NMR in AgPt alloys

Pulsed NMR and susceptibility measurements were performed on dilute silver-rich ${\mathrm{Ag}}_c{\mathrm{Pt}}_{1-c}$ alloys ($c=0.9-1.0\mathrm{}$). The Knight shift, $T_1$ and $T_2$ of ${}_{}{}^{109}{}_{}{}^{}\mathrm{Ag}$ and ${}_{}{}^{195}{}_{}{}^{}\mathrm{Pt}$ could be observed in magnetic fields up to 5.2 T and at temperatures between 4.2 and 60 K. In ${\mathrm{Ag}}_{99.5}$ ${\mathrm{Pt}}_{0.5}$ a SEDOR experiment (spin echo double resonance) was performed. The different temperature dependencies of the ${}_{}{}^{109}{}_{}{}^{}\mathrm{Ag}$ and the ${}_{}{}^{195}{}_{}{}^{}\mathrm{Pt}$ resonances indicate the existence of $d$ holes on Pt sites even in silver-rich alloys. A graphical method in terms of free-electron phase shifts gives a value of 0.33 ± 0.07 $d$ holes which is only about 20% less than in pure platinum. This corresponds closely to the results of Sang and Myers in AgPd alloys where a similar small charge transfer is found and which thus contradicts rigid-band considerations. With a model of Dworin and Narath the ${}_{}{}^{195}{}_{}{}^{}\mathrm{Pt}$ Knight shift and the ${}_{}{}^{195}{}_{}{}^{}\mathrm{Pt}$ relaxation rate $T_1^{-1\mathrm{}}$ could be divided into $s$, $d$, and orbital contributions. Whereas $d$ and orbital terms approximately cancel numerically, $K$ and $T_1^{-1\mathrm{}}$ are dominated by $s$ effects. It is shown why this does not contradict the observed temperature dependencies of $K$, $T_1^{-1\mathrm{}}$, and the magnetic susceptibility $\chi$.