Seyfert Galaxy

What a Seyfert is

A Seyfert galaxy looks at first glance like an ordinary spiral — a disk, arms, dust lanes, a bulge — until you point a spectrograph at the nucleus. There, instead of the absorption-line spectrum of an old stellar population, you see strong, narrow emission lines from highly ionised gas: [O III] 5007, [N II] 6584, [S II], [Ne III], [Ne V], all far stronger than star-forming HII regions produce. In the brighter Type 1 systems, the narrow lines are joined by broad permitted lines (Hα, Hβ, Mg II) with widths of thousands of km/s and a strong featureless blue continuum. The optical spectrum is the diagnostic; everything else — radio, infrared, X-ray — corroborates.

The energy source is accretion onto a central supermassive black hole. The typical Seyfert SMBH mass is 10⁶–10⁸ M_sun; the typical bolometric luminosity is 10⁴³ – 10⁴⁵ erg/s, or 10⁹ – 10¹¹ L_sun — much less than a quasar but enough to outshine the host galaxy's nuclear region.

Type 1 versus Type 2

Property	Seyfert 1	Seyfert 2
Broad permitted lines	Yes (FWHM 1000-10 000 km/s)	No (in total flux)
Narrow forbidden lines	Yes	Yes
UV/optical continuum	Strong, featureless	Weaker, reddened
X-ray absorption (N_H)	< 10²² cm⁻²	10²² – 10²⁵ cm⁻² (often Compton-thick)
Polarised broad lines (Type 2)	—	Often present — hidden BLR
Viewing geometry	Down the torus axis (face-on)	Through the torus (edge-on)
Variability	Strong, on days-years	Modest, mostly in NLR
Example	NGC 4151, NGC 5548	NGC 1068, NGC 4945

Sub-types Sy 1.2, 1.5, 1.8, 1.9 mark systems where the broad component is partially obscured or intrinsically weak. A Sy 1.5 shows clear broad Hα and broad Hβ; a Sy 1.9 shows broad Hα only; a Sy 1.8 has both but with broad Hβ very weak. The continuous sequence is consistent with the unified-model picture of progressively heavier obscuration along the line of sight.

Anatomy of the central engine

Modern reverberation mapping and unified-model fits have resolved the central region into a stack of nested components:

• Supermassive black hole. 10⁶ – 10⁸ M_sun; sets the gravitational potential and the ISCO that launches the disk emission.
• Accretion disk. Geometrically thin, optically thick, 10⁻⁶ – 10⁻³ pc, T = 10⁴ – 10⁵ K. Produces the "big blue bump" UV continuum.
• Hot corona. ~ 10 gravitational radii from the SMBH; kT_e ~ 10⁹ K. Produces the 2–100 keV X-ray power law via inverse Compton scattering of disk photons.
• Broad-line region (BLR). Dense photoionised clouds at 10⁻³ pc orbiting at thousands of km/s. Produces broad permitted lines visible only in Type 1.
• Dusty torus. Molecular gas and dust at ~ 1 pc. Mid-IR thermal bump; X-ray absorber; obscurer in Type 2. Likely clumpy rather than smooth.
• Narrow-line region (NLR). Lower-density (n_e ~ 10³ – 10⁶ cm⁻³) gas at ~ 100 pc, photoionised by an unobscured fraction of the central continuum escaping through the torus opening. Forbidden lines, FWHM 100–1000 km/s.

NGC 1068 and the unified-model test

In 1985, Antonucci and Miller measured spectropolarimetry of NGC 1068, the prototypical Seyfert 2. In total flux the spectrum had no broad lines. But the polarised flux spectrum — selecting photons that had been scattered before reaching us — clearly showed broad Hα, Hβ, and Mg II lines with FWHM ≳ 3000 km/s. The interpretation, now textbook, is that NGC 1068 hides a Seyfert 1 inside its core: the BLR is there, but blocked from direct view by an obscuring torus; UV photons and broad-line photons escape upward through the polar opening, scatter off material above the torus, pick up linear polarisation by the scattering geometry, and reach us as polarised light revealing the hidden engine.

This was the canonical evidence that elevated orientation from one possible factor to the dominant axis of AGN diversity. Hidden BLRs have since been detected in many other Seyfert 2s (e.g. NGC 1068, NGC 7674, IC 5063, NGC 4388) by spectropolarimetric surveys. A small fraction of Seyfert 2s show no polarised broad lines even in deep observations — the "true Type 2" candidates, suggesting the BLR is not a universal AGN feature but an accretion-rate phenomenon.

Identifying Seyferts: the BPT diagram

The Baldwin-Phillips-Terlevich (1981) diagram plots log [O III] 5007 / Hβ against log [N II] 6584 / Hα, both ratios of nearby emission-line pairs (and therefore independent of reddening). Star-forming HII regions occupy the lower-left "starburst" locus; AGN occupy the upper-right "Seyfert" region; an intermediate "LINER" region houses lower-ionisation nuclear emission-line regions. Kewley et al. (2001) and Kauffmann et al. (2003) drew theoretical and empirical boundaries that the modern AGN-classification literature uses.

BPT classification can be done on a single galaxy's nuclear spectrum or on integral-field maps that resolve where the AGN ionisation extends. SDSS and MaNGA have catalogued hundreds of thousands of Seyfert galaxies this way, and the BPT diagram is the workhorse of AGN demography at low redshift.

Famous Seyferts

• NGC 1068 (M77, Sy 2). Type 2 archetype with hidden BLR. Mid-IR resolved torus; H₂O megamaser disk pinning SMBH mass at 1.7 × 10⁷ M_sun.
• NGC 4151 (Sy 1.5). Textbook Type 1; first AGN with reverberation-mapped BLR (Hβ time lag ~ 6 days → R_BLR ~ 1.6 × 10⁻³ pc); SMBH mass ~ 4 × 10⁷ M_sun.
• NGC 5548 (Sy 1). Decades of multi-wavelength monitoring; cornerstone of reverberation-mapping campaigns (AGN Watch, AGN STORM).
• NGC 1275 / 3C 84 (Sy 1.5). Sits at the centre of the Perseus Cluster; radio-loud; embedded in the cluster's cool-core X-ray bubble structure.
• NGC 7469 (Sy 1). Nearby Type 1 with a circumnuclear ~ 1 kpc star-forming ring; classic case of AGN/starburst coexistence.
• NGC 4945 (Sy 2). Compton-thick edge-on Seyfert 2 — extremely heavy X-ray absorption (N_H ~ 4 × 10²⁴ cm⁻²); ALMA-resolved molecular torus.

Changing-look Seyferts

A subset of Seyferts flip between Type 1 and Type 2 on timescales of months to years. Mrk 1018 dropped from Sy 1 to Sy 1.9 over a decade. NGC 2617 turned on its broad lines in 2013. 1ES 1927+654 had a spectacular 2018 event where the X-ray corona vanished, came back, and the optical spectrum changed accordingly. The torus cannot reorient on these timescales (the reorientation timescale is ~ 10⁵ yr). Instead, the leading interpretation is an accretion-rate transition: when ṁ = Ṁ/Ṁ_Edd drops below a threshold (~ 10⁻²), the BLR fades because the inner disk lacks the photons or gas to sustain a broad-line-producing structure. Changing-look events therefore refine the unified model by adding accretion rate as an explicit second axis on top of orientation.

Worked example — broad Hβ FWHM and SMBH mass

The reverberation-mapping virial estimator is M_BH = f × R_BLR × ΔV² / G, where R_BLR is the BLR radius (from light-travel time-lag between continuum and emission-line variations), ΔV is the broad-line velocity (from FWHM), and f ≈ 4.3 is a geometric factor calibrated against the M-σ relation.

For NGC 4151, R_BLR(Hβ) = 6.6 light-days = 1.7 × 10¹⁴ m, and the broad Hβ FWHM corresponds to ΔV ≈ 5800 km/s. Then

M_BH ≈ 4.3 × (1.7 × 10¹⁴ m) × (5.8 × 10⁶ m/s)² / (6.674 × 10⁻¹¹) ≈ 3.7 × 10⁷ M_sun.

This sits within a factor of a few of the more direct stellar-dynamical measurement of ~ 3 × 10⁷ M_sun, confirming the technique. Reverberation mapping in dozens of Type 1 Seyferts establishes the R_BLR–luminosity relation that lets a single optical spectrum yield an SMBH mass for any well-measured Type 1.

Brief history

Carl K. Seyfert in 1943 spotted strong nuclear emission lines in 12 nearby spirals — NGC 1068, NGC 1275, NGC 3516, NGC 4051, NGC 4151, NGC 7469, and others — and proposed them as a distinct class. The work sat fallow until the 1960s, when quasars were discovered and the AGN field exploded. By the 1970s Seyferts were understood as low-luminosity AGN, and X-ray and radio surveys began catalogues of them. Antonucci and Miller's 1985 NGC 1068 paper grounded the unified model. Reverberation-mapping campaigns from the 1990s onward built the R_BLR–luminosity relation. NICER, Chandra, NuSTAR and the upcoming Athena mission continue to refine our X-ray view of Seyfert obscuration and coronae.

Common pitfalls

• Equating Seyfert 2 with "a different engine." Most Seyfert 2s have hidden BLRs and the same central engine as Seyfert 1s; the obscuring torus is what differs. Treating them as physically distinct biases AGN demographics.
• Reading every "narrow-line-only" spectrum as a Type 2 Seyfert. LINERs and low-ionisation AGN have weaker [O III]/Hβ and require a careful BPT diagnostic; not every narrow-line nucleus is Seyfert-class.
• Calling all spirals with AGN "Seyferts." Some spirals host LINERs, low-luminosity AGN, or even mini-quasars; only the BPT-defined Seyfert region counts.
• Assuming Seyferts are always radio-quiet. Most are, but a handful (NGC 1275, 3C 120, NGC 1052) are radio-loud and sit on the unified picture as 'radio-loud Seyferts' or low-luminosity radio galaxies.
• Using broad Hα alone to claim Type 1. Stellar winds, supernova remnants, and other sources can produce broad Hα. Multiple permitted lines plus a strong featureless continuum are needed.
• Treating changing-look events as torus-reorientation. Months-to-years type flips are accretion-rate transitions, not geometry changes.