Kinetic control of structural and magnetic states in LuBaCo${}_4$O${}_7$

The $R$BaCo${}_4$O${}_7$ (R $=$ Ca, Y, Tb, Ho, Tm, Yb, Lu) compounds provide a novel topology for studying the competition between triangular geometry and magnetic order. Here, we report the structural and magnetic behavior of the Lu member of this series via neutron and synchrotron x-ray diffraction, magnetization, and resistivity measurements. We determined sequential phase transitions and a strong competition between a stable and a metastable low-temperature state that critically depends on controlled cooling rates and the associated heat removal kinetics. No evidence for long-range ordered magnetism was detected by neutron diffraction at any temperature. However, very slow spin dynamics are evidenced by time-dependent neutron diffraction measurements and can be explained by several competing magnetic phases with incommensurate short-range correlations coexisting in this material.

Figure 1

Crystal structure of $R$BaCo${}_4$O${}_7$ family in space group $P$31$c$.

Figure 2

The dc magnetization properties of LuBaCo${}_4$O${}_7$ as a function of temperature measured on cooling and warming with 0.3 and 2 K/min cooling rates under an applied magnetic field of 1 T.

Figure 3

(a) The dc magnetization of LuBaCo${}_4$O${}_7$ on cooling as a function of temperature with various cooling rates under 1 T applied magnetic field. (b) $R/R_0$ vs temperature on warming where $R_0$ is resistance at 218 K.

Figure 4

Rietveld refinements of the room temperature structure using the trigonal $P$31$c$ and hexagonal $P$6${}_{3}mc$ symmetries. Neutron data collected at NPDF. Black crosses, red (dark gray) solid, and blue (medium gray) solid lines represent the observed and calculated intensities and their differences, respectively. Insets show the fitting quality of the two models at a select $d$-spacing range. Agreement factors for $P$31$c$ are ${\chi }^2$ $=$ 1.404, $R_{wp}$ $=$ 2.6$\%$, and $R_p$ $=$ 1.76$\%$ and for $P$6${}_{3}mc$ are ${\chi }^2$ $=$ 2.003, $R_{wp}$ $=$ 3.11$\%$, and $R_p$ $=$ 2.19$\%$.

Figure 5

(a) Portions of the neutron diffraction patterns (WISH) at selected temperatures collected on warming after fast cooling to 10 K (5 K/min). (b) Phase fractions of the monoclinic $Cc$ and orthorombic Pbn2${}_1$ phases as a function of temperature. (c) Rietveld refinement in the monoclinic $Cc$ symmetry using neutron data collected at 150 K. Left inset shows the portion of synchotron x-ray data at 300 and 130 K. Right inset shows the unit cell volume as a function of temperature (error bars correspond to 3σ intervals).

Figure 6

(Top panel) Contour plot of synchrotron x-ray diffraction data on slow cooling and warming with a 0.35 K/min cooling rate showing the evolution of the Pbn2${}_1$ phase. (Bottom panel) Magnetization curves on cooling (left; 0.3 K/min cooling rate and under 1 T applied magnetic field) and warming (right) indicating that the anomalies match the structural transitions observed in the x-ray diffraction data. Inset shows the same cooling and warming magnetization curves on the same temperature scale.

Figure 7

Powder diffraction data collected on POWGEN (neutrons) and 11BM (x rays) on fast cooling. Disparity in the data collected on different instruments clearly shows that the sample is very prone to the effectiveness of the cooling process and the heat removal.

Figure 8

Portions of the neutron diffraction data at different timeframes after fast cooling (4.2 K/min) to 70 K showing the evolution of the Pbn2${}_1$ phase. Arrows show the splitting of the (004) peak. The data shown was collected for 3 h and later binned to 20-min segments, which can be done with Spallation Neutron Source (SNS) data due to time event mode data collection.

Figure 9

Neutron powder diffraction data (NPDF) after fast cooling (3.9 K/min) to 15 K. Arrows show the positions of the expected magnetic peaks.

Figure 10

Neutron diffraction patterns (WISH) collected on warming at selected temperatures after (a) fast cooling (5 K/min) and (b) slow cooling (0.35 K/min). Time-dependent patterns collected at 10 K after fast cooling (c). Low-$Q$ background integrated in the 15–30-Å region as a function of time (d). Inset shows the low-$Q$ region of the two patterns recorded immediately after cooling and after 120 min.