1: %\documentclass[manuscript]{aastex}
2: \documentclass[preprint]{aastex}
3:
4: \def\kmps{\ifmmode {\rm\,km\,s^{-1}}\else
5: ${\rm\,km\,s^{-1}}$\fi}
6: \newcommand{\kms}{\kmps}
7: \def\ergcms{\ifmmode {\rm\,ergs\,cm^{-2}\,s^{-1}}\else
8: ${\rm\,ergs\,cm^{-2}\,s^{-1}}$\fi}
9: \newcommand{\fluxunit}{\ergcms}
10: \def\ergcmsa{\ifmmode {\rm\,ergs\,cm^{-2}\,s^{-1}\,\AA^{-1}}\else
11: ${\rm\,ergs\,cm^{-2}\,s^{-1}\,\AA^{-1}}$\fi}
12: \newcommand{\fluxd}{\ergcmsa}
13: \def\ergs{\ifmmode {\rm\,ergs\,s^{-1}}\else
14: ${\rm\,ergs\,s^{-1}}$\fi}
15: \newcommand{\lum}{\ergs}
16: \def\lya{\ifmmode {\mbox{Ly}\alpha}\else
17: {Ly$\alpha$}\fi}
18: \def\mdot{\ifmmode {\, M_\odot\, \rm yr^{-1}}\else
19: $\, M_{odot}\,\rm yr^{-1}$\fi}
20:
21: \slugcomment{ver. 7 May 2007}
22: \shorttitle{Spectroscopy of Extended \lya\ Sources}
23: \shortauthors{Saito et al.}
24:
25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26: %% Beginning of the article itself
27:
28: \begin{document}
29:
30: \title{
31: Deep Spectroscopy of Systematically Surveyed
32: Extended Lyman-$\alpha$ Sources at $z\sim 3-5$
33: \footnote{Based on observations made with ESO telescope at
34: the Paranal Observatory, under the programme ID 074.A-0524}
35: \footnote{Based on data collected at the Subaru Telescope,
36: which is operated by the National Astronomical Observatory of Japan.}
37: }
38:
39: \author{
40: Tomoki Saito\altaffilmark{3,4}, Kazuhiro Shimasaku\altaffilmark{3},
41: Sadanori Okamura\altaffilmark{3}, Masami Ouchi\altaffilmark{5,6},
42: Masayuki Akiyama\altaffilmark{7},
43: Michitoshi Yoshida\altaffilmark{8},
44: Yoshihiro Ueda\altaffilmark{9}
45: }
46: \altaffiltext{3}{Department of Astronomy, University of Tokyo,
47: 7-3-1 Hongo, Bunkyo, Tokyo 113-0033, Japan}
48: \altaffiltext{4}{Present Address:
49: Dark Cosmology Centre, Niels Bohr Institute, University of Copenhagen,
50: Juliane Maries Vej 30, 2100 Copenhagen \O, Denmark,
51: e-mail: tomoki@dark-cosmology.dk}
52: \altaffiltext{5}{Space Telescope Science Institute,
53: 3700 San Martin Drive, Baltimore, MD 21210, USA}
54: \altaffiltext{6}{Hubble Fellow}
55: \altaffiltext{7}{Subaru Telescope, National Astronomical Observatory of Japan,
56: North A'ohoku Place, Hilo, HI 96720, USA}
57: \altaffiltext{8}{Okayama Astrophysical Observatory,
58: National Astronomical Observatory of Japan}
59: \altaffiltext{9}{Department of Astronomy, Kyoto University ,
60: Sakyo-ku, Kyoto 606-8502, Japan}
61:
62:
63: \begin{abstract}
64: Spatially extended \lya\ sources that are faint and/or compact
65: in continuum are candidates for extremely young ($\lesssim 10^7\rm yrs$)
66: galaxies at high redshifts.
67: We present medium-resolution ($R\sim 2000$) spectroscopy of
68: such extended \lya\ sources found in our previous study
69: at $z\sim 3-5$, using VLT/VIMOS.
70: The deep spectroscopy showed that all 18 objects we observed
71: have large equivalent widths (EWs) exceeding 100\AA.
72: For about 30\% of our sample (five objects),
73: we identified conspicuous asymmetry on the profiles of the \lya\ line.
74: They show broad wing emission components on the red side,
75: and sharp cut-off on the blue side of the \lya\ line.
76: Such asymmetry is often seen in superwind galaxies known to date,
77: and also consistent with a theoretical prediction of superwinds.
78: In fact, one of them show systematic velocity structure on the
79: 2-dimensional spectrum suggesting the existence of superwind activity.
80: There are eight objects ($8/18\sim 40\%$) that have large EWs exceeding
81: 200\AA, and no clear signature of superwind activities.
82: Such large EWs cannot be explained in terms of photo-ionization
83: by a moderately old ($>10^7\rm yrs$) stellar population,
84: even with a top-heavy IMF or an extremely low metallicity.
85: These eight objects clearly show a positive correlation between
86: the \lya\ line luminosity and the velocity width.
87: This suggests that these eight objects are good candidates for
88: forming-galaxies in a gas-cooling phase.
89: \end{abstract}
90:
91: \keywords{galaxies:formation --- galaxies:high redshift}
92:
93: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
94: %%%% Beginning of the body of the article %%%%
95: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
96:
97: \section{Introduction}
98: \label{introduction}
99: Galaxies undergoing their initial star formation are predicted
100: to have strong \lya\ emission \citep{partridge, cf93}.
101: Recent imaging surveys using narrowband filters tuned to the
102: redshifted \lya\ line have successfully identified a large
103: number of \lya\ emitters
104: \citep[LAEs, e.g. ][]{ch98, hu98, fynbo, fynbo03, rhoads00, rhoads03,
105: ouchi03, ouchi05, shimasaku03, kodaira, maier, dawson04, kashikawa, iye06}.
106: Some of them are fairly faint in rest-frame UV continuum and have
107: large \lya\ equivalent widths exceeding 240\AA\ \citep[hereafter MR02]{mr2002}.
108: Such large equivalent widths suggest that they are possibly
109: in the initial phases ($\ll 10^7 \rm yrs$) of star formation.
110: The faintness in UV continuum also suggests that their stellar
111: masses are fairly small, implying that a significant star formation
112: has not yet occurred.
113:
114: In such early phases of galaxy formation (age of $\lesssim 10^7\rm yrs$),
115: (proto)galaxies are predicted to radiate spatially extended \lya\
116: emissions. At least three mechanisms are currently
117: proposed for extended \lya\ emissions; {\it i}) cooling radiation
118: from gravitationally heated primordial gas infalling into a dark halo
119: potential
120: \citep{haiman00, fardal, furlanetto04, bd03, db06, dijkstra06a, dijkstra06b},
121: {\it ii}) resonant scattering of \lya\ photons from the central
122: (hidden) ionizing source
123: \citep{moller98, hr01, weidinger04, weidinger05, laursen06},
124: {\it iii}) starburst-driven galactic winds
125: \citep{ts00, tsk01, ohyama, geach05}.
126:
127: Recent narrowband imaging surveys for high-$z$ \lya\ emitters (LAEs) have
128: found several tens of extended \lya\ sources with no
129: significant UV continuum sources sufficient to ionize such a large
130: amount of \ion{H}{1} gas
131: \citep[\lya\ blobs, or LABs:][]{keel99,s00,francis01,m04,palunas04,dey05,nilsson}
132: However, there are still only restricted number of extended \lya\ sources
133: known to date, while thousands of normal LAEs are already known.
134: Most of the extended \lya\ sources known to date are identified in
135: narrowband imaging surveys, which cover very narrow redshift range.
136: Even the largest sample made by Matsuda et al. (hereafter M04),
137: which is constructed from a narrowband survey targeted on LABs
138: of \citet[hereafter S00]{s00},
139: covers a redshift range of only $z=3.06^{+0.04}_{-0.03}$.
140:
141: Due to the lack of a systematic sample and subsequent follow-up studies,
142: nature of such extended \lya\ sources remains unknown.
143: Even for the most famous objects, two LABs of S00,
144: the origins of the extended \lya\ emission are not fully clear.
145: For example, observational studies
146: of S00's LABs are made by many astronomers, ranging from radio to
147: X-ray wavelengths \citep[e.g.][]{chapman01,chapman04,bzs04,geach05,m06}.
148: These studies suggest that obscured ionizing sources such as starburst
149: regions or AGNs are associated with LABs.
150: Detailed optical spectroscopy of the LABs suggests that
151: outflow from the central sources are responsible for the extended
152: \lya\ emission \citep{ohyama, wilman}, while numerical simulation
153: shows that cooling radiation from infalling gas combined with
154: a central ionizing source can also explain the spatial extent,
155: line profiles, and velocity widths of the \lya\ emission
156: \citep{dijkstra06b}. There is also a detailed observational study
157: by \citet{bower04} suggesting that extended \lya\ emission of LAB
158: cannot be simply explained with a single mechanism such as a
159: galactic wind or cooling radiation.
160: They are likely to be massive systems harboring galaxy formation
161: sites \citep{m06}, although the true nature of the \lya\ emission
162: are still unclear.
163:
164: In order to construct a systematic sample of extended \lya\ sources covering
165: a wide range of redshift, we have carried out a deep, wide-field
166: imaging survey of a blank-field, Subaru/{\it XMM-Newton} Deep Field,
167: using Subaru Telescope and seven intermediate-band (IA) filters.
168: This survey enabled us to construct a sample of 41 extended \lya\ sources
169: located at $z\sim 3-5$, and also to show that this kind of objects
170: commonly exist in the early universe far beyond $z\sim 3$
171: \citep[][hereafter Paper1]{paper1}.
172: We here present a deep follow-up spectroscopy of our photometric
173: sample of Paper1 using high-resolution grism of VIMOS on VLT.
174: This gives a three times deeper exposure and a several times higher
175: resolution than our previous spectroscopy presented in Paper1.
176: We constructed a spectroscopic sample of 18 spatially extended \lya\
177: sources with large equivalent widths, located at redshifts
178: of $3.3\lesssim z\lesssim 4.7$. This sample allows us to make a
179: quantitative analysis in terms of equivalent width, velocity width,
180: and \lya\ line luminosity, as well as the velocity structure traced
181: by the \lya\ line profile.
182:
183: The rest of this paper is organized as follows.
184: We briefly introduce our photometrically selected sample of
185: extended \lya\ sources in \S\ref{sample}, based on the previous
186: observations in Paper1.
187: In \S\ref{spec} we present the details of our spectroscopic observations
188: and data reduction.
189: The results of the spectroscopy thus obtained and a discussion based on
190: them are presented in \S\ref{res}.
191: Finally, we summarize the conclusions in \S\ref{con}.
192:
193: All magnitudes are in the AB \citep{oke, fsi95}.
194: We use the standard $\Lambda$-dominated CDM cosmology with
195: $\Omega_M = 0.3$, $\Omega_\Lambda = 0.7$,
196: and $H_0 = 100h = 70\kms\rm Mpc^{-1}$, unless otherwise noted.
197:
198:
199: \section{Our sample of extended \lya\ sources}
200: \label{sample}
201: The targets of our current work were selected from the photometric
202: sample constructed in Paper1, based on an intermediate-band imaging survey
203: (Subaru proposal ID S02B-163: Kodaira et al.) with Subaru Prime-focus
204: Camera (Suprime-Cam) mounted on 8.2m Subaru Telescope.
205: The field of this survey is a blank field called Subaru/{\it XMM-Newton}
206: Deep Field South (SXDF-S) located at $\alpha = 02^h18^m00^s$,
207: $\delta = -05^\circ 25'00''$ (J2000).
208: The selection of extended \lya\ sources was based on photometry
209: in seven intermediate (IA) bands centered at $\sim 5000-7100\rm\AA$
210: \citep{hayashino}, the $B$ band, and a broadband redward of the
211: \lya\ line ($R$, $i'$, or $z'$ band).
212: The broadband data were taken in a large project of Subaru Telescope,
213: Subaru/{\it XMM-Newton} Deep Survey \citep[SXDS: ][]{sekiguchi04}.
214: The details of the selection are reported in Paper1.
215: This sample consists of 41 objects, all of which are basically
216: \lya\ emitters with $B$-dropout, located at redshifts of
217: $3.24\lesssim z\lesssim 4.95$.
218: Their photometric properties are listed in Table 3 of Paper1.
219:
220: Their \lya\ line luminosities measured with the IA-band photometry
221: are typically $\sim 10^{42}-10^{43}\lum$.
222: They are spatially extended in \lya\ line, and faint and/or compact
223: in the continuum redward of the \lya.
224: They can be divided into two minor classes;
225: ``continuum-compact'' objects and ``continuum-faint'' objects.
226: About a half of them, 22 objects, can be regarded as point-sources
227: in the redward continuum, and are classified as continuum-compact objects.
228: They show $B$ band dropout, which is consistent with Lyman break feature
229: of distant galaxies at $z\gtrsim 3$. The remaining 19 objects are
230: continuum-faint objects, which are fainter than $3\sigma$ in the
231: redward continuum band.
232: Their continuum break between the redward band and the $B$ band
233: cannot be measured since they are not detected in the $B$ band.
234: Some of these continuum-faint objects appear to be larger than
235: the PSF size in the redward continuum band,
236: but the sizes measured here are not reliable due to the low signal-to-noise
237: (S/N) ratio.
238:
239: Since our IA bands have relatively low sensitivities to the line emission
240: compared with usual narrowbands, we can select objects with large equivalent
241: widths (EWs). Our selection criteria correspond to the EWs
242: greater than $\sim 55\rm\AA$ in the rest frame.
243: We roughly estimated the EWs with the photometric data, and found
244: that the largest value is $\sim 700\rm\AA$.
245: The \lya\ line luminosities are also relatively large,
246: ranging from several $\times 10^{42}\lum$ to $\sim 1\times 10^{43}\lum$.
247: This corresponds to the luminous-end of the luminosity function (LF)
248: of normal LAEs at similar redshifts.
249: Figure \ref{fig:lf} shows a comparison of the LF between our sample
250: and a sample of LAEs at $z\simeq 3.7$ from a narrowband survey
251: of SXDF (Ouchi et al. in preparation).
252: The completeness of Ouchi et al.'s is $\sim 70-80\%$ at $\le 24.2$ mag
253: ($\gtrsim 3.4\times 10^{42}\lum$) and $\sim 50\%$ at $\le 24.7$ mag
254: ($\gtrsim 5.4\times 10^{42}\lum$).
255: This comparison may suggest that a significant fraction ($\sim 10-20\%$)
256: of luminous LAEs are spatially extended \lya\ sources.
257: However, one object in our sample at $z\simeq 3.7$,
258: IB10-90651, is not included in this LAE sample down to 24.7 mag.
259: This is due to the low surface brightness of this object and the
260: relatively shallow exposure of the narrowband image.
261:
262: Figure \ref{fig:z-hist} shows the redshift distribution of our objects,
263: based on the IA imaging (two double-counted objects are not eliminated).
264: The detection limit of each IA band is also plotted in this figure.
265: This shows that extended \lya\ sources commonly exist in the early
266: universe, almost independently of redshift.
267: The size distribution of our objects is shown in Figure \ref{fig:sizedist}.
268: There is no significant redshift dependence in the spatial extent.
269: Note that the low sensitivity to the line emission also affects
270: the apparent spatial extents of our objects, especially in higher
271: redshift bands. The diffuse emission at the outskirts becomes more
272: difficult to be detected, and the apparent sizes tend to become smaller.
273: This is also suggested from Figure \ref{fig:z-hist}, in which the number of
274: objects declines with redshift.
275: True sizes may be larger than our estimation, and the size distribution
276: may resemble that of M04's LAB sample.
277: Detailed simulations of such effects are presented in Paper1.
278:
279: From the sample of 41 photometrically-selected objects,
280: we selected 21 objects for further follow-up spectroscopy described below.
281: About a half of them (11 objects) are continuum-compact objects,
282: and another half (10 objects) are continuum-faint obejcts.
283: As noted below, we could not obtain the spectra of three objects:
284: IB11-101786 (continuum-faint), IB13-62009 (continuum-faint), and
285: IB14-62116 (continuum-compact). Excluding these three,
286: we obtained a spectroscopic sample of 18 objects (10 continuum-compact,
287: and 8 continuum-faint objects).
288: Photometric properties of the 18 objects are briefly summarized in
289: Table \ref{tab:phot}, and their images are shown in Figure \ref{fig:images}.
290: Seven of them were spectroscopically confirmed to be high-$z$ \lya\
291: emitters (Paper1).
292:
293:
294: \section{VIMOS spectroscopy}
295: \label{spec}
296: \subsection{Observations}
297: \label{spec:obs}
298: We used Visible Multi Object Spectrograph (VIMOS) mounted on the 8.2m
299: Very Large Telescope (VLT) UT3 ``Melipal'' to take deep spectra
300: for the objects selected above, under the program ID 074.A-0524.
301: The observations were made during three dark nights of the visitor
302: mode run on 2004 November 6 to 9 (UT).
303: Two masks were used in multi-object spectroscopy (MOS) mode of VIMOS,
304: covering the 21 objects. Two of them, IB11-101786 and IB13-62009, were
305: too faint to be detected with the current depth of our spectroscopy.
306: For another one object, IB14-62116, the \lya\ line was almost completely
307: overlapped with a strong sky emission line that cannot be correctly
308: subtracted. We thus obtained spectra for 18 objects.
309: In this run, the HR-Orange grism and the GG435 order-cut filter were used.
310: This setting gives a spectral resolution
311: $R=\Delta\lambda/\lambda \approx 2160$ ($\Delta\lambda \approx 2.8\rm\AA$
312: at $\lambda = 6000\rm\AA$) with a $1''.0$ wide slit.
313: This spectral resolution corresponds to a velocity resolution
314: of $\approx 140\kms$.
315:
316: VIMOS has a wide FoV consisting of four quadrants of $\sim 7'\times 8'$
317: each, which can cover about 1/3 of a single Suprime-Cam FoV.
318: The sky distribution of our objects and the VIMOS FoV are shown
319: in Figure \ref{fig:skydist}.
320: The two MOS masks cover ten and eight objects, respectively.
321: The integration time was 9.5 hrs and 8.0 hrs, respectively.
322: During this run, the shutter unit for the second (north-eastern)
323: quadrant often did not work correctly, and several exposures
324: (30 minutes each) were lost. As a consequence, the integration
325: time for each quadrant ranges from 6.5 hrs to 9.5 hrs.
326: The seeing size varied during the three nights from $\lesssim 0''.5$ to
327: $\sim 2''.4$.
328: We flagged the frames with seeing worse than $\sim 1''.0$, so that
329: the effective integration times of objects
330: range from 4.0 hrs to 7.0 hrs.
331: e standard star frames were taken at the end of the run.
332: We used LTT3864, an F-dwarf located at $\alpha = 10^h 32^m 13.90^s$,
333: $\delta = -35^\circ 37' 42''.4$ (J2000) as a standard star.
334:
335: \subsection{Primary reduction and flux measurement}
336: \label{spec:flux}
337: The primary reduction was made with the VIMOS pipeline recipes 1.0
338: provided by ESO. This software package performs flat-fielding,
339: distortion correction, wavelength calibration, and sky subtraction.
340: The spectral resolution was measured to be $R\approx 2200$,
341: using the sky emission lines. This resolution is sufficient to resolve
342: the \lya\ line profiles of our objects, which are known to have velocity
343: widths less than several $100\kms$ (\S 4.2 of Paper1).
344: After primary reduction using the VIMOS pipeline, we performed
345: flux calibration and further analysis using our original software.
346: For the two-dimensional spectra, we smoothed the data with Gaussian
347: kernel with FWHM of $4\times 5$ pixels, which corresponds, to
348: $3.22\rm\AA\, (\mbox{wavelength axis}) \times
349: 1''.0\, (\mbox{spatial axis})$.
350: To measure the fluxes and the velocity width on the one-dimensional
351: spectra, we used 5-pixel smoothing, roughly corresponding to the
352: current spectral resolution.
353: We then measured the central wavelengths, fluxes, and velocity
354: widths of the \lya\ lines.
355: The spectra thus obtained are shown in Figure \ref{fig:spec}.
356:
357: \subsection{Luminosity and velocity width of the \lya\ line}
358: \label{spec:lum-dv}
359: The \lya\ fluxes were calculated by correcting for the fraction
360: of their line fluxes collectable with the $1''.0$ wide slitlets
361: (slit-loss correction).
362: The slit-loss correction was made by the following procedure by
363: combination of photometric and spectroscopic data.
364: We first assumed that (1) the intrinsic continuum spectra are flat within
365: the bandpass of the IA filters, and that
366: (2) the continuum levels are equal to that measured with the
367: broadband redward of the \lya\ line.
368: Then we calculated the effect of the IGM absorption to the continuum
369: using the redshift information obtained with the spectroscopy
370: by following the formulation of \citet{madau95}.
371: The pure \lya\ line fluxes can be estimated by subtracting the continuum
372: contribution from the IA photometry.
373: We made this estimation of the total flux within an automatic aperture,
374: $F_{total}$, and the flux within the slit we used, $F_{slit}$.
375: The fraction of the collectable flux was then estimated by taking the ratio
376: of the flux values measured with the both apertures, i.e., the slit-loss
377: correction was applied by multiplying the line fluxes (from spectra)
378: with $\frac{F_{total}}{F_{slit}}$.
379:
380: The \lya\ line fluxes thus obtained, $F\rm(Ly \alpha)$, are shown in
381: Table \ref{tab:spec},
382: together with the values before the slit-loss correction.
383: The errors listed in Table \ref{tab:spec} are estimated by integrating
384: the $1\sigma$ noise level of the spectra over the range we integrated
385: the flux of the \lya\ line.
386:
387: The \lya\ line luminosities were directly
388: calculated from these flux values by using the redshifts obtained
389: spectroscopically. We fitted a Gaussian function to the line profile
390: of each object, and defined the line center as the central wavelength
391: of the Gaussian. Then we determined the redshift from the wavelength of
392: the line center, and calculated the luminosity, $L\rm(Ly\alpha)$, from
393: the flux and the luminosity distance at the redshift.
394: The velocity widths (FWHM) were measured simultaneously by using the
395: redshifts obtained by the Gaussian fitting.
396: We defined the zero velocity as the line center described above,
397: and calculated the line-of-sight velocities of the wavelengths
398: at which the flux density becomes a half of the peak value.
399: The upper and lower limits of the FWHM were estimated by measuring the
400: full velocity widths at $\mbox{(half maximum)}\pm 1\sigma$.
401:
402: \subsection{Equivalent widths of the \lya\ line}
403: \label{spec:ew}
404: Since most of our objects do not show any continuum emission,
405: we cannot obtain their EWs solely from the spectral data.
406: Instead, we used the photometric data (broadband redward of the \lya\ line)
407: to estimate the continuum flux densities.
408: The continuum flux densities were estimated using a $2''\phi$ aperture.
409: Since we measure the EWs using the total flux of the \lya\ line,
410: this may cause an underestimate of the UV flux densities.
411: For continuum-compact objects, this does not affect the results
412: since the spatial extent is small enough. For continuum-faint objects,
413: the spatial extents are generally larger than the PSF size (see Fig.2
414: of Paper1). However, they are fairly faint in continuum ($<3\sigma$) and
415: the measurement of their size is not so reliable.
416: This implies that using a larger aperture make the measurement
417: more sensitive to the sky fluctuation.
418:
419: Then we divided the \lya\ line fluxes obtained spectroscopically
420: (slit-loss corrected) with the continuum
421: fluxes obtained photometrically. Their rest-frame EWs, $\rm EW_{rest}$,
422: are also shown in Table \ref{tab:spec}.
423: All the objects in our sample have fairly large EWs with a median
424: value of $\rm EW_{rest}\approx 210\AA$.
425: For 9 out of 18 objects, the $\rm EW_{rest}$ exceed 200\AA.
426: Eight out of these nine objects are continuum-faint objects,
427: having yet larger EWs exceeding 240\AA,
428: i.e., all of the continuum-faint objects in our VIMOS sample have large EWs.
429: Another object with $\rm EW_{rest}>200\AA$, IB12-58572,
430: is a continuum-compact object, but its continuum level is the
431: lowest among all the continuum-compact objects in our VIMOS
432: sample ($\sim 3.5\sigma$). We plot the EWs as a function of
433: the \lya\ line luminosity in Figure \ref{fig:lew}.
434: The nine objects with $\rm EW_{rest}>200\AA$ apparently show
435: a positive correlation between the luminosity and the EW.
436: This is thought to be an artifact caused by the detection limits
437: in the redward continuum bands. For these nine objects (or at least
438: eight continuum-faint objects), the EWs we measured are likely
439: to be still larger, since the continuum levels are close to the
440: sky fluctuation level. For the remaining nine objects, our measurement
441: of the EWs is somewhat more accurate than for objects with large EWs,
442: since the continuum levels exceed $3.7\sigma$.
443:
444: The upper and lower limits for the EWs are estimated as follows.
445: Since the denominators in the calculation of EWs are very
446: small values, and the errors in EWs are dominated by the photometric
447: errors in the broadband images.
448: We estimated the error using the limiting magnitudes of the broadband
449: images used to estimate the continuum levels:
450: 27.4 mag for $R$ and 27.0 for $i'$ ($3\sigma$, $2''\phi$ aperture).
451: The upper and lower limits of the EWs, $\rm EW_{upper}$ and $\rm EW_{lower}$,
452: were estimated from
453: \begin{displaymath}
454: {\rm EW_{upper}} = (F(\lya)+{\rm err}[F(\lya)])
455: /(f_\lambda - {\rm err}[f_\lambda]),
456: \end{displaymath}
457: and
458: \begin{displaymath}
459: {\rm EW_{lower}} = (F(\lya)-{\rm err}[F(\lya)])
460: /(f_\lambda + {\rm err}[f_\lambda]),
461: \end{displaymath}
462: where $F(\lya)$ is the \lya\ line flux, $f_\lambda$ is the continuum
463: flux density, and err[] denotes the 1$\sigma$ noise level.
464:
465: Note that our measurements of EWs are likely to underestimate the
466: intrinsic value, since we did not make any correction for the
467: IGM absorption of the \lya\ emission.
468:
469: \subsection{Wing component analysis}
470: \label{spec:wing}
471: Although most of our objects have relatively small velocity widths,
472: some objects have broad wing emission of the \lya\ line.
473: The \lya\ line profile of IB10-90651, for example, has broad wing
474: emission on the red side, while the blue side shows a relatively
475: sharp cut-off. Such features of the line profile must reflect the
476: kinematics of \lya\ emitting gas, and should act as a diagnostic
477: of inflow / outflow activities in the system.
478: In order to identify such high velocity wing components in the \lya\
479: line profiles, we performed a quantitative analysis for all the objects
480: in our spectroscopic sample.
481: To identify such faint wing components, we need to have spectra of
482: sufficiently high S/N ratios.
483: We therefore used 8-pixel smoothing after integrating the spectra
484: along the slit direction. Although the 8-pixel smoothing reduces the
485: spectral resolution to $\sim 5\rm\AA$, this resolution is still high
486: enough to investigate the gas dynamics with the line profile.
487:
488: The procedure of the analysis is similar to that described in
489: \S\ref{spec:lum-dv},
490: except the smoothing width of 8-pixel. After obtaining improved S/N ratios
491: by smoothing the one-dimensional spectra, we fitted a simple Gaussian
492: function to the \lya\ line profile of each object. We then plotted the
493: excess of the spectrum from the Gaussian function against the line-of-sight
494: velocity.
495: If the excess of the spectrum greater than $2\sigma$ has
496: a velocity width greater than $250\kms$ (corresponds to 8-pixels),
497: we here refer to it as the ``excess''. If the excess appears from
498: within the $2\sigma$-width of the Gaussian peak and continuous
499: beyond $\pm 500\kms$ from the line center, we define it as ``wing emission''.
500: The plots thus obtained are shown in Figure \ref{fig:spec},
501: and the results are also summarized in Table \ref{tab:spec}.
502: For at least five objects, IB10-17108, IB10-32162, IB10-54185, IB10-90651,
503: and IB14-52102, we identified high velocity wing emission on the
504: red side of the \lya\ line. On the other hand, none of our objects
505: shows significant wing emission on the blue side.
506: Just two objects, IB12-30834 and IB12-58572, {\em marginally} show
507: a blue wing.
508:
509: \section{Results and discussion}
510: \label{res}
511: Summarizing the results obtained in the previous section,
512: we categorize our objects according to the spectral properties.
513: We can classify our objects in two ways, i.e., in terms of EWs
514: and wing emission in the line profile.
515: In terms of EWs, we divide our VIMOS sample into two EW classes.
516: The moderate EW class has $\rm 100\AA < EW <200\AA$ while the
517: large EW class has $\rm EW>200\AA$ (see Figure \ref{fig:lew}).
518: The value of $\rm EW=200\AA$ is fairly large for ordinary starbursts.
519: The large EW class and the small EW class roughly correspond to
520: continuum-compact objects and continuum-faint objects, respectively.
521: In terms of the line profile, we can also classify our objects into
522: two groups, i.e., those with broad asymmetric wing on the red side,
523: and those with no clear signature of such wing emission.
524:
525: \subsection{Objects with moderate EWs}
526: \label{res:SB}
527: The nine objects in the smaller EW group, IB08-86220, IB10-17108,
528: IB10-32162, IB10-54185, IB12-21989, IB12-71781, IB13-96047,
529: IB14-47257, and IB14-52102, have EWs ranging between 100\AA\ and
530: 200\AA. This EW range is the regime of stellar photo-ionization
531: with a solar metallicity, an upper mass cut-off of $100M_\odot$,
532: and an IMF slope up to 1.5 \citep{cf93}.
533: The EWs of these objects thus can be explained in terms of
534: photo-ionization by moderately old stellar populations,
535: not necessarily requiring an extremely top-heavy IMF or a zero-metallicity.
536: The most conservative explanation for these objects are thus
537: ordinary starbursts. The large spatial extent of \lya\ emission
538: can be understood as a result of resonant-scattering of \lya\
539: photons \citep[e.g.][]{moller98, laursen06}.
540:
541: Assuming that the ionizing sources for these objects are
542: ordinary starburst. we can give rough estimates of the star
543: formation rates (SFRs) of these objects.
544: We here assume Salpeter's IMF, stellar mass range of $0.1-100M_\odot$,
545: solar metallicity, no extinction, and the case B recombination
546: in the low-density limit ($N_e\ll 1.5\times 10^4\rm cm^{-3}$).
547: We can then use a conversion law of
548: $L(\lya)=1\times 10^{42}({\rm SFR}/M_\odot\rm yr^{-1}\lum$)
549: to estimate their SFRs \citep{agn2,kennicutt98}.
550: Their SFRs thus estimated are $3.3-13\mdot$, which are located
551: within the range of normal LAEs known to date
552: \citep[e.g.][]{fynbo03,rhoads03,dawson04}.
553:
554: However, their physical nature, even the SFR, is poorly constrained
555: because of the large uncertainties in dust / IGM absorption,
556: stellar mass function, and the contribution from non-stellar processes.
557: Some of the objects indeed show asymmetric line profiles suggesting
558: superwind activities (see \S\ref{res:wing} below).
559: The possibility of the existence of AGNs, although unlikely,
560: also cannot be eliminated completely.
561: Notes for individual objects in this group are summarized
562: in appendix \ref{notes:modEW}.
563:
564:
565: \subsection{Objects with broad asymmetric wings}
566: \label{res:wing}
567: As described in \S\ref{spec:wing}, five objects,
568: IB10-17108, IB10-32162, IB10-54185, IB10-90651, and IB14-52102,
569: have asymmetric broad wing emission on the \lya\ line
570: profile.
571: These objects have conspicuous high velocity wing components
572: on the red side, and a relatively sharp cut-off on the blue side
573: of the \lya\ line.
574: Their wing emission components are extended up to $\gtrsim 500\kms$ from
575: the line centers, and no clear counterparts are found on the
576: blue side. The origin of such asymmetry, which is commonly seen in
577: high-$z$ \lya\ emission, is thought to be either
578: absorption by intervening neutral hydrogen (Lyman forest absorption),
579: or galactic superwind.
580: Although these two mechanisms cannot be definitely discriminated,
581: the spectra of these five objects show some evidence
582: that favor the existence of superwind activities as shown below.
583:
584: No significant continuum emission is detected in any of the 18 objects
585: in our sample. Only one object, IB12-21989, showed marginal
586: signature of continuum emission with $\sim 2\sigma$ level of the sky
587: fluctuation, but this object does not have the red wing.
588: This implies that the contribution of the continuum to the wing components
589: we identified is negligible, and there should be a high velocity
590: \lya\ emission component within the systems.
591: An asymmetric line profile with a broad red wing and a sharp blue cut-off
592: is commonly seen in superwind galaxies \citep[e.g.][]{ajiki02,dawson02},
593: and well agrees with theoretical predictions of the superwind model
594: \citep{tenorio-tagle}.
595:
596: In fact, one of these five objects with red wings, IB10-90651,
597: shows a systematic velocity
598: structure on the two-dimensional spectrum (second part of Figure
599: \ref{fig:spec}, top-left panel).
600: This object has an extremely extended, diffuse emission component on
601: the red side of the \lya. It extends up to $5''$ (projected distance of
602: 36 kpc at $z=3.68$) toward south from the center of the object,
603: and the velocity extent is $\sim 1000\kms$ redward from the line center.
604: On the blue side, although fairly faint, we can see a counter-image
605: of the diffuse component extending toward north.
606: These velocity structures suggest that this object has a galactic superwind
607: flowing nearly along the position angle of the slit.
608: The faintness of the blue component can be explained with
609: absorption by the near-side shell of \ion{H}{1} gas,
610: while the flux of the red component is enhanced by back-scattering
611: by the far-side shell \citep{tenorio-tagle}.
612:
613: Although the blue counter-image can be seen in the two-dimensional
614: spectrum of IB10-90651, we cannot see blue wing emission on the
615: one-dimensional spectrum.
616: This asymmetric profile is quite similar to other four objects.
617: This suggests that all the five objects with red wings are
618: likely to be superwinds.
619: The fraction of such objects in our sample, $5/18\simeq 30\%$,
620: is similar to that of sub-mm detection in M04's sample \citep{geach05}.
621: Such sub-mm detection suggests the existence of obscured
622: starburst regions responsible for extended \lya\ emission,
623: i.e., superwinds.
624: The lack of diffuse and extended high velocity components
625: may suggest either that the direction of outflow is perpendicular
626: to the slit direction, or that the \lya\ emission is intrinsically faint.
627: The former possibility should be examined by further
628: deep spectroscopy with various slit direction or an integral field
629: spectroscopy. The latter is quite consistent with the
630: relatively small EWs of the four objects with red wings.
631: When the contribution of superwind activities is relatively small,
632: the EWs of the \lya\ emission should be smaller than
633: superwind-dominated sources like IB10-90651.
634:
635: Note that there are some other objects that possibly have
636: red side wings, e.g. IB12-81981 ($\sim 1.5\sigma$ up to $\sim 400\kms$).
637: However, the wing components of such objects, if any, are fairly
638: faint, and affected by neighboring
639: sky emission lines for most cases. If they indeed have wing emission
640: on the red side, they are also likely to have superwind activities
641: within the systems. Other notes for each object are listed in
642: appendix \ref{notes:wing}
643:
644:
645: \subsection{Objects with large EWs and no wings}
646: \label{res:cool}
647: After classifying our objects in two ways,
648: there remains eight objects that have large EWs
649: and no significant wing emission on the red side of the \lya\ line:
650: IB11-59167, IB11-80344, IB11-89537, IB12-30834, IB12-48320, IB12-58572,
651: IB12-81981, and IB13-104299.
652: Except for one, IB12-58572, they are all continuum-faint
653: objects (see \S\ref{sample}) with $\rm EW_{rest} > 240\AA$.
654: The EW of 240\AA\ is predicted for the starbursts with solar metallicity
655: at an age of $10^6\rm yrs$ by \citet{cf93}, and also predicted by
656: MR02 by introducing an extremely top-heavy IMF (slope = 0.5)
657: or zero-metallicity, for starbursts with an age of $\gtrsim 10^7\rm yrs$.
658: As mentioned above in \S\ref{spec:ew},
659: the observed EWs should be smaller than intrinsic values due to the
660: absorption of \lya\ emission by the IGM.
661: Namely, the seven continuum-faint objects
662: have EWs securely exceeding 240\AA.
663:
664: For these eight sources, the origin of the \lya\ emission is hardly
665: thought to be {\it ordinary} starbursts.
666: If their ionizing sources are starbursts, the large EWs require
667: photo-ionization by extremely young stellar populations
668: (age $\ll 10^7\rm yrs$) with an extremely top-heavy IMF
669: and/or very low metallicity.
670: In this case, the stellar components should be dominated by population
671: III stars. Calculations by MR02 show that a stellar population
672: with zero-metallicity and Salpeter's IMF can achieve an extremely
673: large EW, 1122\AA\ at the age of $10^6\rm yrs$. \citet{schaerer03}
674: also suggests by treating various IMFs and stellar mass ranges that,
675: under very metal-poor conditions, the EWs of \lya\ emission
676: from very young starbursts can be highly enhanced, $\sim 1300\rm\AA$
677: at an age of $10^6\rm yrs$. In this picture, the extended \lya\
678: emission can be attributed to resonant-scattering.
679: In addition, it is thought that cooling radiation or supernova explosions
680: also contribute to the \lya\ emission, at least in some part.
681:
682: As for the starbursts, it is also possible that the starburst
683: regions are heavily obscured by dust with somewhat an inhomogeneous
684: distribution. If the dust screen absorbs UV photons
685: traveling along the line of sight, the observed faintness in UV continuum
686: can be reproduced. Then if UV photons traveling along other
687: directions are not significantly absorbed, they can ionize the
688: \ion{H}{1} gas surrounding the starburst region, and can lead to
689: spatially extended \lya\ emission.
690: Since UV continuum emission cannot be scattered in this manner,
691: these processes can at least qualitatively explain the extended
692: \lya\ emission and faint / compact UV continuum emission.
693: This is thought to be a plausible explanation because
694: galaxies in high redshifts are generally surrounded by a
695: larger amount of \ion{H}{1} gas than galaxies in the present
696: universe \citep[e.g.][]{adelberger03}.
697: The \lya\ emitting gas will be spatially more extended than
698: the starburst region, and thus less sensitive to dust absorption.
699:
700: If the dust screen covers the whole direction, even \lya\ photons
701: cannot escape the system. Instead, the star forming activities
702: can lead to shock-ionization by a superwind
703: \citep[e.g.][]{ts00,tsk01,veilleux05}.
704: However, we could not detect significant high-velocity components
705: on the \lya\ line profiles, nor on the two-dimensional spectra.
706: If superwind activities are responsible for the extended \lya\
707: emission for these objects, the shocked region associated with
708: the wind, if any, should be very faint. Their relatively small velocity
709: widths of the \lya\ line, typically $\sim 300-500\kms$, also suggests
710: that powerful superwinds are unlikely. The exception is two objects,
711: IB11-80344 and IB11-89537, which have relatively large velocity widths
712: up to $\sim 740\kms$. These objects might be superwinds,
713: but have smaller velocity widths than the most prominent ones known
714: to date \citep[e.g. $\sim 10^3\kms$, ][]{heckman90}.
715: Additionally, note that there are some objects that have a {\em marginal}
716: signature of wing emission on the red side of the \lya,
717: e.g. IB12-48320 and IB12-81981. The red side spectra of these
718: objects are overlapped with relatively strong sky emission lines
719: which cannot be subtracted completely. If a deeper exposure and/or a
720: more sufficient sky subtraction become available, we can possibly
721: detect a high velocity wing component on the line profile for some
722: objects.
723:
724: Apart from such uncertainties, the eight objects show a remarkable feature.
725: Figure \ref{fig:ldv} plots the velocity width as a function of
726: the \lya\ line luminosity. If the ionizing mechanism is
727: cold accretion, the velocity width will roughly correspond to
728: the circular velocity of the system, and thus will be scaled to
729: the mass of the system.
730: On the other hand, the \lya\ luminosity should be positively correlated
731: with the gravitational energy of the system, and thus scaled to the mass
732: of the system. Therefore, the velocity width and the \lya\ line luminosity
733: should have a positive correlation.
734: Among all the objects in our VIMOS sample,
735: the eight objects with large EWs and no wings clearly show
736: a positive correlation in Figure \ref{fig:ldv}.
737: Although not a definite constraint, this result supports the
738: idea of cooling radiation.
739: Their relatively small velocity widths quite well agree with
740: a theoretical prediction of cooling radiation \citep{fardal}.
741: Even for the two objects with large velocity widths ($\Delta V>500\kms$),
742: these velocity widths can be reproduced by taking into account
743: the effect of resonant-scattering, or perhaps an additional ionizing
744: source in the system \citep{dijkstra06a, dijkstra06b}.
745: Furthermore, two or more objects (e.g. IB12-58572 and IB12-104299)
746: marginally show an odd asymmetry of the line profile, i.e., broader
747: on the blue side than on the red side, which also agrees with the
748: prediction of \citet{dijkstra06a}.
749: Thus, cooling radiation can be a reasonable explanation for these
750: eight objects. It is also true that other mechanisms contribute
751: to the \lya\ emission, together with cold accretion.
752: Such additional mechanisms are discussed in \S\ref{res:other}
753: Notes for individual objects are described in appendix \ref{notes:cool}.
754:
755:
756: \subsection{Constraints from other data}
757: \label{res:other}
758: We have mentioned about four possibilities for the origins of the
759: spatially extended \lya\ emission: stellar photo-ionization,
760: photo-ionization by AGNs, shock-heating by superwinds,
761: and cooling radiation from infalling material (cold accretion).
762: These scenarios predict characteristic emission lines of each
763: ionizing source.
764: In addition to our VIMOS data, we can use the data
765: taken in our previous study and X-ray data for diagnostics
766: of the physical origins.
767: We then first analyzed the low-resolution spectroscopic
768: data taken with FOCAS on Subaru Telescope, and the IA images
769: taken with Suprime-Cam (Paper1) to detect the emission lines.
770:
771: For starbursts, there are two cases described in \S\ref{res:cool},
772: i.e., extremely young starbursts dominated by population III stars,
773: and starbursts heavily obscured by dust.
774: The former case of starbursts has an extremely low metallicity and
775: a top-heavy IMF. Such an extremely young stellar population formed
776: in a primordial condition should have \ion{He}{2} $\lambda 1640$
777: emission \citep{schaerer03}.
778: The \ion{He}{2} line is also predicted for the case of cooling
779: radiation \citep{yang06}. While \lya\ line is optically thick in
780: general, this line should be optically thin.
781: This implies that the \ion{He}{2} line is suitable to probe the
782: gas dynamics, and thus a good probe to discriminate stellar photo-ionization
783: and cooling radiation. However, none of our objects in the FOCAS
784: spectroscopic sample has detectable \ion{He}{2} emission.
785: On the composite spectrum, there is apparently an emission
786: feature at the wavelength of rest-frame 1640\AA\ (see Figure 4 of
787: Paper1).
788: This is just a marginal signature since the wavelengths near the
789: rest-frame 1640\AA\ is largely affected by sky emission lines.
790:
791: For the case of obscured starbursts, there should be characteristic
792: metal lines, such as [\ion{O}{2}] $\lambda 3727$, [\ion{O}{3}]
793: $\lambda\lambda 4959, 5007$, and H$\alpha$ $\lambda 6563$.
794: These lines are less sensitive to dust absorption than the \lya\ line,
795: and should be good diagnostics of starbursts, although they
796: are redshifted beyond the wavelength coverage of our FOCAS data.
797: Such metal lines can also be seen in superwind galaxies.
798: The primordial gas surrounding galaxies should be
799: chemically enriched by supernova explosions, resulting in
800: spatially extended metal lines. Namely, superwind galaxies
801: are expected to have spatially extended \ion{C}{4} line emission
802: with a strength of $\sim 1/10$ of the \lya\ line \citep{heckman91b}.
803: We previously showed in Paper1 that there are no such metal emission
804: lines either in the individual spectra or in the stacked spectrum,
805: down to our current detection limits. This suggests that at least
806: the majority of the FOCAS sample (seven objects) are unlikely to
807: be superwinds.
808: These pieces of evidence from the observations are, however, just weak
809: constraints on the superwind scenario. In fact, IB10-90651 is
810: proved to have extended diffuse emission components by performing
811: deep spectroscopy. Further deep follow-up observations may find
812: diffuse emission components suggesting the superwind scenario.
813:
814: The \ion{C}{4} emission line, as well as \ion{N}{5} or \ion{He}{2}
815: lines, can also be diagnostics of the AGN activities.
816: Similarly to superwinds, AGN scenario is also likely to be ruled out
817: by the absence of the \ion{C}{4} or \ion{N}{5} lines on the
818: FOCAS spectra.
819: Our IA images show that, although they are more extended than point-sources,
820: the sizes of \lya\ emission components are typically $\sim 10-15\rm\, kpc$,
821: and are smaller than those of \lya\ nebulae associated with AGNs
822: \citep[e.g.][]{heckman91a,weidinger04,weidinger05}.
823: The velocity widths of our objects are also smaller than those of
824: \lya\ nebulae associated with AGNs known to date
825: \citep[e.g.][]{vanojik,weidinger04,weidinger05}.
826: Together with the faintness in UV continuum,
827: these facts suggest that if the ionizing sources are AGNs,
828: they must be obscured (type II) AGNs.
829: In order to examine such AGN activities, it is important to compare
830: them with type II QSOs known to date.
831: Type II AGNs are, however, suggested to be quite rare by photometric
832: studies of LAE samples \citep{malhotra03, wang04}.
833: We will then just show the results of our analysis of X-ray data
834: taken with {\it XMM-Newton}.
835:
836: Most of our field is covered by X-ray data with two pointings
837: of {\it XMM-Newton} (Ueda et al. 2007),
838: and the X-ray flux limits can be estimated by using these data.
839: The sensitivity is sufficient to detect bright quasars at $z\sim 4$.
840: In order to put the strongest constraints,
841: we here used the data of 0.5--4.5 keV band,
842: the most sensitive band of {\it XMM-Newton}.
843: We estimated the $1\sigma$ flux limits in the rest-frame 2--10 keV
844: from the count rate limits, by assuming X-ray spectra of
845: $F_\epsilon\propto\epsilon^{-0.8}$ and absorption with a hydrogen
846: column density of
847: $N_{\rm H}=1\times 10^{23}\rm cm^{-2}$ and
848: $N_{\rm H}=1\times 10^{24}\rm cm^{-2}$.
849: The flux limits thus estimated are listed in Table \ref{tab:x}.
850: The typical values of their flux limits in the case of
851: $N_{\rm H}=1\times 10^{23}\rm cm^{-2}$ are a few $\times 10^{-15}\fluxunit$,
852: which leads to X-ray luminosities $\sim 10^{45}\lum$ at $z\sim 4$.
853: These are comparable to the X-ray luminosities of type II QSOs
854: known to date \citep{norman02, stern02, dawson03}.
855: Our analysis ignores the reflection components of X-ray emission,
856: so that the upper limits listed in Table \ref{tab:x} are
857: thought to be the most conservative values, i.e., the X-ray
858: luminosities are likely to be significantly smaller than
859: those of type II QSOs.
860: Note that we cannot put any constraints if our objects are
861: compton-thick, $N_{\rm H}\gg 10^{24}\rm cm^{-2}$.
862: These imply that our objects are unlikely to be harboring
863: type II QSOs, although the possibility of AGNs is not completely
864: ruled out due to the relatively shallow exposure of the
865: X-ray data.
866:
867:
868: \section{Conclusions}
869: \label{con}
870: We have carried out deep, medium-resolution ($R\sim 2000$) follow-up
871: spectroscopy of 18 extended \lya\ sources at $z\sim 3-5$ that are faint
872: and/or compact in UV continuum, using VLT/VIMOS.
873: We found that all 18 objects in our VIMOS sample have fairly large
874: equivalent widths (EWs) of the \lya\ emission with a median value
875: of $\approx 210\rm\AA$ in the rest-frame.
876: A half of our sample (nine objects) have moderately large EWs of
877: $\sim 100-200\rm\AA$, and are accounted for by ordinary starbursts with
878: a solar metallicity. The high resolution and S/N ratio of our data
879: enabled us to make a quantitative analysis of the \lya\ line profiles.
880: The velocity widths were found to be relatively small,
881: typically $\sim 300-500\kms$.
882:
883: For five objects ($\sim 30\%$), we identified conspicuous
884: broad wing emission components on the red side of the \lya\ line,
885: while no significant wing emission was found on the blue side.
886: Such features quite well agree with theoretical predictions of
887: galactic superwind model.
888: Namely, we found diffuse high velocity emission components extended
889: up to $\sim 70\rm\, kpc$ on the two-dimensional spectrum of
890: one of these five objects.
891: The velocity extent of these components is $\gtrsim 2000\kms$
892: with systematic velocity structure, i.e., the southern part
893: is redshifted and the northern part is blueshifted.
894: These features suggest the existence of galactic superwind activities.
895:
896: Excluding the objects above, there remain eight objects, which
897: have large EWs exceeding 200\AA\ (seven have EWs exceeding 240\AA),
898: and no clear signature of high velocity wing emission on the either
899: side of the \lya\ line. These objects are hardly thought to be
900: ordinary starbursts, and our IA images and FOCAS spectra show
901: no clear signature of AGN activities. Non-detection in X-ray
902: data taken with {\it XMM-Newton} also suggests they are unlikely to
903: be AGN-origin like type II QSOs. Their relatively small spatial
904: extents and velocity widths agree well with theoretical predictions
905: of cooling radiation. Furthermore, their velocity widths clearly
906: show a positive correlation with the \lya\ line luminosities.
907: These facts suggest that they are candidates for forming-galaxies
908: in gas-cooling phase, i.e., the very first stage of galaxy formation.
909:
910:
911: \begin{thebibliography}{}
912: \bibitem[Adelberger et al.(2003)]{adelberger03}
913: Adelberger, K. L., Steidel, C. C., Shapley, A. E., \&
914: Pettini, M. 2003, \apj, 584, 45
915: \bibitem[Ajiki et al.(2002)]{ajiki02}
916: Ajiki, M., et al. 2002, \apj, 576, L28
917: \bibitem[Basu-Zych \& Scharf (2004)]{bzs04}
918: Basu-Zych, A., Scharf, C. 2004, \apj, 615, L85
919: \bibitem[Birnboim \& Dekel (2003)]{bd03}
920: Birnboim, Y., \& Dekel, A. 2003, \mnras, 345, 349
921: \bibitem[Bower et al.(2004)]{bower04}
922: Bower, R. G., et al. 2004, \mnras, 351, 63
923: \bibitem[Chapman et al.(2004)]{chapman04}
924: Chapman, S. C., Scott, D., Windhorst, R. A., Frayer, D. T., Borys, C.,
925: Lewis, G. F., \& Ivison, R. J. 2004, \apj, 606, 85
926: \bibitem[Chapman et al.(2001)]{chapman01}
927: Chapman, S. C., Lewis, G. F., Scott, D., Richards, E., Borys, C.,
928: Steidel C. C., Adelberger K. L., \& Shapley A. E. 2001,
929: \apj, 548, L17
930: \bibitem[Charlot \& Fall (1993)]{cf93}
931: Charlot, S., \& Fall, S. M. 1993, \apj, 415, 580
932: \bibitem[Cowie \& Hu (1998)]{ch98}
933: Cowie, L. l., \& Hu, E. M. 1998, \aj, 115, 1319
934: \bibitem[Dawson et al.(2004)]{dawson04}
935: Dawson, S., et al. 2004, \apj, 617, 707
936: \bibitem[Dawson et al.(2003)]{dawson03}
937: Dawson, S., McCrady, N., Stern, D., Eckart, M. E., Spinrad, H.,
938: Liu, M. C., \& Graham, J. R. 2003, \aj, 125, 1236
939: \bibitem[Dawson et al.(2002)]{dawson02}
940: Dawson, S., Sprinrad, H, Stern, D., van Breugel, W., de Vries, W.,
941: \& Reuland, M. 2002, \apj, 570, 92
942: \bibitem[Dekel \& Birnboim (2006)]{db06}
943: Dekel, A., \& Birnboim, Y. 2006, \mnras, 368, 2
944: \bibitem[Dey et al.(2005)]{dey05}
945: Dey, A., et al. 2005, \apj, 629, 654
946: \bibitem[Dijkstra et al.(2006a)]{dijkstra06a}
947: Dijkstra, M., Haiman, Z., Spaans, M. 2006a, \apj, 649, 14
948: \bibitem[Dijkstra et al.(2006b)]{dijkstra06b}
949: Dijkstra, M., Haiman, Z., Spaans, M. 2006b, \apj, 649, 37
950: \bibitem[Fardal et al.(2001)]{fardal} Fardal M., Katz N., Gardner J. P.,
951: Hernquist, L., Weinberg, D. H., Romeel, D. 2001, \apj, 562, 605
952: \bibitem[Francis et al.(2001)]{francis01}
953: Francis, P. J., et al. 2001, \apj, 554, 1001
954: \bibitem[Fukugita et al.(1995)]{fsi95}
955: Fukugita, M., Shimasaku, K., \& Ichikawa, T. 1995, \pasp, 107, 945
956: \bibitem[Furlanetto et al.(2004)]{furlanetto04}
957: Furlanetto, S., Schaye, J., Springel, V., \& Hernquist, L. 2004, \apj, 622, 7
958: \bibitem[Fynbo et al.(2001)]{fynbo}
959: Fynbo, J. U., M\o ller, P., \& Thomsen, B. 2001, \aap, 374, 443
960: \bibitem[Fynbo et al.(2003)]{fynbo03}
961: Fynbo, J. P. U., Ledoux, C., M\o ller, P., Thomsen, B., \& Burud, I. 2003,
962: \aap, 407, 147
963: \bibitem[Geach et al. (2005)]{geach05}
964: Geach, J. F. et al. 2005, \mnras, 363, 1398
965: \bibitem[Haiman et al.(2000)]{haiman00}
966: Haiman, Z., Spaans, M., \& Quataert, E. 2000, \apj, 537, L5
967: \bibitem[Haiman \& Rees (2001)]{hr01}
968: Haiman, Z., \& Rees, M. J. 2001, \apj, 556, 87
969: \bibitem[Hayashino et al.(2000)]{hayashino}
970: Hayashino, T., et al. 2000, Proc. SPIE, 4008, 397
971: \bibitem[Heckman et al.(1990)]{heckman90}
972: Heckman, T. M., Armus, L., Miley, G. K. 1990, \apjs, 74, 833
973: \bibitem[Heckman et al.(1991a)]{heckman91a}
974: Heckman, T. M., Lehnert, M. D., van Breugel, W., \& Miley, G. K. 199a,
975: \apj, 370, 78
976: \bibitem[Heckman et al.(1991b)]{heckman91b}
977: Heckman T. M., Lehnert M. D., Miley G. K., \& van Breugel W. 1991b,
978: \apj, 381, 373
979: \bibitem[Hu et al.(1998)]{hu98}
980: Hu, E. M., Cowie, L. L., \& McMahon, R. O. 1998, \apj, 502, L99
981: \bibitem[Iye et al.(2006)]{iye06}
982: Iye, M., et al. 2006, \nat, 443, 1861
983: \bibitem[Kashikawa et al.(2005)]{kashikawa}
984: Kashikawa, N., et al. 2005, \pasj, 648, 7
985: \bibitem[Keel et al.(1999)]{keel99}
986: Keel, W. C., Cohen, S. H., Windhorst, R. A., \& Waddington I. 1999,
987: \aj, 118, 2547
988: \bibitem[Kennicutt (1998)]{kennicutt98}
989: Kennicutt, R. C., Jr. 1998, \araa, 36, 189
990: \bibitem[Kodaira et al.(2003)]{kodaira}
991: Kodaira, K., et al. 2003, \pasj, 55, L17
992: \bibitem[Laursen \& Sommer-Larsen (2006)]{laursen06}
993: Laursen, P., \& Sommer-Larsen, J. 2006, preprint (astro-ph/0610761)
994: \bibitem[Madau (1995)]{madau95}
995: Madau P. 1995, \apj, 441, 18
996: \bibitem[Maier et al.(2003)]{maier}
997: Maier, C., et al. 2003, \aap, 402, 79
998: \bibitem[Malhotra et al.(2003)]{malhotra03}
999: Malhotra, S., Wang, J. X., Rhoads, J. E., Heckman, T. M., \& Norman, ,C. A.
1000: 2003, \apj, 585, L25
1001: \bibitem[Malhotra \& Rhoads (2002)]{mr2002}
1002: Malhotra, S., \& Rhoads, J. E. 2002, \apj, 565, L71 (MR02)
1003: \bibitem[Matsuda et al.(2006)]{m06}
1004: Matsuda, Y., Yamada, T., Hayashino, T., Yamauchi, R., \& Nakamura, Y. 2006,
1005: \apj, 640, L123
1006: \bibitem[Matsuda et al.(2004)]{m04}
1007: Matsuda, Y., et al. 2004, \aj, 128, 569 (M04)
1008: \bibitem[McCarthy et al.(1999)]{mccarthy99}
1009: McCarthy, P. J., van Breugel, W., Kapahi, V. K., \& Subrahmanya, C. R.
1010: 1999, \apj, 192, 522
1011: \bibitem[Mo \& White (2002)]{mw2002}
1012: Mo, H. J., \& White, S. D. M. 2002, \mnras, 336, 112
1013: \bibitem[M\o ller \& Warren (1998)]{moller98}
1014: M\o ller, P., Warren, S. J. 1998, \mnras, 299, 661
1015: \bibitem[Nilsson et al.(2006)]{nilsson}
1016: Nilsson, K., Fynbo, J. P. U., Sommer-Larsen, J., \& Ledoux, C. 2006,
1017: 452, L23
1018: \bibitem[Norman et al.(2002)]{norman02}
1019: Norman, C., et al. 2002, \apj, 571, 218
1020: \bibitem[Ohyama et al.(2003)]{ohyama}
1021: Ohyama, Y., et al. 2003, \apj, 591, L9
1022: \bibitem[Oke (1974)]{oke}
1023: Oke, J. B. 1974, \apjs, 27, 21
1024: \bibitem[Osterbrock (1989)]{agn2}
1025: Osterbrock, D. E. 1989, ``{\it Astrophysics of Gaseous Nebulae and
1026: Active Galactic Nuclei}'', University Science Books
1027: \bibitem[Ouchi et al.(2005)]{ouchi05}
1028: Ouchi, M., et al. 2005, \apj, 620, L1
1029: \bibitem[Ouchi et al.(2003)]{ouchi03}
1030: Ouchi, M., et al. 2003, \apj, 582, 60
1031: \bibitem[Palunas et al.(2004)]{palunas04}
1032: Palunas, P., Teplitz, H. I., Francis, P. J., Williger, G. M.,
1033: \& Woodgate, B. E. 2004, \apj, 602, 545
1034: \bibitem[Partridge \& Peebles (1957)]{partridge}
1035: Partridge, R. B., \& Peebles, P. J. E. 1957, \apj, 148, 377
1036: \bibitem[Rhoads et al.(2003)]{rhoads03}
1037: Rhoads, J. E., et al. 2003, \aj, 125, 1006
1038: \bibitem[Rhoads et al.(2000)]{rhoads00}
1039: Rhoads, J. E., Malhotra, S., Dey, A., Stern, D., Spinrad, H.,
1040: \& Jannuzi, B. T. 2000, \apj, 545, L85
1041: \bibitem[Saito et al.(2006)]{paper1}
1042: Saito, T., Shimasaku, K., Okamura, S., Ouchi, M., Akiyama, M.,
1043: \& Yoshida, M. 2006, \apj, 648, 54
1044: \bibitem[Schaerer (2003)]{schaerer03}
1045: Schaerer, D. 2003, \aap, 397, 527
1046: \bibitem[Sekiguchi et al.(2004)]{sekiguchi04}
1047: Sekiguchi K., et al. 2004, \apss, 301, 169
1048: \bibitem[Shimasaku et al.(2003)]{shimasaku03}
1049: Shimasaku, K., et al. 2003, \apj, 586, L111
1050: \bibitem[Steidel et al.(2000)]{s00}
1051: Steidel, C., Adelberger, K. T., Shapey, A. E., Pettini, M., Dickinson, M.,
1052: \& Giavalisco, M. 2000, \apj, 532, 170 (S00)
1053: \bibitem[Stern et al.(2002)]{stern02}
1054: Stern, D., et al. 2002, \apj, 568, 71
1055: \bibitem[Taniguchi \& Shioya (2000)]{ts00}
1056: Taniguchi, Y., \& Shioya, Y. 2000, \apj, 532, L13
1057: \bibitem[Taniguchi et al.(2001)]{tsk01}
1058: Taniguchi, Y., Shioya, Y., \& Kakazu, Y. 2001, \apj, 562, L15
1059: \bibitem[Tenorio-Tagle et al.(1999)]{tenorio-tagle}
1060: Tenorio-Tagle, G., Silich, S. A., Kunth, D., Terlevich, E.,
1061: \& Terlevich, R. 1999, \mnras, 309, 332
1062: \bibitem[van Ojik et al.(1997)]{vanojik}
1063: van Ojik R., R\" ottgering H. J. A., Miley G. K., \& Hunstead R. W. 1997,
1064: \aap, 317, 358
1065: \bibitem[Veilleux et al.(2005)]{veilleux05}
1066: Veilleux, S., Cecil, G., \& Bland-Hawthorn, J. 2005, \araa, 43, 769
1067: \bibitem[Wang et al.(2004)]{wang04}
1068: Wang, J. X., et al. 2004, ApJ, 608, L21
1069: \bibitem[Weidinger et al.(2004)]{weidinger04}
1070: Weidinger, M., M\o ller, P., \& Fynbo, J. P. U. 2004, \nat, 430, 999
1071: \bibitem[Weidinger et al.(2005)]{weidinger05}
1072: Weidinger, M., M\o ller, P., Fynbo, J. P. U., \& Thomsen, B. 2005,
1073: \aap, 436, 825
1074: \bibitem[Wilman et al.(2005)]{wilman}
1075: Wilman R. J., Gerssen J, Bower R. G., Morris S. L., Bacon R.,
1076: de Zeeuw P. T., \& Davies R. L. 2005, \nat, 436, 227
1077: \bibitem[Yang et al.(2006)]{yang06}
1078: Yang, Y., Zabludoff, A. I., Dav\' e, R., Eisenstein, D. J.,
1079: Pinto, P. A., Katz, N., Winberg, D. H., \& Barton, E. J. 2006,
1080: \apj, 640, 539
1081: \end{thebibliography}
1082:
1083: \appendix
1084: \section{Notes for individual objects}
1085: \subsection{Objects with moderate EWs}
1086: \label{notes:modEW}
1087: \paragraph{IB08-86220}
1088: $z=3.31$, ${\rm SFR} \simeq 3.3\mdot$, $\rm EW_{rest}\simeq 120\rm\AA$.
1089: Even the upper limit of EW does not exceed 200\AA.
1090: Our analysis may suffer from relatively large residuals of the
1091: sky emission at $\approx 5225\rm\AA$ and $\approx 5240\rm\AA$.
1092:
1093: \paragraph{IB10-17108}
1094: $z=3.79$, ${\rm SFR}\simeq 10\mdot$, $\rm EW_{rest}\simeq 110\rm\AA$.
1095: The EW is the smallest in our sample, and the uncertainty in the
1096: EW is relatively small. Wing emission can be seen on the red side
1097: (see appendix \ref{notes:wing}).
1098:
1099: \paragraph{IB10-32162}
1100: $z=3.74$, ${\rm SFR}\simeq 9.7\mdot$, $\rm EW_{rest}\simeq 150\rm\AA$.
1101: The upper limit of the EW is relatively large for this group of objects,
1102: $\simeq 180\rm\AA$.
1103: The velocity width is relatively large, $\Delta V \simeq 610\kms$ (FWHM).
1104: Wing emission can be seen on the red side (see appendix \ref{notes:wing}).
1105:
1106: \paragraph{IB10-54185}
1107: $z=3.82$, ${\rm SFR}\simeq 6.5\mdot$, $\rm EW_{rest}\simeq 130\rm\AA$.
1108: The velocity width is quite small, $\Delta V\simeq 300\kms$.
1109: Wing emission can be seen on the red side (see appendix \ref{notes:wing}).
1110:
1111: \paragraph{IB12-21989}
1112: $z=4.11$, ${\rm SFR}\simeq 13\mdot$, $\rm EW_{rest}\simeq 180\rm\AA$.
1113: Continuum emission can be seen at $\simeq 1.5\sigma$ level,
1114: and the EW calculated directly from the spectrum is $\simeq 35\rm\AA$.
1115: Our photometry may suffer from neighboring continuum sources.
1116:
1117: \paragraph{IB12-71781}
1118: $z=4.11$, ${\rm SFR}\simeq 4.1\mdot$, $\rm EW_{rest}\simeq 130\rm\AA$.
1119: The upper limit of the EW is fairly large, $\simeq 230\rm\AA$,
1120: so that this object is possibly not categorized as an ordinary starburst.
1121: Our analysis may suffer from the several sky emission lines on the
1122: red side of the \lya\ (e.g. $\approx 6235\rm\AA$).
1123:
1124: \paragraph{IB13-96047}
1125: $z=4.27$, ${\rm SFR}\simeq 10\mdot$, $\rm EW_{rest}\simeq 140\rm\AA$.
1126: Even the upper limit of the EW does not exceed 200\AA.
1127: Our spectral analysis may suffer from neighboring several sky
1128: emission lines on the both sides of the \lya.
1129: Neighboring continuum sources may affect our photometric analysis.
1130:
1131: \paragraph{IB14-47257}
1132: $z=4.66$, ${\rm SFR}\simeq 13\mdot$, $\rm EW_{rest}\simeq 180\rm\AA$.
1133: With an upper limit of the EW $\simeq 240\rm\AA$, this object not be
1134: an ordinary starburst. The spectrum suffers from relatively strong residuals of
1135: the sky emission at $\approx 6865, 6900, \mbox{and } 6910\rm\AA$ near
1136: the \lya\ line. The \lya\ line itself is overlapped with weak sky emissions.
1137:
1138: \paragraph{IB14-52102}
1139: $z=4.47$, ${\rm SFR}\simeq 9.0\mdot$, $\rm EW_{rest}\simeq 150\rm\AA$.
1140: With an upper limit of the EW $\simeq 200\rm\AA$, this object may not
1141: be an ordinary starburst. The line profile has wing emission on the red side
1142: (see appendix \ref{notes:wing}).
1143:
1144: \subsection{Objects with broad asymmetric wings}
1145: \label{notes:wing}
1146: \paragraph{IB10-17108}
1147: $z=4.11$, $\rm EW_{rest}\approx130\rm\AA$, $\Delta V \approx490\kms$.
1148: Also listed in appendix \ref{notes:modEW}.
1149: The wing component is very clear, and extended up to $\gtrsim 700\kms$
1150: from the line center.
1151:
1152: \paragraph{IB10-32162}
1153: $z=3.74$, $\rm EW_{rest}\approx150\rm\AA$, $\Delta V \approx610\kms$.
1154: Also listed in appendix \ref{notes:modEW}.
1155: The velocity width of the main component is fairly large, $\approx610\kms$.
1156: The wing component is not very clear, but the line profile shows
1157: a significant asymmetry.
1158:
1159: \paragraph{IB10-54185}
1160: $z=3.82$, $\rm EW_{rest}\approx130\rm\AA$, $\Delta V \approx300\kms$.
1161: Also listed in appendix \ref{notes:modEW}.
1162: The spectrum may suffer from sky emission lines on the red side
1163: ($\approx 5890\rm\AA$ and a weak one at $\approx 5870\rm\AA$),
1164: but the red side wing exceeds the $3\sigma$ noise level at $\lesssim 450\kms$
1165: from the line center. The wing component can be seen up to
1166: $\sim 600-1000\kms$.
1167:
1168: \paragraph{IB10-90651}
1169: $z=3.68$, $\rm EW_{rest}\approx860\rm\AA$, $\Delta V \approx570$.
1170: The EW is the largest in our sample.
1171: The wing emission is extended up to $\gtrsim 500\kms$.
1172: A systematic velocity structure of diffuse emission can be seen
1173: in the two-dimensional spectrum.
1174:
1175: \paragraph{IB14-52102}
1176: $z=4.47$, $\rm EW_{rest}\approx150\rm\AA$, $\Delta V \approx350\kms$.
1177: Also listed in appendix \ref{notes:modEW}.
1178: The wing component exceeds the $3\sigma$ noise level, and is extended
1179: up to $\sim 500\kms$.
1180:
1181:
1182: \subsection{Objects with large EWs and no wings}
1183: \label{notes:cool}
1184: \paragraph{IB11-59167}
1185: $z=3.94$, $\rm EW_{rest}\approx 270\rm\AA$, $\Delta V \approx270 \kms$.
1186: A systematic (outflow-like) velocity structure may exist,
1187: but the spectrum may suffer from incomplete sky subtraction
1188: at $\approx 6000\rm\AA$.
1189:
1190: \paragraph{IB11-80344}
1191: $z=3.89$, $\rm EW_{rest}\approx 569\rm\AA$, $\Delta V\approx740\kms$.
1192: The lower limit of the EW is $\approx 380\rm\AA$.
1193: The velocity width is large, and the line profile has a double-peaked
1194: shape. The spectrum may suffer from several sky emission lines.
1195:
1196: \paragraph{IB11-89537}
1197: $z=4.03$, $\rm EW_{rest}\approx 460\rm\AA$, $\Delta V\approx 560\kms$.
1198: The lower limit of the EW is $\approx 300\rm\AA$.
1199: Wing emission on the red side may exist.
1200: The spectrum may suffer from a neighboring continuum source.
1201:
1202: \paragraph{IB12-30834}
1203: $z=4.11$, $\rm EW_{rest}\approx 360\rm\AA$, $\Delta V\approx 430\kms$.
1204: The lower-limit of EW is $\approx 240\rm\AA$.
1205: The line profile is quite simple (nearly Gaussian), and the
1206: velocity width is relatively small.
1207:
1208: \paragraph{IB12-48320}
1209: $z=4.04$, $\rm EW_{rest}\approx 320$, $\Delta V\approx 320\kms$.
1210: The lower limit of the EW is $\approx 200\rm\AA$.
1211: A relatively strong residual of sky subtraction remains
1212: on the red side at $\approx 6145\rm\AA$.
1213: The line profile may suffer from a weak sky emission line here.
1214:
1215: \paragraph{IB12-58572}
1216: $z=4.04$, $\rm EW_{rest}\approx 210\rm\AA$, $\Delta V\approx 400\kms$.
1217: The lower limit of the EW does not exceed 200\AA.
1218: This object was categorized as a continuum-compact source in
1219: appendix \ref{sample}.
1220:
1221: \paragraph{IB12-81981}
1222: $z=4.12$, $\rm EW_{rest}\approx 500\rm\AA$, $\Delta V\approx 380\kms$.
1223: The lower limit of the EW is $\approx 260\rm\AA$.
1224: The profile shows asymmetry, but no significant wing emission can be
1225: found. The red side suffers from relatively strong sky emission lines
1226: at $\approx 6235\rm\AA$ and 6260\AA, and several weak ones.
1227:
1228: \paragraph{IB13-104299}
1229: $z=4.42$, $\rm EW_{rest}\approx 300\rm\AA$, $\Delta V\approx 260\kms$.
1230: The lower-limit of EW does not exceed 200\AA.
1231: The spectrum may suffer from several sky emission lines
1232: at $\approx 6560, 6580, \mbox{ and }6600\rm\AA$.
1233: The photometry may suffer from a neighboring continuum source.
1234:
1235:
1236: %%%%%%%%%%%%%%%%%%%%%
1237: %% TABLES
1238: %%%%%%%%%%%%%%%%%%%%%
1239: \input{tab1.tex}
1240: \input{tab2.tex}
1241: \input{tab3.tex}
1242:
1243: %%%%%%%%%%%%%%%%%%%%%
1244: %% FIGURES
1245: %%%%%%%%%%%%%%%%%%%%%
1246:
1247: %Figure: Luminosity function
1248: \begin{figure}
1249: \plotone{f1.eps}
1250: \caption{Luminosity functions (LFs) of our sample (filled squares)
1251: and a sample of normal LAEs at $z\simeq 3.7$ by Ouchi et al. (open circles).
1252: The vertical error bars show the Poisson errors, and the horizontal
1253: error bars represent the size of the luminosity bins.
1254: }
1255: \label{fig:lf}
1256: \end{figure}
1257:
1258: %Figure: Redshift distribution
1259: \begin{figure}
1260: \plotone{f2.eps}
1261: \caption{Redshift distribution of extended \lya\ sources.
1262: The histogram shows the redshift distribution of our photometric
1263: sample of 41 extended \lya\ sources, based on coarse redshift
1264: information obtained from the IA imaging (redshift resolution
1265: $\Delta z\sim 0.2$). The open diamonds show the detection limit
1266: ($2\sigma$) for each IA band in units of rest-frame surface brightness
1267: (right axis).
1268: }
1269: \label{fig:z-hist}
1270: \end{figure}
1271:
1272: %Figure: Size distribution
1273: \begin{figure}
1274: \plotone{f3.eps}
1275: \caption{Size distribution of the photometric sample of 41 extended
1276: \lya\ sources. The top panel is a histogram of the size defined by
1277: twice the half-light-radius (HLR). The bottom panel shows the size
1278: as a function of (roughly estimated) redshift.
1279: For both panels, colors correspond to the bands
1280: where we detected the \lya\ emission. In the bottom panel,
1281: the horizontal error bars indicate the bandwidths of the IA bands,
1282: and the vertical error bar shows the minimum and maximum value. }
1283: \label{fig:sizedist}
1284: \end{figure}
1285:
1286: %Figure: Images
1287: \begin{figure}
1288: \plotone{f4a.eps}
1289: \caption{Images of the 18 extended \lya\ sources.
1290: For each object, the top panel shows the IA band image where the \lya\ line
1291: is detected, and the bottom panel shows the redward broadband image.
1292: The vertical dotted lines show the slit positions.
1293: }
1294: \label{fig:images}
1295: \end{figure}
1296:
1297: \begin{figure}
1298: \figurenum{4}
1299: \plotone{f4b.eps}
1300: \caption{Images of the 18 extended \lya\ sources. (Continued)}
1301: \end{figure}
1302:
1303: \begin{figure}
1304: \figurenum{4}
1305: \plotone{f4c.eps}
1306: \caption{Images of the 18 extended \lya\ sources. (Continued)}
1307: \end{figure}
1308:
1309: %Figure: Sky distribution
1310: \begin{figure}
1311: \plotone{f5.eps}
1312: \caption{Sky distribution of our objects.
1313: The solid and dashed rectangles show the two MOS fields
1314: of the VIMOS spectroscopy. The 41 objects in the photometric sample are
1315: shown in stars, and the 18 objects in the VIMOS sample are
1316: shown in filled symbols.
1317: The large symbols represent objects brighter than 25.5 mag (AB)
1318: in the IA bands with large IA-excess (Cont - IA $>$ 2.0 mag)
1319: The medium large symbols are objects brighter than 25.5 mag in the
1320: IA bands. }
1321: \label{fig:skydist}
1322: \end{figure}
1323:
1324: % Figure: spectra
1325: \begin{figure}
1326: \plotone{f6a.eps}
1327: \caption{Spectra taken with VIMOS. For each object, three panels
1328: show line profile, residual of Gaussian fitting, and two-dimensional
1329: spectrum (from top to bottom).
1330: On the top panel, the solid curve shows line profile, the dotted
1331: curve shows sky spectrum scaled to 1/100, and the vertical dotted
1332: lines show central wavelength and $\pm 1\sigma$ of the Gaussian
1333: function fitted to the line profile. The Gaussian function is shown with
1334: the dashed curve. }
1335: \label{fig:spec}
1336: \end{figure}
1337:
1338: \begin{figure}
1339: \figurenum{6}
1340: \plotone{f6b.eps}
1341: \caption{Spectra taken with VIMOS. (Continued)}
1342: \end{figure}
1343:
1344: \begin{figure}
1345: \figurenum{6}
1346: \plotone{f6c.eps}
1347: \caption{Spectra taken with VIMOS. (Continued)}
1348: \end{figure}
1349:
1350: \begin{figure}
1351: \figurenum{6}
1352: \plotone{f6d.eps}
1353: \caption{Spectra taken with VIMOS (Continued)}
1354: \end{figure}
1355:
1356: \begin{figure}
1357: \figurenum{6}
1358: \plotone{f6e.eps}
1359: \caption{Spectra taken with VIMOS (Continued)}
1360: \end{figure}
1361:
1362:
1363: % Figure: L-EW
1364: \begin{figure}
1365: \plotone{f7.eps}
1366: \caption{Equivalent width of the \lya\ line as a function of the
1367: \lya\ line luminosity. The horizontal dotted line corresponds to 200\AA.
1368: Objects marked with open stars have EWs whose lower limits exceed 200\AA.
1369: Open triangles show objects with EWs larger than 200\AA,
1370: but with lower limits below 200\AA. The remaining objects,
1371: which have EWs less than 200\AA, are marked with open circles.
1372: }
1373: \label{fig:lew}
1374: \end{figure}
1375:
1376: \begin{figure}
1377: \plotone{f8.eps}
1378: \caption{Velocity width (FWHM) as a function of the \lya\ line
1379: luminosity. Objects with large EWs exceeding 200\AA\ are plotted
1380: with the open stars.
1381: }
1382: \label{fig:ldv}
1383: \end{figure}
1384:
1385: \end{document}
1386:
1387: