Pith. sign in

REVIEW 4 major objections 5 minor 33 references

Self-Focusing Control for Depth-Precise Wafer Slicing of 4H-SiC in Femtosecond Laser Processing

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper shows that the first Kerr self-focusing collapse of a femtosecond laser inside a 4H-SiC wafer sets the slicing quality, and that a processability map built on one critical irradiance can land that collapse in a window with separa

desk verdict A solid, honest process-mapping paper whose central engineering claim holds up; the single-surface-Icr normalization and the Fig. 6 validation protocol are the things to probe before trusting the map beyond this setup. read the letter →

arxiv 2608.03814 v1 pith:L44AHTXZ submitted 2026-08-04 physics.optics cond-mat.mtrl-sciphysics.app-ph

Dong Hee Kang , Jaeseung Lim , Mishfaqur Rahman , Seongheum Han , Jae-Hak Lee , Seungman Kim , Jihoon Jeong This is my paper · ORCID
classification physics.opticscond-mat.mtrl-sciphysics.app-ph PACS 42.65.Jx
keywords 4H-SiCfemtosecondlaserslicingKerrself-focusingMarburgerformularayopticssimulationprocessabilitymapseparationstresswaferthinning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

4H-SiC wafers must be thinned to below 100 µm for next-generation power electronics, and femtosecond lasers can slice them without mechanical contact, but Kerr self-focusing makes the cut depth hard to control. The paper's central claim is that the first nonlinear collapse of the beam—not the nominal focus—determines slicing quality, and that its location is governed by the interplay of pulse energy and geometric focal depth. To predict that location, the paper extends the Marburger self-focusing formula to focused beams by replacing the power ratio with an irradiance ratio normalized to the measured surface damage threshold. It then defines a processability map bounded by a modification onset threshold, a self-focusing threshold, and a surface damage threshold, and shows that conditions inside the effective zone separate with stress typically below 10 MPa and comparatively smooth, valley-type surfaces. If the claim holds, a fixed optical system can select pulse energy and focus depth from the map rather than by trial and error.

What carries the argument

The central object is the modified Marburger self-focusing formula extended to focused beams, z_sf/z_f = 0.367 / [(sqrt(I_in/I_cr) − 0.852)² − 0.0219], with I_cr taken as the surface damage threshold. It converts two controllable inputs—pulse energy (through the on-axis irradiance at the would-be focus) and geometric focal depth—into the location of the first nonlinear collapse, and the same normalized irradiance forms the background of the processability map. The map's three thresholds—modification onset threshold (MOT), self-focusing threshold (SFT), and surface damage threshold (SDT)—define an effective processing zone. A ray-optics simulation with a Kerr refractive index supplies the col

What would settle it

Map the bulk modification threshold by fixing the geometric focal depth at several values (e.g., 100, 200, 300, 400 µm) and stepping pulse energy upward until a first internal modification track appears in cross-section. If the offset between that bulk threshold and the surface value of 5.88 kW/µm² changes appreciably across the effective zone—or if samples held at the same I_in/I_cr but different absolute depths show different separation stress and roughness—the map's boundaries will not sort the groups and the universal-normalization claim fails.

Watch

Extended reading notes

Core claim

The paper claims that in femtosecond-laser wafer slicing of 4H-SiC, the event that sets the quality of the cut is the first Kerr-induced self-focusing collapse, which occurs upstream of the geometric focus, and that its depth can be predicted by a modified Marburger relation. The modification replaces the original power ratio with a normalized irradiance I_in/I_cr, using the experimentally measured surface damage irradiance I_cr = 5.88 kW/µm² as the normalizing constant. On this basis the paper constructs a processability map in pulse-energy versus self-focusing-depth space, with three physically defined boundaries: modification onset at I_in/I_cr ≈ 1.9, self-focusing threshold when the geom

Load-bearing premise

The load-bearing premise is that one number measured at the top surface—the irradiance at which the laser just starts to damage the surface—can stand in for the nonlinear collapse threshold at all depths inside the wafer, even though the bulk threshold is higher and depth-dependent.

Editorial extensions

If this is right

  • Laser recipes can be chosen from the map instead of trial-and-error: set pulse energy and geometric focus so that I_in/I_cr lies inside the effective zone, and the first collapse lands at the desired depth.
  • Separation quality becomes a predictable function of beam optics: conditions inside the effective zone produce separation stress below about 10 MPa and low areal roughness, while near-inactive and near-damaged zones produce non-separable or ablated surfaces.
  • The modification width-to-pitch ratio D/W near 0.95 is the practical tuning point: at lower ratios separation stress rises, and once overlaps reach this ratio stress saturates below 5 MPa.
  • The above-threshold depth span L_th predicted from the intensity field explains why deeper self-focusing gives rougher surfaces and more kerf loss, so the map doubles as a warning about planarization cost.
  • Only the first self-focusing collapse is the effective processing point; downstream multifocal peaks seen in simulation are not seen in experiments, so recipes should target the first collapse rather than deeper refocusing events.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the map is normalized by a surface threshold that is known to be lower than the bulk modification threshold, the EPZ boundaries are likely system-dependent: changing the objective NA, wavelength, scan pitch, or surface finish would shift the effective MOT and SDT even if the Kerr collapse physics is unchanged—the paper acknowledges MOT is not universal, and this inference extends that caut
  • A direct test of the paper's core link would be to co-vary geometric focus and pulse energy to hold I_in/I_cr constant at different absolute depths; if the bulk-threshold offset changes with depth, separation stress and roughness should deviate from the map's prediction at depth.
  • The collapse-depth prediction could be combined with in-line monitoring of the modification depth after a first pass to close the loop: adjust the geometric focus on the fly to keep the collapse inside the effective zone, which is the control implication the conclusion gestures toward but does not implement.
  • The same normalized-irradiance reasoning should transfer to other Kerr-positive transparent crystals sliced by focused femtosecond pulses, with I_cr recalibrated per material—an untested extension suggested by the structure of the model.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The manuscript investigates Kerr self-focusing in femtosecond laser slicing of 4H-SiC and proposes that the interplay between pulse energy and processing depth controls the first self-focusing collapse, which in turn correlates with separation stress and surface texture. The key modeling elements are a modified Marburger equation (Eq. 3) that replaces the power ratio with an irradiance ratio normalized by a surface-measured critical irradiance Icr, an energy-conserving ray-optics simulation, and a processability map (Fig. 8) whose zone boundaries are the modification onset threshold (MOT), self-focusing threshold (SFT), and surface damage threshold (SDT). Experiments at fixed optical conditions with varying pulse energy and geometric focal depth support a three-group classification of slicing quality, and quantitative correlations are reported between D/W ratio, separation stress, and roughness.

Significance. If the proposed modified Marburger model and processability map are correct, the paper would provide a practical, physically motivated route to selecting laser parameters for low-stress, low-roughness SiC wafer slicing without trial-and-error. The work is commendable for combining systematic experiments, a semi-empirical analytical model, and a simulation that explicitly conserves energy and matches the qualitative upstream shift of the collapse. The decoupling experiments for pulse overlap and inter-pulse spacing are careful, and the clustering analysis adds quantitative support to the qualitative map. The central limitation is that the adversarial normalization constant Icr is taken from the surface ablation threshold and used as a bulk nonlinearity axis, while the paper itself states the bulk modification threshold is depth-dependent. Since Eq. (3) and the processability map both depend on this axis, the transferability of the quantitative boundaries is not established. The validation in Fig. 6 also uses model-predicted collapse positions to define the horizontal coordinates, weakening the independence of the claimed agreement. These issues are load-bearing but appear addressabl

major comments (4)
  1. [§3.1 and Eq. (3)] The normalized irradiance Iin/Icr is used in Eq. (3) and as the colormap background of the processability map (Fig. 8), but Icr is the surface damage threshold measured at zf=0. The text explicitly states that the bulk modification threshold is higher and varies significantly with depth, propagation distance, and nonlinear losses. If the offset between surface and bulk thresholds is not constant, the x-axis of Fig. 6 and the MOT/SDT boundaries in Fig. 8 are not universal. The paper needs a quantitative test: e.g., measure modification onset at several depths and show that Iin/Icr at onset is constant, or provide a depth-dependent calibration and show that the map's groupings are preserved.
  2. [§3.2, Fig. 6] The validation of Eq. (3) is partly circular. The text states that for each experimental point Iin is calculated at the model-predicted collapse position using w_lin(zsf). Since zsf is itself the solution of Eq. (3), the horizontal coordinate of each experimental point is constructed from the model being tested. The experimental y-coordinate remains independent, but the apparent agreement in x is not an independent test. Recomputing Iin/Icr at the measured zm (or at the geometric focus) would provide a stronger validation.
  3. [Appendix B / Eq. (3)] The replacement of the Marburger power ratio Pin/Pcr by Iin/Icr, while retaining the numerical constants 0.367, 0.852, and 0.0219, is an ad hoc substitution. The original Marburger formula's constants were derived for a collimated beam and a power ratio tied to the Kerr critical power; using an irradiance ratio with an unrelated normalization changes the physical threshold condition. The paper's defense that this 'preserves the correct monotonic dependence' is not sufficient to guarantee quantitative validity, especially because Iin itself is depth-dependent through w_lin(z). The authors should either derive the appropriate focused-beam scaling or explicitly treat the retained constants as fitted parameters with uncertainty.
  4. [§3.3, Fig. 7] The ray optics simulation provides useful qualitative support for an upstream collapse and for the slope of zsf versus zf, but it does not test the Icr normalization or the quantitative form of Eq. (3). The simulation uses n_Kerr and a Gaussian seed, not the surface Icr. This is an independent check of the self-focusing trend, not of the processability-map normalization. The manuscript should clarify this distinction when discussing agreement among model, simulation, and experiment.
minor comments (5)
  1. [Throughout] Some notation is inconsistent: z_f is called 'geometric focal depth' but Eq. (3) uses z_f as both a length and (in z_sf/z_f) a dimensionless scale; the definitions in Appendix E introduce z_m,p with and without an overbar, which is easy to confuse. A notation table would help.
  2. [Fig. 6 caption] The caption states 'all symbols indicate experimental results, showing the normalized modification depth (z_m/z_f) versus normalized irradiance,' but the text says Iin is computed at the model-predicted zsf. The caption should state this model-dependence explicitly.
  3. [§3.4, Table 1] The kurtosis values for group iii are very large (e.g., Sku=26.87) compared with groups i and ii. A brief comment on the statistical robustness of these extreme values and their dependence on the chosen region of interest would strengthen the interpretation.
  4. [Appendix D] The spherical aberration estimate gives d_LA=13.9 µm, which is 'a few micrometers' only in the sense of being much smaller than the observed hundreds-of-micrometer shifts. The wording 'only a few' is slightly misleading; consider saying 'an order of magnitude smaller than the observed shifts.'
  5. [References] Reference [19] contains a typo in the URL ('ttps://'). Also, the text cites [19] for CW laser-assisted splitting and later for microvoid formation; the two uses should be checked.

Circularity Check

1 steps flagged · score 2.0 of 10

One partial construction in the Fig. 6 validation: the experimental x-axis Iin/Icr is evaluated at the model-predicted collapse depth, so the plotted model curve and the data share the model's own output as input; the central processability map otherwise rests on independent stress/roughness measurements and simulation.

  1. self definitional [Section 3.2, near Fig. 6]
    "For each experimental point, I_in = P_in/[π w_lin(z_sf)^2] was calculated at the model-predicted collapse position using the pre-collapse radius w_lin(z) to avoid a circular dependence between the I_in and z_sf."

    Eq. (3) is advertised as predicting zsf/zf from Iin/Icr, but the experimental abscissa Iin/Icr is not an independent measured input: it is evaluated at the model-predicted zsf through wlin(zsf). Thus the x-coordinate of every data point already contains the model's own prediction, and the analytical curve is drawn through those same model-chosen x-values. The comparison therefore reduces to plotting measured zm against the model's zsf at an x-location selected by the model; any error in zsf is partly absorbed into the horizontal coordinate rather than appearing as an independent prediction error. This is a self-definitional construction: the 'input' irradiance ratio is defined in terms of the 'output' collapse depth. Fig. 6 is therefore not an out-of-sample test of Eq. (3), and the display

full rationale

The central derivation is mostly self-contained. Eq. (3) is an explicit extension of the external Marburger formula; the numerical constants are retained, not refit, and Icr = 5.88 kW/µm2 is measured from an independent surface-ablation threshold experiment rather than fitted to collapse-depth data. The processability map's SDT boundary is experimental, SFT is geometric, and MOT is an explicitly empirical cutoff; none of these boundaries is fitted to the separation-stress or roughness data used for quality grouping. The stress/Sa clustering in Figs. 8 and 14 is independent of Eq. (3) and would stand even if the analytical model were removed. Section 3.1's concession that the bulk modification threshold is depth-dependent and higher than the surface Icr is a correctness/transferability risk, not a circularity: the paper explicitly limits Icr to an internal reference for fixed optical conditions, and the ray-optics simulation (using only n0 and nKerr) provides partially independent support that does not rely on the Icr normalization. The only self-citation, Ref. [26], concerns orientation-dependent fracture morphology and is not load-bearing for the main claim. The single genuine circularity is the Fig. 6 validation protocol, where the model's predicted collapse depth is used to construct the experimental x-coordinate; this weakens one validation figure but does not force the central processability conclusions, so the score is 2 rather than higher.

Assumptions & free parameters 5 free parameters · 7 assumptions · 0 invented entities

No new physical entities are postulated; the 'effective processing zone' and 'processability map' are empirical constructs, not mechanisms. The analytical model and map rest on a measured normalization (Icr) and on Marburger constants inherited from collimated-beam theory, plus a set of domain assumptions about neglecting plasma, absorption, and aberration effects.

free parameters (5)
  • Icr (surface damage threshold) = 5.88 kW/µm²
    Measured from the ablation boundary in Fig. 3b; used as the normalization in eq. (3) and as the colormap background for the processability map.
  • MOT threshold (Iin/Icr) = 1.9
    Minimum normalized irradiance that produced separable samples in the fixed-window experiments; defines the EPZ lower boundary in Fig. 8.
  • D/W threshold = 0.95
    Transition value for piecewise linear fits of separation stress vs normalized modification width in Fig. 13; used to define the optimal balance.
  • Ray reconstruction launch width sigma_w = 10 µm
    Hand-chosen Gaussian weight width for ray energy assignment in Appendix C.3; not derived from physics.
  • Affected-region intensity cutoff = 0.05 Imax
    Arbitrary superlevel threshold used to define the hotspot energy fraction in Appendix C.4.
assumptions (7)
  • domain assumption Kerr refractive index model n = n0 + nKerr I with n0=2.55 and nKerr=3.72e-19 m²/W at 1040 nm
    Invoked in Appendix A (eq. A.1) and the simulation (eq. C.1); the nKerr value is taken from ref [32].
  • ad hoc to paper Marburger constants are transferable to the focused-beam irradiance-ratio form
    Appendix B retains 0.367, 0.852, 0.0219 after replacing power ratio with irradiance ratio and z_R with z_f, justified only by agreement with the same experiments and simulation used for validation.
  • domain assumption Primary modification depth zm equals the first self-focusing collapse depth zsf
    Used throughout (e.g., Fig. 6, eq. 3); microvoid positions are read as the location of the first Kerr collapse despite the model neglecting plasma and absorption.
  • ad hoc to paper Surface damage threshold Icr serves as a universal normalization for bulk self-focusing
    Section 3.1 states the bulk threshold is higher and depth-dependent, yet Icr measured at the surface is used as the single normalization in eq. (3) and the map.
  • domain assumption Ray optics without plasma or absorption captures the first collapse location
    Section 2.5/3.3 state absorption and plasma hydrodynamics are neglected; the simulation is used to extract zsf and Lth, which feed the processability discussion.
  • domain assumption Beam is a paraxial Gaussian with M2 nearly 1
    Measured M2_x=1.09, M2_y=1.04; used to justify Gaussian beam radius and simulation seeding.
  • domain assumption Spherical aberration is a small static background
    Appendix D estimates d_LA=13.9 µm, concluded to be negligible compared to hundred-micron Kerr shifts; this isolates the Kerr effect but depends on the geometrical optics formula of ref [33].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-Focusing Control for Depth-Precise Wafer Slicing of 4H-SiC in Femtosecond Laser Processing." pith.science (2026). https://pith.science/paper/L44AHTXZ

@misc{pith2026260803814,
  author       = {Pith},
  title        = {Pith review of: Self-Focusing Control for Depth-Precise Wafer Slicing of 4H-SiC in Femtosecond Laser Processing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L44AHTXZ}},
  note         = {Machine review of arXiv:2608.03814}
}
abstract

4H-SiC has emerged as a third-generation chip material because its superior thermal conductivity and high breakdown field enable the material to achieve high power density and higher switching frequencies in power-electronics applications. As chip architectures evolve toward 3D and heterogeneous integration, the mechanical and thermal design space tightens while yield risks grow. In particular, advanced packages require mid-process wafer thinning to < 100 $\mu$m to shorten interconnects and control thermo-mechanical stress. Femtosecond laser slicing for 4H-SiC wafers offers a non-contact processing approach to produce thin layers with low defects, while strong optical nonlinearities obscure the relationship between the laser parameters and the resulting slicing quality. Here, we systematically investigate Kerr-induced self-focusing using a femtosecond laser in 4H-SiC slicing by combining experiments, a semi-empirical analytical model, and numerical ray optics simulations. We demonstrate that the interplay between pulse energy and processing depth governs the self-focusing behavior, which directly correlates with post-separation surface texture parameters and separation stress, thereby linking nonlinear beam propagation to slicing quality. Based on this relationship, we define a processability map in the pulse energy with self-focusing depth space over a normalized irradiance background. Analytically, the model extends the Marburger formula to focused beams by replacing the power ratio with a normalized irradiance. Ray optics simulations capture the geometric features at the self-focusing point and are validated against experimental observations. Within physically defined thresholds, the processability map directly connects laser parameters to separation stress and surface texture metrics, providing practical guidance for depth control beyond trial-and-error.

Figures

Figures reproduced from arXiv: 2608.03814 by the authors.

Figure 1
Figure 1. (a) Schematics of femtosecond laser slicing in 4H-SiC and (b) the layer separation. (c) Irradiance profile by incident beam power (Pin) along the depth (z) of the wafer from the analytical model. (d) Schematic of beam path with the Kerr-induced self-focusing and plasma generation at the self-focusing depth (zsf), upstream of the geometric focal depth (zf). (e) Beam profile from ray optics simulation model to evaluat… view at source ↗
Figure 2
Figure 2. Conceptual irradiance profile of the focused femtosecond laser beam. The profile depicts a continuous beam propagating without interaction with the medium, illustrating the beam geometry and focusing behavior, despite the actual femtosecond pulse having a very short propagation length. (a) Irradiance distributions along the propagation to depthwise for various pulse energies with the focal point positioned at the ce… view at source ↗
Figure 3
Figure 3. (a) Images of the laser beam-ablated regions on the SiC wafer surface, captured while varying the pulse energy, decreasing the distance between the objective lens and the wafer surface (z) in 10 µm steps from the focal position. (b) Colormap showing the irradiance at the wafer surface as a function of incident pulse energy (Ein) and geometric focal depth (zf). Symbols indicate the investigated regions: circles (o) r… view at source ↗
Figures from the paper (12 more)
Figure 4
Figure 4. Figure 4: Effect of pulse overlap on modification region in self-focusing-driven 4H-SiC slicing. The average laser power (0.6 W) and repetition rate (200 kHz) were fixed, while only the scanning speed was varied: (a) 150 mm/s, (b) 75 mm/s, and (c) 37.5 mm/s. The corresponding ma…
Figure 5
Figure 5. Figure 5: Effect of inter-pulse temporal spacing at constant pulse energy on the modifica￾tion region in self-focusing-driven 4H-SiC slicing. The pulse energy was fixed at 3 µJ by proportionally scaling the average power, repetition rate, and scanning speed: (a) 0.6 W, 200 kHz, …
Figure 6
Figure 6. Figure 6: Relationship between the self-focusing and the geometrical focus as a function of normalized irradiance (Iin/Icr). The solid line represents the modified Marburger equation (the formula indicated inside the graph), and all symbols indicate experimental results, showing…
Figure 7
Figure 7. Figure 7: Results of the ray optics simulation of a femtosecond laser beam path in the 4H-SiC. Simulation results of (a) the beam intensity profile without the Kerr effect, (b) the refractive index profile with the Kerr effect, and (c) the beam intensity profile with Kerr effect…
Figure 8
Figure 8. Figure 8: Processability map for laser slicing with the three theoretical thresholds. The colormap represents the normalized irradiance at the self-focusing depth. Optical micro￾scope images at both low and high magnification show the surface texture of the sliced wafer under th…
Figure 9
Figure 9. Figure 9: Optical microscopic images of the sliced 4H-SiC wafer surface under three rep￾resentative processing conditions: (a) Ein = 3 µJ and zsf ≈ 100 µm, (b) Ein = 6 µJ and zsf ≈ 100 µm, (c) Ein = 6 µJ and zsf ≈ 200 µm. For each condition, the sub-figures for line-averaged sur…
Figure 10
Figure 10. Figure 10: (a,f) Optical micrographs of the sliced 4H-SiC wafer surfaces processed at Ein = 3 µJ and 6 µJ, respectively, with comparable self-focusing depths (zsf ≈ 200 µm) in both cases. (b,g) SEM images taken from the regions indicated in (a) and (f), highlighting the surface …
Figure 11
Figure 11. Figure 11: Raman spectroscopy analysis of the 4H-SiC crystalline-structure decomposition induced by different pulse energies along the laser beam path. (a) Optical microscopic image of the bare 4H-SiC wafer. Modification layer images of samples processed at (b) 4 µJ and (c) 6 µJ…
Figure 12
Figure 12. Figure 12: (a) Relationship between the modification depth (zm) and the modification width (D). Data points show the mean with error bars denoting ±1 SD error bars on both axes from the mean value (n=5). Solid lines represent the calculated beam diam￾eter (Df) at the self-focusi…
Figure 13
Figure 13. Figure 13: (a) Relationship between normalized modification width (D/W) and wafer sep￾aration stress (σ). Mean values with ±1 SD error bars on both axes from the mean value (n=5). All data points represent conditions where separation was performed within the effective processing…
Figure 14
Figure 14. Figure 14: Correlation between the separation stress (σ) and the areal-average surface roughness (Sa) with three clusters overlaid on all separable samples. The color of each symbol represents the normalized irradiance (Iin/Icr). Solid curves denote 95% confidence ellipses that …
Figure 15
Figure 15. Figure 15: (a) Surface roughness at the region of interest of a separated surface (i), and its 2D and 3D surface profiles (ii, iii). (b) Areal-averaged surface roughness at the modification depth by the magnitude of pulse energy. Mean values with ±1 SD error bars on both axes fr…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 31 canonical work pages

  1. [1]

    Sahu, S.T

    S.K. Sahu, S.T. Chacko, O. Parmar, A.S. Rajput, Wide bandgap Materials: Revitalizing Power Electronics with Advances in Power Semiconductor Devices, J. Inst. Eng. (India): B 106 (2025) 1-17, https://doi.org/10.1007/s40031-025-01205-4

    work page doi:10.1007/s40031-025-01205-4 2025

  2. [2]

    Elliott, M.T

    J.W. Elliott, M.T. Lebon, A.J. Robinson, Optimising integrated heat spreaders with distributed heat transfer coefficients: A case study for CPU cooling, Case Stud. Therm. Eng. 38 (2022) 102354, https://doi.org/10.1016/j.csite.2022.102354

    work page arXiv 2022

  3. [3]

    Bar-Cohen, J.J

    A. Bar-Cohen, J.J. Maurer, D.H. Altman, Embedded cooling for wide bandgap power amplifiers: A review, J. Electron. Packag. 141(4) (2019) 040803, https://doi.org/10.1115/1.4043404

    work page doi:10.1115/1.4043404 2019

  4. [4]

    G. Liu, Y. Wu, K. Li, Y. Wang, C. Li, Development of high power SiC devices for rail traction power systems, J. Cryst. Growth 507 (2019) 442-452, https://doi.org/10.1016/j.jcrysgro.2018.10.037

    work page doi:10.1016/j.jcrysgro.2018.10.037 2019

  5. [5]

    Z. Mo, F. Wang, J. Li, Q. Liu, G. Zhang, W. Li, C. Yang, R. Sun, Temporary bonding and debonding in advanced packag- ing: recent progress and applications, Electronics 12(7) (2023) 1666, https://doi.org/10.3390/electronics12071666

    work page doi:10.3390/electronics12071666 2023

  6. [6]

    Domke, B

    M. Domke, B. Egle, S. Stroj, M. Bodea, E. Schwarz, G. Fasching, Ultrafast-laser dicing of thin silicon wafers: strategies to improve front- and backside breaking strength, Appl. Phys. A 123(12) (2017) 746, https://doi.org/10.1007/s00339-017-1374-7

    work page doi:10.1007/s00339-017-1374-7 2017

  7. [7]

    Cheng, Y

    D. Cheng, Y. Gao, R. Liu, Finite element analysis on processing stress of polysilicon cut by diamond multi-wire saw, Mater. Sci. Semicond. Pro- cess 131 (2021) 105860, https://doi.org/10.1016/j.mssp.2021.105860

    work page arXiv 2021

  8. [8]

    J. Qiu, X. Li, R. Ge, C. Liu, Surface formation, morphol- ogy, integrity and wire marks in diamond wire slicing of mono- crystalline silicon in the photovoltaic industry, Wear 488 (2022) 204186, https://doi.org/10.1016/j.wear.2021.204186

    work page arXiv 2022

Show all 33 references
  1. [9]

    E.Kim, Y.Shimotsuma, M.Sakakura, K.Miura, 4H-SiCwaferslicingby using femtosecond laser double-pulses, Opt. Mater. Express 7(7) (2017) 2450-2460, https://doi.org/10.1364/OME.7.002450. 44

    2017 doi

  2. [10]

    Couairon, A

    A. Couairon, A. Mysyrowicz, Femtosecond filamentation in transparent media, Phys. Rep. 441(2-4) (2007) 47-189, https://doi.org/10.1016/j.physrep.2006.12.005

    2007 doi

  3. [11]

    X. Xie, T. Liu, K. Lv, Y. Huang, Y. Huang, W. Hu, J. Long, The internal modified layer structure of silicon carbide induced by ultrafast laser and its application in stealth dicing, Precis. Eng. 88 (2024) 895-904, https://doi.org/10.1016/j.precisioneng.2024.05.006

    2024 doi

  4. [12]

    L. Wang, C. Zhang, F. Liu, H. Zheng, G.J. Cheng, Process mech- anism of ultrafast laser multi-focal-scribing for ultrafine and efficient stealth dicing of SiC wafers, Appl. Phys. A 128(10) (2022) 872, https://doi.org/10.1007/s00339-022-06012-y

    2022 doi

  5. [13]

    L. Wang, C. Zhang, F. Liu, H. Zheng, G.J. Cheng, Ultra- fast pulsed laser stealth dicing of 4H-SiC wafer: Structure evolu- tion and defect generation, J. Manuf. Process. 81 (2022) 562-570, https://doi.org/10.1016/j.jmapro.2022.06.064

    2022 doi

  6. [14]

    Zhang, X

    Y. Zhang, X. Xie, Y. Huang, W. Hu, J. Long, Internal modified structure of silicon carbide prepared by ultrafast laser for wafer slicing, Ceram. Int. 49(3) (2023) 5249-5260, https://doi.org/10.1016/j.ceramint.2022.10.043

    2023 doi

  7. [15]

    Y. Song, S. Zhao, H. He, H. Liang, Z. Dai, X. Lin, G. Zhang, Numerical simulation and experimental investigation of ps pulsed laser modification inside 4H-SiC material, Photonics 11(2) (2024) 189, https://doi.org/10.3390/photonics11020189

    2024 doi

  8. [16]

    S. Han, H. Yu, C. He, S. Zhao, C. Ning, L. Jiang, X. Lin, Laser slic- ing of 4H-SiC wafers based on picosecond laser-induced micro-explosion via multiphoton processes, Opt. Laser Technol. 154 (2022) 108323, https://doi.org/10.1016/j.optlastec.2022.108323

    2022

  9. [17]

    Y. Yao, Q. Chen, B. Li, J. Zhang, R. Wang, M. Bai, R. Liang, L. Ma, T. Ma, J. Jang, J. Xia, H. Nie, B. Zhang, Influence of crystal orientation and incident plane on n-type 4H-SiC wafer slic- ing by using picosecond laser, Opt. Laser Technol. 182 (2025) 112174, https://doi.org/...

    2025

  10. [18]

    Xiangfu, H

    L. Xiangfu, H. Minghui, Micro-cracks generation and growth ma- nipulation by all-laser processing for low kerf-loss and high sur- face quality SiC slicing, Opt. Express 32(22) (2024) 38758-38767, https://doi.org/10.1364/OE.540604

    2024 doi

  11. [19]

    Jiang, S

    L. Jiang, S. Zhao, S. Han, H. Liang, J. Du, H. Yu, X. Lin, CW laser-assisted splitting of SiC wafer based on modified layer by picosecond laser, Opt. Laser Technol. 174 (2024) 110700, ttps://doi.org/10.1016/j.optlastec.2024.110700

    2024

  12. [20]

    Y. Li, Z. Zhang, Q. Song, H. Shi, Y. Hou, S. Yue, R. Wang, S. Cai, Z. Zhang, Surface micromorphology and nanostructures evolution in hybrid laser processes of slicing and polishing sin- gle crystal 4H-SiC, J. Mater. Sci. Technol. 184 (2024) 235-244, https://doi.org/10.1016/j.j...

    2024 doi

  13. [21]

    W. Geng, Q. Shao, Y. Pei, L. Xu, C. Cui, X. Pi, D. Yang, R. Wang, Slicing of 4H-SiC wafers combining ultrafast laser irradiation and bandgap-selective photo-electrochemical exfoliation, Adv. Mater. Inter- faces 10(21) (2023) 2300200, https://doi.org/10.1002/admi.202300200

    2023 doi

  14. [22]

    Boyd, A.L

    R.W. Boyd, A.L. Gaeta, E. Giese, Nonlinear optics, In: G.W.F. Drake (Eds.), Springer Handbook of Atomic, Molecular, and Optical Physics, Springer, 2008, pp. 1097-1110

    2008

  15. [23]

    Bergé, S

    L. Bergé, S. Skupin, R. Nuter, J. Kasparian, J.P. Wolf, Ultra- short filaments of light in weakly ionized, optically transparent me- dia. Rep. Prog. Phys. 70(10) (2007) 1633, https://doi.org/10.1088/0034- 4885/70/10/R03

    2007 doi

  16. [24]

    Marburger, Self-focusing: theory, Prog

    J.H. Marburger, Self-focusing: theory, Prog. Quantum Electron. 4 (1975) 35-110, https://doi.org/10.1016/0079-6727(75)90003-8

    1975 doi

  17. [25]

    Yuan, J.Y

    X.R. Yuan, J.Y. Deng, D.H. Xu, X. Sun, Y. Yu, Q. Chen, Fabrication of microlens arrays with high numerical aperture based on femtosecond laser self-modulating processing method. Acta Opt. Sin. 43(16) (2023) 1623019, https://doi.org/10.3788/AOS230896

    2023 doi

  18. [26]

    Youn, D.H

    H. Youn, D.H. Kang, J. Lim, S. Han, J.H. Lee, J. Jeong, S. Kim, Scanning direction dependence on crystal orientations of a femtosec- 46 ond laser-assisted 4H-SiC wafer slicing. Manuf. Lett. 44 (2025) 466-472, https://doi.org/10.1016/j.mfglet.2025.06.055

    2025 doi

  19. [27]

    Nakashima, T

    S.I. Nakashima, T. Mitani, M. Tomobe, T. Kato, H. Okumura, Raman characterizationofdamagedlayersof4H-SiCinducedbyscratching.AIP Adv. 6(1) (2016) 015207, https://doi.org/10.1063/1.4939985

    2016 doi

  20. [28]

    X. Liu, J. Luo, J. Ye, J. Fu, Q. Yang, Y. Wang, X. Chen, T. Tao, B. Liu, Q. Xu, H. Song, W. Xu, D. Zhou, F. Zhou, X. Xu, F. Reng, F. Xu, H. Lu, R. Zhang, Color centers and crystal structural transformations induced by femtosecond laser writing in 4H-SiC. J. Appl. Phys. 137(12)...

    2025 doi

  21. [29]

    J.Du, X.Lu, L.Jiang, S.Han, X.Li, H.Yu, S.Zhao, X.Lin, Suppressing kerf loss based on multi-focal approach for 4H-SiC laser slicing. Opt. Express 33(16) (2025) 34267-34280, https://doi.org/10.1364/OE.569632

    2025 doi

  22. [30]

    R. Amor, S. Mahajan, W.B. Amos, G. McConnell, Standing-wave- excited multiplanar fluorescence in a laser scanning microscope re- veals 3D information on red blood cells, Sci. Rep. 4(1) (2014) 7359, https://doi.org/10.1038/srep07359

    2014 doi

  23. [31]

    Lenhart, D

    J.L. Lenhart, D. Fischer, S. Sambasivan, E.K. Lin, W.L. Wu, D.J. Guer- rero, Y. Yang, R. Puligadda, Understanding deviations in lithographic patterns near interfaces: Characterization of bottom anti-reflective coat- ings (BARC) and the BARC–resist interface, Appl. Surf. Sci. 2...

    2007 doi

  24. [32]

    X. Guo, Z. Peng, P. Ding, L. Li, X. Chen, H. Wei, Z. Tong, L. Guo, Nonlinear optical properties of 6H-SiC and 4H-SiC in an ex- tensive spectral range, Opt. Mater. Express 11(4) (2021) 1080-1092, https://doi.org/10.1364/OME.415915

    2021 doi

  25. [33]

    Q. Sun, H. Jiang, Y. Liu, Y. Zhou, H. Yang, Q. Gong, Effect of spherical aberration on the propagation of a tightly focused femtosecond laser- pulse inside fused silica. J. Opt. A-Pure Appl. Opt. 7(11) (2005) 655, https://doi.org/10.1088/1464-4258/7/11/006. 47

    2005 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.