---
title: Three-Impurity Anderson Model
url: https://www.emergentmind.com/topics/three-impurity-anderson-model
type: topic
---

# Three-Impurity Anderson Model

Searching arXiv for the cited papers to ground the article in the current record.
The **Three-Impurity Anderson Model** denotes, in the literature considered here, two closely related but distinct Anderson-impurity settings. In one sense, it refers to a **three-orbital Anderson impurity model** with a single correlated impurity carrying three orbital degrees of freedom and coupled to a conduction bath; in another, it denotes a **three-site multi-impurity cluster** or trimer, in which three correlated orbitals or moments are embedded in a host. Both realizations are governed by the same general ingredients—local Coulomb interactions, hybridization to itinerant electrons, and low-energy screening—but they organize the impurity sector differently and therefore generate different infrared phenomena. Representative results include multipolar Kondo exchange and two-stage screening in the three-orbital Kanamori problem, and channel-limited screening, frustration-induced non-Fermi-liquid behavior, and Kosterlitz–Thouless criticality in trimer geometries [1606.07654], [2008.09034].

## 1. Terminological scope and model variants

A persistent ambiguity in usage arises because “three-impurity” may refer either to **three correlated sites** or to **one impurity with three orbitals**. The three-orbital formulation studied in "Low-energy physics of three-orbital impurity model with Kanamori interaction" is a single-impurity Anderson model whose impurity interaction has orbital \(SO(3)\) and spin \(SU(2)\) symmetry and is intended to describe systems with partially occupied \(t_{2g}\) shells [1606.07654]. By contrast, "Strongly correlated multi-impurity models: The crossover from a single-impurity problem to lattice models" treats a genuine **multi-impurity** setting in which several correlated orbitals form a cluster coupled to a host, with the trimer as the minimal nontrivial case [2008.09034].

The distinction matters because the low-energy degrees of freedom are not the same. In the three-orbital single-impurity problem, the impurity carries spin, orbital dipole, and orbital quadrupole moments, and the effective Kondo problem is correspondingly multipolar. In the trimer, the decisive objects are the **effective inter-impurity hoppings** generated by the real part of the hybridization self-energy and the **number of independent screening channels** determined by the rank of its imaginary part. A common misconception is therefore to treat the two problems as interchangeable. The literature summarized here instead supports a narrower conclusion: they belong to the same broad Anderson-impurity family, but their microscopic content and infrared mechanisms are different.

## 2. Three-orbital impurity with Kanamori interaction

The three-orbital Anderson impurity model is defined by
\[
H = H_{\mathrm{bath}} + H_{\mathrm{hyb}} + H_{\mathrm{imp}},
\]
with a flat conduction bath, impurity-bath hybridization, and a local interaction of Kanamori form. In the notation of the cited work,
\[
H_{\mathrm{bath}} = \sum_{k,m,\sigma} \epsilon_k\, c^\dagger_{km\sigma} c_{km\sigma},
\]
\[
H_{\mathrm{hyb}} = V \sum_{m,\sigma} c^\dagger_{m\sigma} d_{m\sigma} + \mathrm{h.c.},
\]
and
\[
H_{\mathrm{imp}} = \frac{1}{2}(U-3J)N_d(N_d-1) - 2J\,\mathbf{S}^2 - \frac{J}{2}\mathbf{L}^2 + \epsilon_0 N_d.
\]
Here \(N_d\) is the impurity occupancy, \(\mathbf{S}\) the total impurity spin, \(\mathbf{L}\) the total impurity orbital angular momentum, \(U\) the Hubbard repulsion, and \(J\) Hund’s coupling. The conduction band is taken as flat with density of states \(\rho_0=1/(2D_0)\), and the hybridization width is \(\Gamma=\pi \rho_0 V^2\) [1606.07654].

The orbital sector has \(SO(3)\) symmetry rather than bare \(SU(3)\) symmetry. For \(M=3\) orbitals, the natural impurity multipoles are the spin \(\mathbf{S}\), the orbital dipole \(\mathbf{L}\), and the orbital quadrupoles \(\mathbf{Q}\). The orbital generators \(\{L,Q\}\) span the eight generators of \(SU(3)\), even though the bare Hamiltonian has only orbital \(SO(3)\) symmetry. The same work introduces an interpolation parameter \(\alpha\) connecting the **Dworin–Narath** model \((\alpha=0)\) and the **Kanamori** model \((\alpha=1)\),
\[
H_{\mathrm{imp}} = -2J\mathbf{S}^2 - \alpha\frac{J}{2}\mathbf{L}^2 + \frac{U-3J}{2}N_d(N_d-1) + \epsilon_0 N_d,
\]
which is useful for isolating the effect of reduced orbital symmetry.

The physically emphasized regime is the two-electron sector \(N_d=2\), relevant to Hund’s metals. In that sector the atomic ground state has
\[
L=1,\qquad S=1,
\]
i.e. the Hund’s-rule high-spin state. This regime is singled out because the local moment is relatively large and its screening is strongly suppressed, producing low coherence scales and strong-correlation effects.

## 3. Schrieffer–Wolff reduction and multipolar Kondo structure

The low-energy physics of the three-orbital Kanamori problem is derived by a **Schrieffer–Wolff transformation** that integrates out charge fluctuations and projects onto the impurity ground-state multiplet. The general form given in the paper is
\[
H_K = -P_n H_{\mathrm{hyb}}
\left( \sum_a \frac{P_{n+1}^a}{\Delta E_{n+1}^a} + \sum_b \frac{P_{n-1}^b}{\Delta E_{n-1}^b} \right)
H_{\mathrm{hyb}} P_n,
\]
where \(P_n\) projects onto the impurity ground multiplet with \(n=N_d\) electrons and \(P_{n\pm1}^{a,b}\) onto virtual charge-excited multiplets [1606.07654].

For the Kanamori case this yields a generalized Kondo Hamiltonian
\[
H_K = J_p N_f + J_s\,\mathbf{S}\cdot\mathbf{s} + J_l\,\mathbf{L}\cdot\mathbf{l} + J_q\,\mathbf{Q}\cdot\mathbf{q}
+ J_{ls}\,(\mathbf{L}\otimes\mathbf{S})\cdot(\mathbf{l}\otimes\mathbf{s})
+ J_{qs}\,(\mathbf{Q}\otimes\mathbf{S})\cdot(\mathbf{q}\otimes\mathbf{s}).
\]
The bath operators entering this expression are the local bath charge density \(N_f\), the bath spin \(\mathbf{s}\), the bath orbital dipole \(\mathbf{l}\), and the bath orbital quadrupole \(\mathbf{q}\). The low-energy problem is therefore not a purely spin-exchange Kondo model; it contains **spin, orbital, quadrupolar, and mixed spin-orbital exchange**.

A central quantitative result is the hierarchy of bare couplings in the \(N_d=2\), \(L=1\), \(S=1\) ground state. The spin-spin coupling \(J_s\) is much smaller than the orbital couplings for most parameters and can even become **ferromagnetic** near the \(N_d=2 \leftrightarrow 3\) valence-fluctuation regime, whereas \(J_l\), \(J_q\), \(J_{ls}\), and \(J_{qs}\) are comparatively large. The hierarchy is summarized as
\[
|J_s| \ll J_l, J_q, J_{ls}, J_{qs}.
\]
Because the Kanamori interaction has only orbital \(SO(3)\) symmetry, the orbital-dipole and orbital-quadrupole couplings are split:
\[
J_q - J_l = \Delta J/2,\qquad J_{qs} - J_{ls} = -\Delta J,
\]
with \(\Delta J \propto \alpha J\) and vanishing when the orbital Hund term is absent. This splitting distinguishes the bare Kanamori problem from the \(SU(3)\)-symmetric Dworin–Narath limit.

## 4. Renormalization, dynamical \(SU(3)\), and two-stage screening

The perturbative renormalization-group analysis is formulated in terms of one-loop scaling equations
\[
\beta_i = \frac{dJ_i}{d\ln D}.
\]
Among the explicit equations given are
\[
\beta_p = 0,
\]
\[
\beta_s = -\frac{1}{9}\left(3J_{ls}^2 + 5J_{qs}^2 + 9J_s^2\right),
\]
\[
\beta_l = -\frac{1}{16}\left(4J_l^2 + 3J_{ls}^2 + 5(4J_q^2 + 3J_{qs}^2)\right),
\]
together with corresponding equations for \(J_q\), \(J_{ls}\), and \(J_{qs}\) [1606.07654]. The qualitative implication emphasized in the paper is that couplings related to orbital degrees of freedom renormalize faster than those associated primarily with spin degrees of freedom.

This flow structure produces a **two-stage screening process**. Orbital moments are screened first at a higher scale, while spin moments are screened later at a much lower scale. The orbital Kondo temperature is larger, whereas the spin Kondo temperature is often an order of magnitude smaller; for \(J\) above the zero-Hund Kondo scale, the separation becomes pronounced. The slow quenching of the local spin moment is thus traced both to the smallness or even ferromagnetic sign of \(J_s\) and to the faster renormalization of the orbital sector.

A major conceptual result is that the distinction between orbital \(SO(3)\) and orbital \(SU(3)\) symmetry becomes irrelevant in the infrared. Even when the bare couplings are split, the renormalization-group flow drives
\[
J_l \to J_q,\qquad J_{ls} \to J_{qs},
\]
so that the splittings of coupling constants flow to zero. The paper concludes that the Kanamori and Dworin–Narath problems therefore describe the **same low-energy physics** and flow to the same **low-energy Fermi-liquid fixed point**.

The perturbative picture is corroborated by **numerical renormalization group (NRG)** calculations. At low temperatures, the finite-size spectra for Kanamori and Dworin–Narath become identical, confirming the same Fermi-liquid fixed point. The spin susceptibility stays larger and saturates later than the orbital susceptibility, directly exhibiting slower spin screening. The occupancy dependence is also significant: near \(N_d=3\), spin screening becomes especially suppressed because the local atomic state is \(L=0, S=3/2\), whereas near \(N_d \lesssim 1\), Hund’s coupling becomes unimportant.

## 5. Low-energy mapping of the genuine three-impurity trimer

For the genuine three-impurity Anderson model, the starting point is a multi-impurity Hamiltonian
\[
H = H_{\rm host} + H_{\rm corr} + H_{\rm hyb},
\]
with a host conduction band, a correlated impurity cluster, and hybridization between them [2008.09034]. The crucial object is the matrix hybridization function
\[
\Delta_{lm,\sigma}(z),
\]
which encodes the influence of the host conduction band on the impurity cluster. In the low-energy, wide-band limit, the central decomposition is
\[
\Delta_{lm,\sigma}(-i0^+) = {\rm Re}\,\Delta_{lm,\sigma}(-i0^+) + i\,\Gamma_{lm,\sigma}.
\]

The **real part** is absorbed into the cluster Hamiltonian as an effective hopping matrix,
\[
t^{\rm eff}_{lm,\sigma} = {\rm Re}\Delta_{lm,\sigma}(-i0^+),
\]
and thereby generates delocalization and **antiferromagnetic RKKY-like exchange** within the cluster. The paper’s formulation is explicit that this mechanism arises directly from the conduction-band structure, without introducing ad hoc Heisenberg couplings. The **imaginary part**
\[
\mathbf{\Gamma}_\sigma = {\rm Im}\,\mathbf{\Delta}_\sigma(-i0^+)
\]
defines the hybridization matrix governing charge fluctuations and screening. After diagonalization, its nonzero eigenvalues are interpreted as the couplings of transformed cluster orbitals to **effective conduction channels**.

This leads to a classification of multi-impurity models into two kinds. In the **first kind**,
\[
{\rm rank}(\mathbf{\Gamma}_\sigma)=N_f,
\]
so there are enough independent screening channels to fully quench all impurity moments by a multi-stage Kondo process. In the **second kind**,
\[
{\rm rank}(\mathbf{\Gamma}_\sigma)<N_f,
\]
so Kondo screening alone is insufficient and the cluster must form its singlet or residual low-energy state through inter-impurity correlations. The paper identifies this distinction as a rigorous replacement for a phenomenological exhaustion criterion and uses it to organize the low-energy physics of finite impurity arrays.

For the trimer with \(N_f=3\), this framework is already rich enough to accommodate magnetic frustration, competition between Kondo screening and intersite exchange, underscreened intermediate-spin regimes, and frustration-induced non-Fermi-liquid fixed points. The decisive control parameters are no longer merely the bare microscopic couplings, but the interplay between the effective hopping generated by \({\rm Re}\Delta\) and the number of effective screening channels encoded in \({\rm rank}(\Gamma)\).

## 6. Frustration, non-Fermi-liquid regimes, and KT criticality in trimers

The trimer acquires especially interesting structure in two settings emphasized by the cited work. In the idealized **equilateral-triangle** or \(C_3\)-symmetric case, the hybridization matrix is constrained to two independent parameters, \(\Gamma^{\rm off}\) and \(t^{\rm eff}\), and the rank remains \(3\) in the parameter regime studied, so this is formally a **first-kind** trimer [2008.09034]. The low-energy behavior then depends sensitively on the magnitude of the effective hopping. Small \(t^{\rm eff}\) gives a fully screened Fermi liquid, intermediate \(t^{\rm eff}\) gives a **frustrated Kondo NFL regime**, and large \(t^{\rm eff}\) destroys the NFL fixed point and restores a Fermi liquid through a crossover.

Within this \(C_3\)-symmetric setting, the paper reproduces two known NFL phases: the **Frustrated Kondo regime**, which is stable against moderate particle-hole asymmetry, and the **Isospin Kondo regime**, which is unstable to particle-hole asymmetry. The mapped Anderson formulation makes clear that these regimes are fragile because the same effective hopping that generates frustration also breaks the symmetry needed to protect the isospin Kondo fixed point.

The more channel-limited situation arises for a **finite trimer cluster in 1d**. There the paper states that
\[
{\rm rank}(\mathbf{\Gamma}_\sigma)\le 2,
\]
so the three-impurity problem is generically of the **second kind**. In this regime the effective bath cannot screen all three impurity spins independently. At or near half filling, the trimer may enter an \(S_{\rm imp}=0\) low-temperature phase, but as the filling is shifted toward the band edges, the cluster develops progressively larger ferromagnetic alignment and unscreened moments. The reported sequence for \(N_f=3\) is a low-temperature regime with singlet behavior, then a transition to a regime where the cluster behaves like an effective spin-\(1/2\), and ultimately, near the band edge, an impurity entropy
\[
S_{\rm imp} = k_B \ln(3),
\]
corresponding to a residual triplet-like degeneracy in the effective low-energy description.

The physical origin of trimer frustration is traced to competition among the effective hoppings \(t_{ij}^{\rm eff}\) generated by \({\rm Re}\Delta_{ij}\). At a special filling these hoppings become comparable, producing maximal magnetic frustration. The isolated cluster then shows a crossing of two different doublets; in the full problem this appears as a cusp in the low-energy crossover scale. Near that point, the NRG data are consistent with **Kosterlitz–Thouless-type quantum phase transitions**. The crossover scale is extracted from the impurity entropy through
\[
S_{\rm imp}(T_0)=\frac{1}{2}\left(S^{\rm LM}_{\rm imp}+S^{\rm LT}_{\rm imp}(0)\right),
\]
and near the critical point obeys
\[
T_0(\epsilon_c)= C_1 \exp\!\left(\frac{C_2}{\sqrt{\epsilon_c^c-\epsilon_c}}\right).
\]
The interpretation given is that an effective collective spin-\(\tfrac{1}{2}\) becomes decoupled from the continuum because ferromagnetic cluster correlations suppress its final-stage screening.

## 7. Conceptual synthesis and significance

Taken together, these results show that the phrase **Three-Impurity Anderson Model** names a small but conceptually diverse class of impurity problems rather than a single universal Hamiltonian. In the three-orbital single-impurity realization, the low-energy description is a **multipolar Kondo model** in which orbital and quadrupolar exchanges dominate over spin exchange, producing two-stage screening and a strongly suppressed spin scale. In the trimer realization, the low-energy description is a **cluster-plus-effective-baths model** in which the real part of the host self-energy generates inter-impurity exchange while the rank of the imaginary part limits the available screening channels [1606.07654], [2008.09034].

The two strands of the literature also identify different symmetry mechanisms. In the Kanamori problem, the central result is **dynamical restoration of orbital \(SU(3)\)** at low energies despite a bare \(SO(3)\) interaction. In the trimer, by contrast, symmetry can stabilize or destabilize non-Fermi-liquid regimes, but the decisive distinction is between first-kind and second-kind models, i.e. whether the number of effective screening channels is sufficient to quench the cluster moments.

A plausible implication is that precise usage should distinguish between **three-orbital Hund impurity physics** and **three-site cluster impurity physics**. The former is organized by multipolar exchange and separated orbital and spin Kondo scales; the latter by frustration, effective hopping, and channel-counting constraints. Within that narrower terminology, the literature summarized here presents the trimer as the minimal setting where frustration, limited screening, and collective-moment formation can generate NFL or KT behavior, and the three-orbital Kanamori model as the canonical impurity realization of Hund’s-metal screening with dynamically emergent \(SU(3)\) infrared structure.

Source: https://www.emergentmind.com/topics/three-impurity-anderson-model