Magnetic order and excitations in Ce$_3$TiBi$_5$ and Ce$_3$ZrBi$_5$

Abstract

The R3MBi5 rare-earth intermetallics (R = rare earth, M = Ti, Zr, Sc) provide a versatile platform to explore how Kondo hybridization, RKKY exchange, magnetic frustration, and broken inversion symmetry may cooperate to generate unusual magnetic behavior. We present a comprehensive neutron scattering investigation of the magnetic structure, crystal electric field (CEF), and low-energy excitations in the locally noncentrosymmetric Kondo-lattice compounds Ce3TiBi5 and Ce3ZrBi5. Powder and single-crystal neutron diffraction reveals incommensurate cycloidal antiferromagnetic order in Ce3TiBi5 with propagation vector k = (0, 0, 0.388) and a reduced ordered moment of m = 0.53(3)$μ_{B}$. Ce3ZrBi5 exhibits a qualitatively similar magnetic diffraction profile, with k $\simeq$ (0, 0, 0.37). Inelastic neutron scattering measurements resolve two clear, well-separated CEF excitations in both compounds with nearly the same profile, confirming a well-isolated Kramers doublet ground state. At low energies, a broad, quasi-elastic magnetic response is observed at T $\simeq$ TN, whose momentum-dependence is inconsistent with that expected from conventional collective excitations of localized moments. This discrepancy, along with a Kondo temperature estimate TK ~ 3--5 K -- comparable to TN -- indicates sizable Kondo hybridization, which accounts for the moment reduction and the spiral magnetic order that appears to involve the magnetic hard direction. Our results place these compounds in a regime where local inversion symmetry breaking, anisotropic CEF effects, and competing Kondo and RKKY interactions collectively give rise to unconventional magnetic order.

Figures (15)

• Figure 1: Crystal structure of Ce$_{3}$(Ti,Zr)Bi$_{5}$. (a) Crystallographic unit cell of Ce$_{3}$(Ti,Zr)Bi$_{5}$. (b) Top and side views of the Ce zigzag chains, which form the primary motif of the magnetic sublattice. (c) In-plane connectivity between Ce sites, forming a distorted corner-shared triangular geometry. Sites below the Ce-Bi plane are shown with a lighter shading of color.
• Figure 2: Integrated scattering intensity as a function of temperature for the (1 0 0.6116) magnetic Bragg peak for Ce$_3$TiBi$_5$. Data are based upon the integration of a fitted Gaussian peak with a constant background as shown in the inset of the figure and described in the text. Error bars in temperature correspond to the standard deviation in the temperature for the duration of the measurement. Solid line is a fit to a power law as described in the text. Inset: measured scattering intensity along $(1, 0, L)$ for $\hbar\omega=0$ meV from the VERITAS measurement as described in the text. Temperature listed in legend is shown to two significant digits.
• Figure 3: Scattering intensity measured along the (1, 1, $L$) direction in Ce$_3$TiBi$_5$ from WAND$^2$ at $T = 10$ K and $T = 1.8$ K.
• Figure 4: In-plane magnetic reflection profile of Ce$_3$TiBi$_5$ from the WAND$^2$ measurement. (a) The difference in scattering intensity between $T = 1.8$ K and $T = 10$ K, integrated over $L = 0.35 \pm 0.1$,rlu. red lines mark the positions of aluminum nuclear Bragg peaks. (b) A line cut along the $(H,0,0.35)$ direction from the data in panel (a) for $K$ integrated between $\pm0.05$ rlu.
• Figure 5: Measured scattering intensity as a function of scattering angle, $2\theta$, for powder Ce$_3$ZrBi$_5$ and Ce$_3$TiBi$_5$ samples. Data for the Ce$_3$ZrBi$_5$ sample are shown offset along the vertical axis for clarity and comparison. Data are shown for $T=20$ and $T=1.5$ K measurements. The scattering angle range is chosen to illustrate the magnetic Bragg peak at $2\theta \approx 21$ degrees. Error bars in scattering intensity are smaller than symbol size.