Chirped optical MEM device

Description

FIELD OF THE INVENTION

The present invention relates to the field of grating light valves arranged in an array of pixels. More particularly, the present invention relates to chirping the widths of various grating light valve components in successive pixels to compensate for unwanted phenomena in a demultiplexed wavelength division multiplexed “WDM” signal.

BACKGROUND OF THE INVENTION

Optical MEM (micro-electro-mechanical) devices have applications in display, print, optical and electrical technologies. An optical MEM device, herein, is a device that is capable of constructively and destructively interfering with an incident light source to produce an optical signal or optical signals. Exemplary optical MEM devices, and methods for making the same, are disclosed in U.S. Pat. No. 5,311,360 entitled “METHOD AND APPARATUS FOR MODULATING A LIGHT BEAM,” U.S. Pat. No. 5,841,579 entitled “FLAT DIFFRACTION GRATING LIGHT VALVE” and U.S. Pat. No. 5,808,797 entitled “METHOD AND APPARATUS FOR MODULATING A LIGHT BEAM” to Bloom et al. and U.S. Pat. No. 5,661,592 entitled “METHOD OF MAKING AN APPARATUS FOR A FLAT DIFFRACTION GRATING LIGHT VALVE” to Bornstein et al., the contents of which are hereby incorporated by reference.

Optical MEM devices can be fabricated from Si-based materials using lithographic techniques. Optical MEM devices can have reflective ribbons that are formed over a suitable substrate structure such that the ribbons are spatially arranged in parallel and are coupled to the substrate structure. In use, a portion of the reflective ribbons can be moved by applying an operating bias voltage, or switching voltage, across the ribbons and the substrate structure. By alternating, or switching, the potential of the bias voltage, the ribbons are alternated between the positions for constructive and destructive interference with the incident light source to generate optical signals.

Variations in the characteristics of optical signals can occur for a variety of reasons including varying polarization of the light, demultiplexing through static grating demultiplexing, light loss through spaces between the ribbons, etc. Variability in the characteristics of optical signals can be detrimental in applications where consistent optical signal characteristics are desired across the wavelength band and over time. Therefore, there is a need for an optical MEM device that is capable of producing stable optical signals exhibiting uniform optical parameters during operation.

SUMMARY OF THE INVENTION

According to embodiments of the present invention, a method of and apparatus for improving the alignment between chirped incident wavelengths of light and the center line of their respective pixels when the WDM signal is demultiplexed through a static grating demultiplexer is disclosed. The present invention is also a method of and apparatus for compensating for the variation of polarization dependent loss among varying wavelengths induced by interaction with a pixel of a grating light valve™ light modulator. According to another aspect of the embodiments, a method of and apparatus for compensating for the variation in the angle of diffraction that occurs in the diffraction patterns of varying wavelengths falling on pixels of a grating light valve™ light modulator is disclosed.

A grating light valve™ light modulator comprises a plurality of ribbons separated by a plurality of gaps. The plurality of ribbons and gaps are grouped into a plurality of pixels including a first end pixel and a second end pixel. Each pixel receives a de-multiplexed signal and the plurality of pixels receive in succession the de-multiplexed signals of increasing wavelengths such that the first end pixel receives the shortest wavelength and the second end pixel receives the longest wavelength. A pixel separation distance is defined as a distance from a center line of a first pixel to a center line of a second pixel adjacent to the first pixel. Pixel separation distances are progressively increased from the first end pixel to the second end pixel. The pixel separation distances are spaced such that a pixel center-line more closely coincides with a center-line intensity of an incident wavelength.

A method of normalizing polarization dependent loss among a plurality of pixels in a grating light valve™ light modulator against a first pixel comprises the steps of configuring a width W_R(n) of a ribbon in the n^th pixel as a function of a width W_R(ref) of a ribbon in a reference pixel according to the equation W_R(n)=W_R(ref)·(λ_n/λ_ref), and configuring a width W_RG(n) of a ribbon gap in the n^th pixel as a function of a width W_RG(ref) of a ribbon gap in the reference pixel according to the equation W_RG(n)=W_RG(ref)·(λ_n/λ_ref), wherein λ_ref is a wavelength designated for the reference pixel and λ_n is a wavelength designated for the n^th pixel, and wherein each pixel comprises at least two ribbons and each ribbon within a pixel is separated from an adjacent ribbon within the same pixel by a ribbon gap.

A method of normalizing diffraction angles in a sequence of pixels in a grating light valve™ light modulator against a first pixel comprises the step of configuring a width W_R2 and W_G2 of a ribbon in the second pixel as a function of a ribbon width W_R1 and gap width W_G1 of a ribbon in the first pixel such that the equation W_R2+W_G2=(λ₂/λ₁·(W_R1+W_G1)) holds, wherein λ₁ is a wavelength designated for the first pixel and λ₂ is a wavelength designated for the second pixel, and wherein each pixel comprises at least two ribbons, each ribbon within a pixel separated by a ribbon gap.

BRIEF DESCRIPTION OF SEVERAL VIEWS OF THE DRAWINGS

FIG. 1 is a perspective view illustrating of a grating light valve™ light modulator.

FIG. 2 is a cross sectional view of a grating light valve™ light modulator comprising a row of undeflected ribbons.

FIG. 3 is a cross sectional view of a grating light valve™ light modulator comprising a row of alternating deflected and undeflected ribbons.

FIG. 4 is an illustration of a WDM signal on a static grating and component wavelengths are directed onto pixels of a grating light valve™ light modulator.

FIG. 5 is a close up illustration of incident wavelength center lines aligned with a pixel center of a corresponding pixel on the grating light type device of FIG. 4.

FIG. 6 illustrates an array of pixels showing various dimensions of pixels and pixel spacing.

FIG. 7 illustrates various dimensions of component parts of a pixel within a grating light valve™ light modulator.

FIG. 8 illustrates a top view of a grating light valve™ light modulator comprised of ribbons and gaps divided into a plurality of pixels.

FIG. 9 graphically illustrates an exemplary relationship of the channel chirping experienced by de-multiplexed wavelengths.

FIG. 10 graphically illustrates an exemplary representation of channel width as a function of channel number.

FIG. 11 illustrates a WDM signal demultiplexed through a static grating, with component wavelengths directed to chirped pixels configured to compensate for wavelength-based chirping.

FIG. 12 illustrates a succession of adjacent pixels within a grating light valve™ light modulator configured for limited diffraction chirping.

FIG. 13 illustrates multiple pixels in a grating light valve™ light modulator exhibiting chirped intra-pixel gaps, chirped ribbon widths, and chirped inter-pixel gaps, wherein the chirping is performed at independently controllable rates.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Reference will now be made in detail to embodiments of the invention, examples of which are illustrated in the accompanying drawings.

FIG. 1 is a perspective view of a grating light valve™ light modulator 71. Parallel ribbons 72-77 rest above a substrate 84 using any conventional technique. The substrate includes conductive members 78-83 which function to exert a field on their respective ribbons 72-77. A voltage potential formed between a ribbon 72-77 and its respective conductive member 78-83 exerts a force sufficient to deflect a ribbon 72-77.

FIG. 2 illustrates a cross section side view of a grating light valve™ light modulator 71. In FIG. 2, the even ribbons 72, 74, 76 and odd ribbons 73, 75, 77 are at the same height above the substrate 84. Incident light I reflects substantially as a specular mirror when a grating light valve™ light modulator is configured as shown in FIG. 2. The ribbons are illustrated as comprising a reflective portion 34 and a resilient portion 36. It will be understood that both the reflective portion 34 and the resilient portion 36 can be formed of one or multiple layers. The reflective portion 34 reflects the incident light rays I, thereby forming reflected light rays R, and the resilient portion 35 of a movable ribbon 73, 75, 77 resists deflection, and returns the ribbon to its original position after deflection, as illustrated further in FIG. 3. In the embodiment of FIG. 2, the six ribbons 72-77 form a single pixel. In wavelength division multiplexed applications, a single pixel is typically assigned to a single wavelength. The conductive members 78-83 are charged to appropriate voltage levels to deflect the movable ribbon 73, 75, 77 but not the fixed ribbons 72, 74, 76.

FIG. 3 illustrates the grating light valve™ light modulator 71 wherein the fixed ribbons 72, 74, 76 are un-deflected, and the movable ribbons 73, 75, 77 are deflected. As shown in FIG. 3, the deflection distance Δ has been established as a distance of λ/4. FIG. 3 illustrates first order diffraction angles D⁺¹ and D⁻¹ resulting from interference of light reflected from the reflective surfaces 34 of the ribbons 72-77 of the grating light valve™ light modulator 71. As discussed above, according to the embodiment of FIG. 3, only ribbons 73, 75, 77 are deflected. However, in other grating light valve™ light modulator embodiments, all ribbons 72-77 are movable, and the separation distance Δ between adjacent ribbons is controlled by controlling the position of both the first and second set of the ribbons. As represented in FIGS. 2 and 3, the six ribbons 72-77 of the grating light valve™ light modulator 71 form a single pixel, such that the movable ribbons 73, 75, 77 are controlled in unison to reflect or diffract the same incident beam of light I to the same degree. It will be understood that this is illustrative only. A pixel can comprise other numbers of ribbons. A pixel must comprise at least two ribbons to generate a diffractive pattern, but may comprise any number of ribbons greater than two.

FIG. 4 illustrates a static grating 25 which functions to demultiplex the incoming wavelength division multiplexed (WDM) signal S_IN into its component wavelength signals λ₀ through λ_n. The wavelengths are scattered by the static grating 25 such that an angle of dispersion θ varies with wavelength λ. The angle of dispersion θ(λ) is measured from the normal to the static grating 25. The optical axis is defined at a dispersion angle θ_axis. The WDM signal S_in is incident to the static grating 25 at an incidence angle θ_IN. The dispersion angle θ(λ), in turn, determines the displacement y(λ) along the y-axis when the component wavelength intersects the transform lens 26. The transform lens 26 preferably is located a distance F_T away from the static grating 25, wherein F_T is the focal length from the transform lens 26. As a result of positioning the static grating 25 a focal length of the transform lens 26, wavelengths λ₀ through λ_n exiting the transform lens 26 are oriented along paths substantially parallel to the optical axis. As the demultiplexed component wavelengths λ₀ through λ_n pass through the lens 26, they are focused on to the grating light valve™ light modulator 27. The grating light valve™ light modulator 27 comprises a plurality of pixels 28-33. The pixels are arranged linearly along the y-axis, with each pixel 28-33 assigned to a respective incoming wavelength λ₀ through λ_n. As illustrated in FIG. 5, component wavelengths of light exhibit a substantially Gaussian distribution, which is represented by the light intensity distribution I(λ). As a result, there is typically some overlap in wavelengths, with a small component of the intensity distribution I(λ) of pixel 31 falling on the adjacent pixels 30 and 32. It is a design parameter to determine the amount of overlap. Preferably, the intensity distribution I(λ) of each adjacent channel intersects at approximately the same intensity level, which is discussed in greater detail below. For purposes of the discussion herein, however, the incident wavelength is often represented as simply being located at centerline of its respective intensity distribution I(λ).

Returning to FIG. 4, as the incoming signal S_IN is diffracted through the static grating 25, the angle between the optical axis and a component wavelengths λ₀, . . . λ_n, is represented by equation 1,
θ(λ)=sin⁻¹ [(λ/Λ_g)−sin θ_IN]  1.
where Λ_g is the static grating period of the static grating 25. Accordingly, the higher the static grating period of the static grating 25, the lower the spread of the incoming component signals λ₀ through λ_n. As discussed above, the static grating 25 is typically positioned a distance F_T from the transform lens 26, wherein F_T is the focal length of the transform lense 26. The linear displacement y(λ) from the optical axis for each component wavelength λ₀, . . . λ_n., is therefore a function of the angle θ(λ) and of the transform length F_T, as illustrated by equation 2,
y(λ)=F_T tan [θ(λ)−θ_axis]  2.
where the “transform length” F_T is the focal length of the transform lens 26 along a path parallel to the optical axis. Equation 2 defines the center position of an incident wavelength λ, as measured from the optical axis, as it impinges the grating light valve™ light modulator 27. The linear displacement of a wavelength along the y-axis is influenced by the wavelength. As the wavelength λ increases so too does the distance between center positions of adjacent wavelengths. If the wavelength increases from λ₀ to λ_n, then the physical distance separating two consecutive de-multiplexed wavelengths along the y-axis is greater from λ_n-1 to λ_n than from λ₀ to λ₁. Embodiments of the present invention include pixels positioned on the grating light valve™ light modulator 27 such that a center line of a pixel P_λ approximately aligns with the center position of a corresponding incident wavelength λ. Such center-line alignment is herein referred to as “pixel chirping”.

Once the location of the pixel center is determined as above, a pixel width is determined. The width for a pixel P_λ is preferably determined as the distance Δx(λ), such that:
Δx(λ)=x(λ)−x(λ−δλ)  3.
where x(λ) is the center line location of the corresponding wavelength λ, x(λ−δλ) is the center line location of the preceding wavelength, and δλ represents the incremental change in the channel wavelength. Since the pixel width is centered about the pixel center line, the pixel width on either side of the pixel center is Δx(λ)/2. In this preferred embodiment, the pixel width is a function of λ, and therefore, the pixel width increases with increasing λ.

Alternatively, the pixel width is determined as the distance Δx(λ), such that:
Δx(λ)=[x(λ−δλ)−x(λ+δλ)]/2  4.
where x(λ+δλ) is the center line location of the wavelength immediately following the wavelength λ. In other words, the pixel width is the average of the distance between the center lines of the two neighboring channels in this alternative embodiment. This averaging method of determining pixel width is not preferred for high contrast operation. However, the averaging method reduces insertion loss and requires less stringent optical alignment methods.

Once the pixel width is determined, the ribbon width of each ribbon within a same pixel is determined. In the case where 1 pixel corresponds to 1 ribbon-pair (two ribbons), a period for a channel corresponding to that pixel is the same as the pixel width. This concept can be generalized to the case where 1 pixel corresponds to n ribbon-pairs. In this general case, the calculated pixel width Δx(λ) is divided by the number of ribbon-pairs in the pixel to determine the period of 1 ribbon-pair. In this manner, the period of a single ribbon-pair, Λ(λ), is determined as:
Λ(λ)=Δx(λ)/n  5.
For a constant ribbon gap width, rg, between ribbons within the same pixel, a ribbon width W(λ) for a given pixel P_λ is determined as:
W(λ)=[Λ(λ)−2rg]/2  6.
It is preferred that the ribbon gap width is constant throughout all pixels. Alternatively, the ribbon gap width can be varied.

FIG. 5 illustrates alignment of pixel centers and incident wavelength center lines for adjacent pixels within a grating light valve™ light modulator. The grating light valve™ light modulator 27 of FIG. 5 comprises a plurality of pixels 30-32, respectively identified by the dotted circles. According to the example of FIG. 5, each pixel 30-32 comprises two ribbons, 40 and 41, 42 and 43, and 44 and 45, respectively. The center line of incident light I(λ) of the component wavelength λ is seen to be approximately aligned with the center line of pixel 31. The center line of the pixel 31 corresponds with the location x(λ) on the grating light valve™ light modulator 27 as measured from the optical axis. The incident beam I(λ) is oriented such that the greatest intensity of the Gaussian distributed light is centered between the ribbons 40 and 41. The ribbon width and spacing between ribbons is selected to ensure that the centerline of the pixel 30 is also aligned with the centerline of the intensity distribution I(λ+δλ) of the next wavelength λ+δλ. As discussed above, however, the linear displacement of each subsequent incident wavelength becomes progressively greater as the wavelength increases, so that the linear displacement x(λ)−x(λ−δλ) is not as great as the linear displacement x(λ+δλ)−x(λ).

FIG. 5 illustrates a portion of a “channel chirped” grating light valve™ light modulator, in which a center line of each pixel of the grating light valve™ light modulator is approximately aligned with a center line of its corresponding incident wavelength. If the grating light valve™ light modulator of FIG. 5 were not pixel chirped, and instead each pixel was evenly spaced across the entire length of the grating light valve™ light modulator, a drift of the incident beams I(λ) from the center line of its respective pixel would occur. In certain cases, the aggregate drift can become so significant that the center line of an incident beam I(λ) can fall between pixels, and eventually cause a pixel to be “skipped” such that n-1 wavelengths are spread over n pixels or more. It can be readily appreciated that such lack of correlation between pixels and wavelengths is not ideal for controlling the diffraction of individual wavelengths.

FIG. 6 illustrates a chirped grating light valve™ light modulator valve type array according to the present invention. The array comprises a first end pixel P₀, through a second end pixel P_N disposed to receive progressively increasing incident wavelengths λ₀ through λ_n. According to the present invention, the array of pixels in FIG. 6 are pixel chirped to reduce or eliminate an offset error between an incident wavelength and its respective pixel. It is recalled from FIG. 4 that the process of demultiplexing a WDM signal by means of a static grating effects the separation distance between successive wavelengths. The wavelength separation distance, measured centerline to centerline of successive Gaussian functions, is not constant, but progressively increases as the wavelength frequency increases. This is referred to as channel chirping. According to FIG. 6, component pixels P₀-P_N are pixel chirped such that the pixel separation distances PS₁-PS_N are not constant. The incremental increase in pixel separation distances PS₁-PS_N is preferably performed at a rate identical to the channel chirping of the respective channels as a result of the static grating demultiplexer 25 (FIG. 4), thereby aligning an incoming component wavelength with its respective pixel. Because, at small angles, the increasing separation between pixels approximates a linear function, according to one embodiment, the rate of pixel chirping selected to match the channel chirping of the incoming component wavelengths is a linear chirping rate. According to this embodiment, successive separation distances increase at the same rate. One example of a linear pixel chirping embodiment of pixel separation is represented by the equation 7.
PS_N=PS₁+NC₀  7.
According to equation 7 above, PS_N is the pixel separation distance from the center of pixel N-1 to the center of pixel N. This distance is equal to the separation distance between the first end pixel P₀ and a second pixel P₁, represented in FIG. 7 as PS₁, plus “n” times the original chirping offset C₀, where “n” is the number of the pixels removed from the first end pixel. The original chirping offset C₀ can be selected as the difference between the first two pixel separation distances, PS₂ minus PS₁.

Those skilled in the art will recognize, however, that the while the tangent function remains near linear for small angles, a linear function does not approximate the tangent function at large angles approaching 90 degrees. Because channel chirping induced by demultiplexing a WDM signal through a static grating is technically non-linear, according to the preferred embodiment, pixels P₀-P_N are chirped at exactly the same rate as the channel chirping of the incoming wavelengths. The advantage of a non-linear pixel chirping is that the pixels can be oriented exactly in line with the incident wavelengths. The advantage of linear pixel chirping is that it may be accomplished easily, and linear pixel chirping approximates the channel chirping induced by the static grating at small angles.

Additional dimensions referenced herein are illustrated in conjunction with the grating light valve™ light modulator 69 of FIG. 7, including the gap width “W_G” between two ribbons, a ribbon width “W_R” and the distance δ between a ribbon 65 and the substrate 67. As discussed in greater detail below, it is possible to distinguish further between gaps within a pixel, and gaps separating ribbons in different pixels.

According to the grating light valve™ light modulator of FIG. 8, a series of ribbons 90-101 are separated by a series of gaps 115-125. According to one embodiment of the present invention, pixel chirping is achieved through progressively widening successive gaps 115-125. According to one embodiment, the width of the respective gaps 115-125 are progressively increased at a rate such that the resulting chirped pixel conforms to the channel chirping induced in the incident component wavelengths by the static grating 25 (FIG. 4). As a result, the center line of each Gaussian function of each wavelength is directed toward the center line of its respective pixel. Because channel chirping increases within increasing incident wavelength λ, as illustrated in conjunction with FIGS. 4-6, according to the preferred embodiment illustrated in FIG. 6, the pixel spacings PS₁, PS₂, . . . PS_N progressively widen from the first end pixel P₀ toward the second end pixel P_N.

Returning to FIG. 8, a pixel must comprise at least two ribbons, but may comprise more than two ribbons. Accordingly, the array of twelve ribbons 90-101 in the grating light valve™ light modulator depicted in FIG. 8 may comprise anywhere from a single pixel to up to six pixels, depending on the application. For exemplary purposes, the grating light valve™ light modulator of FIG. 8 is divided to comprise six pixels 105-110 (circled). A gap 116, 118, 120, 122, 124 disposed between two separate pixels is herein defined as a pixel gap comprising a width W_PG, and a gap 115, 117, 119, 121, 123, 125 separating two ribbons within the same pixel is herein defined as a ribbon gap comprising a width W_RG. Because the pixels may comprise varying numbers of ribbons, however, it can be readily understood that gap 116 functions as a pixel gap in pixel architectures comprising two ribbons per pixel, but will function as a ribbon gap in pixel architectures comprising more than two ribbons. Accordingly, it is envisioned that specific embodiments of the chirped grating light valve™ light modulator according to embodiments of the present invention may incorporate chirping of ribbon gaps exclusive of pixel gaps, or incorporate the chirping of pixel gaps exclusive of ribbon gaps, or chirp pixel gaps and ribbon gaps at different rates of increasing widths.

According to one embodiment, the ribbons 90-101 within the grating light valve™ light modulator are chirped by progressively widening the ribbons 90-101 (not shown) moving from the first end pixel P₀ to the second end pixel P_N (FIG. 7). The progressive widening of the ribbons 90-101 may be done alternatively to, or concurrently with the progressive widening of the gaps 115-125 at the same rate. By chirping the ribbon widths, gap widths, or both, the pixel separation of FIG. 7 is chirped to coincide with the centerline of the incident wavelengths.

FIG. 9 graphically illustrates an exemplary relationship of the channel chirping experienced by de-multiplexed wavelengths. As seen in FIG. 9, linear displacement of an incident wavelength's centerline as it impinges a grating light valve™ light modulator increases as does the wavelength.

FIG. 10 graphically illustrates an exemplary representation of channel width, also referred to as pixel width, as a function channel number. Preferably, channel 0 corresponds to the shortest wavelength. As seen in FIG. 10, the channel width increases as does the channel number. FIG. 10 illustrates the exemplary case in which an increasing pixel width is used to provide pixel chirping.

Polarization dependent loss (“PDL”) is a property inherent in an object that interacts with light. The undesirable effects associated with PDL, however, are often exacerbated as the result of a polarization shift which occurs during signal transmission. A laser generates a wavelength of light according to a particular orientation of the electric and magnetic fields, TE field orientation and TM field orientation, respectively, which is a particular polarization. However, during signal propagation in a fiber optical channel, a continual shift in polarization occurs. This shift is largely induced by characteristics present in the optical fiber.

An optical device such as a lens, slit or mirror can be rated according to the difference in power loss at various field orientations of incoming light. If the power lost in a physical interaction in the TM orientation is identical to the power lost in the TE orientation, the polarization dependent loss of the object is zero. However, if there is a different power loss at different field orientations, the object is said to exhibit a polarization dependent loss, or PDL, which is related to the difference between the loss in the TE field orientation and the loss in the TM field orientation. A power loss is not, in itself, a PDL. If an interaction with a device produces an identical loss in signals of every orientation, the PDL equals zero. There may be an interactive loss, but it is not dependent on or influenced by the field orientations. In interactions where the loss of power or intensity is greater for one field orientation than another, the PDL typically expressed in decibels, which is largely derived from the ratio represented by equation 8 below:
PDL=I_min(P∈p)÷I_max(P∈p)  8.
Equation 8 above states that PDL is the minimum intensity of a signal as a function of polarization from among the set of all possible polarizations divided by the maximum intensity of the same signal as a function of polarization from among the set of all possible polarizations. For example, if the smallest power loss of an apparatus is 3 dB one field orientation, and the maximum power loss of that apparatus is 7 dB in another field orientation, the PDL of that apparatus is 4 dB. Although maximum and minimum intensities typically occur at one of the TM and TE orientations for grating light valve™ light modulators, equation 8 above is true regardless of the field orientation at which the minimum and maximum intensities occur.

Grating light valve™ light modulators are one apparatus known to induce polarization dependent loss in a signal. More significantly, the PDL induced by a grating light valve™ light modulator is a function of the wavelength of the signal. Accordingly, the PDL is not simply a fixed value for a grating light valve™ light modulator, but is also dependent on the wavelength of incoming light.

In the propagation and processing of a wavelength division multiplexed signal, telecommunication standards often require level or constant PDL over a spectrum of frequencies, preferably within 0.3 dB, from the highest PDL to the lowest PDL among component wavelengths, and even more preferably within 0.1 dB among component wavelengths. Accordingly, an apparatus exhibiting high PDL can be unsuited for use in telecommunications. The polarization dependent loss induced by a grating light valve™ light modulator is wavelength dependant.

Referring again to FIG. 7, a pixel 69 of a grating light valve™ light modulator comprises ribbons 65 coated with a reflective coating 66. Each ribbon 65 is separated from other ribbons by a ribbon gap 68. The polarization dependent loss associated with a pixel of a grating light valve™ light modulator is a strong function of the ribbon width W_R of individual ribbons 65, the ribbon gap width W_G between ribbons within a pixel, and the wavelength λ of the incident light beam I_λ. This relationship is depicted by equation 9 below for a given reflectivity coating 66 on the ribbon 65, and for a given reflectivity of the substrate 67:
PDL_GLV(λ)=f(W_R, W_G, λ, Δ, δ)  9.
According to Equation 9 above, the polarization dependent loss experienced by a grating light valve™ light modulator at a given wavelength λ is primarily a function of ribbon width W_R, the ribbon gap width W_G, and the wavelength λ. The distance δ from the ribbon 65 to the substrate 67 and the deflection distance Δ of the ribbon (FIG. 3), have a lesser influence on the PDL experienced by a grating light valve™ light modulator.

Because telecommunication devices are not tolerant to large PDL swings over component wavelengths in a WDM signal, the present invention proposes to equalize the polarization dependent loss of multiple pixels by manipulating select parameters from among W_R, W_G, δ, and Δ within the pixels respectively assigned to the various component wavelengths. The normalization is preferably done by identifying a reference pixel, and manipulating select parameters of equation 9 to normalize the PDL exhibited by the various pixels against a reference pixel. The manipulation of some of these parameters, however, is impractical for a variety of reasons, and is therefore less preferred. For example, it is difficult if not impossible to progressively increase or decrease the reflectivity of the reflective coating or the reflectivity of the substrate, over a sequence of pixels. These reflectivity values are fixed for all practical purposes. Additionally, because the deflection distance Δ is often dictated by system needs for controlling the amount of light that is reflected and diffracted, in many applications it is highly impractical and even counter productive to alter the deflection distance Δ in order to normalize PDL. The separation distance δ between an undeflected ribbon 65 (FIG. 7) and the substrate 67 is a weaker function, and therefore, not an effective means for normalizing PDL of successive pixels against a reference pixel. According to the present invention therefore, the preferred approach for normalizing the PDL of successive pixels 105-110 (FIG. 8) of a grating light valve™ light modulator against a reference pixel is by adjusting the ribbon width W_R and the ribbon gap width W_G in the various pixels of a grating light valve™ light modulator. As discussed above, in fabrications wherein a distinction can be made between ribbon gap widths W_RG and pixel gap widths W_G, the PDL of a pixel is primarily a function of the ribbon gap width W_RG, that is, the gap widths between ribbons comprising a single specific pixel.

According to the mathematical implications of Maxwell's equations, select parameters within equation 9 can be multiplied by a constant to affect PDL of a particular pixel without destroying the functionality of equation 9. Accordingly, the PDL exhibited by pixel “n” is normalized against a reference pixel by multiplying the ribbon gap widths and the ribbon widths in pixel “n” by a ratio of wavelength λ_n divided by a reference wavelength λ_ref, as illustrated in Equations 10 and 11.
W_G(n)=W_G(ref)(λ_n/λ_ref) and  10.
W_R(n)=W_R(ref)(λ_n/λ_ref)  11.
In equations 10 and 11 above, above, W_G(n) is the width of a ribbon gap in the n^th pixel, W_G(ref) is the width of a ribbon gap in the reference pixel, and λ_n and λ_ref are the wavelengths respectively assigned to pixel “n” and to the reference pixel. By progressively widening the widths of both the ribbon gap and the ribbons of the various pixels according to equations 10 and 11, the PDLs of all pixels are normalized against the PDL of the reference pixel. The application of equations 10 and 11 together is herein called “PDL chirping”. The use of the term “widening” is used for verbal economy. It is understood that the present invention may equally use the widest pixel as a reference and narrow the others respectively, or use a reference pixel in the center of a linear array, with pixels on one side progressively narrowed, and pixels on the other side progressively widened.

A grating light valve™ light modulator chirped according to this configuration is illustrated in FIG. 11. A signal S₂ passes through a static grating demultiplexer 127, and the component signals λ₁, λ₂, λ₃ pass through a transform lens 128. Using the PDL of pixel PIX₁ as reference against which the PDL of all other pixels are to be normalized, the PDL of pixel PIX₂ is normalized by multiplying both the ribbon widths W_R2 of pixel PIX₂ and the ribbon gap width (or widths) W_G2 of pixel PIX₂ by a ratio of wavelengths according to equations 10 and 11 above, such that, as applied to pixel PIX₂ of FIG. 11, W_R2=W_R1·(λ₂/λ₁), and W_G2=W_G1·(λ₂/λ₁). Similarly, as applied to pixel PIX₃ of FIG. 11, W_R3=W_R1·(λ₃/λ₁), and W_G3=W_G1·(λ₃/λ₁).

Another phenomena experienced within some optical systems is the wavelength dependent nature of the first order diffraction angle. Because the angle of the diffraction patterns is wavelength dependent, in an unchirped array, as the angle of diffraction changes from pixel to pixel, the bright and dark areas of a diffraction pattern change position. This presents a problem in consistently focusing the bright and dark portions of a diffraction pattern over a range of wavelengths. In contrast, a chirped array of pixels will exhibit a constant angle of first order diffraction patterns if chirped according to a proper mathematical model. FIG. 12 illustrates three pixels properly chirped to produce a constant angle of first order diffraction patterns. The three pixels receive a corresponding incident wavelength, λ_A, λ_B, and λ_C, respectively, which are arranged according to progressively increasing wavelengths from A to C. D_1A and D_−1A represent the first order diffraction patterns created by pixel PIX_A. According to FIG. 18, θ_1A represents the angle between the incident beam λ_A and the first order diffraction pattern D_1A resulting from incident wavelength λ_A falling on pixel PIX_A. The three pixels are chirped to produce constant diffraction angles by linearly increasing the ribbon width with increasing incident wavelengths λ_A, λ_B, and λ_C. In FIG. 12, the pixels are chirped based on PIX_A resulting in constant diffraction angles for PIX_A, PIX_B, and PIX_C according to the diffraction angles of PIX_A. In this case, the intra-pixel gap width, also called the ribbon gap width, remains constant. The resulting pixel-center shift due to the increasing ribbon widths does not necessarily coincide with the incident wavelength center-line. Alternatively, the pixel-center shift can coincide with the incident wavelength center-line to provide channel chirping. The relationship between the ribbon width and the incident wavelength, as well as alternative methods for producing constant diffraction angles, are described in detail below.

The pixel chirping embodiment discussed in conjunction with normalizing the PDL of various pixels against a reference pixel, as illustrated in conjunction with FIG. 11 and equations 10 and 11, has the additional benefit of normalizing the first order diffraction angles exhibited by the various pixels against the diffraction angle of the reference pixel PIX₁. It is recalled, however, that the chirping of a grating light valve™ light modulator according to the formula of equations 10 and 11 together requires the simultaneous chirping of successive ribbon widths W_R and successive gap widths W_G (FIGS. 6 & 7).

A grating light valve™ light modulator type array chirped according to equations 10 and 11 exhibits a substantially constant diffraction angle across a span of wavelengths, whereas an unchirped grating light valve™ light modulator type array exhibits a proportional increase in the diffraction angle of the first order diffraction pattern as wavelength increases.

An alternative embodiment of a chirped grating light valve™ light modulator type array is envisioned according to the present invention wherein the variations in the diffraction angle of the component wavelengths are normalized against the diffraction angle of a reference wavelength by exclusively altering the ribbon width W_R, or exclusively altering the ribbon gap width W_G. Alternative equations describing the chirped ribbon widths or chirped ribbon gap widths according to this embodiment are respectively described in Equations 12 and 13 below.
W_R2=λ₂/λ₁·(W_R1+W_G1)−W_G1  12.
W_G2=λ₂/λ₁·(W_R1+W_G1)−W_R1  13.
According to equations 12 and 13 above, W_R1 represents the width of the ribbons in pixel PIX₁, W_R2 represents the width of the ribbons in pixel PIX₂, and the value of λ₂/λ₁ represents the ratio of the respective wavelengths falling on pixels PIX₂ and PIX₁. W_G1 represents the width of the ribbon gaps between the ribbons in pixel PIX₁, and W_G2 represents the width of the ribbon gaps between the ribbons in pixel PIX₂. Applying equations 12 and 13 above to the illustration of FIG. 11, wherein pixel PIX₁ is the reference pixel, the present invention envisions normalizing the diffraction angle of a pixel PIX₂ to that of pixel PIX₁ by either chirping the width of the ribbons W_R2 in pixel PIX₂ according to equation 12, or by chirping the width of the ribbon gaps W_G2 in pixel PIX₂ according to equation 13.

The application of equations 10 and 11 together is herein called “dual PDL/0 chirping” in that pixel chirping according to these equations simultaneously affects a normalization of both PDL and normalization of the diffraction angle θ. Pixel chirping alternatively performed according to Equations 12 or 13 is herein called “limited diffraction chirping” since either equation is limited exclusively to normalizing diffraction in successive pixels, and is not applicable to normalizing PDL. In either dual PDL/θ chirping or limited diffraction chirping, the pixel gap width (the gap between adjacent pixels) can either remain constant or can vary as a function of the incident wavelength λ.

Not all embodiments implemented to correct for dual PDL/θ chirping and limited diffraction chirping are equally suited to also correct the centerline offset error, using pixel chirping, to account for channel chirping. However, dual PDL/θ chirping is not necessarily incompatible with pixel chirping. By distinguishing between ribbon gaps occurring between ribbons in the same pixel, and pixel gaps occurring between adjacent pixels, it is possible to create a chirped grating light valve™ light modulator type array that corrects for channel chirping and wavelength-based chirping (dual PDL/θ chirping or limited diffraction chirping). The spacing which determines the centerline error offset addressed by channel chirping is not limited to a function of ribbon widths and intra-pixel gap widths. It is exclusively a function of pixel separation PS₁, PS₂, etc. as illustrated in FIG. 6, which is effected by the combination of ribbon widths, intra-pixel gap widths, and inter-pixel gap widths. Specifically, the center-to-center distance between a first pixel and an adjacent pixel can be determined by one of the methods described above. For example, the center-to-center distance between the first pixel and the adjacent second pixel (FIG. 6) is simply ½ of the additive values of the ribbon gaps W_RG and ribbon widths W_R of the first pixel plus ½ the additive values of the ribbon gaps W_RG and ribbon widths W_R of the adjacent pixel plus the width of the pixel gap W_PG separating the two pixels. According to one embodiment therefore, equations 10 and 11 are first applied to perform dual PDL/θ chirping from a reference pixel. The PDL/θ chirping is directed exclusively to “ribbon gaps” (intra-pixel gaps within a single pixel). This establishes the ribbon widths and ribbon gap widths of the various pixels to level the PDL and diffraction angle offset. After these values are calculated, and taking this spacing into account, values are calculated for the inter-pixel gaps W_PG1-W_PGN (FIG. 6) between the respective pixels to adjust the pixel spacing to match the channel chirping induced by static grating, thereby reducing the centerline offset error induced by the static grating demultiplexer. According to another embodiment, the inter-pixel gap remains constant, and the ribbon widths and ribbon gap widths are established to perform dual PDL/θ chirping while simultaneously accounting for pixel spacing to properly align center-lines for channel chirping.

FIG. 13 illustrates multiple pixels PIX₀-PIX₅ in a grating light valve™ light modulator exhibiting chirped intra-pixel gaps 131, chirped ribbon widths 130, and chirped inter-pixel gaps 132, wherein the chirping is performed at independently controllable rates. Calculations are first performed for chirping the ribbon gaps 131 and ribbon widths 130 at a rate necessary to effect PDL normalization and diffraction angle normalization. Subsequently, calculations are performed for adjusting the pixel gap widths in light of the previous calculations. The chirping of the pixel gap widths 132 is at a rate independent of the chirping of the ribbon widths 131, and the pixel gap widths 132 are chirped according to the necessary amount to achieve pixel alignment after accounting for chirping imposed by equations 10 and 11. In this manner, normalization of polarization dependent loss and the diffraction angle is achieved simultaneously to the channel chirping performed to align the various pixels with their respective wavelength.

In the field of telecommunications and many other present day applications, optimum functioning of a grating light valve™ light modulator requires an alignment between component wavelengths of a demultiplexed WDM signal and their respective pixels. Optimum performance further requires the normalizing of polarization dependent loss for the various pixels respectively assigned to various wavelengths. Optimal performance further requires the normalizing of the angle of diffraction induced by the various pixels on their respective wavelengths. The embodiments of the invention as described above, and set forth in the claims herein, are directed to a method of and apparatus for chirping pixels to align component pixels with their respective wavelengths. Embodiments of the present invention are also directed to a method of chirping pixels to normalize the PDL of the various pixels against a reference pixel. Embodiments of the present invention are also directed to a method and apparatus for normalizing the diffraction angle imposed by an array of pixels. Embodiments of the present invention are further directed to independently chirping pixel ribbons, pixel gaps and ribbon gaps to allow the simultaneous implementation of solutions for PDL, diffraction angle, and centerline offset errors experienced in grating light valve™ light modulator type arrays. Embodiments of the present invention are still further directed to singularly, or in any combination, implementing the solutions for dual PDL/θ chirping, limited diffraction chirping and pixel chirping depending on the required application.

Although the aforementioned embodiments of the present invention include a grating light valve™ light modulator, other types of light modulators can be used in alternative embodiments. Such light modulators can include, but are not limited to, tilting mirror arrays and LCD arrays.

The present invention has been described in terms of specific embodiments incorporating details to facilitate the understanding of the principles of construction and operation of the invention. In many cases, the claimed invention can operate without these additional structures or details, or can be made to function with alternative equivalent structures. It will be apparent to those skilled in the art that many of the details found herein are not essential to the manufacture and use of the claimed invention. Accordingly, such reference herein to specific embodiments and details thereof is not intended to limit the scope of the claims appended hereto.