Compton-thick Obscuration in AGN

• Compton-thick obscuration is defined as an AGN state with hydrogen column density exceeding 1.5×10^24 cm⁻², leading to severe X-ray attenuation.
• Spectral diagnostics include a reflection-dominated continuum below 10 keV, strong Fe Kα fluorescence, and a pronounced Compton hump at 20–30 keV.
• Advanced broadband X-ray spectroscopy and torus models, combined with multiwavelength techniques, are critical to mapping AGN growth and explaining the cosmic X-ray background.

Compton-thick obscuration describes the most extreme form of nuclear veiling in active galactic nuclei (AGN), in which the line-of-sight hydrogen column density ($N_{\rm H}$) exceeds the inverse Thomson scattering cross-section, $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$. In this physical regime, not only photoelectric absorption but also multiple Compton scatterings effectively suppress the direct high-energy X-ray continuum, yielding reflection-dominated spectra and characteristic emission-line features. Accurate measurement and modeling of Compton-thick (CT) columns are essential for a complete census of AGN, determination of black hole growth histories, and explanation of the origin and shape of the cosmic X-ray background (CXB).

1. Physical Definition and Thresholds

The defining criterion for Compton-thick obscuration is set by the electron (Thomson) scattering optical depth: $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$ where $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$ is the Thomson cross-section. The Compton-thick threshold is conventionally marked at

$$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$

Above this column, the reprocessor is optically thick to electron scattering. For such columns, X-rays $\lesssim10$ keV are deeply attenuated, and even the 10–100 keV continuum is significantly reduced by Compton down-scattering and energy losses.

Compton-thick obscuration sharply distinguishes itself from the more prevalent Compton-thin regime ($N_{\rm H}\sim10^{22}$–$10^{24}$ cm$^{-2}$) by its photon suppression and spectral signatures.

2. Physical and Spectral Diagnostics

Observational Indicators

Compton-thick AGN manifest several distinctive X-ray spectral characteristics:

• Flat ("reflection-dominated") continuum below $\sim10$ keV: The direct transmitted AGN continuum is suppressed; residual emission is dominated by photons reflected from the far side of the torus or other Compton-thick structures. This yields a hard spectrum with effective photon index $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$0.
• Strong Fe K$N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$1 fluorescence at 6.4 keV: Equivalent widths (EW) of the emission line reach $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$2–$N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$3 keV and beyond, since the suppressed underlying continuum enhances the line/continuum contrast.
• Prominent "Compton hump": A broad excess peaking at $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$4–$N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$5 keV results from the cumulative effect of Compton down-scattering.
• Suppression of 2–10 keV luminosity: $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$6 ratios as low as $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$7–$N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$8 are observed in extreme cases (Boorman et al., 2016).

Spectral Modeling

Accurate quantification of $N_{\rm H}\gtrsim1.5\times10^{24}\,\mathrm{cm}^{-2}$9 and AGN intrinsic properties requires physically motivated torus models, such as MYTorus (Murphy & Yaqoob 2009), borus02 (Baloković et al. 2018), UXCLUMPY (Buchner et al. 2019), and BNTorus (Brightman & Nandra 2011). These frameworks self-consistently treat:

• Photoelectric absorption and energy-dependent attenuation,
• Compton-scattered continuum (angular and energy redistribution),
• Fluorescent line emission (Fe K$\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$0, K$\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$1),
• The geometry and covering factor of the reprocessing medium,
• Allow decoupling of line-of-sight and global average $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$2 to capture clumpy and patchy torus configurations (LaMassa et al., 2023).

The observed flux is typically modeled as: $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$3 where $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$4 parameterizes the reflected spectrum, and $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$5 (1–5%) models soft scattered or leaked emission.

3. Measurement Methodologies

X-ray Spectroscopy

• Broadband (0.3–150 keV) spectral fitting is critical. NuSTAR’s sensitivity in the $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$6 keV regime and combined data from Swift/BAT, XMM-Newton, Chandra, and Suzaku are generally employed to constrain heavily obscured sources (Ricci et al., 2016, Pal et al., 25 Sep 2025).
• Torus model parameters: Covering factor, opening angle, inclination, and both line-of-sight and global $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$7 become fit parameters in advanced models (Pal et al., 25 Sep 2025, Brightman et al., 2015).
• Fe K$\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$8 EW measurement: Large EWs unequivocally indicate reflection-dominated (i.e., Compton-thick) spectra (Boorman et al., 2016, Boorman et al., 2024).
• Luminosity proxies: Isotropic indicators such as [O III] $\tau_{\rm T} = N_{\rm H}\,\sigma_{\rm T}$9 and [O IV] $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$0m line luminosities, as well as mid-IR continuum, provide independent estimates of intrinsic luminosity for sources with deeply buried X-ray continua (0909.3044, Vignali et al., 2010, Vignali et al., 2014).

Submillimeter and Molecular Tracers

A complementary approach uses sub-mm ALMA CO(3–2) mapping to derive nuclear molecular column densities: $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$1 with $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$2 cm$\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$3 (K km s$\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$4)$\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$5 (Musa et al., 26 Feb 2025). However, optically thick CO, multi-phase gas, and beam dilution can lead to $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$6 as measured by X-rays, especially in dense CTAGN.

Multiwavelength Diagnostics

• X-ray to [NeV] and to $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$7m ratios have been shown to robustly distinguish CT sources at intermediate and high redshift (Vignali et al., 2014, Vignali et al., 2010).
• SED decomposition at UV–FIR wavelengths is required to separate AGN and starburst contributions, especially in ULIRGs and highly obscured galaxies (Nardini et al., 2011, Lanzuisi et al., 2015).

4. Prevalence, Demographics, and Host Properties

Local Universe

• Observed CT fractions in hard X-ray (Swift/BAT, 14–195 keV) selected samples are typically $\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$8–$\sigma_{\rm T}=6.65\times10^{-25}\,\mathrm{cm}^2$9 (Ricci et al., 2016, Pal et al., 25 Sep 2025).
• After bias correction (to account for sources missed even above 10–20 keV), the intrinsic CT fraction rises to $$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$0–$$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$1 of local AGN (Akylas et al., 2024, Ricci et al., 2016, Georgantopoulos, 2012).
• Mid-IR selected samples recover higher CT fractions and reveal that even ultra-hard surveys miss a large number of CT AGN (Akylas et al., 2024, Georgantopoulos, 2012).
• Host galaxies: CT AGN are preferentially found in massive, star-forming galaxies, frequently with disturbed morphologies or bars, and often exhibit higher merger fractions than unobscured AGN hosts (Lanzuisi et al., 2014, Lanzuisi et al., 2015). CT AGN tend to have smaller black hole masses and higher Eddington ratios than type-1 AGN at matched luminosity (Lanzuisi et al., 2014).

Cosmic Evolution

• Intermediate redshifts ($$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$2): Space density of CT AGN ($$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$3) is $$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$4 Mpc$$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$5, consistent with XRB synthesis model predictions (Vignali et al., 2014).
• Fractional contribution and evolution: The CT-to-highly-obscured AGN fraction is $$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$6\%, with no statistically significant redshift evolution documented out to $$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$7 (Li et al., 2019).
• High-luminosity regime: CT fractions decline with increasing AGN luminosity, a trend consistent across both local (Ricci et al., 2016, Brightman et al., 2015) and higher-$$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$8 samples.

5. Biases, Limitations, and Uncertainties

Detection Biases

• Energy-dependent bias: Below 10–20 keV, even modestly CT columns ($$N_{\rm H} \geq \sigma_{\rm T}^{-1} \approx 1.5 \times 10^{24}\,\mathrm{cm}^{-2}$$9 cm$\lesssim10$0) attenuate the continuum by >90% (Akylas et al., 2024, Georgantopoulos, 2012). Only broadband or $\lesssim10$1 keV coverage allows robust CT classification.
• Sample selection: Flux-limited and hard X-ray surveys systematically undercount the CT fraction, especially for the deepest and most luminous columns ($\lesssim10$2 cm$\lesssim10$3) (Akylas et al., 2024, Baloković et al., 2014, Lanzuisi et al., 2015).
• Host contamination and line-of-sight variance: Selection via [O III], [O IV], or mid-IR lines can be confounded by host extinction or dilution, while clumpy or non-axisymmetric torus structures may cause global $\lesssim10$4 to exceed the line-of-sight value, obscuring a true census (LaMassa et al., 2023).

Physical Complexity

AGN tori are inhomogeneous, often clumpy and patchy rather than uniform. Decoupling of line-of-sight and average global column densities is routinely observed in advanced torus modeling. Some AGN show reflection from Compton-thick medium globally, even when the sightline is only Compton-thin. Additional physical processes—such as beaming, extended scattering, or partial covering by discrete clouds—challenge the efficacy of simple "screen" models (LaMassa et al., 2023, Yaqoob, 2012).

Model Assumptions

Most torus models make assumptions regarding geometry, element abundances, and illumination. Uncertainties in iron abundance or the CO-to-$\lesssim10$5 conversion factor can affect gas mass determinations (Musa et al., 26 Feb 2025, Boorman et al., 2016). Properly accounting for these factors, and for orientation and multi-phase gas, remains a principal modeling challenge.

6. Broader Implications for Black Hole Growth and the CXB

• Cosmic X-ray Background: CT AGN are required to explain the $\lesssim10$6 keV peak of the CXB. Population-synthesis models require a moderate CT AGN fraction ($\lesssim10$7–30%) to match the observed CXB spectrum (Ricci et al., 2016, Vignali et al., 2014).
• Supermassive Black Hole Accretion: Theoretical models and observed demographics suggest that a substantial—possibly dominant—fraction of black hole growth occurs during Compton-thick phases (Boorman et al., 2024). These are short-lived, heavily obscured stages characterized by high Eddington ratio accretion and associated multiphase outflows (Lanzuisi et al., 2014, Lanzuisi et al., 2015).
• Evolutionary Context: In starburst-rich environments such as ULIRGs, AGN are often completely Compton-thick, with strong links between the intensity of star formation and AGN burial. Such systems are caught at transitional stages in the co-evolution of galaxies and their supermassive black holes (Nardini et al., 2011, Lanzuisi et al., 2014).

7. Future Prospects and Methodological Advancements

• Instrumental advances: Next-generation hard X-ray missions (e.g., HEX-P, Athena, Lynx) offering sub-arcsecond resolution and true $\lesssim10$8 keV sensitivity will significantly enhance the detection efficiency for both faint and extreme CT AGN (Boorman et al., 2024).
• Statistical methodologies: Adoption of probabilistic and machine-learning based obscuration predictors, combining multiwavelength flux ratios, color diagnostics, and spectral hardness, improves the selection and pre-classification of candidate CT AGN (Pal et al., 25 Sep 2025).
• Refined sub-mm approaches: Use of optically thin isotopologues (e.g., $\lesssim10$9CO, C$N_{\rm H}\sim10^{22}$0O), dense gas tracers (HCN, HCO$N_{\rm H}\sim10^{22}$1), and high-resolution beam-matched studies are recommended to break degeneracies in $N_{\rm H}\sim10^{22}$2 estimation from molecular emission (Musa et al., 26 Feb 2025).

A robust census of Compton-thick obscuration remains critical for constraining AGN unification models, cosmic black hole growth, and the energetics of the obscured universe. Combining advanced broadband X-ray spectral models, sensitive multi-wavelength diagnostics, and unbiased selection functions is essential for progress.