We made industrial AGMs and when some failed the initial testing we would discharge them into negative voltage of about -1.0v per cell. Then a slow two day recharge. 50/50 chance of making them pass but some did.
Not too side track this, but do you know if this technique can work for gel cell type batteries(which seem a close cousin of AGM)?
I have a rather niche application(non-automotive, actually photography related) that can use a 6V gel cell among other options, and this was by far and away the most common(the others were a flooded lead acid, which I don't think has been made since the 60s and also isn't great for something designed to be portable, and a Ni-Cd pack which I've never actually seen). I bought a new battery-branded as a "Dry fit" battery in their catalogs-in 2018 and that one still works great but I also quite literally have over a dozen that are totally dead(completely open circuit) and some others that sort of work but will only charge to ~6.5V or so and won't properly operate the equipment they fit.
I've been trying a lot to bring some of these back to life(and mostly have one there-its performance is less than great but it does power on and work the equipment), primarily because, as best as I can tell, it quite literally is not made anymore. The OEM supplied battery was made by Sonnenchein(I have one branded that and Exide) with a small modification to the casing to "key" it. Specifically, if anyone is interested, it's an A506/4.2s but I THINK A506/4.2K could probably be made to work. The issue is that the battery needs to essentially match the original in all 3 dimensions, and something too large in any of those dimensions won't fit at all. There are batteries out there that claim to be an A506/4.2k or to cross to it, but they are too big-the original is nearly square and about 2"x2", and everything I can find with that part number is closer to 2"x3"
So, that's my reason for trying to revive dead ones. I've tried things you're not supposed to do like pop the caps off the cells and try rehydrating them(moderate success) as well as charging at serious overvoltage(like 2.5V/cell or even higher in a few cases-I won't say how high but I found some guidelines from a reputable source and monitored gas evolution closely). I've even dropped battery casings in the ultrasonic since that supposedly breaks up sulfation. I have one that, through a combination of these methods, went from totally dead to being able to charge to ~6.4V, so apparently it's not totally ineffective, just not enough. I've also given up one and sawed the case open, carved out the cell dividers with a die grinder, and dropped 5xsub-C NiCds in it...
Really, though, where I was going with this is I have one that I acquired that would charge but not enough. Through a combination of all of the above, I FINALLY today got it to actually work, meaning it will power the equipment fully but just not quite where a new one will.
I do most of my messing around/rehab of these with a bench PSU that of course gives me a ton of control. At one point, the one that I sucessfully revived to fully working I'd actually inadvertenly reverse charged. IIRC, when I stopped, it was around -2.1V, which would have been about -.7V/cell. I'm wondering if that was part of its sucessful revival.
I have one sitting beside me now that I managed to get charging today. As I sit here it's drawing 12mA at 7.2V, and after about 4 hours sitting there has charged to 3.4V. I'm tempted to try reverse charging it now...