This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

David Hestenes

Arizona State University Abstract: This book provides a synopsis of spacetime calculus with applications to classical electrodynamics, quantum theory and gravitation. The calculus is a coordinate-free mathematical language enabling a unified treatment of all these topics and bringing new insights and methods to each of them. CONTENTS PART 1: Mathematical Fundamentals 1. 2. 3. 4. 5. 6. 7. 8. 9. Spacetime Algebra Vector Derivatives and Differentials Linear Algebra Spacetime Splits Rigid Bodies and Charged Particles Electromagnetic Fields Transformations on Spacetime Directed Integrals and the Fundamental Theorem Integral Equations and Conservation Laws PART II: Quantum Theory 10. 11. 12. 13. 14. The Real Dirac Equation Observables and Conservation Laws Electron Trajectories The Zitterbewegung Interpretation Electroweak Interactions Part III. Induced Geometry on Flat Spacetime 15. Gauge Tensor and Gauge Invariance 16. Covariant Derivatives and Curvature 17. Universal Laws for Spacetime Physics REFERENCES Appendix A. Tensors and their Classification Appendix B: Transformations and Invariants Appendix C: Lagrangian Formulation

1

**PART I: Mathematical Fundamentals
**

1. SPACETIME ALGEBRA.

vVerepresent Minkowski spacetime as a real 4-dimensional vector space M4. The two properties of scalar multiplication and vector addition in j\ll4 provide only a partial specification of spacetime geometry. To complete the specification we introduce an associative geometric product among vectors a, b, c, ... with the property that the square of any vector is a (real) scalar. Thus for any vector a we can write (1.1) where E is the signature of a. and 10.1 is a (real) positive scalar. As usual, we say that a is timelike, lightlike or spacelike if its signature is positive (E = 1), null (E = 0), or negative (E = -1). To specify the signature of j\;14 as a whole, we adopt the axioms: (a) _\114 contains at least one timelike vector; and (b) Every 2-plane in j\ll4 contains at least one spacelike vector. The vector space j\ll4 is not closed under the geometric product just defined. Rather, by multiplication and (distributive) addition it generates a real, associative (but noncommutative), geometric algebra of dimension 24 = 16, called the spacetime algebra (STA). The name is appropriate because all the elements and operations of the algebra have a geometric interpretation, and it suffices for the representation of any geometric structure on spacetime. From the geometric product ab of two vectors it is convenient to define two other products. The inner product a . b is defined by (1.2) o.·b= ~(ab+ba) =b·a, and the outer product a /\ b is defined by a. /\ b

= ~(o.b - bo.) = -b /\ a.

(1.3)

The three products are therefore related by the important identity ab

= a. • b + 0 /\ b ,

(1.4)

which can be seen as a decomposition of the product al: into symmetric and antisymmetric parts. From (1.1) it follows that the inner product o· b is scalar-valued. The outer product a r.b is neither scalar nor vector but a new entity called a bi vec tor, which can be interpreted geometrically as an oriented plane segment, just as a vector can be interpreted as an oriented line segment. Inner and outer products can be generalized. The outer product and the notion of k-vector can be defined iteratively as follows: Scalars are regarded as O-vectors, vectors as I-vectors and bivectors as 2-vectors. For a given k-vector K the positive integer k is called the grade (or step) of K. The outer product of a vector a with K is a (k + I)-vector defined in terms of the geometric product by (1.5) The corresponding inner product is defined by (1.6) and it can be proved that the result is a (k - I)-vector. Adding (1.5) and (1.6) we obtain aK

= a . K + a /\ K ,

(1.7)

which obviously generalizes (1.4). The important thing about (1.7) is that it decomposes a.K into (k - I)-vector and (k + I)-vector parts. 2

Manipulations and inferences involving inner and outer products are facilitated by a host of theorems and identities given in [1], of which the most important are recorded here. The outer product is associative, and (1.8) if and only if the vectors al, a2, ... ,ak are linearly dependent. Since Jvt4 has dimension 4, (1.8) is an identity in STA for k > 4, so the generation of new elements by multiplication with vectors terminates at k = 4. A nonvanishing k-vector can be interpreted as a directed volume element for A,;/4 spanned by the vectors al, a2, ... , ak. In STA 4-vectors are called pseudoecelere. and for any four such vectors we can write (1.9) al /\ a2 /\ ... /\ ak = Ai , where i is the unit pseudoscalar and A is a scalar which vanishes if the vectors are linearly dependent. The unit pseudosceler is so important that the special symbol i is reserved for it. It has the properties (1.10) and for every vector a in ,Vl4 ia

= -ai.

(1.11 )

Of course, i can be interpreted geometrically as the (unique) unit oriented volume element for spacetime. A convention for specifying its orientation is given below. Multiplicative properties of the unit pseudoscalar characterize the geometric concept of duelit». The dual of a k-vector K in STA is the (4 - k)-vector defined (up to a sign) by i.K or Ki, Trivially, every pseudoscalar is the dual of a scalar. Every 3-vector is the dual of a vector; for this reason 3-vectors are often called pseudcvectors. The inner and outer products are dual to one another. This is easily proved by using (1.7) to expand the associative identity (aK)i = a(Ki) in two ways: (a· K

+ a /\ K)i = a·

(Ki)

+ a /\ (Ki)

.

Each side of this identity has parts of grade (4 - k ± 1) and which can be separately equated, because they arc linearly independent. Thus, one obtains the duality identities

O;>

(K)i = a /\ (Ki) ,

(Ll2a) (Ll2b)

**a /\ (K)i = a- (Ki) , Note that (1.12b) can be solved for a·K= [a/\ (Ki)]C1
**

,

(1.13)

which could be used to define the inner product from the outer product and duality. Unlike the outer product, the inner product is not associative, Instead, it obeys various identities, of which the following involving vectors, k-vector K and s-vector B are most important:

(b /\ a) . K = b· (a. K) = (K. b) • a = K· (b /\ a)

a· (K /\ B)

for grade k ?: 2,

(1.14)

= (a. K) /\ B + (_I)k K /\ (a. B).

(1.15)

The latter implies the following identity involving vectors alone: a» (al /\ a2 /\ ... /\ ak)

= (a. al)a2 /\ ... /\ ak - (a· (2)al /\ a3 ... /\ ak+

'" + (_I)k-l(a.

3

ak)al /\ a2'"

/\ ak-l'

(1.16)

21) = a + bi .bi + .(a· c)b = a· bc . (1. the reverse M (1.23b) ii is given by (1. defined for arbitrary multivectors M and N by (MN)~ =NM. (1.f1vl+} of all even multivectors forms an important subalgebra of STA called the cvcn eubelgcbxe: Spinors can bc represented as clements of thc cvcn subalgcbra. while the odd part becomes M_ (1. a and b are vectors and F is a bivector. (1. is the vector part. This convention allows us to drop parentheses on the right side of (1. The simplest case is used so often that it is worth writing down: a· (b 1\ c) = (a· b)c . the decomposition (1. This is a decomposition of 1M into its "k-vector parts" (Ai)k. Any multivector M can be uniquely decomposed into parts of even and odd grade as lvh = ~(M =f iMi). (M)2 = F is the bivector part.19) is the scalar part. (M)3 = bi is the pseudovector part and (M)4 = pi is the pseudoscalar part. (1. Like the decomposition of a complex number into real and imaginary parts. multivectors of mixed grade often have geometric significance that transcends that of their graded parts. Duality has been used to simplify the form of the trivector part in (1.This is called the Laplace expansion.18) where a and b are scalars. Note that bivectors and trivectors change sign under reversion while scalars and pseudoscalars do not.18).20) In terms of the expanded form (1.18) by expressing it as the dual of a vector.. For M in the expanded form (1. and both of those before geometric products. the cvcn part can be written lvh=a+F+pi. On the other hand.16) but not on the left.23a) with a=a for any vector a. (Mh = a.17) Parentheses have been eliminated here by invoking a precedence convention. where (M)o = (M) =a (1. that in ambiguous algebraic expressions inner products are to be formed before outer products.3i . The entire spacetime algebra is generated by taking linear combinations of k-vectors obtained by outer multiplication of vectors in }vI4. Computations are facilitated by the operation of reversion.a· cb. 4 .18) is significant because the parts with different grade are linearly independent of one another and have distinct geometric interpretations. because it generalizes and implies the familiar expansion for determinants. (1.: (Mh k=O 4 . A generic element of the STA is called a multivector.24) = a + a .F .22) The set {.18). Any multivector M can be written in the expanded form M = a + a + F + bi + pi = 2.

29) a complete basis for STA. "Rectangular coordinates" {XV} given by and The IlL generate by multiplication independent elements (1. For manipulating coordinates it is convenient to introduce the reciprocal frame {II"} defined by the equations or (summation convention assumed). 1.STA makes it possible to formulate and solve fundamental equations without using coordinates. many other products can be defined in terms of the geometric product. (1.z. if k is odd.26) VVe shall refer to {II"} &'3 a standard frame.33) = (A x B) x C + B x (A x C) . Let {II". such an expansion is seldom needed. as shown below and in later sections.25) This may be recognized as the defining condition for the "Dirac algebra. without necessarily associating it with the reference frame of a physical observer. For example.BA) = -B x A. Any multivector can be expressed as a linear combination of these elements.31) However.2." which is a matrix representation of STA over a complex number field instead of the reals. The orientation of the unit pseudosceler i relative to the frame {IlL} is set by the equation i = 1011/213 = 10 /\ 11 /\ 12 /\ 13 . we can write (1. The commutator product A x B is defined for any A and B by AxB =~ (AB .IF!""'" II' /\ '"V (1.30) -"2 /V. thus. fl.28) 16 linearly (1. for a vector a. (1. it can be shown that these alternatives are in fact compatible. a bivector F has the expansion F . because it has the "distributive property" A x (BC) = (A x B)C This implies the Jacobi Idcntity A x (B x C) + B(A x C) . with its "scalar components" given by (1. 3} be a right-handed orthonormal frame of vectors in JV!4with 10 in the forward light cone. II'. Although the present interpretation of the {IlL} as an orthonormal frame of vectors is quite different from their usual interpretation as matrix components of a single vector. (1. However.27) of a spacetime point x are then (1.2). 11'. 5 (1.' IlL /\ IV.i.35a) . In accordance with (1.34) which is a derivation on the commutator product. it also facilitates the manipulation of coordinates. consisting of the 24 1. The relation of the commutator product to the inner and outer products is grade dependent.32) Mathematicians classify this product as a "derivation" with respect to the geometric product. = 0. (1. Besides the inner and outer products defined above.

satisfying If p pei6 .39) is the unique decomposition of F into a sum of mutually commuting timelike and spacelike parts. Equation (1. Therefore. If 4) is an even multivector. since ('tI'. the Lie algebra of the Lorentz group.35b) The commutator product is especially useful in computations with bivectors.f = k 1\ e = ke .36) In particular..38c) is its application to a bivector F. An important special case of the decomposition (l.39) with . When F2 = 0.39) where = f 2.40a) . (1.4) for the product of vectors. 6 (1.37) This should be compared with the cor::esponding expansion (1.. F)2 and +F + 1F 1\ 1\ 1\ F)~ F 12 1\ r 1. (l.which is.f = Fe -icp = F(F .. It therefore forms a Lie algebra .40) ." since (if)2 = 1 f 12. (l.38b) RR =RR= Then 1/J can be put in the canonical form 1.:: 2.) ~ = 4!lt'. With any bivector A this product is grade preserving: A x (Mh = (A x M)k.24). (1. with the results 1 1 P (F.39) can be solved directly for sp and f in terms of F. (l. The geometric product of bivector A with NI has the expanded form AM = A· M + A xM + A 1\ M for grade M . in fact. F can still be written in the form (1. (l. this implies that the space of bivectors is closed under the commutator product. F [(F . for any bivector F which is not null (F2 oF 0) we have the invariant canonical form F = f eicp = f (cos cp +i sin cp) . then 1!4) is also even but its bivector part must vanish according to (l. so f is a "timelike bivector. Thus.38a) oF 0 we can derive from 4) an even multivector R (l. F ." The dual if is a "spacclike bivector.38c) We shall see that this invariant decomposition has a fundamental physical significance in the Dirac Theory. F)2 + 1F r (l. especially when applied to an electromagnetic field.39) has important physical and practical implications.F -----'--------'--'1 F) ~ F 12 [(F.if k is even. = where p and j3 are scalars. for which it is convenient to write f = p~ R.. Thus the right side of (1. (l. we can write 4).40b) In later sections it will be seen that the decomposition (1.

" originally discovered by Dirac by seeking a "square root" of the d'Alembertian (2. F is the divergence of F and \7 /\ F is the curl. where the are vectors rather than matrices.3) The matrix representation of (2. In this case.2r)A.F+\7/\F.\7x=rA \7(x /\ A) = A /\ \7x = (4 .28).A)=A.7) As in conventional scalar calculus. an operator \7 = interpreted as the derivative with respect to a spacetime point x can be defined by ax (2. suitable definitions for derivatives and integrals are required.4) \7F=. it provides an appropriate definition for the derivative with respect to any spacetime vector variable.where k is a null vector and e is spacelike.!LA"L = (-1)'"(4 . We can accordingly separate (2. indeed.6) \7·F=J. \7/\F=O.3) in order to find a first order "relativistically invariant" wave equation for the electron.2) The square of \7 is the d' Alembertian (2. It is equally apt for electromagnetic field equations: in STA an electromagnetic field is represented by a bivector-valued function F = F(:r) on spacetime. (2.5) where \7 .a) \71x =a·\7x=a. so that we can write \7F=\7. \7x=\7·x=4 7 .1) can be recognized as the "Dirac operator.r)A \7(Ax) = . . In terms of the coordinates (1. Vector Derivatives and Differentials To extend spacetime algebra into a complete spacetime calculus.I' Since \7 is a vector operator the identity (1. and the exponential factor can always be absorbed in the definition of [. (2. (2.4) into vector and trivector parts: (2. The most useful derivatives are listed here: Table of vector derivatives: \7(x. 2. Contrary to the impression given by conventional accounts of relativistic quantum theory. 7) applies.J. = \7x2 = 2x. it is clear that \7 is a vector operator.8) \7(x.1) where (2. The field produced by a source with spacetime current density J = J(x) is determined by !vlaxwell's Equation (2. the decomposition is not unique. the derivatives of elementary functions are often needed for computations. 12 \7/\x=O. the operator \7 is not specially adapted to spin-j wave equations. In STA however.

and we use the special symbol \7 == a".8): thus. we indicate the variable explicitly by writing Mathematically.1) for defining the vector derivative.9) where F = F(x) is any multivector valued function. Alternatively. The directional derivative a· \7. 4-k In the table.\7F. The second differential has the symmetry property (2.9) produces from F a tensor field called the differential of F. this can be expressed as an operator identity: (a 1\ b) . (2. 8 . (2. the underbar notation serves to emphasize that F(a) is a linear function of a. and obvious singularities at x = 0 are excluded.10) This relation can serve as an alternative to the partial derivative in (2. As already indicated.14) Differentation by \71\\7=0. that is. \7 = ax. Equation (2.10) and the preceding equations above define the derivative with respect to any spacetime vector a. there is no difference between \7 and However.F(b· \7a) = b· VF(a) . denoted variously by (2. Of course. (2. The directional derivative (2. The argument a may be a free variable or a vector field a = a( x). (\7 1\ \7) where the bracket denotes the commutator. form = [a· \7. we reserve the symbol "z" for a vector representing a position in spacetime. for the derivative with respect to this "spacetime point. including the rules for differentiating sums and products. aa and a b puts this identity in the (2.\7 (____::_) I::c Ik ~. independent of x). the "derivative in the direction of vector a" can be obtained from \7 by applying the inner product. while A is a free r-vcctor. the directional derivative can be defined directly by a· \7 F = a· axF(x) = -d d . though it is not a linear transformation unless it is vector valued.c. but the explicit display of the vector variable is advantageous in concept and calculation. Then the general vector derivative can be obtained from the directional derivative by using (2. a is a "free vector" variable (i.15) These last three equations are equivalent formulations of the integrability condition for vector derivatives. The second differential Fb(a) = Fab is defined by Fb(a) == b· \7 F(a) . Of course. the accents can be dropped if a is a free variable.T F(x + aT) I T=O = lim T->O F(x+aT)-F(x) T . there is often an important difference in physical or geometrical interpretation associated with these operators. aa aa' aa' As explained in the next section.14) and (1. (2." When differentiating with respect to any other vector variable a.11) F(a)=Fa==a.17). the directional derivative has the same limit properties as the partial derivative.12) where the accents over V and P serve to indicate that only F is differentiated and a is not.13) Using (l. b· \7] = 0.

This implies that f alters the pseudoscalar i only by a scalar multiple.16) The left side of this identity may be identified as a Lie bracket.5) for any multivector Indeed M.b· \7a = \7. linear algebra is a large subject.f on Minkowski space has a unique extension to a linear function on the whole STA. for a linear operator f acting on a vector a.t because it preserves outer products. either for emphasis or to remove ambiguity.. Other identities will be derived as needed. the parenthesis around the argument of f can be included or omitted. or 9 det f = -if(i). We need a notation which clearly distinguishes linear operators and their products from vectors and their products.3) a1 /\ a2 /\ . It follows that.2) (3... called the outer-morphism of .t for the outermorphism and the operator that "induces" it. (3. As usual in linear algebra.Since the derivative \7 has the algebraic properties of a vector.).b\7· a.6) f(i) = (det f)i .. Linear Algebra The computational and representational power of linear algebra is greatly enhanced by integrating it with geometric algebra. Every linear transformation . distinguishing them when necessary by their arguments. a large assortment of "differential identities" can be derived by replacing some vector by \7 in any algebraic identity. (3. /\ (fa. For example. 3.. The outermorphism is defined by the property f(A/\B) for arbitrary multivectors A. of an r-vector A into vectors. for any scalar a. Since the outermorphism preserves the outer product. It is convenient to use the same notation . a more general concept of the Lie bracket is introduced later on. for any factoring A f(A) = = (fad /\ (fa2) /\ . so we restrict our attention to the essentials needed for gravitation theory. (2. Integration of the two algebras can be achieved with the few basic concepts. The only caveat is to take proper account of the product rule for differentiation. that is (3. geometric calculus makes it possible to perform coordinate-free computations in linear algebra without resorting to matrices. /\ a. B. Thus. and = (fA) /\ (fB) fa =a (3.4) This relation can be used to compute the outermorphism directly from the inducing linear operator. which map spacetime vectors into vectors. Though the approach works for vector spaces of any dimension. notations and theorems reviewed below. However. In fact. To this end. we will be concerned only with linear transformations of Minkowski space. we distinguish symbols representing a linear transformation (or operator) by affixing them with an under bar (or overbar). we write (3. (a /\ b) + a\7· b . the product rule gives whence the algebraic identity (1.1) fa=f(a). it is grade preserving.17) yields a· \7b .

B] = A· . since A . it can be written in the form .f(B). This is called the Lorentz group.f has an a(~joint transformation . Of course. (3. Every linear transformation . and its members are called Lorentz transformations.10) where M and N are arbitrary multivectors and the bracket as usual indicates "scalar part. The defining property for a Lorentz transformation L is (La) .8) from which many other properties of the determinant follow.1 which can be extended to an outer morphism denoted by the same symbol. expressed by (3.7) applies also to their outermorphisms.15) This relation shows explicitly the double duality involved in computing the inverse. Note that the outermorphism makes it possible to define (and evaluate) the determinant without introducing a basis or matrices. it might well be taken as the preferred definition of .f(B) (3.f(a) = [ha· f(b). (3." For vectors a. . However.13) for (grade A) :::.6) that det (flf) = (det fl)(det f). In other words. this immediately gives the general formula for the inverse outermorphism: (3.14) For B = i. such as (3. Although all linear transformations preserve the outer product (by definition of class preserves the inner product.B] = f-l(A) .f N) = (N fM) .which defines the determinant of f.f(a) =a· feb). Differentiating with respect to b we obtain. the inner product is not generally preserved by outerrnorphisms.f [f(A) . (3.8). i' .f from f. b this can be written (3. .f. It follows immediately from (3. (Lb) = a10 (LLb) = a- b. the outermorphism of a product equals the product of outerrnorphisms. i = Ai. it obeys the fundamental transformation law .9) whenever exists.12) This is the most useful formula for obtaining . this law also holds with an interchange of overbar and underbar. The "product" of two linear transformations. (3. (3. If f is invertible. with the help of (2. Indeed.f[A. (grade B). The adjoint outermorphisrn can be defined in terms of f by (M.ll ) b .16) . Unlike the outer product.

for any timelike vector v. (3.24) The rotors form a multiplicative group called the Rotor group. where the composition of two Lorentz transformations LlL2 corresponds to the geometric product ±L1L2.23) where the even multivector R is called a rotor and is normalized by the condition RR = 1.26) Let nl.L(v) > o.This is equivalent to the operator condition (3. which is a double-valued representation of the Lorentz rotation group (also called the restricted Lorentz group).18) where the multi vector L is either even with E = 1 or odd with E -1. (3. A reflection (3.14) when (3.18) it follows immediately that.19) Lorentz transformations therefore preserve the geometric product. according to n(a) = -nan-I.25) This is a reflection with respect to a hyperplane with normal y(a) with respect to a timclikc vector v = vectors which compose the trivector v-I 71. be spacelike (3. The most elementary kind of Lorentz transformation is a reflection n by a (non-null) vector 71. the Lorentz group has a double-valued representation as a multiplicative group of multi vectors. (3. n2. (3. and improper if E = o. (LB) . This defines a doublevalued homomorphism between Lorentz transformations {L} and multivectors {±L}. From (3. (3. B) = (LA) . = 1. From (3. Thus. It is said to be A Lorentz transformation L is said to be proper if orthochronous if. (3. (3.17) is satisfied.16) generalizes to L(A. For a Lorentz rotation 3 the canonical form can be written 3(a) = RaR. This implies that (3. This multivector representation of the Lorentz group greatly facilitates the analysis and application of Lorentz transformations in STA.21 ) -1. L(AB) = (LA) (LB) .22) A proper. orthochronous Lorentz transformation is called a Lorentz rotation (or a restricted Lorentz transformation).17) STA makes it possible to express any L in the simple canonical form (3. 713 = -vav is called a time reflection. for arbitrary multivectors A and B. where E E = detL = ±1.27) 11 .18) it follows easily that L(i) = Ei.20) in agreement with (3.

for a spacelike rotation 9(a) the rotor Q can be factored into a product = QaQ.37a) . first Y and spacelike rotation Q with respect to a given tirnelike vector compute the-vector V = Bvo = RVoR .28) is the spacetime inversion :!lst(a) = :!lsv(a) = -iai-l = -a. of a spacclikc vector with a timelike vector is not a rotor. The composite of the time reflection (3.18). Any Lorentz rotation B can be decomposed into the product B=Y9 Y and of a boost (3.31) is not unique.32) (3. Note that the product.26) and (3. even though it can be expressed as the product of two pairs of vectors. This can be solved for V = 1 (vvo)"2 = 12 vw = wVo . Although :!ls is determined by valone on the right side of (3. The boost is a rotation in the timelike plane containing VI and V2.28).26) with the space inversion (3.30) (3.31). The factorization (3.28). Similarly. (3. because the corresponding Lorentz transformation is not orthochronous.33) of two unit spacelike vectors in the spacelike plane of the rotation. for a given V any timelike vector in the plane can be chosen as VI. for it does not satisfy the rotor condition (3. say n2Vl.35) the timelike vectors V and Vo determine the timelike plane for the boost. (3. (3.A space inversion :!ls can then be defined as the composite of reflections with respect to these three vectors.24). the pseudoscalari is not a rotor.34) Vo = VOl. Indeed. which can therefore be defined by ~ 2 (3. namely timclikc rotations (or boosts) and specclil«: rotations. To compute 9 from B. Two basic types of Lorentz rotation can be obtained from the product of two reflections.28) Note the difference in sign between the right sides of (3. the multivector representation of :!ls must be the trivector iv in order to agree with (3. For a boost Y(a) = VaV. Note that spacetime inversion is proper but not orthochronous.29) which is represented by the pseudoscalari. Likewise.36) V = }(vo = V Vo V = V Vo . the rotor V can be factored into a product (3. (3. so it can be written (3. and V2 then computed from (3.31) of two unit timelike vectors 1h andv2.

and it can be expressed in the factored form (1. Any rotor R can be expressed in the exponential form rotations by vectors terms of initial and are often given.38) 9va = Qvo Q = vo . The set of such rotations is the 3-parameter spatial rotation group (of '(0)' More generally. Since the timelike unit bivector I commutes with its dual ii. More specifically. The positive sign can always be selected when p2 i= 0.43) (3. the rotor R can be decomposed into a product of commuting timelike and spacelike rotors. Thus R=VQ=QV. P can be written in the canonical form P=(a+i. '(0) 1 (3. Equations (3.42) where and v = c~('f = cosh ~a + I sinh ~a.3 . When p2 = 0. (3.37b) ] '2 "bisects the angle" between v and Vo.40) can be reduced to the form R= e~af = 1 + ~al .46) Jj(n) = NnN = n 13 .38).39) This makes (3. Another useful parametrization (Appendix B of [2]).45) represents a lightlike Lorentz rotation. (3. because the velocity vectors are of direct relevance. The parameter .41) a and .32) show how to parametrize boosts and spacelike in the plane of rotation.37a.where the unit vector to == v +vo v +vo Iv +vo I [ 2(1 + v .3 + i] sin~. according to (1.3)1 where (3. Q= e~ij3f (3.36) consistent with (3. The lightlike rotor in (3. (3.b) parametrizes a boost in final velocity vectors.3 is the usual angular meaaure of a spatial rotation. There is no choice of null P which can eliminate the minus sign. the subgroup of Lorentz transformations Jj satisfying (3. The parameter Ct is commonly called the rapidity of the boost. This is especially useful.45) where I is a null bivector. so that the spacclikc rotation satisfies (3.3 being scalar parameters.39). which is a spacelike bivector. The spacelike rotations that preserve a timelike vector Va are commonly called spatial rotations without mentioning the proviso (3.34). in a physical problem. equation (3.35) by virtue of (3.40). The two signs are inequivalent cases.44) = cos~. for any givcn vector n.40) where P is a bivector parametrizing the rotation. The rotor Q can then be computed from Q=VR. (3. or is in terms of angle ±R = C2 1.F =~ L k=O 1 (1 )k k! ':iP .31) and (3. (3. and.

48) On Euclidean spaces every linear transformation has a polar decomposition. where 3 is a Lorentz rotation and by the condition fl' = 3fl 3-1. observations and measurements are usually expressed in terms of variables tied to a particular inertial system. to each event x the single equation (4. so we need to know how to reformulate invariant equations in terms of those variables. In STA a given inertial system is completely characterized by a single future-pointing. If the latter has a well defined square root fl = Uf)! = S. This defines the lo-split of spacetime. then f admits to the polar decomposition (3. However. but on Minkowski space there are symmetric transformations (with null eigenvectors) which do not possess square roots and so do not admit a polar decomposition.1c) = x /\ 10.is called the little group of n.1a) where t= and x X'IO (4. The observer 10 is represented algebraically in STA in the same way as any other physical system. and the spacetime split amounts to no more than comparing the motion of a given system (the observer) to other physical systems.1a) assigns a unique time and position in the 'Yo-system. Refer to the inertial system characterized by the vector 10 as the lo-system.47) . 4.f = 3fl = fl' 3 . so it is convenient to use 10 as a name for tile observer. In "relativistic units" where the speed of light c = 1. (4. With STA we can describe physical processes by equations which are invariant in the sense that they are not referred to any inertial system.1b) is the equation for a one parameter family of spacelike hyperplanes with normal 10. A complete classification of symmetric transformations is given in the next section. Spacetime Splits. x + 10 /\ x = t . Thus. The vector 10 is tangent to the world line of an observer at rest in the lo-system. t is the time parameter for the lo-system.40) composed with timelike and spacelike rotors. (3. (4. Symmetric transformations are.1b) (4. so. The little group of a lightlike vector can be parametrized as a lightlike rotor (3. STA provides a very simple way to do that called a space-time split.1f is a symmetric transformation. indeed. defined S = fl. timelike unit vector.x .:CIO = t+x.1a) is lOx = 10 . The above classification of Lorentz transformations can be extended to more general linear transformations. of course. For any linear transformation f the composite . Equation (4. Note that the reverse of (4. (4. For each spacetime point (or event) x the mapping is specified by .2) 14 . An inertial observer 10 determines a unique mapping of spacetime into the even subalgebra of STA. since 'Y6= 1.1b) assigns a unique time t to every event x.1c) assigns to each event x a unique position vector x in the 'Yosystem. Equation (4.

3.4a) These generate a basis for the relative bivectors: (4. we refer to the elements of p3 as vectors. Note that p3 consists of all bivectors in STA with 10 as a common factor. we use the term Lorentz invariant to mean independent of a chosen spacetime split. Now let us rapidly survey the space-time splits of some important physical quantities. those in JV[4 and those in p3. whence vlO (4. of course!).3) J. The set of all position vectors (4. thus we have proved that the expression t2 . (4. In contrast to (4. Let x = x( T) be the history of a particle with proper time T and proper velocity v = dx/ dt. which are independent of even an indirect reference to an inertial system. In agreement with common parlance. The right sides of (4. equation (4.lb = t' + x' . 10}.6b) is the "time dilation" factor.4d) = Vo (1 + v) ( . relative vectors will be designated in boldface.4b) show how the bivectors for spacetime are split into vectors and bivectors for p3.4b) where the allowed values of the indices {i. The space-time split ofv is obtained by differentiating (4. Corresponding to a standard basis bll} for JV[4.24) in the entire STA by (4.2). this is related to reversion (1.1a).The form and value of this equation is independent of the chosen observer. Also.4c) The geometrical operation of reversion in the algebra of p3 can be defined by and For an arbitrary multivector NI. where lTk = Ik !\ 10 = IklO . we have two kinds of vectors. and v=-=--=--- dx dt £IT dx dt dr 15 V !\ 10 V'IO (4.4a) and (4. By the geometric product and sum the vectors in p3 generate the entire even subalgebra of STA as the geometric algebra of p3. Mostly we shall work with manifestly Lorentz invariant equations. (4.5) The explicit appearance of the timclike vector 10 here shows the dependence of NIt on a particular spacetime split. which we designate by p3 = p3Cfo) = {x = x r./V(4 as proper vectors and to elements of p3 as relative vectors (relative to 10. Remarkably. we may refer to elements of .6c) .2.3} for t». we have a standard basis {lTk.x2 is Lorentz invariant without even mentioning a Lorentz transformation. the right-handed pseudoscalar for p3 is identical to that for M4. indeed.1c) is the 3-dimensional position space of the observer 10. [. k} are cyclic permutations of 1. for a different observer Ib we get the split (4. Thus.k = 1. This is made obvious by constructing a basis. To distinguish between them. (4. thus.1a) is not Lorentz invariant. Henceforth.6a) where Vo = v • 10 = - dt £IT = 1- V 2)_1 2 (4.2.

Let p be the proper momentum (i. An electromagnetic field is a bivcctor-valucd function F = F(x) on spacetime.Ft). as expressed by f /\ '10 = 0 so f = (f .9a) to the invariant decomposition F = fei8 given by (1. lOa) 2z 16 . '1oho = ~(F + Ft). and ExB 1 1 = ---.lOb) (4. Indeed.5). iB = (F /\ '1oho = ~(F .9a) shows that the magnetic field is actually a bivector quantityiB. Thus. then (4.7c) (4. Now consider the relation of the split (4.iB.6b) was obtained from 1 = v2 = (V'Io)(rov) = vo(l + v)vo(l . The last equality in (4.v) = v6(1 .is the relative velocity in the '1o-system. Consequently.B=i(ExB). b) gives us E = f cos tp and B = f sin ip.9a) and. thus F =E+iB. '10. (4. An observer '10 splits it into an electric (relative vector) part E and. we can interpret F = fei'!' physically as a Lorentz invariant decomposition of the field F into parallel (commuting) electric and magnetic parts.e.9b) (4. p = P /\ '10· + p. Ft = E .lOc) (4. it characterizes an intrinsic structural property of the field.9a. and its conventional representation as a vector B is a historical accident in which the duality is hidden in the notion of "axial vector" [3.8) Of course where m is the proper mass of the particle. Equation (4.4]. in accordance with (4. The space-time split of p into energy (or relative mass) E and relative momentum p is given by P'lo = E where E = p. and these fields are the same for all such observers.BE) = C (E /\ B) ( (4. energy-momentum vector) of a particle. If '10 lies in the plane of the timelike bivector t.7a) (4. where E = (F. 7b) (4. but it must be remembered that the imaginaryi here is the unit pseudoscalar and so has a definite geometric meaning. to any such observer F consists of parallel electric and magnetic fields. a magnetic (relative bivector) partiB.39). where E· B = ~(EB + BE) is the usual dot product for Euclidean 3-space. The decomposition is invariant.v2).9c) (4. because it is independent of any observer. Thus. (4.-EB . (4. we obtain EB=E. '10ho. At this point it is worth noting that the geometric product of relative vectors E and B can be decomposed into symmetric and antisymmetric parts in the same way that we decomposed the product of proper vectors.9a) represents the field formally as a complex (relative) vector.

the standard vector algebra of Gibbs is smoothly imbedded in STA and simply related to invariant spacetime relations by a spacetime split.Ex .} given by (4. which.14) R=LU. (4.} into a new frame of vectors {ep. uniquely determines the split (4. Consequently. Note that the split of (4.34). by (4. the combination (4.13) by '10 is accomplished by a split of the rotor R into the product (4.13) into a sequence of two Lorentz rotations determined by U and L respectively. as demonstrated at length in Ref.is usual cross product of Gibbs. To relate that to proper angular momentum.1a) and (4.10c) is commonly used to represent relative angular momentum. The Lorentz rotation (3.18) Hence. where ut = 'IoU'IO = U or (4. thus.s: Note that the split (4. Employing the splits (4.10c) has been used. since U can be computed from U = i.23) transforms a standard frame {lp.2.11) so often employed in the phase of a wave function. (4. the split is done differently than in the previous examples.12) into relative vector and bivcctor parts corresponds exactly to the split (4. The scalar part of this gives the familiar split p .x = Et .16).16) This determines a split of (4.15) and (4. Each such rotation takes a frame of proper vectors 'Ik (for k = 1.Ex . (4.14) of R. translations from STA to vector algebra are effortless. 17 . x .3) into a new frame 'of vectors U'IkU in the 'IO-system.px .17) In particular. This is the group of "spatial rotations" in the 'IO-system. The cross product (4.10) of the dot and cross products into the single geometric product simplifies many aspects of classical nonrelativistic physics.i (x X p) . (4.15) for variable U defines the "little group" of Lorentz rotations which leave '10 invariant. 16.19) This determines L uniquely in terms of the timelike vectors eo and '10.14) is a special case of the decomposition (3. consider a particle with proper momentum p at a spacetime point z. (4. L 2 = eo'lo.15) (4. Equation (4.P .9a) of the electromagnetic field into vector and bivcctor parts. Thus.13) A space-timc split of the Lorentz rotation (4. Our final application of the space-time split is to a Lorentz rotation. in turn. In this case. Moreover. The bivector part gives us the proper angular momentum P 1\ x = pt .x) = Et + pt .7a) we find px = (E + p) (t .12) where (4.

(5.19) in the alternative forms 2 Plo E +p (4. = v = eo = R')'oR . where e3 is identified with the spin direction s .). b) gives L = ( v')'o) 2 = + v')'o [2(1 + v· ')'0)]" 1 rn 1 [2rn(rn + P')'o + P> ')'0)]" rn+E+p 1 [2rn(m + E)]" 1 (4. 5. then (4. Thus.22) L =v. and the spinor R must also be a function of proper time.o=-=--.(T)..37a.7a) enable us to write (4.2) Then {ef.23) Then L represents a boost of a particle from rest in the ')'o-system to a relative momentum p. provided we identify eo with the proper velocity v of the body. at each time T.p.4) .20) From (4.L = ef.37a.:/~R. Since (4.1.} into the comoving frame {e" = e" (T)}. it can be solved for L in the form of (3.3} is a eomoving frame traversing the world line along with the particle.2. ')'0 expresses this as a rotation of relative vectors Uk = ')'k')'O into relative vectors (4. so that dx ~ dT = i.3) Later it will be seen that this corresponds exactly to the spin vector in the Dirac theory where the magnitude of the spin has the constant value I s I = n/2. b). so that.6a) and (4. If eo = v is the proper velocity of a particle of mass m.2.12) it follows that U can be parametrized in the exponential form (4. 5. The rotor equation of motion for R = R( T) has the form R= . The equation can be used to describe the relativistic kinematics of a rigid body (with negligible dimensions) traversing a world line . But the same equations can be used for modeling a particle with an intrinsic angular momentum or spin. we have a rotor-valued function of proper time R = R(T) determining a I-parameter family of Lorentz rotations ef.1) describes a Lorentz rotation of some arbitrarily chosen fixed frame {')'.r = .. rn rn so (3.18) has the same form as (3. The spacelike vectors ek = R')'kR (for k = 1. thus.L(T) = R(Th. 1 (5.L(T). = 0.36).21) where a is a relative vector specifying the axis and angle of rotation.3) can be identified with the principle axes of the body. This approach to spatial rotations is treated exhaustively with many applications to mechanics in Ref. Rigid Bodies and Charged Particles.r( T) with proper time T. so we write (5.Multiplication by ek. 18 . equation (5.

thus. Vl in the (5. It integrates easily to give the history (5. it is usually much simpler to solve the rotor equation (5. For a classical particle with mass m and charge e in an electromagnetic field F. then [l = 0 and (5. Besides the theoretical advantage of being closely related to the Dirac equation. The integration of (5.8) It also determines an invariant decomposition of the initial velocity v(O) into a component f-plane and a component V2 orthogonal to the f-plane.6) This will be recognized as the classical Lorentz force with tensor components rnii" = eFltvvv.7) where Ro = R(O) specifies the initial conditions. Differentiating (5. It should be noted that nonrelativistic rotor equation describes only rotational motion.11) 19 .4) describes rotational and translational motion together. it describes the motion of a "test charge" in an external field F. The dynamics of the rigid body. The corresponding nonrelativistic rotor equation for a spinning top has been analyzed at length in Ref. Though (5. This separates D into commuting parts D1 = f(e/m) cos o and D2 = f( e/m) sin ip. When this is substituted into (5. The single rotor equation (5.10) Note that this is an invariant decomposition of the motion into "electriclike" and "magneticlike" components. the action of external forces and torques on the body is completely characterized by specifying D as a definite function of proper time.22). that is. so (5.4).5) Clearly D can be interpreted as a generalized rotational velocity of the comoving frame.1) and using (5. m So (5.7) we get the explicit T dependence of the proper velocity v. the dynamics is specified by D=!_F. if F is a uniform field on spacetime.2) and (5. For example.8) is used.6) can be solved directly.5).5) gives the particle equation of motion mi: = cF·v. (5. 4.4) has the solution (5. If self-interaction is neglected.9) When this is substituted in (5. we get (5.where D = D(T) is a bivector-valued function. it has the practical advantage of being simpler and easier to solve.2) for the history xCi) is most simply accomplished in the general case of arbitrary non-null F by exploiting the invariant decomposition F = fei<p given by (1.4). while its relativistic generalization (5. we see that the equations of motion for the comoving frame have the form (5. The fact that D = 2RR -D is necessarilv a bivector is easily proved by differentiating RR = 1.4) is equivalent to the set offour frame equations (5.

we find the constants of motion k elL = k· /1'" This includes the constant (5. This can be done by using special properties of F to find constants of motion. (5. kf This implies k2 = O.16) » which can be interpreted as the frequency of the plane wave "seen by the particle. expressing the desired I dependence of F. Any plane wave field F = F(x) with proper propagation vector k can be written in the canonical form F = [z . 6. we get z(k·x) = Z(WI). whence RkR = k.4) reduces to the algebraic condition. So.14) by k and using (5.12b) with (5." Indeed. because it endows these solutions with geometrical properties not possessed by conventional "complex solutions.x(O)) = urt:. 2m where (5. Maxwell's equation (2. (5. which applies for arbitrary initial conditions and arbitrary uniform electric and magnetic fields.15) Thus. Inserting this into (5.13) we find immediately that kR is a constant of the motion.18a) R = exp (cfzI/2m) = 1 + _!:_ fZl. To integrate the rotor equation of motion R=_!:_FR.18b) 20 . Thus. It is crucial to note that the "imaginary" i here is the unit pseudoscalar. It looks far more complicated when subjected to a space-time split and expressed directly as a function of "laboratory fields" in an inertial system. we integrate the rotor equation for a "classical test charge" in an electromagnetic plane wave. as explained in Ref.12b) describe right and left circular polarizations.14) can now be integrated directly. the one parameter family of Lorentz rotations represented by R = R( I) lies in the little group of the light like vector k.15) by (5. we obtain k = kR = Rk = kR.x(O) in the phase factor.12a) where f is a constant null bivector (p = 0). As a second example with important applications.14) it is necessary to express F as a function of I.c). Multiplying (5.) .12b) and absorbing k." Since v we can integrate (5.16) immediately to get = dx / dr (5.12a. Details are given in Ref.1).17) k· (X(. has such a simple form because it is expressed in terms of invariants. Equation (5. 2m (5.b. the orientation ofi determines handedness of the solutions. (5. and the z-dependenoe of F is exhibited explicitly by (5. with the result (5.12c) where o± and P± :2: 0 are scalars. with the initial condition R(O) = 1. the pseudoscalar property of i implies that the two terms on the right side of (5. Multiplying (5.This general result. For the plane wave (5. 3.13) = 0 as well as t? = O.

as (5. and velocityv. it is the Larmor Precession (frequency) of the spin for a particle with a magnetic moment. =vv. It is of interest to separate translational and rotational modes. Note that D . Unfortunately. Then we shall give a specific application to measurement of the g-factor for a Dirac particle. Again. it is a split with respect the instantaneous "rest frame" of the particle rather than a fixed inertial frame.15) and subsequent equations. This can be done by a space-time split by the particle velocity v or by the reference vector 'Yo.3).19a) where D+ = ~(D+vDv) and D_ = (D·v)v + (D!\ v)v. we write D = Dv2 = (D .1).21) in the fixed reference system of 'Yo. These difficulties can be resolved by adopting the 'Yo-split R=LU. though they are generally coupled.2.20) exactly as defined by (2.b. the complete particle history. To facilitate the analysis for any given dynamical model.4) and dynamics by rl = (e/m)F is a geometrically perspicuous and analytically efficient means of characterizing the motion of a classical charged particle. so to speak. (5.19b) shows explicitly. where the problem of motion in a Coulomb field is also solved by the same method.2).9 a. we shall first carry the analysis as far as possible for arbitrary D. This split has exactly the same form as the split (2. The "rnagneticlike" part D_ is completely independent of the particle motion. Moreover. But it is more than that! It also provides us automatically with a classical model of spin precession simply by assuming that the particle has an intrinsic spin characterized by (5.19c) = ~(D-vDv) = (D!\v)v. (5. This includes gravitational precession [7] and electron spin precession in the Dirac theory.We shall consider both ways and how they are related. To split the rotational velocity D by the velocity v.The particle is brought to rest.19b) to express D+ entirely in terms of the proper acceleration i.19a) does not completely decouple precession form translation because rl+ contributes to both. this split determines a "deboost" of relative vectors ekeO = R'Yk'YoR = RCTkR (k = 1. so we can watch it precess (or spin) in one place. (5.4) determines both translational and rotational motions of the comoving frame (5.v)v This produces the split (5. of course. The precession is described by an equation of the form (5. At every time T. we need a way to compare precessions at different points on the particle history. by integrating (5. details are given in Ref.This gives the velocity v and. whatever the frame models physically. However. v = i! was used in (5. Also. In the rest frame the relative velocity of the particle itself vanishes.3) into relative vectors (5.c) of the electromagnetic bivector into electric and magnetic parts corresponding here to rl+ and rl_ respectively. The rotor equation of motion (5.22) 21 . VVehave established that specification of kinematics by the rotor equation (5. 6.19b) (5. any dynamics of spin precession can be characterized by specifying a functional form for rl. so the particle's acceleration is entirely determined by the "electriclike force" D+. so let us refer to it as the Larmor precession in the general case.

10c) to define the cross product yields the familiar equations for a rotating frame (5. . To do that.26) These terms combine to give the well-known Thomas precession frequency wT=((2LL)I\'Yoho=LL-LL= . but by (5.4) gives n = 2RR = zi. + usi.so diffentiation of (5.30) where as defined by (5.I( g me: 2 . (il 1\'11 1\'-"0)')'0 1+ v (5.i.2 ) F_ 1.27) The last step here. but some improvements have been introduced in the present account. Now let us apply the rotor approach to a practica problem of spin precession. In general. expressing the proper vectors in terms of relative vectors.22) gives 2LL = ULL + Linil.20) and use of (5.2)F-l·.22) becomes U= ~wU.5.i.'Yo ~. Then differentiation of (5. - zi. (5.29) The Thomas term describes the effect of motion on the precession explicitly and completely.19a). (5.28) Finally. .. was carried out by differentiating (2.22) and write (5. m.i.31) 22 .~ .~ 'rO I ='l ('112) __ 0_ 1 + va vxv . 1 + o. Solving for wand using the split (5.19c) F_ is the magnetic field in the instantaneous rest frame of the particle. ~ = v 1\ (v + 'Yo) ..18) leads to while differentiation of (2. D g) = .23) The problem now is to express w in terms of the given D and determine the relative contributions of the parts D+ and D_.24) W = -iwvo so (5. for a charged particle with an intrinsic magnetic moment in a uniform electromagnetic field F = F + + F _ . and g is the gyromagnetic ratio.6) to get in: = v 1\ '11= WL v6(v +i(vxv)). we have the desired result W =WT+WL.. writing = L~LL for the transformed Larmor precession.5) the equation of motion for the spin is s= ~ [F + ~(g . 8. (5.6) for the velocity. (5.25) i(vv)L = it + i. (5. (5..e ( F+ + -F_ me: 2 = .3) and (5. we use the time dilation factorrx. This yields the classical equation of motion (5.21) and use of (2. = v· 'Yo = dt/dT to change the time variable in (5.e [ F +. we get w Differentiation of (2. More details are given in Ref.

6a) for the velocity vector. (5.34) where Bo has the constant value Bo = ---.35) where vee = v(O) . For a particle with proper velocity u = L2/0. To conclude this Section.38c) From.(LoFLo)t] -v60 = B + --voX 1 + Voo (Bxvo) + vooExvo.32) -1- - 2 [RF_R _~ (RF_R)]. (5. some general remarks about the description of spin will be helpful in applications and in comparisons with more conventional approaches. The general solution for an arbitrary combination F = E + iB of uniform electric and magnetic fields is easily found by replacing the BMT equation by the rotor equation .3) and alternatively by the relative vector s = e3 where a = Is le3 is defined by (5. like (4. but a is usually 23 . This can be measured experimentally. LoFLo [ 2'l 1 ~ . U(O) = 1.38b) = V· s.36) or.v2)-~. e R=-FR+R-:-(g-2) 2m where . by a A straightforward gives where s = s /\ /0 is the relative spin vector. ~Bo --RF_R 1 2 (e) - 2m iBo. With initial conditions R(O) = Lo. it must be solved for the rate of spin precession.38a) (5. (5.39) where Vo is given by (2. Of course either one can be used.36) and (5. (5. which has been applied to high precision measurements of the g-factor for the electron and muon. Both vectors s and a are sometimes used in the literature. so the last factor gives directly the change in the relative directions of the relative velocity v and the spin s. and some confusion results from a failure to recognize that they come from two different kinds of spacetime split. We have represented the spin by the proper vector s = Is le3 defined by (5.34) has the same effect on both the velocity v and the spin s. To apply the BMT equation.This is the well-known Bargmann-Michel-Telegdi (BMT) equation. the relation of s to a is found to be s=a + (va . these two representations are related by sv = LaL (5. s/o = 80 +s.l)(a' v)v. (5. the solution of (5. t (5.38a). /0 = (1 . and s ·v = 0 implies that 80 = £(8v)L = £sL/o. (5. for a boost without spatial rotation.7) and v = v II v I.32) is (5. equivalently. since one determines the other.33) is the "effective magnetic field" in the /o-system.37) spacetime split of the proper spin vector s.21). The first factor in (5. (5.

(6. However.1) we obtain V F=vJ=v·J+v/\J.6) where A = A(x) and B = B(x) are vector fields. the bivector part of (6.2) is an integral form of Maxwell's equation. a bivector field F = F(x) can be expressed as a derivative with the specific form F=v(A+Bi). (6. Since VA = V . 6.simpler because its magnitude is constant.9) .5) A different field equation is obtained by using the fact that.1. while the scalar and pseudo scalar parts yield the so-called "Lorentz condition" (6. we obtain the charge conservation law v·J=O (6.l.7) v·B=O. This section surveys other simplifications to the formulation and analysis of electromagnetic equations.2) Actually.4) and an alternative equation for the E-M field vF=V/\J.l(x) is the sole source of F.2) provides the unique solution to (6. Separately equating scalar and bivector parts of (6.3) where v is the d'Alembertian (2. (6.1 = . This is a generalization of the well-known "Helmholz theorem" in vector analysis [9]. STA makes it possible to reduce pairs of equations for the electromagnetic field F = F( x) to a single "Maxwell's Equation" vF=J.1) and separating vector and trivcctor parts. 24 2 (6. if the "current" . V-I is an integral operator which depends on boundary conditions on F for the region on which it is defined. As observed in Section 2. 2 (6. ELECTROMAGNETIC FIELDS.1) can be solved for F=v -1 . 2 2 (6.relativistic approximation.3). For example. as shown in Section 9. we obtain the usual wave equation for the vector potential VA=. A + V /\ A with a similar equation for vB. so (6. the operator V has an inverse so (6.1).6) can be written F = V /\ A + (V /\ B)i . Differentiating (6. Note from (5. (6. so F has a "vector potential" A and a "trivector potential" Hi.39) that they are indistinguishable in the non.1) This reduction brings many simplifications to electromagnetic theory.6) into Maxwell's equation (6.3).8) Inserting (6. under general conditions. then (6.

16) Note that the dual Fi of the bivector F is also a bivector. (6.12a). the conservation law \7 . Hereafter we shall restrict our attention to the "physical case" K = O. (6. x) for the electromagnetic field in the compact form (6. Sometimes the source current J can be decomposed into a conduction current JC and a magnetization current \7 . The separation of Maxwell's equation (6.14) and a trivector part \7 AF =iK. the "magnetic current density. an E-M field F and a matter field AI. (6. Using (6. AI = 0.11) we obtain in place of (6.18) where we have defined a new field G=F-AI.15) becomes \7.18) becomes a well defined equation for F. (6. However. (6. so it can be set to zero in (6. in a theory with magnetic charges.17) The Gordon decomposition of the Dirac current is of this ilk. but this symmetry between the roles of electric and magnetic currents is deceptive.20) 25 .11) where K = K(x) is a vector field. STA enables us to write the usual Maxwell energy-momentum tensor T(n) = T(n(x). where the generalized magnetization 1'v1= lvI(x) is a bivector field. because one is vectorial while the other is actually trivectorial.14) by using the duality relation (1. (\7 . in most materials M is a function of the field F. thus. lvI.11) can be separated into a vector part (6. AI) = (\7 A \7) . equation (6. J = 0 implies also that \7 . thus J=J C (6. Maxwell's equation takes the form \7F=J+iI{. However.as well as \7 Bi 2 = O. (6.10) The last equation shows that B is independent of the source J. so when a "constitutive equation" AI = 1\1(F) is known (6. Then (6.19) A disadvantage of this approach is that it mixes up physically different kinds of entities. JC = O.5) again to write \7F=\7·F+\7AF.12) The pseudoscalar i can be factored out to make (6.13) +\7·1\1.12) appear symmetrical with (6.10).11) into electric and magnetic parts can be achieved directly by using (2. Because of the mathematical identity \7 . (6.17). (Fi) = K.9)." On substituting (6.15) This last equation can be made to look similar to (6.14) can be put in the form (6. (6.6).6) into (6.

Recall that the tensor field T(n) is a vector-valued linear function on the tangent space at each spacetime point z describing the flow of energy-momentum through a surface with normal n = n(x), By linearity T(n) = nl,TP, where 71,1' = n 'ip and (6.21) The divergence of T(n) can be evaluated by using Maxwell's equation (6.1), with the result (6.22) Its value is the negative of the Lorentz Force F· .1, which is the rate of energy-momentum transfer from the source .1 to the field F. The compact, invariant form (6.20) enables us to solve easily the eigenvector problem for the E-M energy-momentum tensor. If F is not a null field, it has the invariant decomposition F = fei'P given by (1.39), which, when inserted in (6.20), gives T(n)

= -~fnf

(6.23)

This is simpler than (6.20) because f is simpler than F. Note also that it implies that all fields differing only by an arbitrary "duality factor" ei'P have the same energy-momentum tensor. The eigenvalues can be found from (6.26) by inspection. The bivector f determines a timelike plane. Any vector n in that plane satisfies n /\ f = 0, or equivalently, nf = -fn. On the other hand, if n is orthogonal to the plane, then n» f = 0 and nf = fn For these two cases, (6.23) gives us (6.24) Thus T(n) has a pair of doubly degenerate eigenvalues ±~f2 corresponding to "eigenbivectors" f and if, all expressible in terms of F by (1.40b). This approach should be compared with conventional matrix methods to appreciate the simplifications achieved by STA. The versatility of STA is also illustrated by the ease with which the above invariant formulation of "Maxwell Theory" can be related to more conventional formulations. The tensor components FW' of the E-M field F are given by (1.31), whence, using (2.2), we find (6.25) for the tensor components of Maxwell's equation (6.14). Similarly, the tensor components of (6.15) are (6.26) where the brackets indicate antisymmetrization and f"vnrj =i· CrIJiviO!i,(J)' of the energy-momentum tensor (6.21) are

TIIV

The tensor components

= 't": TV =

-HillFiv

F)(o)

= Crll. F). = FW>: F a

V

(F 'iV)

_

0:,6

h'J 'iV(F2)(0)

(6.27)

19pV F ,Fa,e

2

A space-time split of Maxwell's equation (6.1) puts it in the standard relative vector form for an inertial system. Thus, following the procedure in Section 4, (6.28) splits the current .J into a charge density io-system. Similarly,

i«

= .J. io and a relative current J

.J /\ io in the

(6.29) 26

splits V = into a time derivative = v and spatial derivative \7 = !\ v = with respect to the relative position vector x = z !\ Combining this with the split (4.9a) of F into electric and magnetic parts, we get Maxwell's equation (6.1) in the split form

ax,

at

,0 . ,0.

,0

ax

(at + \7)(E + iB)

= Jo - J .

(6.30)

This can be separated into relative even and odd parts: (6.31a) (6.31b) Equation (4.lOa) gives us the decomposition \7E = \7. E + i(\7xE). (6.32)

This enables us to split (6.31a) into two familiar Maxwell equations. In a similar way the other two equations arc obtained from (6.31b). The vector field TO = TbO) = Tbo) is the energy-momentum density in the ,a-system. The split (6.33) separates it into an energy density TOO = TO .,0 and a momentum density TO the fact that anticommutes with relative vectors, from (6.26) we obtain

,0

= TO !\ ,0. Using

(6.34)

TO,O Whence, the familiar results

= ~FFt = ~(E +iB)(E - iB).

TOO = ~(E2 TO

+ B2),

(6.35a) (6.35b)

= ExB.

Mathematical advantages of writing the E-M field in the complex form F = E + iB have been noted and exploited by many investigators (e.g. [10]), but without recognizing its geometrical basis where the imaginary is the unit pscudoscalar. The spacetime split helps us with physical interpretation. Corresponding to the split F = E+iB, the magnetization field M splits into M=-P+iM, where P is the electric polarization density and M is the magnetic moment densit.y. Writing G=D+iH, we sec that (6.19) gives us the familiar relations D=E+P, H=B-M. (6.38a) (6.38b) (6.37) (6.36)

27

7. Transformations on spacetime This section describes the apparatus of geometric calculus for handling transformations of spacetime and the induced transformations of multivector fields on spacetime. We concentrate on the mappings of 4-dimensional regions, including the whole of spacetime, but the apparatus applies with only minor adjustments to the mapping of any submanifold in spacetime. Throughout, we assume whatever differentiability is required to perform indicated operations, so that we might as well assume that the transformations are diffeomorphisms and defer the analysis of discontinuities in derivatives. \Ve therefore assume that all transformations are invertible unless otherwise indicated. Let I be a diffeomorphism which transforms each point x in some region of spacetime into another point x', as expressed by (7.1) I: x ---+ x' = I (x) . This induces a linear transformation differential of tangent vectors at ;r to tangent vectors at x', given by the

I:

a

---+

a'=j(a)=a.\1.f.

(7.2)

More explicitly, this determines the transformation of a vector field a

**= a(x) into a vector field
**

(7.3)

a'

=

alex')

=

j[a(x);

x]

=

j[a(r1(x'));

r1(x')].

The outermorphism of .f determines an induced transformation particular, where j(i) = ifi, if = det j is the Jacobian of I. The transformation I also induces an adjoint transformation back to tangent vectors at x, as defined by

of specified multi vector fields. In

= -iji

(7.4) tangent vectors at x'

1which takes

b".

1: 1:

b'

---+

b = .1(b') ==

vt b' = O"J(x).

(7.5)

More explicitly, for vector fields

b'(x')

---+

b(:r) = .f[ b'(:r'); :r] = .f[l/(f(x));

»].

(7.6)

The differential is related to the adjoint by b'· j(a) According to (7.15),

= a.- .feb').

(7.7)

1 determines

the inverse transformation (7.8)

Also, however, (7.9) Thus, the inverse of the differential equals the differential of the inverse. Since the adjoint maps "backward" instead of "forward," it is often convenient to deal with its inverse 1-1: a(x) ---+ a'(x') =1-1 [a(f-1(x'))]. (7.10) This has the advantage of being directly comparable to between 1-1 and 1-1.

28

1-

Note that it is not necessary to distinguish

as (7. The transformation of a multivector field can also be defined by the rule of direct substitution: A field P = P(x) is transformed to Thus. (7. and (7.12) The chain rule is more simply expressed as an operator identity a· V' = a· 1(V") = j(a). Each of these two alternatives has something to recommend it. including the transformation law for the differential.15) The transformation rule for the curl of a vector field a V' /\ a = . (B') (A') (7.14) Differentiation with respect to the vector a yields the general transformation derivative: or V' = 1 (V") This is the most basic formulation of the chain rule. we have two kinds of induced transformations for multivector fields: The first. All properties of induced transformations are essentially implications of this rule.11) are related by the chain rule: a· V' P = a.oxP'(. This linking of the vector concept to a transformation law is axiomatic in ordinary tensor calculus.16) = V' /\ 1(a') = 1(V" /\ a') ." The term "vector" is thus associated with the differential while "covector" is associated with the adjoint. however.16) generalizes to or V" /\ A' = f-I(V' /\ A). is said to be covariant. thus. (7. is P'(x') = P'(.f(x)) = P(x). V' x/p'(x') = j(a). the outer morphism of the curl is the curl of an outermorphism. The rule for the induced transformation of the curl is derived by using the integrability condition (2. 29 .15) that the curl of the adjoint outermorphism vanishes. rather it satisfies the "product rule" h(A' /\ B') = h(A') /\ 1 + 1 /\ h(B') . (7. To extend this to multivector fields. while the second.commonly said to be contravariant. for the adjoint of a vector field. from which its many implications are most easily derived. Thus. (7. V" P' = a'· V" P'.although they are associated with different points by changing the functional form of the field. It is very important to note here that the alternative definition P'(x) == P(x') is adopted in [11]. it follows from (7. The first is said to "transform like a vector. V" = a'· V". the values of the field are unchanged . The algebraic concept of vector is determined by the axioms of geometric algebra without reference to any coordinates or transformations.13) shows. the two concepts are kept separate.13) law for the vector (7.f(x)) = (a· V' xf(x)) .11 ) Thus. by 1 or >. Directional derivatives of the two different functions in (7.15) to prove that the adjoint function has vanishing curl." while the second is said to "transform like a covcctor.17) Therefore. by t t. In geometric calculus.18) V' /\ A = 1(V" /\ A') where A = 1(A'). note that the differential of an outennorphism is not itself an outermorphism. Association of a vector or a vector field with a particular transformation law is a separate issue. (7.f (a') is therefore (7.

dkx describes the direction of the tangent space to the manifold at each point. Since the integrals are defined from weighted sums. The whole may be recovered from the parts by using the following generalization of (7. The integrand of any integral over a k-dimensional manifold is a differential k-form (8. using (7. Thus. and [12] discusses the basic concepts at greater length with applications to physics.4) we obtain 1-1(V'. clarifies and generalizes the calculus of differential forms. as indicated by the explicit x-dependence shown on the right side of (2. Of course.The transformation rule for the divergence is more complex. If the surface is not null at x. For this reason it is called a directed measure. Thus. details are given in [1].14) relating them. Only the essentials are sketched here. This formula can be separated into two parts: (7.1) where dkx is a k-vector-valued measure on the manifold. Vi = l(A. Directed Integrals and the Fundamental Theorem In the theory of integration. J(V' /\ (A'i)) = J[(V'. (J.22) 8.It) . The exterior derivative of any "k-form" which is already the exterior derivative of another form necessarily vanishes. and I dkx I is an ordinary scalar-valued measure.It) .18) and (7. A]. A) .A)i. where A' = j(A). (7. A')i] = Then. The exterior differential of a k-form L is a (k + l j-form dL defined by (8. geometric calculus absorbs. V /\ f(A'i) = V /\ [j-1(A').21) V'· j(A) j(V .-1 ![V.17). A+ (V In 1.12) and the transformation law (7. but it can be derived from that of the curl by exploiting the duality of inner and outer products (1. we can write (8. (8.19) Vi j(A) = j[ (V' In . therefore. A] = = (V' In .) . For the divergence. we have the transformation rule V'· A' = V'· !(A) = 1. as is expressed by d 2 L = O.f(i)] = V· (J. accordingly it is a k-form as defined in Section 4.20) (7. j(A) . (7. the values of L may vary with x.3) where the accent indicates that only the implicit dependence of L on x is differentiated. A')J(i).2) where h = Ik(x) is a unit k-vector field tangent to the manifold at x.13) (which can also be derived from (7. the integrand L(dkx) must be a linear function of its k-vector argument.A)] = ![V.14)): !(A) .4) 30 . V) .

then (8. 31 .1Db) Note that duality has changed the curl in (8.This is an easy consequence of the integrability condition (2.6a) where A = A(:r) is a k-vector field.5) This says that the integral of any k-form L over a closed k-dimensional manifold is equal to the integral of its exterior derivative over the enclosed (k + I)-dimensional manifold. Substitution into (8.8) = L(in). For that special case we can write (8. its 3-vector argument can be written as the dual of a vector. thus.I)-form.10b).6b) In this case. In this case. x 1 can be defined by T(n) According to (8.6a).15). However. we do have the divergence (8.4) that this integral vanishes if L = dN where N is a (k . the integrals in (8. Though TC~7)may be called the "divergence of the tensor T. In conventional approaches to differential forms this dependence is disguised and all forms are scalar-valued. and a tensor field T(n) = T[ n(x). where a (8.7) = I4 = i. therefore. An alternative form of the Fundamental Theorem called "Gauss's Theorem" is commonly used in physics.7) implies that T(n)=n. lOa) = a(:r) = A(:r)i is a vector field. It follows from (8. the exterior derivative is equivalent to the curl.5) may be called directed integrel».5) (8.2) we can write and where n is the outward unit normal defined by the relation I3n then gives Gauss's Theorem: (8. The Fundamental Theorem of Integral Calculus (also known as the "Boundary Theorem" or the "Generalized Stokes' Theorem") can now be written in the compact form (8. if L is scalar-valued as in (8.9) where n-1 = En with signature E. If L is a 3-form.6b) into the divergence in (8." it is not generally equivalent to the divergence as defined earlier for multivector fields. To emphasize their dependence on a directed measure. Then (8. (8.a.

18) with K = 0 map into the differential form equations d*w = 32 *0'.22) From the trivector Ji we get a current 3-form *0. we obtain the exterior derivative (8.21) with the help of the duality relation (1.20) Defining a dual form *w for w by *0 = d2x .11) is thus unaffected by the change of variables. Let dx = . Thus.I we can construct a l-forrn 0./fJodx" be an arbitrary "line clement. (8.1) we have (8.14) The exterior derivative and hence the fundamental theorem are likewise unaffected.18) Line elements (hx and d2x determine a directed area element d2x = (iIx /\ (hx. (8. L' (8.11) where or In other words. It has recently become popular to formulate Maxwell's equations in terms of differential forms. so it is important to understand how that articulates with STA. (Pi) . = d3x· (." that is.17) and (6.12a). Now it should be evident that the two Maxwell equations (6. (8. dL' This follows from (8. (8. = dL.19) The exterior derivative dw of the form w can be defined in terms of the curl as the 3-form (8.13) The value of the integral of (8. From the current . more explicitly.A change of integration variables in a directed integral is a transformation on a differential form by direct substitution. Projection of the bivector field F onto an arbitrary directed area defines the electromagnetic 2-form (8. (8.23a) ..15) = 1· dx = lfJodx" . a tangent vector at x.12) = L 1-1 or.16) and (8. for the k-form defined in (8.17) Like (8. (8.Ti). the last equation is a consequence of the integrability condition.4). In other words.

Inserting (8. Again a spacetime split reveals that (8. so we will not use them.27) which.9) are (1) to convert (local) field equations into integral equations and (2) to convert local or (differential) conservation laws into global (or integral) conservation laws.. the exterior differential is completely equivalent to the curl. where the integral on the right is over any 3-manifold with boundary B.20) indicates. Most important: standard form theory does not allow one to combine the two Maxwell forms (8. we deduce that . Gauss' Law.24). they detract from it.26) into (8." It remains to be seen if this The main use of Gauss's Theorem (8. A spacetime split shows that this integral formula is equivalent to Faraday's Law or to "the absence of magnetic poles.28).5). The two integral equations (8. (Indeed.) As (8.8) and (8.25) are deliberate and noteworthy.25) where i is the unit dextral pseudoscalar for spacetime. Integral Equations and Conservation Laws .f (d 2 xFh = J (d 3 xJh.24) for any elosed 2-dimensional submanifold B in spacetime.28) Adding (8. holds for any closed 2-manifolcl B.24) to (8. we use it to derive scalar-valued integral forms for Maxwell's equation. 9. depending on the choice of B.27) are fully equivalent to the vector and trivector parts of Maxwell's equation \7 F = . moreover. or a combination of the two.29) parts.f (Px· F =0 (8. (8. This section shows how..19) and (8. They can be combined into a single equation.5) and using \7 /\F = 0.1.27) by i and use (8. To derived a similar integral formula for the vector part \7 ." or a mixture of the two. we can write the integral version of the whole Maxwell's equation in the form where (. like (8. F = J of Maxwell's equation.24) and (8.23a.4).dw= o.26) to put it in the less familiar form (8.23b) The speeial notations from standard differential form theory (such as are used in (8.26) The different definitions of normal n in (8.22) in the form (8. Insertion of (8. define a unit normal n by writing (8. First multiply (8. h selects only the "invariant (= scalar--pscudoscalar) form has some slick physical applications. the d3x just gets in the way unless one is performing an integration. depending on the choice of B.b)) add nothing of value to the STA formalism.23a.b) into the single equation (6.20) into (8.27) is equivalent to Ampere's Law. 33 .5) yields the integral equation (8. and write (8. (8. As an application of the Fundamental Theorem (8.

l(x).8). (9.l (9. Note that the right side of (9.JM r G(y.2) of (9. .8) A retarded Green's function Gk can be foune! which vanishes on S2. by operating on (9. The field F is bivector-valued while the current . Indeed.5) in the region iv-I for the electromagnetic field F = F(x) "produced by" the charge current (density) . (9. For any two spacetime points z and y. d3x I. timelike vector (field) directed in the forward light cone.3) to solve Maxwell's equation \7F=.5). the Fk satisfy the homogeneous equation (9. is bounded by a/v1 = V(h) + V(t2) + W.9) be given by T(n) = GnF. ta. Let Vet) be a convex 3-dimensional region (submanifold) in Set) which sweeps out a 4-dimensional region A1.7) Because of the condition (9.3) This great formula allows us to calculate F at any point y inside M from its derivative \7 F = oxF(x) and its values on the boundary if G is known. x)J(x) 1 d4x 1 + r.Let A1 be a 4-dimensional region in spacetime. (9. so A1. it is easily shown that F(y) satisfies Maxwell's equation (9. so Set) is a surface of simultaneous t. z) 1L F(x) 1 (9. Then At is the entire region between the hyperplanes Sl = S(h) and S2 = S(t2). Now let u be a constant..1) on the Green's function.6) = • Sk j.6) with \7 = y and using (9.9) yields (9.F2.2) where F = F(x) is any differentiable function.1). a Green's function G = G(y. t S. a 34 .4) The vector u thus determines an inertial frame with time variable t. (9. Let T(n) in (8. G(y. let us enlarge Vet) to coincide with Set) and assume that the integral of F over oV is vanishingly small at spatial infinity. in the time interval i : S.l is vector-valued. and (9. substitution into (8.5). We can use the integral formula (9. The vector u determines a I-parameter family of spacetime hyperplanes Set) satisfying the equation u·x = t.1) where the right side is the 4-dimensional delta function.3) gives us F(y) where Fk(Y) = .l = .6) can be regarded as defining an inverse \7-1 to the vector derivative \7. For simplicity. as asserted previously in writing (6. unit.:r) for Maxwell's equation is defined on M as a solution to the differential equation. in which case FI solves the Cauchy problem for the homogeneous Maxwell equation (9. In this interval the 2-dimensional boundary oV(t) sweeps out a 3-dimensional wall W. If y is an interior point of A1.

IV(t) (9. Inserting this into (9. we show that it gives us an immediate integral formulation of any physics conservation law with a suitable choice of T(n). Energy-Momentum Conservation: = /" T(u)ld3xl. the tensor T(n) represents the flux of energy-momentum through a hypersurface with normal n. [v + 1'· (v (1'.v.6) and integrating.r I.7): (9.23) for the electromagnetic field. This is an invariant form for the classical Licnard.y and I) denotes a l-dimensional delta function with derivative I)'.r I = [. . r dt 1 /" . As the other major application of Gauss's Theorem (8.13) and I = . (9.15) where P(t) Now for some applications. are easily adapted to the present formulation.14) for the region Ai defined above.10) can be integrated to get the field produced by point charge. so it need not be discussed here. equation (9.vl 21'·v ' (9.Green's functions for spacetime have been extensively studied by physicists and the results. For a particle with charge q and world line z = z( T) with proper time T.2_ l' 41r /I. The analysis of retarded and advanced parts of G and their implications is standard. the charge current can be expressed by (9. iJ)] q 41r(1'·v)2 r /I.Wicchart field.6) with (9. Ai j.. v 1 1'vv1'] [ --+--11'/l.IV(t) fl d3. v = dv/dT are all evaluated at the intersection of the backward light cone with vertex at x.z satisfies 1'2 = 0 and z. from [13] we find the following Green's function for (9. In general.9).9) in the form (9.11 ) where v = V(T) = dz/dT.6) and (9.16) We first suppose that T(n) is the energy-momentum tensor for some physical system. 35 .12) where r = x . v)3 /I.10) where r = x . we find that the retarded field can be expressed in the following explicit form F(x) = . Thus. For example. which could be a material medium. we can write (8. Introducing the notations (9. Taking /\/1 to be all of spacetime so Fl and F2 can be set to zero. or some combination of the two. f I d4. it could be the energy-momentum tensor (6. and it could be either classical or quantum mechanical. contained in many books. an electromagnetic field.

1L is the work done on the system and F = .15) can be written dP -d =.22) splits z into a time t = z -u and a position vector x = x 1\ 1L.19) where E = P .23) f· T(n) I dx 2 I. T(n) = T(n) 1\ u is the stress tensor.F 1\ u is the force exerted on it. Angular Momentum Conservation: The "generalized orbital angular momentum tensor" for the system just considered is defined by L(n) = T(n) With (4. Finally.26) 36 .24) These are universal laws applying to all physical systems.13) is the local energy-momentum conservation law.r.F+ t .17) bytz and separate scalar and relative vector parts to get the energy conservation law dE -=w+ dt and the momentum conservation law dp dt = F f + s·nldxl 2 (9. In the limit iz ---+ i : = t.16) is the total energy-momentum of the system contained in V( t) at time t.t characterizes the effect of external influences on the system in question. (9. The quantity I is the total Impulse delivered to the system in the region /vl. and n represents the normal as a "relative vector. Similarly we write (9. its divergence is 1\ x. where the work-force densit'y . (9.F.For the vector field f = f(. We write T(n)lL = n· s + T(n) .17) where .r) specified independently of the tensor field T(n) = T(.25) (9. Equation (9. the conservation law (9.F(t) = (9. We can decompose (9.18) JV(t) is the total work-force on the system. = n 1\ 1L (9.20) Fu=W+F. we multiply (9. The vector pet) given by (9.15) is then the integral energy-momentum conservation law for the system.8V i r T(n)ldxl.21) where n· s = u· T(n) is the energy flux. u is the energy and p = P 1\ 1L is the momentum of the system.tld3xl (9.r)).n(." We also note that XlL = nu = t +x (9.17) into separate energy and momentum conservation laws by using a spacetime split: we write PlL = E + p. where liV = . (9. equation (9. 2 .9).

(9. which satisfies (9. J Now write T(n) = \7 .15) becomes (9.29). x.: (9.28) Then the local angular momentum conservation law for the system is ~f(V) = .13) by (9. (9.f r. But.23) This is the charge conservation equation. F = O. t.30) = n· J and change the notion of (5. (\7 .13) the last term vanishes. 37 .For a symmetric tensor such as (5. (9.12) to Q(t) = {" 'u·JI(Pxl. the spin tensor for the system.15) as an integral law for angular momentum and analyze it the way we did energy-momentum.27) Now define the total angular momentum tensor M(71) = T(71) /\:r + 8(71) . Charge Conservation: From Maxwell's equation we derive the local charge conservation law \7 .29) Replacing (9. in generaL there exits a bivectorvalued tensor 8( n). F) = (\7 /\ \7) .31 ) an expression for the total charge contained in V(t). we can reinterpret (9. Then (9. telling us that the total charge in V(t) changes only by flowing through the boundary aV(t).

(English trans. Brackx ct al. Math. Brackx et al. [2] D. Third printing with corrections (1993). New Foundations for Classical Mechanics. Electrodynamics York. 15. Co. 1013-1028 (1971). J. Vectors. Manifolds and Physics. Phys. Differential Forms in Geometric Calculus. Hestenes. Co. Space-Time Algebra. Dordrecht/Boston (1984). Proper Dynamics of a Rigid Point Particle. Hestenes. [8] D.. Hestenes. Curvature Calculations with Spacetime Algebra. CLIFFORD ALGEBRA to GEOMETRIC CALCULUS. Reidel Publ. In F. 581-88 (1986).. [4] D. Hestenes and G. Pbys. Reidel Publ. Quantum Mechanics. and the classical theory of fields and particles. [10]H.7. Doran & S.References [1] D. A. Hestenes. Kluwcr: DordrcchtyBoston (1993).A. Clifford Algebra to Geometric Calculus. J. Dover. Dordrecht/ Boston (1985). 25. & M. SIAM . [10] H. 1778-1786 (1974). paperback (1987). Sobczyk. Phys. [12]D. [3] D. 589-598 (1986). (cds). 1905). Pbys. [8] D. G. Sobczyk. Hestenes. Clifford Algebras and their Applications in Mathematical Physics. Gravity. C. (1966). Third printing 1992. Gordon & Breach. [7] D. New [5] D. G.7. Press. p. (cds). The Design of Linear Algebra and Geometry. Co. Dc Witte-Moretto North-Holland.M.: 1980. New York. Hestenes. Poincare. 14-144 (1971). Theo. [11] A. [6. Phys. Hestenes. Math.269-285. p. 38 . New Foundations for Classical Mechanics. (1995) submitted. Dover.. Clifford Algebras and their Applications in Mathematical Physics. Co. Math. Thco. Hestenes. Science and Hypothesis. Hestenes..E. [13]Barut. Rev. (1974). A Unified Language for Mathematics and Physics. [9] Y. Analysis. Dordrecht. G Reidel Publ.269-285. New York (1951).. London. 39. A Uni(1984). Kluwcr: DordrcchtyBoston (1993). A Spinor Approach to Gravitational 25. Dordrecht/Boston (1985). Phys. 21. Proper Particle Mechanics. Acta Appliceiuise 23. Motion and Precession. Gauge Theories and Geometric Algebra.. New York (1977). . 15. [6] D. Hestenes and G. Spinors and Complex Numbers in Classical and Quantum Physics. Lasenby. Am. Appl. Gull. Oxford U. Moses. Hestenes./Boston paperback (1985). Differential Forms in Geometric Calculus. Mathematicae [7] D. Hestenes. G. [5] P. J. Int .] D. Dillard-Blcick.65-93 (1991). C. Choquet-Bruhat.1768-1777 [9] D. 4th edition (1958). Reidel Publ. J. fied Language for Mathematics and Physics. In F. Int. Dirac.

[11]D. Hestenes. The Role of Tetrad and Flat-Metric Fields in the Framework of the General Theory of Relativity. of Space-time. Continuous Groups of Transformations. Dover. Common (cds. Davis. [13] D. Herlt & M. E. 589-98 (1986). [15] L. Kramer. Stephani. 227-244 (1986). Cambridge U. 39 . The Large Scale Structure Cambridge (1973). Ellis. Exact Solutions of Einstein's Field Equations. Spinor Approach to Gravitational 25. D. Chisholm & A. Reidel. Int.). Press. [12] W. Clifford Algebras and their Applications in Mathematical Physics. Schmutzer (Ed. E. New York (1961). H. Sobczyk. Cambridge U. Nuovo Cimento 43B. Killing Vectors and Embedding of Exact Solutions in General Relativity. 2816-20 (1966). p. [16] S. MacCallum. Motion and Precession. Cambridge (1980). Press. [14] G.). J. Phys. Hawking and G. R. Eisenhart.. In J. Dordrcchty'Boston. Theo.

P! we simply interpret the II' as matrices. (7 .3) from (7. c) and (5. a column matrix of 4 complex numbers. II' are components of the electromagnetic vector potential.1b) allows us to make it even by multiplying on the right by 10.2) we may assume that ~! is a real even multi vector. and i' as the unit imaginary in the complex number field of the Dirac algebra.1c) enables us to make it real without altering (7. for if any term is odd. THE REAL DIRAC EQUATION.2) to get the standard form (7.1a) (7.1a. The correspondence must be one-to-one. Alternative proofs are given elsewhere. The representation chosen here has the advantages of simplicity and.12 but all representations are. Thus. we might refer to . mathematically equivalent. Furthermore.2) by replacingi' in the term by 1211 on the right. it plays the role of an operator generating observables in the theory. 1 pseudo scalar and 6 bivector dimensions. in (7. then (7.2) where \IF is a matrix which can be expressed as a polynomial in the Ill' The coefficients in this polynomial can be taken as real.1b) 1211U = z·1u . we can write any Dirac spinor \IF = 1{m .4) was much more difficult.PART II: QUANTUM THEORY 10. There are other ways to represent a Dirac spinor in the STA. with 1 scalar.¢! as the operator representation of a Dirac spinor. Now. for if there is a term with an imaginary coefficient. as 4 x 4 Dirac matrices. Now we may reinterpret the IlL in 1{!as vectors in the STA instead of matrices.e 40 . because. because the space of even multivectors (like the space of Dirac spinors) is exactly 8-dimensional. To distinguish a spinor 1{!in the STA from its matrix representation \IF in the Dirac algebra. In terms of the real wave functionto. the Dirac equation for an electron can be written in the form (7.t-" The original converse derivation of (7. then (7. as we shall see. Letu be a fixed spinor with the properties (7. while the A" = A . we begin with a Dirac spinor \IF. the polynomial can be taken to be an even multivector. b.4) This completes the proof. we have established a correspondence between Dirac spinors and even multivectors in the STA. Alternatively. To prove that this is equivalent to the standard matrix form of the Dirac equation. Thus. of course. multiply by tz on the right and use (7. To find a representation of the Dirac theory in terms of the STA. as shown below. ease of interpretation. (7.3) where m is the mass and e = -I e 1is the charge of the electron. let us call it a real spinor to emphasize the elimination of the ungeometrical imaginary i'.1c) In writing this we regard the IlL' for the time being.

timelike unit vector while i is a spacelike unit bivector which commutes with {D.5) where A = AM. In other words.12) where 'Po is a scalar constant.Henceforth. and the notation i = {2{1 = i{3{0 = iU3 (7. It need not be coupled to a change in reference frame.10) This transformation is no more than a change of constants in the Dirac equation. in the matrix formulation it can be interpreted as a mere change in matrix representation.8) {b = {o and i' = i. though it is often convenient to do so.4).8) gives i' = UiU. we can work with the real Dirac equation (7.Ji is the electromagnetic vector potential.4) of the Dirac equation. so UU = 1. Note that U2 = {2{0 anticomrnutcs with both {o and i = iU3.5) on the right by U2 yields (7.2) gives \jf = ~JU =¢' U' . We know that computations in STA can be carried out without introducing a basis. so ~) and (7. (7.14) 41 .6) emphasizes that this bivector plays the role of the imaginarys' that appears explicitly in the matrix form (7. future-pointing. To interpret the theory.13) arc solutions of the same equation. and A corresponding change in the wave function. These constants need not be associated with vectors in a particular reference frame.3) without reference to its matrix representation (7. so multiplication of the Dirac equation (7. it is crucial to note that the bivector i has a definite geometrical interpretation whiles' does not.5) into an equation of the same form: (7. (7.11 ) where u' = Uu. Equation (7.1) to write the real Dirac equation in the coordinate-free form (7. despite the explicit appearance of the constants {o and i = {2{1 in it. in the particular matrices selected to be associated with the vectors {M' for (7. Indeed. For the special case (7. (7. that is. the Dirac equation docs not distinguish solutions differing by a constant phase factor.9) induces a mapping of the Dirac equation (7. so let us use (4.7) where U is a constant rotor. It is only required that {o be a fixed. (7.5) is Lorentz invariant. The constants can be changed by a Lorentz rotation (7.

at each spacetime point x the timelike vector v = v(x) = co(x) is interpreted as the probable (proper) velocity of the electron.. Y IllY ".e. because the pseudoscalar i anticommutes with the "1M" Two of the vector fields in (7. Since the real Dirac wave function 'ljJ = ~)(x) is an even multivector. Thus these fields do not depend on any coordinate system. Thus. like parity conjugation. (7.16) where RR =RR= 1. the definition of charge conjugate is arbitrary up to a constant phase factor such as in (7.18).' II. in accord with the standard Born interpretation. Of course. First.where (7. "1. it will be seen that the cM can be interpreted directly as descriptors of the kinematics of electron motion.18) In other words. proper probability density) that the electron actually is at x.' . we know from (l. Its geometrical meaning is determined by what it does to the "frame of observables" identified below..T ustification for this interpretation comes from angular momentum conservation treated in the next Section.7) and (7. that the rotor field R = R(x) is a descriptor of electron kinematics.16) is an invariant decomposition of the Dirac wave function into a 2-parameter statistical factor (pei(J) ~ and a 6-parameter kinematical factor R. The correspondence of (7.19) Note that that we have here a set of four linearly independent vector fields which are invariant under the transformation specified by (7. Note in Table I that this vector quantity is 42 . we interpret as a probability current.. the transformation l/J ---+ ~p can be interpreted as charge conjugation. the canonical form (7. The second vector field (7.19) are given physical interpretations in the standard Dirac theory.8).20) to the conventional definition of the Dirac current is displayed in Table 1.19). From (7.. without any reference to a complex conjugation operation of obscure physical meaning. We shall see that that the physical interpretation given to the frame field {eM} is a key to the interpretation of the entire Dirac theory.16) does not contribute to (7. and p = p( x) is the relative probability (i.15) charge conjugation. is formulated as a completely geometrical transformation. Thus. Note also that the factor ei. despite the appearance of "1M on the left side of (7. therefore. The main thing to notice here is that in (7./-' "y' "I. .20) ~)'Yo'ljJ = peo = pv is the Dirac current.17) At each spacetime point x.6/2 in (7. which. therefore.21) will be interpreted as the spin vector density.13).1.16) and (7. The factor (pei(3)~ will be given a statistical interpretation. It follows from (7.21c) that it has the Lorentz invariant decomposition (7. the rotor R = R(x) determines a Lorentz rotation of a given fixed frame of vectors hM} into a frame {eM = eM(x)} given by (7.19). the vector field (7.15) The net effect is to change the sign of the charge in the Dirac equation.v . R determines a unique frame field on spacetime. Specifically.18) we find that "/.-y· - pe I" (7.

1c) and (7. Indeed. (7..16) we obtain (7. The hidden relation of spin to the imaginary i' in the Dirac theory can be made manifest in another way. = (pv) . Multiplying (7.2) to translate this into the matrix formalism. 'I". it shows that the bivector ~ih is a reference representation of the spin which is rotated by the kinematical factor R into the local spin direction at each spacetime point. Angular momentum is actually a bivector quantity. '1'5={0'(1/2"(3 for the matrix representation represented as a pseudovector (or axial vector) quantity in the conventional matrix formulation. This establishes an explicit connection between spin and imaginary numbers which is inherent in the Dirac theory but hidden in the conventional formulation.TABLE I: Scalar Vector W'I.22) The right side of this chain of equivalent representations shows the relation of the spin to the unit imaginary i appearing in the Dirac equation (7. we obtain S1]1 = ~ih1]1 .1]1 BILINEAR COVARIANTS = 1]1t'10'1" 1]1= (~''IO. ~i'h is the eigenvalue of the invariant "spin operator" S .23) Then using (7..d! ) (0) t = (~J'IO.lsnf3rywY(J' " -2 43 (7. 'Ill = PV/l Bivector Pseudovector* Pseudoscalar* *Here we use the standard symbol of the unit pseudoscalar i.5). a connection. as seen in a later Section.21) on the right by 1jJ and using (7... as shown below. The spin pseudovector is correctly identified as is. moreover. which remains even in the Schroedinger approximation.25) .) . The spin angular momentum S = Sex) is a bivector field related to the spin vector field s = s(x) by (7.24) Thus..'Iv)(O) = (~) '10 'I.

B) Alternatively. Equation (5. p. The eigenvalue is imaginary because the "spin tensor" Sa.30) reveals that in the Dirac theory the magnetic moment is not simply proportional to the spin as often asserted.28) The last expression shows that the s=' are simply tensor components of the pseudovector is. Their interdependence has been expressed in the literature by a 44 .[ to S is much simpler than any relation of lVIo:. thus EIJVo:.28) with u.3 can be defined simply =ih" /\. The spin S is buried there in the magnetization (tensor or bivector). By the way. 59 of Ref.24) is completely general. Note that the "alternating tensor" as the product of two pseudoscalars. = v . another indication that S is the most appropriate representation for spin.30) The interpretation of J\.a is skewsymmetric. The magnetization IVI can be defined and related to the spin by (5.. There are 1 + 4 + 6 + 4 + 1 = 16 distinct bilinear covariants but only 8 parameters in the wave function.30) provides some justification for referring to (3 henceforth as the duality parameter.27b) we find (5. so the various covariants are not mutually independent. though of course it must be implicated.Otherwise said. Contraction of (5. 'Y./'/\ 'YO: /\ 'Y(3) = (h"'Y'/'YO:'Ya)(O) (7. 22) either explicitly or implicitly introduce the spin (density) tensor (7. 'Yv and use of duality (1. Standard accounts (e.27c) From (5.3 EIWo:. and s«. applying to any Dirac wave function whatsoever. The identification of Sa.3 in (7. We are now better able to assess the content of Table 1. note that (5.1 in Table 1. It should be noted also that (7. the factor i'7i in the Dirac theory is a representation of the spin bivector by its eigenvalue. Note that the spin bivector and its relation to the unit imaginary is invisible in the standard version of the bilinear covariants in Table 1. (5.29) Its significance will be made clear in the discussion of angular momentum conservation. The name is noncommittal to the physical interpretation of (3.26) where use has been made of the identity (7.27a) and the expression for 8.g. It may be appreciated that this relation of J\.3 in the literature.25) as spin tensor is not made in standard accounts of the Dirac theory..27b) = h3 /\ 'Y2 /\ 'Yl /\ 'Yo) • hI' /\ 'YV /\ f /\ . a debatable issue discussed later.3 to svo:.[ as magnetization comes from the Gordon decomposition considered in the next Section.26) and (5. The fact that S = S(::c) specifies a definite spacelike tangent plane at each point x is completely suppressed in the i'7i representation. the two are related by a duality rotation represented by the factor ei.6.16b) gives the desired relation between Svar. (7.

The conventional justification for including thei' is to make antiherrnitian operators hermitian so the bilinear covariants are real. that Evidently Table I tells us all we need to know about the bilinear covariants and makes further reference to Fierz identities superfluous. Perhaps the most significant thing to note about Table I is that only 7 of the 8 parameters in the wave function are involved. we arrive at a geometrical interpretation of the phase of the wave function which is inherent in the Dirac Theory. This parametrization reduces the derivation of any Fierz identity practically to inspection. To understand the significance of this.1) is equally suited to the formulation of Maxwell's equation (4. a challenge that will be met in subsequent Sections. Note that the factor i'h occurs explicitly in Table I only in those expressions involving electron spin. It is literally an angle of rotation in this plane and the spin bivector S = C2Cl = RiR is the generator of the rotation. STA challenges us to ascertain the physical significance of these geometrical facts. the vectors Cl and C2 do not appear in Table I except indirectly in the product C2Cl. One of the miracles of the Dirac theory was the spontaneous emergence of spin in the theory when nothing about spin seemed to be included in the assumptions. OBSERVABLES AND CONSERVATION LAWS. in contrast to the vectors Co and C3 representing velocity and spin directions. We have seen however that this smuggles spin into the expressions..31) where r is any matrix operator. see Ref. 6.24) to derive the general identity (5. 23). But all of this is invisible in the conventional matrix formulation. /3. That can be made explicit by using (5. and we have concluded that the Dirac algebra arises from spacetime geometry rather than anything special about quantum theory. Once it is understood that the two formulations are completely isomorphic. However.system of algebraic relations known as "Fierz Identities" (e. Note. VI" 81"' which constitutes a set of 7 independent parameters. We have already noted that 5 of these parameters are needed to determine the velocity and spin directions Co and C3.18). Thus. By duality. By revealing the geometrical meaning of the unit imaginary and the wave function phase along with this connection to spin. The purpose of Table I is to explicate the correspondence of the matrix formulation to the real (STA) formulation of the Dirac theory. This miracle has been attributed to Dirac's derivation of his linearized relativistic wave equation." However. we have seen that the Dirac operator (4.16) reduces the relations to their simplest common terms. The remaining parameter therefore determines the directions of Cl and C2 in this plane. note also that. so spin has been said to be "a relativistic phenomenon. these vectors also determine the direction C2Cl = iC3CO of the "spin plane" containing Cl and C2. the matrix formulation can be dispensed with and Table I becomes superfluous. since the velocity and spin vectors are constrained by the three conditions that they are orthogonal and have constant magnitudes. Table I shows exactly how the covariants are related by expressing them in terms of p.4). the invariant decomposition of the wave function (5.g. for example. The missing parameter is the phase of the wave function. The origin of spin must be elsewhere. Our objective here is to ascertain precisely what features of the Dirac theory are responsible for its extraordinary empirical success and to establish a coherent physical interpretation which accounts 45 . The missing parameter is one of the six parameters implicit in the rotor R determining the Lorentz rotation (5.

6) v· (ps) = -rnsinf/. The first point to be understood is that there is more to the Dirac theory than the Dirac equation.. This interpretation can be upheld only if the Dirac current is rigorously conserved. 46 .t'l . A set of observables is said to be complete if it supplies a coherent physical interpretation for all mathematical features of the wave function. we shall see that certain features of the Dirac theory conflict with the view that the HOO Principle is a universal principle of quantum mechanics. The geometric insights of STA provide us with a perspective from which to criticize some conventional beliefs about quantum mechanics and so leads us to some unconventional conclusions.. though operators may be used to express the association. Observables of the Dirac theory are associated directly with the Dirac wave function rather than with operators. 2 = Yi pA .18) to get The scalar part of this equation gives us (6. The task now is to specify these observables and their conservation laws unambiguously. energy-momentum and angular momentum. A complete set of observables is determined by the conservation laws for electron position.20). (pv) v· (pel) v· (PC2) = of. There are other conserved currents besides the Dirac current.5) on the right by i)O"i3"if. namely. that is. However.2) Thus we have the four equations v . e2 .5) (6. To establish that. Now. 4. It is contended that the successes of HOO Principle derive from one set of operators only. Let's call this the HOO Principle. = i·1". There is no denying that impressive results have been achieved in quantum mechanics using the HOO Principle. there is a strong tradition in quantum mechanics to associate Hermitian Operators with Observables and their eigenvalues with observed values. so further argument is needed to justify its interpretation as probability current. We must establish the internal and external validity of the interpretation. We assume first of all that the Dirac theory describes the electron as a point particle.. we follow Appendix B of Ref.1) Moreover. (6.4) (6. but the description is statistical and the position probability current is to be identified with the Dirac current (5. the Dirac wave function has no physical meaning at all apart from assumptions that relate it to physical observables. Equation (6. In the approach taken here observables are defined quite literally as quantities which can be measured experimentally either directly or indirectly. . Indeed. we must show that internally it is logically coherent with the interpretation of other observables and externally it agrees with empirical observations. the kinetic energy-momentum operators bt defined in the convention matrix theory by pf. charge.3) is the desired position probability conservation law. = -Yi 2 pA· Cl . The meaning of the other equations remains to be determined.e A u : no. (6.3) (6. it will be seen that STA leads us to a new view on why these operators are so significant in quantum mechanics.t(pv1t) = 0.t:.r and using (5.for all its salient aspects.. multiplying the Dirac equation (5.

we interpret v( x) as the expected proper velocit'y of the electron at z . The velocity vex) defines a local reference frame at x called the electron rest frame. The streamline through a specific point Xo is the expected history of an electron at Xo.3) implies that through each spacetime point there passes a unique integral curve which is tangent to v at each of its points.7) It is normalized so that . that is. Parametrized by proper time T.r . The proper probabilit'y dcnsicy P = (pv) . the streamline x = X(T) is determined by the equation -d dx .T =V(X(T)) .8) where the integral is over the spacelike hyperplane defined by the equation . we pause to consider the main theorem relating local Howto the time development of spatially averaged observables. (6. In keeping with the statistical interpretation of the Dirac current.t = / ddlf) \t (6. V. we use 3.I d x Po 3 = 1.r) to each spacetime point x where Pi. The average value of f at time t in the 10-system is defined by (f) = . that is.13) This result is known as "Reynold's Theorem" in hydrodynamics. and (6.14) 47 .r as observable. Although our chief concern will be with observables describing the How of conserved quantities along streamlines.I d3x pv .9) is the probabilit'y densit. The result is helpful for comparison with the standard operator approach to the Dirac theory.O.12) with the derivative on the right taken along an electron streamline. Assuming that Po vanishes at spatial infinity. Let f be some observable in the Dirac theory represented by a multi vector-valued function f = f(x). of course!).t) dt = .Y in the lo-system. (6. Gauss's theorem enables us to put (6. we have the average position of the electron given by (6. By a well known theorem..f) = po- V.t = df P dT = Po dt ' df (6. In any spacetime region where P i.I d x Pof .10) The physical significance of these predicted electron histories is discussed in the next Section. the probability that the electron actually is at x is given by (6.11) To determine how this quantity changes with time. it is the optimal prediction for the history of an electron which actually is at Xo (with relative probability p(xo). 3 (6. In the 10-system.The Dirac current pv assigns a unit timelike vector v(.J(pv'".12) in the useful integral form !!_ (.0. 10 = t. Let us call these curves (electron) streamlines. v can be interpreted as the probability density in the rest frame. a solution of the Dirac equation determines a famil'y of streamlines which fills the region with exactly one streamline through each point. the probability conservation law (6. Taking the proper position vector . the velocity predicted for the electron if it happens to be at x.

this leads to peculiar and ultimately unphysical concluslons. The (Yk are hermitian operators often interpreted as "velocity operators" in accordance with the HOO Principle. However. It comes from the interpretation of the PM in (6.P" STA resolves the difficulty by revealing that the commutation relations for the (Yk have a geometrical meaning independent of any properties of the electron. usc the space-time splits (2. Then charge conservation V .21) where the underbar signifies a "linear operator" and the operator the bivector i = as defined by .1.16) -(xho from. Having chosen a particle interpretation for the Dirac theory. we have = 1 + (v) (6.1a) and (2.22) 48 .r'' The present definitions of position and velocity (without operators!) are actually equivalent to the most straight. (Yk = IklO = IOlk is the matrix analog of Uk = IklO in STA. as noted before. It shows that the ak are "velocity operators" in only a trivial sense. (6.14). One more assumption is needed to complete the identification of observables in the Dirac theory.19) and Table I leads to two important conclusions:" The hermiticity of the ak is only incidental to their role in the Dirac theory. not in the operators ak.15). and d dt (6. the equivalence of STA representations to conventional operator representations exhibited in (6..and application of (6.forward operator definitions in the standard formulation. Later it will be seen that there is more to this story. Accordingly.·3 d x pv = (dX) dt (6. The role of the ak in (6..19). (6. The (Yk simply pick out its components in (6. specifically.17) !!:.18) to the matrix formulation.18) These elementary results have been belabored here because there is considerable dispute in the literature on how to define position and velocity operators in the Dirac theory.13) gives the average velocity - dt d (x) = . i signifies right multiplication by (6.2. J = 0 is an immediate consequence of probability conservation. To establish that we use Table I to relate the components of in (6.19) is isomorphic to the role of basis vectors Uk used to select components of the vector v. with the result (6. The velocity vector is inherent in the bilinear function \[njJt.20) J = elho~) = epv .15) To sec that this is a sensible result. (6.6) to get (xho = 1 + (x) from (6.1) as kinetic energy-momentum operators. the assumption that the particle is charged implies that the charge current (density) J must be proportional to the Dirac current. (x) = (v) = / dt \ dt dX) . In the STA formulation they arc defined by (6.19) where. and their eigenvalues have no physical significance whatever! These concepts play no role in the STA formulation. Thus.

Indeed. as is evident in (5. which provide clues to a deeper physical meaning. it embodies the fruitful "minimal coupling" rule. as is clear when the equation is written in the form (6. This -leads to the conclusion that the significance of the PI" lies in what they imply about the physical meaning of the wave function. with its emphasis on attributing physical meaning to operators and their commutation rules. the STA formulation tells us even more. We establish below that this connection is a consequence of the form and interpretation of the Pw Thus. The connection of i''h with spin is not inherent in the PI' alone. relating real and matrix versions. the STA formulation reveals the PI'. We have already noted a connection of the -factor in with spin in (5.21) can hardly be overemphasized.23) However. equivalently. (6. spin was inadvertently smuggled into the Dirac theory by the PI'. The operators PI" or. It appears only when the PI' operate on the wave function.v Pv are given a physical meaning by using them to define the components TI"v of the electron energy-momentum tensor: (6.TABLE II: Observables of the energy-momentum operator.25) 49 . pi' = II' '. As mentioned in the discussion of the electromagnetic energy-momentum tensor. Above all.24) Its matrix equivalent is given in Table II. It reveals geometrical properties of the Pp.22). have something important to tell us about the kinematics of electron motion. Energy-momentum tensor Kinetic energy density Kinetic momentum density Angular Momentum tensor Gordon current The importance of (6. Its sudden appearance was only incidentally related to relativity. In this capacity it plays a crucial heuristic role in the original formulation of the Dirac equation. History has shown that it is impossible to recognize this fact in the conventional formulation of the Dirac theory. a fundamental principle of gauge theory which fixes the form of electromagnetic interactions.' hidden in the innocel1t looking iector i'ti.24).

(T!' /\ x) = Til /\ I" .28). because it includes a contribution from the spin.27) as the total energy-momentum conservation law (6.29) To evaluate the first term on the right. .21) in the second term.27) where J is the Dirac charge current (6..26) and (5. momentum and angular momentum conservation laws can be established by direct calculation from the Dirac equation.27) we obtain the angular momentum conservation law 8.8"TI' /\ x.35) .24) of Ti'" in terms of the Dirac wave function. The right side of (6. we find that? (See Appendix B for an alternative approach) (6.34) + pS" (6. In general. by comparison with (2. First.28) To derive the angular momentum conservation law.1er") = . Noting that 8. The observable p = p(x) is the statistical prediction for the momentum of the electron at x. However.4).JIt = (F . we obtain (6. so using (4. From the definition (6.20) and F = V /\ A is the electromagnetic field. Thus from (6.33) is the spin angular momentum tensor already identified in (5.. the pseudoscalar part of this equation proves (6.1) /\ x . we return to the definition (6.fI' = T" /\ x 50 (6. we calculate 8.27) is exactly the classical Lorentz force. the momentum p is not collinear with the velocity. where .M' we can rephrase (6. we identify Tit /\ z as the orbital angular momentum tensor (See Table II for comparison with more conventional expressions).24) and find Summing with IV and using the Dirac equation (6.26) This defines the "expected" proper momentum p = p(x).30) By the way.25) and denoting the electromagnetic energy-momentum tensor (4.::c = I". (6.24) by T}:. will be recognized as defining the relative momentum in the electron rest frame. The energy-momentum density in the electron rest system is T(v) = vltTIt = pp.7c). the bivector part gives the relation we are looking for: (6.is the energy-momentum flux through a hyperplane with normal lit . which. and the scalar part gives the curious result (6.32) where (6.31) Tltl' = Tit '111 = rn cos (3 .23) to evaluate the first term on the right while recognizing the spin vector (5. (6.29) and (6. A measure of this noncollinearity is p /\ v.

equivalently. can be put in the form (6. In the electron rest system. it describes the rotation rate of the frame {ef. by (6.36) where recalling (2. (6.18) have (6. v.26)..t} in the spin plane or." But we have identified the angle of rotation in this plane with the phase of the wave function. the effect of Pv on'lj. therefore. but we still need to ascertain precisely how p is related to the wave function.23).}. of the frame {el. The kinematical import of the operator p" is derived from its action on the rotor R. the angular momentum density is J(v) = p(p!\ x + S). we get the explicit expression (6. VVenow have a complete set of conservation laws for the observebles r. (6. the momentum describes the phase change in all directions of the wave function or. p!\x is recognized as the expected orbital angular momentum and as already advertised in (5.42) This reveals that (apart from the A" contribution) the momentum has a kinematical meaning related to the spin: It is completely determined by the component of It" in the spin plane. The task remaining is to dig deeper and understand its origin. it is easy to prove that the derivatives of the rotor R must have the form (6. This completes the justification for interpreting S as spin. if you will "about the spin axis.is the angular momentum tensor. In other words.22) to get (3"R)i1iR = It"S = 11" .24) for the energy-momentum tensor.37) where Ilf.22) was used to establish that (6. S + Itv!\ S + 3"S. To make that explicit. By an argument used in Section 3.' direction. Sand p./'O .39) vVhence (p" 4) ho.} as it is displaced in the direction 'Iw Now.44) 51 . For that purpose we employ the invariant decomposition 4) = (peiB)~ R. Consequently the derivatives of the e. the momentum components (6.'(x) the form is a bivector field. use (6./ (In p + i/3) + 11" lips - eA.43) where (5.40) Inserting this in the definition (6. (6./ defined by (5. A physical interpretation for this geometrical fact will be offered in Section 8.' = Ilf. S =isv can be indentified as an intrinsic angular momentum or spin. representing the total angular momentum flux in the 'If.12). Thus.38) Thus 11" is the rotation rate of the frame {e. First we need some kinematics.J = [3.41) From this we find.37) and (5. after some manipulations beginning with is = S'O.22). with the help of (5.

we have I ': C ')~ J . /3 is related to J only indirectly through the Gordon Relation (6.qsin. ) 'II) = ( PfL . the significance of the parameter .20) for the Dirac current in terms of the wave function. from the last term in (6.) + zqfL pe ij3 + afL (PSe:ii3) . pseudoscalar and bivector parts explicitly.52) J=K+V·M. Note thafboth Pv and uvS contribute to T/J. in a different guise when it was 52 . rn - (6. = CljY.V in (6.23) for the Dirac charge current. (6. both of which are separately conserved. This suggests that (3 characterizes some feature of the substructure. Evidently if we are to understand this substructure we must understand the role of the parameter (3 so prominently displayed in (6. The problem arose. AI. An auxiliary conservation law can be derived from the Dirac equation by decomposing the Dirac current as follows. K =~ m p(p cos .46) This shows explicitly how the operator Pv relates to kinematical observables. Solving (6.43) in the form (6.47). it should be noted that the last two terms in (6.6 .48) The right side exhibits the scalar. and vector part gives us the so-called "Gordon decomposition" (6.50) and (6./o1.51) When (6.47) The identity (6.6 in the last term remains obscure. but qv does not.41). But how does this square with the physical interpretation already ascribed to J'l It suggests that there is a substructure to the charge flow described by J. However. A curious fact is that (3 does not contribute to the definition (5. It is altogether natural that this flux should depend on the component of uvS as shown. So far we have supplied a physical interpretation for all parameters in the wave function (5.48) we define the magnetization M = ~ pScij3. because the structure of the theory was not analyzed in sufficient depth to identify it.41) describe energy-momentum flux orthogonal to the v direction. By the way.50) As anticipated in the last Section. m (6.49) Or in vector form. although the physical significance of qv is obscure.6).48) is inserted into (6.46) is easily generalized to ( '"( !VP .Introducing the abbreviation or we can put (6. This is ostensibly a decomposition into a conduction current K and a magnetization current V .16) except "duality parameter" (3.p Vi. From the scalar part we define the Gordon current: (6. Let us refer to this as the I)-problem.52). (6. the pseudovector part must vanish. The physical interpretation of (3 is more problematic than that of the other parameters. however.51). This problem has not been recognized in conventional formulations of the Dirac theory.( PfL1.45) (6.p = .

Z'M = -e S' pcoSfJ tri Q (6.6-problem provides a new perspective which suggests that second quantization is unnecessary. The standard Darwin solutions of the Dirac hydrogen atom exhibit a strange . while p sin /3 is the probability that it is an positron. The dependence of P on sini) makes sense.6 unequivocally distinguishes electron and positron solutions. while p = p(x) represents the relative probability of observing a particle at z . However. On this interpretation. since v • s (6.54) is the magnetic moment density. beeause pair creation produees electrie dipoles. where. that . though this is not to deny the reality of pair creation. the solutions also attribute some apparently unphysical properties to the Dirac current. to interpret equation (6. Unfortunately.4) as describing a creation of spin concomitant with pair creation. Then. it appears that a statistical interpretation for . in this Section it will be explained how particle trajectories can be computed in the Dirac theory and how this articulates perfectly with the classical theory formulated in Section 3.3 dependence which cannot reasonably be attributed to pair creation.55) is the electric dipole moment density. On the other hand.) parametrizes an admixture of electron-positron states where cos.3 sri (6. the parameter . On the contrary. Heinz Kruger has recently found hydrogen atom solutions with .3. However. This suggests that . while P= -~ iSpsin. 7. 53 . It may be. it asserted here that the particle concept is not only essential for a complete and coherent interpretation of the Dirac theory.53) = 0.noted that the Dirac equation admits negative energy solutions.3 = 0. It leads also to a plausible interpretation for the f)-dependence of the magnetization in (6.54).3 factor in (6. cancelation of magnetic moments by created pairs may account for the reduction of M by the cos.51). the Gordon current shows a redistribution of the current flow as a function of . The famous Klein paradox showed that negative energy states could not be avoided in matching boundary conditions at a potential barrier.3 is the relative probability of observing an electron. therefore. we can split 1Minto M=-P+iM. recognition of the . ELECTRON TRAJECTORIES In classical theory the concept of particle refers to an object of negligible size with a continuous trajectory. since the basic observables v. It is tempting. Indeed. It is often asserted that it is meaningless or impossible in quantum mechanics to regard the electron as a particle in this sense. there are difficulties with this straight forward interpretation of . Sand p are completely characterized by the kinematical factor R in the wave function. suggesting that they may be superpositions of more basic physical solutions.3as an antiparticle mixing parameter.3 as well as p is appropriate.3 # O. p cos (J is the probability that the particle is an electron. In accordance with (4. also.39). This was interpreted as showing that electron-positron pairs are created at the barrier. and it was concluded that second quantization of the Dirac wave function is necessary to deal with the many particle aspects of such situations.26 In the plane wave solutions of the Dirac equation (next Section). At any rate.3 characterizes a more general class of statistical superpositions than particle-antiparticle mixtures. it is also of practical value and opens up possibilities for new physics at a deeper level.27 It is easy to show that a superposition of solutions to the Dirac equation with /3 = 0 can produce a composite solution with . in the electron rest system determined by J. Indeed. A resolution of the Klein Paradox from this perspective has been given by Steven Gull.

This claim is false.3o It is worth noting.9 exactly as in the Bohr theory! Now comes the experimental evidence. However. though there are some differences in the predicted trajectories.f" Sure enough.e'' He argues that the difference between classical and quantum mechanics is not in the concept of particle itself but in the equation for particles trajectories.David Bohm has long been the most articulate champion of the particle concept in quantum mechanics. But is it equally valid to regard them as "observables" in an atom? Though the Dirac theory predicts a family of orbits (or trajectories) in an atom. however. Here we derive equations of motion for observables of the electron along a single history x = . The same general particle interpretation of the Dirac theory is adopted here. so no electron motion is indicated. the trajectories bunch up at diffraction maxima and thin out at the minima. because the magnetization current is not zero. &'3 should be obvious from the fact that. It will be convenient to use the terms 'history' and 'trajectory' almost interchangeably. except for the fact that the Dirac current does not vanish for an s-state. though it can be inferred more precisely after detection in the final state. This is justification for referring to p and v as "observables.r( T). This shows at least that the particle interpretation is not inconsistent with diffraction phenomena." The probability density Po is literally an observable in a diffraction pattern. because the Schroedinger current is the nonrelativistic limit of the Gordon current rather than the Dirac current. but histories are usually more convenient for theoretical purposes.r'' These remarks apply to the Dirac theory as well as to the Schroedinger theory. But here is some evidence to the contrary that should give the sceptics pause: The hydrogen s-state wave function is spherically symmetric and its Schroedinger (or Gordon) current vanishes. the cause of this phenomenon is the Quantum Force rather than wave interference. This would seem to be no more than a strange coincidence. as we have noted. Moreover. the radial probability distribution has a maximum at Bohr radius." though they are not associated with any operators save the Dirac wave function itself. The claim is only that the electron is likely to travel one of a family of possible trajectories consistent with experimental constraints." He is careful. We have already noted that each solution of the Dirac equation determines a family of nonintersecting streamlines which can be interpreted as "expected" electron histories. it is possible then to infer which slit the electron passed through. not to commit himself to any special hypothesis about the origins of the Quantum Force. He accepts the form of the force dictated by Schroedinger's equation. For double slit diffraction these trajectories have been calculated from Schroedinger's equation. the wave function determines a unique family of electron trajectories.29. and it is often asserted that it is meaningless to say that the electron has a definite velocity in an atom. According to Bohm. The same can be said for the velocities of detected electrons. which trajectory is known initially only with a certain probability. he takes pains to show that all implications of Schroedinger theory are compatible with a strict particle interpretation. most physicists don't take this seriously. The representation of motion by trajectories is most helpful in interpreting experiments. The main objection to a strict particle interpretation of the Dirac and Schroedinger theories is the claim that a wave interpretation is essential to explain diffraction. When negative muons are captured in atomic s-statcs their lifetimes are increased by a time dilation factor 54 . though not in intermediate states of a diffraction experiment. From Schroedinger's equation he derives an equation of motion for the electron which differs from the classical equation only in a stochastic term called the "Quantum Force. though. Indeed. though the origin of the Quantum Force remains to be explained. By a space-time split the history can always be projected into a particle trajectory x( T) = x( T) !\ 'Yo in a given inertial system. However. after flowing uniformly through the slits. that this account has the decided advantage of avoiding the paradoxical "collapse of the wave function" inherent in the conventional "dualist" explanation of diffraction. The obvious objections to this account of diffraction have been adequately refuted in the literature. the average angular momentum associated with this current is n. At no time is it claimed that the electron spreads out like a wave to interfere with itself and then "collapse" when it is detected in a localized region. and the generalization of Bohm's equation derived below provides a new perspective on the Quantum Force.

43): 0. Equation (7. as a function of the electron momentum.20) and (6. Among other things.10) 55 .corresponding to a velocity of .16) can be expressed as a function of proper time (7. this evidence supports the major contention that the electron velocity is more correctly described by the Dirac current than by the Gordon current. therefore. lop Accordingly.3) where the overdot indicates differentiation with respect to proper time. In Bohmian terms. v)iS-1 + (p + eA) . = v . v) + 8.. we find that it "spins" with the ultrahigh frequency (7.6) is part of a more general equation obtained from (6. By (5. v = ~nw.18). . this frequency will be altered by external fields.8) As an interesting aside. From (6. Vep = 0. W = W(X(T)).21) and (5. (5. 1 (7. .S = (p + eA) . that automatically gives us the classical limit formulated as in Section 3.6).. is to investigate what the Dirac theory tells us about 0. the only difference between classical and quantum theory is in the functional form of 0. (7. a limit in which the electron still has a nonvanishing spin of magnitude nj2. = v· V R = "20. (7. ep (7. Besides the idea that an electron in an s-state has a definite velocity. while the spin vector sand bivector S are given as before by (5.1) leads to R .10).R. this can be solved for n= where 8S-1 + (q. Now let us investigate the equations for motion along a Dirac streamline x = x( T). Whence.you guessed it! .7) According to (7.the Bohr velocity.37). this determines a comoving frame (7. differentiation of (7. (7.4) is the rotational velocity of the frame {ep}.6) This defines rate of rotation in the spin plane. In accordance with (6. (7..22). os:' .5) But these equations are identical in form to those for the classical theory in Section 3. Our main task.2) on the streamline with velocity v = co. For a free particle (considered below).v + i(q. S = (p + cA) .1) R = R(X(T)) .42) we immediately obtain 0.9) s :' = ie=!». On this curve the kinematical factor in the Dirac wave function (5. and (7. This is a consequence of the particle interpretation.

(\l /\ v) (7. we obtain (7. Accordingly.This shows something about the coupling of spin and velocity.v. These coupled equations have not been seriously studied. the Dirac equation (5. Inserting this into (5.21) 56 . we get from (7.19) mE. A general expression for n in terms of observables can be derived from the Dirac equation.44) the equations of motion for velocity and spin: mil = e(Feif3) Fx (: Sc • i(3 v ) + Q . This has been done in two steps in Ref. which comes from assuming that only the variation of S along the history can affect the motion. since 1) = v .15) is that the dependence of the Quantum Force on Plank's constant comes entirely from the spin S. The first step yields the interesting result n= -\l /\ v + v· (i\l(J) + (mcos.11) But this tells us nothing about particle streamlines. v is the generalization of Bohm's Quantum Force. Of course.5) and (6. 4. the first term on the right of (7.13) where Q has the complicated form (7. (7.11).20) where the kinematical factor R has been decomposed to explicitly exhibit its spacetime dependence in a phase factor.14) reduces to Q for the limiting classical equations of motion for a particle with intrinsic spin we obtain+' mi. + QxS.EA) is the commutator product. x = p. Except for the surprising factor eif3.5) and using \lp. = S. In the remainder of this Section we examine classical solutions of the Dirac equation.3).16) (7. solutions whose streamlines are classical trajectories.S) x S .14) with (7.15) Inserting (7. (7. second.16) is the classical Lorentz force. (i\l (3) = tri -1 (eFeif3 + Q) . The second step yields . but it is not useful for solving the equations of motion. they should be studied in conjunction with the spinor equation (7. (7. A crucial fact to note from (7.18) (7.=(eF-S).13) in (7. For a free particle (A = 0).12) and the fact that v2 is constant. which can be derived from (1." The classical limit can be characterized first by p --+ 0 and 8" In p ---+ 0. that is. The term Q . This spin dependence of the Quantum Force is hidden in the Schroedinger approximation. but it can be shown to be implicit there nevertheless. v.6+ eA· V)S-l. by 8"S = 'Up-B.17) B= where AxE = ~(AE .12) is a mere identity. = (eF . (7.5) admits planc wave solutions of the form (7.\l /\ v + v . and (7.

Evidently (7. As in (2.27) where T = V· x and w is given by (7.7). If the wave functions are normalized to one particle per unit volume V in the la-system.2. Thus. (7.30) 57 . For applications.26b) As seen below. for k = 1.= p2+n·ae -ip·x/n .25) VVe can identify these as electron and positron wave functions. the factor -iU2 represents a spatial rotation which just "Hips" the direction of the spin vector.24).!. but. we decompose it as follows: Choosing U3 as "quantization axis. (7. the streamlines are straight lines along which the spin is constant and Cl and C2 rotate about the "spin axis" with the ultrahigh frequency (7. Thus we obtain two distinct kinds of plane wave solutions with positive energy E = p. (pv) =- 1 V or p = -- rn EV = --- 1 10'VV . (7.7) as the electron moves along the streamline. but with oppositely directed spins.22) This implies eifJ = ± 1." we decompose U into spin up and spin down amplitudes denoted by U + and U _ respectively.28) The result of calculating L from 10 and the momentum p has already been found in (2.29) This is all we need to characterize spin.J. One solution appears to have negative energy la. Following the procedure beginning with (2. But to make contact with more conventional representations.20). Determining the comoving frame (7.24) (7.13). A similar result is found for the positron solution.26a) where (7. then we have Po = 10 .26a) are both positron solutions.25) and (7.23) and p E= = ±mv corresponding to two distinct solutions. but that can be rectified by changing the sign in the phase of the "trial solution" (7. the constants in the solution must be specified in more detail.'3 = ~nRa/3Ra are constant. so Ci. . we find that the velocity v = RalaRa and the spin . we make the space-time split R=LU where U = Uae-ip.x/n .19) and (3. Indeed. 10: p. the charge conjugate of (7. the two solutions are related by charge conjugation.15). 1.37).24).Solving for p we get (7. (7. and defined by (7.24) is (7. it is convenient to represent the spin direction by the relative vector (7.2) for the electron solution (7. According to (5.i3/2 = 1 ori.

we obtain n= ~ rn F + (m + eA.38) into the Dirac equation.39) obtained by squaring (7.39) As in the plane wave case (7. The classical limit is ordinarily obtained as an "eikonal approximation" to the Dirac equation. v cp - eA = mv .38) gives mV!\v = -eV!\A = -eF (7.47).mv.33) (7.42) . FR-Ri(m+eA. Thus. (7. say in the form (6. (7. Accordingly. (7.12).38) This defines a family of classical histories in spacetime.38).35) (7. (7. v)5-1. the wave function is set in the form (7. inserting (7. Inserting (7. but we see here its geometrical significance. (7. and the two values correspond to electron and positron solutions.34) (7.22) this implies ei(3 = ±l.31 ) Thus (7. On the other hand.6) for each streamline.11). the curl of (7. 2m 58 (7.3) assumes the explicit form . we obtain exactly the classical equation (3.p.40) Dotting this with v and using the identity (7. 0"3 XO" = 2iU_U + .41) Whence the rotor equation (7.37) This decomposition into spin up and down amplitudes is usually given a statistical interpretation in quantum mechanics. For the electron case.or (7. U+U_+U_U+=O. For a given external potential A the phase sp can be found by solving the "Hamilton-Jacobi equation" = A(:r).32) It follows that UU = IU+ 12 + 1U_12 = 1. so the derivatives of 4)0in the Dirac equation can be neglected to a good approximation.36) 0" = U0"3U = {I U+ 12 -I U_ n0"3 + 2U_U +0"3· since 0"0"3 = 0" • 0"3 + i( 0" X 0"3). (7.38) Then the "amplitude" 4)0is assumed to be slowly varying compared to "phase" .eA)e i(3 _ .v)jT!. we obtain (Vcp . e R= _.40) into (7.

Identifying this solution with Ro in (7. That possibility will be taken up in the next Section. the rotor (3. Thus. This is what one would expect classically if there were some sort of localized motion in the spin plane.S' (7. with the interesting result e p·v=rn+-F •. while (7.46) m.47) can be regarded as a generalization of the Cauchy-Riemann equations.1).44) and rp = v • V <p= rn + eA .47) This equation has a whole class of exact solutions where '1/)0 is not constant. This is easily done by integrating (7. e R= -FRo 2m (7. 59 . This class is comparable in richness to the class of analytic functions in complex variable theory.This admits a solution by separation of variables: (7.18). for (7.17) it is readily verified that V Ro = O.45) Equation (7.45) can be obtained from (7. All that remains. we note that it rules out the possibility of finding any spin dependence of the streamlines such as that exhibited in equation (7.45) or solving the Hamilton-Jacobi equation (7.18a. Quantum mechanics also assigns energy to this rotation.39).38) is generalized to a vector field with nonvanishing curl. Evidently the spin dependence appears when the v<P in (7. The eikonal solutions characterized above are exact solutions of the Dirac equation when the 1/-Jo in (7. (7. then.32 Wo have already found the classical solution for this case.41) into (7.43) where .44) is identical with the classical equation in Section 3. An important member of this class is the so-called Volkov solution for an electron in the field of an electromagnetic plane wave. namely. with the help of (3. we can regard wave functions of this class as exact classical solutions of the Dirac equation.38) satisfies (7.38). b).l' Considering the exact correspondence of the eikonal equations with classical theory. in the eikonal approximation the quantum equation for a comoving frame differs from the classical equation only in having additional rotation in the spin plane. is to determine the phase factor <p(x).v . and an explicit expression for it is obtained by inserting (7. As a final observation about the eikonal approximation.43).

4) SO 'U = i can be integrated immediately to get the electron history Z(T) = VT + (enT - l)r'o + Zo. as intended. we are prepared to examine deeper possibilities of the Dirac theory. we now define the electron velocit'y u b'y (8. For p. the (8. the electron can be on anyone of these helixes with uniform probability. there is energy associated with this rotation. indeed. Let us refer to this localized helical motion of the electron by the name zitterbewegung (zbw) originally introduced by Schrocdingcr.7) Finally. Accordingly. These facts suggest that the electron mass.6) The diameter of the orbit is thus equal to an electron Compton wavelength.L = R'/. it projects to a circular orbit of radius I TO I + Zo - TO' = I n -1 I =- ti -13 = 1.16) of the wave function. According to the Born statistical interpretation.5) where TO = n-1e2(O). This is a lightlike helix centered on the Dirac streamline ::C(T) In the electron "rest system" defined by v.24) can be written in the form (8.::c = mv· x = mr . = VT (8.2) it follows that v is constant and (8.1) The choice 112 = 0 has the advantage that the electron mass can be attributed to kinetic energy of self interaction while the spin is the corresponding angular momentum. we recognize that this suggestion can be accommodated by giving the electron a component of velocity in the spin plane. THE ZITTERBEWEGUNG INTERPRETATION. each centered on a unique Dirac streamline.5) describes a 3-parameter family of spacetime filling light like helixes. The Dirac rotor determines a comoving frame {e. For 1'(T) angular momentum of this circular motion is..Ji} which rotates at high frequency in the e2e1-plane. 2m (8.46). Now that we have the geometrical and physical interpretation of the Dirac wave function well in hand. We have seen that the kinematics of electron motion is completely characterized by the "Dirac rotor" R in the invariant decomposition (5. the spin = enTr'o." as the electron moves along a streamline.3) From (8.6) and (7. Moreover. if Zo is varied parametrically over a hyperplane normal to v. spin and magnetic moment are manifestation of a local circular motion of the electron.8.l ' This new identification of electron velocity makes the plane wave solutions a lot more physically meaningful. v of the electron.2) where n is the constant bivector (8.P Accordingly. the kinematical factor for the solution (7. we call w = n· S the zbw ircqucucy and 60 . all the rest energy p . Mindful that the velocity attributed to the electron is an independent assumption imposed on the Dirac theory from physical considerations. equation (8. according to (7. the "spin plane.9 x 10 m.

Henceforth. it completes the kinematical interpretation of R.6) we have v· (pu) = - n pA· 2 e1 . From the present point of view.9) However.The key ingredients of the zbw interpretation are preserved in the nonrelativistic limit and so provide a zitterbewegung interpretation of Schroedinger theory. 14.10) 61 PI'" . The strength of the zbw interpretation lies first in its coherence and completeness in the Dirac theory and second in the intimations it gives of more fundamental physics. the zbw radius is then seen as the radius of curvature for the particle history. the P" computes the phase rotation rates in all spacetime directions and associates them with the electron energy-momentum. characterize features of the electron history. It is also perfectly compatible with everything said about the interpretation of the Dirac theory in previous Sections. because the definition (8. Thus. because the association of the unit imaginary with spin was not established (or even dreamed of). although the phenomenon noticed by Schroedinger certainly appears when wave packets are constructed. the present interpretation was not an option open to Schroedinger.13 But the essential point can be seen by a split of the Dirac wave function y into the factors (8. it is v rather than u. Now that it has been exhumed.21). Perhaps the strongest theoretical support for the zbw interpretation is the fact that it is fundamentally geometrical.1) of the zbw velocity is well defined for any solution of the Dirac equation. this will be referred to as the zitterbew·egung interpretation of quantum mechanics. One need only recognize that the Dirac velocity can be interpreted as the average of the electron velocity over a zbw period. leading through the Pauli theory to the Schroedinger theory. is not necessarily a conserved current. wave packets and interference are not essential ingredients of the zbw. A possible difficulty with the interpretation of 1L as electron velocity is the fact that pu. so all components of R. It will be noted that the zbw interpretation is completely general. where the comoving frame {ell} is interpreted as a Frenet frame.33. The nonrelativistic approximation to the STA version of the Dirac theory. for from (6. that best describes electron motion in most experiments. even the complex phase factor.34whereas here it is associated directly with the complex phase factor of a plane wave.\ = w-1 the zbw radius.. the zbw interpretation explains the physical significance of the mysterious "quantum mechanical operators" .t defined by (6. The present approach associates the zbw phase and frequency with the phase and frequency of the complex phase factor in the electron wave function. it is probably sufficient that pv is conserved. (8. The unit imaginary i appearing in both of these has the dual properties of-representing the plane in which zbw circulation takes place and generating rotations in that plane. Of course. (8. Although the frequency and radius ascribed to the zbware the same here as in Schroedinger's work. so we can refer to that as the zbw phase.15. we can see that the zbw must playa ubiquitous role in quantum mechanics. its role in the theory is quite different. Operating on the phase factor. The key ingredients of the zbw interpretation are the complex phase factor and the energymomentum operators P. Schroedinger attributed it to interference between positive and negative energy components of a wave packet. This kinematical interpretation is made most explicitly in Ref. as expressed by writing v = ii . The phase factor literally represents a rotation on the electron's circular orbit in the i-plane. and the vector e2 needed to form the spacelike component of the zbw velocity 'U was buried out of sight in the matrix formalism. has been treated in detail clsewhere. with vectors e1 and e3 corresponding to first and third curvatures. The phase of the wave function can now be interpreted literally as the phase in the circular motion.8) Since the period is on the order of 1O-21S.

Uzcr. 11. Such success may be surprising from the conventional view of quantum mechanics. Quantum mechanics is characterized by phase coherence over distances very much larger than an electron Compton wavelength defining the dimensions of the zbw. it simply tells us that the zbw exists and describes some of its properties. The zbw fluctuations are much too rapid to observe directly. including barrier penetration and the Aharonov-Bohm effect. This rapidly fluctuating field is a prime candidate for Bohm's Quantum Force. 9. Diffraction may be explained as zbw resonant momentum exchange. The zbw interpretation explains much more than the electron spin and magnetic moment. A speculative analysis of its quantum implications is given in Ref. is neglected so 'l/Jp reduces to the Schroedinger wave function (8. 42-48 (1991). the Dirac theory is not without clues as to what to look for. vVesee now the physical significance of the complex that phase factor e-i('P/It) is kinematical rather than logical or statistical as so often claimed.12) It follows from this derivation of the Schroedinger wave function that just as in the Dirac theory.11) where the kinematical factor U has been broken into a phase factor describing the zbw rotation and a spatial rotation factor Uo which rotates i into the direction of the spin.In the nonrelativistic approximation three of these factors are neglected or eliminated and '1/) is reduced to the Pauli wave function (8. Many aspects of spin and the zbw in the Pauli theory have already been discussed in Ref. the semiclassical approach of imposing quantum conditions on classical dynamics is just of way of meeting the conditions for zbw resonances. does not explain the zbw. Uzer ctc. By what causal mechanism might zbw coherence be established over such large distances? A tantalizing possibility arises by interpreting the circular zbw orbit literally as the orbit of a point charge. P. often surpasing the results from standard quantum mechanical methods. Stationary state solutions of both the Schroedinger and Dirac equations reveal an important property of the zbw. T. though perhaps they have been observed indirectly all along in quantum coherence phenomena. we can regard the Dirac theory and all its successes as evidence that the zbw is a real physical phenomena. or better. One important clue concerns the origin of electron 62 . Considering how well the zbw interpretation fits the Dirac theory. Science 253. Thus we have the possibility. then. This constitutes further evidence for the possibility that standard quantum mechanics is dealing with ensembles of particle orbits with zbw periodicity. That is especially clear in the Schroedinger theory where spin is ignored but the complex phase factor is essential. The singlevaluedness of the wave function implies that the orbital ii"equency is a harmonic of the zbw ii"equency in stationary state". and 8"cp describes the zbw energy and momentum. Farrelly. but from the zbw perspective. This opens the possibility of zbw resonance as a fundamental explanatory principle in quantum mechanics. The observed Coulomb and magnetic dipole fields of the electron are averages of this field over times much longer than a zbw period. For that implies that the electron must be the source of a (nonradiating) electromagnetic field which fluctuates with the zbw frequency.). D. To explain the zbw we must go beyond the Dirac theory to discover new physical mechanisms such as the fluctuating "Quantum Force" proposed in the preceding paragraph. the phase cp In describes the zbw. The Dirac theory. Raines & J. since it is linked to stationary state conditions. Milligan. However. the challenge of finding zbw explanations for all the familiar phenomena of quantum mechanics. Further support for the zbw interpretation comes from recent successes of semiclassical mechanics in molecular dynamics and electronic structure (Ref. In the Schroedinger approximation the factor Ui. Skelton. Celestial Mechanics on a Macroscopic Scale. The Pauli principle may be a consequence of zbw resonance between electron pairs. J.

1. this was done without realizing that the imaginary unit i which generates E-M gauge transformations in the Dirac theory is a spacelike bivector identified with the electron spin. but that must be deferred to another day. the Dirac current is invariant under a 4-parameter group of gauge transformations on the wave function: 1/) ----> 1jJG. (8. 8 parameters are needed to specify the wave function 1/) uniquely. This exhibits explicitly the SU(2)xU(1) structure of the gauge group. this constraint can be easily satisfied in the following way: The Dirac current 1/)101/) is a timelike vector field.mass.13) where Bs is the self-magnetic field presumed to be the origin of the free particle rl. This is a suggestive starting point for a zbw approach to quantum electrodynamics.3) where UU = 1. This opens up possibilities for integrating the zitterbewegung idea with electroweak theory. Thus.3) is a duality factor exactly like the one parametrized by . 9. the in variance group of the Dirac CUITent c~n be identified wii. Therefore. That has been done in Ref. However. the E.§ =1/)13'ljJ invariant is characterized by the additional condition (9. = S· rl = J.1) = G(x) is an even multivector satisfying (9. as expressed by writing m. Note also that the U(l) factor in (9. the generators of the larger E-W group which include it must have related spacetime interpretations. so only 4 parameters are needed to specify it. 10. Evidently that would obviate the need for including Higgs bosons in the theory.W gauge group. where G (9. it may be that the fundamental physical role of /3 is to serve as a gauge parameter in electroweak theory. 63 . Thus. The STA formulation of the Dirac theory has indubitable implications for the Weinberg-Salam CW-S) theory of electroweak interactions. the Dirac equation is not invariant under the entire E-\V gauge group {G}. The very form of the important equation (7. This fact forces a strong geometrical constraint on the W-S theory: Since i has a spacetime interpretation.h.4) The E-M gauge transformations belong to this subgroup. Of course.6) suggests that the electron's mass may be a consequence of magnetic self-interaction. • Bs .2) It follows that (9. The subgroup which leaves the spin density p. but it is easily generalized to one that is by introducing a suitable "gauge invariant derivative" in the standard way. The VV-Stheory generalizes the electromagnetic (E-M) gauge group to the electroweak (E-W) gauge group SU(2)xU(1). where the \Veinberg-Salam model is completely reformulated in terms of STA with the E-VVgauge group defined as above.8/2 in the invariant decomposition of the Dirac wave function (5. ELECTROWEAK INTERACTIONS. since the zitterbewegung provides an alternative mechanism to account for the electron mass. Remarkably.16). However.

Second. S - cAv . however. It follows that the spin S in (10. The analysis has been developed progressively on three levels: reformulation. This has the great advantage over variants of the Copenhagen interpretation of being grounded in the Dirac theory. (10. to see how far we have progressed toward the objective.12)R.3) where [)vR = ~nvR.10. MODIFICATIONS.2) (3. the complex phase factor is interpreted as a direct representation of the zitterbewegung itself.3) attributes energy-momentum to this motion. more coherent and complete interpretation of the theory. REFORMULATION. there is a theoretical challenge to see how far we can go in providing zitterbewegung interpretations for the standard results of quantum mechanics and even quantum electrodynamics. B. Thus it opens up the possibility of a new 64 . REINTERPRETATION. Thus.2) is the angular momentum of the zitterbewegung. the zitterbewegung interpretation of the phase factor carries over to Schroedinger theory and so suggests a reinterpretation of quantum mechanics generally. a physical explanation is given for the appearance of complex numbers in quantum mechanics. reinterpretation and modification. In particular. First. Finally. and (10. By making the geometric structure of the theory explicit. This approach has the great formal advantage of providing the entire rotor R with a kinematical interpretation. vVehave seen that reformulation of the Dirac theory in terms of STA eliminated superfluous degrees of freedom in the Dirac algebra and reveals a hidden geometrical structure in the Dirac equation and its solutions. and it may be the source of an electromagnetic field which fluctuates with the zitterbewegung frequency. A. there is a challenge to probe the zitterbewegung experimentally to see if it can be established as a "literally real" phenomenon.) The factor in in the Dirac equation is a spacelike bivector related to the spin by s= ~R(in)R.) The Dirac wave function has the invariant decomposition (10. the zitterbewegung interpretation presents us with an array of challenges. Let us take stock. C. now. The new zitterbewegung interpretation is imposed on the Dirac theory simply by identifying the electron velocity with the lightlike vector u. If indeed the zitterbewegung is physically real it is probably a consequence of electromagnetic or electroweak self-interaction. there is a challenge to see if the zitterbewegung can lead us beyond present quantum mechanics to deeper physical insights. (10. The general helical character of the zitterbewegung is completely determined by the Dirac equation without further assumption.1) (2. they suggest a new. Moreover.) The electron energy-momentum Pv is related to the spin by Pv = nv . Above all. These results are mathematical facts inherent in the original Dirac theory. The main results are: (l. The objective of this work has been to understand what makes quantum mechanics so successful by analyzing the Dirac theory. CONCLUSIONS. = RCro .

Ph'ys. (October. 18. (Reidel Publ. Crawford.R. Am. SIAM J. J. 17.).A. 21. A Spinor Approach to Gravitational Motion and Precession. 1966). Pbys. 65 . 1013-1028 (1971). Am. Space-Time Structure of Weak and Electromagnetic Interactions.IVI. J. Quantum for Classical Mechanics. 9. 13. H. 14-144 (1971). Pbye. D.S. 11. D. Real Spinor Fields. J. Dordrccht. Hestenes. Quantum Mechanics. 26. Co.. J. Of course. 1986). Co. 15. 893-905 (1973). Chisholm & A. D. On the algebra of Dirac bispinor densities. D. Pauli and Schroedinger Theories. 10. Oxford U. D. In. Dordrecht/ Mechanics. Int. J. Operators and Complex Numbers in the Dirac Theory. Clifford Algebra to Geometric Calculus. P. Dordrecht /Boston (1990). P. Hestenes. such possibilities cannot be explored theoretically without going beyond the Dirac theory. 12. McGraw-Hill. 39. J. In Clifford Algebras and their Applications in Mathematical Physics. On Decoupling Probability from Kinematics in Quantum Mechanics. Hestenes and G.) Kluwcr. Ptiys. Hestenes. 321-346. Hestenes. 12.Reading of Photons and Electrons. Obscrvablcs. Reidel Publ. 15. Local Observables in the Dirac Theory. Jauch and F. 8..approach to the self-interaction problem and actually explaining the phenomenon of quantization rather assuming it. Phys. 16. J. J. Ph'ys. Moses. Phys. Hestenes. Math.Y. Phys.E.. Math. 7.1778-1786 (1974). Hestenes. 798-808 (1967). 15. Math. J. Fougere (Ed. J.. Proper Dynamics of a Rigid Point Particle. Gurtler and D. Hestenes. Relativistic 22. D. G Reidel Publ.. P. 573-583 (1975). Clifford Algebra and the Interpretation of Quantum Mechanics. Spinors and Complex Numbers in Classical and Quantum Physics. 153-168(1982). Bjorken and S. Found. D. Phys./Boston. Phys. K. Rohrlich. 20. Quantum Mechanics from Self-Interaction. 4th edition (1958). 19. Hestenes. p. 23. 1990). 5. J. London. 1439-1441 (1985). Sobczyk. D. Appl. Space-Time 2. R. 25. D.(1964). Addison-·Wesley. Hestenes. 14. 63-87 (1985). Common (cds. Found. Hestenes. G. J. Math. 15. D. D. J. Phy«. D. Math. Vectors. Math.. Co.1768-1777 (1974). Maximum Entropy and Bayesian Methods. 6. 14. Algebra (Gordon & Breach. Dartmouth College 1989. D. London. Hestenes. Hestenes. 16. Thea. 589-598 (1986). Proper Particle Mechanics. 21. The Zitterbewegung Interpretation of Quantum Mechanics. D. (1955). Hestenes. J. 4.339-415 (1979). D. 556-572 (1975). The Theory Mass. N.Dirac. Spin and Uncertainty in the Interpretation of Quantum Mechanics. Hestenes. Phye. Dordrecht/Boston (1984). Drell. REFERENCES 1. Math. Hestenes. 47. D. Phye. A Unified Language for Mathematics and Physics. Found. Press. Hestenes. Consistency in the formulation of the Dirac. New Foundations Boston (1985). 16. Math. 8.

the differential of the k-vector field 0. 32. Schroedinger. Pby». R. Pbye. B. p.418 (1930). Brown & T.R. 50. Localized States for Elementary Systems. Sitzungb. 15-28 (1979). Quantum Electrodynamics. 31. 1987). Quantum Interference and the Quantum Potential. Ptiy«. 24.. Kruger. 26. This Appendix formulates general transformation laws for fields on spacetime and applies the results to establish Poincare invariance of the field equations. 35. Huang. D23. 47-49.24. (1961). Preuss. J. Phys. Dcwdncy. H.2/\ . Causal particle trajectories and the interpretation of quantum mechanics. Peat (cds. a2. from Microscopic to Macroscopic Levels.2) where the derivative is evaluated at x = f-1(X'). Nuovo Cimento 52B. Akad. /\ ak) = (fal) 4 /\ (fa2) /\ . S. Am . 2454 (1981). In Quantum Implications. Bohm & B. 21. It has a natural extension from vector fields to arbitrary multivcctor fields. Kl. P.-P. Barut and A. J. 400-406 (1949). Feynman. Gull. Wigner. Rev. P. Letters 55. ak. 797 (1949). C. Let f be a transformation of a 4-dimensional region (or manifold) R = {x} onto a region R. Benjamin..3) By linearity this determines the differential for any multi vector field M IM = M (x): (A. /\ (fak) . Kypriandis. . Hiley. private communication. J.. It induces a transformation f of a vector field a = a(x) on R into a vector field a' = a'(:r') on R' defined by f :a ---+ a' = Ia = a . Zitterbewegung and the internal geometry of the electron. The transformation law for spinor fields is shown to rest on a convention which can be chosen to make it identical with the transformation law for tensor fields. Rev. L. for vector fields a1. (Routledge and Kegan Paul. 25. Pbye... C. Thus. Phys.1.. M. Philippidis. 705 (1964). D. Holland & A. 34. 29.G. P. . Dewdney and B. (A. Relativistic time dilatation of bound muons and the Lorentz invariance of charge. 27. Unbroken Quantum Realism. C. O. D. Hiley & F. Hiley.T. S. V f . Kibble. A parenthesis has been dropped in writing fa in recognition that f is a linear operator on tangent vectors. London. Newton and E. Bracken.. (A. Wiss. 66 . (A.-Mat11. 251-254 (1982).). 28. Mod. 2511 (1985). 47. Am. N. The transformation f is called the differential of f. A 133. Rev. Silberman.1 /\ 0. D. P.Y. On the Zitterbewegung of the Electron.. (to be published).4) = L I(M)(k) k=O . J. K.l) A transformation is understood to be a differentiable invertible mapping. 30. E.' = {:r'} in a spacetime. A. Rev. Vigier. the pointwise transformation is thus f :x ---+ x' = f (x) . Phy!).W. APPENDIX A: Transformations and Invariants.2 /\ . 33. . The Klein Paradox. /\ ak is defined by f(al /\ 0. T.

67 changes the field . 0 To exhibit the e-dependence of the argument explicitly. 0 0 This gives us immediately the useful expression det f = 1+ v 0 E (A. we obtain (A. for any unit vector Ii. (A.12) Note that k = 0 is not included in the sum since the scalar part of 1\[ is invariant. we make a Taylor expansion of the argument and keep only first order terms to get M'(x) = M(:r) - Eo V M(:r) + [M(:r) 0 v] /\ E. except when E is a null vector. using the duality relations (l.13) fi =i + (i v) /\ E = i(l + v E) . E a is a small quantity.8) It is understood that. k=O (A.1l) MoV= Lf(M)(k)V.M at a designated . Equation (A.14) for the Jacobian of f.7) is the Jacobian of f. An extensive treatment Ref. therefore.where it is understood that (J1\1)(0) = (1\1)(0)' (A.1l) determines a new function M'(x') = M'(:r + E) = M + [M(:r) v] /\ E. for a bivector field a /\ b. As a significant example.b). This is equivalent to the condition that E2 is small. 0 0 0 (A. we have f(a /\ b) = a /\ b + (a '\h) /\ b + a /\ (b '\h) = a /\ b + [(a /\ b) V] /\ E. let us calculate the form of the differential for an arbitrary infinitesimal transformation f(x) = x + E(X). As an application of general interest. (A. where 4 (A.I0) This result generalizes easily to the differential for an arbitrary multi vector field: fM=M+(Mov)/\E.16 a.9) Neglecting second order terms. we apply (A.1l) to the pscudoscalar z and. 15.6) fi where dct f = idet f' = i-1 fi = -ifi (A. (A. IS) This tells us explicitly how the infinitesimal transformation point x. For a vector field the corresponding induced transformation is fa=aoV(X+E) =a+aVoE. 0 (A.5) which is to say that every scalar field is an invariant of of the differential on differentiable manifolds is given in The differential of the unit pseudoscalar is given by f and hence of f.

for Maxwell's equation (4. it can be written in the canonical form (A. From (A.4) we have v'P' = (jv)(jF) = (RVR)(RFR) = RvFR = RJR = . so no argument at all is needed to establish its observer independence. That is clearly not the correct interpretation here.r2. so. Translation invariance implies that spacetime is homogeneous in sense that the same laws of physics are the same at every spacetime point.22).3).cA'¢!)R = R(m'¢rYo)R = ml/.16) l(X) = R.5) can be established in the same way as that of Maxwell's equation.IS) By virtue of (1.r1)2 between every pair of spacetime points .2) we find immediately (A. Indeed. v'¢/i'n . For example.IS) show that they are in the present case. It describes a physical property of the Minkowski model of spacetime. (A.1' (A. /\ b) = R(a /\ b)R . for no reference system has even been mentioned either in the formulation of Maxwell's equation or of its induced transformation (A. (A.r1 and .24).17) For the product of two vector fields this gives the simple result a'b' = (ja)(jb) = RabR . the relation of field F to current J is a Poincare invariant. Poincare transformations are commonly interpreted as relations among different inertial reference systems or observers. According to (A.21) M' = 1M = RM R . From this the Poincare invariance of the basic equations of physics is easily established. Maxwell's equation V F = .19) and (A. The inner product and the geometric product are not generally invariant. according to (1. = 1.3) this can be decomposed into a scalar part a'· b' and a bivector part a' /\ b' = a· b (A.rR +c. since the matter is frequently muddled in the literature. Similarly.eA'¢/ = (RvR) (R'l/JR)(Ri'R)n . The Poincare invariance expressed by (A. The Poincare group is the group of transformations on spacetime which leave invariant the "interval" (X2.1 is manifestly independent of any coordinate system. Every such transformation is the composite of a Lorentz rotation and a translation. however. 68 (A. We are concerned here only with the Restricted Poincare Group for physical reasons discussed below. It is the formal assertion that the laws of physics are the same everywherewhen. Thus.e(RAR)(R¢!R) = R(v#n . Lorentz rotation invariance implies that spacetime is isotropic in the sense that the laws of physics do not favor any particular timelike or spacclikc directions.20) = 1(0.b. Poincare invariance thus provides the theoretical basis for comparing the results of physical experiments and observations made at different times and places.Now let us turn to the question of Poincare invariance of the equations of physics. This is the subgroup of Poincare transformations continuously connected to the identity. It follows that for an arbitrary multivector field NI the transformation law is simply (A. where c is a constant vector and R is a constant rotor with RR the differential a' = fa = RaR. the outer product is an invariant of the differential for any transformation.23) . The physical significance of Poincare invariance deserves some comment.22) Thus.22) should be interpreted as an equivalence of spacetime points rather than an equivalence of observers.19) (A.'. The Poincare invariance of the Dirac equation (5.

in the conventional matrix representation (5.2I) and hence the more general transformation law (A._.26) Nevertheless. (A. Though (A.11).l can be made hermitian by the artifice of introducing a unit imaginary factor ii.25) = R1j.24) and (A. our analysis of the Dirac theory in Part II reveals that the success of this formal procedure should be attributed to the physical interpretation of the Dirac wave function rather than a general physical significance of hermitian operators.25) is therefore a matter of convention. (A.I5) and/or (A.Note that the transformation law for the spinor wave function 1j/ 1/' has been taken to be (A. since they play such a prominent role in the conventional approach to relativistic quantum theory. For an infinitesimal Lorentz rotation. the usual transformation or.I6) reduces to x' = (1 + ~B)x(I Hence. we take R = 1 and E = c in (A.2)..Ii in accordance with (A. as established by (5. transformation law (A.2I) gives M'(x) = [1 . so the translation operators in quantum mechanics are usually defined to be i' w These operators are then identified with momentum operators. It is of interest to consider briefly the infinitesimal Poincare transformations.~B) ::::: x + B· x.30) in (A. the result is for a spinor '1/) subject to the one-sided 1j/(x) = [1.31) This is expressed in a more conventional form by expanding B with respect to a basis to get (A. 'lI1 = R'lI (A.e!B ~ 1 + IB -' ". 2' where B is an infinitesimal bivector. Writing c.B· (x /\ '7) + Bx ]M(x). IS) reduces to M'(:r) = M(:r .27) This applies equally to the electromagnetic field and the Dirac wave function.7) through (5.29) = B· x in (A.24) law for a Dirac spinor is (A. because all observables are bilinear functions of the wave function.25) is simpler. Indeed.24) and (A.2I). However.27).c) = (1 - c- '7)M(:r). the transformation laws (A. For an infinitesimal translation. (A. so (A. However. (A. (A.24) can be transformed away at will.25) are physically equivalent. we recognized the a'l as generators of translations.32) 69 .30) where B X JYf is the commutator product. Then (A.. E (A. It is noted that the a.4). the factor ii on the right of (A. Alternatively. In the conventional formulation.B· (:r /\ '7) + ~B]1/{r). we take c = 0 in (A.'7 = d'al' .24) has the advantage of conformity with (A.28) .I6). The choice between the transformation laws (A.I6) and a R .

Let E = £(x) be the Lagrangian for some field on spacetime. (A.33) are the usual "angular momentum operators" for a Dirac particle. angular momentum operators for the electromagnetic field can be read off (A. to first order we have (B.G) Now for a given L.)= 'I//(x) -lj{1. The Lagrange approach has the advantage of directly relating equations of motion to conservation laws. (B. we adopt the (nonunique) Lagrangian (B..) in the functional form of the wave function which vanishes on the boundary of 'R-.6) can be put in the form Similarly.14) gives us (B.3) x' + E(X). For the Dirac electron.2) A general variation of the action involves both a change in the functional form of £ and an infinitesimal displacement of R producing a new action A' = A For an infinitesimal displacement x ---'t + 8A = r in! £'(x') I d4x'l. Thus.6) involves 70 .5). APPENDIX B: LAGRANGIAN FORMULATION This appendix is concerned with the Lagrangian formulation of the Dirac theory. the STA formulation is sufficiently novel to merit one more version." We derive the Dirac equation by requiring 8A = 0 for an arbitrary variation 8'1/)(.. The derivation employs the scalar-part properties CM) = (M) and (MN) = (NiH).1. the variation of the second term in (B. using (8'1/)) ~ = 8'0 the variation of the last term in (A. writing E' = E + 8£. In this case the boundary is fixed and E = 0 in (B. Though this approach to the Dirac theory has been discussed many times in the literature.6) where (. ) = (.30).l) where the oriented "volume element" for the region is the pseudoscalar (B. both field equations and conservation laws can be derived by requiring the invariance condition 8A = 0 subject to various constraints.. In a similar way.4) Hence. The associated action integral over any region R is (B.where the (A. ')(0) means "scalar part.

.G).21) we have (B. we arrive at (B.12) Inserting this into (B. + 'lj!(x') To first order in small quantities {. However.*A must be added. To evaluate (B.7) must be included and an additional term due to {.6).IO) Applying the same argument to the integrand of (B. This is .'lj!(x) =¢'(x') . we can express the general conservation law in the form (B.8) except that the perfect divergence term in (A. CONSERVATION LAWS Conservation Laws are derived by requiring invariance of the action under infinitesimal displacements preserving the field equations.P)) . we have t5£ Thus.'lj!(x).14) 71 . because t5'lj! vanishes on the boundary.ll) for the electron Lagrangian (B. we note that {j*£ will have the same form as (B.P only if the Dirac equation (5.5) is satisfied. we use t5(vl/') = V (t5'lj}) and ((Vt5l/!)r3.(t5l/!i'Y3(Vl/!)~) = ((V'ljJ)i'Y3t5.j.*'lj!(x') = {.j. since we require that the Dirac equation be satisfied.)(1) . = 0 for any choice of the region R only if t5*£ +V . t5A + £v . From the definition (6. E = t5*£ + E' V £ + £v .9) due to a change in the value of'ljJ and a part due to the shift easily done by writing S¢ = 'lj/(x') .J. E. (B.5).*'lj)(x) and we have (B. (c£) =0.ll).x in the argument.6). the result is simply (B..for evaluate the variation of the first term in (B. (B..) = (V(Sl/!i'Y3.13) It will be helpful to reformulate this in terms of the energy-momentum operators P. For performing the calculation it is convenient to decompose t5'lj! into a part (B.) + v· (t5'lj)i'Y3.8) This vanishes for all values of the arbitrary even multi vector t5.'lj!(x') E = x' . Thus.7) The last term here does not contribute to t5A in (B.ll) This is a Conservation Law for specified E.

Thus.17) But we have already observed in (6.16) which relates one term on the left of (B.6) the term ~(F2). Til = B . momentum conservation is a consequence of the homogeneity of spacetime. note the following: = A • [J • V (B . Hence.13) in the form (B. A) = (B . and t5A = 0. energy- B. For the sake of completeness. 151/)= 0.15) Consequently. = (Bi(~n'¢){3.16) can be written all(E' AJM) = J. this implies the energy-momentum conservation law (6. where F = V /\ A.24) or Table II. Inserting these results into (B. Lorentz Invariance. t5A B· A by (A. this implies the angular momentum conservation law (6. V(E' A) . (B. angular momentum conservation is a consequence of the isotropy of spacetime. we obtain {I') is the spin angular momentum tensor of (6. we note that the last term in (B. Til (rMb¢!'1{3n.23) 72 . and 151/) (1/2)B1/) by (A. x) . In consequence.31).19) where F = V /\ A. A) = .32). For an infinitesimal Lorentz rotation. TI') .22) Since B is an arbitrary bivcctor.J • (B .. Now it is a simple matter to assess the implications of requiring Poincare invariance. (B.hll. A. with the result (B.) where 5" = i(s /\ (B.13) to TM. E = B· x by (B.So from the definition of the energy-momentum tensor TM in (6.34).. (x .27). (pSI'). The electromagnetic part of the Lagrangian is then (B.31) that this vanishes in consequence of the Dirac equation.{. Hence.33). (B.20) reduces to For an infinitesimal translation E is constant. :r) 1 E• = BxA = = (B • J) . The Lagrangian (B.21) Since E is arbitrary.20). and we can put (B. Thus. we obtain (B.6) can also be expressed in terms of Til. Finally.{. Translation Invariance. we note that a complete Lagrangian for electron and E-M fields together is obtained by adding to the Dirac Lagrangian (B.J • (15 .) = B.20) This is the desired final form of the general conservation law. with the help of the identity (B. A = .33)..18) where J = cljJ{o1/) is the Dirac charge current. (B.

27) As before.29) where the E-M spin tensor S{' = Se( ')'/') is given by (B. Thus. Also note that when they are added to the corresponding equations (6.28) And Lorentz invariance yields the angular momentum conservation law (B. Since 8A is arbitrary.34) for the electron. F) + A .24) with 8*A = 8A . where the reverse accents serve to indicate which functions are differentiated by into (B. E(F2) .25) = E .From this the E-M field equation can be derived by the general variational principle.A .28) and (B. 73 .VA = V(FnA) . Inserting this (B.25). V A) - hi' .27) and (6.E' V A and 8*J = 8J . Thus.~n(F2) + A(nv F). we obtain Up [(YI"E . using (B.26).29) can be obtained by direct differentiation of (B. A) V .V F + J) = O. E Let us define the canonical energy-momentum tensor TJ: = TcbP) by Te(n) = -~FnF+n. the internal forces and torques cancel.26) v. V E) .P.(YyI18A) 1 (B. (bJ) . we get the Conservation Law in the form (B.23) into the general Conservation Law (B. translation invariance yields the energy-momentum conseration law (B. Conservation laws for the electromagnetic field can be obtained by inserting (B. we note that (B. both (B. (J .24) The last term vanishes for 8A = 0 on the boundary. this implies Maxwell's equation V F = J. so we have 8£EM = 8A· (.(J .E' V J. (J .30) of course.ll).

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd