Antiferromagnetic Dichroism in a Complex Multisublattice Magnetoelectric ${\mathrm{CuB}}_2{\mathrm{O}}_4$

Magnetic control of the crystal chirality was announced by Saito et al. [Phys. Rev. Lett. 101, (2008)] on the ground of experiments in $\mathrm{Cu}{\mathrm{B}}_2{\mathrm{O}}_4$. This claim has raised a sharp dispute in the literature because it seemed to contradict the fundamental symmetry principles. We settle this dispute on the basis of a high-resolution optical spectroscopy study of excitonic transitions in $\mathrm{Cu}{\mathrm{B}}_2{\mathrm{O}}_4$. We find that a large sublattice-sensitive antiferromagnetic linear dichroism (LD) emerges at the Néel temperature $T_N=21\mathrm{K}$ and show how it could simulate a “magnetic-field control of the crystal chirality.” We prove that the discovered LD is related microscopically to the magnetic Davydov splitting. This LD is highly sensitive to subtle changes in the spin subsystems, which allowed us to observe a splitting of the phase transition into an incommensurate magnetic phase into two transitions ($T_1^{*}=8.5$ and $T_2^{*}=7.9\mathrm{K}$) and to suggest elliptical spiral structures below $T_1^{*}$, instead of a simple circular helix proposed earlier.

Figure 1

Spectra of the α-polarized absorption (lower trace) and the linear dichroism (upper trace) of $\mathrm{Cu}{\mathrm{B}}_2{\mathrm{O}}_4$ in the commensurate magnetic phase at $T=12\mathrm{K}$. ZP lines originating from the ${\mathrm{Cu}}^{2+}(4b)$ and ${\mathrm{Cu}}^{2+}(8d)$ positions are marked in accordance with Ref. [31]. Right inset displays geometry for LD measurements. Lower left inset shows the first ${\mathrm{Cu}}^{2+}(4b)$ ZP line registered in two mutually perpendicular polarizations, $\mathit{E}\parallel [110]$ and $\mathit{E}\parallel [\overline{1}10]$, both for $\mathit{k}\parallel z$, evidencing the Davydov doublet and explaining the observed LD (upper left inset). Pay attention that the ${\mathrm{Cu}}^{2+}(8d)$ ZP lines do not exhibit LD.

Figure 2

Temperature behavior of the lowest-frequency $\mathrm{Cu}(4b)$ and $\mathrm{Cu}(8d)$ ZP $\alpha$-polarized absorption lines. (a),(b) Color-coded contour plots of the absorption coefficient as a function of temperature vs wave number. The transition temperatures are marked as dotted lines. (c),(d) Temperature dependences of the maximum position and full width at half maximum (FWHM). The error bars are smaller than symbols.

Figure 3

(a) Temperature dependence of LD (for the light polarized along and perpendicular to the [110] direction) at $11136{\mathrm{cm}}^{-1}$ (red balls) and $11140{\mathrm{cm}}^{-1}$ (blue circles). $B_{\text{ext}}=0.7\mathrm{T}$ is applied along the [110] direction. A splitting of the phase transition at $T^{*}$ into two transitions (at $T_1^{*}$ and $T_2^{*}$) is evident. Arrows schematically demonstrate proposed magnetic structures. The right Inset shows the “magnetic circular dichroism” data from Ref. [21]. The left Inset displays a hysteresis loop observed in the vicinity of $T_2^{*}$. (b),(c) LD spectra in the region of the first $4b$ ZP line of $\mathrm{Cu}{\mathrm{B}}_2{\mathrm{O}}_4$ at temperatures below $T_N$, (for the light polarized along and perpendicular to the [110] direction; $\mathbf{k}\parallel z$), displayed as (b) color-coded contour plots and (c) shifted plots taken with steps from 0.5 to 0.05 K.