1: \documentclass{nature-jae}
2: \bibliographystyle{naturemag}
3: \usepackage{graphics,psfig,epsfig}
4:
5: \def\la{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
6: \def\ga{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
7: \def\apj{Astrophys. J.}
8: \def\apjl{Astrophys. J.}
9: \def\aj{Astron. J.}
10: \def\pasp{PASP}
11: \def\nat{Nature}
12: \def\ao{Applied Optics}
13: \def\aap{Astron. Astrophys.}
14: \def\araa{Ann. Rev. Astron. Astrophys.}
15:
16: %\title{Resolving Gas and Dust in the Terrestrial Planet Forming
17: %Region of a Protoplanetary Disk}
18:
19: \title{Water Vapor and Hydrogen in the Terrestrial Planet Forming
20: Region of a Protoplanetary Disk}
21:
22: \author{J.A. Eisner}
23:
24: \begin{document}
25:
26: \maketitle
27: \begin{affiliations}
28: \item Department of Astronomy, 601 Campbell Hall,
29: University of California, Berkeley, CA 94720
30: \end{affiliations}
31:
32:
33:
34: \begin{abstract}
35: Planetary systems, ours included, are formed in disks of dust and
36: gas around young stars. Disks are an integral part of the star
37: and planet formation process\cite{SAL87,SAFRONOV69}, and knowledge of the
38: distribution and temperature of inner disk material is
39: crucial for understanding terrestrial
40: planet formation\cite{RQL04}, giant planet migration\cite{LBR96},
41: and accretion onto the central star\cite{KONIGL91}.
42: While the inner regions of protoplanetary disks in nearby star forming regions
43: subtend only a few nano-radians, near-IR interferometry
44: has recently enabled the spatial resolution of these terrestrial
45: zones. Most observations have probed only dust\cite{MILLAN-GABET+07},
46: which typically dominates the near-IR emission.
47: Here I report spectrally dispersed near-IR interferometric observations
48: that probe gas (which dominates the mass and dynamics of the inner disk),
49: in addition to dust, within one astronomical unit of the young star MWC 480.
50: I resolve gas, including water vapor and atomic hydrogen, interior to the edge
51: of the dust disk; this contrasts with results of previous spectrally dispersed
52: interferometry observations\cite{MALBET+07,TATULLI+07}.
53: Interactions of this accreting gas with migrating planets may lead to
54: short-period exoplanets like those detected around main-sequence
55: stars\cite{LBR96}. The observed water vapor is likely produced by the
56: sublimation of migrating icy bodies\cite{CC06}, and provides a potential
57: reservoir of water for terrestrial planets\cite{DRAKE05}.
58: \end{abstract}
59:
60:
61: %\section*{Introduction}
62: %\paragraph*{Introduction}
63:
64: %Protoplanetary disks provide a reservoir of material from which planets may
65: %form, and the abundance and properties of extra-solar
66: %planets\cite{MARCY+05} (as well as the architecture of our own solar system)
67: %suggest that planets frequently form in or migrate through inner regions of
68: %protoplanetary disks. Furthermore, the innermost disk regions represent the
69: %interface between the disk, accretion, and outflow\cite{SHU+94},
70: %and knowledge of the
71: %distribution of material in the inner disk is crucial for understanding the
72: %mass assembly and angular momentum evolution of pre-main-sequence stars.
73: %Interferometry, a technique that synthesizes a large aperture using
74: %two or more smaller, separated apertures, can achieve orders of magnitude
75: %higher angular resolution than conventional telescopes, and can
76: %spatially resolve disk terrestrial regions.
77: %While most interferometric measurements to date have probed only inner disk
78: %dust\cite{MILLAN-GABET+07}, spectrally dispersed interferometric measurements
79: %can also probe the gas, which dominates the mass and dynamics of the disk.
80:
81: %The study of terrestrial planet forming regions in protoplanetary disks is
82: %challenging because of the small angular scales involved. At distances
83: %to nearby star forming regions (e.g., Taurus at 140 parsecs), one astronomical
84: %unit (AU) subtends only a few milli-arcseconds (mas; 1 mas $\approx$ 5
85: %nano-radians). Large ground-based telescopes,
86: %and even the Hubble Space Telescope, can not spatially resolve such small
87: %scales. Most studies of inner disk regions have therefore relied on
88: %spectral information as a proxy for spatial resolution. By
89: %assuming radial temperature\cite{LA92} or velocity\cite{NAJITA+06}
90: %distributions, the spectral dependence of the inner disk emission can be
91: %translated into spatial structure. While such techniques have
92: %provided important insights into the structure of protoplanetary disks in
93: %the terrestrial planet forming region, the inherent assumptions lead to
94: %ambiguous results, and substantial gaps in our understanding
95: %remain due to a lack of high angular resolution observations.
96:
97: %Spatially resolved observations over a range of wavelengths are needed to
98: %determine the distributions of inner disk dust and gas, and to
99: %establish the temperature profiles of these components.
100: %Temperatures in terrestrial disk regions
101: %dictate the radii at which solids like dust and water ice
102: %can exist\cite{HAYASHI81},
103: %with direct implications for Earth-like planet formation and
104: %water delivery to terrestrial planets. Knowledge of the spatial distribution
105: %of inner disk gas is also crucial for understanding the origin of short-period
106: %giant planets (the so-called ``hot Jupiters''\cite{MARCY+05}); these planets
107: %probably migrated from larger radii, and may have ended up in their
108: %current orbits due to interactions with primordial inner disks
109: %\cite{LBR96}. Finally, measuring the spatial distribution of hydrogen and
110: %other gaseous emission in inner disk regions provides a direct test of
111: %whether this emission arises in accreting material or in outflowing
112: %winds\cite{MALBET+05b,TATULLI+06}.
113:
114: I obtained observations with a new grism\cite{EISNER+07b}
115: (a dispersive element consisting of a diffraction grating
116: on the vertex of a prism)
117: at the Keck Interferometer\cite{CW03,COLAVITA+03}
118: that show, for the first time, evidence of spatially resolved gaseous
119: spectral-line emission interior to the edge of the dusty disk around a young
120: star, MWC 480. My observations provide a
121: spectral dispersion of $R=230$ from 2.0 to 2.4 $\mu$m wavelength,
122: and an angular resolution of approximately 1 milli-arcsecond
123: (mas; 1 mas $\approx$ 5 nano-radians), which is $\sim 0.1$ astronomical units
124: (AU) at the 140 parsec distance to the target (the source is located in the
125: Taurus-Auriga star forming complex).
126: MWC 480 has a stellar mass of $2.3$ M$_{\odot}$, an age of
127: $\sim 6$ Myr, and a circumstellar
128: disk of dust and gas in Keplerian rotation\cite{MKS97}. Previous
129: observations found that the dusty component of the disk extended to within
130: $\sim 0.5$ AU of the central star\cite{EISNER+04,MONNIER+06}, and suggested
131: the presence of hot gaseous emission at even smaller stellocentric
132: radii\cite{EISNER+07a}. My observations confirm these suggestions, and
133: resolve hydrogen gas and water vapor in the inner disk.
134:
135: My interferometry data provide a measurement of the angular size of the
136: circumstellar emission from MWC 480 as a function of wavelength (Figure 1).
137: The spectral dependence of the angular size indicates a radial temperature
138: gradient. Since cooler material emits relatively more radiation
139: at longer wavelengths, an outwardly decreasing temperature gradient yields
140: an increasing fraction of emission from larger radii, and hence a larger
141: angular size, at longer wavelengths.
142: %The shape of the broadband spectral energy distribution for MWC 480
143: %strongly suggests that the gradient is not continuous, but rather traces
144: %multiple emission components at different stellocentric radii and temperatures
145: %\cite{EISNER+07a}.
146: %The slope of the size vs. wavelength curve constrains the temperature gradient
147: % between hot and cold components.
148: The temperature gradient probably arises from multiple components at different
149: stellocentric radii\cite{EISNER+07a}: warm dust at radii where temperatures
150: approach the dust sublimation temperature, and hotter gas at smaller radii.
151:
152: Figure 1 also shows more compact emission at 2.165 $\mu$m relative
153: to the surrounding continuum. This dip corresponds to the wavelength of the
154: Br$\gamma$ electronic transition ($n=7 \rightarrow 4$) of hydrogen.
155: The Br$\gamma$ emission observed from MWC 480 traces very hot
156: gas ($\sim 10^4$ K) that has recently recombined from
157: ionized hydrogen\cite{OSTERBROCK89}, presumably in the inner
158: reaches of an accretion column.
159:
160: %Also apparent in Figure 1 is a systematically larger angular size
161: %relative to the hot, short-wavelength continuum
162: %across spectral channels between 2.0 and 2.1 $\mu$m.
163: %This trend (also readily apparent in Figure 2) suggests the presence
164: %of water vapor, which has substantial opacity at these
165: %wavelengths\cite{LUDWIG71}, in the inner disk of MWC 480.
166:
167: To match the features of Figure 1, I constructed a simple physical model and
168: fitted it directly to the measured $V^2$ and fluxes. The model
169: accounts for emission of the unresolved central star, circumstellar
170: continuum emission from hot gas, spectral-line emission from hydrogen
171: (Br$\gamma$), and continuum emission from warm dust.
172: I take each emission component to be a uniform ring whose
173: width is 20\% of its radius. This multiple-ring model approximates a
174: continuous disk where dust, gas, and individual gaseous atoms or molecules
175: contribute emission at different size scales. It has been shown previously
176: that ring models provide reasonable approximations to more complex
177: models of inner disk regions\cite{EISNER+04,IN05}.
178: I also assume that the gaseous continuum and Br$\gamma$ emission originate
179: from stellocentric radii smaller than 0.1 AU; given the $\sim 1$ mas angular
180: resolution of my observations, I can not accurately distinguish
181: between smaller radii, although I can rule out larger radii.
182:
183: As indicated in Figure 2,
184: the model fits the data well over most of the spectral bandpass.
185: However, there are substantial deviations between the model predictions
186: and the data in the region between 2.0 and 2.1 $\mu$m (particularly for the
187: fluxes). These deviations suggest the presence of hot water vapor, which has
188: substantial opacity from 2.0 to 2.1 $\mu$m, in the inner disk
189: of MWC 480. I include water vapor in the model by adding a ring of
190: material with the opacity of water vapor\cite{LUDWIG71}. The superior quality
191: of the fit when water is included clearly demonstrates the presence of H$_2$O
192: in the inner disk of MWC 480 (for further arguments why the model
193: must include water vapor emission in order to match the data, see the
194: Supplementary Information section).
195:
196: The best-fit model includes
197: compact gaseous continuum emission with a temperature of
198: $2410 \pm 125$ K and Br$\gamma$ emission
199: that adds $17\% \pm 7\%$ to this gaseous continuum flux at 2.165 $\mu$m; these
200: components are situated at stellocentric radii $<0.1$ AU. Water vapor
201: emission is located at a radius of $0.16 \pm 0.05$ AU. The temperature of
202: this material is $2300 \pm 120$ K and the column density is
203: $1.2 \pm 0.6 \times 10^{19}$ cm$^{-2}$. Dust emission
204: originates from a radius of $0.28 \pm 0.01$ AU and has a temperature of
205: $1200 \pm 60$ K. Quoted 1$\sigma$
206: uncertainties are quadrature sums of statistical
207: uncertainties in the fits and assumed 5\% systematic errors.
208:
209: % fits with no water vapor
210: %$T_{\rm gas}=2420 \pm 120$ K
211: %$17\% \pm 7\%$
212: %$R_{\rm dust}=0.28 \pm 0.01$ AU
213: %$T_{\rm dust}=1210 \pm 60$ K
214:
215: I illustrate the basic geometry of the innermost regions of MWC 480 in
216: Figure 3. As inferred from the data, gaseous emission is found at
217: substantially smaller stellocentric radii than the dust. The inner edge of the
218: dust disk lies at $\sim 0.3$ AU, where the disk temperature approaches the
219: sublimation temperature of silicate dust\cite{POLLACK+94}.
220: Because I also include contributions from gas in my analysis, the inferred
221: dust size is larger, and the temperature lower, than in previous work where
222: only dust emission was considered\cite{EISNER+04,MONNIER+06}.
223: Emission from H$_2$O is
224: found at $\sim 0.15$ AU, and hot gaseous continuum and Br$\gamma$ emission
225: at still smaller radii ($<0.1$ AU).
226: %Since the inferred temperature of
227: %the gaseous emission is $\sim 2500$ K, well below the excitation temperature
228: %of the Br$\gamma$ transition,
229: The observed Br$\gamma$ emission probably originates in an accretion column,
230: near the shock formed when infalling material
231: impacts high latitude regions of the star\cite{KONIGL91,NCT96,CG98}.
232:
233: My results substantiate models of inner disk structure
234: proposed previously for similar star+disk systems\cite{MUZEROLLE+04}.
235: Comparison with these models suggests that
236: in order to reproduce the ratio of flux in the gas and dust components
237: determined here ($\sim 0.5$), a mass accretion rate $\ga 10^{-7}$
238: M$_{\odot}$ yr$^{-1}$ must be invoked for MWC 480. At such high accretion
239: rates, the inner disk gas becomes optically thick to its own radiation, and
240: can therefore produce relatively more flux. Furthermore, models with this
241: high accretion rate produce strong Br$\gamma$ emission, consistent with my
242: detection. My measurement of gas interior to the
243: edge of the dust disk also confirms previous suggestions of such structure
244: based on modeling of spatially unresolved (but spectrally resolved)
245: observations of gaseous emission lines\cite{NCT96,CTN04,NAJITA+06} or very
246: low dispersion interferometric measurements\cite{EISNER+07a}.
247:
248: The finding that gaseous emission, and Br$\gamma$ emission in particular,
249: lies within the edge of the dusty component in protoplanetary disks contrasts
250: with previous interferometric studies of young stars.
251: Previous work found the Br$\gamma$ emission to
252: lie at radii equal to or larger than the dust, implying an origin of the
253: emission in an outflowing wind\cite{MALBET+07,TATULLI+07}.
254: %In the case where the Br$\gamma$ emission was more extended than the
255: %dust, the central source is very massive ($>20$ M$_{\odot}$), and such
256: %a luminous source may thermally excite Br$\gamma$ emission in an
257: %extended wind\cite{KS81}. However for the object where the Br$\gamma$
258: %and dust emission lie at similar radii, the central star is less luminous
259: %and this argument does not apply.
260: %however for a lower-luminosity star such as MWC 480, recombination is much
261: %more likely.
262: The differing results
263: for MWC 480 and the two previously observed sources may stem from
264: different luminosities of the objects. MWC 480 is orders of magnitude
265: less luminous than MWC 297\cite{EISNER+04}, for which the Br$\gamma$ emission
266: is more extended than the dust emission\cite{MALBET+07}, and approximately half
267: as luminous at HD 104237\cite{VDA+98}, for which the Br$\gamma$ and dust
268: emission have comparable size scales\cite{TATULLI+07}. Br$\gamma$ emission
269: probably traces both accreting and outflowing components in all of these
270: sources to some extent, but the accretion appears to increasingly dominate the
271: emission for lower-luminosity stars.
272:
273: The distribution of gas at stellocentric radii $<0.1$ AU in the inner disk of
274: MWC 480 shows that material is present in this protoplanetary disk at radii
275: comparable to the semi-major axes of short-period exoplanets\cite{MARCY+05}.
276: This finding supports the
277: hypothesis that giant planet migration is halted in resonances with the
278: inner edge of the gaseous disk\cite{LBR96,NAJITA+06}. Future observations of
279: gaseous disk truncation radii in solar-type T Tauri stars will enable
280: direct comparison with exoplanets, which are predominantly found around
281: solar-type stars.
282:
283: The presence of water vapor in the inner regions of the
284: protoplanetary disk surrounding MWC 480 constrains how water
285: (and other material) is transported in the disk terrestrial zone.
286: Because water beyond the ice-line is in solid form, a concentration gradient
287: tends to push water vapor out of the inner disk; the fact that I observe
288: this inner disk vapor suggests continual replenishment through the sublimation
289: of inwardly migrating icy bodies\cite{CC06}. Water vapor in disk
290: terrestrial regions may also enable the {\it in situ} formation of
291: water-rich planets via adsorption of water onto dust grains and subsequent
292: growth of these water-laden grains into planetesimals\cite{DRAKE05}.
293: Finally, as one of the dominant gaseous opacity sources in the near-IR,
294: water vapor can provide an excellent tracer of the temperature and
295: velocity structure of inner disk gas\cite{CTN04}.
296:
297:
298: %As in previous work\cite{MILLAN-GABET+07}
299: %we find dusty material at radii larger
300: %than 0.3 AU; the building blocks of rocky planets are thus
301: %present in the disk terrestrial region\cite{VANBOEKEL+04}.
302: %The fact that gas is found
303: %at still smaller radii, $<0.1$ AU, shows that material is
304: %present in protoplanetary disks at radii comparable to the semi-major axes
305: %of short-period exoplanets\cite{MARCY+05}; this finding supports the
306: %hypothesis that giant planet migration is halted in resonances with the
307: %inner edge of the gaseous disk\cite{LBR96,NAJITA+06}.
308: %Finally, I showed that water, one of the most important building blocks
309: %of potentially habitable terrestrial planets, is present in the inner regions
310: %of the protoplanetary disk surrounding MWC 480.
311: %Given that most of what we
312: %know about planets concerns systems around solar-type stars,
313: %Extending interferometric observations to the terrestrial disk regions around
314: %solar-type pre-main-sequence stars is an important next step in
315: %understanding planet formation and disk accretion in young
316: %analogs of our own planetary system.
317:
318: \bibliography{jae_ref}
319:
320: \begin{addendum}
321: \item Data presented herein were obtained at the W.M. Keck Observatory from
322: telescope time allocated to the National Aeronautics and Space Administration
323: through the agency's scientific partnership with the California Institute of
324: Technology and the University of California. The Observatory was made possible
325: by the generous financial support of the W.M. Keck Foundation. I thank the
326: entire Keck Interferometer team for their invaluable contributions to these
327: observations. I also acknowledge input into this work (and this manuscript)
328: from R. Akeson, E. Chiang, A. Glassgold, J. Graham, J. Najita, and R. White.
329: The author is supported by a Miller Research Fellowship.
330: \item[Competing Interests] The author declares that he has no
331: competing financial interests.
332: \item[Correspondence] Correspondence and requests for materials
333: should be addressed to J.A.E. \\ (email: jae@astro.berkeley.edu).
334: \end{addendum}
335:
336:
337:
338: \clearpage
339:
340: \includegraphics[clip=true,scale=1.0]{fig1.eps}
341: \newline
342: \noindent {\bf Fig. 1.}
343: {\bf Angular size of the near-IR emission from MWC 480 as a function of
344: wavelength.} Sizes are calculated from squared visibility measurements
345: ($V^2$) made with the Keck Interferometer grism on 12 Nov 2006.
346: The $V^2$ (and fluxes; see Fig. 2), were calibrated using
347: observations of unresolved stars of known spectral type and
348: parallax\cite{EISNER+07b}. I applied these calibrations to unresolved
349: ``check stars'' to test the calibration, and thereby estimated
350: channel-to-channel uncertainties ($1\sigma$ standard deviation) in my flux and
351: $V^2$ measurements of 3\%.
352: The shortest-wavelength data point has a larger uncertainty, because
353: atmospheric water vapor absorption
354: leads to lower photon counts in this channel; I estimate the uncertainty
355: in this channel at $\sim 10\%$.
356: There is also a systematic uncertainty of approximately 5\% in the overall
357: normalization of the data, due to errors in
358: the absolute calibration\cite{EISNER+07b}.
359: I assume that the emission morphology is a uniform disk,
360: for which the measured $V^2$ are
361: related to the angular diameter of the source, $\theta_{\rm UD}$, by
362: $V^2 = \left\{{J_1(\pi \theta_{\rm UD} B_{\rm proj}/\lambda)}
363: /({\pi \theta_{\rm UD} B_{\rm proj}/\lambda})\right\}^2,$
364: where $B_{\rm proj}$ is the projected baseline separation of the two telescopes
365: as seen by the source, $\lambda$ is the wavelength, and $J_1$ is a first-order
366: Bessel function. I remove the stellar contribution to the
367: visibilities\cite{EISNER+04}, and the plotted sizes thus represent only
368: the circumstellar emission. The angular size increases from
369: 2.0 to 2.4 $\mu$m, indicating multiple emission components: hot gas
370: at small stellocentric radii and warm dust at larger radii\cite{EISNER+07a}.
371: In addition, the angular size
372: is smaller at 2.165 $\mu$m, the wavelength where hot hydrogen gas emits in the
373: Br$\gamma$ line, than at adjacent continuum wavelengths.
374: %The systematically
375: %larger angular sizes from 2.0 to 2.1 $\mu$m, where water vapor emission lines
376: %blend together to produce substantial opacity\cite{LUDWIG71}, indicates
377: %that H$_2$O is found at slightly larger radii than the hot gaseous
378: %continuum emission.
379:
380: \includegraphics[clip=true,scale=1.0]{fig2.eps}
381: \newline
382: \noindent{\bf Fig. 2.}
383: {\bf Measured $V^2$ and fluxes, compared to the
384: predictions of simple physical models.} Plotted error bars indicate
385: estimated 1$\sigma$ (standard deviation) uncertainties of 3\%; as noted
386: above, the uncertainty in the shortest wavelength $V^2$ measurement is
387: larger, approximately 10\%.
388: The dotted line shows the values computed for a model including continuum
389: emission from warm dust and hot gas, and Br$\gamma$ emission from hydrogen.
390: This model reproduces the overall slopes seen in the data as well as the
391: feature at 2.165 $\mu$m, but deviates from the data between 2.0 and 2.1
392: $\mu$m. Hot water vapor contributes significant emission at these
393: wavelengths\cite{LUDWIG71}, and the solid line shows the predictions of a
394: model that includes H$_2$O. Because the H$_2$O emission lies at a radius only
395: slightly larger than that of the hot gaseous continuum, its effect on the
396: visibilities is more subtle than its effect on the measured fluxes.
397: Absorption due to terrestrial atmospheric
398: water vapor affects the data in the shortest-wavelength channel, but not
399: substantially in longer-wavelength channels since, in contrast to the
400: hot water vapor in our model, water vapor at atmospheric temperature and
401: pressure conditions lacks substantial opacity at these
402: wavelengths\cite{LUDWIG71}. The model including H$_2$O fits the data well
403: over the entire observed wavelength range, and yields substantially smaller
404: residuals between model and data than the model without water vapor.
405: We note that the inferred sizes of the emission components in our model
406: are smaller than the uniform disk sizes plotted in Figure 1; this is
407: due primarily to the different characteristic sizes for uniform disk and
408: relatively narrower ring models\cite{EISNER+04}.
409: Additional details of the modeling and fitting
410: procedure can be found in previous work\cite{EISNER+07a}.
411:
412: \clearpage
413: \includegraphics[clip=true,scale=1.0]{fig3.eps}
414: \newline
415: \noindent {\bf Fig. 3.}
416: {\bf The environment within 1 AU of the young star MWC 480.}
417: Dust in the inner disk around MWC 480 extends inward to
418: approximately 0.3 AU ({\bf D}), where the disk temperature becomes
419: high enough to sublimate the dust. The ring-like appearance of the
420: dust sublimation front is compatible with my data, and with physical models
421: predicting a puffed-up and curved inner edge\cite{DDN01,IN05}. Interior to
422: the dust sublimation radius, gaseous material continues to accrete onto the
423: young star (and may also be blown away in a tenuous
424: wind\cite{SHU+94}, which is not depicted).
425: Based on a sketch for a similar young star\cite{MUZEROLLE+04},
426: I have portrayed a continuous gaseous disk whose optical depth
427: increases at smaller radii. I observe hot continuum
428: emission from the hottest, densest gaseous material at stellocentric radii
429: smaller than 0.1 AU
430: ({\bf B}) and spectral-line emission from hot water vapor at slightly larger
431: radii ({\bf C}). Finally, I observe Br$\gamma$ emission at radii smaller
432: than 0.1 AU, which traces hydrogen gas at high temperatures
433: ($\sim 10^4$ K). I speculate that
434: this emission arises in the hot accretion shock where infalling gas impacts
435: the stellar surface ({\bf A}). The $\sim 1$ mas angular resolution of
436: my observations translates into a spatial resolution of approximately 0.1 AU
437: at the distance to MWC 480.
438:
439: \clearpage
440:
441: %\begin{figure}
442:
443: %\caption{$ $}
444: %\end{figure}
445:
446: %\begin{figure}
447: %\caption{$ $}
448: %\end{figure}
449:
450: %\begin{figure}
451: %\caption{$ $}
452: %\end{figure}
453:
454:
455:
456: \end{document}
457:
458:
459:
460:
461:
462:
463:
464:
465:
466:
467:
468:
469:
470:
471:
472:
473:
474:
475:
476:
477: