1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: \newcommand{\myemail}{riechers@mpia.de}
5: %\usepackage{amssymb}
6:
7: %% -------Author macros----------------------
8: %% DO NOT USE IN MATH MODE!
9: \def\ts {\thinspace}
10: \def\kms {\ts km\ts s$^{-1}$}
11: \def\etal {{\rm et\ts al.}}
12: \def\msol {M$_{\odot}$}
13: \def\lsol {L$_{\odot}$}
14: \def\ci {C\ts {\scriptsize I}}
15: \def\hi {H\ts {\scriptsize I}}
16: \def\aco {{\rm CO}($J$=1$\to$0)}
17: \def\bco {{\rm CO}($J$=2$\to$1)}
18: \def\cco {{\rm CO}($J$=3$\to$2)}
19: \def\dco {{\rm CO}($J$=4$\to$3)}
20: \def\eco {{\rm CO}($J$=5$\to$4)}
21: \def\fco {{\rm CO}($J$=6$\to$5)}
22: \def\gco {{\rm CO}($J$=7$\to$6)}
23: \def\hco {{\rm CO}($J$=8$\to$7)}
24: \def\ico {{\rm CO}($J$=9$\to$8)}
25: \def\jco {{\rm CO}($J$=10$\to$9)}
26: \def\kco {{\rm CO}($J$=11$\to$10)}
27: \def\cho {{\rm H$_2$O}(3$_{13}$$\to$2$_{20}$)}
28: \def\ahcn {{\rm HCN}($J$=1$\to$0)}
29: \def\bhcn {{\rm HCN}($J$=2$\to$1)}
30: \def\chcn {{\rm HCN}($J$=3$\to$2)}
31: \def\dhcn {{\rm HCN}($J$=4$\to$3)}
32: \def\ehcn {{\rm HCN}($J$=5$\to$4)}
33: \def\ehnc {{\rm HNC}($J$=5$\to$4)}
34: \def\ahco {{\rm HCO$^+$}($J$=1$\to$0)}
35: \def\dhco {{\rm HCO$^+$}($J$=4$\to$3)}
36: \def\acn {{\rm CN}($N$=1$\to$0)}
37: \def\ccn {{\rm CN}($N$=3$\to$2)}
38: \def\dcn {{\rm CN}($N$=4$\to$3)}
39: %%-------------------------------------------
40:
41: \slugcomment{draft version \today, accepted for publication in the
42: Astrophysical Journal}
43:
44: \shorttitle{CN($N$=3--2) in the Cloverleaf}
45: \shortauthors{Riechers et al.}
46:
47: \begin{document}
48:
49: \title{Detection of Emission from the CN Radical in the Cloverleaf Quasar at $z$=2.56}
50:
51: \author{Dominik A. Riechers\altaffilmark{1}, Fabian
52: Walter\altaffilmark{1}, Pierre Cox\altaffilmark{2}, Christopher
53: L. Carilli\altaffilmark{3}, \\ Axel Weiss\altaffilmark{4},
54: Frank Bertoldi\altaffilmark{5}, and Roberto Neri\altaffilmark{2}
55: }
56:
57:
58: \altaffiltext{1}{Max-Planck-Institut f\"ur Astronomie, K\"onigstuhl 17,
59: Heidelberg, D-69117, Germany}
60:
61: \altaffiltext{2}{Institut de RadioAstronomie Millim\'etrique, 300 Rue
62: de la Piscine, Domaine Universitaire, 38406 Saint Martin d'H\'eres,
63: France}
64:
65: \altaffiltext{3}{National Radio Astronomy Observatory, PO Box O,
66: Socorro, NM 87801, USA}
67:
68: \altaffiltext{4}{Max-Planck-Institut f\"ur Radioastronomie, Auf dem
69: H\"ugel 69, Bonn, D-53121, Germany}
70:
71: \altaffiltext{5}{Argelander-Institut f\"ur Astronomie, Universit\"at
72: Bonn, Auf dem H\"ugel 71, Bonn, D-53121, Germany}
73:
74: \email{riechers@mpia.de}
75:
76:
77: \begin{abstract}
78: We report the detection of \ccn\ emission towards the Cloverleaf
79: quasar ($z=2.56$) based on observations with the IRAM Plateau de
80: Bure Interferometer. This is the first clear detection of emission
81: from this radical at high redshift. CN emission is a tracer of
82: dense molecular hydrogen gas ($n({\rm H_2}) > 10^4\,$cm$^{-3}$)
83: within star--forming molecular clouds, in particular in regions
84: where the clouds are affected by UV radiation. The HCN/CN intensity
85: ratio can be used as a diagnostic for the relative importance of
86: photodissociation regions (PDRs) in a source, and as a sensitive
87: probe of optical depth, the radiation field, and photochemical
88: processes. We derive a lensing--corrected \ccn\ line luminosity of
89: $L'_{\rm CN(3-2)} = (4.5 \pm 0.5) \times 10^{9}\,$K \kms pc$^2$.
90: The ratio between CN luminosity and far--infrared luminosity falls
91: within the scatter of the same relationship found for low--$z$
92: (ultra--) luminous infrared galaxies. Combining our new results
93: with \cco\ and \ahcn\ measurements from the literature and assuming
94: thermal excitation for all transitions, we find a CO/CN luminosity
95: ratio of 9.3 $\pm$ 1.9 and a HCN/CN luminosity ratio of 0.95 $\pm$
96: 0.15. However, we find that the \ccn\ line is likely only
97: subthermally excited, implying that those ratios may only provide
98: upper limits for the intrinsic 1$\to$0 line luminosity ratios. We
99: conclude that, in combination with other molecular gas tracers like
100: CO, HCN, and HCO$^+$, CN is an important probe of the physical
101: conditions and chemical composition of dense molecular environments
102: at high redshift.
103: \end{abstract}
104:
105: \keywords{galaxies: active, starburst, formation, high redshift ---
106: cosmology: observations --- radio lines: galaxies}
107:
108:
109: \section{Introduction}
110:
111: Investigations of the dense molecular interstellar medium (ISM) in
112: high--redshift galaxies are of fundamental importance to further our
113: understanding of the early phases of galaxy formation and evolution,
114: as it harbors the environments in which the actual star formation is
115: believed to occur. Recent studies of the molecular gas phase in
116: high--$z$ galaxies using CO emission lines have revealed large
117: molecular gas reservoirs with masses in excess of 10$^{10}$\,\msol\
118: (see review by Solomon \& Vanden Bout \citeyear{sv05}, and references
119: therein).
120:
121: However, while the relative brightness of CO emission lines renders
122: this molecule the most common tracer of the requisite material for
123: star formation, the low densities of only $n_{\rm H_2} \sim
124: 10^2-10^3\,$cm$^{-3}$ required to excite its lower--$J$ transitions
125: (due to its low dipole moment of only $\mu_{\rm D}^{\rm CO} = 0.11$)
126: imply that CO is not a specific tracer of dense molecular cloud cores,
127: i.e., the regions where stars are actively formed.
128:
129: In contrast, recent studies of nearby actively star--forming galaxies
130: have shown that the high dipole molecule HCN ($\mu_{\rm D}^{\rm HCN} =
131: 2.98$) is a far better tracer of such dense molecular cores (e.g.\ Gao
132: \& Solomon \citeyear{gao04a}, \citeyear{gao04b}). The critical density
133: of $n_{\rm H_2} \sim 10^5-10^6$\,cm$^{-3}$ to collisionally thermalize
134: its lower--$J$ transitions is much higher than that of CO, and of the
135: same order as the densities found in bright star--forming regions in
136: the Galaxy (e.g.~the Orion Bar, Hogerheijde et al.\ \citeyear{hog95}).
137: A main result in this context is the finding that the HCN luminosity
138: correlates well with the far--infrared (FIR) luminosities (which is
139: commonly used to estimate star--formation rates at high redshift) over
140: 7--8 orders of magnitude, from Galactic dense cores to the highest
141: redshift quasars (Wu et al.\ \citeyear{wu05}).
142:
143: However, in nearby luminous and ultra--luminous infrared galaxies
144: (LIRGs/ULIRGs), it has been found that systems with similar HCN/CO and
145: HCN/FIR luminosity ratios may have quite different dense gas
146: properties regarding their chemical composition, and their gas
147: excitation (e.g., Aalto et al.\ \citeyear{aal02}). To better
148: understand the physical and chemical state of the dense molecular gas
149: phase which directly relates to star formation, it has proven
150: essential to study additional bright tracers of dense gas with
151: properties different from HCN, such as the cyanide radical (CN). Due
152: to its lower dipole moment relative to HCN ($\mu_{\rm D}^{\rm CN} =
153: 1.45$), its critical density is lower by about a factor of 5.
154: Observations of CN emission toward the Orion A molecular cloud complex
155: have shown that CN filaments trace the dense interfaces between the
156: molecular cloud and the major ionization fronts (Rodr\'iguez--Franco
157: et al.\ \citeyear{rod98}). It has also been found that the [CN]/[HCN]
158: abundance ratio is greatly enhanced in the central region of the
159: starburst galaxy M82, being as high as $\sim$5 across the entire
160: nucleus (Fuente et al.\ \citeyear{fue05}). These observations
161: indicate that CN is a good tracer of gas layers which are affected by
162: photochemistry, since this molecule appears to be predominantly found
163: in regions exposed to ionizing stellar UV radiation. It has also been
164: found that, due to the rapid destruction of other dense gas tracers
165: like HCN, the abundance of CN tends to be enhanced in areas where the
166: UV radiation field is only partly attenuated, such as in zones close
167: to the surface of photodissociation regions (PDRs, Fuente et al.\
168: \citeyear{fue93}; Sternberg \& Dalgarno \citeyear{sd95}; Jansen et
169: al.\ \citeyear{jan95}). The CN/HCN intensity ratio can thus be used
170: as a diagnostic for the relative importance of PDRs in a source, and a
171: sensitive probe of optical depth, the radiation field, and
172: photochemical processes (e.g., Boger \& Sternberg \citeyear{bs05}).
173: CN emission may thus in a sense be a more specific tracer for star
174: formation than, e.g., HCN, which only traces regions of dense gas in
175: general. In addition, theoretical studies of the chemical composition
176: of molecular gas for different ionization parameters suggest that the
177: relative abundance of CN may also be enhanced in X--ray dominated
178: regions (XDRs), such as AGN environments (Lepp \& Dalgarno
179: \citeyear{lep96}). This exemplifies why it is desirable to search for
180: CN emission, which is part of our current effort to study molecular
181: tracers other than CO out to high redshifts. Such an investigation is
182: imperative to obtain more meaningful constraints on the physical
183: properties and chemical composition of the dense molecular ISM in
184: distant galaxies.
185:
186: In this paper, we report the first high--$z$ detection of \ccn\
187: emission, which was observed towards the Cloverleaf quasar ($z=2.56$)
188: with the IRAM Plateau de Bure Interferometer (PdBI)\footnote{IRAM is
189: supported by INSU/CNRS (France), MPG (Germany), and IGN (Spain).}.
190: Due to its strong gravitational magnification (magnification factor
191: $\mu_{\rm L} = 11$, Venturini \& Solomon \citeyear{ven03}), the
192: Cloverleaf is the brightest CO source at high redshift (e.g.,
193: Barvainis et al.\ \citeyear{bar94}), and one of the most prolific
194: sources of molecular lines beyond $z$=2. It was the first $z>2$ source
195: to be detected in HCN (Solomon \etal\ \citeyear{sol03}) and HCO$^+$
196: (Riechers et al.\ \citeyear{rie06a}) emission. It was also detected in
197: both \ci\ fine structure lines (Barvainis et al.\ \citeyear{bar97};
198: Wei\ss\ \etal\ \citeyear{wei03}, \citeyear{wei05}). We use a standard
199: concordance cosmology throughout, with $H_0 = 71\,$\kms\,Mpc$^{-1}$,
200: $\Omega_{\rm M} =0.27$, and $\Omega_{\Lambda} = 0.73$ (Spergel \etal\
201: \citeyear{spe03}, \citeyear{spe06}).
202:
203: \section{Observations}
204:
205: We observed the \ccn\ transition line ($\nu_{\rm rest} =
206: 339.4467770-340.2791661\,$GHz for the different fine structure [fs]
207: and hyperfine structure [hfs] transitions of the $v$=0 vibrational
208: state) towards H1413+117 (the Cloverleaf quasar) using the PdBI in D
209: configuration between 2006 July 26 and September 03. At the target
210: redshift of $z$=2.55784 (Wei\ss\ \etal\ \citeyear{wei03}), the line is
211: shifted to $\sim$95.6\,GHz (3.14\,mm). The total integration time
212: amounts to 17.5\,hr using 5 antennas, resulting in 5.8\,hr equivalent
213: on--source time with 6 antennas after discarding unusable visibility
214: data. The nearby sources 1354+195 and 1502+106 (distance to the
215: Cloverleaf: $9.0^\circ$ and $12.0^\circ$) were observed every 20
216: minutes for pointing, secondary amplitude and phase calibrations. For
217: primary flux calibration, several nearby calibrators (MWC\,349,
218: CRL\,618, 3C\,273, 3C\,345, and NRAO\,150) were observed during all
219: runs.
220:
221: The correlator was tuned to a frequency of 95.603\,GHz, which
222: corresponds to the central position between the brightest hfs
223: transitions\footnote{These components are actually blends of the
224: [$N$=3$\to$2, $J$=$\frac{5}{2}$$\to$$\frac{3}{2}$,
225: ($F$=$\frac{7}{2}$$\to$$\frac{5}{2}$,
226: $\frac{5}{2}$$\to$$\frac{3}{2}$, and
227: $\frac{3}{2}$$\to$$\frac{1}{2}$)] hfs components at $\nu_{\rm rest}
228: = 340.0315440-340.0354080\,$GHz and [$N$=3$\to$2,
229: $J$=$\frac{7}{2}$$\to$$\frac{5}{2}$,
230: ($F$=$\frac{9}{2}$$\to$$\frac{7}{2}$,
231: $\frac{7}{2}$$\to$$\frac{5}{2}$, and
232: $\frac{5}{2}$$\to$$\frac{3}{2}$)] hfs components at $\nu_{\rm rest}
233: = 340.2477700-340.2485764\,$GHz (components 2 and 3 in
234: Fig.~\ref{f1}).} of the \ccn\ line at 95.6332 and 95.5736\,GHz. The
235: total bandwidth of 580\,MHz ($\sim$1800\,\kms ) used for the
236: observations is large enough to cover all fs and hfs transitions of
237: the \ccn\ line. It also covers enough channels that are free of line
238: emission to constrain the 3\,mm continuum emission of the Cloverleaf.
239:
240: For data reduction and analysis, the IRAM GILDAS package was used.
241: All data were mapped using the CLEAN algorithm and 'natural' weighting
242: without applying a further taper; this results in a synthesized beam
243: of 5.3\,$''$$\times$4.9\,$''$ ($\sim$42\,kpc at $z$ = 2.56). The
244: final rms in the combined map is 0.25\,mJy beam$^{-1}$ for a 190\,MHz
245: (corresponding to 596\,\kms) channel, 0.4\,mJy beam$^{-1}$ for a
246: 80\,MHz (252\,\kms) channel, and 0.8\,mJy beam$^{-1}$ for a 20\,MHz
247: (63\,\kms) channel.
248:
249: \section{Theoretical Considerations}
250:
251: \begin{figure}[t]
252: \epsscale{1.2}
253: \plotone{f1.eps}
254: \caption{Model spectrum of the \ccn\ emission in the Cloverleaf. The
255: horizontal axes indicate the rest and observed frequencies. The full
256: range of the PdBI bandpass is shown. The vertical axis indicates the
257: predicted intensity, normalized to the ($N$=3$\to$2,
258: $J$=$\frac{7}{2}$$\to$$\frac{5}{2}$,
259: $F$=$\frac{9}{2}$$\to$$\frac{7}{2}$) component. The relative
260: intensities are computed for the LTE case. The crosses indicate the
261: hyperfine structure components. The plus signs indicate the summed
262: intensities of the components in three different frequency bins.
263: The dashed lines are Gaussian fits to the summed intensities,
264: assuming the width of the \cco\ line (Wei\ss\ et al.\
265: \citeyear{wei03}) for the subcomponents (labeled '1'--'3'). The
266: solid line is a sum of all three components, and indicates the
267: model--predicted \ccn\ line shape. The thick dotted line is a
268: single Gaussian (used to fit the observations), fitted to the model
269: line. The dotted vertical line indicates the tuning frequency,
270: corresponding to zero velocity in Fig.~\ref{f4}. \label{f1}}
271: \end{figure}
272:
273: Due to its fine stucture and hyperfine structure splitting, the \ccn\
274: emission line is distributed over 19 lines\footnote{We assume that CN
275: is in its ground electronic state (${}^2\Sigma$), and that the spins
276: couple according to Hund's case (b) coupling scheme: $\vec{N} +
277: \vec{S} = \vec{J}$, (fs coupling) and $\vec{J} + \vec{I_{\rm C}} +
278: \vec{I_{\rm N}} = \vec{F}$ (hfs coupling). Here, $\vec{N}$ is the
279: rotational angular momentum vector, $\vec{S}$ is the electronic
280: spin, and $\vec{I}$ is a nuclear spin.}, in 3 main components
281: separated by more than 200\,MHz from each other in the rest frame.
282: This separation is of the same order as the kinematical broadening of
283: the CO lines in the Cloverleaf (Wei\ss\ et al. \citeyear{wei03}),
284: causing the hfs components to be blended. To analyze the intrinsic
285: line shape of the \ccn\ transition in the Cloverleaf, we calculated a
286: synthetic line profile, assuming optically thin emission in Local
287: Thermodynamic Equilibrium (LTE) to derive the relative intensities of
288: the hfs components. The relative intensities of the components were
289: computed using the laboratory data from Skatrud et al.\
290: (\citeyear{ska83}), and approximation (4) of equation (1) of Pickett
291: et al.\ (\citeyear{pic98}). Assuming that the 3 main \ccn\ components
292: are kinematically broadened in the same way as the CO lines (416 $\pm$
293: 6\,\kms\ FWHM; see Wei\ss\ et al. \citeyear{wei03}), we obtain the
294: synthetic line profile displayed in Fig.~\ref{f1} (solid line). Under
295: the given assumptions, the contribution from all hfs lines of
296: component 1 to the total intensity are negligible, and not detectable
297: at the given signal--to--noise (see below). A single Gaussian fit
298: (dotted line, 484\,\kms\ FWHM, or 116\% of the CO lines) agrees with
299: the more detailed model profile (solid line) within a few percent in
300: its peak position and integral, i.e., well within the observational
301: errors. Note however that the width and peak position of this Gaussian
302: depends on the relative intensities of the hfs components, which in
303: turn depend on the above assumptions, in particular the optical depth.
304: In the optically thin LTE case, components 2 and 3 have a peak
305: strength ratio of 1:1.4. As an example, if they had a ratio of unity,
306: the fitted Gaussian to the line profile would have a FWHM of
307: 603\,\kms, or 145\% of the CO lines. We thus conclude that fitting a
308: Gaussian to the observed line profile contains all relevant
309: information, while minimizing the number of free fit parameters, and
310: thus is preferred over a more complex fitting procedure to describe
311: and analyze the \ccn\ profile at the given signal--to--noise ratio.
312:
313:
314: \section{Results}
315:
316: \subsection{CN Maps and Spectrum}
317:
318: \begin{figure}
319: \epsscale{1.15}
320: \plotone{f2.eps}
321: \caption{Velocity--integrated PdBI map of \ccn\ emission towards the
322: Cloverleaf quasar.
323: At a resolution of 5.3\,$''$$\times$4.9\,$''$
324: (as indicated in the bottom left corner), the source is
325: unresolved. The cross indicates the geometrical center of the CO
326: emission in the Cloverleaf (Alloin et al.\ \citeyear{all97}, see
327: text). Contours are shown at (-3, -2, 2, 3, 4, 5, 6,
328: 7)$\times\sigma$ (1$\sigma$ = 0.25\,mJy beam$^{-1}$).
329: \label{f2}}
330: \end{figure}
331:
332: \begin{figure*}
333: \epsscale{1.15}
334: \plotone{f3.eps}
335: \caption{Channel maps of the \ccn\ emission (same region is shown as
336: in Fig.~\ref{f2}). One channel width is 80\,MHz, or
337: 252\,km\,s$^{-1}$ (at 95.4355, 95.5155, 95.5955, 95.6755, 95.7555,
338: and 95.8355\,GHz; frequencies increase with channel number).
339: Contours are shown at (-3, -2, 2, 3, 4, 5)$\times \sigma$ (1$\sigma$
340: = 0.4\,mJy\,beam$^{-1}$). The beam size (5.3\,$''$$\times$4.9\,$''$)
341: is shown in the bottom left corner; the cross indicates the same
342: position as in Fig.~\ref{f2}.
343: \label{f3}}
344: \end{figure*}
345:
346: We have detected emission from the \ccn\ transition line towards the
347: Cloverleaf quasar ($z=2.56$). The velocity--integrated \ccn\ line map
348: is shown in Fig.~\ref{f2}. The cross indicates the geometrical
349: center position of the resolved \gco\ map at $\alpha=14^{\rm h}15^{\rm
350: m}46^{\rm s}.233$, $\delta=+11^\circ29'43''.50$ (Alloin \etal\
351: \citeyear{all97}). The line emission is clearly detected at 7$\sigma$
352: over a range of 596\,\kms\ (190\,MHz); the source appears unresolved.
353: In Fig.~\ref{f3}, six channel maps (252\,\kms, or 80\,MHz each) of the
354: \ccn\ line emission are shown. At an rms of 0.4\,mJy\,beam$^{-1}$,
355: the line is detected at 4 and 5$\sigma$ in the central channels, and
356: the decline of the line intensity towards the line wings is clearly
357: visible toward the outer channels, as expected.
358:
359: Fig.~\ref{f4} shows the spectrum of the \ccn\ emission at a
360: resolution of 63\,\kms\ (20\,MHz). Zero velocity corresponds to the
361: tuning frequency of 95.603\,GHz. The solid line shows a Gaussian fit
362: to the spectrum. From the fit, we derive a line peak flux density of
363: 1.94 $\pm$ 0.24\,mJy beam$^{-1}$, and a line FWHM of 666 $\pm$
364: 97\,\kms. The fit provides an upper limit for the continuum emission
365: at the line frequency of $<$0.25\,mJy. From the channels assumed to
366: be free of line emission, we derive a formal 3$\sigma$ upper limit to
367: the continuum peak flux density of 0.7\,mJy beam$^{-1}$. Note that
368: the model fit to the dust SED of the Cloverleaf by Wei\ss\ et al.\
369: (\citeyear{wei03}) would suggest a continuum flux of $\sim$0.3\,mJy at
370: the \ccn\ line frequency, which agrees within the errors with the
371: above estimates. As the continuum however is not detected, we do not
372: subtract a continuum component from the observed spectrum. This leads
373: to an integrated \ccn\ line flux of 1.37 $\pm$ 0.17\,Jy \kms. The
374: velocity offset of the Gaussian peak relative to the tuning frequency
375: is -105 $\pm$ 38\,\kms.
376:
377: The FWHM of the Gaussian fit to the \ccn\ emission line suggests that
378: its components 2 and 3 have similar peak strengths (different from
379: 1:1.4 as predicted for the optically thin LTE case; Fig.~\ref{f1}),
380: which may indicate that the emission is optically thick. Within the
381: limited signal--to--noise of the spectrum, the shape of the emission
382: line remains compatible with a two--component structure. It may thus
383: be possible that the two brightest hfs complexes (components 2 and 3)
384: are detected individually, but observations at higher
385: signal--to--noise are needed to confirm this result. In any case, a
386: double Gaussian fit to the line profile gives the same integrated line
387: flux as the single Gaussian fit within the errors.
388:
389: \begin{figure}[b]
390: \epsscale{1.15}
391: \plotone{f4.eps}
392: \caption{ Spectrum of the \ccn\ emission at a resolution of 63\,\kms\
393: (20\,MHz). The velocity scale is relative to the tuning frequency
394: of 95.603\,GHz. The rms per velocity bin is 0.8\,mJy. The solid
395: line shows a Gaussian fit to the data. \label{f4}}
396: \end{figure}
397:
398: \subsection{Line Luminosities}
399:
400:
401: From our observations, we derive a \ccn\ line luminosity of $L'_{\rm
402: CN(3-2)} = (4.5 \pm 0.5) \times 10^9\,$K\,\kms\,pc$^2$ (corrected
403: for gravitational magnification, $\mu_{\rm L} = 11$; see Table 1 for
404: details). This corresponds to 11\% of $L'_{\rm CO(3-2)}$ (Wei\ss\ et
405: al.\ \citeyear{wei03}), 105\% of $L'_{\rm HCN(1-0)}$ (Solomon et al.\
406: \citeyear{sol03}), and 129\% of $L'_{\rm HCO^+(1-0)}$ (Riechers et
407: al.\ \citeyear{rie06a}). Assuming that CN is thermally excited up to
408: the $N$=3$\to$2 transition (i.e., $L'_{\rm CN(3-2)} = L'_{\rm
409: CN(1-0)}$) and that optical depth effects can be neglected, this
410: would mean that CN emission in the Cloverleaf is slightly brighter
411: than that from the other high density probes HCN and HCO$^+$. This is
412: remarkable, as HCN and HCO$^+$ have higher critical densities than CN
413: in the ground--state transition, but the critical density of \ccn\ is
414: higher than that of \ahcn\ and \ahco. This may indicate a higher
415: relative filling factor of CN relative to HCN and HCO$^+$, or even a
416: relatively high chemical abundance of CN. However, with the
417: observations existing at present, it is not possible to disentangle
418: excitation effects from chemical effects.
419:
420: In addition, observations of \dhcn\ towards the Cloverleaf quasar have
421: shown that $L'_{\rm HCN(4-3)}$/$L'_{\rm HCN(1-0)} \leq 0.34$ (Solomon
422: et al.\ \citeyear{sol03}, caption of their Tab.~1, and M.~Gu\'elin
423: 2007, priv.~comm.). This ratio is significantly lower than 1, which
424: implies that the 4$\to$3 transition of HCN is sub--thermally excited.
425: As the \ccn\ transition has a somewhat lower but comparable critical
426: density relative to \dhcn, it is likely that the \ccn\ transition is
427: also sub--thermally excited, and the `intrinsic',
428: thermalization--corrected $L'_{\rm CN(1-0)}$ thus even higher. This
429: would be consistent the finding that CN is clearly subthermally
430: excited in nearby LIRGs and ULIRGs (Aalto et al.\ \citeyear{aal02}).
431: This would also imply that CN is a {\em brighter} tracer of dense gas
432: than HCN and HCO$^+$ in this high redshift quasar.
433:
434: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
435: %%%% Tab.1: Line luminosities
436: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
437: \begin{deluxetable}{ r c c c }
438: \tabletypesize{\scriptsize}
439: \tablecaption{Molecular line luminosities in the Cloverleaf. \label{tab-1}}
440: \tablehead{
441: & $S_{\nu}$ & $L'$ & Ref. \\
442: & [mJy] & [10$^9$\,K\,\kms\,pc$^2$] & }
443: \startdata
444: \ccn\ & 1.94 $\pm$ 0.24 & 4.5 $\pm$ 0.5 & 1 \\
445: \ahcn\ & 0.24 $\pm$ 0.04 & 4.3 $\pm$ 0.5 & 2 \\
446: \ahco\ & 0.19 $\pm$ 0.03 & 3.5 $\pm$ 0.3 & 3 \\
447: \cco\ & 30 $\pm$ 1.7 & 42 $\pm$ 7 & 4 \\
448: \vspace{-2mm}
449: \enddata
450: \tablerefs{${}$ [1] This work, [2] Solomon \etal\ (\citeyear{sol03}),
451: [3] Riechers et al.\ (\citeyear{rie06a}), [4] Wei\ss\ \etal\
452: (\citeyear{wei03}).}
453: \tablecomments{${}$Luminosities are corrected for gravitational
454: magnification. }
455: \end{deluxetable}
456:
457: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
458:
459: \section{Discussion}
460:
461: \begin{figure*}
462: \epsscale{1.15}
463: \plotone{f5.eps}
464: \caption{CN luminosity relations for a sample of local
465: spiral/starburst galaxies, low--$z$ IR--luminous galaxies, and the
466: Cloverleaf. The Cloverleaf luminosities are corrected for
467: gravitational lensing ($\mu_{\rm L} = 11$). The two symbols with star
468: insets denote tentative CN detections. The solid line is a least
469: squares fit to all data except the Cloverleaf. The error bars
470: indicate the statistical errors of the line luminosity measurements.
471: The excitation of the \ccn\ transition in the Cloverleaf may be
472: sub--thermal, which would affect the (extrapolated) \acn\ line
473: luminosity as indicated by the horizontal, shaded regions. As an
474: example, the vertical bars on the shaded regions show a case where
475: $L'_{\rm CN(3-2)} = 0.5 \times L'_{\rm CN(1-0)}$ in the Cloverleaf.
476: See text for more details. \label{f5}}
477: \end{figure*}
478:
479: In the following, we discuss various relationships between the
480: emission observed in CN and other molecules (CO and HCN) and the
481: far--IR continuum for a sample of local spiral/starburst galaxies
482: (Henkel et al.\ \citeyear{hen88}, \citeyear{hen98}, 2007, priv.~comm.;
483: Wang et al.\ \citeyear{wan04}; additional CO/HCN/FIR data from Eckart
484: et al.\ \citeyear{eck90}; Nguyen-Q-Rieu et al.\ \citeyear{ngu92};
485: Mauersberger et al.\ \citeyear{mau96}; Mao et al.\ \citeyear{mao2000};
486: Sanders et al.\ \citeyear{san03}), low--$z$ (U)LIRGs (Aalto et al.\
487: \citeyear{aal02}), and the Cloverleaf (this work; Solomon et al.\
488: \citeyear{sol03}; Wei\ss\ et al.\ \citeyear{wei03}) as shown in
489: Fig.~\ref{f5}. Note that, due to the weak signal--to--noise ratio of
490: 1$\sigma$, we do not include the \dcn\ observations of APM\,08279+5255
491: (Guelin et al.\ \citeyear{gue06}, their Tab.~1) in this discussion.
492:
493: As \aco\ is not yet detected toward the Cloverleaf, we assume in the
494: following that CO is thermalized up to the 3$\to$2 transition, i.e.,
495: $L'_{\rm CO(1-0)} = L'_{\rm CO(3-2)}$ (see Riechers et al.\
496: \citeyear{rie06b} for justification). Both effects of thermal and
497: sub--thermal excitation of the \ccn\ emission line in the Cloverleaf
498: are discussed. CO, CN, and HCN line intensities for all low--$z$
499: galaxies were measured directly in the ground--state transitions, all
500: quoted luminosities thus are 1$\to$0 luminosities. We do not discuss
501: effects of differential lensing\footnote{Note that Chartas et al.\
502: (\citeyear{cha07}) recently reported a likely microlensing event
503: toward the Cloverleaf; however, this event took place close to 2000
504: April, so 3--7\,years before all of the molecular gas observations
505: of this source discussed in this paper (obtained between 2003 March
506: and 2006 September).}, as models indicate similar sizes for
507: molecular and dust emission in the Cloverleaf (Solomon et al.\
508: \citeyear{sol03}).
509:
510: {\em Figure 5a}: Wu et al.\ (\citeyear{wu05}) find a linear
511: correlation between $L'_{\rm HCN}$ and $L_{\rm FIR}$ over 7--8 orders
512: of magnitude in luminosity, which is believed to indicate that the
513: dense gas tracer HCN is also a good estimator for star formation rates
514: out to high $z$. If CN were to be a good tracer of dense, actively
515: star--forming gas, a similar trend may be expected between $L'_{\rm
516: CN}$ and $L_{\rm FIR}$. Figure~\ref{f5}a shows that $L'_{\rm CN}$
517: correlates closely with $L_{\rm FIR}$; a linear least squares fit to
518: all galaxies (excluding the Cloverleaf) yields log$(L_{\rm FIR}) =
519: (0.89 \pm 0.09) \times {\rm log} (L'_{\rm CN})+(4.2 \pm 0.7)$. This
520: slope may suggest a trend of decreasing CN/FIR luminosity ratio
521: towards higher luminosities (see also Aalto et al.\ \citeyear{aal02}).
522: It however also is consistent with unity within the statistical
523: uncertainties (which are quite large due to the limited sample size)
524: and systematical errors (e.g., different beam sizes for CN line and IR
525: continuum observations), and thus with CN being a valuable tracer of
526: star formation. The Cloverleaf agrees remarkably well with the
527: relation defined by the low--$z$ galaxies (even if the \ccn\
528: transition is sub--thermally excited, as indicated by the horizontal
529: bar on the shaded region), extending the observed trend to higher
530: luminosities and out to high redshift.
531:
532: {\em Figure 5b}: Gao \& Solomon (\citeyear{gao04b}) observe a rise in
533: dense gas fraction with $L_{\rm FIR}$. The $L'_{\rm CO}$/$L'_{\rm
534: CN}$ ratio may be considered (the inverse of) a tracer of the dense
535: gas fraction. A significant spread in the $L'_{\rm CO}$/$L'_{\rm CN}$
536: ratio is found for the sample shown in Fig.~\ref{f5}b, ranging from a
537: few to almost a factor of 30. There is no obvious trend of $L'_{\rm
538: CO}$/$L'_{\rm CN}$ with $L_{\rm FIR}$; however, the sample size is
539: too small and heterogenous to come to a definite conclusion. Note
540: that, due to the likely sub--thermal excitation of the \ccn\
541: transition (as indicated by the shaded region), the $L'_{\rm
542: CO}$/$L'_{\rm CN}$ ratio may be considered an upper limit of the
543: `intrinsic', thermalization--corrected ratio of the 1$\to$0
544: transitions for the Cloverleaf.
545:
546: {\em Figure 5c}: Assuming that CN traces the UV field (which falls off
547: rapidly with growing distance from star--forming regions, or an active
548: galactic nucleus), while HCN traces dense gas in general, it would be
549: expected that the HCN/CN ratio stays constant or decreases with
550: increasing $L_{\rm FIR}$. In this scenario, a decrease in the HCN/CN
551: ratio would correspond to an increase in the filling factor of
552: UV--illuminated clouds with $L_{\rm FIR}$. This would however also
553: cause a rise in $L'_{\rm CN}$/$L_{\rm FIR}$, which is not observed
554: (Fig.~\ref{f5}a). If differences in the relative chemical abundances
555: of CN and HCN and optical depth/thermalization effects do not play a
556: major role, one may thus expect $L'_{\rm HCN}$/$L'_{\rm CN}$ to stay
557: fairly constant with $L_{\rm FIR}$. Based on the observations of
558: their subsample, Aalto et al.\ (\citeyear{aal02}) suggest that the
559: HCN/CN intensity ratio may increase slightly with $L_{\rm FIR}$. Such
560: a trend however is not seen in the larger sample displayed in
561: Fig.~\ref{f5}c (note that $L'_{\rm HCN}$/$L'_{\rm CN}$ for the
562: Cloverleaf may again be considered an upper limit due to possible
563: sub--thermal excitation of \ccn, as indicated by the shaded region).
564: Clearly, improved statistics and better models are required to address
565: this issue in more detail.
566:
567: Chemical models suggest that CN is produced quite efficiently in the
568: presence of a strong UV field, which also leads to an enhanced
569: ionization rate (Boger \& Sternberg \citeyear{bs05}). The large
570: observed [CN]/[HCN] abundance ratio of $\sim$5 toward the molecular
571: disk of M82 may be indicative of a large dense PDR bathed in the
572: intense radiation field of the starburst environment (Fuente et al.\
573: \citeyear{fue05}), lending observational support to these models.
574: Other models suggest that the relative abundance of CN may also be
575: enhanced in regions with elevated X--ray ionization rates, such as AGN
576: environments (Lepp \& Dalgarno \citeyear{lep96}; Meijerink \& Spaans
577: \citeyear{mei05}). Especially in the Cloverleaf, where the central
578: quasar is known to contribute a significant fraction to the heating of
579: the gas and dust (Wei\ss\ et al.\ \citeyear{wei03}; Solomon et al.\
580: \citeyear{sol03}), such a scenario would appear reasonable and may
581: explain the relatively high brightness of CN compared to other dense
582: gas tracers in the Cloverleaf. However, note that in Fig.~5c, the pure
583: starburst galaxies show the brightest CN emission relative to HCN, and
584: galaxies with a relatively strong AGN like Mrk\,231 and NGC\,6240 have
585: some of the highest $L'_{\rm HCN}$/$L'_{\rm CN}$ ratios. This may
586: indicate that the strength of the UV field in a starburst environment
587: has a significantly greater impact on the global gas--phase production
588: rate of CN in a galaxy than the strength of the X--ray field, lending
589: further support to the supposition that CN is a good star formation
590: tracer. We thus conclude that the XDR scenario alone does not explain
591: the observed dense gas properties of the (U)LIRGs shown in Fig.~5
592: without any further assumptions.
593:
594: Overall, the Cloverleaf follows the $L'_{\rm CN}$--$L_{\rm FIR}$
595: relation as defined by nearby galaxies of different types remarkably
596: well, and extends this relation to higher luminosities. This relation
597: now appears to hold over almost 3 orders of magnitude. If these
598: findings were to hold for other distant quasars, CN would be an
599: excellent probe of dense molecular environments out to high redshifts.
600: Also, the higher--order transitions of CN appear to be brighter than
601: those of other dense gas tracers like HCN. This may prove to be of
602: particular importance for future high--$z$ studies with the Atacama
603: Large Millimeter/submillimeter Array (ALMA), which will offer the
604: opportunity to probe to fainter galaxy populations in general, but
605: will also be restricted to the higher--order transitions of the most
606: common dense molecular gas tracers (typically 3$\to$2 and higher at
607: $z>2$).
608:
609: \acknowledgments
610: This research is based on observations carried out with the IRAM
611: Plateau de Bure Interferometer. IRAM is supported by INSU/CNRS
612: (France), MPG (Germany), and IGN (Spain). D.~R.\ acknowledges support
613: from the Deutsche Forschungsgemeinschaft (DFG) Priority Programme
614: 1177. C.~C.\ acknowledges support from the Max-Planck-Gesellschaft
615: and the Alexander von Humboldt-Stiftung through the
616: Max-Planck-Forschungspreis 2005. The authors would like to thank
617: Susanne Aalto for fruitful discussions on the subject matter. We also
618: would like to thank the staff at IRAM, in particular Jan Martin
619: Winters, for their assistance in setting up and carrying out the
620: observations, as well as for providing a preliminary reduction of the
621: data. The authors would like to thank the referee for useful comments
622: that helped to improve the manuscript.
623:
624:
625: \begin{thebibliography}{}
626:
627: \bibitem[Aalto et al.(2002)]{aal02} Aalto, S., Polatidis, A.~G.,
628: H\"uttemeister, S., \& Curran, S.~J.\ 2002, A\&A, 381, 783
629: \bibitem[Alloin et al.(1997)]{all97} Alloin, D., Guilloteau, S.,
630: Barvainis, R., Antonucci, R., \& Tacconi, L., 1997, A\&A, 321, 24
631: \bibitem[Barvainis et al.(1994)]{bar94} Barvainis, R., Tacconi, L.,
632: Antonucci, R., Alloin, D., \& Coleman, P.\ 1994, Nature, 371, 586
633: \bibitem[Barvainis et al.(1997)]{bar97} Barvainis, R., Maloney, P.,
634: Antonucci, R., \& Alloin, D.\ 1997, ApJ, 484, 695
635: \bibitem[Boger \& Sternberg(2005)]{bs05} Boger, G.~I., \& Sternberg,
636: A.\ 2005, ApJ, 632, 302
637: \bibitem[Chartas et al.(2007)]{cha07} Chartas, G., Eracleous, M., Dai,
638: X., Agol, E., \& Gallagher, S.\ 2007, ApJ, in press
639: (astro-ph/0702742)
640: \bibitem[Eckart et al.(1990)]{eck90} Eckart, A., Downes, D., Genzel,
641: R., Harris, A.~I., Jaffe, D.~T., \& Wild, W.\ 1990, ApJ, 348, 434
642: \bibitem[Fuente et al.(1993)]{fue93} Fuente, A., Martin-Pintado, J.,
643: Cernicharo, J., \& Bachiller, R.\ 1993, A\&A, 276, 473
644: \bibitem[Fuente et al.(2005)]{fue05} Fuente, A., Garc\'ia-Burillo, S.,
645: Gerin, M., Teyssier, D., Usero, A., Rizzo, J. R., \& de Vicente, P.\
646: 2005, ApJ, 619, L155
647: \bibitem[Gao \& Solomon(2004a)]{gao04a} Gao, Y., \& Solomon, P.~M.\
648: 2004a, ApJS, 152, 63
649: \bibitem[Gao \& Solomon(2004b)]{gao04b} Gao, Y., \& Solomon, P.~M.\
650: 2004b, ApJ, 606, 271
651: \bibitem[Gu\'elin et al.(2007)]{gue06} Gu\'elin, M., Salom\'e, P.,
652: Neri, R., Garc\'ia-Burillo, S., Graci\'a-Carpio, J., et al.\ 2007,
653: A\&A, 462, L45
654: \bibitem[Henkel et al.(1988)]{hen88} Henkel, C., Schilke, P., \&
655: Mauersberger, R.\ 1988, A\&A, 201, L23
656: \bibitem[Henkel et al.(1998)]{hen98} Henkel, C., Chin, Y.-N.,
657: Mauersberger, R., \& Whiteoak, J.~B.\ 1998, A\&A, 329, 443
658: \bibitem[Hogerheijde et al.(1995)]{hog95} Hogerheijde, M.~R., Jansen,
659: D.~J., \& van Dishoeck, E.~F.\ 1995, A\&A, 294, 792
660: \bibitem[Jansen et al.(1995)]{jan95} Jansen, D.~J., van Dishoeck,
661: E.~F., Black, J.~H., Spaans, M., \& Sosin, C.\ 1995, A\&A, 302, 223
662: \bibitem[Lepp \& Dalgarno(1996)]{lep96} Lepp, S., \& Dalgarno, A.\
663: 1996, A\&A, 306, L21
664: \bibitem[Mao et al.(2000)]{mao2000} Mao, R.~Q., Henkel, C., Schulz,
665: A., Zielinsky, M., Mauersberger, R., St\"orzer, H., Wilson, T.~L.,
666: Gensheimer, P.\ 2000, A\&A, 358, 433
667: \bibitem[Mauersberger et al.(1996)]{mau96} Mauersberger, R., Henkel,
668: C., Wielebinski, R., Wiklind, T., \& Reuter, H.-P.\ 1996, A\&A, 305,
669: 421
670: \bibitem[Meijerink \& Spaans(2005)]{mei05} Meijerink, R., \& Spaans,
671: M.\ 2005, A\&A, 436, 397
672: \bibitem[Nguyen-Q-Rieu et al.(1992)]{ngu92} Nguyen-Q-Rieu, Jackson,
673: J.~M., Henkel, C., Truong-Bach, \& Mauersberger, R.\ 1992, ApJ, 399,
674: 521
675: \bibitem[Pickett et al.(1998)]{pic98} Pickett, H.~M., Poynter, R.~L.,
676: Cohen, E.~A., Delitsky, M.~L., Pearson, J.~C., \& M\"uller,
677: H.~S.~P.\ 1998, J.~Quant.~Spectrosc.~Radiat.~Transfer, 60, 883
678: \bibitem[Riechers et al.(2006a)]{rie06a} Riechers, D.~A., Walter, F.,
679: Carilli, C.~L., Wei\ss\, A., Bertoldi, F., Menten, K.~M., Knudsen,
680: K.~K., \& Cox, P.\ 2006a, ApJ, 645, L13
681: \bibitem[Riechers et al.(2006b)]{rie06b} Riechers, D.~A., Walter, F.,
682: Carilli, C.~L., et al.\ 2006b, ApJ, 650, 604
683: \bibitem[Rodr\'iguez--Franco et al.(1998)]{rod98} Rodriguez-Franco,
684: A., Martin-Pintado, J., \& Fuente, A.\ 1998, A\&A, 329, 1097
685: \bibitem[Sanders et al.(2003)]{san03} Sanders, D.~B., Mazzarella,
686: J.~M., Kim, D.-C., Surace, J.~A., \& Soifer, B.~T.\ 2003, AJ, 126,
687: 1607
688: \bibitem[Skatrud et al.(1983)]{ska83} Skatrud, D.~D., de Lucia, F.~C.,
689: Blake, G.~A., \& Sastry, K.~V.~L.~N.\ 1983, J.~Mol.~Spectrosc., 99,
690: 35
691: \bibitem[Solomon et al.(2003)]{sol03} Solomon, P., Vanden Bout, P.,
692: Carilli, C., \& Guelin, M.\ 2003, Nature 426, 636
693: \bibitem[Solomon \& Vanden Bout(2005)]{sv05} Solomon, P.~M., \& Vanden
694: Bout, P.~A.\ 2005, ARA\&A, 43, 677
695: \bibitem[Spergel et al.(2003)]{spe03} Spergel, D.~N., Verde, L.,
696: Peiris, H.~V., \etal\ 2003, ApJS, 148, 175
697: \bibitem[Spergel et al.(2007)]{spe06} Spergel, D.~N., Bean, R.,
698: Dor\'e, O., \etal\ 2007, ApJ, in press (astro-ph/0603449)
699: \bibitem[Sternberg \& Dalgarno(1995)]{sd95} Sternberg, A., \&
700: Dalgarno, A.\ 1995, ApJS, 99, 565
701: \bibitem[Venturini \& Solomon(2003)]{ven03} Venturini, S., \& Solomon,
702: P.~M.\ 2003, ApJ, 590, 740
703: \bibitem[Wang et al.(2004)]{wan04} Wang, M., Henkel, C., Chin, Y.-N.,
704: \etal\ 2004, A\&A, 422, 883
705: \bibitem[Wei\ss\ et al.(2003)]{wei03} Wei\ss, A., Henkel, C., Downes,
706: D., \& Walter, F.\ 2003, A\&A, 409, L41
707: \bibitem[Wei\ss\ et al.(2005)]{wei05} Wei\ss, A., Downes, D., Henkel,
708: C., \& Walter, F.\ 2005, A\&A, 429, L25
709: \bibitem[Wu et al.(2005)]{wu05} Wu, J., Evans II, N.~J., Gao, Y.,
710: Solomon, P.~M., Shirley, Y.~L., \& Vanden Bout, P.~A.\ 2005, ApJ,
711: 635, L173
712: \end{thebibliography}
713:
714: \end{document}
715: