Visiting The Future At CLEO

CLEO – The Conference on Lasers and Electro-Optics – presents the most comprehensive snapshot of laser and photonics applications. Presented by three professional societies (OSA, APS and IEEE Photonics) it was held in San Jose last month. While few of the topics covered are mainstream today for the semiconductor industry, one doesn’t have to look too far to find impact and potential synergies.

One exciting new capability is table-top mid-infrared microscopy. Daylight Solutions, a San Diego laser company that has specialized in quantum cascade lasers for the mid-infra-red won the 2014 CLEO/Laser Focus Innovation Award for its Spero microscope. Intended for life sciences (it can distinguish cell types from false-color images created from infra-red spectra), the Spero can image chip and interposer structures from the backside since 5-11 micron infrared penetrates silicon.

Matthew Barre, Business Development Manager at Daylight, believes they have just scratched the surface of potential inspection applications. Resolution is 1.5um per pixel and each pixel can represent a “cube” of spectroscopic information, allowing specialized algorithms to paint false-color images of the distribution of spectral features or real time imaging of a chosen “color.” The proprietary broadband NA=0.7 objective is diffraction limited – 5um resolution at 5um wavelength – with a 650um field of view. A low resolution I.R. and visible survey object can also be selected on the microscope turret.

Another frontier is quantum cryptography. Banks and others need to send crypto-keys using some communication scheme that eves-droppers cannot tap –or at least not tap undetectably. One way to do this involves entangled photon pairs that together mean something but separately do not. When transmitted through free-space or fiber optics to separated receivers, the quantum information can be collected and decoded using normal communications channels, preserving the secrecy shared only by the two receivers. While quantum, entanglement has been a laboratory fascination for a while, the first commercial high-brightness twin photon source suitable for commercial applications is the TPS_1500 presented by AUREA Technology of Besançon, France at CLEO 2014, earning an Honorable Mention Innovation Award.

Then there is time, or more accurately, pulse-length. These days, it has gotten easy enough to generate pulses only a few femtoseconds long that applications have appeared outside the laboratory. A femtosecond is a millionth of a nanosecond, and powerful femtosecond laser pulses can heat an opaque surface so fast that a layer evaporates before the heat gets transmitted inward. The heating is so nonlinear that the ablated area can be many times smaller than the focused beam. Focused inside a transparent medium, such pulses can change the structure of a sub-micron bubble, creating mark or even a void. Mask makers use this for repair and fine-tuning of finished photomasks, but the most interesting application at CLEO was automated writing of 3-D waveguide structures in borosilicate glass for 1 Tb/s fiber optic communications.

How do you get to 1Tb/sec? And why? That is 125 GB/sec, maybe 5000 times faster than the fastest Internet anyone has in the U.S. Well, it turns out that the growth of internet traffic (mostly Netflix) is so immense that today’s single-core fiber-optic cables may run out of bandwidth even with every known trick. So, the communication guys are developing fibers with arrays of 7-19 cores, which have to be fanned out into planar-processed semiconductor photonics devices. Making a complex 3-D coupler using conventional photonics methods would (at best) require many layers and many lithography steps.

But Optoscribe Ltd., a spin-off from Herriott-Watt University in Scotland, makes 3D Optofan space division multiplexers using femtosecond lasers to write high-index, low scattering waveguides in borosilicate glass. The cores of a standard multicore fiber form a hexagonal array and the 3D Optofan matches that array on one side and then threads the waveguides into a plane of single mode channels that can be further manipulated or directed into conventional single core fibers. This is all done in a single automated step using machinery that writes the cores using laser pulses focused at variable depth and position, according to Nicholas Psaila of Herriott-Watt and Optoscribe.

While semiconductor manufacturing engineers might blanch at such a serial process, even one that replaces many conventional parallel processes, it seems unlikely that the gently rising and falling 3D structures need for these photonic systems can be made using typical planar methods. Might this be a window into a 3D future for electronics?