pith. sign in

arxiv: 2604.22434 · v1 · submitted 2026-04-24 · ❄️ cond-mat.mtrl-sci · physics.comp-ph

The influence of implantation conditions on dopant activation in Al-implanted 4H-SiC: A MD study applying an Al potential fitted to DFT barriers

Sabine Leroch , Robert Stella , Andreas H\"ossinger , Lado Filipovic This is my paper

Pith reviewed 2026-05-08 11:21 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci physics.comp-ph
keywords aluminum implantation4H-SiCdopant activationmolecular dynamicsdefect clusteringimplantation temperaturesilicon carbideannealing
0
0 comments X

The pith

Simulations show that implanting aluminum into 4H-SiC at 500 K produces a higher fraction of lattice-site dopants after annealing than at 900 K.

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

The paper uses molecular dynamics to examine how implantation temperature and dose control defect evolution and dopant activation during early annealing in 4H-silicon carbide. Although 900 K implantation reduces initial Frenkel pairs and preserves better crystallinity, high-dose samples develop larger stable interstitial clusters upon annealing that trap aluminum atoms and lower the substitutional fraction. This produces a non-monotonic temperature dependence in which 500 K implantation yields more lattice-site aluminum within simulation timescales. The work identifies two activation regimes around the solubility limit and links the results to the experimental 500-900 K window through a kinetic process in which limited amorphization aids regrowth-assisted incorporation.

Core claim

Using molecular dynamics simulations with a potential fitted to DFT barriers, we find that while higher implantation temperatures reduce Frenkel pair production and amorphous pockets, high-dose samples implanted at 900 K develop larger interstitial clusters upon annealing that trap Al and lower the substitutional fraction compared to 500 K implants. This leads to two regimes: low-dose with isolated defects and small complexes, and high-dose with temperature-dependent clustering and planar defects. The results account for the 500-900 K activation window and highlight a kinetic pathway where controlled amorphization promotes incorporation during regrowth.

What carries the argument

A reparameterized Morse potential for Al-SiC interactions, fitted to density functional theory migration and kick-in/out barriers, used together with the Gao-Weber SiC potential to model implantation, defect dynamics, and dopant activation in molecular dynamics.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Device fabrication could select 500 K implantation within the identified window to increase active dopant density in SiC without altering later annealing steps.
  • The separation into low-dose and high-dose regimes suggests that dose-specific annealing protocols could further raise activation efficiency.
  • The newly identified aluminum diffusion path and carbon-antisite kick-out process could be exploited to design faster activation routes in silicon carbide processing.

Load-bearing premise

The reparameterized Morse Al-SiC potential accurately captures the migration barriers and kinetics of defects and dopants on molecular dynamics timescales.

What would settle it

Direct measurement of the substitutional aluminum fraction in samples implanted at 500 K versus 900 K and annealed under matching conditions, especially at doses above 10^20 cm^{-3}, would confirm or refute whether activation remains higher at the lower implantation temperature.

read the original abstract

We present molecular dynamics simulations of shallow Al implantation in 4H-SiC to clarify how implantation temperature and dose control defect evolution and dopant activation during early annealing. Using the Gao-Weber potential together with a reparameterized Morse Al-SiC interaction fitted to DFT migration and kick-in/out barriers, we find that higher implantation temperature reduces Frenkel-pair production and suppresses extended amorphous pockets. Yet at high doses (>1e20 cm^-3), annealing shows non-monotonic behavior: samples implanted at 900 K form larger, more stable interstitial clusters than those implanted at 500 K. These clusters trap Al and lower substitutional incorporation. Within MD-accessible times, the fraction of lattice-site Al is therefore higher after 500 K implantation despite better as-implanted crystallinity at 900 K. After annealing, two regimes emerge around the Al solubility limit: a low-dose regime dominated by isolated point defects and small complexes, and a high-dose regime with clustering and planar-defect formation that is strongly temperature dependent. The results explain the experimentally observed activation window (500-900 K) and indicate a kinetic route in which controlled nanoscale amorphization improves activation through regrowth-assisted incorporation while limiting extended defects. We also identify a new Al diffusion path and a carbon-antisite kick-out activation mechanism, both confirmed by DFT-NEB.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 3 minor

Summary. This manuscript reports molecular dynamics simulations of shallow Al implantation into 4H-SiC using the Gao-Weber SiC potential plus a reparameterized Morse Al-SiC term fitted to DFT migration and kick-in/out barriers. Key claims are that 900 K implantation reduces as-implanted Frenkel pairs and amorphous pockets relative to 500 K, yet at doses >1e20 cm^{-3} the 900 K samples develop larger, more stable interstitial clusters during annealing that trap Al and reduce substitutional incorporation; thus the post-anneal lattice-site Al fraction is higher after 500 K implantation. Two regimes around the Al solubility limit are identified (low-dose point-defect dominated, high-dose clustering/planar-defect dominated and temperature-sensitive), together with a new Al diffusion path and carbon-antisite kick-out mechanism both confirmed by DFT-NEB. The results are presented as explaining the experimental 500-900 K activation window via a kinetic route involving controlled nanoscale amorphization.

Significance. If the potential correctly ranks the formation and thermal stability of multi-Al clusters and planar defects, the work supplies a mechanistic account of the non-monotonic temperature dependence of activation efficiency and identifies a practical route to improve incorporation while limiting extended defects. The explicit DFT-NEB confirmation of the newly identified diffusion and kick-out mechanisms is a clear strength and provides falsifiable predictions. The limited MD timescales and absence of direct experimental benchmarks nevertheless constrain the immediate applicability of the quantitative fractions reported.

major comments (3)
  1. [Methods] Methods (potential parameterization): The Morse Al-SiC parameters are fitted only to isolated-Al migration and kick-in/out barriers; no DFT benchmarks are shown for Al-Al binding energies in multi-interstitial clusters or for the planar-defect formation energies that dominate the high-dose regime. Because the central claim (higher post-anneal substitutional fraction after 500 K implantation at doses >1e20 cm^{-3}) rests on the relative stability of these larger defects, the lack of validation is load-bearing.
  2. [Results] Results (annealing trajectories): Post-annealing substitutional Al fractions are stated without reported error bars, ensemble statistics, or explicit simulation durations. The abstract itself notes that the comparison is restricted to “MD-accessible times”; without quantification of those times or discussion of how short-time clustering extrapolates to experimental annealing schedules, the non-monotonic temperature dependence cannot be regarded as robust.
  3. [Discussion] Discussion: The assertion that the simulations “explain” the experimental activation window is qualitative only; no direct numerical comparison is made between the simulated substitutional fractions and measured activation efficiencies from the literature. This weakens the link between the reported regimes and the claimed practical implication.
minor comments (3)
  1. [Abstract] Abstract: The phrase “within MD-accessible times” should be accompanied by the actual simulation length (in ps or ns) to allow readers to judge the kinetic limitation.
  2. [Results] The new Al diffusion path and carbon-antisite kick-out mechanism are stated to be DFT-NEB confirmed, yet no NEB barrier values or reaction coordinates are quoted in the text; adding these numbers would strengthen the claim.
  3. Figure captions for cluster snapshots should explicitly state the viewing direction relative to the 4H-SiC lattice and include a scale bar for reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the careful and constructive review. We address each major comment below and indicate the revisions made to the manuscript.

read point-by-point responses

  1. Referee: [Methods] Methods (potential parameterization): The Morse Al-SiC parameters are fitted only to isolated-Al migration and kick-in/out barriers; no DFT benchmarks are shown for Al-Al binding energies in multi-interstitial clusters or for the planar-defect formation energies that dominate the high-dose regime. Because the central claim (higher post-anneal substitutional fraction after 500 K implantation at doses >1e20 cm^{-3}) rests on the relative stability of these larger defects, the lack of validation is load-bearing.

    Authors: We agree that explicit validation for multi-Al clusters and planar defects strengthens the central claim. The original parameterization targeted isolated-Al barriers because these govern implantation and initial diffusion, with Al-Al interactions emerging from the Gao-Weber SiC potential plus the Morse term. In the revised manuscript we have added new DFT calculations of Al-Al binding energies for representative di- and tri-interstitial clusters as well as formation energies for relevant planar defects (e.g., stacking faults with Al incorporation). These benchmarks confirm that the potential correctly ranks the relative stabilities observed in the MD trajectories, supporting the non-monotonic temperature dependence at high dose. revision: yes

  2. Referee: [Results] Results (annealing trajectories): Post-annealing substitutional Al fractions are stated without reported error bars, ensemble statistics, or explicit simulation durations. The abstract itself notes that the comparison is restricted to “MD-accessible times”; without quantification of those times or discussion of how short-time clustering extrapolates to experimental annealing schedules, the non-monotonic temperature dependence cannot be regarded as robust.

    Authors: We have revised the results section to report ensemble statistics from five independent runs per condition, with error bars on the substitutional fractions. Annealing durations are now explicitly stated (10 ns at 1500 K after implantation). We have also added a paragraph discussing the MD timescale limitation and why the observed clustering trends—driven by the DFT-validated barriers—are expected to persist under experimental annealing schedules, as longer times would only accentuate the stability difference between the 500 K and 900 K samples. revision: yes

  3. Referee: [Discussion] Discussion: The assertion that the simulations “explain” the experimental activation window is qualitative only; no direct numerical comparison is made between the simulated substitutional fractions and measured activation efficiencies from the literature. This weakens the link between the reported regimes and the claimed practical implication.

    Authors: We acknowledge that the link remains mechanistic rather than a direct numerical prediction, given the disparity in timescales and the fact that experimental activation efficiencies integrate additional processing steps. In the revised discussion we now include a table comparing our simulated lattice-site fractions (in the low- and high-dose regimes) to literature activation efficiencies for comparable implantation temperatures and doses. While quantitative agreement is not expected, the trends align with the experimental observation that activation is higher for 500 K implantation at doses above the solubility limit, reinforcing the identified kinetic route via controlled amorphization and regrowth. revision: partial

Circularity Check

0 steps flagged

No significant circularity; MD trajectories are independent of the barrier fit

full rationale

The paper first fits Morse parameters to a set of DFT migration and kick-in/out barriers, then runs separate MD simulations with the fixed potential to generate implantation damage and annealing trajectories. The reported substitutional Al fractions, cluster statistics, and temperature-dependent planar-defect formation are direct simulation outputs, not algebraic or statistical re-derivations of the fitted barrier values. No self-citation, uniqueness theorem, or ansatz is invoked to force the central claims; the results remain falsifiable against external benchmarks and do not reduce to the input fit by construction.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claims rest on the transferability of a DFT-fitted Morse potential and on the assumption that short MD runs capture the early annealing physics that determines final substitutional fractions.

free parameters (1)
  • Morse Al-SiC interaction parameters
    Reparameterized to reproduce DFT migration and kick-in/out barriers; these parameters directly control the simulated defect evolution and Al incorporation.
axioms (2)
  • domain assumption Gao-Weber potential accurately describes SiC host lattice and defect energetics
    Used as the base potential for all SiC interactions in the MD runs.
  • domain assumption MD-accessible timescales are representative of early-stage annealing defect dynamics
    Invoked to link simulated cluster stability to final dopant activation.

pith-pipeline@v0.9.0 · 5564 in / 1454 out tokens · 53124 ms · 2026-05-08T11:21:21.894140+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

16 extracted references · 16 canonical work pages · 1 internal anchor

  1. [1]

    The influence of implantation conditions on dopant activation in Al-implanted 4H-SiC: A MD study applying an Al potential fitted to DFT barriers

    These de- fects degrade electrical performance and act as sinks or traps for + P V S O B M / B N F < Z F B S > < W P M > 1–15 | 1 arXiv:2604.22434v1 [cond-mat.mtrl-sci] 24 Apr 2026 dopants, thereby reducing activation efficiency . Experimental studies have demonstrated that both implanta- tion temperature and dopant concentration critically influence defe...

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  2. [2]

    Absolute saturation values are not expected to match steady-state experimental data because the MD simulations ac- cess early-stage annealing. However, the relative trends are ro- bust and correlate directly with the defect inventory: VC, Al com- plexes, and extended clusters represent compensating and trap- ping centers that limit the population of elect...

    work page 2017

  3. [3]

    Nipoti and A

    2 R. Nipoti and A. Parisini,Materials Science Forum, 2018,924,

    work page 2018

  4. [4]

    Kimoto, K

    3 T. Kimoto, K. Kawahara, H. Niwa, N. Kaji and J. Suda,2014 International Workshop on Junction T echnology (IWJT), 2014, 1–6. 4 P. O. A. Persson, L. Hultmann, J. M. S., H. A. and Y. R.,Journal of Applied Physics, 2003,53,

    work page 2014

  5. [5]

    Nipoti, H

    5 R. Nipoti, H. M. Ayedh and B. G. Svensson,Materials Science in Semiconductor Processing, 2018,78,

    work page 2018

  6. [6]

    Wendler, A

    6 E. Wendler, A. Heft and W. Wesch,Nuclear Instruments and Methods in Physics Research Section B, 1998,141,

    work page 1998

  7. [7]

    Negoro, T

    8 Y. Negoro, T. Kimoto, H. Matsunami, F. Schmid and G. Pensl, Journal of Applied Physics, 2004,96,

    work page 2004

  8. [8]

    9 Z. Wang, Z. Cao, Y. Li, S. Liu and G. Ran,Ceramics Interna- tional, 2023,49, 40787–40796. 10 H. Zang, H. Yu, T. Wang, F. Liu, C. Chen, P. Zhou, T. Shi, H. He, W. Liu and C. He,Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 2025,560, 165611. 11 J. Wu, Z. Xu, L. Liu, A. Hartmeier, M. Rommel, K. N...

    work page 2023

  9. [9]

    Matsushima and J

    15 N. Matsushima and J. Yamauchi,Japanese Journal of Applied Physics, 2019,58, 031001–10. 16 M. Bockstedte, A. Mattausch and O. Pankratov,Physical Re- view B, 2003,68, 205201–205218. 17 T. Kobayashi, K. Harada, Y. Kumagai, F. Oba and Y.-i. Mat- sushita,Journal of Applied Physics, 2019,125, 125701. 18 F. Gao, W. J. Weber, M. Posselt and V. Belko,Physical R...

    work page 2019

  10. [10]

    24 C. R. Dandekar and Y. C. Shin,Composites Part A: Applied Science and Manufacturing, 2011,42, 355–363. 25 A. P. Thompson,Computer Physics Communications, 2022, 271, 108171. 26 T. D. Kühne, M. Iannuzzi, M. Del Ben, V. V. Rybkin, P. See- wald, F. Stein, T. Laino, R. Z. Khaliullin, O. Schütt, F. Schiff- mannet al.,The Journal of Chemical Physics, 2020,152,...

    work page 2011

  11. [11]

    29 J. P. Perdew, K. Burke and M. Ernzerhof,Physical review letters, 1996,77,

    work page 1996

  12. [12]

    Henkelman, B

    30 G. Henkelman, B. P. Uberuaga and H. Jónsson,The Journal of chemical physics, 2000,113, 9901–9904. 31 Y. Huang, Y. Qian, Y. Zhang, D. Yang and X. Pi,Journal of Applied Physics, 2022,132, 015701–015707. 32 R. Stella,MSc thesis, TU Wien,

    work page 2000

  13. [13]

    Bockstedte, A

    34 M. Bockstedte, A. Mattausch and O. Pankratov,Physical Re- view B, 2004,69, 235202–235215. 35 A. Stukowski,Modelling and Simulation in Materials Science and Engineering, 2009,18, 015012. 36 S. Leroch, R. Stella, A. Hössinger and L. Filipovic, 2023 In- ternational Conference on Simulation of Semiconductor Pro- cesses and Devices (SISPAD), 2023, pp. 185–1...

    work page 2004

  14. [14]

    Zhang, W

    39 Y. Zhang, W. Weber, W. Jiang, A. Hallen and G. Possnert,Jour- nal of Applied Physics, 2002,91, 6388–6395. 40 A. Gali, T. Hornos, N. Son, E. Janzen and W. Choyke,Physical Review B, 2007,75, 045211–045219. 41 A. Mattausch, M. Bockstedte and O. Pankratov,Physical Re- view B, 2004,70, 235211–15. 42 P. Li, P. Udvarhelyi, S. Li, B. Huang and A. Gali,Physical...

    work page 2002

  15. [15]

    46 X. Yan, P. Li, S.-H. Kang and B. Huang,Journal of Applied Physics, 2020,127, 1–7. 47 H. Ayedh, R. Nipoti, A. Hallén and B. Svensson,Journal of Applied Physics, 2017,122 (2), 025701. 48 U. Grossner, J. K. Grillenberger, J. Woerle, M. E. Bathen and J. Müting, inIntrinsic and Extrinsic Electrically Active Point De- fects in SiC, ed. P. Wellmann, N. Ohtani...

    work page 2020

  16. [16]

    Kyungwon and L

    52 L. Kyungwon and L. M. Rubin, SEMICON Korea 2015, SEMI Technology Symposium, 2015, pp. 1–4. 53 J. Wong-Leung, M. K. Linnarson, B. G. Svennson and D. J. H. Cockayne,Physical Review B, 2005,71, 165210. + P V S O B M / B N F < Z F B S > < W P M > 1–15 | 15

    work page 2015