Nematic-driven anisotropic electronic properties of underdoped detwinned ${\mathrm{Ba}(\mathrm{Fe}}_{1-x}{\mathrm{Co}}_x{)}_2{\mathrm{As}}_2$ revealed by optical spectroscopy

We collect optical reflectivity data as a function of temperature across the structural tetragonal-to-orthorhombic phase transition at $T_s$ on ${\mathrm{Ba}(\mathrm{Fe}}_{1-x}{\mathrm{Co}}_x{)}_2{\mathrm{As}}_2$ for $x=0\%$, 2.5%, and 4.5%, with uniaxial and in situ tunable applied pressure in order to detwin the sample and to exert on it an external symmetry-breaking field. At $T<T_s$, we discover a remarkable hysteretic optical anisotropy as a function of the applied pressure at energies far away from the Fermi level. Such an anisotropy turns into a reversible linear pressure dependence at $T\ge T_s$. Moreover, the optical anisotropy gets progressively depleted with increasing Co content in the underdoped regime, consistent with the doping dependence of the orthorhombicity but contrary to the nonmonotonic behavior observed for the dc anisotropy.

Figure 1

Left: Three-dimensional schematic view of the pressure device with a cross section along the plane of the incident/reflected optical path. Right: Front view along the light path [45]. By flushing He gas into the spring bellows and evacuating its volume, one can exert and release pressure, respectively, along the direction corresponding to the orthorhombic $b$ axis, as indicated by the black arrow. The optical mask, placed on top of the pressure device, defines equal spots of the sample and reference Au-mirror surface, which are exposed to the electromagnetic radiation polarized along the $a$ and $b$ axes (blue and red arrows, respectively).

Figure 2

Temperature and pressure dependence of the optical reflectivity $R\left(\omega \right)$ for ${\mathrm{Ba}(\mathrm{Fe}}_{1-x}{\mathrm{Co}}_x{)}_2{\mathrm{As}}_2$ with $x=2.5\%$ in the MIR spectral range at $T\ll T_s,T<T_s$, and $T\ge T_s$ and variable $p$ [0 bars: (a)–(c), 1.2 bars: (d)–(f), zero-released pressure: (g)–(i)]. The blue and red curves in panels (a)–(i) refer to data collected with the electromagnetic radiation polarized along the $a$ and $b$ axis, respectively. $R_{\mathrm{ratio}}(\omega ,p)=R_a\left(\omega \right)/R_b\left(\omega \right)$ in panels (l)–(q) emphasizes the optical anisotropy [45], which is explicitly shown for increasing (0-0.8-1.2 bars) and decreasing (1.2-0.8-0 bars) pressures [(l)–(n) and (o)–(q), respectively]. The dashed vertical lines indicate the frequency where $R_{\mathrm{ratio}}$ is read (see text and Fig. 4 below).

Figure 3

(a)–(e) Pressure dependence of $\Delta R_{\mathrm{ratio}}=R_{\mathrm{ratio}}(p,{\omega }^{\prime })-1$ at selected temperatures above and below $T_s$ achieved in the ZPC “pressure-loop” experiment for $x=0\%$ [45] and 4.5% Co doping (full and open symbols denote increasing and decreasing pressure, respectively). The dashed and dotted lines are guides for the eyes. The upper right panel (f) displays the real part ${\sigma }_1\left(\omega \right)$ of the optical conductivity for $x=0\%$ and 4.5% Co doping at $T<T_s$, emphasizing the mid-infrared peak overlapped with the low-frequency tail of the stronger near-infrared absorption at about $4000 \mathrm{cm}{}^{-1}$ (data from Ref. [48] on twinned specimens). The vertical dotted lines mark the position of the mid-infrared peak at ${\omega }^{\prime }=1500\left(1500\right)$ and $900 \mathrm{cm}{}^{-1}$ for $x=0\%(2.5\%$, not shown) and 4.5% Co doping, respectively, where $R_{\mathrm{ratio}}$ is read.

Figure 4

(a) Temperature dependence of $\Delta R_{\mathrm{ratio}}\left({\omega }^{\prime }\right)$ at saturation (i.e., at $p\ge$ 0.8 bars, depending on Co doping) as well as at its remanent state (i.e., at released $p=0$ bars) for $x=0\%$ ($T_s\sim 135$ K), 2.5% ($T_s\sim 98$ K) [45], and 4.5% ($T_s\sim 67$ K) Co doping. The $T$ axis is normalized by $T_s$ and ${\omega }^{\prime }$ is defined by the vertical dotted lines in Fig. 3. (b) Co doping dependence $\left(x\right)$ of $\Delta R_{\mathrm{ratio}}$ at 10 K and at saturation [from panel (a)] and of the orthorhombicity $(a-b)/(a+b)$ [53].