Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
8. Propagation in Nonlinear Media 8.1. Microscopic Description of Nonlinearity. 8.1.1. Anharmonic Oscillator. Use Lorentz model (electrons on a spring) but with nonlinear response, or anharmonic spring f NL x t d 2 x t dx t eE t 2 x t 0 dt 2 dt m m 2 3 f NL a2 x t a3 x t m b) Harmonic spring; Low intensity c) Anharmonic spring (overstretched); High intensity At strong incident linearly polarized fields electron trajectories no longer straight lines 8.1.2. Lorentz Force as Source of Nonlinearity At strong incident linearly polarized fields electron trajectories no longer straight lines Molecule oscillates over other degrees of freedom Lorentz force ev B becomes significant Simple case where electric field polarized along x, magnetic field lies on y E Ex , 0, 0 H 0, H y , 0 f em e E v B y zˆ y xˆ 0 yˆ exB e Ex zB d 2 x t dx t e t 2 x t Ex z (t ) By (t ) 0 2 dt dt m d 2 y t dy t 0 2 y t 0 2 dt dt d 2 z t dz t e 2 z t x (t ) By (t ) 0 2 dt dt m Equations of motion from E and H fields, no force acting in y direction. Solve for x and z displacements. Take Fourier transform, obtain following equations: eEr e i z By m m e z 0 2 2 i i x By . m x 0 2 2 i x D eEx D 2 b 2 m b eEx z 2 . D b 2 m D 0 2 2 i b i Express magnetic field in terms of electric field Equations obtained from method in 5.1 Ignore magnetic field, we get linear response e By . m Order of magnitude relation between x and z displacements z eEx eEx x me 2 mc 2 k E i B By Ex ic Solve for when electric field and associated irradiance where x and z are comparable 1 2 2 mc 2 I Ex E x x z 2 e 1012 V m , 1021 W m2 Using b e Ex mc D we have: e2 z 2 Ex 2 2 mc D b 2 e2 2 Ex 2 2 mc D z Ex eit Ex e it 2 2 Ex 2 1 cos 2t . Electron oscillates at 2 on z axis, DC term is called optical rectification x D eEx 2 m D b 2 eEx 1 . m D 8.1.3. Dropping the Complex Analytic Signal Representation of Real Fields. Real field Complex analytic signal i t U r A cos t U Ae 1 2 it it 2 Ur A e e 2 A2 1 cos 2t U A2 e i 2 t . Re[U ] A 2 cos(2t ) U r Complex analytic signal does not capture DC term. Whenever we deal with fields raised to powers higher than one, we use Ur 1 U U * 2 8.2. Second‐Order Susceptibility d 2 x t dx t 2E t 2 2 x t a x t 0 2 dt 2 dt m Want solution of form where x x x 1 2 is the nonlinear perturbation to linear solution. Simplify by neglecting 2 d 2 x 1 t dx 1 t eE t 2 1 x t 0 2 dt m dt x 2 t x 2 t 0 2 x 2 t a2 x 1 t 2 D x 2 a2 x 1 ⓥx 1 d 2 x 2 t dx 2 t 0 2 x 2 t 2 dt dt 2 and 2 a2 x 1 t 2a2 x 1 t x 2 t a2 x 2 t 0. 2 e E E a2 ⓥ , m D D Illuminating the nonlinear crystal with two monochromatic fields of different frequencies E E1 1 E1* 1 E2 2 E2* 2 2 1 E E e x a2 ⓥ m D D D 2 E ' E ' 1 e d '. a2 m D D ' D ' a ⓥ b a b 2 E E 2 All possible frequencies: 1 1 2 E ⓥE 22 2 1 21 2 1 0 2 1 21 2 1 22 2 1 e x 2 a2 f 1 f 1 f 2 f 2 m D E12 21 E1 1 E1 E2 1 2 f 1 D 1 D 1 * E1 E2 1 2 2 E2 2 22 E2 2 1 E1 E2 1 2 f 2 D 2 D 2 E1 E2* 1 2 2 e x t a2 g1 t c.c. g 2 t c.c. m 2 E1 E12 ei 21t E1 E2 e 1 2 g1 t D 21 D 2 1 D 0 D 1 2 D 1 2 D 1 D 2 2 Fourier Transform → i t E1 E2*e 1 2 c.c. D 1 2 D 1 D* 2 i t E2 E2 2 e i 2 2 t E1 E2 ei1 2 t g2 t D 22 D 2 2 D 0 D 2 2 D 1 2 D 1 D 2 2 E1 E2*e 1 2 c.c. D 1 2 D 1 D* 2 i t Equation 8.26 indicates that, as the result of the second-order nonlinear interaction, the resulting field has components that oscillate at frequencies 22 (second harmonic of 2), 21 (second harmonic of 1), 1 +2 (sum frequency), 1 -2 (difference frequency) and 0 (optical rectification terms). Nonlinear susceptibility from induced polarization P 2 i j 0 2 i j ; i , j E i E j → * * 2 i j ; i , j P 2 i j Nex 2 i j , i, j 1, 2 P 2 r 0 r E r E r 2 r E r E r P r . 0 0 x 2 i i a2 Ne3 g t g 2 t c.c. 0m2 1 Importantly, 2 vanishes in centrosymmetric media 2 Ne P 2 2 r 2 r r Ner Ne r 2 P r . Fulfilled simultaneously only if P 2 r 0 So second‐order nonlinear processes require noncentrosymmetric media 8.2.1. Second Harmonic Generation (SHG) a2 Ne3 1 21 ; 1 , 1 0 m 2 D 21 D 2 1 a2 Ne3 1 . 22 ; 2 , 2 0 m 2 D 22 D 2 2 Nonlinear susceptibility a s function of linear chi: Linear: a) SHG: pumping the chi(2) material at omega yields both the fundamental frequency (omega) and its second harmonic (2omega). b) Description in terms of virtual energy levels. Ne 2 1 . 0 m D 1 Nonlinear: 2 2 a2 0 2 m 1 1 21 ; 1 , 1 2 3 21 1 . N e 8.2.2. Optical Rectification (OR) 2 a2 Ne3 1 0; 1 , 1 2 0 m D 0 D 1 D 1 a2 0 2 m 1 0 1 1 1 1 . 2 3 N e OR: a DC polarization is created in a chi(2) material. 8.2.3. Sum Frequency Generation (SFG) 2 a2 Ne3 1 1 2 ; 1 , 2 2 0 m D 1 2 D 1 D 2 a2 0 2 m 1 1 2 1 1 1 2 . 2 3 N e 8.2.4. Difference Frequency Generation (DFG) 2 a2 Ne3 1 1 2 ; 1 , 2 0 m 2 D 1 2 D 1 D 2 a2 0 2 m 1 1 2 1 1 1 2 . 2 3 N e 8.2.5. Optical Parametric Generation (OPG) The time reverse process of SFG 1 2 3 2 2 3 8.3. Third‐Order Susceptibility Anharmonic oscillator x t x t 0 2 x t a3 x3 t eE t . m Solve using perturbation theory, solution of form eE t 1 1 2 1 x t x t x t 0 m 3 x t x t 0 x t a3 x1 t x 3 t 0. 3 3 2 3 1 3 First term vanishes, approximate a3 x x 3 a3 x 1 3 3 → 3 3 3 1 x t x t 0 2 x t a3 x t , e E E E 1 x t x , m D 1 1 E1* 1 E2 2 E2* 2 E3 3 E3* 3 . 1 For scalar fields perturbation displacement is: e x t x t 0 2 x t a3 m 3 3 3 3 m , n , p 3 Em En E p e i m n p D m D n D p So, e 3 En e int c.c. x t m n 1 D n 1 Induced polarization both for electromagnetic fields in terms of displacement and susceptibility Pi Pi 3 3 Nex 3 q q 3 3 q 0 j , k ,l 1 m , n , p 3 ijkl 3 q ; m , n , p E j m Ek n El p 0 d ijkl 3 q ; m , n , p E j m Ek n El p , jkl Where d is the degeneracy factor 3 General expression for a3 Ne* 1 , ijkl q ; m , n , p d 0 m3 Di q D j m Dk n Dl p 3 Da 0 a 2 2 i Express 3 in terms of 1 a3 m 03 1 1 1 1 ijkl q ; m , n , p p , i q j m k n l 3 4 dN e 3 a 1 Ne 2 1 0 m Da 8.3.1. Third Harmonic Generation (THG) signal contained by terms that oscillate at 3ω The THG susceptibility for one component of the electric field (scalar case) is 3 a3 Ne 4 1 3; , , d 0 m3 D 3 D 3 3 a3 m 03 1 1 3 4 3 , N e D 0 2 2 i 8.3.2. Two‐Photon Absorption (TPA) and Intensity‐Dependent Refractive Index If 1 , 2 , 3 , 3 a3 Ne4 1 ; , , d 0 m3 D 2 D 2 2 2 a3m 03 1 1 3 4 N e 3 2 2 2 a3m 03 1 3 4 R1 I1 i 2 R1 I1 N e R i I 3 3 , Define effective refractive index 1 3 3 E n0 2 1 3 3 E 2 2 n 2 1, n n0 n2 I Find expression for n0 2n2 I 1 ( ) 2 3 3 n2 4n0 2 0 c 3 R3 i I3 4n0 0 c 2 n2 ' in2 '' . • A plane wave undergoes an intensity‐dependent loss of factor e^‐alpha. • b) Energy level diagram for single photon absorption (left) and two‐photon absorption (right). 8.3.3. Four Wave Mixing 3 ; 1 , 2 , 3 1 2 3 a3 Ne 1 d 0 m3 D D 1 D 2 D 3 k k1 k 2 k 3 k‐vector conservation, phase matching condition a) Generic four‐wave mixing process. b) Momentum conservation. PNL 6 0 A1 A2 A3*e 3 i k1 k 2 k 3 r . 8.3.4. Phase Conjugation via Degenerate (all are the same) Four‐Wave Mixing Phase conjugation via degenerated four wave mixing: field E4 emerges as the phase conjugate of E3, i.e. E4=E3*. 8.3.5. Stimulated Raman Scattering (SRS). 8.3.5.1. Spontaneous Raman Scattering Population of the excited vibrational levels obeys the Maxwell‐Boltzmann distribution E 1 k BT PE e , z Spontaneous Raman scattering: ‐a) Stokes shift ‐b) anti‐Stokes shift P E1 P E0 Thus, the Stokes component is typically orders of magnitude stronger ! e h10 k BT 1 p E. the molecular optical polarizability, t 0 xv xv t Av eivt Av*eivt , P t N P xv t . xv 0 N 0 xv xv t E t . xv 0 P t PL t PNL t PL t N 0 AL e iLt c.c. PNL t N * iL v t i L v t c.c. . . . A A e c c A A e v L v L xv 0 Harmonic vibration occurs spontaneously due to Brownian motion 8.3.5.2. Stimulated Raman Scattering To derive expression for SRS susceptibility, solve equation of motion for vibrational mode of resonant frequency and damping , d 2 xv t dxv t F t 2 v xv t , dt 2 dt m 1 p t E t t 2 1 E 2 t t 2 2 1 d 0 xv E t 2 dxv 0 F t W dW Fdxv . t dW dxv 1 d 2 dxv E 2 t . t 0 1 d F E ⓥE 2 dxv 0 E t AL e iLt AS e iS t c.c. E AL L AS S AL* L AS * S P N 0 xv E xv 0 F 1 xv m v 2 2 i 1 d 2m dxv 0 N d PNL 2m dxv AL* AS L ⓥ S v 2 2 i N 0 E N xv E xv 0 PL PNL . 2 AL* AS L ⓥ S AL L AL* L 2 2 v i 0 0 d R AL* AS L S AL L AL* L N d PNL t 2m dxv Induced polarization for SRS i t * * AL AS AL e S AL e L S 2 2 i L S 0 v L S 2 i 2 t nonlinear contribution N dx S t 12m 0 dxv L v S 2 1 2 2 v L S i L S 0 → N dx S t i 12m 0 dxv 2 1 L S 0 By resonantly enhancing the vibration mode the Stokes component can be amplified significantly; in practice, this amplification can be many orders of magnitude higher than for the spontaneous Raman. susceptibility associated with the anti-Stokes component L AS v L S N d AS t i 12m 0 dxv S 2 1 L S 0 t * Strong attenuation a) Raman susceptibility at Stokes frequency; indicates amplification. b) Raman susceptibility at anti‐Stokes frequency; indicates absorption. 8.3.6. Coherent Anti-Stokes Raman Scattering (CARS) and Coherent Stokes Raman Scattering (CSRS) • Coherent Anti-Stokes Raman Scattering (CARS) and Coherent Stokes Raman Scattering (CSRS) are also established methods for amplifying Raman scattering. • These techniques involve two laser frequencies for excitation 3 CARS A 21 2 ; 1 , 1 , 2 3 CSRS S 22 1; 2 , 2 , 1 CARS is a powerful method currently used in microscopy we will hear about it during student presentations. 8.4. Solving the Nonlinear Wave Equation. 8.4.1. Nonlinear Helmholtz Equation B r, t E r, t t D r, t H r, t j r, t t D r, t B r, t 0. 2 D r, t E r, t 0 0, t 2 follow the standard procedure of eliminating B and H from the equations 0 j 0. D r, t 0 E r, t P r, t 0 E r, t PL r, t PNL r, t 0 r E r, t PNL r, t E r, t E 2 E 2 E. 2 2 PNL r, t n 2 E r, t E r, t 2 0 , c dt 2 t 2 2 2E r, 2 E r, 1 0 0 2 PNL r, , Nonlinear wave equation after approximation, D E term negligible Propagation of the Sum Frequency Field SFG nonlinear polarization has the form 2 PNL r; 3 1 2 ; 1 , 2 0 2 3 ; 1 , 2 E1 r, t E2 r, t E1 r, t A1 r e i1t 1z c.c. E2 r, t A2 r e E3 r, t A3 r e 1 n 1 1 c i 2 t 2 z i 3t 3 z c.c. c.c. E3 r, t n 3 2 E3 2 x E3 y 32 c2 E3 r, t 032 0 E1 r, t E2 r, t 2 0 d 2 z i3t 3 z E3 r, t 2 A3 e dt d 2 z i3t 3 z dA3 2 2 A3 2i 3 3 A3 e . dz dz 2 d 2 A3 t dA3 z i 1 2 3 z 2 2 i A A e 2 . 3 0 3 0 1 2 2 dt dz Simplifying approximation, A1 and A2 , do not change with z (do not deplete). d 2 A3 t dA3 t ikz i Be 2 3 dt 2 dt B 32 A1 A2 2 k 1 2 3 , amplitude is slowly varying d 2 A3 z dA3 t 3 . dz 2 dz 2 i ikz dA3 t e dz 23 L iB A3 L eikz dz 23 0 The intensity of SFG field is I3 z iB eikL 1 23 ik kL iB i2kL 2sin 2 e k 23 kL iBL i2kL sin 2 e 23 kL 2 2 1 A3 z 2 0 n B 2 L2 sinc 2 kL , 2 2 83 Net output power can vanish at kL 2 , 2 ,... iB i2kL e sinc kL . 2 23 a) The SFG output field has a phase that depends on the position where the conversion took place. b) The overall SFG intensity oscillates with respect to <eq> Parametric Processes: Phase Matching k 0 3 1 2 n 3 n 1 1 n 2 2 , 3 1 2 c c c 3 n 3 n 1 1 n 2 2 . 1 2 Normal dispersion Abnormal dispersion. Normal dispersion curves for a positive uniaxial crystal. Negative ne no Positive ne no Type I ne 3 3 no 1 1 no 2 2 no 3 3 ne 1 1 no 2 2 Type II ne 3 3 ne 1 1 no 2 2 no 3 3 no 1 1 ne 2 2 Table 8-1. cos 2 sin 2 2 2 n n0 ne 2 1 phase matching can be achieved by angle tuning, that is, selecting the angle that ensures k Type I no 3 3 n 1 , 1 n 2 , 2 Type II no 3 3 no 1 1 n 2 , 2 Type II phase matching by angle tuning in a positive uniaxial crystal: o‐ordinary wave, e‐extraordinary wave, c‐optical axis. 0 Electro‐Optic Effect The electro-optic effect is the charge in optical properties of a material due to an applied electric field that oscillates at much lower frequencies than the optical frequency. Electro‐Optic Tensor a) Linear interaction with a birefringent crystal. b) Electro‐optic (nonlinear) interaction. p is the induced dipole, E(omega) is the optical field, and E(0) is the static field. These sketches should be interpreted as 3D representations. The induced polarization for the Pockels effect Pi ; , 0 0 ijk ; , 0 E j Ek 0 . 2 Kerr effect 3 Pi ; , 0, 0 0 ijkl ; , 0, 0 E j Ek El , Due to the electro‐optic effect, an optical field can suffer voltage‐dependent polarization and phase changes. Pockles effect Pi 2 ; , 0 rijk 0 2 ijk 1 0 ii jj rijk E j Ek 0 . ijk ii jj ni 2 n j 2 rijk rjik Change in the rank of the tensor, from 3 to 2 r11k r1k r22 k r2 k r33k r3k r12 k r21k r6 k r13k r31k r5 k r23k r32 k r4 k . Electro-optic tensor can be represented by a 3x6 matrix. This tensor contraction, allowed by the permutation symmetry, reduces the number of independent elements from 32 27 to 3 6 18 . Electro‐Optic Effect in Uniaxial Crystals Use KDP (KH2PO4, or potassium dihydrogen phosphate) as a specific example of uniaxial crystal 0 0 0 r r41 0 0 0 0 0 0 r41 0 0 0 0 0 0 r63 The KDP refractive index tensor is (in the normal coordinate system of interest) no n 0 0 assume that the voltage is applied only along z, such that only r63 is relevant. 0 no 0 0 0 ne D 0 n2E P Di 0 ni 2 Ei 0 ni 2 n j 2 rijk E j Ek 0 ij E j , electric displacement can be expressed for each component as (i=x, j=y, k=z) Dx 0 no 2 Ex 0 no 4 r63 E y Ez 0 Dy 0 no 2 E y 0 no 4 r63 Ex Ez 0 Dz 0 ne 2 Ez . no 4 r63 Ez 0 0 no 2 no 2 0 . 0 no 4 r63 Ez 0 2 0 0 n e Electro‐optic effect in a KDP crystal. For E(0) parallel to z, the normal axes rotate by 45 degrees around the z‐axis. express ij in a reference system rotated about the z-axis by an arbitrary angle, say ' R R cos 0 sin sin no 2 D cos cos D no 2 sin no 2 D sin 2 0 D cos 2 D cos 2 , 2 no D sin 2 no 2 D 0 0 ' 0 0 no 2 D 0 . 2 0 0 n e sin cos D no 4 r63 Ez 0 Electro‐Optic Modulators (EOMs) defining the normal refractive indices nij ' ij ' n 'x n' 0 0 0 n 'y 0 0 0 n 'z phase retardation n 'x n ' y 2 0 c L no 3r63 Ez 0 L, n 'x no 2 D 1D no 2 no 1 no no 3 r63 Ez 0 2 1 n ' y no no 3 r63 Ez 0 2 n 'z ne . Ez 0 V . d half-wave voltage V 0 d . 3 2no r63 L V V . V a) Longitudinal modulators (electrodes are made of transparent material), d=L. b) Transverse modulator, d<L. a) Electro‐optic (EO) crystal operating as phase modulator: incident polarization parallel to the new normal axis. b) Amplitude modulation: input polarization parallel to original (i.e. when V=0) normal axis; P is a polarizer with its axis parallel to x. KDP longitudinal modulator reads E ' V Ee ino k0 L Ee ino k0 L e e 1 k0 no3 r63V 2 i V 2 . (x’,y’) is rotated by 45° with respect to (x,y) Ex ' V 0 E i ' sin R W V R 45 45 0 x 2 1 Ey ' 2 i2 I E ' ' x x 0 e W0 V i 2 V 0 e2 sin , 2 2 2 2 2 . R 45 2 2 2 2 voltage is modulated sinusoidally, at frequencies, , much lower than the optical frequency, V t V0 sin t t V0 sin t. V phase and intensity modulations 1 V x ' t 0 sin t 2 V 1 V I x ' t sin 2 0 sin t . 2 V phase modulators also modulate the frequency of the light dx ' t dt 1 V t . 2 V 't V 1 I x ' t 1 cos 0 sin t 2 V 2 V0 1 1 sin sin t 2 V 1 V0 sin t. 2 2 V Liquid crystal spatial light modulator: a) phase mode; b) amplitude mode. Acousto‐Optic Effect The acousto-optic effect is the change in optical properties of a material due to the presence of an acoustic wave Elasto‐Optic Tensor the medium acts as a (phase) grating, which is capable of diffracting the light 2 k , v Light wave (wavevector k, frequency omega, speed c) interacts with a travelling grating induced by a sound wave (wavevector script‐k, frequency Omega, speed v). induced polarization in the normal coordinate system Pi 3 3 0 j , k ,l 1 ni 2 n j 2 Pijkl Skl E j . electro-optic tensor, due to symmetry, the elasto-optic tensor can also be used using the following (Voigt) contraction strain tensor 1 u u Skl k l 2 xl xk . Kerr electro-optic effect, induced polarization Pi 3 3 0 ni 2 n j 2 Sijkl Ek 0 El 0 E j , ijkl 1 Pijkl Skl Sijkl Ek 0 El 0 . Pijkl Pmn , i, j , k , l 1, 2,3; m, n 1, 2,..., 6 ij 11, kl 11; m 1, n 1 ij 22, kl 22; m 2, n 2 ij 33, kl 33; m 3, n 3 ij 12, kl 12; m 6, n 6 ij 13, kl 13; m 5, n 5 ij 23, kl 23; m 4, n 4. Photoelastic effect: the sound wave induces strain, which in turn modulates the refractive index. a) Longitudinal sound wave; b) Transverse (shear) sound wave. Lambda is the sound wavelength, with Omega the frequency and v the propagation speed. Acousto‐Optic Effect in Isotropic Media Let us investigate in more detail the acousto-optic effect in isotropic media elasto-optic tensor for isotropic media pmn p11 ' p12 ' p12 ' 0 0 0 p12 ' p11 ' p12 ' p12 ' 0 0 0 0 p12 ' p11 ' 0 0 0 0 1 p11 ' p12 ' 2 0 0 0 0 1 p11 ' p12 ' 2 0 0 0 0 0 0 0 1 p11 ' p12 ' 2 0 0 S kl S33 S3 expressions for the polarization Px 3 0 n 4 S3 p13 Ex p63 E y p53 Ez Dielectric displacement D 0 n 2 E P 3 0 n 4 S3 p63 Ex p23 E y p43 Ez 3 Pz 0 n 4 S3 p53 Ex p43 E y p33 Ez . E , p43 p53 p63 0 dielectric tensor Py 3 Px 3 0 n 4 S3 p13 Ex Py 3 0 n 4 S3 p23 E y Pz 3 0 n 4 S3 p33 Ez . 1 n 2 p12 ' 0 0 0 n 2 S3 0 1 n 2 p12 ' 0 . 2 0 0 1 n p11 ' medium becomes uniaxial, i.e. xx yy zz 1 u z S3 2 z 1 ikAe i t kz 2 plane acoustic wave of amplitude A uz Aei t kz . a) Brogg regime; b) Raman‐Nath regime. Bragg Diffraction Regime wave equation for the total field (incident plus diffracted) reads 2 P 3 r , t n 2 E r, t E r, t 2 0 c t 2 t 2 3 i t k r P r, t BAe E1 r, t 2 E r, t E1 r, t E2 r, t , E2 q, F r, eiqr d 3r V A1 q 1 k k1 , Fourier transform 2 E2 r, n 2 0 2 E2 r, F r, E1 r, F r, 2 BAeik r . A1 1 k 2 k1 2 1 . 2 E1 r, n 2 0 2 E1 r, 0 eit E1 r, t E1 r, . Bragg condition k2 2 1 2 1 k1. 2k1 sin sin 1 2n The Brogg condition: a) k1, k2 and script‐k are, respectively, the incident, diffracted and acoustic wavevectors. b) Triangle that illustrates geometrically k2=k1+script‐k. Raman‐Nath Diffraction Regime a) Bragg diffraction (see solution for k2). b) Raman‐Nath diffraction when sound wave is curved (multiple solutions for k2). scattering potential in the (x,z) coordinates F x, z , 2 BAe i 0 x 2 i 0 z 2z e diffracted field is the Fourier transform of F qx 2 E2 q x , q z , q z 0 . 2 0 q is the scattering wavevector, q k 2 k 1 . k k k2 z k1z 0 2 x 1x 2 0 2 . if the sound beam curvature satisfies k2 x k1x 2 0 2 m 0 , Where m=0,1,2,…, the diffracted beam has multiple solutions k2 z k1z 0 m 0 .