Proceedings of The National Conference
On Undergraduate Research (NCUR) 2016
University of North Carolina Asheville
Asheville, North Caorlina
April 7-9, 2017
Searching for a Connection Between Radio Emission and UV/optical
Absorption in Quasars
Sean Haas1, Abdul Moiz Khatri2,3, Carla Quintero1, Pat Hall3
Department of Physics
1
Humboldt State University
1 Harpst Street
Arcata, CA 95521
2
University of Toronto, 3York University
Faculty Advisor: Dr. Paola Rodriguez Hidalgo
Abstract
Quasars, active galactic cores in distant galaxies, are among the most energetic objects in the universe. The
supermassive black holes that power quasars accrete matter and, in the proccess, cause accretion disks to release
massive amounts of radiation and, on occasion, eject large amounts of matter. These ejections, called outflows, can
be observed in quasar spectra as UV/optical broad absorption lines (BALs). Some of these outflows can reach
speeds larger than 0.1c. In this study, 6760 quasar spectra from the Sloan Digital Sky Survey Data Release 9(SDSS
DR9) were analyzed in search of outflows. Of these spectra, 23 quasars were found to have outflows. These spectra
were then cross-correlated with the Very Large Array Faint Images of radio Sky at Twenty cm (VLA FIRST) survey
in order to search for possible trends between radio properties and outflows. None of the quasars in this sample had
any measured radio flux. In comparison, 3.5% of quasars from the SDSS DR9 quasar catalog have measured radio
flux. These findings suggest that radio emissions are not a prerequisite for outflows in quasars.
Keywords: Quasars, Outflows, Radio
1. Introduction
1.1 Characteristics of Quasars
Quasars are among the most distinctive and most energetic objects in the night sky. Powered by accretion onto an
accretion disk around the central supermassive black hole, quasars emit huge amounts of radiation over the entire
electromagnetic spectrum[1]. Quasars are also very distant, making them difficult to study. Spectral observations are
the main tool used to study quasars, due to this combination of distance and strong emissions. Quasar spectra differ
greatly from the spectra of stellar bodies. Quasar emissions cover a large range of wavelengths instead of resembling
a discrete blackbody curve, as observed in stellar spectra[1].
1.2 Outflows and Jets
One of the more prominent features of quasars are axial radio jets, which are present in roughly 10% of quasars.
These jets, composed of highly accelerated charged particles, emit large amounts of radio waves[1]. These radio
emissions make it possible to detect quasar jets by simply measuring the object’s radio flux. Quasar outflows are
another, more recently documented, phenomenon of note. An outflow is an event where a quasar ejects large
amounts of matter. These outflows can be observed in quasar spectra as blueshifted absorption features[3]. This
blueshift occurs as matter ejected from the quasar travels towards Earth while absorbing light from the quasar. One
problem with this observation method is that only outflows positioned between the quasar and Earth can be detected.
These outflows are sometimes detected as broad absorption lines (BALs), typically with widths over a few thousand
km/s[3]. This broadening of absorption lines is due to the matter in the outflow moving at differing speeds. The
speed of these outflows can be measured by comparing the location of the BAL to its expected location in the
quasar’s rest frame.
1.3 Usage of CIV
The absorption of CIV, triple-ionized carbon, is one of the most commonly used markers for these observations.
This ion is chosen for a few reasons. Firstly, carbon is a very common element, so it can be assumed that any given
outflow will contain at least some amount of carbon. Secondly, CIV absorption features are easy to observe, as some
transitions fall into the UV/optical region of the electromagnetic spectrum. Specifically, the 1548 and 1550Å
transitions produce strong absorption features and are easy to observe in the UV/optical region of the spectrum. This
combination of factors make CIV a useful marker for studying outflows in quasars.
1.4 Goal of this Study
The goal of this study is to investigate if there is any connection between outflows and radio properties in quasars.
By investigating these properties we seek to better characterize quasar outflows and how they relate to jets. We
studied this by first searching for quasar spectra with outflows, and then investigating their radio properties. In
section 2 we will explain the data sources we used for this study. Section 3 will explain, in detail, how this research
was carried out. In section 4 we will cover our results and conclusions, and how our findings relate to past research.
2. Data
Two main data sources were used in this study: the Sloan Digital Sky Survey Data Release 9 (SDSS DR9) and Very
Large Array Faint Images of Sky at Twenty-cm (VLA the Radio FIRST). SDSS is an automated UV/optical survey
of the night sky. Data from this survey is published as numbered data releases, and is made available online in the
form of optical imagery and spectrographic data. SDSS DR9 was used as a source of spectra for the spectral analysis
program.
VLA FIRST is a radiometric survey of the night sky that targets faint objects, and was chosen for this study
because it contains observations for many of the objects surveyed in SDSS. We cross correlated the spectra using the
DR9 Quasar Catalog, a value added catalog of quasar data that contains information from SDSS DR9 and VLA
FIRST[2].
2.1 Data Selection
We chose inputs for the spectrum analysis program using a number of criteria. The first cutoff was redshift (z). The
spectrographic data from SDSS DR9 covers a range of wavelength from 3600 to 10500 A[2]. The maximum speed
of outflows we studied was 0.2c, placing CIV absorption features at around 1250 A in the rest frame. By choosing a
z ≥ 1.9, these CIV absorption features are shifted to 3600 Å, within the survey's wavelength range.
Another cutoff was signal to noise ratio (S/N). Samples with S/N ≥ 10 were chosen because BALs proved to be
hard to detect over noise in spectra with lower S/N. The SDSS DR9 catalog contains 87822 quasar samples. After
applying these filtering criteria, S/N ≥ 10 and z ≥ 1.9, the sample was reduced to 6760 quasars.
3. Methods
The first step in our study was to find a sample set of quasars with outflows so that their radio properties could be
examined. We wrote a Python program to search for quasars with outflows. These quasars were found by searching
quasar spectra from the SDSS DR9 quasar catalog for broad absorption features between the SiIV emission line and
the Lyman alpha forest.
To find BALs we first removed pixels with high error values from the spectra. Each bad pixel was converted to the
flux value of the point next to it. Next, the spectra were normalized using a powerlaw fit. This normalization was
3046
needed because most quasar spectra are bluer at shorter wavelengths, which causes the spectrum to have a negative
slope. After normalization a 3 point median boxcar method was applied to smooth the continuum.
(1)
Absorption was identified using the balnicity index (BI), a measure of the width of absorption features. Equation
(1) shows how the balnicity index is calculated. The f(v) value is the normalized flux value in terms of velocity
taken from the spectra. The C term is set to 1 if the value inside the square brackets is continuously positive over a
given range of velocities, otherwise, C is set to 0[2]. This range used for our C value selects for the minimum
width of broad absorption lines to be measured. We chose a to measure BALs with a width greater than 600
km/s. This allowed the program to detect both BALs and mini-BALs, essentially more narrow broad absorption
lines.
Figure 1. Example of normalized quasar spectra from our final data set, used with permission from Abdul Khatri.
SDSS spectra (green) smoothed and plotted with a guide line at 1 to aid visual inspection (blue). Absorption features
between the Lyman alpha forest and SiIV emission line are marked with a red vertical line. The black vertical line
marks the expected location of CIV absorption if the absorption marked by the left red line was SiIV. The features
marked by the red vertical lines are examples of BALs.
The next step was to rule out spectra where absorption was due to ions other than CIV. We used CIV as a marker
3047
for absorption in this study because of it's relative abundance compared to other ions. This abundance makes it so
that we can reasonably assume any outflow contains CIV. Absorption features in the region between the Lyman
alpha forest and the SiIV emission line can be from SiIV, CII, OI, SiII, or CIV. These other ions had to be ruled out
to be certain that the observed absorption features were due to CIV. The black vertical line in Figure 1 shows where
we would expect to find a CIV absorption feature if the observed absorption features were due to SiIV. Since there is
no absorption feature at this location, we can rule out SiIV. Spectra were eliminated if absorption features were
present in this expected location. We carried out this procedure for CII, OI, and SiII. Spectra not eliminated by this
procedure, then, contain possible evidence of outflows.
The quasars that this program identified as having evidence of outflows were then visually inspected. From this
inspection a set of quasars with definite evidence of outflows were selected. An example of a spectra with evidence
of outflows found by this program is shown in figure 1. These samples were then cross correlated with VLA FIRST
to identify their radio properties. This cross correlation was carried out using the SDSS DR9 quasar catalog[2].
4. Discussion
4.1 Results Of Spectral Analysis
From the filtered set of 6760 spectra from the SDSS DR9 catalog the analysis program identified 41 spectra as
having possible evidence of outflows. Visual inspection was carried out to remove spectra that our program
incorrectly marked as having outflows. This left us with a set of 23 quasars with definite evidence of outflows. Cross
correlation with VLA FIRST radiometric observations were carried out using data in the SDSS DR9 quasar catalog.
From this cross correlation it was found that none of the quasars in the final sample set had any measured radio flux.
Applying Poisson statistics, we get an error of, at most, +/-1 cases. A total of 3076 quasars in the SDSS DR9 quasar
catalog, about 3.5% of the sample, have some radio flux. Our small sample of 23 would imply that +/-1 cases are
within the 3.5% of the general sample.
However, this sample may not be completely representative of the general population of quasars. Originally, SDSS
surveys targeted quasars based on their radio properties. Specifically, some early targets in the SDSS-I/II were
selected due to their radio emissions as measured in VLA FIRST[4]. Due to this targeting there may be a
disproportionately large number of quasars with measurable radio flux in SDSS sample sets than would be found in
the general population of quasars. This sampling bias needs to be considered when using SDSS data.
5. Conclusions
From these results, it is evident that radio emissions are not prerequisite for outflows in quasars. These findings are
consistent with past works that have found a strong negative correlation between BALs in quasars and radio
emissions[5]. There are known cases of quasars with both BAL and radio emissions, however, these radio loud BAL
quasars are less common. Also, in these radio emitting cases BAL features are thinner than typical BALs[6][7]. One
avenue for future research is to investigate lowering the BAL width threshold used in this study to see if we can find
these radio emitting BAL quasars in the sample.
6. Acknowledgements
We would like to thank our research advisor, Dr. Paola Rodriguez Hidalgo. This project was made possible by
funding from the College of Natural Resources and Sciences at Humboldt State University.
7. References
1. Netzer, Hagai. The physics and evolution of active galactic nuclei. Cambridge: Cambridge University Press,
2013.
2. Pâris, I., et al.. "The Sloan Digital Sky Survey Quasar Catalog: Ninth Data Release." Astronomy &
Astrophysics A&A 548 (2012). doi:10.1051/0004-6361/201220142.
3048
3. Rodríguez Hidalgo, Paola and Hamann, Fred and Hall, Patrick. "The Extremely High Velocity Outflow in
Quasar PG0935+417." Monthly Notices of the Royal Astronomical Society 411, no. 1 (2010): 247-59.
doi:10.1111/j.1365-2966.2010.17677.x.
4. Ross, Nicholas P., et al. "The SDSS-III Baryon Oscillation Spectroscopic Survey: Quasar Target Selection For
Data Release Nine." ApJS The Astrophysical Journal Supplement Series 199, no. 1 (2012): 3. doi:10.1088/00670049/199/1/3.
5. Stocke, John T., Simon L. Morris, Ray J. Weymann, and Craig B. Foltz. "The Radio Properties of the Broadabsorption-line QSOs." ApJ ApJL The Astrophysical Journal 396 (1992): 487. doi:10.1086/171735.
6. Kuncic, Zdenka. "Broad Absorption Line Quasars and the Radio‐Loud/Radio‐Quiet Dichotomy." PUBL
ASTRON SOC PAC Publications of the Astronomical Society of the Pacific 111, no. 762 (1999): 954-63.
doi:10.1086/316408.
7. Rochais, T. B., M. A. Dipompeo, A. D. Myers, M. S. Brotherton, J. C. Runnoe, and S. W. Hall. "Radio-loud
and Radio-quiet BAL Quasars: A Detailed Ultraviolet Comparison." Monthly Notices of the Royal Astronomical
Society 444, no. 3 (2014): 2498-506. doi:10.1093/mnras/stu1635.
3049