Back in September 2009 we published a Post titled, “Wetting, Braze Flow and Filler Spreading,” and for those of you who were left hanging, here is our new Post on using phase diagrams to understand Spreading.
Why do we experience excessive spreading with some filler (braze alloys)/substrate interactions and poor spreading in others? Let’s examine phase diagrams to explain the following observations:
• Cusil on Cu blushes (excessive flow)
• Cu/Au on Cu flows well
• Cu on Ni flows poorly
• Cu on Fe flows very well (blush potential)
It is widely known by many in the brazing industry, that “Cusil” an Ag/Cu eutectic, flows like crazy on copper. In this interaction, molten Cusil dissolves copper (Cu) thereby increasing the melt volume, but the solidus temp doesn’t rise at all!
The viscosity of the melt is very low so flow is rapid and solidification is controlled by solid state diffusion of Ag into the Cu substrate, ~ 100X slower than liquid-solid interactions. In this system, a huge amount of Ag needs to diffuse and subsequently a “Freeze Out” condition is virtually impossible. See diagram below:
Looking at a Au/Cu we see a braze alloy like 35Au/65Cu flows well on Cu but not excessively, why is that? In this system there is no eutectic and no solubility limit, and consequently no fixed solidus temperature. Non-eutectic melts have higher viscosity, so flow is slower than with Cusil. As the melt flows away from filler source, it continues to pick up Cu and the composition enters the “mush zone” resulting in a sharp increase in viscosity and reduced flow or what is called a “freeze out.” See diagram below:
Next, let’s look at why Cu flows poorly on Ni? Like Au/Cu, Ni & Cu are a non-eutectic with higher melt viscosity. The raised temperature is only 8 ºC below braze temperature, so the viscosity of the liquid phase is relatively high. Since the solid and liquid phases have similar composition, small increase in Ni pushes filler into the mush zone. High brazing temperature speeds up the solid-state diffusion and here we experience a rapid “Freeze Out”. See diagram below:
Why does Copper flow well on Iron? As copper melts, it dissolves iron, but the solubility of copper in iron is limited. That solubility limit results in a fixed solidus temperature of 1096C for Fe/Cu alloys that are < 91.2% Fe. Once the initial 3% of Fe is in solution, this system acts much like the classical wetting model because of little additional interaction. A considerable amount of copper would need to diffuse into iron otherwise “freeze out” is slow like the Cusil example above. See diagram below:
Let’s look closer at Cusil on copper in a system where the braze temperature is 20 ºC above the liquidus. Here the molten filler dissolves copper changing melt composition to a liquidus/braze temperature intersection. The volume of melt increases ~ 10% by eroding ~ 10% of the filler volume of copper substrate. This happens very fast, within seconds: 9% = 37% – 28% (increase in Cu in an Ag/Cu filler). If we “over-brazed” by 20 ºC: 20% = 48% – 28% (increase in Cu in an Ag/Cu filler), then twice the volume of substrate is eroded. See diagram below:
Here’s a method for predicting erosion potential for a filler/substrate combination:
Using the angle of liquidus line (ºC per % dissolved substrate) and the width of “mush zone” in % dissolved substrate (where narrower implies less erosion), we get a qualitative erosion-potential figure of merit when we take the width of the “mush zone” divided by liquidus angle squared:
Substrate Filler Liquidus Line Angle, C per % Width of Mush Zone Mush over Angle Squared
Cu CuSil 2.2 55.0 11.1 High erosion potential
Ni Cu 3.8 4.0 0.3 Low erosion potential
The erosion behavior of Cusil and 35Au/65Cu are similar but the freeze out behaviors is very different. As braze temperature is held, composition starts to move along the line towards solidus. The solid phase forms by diffusion of either Ag or Au into the Cu substrate and the liquid volume drops accordingly (AKA “mush zone”). See diagram below:
• Ag/Cu solid phase is 8% Ag, whereas filler is 72% Ag.
• Ag must diffuse into a volume of Cu that is 9 times that of the original filler (9 = 72% / 8%).
• Au/Cu solid phase is 16% Au, whereas filler is 35% Au.
• Au must diffuse into a volume of Cu 2 times that of the original filler (~2 = 35% / 16%).
• Cusil must do more diffusion “work” at a lower temperature (810 ºC vs. 1030 ºC for 35/65), which slows solid state diffusion; therefore, Cusil is much slower to “freeze out”.