1. Field of the Invention
The present invention relates to optical communication equipment and, more specifically, to optical amplifiers.
2. Description of the Related Art
Optical communication systems employ optical amplifiers, e.g., to compensate for signal attenuation in optical fibers. One type of amplifier that may be used in a fiber-based optical communication system is an optical fiber amplifier (OFA). Many current optical communication systems employ erbium-doped and/or Raman fiber amplifiers. These amplifiers are examples of phase-insensitive amplifiers (PIAs). More specifically, a PIA produces signal gain that is independent of the signal phase. In contrast, a phase-sensitive amplifier (PSA) produces signal gain that does depend on the signal phase. Potential advantages of using PSAs instead of or in addition to PIAs in optical communication systems include, but are not limited to, noise reduction, the reduction of noise-induced frequency and phase fluctuations, dispersion compensation, and suppression of modulational instability. A more detailed discussion of these and other potential advantages of PSA use in optical communication systems can be found, e.g., in the following publications: (1) R. Loudon, “Theory of Noise Accumulation in Linear Optical-Amplifier Chains,” Institute of Electrical and Electronic Engineers (IEEE) Journal of Quantum Electronics, vol. 21, pp. 766-773 (1985); (2) H. P. Yuen, “Reduction of Quantum Fluctuation and Suppression of the Gordon-Haus Effect with Phase-Sensitive Linear Amplifiers,” Optics Letters, vol. 17, pp. 73-75 (1992); (3) Y. Mu and C. M. Savage, “Parametric Amplifiers in Phase-Noise-Limited Optical Communications,” Journal of the Optical Society of America B, vol. 9, pp. 65-70 (1992); (4) R. D. Li, P. Kumar, W. L. Kath, and J. N. Kutz, “Combating Dispersion with Parametric Amplifiers,” IEEE Photonics Technology Letters, vol. 5, pp. 669-672 (1993); and (5) W. Imajuku and A. Takada, “Reduction of Fiber-Nonlinearity-Enhanced Amplifier Noise by Means of Phase-Sensitive Amplifiers,” Optics Letters, vol. 22, pp. 31-33 (1997), the teachings of all of which are incorporated herein by reference. However, PSAs suitable for use in fiber-based optical communication systems are not yet sufficiently developed.
Problems in the prior art are addressed, in accordance with the principles of the present invention, by a method of and device for generating an amplified optical signal directly in an optical fiber by way of phase-sensitive amplification based on one or more four-wave mixing (FWM) processes. In one embodiment, an input signal and two pump waves are applied to a highly nonlinear fiber (HNLF). The input signal is amplified in the HNLF due to energy transfer from the pump waves to the input signal via a degenerate phase-conjugation (PC) process. The amplified signal is then separated from the pump waves by an optical filter placed at the end of the HNLF.
In another embodiment, an input signal and first and second pump waves are applied to a first HNLF to generate, via a Bragg scattering (BS) process, an idler signal corresponding to the input signal. A first optical filter placed at the end of the first HNLF then filters out the second pump wave while passing the first pump wave and the input and idler signals to a second HNLF, which also receives a third pump wave. In the second HNLF, the first and third pump waves and the input and idler signals interact via a non-degenerate PC process, which results in energy transfer from the pump waves to the input signal. The amplified input signal is then separated from the pump waves and the idler signal by a second optical filter placed at the end of the second HNLF.
Other aspects, features, and benefits of the present invention will become more fully apparent from the following detailed description, the appended claims, and the accompanying drawings in which:
Reference herein to “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the invention. The appearances of the phrase “in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
Optical Parametric Amplification
As known in the art, optical parametric amplification (OPA) is a process that produces a tunable coherent optical signal via nonlinear optical processes, in which, typically, one or two pump-wave photons are converted into two new photons with conservation of photon energy and momentum. The waves corresponding to the two new photons are usually referred to as a signal and an idler. An excellent review of the fundamentals of OPA can be found in a book by G. P. Agrawal, “Nonlinear Fiber Optics” (Third Edition), Academic Press, 2001.
2ωP=ωs+ωi (1)
where ωP, ωs, and ωi are the frequencies of the pump, signal, and idler photons, respectively. The nonlinear medium, in which the FWM process occurs, is characterized by a nonlinearity coefficient (γ) and a set dispersion coefficients, each of which is frequency dependent. The frequency or wavelength at which the second-order dispersion coefficient equals zero is referred to as the zero-dispersion frequency (ω0) or wavelength (λ0). The regions in which the second-order dispersion coefficient is positive and negative are referred to as the normal dispersion region and the anomalous dispersion region, respectively.
The degenerate FWM process of
The following describes the FWM processes that lead to the frequency structure shown in
2ω1=ω1−+ω1+ (2)
A Bragg scattering (BS) process produces a second idler sideband at frequency ω2− according to Eq. (3):
ω1−+ω2=ω1+ω2− (3)
A phase-conjugation (PC) process produces a third idler sideband at frequency ω2+ according to Eq. (4):
ω1+ω2=ω1−+ω2+ (4)
In addition, each of the three idler sidebands is coupled to the other two idler sidebands by an appropriate FWM process, i.e., MI, BS, or PC, that can be expressed by an equation analogous to one of Eqs. (2), (3), or (4).
In addition to the sidebands illustrated in
Phase-Sensitive Amplification in a χ(2) Medium
A representative prior-art scheme for phase-sensitive amplification (and squeezed-state generation) is described in a book by R. Loudon, “The Quantum Theory of Light,” 3rd Ed., Oxford University Press, Oxford, 2000. More specifically, phase-sensitive amplification in a χ(2) medium can occur via degenerate parametric amplification (DPA, also referred to as frequency down-conversion). As known in the art, a χ(2) medium is a nonlinear optical medium characterized in that its nonlinear optical response is governed primarily by the second-order nonlinear optical susceptibility (χ(2), a second-rank tensor). Typically, a χ(2) medium is a non-centrosymmetric crystal, two representative examples of which are (i) a lithium iodate crystal (LiIO3) and (ii) a KTP crystal (potassium titanium oxide phosphate, KTiOPO4).
In a DPA process, a pump wave interacts with signal and idler waves of a lower frequency, where the signal and idler waves have the same (i.e., common) frequency. This interaction is governed by the frequency-matching condition
χ2=2ω1 (5)
where ω2 is the pump frequency and ω1 is the common signal and idler frequency. Amplitude changes of waves <1> and <2> are governed by Eqs. (6) and (7):
dzA1=i2{overscore (γ)}A2Ai*exp(−iβz) (6)
dzA2=i{overscore (γ)}A12 exp(iβz) (7)
where Aj is the amplitude of the corresponding wave, z is the coordinate along the wave propagation direction, {overscore (γ)} is the nonlinear coupling coefficient, and β=2β1−β2 is the linear wavenumber mismatch. One can choose the amplitude units in such a way that |Aj|2 is proportional to the photon flux Pj.
Suppose that wave <2> is a relatively strong pump wave and wave <1> is a relatively weak signal wave. Then, in the small-signal (undepleted-pump) approximation, A2(z)=A2(0), where A2(0) is the wave amplitude at the beginning of the medium. Let
A1(z)=B1 exp(−iβz/2)
Then, B1 obeys the following (linearized) equation:
(dz−iδ)B1=iγB1* (8)
where δ=β/2 and γ=2{overscore (γ)}A2. Eq. (8) has a solution
B1(z)=μ(z)B1(0)+ν(z)B1(0)* (9a)
where the transfer functions μ(z) and ν(z) are:
μ(z)=cos h(κz)+i(δ/κ)sin h(κz) (9b)
ν(z)=i(γ/κ)sin h(κz) (9c)
and the growth rate κis:
κ=(|γ|2−δ2)1/2
The input-output relation described by Eqs. (9) is a property of phase-sensitive amplification. In quantum optics this relation is called a squeezing transformation.
To illustrate important characteristics of phase-sensitive amplification, consider a simple case, in which δ=0, and suppose that A2 is real (measure the signal phase relative to the pump phase). Then, according to Eqs. (9), the in-phase signal quadrature (B1+B1*)/2 is amplified by the factor exp(κz), whereas the out-of-phase quadrature (B1−B1*)/2 is attenuated by the same factor (i.e., multiplied by the factor exp(−κz)). This property of phase-sensitive amplification is largely responsible for the advantages of PSAs over PIAs specified in the background section.
Phase-Sensitive Amplification in a χ(3) Medium
As known in the art, a χ(3) medium is a nonlinear optical medium characterized in that its nonlinear optical response is governed primarily by the third-order nonlinear optical susceptibility (χ(3), a third-rank tensor). Typically, a χ(3) medium is a centrosymmetric crystal or nonlinear glass. One representative example of a χ(3) medium is a Kerr medium, in which the nonlinear response has a special form of χ(3)|E|2 E (where E is the electric field) often referred to as the Kerr nonlinearity. More details on various forms of nonlinear optical interactions in χ(2) and χ(3) media can be found, e.g., in a book by R. W. Boyd, entitled “Nonlinear Optics,” Academic Press, 1992, pp. 1-37 and 57-63, the teachings of which are incorporated herein by reference.
Representative examples of prior-art phase-sensitive amplification in a χ(3) medium are described in: (i) H. P. Yuen and J. H. Shapiro, “Generation and Detection of Two-Photon Coherent States in Degenerate Four-Wave Mixing,” Opt. Letters, vol. 4, pp. 334-336 (1979); (ii) B. Yurke, “Use of Cavities in Squeezed-State-Generation,” Phys. Rev. A, vol. 29, pp. 408-410 (1984); and (iii) P. Kumar and J. H. Shapiro, “Squeezed-State Generation via Forward Degenerate Four-Wave Mixing,” Phys. Rev. A, vol. 30, 1568-1571 (1984), the teachings of all of which are incorporated herein by reference.
The Yuen and Yurke references disclose schemes that utilize a degenerate backward FWM process. More specifically, a degenerate backward FWM process uses a counter-propagating pump configuration, in which two pump waves having the same frequency travel in opposite directions. Due to the phase-matching condition, the signal and idler waves travel in opposite directions as well. Phase sensitivity (and squeezing) is obtained by the use of a semi-transparent (50/50) beam splitter (Yuen) or a partly reflecting cavity mirror (Yurke) placed outside the χ(3) medium, which splitter/mirror combines the signal and idler waves to produce a desired squeezed state.
The Kumar reference discloses a scheme that is similar to the scheme disclosed in the Yuen reference, except that the counter-propagating pump configuration is replaced with a nominally co-propagating pump configuration. More specifically, each of the pump waves propagates at a small angle with respect to a selected axis. To satisfy the phase-matching condition, each of the signal and idler waves also propagates at that small angle with respect to that selected axis such that the plane passing through the wave vectors of the pump waves is orthogonal to the plane passing through the wave vectors of the signal and idler waves. In addition, all waves in the Kumar reference have the same frequency.
It should be noted that the χ(3) medium employed in either of the above schemes is a three-dimensional (3D) medium, i.e., a medium for which the transverse dimensions are comparable with the longitudinal dimension. In contrast, an optical fiber is a substantially one-dimensional medium because the fiber's transverse dimension (e.g., diameter) is typically much smaller than its longitudinal dimension (e.g., length).
First attempts to achieve phase-sensitive amplification in a fiber are disclosed in (1) M. Shirasaki and H. A. Haus, “Squeezing of Pulses in a Nonlinear Interferometer,” Journal of the Optical Society of America B, vol. 7, pp. 30-34 (1990) and (2) M. E. Marhic and C. H. Hsia, “Optical Amplification in a Nonlinear Interferometer,” Electronics Letters, vol. 27, pp. 210-211 (1991). More specifically, these references disclose schemes, each of which utilizes a nonlinear interferometer. The interferometer employs one or more 50/50 beam splitters to (1) direct the pump and signal waves along two interferometer arms, where the waves interact, and (2) combine the waves emerging from the arms to produce the corresponding squeezed states. Note that, in each of these interferometer schemes, (i) a single pump source is used to produce the pump waves that interact with the signal wave(s) in the interferometer arms and (ii) the signal gain increases as a quadratic function of the arm (fiber) length.
Four-Wave Mixing in a χ(3) Medium
The scalar FWM process illustrated in
ω3+ω1=2ω2 (10)
and the amplitude equations given by Eqs. (11):
dzA1=i{overscore (γ)}(|A1|2+2|A2|2+2|A3|2A1+i{overscore (γ)}A22A3*exp(iβz) (11a)
dzA2=i{overscore (γ)}(2|A1|2+|A2|2+2|A3|2)A2+i2{overscore (γ)}A3A1A2*exp(−iβz) (11b)
dzA3=i{overscore (γ)}(2|A1|2+2|A2|2+|A3|2)A3+i{overscore (γ)}A1*A22 exp(iβz) (11c)
where Aj is the amplitude of the corresponding wave, z is the wave propagation direction, {overscore (γ)} is the nonlinear coupling coefficient, and β=2β2−β3−β1 is the wavenumber mismatch. In this scalar FWM process, the signal and idler are substantially identical. The self-phase modulation (SPM) and cross-phase modulation (CPM) coefficients are not exactly equal to {overscore (γ)} and 2{overscore (γ)}, respectively. However, the deviations of these coefficients from the stated values are qualitatively unimportant and, for typical frequencies, are quantitatively insignificant. In one embodiment, ω2 is located within about 10 THz of ω0 (the zero-dispersion frequency).
Suppose that waves <1> and <3> are relatively strong pump waves and wave <2> is a relatively weak signal wave. Then, in the small-signal approximation, the pump photon-fluxes (P1,3) are constant and:
A3(z)=A3(0)exp [i{overscore (γ)}(2P1+P3)z] (12)
A1(z)=A1(0)exp [i{overscore (γ)}(P1+2P3)z] (13)
In a χ(3) medium, the pump waves are subject to SPM and CPM. Let
A1(0)=B1 (14a)
A2(z)=B2(z)exp [−iβz/2+i3{overscore (γ)}(P3+P1)z/2] (14b)
A3(0)=B3 (14c)
Then, the transformed signal amplitude obeys the (linearized) equation:
dz−iδ)B2=iγB2* (15)
where δ=β/2+{overscore (γ)}(P3+P1)/2 and γ=2{overscore (γ)}B3B1. Eq. (15) has the same form as Eq. (8). Hence, the input-output relation for the process of
The vector FWM process of
ω4+ω1=2ω2 (16)
and the following amplitude equations:
dzA1=i{overscore (γ)}(|A1|2+2|A2|2+ε|A3|2+ε|A4|2)A1+i{overscore (γ)}εA22A3A4*exp(iβz) (17a)
dzA2=i{overscore (γ)}(2|A1|2+|A2|2+ε|A3|2+ε|A4|2)A2+i{overscore (γ)}εA3*A4A1 exp(−iβz) (17b)
dzA3=i{overscore (γ)}(ε|A1|2+ε|A2|2+|A3|2+2|A4|2)A3+i{overscore (γ)}εA4A1A2*exp(−iβz) (17c)
dzA4=i{overscore (γ)}(ε|A1|2+ε|A2|2+2|A3|2+|A4|2)A4+i{overscore (γ)}εA1*A2A3 exp(iβz) (17d)
where Aj is the amplitude of the corresponding wave, z is the wave propagation direction, β=β2+β3−β4−β1 is the wavenumber mismatch, {overscore (γ)} is the nonlinear coupling coefficient for co-polarized waves, and ε is the ratio of the coupling coefficients for cross-polarized and co-polarized waves. For (polarization-maintaining) fibers with constant dispersion, ε=⅔; and for (non-polarization-maintaining) fibers with random dispersion, ε=1.
Suppose that waves <4> and <1> are relatively strong pump waves, and waves <2> and <3> are the polarization components of a relatively weak signal. Then, in the small-signal approximation, the pump photon-fluxes (P1,4) are constant and:
A4(z)=A4(0)exp [i{overscore (γ)}(εP1+P4)z] (18)
A1(z)=A1(0)exp [i{overscore (γ)}(P1+εP4)z] (19)
As in the scalar process described above, the pump waves are subject to SPM and CPM. Let
A1(0)=B1 (20a)
A2(z)=B2(z)exp [−iβz/2+i{overscore (γ)}3P1z/2+i{overscore (γ)}(ε−½)P4z] (20b)
A3(z)=B3(z)exp [−iβz/2+i{overscore (γ)}(ε−½)P1z+i{overscore (γ)}3P4z/2] (20c)
A4(0)=B4 (20d)
Then, the transformed signal components obey the following (linearized) equations:
(dz−iδ)B2=iγB3* (21)
(dz+iδ)B3*=−iγ*B2 (22)
where δ=β/2+{overscore (γ)}(P4+P1)/2 and γ={overscore (γ)}εB4B1. It follows from Eqs. (21)-(22) that:
B2(z)=μ(z)B2(0)+ν(z)B3*(0) (23)
B3*(z)=ν*(z)B2(0)+μ*(z)B3*(0) (24)
where μ and ν are defined in Eqs. (9b)-(9c). In quantum optics, the input-output relation defined by Eqs. (23) and (24) is called a two-mode squeezing transformation. If there is no input idler (i.e., B3(0)=0), then the output idler is proportional to the complex conjugate of the input signal, and the FWM process illustrated in
Cascaded Four-Wave Mixing Processes in a χ(3) Medium
The degenerate FWM processes described in the preceding subsection provide phase-sensitive amplification in a fiber because the frequency degeneracies allow the signal amplitudes to interact with their complex-conjugates (see Eqs. (9)). More specifically, in the degenerate PC process of
Consider the BS process. As already explained above, BS is governed by the frequency-matching condition:
ω2+ω3=ω4+ω1 (25)
and the amplitude equations analogous to Eqs. (17). Note that, unlike for the FWM process described in the preceding subsection, for the BS process, waves <1> and <3> are the pump waves (rather than waves <1> and <4> above), wave <2> is the signal wave, and wave <4> is the idler wave (rather than wave <3> above). Because power flows from the signal to the idler, the photon flux of the output idler wave cannot exceed the photon flux of the input signal wave. As a result, BS is intrinsically stable (i.e., does not provide gain).
Suppose that waves <1> and <3> are strong pumps, wave <2> is a weak signal, and wave <4> is a weak idler. Then, in the small-signal approximation, the pump photon-fluxes (P1,3) are constant and:
A1(z)=A1(0)exp [i{overscore (γ)}(P1+εP3)z] (26)
A3(z)=A3(0)exp [i{overscore (γ)}(εP1+P3)z] (27)
As in the PC process described above, the pump waves are subject to SPM and CPM. Let
A1(0)=B1 (28a)
A2(z)=B2(z)exp [−iβz/2+i{overscore (γ)}3P1z/2+i{overscore (γ)}(ε+½)P3z] (28b)
A3(0)=B3 (28c)
A4(z)=B4(z)exp [iβz/2+i{overscore (γ)}(ε+½)P1z+i{overscore (γ)}3P3z/2] (28d)
Then, the signal and idler amplitudes obey the following (linearized) equations:
(dz−iδ)B2=iγB4 (29)
(dz+iδ)B4=iγ*B2 (30)
where δ=β/2+{overscore (γ)}(P1−P3)/2 and γ={overscore (γ)}εB1B3*. It follows from Eqs. (29) and (30) that
B2(z)=μ(z)B2(0)+ν(z)B4(0) (31)
B4(z)=−ν(z)*B2(0)+μ*(z)B4(0) (32)
where the transfer functions are
μ(z)=cos(kz)+i(δ/k)sin(kz) (33)
ν(z)=i(γ/k)sin(kz) (34)
and the wavenumber k=(|γ|2'δ2)1/2. The transfer functions satisfy the auxiliary equation |μ|2+|ν|2=1. In quantum optics, the input-output relation defined by Eqs. (31)-(34) is called a beam-splitter transformation. For a representative case, in which δ=0, k=π/2 and γ is real, B2(z)=B2(0)/21/2 and B4(z)=iB2(0)/21/2. Thus, the output idler wave is a phase-shifted, but non-conjugated image of the signal wave. (Note that the phase shift results from photon-flux conservation.) Because the signal and idler frequencies are distinct, their relative phases can be different in the BS and PC processes (note that, because k4≠k2, propagation effects can change the relative phase naturally).
Now consider the PC process in which waves <5> and <1> are pump waves, wave <2> is a signal wave, and wave <4> is an idler wave. This process is governed by the frequency-matching condition
ω5+ω1=ω2+ω4 (35)
and the amplitude equations given by Eqs. (17), in which the subscripts 3 and 4 are replaced by the subscripts 4 and 5, respectively. The effects of SPM and CPM on the pump waves are described by Eqs. (18) and (19). The signal and idler amplitudes obey the following (linearized) equations:
(dz−iδ)B2=iγB4* (36)
(dz+iδ)B4*=−iγ*B2 (37)
where δ=β/2+{overscore (γ)}(P5+P1)/2, γ={overscore (γ)}εB5B1, and β=β2+β4−β5−β1. It follows from Eqs. (36) and (37) that
B2(z″)=μ(z″−z′)B2(z′)+ν(z″−z′)B4*(z′) (38)
B4*(z″)=ν*(z″−z′)B2(z′)+μ*(z″−z′)B4*(z′) (39)
where μ and ν are defined by Eqs. (9b) and (9c), respectively. Because B2(z′) is proportional to B2(0) and B4*(z′) is proportional to B2*(0), cascaded BS and PC provide PSA in a fiber. By controlling the phase of pump wave <5>, one controls the relative phase of the pump, signal, and idler waves and, hence, the orientation of the squeezing axis. For the representative case described above after Eq. (34), in which |B4(z′)|=|B2(z′)|, Eq. (38) reduces to Eq. (9a). It should be noted that the choice of pump frequencies is not unique. That is, different combinations of pump frequencies can provide phase-sensitive amplification by cascaded BS and PC. Furthermore, BS can also be used to combine a signal wave with the frequency-shifted, and conjugated, idler wave produced by prior PC.
Coupler 804a combines the pump waves generated by pump-wave sources 806a and 806b, and coupler 804b combines the resulting pump wave the optical communication signal received from section 802. The combined pump and signal waves interact in HNLF 808a via a BS process to generate an idler wave, e.g., as described above. Filter 810a placed at the end of HNLF 808a then filters out one of the pump waves (e.g., the pump wave generated by pump-wave source 806b) while passing the other pump wave (e.g., the pump wave generated by pump-wave source 806a) and the signal and idler waves to the second stage.
Coupler 804c adds to the waves received from the first stage a pump wave generated by pump-wave source 806c. These pump, signal, and idler waves interact in HNLF 808b via a PC process to generate an amplified wave, e.g., as described above. Filter 810b placed at the end of HNLF 808b then separates the amplified wave from the other waves present in HNLF 808b for further transmission in the communication system via section 802′. In general, since both the signal wave and the idler wave are amplified in HNLF 808b, either one of them, or both, can be used as an amplified output of OFA 800. In reference to
While this invention has been described with reference to illustrative embodiments, this description is not intended to be construed in a limiting sense. For example, in different OFA implementations, optical fibers having different properties (e.g., polarization maintaining or regular fibers) may be used. Polarization sensitive filters or polarizers can be appropriately inserted at different locations of the OFA to provide a desired polarization configuration. The pump waves may be CW or pulsed. Two, three, or more pump waves may be used. Idler waves may be filtered out or used in the communication system, e.g., as protection signals in a 1+1 protection scheme. Various modifications of the described embodiments, as well as other embodiments of the invention, which are apparent to persons skilled in the art to which the invention pertains are deemed to lie within the principle and scope of the invention as expressed in the following claims.
Although the steps in the following method claims, if any, are recited in a particular sequence with corresponding labeling, unless the claim recitations otherwise imply a particular sequence for implementing some or all of those steps, those steps are not necessarily intended to be limited to being implemented in that particular sequence.