Melting Moly

[HardMetal]

During a CTE test, a stack of pure molybdenum samples (each 0.067 inch thick) was heated to 4000F. Melting of the material was noted in the center of the samples. Moly shouldn't melt until around 4700F. So I figured that there must be an alloy or some contamination of the base metal. We did the CTE test up to 3800F and got melting. We redid it up to 3500F and still had some melting, but much less. My reaction was that there must be some alloy of moly. So I checked the solidified nugget on the SEM/EDS and found only moly with a possible trace amount of silicon (hard to tell since the moly peak overwhelmed the energy level where silicon is present). No other trace elements. Mo-5.5%Si has a eutectic at 3750F. But this doesn't explain why we still got melting at 3500F. The temperature is measured by an optical pyrometer. The material cert papers state the material is 99.95% moly.

Any ideas on what might be causing the low melting point?

Thanks for any help.

 

[Metalguy]

Your pyrometer is probably reading low. Check it at a lower temp. vs. a good TC.

 

[arunmrao]

Get the optical pyrometer calibtated or better use a thermocouple assisted temperature measuring system.

 

[JimMetalsCeramics]

Can you focus the pyrometer's field of view to the center of the stack?

 

[kenvlach]

I don’t believe pyrometer error can explain why melting occurred in the center and not the outer edges (unless if directly heating the Mo stack by electric current).

I suggest that the results could be due to oxygen contamination (not detectable by all EDS systems) in combination with the Si impurity. There is always a bit of oxygen on the surface of the Mo plates (unless you have some means of degassing within your system before stacking). During heating, any oxygen near the outer edges of the pile evaporates as MoO_x or SiO species into your inert gas or vacuum. However, oxygen on the center surfaces cannot escape, so it forms a (Mo,Si)O₂ liquid, which brazes the plates together. I don’t have phase diagrams for this system, but found a reference to the Mo-Si-O system which may be useful:

R. Beyers, “Thermodynamic considerations in refractory metal-silicon-oxygen systems,” J. Appl. Phys., vol. 56, no.1, pp. 147-152 (1984).

 

[JimMetalsCeramics]

It is true that molybdenum oxides have low melting points.

 

[HardMetal]

I believe that moly oxides sublime. There was no significant loss of dimensional propreties from the outside of the samples, so I am not convinced it's oxygen contamination. We're still looking into this, but haven't come up with an answer yet.

Thanks everyone for your suggestions.

 

[kenvlach]

Nobody suggested sufficient oxygen to cause a dimensional change.
I don’t know all of your process details (vacuum or protective gas? gettered? heating method? 99.95% purity on a metals basis or all elements?), but can hypothesize a transient liquid occurring during the heat-up.
1) In most alloys, there is some enrichment of alloying and impurity constituents in the surface layers of atoms, in addition to adsorbed oxygen.
2) The stack of Mo plates heats from the outside inward (presume conventional heating). This drives melting or volatilizing impurities to the cooler center by condensation & wicking & grain boundary diffusion, a sort of zone refining process. The Mo plates are probably polished pretty flat (R_a?) for a CTE experiment, so their interfaces are essentially oversize grain boundaries. This strongly affects the thermodynamics of impurity volatilization, maintaining a liquid to a much higher temperature.
3) The hot liquid causes a sort of brazing between the center surfaces of the plate surfaces, with some Mo dissolving into this liquid film. Continued heating causes this liquid to disappear by interdiffusion with the bulk Mo and volatilization.

Reminds me of a transient liquid brazing process I studied ~10 years ago. An amorphous, thin, nickel-rich Ni-P alloy (either tape or electroless nickel plating) is heated between the pieces of steel being joined within a protective atmosphere furnace. The Ni-P melts at 870^oC due to the L = Ni + Ni₃P eutectic. Holding at temperature causes disappearance of the liquid due to interdiffusion with the bulk steel. This can creates a bond of higher strength and temperature limit than a typical braze, more like a weld.
Did a quick Internet search; the process is referred to as ‘Transient Liquid Phase Diffusion Bonding.’

Physical Chemistry of High Temperature Technology, E. T. Turkdogan (1980) has thermodynamic data for Mo-O, Si-O, Mo-Si-O phases.

Hope this helps.
Ken