Inabstract algebra, amatrix ring is a set ofmatrices with entries in aringR that form a ring undermatrix addition andmatrix multiplication.[1] The set of alln ×n matrices with entries inR is a matrix ring denoted Mn(R)[2][3][4][5] (alternative notations: Matn(R)[3] andRn×n[6]). Some sets of infinite matrices forminfinite matrix rings. A subring of a matrix ring is again a matrix ring. Over arng, one can form matrix rngs.
WhenR is a commutative ring, the matrix ring Mn(R) is anassociative algebra overR, and may be called amatrix algebra. In this setting, ifM is a matrix andr is inR, then the matrixrM is the matrixM with each of its entries multiplied byr.
For any index setI, the ring of endomorphisms of the rightR-module is isomorphic to the ring[citation needed] ofcolumn finite matrices whose entries are indexed byI ×I and whose columns each contain only finitely many nonzero entries. The ring of endomorphisms ofM considered as a leftR-module is isomorphic to the ring ofrow finite matrices.
IfR is aBanach algebra, then the condition of row or column finiteness in the previous point can be relaxed. With the norm in place,absolutely convergent series can be used instead of finite sums. For example, the matrices whose column sums are absolutely convergent sequences form a ring.[dubious –discuss] Analogously of course, the matrices whose row sums are absolutely convergent series also form a ring.[dubious –discuss] This idea can be used to representoperators on Hilbert spaces, for example.
The intersection of the row-finite and column-finite matrix rings forms a ring.
IfA is aC*-algebra, then Mn(A) is another C*-algebra. IfA is non-unital, then Mn(A) is also non-unital. By theGelfand–Naimark theorem, there exists aHilbert spaceH and an isometric *-isomorphism fromA to a norm-closed subalgebra of the algebraB(H) of continuous operators; this identifies Mn(A) with a subalgebra ofB(H⊕n). For simplicity, if we further suppose thatH is separable andAB(H) is a unital C*-algebra, we can break upA into a matrix ring over a smaller C*-algebra. One can do so by fixing aprojectionp and hence its orthogonal projection 1 − p; one can identifyA with, where matrix multiplication works as intended because of the orthogonality of the projections. In order to identifyA with a matrix ring over a C*-algebra, we require thatp and 1 − p have the same "rank"; more precisely, we need thatp and 1 − p are Murray–von Neumann equivalent, i.e., there exists apartial isometryu such thatp =uu* and1 −p =u*u. One can easily generalize this to matrices of larger sizes.
Complex matrix algebras Mn(C) are, up to isomorphism, the only finite-dimensional simple associative algebras over the fieldC ofcomplex numbers. Prior to the invention of matrix algebras,Hamilton in 1853 introduced a ring, whose elements he calledbiquaternions[7] and modern authors would call tensors inC ⊗RH, that was later shown to be isomorphic to M2(C). Onebasis of M2(C) consists of the four matrix units (matrices with one 1 and all other entries 0); another basis is given by theidentity matrix and the threePauli matrices.
A matrix ring over a field is aFrobenius algebra, with Frobenius form given by the trace of the product:σ(A,B) = tr(AB).
TheArtin–Wedderburn theorem states that every semisimple ring is isomorphic to a finitedirect product, for some nonnegative integerr, positive integersni, and division ringsDi.
When we view Mn(C) as the ring of linear endomorphisms ofCn, those matrices which vanish on a given subspaceV form aleft ideal. Conversely, for a given left idealI of Mn(C) the intersection ofnull spaces of all matrices inI gives a subspace ofCn. Under this construction, the left ideals of Mn(C) are in bijection with the subspaces ofCn.
There is a bijection between the two-sidedideals of Mn(R) and the two-sided ideals ofR. Namely, for each idealI ofR, the set of alln ×n matrices with entries inI is an ideal of Mn(R), and each ideal of Mn(R) arises in this way. This implies that Mn(R) issimple if and only ifR is simple. Forn ≥ 2, not every left ideal or right ideal of Mn(R) arises by the previous construction from a left ideal or a right ideal inR. For example, the set of matrices whose columns with indices 2 throughn are all zero forms a left ideal in Mn(R).
The previous ideal correspondence actually arises from the fact that the ringsR and Mn(R) areMorita equivalent. Roughly speaking, this means that the category of leftR-modules and the category of left Mn(R)-modules are very similar. Because of this, there is a natural bijective correspondence between theisomorphism classes of leftR-modules and left Mn(R)-modules, and between the isomorphism classes of left ideals ofR and left ideals of Mn(R). Identical statements hold for right modules and right ideals. Through Morita equivalence, Mn(R) inherits anyMorita-invariant properties ofR, such as beingsimple,Artinian,Noetherian,prime.
IfS is asubring ofR, then Mn(S) is a subring of Mn(R). For example, Mn(Z) is a subring of Mn(Q).
The matrix ring Mn(R) iscommutative if and only ifn = 0,R = 0, orR iscommutative andn = 1. In fact, this is true also for the subring of upper triangular matrices. Here is an example showing two upper triangular2 × 2 matrices that do not commute, assuming 1 ≠ 0 inR:
and
Forn ≥ 2, the matrix ring Mn(R) over anonzero ring haszero divisors andnilpotent elements; the same holds for the ring of upper triangular matrices. An example in2 × 2 matrices would be
Thecenter of Mn(R) consists of the scalar multiples of theidentity matrix,In, in which the scalar belongs to the center ofR.
Theunit group of Mn(R), consisting of the invertible matrices under multiplication, is denoted GLn(R).
IfF is a field, then for any two matricesA andB in Mn(F), the equalityAB =In impliesBA =In. This is not true for every ringR though. A ringR whose matrix rings all have the mentioned property is known as astably finite ring (Lam 1999, p. 5).
In fact,R needs to be only asemiring for Mn(R) to be defined. In this case, Mn(R) is a semiring, called thematrix semiring. Similarly, ifR is a commutative semiring, then Mn(R) is amatrix semialgebra.
Droste, M.; Kuich, W (2009), "Semirings and Formal Power Series",Handbook of Weighted Automata, Monographs in Theoretical Computer Science. An EATCS Series, pp. 3–28,doi:10.1007/978-3-642-01492-5_1,ISBN978-3-642-01491-8