Free-space optical systems for wavelength switching and spectral monitoring applications

Description of equivalent WO 03025630 A2

FREE-SPACE OPTICAL SYSTEMS FOR WAVELENGTH SWITCHING AND SPECTRAL MONITORING APPLICATIONS

INVENTORS Pavel G. Polynkin, Jeffrey P. Wilde, Michael J. Timmons, and Mark H. Garrett

CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority U.S. Patent Application No. 09/961,565, filed September 20, 2001; U.S. Patent Application No. 09/992,778, filed on November 13, 2001, which is a continuation-in-part of U.S. Patent Application No. 09/961,565; and U.S. Patent Application No. 10/022,303 filed on December 14, 2001, which is a continuation-in-part of U.S. Patent Application No. 09/992,778, all of which are fully and completely incorporated herein by reference.

FIELD OF THE INVENTION This invention relates generally to optical systems. More specifically, it relates to free-space optical systems for wavelength switching and spectral power monitoring applications, having alignment compensation and polarization diversity schemes. The present invention can be used to construct a variety of optical devices, such as spectral power monitors, multiplexers and demultiplexers, and optical add-drop multiplexers, which are well suited for WDM optical networking applications, and which may be actively aligned by use of hardware and/or software control, and which may further employ polarization diversity schemes.

BACKGROUND

As all-optical communication networks become increasingly pervasive, a challenge to optical networking equipment makers is to provide optical components and subsystems that are robust, versatile, and cost-effective.

Contemporary optical communication networks commonly employ wavelength division multiplexing (WDM), for it allows multiple information (or data) channels to be simultaneously transmitted on an optical fiber by using different optical wavelengths, thereby significantly enhancing the information bandwidth of the fiber. The prevalence of WDM imposes a particular need for a line of optical systems that are capable of separating a multi- wavelength optical signal into a spatial array of spectral channels according to wavelength, so that these spectral channels can be separately detected by an array of optical power sensors, as in the case of spectral monitors; directed into an array of input/output ports (e.g., optical fibers), as in the case of optical multiplexers/demultiplexers; or dynamically routed by an array of micromirrors according to a predetermined scheme. In such optical systems, it is essential that the requisite alignment between the spectral channels and the designated beam- receiving devices (i.e., optical power sensors or micromirrors) be maintained over the course of operation, and robust with respect to environmental effects such as thermal and mechanical disturbances.

Conventional optical devices in the art, however, typically employ precision alignment, which dictates stringent fabrication tolerances and painstaking alignment during assembly, rendering these devices high in cost and cumbersome in size and operation. Moreover, there are no provisions provided for maintaining the requisite alignment over the course of operation; and no mechanisms implemented for overcoming shift in the alignment due to environmental effects such as thermal and mechanical disturbances. Altogether, these shortcomings render the prior optical devices characteristically high in cost, cumbersome in size and operation, and prone to degradation in performance.

Dense wavelength division multiplexing (DWDM) has also become prevalent in optical communication networks, in response to high bandwidth (or capacity) demand. Along with the deployment of DWDM technology comes a need for a new generation of optical components and subsystems, including optical spectral (or channel) power monitors. A particularly desirable feature for these new optical spectral power monitors is the ability to resolve multiple spectral channels that occupy a broad spectrum range (e.g., C- or L-band) with increasingly narrower frequency spacing (e.g., 50 or 25 GHz). These optical spectral power monitors are also desired to be fast in response time, robust in performance, and cost- effective in construction.

Conventional spectral power monitors typically make use of an architecture in which a diffraction grating separates a multi-wavelength optical signal by wavelength into a spatial array of spectral channels, impinging onto an array of optical power sensors. By detecting the electrical signals thus produced by the optical power sensors, an optical power spectrum of the multi-wavelength optical signal can be derived. In order to provide enhanced spectral resolution in such a system, a diffraction grating with sufficient dispersion capability is required. High-dispersion diffraction gratings commonly known in the art (e.g., holographic gratings), however, are characteristically polarization-sensitive, rendering them unsuitable for the optical spectral power monitors employing the aforementioned architecture.

In view of the foregoing, it is desirable and would be a significant advance in the art to provide for a new line of optical devices which have an optical alignment that may be actively controlled during operation, and/or which may employ a polarization diversity scheme to reduce and/or eliminate polarization sensitivity, in a simple, robust, and cost- effective construction.

. SUMMARY OF THE INVENTION The present invention provides optical systems employing active alignment compensation and polarization diversity schemes. The present invention may include optical devices, such as spectral power monitors, multiplexers and demultiplexers, and optical add-drop multiplexers, which are well suited for WDM and DWDM optical networking applications, and which may be actively aligned by use of hardware and/or software control, and/or which may employ polarization diversity schemes.

In one embodiment, the present invention provides a method and apparatus for servo-based spectral array alignment in optical systems. The optical apparatus of this embodiment of the present invention comprises an input port, for providing a multi-wavelength optical signal along with a reference signal; a wavelength-disperser for spatially separating the multi- wavelength optical signal and the reference signal by wavelength into multiple spectral channels and a reference spectral component in a spectral array with a predetermined relative arrangement; a beam-receiving array including a reference-wavelength-sensing element and a plurality of beam-receiving elements, positioned such to receive corresponding ones of the reference spectral component and the spectral channels; and a servo-control unit for maintaining the reference spectral component at a predetermined location on the reference- wavelength-sensing element and thereby ensuring a particular alignment between the spectral channels and the beam-receiving elements. In the present invention, a "spectral channel" is characterized by a distinct center wavelength and associated bandwidth, and may carry a unique information signal as in WDM optical networking applications. A "reference signal" (and the corresponding "reference spectral component") generally refers to any optical signal characterized by a well-defined (and stable) center wavelength that does not substantially overlap with any of the wavelengths of the spectral channels under consideration. Further, the terms "reference signal" (or "reference spectral component") and "calibration signal" (or "calibration spectral component") may be used interchangeably in this specification.

A beam-receiving element in the present invention should be construed broadly as embodying any optical element that corresponds with at least one spectral channel. By way of example, a beam-receiving element may be an optical power sensor, an optical fiber, a micromirror, a focusing lens, or an optical modulator. The beam-receiving elements may be configured to be in a one-to-one correspondence with the spectral channels. The beam- receiving elements may also be configured such that a subset of beam-receiving elements each corresponds with a plurality of the spectral channels.

In the spectral power monitoring apparatus of the present invention, the optical-sensing array (i.e., a photodiode array) may be configured such that the power levels of the spectral channels impinging on the photodiode array can be related to the electrical signals thus generated by a predetermined conversion matrix, which may be obtained from a calibration. Moreover, selected two (or more) adjacent channel-sensing elements in the optical-sensing array may be utilized such to provide for the reference-position-sensing element.

In one embodiment of the present invention, the alignment compensation unit is servo-based, and in one form may include an alignment-adjusting element for adjusting the alignment of the spectral channels along with the reference spectral component and a processing element. The alignment-adjusting element may be an actuation device coupled to the optical-sensing array for causing it to move, thereby adjusting the relative alignment between the spectral array and the underlying optical-sensing array. The processing element serves to monitor the real-time impinging position of the reference spectral component onto the reference-position- sensing element and to provide control of the alignment-adjusting element accordingly. The alignment compensation unit maintains the reference spectral component at a predetermined location on the reference-position-sensing element by way of servo-control, thereby ensuring the requisite alignment between the spectral array and the underlying optical-sensing array. Such a servo-based alignment compensation unit enables the spectral power monitoring apparatus of the present invention to actively correct for any shift in the alignment that may come about over the course of operation (e.g., owing to environmental effects such as thermal and mechanical disturbances), thereby enhancing the robustness of the apparatus. An additional benefit of using such a servo-based alignment compensation unit is manifested in relaxed fabrication tolerances and precision during initial assembly, rendering the spectral power monitoring apparatus of the present invention simpler and more cost-effective in construction.

The alignment-adjusting element may alternatively be a beam-steering device, such as a dynamically adjustable mirror in optical communication with the input port and the wavelength-disperser, for adjusting the alignment of the input multi-wavelength optical signal along with the reference signal. The alignment-adjusting element may also be an actuation device coupled to the wavelength-disperser (e.g., a diffraction grating), for causing the wavelength-disperser to move (e.g., rotate) and thereby adjusting the alignment of the spectral channels along with the reference spectral component. In the event that a focusing lens is employed as a beam-focuser in an optical apparatus of the present invention, the alignment-adjusting element may also be in the form of an appropriate actuation device coupled to the focusing lens, for controlling the impinging positions of the spectral channels along with the reference spectral component onto the beam-receiving array.

Moreover, the optical apparatus of the present invention may employ one or more auxiliary reference signals, along with corresponding auxiliary-reference-wavelength-sensing elements, to complement the aforementioned function of the reference spectral component. Accordingly, the servo-control unit may advantageously make use of a combination of the alignment-adjustment methods as described above to actively control the position as well the pitch of the spectral array, thereby ensuring a more robust alignment between the spectral channels and the respective beam-receiving elements.

In an alternative embodiment of the present invention, the alignment compensation unit is software-based, and may be in the form of a signal processor in communication with the optical-sensing array. The alignment compensation unit includes a predetermined calibration table containing a plurality of conversion matrices, each relating the electrical signals output from the optical-sensing array to the power levels of the impinging spectral channels at a particular impinging position of the reference spectral component. The alignment compensation unit monitors the real-time impinging position of the reference spectral component onto the reference-position-sensing element. At each impinging position of the reference spectral component thus detected, the alignment compensation unit processes the electrical signals produced by the spectral channels impinging onto the optical-sensing array and looks up a corresponding conversion matrix from the calibration table, thereby providing the power levels of the spectral channels. The spectral power monitoring apparatus thus constructed effectively compensates for any shift in the alignment that may arise over the course of operation by way of software control, without involving any "moving" actuation means. This renders the spectral power monitoring apparatus of the present invention a simpler construction with more robust performance.

The spectral power monitoring apparatus of the present invention may further employ a polarization diversity scheme, for mitigating any undesirable polarization-dependent effects that may be imposed by one or more polarization-sensitive elements in the system. This may be accomplished by disposing a polarization-separating element (e.g., a polarizing beam splitter) and a polarization-rotating element (e.g., a half-wave plate), along the optical path between the input port and the wavelength-disperser. The polarization-separating element serves to decompose the input multi-wavelength optical signal (along with the reference signal) into first and second polarization components, and the polarization-rotating element in turn rotates the polarization of the second polarization component by 90-degrees. For instance, in the event that the wavelength-disperser is provided by a diffraction grating that provides higher diffraction efficiency for p (or TM)-polarization (perpendicular to the groove lines on the grating) than for s (or TE)-polarization (orthogonal to p-polarization), the aforementioned first and second polarization components correspond to the p-polarization and s-polarization components of the multi-wavelength optical signal (along with the reference signal), respectively. The wavelength-disperser separates the first and second polarization components respectively by wavelength into first and second sets of optical beams, which subsequently impinge onto the optical-sensing array. The first and second optical beams (originating from the two polarization components) associated with each spectral channel may impinge at substantially the same location onto the optical-sensing array. Such a polarization diversity scheme has the advantage of maximizing the diffraction efficiency and therefore minimizing the insertion loss of the system.

The employment of the servo-control unit and/or the active alignment compensation hardware and/or software enables the optical apparatus of the present invention to actively correct for shift in alignment owing to environmental effects such as thermal and mechanical instabilities over the course of operation, and therefore be more robust in performance. An additional benefit of using such active alignment compensation is manifested in relaxed fabrication tolerances and precision during initial assembly, rendering the optical apparatus of the present invention a more adaptable and cost-effective construction.

According to one aspect of the present invention, a method is provided for performing spectral alignment of a multi-wavelength optical signal. The inventive method entails combining a multi- wavelength optical signal with a reference signal; spatially separating the multi-wavelength optical signal and the reference signal by wavelength into multiple spectral channels and a reference spectral component having a predetermined relative arrangement; impinging the reference spectral component at a predetermined location, such that the spectral channels impinge onto designated locations in accordance with the predetermined relative arrangement; and maintaining the reference spectral component at the predetermined location by way of servo-control, thereby ensuring that the spectral channels stay aligned at the designated locations.

In the aforementioned method of the present invention, the servo-control mechanism may be accomplished by monitoring the real-time impinging position of the reference spectral component and adjusting the alignment of the reference spectral component along with the spectral channels accordingly, so as to maintain the impinging position of the reference spectral component at the predetermined location and the spectral channels at the respective designated locations.

The method of the present invention may further include the step of focusing the spectral channels along with the reference spectral component into corresponding focused spots. It may additionally include the step of optically detecting the spectral channels at the designated locations, so as to provide a power spectrum of the detected spectral channels; the step of redirecting the spectral channels, so as to dynamically route the spectral channels according to a predetermined scheme; or modulating one or more characteristics of the spectral channels.

According to one aspect of the present invention, a method is provided that uses software- based alignment compensation in spectral power monitoring. This method includes the steps of: providing a multi-wavelength optical signal with a reference signal; spatially separating the multi-wavelength optical signal and the reference signal by wavelength into multiple spectral channels and a reference spectral component having a predetermined relative arrangement; impinging the reference spectral component and the spectral channels onto an array of optical power sensors; and determining an impinging position of the reference spectral component, and looking up a corresponding conversion matrix from a predetermined calibration table that relates output signals from the array of optical power sensors to power levels of the spectral channels impinging onto the array of optical power sensors, thereby providing a power spectrum of the multi- wavelength optical signal.

According to another aspect of the present invention, an apparatus is provided for spectral power monitoring by use of a polarization diversity scheme. The optical spectral power monitoring apparatus of the present invention comprises an input port for a multi-wavelength optical signal; a polarization-separating element that decomposes the multi-wavelength optical signal into first and second polarization components; a polarization-rotating element that rotates the polarization of the second polarization component by 90-degrees; a wavelength-disperser that separates the first and second polarization components by wavelength respectively into first and second sets of optical beams; and an array of optical power sensors (termed "optical-sensing array" herein) positioned to receive the first and second sets of optical beams.

In situations where the first and second optical beams associated with the same wavelength are desired to impinge at substantially the same location (or within the same optical power sensor) on the optical-sensing array, an auxiliary polarization-rotating element may be implemented such that the first and second sets of optical beams are polarized in two orthogonal directions upon impinging onto the optical-sensing array. This eliminates any intensity interference fringes that may arise from the spatial overlap of the optical beams. The auxiliary polarization-rotating element may be disposed between the wavelength- disperser and the optical-sensing array, such that either of the first and second sets of optical beams undergoes a 90-degree rotation in polarization prior to impinging onto the optical- sensing array.

Alternatively, a modulation assembly may he utilized in the present invention to modulate the first and second sets of optical beams prior to impinging onto the optical-sensing array. The first and second sets of optical beams may be modulated to arrive at the optical-sensing array in a time-division-multiplexed sequence. The first and second sets of optical beams may alternatively be modulated in a frequency-division-multiplexed fashion, such that the first and second sets of optical beams impinging onto the optical-sensing array carry distinct "dither" modulation signals. In either case, the use of such a modulation assembly enables the first and second sets of optical beams to be separately detected, whereby an optical power spectrum (optical power level as a function of wavelength) associated with each orthogonal polarization component in the input multi- wavelength optical signal can be independently derived. The modulation assembly may be disposed along the optical path between the polarization-separating element along with the polarization-rotating element and the wavelength-disperser, thereby controlling the first and second polarization components. It may alternatively be implemented between the wavelength-disperser and the optical-sensing array, so as to control the first and second sets of optical beams.

The modulation assembly may comprise liquid crystal shutter elements, MEMS (micro- electro-mechanical-systems) shutter elements, or electro-optic intensity modulating elements known in the art. The modulating assembly may also be provided by an optical beam- chopper (e.g., a rotating disk equipped with at least one aperture), configured to introduce distinct modulations in two incident optical signals.

The employment of the aforementioned polarization diversity scheme enables the optical spectral power monitoring apparatus of the present invention to minimize the insertion loss, while providing enhanced spectral resolution in a simple and cost-effective construction (e.g., by advantageously making use of high-dispersion diffraction gratings commonly available in the art). Further, by introducing distinct modulations to the first and second sets of optical beams prior to impinging onto the optical-sensing array, an optical power spectrum associated with each polarization component in the input multi-wavelength optical signal can be separately determined, which might be desirable in some applications.

As such, a new line of servo-based optical systems, including spectral power monitors and optical multiplexers/demultiplexers, can be constructed according to the present invention, to meet the ever-challenging demands of optical networking applications.

The novel features of this invention, as well as the invention itself, will be best understood from the following drawings and detailed description.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 provides a graphic illustration of an exemplary power spectrum of a reference spectral component and multiple spectral channels, according to the present invention;

FIGS. 2A-2D depict a first embodiment of an optical apparatus according to the present invention;

FIGS. 3A-3B show second and third embodiments of an optical apparatus according to the present invention;

FIGS. 4A-4C depict a fourth embodiment of an optical apparatus according to the present invention;

FIGS. 5A-5B show two flowcharts illustrating a method of performing spectral alignment of a multi-wavelength optical signal, according to the present invention;

FIGS. 6A-6C depict another embodiment of a spectral power monitoring apparatus according to the present invention, employing a servo-based alignment compensation unit;

FIGS. 7A-7B show another embodiment of a spectral power monitoring apparatus of the present invention, employing a software-based alignment compensation unit; FIG. 8 shows a one embodiment of an optical spectral power monitoring apparatus employing a polarization diversity scheme according to the present invention;

FIG. 9 depicts another embodiment of an optical spectral power monitoring apparatus employing a polarization diversity scheme according to the present invention; and

FIG. 10 shows another embodiment of an optical spectral power monitoring apparatus employing a polarization diversity scheme according to the present invention.

DETAILED DESCRIPTION FIG. 1 depicts an exemplary power spectrum, i.e., a plot of optical power P as a function of wavelength λ, of a reference spectral component λ_c and multiple spectral channels λi through λ . In this specification and appending claims, a "spectral channel" is characterized by a distinct center wavelength (e.g., λj) and associated bandwidth, as illustrated in FIG. 1. Each spectral channel may carry a unique information signal, as in WDM optical networking applications. A "reference spectral component" (or "reference signal"), characterized by wavelength λ_c, generally refers to any optical signal with a well-defined (and stable) center wavelength that does not substantially coincide with any of the wavelengths of the spectral channels under consideration. In FIG. 1, by way of example, the reference spectral component is shown to have a wavelength λ_c that is shorter than the wavelengths of the spectral channels. In general, the spectral channels need not be evenly spaced in wavelength (or frequency).

The discussion below describes the present invention employing active alignment compensation and a polarization diversity scheme in the following manner: (i) Section I describes a method and an which employs a servo-based system to achieve active alignment compensation in optical systems; (ii) Section II describes other hardware and software solutions for active alignment compensation in optical systems; and (iii) Section III describes polarization diversity schemes that may be employed within the present inventions.

SERVO-BASED ACTIVE ALIGNMENT COMPENSATION FIG. 2A depicts a first embodiment of an optical apparatus according to the present invention. By way of example to illustrate the principles and the general architecture of the present invention, the optical apparatus 200 comprises an input port 210 for a multi- wavelength optical signal which may be in the form of a fiber collimator; an alignment- adjusting element which in one form may be a steering-mirror 260-1, a wavelength-disperser 220 which may be a diffraction grating; a beam-focuser 230 in the form of a focusing lens; and an beam-receiving array including a reference-wavelength-sensing element 240 and a plurality of beam-receiving elements 250-1 through 250-N. In this specification and appending claims, a beam-receiving element is construed broadly as embodying any optical element that receives one or more spectral channels. It may be, for instance, an optical power sensor, a micromirror, an optical fiber, a focusing lens, or an optical modulator, as will be described in more detail later.

The optical apparatus 200 of FIG. 2A may operate as follows. The input port 210 transmits a multi-wavelength optical signal containing wavelengths λi through λpj along with a reference signal containing wavelength λ_c. The optical signals are then directed onto the diffraction grating 220 by way of the steering-mirror 260-1. The diffraction grating 220 angularly separates the multi-wavelength optical signal and the reference signal by wavelength into multiple spectral channels λj through λw and a reference spectral component λ_c having a predetermined relative arrangement. The focusing lens 230 focuses the reference spectral component and the spectral channels into corresponding focused spots, e.g., in a spectral array in accordance with the predetermined relative arrangement. The beam-receiving array containing the reference-wavelength-sensing element 240 and the beam-receiving elements 250-1 through 250-N is positioned such that upon the reference spectral component λ_c impinging onto the reference-wavelength-sensing element 240 at a predetermined location ΛΓ₀, the spectral channels λi through λ^ impinge onto the beam-receiving elements 250-1 through 250-N at designated locations xi through ΛΓ_N respectively.

It should be noted that the embodiment of FIG. 2 A and the following figures are shown in schematic form, for illustrative purpose only. Various elements and optical beams are not drawn to scale. In general, there can be any number of spectral channels in an optical apparatus of the present invention (so long as they are commensurate with the underlying beam-receiving elements). Moreover, the focused spots of the diffracted optical beams impinging onto the beam-receiving array shown in FIG. 2A (and the following figures) may be unevenly spaced.

The optical apparatus 200 of FIG. 2A may further comprise a servo-control unit 260, which in one form may include the steering-mirror 260-1 and a processing element 260-2. The steering-mirror 260-1 dynamically adjusts the alignment of the multi-wavelength optical signal along with the reference signal, thereby controlling the impinging positions of the spectral channels and the reference spectral component onto the beam-receiving array. The processing element 260-2 monitors the real-time impinging position of the reference spectral component λ_c on the reference-wavelength-sensing element 240 and provides feedback (or servo) control of the steering-mirror 260-1 accordingly, so as to maintaining the reference spectral component λ_c at the predetermined location x₀ and therefore the spectral channels λi through λi at the designated locations xi through x^. As such, the servo-control unit 260 enables the optical apparatus of the present invention to actively correct for shift in alignment owing to environmental effects, such as thermal and/or mechanical instabilities over the course of operation, and therefore be more robust in performance. An additional benefit of using such a servo-control unit is manifested in relaxed fabrication tolerances and precision during initial assembly, rendering the optical apparatus of the present invention a more adaptable and cost-effective construction.

In the above embodiment, the reference-wavelength-sensing element 240 may be a position sensitive detector, a quadrant detector, a split detector, or any other position-sensitive means known in the art, which allows the real-time impinging position (in one or two dimensions) of an optical beam to be monitored by way of electrical (e.g., current or voltage) signals produced by the sensing element. By way of example, FIG. 2B shows a schematic illustration of a position sensitive detector 240-A, upon which an optical beam 241 impinges. The impinging position of the optical beam 241 in the x-direction can be deduced by detecting a pair of output signals, e.g., current signals I_xι and I_x2, whose relative magnitudes provide a measure of the beam spot in the x-direction. Likewise, the impinging position of the optical beam 241 in the y-direction can be obtained by measuring another pair of current signals I_yι and I_y2. Moreover, the deviation of the real-time impinging position of the optical beam 241 from a designated location, e.g., the center point O on the position sensitive detector 240-A, may be monitored by detecting the output signals in an appropriate normalized differential detection scheme (e.g., by measuring (I_xι-I_X2)/(Iχi+Iχ₂) and/or (I_yl- I 2)/(Iyi+I 2))j as commonly practiced in the art. Those skilled in the art will also know how to make use of other types of position-sensitive means known in the art to provide for a reference-wavelength-sensing element in the present invention.

The steering-mirror 260-1 of FIG. 2A may be a dynamically adjustable mirror that is rotatable about one or two axes. For instance, it may be a silicon micromachined mirror with an appropriate actuation mechanism; it may also be provided by coupling an actuation device known in the art to a mirror or beam-deflecting element. The processing element 260-2 of FIG. 2A may include electrical circuits, controllers and signal processing programs for processing the output signals received from the reference-wavelength-sensing element 240 (e.g., the current signals from the position sensitive detector 240-A of FIG. 2B) and deriving from the detected signals the real-time impinging position of the reference spectral component λ_c. The processing element 260-2 accordingly generates appropriate control signals to be applied to the alignment-adjusting element, e.g., the steering-mirror 260-1 of FIG. 2A, so as to adjust the alignment of the reference spectral component along with the spectral channels in such a way that the reference spectral component λ_c is maintained at the predetermined location x₀. The electronic circuitry and the associated signal processing algorithm software for a processing element in a servo-control system are known in the art of electrical engineering and servo control systems.

FIGS. 2C-2D depict two exemplary ways of adjusting the alignment of optical beams in the embodiment of FIG. 2A. In FIG. 2C, a first optical beam 271 containing wavelength λi (which may represent a spectral channel) is incident onto the diffraction grating 220 at an incidence angle θi„, and is diffracted from the diffraction grating 220 as a first diffracted beam 272 at a diffraction angle θ_out, as determined by the grating equation: λ sin^ +sin^, = rø- (1) a where m is the diffraction order and d is the grating pitch (i.e., the spacing between two adjacent groove lines on the grating). Both θi„ and θ₀„_t are measured with respect to the normal axis 270 of the diffraction grating 220. If the incidence angle of the optical beam upon the diffraction grating 220 is varied by Δθj„, as indicated by a second optical beam 273, the diffraction angle of the diffracted beam changes correspondingly by Δθ_out, as indicated by a second diffracted beam 274. Hence, varying the incidence angle of the multi-wavelength optical signal along with the reference signal upon the diffraction grating 220, e.g., by action of the steering-mirror 260-1 in FIG. 2A, causes the diffraction angles of the spectral channels and the reference spectral component to change correspondingly, thereby enabling the reference spectral component λ_c to impinge at the predetermined location x₀ and the spectral channels λi through λN at the designated locations x through JC_N.

Instead of (or in conjunction with) varying the incidence angle of the optical beams upon the diffraction grating 220 (e.g., by way of the steering-mirror 260-1 in FIG. 2A), the diffraction grating 220 itself may be rotated, thereby performing a similar alignment function, as illustrated in FIG. 2D. In this case, a first optical beam 281 containing wavelength λj is incident onto the diffraction grating 220, and is diffracted from the diffraction grating 220 as a first diffracted beam 282. Rotating the diffraction grating 220 by an angle Δθ_g, as indicated by the thus-rotated diffraction grating 221, effectively changes the diffraction angle by Δθ_out, as indicated by a second diffracted beam 283. The rotation of the diffraction grating may be accomplished by coupling to the grating an appropriate actuation device, such as a voice coil actuator, a stepping motor, a solenoid actuator, a piezoelectric actuator, or other types of actuation means known in the art. The actuation device may in turn be controlled by the processing element in a servo-control unit.

In the embodiment of FIG. 2 A, the angular dispersion D of the diffraction grating 220 can be derived from the grating equation in Eq. (1):

<img file="WO03025630A2_D0001.tif" /> Let the focal length of the focusing lens 230 be/ The pitch P - namely, the spacing between any two adjacent spectral spots - of the spectral array formed by the diffracted optical beams can be expressed as

P = f A ^ = ^{f Aλ m 1} (3) d λ d cos#_∞, where λ is the wavelength difference between two adjacent spectral channels. Eq. (3) shows that the pitch of the spectral array generally varies with the diffraction angle θ_out_&gt; unless θ_out is z;ero. The rate of change in P with respect to θ_out can be shown to be dP ₌ f Aλ m sin^, dθ_out d cos² ^ ⁾

In FIG. 2C or 2D described above, since the alignment adjustment is brought about by varying the incidence angle θ_in and therefore the diffraction angle θ_out, Eq. (3) indicates that the pitch of the spectral array may vary as the alignment adjustment takes place, particularly at large values of the diffraction angle θ_out- Hence, in the embodiment of FIG. 2A (where the alignment-adjusting method shown in FIG. 2 C or 2D may be implemented), the constituent beam-receiving elements should be configured such that they can accommodate variations in the pitch of the spectral array. (For example, the sizes of the beam-receiving elements may be such that the variations in the pitch of the spectral array do not substantially alter the correspondence between the spectral channels and the respective beam-receiving elements substantially, and are therefore inconsequential in practice.) The embodiment of FIG. 2A may also be desired in applications where the aforementioned change in the pitch of the spectral array is so small (e.g., in the event that the diffraction angle θ_out is close to zero) that it is inconsequential practically.

Instead of performing alignment adjustment by varying the incidence angle of the input multi-wavelength optical signal along with the reference signal, as exemplified in FIG. 2C or 2D, the reference-wavelength-sensing element 240 along with the beam-receiving elements 250-1 through 250-N in FIG. 2A may be moved in tandem, e.g., by translating and/or rotating the beam-receiving array as a whole, so as to enable the reference spectral component along with the spectral channels to impinge at the designated locations. Alternatively, the focusing lens 230 in the embodiment of FIG. 2A may be moved, e.g., shifted or translated, so as to control the impinging locations of the diffracted optical beams.

FIG. 3A shows a second embodiment of an optical apparatus of the present invention. By way of example, optical apparatus 300 makes use of the architecture as well as many of the elements employed in the embodiment of FIG. 2A, as indicated by those labeled with identical numerals. In this case, the beam-receiving array containing the reference- wavelength-sensing element 240 and the beam-receiving elements 250-1 through 250-N may be integrated into a single structure, e.g., by mounting or fabricating the constituent elements on a substrate, for the ease of operation. Servo-control unit 360 may include an alignment- adjusting element 360-1 which may be a linear actuation device coupled to the beam- receiving array, and a processing element 360-2. The actuation device 360-1 is configured such to cause the beam-receiving array as a whole - therefore the reference-wavelength- sensing element 240 along with the beam-receiving elements 250-1 through 250-N in tandem - to move, e.g., translate along a direction that is substantially transverse to the direction of propagation of the impinging optical beams, thereby adjusting the relative alignment between the spectral array formed by the diffracted optical beams and the underlying beam-receiving array. As in the embodiment of FIG. 2A, the processing element 360-2 serves to monitor the real-time impinging position of the reference spectral component λ_c on the reference- wavelength-sensing element 240 and provides servo control of the actuation device 360-1 accordingly, so as to maintain the reference spectral component λ_c at the predetermined location x₀ and the spectral channels λi through λ_N at the respective designated locations CI through X_N.

FIG. 3B shows a third embodiment of an optical apparatus of the present invention. By way of example, optical apparatus 350 makes use of the architecture as well as many of the elements employed in the embodiment of FIG. 3A, as indicated by those labeled with identical numerals. An alternative servo-control unit 365, including an alignment-adjusting element 365-1 in the form of an actuation device coupled to the focusing lens 230 and a processing element 365-2, may be implemented. The actuation device 365-1 causes the focusing lens 230 to move, e.g., shift, rotate, or translate, thereby controlling the impinging locations of the diffracted optical beams on the reference-wavelength-sensing element 240 and the beam-receiving elements 250-1 through 250-N, respectively. As in the case of the embodiment of FIG. 3A, the processing element 365-2 monitors the real-time impinging position of the reference spectral component λ_c on the reference-wavelength-sensing element 240 and provides servo control of the actuation device 365-1 accordingly, thereby maintaining the reference spectral component λ_c at the predetermined location x₀ and the spectral channels λi through λ_N at the respective designated locations xi through x^.

The actuation device 360-1 in the embodiment of FIG. 3A, or the actuation device 365-1 in the embodiment of FIG. 3B, may be a stepping motor, a solenoid actuator, a piezoelectric actuator, a voice coil actuator, or other types of actuation means known in the art. The processing element 360-2 of FIG. 3A, or the processing element 365-2 of FIG. 3B, may be substantially similar to the processing 260-2 of FIG. 2A in configuration and operation. An advantage of the embodiment of FIG. 3A or 3B is evident in that the underlying alignment- adjustment method does not substantially alter the pitch of the spectral array - that is, only the relative alignment between the spectral array and the beam-receive array is adjusted. It should be appreciated that care should be taken in designing the focusing lens 230 in the embodiment of FIG. 3B; such that aberrations and other imperfections are substantially eliminated. As will be appreciated from the teachings of this specification, one skilled in the art would know how to devise an appropriate alignment-adjusting method along with corresponding servo-control system according to the present invention, to suit a given application.

In the embodiment of FIG. 2A, 3A, or 3B, the multi-wavelength optical signal containing wavelengths λi through λN may be provided by an input optical fiber 201 coupled to the fiber collimator that serves as the input port 210, and the reference signal λ_c may be provided by a reference light source 202. An optical combiner 203, which in one form may be a fiber-optic fused coupler, may be used to couple the reference light source 202 to the input optical fiber 201, such that both the multi-wavelength optical signal and the reference signal are directed into the input port 210. The optical apparatus 200 thus has an independent, internally generated reference light source. Alternatively, the multi-wavelength optical signal itself may include a spectral component (e.g., a service channel in an optical network) that can serve as the reference signal, as in WDM optical networking applications. In such a scenario, the internal reference light source 202 along with the fiber-optic coupler 203 need not be implemented.

In the present invention, one or more auxiliary reference signals, along with corresponding reference-wavelength-sensing elements, may be additionally employed to complement the aforementioned function of the reference spectral component λ_c. FIG. 4A depicts a fourth embodiment of an optical apparatus of the present invention. By way of example, optical apparatus 400 makes use of the architectures as well as a number of the elements used in the embodiments of FIGS. 2 A and 3 A, as indicated by those labeled with identical numerals. In addition, an auxiliary reference light source 402 may be coupled to the input optical fiber 201 by way of an auxiliary optical combiner 403 which may be a fiber-optic coupler, so as to couple an auxiliary reference signal containing wavelength λ_c' to the input port 210. The auxiliary reference signal λ_c', along with the multi-wavelength optical signal and the reference signal λ_c, may then be directed onto the diffraction grating 220 by way of the steering-mirror 260-1. The wavelength λ_c' of the auxiliary reference signal may be selected to be longer than the wavelengths of the spectral channels such that upon diffraction, the auxiliary reference signal λ_c' impinges onto an auxiliary-reference-wavelength-sensing element 441 at a prescribed location x₀'. As such, the reference spectral component λ_c, the spectral channels λi through λN, and the auxiliary reference signal λ_c' form a spectral array with a predetermined relative arrangement. Accordingly, the reference-wavelength-sensing element 240, the beam-receiving elements 250-1 through 250-N, and the auxiliary-reference- wavelength-sensing element 441 form a beam-receiving array, configured such to receive the spectral array. The beam-receiving array may be integrated into a single structure, e.g., by mounting or fabricating the constituent elements on a substrate.

The auxiliary-reference-wavelength-sensing element 441 may be a position sensitive detector, a split detector, a quadrant detector, or other types of position-sensitive means known in the art. It should be appreciated by one skilled in the art that the aforementioned reference signal and auxiliary reference signal may also be termed first and second reference signals; and the reference-wavelength-sensing element and the auxiliary-reference- wavelength-sensing element may be termed first and second reference-wavelength-sensing elements, accordingly. • Moreover, the wavelength-disperser, such as the diffraction grating 220, may direct a (first) reference spectral component λ_c in the first reference signal and a (second) reference spectral component λ_c' in the second reference signal onto the first and second reference-wavelength-sensing elements at first and second predetermined locations, respectively.

The embodiment of FIG. 4A may further include an actuation device 460-1 and a processing element 460-2. By way of example, the actuation device 460-1 may be coupled to the aforementioned beam-receiving array, so as to cause the beam-receiving array as a whole - therefore the reference-wavelength-sensing element 240 along with the beam-receiving elements 250-1 through 250-N and the auxiliary-reference-wavelength-sensing element 441 in tandem - to move, e.g., translate along a direction substantially transverse to the direction of propagation of the spectral channels, and/or rotate as indicated by a curved arrow 470. For instance, the actuation device 460-1 may cause the beam-receiving array to rotate about a pivotal point positioned at the predetermined location x₀. As such, the actuation device 460-1 may be utilized primarily for adjusting the relative alignment between the spectral array formed by the diffracted optical beams and the underlying beam-receiving array. The processing element 460-2 may monitor the real-time impinging position of the reference spectral component λ_c on the reference-wavelength-sensing element 240 and provide servo control of the actuation device 460-1 accordingly, so as to maintain the reference spectral component λ_c at the predetermined location x₀ and the spectral channels λi through λ_N at the respective designated locations Xi through xj .

The processing element 460-2 may additionally monitor the real-time impinging position of the auxiliary reference signal λ_c' on the auxiliary-reference-wavelength-sensing element 441. Such information might be useful for monitoring misalignment between the spectral channels and the respective beam-receiving elements that may not be reflected by the impinging position of the reference spectral component λ_c. By way of example, FIG. 4B illustrates a situation where the reference spectral component λ_c remains at the predetermined location C₀, whereas the impinging position of the auxiliary reference signal λ_c' on the auxiliary- reference-wavelength-sensing element 441 is deviated along the jc-direction from the prescribed location c_o', which may result from a change in the pitch of the spectral array. The x-y plane in this figure (as well as in FIG. 4C) is shown to be substantially transverse to the direction of propagation of the spectral channels. As indicated in the above discussion, the pitch of the spectral array generally varies with the diffraction angle and therefore the incidence angle of the optical signals upon the diffraction grating (e.g., see Eqs. (3) and (4) above). Accordingly, the processing element 460-2 may use the detected deviation of the auxiliary reference signal λ_c' from the prescribed location x₀' to control the steering-mirror 260-1 in a manner effective to bring the auxiliary reference signal λ_c' back to the prescribed location x₀', e.g., by adjusting the incidence angle of the input multi-wavelength optical signal along with the reference signal and the auxiliary reference signal on the diffraction grating 220 in a manner similar to the alignment-adjusting method described in FIG. 2C. The alignment of the reference spectral component λ_c and the auxiliary reference signal λc' at the respective locations x₀, x₀' are indicative of the requisite alignment between the spectral channels and the respective beam-receiving elements.

By way of example, FIG. 4C illustrates another situation where the reference spectral component λ_c remains at the predetermined location C₀, whereas the impinging position of the auxiliary reference signal λ_c' on the auxiliary-reference-wavelength-sensing element 441 is deviated from the prescribed location x₀' as indicated, which may be brought about by a rotational motion of the beam-receiving array relative to the spectral array (or vice versa). Accordingly, the processing element 460-2 may use the detected deviation of the auxiliary reference signal λ_c' from the prescribed location Λ:₀' to control the actuation device 460-1 in a manner effective to bring the auxiliary reference signal λ_c' back to the prescribed location x₀', e.g., by rotating the beam-receiving array relative to the spectral array, thereby restoring the requisite alignment between the spectral channels and the respective beam-receiving elements.

Those skilled in the art will appreciate that the embodiments of FIGS. 4B-4C are provided as a way of example to elucidate the general principles of the present invention. In a practical situation, a deviation of the auxiliary reference signal λ_c' from the prescribed location may be due to a number of effects, e.g., a combination of change in the pitch of the spectral array and rotational motion of the beam-receiving array (relative to the spectral array). Accordingly, the processing element 460-2 may control both the actuation device 460-1 and the steering- mirror 260-1 in a coordinated fashion, in order to bring the auxiliary reference signal λ_c' back to the designated location, while maintaining the reference spectral component λ_c at the predetermined location x₀, thereby restoring the requisite alignment between the spectral channels and the beam-receiving elements. Moreover, the control of the pitch of the spectral array may be accomplished by the alignment-adjusting method described in the embodiment of FIG. 2D, in lieu of (or in conjunction with) the function of the steering-mirror 260-1. The relative alignment between the spectral array and the underlying beam-receiving array may be adjusted by coupling an appropriate actuation device to the focusing lens 230 as described in the embodiment of FIG. 3B, in lieu of (or in conjunction with) the alignment function provided by the actuation device 460-1. Additionally, the impinging position of the auxiliary reference signal λ_c' may be maintained at the prescribed location by way of servo control, whereas the impinging position of the reference spectral component λ_c is being monitored periodically or continuously; or the impinging positions of both the reference spectral component and the auxiliary reference signal may be actively controlled according to an appropriate signal processing and servo-control scheme. In the embodiment of FIG. 4A, the servo-control unit may generally comprise a first alignment-adjusting element (e.g., the actuation device 460-1, or an appropriate actuation device coupled to the focusing lens 230) for adjusting a relative alignment between the spectral array formed by the diffracted optical beams and the underlying beam-receiving array; a second alignment-adjusting element (e.g., the steering-mirror 260-1, or an appropriate actuation device coupled to the diffraction grating 220) for controlling the pitch of the spectral array; and a processing element (e.g., the processing element 460-2) in communication with the first and second alignment-adjusting elements, as well as the reference-wavelength-sensing element 240 and the auxiliary-reference-wavelength-sensing element 441. The processing element 460-2 may monitor the impinging positions of the reference spectral component λ_c and the auxiliary reference signal λ_c' onto the reference- wavelength-sensing element 240 and the auxiliary-reference-wavelength-sensing element 441, respectively, and provide control of the first and second alignment-adjusting elements accordingly, so as to maintain the reference spectral component λ_c and the auxiliary reference signal λ_c' at their respective designated locations, and thereby ensure the requisite alignment between the spectral channels and the respective beam-receiving elements.

As such, the optical apparatus of FIG. 4A advantageously makes use of a combination of appropriate alignment-adjustment methods to actively control the position as well the pitch of the spectral array, therefore being more robust in performance.

In general, one or more auxiliary reference signals in the present invention may be any optical signals with well-defined (and stable) center wavelengths that do not substantially coincide with any of the wavelengths of the spectral channels and the reference spectral component λ_c. In the embodiment of FIG. 4 A, by way of example, the wavelength λ_c' of the auxiliary reference signal is shown to be longer than the wavelengths of the spectral channels, whereas the wavelength λ_c of the reference spectral component is shorter than the wavelengths of the spectral channels, and both the reference signals are provided by the internal reference light sources as shown. It should be noted that the two reference light sources in FIG. 4 A may be coupled to the input fiber by a single optical combiner (e.g., a 3x1 fiber-optic coupler); or the reference signal and the auxiliary reference signal may be provided by a single reference light source that is capable of providing a plurality of reference signals, coupled to the input fiber by an optical combiner. Alternatively, the multi- wavelength optical signal itself may include one or more spectral components (e.g., one or more service channels in an optical network) that can be used as one or more reference signals. One skilled in the art will know how to implement appropriate reference signals in an optical apparatus according to the present invention, to suit a given application.

In the above embodiments, the diffraction grating 220 may be a ruled diffraction grating, a holographic diffraction grating, an echelle grating, or a dispersing prism, all commonly employed in the art for separating a multi-wavelength signal by wavelength. By way of example, the wavelength-disperser in the aforementioned embodiments is shown to be in the form of a reflective diffraction grating. One skilled in the art will appreciate that a transmission diffraction grating, or a dispersing prism, may be alternatively implemented in an optical apparatus of the present invention. The beam-focuser may additionally be an assembly of focusing lenses, or any other suitable beam-focusing means known in the art. The focusing function may also be provided by using a curved diffraction grating that performs a dual function of wavelength-separating and beam-focusing. It should be noted that in applications where the spectral channels along with the reference spectral component are well separated, the beam-focuser, such as the focusing lens 230 in the above embodiments, might not be utilized.

Moreover, the beam-receiving elements 250-1 through 250-N may be optical power sensors, such as photodiodes in the form of pn photodetectors, pin (p-intrinsic-n) photodetectors, or avalanche photo detectors (APDs). The optical apparatus thus constructed constitutes a spectral power monitor with servo-control capability, providing a characteristic power spectrum of the spectral channels of interest. The beam-receiving elements 250-1 through 250-N may also be micromirrors (e.g., silicon micromachined mirrors), each being individually controllable (e.g., pivotable about one or two axes) to dynamically route the spectral channels according to a predetermined scheme. The beam-receiving elements 250-1 through 250-N may alternatively be an array of optical fibers, into which the spectral channels are directed. The optical apparatus thus constructed constitutes a demultiplexer, or a multiplexer upon reversing the propagation of optical beams. The beam-receiving elements 250-1 through 250-N may additionally be in the form of an array of beam-shaping elements, such as focusing lenses, so as to project the spectral channels at desired locations. The beam- receiving elements 250-1 through 250-N can also be in the form of an array of optical modulators, such as liquid-crystal light modulators or optical attenuators, for modulating one or more characteristics (e.g., amplitude and/or phase) of each spectral channel.

In FIG. 2A, 3A, 3B or 4A, by way of example, the beam-receiving elements are shown to have a one-to-one correspondence with the spectral channels. There might be applications where a subset of beam-receiving elements each corresponds with a plurality of the spectral channels, or a plurality of beam-receiving elements are designated to a single spectral channel. For instance, in the event that optical power sensors are used as the beam-receiving elements, one or more optical power sensors may each be assigned to receive a plurality of the spectral channels, so as to provide an integrated power measurement of the received spectral channels.

It is known that the diffraction efficiency of a diffraction grating is generally polarization- dependent, and that the polarization-dependent effect may become considerable for a grating with a large number of groove lines (per unit length). Thus, in the event that a diffraction grating is used as a wavelength-disperser, as is in the embodiment of FIG. 2A, 3A, 3B or 4A, various means/mechanisms may be utilized to mitigate the associated polarization-sensitive effects, such as those mechanisms discussed below in Section III. By way of example, a polarization diversity scheme may be implemented. In this scenario, the input multi- wavelength optical signal (along with one or more reference signals) is first decomposed into a P-polarization portion and an S-polarization portion. Assuming that P-polarization is the preferred direction of the diffraction grating (i.e., the diffraction efficiency is higher for P- polarization than for S-polarization), the S-polarization portion is then rotated by 90-degrees, whereby the optical signals incident onto the diffraction grating all possess P-polarization. Such a polarization diversity scheme has the advantage of maximizing the diffraction efficiency. Alternatively, a suitable polarization-sensitive element (e.g., a leaky beamsplitter) may be implemented, serving to attenuate the P-polarization portion relative to the S- polarization portion in the input multi-wavelength optical signal (along with one or more reference signals) according to a predetermined ratio prior to impinging onto the diffraction grating, so as to compensate for the differential treatment in polarization inflicted by the diffraction grating. The apparatus and methods used for these polarization diversity schemes are discussed below in more detail in Section III. The present invention further provides a method of spectral alignment of a multi-wavelength optical signal. As a way of example to illustrate the general principles of the present invention, FIG. 5A shows an exemplary flowchart, outlining the method of the present invention. Method 500 entails combining the multi-wavelength optical signal along with a reference (or calibration) signal, as indicated in step 510; spatially separating the multi- wavelength optical signal and the reference (or calibration) signal by wavelength into multiple spectral channels and a reference (or calibration) spectral component having a predetermined relative arrangement, as indicated in step 520; impinging the reference (or calibration) spectral component at a predetermined location, such that the spectral channels impinge onto designated locations in accordance with the predetermined relative arrangement, as indicated in step 530; maintaining the reference (or calibration) spectral component at the predetermined location by way of servo-control and thereby ensuring that the spectral channels stay aligned at the designated locations, as indicated in step 540.

The aforementioned method of the present invention utilizes the fact that the reference spectral component and the spectral channels, each characterized by a distinct center wavelength, form a spectral array with a predetermined relative arrangement. Thus, aligning the reference spectral component at a predetermined location ensures the spectral channels simultaneously impinge onto the designated locations in accordance with the spectral array. This provides a simple and effective way of aligning a spectral array formed by a multi- wavelength optical signal. The thus-aligned spectral channels may then be individually manipulated, for example, by an array of beam-receiving elements, as described above.

FIG. 5B illustrates in further detail an exemplary embodiment of the servo-control operation recited in the step 540 of FIG. 5A. It entails monitoring the real-time impinging position of the reference (or calibration) spectral component, as indicated in step 540-A; and adjusting the alignment of the reference (or calibration) spectral component along with the spectral channels accordingly, so as to maintain the impinging position of the reference (or calibration) spectral component at the predetermined location and thereby ensure that the spectral channels stay aligned at the designated locations, as recited in step 540-B.

The method 500 of FIG. 5A (or FIG. 5B) may further include the step of focusing the spectral channels along with the reference (or calibration) spectral component into corresponding focused spots, as indicated in step 550. The method 500 of FIG. 5A (or FIG. 5B) may additionally include the step of optically detecting the spectral channels at the designated locations, so as to provide a power spectrum of the detected spectral channels; the step of redirecting the spectral channels, so as to route the spectral channels according to a predetermined scheme; or modulating one or more characteristics of the spectral channels.

II. OTHER HARDWARE AND SOFTWARE SOLUTIONS FOR ACTIVE ALIGNMENT COMPENSATION IN OPTICAL SYSTEMS

FIG. 6A depicts an exemplary embodiment of a spectral power monitoring apparatus according to the present invention. By way of example to illustrate the principles and the general architecture of the present invention, the spectral power monitoring apparatus 600 comprises an input port 610 for a multi- wavelength optical signal which may be in the form of a fiber collimator; a wavelength-disperser 620 which in one form may be a diffraction grating; a beam-focuser 630 which may be a focusing lens; an array 640 of optical power sensors (termed "optical-sensing array", herein), providing a reference-position-sensing element 640-C and a plurality of channel-sensing elements 640-1 through 640-N. The optical-sensing array 640 may be integrated into a single structure (e.g., by mounting or fabricating the constituent elements on a substrate).

The spectral power monitoring apparatus 600 of FIG. 6A may operate as follows. The input port 610 transmits a multi-wavelength optical signal containing wavelengths λi through λ along with a reference signal containing wavelength λ_c. The diffraction grating 620 angularly separates the incident multi-wavelength optical signal along with the reference signal by wavelength into multiple spectral channels λi through λ_N and a reference spectral component λ_c having a predetermined relative arrangement. The focusing lens 630 focuses the reference spectral component λ_c and the spectral channels λi through λN into corresponding focused spots, e.g., in a spatial array (or "spectral array") with the predetermined relative arrangement. The optical-sensing array 640 may be positioned such that when the reference spectral component λ_c impinges onto the reference-position-sensing element 640-C at a predetermined location C₀, the spectral channels λi through λN impinge onto the channel- sensing elements 640-1 through 640-N at designated locations x_\ through x^, respectively. It should be noted that the embodiment of FIG. 6A and the following figures are shown in schematic form, for illustrative purpose only. Various elements and optical beams are not drawn to scale. In general, there can be any number of the spectral channels in a spectral power monitoring apparatus of the present invention, so long as the number of the channel- sensing elements employed in the system is adequate for determining the power levels of the spectral channels with desired accuracy. Moreover, the focused spots of the diffracted optical beams impinging onto the optical-sensing array shown in FIG. 6A (and the following figures) may not be evenly spaced, and need not necessarily be in a one-to-one correspondence with the underlying channel-sensing elements, as will be described in further detail later.

The spectral power monitoring apparatus 600 of FIG. 6A may further comprise a servo-based alignment compensation unit 660, which in one form may include an actuation device 660-1 coupled to the optical-sensing array 640 and a processing element 660-2. The actuation device 660-1 is configured such to cause the optical-sensing array 640 as a whole - hence the reference-position-sensing element 640-C along with the channel-sensing elements 640-1 through 640-N in tandem - to move (e.g., translate and/or rotate), thereby adjusting a relative alignment between the spectral array formed by the diffracted optical beams and the underlying optical-sensing array 640. The processing element 660-2 monitors the real-time impinging position of the reference spectral component λ_c on the reference-position-sensing element 640-C and provides servo (or feedback) control of the actuation device 660-1 accordingly, so as to maintain the reference spectral component λ_c at the predetermined location x₀ and therefore the spectral channels λi through λ_N at the designated locations xi through Λ: . The servo-based alignment compensation unit thus described enables the optical apparatus of the present invention to actively correct for any shift in alignment that may come about over the course of operation (e.g., owing to environmental effects such as thermal and/or mechanical disturbances), therefore increasing the robustness of the apparatus. An additional benefit of using such an alignment compensation unit is manifested in relaxed fabrication tolerances and precision during initial assembly, rendering the spectral power monitoring apparatus of the present invention simpler and more cost-effective in construction.

By way of example, FIG. 6B illustrates how an array of photodiodes may be implemented as the optical-sensing array 640 in the embodiment of FIG. 6A. Shown in FIG. 6B is an exemplary section of a photodiode array 640 A in a magnified view, comprising a plurality of contiguous photo-sensing elements with varying photo-response characteristics as distinguished by non-hatched and hatched regions illustrated in the figure. As a way of example, photo-response function Rι(x), as illustrated by solid lines in the figure, characterizes the photo-response of a non-hatched photo-sensing region 640-i and its two adjacent hatched regions 640-i-H, 640-j-H. Similarly, photo-response function R_j(x), as illustrated by dashed lines, characterizes the photo-response of a neighboring non-hatched photo-sensing region 640-j and its two adjacent hatched regions 640-j-H, 640-k-H. A photo- response function relates the optical power impinging onto a photo-sensing element to the electrical (e.g., voltage) signal thus generated, as will be described in further detail later. By way of example, each photo-response function in FIG. 6B is shown to be nearly constant across the corresponding non-hatched region and decreases in a nearly linear fashion as moving away from the non-hatched region into the adjacent hatched regions, therefore exhibiting a "trapezoid-like" overall behavior. As such, the photodiode array 640A possesses a continuous overall photo-response function; that is also to say that there are no "dead zones" (or photo insensitive regions) on the photodiode array 640A. The photodiode arrays with the characteristics thus-described are commercially available, for instance, from Sensors Unlimited, Inc., Princeton, New Jersey.

As a way of example, the photodiode arrays contemplated in the ensuing discussion include appropriate detection circuits, such that the output signals may be in the form of voltage signals. It will be appreciated that the general principles and operation of the present invention are also applicable to other photodiode arrays or optical power sensor arrays whose output signals are in the form of current signals. It will also be appreciated that subscript i, j, or k in this specification may assume any integer value between 1 and N.

The photodiode array 640A in FIG. 6B may be configured such that voltage signals are output through the non-hatched regions. By way of example, voltage signal Vi output from the non-hatched region 640-i may be given by V, = \R,(x)I(x,y) dxdy (1) where the integration is taking place over the non-hatched region 640-i and its adjacent hatched regions 640-i-H, 640-j-H, and I (x, y) is the light intensity impinging upon the region of interest in the x-y plane defined in FIG. 6B. The photo-response function Rj(x) is predetermined and based upon the characteristics of the photodiode array employed. Thus, the voltage signal Vj takes into account the total optical power impinging upon the non- hatched region 640-i as well as its adjacent hatched regions 640-i-H, 640-j-H. Likewise, voltage signal Vj output from the non-hatched region 640-j is related to the total optical power impinging upon the non-hatched region 640-j and its adjacent hatched regions 640-j- H, 640-k-H. Moreover, because of the intertwined relationship between two spatially adjacent photo-response functions, such as Ri(x) and Rj(x), the power level of an optical beam impinging upon a hatched region, such as the hatched region 640-i-H sandwiched between the non-hatched regions 640-i, 640-j, can be obtained from the measured voltage signals Vj and Vj. Such is also the case anywhere else in the photodiode array 640 A. Hence, each non-hatched region along with its adjacent hatched regions, such as the non-hatched region 640-i and the adjacent hatched regions 640-i-H, 640-j-H, constitutes a channel- sensing element (or pixel) in the present invention.

In addition, two adjacent channel-sensing elements in the photodiode array 640A may be utilized as a "split detector" to provide for a reference-position-sensing element (e.g., the reference-position-sensing element 640-C in the embodiment of FIG. 6A) for the reference spectral component λ_c. This may be accomplished by measuring the voltage signals Vi, V₂ output from the non-hatched regions 640-1, 640-2 respectively using an appropriate normalized differential detection scheme known in the art, for example, by way of monitoring a position error signal (Vι-V2)/(Vι+V2). Such a normalized differential detection scheme has the advantage of improving the signal-to-noise (SNR) ratio of the detection via common mode rejection of amplitude noise. As a way of example, the impinging location of the reference spectral component λ_c may be positioned on a hatched region 640-2-H sandwiched between two adjacent non-hatched regions 640-1, 640-2, such that either of voltage signals Vi, V₂ output from the non-hatched regions 640-1, 640-2 respectively varies with the position of the reference spectral component λ_c in a nearly linear fashion. In this scenario, a single channel-sensing element, such as that associated with either of the non- hatched regions 640-1, 640-2, may also be used as the reference-position-sensing element.

The spectral power monitoring apparatus 600 of FIG. 6A may be configured such that the spectral channels impinge onto the non-hatched regions of the photodiode array 640A in a one-to-one correspondence; and the spectral spots formed by the spectral channels are confined within the respective non-hatched regions, as illustrated in FIG. 6B. By way of example, the non-hatched region 640-i may be designated for the spectral channel λ„- whereas the non-hatched region 640-j may be assigned to the spectral channel λ_j. In this way, the voltage signals output from the non-hatched regions are proportional to the power levels of their corresponding spectral channels respectively, since in each non-hatched region (e.g., the non-hatched region 640-i) only one photo-response function (such as Ri(x)) is in control. For instance, the voltage signal V; is directly proportional to the power level of the spectral channel λi impinging onto the non-hatched region 640-i, and the associated proportionality factor may be obtained from a calibration, as to be described in further detail later. Such a configuration also takes advantage of the uniform photo-response characteristics in the non- hatched regions, rendering any shift in the impinging position of a spectral channel within the corresponding non-hatched region practically inconsequential. Furthermore, the processing element 660-2 in the embodiment of FIG. 6A may employ a suitable differential detection scheme known in the art to measure the aforementioned voltage signals Vi, V₂, such that the deviation of the actual impinging position of the reference spectral component λ_c from the prescribed location can be readily monitored. The processing element 660-2 may in turn use the detected deviation in the impinging position of the reference spectral component λ_c to generate appropriate control signals to be applied to the actuation device 660-1, so as to maintain the reference spectral component λ_c at the prescribed location and thereby ensure the requisite alignment between the spectral channels and the corresponding channel-sensing elements. As such, the embodiment of FIG. 6B provides one embodiment of the optical- sensor array 640 in FIG. 6A.

In certain situations, it might be difficult to confine the spectral spots of the spectral channels within the non-hatched regions in the corresponding channel-sensing elements (e.g., in the manner illustrated in FIG. 6B). The spectral array formed by the diffracted optical beams may also have a non-uniform pitch, meaning that the spacing between any two adjacent spectral spots may not be constant. Either scenario may result in a situation where one or more channel-sensing elements each receive more than one spectral channel, and in some instances, the spectral spots may overlap. FIG. 2C depicts an exemplary embodiment of how a photodiode array as described in FIG. 6B might be utilized in such applications. By way of example, the photodiode array 640B shown in FIG. 2C may be substantially similar to the photodiode array 640A of FIG. 6B in configuration and operation, hence the elements are labeled with identical numerals. For purpose of illustration and clarity, only three spectral channels λj, λj, λk are explicitly indicated; and the spectral channels are shown to be positioned such that one or more channel-sensing elements may each receive more than one spectral channel. For example, the channel-sensing element associated with the non- hatched region 640-i receives, at least, the spectral channels λi, λ_j; likewise, the channel- sensing element associated with the non-hatched region 640-j receives, at least, the spectral channels λj, λk. Based on Eq. (1), the voltage signal Vi output from the non-hatched region 640-i can be generally expressed as

K^ i, (*)[£/„ (*,y)]ώ«fy (2)

where I„(x,y) is the light intensity associated with the spectral channel λ_n (n = 1 through N) in the region of interest. One skilled in the art will appreciate that Eq.(2) is applicable to any spectral channel of interest (that is, i = 1 through N in the above). Hence, if the power levels Pi through P_N of the spectral channels impinging upon the photodiode array 640B are represented by an optical power vector (P), and the voltage signals Vi through V_M (M &gt; N) thus generated from the photodiode array 640B are represented by a voltage vector (V), (P) and (V) may be related as follows:

(V) = [T](P) (2) where [T] is an (MxN) transfer-matrix. The transfer-matrix [T] generally depends upon the relative alignment between the spectral channels and the underlying channel-sensing elements, as well as the intrinsic properties (such as the photo-response characteristics) of the photodiode array employed. The transfer-matrix [T] is typically band-diagonal, unless one or more channel-sensing elements each receive multiple spectral channels. Those skilled in the art will recognize that Eq. (2) above is also applicable to the embodiment of FIG. 6B, where N = M and the transfer-matrix [T] is truly diagonal.

Based on Eq. (2) above, it follows that

(P)=[C](V) (3) where [C] is an (NxM) conversion matrix and can be derived from the transfer-matrix [T] in

Eq. (2). To determine the transfer-matrix [T], a calibration may be performed (e.g., at the factory), where calibration optical signals characterized by substantially the same wavelengths as the spectral channels to be detected and having known power levels are coupled into the input port 610 in FIG. 6A, thereby traversing substantially the same optical paths as the spectral channels would undertake. (The calibration optical signals may be provided by a tunable laser, for instance.) The output voltage signals of the photodiode array 640B in response to the incident calibration optical signals are then measured. By substituting the measured voltage signals and the known power levels of the calibration optical signals into Eq. (2), the transfer-matrix [T] can be computed. The conversion matrix [C] in Eq. (3) can be further derived from the transfer-matrix [T] by means of a suitable matrix computation algorithm known in the art. The conversion matrix [C] thus obtained may be stored in a system memory, e.g., in a signal processor that may be part of the processing element 660-2 in FIG. 6A. Subsequently in the course of operation, the conversion matrix [C] stays, substantially unchanged, so long as the spectral channels remain impinging at substantially the same locations on the photodiode array 640B as did the calibration optical signals. The requisite alignment may be maintained by the aforementioned servo-based alignment compensation unit 660 in FIG. 6A, for example. This enables the signal processor to readily compute the power levels of the spectral channels impinging upon the photodiode array 640B from the voltage signals thus generated, in a manner according to Eq. (3). One skilled in the art will recognize that if so desired in a practical application, any background contribution (e.g., due to the "dark current" of the photodiode array and/or to "stray light" from the environment) can be independently determined, and subsequently taken into account in the calibration and operation processes described above.

In the embodiment of FIG. 2C, the impinging position of the reference spectral component λ_c may also be monitored by measuring the voltage signals Vi, Vs output respectively from the non-hatched regions 640-1, 640-2 using an appropriate normalized differential detection scheme, for example, by way of detecting the position error signal (Vι-N2)/(Vι+V₂) in a manner as described with respect to FIG. 6B. As such, the embodiment of FIG. 2C may be alternatively implemented in FIG. 6A to embody the optical-sensing array 640.

Referring back to the embodiment of FIG. 6A. The actuation device 660-1 may be a stepping motor, a solenoid actuator, a piezoelectric actuator, a voice coil actuator, or other types of actuation means known in the art. The processing element 660-2 may include electrical circuits, controllers and signal processing algorithms for processing the output signals received from the reference-position-sensing element 640-C (e.g., the voltage signals Vi, V₂ output from the optical-sensing array 640A in FIG. 6B) and deriving from the detected signals the real-time impinging position of the reference spectral component λ_c. The processing element 660-2 accordingly generates appropriate control signals to be applied to the actuation device 660-1, so as to adjust the alignment of the reference spectral component λ_c along with the spectral channels λi through λ_N in such a way that the reference spectral component λ_c is maintained at the predetermined location x₀. The electronic circuitry and the associated signal processing algorithm/software for a processing element in a servo-control system are known in the art of electrical engineering and servo control systems.

Those skilled in the art will appreciate that instead of (or in conjunction with) moving the optical-sensing array 640 as described above, the focusing lens 630 in FIG. 6A may be alternatively (or additionally) moved, e.g., translated or rotated, thereby controlling the impinging locations of the diffracted optical beams and performing a similar alignment function. The translation/rotation of the focusing lens 630 may be accomplished by coupling to it an appropriate actuation device as described above. In some instances, the alignment adjustment may also be brought about (or complemented) by moderate variation in the incidence angle of the input multi- wavelength optical signal (along with the reference signal) upon the diffraction grating 620, e.g., by way of rotating the grating or placing a dynamically adjustable mirror between the input port 610 and the diffraction grating 620, so long as such an adjustment does not substantially alter the pitch of the spectral array formed by the diffracted optical beams. As will be appreciated from the teachings of this specification, one skilled in the art would know how to devise an appropriate alignment-adjusting element and corresponding processing element for a servo-based alignment compensation unit according to the present invention, to best suit a given application.

FIG. 7A depicts a second embodiment of a spectral power monitoring apparatus according to the present invention. By way of example, spectral power monitoring apparatus 700 may make use of the architecture as well as a number of the elements employed in the embodiment of FIG. 6A, as indicated by those labeled with identical numerals. Notice that there is no "moving" alignment-adjusting means employed in this system. Instead, a software-based alignment compensation unit 760 is implemented, which may be a signal processor in communication with the optical-sensing array 640.

FIG. 7B shows in further detail how the optical-sensing array 640 in the embodiment of FIG. 7A may be configured. By way of example, the photodiode array 640C of FIG. 7B may be substantially similar to the photodiode array 640A depicted in FIG. 6B in configuration and operation, hence the elements are labeled with identical numerals. In this case, the impinging position of the reference spectral component λ_c may be monitored by utilizing two or more adjacent channel-sensing elements, occupying a contiguous segment (termed "reference segment" herein) of the photodiode array 640C. For instance, if the reference spectral component λ_c is located within two adjacent channel-sensing elements, such as those associated the non-hatched regions 640-1, 640-2 as illustrated in FIG. 7B, the voltage signals Vi, V₂ output respectively from the non-hatched regions 640-1, 640-2 may be measured using an appropriate normalized differential detection scheme, for example, by way of detecting the position error signal (Vι-V₂)/(Vι+V₂) in a manner as described with respect to FIG. 6B. In the event that the reference spectral component λ_c may undergo a shift in alignment or a "walk-off (e.g., during the calibration or operation to be described below) that extends beyond two channel-sensing elements, the reference segment may contain multiple channel-sensing elements, and their respective output voltage signals are detected. A series of position error signals may be accordingly generated, each associated with the voltage signals output from two adjacent channel-sensing elements in this segment in a manner as described above, from which the impinging position of the reference spectral component λ_c may be deduced. Those skilled in the art will know how to devise an appropriate signal detection and processing scheme, for effectively monitoring the real-time impinging position of the reference spectral component λ_c.

In FIG. 7B, the spectral channels λi through λ_N may impinge onto a plurality of channel- sensing elements located within a section (termed "channel section" herein) of the photodiode array 640C that is well separated from the reference segment designated from the reference spectral component λ_c, so that any voltage signal output from the reference segment does not include contribution from the spectral channels, or vice versa. The correspondence between the spectral channels and the underlying channel-sensing elements may be as depicted in the embodiment of FIG. 6B or 6C. In either scenario, the power levels of the spectral channels impinging onto the photodiode array 640C can be related to the voltage signals thus generated from the channel section of the photodiode array 640C by a predetermined conversion matrix [C], in the manner as indicated in Eq. (3) above.

The spectral power monitoring apparatus 700 in the embodiment of FIG. 7A may be operated as follows. During an initial (or factory) calibration, optical signals having the same wavelengths as the spectral channels to be detected and known power levels are coupled into the input port 610, in a way to "mimic" the spectral channels of interest. (The calibration optical signals may be provided by a tunable laser, for instance.) The calibration optical signals along with the reference signal λ_c emerge from the input port 610, are spatially separated and subsequently impinge onto the photodiode array 640 (e.g., the photodiode array 640C of FIG. 7B) by way of the diffraction grating 620 and the focusing lens 630. The impinging location of the reference spectral component λ_c is then varied incrementally (e.g., along the x-direction shown in FIG. 7B) with sufficient spatial resolution, which may be accomplished by translating the optical-sensing array 640 using a suitable actuation means. The position Λ: of the reference spectral component λ_c is then determined by the voltage signals (e.g., the voltage signal Vi, V2) output from the designated reference section as described above. At each position x of the reference spectral component λ_c, the voltage signals produced by the calibration optical signals are also measured, which are then substituted in Eq. (2) above, along with the known power levels of the calibration optical signals, whereby the corresponding transfer-matrix [T(x)] is computed. The conversion matrix [C(x)] in Eq. (3) may in turn be derived from the transfer-matrix T( ) by using a suitable matrix computation algorithm known in the art. This calibration process thus establishes a matrix-calibration table, containing [C( :)] as a function of x, which may be stored in the alignment compensation unit 760. Subsequently in the course of operation, the alignment compensation unit 760 monitors the real-time impinging position j of the reference spectral component λ_c, and measures the voltage signals produced by the spectral channels impinging onto the channel section of the photodiode array 640C at the corresponding position x. The alignment compensation unit 760 then looks up a corresponding conversion matrix [C(Λ:)] from the predetermined matrix-calibration table, in order to obtain the power levels of the impinging spectral channels from the measured voltage signals by use of Eq. (3) above. As such, the spectral power apparatus 700 effectively compensates for any shift in the alignment that may arise over the course of operation by way of software control, without involving any "moving" actuation means. The employment of such a software-based alignment compensation unit also relaxes fabrication tolerances and precision during initial assembly, enabling the spectral power monitoring apparatus of the present invention to be simpler in construction and more robust in performance.

It should be understood that the exemplary photo-response characteristics of the photodiode array 640A as described in FIG. 6B are provided by way of example, to illustrate the general principles of the present invention. Those skilled in the art will appreciate that other optical power sensor (or photodiode) arrays with different photo-response characteristics may be alternatively implemented in a spectral power monitoring apparatus of the present invention to provide substantially the same function in a substantially equivalent manner (e.g., as described by Eqs. (l)-(3) above). For instance, an optical power sensor array in the present invention need not necessarily possess a continuous overall photo-response function (e.g., there may be one or more "dead zones" between the photo-sensing regions). As will be appreciated from the teachings of the present invention, a skilled artisan will know to design an appropriate spectral power monitoring apparatus, to best suit a given application.

In the embodiment of FIG. 6A or 7A, the multi-wavelength optical signal containing wavelengths λi through λ_N may be provided by an input optical fiber 601 coupled to the fiber collimator serving as the input port 610. The reference signal λ_c may be provided by a reference light source 602, which may be a distributed feedback (DFB) laser, a Fabry-Perot (FP) laser (in conjunction with an appropriate modulation/control system that suppresses the mode hopping and stabilizes the output signal), or any other light source known in the art that can provide an appropriate reference signal with well-defined and stable wavelength. An optical combiner 603 (e.g., a fused fiber-optic coupler) may be used to couple the reference light source 602 to the input optical fiber 601, effective to couple both the multi-wavelength optical signal and the reference signal into the input port 610. The spectral power monitoring apparatus thus has an independent, internal reference light source. Alternatively, the input multi-wavelength optical signal itself may include a spectral component (e.g., a service channel in an optical networking application) that can be used as the reference signal, as in WDM optical networking applications. In such a scenario, the internal reference light source 602 along with the fiber-optic coupler 603 need not be implemented. The input optical fiber 601 may be a single mode fiber, multi-mode fiber, or polarization maintaining fiber. Moreover, the diffraction grating 620 may be a ruled diffraction grating, a holographic diffraction grating, or an echelle grating, all commonly employed in the art for separating a multi-wavelength signal by wavelength. In general, the wavelength-disperser in a spectral power monitoring apparatus of the present invention may also be embodied by other types of wavelength-separating means known in the art, such as a transmission diffraction grating or a dispersing prism. The beam-focuser 630 may alternatively be an assembly of focusing lenses, or any other suitable beam-focusing means known in the art. The focusing function may also be provided by using a curved diffraction grating that performs a dual function of wavelength-separating and beam-focusing. It should be noted that in applications where the spectral channels along with the reference spectral component are well separated, the beam- focuser, such as the focusing lens 630 in FIG. 6A or 7A, need not be utilized.

It is known that the diffraction efficiency of a diffraction grating may be polarization- dependent. For instance, the diffraction efficiency of a grating in a standard mounting configuration may be higher for p (or TM)-polarization that is peφendicular to the groove lines on the grating than for s (or TE)-polarization that is orthogonal to p-polarization, or vice versa. To mitigate such polarization-sensitive effects, a suitable polarization-sensitive element may be implemented in a spectral power monitoring apparatus of the present invention, serving to attenuate one polarization (e.g., p-polarization) component relative to the other polarization (e.g., s-polarization) component in the input multi-wavelength optical signal according to a predetermined ratio prior to impinging onto the diffraction grating, so as to compensate for the polarization dependence of the grating. This may be accomplished, for example, by placing an appropriate weak polarizer (e.g., a leaky beam-splitter) along the optical path between the input port 610 and the diffraction grating 620 in the embodiment of

FIG. 6A or 7A. Alternatively, a suitable polarization diversity scheme may be implemented as discussed below in Section III.

As will be appreciated from the teachings of the present invention, one skilled in the art will know how to design a spectral power monitoring apparatus employing a suitable alignment compensation unit, along with an appropriate polarization diversity scheme, to best suit a given application. For instance, by employing a photodiode array that is InGaAs based (which is particularly sensitive in the wavelength range of 1 - 1.7 μm) as the optical-sensing array in the above embodiments, the present invention provides a new line of spectral power monitors with active alignment compensation that would be particularly suitable for WDM optical networking applications.

III. POLARIZATION DIVERSITY SCHEME

FIG. 8 shows an exemplary embodiment of an optical spectral power monitoring apparatus of the present invention. By way of example to illustrate the principles and the general architecture of the present invention, optical spectral power monitoring apparatus 800 comprises an input port 810 for a multi-wavelength optical signal which may be in the form of a fiber collimator; a polarization-separating element 870 which in one form may be a polarizing beam splitter; a polarization-rotating element 880 which may be a half-wave plate; a wavelength-disperser 820 which in one form may be a diffraction grating; a beam-focuser 830 which may be a focusing lens; and an array of optical power sensors 840 (termed "optical-sensing array," herein).

The principal operation of the optical spectral power monitoring apparatus 800 of FIG. 8 may be as follows. The input port 810 transmits a multi-wavelength optical signal (which may contain wavelengths λi through λ_N, for instance). The polarization-separating element 870 decomposes the multi-wavelength optical signal into a p (or TM) polarization component (perpendicular to the groove lines on the grating) and an s (or TE) polarization component (orthogonal to the p-polarization component) with respect to the diffraction grating 820. (The p-polarization and s-polarization components may also be termed "first and second polarization components.") As a way of example, assuming that p-polarization is the "preferred direction" of the diffraction grating 820 (i.e., the diffraction efficiency is higher for p-polarization than for s-polarization), the polarization-rotating element 880 subsequently rotates the polarization of the s-polarization component (or the second polarization component) by 90-degrees, whereby the optical signals incident upon the diffraction grating 820 all possess p-polarization. The diffraction grating 820 angularly separates the incident optical signals by wavelength respectively into first and second sets of optical beams (wherein each set contains optical beams with wavelengths λi through λ , for instance). The focusing lens 830 may subsequently focus the diffracted optical beams into corresponding focused spots, such that the first and second optical beams associated with the same wavelength (e.g., λj) impinge at substantially the same location (or within the same optical power sensor) on the optical-sensing array 840. (It should be appreciated that in this specification and appending claims, the rotation in polarization produced by a polarization- rotating element (e.g., the polarization-rotating element 880) may be construed as having slight variations about a prescribed angle (e.g., 90-degrees), due to imperfections that may exist in a practical system. Such variations, however, will not significantly affect the overall performance of the invention.)

The aforementioned overlap of the first and second optical beams (polarized in the same direction and characterized by the same wavelength) may give rise to coherent interference that may produce undesirable intensity fringes. To avoid such a situation, an auxiliary polarization-rotating element 890 may be implemented in the embodiment of FIG. 8, whereby the first and second sets of optical beams are polarized in two orthogonal directions prior to impinging onto the optical-sensing array 840. The auxiliary polarization-rotating element may be implemented between the diffraction grating 820 and the optical-sensing array 840, and serves to rotate the polarization of either the first or second set of optical beams by 90-degrees prior to impinging onto the optical-sensing array 840. By way of example in FIG. 8, an auxiliary polarization-rotating element 890 may be disposed between the diffraction grating 820 and the focusing lens 830, such that the first set of optical beams undergoes a 90-degree rotation in polarization prior to impinging onto the optical-sensing array 840. It should be appreciated that the auxiliary polarization-rotating element 890 may alternatively be disposed between the diffraction grating 820 and the focusing lens 830 in such a way that the second set of optical beams undergoes a 90-degree rotation in polarization prior to impinging onto the optical-sensing array 840. In either scenario, the first and second optical beams associated with the same wavelength (e.g., λj) become polarized in two orthogonal directions upon impinging onto the optical-sensing array 840, thereby eliminating any coherent intensity interference.

As such, by advantageously employing the aforementioned polarization diversity scheme, the polarization sensitivity of the diffraction grating 820 becomes inconsequential in the optical spectral power monitoring apparatus 800. This enables the apparatus of the present invention to enhance spectral resolution in a simple and cost-effective construction (e.g., by making use of high-dispersion diffraction gratings commonly available in the art), while providing improved accuracy in optical spectral power detection.

FIGS. 9 depicts another embodiment of an optical spectral power monitoring apparatus employing a polarization diversity scheme, according to the present invention. By way of example, optical spectral power monitoring apparatus 900 may make use of the general architecture along with a number of the elements employed in the embodiment of FIG. 8, as indicated by those elements labeled with identical numerals. In this case, a modulation assembly 985 may be implemented, and configured such that the first and second sets of optical beams impinge onto the optical-sensing array 840 in a time-division-multiplexed (e.g., alternating) fashion. By way of example, the modulation assembly 985 is shown to be in the form of first and second shutter-elements 981, 982 along with a control unit 983, disposed along the optical path between the polarization-separating element 870 along with the polarization-rotating element 880 and the diffraction grating 820, thereby controlling the first and second polarization components, respectively. Either of the first and second shutter- elements 981, 982 may be configured such that it permits an optical signal to pass through under an appropriate control signal (e.g., provided by the control unit 983); and stays closed to the incident optical signal in the absence of any control signal. Hence, by operating the first and second shutter-elements 981, 982 in an alternating manner according to a suitable control scheme by way of the control unit 983, the first and second sets of optical beams impinge onto the optical-sensing array 840 in a time-division-multiplexed sequence, as illustrated by solid and dashed lines in phantom. This enables the first and second sets of optical beams to be separately detected, whereby an optical power spectrum associated with each polarization component in the input multi-wavelength optical signal can be independently derived. One skilled in the art will appreciate that the first and second shutter- elements 981, 982 (along with the control unit 983) may alternatively be implemented between the diffraction grating 820 and the optical-sensing array 840, thereby providing a substantially similar function by controlling the first and second sets of optical beams, respectively.

In the aforementioned embodiment, the modulation assembly 985 may alternatively be provided by an optical beam-chopper (along with associated control unit), such as an opaque rotating disk equipped with at least one aperture, or any other suitable means known in the art that allows two incident optical signals to pass through in an alternating fashion. The optical beam-chopper may be implemented along the optical path between the polarization- separating element 870 along with the polarization-rotating element 880 and the diffraction grating 820, or between the diffraction grating 820 and the optical-sensing array 840, thus providing a substantially similar function in a substantially equivalent manner.

In the embodiment of FIG. 9, the first and second sets of optical beams may each have a predetermined alignment with the underlying optical-sensing array 840. Alternatively, the first and second optical beams associated with the same wavelength (e.g., λi) may impinge at substantially the same location (albeit at different times) on the optical-sensing array 840. The optical-sensing array 840 may be substantially identical to arrays 240 or 640, and may comprise a photodiode array (e.g., a photodiode array from Sensors Unlimited, Inc., Princeton, New Jersey), or other suitable optical power sensing means known in the art. A skilled artisan will know how to implement a suitable optical-sensing array and devise an appropriate detection scheme, to best suit a given application.

Like the embodiment of FIG. 8, the optical spectral power monitoring apparatus 900 is polarization insensitive with respect to the diffraction grating 820, and is therefore capable of providing accurate detection of the multi-wavelength optical signal with enhanced spectral resolution. An additional advantage of the optical spectral power monitoring apparatus 900 is that by impinging the first and second sets of optical beams onto the optical-sensing array in a time-division-multiplexed fashion, an optical power spectrum associated with each polarization component in the input multi-wavelength optical signal can be independently determined, which would be useful in optical networking applications. For instance, polarization multiplexing - the encoding of data streams onto two orthogonal polarization components of a single wavelength channel - has emerged as another way of increasing the information capacity of an optical fiber. Hence, it would be desirable to have a device that can separately detect two orthogonal polarization components of a single wavelength channel.

Those skilled in the art will recognize that the aforementioned function of the modulation assembly 985 may be generalized to modulate the first and second sets of optical beams in a frequency-division-multiplexed fashion, whereby they can be separately identified on the optical-sensing array 840. FIG. 10 shows another embodiment of an optical spectral power monitoring apparatus, pertaining to this situation. By way of example, optical spectral power monitoring apparatus 1000 may make use of the general architecture along with a number of the elements employed in the embodiment of FIG. 8, as indicated by those elements labeled with identical numerals. In this case, a modulation assembly 1085 may be disposed along the optical path between the polarization-separating element 870 along with the polarization- rotating element 880 and the diffraction grating 820, serving to modulate the first and second polarization components, respectively. The modulation assembly 1085 may be in the form of first and second modulating elements 1081, 1082 which may be electro-optic intensity modulators (e.g., liquid crystal based intensity modulators) known in the art, along with a control unit 1083. The first and second modulating elements 1081, 1082 may operate under control of two distinct alternating (or "dither") control signals (e.g., provided by the control unit 1083), which in one form may be sinusoidal functions of time at two distinct frequencies (e.g., first and second "dither frequencies"). Either of the modulating elements 1081, 1082 may be configured to introduce a "dither modulation signal" in the optical power level of its corresponding optical beam that includes a substantially linear response to the control signal to which it is subject. As such, upon emerging from the first and second beam-modulating elements 1081, 1082, the first and second polarization components may carry first and second dither modulation signals (e.g., characterized by the first and second dither frequencies), respectively. Consequently, the first and second sets of optical beams diffracted from the diffraction grating 820 also carry the respective dither modulation signals, impinging onto the optical-sensing array 840. The electrical signals thus generated by the optical-sensing array 840 likewise contain the same characteristic dither modulation signals, which may be detected by a synchronous detection unit 1090 in communication with the optical-sensing array 840. As will be appreciated by those skilled in the art, the synchronous detection unit 1090 may also be in communication with the control unit 1083, if so desired in a practical application.

In the present invention, a "spectral channel" is characterized by a distinct center wavelength and associated bandwidth, and may carry a unique information signal as in WDM optical networking applications. A "dither modulation signal" refers to any modulation in the optical power level of an optical signal produced by the modulation assembly, in contrast with other "intrinsic" modulation signals (e.g., information signals) the input multi-wavelength optical signal may carry. Accordingly, the dither modulation signals are allocated in a spectral range that is sufficiently separated from the frequencies of other "intrinsic" modulation signals the spectral channels may carry.

As in the case of FIG. 9, the first and second sets of optical beams in the embodiment of FIG. 10 may each have a predetermined alignment with the underlying optical-sensing array 840. Alternatively, the first and second optical beams associated with the same wavelength (e.g., λi) may impinge at substantially the same location (or within the same optical power sensor) on the optical-sensing array 840. In either scenario, the distinct dither modulation signals carried by these two sets of optical beams enable them to separately detected, e.g., by use of the synchronous detection unit 1090. To relate the measurements provided by the synchronous detection unit 1090 to the corresponding optical power levels in the input multi- wavelength optical signal, a calibration process may be undertaken, whereby an optical power spectrum associated with each polarization component in the input multi-wavelength optical signal can be derived. From the teachings of the present invention, those skilled in the art will know how to implement a suitable optical-sensing array and devise an appropriate detection scheme, to best suit a given application.

The modulation assembly 1085 may also be provided by an optical beam-chopper (along with an associated control unit), e.g., an opaque rotating disk equipped with two groups of apertures. Each group of apertures effectively "chops" its corresponding optical beam (e.g., the first or second polarization component) at a frequency determined by the spatial arrangement of its constituent apertures. By arranging the two groups of apertures according to a desired scheme, the first and second sets of optical beams arriving at the optical-sensing array 840 may be characterized by distinct modulations, thereby enabling them to be separately detected. It should be appreciated that the modulation assembly 1085 (e.g., the first and second modulation elements 1081, 1082) may be alternatively implemented between the diffraction grating 820 and the optical-sensing array 840, so as to modulate the first and second sets of optical beams, respectively. As will be appreciated from the teachings of the present invention, those skilled in the art will know how to implement an appropriate modulation assembly in an optical spectral power monitoring apparatus according to the present invention, to best suit a given application. In the above embodiments, the polarization-separating element 870 may be a polarizing beam splitter, a birefringent beam displacer, or other types of polarization-separating means known in the art. The polarization-rotating element 880, or the auxiliary polarization-rotating element 890, may be a half-wave plate, a Faraday rotator, a liquid crystal rotator, or any other polarization-rotating means known in the art that is capable of rotating the polarization of an optical beam by a prescribed degree (e.g., 90-degrees). Either of the first and second shutter elements 981, 982 may be a liquid crystal based shutter element, e.g., comprising a liquid crystal rotator that rotates the polarization of an incident optical beam by 90-degrees in the absence of any control signal and leaves the polarization unchanged under an appropriate control signal, in conjunction with a polarizer whose polarization axis is perpendicular to the thus-rotated polarization produced by the liquid crystal rotator. Either of the first or second shutter elements 981, 982 may also be an MEMS (micro-electro-mechanical-systems) based element that acts as a mechanical shutter, or any other shutter-like element known in the art that opens, or remains closed, to an incident optical beam by way of a suitable actuation means. The control unit 983 may include electrical circuits and signal control algorithms known in the art, for controlling the first or second shutter elements 981, 982 according to a desired scheme.

Moreover, either of the first and second modulation elements 1081, 1082 may be an electro- optic intensity modulator, such as a liquid crystal intensity modulator, or any other suitable modulation means known in the art. A skilled artisan will know how to devise an appropriate control unit 1083, such that desired dither modulation signals are produced by the first and second modulating elements 1081, 1082. The synchronous detection unit 1090 generally comprises electrical circuits and signal processing algorithms devised for performing synchronous detection of the dither modulation signals thus produced in the first and second sets of optical beams, respectively.

In the present invention, the wavelength disperser (e.g., the diffraction grating) 820 may be a ruled diffraction grating, a holographic diffraction grating, or an echelle grating, all commonly employed in the art for separating a multi-wavelength signal by wavelengths. In general, the wavelength-disperser 820 in an optical spectral power monitoring apparatus of the present invention may also be embodied by other types of wavelength-separating means known in the art, such as a transmission diffraction grating or a dispersing prism. The beam- focuser 830 may alternatively be an assembly of focusing lenses, or any other suitable beam- focusing means known in the art. The focusing function may also be accomplished by a curved diffraction grating that serves a dual function of wavelength separating and beam focusing. The fiber collimator serving as the input port 810 may be in the form of a coUimating lens (such as a GRIN lens) and a ferrule-mounted fiber packaged together in a mechanically rigid stainless steel (or glass) tube.

It should be appreciated that the foregoing polarization diversity compensating schemes in combination with the spectral monitors described in sections I and II, which also employ active alignment compensation. For example and without limitation, the polarization diversity schemes can be. employed within the spectral monitors by integrating polarization separating element 870, polarization rotating element 880 and/or 890, and modulation assemblies 985, 1085 in corresponding locations within the spectral monitors (e.g., between the input ports and the wavelength dispersers and/or the beam focusers).

Although the present invention and its advantages have been described in detail, it should be understood that various changes, substitutions, and alternations can be made herein without departing from the principle and the scope of the invention. Accordingly, the scope of the present invention should be determined by the following claims and their legal equivalents.