1: %\documentclass[preprint,12pt]{aastex}
2: \documentclass{emulateapj}
3:
4: \newcommand{\co}{\mbox{\rm $^{12}$CO}}
5: \newcommand{\cothree}{{\rm $^{13}$CO}}
6: \newcommand{\cn}{{\rm C$^0$}}
7: \newcommand{\cp}{{\rm C$^+$}}
8: \newcommand{\ci}{\mbox{\rm [\ion{C}{1}]}}
9: \newcommand{\oi}{\mbox{\rm [\ion{O}{1}]}}
10: \newcommand{\cii}{\mbox{\rm [\ion{C}{2}]}}
11: \newcommand{\oiii}{\mbox{\rm [\ion{O}{3}]}}
12: \newcommand{\hi}{\mbox{\rm \ion{H}{1}}}
13: \newcommand{\hii}{\mbox{\rm \ion{H}{2}}}
14: \newcommand{\htwo}{\mbox{\rm H$_2$}}
15: \newcommand{\jone}{\mbox{($1\rightarrow0$)}}
16: \newcommand{\jtwo}{\mbox{($2\rightarrow1$)}}
17: \newcommand{\jthree}{($3\rightarrow2$)}
18: \newcommand{\jfour}{($4\rightarrow3$)}
19: \newcommand{\fsone}{($^3$P$_1\rightarrow^3$P$_0$)}
20: \newcommand{\fstwo}{($^3$P$_2\rightarrow^3$P$_1$)}
21: \newcommand{\fseno}{($^3$P$_0\rightarrow^3$P$_1$)}
22: \newcommand{\fsowt}{($^3$P$_1\rightarrow^3$P$_2$)}
23: \newcommand{\fscii}{($^2$P$_{3/2}\rightarrow^2$P$_{1/2}$)}
24: \newcommand{\percmcu}{cm$^{-3}$}
25: \newcommand{\percmsq}{cm$^{-2}$}
26: \newcommand{\persr}{sr$^{-1}$}
27: \newcommand{\kmpers}{\mbox{km~s$^{-1}$}}
28: \newcommand{\Kkmpers}{\mbox{K~km~s$^{-1}$}}
29: \newcommand{\intunits}{erg~s$^{-1}$~cm$^{-2}$~sr$^{-1}$}
30: \newcommand{\intensity}{erg~s$^{-1}$~cm$^{-2}$}
31: \newcommand{\fluxdens}{erg~s$^{-1}$~cm$^{-2}$~$\mu$m$^{-1}$}
32: \newcommand{\xcounits}{\mbox{cm$^{-2}$ (K km s$^{-1}$)$^{-1}$}}
33: \newcommand{\Ico}{\mbox{\rm I$_{\rm CO}$}}
34: \newcommand{\Lco}{\mbox{\rm L$_{\rm CO}$}}
35: \newcommand{\Lsun}{\mbox{\rm L$_{\odot}$}}
36: \newcommand{\Ici}{\mbox{\rm I$_{\rm [CI]}$}}
37: \newcommand{\Icii}{\mbox{\rm I$_{\rm [CII]}$}}
38: \newcommand{\Lcii}{\mbox{\rm L$_{\rm [CII]}$}}
39: \newcommand{\Fcii}{\mbox{\rm F$_{\rm [CII]}$}}
40: \newcommand{\Ifir}{\mbox{\rm I$_{\rm FIR}$}}
41: \newcommand{\Lfir}{\mbox{\rm L$_{\rm FIR}$}}
42: \newcommand{\Ffir}{\mbox{\rm F$_{\rm FIR}$}}
43: \newcommand{\citoco}{\Ici/\Ico}
44: \newcommand{\ciitoco}{\Icii/\Ico}
45: \newcommand{\ciitoci}{\Icii/\Ici}
46: \newcommand{\xco}{\mbox{\rm X$_{\rm CO}$}}
47: \newcommand{\aco}{\mbox{$\alpha_{\rm CO}$}}
48: \newcommand{\av}{\mbox{\rm A$_{\rm V}$}}
49: \newcommand{\xuv}{\mbox{$\chi_{\rm uv}$}}
50: \newcommand{\vlsr}{\mbox{V$_{\rm LSR}$}}
51:
52: %\received{2004 February 25}
53: \begin{document}
54: \title{The Faintness of the 158 $\mu$m [\ion{C}{2}] Transition \\
55: in the z=6.42 Quasar SDSS J1148+5251}
56: \author{Alberto D. Bolatto}
57: \affil{Department of Astronomy \& Radio Astronomy Laboratory,\\
58: University of California at Berkeley, \\
59: 601 Campbell Hall, Berkeley, CA 94720-3411, USA}
60: \author{James Di Francesco \& Chris J. Willott}
61: \affil{National Research Council of Canada, Herzberg Institute of
62: Astrophysics, \\
63: 5071 West Saanich Road, Victoria, BC, V9E 2E7, CANADA}
64:
65: \begin{abstract}
66: We report the non-detection of the \cii\ \fscii\ 157.74 $\mu$m
67: transition in the $z=6.42$ quasar SDSS J1148+5251 after 37.5 hours of
68: integration with the James Clerk Maxwell Telescope. This transition
69: is the main cooling line of the star-forming interstellar medium, and
70: usually the brightest FIR line in galaxies. Our observed 1 $\sigma$
71: RMS = 1.3 mK in the $T_{A}^{*}$ scale translates to $\Lcii<2.6 \times
72: 10^{9}$ L$_\odot$. Using a recent estimate of the far-infrared
73: continuum of this quasar, we derive for SDSS J1148+5251
74: $\Lcii/\Lfir<5\times10^{-4}$, a ratio similar to that observed in
75: local ultra-luminous infrared galaxies but considerably smaller than
76: what is typical in nearby normal and starburst galaxies. This
77: indicates that the small \Lcii/\Lfir\ ratio observed locally in
78: luminous far-infrared objects also persists at the highest redshifts.
79: \end{abstract}
80:
81: \keywords{quasars: individual (J114816.64+525150.3) --- quasars:
82: emission lines --- early universe --- radio lines: galaxies}
83:
84: \section{Introduction}
85:
86: The most distant quasar in the universe known at the time of this
87: writing is SDSS J114816.64+525150.3 (hereafter SDSS J1148+5251) at a
88: redshift $z=6.42$ (Fan et al. 2003; Walter et al. 2003; Bertoldi et
89: al. 2003a). With a bolometric luminosity of L$_{\rm bol}\sim10^{14}$
90: L$_\odot$, this is an extremely luminous object powered by a
91: $\sim3\times10^{9}$ M$_\odot$ supermassive black hole at its core
92: (Willott et al. 2003). The high far-infrared (FIR) luminosity of this
93: object implies that the host galaxy is forming stars at the prodigious
94: rate of 3000 M$_\odot$ yr$^{-1}$ (Bertoldi et al. 2003b). At this
95: redshift the universe is only 840 million years old and both the black
96: hole mass and star formation rate imply the existence of a very
97: massive galaxy formed from one of the rarest high-density peaks in the
98: matter distribution. Although there are hints that SDSS J1148+5251 may
99: be weakly gravitationally lensed (White et al. 2003), we assume no
100: lensing magnification in this paper.
101:
102: The recent detection of bright molecular and FIR continuum emission
103: from the host galaxy of SDSS J1148+5251 (Walter et al. 2003; Bertoldi
104: et al. 2003a; Bertoldi et al. 2003b) prompted us to observe the \cii\
105: \fscii\ 157.74 $\mu$m fine structure transition in this object. This
106: is the main cooling transition of the star-forming interstellar medium
107: (e.g., Tielens \& Hollenbach 1985), and is commonly the single
108: brightest emission line in galaxies. For example, Stacey et
109: al. (1991) found that most galaxies emit 0.1\% to 1\% of their
110: far-infrared luminosities in this line alone.
111:
112: Given its extremely high luminosity and its relationship to star
113: formation activity, the redshifted \cii\ 158 $\mu$m transition is
114: attractive as a star formation indicator at high-$z$ (e.g., Stark
115: 1997). This transition is conveniently placed into the atmospheric
116: 1~mm window for redshifts $z\gtrsim6.2$, making it accessible to
117: ground-based instrumentation. Furthermore, a spectroscopic approach
118: to detecting high-$z$ sources has the clear advantage of containing
119: redshift information, unlike continuum observations, thus avoiding
120: problems of source confusion that stem from the low angular resolution
121: of single-dish radio telescopes equipped with bolometer arrays. As
122: early as 1997, however, Infrared Space Observatory (ISO) observations
123: suggested potential problems with this method. Indeed, Malhotra et
124: al. (1997) and Luhman et al. (1998) found that the proportionality
125: between the \cii\ and FIR continuum luminosities observed in nearby
126: star forming galaxies broke down for luminous infrared galaxies, where
127: the \cii\ luminosity appeared not to exceed $\Lcii\sim10^9$ \Lsun.
128: These conclusions, confirmed and expanded in a recent analysis of ISO
129: data by Luhman et al. (2003), have since cast doubts on the usefulness
130: of redshifted \cii\ to probe star formation in the distant
131: universe. The observations presented here have bearing on this matter:
132: does the paucity of \cii\ emission observed in local ultra-luminous
133: infrared galaxies (ULIRGs) hold for objects in the early universe?
134:
135: \section{Observations and Results}
136:
137: We observed the quasar SDSS J1148+5251 using the James Clerk Maxwell
138: Telescope (JCMT) at Mauna Kea, Hawaii\footnote{The JCMT is operated
139: by the Joint Astronomy Centre in Hilo, Hawaii on behalf of the parent
140: organizations Particle Physics and Astronomy Research Council in the
141: United Kingdom, the National Research Council of Canada and The
142: Netherlands Organization for Scientific Research.}. The observations
143: were performed in the flexible queue mode over several shifts between
144: 2003 November 15 and 2004 January 30, using 8~Hz beam-position switching
145: over 60 second cycles with offsets of 60\arcsec\ in azimuth. Standard
146: calibration observations, which include pointing and focus, were
147: performed at the beginning and sometimes near the end of each shift.
148: The RxA3 dual sideband heterodyne receiver of the JCMT was tuned to
149: place 256.172 GHz at the center of the lower sideband, which assumes a
150: rest frame wavelength for the \cii\ transition of $\lambda=157.74$
151: $\mu$m and a redshift of $z=6.419$ as reported for the CO detections
152: of this object (Walter et al. 2003; Bertoldi et al. 2003a). We
153: utilized the 1840 MHz bandpass of the Digital Autocorrelation
154: Spectrometer which allowed us to cover a velocity range approximately
155: $\pm1100$ \kmpers\ around the expected emission, equivalent to a
156: redshift range $z\approx6.39-6.44$ (by comparison, the detected CO
157: lines are $\sim300$ \kmpers\ wide). Approximately 48 hours of
158: integration were devoted to this project.
159:
160: Individual spectra obtained from 10 minute-long integrations were read
161: into SPECX, merged, and converted to FITS format. The data were then
162: read into COMB where spectra showing abnormal RMS (due to baseline
163: problems) were expunged, and good quality individual integrations were
164: then combined using the built-in weighting scheme (i.e.,
165: $t_{int}/T^2_{sys}$), then Hanning-smoothed to 10~MHz wide
166: channels. The final spectrum is shown in Figure 1, the result of 37.5
167: hours of integration after the removal of a linear baseline. Its RMS
168: is 1.3~mK in the central 1000 \kmpers\ region, or 32 mJy assuming an
169: aperture efficiency $\approx$ 0.56 for the JCMT at the observing
170: frequency. The integrated intensity of the \cii\ line is
171: $\Icii\approx -1.7\pm1.9$ Jy~\kmpers\ $\approx
172: (-1.4\pm1.6)\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$. The calibration
173: uncertainty of these observations could be as large as $30\%$ due to
174: the poorly characterized sideband ratio of RxA3 at 256 GHz combined
175: with the complex frequency structure of the receiver noise
176: temperature.
177:
178: To convert this measurement to a luminosity in a manner consistent
179: with other recent work on this source, we adopt the WMAP cosmology
180: with H$_0=71$ \kmpers, $\Omega_{\rm m}=0.27$, and
181: $\Omega_\Lambda=0.73$ (Spergel et al. 2003). The luminosity distance
182: to a source at a redshift $z=6.419$ is then
183: $D_l\approx1.97\times10^{29}$~cm. Consequently, the \cii\ luminosity
184: of SDSS J1148+5251 is $\Lcii<2\times10^{9}$ L$_\odot$. Using the FIR
185: luminosity $\Lfir\approx1.2\times10^{13}$ L$_\odot$ determined by
186: Bertoldi et al. (2003b) from 1~mm continuum observations carried out
187: with the MAMBO array, the \cii/FIR ratio is
188: $\Lcii/\Lfir<1.7\times10^{-4}$. Taking into account that changing the
189: assumed 40~K dust temperature by $\pm10$~K will change the derived
190: \Lfir\ by a factor of 2 (Bertoldi et al. 2003b) and the aforementioned
191: uncertainties in the \cii\ calibration, a more conservative limit is
192: $\Lcii/\Lfir<5\times10^{-4}$.
193:
194: \section{Discussion}
195:
196: The limit obtained on the \cii\ emission in J1148+5251 confirms some
197: of the suggested caveats on the usefulness of redshifted \cii\
198: emission as a beacon for star formation in the early universe. The
199: \cii\ line in this very luminous FIR object is disproportionally
200: faint. Indeed, our limit suggests a ratio in this quasar that appears
201: typical of the local sample of 15 ULIRGs studied by Luhman et al.
202: (2003). Their analysis found a median \cii/FIR flux ratio of
203: $\Fcii/\Ffir=3.9\times10^{-4}$ in ULIRGs whereas a median
204: $\Fcii/\Ffir=3.1\times10^{-3}$ was reported for a sample of 60 normal
205: star-forming galaxies by Malhotra et al. (2001). Only one in five
206: ULIRGs approached the lower end of the distribution of \cii/FIR ratios
207: found in normal galaxies, with $\Fcii/\Ffir\sim10^{-3}$. The \cii\
208: luminosities of the ULIRG systems studied by Luhman et al. had a mean
209: value $\Lcii\sim 1\times10^{9}$ L$_\odot$ with none exceeding $3.5
210: \times 10^9$ L$_\odot$. If the brightest \cii\ luminosities that can
211: be expected from objects at $z\sim6.5$ were similar to the
212: luminosities of local ULIRGs, observations that probe such regime need
213: to be $\sim3-10$ times more sensitive than what we have acheived --- a
214: feat probably beyond the current reach of single-dish instruments but
215: which may be attainable by the IRAM Plateau de Bure Interferometer or
216: certainly by the future Atacama Large Millimeter Array.
217:
218: What are the reasons behind the deficit of \cii\ emission relative to
219: the FIR continuum? Malhotra et al. (1997; 2001) find a strong
220: correlation between decreasing \Lcii/\Lfir\ ratio and increasing
221: \Lfir\ and FIR color temperature, indicating a smooth transition
222: between the ratio observed in normal luminosity galaxies and that in
223: ULIRGs, which in turn suggests that the same mechanism is responsible
224: for low ratios in both types of objects. Indeed, about 7\% of the
225: sources in the Malhotra et al. (2001) sample have measured ratios or
226: limits that are smaller than the $\Lcii/\Lfir\sim5\times10^{-4}$ more
227: typical of ULIRGs, although their luminosities are
228: $\Lfir\sim10^{10}-10^{11}$ \Lsun. The physical conditions of the gas
229: in SDSS J1148+5251 may hold clues as to what mechanism is responsible
230: for the low observed ratios. For example, we can place this object in
231: the \cii/FIR vs. CO/FIR plots obtained from photodissociation region
232: (PDR) models (e.g., see Kaufman et al. 1999) using our limit and the
233: CO observations and estimate some of its physical parameters. The
234: observational limit for the luminosity of the CO \jone\ transition in
235: SDSS J1148+5251 reported by Bertoldi et al. (2003a) is
236: $\Lco<7\times10^6$ \Lsun, which yields $\Lco/\Lfir<6\times10^{-7}$.
237: Their multitransition LVG analysis, however, suggests that the
238: excitation conditions of the CO are very similar to those found by
239: Bradford et al. (2003) in the starburst galaxy NGC~253. The expected
240: CO \jone\ luminosity would then be close to the the observational
241: limit, i.e., $\Lco/\Lfir\sim6\times10^{-7}$. Using Fig. 16 of Kaufman
242: et al. we can infer a UV radiation field strength $G_0\gtrsim600$ in
243: units of the local interstellar radiation field (Habing 1968), and
244: densities $n\gtrsim10^5$ \percmcu. In this determination the CO/FIR
245: ratio constrains the radiation field, while the \cii/FIR ratio
246: constrains the gas density. A PDR with an incident radiation field
247: $G_0\gtrsim600$ would have a surface temperature $T\sim200$~K, rapidly
248: decreasing as the gas becomes increasingly molecular. These
249: conditions are actually not peculiar, and are very similar to those
250: found in a variety of starburst galaxies and Galactic star-forming
251: regions (Stacey et al. 1991).
252:
253: Whether or not a simple analysis like that described above can be
254: applied with impunity to a complex system such as an AGN is debatable.
255: Based on their available data, Malhotra et al. (1997; 2001) and Luhman et
256: al. (2003) concluded that the observed deficit of \cii\ emission
257: relative to the continuum emission in FIR-luminous objects is
258: consistent with: 1) low \cii\ emission arising from PDRs with high
259: $G_0/n$, and/or 2) excess FIR emission not related to PDRs, but
260: perhaps arising from a enshrouded AGN component. In the first case,
261: the conversion of UV photons into electrons that heat the gas is
262: inefficient because the dust grains rapidly become positively charged
263: and their photoelectric yield is dramatically reduced. Thus, the gas
264: temperature (and with it the \cii\ brightness) does not increase
265: proportionally with the radiation field but the dust temperature still
266: does, leading to low \Lcii/\Lfir\ ratios. In the second case, the
267: \cii\ emission related to the AGN core is quenched for much the same
268: reasons, as well as because most of the carbon will likely be in a
269: higher ionization state. Copious FIR radiation will be generated,
270: however, if the AGN is at least partially embedded in a dusty
271: envelope. Although this latter scenario naturally explains the deficit
272: of \cii\ emission in a quasar, we caution that there is a dearth of
273: additional evidence for it. For example, one may expect that an
274: excess of FIR emission associated with dust near the ionized core of
275: the AGN will result in a lower than usual $\Lco/\Lfir$ ratio. The
276: $\Lco/\Lfir\sim6\times10^{-7}$ ratio inferred for SDSS J1148+5251,
277: however, is unremarkable when compared to a sample of local starburst
278: galaxies and Galactic star-forming regions (Stacey et al. 1991).
279: Indeed, this quasar falls into a region of the \Icii/\Ifir\ vs.
280: \Ico/\Ifir\ space mostly occupied by ULIRGs (see Fig. 8a of Genzel \&
281: Cesarsky 2000), showing that the observed ratios are not peculiarly
282: associated with AGN activity. Nevertheless, without a detection of
283: the CO \jone\ transition or a stronger limit, it is difficult to
284: develop this analysis any further.
285:
286: \section{Summary and Conclusion}
287:
288: Our redshifted \cii\ observations at $\nu\sim256$~GHz have failed to
289: detect the fine structure 158 $\mu$m \cii\ transition in this high-$z$
290: quasar. We place a 1 $\sigma$ limit on the \cii\ luminosity of SDSS
291: J1148+5251 of $\Lcii<2.6\times10^9$ \Lsun\ (including 30\% calibration
292: uncertainty). Given the observed FIR flux of this source, this places
293: an upper limit on its \cii\ to FIR ratio of
294: $\Lcii/\Lfir<5\times10^{-4}$, substantially lower than what is found
295: in most local star-forming galaxies but similar to what is observed in
296: nearby ultra-luminous IR galaxies. If the cause of this \cii\ deficit
297: in SDSS J1148+5251 was a large FIR component due to partial dust
298: reprocessing of the AGN radiation, thus unrelated to star formation
299: activity, the $\sim$3000 M$_\odot$ yr$^{-1}$ star formation rate
300: obtained from the FIR continuum could be a substantial overestimate.
301: Obtaining a better SED for the FIR continuum emission, to constrain
302: better the FIR luminosity, as well as improving the sensitivity of the
303: CO \jone\ observations, may help to elucidate this matter.
304:
305: Will redshifted \cii\ observations open a new window onto high-$z$
306: galaxies? Our results are based on observations of just one object at
307: high redshift. A larger sample of $z>6$ sources with high FIR
308: luminosities must be observed in the \cii\ transition to establish
309: whether low $\Lcii/\Lfir$ ratios are a general phenomenon. If a
310: maximum luminosity $\Lcii\sim10^9$ \Lsun\ were confirmed for most
311: sources, the Atacama Large Millimeter Array would still be able to
312: detect the \cii\ line and spatially resolve its distribution and
313: kinematics. The answer, then, is a hopeful {\em yes}.
314:
315: \acknowledgements
316: We thank Gerald Moriarty-Schieven, Ming Zhu, the JCMT Telescope
317: Support Specialists, and the numerous visiting JCMT observers who
318: obtained data for this project on our behalf. We would also like
319: to thank Fabian Walter, Gordon Stacey, and the anonymous referee.
320:
321: \newpage
322: \begin{references}
323: \reference{}
324: Bertoldi, F., et al. 2003a, A\&A, 409, L47
325:
326: \reference{}
327: Bertoldi, F., Carilli, C. L., Cox, P., Fan, X., Strauss, M. A., Beelen,
328: A., Omont, A., Zylka, R. 2003b, A\&A, 406, L55
329:
330: \reference{}
331: Fan, X., et al. 2003, AJ, 125, 1649
332:
333: \reference{}
334: Genzel, R., \& Cesarsky, C. J. 2000, ARA\&A, 38, 761.
335:
336: \reference{}
337: Habing, H. J. 1968, \bain, 19, 421
338:
339: \reference{}
340: Kaufman, M. J., Wolfire, M. G., Hollenbach, D. J., \& Luhman, M. L. 1999,
341: ApJ, 527, 795
342:
343: \reference{}
344: Luhman, M. L., et al. 1998, ApJ, 504, L11
345:
346: \reference{}
347: Luhman, M. L., Satyapal, S., Fischer, J., Wolfire, M. G., Sturm, E., Dudley,
348: C. C., Lutz, D., \& Genzel, R. 2003, ApJ, 594, 758
349:
350: \reference{}
351: Malhotra, S., et al. 1997, ApJ, 491, L27
352:
353: \reference{}
354: Malhotra, S., et al. 2001, ApJ, 561, 766
355:
356: \reference{}
357: Spergel, D. N., et al. 2003, ApJS, 148, 175, 194
358:
359: \reference{}
360: Stacey, G. J., Geis, N., Genzel, R., Lugten, J. B., Poglitsch, A., Sternberg,
361: A., \& Townes, C. H. 1991, ApJ, 373, 423
362:
363: \reference{}
364: Stark, A. A. 1997, ApJ, 481, 587
365:
366: \reference{}
367: Tielens, A. G. G. M., \& Hollenbach, D. J. 1985, ApJ, 291, 722
368:
369: \reference{}
370: Walter, F., et al. 2003, Nature, 424, 406
371:
372: \reference{}
373: Willott, C. J., McLure, R. J., \& Jarvis, M. J. 2003, ApJ, 587, L15
374:
375: \reference{}
376: White, R. J., Becker, R. H., Fan, X., \& Strauss, M. A. 2003, AJ, 126, 1
377:
378: \end{references}
379:
380: \clearpage
381: \begin{figure}
382: \plotone{f1.eps}
383: \caption{The redshifted \cii\ spectrum obtained at the JCMT, with a
384: total of 37.5 hours of integration. The RMS in 10 MHz wide channels
385: ($\sim11$ \kmpers) in the central 1000 \kmpers, illustrated by
386: the gray region, is 1.3 mK in the T$_A^*$ scale. This corresponds to
387: $\approx32$ mJy, using an aperture efficiency $\eta_{ap}\sim0.56$.}
388: \end{figure}
389:
390: \end{document}
391: