SED reproduction of NGC 3783, the first type-1 AGN with confirmed polar dust emission

During the last decade, mid-IR interferometry observations of several nearest AGNs reveal that a large fraction of their mid-IR emission comes from the polar dust component, rather than the equatorial torus. In Lyu & Rieke (2018), we show such a behavior can be easily reproduced by obscuring the intrinsic AGN templates with an extended distribution of large dust grains. An example is shown above for NGC 3783. With a similar model, we got consistent measurements of the polar dust emission strength for all type-1 AGNs with robust mid-IR interferometry analysis. In addition, we show that the SEDs of ~60 Seyfert-1 nuclei with AGN-dominated mid-IR emission can be reasonably reproduced by the same model.

A schematic picture of the dusty environment around a typically Seyfert-1 nuclei

With the great success of our simple model, we propose the dust environment around a typical Seyfert nucleus should be like to the figure above. In general, there are three major components: (1) a compact torus (<10 pc) with high optical depth; (2) an extended dust component (<0.1-1 kpc), which is optically-thin in the mid-IR and featured by a distribution of large dust grains; (3) the galactic ISM (>0.1 kpc), where the AGN has limited impact on its dust emission. We argue that the AGN narrow-line-region clouds alone are enough to produce the required extended dust distribution and the corresponding IR-reproduced emission would be observed as the polar dust emission.

Comparison of the properties of dust-free quasars with the low-redshift PG quasars

In Lyu, Rieke & Shi (2017), we show that the SEDs of "dust-free" quasars reported at z~6 (Jiang et al. 2010) do not show difference compared with the low-z Palomar-Green quasars. Although they have very weak IR emission, they are not "dust-free".

DIfferent AGN dust environment and the featured SED shapes

In Lyu et al. 2016, 2017a, 2017b, 2018, we have conducted a systematic research to reveal the connections behind the variations of AGN SEDs through a comparative study of local AGNs with high-z objects, covering broad ranges of luminosity (L_AGN~10^8 – 10^14 L⊙) and redshift (z~0–6). We have shown a “minimalist” two-free-parameter semi-empirical model is good enough to reconcile the SEDs of all major populations of type-1 AGNs, including local Seyfert-1 nuclei, normal blue quasars, extremely red quasars at z ∼2–3, AGNs with mid-IR excess emission at z ∼0.7–2, hot-dust-obscured galaxies at z ∼1.5–4, and dust-free quasars at z ∼ 6. In fact, the commonly seen UV-optical extinction, the unexpected strong mid-IR polar dust emission, and the poorly-understood type-1 AGN IR SED variations can be physically connected in a similar fashion. It is highly likely that most type-1 AGNs, regardless of their luminosities and redshifts, could share similar torus properties but differ mainly due to the extended (sub-kpc to kpc scales) polar dust component. By grouping similar SED behaviors together and summarizing observations in the literature, we have proposed that the AGN environment could be grouped into three general categories:

• Unobcured AGNs : these objects are consistent with the classical torus pimuch dust along the polar direction. Due to their compact size (less than 100 pc), the torus structures are controlled mainly by the BH accretion with possible intrinsic variations. Typical objects are optically-blue quasars;

• Moderately obscured AGNs : besides the small-scale torus, there is some extended distribution of polar dust over a few hundred-parsec scales, possibly driven by AGN winds or outflows. Some Seyfert-1 nuclei like NGC 3783 fall into this group;

• Highly-obscured AGNs : In these objects, the central engine can heat the galaxy ISM over kpc scales. The extended dust distribution has quite a different radial dansity profile andd much higher optical depth compared with typical Seyfert nuclei. Physically, we suggest the ISM is turbulent and clumpy, possibly casued by the collisions beinflows and AGN outflows. Good examples are hot dust-obscured galaxies.

Two populations of dust-deficient quasars

In Lyu, Rieke & Shi (2017), we show that the AGN intrinsic IR emission can be summarized into three major groups: (1) the normal AGNs whose SEDs are well-reproduced by the Elvis et al. (1994)-like template; (2) hot-dust-deficient AGNs who show much weaker emission from the near-IR to the far-IR compared with the normal cases; (3) warm-dust-deficient AGNs; although they present similar hot dust emission peaked around 3 microns, their mid-IR and far-IR is weaker than the normal cases. Similar behaviors are also seen at z~2-6, suggesting that the variations of the quasar IR emission SEDs are a common feature that is independent on redshift.

Proposed AGN far-IR SEDs in the literature

Over the past ~30 years, intensive efforts have been put on the characterization of AGN IR SEDs. However, there are a significant disagreements, especially in the far-IR. As shown above, the discrepancies can be as large as 1 dex. As a result, we have very poor constraints on the outer part of the AGN torus and doubtable measurements of the host galaxy far-IR star formation rates.

Comparison of different AGN intrinsic far-IR templates

In Lyu & Rieke (2017), we provided a critical analysis on the correct form of the AGN far-IR SED. Despite the variations in the near- to mid-IR bands, AGNs in quasars and Seyfert galaxies have remarkably similar intrinsic far-IR SEDs at λ ˜ 20-100 μm, suggesting a similar emission character of the outermost region of the circumnuclear torus.

Simulated NIRSpec spectrum for the quasar rest-frame optical emission at z~6

I have enrolled in several JWST GTO programs. In the figure above, I show the rest-frame optical emission of a typical bright quasar at z~6 as observed by JWST/NIRSpec. The blue line is the intrinsic AGN spectrum, while the black line is the AGN+galaxy spectrum. If we can reach enough S/N, multiple stellar features could be detected!

Example SED decompostions of quasars at z=5-6

In Lyu et al. (2016), we have developed a simple two-parameter model to decompose the quasar IR SEDs at z~5-7. We found that the IR emission of these quasar host galaxies cannot be reasonably matched by the normal star-forming galaxies at low-z (see the poor fits in the middle panel), but by the low-metallicity starburting galaxy like Haro 11 (the left panels). In fact, Haro 11 matches many properties of the first-generation massive galaxies (see De Rossi et al. (2018))

The correlation between the AGN luminosity and host galaxy IR luminosities from z=0 to z=6

With our simple SED model, we re-analyzed all z~5-7 quasars with far-IR constraints. We found that although the quasar host galaxy star formation rates have very large variations from ~10 M⊙/yr to ∼ 4000 M⊙/yr, the average infrared luminosity from star formation and the average total AGN luminosity of the z~5-7 quasar sample follow the correlation defined by quasars at z less than 2.6. With some simple assumptions, we estimate the stellar mass that could form in these z ∼ 5 − 6 quasar host galaxies to be M* ∼ (3 − 5) × 10^11 M⊙. Combining with the black hole (BH) mass measurements, this stellar mass is adequate to establish a BH-galaxy mass ratio M_BH/M∗ at 0.1-1%, consistent with the local relation.

Example optical and mid-IR light curves from a PG quasar

We have used the multi-epoch data from optical transient surveys and the mid-IR WISE/NEOWISE mission to probe the AGN dust structures. A pilot study of Palomar-Green quasars has been published as Lyu, Rieke & Smith 2019. This is the first mid-IR dust reverberation mapping survey of quasars and we provide the first multi-wavelength torus size constraints. You can get a poster here.