J. High Energy Phys. 12 (1999) 010 PreprinttypesetinJHEPstyle. -HYPERVERSION hep-th/9911093 Dyons and Interactions in Nonlinear (Born-Infeld) Electrodynamics 0 by Alexander A. Chernitskii 0 0 St.Petersburg Electrotechnical University, 2 Prof. Popov str. 5, St.Petersburg 197376, Russia n a E-mail: [email protected] J 0 2 Abstract: Born-Infeld nonlinear electrodynamics with point singularities having 3 v both electric and magnetic charges are considered. Problem of interaction between 3 the associated soliton dyon solutions is investigated. For the case of long-range inter- 9 0 action at first order by a small field of distant solitons we obtain that the generalized 1 Lorentzforceisactedonadyonunderconsideration. Short-rangeinteractionbetween 1 9 two dyons having identical electric and opposite magnetic charges is investigated for 9 / an initial approximation. We consider the case when the velocities of the dyons have h t equal modules and opposite directions on a common line. It is shown that the associ- - p ated field configuration has a constant full angular momentum which is independent e h of the interdyonic distance and their speed. This property permits a consideration : v of this bidyon configuration as an electromagnetic model of charged particle with i X spin. We numerically investigate movement of the dyons in this configuration for r the case when the full electric charge equals the electron charge and the full angu- a lar momentum equals the electron spin. It is shown that for this case the absolute value of relation between electric and magnetic charges of the dyons equals the fine structure constant. The calculation gives that the bidyon may behave as nonlinear oscillator. Associated dependence of frequency on the full energy is obtained for the initial approximation. In the limits of the electrodynamic model we obtain that the quick-oscillating wave packet may behave like massive gravitating particle when it move in high background field. We discuss the possible electrodynamic world with the oscillating bidyons as particles. Contents 1. Introduction 1 2. Basic relation for field 2 3. Boundary conditions at the singular points 5 4. Dyon solutions and singularities 6 5. Variational principle for two potentials 7 6. Conservation laws and comments about dimensions 10 7. Method for investigation of interaction between dyons 11 8. Interaction of dyon with a small given field 14 9. Bidyon or an electromagnetic model of particle with spin 16 10. Effect of gravitational interaction 28 11. Nonlinear electrodynamic world with oscillating bidyons 32 12. Conclusions 33 1. Introduction Solitons in nonlinear electrodynamics may behave like real particles. This analogy with classical particles is manifested when we consider a long-range interaction of solitons by a perturbation method [1, 2]. In this case a soliton is subjected to the Lorentz force in the first order by a field of distant solitons. The second order may give the soliton’s trajectory in the form of geodesic line for some effective Riemann space with metric depending on the field of distant solitons. Thus we have also an analogy with gravitating particles. Moreover, light beams distortion under the actionofsomegivenfieldmayappearasgravitationaldistortion[3]. These properties (in particular) provoke interest to nonlinear electrodynamic models in the context of possible unifying description for the real material world. In the framework of 1 this approach it may hope to solve the problem of unification electromagnetism and gravitation. Here we shall consider an interaction of singular solitons in Born-Infeld nonlin- ear electrodynamics1. The investigation of this nonlinear electrodynamic model with point singularities of field is presented in the article [6], where we have considered the singularities with electric charge. But the model’s system of equations for electro- magnetic field has an exact solution which has non-zero divergence for both fields of electric D and magnetic B inductions at the singular point. That is, this point for the solution has both electric and magnetic charges. The particle with both electric and magnetic charges had been named by J.Schwinger as dyon (see [7]). Because the singular solution lookslike point charged particle [6], we namethis field configuration as dyon. In the present article we consider an interaction between these dyons and we consider some quick-oscillating two-dyons field configuration. We also consider like-gravitational interaction of a quick-oscillating wave packet with given field. 2. Basic relation for field Let us state the basic relation for the Born-Infeld nonlinear electrodynamics with singularities [6] generalized to the case with both electric and magnetic charges. We shall use an inertial coordinate system in Minkowskian space: xµ (the Greek { } indexes take values 0,1,2,3). That is, components of metric is independent of time x0 and g = 1, g = 0 (the Latin indexes take values 1,2,3). In this case (see 00 0i | | also [6]) we can write the following nonlinear Maxwell system of equations: divB = 4π¯0 (′) , (a)  divD = 4π0  (′′)  ∂0B+rotE = 4π¯  (′) (2.1) − , (b)  ∂0D−rotH = −4π  (′′) where 1 1 D = (E+α2 B) E = (D α2 H) J − J  L ,  L , 1 1 H = (B α2 E) B = (H+α2 D) − J J (2.2)  L  L = 1 α2 α4 2 , = E2 B2 , = E B , L | − I − J | I − J · q 1 α2 α4 2 , = H2 D2 , = H D , L ≡ | − I − J | I − J · q 1It will be recalled that M. Born and L. Infeld had been considered in their article [4] the electrodynamic model which follows from an action that had been proposed by A.S. Eddington in his book [5]. 2 µ and ¯µ are components of electric and magnetic singular currents such that N N 0 1 dnδ(x an) ,  1 dnVn δ(x an) , ≡ √|g| − ≡ √|g| − nX=1 nX=1 (2.3) N N ¯0 1 nbδ(x an) , ¯ 1 nbVn δ(x an) , ≡ √|g| − ≡ √|g| − n=1 n=1 X X n n d and b are electric and magnetic charges of n-th singularity, n n n n da a = a(x0) is a trajectory of it and V . ≡ dx0 Here we use the definition for the three-dimensional δ-function which is suitable for discontinuous functions f(x): 1 f(x)δ(x an)(dx)3 lim f(x) dσn f(x) , σn dσn . (2.4) ΩZn − ≡ σn→0 |σn| Zσn  ≡ (cid:28) (cid:29)n | | ≡ Zσn   n n where Ω is a region of three-dimensional space including the point x = a, n n n σ is a closed surface enclosing this point, dσ is an area element of the surface σ, n n σ is an area of the whole surface σ. | | Let us define the following functions: +α2D E = +α2B H = α2T00 +1 , H ≡ L · L · (2.5) P (D B) , i = T0i , ≡ × P where Tµν are components of symmetrical energy-momentum tensor µν (see [6] where we designated it as T ). Then, from relations (2.2) we can obtain the following relation (see also [8, 9]): 1 ∂ 1 1 E = H = (D α2P B) = [(1+α2B2)D α2 (D B)B] α2 ∂D − × − ·  H H , (2.6) H = 1 ∂H = 1 (B+α2P D) = 1 [(1+α2D2)B α2 (D B)D] α2 ∂B × − ·  H H where = 1+α2(D2 +B2) +α4 P2 = H q = (1+α2D2)(1+α2B2) α4 (D B)2 . (2.7) − · q Using relations (2.6) we can consider system (2.1) as the system of equations for fields D,B. This representation is best suitable for investigation of the singular dyon solutions. Let us introduce the two electromagnetic potentials A , A¯ . In our case the µ µ potentials have singular line for each singular charged point (in three-dimensional 3 space). We must exclude such line fromspace when consider any ordinarydifferential field model. The alternative way is connected with using distributions or generalized functions. As described by P.A.M. Dirac for monopoles [10], in this case we must include some distributions into definitions of the potentials through derivatives. But here we shall adhere to the first way associated with exclusion of the singular lines. Note that the singular currents into equations (2.1) set boundary conditions at the singular points (see [6]). In contrast, we shall take natural boundary conditions at the singular lines outside of the singular points. Thus we define the potentials outside of any singular set with help the following formulas: F = ∂ A ∂ A , fµν = εµνσρ∂ A¯ , (2.8) µν µ ν ν µ σ ρ − − where ε = g 1/2, ε0123 = g −1/2, and 0123 | | −| | E F Di f0i i i0 ≡ ≡  1  1 Bi εijkF , H ε fjk , (2.9)  ≡ 2 jk  i ≡ 2 ijk F = ε Bk fij = εijkH ij ijk k   where ε = g 1/2, ε123 = g −1/2, ( g = 1, g = 0). 123 00 0i | | | | | | We have the following definition in three-dimensional designations: B = ∇ A × , (a) D = ∇ A¯  × (2.10) H = ∇A¯ +∂ A¯  0 0 − . (b)  E = ∇A ∂ A  0 − 0  From the basic equation of the model outside of the singularities [6] 1 ∂ 1 α2 g fµν = 0 , fµν = Fµν εµνσρF (2.11) g ∂xµ | | − 2 J σρ! | | q L q we can easy obtain also the following equation for the potential A outside of the µ singular set: ∂2A ∂ln g Cµνσρ ρ + L fµν | | = 0 , (2.12) ∂xµ∂xσ 2 ∂xµ where Cµνσρ = gµσgνρ gµρgνσ α2( µν σρ +fµν fσρ) , (2.13) − − F F 1 µν εµναρF . (2.14) αρ F ≡ −2 4 3. Boundary conditions at the singular points Let us integrate system of equations (2.1) over a small four-dimensional space region including only n-th singular point. Using the partial integration in four-dimensional space we obtain the following conditions: ∞ ∞ n n fµν dσ 4π µdΩ dx0 = 0 , µν dσ 4π ¯µdΩ dx0 = 0 , ν ν " − # " F − # −Z∞ Zn Zn −Z∞ Zn Zn σ Ω σ Ω (3.1) n where Ω is a three-dimensional region including only n-th singular point, n dσ is external directed element of the closed (two-dimensional) surface n σ enclosing the singular point (see also [6]), dΩ g (dx)3 , dσn Vn dσn , 0 ≡ | | ≡ − · 1 n q dσ dx0 = ε dxν dxσ dxρ are components of four-vector. µ µνσρ − 6 ∧ ∧ n The surface σ is rigidly coupled with n-th singular point and move with it together. Relations (3.1) will be satisfied if we have in three-dimensional designations n n B dσ = 4πb ·  Zn   B Vn dσn +EZσnσ Ddσn·dσ=n =4π4nbπVndn  , (a)(3.2) · × Zn (cid:20) (cid:18) (cid:19) (cid:21)  Because the surfaceZσnσσn(cid:20)Dm(cid:18)ayVnb·edaσnr(cid:19)bit−raHrily×sdmσna(cid:21)ll,=re4lπatdnioVnns (3.2) a.re boundary(bc)ondi- tions at the n-th singular point. There is the attractive idea, proposed by A.S. Eddington [5], about an invariance oftheoryunderpermutationofpointsofspace. Wecanapplythisideatothesingular points. Let the model be invariant under permutation of any two singular points in n1 n2 n1 n2 the sense of change their charges d d, b b. Because the theory is invariant ↔ ↔ under change of charge’s sign, in this case we have for any singular point n n ¯ ¯ d = d , b = b , (3.3) ± ± ¯ ¯ where d and b are some positive constants of space. ¯ ¯ Thus we have the theory with three dimensional constants α, d and b. For ¯ ¯ the suitable dimensional system we can take α = 1 and d = 1 or b = 1. Thus we ¯ ¯ have the relation d/b as the single dimensionless constant of the theory. Below (in section 9) we shall connect this relation with the fine structure constant. 5 4. Dyon solutions and singularities System of equations (2.1) with relations (2.6) has the following exact dyon solution in a cartesian coordinate system yµ : { } yid yib Di = , Bi = , (a) r3 r3 yid yib Ei = , Hi = , (b) r√r¯4 +r4 r√r¯4 +r4 (4.1) A = dφ (r) , A = bφ(y) , 0 0 (c) A¯ = bφ (r) , A¯ = dφ(y) , 0 − 0 ¯ ¯ where b = b , d = d , ± ± r dr′ 1 r = yiy , φ (r) = , r¯ α2 d¯2 +¯b2 4 . (4.2) i 0 √r¯4 +r′4 ≡ q ∞Z h (cid:16) (cid:17)i The vector function φ(y) may be of two types: with infinite singular line and with semi-infinite one. Its components have the following forms: 1 y2y3 1 y1y3 φ = , φ = , φ = 0 , (a) 1 r ρ2 2 −r ρ2 3 or (4.3) 1 (r +y3)y2 1 (r+y3)y1 φ = , φ = , φ = 0 , (b) 1 r ρ2 2 −r ρ2 3 where ρ = (y1)2 +(y2)2 . Function (4.3qa) has the infinite singular line coinciding with the axis y3 and functions (4.3b) has the semi-infinite singular line in positive direction of y3. In a spherical coordinate system r,ϑ,ϕ the both functions φ(y) have non-zero { } ϕ-components only that can be written in the forms (a) or (b) cotϑ 1 ϑ (4.4) φ = φ = cot ϕ ϕ − r −r 2 accordingly. At the singular line the potentials A,A¯ are devoid of defined direction and their absolute values are infinity. At the singular point r = 0 the vectors E,H,D,B are devoid of defined direction and absolute values of D,B are infinity. The full electromagnetic energy (see (6.3) below) of dyon solution (4.1) is 2 ¯ = α−2βr¯3 , (4.5) E 3 (cid:16) (cid:17) 2 ∞ dr Γ(1) 4 where β = 1.8541 . ≡ √1+r4 h4√πi ≈ Z 0 6 The dyon’s energy has a space localization region. Let us denote the sphere- enclosed dyon’s energy as ′ for radius of the sphere r′. Center of the sphere is at E the origin of the coordinates yi . The numerical calculation gives that ′ = 0.5 ¯ { } E E for r′ = r¯ and ′ = 0.95 ¯ for r′ = 10r¯. E E WithhelpofLorentztransformation, shift, androtationofthecoordinatesystem yi we can obtain the following moving dyon solution in the coordinates xµ : { } { } Di = L¯i Dj εijlV H √1 V2 j − j l − , (a)  (cid:16) (cid:17). Bi = L¯ij Bj +εijlVjEl √1−V2 (cid:16) (cid:17).  Ei = L¯i Ej εijlV B √1 V2 j − j l −  (cid:16) (cid:17). , (b) (4.6) Hi = L¯ijHj +εijlVjDl √1−V2 (cid:16) (cid:17). A = L0 A +Li j A (z) µ .µ 0 .µR.i j , (c) A¯ = L0 A¯ +Li j A¯ (z)  µ .µ 0 .µR.i j  where Di,Bi,Ei,Hi,A ,A¯ are defined by formulas (4.1) ν ν with yi = Li xj Vjx0 aj and zi = i yj , j − − 0 R.j Lν is the Lor(cid:16)entz transformat(cid:17)ion matrix, Li L¯j = δi , .µ j l l 1 V L0 = , L0 = L = i .0 .i i0 − , 1 V2 1 V2 − − q q (4.7) 1 ViV ViV Li = δi + 1 j , L¯i = δi + 1 V2 1 j , j j  −  V2 j j − − V2 1 V2 (cid:18)q (cid:19) − q  aj are components of an initial position, i is a rotation matrix. 0 R.j 5. Variational principle for two potentials Here we propose some action such that the associated variational principle gives system of equations (2.1) with relations (2.6) and boundary conditions (3.2). This action has the following form: S = ( +1) +2πα2 µA ¯µA¯ √g(dx)4 , (5.1) µ µ L − Z h (cid:16) (cid:17)i α2 where + (E D+H B) , (5.2) L ≡ −H 2 · · = (D,B) according to (2.7), H H E,B are represented by A and D,H are represented by A¯ with (2.10). µ µ Using definition (2.5) we can easy write also the following expression for : L 1 = + . (5.3) L −2 L L (cid:16) (cid:17) 7 Thus the action S is invariant under general transformation of coordinates. If we substitute expression (5.3) for into the action S and take into account L ¯ that = (∂ A ), = (∂ A ), then independent formal variation of eight com- µ ν µ ν L L L L ponents of the potentials A , A¯ for the principle δS = 0 gives the following eight µ ν equations: 1 ∂ ∂ 1 ∂ ∂ √|g| ∂xµ "√|g|∂(∂νLAµ)# = 4πν , √|g| ∂xµ "√|g|∂(∂νLA¯µ)# = −4π¯ν . (5.4) For simplicity we didn’t consider here the singular lines of the potentials. Formally we can consider system of equations (2.1) with relations (2.2) as the system of eight equations for the eight unknown functions A , A¯ ((2.1a′),(2.1b′) for µ ν A¯ and (2.1a′′),(2.1b′′) for A ). Such system of equations agree with system (5.4). µ µ Now let us consider the action S (5.1) with the function in form (5.2). In L this case the dependence of Lagrangian on the derivatives of potentials differs from one that we have for the function in form (5.3). Thus the appropriate variational L principles with the function in forms (5.2) and (5.3) are different. L Substituting definitions (2.10), (2.3), (2.4) into (5.1) we obtain the following expression for the action that doesn’t contain δ-functions: S = ( +1) g (dx)4 + (5.5) L | | Z q N n n n n +2πα2 d A +V A b A¯ +V A¯ dx0 , 0 0 · − · n=1Z (cid:28) (cid:18) (cid:19) (cid:18) (cid:19)(cid:29)n X α2 = + ∇A ∂ A ∇ A¯ + ∂ A¯ ∇A¯ ∇ A . (5.6) 0 0 0 0 L −H 2 − · × − · × h(cid:16) (cid:17) (cid:16) (cid:17) (cid:16) (cid:17) (cid:16) (cid:17)i Note, though potentials A , A¯ (4.6c) (cid:0) µ µ (cid:0) for the dyon solutions are infinite at the sin- σ¯˙ (cid:0) (cid:0) (cid:0) (cid:0) (cid:0) (cid:0) gular lines, its averaging at the singular points (cid:0) (cid:0) (cid:0) (cid:0) hAµin, hA¯µin are finite. Thus the action S is '(cid:0)(cid:0) (cid:0)(cid:0)@@R finite for the dyon solution. σn (cid:0) (cid:0) dσ¯ = dσ¯˙ Because definition of the potentials (2.10) r may be used only outside of the singular set, @ @R expression for (5.6) is true only outside of & d%σ¯ = dσn L the singular lines. Let us enclose all singular set in a multi-tuple connected surface σ¯ with Figure 1: Part of the surface σ¯ for external (relative to the singularity) surface the dyon field configuration with semi- element dσ¯ . Let this surface be composed infinite singular line. n of parts of surfaces σ for the N segregated singular points and multi-tuply connected tubular surface σ¯˙ (in three-dimensional space) for parts of the singular lines outside neighbourhoods of the N points. For 8 example, the surface σ¯ for the semi-infinite singular line can have image as it is shown in figure 1. Let us designate the action S in space outside of the singular set as S. We can obtain variation of this action with partial integration. Thus we have ∞ ∂ ∂ δS = dx0 α2∂ D ∇ H δA α2∂ B+∇ H δA¯ dΩ −Z∞ ΩZ " 0 − × ∂B! · − 0 × ∂D! · # α2  n n (D dσ¯)δA (B dσ¯)δA¯ +2πα2 d δA b δA¯ 0 0 0 0 − 2 · − · n − n Zσ¯ h i Xn (cid:18) D E D E (cid:19) ∂ α2 n n + H dσ¯ (E dσ¯ +Bdσ¯ ) δA¯ 2πα2 bV δA¯ (5.7) Zσ¯ "∂D × − 2 × 0 # · − Xn · D En ∂ α2 n n + H dσ¯ (H dσ¯ Ddσ¯ ) δA+2πα2 dV δA "∂B × − 2 × − 0 # · · n Zσ¯ Xn D E + 2πα2 dn E+Vn B +nb H Vn D δan dσ¯ δaσ¯ × − × · − L ·  Xn (cid:20)(cid:28) (cid:18) (cid:19) (cid:18) (cid:19)(cid:29)n (cid:21) Zσ¯   where Ω is three-dimensional space without the region bounded by surface σ¯, σ¯ σ¯ σ¯ da σ¯ dσ¯ V dσ¯ , V , a is a points of the surface σ¯ . 0 ≡ − · ≡ dx0 (cid:18) (cid:19) Now we contract the surface σ¯˙ to the singular lines outside neighbourhoods of n the N singular points. Next we contract the surfaces σ to the N singular points. As result we have δS δS. → Let us find stationary conditions for the action S. By the general principle of n σ¯ calculus of variations [11], we can take δa = δa = 0 at first. We can also take ¯ ¯ continuous variations δA , δA and make, firstly, δA = 0, δA = 0 at the singular µ µ µ µ lines. Then, according to the first line in expression (5.7), we have ∂ ∂ α2∂ B+ ∇ H = 0 , α2∂ D ∇ H = 0 (5.8) 0 × ∂D 0 − × ∂B in the space outside of the singular lines. Using definition of the fields B,D,E,H through potentials (2.10), we obtain from equations (5.8) that 1 ∂ 1 ∂ E = H , H = H . (5.9) α2 ∂D α2 ∂B Nowwetake δA = 0 and δA¯ = 0 atthesingularlinesforcontinuousvariations µ µ 6 6 δA , δA¯ . We assume that the fields D,B,E,H are continuous at the singular lines µ µ outside the segregated singular points. Then we obtain boundary conditions (3.2) from expression (5.7) (line 2-4), using formulas (5.9). 9