Efficient ultrafast laser writing with elliptical polarization

The study investigates how the polarization state of ultrafast laser pulses affects permanent birefringent modifications in silica glass, challenging the common expectation that stronger absorption (and linear polarization) yields stronger permanent change. While linear polarization typically maximizes multiphoton ionization and has been primarily used to form self-assembled nanogratings (type II) and anisotropic nanopores (type X) in silica, the authors ask whether elliptical or circular polarization can more efficiently produce high-transmission birefringence and enhance writing throughput. This question is important for applications such as polarization shaping, geometric phase optics, and 5D optical data storage, where higher birefringence at lower loss and faster writing speeds are desirable. The work shows that, despite lower absorption, elliptically polarized pulses can produce stronger birefringence than linearly polarized pulses and explores the underlying mechanisms and implications for device fabrication and data storage.

Prior research has established ultrafast laser-induced nanostructures in transparent materials, including self-assembled nanogratings (type II) in silica aligned perpendicular to linear polarization, and surface periodic structures in metals, semiconductors, and dielectrics under linearly polarized irradiation. Induced optical anisotropy is widely used in polarization shaping and geometric phase optics, and 5D optical data storage exploits birefringence (slow axis and retardance) for multiplexing. Previous work predominantly used linear polarization; circular/elliptical polarization studies reported nanoparticle formation on crystal surfaces, polarization-dependent orientation and period of surface structures, and ellipticity-controlled capillary formation in silica via laser-assisted etching. Photoionization in fused silica is known to be ~3–4× higher for linear than circular polarization at <15 TW/cm² but becomes higher for circular polarization at >35 TW/cm² due to tunneling; stronger index change under circular polarization at ~45 TW/cm² was reported in silica, while other materials in multiphoton-dominated regimes did not show enhanced modification with ellipticity. Ultra-high-transmission birefringent modifications consisting of randomly distributed oblate nanopores (type X) enable >99% efficient phase/polarization shaping and long-lifetime data storage, but writing speed and transmission (UV/visible) constraints remain challenges.

• Material: Silica glass (fused silica).
• Laser source/conditions: 1030 nm wavelength; pulse durations of 300 fs and 600 fs; repetition rates of 1 MHz (voxel studies) and 200 kHz (raster writing); numerical apertures: 0.3 NA (voxel/absorption studies) and 0.16 NA (raster writing).
• Writing parameters:
  • Voxel imprinting: pulse energies 200 nJ (≈9.8 TW/cm²) and 220 nJ (≈10.8 TW/cm²); pulse counts varied (20, 30, 40, 60, 200, 300 pulses); ellipticity varied from 0 (linear) to 1 (circular); two orthogonal linear polarization orientations used. Voxel spacing 2 µm; background removal algorithm applied for precise retardance measurement.
  • Raster-written square areas: 20×20 µm squares, 1 µm line spacing; energies 0.7–1.0 µJ (notably 0.9 µJ and 1.0 µJ examples), durations 300 fs and 600 fs; scanning speed 10 mm/s.
• Measurements:
  • Retardance and slow-axis orientation maps to quantify birefringence (type X and type II regimes distinguished).
  • Nonlinear absorption measured for single pulse (1p) and the 30th pulse after 29 prior pulses (29+1p) versus ellipticity at 200 nJ, 300 fs, 1 MHz, 0.3 NA.
  • Birefringence magnitude estimated as Δn1 = Ret./l; refractive index change inferred (including negative index change).
  • SEM imaging after polishing and KOH etching to assess nanopore morphology and density for linear vs elliptical polarization-written regions; note that etching inflates apparent pore size.
  • Optical transmission measured at 550 nm for birefringent squares.
• Classification: Type X (random oblate nanopores, ultra-high transmission) vs Type II (nanogratings) identified by retardance trends and SEM; transitions explored as function of ellipticity, pulse number, energy, and duration.

• Elliptical polarization enhances birefringence despite lower absorption: For 200 nJ, 300 fs, 30 pulses, maximum retardance ≈1.5 nm at ellipticity ≈0.6 vs ≈1.0 nm for linear polarization; corresponding birefringence ≈1.5×10⁻⁴ vs 0.8×10⁻⁴ (≈1.9× higher) while absorption is ≈2×–2.5× lower (linear ≈5% vs elliptical (e≈0.6) ≈2%).
• Absorption vs ellipticity: Single-pulse absorption at 200 nJ, 300 fs is ~5% (linear) and ~1% (circular); for the 30th pulse (after 29 prior), absorption increases due to defect accumulation but still decreases with increasing ellipticity.
• Refractive index change: Negative index change estimated as large as −2.5×10⁻⁴ for circular polarization and −0.5×10⁻⁴ for linear (30 pulses, 200 nJ, 300 fs), indicating cavitation-driven density reduction under circular components.
• Ellipticity-controlled modification type: At 220 nJ (10.8 TW/cm²), with >40 pulses, type II (nanograting) forms for ellipticity <0.4, while type X persists for more elliptical polarization; with >200 pulses, only type II observed for ellipticity <0.4 with retardance drop at higher ellipticity. Increasing ellipticity promotes type X and suppresses transition to type II.
• Nanopore morphology and density: SEM reveals larger and denser anisotropic nanopores for elliptical polarization (apparent 30–40 nm after etch; estimated ≈25 nm actual) vs linear (apparent 20–30 nm; ≈15 nm actual). Nanopores are flattened perpendicular to the major axis of the polarization ellipse, producing birefringence.
• High transmission maintained: Birefringent squares show similar transmission at 550 nm for linear (98.6%) and elliptical (98.4%) writing.
• Retardance gains in element writing:
  • 600 fs, ~1.0 µJ (≈8.6 TW/cm²): maximum retardance ≈83 nm at ellipticity ≈0.6, ~1.5× that of linear; at <700 nJ (≈6.0 TW/cm²), no increase with ellipticity.
  • 300 fs: retardance increases with ellipticity; maximum ≈42 nm at ellipticity ≈0.7, about 2.5× linear.
• Application demonstration: Multilayer 5D optical data storage using elliptical polarization achieved with writing speeds of tens of kB/s and nearly 100% readout accuracy; increased throughput in birefringence patterning for polarization shaping and geometric phase optics.

The findings overturn the intuitive link between stronger absorption and stronger permanent anisotropy: despite reduced multiphoton absorption for elliptical/circular polarization, ellipticity enhances the formation of anisotropic nanopores (type X) and increases birefringence. The authors attribute this to a hybrid mechanism wherein the linear component of an elliptically polarized pulse provides near-field anisotropy that flattens nanopores, while the circular component increases the efficiency of nanopore creation. Circular polarization interacts more effectively with randomly oriented bonds and hole polarons (localized states/defects) in the amorphous silica network, with a rotating electric field sampling more dipole orientations and potentially enhancing tunneling ionization of low-excitation-energy defects (<~2 eV) under high intensities (~10 TW/cm²). Although free electrons acquire higher kinetic energy under circular polarization (estimated ~2.15 eV vs ~0.34 eV for linear at 11 TW/cm², 1030 nm), the electron heating rate in solids is not strongly ellipticity-dependent due to collisions; nevertheless, the circular component favors cavitation and density reduction, consistent with the observed larger negative index change. Ellipticity thus shifts the balance towards type X formation and away from self-organized nanogratings (type II), yielding higher-transmission birefringence with larger retardance. These mechanisms directly address the research question by explaining why elliptical polarization can outperform linear polarization in producing efficient birefringence and how this can be tuned for practical applications.

Elliptically polarized femtosecond laser writing in silica glass yields higher birefringence with lower absorption compared to linear polarization, with optimal ellipticities around 0.6–0.7 producing up to ~2× (voxels) to ~2.5× (raster patterns at 300 fs) increases in retardance and enabling control over modification type (favoring high-transmission type X nanopores over type II nanogratings). The enhanced performance is explained by the complementary roles of linear and circular polarization components in anisotropic nanopore formation and by efficient interaction with randomly oriented defects via tunneling ionization. These advances translate to improved throughput for birefringence patterning and enable high-fidelity, faster 5D optical data storage. Future research could quantitatively model defect/tunneling pathways vs ellipticity, extend the approach to other glasses and transparent materials, optimize pulse parameters (duration, repetition rate, energy) for maximum throughput without transition to type II, and investigate long-term stability and environmental robustness of ellipticity-optimized modifications.

• The study is limited to silica glass; generalization to other transparent materials is not demonstrated.
• Exact affiliation of one author (³) is not provided in the excerpted text.
• Mechanistic explanations (defect-mediated tunneling, rotating-field access to dipoles) are interpretative; direct time-resolved measurements of defect populations and ionization pathways are not presented.
• SEM-based nanopore size estimates are inflated by etching; actual pore sizes are inferred.
• At lower pulse energies/intensities (<~700 nJ, ~6 TW/cm²), ellipticity does not increase retardance, indicating a parameter window for the effect.
• To avoid transition to type II at higher doses, writing conditions must be carefully constrained, potentially limiting maximum single-pass retardance.
• Data storage demonstration mentions speeds of tens of kB/s and near-100% accuracy but lacks detailed benchmarking against state-of-the-art and long-term cycling tests.