should metal be scuffed or as flat/polished as possible for brazing/soldering?

[coppercone2]

So assuming you setup the metal correctly (i.e. for a silver solder joint say 5 thousanths), people just say it needs to be clean.

Do you get a better joint if you make it as flat/polished as possible, or do you want some particular surface finish?

I thought the brazing occurs on a microscopic level (where the braze enters the fissures that form in the material as its heated). So if there was a impulse I am thinking that a level polished surface would transfer it better then a scuffed one? Is it possible it would make it stronger towards beatings?

If you made both, how would mechanical wave transfer through both interfaces appear? In terms of the metrology of impulse transmission through metal (its gonna be a small difference)?

I also imagine the flow might be janky if its not polished?

[KL27x]

Yes, you should try to get more fresh air and sleep. Then ask yourself a new question, tomorrow. Do I still care?

IF you do still care, maybe you will try it instead of thinking about it. And when you post, you will have some news.

[coppercone2]

how am I going to try it? You know what kind of equipment would be necessary to test the difference?

I am asking a theoretical question about metallurgy. I am curious about the behavior of braze/solder joints.

I will not be able to detect something like a 0.1% difference. There is no way. Unless its big.

I want to know what happens with the mechanical wave in the say 100 microns (~200 grit finish) between the bulk braze compound and the boundary where the metal dendrites peak average is formed from finishing. Bulk thickness of braze is something liek 150 microns

This boundary gets smaller the more polished it is. I also want to know what difference flatness makes. I am going to assume a very polished finish is something like 1 micron peak (Ra).

Why are you busting my balls?

[David Hess]

My experience is with silver brazing where usually the base metal rips before the joint or filler so other than having clean surfaces, I do not think it matters.

But it would be easy enough to test by brazing some joints under different conditions and measuring the fracture strength in bending or tension.

[coppercone2]

do you think it would be significant enough to detect without very expensive equipment? I am not interested in getting more strength but understanding the mechanism.

I never really busted open a brazed joint (usually their not flat).

Also, would it behave differently with regards to flex cracking ? like from heavy vibration.

[David Hess]

do you think it would be significant enough to detect without very expensive equipment? I am not interested in getting more strength but understanding the mechanism.

If there is a significant difference, then I am sure you could detect it using a torque measurement or perhaps just some weights on a long lever arm.  Now that I think about it, the torque measurement seems particularly easy using a bolt silver soldered to a substrate and then twisted off, unless the bolt breaks first.

Also, would it behave differently with regards to flex cracking ? like from heavy vibration.

That is a completely different question which will depend more on modulus of elasticity so unless the surface preparation creates a crack, I doubt it matters.

Honestly if the surface finish matters, then I think you might be better off using a stronger attachment method like welding although that has its own problems.

[IanB]

I thought the brazing occurs on a microscopic level (where the braze enters the fissures that form in the material as its heated).

This thought is incorrect. In soldering and brazing there is a dissolution, dissolving, alloying of the two metals where they are in contact. The brazing or soldering alloy undergoes a physical/chemical reaction with the metal being joined and at the interface there is a diffusion of metal atoms across the boundary.

For example, if you solder to copper then there is a gradual transition between 100% copper and 100% solder across the joint, with a 50% copper/50% solder alloy at the interface. The copper dissolves into the molten solder just as various metals like gold dissolve into mercury.

Because of this, it scarcely matters whether the surface is smooth or rough. It won't make any essential difference to the joint strength.

[coppercone2]

so its like semiconductor diffusion?

I thought I read somewhere that the metal 'pored up" when it got red hot and the braze flowed into the 'fissures' that don't exist outside of a expansion phase?

And that this was the reason you should not rebraze joints, because you can get excessively deep 'dendrite' formation.

It was in a white paper I think? And that after cooling this results in a braze joint having a 'weakness' similar to hydrogen imbritlement where the crystal structure is being pressured by different tempco stuff. Like kinda a velcro type effect.

I understood what you describe to be soldering. I thought brazing was a bit different. If I ever get lapping plates and a metallurgical microscope I can check.

Are you saying its 100% atomic scale diffusion? And if you do lapping samples you will just get different alloy concentrations?

[IanB]

I understood what you describe to be soldering. I thought brazing was a bit different. If I ever get lapping plates and a metallurgical microscope I can check.

There is no difference between soldering and brazing other than the typical temperatures and alloys involved. Brazing and soldering are essentially the same process.

If done properly, there is no "join" that can be separated afterwards, such as there is with gluing or adhesives. The joint is a continuous transition between base metal and brazing alloy at the atomic scale.

If you separate a brazed joint then the metal will fracture at the weakest point, usually in the filler metal since this is softer and weaker than the base metal. It will not be possible to separate or remove the filler metal from the base metal surface since it will have become one with the base metal.

[SparkyFX]

The flux plays a huge role how the solder flows and really covers the maximum area.
So it might not matter at all, as the solder has a surface tension that can only bridge gaps of a certain size after the flux is gone.

[flolic]

From my experience, base metals to be silver brazed needs to be only reasonably clean. There is absolutely no need for polishing.
You must use enough flux (Borax) and then the filler will flow beautifully.

[coppercone2]

So the depth of diffusion ensures that the roughness is dissolved into a even metalsolution?

I think i got confused because of the terminology. I heard people say penetrate and mix but they should be saying diffusion and dissolving/solution forming.

The terms i heard make me think of non atomically uniform phenomena.

Is this still true if carbide is brazed?

So can you extrapolate that the better the metal solubility is the less flatness and surface texture has an effect on tge concentration gradient of the cooled joined solids?

Does this have something to do with raults law ?

[eKretz]

Yes it is true.  Tungsten carbide bonds via brazing by the same mechanism as steel or any other metal. The layer at the junction is basically a mix of the two -  the braze or solder alloy and the base metal.

[coppercone2]

Hmm so how should i think about it.

1) the solid metal at the molten silver boundary dissolves into the silver and diffuses to the bulk molten silver in an attempt to make a solution (i think so)

2) the silver enters the hot metal like a sponge ?

Do both happen kinda with some grains dissolving faster then others in the silver then silver surrounds them like a salient or peninsula and acts to dissolve them with more surface area or is it a even layer that moves uniformly(if viewed as a transient process)?

Like whats the topology look like if you just graphed the interface geometry over time?

[coppercone2]

Maybe a force diagram of sme test point atoms in a imaginary interface would help me with all concievable forces even if irrelevant (say to 100ppm of peak)

I assume gravity and bouyancy are irrelevant

[T3sl4co1l]

Yes it is true.  Tungsten carbide bonds via brazing by the same mechanism as steel or any other metal. The layer at the junction is basically a mix of the two -  the braze or solder alloy and the base metal.

I don't think this is fair -- but, well, it depends.

Cemented carbide (the most common kind) is WC in a Co matrix.  Co has a little solubility for C and W but not very much even at the melting point.  You'd have to go very high indeed to get a full melt.

I think it's more accurate to think of cemented carbide as WC particles soldered together with a Co filler.  Little diffusion or dissolution takes place, it's primarily a metal-ceramic interface.  (I don't know the mechanics of these sorts of bonds.  They certainly seem strong enough, as many cermets can attest to.  It's some kind of mixed ionic, covalent or metallic bond -- for sure, orders of magnitude stronger than just proximity (van Der Waals) forces.)

In turn, when you braze an insert to a tool, you're using a brazed joint, which will have some Co and filler (and Fe (or other base metal) and filler) interaction on each joint, and that will probably be more intimate than a sharp (non-diffusive) interface.  Typical fillers are in the Cu-Ag-Zn family (including yellow brass, "silver solder"*, and low-melting "silver solder") and Ni-Ag (or Ni-B I've heard of, too) family.  These fillers and base metals all have mutual solubility.

*By convention these are called "silver solder", but they're actually braze.  Go figure.  So I'm quoting them for clarity.

Soft soldering, is typically done at such a temperature and rate, that the diffusion layer is very thin, and this is usually a good thing as a lot of soft solders and base metals have brittle intermetallics that, if allowed to grow, would weaken the joint.  The Cu-Sn system has a lot of these; some are very strong (mostly on the Cu side, hence what's so great about bronze), but the intermediate ones tend to be weaker.  You can also soft solder metals that are very different indeed, e.g., Ti with Sn filler (if you don't mind the extremely toxic and corrosive flux required ).  Although this is a neat example, as Sn is standard in a number of Ti alloys, having an analogous effect as, say... Cr in Fe, maybe?

Intermetallics aren't obligatory, though.  Cu-Pb is a nearly immiscible system (no intermetallics, only a little solubility of Cu in Pb at various temperatures, and vice versa; separate liquid phases until quite high temperatures), but it's just fine for soldering.  Again, it's not that there's necessarily a diffusion reaction (reaction as in, intermetallics are formed), or necessarily dissolution even, just that the atomic bond at the interface is usefully strong.

Now, I could imagine a rough surface -- lots of "tooth", and lots of microscopic surface area -- is more suitable for soft soldering, given the tendency for intermetallics, or for weak interface bonding -- this might be interesting to test, and is pretty easy to do.  Prepare some metal surfaces, of various metals (say, Fe, Cu, Al, Ni would be interesting?), prepare joints with various fillers (Sn100, Pb100 and Sn63 might be interesting datapoints), and test the tensile and shear strength.  Also, for surfaces that are prepped with strong linear grooves (e.g., 150 grit + belt sander), it would be interesting to test them in parallel and crossed orientations; and for shear, testing groove to shear force angles as well.

Polished surfaces are tricky.  It's easy to make a smooth surface (just apply finer and finer grits), but it's a heck of a lot harder to get planar surfaces at the same time.  It might be practical to start with some gage blocks (as awful as it feels to suggest defacing their precision ground surfaces ), since those are lapped and polished to very good planarity.  You would be able to ensure micron-thick filler layers in that case (you'd also run the risk of fusing the blocks together -- essentially liquid phase sintering -- if the diffusion layer is allowed to grow too much!), and can modify the filler or gap or surface or clamping to test thicker layers as well.

Clamping pressure in the soldering fixture would also be relevant.

I do recall it's supposed to be that epoxy bonds best with a couple thou gap, and glass microbeads can be added to ensure this gap.  I don't know if that might vary with material being bonded, or what.  I wonder if solders do the same thing.

Definitely lots of room to test fairly simple things.  No idea if these have been evaluated before.  I'm sure there's a lot of literature out there on soldering; you'd have to peruse the ASTM or other libraries to see.  (May have free access at a school library?)

Tim

[coppercone2]

on a side note, another thing that got me confused is carburization. I think that the carbon actually migrates into the steel some how, I am pretty sure when I see a picture of something carburized cut open and polished, you can see its almost like a ink stain.

That's why a force diagram would help clarify.

You can get a cheap sample flat by sanding it on glass though, thats something I can try. Otherwise lapping plates and granite blocks, which I don't have or have a particular use for at the moment.

I do have gauge blocks i don't really give a fuck about, if other experiments prove promising you can get a data point from that maybe.

A problem with the test setup is to make some kind of micrometer adjustment clamp so you can separate the surfaces consistently. Maybe just make the test samples long and use feeler gauges that you pull out to make a gap but its kinda finicky. I would like a polished thread for adjustment.

What allows carbon to pass into steel like that? Can you do it with other light things like boron, lithium, etc?

[IanB]

What allows carbon to pass into steel like that? Can you do it with other light things like boron, lithium, etc?

Steel is by definition an alloy between carbon and iron (and maybe other things too). Carbon dissolves in iron. When the proportion of carbon is small you have steel. If there is a lot more carbon in the alloy you have cast iron.

[coppercone2]

maybe not lithium but I found this

https://bluewaterthermal.com/boriding-boronizing-diffusion/

[T3sl4co1l]

Carbon diffuses through iron, yes.  Perhaps more perplexing is getting it in there in the first place, apparently mediated by CO and nitrogen (especially in the form of cyanide?).  Contact between powdered carbon (in whatever form) is very poor indeed, in the solid state.

Tim

[coppercone2]

I thought it was vapor pressure diffusion, like a gas sponge. When you carbonize a part you usually put it in a box completely filled with charcoal and heat it glowing red. Like it sublimes.

I actually wonder what would happen if you put it in a pressure cylinder and cranked up the pressure some how during carburization. Maybe it does not work because if you use a gas filler it would lessen the carbon vapor concentration? And its glowing red hot so you can't mechanically compress it.

[IanB]

I thought it was vapor pressure diffusion, like a gas sponge. When you carbonize a part you usually put it in a box completely filled with charcoal and heat it glowing red.

This is referring to a process called case hardening of iron or steel where you pack a carbon rich substance around the part and bake it in a furnace. This causes carbon to diffuse into the surface and produce a thin layer of high carbon steel around the outside.

It is not required to have "pressure" to make this happen. Diffusion of different atoms or molecules can happen when there is a concentration gradient. The concentration gradient itself is the driving force.

[coppercone2]

But I thought diffusion constants change under pressure?

Is it just a sublimed carbon gas going inside? Or do you have something else going on (like teslacoil asked). Like it turning into some kinda liquid stuff?

https://en.wikipedia.org/wiki/Mass_diffusivity

this lists some pressure dependence.

[T3sl4co1l]

It is not required to have "pressure" to make this happen. Diffusion of different atoms or molecules can happen when there is a concentration gradient. The concentration gradient itself is the driving force.

Well yeah but that's the thing, the vapor pressure of carbon is SFA until well into the white-hot range.  How could it be gaseous diffusion if the vapor concentration is ~nil?  Surely it should be evaporating out of the iron, if anything!

The trick is to have something carrying a higher equivalent pressure or concentration of carbon: a catalyst.  Apparently several salts do this, and CO is implicated as well.  The reaction between C, CO, CO2, Fe (at the surface), CN-, CO3(2-) and who knows what else, is likely very complicated, as surface chemistry inevitably is.  (Maybe not so complicated that it hasn't been studied -- this is an important subject, and I bet some very illuminating research exists!)

Actually, come to think of it, carbonate may be a curious case, and not at all relevant.  A lot of anions and cations do have ranges of oxidation states, but carbonate isn't one of them, at least not at these temperatures.

At room temperature, the oxidation states of carbon (using strictly C, H and O) give methane, methanol, formaldehyde, formic acid and CO2.  The acidity (pKa) of these is something like 50, 16, 13, 4 and 6/10 (pKa above about 14 means the anion cannot form in aqueous solution; only CO2 has two ionic charges: bicarbonate and carbonate, and both are given).

At high temperature, most of these decompose, giving mixes of CH4 (or other hydrocarbons, depending on decomposition method and pressure), H2O, OH-, CO, CO2 and CO3(2-).  I'm not aware of there being any salt of reduced carbon between carbonate and carbide.

Point being, carbonate shouldn't be a catalyst, as far as I can think!  Unless it's reduced all the way to a carbide, or weird transient species are present.

Which, on that note, you can reduce, say, potassium carbonate with carbon -- the result is often explosive however, having such strange products as potassium acetylenediolate, K2C2O2 (a linear molecule), or potassium benzenehexolate K6C6O6 (a carbon ring studded with -O-K in all positions, but which is most definitely not all okay!).

Which reminds me of a joke...
A chemistry student asked out another student; it didn't go well.
Friend: "What did she say?"
Student: "Hexanitrosylbenzene. What the hell does that mean?"
Friend: "It's -NO in all positions."

Tim

[eKretz]

Oy. I don't want to quote the book length reply above,  so... Yes,  I was using the colloquial "tungsten carbide" - it is commonly referred to that way.  Cemented carbide is a more accurate description. The base metal in this case that would alloy with the braze is most commonly cobalt, although I have also heard of others being used and have also read that there are other metals added to the cobalt/TC mix in trace amounts depending on the manufacturer and grade.