The concept of diode lasers was born early in the age of laser technology and has come a long way. From a fragile, liquid nitrogen-cooled laboratory device, it has evolved into an integrated, highly energy-efficient laser system. Today, laser diodes are the fundamental solution in most laser materials processing devices. They also drive the internet, enable quantum research, and are a prerequisite for any laser fusion scheme. Diode lasers have been a central technology to several Nobel Prizes and even experienced their own technology bubble when the internet boom swept stock markets around the year 2000. Courtesy of Noda Lab. One might think of such a technology — whose products are sold as commodities — as mature. To some extent, this is true of diode lasers. On the other hand, there has always been a quest for improvement. The simplest approach looks to achieve more power from an emitter at a lower cost per watt. This has led, and continues to lead, to the development of ever-better edge-emitting laser (EEL) diodes. Still, it is well known that these lasers require external focusing, which makes packaging and alignment complex. As a result, efforts to overcome the poor beam characteristics of laser diodes have been on the rise. The concept of vertical-cavity surface emitting laser (VCSEL) diodes, for example, evolved out of this desire. Unfortunately, this concept is power limited. Diodes with special spectral emission characteristics have been developed as well. These DFB lasers achieve distributed feedback from a complex layered structure inside the laser diode’s cavity, enabling them to emit stabilized radiation with a small spectral bandwidth. A laser diode setup that combines these three optimization efforts may offer the best of all worlds: ease of manufacturability, high power, and an effective beam profile. Edge-emitting and vertical-cavity lasers For comprehensive consideration of new diode lasers, one must look closely at the two main types of diode lasers: EELs and VCSELs. These laser types differ fundamentally in their structure, emission direction, and performance characteristics (Figure 1). Figure 1. A comparison of diode lasers: edge-emitting lasers (EELs) (left) and vertical-cavity surface emitting lasers (VCSELs). Courtesy The Photon, CC BY-SA 3.0, www.commons.wikimedia.org/wiki/File:Simple_sch_laser_diode.svg, and www.commons.wikimedia.org/wiki/File:Simple_vcsel.svg, via Wikimedia Commons. A typical EEL consists of low-bandgap-material layers sandwiched between two high-bandgap semiconductor layers. The ends of the horizontal (Fabry-Pérot) resonator are formed by cleaved facets. The laser beam propagates along the plane of the active region in the middle and exits through one of the facets. This architecture enables high output power, efficient heat dissipation along the length of the device, and broad wavelength tunability. However, EELs need precise facet cleaving, which is a requirement that makes manufacturing complex. Their emission exhibits elliptical beam divergence, necessitating external optics for beam shaping. And packaging and alignment can be challenging, especially for high-power applications. In contrast, VCSELs emit light perpendicular to the wafer surface. Their structure relies on multiple layers of alternating high- and low-refractive-index materials forming distributed Bragg reflectors (DBRs), which create a vertical optical cavity. The active region, typically a quantum well or a quantum dot layer, is placed between these reflectors. The DBR typically defines a small emission spectrum. The VCSEL structure allows for on-wafer testing and easy integration into 2D laser arrays, making VCSELs highly scalable for mass production. They also offer circular beam profiles with low divergence, which simplifies optical coupling. Despite these advantages, VCSELs have limitations. Their output power is lower than that of EELs, and their spectral tuning range is relatively narrow due to their reliance on the DBRs. Heat dissipation can also be a challenge, as the current flows vertically through the device. As VCSELs dominate applications requiring low-power, compact, and high-efficiency light sources — such as optical fiber data transmission and 3D sensing — EELs remain the preferred choice for high-power applications such as materials processing, either directly or as the energy source for industrial fiber lasers. Photonic-crystal surface-emitting lasers Members of the Kyoto, Japan-based research group led by Susumu Noda have investigated diode laser setups for several decades and developed a horizontal resonator with vertical outcoupling. They added a 2D structure to the active layer that allows the laser to emit vertically. This structure consists of a periodic pattern of gaps, or holes, called a photonic crystal. Accordingly, these diode lasers are named photonic-crystal surface-emitting lasers (PCSELs, pronounced “pick-cells”, Figure 2). Figure 2. A typical photonic-crystal surface-emitting laser (PCSEL) has an active horizontal layer next to the photonic crystal layer in the center. The horizontal resonator emits vertically to the top and has electrodes on the bottom and the top. A recent article in the journal Nature Reviews Electrical Engineering explains the process in more detail1: “When the current is injected into the active layer through the cladding layers, light is generated within the active layer and is trapped between the active and cladding layers by total internal reflection while also being coupled to the photonic crystal layer. Inside the photonic crystal, light waves propagating in several in-plane directions are coupled with each other, leading to 2D coherent lasing with surface emission.” In other words, in a PCSEL, the optical waves generated by stimulated emission travel parallel to the layers. Feedback is provided by a 2D structure in a layer next to the active layer with a certain periodicity. Crucially, this structure allows the area of the laser to be scaled as well as the output power, all while maintaining high-quality single-mode emission in contrast to EELs and VCSELs. In fact, the periodicity corresponds to second-order Bragg diffraction for wavelengths around the peak of the gain spectrum. Therefore, a part of the optical field is radiated vertically. As in VCSELs, the beam shape in a PCSEL can be circular, with an ultra-narrow divergence angle. A single DBR, such as in a VCSEL, can be implemented beneath the periodic layer to force the outcoupling from one surface. The effect of the photonic crystal on the modes oscillating in the resonator depends on the shape and distribution of photonic crystal features, and on the index contrast between the feature and the surrounding material. For the PCSELs made in Kyoto, and most PCSELs worldwide, the photonic crystal structure consists of air voids embedded in the surrounding semiconductor material, which has a very large index contrast. Alternatively, the photonic crystal can be “all-semiconductor” where the index contrast between different semiconductor materials is much smaller. A number of players are exploring this technological path for more manufacturable PCSELs, including groups at Germany’s Ferdinand-Braun-Institut and at the Glasgow, Scotland-based startup Vector Photonics. The industrial case The Noda group has worked on PCSELs for more than two decades. In November 2024, a workshop at Aston University in Birmingham, England, showed that the interest in this technology has grown far beyond this group. One hundred twenty people attended, including a strong Japanese delegation. Several companies, including Huawei Technologies (via its Ipswich, England, research center), Vector Photonics, nLight, TRUMPF, Lumentum, Mitsubishi Electric, and ASML sent representatives, according to coverage of the workshop2. The Noda group has reported PCSELs suitable for a number of industrial operations. Its reported brightness of 1 GW cm−2 sr−1 for pulsed and continuous-wave operation is a record for laser diodes and could cut metal1. In its paper, the group referred to PCSELs with short-pulse (100 W) operation, as well as short-wavelength (~430 nm) operation1. Vector Photonics, a Scottish startup developing PCSELs, claims that the cost for PCSELs is 50% that of EELs while delivering 10× the power of VCSELs. Vector mentiond another advantage of PCSELs: They can be used to emit a matrix of coherent beams3. In this case, one horizontal resonator would be connected to a matrix of vertical beam outlets. All the beams in this matrix would come from the same modes in the resonator and be inherently coherent as a result. This is an optical phased array in which all beams can be coherently combined. With distributed phase delays, such combined beams can have almost arbitrary shape, and even steer aside depending on the respective phase delay. Such a phased array would be well suited for lidar applications without moving parts, or for materials processing with beam profiles that may change during the process. At the same time, PCSELs can be controlled with multi-gigahertz rates — a quality that would make them ideal for data applications. Future prospects Regarding the future options for PCSEL technology, Noda and his co-authors said, “[Continuous-wave] kilowatt-class operation is expected to become possible by expanding the laser diameter from 3 to 10 mm and optimizing Hermitian and non-Hermitian coupling coefficients, which will enable a further tenfold to hundredfold improvement in the output power and beam brightness.1” All things considered, PCSELs have a number of promising properties already. They unite potentially low manufacturing costs with high output power and high beam quality, and in a circular beam. Acknowledgment The author wishes to thank Ben King of Aston University for valuable discussions supporting the development of this article. References 1. S. Noda et al. (2024). Photonic-crystal surface-emitting lasers. Nat Rev Electr Eng, Vol. 1, pp. 802-814. 2. R. Stevenson. (2024). Promoting the PCSEL. Compound Semiconductor, Vol. 30, No. 9, pp. 30-34. 3. Vector Photonics. The semiconductor laser revolution, www.vectorphotonics.co.uk/technology.