Chandra Grating Spectroscopy of Active Galactic Nuclei Tahir Yaqoob (JHU/GSFC) Collaborators: B. McKernan (UMD), C. Reynolds (UMD), I. M. George (UMBC/GSFC), T. J. Turner (UMBC/GSFC), J. N. Reeves (USRA/GSFC), P. J. Serlemitsos (GSFC), R. F. Mushotzky (GSFC), S. B. Kraemer (CU/GSFC), M. Crenshaw (GSU), J. Gabel (CU/GSFC), U. Padmanabhan (JHU), V. Karas, M. Dovciak (Charles U., Prague), A. Markowitz (GSFC)

Overview  The Chandra grating data and general overview of salient results.  Results from a Chandra HETG sample of Sy 1 galaxies.  Properties of the photoionized outflows & a comparison with other results including XMM RGS. Location of the wind.  Gravitationally redshifted lines. Extreme outflows. Absorption in the IGM.  Emission lines, including Fe K.  What can Suzaku do for AGN winds?  A HETG AGN database of ready-to-go products, suitable for grating non-experts (and experts).

Gratings: best resolution at lowest energies  LETG: ~240 km/s at 0.2 keV [0.16 eV]  MEG: ~280 km/s at 0.5 keV [0.47 eV]  RGS: ~290 km/s at 0.3 keV [0.29 eV] E (keV) V

X-ray Absorption in Ionized Gas in Seyfert 1 Galaxies Before Chandra & XMM, CCDs had best spectral resolution (FWHM ~30, ,000 km/s in soft X-rays, 7,500 at Fe-K energies). ~50% of Sy 1 known with complex (likely photoionized) X-ray absorption. Individual O edges claimed but could have been confused. Chandra & XMM gratings, with FWHM down to ~300 km/s, revolutionalized study of “warm absorbers”, wealth of complexity, unresolved absorption lines, some emission lines. Pre-1999 View

Number of AGN with Chandra Grating Observations

What Do We See? Mkn 509 Chandra HETG+HST STIS campaign [Yaqoob et al. 2003] Mostly He-like, H-like unresolved absorption lines; in most cases blueshifted w.r.t. to systemic (~ km/s FWHM).

Derived column density from X-ray (Chandra LETGS) data versus the ionization parameter for which that ion has the maximum column density. NGC 5548 (Steenbrugge et al. 2003)

Dust vs. Relativistic Lines MCG soft X-ray spectrum: (Left) Lee et al Dusty Warm Absorber; (Right) Relativistically broadened soft X-ray lines, K. Mason & others, Soft X-ray broad lines also claimed in NGC 4051 (Ogle et al. 2004), NGC 5548 (Kaastra et al. 2002), MCG (Salvi et al. 2003), NGC 4593 (Steenbrugge et al. 2003)

MCG Chandra HETG ~500 ks “long-look” (Young et al. 2005)  Fe K absorption features found (~2000 km/s outflow)..absorber model for broad Fe K line predicts Fe K absorption features not found?  Caution..the Fe K absorption features found could be at z~0.

Ton S 180 LETG/ACIS  Weak absorption features..alternative models proposed: No warm absorber plus relativistic lines (Turner et al. 2002); Stratified warm absorber (with identified weak lines) plus disk emission lines (Rozanska et al. 2004).  Very few LETG/ACIS observations. More LETG/HRC observations (but orders can’t be separated), but still much less than HETG/ACIS. Rozanska et al. 2004Turner et al. 2002

X-ray/UV connection X-ray & UV absorbers share kinematics Much higher resolution in UV - why discrete components? Hints of velocity structure in some X-ray profiles, details elusive. Very different UV & X-ray columns (but e.g. Arav et al. ‘04: UV columns much larger?). Two-phase? UV “knots” in an X- ray wind? From Kaspi et al. 2002, & Crenshaw, Kraemer & George 2003, AR&AA, 41, 117 Details of a physical model need to be filled in. Multi-temperature wind (cf. Krolik & Kriss 2001)? What drives it & what is the source of the matter?

A Chandra HETG Sample of Seyfert 1 Galaxies What we want to know:  Where does the outflow originate?  What is its size? Geometry?  Physical, thermal, & ionization structure? Kinematics?  Source of material in the wind?  What determines properties of the outflow?  What is the nature of the connection between the outflow and the central black hole plus accretion disk?  Covering factor/filling factor?  What drives the outflow?  Mass outflow rate (compare to accretion)? Urry & Padovani 1995 Test for correlations between outflow properties & other key properties of the central engine.

A Chandra HETG Sample of Seyfert 1 Galaxies Selection:  Low z (<0.05) Seyfert 1 galaxies  Public HETG data up to 1 July, 2003  Bright (HEG > 0.05 ct/s)  -> 15 AGN; 10/15 exhibit signatures of ionized absorption. (Detailed results in McKernan, Yaqoob, Reynolds 2004, 2005, 2006) What we can measure:  Columns, ionization parameters from modeling; offset velocities of absorption lines; crude kinematic information (widths, profiles).  If sufficient emission-line data, density. Cannot get distance from center without density or variability information.  UV absorption: 4 of the 15 have simultaneous UV/Chandra data. How are UV/X-ray columns, ionization parameters, covering factors and kinematics related?

The HETG Seyfert 1 Sample

Correlations with Outflow Velocity & L ION /L EDD  V appears to have a bimodal distribution. Anti-correlated with L ION /L EDD ?  No correlation of N with L ION /L EDD. Possible anti-correlation of  with L ION /L EDD. NN NVNV V (km/s) L ION /L EDD McKernan, Yaqoob, & Reynolds 2006

Correlations with Black Hole Mass & L ION /L EDD  Appears to be no correlation between mass outflow rate and black-hole mass.  Possible correlation of mass outflow rate with L ION /L EDD. M BH /M SOLAR L ION /L EDD McKernan, Yaqoob, & Reynolds 2006 M OUTFLOW [ 0.01 C M SOLAR / year ]

Blustin et al. (2005) Study

Blustin et al. (2005) study: outflow rates These mass ouflow rates are LARGE, often greater than the accretion rate..but the absolute outflow rates depend on assumptions (Ne, covering & filling factors, and distance to the ionizing source have to be inferred, usually indirectly, based on assumptions which may or may not be true).

Blustin et al. (2005) Warm Absorber Location

Comparison with Blustin et al. Conclusions  Low filling factor (<8%), derived assuming momentum in outflow ~ momentum in radiation intercepted.  Minimum distance of warm absorber calculated assuming outflow exceeds escape velocity - not necessarily true. Leads to conclusion that the absorbers are located at ~ distance of torus or greater.  Kinetic energy of outflow may only be 1% of total energy budget, but so is the accretion rate. If the accretion rate is small yet important for the AGN then the outflow could be too.  Another caveat: Some of the outflow (e.g. closer in to center) may not be observable (e.g. too ionized).

Location of the Wind  Need variability information (continuum and opacity) to constrain the density and hence the location of the photoionized absorber. Any other method of deriving location makes at least one assumption, which may or may not be true.  Early variability studies (e.g. NGC 3227 Ptak et al. 1994; MCG , Reynolds et al. 1995; NGC 4051, McHardy et al. 1995; MR 2251, Otani et al. 1995; NGC 3516, Netzer et al. 2002) with low-resolution (CCD, or PSPC) spectra are difficult to interpret since the spectra are now known to be very complex.  Estimates of the location of the “warm absorber” varied from upper limits of parsecs to the distance of the BLR (e.g. NGC 3516).  From different arguments, location of the very high velocity outflows (e.g. PG ), placed at <0.1pc (more later).  S/N ratio of the higher resolution grating spectra has been too low for time-resolved spectroscopy..except for the extended Chandra HETG observation of NGC 3783…

Opacity Variations in the Fe UTA in NGC A shift in UTA, x2 change in U with x2 change in L -> photoionization eqm., Ne>10 4 cm -3, d 10  Krongold et al. 2005

“Picture” and Structure of the Wind Does the wind originate from the accretion disk (e.g. Elvis 2000, 2003)? Or does it originate in material blown off from the inner edge of the torus (e.g. Krolik & Kriss gas exists in a multi-phase, multi-temperature state in pressure balance but on the thermally unstable part of the cooling curve.)? Mass flux could be large. Large scale cones (of the type directly observed in Sy 2) [e.g. Kinkhabwala 2002]. High resolution X-ray spectra have lead to the view of a wide, radially continuous range of ionization stages of ions. Krongold UTA variability result rules out continuous ionization stage distribution, favoring high degree of clumping. Also rules out kpc scale cone scenario.

E (z=0.297) Evidence for a gravitationally redshifted absorption line, and one the highest redshift broad Fe-K emission lines. Yaqoob & Serlemitsos, 2005

Redshifted Absorption Lines in PG  Redshifted absorption lines at 4.56 & 5.33 keV rest-frame (z=0.0809); ~1.3 x chance probability. FWHM<8000 km/s.  If Fe XXVI, v~0.26c & 0.40c (0.22 & 0.37c for Fe XXV).  Column density is ~4 (+4,-2) x cm -3, but Fe K edge depth < 0.1. log   Pure gravitational shift?[R<6Rg]; Inflow/Failed outflow? Or gravitationally redshifted low-velocity outflow?

Extreme Outflows XMM CCD detections of very high- velocity outflows deduced from (blueshifted) Fe-K absorption lines from highly ionized Fe in some QSOs (Pounds, Reeves et al ). PG : v~ 24,000 km/s PG : v~ 60,000 km/s High optical depths: N H ~ 5 x cm -2 or more..but model-dependent. Other examples: PDS 456 (v~55,000 km/s). These are moderately low redshift QSOs (z<0.2). NOTE: Alternative interpretation of PG by Kaspi et al. (2005) is an outflow with V~3000 km/s, due mainly to alternative line ids plus a detailed model which has emission lines which “fill in” some of the absorption lines.

Search for local (z~0) OVII & OVIII Absorption in the HETG AGN sample Search for WHIGM: McKernan, Yaqoob, & Reynolds (2004, ApJ, 617, 232). + No whigm detected; Ovii & Oviii detected but no FUSE observations; Ovii & Oviii detected AND Ovi detected with FUSE.

Are the Extreme Outflows Due to WHIGM? Local (z~0) OVII absorption from HETG Seyfert 1. PDS 456 3C 273 PKS 2155 PG Mkn 421  Outflows not claimed in 3C 273, PKS 2155, & Mkn 421.  “Outflow” in NGC 4051 with V ~ 600 km/s may be due to local absorption.  Large columns in PG & PDS456 make WHIGM interp. very difficult.

Are the Extreme Outflows Due to WHIGM? Local (z~0) OVII absorption from HETG Seyfert 1. PDS 456 3C 273 PKS 2155 PG Mkn 421 PG does NOT fit on the correlation. The outflow velocity required (from modeling XMM data- Pounds et al. 2004), is ~60,000 km/s, much larger than cz~20,000 km/s. PG most outflows

Narrow Emission Lines In the McKernan et al. Sample, prominent emission lines are seen in NGC 3783, NGC 5548, NGC Weaker emission lines are seen in MCG , NGC 4593, Mkn 509, & Mkn 279. Emission lines are much less common than absorption lines in Sy 1 (but soft X-ray spectra of Sy are line-dominated, e.g. NGC 4151). In general the emission lines are centered on the systemic velocity of the system. NGC 5548 OVII triplet NGC 4151 HETG/MEG

HETG Fe K Lines  Fe K line studied for same HETG sample of Sy 1, in Yaqoob & Padmanabhan (2004).  A narrow, often unresolved core at 6.4 keV in the rest frame is common, but some of the AGN HETG Fe K lines show complexity with the high spectral resolution of the Chandra high-energy gratings.

Peak Energy & FWHM of Fe-K Line Core HETG sample of 15 Sy 1 galaxies. Yaqoob & Padmanabhan Mean peak energy: / keV Mean FWHM: /- 760 km/s At least 8/15 have a broad line. Core resolved in 3/15. HEG resolution is ~1800 km/s FWHM.

Correlation of Ionized Outflow Parameters with the EW of the Fe-K Line Core HETG sample of 15 Sy 1 galaxies: Fe-K core EW from Yaqoob & Padmanabhan Warm absorber (outflow) parameters from McKernan, Yaqoob & Reynolds (2006). NVNV

Variable narrow 6.4 keV line found in Mkn Petrucci et al. (2002).

Ionization balance of Fe (as manifested by the emission line complex between Fe I and Fe XXVI), responds rapidly to the continuum variability (1000s of seconds or less). See Yaqoob et al. (2003).

Cen A Suzaku Spectrum

MCG XIS0XIS1 XIS2 XIS3

MCG

Low Energy Spectral Resolution of the XIS BI CCD is Better than XMM EPIC pn & MOS

Suzaku resolves C, O lines in the Planetary Nebula BD for the first time C/Ne ~ C/O ~ N/C ~ 0 CVI OVII OVIIINe IX Chandra ACIS: C, O lines unresolved. Suzaku BI CCD Suzaku Team/K. Arnaud/K. Makishima

 How does the better low-energy spectral resolution of the XIS BI chip help with studying the warm absorber or wind, compared to Chandra or XMM?  Effective area of the XIS BI chip is much higher than that for the gratings. However, XIS BI is better than XMM pn, making it more suitable for time-resolved spectroscopy.  Method: Folded 830 ks NGC 3783 HETG/MEG spectrum through the Suzaku XIS BI response and then through the XMM response.

Suzaku vs. XMM Total Effective Area

HotGAS: Home of the Grating Archive Spectra

Example page (top half) for NGC 1068

Example page (bottom half) for NGC 1068

Summary  Origin & location of ionized outflows still uncertain.  X-ray columns ~ cm -2,  ~ , sometimes in same source.  Location, N e, X-ray cov/fill factor (C) require variability studies. (Chandra monitoring of NGC 3783 gave only upper limit on R, < 6 pc. Origin may still be at accretion disk, but detection may only be possible for matter further out).  Can only measure absolute mass outflow rate when C is known. If C is similar in different AGN, we find a tentative connection between mass flow and accretion efficiency. For large C, mass outflow rate comparable to accretion rate (~ solar M/year).  Tentative connection between outflow velocity and accretion efficiency, independent of C. Bimodal V?  Some very high V outflows in QSOs may be really be due to z~ 0, local absorption (whigm?). However, the implied column density for whigm is way too high. There are low V interpretations also.  Rich & new phenomenology of Fe K emission with Chandra & XMM: Need to disentangle the disk & DM components.