Zitterbewegung structure in electrons and photons

David Hestenes Department of Physics, Arizona State University, Tempe, Arizona 85287-1504∗

The Dirac equation is reinterpreted as a constitutive equation for singularities in the electromag- netic vacuum, with the electron as a point singularity on a lightlike toroidal vortex. The diameter of the vortex is a Compton wavelength and its thickness is given by the electron’s anomalous mag- netic moment. The photon is modeled as an electron-positron pair trapped in a vortex with energy proportional to the photon frequency. The possibility that all elementary particles are composed of similar vortices is discussed.

PACS numbers: 10,03.65.-w Keywords: pilot waves, Hopf fibrations, Hall effect, zitterbewegung, spacetime algebra, QED

I. INTRODUCTION magnetic theory in terms of Spacetime Algebra (STA) to provide a context for two important new developments. The spectacular success of quantum electrodynamics The first is Blinder’s concept of a classical vacuum sin- (QED) gives physicists great confidence in Maxwell’s gularity. The second is Antonio Ra˜nada’sdiscovery of equation on the one hand and Dirac’s equation on the toroidal solutions to Maxwell’s equation. The two pro- other, yet something is missing in relations between vide complementary inputs to a new theory of the elec- them. With his usual penetrating insight, Einstein fo- tron and the electromagnetic vacuum. cused on the crux of the problem [1, 2]: “It is a delusion Section III begins with a review and extension of the to think of electrons and the fields as two physically dif- zitter particle model for the electron clock in [3] and up- ferent, independent entities. Since neither can exist with- dated in the preceding paper [4]. Then the model is out the other, there is only one reality to be described, grounded with fundamental constants. That leads to which happens to have two different aspects; and the the- Oliver Consa’s calculation [5] for the electron anomalous ory ought to recognize this from the start instead of doing magnetic moment and for its physical explanation. things twice.” The stage is set in Section IV for the main subject This paper proposes a synthesis of Maxwell and Dirac of this paper, namely, reconstruction of the Dirac equa- theories based on a new model for singularities in the tion as a constitutive equation for the vacuum with the electromagnetic vacuum. The model is suggested by a electron as a point singularity. The singularity generates remarkable relation between electron mass and vacuum the electron’s Coulomb field with toroidal zitterbewegung polarization proposed by Seymour Blinder. The only re- and magnetic field with the spin vector as its axis. I quirement is consistency with Maxwell’s equation. No call this Maxwell-Dirac theory. It is complementary to changes in the form of Dirac or Maxwell equations are the conventional Born-Dirac theory discussed in [4] in necessary, but the two are fused at the source. The so- the sense that motion is described by the same Dirac lutions seamlessly integrates electron field and particle equation in both. In Born-Dirac the electron charge is properties along lines proposed by de Broglie. They an- inert and responds only to action of external fields. In swer Einstein’s call for a unified electron theory with a Maxwell-Dirac the charge is active and generates an elec- unified Maxwell-Dirac theory. tromagnetic field. In this sense, Maxwell-Dirac may be Singular toroidal solutions of the Dirac equation con- regarded as an alternative to second quantization, though stitute a new class of wave functions, fairly called ontic we do not directly consider its relation to standard QED. states (or “states of reality” as Einstein might have put Section V proposes a new model for the photon as an it), because they have a definite physical interpretation electron-positron pair bound in a toroidal ring with the in terms of local observables of the electron and associ- diameter of a reduced Compton wavelength. The ring ated deformation of the vacuum. No probabilities are in- has quantized internal states that have been observed in volved. Electron states are thus characterized by a literal arXiv:1910.11085v2 [physics.gen-ph] 24 Jan 2020 photon diffraction [6]. That invites reconsidering in Sec- field-particle duality: field and particle coexist as a real tion VI the possibility that all elementary particles can physical entity. This appears to finesse the notorious self- be constructed from leptons, as proposed in a seminal energy problem. It implies there is no such thing as the analysis by Asim Barut [7, 8]. electron’s own field acting on itself, because particle and field are two different aspects of one and the same thing. Section VII discusses many particle systems and Section II reviews the formulation of classical electro- diffraction. Considering the enormous scope of Maxwell- Dirac theory, our treatment has many loose ends and is best regarded as a guide for further research, though the model of electron as a point particle with an inherent ∗Electronic address: [email protected]; URL: http://geocalc. periodicity is an essential feature in all variants of the clas.asu.edu/ theory. 2

2 II. CLASSICAL ELECTROMAGNETICS  C = Jm. (11)

To place our analysis in the most general context, we Though we shall dismiss the magnetic monopole current begin with an STA formulation of Maxwell’s electrody- Jm as unphysical, we shall see good reason to keep the mamics in accord with the authoritative presentation by complex vector potential for radiation fields. Sommerfeld [9]. Squaring the Faraday gives us scalar and pseudoscalar In STA an electromagnetic field is represented by a invariants which can be expressed in terms of electric and bivector-valued function F = F (x) on spacetime, appro- magnetic fields: priately called the Faraday. Its split into electric and F 2 = (E + iB)2 = E2 − B2 + 2iE · B. (12) magnetic fields is given by As first shown in [10], if either of these invariants is F = E + iB. (1) nonzero, they can be used to put the Faraday in the In a polarizable medium, the electromagnetic field density unique invariant form: is a bivector field G = G(x) with the split into electric F = feiϕ = f cos ϕ + if sin ϕ, (13) and magnetic densities given by

G = D + iH. (2) where the exponential specifies a duality transformation through an angle given by The important distinction between “field” and “field den- 2E · B iF ∧ F sity” or “excitation” is emphasized by Sommerfeld. We tan 2ϕ = 2 2 = . (14) are interested in G for describing properties of the vac- E − B F · F uum. Note that (13) determines a rest frame in which the The most general possible version of Maxwell’s equa- electric and magnetic fields are parallel without using a tion for the electromagnetic field is Lorentz transformation. In addition, the squared magni- tude of f is F = Je + iJm , (3) 1 2 2 2 2 2 2 where Je = Je(x) is the electric charge current and Jm = f = [(E + B ) − 4(E × B) ] , (15) Jm(x) is a magnetic charge current. Separating vector and pseudovector parts, we get which is an invariant of the Poynting vector for F . A null field can also be put in the form  · F = Je (4) F = feiϕ with f 2 = 0, (16) and so we can write  ∧ F = iJm. (5) f = e + ib = e(1 + kˆ) = fkˆ. (17) Using the duality between divergence and curl Hence we can put the null field into the form ( ∧ F )i = · (iF ) (6)   ˆ F = feiϕ = feikϕ, (18) and the anticommutivity of the pseudoscalar with vec- tors, the latter equation can be written showing that the duality rotation is equivalent to a ro- tation of vectors in the null plane. Of course, this is  · (iF ) = Jm. (7) significant for description of radiation fields. In the most general case, the Faraday F can be derived from a “complex” pair of vector potentials A = A(x) and A. Conservation Laws C = C(x), so we have

F = (A + Ci). (8) Using the identity The scalar and pseudoscalar parts of this equation give  · ( · F ) = ( ∧ ) · F = 0, (19) us from (4) and (7), we get the current conservation laws  · A = 0 =  · C, (9)  · Je = 0 and  · Jm = 0. (20) while equations (4) and (7) give us separate equations for fields produced by electric and magnetic charges: An energy-momentum tensor T (n) = T (n(x), x) de- scribes the energy-momentum flux in direction of a unit 2  A = Je, (10) normal n at spacetime point x. As discussed elsewhere 3

[11, 12], the electromagnetic energy-momentum tensor is B. The Classical Vacuum given by 1 1 In a thorough analysis of constitutive equations in T (n) = hF nGei1 = [F nGe + GnFe]. (21) Maxwell’s electrodynamics, E. J. Post [13] identified a 2 4 hitherto unrecognized degree of freedom in Maxwell’s where n is a unit normal specifying the direction of flux. equation for the vacuum. Regarding the vacuum as a With some algebra, the tensor can be expressed in the dielectric medium with variable permittivity ε = ε(x) alternative form, which separates normal and tangential and permeability µ = µ(x) at each spacetime point x, fluxes: Maxwell’s condition for the propagation of light in a vac- uum is given by 1 T (n) = [(G · Fe)n + G · (n · Fe) + (G · n) · Fe]. (22) 2 2 εµ = 1/c = ε0µ0. (31) Whence, Obviously, this leaves the impedance

∂ T (n) = ∂ · T (n) + ∂ ∧ T (n) = 0. (23) rµ n n n Z = = Z(x) (32) ε The vanishing of both scalar and bivector parts in this expression tells us the linear function T (n) is traceless as an undetermined function. To ascertain its value, Post and symmetric. further argues that charge should be regarded as an in- The divergence of the energymomentum tensor is given dependent unit qe rather than the derived unit e. The by standard rule for changing units is 2 ˙ ˙ 1 ˙ ˙ 1 2 qe T () = 2 hF Ge + F Gei1 = 2 [G · Je + F · Jf ], (24) e = . (33) 4πε0 where current Jf includes any polarization or magneti- Accordingly, the fine structure constant is given by zation currents. Note use of the overdot to indicate dif- 2 2 2 r ferentiation to the left. e qe qe µ0 Physical interpretation of the energymomentum tensor αe = = = . (34) ~c 4πε0~c 4π~ ε0 is perhaps facilitated by using a v-split to put it in the form Hence, as Post observes, the fine structure constant can be expressed as a ratio of two generally invariant 1 T (v) = [FG† + GF †]v. (25) impedances: 4 r Z0 µ0 2h From this we find the energy density αe = , where Z0 = ,ZH = 2 . (35) ZH ε0 qe 1 1 T (v) · v = 2 F · Ge = 2 (E · D + B · H), (26) This suggested to Post that the Hall impedance ZH is an intrinsic property of the electromagnetic vacuum. and the momentum density Blinder [14, 15] has shown that polarization of the vacuum in the neighborhood of a classical electron is T (v) ∧ v = 1 (D × B + E × H). (27) 2 uniquely determined by the very simple assumptions that For a unit normal n = nv orthogonal to v, we have (1) the energy density of the electron field is proportional to the charge density, and (2) the total energy in the field 1 determines the electron mass. We review Blinder’s argu- T (n)v = [F nG† + GnF †], (28) 4 ment to serve as a guide for generalizing it to the Dirac so electron. For a point charge in its rest frame the electric field E 1 and electric displacement D are given by T (n) · v = 2 (D × B + E × H) · n, (29) qe r qe r and the spatial flux in direction n is E = , D = ε(r)E = , (36) 4πε(r) r3 4π r3 1 T (n) = T (n) ∧ v = hEnD + BnHi . (30) where r = |r|. The total energy in the field is 4 1 1 Z 1 Z ∞ 1 q2 The treatment of energymomentum conservation in this 3 e 2 W = E · D d r = 2 4 4πr dr. (37) subsection is completely general, applying to models of 2 32π 0 ε(r) r the electromagnetic vacuum considered next as well as The charge density % is determined by material media. It sets the stage for specific applications  0 3  considered next as well as extensions to be considered in % −qe ε (r) δ (r) = ∇ · E = 2 2 + (38) the future. ε0 4π r [ε(r)] ε(0) 4

0 with ε = ∂r ε and the normalization subject to the constraint

Z ∞ Z ∞ 0 2 2 ε0ε (r)dr (x − z(τ)) = 0. (48) %(r) 4πr dr = −qe 2 = qe, (39) 0 0 [ε(r)] Hence, it generates the electromagnetic field which require that ε(∞) = ε0 and ε(0) = ∞. 2   Finally, assuming that W = mec and the charge den- qe 1  ∧ v F =  ∧ A = v ∧  + . (49) sity in (38) is proportional to the energy density in (37), 4πε(r) r λ0 we get To evaluate the derivatives, we use the constraint (48), 0 2 −ε0ε (r) qe which implies proper time τ = τ(x) as a function of po- 2 2 = 2 2 4 . (40) r [ε(r)] 32π mec ε(r)r sition with gradient Whence, x − z τ = ≡ k, (50)  c r λ  ε(r) = ε exp 0 , (41) 0 r where null vector k is independent of distance r. Conse- quently, the curl of v = v(τ(x)) has the simple form where 2 2  ∧ v = τ ∧ ∂τ v = c k ∧ v.˙ (51) qe 1 1 e λ0 = 2 = 2 (42) 8πε0 mec 2 mec Similiarly, the gradient of (47) gives us is recognized as half the classical electron radius and puts r = v(v ∧ r) + τ (x − z) · vc˙ that quantity into new perspective as a radius of vacuum = vˆr + crkk · v.˙ (52) polarization. Now we have an explicit expression for the vacuum Inserting all this into (49), we get charge density: F =  ∧ A qe λ0 %(r) −λ0/r   ∇ · E = 4 e = . (43) qe r ck · v˙ ck ∧ v˙ 4πε0 r ε0 = 3 + v ∧ k + . (53) 4πε(r) r r λ0 Sommerfeld [9] emphasizes that this does not have the dimensions of charge, and he interprets it simply as “di- This differs from the classical retarded field [16] only in vergence of the electric field.” Of course, he was not privy the last term, wherein the distance r in the denominator to our notion of vacuum polarization or its expression as is replaced by λ0. At first sight this seems wrong, be- a manifestly nonsingular quantity. However, the dimen- cause a spherical electromagnetic wave surely attenuates sions of charge and its singularity are explicit in with distance. That may be why Blinder did not asso-   ciate his vacuum impedance with the electron’s vector −qe 2 1 3 potential. On the other hand, if Blinder’s idea has any ∇ · D = ∇ = qeδ (r). (44) 4π r relevance to quantum mechanics it must be expressed through the vector potential, because that is the only Also note that E = −∇ϕ , where e mechanism for electromagnetic interaction. Indeed, if the −qe last term contributes to photon emission, it should not ϕe(r) = . (45) 4πε(r)λ0 depend on distance. Let us therefore withhold judgment until we examine how the vacuum impedance might fit This suggests a straightforward generalization of Blin- into quantum mechanics. der’s argument. To wind up our discussion of classical theory, we define By interpreting the variable r as the retarded distance a generalized displacement field by G = εF . Whence its between each spacetime point x and the path z = z(τ) of divergence for a charge at rest is a point charge with velocity v =z ˙ = c−1dz/dτ, the sim- Z ∞ ple form for the scalar potential in (45) leads immediately 3 4  · G = qeδ (r)v = qe dzδ (x − z(τ)) = Je, (54) to the following generalization of the Li`enard-Wiechert −∞ potential: in agreement with (44) and (4). These expressions for q v A(x) = e . (46) F and G = εF suffice to fit Blinder’s model for a point 4πε(r) λ0 charge in the vacuum into the general formulation of clas- Let’s call this the “Coulomb vector potential” to empha- sical electrodynamics in the preceding section. size its relation to the Coulomb scalar potential. The For a more realistic model of the electron, we need to retarded distance is defined explicitly by incorporate electron spin and magnetic moment as well. Blinder [15] tried that with a dipole model of the mag- r = (x − z(τ)) · v = |r| with r = (x − z(τ)) ∧ v (47) netic moment, and he deduced an exponential form for 5 the magnetic permeability µ analogous to that for ε, but Inserting this into Maxwell’s equation and using an over- with a different functional dependence. However, that dot for the time derivative, we have approach is inconsistent with Maxwell’s condition (31) relating µ and ε to the speed of light. Instead, from now (v˙ + iϕ˙v)eiϕ = −(∇v + i(∇ϕ)v)eiϕ. (59) on, we impose Maxwell’s condition by introducing the dimensionless vacuum impedance Cancelling exponentials and separately equating even and odd parts, we get r r ε µ0 ε µ0 +iϕ˙v = −∇v, (60) (x) = = = = eλ0/r (55) µ ε0 ε0 µ and I call the inverse of this function with some given value for v˙ = −i(∇ϕ)v. (61) λ0 a Blinder function in recognition of Blinder’s seminal contribution. Writing κ = −ϕ˙, we put (60) in the form

C. Radiation Fields +κv = −i∇v = ∇ × v. (62) As advertised, this is precisely the Beltrami equation, A remarkable new family of null solutions to Maxwell’s with a side condition ∇ · v = 0. If κ is constant, we equation was discovered by Antonio Ra˜nada in 1989 [17, can differentiate again to get an equation of familiar 18] and subsequently reviewed and extended in [19, 20]. “Helmholtz type:” He called them “knotted radiation fields” because electric and magnetic field lines are interlocked in toroidal knots (∇2 + κ2)v = 0. (63) that persist as fields propagate. As beautifully described in [21, 22], this has opened up an exciting new thread of Time dependence of the vector field is governed by eq. research on electromagnetic radiation. (61), which can be written Many researchers have discovered advantages in for- mulating the radiation field as a “complex vector” F = v˙ = −v × ∇ϕ, (64) E + iB with the null condition F 2 = F ◦ F = 0, but only Enk [23] has formulated it with geometric algebra with the side condition v · ∇ϕ = 0. Besides identify- to show that the imaginary unit must be interpreted as ing a role for Beltrami’s equation, this analysis serves to the unit pseudoscalar of spacetime. We shall demonstrate demonstrate some advantages of geometric calculus and that STA has further advantages preparing the way for the way it incorporates standard vector calculus. extending the theory to include toroidal fields described As emphasized by Enk [23], it follows immediately from Maxwell’s equation (57), that there is a whole hier- by the Dirac equation in a later Section. n As emphasized by Kholodenko [24, 25], a crucial el- archy of “parabivector” fields Fn ≡ ∇ F that satisfy ement in an electromagnetic knot is a “self-generating” ∂ F = −∇F . (65) vector field v = v(x) described by the “eigenvector equa- 0 n n tion” In particular, with F = ∇F0 we can define a vector po- tential ∇ × v = κv. (56)

This equation has been known since the nineteenth cen- F0 = A + iC (66) tury as the Beltrami equation and employed to model so that A is the usual magnetic vector potential given by vorticity in fluids. It appears again in magnetohydrody- namics, where it is called the force-free equation. And iB = ∇A = i(∇ × A) (67) in superconductivity it is known as the London equation. Kholodenko [26] has reviewed the vast literature on the with ∇ · A = 0, and C is an analogous vector potential subject across mathematics as well as physics from the for the electric field E. unifying perspective of contact geometry. For radiation fields the relation to Beltrami’s equation The Beltrami equation is even inherent in the free field is especially simple. When F 2 = 0, we can write Maxwell equation [25]. Indeed, using the spacetime split, Maxwell’s equation can be written F = E + iB = ρ(e + ib)eiϕ, (68)

∂0F = −∇F. (57) where ρ is a single scale factor for both electric and mag- netic fields so we can set e2 = b2 = 1. Then, from (25), For the unique reference frame and form specified by (13), the Poynting paravector P is given by we can write 1 1 iϕ(x,t) P = FF † = ρ2[ (e2 +b2)−ieb] = ρ2(1+e×b). (69) F (x, t) = v(x, t)e . (58) 2 2 6

For monochromatic radiation, all the time dependence Let’s call this toroidal gauge invariance. This implies † † † is in the phase, so we can write κ = −ϕ˙ as before, and that n = UUϕσUϕU = UσU , and thus reduces the de- Maxwell’s equation (57) becomes grees the freedom for n(r) from three to two, while main- taining its smooth covering of the entire sphere. This ∇F = iκF. (70) completes our formulation of the Hopf map using rotors in geometric algebra. Defining “complex” inner and outer products in terms of Ra˜nadahad the great insight to use the Hopf map to commutator and anticommutator parts, we can split this model magnetic field lines. The essential idea is already into two equations contained in Hopf’s original example for a map. Hopf 3 3 ∇ × F = κF, (71) recognized that S is isomorphic to R by stereographic projection, as expressed by ∇ ◦ F = 0. (72) u 2x/λ0 = , (78) 1 − u0 Thus we see Beltrami’s equation in a more fundamental n where λ0 is a length scale factor. Using (74) and writing role. Further, defining Fn = ∇ F , we get 2r = x/λ0, this can be inverted to give n ∇Fn = (iκ) Fn (73) 2r r2 − 1 u = , u = with r2 = r2. (79) from (70) with constant κ. Thus we have a whole nest of r2 + 1 0 r2 + 1 Beltrami fields. With this prelude on the structure of radiation fields, Thus, we have the explicit function we turn to Ra˜nada’sseminal insight into the toroidal U(r) = (r2 + 1)−1[(r2 − 1) + 2ir]. (80) structure of magnetic fields provided by the celebrated Hopf fibration. This can be inserted into (76) to give us an explicit ex- Hopf studied smooth maps from the 3-sphere S3 onto ample of a Hopf map, which has been thoroughly studied the 2-sphere S2 using classical techniques of complex in [28]. variable theory. However, it is simpler and more informa- Actually, we can eliminate the stereographic projection tive to exploit the fact that S3 is a 3-dimensional man- to produce an even simpler version of the Hopf map where ifold isomorphic to the group SU(2) = {U} of unitary the rotor is normalized by quaternions or rotors, to use a more descriptive term. √ Accordingly, we define a Hopf map as a rotor-valued U(r) = λ(1 + ir) with UU † = λ(1 + r2) = 1. (81) function U = U(r) defined on a dimensionless represen- 2 tation of physical space R3 with a fixed origin r = 0. Then, the normal map n(r) on S is given by Expressed in terms of Euler variables (u0, u) defined and n = UσU † = λ{σ + 2σ × r + rσr} discussed in [27], the rotor is normalized by = λ{(1 − r2)σ + 2[σ × r + (r · σ)r]}. (82) † 2 2 U = u0 + iu with UU = u0 + u = 1. (74) Thus, the normalization for the Hopf map is completely This determines the orientation of an orthonormal frame determined by the simple scaling factor λ(r) = 1/(1 + of vectors r2) = 1/(1+r2). The significance of this remarkable scal- ing factor was implicit in Ra˜nada’streatment of knotted † ek = UσkU , (75) radiation fields from the beginning, as is evident in his treatment of a static magnetic field in [19, 28]. In fact, though the Hopf map requires only that one of them, say Hopf’s original example already suffices to model a mag- σ (designated by dropping the subscript), serves as a 3 netic field with a suitable power of the scale factor, as we pole σ and unit normal n = UσU † for the sphere S2. now demonstrate for purposes of comparison. Accordingly, fixing the pole σ on S2 determines a Consider the following candidate vector potential: smooth mapping n(r) of a unit normal on the surface specified by the rotor function U(r): 1 2 A(r) = 2 λ (σ × r + σ). (83) † 2 n = UσU = u0σ + 2u0σ × u + uσu Differentiating = (u2 − u2)σ + 2[u σ × u + (u · σ)u], (76) 0 0 ∇λ2 = −2λ3∇r2 = −4λ3r (84) where the identity uσ = −σu + 2u · σ has been used to and reorder noncommuting vectors. It is crucial to note, how- ever, that the function n(r) in (76) is uniquely specified ∇(σ × r) = i∇(σ ∧ r) = 2iσ, (85) by the rotor function U(r) only up to a rotation about the pole, as specified by we obtain

† iσϕ/2 3 2 UϕσUϕ = σ where Uϕ = e . (77) ∇A = −2λ r(σ × r + σ) + λ iσ. (86) 7

Whence D. Geometric Calculus and differential forms

3 2 2 ∇A = −2λ i[r σ − r(r · σ) + irσ] + λ iσ. (87) Cartan’s calculus of differential forms is used widely in mathematics and increasingly in physics, despite some The scalar part of this equation gives us a single term significant drawbacks. Consequently, it is worth pointing ∇ · A = 2λ3r · σ, (88) out here that there is a more general Geometric Calculus (GC) that articulates smoothly with standard vector cal- which we will need to eliminate to achieve toroidal gauge culus and applies equally well to spinor-valued functions. invariance. For the moment, though, we are only inter- As a detailed exposition of GC is given in [16, 29], it suf- ested in the curl of the vector potential ∇×A = i∇∧A. fices here to illustrate how it relates to differential forms Accordingly, from (87) we get in the simplest case of applications to electrodynamics. In GC, the concept of directed integral is fundamental, ∇ × A = −2λ3[r2σ − r(r · σ) + σ × r] + λ2σ and the volume element for a k-dimensional integral is k k k 3 2 2 a (simple) k-vector d r = Ikd r with magnitude |d r| = = λ [2σ × r + 2r(r · σ) − 2r σ + (1 + r )σ] k d r and direction at r given by unit k-vector Ik = Ik(r). = λ3[2σ × r + 2r(r · σ) + (1 − r2)σ]. (89) For a closed k-dimensional surface Sk with boundary Bk, The Fundamental Theorem of Integral Calculus has the This is proportional to the normal field n(r) defined in general form (82). Hence, with a suitable choice of units, we can iden- Z I tify it with a magnetic field dkr0 · ∇0f(r0 − r) = dk−1r0f(r0 − r), (97) Sk Bk B = ∇ × A = λ2n(r) = λ2UσU †. (90) where f(r0 − r) is an arbitrary (differentiable) function, To make the vector potential (83) gauge invariant, we not necessarily scalar-valued. This reduces to the fun- simply need to eliminate terms orthogonal to the pole σ. damental theorem for differential forms when the inte- This is achieved by grands are scalar-valued. Note that the inner product with the volume element projects away any component A(r) = 1 λ2(σ + σ × (σ × r)), (91) of the vector derivative normal to the surface. 2 For a multivector-valued function A = A(r) with k- where the last term is just a fancy way of writing vector parts Ak =< A(r) >k, a differential k-form, or just a “k-form,” can be defined by

σ × (σ × r) = σ · (σ × σ) = r − (σ · r)σ ≡ r⊥. (92) k k αk =< d rA(r) >= d r · Ak. (98) This now satisfies the condition (77) for toroidal gauge The exterior derivative of a k-form is a (k − 1)-form de- invariance: fined by

† k k UϕA(r)Uϕ = A(r). (93) dαk =< d r · ∇A >= (d r · ∇) · Ak−1. (99) And the toroidal magnetic field is given still by The Hodge dual can be defined (up to a choice of sign for the pseudoscalar i) by 2 † iB = ∇ ∧ A = λ UiσU . (94) k k ∗αk =< d rA(r)i >= d r · An−k. (100) This prepares us for a straightforward generalization to Here of a couple of important examples of differential toroidal radiation fields. forms. As Ra˜nadahas demonstrated [18], the structure in a The unit “outward” normal n of the directed area el- monochromatic radiation field can be generated by an ement d2r is defined by d2r = −in d2r. Accordingly, the orthogonal pair of static vector potentials, as specified element of flux for a magnetic field B(r) is given by by (66), where d2r · (iB) = B · n d2r = d2r · (∇ ∧ A), (101) 2 2 F0 = (A + iC) = 2iAC = 2C × A. (95) which integrates to a familiar form of Stokes’ theorem. Whence the radiation field has the form An important example of a 3-form is the magnetic he- licity hm [23], defined as an integral over all space: iϕ F = (∇F0)e . (96) Z 3 hm = d r A · B. (102) Despite its appearance, this quantity is a relativistic in- variant. Our main interest in this result is its relevance to To make the directed volume element and the vector po- modeling the photon, which is considered in a subsequent tential explicit, we write Section. . A · (∇ × A) = A · (−i∇ ∧ A) = −i(A ∧ ∇ ∧ A). (103) 8

Whence, Z 3 † hm = d r · (A ∧ ∇ ∧ A) . (104)

Ra˜nada[18, 19] recognized that the hm can be identified with the Hopf index for a multi-valued vector potential with integer values. In Section IV we propose further to identify the Hopf index with the principal quantum number in atomic physics.

III. DIRAC THEORY OF THE ELECTRON CLOCK

This Section reviews key features of Dirac electron the- ory developed in the preceding paper [4] to prepare for extension to a unified Maxwell-Dirac Theory of the elec- tromagnetic vacuum in the following Section. Then we discusses empirical evidence the that the electron is a quantum oscillator that serves as a digital clock in physi- cal processes. Finally, the electron’s anomalous magnetic moment is explained and calculated to seven significant figures. FIG. 1: The “spinning frame” of local observables along an electron path is depicted in a direction orthogonal to the spin vector. A. Local observables on Dirac streamlines and the Dirac equation implies this is a conserved vector We have seen that, with Spacetime Algebra, the stan- current with dard matrix version of the Dirac equation can be refor- mulated as a real Dirac equation:  · (ρv) = 0. (110) e The Dirac current fills a given region of spacetime with Ψi − AΨ = m cΨγ . (105)  ~ c e 0 a congruence of non-intersecting timelike paths. Every µ path (streamline or fiber) x(τ) has a velocity v =x ˙ = Here A = A(x) = Aµγ is the electromagnetic vector po- e0(x(τ)) and a comoving frame of “local observables” tential for external sources and the unit bivector i = iσ3 encodes the crucial property of spin. The spinor “wave eµ = eµ(x) = eµ(x(τ)) (111) function” Ψ = Ψ(x) has the Lorentz invariant decompo- sition with µ = 0, 1, 2, 3. The vector 1 1 1 iβ s = e = e (x(τ)) (112) Ψ = ρ 2 Re , (106) 2 ~ 0 2 ~ 0 is identified with the electron spin vector on the path, where ρ = ρ(x) is scalar-valued density and “rotor” R = while the spin bivector R(x) is normalized to RRe = RRe = 1. The rotor R has a 1 unique decomposition into the product S = isv = 2 ~ e2e1 (113)

R = V Ue−iσ3ϕ, (107) specifies the plane of electron phase ϕ = ϕ(x(τ)). (See Fig.1). This characterization of local observables in Dirac 1 theory is completely general, without any approximation, where rotor V = (vγ ) 2 defines a boost to the electron’s 0 so it is equally applicable to Copenhagen or Pilot Wave center of mass, the spatial rotor interpretations of QM. Differences arise in the physical −iσ3ϕ1 −iσ1ϕ2 interpretation ascribed to electron paths. U = U1U2 = e e (108) describes kinematics of spin, and ϕ = ϕ(x) is identified as the quantum mechanical phase of the wave function. B. Pilot particle and electron clock The Dirac wave function determines a timelike velocity v = v(x) defined by Physical implications of the Dirac wave function Ψ(x) are studied in [4] by applying conservation laws of mo- Ψγ0Ψe = ρRγ0Re = ρV γ0Ve = ρv, (109) mentum and angular momentum to construct equations 9 of motion for local observables on a particle path. One outcome is the Pilot Particle Model (PPM), which pro- vides a new approach to studying the Hydrogen atom. Another outcome is the Zitter Particle Model (ZPM), which requires a subtle modification of Dirac equation so the electron path is a lightlike helix. The ZPM pro- vides the core for electron theory in all that follows. As explained in [4], the ZPM regards the worldline of the electron as a lightlike circular helix ze = ze(τ) with ve- locity u =z ˙e (Fig.2). The helix is centered on a timelike path z = z(τ) with velocity v =z ˙ normalized to v2 = 1. This sets a time scale for the proper time parameter τ. A length scale is set by identifying the radius of the helical path with

|ze(τ) − z(τ)| = λe = c/ωe, (114) where 2λe = ~/mec is the reduced Compton wavelength and ωe is the frequency of the circular motion called zit- ter. We refer to the point ze(τ) as the center of charge (CC) and to z(τ) as the center of mass (CM). As first proposed in [30, 31] and extended to lightlike particles in [3, 32], we define a comoving frame of local observables attached to the CM by

eµ(τ) = RγµR,e (115) where rotor R = R(τ) is normalized to RRe = 1. Then FIG. 2: In the Zitter Particle Model the electron path is we identify e0 = v and define re = ze − z = λee1 as the an oriented lightlike helix with an opposite orientation for radius vector for the zitter. Thus we can regard e1 as the a positron. “hand of the electron clock.” We identify the unit vector e as the tangential veloc- 2 As shown in [4], zitter mechanics can be described in ity of the zitter. There are two distinct senses for the terms of mass current m v, momentum p and spin an- circulation, which we identify with the electron/positron e gular momentum S. The chiral spin bivector S can be distinction called “chirality.” Accordingly, we have two expressed in several equivalent forms: null vector tangential velocities: S = ud = v(d + is) = ius, (121) e± = v ± e2 = Rγ±R,e with γ± = γ0 ± γ2. (116) Note that the null velocity u2 = 0 implies null spin bivec- Unless otherwise noted, we restrict our attention to the tor S2 = 0. It follows that the free particle zitter motion electron case here, and our choice of sign for the chirality can be reduced to a single equation: is in agreement with [32]. Accordingly, we define the ΩS = pv. (122) electron “chiral velocity” u by The bivector part of this expression gives us the spin u = Rγ+Re = v + e2. (117) equation of motion: The rotational velocity of the zitter is a spacelike bivector S˙ = Ω × S = p ∧ v. (123) defined by And the scalar part gives us ˙ Ω = 2RR,.e (118) mec = p · v = ±Ω · S, (124) so that where the sign distinguishes electron/positron chirality. Note that these equations are consistent with identi- e˙µ = Ω · eµ. (119) fying momentum p with either the timelike vector mecv or the null vector mecu. This ambiguity is resolved in In particular, the next Section, where eu is identified as a charge cur- rent. On the other hand, the opposite choice was tacitly r˙e = Ω · re = λeΩ · e1 = (e1e2) · e1 = e2, (120) introduced in [4] by dividing (123) with v to get in agreement with (117). p = mecv + S˙ · v. (125) 10

Of course, we could not instead divide by the null vector 1 fe = √ = fe. (132) u. The fact that introduction of electron charge seems to LeCe be needed to resolve this ambiguity about mass may be an important clue about the relation to charge and mass While the energy of the particle oscillates between elec- in the theory. tric and magnetic energy, the average energy has the con- stant value

LI2 CV 2 hf hf C. Electron clock and fundamental constants E = + = + = hf = ω. (133) 2 2 2 2 ~ The Josephson effect gives very accurate data for the As Consa observes, the above calculations are valid for quantum flux unit h/e, while the quantum Hall effect any elementary particle with charge e, vibration fre- 2 gives very accurate data for the Hall impedance h/e . To- quency f, and energy given by Planck’s equation E = gether they measure the fundamental constants [h] and hf = ~ω. [e] to an accuracy approaching 9 decimal places. As ar- gued by Post [33], this provides ample reason to replace the standard system of units [l, t, m, q] for [length, time, D. Electron anomalous magnetic moment mass, charge] by a system [l, t, h, e] with metric units [l, t] for length and time and topological units [h, e] for action and charge with integer values. Moreover, it supports The basic free particle solution of the zitter equations identification of the electron clock as the fundamental of motion is specified by a constant spacelike bivector mechanism grounding the units of action and charge. Ω = ω e e , (134) The zitter model of the electron clock has been mem- e 2 1 orably described by Consa [5] as a superconducting LC where ω = 2m c2/ is the zitter frequency. There is, circuit composed of two indivisible elements: “a quantum e e ~ however, a more subtle solution that was recognized only of electric charge and a quantum of magnetic flux, the recently by Oliver Consa [5], who also fully recognized product of which is equal to Planck’s constant. The elec- its extraordinary significance in explaining the intrinsic tron’s magnetic flux is simultaneously the cause and the anomalous magnetic moment of the electron. consequence of the circular motion of the electric charge:” When dealing with free particle solutions, it is conve-

eΦe = h. (126) nient to work with the spacetime split of the comoving frame (115) defined by v = e0 to introduce a frame of Motion of the charge causes a frequency dependent elec- relative vectors tric current I = efe and an electric voltage Ve = hfe/e † with a fixed ratio given by the Hall impedance: ek = ek(τ) = UσkU , (135)

Ve hfe/e h so that Ze = = = 2 . (127) Ie efe e e˙ = ω × e , where − iω = 2UU˙ † (136) The electron’s circular orbit at the speed of light consti- k k tutes the most elementary superconductor possible, gen- is the angular velocity. In the simplest solution, called erating the flux quantum circular zitter, the particle orbit is generated by rotating hf 1 h vector re = λee1 with period τe = 2π/ωe. Φ = V T = e = . (128) e e e e f e In the new solution, called toroidal zitter, the elec- e tron’s circular orbit precesses around a fixed axis in a Analysis of the quantum LC circuit is straightforward plane inclined at a fixed angle specified by the anoma- with standard techniques. The Capacitance (C) and Self lous magnetic moment. As the orbit precesses, it sweeps Inductance (L) are defined by out a toroidal surface as depicted in Fig.3. The precess- ing motion generates two orthogonal electric currents, as Φe h indicated in Fig.3. The first is a circular current with ra- Le = = 2 , (129) Ie e fe dius r1. The second is a poloidal current that circulates around the poloidal cross section with radius r2. e e2 In the zitter model the electron magnetic moment is Ce = = . (130) generated by the charge current circulating in electron Ve hfe zitter. The poloidal component of this current cancels Applying the formulas for an LC circuit, we get expres- out, thus reducing the net magnetic moment. Such can- sions for the impedance and natural frequency in the form cellation of poloidal currents in a solenoid is well known in the design of electrical antennas. Here it explains the ori- r Le h gin of the electron’s anomalous magnetic moment. Next Ze = = 2 , (131) Ce e we show how to calculate its value. 11

with ω1 = ω1σ3 and ω2 = ω2σ2. This differentiates to ˙ ˙ ˙ 1 U = U1U2 + U1U2 = − 2 [iω1U + Uiω2], (142) which, in accord with (136), gives us an explicit expres- sion for the rotational velocity of the motion:

˙ −1 † 2UU = −i[ω1 + U1ω2U1 ]

iω1τ = −i[ω1 + ω2 e ] = −iω. (143) Whence, the particle velocity is given by r˙ = ω × r. Since the electron circulates at the speed of light, we have  dr 2 r˙ 2 = = −(r ∧ ω)2 = r2ω2 − (r · ω)2 = c2, (144) dτ where r·ω = 0 for a free particle with the CM at rest. For an orbit precessing about the CM as depicted in Fig.3, the average velocity is reduced to c/g by a factor g called the helical g-factor by Consa. Accordingly, the radius r = |r| projects to a mean radius r1 with mean velocity ωr1 = c/g, and (144) gives us

r2ω2  r 2 2 = . (145) c gr1

FIG. 3: In a plane rotating with angular velocity ω1σ3 around Solving for g we get the CM, the electron’s lightlike circular orbit lies on a toroidal 2 2 −1 surface called the electron energy shell that projects to the g = 1 + (r2/r1) + 2r2 · r1 , (146) orbital plane as an ellipse with antipodes a and b. and averaged over one zitter cycle, the cross term gives

−1 2 1 2 We consider a particle orbit r = r(τ) composed of a 2 r2 · r1 = (r2/r1) cos θ = 2 (r2/r1) (147) pair of orthogonal rotating vectors with the combined result r = r1 + r2 = r1e1 + r2e3, (137) p 2 g = 1 + (3/2)(r2/r1) . (148) so the orbit circulates at the speed of light on a torus with radius r2. Depiction of this path in a frame rotating with Then Consa evaluated this quantity by comparing it with angular velocity ω1σ3 around the CM is given in Fig.3. Schwinger’s celebrated result from quantum electrody- The toroidal path can by parametrized with angles θ namics: and φ as follows. First, describe a circle in the σ1σ3 = g(QED) = 1 + α/2π = 1.0011614, (149) −iσ2 plane displaced along a line in direction σ1 by

−iσ2φ where g/2 corresponds to the standard definition for the r0(φ) = (r1 + r2e )σ1 † anomalous g-factor a0. = r1σ1 + r2U2σ1U2 . (138) Accordingly, by expanding the radical in (148) to first order we get Then rotate the circle around the σ3 axis to construct a torus given by 2 α g − 1 = (3/4)(r2/r1) = . (150) † 2π r(θ, φ) = U1r0(φ)U1 (139) The right side of this expression gives g a quantitative and generated by rotors value while Fig.3 gives it a physical interpretation, where

−iσ3θ/2 −iσ2φ/2 r1 and r2 are outer and inner radii of an electron energy U1 = e and U2 = e . (140) shell. The outer radius is a known quantity, the zitter ra- A closed curve r(τ) with constant frequencies is gener- dius λe = r1, while the inner radius r2 is a free parameter in the model. ated by setting θ = ω1τ and φ = ω2τ. The rotors com- bine to define a single spinor generating toroidal zitter: According to (124), the energy in free electron zitter is given by mec = Ω · S = ω · s, so with constant spin −iω1τ/2 −iω2τ/2 U = U1U2 = e e = U(τ), (141) S = is, the g-factor describes a zitter frequency shift due 12 to free precession producing an energy (mass) shift given the Dirac wave function Ψ that seems to do the trick. The by essential idea is to give back to the electron its charge and electromagnetic field, which were ignored in the original (me + ∆me)c = Ω · S. (151) neutered version of Dirac theory. This necessarily local- izes the electron to the point source of its Coulomb field. The same result has been found by others [34]. The trick is to do it in a way that is consistent with well Conventional QED provides no physical explanation established features of Dirac theory. whatsoever for the Schwinger result (149), despite its mathematical sophistication and claims for unprece- According to the Born rule, the Dirac current Ψγ0Ψe = dented accuracy. Indeed, I heard Feynman himself de- ψγ0ψe = ρv is to be interpreted as a probability current clare that such lack of physical insight would not be tol- so the dimensionless function ρ = ρ(x) = ψψe must be a erated in any other branch of physics. probability density; therefore eρ(x) should be interpreted But now, with (150) and (151), we have fresh theoret- as a “probable” charge density. In the early days of QED, ical insight explaining the electron anomalous magnetic Furry and Oppenheimer [37] called this proportionality moment as free precession of electron zitter, quite anal- into question by asserting that eρ(x) must be interpreted ogous to free precession of a rigid body in classical me- as a physically real charge density to enable comparison chanics. That should drive a new round of research into with “real charges” in classical electrodynamics that pro- its implications. duce real electromagnetic fields. Indeed, second quan- In the meantime, there is plenty of room for design of tization was soon invented to do that, but it involves experiments to probe related issues, especially on the role some dislocation from the original Dirac theory. How- of magnetic resonance. For example: ever, there is a subtle alternative to this approach that Rotor methods for magnetic resonance measurement has not been heretofore considered. are given in [27], along with many other coordinate-free Our formulation of Born-Dirac as a classical field the- applications to rotational dynamics. Though a prop- ory in [4] facilitates comparison with Maxwell’s electro- erly tuned magnetic resonance measurement may acti- magnetics. To coordinate the two to produce a fully in- vate toroidal zitter in the electron, external fields are not tegrated Maxwell-Dirac theory, we need some educated necessary to maintain it. guesses to guide us. To that end, I propose three funda- If the toroidal zitter can be quenched or activated at mental anzatz’s, will, then the electron has at least two distinguishable First, we invoke “de Broglie’s clock ansatz” already internal states and might thereby serve as the ultimate articulated in the ZPM model of the electron clock in magnetic storage device. On the other hand, if toroidal the preceding Section. zitter is a persistent intrinsic property of the electron, Second, we note from Section II that Blinder’s assump- then circular zitter should simply be regarded as an ap- tion that charge density is proportional to mass density proximation. in Maxwell theory can be carried over to Dirac theory Such issues about toroidal zitter are likely to be central where both quantities are associated with the Dirac cur- to elementary particle theory as proposed in Section VI. rent. Hence Blinder’s argument relating electron mass to impedance of the vacuum should apply to the interpre- tation of the Dirac wave function in some way. Let me IV. MAXWELL–DIRAC THEORY call this the “Blinder ansatz.” Third, we recall London’s assumption that for elec- Born-Dirac theory supports the “Pilot–Wave” inter- trons in a superconductor the magnetic vector potential pretation of quantum mechanics originally proposed by is proportional to the charge current [38]. As all interac- de Broglie [4]. But that is only half of de Broglie’s pro- tions in the Dirac equation are mediated by vector poten- posal, which he called double solution theory [36]. In con- tials, we look for a comparable relation to the electron’s sonance with his realist perspective on quantum mechan- magnetic vector potential. Let’s call that the “London ics, he proposed that there must be two distinct kinds of ansatz.” solution to the wave equation. Besides the pilot wave, Now, to incorporate zitter into the Dirac equation in there must be some kind of singular solution describing accord with “de Broglie’s clock ansatz,” we simply re- a real particle without involving probability. In other place the timelike velocity v = Rγ0Re in the Dirac current words, he proposed a unique kind of wave–particle dual- (110) with the lightlike velocity of the electron’s helical ity that might better be called field–particle duality. path: De Broglie insisted that relativity is an essential ingre- dient of fundamental quantum mechanics, and he noted u = R(γ0 + γ2)Re = e0 + e2 =z ˙e(τ), (152) that monochromatic plane wave solutions of the relativis- This decomposes the zitter into the timelike CM velocity tic wave equation determine well–defined particle paths, just as we have seen for the Dirac equation in Born-Dirac v = Rγ0Re =z ˙(τ) =u ¯(τ) (153) theory. However, he never found a convincing way to de- fine a singularity that represents a physical particle. In and the fluctuating spacelike velocity e2 = Rγ2Re of the this section we introduce a new physical interpretation of circular zitter withe ¯2 = 0. In a similar way we identify 13 the spacelike spin bivector in (112) as the zitter average undetermined, so it can be multivalued. This mechanism S¯ of the lightlike spin bivector (121). can also serve to pick out a particle path in Pilot Wave With the electron’s helical path embedded in the wave theory. function Ψ(x), its functional form reduces to Ψ(x−ze(τ)) If indeed the electron’s Coulomb potential is already and factors into the product: inherent in the Dirac equation as the Blinder ansatz re- quires, then we should expect to find the electron’s mag- Ψ = V Uψ, (154) netic potential there as well. This leads us to examine the Gordon current, where we note that it includes a “magne- where zitter kinematics is incorporated in the rotor tization current.” Proposing this as a specific realization of the “London ansatz,” we reinterpret that term as a U = U(τ) = U (τ)U (τ) (155) 1 2 “magnetic potential.” Thus, in analogy with the Blinder with form specified by (140) potential (159) we write e 1/2 −iϕ −(α+iσ3ϕ) A =∼ − · (ρS). (161) ψ = ρ e = e . (156) c m  has the familiar form for a Schr¨odinger wave function, The congruence sign =∼ serves to indicate that the two but, of course, with an imaginary unit specified by the quantities are mathematically equivalent but have differ- bivector i = iσ3. ent physical interpretations. In Born-Dirac theory pre- A truly remarkable implication of the wave function sented in [4] the spin density ρS represents a distribution ψ = ψ(x − z˙(τ)) given by (156) is its reduction of cou- of spins associated with distinct particle paths. As will pling to the electromagnetic vacuum to a pair of retarded become evident, in the present Maxwell-Dirac theory the electric and magnetic scalar functions α(x − z˙(τ)) and analogous quantity represents a density of magnetic field ϕ(x − z˙(τ)). Details are provided by the complementary lines. Combining the electron’s electric and magnetic ansatzes of Blinder and London. vector potentials, we have a complete analogy with the The Blinder ansatz identifies the Dirac current mecρv entire Gordon current: with the Blinder form for the retarded potential of a point e e A = (A + A ) =∼ m cρu − · (ρS). (162) charge (46) by assuming that the Dirac density ρ = ρ(x) c e c c m e  is reciprocal to tne impedance  of the vacuum. Thus, it holds that According to (159), therefore, this reduces the Dirac equation to: −1 −α ρ =  = e where α = λc/r, (157) e e ρp = ρ(P − A) = Ae. (163) 2 2 c c while the classical electron radius λc = e /mec serves as a charge/mass scaling length and r = (x − ze(τ)) · v This equation describes a remarkable duality between the is the classical retarded distance from a point singularity charge current along the electron path and the electro- at the position ze(τ) of the electron. In other words, we magnetic field it generates and propagates by Maxwell’s identify equation: −1 e e2 (ρ Ae) = 0. (164) Ac ≡ u = mecρu, (158) c cλc Recall that the Blinder function for ρ, given by (157), or, more simply, vanishes on the electron path, so the zero can be factored out of the momentum balance equation (163). With (163) we have boiled down the Dirac equation λcAc = eρu, (159) to a relation between the electron’s vector potential Ae with the classical “Coulomb vector potential” Ac for a and what we can now identify as an energymomentum point charge. Acceptance of this “Blinder ansatz” com- density of the vacuum ρP . We note that, as an equation mits us to solutions of the Dirac equation as a function of for momentum balance, it comes close to the ideal of retarded position. Note that the source CC of the electric putting the electron vector potential Ae on equal footing field ze(τ) is displaced from the CM z(τ) by the radius with the vector potential A for external interactions. The vector only difference is that the density ρ is an essential part of Ae but not of A. Later on, we take this as a clue to re = ze(τ) − z(τ) = λee1, (160) a many particle generalization. For present purposes, we can ignore external interactions, though we maintain the where e1 is the electron “clock vector” and λe = ~/2mec electron’s interaction with the vacuum. is the zitter radius. Electric and magnetic components of the electron po- A crucial feature of the Blinder function is that tential Ae are separated by a spacetime split with respect ρ(ze(τ)) = 0 everywhere along the electron path ze(τ), to the CM velocity v; thus and thus at a single point on any 3-D spacelike hyper- surface. At that point the phase in the wave function is Aev = Ac · v + Am ∧ v = eρ/λc + Am. (165) 14

Then a v-split of the canonical momentum P gives us (104) we identified it with the Hopf index and the prin- cipal quantum number in atomic physics. It can be re- P v = P · v + P ∧ v = P0 + P = mec + ~∇ϕ. (166) garded as one more prediction of Dirac theory, provided the above identification of the electron vector potential Hence, electron energy is identified with the electric po- with the Gorden current is accepted. In that case, the tential flux quantum ranks with electric charge as a fundamental e property of the electron, as, indeed, many since London P = m c = A · vρ−1, (167) 0 e c c have suggested. More precisely, it supports interpreting Dirac theory as modeling the electron as a fundamental while momentum is identified with the magnetic poten- singularity of the vacuum. tial: e P = ∇ϕ = A ρ−1. (168) A. Electron Self-Energy and Zitter ~ c m Note that, though electric and magnetic fields are sepa- Zitterbewegung is said to contribute to electron self- rable in the Gordon current (162), they are not indepen- energy in QED, though that can be questioned because dent. They are intimately coupled by the vacuum density the integrals are divergent and must be removed by renor- ρ = ρ(x), as we now show. malization. Weisskopf [40] was the first to discuss the Regarding the zitter average of spin s as constant, we role of zitterbewegung in QED explicitly. Expressed in calculate the magnetic vector potential from: our lingo, he argues that zitter generates a fluctuating electric field. But when he calculates the zitter contribu- −1 −c c tion to the energy in the field he gets a divergent result. Amρ = ∇ · (ρis) = ∇ × (ρs) eρ eρ In contrast, we show here that calculation with the zit-   c −λcs r × s ter model is not only simpler, but the result is finite and = ∇ × = ge , (169) 2 e r r3 equal to the expected result mec . This is one reason to suspect that the zitter model may generate finite results which is the familiar potential for a magnetic dipole, in QED. 2 2 where λc = e /mec is the “classical electron radius” According to the Blinder ansatz, the vacuum is char- and acterized by the Blinder form (157) for the Dirac density ρ = ρ(x) given by ge = cλc/e = e/mec (170) ρ = −1 = e−λc/r, (173) is the correct g-factor for a Dirac electron. What a sur- 2 2 prise! For ge has been derived here from a singular wave where the polarization radius λc = e /mec is a charge function unknown in conventional Dirac theory. Indeed, to mass scaling length, and conventional free particle solutions do not even have a r = (x − z (τ)) · v (174) magnetic moment! Moreover, the derivation connects the e magnetic vector potential to the Coulomb potential, as is the classical retarded distance from a point singularity expected in an integrated model of electric and magnetic at the position of the electron. forces in an electron. To study zitter fluctuations, we shift electron position Comparison of equation (168) with (169) shows that ρ to the zitter CM. That shift must be done in a way that serves as a Lagrange multiplier relating gradient to curl preserves the ”retarded distance” property of r. in the form With the backward Taylor expansion z(τ −τc) ≈ z(τ)+ v(τ)τc, we get ~∇ϕ = ∇ × π. (171) ze(τ) = z(τ − ze) + re(τ), (175) The general question of when the curl of a vector field π is equivalent to the gradient of a scalar has been studied by hence Kleinert [39]. He shows that is possible when the scalar function is the projection of an area integral – in physical ze = z(τ) + vτc + re, (176) terms, a flux generated by a current loop. Moreover, in where all functions are evaluated at time τ. Requiring general, the flux is multivalued, so we can conclude that 2 [vτc + re] = 0, we find I I I e −1 dx · A ρ = dx · P = dϕ = nh, (172) τc = λe/c (177) c m ~ as the time for a light signal to propagate from the zitter This assigns a definite quantum of flux Φ = ~/e to the center to the circulating particle. Accordingly, we have electron’s magnetic field, as anticipated by London and evidently measured in the quantum Hall effect [33]. In ze = z + vτc + re = z + λe(v + e1). (178) 15

2 2 2 Then requiring (x − ze) = r − r = 0, for the the zitter B. Zitter fluctuations in QED retarded position we get Without trying to be comprehensive, let us consider r = (x − ze) ∧ v = x − (z + re) (179) how zitter fluctuations can account for important QED with retarded distance results. The simplest case has already been considered in Sec-

r = |r| = (x − ze) · v = (x − z) · v + λc. (180) tion III D, where the electron anomalous magnetic mo- ment was explained as force free precession of the elec- This completes our characterization of the zitter vacuum. tron’s zitter orbit, and described by a single rotor with Now, to incorporate the effect of zitter into the elec- form given by (141) as tron’s electromagnetic field, we simply replace velocity v −iω1τ/2 −iω2τ/2 with u = v+e2 in (152) to get a Coulomb vector potential Ug = e e . (187) of the form Now note that his rotor can be incorporated into the e e2 u electron wave function Ψ by direct multiplication of the Ac ≡ = mecρu (181) rotor part U in (155) to get c cλc  0 To ascertain the implications of this change on the elec- U = UgU. (188) tron self energy, we need only consider how it modifies The net effect on the electron equation of motion is to the Coulomb field of a free electron: reduce the g-factor ge specified in (170) to 2 mec rˆ 0 F = ( ρ) ∧ u = eρ ∧ u = E + iB. (182) g = ge/g, (189) e  r2 e where the scaling factor g is given by (150) and related Note thatr ˆ · u =r ˆ · (v + e2) = 0, so we haver ˆ ∧ u =ru ˆ . equations. Hence, To make another point, it appears that zitter fluctua- F 2 = E2 − B2 + iE · B = 0. (183) tions will not alter the quantum conditions proposed for stationary states of a Pilot particle as long as they are Since G = ρ−1F = D + iH, this implies that resonant with the quantum periods. Indeed, resonance of zitter fluctuations with orbital motion may be the most 1 1 2 E · D = 2 B · H. (184) fundamental criterion for stationary states. Though zitter will not alter quantum conditions, it Hence, the total energy in the field is should alter energy levels by smearing out the Coulomb Z potential over the zitter radius λe. This kind of explana- 1 3 tion for the Lamb shift was first proposed by Welton [42]. W = 2 (E · D + B · H) d r Moreover, in s-states discussed in [4] the Coulomb oscil- Z 3 2 lator solutions of the zitter will carry the electron around = E · D d r = mec , (185) the nucleus at the distance λe instead of right through it. In the ground state of hydrogen the nucleus just sits exactly twice the result obtained if the electron velocity inside the zitter circle. This is an analog of the Darwin were v instead of u. As first suggested by Slater [41], term in wave mechanics. Calculations of the ground state the reason for the difference is expressed by (184), which energy are therefore especially sensitive to the model for tells us the potential energy density of the circulating the nucleus. charge is equal to its kinetic energy density. However, At a deeper level, if zitter resonances are characteris- this cannot be regarded as a fully satisfactory resolution tic of quantized states, they may play a role in electron- of the notorious electron self-energy problem until the electron interactions. Indeed, it has been suggested [43] relation of gravitational to inertial mass is understood. that the Pauli principle and Exchange forces may be ex- More generally, we note thatruvu ˆ rˆ = uvu = 2u, hence plained by zitter resonances. Possibilities for experimen- the field F has an energymomentum density tal test of those ideas are noted in the many electron e2ρ theory proposed below. T (v) = 1 F v Ge = u, (186) 2 r4 C. Gravitational binding of the electron where u = v +r ˙e exhibits momentum fluctuations due to the rapidly rotating vectorr ˙e. Note that these zitter fluctuations are not radiating, because the zitter velocity This Section has presented a well-defined model of is orthogonal to the zitter radius vector. It remains to be the electron as a point charge confined to a supercon- seen if they can be identified with vacuum fluctuations of ducting ring. But it raises a natural question: “What QED or the ubiquitous ground state oscillators proposed holds the ring together?” Remarkably, the obvious an- by Planck. swer “Gravity!” has strong independent support from 16

General Relativity. The main fact is that the celebrated These restrictions leave open the possibility of a more Kerr-Newman solution of Einstein’s equation involves a complex functional form for ψ(r) due to short range in- charged ring singularity with spin and g-factor just like teractions in the neighborhood of the singularity. our zitter model of the electron [44]. That fact has of- We note that the Blinder function has an alternative ten been dismissed as a coincidence, but has been given formulation suggesting a general property of vacuum sin- new currency by arguments for its relevance in elemen- gularities not limited to the electron or even to charged tary particle theory [45]. Without going into details, we particles. We write the electron’s Blinder exponent in note here that our treatment of electron charge density the form

(173) has been generalized to Kerr-Newman by Blinder 2 e /~c αe himself [46]. That promises to be a significant step in λc/r = = . (195) integrating gravitation with electrodynamics. mec/~v · (x − ze) ke · (x − ze(τ)) This suggests that any particle with kinetic momentum V. FROM ELECTRON TO PHOTON k = p/~ and position z(τ) will have a Blinder function of the form

We have seen how a singularity of the vacuum Dirac ρ = e−αe/k·(x−z(τ)), (196) equation specified by the Blinder function specifies the seat of the electron’s electromagnetic potential. For the so the particle is located at ρ = 0, and we drop the sub- sake of possible modification or generalization, let us script on ke to allow k to be a null as well as timelike. frame our assumptions in the most general terms. We There is no longer a suggestion here that the exponent saw in (154) that the electron wave function Ψ(x) = is the Coulomb potential of a charged particle. Here the Ψ(x − ze(τ)) can be cast in the general form fine structure constant αe acts as a kind of general scaling −1 constant for vacuum singularities, so it may play that role Ψ = Rψ = VUZ , (190) even in strong interactions, as argued by MacGregor [49]. Throughout this article we will identify the presence of an where zitter kinematics is incorporated in the rotor electron (or positron) with a zero of its Blinder function. As explained in [4], the positron wave function Ψ0 is U = U(τ) = U1(τ)U2(τ) (191) obtained from electron wave function Ψ by antiparticle given by (155), while conjugation:

1/2 −iϕ/2 −1 0 ψ = ρ e = Z , (192) Ψ → Ψ = Ψσ1. (197) with i = iσ3, and rotor R describes the spacetime kine- This generates a reflection of the electron’s comoving matics of the electron singularity. Here we propose to frame (154) given by: interpret Z as a complex impedance of the vacuum, so 0 0 0 ψ is the vacuum admittance. Considering the role of the eµ = eµ(τ) → eµ = R γµRe , (198) Blinder function (173) relating mass to the vacuum, we 0 propose further to identify ψ with the Higgs field, though with R = Rσ1 = VUσ1 given by (190). The net result we do not elaborate that suggestion in this paper. is a reflection of rest frame observables along the clock Let us refer to the multiplicative form (190) for a singu- vector axis e1 = e1e0 given by: lar solution of the Dirac equation as vacuum separability. e0 = −e e e . (199) It is noteworthy that the complex admittance ψ has the k 1 k 1 form of the Schr¨odingerwave function, which is indeed This result is needed to relate electrons and positrons to an approximation to the Dirac wave function [47, 48]. the structure of photons. That suggests that Schr¨odinger theory is fundamentally about vacuum singularities. Later on we shall consider possibilities for generalizing A. Photon structure the “vacuum impedance.” To retain essential physical features that we have already identified, we place the Hard on the spectacular successes of Dirac’s electron following two restrictions on the functional form of ψ. theory, de Broglie applied it to model the photon. De First, vacuum positivity: Broglie must have had high hopes for his photon theory, because he took the unusual step of announcing it with ρ(r) = ψψe ≥ 0, (193) fanfare and immediately translating it into English [50]. with ρ(r) = 0 only at r = 0. Second, for agreement with He argued cogently that the photon must be composed electrodynamics, we require that ρ reduce to the Blinder of an electron-positron pair described by the Dirac equa- function in the asymptotic region, that is: tion. Nevertheless, despite his deep insight and clever analysis, he was never able to bring his argument to a −λc/r ρ(r) = e for r  λc. (194) successful conclusion. Here we show how to realize de 17

FIG. 4: Photon structure: A photon is composed of an electron-positron pair circulating at the speed of light with a fixed angular separation ϕ0, where 0 ≤ ϕ0 ≤ ±π designates FIG. 5: Photon propagation: (a) Feynman diagram for pair left and right circular polarizations. creation (b) Photon propagation is centered along a lightlike path with tangent vector k.

Broglie’s proposal with the zitter particle model depicted in Fig. 2 for both electron and positron. We simply as- sume that both electron and positron have helical paths with a common center, but separated with a spacelike interval less than a Compton wavelength. As depicted in Fig. 4, in the rest frame of this electron- positron system, the particle motion projects to a circle with zitter radius λe where electron and positron are lo- cated with fixed angular separation ϕ0 and spin angular momentum 2s = ~σ3. Lets call this model of photon structure the photon ring. Actually, since electron and positron have opposite chi- rality (as indicated in Fig.5a), the timelike component of FIG. 6: Picture the photon as a moving ring with angular their paths exactly cancel, leaving only the spacelike seg- momentum ~ that generates an electromagnetic wave with ment e−e+ depicted in Fig 4. amplitude normalized to its energy. (Figure from [51] with a Now we assume that the photon ring with the e−e+ different but related interpretation) segment, and the spin and charge stored within it, prop- agates as a photon along a lightlike path with tangent model we conclude that the momentum of the emitted vector k, as depicted in Fig.5b. It is amusing to think of photon p = k must be collinear with with electron spin the e− and the e+ as terminals of a battery that drives ~ s. Among other things, this general result accounts for a current of constant energy ω around the ring. ~ the headlight effect in cyclotron radiation. As the photon moves, it generates an electromagnetic We complete our picture of the photon by project- field that can be represented algebraically by a vector ing the lightlike path described in Fig.5b into a space- potential like path shown in Fig.6. That depicts the photon as a

ikˆϕ ring that generates a circularly polarized electromagnetic A = ρ e e , (200) wave with some given frequency as it propagates with ve- locity ckˆ. It is tempting to picture the field generated in with amplitude ρ and phase ϕ(t) = k · z + ϕ0 with each cycle as akin to a smoke ring, so the whole wave k ·z = ωt−k ·z(t). Note that the variables for time t and train consists of a chain of discrete “circulating smoke frequency ω are set as initial conditions in the instanta- rings” much like the “vortex atoms” proposed by Lord neous rest frame of the electron emitting the photon. We 2 Kelvin in the nineteenth century. That would be con- also assume that the photon carries energy ~ω ≤ 2mec sistent with the experimentally observed countability of with the limiting value given for pair production (Fig.5a). photons, but it is a step beyond our present model. The photon propagates at the speed of light with mo- 2 The problem remains to square our model of the pho- mentum p = ~k, so p = 0. The photon energy ~ω = p · v ton with what is known about electromagnetic radiation is determined by the proper velocity v = v(τ) of the discussed in Section IIC. As explained there, the motion source when it is emitted. Therefore, without loss of of a photon is governed by Maxwell’s equation generality, we can describe photon emission in the in- stantaneous rest frame of the electron given by v = γ0, so ∂0F = −∇F, (201) we have the momentum spacetime split pγ0 = ~(ω + k). Then from the rest frame independence of our photon with the constraint F 2 = 0. For a photon we can write 18

moment µe. That may have theoretical implications for binding of the photon, but the more significant point is that it increases the degrees of freedom for the vector potential A specified in (200) from 1 to 2. Whence, with e · k = 0 the polarization vector e can be generated by rotor U = U1U2 and written

e = Uσ1U.˜ (205)

as specified before by (140) and (141). This provides strong theoretical grounds for predicting the existence of quantized toroidal states for individual photons. Experi- FIG. 7: The photon can be modeled as an electron-positron mentalists will be proud to announce that they got there pair located on a toroidal ring (or energy shell) with a fixed first! They have already detected toroidal states in the angular separation ϕ0 designating its polarization. diffraction of individual photons! [6] However, that may be a beginning rather than the end of the story, a presage of a richer landscape of toroidal ˆ F = E+iB = E(1+k), where the electric field E = ∂0A states in elementary particle theory to be considered in is determined by the vector potential given by (200). the next Section. The photon field F = F (ϕ) has the same functional form as a plane wave, but its phase function ϕ = k · z(τ) is centered on a lightlike curve z = z(τ). Hence, the B. Compton Scattering phase is given by ϕ = ωt − k · z. and Maxwell’s equation (57) reduces to the eigenvalue equation Compton scattering by a photon on an electron is gov- erned by the conservation law kF = ±F ω, (202) 0 0 where the signs correspond to states of left/right circular ~k + p = ~k + p . (206) polarization. Moreover, we can decompose F into the For an electron at rest we have p = m cv with v = γ . canonical form e 0 With a simple calculation in [53] this gives us a shift in F = E + iB = fZ, (203) the photon wavelength:

iϕ 0 ˆ ˆ0 where Z = Z(ϕ) = ρe can be regarded as a complex ∆λ = λ − λ = λe(1 − k · k ), (207) impedance of the photon singularity, and f is a polar- ization bivector with various forms given by Eqns. (17), where λe = h/mec is the Compton wavelength, the di- (18) and (68). In particular, we can write ameter of the electron zitter cycle. This is still the most direct and convincing evidence for zitter in electron and ˆ ikϕ0 ˆ photon structure. f = eˆ e = eˆ(cos ϕ0 + ik sin ϕ0) = e + ib, (204) It is noteworthy that Compton’s equation (207) seems † where ϕ0 is the polarization angle in Fig. 5, and ff = 1. to be independent of photon polarization. However, the Like the electron, the photon is a singularity in the propagation vector k is orthogonal to polarization as ex- electromagnetic vacuum field with density ρ = ρ(z) plau- pressed by < kf >= k · f = 0. Hence, the change of sibly described by (196). Moreover, it has singularity polarization in scattering is given by structure described as a circular ring in Fig.4 and Fig.6. That gives the photon a size and shape. One might worry kˆ0 · kˆ = f 0 · f † =< f 0f † > . (208) that the photon density (196) could propagate like the electron’s Coulomb potential to influence the photon’s This agrees with polarization factors in more elaborate surroundings in a way that has not been observed. How- STA treatments of scattering given in [54]. ever, it is a general theorem [52] that influence from a null surface, like the boundary of a photon path in the present model, will propagate only along that boundary. VI. VACUUM UNIVERSALITY In free space the photon moves in a straight line. How- ever, in a wave guide or optical fiber, the path is shaped The simplicity and power of modeling the electron as by the material walls that modify the parameter ρ. a vacuum singularity strongly suggests that the space- To complete our model of photon structure, we note time vacuum can be regarded as a universal medium for that it can be refined by replacing the circular ring in the physical world, so all elementary particles can be re- Fig.4 with a toroidal ring as depicted in Fig.7. As ex- garded as vacuum singularities of various types. Success plained in Section IIIC, the width of the toroidal energy in explaining quantum mechanics and QED for the elec- shell is determined by the electron anomalous magnetic tron promises strong support for the general thesis that 19 the Dirac equation describes spacetime dynamics of vac- ton has the form uum singularities. Thus we have new prospects for a unified vacuum field theory of elementary particles. −iσ3ωkτ −iσ2ωkkτ With electrons modeled as vacuum singularities, it is U1(τ)U2(τ) = e e , (210) natural to consider the topology of more complex vac- uum singularities to model the whole zoo of elementary particles, including photons. A promising possibility is where ωk = mkc/~ is a constant frequency specified by based on the fact that, in a certain sense, the electroweak each leptonic mass mk, and each generation {e, µ, τ} is gauge group is already inherent in the Dirac equation, indexed by its writhe number k = 1, 2, 3. Why the num- and a gauge theory version of gravitational interactions ber of generations is limited to three is unknown, though is readily included as well. A unified “gravelectroweak it may be due to a mass constraint. A natural geometric theory” of that kind has already been described in [55]. explanation for the quantum number k has been already It suffices to summarize its main features here and discuss noted in Consa’s model for toroidal zitter. what Vacuum Dirac theory has to add. A fundamental feature of Electroweak Theory is that The general idea is that gravity is about deformation each charged lepton is paired with a unique neutrino com- of the vacuum due to presence and propagation of singu- panion. As noted in [43], this fits with the fact there larities described by the Dirac equation. The implication are only two possibilities for the path of a zbw center: that all elementary particles and their interactions can timelike or lightlike. It conforms nicely with Barut’s pro- be described by variations and excitations of the vacuum posal [8] that the electron carries the neutrino within impedance promises closure to the search for a Unified itself along with its electromagnetic field. Field Theory. An attractive possibility is that the energy in the electrons anomalous magnetic moment might be carried away by a neutrino much like beta-decay. Indeed, if cir- A. Lepton wave functions cular zitter is an intrinsic geometric property of the elec- tron, as suggested by (136), then it may well be decoupled The electron kinematic rotor (191) can be expressed in from charge to produce a neutral particle. There seems the more explicit form to be no alternative account for the neutrino, because we have used up all degrees of freedom in the wave function. U(τ) = e−iσ3φ e−iσ2θ. (209) Moreover, this leads by analogy to testable experimental predictions about muon and tau neutrinos. Now suppose that both angles are harmonically related functions of proper time φ(nτ) and θ(kτ), where (n, k) This completes our discussion of kinematic rotors for is a pair of coprime integers known in knot theory as modeling leptons. But a few words are in order to fit it rotation and writhe numbers respectively. The angle ϕ into Barut’s account of elementary particles [7, 8]. generates circular zitter in the spin plane, while the angle In the Standard Model of elementary particles, the θ tilts the plane, so together the two angles can be ad- three generations of leptons stand in obvious analogy to justed to generate a family of closed toroidal curves (or the three generations of quarks. That suggests that they helices) with periods in ratio k/n. must be related in some way. Since the leptons are ob- Regarding each helix as the path of a point singular- servable particles while quarks are not, the simplest pos- ity like the electron, we have here a family of singularity sibility is that all hadrons are composed of leptons. The types distinguished by quantum numbers (n, k). Oth- quarks then represent symmetries in the lepton compo- ers have proposed knot theory for classifying elementary sition rather than actual particles. particles, though not with such a direct tie to the Dirac equation and geometry of the vacuum. Jehle [56, 57], As noted by Barut, to explain nucleons as composed of for one, proposed a classification based on quantized flux leptons, we must explain isospin and strangeness quan- tum numbers in terms of leptons. To do that, Barut that has much in common with the present approach, but ± it is weak on connection with field equations. Finkelstein observed that the number of µ mesons in hadrons is [58, 59] has developed a detailed match of knot topology exactly equal to the “strangeness” quantum number in with structure of the Standard Model. Knot topology can hadrons. Evidently “strange” hadrons decay into ordi- be taken as supplementing the present approach, which nary hadrons if the µ inside the hadron decays. In strong is based on differential geometry using STA. interactions µ is stable, hence strangeness is conserved. In the following we take standard Electroweak The- Barut goes on to generalize the electron’s Coulomb po- ory as given, at least in the main. The lepton family tential to include magnetic interaction that dominates in − − − is composed of three generations: (e νe)(µ νµ)(τ ντ ) ranges less than a Compton wavelength. This remains and their antiparticles. Our aim is to incorporate it into a work in progress to describe fundamental interactions Maxwell-Dirac theory by identifying the various leptons of the electron in electromagnetic terms. Presumably, it with their kinematic rotors in the form given by (209). can be incorporated into a generalization of the Blinder Like the electron itself, the kinematic rotor for each lep- function. 20

B. Gravelectroweak Gauge Theory circulating charges are distributed on the torus in some symmetrical way that generates the proton’s magnetic A natural extension of the Dirac equation to include moment. The similarity of this toroidal coupling with weak interactions rests on the unique fact that the elec- the three quark model of the proton is noteworthy, and troweak gauge group SU(2)⊗U(1) is the maximal group would seem not to require gluons for binding. of gauge transformations Ψ → Ψ0 = ΨU that leave the velocity observable invariant: VII. MANY PARTICLE THEORY 0 0 ρu = Ψγ0Ψe = Ψ γ0Ψf, (211) As a minimal generalization of Vacuum Dirac Theory This gives the gauge group geometric significance as the to a many particle theory, we consider a system of N invariance group of the Dirac current, thereby insuring electrons regarded as particle singularities in the vacuum a spacetime path for the zitter center. To incorporate with zitter velocities uk and CM spacetime paths zk = both gravitational and electroweak interactions in vac- zk(τk) with proper velocities vk =z ˙k. We suppose their uum Dirac theory, we require invariance under the group motions are determined by a spinor wave function for the

0 vacuum Ψ = Ψ(x, z1, z2, ..., zN ). The vacuum density is Ψ → Ψ = LΨU, (212) then given by where LL = 1. The corresponding gauge invariant N e Y derivative is ΨΨe = ρ = ρk =ρ ˇkρk, (214) k=1 1 DµΨ = (∂µ + 2 ωµ)Ψ − ΨiWµ, (213) whereρ ˇk designate the product with the kth factor omit- where the geometric “connexion” ωµ expresses gravita- ted, and, as before, tional interaction and the Wµ express electroweak inter- −αk actions. See [55] for an extensive account of the details. ρk = e , (215) Applied to modeling the electron, the gauge derivative where α = α(r ) is a Blinder potential or its generaliza- (213) leads [55] to a generalization of the Kerr-Neuman k k tion with retarded position rk = (x − zk) · vk. solution of Einstein’s equation with a Blinder function As the basic equation for energymomentum density in yet to be determined. This promises to be a pathway the vacuum, we consider a straightforward generalization uniting gravity with electromagnetism. of the Gordon current in Born-Dirac theory, namely

N N e X X C. Stable Composite Particles ρP = A ρˇ = ρ p , (216) c k k k k=1 k=1 The upshot of Electroweak Theory is that all leptons where, as before, are just different states of the electron. This suggests that all elementary particles and resonances can be built e A = ρ p = m cρ u − · (ρ S ) (217) out of electrons. That possibility had been studied at c k k k e k k  k k some depth by Asim Barut [7, 8] until his unfortunate is the electromagnetic vector potential (Gordon current) sudden death. Now the insights into electroweak gauge of the kth particle modeled with or without zitter, and theory outlined in the preceding Section open up new possibilities. ρPµ = h(∂µΨ)iΨei (218) As Barut observed, aside from the electron and neu- ~ trino, the only stable elementary particles are the photon defines components Pµ = γµ · P of the canonical momen- and proton. It follows that the whole spectrum of el- tum. ementary particles (and antiparticles) can be generated Since equation (216) is the core synthesis of Maxwell’s by leptons if the photon and proton can. Let us consider electrodynamics with Dirac’s electron theory, consolidat- how the zitter model of electron structure may make that ing what we have discussed already and providing a plat- possible. form for extensions to follow, let me christen it with the name Maxwell-Dirac Equation. Generalization to include other fermions has been dis- 1. Proton structure cussed already. Restricting our attention to electrons for the moment, we note that equation (216) has obvious im- Given the model of the photon as an (e+e−) pair, we plications for the Helium atom, where it will treat both have an obvious extension to a proton model (e+e−e+) electrons on equal footing and imply correlations similar simply by inserting a positron centered at the CM with to the “exchange interaction.” Carrying out the calcula- its circulating charge located on the torus of the pho- tions would provide a stringent test of (216) with impli- ton. Presumably, in the proton bound state the three cations for the Pauli principle. 21

Equation (216) also meets Carver Mead’s objective for diffraction in general. We point them out here without a many electron quantum state determined entirely by delving into detailed calculations or experimental tests. vector potentials of all particles in the system [2, 60]. It The first and most important point is that, according goes beyond Mead in anchoring the electron state in a to (214), the density ρ = ρ(x) of a single electron factors Dirac wave function, in principle including the contribu- into a product tion of positive charges in the lattice of a superconductor. N Y Note that P in (216) can be regarded as kind of “super- ρ = ρ =ρ ˇ ρ , (224) potential” for the entire system. It follows, then, that k e e k=1 N X where ρe(x) is the Blinder function of Maxwell-Dirac F ≡ ∧ P = ∧ p (219)   k theory, andρ ˇe(x) =ρ ˇe(x, x1, . . . , xN−1) is the den- k=1 sity expressed with Blinder functions of all other par- can be regarded as the total electromagnetic field for the ticles with influence. Since the Blinder function satisfies entire system. 0 ≤ |ρk| ≤ 1, we also have 0 ≤ |ρ(x)| ≤ 1. So there are Inside a superconductor we have F = E + iB = 0 no normalization issues, and sufficiently distant particles (Meissner effect). Hence, as we have seen before, Stokes automatically have insignificant influence. Theorem implies that for any closed curve in the region One consequence is that the “Quantum potential” in the Pilot Wave guidance law must have a causal source I in matter composing the diffraction slits. To get the P · dx = 0. (220) “acausal density” of Pilot Wave theory, the matter coor- dinates must be integrated out with some sort of average And, as in the single particle case, we get a many particle hρi. That leaves the possibility open for fluctuations in quantization condition path density, for example, from heating material in the slits. Z Tn I 1 We still have the problem of identifying a plausi- P0 dt = P · dx = (n + 2 )h, (221) 0 ble mechanism for momentum exchange between each diffracted particle and the slits, a causal link which is with integer n. This agrees with Mead’s formulation of missing from all accounts of diffraction by standard wave phase and flux quantization in a superconductor [2, 60]. mechanics or by Pilot Wave theory. Note that mo- The relevance of this argument to the Aharonhov–Bohm mentum transfer is observable for each scattered parti- effect is also worth noting [61]. cle, whereas the diffraction pattern conserves momentum Specific application to superconductors is beyond the only as a statistical average. Evidently the only way to purview of this exploratory discussion. However, before account for this fact is by reducing diffraction to quan- dropping the subject, it is worth noting that the present tized momentum exchange between each particle and slit. model satisfies the additivity of electron phases ϕk = To that end, [62] provides a detailed analysis of optical ϕk(x − zk) that is essential for superconductivity. That diffraction patterns explained by photon momentum ex- can be made manifest by writing the wave function in change. the form Duane was the first to offer a quantitative explanation N for electron diffraction as quantized momentum exchange Y Ψ = R ΨkΛ, (222) [63]. A more general argument using standard quantum k=1 mechanics has been worked out by Van Vliet [64, 65]. These explanations suffer from the same disease as Old where RRe = 1, Λ2 = 0 and Quantum Mechanics in failing to account for the density distribution in the diffraction pattern. However, we now −αk−iϕk Ψk = e . (223) have the possibility of curing that disease with relativistic Pilot Wave theory. We only need to explain how the Evidently, the Pauli principle can be incorporated in momentum exchange is incorporated into the Pilot Wave symmetries of the wave function in the usual way, and guidance law. that would identify it as a property of vacuum singu- Now, presuming vanishing electric and magnetic fields larities! The symmetries need not apply to all particle outside the diffraction slits as before, we have ∧ A = 0, variables, but only to particles whose motions are res-  so locally, at least, A is a gradient. Assuming the same onant in some sense, as in the quantized atomic states for P , we have a gauge invariant phase gradient discussed earlier. e Φ = P − A. (225)  c A. Particle Diffraction This provides a promising mechanism for quantized mo- mentum transfer in diffraction. For we know that quan- Maxwell-Dirac theory has unique implications for the tized states in QM are determined by boundary condi- problem of electron diffraction, indeed, for particle tions on the phase. Successful calculation of diffraction 22 patterns along these lines would provide strong evidence of any charge. Evidently it applies to photons as well for the following claim: the vacuum surrounding electro- as electrons and protons, so it should be regarded as a magnetically inert matter is permeated by a vector po- universal property of the vacuum. This suggests associ- tential with vanishing curl. Remarkably, the same mech- ation with a gravitational field. That possibility is best anism would explain the extended Aharonov-Bohm (AB) approached by a gauge theory as proposed in Section VB. effect [66]. Evidently, then, the causal agents for diffrac- Strictly speaking, the density (impedance) of the vac- tion and the AB effect are one and the same: a universal uum should be incorporated into any vector potential vector potential permeating the vacuum (or, Aether, if by writing A = ρA, with a new notation to distinguish you will) of all spacetime, much as proposed by Dirac it from the usual vector potential, whether or not it is [67]. the gradient of a scalar field. The Aether can then be Considering the similarity of electron and photon regarded as a conserved fluid (with  ·A = 0) flow- diffraction patterns, we should expect the same mecha- ing through spacetime with particle singularities (elec- nism to explain both, especially if photons are composed tron, photon or whatever) in the density swept along. of electron-positron pairs as proposed in Section V. In- This picture has a beautiful macroscopic analog describ- deed, the evolution of path density for the electron is ing diffraction of a macro particle in a classical fluid [68]. determined by the Dirac equation, which gives

2  Φ = −mecz˙ ·  ln ρ. (226) B. Vacuum Topology For a photon with propagation vector k, the analog is

2 k ·  ln ρ =  Φ/~, (227) Derivation of equations of motion for singularities from gravitational field equations has been an important the- where, of course, ρ is the path density for photons, just as oretical objective since Einstein, Infeld and Hoffman at- it is for electrons. Accordingly, we conclude that diffrac- tacked it [69]. We have seen something like that for Dirac tion is “caused” by the vacuum surrounding material ob- theory. Conversely, if we develop a rich theory of singu- jects. In other words, diffraction is refraction by the vac- larities along the lines suggested here, that might require uum! modification of the field equations. Gauge theory may We have seen that the Blinder form for the vacuum then be regarded as a means for coordinating singulari- density ρ = ρ(x), which was originally introduced to gen- ties with equations of motion. eralize the Coulomb potential, is actually determined by “Still keeping one principal object in view, to preserve the momentum at each vacuum singularity independent their symmetrical shape!”

[1] E. T. Jaynes. Clearing up mysteries — the original Am. J. Phys., 71:691–704, 2003. goal. In W. H. Zurek, editor, Complexity, Entropy and [12] A. Lasenby, C. Doran, and S. Gull. A multivector deriva- the Physics of Information, Reading MA, 1990. Addison- tive approach to Lagrangian field theory. Found. Phys., Wesley. Revised and extended version 1996. 23:1295–1327, 1993. [2] C. Mead. Collective Electrodynamics I. Proc. Natl. Sci. [13] E. J. Post. The constitutive map and some of its ramifi- USA, 94:6013–6018, 1997. cations. Annals of Physics, 71:497–518, 1972. [3] D. Hestenes. The zitterbewegung interpretation of quan- [14] S. Blinder. Classical electrodynamics with vacuum polar- tum mechanics. Found. Phys., 20:1213–1232, 1990. Un- ization: electron self energy and radiation reaction. Rep. fortunately, the appendix is marred by many misprints. Math. Phys., 47:269–275, 2001. [4] D. Hestenes. Quantum mechanics of the electron particle [15] S. Blinder. Singularity-free electrodynamics for point clock. arXiv:1910.10478, 2019. charges and dipoles: a classical model for electron self [5] O. Consa. Helical solenoid model of the electron. Progress energy and spin. European Journal of Physics, 24:271– in Physics, 14(2):80–189, 2018. 275, 2003. [6] S. Johnstone S. Tempone-Wiltshire and K. Helmerson. [16] D. Hestenes. Differential forms in geometric calculus. In Optical vortex knots–one photon at a time. Nature Sci- F. Brackx et al., editor, Clifford Algebras and their Appli- entific Reports, 6:24463:1–6, 2016. cations in Mathematical Physics, pages 269–285. Kluwer, [7] A. O. Barut. Stable particles as building blocks of matter. Dordrecht/Boston, 1993. Surveys in High Energy Physics, 1:113–140, 1980. [17] A. Ra˜nada. Topological theory of the electromagnetic [8] A. O. Barut. Unification based on electromagnetism. a field. Lett. Math. Phys., 18:97–106, 1989. simple composite model of particles. Annalen der Physik, [18] A. Ra˜nada. Topological electromagnetism. J. Phys A: 498:83–92, 1986. Math. Gen., 25:1621–1641, 1992. [9] A. Sommerfeld. Electrodynamics. Academic Press, New [19] A. Ra˜nadaand A. Tiemblo. A topological structure in York, 1952. the set of classical free radiation electromagnetic fields. [10] D. Hestenes. Space Time Algebra. Gordon and Breach, 2014, arXiv 1407.8145v1. New York, 1966. 2nd Ed.: Birkh¨auser,New York 2015. [20] M. Array´as, J. Trueba, and A. Ra˜nada. Topological [11] D. Hestenes. Spacetime physics with geometric algebra. electromagnetism: Knots and quantization rules. In 23

V. Barsan and R. Lungu, editors, Trends in Electromag- [45] A. Burinskii. Illusion about weakness of gravity hides netism – From Fundamentals to Applications, pages 71 – new way to unify gravity with particle physics. Journal 88. InTech, Internet: Open Access, 2012. of Physics: Conf. Series 942 (2017) 012006, Conf. Series [21] W. Irvine and D. Bouwmeester. Linked and knotted 942 (2017) 012006:1–8, 2017. beams of light. Nature Physics, 4:716–720, 2008. [46] S. Blinder. General relativistic models for the electron. [22] H. Kedia, I. Bialynicki-Birula, D. Peralta-Salas, and Rep. Math. Phys., 47:279–287, 2001. W. Irvine. Tying knots in light fields. Phys. Rev. Lett., [47] R. Gurtler and D. Hestenes. Consistency in the formu- 111:150404, 2013. lation of the Dirac, Pauli and Schr¨odingertheories. J. [23] S. van Enk. Covariant description of electric and mag- Math. Phys., 16:573–584, 1975. netic field lines of null fields: application to Hopf-Ra˜nada [48] D. Hestenes. Spin and uncertainty in the interpretation solutions. J. Phys. A: Math. Theor., 46:175204, 2013. of quantum mechanics. Am. J. Phys., 47:399–415, 1979. [24] A. Kholodenko. Optical knots and contact geometry I. [49] M. MacGregor. The Power of Alpha. World Scientific, From Arnold inequality to Ra˜nadasdyons. Annals of London, 1992. Mathematics and Physics, 6(2):163–198, 2016. [50] L. de Broglie. A New Conception of Light. Hermann and [25] A. Kholodenko. Optical knots and contact geometry II. Co., Paris, 1934. Translated by D. H. Delphenich. from Ra˜nadadyons to transverse and cosmetic knots. [51] A. Di Tommaso and G. Vassallo. Structure, ultra-dense Annals of Physics, 371:77–124, 2016. hydrogen and low energy nuclear reactions. J. Condensed [26] A. Kholodenko. Applications of Contact Geometry and Matter Nucl. Sci., 29:525–547, 1999. Topology in Physics. World Scientific, Singapore, 2013. [52] C. Doran, A. Lasenby, S. Gull, S. Somaroo, and A. Challi- [27] D. Hestenes. New Foundations for Classical Mechanics. nor. Spacetime Algebra and Electron Physics. Adv. Imag. Kluwer: Springer, New York, 1986. and Electron Phys., 95:271, 1996. [28] M. Array´aand J. Trueba. On the fibration defined by [53] D. Hestenes. New Foundations for Classical Mechanics. the field lines of a knotted class of electromagnetic fields Kluwer: Springer, New York, 1999. at a particular time. Symmetry, 9:9100218, 2017. [54] A. Lewis, A. Lasenby, and C. Doran. Electron scattering [29] D. Hestenes and G. Sobczyk. Clifford Algebra to Geomet- without spin sums. Int. J. Theo. Phys., 40:363 – 375, ric Calculus. Kluwer: Springer, New York, 1984. Second 2001. Edition: 1986. [55] D. Hestenes. Gauge gravity and electroweak theory. In [30] D. Hestenes. Proper particle mechanics. J. Math. Phys., R. Jantzen, H.Kleinert, and R. Ruffini, editors, Proceed- 15:1768–1777, 1974. ings of the Eleventh Marcel Grossmann Meeting. World [31] D. Hestenes. Proper dynamics of a rigid point particle. Scientific, 2008. arXiv:0807.0060v1 [gr-qc] 1 Jul 2008. J. Math. Phys., 15:1778–1786, 1974. [56] H. Jehle. Relation of flux quantization to charge quanti- [32] D. Hestenes. Zitterbewegung in quantum mechanics. zation and the electromagnetic coupling constant. Phys. Found. Phys., 40:1–54, 2010. The treatment of Quantiza- Rev. D, 3:306–345, 1971. tion in Section X is defective, but the error is instructive. [57] H. Jehle. Flux quantization and fractional charges of [33] E. J. Post. Copenhagen’s single system premise prevents quarks. Phys. Rev. D, 11:2147–2178, 1971. a unified view of integer and fractional quantum Hall [58] R. J. Finkelstein. The elementary particles as quantum effect. Annals of Physics (Leipzig), 8:405–423, 1999. knots in electroweak theory. International Journal of [34] K. Muralidhar. Complex vector formalism of harmonic Modern Physics A, 22:4467–4480, 2007. oscillator in geometric algebra: Particle mass, spin and [59] R. J. Finkelstein. The slq(2) extension of the stan- dynamics in complex vector space. Foundations of dard model. International Journal of Modern Physics Physics, 44:266–295, 2014. A, 30(16):1–50, 2015. [35] B. Lautrup and H. Zinkernagel. g-2 and the trust [60] C. Mead. Collective Electrodynamics. MIT Press, Cam- in experimental results. Stud. Hist. Phil. Nod. Phys., bridge, 2000. 30(1):85–110, 1999. [61] E. J. Post. Quantum Reprogramming.Ensembles and Sin- [36] L. de Broglie. Interpretation of quantum mechanics by gle Systems: A Two-Tier Approach to Quantum Mechan- the double solution theory. Annales de la Fondation Louis ics. Kluwer, Dordrecht, 1995. de Broglie, 12(4):1–23, 1987. [62] M. J. Mobley. Momentum exchange theory of pho- [37] W. Furry and J. R. Oppenheimer. On the theory of the ton diffraction. Optical Engineering, 57(1):015105–1–19, electron and positron. Physical Review, 45:245–262, 1934. 2018. [38] F. London. Superfluids, volume I. Wiley, New York, [63] W. Duane. The transfer in quanta of radiation momen- 1950. especially p.152. tum to matter. Proc. Nat. Acad. Sci., 9:158–164, 1923. [39] H. Kleinert. Multivalued Fields in Condensed Matter, [64] K. M. Van Vliet. Linear momentum quantization in pe- Electromagnetism and Gravitation. World Scientific, Sin- riodic structures. Physica, 35:97–196, 1963. gapore, 2008. [65] C. M. Van Vliet. Linear momentum quantization in pe- [40] V. Weisskopf. On the self-energy of the electromagnetic riodic structures II. Physica A, 389:1585–1593, 2010. field of the electron. Physical Review, 56:72–85, 1939. [66] H. Batelaan and A. Tonomura. The Aharonov-Bohm [41] J. C. Slater. Spinning electrons and the structure of spec- effects: Variations on a subtle theme. Physics Today, tra. Nature, 117:587, 1926. 62:38–43, 2009. [42] T. Welton. Some observable effects of the quantum- [67] P. A. M. Dirac. Is there an aether? Nature, 168:906–907, mechanical fluctuations of the electromagnetic field. 1951. Physical Review, 74:1157–1167, 1948. [68] J. Bush. Pilot-wave hydrodynamics. Annual Review of [43] D. Hestenes. Quantum mechanics from self-interaction. Fluid Mechanics, 47:269–292, 2015. Found. Phys., 15:63–87, 1985. [69] A. Einstein, L. Infeld, and B. Hoffmann. The gravita- [44] C. Misner K. Thorne & J. Wheeler. Gravitation. W. H. tional equations and the problem of motion. Ann. Math., Freeman, San Francisco, 1970. 39:65–100, 1939.