Tel: 0086 755 83545255
Fax: 0086 755 83545271
Zip Code: 518000
Silicon technology gradually matures, chip technology bottleneck breakthrough, packaging still needs to be upgraded.
Silicon optics technology has entered the industrialization stage. Intel, Luxtera, Acacia, Optics, Rockley and other enterprises have introduced chip-level, module-level products, and gradually realize small-scale commercial shipment. Chip technology requires a long time of iterative trial and error, with the development of the industry chain for many years, the accumulation of silicon-optic technology has reached the level of qualitative change, silicon-optic technology station on the scale of commercial starting line.
Silicon optical technology has entered the commercial stage, and the industrial chain has gradually matured.
Based on the waveguide theory put forward around 1985, silicon optical technology began to develop from theory to industrialization around 2005-2006. Luxtera, Kotura and other pioneers promoted the development of technology and industrial chain, forming a silicon optical chip substitutes factory (Global Foundries, Italian Semiconductors, AIM, etc.), laser chip substitutes factory (Lianya Electronics, etc.). (Luxtera, Kotura, etc.) Fabless industry chain model is more mature, and there are also IDM model represented by Intel. In addition to laser chip, design, silicon chip processing, sealing and testing are completed by themselves.
The traditional optical module uses a discrete structure, optical chip through a series of passive coupler devices, and fiber alignment coupling, complete optical path packaging. The whole packaging process needs more material and labor costs, and the packaging and testing process is more complex, and the automation rate of packaging process is lower. In the test, the optical modules need to be tested by hand. The time cost and labor cost are higher.
With the maturity of waveguide theory and the design and production of a series of new devices, silicon optical technology based on CMOS manufacturing technology has been proposed.
Silicon light can fabricate large-scale waveguide devices by etching on silicon substrates. Key devices such as modulators and receivers can be fabricated by epitaxial growth, and modulators, receivers and passive optical devices can be finally realized. Highly integrated.
Compared with traditional discrete devices, traditional technology needs to package electrical chips, optical chips, lenses, alignment components, optical fiber end faces and other devices in turn. Silicon optical volume is greatly reduced, material costs, chip costs, packaging costs are expected to be further optimized. At the same time, silicon optical technology can be measured in batches by wafer testing and other methods. The test efficiency is significantly improved.
The silicon optical technology can realize the integrated manufacturing of all kinds of devices in the optical module except the laser.
The devices manufactured by the aforementioned silicon optical technology cover most of the components within the optical module, but do not contain laser chips. Because silicon is an indirect band gap, the minimum of conduction band (bottom of conduction band) and the maximum of full band are in different positions in k-space, the electron transition needs to change the potential energy and kinetic energy at the same time, and the generation of laser needs the action of phonons. The hole recombination efficiency is very low and the luminescence efficiency is very low.
In contrast, III-V materials (such as InP) have direct band gaps and only need to absorb the external energy for the electron energy transition. The luminescence efficiency is high. So at present, the mainstream of silicon-based hybrid integration is still silicon-based hybrid integration. The traditional III-V family materials are still used in the optical chips, and the III-V family lasers are fabricated together with the modulation and coupling optical paths integrated on silicon by discrete mounting (such as optical communication, Luxtera, etc.) or wafer bonding (Intel, etc.).
At present, chip-level devices based on silicon optics technology can be mature and processed, including optical waveguides, split-wave devices, external modulator devices, APD receivers and so on.
However, the design and process routes of mainstream manufacturers are still quite different, and there are many technical routes. From this point of view, silicon optic technology is still in the early stage of development of a hundred schools of contention. The scheme with the highest performance-price ratio and technical stability has not yet emerged. Silicon optic technology still needs a period of precipitation and development. Only in this way can we focus on the ultimate winning mainstream technology, and further give full play to the scale effect of CMOS process, and the cost and yield can be continuously optimized.