1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass{emulateapj}
3: %\usepackage{apjfonts}
4: %\usepackage{epsf}
5: %\usepackage{includegraphics}
6: %\usepackage{times}
7:
8: \newcommand{\etal}{{et al}\/.}
9: %\newcommand{\etal}{et al.\,}
10: \newcommand{\hh}{^{\rm h}}
11: \newcommand{\mm}{^{\rm m}}
12:
13:
14: \begin{document}
15:
16: \slugcomment{Draft of \today}
17:
18: \shorttitle{Deep VLT $V$-band Imaging of a $z = 10$ Candidate}
19:
20: \shortauthors{M. D. Lehnert \etal}
21:
22: \title{Deep VLT $V$-band Imaging of the Field of a $z=10$ Candidate Galaxy:
23: Below the Lyman Limit?\altaffilmark{1}}
24:
25: \altaffiltext{1}{Based on Director's Discretionary time observations
26: collected at ESO-VLT under program 273.A-5028(A)}
27:
28: \author{M. D. Lehnert, N. M. F\"{o}rster Schreiber}
29: \affil{Max-Planck-Institut f\"ur extraterrestrische Physik,
30: Giessenbachstra\ss e, 85748 Garching bei M\"{u}nchen, Germany}
31: \and
32: \author{M. N. Bremer}
33: \affil{Department of Physics, University of Bristol, Tyndall Avenue,
34: Bristol BS8 1TL, UK}
35:
36:
37: \begin{abstract}
38:
39: We present a deep 16.8~ks $V$-band image of the field of a candidate
40: $z{=}10$ galaxy magnified by the foreground ($z{=}0.25$) cluster A1835.
41: The image was obtained with FORS1 on VLT-Kueyen to test whether the
42: $V$-band lies below the Lyman limit for this very high redshift candidate.
43: A detection would unambiguously rule out that the source is at $z{=}10$.
44: The $3\sigma$ detection limit of the image in the area of the $z = 10$
45: candidate is $V_{\rm AB} = 28.0~{\rm mag}$ in a 2\arcsec -diameter
46: aperture (about 3 times the seeing FWHM of 0\farcs 7). No source at
47: the position of the candidate galaxy is detected down to this limit.
48: Formally, this is consistent with the $V$-band probing below the Lyman
49: limit in the rest-frame of a $z = 10$ source. However, given the
50: recent non-detection of the object in a deep $H$-band exposure with
51: NIRI on Gemini North down to $H_{\rm AB} = 26.0~{\rm mag}$ ($3\sigma$
52: in a 1\farcs 4 aperture) and concerns about the detection of the reported
53: associated emission line, it may be possible that this source is spurious.
54: We discuss several astrophysical possibilities to explain the puzzling
55: nature of this source and find none of them compelling.
56:
57: \end{abstract} \keywords{cosmology: observations - early universe -
58: galaxies: distances and redshifts - galaxies: evolution - galaxies:
59: formation}
60:
61: \maketitle
62: \section{Introduction}
63: \label{intro}
64:
65: Motivated by the discovery of high redshift, $z \sim 6$ quasars with what
66: appeared to be Gunn-Peterson troughs \citep{becker01, djorgovski01}, many
67: research groups began to search for the sources responsible for reionization
68: \citep[see {\it e.g.,}][]{L03,bremer04a, S04a, S04b, bunker03, ajiki03,
69: rhoads03, bouwens04, hu04}. The WMAP result of the surprising detection
70: of a large Thompson electron optical depth of $\tau = 0.17 \pm 0.04$
71: \citep{kogut03} and questions about whether the intergalactic opacity at
72: $z \approx 6$ is due to the neutral intergalactic medium (IGM) or to discrete
73: absorbers \citep[e.g.,][]{sc02} have led observers to try and push discovery
74: techniques into the near-infrared (NIR) and beyond the redshift of the most
75: distant Sloan quasars \citep{pello04, kneib04}.
76:
77: To reconcile the possible Gunn-Peterson troughs observed in high redshift
78: quasar spectra and the WMAP results, the fact that the ionizing photon density
79: at high redshift appears to be declining \citep[e.g.,][]{L03,bunker03} and
80: the rapidly increasing density necessary to ionize the IGM at successively
81: higher redshifts \citep[e.g.,][]{madau99} suggest that the universe may
82: have had a complex reionization history. Indeed, these arguments led some
83: researchers to propose complex models such as extended partial reionization
84: \citep[e.g,][]{madau03, ricotti04a} or twice reionization
85: \citep[e.g.,][]{cen03, ciardi03, wyithe03}. In these complex scenarios,
86: the discovery of even one high redshift star-forming galaxy provides
87: powerful constraints on the sources of reionization and on how star
88: formation proceeded in the early universe \citep[e.g.,][]{ricotti04b}.
89:
90: But discovering galaxies at redshifts beyond $z\approx 6$ becomes
91: increasingly challenging. Rest-frame emission longwards of the Lyman
92: limit is redshifted to observed $\lambda > 7500$~\AA\ and the faintness
93: of the galaxies makes them extremely difficult to detect. The essential
94: spectroscopic confirmation is hampered by the dramatic increase in the
95: density of telluric OH emission bands at $\lambda > 7500$~\AA. There are
96: regions $100 - 200$~\AA\ wide that are relatively devoid of OH lines
97: and narrow-band surveys for high redshift sources within these windows
98: have been successful \citep[e.g.,][]{hu04}. Outside of these windows,
99: even determining redshifts of color selected galaxies is generally very
100: difficult at $z \ga 6$.
101:
102: One technique developed to overcome the difficulty of detecting the most
103: distant galaxies takes advantage of gravitational lensing by an intervening
104: galaxy cluster to boost the apparent brightness of background sources.
105: This boost can be as much as a factor of $10 - 100$ along the critical lines
106: for lensing. \cite{santos04} proved the feasibility of this technique
107: out to $z = 5.6$ and \cite{kneib04} discovered a probable lensed $z \sim 7$
108: Lyman break galaxy behind A2218. The efficiency of gravitational lensing
109: assisted searches compared to blank field searches is, however, sensitive
110: to the slope of the luminosity function and thus its overall utility in
111: finding large numbers of high redshift galaxies has yet to be assessed.
112:
113: \cite{pello04} reported the identification of a highly magnified
114: galaxy lying on a critical line of the $z = 0.25$ cluster A1835 (which
115: they denoted A1835-1916). Their data set included broad-band optical
116: imaging from the {\em HST\/} and the CFHT along with NIR imaging and
117: spectroscopy with ISAAC at the VLT. The object was undetected in the
118: $V,R,I$ optical bands, and only detected at $4\sigma$ in $H$ and $3\sigma$
119: in $K$. The $J$-band detection quoted by \citet{pello04} is formally an
120: upper limit. The optical non-detection, the large break between the $J$
121: and $H$ bands, and the blue $H - K_{\rm s}$ color ($H_{\rm AB} - K_{\rm
122: s,AB} < 0$) found by \citeauthor{pello04}could possibly indicate a young
123: galaxy at extremely high redshift. In their $J$-band spectroscopy
124: \citeauthor{pello04}reported an emission line at 1.33745~\micron,
125: detected in two separate wavelength settings of the spectrograph and with
126: a combined significance of $4 - 8\,\sigma$. The photometry, together
127: with the lensing model suggesting the source lies on a caustic for very
128: high redshift (with a magnification factor between 25 and 100 as being
129: most likely), led \citeauthor{pello04}to argue that the line is most
130: likely Ly$\alpha$ at $z = 10.0175$. Unfortunately, the signal-to-noise
131: ratio of the spectrum is insufficient to show the telltale signature
132: of the line profile asymmetry of highly redshifted Ly$\alpha$ to rule
133: out other line identifications. Given the uncertainties in the lensing
134: model, the most important piece of evidence upon which the conclusion
135: of $z = 10$ rests is the shape of the spectral energy distribution (SED)
136: measured from the imaging data.
137:
138: However, the high redshift nature of this source has been recently
139: questioned. Based on new independent $H$-band data obtained with NIRI at
140: the Gemini North telescope, \citet{bremer04b} did not detect the $z = 10$
141: candidate down to the $3\sigma$ limit of $H_{\rm AB} = 26.0~{\rm mag}$.
142: This limit is 1 magnitude deeper than the $4\sigma$ detection quoted by
143: \citet{pello04} in their ISAAC data. This significantly weakens the main
144: evidence for a redshift of 10, which relies on the strength of the break
145: between the optical and $J$ bands and the $H$ band. The photometry no
146: longer constrains the redshift and other identifications for the emission
147: line from a lower redshift source remain possible.
148:
149: Before the \cite{bremer04b} results were obtained, we were awarded
150: Director's Discretionary Time with FORS1 at the VLT to conduct
151: $V$-band imaging of the field around the $z = 10$ candidate and push
152: the sensitivity to fainter levels than presented by \citet{pello04}.
153: A $V$-band detection would be decisive: it would demonstrate beyond
154: any doubt that the source is {\em not\/} at $z = 10$. Besides probing
155: rest-frame wavelengths $\sim 500$~\AA\ for $z = 10$, well below the Lyman
156: limit, we chose the $V$-band to reach within a reasonable observing
157: time very sensitive limits compared to other optical or NIR bands and
158: thus provide the best chance of detecting a reddened $z = 1-2$ object,
159: which seems a likely alternative \citep[see][]{bremer04b}.
160:
161: \section{Data and analysis}
162:
163: \subsection{Observations and Data Reduction}
164:
165: The $V$-band observations of A1835 were carried out during Director's
166: Discretionary Time on the nights of 2004 July 9, 16, and 19 (UT). We used
167: FORS1 on the VLT-Kueyen telescope in imaging mode with a projected scale
168: of $\rm 0\farcs 200~pixel^{-1}$. Forty-eight separate frames of 350~s
169: each were taken, for a total integration time of 16.8~ks. Half of the
170: integration time was taken on 2004 July 16 and one-third on 2004 July
171: 19, with only a small fraction (1/6) on 2004 July 9. The conditions
172: were clear and the seeing was typical for the VLT in the optical,
173: varying from 0\farcs 5 to 1\farcs 0 with a median of about 0\farcs 7.
174: Each individual dithered exposure had a unique (non-redundant) pointing
175: position, centered around the location of the $z = 10$ candidate.
176:
177: We reduced the data within IRAF as follows. We subtracted the bias
178: from each frame using bias frames taken at the beginning and end of
179: each night. We flattened each frame initially with flat field images
180: produced by a combination of dithered sky images taken during twilight.
181: From these bias-subtracted and flat-fielded images, we created a master
182: sky flat by combining all of the dithered exposures of the target
183: field, masking out all sources above $1\sigma$ of the average rms of
184: each frame (the rms calculated including the whole frame). This flat
185: was divided into each dithered frame. From the master sky flat, we
186: also generated the bad pixel map used to exclude the bad pixels in each
187: frame during the final combination. We determined the offsets between
188: individual exposures by fitting the position of about 30 bright,
189: unsaturated stars. We co-averaged the reduced, registered frames using
190: three different weightings: \footnote{The weight $w$ is applied
191: as multiplicative factor $w^{-2}$.} (1) uniform weighting, (2) weighting
192: by the seeing FWHM of each frame, which gives most weight to the frames
193: with best seeing, and (3) weighting by the FWHM times the squared root
194: of the average sky counts in each frame, which accounts for the rms
195: noise within the seeing disk and optimizes for point-source detection.
196: There is little variation in the sensitivity of the final images between
197: the three different schemes. The latter two produce final combinations
198: with the best FWHM of $\sim 0\farcs 7$. For the analysis, we adopted
199: the image combined following the third weighting scheme.
200:
201: To flux-calibrate the data we measured the photometric zero points using
202: exposures of standard stars taken at the beginning and end of each night.
203: We compared these measurements to zero points estimated throughout
204: the month encompassing the range of observing dates. All zero points
205: were in excellent agreement (within $\rm \lesssim0.02~mag$) and the final
206: calibration is based on the average measured by ESO during 2004 July when
207: our data were taken. \footnote{The ESO FORS1 zero points are available at
208: http://www.eso.org.} The average sky brightness in the final image is $V =
209: 21.5~{\rm mag~arcsec^{-2}}$, and varies among the individual frames by a
210: few tenths of a magnitude. \footnote{We refer to the Vega photometric
211: system when not explicitely indicating AB magnitudes. For the $V$
212: band, the AB correction is very small, with $V_{\rm AB} - V_{\rm Vega}
213: = 0.014~{\rm mag}$ for FORS1 accounting for the full system transmission.}
214:
215: \subsection{Photometry}
216:
217: Figure~\ref{fig-Vmap} shows the $20^{\prime\prime} \times
218: 20^{\prime\prime}$ region in our $V$-band image around the reported
219: location of the candidate $z = 10$ galaxy. We do not detect any source
220: at this position down to the faintest levels reached in our data.
221: This implies a $3\sigma$ limit of $V_{\rm AB} > 28.0~{\rm mag}$ in a
222: 2\arcsec -diameter aperture, as described below. \footnote{This aperture
223: size corresponds to 3 times the seeing FWHM in the combined data and has
224: a negligible aperture correction except for the brightest point sources
225: and the most extended objects. We note that it leads to conservative
226: limits for faint point sources, for which the S/N is optimized with an
227: aperture of 1.5 times the FWHM.} Our limit is 0.6~mag deeper than the
228: $3\sigma$ limit of \citet{pello04} given for a 0\farcs 6 aperture, or
229: about 3.5~mag deeper for an aperture of 3 times the FWHM (0\farcs 76)
230: of their $V$-band data, assuming uncorrelated Gaussian noise.
231:
232: We assessed our ability to reliably set a meaningful upper limit to
233: any possible source at the position of the $z = 10$ candidate using
234: three different techniques. First we established the pixel-to-pixel rms
235: variations in the background across the entire image. For this purpose,
236: we used the Sextractor software \citep{bertin96}. Sextractor estimates
237: the local background level in a mesh grid over the image. These local
238: background estimates are iteratively clipped until they converge to
239: $\pm 3\sigma$ around the median of all meshes. The histogram of all
240: remaining unclipped pixels is then used to determine the rms variations
241: of the background noise in the image. This procedure yielded an
242: overall $3\sigma$ detection limit of $V_{\rm AB} = 27.6~{\rm mag}$
243: in a 2\arcsec -diameter aperture. However, this limit is pessimistic
244: because our image is very crowded with cluster and background galaxies
245: at the depths reached: the surface density of faint sources is such
246: that there is typically only about 4\arcsec\ between adjacent objects.
247: Thus, the rms variations derived in this way partly reflect extended
248: light profiles of the myriad of galaxies in the image.
249:
250: To obtain more realistic estimates of our detection limits in the area
251: around the $z = 10$ candidate, we focussed our analysis to the region
252: within $\pm 20$\arcsec\ of its reported location. We placed 30-60
253: apertures 2\arcsec\ in diameter all away from the outer light profile of
254: any obvious object in this $\rm 1600~arcsec^{2}$ area. This procedure was
255: repeated several times with differing numbers of apertures and positions
256: to ensure that our final estimates were not critically dependent on the
257: number of apertures or their exact placement. The average pixel-to-pixel
258: rms within these apertures implies a $3\sigma$ detection limit of
259: $V_{\rm AB} = 28.0~{\rm mag}$, with variations of $\approx 0.1~{\rm
260: mag}$ between the apertures. The variation in the total residual flux
261: within each aperture implies a $3\sigma$ detection limit of $V_{\rm AB}
262: = 27.8~{\rm mag}$.
263:
264: To further determine the robustness of our detection limits, we placed
265: 20 artificial point sources each at 3 different AB magnitudes, 27.6,
266: 27.8 and 28.0, within $\pm 20$\arcsec\ of the reported location of the
267: $z = 10$ candidate and avoiding all obvious sources in this region.
268: The high degree of crowding limits the number of point souces that
269: can be placed non-redundantly close to the position of the $z = 10$
270: candidate to about 20. We recovered the sources down to $V_{\rm AB}
271: = 28.0~{\rm mag}$ at a rate similar to that expected for a $3\sigma$
272: detection limit (i.e. 50\%), and the dispersion between the measured and
273: true input brightness was $\approx \pm 0.3-0.4~{\rm mag}$. Given the
274: excellent agreement between the two most robust methods, we adopted
275: $V_{\rm AB} = 28.0~{\rm mag}$ as the detection limit in a 2\arcsec\
276: aperture at the position of the $z = 10$ candidate.
277:
278: \section{Discussion}
279: \label{discuss}
280:
281: The non-detection in our VLT FORS1 $V-$band data down to a faint limit has
282: implications for the nature of the source investigated by \citet{pello04}.
283: Formally, a non-detection is consistent with the candidate having a
284: redshift of 10. The Bessel $V$ band filter has an effective wavelength
285: of 5540~\AA\ and half maximum transmission at about 5000 and 6000~\AA.
286: For $z = 10$, these wavelengths correspond to 500~\AA\ and the range
287: $\approx 450 - 550$~\AA\ in the rest-frame --- well below the Lyman limit
288: and completely opaque. The lower limit to the redshift, if the $V$-band
289: non-detection is caused solely by the IGM opacity below the Lyman limit,
290: is $\sim5.6$. However, there is substantial opacity within the Lyman
291: forest at these redshifts and thus $V$-band Lyman break or drop-out
292: galaxies have redshifts that are usually less than this \citep[see
293: e.g.,][]{bremer04a}.
294:
295: Although our $V$-band non-detection may allow for $z \ga 5.6$, or even $z =
296: 10$, this is not the only interpretation possible in view of other lines
297: of evidence that have recently come to light. There are certainly two
298: other hypotheses that are equally plausible, and perhaps significantly
299: more likely.
300:
301: The strongest evidence for the $z = 10$ interpretation presented by
302: \citet{pello04} relied on the large break among the optical, $J$,
303: and $H$ bands, and the subsequent detection of an emission line
304: at 1.33745~\micron. However, this evidence has now been called into
305: question. \cite{bremer04b} did not detect the candidate object in their
306: new independent and $\rm 1~mag$ deeper $H$-band image from NIRI down
307: to $H_{\rm AB} = 26.0$ ($3\sigma$). This greatly weakens the argument
308: based on the break strength (as well as the blue $H - K_{\rm s}$ color)
309: that supported the claim for the emission line being Ly$\alpha$ at $z
310: \approx 10$. Furthermore, \cite{Weatherley04} have recently questioned
311: the robustness of the line detection. However, this was based on a
312: re-analysis of the spectroscopic data used by \cite{pello04} and subtle
313: differences in reduction techniques could lead to contentious results.
314: It is probably reasonable to conclude that the significance of the line
315: is difficult to judge.
316:
317: If the object is not at a redshift of 10, then what could it be? One way
318: forward is to assume the reality of the line, use the sensitive $H$-band
319: limit of \cite{bremer04b}, and assume an alternative line identification
320: to determine if the emission properties are reasonable for a galaxy at
321: a lower redshift. \citeauthor{bremer04b}argued that the source could
322: be a dwarf/HII galaxy at intermediate redshift and that lines such as
323: the [\ion{O}{2}]$\lambda\lambda 3726,3729$ doublet, [\ion{O}{3}]$\lambda
324: 5007$, or H$\alpha$ at $1.0 < z < 2.6$ are the most likely alternatives to
325: Ly$\alpha$. If it were [\ion{O}{3}], then the line width and luminosity
326: would be consistent with the local relationship for dwarf/\ion{H}{2}
327: galaxies between these quantities \citep[e.g.,][]{melnick00}. In light
328: of our new $V$-band results, we wish to further this argument.
329:
330: Assuming the line is [\ion{O}{3}] at $z = 1.67$ and adopting reasonable
331: cosmological parameters ($H_{0} = 70 {\rm km\,s^{-1}\,Mpc^{-1}}$,
332: $\Omega_{\rm m} = 0.3$, and $\Omega_{\Lambda} = 0.7$), our $V$-band limit
333: implies an unlensed absolute magnitude fainter than $-16.4~{\rm mag}$
334: at 2080~\AA. Since there is no published detailed analysis of the mass
335: model of the intervening cluster, it is difficult to quantitatively
336: assess the magnification factor for $z = 1.67$ (\citeauthor{pello04}
337: suggested a magnification factor of between 25 and 100 for a source
338: at $z = 10$). Conservatively, we assume no lensing; lensing will
339: only make the source intrinsically fainter. The co-moving density
340: of galaxies in the local universe with absolute magnitude $-16.5$ at
341: 2000~\AA\ is $\phi_{2000} \approx 10^{-1.5}~{\rm Mpc^{-3}\,mag^{-1}}$
342: \citep{treyer98}. The number of such sources on the sky with $1.45 <
343: z < 1.95$ is then $\approx 50~{\rm arcmin^{-2}}$ if the UV luminosity
344: function does not evolve out to $z = 1.7$. The luminosity function of
345: $z \approx 3$ Lyman break galaxies implies a roughly similar surface
346: density for such faint objects \citep{SLBGs4}. Comparable densities
347: would be inferred to a factor of $\sim$1.5 if the line were H$\alpha$
348: at $z = 1.04$ or [\ion{O}{2}] at $z = 2.59$.
349:
350: The surface density of sources down to the detection limit of our image
351: is $>$50 arcmin$^{-2}$ (modulo the incompleteness). Therefore the
352: probability of finding and perhaps even mistaking it for a very high
353: redshift source is consequently high. The immediate explanation is
354: that the source might simply be fainter. However, this poses problems
355: when trying to reconcile the broad-band flux limits with the reported
356: emission line flux.
357:
358: The $3\sigma$ $H$-band limit obtained by \citet{bremer04b} coupled with
359: the line flux given by \citet{pello04} implies an observed equivalent
360: width (EW) of $W_{\rm obs} > 170$~\AA\ (assuming a flat $f_{\nu}$
361: continuum between 1.65 and 1.34~\micron; if the source were redder
362: [bluer] as suggested by \citeauthor{bremer04b}, the EW limit would
363: be higher [lower]). Assuming the line is [\ion{O}{3}] at $z = 1.67$,
364: the corresponding rest-frame EW is $W_{\rm rest} \ga 60$~\AA. For the
365: alternative case of H$\alpha$ at $z = 1.04$, $W_{\rm rest} \ga 80$~\AA,
366: and of [\ion{O}{2}], $W_{\rm rest} \ga 45$~\AA. Locally, such large
367: EWs are only found in galaxies with low metallicity and blue colors
368: \citep[][]{sullivan00}.
369:
370: Although the \citeauthor{bremer04b}$H$-band limit sets a reasonably
371: stringent lower limit on the EW of the line, our $V$-band limit may
372: provide an even stronger constraint. Locally, sources with large
373: emission line EWs are very blue, with ${\rm UV} - B$ colors always bluer
374: than for a flat $f_{\nu}$ spectrum \citep{sullivan00}. For sources
375: at $z\approx1.7$ this implies observed $V_{AB}-H_{AB}\lesssim 0$.
376: Supposing again the source is at $z = 1.67$ and assuming conservatively
377: it has a flat $f_{\nu}$ spectrum, our limit of $V_{\rm AB} > 28.0~{\rm
378: mag}$ implies an $H$-band limit of $H_{\rm AB} > 28.0~{\rm mag}$ and,
379: repeating the reasoning above, a rest-frame EW for [\ion{O}{3}] of
380: $W_{\rm rest} > 400$~\AA ($\gtrsim$500~\AA\ if H$\alpha$, $\gtrsim$300~\AA\
381: if [\ion{O}{2}]). For a bluer spectrum as observed locally, the line
382: EW limit only increases. A similar argument would apply to any other
383: optical line. Such high EWs are rarely observed, would require a very
384: young age ($\rm < 10~Myr$) and perhaps an initial mass function heavily
385: skewed toward massive stars.
386:
387: Summarizing the above arguments, our sensitive $V$-band limit
388: implies a faint intrinsic UV ($\sim 2000$~\AA) absolute magnitude.
389: Plausible luminosity functions predict a high surface density of
390: such objects and thus a high probability of detection if the source
391: were just bright enough to match the $3\sigma$ limiting magnitude of
392: our $V$-band image. No object is detected at the expected location
393: despite the high density $\ga 50~{\rm arcmin^{-2}}$ of $3 - 4\sigma$
394: sources in the data. Postulating then that the object could be fainter,
395: the lower limits on the rest-frame EW of the emission line become very
396: large and the probability of detecting such a source becomes very low.
397: Thus, in view of the simple astrophysics of line emission, the scenario
398: of an intermediate redshift dwarf does not appear very likely given
399: the difficulties in reconciling the broad-band limits with the emission
400: line flux. Additionally, although largely circumstantial, the arguments
401: could perhaps cast doubt on the reported line flux \citep{Weatherley04}.
402:
403: The other possible scenario is that the source may be transient or
404: variable. The NIR ISAAC images and spectra presented by \citet{pello04}
405: were taken in 2003 February and late June to early July, respectively.
406: The NIRI $H$-band data were collected during two nights in 2004 late
407: May and early June. The $V$-band data discussed here were obtained in
408: 2004 mid-July. Overall, these observations span a period of about a
409: year and a half.
410:
411: Given the ecliptic latitude of the A1835 field at about 14$^{\circ}$,
412: one hypothesis could be that the source is a solar system object with
413: a large proper motion. The tightest constraint comes from the time
414: over which all of the \citet{pello04} ISAAC $H$-band data were taken.
415: The observations were carried out over 24 hours on 2003 February 15
416: and 16 and the seeing in the final $H$-band image is $\sim 0\farcs 5$.
417: Conservatively, for the source not to appear extended would imply
418: an angular velocity of $< 0\farcs 5 / 24~{\rm h}$ or $< 0\farcs 02
419: {\rm h}^{-1}$. If the source is bound to the Sun and consistent with
420: a distant object like Pluto, for example, with a tangential velocity
421: of $5~{\rm km\,s^{-1}}$, it would lie at more than D=1240~AU. To be
422: unresolved in the 0\farcs 5 $H$-band image at the distance derived
423: from the proper motion limit, the source would have to be smaller than
424: $\sim 10$ Jupiter diameters. The brightness of an object illuminated
425: by reflected sunlight and assuming an albedo of one, is related to its
426: angular diameter, $\theta$, by m = m$_{\odot}$ - 5 log($\theta$/2D).
427: From the size calculated above the source would have $H\approx$55 mag.
428: An object with the magnitude as claimed by \cite{pello04} would have to
429: be a million times larger than what we roughly estimated and thus would
430: violate the size constraint by 6 orders of magnitude. It is possible to
431: relax these constraints by an order of magnitude, but even given more
432: optimistic estimates, a solar system object seems highly unlikely.
433:
434: It is more difficult to estimate the probably that the source is a
435: high redshift supernova, $\gamma$-ray burst or perhaps something more
436: exotic \citep[e.g.,][and references therein]{stern04}. \cite{dahlen04}
437: determined a SN rate of about 5.5 $\times$ 10$^{-4}$ SN yr$^{-1}$
438: Mpc$^{-3}$ at $<z>$$\approx$0.7-0.8 and using a model for the
439: star-formation rate density evolution, they estimate about a factor
440: of 2 increase in the SN rate at z$\approx$2. The magnitude at which
441: \citeauthor{pello04}claim to have detected the source is $H_{AB}=25.0$.
442: It is difficult to know what the redshift of the source might be,
443: especially since the photometry for a transient provides no constraints
444: on the redshift. It is worth noting that distant SNe generally have
445: optical magnitudes fainter than the \citeauthor{pello04}detection
446: \citep[e.g.,][]{stolger04,dahlen04} and given the relatively blue
447: SED of SNe this means that the $z{=}10$ candidate would have been one
448: of the brighter SNe (it is of course magnified by an unknown amount
449: which could account for its relative brightness). At any rate, we can
450: estimate the chances of finding a SNe serendipitously in their field.
451: Assuming that the SN was caught within 30 days of its peak, that the
452: redshift range of the SN was 0.5-1.0 \citep[roughly consistent with the
453: range in][]{dahlen04}, and that the area in the detection image was 6
454: arcmin$^2$, we would have expected 0.22 SN in the field based on the
455: \cite{dahlen04} SNe rate density. Thus the candidate as an intermediate
456: redshift SN cannot be ruled out. However, this is surely optimistic given
457: the magnitude of the detection claimed in the \citeauthor{pello04}$H$-band
458: image and the unknown magnification. Lensing would decrease the
459: probability by decreasing the effective area sampled. Requiring the
460: SN to be observed more closely to its peak in order make detection more
461: likely would (linearly) lower the relative probability of observing it.
462:
463: Neither of the transient hypotheses can explain the line detection.
464: One would have to postulate that it comes from the host galaxy in the case
465: of a SNe from which there has been no subsequent continuum detection.
466: Given the EW arguments we made previously, this would have to be a
467: very unusual object to escape detection. For a solar system object,
468: the gap between the dates of the imaging and spectroscopy means that
469: the source would have moved a significant distance from the discovery
470: position and not within the subsequent slit spectroscopy. Therefore,
471: none of the possible transient hypotheses explain simply or logically
472: all of the claims in \cite{pello04}. Given the $H$-band non-detection
473: of the source down to faint limits by \citeauthor{bremer04b}and now in
474: the V-band, and, the unlikely nature of any galaxy with the properties
475: (lower-limits) observed, it is tempting to conclude that the most likely
476: explanation is that the source is spurious.
477:
478: \acknowledgments
479:
480: We would like thank the referee, Pat McCarthy, for his comments, and
481: Roser Pello and Daniel Schaerer for their arguments and insights.
482: We wish to thank the ESO DG, Catherine Cesarsky for the generous
483: allocation of observing time and Bruno Leibundgut for his curiosity
484: and support. We are grateful to the ESO USG, DMD, and Paranal staff for
485: conducting these observations, Mario van den Ancker and Roberto Mignani,
486: in particular, for their dedication and effort.
487:
488: \begin{thebibliography}{}
489:
490: \bibitem[Ajiki et al. (2003)]{ajiki03}Ajiki, M. et al. 2003, \aj,
491: 126, 2091
492:
493:
494: \bibitem[Becker et al. (2001)]{becker01}Becker, R. H. et al. 2001, \aj,
495: 122, 2850
496:
497: \bibitem[Bertin \& Arnouts(1996)]{bertin96} Bertin, E., Arnouts, S.
498: 1996, \aaps, 117, 393
499:
500: \bibitem[Bouwens et al. (2004)]{bouwens04}Bouwens, R. J et al. 2004,
501: \apj, 606, 25
502:
503: \bibitem[Bremer et al. (2004a)]{bremer04a}Bremer, M. N., Lehnert, M. D.,
504: Waddington, I., Hardcastle, M. J., Boyce, P. J., \& Phillipps, S. 2004a,
505: \mnras, 347, 7
506:
507: \bibitem[Bremer et al. (2004b)]{bremer04b} Bremer, M. N., Jensen, J.,
508: Lehnert, M. D., F\"{o}rster Schreiber, N., Douglas, L. 2004b, \apj,
509: 615, L1
510:
511: \bibitem[Bunker et al. (2003)]{bunker03}Bunker, A. J., Stanway, E. R.,
512: Ellis, R. S., McMahon, R. G., \& McCarthy, P. J. 2003, \mnras, 342, 47
513:
514: \bibitem[Cen (2003)]{cen03} Cen, R. 2003, \apj, 591, 12
515:
516: \bibitem[Ciardi et al. (2003)]{ciardi03} Ciardi, B., Ferrara, A., \&
517: White, S. D. M. 2003, \mnras, 344, L7
518:
519: \bibitem[Dahlen et al. (2004)]{dahlen04} Dahlen, T. et al. 2004, ApJ, 613, 189
520:
521: \bibitem[Djorgovski et al. (2001)]{djorgovski01}Djorgovski, S. G., Castro, S., Stern, D., \& Mahabal, A. A. 2001, \apj, 560, 5
522:
523: \bibitem[Hu et al. (2004)]{hu04}Hu, E. M., Cowie, L. L., Capak, P.,
524: McMahon, Richard G., Hayashino, T., \& Komiyama, Y. 2004, \aj, 127, 563
525:
526: \bibitem[Kneib et al. (2004)]{kneib04}Kneib, J.-P., Ellis, R. S., Santos,
527: M. R., \& Richard, J. 2004, \apj, 607, 697
528:
529: \bibitem[Kogut et al.(2003)]{kogut03} Kogut, A., et al.\ 2003,
530: \apjs, 148, 161
531:
532: \bibitem[Lehnert \& Bremer (2003)]{L03} Lehnert, M. D. \& Bremer, M. N.
533: 2003, \apj, 593, 630
534:
535: \bibitem[Madau, Haardt, \& Rees (1999)]{madau99} Madau, P., Haardt, F.,
536: \& Rees, M. J. 1999, \apj, 514, 648
537:
538: \bibitem[Madau et al. (2004)]{madau03} Madau, P., Rees, M. J., Volonteri,
539: M., Haardt, F., \& Oh, S. P. 2004, \apj, 604, 484
540:
541: \bibitem[Melnick, Terlevich, \& Terlevich (2000)]{melnick00}Melnick,
542: J., Terlevich, R., \& Terlevich, E. 2000, \mnras, 311, 629
543:
544: \bibitem[Pell\'{o} et al. (2004)]{pello04} Pell\'{o}, R., Schaerer, D.,
545: Richard, J., Le Borgne, J.-F., \& Kneib, J.-P. 2004, \aap, 416, 35
546:
547: \bibitem[Ricotti \& Ostriker (2004)]{ricotti04a} Ricotti, M., \& Ostriker,
548: J. P. 2004, \mnras, 350, 539
549:
550: \bibitem[Ricotti et al. (2004)]{ricotti04b}Ricotti, M., Haehnelt, M. G.,
551: Pettini, M., \& Rees, M. J., 2004, \mnras, 352, L21
552:
553: \bibitem[Rhoads et al. (2003)]{rhoads03}Rhoads, J. E. et al. 2003, \aj,
554: 125, 1006
555:
556: \bibitem[Santos et al. (2004)]{santos04}Santos, M. R., Ellis, R. S.,
557: Kneib, J.-P., Richard, J., \& Kuijken, K. 2004, \apj, 606, 683
558:
559: \bibitem[Songaila \& Cowie (2002)]{sc02} Songaila, A., \& Cowie,
560: L. L. 2002, \aj, 123, 2183
561:
562: \bibitem[Stanway et al. (2004a)]{S04a} Stanway, E. R., Bunker, A. J.,
563: McMahon, R. G., Ellis, R. S., Treu, T., \& McCarthy, P. J. 2004, \apj,
564: 607, 704
565:
566: \bibitem[Stanway et al. (2004b)]{S04b}Stanway, E. R. et al. 2004, \apj,
567: 604, 13
568:
569: \bibitem[Steidel et al. (1999)]{SLBGs4}Steidel, C. C., Adelberger, K. L.,
570: Giavalisco, M., Dickinson, M., \& Pettini, M. 1999, \apj, 519, 1
571:
572: \bibitem[Stern et al. (2004)]{stern04} Stern, D. et al. 2004, \apj,
573: 612, 690.
574:
575: \bibitem[Stolger et al. (2004)]{stolger04}Stolger, L.-G., et al. 2004,
576: \apj, 613, 200
577:
578: \bibitem[Sullivan et al. (2000)]{sullivan00} Sullivan, M., Treyer,
579: M. A., Ellis, R. S., Bridges, T. J. Milliard, B., \& Donas, J. 2000,
580: \mnras, 312, 442
581:
582: \bibitem[Treyer et al. (1998)]{treyer98} Treyer, M. A., Ellis, R. S.,
583: Milliard, Donas, J., \& Bridges, T. J. 1998, \mnras, 300, 303
584:
585: \bibitem[Weatherley et al. (2004)]{Weatherley04}Weatherley, S.J., Warren,
586: S.J., Babbedge, T.S.R. 2004, astro-ph/0407150, to appear in \aap Letters
587:
588: \bibitem[Wyithe \& Loeb (2003)]{wyithe03} Wyithe, J. S. B., \& Loeb,
589: A. 2003, \apj, 588, L69
590:
591: \end{thebibliography}
592:
593: \begin{figure}
594: \vspace{2cm}
595: \begin{center}
596: \includegraphics[scale=0.8]{f1.eps}
597: \end{center}
598: \caption{Deep $V$-band FORS1 image of the region around the $z{=}10$
599: candidate object reported by \cite{pello04}. The image is approximately
600: $20^{\prime\prime} \times 20^{\prime\prime}$ with seeing of $\sim 0\farcs
601: 7$. The circle indicates the location of the candidate object, and is 2.0
602: arcsec in diameter. There is no evidence for emission from the candidate
603: in the V-band down to a 3$\sigma$-limit of $V_{\rm AB}=28.0~{\rm mag}$.
604: \label{fig-Vmap}}
605: \end{figure}
606:
607: \end{document}
608: