VCSELs have become essential components for 3D sensing in smartphones ever since their integration in 2017 Apple’s flagship. While Apple’s competitors are now following the same trend, the explosion of VCSELs is just at its beginning as new growth drivers in the automotive applications are also expected to generate high volumes. VCSEL market revenue reached ~$330M in 2017 and Yole Développement (Yole) is expecting a CAGR of +48% for the next 5 years, as shown in the recently released report “VCSEL – Technology, Industry and Market Trends 2018”.
The current shift from the Datacom industrial era to the 3D sensing era is likely to have a significant impact on manufacturing expertise related to VCSELs. In this context Pierrick Boulay, Technology & Market Analyst, Solid-State Lighting at Yole had the opportunity to interview Stéphanie Baclet, Technical Writer at Oxford Instruments, to discuss the challenges related to VCSEL manufacturing.
Pierrick Boulay (PB): Can you please introduce yourself, your activities and the activities of Oxford Instruments, to our readers?
Stephanie Baclet (SB): My name is Stephanie Baclet and I work for Oxford Instruments Plasma Technology as a technical writer. I work closely with the optoelectronics device manufacturers to translate requirements of their devices into nanofabrication requirements for plasma processing products. I have worked as a Senior Application Engineer focused on new product introduction and developed processing techniques for various technologies, such as LED, Laser diodes and diffractive optical elements.
Oxford Instruments Plasma Technology is a leading provider of etching and deposition plasma process solutions for nanometre-sized features, nanolayers, and the controlled growth of nanostructures. These solutions are based on core technologies in plasma, ion beam and atomic layer deposition and etching. Products range from clustered cassette-to-cassette platforms for high-throughput production processing to compact stand-alone systems for R&D.
PlasmaPro100 Polaris MMX
(Source: Courtesy of Oxford Instruments)
PB: What are the main challenges related to the VCSEL manufacturing process? How do Oxford Instruments overcome them?
SB: The performance and characteristics of a device are always the result of multiple elements within its design and fabrication. For this reason, establishing the right process flow and executing each step with-in control tolerances is paramount for reliable manufacturing. When it comes to VCSEL, I would say that the aperture is a critical element of the design since it directly defines key parameters of the laser, such as the threshold voltage. Now because the size of the aperture is partly defined by the mesa diameter as well as the Al content and the Oxygen exposure, it takes each of these factors to be well controlled to achieve the desired aperture. Overall, these 3 elements are exactly where the main challenges of VCSEL manufacturing lies. VCSEL technology requires epi structures with well controlled Al content within the stack, reliable manufacturing of mesa structures and oxidation furnace with tight flow control.
As applications such as 3D sensing are driving the need for a high volume of VCSELs, there is a focus on maximizing the yield and moving the technology toward 150mm. Oxford Instruments supplies a solution for processing VCSEL mesas as well as several other laser elements. We have been working with VCSEL manufacturers for many years and our expertise in designing plasma processing solutions for III-V material has allowed us to produce a solution generating maximal yield at 100mm and 150mm wafer sizes. Additionally, the world of nanofabrication is evolving. Techniques such as atomic layer deposition and Atomic layer etching are enabling new device architectures and maximizing device performance. We are constantly evolving our solution portfolio to maximize the performance of our customers’ devices.
(Source: VCSELs – Technology, Industry and Market Trends, Yole Développement, July 2018)
PB: The etching step seems to be a critical step in the manufacturing process. Can you explain to our readers why?
SB: There are several aspects of the mesa etching process which are critical for the good behavior of a VCSEL. First and foremost is the quality of the sidewalls. The mesa sidewalls are where oxidation starts. You must have a smooth and clean surface to allow your aperture to form uniformly. This is a challenge since you also need to control the profile angle of your mesa which can generate roughness depending on your processing strategy. Additionally, for the standard GaAs/AlGaAs DBR structure, the Al present in AlGaAs can also introduce selective etching between layers of each pair. Another critical result of the etching step is how well you can define the end layer. Failing to stop the etch at the target end layer within your epi stack will lead to oxidation of undesired layers when forming the apertures. Controlling where the process stops will not only result from the accuracy of your endpoint technique, but also from the footing you can achieve and the uniformity of the etching rate across the wafer. Obviously, all these factors become more difficult to control at large wafer size.
PB: What kind of tools can be used to control the different manufacturing steps? Can you describe them?
SB: For the plasma process solution steps, you have tools which control the process itself, such as automatic end point technology, and then you have tools which control the equipment. As compound manufacturing matures many of the methodologies used for years in the Si industry will begin to appear e.g. SECS/GEM and factory automation. These will drive yield improvements and enable the reduced Cost of Ownership required for more mainstream adoption of VCSELs. Oxford Instruments Plasma Technology has implemented SECS/GEM across multiple customer sites and is fully ready for this next stage in VCSEL manufacturing.
(Source: Courtesy of Oxford Instruments)
PB: Why is it so difficult to maintain the same production uniformity from wafer to wafer?
SB: One of the great advantages of VCSEL is the ability to manufacture arrays to scale up the power. However, to drive an array of VCSELs with a single electrical input, each single VCSEL within the array ideally needs to have identical electro-optical properties. For example, you need all VCSELs within your array to have similar threshold voltages, so they all switch on simultaneously. That becomes even more critical when operating under pulse conditions. So, overall, the requirements for production uniformity are tight. Currently, the yield is pretty much defined by the yield of the epiwafer. The complexity and thickness of the epi structure is a challenge for epiwafer manufacturers. The DBR structure has to be very thick to achieve the required reflectivity and therefore the Al content has to be accurately controlled through a stack of 50+ layers and across a large wafer area.
PB: What are the key differences between VCSELs for Datacom and VCSELs for 3D sensing applications including consumer and automotive?
SB: Datacoms is the application where VCSEL originally started. Today it is still of interest for applications such as optical interconnects and optical HDMI cables. For these applications, the laser wavelength is typically at 850nm multimode emission on large chips operating at low power in the mW range. Since the laser is modulated at high frequencies, the design is often tailored for low electrical parasitic. In 3D sensing applications, such as gesture recognition, the laser wavelength is more often in the order to 940nm and designs are typically arrays running at higher power. Depending on the application, this power can be around 10-50W for LiDAR or around 0.5W for gesture recognition. Array density and apertures size are adjusted to power requirements.
PB: What do you expect from VCSEL use in the next five years?
SB: As the yield improves and the cost comes down, we should see more applications choosing VCSEL as a light source. IR light sources are at the heart of some of the most exciting applications, such as IoT, smart homes or gesture recognition. These applications are set to become common-place in our daily life and are perfect areas of application for VCSEL. That is not to say that VCSEL will be the only solution for the IR light source, but that it will become a technology of choice for applications where compact size, high beam stability and low power consumption matter.
Stephanie Baclet is part of Oxford Instruments Marketing team. She works close to device manufacturers in order to define their manufacturing process and identify solutions to maximise their device performances. Prior to joining OIPT Marketing team, she held several technical positions as lead process engineer in new product introduction and application engineer. Her work has been focused on nanofabrication of Optoelectronic and RF devices where she advised on and demonstrated optimum plasma processing solutions for a wide range of materials such as GaN. Her experience in plasma processing equipment ranges from VHF CCP, TCP to ICP technology. Academically, Stephanie studied physics. She entered selective postgraduate engineering schools in Orleans and holds a Double Master’s Degree in Optics Laser and Plasma as well as Space sciences.
As part of the Photonics, Sensing & Display division at Yole Développement (Yole), Pierrick Boulay works as Market and Technology Analyst in the fields of LED, OLED and Lighting Systems to carry out technical, economic and marketing analysis. He has experience in both LED lighting (general lighting, automotive lighting…) and OLED lighting. In the past, he has mostly worked in R&D department for LED lighting applications. Pierrick holds a master degree in Electronics (ESEO – France).
3D sensing – and more – in smartphones will drive the VCSEL market for the next five years. – Get more here
Related Reports and Monitors
Automotive Semiconductor Trends 2021
Market & Technology