From the fall of α-Si and the choice of a PMOS driving structure to LTPO’s 1–120Hz adaptivity, six core questions explain why mobile backplane technology converged on LTPS/LTPO — plus a "ZJWY view" on what these principles mean on the repair bench.
Look at the display panels in today's mainstream mobile devices and one feature stands out: whether it is a flagship high-refresh screen, a flexible foldable, or a mid-range panel, the backplane driving technology is almost invariably LTPS (Low Temperature Poly-Silicon) or LTPO (Low Temperature Polycrystalline Oxide) — not the older, cheaper α-Si (amorphous silicon).
This article answers why through six core technical questions.
| Concept | What it is | Relationship |
|---|---|---|
| TFT (thin-film transistor) | The pixel driving unit of a display panel — it switches pixels on/off and controls brightness and colour | The starting point of everything |
| α-Si / LTPS / LTPO | Three different TFT backplane technology branches | Parallel technology routes |
| PMOS / NMOS | P-channel / N-channel metal-oxide-semiconductor transistors — two driving structures of a TFT | A structural choice within the same route |
So the question reduces to one sentence: which combination of TFT technology and driving structure best fits the mobile display scenario?
α-Si was the first TFT technology to reach industrial maturity. Its process is simple and its cost low, and it is still widely used in TVs, computer monitors and low-end tablets — large-size, low-resolution devices. But it has two weaknesses that make it untenable for phones.
Weakness 1: low carrier mobility. Carrier mobility measures how fast charge moves inside the material — the higher it is, the faster the TFT switches and the stronger its driving capability. α-Si has an electron mobility of only 0.5–1 cm²/(V·s). Today's mainstream phones are 2K, some flagships 4K, with pixel densities above 500 PPI and every pixel driven by its own TFT. Insufficient mobility directly delays pixel switching, showing up as image smearing, slow response and display misalignment.
Weakness 2: oversized TFT devices. Space in a phone panel is extremely constrained. At high PPI the usable area of a single pixel is tiny, and TFT size directly determines aperture ratio, brightness and power. α-Si's loose lattice cannot be effectively miniaturised; forcing it into a phone panel only brings lower aperture ratio, insufficient brightness and a power spike — and battery life collapses with it.
💡 In one line: α-Si is not impossible to build — it simply cannot reach the resolution, refresh rate and power levels a phone demands.
LTPS converts amorphous silicon into poly-silicon through a low-temperature process. The poly-silicon lattice is more complete with more regular atomic arrangement, giving two direct advantages:
The core substrate of a TFT is silicon (Si). Pure silicon is an intrinsic semiconductor with weak conductivity; it must be doped to change its conduction characteristics and form a controllable transistor.
Boron (B) is a group III element with only 3 outer electrons, while silicon has 4. When boron atoms enter the silicon lattice they replace some silicon atoms; being one outer electron short, they leave an empty electron orbital — a hole. Holes act as positive charge carriers, and a transistor that conducts by holes is a PMOS (P-channel Metal-Oxide-Semiconductor) transistor.
Correspondingly, doping with phosphorus (P, a group V element with 5 outer electrons) produces an NMOS: phosphorus has one more outer electron than silicon, and that surplus electron becomes a free electron acting as a negative charge carrier.
So the "P" in PMOS essentially refers to the fact that the poly-silicon substrate is doped with boron — which is exactly why LTPS display panels are predominantly PMOS-driven. Distinguishing PMOS from NMOS conduction is the premise for everything that follows.
In semiconductor characterisation, the transfer characteristic (Id-Vg curve) is the core metric for driving performance, describing how drain current (Id) responds to gate voltage (Vg). Its steepness reflects the transistor's sensitivity to gate voltage: the steeper the curve, the more violently drain current swings for a tiny change in gate voltage — and the harder the device is to control.
NMOS is inherently steeper than PMOS, and the core reason is that the carrier in NMOS is the electron, whose mobility far exceeds that of the hole.
An electron has an extremely small mass. Moving through the silicon lattice it is barely disturbed by lattice vibration or impurity scattering, so it travels almost freely and at high speed. A hole is not a real particle: its "movement" is the equivalent process of valence electrons hopping between atoms to fill vacancies. Those valence electrons are strongly bound by chemical bonds, have a large effective mass, and move markedly slower than electrons.
This difference appears directly in the curve: when gate voltage changes, free electrons in NMOS respond quickly and drain current rises steeply, whereas holes in PMOS respond more slowly, producing a gentle slope. That difference in current-control precision is decisive for AMOLED display quality.
This needs the energy-band picture from solid-state physics.
Electron energies in a semiconductor occupy different levels: higher-energy electrons able to break free of the nucleus sit in the conduction band; lower-energy, bound electrons sit in the valence band; between them lies the band gap, which an electron must absorb enough energy to cross.
For the NMOS carrier (the electron), it sits in the conduction band, already free of the nucleus. There it is barely constrained by chemical bonds and only lightly scattered by lattice vibration and impurities — the overall binding is very weak, so it moves fast and its mobility is high.
For the PMOS carrier (the hole), movement is really valence-electron migration. Valence electrons are bound in the valence band; when external energy (such as gate voltage) is applied, they hop from one atom's valence band to another atom's empty valence band, producing the equivalent motion of a hole. Because those electrons remain strongly bound, they cannot move as freely as conduction-band electrons — hence hole mobility is much lower.
Industry test data: in silicon, electron mobility is about 1500 cm²/(V·s) while hole mobility is only about 500 cm²/(V·s) — a gap of roughly 3×.
This natural difference in carrier mobility determines the performance gap between PMOS and NMOS and provides the core basis for choosing the driving structure of a mobile display panel.
The contradiction is obvious: NMOS has higher mobility, faster switching and better driving efficiency, yet it did not become the core driving structure of mobile AMOLED. The reason is that AMOLED demands current-control precision far more than switching speed.
AMOLED is a current-driven device: each pixel is an organic light-emitting diode whose brightness correlates with the current through it. Phones need fine grey-scale transitions and stable brightness — especially at low brightness, in always-on display (AOD) and in HDR dimming — which requires precise control of microamp-level currents. Even a slight fluctuation causes brightness flicker, grey-scale distortion and colour shift.
On this point the two devices sit at opposite ends:
Low-power requirements reinforce the advantage further: PMOS leakage current in the off state is significantly lower than NMOS, which meaningfully cuts panel power in low-power scenarios such as always-on display and extends battery life. For these reasons, mainstream mobile AMOLED panels are predominantly PMOS-driven.
Building on LTPS, the industry developed LTPO (Low Temperature Polycrystalline Oxide). Today's high-end flagships and flexible foldables treat LTPO as standard, because it combines the high mobility of LTPS with the low power of an oxide semiconductor such as IGZO.
The logic of LTPO is a hybrid "LTPS + oxide" driving structure: high-refresh, high-resolution areas are driven by LTPS for response speed and driving strength; low-refresh, low-brightness areas are driven by oxide TFTs to cut leakage and overall power. This enables adaptive refresh from 1 to 120Hz — smooth in games and video, efficient in reading and always-on states — matching the diverse usage of a phone.
| Technology | Carrier mobility | Leakage / power | Refresh capability | Typical use |
|---|---|---|---|---|
| α-Si amorphous | 0.5 – 1 cm²/(V·s) | Low leakage, but poor brightness raises overall power | Low | TVs, monitors, low-end tablets |
| LTPS | ≥ 100 cm²/(V·s) | Moderate | High (up to 120 / 144Hz) | Mid-to-high-end phones, tablets |
| LTPO | High in LTPS zones, low in oxide zones | Low (little leakage on static content) | 1 – 120Hz adaptive | Flagships, flexible foldables |
The step from LTPS to LTPO is ultimately a pursuit of "high performance and low power together" — and its foundation remains LTPS's core advantages: high carrier mobility, miniaturised TFT devices, and the precise current control of a PMOS driving structure.
Display panel technology choice is fundamentally the deep matching of semiconductor material properties, process capability and end-application scenario. The point was never to chase extreme specs, but to achieve optimal fit to the scenario.
For the screen-repair industry the value of this article is not memorising three letters — it is understanding that a screen is layered, and which repair method maps to which layer. From outside in, a phone screen is roughly: polariser (optical layer) → emitting layer / liquid crystal → TFT backplane array → driver IC (DDIC) → flex cable (FPC / COF) → traces at the screen root. LTPS and LTPO belong to the backplane layer alone — and that layer determines whether a fault is backplane-level or connection-level.
1. Layer first, then pick the machine. Black screen, white screen, green screen, vertical lines — the fault point may be the flex cable, the ITO traces, the DDIC, or the screen root. Flex faults go to flex-replacement machines (EN-580 / EN-50C series bonding machines); ITO faults go to wire jumping; DDIC faults go to IC pre-bonding equipment (such as the EN-801S); screen-root faults go to laser machines (such as the EN-LS26 series). Skipping the layer analysis and cutting straight in is the single biggest driver of rework.
2. Why high-refresh and foldable models are more fragile to repair. An LTPO panel carries both LTPS and oxide zones with different TFT characteristics, and at high refresh rate the per-frame charging time is compressed, so insufficient TFT charging amplifies process variation — the same panel can show different Mura distribution and Gamma curve shape at normal versus high refresh. This explains a common sight on the bench: after flex bonding, cable replacement or polariser re-lamination, low refresh is fine but high refresh flickers, or local Mura appears. That is usually not a wrong repair — it is process consistency being magnified at high refresh, which in turn raises the bar for alignment accuracy, pressure and temperature profile stability.
3. Does high mobility make a screen easier or harder to repair? LTPS's high mobility, strong driving and high aperture ratio are exactly why phone panels must miniaturise their TFTs. Conversely, once the TFT array in some region of the backplane itself fails — for example from ESD or excessive laser energy — repair is far harder than a cable-level fault. Those screens are better served by microscopic confirmation before cutting, so that a repairable screen is not turned into scrap.
4. Inspection must keep up with device scale. A miniaturised backplane means sub-pixel-scale defects. Confirming the cause of bright/dark dots or line defects needs sufficient magnification and imaging resolution — metallurgical and electronic microscopes inspect ITO traces, bonding zones and sub-pixel structures, while the machine-vision alignment optics on bonding and IC pre-bonding machines address lamination and bonding accuracy. One is about seeing, the other about doing.
The conclusion is plain: understanding backplane technology is not about memorising specs — it is about deciding within thirty seconds whether a screen is worth repairing, and which machine to put it on.
← Back to Support