Patented fiber switch revs speed, cuts cost
george kotelly
Incorporating more than 30 years` experience in electro-optical technology and product research, an independent designer has been awarded two U.S. patents for a novel approach to all-optical switching by mastering frustrated total internal reflection (FTIR) techniques. The patented concepts represent a virtual solid-state switch in that a flexible glass plate acts as a leaf-spring in making and breaking contact with a prism. This microscopic movement holds the promise of faster than 10-microsecond optical switching.
Richard H. Laughlin, president of Omega Technologies, an engineering and patent development consultancy in Richardson, TX, developed the patented fiber-optic switch technology based on his FTIR design studies performed during the past 15 years in attempting to solve optical identification friend or foe problems for the military. In these applications, an omnidirectional transponder remains covert until interrogated with the proper optical beam. Under some operating conditions, though, identification friend or foe devices are prone to return optical energy in "off" mode, thus destroying their covert purpose.
In previous FTIR applications, two optical parts--primary and secondary--worked in tandem to accomplish switching or modulating a beam of light. When the two parts are separated by a small air gap, an input beam at an appropriate incidence angle and wavelength delivered to the primary part is totally and internally reflected by the primary part. But when the two parts are forced into contact, then the same beam is totally and internally reflected only by the secondary part. Consequently, the secondary part "frustrates" or cancels the total internal reflection characteristic of the primary part.
Scientists and engineers have been struggling with frustrated total internal reflection for more than 30 years. The theory for FTIR was developed in "Frustrated Total Internal Reflection and Application of its Principle to Laser Cavity Design," by Ian N. Court and Frederick K. von Willisten, Applied Optics, June 1964, page 719.
The practical application of FTIR technology has been historically plagued by several implementation problems. The primary problem has been achieving total contact when the air gap is being closed. In general, a small gap (less than 0.1 micron) remained, resulting in a nominal 10% reflection at the interface. If sufficient force was applied in an attempt to reduce this gap, the components sustained damage over a period of time. Once the components were brought together, it was difficult, if not impossible, to get them apart. This new patented approach overcomes these problems.
The evolution
Laughlin expains how the patented switch design evolved. "The key to accomplishing positive identification friend or foe results involves the ability to switch a corner-cube-type retroreflector on and off with no residual reflection. This reflection generally arises because of a lack of absolute contact between two mating glass parts," he says. "Solving the residual reflection problem resulted mainly from implementing peeling forces while separating the parts as opposed to shear forces." As he was solving identification friend or foe problems with FTIR technology, Laughlin realized its potential for fiber-optic switching.
The patented switch is based on the concept of an input lens that encodes the spatial positions of fibers into an angular position in collimated space. Then, this angular position is changed into distinct positions at the output focal plane by both total internal reflection and frustrated total internal reflection.
An optical signal travels over an input fiber and through an input collimating lens into a basic 1ٴ FTIR switch setup, comprising a right-angle prism separated from a relatively flat glass switch plate by a microscopic air gap. The input lens collimates and projects the optical signal into the prism. Reflected at the hypotenuse of the prism by the total internal reflection effect, the light energy passes through an output or decollimating lens and into one of the two output optical fibers.
When the glass switch plate is brought into contact with the hypotenuse of the prism, a second or frustrated total internal reflection effect takes place. The input collimated beam now becomes reflected only from the back surface of the glass switch plate. Because the back surface of the plate is structured at a very small angle with respect to the plate`s front surface, it changes the angle of the collimated beam projected onto the decollimating lens. The output light beam is therefore deflected or "switched" to the other or second output fiber. To produce a multiple input/output FTIR switch, the input fiber can be replaced by a fiber array; the two output fibers can be replaced by two fiber arrays.
A gradient refractive index lens is used for both the collimating and the decollimating lenses because its blur circle is much smaller than that of a conventional lens. Moreover, these lenses can be mounted directly to the prism without precision parts, which helps to reduce the switch`s manufacturing cost. They translate the spatial position of the fibers into an angular position in collimated space.
Long-term problems
According to Laughlin, frustrated total internal reflection has been investigated for more than 30 years, but "previous studies concentrated on virtual contact. This type of contact usually meant that the plate and prism surfaces were moved to within 0.1 micron of each other." But even this diminutive spacing would result in a 15% residual reflection, which equates to -8-decibel crosstalk.
Improving the crosstalk took five prototype iterations and the assistance of a precision optical parts manufacturer, remarks Laughlin. Finally, he discovered how to force the glass plate into total contact--within 50 to 100 angstroms --with the prism. The residual reflection improved to -50- to -60-dB crosstalk. The FTIR optical switch characteristics have been verified in the laboratory, claims Laughlin.
He adds, "After two years of development, I accomplished the patented techniques of using a bending motion and peeling forces to separate and join the two glass surfaces. Peeling the two surfaces apart instead of pulling them apart involves much less effort and losses."
To do that, Laughlin placed a piezo-electric transducer in a slot at the top of the prism to control the separation between the prism and the glass plate. Both optical parts are held in place by a frame-and-spring device. The spring generates a force on the glass plate to hold it in contact with the prism when the transducer is not activated.
In "off" mode, the air gap between the two contacted parts is substantially zero. In this mode, the total internal reflection of the prism is frustrated (canceled), and the incoming light beam passes on to the back surface of the glass plate, where it can be absorbed, scattered or reflected at a different angle.
In "on" mode, a voltage is applied to the piezo-electric transducer and causes the transducer to elongate. This effect lifts the outer edges of the flexible glass plate. In turn, air flows into the gap, breaks the vacuum and separates the two parts. As a result, the incoming beam is totally and internally reflected by the prism at another angle.
"Furthermore," says Laughlin, "the basic switch can be cascaded." The first pair of glass plates causes a translation in the z-axis. A second pair of plates can be rotated 90 degrees to effect a translation in the y-axis. Presently, claims Laughlin, "enough plates can be cascaded to 8 switch modules, encompassing 256 combinations.
"These design requirements call for relatively easy tolerances to achieve in volume optical parts." To corroborate, Laughlin has performed a detailed 1ٴ switch-pricing study. He estimates the projected manufacturing cost of the FTIR switch at 1/2 to 1/5th of the cost of conventional mechanical switches. A 1¥n, FTIR switch is assessed by Laughlin to nominally cost approximately $40 per port. Commercially available optical switches are now priced at $300 per port. Laughlin is trying to find a company that is willing to license the design and manufacture the part. He can be reached by phone/fax (214) 437-5478. q