Solar cell laser processing covers two categories of application that share a name but have almost nothing else in common. Getting the source specification wrong because you conflated them is a common and expensive mistake. This guide separates the two applications, explains the P1/P2/P3 thin-film scribing architecture in detail, and gives quantified source requirements for each step and each solar technology type.
Two very different applications share the same name
The first application is crystalline silicon solar cell scribing — specifically, PERC (Passivated Emitter and Rear Cell) back-contact scribing. This is a high-speed, high-throughput process that uses a laser to open lines in the passivation layer on the back of a silicon wafer. The wafer is approximately 180 µm thick, and the scribing process is fast, powerful, and relatively forgiving of heat input because the silicon substrate underneath can withstand significant thermal load without damage.
The second application is thin-film solar module scribing — the P1/P2/P3 three-step laser patterning process used in CIGS, CdTe, and perovskite solar modules. This is a slow, precise, thermally sensitive process that uses laser ablation to selectively remove individual layers from a stack that may be as thin as 200 nm. A single degree of error in thermal management can form shunts that reduce module efficiency. The laser source requirements for this application are almost opposite to those for PERC scribing.
Before specifying a fiber laser source for any solar cell application, establish which of these two categories the application falls into. The sections below address each in sequence.
Crystalline silicon (PERC) scribing — high-speed, high-power, HAZ-tolerant
What PERC scribing is and why it’s different from thin-film
PERC solar cells use a passivation layer on the cell’s rear surface — typically silicon nitride (SiNx) or aluminum oxide (Al₂O₃) — to reduce carrier recombination and increase conversion efficiency. The laser scribing step opens local contact openings in this passivation layer, allowing aluminum from the rear metallization to form direct contact with the silicon. These openings are small, precise, and numerous — but they are made in a silicon wafer substrate, not in nanometer-scale thin films.
The tolerance for heat input in PERC scribing is fundamentally different from thin-film scribing. A silicon wafer can absorb significant thermal energy without phase change or chemical modification, and the passivation layers being removed are silicon-based materials with well-understood ablation thresholds. PERC scribing is a volume removal process on a robust substrate, not a selective layer-removal process on thermally fragile films.
Source requirements for PERC scribing — speed and throughput define the specification
These requirements — 25 m/s scan speed, thousands of lines per wafer, processing rates measured in thousands of wafers per hour — define the source and scanner architecture. Fast galvo scanners or polygon mirror scanners are required to achieve these speeds; conventional linear motion stages cannot. The laser source must provide sufficient average power at a high enough repetition rate that individual pulses overlap at the scan speed to produce continuous scribe lines, and pulse energy must be stable enough to prevent variation in scribe line width across the full 155 mm line length.
Wavelength for PERC scribing is typically 532 nm (green) for passivation layer ablation, where the shorter wavelength provides better absorption in silicon-based dielectrics, or 1064 nm for certain contact opening geometries. Average power requirements are modest by cutting standards — typically in the range of 2–15 W — because the material being removed is thin and the goal is selective ablation of the passivation layer, not bulk material removal.
Thin-film solar scribing — the P1/P2/P3 architecture
Thin-film solar modules — CIGS (copper indium gallium selenide), CdTe (cadmium telluride), and perovskite — achieve their monolithic series interconnection through three sequential laser scribing steps. This architecture is fundamental to thin-film module manufacturing: it converts a large-area monolithic thin-film stack into a series-connected array of individual subcells, each contributing its voltage to the module output.
The P1, P2, and P3 scribing steps sequentially pattern different layers of the thin-film stack to create the subcell boundaries and electrical interconnections. Each step targets a different layer, requires a different fluence range, and has different consequences for module performance if executed incorrectly. The laser source must provide the right combination of wavelength, pulse width, pulse energy, and beam quality for each step — and these requirements differ enough between steps that using the same unoptimized parameters for all three is a recipe for efficiency loss.
P1 scribing — removing the transparent conductive oxide layer
What P1 requires from the laser source
P1 is the first scribing step, performed after deposition of the transparent conductive oxide (TCO) layer on the glass substrate and before deposition of any semiconductor material. Its purpose is to scribe through the TCO layer — typically indium tin oxide (ITO) or fluorine-doped tin oxide (FTO), 200–500 nm thick — down to the glass surface, creating the electrical isolation boundaries between adjacent subcells.
P1 uses near-infrared wavelengths at approximately 1064 nm, with power levels up to approximately 8 W. At this wavelength, TCO materials have sufficient absorption for efficient ablation, and the glass substrate below is essentially transparent to 1064 nm — meaning the substrate is self-limiting for P1: even if the pulse energy slightly exceeds the TCO ablation threshold, the glass does not absorb the excess energy and is not damaged.
This self-limiting behavior makes P1 the most forgiving of the three scribing steps. The process window for P1 is comparatively wide — pulse energy must exceed the TCO ablation threshold but the upper limit is defined by collateral effects (TCO redeposition, edge quality) rather than substrate damage. Standard Q-switched or MOPA fiber laser sources operating at 1064 nm with nanosecond pulse widths are appropriate for P1 in most thin-film material systems.
Beam quality requirements for P1 — the dead zone connection
The width of the P1 scribe line is the first factor in determining the dead zone — the non-generating area between subcell boundaries where no light-to-electricity conversion occurs. To reduce scribe width, using lasers with beam quality factor M² less than 1.3 is generally recommended for thin-film solar scribing. A tighter, better-quality beam focuses to a smaller spot, produces a narrower scribe line, and contributes to a narrower total dead zone.
The economic significance of dead zone width is direct and quantifiable. The dead zone is the combined width from P1 through P3, typically 150–250 µm in production thin-film modules. Reducing the total dead zone width increases the geometric fill factor — the fraction of the module area that actively generates electricity — and new laser scribing technology can reduce dead area width to 200 µm or below, significantly reducing power loss. Every 100 µm reduction in total dead zone width across all three scribes translates to approximately 0.5–1% improvement in module-level conversion efficiency for a module with 4–5 mm subcell pitch. At production scale, this efficiency gain has direct revenue implications that justify investment in higher beam quality sources.
This is the quantitative link between fiber laser source beam quality specifications and solar module commercial performance: M² → focused spot diameter → P1 scribe width → dead zone contribution → geometric fill factor → module efficiency.
P2 scribing — the most demanding step
Why P2 is categorically harder than P1
P2 scribing is performed after deposition of the semiconductor absorber layer (CIGS, CdTe, or perovskite) on top of the P1-scribed TCO. Its purpose is to scribe through the absorber layer down to the TCO layer below, creating the conductive pathway that will connect adjacent subcells in series. The challenge is selectivity: the absorber layer must be completely removed in the scribe line, but the TCO layer immediately below must not be damaged.
The absorber layer is typically 1–3 µm thick. The TCO layer below it is 200–500 nm thick. The thermal diffusion length in a nanosecond-scale pulse is on the order of 100–500 nm. If the laser pulse deposits energy slowly enough that heat diffuses through the absorber layer into the TCO before the absorber has been removed, the TCO will be melted or damaged — creating a shunt resistance path that reduces the fill factor of the affected subcell. Under high-power and high-overlap conditions, substantial fill factor loss and up to a 75% increase in dead zone area have been observed in CIGS solar cells from P2 scribing damage. P2 is where module efficiency is most easily lost through suboptimal laser source selection.
Why pulse width below 5 ns is the critical source requirement for P2
The requirement for ultrashort pulses in P2 scribing is not arbitrary — it follows directly from the thermal diffusion physics of the multi-layer stack.
Research using a pulse-programmable fiber laser for CIGS P1, P2, and P3 scribing established that the optimal P2 and P3 scribing parameters require pulse duration less than 5 ns. Thermal dynamics dictate that slow rise and fall times spoil the brittle fracture ablation process in CIGS — the pulse must deposit energy faster than heat can diffuse to the underlying layer. Using a pulse-programmable fiber laser with precisely controlled sub-5 ns pulses, P2 and P3 scribing achieved zero heat-affected zone at the edges of the scribe — confirming that the pulse duration control is the mechanism enabling HAZ elimination.
The mechanism is thermal confinement. When a laser pulse is short enough relative to the thermal diffusion time of the absorber layer, the energy is deposited before significant heat transport to adjacent layers occurs. The absorber material reaches its ablation threshold and is removed; the TCO layer below sees only a negligible thermal pulse because the ablation event is complete before heat can diffuse through the removed material. This is sometimes called the “thermal confinement regime” of laser ablation, and achieving it for 1–3 µm absorber layers at the relevant thermal diffusivities requires pulse widths in the sub-5 ns range.
This is the fundamental argument for MOPA architecture in thin-film solar scribing: a MOPA source can independently set pulse width below 5 ns while maintaining the repetition rate and average power needed for production throughput. A Q-switched source with coupled pulse parameters — where changing repetition rate changes pulse energy and pulse width simultaneously — cannot maintain sub-5 ns pulses across the range of operating conditions needed for production P2 scribing.
The lift-off vs. direct ablation choice for P2
Two ablation mechanisms are available for P2 scribing, with different implications for edge quality and laser source requirements.
Direct ablation deposits energy primarily in the absorber layer, which is heated, ablated, and removed from the scribe path. This mechanism requires precise control of fluence — enough to completely remove the absorber but not enough to damage the underlying TCO — and produces scribe line edges that are defined by the thermal boundary of the ablation zone.
Lift-off ablation works by depositing energy primarily at the TCO/absorber interface rather than in the absorber bulk. The TCO absorbs the laser energy, and the resulting pressure impulse at the interface mechanically separates and ejects the overlying absorber material without directly heating it to its ablation temperature. This mechanism produces cleaner scribe edges with lower thermal damage because the absorber material is removed mechanically rather than thermally. Lift-off requires more precise pulse energy control — the fluence must exceed the lift-off threshold but remain below the TCO damage threshold — which is the reason MOPA architecture with accurate pulse energy setting is advantageous for this mechanism.
P3 scribing — back electrode isolation
P3 is the final scribing step, performed after deposition of the back electrode layer (aluminum, molybdenum, or other metals). P3 scribes through the back electrode, isolating the rear contact of one subcell from the rear contact of the adjacent subcell and completing the series interconnection structure.
P3 scribe lines are positioned between the P1 and P2 scribes, and the separation between the P2 and P3 scribes is approximately 125 µm in production CIGS modules. Precision of placement relative to the P1 and P2 scribes is a mechanical alignment requirement — handled by the motion system and vision alignment of the scribing tool — rather than a laser source specification.
P3 source requirements are similar to P2: sub-5 ns pulse width to prevent thermal damage to the absorber layer and TCO structure underneath the back electrode being removed, and precise pulse energy control to ensure complete back electrode removal without excessive heat propagation downward. P3 processes using the same pulse-programmable fiber laser as P2 also achieved zero heat-affected zone under optimized parameters. The operational insight is that P2 and P3, despite scribing different layers, share the same fundamental source requirement: short pulses with independently controllable parameters.
The total dead zone is the combined width of P1, P2, and P3 scribe lines plus the required spacing between them. In silicon thin-film solar modules, a total dead zone of 150 µm has been achieved — with 30, 45, and 50 µm separation for P1, P2, and P3 ablations respectively — by using carefully optimized laser peak power, focus, and pulse width parameters. Achieving dead zone widths at this level requires all three scribing steps to be executed with beam quality and pulse parameters that minimize the HAZ contribution of each step.
Perovskite solar modules — the emerging scribing challenge
Why perovskite is more thermally sensitive than CIGS or CdTe
Perovskite solar cells have achieved certified efficiencies exceeding 26.1% in small-area cells, making them the most rapidly advancing solar technology. The sequential ablation of wide-band-gap oxide materials and metal-organic films in perovskite scribing requires pulsed laser techniques, because the layer-by-layer selectivity required cannot be achieved with CW laser or insufficiently short pulses.
Perovskite absorber materials (typically methylammonium lead iodide or formamidinium-based halide perovskites) are chemically sensitive to thermal decomposition in ways that CIGS and CdTe are not. The organic components of the perovskite begin to decompose at temperatures well below the ablation threshold of the material, meaning that any thermal damage zone surrounding the scribe line represents not just altered morphology but altered chemical composition. A chemically degraded zone at the scribe boundary increases leakage current and reduces open-circuit voltage in the adjacent active area.
This thermal sensitivity makes perovskite scribing the most demanding application among current commercial thin-film technologies, and it places a premium on ultrashort pulse capability — sub-5 ns pulses that confine energy deposition to the target layer without thermal spread to adjacent material.
Single-source scribing — the cost reduction opportunity
A significant challenge in thin-film solar manufacturing has been that P1, P2, and P3 have traditionally required different laser wavelengths or pulse characteristics, leading to multi-source scribing tools with higher capital cost and maintenance complexity.
A single infrared fiber laser source can perform all three scribes — P1, P2, and P3 — in a perovskite solar module through an indirect lift-off mechanism. By using a TCO-mediated selective lift-off approach, the same 1064 nm source with nanosecond-to-microsecond pulse duration capability can complete all three scribing steps while avoiding residual thermal damage. This dramatically reduces equipment cost and process complexity for perovskite module manufacturing.
The enabling technology for single-source scribing is MOPA architecture: a pulse-programmable MOPA fiber laser source can independently set the pulse duration for each scribing step — relatively longer pulses for P1 where thermal tolerance is higher, progressively shorter pulses for P2 and P3 where thermal confinement is critical. This flexibility, which is not available in Q-switched sources with coupled pulse parameters, allows a single fiber laser source to cover the full scribing sequence.
Source architecture comparison — Q-switched vs. MOPA for solar scribing
The choice between Q-switched and MOPA fiber laser source architecture directly determines what solar scribing applications can be addressed and with what performance.
| Attribute | Q-switched | MOPA |
|---|---|---|
| Pulse width control | Coupled to repetition rate — changing one changes the other | Independent — pulse width and repetition rate set separately |
| Pulse width achievable | Typically 5–50 ns depending on operating point | 2–500 ns, set electronically per application |
| HAZ at P2/P3 | Variable — sub-5 ns may not be achievable across full operating range | Achievable at production repetition rates |
| Single-source P1+P2+P3 | Not feasible — parameters cannot span all three requirements | Feasible — different parameter sets for each step |
| PERC crystalline silicon | Adequate — pulse precision less critical | Adequate — overkill for this application |
| Thin-film CIGS P1 | Adequate | Adequate, with more flexibility |
| Thin-film CIGS P2/P3 | Marginal to inadequate — sub-5 ns coupling is the limitation | Well-suited — sub-5 ns independently controllable |
| Perovskite P1+P2+P3 | Inadequate for single-source scribing | Enables single-source scribing |
| Cost | Lower | Higher |
Source specification requirements — a buyer’s reference
| Application | Wavelength | Pulse width | Average power | M² | Repetition rate |
|---|---|---|---|---|---|
| PERC crystalline silicon | 532 nm or 1064 nm | 5–50 ns | 2–15 W | < 1.5 | 50–500 kHz |
| Thin-film P1 (TCO scribing) | 1064 nm | 10–50 ns | 1–8 W | < 1.3 | 20–200 kHz |
| Thin-film P2 (absorber scribing) | 1064 nm | < 5 ns | 0.5–3 W | < 1.3 | 100–500 kHz |
| Thin-film P3 (back contact) | 1064 nm | < 5 ns | 0.5–3 W | < 1.3 | 100–500 kHz |
| Perovskite single-source | 1064 nm | 2 ns – 1 µs (programmable) | 1–5 W | < 1.3 | 20–500 kHz |
Notes on the table: The average power figures for thin-film scribing are low by cutting machine standards — this is not a power-limited application. The critical source specifications for thin-film scribing are pulse width control (particularly for P2/P3) and beam quality (which determines scribe line width and dead zone). Average power is almost always adequate with standard MOPA fiber laser sources; pulse width and M² are where selection decisions are made.
For PERC crystalline silicon, the high throughput requirement (3,600 WPH, 25 m/s) makes the scanner system the throughput-limiting component rather than the laser source — any source that provides adequate average power at the required repetition rate enables the throughput target, provided the scanner can sustain 25 m/s with appropriate laser control.