Spinors in Relativity
Jörg Frauendiener
December 13, 2018
Department of Mathematics and Statistics University of Otago Since this script is available online I would appreciate any feedback from whoever reads it. Please send comments, suggestions, corrections to
[email protected]. Contents
1 Introduction and Literature 4
2 The geometry of Minkowski space-time 8 2.1 Affine space ...... 8 2.2 Metric space ...... 10 2.3 Basic notions ...... 11
3 Spin transformations 15 3.1 Representation of the sphere ...... 15 3.2 Null directions and the celestial sphere ...... 17 3.3 Lorentz and spin transformations ...... 18 3.4 Lorentz transformations: a closer look ...... 21
4 Spinorial objects and spin vectors 23 4.1 Null flags ...... 23 4.2 Spin vectors ...... 27 4.3 Some global considerations ...... 28
5 Spinor algebra 30 5.1 Foundations of tensors ...... 30 5.2 Spinors ...... 35 5.2.1 ε spinors ...... 36 5.2.2 Complex conjugate spinors ...... 37 5.2.3 Spin frames ...... 39
6 4-vectors as spinors 40 6.1 Spin frames and their tetrads ...... 40 6.2 Van-der-Waerden symbols ...... 42 6.3 Clifford algebra, Dirac spinors, Weyl spinors ...... 46 6.4 Null vectors, null flags, etc...... 50
3 1 Introduction and Literature
Einstein’s theory of special relativity is of fundamental importance in today’s mod- ern Physics. It is not a theory of interaction, i.e., describing a fundamental force such as for instance electrodynamics. Rather it is a kinematical theory which describes the framework within which Physics takes place. It takes the place of classical Newtonian Physics which was shaped by Galilei’s principle of relativity. This principle postulates that all reference frames which are related by Galilei transformations are equivalent for the description of physical phenomena. This stipulation distinguishes a class of pre- ferred reference frames, called inertial frames. It was expected that all laws of Nature “look the same” with respect to each of these inertial frames, i.e., that they are invariant under Galilei transformations. However, after the realisation that Maxwell’s equations of electrodynamics are not Galilei invariant, it was Einstein to dissolved the ensuing dilemma (parts of Physics being Galilei invariant, others not) by postulating a different principle of relativity. According to his principle of relativity not the Galilei transfor- mations map inertial frames to inertial frames but it is Poincaré transformations, also known as inhomogeneous Lorentz transformations.
The space-time continuum of relativistic (quantum-) mechanics is the Minkowski space- time, an affine space which is equipped with a bilinear form of signature −2. The invari- ance group of this space is the Poincaré group, the group of inhomogeneous Lorentz transformations. The Lorentz group itself is the subgroup of the Poincaré group which consists of transformations which keep one selected point fixed. It can be regarded as the invariance group of the Minkowski vector space of position vectors based on the selected point. The importance of the Poincaré group is highlighted by the fact that elementary particles are classified in relativistic quantum mechanics according to its irreducible representations.
The fact that the Poincaré group is intimately connected to the global structure of Minkowski space (which can even be defined as a homogeneous space on which the Poincaré group acts) already suggests that there might be difficulties in the transition to Einstein’s theory of gravity. Now the space-time is no longer an affine space but a more general (curved) manifold whose curvature is determined by the matter contents in the space-time. The only commonality with Minkowski space is the fact that at every point there is attached a Minkowski vector space which also carries a bilinear form of signature −2. Accordingly, the importance of the Poincaré group evaporates because only in the rarest cases will there exists a global invariance group for the space-time which in addition will not agree with the Poincaré group. Furthermore, in this general case the notion of an elementary particle gets blurred. Together with other difficulties
4 this is a manifestation of the current incompatibility of quantum theory and general rel- ativity. However, the Lorentz group does survive in the transition from flat to curved space. It remains the invariance group of the Minkowski vector space and, hence, only acts at each point individually. It is the truly fundamental group of relativistic physics and its representations should be of physical relevance in all aspects of Physics. The representations of the Lorentz group are treated at length in [4].
The best known representation of the Lorentz group is its defining representation in the 4-dimensional Minkowski vector space. It is the basis for the usual tensorial repre- sentations and the mathematical process leads naturally to the development of tensor algebra. The Lorentz group is a semi-simple Lie group whose finite dimensional rep- resentations are fully reducible. Hence, it is enough to find and classify the irreducible representations. Among those there are two distinguished ones: the two 2-dimensional representations are fundamental representations, i.e., all representations of the Lorentz group can be obtained as appropriate tensor products of only those two. The vectors of these representation spaces are the primed and unprimed spin-vectors and the ten- sor algebra built upon the fundamental representations is called spinor algebra. This means that spin-vectors can be used to construct all other tensors.
The 2-dimensionality of the spin-vectors makes calculating with them quite easy. This is a peculiarity of the four dimensions of the world we live in. In higher dimensions one can find analogous spinor representations related to the corresponding invariance groups. However, their dimension grows exponentially with the dimension of the space-time so that already for space-time dimension 8 the corresponding spinor rep- resentation is also 8-dimensional. Only for space-time dimensions 3 and 4 can we ex- pect to gain simplifications in the study of space-time geometry using the calculus of spinors.
The theory of spinors was already developed and applied at the beginning of the last century. The general notion of a spinor was introduced by Cartan [1, 2] while van der Waerden [11] developed the calculus of two-component spinors. Dirac discovered the fundamental importance of spinors for relativistic physics when he found the equation the electron. With the work of Penrose [7] the calculus of spinors was firmly intro- duced into General Relativity. This formalism as well as the related Newman-Penrose (NP) formalism was very successfully applied in many different areas of gravitational theory, in particular in the treatment of asymptotically flat space-times, the analysis of gravitational waves and last but not least in the proof of the positive-energy-theorem of General Relativity, one of the most fundamental results in that area.
The two volumes by Penrose and Rindler [8] contain an impressive introduction into the theory of spinors and many application thereof. They also explain the relationship between spin-vectors and space-time geometry. In the book by Stewart [10] one can find the treatment of asymptotically flat space-times using the calculus of two-component spinors. Spinors are also discussed in the book by Naber [6].
5 Another good source for the mutual relations between spinors, relativity and quantum theory is the book by Sexl and Urbantke [9]. It discusses in particular the group theoret- ical aspects. Finally, it is worthwhile to point out a classic which does not deal explicitly with spinors but presents a very good treatment of the foundations of General Relativ- ity: the book by Hermann Weyl [12] is warmly recommended. It appeared first in 1918 and has been published now in its 7th edition. The following course will provide an introduction into the calculus of two-component spinor-vectors. The aim of the course is to reduce the timidity towards and overcome the hesitation to apply the formalism which may seem daunting at the first encounter. The first chapters deal with the fundamental properties of Minkowski geometry. How- ever, the point of view is a bit different from other approaches. We regard the geometry of light rays as fundamental. The “celestial sphere” of past directed light rays is intro- duced and with its aid the null flags are defined. These are the geometric imprints of spin-vectors. Then we adopt a more abstract point of view and introduce the spinor algebra. After a chapter on the main tools of spinor calculus (principal null directions, dualisation, symmetry operations) the results are transformed to manifolds. The last three chapters are devoted to applications, the classification of the Weyl tensor, the NP formalism, and finally a discussion of some spinor equations. Except for some parts of the last chapter the whole material is standard and can be found in the cited litera- ture.
6 Bibliography
[1] É. CARTAN, Les groupes projectifs qui ne laissent invariante aucune multiplicité plane, Bull. Soc. Math. France, 41 (1913), pp. 53–96. [2] , The Theory of Spinors, Hermann, Paris, 1966.
[3] T. FRIEDRICH, Dirac-Operatoren in der Riemannschen Geometrie, Verlag Vieweg, Braunschweig, 1997.
[4] I. M. GELFAND,R.A.MINLOS, UND Z. Y. SHAPIRO, Representations of the Rotation and Lorentz Group and Their Applications, Pergamon Press, Oxford, 1963. [5] H. B. Lawson und M.-L. Michelsohn, Spin Geometry, Princeton University Press, 1989.
[6] G. L. NABER, The Geometry of Minkowski Spacetime, Springer-Verlag, Berlin, 1992.
[7] R. PENROSE, A spinor approach to general relativity, Ann. Phys., 10 (1960), pp. 171– 201.
[8] R. PENROSEUND W. RINDLER, Spinors and Spacetime, vol. 1,2, Cambridge Univer- sity Press, 1984, 1986.
[9] R. SEXLUND H.URBANTKE, Relativität, Gruppen, Teilchen, Springer-Verlag, Wien, 3. Auflage, 1992.
[10] J. STEWART, Advanced general relativity, Cambridge University Press, Cambridge, 1990.
[11] B. VAN DER WAERDEN, Spinoranalyse, Nachr. Akad. Wiss. Götting. Math.-Physik. Kl., (1929), pp. 100–109.
[12] H. WEYL, Raum, Zeit, Materie, Springer-Verlag, Berlin, 7. Auflage., 1988.
7 2 The geometry of Minkowski space-time
In this first chapter we will present a short representation of the geometrical founda- tions of special relativity. This theory lives in the simplest model of a space-time contin- Minkowski space-time uum, the so called Minkowski space-time, named after HERMANN MINKOWSKI who was the first to describe its geometric properties.1
The structure of the world of special relativity follows from Einstein’s empirical princi- ple of relativity which, mathematically formulated, is formulated as follows
The world is a 4-dimensional affine space equipped with an indefinite quadratic form with signature −2 which defines a metric.
This formulation is taken from the book Space - Time - Matter by HERMANN WEYL, which despite of its age (first published in 1918) even today provides a clear and trans- parent presentation of the mathematical and physical notions underlying both theories of relativity.
2.1 Affine space
What is the contents of the principle mentioned above? We first remind ourselves about the notion of an affine space. It is well known that there are two categories of “things” point in affine geometry — points and vectors (translations) — and relations between them. vector An affine space M can be characterised by the following three requirements
1. Vectors: The set of vectors is a real vector space V of dimension n < ∞.
2. Points: −→ a) For any two points P, Q there exists exactly one vector, denoted by PQ.
b) If u is any vector and P is any point then there exists a unique point Q, so −→ that u = PQ holds. −→ −→ −→ c) If u = PQ and v = QR, then PR = u + v.
1H. Minkowski, Raum und Zeit, Talk at 80. Versammlung Deutscher Naturforscher und Ärzte zu Cöln on 21. September 1908
8 Strictly speaking we need to distinguish between the space of points and the space of vectors. This is often suppressed. The relationship between these two spaces is charac- −→ terised by the property (2). Every vector PQ may be regarded as a translation which sends the point P to the point Q. The main property of affine space is its homogeneity: homogeneity no point is preferred over any other. This is contrary to a vector space where the zero- vector is different from all others since it is the only one with vanishing components in every basis. At every point O the vector space V can be represented as the set of −→ position vectors OP. In this sense, we can regard V to be attached to every point of the position vector affine space.
An affine coordinate system C consists of a point O together with a basis (e1,..., en) of coordinate system V. Each vector v can be expressed in exactly one way as a linear combination
i v = v ei. (2.1.1)
(We use the Einstein summation convention throughout). The numbers vi are the co- ordinates of V with respect to C. The coordinates of a point P with respect to cC are coordinates −→ defined as the coordinates of the vector OP.
0 0 0 0 If C = (O ; e1,..., en) is another affine coordinate system then equations of the form
i 0 ek = A kei, (2.1.2)
i hold, where the number A k form a matrix with non-vanishing determinant. For the coordinates of the vector v with respect to C0 we obtain the expressions
0i i k v = A kv . (2.1.3)
When xi are the coordinates of a point P with respect to C then the coordinates of the −→ −−→ −→ same point with respect to C0 are obtained from the relation O0P = O0O + OP as
0i i k i x = A kx + α . (2.1.4)
Here, the αi are the coordinates of O with respect to C0. We see again that there is indeed a difference between points and vectors, they behave differently under the change of coordinate system: the coordinates of points transform inhomogeneously, while those of vectors are subject to homogeneous transformations.
The formulae (2.1.4) indicate to transition from one coordinate system to another. This is the passive point of view of a transformation in which the description of an object is changed. But there is another interpretation of the same formulae: according to the ac- tive point of view they define an affine transformation. This is a bijective map of affine affine transformation space M onto itself which induces a bijective linear transformation on the vector space V. For, if xi are the coordinates of P with respect to C then (2.1.4) defines the coordi- nates of another point P0 with respect to the same coordinate system C. The assignment
9 P 7→ P0 defined in this way gives a map f : M → M and a related map fˆ : V → V. It is easy to see that this map has the property that −→ −−−−−−→ v = PQ ⇐⇒ fˆ(v) = f (P) f (Q).
Indeed, every map with this property can be expressed by (2.1.4) with respect to an affine coordinate system C. The set of affine transformations form a group, the affine affine group group of dimension n. It is the invariance group of affine space M. invariance group In a sense, the affine geometry abstracts the measurement process. This is most easily seen in the 1-dimensional case. Measuring the distance between two points consists of comparing the segment between the points to a given standard ruler which is freely moving in space. In a sense this ruler corresponds to the vector between the endpoints of the standard segment. By extending and subdividing the standard ruler one can obtain the law of addition of vectors and multiplication of vectors with rational num- bers. The extension to real numbers is then a mathematical sophistication. In higher dimensions one can also orient the standard ruler in different directions. With the affine framework there is no criterion to select a particular class of standard rulers. Therefore, there is no “absolute” notion of length or distance in this geometry, only a “relative” one: for any two segments it is not possible to determine their length individually, one can only say by which factor they differ. To make such a statement it is not relevant which standard ruler is used to measure the two segments as long as it is the same for both. These considerations extends in an analogous way to the higher dimensional notions such as area and volume. In order to go from the affine to the metric geometry one needs an additional structure which allows one to assign to each segment between two points a unique length and in higher dimensions to any two vectors an angle.
2.2 Metric space
A metric space in an affine space M on whose vector space V there is defined a real scalar product. To any pair of vectors U, V is assigned a real number (U · V) in such a way that this map is linear in both arguments and symmetric, i.e., (U · V) = (V · U) regular for all U, V ∈ V. Finally, we require that this bilinear form is regular in the sense that U · V = 0 for all V ∈ V implies U = 0. After choosing a coordinate system C the bilinearity implies
i k i k (U · U) = (ei · ek)U U = gikU U . (2.2.1)
The number gik = (ei · ek) form a symmetric matrix with non-vanishing determinant. From (2.2.1) we see that the scalar product define a quadratic form of the coordinates of a vector.
10 Since the choice of a coordinate system is entirely arbitrary one can ask whether there are any particularly useful choices. Indeed, it is well known that one can always find a linear transformation of the coordinates Ui (i.e., a basis of V) of U so that
2 2 2 2 2 (U · U) = U1 + U2 + ··· + Ur − Ur+1 − · · · − Un (2.2.2) for some 1 ≤ r ≤ n holds. This means that one can always find a coordinate system so that the matrix of the gik has the diagonal form diag(1, . . . , 1, −1, . . . , −1). The signature signature | {z } | {z } p q of a quadratic form is the difference p − q between the number of positive and negative entries in this diagonal form. Together with the dimension n = p + q a regular bilinear form is uniquely fixed by its signature. All bilinear forms on a metric space with the same signature are equivalent and cannot be distinguished. The bilinear form on V allows us to define a distance between any two points of M. This is achieved by assigning to any two points P and Q their squared distance squared distance −→ −→ Φ(P, Q) = PQ · PQ. In space with a definite bilinear form (signature ±n) this expres- sion has always the same sign and one can take the square root without any problem. This yields the geometric distance between the points. In a similar way one can also define angles.
The existence of such a bilinear form, also called metric, restricts the invrariance group metric of the space. Now only those affine transformations are allowed which in addition leave the metric invariant. If the corresponding linear transformation is given by (2.1.3) then we need to require in addition that
l m glm A i A k = gik (2.2.3) holds. The affine group is reduced to the subgroup of isometries. isometries
A special example for a metric space is Euclidean space for which the signature is Euclidean space equal to the dimension of the space, and Lorentzian space with signature ±(n − 1). Lorentzian space In Euclidean spaces we have Euclidean geometry and the invariance group is called the Euclidean group. In Lorentz spaces we have Lorentzian geometry with the Poincaré group or inhomogeneous Lorentz group as invariance group. The Lorentz space with Poincaré group dimension 4 and metric signature ±2 is called Minkowski space-time. We will choose Lorentz group signature −2, i.e., the metric can be put into diagonal form diag(1, −1, −1, −1). The reason for this will become clear later. Minkowski space-time
2.3 Basic notions
First, we want to have a closer look at the Minkowski vector space V. It appears at several places in Relativity, not only the space of position vectors in Minkowski space- time but also as tangent space to a space-time manifold or as the space of 4-momenta.
11 As we have already pointed out the 4-dimensionality of space-time means nothing else than the existence of four linearly independent vectors (e0,..., e3). Occasionally, we will denote these also by (t, x, y, z) in accord with the meaning of the four dimensions. tetrad A basis in Minkowski vector space is also often called a tetrad. Each vector U can be written as a linear combination with respect to a basis
0 1 2 3 i U = U e0 + U e1 + U e2 + U e3 = U ei. (2.3.1)
0 1 2 3 contravariant The numbers U , U , U , U are the contravariant coordinates of U with respect to the tetrad (e0,..., e3).
Any two tetrads (ei), (fk) are related by an invertible linear map
i fk = A kei. (2.3.2)