Crossover from Kondo semiconductor to metallic antiferromagnet with $5d$-electron doping in ${\mathrm{CeFe}}_2{\mathrm{Al}}_{10}$

We report a systematic study of the $5d$-electron-doped system $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$ $(0\le x\le 0.15)$. With increasing $x$, the orthorhombic $b$ axis decreases slightly while accompanying changes in $a$ and $c$ leave the unit cell volume almost unchanged. Inelastic neutron scattering, along with thermal and transport measurements, reveal that for the Kondo semiconductor ${\mathrm{CeFe}}_2{\mathrm{Al}}_{10}$, the low-temperature energy gap, which is proposed to be a consequence of strong $c\text{−}f$ hybridization, is suppressed by a small amount of Ir substitution for Fe and that the system adopts a metallic ground state with an increase in the density of states at the Fermi level. The charge or transport gap collapses (at $x=0.04$) faster than the spin gap with Ir substitution. Magnetic susceptibility, heat capacity, and muon spin relaxation measurements demonstrate that the system undergoes long-range antiferromagnetic order below a Néel temperature $T_{\mathrm{N}}$ of 3.1(2) K for $x=0.15$. The ordered moment is estimated to be smaller than 0.07(1) ${\mu }_{\mathrm{B}}$/Ce, although the trivalent state of Ce is confirmed by Ce $L_3$-edge x-ray absorption near edge spectroscopy. It is suggested that the $c\text{−}f$ hybridization gap, which plays an important role in the unusually high ordering temperatures observed in $\mathrm{Ce}{T}_2{\mathrm{Al}}_{10}$ ($T$ = Ru and Os), may not be necessary for the onset of magnetic order with a low $T_{\mathrm{N}}$ seen here in $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$.

Figure 1

Fitted powder neutron diffraction patterns of $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}(0\le x\le 0.15)$ obtained at room temperature using the GEM diffractometer. The observed and calculated intensities and their difference are plotted as red symbols, a solid black line, and a solid blue line, respectively. The vertical bars show the positions of the Bragg reflections.

Figure 2

Variation of the orthorhombic lattice parameters $a$, $b$, and $c$, and unit cell volume $V$ as a function of Ir concentration $x$ for $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$. The lines are guides for the eye.

Figure 3

Electrical resistivity vs temperature for $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}(0\le x\le 0.15)$. The inset in (a) shows the thermal activation behavior of the resistivity, while the inset in (d) shows an enlarged view of the low temperature upturn of $\rho \left(T\right)$ for $x=0.15$.

Figure 4

Thermoelectric power versus temperature for $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$ with $x=0$ (left axis) from Ref. [4] and $x=0.15$ (right axis).

Figure 5

(a) Temperature dependence of $C/T$ for polycrystalline samples of $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$ ($x=0$, 0.04, 0.08, 0.15) and ${\mathrm{LaFe}}_2{\mathrm{Al}}_{10}$, on a log-log plot. Lower inset: Variation of $\gamma$ [calculated in the temperature range $(7\lesssim T\lesssim 20\mathrm{K})]$ as a function of Ir concentration $x$ for the polycrystalline samples. (b) Magnetic entropy $S_{\mathrm{mag}}$ as a function of temperature for the polycrystalline samples. (c) $C/T$ versus $T$ (left axis) and $S_{\mathrm{mag}}$ versus $T$ (right axis) for a single crystal sample of $x=0.15$.

Figure 6

Temperature dependence of magnetic susceptibility $\chi (T$) for polycrystalline samples of $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$ ($x=0$, 0.04, 0.08, 0.15) and ${\mathrm{LaFe}}_2{\mathrm{Al}}_{10}$ measured at $H=10$ kOe. Upper left inset shows the peak in $\chi \left(T\right)$ for $x=0.15$. Lower right inset shows ${\chi }^{-1}$ vs $T$ for $x=0$, 0.04, 0.08, 0.15. The solid lines are fits to the data with the modified Curie-Weiss behavior including a temperature independent term ${\chi }_0$.

Figure 7

Temperature dependence of magnetic susceptibility $\chi (T$) of single crystal $\mathrm{Ce}{\left({\mathrm{Fe}}_{0.85}{\mathrm{Ir}}_{0.15}\right)}_2{\mathrm{Al}}_{10}$ in a field of 10 kOe applied along the three principal axes. Upper insets: $\chi$ vs $T$ at low temperature. Lower inset: ${\chi }^{-1}$ vs $T$ along the three principal axes. The dotted and solid lines are fits to the susceptibility data with CEF model 1 and model 2, respectively, discussed in the text.

Figure 8

Inelastic neutron scattering data at 7 K after subtracting the phonon contribution for (a) ${\mathrm{CeFe}}_2{\mathrm{Al}}_{10}$ (data from Ref. [34] are given for comparison), (b) $\mathrm{Ce}{\left({\mathrm{Fe}}_{0.96}{\mathrm{Ir}}_{0.04}\right)}_2{\mathrm{Al}}_{10}$, (c) $\mathrm{Ce}{\left({\mathrm{Fe}}_{0.92}{\mathrm{Ir}}_{0.08}\right)}_2{\mathrm{Al}}_{10}$, and (d) $\mathrm{Ce}{\left({\mathrm{Fe}}_{0.85}{\mathrm{Ir}}_{0.15}\right)}_2{\mathrm{Al}}_{10}$ at 6 K. The $Q$-integration range for the 20 meV data in (a) is 0 to $2{\AA }^{-1}$, for the 38 meV data in (b)–(d) is 0 to $1.5{\AA }^{-1}$, and for the 100 meV data in (a)–(d) is 0 to $4{\AA }^{-1}$. The solid black curves represent fits to the INS peaks with a triple-Lorentzian function (dashed curves represent the components of the fit) multiplied by the population factor (i.e., Bose factor). The fit parameters, i.e., the peak positions, widths, and amplitudes are given in Table S2 of the Supplemental Material [52]. The solid olive curve in (d) shows the fit using model 1.

Figure 9

Zero-field $\mu \mathrm{SR}$ spectra for polycrystalline $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$ ($x=0.15$) at various temperatures. The solid lines are least-squares fits to the data as described in the text.

Figure 10

Temperature dependence of (a) the initial asymmetry, (b) the internal field at the muon stopping site, obtained by zero-field $\mu \mathrm{SR}$ experiment on $\mathrm{Ce}{\left({\mathrm{Fe}}_{0.85}{\mathrm{Ir}}_{0.15}\right)}_2{\mathrm{Al}}_{10}$. The solid line represents the fit using Eq. (5). (c) Temperature dependence of the distribution width of the local Gaussian field and (d) the depolarization rate.

Figure 11

Ce $L_3$ x-ray absorption edge spectra of $\mathrm{Ce}{\left({\mathrm{Fe}}_{1-x}{\mathrm{Ir}}_x\right)}_2{\mathrm{Al}}_{10}$ for $x=0.04$ and 0.15 at 7 K.