Measuring the properties of QSO broad- line regions with the GMOS IFU. Randall Wayth with Matt O'Dowd & Rachel Webster.

Measuring the properties of QSO broad-line regions with the GMOS IFU.

Randall Waythwith Matt O'Dowd & Rachel Webster.

Outline

Motivation Introduction to 2237+0305 Observations & Data reduction Emission line flux ratios Microlensing of QSO emission regions Constraints on emitting region size

Motivation - QSO emission regions QSO continuum/line emission regions are too

small to resolve Reverberation mapping suggests they are very

small Gravitational lensing magnifies objects. If a

source is resolved in a lensed image, then we can directly determine its true size & surface brightness

If not resolved, then microlensing should create different magnifications for the continuum and broad-line regions of the QSO.

2237+0305

Barred spiral (Sbc) galaxy at low redshift (z=0.04) lensing z=1.69 radio quiet QSO.

Four images of QSO formed around galaxy bulge with separation ~1-2 arcsec

Lensing offers unique opportunity to study

• QSO continuum and emission line region size/structure

• Properties of dark matter halo (shape, cuspiness, clumpiness)

• Mass function of galaxy bulge stars, and more...

2237+0305

15 second r-band acquisition image

N

E

2237+0305

Same image, different contrast

2237+0305

A

D

B

C

Galaxy centre

QSO image labels follow Yee (1988)

Is the CIII] emission region in 2237+0305 resolved?

Mediavilla et al. (1998) claimed seeing an arc of resolved CIII] emission using INTEGRAL IFU on WHT. (0.5” separation, rectangular array, 0.45” fibre diameter)

If real, we can “undo” the effects of lensing and create a true image of the emission region.

From Mediavilla et al. (1998) ApJ 503 L27

Data - GMOS IFU

IFU is a hexagonal lenslet array with separation 0.2”

R400 grating in “one slit” mode. Useful wavelength range ~500-850nm. Object coverage is 5”x3”.

8 x 30min exposures taken on 16/17 July, 2002. We use 5 of the 8 frames. Seeing 0.6”

ObjectSky

Aims

Confirm/refute existence of arc of emission If real, make an image of the QSO BELR! If not real, examine effect of lensing on the

relative strengths of continuum and broad-line emission from the QSO

DataQSOspectrum

Galaxyspectrum

AB

D

C

CIII]

MgII

Line flux extraction

CIII] line

Continuum

MgII line

Images of the broad-line flux Subtracting surrounding continuum from the

emission lines leaves the line flux Subtraction is quite clean Notice difference in brightness of QSO images

CIII] continuum CIII] - emission lineMgII continuum MgII – emission line

Arc or PSF overlap?

PSF modelling & subtraction We are looking for a faint arc, so we need to

create an accurate PSF and subtract the QSO images.

Method

• combine line images for the 5 good frames

• define a mask around each QSO image including a region which is uncontaminated by other images

• cut out, rescale and combine sections from each image

• use this PSF, to subtract QSO images, iterate a few times

PSF model

Uncontaminated regions

Combined PSF (MgII)

Line images with QSOs subtracted

Unresolved! - No arc in MgII or CIII]! Peak residual ~10%

MgIICIII]

Microlensing and the BELR Microlensing by stars in

the lens galaxy's bulge project a network of “caustics” onto the QSO.

Parts of the source crossing caustics (red/yellow) are highly magnified.

The QSO can be differentially magnified depending on its size relative to the caustic network.

Microlensing caustic networkImage courtesy Joachim Wambsganss

Microlensing and the BELR Small source = no

differential magnification

All parts of the source are equally magnified

Microlensing and the BELR Medium source =

differential magnification!

If the QSO's continuum region is much smaller than the BELR, then the continuum should be more highly magnified.

Flux ratios

A

DB

C

Galaxy centre

Extinction corrected flux ratios for continuum and broad-lines are certainly different!

Without microlensing, all images should have approx same magnitude, so BELR is also microlensed.

Because MgII and CIII] lines have same flux ratio, they must be similar size.

BELR size ~0.06pc based on simulations of Wyithe et. al 2002 (MNRAS 331)

Next: de-dispersed spectral ratios

After correcting for atmospheric dispersion, take ratios of image spectra

Broad-line magnifications clearly visible

Shape of continuum is a function of source morphology, microlensing and extinction

Shape/location of lines depends on BELR structure!

C/A

D/A

B/A

5000 6000 7000 8000

Summary

Using GMOS-N IFU we have taken the best spectroscopic data of 2237+0305 to date

We find no arc of emission in either the CIII] or MgII line, contrary to previous claims

Magnification ratios of the images in both the continuum and broad-lines show microlensing

BELR is measured from flux ratios to be ~0.06pc. This estimate will improve using de-dispersed data.

MNRAS 359 561 (2005)