Do you have rust stains? Do you not want rust stains? Come on in!

Disclaimer: This does NOT remove rust, it just removes rust stains on vertical surfaces.

Dad is a restoration expert by trade, antique furniture.  Trained in the mystic restoration arts by those in the darkest corners of the Smithsonian institute.  His knowledge is vast and strange and heavily focused in chemistry.  He worked this up to help with some old rust stains on my Herter's boat.  He tested everything on his VW van.  I'll probably test it on multiple rusty projects we have
We have a lot of rust in the family.

I will now post it verbatim with the pictures he sent to help describe the process. The only edit I will make is separating the final paragraph into the three separate tests for clarity.

To remove rust on vertical surfaces.  Mixed approximately equal parts (by bulk) of methylcellulose powder and oxalic acid crystals, wet it with a supersaturated solution of more oxalic acid, let it stand overnight to gel.  I tried different methods of application, I got it slightly too liquid, it sags on the surface.

I know the lower part of the Van below the windows had been repainted; I figure the big irregular oval on the right side was a dent, the ring of rust around it is improperly-primed metal around bondo, the metal was bare when they sanded down the bondo in the middle.  Bondo is shrinkage/age-checked, but you can only see that up close and in raking light.

1. The first test spot with the long run below it was left on overnight; that was too long, the acid sucked the red out of the orange paint, and it feels blistered to the touch, I assume that's rust spots under the paint. 
2. The second spot was perfect (for getting the rust stain out of the paint, it doesn't fix the problem of rust underneath it); I just dabbed some on with my fingertip, let it dry, about two hours. 
3. I put Test 3 on thickly, with a palette knife.  That was unnecessary, and gave me a lot more gel to wash off; the acid does quite well in a thin layer, at least on removing rust stains from the paint.  Removing actual rust from the steel would probably take more, and longer.  Sandblasting is more direct, more thorough for rusted steel.  But it's nice to have another arrow in the quiver for rust stains.

The mixed gel

Test 1

Test 2

Test 3

Follow up email from Dad after I asked if he minded if I posted his email.

Pic of the ingredients.  Mixing it to make a gel can be finicky.  First thoroughly mix the dry ingredients; when you add water, the oxalic acid crystals will dissolve immediately and will help the water surround the methylcellulose particles, which is necessary for a smooth gel.  You know how when you're making a flour thickener you don't just dump water into the flour, that'll make lumps.  Same thing here; add just a small amount of water, stir to mix thoroughly into a thick paste.  Add a little more, until it starts to smooth out a little.  Resist the temptation to keep adding water, the methylcellulose particles haven't absorbed any yet, haven't gelled.  Set it aside overnight, see if it's too thick in the morning.  If yes, add tiny amounts of water until it's the consistency to not slump on a vertical surface.  If you get too much, you can't take the water out, it will only evaporate from the surface, leaving a skin.  And you can't heat it very much, the methylcellulose will turn solid (but I don't know at what temperature).  Also, I haven't had any luck making a methylcellulose gel with bases, only acids.

I have a rusty set of long tube headers I want to paint.

Do you think that I could hang them and coat them with this rather than trying to immerse one end at a time in a solution of citric acid?

How rusty are rusty?  Like I mentioned, this is more for rust stains than straight rust.

In reply to Mr_Asa :

Good point. I think that immersion would work better. 

Personally, I like EvapoRust for straight rust more than any acids.  Personal preference.