BCTO: Random-Singlet Perovskite Oxide

• BCTO is a disordered perovskite oxide where Cu²⁺ and Ta⁵⁺ randomly occupy the B-sites, producing a constrained random-singlet magnetic network.
• It features multiple magnetic exchange pathways including a dominant Cu–O–Ta–O–Cu interaction (~4 K), rare Cu–O–Cu dimers (~70 K), and a low-J pathway (~0.1 K).
• Thermodynamic measurements show no long-range magnetic order above 0.1 K, highlighting a departure from infinite-randomness criticality with finite RS scaling.

BaCu$_{1/3}$Ta$_{2/3}$O$_3$ (BCTO) is a disordered spin-$\frac{1}{2}$ perovskite oxide in which Cu$^{2+}$ ($S=\frac{1}{2}$) and Ta$^{5+}$ randomly occupy the B-site of a pseudo-cubic ABO$_3$ lattice. The material exhibits a locally correlated but macroscopically homogeneous B-site distribution, featuring Jahn–Teller distorted CuO$_6$ octahedra and constrained exchange pathways. BCTO’s magnetic ground state is characterized by the emergence of a broad, bounded random-singlet network, with a distribution $P(J)$ of exchange couplings spanning nearly four decades in energy, yet truncated at low $J$, distinguishing it from an infinite-randomness fixed point regime. No long-range magnetic ordering or spin freezing is observed down to 0.1 K, and thermodynamic anomalies point to a departure from conventional random-singlet (RS) models (Mahapatra et al., 24 Jan 2026).

1. Lattice Structure and Local B-site Disorder

BCTO crystallizes in a pseudo-cubic perovskite structure (ABO$_3$; A = Ba, B = Cu/Ta), with two distinct but related tetragonal symmetry variants (P4/mmm and P4mm) determined via high-resolution synchrotron powder X-ray diffraction (typically $\lambda \approx 0.5$ Å). The refined lattice constants yield $a \approx b \approx 3.93$ Å and $c \approx 3.98$ Å, giving $c/a \approx 1.013$. A minute off-centering of the B-site in the P4mm phase further enhances the range of Cu–O–Ta bond angles, broadening their distribution from 169$^\circ$ (distorted) to the ideal 180$^\circ$.

On the B-sublattice, Cu$^{2+}$ and Ta$^{5+}$ occupy corner sites in a statistical 1:2 ratio. EXAFS (Cu K-edge, Ta L$_{III}$-edge) analyses confirm no macroscopic phase separation and demonstrate local correlation in Cu/Ta occupancy.

Key local bond parameters (from EXAFS):

Atom Pair	Mean Distance (Å)	Variance (Å$^2$)	Coordination Number
Cu–O$_1$	2.03(1)	0.0087	4
Cu–O$_2$	2.32(1)	0.0087	2
Ta–O	1.98(1)	0.0089	6

Neighbor-counting analysis evidences strong heteroatomic coordination: on average, each Cu sites have $\approx$ 5.6 Ta and $\approx$ 0.4 Cu B-site neighbors, while Ta sites are coordinated by $\approx$ 3.1 Ta and $\approx$ 2.9 Cu.

2. Magnetic Exchange Pathways

The local chemical order in BCTO constrains possible superexchange interactions. The dominant exchange pathways identified by structural analysis are:

• Cu–O–Ta–O–Cu: Provides the main magnetic exchange coupling, $J \sim 4$ K.
• Cu–O–Ta–O–Ta–O–Cu: Facilitates longer-range connections with $J \sim 0.1$ K.
• Cu–O–Cu dimers: These provide a strongly suppressed but finite population with $J \sim 70$ K; they are rare but account for the high-$J$ shoulder in the reconstructed $P(J)$.

These superexchange motifs, enforced by the compositional disorder, do not permit extensive Cu–O–Cu chains, and the distribution of exchange geometries is intrinsically bounded. This results in a broad but finite spectrum of exchange strengths.

3. Bulk Magnetism and Thermodynamic Behavior

BCTO exhibits no detectable magnetic ordering or spin freezing down to 0.1 K, indicated by the lack of zero-field-cooled/field-cooled (ZFC/FC) splitting in susceptibility data. Bulk susceptibility $\chi(T)$ above 50 K follows a Curie–Weiss law with effective moment $\mu_\text{eff} = 1.85\,\mu_B$ and Curie–Weiss temperature $\theta_\text{CW} \approx -35$ K. Below $\sim$10 K, the susceptibility diverges as a fractional power law, $\chi(T) \propto T^{-\gamma}$ with $\gamma \approx 0.67$, verified down to $\sim$4 K.

Field- and temperature-dependent magnetization collapses for $3~\text{K} < T < 9~\text{K}$ under the RS scaling:

$M \cdot T^{\gamma-1} \text{ vs. } \mu_0 H / T$

with $\gamma \approx 0.67$.

The magnetic specific heat $c_\mathrm{mag}(T)$ (phonon background subtracted) displays a broad Schottky-like anomaly shifting with applied field, and a nuclear upturn below $\sim$0.2 K. Distinctly, below $1$ K, $c_\mathrm{mag} \propto T$ rather than $T^{1-\gamma} = T^{0.33}$, and the RS scaling collapse for $(\mu_0 H)^\gamma c_p/T$ vs. $T/(\mu_0 H)$ fails. The integrated magnetic entropy saturates at $\approx 0.4\,R\,\ln 2$ by 20 K, indicating $\sim$60% of spins remain dynamic at 0.1 K (Mahapatra et al., 24 Jan 2026).

4. Exchange Coupling Distribution: Reconstruction and Physical Consequences

The exchange distribution $P(J)$ in BCTO is directly constrained and reconstructed by fitting bulk magnetic observables. Instead of an assumed power-law, $P(J)$ is represented as a sum of log-normal components:

$$P(J) = \sum_{k=1}^K w_k \frac{1}{J \sigma_k \sqrt{2\pi}} \exp\left[-\frac{(\ln J - \ln J_{0k})^2}{2\sigma_k^2}\right], \quad \sum_k w_k = 1.$$

Empirical analysis yields a distribution very broad, spanning $J \sim 0.01$ K – $100$ K, with three dominant features:

• High-$J$ shoulder at $J \approx 70$ K (Cu–O–Cu dimers, rare).
• Main peak at $J \approx 4$ K (Cu–O–Ta–O–Cu).
• Low-$J$ bump at $J \approx 0.1$ K (double Ta via Cu–O–Ta–O–Ta–O–Cu), plus a nonsingular tail for $J \rightarrow 0$.

Over $J \gtrsim 4$ K, the data approximate $P(J) \propto J^{-0.7}$, accounting for the observed RS-like scaling of $\chi$ and $M$ for $T > 3$ K. At lower $J$, however, $P(J)$ is sharply cut off, preventing the realization of true infinite-randomness scaling.

Global fits including a small monomer (“orphan”) fraction $f \approx 0.04$ of $S=\frac{1}{2}$ spins and the reconstructed $P(J)$ achieve rms deviations below 3% for all observed $\chi(T)$, $M(H,T)$, and $c_\mathrm{mag}(T,H)$.

5. Departure from Infinite-Randomness Random-Singlet Physics

In canonical one-dimensional RS phases at an infinite-randomness fixed point, exchange couplings follow a scale-free distribution $P(J) \propto J^{-\gamma}$ down to $J\to0$ and thermodynamics obey universal scalings:

• $\chi \propto T^{-\gamma}$,
• $c_\mathrm{mag} \propto T^{1-\gamma}$,
• $M/T^{1-\gamma}$ vs. $H/T$,
• $(H^\gamma c_\mathrm{mag}/T)$ vs. $T/H$.

In BCTO, these scaling relations are respected only over an intermediate regime; at lower energies and temperatures, the boundedness of $P(J)$ leads to a breakdown of RS scaling, especially in the low-$T$ behavior of $c_\mathrm{mag}$ and the field scaling of thermodynamic quantities.

This establishes BCTO as a three-dimensional, broad but finite, random-singlet network—not an infinite-randomness RS phase. The absence of scale invariance at the lowest energies is intrinsic to the locally correlated but globally random Cu/Ta order, which precludes divergent sequences of weak links (Mahapatra et al., 24 Jan 2026).

6. Key Equations and Experimental Observables

The primary equations governing the low-temperature magnetism of BCTO are:

Exchange distribution:

$$P(J) = \sum_{k=1}^K w_k \frac{1}{J \sigma_k \sqrt{2\pi}} \exp\left[-\frac{(\ln J - \ln J_{0k})^2}{2\sigma_k^2}\right]$$

Low-temperature susceptibility of dimers:

$$\chi_d(T) = \int_0^\infty P(J) \cdot \frac{(g\mu_B)^2}{k_B T} \frac{e^{-J/T}}{3+e^{-J/T}}\,dJ + f\cdot C/T$$

Magnetic specific heat of dimers:

$$c_d(T) = k_B \int_0^\infty P(J) \frac{(J/T)^2 e^{-J/T}}{(3+e^{-J/T})^2} \, dJ$$

Critical figures include the pseudo-cubic unit cell (random B-site occupation), EXAFS quantification of neighbor probabilities, $T^{-\gamma}$ scaling in susceptibility and magnetization, and the reconstructed $P(J)$ spanning multiple orders of magnitude with annotated peaks.

7. Broader Significance and Implications

BCTO serves as a model system for exploring quantum disordered ground states in three dimensions where constrained crystal chemistry drives broad yet bounded randomness in magnetic exchange. The system demonstrates that correlated disorder can generate a random-singlet-like phase with finite randomness, deviating sharply from infinite-randomness criticality typically seen in lower-dimensional or structurally simpler systems. A plausible implication is that similar compositional templates may provide routes to stabilize novel quantum paramagnetic ground states with tunable singlet networks and defect-induced dynamical spins (Mahapatra et al., 24 Jan 2026). The methodology—combining high-resolution XRD/EXAFS, thermodynamic measurements, and numerical reconstruction of $P(J)$—establishes a framework for studying broader classes of quantum magnets with engineered randomness.