Quantum Euclidean spaces with noncommutative derivatives (Q2140249)

Quantum Euclidean spaces have been studied in noncommutative geometry (and quantum physics) under the names of CCR algebras, noncommutative Euclidean spaces and Moyal planes. These can be identified with the von Neumann algebra \(\mathbb R_\theta\) spanned by projectors \(x_1,\ldots,x_d\) that satisfy \( [x_i,x_j]=-\mathrm{i} \theta_{i,j}\) with \(\mathrm{i}=\sqrt{-1}\) and \(\theta\) a skew-symmetric \(d\times d\)-matrix. Also, differentiation operators \(D_1,\ldots,D_d\) are constructed in detail in \S 2 that satisfy \[ [x_k,x_j]=-\mathrm{i} \theta_{k,j} \quad [D_j,D_k]=0 \quad \text{and} \quad [x_k,D_j]= \mathrm{i}\, \delta_{k,j}\text{ for all } j,k=1,\ldots,d.\, \] The \(\Psi\)DO-calculus under these relations was established as operators on \(L_2(\mathbb R_\theta)\) by [\textit{A. M. González-Pérez} et al., Singular integrals in quantum Euclidean spaces. Providence, RI: American Mathematical Society (AMS) (2021; Zbl 1495.42001)]. In the article under review, the authors replace the CCR's by more general rules \[ [x_j,x_k]=-\mathrm{i}\,\theta_{j,k} ,\quad [\xi_j,\xi_k]=-\mathrm{i}\,\theta'_{j,k},\quad [x_j,\xi_k]=\mathrm{i}\,\delta_{j,k} \text{ for all } j,k=1,\ldots,d, \] for a certain additional skew-symmetric \(d\times d\)-matrix \(\theta'\), and thoroughly develop the theory of \(\Psi\)DO's on \(L_2(\mathbb R_{\Theta})\), with \(\Theta=\big(\begin{smallmatrix} \theta & I \\ I & \theta' \end{smallmatrix}\big) \). On top of being mathematically natural, an additional asset of these generalized CCR's is their applicability to quantum dynamics of charged particles on \(\mathbb R^d\) in the presence of a constant magnetic field normal to that plane. Each \(\xi_j\) is a covariant derivative generalizing \(D_j\) (`ordinary partial derivative') but, crucially, such that \([\xi_i,\xi_j]\) is not identically \( 0\) (whence `noncommutative derivatives' in the title). The \(\Psi\)DO calculus developed here allows the authors also to prove a local index formula for a spectral triple over the algebra \(\mathcal A_\Theta=W^{1,\infty}(\mathbb R_\theta)\) (a noncommutative Sobolev space; cf. \S 5), which acts on the Hilbert space \(\mathcal H_\Theta=L_2(\mathbb R_\Theta)\otimes \mathbb C^{N}\). There, the following Dirac operator is defined \[ D_{\Theta}=\sum_{j=1}^{d} \xi\otimes \gamma_j, \] where \(\gamma_1,\ldots, \gamma_d\) generate the Clifford algebra \(\mathrm{Cl}_d\) and \(\mathbb C^N\) is the spinor space. The authors prove that on an even-dimensional quantum space (say \(d=2n\), so the spinor space is \(\mathbb C^{2^n}\)), the spectral triple \((A_\Theta,H_\Theta,D_\Theta)\) is even, smoothly summable, semifinite nonunital (in the sense of [\textit{A. L. Carey} et al., Adv. Math. 202, No. 2, 451--516 (2006; Zbl 1118.46060)]) and that it has isolated spectrum dimension. They prove the following index pairing formula of a \(K_0(\mathcal A_\Theta^{\sim}) \)-element \([e]-[1_e]\) with the previous spectral triple \((A_\Theta,H_\Theta,D_\Theta)\). Namely, \[ \frac1{\pi^n}\big\langle [e]-[1_e],(A_\Theta,H_\Theta,D_\Theta)\big\rangle =\tau_\theta\otimes \mathrm{tr}\Big(\gamma (e-1_e) \frac{\omega^n}{n!}\Big)+\sum_{m=1}^{n}\frac{1}{(2m)!}\tau_\theta\otimes \mathrm{tr}\Big(\gamma e(\mathrm de)^{2m} \frac{\omega^{n-m}}{(n-m)!}\Big), \] where \(\omega\) is the curvature form \(\omega=\frac{\mathrm{i}}{2}\sum_{j,k=1}^d \theta_{j,k} \gamma_j\gamma_k\). Above, \(\gamma\) is the \(\mathbb Z_2\)-grading of \(\mathcal H_\Theta\), and \(\tau_\theta\) is the canonical trace associated to \(\mathbb R^d\). The text is insightful and contains an important example worked out in depth at the end of the article for the case \(d=2\) of this formula.