How to tell image retention from burn-in, how burn-in actually happens, what manufacturers do and what you can do. Starting from "every pixel is its own lamp", four technical sections unpack OLED burn-in, plus a "ZJWY view" on what this class of problem means on the repair bench and how to classify it at intake.
Abstract: Anyone who used an OLED phone in the early days has probably had the feeling that the screen is broken — after a year or two, sometimes only months, a mark appears that will not wash off. There is also a second phenomenon that is easily confused with it: a faint after-image left behind after a fixed picture has been shown for a short while, which is entirely normal. This article separates the two: how to tell image retention from burn-in, how burn-in actually happens, what manufacturers do about it, and what you as a user can do.
A quick test to start with: open a plain white or plain grey image, adjust the brightness and look carefully for a shadow in a fixed position. If there is one, that is burn-in.
The difference between the two is, at bottom, whether it can recover.
| Image persistence | Burn-in | |
|---|---|---|
| Nature | Temporary | Permanent damage |
| Cause | After TFT stress, the previous frame's signal has not been fully cleared | Some pixels have aged noticeably more than their neighbours |
| Symptom | A faint after-image briefly visible after switching pictures | A mark in a fixed position that will not wash off |
| Duration | Fades within minutes of normal use | Still there after days or months |
| Action | None needed | No software fix exists |
| Analogy | Fresh dust — a puff of air and it is gone | An ageing bulb — it cannot be restored |
One sentence is enough to remember: what fades by itself within minutes is retention; what sits in a fixed position for days is burn-in.
Whether a panel can burn in depends on where the "lamp" is fitted.
LCD is one sheet of backlight plus liquid-crystal valves. The backlight layer lights and dims as a whole; the liquid-crystal layer only decides how much light each position passes. Because the lamp is one sheet, it also ages evenly — there is no question of one area ageing first.
OLED has no backlight. Every pixel emits its own light, and the material is an organic light-emitting compound. That gives two consequences: when showing black, a pixel can be switched off entirely (blacker blacks, lower power) and the panel can be made thinner — but organic materials have a physical lifetime: every time it lights up, it ages a little.
Within one panel, whichever pixels have been lit longest and brightest will dim first. Once that difference in ageing becomes fixed, what you see is burn-in.
One batch of pixels, lit continuously by the same picture, dims ahead of its surroundings. You see these situations every day:
They share three traits: fixed position, reasonably high brightness, long display time. The pixels underneath keep emitting, and consume their lifetime faster than the rest. Open a full-white image and the over-used area cannot keep up in brightness, so a faint dark shadow surfaces.
💡 Turned around, those same three variables are the prevention strategy: hold down any one of them and you noticeably delay burn-in.
Besides user habits, manufacturers have taken four measures in hardware and software.
Special algorithms watch how each pixel is being used, generate calibration data and automatically adjust pixel brightness — levelling out part of the ageing difference and making burn-in less visible.
The system shifts long-displayed UI elements by a few pixels every so often, spreading the wear across more pixels instead of letting one fixed batch carry it all. The shift is small enough to be essentially imperceptible.
Models that support always-on keep only a very small number of pixels refreshing at low frequency; most pixels are physically switched off, thinning out the wear from a static picture.
Material stability improvements raise the constant-lighting lifetime of the organic emissive material. Today's newest materials should last more than twice as long as those of three or four years ago. There is also the tandem (stacked) OLED approach, which uses multiple emissive layers to reduce the load on any single layer.
This is also why genuinely burnt-in new phones are noticeably rarer in recent years — not because the principle changed, but because the material and algorithm headroom grew.
Two cases.
Playing a video with rapidly alternating pixels helps to a degree, but the effect depends on how far the ageing has gone; some cases need the screen to be lit for a long time before they improve.
At present no software method can make aged organic material young again. Third-party repair shops normally have no means of restoring it either — running Demura again should work in principle, but nobody is known to be doing it.
One more idea: display the very pattern that burnt your screen inverted, for a long time, so that the bright parts are brought down. It makes sense in theory, but it is hard to control in practice and not without risk — not something for an ordinary user to try.
Working from brightness, duration and stillness, here are five things you can do straight away:
For those of us who build screen repair equipment, burn-in is a category that has to be explained up front, not after the job is done. Unlike a broken flex or a corroded ITO trace, it has no clear repair path — it happens in the emissive layer itself and it is material ageing.
1. First step at intake: classify, then quote. What a customer calls "there is an after-image" or "the mark will not come off" can be three entirely different things, with very different work and very different prices:
| Symptom | Classification | Action |
|---|---|---|
| Appears briefly after switching pictures, gone within minutes | Image retention (residual TFT signal) | No repair. Let the customer use it normally for a few days |
| Fixed position, visible on white and grey, still there after days | Burn-in (emissive layer ageing) | Replace the panel. The emissive layer cannot be repaired |
| Uneven brightness or colour, dim edges, blotchy patches | Mura, or a backlight / encapsulation issue | Go through Demura or an encapsulation-side assessment |
2. Standardise the criteria — do not rely on a visual impression. A fixed check works well: plain white field + 50% grey + a low-light environment, looking at 100% brightness once and 30% once. Burn-in shows up most clearly at low brightness on grey, and is almost invisible on a colourful interface — which is why customers often never noticed it until someone pointed it out. Photograph it for the record: evidence for the customer and protection against disputes.
3. State plainly: burn-in cannot be repaired, and any promise otherwise is over-promising. This is the most important point. When a customer asks "can it be fixed", the answer is that ageing of the emissive layer is irreversible, and Demura only corrects brightness/chroma non-uniformity (Mura) — it cannot restore organic material that has already decayed. Put that into the intake script; it is far more effective than explaining it after the repair.
4. Incoming inspection of pulled and second-hand panels must include this check. Burn-in is a usage mark, not shipping damage, so it is common on pulled and used panels. Before fitting one, run white field + grey scale + fixed-UI positions (status bar, navigation bar, dock) and confirm there is no shadow. Otherwise the customer finds "the panel you fitted has marks too", and the cost of explaining that is very high.
5. Be more restrained with heat when opening an OLED. OLED organic materials are temperature-sensitive; both the heating temperature and the dwell time during separation and adhesive removal have to stay inside the material window. Too high a temperature, or holding heat on one spot too long, accelerates emissive-layer decay and can even cause fresh irreversible damage. Even heating + a capped upper temperature + no long dwell on one spot — these three are the biggest operational difference between opening an OLED and opening an LCD.
6. Do not treat image retention as a defect and replace a good panel. When a customer comes back saying the panel they just had fitted "has marks", ask them to use it normally for ten or twenty minutes and look again. Genuine retention fades, and once it has faded there is no need to open anything. This cuts unnecessary rework and stops good panels from being written off.
7. Write the warranty boundary down. Burn-in is wear from use and, like a cracked panel or liquid ingress, is normally outside warranty. State on the job sheet or warranty card that the replacement panel passed white-field and grey-scale inspection with no burn-in before fitting, and that any fixed-position shadow appearing afterwards arises from subsequent use.
In short: burn-in is the one screen problem that cannot be repaired. Identifying it and explaining it at intake saves more cost than any repair technique — for the customer and for the shop alike.
← Back to Support