pith. sign in

arxiv: 2604.21438 · v2 · pith:XGHCOIKRnew · submitted 2026-04-23 · 🌌 astro-ph.GA · astro-ph.IM· astro-ph.SR

SFUMATO#: A GPU-accelerated code for self-gravitational radiation hydrodynamics simulation with adaptive mesh refinement

Hajime Fukushima , Tomoaki Matsumoto This is my paper

Pith reviewed 2026-05-22 10:26 UTC · model grok-4.3

classification 🌌 astro-ph.GA astro-ph.IMastro-ph.SR
keywords GPU-accelerated codeadaptive mesh refinementself-gravitational radiation hydrodynamicsnon-equilibrium chemistrylinearized implicit solvermultigrid self-gravityM1 closure radiation transferastrophysical simulation
0
0 comments X

The pith

SFUMATO# implements self-gravitational radiation hydrodynamics on GPUs with adaptive mesh refinement using a linearized chemistry solver.

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

The paper introduces SFUMATO#, an updated version of the SFUMATO code designed for simulations involving self-gravity, radiation, and hydrodynamics on grids that refine adaptively. It employs CUDA and HIP for GPU acceleration and MPI for multi-device runs. A central innovation is the linearized implicit method for solving non-equilibrium chemistry and thermal evolution, which is shown to stay accurate even when the dust heat capacity is made much larger to speed up calculations. This matters because such simulations are computationally intensive, and efficient codes are needed to model processes like star formation in molecular clouds.

Core claim

The authors present SFUMATO# as a new implementation that solves self-gravitational radiation hydrodynamics problems using adaptive mesh refinement with the CUDA/HIP programming frameworks. The code includes a multigrid solver for self-gravity, radiation transfer with M1 closure and reduced speed of light, non-equilibrium chemistry, thermal evolution, and sink particles. New solvers based on a linearized implicit method are developed and validated by comparison with Newton-Raphson solutions, and increasing the pseudo dust heat capacity is shown to accelerate the chemistry solver while preserving accuracy up to three orders of magnitude.

What carries the argument

Linearized implicit method for non-equilibrium chemistry and thermal evolution combined with increased pseudo dust heat capacity

If this is right

  • The code supports efficient multi-GPU execution via MPI for simulations of giant molecular clouds.
  • The self-gravity solver cost grows with the number of MPI processes, so balanced device allocation is required for good scaling.
  • Test problems confirm validity of radiation transfer, AMR, and other components alongside the new chemistry solver.
  • Accuracy holds for the chemistry and thermal evolution even after the heat capacity adjustment.

Where Pith is reading between the lines

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

  • The boosted heat capacity trick may apply to other astrophysical codes that couple chemistry to hydrodynamics.
  • Optimal performance likely requires matching the number of GPUs to the relative costs of gravity and hydro steps.
  • Larger domain simulations of star-forming regions become feasible if the scaling behavior generalizes beyond the tested cases.

Load-bearing premise

The linearized implicit method continues to produce accurate solutions even when the pseudo dust heat capacity is increased substantially.

What would settle it

A direct comparison of chemical abundances and temperatures from the chemistry solver on a standard test problem using the realistic dust heat capacity versus the value increased by three orders of magnitude.

Figures

Figures reproduced from arXiv: 2604.21438 by Hajime Fukushima, Tomoaki Matsumoto.

Figure 1
Figure 1. Figure 1: Parallelization of numerical flux calculations at cell interfaces within a block. Each block contains 8 3 cells, and the computation is parallelized using 64 threads for each group of 8 cells aligned along the update direction. The thick arrows indicate the numerical fluxes at the cell interfaces. Alt text: Schematic illustration showing the strategy for calculating numerical fluxes in our code. coarsened … view at source ↗
Figure 2
Figure 2. Figure 2: Density (ρ, blue), velocity (v, orange), and pressure (p, green) in the shock tube test. Lines show the results of numerical simulations, and dots represent the exact solutions of the shock tube test. Alt text: Three￾line graph showing density, velocity, and pressure in the shock tube test. examined in Section 3.3. Radiation transport with M1-closure is validated in Section 3.4. In Section 3.5, we assess t… view at source ↗
Figure 3
Figure 3. Figure 3: Density distribution for the double Mach reflection problem at t = 0.2 (top), and the corresponding grid structure (bottom). In the bottom panel, the black lines indicate the block boundaries. The top panel shows 30 contour lines in the range 2 < ρ < 24. Alt text: Two-panel figure. 8 3 cells. The minimum and maximum resolutions are h = 1/64 and 1/1024, respectively. Refinement is triggered according to the… view at source ↗
Figure 4
Figure 4. Figure 4: shows the relative error of the numerically computed gravitational force, |g − gex|/|gex| where gex denotes the exact gravity force calculated analytically. The maximum error occurs near the spheres due to the sharp density gradient at their edge. Elsewhere, the error remains below 10−3 , and the error distribu￾tion is smoothly connected across the interfaces between coarser and finer grids. This indicates… view at source ↗
Figure 6
Figure 6. Figure 6: The time evolution of chemical abundances in a test problem of molecular hydrogen formation. The colored lines show the abundances of H (blue), H + (orange), and H2 (green). The black dashed line shows the analytical solution given by Equation (37). Alt text: Time evolution of the abundances of H2, H, and H + shown as three colored lines with the analytical solution given by Equation (37) overplotted as a … view at source ↗
Figure 5
Figure 5. Figure 5: Photon densities in the beam experiments at t = 3.2 × 10−7 . The top panel shows the result for an off-axis beam with an angle of 30◦ , while the bottom panel presents the result for a beam aligned with the x-axis. The white lines delineate the block boundaries. Alt text: Two-panel figure. 3.5 Time evolution of chemical abundances As described in Section 2.2.8, we adopt a newly developed LI scheme to evolv… view at source ↗
Figure 7
Figure 7. Figure 7: Time evolution of the relative error of yH2 between the numerical results and the analytical solution given by Equation (37). The dashed, solid, and dot-dashed lines show the cases with fchem = 0.1, 0.03, and 0.01, respectively. Alt text: Graph showing the time evolution of the relative error in the molecular hydrogen abundance for different values of fchem. 3.5.2 Case 2: Photoionization of atomic hydrogen… view at source ↗
Figure 9
Figure 9. Figure 9: Time evolution of the relative error of yH+ between the numerical results and the analytical solution given by Equation (39). The dashed, solid, and dot-dashed lines show the cases with fchem = 0.1, 0.03, and 0.01, respectively. Alt text: Graph showing the time evolution of the relative error in the ionized hydrogen abundance for different values of fchem [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Gas temperature (top) and dust temperature (bottom) as a func￾tion of number density at the epoch of sink formation in the simulations of Bonnor-Ebert spheres. The colored lines represent the results obtained with the LI method using Cd = 107 erg g−1 K−1 . Each color corresponds to a different metallicity: Z = 1 Z⊙ (blue), 10−2 Z⊙ (orange), and 10−4 Z⊙ (green). The dashed lines show the results calculated… view at source ↗
Figure 12
Figure 12. Figure 12: shows the abundances of H, H2, and H + obtained us￾ing both of the LI and NR methods at Z = 1 Z⊙. At this metallicity, gas is almost fully molecular owing to reactions on the surface of dust grains. The new solver accurately reproduces the NR results, even for the low-abundance species H and H +. In low-metallicity environments, the reduced dust abundance decreases the total thermal energy reservoir of du… view at source ↗
Figure 13
Figure 13. Figure 13: The relative errors in the gas and dust temperature between the LI and NR methods at Z = 10−2 Z⊙ and 10−4 Z⊙. We adopt Cd = 107 erg g−1 K−1 for the LI method. The blue and orange lines cor￾respond to the cases with Z = 10−2 Z⊙ and 10−4 Z⊙, respectively. The solid and dashed lines represent the relative errors in the gas and dust temperatures. Alt text: Graph showing the relative errors of gas and dust tem… view at source ↗
Figure 14
Figure 14. Figure 14: Radial profiles of gas density (top) and temperature (bottom) at the epoch of sink formation for Z = 1 Z⊙. Each line corresponds to a simulation using either a synchronous (blue) or adaptive (black) timestep. The vertical dashed line indicates the sink radius, rsink = 34 au. Alt text: Radial profiles of gas density and temperature comparing simulations with synchronous and adaptive timestep. 10 1 10 2 10 … view at source ↗
Figure 15
Figure 15. Figure 15: Radial distribution of the relative errors in number density (nH, dashed line) and gas temperature (T, solid line) between the synchronous and adaptive timestep models. As in [PITH_FULL_IMAGE:figures/full_fig_p012_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: shows the radial profiles of the temperatures and chem￾ical abundances at t = 2 × 105 yr. The colored and black dashed lines show the results obtained with the LI and NR methods. For the LI method, we adopt Cd = 107 erg g−1 K −1 . The ionization front is located at ∼ 2.7 pc, which coincides with the Strömgren radius estimated from Equation (42). Within the HII regions, the gas is heated to Tg ∼ 104 K and … view at source ↗
Figure 17
Figure 17. Figure 17: Relative errors in the gas temperature (top) and dust temperature (bottom) between the LI and NR methods in the HII region formation tests. Each colors denotes a different pseudo-specific capacity:Cd = 104 (blue), 105 (orange), 106 (green), 107 (red), and 108 erg g−1 K−1 (purple). Alt text: Graph showing the relative error in gas and dust temperatures be￾tween the LI and NR methods for different values of… view at source ↗
Figure 18
Figure 18. Figure 18: shows the performance for the 8 3 and 163 blocks (see also [PITH_FULL_IMAGE:figures/full_fig_p015_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Same as [PITH_FULL_IMAGE:figures/full_fig_p016_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: shows the performance of the hydrodynamics solver with the 8 3 and 163 blocks (see also [PITH_FULL_IMAGE:figures/full_fig_p017_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Same as [PITH_FULL_IMAGE:figures/full_fig_p018_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: Density structure in the GMC simulation at t = 1.73 Myr with the 8 3 blocks. Alt text: Density structure in the GMC simulation at t= 1.73 Myr [PITH_FULL_IMAGE:figures/full_fig_p018_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: Same as [PITH_FULL_IMAGE:figures/full_fig_p019_23.png] view at source ↗
read the original abstract

We present a new implementation of the SFUMATO code, called SFUMATO#, for solving self-gravitational radiation hydrodynamics problems using adaptive mesh refinement (AMR) with the CUDA/HIP programming frameworks. The code incorporates a multigrid solver for self-gravity, radiation transfer with M1 closure and reduced speed of light approximation, non-equilibrium chemistry, thermal evolution, and sink particle schemes. We develop new non-equilibrium chemistry and thermal solvers based on a linearized implicit method, whose accuracy is validated through a series of test problems by comparison with solutions obtained using the Newton-Raphson method. By incorporating the heat capacity of dust grains, the dust temperature can be evolved without iterative energy-balance calculations. From the perspective of computational cost, we demonstrate that adopting an increased pseudo dust heat capacity accelerates the chemistry solver while preserving accuracy, even when the value is increased by up to three orders of magnitude relative to the realistic value. In addition, we perform a suite of test problems to confirm the validity of the other components of our implementation. The code supports multi-GPU execution via MPI-based parallelization. We measure the strong-scaling performance of the hydrodynamics and self-gravity solvers on both uniform and AMR grids, as well as the overall code performance using a giant molecular cloud simulation. We find that the computational cost of the self-gravity solver increases with the number of MPI processes, indicating that efficient parallel performance is achieved only when the number of devices is chosen such that the cost of the self-gravity solver remains comparable to that of the other components.

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

1 major / 1 minor

Summary. The manuscript describes SFUMATO#, a new GPU-accelerated implementation of the SFUMATO code for self-gravitational radiation hydrodynamics simulations with adaptive mesh refinement (AMR) using CUDA/HIP. It includes a multigrid self-gravity solver, radiation transfer with M1 closure and reduced speed of light approximation, non-equilibrium chemistry and thermal evolution via a new linearized implicit method (validated against Newton-Raphson), sink particles, and multi-GPU MPI parallelization. Accuracy of the chemistry/thermal solvers is claimed to be preserved even with pseudo dust heat capacity increased by up to three orders of magnitude; strong-scaling performance is measured on uniform/AMR grids and demonstrated in a giant molecular cloud simulation.

Significance. If the validations hold, this provides a practical tool for efficient large-scale simulations of molecular clouds incorporating radiation, chemistry, and self-gravity on GPU architectures. The suite of test problems for component validation and the multi-GPU performance measurements are positive features that support reproducibility and usability.

major comments (1)
  1. [Abstract (validation of linearized implicit method)] Abstract (validation of linearized implicit method): The central efficiency claim—that increasing the pseudo dust heat capacity by up to three orders of magnitude accelerates the chemistry solver while preserving accuracy—is load-bearing for the new solver's utility. However, no quantitative error metrics (e.g., maximum relative errors in abundances or temperature), specific test conditions (density/temperature/optical depth ranges relevant to molecular clouds), or detailed comparison tolerances versus Newton-Raphson are supplied, leaving unclear whether the approximation remains robust when coupled to radiation transfer and self-gravity.
minor comments (1)
  1. [Performance measurements] The scaling discussion notes that self-gravity solver cost increases with MPI processes but could be strengthened by explicit tables or figures comparing relative costs of hydrodynamics, gravity, and chemistry components across device counts.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments on our manuscript. We address the major comment regarding the abstract's description of the linearized implicit solver validation below.

read point-by-point responses

  1. Referee: [Abstract (validation of linearized implicit method)] The central efficiency claim—that increasing the pseudo dust heat capacity by up to three orders of magnitude accelerates the chemistry solver while preserving accuracy—is load-bearing for the new solver's utility. However, no quantitative error metrics (e.g., maximum relative errors in abundances or temperature), specific test conditions (density/temperature/optical depth ranges relevant to molecular clouds), or detailed comparison tolerances versus Newton-Raphson are supplied, leaving unclear whether the approximation remains robust when coupled to radiation transfer and self-gravity.

    Authors: We agree that the abstract would be clearer with explicit quantitative support for the efficiency claim. The main text already presents detailed comparisons of the linearized implicit method against Newton-Raphson solutions, including maximum relative errors in chemical abundances and gas/dust temperatures across density and temperature ranges relevant to molecular clouds, as well as the specific tolerances used. These validations are performed within the context of the radiation hydrodynamics framework. To address the concern directly, we will revise the abstract to include a concise statement summarizing the quantitative error levels and test conditions. The robustness under coupling is further supported by the full giant molecular cloud simulation, which incorporates radiation transfer, self-gravity, and the chemistry solver simultaneously. revision: yes

Circularity Check

0 steps flagged

Software implementation and validation paper exhibits no circularity in derivation chain

full rationale

This is a code development and performance paper describing the SFUMATO# implementation for self-gravitational radiation hydrodynamics with AMR, CUDA/HIP, multigrid gravity, M1 radiation, and new linearized implicit chemistry/thermal solvers. The central claims rest on direct comparisons of the new solvers to Newton-Raphson solutions on test problems and on measured scaling in a giant molecular cloud run; these are external benchmarks rather than quantities derived from the same fitted inputs. No equations reduce by construction to prior results, no parameters are fitted to data then relabeled as predictions, and no uniqueness theorems or ansatzes are smuggled via self-citation. Prior SFUMATO references establish code lineage but do not carry the load for the accuracy or performance assertions, which are independently tested. The paper is therefore self-contained against external test benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The implementation rests on standard numerical astrophysics techniques plus a small number of modeling choices whose impact is explored in the tests.

free parameters (1)
  • pseudo dust heat capacity multiplier = up to 1000
    Increased by up to three orders of magnitude relative to the physical value in order to accelerate the chemistry solver while preserving accuracy.
axioms (2)
  • domain assumption M1 closure for radiation transfer
    Standard closure relation adopted for the radiation moment equations.
  • domain assumption reduced speed of light approximation
    Approximation used to make the radiation transport timestep feasible.

pith-pipeline@v0.9.0 · 5827 in / 1310 out tokens · 37058 ms · 2026-05-22T10:26:34.537977+00:00 · methodology

Review history (2 revisions) →

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

68 extracted references · 68 canonical work pages

  1. [1]

    Abel, T., Anninos, P., Zhang, Y ., & Norman, M. L. 1997, New Astronomy, Publications of the Astronomical Society of Japan(2024), Vol. 00, No. 021 2, 181

    work page 1997

  2. [2]

    Bakes, E. L. O., & Tielens, A. G. G. M. 1994, ApJ, 427, 822

    work page 1994

  3. [3]

    J., & Colella, P

    Berger, M. J., & Colella, P. 1989, Journal of Computational Physics, 82, 64

    work page 1989

  4. [4]

    J., & Oliger, J

    Berger, M. J., & Oliger, J. 1984, Journal of Computational Physics, 53, 484

    work page 1984

  5. [5]

    2021, A&A, 646, A123

    Bloch, H., Tremblin, P., González, M., Padioleau, T., & Audit, E. 2021, A&A, 646, A123

    work page 2021

  6. [6]

    L., Norman, M

    Bryan, G. L., Norman, M. L., O’Shea, B. W., et al. 2014, ApJS, 211, 19

    work page 2014

  7. [7]

    I., Franco, J., & Shore, S

    Diaz-Miller, R. I., Franco, J., & Shore, S. N. 1998, ApJ, 501, 192

    work page 1998

  8. [8]

    Draine, B. T. 2011, Physics of the Interstellar and Intergalactic Medium (Princeton: Princeton University Press)

    work page 2011

  9. [9]

    C., & Klessen, R

    Federrath, C., Banerjee, R., Clark, P. C., & Klessen, R. S. 2010, ApJ, 713, 269

    work page 2010

  10. [10]

    J., Peterson, B

    Ferland, G. J., Peterson, B. M., Horne, K., Welsh, W. F., & Nahar, S. N. 1992, ApJ, 387, 95

    work page 1992

  11. [11]

    Forrey, R. C. 2013, ApJL, 773, L25

    work page 2013

  12. [12]

    2000, ApJS, 131, 273

    Fryxell, B., Olson, K., Ricker, P., et al. 2000, ApJS, 131, 273

    work page 2000

  13. [13]

    2020, MNRAS, 497, 829

    Fukushima, H., Hosokawa, T., Chiaki, G., et al. 2020, MNRAS, 497, 829

    work page 2020

  14. [14]

    2018, MNRAS, 473, 4754

    Fukushima, H., Omukai, K., & Hosokawa, T. 2018, MNRAS, 473, 4754

    work page 2018

  15. [15]

    2021, MNRAS, 506, 5512 —

    Fukushima, H., & Yajima, H. 2021, MNRAS, 506, 5512 —. 2022, MNRAS, 511, 3346 —. 2023, MNRAS, 524, 1422

    work page 2021

  16. [16]

    1998, A&A, 335, 403

    Galli, D., & Palla, F. 1998, A&A, 335, 403

    work page 1998

  17. [17]

    E., & Langer, W

    Glassgold, A. E., & Langer, W. D. 1974, ApJ, 193, 73

    work page 1974

  18. [18]

    Glover, S. C. O., & Jappsen, A. K. 2007, ApJ, 666, 1

    work page 2007

  19. [19]

    Y ., & Abel, T

    Gnedin, N. Y ., & Abel, T. 2001, New Astronomy, 6, 437

    work page 2001

  20. [20]

    C., & Wolfire, M

    Gong, M., Ostriker, E. C., & Wolfire, M. G. 2017, ApJ, 843, 38

    work page 2017

  21. [21]

    2019, A&A, 621, A85

    Hainich, R., Ramachandran, V ., Shenar, T., et al. 2019, A&A, 621, A85

    work page 2019

  22. [22]

    D., & Leer, B

    Harten, A., Lax, P. D., & Leer, B. v. 1983, SIAM review, 25, 35

    work page 1983

  23. [23]

    D., & Krumholz, M

    He, C.-C., Wibking, B. D., & Krumholz, M. R. 2024, MNRAS, 531, 1228

    work page 2024

  24. [24]

    1990, Numerical computation of internal and external flows

    Hirsch, C. 1990, Numerical computation of internal and external flows. V ol. 2 (New York: John Wiley & Sons) - Computational methods for inviscid and viscous flows

    work page 1990

  25. [25]

    W., & Eastwood, J

    Hockney, R. W., & Eastwood, J. W. 1981, Computer Simulation Using Particles (New York: McGraw-Hill)

    work page 1981

  26. [26]

    Hollenbach, D., & McKee, C. F. 1979, ApJS, 41, 555

    work page 1979

  27. [27]

    2016, ApJ, 824, 119

    Hosokawa, T., Hirano, S., Kuiper, R., et al. 2016, ApJ, 824, 119

    work page 2016

  28. [28]

    2009, ApJ, 691, 823

    Hosokawa, T., & Omukai, K. 2009, ApJ, 691, 823

    work page 2009

  29. [29]

    2020, MNRAS, 497, 5061

    Inoguchi, M., Hosokawa, T., Mineshige, S., & Kim, J.-G. 2020, MNRAS, 497, 5061

    work page 2020

  30. [30]

    2015, A&A, 580, A49

    Inutsuka, S.-i., Inoue, T., Iwasaki, K., & Hosokawa, T. 2015, A&A, 580, A49

    work page 2015

  31. [31]

    2026, ApJ, 999, 257

    Kimura, K., Sugimura, K., Hosokawa, T., Fukushima, H., & Omukai, K. 2026, ApJ, 999, 257

    work page 2026

  32. [32]

    2010, Science, 329, 69

    Kreckel, H., Bruhns, H., ˇCížek, M., et al. 2010, Science, 329, 69

    work page 2010

  33. [33]

    R., Klein, R

    Krumholz, M. R., Klein, R. I., McKee, C. F., & Bolstad, J. 2007, ApJ, 667, 626

    work page 2007

  34. [34]

    Laor, A., & Draine, B. T. 1993, ApJ, 402, 441

    work page 1993

  35. [35]

    Larson, R. B. 1985, MNRAS, 214, 379 —. 2005, MNRAS, 359, 211

    work page 1985

  36. [36]

    1985, A&A, 144, 147

    Leger, A., Jura, M., & Omont, A. 1985, A&A, 144, 147

    work page 1985

  37. [37]

    1997, A&AS, 125, 229

    Lejeune, T., Cuisinier, F., & Buser, R. 1997, A&AS, 125, 229

    work page 1997

  38. [38]

    Levermore, C. D. 1984, J. Quant. Spectrosc. Radiat. Transfer, 31, 149

    work page 1984

  39. [39]

    N., Inutsuka, S.-i., & Matsumoto, T

    Machida, M. N., Inutsuka, S.-i., & Matsumoto, T. 2010, ApJ, 724, 1006

    work page 2010

  40. [40]

    G., Schwarz, D

    Martin, P. G., Schwarz, D. H., & Mandy, M. E. 1996, ApJ, 461, 265

    work page 1996

  41. [41]

    2007, PASJ, 59, 905

    Matsumoto, T. 2007, PASJ, 59, 905

    work page 2007

  42. [42]

    2015, ApJ, 801, 77

    Matsumoto, T., Dobashi, K., & Shimoikura, T. 2015, ApJ, 801, 77

    work page 2015

  43. [43]

    2003, ApJ, 583, 296

    Matsumoto, T., & Hanawa, T. 2003, ApJ, 583, 296

    work page 2003

  44. [44]

    H., Federrath, C., & Krumholz, M

    Menon, S. H., Federrath, C., & Krumholz, M. R. 2022, MNRAS, 517, 1313

    work page 2022

  45. [45]

    Mihalas, D., & Mihalas, B. W. 1984, Foundations of radiation hydrodynam- ics

    work page 1984

  46. [46]

    2000, ApJ, 534, 809

    Omukai, K. 2000, ApJ, 534, 809

    work page 2000

  47. [47]

    2010, ApJ, 722, 1793

    Omukai, K., Hosokawa, T., & Yoshida, N. 2010, ApJ, 722, 1793

    work page 2010

  48. [48]

    2005, ApJ, 626, 627

    Omukai, K., Tsuribe, T., Schneider, R., & Ferrara, A. 2005, ApJ, 626, 627

    work page 2005

  49. [49]

    Osterbrock, D. E. 1989, Astrophysics of gaseous nebulae and active galactic nuclei (Mill Valley: University Science Books)

    work page 1989

  50. [50]

    E., & Stahler, S

    Palla, F., Salpeter, E. E., & Stahler, S. W. 1983, ApJ, 271, 632

    work page 1983

  51. [51]

    Petrosian, V ., Silk, J., & Field, G. B. 1972, ApJL, 177, L69

    work page 1972

  52. [52]

    Roe, P. L. 1981, Journal of Computational Physics, 43, 357

    work page 1981

  53. [53]

    2013, MNRAS, 436, 2188

    Rosdahl, J., Blaizot, J., Aubert, D., Stranex, T., & Teyssier, R. 2013, MNRAS, 436, 2188

    work page 2013

  54. [54]

    2015, MNRAS, 449, 4380

    Rosdahl, J., & Teyssier, R. 2015, MNRAS, 449, 4380

    work page 2015

  55. [55]

    A., Goldbaum, N

    Schive, H.-Y ., ZuHone, J. A., Goldbaum, N. J., et al. 2018, MNRAS, 481, 4815

    work page 2018

  56. [56]

    E., & Robertson, B

    Schneider, E. E., & Robertson, B. E. 2015, ApJS, 217, 24

    work page 2015

  57. [57]

    R., & Kang, H

    Shapiro, P. R., & Kang, H. 1987, ApJ, 318, 32

    work page 1987

  58. [58]

    A., & Ostriker, E

    Skinner, M. A., & Ostriker, E. C. 2013, ApJS, 206, 21

    work page 2013

  59. [59]

    M., Tomida, K., White, C

    Stone, J. M., Tomida, K., White, C. J., & Felker, K. G. 2020, ApJS, 249, 4

    work page 2020

  60. [60]

    M., Mullen, P

    Stone, J. M., Mullen, P. D., Fielding, D., et al. 2024, arXiv e-prints, arXiv:2409.16053

    work page arXiv 2024

  61. [61]

    2017, MNRAS, 469, 62

    Sugimura, K., Hosokawa, T., Yajima, H., & Omukai, K. 2017, MNRAS, 469, 62

    work page 2017

  62. [62]

    2002, A&A, 385, 337

    Teyssier, R. 2002, A&A, 385, 337

    work page 2002

  63. [63]

    Tielens, A. G. G. M., & Hollenbach, D. 1985, ApJ, 291, 722

    work page 1985

  64. [64]

    Tomida, K., & Stone, J. M. 2023, ApJS, 266, 7

    work page 2023

  65. [65]

    F., Spruce, M., & Speares, W

    Toro, E. F., Spruce, M., & Speares, W. 1994, Shock waves, 4, 25

    work page 1994

  66. [66]

    D., & Krumholz, M

    Wibking, B. D., & Krumholz, M. R. 2022, MNRAS, 512, 1430

    work page 2022

  67. [67]

    1984, Journal of Computational Physics, 54, 115

    Woodward, P., & Colella, P. 1984, Journal of Computational Physics, 54, 115

    work page 1984

  68. [68]

    2011, ApJS, 196, 20 Appendix 1 Chemical network In Table 11, we summarize the chemical network consisting ofH, H2,H +, ande −

    Zhang, W., Howell, L., Almgren, A., Burrows, A., & Bell, J. 2011, ApJS, 196, 20 Appendix 1 Chemical network In Table 11, we summarize the chemical network consisting ofH, H2,H +, ande −. The abundance ofH − is obtained as an interme- diate species, following the treatment described in Appendix A of Omukai et al. (2010). The photoionization rate of hydroge...

    work page 2011