Martensitic transformation in a Cu-Zn-Al alloy studied by 63Cu and 27Al NMR.

Abstract

$^{27}\mathrm{Al}$ and $^{63}\mathrm{Cu}$ line shape, Knight shift, and relaxation rates over a wide range of temperature and external magnetic field are reported for a Cu-Zn-Al alloy displaying a martensitic phase transformation (MPT) at ${\mathit{M}}_{\mathit{S}}$=152 K. Changes in line shape, linewidth, and ${\mathit{T}}_{2}^{\mathrm{\ensuremath{-}}1}$ at the MPT are detected for both nuclei, and are found to be consistent with the local atomic rearrangement occurring at the transformation. A double structure for the $^{27}\mathrm{Al}$ NMR line is observed in a small range of temperature below ${\mathit{M}}_{\mathit{S}}$, and interpreted as the superposition of the signals arising from the two coexisting phases. It is shown that the growth of the martensitic phase during the cooling can be monitored by means of the deconvolution of the $^{27}\mathrm{Al}$ spectrum into the two components. From the analysis, it is inferred that a sudden formation of extensive regions in the martensitic phase occurs at the transition. The Knight shift and the Korringa term (${\mathit{T}}_{1}$T${)}^{\mathrm{\ensuremath{-}}1}$ are slightly different in the two phases, indicating a small increase of the density of s electrons at the Fermi surface at the nuclear sites. The enhancement factors of the susceptibility and of the spin-lattice relaxation rate do not seem to be affected by the MPT but are different when measured at the Al or Cu site, indicating a local nonuniform charge-density distribution in the unit cell. A small enhancement of ${\mathit{T}}_{1}^{\mathrm{\ensuremath{-}}1}$ is observed for both nuclei in the temperature interval in which the growth of the martensite within the austenite is detected. The anomalous contribution to the relaxation is interpreted as due to strong local charge-density fluctuations caused by atomic motion at the interfaces between the two phases. No precursor effects were detected on the NMR parameters above ${\mathit{M}}_{\mathit{S}}$, indicating the absence of a static or long-lived microstructure of the product phase and of a static short-wavelength modulation of the lattice.