Electro-Optically Tuned Rejection Filter

This invention relates to tunable optical filters utilizing electro-optic materials.

A number of uses has been developed for solid state spectral filters whose transmission characteristics can be controlled solely by the application of an electric field. Research into such devices was based on the premise that the electro-optic effect in certain crystals could be employed to replace the mechanical tuning sections of a birefringent interference filter. The initial motivation for tunable filter development came from several potential applications, ranging from color projection systems to astronomical telescopes, for which simultaneous spatial and spectral resolution was required. Subsequently, the need for remote sensing capabilities and, in particular, space-based platforms, has motivated the development of devices which are both compact and free of the difficulties inherent in mechanically tuned spectral filters. In many of these applications restrictions on size, weight, and power consumption have made the use of compact, solid state tunable filters highly desirable. Tunable filters are incorporated in these systems to enhance the probability of target detection by identification of spectral signatures and to eliminate undesirable optical signals from the sensor.

Optical filters can be spectrally tuned by the use of optical materials whose refractive indices are modified by the application of an electric field. Such materials include those which exhibit linear (Pockels) or quadratic (Kerr) electro-optic effects in which an index of refraction change results from a non-linear response of the crystalline material. In addition, liquid crystal materials, in which large index changes result from an electric field-induced rotation of optically anisotropic molecules, may be employed. Among the tunable filter designs available in the art are electro-optic birefringent interference filters, Fabry-Perot devices, and electro-optic coupled-wave filters. These filters can be employed as passband filters, in which a narrow band of wavelengths is transmitted through the filter and the remainder of the incident light beam is rejected, or as stopband filters, which pass all wavelengths except for a narrow band of light. In the passband configuration, such filters can be made highly efficient, so that nearly 100% of the light at the desired wavelength is transmitted through the filter. Where it is necessary to exclude a particular wavelength, however, these filters are not as successful, since the percentage of light excluded by such a filter at stopband wavelengths cannot readily be made close to zero. Consequently, a technique which would permit the highly efficient separation of a particular wavelength from a beam of light would be welcome in the optical filter art.

It is a general objective of this invention to provide an improved technique for removing electromagnetic energy at a particular wavelength .lambda. from a beam of electromagnetic energy.

In one embodiment, an apparatus constructed according to this invention includes an electro-optically tunable interference filter for intercepting the beam, with the transmission of the filter controlled by the application of an electrical potential to the filter. A detector produces a signal in response to the electromagnetic energy transmitted by the filter, and a control circuit adjusts the electrical potential applied to the filter in accordance with the signal to maximize the transmission of electromagnetic energy at the wavelength .lambda..

In a more particular embodiment, the filter is a Fabry-Perot filter with an electro-optic crystal having a non-vanishing linear electro-optic coefficient, such as Z-cut LiNbO.sub.3. The control circuit includes a lock-in amplifier for receiving an electrical signal from the detector as an input. The output of the lock-in amplifier and a source of alternating current are supplied as inputs to a control amplifier, the output of which is applied to a high voltage amplifier. The output of the high voltage amplifier is then applied to the filter.

The technique of this invention also includes a method of sensing electromagnetic energy at a wavelength .lambda. from a beam of electromagnetic energy. An electro-optically tunable Fabry-Perot filter is placed in the path of the beam and a source of electrical potential is applied to the filter so that 2nd Cos .theta.=m.lambda., where n is the refractive index of the filter, d is the thickness of the filter, .theta. is the internal angle of refraction of the beam in the filter, and m is an integer.

Additional objectives, features, and advantages of the invention are detailed below in conjunction with the drawings, where:

FIG. 1 is a schematic illustration of an apparatus for removing electromagnetic energy at a particular wavelength from a beam of electromagnetic energy;

FIG. 2 is a plot of the calculated reflectance for a Fabry-Perot filter;

FIG. 3 is a plot of the calculated transmission for a Fabry-Perot filter;

FIG. 4 is a cross-sectional side view of the electro-optic filter shown in FIG. 1; and

FIG. 5 is a schematic diagram illustrating a particular embodiment of the control circuit shown in FIG. 1.

The present invention provides an improved technique for eliminating an intense narrowband optical signal from a broadband background. This new approach is particularly versatile because actual knowledge of the narrowband wavelength is not required.

FIG. 1 provides a schematic illustration of an apparatus for removing electromagnetic energy at a particular wavelength .lambda. from an incoming beam 10. The beam 10 is intercepted by an electro-optically tunable interference filter 12, which is placed so that the beam 10 makes an internal angle .theta. with the plane of the filter. The filter separates the beam 10 into a transmitted portion 14 and a reflected portion 16, with the transmission properties of the filter 12 being controllable by the application of an electrical potential to the filter. The transmitted portion 14 of the incoming beam impinges on a detector 18, which responds by producing a signal 20. Control circuit 22 receives the signal 20 and adjusts an electrical potential which is applied to the filter 12, as indicated by the line 24. In this manner, the signal 20, control circuit 22, and line 24 cooperate as a feedback loop to maximize the transmission through the filter 12 of light at the wavelength .lambda.. The reflected portion 16 may thus be efficiently separated from a selected narrow band component at the wavelength .lambda., the reflected portion then being available for further processing.

A number of different filter designs are available for use as tunable filters in a narrowband rejection device. The electro-optic Fabry-Perot filter, however, is the filter of choice for the preferred embodiment of the present invention because of its potential for rapid tuning speed, high rejection efficiency, and relative ease of fabrication. A Fabry-Perot filter is a comparatively simple structure consisting of two plane parallel partial reflectors separated by a suitable transparent medium. When the optical path length between the reflecting surfaces is an integer number of half waves, the structure becomes optically resonant, with a zero electric field intensity at the boundaries and energy coupled through the filter, ideally without loss. Other wavelengths not meeting the resonant condition are reflected.

The condition for maximum transmission in the Fabry-Perot filter is given by

where n is the index of the medium within the cavity, d is the mirror separation, .theta. is the angle of incidence within the cavity medium, m is an integer, and .lambda. is the free space wavelength. The spectral transmission T is

where

and

R is the reflectance of the mirrors, which are assumed to have identical properties. The spectrum for this filter consists of a series of transmission peaks at wavelengths and angles of incidence which satisfy equation (1). The spacing between these peaks is referred to as the free spectral range (FSR) and is given approximately by

As the reflectance of the Fabry-Perot mirrors is increased, the width of the individual transmission peaks will be reduced.

The finesse, F which describes the ratio of the FSR to the band-width, is given by ##EQU1## In the absence of absorption in the cavity or the reflectors, the peak transmission will be 100%, while the reflectance, integrated over a broadband spectrum, will be determined by F. FIG. 2, for example, is a plot of the calculated reflectance for a Fabry-Perot filter 200 .mu.m thick with a finesse of 5 and an FSR of 135 .ANG., while FIG. 3 is a plot of the calculated transmission spectrum for the same filter. Therefore, for a Fabry-Perot filter to be used effectively for band rejection, the narrowband signal must be rejected by transmission into an absorber, while the broadband signal may be substantially reflected for collection and analysis by an optical sensor. Over the broadband spectrum, the intensity of the reflected signal will be modulated by the presence of the transmission bands. The band structure, however, will typically occur on a scale which is much finer (.about.100 .ANG. or less) than that of any significant signal structure which would be found in a typical optical sensor system.

The filter of the present invention is tuned by the application of a DC voltage across the filter. FIG. 4 is a cross-sectional side view of the filter 12 of FIG. 1, which includes an electro-optic crystal 26, sandwiched between an incident reflecting layer 28 and an exit reflecting layer 30. A pair of transparent electrodes 32 and 34 contact opposite faces of the crystal so that an electric potential can be applied across the crystal. Linear electro-optic materials having specific crystal symmetries must be used in such a design. In order for the filter to operate without polarizers, for example, the direction of light propagation must be along the uniaxial (C-) axis, where the ordinary and extraordinary modes are degenerate. The direction of the applied electric field must also be in this direction, so that the polarization mode degeneracy is not lifted. This condition alone is not sufficient, since the degeneracy can still be lifted in this geometry for certain crystal classes. In general, when an electric field is applied to a noncentrosymmetric crystal, the refractive index ellipsoid is represented by ##EQU2## where r.sub.ijk are the linear electro-optic coefficients, x.sub.i and x.sub.j are the refractive indices, and the E.sub.k are the applied electric fields. By symmetry, r.sub.ijk =r.sub.jik, so that ij subscripts are usually contracted using the equivalent notation 1=11, 2=22, 3=33, 4=23, 5=13, and 6=12. In general, many of the r.sub.mk are identically zero, as dictated by crystal symmetry.

If an electric field is applied along the C-(3) axis and light propagates in the same direction, the index ellipsoid cross section becomes

If r.sub.13 =r.sub.23, then the field induced change in the refractive index will be independent of the polarization direction. Assuming the index change .DELTA.n to be small,

Among the various crystal classes which have nonvanishing and equal coefficients r.sub.13 and r.sub.23, the possible groups are then 4, 4 mm, 3, 3 m, 6, or 6 mm. These groups include most of the ferroelectric oxides, such as LiNbO.sub.3 and Sr.sub.x Ba.sub.1-x Nb.sub.2 O.sub.6.

The voltage required to tune a Fabry-Perot filter through one complete FSR is that voltage which induces a change of the optical path length by one half wave. This half wave voltage is given by

An important feature of the present invention is the capability for automatically tuning to the wavelength of a narrowband source. This feature is possible because of the non-zero transmission of a low finesse Fabry-Perot filter for any wavelength within its FSR and because of the use of closely spaced transmission peaks. A detector sensing the transmission of the filter can detect the presence of an intense narrowband source even if the filter is not exactly tuned to the wavelength of that source. By modulating the voltage applied to the filter and detecting the phase of the filter's transmission, the control electronics may tune the filter to maximize the transmitted signal. This approach, which eliminates the need to determine the wavelength of the source, also avoids the strict temperature control requirements which are usually encountered when operating such filters.

The transmission characteristics of an electro-optic Fabry-Perot filter are given by a function T(.lambda.,V) where .lambda. is the wavelength and V is the voltage applied to the etalou. If the filter is illuminated by a monochromatic source .lambda..sub.L, then the transmission of the filter in the vicinity of a transmission maximum is a function of voltage, which can be expanded as ##EQU3## If the voltage is given by

where a <<V.sub.o, then T(.lambda..sub.L, V) becomes ##EQU4##

In the frequency domain, as can be seen from Equation 11, the transmission includes a component at frequency .omega. with amplitude a (.differential.T/.differential.V) and a second harmonic with frequency 2.omega. and amplitude a.sup.2 /4(.differential..sup.2 T/.differential.V.sup.2). Thus, by monitoring the first and second harmonic components of the transmission, it is possible to determine the first and second derivatives, with respect to voltage, of the transmission function. By determining the voltage which makes .differential.T/.differential.V=0 and .differential..sup.2 T/.differential.V.sup.2 <0, the filter transmission at the monochromatic source wavelength .lambda..sub.L can be maximized.

Illustrated in FIG. 5 is a particular embodiment of the control circuit 22 of FIG. 1 which uses the first derivative signal to generate a DC correction signal. In this approach, the signal from the detector 18 (shown in FIG. 1) is applied to a lock-in amplifier 36, which is tuned to the frequency .omega.. The output of the amplifier 36 provides the first derivative .differential.T/.differential.V, which is integrated by an operational amplifier 38 and summed, at junction 40, with a DC bias signal from an amplifier 42 and an AC modulation a Sin .omega.t.

Depending upon the sign of the first derivative, a signal integrator 44 generates a correction signal which is used as the input to a high voltage amplifier 46, which in turn is connected to the electrodes of the filter 12 (shown in FIG. 1). The correction signal continues to increase or decrease, depending upon the sign of .differential.T/.differential.V, until .differential.T/.differential.V becomes zero, and the voltage is then maintained by this circuitry at a constant level. In this manner, the circuit compensates automatically for any shift from the transmission maximum, whether caused by a change in the source wavelength or a change in the filter temperature. The presence of a broadband signal should not have a significant effect on the performance of this circuit, assuming that the broadband signal structure is not within the resolution of the filter. The finesse of the filter should be made relatively low in order to provide sufficient transmission of a narrowband signal at any point within the FSR of the filter.

In one particular example of a filter which was constructed according to the present invention, an electro-optic crystal was fabricated from LiNbO.sub.3 with a thickness of 0.5 mm and measuring 15 mm on each side. A three layer germanium and zinc sulfide quarter-wave stack of coatings was applied to each face of the crystal by electron beam deposition to provide both the reflecting layers and the transparent electrodes. Gold leads were attached using silver filled epoxy, and the filter was mounted into a glass filled epoxy insulating holder with a small amount of silicon rubber cement. This device was used to demonstrate the concept of this invention for a narrowband source at a wavelength of 3.39 .mu.m in the infrared. Another device, using silver reflectors, was used to demonstrate the operation of this filter at 6328 .ANG. in the visible spectrum.

In conclusion, the rejection filter of this invention is of a relatively simple design, with straightforward construction. Among the outstanding advantages of this filter are its immunity to thermal variations and the lack of a requirement for a mechanically stable enclosure. In addition, complicated tuning algorithms are not required and it is not necessary to determine the wavelength of the narrowband source which is to be rejected. Although the particular embodiments discussed above are exemplary, modifications and additional embodiments may be apparent to those skilled in the art. Consequently, the full scope of the invention should be measured by the claims which follow.