Liquid Metal Corrosion of Mo and TZM in Al Melt: Mechanism and Kinetics

Abstract

Liquid metal corrosion of Mo and TZM (typically Mo with 0.5 pct Ti, 0.1 pct Zr, 0.02 pct C) in aluminum melt was investigated at temperatures between 680 °C and 720 °C for 0.17 to 10 hours using a dynamic immersion test with rotating specimen. A layer of intermetallic phases identified as Al₁₂Mo, Al₅Mo, Al₂₂Mo_5, and Al₈Mo₃ using energy-dispersive X-ray spectroscopy (EDX) is formed. Layer growth was found to be reaction-controlled for the first 8 hours with a growth constant of 2.0 × 10⁻⁸ m/s for Mo in Al at 700 °C. A solubility of 0.23 wt pct Mo in Al at 700 °C and a corresponding dissolution-rate constant of 3.3 × 10⁻⁵ m/s are found. For processes where the melt is permanently renewed, a more severe mass loss by dissolution is expected.

Appendix

By normalizing all mass losses to an immersion depth of 39 mm (≙ immersed area à of 13.04 cm²), the effect of different immersion depths and thus varying areas of attack can be eliminated (see Eqs. [A1] through [A3]). The volume of the growing layer is proportional to the immersed area A and thus the mass of Mo in the layer can be normalized by Eq. [A1].

$$\begin{array}{c}{\widetilde{m}}_{\text{layer}}^{\text{Mo}}=\frac{{m}_{\text{layer}}^{\text{Mo}}}{A}\cdot \tilde{A}\end{array}$$

(A1)

$${\widetilde{m}}_{\text{melt}}^{\text{Mo}}=\left\{\begin{array}{cc}{m}_{\text{melt}}^{\text{Mo}},& c\ge {c}_{\text{s}}\\ {c}_{\text{s}}-\left({c}_{\text{s}}\cdot {\text{e}}^{\frac{\tilde{A}}{A}\cdot \text{ln}\left(1-\frac{c}{{c}_{\text{s}}}\right)}\right){\cdot m}_{\text{melt}}^{\text{Al}},& c<{c}_{\text{s}}\end{array}\right.$$

(A2)

$${\widetilde{m}}_{\text{loss}}^{\text{Mo}}=({m}_{\text{loss}}^{\text{Mo}}-{m}_{\text{melt}}^{\text{Mo}})\frac{\tilde{A} }{A}+{\widetilde{m}}_{\text{melt}}^{\text{Mo}}$$

(A3)

For the dissolved material in the melt, the normalized value cannot be calculated exactly. Two cases can be distinguished (Eq. [A2]). When the measured concentration of Mo in the melt is equal or greater than the maximal solubility, the melt is saturated and the influence of immersion depth on dissolved material is negligibly small. When the concentration of Mo in the melt is smaller than the maximal solubility, the normalized value can be derived from Eq. [9]:

$$\begin{array}{c}\text{ln}\left(\frac{{c}_{\text{s}}}{{c}_{\text{s}}-c}\right)= {k}_{\text{dis}}\cdot \frac{A}{V}t\end{array}$$

(A4)

$$\begin{array}{c}\frac{{k}_{\text{dis}}}{V}t=\frac{{\text{ln}}\left(\frac{{c}_{\text{s}}}{{c}_{\text{s}}-c}\right)}{{A}}\end{array}$$

(A5)

When immersing the normalized area Ã, Eq. [8] yields the concentration $\widetilde{c}$:

$$\widetilde{c}= {c}_{\text{s}}\left(1-{e}^{- \frac{{k}_{\text{dis}}\tilde{A}}{V}t}\right)$$

(A6)

Inserting of Eq. [A6] yields

$$\widetilde{c}= {c}_{\text{s}}\left(1-{e}^{- \frac{{\text{ln}}\left(\frac{{c}_{\text{s}}}{{c}_{\text{s}}-c}\right){\tilde{A}}}{A}}\right)$$

(A7)

and applying of Eq. [18]

$${\widetilde{m}}_{\text{melt}}^{\text{Mo}}={c}_{\text{s}}\left(1-{e}^{ \frac{\tilde{A}}{A}{\text{ln}}\left(1-\frac{c}{{c}_{\text{s}}}\right)}\right)\cdot {m}_{\text{melt}}^{\text{Al}}.$$

(A8)

For the total mass loss, the contribution of dissolution and layer growth must be normalized separately (Eq. [A3]). The measured mass dissolved in the melt is subtracted from the total mass loss, leaving the mass in the layer that is normalized with Eq. [A1]. Afterward, the material dissolved in the melt, normalized with Eq. [A2], is added. For some samples with long dipping times (6, 8, 10 hours), the composition of the melt was not measured. Here, it was assumed that the melt is saturated with the maximum solubility.