Jean-François Dionne en Ingénieurs et Techniciens, Directors and Executives, Engineers and Technicians Metallurgical Engineer • Laboratoire Dubois SA Hace 4 d · 2 min de lectura · +300

A side order of chips with that?

A side order of chips with that?

Additive Manufacturing is taking most of the place in the news. This recent fabrication process is, for now, dedicated to the much more complex (and expensive) shapes. The cost of powder is very high compared to wrought material and the infrastructures needed for AM are also costly. It's a young business with intricate microstructural details that are not always fully understood (or communicated) by the stakeholders. I'm also fascinated by AM but I would like to direct your attention here to a more "classical" problem.

For now and the near future, traditional substractive processes are very useful and cost effective to make parts. Moreover, computers are also not always needed. A good hand drawing can get you very far with a good mechanic on the lath.

Then comes machinability. Copper, steel, alumin(i)um and titanium all have different characteristics making them more or less difficult to machine. Machinability is both process and material dependant.

Improving the machinability can be achieved by different ways. My personal favourite is machining parameter optimisation including proper tool selection. Many people usually prefer chip breakers. Leaded copper alloys and resulphurised steels are good examples of materials modified to improve machinability.

These impurities in the metal (Pb and MnS for example) make wonders for the machining part of the whole process. But what about the following processing steps like plating? The corrosion properties? The mechanical properties? 

The perfect material does not exist, it's always a question of compromise. These machinability improvement elements have physical and chemical effects on the final component whether you like it or not. I'm no guru on this matter, read more on the topic here.

I wan't to present you an example from my recent experience. I was asked to explain the brittle behaviour of a component made in a martensitic stainless steel (1.4197, X20CrNiMoS13-1, ~420F). 

The parts are used in the quenched and tempered state. The geometry hasn't changed in years, yet the last production batch breaks unexpectedly: during assembly, after assembly, during shipping, many parts break but not necessarily all of them.

Observation of the fracture surface shows a intergranular failure (above pictures). Hardness is good and chromium carbides are absent at the grain boundaries. I was not expecting this. 

So I looked into the chemical composition: in specification. What happened?

Reading on steel metallurgy we can find that sulphur has been a great deal since long (Titanic). Manganese is nowadays added to steels to improve its properties but also to capture sulphur. Manganese avoids the formation of the notorious iron sulphides that decrease the properties of steel. Iron sulphides precipitate at grain boundaries, have a low melting point and have poor mechanical properties.

Sulphur in the molten steel won't just react with manganese because you asked politely. There needs to be a certain amount for it to react, else it would be too easy.

In the books, there is mention of a critical Mn:S ratio needed to suppress the formation of iron sulphide. I found a reference giving 8:1. This value is probably valid for low alloyed steels and I'm unaware how the ratio changes for 13%Cr steel. I'll let my good friend Jean-Philippe do the chemical-thermodynamics calculations/lesson on this (I hope he won't mind). The cast certificate in my case gave 6:1.

Back in the lab, I took a deeper look at my cross-section in the SEM to find that the sulphides give a significant EDX iron peak. This was a surprise because manganese sulphides usually give almost nothing else than manganese and sulphur. 

So there is my theoretical failure mechanism: iron sulphide precipitation. I never would have thought. Of course I can't be 100% sure this is the only reason for failure in this case. Failure problems have many components and some are more important relative to others. But at least this element of knowledge I've learned goes in the right direction to explain the occurrence of the problem given the elements at hand.

End note:

Machining grades are often fine, optimising your processes to avoid them is even better. You get to learn many things from the trials, and your part quality gets globally better. 


This article was originally published on LinkedIN a couple of days ago.