Canonical: https://knowlpedia.com/operator-algebras/full-hereditary-subalgebra/

# Full hereditary C*-subalgebra

A hereditary C*-subalgebra whose generated closed two-sided ideal is the whole ambient algebra.

Let \(A\) be a [\(C^*\)-algebra](https://knowlpedia.com/operator-algebras/cstar-algebra/) and
\(B\subseteq A\) a
[hereditary \(C^*\)-subalgebra](https://knowlpedia.com/operator-algebras/hereditary-cstar-subalgebra/). The subalgebra \(B\) is **full in \(A\)** if it is
contained in no proper
[closed two-sided ideal](https://knowlpedia.com/operator-algebras/closed-two-sided-ideal/) of \(A\).
Equivalently, the ideal generated by \(B\) is all of \(A\):
\[
\overline{\operatorname{span}}\,ABA=A.
\]
Fullness is therefore an ambient property of the inclusion \(B\subseteq A\);
the same \(C^*\)-algebra embedded hereditarily in two different ambient
algebras may be full in one and not in the other.

## Full corners

If \(p\) is a [projection](https://knowlpedia.com/operator-algebras/projection-cstar-algebra/) in
\(A\), or suitably in its
[multiplier algebra](https://knowlpedia.com/operator-algebras/multiplier-algebra/), then \(pAp\) is
hereditary. It is a **full corner** exactly when \(p\) is a full projection, meaning
\(\overline{\operatorname{span}}\,ApA=A\). Every full corner is thus a full
hereditary subalgebra, but a general hereditary subalgebra need not be
presented by a projection belonging to \(A\).

## Morita-theoretic significance

A full hereditary \(C^*\)-subalgebra \(B\) is
[strongly Morita equivalent](https://knowlpedia.com/operator-algebras/strong-morita-equivalence/)
to \(A\), with the completion of \(BA\) providing the equivalence module.
Under \(\sigma\)-unital hypotheses, the Brown–Green–Rieffel theorem further
relates this equivalence to stable isomorphism.

## Examples and non-examples

For a nonzero projection \(p\in\mathcal K(H)\), the corner
\(p\mathcal K(H)p\) is full because the
[compact-operator algebra](https://knowlpedia.com/operator-algebras/compact-operator-cstar-algebra/) is simple.
By contrast, if \(A=A_1\oplus A_2\), then \(A_1\oplus0\) is hereditary but
not full: the ideal it generates is only \(A_1\oplus0\).

## References

1. Marc A. Rieffel, “Morita Equivalence for \(C^*\)-Algebras and \(W^*\)-Algebras,” *Journal of Pure and Applied Algebra* 5 (1974), 51–96. [DOI record](https://doi.org/10.1016/0022-4049%2874%2990003-6). Relevant: §§1–2 on full hereditary subalgebras and equivalence bimodules.
2. Lawrence G. Brown, Philip Green, and Marc A. Rieffel, “Stable Isomorphism and Strong Morita Equivalence of \(C^*\)-Algebras,” *Pacific Journal of Mathematics* 71 (1977), 349–363. [Journal PDF](https://msp.org/pjm/1977/71-2/pjm-v71-n2-p06-s.pdf). Relevant: Theorem 1.2 and the discussion of full hereditary subalgebras.
