Note: It's been a good 6 months since I've posted an update, an inexcusably long time it seems. I'll be writing up and sharing some of the things that've crossed my brain-space in that time in the next few weeks, more big news to come.
Warning: Hydrogen peroxide in even modest concentrations is corrosive, be absolutely sure to wear gloves and goggles when handling it. It will rapidly bleach and deteriorate clothing and some metals. Never handle an industrial chemical you're not familiar with before reading though (and yes, I mean actually reading) the material safety data sheet, or MSDS.
I recently fell down a rabbit hole on YouTube, watching videos of people carefully restoring the vintage game consoles of my early years. Apparently an offshoot of the other tear I’d been on recently, digging into interesting glitches in older games; more on that in a later post. There’s something compelling about all restoration videos, from rusted barbell plates to yellowed-varnish on a painting, but that’s not what piqued my particular interest yesterday. In a lot of these game console restoration videos they mentioned retrobrite (stylized “retr0brite” with a zero in print), and at first I’d assumed it was just a commercial product, since the name honestly fit the bill. It was only when I came across a video with the process in action that I became confused.
Retr0brite isn’t a product, but rather a process. It boils down to a single step, submerging an aged plastic part (sans metal and rubber pieces) in a bath of roughly 10% hydrogen peroxide while flooding with ultraviolet light, typically sunlight or a UV LED strip. The catalyst I talk about later is entirely optional. In video after video I saw this process covert yellowed age-ridden plastic into what looked to be factory-fresh versions of the same parts. Usually I’d take this as a neat trick, but what I was seeing and the claims that accompanied it (namely that the process was "restoring" or "de-aging" the plastic) completely contradicted what I knew about plastics. To be fair, while I am trained as a materials scientist, polymers were never my specialty. One of the few practical facts I've retained about polymers is that they almost always become damaged and brittle under prolonged UV exposure. So the idea of reversing a degradation process by flooding them with UV, and soaking them in an oxidizing agent to boot, felt absolutely bonkers. I had to know more. Briefly, let me just drop here the exact protocol - it was originally suggested that it be mixed up in a paste, but it seems easier and more straightforward to deal liquids hydrogen peroxide, and the restoration community seems to have settled on that method as well.
The (Modern) Retr0brite Protocol
First thoroughly wash and scrub the yellowed plastic, be sure to remove any polishes or coatings that might block the UV illumination. Fill a container with hydrogen peroxide with a concentration of 10-15%, 12% is sold to consumers as a hair bleaching solution and should work well. For every 500ml of hydrogen peroxide add 0.5-1 teaspoon of “Oxy” laundry detergent (a consumer source for TEAD, our catalyst, more on that later). Wearing gloves, submerge the yellowed plastic parts completely. Expose the container to 1-2 hours of bright sun light or 12-48 hours of UV LED lightning, checking the color periodically. It’s a good idea to check the temperature of the bath periodically as temperatures above 60 C can lead to warping of the plastic parts. Remove them from the bath when they look like their old selves, and rinse thoroughly with water before re-assembling the device. The hydrogen peroxide will eventually be depleted as it breaks down into water, so replenish the bath as needed if you're re-using it.
Okay, so what is going on here?
Most of what I found while searching was simply people providing the recipe, attempting to sell some kit or product based on the process, or running a console restoration service. After a good long while I finally came across the original source: A series of write-ups by Dave Stevenson from Manchester, UK (aka Merlin on the AmiBay, English Amiga Board and Vintage Computer Forums), now only viewable through the Internet Archive. In these posts they shed some light on how the process was discovered by accident in a German museum for Commodore computers sometime in 2008, which seems fitting. Much of their explanation was repeated on Vintage Computers and Gaming shortly thereafter, as well as many many forum threads. In the years since then this explanation has made it to basically every corner of the internet that is interested in this sort of work, but it struck me after awhile that the mechanism they describe cannot be quite how the retr0brite protocol de-yellows plastic. Let's start at the beginning.
ABS Plastic
The plastic used to make computer cases and many of these retro electronics enclosures is called acrylonitrile butadiene styrene or ABS. ABS is actually a terpolymer and is made up of the following chemicals: acrylonitrile, 1,3-butadiene, and styrene. The acrylonitrile provides thermal and chemical strength, the butadiene provides flexibility and impact resistance, and the styrene provides the glossy surface as well as making the polymer inexpensive and moderately machinable. This mixture, when heated in the presence of catalysts produce the ABS polymer.
The monomers form a cross-link lattice, and it is this lattice that gives the plastic the required strength, as styrene is a brittle material; the styrene partly fills in the voids within this lattice work as well as cross-linking to the acrylonitrile. The butadiene has a similar effect, as well as dispersing in the molten polymers. ABS was one of the first commercial plastics and was developed in 1948, and remains one of the cheapest injection-moldable plastics.
Flame Retardant Theory
Untreated ABS plastic is combustable, which is a bad property for consumer electronics to have. Chemists found that the combustibility of ABS could be reduced by adding poly brominated diphenyl ethers (PBDEs), compounds containing a good amount of the Group VII halogen Bromine. Halogens had a long history of use in fire suppression before their negative environmental impacts were known; the bromine in ABS is contained within the polymer and presents no such hazard.
The primary PBDE used in consumer electronics plastics back in the 1980s and 90s was tetrabromobisphenol-A (or TBBP-A for short). The theory put forward back in 2008 held that these bromine atoms would eventually be implicated in the yellowing process.
The thinking was that incoming UV light would break these bromine free from the TBBP-A, and these free radicals would then stabilize by forming semi-stable coordinate covalent bonds with oxygen, causing yellowing in the surface when this occurred near the surface. The use of UV light in retr0brite was claimed to destabilize those coordinate bonds between bromine and oxygen, allowing a hydrogen from the hydrogen peroxide to instead bond with the bromine and reduce it to a colorless substance. The tetra-acetyl-ethylene-diamine (TAED) the protocol calls for serves as a catalyst in this process, reducing the energy needed.
Misgivings and Clarity
This explanation almost fit with the observations, but something still didn't sit quite right. My first thoughts were; I remember handling bromine, it evaporates quickly at normal temperatures and pressures, it is only darkly colored in it's covalent form (Br2) while bromates are usually colorless, also it is really quite pungent. If ABS plastic gave elemental bromine off in sunlight we'd almost certainly smell it. After a bit more searching I came across an excellent post by a chemistry student on Reddit by the name Grogglebob, that put a finger on the two major inconsistencies in this proposed reasoning. First, that one needs a cation counterpart to coincide with the bromide anion being created, and that there are dozens of industrial studies of ABS plastics yellowing even in the total absence of TBBP-A or other bromine-containing additives.
It seems more likely on inspection that the yellowing is due to photooxidation of the polymer itself, not the additives. In that post they cite a paper, Clarification and Discussion of Chemical Transformations Involved in Thermal and Photo-oxidative Degradation of ABS by J. B. Adeniyi (Eur. Polym. J. 1984, 20, 291-299), which lays out a reaction scheme showing how unadulterated ABS polymer will photo-oxidize to eventually produce a stable compound called α,β-unsaturated carbonyl. They suggest that since conjugation tends to reduce the energy gap of compounds, it's not unrealistic that this compound might have a yellow color.
The conjecture now is; the butadiene chain is full of double-bonds, these are the sites where this carbonyl forms, and they themselves contain a double bond. If the hydrogen peroxide is a strong enough oxidizer (or is so with the help of TEAD) to break these double bonds, then it's possible it's breaking the double bond on the carbonyl, effectively bleaching it.
It's just bleaching
The major difference between these two interpretations is that the first proposes that retr0brite is reversing the degradation and stabilizing the free radicals, while the second proposes that retr0brite is actually further degrading the material, but simply degrading the yellow-tinged chromophore first. Given that the solution itself, hydrogen peroxide plus a non-chlorine bleaching activator (the TEAD), and catalyzed by UV light, seems a lot more like a traditional bleaching solution than a hypothetical bromine soluablizing solution. If the goal really were to solubilize and remove bromine and borates, water would probably work, but it really seems like they're not the problem we're dealing with.
This interpretation is furthered by the demonstration another user on YouTube provided, using methanol, hydrochloric acid as a catalyst, and bubbled chlorine gas as their bleaching agent, and achieved similar results (though it did more aggressively attack the plastic) with this alternative bleaching solution. Our hydrogen peroxide, TEAD, and UV treatment is just a normal bleaching process, albeit one that is more gentle than chlorine based bleaches, and one that does seem to degrade the chromophores first.
Take aways
All that said, is there any reason not to use this process to de-yellow aging electronics? I honestly don't think so. Anecdotally at least, the change in mechanical properties appears to be quite small, and experiments attempting to demonstrate that re-yellowing occurs faster after treatment haven't been terribly consistent. It is possible that the mixture, in addition to breaking down the chromophores, also breaks down light-stabilizers intentionally added to the plastic, which could lead to a faster second yellowing, but users have reported good results with a thin coat of UV resistant varnish. It's true though that it's certainly not "de-aging" or "stabilizing" the plastic, and anyone claiming otherwise is talking nonsense. It won't be any less brittle after the treatment than before, but will be less yellow, which is probably more than enough.
Only after writing this all out did I finally stumble upon this excellent write-up by Josh Velson over at Quora, a chemical industry consultant, and this piece over at Medium by David Flood . Thankfully it seems I came to the same conclusion as people who are probably more suited to the question - namely that the originally proposed mechanism for the action of retr0brite is nonsense. Less fortunately, it seems that the bromine hypothesis has basically been taken as gospel by many users of the protocol in the electronics restoration hobby. That all said, I suppose it doesn't really matter - gently bleaching old plastic still makes it less yellow, regardless of how you think it gets it done.