1: \def\astroph{1}
2: \ifnum\astroph=0
3: \documentclass[manuscript]{aastex} % Single column, double spaced
4: %\documentclass[12pt,preprint]{aastex}
5: \else
6: \documentclass{emulateapj}
7: % \documentclass[a4paper]{emulateapj}
8: % \usepackage{emulateapj5}
9: \fi
10:
11: \usepackage{epsfig,amsmath,natbib}
12: \usepackage[dvips]{color}
13: \usepackage{verbatim}
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15:
16: \begin{document}
17: %
18: \shorttitle{Ly$\alpha$ Resonant Scattering in Young Galaxies}
19: \shortauthors{Laursen et al.}
20: %
21: \title{Lyman {\Large $\alpha$} Resonant Scattering in Young Galaxies --- Predictions
22: from Cosmological Simulations}
23: %
24: \author{Peter Laursen\altaffilmark{1}
25: and Jesper Sommer-Larsen\altaffilmark{1,2}}
26: %
27: \email{pela@dark-cosmology.dk, jslarsen@tac.dk}
28: %
29: \altaffiltext{1}{Dark Cosmology Centre, Niels Bohr Intitute, University of
30: Copenhagen, Juliane Maries Vej~30, DK-2100, Copenhagen {\O},
31: Denmark.}
32: \altaffiltext{2}{Institute of Astronomy, University of Tokyo, Osawa 2-21-1,
33: Mitaka, Tokyo, 181-0015, Japan.}
34: %
35: \begin{abstract}
36: We present results obtained with a three-dimensional, Ly$\alpha$ radiative
37: transfer code applied to a fully cosmological galaxy formation
38: simulation. The developed Monte Carlo code is capable
39: of treating an arbitrary distribution of source Ly$\alpha$ emission, neutral
40: hydrogen density, temperature, and peculiar velocity of the interstellar
41: medium.
42: We investigate the influence of resonant scattering on the appearance and
43: properties of young galaxies by applying the code to a simulated ``Lyman-break
44: galaxy'' at redshift
45: $z = 3.6$, and of star formation rate 22 $M_{\odot}$ yr$^{-1}$ and total
46: Ly$\alpha$ luminosity $2.0\times10^{43}$ erg s$^{-1}$. It is found that
47: resonant scattering of Ly$\alpha$
48: radiation can explain that young galaxies are frequently observed to be more
49: extended on the sky in Ly$\alpha$ than in the optical. Moreover, it is
50: shown that, for the system investigated, due to the anisotropic escape of the
51: photons, the observed maximum surface brightness can differ by more than an
52: order of magnitude, and the total
53: derived luminosity by a factor of $\sim 4$, depending on the orientation of
54: the system relative to the observer.
55: \end{abstract}
56: %
57: \keywords{galaxies: formation --- line: formation --- line: profiles ---
58: radiative transfer --- scattering}
59:
60: \section{Introduction}
61: %
62: The Ly$\alpha$ line is a very important diagnostic in a wide range of fields
63: of astrophysics, not the least of which is galaxy formation, providing us with
64: extensive
65: information on redshift, dynamics, kinematics, morphology, etc. Three
66: distinct physical processes result in Ly$\alpha$ source emission in the
67: context of galaxies:
68: First, Ly$\alpha$ emission due to photo-ionization
69: of hydrogen atoms by UV radiation from nearby, massive stars
70: %\citep[and, possibly, a quasar; see, e.g.,][]{hai01,wei04}
71: and subsequent recombinations may contribute as much as 10\% of the
72: total luminosity of the galaxy \citep{par67}.
73: Second, part of the potential energy gained by gas falling into galactic
74: potential wells is converted into cooling radiation. \citet*{far01} find
75: that, at high redshifts, most of this radiation is emitted by gas with
76: $T < 20\,000$~K, and consequently $\sim 50\%$ in Ly$\alpha$ alone.
77: Finally, the external, metagalactic UV field, penetrating some (in case of
78: damped
79: Ly$\alpha$ systems) or all (in case of Lyman-limit systems) of the
80: outer parts of galactic hydrogen ``envelopes'' will also produce some
81: Ly$\alpha$ radiation through case B
82: recombination.
83: %In fact, approximately two-thirds of such recombinations
84: %will result in a Ly$\alpha$ photon \citep{spi78}.
85: Moreover, this UV field can also photo-heat non-self-shielded gas,
86: which subsequently cools, radiating Ly$\alpha$ \citep{fur05}.
87:
88: Over the past years it has become possible to actually resolve
89: observationally these young, Ly$\alpha$ emitting galaxies. In several cases,
90: the galaxies have been
91: found to be significantly more extended on the sky when observed in Ly$\alpha$
92: as opposed to optical bands \citep[e.g.,][]{mol98,fyn01,fyn03}.
93: Due to the complexity and
94: diversity of the systems, in order to correctly interpret observations and
95: make predictions about the properties of young galaxies, it is desirable to
96: develop realistic theoretical and numerical models.
97:
98: For a number of idealized cases analytical solution are obtainable.
99: %Inspired by \citet{unn55},
100: \citet{har73} investigated the emergent spectrum of resonantly scattered
101: radiation in the case of a highly optically thick slab of
102: finite thickness but infinite extension, and
103: uniform temperature and density.
104: %He also gave a quantitative explanation of the fact discovered by
105: %\citet{ada72} that the total number of scatterings that a photon undergoes
106: %before escaping the medium scales proportionally to the optical depth $\tau_0$
107: %in the line center and not, as one would naively expect from random walk
108: %considerations, as $\tau_0^2$.
109: \citet{neu90} extended this solution to include the possibility of the photons
110: being destroyed (as by dust) and injected with arbitrary initial frequency.
111: \citet{dij06} derived a similar analytical expression for spherical symmetry,
112: allowing for isotropic expansion or collapse of the gas, and
113: \citet{loe99} examined the spectrum for an isotropically expanding or
114: contracting medium with no thermal motion.
115:
116: However idealized these configurations may seem, they provide valuable and
117: at least qualitative insight into the characteristics of young galaxies.
118: Moreover, they offer direct means of testing numerical methods.
119:
120: The Monte Carlo (MC) method %was developed in the late 1940s \citep{met49}, and
121: has been used for solving radiative transfer (RT) problems since the
122: beginning of the 1960s.
123: %\citep{fle63}.
124: Nevertheless, though conceptually simple, the demand for strong
125: computer power until quite recently restricted this technique to deal with
126: more or less the same idealized configurations that had already been dealt with
127: analytically. Thus, the majority of previous attempts to
128: model RT in astrophysical situations have been based on strongly simplified
129: configurations of the physical parameters. Only a few authors
130: \citep{can05,tas06,ver06} considered more general cases.
131:
132: With the aim of predicting the appearance and properties of
133: Ly$\alpha$ emitting galaxies, we developed an MC code capable of treating an
134: arbitrary distribution of source Ly$\alpha$ emission, neutral
135: hydrogen density, temperature, and peculiar velocity of the medium and
136: subsequently applied it to a simulated Lyman-break galaxy (LBG) from a fully
137: cosmological simulation.
138: \ \\
139: \ \\
140:
141: \section{Simulations}
142: \label{sec:code}
143:
144: \subsection{The Cosmological Simulation and Monte Carlo Code}
145: \label{sec:ini}
146:
147: The cosmological simulation of the formation and evolution of an individual
148: galaxy was performed using the N-body/hydrodynamical
149: TreeSPH code of \citet[][see also \citet{som06}]{som03}.
150: The system becomes a Milky Way/M31-like disk galaxy at $z = 0$ and is
151: simulated using $\sim2.2\times10^6$ particles in total, comprising only
152: smoothed particle hydrodynamic (SPH) and dark matter (DM) particles at the
153: initial redshift $z_i = 39$.
154: The masses and gravity softening lengths
155: of SPH and star particles are $9.9 \times 10^4 h^{-1} M_{\odot}$ and
156: $200 h^{-1}$ pc, respectively ($h = 0.65$). For the DM particles the
157: corresponding
158: values are $5.7 \times 10^5 h^{-1} M_{\odot}$ and $370 h^{-1}$ pc. The minimum
159: SPH smoothing length in the simulation is about $13 h^{-1}$ pc.
160: For the purposes of the present study, a single simulation output, at
161: $z = 3.6$, was chosen.
162:
163: The MC code, used to propagate the photons through the medium, by and
164: large resembles those recently developed by other authors
165: \citep[e.g.,][]{can05,tas06}. Since it is grid-based --- the number of cells
166: typically being 512$^{3}$ --- the physical parameters of interest are first
167: interpolated from the
168: %gas particles of the SPH simulation
169: SPH particles to
170: the cells of the computational box. These parameters are the Ly$\alpha$
171: emissivity, the temperature $T$, the density
172: $n_{\textrm{{\scriptsize H}{\tiny I}}}$ of neutral hydrogen, and
173: the three-dimensional velocity field $\mathbf{v}_{\mathrm{bulk}}$ of the gas.
174:
175: \subsection{Determination of Source Ly$\alpha$ Emission}
176: \label{sec:em}
177:
178: The emission cell of a given Ly$\alpha$ photon is found by setting the
179: probability of being emitted from a given cell equal to the ratio of the
180: luminosity in that cell to the total luminosity $L_{\mathrm{tot}}$. The photon
181: is then injected in the line center (in the reference frame of the fluid
182: element) from a random point $\mathbf{x}_i$ in the
183: cell.
184: %Following most earlier work on the subject,
185: The frequency $\nu$ of the photon is parametrized through
186: $x = (\nu - \nu_0)/\Delta\nu_{\mathrm{D}}$, where $\nu_0 = 2.466\times10^{15}$
187: Hz is the line center frequency and
188: $\Delta\nu_{\mathrm{D}} = (v_{\mathrm{th}}/c)\nu_0$ is the Doppler frequency,
189: with $v_{\mathrm{th}} = (2 k_B T / m_{\mathrm{H}})^{1/2}$ being the thermal
190: atom velocity dispersion (times $\sqrt{2}$) and the rest of the variables
191: having their usual meanings. In terms of this variable the injected photon
192: obviously has a frequency of $x = 0$.
193: The initial direction
194: $\mathbf{\hat{n}}_i$ of the photon follows an isotropic probability
195: distribution.
196:
197: \subsection{Propagation of the Radiation}
198: \label{sec:prop}
199:
200: The optical depth $\tau$ covered by the photon before it is scattered is
201: drawn randomly from the probability distribution $e^{-\tau}$, and subsequently
202: converted into a physical distance
203: $r = \tau / n_{\textrm{{\scriptsize H}{\tiny I}}} \sigma_x$, where
204: the physical parameters are given by the present cell. The
205: cross section $\sigma_x$ is given by a Voigt profile, i.e.~the convolution of
206: the Lorentzian natural line profile and the Gaussian thermal broadening of the
207: atoms, resulting in
208: %
209: \begin{equation}
210: \label{eq:xsec}
211: \sigma_x = f_{12} \frac{\sqrt{\pi} e^2}{m_e c \Delta\nu_{\mathrm{D}}} H(a,x),
212: \end{equation}
213: %
214: where $f_{12} = 0.4162$ is the Ly$\alpha$ oscillator strength, and
215: %
216: \begin{equation}
217: \label{eq:H}
218: H(a,x) = \frac{a}{\pi} \int_{-\infty}^{+\infty}
219: \frac{e^{-y^2}}{(x-y)^2 + a^2} dy
220: \end{equation}
221: %
222: is the Voigt function with $a = \Delta\nu_{\mathrm{L}}/2\Delta\nu_{\mathrm{D}}$
223: the ratio between the natural line width
224: $\Delta\nu_{\mathrm{L}} = 9.936\times10^{7}$ Hz and the Doppler width.
225: Since eq.~(\ref{eq:H}) is not analytically
226: integrable, we use the analytic fit given by \citet{tas06}.
227:
228: If the final position $\mathbf{x}_f$ of the photon is outside the original
229: cell, the photon is placed at the point $\mathbf{x}_{\mathrm{int}}$
230: of intersection with the face of the
231: cell and the above calculation is redone with the parameters of the new cell,
232: an optical depth $\tau' = \tau - |\mathbf{x}_{\mathrm{int}} - \mathbf{x}_i|\,
233: ( n_{\textrm{{\scriptsize H}{\tiny I}}}\,
234: \sigma_x )_{\mathrm{prev.cell}}$, and a frequency Lorentz transformed to the
235: bulk velocity of the new cell. This procedure is repeated until either the
236: originally assigned value of $\tau$ is spent, and the photon is scattered, or
237: it leaves the computational box.
238:
239: \subsection{Scattering}
240: \label{sec:scat}
241:
242: Except for a small recoil effect, the scattering is coherent in the reference
243: frame of the atom. However, to an external observer, the non-zero velocity
244: $\mathbf{u} = \mathbf{v}_{\mathrm{atom}}/v_{\mathrm{th}}$ of
245: the atom will shift the frequency of the photon. In the directions
246: perpendicular to $\mathbf{\hat{n}}_i$,
247: the velocities $u_{\perp1,2}$ will follow a Gaussian
248: distribution. When $x \sim 0$, the photon barely diffuses spatially. Only when
249: it has diffused sufficiently far in frequency space, will it be able to make
250: a large journey in real space. To skip these non-important core scatterings and
251: thus accelerate the code, if $|x|$ is less than some critical value
252: $x_{\mathrm{crit}}$, following \citet{dij06} $u_{\perp1,2}$ is drawn from a
253: truncated Gaussian so as
254: to favor fast moving atoms and artificially push the photon back in the wing.
255: $x_{\mathrm{crit}}$ is determined according to $a \tau_0$ in the given cell. If
256: $a \tau_0 \le 1$, a proper Gaussian is used; otherwise we find that
257: $x_{\mathrm{crit}} = 0.02\,e^{\xi \ln^{\chi} a\tau_0}$, where
258: $(\xi,\chi) = (0.6,1.2)$ or $(1.4,0.6)$ for
259: $a\tau_0 \le 60$ or $a\tau_0 > 60$, respectively, can be used without
260: affecting the final result. The effect of this acceleration scheme is a
261: speed-up of several orders of magnitude.
262:
263: Due to the resonance nature of the scattering event,
264: the velocity $u_{||}$ parallel to $\mathbf{\hat{n}}_i$ depends on $x$, and is
265: generated following \citet{zhe02}.
266:
267: % for small values of $|x|$, being scattered by an atom with $u_{||} = x$ is
268: % highly favored. For large values of $|x|$ the abundance of these atoms
269: % decrease as $e^{-x^2}$, so that scattering by ``slow'' atoms becomes more
270: % probable, even though in reference frame of these atoms the photon is far in
271: % the wing. The resulting normalized probability distribution is
272: % %
273: % \begin{equation}
274: % \label{eq:fu}
275: % f(u_{||}) = \frac{a}{\pi H(a,x)} \frac{e^{-u_{||}^2}} {(x-u_{||})^2 + a^2}.
276: % \end{equation}
277: % %
278: % $u_{||}$ is generated from this distribution following \citet{zhe02}. As
279: % the value $u_0$ separating the two comparison functions \citep{zhe02}, for the
280: % wide range of temperatures and frequencies involved, we find that a
281: % satisfactory average acceptance-to-rejection ratio of approximately $1/4$ is
282: % achieved for
283: % %
284: % \begin{equation}
285: % \label{eq:u0}
286: % u_0 = \left\{ \begin{array}{ll}
287: % 0 & \textrm{for } 0 \phantom{.2}\le x < 0.2\\
288: % x - a e^{1.2x + b - 0.5\sin(2\pi x/5)} & \textrm{for } 0.2 \le x < Q(a)\\
289: % 4.5 & \textrm{for } \phantom{0.2| \le } x \ge Q(a),
290: % \end{array}
291: % \right.
292: % \end{equation}
293: % %
294: % where $b = -0.84 \ln a - 4.4$, and
295: % %
296: % \begin{equation}
297: % \label{eq:Qa}
298: % Q(a) = -0.0283 \ln^2\! a - 0.514\ln a + 3.648.
299: % \end{equation}
300: % %
301: The final frequency $x'$ of the scattered photon (in the reference frame of the
302: fluid element)
303: depends on direction in which the photon is scattered. For scattering in the
304: line center, transitions to the $2P_{1/2}$ state results in
305: isotropic scattering, while the $2P_{3/2}$ transition causes some polarization,
306: resulting in a probability distribution
307: $W(\theta) \propto 1 + \frac{3}{7}\cos^2\theta$, where $\theta$ is the angle
308: between $\mathbf{\hat{n}}_i$ and the outgoing direction $\mathbf{\hat{n}}_f
309: $\citep{ham40}. Since the spin multiplicity is
310: $2J + 1$, the probability of being excited to the $2P_{3/2}$ state is twice as
311: large as being excited to the $2P_{1/2}$ state\footnote{For the environments
312: produced here, transitions to the $2S$ state and subsequent
313: destruction of the photon can be neglected.}. For scatterings in the wing,
314: polarization for $\pi/2$ scattering is maximal, resulting in a dipole
315: distribution $W(\theta) \propto 1 + \cos^2\theta$ \citep{ste80}. The
316: transition from core to wing scattering is given by the value of $x$ where
317: the Lorentzian starts dominating the Gaussian profile, i.e.~where
318: $a/\pi x^2 \sim \sqrt{\pi} e^{-x^2}$, or $x \sim 3$ (the exact value is not
319: crucial). In all
320: cases, the scattering is isotropic in the azimuthal angle $\phi$. In the
321: observers frame, the final frequency is then given by
322: $x' = x - u_{||} + \mathbf{\hat{n}}_f \cdot \mathbf{u}
323: + g(1 - \mathbf{\hat{n}}_i \cdot \mathbf{\hat{n}}_f)$,
324: % %
325: % \begin{equation}
326: % \label{eq:x}
327: % x' = x - u_{||} + \mathbf{\hat{n}}_f \cdot \mathbf{u}
328: % + g(1 - \mathbf{\hat{n}}_i \cdot \mathbf{\hat{n}}_f),
329: % \end{equation}
330: % %
331: where the factor $g = h_{\mathrm{Pl}} \nu_0 / m_{\mathrm{H}} c v_{\mathrm{th}}$
332: \citep{fie59} accounts for the recoil effect.
333:
334: \subsection{Observable Surface Brightness Maps}
335: \label{sec:obs}
336:
337: Following the above scheme, the photons are propagated through the medium, one
338: by one, until they escape the computational box. For each scattering
339: (as well as the emission) of a photon, the probability that the photon will
340: escape in the direction of
341: six virtual observers situated in the negative and positive directions of the
342: three principal axes is calculated. This probability is added as a weight to a
343: three
344: dimensional array (the two spatial dimensions of the projected image of the
345: galaxy, plus a spectral dimension for each pixel). The array is finally
346: collapsed along the spectral dimension and along the spatial dimensions to
347: yield the image and spectrum, respectively, that an external observer would
348: see.
349: The contribution of each pixel element to the surface
350: brightness (SB) is then
351: %\footnote{Measured in erg s$^{-1}$ cm$^{-2}$ arcsec$^{-2}$ at the location of
352: %the observer.}
353: %
354: \begin{equation}
355: \label{eq:SB}
356: \mathrm{SB_{\mathrm{pix}}} = \frac{L_{\mathrm{tot}}}{n_{\mathrm{ph}}}
357: \frac{1}{d_L^2}
358: \frac{1}{\Omega_{\mathrm{pix}}}
359: \sum_{\mathrm{ph.,scat.}}
360: e^{-\tau_{\mathrm{esc}}} W(\theta),
361: \end{equation}
362: where $n_{\mathrm{ph}}$ is the number of photons, $d_L$ is the luminosity
363: distance to the observer, $\Omega_{\mathrm{pix}}$ is the solid angle subtended
364: by the pixel, and $\tau_{\mathrm{esc}}$ is the optical depth of the gas lying
365: between the scattering event and the edge of the computational box in the
366: direction of the observer.\\[.3cm]
367: %
368: The MC code was tested on various simple configurations for which analytic
369: solutions exist, discussed in the introduction. Our code exquisitely passes all
370: tests. The results of these tests will be presented in a future paper.
371:
372: \section{Results}
373: \label{sec:res}
374: %
375: The MC code was applied to a proto-galaxy at $z = 3.6$, consisting of two small,
376: star-forming ``disks'' separated by approximately 2 kpc, on one of which the
377: computational box is centered, and a third more extended disk of lower star
378: formation rate (SFR),
379: located about 15 kpc from the center. The star-forming
380: regions are embedded in a significant amount of more diffuse, non-star-forming
381: H\textsc{i} gas in a 10--15 kpc
382: thick, sheet-like structure, taken to constitute the $x$-$y$ plane.
383: The total SFR of the system is 22 $M_\odot$ yr$^{-1}$. Observed
384: LBGs have SFRs in the range
385: 10--1000 $M_\odot$ yr$^{-1}$ \citep[e.g.,][]{rig06}, so the
386: simulated galaxy corresponds to a small LBG.
387: The Ly$\alpha$ emissivity is produced according the three different scenarios
388: described in the introduction. Specifically, the luminosity originating from
389: the H\textsc{ii} in the vicinity of massive stars (accounting for
390: approximately 90\% of $L_{\mathrm{tot}}$) is determined following
391: \citet{far01}, using the code Starburst99 \citep{lei99} to yield the Lyman
392: continuum and
393: assuming a Miller-Scalo initial mass function,
394: a mean Lyman continuum photon energy of 1.4
395: Rydberg, and that 0.68 Ly$\alpha$ photons are emitted per photonionization.
396: The length of the box used for the MC calculations is 50 kpc (physical),
397: and $d_L \sim 34$ Gpc.
398:
399: Fig.~\ref{fig:SBmap} shows the results obtained (all with a resolution of
400: $512^3$ cells), viewed from two different
401: directions --- from the negative $x$- and $z$-direction,
402: corresponding to an edge-on and a face-on view of the sheet-like structure,
403: respectively. Upper panels assume that the gas is optically thin to the
404: Ly$\alpha$ line, lower show the corresponding results with resonant
405: scattering included.
406:
407: %
408: \begin{figure}
409: \epsscale{1.2}
410: \plotone{f1.eps}
411: \caption{Bolometric surface brightness map of a simulated galaxy at $z = 3.6$.
412: Left and right column shows the system when viewed edge-on
413: and face-on, respectively. The top panel displays the galaxy as if
414: the Ly$\alpha$ radiation was able to escape directly, without
415: scattering. The bottom panel shows the effect of the scattering.}
416: \label{fig:SBmap}
417: \end{figure}
418: %
419:
420: The effect of the scattering is incontestable: although the original
421: constellation of the principal emitters is still visible, the surface
422: brightness distribution is clearly much more extended. Moreover,
423: we notice the effect of the viewing angle. Qualitatively, we expect the
424: photons to escape more easily
425: from the face of the sheet than from the edge and hence that the system should
426: have a higher surface brightness than when viewed edge-on. Here, this
427: anticipation is quantified: Fig.~\ref{fig:ProfSpec} shows the azimuthally
428: averaged SB profiles.
429: %
430: \begin{figure*}
431: \epsscale{1}
432: \plotone{f2.eps}
433: \caption{\emph{Left}: Bolometric surface brightness (SB) profiles of the galaxy
434: when viewed edge-on (\emph{orange}) and face-on (\emph{green}). Both the
435: true (\emph{dotted}) and the smoothed (\emph{solid}, see text) profiles are
436: displayed.
437: Also shown are the SB profiles of the galaxy LEGO2138\_29 \citep[SFR
438: $\sim 15$ $M_\odot$ yr$^{-1}$ if extinction is negligible,][]{fyn03} in
439: Ly$\alpha$ (\emph{blue}) and in the $R$-band (\emph{red}, normalized to the maximum
440: observed Ly$\alpha$ SB). In particular the SB of the $y$-$z$ plane
441: nicely reproduces the observed SB.
442: Left axes measure the SB at the source, while right axes measure the
443: SB observed at $z = 0$.
444: \emph{Right}: Spectrum of the emergent radiation, clearly displaying
445: the characteristic double-peaked profile. The enhanced blue peak
446: indicates a net inward bulk velocity of the gas.}
447: \label{fig:ProfSpec}
448: \end{figure*}
449: %
450: To allow for direct comparison with observations, the profiles are also
451: shown not including
452: the luminosity of the emitter lying in the outskirts of the image and
453: smoothed with a PSF corresponding to a seeing of $0.8''$. The
454: maximum SB of the $x$-$y$ plane
455: is found to be $6.1 \times 10^{-2}$ erg s$^{-1}$ cm$^{-2}$, or $\sim14$ times
456: higher than that of the $y$-$z$ plane. The average
457: SB, from which the total luminosity is usually derived assuming isotropic
458: emission, is $3.8$ times higher.
459:
460: Finally, we show the emergent spectrum. Though not as clear for the edge-on
461: view, both profiles exhibit the characteristic double-peaked profile
462: \citep[see, e.g.,][]{ven05}.
463: Obviously, this is the result of the high opacity for photons near the line
464: center; diffusing to either side of $\nu_0$ quickly decreases $\tau$ so as to
465: make escape more probable. Furthermore, both profiles imply a net inward
466: velocity of the gas; for infalling gas, red photons are shifted into resonance,
467: while blue photons escape even more easily, thus enhancing the blue peak and
468: diminishing the red.
469:
470: Fitting a ``Neufeld profile'' to the observed spectra can give us an idea of
471: the intrinsic characteristica of the system. Unfortunately, there is a
472: degeneracy in that the profile is dependent only on the parameter $a\tau_0$,
473: Assuming some temperature, e.g.~$10^4$ K, representative of most of the
474: Ly$\alpha$
475: emitting gas, one could in principle deduce the equivalent column density
476: $N_{\textrm{{\scriptsize H}{\tiny I}}}$.
477: Alternatively, if $N_{\textrm{{\scriptsize H}{\tiny I}}}$ is
478: obtainable due to, e.g., the presence of a background quasar, constraints can
479: be put on the temperature.
480:
481: A bulk rotational motion of the gas will also alter the profile.
482: Since in fact a full spectrum is obtained for every pixel element, this effect
483: can be studied through long-slit spectroscopy.
484:
485: \section{Discussion and Conclusion}
486: \label{sec:disc}
487: %
488: The present MC calculations do not include the effect of dust.
489: Effectively, dust will act as a
490: photon sink and an extra scattering possibility. Since, on average, each
491: photon scatters $\sim 4 \times 10^{8}$ times and travel a distance of
492: $\sim 40$ kpc before escaping (determined from a
493: non-accelerated run of the code with a few
494: $10^3$ photons), approximately half of which is in the high
495: density ($n_{\textrm{{\scriptsize H}{\tiny I}}} \gtrsim 0.1$
496: cm$^{-3}$), cooler ($T \sim 10^4$ K) regions,
497: even a small amount of dust may be expected to be capable of causing a
498: significant decrease
499: in the observed intensity. However, it is not clear to which extend the dust
500: will affect the observations. If the medium is clumpy, the ratio of Ly$\alpha$
501: to continuum radiation may in fact be increased \citep{neu91,han06}. Moreover,
502: \citet{tas06} argued that dust acting as catalysator for hydrogen molecules
503: will lower $n_{\textrm{{\scriptsize H}{\tiny I}}}$, making the
504: medium more transparent. We will study these effects in future work,
505: implementing a realistic model of the dust based on the 8 different metal
506: species kept track of by the cosmological simulation.
507:
508: As a very recent improvement of the cosmological simulation, a post-processed
509: RT scheme of UV radiation from star-forming regions was developed by
510: \citet{raz06}. They found that up to 5--10\% of the ionizing photons escape these
511: regions at $z \sim 3.6$. However, a very preliminar analysis indicates that
512: the results presented in this work are not significantly changed by the
513: inclusion of H-ionizing photon RT. The quantative effects of this on
514: Ly$\alpha$ RT will also be discussed in a forthcoming paper.
515:
516: The developed Monte Carlo code has reproduced qualitatively
517: \emph{and} quantitatively the observation that young galaxies often appear
518: significantly more extended on the sky in Ly$\alpha$ than in the optical.
519: Furthermore, we investigated the impact of the viewing angle on the
520: observed surface brightness. Future simulations of a large statistical sample
521: of galaxies, taking properly into account dust and H-ionizing UV photon
522: radiative transfer, will allow us to learn more about the enigmatic Ly$\alpha$
523: emitters.
524: %
525: %--------------------------------------------------------------------------
526: %
527: \acknowledgments
528: %
529: The authors are very grateful to J.~Fynbo and T.~Haugb{\o}lle for beneficial
530: discussions, to A.~Tasitsiomi for quick response to questions, and to
531: J.~Schaye for useful comments. We also thank the anonymous referee for
532: constructive remarks. The simulations were performed
533: on the SGI Itanium II facility provided by DCSC. The Dark Cosmology Centre is
534: funded by the DNRF.
535:
536: \begin{thebibliography}{}
537: %
538: %\bibitem[Adams(1972)]{ada72} Adams, T.~F. 1972, \apj, 174, 439
539: \bibitem[Cantalupo et al.(2005)]{can05} Cantalupo, S., Porciani, C., Lilly,
540: S.~J., \& Miniati, F. 2005, \apj, 628, 61
541: \bibitem[Dijkstra et al.(2006)]{dij06} Dijkstra, M., Haiman, Z., \& Spaans, M.
542: 2006, ApJ, 649, 14
543: \bibitem[Fardal et al.(2001)]{far01} Fardal, M.~A., Katz, N., Gardner, J. P.,
544: Hernquist, L., Weinberg, D. H., \& Dav\'e, R. 2001, \apj, 562, 605
545: \bibitem[Field(1959)]{fie59} Field, G. 1959, \apj, 129, 551
546: %\bibitem[Fleck(1963)]{fle63} Fleck, J.~A., Jr. 1963, \emph{Methods in
547: % Computational Physics}, Vol. I, ed. B Alder, S. Fernbach, \& M.
548: % Rotenberg (New York: Academic Press)
549: \bibitem[Furlanetto et al.(2005)]{fur05} Furlanetto, S.~R., Schaye, J.,
550: Springel, V., \& Hernquist, L. 2005, ApJ, 622, 7
551: \bibitem[Fynbo et al.(2001)]{fyn01} Fynbo, J.~P.~U., M{\o}ller, P., \& Thomsen,
552: B. 2001, A\&A, 374, 443
553: \bibitem[Fynbo et al.(2003)]{fyn03} Fynbo, J. P. U., Ledoux, C., M{\o}ller, P.,
554: Thomsen, B., \& Burud, I. 2003, A\&A, 407, 147
555: \bibitem[Hamilton(1940)]{ham40} Hamilton, D.~R. 1940, Phys.~Rev, 58, 122
556: \bibitem[Hansen \& Oh(2006)]{han06} Hansen, M. \& Oh, S.~P. 2006, \mnras, 367,
557: 979
558: \bibitem[Harrington(1973)]{har73} Harrington, J.~P. 1973, \mnras, 162, 43
559: \bibitem[Leitherer et al.(1999)]{lei99} Leitherer, C. et al. 1999, \apjs, 123,
560: 3
561: \bibitem[Loeb \& Rybicki(1999)]{loe99} Loeb, R. \& Rybicki, G.~B. 1999, \apj,
562: 524, 527
563: \bibitem[M{\o}ller \& Warren(1998)]{mol98} M{\o}ller, P. \& Warren, S.~J.
564: 1998, \mnras, 299, 611
565: \bibitem[Neufeld(1990)]{neu90} Neufeld, D. 1990, \apj, 350, 216
566: \bibitem[Neufeld(1991)]{neu91} Neufeld, D. 1991, \apj, 370, L85
567: \bibitem[Partridge \& Peebles(1967)]{par67} Partridge, R.~B. \& Peebles,
568: P.~J.~E. 1967, \apj, 147, 868
569: \bibitem[Razoumov \& Sommer-Larsen(2006)]{raz06} Razoumov, A.~O. \&
570: Sommer-Larsen, J. 2006, \apj, 651, 81
571: \bibitem[Rigopoulou et al.(2006)]{rig06} Rigopoulou, D. et al 2006,
572: \apj, 648, 81
573: \bibitem[Sommer-Larsen et al.(2003)]{som03} Sommer-Larsen, J., G\"otz, M., \&
574: Portinari, L. 2003, \apj, 596, 47
575: \bibitem[Sommer-Larsen(2006)]{som06} Sommer-Larsen, J. 2006, \apj, 644, L1
576: %\bibitem[Spitzer(1978)]{spi78} Spitzer, L. 1978, \emph{Physical Processes in
577: % the Interstellar Medium} (New York Wiley-Interscience)
578: \bibitem[Stenflo(1980)]{ste80} Stenflo, J.~O. 1980, A\&A, 84, 68
579: \bibitem[Tasitsiomi(2006)]{tas06} Tasitsiomi, A. 2006, \apj, 645, 792
580: %\bibitem[Unno(1955)]{unn55} Unno, W. 1955, \pasj, 7, 81
581: \bibitem[Venemans et al.(2005)]{ven05} Venemans, B. et al. 2005,
582: A\&A, 431, 793
583: \bibitem[Verhamme et al.(2006)]{ver06}Verhamme, A., Schaerer, D., \& Maselli,
584: A. 2006, A\&A, 460, 397
585: \bibitem[Zheng \& Miralda-Escud\'e(2002)]{zhe02} Zheng, Z. \&
586: Miralda-Escud\'e, J. 2002, \apj, 578, 33
587: %
588: \end{thebibliography}
589:
590: \end{document}
591: