$\beta$-NMR study of isolated ${}^8$Li${}^{+}$ in the enhanced paramagnet platinum

We report $\beta$-detected nuclear magnetic resonance ($\beta$-NMR) measurements of ${}^8$Li${}^{+}$ implanted into high purity Pt. The frequency of the ${}^8$Li $\beta$-NMR resonance and the spin-lattice relaxation rates $1/T_1$ were measured at temperatures ranging from 3 to 300 K. Remarkably, both the spin-lattice relaxation rate and the Knight shift $K$ depend linearly on temperature $T$ although the bulk susceptibility does not. $K$ is found to scale with the Curie-Weiss dependence of the Pt susceptibility extrapolated to low temperatures. This is attributed to a defect response of the enhanced paramagnetism of Pt, i.e., the presence of the interstitial Li${}^{+}$ locally relieves the tendency for the Curie-Weiss susceptibility to saturate at low $T$. We propose that the low temperature saturation in $\chi$ of Pt may be related to an interband coupling between the $s$ and $d$ bands that is disrupted locally by the presence of the Li${}^{+}$.

Figure 1

The temperature dependence of the frequency shift $\delta$ of ${}^8$Li in Pt at an applied field of 4.1 T. The inset shows the resonance spectra of ${}^8$Li in Pt at 6 K (red circles) and 295 K (black squares) with fits described in the text.

Figure 2

The measured SQUID magnetic susceptibility versus temperature for the Pt foil sample for (a) temperatures below 300 K and (b) to high temperatures (filled black circles) in comparison with literature values (Ref. 32). The small Curie-like upturn below 20 K is likely due to a low level of residual magnetic impurities. The curves are fits to the high-temperature CW law (blue dashed) and to a form that saturates at low temperature (red solid) (Ref. 33).

Figure 3

The temperature dependence of the (a) ${}^8$Li and (b) muon Knight shifts in the same Pt sample. (a) A linear fit to $\chi \left(T\right)$ above 250 K [Eq. (3)] yields the heavy red curve which deviates from the data. Using the same fit, but substituting ${\chi }_{\mathrm{CW}}\left(T\right)$, reproduces the data well over the whole range (blue dashed curve). A similar scaling with the ${}^{195}$Pt Knight shift[34] yields the dot-dashed black curve. The arrow indicates the measured Curie-like contribution to $\chi$, which, if due to dilute magnetic impurities, will not contribute to the shift. (b) In contrast, $K\left(T\right)$ for ${\mu }^{+}$ fits the bulk ${\chi }_p\left(T\right)$ (red curve). Using the same fit parameters with ${\chi }_{\mathrm{CW}}$ yields the steep dashed blue curve, while fitting instead over the full temperature range provides nearly as good a fit as ${\chi }_p\left(T\right)$ (shallow blue dashed curve).

Figure 4

The temperature dependence of the $T_1^{-1}$ of ${}^8$Li in Pt. The fit of the data to a straight line through the origin yields $T_1^{-1}=7.78\left(9\right)\times {10}^{-4}$ s${}^{-1}$ K${}^{-1}\mathrm{T}$. The inset shows the ${}^8$Li time-dependent asymmetry at 25 K (blue circles) and 295 K (black squares); the lines indicate the fit to Eq. (4).