WO2005041423A2

WO2005041423A2 - Wavefront polarization and phase scrambler

Info

Publication number: WO2005041423A2
Application number: PCT/US2003/033179
Authority: WO
Inventors: Nicolae Miron
Original assignee: Stockeryale Inc.
Priority date: 2002-10-17
Filing date: 2003-10-17
Publication date: 2005-05-06
Also published as: AU2003299551A1; WO2005041423A3; AU2003299551A8

Abstract

There is provided herein a method and a device to scramble simultaneously the polarization and the phase of the wavefront of the wavefront of a light beam by using a wavefront sampling method and a device that can change the polarization state and the phase retardation of individual samples of the wavefront of the light beam (102 in Fig. 1). An exemplary method consists of sampling the wavefront of a beam having a certain polarization, changing the polarizatoin state of some samples only, leaving the remaining samples with the same polarization state but inducing instead a phase retardation, thus generating another wavefront with different polarization states and phases on its surface (611). The overall polarization and phase of the resulting wavefront will be randomized with respect to the polarization of the initial beam; that is, the polarization state and the phase will be highly scrambled.

Description

WAVEFRONT POLARIZATION AND PHASE SCRAMBLER

This application claims priority of US Application 60/419,154, filed on October 17, 2002 and which is incorporated herein by reference in its entirety.

TECHNICAL FIELD OF THE INVENTION

The present invention relates to a method and a device to scramble the polarization and the phase of a wavefront of a light beam by sampling the wavefront of the beam and by changing the polarization state on some of the samples and the phase of the wavefront on the remaining samples so that the resulting wavefront will have overall no preferential state of polarization and undefined phase, while keeping its original frequency spectrum (i.e. monochromatic). In essence, the method is static, which means that there is no need to have a time varying pattern of the wavefront samples n order to change the polarization state, but if some certain applications require this feature, the polarization-scrambling device can change the pattern with respect to time. The polarization of the original beam can be either linear or elliptic and the wavelength range is approximately a few tens of nanometers around an optimum wavelength at which the effect reaches its maximum for a given applied voltage. However, the optimum wavelength can be selected anywhere in the visible and in the near infrared part of the spectrum, by adjusting the drive voltage of the device. . In contrast to the present invention and its operating principle, the prior art operates with the beam as a whole in order to obtain a time-averaged polarization scrambling. The application of the wavefront polarization scrambler according to the present invention can be in laser illumination products to reduce the speckle and in optical communications such as to minimize the polarization dependence properties of Erbium-doped fiber amplifiers (EDFA), Mach-Zehnder interferometers, arrayed waveguide gratings (AWG), fiber optic couplers, fiber optic splitters and fiber Bragg gratings (FBG), to mention just a few.

DESCRIPTION OF THE PRIOR ART

Light polarization is an important issue in interferometry and in optical communications. It is well known to those knowledgeable in the art that two monochromatic beams with the same polarization, here considered having linear polarization for simplicity, interfere optimally if their polarization vectors are collinear. The interference is less efficient (less fringe contrast) if the polarization vectors are not collinear, the efficiency decreasing by increasing the angle between the polarization vectors from zero to 90° (cross-polarized beams) at which angle it reaches its minimum value. Those skilled in machine vision applications know that local interference between the light beams scattered by adjacent points of a rough surface produce the speckle, which is the major source of optical noise when a laser source (coherent light) is used for illumination. On the other hand, laser illuminators are preferred for their monochromatic light and high brightness, which allows the use of interference filters to reject the ambient light. Therefore it is desirable to produce a polarization and phase modulator that will permit the use of laser illuminators, while reducing the optical noise generated by speckle.

In optical communications, all semiconductor lasers generate linear polarized beams with high degree of polarization, usually better than 120:1. The orientation of the polarization vector is unknown at the entrance in the optical fiber and is also unknown at the entrance port of the polarization sensitive components used in the optical networks such as AWG, fiber optic splitters, fiber optic couplers, Mach-Zehnder interferometers, EDFA, FBG. Those skilled in the art know that EDFA gain depends on a polarization, effect known as polarization hole burning. AWG are sensitive to polarization because of the rectangular geometry of their waveguides. Mach-Zehnder interferometers are polarization sensitive because of the interference principle and, if there are built into an integrated structure, the waveguides are sensitive also to polarization. FBG have their insertion loss dependent on polarization because of the geometry of the volume grating with respect to the optical waveguide. Fiber optics couplers and splitters are also sensitive to the polarization state because of their geometry, too.

Prior art U.S. Patent No. 4,511,220, will be summarized in connection with FIG. 1. A laser 101 generates a monochromatic light beam 102 having a linear vertical polarization state 103. The beam 102 goes through a polarization rotator 104 generating at the output a linearly polarized beam 105 with polarization 106 rotated with 45° with respect to the polarization state 103. Further, the beam 105 is incident into a polarization beam splitter 107 that generates two beams: beam 112 that keeps the same direction of propagation and the same polarization state as the incident beam 105, and beam 108 with horizontal polarization that propagates perpendicular to the incident beam 105. The beam 108 is reflected two times at right angle by the roof prism 110, to obtain the beam 111 parallel with the beam 108 and having also a horizontal polarization. A beam combiner 113 receives the two beams 111 and 112 propagating in perpendicular directions having respectively perpendicular polarizations and also an optical path difference D between them. The beams 111 and 112 are merged in the recombination point 117, into a single beam 116 having both the horizontal polarization state 114 and the vertical polarization state 155. The optical path difference D is due to a geometrical path difference between the splitting point 117 and the recombination point 118. This prior art method however fails to eliminate the speckle in the laser target. In fact, the speckle is only partially eliminated, for the reasons explained later. The resulting output beam 116 contains just the superposition of the beams 111 and 112, both coherent and having respective orthogonal polarizations and a phase shift between them given by the difference in their geometric paths. These differences in optical paths and in polarization do not prevent each of the beams to produce its own speckle pattern because each beam is spatially coherent on its wavefront. Instead of having a single speckle pattern produced by one beam 102 with single polarization, the optical setup according to the prior art produces two speckle patterns respectively generated by the overlapping beams 111 and 112 with crossed polarizations, but each of them still spatially coherent. Therefore, the prior art does not eliminate the speckle; it just generates two speckle patterns with orthogonal polarizations.

Other speckle-reduction methods use the time averaging to obtain a lower value of the optical noise. U.S. Pat. No. 5,621,529 (the "529" patent) claims a speckle reduction method comprising a monochromatic light source, such as a laser, and a pattemiser for generating a pattern of projected laser light onto a surface. Typical patterns are parallel stripes of narrow bands of light and dark. The apparatus also includes a means for causing the projected pattern to move relative to the surface, parallel to the parallel stripes of the pattern. Thus, the pattern that is projected with the moving apparatus engaged is elongated relative to a pattern that would be projected with the moving apparatus disengaged. The resultant speckle-type interference in the projected pattern is reduced from that normally present in patterns of projected monochromatic, coherent light. The same image of low-speckle line pattern but with smaller fill factor is obtained by moving linearly a distribution of dots, perpendicular on their alignment direction. The apparatus according to the 529 patent can generate only lines with low content of speckle, realized by time averaging. For the highspeed machine vision systems, where the acquisition time is usually much shorter than the averaging time, this method it is not suitable because the image acquisition equipment practically sees the same amount of speckle. U.S. pat. No. 5,274,494 (the "494" patent) presents an optical setup to reduce the speckle by broadening the frequency spectrum of the laser beam using the Raman effect, in order to reduce the temporal coherence of the beam. For those knowledgeable in the art is well known that Raman effect is an inelastic interaction between photons and the molecules of a transparent medium where the light is propagating, resulting a number of photons with smaller energy than the photons of the incident beam. This results in the generation of a beam having wavelengths smaller than the wavelength of the incident beam and distributed into a relatively broad spectral range. Raman- generated wavelengths broaden the spectrum of the incident light producing a resulting beam with lower temporal coherence than the incident beam that will produce less speckle. While using the Raman effect to reduce the speckle by reducing the temporal coherence of the beam is somewhat effective, it has several drawbacks, it requires an initial beam (pump source) with high intensity generated usually by a high power laser, and the intensity of the Raman radiation is quite low because of the conversion efficiency.

U.S. Pat. No. 5,313,479 describes a method to reduce the speckle by using a rotating diffuser placed into the laser beam, between the laser and the object illuminated by the laser beam. The spinning diffuser makes a constant change in speckle distribution, which produces a low speckle image for the viewer's eyes, by time averaging.

U.S. Pat. No. 4,360,372 describes an optical element able to reduce the speckle produced by the laser beam generated by a semiconductor laser. The optical element that reduces the speckle, according to this invention, consists of an optical guide made of optical fibers with different refractive indices bundled together, preferably in random manner, that are fused together to form a multimode fiber having different modes of light propagation, due to different refractive indices of the initial fibers. According to the author, the random distribution of the refractive indices in the resulting multimode core produces mode scrambling that reduces the speckle. The description of the patent does not give any number relative to speckle-reduction effect. In reality, the twisted bundle of optical fibers with different refractive indices that are fused together produces at the output a beam with a still strong coherent wavefront, because the differences between the refractive indices of the initial optical fibers used to build the multimode core are very small and the phase relationships between the different points of the wavefront emerging from the resulting multimode fiber are constant. There are some practical limitations related to the optical element claimed by the invention, such as: difficulty of realizing the described multimode optical fiber and insertion losses related to the coupling with the source.

U.S. Pat. No. 6,046,839 (the "839" patent) describes a polarization scrambler based on time averaging, for optical communication purposes. The device has two sections. The first section rotates the linear polarization of the incident beam and the second section transforms the linear polarized beam generated by the first section, into an elliptic polarized beam. Both the polarization rotation and the conversion from a linear polarized beam to an elliptic polarized beam are variable in time, which generate according to the authors, a polarization scrambling. Considering that these two modulations of the polarization state of the beam can be done at frequencies in MHz range or beyond, the device according to this invention can eventually work as a polarization scrambler to reduce the speckle in some machine vision applications. However, for optical communications, where the transmission rates are much higher than the modulation frequency of the polarization state, the device according to the '839 patent is unsuitable, because it can induce a dangerous parasitic amplitude modulation for some optical components sensitive to polarization, such as AWG

or FBG. U.S. pat. No. 5,933,555 claims a polarization scrambler built with a 2x2 ports fiber optic coupler that acts as a standard Y coupler for the input beams and as a 2:1 splitter for the output beam. One input port is connected to one output port by an external optical fiber, forcing part of the output light beam to re-enter into the splitter into an endless loop. The incident beam enters into the coupler on the only available input port of the Y coupler, combines with a part of the output beam routed to the other input of the coupler by the external fiber and the two beams are propagating as a single combined beam through a birefringent fiber that is part of the 2x2 coupler, comiected between the Y input coupler and the output 1 :2 beam splitter. Inside the birefringent fiber, the combined beam changes its polarization state, arriving at the 1:2 beam splitter with a different polarization state than at the entrance into the Y birefringent coupler, the beam available at the recirculating output having another polarization state than at the input into the birefringent fiber. Because of the infinite number of such loops made by part of the input beam entering at one input of the 2x2 coupler, to one output of the 2x2 coupler and back to the other input of the 2x2 coupler, the beam available at the other output of the 2x2 coupler will have the polarization scrambled. The method claimed by the prior art and the device built using the method have the advantage of being totally passive (no drive voltage necessary), but with the drawback that an important part of the input beam is always present in the output beam, with its polarization state. Moreover, the polarization scrambling depends on the polarization state of the incident beam at the input of the 2x2 coupler, because both the Y coupler existing at the input of the device and the 1 :2 splitter at the output of the 2x2 coupler are sensitive to polarization, favoring a certain polarization state. Many methods have also been developed to reduce the speckle, based upon the idea to reduce the interference between the scattered beams at the object surface level, however none have been powerful enough to be completely effective.

It is therefore an object of the present invention to provide a polarization scrambling method based on two-dimensional sampling of the wavefront with a matrix of polarization active elements (PAE) and to control the polarization state of each sample of the wavefront, such that the average polarization for the whole wavefront should be random or undefined.

It is a further object of the present invention to provide a liquid crystal modulator (LCM) with a plurality of elements (modulation pixels or MP) as a preferred embodiment to implement the polarization scrambling method based on two-dimensional sampling of the wavefront, each modulation pixel being able to change the polarization state of a beam going through it when no voltage is applied on a MP (polarization active MP) or to keep the same polarization of the beam but to increase the retardation of the beam when a certain voltage is applied on MP (polarization passive MP), by individual electrical addressing of each MP, such as LCM as a whole provides simultaneously a polarization and a phase scrambling of the wavefront.

It is a further object of this invention to provide different embodiments for the LCM either to keep constant in time the two-dimensional distribution of the polarization active MP and of the polarization passive MP, or to make this two-dimensional distribution variable in time into a user- defined manner. It is yet another object of this invention to make programmable the two-dimensional sampling rate of the beam.

It is yet another object of this invention to provide an optional optical setup with spatial filtering properties in order to remove the high order spatial frequencies eventually introduced by LCM to eventually eliminate the effect of phase-only diffraction grating produced by LCM.

It is yet another object of this invention to realize beam smoothening simultaneous with polarization and phase scrambling.

It is yet another object of this invention to use the combination of LCM and of the spatial filter as a polarization and phase scrambler (PPS) unit to reduce the speckle produced by a laser beam in illumination applications and to scramble the polarization of a beam propagating through an optical fiber, for optical communication purposes, where phase scrambling is not an important issue.

SUMMARY OF THE INVENTION

According to the present invention, these and other objectives and advantages are achieved using a method and an appropriate device to scramble the polarization of a light beam. In a preferred embodiment a waveform polarization and phase scrambler samples a wavefront of a light beam preferably collimated or having a small divergence or convergence, at normal incidence on a two-dimensional matrix of MP acting on polarization and phase of the incident beam, each MP being able to change the polarization state and the phase of a beam going through, depending on the value of a control voltage applied to it, and a spatial filter capable to reject the high diffraction orders eventually generated by the two-dimensional matrix, the spatial filter being optionally used also to smoothen the beam in highly demanding applications.

According to an important feature of this invention, to sample the wavefront of a beam at normal incidence there is provided a liquid crystal modulator with a rectangular matrix of MP. The wavefront of the beam is projected onto the matrix and is thereby sampled.

According to another important feature of this invention, each MP with low or eventually zero control voltage is in polarization active state (PAS) changing the polarization of the beam going through. Conversely, each MP with high or eventually optimum control voltage is in polarization passive state (PPS) not changing the polarization of the beam passing through the matrix. The two- dimensional combination of the elements of the liquid crystal modulator: some in PAS and others in PPS, generates at the output a wavefront with different polarization states on it, or globally with an undefined or a scrambled polarization.

According to another important feature of this invention, each MP with a high control voltage applied on it changes the optical retardation for the beam going through it, the plurality of MP in PAS and in PPS producing also a phase scrambling of the whole wavefront.

According to another feature of this invention, the two-dimensional distribution of MP in PAS and in PPS can be either constant in time, or variable in time. The phase scrambling may produce also diffraction orders generated by a phase only diffraction grating resulting from the two- dimensional distribution of MP in PAS and in PPS. According to another feature of this invention, a spatial filter can suppress the high diffraction orders.

Accordingly there is provided a method of wavefront polarization and scrambling comprising providing a collimated light beam having a wavefront, the wavefront having a first polarization, isolating the wavefront into discrete portions, and simultaneously changing the first polarization of at least one of said discrete portions, and introducing a phase retardation to the discrete portions having the first polarization.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1. is a schematic drawing showing a polarization scrambler according to the prior art.

FIG. 2 is a schematic drawing showing the operating principle and the device of the wavefront polarization and phase scrambler according to the preferred embodiment of this invention.

FIG. 3 is a schematic representation of a detail of the wavefront emerging from the polarization and phase modulator.

FIG. 4 shows schematically the operation of the low-pass spatial filtering according to the preferred embodiment of the invention.

FIG. 5 shows schematically the electrical connections of LCM to the liquid crystal drivers and to a controller to generate the two-dimensional pattern of MP in PAS and in PPS.

FIG. 6 shows schematically an optical setup using PPS that scrambles the polarization of the pump laser of an Erbium-doped fiber amplifier (EDFA). PAGE INTENTIONALLY LEFT BLANK

DETAILED DESCRIPTION

The following description of the preferred embodiment is merely exemplary in nature and it is in no way intended to limit the invention or its application or uses.

The polarization scrambling done by the preferred embodiment of the present invention operates by sampling the wavefront of a light beam and by producing simultaneous changes of the polarization state on some samples of the wavefront, leaving the remaining samples of the same wavefront in the initial polarization state and also introducing a phase retardation in the remaining samples, which has the effect of phase scrambling the wavefront, the total result on the' emerging wavefront being a simultaneous scrambling of the polarization state and of the phase of the beam, without any time-averaging.

The preferred embodiment of the present invention will be explained further, in connection with the FIG. 2. The polarization and phase scrambler (PPS) according to the preferred embodiment of this invention consists of a polarization and phase modulator (PPM) 201, two converging lenses 202 and 203, and a pinhole 204. PPM consists of two transparent glass plates 205 and 206, being in parallel alignment with respect to each other and flat. In the preferred embodiment of the invention, the gap between the glass plates 205 and 206 is filled with a twisted nematic liquid crystal (LC) 207, but some other types of liquid crystals can also be used. The glass plate 205, on the side facing the glass plate 206, has a series of R parallel and transparent row stripe electrodes 208_1; 208₂, ... 208_R each of them connected to a series of electrical terminals 210^ 210 , ... 21 O_R. On top of the electrodes 208₁, 208₂, ... 208_R is a dielectric alignment layer 212, which is orienting the LC molecules in contact with it, along with the alignment direction 214. Further, the explanation will use either the word molecule, or the molecular vector that is collinear with the direction of the molecule, depending on the context. The glass plate 206, on the side facing the glass plate 205, has a series of C parallel and transparent column stripe electrodes 209_l3 209₂, ... 209c, each of them connected to a series of electrical terminals 211^ 211₂, ... 211c. On top of the electrodes 209_l5 209₂, ... 209c is a dielectric alignment layer 213, which is orienting the LC molecules in contact with it, along with the alignment direction 215. U.S. Patent No. 5,091,794 describes very well the alignment of LC molecules between the plates 205 and 206 in the absence and in the presence of the electrical field applied between the row electrodes 210 and the column electrodes 209. When the electric field in the gap between 205 and 206 is very small or zero (low or zero drive voltage), LC molecules are arranged in stacks between the two glass plates, with their molecular vectors parallel with the plates 205 and 206, LC molecules in contact with the alignment layer 212 being oriented along the direction 214, LC molecules in contact with the alignment layer 213 being oriented along the direction 215, all the other molecules lying between them being stacked in layers on top of each other, parallel with the glass plates and tilted from one layer to the next layer, realizing a continuous twist between the two alignment layers. Finally, the long molecules of the twisted nematic LC 207 are arranged in twisted stacks between the glass plates 205 and 206, with 90° total twist angle. When the drive voltage is low or zero, every such LC molecular stack rotates with 90° the polarization plane of a light beam at normal incidence on the plate 205 so the beam exits through the plate 206 with the polarization state perpendicular to the polarization state at the entrance in the plate 205. This is called the polarization active state (PAS) of the MP, and characterized as having VLC=0 (FIG. 2). When the drive voltage between the row and column stripe electrodes that define a MP is high or optimal for LC 207, LC molecules tend to be aligned with their molecular vector parallel with the electric field created by the drive voltage and will rotate less or with a minimal value the polarization plane of the beam incident on the plate 206. There is a certain optimum value Nopt, which in the preferred embodiment of the invention can be in a few volts range, wherein NLC^Nopt, for which the polarization state of the emerging beam will be changed a small amount if at all, keeping almost the same state as of the incident beam. This state is referred to as the polarization passive state (PPS) of MP. Moreover, in PPS the refractive index of LC is significantly higher than in PAS, which introduces also a retardation of the beam going through an MP having a high voltage applied to it. For those knowledgeable in the art, the electro-optical property of LC to monitor the polarization state of a light beam using a control voltage is widely used in liquid crystal displays (LCD), where in order to visualize the difference between the PAS and PPS of the LC, the whole device 201 is placed between two crossed polarizers: one in front of the plate 205 having the polarization direction parallel with 214, and the other behind plate 206 having the polarization direction parallel with 215. The polarization scrambler according to the present invention is very similar with a regular dot matrix LCD working in transmission, but without polarizers.

one embodiment of the invention, a parallel beam 116 of monochromatic light with a plane wavefront 117 and with a polarization state Px collinear with the direction 214 of the alignment layer 212 is incident on the glass plate 205. Into one possible embodiment of the invention, MP of the PPM have alternating PAS and PPS on both horizontal and vertical directions, as a checkerboard pattern, which changes alternatively the polarization and the phase shift of the light beam passing through the whole device 201. On all MP in PAS, the polarization state of the incident beam will be rotated with 90°, being aligned with the direction 215 of the alignment layer 213, as denoted by 222 (FIG. 2). On all MP in PPS, the polarization state of the incident beam will be kept in the same state as of the incident beam, being aligned with the direction 214 of the alignment layer 212, as denoted by 223 (FIG. 2). Each MP makes a spatial sampling of the incident wavefront and for a large enough number of MP contained in the wavefront 217 of the incident beam, the polarization will have a quite uniform distribution of samples in the initial polarization state 214 but with a phase retardation and also a complimentary distribution of samples in the induced polarization state 215, the samples of the wavefront in the two states being interleaved, as is shown in the schematic drawing of the emerging wavefront 219 (FIG. 2). A detail 221 of the emerging wavefront 219 is shown in FIG. 3. For those knowledgeable in the art it is well know that a large enough number of MP's in the wavefront of the beam means a sufficient number of samples of the wavefront will be made in accordance with the sampling theorem for that beam, in terms of two- dimensional intensity distribution function in the cross section of the beam. As a practical value, a number between 30 and 100 MP can sample the beam quite well for most applications. The resulting distribution of the polarization state of the wavefront 219 at the output of the device 201 will have no preferred direction, i.e. the polarization is scrambled. Into another preferred embodiment of the invention, by grouping a few MP together into a single modulation pixel cluster MPC having a b modulation pixels where a = 1, 2, ...C and b = 1, 2, ... R, the spatial sampling rate of the wavefront can be adjusted and accordingly the degree of randomization of the polarization state averaged over the whole wavefront can be adjusted, too.

The real LCM always has a gap 224 between the stripe electrodes 208x and 209y, which is needed to isolate them electrically. There is no electrical control on polarization inside the gap, and accordingly the polarization state of the incident beam going through the gap will always be rotated with 90°. This 90° rotation will introduce a residual polarization in the output beam 218. Generating a pattern containing more MP's in PPS than in PAS can compensate for the residual polarization. To give an idea about some numbers associated with the residual polarization of a preferred embodiment of an LCM manufactured by regular, low cost technologies of LCD manufacturing, the minimum stripe widths Dx, Dy are in 0.250mm range and the minimum gap value g is in 0.025mm range. For these numbers, the relative value of the gap versus the total area of a pixel including a two-dimensional gap is:

(Dx + g) - (Dy + g) -Dx - Dy _{Q 17} (Dx + g) - (Dy + g) which means that about 17% of the cross-section of the beam must be compensated for a full polarization scrambling.

In the preferred embodiment of the invention described herein, the residual polarization can be compensated by having about 8% more MP in PPS than in PAS. The preferred embodiment of LCM allows this compensation, by driving more MP's with higher voltage than with lower voltage, thus resulting in a greater number of MP's in PPS than in PAS.

In the description of the principle and of the operation of PPS given herein and above, the polarization state of the input beam was considered linear and also collinear with the direction 214 of the alignment layer 212 (component Ex) that does not restrict at all the preferred embodiment of this invention to this polarization state only. Generally, the incident beam has an elliptic

polarization state represented by the polarization vector Pi making an instantaneous angle α with

the direction Ox collinear with the orientation 214 (FIG. 2). As it is generally well known by those skilled in the art (see G.H. Gooch, H.A. Tarry: "The Optical Properties of Twisted Nematic Liquid

Crystal Structures with Twist Angles < 90° ", J. Phys. D: Appl. Phys., vol. 8, 1975), both

components Px and Py of Pi will be twisted by MP in PAS or, in other words, the whole Pi will be twisted with 90°, which make valid all the explanations given herein and above for all polarization states of the input beam. In the preferred embodiment of this invention shown schematically in FIG. 2, the liquid crystal modulator 201 works also as a phase grating, due to the increases of the refractive index of MP's in PPS, which may produce a central beam as presented above, and may result in diffraction orders (i.e. +1, -1, +2, -2, ...) of the wavefront output from the modulator. All diffraction orders may appear as multiple output beams in the far field, which are unwanted for certain applications. According to the preferred embodiment of this invention, a spatial filter can filter out the high diffraction orders generated by LCM. An exemplary embodiment of the spatial filter is represented also in FIG. 2, and consists of a converging lens 202 having the focal distance FI, an opaque plate 225 with a pinhole 204 and a converging lens 203 having the focal distance F2. The lenses 202 and

203 are centered on the same optical axis 226 that is normal to the LCM 201 in its geometrical center, the two lenses being separated at a distance FI + F2, the opaque plate 225 being normal to the same optical axis 226 having the pinhole 204 centered to the axis 226 and located at the common focal point of the two lenses 202 and 203. The exemplary spatial filter described above is a low-pass spatial filter having the cut-off frequency determined by the diameter Dp of the pinhole

204 which is well known to those skilled in the art.

A detailed theory of how the lenses 202, 203 and the pinhole 204 perform spatial filtering in the configuration shown in FIG. 2 can be found in " J. Goodman: Analog Optical Signal and Image Processing, Handbook of Optics, Editor Michael Baas, Second Edition, McGraw-Hill, Inc., vol. 1, Chapter 30, pp. 30.3-30.8. "

Another explanation about how the lenses 202, 203 and the pinhole 204 perform spatial filtering in the preferred embodiment of this invention shown in FIG. 2, will be given in connection with ^• FIG. 4. The beam emerging from the polarization and phase modulator 201 that is also a phase-only grating may have the intensity distribution 226 in the focal plane of the lens 202. This distribution has a central maximum (zero order) 232 with intensity I₀, the first-order maxima 233 with intensities I₊₁ and respectively L_l5 the second-order maxima 234 with intensities I₊₂ and respectively L₂, and some other higher-order maxima, considered having negligible intensities in the preferred embodiment of this invention. For those knowledgeable in the art, it is obvious that I_+1} I₊₂, and their symmetric orders Li and I- decrease rapidly with respect to I₀, and their contribution to the output beam can be neglected. However, a limited number of higher diffraction orders can be used in the preferred embodiment of this invention. The pinhole 204 with a diameter D of the opaque plate 225 is centered on the maximum M₀ of the zero diffraction order to allow going through it only a part 227 of the maximum 232 (zero-order clipping). All the other maxima 233, 234 and higher are stopped by the opaque screen 225, the beam emerging from the pinhole 204 being free of higher diffraction orders. The pinhole 204 is in the focal plane of the collimating lens 203, which generates a collimated beam 220 with the intensity profile 227, free of diffraction orders 233, 234 and higher. Moreover, for those knowledgeable in the art it is an obvious fact that the embodiment of the converging lenses 202 and 203 having the same optical axis 226 perpendicular on PPM 201 and centered to it, with the focal planes of the lenses coincident and with the pinhole 204 centered on the optical axis on the common focal plane of the lenses, is a spatial filter for the beam 218, so the embodiment described above provides simultaneously a rejection of higher diffraction orders and eventually a smoothening (cutting-off irregularities in beam intensity) of the incident beam 216; a smaller value of D providing a smoother beam 220, with the corresponding loss in the intensity of the output beam.

According to the preferred embodiment of the invention, the polarization and phase modulator 201 is very similar from the electro-optical standpoint, with a regular dot matrix graphic LCD. It is very well known for those knowledgeable in the art that LCD elements must be driven with a voltage having a time changing polarity averaged on zero, to avoid freezing the position of their molecules into a certain position that may induce permanent damage to the device. It is also very well known by those knowledgeable in the art that a two-dimensional pattern needs to be updated on a graphic LCD. In the preferred embodiment of this invention, the update of the pixel activation pattern is made in row scan mode, by selecting one row at a time with a row driver 227 (FIG. 5) and activating all the necessary columns for the selected row with a column driver 228, then move to the next row and activate all the necessary columns for that row and so on until all rows are updated, the process begins again with the first row, the whole process being periodic. The duty cycle for updating is l/(number of columns), which can have a value of 1/32. The pattern with the pixels that need to be electrically activated is downloaded from a microcontroller 229 to the row driver 227 and to the column driver 228, using a serial communication link 230, and the link 231 between the two LC drivers. Into one possible embodiment of this invention, PPM 201 has 32 rows and 40 columns, driven by a row driver Philips PCF8578 and a column driver Philips PCF8579, supervised by a microcontroller Microchip PIC16LF872 using I²C serial communication bus. This embodiment does not restrict at all any other embodiments using LC modulators with a different number of pixels, driven by any type of LC drivers compatible with the PPM, under the supervision of any type of controller communicating with the LC drivers by any physical connection and using any communication protocol. In another embodiment of the present invention, the pattern of pixels in PAS and in PPS can be fixed and this can be eventually achieved by "freezing" the LC molecules in positions similar with those induced by internal electric fields generated by LC display drivers.

In the preferred embodiment of the invention described above, the polarization scrambling is achieved by inducing changes on the polarization state of some samples of the wavefront, their spatial distribution not necessarily varying in time, so the polarization scrambling is fully static. This preferred embodiment can generate monochromatic, spatially noncoherent and unpolarized light without time averaging that can be used as an illumination source for machine vision applications. The light beam with such properties produces a very low speckle with no amplitude or phase modulation (variation in time) of the beam. The same preferred embodiment of the present invention can also find applications to scramble the polarization of the light beam at the input of some optical components sensitive to polarization, used in dense wavelength division multiplexing (DWDM) technology for advanced optical communications.

In another preferred embodiment of this invention, the degree of scrambling for the polarization and phase of the beam can be adjusted by selecting a certain spatial distribution of MP's in PPS versus the total number of MP's that are sampling the beam, this selection being dependent on application.

The preferred embodiment of the present invention uses the electro-optical properties of the twisted nematic LC, but this does not limit the applicability of the invention to these types of LC only, any type of LC that induces changes in the polarization state could be used, such as cholesteric type or chiral type LC.

If some of the total number of MP's that are contained in the cross section of a light beam change the polarization state, the resulting polarization of the whole wavefront of the beam will be scrambled at any moment in time to an extent proportional to the ratio between the number of MP in PPS and the total number of MP that are sampling the beam, without the need to change in time this ratio. If, for some reasons, the average polarization on the wavefront of the beam needs to be time-variable, this requirement can be met also by changing the spatial distribution of electrically activated MP's (in PPS) with certain time dependence, either periodic or random. The extent of the polarization scrambling considered over the whole wavefront depends on how many MP's sample the beam in its cross section and how many of these sampling MP change the polarization state. Because the refractive index of LC is increasing when an electric field is applied to it, all electrically activated MP (VLC 0) effect a phase scrambling of the wavefront, simultaneously with a polarization scrambling.

A preferred embodiment of the present invention that generates an unpolarized laser beam to pump the optical amplifying medium of an Erbium-doped fiber amplifier (EDFA) is shown schematically in FIG. 6. The semiconductor pump laser 601 generates a divergent beam 609 that is transformed by the aspheric lens 602, into a collimated beam 610. The polarization and phase modulator 201 scrambles the polarization and phase of the beam 610 that becomes the beam 611 emerging from PPM with a random polarization and phase. The aspheric lens 604 with focal length fl, and the aspheric lens 606 with focal length f2 (fl>f2) have their focal planes coincident. This makes an afocal optical system that shrinks the collimated beam 611 to the collimated beam 614, having a smaller cross-section that fits the size and the numerical aperture of the fiber optic collimator 607. The fiber optic collimator further focuses the beam to the optical fiber 608. The beam 615 with random polarization carried by the optical fiber 608 does the optical pumping of EDFA, making the optical amplifier insensitive to the polarization state of the beam 609 generated by the pump laser 601. The pinhole 613 in the opaque plate 605 makes a spatial filtering of the beam in order to remove the high diffraction orders eventually generated by PPM 201. The present wavefront polarization and phase scrambler has a number of advantages in comparison with the prior art. The disclosed scrambler can work fully static, i.e. it scrambles the polarization and phase instantly, not on time averaging basis only, as the prior art does. It is programmable, being able to produce a certain degree of polarization scrambling either in the whole cross section of the beam, or only in selected parts of the beam, depending on the electrical activation of the modulation pixels. Another advantage is that the beam emerging from the disclosed wavefront polarization and phase scrambler also reduces the polarization-induced effects appearing during the beam propagation in monomode optical fibers used in optical communications.

While the invention has been particularly shown and described with respect to a preferred embodiment thereof, it will be understood by those skilled in the art that the foregoing and other changes in form and details may be made therein without departing from the spirit and scope of the invention.

Claims

CLAIMS What is claimed is:

1. A wavefront polarization and phase scrambler system comprising: a collimated light beam source, a liquid crystal matrix in optical alignment with said collimated light beam source, a spatial filter in optical alignment with said collimated light beam source.

2. The wavefront polarization and phase scrambler of claim 1, wherein said spatial filter further comprises: a first converging lens having a predetermined focal distance, a second converging lens having a predetermined focal distance in optical alignment with said first converging lens, and an opaque plate with a pinhole located between said first and second converging lenses along the optical axis of said collimated light source and in optical alignment with said first and second converging lens.

3. The wavefront polarization and phase scrambler of claim 2, wherein said first and second converging lenses are centered on a single optical axis and are separated by a separation distance, wherein the opaque plate is nonnal to said single optical axis having said pinhole centered to the axis and located at the common focal point of said first and second converging lenses.

4. The wavefront polarization and phase scrambler of claim 3, wherein said separation distance is equal to the sum of the focal distances for said first and second converging lenses.

5. A method of wavefront polarization and scrambling comprising: providing a light beam having a wavefront, wherein said wavefront has an incident polarization state, sampling said incident polarization state for a plurality of discrete portions of said wavefront, and simultaneously changing said incident polarization state of at least one of said discrete portions, and introducing a phase retardation to said discrete portions wherein said incident polarization is unchanged.

6. The method of wavefront polarization and scrambling of Claim 5 further comprising: filtering any high diffraction orders of said polarization scrambled light beam.

7. The method of wavefront polarization of Claim 5 further comprising: varying the change in polarization state of said light beam in at least one discrete portion with respect to time.

8. The method of wavefront polarization of Claim 5 further comprising: varying at a particular instant, the change in polarization state for each of said discrete portions with respect to the two dimensional face of said wavefront.

9. The method of wavefront polarization of Claim 5 further comprising: varying the change of said incident polarization state with respect to the wavelength of said wavefront.

10. The method of wavefront polarization of claim 5 including: providing a collimated light beam.