Gas properties in the Early Universe deciphered from spectral surveys of high-z objects: the Cloverleaf quasar

Observations of molecular lines in very-high redshift objects offer a unique chance to investigate the gas dynamics and physico-chemical conditions in early galaxies, hence to understand galaxy evolution. densities radiatively gas Thirty rotational transitions


Introduction and Observations
The Cloverleaf Quasar (H1413+117) is the archetype of gravitationally lensed quasars and one of the most luminous and best studied objects in the distant, hence young Universe. The quasar redshift (z=2.5579, as derived from a Gaussian fit to the CO J=3-2 line of figure 2) implies 11 billion years of light travel time to the Sun and an age of about 2 Gyr after the appearance of the very first stars. Located at the center of a starburst galaxy, its radiation, viewed from the Earth, is magnified by a factor as large as 11 (according to lensing models [8]) by 2 galaxies, with redshifts z≃1 and 2, respectively, that intercept the line of sight and split the quasar image into 4 distinct components (figure 1). The quasar and its host galaxy have been studied in the X-ray, optical and radio domains, and more particularly at sub-millimetre wavelengths. The detection of low-J lines of CO, HCN and HCO + was reported in the early 2000s [1,2,5,7], whereas a low spectral resolution survey of the 680-1100 GHz band (rest frequencies), carried out at the CSO ( [3]), allowed to detect the J=6-5 through J=9-8 lines of CO. We report here the results of an on-going spectral line survey of this quasar in the mm-sub/mm domain (from 100 to 800 GHz, rest frequencies) aimed at studying the physical conditions and chemical composition of the gas in the quasar surroundings.
The survey is carried on with NOEMA, ALMA and the IRAM 30-m and Yebes 40-m telescopes. It benefits from a high spectral resolution (few MHz) and a high sensitivity (10 times higher than the pioneering work of Barvainis et al. [2]). So far, it covers the 110-180 GHz (rest frequencies) band, observed with the 40-m telescope, the 340-370 GHz band, observed with the 30-m telescope, NOEMA and ALMA, as well as the 525-625 GHz band and most of the 700-785 GHz band, observed with NOEMA. Figures 2 and 3 show some of the ALMA and NOEMA spectra. A degree 1 baseline has been subtracted from each spectrum to remove the continuum. The r.m.s. noise per 80 kms −1 channel is 0.3-0.5 mJy. The NOEMA on-source integration time was 8 hours per pair of 8 GHz-wide spectra.

Results
Fourteen molecular species, including the rare CO isotopologues 13   HCN and HCO + was reported in the early 2000s [1,2,5,7], whereas a low spectral resolution survey of the 680-1100 GHz band (rest frequencies), carried out at the CSO ( [3]), allowed to detect the J=6-5 through J=9-8 lines of CO. We report here the results of an on-going spectral line survey of this quasar in the mm-sub/mm domain (from 100 to 800 GHz, rest frequencies) aimed at studying the physical conditions and chemical composition of the gas in the quasar surroundings.
The survey is carried on with NOEMA, ALMA and the IRAM 30-m and Yebes 40-m telescopes. It benefits from a high spectral resolution (few MHz) and a high sensitivity (10 times higher than the pioneering work of Barvainis et al. [2]). So far, it covers the 110-180 GHz (rest frequencies) band, observed with the 40-m telescope, the 340-370 GHz band, observed with the 30-m telescope, NOEMA and ALMA, as well as the 525-625 GHz band and most of the 700-785 GHz band, observed with NOEMA. Figures 2 and 3 show some of the ALMA and NOEMA spectra. A degree 1 baseline has been subtracted from each spectrum to remove the continuum. The r.m.s. noise per 80 kms −1 channel is 0.3-0.5 mJy. The NOEMA on-source integration time was 8 hours per pair of 8 GHz-wide spectra.

Results
Fourteen molecular species, including the rare CO isotopologues 13 CO and C 18 O and the rare HCO + isotopologue H 13 CO + , are detected. Three species, CH, CCH, and H 2 O + , are The low noise level and the flat baselines provided by the interferometers allow an accurate determination of the shapes and strengths of the strongest lines (figure 4). The CO  line, which has the highest S/N, exhibits an almost perfect gaussian shape with a full width at half power of 421 ± 50 kms −1 . This is the same width as the CO(1-0) line and, within the errors, as the much weaker, but optically thin, 13 CO(5-4) line (440 ± 100 kms −1 ) -note that the CO(7-6) line reported by [1] was observed with an insufficient bandwidth and that the higher-J CO lines were observed with a too coarse spectral resolution [3].
The lines of the high gas density tracers HCN, HCO + and H 2 O seem to be marginally wider (FWHP ≃ 440 ± 100 kms −1 ), but still compatible with the CO linewidths. The line intensities are far better determined than in previous measurements. The CO(3-2) line luminosity (hence the observed source brightness temperature) is found to be definitely stronger (by 15%) than that of the (1-0) line. This and the near equality of the 13 CO(5-4) and (7-6) line brightness temperatures imply a warmer and/or denser gas than previously assumed (30-50 K, 10 3−5 cm −3 according to [6,9]) and refute the hypothesis that CO lines arise from an homogeneous gaseous disk. As for the dust grains, the SED of the mm/sub-mm continuum [9] is best fitted by a 3-component model with a cold component at 30 K, a lukewarm component at 50 K and a hot component at 115 K, the latter representing only ≃ 1% of the total dust mass. The detections of several transitions of the high density tracers HCO + , HCN, HNC and CN, as well as that of the rare isotopologues 13 CO and H 13 CO + (whose lines have opacities <1) allow us to disentangle the relative contributions of the gas kinetic temperature, H 2 volume density, infrared pumping and line opacity to the line strengths, hence to better constrain the molecular gas properties. Figure 5 shows the HCO + , HCN and HNC velocity-integrated line flux density divided by J 2 upper (i.e. a quantity that scales with the line luminosity L'), as a function of J upper . Two main results stem from this figure. The first is that fluorescent excitation, via infrared pumping, does not play a major role througout the molecular disk, contrary to what is argued for some superluminous infrared galaxies (e.g. APM08279+5255). The HNC molecule, which has the lowest vibrationally excited state (ν2(1 −0) = 21.7µm above the ground state, a wavelength close to the peak of the dust continuum SED), should be the most affected by such a pumping. In contrast, the lowest bending states of HCN (ν2(1 − 0) = 14.0µm), and HCO + (ν2(1 − 0) = 12.1µm), are shifted to wavelengths at which the continuum flux is reduced by factors 6-8. Yet, HCN is the molecules whose luminosity decreases the slowest. In the same vein, despite the lack of bending states for CN, the HCN/CN luminosity ratio remains constant from 350 GHz to 600 GHz, when IR pumping should have strongly increased it.
The second result is that the HCN line luminosity remains almost constant between the J=6-5 and the J=8-7 lines, despite the fact that the latter levels lie twice higher in energy than the former ones. Similarly, the 13 CO(5-4) line luminosity is not significatively lower than that of 13 CO . This contrasts with the factor of 2 decrease in luminosity between HCN(4-3) and HCN , and the factor of 1.7 decrease between CO(3-2) and CO . A probable explanation is that the gas kinetic temperature and the dust temperature increase from the edges of the extended starburst region (whose radius is 700 pc according to [8]) to the quasar immediate surroundings. This is illustrated by the fits to the HCO + lines in Fig. 5 with an LVG model, which require at least two different temperatures. That the linewidths do not significantly change from line to line suggests that the starburst region is a clumpy gaseous disk rotating around the central quasar with a nearly flat rotation curve.