betaFeedback

WorksPhysics and Astronomy

Non-magnetic impurities as a quantum microscope of finite size effects in spin chains

Abstract

Finite-size effects are ubiquitous in quantum matter, yet typically invisible in the thermodynamic limit; when system size is restricted, boundaries and confinement reveal modified spectra, correlations, and emergent phenomena. With growing access to well-controlled few-spin systems, these effects are now directly observable and offer a route to probing universal physics and enabling quantum information applications. Here, we study finite-size physics in the quasi-one-dimensional Ising-like quantum antiferromagnet Ba(Co1−x_{1-x}Mgx_x)2_2V2_2O8_8 with x=2%x = 2\% and x=5%x = 5\%, where non-magnetic Mg2+^{2+} ions fragment the Co2+^{2+} spin chains into finite segments. Combining inelastic neutron scattering and numerical calculations, we identify distinct regimes with fundamentally different low-energy quantum excitations: spinons confined to chain boundaries, and a cluster mode in shorter chains corresponding to a collective reversal of the Néel-ordered segment without classical analogue. Our results demonstrate how chemical substitution acts as a quantum microscope providing direct access to boundary-shaped excitations in strongly correlated spin systems.

Introduction

Quantum spin chains constitute a central platform for exploring strong correlations and emergent quasiparticles in one dimension. Their reduced dimensionality enhances quantum fluctuations, suppresses classical long-range order, and stabilizes excitation continua arising from fractionalized spinons—mobile domain walls whose neutron-scattering signatures differ fundamentally from conventional magnons [7, 9, 6]. Although most theoretical insights originate from the thermodynamic limit, growing attention has shifted toward finite chains, where boundaries, discrete spectra, and length-dependent coherence give rise to qualitatively new physics.

Beyond their role in model magnetism, finite spin chains are also relevant for quantum information science. Entire segments can encode and transmit quantum information through collective degrees of freedom, enabling proposals for coherent state transfer without single-site control [41, 35, 36, 42, 44, 43]. The transverse-field Ising model (TFIM), a cornerstone of quantum many-body physics [32, 48], provides a paradigmatic example: via the Jordan–Wigner transformation, it maps onto free fermions, allowing an exact description of quantum criticality and serving as a benchmark for quantum simulation. This has motivated high-fidelity experimental realizations in Rydberg arrays, trapped ions, and cold atoms [60, 45, 47, 46, 52, 51], where finite size is not a correction but a defining feature governing entanglement and low-energy structure.

The Jordan–Wigner transformation also provides a natural interface with 1D topological models such as the Su-Schrieffer-Heeger model, which first predicted topological edge states and the Kitaev chain, where these boundary modes become Majorana zero modes [54, 53, 50]. In particular, the TFIM and XYXY models map to quadratic fermions with pairing terms, linking magnetic boundary modes to Majorana-like physics [38, 55]. Away from these limits, additional interactions, as in XXZXXZ systems for instance, move the model beyond the free-fermion regime, but boundary physics can persist in the form of edge modes, gapped excitations, and parity-dependent splittings [39, 37, 40]. Spin chains therefore offer a versatile setting to explore boundary-localized excitations beyond idealized one-dimensional topological models.

Controlled realizations of finite chains are now achievable in the solid state; atomically precise nanographene synthesis enables chains with tunable length and edge termination [27, 59]. A highly flexible alternative is provided by nonmagnetic doping in bulk quantum magnets, which segments long chains into finite pieces. Owing to their one-dimensional nature, such systems are remarkably sensitive to intrachain disorder [16]. Theory predicts even-odd effects [30], impurity-pinned edge modes [19, 31], and boundary-induced Kondo phenomena [28], while experimentally, impurity segmentation produces notable modifications of ground states and excitation spectra in spin-1/2 and Haldane chains [13, 14, 15, 12, 11, 17, 10, 20, 29],

In this study, we investigate finite-size quantum effects in the Ising-like (XXZXXZ) S=1/2S=1/2 antiferromagnetic chain compound BaCo2_{2}V2_{2}O8_{8}. By replacing a fraction of the magnetic sites with non-magnetic impurities, we segment the chains into finite pieces whose length distribution is tunable through impurity density. Combining inelastic neutron scattering in this material with numerical simulation of the corresponding model, we identify three distinct low-energy regimes governed by the segmented chain length. While the shortest-chain Anderson-tower regime requires strong dilution and is accessed theoretically, our experiments access the intermediate and long-chain regimes. In these regimes, the lowest-energy excitations take the form of collective cluster modes – a coherent reversal of the Néel states – at intermediate lengths, and of edge-localized spinon modes stabilized by open boundaries in longer chains. These results demonstrate how impurity engineering in bulk materials provides a powerful route to resolve finite-size quantum phenomena typically accessible only in engineered nanoscale or quantum-simulator platforms.

Results

BaCo2_{2}V2_{2}O8_{8}: a quasi-one-dimensional Ising quantum antiferromagnet

Before presenting our results, we first introduce the key properties of the pure BaCo2_{2}V2_{2}O8_{8} system. It consists of one-dimensional screw chains of Co2+^{2+} ions aligned along the crystallographic cc-axis. These Co2+^{2+} ions carry an effective spin-12\tfrac{1}{2} and are coupled via antiferromagnetic (AF) exchange interactions [Figure 1a]. Owing to a pronounced Ising-like anisotropy and weak interchain interactions, BaCo2_{2}V2_{2}O8_{8} stabilizes a Néel-ordered AF ground state below the transition temperature TN=5.5T_{\mathrm{N}} = 5.5 K, with magnetic moments aligned along the cc-axis [2] [Figure 1b–c]. Unlike conventional antiferromagnetic spin chains, which host a continuum of spinon excitations, BaCo2_{2}V2_{2}O8_{8} exhibits a discrete sequence of gapped excitations, so-called Zeeman ladders. They correspond to bound states of spinons confined by the effective staggered field emerging from the Néel order and the interchain coupling of BaCo2_{2}V2_{2}O8_{8} [Figure 1f]. The Zeeman ladders comprise both transverse modes, corresponding to fluctuations perpendicular to the ordered moments and involving an odd number of spin flips, and longitudinal modes associated with fluctuations parallel to the ordered moment direction and involving an even number of spin flips [Figure 1d-e]. We denote the transverse and longitudinal modes as ∣NT⟩\ket{N T} and ∣NL⟩\ket{N L}, respectively. In the Ising limit, these modes are associated with the number of flipped spins NN. The transverse mode contributes to the transverse spin dynamical susceptibility Sxx{\cal S}_{xx} (equivalently Syy{\cal S}_{yy}), and the longitudinal modes contributes to the longitudinal spin dynamical susceptibility Szz{\cal S}_{zz}. The dynamical susceptibilities are the Fourier transforms of the corresponding spin pair correlation functions and can be measured by neutron scattering [Figure 1d–f]. These excitations were previously observed in both BaCo2_{2}V2_{2}O8_{8} and its structurally analogue compound SrCo2_2V2_2O8_8 [1, 3, 4].

The spin model Hamiltonian of BaCo2_{2}V2_{2}O8_{8} consists of intrachain and interchain parts [57]. The former reads

Hintra=J∑j∑μ[ϵ(Sj,μxSj+1,μx+Sj,μySj+1,μy)+Sj,μzSj+1,μz],\begin{align} \mathcal{H}_{\mathrm{intra}} = J \sum_{j}\sum_{\mu} \left[\epsilon \left(S_{j,\mu}^{x}S_{j+1,\mu}^{x}+S_{j,\mu}^{y}S_{j+1,\mu}^{y}\right) +S_{j,\mu}^z S_{j+1,\mu}^z\right], \tag*{(1)} \end{align}

where JJ is the intrachain spin coupling, ϵ\epsilon the Ising-like anisotropy, and Sj,μ\mathbf{S}_{j,\mu} a spin-1/2 operator on jj-th site of μ\mu-th chain. The interchain part reads:

Hinter=J′∑j∑⟨ν,μ⟩Sj,μSj,ν,\begin{align} \mathcal{H}_{\mathrm{inter}} = J'\sum_{j}\sum_{\braket{\nu,\mu}} \boldsymbol{S}_{j,\mu} \boldsymbol{S}_{j,\nu}, \tag*{(2)} \end{align}

where J′J' is the interchain spin coupling and ∑⟨μ,ν⟩\sum_{\braket{\mu,\nu}} denotes the summation over the pairs of nearest neighbor chains. For BaCo2_{2}V2_{2}O8_{8}, we used the parameters J=5.8J = 5.8 meV, ϵ=0.53\epsilon = 0.53, and J′=0.17J' = 0.17 meV, determined previously since they best reproduce the excitation spectra of the undoped compound [57]. BaCo2_{2}V2_{2}O8_{8} is a quasi-1D system satisfying J≫J′J \gg J'.

For the analysis, we treat the interchain coupling by the mean field theory. In the presence of the Néel order along the cc axis, Eq. (2) is approximated as Hinter≃J′∑j∑⟨μ⟩Sj,μ⟨Sj,ν⟩\mathcal{H}_{\mathrm{inter}} \simeq J'\sum_{j}\sum_{\braket{\mu}} \boldsymbol{S}_{j,\mu} \braket{\boldsymbol{S}_{j,\nu}}. Although the actual structure of microscopic interchain coupling is likely to be more complex, this modeling was shown to give an excellent description of the pure system up to very large magnetic fields [56, 58]. It also allows to deal with a 1D problem.

Realization of finite-size spin chains by non-magnetic impurities

We now consider the effects of doping by substituting magnetic Co2+^{2+} ions with non-magnetic Mg2+^{2+} ions. We assume that this substitution fragments the spin chains into finite segments of various lengths [Figure 1g], each described by a Hamiltonian with open boundary conditions:

HN=J∑j=1N−1[ϵ(SjxSj+1x+SjySj+1y)+SjzSj+1z]+hMF∑j=1N(−1)jSjz,\begin{align} \mathcal{H}_{\mathcal{N}} = J \sum_{j=1}^{\mathcal{N}-1} \left[\epsilon \left(S_{j}^{x}S_{j+1}^{x}+S_{j}^{y}S_{j+1}^{y}\right) +S_{j}^{z} S_{j+1}^{z}\right] +h_{\mathrm{MF}} \sum_{j=1}^{\mathcal{N}}(-1)^{j}S_{j}^{z}, \tag*{(3)} \end{align}

with hMF=J′∣⟨Sjz⟩∣h_{\mathrm{MF}}=J'|\braket{S_{j}^{z}}| the effective staggered magnetic field from neighboring chains. The probability that a chain segment has a length N\mathcal{N}, given that xx is the Mg2+^{2+} concentration, is obtained by the geometric distribution [18, 22], P(N)=x(1−x)NP(\mathcal{N}) = x(1-x)^{\mathcal{N}}, with mean Nˉ≈1/x\bar{\mathcal{N}} \approx 1/x in the dilute limit [Figure 1h].

Figure 1. Magnetic model for BaCo2_{2}V2_{2}O8_{8} and its Mg-doped variant. (a) One-dimensional screw chain of edge-sharing Co2+^{2+} octahedra along the crystallographic cc-axis in BaCo2_{2}V2_{2}O8_{8}. (b) Projection onto the (a,b)(a,b) plane, showing neighboring chains; the complex interchain exchange paths yield the average effective antiferromagnetic diagonal coupling J′J' (dashed lines). (c) Minimal magnetic model consisting of two antiferromagnetic S=12S=\tfrac{1}{2} chains with intrachain coupling JJ and effective interchain coupling J′J' (spins rotated by 90∘90^{\circ} for clarity). (d,e) Lowest confined excitations: the first transverse mode ∣1T⟩\ket{1T} and the first longitudinal mode ∣2L⟩\ket{2L} with ΔSz=±1\Delta S^{z}=\pm 1 and 00 respectively with respect to the ground state. (f) Schematic representation of Zeeman-ladder spectra of confined spinons, showing the transverse response Sxx{\cal S}_{xx} (top) and the combined response Sxx+Szz{\cal S}_{xx}+{\cal S}_{zz} (bottom). The dashed line corresponds to the gapped spinon-continuum expected for an Ising spin chain in absence of interchain coupling. (g) Mg substitution segments the chains into finite lengths N\cal{N}, suppressing exchange near the boundaries (J→0J\rightarrow0) and realizing open boundary conditions. (h) Geometrical distribution of segment lengths for diluted spin chains. Plain and dashed lines denote P(N)P(\cal{N}) and NP(N)\cal{N} P(\cal{N}) respectively while green and purple colors denote x=2%x = 2\% and x=5%x = 5\% respectively.

To experimentally realize such finite-size chain segments, Ba(Co1−x_{1-x}Mgx_x)2_2V2_{2}O8_{8} single crystals at x=2%x = 2\% (⇔Nˉ=50\Leftrightarrow \mathcal{\bar{N}} = 50) and 5%5\% (⇔Nˉ=20\Leftrightarrow \mathcal{\bar{N}} = 20) were synthesized via the optical floating-zone technique [5] and characterized by neutron diffraction (see Supplementary Material). Both the crystallographic structure and the zero-field antiferromagnetic order remain essentially intact upon doping, confirming that Mg2+^{2+} substitution cleanly realizes open boundary conditions without structural distortion. The Néel temperature and ordered magnetic moment decrease linearly with xx, from TN≃5.5T_{\mathrm{N}} \simeq 5.5 K and 2.2 μB2.2~\mu_{\mathrm{B}} in the parent compound to TN=3.8T_{\mathrm{N}} = 3.8 K and 1.3 μB1.3~\mu_{\mathrm{B}} at x=5%x = 5\%, consistent with previous macroscopic characterizations [8].

Emergence of localized excitations revealed by inelastic neutron scattering

Figure 2(a–b) present the INS intensity maps along the intrachain direction Q=(2,0,QL)\boldsymbol{Q} = (2, 0, Q_{L}) for x=5%x = 5\% (Nˉ=20\bar{\mathcal{N}}=20) and 2%2\% (Nˉ=50\bar{\mathcal{N}}=50) at T=1.5T = 1.5 K. For comparison, the undoped compound [Figure 2(c)] displays the Zeeman ladder spectrum with a spin gap Δ≃1.75\Delta \simeq 1.75 meV at the antiferromagnetic (AF) point Q=(2,0,1)\boldsymbol{Q} = (2, 0, 1) [1]. In the doped samples, the Zeeman ladder modes persist: magnetic modes corresponding to the first transverse and longitudinal excitations are indeed observed around 1.7 and 1.9 meV for x=2%x = 2\%, and merged in a broad excitation around 1.6 to 1.8 meV for x=5%x = 5\% at the AF point [Figure 2(d–e). They disperse up to a maximum energy Em≃4E_m \simeq 4 meV at Q=(2,0,1.5)\boldsymbol{Q} = (2, 0, 1.5). More strikingly, three additional features emerge below the first Zeeman ladder at around 0.7, 1.0, and 1.2 meV [Figure 2(d–e)], with intensity increasing with energy and fading away from AF and zone-center points. The absence of dispersion indicates a localized origin.

Figure 2. Spin dynamics of Ba(Co1−x_{1-x}Mgx_x)2_2V2_{2}O8_{8}. (a–c) Inelastic neutron scattering intensity maps of the intrachain dispersion measured along Q=(2,0,QL)\boldsymbol{Q}=(2,0,Q_{\rm L}) at T=1.5T=1.5 K for x=5%x=5\%, 2%2\%, and 00, corresponding to chains with average lengths Nˉ≃20\bar{\mathcal{N}}\simeq20, 5050, and ∞\infty, respectively. Although the Brillouin zones displayed for the doped and undoped compounds differ, they are crystallographically equivalent. For clarity, the spurious (parasitic) intensity previously reported in Ref. [58], extending from Q≈(2,0,1.1)\mathbf{Q}\approx(2,0,1.1) at E≈0.5E\approx0.5 meV to Q≈(2,0,1.25)\mathbf{Q}\approx(2,0,1.25) at E≈1.1E\approx1.1 meV, has been removed (see the Supplemental Material for details). (d–f) Enlarged views of panels (a–c) in the low-energy range between 0 and 3 meV. (g–i) Representative constant-Q\boldsymbol{Q} energy scans used to construct the dispersion maps (taken at the vertical dashed lines in panels (a,b)). In the doped samples, additional nearly non-dispersive modes emerge at approximately 0.70.7, 1.01.0, and 1.21.2 meV (indicated by the dashed black arrows in panels (g,h)), whereas the undoped compound exhibits only the Zeeman ladder excitations. Owing to the absence of magnetic intensity below the first transverse excitation ∣1T⟩\ket{1T} observed in (i), the full low-energy dispersion was not measured for the undoped sample in panel (c).

To characterize the transverse and longitudinal nature of these excitations, we compare INS spectra along Q=(0,0,QL)\boldsymbol{Q} = (0, 0, Q_{\rm L}), sensitive to Sxx{\cal S}_{xx} only, and Q=(2,0,QL)\boldsymbol{Q} = (2, 0, Q_{\rm L}), probing both Sxx{\cal S}_{xx} and Szz{\cal S}_{zz}, with density matrix renormalization group (DMRG) calculations [Figure 3(a–d)]. The calculations successfully reproduce the experimental dispersion maps, including the low-energy non-dispersive features and their momentum dependence. A direct comparison of energy scans at Q=(0,0,2)\boldsymbol{Q} = (0, 0, 2) and Q=(2,0,1)\boldsymbol{Q} = (2, 0, 1) is shown in Figure 3(e–h). The main spectral features are well reproduced: at Q=(0,0,2)\boldsymbol{Q} = (0, 0, 2) [Figure 3e], a localized transverse mode is calculated around 1.2 meV (compared to 1 meV in the experiment), below the first transverse Zeeman ladder at 1.9 meV (compared to 1.7 meV); at Q=(2,0,1)\boldsymbol{Q} = (2, 0, 1) [Figure 3g], two additional longitudinal modes appear at 0.7 (compared to 0.7 meV) and 1.2 meV (compared to 1.4 meV). While the overall agreement is excellent, some discrepancies remain in the mode energies, slightly shifted upward in the calculations, and in relative intensities. This may arise from (i) mean-field treatment of the interchain coupling, neglecting weak interchain dispersion [57]; (ii) impurity effects accounted for only in the main chain, overestimating magnetic order due to interchain coupling; (iii) JJ and J′J' fixed from the undoped compound, although doping likely affects these parameters, as suggested by the reduction of the ordering temperature and first transverse Zeeman gap when increasing xx [Figure 2(a–c) and Supplementary Material].

Figure 3. Transverse and longitudinal character of the emergent modes: (a-b) Inelastic neutron scattering intensity maps of the intrachain dispersion along Q=(0,0,QL)\boldsymbol{Q} = (0, 0, Q_L) and Q=(2,0,QL)\boldsymbol{Q} = (2, 0, Q_L) for x=2%x = 2\% (Nˉ≃50\bar{\mathcal{N}} \simeq 50) and measured at T=1.5T=1.5 K. The former is sensitive to transverse fluctuations Sxx{\cal S}_{xx} only while the later is related to both transverse Sxx{\cal S}_{xx} and longitudinal Szz{\cal S}_{zz} fluctuations. (c-d) Numerically calculated DMRG (see text) corresponding dispersion maps. (e-h) Comparison between experimental (blue points) and calculated (orange points) intensity for Q\mathbf{Q}-constant energy scans at Q=(0,0,2)\boldsymbol{Q} = (0, 0, 2) (e,g) and Q=(2,0,1)\boldsymbol{Q} = (2, 0, 1) (f,h) (corresponding to the vertical dashed lines in panels (a–d)).

Cluster and edge excitations: even-odd parity and chain-length dependence

To investigate the impurity effects in detail, we performed exact diagonalization (ED) and DMRG calculations on finite-length chains with open boundary conditions. Figure 4b shows the numerically calculated sum of transverse Sxx{\cal S}_{xx} and longitudinal Szz{\cal S}_{zz} dynamical susceptibilities as a function of the chain length N\mathcal{N}, with the individual components for all, even, and odd N\mathcal{N} shown in Figure 4(c–h). Low-energy modes are observed in the spectra. While the states depicted in Figure 4a are well-defined configurations in the Ising limit, BaCo2_{2}V2_{2}O8_{8} lies away from this limit (ϵ=0.53\epsilon=0.53) and the transverse XXXX exchange term mixes different spin configurations, as is also the case for the Zeeman ladders. Consequently, the actual eigenstates are linear combinations of several basis states. Here, for clarity, we restrict the discussion to the Ising-limit configurations shown in Figure 4a, which provide a simple picture of the physics at play. The detailed weights of the different configurations and their evolution with N\mathcal{N} are presented in the Supplementary Material.

For N=even\mathcal{N}=\mathrm{even} chains, the two Néel-ordered configurations ∣↑↓⋯↑↓⟩\ket{\uparrow\downarrow\cdots\uparrow\downarrow} and ∣↓↑⋯↓↑⟩\ket{\downarrow\uparrow\cdots\downarrow\uparrow} both belong to the Stotz=0S_{\mathrm{tot}}^{z}=0 sector and are entangled by the XYXY exchange term, which induces quantum fluctuations via successive two-spin flips ↑↓  ⇔  ↓↑\uparrow\downarrow \;\Leftrightarrow\; \downarrow\uparrow. This hybridization is maximal for short chains: for N=2\mathcal{N}=2 and small interchain mean-field term hMFh_{\mathrm{MF}}, it yields the antibonding singlet ground state ∣ABN=2⟩≃(∣↑↓⟩−∣↓↑⟩)/2\ket{\mathrm{AB}_{\mathcal{N}=2}} \simeq (\ket{\uparrow\downarrow}-\ket{\downarrow\uparrow})/\sqrt{2} and the bonding triplet excited state ∣BN=2⟩≃(∣↑↓⟩+∣↓↑⟩)/2\ket{\mathrm{B}_{\mathcal{N}=2}} \simeq (\ket{\uparrow\downarrow}+\ket{\downarrow\uparrow})/\sqrt{2}, both maximally entangled with negligible Néel order. As N\mathcal{N} increases, two effects progressively suppress the entanglement: the XYXY-induced fluctuations require an increasing number of successive two-spin flips to connect the two configurations, and the mean-field term hMFh_{\mathrm{MF}} increasingly splits their energies. As a result, the system displays a behavior reminiscent of Anderson’s tower of excited states [33, 21]: the quantum superpositions ∣ABN⟩\ket{\mathrm{AB}_{\mathcal{N}}} and ∣BN⟩\ket{\mathrm{B}_{\mathcal{N}}} progressively evolve toward classical-like Néel states. The antibonding state ∣ABN⟩\ket{\mathrm{AB}_{\mathcal{N}}} asymptotically approaches the ground state ∣↑↓⋯↑↓⟩\ket{\uparrow\downarrow\cdots\uparrow\downarrow}, stabilized by hMFh_{\mathrm{MF}}, while the bonding state ∣BN⟩\ket{\mathrm{B}_{\mathcal{N}}} asymptotically approaches the cluster state ∣Cℓeven⟩≡∣↓↑⋯↓↑⟩\ket{\mathcal{C\ell}^{\mathrm{even}}} \equiv \ket{\downarrow\uparrow\cdots\downarrow\uparrow}, the reversed Néel configuration with respect to the ground state. Crucially, unlike the ideal Anderson tower scenario where the singlet-triplet gap closes in the thermodynamic limit, here hMFh_{\mathrm{MF}} explicitly breaks the symmetry between the two Néel configurations, so that the gap never closes. As seen in Figure 4f, the gap decreases as ∼1/N\sim 1/\mathcal{N} for N≲14\mathcal{N} \lesssim 14 (disentanglement regime) and reaches a minimum near N∼14\mathcal{N} \sim 14, then grows linearly as ∝NhMF\propto \mathcal{N} h_{\mathrm{MF}} for 14≲N≲3614 \lesssim \mathcal{N} \lesssim 36, reflecting the interchain mean-field energy cost of the transition from ∣GS⟩\ket{GS} to ∣Cℓeven⟩\ket{\mathcal{C\ell}^{\mathrm{even}}}. Since this transition involves ΔStotz=0\Delta S_{\mathrm{tot}}^z = 0, it is visible in Szz\mathcal{S}_{zz}.

For N=odd\mathcal{N}=\mathrm{odd} chains, the two lowest Néel configurations ∣↑↓⋯↑↓↑⟩\ket{\uparrow\downarrow\cdots\uparrow\downarrow\uparrow} and ∣↓↑⋯↓↑↓⟩\ket{\downarrow\uparrow\cdots\downarrow\uparrow\downarrow} belong to different StotzS_{\mathrm{tot}}^z sectors (+1/2+1/2 and −1/2-1/2 respectively) and cannot be connected by any number of two-spin flips, so that they do not entangle. They correspond to the ground state ∣GS⟩\ket{GS} and the cluster state ∣Cℓodd⟩\ket{\mathcal{C\ell}^{\mathrm{odd}}}, already well-defined even for the shortest chains, with a gap growing linearly with N\mathcal{N} throughout N≲36\mathcal{N} \lesssim 36 [Figure 4g]. The transition between these states yields ΔStotz=±1\Delta S_{\mathrm{tot}}^z = \pm 1 and appears in Sxx\mathcal{S}_{xx}.

Edge states arise because locating one domain wall at the chain boundary suppresses the energy cost of a spin-flip excitation, placing them below the Zeeman ladder modes. For N=even\mathcal{N}=\mathrm{even}, the lowest edge mode ∣1Eeven⟩\ket{1\mathcal{E}^{\mathrm{even}}} (ΔStotz=±1\Delta S_{\mathrm{tot}}^z = \pm 1, visible in Sxx\mathcal{S}_{xx}) is generated by a single spin flip from the ground state; its energy decreases with system size up to N∼24\mathcal{N} \sim 24 and then converges to ≈1.2\approx 1.2 meV [Figure 4e]. A second edge mode ∣2Eeven⟩\ket{2\mathcal{E}^{\mathrm{even}}} involves a two-spin-flip configuration (ΔStotz=0\Delta S_{\mathrm{tot}}^z = 0, visible in Szz\mathcal{S}_{zz}) and carries the second longitudinal spectral weight [Figure 4f]. Its energy decreases monotonically with N\mathcal{N} until an anticrossing with ∣Cℓeven⟩\ket{\mathcal{C\ell}^{\mathrm{even}}} occurs near N≃36\mathcal{N} \simeq 36, marking the point at which a single edge spin flip becomes energetically favorable over reversing the entire chain, i.e. when NJ′≃J\mathcal{N}J' \simeq J, giving N≃34\mathcal{N} \simeq 34 for J=5.8J=5.8 meV and J′=0.17J'=0.17 meV, consistent with the numerical result. For longer chains, its energy converges to ≈1.4\approx 1.4 meV. The odd-chain case mirrors this picture: ∣1Eodd⟩\ket{1\mathcal{E}^{\mathrm{odd}}} (ΔStotz=±1\Delta S_{\mathrm{tot}}^z = \pm 1, visible in Sxx\mathcal{S}_{xx}) undergoes an anticrossing with ∣Cℓodd⟩\ket{\mathcal{C\ell}^{\mathrm{odd}}} near N≃36\mathcal{N} \simeq 36 and converges to ≈1.2\approx 1.2 meV [Figure 4g], while the two-spin-flip edge mode ∣2Eodd⟩\ket{2\mathcal{E}^{\mathrm{odd}}} (ΔStotz=0\Delta S_{\mathrm{tot}}^z = 0, visible in Szz\mathcal{S}_{zz}) shows no anticrossing and converges directly to ≈1.4\approx 1.4 meV for N≳24\mathcal{N} \gtrsim 24 [Figure 4h].

Figure 4. Quantum nature of cluster and edge modes revealed by ED and DMRG calculations. (a) Relevant Ising-limit basis configurations in the total-spin sectors Stotz=0,±1S^z_{\mathrm{tot}} = 0, \pm 1 for even N\mathcal{N}, and Stotz=±12S^z_{\mathrm{tot}} = \pm \tfrac{1}{2} for odd N\mathcal{N}. These configurations provide a simple picture of the lowest-energy excitations, whose detailed weights are obtained from ED (see Supplementary Material). In regimes 2 and 3, the ground states ∣GS⟩\ket{\mathrm{GS}} (for both even and odd chains) are dominated by one of the Néel configurations due to symmetry breaking induced by the interchain coupling (see text). Transitions with ΔStotz=±1\Delta S^z_{\rm tot}=\pm1 (ΔStotz=0\Delta S^z_{\rm tot}=0) correspond to transverse (longitudinal) correlations, indicated by blue (red) arrows. (b) DMRG calculations of the summed transverse Sxx{\cal S}_{xx} and longitudinal Szz{\cal S}_{zz} dynamical susceptibilities as a function of the total number of sites N\mathcal{N}, evaluated at Q=(2,0,2)\boldsymbol{Q} = (2, 0, 2). White dotted lines delimit regimes 1–3, and orange arrows mark the experimental impurity concentrations x=2%x = 2\% and x=5%x = 5\%. (c–h) Transverse Sxx{\cal S}_{xx} (upper panels) and longitudinal Szz{\cal S}_{zz} (lower panels) for: (c, d) all lengths, (e, f) even N\mathcal{N}, and (g, h) odd N\mathcal{N}. In (e–h), blue and red open circles denote ED results; lines are guides to the eye. Colored arrows indicate the excitations shown in (a).

From this analysis, three distinct regimes of chain-length dependence emerge, delimited by the white dotted lines in Figure 4b. (1) Short-chain regime (N≲14\mathcal{N} \lesssim 14): the spectrum is strongly parity-dependent; for even N\mathcal{N}, bonding and antibonding states dominate and progressively disentangle, while for odd N\mathcal{N}, the Néel and cluster states already constitute the dominant contributions to the ground state and first excited state, respectively, even for the shortest chain lengths. (2) Intermediate regime (14≲N≲3614 \lesssim \mathcal{N} \lesssim 36): well-defined cluster and edge states coexist with length-dependent energy levels; even/odd distinctions persist but are reduced. (3) Long-chain regime (N≳36\mathcal{N} \gtrsim 36): the anticrossings have occurred, parity dependence vanishes, and the system approaches the thermodynamic limit with size-independent edge mode energies. In the infinite-chain limit, spectral weight transfers entirely to the Zeeman ladder modes ∣1T⟩\ket{1T} and ∣2L⟩\ket{2L} at ≈2.0\approx 2.0 and 2.22.2 meV, and edge-mode weight vanishes.

Neutron spectroscopic identification of finite-chain excitations

Based on the above analysis of our numerical work, which reveals the evolution of the excitation spectrum, from a quantum cluster mode to edge-localized states, we can now interpret the results of INS experiments. In actual doped materials, the experimental dynamical susceptibility naturally includes contributions from segments of different lengths. The measurements performed for x=5%x = 5\% and 2%2\% thus probe two different chain-length regimes. This summation is also taken into account in calculations, each individual chain segment being weighted by the geometric distribution of lengths (and dominated by segments of length Nˉ≈1/x\bar{\mathcal{N}} \approx 1/x). A direct comparison between theory and experiment makes it possible to identify the nature of the observed excitations [Figure 5].

For x=5%x=5\%, corresponding to Nˉ≃20)\bar{\mathcal{N}}\simeq 20) [Figure 5(a,b)], the system lies in the intermediate chain regime, where the excitation spectrum retains a clear chain-parity dependence. The Sxx{\cal S}_{xx} component probed at Q=(0,0,2)\mathbf{Q}=(0,0,2) [Figure 5a] reveals three transverse excitations at approximately 0.7, 1.0 and 1.25 meV, assigned to ∣Cℓodd⟩\ket{\mathcal{C \ell}^{\rm odd}}, ∣1Eeven⟩\ket{1\mathcal{E}^{\rm even}}, and ∣1Eodd⟩\ket{1\mathcal{E}^{\rm odd}} respectively, based on calculations for ideal chains of N=20\mathcal{N}=20 and N=21\mathcal{N} = 21 sites [Figure 4c]. Notably, both the well-defined cluster-mode feature and the strong intensity of the 1.25 meV excitation exceed expectations from calculations averaging over a geometrical chain-length distribution [Figure 5a]. This may indicate a higher proportion of chains around the mean length of 20 sites, possibly related to the small difference in ionic radii between Mg2+^{2+} and Co2+^{2+}, although this interpretation remains speculative. At Q=(3,0,1)\mathbf{Q}=(3,0,1) [Figure 5(b)], both Sxx{\cal S}_{xx} and Szz{\cal S}_{zz} contribute. The most prominent additional feature is the even-chain cluster mode ∣Cℓeven⟩\ket{\mathcal{C \ell}^{\rm even}} at around 0.7 meV, which carries significantly larger spectral weight than the odd-chain cluster mode, consistent with the DMRG results of Figure 5b. Additional longitudinal excitations associated with ∣2Eeven⟩\ket{2\mathcal{E}^{\rm even}} and ∣2Eodd⟩\ket{2\mathcal{E}^{\rm odd}} are observed around 1.25 meV. Although their parity-dependent energy splitting is difficult to resolve, an additional broader Gaussian component is required to reproduce the data, reflecting the distribution of short chain segments and the associated energy broadening. Above 1.5 meV, the response includes contributions from a distribution of transverse Zeeman ladders, also reproduced by the distribution-averaged calculations. Although a full separation of transverse and longitudinal contributions would require polarized neutron scattering, the overall mode assignment is supported by the DMRG calculations in Figure 5b. Altogether, these observations provide a clear experimental signature of the parity-dependent excitation spectrum characteristic of the intermediate chain regime as well as unambiguously establishing the presence of a cluster mode.

For x=2%x=2\%, corresponding to Nˉ≃50\bar{\mathcal{N}}\simeq50 [Figure 5(c,d)], the system lies in the long chain regime where the excitation spectrum becomes essentially independent of chain parity. At Q=(0,0,2)\mathbf{Q}=(0,0,2), a transverse excitation is observed near 1.05 meV, which corresponds to ∣1Eeven⟩\ket{1\mathcal{E}^{\mathrm{even}}} and ∣1Eodd⟩\ket{1\mathcal{E}^{\rm odd}}, now indistinguishable as the system approaches the thermodynamic limit. At Q=(3,0,1)\mathbf{Q}=(3,0,1), an additional measured excitation around 1.25 meV matches the now-degenerate ∣2Eeven⟩\ket{2\mathcal{E}^{\rm even}} and ∣2Eodd⟩\ket{2\mathcal{E}^{\rm odd}} modes. A residual peak around 0.7 meV, visible in both channels and more clearly in Figure 3f and Figure 3h, is the remnant cluster-mode signal from short segments, which still exist with small probability in the geometric distribution. The second lowest excitation around 1.7 meV corresponds to the first Zeeman ladder state ∣1T⟩\ket{1T}. These observations confirm that the parity dependence characteristic of the intermediate chain regime disappears in the long chain regime.

Finally, in the pure compound (x=0x=0), corresponding to the infinite-chain limit, the in-gap excitations vanish and the spectrum recovers the Zeeman ladder structure [Figure 5(e,f)].

Figure 5. Evolution of the lowest-energy modes with increasing chain length. (a–f) INS spectra (open circles) and DMRG calculations (thick solid lines) for x=5%x=5\%, 2%2\%, and 00, corresponding to average chain lengths Nˉ≈20\bar{\mathcal{N}}\approx20, 5050, and ∞\infty, respectively. Thin solid lines in the experimental (Exp.) panels show fits to the INS data, highlighting the excitation modes identified by DMRG. Dashed lines below the DMRG spectra show the even- and odd-chain contributions to Sxx{\cal S}_{xx} (light and dark blue) and Szz{\cal S}_{zz} (orange and brown), respectively, with intensities multiplied by a factor of 3 for clarity. Measurements and calculations at Q=(0,0,2)\mathbf{Q}=(0,0,2) (blue) and Q=(3,0,1)\mathbf{Q}=(3,0,1) (red) probe Sxx{\cal S}_{xx} and Sxx+Szz{\cal S}_{xx}+{\cal S}_{zz}, respectively. Colored arrows indicate the excitations identified by ED and DMRG, following the color code of Figure 4; black arrows mark the Zeeman ladders. (g) Schematic evolution of the lowest-energy excitations with chain length. At x=5%x=5\% (regime 2), the cluster mode ∣Cℓodd⟩\ket{\mathcal{C \ell}^{\rm odd}} and ∣Cℓeven⟩\ket{\mathcal{C \ell}^{\rm even}}, with transverse or longitudinal character depending on chain parity. At x=2%x=2\% (regime 3), bound edge spinons ∣1Eeven⟩\ket{1\mathcal{E}^{\rm even}} and ∣1Eodd⟩\ket{1\mathcal{E}^{\rm odd}} (now indistinguishable) , corresponding to a single spin flip at the chain boundary and therefore a transverse transition (ΔSz=±1\Delta S^z=\pm1). In the infinite-chain limit (x=0x = 0), the first transverse Zeeman ladder, ∣1T⟩\ket{1T} (ΔSz=±1\Delta S^z=\pm1).

Discussion

Previous theoretical work predicted edge-mode physics in anisotropic open chains [19, 31, 39], and discrete excitations were experimentally observed in Mg-doped quasi-one-dimensional Ising systems [26, 24, 23]. In this context, the key finding of our work is that by varying the Mg-doping concentration, it becomes possible to reveal three distinct regimes in which chain parity, spin entanglement, and the nature of the dominant excitations all vary: a cluster mode corresponding to coherent reversal of the Néel segment, and edge states corresponding to spinons localized at segment boundaries. The INS spectra provide direct evidence for these distinct modes associated with even and odd chain lengths.

The experimental spectra are qualitatively captured by XXZXXZ calculations with open boundary conditions and mean-field interchain coupling. Although approximate — neighboring chains are treated as infinite — this approach captures the essential low-energy physics. Future work could involve cluster mean-field or two-dimensional DMRG approaches. On the experimental side, higher dopings (∼10\sim 10–15%15\%) would access the short-chain regime, where the two Néel configurations of an even-length chain become strongly entangled, potentially forming a mesoscopic Schrödinger cat state. Substitution with magnetic impurities may further open a route to Kondo-like physics in segmented quantum magnets [18, 28].

At a fundamental level, our results demonstrate how finite size, anisotropy, and open boundaries conspire to generate emergent boundary-localized excitations. The cluster mode in particular underscores the essential role of interchain coupling, which acts as an effective staggered field. More broadly, our findings could motivate studies in engineered quantum platforms — ultracold atoms, Rydberg arrays, and trapped-ion chains [60, 45, 47] — where finite-length spin systems with tunable boundaries and staggered fields could directly realize edge spinons and cluster modes. Related boundary fractionalization has recently been observed in atomically engineered nanographene Haldane chains [27]. From a quantum-information perspective, edge spinon modes offer localized, boundary-confined excitations suited for robust quantum state storage, while the cluster mode provides a segment-wide collective excitation for encoding quantum states across a finite chain, complementing efforts toward protected qubits in one-dimensional systems [46, 52, 51].

Our observations also raise the question of whether these modes can be interpreted within the framework of Majorana fermions. Via a Jordan–Wigner transformation, the system maps onto an interacting Majorana chain with spatially varying masses (see Supplementary Material), and BaCo2_{2}V2_{2}O8_{8} has recently attracted attention as a platform for Kitaev-type anisotropic interactions [34, 25, 49]. A detailed analysis of these connections is left for future work. Finally, doping-induced segmentation in higher-dimensional correlated materials may similarly generate confined low-energy excitations, offering new perspectives on emergent phenomena in complex quantum matter.

In summary, combining neutron spectroscopy with numerical calculations, we demonstrate that chemical segmentation of a quasi-one-dimensional Ising antiferromagnet fundamentally reshapes its excitation spectrum through controlled finite-size effects. We identify three regimes governed by chain length: a short-chain regime characterized by strong entanglement and Anderson-tower physics, an intermediate regime dominated by a collective cluster mode, and a long-chain regime hosting edge-localized spinon modes. Our experiments directly probe the latter two, providing spectroscopic evidence for nondispersive excitations with clear chain-parity signatures, in close agreement with theory. These results establish segmented spin chains as a model platform for mesoscopic quantum magnetism and demonstrate that chemical substitution provides a controlled route to engineering collective and localized quantum degrees of freedom within a single bulk material.

Methods

Sample growth

Ba(Co1−x_{1-x}Mgx_x)2_2V2_{2}O8_{8} (x=2%,5%)(x = 2\%, 5\%) single-crystals were grown at Institut Néel by floating zone method. For both concentrations, a 5 cm long cylindrical crystal rod of about 4 mm diameter was obtained by imposing the growth axis to be along the b{\bf b} crystallographic axis. One crystal piece of 5 mm long was cut from the rod for the diffraction experiment while two crystal pieces of about 2 cm long were cut for the inelastic neutron scattering (INS) experiments.

Neutron diffraction

Two single-crystal diffraction experiments were carried out on the two thermal single crystal diffractometers 6T2 (LLB) for x=5%x = 5\% and D23 (ILL) for x=2%x = 2\%. These experiments were performed with a 6 T and a 12 T cryomagnet respectively and with a dilution insert allowing to reach temperatures down to 50 mK. For 6T2 and D23, we used a Cu (200) monochromator with wavelength of 0.9 Åand 1.28 Årespectively.

Inelastic neutron scattering experiment

The INS experiments for both x=2%x = 2\% and x=5%x = 5\% were carried out on the cold-neutron triple-axis spectrometer FZJ-CRG IN12 at ILL. We used a combination of PG(002) monochromator and analyzer allowing to get an energy resolution of the order of 0.035 meV and 0.1 meV with final wavelength 1.05 Å and 1.3 Å respectively. The higher order contamination was suppressed using a velocity selector. The sample was oriented with the b\bm{b} axis vertical yielding a (a∗,c∗\bm{a}^*, \bm{c}^*) horizontal scattering plane and was inserted in an orange cryostat allowing to reach a lowest temperature of 1.5 K.

Inelastic neutron scattering formalism

The spectra obtained in the INS experiments are related to the dynamical susceptibility,

Sαβ(Q,ω,N)=−Im∫dt∑j=1Nei(ωt−Q⋅rj)CαβR(rj,t),\begin{align} S_{\alpha\beta}(\boldsymbol{Q},\omega,\mathcal{N}) =-\mathrm{Im}\int dt\sum_{j=1}^{\mathcal{N}} e^{i(\omega t-\boldsymbol{Q} \cdot \boldsymbol{r}_{j})} C_{\alpha\beta}^{\mathrm{R}}(\boldsymbol{r}_{j},t), \tag*{(4)} \end{align}

which is the Fourier transform of the retarded spin-spin correlation,

CαβR(rj,t)=−iϑ(t)⟨[Sα(rj,t),Sβ(0,0)]⟩,\begin{align} C_{\alpha\beta}^{\mathrm{R}}(\boldsymbol{r}_{j},t) =-i\vartheta (t) \braket{[S^{\alpha}(\boldsymbol{r}_{j},t),S^{\beta}(\boldsymbol{0},0)]}, \tag*{(5)} \end{align}

where rj\boldsymbol{r}_{j} is the position of Co2+^{2+} ions, α,β=x,y,z\alpha, \beta = x, y,z and ϑ(t)\vartheta (t) is the step function. After taking the summation over the ensemble of chain segments, we obtain

Sαβ(Q,ω)=∑NNP(N)Sαβ(Q,ω,N),\begin{equation} \mathcal{S}_{\alpha\beta}(Q,\omega) = \sum_{\mathcal{N}} \mathcal{N} P(\mathcal{N}) S_{\alpha\beta}(\boldsymbol{Q},\omega,\mathcal{N}), \tag*{(6)} \end{equation}

Note that the contribution from each segment is proportional to the length N\mathcal{N} of the segment. The graph of the weight NP(N)\mathcal{N} P(\mathcal{N}) is given in Figure 1h.

At a given momentum transfer, the measured neutron instensity is proportional to the scattering function

S(Q,ω)=∑α,β=x,y,z(δαβ−QαQβ∣Q∣2)∣f(Q)∣2Sαβ(Q,ω),\begin{align} \mathcal{S}(\boldsymbol{Q},\omega) =\sum_{\alpha,\beta=x,y,z} \Big(\delta_{\alpha\beta}-\frac{Q_{\alpha}Q_{\beta}}{|\boldsymbol{Q}|^{2}}\Big) |f(\boldsymbol{Q})|^{2} \mathcal{S}_{\alpha\beta}(\boldsymbol{Q},\omega), \tag*{(7)} \end{align}

where f(Q)f(\boldsymbol{Q}) is the Co2+^{2+} magnetic form factor and Q\boldsymbol{Q} the scattering vector. For Q∥\boldsymbol{Q} \parallel zz axis, only the transverse component Sxx(Q,ω)\mathcal{S}_{xx}(\boldsymbol{Q},\omega) (=Syy(Q,ω)=\mathcal{S}_{yy}(\boldsymbol{Q},\omega)) is captured. Otherwise, both longitudinal Szz(Q,ω)\mathcal{S}_{zz}(\boldsymbol{Q},\omega) and transverse Sxx(Q,ω)\mathcal{S}_{xx}(\boldsymbol{Q},\omega) excitations are observed.

Numerical calculations

The inelastic neutron scattering spectra were calculated using the Chebyshev matrix product state (CheMPS) method, which combines density matrix renormalization group (DMRG) calculations for the ground state with a kernel polynomial expansion of the dynamical response. The many-body wavefunction was represented as a matrix product state. Ground states were obtained by finite-system DMRG on clusters of length N\mathcal{N} with open boundary conditions. Dynamical spin structure factors were then evaluated at zero temperature by applying the corresponding spin operator to the ground state and generating the Chebyshev vectors recursively after rescaling the Hamiltonian to the required interval. The spectra were reconstructed from the Chebyshev moments using a Jackson kernel. Exact diagonalization (ED) calculations were performed for chains of up to 24 sites to characterize the eigenstates in terms of their Ising-limit basis states. The principle of the numerical calculations and the used parameters are given in Supplementary Material. All calculations were performed using an in-house code developed by the authors.

Acknowledgments

We acknowledge insightful discussions with Nicolas Laflorencie and Pierre Pfeuty. The neutron beam time was provided by the ILL and LLB neutron sources. ST is supported by a Grant-in-Aid for Scientific Research from JSPS, KAKENHI Grant Nos. JP24H00191, JP24K06891, and JP26K00662. This work was supported in part by the Swiss National Science Foundation under Division II (Grant No. 200020- 219400).

Author contributions

All authors contributed significantly to this work. In details, sample preparation by PL, neutron scattering experiments and analysis by QF, BG, SP, and VS with the support of SR, ER and XF. Exact diagonalization calculations by ST, QF, SP and DMRG by ST. Physical discussions with TG; Manuscript written by QF with constant feedback from the other co-authors.

Competing interests

The authors declare no competing financial interests.

Data availability

Data are available upon request.

References

  1. [1]Grenier, B., Petit, S., Simonet, V., Canévet, E., Regnault, L.-P., Raymond, S., Canals, B., Berthier, C., & Lejay, P. (2015). Longitudinal and Transverse Zeeman Ladders in the Ising-Like Chain Antiferromagnet BaCo$_2$V$_2$O$_8$. Phys. Rev. Lett., 114, 017201. https://doi.org/10.1103/PhysRevLett.114.017201
  2. [2]He, Z., Fu, D., Kyômen, T., Taniyama, T., & Itoh, M. (2005). Crystal Growth and Magnetic Properties of BaCo$_2$V$_2$O$_8$. Chem. Mater., 17(11), 2924–2926. https://doi.org/10.1021/cm050760e
  3. [3]Bera, A. K., Lake, B., Stein, W.-D., & Zander, S. (2014). Magnetic correlations of the quasi-one-dimensional half-integer spin-chain antiferromagnets SrM$_2$V$_2$O$_8$ ($M$ = Co, Mn). Phys. Rev. B, 89, 094402. https://doi.org/10.1103/PhysRevB.89.094402
  4. [4]Bera, A. K., Lake, B., Essler, F. H. L., Vanderstraeten, L., Hubig, C., Schollwöck, U., Islam, A. T. M. N., Schneidewind, A., & Quintero-Castro, D. L. (2017). Spinon confinement in a quasi-one-dimensional anisotropic Heisenberg magnet. Phys. Rev. B, 96, 054423. https://doi.org/10.1103/PhysRevB.96.054423
  5. [5]Lejay, P., Canevet, E., Srivastava, S. K., Grenier, B., Klanjsek, M., & Berthier, C. (2011). Crystal growth and magnetic property of MCo$_2$V$_2$O$_8$ ($M$ = Sr and Ba). J. Cryst. Growth, 317(1), 128–131. https://doi.org/10.1016/j.jcrysgro.2011.01.016
  6. [6]Lake, B., Tennant, D. A., Caux, J.-S., Barthel, T., Schollwöck, U., Nagler, S. E., & Frost, C. D. (2013). Multispinon Continua at Zero and Finite Temperature in a Near-Ideal Heisenberg Chain. Phys. Rev. Lett., 111, 137205. https://doi.org/10.1103/PhysRevLett.111.137205
  7. [7]Giamarchi, T. (2003). Quantum Physics in One Dimension. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198525004.001.0001
  8. [8]Niesen, S. K., Breunig, O., Salm, S., Seher, M., Valldor, M., Warzanowski, P., & Lorenz, T. (2014). Substitution effects on the temperature versus magnetic field phase diagrams of the quasi-one-dimensional effective Ising spin-$ 12$ chain system BaCo$_2$V$_2$O$_8$. Phys. Rev. B, 90, 104419. https://doi.org/10.1103/PhysRevB.90.104419
  9. [9]Mourigal, M., Enderle, M., Klöpperpieper, A., Caux, J.-S., Stunault, A., & Rønnow, H. M. (2013). Fractional spinon excitations in the quantum Heisenberg antiferromagnetic chain. Nat. Phys., 9(7), 435–441. https://doi.org/10.1038/nphys2652
  10. [10]Grenier, B., Monod, P., Hagiwara, M., Matsuda, M., Katsumata, K., Clément, S., Renard, J.-P., Barra, A. L., Dhalenne, G., & Revcolevschi, A. (2002). Electron spin resonance of Ni-doped CuGeO$_3$ in the paramagnetic, spin-Peierls, and antiferromagnetic states: Comparison with nonmagnetic impurities. Phys. Rev. B, 65, 094425. https://doi.org/10.1103/PhysRevB.65.094425
  11. [11]Masuda, T., Fujioka, A., Uchiyama, Y., Tsukada, I., & Uchinokura, K. (1998). Phase Transition between Dimerized-Antiferromagnetic and Uniform-Antiferromagnetic Phases in the Impurity-Doped Spin-Peierls Cuprate CuGeO$_3$. Phys. Rev. Lett., 80, 4566–4569. https://doi.org/10.1103/PhysRevLett.80.4566
  12. [12]Hase, M., Terasaki, I., Sasago, Y., Uchinokura, K., & Obara, H. (1993). Effects of substitution of Zn for Cu in the spin-Peierls cuprate, CuGeO$_3$: The suppression of the spin-Peierls transition and the occurrence of a new spin-glass state. Phys. Rev. Lett., 71, 4059–4062. https://doi.org/10.1103/PhysRevLett.71.4059
  13. [13]Azuma, M., Fujishiro, Y., Takano, M., Nohara, M., & Takagi, H. (1997). Switching of the gapped singlet spin-liquid state to an antiferromagnetically ordered state in Sr(Cu$_1-x$Zn$_x$)$_2$O$_3$. Phys. Rev. B, 55, R8658–R8661. https://doi.org/10.1103/PhysRevB.55.R8658
  14. [14]Fujiwara, N., Yasuoka, H., Fujishiro, Y., Azuma, M., & Takano, M. (1998). NMR Study of Zn Doping Effect in Spin Ladder System SrCu$_2$O$_3$. Phys. Rev. Lett., 80, 604–607. https://doi.org/10.1103/PhysRevLett.80.604
  15. [15]Ohsugi, S., Tokunaga, Y., Ishida, K., Kitaoka, Y., Azuma, M., Fujishiro, Y., & Takano, M. (1999). Impurity-induced staggered polarization and antiferromagnetic order in spin-$ 12$ Heisenberg two-leg ladder compound SrCu$_2$O$_3$: Extensive Cu NMR and NQR studies. Phys. Rev. B, 60, 4181–4190. https://doi.org/10.1103/PhysRevB.60.4181
  16. [16]Alloul, H., Bobroff, J., Gabay, M., & Hirschfeld, P. J. (2009). Defects in correlated metals and superconductors. Rev. Mod. Phys., 81, 45–108. https://doi.org/10.1103/RevModPhys.81.45
  17. [17]Grenier, B., Renard, J.-P., Veillet, P., Paulsen, C., Dhalenne, G., & Revcolevschi, A. (1998). Scaling in dimer breaking by impurities in CuGeO$_3$: A comparative experimental study of Zn-, Mg-, Ni-, and Si-doped single crystals. Phys. Rev. B, 58, 8202–8205. https://doi.org/10.1103/PhysRevB.58.8202
  18. [18]Eggert, S., Affleck, I., & Horton, M. D. P. (2002). Néel Order in Doped Quasi-One-Dimensional Antiferromagnets. Phys. Rev. Lett., 89, 047202. https://doi.org/10.1103/PhysRevLett.89.047202
  19. [19]Wessel, S., & Haas, S. (2000). Excitation spectra and thermodynamic response of segmented Heisenberg spin chains. Phys. Rev. B, 61, 15262–15268. https://doi.org/10.1103/PhysRevB.61.15262
  20. [20]Augier, D., Sørensen, E., Riera, J., & Poilblanc, D. (1999). Soliton bound states in the Raman spectrum of pure and doped spin-Peierls chains. Phys. Rev. B, 60, 1075–1081. https://doi.org/10.1103/PhysRevB.60.1075
  21. [21]Sotnikov, O. M., Stepanov, E. A., Katsnelson, M. I., Mila, F., & Mazurenko, V. V. (2023). Emergence of Classical Magnetic Order from Anderson Towers: Quantum Darwinism in Action. Phys. Rev. X, 13, 041027. https://doi.org/10.1103/PhysRevX.13.041027
  22. [22]Sirker, J., Laflorencie, N., Fujimoto, S., Eggert, S., & Affleck, I. (2007). Chain Breaks and the Susceptibility of Sr$_2$Cu$_1-x$Pd$_x$O$_3+ $ and Other Doped Quasi-One-Dimensional Antiferromagnets. Phys. Rev. Lett., 98, 137205. https://doi.org/10.1103/PhysRevLett.98.137205
  23. [23]Yang, Y. S., Marsiglio, F., Madsen, M., Gaulin, B. D., Rogge, R. B., & Fernandez-Baca, J. A. (2002). Spin-wave response in the dilute quasi-one-dimensional Ising-like antiferromagnet CsCo$_0.83$Mg$_0.17$Br$_3$. Phys. Rev. B, 65, 212408. https://doi.org/10.1103/PhysRevB.65.212408
  24. [24]Itoh, S., Ikeda, H., Sugimoto, D., Iwasa, K., & Eccleston, R. S. (1995). Spin Dynamics in an Ising-Like Quantum Spin Chain with Nonmagnetic Impurities. J. Phys. Soc. Jpn., 64(3), 927–933. https://doi.org/10.1143/JPSJ.64.927
  25. [25]Bhullar, M., Richard, P., & Kee, H.-Y. (2025). Microscopic Theory Revealing Dual Field-Induced Transitions in Spin-$1/2$ Screw-Chain Magnets. https://arxiv.org/abs/2512.09039
  26. [26]Nagler, S. E., Buyers, W. J. L., Armstrong, R. L., & Briat, B. (1983). Ising-like spin-$ 12$ quasi-one-dimensional antiferromagnets: Spin-wave response in CsCoX$_3$ salts. Phys. Rev. B, 27, 1784–1799. https://doi.org/10.1103/PhysRevB.27.1784
  27. [27]Mishra, S., Catarina, G., Wu, F., Ortiz, R., Jacob, D., Eimre, K., Ma, J., Pignedoli, C. A., Feng, X., Ruffieux, P., Fernández-Rossier, J., & Fasel, R. (2021). Observation of fractional edge excitations in nanographene spin chains. Nature, 598(7880), 287–292. https://doi.org/10.1038/s41586-021-03842-3
  28. [28]Laflorencie, N., Sørensen, E., & Affleck, I. (2008). The Kondo effect in spin chains. J. Stat. Mech.: Theory Exp., 2008(02), P02007. http://stacks.iop.org/1742-5468/2008/i=02/a=P02007stacks.iop.org/1742-5468/2008/i=02/a=P02007
  29. [29]Kenzelmann, M., Xu, G., Zaliznyak, I. A., Broholm, C., DiTusa, J. F., Aeppli, G., Ito, T., Oka, K., & Takagi, H. (2003). Structure of End States for a Haldane Spin Chain. Phys. Rev. Lett., 90, 087202. https://doi.org/10.1103/PhysRevLett.90.087202
  30. [30]Eggert, S., & Affleck, I. (1992). Magnetic impurities in half-integer-spin Heisenberg antiferromagnetic chains. Phys. Rev. B, 46, 10866–10883. https://doi.org/10.1103/PhysRevB.46.10866
  31. [31]Haque, M. (2010). Self-similar spectral structures and edge-locking hierarchy in open-boundary spin chains. Phys. Rev. A, 82, 012108. https://doi.org/10.1103/PhysRevA.82.012108
  32. [32]Pfeuty, P. (1970). The one-dimensional Ising model with a transverse field. Ann. Phys., 57(1), 79–90. https://doi.org/10.1016/0003-4916(70)90270-8
  33. [33]Anderson, P. W. (1952). An Approximate Quantum Theory of the Antiferromagnetic Ground State. Phys. Rev., 86, 694–701. https://doi.org/10.1103/PhysRev.86.694
  34. [34]Morris, C. M., Desai, N., Viirok, J., Hüvonen, D., Nagel, U., Rõõm, T., Krizan, J. W., Cava, R. J., McQueen, T. M., Koohpayeh, S. M., Kaul, R. K., & Armitage, N. P. (2021). Duality and domain wall dynamics in a twisted Kitaev chain. Nat. Phys., 17(7), 832–836. https://doi.org/10.1038/s41567-021-01208-0
  35. [35]Meier, F., Levy, J., & Loss, D. (2003). Quantum Computing with Spin Cluster Qubits. Phys. Rev. Lett., 90, 047901. https://doi.org/10.1103/PhysRevLett.90.047901
  36. [36]Meier, F., Levy, J., & Loss, D. (2003). Quantum computing with antiferromagnetic spin clusters. Phys. Rev. B, 68, 134417. https://doi.org/10.1103/PhysRevB.68.134417
  37. [37]Kattel, P., Tang, Y., Pixley, J. H., & Andrei, N. (2025). Edge spin fractionalization in open one-dimensional spin-$S$ quantum antiferromagnets. Phys. Rev. B, 111, L220402. https://doi.org/10.1103/PhysRevB.111.L220402
  38. [38]Fendley, P. (2012). Parafermionic edge zero modes in $ Z_n$-invariant spin chains. J. Stat. Mech.: Theory Exp., 2012(11), P11020. https://doi.org/10.1088/1742-5468/2012/11/P11020
  39. [39]Kattel, P., Pasnoori, P. R., Pixley, J. H., & Andrei, N. (2025). Edge modes and boundary impurities in the anisotropic Heisenberg spin chain. Phys. Rev. B, 111, 174430. https://doi.org/10.1103/PhysRevB.111.174430
  40. [40]Pasnoori, P. R., Tang, Y., Lee, J., Pixley, J. H., Andrei, N., & Azaria, P. (2025). Spin fractionalization and zero modes in the spin-$ 12$ XXZ chain with boundary fields. Phys. Rev. B, 112, 075121. https://doi.org/10.1103/thlq-h58t
  41. [41]Bose, S. (2007). Quantum communication through spin chain dynamics: an introductory overview. Contemp. Phys., 48(1), 13–30. https://doi.org/10.1080/00107510701342313
  42. [42]Burgarth, D., & Bose, S. (2005). Conclusive and arbitrarily perfect quantum-state transfer using parallel spin-chain channels. Phys. Rev. A, 71, 052315. https://doi.org/10.1103/PhysRevA.71.052315
  43. [43]Kay, A. (2006). Perfect state transfer: Beyond nearest-neighbor couplings. Phys. Rev. A, 73, 032306. https://doi.org/10.1103/PhysRevA.73.032306
  44. [44]Christandl, M., Datta, N., Ekert, A., & Landahl, A. J. (2004). Perfect State Transfer in Quantum Spin Networks. Phys. Rev. Lett., 92, 187902. https://doi.org/10.1103/PhysRevLett.92.187902
  45. [45]Bernien, H., Schwartz, S., Keesling, A., Levine, H., Omran, A., Pichler, H., Choi, S., Zibrov, A. S., Endres, M., Greiner, M., Vuletić, V., & Lukin, M. D. (2017). Probing many-body dynamics on a 51-atom quantum simulator. Nature, 551(7682), 579–584. https://doi.org/10.1038/nature24622
  46. [46]Monroe, C., Campbell, W. C., Duan, L.-M., Gong, Z.-X., Gorshkov, A. V., Hess, P. W., Islam, R., Kim, K., Linke, N. M., Pagano, G., Richerme, P., Senko, C., & Yao, N. Y. (2021). Programmable quantum simulations of spin systems with trapped ions. Rev. Mod. Phys., 93, 025001. https://doi.org/10.1103/RevModPhys.93.025001
  47. [47]Schauss, P. (2018). Quantum simulation of transverse Ising models with Rydberg atoms. Quantum Sci. Technol., 3(2), 023001. https://doi.org/10.1088/2058-9565/aa9c59
  48. [48]Sachdev, S. (2000). Quantum Phase Transitions. Cambridge University Press.
  49. [49]Tang, Y., Kattel, P., & Andrei, N. (2026). On the topological dual of the XXZ spin chain. Phys. Rev. B, 113, L041113. https://doi.org/10.1103/blkg-3t1v
  50. [50]Read, N., & Green, D. (2000). Paired states of fermions in two dimensions with breaking of parity and time-reversal symmetries and the fractional quantum Hall effect. Phys. Rev. B, 61, 10267–10297. https://doi.org/10.1103/PhysRevB.61.10267
  51. [51]Mi, X., Michailidis, A. A., Shabani, S., et al. (2024). Stable quantum-correlated many-body states through engineered dissipation. Science, 383(6689), 1332–1337. https://doi.org/10.1126/science.adh9932
  52. [52]King, A. D., Suzuki, S., Raymond, J., et al. (2022). Coherent quantum annealing in a programmable 2000-qubit Ising chain. Nat. Phys., 18(11), 1324–1328. https://doi.org/10.1038/s41567-022-01741-6
  53. [53]Kitaev, A. Y. (2001). Unpaired Majorana fermions in quantum wires. Phys.-Usp., 44(10S), 131. https://doi.org/10.1070/1063-7869/44/10S/S29
  54. [54]Su, W. P., Schrieffer, J. R., & Heeger, A. J. (1979). Solitons in Polyacetylene. Phys. Rev. Lett., 42, 1698–1701. https://doi.org/10.1103/PhysRevLett.42.1698
  55. [55]Fendley, P. (2016). Strong zero modes and eigenstate phase transitions in the XYZ/interacting Majorana chain. J. Phys. A: Math. Theor., 49(30), 30LT01. https://doi.org/10.1088/1751-8113/49/30/30LT01
  56. [56]Takayoshi, S., Faure, Q., Simonet, V., Grenier, B., Petit, S., Ollivier, J., Lejay, P., & Giamarchi, T. (2023). Phase transitions and spin dynamics of the quasi-one-dimensional Ising-like antiferromagnet BaCo$_2$V$_2$O$_8$ in a longitudinal magnetic field. Phys. Rev. Res., 5, 023205. https://doi.org/10.1103/PhysRevResearch.5.023205
  57. [57]Faure, Q., Takayoshi, S., Petit, S., Simonet, V., Raymond, S., Regnault, L.-P., Boehm, M., White, J. S., Månsson, M., Rüegg, C., Lejay, P., Canals, B., Lorenz, T., Furuya, S. C., Giamarchi, T., & Grenier, B. (2018). Topological quantum phase transition in the Ising-like antiferromagnetic spin chain BaCo$_2$V$_2$O$_8$. Nat. Phys., 14(7), 716–722. https://doi.org/10.1038/s41567-018-0126-8
  58. [58]Faure, Q., Takayoshi, S., Grenier, B., Petit, S., Raymond, S., Boehm, M., Lejay, P., Giamarchi, T., & Simonet, V. (2021). Solitonic excitations in the Ising anisotropic chain BaCo$_2$V$_2$O$_8$ under large transverse magnetic field. Phys. Rev. Res., 3, 043227. https://doi.org/10.1103/PhysRevResearch.3.043227
  59. [59]Zhao, C., Yang, L., Henriques, J. C. G., Ferri-Cortés, M., Catarina, G., Pignedoli, C. A., Ma, J., Feng, X., Ruffieux, P., Fernández-Rossier, J., & Fasel, R. (2025). Spin excitations in nanographene-based antiferromagnetic spin-$1/2$ Heisenberg chains. Nat. Mater., 24(5), 722–727. https://doi.org/10.1038/s41563-025-02166-1
  60. [60]Simon, J., Bakr, W. S., Ma, R., Tai, M. E., Preiss, P. M., & Greiner, M. (2011). Quantum simulation of antiferromagnetic spin chains in an optical lattice. Nature, 472(7343), 307–312. https://doi.org/10.1038/nature09994

Paper details

vlnc:W_ax66ky53
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 2

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents