ATTACK!!! (epoxy remover)

>>>ATTACK!!! (epoxy remover)

Glass sticks to almost anything when it’s hot and practically nothing when it’s cold, which is why I love it as a work surface. That’s also why it’s such a pain when you want to attach glass TO something. Once I tried attaching a hanger to glass with GE Silicone II.

Let’s just say that I’m very, very glad that pate de verre bounces.

My favorite pate de verre hangers are small acrylic French cleats (right) –they don’t make a shadow on the translucent glass, and they push the work out far enough from the wall that light can get in through the back. I bought the originals from my friend Barbara Muth; since then I’ve had my own cut at the local TAP Plastics.*

I first tried gluing them on with silicone; I will NEVER use it again (at least on glass hangers). VHB tape didn’t work much better (although the massive roll I got free from my metals guy was probably outdated). Most of the regular glass glues I’ve used aren’t fond of acrylic.

The adhesives website ThistoThat recommended Loctite Impruv 349. Took five phone calls and a very long drive to find it (you now can buy it on Amazon), but it does indeed provide a permanent bond between acrylic and glass.

Unfortunately, they’re not kidding when they say permanent. Just try UNgluing the stuff if you make a mistake. Which I did.

Backstory: Loctite recommended a $1,500 UV light to cure the Impruv. EEEEEK. My glassist friends scoffed and advised buying a UV bulb from from Home Depot for maybe $6. No contest there; I headed for Home Depot, found a nice swirly blacklight bulb and screwed it into a desk lamp. Then I turned it, full glare, onto cleat+glass.

18 hours later I tested the bond. That cleat slipped off the glass as easily as if I’d stuck it on with vaseline. Frustrated, I set the glass, slippery cleat and all, outside and went back in to drown my sorrows in lemonade.

5 minutes in the drizzly Oregon no-sun cured that adhesive beautifully, making a strock, rock-hard join. It was just too bad that the cleat had slipped down and was crooked, facing the wrong direction.

As promised, the Loctite 349 bond was pretty daggone permanent, and not even Loctite held out much hope for getting it off mechanically.

Someone suggested placing the sculpture in the oven, heating it to 500F and whacking it a couple of times to see if the cleat loosened. The glass, he said, probably wouldn’t break.


While browsing the HIS Glassworks website, I found Attack, an epoxy solvent, that promised to dissolve epoxies and cured polyester resins, ordered an 8-ounce can.

The instructions (which by the way caution you up one side and down the other that this stuff is NOT human-friendly) said to fill a container with Attack and immerse the piece, naughty glue joint and all. You then sealed the container and let it sit overnight. In the morning, you’d peel the pieces apart.

Yeah, right.

This was a largish sculpture; I didn’t have enough Attack, or an Attack-safe container big enough, to immerse the whole sculpture. And on the outside chance that this stuff could mess up pate de verre, I wanted it to confine it to the back area around the cleat.

I made a well of thick clay around the cleat and dumped in the Attack. It didn’t seem to do much, and the clay was holding it, so I covered the clay with a scrap sheet of clear acrylic to hold in the Attack.

In retrospect, that was pretty dumb thing to do.

In the morning, the Attack had dissolved about a quarter-inch of the cleat all around, turning it into a gummy, smelly goo. And it had turned my acrylic lid into a jello-like substance.

Unfortunately, the cleat was attached as permanently as ever. That spoke well for the original bond, but wasn’t getting me very far.

I scraped out the goo and dumped more Attack into the reservoir, and slowly developed a routine. Every 8-12 hours I’d check the mess, scrape out the old oozy acrylic, pour in new, and seal it up. The Attack was attacking the OUTside of the cleat, never actually getting to the joint itself.

After four days, it had dissolved all but maybe a half-inch square of the original 2x3x.5 inch cleat. I applied gentle but determined elbow grease and finally scraped off that little bit. The residue solidified into something like rubber cement once the solvent evaporated, making it easy to roll off the glass.

I scrubbed the glass with soap and water, then denatured alcohol, then acetone (which hopefully got rid of any residual Attack). Then I reattached the cleat in the right place with fresh Impruv, cured it, and delivered it (barely on time) to the gallery.

(BTW, I used a Feit UV bulb to cure the new Loctite. It works like a charm in a plain old desklamp for $10. Thanks, Gary.)

So, does Attack work? Yup. Eventually. If you’re patient, have four or five days and are willing to do a LOT of scraping and repouring.

At a single bottle per joint it’s a pretty expensive glue remover, but it did get a fully cured UV-adhesived chunk off my glass. The alternative–casting a new glass sculpture–would have cost a lot more in dollars and time.

I’m not sure how Attack would work in a glass-to-glass bond. In my case, it clearly attacked the cleat, not the joint, and worked from the outside in. I saw little or no capillary action pulling the remover into the glue joint. I suspect it would take quite awhile (although much less Attack) to do the job.

Ah well. At least I learned a lesson: “Think first, glue later.”


*Word of caution: If you have your local plastics supply store make a supply of these acrylic cleats, EMPHASIZE that the 45-degree angle on one edge must be perfectly straight and parallel to the cleat. My first batch was cut any which way, and absolutely useless.



  1. Cynthia January 31, 2012 at 12:12 pm - Reply

    Tony, I’d be really REALLY careful about this, and test on pieces you don’t care about losing first. I don’t think Attack would hurt quartz or carborundum (though I most certainly could be mistaken).

    HOWEVER…soft, absorptive materials like turquoise, tiger-eye, malachite, etc., might be harmed by it. I’ve ruined the finish on turqoise and opal by not testing first, and once they’re messed up it’s a lot of work to fix, assuming you can.

  2. Tony January 29, 2012 at 11:01 pm - Reply

    Many years ago, I took a lapidary class and we cut and polished many nice stones — trouble was, we used cheap settings to which we epoxied the cabochons to …. we had cut tiger-eye, quartz, picture stones, turquoise, malachite and similar gemstones….

    Would love to remove the epoxy so I could then reset them on nice quality settings.

    The stones have flat backs and were glued to the flat metal backing of the settings…. would this “Attack” dissolve or soften the epoxy so I could eventually remove stone???

  3. Kris November 9, 2011 at 4:56 pm - Reply

    I’ve, unfortunately, had to use Attack on many laminated pieces and glued pieces of glass. Nikki is right on the laminated part. I’d rather cut it apart for some things. Also I have had pieces of glass break sitting in baths of Attack. But for smaller simple joints it is great. Provided it doesn’t eat off the gloves!

  4. Nikki October 17, 2011 at 3:59 pm - Reply

    Using Attack to de-laminate a thick Hxtal-laminated piece (thin joins) doesn’t work. After immersion in a couple quarts for a few days, the seams were not “attacked.” No capillarity here. Best way to de-laminate such thin joins is heat, like 700F for an hour for a 20 sq. in. piece. My friend and I went through that. Note that the methylene chloride-based solvent is volatile and evaporates quickly. So cover tightly.

    It’s all a process..

  5. GBC October 14, 2011 at 9:51 pm - Reply

    I don’t have a can of Attack around right now I ran out & haven’t need any lately.

    Check the label I think it is “methylene chloride” a highly Cancer causing chemical. Use beware.

    Look for other MC solvents / strippers that will work in a pinch or to save hazardous shipping cost.
    Also containing MC are most Automotive and heavy duty paint strippers.
    Home Depot sells KS-3 as their max power gel stripper they also have a sprayable thinner label by same brand.

  6. GBC October 14, 2011 at 9:41 pm - Reply

    Attack will work on all epoxy and UV glues but your correct it has no capillary action at all, it’s only designed to work at point of contact.

    Suggestion. cut slots in or grind down the part to remove to make more direct contact with the glue. I know you say that defeats the need for the solvent… no not really. The idea is to remove one part with out damaging the main product.

    Mainly used in jewelry repair and watch crystal replacement. When using on a watch all the glass must be cracked or chipped off then the glue is devolved with Attack leaving behind a perfect metal surface… gold, stainless, silver what ever… not 100% sure about aluminum.

    It always helps to know what the product is intended to be used for to know how it can help you and best be applied to you needs.

    Hope this sheds some prospective,

  7. cynthia March 19, 2010 at 7:20 am - Reply

    Well….perfectly willing to try it again, but except for the “immerse the whole piece bit” I was following the directions. What do you folks do differently?

    Amanda, I’ll have to give GE Silicone I a try. Thanks.

  8. Amanda Taylor March 19, 2010 at 3:50 am - Reply

    We use the Attack for our UV glues and love it. As for silicone use we stopped using GE Silicone II and use only the GE Silicone I (with no additives), on Marc Leva and Ellen Abbott’s suggestions. I was finding that the GE Silicone II never really setup properly. But since we switched to GE Silicone I… additives… it works very, very well and sets up solid. Just a thought.

  9. Jerome March 16, 2010 at 11:39 am - Reply

    Thats Funny… sorry to Hear you horrable exerience to the solution. I’ve been using Attack for many years now and it’s the greates stuff invented, if used right.
    I’m a Navajo Silversmith and it works great for my Inlay jewelry repair

  10. cynthia February 4, 2010 at 9:38 pm - Reply

    Thanks, guys. (And thanks, Sam, for the nice compliment!)

    I’m glad I found the Attack…I hope never to have to use it again!

  11. Sylvanye "Sam" Roh February 4, 2010 at 5:01 pm - Reply

    Cynthia, I am so happy to have discovered your blog, it is great.

  12. Sylvanye "Sam" Roh February 4, 2010 at 5:01 pm - Reply

    You are so right, Attack is the best product I’ve found for removing even UV cured glues from my glass and metal findings. It will take awhile so one needs to be patient but in time you will have your freed piece.

  13. chaniarts February 3, 2010 at 10:46 am - Reply

    acetone and gasoline will dissolve acrylic too, probably not as fast as attack though.

  14. ellen abbott February 3, 2010 at 7:36 am - Reply

    What a nightmare.

Comments welcome! (thanks)

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: