Tikalon Blog by Dev Gualtieri

Tikalon Header

Optical Routers

April 29, 2013

Mechanical systems are generally slow, they require frequent maintenance , they're subject to wear , and they often have a short lifetime . After vacuum tubes were replaced by transistors , electronic systems became more reliable than mechanical systems, they didn't need maintenance, and they were far less expensive.

Although the signals switched were electrical , the US telephone switching system of the early twentieth century was mechanical. Stepping relays would convert the pulses from rotary dials to connections between wires . When transistor technology had reached a critical performance point, the telephone company began the phase-out of mechanical switching components.

Bell Labs is located in New Jersey , having moved from the Bell Laboratories Building in New York City in 1966. New Jersey had become a hotbed of telephone company activity at that time. Bell Telephone field-tested its first electronic switching system, the 1ESS switch , in Succasunna, New Jersey , less than a mile from my house, in 1965.

Succasunna, New Jersey, Post Office

The Succasunna, New Jersey , Post Office , on a fine spring day in 2013.

Succasunna was the site for the field test of the 1ESS telephone switching system , starting May, 1965, somewhat coincident with the move of Bell Labs from New York City to New Jersey.

(Photo by the author.)

Succasunna is the word for "black rock" in the language of the region's Native Americans . The black rock in this case is iron -bearing magnetite , which is abundant in the area. I wrote about New Jersey magnetite in a previous article about Thomas Edison's venture into iron mining (Edison's Iron Mine, September 20, 2010).

Although the 1ESS switching system replaced the rotary relays, it still contained miniature reed relays to switch the connections. Transistors at the time couldn't tolerate the high voltages carried on telephone lines. Although the speech signals of POTS (plain-old-telephone-service) are less than a volt , there are 48 volts present to determine the hook status of the telephone handset ; and a ring signal of 90 volts, 20 Hz , alternating current superimposed on the DC hook voltage.

The interesting thing about standards is that they linger for too long a period. My residential telephone service comes to me via a fiberoptic cable , but there's a big box in my cellar that converts digital signals to the high voltages traditionally used by telephones. While antique telephones used the 90 volt signal to ring a mechanical bell, the telephones in my house use this voltage to generate a tone signal.

As my home example shows, world communications are migrating to optical fiber , even in "the last mile ." Often along the course of a message, light signals are converted to electrical signals, switched to another channel, and then converted to optical signals again. This is quite a roundabout way of doing things. It would be far more efficient to just route the light.

This fact was noted by Bell Labs scientists many years ago, and I attended quite a few presentations about their solution. This was a peak period for microelectromechanical systems (MEMS) devices, so their solution involved miniature mirrors popping out of substrates to change the direction of light signals. It worked, it was high-tech , but it took a lot of circuit area, a lot of chemical processing , it was mechanical, and it was not implemented.

Now, an international research team comprised of members from Harvard University , Singapore's Nanyang Technological University , the Singapore Institute of Manufacturing Technology , and Nankai University , China , have developed a nanoscale optical router based on surface plasmon polaritons excited in a nanostructured metal surface layer.[1-3] I wrote about plasmons in nanostructured metal layers as a means of increasing solar cell efficiency in a previous article (Light Trap, January 7, 2013). The principal investigator of this study was Federico Capasso , a Professor of Applied Physics at Harvard and a Bell Labs Alumnus.

An optical splitter based on surface plasmons

A circularly-polarized light beam is split into two components by a surface plasmon array.

The splitter is an array of herringbone slits in a gold layer.

(A portion of a Harvard University image by Jiao Lin and Samuel Twist.)

Plasmon control of light is not new, but the typical structure used in the past was a grating , which is not that efficient. The surface layer of the Harvard device is a thin sheet of gold perforated with a herringbone pattern of slits.[2] Since the pattern elements are smaller than a wavelength of light, it would be easy to integrate such structures onto integrated circuits.[2]

The surface plasmon polaritons, created by incident light, are waves in the electron sea that exists in metals, and they inherit the polarization of the light source.[2-3] A perpendicularly incident light wave will be be sent in different direction depending on its polarization (see figure). The polarization can be linear , left-hand circular , or right-hand circular .[2] As Andrey E. Miroshnichenko and Yuri S. Kivshar of the Australian National University remarked in a perspective on this research, this is a somewhat unexpected property of surface plasmons.[3]

Operation of the Harvard plasmon router

Plasmonic waves in the Harvard University optical router. Left, left-hand circularly polarized light is routed only to the left; center; linearly polarized light sent both left and right; right, right-hand circularly polarized light is routed only to the right. These images were obtained with a near-field scanning optical microscope . (Harvard University image by Jiao Lin and Balthasar Müller.)

This research was supported by the U.S. Air Force Office of Scientific Research , A*STAR of Singapore, the National Natural Science Foundation of China , and other sources.[2]

References: