JPH0361174B2 - - Google Patents

Description

BACKGROUND OF THE INVENTION This invention relates generally to frequency shifters, and specifically to optical frequency shifters.

More particularly, the present invention provides a waveguide suitable for use in an optical rotation detection system that uses an electro-optic rotating half-wave plate to shift the frequency of the optical input to a frequency suitable for angular rotation rate detection. It relates to optical frequency shifters.

Frequency shifters can be used to implement optical rotation detection systems with the necessary bandwidth for aircraft navigation. Such a system must be capable of rotating at rotational speeds as low as 0.01 degrees per hour and as high as 1000 degrees per second. The ratio of the upper and lower limits to be measured is approximately 10 ⁹ .

Optical fiber ring interferometers have been found to be particularly useful for rotation detection. Fiber optic ring interferometers typically include a loop of optical fiber material having a light wave counterpropagating therethrough. After traversing the loop, the counter-propagating waves combine so that they interfere to form the output signal. The output signal and intensity depend on the relative phase of the two waves. Rotation of the loop creates a relative phase difference between the two waves according to the well-known Sagnac effect. The amount of phase difference is a function of the angular velocity of the loop, so that the optical output signal produced by the interference of the counterpropagating waves varies in intensity as a function of the rotational velocity of the loop.

Rotation detection is accomplished by detecting the optical output signal and processing the signal to determine the rotation rate. A frequency shifter can be included in a device that processes the optical output signal to determine the value of the sagnac phase shift between the two waves.

The frequency shifter must have a bandwidth comparable to that of the appropriate light source used to introduce the counterpropagating light wave into the loop. Superluminous diodes with coherent light output in a wavelength band of approximately 100 nm have been found to be suitable light sources for optical rotation detection systems. Therefore, a suitable frequency shifter is
It must be possible to shift the input optical beam above or below the carrier wavelength by at least 50 nm. Previously used frequency shifters have too narrow a bandwidth for aircraft guidance systems.

Typical optical frequency shifters use acoustic waves to form a moving grating. The light whose frequency is to be shifted is applied to an acoustic wave. The regions of alternating compression and rarefaction in the acoustic wave act as a diffraction grating with a grating spacing equal to the wavelength of the acoustic wave. The grating moves at acoustic speed and shifts the frequency of the incoming optical beam due to the Doppler effect. The acousto-optic frequency shifter has a bandwidth of only about 2 = 0.2 nm. Acousto-optic frequency shifters typically have an efficiency of about 80%, but the efficiency varies with frequency and amount of frequency shifter. The frequency shifter of the optical rotation detection system is invariant to changes in the amount of frequency shift required.

SUMMARY OF THE INVENTION The present invention provides a frequency shifter that overcomes the shortcomings of conventional frequency shifters in optical rotation detection systems. The frequency shifter of the present invention has a switching time of about 1 nanosecond and provides a means for shifting the frequency of an input optical signal by values ranging from 0 to about 1.3 GHz. Efficiency varies by about 1dB over a 1GHz bandwidth. Therefore, the frequency shift covers the entire range of interest in optical rotation detection systems.

The invention includes an optical waveguide and a plurality of electrodes formed in a substrate of electro-optically active material, such as lithium niobate. The electrodes are arranged to apply two orthogonal electric fields along the length of the optical waveguide. The phases of these electric fields are controlled such that the combination of two orthogonal electric fields becomes a rotating electric field vector. Application of an electric field to electro-optically active lithium niobate changes its refractive index. By appropriate selection of the electrode length, spacing and voltage applied thereto, the optical waveguide behaves as a half-wave retardation plate rotating at half the frequency of the voltage applied to the electrodes.

The invention further includes a quarter-wave plate at each end of the half-wave plate to ensure that the optical signal input to the half-wave plate is circularly polarized. The circular polarizer input to the half-wave plate is shifted in frequency from the carrier frequency by a value of f, where f is the frequency of the voltage applied to the electrodes. The frequency shift is
Depending on the orientation of the circularly polarized input wave and the direction of rotation of the half-wave plate, it produces sidebands with frequencies greater or less than the carrier frequency.

In a preferred embodiment, the first electrode cooperates with the ground electrode to create an electric field through the optical waveguide and parallel to the surface of the substrate. A second electrode cooperates with the ground electrode to create an electric field through the optical waveguide and normal to the surface of the substrate. The magnitude of the voltage applied to the electrodes is adjusted to ensure that the two electric fields are of equal magnitude. Preferably, the magnitude of the electric field is chosen to produce a half-wavelength delay in the input optical signal at the selected wavelength.

The frequency shifter of this invention also has application as a sensor and local oscillator in optical communication systems.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Frequency Shifter Device Referring to FIG. 1, a frequency shifter 10 according to the present invention includes a substrate, preferably constructed from an electro-optically active material, such as lithium niobate. For ease of illustration, substrate 12 is shown as having a conventional rectangular surface. Optical waveguide 14 may be formed by doping a portion of substrate 12, as described below, and extends the length of the substrate. For convenience, optical waveguides 14 are shown along parallel lines and equidistantly spaced from a pair of opposite sides of substrate 12 . The optical waveguide 14 shown in FIGS. 1, 7E, and 11 preferably has a generally semicircular cross-section with a curved side 11 and a flat side 13 that spans the plane of the substrate 12. .

A plurality of electrodes 16, 18 and 20 are attached to a central portion of substrate 12. electrode 16
is preferably centered across the flat surface 13 of the semicircular optical waveguide 14, as shown in FIGS. 1-3. As shown in FIGS. 1 and 2, electrodes 18 and 20 are spaced apart from and generally flush with electrode 16. As shown in FIGS. Electrodes 16, 18 and 20 are preferably rectangular of approximately equal length L and are positioned in parallel alignment with their generally collinear ends. Preferably, electrodes 18 and 20 are each spaced a distance d from central electrode 16. In a typical construction of frequency shifter 10, length L is approximately 1.0 cm and separation distance d is approximately 10 μm.

As shown in FIG. 1, electrode 18 is preferably connected to ground potential. Signal generator 22 directs power to electrodes 16 to 22 via power divider 24.
supply to. Power divider 24 includes means well known in the art to provide two output voltage signals that are preferably approximately equal in magnitude and 90 degrees out of phase. Power divider 24 has a first output terminal 26 that connects end 2 of electrode 16 via attenuator 30.
Connected to 8. Output 32 of power divider 24 is connected directly to electrode 20 .

A resistor 36 is connected between an end 38 of electrode 16 and ground and a resistor 40 is connected between an end 42 of electrode 20 and ground. Resistors 36 and 40 connect the electrical loads of electrodes 16, 18 and 20 to signal generator 2.
It works to match the impedance of 2. Resistors 36 and 40 are typically each about 50 ohms.

A pair of electrodes 44 and 46 are positioned on opposite sides of optical waveguide 14 and are connected to electrodes 16, 18 and 2.
0 along its length. Electrode 46 is shown as grounded, and electrode 46 is connected to a power source 48, which is preferably an adjustable DC power source. A pair of electrodes 50 and 5
2 are positioned on opposite sides of optical waveguide 14 from electrodes 44 and 46 to opposite ends of electrodes 16, 18 and 20. Electrode 52 is grounded and electrode 5
0 is connected to a power supply 54 similar to power supply 48.

Referring to FIG. 2, applying a positive voltage to the electrode 20 produces an electric field indicated by E _Y , which is applied to the electrode 20.
Electrodes 20 and 18 are arranged to extend from 0 through optical waveguide 14 to electrode 18 . Conductor-dielectric interfaces, e.g. electrodes 16, 18 and 2
It is a well-known boundary condition from electromagnetic field theory that at the interface between 0 and 0, the electric field is perpendicular to the conductor. Therefore, the electric field near electrodes 18 and 20 is perpendicular thereto. electrodes 18 and 20
are coplanar, so the electric field E _Y in the region of the optical waveguide 14 is essentially parallel to the surface of the substrate 12.

Referring to FIG. 3, between electrodes 16 and 18
The electric field, designated _EX , is also perpendicular to the interface between substrate 12 and electrodes 16 and 18. Since the electrode 16 is directly above the flat surface optical waveguide 14, the electric field _EX is essentially perpendicular thereto. Electrodes 16, 18 and 20 thus provide a means of creating two mutually perpendicular electric fields in optical waveguide 14. The spacing between electrodes 16 and 18 is smaller than that between electrodes 18 and 20. To form the frequency shifter 10, an electric field
It is desirable that E _X and E _Y have equal magnitude. An attenuator 30 is used to adjust the voltage applied to the electrodes 16 to equalize the electric field components.

As a result of the above structure of frequency shifter 10, equal magnitude orthogonal electric field components that are 90° out of phase are applied to optical waveguide 14. Such electric field components are transmitted to electrodes 16, 18 and 20 by signal generator 22.
has a composite, which is an electric field vector rotating with each frequency equal to the frequency of the signal applied to it. Because the doped lithium niobate optical waveguide is electro-optically active, the rotating electric field causes a rotational change in the refractive index of the optical waveguide 14. By appropriately selecting the electric field magnitude, electrode length and electrode spacing, the change in refractive index will cause the portion of the optical waveguide 14 covered by the electrode 16 to behave as a rotating half-wave plate. .

Power supplies 48 and 52 are adjusted such that power supply pairs 44, 46 and 50, 52 each operate as a quarter wave plate.

Theory and Method of Operation The elliptic refractive index of uniaxially crystalline lithium niobate is given by: (1/n ² _p ) x ² + (1/n ² _p ) y ² + (1/n ² _e ) z ² =1(1) where n is the normal refractive index and n _E is the extraordinary refractive index. If the laser light propagates along the Z-axis, only one permissible refractive index depends on the two orthogonal polarization states propagating in the optical waveguide 14, which means that the externally applied If no electric field is present, it means that there is no birefringence in the optical waveguide 14. The elliptical index of refraction is deformed and rotated when an external electric field is applied to the crystal. Therefore, applying an external electric field to the crystal changes both the refractive index of the crystal and the polarization of laser light propagation.

The new elliptic refractive index calculated from the lithium niobate electro-optic sensor is expressed as: (1/n ² _p − γ ₂₂ E _y ) x ² + (1/n ² _p + γ ₂₂ E _y ) y ² + (1/n ² _e )z ² +2γ ₄₂ E _y yz + 2γ ₄₂ E _x xz−2γ ₂₂ E _x xy=1 (2) where y _IJ is the linear electro-optic coefficient and E _x and E _y are respectively The applied electric field in the x and y directions. Assuming that the propagation direction is the z-axis, z=
0, and equation (2) is as follows: (1/n ² _p −γ ₂₂ E _y )x ² +(1/n ² _p +γ ₂₂ E _y ) y ² −2γ ₂₂ E _x xy=1 (3) The coordinate system is then transformed into a principal axis system and eliminated in the last term of equation (3) to obtain an elliptic representation, which then contains products for different variables x and y. There are no terms. The desired coordinate transformation is x- through the angle θ
This is the rotation of the coordinates in the y plane, and we get the following equation: (1/n ² _p −γ ₂₂ E _y cos2θ−γ ₂₂ E _x sin2θ) x′ ² + (1/n ² _p + γ ₂₂ E _y cos2θ + γ ₂₂ E _x sin2θ)y'=1 (4) The rotation angle is given by the following formula: θ=+1/2tan ^-1 (E _x /E _y ) (5) Therefore, the refractive index of the optical waveguide 14 is It is given by: n _x ′=n _p −1/2n _p (γ ₂₂ E _y cos2θ + γ ₂₂ E _x sin2θ) (6) n _y ′=n _p −1/2n ³ _p (γ ₂₂ E _y cos2θ +γ ₂₂ E _x sin2θ) (7) The different expressions of Equations 6 and 7 for the refractive index in the new principal axis coordinate system show that birefringence occurs when an electric field is applied to an electro-optically active material. . The light wave propagating through the birefringent optical waveguide 14 undergoes a phase delay of the following equation: T=(2π/λ)(n _x ′−n _y ′)・L (8) T=2πn ³ _p γ ₂₂ (L /λ)(E _y cos2θ +E _x sin2θ (9) where L is the interaction length of electrodes 16, 18 and 20 and λ is the optical signal vacuum wavelength.

The half-wave voltage V _o is defined as the voltage required to achieve a phase delay of radians. Therefore, if we set = and E _x = E _y = V _o /d, then
The half-wave voltage is: V = λd (2n ³ γ ₂₂ L) ^-1 (10) If the electric field components are of the same amplitude E required to produce a composite rotating electric field vector in the optical waveguide 14. However, if the waveform is a sine wave whose angle differs by 90°,
The electric field component can be expressed as follows.

E _x = Esinωt (11) E _y = Ecosωt (12) Using equations (11) and (12) in equation (5) for the rotation angle, the following equation is obtained: θ=1/2tan ^{- 1} (sinωt/kcosωt) (13) θ=1/2tan ^-1 tanωt (14) θ=1/2ωt (15) Therefore, the principal axis system for the state in which propagation in the optical waveguide is allowed is The rotation angle is determined by the frequency of the signals applied from signal generator 22 to electrodes 16, 18 and 20. Differentiating Equation 15 with respect to time shows that the angular velocity of the rotating wave plate is as follows: θ = 1/2ω (16) If the rotating electric field with a voltage corresponding to the half-wave voltage A frequency shifter 10 responsive to an input optical signal having a frequency f when applied to the wave tube 14
The output of is ±fm from the input, or carrier, signal.
Here is the optical signal that is frequency shifted by
fm is the frequency of the applied voltage from the signal generator 22. The frequency of the sidebands depends on the direction of rotation of the wave plate relative to the direction of circular polarization of the input waveform. If the voltage is equal to the half-wave voltage, there is also complete conversion of the carrier into sidebands. FIG. 6 shows the relative optical power of the carrier and sidebands as the voltage changes from zero to the voltage required for half-wavelength delay.

FIG. 4 shows the frequency shifter 10 when the rotation direction of the wave plate is clockwise or to the right when viewed from the left.
FIG. A left-handed circularly polarized input signal from the left with frequency f ₀ produces a right-handed circularly polarized sideband with frequency f ₀ +f. A right-handed circularly polarized signal input from the right has frequency f ₀ +fm and is shifted to a sideband with left-handed circularly polarized light. FIG. 5 shows a sideband shifted from the left with left-handed circular polarization and frequency f ₀ −
FM right-handed circularly polarized input is shown. The left-handed circularly polarized input from the right has right-handed circularly polarized light and is shifted to a sideband with frequency f ₀ -fm. If the back-propagating beam is
For example, in the case of optical gyroscopes,
If input to the frequency shifter 10, the rotational orientation of the wave plates for the two beams are mirror images of each other so that one wave is shifted up in frequency and the other shifted down.

The bandwidth of frequency shifter 10 is given by: f _3db =1/(πRC) (17) where C is the capacitance related to electrodes 16, 18 and 20 and R is the capacitance related to electrodes 16 and 20. 18 and ground potential. Typically,
The capacitance is approximately 1pf, so the terminating resistor is
Resulting in a bandwidth of 1.3GHz. Frequency shifter 1
0 drive power is given by: P=V ² /(2R) (18) If 100% conversion from carrier to sideband is desired, drive power is given by: P= V ² 〓/(2R) (19) It must be applied to electrodes 16, 18 and 20, where V _o is the 10th
It is given by Eq. The wavelength λ is 6328 Å. Electrode 1
The distance d between 6 and 18 is approximately 10 μm. Lithium niobate has a refractive index of 2.29 and its linear electro-optic coefficient γ ₂₂ is 3.4×10 m/V. Electrodes 16, 18 and 20 are each approximately 1.0 cm in length. The above parameters provide a half-wave voltage of about 7.7 volts, resulting in a drive power of about 593 mw per electrode, which are connected to signal source 22. Therefore, approximately 1 watt of total power is required for frequency shifter 10 operation. The above parameters are exemplary only and are not critical to the operation of frequency shifter 10. The power requirement is
It can be reduced by reducing the distance between electrodes 16, 18 and 20 and by increasing their length.

Preferably, the optical wave input to the half-wave plate is circularly polarized. A signal with linear polarization is divided into four
Passing through one of the wave plates 56 or 58 ensures that the input to the half wave plate has the desired circular polarization. If, for example, the input to quarter-wave plate 56 is linearly polarized such that its polarization is 45° from the positive X-axis, then
A quarter wave plate 56 introduces a 90° phase difference with waveforms of equal polarization along both the X and Y axes.

Construction Methods FIGS. 7 and 9-11 illustrate process steps that may be used to construct frequency shifter 10 of FIG.

Figures 7A-7E illustrate forming optical waveguide 14 in a substrate. At first,
Substrate 12 is coated with a layer of photoresist 70 . Glass plate 71 called photomask
is placed on top of the photoresist layer 70. Board 71
is prepared using standard optical reproduction techniques to create the desired pattern, where portions 72 and 74 are opaque and spaced apart leaving an elongated rectangular portion 76. The arrow in FIG. 7A indicates the glass plate 71.
The ultraviolet (UV) light incident on the photoresist layer 70 and the exposed portion 76 of the photoresist layer 70 are shown. Portions 72 and 74 are opaque to UV light and therefore
UV light is applied to exposed portions 76 of photoresist layer 70.
incident only on Referring to FIG. 7B, by placing the photoresist portions 76 in a suitable developer, only the pair of photoresist portions 77 and 78 remain attached to the substrate 12, and these portions are attached to the glass plates 72 and 78, respectively. It's right next to 74.

Referring to FIG. 7C, a titanium layer 80 is formed on the substrate 12 from which photoresist portions 76 have been removed. The titanium layer 80 is formed by a process such as vapor deposition to form a thin metal film.
It can be formed by any suitable method. Other photoresist portions 77 and 78 also have titanium layers 82 and 84 formed thereon, but layer 80 has essentially parallel sides to photoresist layers 77 and 78.
sharply defined by the edge of

Photoresist portions 77 and 78 are removed by placing substrate 12 in a solvent such as acetone.
is removed and a seventh layer is placed on the substrate 12.
As shown in Figure D, only a well-defined layer 80 of titanium is left. Layer 80 has a generally rectangular cross-section when viewed from the end, as shown in Figure 7D. Substrate 12 having titanium layer 80 thereon is placed in a high temperature oven, as is well known in the art, and diffusion of Ti ⁺⁺ ions occurs in substrate 12, forming the normal semicircular shape shown in FIG. 7E. sufficient time for the waveguide 14 to form;
Fired.

Frequency shifter 10 is coupled between a pair of ends 88 and 90 of optical fiber material, one end;
For example, it receives an optical signal from end 88, shifts the frequency of the signal, and couples the frequency shifted output to the other end 90. FIG. 8 illustrates the major considerations involved in joining optical waveguide 14 and fiber end 88 with core 91 and cladding 92. It is typically desirable to minimize insertion loss caused by slicing frequency shifter 14 between ends 88 and 90. In this application, insertion loss is defined as the ratio of the power carried by the sideband signal to the power in the input carrier signal. Optical waveguide 14
and core 90 are dimensioned to maximize the area of overlap therebetween to minimize insertion loss. A small portion 94 of the fiber extends beyond the linear end of the semicircular cross section of the optical waveguide 14, and a portion 98 of the optical waveguide 14 extends beyond the circular periphery of the core 91. For a given diameter of core 91, careful control of the process used to form optical waveguide 14 will reduce the insertion loss.
Limited to 5dB or less. Insertion loss can be maintained at a relatively low value because the distribution of optical power in the fiber 88 and optical waveguide is a Gaussian distribution of similar shape as shown in FIG.

The diameter of cladding 92 is gradually larger than that of core 91, which facilitates manipulation of fiber 88 to align core 90 with optical waveguide 14. Referring to FIG. 10, the fibers 88 are conveniently disposed in V-shaped grooves 100 in a substrate 102, which may include a plurality of alignment ridges 104. As shown in FIG. 11, substrates 12 and 102 can be positioned to bring core 91 and optical waveguide 14 into the desired alignment for butt-coupling. A suitable adhesive, such as an epoxy resin, may be used to maintain the butt bond between fiber 88 and optical waveguide 14.

[Brief explanation of drawings]

FIG. 1 is an isometric view of a frequency shifter of the present invention including electrodes formed on a substrate and includes a schematic diagram of the electrical circuitry used to apply voltages to the electrodes. FIG. 2 is a cross-sectional view of the electrode of FIG. 1 showing an electric field parallel to the substrate. FIG. 3 is a partial cross-sectional view taken along line 2--2 of FIG. 1 showing the electric field perpendicular to the substrate. Fourth
The figure is a schematic diagram of a rotating half-wave plate. FIG. 5 is a second schematic diagram of a rotating half-wave plate. FIG. 6 is a graph showing the carrier and sideband percentages of the frequency shifter of FIG. 1 as a function of the voltage applied to the electrodes. 7A-7E illustrate steps that may be used to fabricate the frequency shifter of FIG. 1. FIG. 8 shows the overlap of the optical waveguide formed in the substrate of FIG. 1 and the core of the optical fiber. FIG. 9 diagrammatically shows the distribution of power in the optical fibers and optical waveguides of FIGS. 1 and 8. FIG. FIG. 10 shows an apparatus and method for positioning a fiber optic waveguide for butt coupling to the optical waveguides of FIGS. 1 and 8. FIG. FIG. 11 is a front view illustrating butt coupling of the optical fiber waveguide of FIG. 10 to the optical waveguide of FIG. 1; In the figure, 12 is a substrate, 14 is an optical waveguide, 16, 18, 20 are electrodes, 22 is a signal generator, 24 is an output divider, 30 is an attenuator, 36
and 40 is a resistor, 48 is a power supply, 50 and 52
is an electrode.

Claims (1)

Claims: 1. A frequency shifter that receives an optical carrier signal having a carrier frequency and provides a sideband output shifted in frequency from the carrier frequency, the optical signal being disposed within a substrate formed of an electro-optically active material. an optical waveguide for guiding the optical waveguide; a first electrode attached to a surface of the substrate and covering a length of the optical waveguide; and proximate the first electrode on a first side of the waveguide. a second electrode attached to the surface of the substrate; a third electrode attached to the surface of the substrate proximate the first electrode on a second side of the waveguide; and rotation of the optical waveguide. means for applying a first voltage between the first electrode and the second electrode to form a first electric field across the wave plate region; and means for applying a first voltage between the first electrode and the third electrode. applying a second voltage and having the same magnitude as the first electric field;
forming a second electric field through a rotating waveplate region perpendicular to the electric field of the optical waveguide, the first voltage and the second voltage being such that the resultant electric field rotates relative to the optical waveguide. A quarter-wavelength retardation plate formed in the optical waveguide with a phase difference of 90° to adjust the polarization of the optical signal thereto for input into the rotating waveplate region. an apparatus for providing a circularly polarized optical signal; a signal generator; and a first apparatus connected to receive the output of the signal generator;
a power divider providing an output voltage to a first electrode and a second output voltage to a second electrode, the first and second output voltages having a 90° phase difference; and a device for adjusting the relative magnitudes of the first and second voltages such that the electric fields of the first and second voltages have equal magnitudes. 2. A method for shifting the frequency of an optical carrier signal from a carrier frequency to a sideband frequency, the method comprising: forming an optical waveguide within an electro-optically active material to direct the optical signal within the electro-optically active material; forming a first electrode on a surface of the substrate over the length of the optical waveguide; and on a surface spaced apart from the first electrode on a first side of the waveguide. forming a second electrode on a second side of the waveguide and a third electrode on a spaced apart surface from the first electrode; and passing through the optical waveguide. applying a first voltage to a first electrode to form a first electric field; and passing a first voltage through an optical waveguide having the same magnitude as the first electric field and perpendicular to the first electric field. applying a second voltage between the first and third electrodes to form an electric field of two, the first and third electrodes so that the resultant electric field rotates relative to the optical waveguide. and a step in which the second voltage has a phase difference of 90° and a quarter retardation plate formed in the optical waveguide to adjust the polarization of the optical signal thereto and input to the rotating waveplate region. receiving the output of the signal generator and applying a first output voltage to the first electrode and a second output voltage to the second electrode, the output voltage being 90°; and adjusting the relative magnitudes of the first and second voltages to equalize the magnitudes of the first and second electric fields. Method.