|
Ever since Apple Inc. first introduced Face ID on the iPhone X in 2017, it has begun replacing the longstanding Touch ID fingerprint recognition tool. Now various mobile phone manufacturers are finally beginning to pay attention to this technology that appeared only in movies in the past. This technology, which makes electronic products seem as though they have vision, uses a 3D sensor module that can identify the user’s three-dimensional profile. One of the most critical components of this module is the VCSEL. |
 Figure 1. Face Recognition Function (Image Source: Science and Technology News) |
VCSEL is short for Vertical Cavity Surface Emitting Laser (Footnote 1). Since the demonstration of the first semiconductor laser diode in 1962, components with similar structures have been published one after another. The man now recognized as the inventor of the VCSEL is Professor Kenichi Iga of the Tokyo Institute of Technology, Japan. He drew the first VCSEL diagrams in a laboratory notebook in 1977 (Figure 2). Then in 1979, the InGaAs/InP material VCSEL was produced via Liquid Phase Epitaxy (LPE). This inspired a series of follow-up academic research studies as well as commercial applications.
 Figure 2. Hand Drawn VCSEL Diagram (Image Source: Yinyueguang Technology) |
|
Human Vision vs. Machine Vision |
The biggest difference between human vision and machine vision lies in our human ability to directly interpret the three-dimensional spatial information and depth of the pictures we see. Machine vision, on the other hand, encodes every pixel of the captured image. As only 2D images can be obtained, there is no information about the distance of objects. Therefore, 3D sensing technology is about improving the capabilities of machine vision. It may even be the key to further empowering machines to interact with their environments.
3D sensing modules consist of two parts: the “emission source” and the “detector”. Detectors can be roughly divided into Si-based CMOS and III-V types. Examples include InGaAs-like materials or quantum dots and other technologies. Suitable detectors are selected according to the different receiving bands. At present, mainstream emission sources use Infra-Red (IR) light because, compared to visible light, infrared light has a better Signal-to-Noise Ratio (SNR). Furthermore, there are several suitable wavelengths in the infrared light band, such as the 850nm, 905nm, and 940nm bands of Near-IR (NIR) light and the 1,350nm and 1,550nm bands of Short-Wave IR (SWIR) light. Therefore, currently, each manufacturer chooses a frequency band then develops the corresponding module.
|
The Structures of Three Competitive 3D Sensor Modules—Infrared Light Emitting Source Components |
|
Infrared light emission sources can be divided into three categories using the component structure, as shown in Figure 3. These categories include the LED (Light-Emitting Diode), EEL (Edge Emitting Laser) and VCSEL. Although LEDs and EELs still have advantages in terms of cost, when considering component stability and mass production, VCSELs maintain wavelength stability over a wider temperature range and are easier to incorporate into array packaging. As such, in an overall comparison, the VCSEL’s high output power, high conversion efficiency and high quality beam wins all around. Therefore, it is widely used in applications in the field of 3D sensing. |
Figure 3. Types of Infrared Light Sources for 3D Imaging Modules |
|
New VCSEL Development—Self-Driving Automotive Systems |
In addition to its applications in optical communications and smart phone light sensors, the VCSEL is a Key sensor component in the automotive autopilot system’s optical radars (or LiDAR). This is the next major application driving the development of VCSELs. LiDAR is short for “Light Detection and Ranging”. It is a method of using a laser to detect distance by emitting laser pulses, collecting the time differences between the signals reflected from the objects ahead, and using that information to calculate distances. Automotive LiDAR systems transmit and process data from multiple laser pulses simultaneously. A 3D environment model with depth information can thus be constructed. By recognizing the positions and actions of road signs, automobiles, locomotives, pedestrians and other static and dynamic objects, self-awareness is achieved. This helps vehicles detect obstacles in their paths, enabling the realization of obstacle avoidance, braking, path planning and other goals of autonomous driving. In addition, the depth-sensing system can be installed in the interior of the car for use in cabin monitoring and passenger gesture sensing to assist the driver and enhance passenger comfort.
The automotive industry has been moving towards intelligent, self-driving cars with Advanced Driver Assistance Systems (ADAS). It is an inevitable trend for cars to start being equipped with LiDAR and depth-sensing systems. However, VCSEL components and modules must meet the AEC-Q102 vehicle regulations’ high verification standards to be used in the automotive field. The scope of these regulations covers wafer fabrication, component packaging, electrical and optical functions, reliability verifications and more. In the area of automotive component verification, the focus is on how to effectively reduce the failure rate. The ultimate goal is the zero failures required by the AEC-Q004. The overall discreet optical vehicle component verification process is shown in Figure 4. The failure rate can be effectively controlled by applying the 6δ, SPC (Statistical Process Control) and other quality control methods. In addition, increasing the number of verification samples helps ensure a lower failure rate.
|
Figure 4. AEC-Q102 Verification Process Flowchart (Image Source: AEC-Q102) |
|
Doesn’t Strict Vehicle Specification Verifications Take a Long Time? MA-tek Offers a One-Stop Solution |
|
The verification of the reliability of VCSEL components for automotive use includes five major areas: environmental stress testing, accelerated life testing, package strength testing, structural reliability testing and optical and electrical verification. To expedite this verification process, MA-tek has developed a one-stop One-Site Service. With the introduction of VCSEL optical characteristic measurement equipment (Figure 5), there is now no need to send samples back to the client for function verification during the AEC-Q102 verification process as it can be done directly at MA-tek. This reduces the time needed for deliveries, thus reducing overall verification time. |
 Figure 5. VCSEL Optical Characteristics Measurement Equipment |
MA-tek has an abundance of experience in AEC-Q automotive electronic verification. We have coached several customers to successfully pass strict vehicle regulation verifications. As for customers who want to enter the automotive market, MA-tek is able to provide a full range of vehicle verification services and even aid with overall planning. Our services cover AEC-Q100, Q101, Q102, Q104, Q200 and more. MA-tek has the knowledge and equipment to assist customers to successfully obtain tickets into the automotive market and make products that are safe to install in cars so that they can become a part of the automakers’ supply chain.
 Figure 6. MA-tek’s Full Range of Vehicle Verification Services |
Footnote 1: Taken from the bilingual vocabulary list issued by the National Institute of Education’s academic terms and dictionary information network. https://terms.naer.edu.tw/detail/11562042/
