Quan%fying the diffuse con%nuum contribu%on from BLR gas: a - - PowerPoint PPT Presentation

quan fying the diffuse con nuum contribu on from blr gas
SMART_READER_LITE
LIVE PREVIEW

Quan%fying the diffuse con%nuum contribu%on from BLR gas: a - - PowerPoint PPT Presentation

Quan%fying the diffuse con%nuum contribu%on from BLR gas: a modeling approach Mike Goad, Daniel Lawther, Kirk Korista, Otho Ulrich, Marianne Vestergaard Mike Goad Atlanta , USA 2017 Our approach: q Build a model BLR match the intensi:es (


slide-1
SLIDE 1

Atlanta , USA 2017 Mike Goad

Quan%fying the diffuse con%nuum contribu%on from BLR gas: a modeling approach

Mike Goad, Daniel Lawther, Kirk Korista, Otho Ulrich, Marianne Vestergaard

slide-2
SLIDE 2

Atlanta 2017 Mike Goad

Our approach:

q Build a model BLR match the intensi:es (variability 9mescale/amplitude)

  • f strongest UV/op%cal emission lines in NGC 5548

(***For objects of interest, no such thing as a steady-state model***)

q Compute wavelength-dependent (UV-op%cal-IR) flux and variability %mescale of DC arising from the same gas q Scale delays according to the frac%onal flux contribu%on

DC/(INCIDENT + DC)

q Drive with model con%nuum light-curves

es%mate sta%s%cally likely delays (CCF/JAVELIN)

+ dependence on characteris9c %mescale & variability amplitude

  • f driving con%nuum (MC approach)
slide-3
SLIDE 3

Atlanta 2017 Mike Goad

Types of model:

q Pressure law model : Rees, Netzer and Ferland 1989, Goad, O’Brien, Gondhalekar 1993 Kaspi and Netzer 1999 Netzer 2000 q Local Op1mally emi4ng Clouds : Baldwin, Ferland, Korista, Verner 1995, Korista and Goad 2000, 2001, 2004.

slide-4
SLIDE 4

Atlanta 2017 Mike Goad

(1) Pressure Law models :

Lawther, Goad, Korista, Ulrich, Vestergaard 2017, in prep

Run of physical condi%ons with radius specified by simple radial pressure law P(r) ∝ r−s

nH(r) ∝ r−s

U(r) ∝ rs−2

Ac(r) ∝ R2

c ∝ r2s/3

dC(r) ∝ Ac(r)nc(r)dr ∝ r2s/3−3/2dr

L = 4⇡ Z rout

rin

✏(r)Ac(r)nc(r)r2dr

Assume mass conserva%on + spherical clouds Differen%al covering frac%on Line luminosity

Ncol(r) ∝ RcnH ∝ r−2s/3

const Temp

slide-5
SLIDE 5

Atlanta 2017 Mike Goad

Normaliza%on condi%on : specify

ΦH, nH, Ncol at some r

+ inner and outer radius, and total covering frac%on Ctot Choose line radia%on pagern – we assume clouds are spherical

✏(r, ✓) = ✏totl[1 − (2F(r) − 1) cos ✓] F(r) = ✏inwd/✏totl

slide-6
SLIDE 6

Atlanta 2017 Mike Goad

−2 2 4 6

log[Energy / 1Ryd]

40 41 42 43 44 45

log[λLλ /1 erg s−1]

Two cases: s=0 , constant density nh, constant column density Nh + s=2 , constant ioniza%on parameter U

Mehdipour et al. 2015

slide-7
SLIDE 7

Atlanta 2017 Mike Goad

slide-8
SLIDE 8

Atlanta 2017 Mike Goad

7 8 9 10 11 12 13 14

log(nH/1cm−3)

20 40 60 80 100 120 140 160 180

r (lightdays) Ly-α C IV Hα Hβ He II 4686 He II 1640 log(ncol) = 22.5

emissivity-weighted radii can include effects of anisotropy/responsivity tends to increase delays further R=14.8 lt-days R=148 lt-days R=1.48 lt-days

slide-9
SLIDE 9

Atlanta 2017 Mike Goad

Similarly – constant U models

−2.0 −1.5 −1.0 −0.5 0.0 0.5

log(U)

1040 1041 1042 1043 1044

Lline (erg s−1)

Ly-α C IV Hβ He II 4686

Broad range in ioniza%on for which we can exceed the measured line luminosi%es

slide-10
SLIDE 10

Atlanta 2017 Mike Goad

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

1042 1043 1044

Continuum νLν [erg s−1]

Diffuse continuum Ionizing continuum Total continuum

Model 1 (s =0,log(nH) = 10.75)

2000 4000 6000 8000 10000

Wavelength[ ˚ A]

0.0 0.2 0.4 0.6 0.8 1.0

Diffuse continuum fraction, Fdiff

Model 1

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

10 20 30 40 50 60 70 80

Centroid [lightdays]

r rη τ (anisotropic) τη (anisotropic)

Model 1 (s =0,log(nH) = 10.75)

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

10 20 30 40 50 60 70 80

Lag (days)

CCF Peak CCF Centroid

Model 1 (s =0,log(nH) = 10.75)

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

2 4 6 8 10 12 14

Diluted lag (days)

CCF Peak × Fdiff CCF Centroid × Fdiff

Model 1

slide-11
SLIDE 11

Atlanta 2017 Mike Goad

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

10 20 30 40 50 60 70 80

Centroid [lightdays]

r rη τ (anisotropic) τη (anisotropic)

Model 2 (s =2,log(U) = −1.23)

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

2 4 6 8 10 12 14

Diluted lag (days)

CCF Peak × Fdiff CCF Centroid × Fdiff

Model 2 (s =2,log(U) = −1.23)

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

10 20 30 40 50 60 70 80

Lag (days)

CCF Peak CCF Centroid

Model 2 (s =2,log(U) = −1.23)

2000 4000 6000 8000 10000

Wavelength[ ˚ A]

0.0 0.2 0.4 0.6 0.8 1.0

Diffuse continuum fraction, Fdiff

Model 2 (s =2,log(U) = −1.23)

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

1042 1043 1044

Continuum νLν [erg s−1]

Diffuse continuum Ionizing continuum Total continuum

Model 2 (s =2,log(U) = −1.23)

slide-12
SLIDE 12

Atlanta 2017 Mike Goad

2000 4000 6000 8000 10000

Wavelength[ ˚ A]

0.5 0.6 0.7 0.8 0.9

Anisotropy, F = in(λ)/tot(λ)

Model 1 (s =0,log(nH) = 10.75)

2000 4000 6000 8000 10000

Wavelength[ ˚ A]

0.4 0.5 0.6 0.7 0.8 0.9 1.0

Anisotropy, F = in(λ)/tot(λ)

Model 2 (s =2,log(U) = −1.23)

Inward frac%ons

slide-13
SLIDE 13

Atlanta 2017 Mike Goad Density dependence - constant density models

10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 2000 4000 6000 8000 10000 Wavelength [ ˚ A] 10 20 30 40 50 60 τCCF,cent. [days] 2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 8

2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 9

2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 10

2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 11

2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 12

2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 13

2000 4000 6000 8000 10000 Wavelength [ ˚ A] 2 4 6 8 10 12 14 τCCF,cent × Fdiff.

log(nH) = 14

slide-14
SLIDE 14

Atlanta 2017 Mike Goad

1041 1042 1043 1044 1045

Lν [erg s−1]

1041 1042 1043 1044 1045

Lν [erg s−1]

1041 1042 1043 1044 1045

Lν [erg s−1]

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

1041 1042 1043 1044 1045

Lν [erg s−1]

20 40 60 80 100

τη [days]

log(U) = 0.52

20 40 60 80 100

τη [days]

log(U) = −0.48

20 40 60 80 100

τη [days]

log(U) = −1.48

2000 4000 6000 8000 10000

Wavelength [ ˚ A]

20 40 60 80 100

τη [days]

log(U) = −1.98

Constant ioniza%on models

slide-15
SLIDE 15

Atlanta 2017 Mike Goad

100 200 300 400 500

Time [lightdays]

0.0 0.5 1.0 1.5 2.0 2.5

Relative luminosity

DRW Continuum C IV, -weighted resp. C IV, η-weighted resp.

100 200 300 400 500

Time [lightdays]

0.0 0.5 1.0 1.5 2.0 2.5

Relative luminosity

DRW Continuum Mg II, -weighted resp. Mg II, η-weighted resp.

Aside :

slide-16
SLIDE 16

Atlanta 2017 Mike Goad

10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 10 20 30 40 50 60 τCCF,cent. [days] 2000 4000 6000 8000 10000 Wavelength [ ˚ A] 10 20 30 40 50 60 τCCF,cent. [days] 2 4 6 8 10 12 14 τCCF,cent. × Fdiff.

Tchar = 5

2 4 6 8 10 12 14 τCCF,cent. × Fdiff.

Tchar = 10

2 4 6 8 10 12 14 τCCF,cent. × Fdiff.

Tchar = 20

2 4 6 8 10 12 14 τCCF,cent. × Fdiff.

Tchar = 40

2 4 6 8 10 12 14 τCCF,cent. × Fdiff.

Tchar = 80

2000 4000 6000 8000 10000 Wavelength [ ˚ A] 2 4 6 8 10 12 14 τCCF,cent. × Fdiff.

Tchar = 160

Dependence on Tchar Tchar (days) 1989 ~ 80-120 1993 ~ 40 2014 ~ 10-20

RAW DILUTED

slide-17
SLIDE 17

Atlanta 2017 Mike Goad

(2) Local op%mally emilng clouds (LOC) models

Korista and Goad 2000,2001

At any given radius there exists a range of gas densi%es/column densi%es (or simply …..more than one pressure-law!) Spectrum dominated by selec%on effects introduced by atomic physics and general radia%ve transfer within the large pool of line-emilng en%%es Strengths: Summa%on over cloud distribu%on leads to: (i) typical AGN spectrum (ii) Ioniza:on stra:fica:on (iii) Luminosity-Radius rela:on arises naturally

slide-18
SLIDE 18

Deriving the spectrum:

Give each line emilng en%ty a weight in 2-dimensions: gas density and distance Assume: analy9c, separable, and a power-law in each variable

f(R) ∝ RΓ g(nH) ∝ nβ

H

Baldwin 1997 – composite quasar spectra best fit if indices in both are close to -1. In our models assume :

β = −1

and fit for Γ Korista and Goad (2000) found a value of -1.2 gives an Acceptable fit to the line luminosi%es For NGC~5548 Atlanta 2017 Mike Goad

slide-19
SLIDE 19

Atlanta 2017 Mike Goad Log Nh=22 Balmer Jump

slide-20
SLIDE 20

Atlanta 2017 Mike Goad Log Nh=23

slide-21
SLIDE 21

Atlanta 2017 Mike Goad Log Nh=24

slide-22
SLIDE 22

Atlanta 2017 Mike Goad

2000 4000 6000 8000 2 4 6 8 10 12 14 Korista and Goad 1993 campaign Korista and Goad 2017 AGN STORM (Nh22) Centroid 2000 4000 6000 8000 2 4 2000 4000 6000 8000 2 4 6 8 10 Korista and Goad 1993 campaign Korista and Goad 2017 AGN STORM Centroid 2000 4000 6000 8000 2 4

slide-23
SLIDE 23

Atlanta 2017 Mike Goad

2000 4000 6000 8000 2 4 6 8 10 Korista and Goad 1993 campaign Korista and Goad 2017 AGN STORM (Nh24) Centroid 2000 4000 6000 8000 2 4

slide-24
SLIDE 24

Atlanta 2017 Mike Goad

2000 4000 6000 8000 2 4 6 8 10 12 Korista and Goad 1993 campaign Korista and Goad 2017 AGN STORM (Nh23 - low den) Centroid 2000 4000 6000 8000 2 4 2000 4000 6000 8000 2 4 6 8 10 12 Korista and Goad 1993 campaign Korista and Goad 2017 AGN STORM (Nh23 - high den) Centroid 2000 4000 6000 8000 2 4

slide-25
SLIDE 25

Atlanta 2017 Mike Goad

Summary:

(1) At typical Nh,nh, U appropriate for BLR there exists a significant diffuse con%nuum arising from the same gas that emits the broad emission-lines (2) Form of the delay spectrum (including underlying powerlaw ) approximately matches that observed, especially around Balmer and Paschen jumps. (3) Even when included, disks s%ll appear too large for their luminosity(?) (4) We need to find new/improved ways (fourier analysis. PCA) of isola%ng the major variable contribu%ons to the observed con%nuum bands