# Ag-Sn-Cu Phase diagram to formulate homogenization temperature / duration

I am working on Ag-Sn-Cu alloy phase diagram to figure out the best temperature / duration to homogenize the alloy (annealing). The %age of elements in the alloy are

1. Ag - 40 %
2. Cu - 27.8 %
3. Sn - 32.2 %

(Composition can be varied if necessary but not more than 5 %)

Now the main goal is:

1. Find out the temperature / time duration at which the ingot of the above said alloy should be annealed to get the maximum Ag3Sn phase in the final ingot. Ag3Sn is the most desired phase with Cu3Sn being the second desired phase. Anyhow the main goal is to make as much Ag3Sn as possible.
2. The time duration for which the ingot should be placed at the desired temperature. (The total ingot is usually 6–7 kg)
3. [Supplementary question] Should the temperature be gradually reduced to the desired level of maximum Ag3Sn yield (or gradually increased)?

The phase diagram of Ag-Sn-Cu is as follows:

• I'm away from my relevant databases now, but two comments: phase diagrams will not help with kinetic questions, and that particular phase diagram is not the best for your question. Early next week I'll have s real answer for you. Jul 26 '15 at 17:07
• thanks @JonCuster .. Any tip to lead me in the right direction is much appreciated. Jul 27 '15 at 6:20

Much of this answer is based on a document available from NIST at Properties of Ternary Copper-Silver Systems [originally J. Phys. Chem. Ref. Data, 6(3) 621-673 (1977)] - this has an equivalent to the diagram in the question (which is just the surface of the liquidus). Now, since you are talking about balances of solid phases, it is also worth looking at the binary diagrams to discuss those.

For Ag-Sn, one gets:

For Cu-Sn there is:

Now I'll throw out a few observations: Ag$_{3}$Sn is a line compound, which in the binary system is not stable above 480C. In the NIST report this phase is referenced as $\theta$, but called Ag$_{3}$Sn($\epsilon$), causing great confusion since the $\zeta$ Ag-Sn phase is designated $\epsilon_{2}$, and Cu$_{3}$Sn($\epsilon$) is designated $\epsilon_{1}$. So, watch it...

The Cu$_{3}$Sn line compound converts to the fcc $\gamma$ phase at around 676C. The $\gamma$ phase is sometimes also referred to as Cu$_{3}$Sn (yet more confusion possible). The Cu$_{3}$Sn and Ag$_{3}$Sn line compounds are different crystal structures, and do not form a solid solution across the ternary.

For both Ag-Sn and Cu-Sn, another thing to notice is the various other phases that are in close proximity to the Ag$_{3}$Sn and Cu$_{3}$Sn line compounds. Going into the ternary system, you get questions like how adding Cu into $\zeta$Ag-Sn impacts the free energy.

So, what now? In the NIST report, on page 664 is a pseudo-binary diagram where the Sn wt% is held at 20%, and Cu wt% is varied from 0 to 80 (balance Ag). Personally I hate weight percent, but... The point that one can eventually pick out is that the Ag$_{3}$Sn phase, designated $\theta$ appears almost nowhere - there is just a little sliver way over near 0 wt% Cu and under 250C. It appears that with the addition of Cu, the $\zeta$Ag-Sn phase rapidly replaces the Ag$_{3}$Sn phase.

So, as far as I can tell, I'm not sure that the kinetics even matter here. You basically can't get much Ag$_{3}$Sn to form within the ternary parameters you have chosen - you will get $\zeta$Ag-Sn instead.

• Insightful. That explains a lot. I have been looking into other documentations and here is something which I found can be a little importance. This paper on page 735 explains the behaviour and existence of phases. Ag3Sn & Cu3Sn & liq would exist between 385 - 435. Though the Ternary parameters are not exactly wo what I'd like but still .... What is your take? Jul 27 '15 at 22:24
• And the main goal of the question is to maximize the Ag3Sn .. We can compromise on the Cu3Sn phase. Jul 28 '15 at 2:02
• Interesting. There would appear to be some contradictions between your reference and mine. The NIST ternary and pseudo-binaries are quite complicated compared to the Fima paper you cite. On the other hand, all of their data (on, say Figure 3) are in the simplest part of the ternary, away from ugliness of the low Sn content areas of the Ag-Sn and Cu-Sn diagrams. But, take their Figure 3 as fact - where does that leave you? If you try to solidify at ~25 at.% Sn and Ag rich, you will first form a liquid+$\zeta$ mix, and then hope that you can convert the $\zeta$ to Ag$_{3}$Sn. Not easy. Jul 28 '15 at 19:20
• You can't avoid passing through the liquid-zeta region unless you splat cool (perhaps). So you will get some zeta forming. Then the liquid will get richer in Sn. Once you finally hit the top of the Ag$_{3}$Sn line you will start forming it and stuff to the Sn-rich side of the curve. To convert whatever zeta you made will require Sn to diffuse back into the zeta and have the zeta nucleate and transform into zeta. A tough problem. Jul 28 '15 at 23:29
• So, you have made a good percentage of solid, but not the right phase, and if that won't transform quickly (and it won't) you have to wait for diffusion. Your approach is about as good as it gets for an ingot, but it might take a long time to convert. Jul 28 '15 at 23:51