1: \documentclass[12pt,preprint]{aastex}
2:
3: \def\arcsecpoint{$''\!.$}
4: \def\deg{$^{\rm o}$}
5: \def\ltsim{\raisebox{-.5ex}{$\;\stackrel{<}{\sim}\;$}}
6: \def\gtsim{\raisebox{-.5ex}{$\;\stackrel{>}{\sim}\;$}}
7:
8: %\journalid{}{}
9: %\articleid{}{}
10:
11: %\slugcomment{submitted to {\it The Astrophysical Journal}}
12:
13: \shortauthors{Crenshaw et al.}
14: \shorttitle{Mass Outflow from NGC~4151}
15:
16: %\received{}
17: \begin{document}
18:
19: \title{Mass Outflow from the Nucleus of the Seyfert 1 Galaxy
20: NGC~4151\altaffilmark{1}}
21:
22: \author{D.M. Crenshaw\altaffilmark{2} \& S.B. Kraemer\altaffilmark{3,4}}
23:
24: \altaffiltext{1}{Based on observations made with the NASA/ESA Hubble Space
25: Telescope, obtained from the Data Archive at the Space Telescope Science
26: Institute, which is operated by the Association of Universities for Research in
27: Astronomy, Inc., under NASA contract NAS 5-26555.}
28:
29: \altaffiltext{2}{Department of Physics and Astronomy, Georgia State
30: University, Astronomy Offices, One Park Place South SE, Suite 700,
31: Atlanta, GA 30303; crenshaw@chara.gsu.edu}
32:
33: \altaffiltext{3}{Institute for Astrophysics and Computational Sciences,
34: Department of Physics, The Catholic University of America, Washington, DC
35: 20064; kraemer@yancey.gsfc.nasa.gov}
36:
37: \altaffiltext{4}{Exploration of the Universe Division, Code 667, NASA's
38: Goddard Space Flight Center, Greenbelt, MD 20771}
39:
40: \begin{abstract}
41:
42: We present an analysis of UV and optical spectra of NGC~4151 obtained at high
43: spectral and angular resolutions with the {\it Hubble Space Telescope's} {(\it
44: HST's)} Space Telescope Imaging Spectrograph (STIS). We identify a kinematic
45: component of the emission lines that has a width of 1170 km s$^{-1}$ (FWHM),
46: intermediate between those from the broad and narrow (emission) line regions
47: (BLR and NLR). We present evidence that these emission lines arise from the same
48: gas responsible for most of the high-column UV and X-ray absorption (component
49: ``D+E'') that we see in outflow at a distance of $\sim$0.1 pc from the central
50: nucleus. The gas in this intermediate-line region (ILR) shields the NLR and has
51: a global covering factor of $\sim$0.4, based on the observed C~IV fluxes,
52: indicating mass outflow over a large solid angle centered on the accretion
53: disk's axis. A large transverse velocity ($v_T \gtsim 2100$ km s$^{-1}$)
54: compared to the radial velocity centroid ($v_r = -490$ km s$^{-1}$)
55: indicates that the kinematics is dominated by rotation at this distance, but has
56: a significant outflow component. The mass outflow rate at 0.1 pc is
57: $\sim$0.16 M$_{\odot}$ yr$^{-1}$, which is about 10 times the accretion rate.
58: Based on physical conditions in the gas and dynamical considerations, models
59: that invoke magnetocentrifugal acceleration (e.g., in an accretion-disk wind)
60: are favored over those that rely on radiation driving or thermal expansion as
61: the principal driving mechanism for the mass outflow.
62:
63: \end{abstract}
64:
65: \keywords{galaxies: individual (NGC 4151) -- galaxies: Seyfert}
66: ~~~~~
67:
68: \section{Introduction}
69:
70: Mass outflows of ionized gas from active galactic nuclei (AGN) are detected in
71: the form of UV and X-ray absorption lines that are intrinsic to the AGN and
72: blueshifted with respect to the systemic velocities of their host galaxies.
73: While the radial velocities and physical conditions of the outflows have been
74: characterized through intensive ground- and space-based observations over the
75: past decade, their origin(s) and means of outward acceleration are still not
76: well understood. The most popular dynamical models invoke accretion-disk winds
77: and one or more acceleration mechanisms, including radiation driving, thermal
78: expansion, and/or magnetocentrifugal winds (Crenshaw, Kraemer, \& George 2003,
79: and references therein; Proga 2003; Chelouche \& Netzer 2005; Everett 2005).
80: Another possible source of the outflows is the putative obscuring torus (Krolik
81: \& Kriss 2001), although it is possible that the torus is just the outward
82: extension of a dusty accretion-disk wind (K\"{o}nigl \& Kartje 1994; Elitzur \&
83: Shlosman 2006).
84:
85: To distinguish between the dynamical models and determine the nature and origin
86: of the mass outflows, we need to know the locations and complete kinematics
87: (transverse as well as radial velocities) of the intrinsic absorbers. Several
88: in-depth studies of Seyfert galaxies, the nearest bright AGN, have placed the
89: absorbers at distances of tenths to tens of parsecs ($\sim$100 -- 10,000 light
90: days) from the central continuum source (i.e., accretion disk plus X-ray corona)
91: (Crenshaw \& Kraemer 2005, and references therein). On the other hand, the
92: high-ionization broad (emission) line region (BLR) is much closer, typically at
93: a distance of only 3 -- 10 light days in these same Seyfert galaxies (Peterson
94: et al. 2004; Metzroth et al. 2006). Furthermore, the radial velocities of the
95: absorbers are nearly always $<$ 2000 km s$^{-1}$ (Crenshaw et al. 2003), whereas
96: the BLR velocities typically extend out to at least 10,000 km s$^{-1}$. Thus, we
97: have suggested that most of the intrinsic absorbers observed in Seyfert galaxies
98: are located in the inner narrow (emission) line region (NLR) (Crenshaw \&
99: Kraemer 2005), based on similarities in locations and velocities, although this
100: certainly does not rule out an origin much closer to the nucleus.
101:
102: If we want to detect and characterize mass outflows close to the nucleus, the
103: best candidate is probably NGC~4151, the nearest ($cz =$ 995 km s$^{-1}$) and
104: apparently brightest Seyfert 1 galaxy. It's UV spectrum shows intrinsic
105: absorption lines, spanning a wide range in ionization (O~I to N~V) (Bromage et
106: al. 1985), from a number of kinematic components that cover a large range in
107: radial velocity centroid (v$_r =$ 0 to $-$1600 km s$^{-1}$) and velocity width
108: (FWHM $=$ 15 to 940 km s$^{-1}$) (Kraemer et al. 2001). Component ``D$+$E''
109: ($v_r = -$490 km s$^{-1}$, FWHM $=$ 435 km s$^{-1}$), which is a blend of two
110: kinematic components identified by Weymann et al. (1997), dominates the
111: absorption in the UV with its high column density. This component shows variable
112: ionic column densities that respond to continuum changes and strong metastable
113: C~III and Fe~II absorption lines (Crenshaw et al. 2000), indicating high number
114: densities ($n_H \approx$ 10$^{7 - 9}$ cm$^{-3}$) and proximity to the nucleus
115: (Kraemer et al. 2001).
116:
117: To further explore the properties of mass outflow in NGC~4151, we obtained
118: contemporaneous high-resolution UV and X-ray spectra in 2002 May with the {\it
119: HST} Space Telescope Imaging Spectrograph (STIS), the {\it Far Ultraviolet
120: Spectroscopic Explorer} ({\it FUSE}), and the {\it Chandra X-ray
121: Observatory} ({\it CXO}), and retrieved additional archive spectra from
122: these missions. We give the details of the observations and analysis in
123: Kraemer et al. (2005, Paper I) and Kraemer et al. (2006, Paper II) for the
124: X-ray and UV data respectively. To summarize our findings, in order to match
125: the ionic column densities of D$+$E and the observed covering factors of the
126: background continuum and line emission in the line of sight ($C_{los}$), we
127: needed four physical subcomponents (D$+$Ea, b, c, and d) characterized by
128: different ionization parameters ($U$) and hydrogen column densities ($N_H$),
129: as shown in Table 1. Based on the ionization parameters and the number
130: densities ($n_H$) from C III metastable absorption, we found that the D$+$E
131: absorbers were only $\sim$0.1 pc from the central continuum source. We also
132: found that D+Ea, the UV absorber with the highest $U$ and $N_H$, was
133: responsible for the bulk of the X-ray absorption, although we needed to
134: include an additional, more highly ionized component (``X-High'') to account
135: for all of the X-ray absorption , which is also listed in Table 1.
136: Furthermore, we found that the D+E absorbers showed strong evidence for bulk
137: motion across the BLR. From a {\it CXO} spectrum in 2000 March, we found a
138: lower $U$ and higher $N_H$ for D+Ea (Table 1) than in 2002 May; the latter
139: yielded a transverse velocity of $v_T$ $\gtsim$ 2500 km s$^{-1}$ (Paper I).
140: Between 2001 April and 2002 May, we noticed a change in the $C_{los}$ of D+Ed
141: (Table 1), which gave $v_T$ $\approx$ 2100 km s$^{-1}$ (Paper II), close to
142: the previous lower limit.
143:
144: The proximity of the D+E absorbers to the central SMBH in NGC~4151 and their
145: large transverse velocities, possibly reflecting orbital motion, suggested to us
146: that further study of this component could lead to more insight into the
147: connection between accretion disks and mass outflows in AGN. In particular, the
148: gas from this component must contribute in some fashion to the emission lines,
149: and if the contribution is significant and can be isolated, it should
150: provide additional constraints on the geometry, kinematics, and physical
151: conditions of the gas. In Paper II, we identified a distinct component of the
152: He~II and C~IV emission lines in the STIS echelle spectra of NGC~4151 with a
153: FWHM $=$ 1170 km s$^{-1}$, which is intermediate between that of the C~IV broad
154: (FWHM $\approx$ 10,000 km s$^{-1}$) and narrow (FWHM $=$ 250 km s$^{-1}$)
155: components. We noted that the emission from this ``intermediate line region''
156: (ILR) was absorbed by D+E. Here, we investigate the properties of the ILR in
157: more detail, using additional emission lines from the six epochs of STIS
158: high-resolution echelle observation. For convenience, we list the dates of
159: observations and continuum fluxes at 1350 \AA\ in Table 2.
160:
161: \section{Analysis}
162:
163: The available STIS E140M spectra of NGC~4151 provide a unique opportunity to
164: deconvolve the emission-line components of NGC~4151, due to the high spectral
165: resolution of the grating (R $=$ $\lambda$/$\Delta\lambda$ $\approx$ 45,000)
166: and the high angular resolution of {\it HST}. To isolate the ILR contribution,
167: we selected a low continuum-flux spectrum with the best signal-to-noise, which
168: was obtained on 2000 June 15. The low state minimizes the contribution of the
169: BLR to the profile, and the small aperture used for all of the E140M spectra
170: (0\arcsecpoint2 $\times$ 0\arcsecpoint2) minimizes the contribution from the
171: NLR. By happenstance, there was a contemporaneous high-resolution observation of
172: NGC~4151 on 2000 July 2 through a 52 $\times$ 0\arcsecpoint2 slit with the STIS
173: G430M grating (R $\approx$ 10,000), covering the H$\beta$ and [O~III]
174: $\lambda\lambda$4959, 5007 lines.
175:
176: We show the central regions of the emission line profiles from these
177: observations in Figure 1. The strong blueshifted absorption in the UV resonance
178: lines (Ly$\alpha$ and the C~IV and N~V doublets) is due to the high-column
179: absorption component D+E (the other absorption components are identified in
180: Paper II). H$\beta$ also shows strong D+E absorption in the low continuum flux
181: state, whereas a STIS G430M spectrum obtained in a high state on 1997 July 17
182: shows only weak H$\beta$ absorption (Hutchings et al. 2002), consistent with the
183: weakness or absence of UV low-ionization lines from component D+E in other high
184: continuum states (Kraemer et al. 2001, Paper II).
185:
186: Figure 1 shows that the He~II $\lambda$1640 line is unaffected by intrinsic
187: absorption and has two distinct emission components: one from the NLR (FWHM $=$
188: 250 km s$^{-1}$) and one from the ILR (FWHM $=$ 1170 km s$^{-1}$). There is
189: likely a contribution from the BLR, but it is too faint to be seen in these data
190: and is therefore included as part of the continuum fit (spline fits over large
191: wavelength regions). We fit the profile of He~II with two Gaussians, and used
192: these profiles as templates to determine the NLR and and ILR contributions to
193: the other lines. We also detected faint BLR components in the wings of C~IV,
194: Ly$\alpha$, and H$\beta$, and fit these with splines. To model the observed
195: emission-line profiles, we reproduced the NLR and ILR templates from He~II at
196: the expected positions of the lines (including the doublet lines of C~IV
197: $\lambda\lambda$1548, 1551 and N~V$\lambda\lambda$ 1239, 1243), retaining the
198: same velocity widths, and scaled them in intensity until we obtained a suitable
199: match. The fits to the NLR components were straightforward, since they they are
200: unaffected by absorption, except for that in the cores of the UV resonance lines
201: due to the host galaxy's disk and/or halo (components F and F$'$ in Weymann et
202: al. 1997). The fits to the ILR components were also relatively easy to
203: accomplish, relying on the observed red wings of the profiles for lines with
204: strong blueshifted absorption. Overall, this method of template scaling provided
205: excellent fits to the profiles.
206:
207: As shown in Figure 1, the use of high spatial- and spectral-resolution STIS
208: spectra at a low continuum state allows us to easily detect and deconvolve the
209: ILR profiles, which are essentially symmetric around zero km s$^{-1}$.
210: Inspection of the profiles and fits reveals a couple of interesting results.
211: 1) Component D+E absorbs the emission from not only the continuum and BLR, but
212: the ILR as well, as can be seen in all of the UV resonance lines. 2) The
213: maximum blueshifted velocity of D+E ($\sim$ $-$1400 km s$^{-1}$) is close to
214: the maximum velocity of the ILR, which suggests self-absorption by the ILR.
215: Hence the ILR contribution resembles a P-Cygni profile, as seen for example in
216: the H$\beta$ profile.
217:
218: As discussed in Paper II, if the velocity width of the ILR component is due
219: primarily to gravitational motion, similar to the case for the BLR (Peterson et
220: al. 2004), then the FWHM of the ILR profile (1170 km s$^{-1}$) and the mass of
221: the supermassive black hole (4.1 x 10$^{7}$ M$_{\odot}$, Metzroth et al. 2006)
222: give a distance from the ILR to the central continuum source of $\sim$0.1 pc.
223: This is consistent with the distance of the D+E absorbers from the central
224: source based on their densities and ionization parameters (Paper II).
225:
226: All of the above suggests a connection between the D+E absorption and the ILR,
227: including the possibility that the ILR emission lines originate directly from
228: gas responsible for the D+E absorbers. To test this idea, we measured the
229: emission-line fluxes from the ILR using the fits in Figure 1. Other UV
230: emission lines were too weak and/or absorbed to give reliable fluxes. We
231: determined uncertainties in the fluxes by varying the scale factors applied to
232: each template until we obtained fits that were clearly no longer acceptable.
233: Table 3 show the observed line ratios relative to H$\beta$, as well as
234: reddening-corrected ratios based on the very small reddening of E(B$-$V) $=$
235: 0.02 in our line of sight to the nucleus, which is entirely from our Galaxy
236: (Crenshaw \& Kraemer 2005). We present photoionization models of the line
237: ratios in the next section.
238:
239: The strong C~IV line provides the best opportunity for deconvolving the
240: emission-line components in the other five epochs of STIS echelle
241: observations, so that we can characterize the variability in the components.
242: The other UV lines shown in Figure 1 are noisier and more difficult to
243: deconvolve in the other epochs, particularly in high states when there is a
244: strong BLR contribution. An attempt to deconvolve these lines resulted in
245: error bars that were larger than the measured variations. To determine the
246: emission-line fluxes of the separate C~IV components in the six epochs of STIS
247: E140M observations, we followed the same procedure as above; we scaled the
248: separate C~IV BLR, ILR, and NLR templates from 2000 June 15 until we obtained
249: matches to the observed profiles.
250:
251: Figure 2 shows the light curves for the C~IV emission components and the
252: continuum flux at 1350 \AA. Unsurprisingly, the narrow component remained
253: constant to within the uncertainties, which is to be expected from emission from
254: an extended region (constrained by the aperture to be $\sim$14 $\times$ $\sim$14
255: pc in the plane of the sky and possibly larger in the line of sight). The BLR
256: shows large amplitude C~IV variations similar to those of the UV continuum
257: (maximum/minimum flux $\approx$ 10), consistent with previous results from more
258: intensive UV monitoring that indicate the BLR is small ($\sim$6.8 light days in
259: diameter, Metzroth, et al. 2006) and therefore responds rapidly to continuum
260: variations. The ILR shows a positive correlation with continuum flux as well,
261: although it is not possible to determine a time lag from these undersampled
262: light curves. However, the timescales of the variations do put an upper limit of
263: $\sim$140 light days ($\sim$ 0.12 pc) on the radius of the ILR (i.e., from the
264: time interval between the third and fourth observations), consistent with the
265: value that we derived above. Interestingly, the amplitude of the ILR light curve
266: is significantly smaller (maximum/minimum flux $\approx$ 4) than that of the
267: BLR, which may be be expected from a much larger region in which the amplitude
268: is reduced by the time-delayed response of different parts of the ILR to
269: continuum variations (Peterson 1993).
270:
271: \section{Photoionization Models}
272:
273: In Papers I and II, we determined that D+Ea has the highest $C_{los}$, $U$, and
274: $N_H$ of all the UV absorbers. If this subcomponent has a non-negligible global
275: covering factor, then it must dominate the line emission from the absorbers,
276: and it is therefore a likely source of the ILR emission. Here, we test
277: this hypothesis by comparing the predicted emission-line fluxes from our
278: photoionization models of D+Ea with the observed values from the ILR.
279:
280: The details of the photoionization models used for this study are given in Paper
281: I. To summarize, the models were generated using the Beta 5 version of Cloudy
282: (Ferland et al. 1998), which includes estimated $\Delta$n$=$0 dielectronic
283: recombination (DR) rates for the M-shell states of Fe and the L-shell states of
284: the third row elements (Kraemer, Ferland, \& Gabel 2004). We assumed roughly
285: solar elemental abundances (see Paper II) and that the absorbing gas is free of
286: cosmic dust, consistent with our previous result that there is essentially no
287: reddening in our line of sight to the nucleus outside of the Galaxy (Crenshaw \&
288: Kraemer 2005). As per convention, the models are parameterized in terms of $U$
289: and $N_{H}$. We modeled the intrinsic spectral energy distribution as a broken
290: power law of the form $L_{\nu} \propto \nu^{\alpha}$ as follows: $\alpha = -1.0$
291: for energies $<$ 13.6 eV, $\alpha = -1.3$ over the range 13.6 eV $\leq$ h$\nu$
292: $<$ 0.5 keV, and $\alpha = -0.5$ above 0.5 keV. We included a low energy
293: cut-off at $1.24 \times 10^{-3}$ eV (1 mm) and a high energy cutoff at 100 keV.
294: The luminosity in ionizing photons for the 2002 May epoch was $Q =
295: 1.1\times10^{53}$ photons s$^{-1}$.
296:
297: In order to determine whether D+Ea could be the source of the ILR emission, we
298: scaled the ionization parameter from the 2002 May observation to match the
299: lower continuum flux observed in 2000 June, which yields $U = 10^{-0.92}$. We
300: then fixed the other parameters to the 2002 May epoch and ran a new model; we
301: list the predicted line ratios from this model (``Low-N'') in Table 3.
302: Overall, the correspondence between the model and observed values is quite
303: good, considering that no attempt was made to adjust the parameters derived
304: from the D+Ea absorber to match the ILR ratios. The low [O~III]/H$\beta$ ratio
305: in particular confirms that we have the correct density ($n_H$).
306:
307: Table 3 shows that C~IV $\lambda$ 1550 is slightly underpredicted and N~V
308: $\lambda$ 1240 is too low by a factor of $\sim$2.4. While increasing $n_{H}$ by
309: a factor of $\sim$ 10 brings the predicted C~IV emission inside the error bars,
310: the [O~III] $\lambda$ 5007 is then collisionally suppressed. This scenario would
311: therefore require an additional component of lower density gas to provide the
312: [O~III] emission. Moreover, N~V is still underpredicted at the higher density.
313: Although strong N~V emission is often cited as evidence of super-solar nitrogen
314: abundances (e.g., Hamann et al. 2002), we find that increasing N/H in the models
315: results in an overprediction of N~IV] $\lambda$ 1486/H$\beta$ (the observed
316: value is $\sim$1.2 for the BLR, ILR, and NLR combined). The underprediction of
317: N~V is not unique to this case, as we have encountered it on a number of
318: occasions when modeling the physical conditions in the NLR (e.g., Kraemer et al.
319: 2000; Kraemer \& Crenshaw 2000).
320:
321: Having established that it is plausible that D+Ea is the source of the ILR
322: emission, we generated a grid of models to determine if the observed
323: variations in the C~IV flux could result from the response of this component
324: to changes in the strength of the ionizing continuum. As with our model of the
325: emission line ratios, we assumed that the ionizing flux, and therefore $U$,
326: scale linearly with the continuum flux at 1350 \AA. Figure 3 shows that Low-N,
327: with $N_{H}$ = 10$^{22.46}$ cm$^{-2}$, cannot replicate the behavior of the
328: observed C~IV fluxes, as the model curve turns over too quickly, due to the
329: fact that too much C$^{+3}$ is ionized into C$^{+4}$ at the higher flux
330: states.
331:
332: We were able to remedy this problem by increasing $N_{H}$ to 10$^{22.93}$
333: cm$^{-2}$, the column density of D+Ea in 2000 March (Table 1). In Table 3,
334: we list the emission line ratios for this model (``High-N''). The differences in
335: line ratios between the low- and high column density models are negligible.
336: Figure 3 shows that High-N is able to reproduce the observed positive
337: correlation of continuum and C IV ILR fluxes. Thus, we conclude that the higher
338: column density is more appropriate globally, and that the low column in 2002 May
339: due to transverse motion is atypical. This is consistent with our previous claim
340: that variations in the equivalent widths of the absorption lines from D+E over
341: the time period of the STIS echelle observations were primarily due to changes
342: in the ionizing continuum, until the large drop in column density seen in 2002
343: May (Paper II). Interestingly, we found that varying the ionization parameter
344: for X-High suppresses $F_{CIV}$ for the lowest values of $U$, which suggests
345: that the ILR may not be fully screened by this component.
346:
347: The global covering factor ($C_g$) of the ILR gas can be obtained from the scale
348: factor needed to match the model curve in Figure 3 with the general trend of the
349: observed C~IV fluxes. For model High-N, we find that $C_g =$ 0.4, which is
350: significant, and yet reassuringly does not exceed one. The predicted C~IV flux
351: initially increases with continuum flux, but begins to decrease after peaking at
352: F(1350) = 1.3 $\times 10^{-13}$ ergs cm$^{-2}$ s$^{-1}$. Thus, our highest
353: observed C~IV flux comes after the peak, and the model predicts that even higher
354: continuum levels would result in significantly lower C~IV fluxes from the ILR.
355:
356: Although the High-N model matches the behavior of the ILR C~IV emission
357: reasonably well, the agreement is not perfect. We attribute these discrepancies
358: to the size of the ILR and light travel time effects, which are difficult to
359: disentangle since the light curves in Figure 2 are severely undersampled. For
360: example, the most discrepant C~IV flux, obtained during 2002 May, may simply be
361: due to the fact that the higher continuum flux in our line-of-sight has yet to
362: reach the bulk of the ILR gas, which is irradiated by weaker continuum flux from
363: an earlier epoch as we see it. Another possible problem is that the global
364: covering factor required for the ILR models exceeds, by a factor of $\sim$ 4,
365: the constraint that we derived from the forbidden O~VII 22.1 \AA~ emission line
366: in Paper I. A likely explanation is that most of the O~VII emission is absorbed
367: across the entire profile by the large soft X-ray opacity (primarily the He~II
368: edge) of D+Ea, whereas the UV resonance lines are only self-absorbed in part of
369: their blue wings.
370:
371: \section{Summary and Discussion}
372:
373: We have discovered a distinct component of the emission lines in NGC~4151 that
374: has a width (FWHM $\approx$ 1170 km s$^{-1}$) between those of the classic
375: broad and narrow emission lines. We were able to isolate this component by using
376: {\it HST} spectra at high spectral and angular resolutions, which reduce the
377: contribution of the NLR to the lines, of NGC~4151 in a low continuum-flux state,
378: which reduces the BLR contribution. This intermediate-line region (ILR) would be
379: difficult to detect in ground-based observations, especially in moderate- to
380: high-flux states, and its connection to previous claims of ILRs in Seyfert
381: galaxies and quasars (e.g., Brotherton et al. 1994; cf. Sulentic \& Marziani
382: 1999) is therefore uncertain.
383:
384: The properties of the ILR lead us to believe that we have found the emission
385: from at least the major component (D+Ea) of the high-column outflowing absorbers
386: in NGC~4151, based on the following evidence. 1) The ILR emission and
387: blueshifted absorption lines extend to about the same velocity range ($-$1400 km
388: s$^{-1}$), and together they resemble P-Cygni profiles, suggesting
389: self-absorption. 2) The approximate location of the ILR ($\sim$0.1pc), derived
390: from the width of the emission lines, is consistent with that of the D+E
391: absorbers based on their ionization parameters and densities. 3) A
392: photoionization model of the D+Ea absorber that is completely constrained in
393: terms of $U$, $N_H$, and $n_H$ gives line ratios that are close to the observed
394: ILR values.
395:
396: Given that the ILR emission lines are from the same kinematic component that
397: produces the D+Ea absorption, the emission lines provide additional constraints
398: on the mass outflow. Our photoionization models are able to reproduce the
399: observed correlation of continuum and C~IV fluxes in the STIS data if we use the
400: higher column density from 2000 March ($N_H$ $=$ 10$^{22.93}$ cm$^{-2}$ ), as
401: opposed to that from 2002 May ($N_H$ $=$ 10$^{22.5}$ cm$^{-2}$), suggesting that
402: the former value is a more appropriate global average. Furthermore, the scale
403: factor needed to match the observed ILR C~IV fluxes gives us a global covering
404: factor of $C_g$ $=$ 0.4 for the bulk of the mass outflow. This value derived for
405: an individual Seyfert is similar to that derived from the frequency of absorbers
406: in Seyfert galaxies, $C_g$ $\approx$ 0.5, assuming that all Seyfert 1 galaxies
407: have mass outflows (as opposed to $\sim$50\% of Seyferts having outflows with
408: C$_g$ $=$ 1, Crenshaw al. 1999). However, we note that the D+Ea absorption in
409: NGC~4151 has an unusually large column density and small distance from the
410: continuum source compared to most UV absorbers (Crenshaw et al. 2003). One
411: possible interpretation is that the majority of the mass outflow in NGC~4151 is
412: at an early stage of evolution. Another possibility is that the high column is
413: due to a special viewing angle, close to the edge of the NLR bicone, as
414: discussed below.
415:
416: Previous photoionization studies (Alexander et al. 1999; Kraemer et al. 2000)
417: have concluded that the ionizing radiation incident on the NLR in NGC~4151 is
418: filtered by absorption close to the continuum source. In Kraemer et al.
419: (2000), we predicted an absorber with $U = 1$ and $N_H = 10^{22.5}$, which are
420: remarkably close to the values for D+Ea and the ILR. Thus, the ILR is likely
421: shielding the NLR, which places some interesting constraints on its geometry.
422: Based on kinematic models of {\it HST} long-slit spectra of NGC~4151, we found
423: that the velocity field of the NLR can be explained by radial outflow in a
424: bicone with an axis inclined by $\sim$45\deg\ with respect to our line of
425: sight, with the outflowing gas roughly confined between inner and outer
426: half-opening (polar) angles of $\theta =$ 15\deg\ and 33\deg\ (Das et al.
427: 2005). Although this geometry places our line of sight just outside of the
428: bicone, we can still see the nucleus of this Seyfert 1 galaxy because the
429: source of obscuration, and therefore the bicone edges, do not have sharp edges
430: (Evans et al. 1993; Schmitt et al. 2006). In order to shield the NLR and be in
431: our line of sight as well, the ILR must cover at least polar angles from
432: 15\deg\ to 45\deg\, which yields a global covering factor of $C_g \geq 0.26$.
433: Extending the coverage to $\theta =$ 15 -- 55\deg\ increases $C_g$ to the
434: observed value of 0.4, whereas filling in the cone so that $\theta =$ 0 --
435: 55\deg\ increases it slightly more to 0.43. We conclude that mass outflow in
436: the ILR covers a large solid angle that is likely focused in the polar
437: direction. Furthermore, the high column density in our line of sight ($\theta
438: = 45$\deg) compared to other Seyferts that are more likely viewed pole-on,
439: suggests that the $N_H$ is increasing with polar angle, and assuming this
440: trend continues, it is consistent with previous suggestions that the source of
441: obscuration in the unified scheme is an optically thick wind more focused
442: toward the plane of the accretion disk ($\theta = 90$\deg) (K\"{o}nigl \&
443: Kartje 1994; Elitzur \& Shlosman 2006).
444:
445: In Papers I and II, we determined transverse velocities of the D+E absorbers
446: that were on the order of $v_T \approx 2100$ km s$^{-1}$ or larger. These
447: values are considerably higher than the radial velocity centroid at $v_r =
448: -490$ km s$^{-1}$, which suggests that the kinematics of the gas at $\sim$0.1
449: pc is dominated by rotation, but with a significant outflow component. A
450: transverse velocity of 2100 km s$^{-1}$ leads to a somewhat smaller distance
451: ($\sim$0.04 pc), but that assumes purely gravitational motion at this
452: distance, which may not be the case, as discussed below.
453:
454: It is interesting that the transverse velocity of the absorber is higher than
455: the largest emission-line radial velocity, given by the half-width at zero
456: intensity (HWZI) of the emission lines ($\sim$1400 km s$^{-1}$), which is
457: possible for a number of different outflow geometries. For example, in Figure
458: 4, we show perhaps the simplest model that can satisfy our geometric and
459: kinematic constraints. In this model, the line of sight velocity is the same
460: as the velocity in the ``r'' direction ($v_r$), which can also be equated to
461: the outflow velocity. For simplicity, we assume constant $v_{r}$ and
462: $v_{\phi}$ (in standard spherical coordinates) and that the flow is at a
463: constant polar angle ($v_T = v_{\phi}$, $v_{\theta} = 0$), although in general
464: $\bf{v_T}$ is the vector sum of $\bf{v_{\phi}}$ and $\bf{v_{\theta}}$. The
465: resulting curve of motion is a spiral along a cone with a half-opening angle
466: of $\theta$ = 45\deg. If $v_{\phi} = 2100$ km s$^{-1}$ holds for other ILR
467: clouds (i.e., those not seen in absorption), we calculate that the highest
468: emission-line radial velocity in this model is 1550 km s$^{-1}$, which is
469: indeed smaller than the transverse velocity and close to the observed HWZI of
470: the ILR profile. For comparison, we also plot a curve in Figure 4 with
471: $v_{\phi} = 10,000$ km s$^{-1}$, which helps to show the conical surface that
472: the clouds follows. However, this case results in emission-line velocities
473: much higher than those seen in the ILR. Thus, at least in this simple model,
474: $v_T$ cannot be much higher than $\sim$2100 km s$^{-1}$.
475:
476: It is easy to think of D+Ea as a cloud, as drawn in Figure 4, since the radial
477: thickness of D+Ea is $N_H/n_H = 1.3 \times 10^{16}$ cm $= 0.004$ pc, which is
478: only $\sim$4\%\ of its radial distance from the central source. Thus, the ILR
479: can then be thought of as a collection of ``clouds'' that altogether cover
480: $\sim$ 40\% of the sky as seen from the central source. However, it could also
481: take the form of a relatively thin (partial) shell or flow tube with column
482: density inhomogeneities in the radial direction. The relative thickness of
483: X-high is $\sim$30\% if it is near D+E (but a smaller percentage at smaller
484: distances), so it could be more like a semi-continuous ``wind''.
485:
486: Assuming that the radial velocity centroid of D+Ea is the approximate outflow
487: velocity, the mass outflow rate due to this component is $\dot{M}_{out} = 4 \pi
488: r N_H m_H C_g v_r = 0.16$ M$_{\odot}$ yr$^{-1}$, where $m_p$ is the mass of a
489: proton. If we include the other X-ray and UV components (dominated by X-high)
490: and assume the same $C_g$, the outflow rate increases to $\dot{M}_{out} = 0.19$
491: M$_{\odot}$ yr$^{-1}$. On the other hand, the inferred accretion rate is only
492: $\dot{M}_{acc} = 0.013$ M$_{\odot}$ yr$^{-1}$, based on the average bolometric
493: luminosity of NGC~4151 ($L_{bol} = 7.3 \times 10^{43}$ ergs s$^{-1}$, Kaspi et
494: al. 2005) and an expected efficiency of $\sim$0.1 in the conversion of mass
495: infall to energy (Peterson 1997). This difference can be explained by one of two
496: scenarios, or a combination of the two). 1) In a steady-state situation, most
497: (more than 90\%)of the inflowing mass at $\sim$0.1 pc is not eventually
498: accreted, but is instead driven outward, or 2) the current episode of mass
499: ejection in NGC~4151 is an unusual or infrequent event, preceded by a long
500: interval of mass accumulation into a reservoir (e.g., accretion disk).
501:
502: In Paper I, we examined the possible dynamical forces (radiation driving,
503: thermal expansion, magnetocentrifugal acceleration) that could possibly drive
504: the outflow in NGC~4151. Here we review our arguments using a revised value for
505: the black-hole mass ($M = 4.1 \times 10^{7}$ M$_{\odot}$, Metzroth et al. 2006),
506: which yields an Eddington ratio of $L_{bol}/L_{E} = 0.014$ . In order for
507: radiation pressure to drive the gas outward from the nucleus, the force
508: multiplier must be $FM \geq (L_{bol}/L_{E})^{-1} = 71$. From our photoionization
509: models, we found that $FM \leq 2 $ for X-High, $FM \leq 41 $ for D+Ea in
510: 2002, and $FM \leq 11 $ for D+Ea at historically high continuum flux levels
511: (four times the 2002 level, see Paper I). These values are upper limits because
512: they were calculated for the optically thin case, whereas the force multiplier
513: decreases into a cloud as it becomes optically thick (Chelouche \& Netzer 2005).
514: Dust grains are efficient absorbers of radiation and could greatly increase the
515: effective force multipliers, but the lack of reddening indicates that dust is
516: essentially absent in the high-column absorbers (e.g., a Galactic dust/gas ratio
517: would result in a reddening of E$_{B-V} \approx 16$ mag for D+Ea). The lack of
518: dust is not surprising since the absorbers are near and likely originated inside
519: the dust sublimation radius, $r_{sub} \approx 1.3L_{46}^{1/2}~T_{1500}^{-2.8}$
520: pc $\approx 0.11$ pc (Barvainis 1987). Thus, X-high cannot be radiatively driven
521: and D+Ea is only marginally susceptible to radiation driving in low to medium
522: flux states. However, we cannot rule out the possibility that the absorbers
523: had lower $U$ when they were closer to the central nucleus and were therefore
524: more susceptible to radiation driving.
525:
526: For a thermal wind, the radial distance at which the gas can escape is given by
527: $r_{esc} \geq G M m_H / T_g k$ (Crenshaw et al. 2003, and references therein),
528: where $T_g$ is the gas temperature and k is the Boltzmann constant. Our
529: photoionization models give $T_g =$ $5 \times 10^4$ K and 3 $\times 10^{6}$ K
530: for D+Ea and X-High, which yield $r_{esc} \geq 1.3 \times 10^{21}$ cm and $\geq
531: 2.2 \times 10^{19}$, respectively. These are much greater than the radial
532: locations of D+Ea and X-high, so these high-column components cannot be
533: attributed to thermal winds.
534:
535: An MHD flow is therefore a likely possibility, at least by comparison to the
536: current alternatives. Accretion disk winds that invoke magnetocentrifugal
537: acceleration (Emmering, Blandford, \& Shlosman 1992; Bottorff, Korista, \&
538: Shlosman 2000) predict that the gas will flow along rotating magnetic field
539: lines, which would result in significant transverse velocities as seen by an
540: outside observer, as illustrated in Figure 4. In the presence of
541: non-dissipative MHD waves, the line widths can greatly exceed the thermal widths
542: (Bottorff et al. 2000), which could explain the large width of D+Ea (FWHM $=$
543: 450 km s$^{-1}$). Everett (2005) presents an accretion-disk wind model that
544: incorporates both radiative and magnetocentrifugal acceleration. Further clues
545: to the nature of mass outflows in AGN will come from comparing predictions from
546: these models with detailed observational constraints, such as those we have
547: presented for NGC~4151.
548:
549: \acknowledgments
550:
551: We thank Brad Peterson, Gary Ferland, Alvin Das, and John Everett
552: for helpful discussions. Some of the data presented in this paper were obtained
553: from the Multimission Archive at the Space Telescope Science Institute (MAST).
554: STScI is operated by the Association of Universities for Research in Astronomy,
555: Inc., under NASA contract NAS5-26555. Support for MAST for non-HST data is
556: provided by the NASA Office of Space Science via grant NAG5-7584 and by other
557: grants and contracts. This research has made use of the NASA/IPAC Extragalactic
558: Database (NED) which is operated by the Jet Propulsion Laboratory, California
559: Institute of Technology, under contract with the National Aeronautics and Space
560: Administration.
561:
562: \clearpage
563: \begin{references}
564:
565: \reference{ale1999}Alexander, T., Sturm, E., Lutz, D., Sternberg, A., Netzer,
566: H., \& Genzel, R. 1999, \apj, 512, 204.
567:
568: \reference{bar1987}Barvainis, R. 1987, \apj, 320, 537.
569:
570: \reference{bro1985}Bromage, G.E., et al. 1985, \mnras, 215, 1.
571:
572: \reference{bot2000}Bottorff, M.C., Korista, K.T., Shlosman, I. 2000, \apj, 537,
573: 134.
574:
575: \reference{bro1994} Brotherton, M.S., Wills, B.J., Francis, P.J., \& Steidel,
576: C.C. 1994, \apj, 430, 495.
577:
578: \reference{che2005}Chelouche, D. \& Netzer, H. 2005, \apj, 625, 95.
579:
580: \reference{cre2005}Crenshaw, D.M. \& Kraemer, S.B. 2005, \apj, 625, 680.
581:
582: \reference{cre1999}Crenshaw, D.M., Kraemer, S.B., Boggess, A., Maran, S.P.,
583: Mushotzky, R.F., \& Wu, C.-C. 1999, \apj, 516, 750.
584:
585: \reference{cre2003}Crenshaw, D.M., Kraemer, S.B., \& George, I.M. 2003, \araa,
586: 41, 117.
587:
588: \reference{cre2001}Crenshaw, D.M., et al. 2000, \apj, 545, L27.
589:
590: \reference{das2005}Das, V., et al. 2005, \aj, 130, 945.
591:
592: \reference{emm1992}Emmering, R.T., Blandford, R.D., \& Shlosman, I. 1992,
593: \apj,
594: 385, 460.
595:
596: \reference{eli2006} Elitzur, M. \& Shlosman, I. 2006, \apj, in press
597: (astro-ph/0605686).
598:
599: \reference{eva1993}Evans, I.N., Tsvetanov, Z., Kriss, G.A., Ford, H.C.,
600: Caganoff, S., \& Koratkar, A.P. 1993, \apj, 417, 82.
601:
602: \reference{eve2005}Everett, J.E. 2005, \apj, 631, 689.
603:
604: \reference{fer1998}Ferland, G.J., Korista, K.T., Verner, D.A., Ferguson, J.W.,
605: Kingdon, J.B., \& Verner, E.M. 1998, \pasp, 110, 749.
606:
607: %\reference{gre1989}Grevesse, N., \& Anders, E. 1989, in AIP Conf. Proc. 183,
608: %Cosmic Abundances of Matter, e.d. C.J. Waddington (New York: AIP), 1
609:
610: \reference{ham2002}Hamann, F., Korista, K.T., Ferland, G.J., Warner, C., \&
611: Baldwin, J. 2002, \apj, 564, 592.
612:
613: \reference{hut2002}Hutchings, J.B., Crenshaw, D.M., Kraemer, S.B., Gabel, J.R.,
614: Kaiser, M.E., Weistrop, D., \& Gull, T.R. 2002, \aj, 124, 2543.
615:
616: \reference{kas2005}Kaspi, S., Maoz, D., Netzer, H., Peterson, B.M.,
617: Vestergaard, M., \& Jannuzi, B.T. 2005, \apj, 629, 61.
618:
619: \reference{kon1994}K\"{o}nigl A., \& Kartje J. 1994, \apj, 424, 446.
620:
621: \reference{kra2000b}Kraemer, S.B. \& Crenshaw, D.M. 2000, \apj, 544, 763.
622:
623: \reference{kra2000b}Kraemer, S.B., Crenshaw, D.M., Hutchings, J.B., Gull, T.R.,
624: Kaiser, M.E., Nelson, C.H., \& Weistrop, D. 2000, \apj, 531, 278.
625:
626: \reference{kra2004}Kraemer, S.B., Ferland, G.J, \& Gable, J.R. 2004, \apj, 604,
627: 56.
628:
629: \reference{kra2001}Kraemer, S.B., et al. 2001, \apj, 551, 671.
630:
631: \reference{kra2005} Kraemer, S.B., et al. 2005, \apj, 633, 693 (Paper I).
632:
633: \reference{kra2005} Kraemer, S.B., et al. 2006, \apj, in press
634: (astro-ph/0608383) (Paper II).
635:
636: \reference{kro2001}Krolik, J.H. \& Kriss, G.A. 2001, \apj, 561, 684.
637:
638: \reference{met2006}Metzroth, K.G., Onken, C.A., \& Peterson, B.M. 2006, \apj,
639: 647, 901.
640:
641: \reference{pet1993}Peterson, B.M. 1993, \pasp, 105, 247.
642:
643: \reference{pet1997}Peterson, B.M. 1997, An Introduction to Active Galactic
644: Nuclei (Cambridge, UK: Cambridge University Press).
645:
646: \reference{pet2004}Peterson, B.M., et al. 2004, \apj, 613, 682.
647:
648: \reference{pro2003}Proga, D. 2003, \apj, 585, 406.
649:
650: \reference{sch2006}Schmitt, H.R., Kraemer, S.B., Crenshaw, D.M., \& Hutchings,
651: J.B. 2006, in The X-ray Universe 2005, ed. A. Wilson, ESA SP-604, 661.
652:
653: \reference{sul1999}Sulentic, J.W. \& Marziani, P. 1999, \apj, 518, 9.
654:
655: \reference{wey1997}Weymann, R.J., Morris, S.K., Gray, M.E., \& Hutchings, J.B.
656: 1997, \apj, 483, 717.
657:
658: \end{references}
659:
660: \clearpage
661:
662: \figcaption[f1.eps]{Emission-line profiles from the STIS E140M observation on
663: 2000 June 15 (top four panels) and G430M observation (bottom two panels) on 2000
664: July 2. The solid black line shows the observed flux as a function of radial
665: velocity in the rest frame of NGC~4151 (z $=$ 0.00332). Components of the
666: profiles fits are: continuum + NLR (dotted blue), continuum + BLR (dotted-dashed
667: green), continuum + BLR + ILR (dashed red), and continuum + BLR + ILR + NLR
668: (dashed black). The full extent of absorption component D+E (including doublets)
669: is given below the profiles. The absorption at more negative velocities in the
670: L$\alpha$ profile is due to Galactic L$\alpha$, and the absorption in the cores
671: of the narrow UV resonance lines is due to components F and F$'$ from NGC 4151's
672: ISM and halo. The slight offsets in the apparent velocities of the narrow and
673: intermediate profiles are due to blending of the doublets and/or to
674: different shapes and contributions of the underlying broad and continuum
675: emissions.}
676:
677:
678: \figcaption[f2.eps]{Light curves of the continuum flux at 1350 \AA\ (bottom
679: panel, in units of 10$^{-14}$ ergs s$^{-1}$ cm$^{-2}$ \AA$^{-1}$) and the C~IV
680: emission components (top three panels, in units of 10$^{-12}$ ergs s$^{-1}$
681: cm$^{-2}$) from the STIS E140M spectra.}
682:
683:
684: \figcaption[f3.eps]{The model predictions of ILR C~IV flux as a function of
685: continuum flux at 1350 \AA\, for $N_{H} = 10^{22.46}$ cm$^{-2}$ (dotted line)
686: and $N_{H} = 10^{22.93}$ cm$^{-2}$ (dashed line), compared to the observed
687: fluxes. The details of the scaling of the model results are given in the text.
688: The continuum flux is in units of ergs s$^{-1}$ cm$^{-2}$ \AA$^{-1}$ and the
689: C~IV emission flux is in units of ergs s$^{-1}$ cm$^{-2}$).}
690:
691: \figcaption[f4.eps]{Schematic diagram of a simple model for the geometry and
692: kinematics of the ILR. The line of sight is in the y-z plane, at a polar angle
693: of $\theta = 45$\deg\ from the rotation axis (z) of the accretion disk. Unit
694: vectors for the velocity components are shown at the location of the D+Ea
695: absorber, at a radial distance of 0.1 pc from the inner accretion disk and BLR.
696: The solid red line shows the curve of motion for the D+Ea absorber over 1000
697: years assuming $v_{r} = 490$ km s$^{-1}$, $v_{\phi} = 2100$ km s$^{-1}$, and
698: $v_{\theta} = 0$. The grey curve shows the curve of motion over the same time
699: period for $v_{r} = 490$ km s$^{-1}$, $v_{\phi} = 10,000$ km s$^{-1}$, and
700: $v_{\theta} = 0$. Additional ILR clouds moving in a similar fashion are assumed
701: to cover the BLR at a distance of $\sim$0.1 pc within the cone at $\theta \leq
702: 45$\deg, and possible extending to $\theta \approx 55$\deg.}
703:
704:
705: \clearpage
706:
707: \begin{deluxetable}{lccccc}
708: \tablecolumns{6}
709: \footnotesize
710: \tablecaption{NGC 4151 Absorption Components}
711: \tablewidth{0pt}
712: \tablehead{
713: \colhead{Name} & \colhead{Date} & \colhead{$C_{los}$} & \colhead{log $U$} &
714: \colhead{log $N_H$} & \colhead{log $n_H^{a}$}\\
715: \colhead{} & \colhead{} & \colhead{} & \colhead{} & \colhead{(cm$^{-2}$)} &
716: \colhead{(cm$^{-3}$)}
717: }
718: \startdata
719: X-high & 2002 May & 1.0 & 1.05 & 22.5 & $\geq$5.5 \\
720: D+Ea & 2002 May & 0.9 & $-$0.27 & 22.46 & 6.8 \\
721: & 2000 Mar & 0.9 & $-$0.57 & 22.93 & 6.8 \\
722: D+Eb & 2002 May & 0.8 & $-$1.67 & 20.8 & 8.1 \\
723: D+Ec & 2002 May & 0.5 & $-$1.08 & 21.6 & 7.4 \\
724: D+Ed & 2002 May & 0.2 & $-$3.1 & 19.5 & 9.5 \\
725: & 2001 Apr & 0.7 & $-$4.0 & 19.5 & 9.5 \\
726: \enddata
727: \tablenotetext{a}{From profile fits to the metastable C~III absorption
728: for subcomponents a, b, and c (Paper II). The number density for D+Ed is
729: obtained by assuming co-location with the other subcomponents, and the upper
730: limit for X-high is derived from evidence that it is interior to D+E.}
731: \end{deluxetable}
732:
733: \begin{deluxetable}{lcc}
734: \tablecolumns{3}
735: \footnotesize
736: \tablecaption{STIS E140M Spectra of NGC 4151}
737: \tablewidth{0pt}
738: \tablehead{
739: \colhead{UT Date} & \colhead{Julian Date} & \colhead{F(1350)$^a$}
740: }
741: \startdata
742: 1999 July 19 & 2,451,380 & 9.4 \\
743: 2000 March 3 & 2,451,607 & 4.3 \\
744: 2000 June 15 & 2,451,711 & 2.5 \\
745: 2000 November 5 & 2,451,854 & 17.3 \\
746: 2001 April 14 & 2,452,014 & 1.4 \\
747: 2002 May 8 & 2,452,403 & 11.3 \\
748: \enddata
749: \tablenotetext{a}{Units of 10$^{-14}$ ergs s$^{-1}$ cm$^{-2}$ \AA$^{-1}$.}
750: \end{deluxetable}
751:
752: \begin{deluxetable}{lrrcc}
753: \tablecolumns{5}
754: \footnotesize
755: \tablecaption{NGC 4151 -- ILR Line Ratios
756: (Relative to H$\beta$$^a$)
757: \label{tbl-1}}
758: \tablewidth{0pt}
759: \tablehead{
760: \colhead{Emission Line} & \colhead{Observed}& \colhead{Dereddened$^b$}
761: & \colhead{Model Low-N$^c$} & \colhead{Model High-N$^d$}
762: }
763: \startdata
764: Ly$\alpha$ $\lambda$1216 & 41.7 & 46.9$\pm$13.2 & 42.3 & 39.4 \\
765: N~V $\lambda$1240 & 6.2 & 6.9$\pm$1.9 & 2.9 & 2.7\\
766: C~IV$\lambda$1550 & 37.5 & 40.6$\pm$9.0 & 28.7 & 26.7 \\
767: He~II $\lambda$1640 & 1.4 & 1.5$\pm$0.4 & 1.9 & 1.8\\
768: $[$O~III] $\lambda$5007 & 4.5 & 4.5$\pm$0.8 & 5.4 & 5.1
769: \enddata
770: \tablenotetext{a}{Observed flux of ILR H$\beta$ $=$ 1.89 ($\pm$0.32) $\times$
771: 10$^{-13}$ ergs s$^{-1}$ cm$^{-2}$}
772: \tablenotetext{b}{From E$_{B-V}$ $=$ 0.02 and the standard Galactic reddening
773: curve of Savage \& Mathis (1979).}
774: \tablenotetext{c}{Model parameters: $U = 10^{-0.92}, N_{H} = 10^{22.46}$
775: cm$^{-2}$, and $n_{H} =
776: 10^{6.75}$ cm$^{-3}$.}
777: \tablenotetext{d}{Model parameters: $U = 10^{-0.92}, N_{H} = 10^{22.93}$
778: cm$^{-2}$, and $n_{H} =
779: 10^{6.75}$ cm$^{-3}$.}
780: \end{deluxetable}
781:
782: %\end{document}
783:
784: \clearpage
785: \begin{figure}
786: \plotone{f1.eps}
787: \\Fig.~1
788: \end{figure}
789:
790: \clearpage
791: \begin{figure}
792: \plotone{f2.eps}
793: \epsscale{1.0}
794: \\Fig.~2
795: \end{figure}
796:
797: \clearpage
798: \begin{figure}
799: \plotone{f3.eps}
800: \\Fig.~3
801: \end{figure}
802:
803: \clearpage
804: \begin{figure}
805: \plotone{f4.eps}
806: \\Fig.~4
807: \end{figure}
808:
809: \end{document}
810: