1: %\documentstyle[11pt,aasms4,flushrt]{article}
2: %
3: %\documentclass[12pt,preprint]{aastex}
4: %
5: %\documentclass[12pt,onecolumn]{pasj00}
6: %\draft
7: \documentclass{aastex}
8: \usepackage{emulateapj5,graphicx}
9: %\usepackage{epsfig}
10: %\usepackage{eclepsf}
11: %\usepackage{graphicx}
12: \begin{document}
13:
14: \newcommand{\bm}[1]{\mbox{\boldmath $#1$}}
15: %%%%%%%%%%%%%%%%%
16: % (1)TITLE PAGE %
17: %%%%%%%%%%%%%%%%%
18: \title{
19: Regulated star formation in forming disk galaxies\\
20: under ultraviolet radiation background
21: }
22: \author{Hajime Susa\altaffilmark{1}
23: \vskip 0.2cm
24: \affil{Department of Physics, Konan University, Okamoto, Kobe, Japan}
25: \altaffiltext{1}{susa@konan.ac.jp}
26: }
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28: % (2)Abstract & Subject Headings %
29: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
30: \begin{abstract}
31: We perform radiation hydrodynamics simulations on the evolution of
32: galactic gas disks irradiated by ultraviolet radiation background.
33: We find gas disks
34: with $N_{\rm H} \ga 10^{21} {\rm cm^{-2}}$ exposed to
35: ultraviolet radiation at a level of $I_{\rm 21}=1$ can be self-shielded
36: from photoheating, whereas the disk with $N_{\rm H}
37: \la 10^{21} {\rm cm^{-2}}$ cannot. We also find that the unshielded
38: disks keep smooth density distribution without any sign of
39: fragmentation, while the self-shielded disks easily fragment into small
40: pieces by self-gravity, possibly followed by star formation.
41: The suppression of star formation in unshielded disks
42: is different from
43: photoevaporation effect,
44: since the assumed dark halo potential is deep enough to keep the
45: photoheated gas.
46: Presence of such critical threshold column density would be one of the
47: reason for the so-called down-sizing feature of present-day galaxies.
48: \end{abstract}
49: \keywords{galaxies: formation --- radiative transfer --- hydrodynamics}
50:
51: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
52: %(3)TEXT & Acknowledgements %
53: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
54:
55: \section{Introduction}
56: \label{intro}
57:
58: Recent advances of available computational resources enable us to
59: simulate the entire disk galaxy to investigate parsec scale structures in the
60: disk\citep{Wada07,Robertson07,Tasker08,Saitoh08}.
61: In addition, numerical techniques on Radiation HydroDynamics (RHD)
62: simulations are now being developed, especially in the field related to
63: the cosmic
64: reionization\citep{GA01,C2ray,Saru,Susa06,TSU3,Yoshida07,Qiu07,Whalen08,Altay08}.
65: Using these schemes, we are now ready to tackle the
66: numerical simulations of the detailed structures in galactic disks
67: coupled with radiation transfer.
68:
69: The importance of ultraviolet radiation transfer effects on the galaxy
70: formation/star formation in galaxies
71: have been pointed out by several authors.
72: First of all,
73: the observed star formation threshold in galactic disks \citep{K89,MK01}
74: could be explained by the self-shielding effects of galactic disks from
75: external ultraviolet radiation field\citep{Schaye01,Schaye04}.
76: \citet{Schaye04} obtained the physical state of the disks exposed to
77: external ultraviolet field, using one dimensional calculation with CLOUDY\citep{CLOUDY}.
78: Based upon the liner stability arguments\citep{Toomre},
79: he has demonstrated that the inner parts of the galactic disks shielded
80: against external radiation field by dust absorption could be
81: gravitationally unstable, whereas the unshielded outer parts of the disks
82: are stable. These results nicely explain the observed features of
83: galactic disks, however, it has not been tested
84: by multi-dimensional RHD simulations.
85:
86: Secondly, in the context of galaxy formation, radiative feedback by
87: ultraviolet background is expected to be very important especially for
88: low mass galaxies. The photoheating can heat the gas up to
89: $10^4$K, which prevent the gas from collapsing, in case the gravitational
90: potential of the dark matter halo is not deep enough to retain the
91: photoheated gas. Such feedback
92: effects are quoted as photoevaporation, which is well studied at various
93: levels\citep{UI84,Efs92, BR92,TW96,BL99,FT00,Gne00c,Kita00,Kita01,SU04a,SU04b}.
94: In addition, there are some evidences in dwarf galaxies that the
95: star formation is suppressed in ``hot'' ($\sim 10^4$K) phase \citep{YL97a,YL97b},
96: which infer that photoheating can suppress the
97: star formation in galaxies, even if the potential of
98: the dark halo is deep enough to prevent the gas from evaporating.
99: This issue could have a concern to the so-called down-sizing
100: problem in nearby galaxies\citep{Cowie96,downsizing}. They found
101: that old galaxies are massive, while the young ones are less massive.
102: Such trend continues to higher
103: redshift\citep[e.g.,][]{downsizing_subaru,Erb06,Reddy06,Papovich06}.
104: One of the possible interpretation of this feature is
105: that star formation proceed very rapidly in
106: massive galaxies, whereas it is a slow process in less
107: massive galaxies for some reason.
108: \citet{downsizing} also pointed out that these two groups are well
109: defined, i.e. they have a clear boundary at $\sim 10^{10}M_\odot$ in stellar mass. This critical mass scale is
110: too large to be related to the photoevaporation mechanism, however,
111: photoheating could still be the candidate to explain the down-sizing
112: mass if we can show the mechanism to suppress the star formation even
113: in halos in which the photoionized gas can be inherent.
114:
115:
116: In this paper, we examine the fragmentation of gas disks embedded in dark
117: halo potential under the ultraviolet radiation field, using
118: the recently developed RHD code RSPH\citep{Susa06}. This paper is organized as
119: follows. In the next section, the numerical scheme is briefly
120: summarized. In section \ref{setup}, the setup of numerical simulations
121: are described. We show the results of numerical simulations as well as
122: the analytic estimate in section \ref{results}.
123: Sections \ref{discussion} and \ref{conclusion} are
124: devoted to discussions and conclusion.
125:
126: \section{Methodology}
127: \label{scheme}
128: We perform numerical simulations by Radiation-SPH(RSPH) code \citep{Susa06}.
129: % In this section, we briefly summarize the scheme.
130: The code can compute the fraction of primordial chemical species e$^-$, H$^+$, H,
131: H$^-$, H$_2$, and H$_2^+$ by fully implicit time integration. It also
132: can deal with multiple sources of ionizing radiation, as well as
133: the radiation at Lyman-Werner band.
134:
135: Hydrodynamics is calculated by Smoothed Particle Hydrodynamics (SPH)
136: method. We use the version of SPH by \citet{Ume93} with the modification
137: on the SPH kernel function and symmetrization of equation of motions
138: according to \citet{SM93}. We also adopt the particle resizing
139: formalism by \citet{Thac00} in which the number of neighbor SPH
140: particles are kept almost constant without sudden changes.
141:
142: The non-equilibrium chemistry and radiative cooling
143: for primordial gas are calculated by the code
144: developed by \citet{SuKi00}, where H$_2$ cooling and
145: reaction rates are mostly taken from \citet{GP98}.
146: As for the photoionization process,
147: we employ so-called on-the-spot approximation
148: \citep{Spitzer78}.
149: In this paper, we also added the radiative cooling due to metals employing the
150: formula given in \citet{DM73}, assuming $Z=10^{-2}Z_\odot$.
151: The metallicity of nearby disk galaxies are normally larger than
152: $Z=10^{-2}Z_\odot$,
153: however, the observed metallicity of the IGM is $Z
154: \sim 10^{-2}Z_\odot-10^{-3}Z_\odot$ \citep[e.g.,][]{Son01}
155: which is the ingredient of the galaxies. Since we are interested in the
156: forming disk galaxies, we employ $Z=10^{-2}Z_\odot$ in the present set
157: of simulations.
158: Remark that the radiative transitions of heavy elements
159: are not the dominant process of radiative cooling for $\ga
160: 1000$K \citep{BH89,SU04a} in case $Z=10^{-2}Z_\odot$ is assumed.
161: H and H$_2$ molecules are the main coolant of such
162: gas\footnote{We have to keep in mind that for low temperature ($T\la
163: 500$K), or high density ($n_{\rm H}\ga 10^4{\rm cm^{-3}}$) realm, even
164: $Z\sim 10^{-4}Z_\odot$ could be important to decide the fate of cooling gas\citep{Omukai00}. }.
165:
166: The optical depth is integrated utilizing the
167: neighbor lists of SPH particles.
168: In our old scheme \citep{SU04a}, we create many grid points on the light
169: ray between the radiation source to an SPH particle.
170: In the present scheme, we do not create so many grids.
171: Instead, we create one grid point on the light ray per one SPH
172: particle on its upwind.
173: We find the neighboring 'upstream' particle for each SPH particle on its
174: line of sight to the source, which corresponds to the grid point.
175: Then the optical depth from the source to
176: the SPH particle is obtained by summing up the optical depth at the 'upstream'
177: particle and the differential optical depth between the two
178: particles. The more details are described in \citet{Susa06}.
179: We assess the optical depth for ionizing photons as well as the
180: Lyman-Werner band photons by the method described above.
181: In the present version of the code, the dust opacity is not included in
182: the calculation.
183:
184:
185: The code is tested for various standard RHD problems\citep{Susa06}, and
186: already applied to the issues on the radiative feedback effects of first
187: generation stars \citep{SU06,Susa07}.
188: We also take part in the code comparison
189: project with other radiation hydrodynamics codes \citep{TSU3} in
190: which we find reasonable agreements with each other.
191:
192: \section{Setup of Numerical simulations}
193: \label{setup}
194:
195: We perform numerical simulations of galactic gas disk embedded in a
196: dark halo potential. The dark halo potential is fixed as
197: \begin{equation}
198: \Phi_{\rm DH}(r) \equiv
199: -\left(\frac{27}{4}\right)^{1/2}
200: \left[
201: \frac{av_1^2}{\left(r^2+a_1^2\right)^{1/2}} +
202: \frac{av_2^2}{\left(r^2+a_2^2\right)^{1/2}}
203: \right],
204: \label{eq:DHpot}
205: \end{equation}
206: where $r$ denotes the distance form the center,
207: $a=1{\rm kpc},a_1=0.3{\rm kpc},a_2=5{\rm kpc}$ and $v_1=v_2=100{\rm km/s}$.
208: Thus the assumed potential is similar to the one employed by
209: \citet{Wada07}, except that the rotation velocity at given radius
210: is smaller by a factor
211: of 2,
212: since we are interested in the forming galaxies relatively
213: less massive than our Galaxy.
214:
215: In the present set of simulations, we assume initially uniform disk
216: (Fig. \ref{fig:model})
217: with slight perturbations (displacements of SPH particles are $\la$ 10 \%).
218: The initial thickness and radius of the disks, $H_{\rm i}$, $ R_{\rm
219: disk}$ are 100pc and 3kpc for all of the runs, respectively.
220: The initial velocity of the gas particles in the disk are assumed to be the
221: Kepler rotation velocity around the dark halo potential center.
222: Using equation (\ref{eq:DHpot}), the rotation velocity at the edge
223: of the disk $r=R_{\rm disk}$ is as large as $\sim 100$km/s. Thus, the
224: dark halo potential is deep enough to retain the gas even if it is
225: photoheated to $\sim 10^4$K.
226: The initial densities of the disks, $\rho_{\rm i}$, are changed depending
227: on the models, those are listed in Table \ref{tab:models}.
228:
229:
230: The light rays of external ultraviolet radiation field are assumed to be
231: perpendicular to the gas disk. We consider two
232: directions (upward/downward) as shown in Fig. \ref{fig:model}.
233: This particular choice of the ray-direction is an approximation
234: if we consider background radiation field exactly, since background
235: photons are coming from all angles. If we take into account
236: such effects, the boundary between the shielded and unshielded
237: regions are softened, which might change the self-shielding effects
238: slightly. However, as far as we keep the number of photons coming
239: into the disk unchanged, the self-shielding condition is also
240: almost unchanged, since it is determined by the balance between
241: the number of recombining hydrogen atoms and the number of photons.
242: The
243: flux of the field at the Lyman limit is given so that the mean intensity
244: at the midplane of the disk equals to $I_{\rm 21}\times 10^{-21} {\rm erg
245: s^{-1}cm^{-2}Hz^{-1}str^{-1}}$ in case the opacity of gas disk is ignored.
246: We consider an AGN type spectrum, which is proportional to $\nu^{-1}$.
247:
248: For all of the runs, we employ common physical/numerical parameters
249: of gas disk listed in Table\ref{tab:common}, while the model dependent
250: parameters are listed in Table \ref{tab:models}. As shown in the table,
251: we perform ten runs changing the density of the disk and the intensity
252: of the radiation field.
253:
254: %minimum temperature
255: %softening, limit
256: Since the resolution of the simulation is constrained by the mass of an SPH
257: particle, we have the critical density above which the gravitational
258: fragmentation
259: are not captured properly. According to \citet{BB97}, the density $n_{\rm H,res}$ is given as
260: \begin{equation}
261: n_{\rm H,res}=\frac{3}{4\pi m_{\rm p}}\left(\frac{5k_{\rm B}T_{\rm min}}{2 G
262: m_{\rm p}}\right)^3\frac{1}{\left(2 m_{\rm sph} N_{\rm nei}\right)^2}\label{eq:den_res}
263: \end{equation}
264: where $m_{\rm sph}, N_{\rm nei}, T_{\rm min}$ denote the mass of each
265: SPH particle, number of neighbor particles and the minimal temperature
266: set in the simulations, respectively.
267: The symbols $k_{\rm B},G,m_{\rm p}$ have ordinary meanings, i.e., they represent
268: Boltzmann constant, gravitational constant and proton mass, respectively.
269: %%
270: In the present set of simulations, $n_{\rm H,res}$ equals to $235{\rm
271: ~cm^{-3}}$, except the runs B/2,Br/2,B/8 and Br/8 in which the mass of a SPH
272: particles is larger than other regular runs
273: by a factor of two or eight (see Table \ref{tab:models}).
274: %%
275: The softening length of
276: gravitational interaction between SPH particles , $\epsilon$, is set so
277: as to satisfy
278: \begin{equation}
279: \epsilon=\frac{1}{2}\left(\frac{3N_{\rm nei}m_{\rm SPH}}{4\pi m_{\rm p}n_{\rm H, res}}\right)^{1/3}.\label{eq:softening}
280: \end{equation}
281: This expression guarantees that the self-gravity of a dense clump above $n_{\rm H, res}$ is softened.
282:
283: %Q value (fig6 cs=const 100K, 10000K)
284: Given these configurations, we can assess the Toomre's $Q$ value to
285: understand the stability of the initially uniform disk.
286: Toomre's $Q$ is defined as
287: \begin{equation}
288: Q=\frac{c_{\rm s}\kappa}{\pi G \Sigma}.
289: \end{equation}
290: Here $c_{\rm s},\kappa$ and $\Sigma$ denote the sound speed of the gas,
291: epicyclic frequency and the surface density of the disk. In case $Q>1$
292: holds, the disk is gravitationally stable because of the Coriolis
293: force/thermal pressure,
294: whereas it is unstable if $Q<1$ is satisfied\citep{Toomre}.
295:
296: The actual
297: $Q$ parameters for three models A, B and C (see Table \ref{tab:models}) are plotted in Fig.\ref{fig:Q}. Here
298: $c_s$ is assumed to be the sound speed corresponding to $T_{\rm min}=300$K.
299: Since we assume uniform disks for all models, the disks are more
300: unstable at the outer radii. In the present simulations, the radius of the disks are 3kpc for all models. Thus, there exist critical radii
301: above which the disks are unstable if those are cooled down to $T_{\rm
302: min}$,
303: although the unstable region is narrow for model A ( it is unstable only
304: at the edge of the uniform disk).
305: %In fact, in
306: %the absence of radiation field, those disks are cooled down to $300$K.
307: It is worth noting that if we employ $T_{\rm min}$ lower than $300$K, the
308: unstable region should expand, since the gas temperature will go down to
309: $T_{\rm min}$, that results in smaller $Q$.
310:
311: \section{Results}
312: \label{results}
313: \subsection{Stability of the disks without Radiation}
314: First, we present the results for $I_{21}=0$,
315: i.e. no external radiation field. The face-on/edge-on view of the
316: snapshots for the disks in models A, B and C are shown in
317: Fig.\ref{fig:montage_norad}.
318: The snapshots are taken at $t=300$Myr for model A, $t=120$Myr
319: for model B and $t=40$Myr for model C.
320: The particular choices of these times when the snapshots are taken
321: basically corresponds to the time when the time steps at dense gas clumps
322: collapsed below $n_{\rm H,res}$ becomes so short ( $\la 10^3$yr ) that
323: physical time in numerical computation evolve very slowly. Since the
324: growth time scale of gravitational instability is shorter in denser
325: disks, we take the earlier snapshot in model C than others.
326:
327: It is clear that the disks in all models fragment into small
328: filaments/knots. At the same time, the inner parts of the disks are
329: stable. As shown in Fig.\ref{fig:Q}, $Q$ values are
330: larger than unity at smaller radii, whereas they are smaller than unity at
331: larger radii. Therefore, the inner parts of the disks are stable, which is consistent
332: with the present results. In addition, the critical radii outside which
333: the disks are unstable predicted by Toomre criterion are also
334: consistent with numerical results. Indeed, the critical radii read from
335: Fig.\ref{fig:Q} are 2.8kpc, 1.6kpc and 0.8kpc for models A,B and C, those
336: are almost consistent with the boundary radii of smooth parts of the
337: disks.
338:
339: In Fig.\ref{fig:phase_norad}, the phase diagram ($n_{\rm H} - T$ plane) of the three runs at the
340: same snapshots as in Fig.\ref{fig:montage_norad} are shown.
341: The gas temperature in
342: all of the models are cooled efficiently,
343: which are almost close to $T_{\rm min}=300$K at high
344: density realm ($n_{\rm H} \ga 10^3 {\rm cm^{-3}}$). Such efficient
345: cooling justify the stability argument based upon the $Q$ values
346: plotted in Fig.\ref{fig:Q} in which $T=T_{\rm min}$ is assumed.
347:
348: These phase digram also show that very dense regions above $n_{\rm H,
349: res}$ are formed in all of the models. These dense regions correspond
350: to the fragments presented in Fig. \ref{fig:montage_norad}.
351:
352: \subsection{Effects of Ultraviolet Radiation on the Disk Stability}
353: Now, we show the snapshots of the runs with $I_{21}=1$ (Ar, Br and Cr)
354: in Fig.\ref{fig:montage_rad}, i.e. $t=300$Myr for model Ar, $t=120$Myr
355: for model Br and $t=40$Myr for model Cr.
356: The snapshots are synchronized to the corresponding ones in
357: Fig.\ref{fig:montage_norad}.
358: At a first glance we find clear difference from the runs without
359: ultraviolet radiation. In model Cr, we find small filaments and knots
360: as were found in model C, whereas we cannot find any sign of
361: fragmentation in models Ar and Br. In addition, the density of the disks
362: in models Ar and Br becomes lower than that in models A and B. We also
363: observe that the disks are geometrically thicker than the previous ones.
364: In Fig.\ref{fig:phase_rad}, the phase diagrams are shown for models Ar,
365: Br and Cr. In all of the runs, significant amount of materials are
366: photoheated up to $\ga 10^4$K in low density region, that contribute to
367: the disk thickening. Additionally,
368: no very dense gas component above $n_{\rm H,res}$ is found in models Ar
369: and Br, while we find it again in model Cr as was found in model C.
370: Thus, these results indicate that the presence of external ultraviolet radiation field suppress the
371: fragmentation of the disks with low density models (Ar and Br), although
372: it does not for high density model (Cr).
373:
374: \subsection{Self-Shielding from Ultraviolet Radiation}
375: The difference between low density models(Ar, Br)
376: and high density model(Cr) comes from the self-shielding
377: effects. Fig.\ref{fig:z-X} shows the density (top), temperature
378: (middle) and HI fraction $y_{\rm HI}$(bottom) distributions along the direction
379: perpendicular to the disks. In model Cr (right column), the gas near the
380: midplane ($z=0$) is neutral, cold and dense, because of the self-shielding.
381: On the contrary, in models Ar and Br,
382: cold and dense regions near the midplane is
383: relatively smaller than that in model Cr. To be more quantitative,
384: Fig.\ref{fig:pdf_tmp} shows the temperature probability
385: distribution functions (PDFs)for models Ar,Br,Cr (top) and A,B,C(bottom).
386: The temperature PDFs show the mass fraction of the disk
387: found within a logarithmic temperature bin $\Delta(\log T)=0.06$.
388: In the runs without radiation field (A,B and C), most of the mass is
389: condensed in the coldest phase at $T_{\rm min}$.
390: On the other hand,
391: the temperature PDFs are broader in the runs with radiation.
392: In models Ar and Br, most of the mass in the disks exist in the hot
393: phase ($\sim 10^4$ K), while cold phase ($< 10^3$K) is dominant in
394: model Cr.
395: Such large difference in temperature distribution results in different stability of
396: the disks.
397: According to Fig. \ref{fig:Q}, $Q$ values for models Ar and Br satisfy
398: $\ga 0.5$ if $r\le 3$kpc and $T=T_{\rm min}$. In actual
399: simulations, we find $T\sim 10^4$K, which means that $Q$ values are larger by a factor
400: of $\sqrt{10^4/300}=5.77$ than 0.5, i.e. $Q > 1$ is achieved everywhere
401: in the disks for these models. Thus,
402: these disks are stable, while the disk in model Cr is unstable because
403: of its coldness.
404:
405: A rough estimate of
406: self-shielding criterion for primordial gas has been derived by
407: \cite{SU00a}, who evaluate the shielded photoheating rate to
408: compare with the peak H$_2$ cooling rate below $10^{4}$K. Remark that
409: heavy elements are not the dominant coolant for $10^3 {\rm K} \la T\la
410: 10^4{\rm K}$, in case $Z\la 10^{-2}Z_\odot$\citep{SU04a}.
411:
412: Similar
413: argument is also found in \cite{Corbelli}, which describes the thermal
414: instability of the primordial gas.
415: According to \citet{SU00a}, the photoheating rate per unit volume
416: at the midplane of the disk for
417: the optically thick limit is given as $n_{\rm H,c}y_{\rm
418: HI,c}\mathcal{H}$ where
419: \footnote{The expression $\mathcal{H}$ is 2
420: times larger than the equation (A12) in \citet{SU00a}, since the
421: radiation field irradiate the right side of the disk as well as the
422: reverse side in the present simulations.}
423: \begin{equation}
424: \mathcal{H}=\frac{4\pi I_{\nu_{\rm L}}\nu_{\rm
425: L}\sigma_{\nu_{\rm
426: L}}}{3}\frac{\Gamma\left(\beta\right)}{1+\beta}\tau_{\nu_{\rm L}}^{-\beta}
427: \label{eq:photoheating}
428: \end{equation}
429: Here $n_{\rm H, c}$ and $y_{\rm HI,c}$ are the
430: number density of the hydrogen nucleus and HI fraction at the midplane
431: of the disk, $\sigma_{\nu_{\rm L}}$ is the photoionization cross section
432: at Lyman
433: limit, $\nu_{\rm L}$ is the Lyman limit frequency, $I_{\nu_{\rm L}}$
434: denotes the incident ultraviolet intensity at Lyman limit, $\Gamma(x)$
435: is the gamma function. $\beta$ used in this
436: equation is defined as $\beta\equiv 1+\left(\alpha-1\right)/3$,
437: where
438: $\alpha$ denotes the spectral index, i.e., $I_{\nu}\propto \nu^{-\alpha}$
439: is assumed. $\tau_{\nu_{\rm L}}$ denotes the optical depth at Lyman
440: limit, which is written as,
441: \begin{equation}
442: \tau_{\nu_{\rm L}} = \frac{\left\langle y_{\rm HI}\right\rangle N_{\rm
443: H} \sigma_{\nu_{\rm L}}}{2}.
444: \label{eq:tau}
445: \end{equation}
446: Here $N_{\rm H}$ denotes the column density of the disk, $\left\langle
447: y_{\rm HI}\right\rangle$ is the HI fraction averaged along the direction
448: perpendicular to the disk.
449: The peak H$_2$ cooling rate per unit volume below $10^4$K is described as
450: $n_{\rm H,c}^2\mathcal{C}_{\rm H_2}$, where
451: $\mathcal{C}_{\rm H_2}\simeq 10^{-26}{\rm erg~s^{-1}~cm^{3}}$\citep{SK87,SU04a}.
452: In addition, we assume hydrostatic equilibrium in the vertical direction.
453: As a result, the number density of hydrogen nucleus at the midplane of
454: the disk is related to the column density as follows:
455: \begin{equation}
456: n_{\rm H,c} = \frac{\pi G N_{\rm H}^2 \mu m_{\rm p}}{2 c_{\rm s}^2}.
457: \label{eq:zbalance}
458: \end{equation}
459: Here the gas disk is assumed to be isothermal with sound velocity
460: $c_{\rm s}$.
461: Letting the photoheating rate be equal to the
462: H$_2$ peak cooling rate, combined with the equations (\ref{eq:photoheating}), (\ref{eq:tau})
463: and (\ref{eq:zbalance}), we obtain the threshold column density $N_{\rm
464: H,sh}$, above which the photoheating is shielded enough for the midplane
465: to cool down to $\la 1000$K:
466: \begin{equation}
467: N_{\rm H,sh}=\left(\frac{2^{\beta+3} c_{\rm s}^2 I_{\nu_{\rm L}}\nu_{\rm
468: L}\sigma_{\nu_{\rm L}}^{1-\beta}\Gamma\left(\beta\right) }{3G \mu
469: m_{\rm p}\mathcal{C_{\rm
470: H_2}}y_{\rm HI,c}\left\langle y_{\rm HI}\right\rangle^\beta\left(\beta+1\right)}\right)^{1/\left(2+\beta\right)}
471: \end{equation}
472: Using the present assumptions ($\alpha=1$, $I_{21} = 1$), we have
473: \begin{eqnarray}
474: N_{\rm H,sh}\simeq 2.5 &\times& 10^{21} {\rm cm^{-2}} \nonumber \\
475: &\times &\left(\frac{I_{21}}{1}\right)^{1/3}\left(\frac{c_{\rm s}}{10{\rm
476: km s^{-1}}}\right)^{2/3}\left(\frac{\left\langle y_{\rm
477: HI}\right\rangle }{0.5}\right)^{-1/3}
478: \label{eq:Nsh}
479: \end{eqnarray}
480: Here we also assume $y_{\rm HI,c}=1$ in the above expression, since it
481: is true for all models Ar, Br and Cr.
482:
483: On the other hand, the initial column density of the disks in models
484: Ar, Br and Cr are $6.3\times 10^{20}{\rm cm^{-2}}$, $1.3\times
485: 10^{21}{\rm cm^{-2}}$ and $3.8\times 10^{21}{\rm cm^{-2}}$,
486: respectively. Thus, the numerical simulations indicate $1.3\times
487: 10^{21}{\rm cm^{-2}} \la N_{\rm H,sh} \la 3.8\times 10^{21}{\rm
488: cm^{-2}}$, which almost agrees with the rough analytic estimate shown above.
489:
490:
491: \subsection{Probability Distribution Functions of Density Field}
492: %pdf(fig2 for 2models, 2panels, including time evolution)
493: The differences among the six models (A,Ar,B,Br,C and Cr) are clarified
494: in Fig. \ref{fig:pdf} in terms of density probability distribution
495: functions(PDFs). Six panels represent the snapshots of the density PDFs
496: for six models. Here the density PDFs are defined as the mass fraction
497: found within a logarithmic density bin $\Delta(\log n_{\rm H})=0.18$.
498: The histograms drawn by thin lines represent the initial PDFs, while
499: the thick lines show the evolved ones.
500:
501: It is clear that in the models A,B,C, and Cr,
502: the PDFs extend beyond the resolution limit $n_{\rm H, res}$. Thus, dense
503: self-gravitating fragments formed in these disks, which would lead
504: to the star formation activities.
505:
506: It looks strange at first that very dense clumps form beyond $n_{\rm H,
507: res}$, although the gravitational force is softened for $n_{\rm H} \ga n_{\rm H,
508: res}$ (see eq. (\ref{eq:softening})).
509: In fact, even if gravitational force could be
510: neglected at $n_{\rm H} \ga n_{\rm H, res}$, matters accrete from the
511: outer envelope.
512: As a result, density (or
513: pressure) of the clump becomes higher than $n_{\rm H, res}$, in order to
514: compensate for the increasing ram/thermal pressure of accreting material.
515:
516: On the other hand, in the models Ar and Br, PDFs have very sharp cut-off below
517: $n_{\rm H, res}$ for the snapshot at $t=300$Myr(Ar) and $t=120$Myr(Br),
518: although models A and B have
519: clear fragmentation signature at the same snapshots.
520: Thus, the impression from the
521: montage of the disk is correct, i.e. the disk in models Ar and Br
522: do not fragment into dense clouds. This result could be interpreted
523: that the star formation in these disks are heavily suppressed by ultraviolet
524: radiation field.
525:
526:
527: \subsection{Convergence of PDF}
528: The convergence of density PDF is checked for models B and
529: Br. We perform runs with larger SPH particles mass by a factor of two
530: and a factor of eight.
531: These runs are
532: tagged
533: as B/2, Br/2 and B/8, Br/8
534: (Table \ref{tab:models}). Fig.\ref{fig:conv} shows the density PDFs for models
535: B,B/2,B/8 (top) and models Br,Br/2,Br/8 (bottom) at
536: $t=40,100,120$Myr. In the half resolution runs (B/2 and Br/2), the density
537: resolution limits are four times smaller than those in B and Br, because
538: $n_{\rm H, res}$ inversely proportional to the square of mass resolution
539: (see equation (\ref{eq:den_res})). Similarly, the density resolution
540: limits in B/8 and Br/8 are 64 times smaller than B and Br.
541: The PDF histograms basically agree
542: very well below $n_{\rm H,res}$ as expected.
543:
544: In the runs B/8 and Br/8, the density resolution limits are so low
545: that the high density regions ($\ga 100{\rm cm^{-2}}$) are not captured
546: properly. As a result, it seems to be difficult to distinguish the PDFs for
547: models B/8 and Br/8 from each other.
548: On the other hand, the results from the runs B/2 and Br/2
549: indicate the same conclusion as we found in the runs B and Br,
550: since the density resolution is enough to capture the self-shielded dense
551: regions.
552: Thus, the disk in B/2 fragments due
553: to gravitational instability, whereas that in Br/2 is stable. Therefore we
554: conclude that the resolution of the present regular simulations
555: are enough to capture
556: the fragmentation of the disks under the assumptions we employed,
557: whereas in lower resolution runs such as B/8 and Br/8 are not able to
558: describe the present physical situation.
559:
560: \section{Discussion}
561: \label{discussion}
562: It should be emphasized that present results directly prove the
563: suppression of star formation activities by ultraviolet background radiation
564: {\it in halos with $T_{\rm vir} > 10^4$K.} It has already been
565: known that the star formation is suppressed in less massive halos with
566: $T_{\rm vir} \la 10^4$K, since the gas in the dark halo potential
567: evaporates due to the high thermal pressure of photoheated gas
568: \citep{UI84,Efs92,BR92,TW96,BL99,FT00,Gne00c} if the self-shielding effect is not important \citep{Kita00,Kita01,SU04a,SU04b}.
569: On the other hand, in the present simulations, the gas in the dark halo
570: do not evaporate because of the deep gravitational
571: potential. We find that even in such halos, star formation could be heavily
572: suppressed in case gas is configured to form disks with low column density.
573:
574: Another important result found in the present calculation is the
575: presence of a clear boundary in the column density of the disk, below which
576: star formation is heavily suppressed.
577: According to the numerical results, the critical column density is
578: $N_{\rm H}\sim 1-4 \times 10^{21}{\rm cm^{-2}}$. This threshold is very interesting
579: mainly for two reasons.
580: First, although the present simulations are performed with fixed dark
581: halo mass, we can try to convert the critical column density into dark halo
582: mass.
583: The mass of the uniform disk with given $N_{\rm H}$ and disk radius $R_{\rm disk}$ is
584: \begin{equation}
585: M_{\rm disk} = 6\times 10^9M_\odot
586: \left(\frac{R_{\rm disk}}{10{\rm kpc}}\right)^2
587: \left(\frac{N_{\rm H}}{2.5\times10^{21}{\rm cm^{-2}}}\right).
588: \end{equation}
589: Therefore,
590: if the mass of the host dark halo is 7($\simeq\Omega_{\rm B}/\Omega$) times the disk mass,
591: the critical dark halo mass is $\sim 4 \times 10^{10}M_\odot$.
592: This value is still smaller than the critical scale found by \citet{downsizing}
593: by a few factor, but
594: if we take into account the internal feedback effects such as UV
595: radiation from internal sources (AGN, massive stars) or supernova, the
596: threshold might account for the observed critical down-sizing mass.
597: We also point out that if we start the simulation from cosmological
598: setup, the threshold mass could be raised more, since photoheating might
599: be able to penetrate deeper into the disk because of initially less
600: dense configurations.
601: Thus, inclusion of such physics as well as starting simulations from
602: cosmological setup are necessary to obtain more precise understanding of
603: star formation in forming disk galaxies.
604:
605: Secondly, the critical column density is as large as the star formation
606: threshold found in local disk galaxies \citep{K89,MK01}. In fact,
607: the observed star formation threshold of galactic disks has been
608: investigated by \citet{Schaye01,Schaye04} from theoretical side.
609: He suggest that the galactic disk
610: could be gravitationally stable if it is unshielded from the external
611: UV radiation, since Q value of the photoheated disk
612: easily exceeds unity.
613: As a result, gravitational fragmentation of
614: the disk is suppressed, so the star formation activities do.
615: \citet{SD08} performed simple hydrodynamics simulations taking into
616: account the star formation threshold column density using effective
617: equation of state in multiphase medium,
618: although they do not solve radiation transfer equations explicitly. They
619: found the threshold column density is as large as
620: $4M_\odot {\rm pc}^{-2}$ (Fig.4 and 5 in \citet{SD08}), which is
621: $5\times 10^{20}{\rm cm^{-2}}$ in cgs unit. This value is smaller
622: than that obtained in this paper by a few factor($\sim 2.5 \times 10^{21}{\rm
623: cm^{2}}$).
624: The basic mechanism they propose to suppress the star formation of the
625: disk is same as the one found in the present simulations, except that we do
626: not include the effects of dust extinction required
627: especially for present-day disk galaxies.
628: The gas disk is more
629: easily self-shielded by dust absorption, since the dust opacity at the
630: Lyman limit frequency is as large as HI continuum for solar metallicity.
631: Thus, it is it is reasonable that we have larger critical column
632: density than obtained in \citet{SD08}.
633: On the other hand, in the present simulations, we assume $Z=10^{-2}Z_\odot$.
634: Therefore, the dust opacity at Lyman limit is smaller than HI continumm
635: opacity by two orders of magnitude. As a result, dust opacity has much
636: smaller impact on the self-shielding effects at such low metallicity.
637: In any case, we would dare to mention that present
638: results almost succeeded to probe the presence of star formation threshold column density
639: proposed by \citet{Schaye04} utilizing the full 3D radiation
640: hydrodynamics simulations. More realistic calculations for present-day
641: galaxies including the effects of dust extinction are left for future
642: works.
643:
644: It is also worth noting that star formation rate (SFR) of the disk above the
645: threshold column density is consistent with the observed value, although
646: it is difficult to compare the present results directly with observed
647: SFR, since we do not take into account the local
648: stellar feedbacks, the effects of dust particles, metal enrichment
649: followed by the radiative cooling by abundant metals.
650: Despite of such issues,
651: SFR cloud be obtained by the density PDF assuming the stars are formed in
652: dense clumps within local free-fall time. It is given as
653: \begin{equation}
654: \dot{\Sigma_*}=\epsilon_{\rm c}\sqrt{G\rho_{\rm c}}\Sigma_{\rm disk}f_{\rm c}
655: \end{equation}
656: where $\epsilon_{\rm c}$ is the star formation efficiency, $\rho_{\rm
657: c}$ is the threshold density above which the gas is converted to
658: stars, $\Sigma_{\rm disk}$ denotes the surface density of the disk, and
659: $f_{\rm c}$ denotes the mass fraction of gas in the disk condensed into
660: dense clumps with $\rho > \rho_{\rm c}$. $f_{\rm c}$ is the quantity
661: which can be obtained by integrating the density PDF above $\rho_{\rm
662: c}$. In the present set of simulations with radiative feedback, model Cr
663: is the only one in which star formation is expected since self-shielded
664: cold fragments emerges in the run. If we use $\epsilon=0.1$ and $n_{\rm
665: H,res}$ to assess the threshold density $\rho_{\rm c}$,
666: we obtain $f_{\rm c}\simeq 0.066$. Using the surface density of the disk
667: in model Cr, the SFR is evaluated as
668: $\dot{\Sigma_*}=0.03 {M_\odot {\rm yr}^{-1} {\rm kpc}^{-2}}$.
669: Although we cannot discuss the dependence of SFR on $\Sigma_{\rm disk}$,
670: the value is consistent with the observation \citep{K98}. This
671: reasonable agreement infer the validity of the present numerical models.
672:
673:
674:
675: We also point out another standpoint on this issue by
676: \citet{SU00a}, in which they suggested that
677: self-shielding from ultraviolet background
678: could be a key mechanism to determine the morphology of
679: galaxies\citep{SU00b}, although their arguments based upon 1-dimensional
680: radiation hydrodynamics calculations. Unfortunately, it is impossible to
681: relate the present results to morphology bifurcation of galaxies,
682: since we assume a disk by hand in our simulations performed so far. From
683: this point of view, again we need to perform simulations from cosmological
684: initial conditions.
685:
686: % PDF recovery ?
687:
688: %High frequency photons?
689:
690:
691: \section{Conclusion}
692: \label{conclusion}
693: In this paper, we perform radiation hydrodynamics simulations on the
694: fragmentation of galactic disks under the ultraviolet
695: radiation background.
696: We find that ultraviolet radiation field strongly suppress
697: the star formation in the disks in case the photoheating is not shielded
698: enough. We emphasize that this suppression is different from
699: photoevaporation effect,
700: %which is important for low mass galaxies with
701: %$T_{\rm vir}\la 10^4$K,
702: because the rotation velocities at the outer boundary of the disks
703: in the present set of
704: simulations are $\sim 100$km/s,
705: which are fast enough to retain the photoheated gas.
706: In our simulations, we find a threshold column density of the
707: disk ($\sim 10^{21}{\rm cm^{-2}}$) above which the fragmentation is not suppressed. It is similar to the star formation threshold
708: column density observed in nearby galaxies.
709: Presence of such critical threshold would
710: be one of the reason for the so-called down-sizing problem in nearby
711: galaxies.
712:
713:
714:
715: \bigskip
716: HS thanks the anonymous referee and
717: K. Wada for helpful comments and discussions.
718: HS also thanks
719: M. Umemura, M. Ohta and H. Sato for their continuous encouragements.
720: The analysis has been made with computational facilities
721: at Center for Computational Science in University of Tsukuba, Konan
722: University, and Rikkyo University.
723: This work was supported in part by Inamori Research Foundation as well
724: as Ministry of Education, Culture,
725: Sports, Science, and Technology (MEXT), Grants-in-Aid, Specially
726: Promoted Research 16002003.
727:
728: %%%%%%%%%%%%%%%%%%
729: % (4) Appendices %
730: %%%%%%%%%%%%%%%%%%
731: %No Appendix
732: %%%%%%%%%%%%%%%%%
733: % (5)References %
734: %%%%%%%%%%%%%%%%%
735: \begin{thebibliography}{}
736: %\bibitem[Abel, Bryan, Norman(2000)]{Abel00}
737: %Abel, T., Bryan, G. L., \& Norman, M. L. 2000, \apj, 540, 39
738: %\bibitem[Ahn \& Shapiro (2006)]{AS06}
739: %Ahn, K. \& Shapiro, P.R. 2006, astro-ph/0607642
740: \bibitem[Altay et al.(2008)]{Altay08} Altay, G., Croft,
741: R.~A.~C., \& Pelupessy, I.\ 2008, ArXiv e-prints, 802, arXiv:0802.3698
742: %\bibitem[Baraffe, Heger \& Woosely (2001)]{Baraffe01}
743: %Baraffe, I., Heger, A. \& Woosely, S.E. \ 2001, \apj, 550, 890
744: \bibitem[Barkana \& Loeb(1999)]{BL99}
745: Barkana, R., \& Loeb, A.\ 1999, \apj, 523, 54
746: %\bibitem[Barnes \& Hut (1986)]{BH86}
747: %Barnes, J. \& Hut, P., 1986, Nature, 324, 446
748: \bibitem[Bate \& Burkert(1997)]{BB97}
749: Bate, M.~R., \& Burkert, A.\ 1997, \mnras, 288, 1060
750: \bibitem[Boehringer \& Hensler(1989)]{BH89}
751: Boehringer, H., \& Hensler, G.\ 1989, \aap, 215, 147
752: %\bibitem[Bromm, Coppi \& Larson (2002)]{BCL02}
753: %Bromm, V., Coppi, P., \& Larson, R. \ 2002, \apj, 564, 23
754: %\bibitem[Bromm \& Larson (2004)]{BL04}
755: %Bromm, V \& Larson, R. 2004, \araa, 42, 79
756: \bibitem[Babul \& Rees(1992)]{BR92} Babul, A.~\& Rees, M.~J.\ 1992, \mnras, 255, 346
757: \bibitem[Baldry et al.(2006)]{Baldry06}
758: Baldry, I.~K., Balogh, M.~L., Bower, R.~G., Glazebrook, K., Nichol, R.~C., Bamford, S.~P.,
759: \& Budavari, T.\ 2006, \mnras, 373, 469
760: %\bibitem[Ciardi, Ferrara \& Abel (2000)]{CFA00}
761: %Ciardi, B., Ferrara, A., \& Abel, T. 2000, \apj, 533, 594
762: \bibitem[Corbelli et al.(1997)]{Corbelli} Corbelli, E., Galli,
763: D., \& Palla, F.\ 1997, \apjl, 487, L53
764: \bibitem[Cowie et al.(1996)]{Cowie96}
765: Cowie, L.~L., Songaila, A., Hu, E.~M., \& Cohen, J.~G.\ 1996, \aj, 112, 839
766: \bibitem[Dalgarno \& McCray (1972)]{DM73} Dalgarno, A. \& McCray, A. 1972, ARA\&A, 10, 375
767: \bibitem[Draine \& Bertoldi (1996)]{DB96}
768: Draine, B. T., \& Bertoldi, F. 1996, \apj, 468, 269
769: %\bibitem[Dubinski (1996)]{Dubinski96}
770: %Dubinski, J. 1996, NewA, 1, 133
771: \bibitem[Efstathiou(1992)]{Efs92} Efstathiou, G.\ 1992, \mnras, 256, 43P
772: \bibitem[Erb et al.(2006)]{Erb06}
773: Erb, D.~K., Steidel, C.~C., Shapley, A.~E., Pettini, M., Reddy, N.~A.,
774: \& Adelberger, K.~L.\ 2006, \apj, 647, 128
775: \bibitem[Ferrara \& Tolstoy(2000)]{FT00} Ferrara, A.~\& Tolstoy, E.\ 2000, \mnras, 313, 291
776: \bibitem[Ferland(2000)]{CLOUDY} Ferland, G.~J.\ 2000, Revista
777: Mexicana de Astronomia y Astrofisica Conference Series, 9, 153
778: %\bibitem[Fukushige, Makino \& Kawai (2005)]{FMK05}
779: %Fukushige, T., Makino, J. \& Kawai, A. 2005, astro-ph/0504407
780: \bibitem[Galli \& Palla (1998)]{GP98}
781: Galli D. \& Palla F. 1998, \aap, 335, 403
782: %\bibitem[Glover \& Brand (2001)]{GB01}
783: %Glover, S \& Brand, P. 2001, \mnras, 321, 385
784: \bibitem[Gnedin (2000)]{Gne00c} Gnedin, N.~Y.\ 2000, \apj, 542, 535
785: \bibitem[Gnedin \& Abel (2001)]{GA01}
786: Gnedin, N. Y.\& Abel, T , 2001, NewA, 6, 437
787: %\bibitem[Grief \& Bromm (2006)]{GB06}
788: %Grief, T. \& Bromm, V. 2006, preprint, astro-ph/0604367
789: %\bibitem[Hiaman, Rees \& Loeb (1997)]{HRL97}
790: %Haiman, Z., Rees, M. \& Loeb, A., 1997, \apj, 476, 458
791: %\bibitem[Hieneman et al. (2005)]{Hienemann05}
792: %Heinemann, T., Dobler, W., Nordlund, A.,Brandenburg, A.\ 2006, \aa ,448, 731
793: %\bibitem[Hosokawa \& Inutsuka (2005)]{Hosokawa05}
794: %Hosokawa, T. \& Inutsuka, S. \ 2005, \apj, 623, 917
795: \bibitem[Iliev et al.(2006)]{TSU3} Iliev, I.~T., et al.\
796: 2006, \mnras 371, 1057
797: %\bibitem[Jhonson \& Bromm (2006)]{JB06}
798: %Jhonson, J.L. \& Bromm, V. 2006, \mnras, 366, 247
799: \bibitem[Kauffmann et al.(2003)]{downsizing} Kauffmann, G., et
800: al.\ 2003, \mnras, 341, 54
801: %\bibitem[Kang \& Shapiro (1992)]{KS92}
802: %Kang, H., \& Shapiro, P., \apj, 386, 432
803: \bibitem[Kennicutt(1989)]{K89} Kennicutt, R.~C., Jr.\ 1989, \apj, 344, 685
804: \bibitem[Kennicutt(1998)]{K98} Kennicutt, R.~C., Jr.\ 1998,
805: \apj, 498, 541
806: \bibitem[Kitayama et al.(2000)]{Kita00} Kitayama, T., Tajiri,
807: Y., Umemura, M., Susa, H., \& Ikeuchi, S.\ 2000, \mnras, 315, L1
808: \bibitem[Kitayama et al.(2001)]{Kita01} Kitayama,T., Susa, H., Umemura, M., \& Ikeuchi, S. 2001, \mnras, 326, 1353
809: %\bibitem[Kitayama et al. (2004)]{Kitayama04}
810: %Kitayama,T., Yoshida, N., Susa, H. \& Umemura, M., 2004, \apj, 613, 631
811: \bibitem[Kodama et al.(2004)]{downsizing_subaru} Kodama, T., et al.\
812: 2004, \mnras, 350, 1005
813: %\bibitem[Larson (1969)]{Larson69}
814: %Larson, R. 1969, \mnras, 145, 271
815: \bibitem[Martin \& Kennicutt(2001)]{MK01} Martin, C.~L., \& Kennicutt, R.~C., Jr.\ 2001, \apj, 555, 301
816: %\bibitem[Machacek, Bryan \& Abel (2001)]{Macha01}
817: %Machacek, M.E.,~ Bryan, G. L. \& Abel, T. \ 2001, \apj, 548, 509
818: %\bibitem[Meisinger, Bryan, \& Haiman (2006)]{MBH06}
819: %Meisinger,A., Bryan, G., \& Haiman, Z. \ 2006, \apj, 648, 835
820: %\bibitem[Maselli, Ferrara \& Ciardi (2003)]{MFC03}
821: %Maselli, A., Ferrara, A. \& Ciardi, B. 2003, \mnras, 345, 379
822: \bibitem[Mellema et al.(2006)]{C2ray} Mellema, G., Iliev,
823: I.~T., Alvarez, M.~A., \& Shapiro, P.~R.\ 2006, New Astronomy, 11, 374
824: %\bibitem[Monaghan (1992)]{Monaghan92}
825: %Monaghan, J.J. 1992, ARA\&A, 30, 543
826: %\bibitem[Nagakura \& Omukai (2005)]{Nagakura05}
827: %Nagakura, T. \& Omukai, K. 2005, \mnras, 363, 1378
828: %\bibitem[Nakamoto, Umemura, \& Susa (2001)]{NUS01}
829: %Nakamoto, T., Umemura, M., \& Susa, H.\ 2001, \mnras, 321, 593
830: %\bibitem[Nakamura \& Umemura(1999)]{NU99}
831: % Nakamura F. \& Umemura M. 1999, \apj, 515, 239
832: %\bibitem[Nakamura \& Umemura(2001)]{NU01}
833: % Nakamura, F., \& Umemura, M. 2001, \apj, 548, 19
834: %\bibitem[Oh \& Haiman (2002)]{OhH02}
835: %Oh, P \& Haiman, Z. \ 2002, \apj, 569, 558
836: \bibitem[Omukai(2000)]{Omukai00} Omukai, K.\ 2000, \apj, 534,
837: 809
838: %\bibitem[Omukai \& Nishi (1998)]{ON98}
839: %Omukai, K. \& Nishi, R. \ 1998, \apj, 508, 141
840: %\bibitem[Omukai \& Nishi (1999)]{ON99}
841: %Omukai, K. \& Nishi, R. \ 1999, \apj, 518, 64
842: %\bibitem[O'shea, Abel, Whalen \& Norman (2005)]{Oshea05}
843: %O'Shea, B.~W., Abel, T., Whalen, D., Norman, M.~L. \ 2005, \apj, 628, 5L
844: %\bibitem[O'shea \& Norman (2006)]{Oshea06}
845: %O'Shea, B.~W. \& Norman, M.~L. \ 2006, \apj, 648, 31
846: \bibitem[Papovich et al.(2006)]{Papovich06}
847: Papovich, C., et al.\ 2006, \apj, 640, 92
848: %\bibitem[Penston(1969)]{Penst69}
849: %Penston, M.~V.\ 1969, \mnras, 144, 425
850: \bibitem[Qiu et al.(2007)]{Qiu07}
851: Qiu, J.-M., Feng, L.-L., Shu, C.-W., \& Fang, L.-Z.\
852: 2007, New Astronomy, 12, 398
853: \bibitem[Reddy et al.(2006)]{Reddy06}
854: Reddy, N.~A., Steidel, C.~C., Fadda, D., Yan, L., Pettini, M., Shapley, A.~E., Erb, D.~K., \& Adelberger, K.~L.\ 2006, \apj, 644, 792
855: %\bibitem[Ricotti, Gnedin, \& Shull(2001)]{Ricotti01}
856: %Ricotti, M. Gnedin, N.~Y., Shull, M. \ 2001, \apj, 560, 580
857: \bibitem[Rijkhorst et al.(2006)]{Saru} Rijkhorst, E.-J., Plewa, T., Dubey, A., \& Mellema, G.\ 2006, \aap, 452, 907
858: \bibitem[Robertson \& Kravtsov(2007)]{Robertson07} Robertson, B., \& Kravtsov, A.\ 2007, ArXiv e-prints, 710, arXiv:0710.2102
859: \bibitem[Saitoh et al.(2008)]{Saitoh08} Saitoh, T.~R., Daisaka,
860: H., Kokubo, E., Makino, J., Okamoto, T., Tomisaka, K., Wada, K.,
861: \& Yoshida, N.\ 2008, ArXiv e-prints, 802, arXiv:0802.0961
862: \bibitem[Schaye(2001)]{Schaye01} Schaye, J.\ 2001, \apjl, 562, L95
863: \bibitem[Schaye(2004)]{Schaye04} Schaye, J.\ 2004, \apj, 609, 667
864: \bibitem[Schaye \& Dalla Vecchia(2008)]{SD08} Schaye, J., \& Dalla Vecchia, C.\ 2008, \mnras, 383, 1210
865: \bibitem[Shapiro \& Kang (1987)]{SK87}
866: Shapiro, P.R., \& Kang, H., 1987, \apj, 318, 32
867: \bibitem[Spitzer(1978)]{Spitzer78}
868: Spitzer, L. Jr. 1978, in Physical Processes in the Interstellar Medium
869: (John Wiley \& Sons, Inc. 1978)
870: \bibitem[Songaila (2001)]{Son01} Songaila, A. 2001, \apj, 561, 153L
871: \bibitem[Steinmetz \& M\"uller(1993)]{SM93} Steinmetz, M.~\& M\"uller, E.\ 1993, \aap, 268, 391
872: \bibitem[Susa \& Kitayama (2000)]{SuKi00}
873: Susa, H. \& Kitayama, T. 2000, \mnras, 317, 175
874: %\bibitem[Susa et al.(1998)]{SUNY98}
875: %Susa, H., Uehara, H., Nishi, R., \& Yamada, M. 1998,
876: %Prog. Theor. Phys., 100, 63
877: \bibitem[Susa \& Umemura(2000a)]{SU00a} Susa, H., \& Umemura, M.\ 2000, \apj, 537, 578
878: \bibitem[Susa \& Umemura(2000b)]{SU00b} Susa, H., \& Umemura, M.\ 2000, \mnras, 316, L17
879: \bibitem[Susa \& Umemura(2004a)]{SU04a}
880: Susa, H. \& Umemura, M. 2004, \apj, 600, 1
881: \bibitem[Susa \& Umemura(2004b)]{SU04b}
882: Susa, H. \& Umemura, M. 2004, \apj, 610, 5L
883: \bibitem[Susa (2006)]{Susa06}
884: Susa, H. 2006, \pasj, 455, 58
885: \bibitem[Susa \& Umemura(2006)]{SU06}
886: Susa, H. \& Umemura, M. 2006, \apj, 645, 93L
887: \bibitem[Susa(2007)]{Susa07} Susa, H.\ 2007, \apj, 659, 908
888: \bibitem[Tasker \& Bryan(2008)]{Tasker08} Tasker, E.~J., \& Bryan, G.~L.\ 2008, \apj, 673, 810
889: \bibitem[Thacker et al. (2000)]{Thac00}
890: Thacker, R.J., Tittley, E.R., Pearce, F.R., Couchman, H.M.P. \& Thomas, P.A. 2000,
891: \mnras 319, 619
892: \bibitem[Thoul \& Weinberg(1996)]{TW96} Thoul, A.~A.~\&
893: Weinberg, D.~H.\ 1996, \apj, 465, 608
894: \bibitem[Toomre(1964)]{Toomre} Toomre, A.\ 1964, \apj, 139, 1217
895: %\bibitem[Uehara \& Inutsuka(2000)]{UI00}
896: %Uehara, H. \& Inutsuka, S., 2000, \apj, 531, L91
897: \bibitem[Umemura \& Ikeuchi(1984)]{UI84} Umemura, M.~\& Ikeuchi, S.\ 1984, Progress of Theoretical Physics, 72, 47
898: \bibitem[Umemura(1993)]{Ume93} Umemura, M.\ 1993, \apj, 406, 361
899: %\bibitem[Yoshida et al. (2003)]{Yoshida03}
900: \bibitem[Wada \& Norman(2007)]{Wada07} Wada, K., \& Norman, C.~A.\ 2007, \apj, 660, 276
901: %\bibitem[Whalen
902: %\& Norman(2006)]{Whalen06} Whalen, D., \& Norman, M.~L.\ 2006, \apjs, 162, 281
903: \bibitem[Whalen \& Norman(2008)]{Whalen08} Whalen, D.~J., \& Norman, M.~L.\ 2008, \apj, 672, 287
904: \bibitem[Young \& Lo (1997a)]{YL97a} Young, M. \& Lo, Y. 1997, \apj, 476, 127
905: \bibitem[Young \& Lo (1997b)]{YL97b} Young, M. \& Lo, Y. 1997, \apj, 490, 710
906: %Yoshida, N., Abel, T., Hernquist, L. \& Sugiyama, N., 2003, \apj, 592, 645
907: %\bibitem[Yoshida et al. (2006)]{Yoshida06}
908: %Yoshida, N., Oh, P., Kitayama, T., \& Hernquist, L., 2006, preprint, astro-ph/0610819
909: \bibitem[Yoshida et al.(2007)]{Yoshida07} Yoshida, N., Oh, S.~P.,
910: Kitayama, T., \& Hernquist, L.\ 2007, \apj, 663, 687
911: \end{thebibliography}
912:
913: %\end{document}
914: \begin{table}[ht]
915: \begin{center}
916: \caption{Common Parameters\label{tab:common}}
917: \begin{tabular}{ccc}
918: \tableline\tableline
919: symbol & numerical value employed & description \\
920: \tableline
921: $N_{\rm nei}$ & 50 & Number of neighbor particles \\
922: $H_{\rm i}$ & 100pc & Initial disk thickness \\
923: $R_{\rm disk}$ & 3kpc & Initial disk radius \\
924: $T_{\rm i}$ & $10^4$K & Initial disk temperature \\
925: $T_{\rm min}$ & $300$K & minimum temperature \\
926: \tableline
927: \end{tabular}
928: \end{center}
929: \end{table}
930:
931: \begin{table}[ht]
932: \begin{center}
933: \caption{Model dependent Parameters \label{tab:models}}
934: \begin{tabular}{ccccccc}
935: \tableline\tableline
936: Label& $I_{21}$ & $\rho_{\rm i}[M_\odot/{\rm pc^{-3}}]$ & simulated time
937: & \# of SPH particles &$n_{\rm H,res}[{\rm cm^{-3}}]$&$\epsilon[{\rm pc}]$\\
938: \tableline
939: A & 0 & 0.05 & 300 Myr & $1.28\times 10^6$ & 235 & 3.05\\
940: B & 0 & 0.1 & 120 Myr & $2.56\times 10^6$ & 235 & 3.05\\
941: C & 0 & 0.3 & 40 Myr & $7.68\times 10^6$ & 235 & 3.05\\
942: Ar & 1 & 0.05 & 350 Myr & $1.28\times 10^6$ & 235 & 3.05\\
943: Br & 1 & 0.1 & 200 Myr & $2.56\times 10^6$ & 235 & 3.05\\
944: Cr & 1 & 0.3 & 40 Myr & $7.68\times 10^6$ & 235 & 3.05\\
945: \tableline
946: B/2 & 0 & 0.1 & 120 Myr & $1.28\times 10^6$ &58.8& 6.09\\
947: Br/2 & 1 & 0.1 & 120 Myr & $1.28\times 10^6$ &58.8& 6.09\\
948: \tableline
949: B/8 & 0 & 0.1 & 120 Myr & $3.20\times 10^5$ &3.67& 24.4\\
950: Br/8 & 1 & 0.1 & 120 Myr & $3.20\times 10^5$ &3.67& 24.4\\
951: \tableline
952: \end{tabular}
953: \end{center}
954: \end{table}
955:
956: %%%%%%%%%%%%%%%%%%%%%%%
957: % (6) Figure Captions %
958: %%%%%%%%%%%%%%%%%%%%%%%
959:
960: \setcounter{figure}{0}
961:
962: \clearpage
963: %
964: \begin{figure}[ht]
965: \begin{center}
966: \includegraphics[angle=0,width=8cm]{f1.eps}
967: \caption{Schematic view of the assumed configuration. Rotating gas disk is
968: embedded in a dark halo potential, which is irradiated by the
969: ultraviolet background radiation. The light rays are assumed to be
970: perpendicular to the disk.
971: }\label{fig:model}
972: \end{center}
973: \end{figure}
974: %
975: \begin{figure}[ht]
976: \begin{center}
977: \includegraphics[angle=0,width=8cm]{f2.eps}
978: \caption{
979: $Q(=c_{\rm s}\kappa/\pi G \Sigma)$ value is plotted as functions of radial
980: coordinate in the disk. In case $Q > 1$ is satisfied in a certain region
981: of a disk, the region is gravitationally stable, otherwise not.Three
982: curves correspond to the three models A,B and C assuming $T=T_{\rm min}$.
983: }\label{fig:Q}
984: \end{center}
985: \end{figure}
986:
987: %
988:
989: \begin{figure}[ht]
990: \begin{center}
991: \includegraphics[angle=0,width=15cm]{f3small.eps}
992: \caption{Face-on/edge-on view of the disks in three models A B
993: and C are shown. These are the snapshots taken at $t=300$Myr for model
994: A, $t=120$Myr for model B and $t=40$Myr for model C, respectively. The
995: color represents the density contrast, as shown in the legend.
996: }\label{fig:montage_norad}
997: \end{center}
998: \end{figure}
999:
1000: \begin{figure}[ht]
1001: \begin{center}
1002: \includegraphics[angle=0,width=15cm]{f4.eps}
1003: \caption{Phase diagram for models A,B and C. Each dot represents each
1004: SPH particle. The vertical dotted lines represent $n_{\rm H,res}$.
1005: }\label{fig:phase_norad}
1006: \end{center}
1007: \end{figure}
1008:
1009: %
1010: \begin{figure}[ht]
1011: \begin{center}
1012: \includegraphics[angle=0,width=15cm]{f5small.eps}
1013: \caption{Same as Fig.\ref{fig:montage_norad}, except that the models are
1014: Ar, Br and Cr.
1015: }\label{fig:montage_rad}
1016: \end{center}
1017: \end{figure}
1018: %
1019: \begin{figure}[ht]
1020: \begin{center}
1021: \includegraphics[angle=0,width=15cm]{f6.eps}
1022: \caption{Same as Fig.\ref{fig:phase_norad}, except that models are Ar,
1023: Br and Cr.
1024: }\label{fig:phase_rad}
1025: \end{center}
1026: \end{figure}
1027: %
1028: \begin{figure}[ht]
1029: \begin{center}
1030: \includegraphics[angle=0,width=13cm]{f7small.eps}
1031: \caption{Various quantities in the disks are plotted as functions of
1032: vertical coordinate, $z$. The horizontal axes denote $z$, whereas
1033: vertical axes denote density (top), temperature (middle) and HI
1034: fraction $y_{\rm HI}$ (bottom). These are the snapshots taken at
1035: $t=300$Myr for model A, $t=120$Myr for model B and $t=40$Myr for model
1036: C, respectively.
1037: }\label{fig:z-X}
1038: \end{center}
1039: \end{figure}
1040: %
1041: \begin{figure}[ht]
1042: \begin{center}
1043: \includegraphics[angle=0,width=13cm]{f8.eps}
1044: \caption{Temperature Probability Distribution Functions(PDFs) are shown
1045: for six models Ar,Br,Cr,A,B and C.
1046: Horizontal axes denote the temperature, whereas the
1047: vertical axes show the differential mass fraction $\Delta M/M_{\rm total}$
1048: found within a logarithmic temperature bin $\Delta(\log T)=0.06$.
1049: Here $M_{\rm total}$ is the total gas mass of the
1050: disk. The histograms denote temperature PDFs at $t=300$Myr for models
1051: Ar and A, $t=120$Myr for models Br and B, $t=40$Myr for models Cr and C,
1052: respectively.
1053: }\label{fig:pdf_tmp}
1054: \end{center}
1055: \end{figure}
1056: %
1057: \begin{figure}[ht]
1058: \begin{center}
1059: \includegraphics[angle=0,width=15cm]{f9.eps}
1060: \caption{Density Probability distribution functions(PDFs) are shown for models A,B,C,Ar,Br and Cr. Horizontal axes denote the density, whereas the
1061: vertical axes show the differential mass fraction $\Delta M/M_{\rm total}$
1062: found within a logarithmic density bin $\Delta(\log n_{\rm H})=0.18$.
1063: The histogram drawn by thin solid lines denote the initial PDFs in
1064: the models. The thick solid lines denote the evolved PDFs at $t=300$Myr for models A and
1065: Ar, $t=120$Myr for models B and Br, $t=40$Myr for models C and Cr,
1066: respectively. The vertical dotted thin lines show the resolution limits of
1067: the runs.
1068: }\label{fig:pdf}
1069: \end{center}
1070: \end{figure}
1071: %
1072: \begin{figure}[ht]
1073: \begin{center}
1074: \includegraphics[angle=0,width=15cm]{f10.eps}
1075: \caption{PDFs in models B and Br are compared to the results from the
1076: low resolution runs(B/2,Br/2,B/8,Br/8). The axes are the same as
1077: Fig.\ref{fig:pdf}. Top three panels :the PDFs for models Br(black histogram)
1078: , Br/2(red), Br/8(blue) at $t=40$Myr(left), $100$Myr(middle) and $120$Myr(right).Density resolution limits, $n_{\rm H, res}$, corresponding to the runs
1079: Br, Br/2 and Br/8 are shown as the vertical dotted lines with colors
1080: same as the histgrams.
1081: Bottom panels : PDFs for models B, B/2 and B/8 at three epochs. The
1082: colors are same as the top panels.
1083: }\label{fig:conv}
1084: \end{center}
1085: \end{figure}
1086: %
1087:
1088: \end{document}
1089: