On Essential and Topologically Essential Submodules of Hilbert C ∗ C^{*} -Modules
K. Kartvelishvili · 2026
Read On Essential and Topologically Essential Submodules of Hilbert C ∗ C^{*} -Modules by K. Kartvelishvili (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-10206.txt.
Introduction
Kartvelishvili weighs two notions of size for closed submodules of Hilbert C*-modules: essentiality and topological essentiality. For closed two-sided ideals of C*-algebras the two traits coincide, but the one-sided case is subtler. He proves they remain equal for closed right ideals of any C*-algebra. Via the link between submodules and right ideals of compact operators, he extends the result to closed submodules of Hilbert C*-modules. When things commute, he gives a shape-based account and a per-fiber test.
Catalog record
| CREATOR | K. Kartvelishvili |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21357 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.10206
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On Essential and Topologically Essential Submodules of Hilbert C ∗ C^{*} -Modules K. Kartvelishvili Abstract . We study two notions of largeness for closed submodules of Hilbert C ∗ C^{*} -modules: essentiality and topological essentiality. While the analogous properties are known to be equivalent for closed two-sided ideals of C ∗ C^{*} -algebras, the one-sided case is more subtle. We prove that these two notions remain equivalent for closed right ideals of an arbitrary C ∗ C^{*} -algebra. Next, using the correspondence between submodules and right ideals of the algebra of compact operators, we extend this result to closed submodules of Hilbert C ∗ C^{*} -modules. In the commutative case, where a Hilbert module can be realized as a continuous field of Hilbert spaces, we give a geometric reformulation of essentiality and derive a fiberwise criterion. 1. Introduction For an ideal of a C^{*} -algebra, there are several natural ways to say that it is large. Some of these notions, together with their natural generalizations to the case of submodules of Hilbert modules, were considered in [ 3 ] . It is well known that if one restricts attention to closed two-sided ideals, then all these properties are equivalent. For one-sided ideals, however, this equivalence fails, and that paper establishes some relations between different definitions of a “large” closed one-sided ideal. For general background on Hilbert C^{*} -modules, we refer the reader to [ 2 ] . We focus on two notions for closed submodules of Hilbert C^{*} -modules, namely essentiality and topological essentiality. Definition 1.1 . Let \mathcal{M} be a Hilbert module over a C^{*} -algebra A . A closed submodule \mathcal{N}\subset\mathcal{M} is called (E) essential if for every nonzero (not necessarily closed) submodule \mathcal{K}\subset\mathcal{M} , the intersection \mathcal{N}\cap\mathcal{K} is nontrivial; (TE) topologically essential if for every nonzero closed submodule \mathcal{K}\subset\mathcal{M} , the intersection \mathcal{N}\cap\mathcal{K} is nontrivial. In the particular case where the algebra A is regarded as a right Hilbert C^{*} -module over itself, this yields the definitions of topologically essential (TE) and essential (E) right ideals of a C^{*} -algebra. The property (E) may seem somewhat unnatural in the context of C^{*} -algebras and Hilbert modules, since it involves nonclosed submodules or ideals. However, it admits a reformulation that looks less algebraic and is more convenient in practice (see Lemma 3.3 ). Our main goal is to prove that these definitions are equivalent. In Section 2 we show that essentiality and topological essentiality are equivalent for closed right ideals of an arbitrary C^{*} -algebra. In Section 3 the result of Section 2 is generalized to submodules of Hilbert C^{*} -modules. In Section 4 we consider the important special case of a Hilbert module over a commutative C^{*} -algebra. In this setting, the module can be viewed as a continuous field of Hilbert spaces over a locally compact Hausdorff space, and essentiality of a submodule can be reformulated in geometric terms. 2. (TE)\iff(E) for one-sided ideals of a C*-algebra Clearly, (E)\Rightarrow(TE) . To prove the converse implication, we need several standard facts from the operator algebra theory. Let A be an arbitrary C^{*} -algebra. By means of the universal representation (see [ 1 , Theorem I.9.12] ), we may identify it with a closed * -subalgebra of \mathbb{B}(H) , the algebra of bounded operators on a Hilbert space. Denote by A^{\prime\prime} the bicommutant of A , that is, the smallest von Neumann subalgebra of \mathbb{B}(H) containing A . A projection p\in A^{\prime\prime} is called open if there exists a net of positive elements a_{s}\in A_{+} increasing monotonically to p in the strong operator topology. It is known (see Theorem 1.1 in [ 5 ] ) that closed one-sided ideals of A admit a description in terms of open projections. More precisely, for every open projection p\in A^{\prime\prime} , the set J=pA^{\prime\prime}\cap A is a closed right ideal of A , and every closed right ideal is of that form. Every von Neumann algebra M admits the Borel functional calculus for self-adjoint elements (see [ 1 , Corollary II.2.6] ). That is, for every a=a^{*}\in M , there is a ∗ -homomorphism from the algebra of bounded Borel functions on the spectrum \sigma(a) to the algebra M , sending the function g(t)=t to a . The following two technical lemmas are standard. Lemma 2.1 . Let g_{k} be an increasing sequence of continuous functions on the spectrum of a self-adjoint element a\in A . Suppose that g_{k}\to g pointwise as k\to\infty . Then the sequence g_{k}(a) converges increasingly to g(a)\in A^{\prime\prime} in the strong operator topology. For a proof one may refer to Lemma I.6.4 in [ 1 ] . Lemma 2.2 . Let a=a^{*}\in A , \varepsilon>0 , and let \chi_{(\varepsilon,+\infty)} be the characteristic function of the ray t>\varepsilon . Then p=\chi_{(\varepsilon,+\infty)}(a) is an open projection in A^{\prime\prime} . Proof. The Borel function g:=\chi_{(\varepsilon,+\infty)} can be approximated by an increasing sequence of continuous functions on the real line \mathbb{R} . For example, g_{n}(t)=\begin{cases}0,\;t\leqslant\varepsilon\\ n(t-\varepsilon),\;t\in(\varepsilon,\varepsilon+\frac{1}{n})\\ 1,\;t\in[\varepsilon+\frac{1}{n},+\infty)\end{cases} Hence, by the previous lemma, the sequence g_{n}(a) increases to g(a) in the strong operator topology. Therefore g(a)=p is an open projection. ∎ We now prove the equivalence of the properties (TE) and (E) for one-sided ideals of a C^{*} -algebra. Theorem 2.3 . Every topologically essential right ideal J of a C^{*} -algebra A is essential. Proof. Let I be a nonzero right ideal of A , not necessarily closed. Choose x\in I , x\neq 0 . Then a=xx^{*}\neq 0 is a positive element of I . Using the Borel functional calculus in A^{\prime\prime} , set (a-\varepsilon)_{+}=\max(a-\varepsilon,0)=(a-\varepsilon)\chi_{(\varepsilon,+\infty)}(a). There exists \varepsilon>0 such that (a-\varepsilon)_{+}\neq 0 . Indeed, otherwise for every \varepsilon>0 we would have a-\varepsilon\leqslant 0 , hence a\leqslant 0 , which implies a=0 . Thus p=\chi_{(\varepsilon,+\infty)}(a) is a nonzero projection, and by Lemma 2.2 it is open. Hence K=pA^{\prime\prime}\cap A is a nonzero closed right ideal of A . Let g be a continuous function on the real line \mathbb{R} such that g(t)=0 for t<\frac{\varepsilon}{2} and g(t)=\frac{1}{t} for t\geqslant\varepsilon . Put f(t)=tg(t) and consider f(a)=ag(a)\in I . We have f(a)p=f\chi_{(\varepsilon,+\infty)}(a)=\chi_{(\varepsilon,+\infty)}(a)=p, since f(t)=1 for t\geqslant\varepsilon . Now take any b\in K . Then b=pb=f(a)pb\in I. Thus K\subset I and by topological essentiality of the ideal J we have J\cap I\supset J\cap K\neq 0.\qed Remark . The proof also yields the following interesting fact. Every nonzero right ideal I\subset A contains a norm-closed right subideal K\subset I . 3. Generalization to Hilbert modules Let \mathcal{M} be a right Hilbert C^{*} -module over a C^{*} -algebra A . Denote by \mathbb{K}(\mathcal{M}) the C^{*} -algebra of compact operators on \mathcal{M} , that is, the closure of the linear span of the elementary operators \Theta_{x,y}(z):=x\langle y,z\rangle over all x,y\in\mathcal{M} . It is well known (see [ 4 , Corollary 4.2] ) that closed submodules of \mathcal{M} are in one-to-one correspondence with hereditary subalgebras, and hence with closed right ideals of \mathbb{K}(\mathcal{M}) . Namely, for a closed submodule \mathcal{N}\subset\mathcal{M} define the ideal J_{\mathcal{N}}=\{T\in\mathbb{K}(\mathcal{M})\mid\operatorname{Ran}T\subset\mathcal{N}\} . Conversely, from a closed right ideal J\subset\mathbb{K}(\mathcal{M}) one recovers the submodule \mathcal{N}=\overline{J\mathcal{M}} . Theorem 6.1 of [ 3 ] states that this correspondence preserves essentiality, that is, \mathcal{N} is an (E) -submodule if and only if J_{\mathcal{N}} is an (E) -ideal. We now show