological University in 2015; email: kkruse@vixarinc.com.