A Search for Resolved Inverse-Compton X-ray Jets from Radio Core-dominated Quasars

Abstract

An analysis has been undertaken of 186 candidates narrowed down by cross-referencing to identify and study resolved X-ray jets from radio-loud core-dominated quasars. This is motivated by the discovery of B3 0727+409 – a quasar with an X-ray jet with no radio counterpart – found by Simionescu et al. (2016). Objects such as these would provide evidence for the beamed inverse-Compton scattering of the Cosmic Microwave Background model for X-ray emission in jets (iC-CMB).

Various methods were utilised to ensure that jet candidates were truly jets – including Point Spread Function (PSF) simulations and utilisation of full (unprocessed) Chandra images – but none were found. The quasar B3 0727+409 was recovered by these methods – demonstrating their validity. New observations of B3 0727+409 were mosaic-ed – revealing new details in the jet, including the disappearance and subsequent re-appearance of the jet in the X-ray.

Contents

Introduction

Background Theory and Objectives

Jets – What do we know?

Energy Source

Composition

Emission

X-ray Observations and Chandra

Point Spread Functions (PSFs)

Emission Mechanisms

Synchrotron Emission

Inverse Compton Scattering

Critical Analysis of iC-CMB

Objectives

Experimental Method

Results

Discussion

Conclusion

Appendices

Appendix A – Distortion of Source Spectrum

Appendix B – PSF Simulation

References

Introduction

Active Galactic Nuclei (AGN) are some of the most extreme objects in the universe. Powered by supermassive black holes, their luminosity can exceed their host galaxies’ by many orders of magnitude (Longair, 2011). AGN exhibit some characteristic physics only possible in such an extreme and exotic environment. First identified as Quasi-Stellar Objects (QSOs) or quasars in the 1960s (Sandage, 1965), they were later unified into a single scheme of AGN along with blazars and radio loud galaxies (Barthel, 1989).

A characteristic component of AGN are two antipodal jets oriented along the axis of the accretion disk. These jets consist of ionised matter (Wardle et al., 1998) travelling at relativistic speeds and can extend up to lengths of hundreds of kiloparsecs (Schwartz et al., 2003). They have been an object of research and discussion for several decades and still much of their nature including formation, composition, collimation and emission mechanism remains a mystery.

Although first noticed as early as 1918 (Curtis), jets were first looked at in more depth with the advent of advanced radio interferometry in the 1970s. The high resolution afforded by Very Large Baseline Interferometry allowed the detailed study of the structures and spectra of these jets. However, studies using radio alone could not give a complete picture of the nature of jets. In 1999, two flagship X-ray telescopes were launched into space – namely Chandra by NASA and XMM-Newton by the ESA. The Chandra mission in particular had an precedented angular resolution of less than 0.5” which allowed for the resolved observation of X-ray jets. Although they were not launched with the primary intent of observing quasars, a quasar was used for in-flight calibration of Chandra (Worrall, 2009). This led to the observation of a resolved X-ray jet (Chartas et al., 2000) with a flux which did not follow the predictions of previous models based on radio and optical measurements (Figure 1).

This and subsequent observations of X-ray jets sparked discussion regarding their emission mechanism. Before the Chandra observations the jets were thought to emit via synchrotron emission from a single continuum of electrons; however, flux upper bounds in the optical set by the Hubble Space Telescope (HST) make this highly unlikely (Chartas et al., 2000). There are two main competing theories to explain this – Synchrotron emission with a modified power law and beamed inverse Compton scattering of the Cosmic Microwave background (iC-CMB). The debate on the emission mechanism is yet unresolved.

The two emission mechanisms imply different properties for the jet. The iC-CMB model requires highly relativistic bulk velocities in the jet. The synchrotron model has no such requirement, which posits a second electron population only at higher energies. However, this is still an unattractive explanation as current models for electron power laws do not sufficiently explain such a population. A resolution to this debate would thus be helpful in improving our understanding of AGN jets.

In 2016 Simionescu et al. discovered a high-redshift quasar with an extended jet in the X-ray but no observable counterpart in the radio. This is interesting because the iC-CMB model predicts increased X-ray flux relative to the radio with increasing redshift. Investigation into the presence of these objects would provide valuable evidence to aid in the resolution of the debate. This project thus focuses on using existing quasar catalogues to find objects like the aforementioned one in the background of Chandra observations.

Figure 1

One of the first Chandra image of a resolved X-ray jet (of PKS 0637-752), the false colour image represents the X-rays and the contours represent the counterpart in radio. Following observations show similar results with a close correlation between the X-ray and radio emission (Chartas et al., 2000)

Background Theory and Objectives

Jets – What do we know?

As mentioned previously, many properties of jets are yet unknown, however observations have identified key characteristics as well as slowly constrained possible explanations.

Energy Source

Jets carry away a significant amount of energy that a supermassive black hole generates. It is not known how the jets are powered, but they are theorised to extract energy from the rotation of the central black hole. It is hard to test current theories regarding jet energy sources but some explanations have been put forth, including the extraction of energy by means of the Penrose Mechanism (Penrose and Floyd, 1971).

Composition

Jets are composed of some form of ionised plasma – some mixture of protons and electrons/positrons – and a magnetic field (Krawczynski, Böttcher and Harris, 2012). The exact composition is yet to be determined, but many studies have constrained their relative numbers in the jets (Celotti and Fabian, 1993; Wardle et al., 1998; Sikora and Madejski, 1999; Hirotani et al., 2000).

Emission

Emission from jets has been observed across the spectrum, including radio, optical, X-ray and gamma-ray observations. Due to the presence of high-energy electrons in a magnetic field, we expect to observe synchrotron emission from the jets. It was confirmed that the radio jets seen from AGN were most likely synchrotron due to the observation of linear polarisation in the jet of M87 (Baade, 1956). X-ray emission from the jets was observed and was attributed to synchrotron. However, the angular resolution was too low to discern enough meaningful information from their observation until the launch of Chandra in 1999.

X-ray Observations and Chandra

Typically, optical telescopes use mirrors, lenses or both to focus light to be viewed. In the case of X-rays though, mirrors would absorb the X-rays at angles closer to normal incidence on them. There are also no analogues to lenses in the X-rays. To focus X-rays using reflection then, a grazing angle of incidence is required.  Wolter (1952a, 1952b) proposed 3 designs for grazing incidence mirror systems to focus X-rays which would minimize the problems otherwise present with grazing incidence angles[1].

The Chandra X-ray Observatory uses nested mirrors in a Wolter Type 1 configuration, which consists of paraboloid and hyperboloid segments. Chandra has an angular resolution of 0.5” which allows for unprecedented levels of detail in X-ray observations. It is the most advanced X-ray telescope till date. Due to this high resolution, in the cases of interest to the project, the PSF is often the limiting factor in terms of understanding observational details.

Point Spread Functions (PSFs)

As best explained by the Chandra X-ray Centre (2017) “The point spread function (PSF), also known as the point response function (PRF), describes the shape and size of the image produced by a delta function (a point) source.”. In practice, this is seen as a ‘smearing’ effect in observations. The PSF may be affected by many factors, including atmospheric effects, optical equipment quality and even off-axis angle among others. Due to the large number of point-like sources observed by telescopes (and, by extension, Chandra) the PSF can be modelled and re-produced for a wide range of variables and object characteristics. Accurately doing so can dramatically improve our understanding of observational data from Chandra.

Emission Mechanisms

Synchrotron Emission

Synchrotron emission is the emission of EM radiation by highly-relativistic charged particles in a magnetic field. The energy loss rate per electron for synchrotron emission is defined as follows (Longair, 2011).

-dEdt=2σTcγ2β2uBsin2⁡θ (1)

Where

σT is the Thompson cross-section,

γ is the Lorentz factor,

β is the ratio of velocity to the speed of light and

θ is the angle made by the electron with the magnetic field.

uB is the energy density of the magnetic field given by

B22μ0.

At highly relativistic velocities

β~1and the energy loss rate becomes proportional to the square of electron energy. The observed electron energy spectrum can be determined using this knowledge of the energy loss rate. Generally, electrons have a power law electron distribution of say

QE=kE-p. The observed energy spectrum is modified by energy loss mechanisms as follows (Longair, 2011)

nE=-kE-p-1p-1dE/dt (2)

While this assumes an infinite and uniform distribution of electron sources, it is sufficient to see that synchrotron emission will have a broken power-law spectrum which steepens and is cut off at higher energies. Since for synchrotron emission -dE/dt∝E2 it is immediately clear that the observed spectrum n(E) will be steeper than the source spectrum by 1. What may not be as clear is that this steepening only occurs when the energy loss timescale by synchrotron is high enough for the effects to be apparent. This means that the straight-line spectrum ‘breaks’ when the energies become high enough for synchrotron energy loss to become significant with respect to electron residence time in the jet (See Figure 2).

Another characteristic of synchrotron emission that it has high linear polarisation. This allows for the identification of synchrotron emission by measuring polarisation and indeed, this polarisation can be seen in many radio sources. These two characteristics have been widely observed in jets; and it is generally accepted that the radio-optical spectrum of the jets arises from synchrotron emission.

The problem arises when considering the X-ray emission measured by Chandra. As seen in Figure 2 a single component synchrotron spectrum cannot explain the spectrum of the jet. As previously mentioned are two main explanations for this – a modified synchrotron spectrum with two electron energy distributions, or beamed inverse Compton scattering of the Cosmic Microwave Background (iC-CMB).

The synchrotron explanation simply posits that there is a second electron population exclusively at higher energies. This would satisfy the observed spectra from these jets, but there are no concrete explanations as to why or how such an electron population would come to exist.

Figure 2

The spectrum of M87 (left) and PKS 0637-752 (right). Note the steepening of the spectrum at a certain critical frequency. The initial slope is -α=-(p-1)/2, which then becomes steeper when synchrotron loss timescales become significant. On the left it is clear that a single component electron distribution can describe the observed fluxes across the spectrum. However, for PKS 0637-752, clearly a single broken spectrum cannot explain the X-ray emission. From the review of X-ray jets by Worrall (2009).

Inverse Compton Scattering

 

Figure 3

Two images showing a visual representation of the beaming effect of light at relativistic speeds. The left image shows a blob selected from the jet, and the right image shows the rest frame of the blob. The red background in the rest frame represents the CMB, and the shifted colours on the right show the relative beaming effect of the CMB. These are not the true colours that would be seen, but a representation of the effect. This was simulated in Python using the relativistic Doppler effect equations at

Γ=10.

Inverse Compton scattering in a nutshell is the up-scattering of photons by relativistic electrons. The energy loss rate per electron has a very similar form to synchrotron emission (1). Instead of being dependent on the energy density of the magnetic field uB it depends on the energy density of the photon field urad.

-dEdt=43σTcγ2β2urad  (3)

In the case of iC-CMB, the photons that are scattered are the Cosmic Microwave Background photons and thus the energy density needed is that of the CMB. The CMB energy density is dependent on redshift as 1+z4 so it is to be expected that this effect would be more prominent at higher redshifts. The observed ratio of flux density between synchrotron and iC-CMB would thus go as follows.

SiC-CMBSsyn∝uraduB∝1+z4B2 (4)

For the special case of jets however, there is an additional beaming factor. As jets have high bulk velocities, the previously stated equations would not be sufficient as they assume random velocities and an isotropic photon field. For a more rigorous treatment of inverse Compton scattering in jets, first a Lorentz transformation is made to a ‘blob’ in the jet frame (Figure 3). Now the effects can be calculated using the case of a population of electrons with zero bulk velocity. It must be kept in mind that due to the Lorentz transformation the CMB is now strongly beamed in the direction of motion of the jet. The scattering will then be primarily beamed in its direction of travel. Once the effects are calculated in this frame, this must once again be Lorentz transformed back into the observer frame. Though the full derivation is beyond the scope of this project, the result is important — the flux density ratio is boosted by a factor of approximately

Γ2, where

Γ is the bulk Lorentz factor. Equation (4) is then modified as follows.

SiC-CMBSsyn∝Γ21+z4B2 (5)

For the sake of convenience, an angular dependence factor of (1+cos⁡θ)2 is omitted as it is difficult to ascertain both θ and Γ independently.

Although there is a general idea of how jet characteristics affect iC-CMB flux, there is still debate about the correct mathematical treatment of the model. In general, for the observed flux in the X-ray, a high jet bulk Lorentz factor Γ ~ 10-20 is required. Some models posit that there is a fast-moving spine that produces the iC-CMB X-rays and a sheath of slower-moving electrons that are the cause for synchrotron radio emission.

Critical Analysis of iC-CMB

With this knowledge of the iC-CMB and synchrotron models at hand, the next question to ask would be how do they stand up to scrutiny and what do observations tell us about them? For iC-CMB, we know that the effect must be present, as there exist relativistic electrons that scatter the CMB. The question is whether it is significant enough for it to be the cause of the observed X-ray emission. From Equation (4) we know that the iC-CMB effect has a 1+z4 dependence on redshift, so we would expect the effect to be more pronounced at higher redshifts and there have been a few studies which have investigated this z-dependence.

Most recently Marshall et al. (2018) conducted a survey of 56 jets and measured the dependence of αrx– a value related to the flux density ratio between the X-ray and radio emission of the observed jets. Their findings rejected the iC-CMB model for αrx dependence on redshift at 99.5% confidence. Most of the jets were, however, were at z<2. At higher redshifts, as found by McKeough et al. (2016), the measured flux ratios are higher than that at lower redshifts. Their study only included 12 jets however, which could be argued to be too small a sample size. It does however suggest that there could be some missing information on the matter, and studies in the future with a larger sample size could shed more light on the topic.

Using Fermi measurements of the Gamma-ray spectrum of PKS 0637-752 and constraining the model using ALMA measurements in the radio, Meyer et al. (2017) rejected the iC-CMB model at a 8.7σ level. Meyer et al. (2016) also measured the proper the proper motion of knots in the jet of 3C 273 and found that they had Γ<2.9, which is too low for the observed flux to be from iC-CMB. These are hard to explain using current models, but there still exist other objects which are probably best explained by it.

In 2016, Simionescu et al. published a paper reporting the serendipitous observation of an X-ray jet from the high redshift (z=2.5) background quasar B3 0727+409 without a radio counterpart. From Equation (4), objects of this nature could be expected due to the high redshift dependence of iC-CMB. This observation is thus well in line with the predictions of iC-CMB and provides valuable insight into a new set of potential objects to search for, which could provide new insight into the evolution of AGN.

Objectives

The primary objective of this project is to find objects similar to the one found by Simionescu et al. (2016) in the background of existing Chandra observations. A population of jets with similar high X-ray to radio flux density ratios would align with the predictions of iC-CMB and contribute in a meaningful way to the current discussion of X-ray emission mechanisms.

Experimental Method

The bulk of this experiment used the software Topcat (Taylor, 2005, 2006) to cross reference multiple catalogues. First a quasar catalogue called the Million Quasars catalogue (Milliquas) (Flesch, 2017) was used to identify quasars for cross-referencing. This catalogue consists of the Sloan Digital Sky Survey (SDSS) DR14 Quasar catalogue and a list of high-confidence (>80%) quasar candidates. This was then cross-referenced with the FIRST Radio Survey of the sky by the VLA. The data from the FIRST survey is then used to identify a set of candidates with angular extent of the resolution of the FIRST survey (5 arcseconds). This set is then limited to candidates with redshift greater than 2, and then cross-referenced again with the list of Chandra Observations. The cross-referencing with Chandra observations is done so that the candidates may lie within the angular extent of the Chandra images.

The Chandra database for each observation is accessed to retrieve a pre-processed image of each of the candidates. These images are then manually combed through for signs of an extended jet. The Point Spread Function (PSF) of these objects may vary depending on their location on the image and how light is reflected by the mirrors. To ascertain that the candidate is in fact displaying signs of a jet, the PSF is simulated using the software package ‘CIAO’ provided by the Chandra X-ray Centre (CXC) (Fruscione et al., 2006). For the final jet candidates, the full images are downloaded from the Chandra website and the PSF can be simulated. The PSF can cause distortion anisotropically, which could otherwise be misconstrued as a jet.

In addition to this, new data from the object discovered by Simionescu et al. that was previously unavailable for public use became available in late December 2017. This new image has much higher exposure time, spread between 3 images. A script provided by the Project Supervisor which called several tasks of the CIAO software was used to combine these images to reveal more detail in the jet.

Results

Figure 4

Representative image showing the problem with certain Chandra images. Sometimes, for the purpose of making a quick observation, only the middle pixels of the sensor will be activated, and large portions of the background are missed, leading to quasars outside the field of vision of Chandra. The greys have been brightened for representative purposes. Each white dot represents photon counts.

The final list of candidates found by cross referencing was a list with 186 quasars. Of these, 42 were found to be outside of the frame. This is due to the cross-referencing method using a radius input, while the Chandra image is a square. Due to this there are regions outside of Chandra’s frame where quasars may be located. In addition to this, some observations are made only using the middle portion of the sensors, further reducing the area captured in the background (Figure 4). With these factors and the geometry of the sensor array in mind, the probability of the object being outside the Chandra image is (very) roughly 15%-25%. The observed percentage of objects outside of the image is 22.5%. This follows our expectations of how many objects would be outside the frame.

 

Figure 5

Side-by-side comparison of B3 0727+409 in image found using project files and cropped image published by Simionescu et al. (2016). The two have dissimilarities due to the colouring scheme difference between the images. The hot spot that would later be confirmed has been cropped out for better visibility of short-range structure. The false colour in the right-hand image represents counts.

The object discovered by Simionescu et al.(2016) was recovered using the parameters set for the candidate list. The first archival image of the object is shown in Figure 5, adjacent to the image used in the paper. The presence of B3 0727+409 in the candidate list is a good indicator that the parameters used for cross-referencing were appropriate to find similar objects.

To demonstrate the potential effect of PSF on an object, one is simulated for an object that could be erroneously identified as an extended object (Figure 6). The simulated PSF and observed image show remarkable consistency in the appearance and the 2-D ‘slice’ of the PSF shows that they have a very similar spread as well. This reveals that it is almost certainly a point source. Although for other sources it isn’t as pronounced as Figure 6, it clearly demonstrates the utility of the method.

Figure 6

Top – PSF simulation. Bottom – Chandra image. On the left half of the image, the line plots between the same coordinates on both the simulated and observed image are shown. On the right, the shape of the PSF can be seen alongside the Chandra image where it is more prominent in one direction over the other. The remarkable similarity between the shapes demonstrates how the PSF may smear a point source into what may be misinterpreted as a jet.

The FIRST data was also used to identify potentially interesting features; however, the resolution was too low to be of interest except in a few isolated circumstances. One of the images with the radio contours overlaying the Chandra image can be seen in Figure 7.

Figure 7

A Chandra image of a potential jet candidate overlaid with the radio contours in green. There is a slight correlation between points with 2 counts to the bottom right. This was investigated further using the rest of the observations of the object, but that revealed that it was simply background noise that happened to line up with the FIRST data.

In late 2017, a second observation of B3 0727+409 was made publicly available. It consists of 3 observations totalling 130 kiloseconds. These images were combined as detailed in the Experimental Method. The image reveals that the jet disappears and then reappears, a feature previously unclear in the original observation (Figure 8). This may be of interest in terms of the discussion on iC-CMB. In addition, the overall count rate in the image has improved significantly as would be expected.

 

Figure 8

Left – Combined image made using new Chandra observations of B3 0727+409. Note the clear re-appearance of the jet observable in the new observation. Right – Image of B3 0727+409 (with a gaussian smooth function) from the paper by Simionescu et al. (2016) with green radio contours.

Discussion

Although a large sample of compact radio-loud quasars was found, there were no conclusive candidates that would merit a future, more detailed study by Chandra. Out of 186 candidates produced by the cross-referencing, fewer than 20 could be considered interesting for investigation beyond visual identification.

If a large percentage of these turned out to be viable candidates for further investigation, a preliminary analysis could be done on their flux dependence compared to redshift. In addition, an attempt could be made to compare the X-ray images with higher resolution radio charts. This could then be potentially used for further investigation by means of a longer duration Chandra image. This was not the case however, as under further investigation any signs of an extended jet are found to be caused by other effects. They were primarily investigated primarily using comparisons with PSF simulations and/or using different full images available on the CXC website.

The question would then be, why do the predictions of iC-CMB not hold up? It is possible that the data used is insufficient to identify some jets. A significant problem with the project methodology is that the Chandra images used have insufficient count rates to meaningfully distinguish possible jet features from background noise. This is mostly due to the targets of observations being at lower redshifts than the quasar candidates and thus typically require shorter exposure times. Perhaps an extensive survey of high-redshift quasars with higher exposure times could reveal more information about the existence of similar objects. This could then even focus on better candidates selected using higher resolution radio data. The radio data provided by FIRST, though extensive, has a relatively low angular resolution of 5”. This is too low to distinguish features that could potentially identify a jet, such as a knot. Using higher resolution radio data could be used to identify jets from quasars which would otherwise slip past other testing methods.

The lack of observed jets could also indicate that B3 0727+409 is an anomalous object. Perhaps it has an unusually high jet bulk velocity, or some other effect is at play. It is still most likely that the X-ray emission from it is iC-CMB emission, as it closely aligns with the predictions of the model. This can especially be seen with the new longer-exposure image as it follows the iC-CMB prediction of low loss of speed on kiloparsec scales. Further spectral analysis of the jet of B3 0727+409 could also reveal new and interesting information, however, that was not possible in this project due to time constraints.

In the future, measuring polarisation of X-rays jets could shed new light on their emission mechanism. In the past, X-ray polarimetry has been unsuccessful for most extragalactic sources, however new technology might change that soon. The Imaging X-ray Polarimetry Explorer (IXPE) due to launch in 2021 (Soffitta, 2017) aims to measure X-ray polarisation from, among other things, quasars. It utilises gas filled X-ray detectors to measure polarization and has a high spatial-resolution compared to past missions. Measurement of polarisation in X-ray jets could provide conclusive evidence to resolve the debate on their emission mechanism. X-ray polarimetry and its future has been reviewed by Kaaret (2014).

Conclusion

Of 186 candidates narrowed down by cross-referencing and narrowing down high-redshift compact radio source, none were found to have resolved X-ray jets. Various methods were utilised to ensure that jet candidates were truly jets – including PSF simulations and utilisation of full (unprocessed) Chandra images – but none were found that were not PSF smearing effects or background noise. The object found by Simionescu et al. (2016), was recovered by these methods – demonstrating their validity. The lack of observations could indicate problems with iC-CMB model, or simply indicate that jets with high X-ray to radio flux ratios are rare.

A mosaic-ed image of B3 0727+409 from the new observation revealed new details in the jet, including the disappearance and subsequent re-appearance of the jet in the X-ray. This was previously unclear in the original observation and could shed new light on the nature of emission in the jet, however, there wasn’t been enough time to study it more. This could be analysed in future projects.

Future projects could focus on the spectral energy distributions of known jets and conduct statistical analyses of spectral characteristics of high-redshift X-ray jets. Finding a way to implement higher-resolution radio data would also be worth exploring as the FIRST data is not sufficient to describe the morphology of high-redshift quasars. This can especially be seen in the B3 0727+409 radio data, in which one of the hotspots in the jet is unseen in FIRST data. This could provide some much-needed evidence in the debate regarding emission mechanism.

Appendices

Appendix A – Distortion of Source Spectrum

Derivation from (Longair, 2011)

From the diffusion loss equation:

dN(E)dt=D ∇2 NE+∂∂EbEnE+Q(E)

Where

Q(E)is the source spectrum term,

bE=-(dE/dt) and

n(E) is the observed spectrum term. The first term on the RHS is a diffusion term.

For steady state, LHS goes to 0. We also assume diffusion term is 0 for simplicity. Then the equation simplifies to.

ddEbEnE=-Q(E)

Assuming standard source spectrum term of

QE=kE-p, this can be integrated as follows

nE=kE-(p-1)(p-1)b(E)

… QED

Appendix B – PSF Simulation

The PSF simulations are done using the CIAO software suite. First, ray-tracing of the reflections along Chandra’s High-Resolution Mirror Assembly or HRMA is done using the ChaRT (Chandra Ray Tracer) online interface. This takes inputs of the source characteristics, pointing angle of Chandra and coordinates of the target. This is then projected onto the detector using the MARX tool in CIAO. Finally, this is compiled into a fits file to return the simulation of the PSF which can be viewed through the viewing software ds9.

References

Baade, W. (1956) ‘Polarization in the Jet of Messier 87.’, The Astrophysical Journal, 123, p. 550. doi: 10.1086/146194.

Barthel, P. D. (1989) ‘Is every quasar beamed?’, The Astrophysical Journal, 336, p. 606. doi: 10.1086/167038.

Celotti, A. and Fabian, A. C. (1993) ‘The kinetic power and luminosity of parsec-scale radio jets – an argument for heavy jets’, Monthly Notices of the Royal Astronomical Society, 264(1), pp. 228–236. doi: 10.1093/mnras/264.1.228.

Chandra X-ray Centre (2017) Welcome to PSF Central – CIAO 4.10. Available at: http://cxc.harvard.edu/ciao/PSFs/psf_central.html (Accessed: 23 April 2018).

Chartas, G. et al. (2000) ‘The Chandra X‐Ray Observatory Resolves the X‐Ray Morphology and Spectra of a Jet in PKS 0637−752’, The Astrophysical Journal. IOP Publishing, 542(2), pp. 655–666. doi: 10.1086/317049.

Curtis, H. D. (1918) ‘Descriptions of 762 Nebulae and Clusters Photographed with the Crossley Reflector’, Publications of Lick Observatory, vol. 13, pp.9-42, 13, pp. 9–42. Available at: http://adsabs.harvard.edu/abs/1918PLicO..13….9C (Accessed: 10 April 2018).

Flesch, E. W. (2017) ‘VizieR Online Data Catalog: The Million Quasars (Milliquas) catalog (V5.2) (Flesch, 2017)’, VizieR On-line Data Catalog: VII/280. Originally published in: 2015PASA…32…10F, 7280. Available at: http://adsabs.harvard.edu/abs/2017yCat.7280….0F (Accessed: 19 April 2018).

Fruscione, A. et al. (2006) ‘CIAO: Chandra’s data analysis system’, in Silva, D. R. and Doxsey, R. E. (eds) Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, p. 62701V. doi: 10.1117/12.671760.

Hirotani, K. et al. (2000) ‘Pair Plasma Dominance in the Parsec-Scale Relativistic Jet of 3C345’. doi: 10.1086/317769.

Kaaret, P. (2014) ‘X-Ray Polarimetry’, eprint arXiv:1408.5899. Available at: http://arxiv.org/abs/1408.5899 (Accessed: 24 April 2018).

Krawczynski, H., Böttcher, M. and Harris, D. E. (2012) Relativistic jets from active galactic nuclei. Weinheim: Wiley-VCH.

Longair, M. S. (2011) High energy astrophysics. Cambridge University Press.

Marshall, H. L. et al. (2018) ‘An X-ray Imaging Survey of Quasar Jets — The Complete Survey’, The Astrophysical Journal, Volume 856, Issue 1, article id. 66, 25 pp. (2018)., 856. doi: 10.3847/1538-4357/aaaf66.

McKeough, K. et al. (2016) ‘DETECTING RELATIVISTIC X-RAY JETS IN HIGH-REDSHIFT QUASARS’, The Astrophysical Journal. IOP Publishing, 833(1), p. 123. doi: 10.3847/1538-4357/833/1/123.

Meyer, E. T. et al. (2016) ‘AN HST PROPER-MOTION STUDY OF THE LARGE-SCALE JET OF 3C273’, The Astrophysical Journal. IOP Publishing, 818(2), p. 195. doi: 10.3847/0004-637X/818/2/195.

Meyer, E. T. et al. (2017) ‘New ALMA and Fermi/LAT Observations of the Large-scale Jet of PKS 0637−752 Strengthen the Case Against the IC/CMB Model’, The Astrophysical Journal Letters. IOP Publishing, 835(2), p. L35. doi: 10.3847/2041-8213/835/2/L35.

Penrose, R. and Floyd, R. M. (1971) ‘Extraction of Rotational Energy from a Black Hole’, Nature Physical Science. Nature Publishing Group, 229(6), pp. 177–179. doi: 10.1038/physci229177a0.

Sandage, A. (1965) ‘The Existence of a Major New Constituent of the Universe: the Quasistellar Galaxies.’, Astrophysical Journal, 141, p. 1560. doi: 10.1086/148245.

Schwartz, D. A. et al. (2003) ‘CHANDRA Observations of X-ray jet structure on kpc to Mpc scales’, New Astronomy Reviews, 47(6–7), pp. 461–465. doi: 10.1016/S1387-6473(03)00072-1.

Sikora, M. and Madejski, G. (1999) ‘On Pair Content and Variability of Sub-Parsec Jets in Quasars’, The Astrophysical Journal, Volume 534, Issue 1, pp. 109-113., 534, pp. 109–113. doi: 10.1086/308756.

Simionescu, A. et al. (2015) ‘Serendipitous discovery of an extended X-ray jet without a radio counterpart in a high-redshift quasar’. doi: 10.3847/2041-8205/816/1/L15.

Simionescu, A. et al. (2016) ‘Serendipitous discovery of an extended X-ray jet without a radio counterpart in a high-redshift quasar’. doi: 10.3847/2041-8205/816/1/L15.

Soffitta, P. (2017) ‘IXPE the Imaging X-ray Polarimetry Explorer’, in Siegmund, O. H. (ed.) UV, X-Ray, and Gamma-Ray Space Instrumentation for Astronomy XX. SPIE, p. 19. doi: 10.1117/12.2275485.

Taylor, M. ~B. (2006) ‘STILTS – A Package for Command-Line Processing of Tabular Data’, in Gabriel, C. et al. (eds) Astronomical Data Analysis Software and Systems XV. (Astronomical Society of the Pacific Conference Series), p. 666.

Taylor, M. B. (2005) ‘TOPCAT & STIL: Starlink Table/VOTable Processing Software’, 347. Available at: http://articles.adsabs.harvard.edu/cgi-bin/nph-iarticle_query?2005ASPC..347…29T&data_type=PDF_HIGH&whole_paper=YES&type=PRINTER&filetype=.pdf (Accessed: 20 April 2018).

Wardle, J. F. C. et al. (1998) ‘Electron-positron jets associated with the quasar 3C279’, Nature. doi: 10.1038/26675.

Wolter, H. (1952a) ‘Spiegelsysteme streifenden Einfalls als abbildende Optiken für Röntgenstrahlen’, Annalen der Physik. Wiley-Blackwell, 445(1–2), pp. 94–114. doi: 10.1002/andp.19524450108.

Wolter, H. (1952b) ‘Verallgemeinerte Schwarzschildsche Spiegelsysteme streifender Reflexion als Optiken für Röntgenstrahlen’, Annalen der Physik. Wiley-Blackwell, 445(4–5), pp. 286–295. doi: 10.1002/andp.19524450410.

Worrall, D. M. (2009) ‘The X-ray jets of active galaxies’, Astron Astrophys Rev, 17, pp. 1–46. doi: 10.1007/s00159-008-0016-7.

---

[1] Namely coma, a ‘tail’ seen due to the inability to focus light effectively at off-axis angles.