Single transistor to the evolution of integrated circuits, directly contributed to the outbreak of the human information revolution, opened the prelude to consumer electronics, resulting in nearly 50 years of countless scientific and technological miracles and countless great enterprises. Based on the long-term development of the semiconductor industry statistics, the semiconductor industry summarized the so-called & ldquo; Moore's law & rdquo; & mdash; & mdash; & mdash; integrated circuits on the number of transistors that can be accommodated in about every 18 months will be doubled, and the performance has doubled.
But along with Moore's law to the limit, Moore's law is drawing a perfect & ldquo; S curve & rdquo; towards the ceiling: the electric chip works by manipulating the electrons in the device (transistors, resistors, and capacitors, etc.) to transfer information. But when the device reaches the nanometer scale, the electrons produce tunneling effects, making the electrons not easily controlled, which is a fatal blow to the device. That's why some experts say that the limits of manipulating electrons are gradually approaching, and performance can no longer be enhanced simply by reducing device size and improving integration.
Unlike electrons, photons have parallel, high-speed characteristics. Optical path in the air cross-transmission and mutual non-interference, while optical computing has a natural parallelism, can be in a time period at the same time for multiple calculations, and its own energy consumption is very low, to the current booming artificial intelligence, through the photon can be instantaneous realization of the convolutional algorithms. That is to say, from the point of view of future technological development and application, optical computing is expected to replace electrical computing, evolving into the next generation of high-performance computing processor.
But has always been, for & ldquo; optical, mechanical and electrical computing & rdquo; four major engineering fields, the volume of light is less than the electricity of one thousand. From the C-end application point of view, the realization of optics is mainly limited to the geometric optics design theory category and cold optics process manufacturing category, such as lenses, imaging, etc., the accuracy stays in the sub-millimeter and micron level category; from the B-end application point of view, the field of optical communication faster witnessed & ldquo; light & rdquo; generation of & ldquo; electricity & rdquo; trend, silicon photonics technology is gradually Making light and electricity in the accelerated fusion. From the future trend, we believe that in the near future:
Back to the development of consumer electronics, & ldquo; nano-scale & rdquo; and & ldquo; scale low-cost & rdquo; is the integrated circuit technology to make electronics into the two major features of the consumer level. Similarly, & ldquo; consumer photonics & rdquo; prologue of the real open will also be accompanied by these two characteristics of the claim. Throughout the current state of optical development: active optics with silicon photonics technology as the flagship of rapid development in recent years; but on the contrary, the volume is larger, more closely related to the consumer passive optics is still stuck in the traditional & ldquo; cold optics & rdquo; system & ndash; component size, precision by the process limitations, resulting in application limitations.
Realize & ldquo; consumer photonics & rdquo;, means that optics also need to step into & ldquo; nano-scale & rdquo;, need to step into & ldquo; great scale low-cost & rdquo; made from the higher the cost of precision made of a single unit, means that in the need to & ldquo; wafers & rdquo; level to achieve optical design and Made.
Wafer-level optics is a cornerstone area for consumer photonics. Wafer-level optics allows optics to improve accuracy by an order of magnitude while decreasing cost by an order of magnitude, which in turn makes possible many emerging needs and commercial value, including 3D depth imaging and unmanned, AR/MR display, short-range all-optical transmission between chips, medical imaging, automated security and so on. A typical application is 3D depth imaging.
The emergence of 3D depth imaging is both accidental, but also inevitable. On the one hand, it is triggered by the new needs of the terminal, but in turn, profoundly triggered the industry's deep changes. 3D imaging module core components, DOE as a new passive device, its effect is to receive the incoming light field, light re-calculation, and the output of a specific distribution of the light field.
From the optical point of view, DOE is like a new type of lens designed based on the theory of optical fluctuation; and from the point of view of signal processing, as DOE realizes the specific calculation and modulation of the incident light field (which is in fact very similar to the modulation of electrical signals. Electrical signal processing is to calculate the signal in the time domain; optical signal processing is to realize the calculation and modulation of the signal in the spatial domain), so DOE is more like a chip based on optical signal processing. At the same time, in order to realize this purpose, the size and precision of DOE optical components need to reach the realization of nanometer scale on the wafer. Therefore, the emergence of DOE in depth imaging has pioneered the application of fluctuating optics and wafer-level fabrication to very large-scale consumer products.
But trying to make nanoscale optical components is no easy task.
One is the difficulty of design. Optical components with computational capability are designed by introducing micro-nano optical structures into materials to realize new optical properties through the fluctuation effect of light and then controlling the rearrangement of the optical field. From this point of view, the design and manufacture of micro-nano optical structures is a key technical issue in the development of micro-nano optics, and the design of micro-nano structures is a cutting-edge field.
Secondly, the difficulty of production, as mentioned above, the previous optical applications are mostly limited to the scope of geometric optics, the size of the device requirements are generally in the sub-millimeter or micron level, the processing process involved are grinding, casting and molding and other cold processes. Optical chips need to be optical components precision to the nanometer level, obviously the original cold process is not applicable, but the optical industry and the lack of similar semiconductor kind of mass production of ultra-high-precision devices of the mature process.
Therefore, the current more feasible way to develop, is the integration of optical and semiconductor processes, with semiconductor ideas to do nanoscale optical components. That is, the world over five decades, invested hundreds of billions of dollars to build a microelectronic chip manufacturing infrastructure into the passive optical field, the mature, developed semiconductor integrated circuit process applied to wafer-level optics, to rapidly enhance the manufacturing level of the optical industry.
However, on this basis also need to address the yield and cost issues, due to the nanometer feature scale, making the back-channel processing will inevitably cause random errors, so the need to compensate for the front process, the need to compensate for the design, the need to & ldquo; Design to Manufacture & rdquo;. And to ensure that each piece of optical chip yields, testing links also need to build a complete process steps. In addition, because it is & ldquo; Consumer Photonics & rdquo; the basis of a link, low cost is more often than not the decisive requirement. Therefore, from the production point of view, its & ldquo; impossible task & rdquo; In summary, it is to & ldquo; precision to the extremely high standard outside the nanometer scale, while to achieve high yield, high consistency, very low cost of very large-scale production. ”.
The design and manufacturing technology of wafer-level optics can be applied to the field of 3D perception in addition to VR/AR products and 5G industry, mainly involving the design and manufacture of optical field display optical waveguide, high-speed optical communication links.
Although the technical difficulty is high, due to the broad application prospects, there are domestic startups engaged in the design and production of wafer-level optics, such as Kunyou Optoelectronics. However, for startups, in addition to the pressure of capital investment, but also face the technical problems of R & D in the early stage, and the management of large-scale production problems in the later stage. Therefore, for startups, efficient integration of upstream and downstream industry chain resources is the only way to succeed.
This article is an interview with Mr. Tao Lin, Chairman of KunYou Optoelectronics:
KunYou Optronics is a wafer-level optics startup. Wafer-level optics, as an extremely cross-cutting discipline, cannot be separated from the support of resources from all walks of life. In addition to an experienced and diversified team, KUNYU Optronics has gathered shareholders including Yuanjing Capital, Walden International, CSCI, Sunny Optical, Kunzhong Capital, Chen Hui Venture Capital, and Zhongheng Starlight, covering optics, semiconductors, Chinese Academy of Sciences, and downstream, etc. All of these parties have been working together.
Wafer-level optical industry is at a point of rapid development, on the one hand, the prospect is broad, but on the other hand, the upstream and downstream with the ecology is still not perfect. Only more upstream and downstream enterprises and friends to join, in order to make this field bigger and stronger.
