CA1292089C - Apparatus using two-mode optical waveguide with non- circular core - Google Patents

Abstract

APPARATUS USING TWO-MODE OPTICAL
WAVEGUIDE WITH NON-CIRCULAR CORE
An apparatus utilizes a two-mode optical waveguide with a non-circular core to provide stable spatial intensity patterns in both propagation modes for light propagating therein. The light has a wavelength, and the non-circular core has cross-sectional dimensions selected such that (1) the waveguide propagates light at that wavelength in a fundamental mode and a higher order mode, and (2) substantially all of the light in the higher order mode propagates in only a single, stable intensity pattern. Embodiments of the invention include, for example, modal couplers, frequency shifters, mode selectors, and interferometers. One of the interferometer embodiments may be used as a strain gauge.

UA3-996:ns/cc3

Description

STANF. C76 --1-- PATEaiT

APPARATUS USING TWO--MODE OPTICAL
WAVEGUIDE WITH NON-CIRCULAR CORE
Field of the Invention The present invention relates generally to optical waveguide devices and, more specifically, to devic s which incorporate two-mode optical waveguides to control the propasation of optical energy in the two modes of the waveguides.
Backqround of the Invention An optical fiber is an optical waveguide having a central core surrounded by an outer cladding. The refractiYe indices of the core and cladding are s~lected so that optical energy propagating in the optical fiber is well-guided by the fiber.
As is well known in the art, a single optical fiber may provide one or more propagation paths under certain conditions. These propagation paths are commonly referred to as the normal modes of a fiber, which may be conceptualized as independent optical paths through the fiber. Normal modes have unique electric field distribution patterns which remain unchanged, except for amplitude, as the light propagates through the fiber.
Additionally, each normal mode will propagate through the fiber at a unique propagation velocity.
The number of modes which may be supported by a particular optical fiber is determined by the wavelength of the light propagating therethrough. If the wavelength is greater that a l'second order mode cutoff" wavelength (i.e., the frequency of the light is less than a cutoff frequency), the fiber will support only a single mode. If the wavelength is less than cutoff (i.e., the frequency is greater that the cutoff frequency), the fiber will begin to support higher order modes. For wavelengths less than, but near cutoff, the fiber will support only the fundamental, of first order mode, and thP next, or second order mode. AS the wavelength is decreased, the fiber C? 1~3 will support additional modes, for example, third order, fourth order, etc.
Each of the normal modes (e.g., first order, second order, etc.) are orthogonal, that is, ordinarily, there is no coupling between the light in these modes. The orientation of the electric field vectors of the modes defines the polarization of the light in the mode, for example, linear vertical or linear horizontal. A more complete discussion of these modes, and their corresponding electric field patterns, will be provided below.
A number of devices have been constructed to utili~e the orthogonality of the modes of an optical fiber to provide selective coupling between the modes. For example, U.S. Patent No. 4,768,851, entitled "Fiber Optic Modal Coupler," assigned to the assignee of this invention, describes a device which couples optical energy from the first order mode to the second order mode, and vice versa. Canadian Application No. 527,487, filed on January 16, 1987, entitled "Fiber Optic Inter-Mode Coupling 5ingle-Sideband Fxequency Shifter," and ~ssigned to the assignee of this invention, discloses frequency sifters which couple optical energy from one propagation mode to another propagation mode while shifting the frequency of the optical energy. Canadian Application No. 527,402, filed on January 15, 1987, entitled "Fiber Optic Mode Selector," assigned to the assignee of the present invention, discloses a devic~ which separates optical energy propagating in one of the first order and second order propagation modes from other of the first order and second order propayation modes.
Summary of the Invention The present invention is an optical apparatus comprising a source of light and a waveguide having a core with a non-circular cross section. The source of light is 1~3;~

arranged to introduce light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of the waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode. The waveguide is sized to provide a second predetermined cutoff wavelength for the signals, less than the first predetermined cutoff wavelength. The non-circular cross section of the core has cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable intensity pattern. Substantially all of the signals introduced into the waveguide by the source of light are at one or more wavelengths greater than the second predetermined cutoff wavelength to cause the light signals to propagate in only the single, spatial intensity pattern of the higher order mode.
The fundamental spatial mode includes two polarization modes, and preferably, the cross-section~l dimensions of the core are further selected to cause the polarization modes of the fundamental mode to be nondegenerate. The single intensity pattern of the higher order spatial mode also includes two polarization modes, and the cross-sectional dimension of the core are also preferably selected to cause these polarization modes to be nondegenerate. For many applications, the nondegeneracy between polarization modes of the fundamental mode and higher order mode preferably produces a beat length between polarization modes on the order of 10 cm, or less, for both ~ets of polarization modes.
In preferred embodiments of the invention, the waveguide comprises an optical fiber having a core with an elliptical cross section. The fundamental mode is the ~ ;

?2~8~

LPo1 mode of the optical fiber and the higher order mode is the LPl1 mode of the optical fiber. The single intensity pattern is the even mode intensity pattern of the ~P11 mode.
In one of the preferred embodiments, the invention includes a device for inducing a periodic stress in the optical fiber at intervals related to the b~at length between the fundamental mode and the higher order mode such that light is cumulakively coupled between the fundamental and higher order modes at the intervals.
Preferably, the device induces the stress by producing microbends in the fiber.
Frequency shifting is achieved in another embodiment of the invention by utilizing a generator to produce a traveling flexural wave which propagates in the fiber.
The energy of the traveling flexural wave is confined to the optical fiber and has a wavelength in the direction of propagation selected in accordance with a beat length for two modes of the fiber to cause light to be cumulatively coupled from one of the modes to the oth~r of the modes and shifted in frequency.
In a further embodiment of the invention, the waveguide comprises a first optical fiber, and the apparatus additionally comprises a second optical fiber which is dissimilar to the first fiber, and has at least one spatial propagation mode. Only two of the modes of the fibers have matching propayation velocities, one of the matching modes being in the first fiber, and other in the second fiber. The fibers are juxtaposed to form an interaction region in which light is transferred between their cores. The proximity of the fiber cores at the interaction region are selected such that light propagating in one of the matching modes in one of the ~ibers is coupled to the other of the fibers. The remainder of the modes all have mismatched propagation velocities such that the propagation velocity of each of .~

. i ~;~9;~

the mismatched modes differ sufficiently from all the other modes to prev~nt substantial optical coupling between any of the mismatched modes. Preferably, the cores of each of the fibers have a cross section which is substantially the same inside the interaction region as outside the interaction region, and the lPngth of the interaction region is at least an order of magnitude larger than the maximum cross-sectional core dimension of either of the fibers. This embodiment is highly advantageous for use as a mode selector.
The invention is also useful for interferometry. In an interferometer embodiment of the invention, a source of light is used to introduce light into the waveguide such that the light propagates in two spatial modes of the waveguide, and such that the light propagates through a sensing section of the waveguide for exposure to an ambient effect. The light output from the sensing section is detected. A photodetector is arranged to intercept only a selected portion of the spatial intensity pattern d~fined by a superposition of the spatial intensity patterns of the two modes. The selected portion of the spatial intensity pattern includes substantial portions of light from both of the two spatial modes such that the intensity pattern of the selected mode varies in response to the ambient effect. In accordance with another embodiment, another photodetector is arranged to intercept another portion of the spatial intensity pattern, and a comparing apparatus is used to compare the output of the two photodetectors to sense the ambient effect. In one particularly preferred embodiment, a splitting device is used to split the selected portion of the spatial intensity pattern into two light beams. A first polari~er is used to polarize on~ of the beams to produce a first liyht signal. The photodetector is positioned to receive this first light signal. A second polarizer is used to polarize the other of the beams to produce a second light ~J
~. i, ~z~

signal, and a second photodetector receives this second lighe signal. The polarizers ar oriented such that the first and second signals have orthogonal polarizations, and a comparing device is used to compare the outputs of the photodetectors. This embodiment includes a requency shif~er which couples light from one mode to the other, and frequency shifts the coupled light. The c~oss-sectional dimensions of the core are selected to cause the polari2ation modes for the fundamental mode and ~he 13 polarization modes for the higher order mode to be nondPgenerate .
The invention also includes a ~ethod of propagating light through an optical apparatus which comprises a sourre of light and a waveguide having a core with a non-circular cross section. The method includes selecting the wavelength of the light and the cross-sec~ional dimensions of the non-circular core such that ~1) the waveguide propagates light in a fundamental spatial propagation mode and a higher order 6patial propagatisn mode and (2) 2~ ~ubstantially all sf the light in the higher order mode propagate~ in only a single, s~able intensity paetern.
Brief Desc~ tion of the Drawines . . _~
Figure 1 is a crosR-sec~ional view of an e~emplary circular core opti al fiber.
Figures 2a and 2b illustra~e the electric f ield intensi~y distrib~tion patterng for the vertically polarized and horizontally polarized HEl 1 (fundamental) propagation modes of the circular core optiral fiber of Figure t.
Figure 2c is a graph of the electric field am~litude distribution corresponding to the ineensity distribution patterns of Figures 2a and 2bo Figures 2d, ~e, 2f, and 2g illustr~te the electric field inten ity distribution pattern~ for the TEo1, TMo1.
even HE21 and odd HE21 (~econd order) propaga~ion modes, ~ 89 respectively, of the circular core optical fiber of Figu e 1.
Figure 2h is a graph of the electric field amplitude distribution patterns for the second order modes of the optical fiber of Figure 1.
Figures 3a and 3b illustrate the LP~l approximations for the f ir6t order propagation modes of the optic~l fiber of Figure 1.
Figures 3c, 3d, 3e, and 3f illustrate the LPll l~ approximations for the second order propagation modes of the optical fiber o~ Figure 1.
Figure 4 is an unscaled graph of the propagation constant of an optical waveguide versus the elliptirity of the core of the optical waveguide.
Figure 5 is a cross sectional view of an exempla-y optical fiber of the present invention having a highly elliptical core.
Figures 6a and 6b illustrate the electric field intensity patterns for the LPo1 (fundamental) p-opa~a~ion modes of the elliptical core optical fiber of Figure 5.
Figure 6c i8 a graph of he electric field amplitude di~tribueion for the LPo1 propagation mode of the elliptical core optical fiber of Figure 5 Figures 6d ~nd 6e illu~trate ~he electric fiQld inten~ity pat~erns for ~he even LP11 propagation modes of the ellip~i~al core optical fiber of Figure 5O
Figure 6f is a graph of the electric field amplitude distribution for the even LPll propagation modes of the ~lliptical core opeical fiber of Figure 5.
Figures 6g and 6h illustrate the electric field intensity patterns for the odd LP11 propagRtion modes of the elliptical core optical fiber of Figure 5.
Figure 7 i8 ~ partial cross sectional view of a bend in an optical fiber pictorially illustrating t~e coupling 3~ effect from the LPol optical mode to the LP1l opcical mode.

~2~8~3 Figure 8 is a pictorial illustration of the phase relationships between the optical signals traveling in the two propagation modes of the optical fiber of Figure 7.
Figure 9 is a pictorial representation of an optical fiber formed into a series of static microbends to form an inter-modal coupler.
Figure 9a is a cross section of the optical fiber of Figure 9 taken along the lines 9a-9a, showing the orientation of the elliptical coreO
10Figure 10 is a perspective illustration of the present invention showing an optical fiber and a transducer mechanically connected to the optical fiber to induce a series of travsling microbends in the optical fiber.
Figure lOa is a cross sectional view taken along the lines lOa-lOa in Figure 10 showing the attachment of the optical fiber to the transdu~er.
Figures lla-lle are partial elevational views showing the operation of the transducer to generate the traveling microbends in the optical fiber.
20Figure 12 is a perspective illustration of a preferred embodiment of a transducer to generate the traveling microbends in the optical fiber.
Figure 13 is a cross sectional view of a fiber optic mode selector using the highly elliptical core fiber of the present invention, showing a two-mode fiber and a single mode fiber juxtaposed in a directional coupler.
Figure 14 is a perspective view of the coupler halves which cvmprise the present invention, and shows the facing sur~aces formed on the fibers mounted in each coupler half.
30Figurs 15 is a cross-sectional end view taken along the lin~s 15-15 in Figure 13 showing the positional relationship between the cores and the claddings of the two ~ibers in Figure 13.

2~1~9 Figure 16a is a rross-se~tional view of the single-mode optical fiber ~aken along the lines 16a-16s in Figure 13.
Figure 16b i3 a graphical ~epresentation of ~he electric field energy distribution of an optical signal propagating in the LPol mode of thé single-mode optical fiber in Figure 17a.
~ igure 17a is a cross-sectional view of the double mode fiber taken along the lines 17a-17a in Figure 13.
Figure 17b is a graphical representation o~ the electrical field energy distribution of the LP~l propagation mode and ehe L~11 pro.pagation mode of an optical signal propagating in the two-mode tiber of Figu-e 17a~ :
Figure 18a shows the single-mode optical fiber and the two-mode optical fiber in juxtaposed relationship ae their facing surfaces.
Figure 18b graphically illustrates the interaction of the evanescent fields of the LP11 propagation mode of the two-mode optical fiber with the LPo1 propagation mode of the single-mode optical fiber.
Figure 19a is a graphical representation of the waveform of an optical sl~nal propagating in che LP~l mode of the single-mode op~ical fiber.
Figure 19b is a graphical representation of the waveform of an optical signal propagating an ~he LPol mode of the two-mode optical fiber.
Figure l9c is a graphical representation of the waveform of an optical signal propagating in the LP11 mode of ~he two-mode op~ical fiber.
Figure 2~a pictorially illustrates mi~matched phase propagation velocities of ehe single-mode and two-mode optical fiber~ when the core axes are parallel.
Figure 20b pictorially illustrates the matching of 3~ phase propagation velocities by positioning one of the ~2~

-1 () o~tical fibers a~ an angle with respect to the core axis of t~e other optical fiber.
Figure 21 is a pictorial illustration of an inter-mode frequency modul~tor that uses the apparatus of the present invention to separate frequency-shifted light in the LP
propa~ation mode from unshifted light in ~he LP
pr~pagation mode.
Figure 22 is a ~ystem inco~porating the mode selector of the present invention and an evanescen~ field grating reflector that separate~ ligh~ propagating at a pa-ticula_ frequency from light at other frequencies~
Figure 23 illustra~es an exemplary 6egment of a ~wo-mode optical fiber of the present invention showing the interrelationship of the optical beat length between tne LPol propagation modes and the L~11 even propagation modes and the phase delay between the L~o1 modes and the L~
even modes.
Figure 24a illustrates ~he field intensity patterns of ~he LP3l mode3 and the LPll even modes for the optical fiber o~ Figure 23.
Figure 24b illustrates the field intensity patterns resulting from ~he superposition of the LP~1 and the L~ll modes of the optical fiber of Figure 23 for three different phase delays between the modes~
Figure 25~ illustra~es an embodiment of an interferometer utilizing a single in~er-modal coupler and having ~ separa~se detector for detecting the upper and lower field intensity patterns of the highly elliptical core optical fiber.
3~ Figure 25b i8 an alternative embodiment of the in~erferometer of Figure ~5a u~i lizing an offset ~plice in place of the inter-modal coupler.
Figure 26 illustratles an exemplary of~set splice used in the inter~erometer of Figure 25b.

Figure ~7a illustrates an alternative embodiment of an interferometer utilizing the highly elliptical core optical fiber and a pair of inter-modal couplers.
Figure 27b illustrates an alternative embodiment o~
the interferometer of Figure 27a that utilizes a modal filter or mode selec~or to separa~e the optical energy in the two propagation modes.
Figure 28a illustrates an alternative embodiment of the interferome~er of Figure 27a that utilizes a modal 1~ filter or mode selector to separate the optical energy in the two propagation modes.
Figure 28b illust ates an alternative embodiment of the interferometer of Figure 282 in which the first inter-modal coupler is re~laced ~ith an offset splice.
Figure 28c illustrates an alternative embodiment of the interferometer of Figure ~8a which includes an inter-modal frequency shif~er and a ~ynchronous ~lock-in) amplifier to reduce or eliminate signal fadingO
Figure 29a illus~rates an alternative embodiment o~
the interferometer of Figure 27a tha-t includ~s a reflec~ive sur~ace at one snd of the ~ensing portion of the two-mode optical fiber 80 ehat only one inte--modal coupler i8 needed.
Figure 29b is an alternative embodimene of the interferometer of Figure 29a in whic~ a 50% coupler and an offset splice are used in place of ehe beam splitter and the inter-modal coupler in Figure 29a~
Figure 30 illustrates the effect of increasing ellipticity on the propagation constants of the polarization modes wi~hin the spatial propagation modes of the highly elliptical core optical fiber.
Figure 31 illustr~tes an embodiment of a Rtrain gauge conseructed in accordance with the present invention.
~igure 3~ illustrates t~e ~ensing portion o~ the strain gauge of Figure 31 wrapped around a mandrel to provide a temperature sensing function.

2 fi~

Detailed Descri~tion of ~he Preferred Embodiments The presen~ invention utilizes an optical waveguide ~hat operates at a wavelength below cutof such ~hat the waveguide supports both fundamental and ~econd order guided modes. The ~undamental and ~econd order guided modes provide two orthogonal paths through the optical waveguide which permits the device to be used as a two-channel optical propagation medium. The embodiments of ~he present invention utilize an optical waveguide having the geometry of the core ~elected 80 that only one stable spatial orientation of the second order mode is supported in the waveguide.
Before discussing the specific embodiments of the present invention, a detailed description of the optical waveguide and a brief summary of the applicable mode theory will be presented to provide background for more fully understanding the invention.
Mode Theory Although described below in connection ~ith a ~ilica glass optical fiber waveguide, one skilled in the art will understand that the conceptC presented are also applicable to other optical waveguides, ~uch as a LiNbO3 optical fiber, integrated optic~, or the like.
An exemplary cro~s-section of a ~ilica gla~s optical fiber 100 i8 illu~trated in Figure 1. The fiber 100 compri~es an inner core 102 and ~n ou~er cladding t~4.
The inner core 102 ha~ a radi U8 of r. In ehe exemplary fiber 100, the core has a refractive index nCO and the cladding has B refractive index nCl. A is well known in ~he art, the core refract~ve index nCO i8 greater than the cl~dding index nCl 80 tha~ an optical signal propagating in the optical fiber 100 is well-guided. The number of modes guided by the optioal fiber 100 depends upon the ~iber g2sme~ry snd upon the waveleng~h of the optical 3~ ~ignal propagating therethrough. Typically, the wavelen~th sbove which ~n optical fiber will propagate ~ ~3~

only the fundamental or first order mode i8 referred to as the "second order mode cutoff" wavelength ~c~ which may be calculated for a circular core fiber utilizing the following equation:

~c ~.4U3 (1) If the wavelength of the optical ~ignal is grea~er than the wavelength ~c (i.e~, ehe frequen y of the sptical ~ignal i5 less than a cutoff frequency), only the first order or fundamental propagation mode of the optical signal will be well-~uided by the fiber and will be propagated by the fiber. If the wavelength of an optical ~ignal i8 less than ~c (i.e., ehe frequency of the optical ~ignal i8 greater than the cu~off freguency), higher order modes, such as the ~econd order modes, will begin to propagate.
The true first and ~econd order modes of a circular core opt~cal fiber and their respective electric field amplitude diqtributions are illustrated in Figures 2a-2hO The ~wo first order modes are the vertic~lly polarized HEl 1 mode represented by an electric field pattern 110 in Figure 2a, and ehe horizontally polarized HE11 mode, represented by a~ eler~ric field pattern 112 in Figure 2b~ The outer circle in each figure represents the boundary of the core 102 of the fiber 1~U
of Figure 1.
As illugtrated in Figure 2c, the LPo1 modes have an electric f$eld amplitude distribution 116 that is ~ub~tantially symmetrical around the center line of the core 10~. The electric field amplitude di~tribution 116 is concentrated in the center of the core 102 and decreases ~ the distance from the center of the core 102 ~ncreases. A small portion of the electric field amplitude distribution 116 often extends beyond the l ~Z~

bounda!ies of the core. This extended electric field is commonly referred to a~ the evanescent field of the guided modes.
The four true second order modes are illustrated in E`igures 2d-2g. These four true modes are distinguished by t~e orien~ation of the transverse el~ctric field, denoted by the directions of ~he arrows in Figures 2d-2g, and are commonly referred to a~ the TEo1 mode, represented by an electric field pattern 12~ in Figure ~d; the TM~l mode, l~ represented by an electric field pa~tern 122 in Figu-e 2e;
the HE21 even mode, represented by an electric fieL~
pattern 124 in Figure 2f; snd the. ~E21 odd mode, represented by an electric field pa~P~e~n 126 in Figure ~g.
An electric field amplitude di tribution 130 for an exemplary optical signal propagating in the second order modes is illustrated in Figure 2h. As illustrated, the eleceric field amplitude distribution 130 is ubstantially equal to zero at the central line of the core, snd has two maximum amplitudes 132 and 134 near the boundary of the core. As further illustrated, the two amplitude maxima 132 and 134 are 180 out of phase. Further, a gre~ter portion of the electric field dis~ribution extends beyond ~he boundary of the core in the gecond order modes, thus providing a larger evanescent field th n for the HE11 modes.
Each of the four true second srder modes has a slightly different propagation velocity from the other of the four s2cond order modex. Thus, when ~wo or more of the true second order modes are co-propagaeing in a two-3~ mode fiber, the intensiey distribution of the second order mcde v~ries as a function of the length of the fiber as a result of changes in the phase differences be~ween the four modes as ~hey propagate. The cross-sectional intensity distribution of the second order mode changes in re~poDse ~o environmenll;al changes that induce differential phase shift~ between the almose degenerate four modes.

~2~8~

In order to more easily analyze the characteristics of optical signals propagating in the second order propagation modes, the characteristics of the modes are analyzed using the LP approximations for the modes defined and d~scribed in detail in D. ~loge, "Weakly Guiding Fibers," Applied optics~ Vol. 10, No. 10, October 1971, pp. 2252-2258.
A better understanding of the mode theory of optical propagation in an optical fiber or other circular core waveguide can be obtained by referring to Figures 3a-3f, wherein the first and second modes are represented in accordance with the LP approximations described by &loge in his paper. The outer circles in each of the illustrations again represent the cross section of the core 102 of the optical fiber 100 of Figure 1. The outlines within the core circles represent the electric field distributions. Arrows with the inner outlines represent the direction of polarization.
Figures 3a-3b show the field patterns of the two polarization modes in the fundamental LPo1 set of modes.
A field pattern 140 in Figure 3a represents vertically polarized light in the LPol fundamental mode, and a field pattern 142 in Figure 3b represents horizontally polarized light in the fundamental LPol mode.
Z5 Figures 3c-3f illustrate the LP11 approximations for the second order modes. As illustrated in Figures 3c-3f, there are four LP11 modes, each having two lobes for the electric field distribution. Two of the modes, represented by an LPll mode pattern 150 i~ Figure 3c and an LPll mode pattern 152 in Figure 3d, are referred to herein as the LPl1 ev~n modes. The other two LPll modes, represented by an LPl1 mode pattern 154 in Figure 3e and an LPll mode pattern 156 in Figure 3f, are referred to as the LPll odd modes. The four LP11 modes are distinguished by the orientation of the lobe patterns and the orientation of the electric field vectors (i.e., the ~:J
d~ , i, . ,.i polarization vectors) within the lobe patterns. For example, the first LP11 even mode field pattern 150 (Figure 3c) has two lobes that are symmetrically located about a horizontal zero electric field line 160. Within the two lobes, the electric field vectors are parallel to and antisymmetri~ about the zero electric field line 16~. For convenience, the LP~1 mode represented by ~he lobe pattern 150 will be referred to as the horizontally polarized LP11 even mode.
The ~econd LP~1 even lobe pat~ern 152 (Figure 3d) is symmetrically located about a horizontal zero electric field line 142. Within the two lobes of the field pattern 152, the eleceric ield vector~ are perpendicular to and antisymmetric about the zero electric field line 162. The LP11 mode represented by the electric field pattern 152 will be referred to as the vertically polarized LP11 even mode.
Tne firs~ L~11 odd mode field pattern 154 has two lobes ~hat are 3ymmetrically loca~ed about a vertically orien~ed zero electric ield line 164. Within the two lobes, the electric field vector is perpendicular to and antisymmetric about the zero electric field line 164, and are thus oriented horizontally. The LP11 mode r~presented by the field pattern 154 will thus be referred to as the horizontally polarized LP11 odd mode.
The electric field pattern 156 of the second LP11 odd mode has two lobes that are symmetrically located about a ve_tically oriented zero electric field line 166. Within the two lobes, the electric field vectors are parallel to and antisymmetric about ehe zero electric field line ~66. Thus, the LP1 1 ~ode represen~ed by the electric field pattern 156 will be referred to as the vertically polarized LP11 odd mode.
In the LP-mode approximations, ~ach of ~e six elec~ric field patterns in Figures 3a-3f, nsmely, the two L~Ul pateerns and the four LP11 paeterns, are orthogonal ~z~

to each other. In other words, in the absence o~
perturbations to the optical waveguide, there is substantially no coupling of optical energy from one of the field patterns to any of the other field patterns.
Thus, the six electric field patterns may be viewed as independent optical paths through the optical waveguide, ~hich ordina ily do not couple with each o~he If the indices of the core 102 and the cladding 104 of the optical fiber 1~0 are approximately equal, the two LPo1 modes will travel th-ough ~he fiber at ~pproximately the ~ame propagation velocity, Pnd the four ~econd srder LP11 modes will travel ~hrough the fiber at approximately the ame propagation velocity. ~wever, the propagation velocity for the fundamental LPo1 set of modes will be slower than the propagation velocity for the second o der LP11 set of modes. Thus, the two sets of modes, LPo1 and LP11, will move in and out of phase with each ot~er as the light propagates through the fiber. The propagation distance required for the two 8et~ of modes ~o move out of phase by 360 (i.e., 2~ radians) is commonly referred to as the beat len~th of the fiber, which may be ma~hematically expressed as:

L . ~ 3 2 (2) where L~ i~ the bea~ leng~h, ~ i8 the optical wavelength in a v~cuum, ~n is the difference in ehe effective refractive indices of the two sets of modes, and ~ is th~
difference in the propagation constants for the two sets of modes.
It ha~ been previously shown ~hat coherent power transfer between the two sets of ~he modes, LPo1 and LP11, can be achieved by proclucing periodic perturbations in the optical fiber that match the beat length of the two modes, A number of optical devices have been constructed 8~

_1 ~

to cont ol the coupling of op~ical energy between the ~wo modes to provide useful devices for select$ve coupling, filtering and frequency shifting o~ an optical signal.
See, for example, W. V. Sorin, et al., "Highly selective evanescent modal filter for two-mode optical fibers,"
OPTICS LETTERS, Vol. 11, No. 9, Septembe_ 1986, pp. 581-5~3; R. C. Youngquist, et al,, "All-fibre components using periodic coupling," IEEE Proceedin~s, Vol. 132, Pt. J, No. 5, October 1985, pp~ 277-2B6; R. CO
Youngquist, et al., "Two-mode fiber modal coupler," OPTICS
LETTE~S, Vol. 9, No. 5; May 1984, pp. 177-179; J. N.
Blake, et al., "Fiber-optic modal coupler using pe~iodic microbending," PTI~5 LETTERS, Vol. 11, No. 3, March 1986, pp. 177-179; B.Y. Kim, et al., "All-fiber acousto-optic t5 frequency ~hifter," OPTI~ L~TT~KS, Vol. 11, No. 6, June 1986, pp. 389 391; and J. N. Blake, et al., "All-fiber acousto-optic frequency shifter using two-mode fiber,"
Proceedingq of the SPIE, Vol. 719, 1986. The present invention provides ~ubstantial improvement to many of ~0 those devices and provides a number of new devices that utilize coupling between the modes to fur~her control an optical signal.
Although the four LP11 msd*s provide four orthogonal ehannels for the propagation of optical energy through an optical fiber or other waveguide/ it has often been found to be difficult to fully utilize the four channels independently. As se forth above, the LP11 modes a e approximations of real modes and are nearly degenerate in a circular core fiber 100. This makes the LP11 modes very sen~itive ~o couplings caused by perturba~ions in ~he op~ieal fiber, such ~s bending, twi~ing and lateral stressing. Furthermore, since the LP~1 modes are only an approximation of the real modes, there will be a slight amount of coupling ~ven in the absence of perturbations of the fiber 100. The net re~ult i8 that the propagation o~
an LP11 mode electric field pattern in ~ given mode is not ~table. In like manner, the electric field patterns o~
~he two LPo1 polarization modes are likewlse unstable.
It has been previously shown that the use of an elliptical core cross-section in an optical fiber or sther waveguide can introduce birefringence and separate the prop~gation constants for the EwO polarizations of the LPol firs~ orcler mode. The separation of the propagation constants looks the polarizatio~ of the signal to a principle axis of the oore cross-section. I~ has also been shuwn that an ellipti~al co e also increases the ~eparation between the propagation constantR of the LP11 mode pat~erns. This tends to enhance modsl stability.
This is illustrated in Figure 4 which is an unscaled representation of the propagation constant ~ versus the ellipticity of the core of an optical waveguide. As illustrated, the LP~1 propagation mode has a larger propagation constant than the LP11 propagation mode. From Equation (2), this difference in the propaga~ion constants i~ related to the beat length LB bet~een the LPo1 and LP
propagation modes as follows:

~01 ~ (3) where ~01 iQ the difference in the propaga~ion constants between the LPo1 mode and ~he LP11 mode and L~ol is the beat length be~ween the LPo1 and LP11 modes.
As illustrated in the left-hand portion of Figure 4, when the core of ~he optical waveguide is ~ubstantially circular, the LPll odd and even mode~ have su~staneially the ~ame propagat~on con~ean~. ~owever, when the core of the optical waveguidle is elliptical, the propagation constants of the odd and even LP11 modes are different.
This is illustrated by the propagation constant difference ~11 in the right hallE of Figure 4. As illustrated. the difference in the propagation constants of the odd and -2~-even LP11 modes (Q~11) increases a ~he ellip~icity increases. The use of an elliptical core optical fiber has been suggested as a means of avoiding the degeneracy of the orthogonal lobe orientations of the LP11 modes.
See, for example, J. N. Blake, et al~, "All-fiber acousto-optic frequency shifter using two-mode fiber," Proceedings of the S~IE, Vol. 719, 1~86, The foregoin~ differences in the propagation constants between the LPo1 ~ode and the odd and even LP11 modes when the core of the optical fiber is elliptical, also results in a change in the cutoff wavelength and the corresponding cutoff frequency. For exa~ple, ~or- ~ circular core optical fiber, the cutoff wavelength is rela~ed ~o the radiu~ of the fiber core, as set forth in Equation (1) above. Thus, optical 8 ignals having wavelengths above the second order mode cutoff wavelengeh ~c (i.e., frequencies below the second order mode cutoff frequency) will not propagate in the second order or higher mod~s in the optical fiber. Optical signals having wavelengths less than the cutoff wavelength ~c will propagate in the secsnd order modes. If the wavelength i~ fur~her reduced to a wavelength ~c~ third order and higher modes will be suppor~ed by the opeical waveguide. For a circular core optical wa~eguide, ~c2 can be found by ehe following equation:

~ 2 ' ~c~3~cl (4) where r, ncO and nCl are as set forth above for Equation (1). One skilled in the art will under~tand that the foregoing can also be represented by cutoff requencies.
For example, the fir~t cutoff wavelength ~c corresponds to a first cutoff requency fc~ and the Recond cutoff wa~elength ~c2 correspond~ to a second cutoff frequency fc2 that i~ greater than the first cutoff frequency fc~

~peci~ically, for the circular core optical waveguide, ir the first cutoff frequency fc is normalized to 2.4~5, the second cuto~f frequency fc2 will be normalized to 3.832.
In other words, the second cutoff frequency will be 1.5g times grea~er than the first cutoff frequency (e.g., fc2/fc = 3-832/2.405 = 1.59) Thus, an optical signal having a normalized frequency less than 2.405 will propagate in the optical waveguide only in the LPo1 mode. An optical signal having a normalized frequency in the range of 2.405 to 3.832 will also propagate in the second order LP11 ~ode. An optical ~ignal having a normalized frequency greater than 3.832 will propaga~e in higher order modes.
The foregoing relationships also apply when the core of ehe optical waveguide is elliptical or has ~ome othe_ noncircular geometry. For example, Allan W. Snyder and Xue-Heng Zheng, in "Optical Fibers of Arbitrary Cross-Sections," Journal of the Optical Society of America A, Vol. 3, ~o. 5, May 1986, pp. 6~0-609, set forth the 2~ normalization ~actors for a number of different waveguide cross-section~. For example, an elliptical core waveguide having a major ~xi~ tha is twice ~he length of tne minor axis, will have a normalized cutoff frequency f~ of 1.889 when the minor axis has th~ same length as the diameter of a corresponding circular core opti~al fiber of the same material construction. In other words, below the normalized frequency of 1.889, only the first order LPo1 modes will propagate. Similarly, Snyder ~nd Zheng suggest that the LP11 even mode will have a normalized cutoff frequency of 2.505, and ehe LP11 odd mode will have a normalized cutoff frequency of 3.426~
Snyder and Zheng generalize the foregoing concept for an elliptical core optical waveguide wirh varying ratios between the length of Ithe minor axi3 and the length of the major ~xis as follows:

~

fc = 1-70~ (1+(b/a)2)1/2 (5a) fc2even = 1.916 (1~3(bla)2)l/2 (5b) fc2odd 1.916 ~3+(b/a~ (5c) where fc is the no-malized cutoff frequency for the LPo1 mode, below which optical energy will propagate only in the LPo1 mode in the elliptical core optical fiber; whe e fc2even is the normalized cutoff frequency for optical energy propagating in the LP11 even mode, below which optical energy will propagate only in the LP11 even mode but not in the LP11 odd mode; and where fc20dd is the normalized cutoff frequency for the LP1~ odd mode, below which optical energy will propagate in the LP11 odd mode aQ well as ~he LPl1 even mode, but not in any of the higher order modes; b is one-half the length of the minor axi~ of the elliptical core; and a is one-half the length of the major axis of ~he ellip~ical core. Equations (5a), ~5b) and (5c) can be evalua~ed for an elliptical core fiber having a major axis leng~h 2a of twice the mino-axis length 2b ~o obtain the normal~zed requencies 1.889, 2.505 and 3.4~6, set forth above. Equations (5a), ~5b) and (5c) can be further evaluated for b o a (i-e., for a circular core) ~o obtain ~he LPol cutoff frequency of 2.405 and the LP11 cutoff frequency of 3.832 for both the odd and even modes, as ~et forth above~
The foregoing properties of the elliptical core optical waveguide are atvantageously utilized in the 3~ present invention to improve the operating characteristics of the optical waveguide by eliminating the LP1l odd propagation mode and thus provide only one spatisl orientation for the electric field pattern of the second order ~ode. Thi~ i~ illus~rated in Figures 5 and 6a-6g.
Figure 5 illustral:es an exemplary optical fiber 200 ~aving sn elliptical core 202 and a ~urrounding cladding 204. The dimensions of the elliptical core 202 a-e ~,elected ~o that the cutoff wavelength~ and frequencies for the two orthogonal lobe patterns of the second order mode are well separated. An optical sign~l i3 applied to tne fiber 200 that is within a frequency range selected to be above ~he cutoff frequency fc2even dnd to be below the frequeney fc2odd. For example, in an exemplary optical fiber, having a fir6t cutoff frequency f that is normalized to 1.8~9, and a second frequency fc2even of 2n505~ the frequency of the input optical signal is ~elected to have a normalized frequency in the range of 1.889 to 2.505. Thus, a light source is selected so ehat substantially all of ~he light produced by the light source has a normalized frequency that is substantially less than the second cutoff frequency fc2even. and that ha~ a 6ubstantial portion of the light that has a normalized frequency that i8 greater than the first cutoff frequency fc- In te~ms of wavelength, substantially all of the light produced by the light source has one or more wavelength~ that are greater than the second cutoff gth ~c2~ven~ and wherein a ~ubstantial portion of the light has at least one wsvelength that i8 less ~han the first cutoff wavelen~th 1~c~ Thus, the light er.~ering the optical fiber is caused to propagate only in either the first order LPol mode or the LPll even mode. Since the frequency of the optical signal i5 selected ~o be less than the cutof ~av21ength for the L~11 odd mode, substantially no light propagates in the LP11 odd mode.
The foregoing i5 illuseraeed in Figures 6a-6g. In Figures 6a and 6b, the two polarization modes for the LPol, firsc order mode are illustrated. ~n electric field pattern 210 in Figure 6,a represents the electric field for the vertically polarized LPol mode, and an electric field pattern 2t2 in Figure 6b represents the electric field for the horizontally polarized LPo1 mode. One skilled in the ~rt will under3tand that the opcical fiber 200 ~Figure 5) is birefringent for the first order LPo1 mode, and that the horizontally polarized LPol mode will propagate at a greater velocity than the vertically polarized LPo1 mode.
An electric field amplitude distribution 214 for the LPo1 5 propagation modes is illustrated in Figure 6c. As illustrated, the electric field amplitude distribution 214 is similar to the electric field amplitude distribution 116 in Figure 2b, for a circular core fiber and has a peak amplitude 216 proximate to the center line of the core 203.
Figu~es 6d and 6e illustrate the LP11 even modes for the elliptical core fiber 200. As illustrated in Figure 6d and Figure 6e, respectively, a vertically polarized even mode electric field pattern 220 and a horizontally polarized even mode electric field pattern 222 are both well-guided by the optical fiber 200. As illustrated in Figure 6f, the LP11 even modes have an electric field amplitude distribution, represented by a curve 224, that has a first maxima 226 proximate to one boundary of the core, and that has a second maxima 228 proximate to an opposite boundary of the core, and wherein the first maxima 226 and the second maxima 228 are 180 out of phase.
The LP11 odd vertical polarization mode, represented by an electric field pattern 230 (Figure 6f), and the LP11 odd horizontal polarization mode, represented by an electric field pattern 232 (Figure 5g), are not guided by the optical fiber 200 when the optical wavelength is selected to be above the second cutoff wavelength ~c2even Thus, the optical energy in the LPll odd modes, represented by the field patterns 230 and 232, will not propagate.
Thus, rather than providing four degenerate optical communication channels, such as provided by a circular core waveguide or a slightly elliptical core waveguide, the highly elliptical core 202 of the optical fiber 200 provides only two LPol mode propagation channels and two ~ , ~, ..1 LP11 even mode propagation channels. Furthermore, tne communication channels are well-defined and stable and, in ~he absenee of a perturbation in ~he optical fiber 2~V, there is no coupling between any of the four channels.
Therefore, an optical signal can be launched in the second order LP11 mode and it will propagate only in the LP11 even mode. It is not necessary to avoid exci~ing the odd lobe patterns of the second order LP11 mode be~ause optical energy in those lobe patterns will not propagate. Furthermore, optical energy ~ill not be coupled to the odd lobe patterns.
Because of the stability of the electric field intensity pat~erns of the LP91 mode and the LP11 even modes, the performances of fiber optic devices previously developed to ueilize the second order LP~1 mode will be increased. Specific examples o~ devices u~ilizing the highly elliptical core waveguide ~ill be set forth hereinafter.
escription of an Inter-Modal Coupler It has been found that if ~ fiber 300, having a core 30~ and ~ cladding 304, i8 bent, as illustrated in c-oss section in Figure 7, a portion of ehe optical energy entering the bent portion of the fiber in one mode (e~g., the first order LPo1 mode) i3 coupled to ~he orthogonal mode (eOg., the second order LP11 mode) as the optical energy propagaees through the bent portion of the fiber 300. One explanation for this effece is that the optical energy traveling ~hrough the core 302 of the fiber 300 on the inside of the bend has a ~horter path than the lighe 3~ traveling on the oue~ide of ~he bend. Xeferring to Figure 7, loca~ion 306 designa~es the beginning of ~he ~ent portion of the fiber 300. Location 308 design~tes the end of the bent portion of the fiber 300. An electric field flmplitude discribution curve 310 is superimposed upon ~he cross section of the fiber 300 at the locacion 306 and illu~trates thae the optical energy is in the LPQ1 mode (i.e., the elec~-ric field amplitude di~tribution is symmetrical about the center of the fiber). The curve 312 generally corresponds to the curve 214 in Figure 6c. A
second optical amplitude curve 312 illustrate~ the amplitude of the optical energy in the LP11 mode. In this example, it will be assumed that there is no light in the second order LP11 mode at the location 306, and thus, the electric field distribution ampli~ude curve 312 is shown as having zero magnitude. Thus, all of ~he optical energy is concen~rated in the LPo1 mode at the location 306 sf ~he fiber 300.
In the straight portion of ~he-optical fiber before the location 306, the LPo1 and LP11 modes are orthogonal and no couplin~ occurs. As the optical signal eravels from the location 306 to the location 308, a portion of the optical signal travels along ~he center of the core, illustrated in phantom lines by a path 320. A portion of the optical signal also travels along an inner path 322, illustra~ed in dashed lines, which has a shorter r~dius 2~ than the pa~h 320 in the center of the fiber core and thus has a shorter path len~th. Additionally, a portion of the optical ~ignal travels along a path 324, also illu~trated by dashed line~, which has a larger radius than the path 320 and ~hus has a longer path length. Thus, an optic l signal traveling alon~ the path 322 or any other path having a radius smaller than the radius of the cen~e_ of the core ~ill travel a ~hor~er distance from the location 306 to the location 308 ~han an optical ~ignal traveling along the path 324 or any other pa~h having radius greater ~han the radius of the path 320. Because of the difference in the length3 of the path~ from the loca~ion 306 to the location 308, the optical signal which was in phase across a cros~-section of the fiber 300 ~t the location 306, i~ no lon~er in phase when it reaches the location 308. Thus, the amplitude distribution of the optical signal at the Loca~ion 308 does not correspond to ~2~

~he symmetrical distribution shown in ~'igure 6c.
Therefore, the optical signal is no longer entirely ortnogonal to the LP11 mode, and a portion of the op~ical ~ignal is coupled to the LP11 mode. As illustrated in Figure 7, at location 3~8 the amplitude of the signal in the LPol mode, depicted by a curve 310' has been reduced in amplitude. Furthermore, a curve 312', representing the optical amplitude in the LP11 mode, no longer has a zero magnitude. Thus, a portion of the op~ical energy is transferred from the LPol mode to the LP11 mode. The fraction of energy transferred from the LPo1 mode to the LP11 mode depends upon the radius of the bend of the fiber core 302 and upon the length of the fiber core 302 which is ~o bent. The foregoing effect is reciprocal in that light energy input into the fiber such that it is initially eraveling in the LP11 mod~ is coupled to ~he LPol mode.
When an optical ~ignal is traveling in the core 3V~ of the fiber 300 in tws different propagation modes, light traveling in the first order LPo1 mode eravels at a slower pha3e propagation velocity than light traveling in the second srder L~11 modeO Thu~, if ~he li~t in the two modes i8 from the same source and ha~ ~he same frequency, light traveling a distance LB in the f{r~t order LP01 mode will take more time to travel the distance L than the light travelin~ the ~ame di~tance in the ~econd order LP11 mode. Thus, the phase of the ligh in the LP~l mode will lag the pha~e of the light in the LPl~ mode throu~h the di3~ance LB~ This i piceori211y lllustrated in Figure 8.
3~ The light traveling in the LP11 mode i~ represented as a fferies of waves 350 and the light tr veling in the LPol mode is represented a~ a ~eries of waves 352. The length L~ is selected such th,at if an optical w~vefront traveling in the LPo1 mode completes exactly n ycles in traveling the distance LB, the light traveling in ehe LPll mode will complete exactly n-l cycles. This is illustrated in ~32~9 Figure 8. Thus, when the light in the LPll mode is exactly in phase with the light in the LPol mode at the beginning of the distance LB, designated as the locations 354 and 356 on the curves 350 and 352, respectively, the light will also be in phase at the end of a distance Lg, illustrat~d as locations 358 and 360 on the curves 350 and 352, respectively. Similarly, when the light has traveled a distance of LB/2, the light in the LPll mode is lBOD (~
radians) out of phase with the light in the LPol mode, as illustrated by tha locations 362 and 364 on the curves 350 and 352, respectively. The distance LB is referred to as the beat length of the fiber 300 for the two propagation modes at a selected frequency. The distance LB is calculated as set forth above in Equation (2) as:

B Q~ (2) where ~ is the difference in the propagation constants of the two modes along the fiber. As is well known, the propagation constant, ~, is 2~ times the number of cycles of a signal in a unit length, and is calculated as follows:
~ = 2 ~6) where ~ is the wavelength in the medium in which the signal is pxopagating. As set forth above, a signal propagating in the first order LPol mode propagates at a lower velocity and thus has more cycles per unit length than the second order or LPll mode. Thus, a given signal at a given frequency propagating in the first order LPol mode will have a higher propagation constant ~01 than a propagation constant ~11 Of the same signal propagating at the same frequency in the second order LPll mode.
Returning to Equation (2), above, the beat length LB is .,~

~Z~ ~

thus inversely proportional to the difference ~ 01 ~ ~11) in the propagation constants in the two modes. A greater difference ~ in the propagation constant results in a smaller beat length, and vice versa. Typically, the difference in the propagation constants between the first order LPo1 propagation mode and the second order LP11 propagation mode is greater than the differences in the propagation constants between two polarization modes of a signal in a birefringent fiber.
Thus, the beat lengths of the two spatial propagation modes are shorter than the beat lengths of the two polarization modes.
It has been found that if an optical frequency, referred to as a center frequency, is selected to provide a minimum beat length for the first and second order propagation modes, the frequency of the optical signal can be varied substantially above and below the center frequency without causing a significant change in the difference between the propagation constants of the two modes. Thus, the beat length of the two propagation modes does not vary significantly at frequencies near the center frequency. Therefore, the beat length is relatively insensitive to changes in optical frequency over a relatiYely broad sptical frequency range in comparison to the sensitivity of the beat length between two polarization modes of a birefringent fiber.
It has been discovered that when an optical fiber is formed into a series of periodic bends which are spaced by a beat length, then the coupling between the two spatial propagation modes of an optical signal traveling through the fiber will have a cumulative effect. As illustrated in Figure 9, a length of a fiber 400 has a series of small bends 402, 404, 406, 408, 410, 412, 414, 416, 418 and 420, referred to as microbends, which are spaced apart 3~ ~uch that the distance between corresponding bends (i.e., between bends in the same direction) is substantially equal to LB, the beat length of an optical signal passing through the fiber 400. The effect of each section having a length of LB is cumulative with each other section having an length of LB to cause a cumulative coupling of optical energy from one mode to another mode in the fiber ~00. This effect was demonstrated in theory in "Bending Effects in Optical Fibers," Henry F. Taylvr, Journal of Lightwave Technology, Vol. LT 2, pp. 616-633 (1984). In that paper, the periodic microbends were introduced by statically positioning the fiber between opposing periodic structures. Thus, the coupling between the modes was a static coupling which did not effect any change in the frequency of te optical signal in the coupled mode.
The optical fiber 400 of Figure 9 is preferably an optical fiber having a highly elliptical core 420 (see Figure 9a) such as the optical fiber 200 that was described above in connection with Figures 5 and 6a-6h.
The bends 402, 404, 406, etc~ can be advantageously formed by bending the optical fiber 400 between two ridge structures 430 and 432. For example, the two ridge structures 430 and 432 can be formed by wrapping plural turns of copper wires 434, 436, or the like around respective supporting frames 440, 442. The center-to-center spacing of the turns of the copper wires 434, 436 determines the spacing of the bends and is preferably closely matched to the beat length LB f the optical fiber 400. As illustrated in Figure 9, the small bends 402, 404, 406, etc. formed in the optical fiber 400 lie in a plane. The optical fiber 400 i5 positioned so that the elliptical core 422 has its major axis lying in the plane of the bends and oriented in the direction of the bends.
Thus, referring to Figures 5d and 6e, the lobes of the hPll field pattern will propagate through the optical fiber 400 along the insides and outsides of the curves formed by the bends, thus maximizing the effects of the , ~ .,li l~Z~

bends on the coupling of optical energy between the fundamental LPo1 and second order LP11 modes.
The amount of coupling between the two modes is dependent upon a number of factors such as bend radius, fiber ~onstruction, the number of bends, and the lateral pressure applied to the fibers. Preferably, a combination of these factors are varied to achieve a desired percentage of coupling, such as 50% coupling.
Description of a Fxequency Shifter Usinq the Pres~nt Invention Figure 10 illustrates an embodiment of the present invention in which a traveling periodic microbend is introduced into a multimode fiber to cause light to be coupled from one mode to another and to be shifted in frequency. The present invention comprises an optical fiber 500, having a highly elliptical core as illustrated in Figures 5 and 6a-6h. The optical fiber 500 has a first end portion 502 into which an optical signal, represented by an arrow 504, is introduced at a first angular frequency ~o (i.e., ~o = 2~fo). The fiber 500 is secured to a transducer 510. In Figure 10, th~ transducer 510 is preferably a shear transducer comprising PZT
(lead-æirconium-titanate), lithium niobate (LiNbO3) or another piezoelectric material. As shown in Figure 10a, the fiber 500 may advantageously have a small portion of its outer cladding removed to form a flat surface 512 which rests on a top surface 514 of the transducer 510 to thereby proYide additional mechanical contact between tne transducer 510 and the fiber 500. The fiber 500 can be 33 secured to the transducer 510 by epoxy 516 or other securing means.
The transducer 510 is driven by an electrical signal source 5Z0 (shown schematically~, which, in the preferred embodiment, is an a.c. source. When activated by the 35 source 520, the transducer 510 operates in the shear mode as illustrated in Figures lla-lle. The transducer 510 is '" ~ ~
,?~

'8 ~hown in cross section in Figure 11a wi~h the,fiber 500 mounted to the top surface 514 of the transducer 510. The t~ansducer 510 has a fi~st side 524 and a second side 526. At rest, the cross ~ection of the tran~ducer 510 is ~ubstantially rectan~ular. When the electrical 3ignal 520 i~ applied to the transducer 510 with a first polarity, the ~ransducer 51~ operates in the shear mode causing the first side 524 and the second side 526 to be di~placed in opposite directions indicated by the arrows 528 and 53~, 1~ respectively, in Figure 11b. This causes the fiber 500 to be displaced at an angle with respect to the rest position shown in Figure 11a, When the a.c.- electrical signal applied to the transducer 510 reaches,a zero crossing, the first side 524 and second side 526-return to their re~t positions as illustrated in Figure 11c, thu~ returning the fiber S00 ~o its rest position.. When the a.c. electrical ~ignal is applied to the transducer 510 with the opposite polarity to the polarity applied in Figure 1lb, the first Ride 524 and the ~econd side 526 are displaced in directions indicated by the arrows 532, 534 in Figure 11d.
This displacement is opposite to the displacement illustrated in Figure llb. Thus, ~he fiber 500 is dLsplaced at an angle opposite the angle of displacement in Figure 11b. When ~he a.c~ electrical ~ignal again reacnes zero cro~sing, the first side 524 and the second side 526 again reeurn to ~heir rest posi~ions and the fiber 500 thus return~ to its re~t position as illust_ated in Figure 11eO In the preferred embodi~ent, the a.c.
electrical signal is applied to the tran~ducer 510 so that the fiber 5~0 i~ periodically displaced ~o ~hereby induce a flexural wave in the fiber 500 which propagates as a ~eries of traveling ~icrobends along ~he length of the fiber 500 away from the eransducer 510. The traveling microbends have a :Erequency that is determined by frequency of the a.c. ~!~ource 520.

~92~

The optical fiber S0~ is preferably oriented so that the major axis of the elliptical core i~ ~ligned with the movement of the transducer 510 and thu~ the major axis lies on the plane of the traveling microbends. A~ set for~h above, this maximizes the coupling induced by ehe microbends .
The presen~ invention preferably includes a fi-st damper 536 fo_med of damping material which ~urrounds the fiber 500 at a location proxima~e to ~he side 524 of the ~1~ transducer 510. Thus, any flexural wave ~hich travels ~way from the tran~ducer 510 in tne direction eowards the damper 536 is suppres~ed. Therefore, the flexural waves travel away from the transducer 510 in one direction only, as indicated by an arrow 540 in Figure 10. Tne damper 536 can advantageously be ~upported by a first support block 542. Conventional optical fibers often have an outer plastic jacket that protects the cladding of the fiber.
In the present invention, the plastic jacke~ is removed to expose the cladding that is to propagate the acoustic wave. It has been found that the damper 536 can be advan~ageously formed by leaving a portion of the plastic ~acket on the fiber oueside the interaction region of the fiber with the aCouetic wave. The first ~upport block 542 can also ~erYe a~ ~ mounting block for ~he transducer 510 tO hold the transducer 510 in a fi~ed relationship to the damper 536. The present invention also preferably includes a seccnd damper 544, formed of damping material (such a~ the plastic ~acket of ehe fiber) through which ~hP fiber 500 pas~e~, to thereby suppre~s any further propagation of the trsveling microbends ~o that the microbend~ have no fur~her ef~ect beyond the second damper 544. The second damper 544 i8 preferably ~upported by a second support block 546. The second damper 544 is positioned 80 that only a s~lected length of the optical fiber S00 is affected by the traveling microbends. thus defining an interaetion length of the optical fiber 500~

In some applications, in which R long interaction length may be desired, the second damper 544 may not be necessary as the traveling microbend wave will be attenuated by the length of the fiber 50~. The opeical fiber 5~ can be suspended in air, vacuum, or another medium between the first support block 542 and the ~econd support block 546. Tne medium can be any material which does not attenuate the traveling microbend waves and which does not conduct any of the energy away from the optical fiber 5~0. It is not neces~ary that the fiber 500 be taut between the first and second ~upport blocks 54~, 546, nor is it necessary that the fiber be straight so long as the fiber 50~ is not bent with a radius sufficiently small so that the optical signal in the fiber 500 is perturbed by the bend.
The frequency of the electrical signal applied to the transducer 510 is chosen 80 that the flexural wave thus produced has a wave length along the fiber which i~
substan~ially equal ~o the beat length LB as indicated in Figure 10. Thus, a~ discus~ed above, ~he coupling of op~ical energy from one propagation mode to the other propagation mode will be reinforced in each section of the fiber. ~owever, unlike ehe previously discussed ~tatic microbend device, the microbend~ in the f iber 501U
propaga~e along the length of the fiber 500 at a velocity vp. The propagation velocity vp i~ determined by the particular characteristics of the fiber 500. The frequency of the elec~rical signal applied ~o rh2 transducer 510 (referred to hereinafter as fa) is ~elected ~o that the wave length (referred to hereinaf~er a3 ~a) of the propagating microbend i8 sub3tantially equal to the beat length LB~ Since the frequency fa i~ equal to the velocity vp divided by the wave length ~a~ then the frequency fa i8 determined by:
f vp ~ vp ~Z~ ~9 The angular frequency ~a of the electrical signal is ~a = 2~fa-It has been shown that when a propagating acousticwave causes a periodic, traveling stress on an optical fiber, the effect of the traYeling acoustic wave is to cause light to be coupled from one polarization mode to another polarization mode and be shifted in frequency.
See, for example, I'Single-Sideband Frequency Shifting in Birefringent Optical Fiber," W. P. Risk, et al., SPIE Vol.
473 - _iber optic and Laser Sensors II (1984~, pp. 91-97, in which this effect is discussed with respect to coupling between polarization modes in a birefringent fiber. A
similar effect has been described for multimode fibers for an externally applied stress to the fiber. See, for example, copending U.S. Application Serial No. 556,~36, "Single~mode Fiber optiz Single-sideband Modulator," filed November 30, 1983, and assigned to the same assignee as the present application, now U.S. Patent No. 4,684,215, issued on August 8, 1987. Thus, an optical signal, illustrated as an arrow 550, exitiny from a second end portion 552 of the fiber 500 exits at an angular frequency ~s, which is shifted in frequency from the angular frequency ~o which was input at the first end portion 502 of the fiber 500. The frequency ~s is equal to the angular frequency ~o plus or minus the angular frequency ~a f the signal applied to the transducer 510 (i.e., ~s = ~0 + ~a) Whether the frequency ~a is added to or subtracted from the frequency ~o is determined by whether the signal is input in the first order LPo1 mode or the ~econd order LPll mode and whether or not the optical signal is propagating in the same direction as the propagating microbend. The embodiment of Figure 10 is bidirectional in that the optical signal ~0 can be introduced into the second end portion 552 and thereby be caused to propagate towards the first end portion 502 in a '~

8~

direction opposite the direction of propagation of the traveling microbend.
As set forth in the above-referenced paper, "Single-Sideband Frequency Shifting in Bir~fringent Optical Fiber," by W. P. Risk, et al., when a traveling acoustic wave stresses an optical fiber having an optical signal propagating therein in the same direction as the traveling acoustic wave, the frequency o~ the traveling acoustic wave will be subtracted from the frequency of the optical signal if the optical signal is initially traveling in the slow optical mode (a first polarization mode in the Risk paper~. On the sther hand, if the optical signal is initially traveling in the fast optical mode (a second polarization mode in the Risk paper), the frequency of the acoustic wave is added to the original frequency ~0 of the optical signal. A similar effect occurs when the fiber is flexed by the traveling microbend having a frequency ~a in the present invention. The light input in the slow LPol optical mode at the frequency ~0 is coupled from the slow LPol optical mode to the fast LPll optical mode and is downshifted in frequency by an amount ~a to a frequency shown as ~11 (i.e., ~ o ~ ~a). On the other hand, light initially input at the frequency ~o in the fast LPll optical mode is shifted upward in frequency by an amount ~a to a frequency ~01 (i.e., ~01 = ~0 ~ ~a) in the LPo optical mode.
When the acoustic wave is traveling in the opposite direction of the light wave, the coupling from the fast LPl1 optical mode to the slow LPol optical mode causes a downshift in the frequency from the original frequency ~o to a new frequency ~01 (i.e., ~01 = ~0 ~ ~a)- This is the opposite effect from the frequency shift that occurs when the optical signal and the micr~bends are propagating in the same direction. Similarly, when the light initially travel~ in the slow LPol optical mode~ the light is coupled to the fast LP11 optical mode and shifted upward ~L~ ,. _.`.`i 2 ~ 8 in frequency. The csupled light has a frequen~y ~0 (i.e., o ~ ~a)-The foregoing can ~lso be considered in ~erms of the summation of the propagation constants of the two optical modes and the traveling microbends. For proper phase matching be~ween the traveling microbends and the optical signal, ~e propagation constants mus~ satisfy the following mathematical relationship:

~11 + ~a ~01 ~8) Thu~, when the optical signal is initially traveling in ~he LPll mode, the propagation constant ~a of the travellng microbends is added to ehe propaga~ion constan~
~11 of the optical ~ignal in ~he LPl1 optical mode to ob~ain the propagation constant ~ f the LPol op~ical mode a3 set forth in Equation (8) above. Simila~ly, when ehe optical signal is initially traveling in the LP~l optical mode, the propagation constant ~a of the eraveling microbends is subtracted from the propagation constant ~01 of the LPo1 mode to obtain the propagation con3tant ~11 of the LP~1 sp~ical mode a~ follows:

~01 ~ ~a ~ 9) The resultin~ frequency of he coupled optical signal depends upon whether the velocity o~ the traveling microbends i8 ~n the same direction ~s the direction of propagation of the optical ~ignal~ or in ~h~ oppo~ite direction of the propagation o~ the op~ical ~ignal.
The frequency shiting can be e~cpres~ed ~aehematically by represeneing eh2 light in the fast LP1l opeical mode as co~(~Ot - ~11Z)~ where ~0 i8 the initial frequency of the lnput light, t i8 ti~e, and Z is the distance in the direction of propagation 540 of the traveling microbends along the fiber 500. The traveling microbends may be -3~-represented as css(~at - ~aZ~ where wa is the frequency of the traveling microbends . ~a i8 the propagation con tant of the traveling microbends and ~ i5 the distance alon~ the axis of ~he fiber 500 in the direction of propagation of the microbends. The interaction of ~he optical signal with the traveling microbends leads to a product term proportional to ~he following expression:

1/2{cos[~wo+~a)t~ a)zl+cos[(~u-~a)~ a)z]} (10) The &ecood term of Expression (10) does not satisfy the phase matching condition of ei-ther Equation (8) or Equation (9) above. The first ter~- in the expression is phase matched in accordance wi~h Equation (8), This match explicitly indicates that the optical signal in the LPo1 mode i8 upshifted in frequency to the frequency ~01 ~ ~0 ~ ~a. A sim~lar analysis for interac~ion of an optical si~nal in the LPol mode leads to a product term propor~ional to the ~ollowing expression:
~0 l/2~cos~(ou-~a)t-(~o~ )z]~cos~(wo~5~a)~ ol~a)z3}

The ~econd term in Expression (11) is not phase maeched in accordance with either Equation (8) or Equation (9) above. The first term does meet ~he phase matching requirements of Equa~ion (9). This explicitly indicates t~at the LP11 mode i8 downshifted in frequency to the frequencY w~ o ~ ~a-If ~he optical ~ign~l propagates in ~e opposite 3g direc~ion as the travelin~ microbend~, t~e traveling microbend can be represented as cos(~a + ~aZ) The interaction of an optical signal in ~he LP1~ mode with t~e ~raveling microbend lead~ eo a product term proportional to ~he following expression:

l2{co5l(~o~a)t~ a)z~+cos~o-~)t~ +~a)z]} (12 9 ~ ~ 9 The first term in Expression (12) does not provide prope~
phase matching in accordance with either Equation (~) sr Equation (9). The ~econd term doe~ provide proper phase matching in accordance with Equation (8). Thus, when the optical signal propagates in the opposite directlon to the traveling microbends, the light coupled to the LP~1 mode is downshifted in frequency from the lignt input in the LP11 mode to a frequency ~ 0 ~a ins~ead of being '10 upshifted as discussed above ~ith regard ~o Expression (10) for lîght propagating in the same direction. When the optical 3ignal i8 initially in the LPo1 mode and propagaees in ~he opposite direc~ion as the t aveling microbends, ehe interaceion of the optical signal and the microbends lead~ to a product term proportional to the following e~pression:

1/2{cost~o-~a)t-(~o1+~a)z]+cos[(~o~a)~ ol-~a)z]} (13) The fir~t term. in Expression (13) i8 not properly phase maeched in accordance wi~h either Equation (~) or Equation (9). The second term is properly phase matched in accordance with Equation (9), Thus, the op~ical energy coupled from ~he LPol mode to the LP11 mode is shifted upward in frequency ~o a frequency ~ wo ~ ~a.
The pre~ent invention has many advantagesO For example, the presen~ inven~ion u~es a highly elliptical core ~wo-mode fiber, ~nd i~ operated at a frequency wherein ~he L~ll odd mode is cut off. Thus, optical energy in the LP11 mode will propaga~e onLy in the LPll even mode. Thu~, the pre~ent invention does not require precise ~lignment of the major axi~ of the elliptical core with the input light source.
Tbe presen~ inventi.on is particularly advantageous in that it operate~ sver a broad range of optical frequencies for the input optical signal. This advantage esults from the use of the spatial propagation modes for coupling. A5 set forth above, the beat length of the two sp~tial propagation modes (e.g., the LPo1 and LP1l modes) does not vary significantly over a broad optical frequency range.
S Thus, the beat length will match the wavelength of the traveling microbend wave even when the optical frequency of the input optical signal (i.e., fo ~ ~vl2~) i3 va~ied over a broad frequency range about the selected cen~er frequency where the optical beat leng~h and the wavelength 1~ of ~he ~raveling microbends match exactly. In order to take full advantage of the characteristics of the highly elliptical core optical fiber 500, the center frequency fo i5 preferably selected to be in the central portion of the frequency range between ~he fir~t cutoff frequency fc and 1S the second cutoff frequency fc2~ and the fre~uency shifter is operated well within the range of frequencies between f~ and fc2 The frequency of the a.c. Rignal modulation applied to the transducer 510 (Figure 10) can also be varied over a relatively broad range in the present invention. The broad range of ~odulation frequencies results from two features of the present invention. The first feature of the present inveneion that provides for operation over a broad modulation frequency range is that it couples optical energy bet~een the ~patial propagaeion modes. The beat length of the two spatial propagation ~odes is ubstantially ~maller than the bea~ length between the polarization modes of an optical ~ignal at the same frequencyO Thus, the present invention operates at a 3~ higher absolute modulation frequency than an exemplary deYice which couples optical energy between polarization modes. The present invention operates with optical beat lengths of appro~imately 50~m to 500~m and thus can operate with a microbend frequency of approximately 3 MHz 3j; to 50 M~z-Z~9 The second feature of the p_esent invention that provides for operation over a broad modulation frequPncy range is that the modulation energy from the transduce-510 (Figure 10) i~ coupled direc~ly into ~he fiber 500 to induce ~he traveling, periodic microbend wave. The present invention does not require a substrate or other medium external ~o the fiber 500 to conduc~ ~he ~odulation energy to the fiber 500. Thus, a relatively large percentage of the modulation energy acts upon the fiber 500 to create the eraveling microbend ~ave. Therefore, for a given modulation energy input, it is believed that a larger percentage of the optical energy i~ transferred from one spatial propagation ~ode to the other spatial propagation ~ode in each beat length. Thus, relatively fewer beat length~ of interaction between ~he optical signal and the traveling microbend wave are required to couple substantially all of the optical energy from one spatial propagat~on mode to the other 8patial propagation mode. It has been shown that coupling between the spatial propagation modes will occur even wh~n there is a small percentage of deviation from the modulation wavelength which corresponds ex~cely wi~h the beat leng~h of the two spatial propagation modes 80 long a the deviation i~ not allowed to accumulate to a large eotal percentage of deviation over a large number of beat len~ths. Thus, since relatively few beat len~thQ are required in the present invention to couple the optical energy from one spatial propagation mode to ~he other spa~ial propa~ation mode, the acceptable percentage of deviation in one beat lengeh can be relatively large (relative to a device requiring a large number of beat lengehR). The relatively large percenta~e of accep~able deviation in wavelength combined with ~he relatively large ab~olute ~odulation frequency results in a relatively broad range for ~he modulation frequency.

2~89 The present invention is particularly advantageous in that ehe energy required to cause the coupl~ng between the modes is concentrated entirely within the fiber 500.
Substantially all of ~he energy applied to the transduce 510 is transferred to the optical fiber 500 to produce the pe' iGdiC microbends. Thus, very little, if any energy is wasted in the present invention. Thus, the mechanical ene_gy generated by the transducer 510 i~ utilized very efficiently.
Figure 12 illustrates a preferred embodiment of the frequency shifter of ~he present invention in which ~,e transducer 51~ of Figure 10 i8 replaced with a transducer 700 particularly adapeed to generate the periodic microbends described above. The transducer 700, in the preferred embodiment comprises a rod of fused quarez having ~ substantially circular cross 3ection throughout its length. A first end 702 of the transducer 700 has a diameter substantially equal to the diamee2r of the fiber 500. For example, in one embodiment of the present invention, the diameter of the fiber 502 and of the first end 702 of the transducer 700 is ~pproximately equal to 100~m. Preferably, the ~ransducer 700 and the optical fiber 500 are fused ~ogether at a location 704 to provide good acous~ic eon~act between the fiber 500 and the second end 702 of the tran~ducer 70~.
The transducer 700 has a seeond end 710 which has a diameter which i~ substantially larger than the diameter of the first end 702~ For exa~ple, the second end 710 can have a diameter of appro~imately ~wo millimeters. In the preferred embodimen~, the ~ransducer 700 ~ formed from a hollow tube of fused quartz having an ini~ial diameter of t~o millimeters or larger and by drawing the quartz tube into a form which gradually tapers from the ~econd end 710 eo the smaller first end 70~. Thus, the transducer 700 is 3~ hollow at the ~econd end 710 and i~ subs~aneially closed off (i.e., solid) at t:he fir~t end 702. Fur~her details 2~8~

of the construction of the preferred embodiment of the transducer 700 are set forth in copending Canadian Application No. 527,487, entitled "FIBER OPTIC INTER-MODE
COUPLING SINGLE-SIDEBAND FREQUENCY SHIFTER," fil~d on January 16, 1987, and assigned to the assignee of the present application.
A piezoelectric material 712, such as PZT, is bonded to the second end 710 of the transducer 700 in a manner well-known to the art. When an electrical signal, represented schematically as a signal generatsr 714, is applied to the piezoelectric material 712, the piezoelectric material 712 expands and contracts in the directions indicated by the double-headed arrow 716 and generates a series of acoustic wavefronts which propagate through the transducer 700 from the second end 710 to the first end 702, as indicated by an arrow 718. At the first end 702, the acoustic energy in the transduser 700 is coupled directly to the optical fiber 500 at the location 704 to cause up and down movement of the fiber 500, thus inducing a vibration in the fiber 500 which propagates away from the location 704 as a traveling flexural wave or traveling microbend as described above with respect to Figure 10. The surface of the quartz transducer 700 acts as an acoustic funnel which concentrates the acoustic energy developed at the second end 710. Furthermore, substantially all of the acoustic energy applied to the ~econd end 710 is conducted to the first end 7~0 and is used to induce the traveling microbend in the fiber 700.
As in Figure 10, the embodiment of Figure 12 further includes the first damper 536, proximate to the location 704, to limit the travel of the microbend in the fiber 502 to one direction, indicated by an arrow 720, away from the damper 536. The embodiment of Figure 12 also preferably includes the second damper 544 to suppress propagation of ~ ~',~'i ~2(~89 the microbends beyond a selected length of the fiber 500 as discussed above.
Detailed Description of the Mode Selector As shown in Figures 13-20b, an inter-mode selector comprises a first optical fiber 1100 and a second fiber 1110. The first optical fiber 1100 has an inner core 1102 and an outer cladding 1104. The second optical fiber 1110 has a highly elliptical inner core 1112 and a outer cladding 1114. The core 1102 of the first optical fiber 1100 has a core refractive index nCorel~ and the cladding 1104 has a cladding refractive index ncladdingl The core refractive index and the cladding refractive index of the first fiber 1100 are chosen such that the core refractive index is greater than the cladding refractive index ~i.e., nCorel > ncladdingl) Therefore, light propagating in the core 1102 will propagate at a slower phase v210city than light propagating in the cladding 1104. In like manner, the core 1112 of the second optical fiber 1110 has a core refractive index nCore2 and the cladding 1114 has a cladding refractive index cladding ncladding2 which are selected so that the core refractive index is greater than the cladding refractive index (i-e~, ncore2 >
ncladding2) In the preferred embodiment, the core refractive index, the cladding refractive index, and the diameter of the core of the first optical fiber 1100 are selected so that the first optical fiber 1100 is a single-mode optical fiber at a selected optical *requency fo. The core refractive index, the cladding refractive index, and the diameter of the core of the second optical fiber 1110 are selected so that the second optical fiber 1110 is a multimode (i.e., a two-mode) optical fiber at the same selected optical frequency fo. Thus, the first optical fiber 1100 will propagate only light in the first order LPol mode at the selected optical frequency. The second optical fiber 1110 will propagate light in the first order LPo1 mode and will also propagate light in a higher order ~2~

mode, namely, the second order LPll propagation mode. As illustrated in Figures 15, 17a, and 18a, the core 1112 of the second optical fiber 1110 preferably is highly elliptical in accordance with Figure 5, and Figures 6a-6h, above. The ellipticity of the core 1112 is selected so that, at the frequency fo, only the even lobe patterns of the LPll propagation mode will propagate in the optical fiber 1110. The wavelength of the optical signal at fO is above ~he cutoff wavelength for the LP11 odd modes so that the LPl1 odd modes do not propagate.
The first optical fiber 1100 is arcuately mounted in a first mounting block 1120. The second optical fiber is arcuately mounted in a second mounting block 1139. In the preferred embodiment, the first and second mounting blocks 1120, 1130 are constructed in accordance with the teachings of U.S. Patent No. 4,536,058, which is incorporated herein by reference. The first mounting block 1120 has a flat mounting surface 1140 into which an arcuate slot 1142 is cut to provide a guide for the first optical fiber 1100. As described in U.S. Patent No.

4,536,058, the slot 1142 has a depth with respect to the mounting surface 1140 at each of two ends of the first mounting block 1120 that is greater than the depth at the middle of the mounting surface 1140 so that when the first optical fiber 1100 is positioned in the slot 1142, a portion of the cladding 1104 on one side of the first optical fiber 1100 is proximate to the mounting surface 1140. The mounting surface 1140 is polish~d so that the cladding 1104 of the first optical fiber 1100 is gradually removed with the surface 1140 to form a facing surface 1144 on the cladding 1104 which has a general oval shape that is coplanar with the surface 1140, as illustrated in Figure 14. The polishing is continued until a sufficient amount of the cladding 1104 is removed so that the facing ~urface 1144 is within a few miorons of the core 1102 of ~, ,.~,, g the first optical fiber 1100. In like manner, an arcuat~
slot 1152 is formed in a mounting surface 1150 of the second mounting block 1130 and the second optical fiber 1110 i5 positioned in the slot 1152. The mounting surface 1150 and the cladding 1114 of the second fib r 1110 are polished in the above-described manner to form a facing surface 1154.
As illustrated in Figure 15, the facing surface 1144 of the first optical fiber 1100 is positioned in juxtaposed relationship with the facing surface 1154 of the second optical fiber 1110 to form an interaction region 1156 (labelled in Figure 14) for transferring light between the fibers. The rore 1102 of the first optical fiber 1100 and the core 1112 of the second optical fiber 1110 are spaced apart by the thin layer of the claddinq 1104 remaining between the facing surface 1144 and the core 1102, and the thin layer of the cladding 1114 remaining between the facing sur~ace 1154 and the core 1112. The removal of the cladding is preferably performed in accordance with the method described in U.S. Patent No.
4,536,058. The oil drop test described in U.S. Patent No.
4,536,058 is advantageously used to determine the amount of cladding remo~ed and the proximity of the facing surfaces 1114, 1154 to the cores 1102, 1112, respectively. As discussed hereinafter, cladding is removed from the first ~iber 1100 and the second fiber 1110 until the evanescent field penetration of the facing surfaces 1144, 1154 for the ~elected guided modes of the fibers is sufficient to cause coupling of light between tha two guided modes.
As further illustrated in Figure ~6, the two-mode optical fiber 1110 is preferably oriented so that the major axis of the elliptical core 1112 is normal to the ~acing surface 1154.
When the facing surface 1144 and the ~acing surface 1154 are superimposed, as illustrated in Figure 13, the ~ ;.. ~

1 ~2~?8~

first fiber 1100 and the second fiber 1110 converge near the center of the mounting blocks 1120 and 1130 and diverge gradually as the distance away from the center of the blocks 1120 and 1130 increases. The rate of convergence and divergence of the two fibers is ~etermined by the radius of ~urvature of the two arcuate grooves, which, in one preferred embodiment, is selected to be 25 centimeters. This radius of curvature permits the cores of the two fibers to be positioned in close proximity to permit the evanescent fields to interact while limiting the 1 ngtn of the interaction region 1156. As explained in detail by Digonnet, et al., in "Analysis of Tunable Single Mode Optical Fiber Coupler," IEEE Journal of ~uantum Electronics, Vol. QE-18, No. 4, April 1982, pp.
746-754, and in U.S. Patent No. 4,556,273, the teachings of which are incorporated herein by reference, the length of an interaction region of two juxtaposed fibers is defined principally by the radii of curvature of the fibers, while the strength of coupling is defined principally by the proximity of the cores in the interaction region, particularly the minimum core spacing (i.e., the distance between the cores at the centers of the facing surfaces 1144 and 1154). The length of the interaction region is preferably at least an order of magnitude larger than the maximum cross-sectional dimension of the core of either of the fibers so that there is a substantial amount of light transferred, and preferably a complete transfer of light between the two fibers. The interaction region length increases with increasing radii of curvature and the strength of coupling increases with decreasing core spacing. The radii of curvature are preferably selected to be sufficiently large ~o that little, if any bending effect is introduced into the fibers. Further, the core spacing is preferably no less than zero so that the diameters of the cores 1102, 1112 are uniform throughout the length of the apparatus of ~ j ; .

~he invention, and, thus, no modal pertu~bat~ons are introduced by changes in the characteri~tics of ~he fiber~
The operation of the present invention can be more fully understood by referring to Figu-es 16a-b, 17a-b, 18a-b, 19a-c, and 20a-b, Figures 16a and 16b illustrate ~he elec~ric field intensi~y distribution for the LPol propagation mode of an op~ical si&nal propagating in the single-mode optical fiber 1100~ A~ illus~rated in Figure 16b, the field in~ensity distribution iQ defined by an electric field intensity distribution graph 1200 which has a maximum 1202 ~ubstantially in the center of ~he core 1102. The intensity decreases as the distance from the center of the core 1102 increases. Most of the electric field ene-gy of the LPo1 propagation ~ode iB concentrated in ~he core 1102 of the fiber 1100. ~owever, a portion of the electric field energy e~ends into the cladding 1104, as illustrated by B field portion.1204 and a field portion 1206 of the graph 1200, The portion of the elec~ric field energy outside the core 1102 i8 referred to as the evane~cent field. The interaction of the evanescent fields of two guided modes within the interac~ion region 1156 provides the means for coupling opeical energy be~ween the t~o fibers 1100, 1110, as described in U.S.
Paten~ No. 4,536,058. The amount of coupling i8 determined principally by the leng~h 6f ~he in~eraction region and the dis~ance between the respective co_es of the two fibers.
Figures 17a and 17b illus~rate ~he electric field intensity tiRtribution for lighe propagating in the LPol and LP11 modes of the ~wo-mode fiber 11 lO at the ~elected requency~ The electric field intensity distribution of the LPol propagation mode i~ illu~trated by a graph 1220, and ~he electric f:ield distribution of the LP11 prop~gation mode ~s illustrated by grap~ 1222. As di~cus~ed above with regard to the graph l~OO in Figure 16b, the electric field of an optical signai propagating in ~he LPo1 propagation mode is concentrated in the core 1112 near the center of the fiber 1110, as illustrated by a maximum 1224, and the intensity diminishes as the dis~ance from the cen~er of the core 1112 increa.~es. A
portion of ~he electric field intensity distribution of the LP~1 propagation mode extends into the cladding 1114 as an evanescen~ field, designated aQ portions 1226 and 1228.
The electric field intensity distribution g aph 1222 for the light propagating in the LP11 propagation mode has a minimum intensity 1230 near the center of the core 1112 and the intensity increases gradually as the distance from the core 1112 increases. The maximum electric field intensity occur~ ae maxima 1232 and 1234 which are proximate to the interface between the core 1112 and the cladding 1114 of the optical fiber 1110. Note ~hat the intensity distribution is proportional ~o the square of ~he mplitude distribution 80 tha~ ~he polarity of both maxima are the same. Thus, ~he evanescent field of an optical signal propagating in the LPl1 propagation mode, represented by electric field energy di~tribution portions 1 36 and 1238, extends fur~her ~nto the cladding than the evanescent field of ~he por~ion of an op~ical signal propagating in the LPo1 mode of the 6econd optical fibe_ 1 1 1 0 .
As illustrated in Figure 17b, the optical jignal propagating in ehe LPo1 propagation mode of the second optical fiber 1110 has very little evanescent field 3~ ~xtending beyond ~he core 1112 of the fiber 1 l l O (see the portions 1226 ~nd 1227 of the graph 122~. In contrast, t~e evanescent field of an optical signal propagating in the LP11 propagation mode e~tends well into the clsdding 1114 of the second optical fiber 1110 ~see the portions 1236 and 1238 of the graph 1222). Thus, when a portion of the cladding is removed from e~ch of the two optical ~2C8 fibers so that the cores 1112 and 1102 can be closely j~x~posed, there is a relatively large evanescent field intensity for the LP11 propagation mode ~vailable to interact with the core 1102 of ~he first ~ptical fiber 1100. This is illus~rated in Figures 1 8a and 18b, wherein the electric field intensity distribution graphs 1220 and 12~ of the second optical fiber 111~ a-e ~hown in relation to the electri~ field intensity distribution graph 1200 of the firs~ optical fiber 11U0 when the two fibers are jux~aposed at their respective facing su~aces 1154 and 1144. As illustrated, there is ~ubstantially no overlap of the LPo1 electric field intensity distribution 1220 of the ~econd optical fiber 1110 with the LPo1 electric field intensity distribution 1200 for of the first optical fiber 1100. ~owever, there is relatively large overlap of the evanescent field portion 1238 of the LP11 electric field intensity distribution 1222 of the econd optical fiber 1110 with the evanescen~ field portion 1204 of the LPol electric field intensity 2~ distribution 1200 of the first optical fiber 1100. Thus, ~he LPll prop~ga~ion mode of the second optical fiber 1110 and the LPo1 propagation ~ode of ~he first optical fiber 1110 interact ~trongly as compared ~o ~he interaction between the LP~1 modes of ~he ~wo ~ibers. Due to the strong interac~ion be~ween the LP11 mode of the two-mode fiber 1110 and the LPol mode of the single-mode fiber 1100, a relatively large amount of op~ical energy is coupled between the LP~ 1 propagat ion mode of the f iber 111(~ and the I.Po1 propagatic~n mode o~ ~he fiber 1100.
~owever, since there is substantially no interaction between the LPol ~odes of the two fibers 1100, 1110, there is substantially ~o optical ener~y coupled between the LPo1 propagation mode of ~he fiber l 110 and the LPol mode of the fiber 1100. As di~cussed hereinafter, coupling between the LPll mode of the fiber 1110 and the LPo1 mode of the fiber 1100 is enhanced by electing the fibers such ~2~8 that the propagation constants of these two ~odes a~e the ~ame. Similarly, coupling be~ween the LPo1 modes of the fibers is inhibited by ensuring that the LP~1 ~odes of the fibers have ~ubstantially different propagation constants.
As set forth above, the two-~ode optical fiber 1110 has a highly elliptical oore 1112 in which only the LP11 even m~de lobes can propagace in ehe fiber 1 110 at the selected frequency of the optical signal. By orienting the major axes of the elliptical core 1112 go that it is normal to the facing surface 154, maximum coupling will be provided from the LP11 mode, two-mode fiber 1110 to the LPo1 mode of ehe single-mode f iber 1~10.
As discussed above, the energy`~~of an optical Rignal propag2ting in the LPo1 mode of a fiber is confined mos~ly within the core of ~he fiber. Thus, an optical signal in the LPo1 propagation mode propagates at a phase propagation velocity which is largely dete~mined by the refractive index of the core of the fiber. The effective refractive $ndex seen by the light propagating in the LPo1 propagation mode ha~ a value close ~o the value of the refractive index of the core. In contrast to the LPUl mode, the LP11 propagation mode propagates a larger fraction of optical energy in the cladding of the fiber.
Thus, the phase propagation veloci~y of ~n optical signal propagating in the LP11 propagation mode is determined by an effectlve refractive index ~hich has a value less than the refractive inde~ 9f the oore and greater than the refractive inde~ of the cladding. Accordingly, the ef~ective refractive index for light in the LP11 3~ propagation ~ode has a value which is less than ~hat for light in the LPo1 propagation mode. 'l'hu~, the phase propa&ation ~elocity of an opticsl signal propagating in the LP11 propagation mode is faseer than the phase propagation veloci~y o~ an op~ical ignal propagating in 3!; the LP~l mode. Due to the difference in phase velocities o the LP~1 and LP11 ~odes, these modes do not easily 2 ~ 8 9 couple; rather, they ~end to function ~Q two independe~t opticsl paths. In the preferred embodiment, elective coupling between the LP11 mode of the fiber 111~ and the LPo1 mode of the fiber 1100 is accomplished by utilizing dissimilar fibers such that ~he phase propagation velocity of the LP11 mode of the fiber 1110 matches the phase propagation velocity of the LPU1 mode of ~he fiber 11U0, while the phase propagation velocities of all othe~ pairs of modes are not ma~ched.
'1~ From the foregoing, it will be understood that the propa~ation velocity of an optical signal propagating in the LPo1 mode of the fi_~t (single-msde) op~ical fiber 1100 is determined in pare by a combination of the refractive index of the core 1102, the refractive index of the cladding 1104, and the diameter of the core 1102.
Similarly, the phase propagation velocities of op~ical signals propagating in the LPo1 mode and ~he LP11 mode of the second (two-mode) optical fiber 1110 are determined by the combination of the refractive inde~ of the oore 1122, - 20 the refractive index of the cladding 1124, and the diameter of the core 1 122. As set forth above, the phase propagation velocity of an op~ic~l ~ign~l prop~g~ting in the LP11 mode in the opeic~l fiber 1110 will be greater than the pha~e propagation veloci~y of an optical signal propagating in the LPo1 mode of ~he optical fibe~ 1110.
In the present invention, the first optical fiber 110~ and ~he ~econd optical fiber 1110 are ~elected to have ch~racteristics (i.e., refractive indices and core dimensions) such that the phase propagation velori~y of an 3~ optical si~nal in the LPl1 propagation ~ode of the second opeical fiber 1110 is mstched to the phase propagation velocaty sf qn optical signal in the LPo1 propagation mode ~f ~he ~irst opeical fiber 110U. Tnus, the phase propagation velocity of an optical 8ign81 propagating in ~he LPo1 mode in the ~econd optical ~iber will be slower than the phase propagation velocity of an optical ~ignal ~8 propagating in the LPl 1 mode of the second optical fiber 1110 and will al~o be slower than the phase propagation velocity of an optical ~ignal propagating in the LPol mode of the firs~ optical fiber 1100. The relationship among the phase propagation velocities is illust ated in Figures 19a, 19b, and 19c.
In Figur4 19a, a t-aveling wave 125U represents the phase propagation of a signal propagating in the LPo1 propagation mode of the firsc (single-mode~ optical fiber 1100. The traveling ~ave 1250 has a first wavelength ~1 which is defined by a velocity V1 of the optical ~ignal in the LPo1 propagation mode of the fiber 1100 and the ~elected optical frequency fo (i-e-, ~1 ' Vl/fo).
Similarly, Figure 19b illustrates a traveling wave 125~
that represents the pha~e propagation of a signal traveling in the LPo1 propagation mode of the second (two-mode) optical fiber 1110 at the ~elected optical frequency fO~ The traveling wave 1252 has a wavelength ~2 which is defined by a velocity V2 of the optical signal in the LPo1 propagation mode of the second optical fiber 1110 and the ~elected optical fr~quenc9 fO (i~ 2 - V2/fo)- Figure 19c illustrates a traveling wave 1254 which repres2nts the pha~e p~opagation of a signal traveling in the LP11 propagation mode of the optical fiber 1110 at the selected optical frequency fO. The travelling wave 1~54 has a wavelength ~3 which i5 defined by a velo ity V3 of the optica~ signal in the L~11 propagation mode of the second optical fiber 1110 and ehe selected optical frequency fO
(i-e-, ~3 c V3/fo)- As ~et forth above, the characteri~tics of the fir~t optical fiber 1100 and che ~econd optical fiber 111~ are chosen ~o that the velocity V1 i8 substan~ially equal to the velocity V3. T~us, the velocity V2, which mus1; be less than the velocity V3, is le~ than the velocity Y1. Therefore, the wavelength ~3 3~ matches the wavelength ~1 and the wavelength ~2 does not match the wavelength ~. Accordingly, as illustrsted in z~9 ~igures 19a and 1~c, light travelling in the LP11 propagation mode of the econd optical fiber 1110 will have a ~ubstantially fixed phase relation to light travelling in the LP~1 propagation mode of the fir~t optical fiber 1190- ~n the other hand, as illustrated in Figures 19a and 19b, light traveling in the LPo1 propagation mode of the ~econd opti~al fiber 1110 will have a continually varying phase relation to light travelling in the LPo1 propagation mode of the first optical fiber 1100.
It has been ~hown tha~ to couple light from a mode of one fiber to a mode of another fiber, ehe modes l3hould interact through their evanescent fields over an interaction length, and the phase propaga~ion velocities of the two opticsl signals should be substantially equal. In the preferred embodiment, the phase propagation velocity V3 of ehe LP11 propagation mode of the l3econd optical fiber 1110 at the selected frequency is substantially equal to the phase propagation velocity V1 2~ of the LPo1 propagation mode of the first optical fiber 1100. Thus, the hpparatus of ~he present invention provides a means for coupling optical energy between the LP~1 propagation mode in the 6econd optic~l fiber 1110 and éhe LPo1 propagation mode in the first optical fiber 1100. An optical ~3ignal propagating in the LPo1 propagation mode in the second optical fiber 1110 is not coupled to the firs~ optical fiber 1100 because the phase propagation velocity V2 for light propagating in the LPo1 mode in the second optical fiber 1110 does not match the phase propagation veloci~y V2 for light propagating in t~e LPo1 propagation mode of the first optical fiber 1100.
Thus, the appar~tus of the present ~nvention provides ~election between opti.cal energy propagating in the LPo~
mode of the ~econd optical fiber 1110 and optical signal propagating in the LP11 mode of the ~econd optical fiber 1110. Accordingly, as shown in Figure 13, when optical 2 ~ 8 energy is input in~o a fir ~ end 1300 o~ the ~econd optical fiber 1110 in ~he LPol propagation mode, represented by an arrow 1302, and in the LP11 propagation mode, represented by an arrow 1304, the optical energy in the LP11 propagation mode will be coupled to th~ firs~
optical fiber 1100 and will exit from an end 1310 of the first optical fiber 1100 in the LPU1 pr4pagation mode of the first optical fiber 110U, as represented by an a row 1312. In contrast, the optical energy in the LPo1 prspagation mode of the second optic~l fiber 1110 will remain in ~he second optical fiber 111 U and will exit from a second end 1320 of the ~econd optical fiber 1110, as represented by an arrow 1322. The first optical fiber 1100 therefore provides an output signal which corre~ponds to the optical signal incident to the interac~ion region 1156 of the two fibers propagating in the LP11 propagation m~de of the second optical fiber 111U.
The foregoing properties of the present invention are reciprocsl in that an optical ~ignal introduced into an end of the first optical fiber 1100 in the LPo1 propagation mode of the first opticsl fiber 1100 interacts wi~h the LP11 propagatioh mode of the second optical fiber 1110 in ~he interaction region be~ween ~he first fscing surface 1144 and the second facing ~urface 1154 to couple optical energy into the LP11 propagation mode of the ~econd optical fiber 1110. Thi~ reciprocal effect has advantages in many applications.
The apparatus of ~he present invention can be fine-tuned by orien~ing the two fibers 1100, 111U ~o ehat the longitudinal axes of ~he two fiber~ are not exactly parallel. Thus, if the phase propagation velocities V7 and V3 are not preei~ely the same at the selected frequency, the ~econd optical fiber 1110, for example, can be pO8 itioned 80 that the longitudinsl axis of the core 3,; 111~ of the second opt.ical fiber 1110 i3 at an angle with re~pect to the longitudinal axi~ of the core 1102 of t~e 2 ~ ~9 fi~st optical fiber 1100 such that the magnitude of the oomponent of the phase velocity V3 in the direction of the longitudinal axis o the core 1102 o~ the first optical fiber 1100 is the same as the phase velocity V1 along the longitudinal axis of the core 1102 of the fir~ optical fibe 1100. This is illuqtrated in Figures 20a and 20b, wherein a vector 1402 represent~ the phase propagation velocity Y3 of the LP11 propagation mode of the second optical fiber 1110 ~hown in phantom) ~nd a ve~tor 1400 ~0 represents the phase propagation velscity V1 of the propagation mode LPo1 in the first optical fiber 1100 (~ho~n in phantom). In Figure 20a, the two vectors a e positioned in parallel ~nd the magnitude of the vector 1400 is ~maller than the magnitude of the vecto~ 1402.
Thus, the phase propagation velocities V3 and V1 do not match and efficient coupling cannot occur. In Figure ~Ob, the ~econd optical fiber 1110 i~ or~ented at a small angle with respect to the firAt optical fiber 1100 so that the phase propagation velocity vector 140~ is oriented at a ~mall angle with respect to the phase propaga~ion velocity vector 1402. The component of the phase prspagation velocity vector 1402 in the direction of the phase pr~pagation velocity vector 14DO i8 ~hown in dashed lines as a Yec~or 1402' that represents a ~elocity V3'- The vector 14G2' has subRtan~ially the ~ame magnitude as the vector 1400 and thus the phase psopaga~ion velocities V3' and V1, represented by the vectors 1402' and 1400, match at the angle ~hown. Thus, efficient coupling can occur from the LP11 prop~gation mode of the second optical fiber 1110 to the LP~1 propagation ~ode of ~he first optical fiber 1100~ Al~hough, for clarity of ill~seration, the angle between ehe fibers i~ shown as a signif~cant angle in Figure 20b, ~t is prleferable that the angle between the t~o fibers 1100 and 1110 be ve_y small 80 ~hat ~he fibers are substantially par,allel at the interaction region 1156. In rhe present invention, it is de~irable ~o avoid ~Z~89 large angles between the fiberR to prevent a. ~ignificant reduction in the interaction length. As u~ed herein, the interaction length means the length in the direction o the fiber axis of one of the fiber~ (e.g., the fiber 1100) through which the core of the one fiber (e.g., the fiber 51100) is positioned within the evanescent field of the o~her fiber (e.g., ~he fiber 111 O) J
An exemplary application for the frequency shifter and mode selector of the present invention i8 illustrated for an inter-mode frequency ~hifter 1500 in Figure 21~ The 10inter-mode frequency shifter 1500 preferably includes an input optical fiber 1600 which i8 advantageously a single-mode optical fiber. The input optical fiber 1600 is butt-6pliced to a two-mode optical fiber 1602. A first portion of the t~o-mode optical fiber 1602 i8 formed into a coil 151604 which operates as a mode Btripper in a ~nne~ known to ~he art. Alternative mode 6trippers could be used. A
second portion of the ~wo-mode optical fibe_ 1602 lnterconnect~ the mode stripper 1604 with an lnter-mode frequency shifter 16060 The inter-mode requency shifter 201606 is preferabl~ constructed in accQrdance with the frequency shif~er described above in conneceion with Figures 9-12. A transducer 1610, driven by a modulation source 1612, induce~ vibration~ into the fiber 1602 which propagate in a ~ingle disection 1614 away from the 25tran6ducer 1610 ~& ~ flexure wave compri~ing a series of traveling microbends in an in~eraction region defined between ~ fir~t damper 1622 and a second damper 1624, suppor~ed by a fir~t support 1626 and a ~econd support 1628, respectivel~. The inter-mode frequency shifter 1606 30vperate~ to couple light from a first propagation mode (e.g., the LPol mode) to a aecond propagation ~ode (e.g., the LP11 ~ode) and ~o cause ehe light to be shifted in frequency by an amount determined by a modulation signal applied to the transducer 1610 of the frequency shifter 1606 from the ~odulation source 1612. After passing -5~-~hrough the frequency shifter 1606, the two-mode optical fiber 1602 interconnec~s the frequency ghifter 16~6 with a mode selector 1640 constructed in accordance with ~'igu es 1 3-2~b . In the mode selector 1640, the two-mode optical fiber 16~2 is juxtaposed with a single-mode optical fiber 16~ at an interaction region 1652 in the manner described above. The ~wo-mode optical fiber 1602 has an output end por~ion 1654. The ~ingle-mode opcical f iber 1650 has an output end portion i656.
~10 The inter-mode frequency modulator operates in the following manner. An optical ~i~nal, repr~sented by an arrow 166~, is introduced into the input optical fiber 1600. Preferably, the optical signal is propagating solely within the LPol propagation mode for the input optical fiber 1600. The optical signal propagates through the input optical f iber 1600 and is coupled to the two-mode optical fiber 1602 and propagates within the two-msde optical f iber 1602 in ~he LPo1 prop~gation mode. Any optical signal propagating in the LP11 propa~ation mode in the fiber 160~ i8 stripped from the ~wo mode optical fibe 1602 in the mode ctripper 1604 in a manner known to the art . Thus, af ter pass ing through ~che mode ~tripper 1604, the op~ical signal remaining in the ~wo-mode optical f iber 16()2 is propagating 80~ y in ehe LPo1 propagation mode.
The optical signal propagates in the LPo1 propagation mode through the ~wo~mode optical fiber 1602 to etle frequency ~hifter 1606. In 'che frequency ~hif~er 1606, the opeical signal propagating in the l.Po1 propagaeion mode is coupled to the LP1 1 propagation mode and i~ 6hifted in frequency by ~n amount de~ermined by the frequency of the modulation soùrce 161~. A~ set forth above, ~he optical signal coupled to ehe LP11 propsgation mode can be shifted upward in frequency or downward in frequency in accordance with the direction of propagation of traveling flexure waves in the frequency ~hifter 1606, and thus in accordance with the loc tion of the tr,~nsducer 1610 on the optical fiber ~2~2(: ~39 1602~ The frequency shifted light in ~he LPlf propagation mode and any li~ht remaining in the LPo1 propagation mode propagates from the frequency shifter 1606 to the mode selector 1640 through the two-m~de optioal fiber 1602.
Within the mode selector 1640, the light propagating in ~he LP11 propagation mode is coupled to the LPo1 propagation m~de o the ~ingle-mode op~ical fiber 1650 and propagates to the end portion 1656 where it exi~s as frequency shifted light, represented by the arFow 1662.
The ligh~ propagating in the ~PVl mode in the fiber 1602 at the original unshifted frequency propagates to ~he end portion 1654 of the two-mode optical fiber 1602 and exit~
as unshifted lignt, represented by an arrow 16S4.
As set forth above, the mode selector 1640 of the 1~ present invention is reciprocal so tha~ an optical signal can be input into the end portion 1656 of ~he single-mode optical fiber 1650 in the LPo1 propagation mode. The light in the LPol propagation mode of the ~ingle-mode optical fiber 1650 is coupled to the LP11 propagation mode of the two-mode optical fiber 1602 within the mode selec~or 1640~ Thereafter, ehhe light propagates through the optical f iber 1602 to the frequency shi~ter 1606 where i~ i5 ~hifted in fre~uency and is coupled to the LPo1 mode of the optical fiber 1602. The optical 6ignal propagates rom the frequency shif~er 1606 through the two-mode optical fiber 1602 to the mode stripper 1604 whPrein any optical signal remaining in the LP1 1 propagation mode is stripped from the optical fiber 1602. Thu~, the optical ~ignal propagating from the mode ~tripper 1604 ~o the single-mode op~ical fiber 1600 (which now operates as an output optical fiber) is propagated ~olely in the ~01 propagation mode and i~ shifted in frequency from the optical signal input at the end portion 1656 of the optical fiber 1650~
Figure 22 illugtrates a system which incorporates a mod ~elec~or built in accordance with the present 1~2C ~9 invention into a system that separates light propagating at a plurality of frequencies in one fiber into a plurality of light signals at discrete frequencies propagating in separate fibers. The system includes a first optical fiber 1800 which, in the embodiment shown, has only two propagation modes and has a highly elliptical core. This fiber 1800 includes a first end 1802 which receives input light, represented by an arrow 1804. The input light 1804 is comprised of a plurality of optical signals having discrete optical wavelengths ~ 2,---~n- Preferably, the optical signals are propagating in the LPol propagation mode of the fiber. A portion of the optical fiber 1800 is positioned in a coupler half 180~ constructed in accordance with the present invention, and it is preferably oriented in the coupler half 1806 so that the major axis of its elliptical core is norm~l to the facing surface of the coupler half 1806. A second optical fiber 1810, which is preferably a single-mode optical fiber, is positioned in a second coupler half 1812, also constructed in accordance with the present invention. The two coupler halves 1806, 1812 are positioned so that facing surfaces formed on fibers 1800 and 1810 are juxtaposed to provide coupling between the propagation modes of thé two fibers and thereby form a mode selector 1814 of the present invention. A
second portion of the first two-mode optical fiber 1800 is incorporated into an evanescent field grating reflector 1820, constructed in accordance with copending Canadian Application No. 513,681, entitled "Optical Fiber Evanescent Grating Reflector," filed on July 14, 1986, and assigned to the assignee of the instant application. This application is incorporated herein by reference. The two-mode optical ~iber 1800 is preferably oriented so that the major axis of its elliptical core is normal to the surface of the grating reflector 1820.

,~

8~

As disclosed in the copending application; by properly selecting the periodicity of ~he grating of the grating reflector 1820, light incident on the grating re~lec~o in ~he LPo1 propagation mode of the optical fiber 1800 as represented by an arrow 1830, is reflected by the grating reflector 1820 and is caused to propagate in the reverse di ection in the LP11 propagation mode, as illustr~ted by an arrow 1840. As set forth in the copending patent application, ~he grating reflector 1820 can be construc~ed to have a series of periodic gratings with a periodicity of ~, (wherein ~ is the dis~ance between parallel lines forming the gra~in~) 60 ehat light inrident upon the grating reflector 18Z0 ~n the LYo1 mode with a wavelength of ~01 (wherein ~01 refers to ~he wavelength at the input frequency in the LPo1 propagation mode3 will be reflected back into the fiber 1800 in ~he opposite direction at a wavelength ~11 which corresponds to the wavel~ngth of 8 ~ignal a~ ~he ~ame frequency in ~he LP11 propagation mode. In order to achieve ~his reflection characteristic, the periodicity ~ of the grating reflec~or 1820 is selected in accordance with the following equa$ion:

A ~ 143 As set for~h in the copending pa~ent application, the grating reflector 1820 acts as an inter-~ode couple- when 3~ the periodicity of the grating is ~elected in accordance with Equ~tian ~14). The grating reflec or 1820 is particularly advantageous in ~hat it is frequency selective and can ~leparate light propagating in a plur~lity of frequenciles oeuch that t~e light propagating i~ th2 rever~e direction, represented by the arrow 184~, will only have a selec~ed one of the plurality of g~ ~

frequencies~ That frequency Shaving ~ wavelength ~1~
representing ~he free space wavelength of an optical signal at the selected frequency) is the frequency wherein the LP~1 mode wavelength ~01 and the LP11 mode wavelength ~ atisfy Equation (14) for the periodicity ~ of the grating reflector 1820. Thus, although the input light signal incident upon the fi-st end 1802 of the fir~t multimode optical fiber 1800, as represented by the arrow 1804, has a plur~lity of optical frequencies (i.e., ~1~
. ~ . ~n~ only the optical signal having the wavelength ~1, having propagation mode wavelengths ~01 and ~11 aatisfyin~ the Equation ~14), will be reflected by the grating reflector 1820.
The ~ystem in Figure 22 opera~e~ as ~ollows. The light incidene to the firs~ end 1802 of the first optical fiber 1800 in the LPol mode propagates to the mode ~elector 1814. The mode selector 1~14 is constructed with the fibers 1800 and 1810 ~elected so that the phase propagation velocities are matched only for coupling from ~he LPll moae of the first two-mode op~ical fibPr 1B00 to the ~econd LP~l mode of the ~econd ~ingle-mode optical fiber 1810. Thus, optical ~ignals prop~ga~ing in the LPol mode vf ~he fir6t optical fiber 1810 pa~s through the interaction segion of the mode ~elector 1814 with little coupling ~o the second op~ical fiber 1810. The light inciden~ to the ~rating reflec~or 1820, repre~ented ~y the arrow 1830, thus comprise~ substantially all of the input light at the input frequencies ~ 2~ ~3~ n in ~he LPo1 modeO A~ the gr~ting reflector 1820 the inpu~ ligh~
3g at the frequency ~1, which satisfies ~he Equa~ion (14) for the waveleng~h ~01 for the LPol propagat$on mode and the wavelen~th ~11 for the LPl~ propagation mode is reflected by the grating reflec~or 1820 and propagates in the reYerse direction in the LP11 propagation mode as indicated by an arrow 1840. When this light is incident up~n ~he mode ~elector 1814, the optical signal in the LP11 propagation mode of the first optical f.iber 18~ is coupled to ~he L~1 propagation mode of the second optical fiber 1810 and is p~ovided as an output signal, represented by an arrow 1850, from an end 1852 of the second optical fiber 181~. The input optical signals at the other frequencies (represent~d by ~2~ ~3~ n a e not reflected by the gra~ing reflector 1~20 and continue to propagate in ~he LPol propagation mode of the first optical fiber 1800 in the original forward direction as indica~ed by ~n arrow 18600 Thus, the mode selector 1814 and the grating reflector 1820 act to~e~her to select the optical signal at the frequency .corresponding to the wavelength ~1, and p-ovide it ~s a ~iscrete ou~pu~ signal from the end 1852 of ~he second optical fiber 1810 in the LYo1 propagation mode. Additional pai~s of mode selectors and grating reflectors (not shown) constructed for the other wavelengths (i.e., ~ 3. . . . ~n) can be formed on the first optical fiber 1800 to select the other ~aveleng~hs and provide them as discrete output signals.
Thus, the embodiment of Figure 22 provides an advantageous means for ~eparatin~ optioal frequencie from an input light having 8 plur lity of input frequencies.
From the discu~sions rela~ing ~o the emb~diments set forth above, it will be understood that the mode selector of the present invention utilizes two dissimilar optical fibers, one of which is multimode, while ~he other may be either single-mode or multimode. The multimode fibers of the present invention, however, are a ~pecial class of multimode fibers, referred to herein as "few-mode fibersl"
which propagate light in no ~ore than ~bout five to ten spatial modes7 Those skilled in the art ~ill recognize that it is difficult to accomplish ehe propagation velocity matching andl mismatching dis ussed above in fiber~ that have more than about five ~o ten modes, since the dif~erence between the propagation velocities decreases as the nu~ber of modes increases. Further, tne propagation velocity difference between the 99th order mode and the 100th order mode of a hundred mode fiber is very small compar~d to the propagation velocity difference between the first and second order modes of a double mode fiber. Thus, it is particularly advantageous to use few mode fibers in the present invention.
Descri~tion of Interferometer Usina Highly Elliptical Core Wavequides The fundamental and second order guided modes of a highly ellipti~al core optical fiber provide two orthogonal paths through the fiber which permits the device to be used as a two-channel medium, e.g., as an in-line Mach-Zehnder interferometer, and as a two channel medium in data systems.
The principle of using the highly elliptical core optical fiber as an interferometer is illustrated in Figures 23 and 24a-24b. An exemplary section of an optical fiber 2100 having a highly elliptical core, as described above, is illustrated in Figure 23. An optical signal is input into the optical fiber 2100 with energy in the LPol modes and the LPll even modes, as illustrated by the arrows 2101 and 2104, respectively. Both modes propagate in the optical fiber 2100. The two modes have a beat length Lg, as discussed above. Five locations 2110, 2112, 2114, 2116 and 2118, that ara spaced apart by one-quarter beat length (i.e., Lg/4) are indicated in Figure 23.
In Figure 24a, a first field intensity pattern 2200 represents the distribution of optical energy in the core ~or thP LPol propagation mode, and a second field intensity pattern 2202 represents the distribution of optical energy in the core for the LPll propagation mode. In each cas~, the optical fiber is aligned so that the major axis of the elliptical core is vertically z~

-~5-aligned, as viewed in Figures 24a and 24b. When light is input into the highly elliptical core optical fiber 2100 with equal intensities in the LPol and LPll propagation modes, the liqht in the optical fiber 2124 at any particular location will have a field intensity pattern that represents the superposition of the LPol and LP11 modes, and the shape and intensity of the radiation patt~rn will be dependent upon the relative phase of the two modes at that location. The field intensity patterns for the optical fiber at three locations representing three different phases are illustrated in Figure 24b. A
first output field intensity pattern Z210, having the field intensity concentrated in the upper half of the optical fiber 2124 ~when oriented as shown) represents a phase difference (~ between the two modes of 2N~ (i.e., 0, 2~, 4~, etc.). Approximately half of the first output field intensity pattern 2210 is light and approximately half of the first output field intensity pattern 2210 is dark. (For convenience, the illustrations of the field intensity patt~rns in Figures 24a and 24b have the areas of maximum light intensity shaded and the areas of minimum light intensity unshaded). For purposes of illustration, it is assumed that the LPo1 modes and the LP11 even modes are in phase at the location 2110 (Figure 23). Thus, the two modes will be in phase again at the location 2120 that is separated by one beat length LB from the location 2110.
A second output field inten~ity pattern 2212, having the field intensity distributed between the upper and lower half of the optical fiber, represents a phase difference (~ between the two modes of (N ~ (i.e., ~/2, 3~/2, 5~/2, etcO). This intensity pattern will occur, for example, at the location 2112 (LB/4 from the location 2110) and the location 2116 (3LB/4 from the location 2110).
. .
,',..'~

~9Z~83 A third output field intensity pattern 2214, having the field intensity concentrated in the lower half of the optical fiber, represents a phase different (~) between the two modes of (2N ~ e., ~, 3~, 5~, etc.). This intensity pattern will occur at the location 2114 (LB/2 away from the location 2110).
Thus, by monitoring the output field intensity patterns in the upper and lower halves of the optical fiber 2100, the changes in the optical phase difference (i.e., the differential phase shift) between the two modes can be measurad. The differential phase shift between the two modes can result from perturbations to the optical fiber 2100, such as axial strain of the fiber, twists in the fiber, bending of the fiber, changes in the temperature of the fiber, lateral stress of the fiber, acoustic pressure on the fiber, and the like. Exemplary interferometers for measuring the differential phase shifts caused by external perturbations to the optical fiber are illustrated hereinafter.
Figure 25a illustrates an interferorl~eter having a highly elliptical core two mode optical fiber 2300. The optical fiber 2300 has a first end portion 2302 and a second end portion 2304. A mode stripper 2310 is formed in the optical fiber 2300 proximate to the first end portion 2302 by wrapping a number of turns of the fiber 2300 around a mandrel, or the like. An inter-modal coupler 2312, such as described above in connection with Figure 9 and Figure 9a, is formed on the optical fiber 2300 proximate to the mode stripper 2310 and between the mode stripper 2310 and the second end portion 2304. An intermediate portion 2314 of the optical fiber 2300 between the inter-modal coupler 2312 and the second end portion 2304 is exposed to an external perturbation such as an acoustic wave, a strain, or the like. The external pertur~ation is represented in general by a pair of jagged arrows 2320.

,~ ....

S~ 9 A first optical detector 2330 and a second optical detector 2332 are positioned proximate to khe second end portion 230~ of the optical fiber 2300. The first optical detector 2330 and the second optical detector 2332 can be conventional photodetectors, each of which provides an electrical output signal that is responsive to the intensity of the optical energy incident on it.
Preferably, the first detector 2330 is positioned to receive optical energy emitted from the upper half of the second end portion 2304 of the optical fiber 2300, and the second detector 2332 is positioned to receive optical energy emitted from the lower half of the second end portion 2304 of the optical fiber 2300. "Upper half" and "lower half," as used herein, refer to the upper half of the second end portion 2304 of the optical fiber 2300 when the second end portion 2304 is aligned so as to provide the intensity patterns in the orientations shown in Figures 24a and 24b (i.e., the major axis of the elliptical core is vertical). of course, thP second end portion 2304 can be aligned so that the major axis of the elliptical core is other than vertical so long as the two detectors 2330 and 2332 are aligned with the major axis.
The electrical output of the first detector 2330 is electrically connected to the negative input of a differential amplifier 2340, and the electrical output of the second detector 2332 is electrically connected to the positive input of the amplifier 2340. The differential amplifier 2340 compares the two inputs and provides an output on an output line 2342 that is proportional to the difference in intensity of the optical energy incident upon the first detector 2330 and the second detector 2332.
Although two detectors are shown, one skilled in the art will understand that a single one of the two detectors can be ussd to detect only the upper or the lower signal output to obtain the same information. The two detectors 2330 and 2332 and the differential amplifier 2340 are advantageously , , ~,,'-J;~

~ ~2~8~

used to increase the sensitivity of the embodiment of Figure 25a.
A light source 2350, which can be a laser light source, a broadband source (such as a superluminescent diode), or the like, prvvides a source output signal, represented by a line 2352, and is positioned to direct its output into the input end portion 2302 of the source optical fiber 2300. The light source 2350 is selected so that the wavelength of the source output signal 2352 is less than the cutoff wavelength for the LPl1 even mode of th~ optical fiber 2300, and greater than the cutoff wavelength for the LPl1 odd mode of the optical fiber 2300. Thus, only the LPo1 modes and the LPl1 even modes of th~ optical signal entering the first end portion 2303 of the optical ~iber 2300 will be supported by the optical ~iber 2300. Preferably, the light source 2350 is oriented so that substantially all of the sourc~ optical energy entering the first end portion 2302 of the optical fiber 2300 is in the LPol mode. However, any optical energy in the LPll even mode will be stripped by the mode stripper 2310. Thus, substantially all of the optical energy in the portion of the optical fiber 2300 between the mode stripper 2310 and the inter-modal coupler 2312 will be in the LPo1 propagation mode. The inter-modal coupler 2312 is preferably adjusted so that approximately 50% of the optical energy in the LPo1 propagation mode is coupled to the LP11 propagation mode. Thus, the optical energy entering the portion of the optical fiber 2303 between the inter-modal coupler 2312 and the second end portion 2304 will initially have approximately equal intensities in the LPol and LP11 propagation modes. The optical energy will propagate tv the second end portion 2304 and will be emitted there~rom onto the first detector 2330 and the second detector 2332. The optical intensity pattern of 2 ~ ~ 9 the opti~.al energy emitted by the ~econd end-portion 23U4 will depend upon the phase delay between thP two propagation modes caused by the differenre in phase propagation velocities between the LPo1 and the LP11 propagation modes. So long as the optical fibPr 2300 is not perturbed by an acou~tic signal, a temperature change, or the like, the phase delay ~ill remain constant, and the electrical output signal on the line 2342 from ~he differential amplifier 2340 will remain ~table. When a perturbation occurs, the leng~h of the optical path be~ween ~ne in~er-modal coupler 2312 and ehe ~econd end portion 2304 ~ill change, thus causing a change in th phase difference between the LP01 and the LP11 propagation ~odes. This phase difference causes a change in the elec~rical output signal on the line 2342 from the differential amplifier 2340.
The use of ~wo-mode optical fiber in the interferometer of Figure 25a is advantageous because of the unique characteri~tics ~f the group propaga~ion velocities and the phase propagation velocitie~ of the optical ~ignal in the fiberJ Although the phase propagation velorities are different for the wavelength of li~ht selected to propagate only ~he LPo1 mode and the LP11 e~en mode, the group propagation velocities of $he ~wo modes ~ill be ~ubstantially the ~ame near 8 particular optiral wavelength. Thus, the op~ical path length of the two transmi~ion paths provided by the two propAgation ~odes i8 substantially the same. For conventional in~erferometers h~ving two ~epara~e ~rms to provide the 3~ two opticnl paths, care must be taken ~o a~sure that the coherence length of the so~rce of optical ener~y is gseater .han the optLcal path difference in the two arms. This optical path difference in the two-arm interferometers is caused by differences in the group delay in ~he optical signals propagating in the two arms. In the present inven~ion, the gro~p delay for both -7~-modes is ehe game, and an optical signal ~ource having a sbor~ coherence length can be u~ed, even when the phase delays are large. Thu~, a longer oytical fiber can be used to increase the sensitivity of the interferometer without causing a ~ignificant increase in noise. The use of the highly elliptical core optical fiber 230~ prsvides a means for maintaining a stable orientation of the field intensiey patterns of the second order mode, which is very difficult to do when using circular core fiber.
-1~ Because the op~ical energy propagating in the two ~odes propaga~es at the same group velocity, an optical pulse containing energy propagating in the ~wo modes will not undergo modal di persion as it propagates. Thus, the optical pulse will not ~pread in time as it propagates down the fiber 2300. This allows the present invention eo be used with a pulsed lighe ~ource as well as with a continuous light source, since light propagating in both modes will reach the ~econd end portion 2304 of the optical fiber 2300 simultaneously and will thus interfere.
Figure 25b illustrates sn alternative embodi~ent of the interferometer of Figure 25a. In Figure 25b, the light source 235~, the source optical signal 2352, the first detector 2330, the ~econd detector 2332, and the differencial ~mplifier 2342 operate as d*scribed above for Figure 25a. In Figure 25b, the opti~al waveguide portion of the interferome~cer is formed from e-to c~ptical fibers, a fi~st optical fiber 2360 thae is preferably a ~ingle-mode optical fiber having, and a second optical fiber 2362 that i8 preferably a two-mode optical fiber having a ~ighly elliptical core. The firs~c op~ical fiber ~360 has a first end portion 2364 and a second end por~ion 236S. The first end portion 2364 iQ positioned proximate to the light source 2350 to receive the source optical signal 235?
generated by the light source 2350. The 6econd optical 3!; f iber 2362 has ~ fir~ end portion 2370 and a ~econd end poreion 2374. The 6ec:0nd end portion 2374 of the second ~8 optic~l fiber is positioned proximate to the first and ~econd detectors 2330 and 2332 th~t ~re vertically aligned as before. The major axis of the elliptical core of the 6econd optical fiber 2362 i~ preferably ali~ned with the vertical alignment of the first and ~econd deteceor~ 2330 and 2332.
The second end portion 2366 of the fir~t op~ical fiber 2360 And the first end portion 237~ of the second optical fiber 2362 are butt-~pliced to form an off~et 8plice l~ 2376. The offset ~plice 2376 is shown ln more detail in Figure 26, wherein the cores of the firRt optical fiber 2360 and the ~econd optical fiber 2362 are represented by outlines 238~ and 2382, respectively. The rore 2380 of the first optical fiber has a eente-line 2384 ~shown in phanto~) and the core 2382 of the second optical fiber has a center 2386 (~lso shown in phantom). As illustrated the centerlines of the two cores are offset with respect to each other so that the centerlines 2384, 2386 are not aligned. Preferably~ the ~wo ceneerlines are offset along the major axi3 of the ellip~ical core- of the second optical fiber 2362.
The effect of the offset centerlines of the two cores i~ to cau~e op~ical energy propagsting in the LP~1 mode of the first optical fiber 236Q to be coupled to the LPol and the LPl1 mode of the ~econd optical fiber. this is illus~ra~ed in Figure 26 by the 6uperposition o~ exemplary electric field amplitude distribution graphs on the core outlines 2380 ~nd 2382~ The fir~t distri~u~ion graph 2390 represent~ the ~mpli~ude distribution of ~he LPol propagation ~ode in the core 2380 of the first optical fiber 2360 proximate the offset splice 2376. As illustrated, the amplitude diseribueion in ~he first (single-mode) fiber 2360 is symmetrical about the centerline 23~2. When the optical energy in the first optical fiber 2360 cro~sefi the off et splice 2376 into the ~econd ~two-mode) ~ptical fiber 2362, it enters the ~econd optical fiber 2362 w~ th the optic~l energy asymmetrically distributed with respect to the centerline 23~6 of tne 6econd optical fiber 2362, as represented by an amplitude dis~ribution graph 2392. The amplitude distributaon graph 23~2 r presents the Qum of the optical ener~y in the LPo1 propagation mode, repre~ented by an amplitude distribution graph 2394, and the optical energy in the LP~1 propagation mode, represented by ~n amplitude distribution graph 2396. The amount of the off6et of the two center lines 2384 and 2386 is selec~ed 60 ~hat the optical energy in the 6econd optical fiber 2362 is substantially evenly distributed between the two propagation modes (i.e., approximately S0% of the optical energy is the LP
propagation mode and approximately 50% is in the LP11 propagation mode). Because of the alignment of the offset with the major ~xis of the elliptical core of the second optical fiber 2362, the optical energy will preferentially enter the LP11 even mode of the cecond optical fiber 2362. Furthermore, ~ny optical energy entering into the 20, LPll odd mode will not be ~uppor~ed by the secsnd optical fiber 2362 and thus will not propagate.
The apparatus of Figure 25b operates in substantially the 6ame manner as ~he apparatu of Figure 25a. The optical energy propagates in the single mode fiber 2360 in only the LPo1 mode. Thus, ie i3 not necessary to use a mode ~tripper to eliminate the LP11 mode. The offset ~plice 2376 i~ adjusted to couple 50% of the optical energy in~o eaoh of the LPol mode and the LP11 even mode. The detection in changes in the phase difference 39 be~ween the two mode~ caused by perturb~tions to the second optical fiber 2362 is accomplished in the 6~me ~anner as described above for Figure 25a.
E'igure 27~ illustrates an alternative embodiment of the interferometer of Fi~ure 25a that requires only one detec~or 2400. As in Figure ?5a, the interferometer of Figure 27~ includes the light source 2350 that produces 8~

the source optical input signal 2352. The optical input signal is incident upon the first end portion 2302 of the optical fiber 2300. Any optical energy in the LPll propagation mode entering the first end portion 2302 is stripped by the mode stripper 2310. Thereafter, the inter-modal coupler 2312 causes 50% of the optical energy to be coupled to the LP11 propagation mode so that the optical energy propagating in the intermediate fiber portion 2314 after the inter-modal coupler 2312 has substantially equal intensity in the two propagation modes. The intermediate fiber portion 2314 of the optical fiber 2300 is subjected to the perturbations 2320, as before. ~nlike the interferometer of Figure 25a, the interferometer of Figure 27a includes a second inter-modal coupler 2410 that is formed on the optical fiber 2300 between the intermediate fiber portion 2314 and the second end portion 2304. The interferometer of Figure 27a further includes a second mode stripper 2412 formed on the optical fiber 2300 between the second inter-modal coupler 2410 and the second end portion 2304. The second inter-modal coupler 2410 is preferably adjusted to provide 50%
coupling and operates to combine the optical energy from the two propagation modes. The optical intensity in the LPol propagation mode and in the LPll propagation mode after passing through the ~econd inter-modal coupler 2410 is determined by the differential phase shift of the two modes caused by the perturbations of the intermediate fiber portion 2314. Ths second mode stripper 2412 strips off the optical energy propagating in the LPll mode so that only the optical energy propagating in the LPol mode is ~mitted from the second end portion 2304 of the optical fiber 2300. This optical energy is detectsd by the detector 2400 and will have a sinusoidal dependency upon the differential phase shift between the two modes. Thus, the electrical output of the detector 2400 can be monitored and analyzed to determine the amount of phase J

shift snd thus the magnitude of the perturbations to the intermediate por~ion 2314 and the optical fiber 23U0.
Figure 27b illustrates the interferometer of Figu~e ~7a wherein a first single mode optical fiber 2420 and an off~et splice ~422 replace the firs~ mode stripper 231 and the first inter-modal coupler 2312 in Figure 27a, and a second -ingle-mode optical fiber 2424 and a second offset splice 2426 replace the second inter-modal coupler 2410 and ~he second mode ~tripper 2412. The fir8t ~ingle-mode fiber 2420 has ~ first end portion 2430 positioned to receive the optical signal 2352 generated by the light ~ource 2350 and has a seeond end portion 2432 fo~ming pa t of the first off~et splice 2~22. The ~econd single-mode fiber 2424 has a first end portion 2434 forming part of the second offset ~plice 2426 and has a ~econd end portion 2428 positioned proximate the detector 2400. A two-mode optical fiber 2440, having a highly elliptical core, has ~
first end portion 2442 forming a part of the firs~ offset splice 2422 and has a second end portion 2444 forming a part of che second offse~ splice 2426. The first single-mode fiber 24~0 and the fir~ offset splice 2422 operate in ~he manner described above for Figure 25b to cause the light entering th~ first end portion 2442 of the two-mode o,ptical fiber 2440 to have substantially equal optic~l in~censities in each of ~he LPol and the LPl 1 propagation modes. The light propagating in each of ~che two modes ~ill e~cp~rience a phase difference that deperlds upon the perturbations to the two-mode fiber 2440, represented by the arrow~ 2320 . At the second off se~ 8plice 2426, the optical energy in the ~wo-mode op~ical f~ber 2440 is coupled to the LPo1 propagation mode of the ~econd single-mode optical fiber 2424 and the optical energy in the second 6ingle-mode optic~l fiber 2424 will have an intensity that varies in accordance with va_iations in the pha~e difference of the two propagation modes in the two`
mode optica~ fibPr 2440 at the second offset splice ~ZQ89 24~6. 'lhe intensity of the optical energy in the second single-mode optical fiber 24~4 is detected by the detector 240U that provides an Rlectrical output ~ ign~l tnat i8 responsive to changes in the intensity snd thus to changes in the phase difference in the two propagation modes in the two-mode fiber 244~.
Figure 28a illustrates a further al~ernative interferometer utilizing the highly elliptical core two-msde optical fiber. In Figure 28a, the sensing portion of the interferometer comprises the light source 2350, the two-mode optical fiber 2300, the mode ~3tripper 2310, the first inter-modal coupler 2312, th~ intermediate portion 2314 of the optical fiber 2300, and~he second inter-~odal coupler 2410, interconnected ~s in- Figure 27a. ~oweve_ ra~her than including the second mode stripper 241~ of Figure 27a, the embodiment of Figure 28a includes a ~odal filter or mode ~elector 2450 ~3uch as was desc_ibed above. The modal filter 2450 is used to separate the optical energy in the two propaga~ion modes. The optical energy in the LPUl propagation ~ode iB provided as an output on a single-mode optical fiber 2452 and i~ directed to a first detector 2454. The optical energy in the LPl 1 propagation mode is proYided a~ an output on a ~wo-mode optical fiber 2456, that can advantageously be a continu tion of ~he optical fiber 23~)0. Tne optical energy output from the two mode optical fiber 2456 is directed ~o a second detec~or 2460. The first detector 2454 provides an electrical output ~ignal tha~ is provided to the negative input of a differential amplifier 2462, and the second de~ector 2460 provides an electrical OU~pUt that i8 provided to ehe po~itive input of ~he differential ~mplifier 2462. The differential ampl;fier provides an output on a line 2464 that represents the difference in intensity of the optical energy detected by the first and second detector6 2454, 246~, Rnd is thus responsive to hanges in the pha~3e difference between the two propagation modes in the intermediate port~on 2314 of the optical fiber 2300 caused by perturbat$ons 2320.
Figure 2~b i~ an alternative embodiment sf the interferometer of Figure 28a in wnich the mode strippe~
231U and the first inter-modal coupler 2312 ~re replaced wi~h the single-mode optical fiber 2420 and the offse~
splice 2422, as in Figure 27b. The interferométer o~
Figure 28b operate~ in a 8 imilar manner to the interferometer of Figure 28a~
~agure 28c is.a further altPr~ative embodiment of the interferometer of Figure 28a wherein an optical frequency ~hifter, such as ~he frequency shifter described above in conneccion with Figure 12, is 3ubstituted for the first inter-modal coupler 2312. As set forth ~bove, the ~requency shifter comprises the transducer 700 driven by a piezoelectric acoustic generaeor 712. The ~mall end of the transducer 70~ is secured to the optical ~iber 23~0 as xet forth in csnnection with Figure 12. The first ~coustic damper 536 is positioned on the optical fiber 2D 2300 on one side of the connection between the optical fiber 2300 and the transducer 700. The second acoustic damper 544 is positioned on ~he optical fiber 2300 at a location displaced away rom the connection between the ~ransducer 70U and the op~ical fiber 2300 ~o as to provide frequeDcy ~hifeer portion 2466 of ~he opt~cal fiber 2300 ~hat is po~itioned in the effec~ of the acoustic waves - produced by the transducer 7~0. The piezoelectric generator 712 of ~he transducer 700 i~ electricAlly driven by the output of the signal source 520, as befQre. The output of the signal source 520 is sl~o provided as one ~nput to a lock-in amplifier 2468. The lock-in amplifier 246B has a ~econd input that is connected to the output line 2464 of the differential amplifier 2462.
The frequency shifter in Figure 28c operates to cause the light coupled froDl the LPol propagation mode to the LP~l propagation mode in ~he frequency 6~ifter portion 2466 of the opticsl fiber 2300 to vary -in ti~e in ~ccordance with the frequency of the signal source 520.
Tnus, the optical intensities detected by the fi st detector 2454 and the xecond detecto 2460 will each have a component that varies in aceordance with the f equency of the signal ~ource 520 ~s well as ~ component that varies in accordsnce with ~he per~urbations, represented by ~he arrows 2320. The lock-in amplifier 2468 is synchronized with the ~ignal ~ource 52~ and ~hus provides an oueput signal that varies only in accordance ~ith the changes in optical intensity eaused by the perturbations 2320. The use of heterodyne detection ~uch as this ~ubstantially reduces or eliminates any signal fading that may occur as a result of environmentally induced phase drift.
Figure 29a illustra~es an alternative interferometer in which only one inter-modal coupler 2312 and one mode 6tripper is required. In Figure 29a, the source optical signal 2352 from the light 60urce 2350 i6 directed toward a beam splieter 2470. ~ portion of the~ souree optieal ~ignal 2352 passes through the beam ~plitter 2470 and is provided as an input ~o a fir~t end portion 2480 of a highly elliptical core optical fiber 2482. A portion of the optical fiber 248~ proximate to the first end portion 2480 is formed into a mode ~tripper 2484, as desc~ibed above. Another portion of the opti~al fiber 2482 is formed into an inter-modal coupler 2486 that is adjus~ed for 50% coupling. The msde splitter 2484 and the inter-modal coupler 2486 operate as described above such t~at op~ical energy thae has passed through the mode spl~tter 2484 ~nd the inter modal coupler 2486 has 3ubstantially equal inten~itie~ in each of the LPo1 and the LP11 even propag~tion modes. After passing through the inter-modal coupler 2486, the optioal energy propsgates in a sensing portion 249U of the t:wo-mode optical ~iber 2482. The ~ensing portion 2490 is positioned 90 that it i8 perturbed ~Z9;i~

by an external perturbation, such as an acoustic signal, temperature, or the like, represented by the arrows 2320, as before. The two-mode optical fiber 2482 has a second end portion 2492 that is terminated at a highly reflective surface 2494. The highly reflective surface 2494 can be a mirror, or the like, or it can advantageously he formed by pslishing the second end portion 2492 of the optical fiber 2482 so that substantially all of the optical energy reaching the second end portion 2492 is reflected back into the sensing portion 2490 of the optical fiber 2482 and propagates toward the first end portion 2480. The re~lected optical energy passes through the inter-modal coupler 2486 and the mode stripper 2484 and is emitted from the first end portion 2480. The optical energy emitt~d from the first end portion 2480 is directed by the beam splitter 2470 onto a detector 2496. The effect of the passage of the optical energy back through the inter-modal coupler 2486 and the mode stripper 2484 is substantially the same as the effect of the passage of the optical energy through the second inter-modal coupler 2410 and the second mode stripper 2412 in Figure 27a. Thus, the optical energy detected by the detector 2496 will have a sinusoidal dependency upon the differential phase shift between the modes~ The electrical output of the detector 2496 can be monitored and analyzed to determine the amount of phase shift and thus the magnitude of the perturbations to the sensing portion 2490 of the optical fiber 2482.
Figure 29b illustrates a further alternative embodiment of an interferometer similar to the interf~rometer of Figure 29aO In Figure 29b, a single-mode optical fiber 2500, having a first end portion 2502 and a second end portion 2504, is provided. Intermediate the fir~t end portion 2502 and the second end portion 2504, a portion of the first single-mode optical fiber 2500 is formed into a coupler half 2506 and is juxtaposed with a coupler half 2510 ~ormed on a second single-mode ~Z~2C~

optical fiber 2512. The coupler halves 2506 and 2510 are advantag~ously constructed in accordance with U.S. Pat~nt No. 4,536,058, as described above, and are adjusted to provide approximately 50% coupling between the first single-mode optical fibe~ 2500 and the second single-mode optical fiber 2512. The second end portion 2504 of th~
first single-mode optical fiber 2500 is juxtaposed with a first end portion 520 of a highly elliptical core two-mode optical fiber 2522 at an offset splice 2524, such as was described above. The two-mode optical fiher 2522 has a second end portion 2530 that i~ terminated at a highly reflective surface 2532 that may advantageously be formed by polishing the second end portion 2530. In operation, the source optical signal 2352 from the light source 2350 is input into the first end portion 2502 of the rirst single-mode optical fiber 2500. At the coupler halves 2506, 2510, approximately 50% of the optical energy in tne first single-mode optical fiber 2500 remains in the first single-mode optical fiber 2500 and is propagated to the second end portion 2504 at the offset splice 2524. At the offset splice 2524, the optical energy is coupled into the first end portion 2520 of the two-mode optical fiber 2522 with approximately equal intensities in each of the LP
and LPl1 even propagation modes of the two-mode optical fiber 2522. The optical energy propagates to the second end portion 2530 and is reflected back to the offset splice 2524 where the optical energy is coupled back to the first single-mode optical fiber 2500~ The intensity of the optical energy coupled back to the first single-mode sptical fiber 2500 will vary in accordance with the change~ in the phase difference between the LPol and the LP1l even propagation modes in the two-mode optical fiber 2522 caused by the ~xternal perturbations presented by the arrows 2320. Approximately 50% of the optical energy in the fir~t single-mode sptical fiber ~500 is coupled to the second single-mode optical fiber 2512 at the coupler "~,~,..~

~LZ~3Z~8~

halves 2506, 2510, and the coupled optical energy is emitted from the second single-mode optical fiber 2512 onto a detector 2540. The detector 2540 provides an electrical output signal that varies in accordance with the intensity of the optical energy incident upon it and thus varies in accordance with changes in the phase difference between the LPol and LPll even propagation modes in the two-mode optical fiber 2522 caused by the perturbations.
Description of a Strain Gauge Usin~ a Highly Elliptical Core Optical Flber Heretofore, only the differentiation in the propagation constants and thus the propagation velocities of the LPol mode and the even and odd LPl1 modes ha~e been consideredO However, within each of the spatial modes, there is also a difference between the propagation constants and velocities of the two polarization states within the modes that hecomes more apparent as the ellipticity or other asymmetry in the core of the optical waveguide increases. This is illustrated in Figure 30 which is an unscaled graph of the propagation constants versus ellipticity for the modes in a highly elliptical core optical waveguide such as was described in Figure 5 and in Figures 6a-~h. As previously explained, the propayation constant for the LPol mode is greater than the propagation constant for the LPl1 mode. Furthermore, within the LP1~ mode, the propagation constant for the LP11 even mode is greater than the propagation constant for the LP11 odd mode. The difference in the propagation constant for the LPll odd and even modes allows an optical siynal to be selected that has a wavelength that can propagate in the LPll even mode although it cannot propagate in the LPll odd mode. As further illustrated in the graph of Figure 30, the propagation constant of the vertically polarized LPol mode of the elliptical core waveguide is larger than the propagation constant of the ,~ ," i ~
" "

horizontally polarized LP~l mode. Similarly, the propagation constan~ of ~he vertically polarlzed LP11 even mode is larger than the propagation constant of the h~rizontally polarized LP11 even mode. Although exaggerated in Figure 30 to emphasize the differences in the propagation constants, it should be underatood ~hat the differences in the propagation cons~ants for the two polariYation modes within each spatial propagation mode is much smaller than the diff~rence in the propagation constants for the LPo1 and LP1l ~patial propaga~ion modes. The two polarization modes in the LPo1 propagation mode and the two polarization modes..in the LP11 even mode provide a total of four propBgation- paths in ~n optical waveguide for an optical signal having a selected wavelength between ehe LP11 even and the LPll odd cutoff wavelengths. Each of these fourth paths has a diffe ent propagation velocity and provides a dif~erent amount of optical phase delay fsr an optical signal propagating through the waveguide. The~e properties of the optical waveguide at the wavelength selected as set forth above provide the basis for the construction of an optical strain gauge that discriminates between ~ignal changes c~used by strain and signal changes caused by temperature changes.
An exemplary ~train g~uge constructed in ac~ordance with the present iDvention is illustra~ed in Fi~ure 31.
The strain gauge includes a light source 2600 (e.g., a laser light source, a brvadband light ~ource, ~uch as a super-luminescent diode, or the like) that produces a ~ource optic~l signal 2602. The ~ource optical signal 26~ i8 directed in~o the firs~ end portion 2610 sf a first ~ingle-mode optica~ fiber 2612. The first single-mode optical fiber 2612 has a second end p~r~ion 2614 that i~ butt-spliced to a first end por~ion 2620 of a two-mode optical fiber 2622 that has a highly ellip~ical core ~uch a~ has been pre~iou~ly discussed. Preferably, the first ~Z92~

single-mode optical fiber 2612 and the two-mode optical fiber are aligned 60 that Rubstantially all of the optical energy coupled to the two-mode optical fiber 2622 from the first single-mode optical fiber 2612 iB in the LP~1 propagation mode; however, a portion of the two-mode optical fiber 2622 proxim~te to the first end portion 2620 is formed in~o a mode stripper 2624 to remove subst~ntially all of ~ny optical energy that may be coupled to the LP11 propagstion mode.
1~ After fo~ming the mode 6tripper 2624, a portion of the two-mode optical fiber 2622 is formed into a frequency ~hifter by bonding the opt9 cal fiber 2622 ~o a transducer 2630 at a location ~632. The transducer 2630 is preferably cons~ructed in accordance with ~igure 12 above and includes a piezoelectric acoustic signal generator 2634 tha~ i~ electrically driven by the output of ~ signal 60urce ~64U via a line 2642. The acoustic signal generator 2634 generaees acoustic signals that are propagated by the ~ran~ducer 2630 to tne loca~ion 2632 on the optical fiber 2622. The acou tic waves induce traveling microbends into the optical fiber 2622 that ravel away from the location 2632. The traveling microbends are suppre~sed in the portion of the optical fiber 2622 between the location 2632 and the mode st ipper 2624 by a fir~t acoustic damper 2650. A second acoustic damper 26~2 i~ located on ehe optical fiber 2622 at a diatance away from the location 2632 in the opposi~e direction from the f~rst aroustic damper 2650. A
frequency shif~er portion 2654 of the optical fiber 2622 i$ defined be~ween ~he location 2632 and the second acou~tic damper 2652 ~hat provides a propagation medium or the traveling microbends induced in the optical fiber 26~2 by thP tran~ducer 2630. The optic~l energy traveling in the LPo1 propaga~ion mo~e in the optical fiber 2622 upon en~ry to the frequency ~hifter portion 2654 is coupl2d to the LPtl propagation mode And is shifted in il59 frequency by an amount determined by the frequency of the electrical signal generated by the xignal ~ource 2640.
After passing through ~he frequency Rhifter portion 2~54, the frequency Rhifted optical energy in the LPll propagation mode then enter~ a strain gauge po~tion 2660 of the optical fiber 2622 between the second acsustic damper 2652 and a second end portion 2662 of the tws-mode optical fiber 2622. The strain ~auge portion 2660 is ~ubjec~ed to a ~train, represented s S in Figure 33, and may also be subjected to an additional environ~ental perturbatisn, such as te~perature changes, repre~ented by the arrows labeled as T.
The second end psrtion 2662 of the two-mode optical fiber 2622 fo ms part of an offse~ splice 2664 along with lS a first end psrtion 2670 of a secsnd single-msd2 sptical fiber 2672~ The ~econd ~ingle-mode optical fiber 2672 has a ~ecsnd end portion 2674. A first coupler half 2676 is formed on ~he secsnd single-mode optical fiber 267~ at a location between the first end portion 2670 and the ~econd end portion 2674. The first coupler half 2676 is ju~taposed with a ~econd coupler half ~6~0 fo~med on a third ~ingle-mode optical fiber 2682 between a fir~t end por~ion 2684 and a ~ecsnd end portion 2686. The first coupler half 2676 and ~he ~econd ~oupler half 2680 are preferably ~djusted 80 that approximately 50% of the optical energy propagating in the 6econd single-mod2 optical fiber 2672 from the firs~ end portion 2670 is coupled ~o the third sin~le-mode op~ical fi~er ~6~2 and approximately 50% of the optical energy remains in the ~econd ~ingle-mode optical fiber 2672.
The optical energy propagating in each of the second ~in~le-mode optical fiber 2672 and the ~hird ~ingle-mode op~i~al fibers 2682 after passing through the coupler halves 2676, ~680 compsise6 optical energy in both the vertically polarized LPo1 mode and the horizontally p~larized LPot ~ode. A first polsrization filter (HPF) ~ 2 ~ 3 -~4-2690 is posi~ioned on the ~econd ingle-mode optical fiber 2672 between the first coupler half 2676 ~nd thè ~econd end portion 2674. The first polarization filter 2690 is oriented to pass only optical energy in the ho~izontally pvlarized LPol mode. Similarly, ~ secDnd polarization filte~ (VPF) 2692 is positioned on the third single-mode optical fiber 2682 between the second ooupler half 2680 and the second end portion 2686, ~nd is oriented ~o pass only optical energy in the vertically polarized LPU1 1~ mode. The coupler halves 2676, 2680 and the polarization fil~ers 2690, 2692 an be replaced with other known polarization beam splitters.
The second end portion 2674 of the second single-mode optical fiber 2672 is positioned proximate to ~ first detector 2700, and the ~econd end portion 2686 of the third single-mode optical fiber 2682 ic positioned pro~imate to a ~econd detector 2702. The first detector 2700 provides an elec~rical oueput signal on a line 2704 that is responsive ~o the intensity of the optical energy 2~ in the horizontally polarized LPol propagation mode in the second single-mode optical fiber 2672. The second detector 2702 provides an electrical output ~ignal on a line 2706 that i~ responsive ~o the intensity of the optical energy in the vertically pola i~ed LPol propagation mode in the third 3ingle-mode optical fiber 2682. ~oth of the electrical output signals will in~lude components cau~ed by the effects of the frequency shifter, the effects of any strain ~pplied to t~e ~train gauge portion 2660, and the effect~ of any changes in the te~per~ure. The electr~cal oueput of the first detector 2700 on the line 2704 is prsvided as an input to a first lock-in amplifier 2710, and the electrical output of the second detector 2702 on the line 2706 is provided as an input ~o a second lock-in amplifier 2712. The first loc~-ln amplifier 271~ ~as an input fro~ the ~ignal source 2640 and provides fln output signal ~H on B line 2720 that ~92~g -~5-represents the changes in optical phas~ in the horizontally polarized L~1 mode caused by the effects of strain and the effects of tempera~ure changes. Similarly, the ~econd lock-in amplifier 2712 has an inpue from the signal ~ource 2640 and provides an output signal ~V on a line 2722 that represents the changes in the optical phase in the vertically polarized LPo1 mode caused by the effects of strain and the effects of ~emperature ch~nges.
Both the strain S and the temperature T will cause changes in the length of the optical paths through the strain g uge portion 2660 and will thus cause chan~es in the phase delays through the strain gauge portion 2660.
Thus, it would be expected that one would not be able to differentiate the effects of changes in the temperature T
from che effects of the ~train. ~owever, it has been found that the changes in the phase delay caused by temperature changes affect both the vertically polarized optical ene_gy and the horizontally polarized opeical energy substantially equally over a rela~ively wide range of temperatures. In contrast, the changes in phase delay caused by strain (i.e., by changing the ~verall length of the ~train gauge portion 2660 of the optical fiber 2614) have a grea~er effec~ on one polarization ~han on ~he other polarization. This can be bet~er underseood by the following equations:

I~H = AbL ~ B~T ~15) ~y - C~L ~ D~T (16) where ~L i~ the chan~e in the length csused by strain ~pplied to the ~train gauge portion 2660 of ehe two-mode ~2gZ~89 optical fiber 2614; ~T is the change in the temperature spplied to the ~train gauge portion 2660 of ~he two-mode optical fiber 2614; and the constants A, B, C, ~nd D are experimentally determined as part o~ ~ calibration procedure for the ~train gauge of Figure 31, The output ~ignals ~ and ~y on the lines 2720 and 2722, respectively, are provided as inputs to a processo_ 2730 that calculates ~L and ~T using Equations (15) and (16), above, and provides an ou~put 8 ignal representing ~L
on a line 2732 and an output repre~2nting ~T on a line 2734. The processor 2730 can advantsgeously be a digital - proces~or havin~ one or more ~nalog-to-digital convertor~
to convert the analog input ~ignals on ~he lines 2720 and 2722 to digital repre6enta~ionsO The processor 2730 can 1S apply one of a number of known algorithm~ for 601ving the two equations (15 and 16) for two unknown values (~ and ~T), provided that AD-BC i8 non-zero. AD-BC should be non-zero for the pre6ent invention because the temperature affec~ H and ~V differently than he ~train affects a~H and ~V-As set for~h above, although ~he ~train gauge portion 2660 is sensitive to temperature, the ~ensitivity is relatively small compared to its sensitivity to strain.
Thu~, ~he ~T outpu~ on ~he l~ne 2734 doe not provide Yery sensitive temperature indicationO The ~train gauge of ~igure 31 can also be used as a 6ensitive tempera~ure ~ensor by wrapping the strain gauge portion 2660 of the optical fiber 2614 around a mandrel 2740 a8 illustrated in Figure 32. The ~andrel 2740 is constructed from a material tha~ e~pand~ and contract~ in accordance with its eemper~ture~ The expansion ~nd contraction of the mandrel 2740 caus es change~ in ~he length of ~che strain gauge portion 2660 of the optical fiber 2614 th~t are detectable as Get forth above., The proce sor 2730 can be calibrated 3~; 80 that the ~L output is respon~ive to the changes in temperature of the mandrel 2740.

-~7-While preferred embodiments of this in~ention have been disclosed herein, those skilled in the art will appreciate that changes and modifications may be made tAerein without departing from the spiri~ and the scope of this invention, as defined in the appended claims, 2~

3~

Claims (63)

1. An apparatus, comprising:
an optical waveguide having a core with a non circular cross section; and a source of light for introducing light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of said waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode;
said waveguide being sized to provide a second predetermined cutoff wavelength for said signals, less than the first predetermined cutoff wavelength, the noncircular cross section of the core having cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable, spatial intensity pattern, substantially all of the light signals introduced into the waveguide by said source of light being at one or more wavelengths greater that the second predetermined cutoff wavelength to cause said light signals to propagate in only one spatial intensity pattern for the higher order mode.

2. The apparatus as defined in Claim 1, wherein the fundamental spatial mode includes two polarization modes, the cross-sectional dimensions of the core further selected to cause the polarization modes of the fundamental mode to be non-degenerate.

3. The apparatus as defined in Claim 2, wherein the single intensity pattern of the higher order spatial mode includes two polarization modes, the cross-sectional dimensions of the core further selected to cause these polarization modes to be non-degenerate.

4. An apparatus, comprising:
an optical waveguide having a core with a noncircular cross section; and a source of light for introducing light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of said waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode;
wherein said waveguide is sized to provide a second predetermined cutoff wavelength for said signals, less than the first predetermined cutoff wavelength, the noncircular cross section of the core having cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable, spatial intensity pattern, substantially all of the light signals introduced into the waveguide by said source of light being at one or more wavelengths greater than the second predetermined cutoff wavelength to cause said light signals to propagate in only one spatial intensity pattern for the higher order mode; and wherein the waveguide comprises an optical fiber, the fundamental mode being the LP01 mode of the optical fiber and the higher order mode being the LP11 mode of the optical fiber, the single intensity pattern being the even mode intensity pattern of the LP11 mode.

5. The apparatus as defined in Claim 4, wherein the core of the optical fiber has an elliptical cross section.

6. An apparatus, comprising:
an optical waveguide having a core with a noncircular cross section; and a source of light for introducing light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of said waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode;
wherein said waveguide is sized to provide a second predetermined cutoff wavelength for said signals, less than the first predetermined cutoff wavelength, the noncircular cross section of the core having cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable, spatial intensity pattern, substantially all of the light signals introduced into the waveguide by said source of light being at one or more wavelengths greater than the second predetermined cutoff wavelength to cause said light signals to propagate in only one spatial intensity pattern for the higher order mode;
wherein the fundamental spatial mode includes two polarization modes, the cross-sectional dimensions of the core further selected to cause the polarization modes of the fundamental mode to be non-degenerate;
wherein the single intensity pattern of the higher order spatial mode includes two polarization modes, the cross-sectional dimensions of the core further selected to cause these polarization modes to be non-degenerate; and wherein the non-degeneracy between the polarization modes of the fundamental mode and higher order mode produces a beat length between polarization modes on the order of 10 cm or less, for both sets of polarization modes.

7. The apparatus as defined in Claim 4, additionally comprising means for inducing a periodic stress in the optical fiber at intervals related to the beat length between the fundamental mode and the higher order mode such that light is cumulatively coupled between the fundamental and higher order modes at said intervals.

8. The apparatus defined by Claim 7, wherein the stress inducing means induces microbends into the fiber.

9. The apparatus defined by Claim 4, additionally comprising a periodic structure for applying force to the fiber to induce static microbends in the fiber at intervals defined by the beat length between the fundamental mode and the higher order mode such that light is cumulatively coupled between the modes at said intervals.

10. An apparatus, comprising:
an optical waveguide having a core with a noncircular cross section; and a source of light for introducing light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of said waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode;
wherein said waveguide is sized to provide a second predetermined cutoff wavelength for said signals, less than the first predetermined cutoff wavelength, the noncircular cross section of the core having cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable, spatial intensity pattern, substantially all of the light signals introduced into the waveguide by said source of light being at one or more wavelengths greater than the second predetermined cutoff wavelength to cause said light signals to propagate in only one spatial intensity pattern for the higher order mode; and wherein the waveguide comprises an optical fiber, said apparatus additionally comprising:
a generator for producing a traveling flexural wave which propagates in said optical fiber, the energy of said traveling flexural wave confined to said optical fiber and having a wavelength in the direction of propagation selected in accordance with a beat length for two modes of the fiber to cause light to be cumulatively coupled from one of the modes to the other of the modes and shifted in frequency.

11. The apparatus as defined in Claim 10, wherein said flexural wave forms a series of microbends which propagate in the fiber.

12. An apparatus, comprising:
an optical waveguide having a core with a noncircular cross section; and a source of light for introducing light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of said waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode;
wherein said waveguide is sized to provide a second predetermined cutoff wavelength for said signals, less than the first predetermined cutoff wavelength, the noncircular cross section of the core having cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable, spatial intensity pattern, substantially all of the light signals introduced into the waveguide by said source of light being at one or more wavelengths greater than the second predetermined cutoff wavelength to cause said light signals to propagate in only one spatial intensity pattern for the higher order mode; and wherein said waveguide comprises a first optical fiber, said apparatus additionally comprising a second optical fiber which is dissimilar to the first fiber, said second fiber having at least one spatial propagation mode, only two of the modes of the fibers having matching propagation velocities, one of the matching modes being in the first fiber and the other being in the second fiber, said fibers juxtaposed to form an interaction region in which light is transferred between their cores, the proximity of the fiber cores at the interaction region selected such that light propagating in one of the matching modes in one of the fibers is coupled to the other of the fibers, the remainder of the modes all having mismatched propagation velocities such that the propagation velocity of each of the mismatched modes differs sufficiently from all of the other modes to prevent substantial optical coupling between any of the mismatched modes.

13. The apparatus defined by Claim 12, wherein the cores of each of the fibers have a cross section which is substantially the same inside the interaction region as outside the interaction region.

14. The apparatus defined by Claim 13, wherein the second fiber is a single mode fiber.

15. The apparatus defined by Claim 13, wherein the length of the interaction region is at least an order of magnitude larger than the maximum cross sectional core dimension of either of the fibers.

16. An apparatus, comprising:
an optical waveguide having a core with a noncircular cross section; and a source of light for introducing light signals having at least one wavelength into the waveguide for propagation therein, such that a substantial portion of the light is at one or more wavelengths less than a first predetermined cutoff wavelength of said waveguide to cause the waveguide to guide light in both a fundamental spatial propagation mode and a higher order spatial propagation mode;
wherein said waveguide is sized to provide a second predetermined cutoff wavelength for said signals, less than the first predetermined cutoff wavelength, the noncircular cross section of the core having cross-sectional dimensions selected such that light guided by the waveguide in the higher order mode at wavelengths greater than the second predetermined cutoff wavelength propagates in only a single, stable, spatial intensity pattern, substantially all of the light signals introduced into the waveguide by said source of light being at one or more wavelengths greater than the second predetermined cutoff wavelength to cause said light signals to propagate in only one spatial intensity pattern for the higher order mode; and wherein said waveguide comprises a first optical fiber, said apparatus additionally comprising a second optical fiber which is dissimilar to said first optical fiber, said fibers juxtaposed to form an interaction region for coupling light exclusively between a selected spatial mode in the first fiber and a selected spatial mode in the second fiber, one of the selected modes being of a higher order than the other of the selected modes, the length of the interaction region being at least an order of magnitude greater than the maximum cross-sectional dimension of the core of either of the fibers, and each of the fibers having a cross-sectional area which is substantially the same within the interaction region as outside the interaction region.

17. The apparatus as defined in Claim 12, additionally comprising a grating reflector disposed at a location on the first fiber, the grating reflector having a spatial periodicity selected to couple light having a predetermined wavelength between one of the mismatched modes and one of the matched modes, both of these modes being in the first fiber.

18. The apparatus as defined in Claim 1, additionally comprising:
means for introducing light into said waveguide such that said light propagates in two spatial modes of said waveguide, and such that said light propagates through a sensing section of the waveguide for exposure to an ambient effect; and means for detecting light output from said sensing section, said detecting means including a photodetector arranged to intercept only a selected portion of the spatial intensity pattern defined by a superposition of the spatial intensity patterns of the two modes, said selected portion of said spatial intensity pattern including substantial portions of light from both of said two spatial modes such that said intensity pattern of said selected portion varies in response to said ambient effect.

19. The apparatus as defined in Claim 18, wherein said detecting means further includes another photodetector arranged to intercept another portion of the spatial intensity pattern, and a comparing apparatus for comparing the output of the two photodetectors.

20. The apparatus as defined in Claim 18, wherein said waveguide comprises a multimode optical fiber.

21. The apparatus as defined in Claim 20, wherein said introducing means comprises a single mode optical fiber, butt coupled to introduce light into said multimode fiber, said single mode fiber having a central axis which is offset from the central axis of said multimode fiber to cause the light entering the multimode fiber to be split between said two modes of said multimode fiber.

22. The apparatus as defined in Claim 21, additionally comprising a coupler, positioned on said single mode fiber, and a reflector positioned at the output of the sensing section for reflecting said light after propagation through said sensing section such that it passes back through said sensing section, and through said coupler, said photodetector positioned to receive light from said coupler.

23. The apparatus as defined in Claim 20, wherein said introducing means comprises a modal coupler at the input to said sensing section of said multimode fiber to split said light between said two modes of said multimode fiber.

24. The apparatus as defined in Claim 23, additionally comprising a reflector for reflecting said light after propagation through said sensing section such that it passes back through said sensing section and back through said modal coupler.

25. The apparatus as defined in Claim 24, wherein said detecting means additionally comprises a mode stripper for stripping one of said modes from said fiber prior to reaching said photodetector.

26. The apparatus as defined in Claim 20, wherein said detecting means further includes a mode selector for selectively coupling one of said two modes from said multimode fiber, after propagation of said light through said sensing section, said photodetector positioned to receive light from the single mode fiber.

27. The apparatus as defined in Claim 25, wherein said detecting means further includes a modal coupler at the output of said sensing section for coupling light between said two modes prior to reaching said mode selector.

28. The apparatus as defined in Claim 25, wherein said detecting means further includes a single mode fiber butt coupled to receive light from the multimode fiber, said single mode fiber having 2 central axis which is offset from the central axis of the multimode fiber to cause the light entering the single mode fiber to include substantial portions of light from both of said two modes.

29. The apparatus as defined in Claim 18, wherein said introducing means comprises an optical frequency shifter which couples light input into one of the two modes to the other of the two modes and shifts the coupled light in frequency.

30. The apparatus as defined in Claim 29, wherein said frequency shifter comprises an optical fiber and a generator for introducing a flexural wave in said fiber.

31. The apparatus as defined in Claim 28, wherein said detecting means further includes a signal source for driving said frequency shifter and an amplifier for synchronously detecting the photodetector output in accordance with the frequency of the signal source.

32. The apparatus as defined in Claim 19, wherein said introducing means comprises an optical frequency shifter and a signal generator for driving the frequency shifter, and wherein said detecting means further comprises a comparing apparatus for comparing the output of the two photodetectors.

33. The apparatus as defined in Claim 32, wherein said detecting means further comprises an amplifier which detects the output of the comparing apparatus in accordance with the frequency of the signal source.

34. The apparatus as defined in Claim 20, wherein said ambient effect comprises a strain on said optical fiber.

35. The apparatus as defined in Claim 18, additionally comprising:
a splitting device which splits said selected portion of said spatial intensity pattern into two light beams;
a first polarizer for polarizing one of said beams to produce a first light signal, said photodetector positioned to receive said first light signal;
a second polarizer for polarizing the other of said beams to produce a second light signal;
a second photodetector for receiving said second light signal, said polarizers oriented such that said first and second signals have orthogonal polarizations; and a comparing device for comparing the outputs of the photodetectors.

36. The apparatus as defined in Claim 35, wherein said two spatial modes comprise said fundamental mode and said higher order mode, said single intensity pattern of the higher order mode including two polarization modes, said fundamental mode also including two polarization modes, the cross-sectional dimensions of the core further selected to cause both sets of polarization modes to be non-degenerate.

37. The apparatus as defined in Claim 36, wherein said introducing means comprises a frequency shifter for coupling frequency shifted light from one of the two spatial modes to the other, a generator for driving the frequency shifter at a frequency, and a device for synchronously detecting the output of the comparing device in accordance with the frequency of the generator.

38. The apparatus as defined in Claim 37, wherein said waveguide comprises an optical fiber and said ambient effect comprises a strain on said optical fiber.

39. The apparatus as defined in Claim 38, wherein said optical fiber is wrapped on an expandable member to cause said strain.

40. The apparatus as defined in Claim 39, wherein said member expands in response to temperature.

41. A method of propagating light from a source of light through an optical waveguide having a core with a non-circular cross-section, comprising the steps of:
selecting the wavelength of the light and the cross-sectional dimensions of the non-circular core such that (1) the waveguide propagates light in a fundamental spatial propagation mode and a higher order spatial propagation mode, and (2) substantially all of the light in the higher order mode propagates in only a single, stable intensity pattern.

42. The method of Claim 41, wherein the fundamental spatial mode includes two polarization modes, the method additionally comprising further selecting the cross-sectional dimensions of the core to cause the polarization modes of the fundamental mode to be nondegenerate.

43. The method of Claim 42, wherein the single intensity pattern of the higher order spatial mode includes two polarization modes, the method additionally comprising further selecting the cross-sectional dimensions of the core to cause these polarization modes to be nondegenerate.

44. The method of Claim 43, wherein the nondegeneracy between the polarization modes of the fundamental mode and higher order mode produce a beat length between polarization modes on the order of 10 cm or less, for both sets of polarization modes.

45. The method of Claim 42, additionally comprising the step of inducing a periodic stress in the waveguide at intervals related to the beat length between the fundamental mode and the higher order mode such that light is cumulatively coupled between the fundamental and higher order modes at the intervals.

46. The method of Claim 42, wherein the waveguide comprises an optical fiber, the method additionally comprising the step of propagating a traveling flexural wave in the optical fiber and utilizing the flexural wave to couple light from one of the modes to another of the modes.

47. The method of Claim 46, wherein the flexural wave forms a series of microbends which propagate in the fiber.

48. A method of propagating light from a source of light through an optical waveguide having a core with a non-circular cross section, comprising the steps of:
selecting the wavelength of the light and the cross-sectional dimensions of the non-circular core such that (1) the waveguide propagates light in a fundamental spatial propagation mode and a higher order spatial propagation mode, and (2) substantially all of the light in the higher order mode propagates in only a single, stable intensity pattern, wherein the fundamental spatial mode includes two polarization modes, the method additionally comprising further selecting the cross-sectional dimensions of the core to cause the polarization modes of the fundamental mode to be nondegenerate;
wherein the waveguide comprises a first optical fiber, said method additionally comprising the step of coupling the light from a selected one of the fundamental and higher order modes of the first fiber to a selected mode of a second fiber, without coupling light from the other of the fundamental and higher order modes to the selected mode of the second fiber.

49. A method of propagating light from a source of light through an optical waveguide having a core with a non-circular cross section, comprising the steps of:
selecting the wavelength of the light and the cross-sectional dimensions of the non-circular core such that (1) the waveguide propagates light in a fundamental spatial propagation mode and a higher order spatial propagation mode, and (2) substantially all of the light in the higher order mode propagates in only a single, stable intensity pattern, wherein the fundamental spatial mode includes two polarization modes, the method additionally comprising further selecting the cross-sectional dimensions of the core to cause the polarization modes of the fundamental mode to be nondegenerate;
said method additionally comprising the step of introducing light into the waveguide such that the light propagates in two spatial modes of the waveguide, and such that the light propagates through a sensing section of the waveguide for exposure to an ambient effect, and detecting only a selected portion of the spatial intensity pattern defined by a superposition of the spatial intensity patterns of the two modes, said selected portion of the spatial intensity pattern including substantial portions of light from both of the two spatial modes.

50. The method of Claim 49, additionally comprising the step of detecting another portion of the spatial intensity pattern and comparing the detected portions of the spatial intensity patterns to measure the ambient effect.

51. The method of Claim 49, additionally comprising the step of coupling light input into one of the two modes to the other of the two modes and shifting the coupled light in frequency.

52. The method of Claim 49, wherein the waveguide comprises an optical fiber, the method additionally comprising the step of stretching the optical fiber to produce the ambient effect.

53. The method of Claim 49, wherein the two spatial modes comprise the fundamental mode and the higher order mode, the single intensity pattern of the higher order mode including two polarization modes, the fundamental mode also including two polarization modes, the method additionally comprising the step of further selecting the cross-sectional dimensions of the core to cause both sets of polarization modes to be non degenerate.

54. The method of Claim 53, additionally comprising the steps of splitting the selected portion of the spatial intensity pattern into two light beams, passing one of the beams through a polarizer to produce a first light signal, passing the other of the beams through a second polarizer to produce a second light signal, orienting the polarizers such that the first and second signals have orthogonal polarizations, and comparing the first and second signals to measure the ambient effect.

55. The method of Claim 54, additionally comprising the step of coupling light between the fundamental mode and the higher order mode and frequency shifting the coupled light.

56. An apparatus comprising:
an optical waveguide having a non-circular core, said core being sized to have dimensions that are selected so that said waveguide propagates light in a predetermined wavelength range such that the light having a wavelength within said predetermined wavelength range propagates in a fundamental spatial propagation mode and in a second order spatial propagation mode, each of said fundamental spatial propagation mode and said second order spatial propagation mode supporting the propagation of light in first and second orthogonal polarization modes;
a light source that introduces light into said optical waveguide at a wavelength within said predetermined wavelength range to propagate in said waveguide in both of said fundamental and second order spatial propagation modes and in both of said two polarization modes within each of said spatial propagation modes:
means optically coupled to said waveguide for separating said light after passing through said optical waveguide into a first portion having said first polarization and a second portion having said second polarization, each of said first and second portions including light in said first and second spatial propagation modes; and a detection system that detects the intensity of the light in said first and second portions, the intensity of the light in said first and second portions varying in response to a perturbation of said waveguide, said detection system providing a first output signal responsive to the intensity of the light in said first portion and a second output signal responsive to the intensity of the light in said second portion.

57. The apparatus as defined in Claim 56, wherein the perturbation is a strain applied to said waveguide that causes the length of said waveguide to change.

58. The apparatus as defined in Claim 57, wherein the intensity of the light in said first and second portions varies in response to changes in the temperature of said waveguide, the variations in intensity caused by changes in the temperature differing from the variations in intensity caused by strain, said apparatus further including means for processing said first and second output signals to determine the magnitude of the strain and the magnitude of the temperature change.

59. The apparatus as defined in Claim 58, wherein said processing means determines the magnitude of the strain and the magnitude of the temperature change in accordance with the following equations:
.DELTA.?H = A.DELTA.L + B.DELTA.T
.DELTA.?V = C.DELTA.L + D.DELTA.T
wherein .DELTA.?H is the change in phase between the fundamental spatial propagation mode and the second order spatial propagation mode for the light propagating in one of said first and second polarization modes as measured by the changes in intensity in a respective one of said first and second portions: .DELTA.?L is the change in phase between the fundamental spatial propagation mode and the second order spatial propagation mode for the light propagating in the other of said first and second polarization modes as measured by the changes in intensity in the respective other of said first and second portions; A.DELTA.L is the change in length of the waveguide responsive to the applied strain; BAT is the change in the temperature of the waveguide; and A, B, C and D are constants determined when said apparatus is calibrated with known applied strains and known temperatures.

60. The apparatus as defined in Claim 56, wherein said non-circular core has a geometry that is selected so that the light having said wavelength within said predetermined range of wavelengths propagates only in a single, stable intensity pattern in said second order propagation mode, said core geometry suppressing propagation of light in a second intensity pattern in said second order propagation mode.

61. The apparatus as defined in Claim 60, wherein said waveguide is a two-mode optical fiber.

62. An apparatus comprising:
a light source coupled to a single core optical waveguide such that light from said light source propagates in said optical waveguide in first, second, third and fourth propagation paths as first, second, third and fourth light signals; and a detection system that (1) detects the phase difference between said first and second light signals propagating in said first and second propagation paths and (2) detects the phase difference between said third and fourth light signals propagating in said third and fourth propagation paths, said detection system utilizing (1) said detected phase difference between said first and second light signals and (2) said detected phase difference between said third and fourth light signals to provide an output signal responsive to perturbation of at least one of said first, second, third and fourth propagation paths of said optical waveguide.

63. The apparatus as defined in Claim 62, wherein said optical waveguide is a multimode optical fiber, said first and second propagation paths are the fundamental and second order spatial propagation modes of light having a first polarization, and said third and fourth propagation paths are the fundamental and second order spatial propagation modes of light having a second polarization orthogonal to said first polarization.