Twisted absorption for pattern-equivariant Heisenberg modules

Readers coming from tiling dynamics can begin with The proof without technical machinery and the tiling-theory dictionary. The complete PV/Barlak--KK comparison and Hilbert-module calculations are retained in Part III as technical certification chapters, so they need not be mastered before the construction itself is understood.

The introductory part also contains a short construction of Rieffel--Heisenberg modules, including their phase-space lattice, Gabor interpretation, projectivity, and trace. It explains the magnetic input used by the theorem without requiring the analytic descent calculations of Part III.

This book explains a constructive theorem about graded finite-projective representatives over twisted crossed products

\[ A_{\Sigma,\Theta} =C(\Sigma)\rtimes_{\sigma_\Theta}\mathbb Z^p, \]

where \(\Sigma\) is a Cantor dynamical system arising from an aperiodic pattern and \(\Theta\) is a constant magnetic field.

Why gap labels are K-theoretic

The observable algebra of an aperiodic solid is assembled from two kinds of information: the local patterns seen in the hull and the translations which move between them. In the Cantor-transversal model these become the coefficient algebra \(C(\Sigma)\) and the action of \(\mathbb Z^p\). A constant magnetic field changes the translation law: magnetic translations commute only up to the multiplier \(\sigma_\Theta\). The twisted crossed product \(A_{\Sigma,\Theta}\) is therefore the natural receptacle for covariant finite-range Hamiltonians and their norm limits.

If a Fermi energy lies in a spectral gap of such a Hamiltonian, continuous functional calculus produces a Fermi projection and hence a class

\[ [P_F]\in K_0(A_{\Sigma,\Theta}). \]

For every invariant probability measure \(\mu\) on the hull, the integrated density of states in that gap is

\[ \operatorname{IDS}(E_F) =\tau_\mu([P_F]). \]

Thus the universal group of possible gap labels is controlled by the ordered, traced \(K_0\)-group, and in particular by

\[ \tau_\mu\left(K_0(A_{\Sigma,\Theta})\right). \]

This statement concerns the allowed label group. It does not assert that every class in that group is realized as a gap of one fixed Hamiltonian. It separates the topological constraint on all covariant Hamiltonians from the additional spectral question of which gaps a particular model actually opens.

In the untwisted theory, the trace range is governed by frequencies of finite patches, encoded by integrals of integer-valued locally constant functions on \(\Sigma\). Magnetic twisting introduces new topological coefficients. The expected magnetic frequency group has the schematic form

\[ \mathcal G_\Theta(\mu)={} \sum_{\substack{I\subseteq\{1,\ldots,p\}\\|I|\text{ even}}} \operatorname{Pf}(\Theta_I)\mathbb Z_I[\mu], \]

where the groups \(\mathbb Z_I[\mu]\) are obtained from pattern frequencies after taking the appropriate directional coinvariants and invariants. Benameur and Mathai's magnetic gap-labelling conjecture is, in its general form, the upper containment

\[ \tau_\mu\left(K_0(A_{\Sigma,\Theta})\right) \subseteq \mathcal G_\Theta(\mu). \]

This upper-bound statement says that no unexpected trace values can occur. Equality, or even the reverse inclusion for a specified summand, is a separate existence problem: one must produce enough \(K_0\) classes with the predicted traces.

Why constructive representatives matter

Index theory, assembly maps, and spectral sequences can detect trace values without producing a finite pattern-equivariant projection or module which realizes them. For gap labelling this abstract information is already powerful, but a constructive representative answers several additional questions:

• how the patch-frequency class is incorporated into the magnetic module;
• whether the representative can be chosen local, finite propagation, and pattern equivariant;
• how the magnetic and aperiodic pieces interact before taking the trace; and
• which cohomological obstruction prevents an abstract class from descending to a strict module of the desired form.

The construction in this book makes the mixed nature of the class visible. If \(f\) is directional coinvariant data and \(x\in K_0(A_{\Theta_G})\) is a transverse noncommutative-torus class, the resulting graded module has

\[ \tau_\mu([P_{f,x}]) =\mu(f)\tau_{\Theta_G}(x). \]

When the torus pairing isolates a Pfaffian term, this becomes \(\operatorname{Pf}(\Theta_I)\mu(f)\). If \(f\) records nonconstant clopen pattern data, the resulting class is genuinely quasicrystalline: it is not merely the pullback of a class from the noncommutative torus. The formula is therefore not only a numerical coincidence in a trace calculation; it is implemented by a pattern-equivariant Rieffel--Heisenberg module.

This is the sense in which the theorem supplies a constructive lower-bound mechanism for magnetic gap labelling. Combined with an independently proved upper containment, such a construction can yield equality for the subgroup or dimension under consideration. The construction alone is not an upper-bound theorem and is not a complete calculation of \(K_0\).

Place in the K-theory program for quasicrystals

Understanding a quasicrystal \(C^*\)-algebra involves more than computing an abstract abelian group. A fuller program has at least four interlocking parts:

1. compute \(K_0\) and \(K_1\), using crossed-product exact sequences, groupoid methods, assembly, or spectral sequences;
2. identify geometric or pattern-equivariant representatives of the classes;
3. compute their pairings with invariant traces and higher cyclic cocycles; and
4. compare the resulting lower bounds with index-theoretic upper bounds to determine the actual gap-labelling group.

Twisted algebras sharpen this program. Even when deformation methods relate the abstract \(K\)-groups of twisted and untwisted crossed products, they do not by themselves identify finite PE representatives, compute the magnetic trace pairing, or explain the appearance of Pfaffian coefficients. Conversely, a trace calculation does not determine all of \(K_0\), since the trace can have a large kernel. Explicit modules provide a bridge between these two levels: they retain the local pattern data while carrying the global magnetic topology.

The present height-one theorem is one step in that larger program. It constructs a substantial family of mixed classes, removes the finite sheet condition from the block-supported height-one representative problem, works in arbitrary ambient dimension for every transverse Rieffel height, and isolates the remaining diagonal phase obstruction when the magnetic field has height--transverse components. It also shows precisely what the method does not yet supply. A full realization of all expected Pfaffian layers would require higher-rank absorption, in which several coefficient directions are absorbed simultaneously into a higher-dimensional Heisenberg module. A full equality theorem would additionally require the corresponding analytic or topological upper containment.

Conceptually, the book connects three constructive traditions: Rieffel's projective modules over noncommutative tori, Gabor modules for quasicrystals, and the cohomological magnetic gap-labelling framework of Benameur and Mathai. The new ingredient is to absorb clopen coinvariant data into finite PE transport and then strictify its dynamical defect before performing Rieffel descent.

The problem addressed here

The problem begins with a clopen coinvariant class. In three dimensions, Benameur and Mathai constructed the desired magnetic \(K_0\)-class when the clopen set admits a finite sheet decomposition, their Hypothesis (H). The construction developed here replaces those sheets by finite pattern-equivariant transport along one minimal “height” direction.

Current proof status. The operator construction, the PV/Barlak comparison identifying the transfer-function curl with \(d_2\), the GPS reduction under the normalizer hypothesis, and the descent/trace calculation are proved. The block-supported height-one theorem and its stated lower bounds are unconditional under the displayed minimal-height and normalizer hypotheses. A full magnetic matrix still requires the separate finite-stage diagonal phase-coboundary hypothesis.

The central result has two forms.

1. Block-supported magnetic field. If the magnetic multiplier has no height–transverse entries, every invariant height-coinvariant class admits a strict finite PE equivariant representative when the height homeomorphism is minimal.
2. General magnetic field. Under the same minimal-height hypothesis, the norm-continuous phase-scaling path (A.20)--(A.22) identifies its primary \(K_1\)-curvature with the block-supported one. The GPS construction then removes the nonabelian permutation defect. What remains in a fixed corner is an explicit locally constant diagonal \(2\)-cocycle. Its vanishing is necessary and sufficient for diagonal strictification in that corner. Across representatives and stabilizations, the book uses only an existential vanishing hypothesis and does not claim a choice-independent stable invariant.

The proof is constructive. It produces matrices whose entries are clopen-weighted powers of the height shift. After Rieffel descent, each generator acts as

\[ \text{ordinary time-frequency shift} \quad\times\quad \text{finite PE height transport}. \]

Technical Engine B proves the balanced correspondence, compactness, and product trace formula. The native time-frequency realization works for every permitted Rieffel height, not only the top-dimensional Pfaffian class.

flowchart LR
    f["Invariant clopen coinvariant f"]
    P["Full clopen projection P"]
    W["Individual PE transports W_i"]
    K["Pairwise K_1 curvature"]
    GPS["GPS normalizer retraction"]
    phase["Diagonal magnetic phase ν"]
    module["Mixed Heisenberg module"]

    f --> P --> W --> K
    K -->|"proved comparison + minimal height"| GPS
    GPS --> phase
    phase -->|"block support: ν = 1"| module
    phase -->|"general Θ: coboundary at an allowed finite stage"| module

Main references

The proof uses:

• Benameur–Mathai’s explicit magnetic gap-label construction;
• Barlak’s formula for the \(d_2\)-differential;
• the normalizer decomposition of Giordano–Putnam–Skau;
• Rieffel’s Heisenberg modules and their Gabor realizations.

Precise links and theorem locations appear in Sources. The consequences for magnetic gap labelling are developed separately in dimension three and in higher dimensions.