Carburizing experiments

This dwells more into the realm of metalworking, but perhaps it is interesting to some people...


Since I have access to a decent electrical furnace with good temperature control, I thought to try some case hardening experiments just for the fun of it. Basically one takes a piece of low carbon steel and chucks it into a reasonably tightly closed metal container that is filled with a mixture of carbon bearing material and some activators. Then heats the lot up to a predetermined tempetarture in the range of 700-1000C and keeps it soaking there for a period of time. Usually from a few hours to in excess of ten hours depending on the expected case depth.
What will happen is that carbon begins to diffuse into the metal aided by the heat and activators and turns the simple cheap and soft material into a rather interesting composite that has a heat treatable outer case which is hard, but the inner core shall mostly remain soft and ductile.
Depending on the choice of activators and the heating time, the carbon enriched zone depth can be controlled and thus one can turn a simple mild steel component into something that has a little more function and value.


Fig.1 A simple mild steel compartment to house the carburizing mixture and the workpiece(s).




Fig2. Pieces of old utility knife blade and a mild steel pipe were chosen as the victims of carbon enrichment

The carburizing vessel was packed with the following concoction:

10% wood charcoal
10% SrCO3
80% Na2CO3

The carbonates act as activators for the carbon diffusion process and help it along.

The thermal carburizing treatment was conducted according to following
regimen:
700C for 2 hours
raising of temperature to 950C during 1 hour
keeping at 950C for 3 hours
slow cool off in the furnace (about 12h)



Fig. 3 Specimens being cooked at 950C in the furnace

Since both the mild steel pipe and the blades had a thickness of only 1mm, I expected them to carburize completely with a uniform high carbon content throughout the whole of the material. I did not have to be dissappointed, all specimens had acquired a very high carbon content (probably well in excess of 1% C). This was confirmed with a quick water quench from about 900C and scraping /spark tests. Specimens were extremely brittle after the quench, even to the point where it was impossible to form a cutting edge on them. The material would simply fall apart during sharpening. A bit of annealing fixed that nicely. Metal files could not touch the quenched samples and they were hard enough to leave a scrape mark on lathe hard jaws.
The utility knife blades already had a decent carbon content before carburizing (0,8%C SK-5 steel), but after the treatment their carbon content had clearly risen even higher.
The blade would distort during water quench heavily, that will not happen with original SK-5 steel (at least not to my experience so far), but is prevalent in steels that have a carbon content in excess of 1% (like russian Y13A 1,3%C).



Fig. 4 Carburized and water quenched utility blade sample displaying a massive deformation after quench




Fig. 5 Carburized mild steel pipe being formed into a circular cutter on lathe and displaying a characteristic high carbon steel spark


To assess the formed case depth in the current heating regime I also conducted a test with a beefier metal sample. A mild construction steel round stock of 19mm diameter.




Fig. 6 The mild steel round stock post carburizing treatment, displaying a visible carbon enriched zone at the outer edge after facing off and being etched with nitric acid

The visible darker edge has a depth of about 1,5-1,6mm when measured, but spark testing revealed that the higher carbon zone extended several millimeters deeper than the visible etchant marking would suggest. That is quite surprising as I would have expected the case depth to be much shallower under these conditions. On the order of about a millimeter at best.....go figure


To try something more exotic I also decided to put a piece of stainless alloy through the carburizing treatment, but on a modified heat regime:

950C 2 hours
slow cool off in furnace (about 12h)

The reason being that I chucked the additional sample into the furnace when the mild steel samples had already gone through the 700C preheat and were an hour into the 950C treatment.

The exact stainless alloy designation was unknown, but it was an austenitic (non magnetic) type like 304.

The result was quite interesting....literature suggests that stainless alloys form a carbon diffusion barrier at higher carburizing temperatures (in excess of 600C) and hence the achieved case depth would be very shallow. To counter that and allow a deeper case to be formed, lower temperatures and very agressive activators (halides, hydrogen chloride etc.) are used. Furthermore the formation of chromium carbide on the surface is known to deplete the alloy of chromium and hence promote corrosion of the treated part as no more is available to form the protective oxide layer on stainless surface.


The sample came out of the carburizing vessel after the slow cool off bearing a dark grey hard coating without any further heat treatment. The surface could not be scratched with a metal file. Further heat treatment at 900C following a water quench was tried on a separate piece cut off from the sample, but seemingly it did not alter the hardness of the coat or substrate material. This suggests that the outer hard coating is mainly composed of carbides.
The sample had not formed any scale, pitting or other surface defects during the carburizing treatment. The hard coat had preserved all the previous details...light scratches, cutting marks etc and except for the color change one could not tell that it had been through a 950C heat cycle.


Fig. 7 The stainless alloy sample after carburizing with one edge ground off on a diamond wheel and polished to reveal inner details


Fig. 8 The stainless sample post carburizing and grinding to reveal the case/base alloy interface

The camera does not capture the case depth and to be honest it was really hard to notice, but the visually affected depth of the sample was about 0,5mm. There was a faint frame around outer edges of the sample extending below the actual carbide bearing hard layer which was very thin...perhaps on the order of 100 microns, but without special instruments it is difficult to assess.
Which was also ineresting was the effect that the hardness of the inner core material of the sample below the carbide layer and the faint frame indicating possibly carbon diffusion or other structural changes had risen quite considerably also and seemingly uniformly throughout the cross cut of the sample. The initial stainless alloy was quite soft (comparable to mild steel) and could be worked with a metal file easily. The carburized sample was much tougher in the core region and one had to apply quite a bit more effort to leave file marks on the sample.

I do not have the proper means at the moment to measure surface hardness quantitatively, so please forgive the anecdotal comparative methods that I had to use, but it is amateur metal research after all

Quote: Originally posted by markx

10% wood charcoal 10% SrCO3 80% Na2CO3

Quote: Originally posted by Fulmen

Quote: Originally posted by markx   10% wood charcoal 10% SrCO3 80% Na2CO3 I do wonder how you ended up with your mix, usually it's 90% charcoal and 10% activators. But it seems to work... As for testing you should try 5% nital etchant, this will color the carburized layer dark. r

I like to use such rogue approach from time to time (also in my daily development job): reverse ratios and head in the general opposite direction of previous attempts.
Not too seldom it has revealed interesting results

I also plan to do some boriding experiments in borax/reducer mix (Al powder should work as the reducing agent). Of course the sale of borax has just been prohibited to general public....for gods sake where does it all end.

Quote: Originally posted by MrHomeScientist

Really? What's wrong with borax?

Quote: Originally posted by MrHomeScientist

Really? What's wrong with borax?

Update regarding boriding attempt:

I took a piece of the same mild steel pipe that was also used as the precursor in last carburizing attempts and added another machined mild steel part and a disc of 304 stainless.
All three specimens were packed into the carburizing vessel and packed with the following mixture:

10% Al powder (PAP-1)
90% Borax

The vessel was placed into the electric furnace, heated to 1000C and kept there for 2 hours. After that the lot cooled down overnight in the furnace.

During the heating phase one could notice faint green flames escaping the boriding vessel through the sealing surface of the lid under slight pressure when temperature had reached an excess of 900C. Clearly there was a reaction going on inside. The furnace door was kept shut for the rest of the heating cycle to avoid boron bearing fumes entering the lab. Furnace exhaust took care of the off gases. Hopefully

The next morning everything had cooled down to 50C, the vessel was removed from furnace and pried open. The metal parts on the inside of the vessel had melted into a solid cake of borax on the bottom and getting them out from this mess proved to be a challenge.
After about an hour of sokaing the vessel in hot flowing water it became possible to remove the parts.




Fig. 1 The 304 stainless disk and a mild steel piece were removed first (the stainless looks very much like after the carburizing treatment: a light grey hard coat on surface)




Fig. 2 The mild steel pipe was lodged tight in the vessel and was slightly deformed by the violent removal process




Fig. 3 The mild steel pipe section after removal and clening with a wire brush on angle grinder (there is ununiform surface appearance with shiny and dull areas. The dull areas seem harder and wire wheel does not polish them)


Fig. 4 The lid of the boriding vessel displaying a neat dark grey coating on the side facing into the boriding vessel




Fig. 5 The lid of the vessel after an attempt to grind with a coarse dremel emery wheel (no success in the middle area that had the grey coating apparent)

All the parts were removed and cleaned in the soft annealed state that they had acquired during the slow cooling of the furnace and were not heat treated to alter their hardness.
All parts had a very hard outer layer formed on them although the core was annealed and soft. The pipe section could be deformed even pressing between bare fingers. But metal file would not cut into the surface and the hard layer could not be removed with a high speed dremel emery wheel. The abrasive would simply skate over the surface without striking a single spark and just leaving a shiny spot on the part. The dremel wheel would slightly deform the softer core under the hard surface layer and thus flatten some fine details previously visible on the part, but would not cut into the metal.
So it would seem that a boride bearing layer has indeed formed on the parts under the conditions employed. Although the layer is inherently thin, I would believe perhaps on the order of just some microns considering the rather short residence time of just 2 hours at 1000C.
Perhaps it would also be beneficial to place the boriding vessel into a preheated furnace at already on the working temperature to minimize possible side reactions and losses during heat up time.

Quote: Originally posted by Fulmen

Boriding is comparable to nitrocarburizing in that it forms hard iron boride rather than martensite like carburizing (and regular heat treatment) does. I don't know if you can get the process to work at lower temperatures, but it would be worth a try. This would allow you to surface harden heat treated steel without destroying the original HT.

Quote: Originally posted by Fulmen

Actually many low alloy steels are tempered at those temperatures. Currently I'm working on a rifle bolt made from 4340, tempered at 550°C it has a yield strength of appr 1100MPa. Adding a tough wear-resistant coating on such parts can be very beneficial, that's why ferritic nitrocarburizing has become so popular.

The downside being that nitrocarburizing often utilizes molten cyanide as the main component of the bath....although I think that some cyanide free compositions are also available.
Retaining the original heat treat with boriding techniques might still be tough, but if the process could be driven on lower temperatures then that would be great

Attached article describes electrochemical techniques of producing borided surfaces at lower temperature ranges using composite molten salt baths:

Attachment: MoltenSaltBathsElectrochemicalBoriding.pdf (739kB)
This file has been downloaded 598 times

Quote: Originally posted by macckone

I wonder if cyanoacrylates might be good for nitrocarburization. Acrylic plastics are readily available as is superglue although methylcyanoacrylate has a lower amount of hydrogen than ethylcyanoacrylate.

Allrighty then....I did try the nitrocarburisation experiment with cyanoacrylates.

A dead center is machined out of mild steel, the tip is covered in cyanoacrylate and cured with sodium carbonate in several layers. The lot is then packed into the carburizing vessel, buried in sodium carbonate layer (tip downwards) and off into the furnace it all goes....











I did not have time today to wait for countless hours, so the merry hell was fired for two hours at 950C with the vessel placed into the furnace when it had reached the setpoint temperature. For practical purposes all of the organics shall decompose way before the 2 hours are over. All that wants to fly out will be gone and all that remains shall be in the vessel. So I figured that for first try the 2 hour heat cycle should be ok.




The vessel was removed from the hot furnace and placed into a bowl of water to cool it down. Not a full quench, but rather a slow boil down to room temperature. Then the lot was washed under flowing water to loosen the lid and out came such a "thing".



The whole contents of the vessel smelled precariously like almonds, so I flushed everything with coupious amounts of water to clean up any possible cyanide that may have formed in the vessel and gotten trapped there.
In the vessel bottom one can see the charred shadow of the cyanoacrylate resin in a solidified pool of sodium carbonate. So if nothing else then at least a thin high carbon case has been formed on the dead center's tip area.
There is a slight coloration difference between the area that was covered in cyanoaclrylate and the rest....so something. Sure enough when I tried the file test, both the tip and the tail of the dead center were soft. Does a nitrocarburized layer need additional heat treatment or should it be hard even without a oil or water quench? I'm not sure....



Sooo....I heated the dead center up to about 850C in a propane flame and water quenched it. Well the tip area hardened and the tail remained soft. So there is a case, but I have no idea wether it is a nitride/carbon case or just a carbon case.....

Below one can observe two dead centers I made...the smaller one is machined from high carbon steel and heat treated with the same water quench, the bigger one is the subject of the nitrocarburizing experiment. For practical purposes both are very hard in the tip area after a water quench and with simple methods I can not really evalutate a difference between them.



Sigh.....well it is inconclusive at best. I'm sure the same outcome would have been reached if I had wrapped a garbage bag around the tip of the mild steel dead center...or any other carbon source of organic nature for that matter.

Since I was really not inclined to destroy the functional heat treated and cleaned up center with grinding and etching tests, I opted to black oxide coat them both and just use them in the lathe to turn stuff between centers as intended.


Two dead centers and two cutting tool holders for mini lathe post degrease and pre oxide coat...into the magnetite bath....and then out. A thing of beauty




In fact I hoped that the black oxide coating process would reveal the differences in surface structure if there was a nitiride case at the tip of the bigger center ( a difference in the coloration shade of the coat, the sheen or dynamics of formation). But really it all just coated black quite evenly ....





More experimentation is needed....and in milder conditions! The 950C setpoint is really overkill for nitrocarburizing and really it should be done around 550-750C region...

Quote: Originally posted by royalallen

The temperature is way too high for nitrogen induction

Indeed....I also repeated the procedure with mild steel gadgets at 500C (2h in furnace). As far as I can recognize from a spark test a carbon case has definitely formed at the surface of the metal samples. The parts became heat treatable and hardened considerably on the surface after a water quench.
Straight out of the furnace and after cooling (without any further heat treatments) the parts were perhaphs slightly harder at the surface than untreated metal from the same source, but the layer was definitely very thin.
I shall add pictures later.
Surface appearance hinted little to nothing about a definite change having occurred in surface layer....part were perhaps a little darker than before treatment, but still mostly shiny and seemed uaffected as far as any scale forming or other changes. Which is a good thing as far as keeping dimensional tolerances goes

Fig.1 The test bodies pre cyanoacrylate treatment

Two sections of mild steel pipe (only one is used in experiment) and a disc of high strength threaded rod center section with the tread machined off. I do not know the exact compositions of the alloys, but the mild steel pipe I trust is mostly iron with less than 0,3% carbon as it is not heat treatable. The high strength threaded rod is definately a high carbon alloy judging by spark test and it definately also contains alloying elements of carbide forming nature (most possibly chromium). It is one of the least machinable steels that I have encountered. It just eats cutting tools....tool steel and carbide alike. It also work hardens to the degree where even honed carbide is unable to touch the surface once it has been hardened by a dulled tool skidding over the workpiece.


Fig.2 A view into the sodium carbonate filled reaction vessel after 2h in 500C furnace. The cyanoacrylate encrusted samples are buried in the depth of the vessel.


Fig 3. The samples after the supposed nitrocarburizing treatment. Both look very little affected by visual assessment.

The samples were "scratch tested" with file and showed a slightly increased surface hardness to pre treatment condition. It was still possible to cut both with a file, but a very thin hard surface layer was covering them (supposedly nitride bearing layer).
The mild steel pipe was spark tested and showed a definite high carbon signature. Unless nitrogen inclusion produces the same spark (not that I know of) then there must have been some carbon migration into the samples even under these mild conditions.
The mild steel sample was also subjected to a water quench from around 850C after which the surface of the sample hardened to a degree where it was no longer possible to cut into it with a metal file (most probably a carbon case?).
The high carbon steel sample was not subjected to additional heat treatments as i would not have shown any recognizable effects (it was highly water hardenable already in the pre nitrocarburizing state).

Quote: Originally posted by macckone

This is really cool work. Obviously ammonia nitriding would be more reliable than this if you could work with ammonia gas. But the fact that you do get nitriding with the methylcyanoacrylate is cool.

Thanks for the kind words!
I'm not totally sure that I get a nitriding action here....it looks more like a carbon enrichment type of process is mostly happening. With perhaps a very slight nitriding because the surface layer of the samples was harder than before the treatment. It is still not nearly comparable to the boriding treatment...that case was epic in it's pure hardness.

I think the main cause for the weak nitriding action is that the gaseous reactants just escape the vessel too quickly and not much interaction occurs. I'm not inclined to use a closed vessel for the process and create a possible rupture in the furnace spilling cyanide gas into the lab. I trust my lab assistant would not hesitate to cast me into concrete for that kind of a stunt