In a word, 'polymers'

A Discussion With Louay Eldada of Telephotonics

Louay Eldada founded Telephotonics (Wilmington, MA) and serves as chief technology officer and vice chairman. Previously he managed device development in the polymer photonics division of Corning and held several senior positions at Honeywell, where he specialized in polymer photonic components. He holds Ph.D., master's, and bachelor's degrees in electrical engineering from Columbia University.

Eldada: I have been working in the field of integrated optics for the last 12 years, after having been involved with electronic integrated circuits for six years. Over these 18 years, I have developed an understanding for VLSI [very large-scale integrated] circuitry, optical componentry, and material systems based on silicon, III-V semiconductors, garnets, lithium niobate, glass, and polymers.

Experience in these areas allows us to produce integrated optical chips that are highly dense and where each element is produced in the most suitable material system. Furthermore, we understand what it takes to fabricate every type of optical building-block function in integrated form. We decided from day one—before procuring any material-specific or process-specific equipment, and before recruiting any specialized experts—to look at all optical component technologies that have been developed or conceived worldwide.

We selected the most robust integrated approach for each optical function, regardless of the material system or technology involved, and we licensed all the patents that we needed to realize these functions. Likewise, we did not shy away from material hybridization, as long as performance and reliability were not compromised.

Eldada: The advantages of integrated optics are numerous, with the most notable being size reduction, reduction in cost, performance improvement in the form of things such as lower loss due to fewer interfaces and a reduction in transient errors when all interconnects are printed. There are also advantages from an increase in yields because of the increase in uniformity and repeatability, the increase in production volumes, the reduced need for manual labor, and the faster time to market.

Our goal has been to develop the technology, expertise, and intellectual property to build, over time, every single optical function on a chip, including couplers, switches, attenuators, filters, isolators, circulators, modulators, lasers, and detectors. All of these can be tunable when needed.

Highly integrated optical components will become a reality in the future. They'll delight system integrators with the improved performance and reduced cost, and will eventually become the norm. Integrated optics is in its infancy today, and if it continues to grow at its current rate, we could have in ten years the level of integration that the electronics industry reached ten years ago.

Eldada: There is a strong desire in the telecom industry for optical polymers to succeed. The manufacturing simplicity—and therefore the potential to scale to large volumes at low costs—are intuitively obvious. In past years, several laboratories announced work on optical polymers once one or two desirable effects had been demonstrated, only to see them fail for some reason or another. In fact, there are 25 essential technical requirements that must be met before a polymer can truly be considered a viable optical polymer for carrier-class applications. Twenty four out of 25 is not good enough. These requirements range from low absorption loss and low dispersion characteristics, to stability, curability, and ease of machining.

Our polymers are synthesized by a team of polymer chemists who are the only chemists globally to have produced optical polymers that have met Telcordia telecom requirements and all 25 technical requirements. Our polymers have the highest level of optical transparency, meaning the least amount of optical loss, and unmatched environmental stability, while allowing a high level of tunability.

Eldada: Planar integrated optics are in the very early days. The simplest forms of these circuits are arrays of a single optical function, such as VOA arrays or optical switch arrays. Although these arrays demonstrate integration, they are simple products that have the possibility of offering lower cost, smaller footprint, better performance, and simpler fiber management than the discrete predecessors. The obvious next step is to produce multistage integrated optical circuits by integrating multiple functions into a single chip. One example would be the integration of switches, VOAs, and taps, allowing the creation of monolithic automatic gain-flattening components and monolithic switchable OADM [optical add/drop multiplexer] functions.

Eldada: The vision of Telephotonics is the production of next-generation DWDM components based on solid-state tunable filters integrated with switches, VOAs, taps, and intelligent controls. Components such as a dynamic OADM perform the filtering in a hitless fashion. In other words, channels other than the source channel and the destination channel are not disturbed during tuning or reprovisioning. They operate over the C-band and are process-compatible with all other building-block elements that we fabricate on a wafer.

Another area of interest in network engineering is the zero-loss node in the all-optical network, where the light is not attenuated as it passes through a ring node. This can be elegantly achieved with integrated optics.

Eldada: Unlike other optical material systems, polymers are designed and synthesized to have the desired characteristics. Other material systems must be manipulated as part of the fabrication process, complicating the idea of integrating multiple functions on a chip. There are several ways you can work with polymers such as direct photolithographic imaging, reactive ion etching, embossing, molding, and casting. This flexibility makes it an ideal integration platform where foreign material systems such as YIG [yttrium iron garnet] and lithium niobate can be inserted into an etched groove in a tunable waveguide to make very exciting integrated optical circuits. You could consider this platform a polymer optical bench.

Eldada: In general, we will tend to only outsource the volume production of those processes where we do not add value. For example, material synthesis and wafer fabrication are core competencies and will not be outsourced.

Optical component packaging will be developed in-house and we will always maintain an in-house capability for process development and process improvements; however, we will outsource high-volume packaging to qualified contract manufacturers.

As for production capacity, fabricating planar waveguide devices can be an extremely fast process. It takes less than one shift to process a wafer from a blank 6-in. substrate to diced chips. Furthermore, we have automated this process into a cassette-to-cassette operation using a system. Our wafer fab can produce up to 1000 6-in. wafers per month, each one containing a substantial number of optical components.

Eldada: Highly integrated tunable optical components require more than traditional optical-component design and manufacturing expertise. We have exclusively acquired intellectual property pertaining to three material systems and we built our scientific team to "productize" these over time, starting with the polymer thermo-optics.

Integrated optics is in its infancy and much of the equipment and services required to manufacture high volumes simply does not exist, very much like the early days of semiconductor integrated circuits. For example, initially we had to build our own in-house automation team with specific expertise in automated wafer fab equipment and automatic test equipment—which gave us the capability to mass-produce with lot-to-lot consistency.

Eldada: Any optical component must meet the minimum technical requirements for functionality, quality, and reliability to be considered. Any optical component that can achieve these and still be a cost leader will get serious attention. We've specifically focused our attention on metro, access, and enterprise opportunities.

These specific segments have a need for low cost, dynamic optical components. Volumes are much higher in these segments when compared to long haul or submarine; therefore the underlying technology must scale to high volumes as well as low cost. Integrated optics in general have the promise of lower costs through integration, and with tunable integrated optics, there is the promise of providing new functionality at lower costs than the previous generation.