Could not include higher spin geometry -# P contents
A quadratic -module is an -module equipped with a quadratic form: an -quadratic function on with values in .
Equivalently, it is the initial object in the category whose objects are pairs where is an associative unital -algebra, and is an -linear map satisfying for all , and whose morphisms are the associative -algebra maps such that .
Examples in low rank can be calculated easily. If is freely generated by a single element , with quadratic form , then . Note that the opposite sign convention is often used in the differential geometry literature, so one may see the assertion that the Clifford algebra of the real line with a positive definite metric is isomorphic to the complex numbers . Similarly, the Clifford algebra of a negative definite two dimensional real vector space is isomorphic to the (non-split) quaternions in our convention, but one may see the assertion that it is isomorphic to . Complexification removes the difference between positive definite and negative definite, and the two complexified algebras are isomorphic.
Let for projective of rank over , and the dual module. One can define the canonical quadratic form for and . In this case, . In general, the Clifford algebra arising from a nondegenerate form is flat-locally (on ) isomorphic to a matrix algebra (when rank is even) or a direct sum of two matrix algebras (when rank is odd).
If is a projective -module of rank , then independently of , the Clifford algebra is projective of rank , and is (noncanonically) isomorphic to as an -module equipped with a map from . The Clifford algebra is isomorphic to the exterior algebra (as algebras equipped with -module maps from ) if and only if .
If is the ring of smooth functions on a pseudo-Riemannian manifold , and is the -module of sections of the tangent bundle, then the metric endows with a quadratic structure, and one can form the Clifford algebra of the tangent bundle.
This is one of the incarnations of Bott periodicity.
When is projective of rank , each homogeneous piece is projective of rank . When is nondegenerate, the even part of the Clifford algebra is also flat-locally isomorphic to a matrix ring or a sum of two matrix rings.
One can view the Clifford algebra multiplication as a quantization of the commutative super algebra .
The Clifford group is the subgroup of elements for which twisted conjugation stabilizes the submodule . Here, twisted conjugation is defined by , where is the automorphism of induced by the map on . Since twisted conjugation by -stabilizing elements amounts to reflection , there is a canonical map , and the Clifford group is in fact a central extension of the orthogonal group by .
The Clifford group is made of homogeneous elements in the -grading, and the subgroup of even elements is a normal subgroup of index two. One also has a spinor norm on the Clifford group, defined by , where is the anti-involution of the Clifford algebra defined by opposite multiplication in the tensor algebra.
The restriction of the map to the Pin group may not be surjective, but it is for positive definite real vector spaces. The kernel is the group of elements of that square to 1. Similarly, the Spin group has a map to the special orthogonal group with kernel , but it may not be surjective in general.
One can use base change to define the groups given above as functors on commutative -algebras.
For nondegenerate quadratic forms on real vector spaces, spinors/spin representations are distinguished linear representations of Spin groups? that are not pulled back from the corresponding special orthogonal groups. In other words, the central element acts nontrivially. They can be realized as restrictions of representations of the even parts of Clifford algebras. Since even parts of Clifford algebras are (up to complexification) the sum of one or two matrix rings, their representation theory is quite simple.
The specific nature of spinor representations possible depends on the signature of the vector space modulo 8. This is a manifestation of Bott periodicity. One always has a Dirac spinor - the fundamental (spin) representation of the complexified Clifford algebra. In even dimensions, this splits into two Weyl spinors (called half-spin representations). One may also have real representations called Majorana spinors, and these may decompose into Majorana-Weyl spinors.
There are infinite dimensional Clifford algebra constructions that appear in conformal field theory. One may extend the above discussion to topological -modules and continuous quadratic forms, and one obtains canonical central extensions of infinite dimensional groups and algebras by a relative determinant construction. Semi-infinite wedge spaces are spinor modules for Clifford algebras of quadratic Tate -modules.
There is a difference of sign convention between differential geometers (following Atiyah) and everyone else.
Clifford algebras are often defined using bilinear forms instead of quadratic forms (and one often sees incorrect definitions of quadratic forms in terms of bilinear forms). Such definitions will yield wrong (or boring) objects when 2 is not invertible.
Special orthogonal groups are often defined as the kernel of the determinant map on the corresponding orthogonal groups, but in characteristic 2, the determinant is trivial, while the Clifford grading (called the Dickson invariant) is not.
The Clifford group is sometimes defined without the twist in the conjugation, and this means the map to the orthogonal group may not be a surjection, and the action on is by negative reflections.
The spinor norm is sometimes defined with the opposite sign.
Special orthogonal groups over the reals are sometimes defined to be the connected component of the identity in the orthogonal group. In indefinite signature, this defines an index two subgroup of the special orthogonal group.
Spin groups in signature for have fundamental groups of order two. They are simply connected when or is at most one.
For some standard material see for instance
H. Blaine Lawson, Marie-Louise Michelsohn, (1989), Spin Geometry, Princeton University Press, ISBN 0-691-08542-0.