 Division algebra

In the field of mathematics called abstract algebra, a division algebra is, roughly speaking, an algebra over a field, in which division is possible.
Contents
Definitions
Formally, we start with an algebra D over a field, and assume that D does not just consist of its zero element. We call D a division algebra if for any element a in D and any nonzero element b in D there exists precisely one element x in D with a = bx and precisely one element y in D such that a = yb.
For associative algebras, the definition can be simplified as follows: an associative algebra over a field is a division algebra if and only if it has a multiplicative identity element 1≠0 and every nonzero element a has a multiplicative inverse (i.e. an element x with ax = xa = 1).
Associative division algebras
The bestknown examples of associative division algebras are the finitedimensional real ones (that is, algebras over the field R of real numbers, which are finitedimensional as a vector space over the reals). The Frobenius theorem states that up to isomorphism there are three such algebras: the reals themselves (dimension 1), the field of complex numbers (dimension 2), and the quaternions (dimension 4).
Wedderburn's little theorem states that if D is a finite division algebra, then D is a finite field. (T. Y. Lam, A First Course in Noncommutative Rings.)
Over an algebraically closed field K (for example the complex numbers C), there are no finitedimensional associative division algebras, except K itself of course.
Associative division algebras have no zero divisors. A finitedimensional unital associative algebra (over any field) is a division algebra if and only if it has no zero divisors.
Whenever A is an associative unital algebra over the field F and S is a simple module over A, then the endomorphism ring of S is a division algebra over F; every associative division algebra over F arises in this fashion.
The center of an associative division algebra D over the field K is a field containing K. The dimension of such an algebra over its center, if finite, is a perfect square: it is equal to the square of the dimension of a maximal subfield of D over the center. Given a field F, equivalence classes of simple (contains only trivial twosided ideals) associative division algebras whose center is F and which are finitedimensional over F can be turned into a group, the Brauer group of the field F.
One way to construct finitedimensional associative division algebras over arbitrary fields is given by the quaternion algebras (see also quaternions).
For infinitedimensional associative division algebras, the most important cases are those where the space has some reasonable topology. See for example normed division algebras and Banach algebras.
Not necessarily associative division algebras
If the division algebra is not assumed to be associative, usually some weaker condition (such as alternativity or power associativity) is imposed instead. See algebra over a field for a list of such conditions.
Over the reals there are (up to isomorphism) only two unitary commutative finitedimensional division algebras: the reals themselves, and the complex numbers. These are of course both associative. For a nonassociative example, consider the complex numbers with multiplication defined by taking the complex conjugate of the usual multiplication:
This is a commutative, nonassociative division algebra of dimension 2 over the reals, and has no unit element. There are infinitely many other nonisomorphic commutative, nonassociative, finitedimensional real divisional algebras, but they all have dimension 2.
In fact, every finitedimensional real commutative division algebra is either 1 or 2 dimensional. This is known as Hopf's theorem, and was proved in 1940. The proof uses methods from topology. Although a later proof was found using algebraic geometry, no direct algebraic proof is known. The fundamental theorem of algebra is a corollary of Hopf's theorem.
Dropping the requirement of commutativity, Hopf generalized his result: Any finitedimensional real division algebra must have dimension a power of 2.
Later work showed that in fact, any finitedimensional real division algebra must be of dimension 1, 2, 4, or 8. This was independently proved by Michel Kervaire and John Milnor in 1958, again using techniques of algebraic topology, in particular Ktheory. Adolf Hurwitz had shown in 1898 that the identity held only for dimensions 1, 2, 4 and 8.^{[1]} (See Hurwitz's theorem.)
While there are infinitely many nonisomorphic real division algebras of dimensions 2, 4 and 8, one can say the following: any real finitedimensional division algebra over the reals must be
 isomorphic to R or C if unitary and commutative (equivalently: associative and commutative)
 isomorphic to the quaternions if noncommutative but associative
 isomorphic to the octonions if nonassociative but alternative.
The following is known about the dimension of a finitedimensional division algebra A over a field K:
 dim A= 1 if K is algebraically closed,
 dim A= 1, 2, 4 or 8 if K is real closed, and
 If K is neither algebraically nor real closed, then there are infinitely many dimensions in which there exist division algebras over K.
See also
References
 ^ Roger Penrose (2005). The Road To Reality. Vintage. ISBN 0099440687.
Categories: Algebras
 Ring theory
Wikimedia Foundation. 2010.