Jump to content

Retrobrite

From RetroTechCollection
Revision as of 02:27, 23 September 2026 by Josh (talk | contribs) (Fix the 6 per cent row of the hazard table: at 6 per cent the classification is Eye Irrit. 2 (H319), not Eye Dam. 1, which applies from 8 per cent)
(diff) โ† Older revision | Latest revision (diff) | Newer revision โ†’ (diff)

Retrobrite (also written Retr0bright or retrobright) is a hydrogen-peroxide process used to reduce the yellow discolouration of aged ABS plastic on vintage computers and consoles. Peroxide is applied to the plastic, usually with ultraviolet light or heat, and the surface lightens.

Two things about it are worth knowing before you start.

It is a chemical process using an oxidiser that causes serious eye damage. At the concentrations everyone actually uses โ€” salon developer cream at 9 % to 12 % โ€” hydrogen peroxide is classified under the GB and EU CLP harmonised entry as Eye Dam. 1, H318, "causes serious eye damage", a classification that applies from 8 % upwards.[1] That is not "an irritant". Goggles, not glasses.

The long-term result may be worse than doing nothing. In December 2025 Shelby Jueden (Tech Tangents) published the outcome of an unintentionally well-controlled ten-year test: a Dreamcast case retrobrited in 2015 with peroxide and sunlight, with part of the underside masked off with tape so that it received no treatment. Ten years later the treated areas were more yellow than the masked, untreated areas, and also showed streaking and blotching the untreated plastic did not have.[2] This is a single observation on a single console rather than a controlled study, and it should be read that way; but it is the only ten-year masked-control comparison anyone has published, and it points the opposite way to the usual advice.

Retrobrite comparison

What actually causes the yellowing

[edit | edit source]

The account repeated almost everywhere is that yellowing is the photo-oxidation of a brominated flame retardant in the plastic, and that peroxide converts the coloured bromine species back to colourless ones.

The published polymer-degradation literature describes something broader. Yellowing of ABS is studied as photo-oxidation of the polymer itself, and specifically of the polybutadiene rubber phase: irradiation produces chain scission and cross-linking in the rubber, loss of the 1,4-polybutadiene unsaturation, and the formation of carbonyl species โ€” which is where the colour comes from.[3][4] Brominated additives, where they are present, are understood to accelerate UV degradation of ABS rather than to be the sole cause of it.

No source has been found for the single-cause bromine explanation, so this page does not present it as fact. The practical consequence is not academic: if the colour comes partly from oxidised polymer rather than only from a mobile additive, then bleaching the surface does not undo the degradation underneath it, which is consistent with what the ten-year test found.[2]

Peroxide concentrations and what they mean

[edit | edit source]
Concentration Where you meet it CLP classification at that strength[1]
3 % Pharmacy hydrogen peroxide solution Eye Irrit. 2 (H319) from 5 %; below that, not classified for eye effects
6 % (20 volume) Salon developer Eye Irrit. 2 (H319) โ€” the 5 % to 8 % band. The step up to serious eye damage happens at 8 %
9 % (30 volume) Salon developer Eye Dam. 1 (H318) โ€” causes serious eye damage
12 % (40 volume) Salon developer cream; the usual Retrobrite strength Eye Dam. 1 (H318) โ€” causes serious eye damage. 12 % is also the legal maximum for hydrogen peroxide present or released in a hair product under the cosmetics regulation[5]
30โ€“35 % Laboratory chemical; sometimes sold for "vapour chamber" methods Skin Irrit. 2 (H315) from 35 %, Eye Dam. 1 (H318), STOT SE 3 (H335) from 35 %. Acutely toxic if swallowed or inhaled (H302, H332)
50 % and above Industrial Oxidising liquid (H272), Skin Corr. 1B (H314) โ€” severe skin burns. From 70 %: Ox. Liq. 1 (H271) and Skin Corr. 1A

The "volume" scale on salon products is the oxygen volume, not a percentage. 40 volume is 12 % w/v, 30 volume is 9 %, 20 volume is 6 %. A bottle labelled only "40 vol" is the strongest hydrogen peroxide you can buy without a laboratory account, and the cosmetics regulation is the reason it stops there.[5]

Safety

[edit | edit source]
  • Goggles, not safety glasses. At 8 % and above this is an H318 "causes serious eye damage" chemical.[1] A splash from a stirred bath or a squeezed gel bottle goes in from the side.
  • Nitrile gloves and long sleeves. Peroxide bleaches skin white on contact; the white patch is a chemical burn of the outer layer and it stings for hours.
  • Rinse immediately and at length. Eye contact: flood with water for at least fifteen minutes and get medical advice. Skin: wash off at once.
  • Ventilate. From 35 % upwards the classification includes STOT SE 3 (H335), "may cause respiratory irritation".[1]
  • Never mix with bleach, ammonia, acetone, or metals and metal salts. Peroxide decomposes violently in contact with many metals and with catalytic contaminants; the reaction generates oxygen and heat.
  • Do not decant into a sealed container. Peroxide decomposes slowly and builds pressure. Commercial bottles have vented caps for this reason.
  • UV lamps. If you use a UV source rather than sunlight, it is a source of UV exposure for your eyes and skin as well as for the plastic. Enclose it.
  • Dispose of it diluted. Dilute heavily with water and flush; do not pour concentrated peroxide down a drain with bleach residue in it.

Methods

[edit | edit source]

These are the approaches in common use. Durations are deliberately not tabulated on this page: no controlled study of treatment time against UV intensity, temperature and initial discolouration has been published, and the hour figures that circulate are guesses. Work by inspection โ€” check every 20โ€“30 minutes, and stop when the colour matches, not when a timer says so.

Peroxide cream and UV

[edit | edit source]

The common method. Salon developer cream at 12 % (40 volume) is brushed evenly over the part and left in sunlight or under a UV lamp. The cream is viscous, so it stays where it is put and wastes less than a liquid.[5] Wrapping the part in cling film keeps the cream from drying; if it dries, the reaction stops. Uneven thickness is the main cause of streaking.

Immersion bath

[edit | edit source]

The part is fully submerged in liquid peroxide and exposed to UV. This gives the most even result and is the only method that reaches complex mouldings properly, but it needs several litres of peroxide, and a part that traps air will float and band.

Sodium percarbonate

[edit | edit source]

Oxygen laundry boosters are sodium carbonate and hydrogen peroxide in solid form; dissolved in warm water they give a dilute peroxide solution. Slower and weaker than developer cream, cheap, and easier to handle. The original Retr0bright formula added TAED (tetraacetylethylenediamine), the activator sold in the same laundry products, to the peroxide.

Sunlight alone

[edit | edit source]

UV with no peroxide will lighten yellowed ABS over days to weeks. It is the slowest method and the only one that adds no oxidiser to a polymer that is already oxidised. Given the ten-year finding, it deserves more consideration than it usually gets.[2]

Heat instead of UV

[edit | edit source]

Warming the peroxide accelerates the reaction without UV. This trades UV exposure for a warping risk, and ABS softens well below the temperature people assume. Keep well away from the glass transition of the moulding and do not use an oven.

Vapour and ozone

[edit | edit source]

Peroxide vapour in a sealed chamber, or ozone plus UV, are occasionally used. The vapour method depends on concentrated (30โ€“35 %) peroxide, which is a materially more dangerous chemical than developer cream โ€” see the table above โ€” and is not recommended for hobby use.

What goes wrong

[edit | edit source]
  • Streaking and blotching. Uneven cream thickness, or the cream drying in patches. The ten-year test found streaking and blotching still visible a decade later, so this is not necessarily a defect you can re-treat away.[2]
  • Chalky or matte surface. The oxidiser attacks the surface as well as the colour. Textured mouldings lose their texture.
  • Brittleness. ABS that is already degraded is being further oxidised. Screw bosses and thin clip fingers are what break afterwards.
  • Re-yellowing. Expect it. The treated Dreamcast surfaces re-yellowed past the untreated ones within ten years.[2]
  • Damage to anything that is not bare ABS. Painted legends, printed labels, silk-screened keycaps, rubber feet, light pipes and adhesive-backed badges are all attacked. Remove them, or do not treat the part.

What to do instead, or first

[edit | edit source]

Most of what people call yellowing is not yellowing:

  1. Wash it. Warm water and washing-up liquid, a soft brush, and patience. Forty years of nicotine, grease and handling film removes a surprising amount of "yellow".
  2. Try a magic eraser (melamine foam) on a test area. It is a fine abrasive and it will dull a textured surface, so test somewhere hidden.
  3. Compare against a shielded surface. The underside of the case, or the area under a badge, shows you the original colour. If the exposed surface already matches it, the plastic is not yellowed and you have a cleaning job.
  4. Consider leaving it. A machine that is evenly aged looks like what it is. Retrobrite is not reversible, the ten-year evidence is not encouraging, and an uneven bleach job is permanent.[2]

Material notes

[edit | edit source]
  • ABS โ€” what almost all of this applies to. Disassemble fully; treat whole parts, never half a part.
  • Polystyrene and polycarbonate โ€” behave differently and can craze. Test on a hidden area.
  • Coloured plastics โ€” pigments bleach too. Grey, beige and platinum machines are the usual subjects for good reason.
  • Painted, printed or stickered surfaces โ€” do not treat.
  • Rare or high-value items โ€” an irreversible chemical treatment on an irreplaceable object is a collections decision, not a cleaning decision.
[edit | edit source]

References

[edit | edit source]
  1. โ†‘ 1.0 1.1 1.2 1.3 Hydrogen peroxide, harmonised classification under Regulation (EC) No 1272/2008 (CLP) / GB CLP, index number 008-003-00-9, EC 231-765-0, CAS 7722-84-1, including the specific concentration limits. As reproduced in Carl Roth GmbH + Co. KG safety data sheet for article 9790, Hydrogen peroxide 30 %, VLSI Grade, GB version 7.1, revision 2024-10-10, sections 2 and 3. Specific concentration limits quoted: Ox. Liq. 1 H271 at C โ‰ฅ 70 %; Ox. Liq. 2 H272 at 50 % โ‰ค C < 70 %; Skin Corr. 1A H314 at C โ‰ฅ 70 %; Skin Corr. 1B H314 at 50 % โ‰ค C < 70 %; Skin Irrit. 2 H315 at 35 % โ‰ค C < 50 %; Eye Dam. 1 H318 at C โ‰ฅ 8 %; Eye Irrit. 2 H319 at 5 % โ‰ค C < 8 %; STOT SE 3 H335 at C โ‰ฅ 35 %. Also classified Acute Tox. 4 H302 and H332.
  2. โ†‘ 2.0 2.1 2.2 2.3 2.4 2.5 Jowi Morales, "A 10-year study indicates that treated surfaces turn yellow more severely than untreated plastic parts โ€” retrobrighting does more harm than good to consoles", Tom's Hardware, 6 December 2025, reporting Shelby Jueden's (Tech Tangents) ten-year masked-control test on a Sega Dreamcast case treated in 2015. Cited for the observation that treated surfaces ended up more yellow than the masked untreated surfaces on the same part, and showed streaking and blotching. This is one person's long-duration observation on one console, reported in the trade press โ€” not a controlled multi-sample study โ€” and is presented here as such.
  3. โ†‘ X. Jouan and J. L. Gardette, "Photo-oxidation of ABS: Part 2 โ€” Origin of the photodiscoloration on irradiation at long wavelengths", Polymer Degradation and Stability, volume 36, issue 1 (1992), pages 91โ€“96, doi:10.1016/0141-3910(92)90054-9. Cited for the existence of published work locating the origin of the photodiscolouration of ABS in the photo-oxidation of the polymer.
  4. โ†‘ R. M. Santos, G. L. Botelho, C. Cramez and A. V. Machado, "Outdoor and accelerated weathering of acrylonitrile-butadiene-styrene: A correlation study", Polymer Degradation and Stability, volume 98, issue 10 (October 2013), pages 2111โ€“2115, doi:10.1016/j.polymdegradstab.2013.07.016. Cited for the weathering behaviour of ABS โ€” degradation and yellowing governed by cross-linking and chain scission of the rubber phase, with loss of butadiene unsaturation and growth of carbonyl and hydroxyl groups.
  5. โ†‘ 5.0 5.1 5.2 Regulation (EC) No 1223/2009 on cosmetic products, Annex III, entry 12 (hydrogen peroxide and other compounds or mixtures that release hydrogen peroxide), retained in GB law. Hair products: maximum 12 % H<sub>2</sub>O<sub>2</sub> (40 volumes), present or released. Cited for the 12 % / 40 volume ceiling on consumer hair products, which is why salon developer does not go above that strength.