Varying kinetic stability, icosahedral ordering, and mechanical properties of a model Zr-Cu-Al metallic glass by sputtering

${\mathrm{Zr}}_{65}{\mathrm{Cu}}_{27.5}{\mathrm{Al}}_{7.5}$ metallic glass thin films with widely varying kinetic stability as a function of deposition rate were synthesized by single-target direct current magnetron sputtering. Fluctuation electron microscopy and angular correlations in coherent electron nanodiffraction show that glasses with increased stability have increased nanoscale structural order, particularly of icosahedral character. The most kinetically stable film's reduced modulus was 22% higher than a glass of the same composition rapidly quenched from the liquid, consistent with increased density and improved thermodynamic stability. These results suggest that enhanced nanoscale icosahedral order contributes both to the kinetic stability of the glass and its resistance to mechanical deformation.

Figure 1

(a) X-ray diffractograms are shown for ribbon and thin films (offset for clarity). (b) FWHM and (c) first broad diffraction peak positions for the thin film and ribbon samples.

Figure 2

Calorimetry data at a heating rate of 5000 K/s showing (a) $T_{\mathrm{onsets}}$ values. (b) Kissinger analysis for thin film grown at 0.24 nm/s and ribbon sample is shown. Panel (c) shows crystallization behavior at 5000 K/s. Raw data have been smoothed and background-subtracted as described in the Experimental Details section.

Figure 3

(a) $V$($k$) measured ${\mathrm{Zr}}_{65}{\mathrm{Cu}}_{27.5}{\mathrm{Al}}_{7.5}$ MG films deposited at different deposition rates. Normalized average power spectral density of the (b) fourfold, (c) sixfold, and (d) tenfold rotational symmetry in nanodiffraction for all samples. (e) An example nanodiffraction pattern showing the polar coordinates ($k$, φ) used in Eq. (2). (f) The power spectrum of the pattern in (e) showing strong tenfold symmetry.

Figure 4

(a) Reduced elastic moduli and (b) hardness vs $T_{\mathrm{onset}}$ for thin films and ribbon samples are shown.

Figure 5

Panels (a) and (b) show gradient force images (scanning set point error) while (c) and (d) show two-dimensional surface maps for ribbon and the ∼3-μm-thick film grown at a deposition rate of 0.83 nm/s respectively.

Figure 6

Compositions of ribbon, thin films, and target source as measured by SEM-EDS are shown. Dashed lines represent nominal alloy composition.

Figure 7

Plan view SEM (top) and AFM images (bottom) of thin films grown at deposition rates of (a) 0.24, (b) 0.83, and (c) 1.19 nm/s.

Figure 8

Average angular power spectrum for $n=1\text{–}10$ for all three samples with the same vertical scale for every order.