Contact

Making Gigabyte Networks a Reality. Automated Active Alignment.

The rapidly evolving landscape for capital equipment to support photonics manufacturing is taking shape. New players, new entrants, new approaches and new procedures are required, and innovators are responding. At the beginning of a silicon photonic device’s manufacture, quality assurance at wafer level is essential for managing costs and yield. As with conventional chips, the biggest costs are in packaging, so it is imperative not to allow bad chips to proceed to assembly. And photonic chips are much costlier to package than conventional chips, due to the necessity to align optical fibres and other elements to each other and to the photonic chip. This process must be accurate to sub-submicron levels and must often be repeated multiple times as the device proceeds through assembly. The alignment process must not only be economical, but ensure series production with a high productivity and consistency. And it must be compatible with the array-based device formats typical of this new era.

Beyond Connectivity: Active Alignment as an Enabling Technology

The decisive time and cost factor in the production of PICs is the necessity to repeatedly align signal-carrying glass fibers, active and passive components, and the individual chips. From wafer probing through chip testing to packaging, this demanding positioning task must be repeated multiple times for each device. The enabling technology of the active alignment (FMPA = Fast Multichannel Photonic Alignment) developed by PI is the key to speeding up this process immensely: Ultrafast, precise, and active alignment where the orientation of the optical components is optimized for best coupling efficiency.

The FMPA system combines a hybrid active alignment mechanism with up to six degrees of freedom in a modular architecture, with fast algorithms that are integrated in the firmware of the controller. These algorithms provide command-level functions such as the search for first light, local-maximum rejection, and especially the parallel gradient search. That unique technology provides a fast optimization across multiple inputs, outputs, channels and degrees of freedom between photonic elements, even when these influence each other optically or geometrically.

Examples: Leveraging the Possibilities of FMPA

FormFactor Inc. integrated PI FMPA technology early on into its Cascade Wafer Probers. With the introduction of these groundbreaking tools, entire SiPh wafers could be tested within hours instead of weeks or months, ensuring that processes can be rapidly optimized and only good chips proceed to packaging. Nowadays, the FormFactor's wafer probers are being successfully employed by more than two dozen manufacturers of PICs, i.e. SiPh components. The newest product, the Cascade CM300xi-SiPh, also offers, for the first time, the possibility of horizontal die-level edge coupling as well as wafer-level edge coupling.

Recently, the Dutch system integrator TEGEMA B.V. presented an assembly and packaging platform for PICs, transceivers and other photonic devices that is also based on the FPMA technology from PI. The cycle times for one process cycle are typically less than 30 seconds-- an order of magnitude faster than legacy tools. The system, with a floor space of less than 1 m², is highly space efficient – which is particularly important when they are used in expensive clean rooms. Its modular concept is a special feature that makes it possible to adapt the system to increasing unit numbers and to achieve a higher degree of automation.

Broad Application Potential Still Untapped

Further application examples for PI´s FMPA technology are the assembly of optical cables for end user applications that require large cable lengths. This emerging field is being driven by displays with high pixel counts, high frame rate, and new high bandwidth bus systems like Thunderbolt 3 and USB 4 and beyond. Again, the challenge for assembling duplex and higher fiber-count cables is the multi-degree-of-freedom, double sided alignment. PI systems deployed in this field make short work of this challenge and allow assembly of both ends of the cable simultaneously due to FMPA’s inherent parallelism.

Do you have questions about our solutions? Our specialists are happy to help you!