Cosmic Static and X-ray Vision: Studying GPS/CSS sources at radio and

X-ray wavelengthsJoe Callingham

The University of Sydney / CASSMulti-Wavelength Emission Conference, Sydney – February 2015

NR

AO

/AU

I

What are CSS/GPS sources?› Originally empirical classification:

- Powerful AGN with concave spectrum

- GPS turnover ~ 1 GHz, CSS turnover ~ 150 MHz (?)

- Small physical sizes. GPS < 1 kpc, CSS ~ 1 – 10 kpc

- Hosts vary - quasars, radio galaxies and Seyferts

0710+439, O

wsianik et al 1998

Radio detected galaxy

GPS

CSOCore-

jet

CSS

CSOCore-

jet

Acronym Soup

Spectral classification

Morphology classification

Core-jet

Acronym Spaghetti

Morphology classificationKunert-Bajraszewska et al. 2010

Why Study CSS/GPS Sources?

› Unique view of early stages of AGN activity. Probe of environment to tens pc scale.

› How many sources go from birth to A team sources (Cyg A, Her A etc)?

› Are they confined to small spatial scales due to ‘youth’ or ‘frustrated’?

› Cause of the turnover in spectrum?

NASA, ESA, S. Baum and C. O'Dea (RIT), R. Perley and W. Cotton (NRAO/AUI/NSF), and the Hubble Heritage Team

Despite three decades since the discovery of GPS sources as a

significant population, the nature of the absorption mechanism is still

unresolved and restricting progress with these sources.

Stawarz et al. (2008)

The spectral revolution has begun!

› ~10% of MWA radio population?

- turnover solely due to synchrotron self-absorption or free-free absorption?

- are CSS/GPS properties only because of youth?

- or are some “frustrated” sources confined by dense gas?

› Three absorption models:

- Homogenous free-free

- Inhomogeneous free-free

- SSA

Models of GPS Radio Spectra

Homogeneous free-free model

Tingay & De Kool (2003)

Models of GPS Radio Spectra

Inhomogeneous free-free model (Bicknell et al. 1997, Begelman 1999)

Models of GPS Radio Spectra

Synchrotron self-absorption (SSA) model (Kellermann 1966)

Prediction of 2.5 slope – never seen

New Extreme GPS Source PKS B0008-42

› Low frequency data has a gradient of ~2.4 – steepest known. Spectral width ~0.6 decade of freq. – smallest known.

› Test bed for models of GPS/CSS spectra.

› ~120 mas scale, ~1000 pc

› MWA and CABB data vital in disentangling competing models.

Flu

x D

ensi

ty (

Jy)

Frequency (GHz)

Flu

x D

ensi

ty (

Jy)

Frequency (GHz)

χ

Flu

x D

ensi

ty (

Jy)

Frequency (GHz)

χ

Spectral break

Flu

x D

ensi

ty (

Jy)

χ

Frequency (GHz)

Spectral Time Evolution

Adapted from Lindfors et al. (2006)

Frequency

Flu

x D

ensi

ty

Going from one source…

› Spectral modelling of PKS 0008-42 shows power of MWA and CABB in constraining spectrum.

› Ruling out SSA.

› Old and dying GPS sources? Will we see more?

› Implications break moved below the spectrum for USS source

Klamer et al. 2006

MWA and GLEAM Survey

› GLEAM is the MWA all sky survey

› Two minute snapshot observations

› Five declination bands (+18 to -72 degrees).

› Continuous coverage between 80 and 230 MHz.

› Confusion limited and ~half a million sources detected down to ~10 mJy.

Natasha Hurley-Walker

New GPS SourcesF

lux

Den

sity

(m

Jy)

Frequency (MHz)

χ

New GPS SourcesF

lux

Den

sity

(m

Jy)

Frequency (MHz)

χ

New GPS SourcesF

lux

Den

sity

(m

Jy)

Frequency (MHz)

χ

H I Absorption

› Provides a unique insight into the idea of jets being ‘frustrated’.

› Variability in H I profile? Thoughts of clouds moving through lobes/jets.

Allison et al.

Araya et al. (2010)

Who are you calling confined?

› Frustrated sources since as linear size increased NHI decreased?

Pihlstrom et al. 2003

X-ray Vision

› Insights into youth vs frustration

› Origin of X-ray emission is still debated as resolution limiting progress. From:

- Shocks from expanding radio lobes?

- Accretion disk?

- IC scattering?

- Synchrotron from jets?

Cen A; ESO; NASA/CXC/CfA

X-ray Vision

› Vink et al. (2006) found average NH 10∼ 22 cm−2 – similar to column densities observed toward other types of radio galaxies

› Results argue against the frustration scenario since much higher X-ray absorption column densities expected if the expansion of the radio jets is confined.

› Small, biased sample. Only X-ray detected sources! Those with greatest HI densities may not be detected.

Vink et al. 2006

X-ray Column Density

› Where is the absorption occurring? Co-spatial not disc since 1:1.

› Correlation between column density of HI from radio and H from X-ray.

› Column densities much larger than thought with ~1023 cm−2.

Ostorero et al. 2010

Conclusions

› Spectral modelling of PKS 0008-42 shows power of MWA and CABB in constraining spectrum.

› Will MWA reveal a new population of dying sources?

› X-ray and radio observations help constrain environmental properties.

› Ruling out SSA.

› Nail down what the absorption mechanism - vital for evolutionary models.

Death to the Power Law

Days of the power law are numbered!

Why Bayes?

Aesthetics:

› Philosophy – accepting a theory rather than rejecting a hypothesis.

› Chi-squared evaluates the significance of the mismatch between theory and experiment, not whether the hypothesis is true.

› Rigorous theoretical framework

Why Bayes?

Positives - Practical:

› Full PDFs for each model parameter

› Prior knowledge can be used to get a more accurate result and place physical constraints.

› Can deal with non-Gaussian uncertainties (e.g. calibration errors)

› Marginalise over nuisance parameters (e.g. noise floor.)

› Objective model selection more robust than reduced chi-squared.

› Less likely to get stuck in a local minimum due to implementation.

› Hyperparameters

Why Bayes?

Negatives:

› Less ‘natural’ to think about – integrals, baggage of another statistical language etc.

› More computationally expensive

› In simple cases, often converges to the same parameter values as less computationally expensive methods do.

› More difficult and time consuming to code.

› Can be influenced by prior knowledge.

Application on PKS 0008-42

Inhomogeneous free-free model (Bicknell et al. 1997)

PKS 0008-42