---
title: Generalized Real-Space Quadrupole Moment
url: https://www.emergentmind.com/topics/generalized-real-space-quadrupole-moment
type: topic
---

# Generalized Real-Space Quadrupole Moment

Searching arXiv for the cited topic and papers.
arxiv_search(query="generalized real-space quadrupole moment", max_results=10)
arxiv_search(query="quadrupole moment real-space operator multipole insulator", max_results=10)
arxiv_search({"query":"quadrupole moment real-space operator multipole insulator", "max_results": 10})
Searching arXiv for related papers: 1805.08257, 1812.06990, 2208.05058, 2503.16897, 2508.00277, 1912.12487.
The generalized real-space quadrupole moment denotes a family of coordinate-space constructions that extend ordinary dipole observables to second-rank moments without relying exclusively on momentum-space topology. In one line of work, it is a geometric tensor built from the unit normal field of a two-dimensional surface embedded in $\mathbb R^3$, designed to probe the anisotropy of quantum surfaces in loop quantum gravity (LQG). In another, it is a many-body or projector-based observable for crystalline and disordered lattice systems, formulated directly in real space so that quadrupole moments remain computable under periodic, open, infinite-lattice, or finite-temperature settings. Across these settings, the common purpose is to isolate shape information beyond total area or total charge, while preserving the appropriate invariances of the problem [1805.08257] [1812.06990] [2503.16897] [2508.00277].

## 1. Geometric definition on embedded surfaces

For a smooth two-dimensional surface $S\subset\mathbb R^3$ with local coordinates $\sigma=(u,v)\mapsto x(u,v)$, the induced metric is
$$
g_{AB}(\sigma)=\partial_A x\cdot \partial_B x,
$$
and the oriented surface element is determined by
$$
\vec N(\sigma)=\partial_u x\times \partial_v x,\qquad N(\sigma)=|\vec N(\sigma)|,\qquad \vec n(\sigma)=\vec N(\sigma)/N(\sigma).
$$
The area element is $dA=N\,du\,dv$. The dual quadrupole tensor is then defined by
$$
Q_{ij}=\int_S dA\, n_i(\sigma)\,n_j(\sigma)
=\int du\,dv\,N(\sigma)\,n_i(\sigma)\,n_j(\sigma)
=\int du\,dv\,N^{-1}(\sigma)\,N_i(\sigma)\,N_j(\sigma).
$$
Its trace is
$$
\mathrm{Tr}\,Q=Q_{ii}=\int_S dA=\mathrm{Area}(S).
$$
In this formulation, the monopole is the area, the dipole is the closure vector, and the quadrupole captures the leading anisotropy of the surface, distinguishing, for example, ellipsoidal deformations from the round sphere [1805.08257].

The construction is explicitly reparametrization-invariant. Under a change of coordinates $\sigma\mapsto \tilde\sigma(\sigma)$, the measure $dA=N\,du\,dv$ is invariant, while the unit normal $\vec n=\vec N/N$ behaves as a scalar under reparametrizations. Consequently, the integrand $N\,n_i n_j\,du\,dv$ is a scalar times the invariant measure, and $Q_{ij}$ is coordinate-independent. This invariance is not a secondary technicality: it is the condition that makes the quadrupole a genuine observable of shape rather than of parametrization [1805.08257].

## 2. Real-space quadrupole operators in lattice systems

In crystalline insulators, the direct position operator and its products are not well-defined on periodic Hilbert spaces, so real-space quadrupole observables are typically introduced in exponentiated form. A normalized lattice quadrupole operator may be written as
$$
\hat q_{xy}=\frac{1}{L_xL_y}\sum_j \hat n_j\,x_j\,y_j,
$$
with corresponding many-body unitary
$$
\hat U_2=\exp\!\bigl[i\,2\pi\,\hat q_{xy}\bigr],
$$
and quadrupole moment
$$
q_{xy}=\frac{1}{2\pi}\,\mathrm{Im}\,\ln\langle\Psi_G|\hat U_2|\Psi_G\rangle\quad (\mathrm{mod}\;1).
$$
A more general multipole construction on periodic lattices uses
$$
\hat{\mathcal U}^{(Q)}_{ab}
=\exp\!\Bigl[\frac{2\pi i}{N_aN_b}\,b_a^i\,\hat q^{ij}\,b_b^j\Bigr]
=\exp\!\Bigl[\frac{2\pi i}{L_aL_b}\sum_n(\mathbf b_a\!\cdot\!\hat{\mathbf x}_n)(\mathbf b_b\!\cdot\!\hat{\mathbf x}_n)\Bigr],
$$
from which the bulk quadrupole density is extracted through the phase of $\langle\Phi_0|\hat{\mathcal U}^{(Q)}_{ab}|\Phi_0\rangle$ in the thermodynamic limit [1812.06990] [1912.12487].

A central consistency condition is that lower multipoles vanish modulo their own quanta. The many-body quadrupole operator transforms trivially under lattice translations only if the filling per $ab$-plane is an integer and the dipole moment vanishes mod $e$; if the lower moments are nontrivial, the quadrupole expectation value must vanish. This constraint is fundamental to the interpretation of a bulk quadrupole as an invariant rather than an artifact of unresolved dipolar structure [1812.06990].

A distinct real-space route avoids momentum-space discretization altogether by using the spectral projector $P$ and a square-root projector $Q$. On an infinite lattice, the quadrupole is represented by
$$
q_{xy}
=
-\frac12\,\mathrm{Tr}\bigl[Q^\dagger\{\Sigma_{xy},Q\}\bigr],
$$
where $\Sigma_{xy,jl}=\sigma_{j_x}\sigma_{j_y}\delta_{jl}$ and $\sigma_j=\tfrac12\,\mathrm{sgn}(j)$. For a finite system under open boundary conditions,
$$
q_{xy}=-\,\mathrm{Re}\,\mathrm{Tr}[P\,\Sigma_{xy}].
$$
The infinite-lattice expressions assume translational invariance of $P$ or $Q$, while the open-boundary formula is numerically useful, explicitly gauge-invariant, and emphasizes the role of edge states in polarization and quadrupole moments [2503.16897].

## 3. Discrete and quantum-surface formulations in LQG

In LQG, a continuous surface is replaced by a polyhedral patch with $P$ faces, each carrying a normal $\vec N_f\in\mathbb R^3$ and area $|\vec N_f|$. The discrete quadrupole is
$$
Q_{ij}^{(\mathrm{disc})}
=
\sum_{f=1}^P \frac{1}{|\vec N_f|}\,N_f^i\,N_f^j.
$$
The inverse-norm weighting is the discrete analogue of the continuum factor $N^{-1}N_iN_j$ and is what preserves the shape-sensitive character of the construction [1805.08257].

Upon quantization, the face normals become SU(2) generators $J_i^a$ acting on the intertwiner Hilbert space
$$
H_P=\bigotimes_{i=1}^P\Bigl(\bigoplus_{j_i}V^{j_i}\Bigr).
$$
Because $1/|\vec N_i|=1/j_i$ must be regularized, a natural choice is the shift $j_i\to j_i+1$. Two operator orderings are then considered:
$$
\hat T^{ab}
=
\sum_{i=1}^P \frac{1}{j_i+1}\,J_i^aJ_i^b,
$$
and
$$
\hat{\mathcal T}^{ab}
=
\sum_{i=1}^P \frac{1}{j_i+1}\,\{J_i^a,J_i^b\}/2.
$$
Both satisfy
$$
\mathrm{Tr}\,\hat T\equiv \sum_a \hat T^{aa}=\sum_i j_i,
$$
so the trace reproduces the area operator. Higher invariants such as $\mathrm{Tr}(\hat T^2)$ and $\mathrm{Tr}(\hat T^3)$ are likewise well-defined SU(2)-invariant operators on the intertwiner space [1805.08257].

This quantum-surface quadrupole is not redundant with area or volume observables. The standard LQG toolkit already contains area, volume, and holonomy operators, and the motivation for introducing dual multipoles is precisely that those operators do not carry information on the global shape of intertwiners. In this sense, the generalized real-space quadrupole moment functions as a shape observable for fundamental quanta of geometry [1805.08257].

## 4. Relation to coherent intertwiners, Wannier topology, and bulk invariants

For Livine–Speziale coherent intertwiners,
$$
|\{j_i,z_i\}\rangle=\int_{SU(2)} dg\,\bigotimes_{i=1}^P D^{j_i}(g)\,|j_i,z_i\rangle,
$$
the norm takes the form
$$
\langle\{j_i,z_i\}|\{j_i,z_i\}\rangle=\int_{SU(2)} dg\,e^{S[g]},
$$
with
$$
S[g]=\sum_i 2j_i\ln\langle z_i|g|z_i\rangle.
$$
Expanding about the stationary point $g=I$, the first derivative enforces closure,
$$
2\sum_i j_i\vec n_i=0\qquad\Longleftrightarrow\qquad \sum_i \vec N_i=0,
$$
and the second derivative yields the Hessian
$$
H_{ab}
=
2\sum_i j_i(\delta_{ab}-n_i^an_i^b)
=
2[\delta_{ab}-Q_{ab}],
$$
where $Q_{ab}=\sum_i n_i^a n_i^b$ is the continuum-inspired quadrupole. The Gaussian approximation becomes
$$
\langle\{j_i,z_i\}|\{j_i,z_i\}\rangle
\sim
\mathrm{const}\cdot [\det H]^{-1/2}
=
\mathrm{const}\cdot \det[I-Q]^{-1/2}.
$$
The quadrupole therefore controls the Gaussian spread of coherent intertwiners and enters directly into the semiclassical regime of spinfoam amplitudes [1805.08257].

In higher-order topological insulators, an analogous structural role is played by nested Wilson loops and Wannier-sector polarizations. In the generalized separable BBH model,
$$
q_{xy}\equiv \sum_{\nu=\pm} p_x^\nu p_y^\nu
=\frac{\mathcal N_x\mathcal N_y}{2}\;\mathrm{mod}\;1
\in\Bigl\{0,\frac12\Bigr\},
$$
where $p_\mu^\nu$ are Wannier-sector polarizations and $\mathcal N_x,\mathcal N_y\in\mathbb Z$ are the winding numbers of the constitutive one-dimensional chains. The same quantity appears in a Bloch representation,
$$
Q_{xy}
=
- \sum_{n\in \mathrm{occ}}
\int_{\mathrm{FBZ}}\frac{d^2k}{(2\pi)^2}\,
\Psi_n^\dagger(k)\,\partial_{k_x}\partial_{k_y}\Psi_n(k)
=
\frac{\mathcal N_x\mathcal N_y}{2},
$$
and its fractional part agrees with the nested-Wilson-loop quadrupole. In that model, translation symmetry, chiral symmetry, and separability are sufficient; no mirror or $C_4$ rotation is required [2208.05058].

The many-body operator and the nested-Wilson-loop invariant are related but not interchangeable in all regimes. In generalized electric quadrupole insulators, the nested-loop construction can fail to diagnose certain real transitions, particularly when edge loops close and reopen while the bulk remains gapped or when an indirect-gap phase hides corner modes in bulk bands. By contrast, the real-space many-body formula tracks the presence of fractional corner charge and vanishes when the many-body gap closes, even indirectly [1912.12487].

## 5. Symmetry, finite temperature, and disorder

The quantization mechanism for a real-space quadrupole moment depends on the framework. Mirror symmetries $M_x$ or $M_y$ enforce $Q_{xy}\to -Q_{xy}$ and thus quantize $Q_{xy}$ to $0$ or $1/2$ modulo $1$. Chiral symmetry can also quantize the nested-loop result to half-integers in certain models, and generalized BBH constructions show that quadrupole moments can remain quantized even when mirror symmetries are absent. In a separable chiral-symmetric model, the quantization into $\{0,1/2\}$ follows from the winding numbers of the one-dimensional building blocks, while in extended BBH models one numerically observes $Q_{xy}=1/2$ and half-quantized edge polarizations even after breaking both $M_x$ and $M_y$ but preserving inversion $C_2$ [2208.05058] [1912.12487].

A finite-temperature generalization replaces ground-state expectation values by ensemble averages. For spinless fermions on an $L_x\times L_y$ lattice with $N=4L_xL_y$ orbitals,
$$
\hat Q_{xy}
=
\sum_{j=1}^N \frac{x_jy_j}{L_xL_y}\,\hat n_j,
$$
and with
$$
D=\mathrm{diag}\Bigl\{\exp\!\bigl(2\pi i\,x_jy_j/(L_xL_y)\bigr)\Bigr\}_{j=1}^N,
$$
the thermal density matrix
$$
\rho=\frac{1}{Z}\exp\!\Bigl(-\beta\sum_{i,j}c_i^\dagger H_{ij}c_j\Bigr),\qquad
Z=\det(I+e^{-\beta H}),
$$
leads to
$$
q_{xy}(T)
=
\frac{1}{2\pi}
\arg\!
\Bigl[
\frac{\det(I+e^{-\beta H}D)}
{\det(I+e^{-\beta H})\,\det(D^{1/2})}
\Bigr],
\qquad 0\le q_{xy}<1.
$$
If the single-particle Hamiltonian obeys chiral symmetry $\{\Pi,H\}=0$ with $\Pi^2=I$, then $\det(I+e^{-\beta H}D)\in\mathbb R$, every factor inside the argument is real, and
$$
q_{xy}(T)\in\Bigl\{0,\frac12\Bigr\}\qquad \text{for any }T>0.
$$
At $T\to 0$ this reproduces the zero-temperature formula in terms of occupied states, while at $T\to\infty$ one finds $q_{xy}\to 0$ [2508.00277].

Temperature and disorder alter the phase structure in ways not visible at zero temperature alone. In the isotropic BBH case $t_x=t_y$, there is a single transition from $q_{xy}=1/2$ at low $T$ to $q_{xy}=0$ at high $T$, with critical temperature decreasing to zero as $t_x\to 1^-$. In the anisotropic case $t_x\ne t_y$, there can be a reentrant nontrivial window at intermediate-high temperature. With quasi-periodic disorder $t_n=t_x+W\cos(2\pi\alpha n)$ and $\alpha=(\sqrt5+1)/2$, an initially trivial system can be driven into a topological phase for moderate $W$, then returned to $q_{xy}=0$ at larger $W$; at intermediate temperatures the phase can exhibit multiple re-entries as $W$ is swept [2508.00277]. In the infinite-lattice projector formalism, disorder may also be included in hopping or onsite energies without destroying the validity of the finite-system formulas, provided a bulk gap remains [2503.16897].

## 6. Physical meaning, bulk–boundary correspondence, and common pitfalls

The physical interpretation of the generalized real-space quadrupole moment is consistently shape-theoretic, but the object whose shape is being probed differs by field. In LQG, the quadrupole tensor measures the anisotropy of a quantum surface, its eigenvalues becoming quantized shape observables, and its appearance as a Hessian means it governs Gaussian shape fluctuations around a classical geometry. The proposed extension to spin-network states is to correlate quadrupole operators at different vertices so as to describe propagating shape fluctuations, viewed as discrete, gauge-invariant “shape-waves” and as a possible avenue toward modeling quantum gravitational waves in the kinematical framework [1805.08257].

In higher-order topological matter, the quadrupole moment is tied to edge polarizations, corner charges, and adiabatic response to electric-field gradients. Changes in the phase of the many-body quadrupole operator are tied to flows of dipole current, and the operator encodes the adiabatic evolution of the system in the presence of an $n-1$st gradient of the electric field. In the BBH family, the quadrupole pump evolves the ground-state quadrupole from $0\to e/2\to 0$, and the many-body operator tracks the corner charge and edge polarization throughout the cycle [1812.06990].

Two recurrent misconceptions are explicitly ruled out by the literature. First, fractional corner charge is not by itself sufficient to establish a nontrivial bulk quadrupole: there are mirror-symmetric models with fractional corner charge but vanishing bulk quadrupole, for which the operator $\hat{\mathcal U}^{(Q)}_{xy}$ remains trivial [1812.06990]. Second, symmetry requirements are model-dependent rather than universal. Early quadrupole-insulator constructions emphasized crystalline symmetries, but later results show that translation symmetry plus chiral symmetry and separability can already yield quantized $q_{xy}\in\{0,1/2\}$, and finite-temperature quantization can be enforced by chiral symmetry alone [2208.05058] [2508.00277].

Taken together, these formulations establish the generalized real-space quadrupole moment as a next-to-leading observable beyond area or dipole order. In geometry, it resolves global shape information not captured by area and volume. In condensed matter, it provides a bulk invariant and response diagnostic that survives real-space disorder, open boundaries, and thermal ensemble averaging. The common structure is the extraction of anisotropy from a coordinate-space second moment while preserving the invariances—reparametrization, lattice translation, gauge covariance, or symmetry quantization—required by the underlying theory.

Source: https://www.emergentmind.com/topics/generalized-real-space-quadrupole-moment