astro-ph0108521/ms.tex
1: \documentstyle[12pt,aasms4]{article}
2: 
3: \def\arcsecpoint{$''\!.$}
4: \def\deg{$^{\rm o}$}
5: 
6: \received{}
7: \accepted{}
8: %\journalid{}{}
9: %\articleid{}{}
10: 
11: \lefthead{Crenshaw et al.}
12: \righthead{Kinematics in NGC 4151}
13: 
14: \begin{document}
15: 
16: \title{Kinematics of the Narrow-Line Region in the Seyfert 2 Galaxy Mrk 3\altaffilmark{1}}
17: \author{Jose R. Ruiz\altaffilmark{2},
18: D.M. Crenshaw\altaffilmark{2,3},
19: S.B. Kraemer\altaffilmark{2,3},
20: G.A. Bower\altaffilmark{4},
21: T.R. Gull\altaffilmark{3},
22: J.B. Hutchings\altaffilmark{5},
23: M.E. Kaiser\altaffilmark{6},
24: \& D. Weistrop\altaffilmark{7}}
25: 
26: \altaffiltext{1}{Based on observations made with the NASA/ESA Hubble Space 
27: Telescope. STScI is operated by the Association of Universities for Research in 
28: Astronomy, Inc. under NASA contract NAS5-26555.}
29: \altaffiltext{2}{Catholic University of America}
30: \altaffiltext{3}{Laboratory for Astronomy and Solar Physics, NASA's Goddard
31: Space Flight Center, Code 681, Greenbelt, MD 20771}
32: \altaffiltext{4}{NOAO, 950 N. Cherry Street, Tucson, AZ 85726-6732}
33: \altaffiltext{5}{Dominion Astrophysical Observatory, National Research Council of
34: Canada, 5071 West Saanich Rd., Victoria, BC V8X 4M6, Canada}
35: \altaffiltext{6}{Department of Physics \& Astronomy, Johns Hopkins University,
36: 3400 North Charles St., Baltimore, MD 21218}
37: \altaffiltext{7}{Department of Physics, University of Nevada at Las Vegas, 
38: 4505 South Maryland Parkway, Las Vegas, NV 89154-4002}
39: 
40: \begin{abstract}
41: We present measurements of radial velocities for the narrow-line region (NLR) 
42: gas in the Seyfert 2 galaxy Mrk 3 out to $\sim$1 kpc from the nucleus.  The
43: observations consist of two datasets, both using the Space Telescope Imaging 
44: Spectrograph (STIS) on board the $Hubble$ $Space$ $Telescope$ $(HST)$:
45: 1) An [O~III] slitless spectrum with the G430M grating of the inner 3$''$ 
46: around the nucleus, and 2) a long-slit observation centered on the nucleus 
47: (PA = 71\deg) using the G430L grating and the 52$''$ x 0$''$.1 aperture.  
48: Our results produce radial velocity maps of the emission-line gas.  These maps
49: indicate general trends in the gas motion, which include: blueshifts and 
50: redshifts on either side of the nucleus, steep velocity rises from systemic up
51: to $\sim$$\pm$700 km s$^{-1}$ taking place in the inner 0$''$.3 (0.8 kpc) both
52: east and west of the nucleus, gradual velocity descents back to near-systemic
53: values from 0$''$.3-1$''$.0, slightly uneven velocity amplitudes on each
54: side of the nucleus, and narrow velocity ranges over the entire observed region.
55: 
56: When fitted to kinematic modeling programs for the NLR gas, the data clearly favor
57: a model where the gas exists in a partially filled bicone, is accelerated 
58: radially away from the nucleus, and is followed by a constant deceleration
59: (possibly due to collision with an ambient medium).  This geometry and general 
60: kinematic model is in agreement with previous work done on the NLR gas of NGC
61: 1068 and NGC 4151.  On scales of hundreds of parsecs, we conclude that radial 
62: outflow may be a common feature of Seyfert galaxies.
63: \end{abstract}
64: 
65: \keywords{galaxies:individual (Mrk 3) -- galaxies: Seyfert}
66: 
67: \section{Introduction}
68: In many Seyfert 2s, the clouds appear to lie in a biconical or roughly linear 
69: configuration surrounding the nucleus (Schmitt \& Kinney 1996, etc).  Various 
70: kinematic models have been proposed to explain NLR cloud motion.  Capetti et al. 
71: (1995) have compared optical and radio measurements of the NLR of Mrk 3 and 
72: concluded that the NLR clouds are the result of radio jet plasma expanding away 
73: from the bicone axis. Winge et al. (1997), (1999) postulate gravitational motions 
74: for the NLR in NGC 4151.  Recently, Crenshaw \& Kraemer (2000) and Kaiser et al. 
75: (2000) have determined radial velocities as a function of position in the NLRs 
76: of NGC 1068 and NGC 4151 (the brightest Seyfert 2 and Seyfert 1, respectively) 
77: with the STIS on $HST$.  Crenshaw et al. (2000) have proposed a model where clouds 
78: on the surface of a bicone are radially accelerated from the nucleus by wind
79: pressure or radiation pressure, encounter and collide with an ambient medium, 
80: then decelerate to near-systemic values.  It explains the general trends seen 
81: in the radial velocity as a function of position in the inner kiloparsec around 
82: the nuclei of these galaxies.
83: 
84: Mrk 3 is a well-studied Seyfert 2 galaxy, which shows evidence for a hidden
85: Seyfert 1 nucleus from broad polarized emission lines (Schmidt \& Miller 1985). 
86: The host galaxy is classified as an elliptical or S0 galaxy type.  It lies 53 
87: Mpc away (H$_{0}$=75 km s$^{-1}$ Mpc$^{-1}$, z$=$0.0135, 3.82$''$ kpc$^{-1}$). It 
88: has been studied in every wavelength regime, from the X-ray (Griffiths, 1998; 
89: Georgantopoulos et al, 1999) through the radio (Kukula et al. 1999).  Mrk 3 has 
90: bright [O III] emission-line clouds that lie in a biconical configuration (apex 
91: of the two cones coincident with the nucleus) along PA$=$80\deg, with a 
92: half-opening angle of 22.5\deg\ (Schmitt \& Kinney 1996).  At the end of the 
93: western cone, a large, diffuse knot appears, while on the end of the eastern cone, 
94: a bright knot appears out of the bicone, giving the entire structure an `S' shape 
95: (Kukula et al. 1993).  Schmitt \& Kinney measure the clouds as extending 280 pcs 
96: on either side of the nucleus.  Recent X-ray observations of Mrk 3 using 
97: $Chandra$ (Sako et al. 2000) have found soft X-ray extended emission which lies 
98: along this PA.  Radio jets have also been observed (Axon et al. 1998, 
99: Kukula et al. 1999) along the same PA; these appear to follow the biconical structure, 
100: although the half-opening angle is far less ($\sim$~8-10\deg).  The jets also have 
101: an `S' shape (though much less pronounced) in the same regions as the [O~III] emission.  
102: Though they lie close to the emission-line clouds, they are not exactly coincident; 
103: Axon et al. (1998) suggest that they lie along the convex edge of the S-shaped curvature.  
104: 
105: A fainter set of clouds, situated along P.A.$\sim$100\deg, are seen further 
106: out in the ENLR.  These clouds extend from 1.$''$0 on either side of the nucleus 
107: to about 3$''$.2.  This group is more diffuse, fainter in surface brightness
108: by an order of magnitude, and follows the [O~III] emission contour lines seen in 
109: the ground-based observations of Pogge \& De Robertis (1993).  
110: 
111: In this paper, we present two $HST$ spectra (one long-slit, one slitless) of the 
112: Mrk~3 NLR clouds.  They provide consistent values of radial velocities as functions 
113: of NLR cloud positions.  The two datasets are then fitted to a kinematic modeling 
114: program that provides a radial velocity map of clouds within a bicone, given a 
115: velocity law that directs their motion.  When fitted to the radial acceleration + 
116: constant deceleration velocity law, the trends seen in both datasets are matched well.  
117: In \S 2 we detail the observations, while the data analysis is presented in \S 3.  
118: The results from the two datasets are given in \S 4.  The discussion of the model 
119: is given in \S 5, with \S 6 providing the overall discussion.  \S 7 presents the 
120: conclusions.
121: 
122: \section{Observations}
123: In order to find the NLR knots in the slitless observations, and hence, determine
124: their velocities, a companion [O III] image, as well as a continuum image are 
125: required.  To that end, the archival images of Mrk 3 were obtained.  The first of 
126: these is a Wide Field Planetary Camera (WFPC) [O III] observation.  This image 
127: served to match the bright NLR clouds.  A Wide Field Planetary Camera 2 (WFPC 2) 
128: continuum image was also retrieved, as well as a Faint Object Camera (FOC) [O III] 
129: image.  The FOC image is shown in Figure 1.  It served to match the faint ENLR 
130: clouds.  A summary of all the observations is given in Table 1.
131: 
132: The new slitless observations take advantage of STIS's spatial resolution (0$''$.1) 
133: and the G430M spectral resolution ($\lambda$/$\Delta$$\lambda$$\approx$10,000).  
134: The observations were centered at 5093 \AA, with a bandwidth of 286 \AA.  In addition 
135: to [O III] $\lambda$5007, [O III] $\lambda$4959 and H$\beta$ were also observed.  
136: The spectral region around [O III] $\lambda$5007 is shown in Fig. 2.  The horizontal 
137: axis is the dispersion axis, while the vertical axis is the spatial axis, as indicated 
138: by the scale.  The bright NLR clouds are smeared out along the dispersion axis, 
139: indicating large velocity dispersions.  They are also shifted along this axis, from 
140: which their radial velocities can be calculated.  The faint ENLR clouds are also 
141: seen with apparently smaller dispersions. 
142: 
143: We also obtained long-slit observations of Mrk 3 using the G140L, G230L, G430L 
144: and G750L gratings and the 52$''$ x 0$''$.1 aperture.  The complete observations 
145: covered the wavelength range from 1150-10,000 \AA.  A full analysis of these 
146: data will be presented separately (Collins et al. 2001, in preparation).  
147: For this work, the single emission line of [O~III] $\lambda$5007 from these 
148: observations was used.  The slit had a PA of 71\deg, and was chosen to pass 
149: through the nucleus.  The position of the long-slit is shown in Fig. 1, overlaid 
150: on the FOC image for comparison with the [O III] clouds.  
151: 
152: \section{Data Analysis}
153: The data reduction was done with IDL software developed for the STIS Instrument 
154: Definition Team at Goddard Space Flight Center (Lindler et al. 1999).  The spectral 
155: images were cleaned of cosmic rays during the image processing.  Once the archival 
156: images were retrieved, they were also cleaned of cosmic rays.  Determining radial 
157: velocities of NLR clouds using slitless spectra has been done previously using 
158: NGC 4151 (Hutchings et al. 1998, Kaiser et al. 2000) and NGC 1068 (Crenshaw et al. 2000).  
159: These authors describe the STIS slitless data analysis; a brief synopsis follows here.
160: 
161: The data analysis consists of matching [O III] undispersed clouds (from an FOC,
162: WFPC or WFPC2 image) with their counterparts in a STIS dispersed image.  In order 
163: to make the one-to-one correspondence between clouds, the non-STIS images must be 
164: rotated, aligned, and corrected for anamorphic magnification with respect to the
165: STIS image (spatial scale of 0$''$.0507 pixel$^{-1}$).  Once the spatial scale
166: is set, then the wavelength scale (ultimately a velocity scale) must be set using 
167: the calibration spectra.  To set the velocity scale, the STIS dispersed image is
168: aligned with the direct [O III] images such that there is no shift for a cloud at
169: the systemic redshift (z=0.0135, Tifft \& Cooke 1988) of Mrk 3.  The redshifted 
170: wavelength is taken from 21 cm measurements of H I, assumed to be at rest with 
171: respect to the central galactic region.  We note that the WFPC [O III] observation 
172: and WFPC2 continuum image were taken at different times.  Despite this, we were able 
173: to identify the bright, distinct emission line knots in both images.
174: 
175: Once the STIS dispersed image and the non-STIS undispersed image are aligned, 
176: Gaussian fits are made to the NLR knots in both images on each row parallel 
177: to the dispersion axis.  This is to accurately determine the positions and 
178: FWHMs for each knot.  The difference in position (in pixels) between an 
179: undispersed and dispersed knot is converted to a difference in \AA, then to
180: one in velocity.  The difference in width (in quadrature) gives the velocity 
181: dispersion of the emission knot.  For some purposes, the individual line 
182: measurements are averaged over each cloud in order to obtain a single value of 
183: radial velocity and dispersion, and the standard deviations are used as 
184: uncertainties. 
185: 
186: For the long-slit data, Gaussian fits were made to the NLR knots on each row 
187: perpendicular to the dispersion (in 0$''$.05 intervals).  Once the positions and 
188: FWHMs were calculated, the radial velocities and dispersions for each row could 
189: be calculated using the same fitting routines as for the slitless spectrum.  The 
190: FWHMs were corrected for the instrumental broadening of 300 km s$^{-1}$.  For the 
191: purpose of comparing slitless and long slit data, we determined which clouds lay 
192: in the slit from the PA of the slit.  We then averaged the long-slit radial 
193: velocities for these clouds over each individual cloud, thus obtaining two values 
194: of radial velocity from the two different methods and plotted them against each 
195: other.  For both sets of radial velocities, the standard deviations calculated 
196: were used for the uncertainties. Figure 3 shows averaged slitless radial velocities 
197: plotted against the corresponding averaged long-slit radial velocities.  The plot 
198: shows very good agreement, to within $\sim$50 km s$^{-1}$, indicating that there 
199: are no systematic errors with the slitless technique.   
200: 
201: \section{Results}
202: The entire dataset is seen in Figure 4, which shows a plot of radial velocity 
203: versus distance.  Each point represents an average over each cloud.  The inner 
204: region clouds are seen to have a wide range of radial velocities, from $-$1000 
205: km s$^{-1}$ to +600 km s$^{-1}$.  There are roughly equal numbers of clouds that 
206: show redshifts and blueshifts.  As noted in the previous section, the long-slit 
207: cloud values are consistent with the slitless spectral values, both in terms of 
208: location and velocity.  The fainter, ENLR clouds stretch for about 1$''$.5 
209: further out in either direction than the inner region clouds.  The fainter 
210: clouds show a narrower range of radial velocity, from +400 km s$^{-1}$ down to 
211: $-$200 km s$^{-1}$, with the majority of the clouds (75\%) showing redshifts.
212: 
213: Figure 5 shows the velocity distribution of all the clouds.  The NLR clouds are 
214: difficult to distinguish due to their proximity to each other, so we present this 
215: figure in order to show the ENLR clouds.  The eastern half of these clouds is 
216: exclusively redshifted, while the majority of the western half is so with lower 
217: magnitudes.  Figure 6 is an expanded view of the central region of Figure 5 to 
218: show the inner region clouds' velocity distribution in greater detail.  The inner 
219: region clouds, in contrast to the extended region clouds, have equal numbers of 
220: blueshifted and redshifted clouds on either side of the nucleus.  The blueshifted 
221: clouds to the east tend to have greater velocities, while to the west the velocities 
222: are similar for redshifts and blueshifts.  On either side, maximum values of 
223: radial velocities are reached $\sim$0$''$.3 away from the center for both sets 
224: of data.
225: 
226: Figure 7 shows the unbinned NLR radial velocities (relative to systemic) plotted 
227: against distance from the optical continuum center.  The unbinned velocities are
228: used for the model in order to take advantage of $HST's$ spatial resolution.  
229: Long-slit points are shown, along with the slitless spectral points that 
230: lie within the slit.  This set of points is used because the modeling program we
231: use (to be discussed in the next section) simulates a long-slit, so we choose only
232: those points within the slit.  The velocity errors are only measurement errors 
233: in this case, and are comparable to the size of the symbols.  The two sets of data 
234: points are seen to be compatible.  They show properties that must be duplicated 
235: by any model fit.  Now we briefly discuss each of these properties. 
236: 
237: Both sets of data points show fast rises in velocities (from systemic values at
238: the nucleus to $\sim$$\pm$600 km s$^{-1}$) out to $\sim$0$''$.25 of the nucleus 
239: on each side.  The climbs in velocities are seen in both redshifts and blueshifts. 
240: All the rises are then followed by shallower velocity downturns, so that at 
241: $\sim$0$''$.7-1$''$.0, the velocities have returned to near systemic values.  
242: The amplitudes of the maximum velocities are not equal.  The amplitudes range
243: from $\sim$300 km s$^{-1}$ on the blueshifted west side to $\sim$800 km s$^{-1}$ 
244: on the blueshifted east side (ignoring a few high-velocity points).  This variation 
245: in amplitude implies that any fitted cone is tilted, and in fact, the angle of 
246: inclination can be calculated by the amplitude difference.  Finally, the range of 
247: velocities is fairly narrow.  For example, $\sim$0$''$.30 west of the nucleus, 
248: blueshifted velocities are seen exclusively from -200 to -300 km s$^{-1}$, while 
249: redshifted velocities are seen spanning a narrow range from 300 to 500 km s$^{-1}$.  
250: At this distance, there are no velocities seen from -200 to 300 km s$^{-1}$. 
251: Any model fit must be able to match these narrow velocity ranges.  
252: 
253: \section{Model Fitting}
254: Once the spatial orientations and radial velocities of the clouds were obtained,
255: we attempted to fit these observations using kinematic modeling programs.  These 
256: programs calculate the radial velocities and spectral lines of material on the
257: surface of a thin disk, or a bicone, either filled or hollow.  Each geometry can 
258: assume one of various velocity laws that control the material's movement.  We 
259: concentrate here on the inner NLR out to $\sim$1$''$.0 on either side of the 
260: nucleus.  Later, we will discuss briefly the ENLR clouds and their motion.
261: 
262: We can immediately rule out gravitational rotation models by calculating 
263: the mass required to impart radial velocities on the order of 
264: 500-1000 km s$^{-1}$ at a distance of 100-200 pcs away from the nucleus.
265: This mass is of the order of 10$^{9-10}$ M$_\odot$.  Typical masses for 
266: black holes in Seyfert nuclei are 10$^{6-8}$ M$_\odot$ (Peterson \& Wandel 2000).  
267: Observationally, the observed morphology of the NLR does not suggest a disk 
268: geometry, while the redshifts and blueshifts on either side of the nucleus 
269: cannot be the result of simple Keplerian rotation.  Thus we are left with outflow 
270: models, or models where material flows tangentially outward from the radio axis. 
271: 
272: The bicone program has been used previously to model the NLR emission-line 
273: clouds of NGC 1068 and NGC 4151 (Crenshaw \& Kraemer 2000, Crenshaw et al. 2000).
274: The two cones (one on either side of the nucleus) are assumed to possess identical 
275: properties, including geometry, size and velocity law.  In addition, the cones are 
276: assumed to have a filling factor of 1 within the minimum and maximum half opening
277: angle, and not to absorb [O~III] photons.  We adjust certain parameters, shown in 
278: Table 2, to obtain the best fit.  The program creates a two-dimensional velocity map, 
279: which is sampled through a simulated slit.  We applied the models to unbinned 
280: velocities in order to obtain the best spatial resolution.
281: 
282: Several model input variables can be constrained from the observations.  The first 
283: of these was the extent of the NLR.  Based on the approximate placing of a bicone
284: on the Mrk 3 NLR by Schmitt \& Kinney (1996), we measured its maximum extent as 
285: $\sim$0.75$''$.  In addition to a minimum and maximum distance (in pixels) of the 
286: cones, the program requires a minimum and maximum half-opening angle.  The maximum 
287: angle was measured from the images, giving a value of 25\deg.  This agrees with 
288: Schmitt \& Kinney (1996), who measure a maximum half-opening angle of 22.5\deg.  
289: The minimum half-opening angle is not visible in the [O III] images, so it was 
290: varied to match the data.  The optimum value for the models was 15\deg.  This value 
291: places the emission-line material outside the observed radio jet cone (half-opening
292: angle $\sim$7-8\deg) (Capetti et al. 1995).  
293: 
294: The inclination angle was calculated based on the differences between the radial 
295: velocity maxima on the W and E sides of the cones.  The maximum blueshifts are 
296: higher by $\sim$300 km s$^{-1}$ than the maximum redshifts on the east side.  
297: The NLR inclination angle was then calculated as $\sim$5-10\deg, using simple 
298: trigonometry.  Finally, the value for the maximum deprojected velocity of the NLR 
299: gas was chosen so that it would match the observed NLR radial velocity peak 
300: ($\sim$-800 km s$^{-1}$).  The best fit parameters of all the models are shown 
301: in Table 3.  The results of the models are summarized below.
302: 
303: 1. The radial acceleration (RA) outflow model consists of NLR clouds being driven 
304: away from the nucleus, perhaps by winds or jets.  The acceleration is along 
305: the bicones' entire length.  The best fit was not able to match the high 
306: velocities near the center, given the observed parameters.  The only way to
307: marginally match these velocities was to widen the half-opening angle 
308: past $\sim$40\deg, but the sharp downturns cannot then be fit.  It is clear from
309: the images that the ionization cone's half-opening angle cannot be more than 
310: $\sim$30\deg.  If there is acceleration along the bicone, it cannot take place
311: along the entire length of the NLR.
312: 
313: 2. The constant velocity (CV) model consists of clouds with a negligible drag 
314: force, having been accelerated out to some distance (small compared to the NLR),
315: then proceeding with constant velocity.  This model is able to match the high central 
316: velocities $\sim$0$''$.3 from the nucleus.  Further out, however, the modeled 
317: velocities remain at a constant value out to the ENLR, whereas the observed 
318: velocities drop to near systemic values by $\sim$1$''$.0 out from the center.
319: 
320: 3. The constant tangential (CT) model consists of NLR clouds moving radially away 
321: from the central radio axis.  This would be seen if the radio plasma expanded within 
322: the emission-line bicone.  This fit resembles the CV model, except that it predicts
323: equal magnitude redshifts and blueshifts on either side of the nucleus.  This is 
324: certainly not the case, as seen in the two datasets.  Note that this model predicts 
325: velocity magnitudes substantially less (1/3 to 1/6) than the other models (see Table 3).  
326: These velocities appear to be too low.  There are a number of other inconsistencies with
327: this model, which are discussed in \S 6. 
328: 
329: 4. The model that fit the most data points is the radial acceleration + constant 
330: deceleration (RA+CD).  The model can be visualized as material first accelerated by 
331: wind or radiation pressure from the nucleus, which then impacts an ambient medium 
332: and then decelerates at a constant rate.  This model implies that the emission-line 
333: clouds originate from a region closer to the nucleus and move outward from there.
334: Figure 8 shows the long-slit and slitless data points overlaid with the shading from 
335: this model.  Obviously, this model does not perfectly fit every point, but it fits 
336: the gross features of the observations well.  Many of the discrepant points can be 
337: ascribed to slightly different acceleration or deceleration laws in different quadrants.  
338: The discrepant high-velocity points suggest clouds that perhaps do not encounter the 
339: ambient, possibly patchy medium, or encounter it in a region of lower density and 
340: do not decelerate as much.  
341:   		                                             
342: \section{Discussion} 
343: The slitless spectral method of determining radial velocity gives consistent 
344: values with the long slit method, as shown in Figures 3 and 7.  This result 
345: gives confidence in future work using the slitless method, and has been shown 
346: before for NGC 4151 (Hutchings et al. 1998, Crenshaw et al. 2000).  The best fit 
347: RA+CD model shows, in addition to fitting all of the trends seen in the data, 
348: some discrepancies.  These can best be explained by slightly different
349: acceleration/deceleration laws in different directions. 
350: 
351: While the NLR clouds are fitted most closely by the RA+CD model, the ENLR 
352: clouds require a different model.  Those clouds (with velocities 
353: $\lesssim$ 350 km s$^{-1}$) appear to be influenced by the gravitational 
354: potential of the supermassive black hole (SMBH) and inner galaxy, rather than 
355: the outflowing material.  This hypothesis agrees with surface photometry on Mrk 3 
356: done by one of us (Bower).  Ellipsoids were fit to the 
357: surface brightness of Mrk 3, from 0.01$''$ out to 100$''$.  From these 
358: ellipsoid fits, a spherical dynamical model was used to predict a rough upper 
359: limit on the rotational radial velocity induced by the gravitational potential. 
360: For the range from 1$''$ to $\sim$5$''$, where the ENLR clouds reside, the 
361: projected radial velocities are predicted to be $\lesssim$200 km s$^{-1}$.  
362: This heuristic result agrees roughly with the observed ENLR cloud velocities, 
363: although we cannot explain the preponderance of redshifted clouds with this model. 
364: They may be due to a lack of ionized gas at the positions that would produce 
365: blueshifts. 
366: 
367: The orientation of the host galaxy has been previously reported as 27\deg\ out of
368: the plane of the sky (Schmitt \& Kinney 1996).  If this orientation extends down to 
369: kiloparsec scales, then the plane of the galactic disk would lie within the angular
370: range of one side of each cone (15\deg\ to 25\deg, tilted out 5\deg\ of the plane of 
371: the sky).  The situation then resembles NGC 1068 and NGC 4151 (Crenshaw et al. 2000,
372: Crenshaw \& Kraemer 2000), which also seem to have the galactic disk and one side 
373: of the bicone in the same plane.  Crenshaw et al. (2000) propose that the galactic
374: disk's ionization (by the nucleus) contributes to ENLR gas.  We propose that the
375: same geometry exists in this galaxy.  
376:  
377: The radio jet and the NLR emission share a similar axis, and are nearly coincident.
378: However, other than their spatial coincidence, there do not appear to be any other
379: correlations, as would be expected if the radio plasma's expansion were the source 
380: of the NLR velocities.  Firstly, in the data itself, there are no bright NLR clouds
381: that correspond to jet flux maxima.  This lack of correspondence has been noted
382: in other objects (NGC 4151, Kaiser et al. 2000). In addition, there is no evidence for
383: peculiar velocities at the positions of the radio lobes.  In terms of the dataset
384: velocities, there is no physical reason given by this model to explain the velocity 
385: trends (increasing to some turnover distance, then steadily decreasing) that we see 
386: in the data.  In terms of the modeling, the CT and other transverse velocity models,
387: predict equal blueshift/redshift amplitudes no matter what inclination angle the 
388: bicone is tilted.  The data show a definite difference (200-400 km s$^{-1}$) 
389: in velocity maxima between redshifts and blueshifts, consistent with a biconical 
390: geometry.  
391:      
392: \section{Conclusions}
393: Two STIS spectra were obtained of the NLR of the Seyfert 2 galaxy Mrk 3.  Radial 
394: velocities were determined of the emission-line gas as a function of position (out
395: to $\sim$1 kpc from the nucleus).  The velocity maps indicate general trends in the
396: gas motion.  These include: blueshifts and redshifts on either side of the nucleus,
397: steep velocity rises from systemic up to $\sim$$\pm$700 km s$^{-1}$ taking place
398: in the inner 0$''$.3 (0.8 kpc) both east and west of the nucleus, and gradual velocity
399: descents back to near-systemic values from 0$''$.3-1$''$.0.  
400: 
401: The data were then fitted to kinematic modeling programs for the NLR gas on the
402: surface of the bicone.  The data sets were fit best with a radial acceleration + 
403: constant deceleration model.  In the model, the cones extend out to a radius of 
404: 0$''$.75 from the nucleus, with a half-opening angle between 15\deg\ and 25\deg.  
405: The modeled material reaches a maximum deprojected velocity of 1750 km s$^{-1}$, 
406: reaching this velocity at a distance of 0$''$.3-0$''$.43 from the nucleus, close 
407: to the observed distance of 0$''$.2-0$''$.3 from the nucleus.  The fit could be 
408: improved by positing different turnover radii and/or acceleration/deceleration laws 
409: for each quadrant.  Also, the high velocity data points not fit by the model appear 
410: to be clouds that do not encounter any dense medium and maintained their acceleration.  
411: Nevertheless, our goal of being able to explain all the basic trends in the data 
412: with a simple model was accomplished.  We have ruled out gravitational and constant 
413: velocity models.  We show that a model where the NLR emission is produced by expansion 
414: of radio jet plasma away from the radio axis does not fit the data well.
415: 
416: An important observational result is that the two distinct methods of obtaining 
417: radial velocities each gave similar results.  This has been shown previously for
418: NGC 4151 (Hutchings et al. 1998, Kaiser et al. 2000).  The slitless spectral procedure 
419: of obtaining radial velocities has proven to be a useful and efficient tool for 
420: quickly examining and mapping nearby galaxies with clumpy NLRs and ENLRs.  We will 
421: take advantage of this technique in the future to map the kinematics of the NLR in 
422: nine other Seyfert galaxies.
423: 
424: This work was supported by NASA Guaranteed Time Observer funding to the STIS 
425: Science Team under NASA grant NAG 5-4103 and by {\it HST}. 
426: Additional support for this work was provided by NASA through grant number 
427: HST-GO-08340.01-A from the Space Telescope Science Institute, which is operated 
428: by AURA, Inc., under NASA contract NAS5-26555. 
429: \begin{references}
430: 
431: \reference{axo1998}Axon, D.J., Marconi, A., Capetti, A., Macchetto, F.D.,
432: Schreier, E., \& Robinson, A. 1998, \apj, 496, L75
433: 
434: \reference{bic1998}Bicknell, G.V., Dopita, M.A., Tsvetanov, Z.I., 
435: Sutherland, R.S. 1998, \apj, 495, 680
436: 
437: \reference{cap1995}Capetti, A., Macchetto, F., Axon, D.J., Sparks, W.B.,
438: \& Boksenberg, A. 1995, \apj, 448, 600
439: 
440: \reference{col2001}Collins, N.R., et al., 2001, in preparation
441: 
442: \reference{cre2000}Crenshaw, D.M., Kraemer, S.B. 2000, \apj, 532, L101
443: 
444: \reference{cre2000}Crenshaw, D.M., Kraemer, S.B., Hutchings, J.B., Bradley, 
445: L.D., II, Gull, T.R., Kaiser, M.E., Nelson, C.H., Ruiz, J.R., Weistrop, D. 2000,
446: \aj, 120, 1731
447: 
448: \reference{geo1999}Georgantopoulos, I., Papadakis, I., Warwick, R.S., Smith, D.A.,
449: Stewart, G.C., \& Griffiths, R.G. 1999, \mnras, 307, 815
450: 
451: \reference{gri1998}Griffiths, R.G., Warwick, R.S., Georgantopoulos, I., Done, C., 
452: \& Smith, D.A. 1998, \mnras, 298, 1159
453: 
454: \reference{hut1998}Hutchings, J.B., Crenshaw, D.M., Kaiser, M.E., Kraemer, S.B.,
455: Weistrop, D., Baum, S., Bowers, C.W., Feinberg, L.D., Green, R.F., Gull, T.R., 
456: Hartig, G.F., Hill, G., and Lindler, D.J. 1998, \apj, 492, L115
457: 
458: \reference{kai2000}Kaiser, M.E., Bradley, L.D. II, Hutchings, J.B., Crenshaw, D.M.,
459: Gull, T.R., Kraemer, S.B., Nelson, C.H., Ruiz, J., and Weistrop, D. 2000, \apj,
460: 528, 260 
461: 
462: \reference{kuk1993}Kukula, M.J., Ghosh, T., Pedlar, A., Schilizzi, R.T., 
463: Miley, G.K., de Bruyn, A.G., \& Saikia, D.J. 1993, \mnras, 264, 893
464: 
465: \reference{kuk1999}Kukula, M.J., Ghosh, T., Pedlar, A., \& Schilizzi, R.T. 1999,
466: \apj, 518, 117
467: 
468: \reference{lin1999}Lindler, D. 1999, CALSTIS Ref. Guide (version 6.4; Greenbelt:
469: NASA/GSFC)
470: 
471: \reference{nel2000}Nelson, C.H., Weistrop, D., Hutchings, J.B., Crenshaw, D.M.,
472: Gull, T.R., Kaiser, M.E., Kraemer, S.B., and Lindler, D. 2000, \apj, 531, 257
473: 
474: \reference{pet2000}Peterson, B.M., Wandel, A. 2000, \apj, 540, L13
475: 
476: \reference{pog1993}Pogge, R.W., \& De Robertis, M.M. 1993, \apj, 404, 563
477: 
478: \reference{sak2000}Sako, M., Kahn, S.M., Paerels, F., Liedahl, D.A. 2000, \apj, 
479: 543, L115
480: 
481: \reference{sch1985}Schmidt, G.D., \& Miller, J.S. 1985, \apj, 290, 517
482: 
483: \reference{sch1996}Schmitt, H.R., \& Kinney, A.L. 1996, \apj, 463, 498
484: 
485: \reference{tif1988}Tifft, W.G., \& Cocke, W.J. 1988, \apjs, 67, 1
486: 
487: \reference{win1997}Winge, C., Axon, D.J., Macchetto, F.D., \& Capetti, A. 1997,
488: \apj, 487, L121
489: 
490: \reference{win1999}Winge, C., Axon, D.J., Macchetto, F.D., Capetti, A., \&
491: Marconi, A. 1999, \apj, 519, 134
492: 
493: \end{references}
494: 
495: \clearpage
496: \begin{deluxetable}{llllll}
497: \tablecolumns{6}
498: \tablecaption{Observations \label{tbl-1}}
499: \tablewidth{0pt}
500: \tablehead{\colhead{Date\tablenotemark{a}} &\colhead{Instrument} 
501: &  \colhead{Root Name} &\colhead{Filter/Grating} & \colhead{Exposure (s)} 
502: & \colhead{Slit}} 
503: \startdata
504: 1991 July 18 (A)  &	PC	&W0MW0601T      &F502N   &1800	&  \\
505: 1992 Dec 11  (A)  &	FOC	&X14W0301T      &F501N	 &1197	&  \\
506: 1997 Oct 20  (A)  &	WFPC2	&U2E62A01T      &F606W	 &500	&  \\
507: 2000 Jan 16  (N)  &	STIS	&O5F403010	&clear	 &20	&  open\\
508: 2000 Jan 16  (N)  &	STIS	&O5F403020	&G430M   &2154	&  open\\
509: 2000 Aug 22  (N)  &	STIS	&O5KS01010	&G430L	 &1080	& 52$''$x0$''$.1 \\
510: \enddata
511: \tablenotetext{a}{A-Archival, N-New}
512: \end{deluxetable}
513: 
514: \clearpage
515: \begin{deluxetable}{ll}
516: \tablecolumns{2}
517: \footnotesize
518: \tablecaption{Modeling Parameters \label{tbl-2}}
519: \tablewidth{0pt}
520: \tablehead{\colhead{Parameter}  &   \colhead{Symbol or constant value}}
521: \startdata
522: Min \& Max distance of cones (pcs)  &   Min. = 0 , Max = D    \\
523: Min \& Max half-opening angle  &   $\theta$$_{min}$, $\theta$$_{max}$  \\
524: Inclination angle	       &   $i$ \\
525: Deprojected Maximum velocity 
526: of NLR gas (km s$^{-1}$)   &   V$_{max}$	 \\
527: Velocity Laws		       &   \it Constant velocity (CV)   \\  
528: 			       &   \it Radial Acceleration (RA)   \\
529: 		 &   \it Radial Acceleration plus Constant Deceleration (RA+CD)  \\
530: 			       &   \it Constant Tangential Flow (CT)   \\
531: 			       &   \it Gravitational Infall (GI)   \\
532: Center of Slit 		       &   Centered on optical continuum peak  \\
533: Position angle of the long slit	 &    $\sim$71\deg    \\
534: Slit Width (in pixels)	       &   0$''$.1   \\
535: \enddata
536: \end{deluxetable}
537: 
538: \clearpage
539: \begin{deluxetable}{lllll}
540: \tablecolumns{5}
541: \footnotesize
542: \tablecaption{Parameters of each best fit model}
543: \tablewidth{0pt}
544: \tablehead{\colhead{Parameter} &  \colhead{CV} &  \colhead{RA}  &  \colhead{RA+CD} & 
545: \colhead{CT}}
546: \startdata
547: D (pcs)	&   80	&   80    &     80	&   80	\\
548: $\theta$$_{min}$, $\theta$$_{max}$  &  15, 25  &   15, 25   &   15, 25	 &  15, 25\\
549: $i$	&   5\deg  &  5\deg   &  5\deg   &   5\deg	\\
550: V$_{max}$ (km s$^{-1}$) &  1400	&   3000   &   1750   &  550  \\
551: \enddata
552: \end{deluxetable}
553: 
554: \clearpage
555: \figcaption[ruiz.fig1.ps]{FOC image of the bright NLR clouds of Mrk 3.  The 
556: position of the long slit is seen to pass through the nucleus, and through
557: the clouds in spots.  The cross depicts the position of the nuclear continuum
558: center.  The backward `S' shape of the main clouds is seen.}
559: 
560: \figcaption[ruiz.fig2.ps]{STIS slitless spectrum showing the region around 
561: [O III] $\lambda$5007.  The horizontal axis is along the dispersion, while the
562: vertical scale is the spatial axis.  Note the high dispersions of the NLR clouds.
563: The clouds are also shifted slightly along the dispersion axis, the shift
564: allowing the radial velocity to be calculated.  The fainter ENLR clouds can
565: also be seen $\sim$1'' above and below the NLR.  Their dispersion can also be
566: seen.}
567: 
568: \figcaption[ruiz.fig3.ps]{Radial velocities averaged over bright clouds in the long slit 
569: are shown plotted against their corresponding values obtained with the 
570: slitless spectrum.  The error bars represent the standard deviations from 
571: the averages.} 
572: 
573: \figcaption[ruiz.fig4.ps]{Radial velocities (with respect to systemic) 
574: averaged over each individual cloud vs each cloud's distance from the optical 
575: continuum center.  Standard deviations in the position and velocity averages 
576: are shown as the error bars.  The slitless clouds (marked with open circles) 
577: include all clouds, including those not lying in the slit, and those extending 
578: out into the ENLR (from 1$''$.5 west of the nucleus outwards, and from 1$''$.0 
579: east of the nucleus outwards).}
580: 
581: \figcaption[ruiz.fig5.ps]{This figure shows the positions in the sky of the 
582: clouds, with their radial velocity magnitudes symbolised by the size of the 
583: figure used to identify them.  The blueshifted clouds are symbolized by a 
584: ``+'', while redshifted clouds are an ``x''.  The radial velocity measurements 
585: are binned over each cloud.  The ENLR clouds can be seen to be mostly 
586: redshifted.  The long slit clouds are differentiated by being in boldface.}
587: 
588: \figcaption[ruiz.fig6.ps]{Same as Figure 5, but zoomed in on the 
589: NLR clouds.}
590: 
591: \figcaption[ruiz.fig7.ps]{Unbinned radial velocity of each row 
592: perpendicular to the dispersion direction plotted against the distance 
593: from the center.  The slitless spectral data points plotted here are only 
594: those observed to lie within the slit.  Errors are roughly equal to the 
595: size of the symbols.}
596: 
597: \figcaption[ruiz.fig8.ps]{Same as Fig. 7, but overlaid with the best fit model
598: of the radial acc + constant decel model.  Parameters of the model are given
599: in Table 3.}
600: 
601: \clearpage
602: \begin{figure}
603: \plotone{ruiz.fig1.ps}
604: \\Fig.~1.
605: \end{figure}
606: 
607: \clearpage
608: \begin{figure}
609: \plotone{ruiz.fig2.ps}
610: \\Fig.~2.
611: \end{figure}
612: 
613: \clearpage
614: \begin{figure}
615: \plotone{ruiz.fig3.ps}
616: \\Fig.~3.
617: \end{figure}
618: 
619: \clearpage
620: \begin{figure}
621: \plotone{ruiz.fig4.ps}
622: \\Fig.~4.
623: \end{figure}
624: 
625: \clearpage
626: \begin{figure}
627: \plotone{ruiz.fig5.ps}
628: \\Fig.~5.
629: \end{figure}
630: 
631: \clearpage
632: \begin{figure}
633: \plotone{ruiz.fig6.ps}
634: \\Fig.~6.
635: \end{figure}
636: 
637: \clearpage
638: \begin{figure}
639: \plotone{ruiz.fig7.ps}
640: \\Fig.~7.
641: \end{figure}
642: 
643: \clearpage
644: \begin{figure}
645: \plotone{ruiz.fig8.ps}
646: \\Fig.~8.
647: \end{figure}
648: 
649: \pagestyle{empty}
650: \end{document}
651: 
652: 
653: 
654: 
655: 
656: 
657: 
658: 
659: 
660: 
661: 
662: 
663: 
664: 
665: