1: %\documentclass{aastex}
2: \documentclass[12pt,preprint]{aastex}
3:
4: %% manuscript produces a one-column, double-spaced document:
5:
6: % \documentclass[manuscript]{aastex}
7:
8: %% preprint2 produces a double-column, single-spaced document:
9:
10: % \documentclass[preprint2]{aastex}
11:
12: \newcommand{\vdag}{(v)^\dagger}
13: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
14:
15: \shorttitle{Chandra observations of X-ray weak Quasars}
16: \shortauthors{Risaliti et al.}
17:
18: \begin{document}
19:
20: \title{A Chandra mini-survey of X-ray weak quasars
21: }
22:
23: \author{G. Risaliti\altaffilmark{1,2}, M. Elvis\altaffilmark{1}
24: }
25:
26: \author{
27: %A. Marconi\altaffilmark{2}, R. Maiolino\altaffilmark{2},
28: R. Gilli\altaffilmark{2}, M. Salvati\altaffilmark{2}}
29: \email{grisaliti@cfa.harvard.edu}
30:
31: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics, 60 Garden st.
32: Cambridge, MA 02138 }
33: \altaffiltext{2}{INAF - Osservatorio Astrofisico di Arcetri, L.go E. Fermi 5,
34: I-50125 Firenze, Italy}
35:
36: \begin{abstract}
37: We present {\em Chandra} observations of 18 spectroscopically selected quasars,
38: already known to be X-ray weak from previous ROSAT observations.
39: All the sources but one are detected by {\em Chandra}, and spectral analysis
40: suggests that most of them are intrinsically underluminous in the X-rays (by a
41: factor from 3 to $>100$).
42: These objects could represent a large population of quasars with
43: a Spectral Energy Distribution different from that of standard blue
44: quasars. We discuss the possibility that a significant fraction of the obscured AGN needed
45: in Synthesis models of the X-ray background could be instead optically
46: broad-line, X-ray weak quasars.
47: \end{abstract}
48:
49: \keywords{Galaxies: AGN --- X-rays: diffuse background}
50:
51: \section{Introduction}
52:
53: The blue optical spectrum
54: has normally been used to distinguish quasars
55: from stars and normal galaxies
56: in optical
57: surveys (BQS, Schmidt \& Green 1986; LBQS, Foltz et al.
58: 1990; 2dF, Boyle et al. 2000). As a
59: consequence, our knowledge of quasars is by definition limited to ``blue''
60: quasars.
61: However, many different quasar SED could exist,
62: undiscovered because of the limits of the available
63: instruments and selection criteria (Elvis 1992).
64: The X-ray properties of optically selected quasars show
65: a similar homogeneity: the 1-10 keV spectrum is well represented by a
66: power law with photon index $\Gamma\sim1.8-2$ (Laor et al. 1997, Reeves \& Turner
67: 2000); the average optical to X-ray slope for optically
68: selected samples is $\alpha_{\rm OX}=1.55$ (Laor et al. 1997\footnote{
69: $\alpha_{\rm OX}$ is defined as the spectral index of a power law connecting the
70: points at 2500~\AA~and at 2 keV (rest frame) of the quasar SED in the ($\nu,f_\nu$) plane.})
71: Only a minority of objects ($<10\%$) are significantly weaker
72: ($\alpha_{\rm OX}>1.8$) than the
73: average in the 0.5-2~keV X-ray band (Yuan et al. 1998).
74:
75: When quasars are searched with selection criteria other than the U-B
76: color, different properties emerge. Webster et al. (1995) found a large
77: population of red broad-line quasars in radio surveys. Kim \& Elvis (1998)
78: discovered red quasars in soft X-ray selected samples. Most strikingly,
79: the 2MASS survey (Skrutskie et al. 1997)
80: found a large number of red quasars, whose density
81: is of the same order of that of local color-selected quasars (Cutri et al.
82: 2001). A {\em Chandra} survey of these objects revealed X-ray properties completely
83: different from ``normal'' quasars: the X-ray emission is much weaker and the
84: spectra are flatter, suggesting that these objects suffer significant
85: absorption by circumnuclear gas (Wilkes et al. 2002). The existence of these
86: populations is important. Synthesis models for the X-ray (Comastri et al.
87: 1995, Gilli, Salvati \& Hasinger 2001) and FIR (Risaliti, Elvis \& Gilli 2002)
88: backgrounds depend sensitively on knowing the true AGN population. Similarly,
89: the total accretion luminosity of the Universe (Fabian \& Iwasawa 1999) and
90: the average efficiency of black hole accretion (and hence spin, Elvis,
91: Risaliti \& Zamorani 2002) depend primarily on these ``hidden'' populations.
92:
93: Another indication of unexpected X-ray properties of non-color selected
94: quasars comes from the work of Risaliti et al. 2001 (hereafter R01), where a sample of
95: spectroscopically selected quasars of the Hamburg survey (Hagen et al. 1995)
96: has been cross-correlated with the WGA Catalogue of ROSAT pointed observations
97: (White, Giommi \& Angelini 1995). More than half of the resulting sample
98: is underluminous in the X-rays, by a factor from $\sim 5$ to $>100$.
99: Interestingly, most of these objects are somewhat redder
100: than ``normal'' quasars ($\Delta(B-R)\sim~1$ vs. $\Delta(B-R)\sim~0.5$, R01), and
101: would have been probably missed in standard color-based surveys, since we
102: expect their U-B color also to be redder than the average of standard blue
103: quasars.
104: Almost all the objects in this sample were not detected by ROSAT, therefore
105: the claim on their X-ray weakness is based on upper limits. As a consequence,
106: nothing is known about their X-ray spectral properties.
107:
108: The subarcsecond beam size of the {\em Chandra} mirrors
109: (van Speybroeck et al.
110: 1997) and the large collecting area
111: endows {\em Chandra} with far greater sensitivity than ROSAT. Hence, a
112: {\em Chandra} survey can explore the X-ray sample of quasars
113: far better than ROSAT.
114: Here we present the results of {\em Chandra} observations of 18 objects selected
115: from the X-ray weak sample of R01.
116: \section{Sample selection and observations}
117: The parent sample was obtained by R01
118: from the cross-correlation of the Hamburg Quasar Survey (hereafter HS,
119: Hagen et al. 1995) with the WGA
120: Catalogue. The HS sample consists of 397 quasars, with redshift between 0 and
121: 3, and limiting magnitude B$\sim 19-19.5$. The selection criteria are either
122: the standard U-B colour or the presence of broad emission lines (or both) in grism
123: optical spectra. In this way it is possible to discover objects with
124: intrinsically red continua, or with moderate extinction.
125:
126: The R01 sample contains 85 sources, of which only 31 were
127: detected by ROSAT. In Fig.~1 we show the distribution of X-ray to
128: optical ratios for this sample, compared with the one of PG quasars.
129: We adopted an index, defined in R01 as
130: I$_{\rm OX}=\frac{20-B}{2.5}-\log~\phi$, where $\phi$ is the ROSAT 0.5-2.4 keV count rate.
131: The reason for using this new index, instead of $\alpha_{OX}$, is that
132: it uses the observed optical data since, with these redder quasars,
133: an extrapolation to
134: 2500~\AA~may be problematic.
135: For quasars with a standard SED, I$_{\rm OX}\sim
136: 2.6\alpha_{\rm OX}-1.3$. \\
137: The detected sources in R01 have a rather normal X-ray to optical
138: ratio, while the 54 non-detections are X-ray weaker than most PG quasars.
139: The dashed vertical line in Fig. 1 (I$_{\rm OX}=3.2$) represents the value at which
140: the underluminosity in the X-ray is a factor of 5 with respect to the average
141: of PG quasars. In the following we refer
142: to I$_{\rm OX}>3.2$ as ``X-ray weak'' sources.
143:
144: We randomly selected 17 sources from the X-ray weak half of the sample.
145: As a control sample we included three HS quasars not detected by ROSAT
146: but with upper limits on I$_{\rm OX}$ lower than 3.2.
147:
148: As shown by R01, there is a strong correlation between the optical O-E
149: color, obtained from the POSS plates, and the X-ray to optical ratio. As a
150: consequence, our 17 X-ray weak sources, being a representative sub-sample
151: of the X-ray weak quasars of R01, are automatically also a representative sub-sample of the
152: red quasars present in the Hamburg Survey, i.e. those objects that would have not been
153: selected by optical color-based surveys.
154:
155: 18 out of the 20 sources of the sample have been observed with the ACIS-S
156: detector on {\em Chandra} (Weisskopf et al. 2001) in the
157: year 2002.
158: Out of these 18 objects, 16 are from the X-ray weak group, and the remaining 2
159: are from the small group of 3 control sources.
160: The observing times vary from 4 to 10 ksec, given the B magnitude of
161: the sources. The observing times were chosen in order to have the
162: same lower limit on $\alpha_{\rm OX}$ for all the sources, in case of non-detection
163: with {\em Chandra}.
164: All the sources but one have been detected. The net source counts range from a few
165: tens counts to $\sim 1000$ for a few bright objects (Tab. 1).
166:
167: \section{Analysis and Results}
168:
169: The data were analyzed using the latest ACIS calibrations provided by the
170: {\em Chandra} X-ray Center.
171: A correction was applied to the response matrices to account for
172: the low energy quantum efficiency
173: degradation of the ACIS detector\footnote{URL:http:$//$asc.harvard.edu$/$ciao$/$threads$/$apply\_acisabs$/$index.html}.
174: We analyzed the ACIS spectra using a simple model, consisting of an
175: absorbed power law.
176: Both the photon index, $\Gamma$ and the absorbing column
177: density $N_H$ were left free. (Table 1)
178: In Fig. 2 we show the results for the 16 objects of the X-ray weak sample, compared with the
179: spectra estimated from the B magnitudes, assuming the normal UV-selected value of
180: $\alpha_{\rm OX}=1.55$ (shaded
181: band in each panel of Fig. 2).
182: %We also estimated the $\alpha_{\rm OX}$ for the objects in two different ways (Table 1).
183: We calculated $\alpha_{\rm OX}$ deriving the monochromatic flux at 2500~\AA~from the B
184: magnitude, assuming a spectral shape $f_\nu \propto \nu^{-0.5}$.
185: This is a conservative assumption because the optical colors
186: of these objects are on average redder than in normal quasars. Since for all but three
187: sources the B magnitude central wavelength ($\lambda_B=$4400~\AA) is greater than the
188: redshifted 2500 \AA~wavelength, using flatter optical spectra would imply higher
189: extrapolations of $f_\nu$ at 2500 \AA, and so even higher values of $\alpha_{\rm OX}$.
190: The rest-frame 2 keV flux
191: was directly measured from the spectrum.
192: %In the second
193: %case we fixed the photon index to $\Gamma=1.8$ and left N$_H$ free. We then
194: %extrapolated the intrinsic flux at 2 keV, assuming no absorption.
195: %The resulting index is therefore a measure of the intrinsic $\alpha_{\rm OX}$ of
196: %the source.
197: The results can be summarized as follows:\\
198: $\bullet$ 12 out of the 16 X-ray weak sources are confirmed to be
199: extremely faint in the X-rays, with $\alpha_{\rm OX}$ ranging from 1.7 to 2.3.
200: Another object, HS 1417+4522, is only marginally weaker than the average at 2 keV
201: ($\alpha_{\rm OX}=1.58$), but is significantly weaker than normal AGNs at higher energies
202: (Fig. 2).
203: The spectra are on average slightly flatter than the canonical $\Gamma=1.8$ (Risaliti 2002):
204: we computed a stacked spectrum of these 13 sources and we obtained
205: an average photon index $\Gamma=1.5$.\\
206: $\bullet$ 3 out of 16 objects have a ``normal'' $\alpha_{\rm OX}$, significantly
207: higher than the estimate from the ROSAT upper limit. This implies strong
208: variability (of at least a factor $\sim 10$ in the 0.5-2 keV band).
209: Interestingly, these three objects are the only ones
210: in our sample at redshift lower than unity. We will further discuss these sources
211: in a forthcoming paper.\\
212: $\bullet$ The two ``control sources'' both show a ``normal'' X-ray spectrum.
213: As a consequence, there is no indication that a significant fraction of the
214: sources in the left part of the histogram in Fig. 1 are X-ray weak. Our best
215: estimate of the fraction of X-ray weak quasars in the Hamburg quasars remains
216: the one inferred from the ROSAT observations, i.e. $\sim 50\%$.
217: \section{Discussion}
218: Since 13 out of 16 objects are confirmed to be X-ray weak by {\em
219: Chandra}
220: observations,
221: the fraction of X-ray weak sources in the
222: parent sample of R01 is as high as 13/16 of 50\%, i.e. $\sim~40$\%.
223: Since the space density of these sources is of the
224: same order of that found in color-selected surveys with similar limiting
225: magnitudes, these objects represent a significant part of the AGN population.
226: From the X-ray point of view, they appear to be completely different from
227: standard blue quasars both in $\alpha_{\rm OX}$ and $\Gamma$.
228: In principle, two interpretations are possible for the above results:
229: the sources can be either (1) heavily absorbed or (2) intrinsically X-ray weak.
230:
231: (1) If absorption plays a crucial role, the observed radiation could be due to
232: warm scattering and/or cold reflection\footnote{Note that diffuse emission from the
233: host galaxies is expected to give little contribution, the observed luminosity
234: being higher than 10$^{44}$ erg s$^{-1}$.} while the intrinsic emission would be
235: absorbed by a column density $N_H>10^{24}$~cm$^{-2}$.
236: These objects are somewhat redder than normal ``blue
237: quasars''
238: %(although
239: %enough to prevent these objects from being missed by classical optical color-based
240: %surveys)
241: and are all broad line quasars. We would then have an unlikely
242: column density distribution, with all the optically selected blue quasars
243: having
244: $N_H<10^{21}$~cm$^{-2}$, all the redder quasars having
245: $N_H>10^{24}$~cm$^{-2}$, and nothing in between.
246: This is an argument favoring the alternative hypothesis
247: of intrinsic X-ray weakness.
248: %Even if the weakness at high energies (E$>$ 2~keV) cannot be ascribed to absorption,
249: %part of our sources show a decrease at $\sim 1$~keV greater than that expected
250: %from Galactic absorption.
251: %This effect is likely due to nuclear absorption by a column density in the range
252: %$10^{21}-10^{22}$~cm$^{-2}$. This issue will be further discussed in a
253: %forthcoming paper dedicated to the detailed spectral analysis of this sample.
254: %(Risaliti et al. 2003a in preparation).\\
255:
256: (2) If the quasars are intrinsically X-ray weak, the accretion disk/corona
257: system could be in a different state than in normal quasars. For example,
258: a weaker corona would naturally produce a weaker X-ray emission.
259: A widely accepted model is that an X-ray emitting corona is generated by the
260: Magneto-Rotational Instability (MRI, Balbus \& Hawley 1991).
261: This MRI generates the viscosity in the accretion disk, so if MRI is
262: ineffective little energy should be liberated and the UV continuum and
263: the emission lines should be weak. Yet most
264: most of our objects
265: have been selected through the CIV 1549~\AA~line, and therefore it is unlikely
266: that they have weak ionizing continua. This paradox clearly needs investigating
267: theoretically. Perhaps MRI does not produce disk viscosity or the bulk of the X-rays
268: in normal quasars have another origin.\\
269: Our results show that the current view of the X-ray
270: properties of quasars could be strongly biased by the optical selection
271: towards X-ray loud and
272: steep-spectra objects. This is supported by the fact
273: that the other known red quasars have properties similar to our objects (see
274: for example Wilkes et al. 2001 for {\em Chandra} observations of 2MASS quasars).
275:
276: As can be seen from Table 1, the X-ray luminosities of our sources are in the range
277: 10$^{44}-10^{45}$ erg~s$^{-1}$, only slightly higher
278: than the typical luminosities where the bulk of the X-ray background is
279: made, according to synthesis models (Gilli et al. 2001). Also, the
280: average spectral
281: properties ($\Gamma_{AV}=1.5$) are close to those needed by these models.
282: It could well be that a fraction of the sources predicted to have N$_H\sim
283: 10^{22}-10^{23}$~cm$^{-2}$, used by current synthesis models are instead
284: intrinsically weak, flat spectrum sources, with normal broad quasar emission lines.
285: If this is the case, the optical/infrared counterparts of these objects would be
286: completely different from those of the standard type 2 AGNs: the optical emission
287: would not be obscured, but could be redder than standard blue quasars, and therefore less
288: readily distinguished from stellar emission.
289: %This is in agreement with the discovery of
290: %many X-ray weak, optically ``normal'' galaxies in shallow surveys such as the
291: %HELLAS (Fiore et al. 1999).\\
292: Another important difference with respect to the standard view is that
293: the direct optical/UV emission would not be reprocessed
294: at mid/far IR wavelengths.
295: This would significantly lower the expected contribution of AGNs to the far IR background (Risaliti,
296: et al. 2002) and allowing a larger population of such quasars to be present.
297:
298: A key test for this hypothesis will be the comparison between the X-ray and optical properties
299: of the sources with X-ray flux $\sim 10^{-15}$ erg~s$^{-1}$~cm$^{-2}$ in the
300: {\em Chandra} Deep Surveys. These objects have luminosities of the order of 10$^{43}-10^{44}$~erg~s$^{-1}$,
301: where the bulk of the XRB is made, and their observations have enough S/N to distinguish
302: between intrinsically weak and absorbed spectra. This will make clear whether
303: the X-ray sources described in this work are also common at lower luminosities.
304:
305: Another important step to improve the understanding of this class of sources,
306: will be the study of their optical and near-IR spectra:
307: since they are completely different from
308: normal blue quasars in the X-rays, they could also have quite different
309: optical and near-IR SEDs. To explore this issue, we are undertaking
310: optical and near-IR observations of several of the sources in our sample at
311: 4-meter class telescopes.
312: \acknowledgments
313:
314: We are grateful to Alexey Vikhlinin for useful comments, and to the referee,
315: Dr. R. Cutri, for a careful reading of the manuscript.
316: This work was partially supported by NASA Grant GO2-3142X.
317:
318: \begin{thebibliography}{}
319:
320: \bibitem{a5} Balbus, S.A., \& Hawley, J. F. 1991, ApJ, 376, 214
321: %\bibitem{a10} Bechtold, J., et al. 2002, ApJ, in press (astro-ph 0204462)
322: \bibitem{a20} Boyle, B. J., Shanks, T., Croom, S. M., Smith, R. J.,
323: Miller, L., Loaring, N., \& Heymans, C. 2000, MNRAS, 317, 1014
324: \bibitem{a50} Cutri, R. et al. 2001, in ASP Conf. Ser. 232, ed. R. Clowes, A.
325: Adamson \& G. Bromage (San Fransisco: ASP), 78
326: \bibitem{a60} Elvis, M. 1992, in Frontiers of X-ray astronomy
327: (Universal Academy press, Tokyo) eds. Y. Tanaka, K. Koyama, p. 567
328: \bibitem{a63} Elvis, M., Risaliti, G., \& Zamorani, G. 2002, ApJ 565, L75
329: %\bibitem{a65} Elvis, M., et al. 1994, ApJS, 95, 1
330: \bibitem{a75} Gilli, R., Salvati, M., \& Hasinger, G. 2001, A\&A, 366, 407
331: \bibitem{a90}
332: Hagen, H.-J., Groote, D., Engels, D., \& Reimers, D., 1995, A\&AS 111, 195
333: \bibitem{a94}
334: Kim, D. W., \& Elvis, M., 1999, ApJ 516, 9
335: \bibitem{a95} Foltz, C.B., Chaffee, F.H.,
336: Hewett, P.C.,, MacAlpine, G.M.,
337: Turnshek, D.A., Weymann, R.J., \&
338: Anderson, S.F. 1987, AJ, 94, 1423
339: \bibitem{a96} Laor, A., Fiore, F., Elvis, M., Wilkes, B.J., \& McDowell, J.C.
340: 1997, ApJ, 477, L93
341: %\bibitem{a99} Malkan, M.A., \& Sargent, W.L.W. 1982, ApJ, 254, 22
342: \bibitem{a100} Reeves, J.N., \& Turner, M.J.L. 2000, MNRAS, 316, 234
343: \bibitem{a101} Risaliti, G., Marconi, A., Maiolino, R., Salvati, M., \& Severgnini, P. 2001,
344: A\&A, 371, 37 (R01)
345: \bibitem{a104} Risaliti, G., Elvis, M., \& Gilli, R. 2002, Apj, 566, L67
346: \bibitem{a105} Risaliti, G. 2002, A\&A, 386, 379
347: \bibitem{a107} Scmidt, M., \& Green, R.F. 1983, ApJ, 269, 352
348: \bibitem{a110} Skrutskie, M.F., et al., 1997,
349: in The Impact of Large Scale Near-IR Sky Surveys, eds. F. Garzon et al. (Kluwer
350: (Netherlands), 25.
351: \bibitem{a120} van Speybroeck, L.P.; Jerius, D.,
352: Edgar, R.J., Gaetz, T.J., Zhao, P.,
353: \& Reid, P.B. 1997, Proc. SPIE Vol. 3113,
354: Grazing Incidence and Multilayer X-Ray Optical Systems, Eds. R.B. Hoover, A.B. Walker, p. 106
355: \bibitem{a150}
356: Webster, R.L., Francis, P.J., Peterson, B.A., Drinkwater, M.J., \& Masci, F.J., 1995,
357: Nature 375, 469
358: \bibitem{a170} Weisskopf, M.C., Brinkman, B., Canizares, C.,
359: Garmire, G., Murray, S., Van Speybroeck, L. P. 2002, PASP, 114, 1
360: \bibitem{a180}
361: Wilkes, B. J., Schmidt, G. D., Cutri, R. M.,
362: Ghosh, H., Hines, D.C., Nelson, B.,
363: \& Smith, P. S., 2002, ApJ 564, L65
364: \bibitem{a200}
365: White, N.E., Giommi, P., \& Angelini, L., 1995, \\
366: http://heasarc.gsfc.nasa.gov/W3Browse/all/wgacat.html
367: \bibitem{a220}
368: Yuan, W., Brinkmann, W., Siebert, J., \& Voges, W., 1998, A\&A 330, 108
369: \end{thebibliography}
370:
371: \begin{deluxetable}{lccccccc}
372: \tablecaption{Data Analysis Results}
373: \tablehead{
374: \colhead{Source} & \colhead{z} &
375: \colhead{Exp. time\tablenotemark{a}} & \colhead{Counts} &
376: \colhead{$\alpha_{\rm OX}$\tablenotemark{b} } &
377: \colhead{$\Gamma$\tablenotemark{c}} &
378: \colhead{N$_H$\tablenotemark{d}} &
379: \colhead{L(2-10 keV)\tablenotemark{e}} }
380: \startdata
381: HS 0017+2116&2.02 &10130& 62 & 1.77 & 1.79$^{+0.52}_{-0.36}$ & $<0.81$ & 43 \\
382: HS 0810+5157&0.38 &6940 & 988 & 1.40 & 1.56$^{+0.15}_{-0.07}$ & $<0.14$ & 2.7 \\
383: HS 0830+1833&2.27 &6480 & 63 & 1.83 & 1.91$^{+0.53}_{-0.33}$ & $<0.65$ & 4.9 \\
384: HS 0848+1119&2.62 &6120 & 47 & 1.76 & 1.39$^{+0.53}_{-0.38}$ & $<3.20$ & 10.7\\
385: HS 0854+0915&1.05 &3760 & 37 & 2.07 & 0.67$^{+0.60}_{-0.53}$ & --- & 1.5 \\
386: HS 1036+4008&1.96 &6060 & 51 & 2.12 & 1.04$^{+0.66}_{-0.56}$ & --- & 2.5 \\
387: HS 1111+4033&2.18 &9760 & 168 & 1.78 & 2.07$^{+0.28}_{-0.25}$ & $<0.43$ & 11.3\\
388: HS 1202+3538&2.28 &6760 & 52 & 1.79 & 2.28$^{+1.18}_{-0.50}$ & $<2.32$ & 3.2 \\
389: HS 1229+4807&1.37 &6750 & 100 & 1.58 & 2.17$^{+0.38}_{-0.31}$ & $<0.15$ & 2.9 \\
390: HS 1237+4756&1.55 &4750 & 393 & 1.38 & 1.55$^{+0.22}_{-0.09}$ & $<0.34$ & 12.1\\
391: HS 1415+2701&2.50 &8720 &$<$15& $>$2.3 & --- & --- & $<$3\\
392: HS 1417+4722&2.21 &7760 & 132 & 1.58 & 2.19$^{+0.46}_{-0.28}$ & $<1.55$ & 12.0\\
393: HS 1422+4224&2.21 &5960 & 157 & 1.71 & 2.45$^{+0.33}_{-0.44}$ & $<1.10$ & 9.4 \\
394: HS 1824+6507&0.30 &6950 & 733 & 1.47 & 1.80$^{+0.14}_{-0.12}$ & $<0.11$ & 1.5 \\
395: HS 1939+7000&0.12 &4970 &1103 & 1.54 & 0.95$^{+0.16}_{-0.07}$ & $<0.02$ & 0.74\\
396: HS 2135+1326&2.29 &5990 & 53 & 1.88 & 1.98$^{+0.83}_{-0.63}$ & $<2.25$ & 5.8 \\
397: HS 2146+0428&1.32 &7610 & 170 & 1.73 & 1.88$^{+0.39}_{-0.30}$ & $<0.55$ & 4.1 \\
398: HS 2251+2941&1.57 &7110 & 35 & 1.97 & 1.50$^{+0.56}_{-0.52}$ & $<2.57$ & 1.5 \\
399: \enddata
400: \tablenotetext{a}{Exposure time in seconds.}
401: \tablenotetext{b}{$\alpha_{\rm OX}$ obtained using the best fit model.}
402: \tablenotetext{c}{Photon index.}
403: \tablenotetext{d}{Column density in units of 10$^{21}$~cm$^{-2}$.}
404: \tablenotetext{e}{2-10 keV luminosity in units of 10$^{44}$ erg s$^{-1}$.}
405: \end{deluxetable}
406:
407: \begin{figure}
408: \plotone{f1.eps}
409: \figcaption{Optical to X-ray ratio as inferred from ROSAT observations
410: for the color-selected PG quasars (shaded hystogram) and the sample of R01.
411: $I_{\rm OX}$ is a logarithmic measure of the ratio between optical (B band)
412: and soft X-ray flux.}
413: \end{figure}
414:
415: \begin{figure}
416: \begin{center}
417: \plotone{f2.eps}
418: \figcaption{Chandra spectra of the sample of X-ray weak quasars.
419: The shaded region represent the X-ray spectrum expected
420: assuming $\alpha_{\rm OX}=1.55$ and $\Gamma=1.8$. ROSAT
421: upper limits are shown for the three sources having fluxes significantly
422: higher than in ROSAT observations.}
423: \end{center}
424: \end{figure}
425:
426: \begin{figure}
427: \plotone{f3.eps}
428: \figcaption{$\alpha_{\rm OX}$ versus luminosity compared with (a) the average values
429: found by Yuan et al. (1998) for a sample of $\sim~1000$ optically selected
430: quasars (shaded lines), (b) the average $\alpha_{\rm OX}$ of PG quasars (dashed line) and (c)
431: the average $\alpha_{\rm OX}$ of X-ray selected quasars, according to Elvis et al. 1994 (bottom
432: continous line).
433: The three low-luminosity objects are those with {\em Chandra} fluxes
434: significantly higher than ROSAT upper limits.
435: }
436: \end{figure}
437:
438: \end{document}
439: