Okay, so carbon’s main goal is to decrease part oxidation. I usually say that this is because it reacts with oxygen, making CO2, and displacing more oxygen from coming in. But that’s actually only part of the story:
Take a look at/squint at this diagram for about 20 minutes. The Y axis on the left side is the favorability of the reaction. The more negative it is, the more favorable. Meanwhile, the X axis is temperature. Basically, what this chart does it show how likely a reaction is to occur depending on temperature.
Note that every single reaction on this particular chart is negative on the Y axis (it goes from 0 to -1200): Regardless of the temperature, all of the reactions will happen to some degree. Aluminum will become aluminum oxide, copper will become copper oxide, carbon will become carbon dioxide, etc. Oxygen is a real jerk.
Note, though, that the slope of the favorability of the metals reacting with oxygen is positive: With increasing temperature, the favorability of the reaction with oxygen actually decreases (but stays negative)! Why, then, does oxygen react with metals faster at higher temperatures, even though the reaction is less favorable? Well, because favorability isn’t the same as the rate of the reaction. It is favorable for hydrogen to react with oxygen, but this does not happen unless some activation energy (heat, spark) is provided in the first place.
Now take a look at the reaction of Aluminum with oxygen versus copper with oxygen. Aluminum is way way lower, so the reaction is more favorable. The opposite case is also true: It is way more difficult to remove the oxygen from aluminum than from copper. This explains why copper metal has been used for thousands of years but aluminum only since the 1800s (and why it still takes so much energy to produce today).
Okay, now take a look at the lines for carbon: Carbon to CO2, carbon to CO, and CO to CO2.
We can see that the C to CO slope is actually negative (unlike virtually all the other lines). This means that with increasing temperature, it becomes more likely for carbon to be found in the monoxide form than in the dioxide form. The switchover happens at about 600 C. Beyond this point, in a semisealed container we’re dealing with CO instead of CO2. And that’s fantastic news! Why?
Because if something is lower on the chart, it can reduce (remove oxygen) from the thing above it! We can see that at sintering temperatures ~1200C or so, the line for formation of carbon monoxide is lower than that of iron and nickel, found in stainless, and well below that of copper. so at this temperature, the carbon (in the form of CO) can not only remove oxygen from the air, it also can remove it from the part itself!*.
The problem is that this does not work for every metal on the chart. If you look at the line for titanium, for instance, We can see that it only crosses the carbon-carbon monoxide line at about 1600 C (Beyond the melting point). Useful for purifying the raw ore, but not for sintering. At temperatures lower than this, it is more favorable to oxidize the titanium over the carbon. Also, according to the chart oxygen will still attack steel at temperatures less than 500 or so, and copper at less than 200 or so.
My idea then, and I apologize for so much exposition, is to use metal powders found even lower on the chart to act as oxygen scavengers and reducing agents at both low and high temperatures.
For instance, magnesium is way lower than almost everything else on the chart. So if it were found in the media, in loose form, it should scavenge oxygen at all stages of the heating cycle, and maybe even reduce reactive metals like titanium.
Anyway, I tried a quick experiment once with copper prints and magnesium powder:
I had three crucibles containing copper disks. One with carbon as usual, one with a small amount (1.5 grams) of magnesium powder mixed into the refractory and a bit more on top, and one with nothing at all. I debound and sintered them (final temp 1060) as usual.
One thing that was interesting was the color produced in the no-carbon condition. I used old refractory that was a bit black and it turned brown/yellow. Whereas the magnesium on the top was white (oxidized).
Here are the disks:
We can clearly see the carbon condition (bottom) is good sintered copper. Top left is magnesium condition, top right is no carbon or magnesium. The top two are very clearly not shrunk/sintered. But they’re not quite the same, either…
They snap easily and here’s what’s inside:
The nothing condition is perfectly preserved and black. The magnesium condition is delaminated and shiny and purple! I remember seeing this with my very first attempt at sintering — I ran out of carbon partway through and I got this glass-like not really sintered part. So maybe this is what happens when oxidation is prevented early on but not in the late stages.
What I would try next is a combination of more magnesium, and then both magnesium mixed in and carbon on top. That, and other metals like aluminum or even iron. Iron wouldn’t act as a reducing agent, but it would still act as a competitor for oxygen for our parts, and would be more effective at this job than carbon at low temperatures (according to the crazy chart above).
Finally, This paper has various thoughts on the topic. I think especially we should be trying to figure out some kind of kiln configuration that seals as much as possible as opposed to hacking together random lids and whatnot.