New paths needed to breakthrough Designs
Dr. Boudewijn Docter is Chief Technical Officer of a truly disruptive high-tech startup. Effect Photonics is located in a converted building that once housed Philips lighting division in Eindhoven, South-East Netherlands, Effect Photonics is developing a micro-photonics technology that will be needed soon by several industries. “A bandwidth capacity crunch is looming”, Boudewijn explains “and we have the answer that’s ready now to scale up”.
“Most of us are not aware of the growing challenge facing datacentres around the world. The public is making increasing demands on the Internet and datacentres to drive their intelligent devices, like smartphones, tablets, or wearable technologies. They assume bandwidth is unlimited, expect speeds to multiply and prices to plummet.”
Sales of smartphones have grown exponentially with sales in 2014 (from all manufacturers combined) exceeding 1.2 billion units.
“Most handset users want to be able to access and share video and photos with others. At the same time, street navigation and voice-activated apps are rising in popularity.”
As a result, the European data centre services market is expected to grow by 16% up to 2018, thanks to the growth of these rich-content applications. Bandwidth demands are also under pressure as enterprise cloud computing takes off, together with machine-to-machine connectivity – which forms the basis of the Internet of Things. The UK, Germany, France and Benelux will be the largest data centre markets in the region.
“All this means that data traffic through cell-towers and between data centres across the globe is set to rise by 10 times over the next four years. It’s an exponential growth in data handling which cannot be easily delivered by existing technologies. And it comes with a huge rise in energy demand to keep these servers up and running.”
Data centers are one of the fastest-growing consumers of electricity in developed countries. They are one of the key drivers in the construction of new power plants.
A recent report by the US Natural Resources Defense Council noted that U.S. data centres used 91 billion kilowatt hours of electricity when surveyed in 2013, "enough to power all of New York City's households twice over and growing. Data centre electricity consumption is projected to increase to roughly 140 billion kilowatt hours annually by 2020, the equivalent annual output of 50 power plants. It is costing US businesses alone $13 billion annually in electricity bills and emitting nearly 100 million metric tons of carbon pollution a year.
Light brings a more efficient answer
Light is playing an increasingly important role in solving this challenge. It started in the 1980’s when engineers realised that optical fibre could carry many times more data than copper wires. And by using different wavelengths (i.e. colours) of laser light, it has become possible to feed many separate signals over the same optical fibre. This technology is known as wavelength-division multiplexing (WDM). Modern systems can handle up to 160 signals and can thus expand a basic 10 Gigabit/s system over a single fibre pair to over 1.6 Terabit/s. Not surprisingly, WDM systems are popular with telecommunications companies because they allow them to quickly expand the capacity of the network without laying more fibre.
What EFFECT Photonics has done is to take our knowledge of the science of photonics and apply it at a chip level. We’re designing and building chips that can handle huge bandwidths of data in a far more energy efficient and cost-effective way. Just as applications of electronics have expanded dramatically since the first transistor was invented in 1948, many unique applications of photonics continue to emerge.”
“We’ve built what’s called a “Dense Wavelength Division Multiplexing Optical System. We’re using another substrate called Indium Phosphide instead of silicon, because you can build light sources as well as the switches and the filters all on the same one chip. You can’t generate light sources directly on silicon.”
What’s been holding you back as a high-tech startup?
The challenge for the global photonics industry has been to scale up. In micro photonics, most integration technologies have been developed and optimized for a specific application. As a result, there are almost as many technologies as applications. This huge fragmentation means that the market for many of these application-specific technologies is too small to justify scaling up into a low-cost industrial volume manufacturing process.
But a joint European programme led by the Eindhoven University of Technology has come up with a different approach.
“It costs a small fortune to design and build a micro photonics chip if you do it on your own– anything up to half a million Euros.” explains Docter. “But the COBRA research institute at the Eindhoven University has developed a new process where several photonics design companies can collaborate on a single production run, bringing the costs down to around €10,000 per team. That is far more interesting to a hardware startup like us.”
“I believe we’ve reached a tipping point where photonics can really scale up, from being an exciting experiment in the lab, to a serious global business. There’s plenty of photonics expertise in Eindhoven we can tap into. But the biggest advantage is a dedicated photonics foundry Smart Photonics, who have cleanroom facilities at Eindhoven University as well as a production facility. They gather together the various photonic design projects and make the chips in batches. This approach is unique in the world, putting us two years ahead of the US and it have proved extremely useful to have it on our doorstep.”
Building a career in Photonics – From Twente to Eindhoven
I graduated at a time when photonics was still mainly confined to experiments in the clean room. I had studied for a masters in integrated photonics at the University of Twente, coming to Eindhoven in 2004 to check out the possibilities. In that year, I decided that research was the only path. I did a PhD with Eindhoven University of Technology, where I was introduced to the interesting properties of Indium Phosphide. It is the only material system where you can have all the functionalities in the same platform. So you can build light sources as well as the switches and the filters all in one chip. (There are no light sources in silicon). During the Ph.D, I spent a lot of time in the clean room learning the process from design, to fabrication to testing.
“I discovered I liked working in a small company. But the job prospects in Photonics were very different in those days. Many of my contemporaries also had difficulty in applying the skills they learned in photonics. They went off to work for ASML in silicon chip design.”
“In the late 1990's there had been a thriving photonics business in Eindhoven, because Philips had their opto electronics lab, later acquired by Uniphase. They actually built a good production line for telecom lasers, so that you can put multiple data streams on a single fibre. It proved to be a very successful exit for Philips, who made something like US$3 billion I believe. But fortunately that knowledge did not leave this city. And most of the technology ended up in Eindhoven University. They got national grants to build a superb world-class clean room facility. It is still at the top of its league today.”
How did you try turning this knowledge into business?
“At the time, people were not willing to take the risk and just try things out. There was a business plan competition organised by the High Tech Campus in 2008, where people like Bert-Jan Woertman saw the potential for what we were doing. We won the final.”
“I developed a novel building block in some type of special mirror which can be integrated with the rest of the circuit. And to demonstrate the capabilities of the mirror, we developed a special type of switchable laser which can operate at different colours of light, i.e. different frequencies, very rapidly. So with a patent on this, we decided to see if we could build a business. We quickly discovered there were other tuneable lasers out there. And although our laser was far more frequency agile, customers were not prepared to pay extra for it.”
Lessons Learned
So this experience taught us they the earlier you connect with the market, the better. You need a mature supply chain, so you can scale up quickly. Making a one-off in a research lab is one thing. But a customer wants volume, say 10,000 a month, and to know there is a guaranteed delivery date. The advice to “get out of the building” and talk to customers as soon as possible is very important advice, especially if you’re working with emerging technology on hardware.”
So far, there are no photon microchip manufacturing plants in China (often called "fabs"). There are laser factories, as well as packaging and assembly, but not the type of integration where you can make photonics chips. It is basically being led by Berlin, Oclaro in the UK and Eindhoven, so there’s no reason for us to move from this area.
What is Effect Photonics Today?
EFFECT Photonics recently announced the close of investment to bring its Optical System-on-Chip technology to market to meet the soaring demand for bandwidth to cell towers and between data centres.
“We decided to operate as a design technology company because we understand what's possible. We want to do more than just the chip design. Customers quickly see the benefits of integrating everything onto one chip. But they also want the ability to be able to interface with that chip. So it needs to be packaged and integrated into a larger design. You need to be able to provide a module in which the chip is the “optical heart”, but you also need the right driver electronics around it to control it, as well as connectors so you can plug into something else.
The fundamental demonstration of photonic integration on our chip (system on chip) has been proven. We have shown investors that you can scale production, so that if you make 1000 chips, a sufficient number of chips will be within spec. We still need to show that the technology is reliable enough so that the failure rate during manufacture is as low as possible. It is similar to the reliability demanded by spacecraft, with an expected lifespan of 30 years.
This “reliability engineering” turns out to be expensive but we’re getting there. We always co-design our product with the launching customers. Most of the customization work is done in the electronics that are around the photonics chip, not on the chip itself.
Team Dynamics Important
Last, but by no means least, spend time on building the right team. Hiring the wrong people can destroy a start-up. My cofounder, Tim Koener, comes from an electronics background and saw a market for a control company that understood photonics. But we quickly realised the need for commercial people in our team. Customers want to know you can scale your production. So, James Regan became our COO. He has a long history in commercial telecoms.”
Mark your calendar for September 23rd
On Wednesday September 23rd 2015 there will be a one-day briefing at High Tech Campus Eindhoven dedicated to the huge business opportunities that lie ahead for Photonics. The event will explore the journey of Photonic’s startup companies like EFFECT photonics on their journey to becoming a global company. The meeting will also showcase surprising examples of what’s just around the corner. The conference programme is in the advanced planning stage. More details at www.phiconference.com.
