THEMIS Dust Evolution Framework

• THEMIS is a comprehensive framework that models the evolution, composition, and observational signatures of interstellar dust across diverse astrophysical environments.
• It employs detailed dust components such as amorphous silicates, hydrogenated amorphous carbons, and core-mantle structures, leveraging laboratory data and computational methods like Mie theory.
• The model accurately reproduces extinction, emission, and polarization trends in the diffuse ISM, dense clouds, and star-forming regions, validated against multiwavelength astronomical observations.

THEMIS (The Heterogeneous dust Evolution Model for Interstellar Solids) is a comprehensive, physically-motivated framework for modeling the evolution, composition, and observational signatures of interstellar dust in a wide range of astrophysical environments. THEMIS supersedes empirical or ad hoc models by integrating laboratory-derived optical constants, well-constrained dust mineralogy, explicit treatment of evolutionary processes, and predictive calculations directly comparable with multiwavelength astronomical data. The framework is designed for extensibility and self-consistency, offering tools for modeling extinction, emission, and polarization from the diffuse interstellar medium (ISM) through dense molecular clouds into star-forming regions.

1. Structural Components of the THEMIS Dust Model

The THEMIS core is built on a physically justified inventory of dust populations distinguished by composition, structure, and grain size distribution:

• Amorphous Silicates (a-SilFe,FeS): Grains comprise a 1:1 mass mixture of olivine-type and pyroxene-type amorphous silicates with $7\%$ metallic Fe and $3\%$ FeS nano-inclusions (by volume). This composition is motivated by both laboratory data and depletion patterns in the ISM.
• Hydrogenated Amorphous Carbons (a-C(:H)): THEMIS features an ensemble of hydrogenated amorphous carbon materials spanning the full range from H-rich (aliphatic; wide band gap) to H-poor (aromatic; narrow band gap), with their optical constants derived from extended Random Covalent Network (eRCN/optEC(s)(a)) models constrained by laboratory measurements.
• Core–Mantle Structures: Large silicate grains are modeled as cores with an amorphous carbon mantle ($5$–$10\;\mathrm{nm}$), while large a-C(:H) grains carry an UV-processed outer a-C mantle ($7.5$–$20\;\mathrm{nm}$). A distinct nanoparticle population of a-C grains ($a \lesssim 20 \;\mathrm{nm}$) is also treated.

Grain size distributions are characterized by:

• Nano-particles: $n(a) \propto a^{-\alpha}$, $\alpha \approx 5$, for $a$ up to $20\;\mathrm{nm}$.
• Large grains: Log-normal forms, e.g., $n(a) = C \exp[-(\ln(a/a_0))^2 / (2\sigma^2)]$ with $a_0 \approx 140$–$160\;\mathrm{nm}$, $\sigma \approx 0.4$–$0.5$ dex.

Optical properties are determined using:

• Complex refractive indices $m(\phi, \lambda) = n(\lambda) + i k(\lambda)$, sourced from laboratory or eRCN models for each material type.
• Cross-section computation: $C_i(a, \lambda) = \pi a^2 Q_i(a, \lambda)$, with $Q_i$ (for extinction, absorption, scattering) from Mie theory (homogeneous/coated spheres) or DDSCAT (core–mantle/aggregates).

2. Physical Processes and Environmental Dependence

THEMIS incorporates essential dust processing mechanisms as a function of environmental parameters:

• Diffuse ISM: The default model reproduces observed extinction curves UV to millimeter, including the 217\,nm bump, FIR/submm emission, and low albedo. Small a-C grains dominate FUV extinction and mid-IR emission; large CM grains dominate in optical/NIR extinction and FIR emission.
• Dense Regions: As density increases:
  • Accretion: Gas-phase C (and H) forms new a-C:H mantles (CMM) on large grains; nano-particles aggregate onto larger grains, reducing small-grain numbers.
  • Coagulation: CMM grains form aggregates (AMM), gain ice mantles (AMMI) at $T \lesssim 20$ K and $A_V \gtrsim 3$. This boosts FIR emissivity by factors of $2$–$4$, decreases typical dust temperatures, steepens the sub-mm spectral index, and enhances scattered light (“cloud-shine”/“core-shine”).
• PDRs and HII Regions: UV photolysis aromatizes a-C(:H), eroding nanoparticles, altering optical properties, and releasing small hydrocarbons and H$_2$. The photoelectric effect on a-C(:H) nano-grains affects gas heating and charge states.

THEMIS accounts for the full lifecycle hysteresis: destruction and renewal routes in harsh environments are not the inverse of accretion/coagulation in shielded regions.

3. Model Implementation, Computational Methods, and Output

THEMIS is realized via the DustEM modeling tool, which orchestrates all subcomponents:

• Materials library: Manages complex refractive indices across wavelength grids.
• Size-distribution/aggregation modules: Parameterizes and evolves $n(a)$ for each population; supports accretion/coagulation prescriptions based on environmental input.
• Optical property solvers: Implements Mie theory for spherical/coated grain populations and DDSCAT for aggregates or non-spherical grains. Core–mantle structure, nano-inclusions, and shape/porosity variations are explicitly handled.
• Radiative solver: Computes the thermal balance and resulting spectral energy distributions (SEDs), with stochastic heating for small grains and equilibrium temperatures for large grains.

Input requirements:

• Interstellar radiation field (Mathis-MMP83 or user-specified).
• Gas density $n_H$, temperature, radiation field intensity scaling $\chi$.
• Elemental abundances: C, O, Si, Mg, Fe, S.

Output diagnostics:

• Extinction curves $A(\lambda)/N_H$, albedo, phase function.
• Emission SEDs ($I_\nu(\lambda)$) across UV-cm range.
• Dust temperature distributions $T(a)$, charge state distributions.

4. Model Validation, Observational Comparison, and Utility

THEMIS validation is based on a range of multiwavelength diagnostics:

• Diffuse ISM: Matches Planck-IRAS correlations, FIR/sub-mm SEDs [Ysard et al. 2015, 2016], and the mean Milky Way extinction curve ($R_V=3.1$), including the UV slope and 217 nm bump [Jones et al. 2013].
• Dense clouds: Accurately reproduces the transition in SEDs from $3\,\mu$m to $500\,\mu$m as well as scattered light phenomena (“coreshine”).
• Comparative strengths: Unlike MRN [Draine & Li] or PAH+silicate models (ZA11/C11), THEMIS uses laboratory-derived optical constants and realistic mixed chemistry rather than empirical “astronomical” values, and more faithfully tracks trends across extinction, emission, polarization, and depletion in varied environments.

Environment	Diagnostic	THEMIS performance/claim
Diffuse ISM	$A(\lambda)/N_H$, SED	Matches $<10\%$ up to mm; UV bump accurate
Dense cloud	SED, coreshine	Accurately reproduces SED flow, scattered light
PDR/HII	Aromatic features, H$_2$	Predicts aromatization, hydrocarbon release

5. Ongoing Developments and Future Directions

THEMIS continues to expand along several axes:

• DustPedia integration: THEMIS serves as the reference dust module in the DustPedia project, enhancing cross-comparison of dust modeling and galaxy SED databases.
• Polarization and alignment: Advances in treating grain alignment/polarization for core–mantle and aggregate populations are underway.
• Improved long-wavelength optical constants: Laboratory programs are providing improved $n,k$ data at $\lambda \gtrsim 60\,\mu$m.
• Time-dependent and coupled chemistry: Extension to dynamic accretion/coagulation and explicit coupling to gas-phase chemistry.
• Expanding into PDRs, HII regions, and star-formation contexts: Current developments target the aromatization pathways, destruction and renewal in harsh UV environments, and integration with star-formation models.

6. Resources, Adoption, and Community Use

• The THEMIS model, documentation, and DustEM input files are publicly accessible at THEMIS and DustEM.
• Researchers may download the complete input suite, specify custom environmental parameters, and compute self-consistent extinction/emission properties over the UV–cm spectral domain.
• THEMIS is widely employed in ISM studies and has been foundational for Planck and Herschel dust analyses, with continued validation against new observations and laboratory constraints.

Key references: Jones et al. 2013, 2014, Köhler et al. 2015, Ysard et al. 2015, 2016, Jones 2012a, 2012b (Jones et al., 2017).