NMR investigation of the Knight shift anomaly in CeIrIn${}_5$ at high magnetic fields

We report nuclear magnetic resonance Knight shift data in the heavy-fermion material CeIrIn${}_5$ at fields up to 30 T. The Knight shift of the In displays a strong anomaly, and we analyze the results using two interpretations. We find that the Kondo lattice coherence temperature and the effective mass of the heavy electrons remain largely unaffected by the magnetic field, despite the fact that the Zeeman energy is of the order of the coherence temperature.

Figure 1

(a) Field-swept spectrum of CeIrIn${}_5$ oriented along the $c$ axis at 4 K at a fixed frequency of 246.15 MHz; 1 and 2 refer to In(1) and In(2) resonances. (b) The Knight shift of the In(1) and In(2) in CeIrIn${}_5$ as a function of the temperature and field along the $c$ axis. The susceptibility shown as the solid line was measured in a field of 0.1 T but exhibits little field change up to 9 T at these temperatures. Inset: Crystal stucture of CeIrIn${}_5$, with Ce in yellow, Ir in gray, In(1) in blue, and In(2) in brown.

Figure 2

Knight shifts (a) $K_1$ and (b) $K_2$ versus field at 10 K. Solid lines are guides for the eye. (c) $K_{cf}^0$ and (d) $T^{*}$ versus applied field, as determined from fits to $K_{cf}\left(T\right)$ as described in the text.

Figure 3

$\chi$ ($\circ$) at 0.1 T versus temperature, compared with calculations based on the crystal field potential of the Ce at several different external fields (solid lines). Knight shifts (dashed lines) are shown assuming that the hyperfine coupling to the ground-state doublet vanishes. The temperature at which $K$ and $\chi$ deviate clearly increases with increasing field, in contrast with experimental observations.

Figure 4

(a) Knight shifts of the In(1) and In(2) at 12 T versus the susceptibility (at 0.1 T). Solid lines are best fits to the high-temperature regime, with fit parameters $K_1^0=0.96\pm 0.08$%, $B_1=14.5\pm 0.4$ kOe/${\mu }_B$, $K_2^0=3.4\pm 0.2$%, and $B_2=20.8\pm 0.9$ kOe/${\mu }_B$. (b) The Knight shift of the In(1) versus that of the In(2) at various fields. The solid line is calculated using the fit parameters in (a) as discussed in the text.

Figure 5

$K_{cf}$ versus temperature for several applied fields. Solid lines are fits to the data as described in the text.