How PMOLED is Made | Manufacturing Process, Materials

A PMOLED is made by patterning a transparent anode on glass, building insulating pillars between the cathode columns, evaporating a stack of organic layers about 100–200 nm thick in vacuum, and depositing a metal cathode that the pillars break into separate columns. The panel is then sealed against moisture. There is no thin-film transistor array anywhere in the flow.

Process at a glance

The sequence runs in one direction, from a coated glass sheet to a sealed panel: clean and pattern the indium tin oxide (ITO), form the isolation pillars, evaporate the organic stack, deposit the cathode, encapsulate, then scribe and test.

Organic semiconductors are the emissive medium here, a route that displaced earlier inorganic compound emitters for small thin panels^[1].

Two features separate this from an active-matrix line. First, no TFT array is deposited, so there are no amorphous-silicon or oxide semiconductor steps and no pixel storage capacitors. Second, the cathode is patterned by shadowing rather than by photolithography, because a wet process would destroy the organic layers underneath.

The trade-off is written into the architecture. Rows are driven one at a time, so each pixel emits for a fraction of the frame and must be driven harder to reach the same average brightness. That is why PMOLED suits small, low-to-medium information displays rather than large panels.

Anode patterning

Most PMOLED lines start with soda-lime or alkali-free glass at 0.5 mm or 0.7 mm, pre-coated with ITO by magnetron sputtering. A typical anode film is 120–150 nm thick with sheet resistance in the 10–15 Ω/sq range, which keeps the voltage drop along a column tolerable at the currents a passive matrix needs.

Photolithography follows: spin-coat positive resist, soft bake near 90–100 °C, expose through a chrome mask at i-line, develop in TMAH, then wet-etch the exposed ITO in a hydrochloric/nitric acid mix held at a controlled temperature. The resist is stripped and the surface inspected.

Edge profile matters more than line width. A near-vertical ITO step can leave the later organic films discontinuous over the edge, so lines are usually etched to a tapered slope rather than a wall. Surface roughness is checked as well; a spike taller than the organic stack can short the device at that pixel.

All of the values above are line targets rather than universal constants. Sheet resistance, etch chemistry and slope angle are chosen per product, and any supplier should be able to state the window it actually controls.

Isolation pillars

The pillars are the part of the process that makes a passive matrix possible. A thick negative photoresist, several microns high, is exposed and developed so that its sidewalls slope inward, leaving an overhanging profile with the top wider than the base.

The mechanism is straightforward. UV is absorbed as it travels down through the resist, so the top cross-links more heavily than the bottom. The developer removes the less-cured base faster, and the undercut appears without any extra etch step.

A hard bake follows, in vacuum or nitrogen, to drive off residual solvent and finish the cross-linking. This matters because the pillars sit inside the evaporation chamber: anything that outgasses there becomes a contaminant in the organic layers.

Pillar height is set well above the combined thickness of the organic stack and the cathode, so that the overhang casts a genuine shadow during metal deposition rather than merely a partial one.

Organic evaporation

The organic layers are deposited by vacuum thermal evaporation. Small-molecule materials are heated in ceramic or tantalum crucibles until they sublimate, and the vapour travels ballistically to the substrate because the mean free path at the chamber pressure is far longer than the source-to-substrate distance.

A typical stack is a hole injection layer, a hole transport layer, an emissive layer, an electron transport layer and an electron injection layer, with total organic thickness usually between 100 nm and 200 nm. Some designs add blocking layers or a doped charge-generation structure.

Deposition rate is monitored by quartz crystal microbalances and held by closed-loop control, generally in the range of a fraction of an ångström to about 2 Å/s per source. Faster rates tend to give less dense films; slower rates lengthen the substrate's exposure to radiant heat.

Host and dopant are usually co-evaporated from separate sources so their ratio can be trimmed independently. Cluster tools move the glass between chambers without breaking vacuum, which is essential: water and oxygen quench organic electroluminescence.

Cathode and separation

The cathode is a low-work-function metal or alloy, commonly aluminium or a magnesium-silver blend, evaporated to roughly 100–150 nm at the end of the vacuum run.

Because the deposition is line-of-sight, the overhanging pillar tops shield the substrate immediately beside them. Metal lands on the pillar tops and on the pixel areas between pillars, but not on the sidewalls or the shadowed strips. The cathode is therefore severed into electrically independent columns in a single step.

The geometry has to be respected. If the source subtends too large an angle, or the substrate sits too close to it, arriving atoms can come in at a shallow enough angle to creep under the overhang and bridge adjacent columns. Source distance, source size and substrate rotation are all set to keep the arrival cone narrow.

Sealing and singulation

Unsealed organic layers degrade within hours in ordinary air, so the panel is transferred to a nitrogen glove box and sealed there. The usual construction is a glass cap bonded to the substrate with a cured epoxy edge seal, with a getter or desiccant inside the cavity to absorb what permeates over time.

Thin-film encapsulation deposited directly over the cathode is the alternative, and it is what makes flexible PMOLED prototypes possible: a multilayer of inorganic barrier and polymer planarisation replaces the rigid cap. Barrier performance, not thickness, is the specification that matters.

After sealing, the sheet is scribed and broken into individual panels, driver ICs are bonded, and the module is tested for luminance, current draw at a fixed pattern, and visual defects. A current reading above the expected value at the same drive voltage usually points to a shorted column.

Where yield is lost

Most PMOLED losses trace to a short list. Particles left after cleaning or introduced at mask handling create pinholes or dark pixels. Mask alignment error shifts the anode columns relative to the pillars and distorts the aperture. Etch non-uniformity changes line width across the sheet.

Shadowing failures produce the most expensive defect: a bridge between adjacent cathode columns shows up as a row or block that lights when it should not. Sealing defects are slower to appear, and usually surface only after damp-heat or storage testing.

For buyers, the practical question is not which materials a factory lists but which of these parameters it monitors in line, with what sampling rate, and what it does with out-of-window sheets. Our PMOLED module range is built on that control discipline, and the same stack knowledge carries into our character OLED and graphic OLED modules.

PMOLED manufacturing is a short flow with narrow windows: pattern the anode, build the pillars, evaporate the stack, break the cathode by shadowing, seal. Every step is chosen so the next one can work, which is why process control, not equipment count, decides the yield.

References

1. Hilsum, C. (2010). Flat-panel electronic displays: a triumph of physics, chemistry and engineering. Philosophical Transactions of the Royal Society A, 368(1914), 1027–1082. https://doi.org/10.1098/rsta.2009.0247