Laser-Induced Amorphization: Mechanisms & Control

• Laser-induced amorphization is the process of converting crystalline materials to amorphous states via precise ultrafast laser pulses, driven by thermal, nonthermal, and shock mechanisms.
• The studies reveal that variations between retained and transient amorphization depend critically on energy deposition, fluence, wavelength, and substrate temperature, which govern melt depth and phase stability.
• Advanced theoretical models—including Drude–Kerr, two-temperature, and hybrid nTTM–MD approaches—quantitatively link laser parameters to ultrafast structural transitions and controlled phase-change applications.

Laser-induced amorphization is the formation of a non-crystalline state in an initially crystalline material under laser irradiation. In the reported literature, it includes a $\sim 30 \; \mathrm{nm}$ amorphous germanium cap formed concurrently with high spatial frequency laser induced periodic surface structures (HSFL) under mid-IR femtosecond pulses, amorphous silicon layers generated by 1030 nm and 800 nm ultrafast excitation with strong dependence on substrate temperature and quench kinetics, reversible amorphization of Sb$_2$S$_3$ thin films under 532 nm single pulses, shock amorphization of Fe$_2$O$_3$ between 122(3) and 145(10) GPa, and transient collapse/amorphization of titanium in ultrashort-laser-driven shock fronts (Austin et al., 2016, Kuz et al., 19 May 2026, Laprais et al., 2024, Crépisson et al., 2024, Zhakhovsky et al., 2023, Ivanov et al., 2024). Across these systems, the decisive variables are the energy-deposition pathway, the pressure or carrier-density history, and the competition between melting, quenching, recrystallization, and damage.

1. Phenomenological regimes

The reported cases separate into retained amorphization and transient amorphization. Retained amorphization is observed as a surface-confined or film-confined non-crystalline region after irradiation: a $\sim 30 \; \mathrm{nm}$ cap on crystalline Ge, an amorphous ring or thin layer on crystalline Si, and reversible amorphization of a 42 nm Sb$_2$S$_3$ film. Transient amorphization is observed under shock loading, where the non-crystalline state may exist only along part of the pressure–temperature path before transforming further into a liquid or recrystallized solid; this is the case for Fe$_2$O$_3$ under laser-driven shock compression and for Ti in the shock front (Austin et al., 2016, Kuz et al., 19 May 2026, Laprais et al., 2024, Crépisson et al., 2024, Zhakhovsky et al., 2023).

The mechanisms also differ. In Ge and Si, nonlinear absorption, dense carrier generation, nonthermal melting, and ultrafast solidification are central. In Ge, two photon absorption and defect-induced amorphization are discussed as probable causes of the amorphous layer. In Si, both a carrier dependent two-temperature model and a hybrid nTTM–MD treatment identify melting and quench kinetics as the discriminating factors. In Sb$_2$0S$_2$1, the reported amorphization pathway is explicitly tied to optical absorption and transient heating through the film stack. In Fe$_2$2O$_2$3 and Ti, the dominant drivers are shock compression, high strain rate, and the evolution of short-range order under extreme pressure (Austin et al., 2016, Kuz et al., 19 May 2026, Ivanov et al., 2024, Laprais et al., 2024, Crépisson et al., 2024, Zhakhovsky et al., 2023).

This suggests that “laser-induced amorphization” is not a single microscopic process but a family of nonequilibrium phase-transition pathways. The common endpoint is loss of crystalline long-range order, but the route to that endpoint may be optical, electronic, thermal, mechanical, or coupled.

2. Surface-confined ultrafast amorphization in germanium and silicon

In germanium, femtosecond mid-IR pulses with wavelengths between $_2$4 and $_2$5 generate HSFL whose period varies from $_2$6 to $_2$7. Under peak fluence $_2$8–$_2$9, high enough to produce HSFL but below LSFL-forming fluences, cross-section TEM reveals a $_3$0 amorphous germanium layer immediately beneath a native Ge-oxide of $_3$1 and a $_3$2 Au protection layer, with well-ordered crystalline Ge beneath. Electron diffraction through the top $_3$3 shows diffuse rings, whereas diffraction through the bulk shows discrete spots. The amorphous cap is reported for $_3$4 cross sections, and HSFL together with the cap appear independent of polarization. The proposed mechanisms are two-photon absorption, described by $_3$5, and defect-induced amorphization under multi-pulse accumulation; using $_3$6 scaled to $_3$7, the estimated energy deposition extends $_3$8 and exceeds the stated melting criterion $_3$9 (Austin et al., 2016).

In crystalline silicon irradiated by a single 300 fs, 1030 nm pulse, the reported pathway begins with nonlinear absorption, predominantly two-photon absorption, producing a dense $_2$0–$_2$1 plasma with $_2$2–$_2$3 on a sub-ps timescale. At such carrier densities the interatomic potential is strongly renormalized, and the lattice disorders before significant lattice heating by electron-phonon coupling can occur. Cryogenic temperature strengthens the retained amorphous outcome. At 300 K, above $_2$4 the damage site exhibits a shallow central pit with only a faint, narrow amorphous periphery; at 77 K the amorphous ring is clearly visible at all fluences from $_2$5 to $_2$6; and at 24 K the ring is brightest and widest. At $_2$7, the ring width grows from $_2$8 at 300 K to $_2$9 at 24 K. The outer-ring threshold shifts from $_3$0 at 300 K to $_3$1 at 77 K and $_3$2 at 24 K. Raman spectroscopy identifies a pristine Si LO phonon at $_3$3 with $_3$4 and amorphous Si as a broad peak near $_3$5 with $_3$6–$_3$7; the ratio $_3$8 grows from $_3$9 at 300 K to $\sim 30 \; \mathrm{nm}$0 at 24 K, while the a-Si layer thickness in the ring reaches $\sim 30 \; \mathrm{nm}$1 at 24 K, drops to $\sim 30 \; \mathrm{nm}$2 by 120 K, and vanishes above 210 K. AFM shows the ring is $\sim 30 \; \mathrm{nm}$3 depressed relative to surrounding Si, with RMS roughness $\sim 30 \; \mathrm{nm}$4, and KPFM records a $\sim 30 \; \mathrm{nm}$5 surface-potential shift between 300 K and 24 K (Kuz et al., 19 May 2026).

A second Si study, using a hybrid atomistic-continuum model for a 100 fs pulse at 800 nm, resolves the kinetic side of the problem. For 100 fs, 800 nm pulses on $\sim 30 \; \mathrm{nm}$6 Si, the MD–nTTM predicts a melting threshold $\sim 30 \; \mathrm{nm}$7, whereas a stand-alone nTTM gives $\sim 30 \; \mathrm{nm}$8. At $\sim 30 \; \mathrm{nm}$9, the total melt depth is 60 nm for $_2$0 and 45 nm for $_2$1. The model further reports that if the cooling rate exceeds $_2$2, the solid–liquid interface becomes immobile even when $_2$3, trapping a presurface amorphous layer, and MD quenching experiments from $_2$4 yield an $_2$5 amorphous film at $_2$6 (Ivanov et al., 2024).

3. Reversible amorphization in Sb$_2$7S$_2$8 thin films

For Sb$_2$9S$_3$0 thin films, laser-induced amorphization is reported as a reversible phase-change process constrained by optical absorption, transient heat flow, and the surrounding stack. The investigated geometry is a Si substrate / Sb$_3$1S$_3$2 / SiO$_3$3 cap stack under normal incidence at $_3$4 nm. For a 42 nm film, the melting point is first reached $_3$5 nm below the cap interface at a fluence $_3$6, and the very surface reaches $_3$7 at $_3$8. For $_3$9 the entire 42 nm film can be amorphized, whereas beyond $_2$0 the peak temperature would exceed the evaporation point and cause damage. The measured threshold for the 42 nm film is $_2$1, and the damage threshold is observed around $_2$2 (Laprais et al., 2024).

The same study reports partial amorphization and partial recrystallization as controllable states. Recrystallization is driven by a CW 532 nm beam with the same spot size, and the minimum crystallization time decreases from $_2$3 to $_2$4. Because the total energy $_2$5 is not constant, the reported interpretation is that crystallization kinetics, taken as grain-growth limited, also participate; the grain-growth rate is written as

$_2$6

By stopping the CW exposure before completion, for example at $_2$7 and $_2$8, one can program a partially recrystallized depth (Laprais et al., 2024).

Polycrystallinity and optical anisotropy materially broaden the switching window. Thermally crystallized Sb$_2$9S$_3$0 is orthorhombic and forms $_3$1-scale grains with different in-plane orientations, and Mueller-matrix imaging reveals grain-dependent absorption coefficients. Experimentally, thresholds on a 42 nm film vary locally as $_3$2, described as a $_3$3 spread purely from grain-to-grain anisotropy. Thick-film operation is limited not by a quoted thermal thickness estimate but by the optical absorption profile: above $_3$4 nm, simulations and TEM show that one cannot fully melt all the way to the substrate without damaging the surface. For a 93 nm film at 532 nm, the bottom $_3$5 nm never reach $_3$6, and at the maximum non-damaging fluence only $_3$7 nm are amorphized. A proposed strategy is to shift to 660 nm, where $_3$8 is smaller; this flattens the absorption profile and raises the amorphized depth to $_3$9 nm, but requires higher pulse fluence, for example from 39 to $_2$00 to reach the same peak temperature. At least 10 write/erase cycles were demonstrated optically without drift (Laprais et al., 2024).

4. Shock-driven amorphization and mechanical melting

In Fe$_2$01O$_2$02, the reported laser-induced route to amorphization is indirect: the laser drives a shock, and the compressed material is probed in situ by femtosecond XFEL diffraction. Crystalline $_2$03-Fe$_2$04O$_2$05 Bragg peaks vanish above 122$_2$063 GPa and are replaced by two broad maxima at $_2$07–$_2$08 and $_2$09–$_2$10. This diffuse pattern persists up to 145(10) GPa, defining a shock-induced amorphous regime. Between 145$_2$1110 GPa and 151$_2$1210 GPa, the intensity ratio $_2$13 climbs rapidly, $_2$14 grows to exceed $_2$15, and both peaks shift by $_2$16 and broaden. The reported interpretation is an amorphous-to-liquid transition above 151(10) GPa rather than a second amorphous polymorph. Independent DFT+$_2$17 calculations of the Fe$_2$18O$_2$19 Hugoniot agree with SESAME 7440 and indicate temperatures below 2000 K up to 150 GPa, supporting the claim that the amorphous regime is not simply static melting. Upon release, a non-crystalline phase is observed alongside crystalline $_2$20-Fe$_2$21O$_2$22, and at 1.9 ns the extracted $_2$23 shows peaks at $_2$24 and $_2$25, with $_2$26 indicating Fe–O $_2$27 and Fe–Fe $_2$28 and a smeared coordination shell beyond 3 Å (Crépisson et al., 2024).

In titanium, the amorphous state is reported as a transient stage of shock-front dynamics rather than a retained final layer. A chromium-forsterite femtosecond laser at $_2$29 nm and $_2$30 fs, with pulse energy on target $_2$31 mJ and $_2$32 at beam center, launches a shock wave with front speed $_2$33. Experimentally, two modified subsurface layers form. A top surface polycrystalline layer of nanoscale grains is formed from a shock-molten layer via rapid crystallization, and in a deeper subsurface layer recrystallization of plastically deformed titanium changes the grain size. Molecular dynamics reveals that after $_2$34 ps, the shock front continues to melt or liquefy even when $_2$35: the HCP lattice collapses or amorphizes under high shear stress, producing a metastable supercooled liquid. Mechanical melting ceases when the maximum shear stress falls below $_2$36 GPa, corresponding to $_2$37–$_2$38. The equilibrium shock melt ceases at $_2$39 with depth $_2$40 nm, whereas cold melt extends further to $_2$41–$_2$42 nm depending on orientation. The resulting layer B is $_2$43 thick in coarse-grained Ti, and the deeper layer C is also $_2$44 thick; below $_2$45–$_2$46 the shock drops below the Hugoniot elastic limit and the material no longer recrystallizes (Zhakhovsky et al., 2023).

Taken together, these studies show that laser-induced amorphization can be pressure-mediated even when the final observed microstructure is crystalline. In Fe$_2$47O$_2$48, the amorphous regime is directly retained over a finite pressure interval. In Ti, amorphization is a transient part of a shock-front pathway that terminates in nanocrystalline and recrystallized layers.

5. Theoretical descriptions and scaling laws

The theoretical descriptions span optical interference models, coupled carrier–lattice transport models, atomistic–continuum hybrids, transfer-matrix heat transport, and reciprocal-space analysis. For Ge HSFL, the reported Sipe–Drude–Kerr framework writes the dielectric response as

$_2$49

with

$_2$50

and

$_2$51

Using $_2$52 and $_2$53 fs, with $_2$54 for $_2$55, the efficacy factor $_2$56 predicts HSFL periods $_2$57 that agree with $_2$58–$_2$59 and yields optically excited electron densities $_2$60–$_2$61, where $_2$62 solves $_2$63 (Austin et al., 2016).

For Si at 1030 nm, the carrier dependent two-temperature model evolves the free-carrier density $_2$64, electron temperature $_2$65, and lattice temperature $_2$66 using coupled balance equations with generation by $_2$67 and $_2$68, loss by Auger recombination, carrier diffusion, electron-phonon coupling, and attenuation

$_2$69

The reported temperature-sensitive parameters are the plunge of $_2$70 at low temperature due to phonon freeze-out, the strong rise of $_2$71 below 50 K, and a more modest temperature dependence of $_2$72. The simulations reproduce the experimental trends and report that at 30 K the peak carrier density and $_2$73 are $_2$74 smaller than at 300 K, yet the post-pulse lattice quench rate is $_2$75 larger (Kuz et al., 19 May 2026).

For Si at 800 nm, the hybrid nTTM–MD model replaces the lattice equation by direct MD with the Stillinger–Weber potential while retaining continuum carrier transport. The free-carrier density equation includes one- and two-photon absorption, impact ionization, Auger recombination, and diffusion; the source term for a 100 fs Gaussian pulse is written as

$_2$76

This framework captures the kinetic distinction between heterogeneous and homogeneous melting and explicitly links retained amorphization to cooling rates above $_2$77 (Ivanov et al., 2024).

For Sb$_2$78S$_2$79, the optical problem is treated by the transfer-matrix method and the thermal problem by a one-dimensional transient heat equation,

$_2$80

with

$_2$81

Latent heat of fusion $_2$82 is incorporated through an enhanced $_2$83 near the melting point, and the optical plus thermal sequence is rerun whenever a 1 nm slice first exceeds $_2$84 (Laprais et al., 2024).

For shock-amorphized Fe$_2$85O$_2$86, the structure is characterized through total structure factors and pair distributions:

$_2$87

and

$_2$88

The Hugoniot path is constrained with the Rankine–Hugoniot relations

$_2$89

combined with SESAME 7440 tables and DFT+$_2$90 (Crépisson et al., 2024). For Ti, the reported two-temperature hydrodynamics and MD resolve how shock attenuation moves the system from equilibrium melting to cold mechanical melting and then to plasticity-limited recrystallization, with the stop criterion $_2$91 GPa (Zhakhovsky et al., 2023).

6. Control parameters, recurring constraints, and misconceptions

A first recurring constraint is that amorphization is not synonymous with any visible laser damage. In Ge, below $_2$92 the reported morphology is “no LSFL, only HSFL and shallow amorphous caps,” so the amorphous layer is coupled to a specific sub-LSFL regime rather than to gross ablation. In cryogenic Si, the amorphous ring is explicitly distinct from the central ablation crater, where LIPSS and resolidified melt appear. In Sb$_2$93S$_2$94, the relevant operating space is a reversible window bracketed by amorphization and damage thresholds, and crossing the upper boundary causes ablation or irreversible swelling rather than better amorphization (Austin et al., 2016, Kuz et al., 19 May 2026, Laprais et al., 2024).

A second recurring constraint is that higher deposited energy does not automatically produce deeper or more useful amorphization. In Sb$_2$95S$_2$96, full-depth amorphization above $_2$97 nm is blocked by the optical absorption profile, not by the quoted thermal thickness estimate, and for a 93 nm film the bottom $_2$98 nm remain below $_2$99 at 532 nm. In Si, the hybrid calculation shows that stand-alone continuum models underestimate the melting threshold and overestimate melt depth below $_3$00 because they omit latent-heat and interface-motion kinetics. In Ge, the observed amorphous cap thickness remains $_3$01 even though the TPA-based estimate of deposited energy extends to $_3$02, which is consistent with the reported discussion that defect accumulation may also contribute (Laprais et al., 2024, Ivanov et al., 2024, Austin et al., 2016).

A third misconception is to treat amorphization as necessarily thermal melting at ambient-pressure criteria. The Fe$_3$03O$_3$04 study reports a shock-induced amorphous regime between 122(3) and 145(10) GPa while DFT+$_3$05 and SESAME indicate temperatures below 2000 K up to 150 GPa. The Ti study reports continued collapse/amorphization of the lattice when $_3$06, provided the shock-front shear stress exceeds $_3$07 GPa. These results indicate that a non-crystalline state can be induced by high-pressure kinetics and shear even when conventional static melting expectations are not met (Crépisson et al., 2024, Zhakhovsky et al., 2023).

The control parameters that recur across the literature are wavelength, fluence, pulse duration, crystal orientation, grain orientation, substrate temperature, and pressure history. In Ge, the HSFL period scales approximately linearly with $_3$08 but deviates from $_3$09 because Drude–Kerr excitation modifies $_3$10. In cryogenic Si, low temperature suppresses 1030 nm single-photon absorption, enhances thermal conductivity below $_3$11 K by a factor of 3–4, and broadens the amorphous window. In 800 nm Si, the difference between $_3$12 and $_3$13 affects melt depth. In Sb$_3$14S$_3$15, grain-to-grain anisotropy produces a $_3$16 threshold spread. In Fe$_3$17O$_3$18, the decisive variables are the peak pressure and subsequent release. In Ti, the decisive quantity is the attenuation of the shock front to the point where $_3$19 drops below the cold-mechanical-melting criterion (Austin et al., 2016, Kuz et al., 19 May 2026, Ivanov et al., 2024, Laprais et al., 2024, Crépisson et al., 2024, Zhakhovsky et al., 2023).

The reported applications follow directly from the retained or transient character of the non-crystalline state. In Si, the cited uses are buried waveguides and refractive-index-contrast features in integrated photonics, localized amorphization for dopant activation or selective etching in microelectronics, and surface texturing and passivation in thin-film and crystalline-Si solar cells. In Sb$_3$20S$_3$21, reversible switching and partial recrystallization are relevant to nanophotonic tuning and multilevel patterning. In Ti, the shock-induced pathway is connected to surface hardening through nanocrystalline and recrystallized subsurface layers. A plausible implication is that the practical value of laser-induced amorphization depends less on the mere creation of disorder than on whether the disorder can be retained, spatially confined, and cycled without crossing into ablation or uncontrolled recrystallization (Kuz et al., 19 May 2026, Laprais et al., 2024, Zhakhovsky et al., 2023).