1: \documentclass[12pt,preprint]{aastex}
2:
3:
4: \def\chandra {{\it Chandra}}
5: \def\kms{\rm\,km\,s^{-1}}
6: \def\ergcms{\rm\,erg\,cm^{-2}\,s^{-1}}
7: \def\l5100{L_{\rm 5100\AA}}
8: \def\l1kev{L_{\rm 1 keV}}
9: \def\mbh{M_{\rm BH}}
10: \begin{document}
11:
12: \title{\chandra\ Observations of X-Ray Weak Narrow-Line Seyfert 1 Galaxies}
13: \author{Rik J. Williams, Smita Mathur, and Richard W. Pogge}
14: \affil{Department of Astronomy, Ohio State University, 140 West 18th Avenue, Columbus, OH 43210-1173, USA}
15: \email{Send comments to: williams@astronomy.ohio-state.edu}
16: \shorttitle{X-Ray Weak Narrow-Line Seyfert 1 Galaxies}
17: \shortauthors{Williams, Mathur, \& Pogge}
18:
19: \begin{abstract}
20:
21: We present \chandra\ observations of 17 optically-selected,
22: X-ray weak narrow-line Seyfert 1 (NLS1) galaxies. These objects were
23: optically identified by Williams et al.~(2002) in the Sloan Digital Sky
24: Survey Early Data Release,
25: but were not found in the ROSAT All-Sky Survey (RASS) despite having
26: optical properties similar to RASS--detected NLS1s. All objects in this
27: sample were detected by \chandra\ and exhibit a range of 0.5--2~keV
28: photon indices
29: $\Gamma=1.1-3.4$. One object was not detected in the soft band, but has
30: a best--fit $\Gamma=0.25$ over the full 0.5--8~keV range. These
31: photon indices extend to values far below what are normally observed in NLS1s.
32: A composite X-ray spectrum of the hardest objects in this sample
33: does not show any signs of absorption, although the data do not prohibit
34: one or two of the
35: objects from being highly absorbed. We also find a strong
36: correlation between
37: $\Gamma$ and $\l1kev$; this may be due to differences in
38: $L_{\rm bol}/L_{\rm Edd}$ among the NLS1s in this sample. Such variations
39: are seemingly in conflict with the current paradigm that NLS1s accrete
40: near the Eddington limit. Most importantly,
41: this sample shows that strong, ultrasoft X-ray emission is not a universal
42: characteristic of NLS1s; in fact, a substantial number may exhibit weak
43: and/or low--$\Gamma$ X-ray emission.
44:
45: \end{abstract}
46:
47: \keywords{galaxies: Seyfert---galaxies: active---quasars: general---X-rays: galaxies}
48:
49: \section{Introduction}
50:
51: \citet{ostpog85} initially defined narrow-line Seyfert 1 galaxies
52: (NLS1s) by their striking optical spectral characteristics: strong,
53: narrow H$\beta$ emission \citep[later formally defined to be FWHM~$\leq
54: 2000 \kms$ by][]{goodrich89}, weak [\ion{O}{3}] relative to H$\beta$,
55: and strong \ion{Fe}{2}. These properties put NLS1s at one extreme end
56: of the so--called \citet{bg92} ``Eigenvector 1,'' thought to correspond
57: to emission from lower-mass nuclear black holes coupled with
58: near--Eddington accretion rates \citep{boroson02}. X-ray observations
59: have revealed a strong soft X-ray ``excess'' in NLS1s
60: \citep[e.g.,][]{leighly99}, further bolstering the high $L_{\rm bol}/L_{\rm
61: Edd}$---low--mass black hole hypothesis \citep{pounds95,wang96}.
62: Indeed, \citet[][hereafter BBF96]{bbf96} found a possible anticorrelation
63: between X-ray spectral slope and H$\beta$ line width, with NLS1s generally
64: having softer X-ray spectra than other AGN. Ultrasoft X-ray
65: selection has consequently proven to be an essential tool for the
66: discovery of large numbers of NLS1s \citep[e.g.][]{grupe00,grupe04}.
67:
68: The disadvantage of selecting NLS1s solely upon their X-ray
69: properties is that it can introduce into NLS1 samples a strong bias toward
70: those exhibiting an ultrasoft excess \citep{forster99}.
71: Since NLS1s are primarily
72: defined by their optical properties, the true nature of their
73: X-ray emission is thus difficult to determine. Though it is well known
74: that \emph{some} NLS1s are ultrasoft X-ray sources, previous samples of
75: optically selected NLS1s were simply too sparse to determine how many,
76: as well as whether or not a significant number of NLS1s have hard X-ray
77: spectra. With the advent of the Sloan Digital Sky Survey
78: \citep[SDSS;][]{york2000}, it is possible to build large catalogues of
79: NLS1s with homogeneous selection criteria based on their optical spectra
80: alone \citep[see][hereafter WPM02]{wpm02}.
81:
82: Of the 150 NLS1s listed in WPM02, 52 were detected in the ROSAT All-Sky
83: Survey \citep[RASS;][]{rass}. Forty--five of these had sufficient
84: counts in the 0.1--2.0~keV range to derive power-law photon indices
85: ($\Gamma$, where $N(E)\propto E^{-\Gamma}$) based on hardness ratios.
86: Most of these objects were optically bright ($g \la 18.5$),
87: low--redshift ($z \la 0.4$), and exhibited typical NLS1 photon indices
88: of $\Gamma \ga 2.0$. However, a substantial number of
89: optically--bright, low--redshift NLS1s did \emph{not} have RASS source
90: identifications. The optical spectra of these objects appear completely
91: normal (within the limitations of the SDSS resolution and
92: signal--to--noise) and only one or two are in regions of high Galactic
93: \ion{H}{1} column density, which could potentially obscure the X-ray
94: flux.
95:
96: It is thus possible that these objects represent a subset of NLS1s which are
97: optically normal but X-ray weak. Only a few such objects have
98: previously been found; for example,
99: RX J2217.9-5941 \citep{grupe01b} and PHL 1811 \citep{leighly01}.
100: Since these SDSS NLS1s were not detected in
101: the RASS, however, the nature and extent of their X-ray weakness could not be
102: determined. They could emit significant flux at
103: higher energies not covered by the 0.1--2.4~keV ROSAT band, or they
104: might also be ultrasoft X-ray sources but with substantially lower
105: overall X-ray fluxes than typically seen in NLS1s (i.e., much higher
106: $\alpha_{\rm{ox}}$). Another possible explanation is variability, but
107: it seems unlikely that such a large fraction ($\sim 40\%$) would be in
108: an exceptionally low state during the RASS observations. In reality,
109: all three of these factors probably have some bearing on the X-ray
110: weakness of these objects, but we cannot infer how many are affected by
111: which factors, if any, from the existing data.
112:
113: As a first step toward solving this puzzle, we have observed 17 of these
114: optically--selected but RASS--undetected NLS1s with \chandra. Due to its
115: excellent sensitivity and low background levels, \chandra\ is able to
116: detect objects at far lower flux levels than ROSAT, and its large
117: (0.5--8~keV) energy range allows detection of objects with harder X-ray
118: spectra as well. Our primary goals are (1) to detect these objects in
119: X-rays, or set upper limits to their X-ray emission, and (2) to obtain
120: rough estimates of $\Gamma$ for the \chandra--detected NLS1s. Given
121: this information, we can gain some insight as to which of the
122: aforementioned scenarios (if any) sufficiently explain these
123: ROSAT--unobserved NLS1s. In this paper we present the results of these
124: \chandra\ observations, and the possible implications for the NLS1s in
125: our sample.
126:
127:
128: \section{Target Selection}
129:
130: A full description of the spectroscopic selection and preliminary X-ray
131: analysis can be found in WPM02; a brief summary follows. The NLS1s in
132: the WPM02 sample were initially selected from the SDSS Early Data
133: Release \citep{stoughton02} solely on the basis of narrow H$\beta$
134: emission. Subsequently, each spectrum was visually inspected and a more
135: accurate measurement was taken of the H$\beta$ line width. Objects
136: which did not fulfill the criteria of \citet{ostpog85} and
137: \citet{goodrich89} were discarded. X-ray power law slopes and
138: luminosities were estimated for the 45 objects with adequate RASS data
139: available.
140:
141: Figure~\ref{fig_gz} shows the SDSS $g$~magnitude--redshift distribution
142: for the WPM02 sample, differentiating between NLS1s detected in RASS,
143: those not detected, and those chosen for \chandra\ follow-up. In this
144: diagram, the fainter, higher--redshift NLS1s are detected in the RASS
145: less frequently, and most NLS1s with $g \la 18.5$ and $z \la 0.4$ were
146: detected by RASS. We adopted these as rough limits for our follow--up
147: sample. To rule out the possibility that high foreground extinction
148: prevented some objects from being detected in the RASS, we restricted
149: our sample to objects in regions of low Galactic \ion{H}{1} column
150: density\footnote{Found for each location using the {\tt nh} utility,
151: part of the HEAsoft package, available at
152: \url{http://heasarc.gsfc.nasa.gov/lheasoft/}.} ($N_{\rm H} < 4\times
153: 10^{20}$~cm$^{-2}$, although one relatively bright NLS1 in the sample
154: has $N_{\rm H} = 5.7\times 10^{20}$~cm$^{-2}$). The resulting 17
155: objects which comprise this follow--up sample appear normal in all other
156: respects.
157:
158: %To double--check our selection
159: %methods, we derived an optical--X-ray relation for the RASS--detected
160: %NLS1s and interpolated expected X-ray fluxes for the undetected objects
161: %based on their $g$ magnitudes. Assuming typical values of $\Gamma = 2.5$,
162: %most of these NLS1s should have been above the nominal RASS Bright Source
163: %Catalog (BSC) limiting flux
164: %of $\sim 3\times 10^{-13}\ergcms$ in the 0.5--2 keV band
165: %\citep[0.05~counts~sec$^{-1}$;][]{rass}.
166: %Since the BSC contains only the brightest 13\% of the total RASS objects
167: %cross--referenced in the SDSS catalog,
168: %It should be noted that there is a fair amount
169: %of scatter in the optical--X-ray flux relationship. Even so, the scatter
170: %is unable to account for so many of these NLS1s falling below the X-ray flux
171: %limit; i.e., these objects exhibit X-ray fluxes well below
172: %average for RASS--detected NLS1s of similar optical flux levels.
173:
174:
175: \section{Observations and Data Reduction}
176:
177: We observed the 17 NLS1s in this sample with \chandra\ for $\sim 2$~ksec
178: each between 3 January and 15 September 2003. The Advanced CCD Imaging
179: Spectrometer Spectroscopic array \citep[ACIS-S;][]{garmire03} was
180: employed, and all observations were offset by 3\arcmin\ from the nominal
181: aimpoint along each detector axis to mitigate the effects of pileup
182: should any of the sources be unexpectedly bright. This resulted in a
183: slight ($\sim 10\%$) reduction in detector efficiency as well as a
184: broadening of the point--spread function. Neither of these effects
185: significantly hinders our source detection efficiency, thanks to
186: \chandra's exceptionally low background levels. These observing
187: parameters allow $3\sigma$ detections of point sources down to
188: $2-5\times 10^{-14}\ergcms$ (for $\Gamma = 3.0-1.0$ respectively) in the
189: 0.5--8 keV band.
190:
191: \subsection{Data preparation}
192: All data were reprocessed with the newest version of the \chandra\
193: Interactive Analysis of Observations (CIAO 3.0.1) software employing the
194: most recent calibration files (CALDB version 2.23).\footnote{Both
195: available at \url{http://cxc.harvard.edu/ciao/download\_ciao\_reg.html}}
196: We then defined circular extraction regions that were centered on each
197: source and of sufficient radius to encompass the PSFs (typically about
198: 4\farcs 5). Background regions were defined as annuli with inner and
199: outer radii of 7\farcs 5 and 15\arcsec\ respectively. The CIAO tool
200: {\tt psextract} was used to extract source and background spectra for
201: those objects with sufficient flux ($\ga 25$ total counts).
202: Additionally, we used {\tt dmextract} to determine the raw number of
203: counts for each source in the 0.5--0.9~(S), 0.9--2.0~(M), and
204: 2.0--8.0~keV~(H) bands. Finally, the {\tt apply\_acisabs} script was
205: applied to each ancillary response file (ARF) in order to account for
206: the time--dependent ACIS quantum efficiency degradation.\footnote{See
207: \url{http://cxc.harvard.edu/cal/Acis/Cal\_prods/qeDeg/} for details.}
208:
209: All 17 NLS1s in this sample were detected by \chandra, 16 with high
210: ($\geq 4\sigma$) significance. Table~\ref{tab_obslog} lists the net,
211: background subtracted, number of counts per energy bin for each object.
212: As expected, the background was found to be low, with 0--2 counts per
213: band detected in the background extraction regions (i.e. less than 0.25
214: counts expected in the source region on average), for most of the
215: observations. In only three cases were one or more background counts
216: expected in the source region. Due to the low background levels, we
217: assume the errors are simply Poisson errors on the raw number of
218: source counts.
219:
220: \subsection{Determination of Power-Law Slope}
221: We used the CIAO spectral fitting tool {\tt Sherpa} to analyze the resulting
222: data. Background levels are negligible ($\leq 1$ count expected in the
223: source extraction region) in all but two observations; for these two, we
224: fit the background simultaneously with a power law model.
225: For the twelve objects with sufficient counts to obtain binned
226: spectra (with at least 5 counts per bin), we fit a simple power law
227: with foreground Galactic absorption taken into account\footnote{
228: For fitting we used the $\chi^2$ statistic with the \citet{gehrels86}
229: variance function, which is the {\tt Sherpa} default and applicable
230: to data with few counts.}. This model can be represented by the equation
231: \begin{equation}
232: dN(E) \propto E^{-\Gamma} e^{-N_H \sigma(E)} A(E) dE
233: \end{equation}
234: where $\sigma(E)$ is the photoelectric absorption cross section of
235: Galactic gas with effective column density $N_H$, and $A(E)$ is the
236: observation--specific effective detector area encoded within each
237: ancillary response function (ARF) file used within {\tt Sherpa}.
238: Knowing $N_H$, $\Gamma$ and the overall normalization can be found
239: through a simple two--parameter fit to the data.
240:
241: For those objects too faint to be analyzed as binned spectra in {\tt
242: Sherpa} (and to check the consistency of our spectral fits) we employed
243: a hardness ratio fitting method. This method reduces the problem to a
244: one--parameter fit by disregarding the overall normalization of the
245: spectrum (which can be determined afterwards from the total count rate,
246: $(H+S)/t_{\rm exp}$).
247: The hardness ratio is defined as ${\rm HR} = (H-S)/(H+S)$,
248: where $H$ and $S$ are now the net counts in some arbitrary hard and soft
249: bands respectively. Assuming Poisson errors, $\sigma_{\rm
250: HR}=2\sqrt{S^2 H+H^2 S}/(H+S)^2$ (again using the {\it raw} number of
251: counts per band to calculate errors, for those observations with
252: non--zero background levels). We employed two different hardness ratios
253: in this analysis, which we call HR$_a$ and HR$_b$. HR$_a$ is analogous
254: to the ROSAT ``hardness ratio 2'' parameter, where the soft and hard
255: bands are 0.5--0.9 and 0.9--2~keV respectively. HR$_b$ takes advantage
256: of the full \chandra\ energy range with soft and hard bands covering
257: 0.5--2 and 2--8~keV, respectively. Again using the ARF and $N_H$ value
258: specific to each observation, we used {\tt Sherpa} to calculate HR$_a$
259: and HR$_b$ for a test value of $\Gamma$. By iterating $\Gamma$ until
260: the correct values of the hardness ratios were reached, we derived
261: photon indices independently for each of HR$_a$ and HR$_b$.
262:
263:
264: \subsection{Consistency}
265: For 15 of these NLS1s, $\Gamma$ was derived by using two or three of the
266: methods discussed in the previous section. As a consistency check, we
267: compared the results for individual objects. For the most part,
268: especially for the brighter objects, the three measurements produced
269: results well within $1\sigma$ of each other. However, there are
270: systematic offsets. Values of $\Gamma$ derived from HR$_a$ are nearly
271: always higher than those found using HR$_b$, while {\tt Sherpa} fits
272: typically lie somewhere in between. This is as expected since the soft
273: X-ray power-law slopes of NLS1s are known to be steeper than those in
274: the hard band due to the commonly--observed soft excess
275: (e.g. ROSAT: BBF96 vs. ASCA: Brandt, Mathur \& Elvis 1997, although some
276: of this offset may be due to calibration issues as described in Iwasawa,
277: Fabian, \& Nandra 1999).
278: A slope over the entire energy range, as measured
279: by {\tt Sherpa}, would be in between the two. Since they contain the
280: average spectral information, we list the $\Gamma$ from {\tt Sherpa}
281: fits in Table~\ref{tab_objects}. If there are not enough counts for
282: such a fit, we list the 0.5--8~keV (HR$_b$) measurement since it covers
283: the full energy range; for the lone observation with no 2--8~keV counts,
284: the 0.5--2~keV measurement of $\Gamma$ is listed.
285:
286: % \section{Analysis}
287:
288: \section{X-ray Spectral Properties}
289:
290: \subsection{Detectability in the RASS\label{sec_rass}}
291: The NLS1s in this sample span a broad range of $\Gamma$ (best fit values
292: of 0.25--3.15) and X-ray fluxes from near the detection limit of about
293: 0.002 counts~s$^{-1}$ to almost 0.2 counts~s$^{-1}$. Taking the
294: 0.5--2~keV flux derived from {\tt Sherpa} for the twelve brightest
295: objects, we used the PIMMS software\footnote{Portable, Interactive
296: Multi-Mission Simulator Version 3.4, from NASA's High Energy
297: Astrophysics Science Archive Research Center, currently available at
298: \url{http://heasarc.gsfc.nasa.gov/docs/software/tools/pimms.html}.} to
299: calculate the count rate expected in ROSAT.
300: Exposure times for each position were taken from the RASS exposure
301: maps\footnote{Available at
302: \url{http://www.xray.mpe.mpg.de/cgi-bin/rosat/rosat-survey}.} and the
303: expected number of RASS counts was then computed for each object. Three
304: of the sources, J1013$+$0102, J1214$+$0055, and J1449$+$0022 should have
305: been easily detected with about 17, 17, and 22 counts, respectively.
306: The second object may actually correspond to a nearby RASS source which
307: was marginally detected and not cross-referenced in the SDSS database.
308: Four other objects in this sample should be marginally detected by the
309: RASS with $\sim 10$ counts; their nondetection may be due to Poisson
310: error or uncertainty in the flux determination. The remaining ten
311: objects were well below the RASS detection limit.
312:
313: The X-ray luminosity of the three bright objects may have varied by a
314: factor of two or more between their RASS and \chandra\ observations,
315: which could account for their lack of detection in the RASS. Such
316: long--timescale variability is not surprising, as it is commonly seen in NLS1s
317: \citep[e.g.,][]{grupe01}. It is impossible to ascertain the
318: degree of variability in the fainter objects; however, most of this
319: sample could not have been detected in the RASS at the flux levels
320: observed by \chandra. It is unlikely that all 14 of these 17 NLS1s were
321: in an exceptionally low luminosity state during both the RASS and
322: \chandra\ observations; thus, at least some of these objects must be
323: intrinsically faint at ROSAT energies.
324:
325: \subsection{Photon Index--Luminosity Relation}
326:
327: It is likely that some combination of variability, X-ray hardness, and
328: low X-ray luminosity resulted in these objects not being detected in the RASS.
329: The latter two factors can fortunately be determined from these
330: observations. The data in Table~\ref{tab_objects} show that many of
331: the NLS1s with low $\Gamma$ are also the faintest X-ray sources.
332: To determine whether this is a true effect, we calculated the monochromatic
333: 1~keV (rest-frame) fluxes using the models derived through spectral
334: fitting in {\tt Sherpa}. For the fainter objects we calculated the
335: expected flux from the best--fit hardness ratio models using {\tt Sherpa},
336: with a fractional
337: flux error equal to $N^{-1/2}$ where $N$ is the total number of counts
338: in the 0.5--8~keV band.
339: Using the SDSS redshifts and an $\Omega_m=0.3$, $\Omega_\Lambda=0.7$,
340: $H_0=70\,\rm{km}\,\rm{s}^{-1} \rm{Mpc}^{-1}$ cosmology, the corresponding
341: rest--frame luminosities were then calculated. These are given in
342: Table~\ref{tab_objects}, along with the optical $\lambda 5100$\AA\
343: monochromatic flux measured from the original SDSS spectra.
344:
345: The relation between $\Gamma$ and $\l1kev$ is shown in Figure~\ref{fig_gamlum}.
346: Luminosities derived from {\tt Sherpa} fitting are plotted as filled
347: circles, and those estimated from the hardness ratio models are plotted
348: as open circles. This figure shows a strong correlation between X-ray
349: spectral slope and 1~keV luminosity. \citet{grupe99} found such a
350: correlation among soft X-ray selected AGN, with a stronger correlation
351: among NLS1s noted in \citet{grupe04-2}. No such correlation is seen
352: in their broad--line AGN sample; additionally, \citet{laor97} do not find
353: a significant correlation in their sample of bright, optically--selected
354: quasars (although this may be due to the limited range of $\Gamma$ in these
355: samples).
356:
357: This relation has a slope of $b=0.66\pm 0.13$ (where $\Gamma
358: \propto b \log \l1kev$); if the two hardest objects are excluded
359: from the fit, then $b$ decreases to $0.55\pm 0.13$. Similar relations
360: have been measured for NLS1s by \citet{forster96} with $b=0.32$
361: (where the luminosity
362: is measured over the 0.1--2~keV range), while the \citet{dai04} quasar sample
363: has $b=0.55\pm 0.11$ (with $L=L_{0.2-2\,\rm{keV}}$). When the luminosities
364: of our sample are extrapolated to the 0.1--2\,keV range using the best--fit
365: models and the two lowest--$\Gamma$ points are excluded, the slope
366: becomes $b=0.48\pm 0.08$, in agreement with \citet{dai04} but about
367: 2$\sigma$ higher than the \citet{forster96} result (note that the choice of
368: 0.1\,keV or 0.2\,keV as the lower bound of the energy range
369: does not significantly affect the extrapolated luminosity).
370: This latter discrepancy may be due to their sample being selected from the
371: RASS, which is not as sensitive to flat--spectrum objects as \chandra.
372:
373: For contrast, the $\Gamma-\l1kev$ relation caused by pure
374: absorption (i.e. assuming a typical source with $\Gamma\sim 2.5$, determining
375: what hardness ratio and $\l1kev$ would be observed with
376: various intrinsic $N_H$ values, and recalculating $\Gamma$ using the simulated
377: HR$_b$ assuming \emph{no} prior knowledge of the intrinsic absorption)
378: produces a slope closer to $b=2$ for $N_H\la 5\times 10^{21}$\,cm$^{-2}$.
379: Most of the observed correlation thus
380: cannot be explained simply by intrinsic absorption, although it could
381: explain the hardest objects in this sample ($\Gamma < 1$). However, this
382: would require large neutral column densities ($\ga 10^{22} \rm{cm}^{-2}$),
383: despite there being no sign of optical dust reddening. Moreover, it is
384: unlikely that all four NLS1s with $\Gamma < 1$ are strongly absorbed
385: (see \S\ref{sec_abs}). A dust--poor absorber could, in
386: principle, account for the extreme hardness and low X-ray luminosity of
387: one or two of these objects.
388:
389: The observed $\Gamma-\l1kev$ correlation may be a result of
390: differing Eddington ratios. To test this, we estimated black hole masses
391: and Eddington luminosities using the \citet{kaspi00}
392: relation between $R_{\rm BLR}$, $\lambda L_\lambda$(5100\AA), and
393: FWHM(H$\beta$) and applying the virial theorem. It should be noted that
394: these estimates may be subject to systematic error since the \citet{kaspi00}
395: sample contains very few NLS1s; however, this method should be sufficient
396: to show general trends in the data. Values of FWHM(H$\beta$)
397: were taken directly from WPM02, and the 5100\AA\ flux was measured from
398: the SDSS spectra used in that study. Figure~\ref{fig_gamledd} demonstrates
399: a strong correlation between $\Gamma$ and $\l1kev/L_{\rm Edd}$, which
400: further indicates that the objects with the softest X-ray spectra
401: have the highest relative accretion rates if $\l1kev\propto L_{\rm bol}$.
402: This is probably a reasonable assumption. Figure~\ref{fig_optx} shows
403: that the 1~keV luminosity is correlated with the 5100\AA\ luminosity
404: for the bright and faint objects in this sample (with the notable exception
405: of one point, which has $\Gamma\sim 0.25$ and may be strongly absorbed).
406: We can thus assume that both the optical and X-ray luminosities
407: provide some indication of the bolometric luminosity. As expected, a
408: correlation is also seen between $\Gamma$ and $\lambda L_\lambda$(5100\AA),
409: though it is somewhat weaker than that between $\Gamma$ and $\l1kev$.
410: This weakness is not surprising, since X-ray properties should physically be
411: more closely linked with each other than with optical measurements.
412:
413: Although $\Gamma$ is directly proportional to $L_{\rm bol}/L_{\rm Edd}$,
414: it is thought to be inversely proportional to $\mbh$ \citep[e.g.,][]{kuras00},
415: which could wash out the observed $\Gamma-\l1kev$
416: correlation for some samples. For this \chandra\ sample, the $\mbh$
417: estimates span approximately two orders of magnitude, while
418: $\l1kev/L_{\rm Edd}$ spans three, so the correlation is
419: observed. In samples with a larger range of black hole masses and luminosities
420: (such as the Grupe 2004 BLS1s), this correlation would not be expected
421: and, indeed, is not seen.
422: The $\Gamma-\l1kev$ relation would also not be seen in samples with
423: a smaller range of $L_{\rm bol}/L_{\rm Edd}$, unless the black hole mass
424: range is correspondingly larger, as proposed for the \citet{dai04} sample.
425: When the same analysis is performed
426: for the WPM02 RASS objects, no correlation is observed. Instead, they cluster
427: almost exclusively around the highest--$\l1kev/L_{\rm Edd}$,
428: highest--$\Gamma$ members of the \chandra\ sample. This is not unexpected
429: since the WPM02 RASS sample consists mostly of NLS1s near the RASS
430: detection limit; they do not exhibit the range in luminosities
431: (and hence $\l1kev/L_{\rm Edd}$) or $\mbh$ required to see the
432: $\Gamma-\l1kev$ correlation. If the observed $\Gamma-\l1kev$ correlation is
433: indeed due to differences in the Eddington ratio, then it is
434: quite likely that this \chandra\ sample includes some NLS1s which are
435: accreting far below the Eddington limit.
436:
437: \subsection{Spectral Features in J1449+0022\label{sec_771}}
438: As noted in \S\ref{sec_rass}, J1449+0022 is the brightest object (in
439: X-rays) in this sample and may have gone undetected in the RASS due to
440: strong variability. Indeed, there appears to be a slight decrease in
441: the \chandra\ count rate over the course of the 2~ksec observation.
442: More interesting, however, is that the spectrum of this object is not
443: well fit by a simple power law with Galactic foreground absorption, as
444: seen in Figure~\ref{fig_771}; specifically, there appears to be an
445: excess of soft photons over that expected from a power law. The X-ray
446: continuum can be better fit using a slightly more complicated model;
447: either a power law plus a blackbody (which yields $\Gamma=1.58\pm 0.22$
448: and $kT=0.11\pm 0.02$), or a broken power law with
449: $\Gamma_1=3.40^{+0.73}_{-0.44}$, $\Gamma_2=1.64\pm 0.19$, and
450: $E_b=1.04^{+0.12}_{-0.19}$\,keV. There is also a significant dip in
451: flux at $1.10\pm 0.03$\,keV with equivalent width $0.12^{+0.06}_{-0.04}$\,keV.
452: This absorption is similar to that seen in some other NLS1s
453: \citep{leighly97}. Such a feature could be due OVII/OVIII
454: absorption in a highly relativistic outflow \citep[e.g.,][]{pounds03},
455: but a more plausible model
456: may be the \citet{nicastro99} hypothesis of a strong Fe~L complex at
457: this energy.
458:
459: \subsection{Intrinsic Absorption \label{sec_abs}}
460: Four of the NLS1s in this sample exhibit spectra with $\Gamma < 2$.
461: Since for low count rates $\Gamma$ and $N_H$ are highly degenerate, no
462: individual object is bright enough to place a meaningful limit on its
463: intrinsic absorption; thus, we cannot determine with confidence whether
464: this X-ray hardness is intrinsic to the accretion process or merely
465: caused by a high degree of absorption. The spectra of these four NLS1s
466: were added together to determine whether they exhibit strong absorption
467: as a group. We assumed all have typical $\Gamma =2.5$ and similar
468: $N_H$ values at each redshift, and fit a corresponding model to the coadded
469: spectrum. This resulted in a best-fit value of
470: $N_H=1.5^{+1.6}_{-0.9}\times 10^{21} \rm{cm}^{-2}$, but the fit is not
471: particularly good and shows strong residuals. If we leave both $\Gamma$
472: and $N_H$ as free parameters, $N_H$ becomes zero (with a $2\sigma$ upper
473: limit of $2\times 10^{21} \rm{cm}^{-2}$) and $\Gamma = 1.2\pm 0.3$. The
474: fit with no absorption is also much better ($\Delta \chi^2 = -6.6$
475: compared to the absorption model). Based on this, it appears as though
476: these flat--spectrum sources are not strongly absorbed as a group.
477:
478: Of course, it is possible that one or two of these NLS1s could be
479: strongly absorbed, particularly the faintest object (J1259$+$0102) since
480: with only five total counts its influence on the coadded spectrum is
481: small. If this object has an intrinsic X-ray luminosity
482: $\l1kev\sim 10^{44}\,\rm{erg}\,\rm{s}^{-1}$ and $\Gamma\sim 3$
483: (similar to the brightest objects in our sample), an
484: intrinsic neutral column density of $\sim 2.5\times 10^{22}
485: \rm{cm}^{-2}$ would be required to reproduce the observed flux and
486: $\Gamma$. The input $\Gamma$ and luminosity of such a model can be
487: adjusted to reproduce the other faint, hard sources as well, but
488: somewhat high column densities ($N_H \ga 5\times 10^{21} \rm{cm}^{-2}$)
489: are typically required. Moreover, there is no indication of absorption
490: in the optical spectra of these objects. Thus, we conclude that these
491: objects most likely have intrinsically hard X-ray spectra, although the
492: two hardest may be heavily obscured in the X-rays but not at optical
493: wavelengths \citep[cf.][]{risaliti01,nandra04}.
494:
495: \section{Comparison to Other NLS1 Samples}
496: \subsection{X-ray Spectral Slope}
497: At first glance, this sample appears to exhibit a range of X-ray
498: spectral slopes extending to much lower values than previously seen for
499: NLS1s. However, the $\Gamma$ values listed are, for the most part,
500: derived over a larger energy range than used by previous studies. For
501: example, $\Gamma$ is measured by ROSAT over the 0.1--2.4~keV energy
502: range for the 52 narrow- and broad-line Seyfert~1 objects in BBF96,
503: while the 0.1--2.0~keV hardness ratio is used with the WPM02 sample. In
504: order to ensure the energy bands overlap, we employ the HR$_a$
505: measurement of $\Gamma$ for comparison to these two samples.
506:
507: It should be noted that the energy ranges used by these samples are
508: slightly different nonetheless. Since HR$_a$ covers the 0.5--2.0~keV
509: band, it does not extend to the 0.1~keV minimum energy of ROSAT, or to
510: the 2.4~keV maximum energy used by BBF96. However, 28 of the WPM02
511: NLS1s detected in ROSAT have measurements for both HR1 (0.1--2.0~keV)
512: and HR2 (0.5--2.0~keV). We rederived $\Gamma$ using the HR2
513: measurements and compared it to the values from HR1. Although the
514: errors are larger, the HR2 $\Gamma$s are consistent with those from HR1
515: in all but four objects. No systematic offset is seen between these two
516: measurements. Thus, we assume that the $\Gamma$s derived from ROSAT HR1
517: are good estimators of the spectral slopes over the 0.5--2~keV range.
518: As for the BBF96 measurements extending to 2.4~keV while ours extend
519: only to 2.0~keV, we expect this to be negligible for two reasons: first,
520: few photons are emitted in the high--energy tail ($N_{2.0-2.4{\rm keV}}
521: \sim 0.1N_{0.5-0.9\rm{keV}}$ for $\Gamma=2$); and second, the ROSAT
522: sensitivity decreases quite rapidly in this regime, so even fewer
523: photons should be detected. Thus, the differing energy bands should
524: have little effect on comparing these various samples.
525:
526: Figure~\ref{fig_fgamma} shows the relation between $\Gamma$ and the
527: H$\beta$ velocity width for all three samples. Crosses denote
528: BBF96 data, blue circles are the WPM02 RASS data, and the 16 \chandra\
529: objects with HR$_a$ measurements are shown as red circles with errorbars.
530: The horizontal lines show the mean and sample standard deviation of
531: $\Gamma$ for only the BBF96 data; NLS1 and Sy1 averages are computed
532: separately. It is immediately apparent that the WPM02 and \chandra\
533: NLS1s do not fall within the same range as those in BBF96; in fact,
534: a significant number appear to exhibit \emph{Seyfert 1}--type
535: X-ray spectra. Table~\ref{tab_gamma} lists the mean and sample standard
536: deviation of $\Gamma$ for the three samples. An interesting progression
537: emerges: as we move from NLS1s which were found through a mixture of
538: optical and X-ray selection (BBF96), to those which were selected solely
539: based on their optical properties (WPM02), to the \chandra\ sample
540: presented herein which was selected on the basis of weak X-ray emission,
541: $\langle\Gamma\rangle$ and the extremes of the $\Gamma$ distribution become
542: harder. In fact, the $\Gamma$ distribution of this sample extends to
543: lower values than that of the BBF96 Seyfert~1 galaxies. This illustrates
544: the effect of selection methods on building NLS1 samples, as well as
545: demonstrating the existence of a significant population of NLS1s which
546: are {\emph not} ultrasoft X-ray sources.
547:
548: \subsection{X-ray vs. Optical Luminosity}
549: This \chandra\ sample was originally selected as a set of X-ray weak
550: NLS1 candidates, i.e. NLS1s which were not detected in the RASS but
551: had $g^\prime$ magnitudes similar to detected objects. To determine
552: whether these objects are truly X-ray weak, we used the WPM02 RASS--detected
553: NLS1s as a comparison sample. Rest--frame luminosities at
554: 1\,keV were calculated using PIMMS along with ROSAT count rates and
555: the best--fit $\Gamma$ values given in WPM02, and 5100\,\AA\ luminosities
556: were measured from the original SDSS spectra. The best--fit relation
557: between the monochromatic optical and X-ray luminosities has a slope
558: consistent with unity, and is plotted on Figure~\ref{fig_optx} along
559: with the data points from the \chandra\ sample.
560:
561: Interestingly, the RASS relation corresponds very well with the
562: upper bound on the X-ray--optical relation for the \chandra\ NLS1s.
563: Eight of these objects are consistent with the RASS fit and another
564: three are consistent within $\sim 1-2\sigma$, while the
565: rest are significantly below the line. The faintest objects in X-rays
566: appear to exhibit a $L_X-L_{\rm opt}$ slope similar to that of the
567: brighter objects (if the point at the lower right of the plot, which
568: is probably absorbed in X-rays, is excluded), but their X-ray luminosities
569: are roughly a factor of 5 lower than those seen in the WPM02 RASS sample.
570: Thus, although most of the objects in this sample are not particularly
571: X-ray faint compared to NLS1s in the WPM02 RASS sample with similar optical
572: luminosities, six of them do fall well below the $L_X-L_{\rm opt}$
573: relation due to intrinsic X-ray faintness or obscuration.
574:
575:
576: \section{Discussion and Conclusions\label{conclusions}}
577:
578: Through short--duration \chandra\ observations of RASS--undetected NLS1s,
579: we have determined that six of the 17 objects have X-ray luminosities
580: substantially lower than NLS1s with similar optical properties. Of
581: the brighter objects, at least two exhibit flux levels which should
582: have been detectable by the RASS, indicating that their luminosities may
583: have increased by a factor of two or more between the RASS and \chandra\
584: observations. Many of the remaining bright objects were near or just
585: below the RASS detection limit, and were most likely not seen due to
586: smaller luminosity variations or Poisson noise.
587: Across the entire sample, a strong correlation is seen
588: between the X-ray spectral slope $\Gamma$ and $\l1kev$. This is probably
589: not entirely due to intrinsic absorption, since individual spectra of
590: bright objects
591: as well as a coadded spectrum of the faintest objects do not indicate
592: high degrees of absorption (although one or two of the faintest hard--spectrum
593: objects may be absorbed in X-rays but not at optical wavelengths).
594: If $\Gamma$ is indeed correlated with $L_{\rm bol}/L_{\rm Edd}$, then
595: the $\Gamma-\l1kev$ relation suggests that variations in $\l1kev$ are
596: primarly due to
597: differences in $L_{\rm bol}/L_{\rm Edd}$ among objects with comparatively
598: similar black hole masses. This interpretation is complementary to that
599: of \citet{dai04}, who find a similar relation but whose sample more
600: likely includes objects with a large
601: range of $\mbh$ but $L_{\rm bol}/L_{\rm Edd}\sim 1$.
602:
603: These observations may hold important implications for the ``Eigenvector
604: 1'' (Principal Component 1; PC1) paradigm
605: posited by \citet{bg92} and reinforced by \citet{boroson02}. In this
606: picture, PC1 (which is primarily driven by an anticorrelation between
607: [\ion{O}{3}] and \ion{Fe}{2}) is an indicator of $L_{\rm bol}/L_{\rm Edd}$.
608: NLS1s typically lie at one extreme end of PC1---the
609: end thought to correspond to the highest relative accretion rates. Since
610: $\Gamma$ is thought to be related to $L_{\rm bol}/L_{\rm Edd}$, $\Gamma$ and
611: PC1 should be correlated; indeed, \citet{brandt98} find such a correlation.
612: However, the sample presented herein contains objects which from their
613: optical spectra are at the
614: supposed high--$L_{\rm bol}/L_{\rm Edd}$ end of PC1, yet also exhibit very low
615: values of $\Gamma$, as well as low inferred $\l1kev/L_{\rm Edd}$.
616:
617: These extreme objects may indicate that while PC1 is usually
618: correlated with $L_{\rm bol}/L_{\rm Edd}$, it may also be
619: affected by orientation, black hole mass, or other physical drivers
620: \citep[as noted by][]{boroson04}. This is not
621: a completely new phenomenon; for example, BBF96 note that Mrk~507, with
622: $\Gamma=1.6\pm 0.3$, has an unusually flat X-ray spectrum for a NLS1.
623: This sample simply demonstrates
624: that Mrk~507 is not an isolated case, and in fact a small but interesting
625: subset of NLS1s do not appear to fit within the PC1 framework. The
626: apparent lack of strong absorption in some of these flat sources indicates
627: that their X-ray spectra actually are intrinsically flat.
628: Further studies of X-ray weak NLS1s, as well as much larger samples from
629: surveys such as the SDSS, should offer greater insight into the
630: mechanism(s) behind PC1.
631:
632: Due to the short exposure times of these observations ($\la 2$~ksec), we
633: cannot infer much outside
634: of $\Gamma$ and luminosity estimates for individual objects; indeed, this
635: program was intended to study the group properties of an X-ray weak
636: NLS1 sample. However, there are several objects in this sample with
637: exceptionally low $\Gamma$ which may be worthy of further study.
638: These hard X-ray NLS1s may represent a new, rare
639: subclass which are optically normal but highly absorbed in the X-rays,
640: or which exhibit abnormally low $L_{\rm bol}/L_{\rm Edd}$, or both.
641:
642:
643: \acknowledgements
644: The authors are grateful to Th. Boller for providing the original BBF96 data
645: for our Figure~\ref{fig_fgamma}. We also thank the anonymous referee
646: for helpful comments and suggestions.
647: Support for this work was provided by NASA through Chandra Award Number
648: GO3-4145X issued by the Chandra X-ray Observatory Center, which is
649: operated by SAO for NASA under contract NAS8-39073.
650:
651:
652: \begin{thebibliography}{}
653: \bibitem[Boller, Brandt, \& Fink(1996)]{bbf96} Boller, Th., Brandt, W. N., \& Fink, H. 1996, \aap, 305, 53
654: \bibitem[Boroson(2002)]{boroson02} Boroson, T. A. 2002, \apj, 565, 78
655: \bibitem[Boroson(2004)]{boroson04} Boroson, T. A. 2004, AGN Physics with the Sloan Digital Sky Survey, ed. G. T. Richards and P. B. Hall (San Francisco: ASP), in press
656: \bibitem[Boroson \& Green(1992)]{bg92} Boroson, T. A. \& Green, R. F. 1992, \apjs, 80, 109
657: \bibitem[Brandt \& Boller(1998)]{brandt98} Brandt, N. \& Boller, T. 1998, Astron. Nachr., 319, 163
658: \bibitem[Brandt, Mathur, \& Elvis(1997)]{brandt97} Brandt, W. N., Mathur, S., \& Elvis, M. 1997, \mnras, 285, L25
659: \bibitem[Dai et al.(2004)]{dai04} Dai, X., Chartas, G., Eracleous, M., \& Garmire, G. P. 2004, \apj, in press (astro-ph/0401013)
660: \bibitem[Forster(1999)]{forster99} Forster, K. 1999, Ph.D. Thesis, Columbia University (New York)
661: \bibitem[Forster \& Halpern(1996)]{forster96} Forster, K., \& Halpern, J. P. 1996, \apj, 468, 565
662: \bibitem[Garmire(2003)]{garmire03} Garmire, G. P. 2003, Proc. SPIE, 4851, 28
663: \bibitem[Gehrels(1986)]{gehrels86} Gehrels, N. 1986, \apj, 303, 336
664: \bibitem[Goodrich(1989)]{goodrich89} Goodrich, R. W. 1989, \apj, 342, 234
665: \bibitem[Grupe et al.(1999)]{grupe99} Grupe, D., Beuermann, K., Mannheim, K., \& Thomas, H.-C. 1999, \aap, 350, 805
666: \bibitem[Grupe(2000)]{grupe00} Grupe, D. 2000, NewAR, 44, 455
667: \bibitem[Grupe, Thomas, \& Beuermann(2001)]{grupe01} Grupe, D., Thomas, H.-C., \& Beuermann, K. 2001, \aap, 367, 470
668: \bibitem[Grupe, Thomas, \& Leighly(2001)]{grupe01b} Grupe, D., Thomas, H.-C., \& Leighly, K. M. 2001, \aap, 369, 450
669: \bibitem[Grupe et al.(2004)]{grupe04} Grupe, D., Wills, B. J., Leighly, K. M., Meusinger, H. 2004, \aj, in press (astro-ph/0310027)
670: \bibitem[Grupe(2004)]{grupe04-2} Grupe, D. 2004, AJ, in press (astro-ph/0401167)
671: \bibitem[Iwasawa, Fabian, \& Nandra(1999)]{iwasawa99} Iwasawa, K., Fabian, A. C., \& Nandra, K. 1999, \mnras, 307, 611
672: \bibitem[Kaspi et al.(2000)]{kaspi00} Kaspi, S., Smith, P.S., Netzer, H., Maoz, D., Jannuzi, B.T., \& Giveon, U., 2000, ApJ, 533, 631
673: \bibitem[Kuraszkiewicz et al.(2000)]{kuras00} Kuraszkiewicz, J., Wilkes, B. J., Czerny, B., \& Mathur, S. 2000, \apj, 542, 692
674: \bibitem[Laor et al.(1997)]{laor97} Laor, A., Fiore, F., Elvis, M., Wilkes, B. J., \& McDowell, J. C. 1997, \apj, 477, 93
675: \bibitem[Leighly(1999)]{leighly99} Leighly, K. M. 1999, \apjs, 125, 317
676: \bibitem[Leighly et al.(1997)]{leighly97} Leighly, K. M., Mushotzky, R. F., Nandra, K., \& Forster, K. 1997, \apj, 489, L25
677: \bibitem[Leighly et al.(2001)]{leighly01} Leighly, K. M., Halpern, J. P., Helfand, D. J., Becker, R. H., \& Impey, C. D. 2001, \aj, 121, 2889
678: \bibitem[Nandra et al.(2004)]{nandra04} Nandra, K., Georgantopoulos, I., Brotherton, M., \& Papadakis, I. E. 2004, \mnras, in press (astro-ph/0311029)
679: \bibitem[Nicastro, Fiore, \& Matt(1999)]{nicastro99} Nicastro, F., Fiore, F., \& Matt, G. 1999, \apj, 517, 108
680: \bibitem[Osterbrock \& Pogge(1985)]{ostpog85} Osterbrock, D. E. \& Pogge, R. W. 1985, \apj, 297, 166
681: \bibitem[Pounds, Done, \& Osborne(1995)]{pounds95} Pounds, K. A., Done, C., \& Osborne, J. P. 1995, \mnras, 277, L5
682: \bibitem[Pounds et al.(2003)]{pounds03} Pounds, K. A., King, A. R., Page, K. L., \& O'Brien, P. T. 2003, \mnras, 346, 1025
683: \bibitem[Risaliti et al.(2001)]{risaliti01} Risaliti, G., Marconi, A., Maiolino, R., Salvati, M., \& Severgnini, P. 2001, \aap, 371, 37
684: \bibitem[Stoughton et al.(2002)]{stoughton02} Stoughton, C., et al. 2002, \aj, 123, 485
685: \bibitem[Voges et al.(1999)]{rass} Voges, W., et al. 1999, \aap, 349, 389
686: \bibitem[Wang, Brinkmann \& Bergeron(1996)]{wang96} Wang, T., Brinkmann, W., \& Bergeron, J. 1996, \aap, 309, 81
687: \bibitem[Williams, Pogge \& Mathur(2002)]{wpm02} Williams, R. J., Pogge, R. W., \& Mathur, S. 2002, \aj, 124, 3042
688: \bibitem[York et al.(2000)]{york2000} York, D. G., Adelman, J., Anderson, J.E., et al. 2000, \aj, 120, 1579
689:
690:
691: \end{thebibliography}
692:
693: \clearpage
694:
695: \begin{figure}
696: \plotone{f1.eps}
697: \caption{SDSS g$^\prime$ magnitude--redshift relation for the WPM02
698: SDSS--selected NLS1s.
699: Crosses indicate objects undetected in the RASS, open circles are
700: RASS--detected objects,
701: and filled circles are NLS1s undetected in the RASS which were
702: chosen for \chandra\ follow--up.
703: The dashed lines show the nominal brightness and redshift limits for
704: the \chandra\ follow--up sample. \label{fig_gz}}
705: \end{figure}
706:
707: \begin{figure}
708: \plotone{f2.eps}
709: \caption{$\Gamma$ (from Table~\ref{tab_objects}) vs. 1~keV rest--frame
710: luminosity for the \chandra\ follow--up NLS1 sample. Filled points
711: are bright objects with spectra fit in Sherpa, while for open circles
712: the luminosity was estimated using HR$_b$ (HR$_a$ for J1311$+$0031) and
713: the 0.5--8~keV count rate. \label{fig_gamlum}}
714: \end{figure}
715:
716: \begin{figure}
717: \plotone{f3.eps}
718: \caption{Same as Figure~\ref{fig_gamlum}, but with photon index plotted
719: against $\l1kev/L_{\rm Edd}$. The x-axis errorbars only reflect
720: the uncertainty on $\l1kev$ since the uncertainty in
721: $L_{\rm Edd}$ is not well-known. \label{fig_gamledd}}
722: \end{figure}
723:
724: \begin{figure}
725: \plotone{f4.eps}
726: \caption{X-ray vs. optical luminosity for the objects with $\Gamma$ derived
727: in {\tt Sherpa} (solid points), or using hardness ratios (open circles).
728: The solid line denotes the best--fit $\l1kev-L_{\rm opt}$ relation
729: from the WPM02 RASS sample.
730: The high--optical, low--X-ray luminosity point is the object with
731: $\Gamma=0.25$, and may be highly absorbed. \label{fig_optx}}
732: \end{figure}
733:
734: \begin{figure}
735: \plotone{f5.eps}
736: \caption{\chandra\ count spectrum for J1449+0022, along with the best-fit
737: power-law model and residuals. This spectrum has been binned to 10
738: counts per bin. The observed flux is systematically high at $E<0.8\rm{keV}$,
739: and there is an absorption feature just above 1\,keV. \label{fig_771}}
740: \end{figure}
741:
742: \begin{figure}
743: \epsscale{0.8}
744: % ** Figure should be rotated 90 degrees clockwise for publication.
745: \plotone{f6.eps}
746: \caption{Soft X-ray photon index vs. H$\beta$ width, for NLS1s and Seyfert~1
747: galaxies (on the left and right of the dashed vertical line, respectively).
748: Crosses are data from BBF96, kindly provided by Th. Boller. Blue circles
749: are NLS1s from WPM02 (with $\Gamma$ derived from RASS measurements) and
750: red circles are the sample presented herein. All values of $\Gamma$ for the
751: \chandra\ follow--up sample were found using the 0.5--2~keV hardness ratio
752: HR$_a$. The
753: horizontal solid and dashed lines show the means and sample standard deviations
754: of $\Gamma$ for the BBF96 NLS1s and Sy1s. \label{fig_fgamma}}
755: \end{figure}
756:
757:
758: \clearpage
759:
760: \begin{deluxetable}{lccccccc}
761: \tablecolumns{8}
762: \tablewidth{470pt}
763: \tablecaption{\chandra\ observing log\label{tab_obslog}}
764: \tablehead{
765: \colhead{SDSS Name\tablenotemark{a}} &
766: \colhead{Date} &
767: \colhead{$t_{\rm exp}$} &
768: \colhead{S\tablenotemark{b}} &
769: \colhead{M\tablenotemark{b}} &
770: \colhead{H\tablenotemark{b}} &
771: \colhead{$N_H$\tablenotemark{c}} &
772: \colhead{CR\tablenotemark{d}}
773: \\
774: \colhead{} &
775: \colhead{} &
776: \colhead{(s)} &
777: \colhead{} &
778: \colhead{} &
779: \colhead{} &
780: \colhead{($10^{20}$\,cm$^{-2}$)} &
781: \colhead{($s^{-1}$)}
782: }
783:
784: \startdata
785: J002305.03$-$010743.5 &25--08--2003 &1940 &30 &44 &13 &2.81 &$0.045\pm 0.005$\\
786: J002752.39$+$002615.8 &03--09--2003 &1560 &7 &21 &14 &2.72 &$0.027\pm 0.004$\\
787: J015652.43$-$001222.0 &12--09--2003 &1710 &8 &6 &3 &2.62 &$0.010\pm 0.002$\\
788: J022756.28$+$005733.1 &23--06--2003 &1730 &6 &10 &7 &2.84 &$0.013\pm 0.003$\\
789: J022841.48$+$005208.6 &24--06--2003 &2000 &51 &65 &22 &2.76 &$0.069\pm 0.006$\\
790: J031427.47$-$011152.4 &03--09--2003 &1910 &48 &43 &9 &5.70 &$0.052\pm 0.005$\\
791: J101314.86$-$005233.5 &08--01--2003 &1990 &63 &70 &14 &3.64 &$0.074\pm 0.006$\\
792: J104230.14$+$010223.7 &20--02--2003 &1730 &33 &42 &12 &3.72 &$0.050\pm 0.005$\\
793: J121415.17$+$005511.4 &07--02--2003 &1960 &49 &42 &12 &1.94 &$0.053\pm 0.005$\\
794: J125943.59$+$010255.1 &04--03--2003 &1940 &0 &2 &3 &1.62 &$0.003\pm 0.001$\\
795: J131108.48$+$003151.8 &10--03--2003 &1730 &7 &6 &0 &1.90 &$0.008\pm 0.002$\\
796: J141234.68$-$003500.0 &07--01--2003 &2090 &36 &48 &16 &3.29 &$0.048\pm 0.005$\\
797: J143030.22$-$001115.1 &23--04--2003 &1950 &5 &10 &9 &3.15 &$0.012\pm 0.003$\\
798: J144932.70$+$002236.3 &09--07--2003 &2150 &129 &165 &68 &3.75 &$0.168\pm 0.009$\\
799: J145123.02$-$000625.9 &22--04--2003 &2120 &52 &74 &14 &3.84 &$0.066\pm 0.006$\\
800: J170546.91$+$631059.1 &17--09--2003 &1940 &6 &8 &2 &2.57 &$0.008\pm 0.002$\\
801: J233853.83$+$004812.4 &28--08--2003 &1910 &28 &23 &11 &3.88 &$0.032\pm 0.004$\\
802:
803: \enddata
804:
805: \tablenotetext{a}{Format: SDSS JHHMMSS.ss$\pm$DDMMSS.s}
806: \tablenotetext{b}{Net counts in soft (0.5--0.9~keV), medium
807: (0.9--2.0~keV), hard (2.0--8.0~keV) bands.}
808: \tablenotetext{c}{Galactic foreground \ion{H}{1} column density,
809: found with the {\tt nh} utility.}
810: \tablenotetext{d}{Net \chandra\ 0.5--8\,keV count rate.}
811:
812: \end{deluxetable}
813:
814: \begin{deluxetable}{lccccccc}
815: \tablecolumns{8}
816: \tablewidth{500pt}
817: \tablecaption{NLS1 Optical and X-ray Properties\label{tab_objects}}
818: \tablehead{
819: \colhead{Name\tablenotemark{a}} &
820: \colhead{z\tablenotemark{b}} &
821: \colhead{FWHM(H$\beta$)\tablenotemark{c}} &
822: \colhead{$\Gamma$\tablenotemark{d}} &
823: \colhead{$\log(E L_E)$\tablenotemark{e,g}} &
824: \colhead{$\log(\lambda L_\lambda)$\tablenotemark{f,g}} &
825: \colhead{$\log(\frac{\mbh}{M_\sun})$\tablenotemark{h}} &
826: \colhead{Note}
827: \\
828: \colhead{} &
829: \colhead{} &
830: \colhead{(km\,s$^{-1}$)} &
831: \colhead{} &
832: \colhead{(erg\,s$^{-1}$)} &
833: \colhead{(erg\,s$^{-1}$)} &
834: \colhead{} &
835: \colhead{}
836: }
837:
838: \startdata
839: J0023$-$0107 &0.166 &1160 &$2.21\pm 0.41$ &$42.89\pm 0.20$ &$43.89$ &6.73 &\nodata\\
840: J0027$+$0026 &0.205 &1830 &$1.61\pm 0.50$ &$42.80\pm 0.29$ &$43.90$ &7.14 &\nodata\\
841: J0156$-$0012 &0.163 &1320 &$2.04^{+0.75}_{-0.49}$ &$42.24\pm 0.24$ &$43.84$ &6.82 &1\\
842: J0227$+$0057 &0.128 &770 &$1.43\pm 0.44$ &$42.04\pm 0.21$ &$43.40$ &6.04 &1\\
843: J0228$+$0052 &0.186 &990 &$2.46\pm 0.25$ &$43.27\pm 0.09$ &$43.85$ &6.57 &\nodata\\
844: J0314$-$0111 &0.387 &1810 &$3.15\pm 0.44$ &$43.97\pm 0.16$ &$44.70$ &7.69 &\nodata\\
845: J1013$-$0052 &0.276 &1580 &$2.71\pm 0.34$ &$43.74\pm 0.08$ &$44.43$ &7.38 &\nodata\\
846: J1042$+$0102 &0.116 &1010 &$2.38\pm 0.31$ &$42.67\pm 0.14$ &$43.66$ &6.46 &\nodata\\
847: J1214$+$0055 &0.396 &1980 &$2.70\pm 0.28$ &$43.93\pm 0.13$ &$44.50$ &7.62 &\nodata\\
848: J1259$+$0102 &0.394 &1460 &$0.25^{+0.80}_{-1.01}$ &$41.85\pm 0.45$ &$44.54$ &7.39 &1\\
849: J1311$+$0031 &0.429 &1640 &$3.06\pm 0.96$ &$43.27\pm 0.28$ &$44.70$ &7.60 &2\\
850: J1412$-$0035 &0.127 &1100 &$2.34\pm 0.38$ &$42.74\pm 0.14$ &$43.36$ &6.31 &\nodata\\
851: J1430$-$0011 &0.103 &1740 &$0.92\pm 0.64$ &$41.62\pm 0.47$ &$43.19$ &6.60 &\nodata\\
852: J1449$+$0022 &0.081 &1070 &$2.12\pm 0.16$ &$42.82\pm 0.08$ &$43.38$ &6.31 &3\\
853: J1451$-$0006 &0.139 &1120 &$2.56\pm 0.22$ &$42.96\pm 0.12$ &$43.57$ &6.48 &\nodata\\
854: J1705$+$6310 &0.119 &1660 &$2.40^{+1.02}_{-0.55}$ &$41.96\pm 0.25$ &$43.55$ &6.81 &1\\
855: J2338$+$0048 &0.170 &1010 &$2.69\pm 0.69$ &$42.86\pm 0.18$ &$43.63$ &6.43 &\nodata\\
856:
857: \enddata
858:
859: \tablenotetext{a}{Truncated to JHHMM$\pm$DDMM}
860: \tablenotetext{b}{Redshift from the SDSS catalog, as listed in WPM02}
861: \tablenotetext{c}{From WPM02}
862: \tablenotetext{d}{Derived from spectral fitting in {\tt Sherpa}, unless noted
863: otherwise.}
864: \tablenotetext{e}{Monochromatic, rest-frame 1~keV luminosity inferred from the
865: best--fit {\tt Sherpa} or HR$_{a/b}$ model flux, with quoted errors from
866: {\tt Sherpa} or Poisson errors on the total number of counts respectively.}
867: \tablenotetext{f}{Monochromatic, rest-frame 5100\AA\ luminosity.
868: Fluxes are measured from
869: the original SDSS spectra used in WPM02; errors are
870: considered negligible for $\mbh$ estimates.}
871: \tablenotetext{g}{All luminosities are calculated assuming
872: $H_0=70$\,km\,s$^{-1}$\,Mpc$^{-1}$, $\Omega_m=0.3$, $\Omega_\Lambda=0.7$.}
873: \tablenotetext{h}{Estimated from the \citet{kaspi00} relation.}
874:
875:
876:
877: \tablecomments{
878: (1) $\Gamma$ derived from HR$_{\rm b}$;
879: (2) $\Gamma$ derived from HR$_{\rm a}$;
880: (3) Spectrum not well fit with power law $+$ Galactic absorption model; see
881: \S\ref{sec_771}.
882: }
883:
884: \end{deluxetable}
885:
886:
887: \begin{deluxetable}{lccl}
888: \tablecolumns{4}
889: \tablewidth{250pt}
890: \tablecaption{Statistics of $\Gamma$ for NLS1 X-ray Samples \label{tab_gamma}}
891: \tablehead{
892: \colhead{Sample} &
893: \colhead{$\langle\Gamma\rangle$} &
894: \colhead{$\sigma_\Gamma$} &
895: \colhead{$\Gamma$ Range}
896: }
897: \startdata
898: BBF96 NLS1s &3.22 &0.73 &1.8--5.2\\
899: WPM02 (RASS) &2.75 &0.31 &1.5--4.4\\
900: This Paper &2.60 &0.39 &1.1--3.4\\
901: BBF96 Sy1s &2.31 &0.28 &1.4--3.4\\
902: \enddata
903: \end{deluxetable}
904:
905:
906: \end{document}
907: