Mixing effects in a ternary Hf-Zr-Ni metallic melt

We study the effect of the substitution of Zr by Hf on the dynamical behavior in the ${\mathrm{Zr}}_{36}{\mathrm{Ni}}_{64}$ melt. A reduced measured self-diffusion coefficient and a higher measured melt viscosity for an increased amount of Hf were observed. The ternary ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ melt, which exhibits a pronounced deviation from Arrhenius behavior over a studied temperature range of 550 K, can be accurately described by the scaling law of mode-coupling theory (MCT) with almost equal parameters for the self-diffusion and the viscosity. Although we only substitute alloy components with a nearly equal atomic size and the measured overall packing fraction remains almost unchanged, the dynamics in ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ are slower compared to ${\mathrm{Zr}}_{36}{\mathrm{Ni}}_{64}$. This corresponds also to a higher critical temperature $T_c$ and might be induced by different chemical interactions in the melts. The increased $T_c$ results in a significantly smaller difference between liquidus and critical temperature $\mathrm{\Delta }{T}_{\mathrm{LC}}=T_{\mathrm{L}}-T_c$ for the ternary melt in comparison with ${\mathrm{Zr}}_{36}{\mathrm{Ni}}_{64}$, which may favor the glass formation in the ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ melt.

Figure 1

Upper panel: Normalized intermediate scattering functions for ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ at 1550 K and $q=0.85{\AA }^{-1}, q=1.05{\AA }^{-1}$, and $q=1.25{\AA }^{-1}$. The solid lines represent the stretched exponential fit by formula (1). Lower panel: Normalized intermediate scattering function at 1355, 1400, 1470, 1550, 1590, 1675, 1750, and 1905 K for $q=1.25{\AA }^{-1}$.

Figure 2

Time-temperature scaling (master curve) at $q=1.05{\AA }^{-1}$ for ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$. $S(q,t)$ and the time $t$ have been normalized by the calculated $f_{\mathrm{q}}$, respectively, $\langle {\tau }_{\mathrm{q}}\rangle$ from the fit with equation (1).

Figure 3

Inverse structural relaxation time for ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ at 1355, 1400, 1470, 1550, 1590, 1675, 1750, and 1905 K. The linear fit (solid lines) has been performed using (0,0) and values between $q^2=0.56{\AA }^{-2}$ and $q^2=1.69{\AA }^{-2}$.

Figure 4

(a) Self-diffusion coefficients for ${\mathrm{Hf}}_{35}{\mathrm{Ni}}_{65}$ [11], ${\mathrm{Zr}}_{36}{\mathrm{Ni}}_{64}$ [39], and ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ measured by QENS. The black dashed-dotted lines represent the Arrhenius fit and the blue dashed lines the scaling law fit of MCT. (b) Measured viscosity for ${\mathrm{Zr}}_{36}{\mathrm{Ni}}_{64}$ [28], ${\mathrm{Hf}}_{35}{\mathrm{Ni}}_{65}$, and ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ by using the oscillating drop technique in an electrostatic levitator. The black dashed-dotted line represents the Arrhenius fit and the blue lines the scaling law fit of MCT. (c) Packing fraction for ${\mathrm{Zr}}_{36}{\mathrm{Ni}}_{64}$ [40], ${\mathrm{Hf}}_{35}{\mathrm{Ni}}_{65}$ [11], and ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$. The blue vertical line reflects the experimental error of the measurement.

Figure 5

(a) $D·\eta$ for ${\mathrm{Hf}}_{10}{\mathrm{Zr}}_{25}{\mathrm{Ni}}_{65}$ as a function of temperature. The red dashed-dotted line is the calculated mean value. (b) Hydrodynamic radius $a_{\mathrm{hydr}}$ derived from the Stokes-Einstein relation. The red dashed-dotted line is the mean value.