[Paper Review] H I spin temperature in the Fermi-LAT era

G. Jóhannesson, I. V. Moskalenko+1 more|arXiv (Cornell University)|Feb 1, 2010

Astrophysics and Cosmic Phenomena1 references3 citations

TL;DR

This paper proposes improved modeling of Galactic diffuse gamma-ray emission by replacing the traditional assumption of a constant H I spin temperature with spatially varying estimates derived from Fermi-LAT data and direct H I absorption observations. It demonstrates that using a linearly varying spin temperature model significantly improves fit quality over constant assumptions, with a log-likelihood ratio improvement of 28 in the Galactic plane, highlighting the critical role of accurate spin temperature treatment in cosmic-ray propagation modeling.

ABSTRACT

The diffuse high-energy gamma-ray emission of the Milky Way arises from interactions of cosmic-rays (CRs) with interstellar gas and radiation field in the Galaxy. The neutral hydrogen (H I) gas component is by far the most massive and broadly distributed component of the interstellar medium. Using the 21-cm emission line from the hyperfine structure transition of atomic hydrogen it is possible to determine the column density of H I if the spin temperature (Ts) of the emitting gas is known. Studies of diffuse gamma-ray emission have generally relied on the assumption of a fixed, constant spin temperature for all H I in the Milky Way. Unfortunately, observations of H I in absorption against bright background sources has shown it to vary greatly with location in the Milky Way. We will discuss methods for better handling of spin temperatures for Galactic diffuse emission modeling using the Fermi-LAT data and direct observation of the spin temperature using H I absorption.

Motivation & Objective

• To improve the accuracy of Galactic diffuse gamma-ray emission (DGE) modeling by addressing the limitations of assuming a constant H I spin temperature across the Milky Way.
• To evaluate the impact of spatially varying spin temperatures on DGE model fits using Fermi-LAT data.
• To assess the value of direct H I spin temperature measurements from absorption lines against bright radio sources for refining gas column density estimates.
• To quantify the improvement in model fit quality when incorporating observed spin temperatures in key Galactic regions, particularly the Galactic plane.
• To identify systematic uncertainties in spin temperature modeling and their implications for DGE and cosmic-ray propagation studies.

Proposed method

• Utilized the GALPROP code to simulate cosmic-ray propagation and diffuse gamma-ray emission, assuming a conventional model with CR injection spectra and diffusion parameters tuned to match local observations.
• Constructed Galactocentric annuli for gas distribution using the 21-cm LAB survey for H I and CO (J=1→0) survey for H₂, with H₂ column density derived via the X_CO factor.
• Applied the standard column density formula involving spin temperature (T_S) and brightness temperature (T), with correction for optical depth via the relation N_HI(v, T_S) = -log(1 - T/(T_S - T_bg)) * T_S * C.
• Tested multiple T_S assumptions: a constant global value (110–150 K in 5 K steps), a linearly varying model T_S = max(T_S,min, T_max + ΔT_S), and a hybrid model incorporating direct T_S measurements from 500 lines of sight.
• Performed full-sky maximum likelihood fits using the GaRDiAN package on Fermi-LAT data with enhanced background suppression, including the 1-year LAT source list and isotropic emission component.
• Compared model fits via log-likelihood ratio tests to evaluate statistical significance of improvements, especially in the Galactic plane region with high T_S observation density.

Experimental results

Research questions

• RQ1 How does assuming a constant H I spin temperature affect the accuracy of Galactic diffuse gamma-ray emission modeling with Fermi-LAT data?
• RQ2 Can a spatially varying spin temperature model, based on observed brightness temperatures, significantly improve the fit quality of the diffuse gamma-ray emission?
• RQ3 To what extent do direct H I spin temperature measurements from absorption lines against bright radio sources enhance the fidelity of gas column density estimates and DGE modeling?
• RQ4 What is the statistical significance of using observed T_S values in a constrained region of the Galactic plane (|b| < 10°, 15° < l < 165°) compared to global assumptions?
• RQ5 How do systematic uncertainties in gas distribution and CR propagation affect the inferred spin temperature values and model reliability?

Key findings

• The constant spin temperature assumption (T_S = 130 K) yields the best fit among global constant models, but this value is below the maximum observed brightness temperature (≈150 K), violating the physical condition T_S > T + T_bg.
• A linearly varying spin temperature model (T_S,min = 110 K, ΔT_S = 10 K) improves the model fit by a log-likelihood ratio of 1000 compared to the constant T_S model, indicating a highly significant improvement.
• Incorporating direct T_S measurements from 500 lines of sight results in a log-likelihood ratio of –105, indicating a statistically worse fit than the linear model, likely due to inconsistent T_S values and lack of adjustment in the model parameters.
• Fitting only in the high-coverage Galactic plane region (|b| < 10°, 15° < l < 165°) with observed T_S values yields a log-likelihood ratio improvement of 28, demonstrating a statistically significant enhancement despite only 25% of the region being covered.
• Residuals in the best-fitting models still show structured features, indicating that current models remain imperfect and that further refinement of gas distribution and T_S treatment is necessary.
• The study confirms that spin temperature variations significantly affect H I column density estimates, with non-linear corrections reaching up to a factor of 2 in regions where T ≈ T_S, underscoring the need for spatially resolved T_S modeling.