Two processes, two physics
Sol-gel deposition is a wet-chemistry process. A metal-alkoxide precursor in a solvent is applied to the substrate by gravure, slot die, dip, spray, spin, or roll. The solvent evaporates, the precursor hydrolyses and condenses into an oxide network, and heat or UV completes the cure. The film grows from a liquid that touches the substrate. Every point the liquid wets becomes coated.
PVD covers a family of vacuum-deposited processes — sputtering, e-beam evaporation, thermal evaporation, ion-assisted deposition. A target material is vapourised inside a vacuum chamber and condenses onto the substrate. The film grows atom by atom from a directional vapour flux. Line-of-sight from source to substrate determines whether a point is coated and how thick.
That single difference — liquid wetting versus directional vapour — explains most of the practical comparison that follows.
Process specification comparison
| Process attribute | Sol-gel (wet-chemistry) | PVD (vacuum) |
|---|---|---|
| Operating environment | ambient pressure | high vacuum, 10⁻⁶ to 10⁻³ mbar |
| Substrate temperature | ambient to 700 °C (chemistry-dependent) | 50-300 °C typical; 400+ °C for hard dielectrics |
| Coverage | conformal on any wetted geometry | line-of-sight only; shadowing on 3D parts |
| Process mode | continuous R2R or sheet-fed | batch; some R2R sputter on flexible web |
| Typical web width | up to 2 m+ at production speed | up to 1.6 m R2R sputter; chamber-limited |
| Line speed | 10-50 m/min typical R2R | 0.5-5 m/min R2R sputter |
| Layer thickness range | 50 nm to several µm | 5 nm to ~2 µm |
| Layer thickness control | ±3-5% via wet-on-wet metrology | ±1-2% via quartz crystal monitor |
| Surface roughness contribution | smoothing effect on substrate | replicates substrate roughness |
| Multi-functional stacking | AR + hardcoat + anti-static in one stack | requires separate coating steps |
| Defectivity sensitivity to particles | filter-controlled (0.8 µm) | chamber-cleanroom-dependent |
Capex and throughput
A wet-coat sol-gel line of useful production capacity can be installed for a small fraction of the cost of an equivalent PVD line. The numbers below are indicative of commercial mid-range systems and will vary with width, speed, and integrated metrology. Treat them as orders of magnitude.
| Manufacturing factor | Sol-gel wet-coat line | PVD vacuum line |
|---|---|---|
| Capex (mid-range, 1 m web) | ~€0.5-2M | ~€10-40M |
| Footprint | ~150-400 m² | ~400-1500 m² |
| Throughput (1 m web, full stack) | ~600-3000 m²/h | ~30-300 m²/h |
| Energy intensity per m² | low (drying + UV cure) | high (vacuum pumps, plasma) |
| Time to first sample after retooling | hours | days (target change, vent cycle) |
| Material utilisation | >90% of formulation reaches substrate | 15-40% of target material reaches substrate |
| Indicative cost per m² (multi-layer AR) | low single-digit €/m² | double-digit €/m² |
The throughput delta is structural. Vacuum deposition is rate-limited by deposition physics and pump-down time. Wet-coat speed is rate-limited by drying and cure kinetics. The gap is approximately one order of magnitude in continuous production.
Achievable RI and stack design
Both routes can produce the high-RI / low-RI alternation needed for quarter-wave AR stacks. The achievable ranges differ.
| RI / stack parameter | Sol-gel (Kriya platform) | PVD typical |
|---|---|---|
| Low RI floor | 1.16 (porous SiO2 hybrid) | 1.38 (MgF2) |
| High RI ceiling | 2.00 | 2.40 (TiO2 dense) |
| RI tunability within one chemistry | continuous, formulation-controlled | discrete target materials |
| Broadband reflection floor | <0.15% | <0.1% |
| Number of layers for <0.5% AR | 2-3 | 3-5 |
| Graded-index layers | natively supported via formulation gradient | co-deposition required |
Kriya's RI 1.16 floor is meaningful because it raises the index contrast between high and low layers. Higher contrast means fewer layers are needed to hit a given reflection target. A 2-layer wet-coat AR stack with RI contrast of 1.95/1.16 reaches performance that requires 4 or more PVD layers to match.
Substrate compatibility
This is where the choice often resolves itself. PVD is excellent on rigid, thermally stable substrates that can survive vacuum cycling and process heat. Sol-gel is the practical choice for thermally limited polymers, flexible web, and complex 3D shapes.
| Substrate | Sol-gel | PVD |
|---|---|---|
| Flat float glass | excellent | excellent (industry standard) |
| Curved automotive glass | excellent (conformal) | limited; multi-axis fixturing required |
| PET / PC / PMMA film | excellent | feasible R2R sputter; thermal limits |
| TAC polariser cover | excellent (R2R) | marginal; substrate deformation |
| Foldable polymer (bend radius <3 mm) | excellent (hybrid flexibility) | brittle dielectric stack cracks on fold |
| Complex 3D shapes (lenses, domes) | good (dip, spray) | limited by line-of-sight shadowing |
| Silicon wafer / wafer-level optics | good (spin coat) | excellent (industry standard) |
| Architectural glass (large area) | excellent (jumbo size) | expensive at large scale |
Application-fit matrix
The summary below maps the most common optical-coating applications to the more practical of the two routes for volume production. “Either” means the choice depends on capex strategy and incumbent equipment.
| Application | Better fit | Why |
|---|---|---|
| Display AR on polymer film | sol-gel | R2R speed, no thermal damage |
| Polariser ULR film | sol-gel | TAC substrate, jumbo width, cost |
| Foldable display hardcoat + AR | sol-gel | flexible hybrid network, no crack |
| High-spec laser optics | PVD | layer thickness precision, dense oxides |
| Camera lens elements | PVD | small parts, fixture cost amortised |
| Automotive HUD combiner | either | size and curvature dominate decision |
| LIDAR window AR | sol-gel | tunable narrowband, large parts |
| AR/VR waveguide cores and cladding | sol-gel (UV-curable) | NIL replication, RI range, R2R |
| Solar panel front-glass AR | sol-gel | jumbo glass, cost per m² |
| Metalens master replication | sol-gel (high-RI UV) | NIL-compatible, RI up to 2.00 |
When PVD is still the right call
We are biased toward wet-coat — the chemistry catalogue we ship covers exactly the applications where sol-gel wins. There are clear cases where PVD remains the practical choice. Layer thickness control below 2 nm precision. Dense oxide stacks for high-fluence laser optics. Small-format precision optics where target material cost is negligible. Wafer-level processes integrated into existing fab equipment.
For everything that is large, flexible, geometrically complex, thermally limited, or cost-driven by square metres rather than parts — the sol-gel route wins on every dimension that matters in volume.
Process integration questions to answer
- Substrate temperature ceiling. Below 100 °C rules out most PVD options without active cooling. Sol-gel UV-curable systems cure at ambient.
- Geometry. If line-of-sight masking is required, sol-gel coats wherever the liquid touches.
- Web width and speed. Above 1.6 m or 5 m/min, R2R sputter struggles. Sol-gel R2R routinely runs wider and faster.
- Stack count. Beyond 5 layers, vacuum cycling cost compounds. Sol-gel multi-layer cures in sequence at line speed.
- Multi-functionality. If hardcoat, anti-static, or anti-smudge ride with the AR, sol-gel integrates them in one stack.