Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://hea-www.harvard.edu/QEDT/Papers/joasia1.ps Äàòà èçìåíåíèÿ: Thu Sep 3 22:57:26 1998 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 23:46:08 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ion drive

SEDs vs. Emission­Line Correlations in Low Redshift
Quasars
Joanna Kuraszkiewicz 1 , Belinda J. Wilkes, Paul J. Green, Smita
Mathur, & Jonathan C. McDowell, Bo — zena Czerny 2
SAO, Cambridge, MA, USA
Abstract. We investigate the relations between the observed emission
line strengths, widths and continuum properties of a sample of low red­
shift quasars for which contemporaneous IR­soft­X­ray spectral energy
distributions (SEDs) are available. This includes investigating correla­
tions between optical/UV lines with both the luminosity and the shape
of the quasars' continua, as well as correlations between the various lines.
Our data do not favor a model in which changes in continuum shape
(due to e.g. ionization parameter decreasing with luminosity) cause the
Baldwin effect. The data can instead be explained by an accretion disk
(AD) model in which limb darkening and the projected surface area of an
optically thick, geometrically thin disk combine to cause a viewing­angle
dependent apparent optical/UV (OUV) luminosity and a more isotropic
X­ray luminosity. The scatter in our correlations is larger than that
expected from this AD model, suggesting the presence of dust which red­
dens both the continuum and the broad emission lines. We also look in
greater detail at the line properties of NLS1 galaxies which do not follow
the Baldwin effect for the remainder of the sample and discuss reasons
for the discrepancy.
1. Introduction
The strong, broad emission lines which characterize quasar spectra are generally
believed to be generated in a large number of small gas clouds photoionized by
the central continuum source of a quasar. To date photoionization models have
been reasonably successful in predicting the average emission line properties of
a quasar using an average continuum shape. However it has become clear that,
while the emission line properties are largely similar from quasar to quasar, the
observed spectral energy distributions (SEDs) are not (Elvis et al. 1994). If
photoionization models are generally applicable, we would expect systematic
relations between the observed lines and continuum in different objects (Krolik
& Kallman 1988) and, at first glance at least, the dichotomy between continuum
and the line behavior looks surprising. We will show that this behavior can be
1 also N.Copernicus Astronomical Center, Warsaw, Poland
2 N.Copernicus Astronomical Center, Warsaw, Poland
1

explained by an accretion disk model, surrounded by a dusty torus, with the
addition of small amounts of dust reddening both the lines and the continuum.
2. Sample
We have obtained a small sample of uniform, high quality continuum and emis­
sion line data. The sample consists of 41 low redshift (z ! 1) quasars, among
which 18 are radio­loud, selected to have high­quality X­ray data in the Einstein
archives. We have obtained far­infrared (IR) through soft X­ray (100 ¯m -- 3.5
keV) continuum data and low­resolution (5­20 š A) optical (our observations) and
ultraviolet spectra (IUE archives). We have measured all the line and continuum
parameters ourselves to minimize the scatter generally introduced by combining
datasets from different techniques and different analysis. The range in shapes
of the IR--ultraviolet continuum is large even in this small, low­redshift sample,
allowing investigation of whether the range in continuum shapes produces any
corresponding range in the emission line properties. The X­ray selection results
in a bias towards low ff ox quasars.
3. Analysis and Results
We investigate the relations between various emission line parameters (fluxes,
equivalent widths (EW) and full­width at half maximum intensity (FWHM) of
all the prominent optical and UV emission lines) and continuum parameters
(L opt , LX , decade and octave luminosities, broad­band luminosities: LUVOIR ,
LBOL , Lycon, L Ion and spectral slopes: ff ox , ff x , ff OUV ). We use the ASURV
statistical package (Isobe, Feigelson, & Nelson 1986) which includes allowance
for the presence of upper limits in the sample. Specifically we applied the fol­
lowing tests to each pair of parameters: the Generalized Kendall Rank and the
Spearman Rank test, which is insensitive to outlying points. We considered a
correlation real only if the probability of it occurring by chance was !2% in both
of these tests. We also look for primary correlations among the line/continuum
relations ie. test which correlation is strongest, while other variables are held
constant.
We find anticorrelations between the equivalent­width (EW) and various
OUV luminosities for the Lyff and Hfi lines (Fig.1a,b) and a marginal anticorre­
lation for CIII]. Here we use the expression ``Baldwin effect'' (BEff) in a general
sense to refer to an anticorrelation of line EW with any OUV luminosity. The
exclusion of seven narrow line, low luminosity AGN reveals similar BEff relation
for both CIII] and CIV lines (Fig. 1c). This suggests that NLS1s have system­
atically low carbon line EW for their luminosity. We will return to this problem
in Section 5.
A significant anticorrelation with ff ox is seen for the EW of CIV line (Fig. 1d),
and a marginal anticorrelation for Hfi and OVI. We do not find any correlations
between lines and the X­ray luminosity or X­ray slope. The FeII optical mul­
tiplet also does not show a simple relationship with luminosity or continuum
slope, however there is a tendency for objects with flat X­ray spectra and/or
strong X­ray luminosities to have weak FeII.
2

Figure 1. a,b,c ­ Correlations between line equivalent widths and
luminosity. Radio­loud quasars are indicated by stars and radio­quiet
by circles; upper limits by arrows. In Fig. 1c filled circles denote NL S1
objects which destroy the CIV Baldwin effect. d ­ Correlation between
CIV equivalent width and ff ox .
4. The Baldwin Effect
The inverse correlation between the emission line equivalent widths and the
UV luminosity is commonly known as the Baldwin effect (Baldwin 1977). It
indicates that the line flux is increasing more slowly than the local continuum
(i.e. is constant) as the quasar luminosity increases. It has been shown that
the Baldwin effect in CIV and Lyff lines covers 7 orders of magnitude in quasar
luminosity (Kinney et al. 1990). In comparison our sample covers ¸ 2 orders of
magnitude and shows a steeper slope in the EW (L) relation. There have been
a number of explanations for the BEff suggested:
Zheng & Malkan (1993) ­ explain it as due to the shift of the big blue bump
towards lower energies with increasing luminosity. Hence the higher luminosity
quasars have a lower fraction of higher energy ionizing photons. This causes the
equivalent widths of high ionization emission lines sensitive to the X­ray contin­
uum (CIV and HeII) to decrease at high luminosities relative to lower­ionization
3

emission lines (Lyff, CIII], and the Balmer lines). The scenario thus predicts
a stronger and more easily detectable Baldwin effect for higher­ionization lines.
Our data do not favor this model as we see the Baldwin effect equally for both
high and low­ionization lines.
Mushotzky & Ferland (1984) ­ explain the Baldwin effect as due to a system­
atic decrease in ionization parameter, U, as the luminosity increases. As the
ionization parameter depends on the number of ionizing photons, U depends on
the continuum shape.
The model predicts the Baldwin effect for the CIV line only, as the CIV
line luminosity increases rapidly with increasing U (decreasing UV luminosity
and ff ox ­ see Fig.1 in Mushotzky & Frland 1984). The Lyff, CIII], Hff, and Hfi
do not show any Baldwin effect in this scenario. Additionally equivalent widths
of Lyff, CIII], CIV should be, according to the model, relatively independent of
both ff ox and ff x as the lines do not originate in the X­ray heated zones deep in
the emission­line clouds. The model is not consistent with our data.
The Accretion Disk (AD) Model ­ Netzer (1985,1987), Netzer et al. (1992) If
the optically thick accretion disk, in the center of an AGN, is geometrically thin,
then limb darkening and change in projected surface area will result in a range of
apparent OUV luminosities if the disk is viewed from different angles. The UV
emission is strongest when the disk is viewed face­on, while the X­ray emission
is more isotropic. The BELR radiates isotropically as the clouds are at distances
greater than the size of the UV­emitting part of the disk. The differing viewing
angle dependence of the lines and the continuum results in an anticorrelation of
line equivalent widths with OUV luminosity and ff ox . For the more isotropic X­
ray emission, such correlations are not expected. The predictions of this model
are consistent with our data.
The AD model predicts both the slope of the line vs. driving continuum
(ie. heating/ionizing continuum defined after Krolik & Kallman 1988) correla­
tions and scatter around the mean. Both are dependent on the range of disk
inclination angles seen by the observer and the luminosity range of the observed
sample. For a larger range in both disk inclinations and/or observed range of
luminosities the slopes are flatter and the scatter larger. Most of the lines vs.
driving continuum correlations (except for CIII] and CIV which are also colli­
sionally excited) have slopes as predicted by the model, with the assumption
that a dusty torus with a standard opening angle of 60 0 obscures the AD (and
prevents the observer seeing the AGN at larger viewing angles).
The scatter around the mean slope in our sample is larger than that ex­
pected from the model. We interpret this as due to random obscuration of lines
and continuum by dust lying outside the BLR.
4.1. The role of dust
Several properties of our sample suggest the presence of dust: first we see larger
Balmer decrements than the canonical ``case B'' value, second the primary cor­
relations between lines and continua are not with their driving continuum.
To check how dust can change the primary correlations and how much scat­
ter it can produce, we have constructed a hypothetical sample of objects, having
a range in bolometric luminosity matching our real sample, and reddened by a
4

random amount of dust, the maximum reddening was based on the range of
big blue bump (BBB) strengths in the real sample. Before being reddened each
hypothetical object was assumed to have a SED matching the quasar with the
largest BBB in the real sample (PG 1426+015). We have chosen two extinc­
tion curves: the standard Galactic extinction curve (Seaton 1979) and the Small
Magellanic Cloud (SMC) extinction curve (Pr'evot et al. 1984), which corre­
sponds well to the extinction by dust dominated by small amorphous carbon
grains (Czerny et al. 1995). Two lines widely separated in wavelength, Lyff and
Hfi, were studied. We found that the hypothetical sample randomly reddened
by dust lying outside the BLR, with maximum extinction E(B \Gamma V ) = 0:2, and
following the SMC extinction curve, has the same primary correlations for Lyff
and Hfi lines as the real sample. The scatter in the line vs. driving continuum
correlations is comparable to the additional scatter required in comparing our
results with the accretion disk model.
We conclude that the Netzer AD disk model surrounded by an optically
thick dusty torus with a 60 0 opening angle, and addition of small amounts of
dust outside the BLR (reddening both the lines and continuum) explains well
our observed line­continuum correlations.
4.2. How does our sample fit into the global Baldwin effect?
The BEff has been seen to cover many orders of magnitude in LUV . A com­
parison with the sample of Kinney et al. (1990), which shows a BEff for Lyff
and CIV lines over 7 orders of magnitude in luminosity, reveals that our sample,
which covers only a factor of 100 in LUV , has a steeper slope of the BEff. This
is in agreement with the AD model predictions, as samples with smaller lumi­
nosity range should show in the EW vs LUV diagram mostly the dependence
from the disk inclination (hence having steeper slopes ­ see Netzer et al. 1992)
which in larger luminosity range samples is superposed with other processes and
adds only scatter to the general BEff. So one possible way of reducing the BEff
scatter, and hence opening the possibility of using the BEff as a standard candle
would be to select sources with a small range of orientation.
As has also been shown in Fig.1c NLS1s are sampling a different region
of the BEff, suggesting that selection of objects with the similar FWHM could
reduce the scatter in the BEff.
5. NLS1
Having found that NLS1 galaxies follow a different BEff relations for CIV (Fig.1c)
we decided to look more carefully at the UV spectra of this class of objects. We
have compiled from the literature a list of all currently known NLS1s and looked
for their UV spectra in the HST and IUE archives. We ended up with a sam­
ple of 11 objects, for which emission line properties were compared with typical
broad­line AGN (Seyfert 1 and quasars). The CIV equivalent widths of NLS1s
were found to be significantly smaller than the EW of all other broad­line AGN.
Also the CIII]/Lyff, CIV/Lyff ratios, when compared to these AGN, appeared
smaller (Fig.2). As the carbon lines are collisionally excited, they are sensitive
to the cloud densities at which they form. They are also a sensitive function of
the ionization parameter U. Therefore to find what parameters can cause such
5

Figure 2. Comparison of CIII/Lyff and CIV/Lyff ratios for NLS1
objects (filled squares) and Seyfert 1 galaxies from Wu et al. (1983,
circles) and QSOs from Laor et al. (1995, stars), Christiani & Vio
(1990, crosses), Wilkes et al. (1998, triangles) and narrow line quasars
from Baldwin et al. (1988, diamonds). (For CIII we used the sum of
CIII+SiIII+AlIII to allow comparison with other samples were broad
lines unabled the authors to separate these components.)
6

small ratios and small EW, we calculated the CIII]/Lyff and CIV/Lyff ratios
for different densities and U using the ionization code Cloudy (Ferland 1991)
and the input ionizing continuum of one of our NLS1s (PG 1211+143). The
results are presented in Fig.3, where filled circles represent values of line ratios
for logU = \Gamma3 (where U =
R 1
1Ryd
Lš
hš
4úr 2 cn H
), open squares logU = \Gamma2, and triangles
logU = \Gamma1; horizontal lines show the observed values. It is clear from the figure
that in order to produce such small carbon to Lyff ratios as seen in NLS1s, a
smaller ionization parameter (logU = \Gamma3) and densities larger (by factor 10­
100) than standard (logU = \Gamma2; ae = 10 9:5 cm \Gamma3 ) are needed. We have also
found that the SiIII]–1892/CIII]–1909 and SiIV+OIV]–1400/CIV–1549 ratios
are larger and that AlIII–1857 is stronger than in other broad­line AGN. This is
also consistent with higher densities of the emitting line region (see Rees, Net­
zer, Ferland 1989, Korista et al. 1996). SiIII] is thermalized at a higher critical
density than CIII], so higher SiIII]/CIII] ratios indicate higher densities. Also at
densities higher than critical CIV and Lyff are no longer efficient coolants and
other ions such as AlIII–1857 take over the cooling resulting in stronger lines.
5.1. Why do NLS1 BLR cloud have higher densities?
The broad line region (BLR) gas which is Compton heated will form two phases:
a cool phase with T ¸ 10 4 (the BLR clouds) and a hot phase with T ¸ 10 8 (the
intercloud medium) (see eg. Krolik & Kallman 1988, Czerny & Dumont 1998,
Wandel & Liang 1991). Precise values of these temperatures depend on the
shape of the continuum.
It is known (eg. Puchnarewicz et al. 1995, Boller, Brandt & Fink 1996)
that the big blue bumps (BBBs) of NLS1s are shifted towards higher energies
compared to other AGN. It has been suggested that this is due to higher ratios
of their luminosity to the Eddington luminosity ie. systematically lower masses
(factor of 100) at a given luminosity range than other AGN (Pounds, Done &
Osborne 1995, Wandel 1997). Krolik & Kallman (1988) calculated that for a
NLS1 spectrum the Compton temperature of the hot (intercloud) medium is
smaller by a factor of 4, when compared with a ``normal'' AGN spectrum, while
the temperature of the cold phase is 10 times larger. The density of the BLR
clouds should also change due to the change in the intrinsic spectrum and we
will now estimate exactly by how much. We use the ionization parameter of
Krolik, McKee & Tarter (1981):
\Xi = 2:3F ion
cp
= 2:3F ion
ckaeTc
¯H
(1)
where F ion is the flux above 1 Ryd determined by the ionizing luminosity of the
central source, L ion , and the current radius r (where effects of geometry have
been neglected):
F ion = L ion =4úr 2 : (2)
The two phases coexist at the value of the ionization parameter \Xi h which
scales with the hot phase temperature T h in the following way (Begelman 1983):
\Xi h = 0:65
`
T h
10 8
' \Gamma3=2
(3)
7

Figure 3. Calculated line ratios a) CIV/Lyff, b) CIII]/Lyff, c) Si­
III]/CIII], d) SiIV+OIV]/CIV for different ionization parameters. Tri­
angles denote U = \Gamma1, squares U = \Gamma2 and filled circles U = \Gamma3.
Horizontal lines show the values of the observed line ratios
8

The BLR is most probably radially extended. However, for the purposes
of this estimation we determine a representative radius. If the cloud number
density profile is flatter than r \Gamma2 , then most of the emission would come from
the outer radii of the BLR. As in the case of the Inverse Compton heated coronae
discussed by Begelman et al. (1983), a nearly hydrostatic corona will exist up
to a radius where the temperature of the hot medium is equal to the ``escape''
temperature (ie. the virial temperature). For larger radii the temperature of
the corona is larger than the virial temperature, the corona becomes unstable
and is blown away in a form of a wind. We therefore identify the outer edge of
the BLR r BLR with the radius at which the hot medium temperature is equal
to the virial temperature:
kT h = GMmH
r BLR
: (4)
Combining equations (1)­(4) we estimate the cloud density:
ae cold ¸
L
M 2
\Theta
T 7=2
h
T cold
(5)
Remembering that for the same luminosity range masses of the central black
hole in NLS1s are a factor of 100, T h a factor 4 smaller as compared to other
AGN, and T cold is a factor 10 larger, this implies that the densities of the BLR
should be an order of magnitude higher in NLS1 than in typical AGN. This is
consistent with the results previously obtained from line ratios.
Acknowledgments. BJW, JCM, PJG acknowledge support provided by
NASA through Contract NAS8­39073 (ASC), JK a Smithsonian pred­doc. fel­
lowship at Harvard­Smithsonian Center for Astrophysics and KBN grant 2P03D00410.
SM acknowledges support by NASA grant NAG5­3249 (LTSA) and PJG grant
HF­1032.01­92A awarded by the Space Telescope Science Institute, which is op­
erated by the Association of Universities for Research in Astronomy, Inc., under
NASA contract NAS5­26555.
References
Baldwin, J.A., 1977, ApJ, 214, 679
Baldwin, J.A., McMahon, R., Hazard, C., Williams, R.E., 1988, ApJ, 327, 103
Begelman M. C., McKee, C. F., Shields, G. A., 1983, ApJ, 271, 70
Boller, Th., Brandt, W.N., Fink, H., 1996, AA 305, 53
Christiani, S. & Vio, R., 1990, AA, 227, 385
Czerny, B., Loska, Z., Szczerba, R., Cukierska, J., Madejski, G. 1995, Acta
Astron. 45, 623.
Czerny, B., Dumont, A.­M., 1998, astro­ph/9808014, to appear in AA
Elvis, M., Wilkes, B.J., McDowell, J.C., Green, R.F., Bechtold, J., Willner, S.P.,
Oey, M.S., Polomski, E., Cutri, R., 1994, ApJS, 95, 1
Ferland, G.F., 1991, ``HAZY'', OSU Astronomy Department Internal Report
Isobe, T., Feigelson, E.D., & Nelson P.I. 1986, ApJ, 306, 490.
9

Kinney, A.L., Rivolo, A.R., & Koratkar, A.P. 1990, ApJ, 357, 338.
Korista, K., Baldwin, J., Ferland, G., Vermer D., 1996, ApJS, 108, 401
Krolik, J.H., McKee, C.F., Tarter, C.B., 1981, ApJ, 249, 422
Krolik, J.H., & Kallman, T.R. 1988, ApJ, 324, 714.
Laor, A., Bahcall, J.N., Jannuzi, B.T., Schneider, D.P., Green, R.F., 1995, ApJS,
99, 1
Mushotzky, R., & Ferland, G.J. 1984, ApJ, 278, 558.
Netzer, H. 1985, MNRAS, 216, 63.
Netzer, H. 1987, MNRAS, 225, 55 .
Netzer, H., Laor, A., & Gondhalekar, P.M. 1992, MNRAS, 254, 15.
Pounds, K.A., Done, C., & Osborne J., 1995, MNRAS, 277, L5
Pr'evot, M.L., Lequeux, J., Maurice, E., Pr'evot, E., & Rocca­Volmerange, B.
1984, A&A, 132, 389.
Puchnarewicz, E.M., Mason, K.O., Siemiginowska, A., Pounds, K.A., 1995, MN­
RAS, 276, 20
Rees, M.J., Netzer, H., Ferland G.J, 1989, ApJ, 347, 640
Seaton, M.J. 1979, MNRAS, 187, 73P.
Wandel, A., 1997, ApJ, 490, L131
Wandel, A., & Liang, E.P., 1991, ApJ, 380, 84
Wilkes B.J., Kuraszkiewicz, J. Green, P.J., Mathur, S., McDowell, J.C., 1998,
submitted to ApJ
Wu C.­C., Boggess, A., Gull, T.R., 1983, ApJ, 266, 28
Zheng, W., & Malkan, M.A. 1993, ApJ, 415, 517.
10