SuperLite: Monte Carlo Radiative Transfer

• SuperLite is a specialized Monte Carlo radiative transfer code designed to compute synthetic spectra for astrophysical transients affected by circumstellar interaction.
• It employs a steady-state, semi-implicit, and semi-relativistic framework with detailed opacity treatments and novel non-homologous velocity field handling.
• Validated against benchmark models, SuperLite effectively simulates spectra for various events including Type Ia, IIP, IIn, and superluminous supernovae.

to=arxiv_search.search 彩神争霸平台json սխալ {"query":"SuperLite Monte Carlo radiation transport astrophysical transients circumstellar interaction", "max_results": 10} to=arxiv_search 彩神争霸.factory"} SuperLite is an open-source Monte Carlo radiation transport code designed to produce synthetic spectra for astrophysical transient phenomena affected by circumstellar interaction. It is built by significantly modifying the Los Alamos code SuperNu, but removes key assumptions—especially homologous expansion—that break down in shocked, interacting outflows. In its stated niche, SuperLite is a steady-state, semi-implicit, semi-relativistic, multigroup Monte Carlo transport code optimized for fast, high-fidelity post-processing of hydrodynamic or radiation-hydrodynamic profiles, with particular emphasis on high-velocity shocked outflows, superluminous supernovae, pulsational pair-instability supernovae, and other peculiar transients in which circumstellar material reshapes the emergent spectrum (Wagle et al., 2023).

1. Definition and scope

SuperLite is formulated as a post-processing code rather than a full explosion-evolution solver. Its inputs are a radial grid together with density, velocity, temperature, and composition profiles extracted from an external hydrodynamic or radiation-hydrodynamic calculation. The code then computes emergent spectra from that snapshot using Monte Carlo packet propagation, multi-group structured opacity calculations, and detailed line formation physics. This division of labor is central to its design: hydrodynamics and radiation–matter coupling are assumed to have already been accounted for in the input structure, while SuperLite focuses on spectral synthesis.

The code was developed for regimes in which circumstellar interaction invalidates assumptions commonly used in supernova spectrum synthesis. In particular, the shocked ejecta between forward and reverse shocks is non-homologous, the velocity field may contain regions with $dv/dr < 0$, and line formation occurs in cool dense shells and interaction-powered outflows. SuperLite therefore targets cases in which $v \sim 0.01\text{–}0.1\,c$, the flow is semi-relativistic, and the spectral problem is shaped by arbitrary 1D velocity profiles rather than by the simplifications of homologous expansion.

Its intended astrophysical domain includes Type Ia, Type IIP, and Type IIn supernovae, as well as superluminous supernovae, pulsational pair-instability supernova shell collisions, and other peculiar transients where pre-explosion mass ejection or mergers produce dense circumstellar material. This scope places SuperLite between general radiation-hydrodynamics packages, which evolve the explosion but typically do not provide high-resolution synthetic spectra, and specialized spectrum codes that assume homologous expansion.

2. Transport formalism and numerical scheme

SuperLite uses Implicit Monte Carlo with Discrete Diffusion Monte Carlo acceleration in optically thick regions. Its transport starting point is the time-dependent radiative transfer equation in the comoving frame, expanded to $O(v/c)$, so that Doppler shifts, aberration, and compressive or expansive work are retained without invoking a fully relativistic formalism (Wagle et al., 2023). The normalized Planck function is written as

$$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$

and the Fleck factor is

$$f = \frac{1}{1 + 4 a T^3 c \sigma_{P,e}\,\Delta t / C_V}.$$

Because SuperLite is a steady-state post-processor, it adopts $\Delta t \to \infty$, implying $f \to 0$. In that limit, all collisions are treated as effective scattering, the comoving-frame packet energy is conserved through interactions, and SuperLite does not evolve the material energy.

This steady-state reduction is paired with a Monte Carlo representation in which packets carry luminosity rather than an energy tied to an artificial time step. Packet weights are normalized so that the sum of packet luminosities reproduces the bolometric luminosity of the input hydrodynamic profile. Packets are initialized in radius, direction, and frequency according to the local radiation temperature, and then propagated until escape, scattering, cell-boundary crossing, or Doppler shifting to a group edge.

A packet step is determined by

$$d_p = \min(d_B, d_{\rm thm}, d_{\rm col}, d_{\rm dop}),$$

where $d_B$ is the distance to the next cell boundary, $d_{\rm thm}$ the Thomson-scattering distance, $v \sim 0.01\text{–}0.1\,c$0 the effective-collision distance, and $v \sim 0.01\text{–}0.1\,c$1 the distance to a group boundary by Doppler shift. In 1D spherical geometry, packet kinematics are updated analytically, and interaction distances are sampled from exponential laws.

The distinguishing algorithmic feature is the treatment of non-homologous expansion. In each cell, the radial velocity is approximated as linear,

$v \sim 0.01\text{–}0.1\,c$2

with

$v \sim 0.01\text{–}0.1\,c$3

This permits an analytic expression for the Doppler distance to a group edge,

$v \sim 0.01\text{–}0.1\,c$4

and a Newton–Raphson fallback when the $v \sim 0.01\text{–}0.1\,c$5 approximation in the auxiliary frame may break down. The result is a transport scheme that can handle arbitrary 1D velocity fields, including sign changes in $v \sim 0.01\text{–}0.1\,c$6 and shock structures, rather than only homologous flows.

3. Opacities, atomic physics, and NLTE treatment

SuperLite uses multi-group structured opacities with subgroups, inherited from the SuperNu methodology but adapted to the steady-state interacting-transient context (Wagle et al., 2023). Groups are logarithmically spaced in wavelength, while subgroups within each group have uniform linear spacing. By default, ionization and excitation are treated in local thermodynamic equilibrium. Ionization follows the Saha equation, and excitation follows the Boltzmann relation, yielding the Saha–Boltzmann ionization relation in terms of partition functions:

$v \sim 0.01\text{–}0.1\,c$7

For a bound–bound transition $v \sim 0.01\text{–}0.1\,c$8, the LTE line opacity is

$v \sim 0.01\text{–}0.1\,c$9

with emissivity supplied by Kirchhoff’s law,

$O(v/c)$0

Line opacity is computed from Kurucz line lists up to $O(v/c)$1, with hundreds of levels per species and about 786,000 transitions in total. Bound–free opacity uses analytic fits following Verner et al. (1996), free–free opacity uses Gaunt factors from Sutherland (1998), and Thomson scattering is included as elastic electron scattering.

The default opacity calculation is therefore LTE, but the code also includes an experimental NLTE treatment for hydrogen. In that module, ionization balance is still taken from LTE Saha equilibrium, while excited-state populations are obtained by solving a rate-matrix system,

$O(v/c)$2

for hydrogen levels up to principal quantum number $O(v/c)$3, subject to the population-conservation constraint

$O(v/c)$4

The included rates are photo-ionisation and radiative recombination, together with electron-impact collisional excitation and de-excitation. Other species remain treated in LTE. The hydrogen-only NLTE extension is explicitly described as under active testing, but it is motivated by the low-density, line-forming regions of interacting supernovae, particularly the Balmer line problem in SNe IIn.

4. Verification and benchmarking

The code paper presents verification against semi-analytic transfer tests and cross-code comparisons with established transport tools. These tests are central because SuperLite combines three nontrivial ingredients—Monte Carlo transport, line-by-line multi-group opacity treatment, and non-homologous velocity fields—that are rarely implemented together in an open-source package (Wagle et al., 2023).

Case	Setup	Outcome
Line-transfer test	Uniform sphere with a single line at 1 eV	Matches the semi-analytic P-Cygni profile within Monte Carlo noise
W7 SN Ia	Checkpointed SuperNu profile at 10 days	Spectra agree very well
s18.0 SN IIP	Checkpointed SuperNu profile at 20 days	Spectra agree very well

The line-transfer verification uses the Roth et al. (2015) setup: a uniform-density sphere with a single line and homologous expansion, followed by a non-homologous modification that adds a constant-velocity wind component and produces a cell with a negative velocity gradient. In the homologous case, SuperLite reproduces the standard P-Cygni profile within Monte Carlo noise. In the non-homologous cases, it captures the expected enhancement of emission and absorption features produced by blue-shifting back into resonance or by pure absorption in the modified region, demonstrating that the generalized Doppler algorithm remains functional when the monotonic-velocity assumption fails.

The W7 comparison uses a checkpointed SuperNu simulation of the Nomoto et al. Type Ia model at 10 days, with the bolometric luminosity from SuperNu used to normalize SuperLite packet weights. The resulting spectra agree very well, and the test also shows that SuperLite can reach comparable signal-to-noise with fewer packets because packets are not absorbed or censused for future time steps. A related truncation experiment shows that spectra remain nearly unchanged when the inner boundary is placed at Rosseland optical depth $O(v/c)$5, change only slightly at $O(v/c)$6, and differ significantly at $O(v/c)$7. The recommended practical choice is to include layers down to at least $O(v/c)$8.

The s18.0 Type IIP comparison again uses checkpointed SuperNu profiles and shows strong agreement at 20 days during the plateau. The code paper also presents application-level comparisons to STELLA- and HERACLES-based models and a qualitative comparison with CMFGEN for a luminous SN IIn analogue, extending validation from idealized tests to realistic interacting-transient calculations.

5. Astrophysical applications

SuperLite has been used as the spectral-synthesis engine in several interacting-transient studies. In its foundational applications, it was applied to a SN 1999em-like Type IIP model, to a classical SN IIn model with a dense H-rich circumstellar medium, and to a very luminous SN 2017hcc-like IIn/SLSN-II configuration extracted from HERACLES calculations (Wagle et al., 2023). These calculations illustrate the range of conditions SuperLite was designed for: homologous and non-homologous outflows, plateau-phase hydrogen-rich ejecta, shocked interaction regions with forward and reverse shocks, and very luminous interaction-powered events.

In the rotating PPISN study, SuperLite is used as the final detailed radiative-transfer step after MESA and STELLA. There it is described as “a 1D, multi-group, Monte Carlo radiative transfer code that post-processes a single snapshot profile obtained from any hydrodynamic evolution code,” applying IMC–DDMC to shell-collision profiles from 10 days before to 30 days after the light-curve peak. The specific calculation uses 6000 wavelength groups covering $O(v/c)$9, about $$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$0 Monte Carlo source particles, and a truncation at Rosseland mean optical depth $$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$1. The resulting spectra are H-free and dominated by He, C, O, Mg, and Ca, with broad C II and O I/II features at early times and stronger absorption features as the ejecta cool and recombine; the study concludes that shock-heated H-poor PPISN shell collisions from rapidly rotating progenitors can lead to moderately luminous H-poor transients that share some similarities with observed SLSN-I events (Huynh et al., 28 Jul 2025).

A later study applies SuperLite to detailed spectroscopic modeling of circumstellar interaction in both hydrogen-rich and hydrogen-poor superluminous supernovae, systematically varying circumstellar density, composition, and geometry. In that work, synthetic spectra generated with SuperLite show that hydrogen and helium lines vary strongly with circumstellar mass and composition; hydrogen-rich SLSN-II models exhibit pronounced hydrogen emission lines correlated with dense, extended circumstellar material, while hydrogen-poor SLSNe recover mostly featureless early spectra, with weak hydrogen lines appearing only in very early phases. The analysis emphasizes $$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$2 and $$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$3 as spectroscopic indicators of circumstellar characteristics and of the progenitor’s mass-loss history (Wagle et al., 2024).

6. Position in the software ecosystem, implementation, and outlook

SuperLite occupies a specific position among supernova and transient radiative-transfer codes. Relative to SuperNu, it is a steady-state post-processor and does not assume homologous expansion. Relative to CMFGEN, it uses Monte Carlo transport rather than a deterministic co-moving-frame solution of the radiative transfer equation, and it remains much less elaborate in microphysics because its default is LTE with experimental NLTE only for hydrogen. Relative to TARDIS, it does not assume homologous expansion or a Sobolev line treatment, and it is not tied to a sharp-photosphere formalism. These differences are particularly consequential in circumstellar-interaction problems, where shock structures, non-monotonic velocity fields, and line-by-line opacity effects are all spectroscopically important (Wagle et al., 2023).

The implementation is parallelized with MPI and OpenMP and released under GNU GPL v3 at https://github.com/gururajw/superlite. A representative benchmark in the code paper reports that a W7 SN Ia spectrum at 10 days with about $$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$4 packets completes in a few minutes on a typical multi-core desktop using 6 MPI ranks. The practical workflow is correspondingly simple: an external hydro or radiation-hydrodynamics code supplies the profile; the user truncates the inner region at an appropriate optical depth; SuperLite ingests the snapshot, assigns packet luminosities so that $$b(\nu) = \frac{15}{\pi^4}\frac{h}{kT}\frac{(h\nu/kT)^3}{e^{h\nu/kT}-1},$$5, and produces emergent spectra over the specified wavelength grid.

Its limitations are equally explicit. SuperLite is currently 1D and spherically symmetric; 2D and 3D support is planned. The default microphysics is LTE, and the NLTE implementation is presently restricted to hydrogen. The code is steady-state, so it does not evolve hydrodynamics or radiation fields in time and depends on external simulations for density, temperature, composition, and luminosity histories. The planned extensions—to 2D and 3D geometries, to more complete NLTE treatments, and potentially to coupling with a hydrodynamic solver—suggest a trajectory from rapid post-processing toward a broader radiative-transfer platform for interacting transients. A plausible implication is that SuperLite’s principal long-term value lies not only in individual spectral calculations, but in providing an open, extensible bridge between modern radiation-hydrodynamics models and observation-facing spectroscopy of circumstellar-interaction phenomena.