A&A 394, 763-768 (2002)
DOI: 10.1051/0004-6361:20021231

## A search for molecules in damped Lyman-alpha absorbers occulting millimetre-loud quasars

S. J. Curran1 - M. T. Murphy1 - J. K. Webb1 - F. Rantakyrö2,3 - L. E. B. Johansson4 - S. Nikolic4,5

1 - School of Physics, University of New South Wales, Sydney NSW 2052, Australia
2 - European Southern Observatory, Casilla 19001, Santiago 19, Chile
3 - Observatorio Cerro Calan, Universidad de Chile, Santiago, Chile
4 - Onsala Space Observatory, Chalmers University of Technology, 439 92 Onsala, Sweden
5 - Astronomical Observatory, Volgina 7, 11160 Belgrade, Serbia

Received 21 June 2002 / Accepted 27 August 2002

Abstract
We have used the SEST 15-metre and Onsala 20-metre telescopes to perform deep (  mJy) integrations of various molecular rotational transitions towards damped Lyman-alpha absorption systems (DLAs) known to occult millimetre-loud quasars. We have observed 6 new systems and improved the existing limits for 11 transitions. These limits may be approaching the sensitivities required to detect new systems and we present a small number of candidate systems which we believe warrant further observation.

Key words: quasars: absorption lines - cosmology: observations - cosmology: early Universe

### 1 Introduction

Millimetre-band molecular absorption systems along the line-of-sight toward quasars can provide a powerful probe of cold gas in the early Universe. Wiklind & Combes (1994, 1995, 1996b) have used such absorption lines to study a variety of properties of the absorbers themselves (e.g. relative column densities, kinetic and excitation temperatures, filling factors). Constraints on the cosmic microwave background temperature can also be obtained by comparing the optical depths of different rotational transitions (e.g. Wiklind & Combes 1996a). If the background quasar is gravitationally lensed, time delay studies can yield constraints on the Hubble constant (e.g. Wiklind & Combes 2001). However, these studies have so far been limited by the paucity of mm-band molecular absorbers: only 4 such systems are currently known-towards TXS 0218+357 (Wiklind & Combes 1995), PKS 1413+135 (Wiklind & Combes 1997), TXS 1504+377 (Wiklind & Combes 1996c) and PKS 1830-211 (Wiklind & Combes 1998).

Recent attention has focused on using molecular absorption lines as a probe for possible changes in the fundamental constants on cosmological time-scales. Detailed studies of the relative positions of heavy element optical transitions compared with laboratory spectra favour a smaller fine structure constant ( ) at redshifts 0.5 < z < 3.5 at the 4.1 significance level (Murphy et al. 2001b; Webb et al. 2001). The observed fractional change in [ ] is very small, so systematic errors have to be carefully considered. However, a thorough search for systematics has not revealed a simpler explanation of the optical results (Murphy et al. 2001c) and so independent constraints at similar redshifts are required.

Comparison of molecular rotational (i.e. mm-band) and corresponding H  I 21-cm absorption line frequencies has the potential to constrain changes in with a fractional precision 10-6 per absorption system. The ratio of the hyperfine (H  I) transition frequency to that of a molecular rotational line is and so any variation in this will be observed as a difference in the apparent redshifts of these lines (Drinkwater et al. 1998).

 Quasar Coordinates (J2000) Radio flux densities (Jy) Transition h  m  s d   '   '' S1.4 S5.0 S22 S90 S230 S350 B 0235+1624 02 38 38.9 16 36 59 2.36 1.64 2.48 1-6 1-4 - 0.523869 CS B 0248+430 02 51 34.5 43 15 16 1.43 0.66 0.7 0.32 0.08 - 0.3939 CS B 0738+313 07 41 10.7 31 12 00 2.05 - 1.3 0.27 0.10 - 0.2212 CO B 0827+243 08 30 52.1 24 11 00 0.84 0.89 1.3 2.4 1.3 - 0.5247 CS B 08279+5255 08 31 41.6 52 45 18 0.001 - - - - 0.08 2.97364 *HCO+ ... ... ... ... ... ... ... ... ... ... CO B 1017+1055 10 20 08.8 10 40 03 0.22 - - - 0.004 - 2.380 *CS ... ... ... ... ... ... ... ... ... ... CO B 1328+307 13 31 08.3 30 30 33 14.7 1.41 2.7 0.75 0.33 - 0.69218 HCO+ ... ... ... ... ... ... ... ... ... ... CS B 0458-020 05 01 12.8 -01 59 14 2.2 3 - 0.8 0.5 - 2.0399 HCO+ ... ... ... ... ... ... ... ... ... ... CO B 0834-201 08 36 39.2 -20 16 59 1.97 1.5 - 0.85 0.28 - 1.715 HCO+ ... ... ... ... ... ... ... ... ... ... CO B 1229-0207 12 32 00.0 -02 24 05 1.90 0.90 1 0.4 0.2 - 0.3950 CO ... ... ... ... ... ... ... ... ... ... CO B 1451-375 14 54 27.4 -37 47 33 1.57 1.84 - 1.5 0.5 - 0.2761 CO ... ... ... ... ... ... ... ... ... ... HCO+

From spectra of two of the known high redshift absorbers, PKS 1413+135 and TXS 0218+357, Carilli et al. (2000) and Murphy et al. (2001a) have obtained constraints on consistent with zero y-variation at redshifts and 0.24671, respectively. However, the major uncertainty in the mm/H  I comparison is that intrinsic velocity differences between the mm and H  I absorption lines can be introduced if the lines-of-sight to the millimetre wave and radio continuum emission regions of the quasar differ, as is certainly the case for PKS 1413+135 and TXS 0218+357 (Carilli et al. 2000). Thus, a statistical sample of mm/H  I comparisons is required to independently check the optical results.

One systematic approach to finding new high redshift molecular absorbers is to target high column density systems with known redshifts. A convenient sample is therefore the damped Lyman-alpha absorbers defined to have neutral hydrogen column densities  cm-2. We therefore compiled a catalogue of all known DLAs and shortlisted those which are illuminated by radio-loud quasars (Curran et al. 2002). Of these, seven in the northern sky have measured millimetre fluxes, and in the south there are around a dozen DLAs illuminated by millimetre-loud quasars, of which we observed the four loudest (Table 1).

We note that Wiklind & Combes (1994b, 1995, 1996b) searched with null results for redshifted molecular emission and absorption towards 12 and 46 quasars, respectively. However, only 11 are occulted by DLAs and not all of these are radio-loud (Table 2). This motivates a more systematic search for millimetre absorption systems associated with DLAs. In this paper we present the results of our first search: the DLAs which occult known millimetre-loud quasars with the SEST and Onsala 20-m telescopes.

### 2 Observations

The northern 3 mm and 7 mm observations were performed in February and April 2002, respectively, with the 20-m telescope at Onsala Space Observatory, Sweden. For the 3 mm observations the SIS receiver was tuned to single-sideband mode and the backend was a hybrid correlator with a bandwidth of 1280 MHz and a channel separation of 0.8 MHz. We used dual-beam switching with a throw of in azimuth, and pointing errors were typically rms on each axis. The intensity was calibrated using the chopper-wheel method and typical system temperatures, on the -scale, were around 300 K. For the 7 mm observations we used the 43 GHz SIS receiver and again the backend was the hybrid correlator with a bandwidth of 640 MHz and a channel separation of 0.4 MHz. Since there is no beam switch capability at these frequencies, we removed the ripple caused by standing waves by subtracting the Fourier components in a transform of the spectra. Typical system temperatures, on the -scale, were around 200 K.

 Quasar Transition Radio flux densities (Jy) [mK] Reference S1.4 S5.0 S22 S90 S230 B 0235+1624 0.5238 CO 2.36 1.64 2.48 1.2* 1.3* 2.3 Wiklind & Combes (1995) HCO+ ... ... ... ... ... 3.6 Wiklind & Combes (1995) B 0458-020 2.0397 CO 2.2 3 - 0.8 0.5 3.0 Wiklind & Combes (1994b) HCO+ ... ... ... ... ... 7.1 Wiklind & Combes (1996b) B 0528-2505 2.1408 CO 1.50 1.13 - - - 1.3 Wiklind & Combes (1994b) B 0834-201 1.715 HCO+ - 1.5 - 0.85 0.28 3.6 Wiklind & Combes (1996b) HCO+ ... ... ... ... ... 3.9 Wiklind & Combes (1996b) B 1215+333 1.9984 CO 0.18 0.08 - - - 4.4 Wiklind & Combes (1994b) B 1229-0207 0.39498 CO 1.90 0.90 1 0.4 0.2 12 Wiklind & Combes (1995) B 1328+307 0.69215 CO 14.7 1.41 2.7 0.6* 0.24* 6.1 Wiklind & Combes (1995) CO ... ... ... ... ... 6.5 Wiklind & Combes (1995) HCO+ ... ... ... ... ... 5.8 Wiklind & Combes (1995) B 2136+142 2.1346 CO - 1.11 1.6 0.59* 0.25 2.8 Wiklind & Combes (1996b) CO ... ... ... ... ... 2.3 Wiklind & Combes (1996b) CO ... ... ... ... ... 3.9 Wiklind & Combes (1996b) HCO+ ... ... ... ... ... 3.4 Wiklind & Combes (1995)

The southern sources in the sample were observed in April 2002 with the 15-m SEST at La Silla, Chile, using the 100 GHz and 150 GHz SESIS receivers. These were tuned to single-sideband mode and typical system temperatures, on the -scale, were 200 K for the RX100 and 250 K for the RX150. The backends were acousto-optic spectrometers with 1440 channels and a channel width of 0.7 MHz. As with the Onsala 3 mm observations, we used dual-beam switching with a throw of about in azimuth, and pointing errors were typically rms on each axis. Again, the intensity was calibrated using the chopper-wheel method.

### 3 Results and discussion

Upon the removal of a low order baseline and subsequent averaging of the data for each quasar, no absorption features of /channel were found. In Table 3 we summarise the derived upper limits together with the previously published results. Comparing the optical depth limits with those in the four known absorbers, the strongest absorber, towards PKS 1830-211, has for HCO+ and for HCO+ . The weakest, towards PKS 1413+135, has for both CO and HCO+ . The two remaining systems have optical depths similar to that towards PKS 1830-211 and so our new observations and the previous surveys should have been sensitive to (at least) CO and HCO+ at similar strengths to those in 3 of the 4 known systems. Only for the observations of B 0235+1624, HCO+ towards B 2136+141 (IRAM) and CO towards B 1451-375 are the surveys sensitive enough, particularly for resolutions of 10  , to detect an absorption system of a strength similar to that towards PKS 1413+135 (see Table 4).

 DLA Transition V B 0235+1624 0.52398* CO 15.5 0.05-0.5a <0.03 <0.09 < ... ... HCO+ ... ... ... <0.03 <0.3 < ... 0.523869 CS ... ... ... <0.2 <0.9 < B 0248+430 0.3939 CS 17.7 0.20 <3 - - B 0458-020 2.0397*/9 HCO+ 18.4 0.30b <0.3/<0.2 <0.4 < ... 2.0397* CO ... ... ... <0.4 - - ... 2.0399 CO ... ... ... <0.1 <1 < B 0528-2505 2.1408 CO 19.0 <0.2 No published millimetre fluxes B 0738+313 0.2212 CO 16.1 0.07 <1 - - B 0827+243 0.5247 CS 17.3 0.007 <0.1 <0.4 < B 08279+5255 2.97364 HCO+ 15.2 -c - - ... ... CO ... ... ... -c - - B 0834-201 1.715 HCO+ 18.5 <0.1/<0.2 <0.4 < ... ... HCO+ ... ... ... <0.2 <0.6 < ... ... CO ... ... ... - - - B 1017+1055 2.380 CS 17.2 -d - - ... ... CO ... .. ... -d - - B 1215+333 1.9984 CO 18.1 -I,e - - B 1229-0207 0.3950 CO 16.8 <0.3 <1 < ... ... CO ... ... ... <1 - - ... 0.39498* CO ... ... ... - - - B 1328+307 0.69215 HCO+ 17.3 0.11 <0.2/<0.4 <0.7 < ... ... CS ... ... .... <0.3 <2 < ... ... CO ... ... ... <0.3 <1 < ... ... CO ... ... ... <0.6 - - B 1451-375 0.2761 HCO+ 16.7 <0.006 <0.2 <0.6 < ... ... CO ... ... ... <0.09 <0.3 < B 2136+141 2.1346 HCO+ 18.9 <0.08 <0.3 < ... ... CO ... ... ... <0.2 <0.6 < ... ... CO ... ... ... <0.2 <0.8 < ... ... CO ... ... ... <0.6 - -

Notes: where we have observed towards the same quasar *denotes Wiklind & Combes results (see Table 2 for details). aH  I absorption at z=0.52385(Briggs & Wolfe 1983). bH  I absorption at z=2.03945 (Wolfe et al. 1985; Briggs et al. 1989). cAssumed flat spectrum, i.e. S45=S87=S350=0.08 Jy. dAssumed  Jy and  Jy; e  Jy, based on the two available measured values (Table 1).

From the optical depths we may estimate column density limits for each transition. In Table 3 we give the optical depth limits according to a velocity resolution of 10   as well as at a resolution of 1  . We give the former in order to show the limits corresponding to a visual inspection of the spectra3 and facilitate a more direct comparison with the previous surveys. By multiplying the latter with the expected width of a line, the velocity integrated optical depth for a detection at   is obtained. For all of the limits, assuming LTE conditions, we can estimate the total column density of each transition from

 (1)

where is the rest frequency of the transition, and are the statistical weight and the Einstein A-coefficient of the transition, is the velocity integrated optical depth and f is the product of the partition function for an excitation temperature, Tx (assumed 10 K), and (see Wiklind & Combes 1995, 1996b, 1999 for details). From the optical depth limits, we derive limits on the total column density per unit velocity, thus normalising the limits which depend on the spectral resolution to which the data have been smoothed.

 DLA Transition [GHz] B 0248+430 CS 105.3701 4.2 0.39479 0.3941 B 0458-020 HCO+ 87.9994 3.8 2.04045 2.03945 ... ... 88.0229 3.2 2.03964 ... ... CO 151.7144 4.3 2.03888 ... B 0738+313 CO 94.3116 4.4 0.22224 0.2212 B 0834-201 HCO+ 98.5135 4.4 1.71595 - PKS 1413+135 HCO+ 214.6110 4.6 0.24671 0.24671 -

Note: the last entry is a known absorption system included as a guide. A spectrum was obtained after 30 hours with SEST (August 2001) which may well have been rejected upon visual inspection3.

Since we have only upper limits, and thus no knowledge of the width of any line which may be hidden in the noise, as well as uncertainties in the conversion to H2 column densities for molecules other than CO, it is difficult to draw meaningful comparisons between and . However, by assuming an absorption line of   (as in the case of 3 of the 4 known absorbers), we estimate for the lowest optical depth limit (CO in B 0235+1624) a value of  at the level and a resolution of 1  , i.e. . According to (e.g. Wiklind & Combes 1998), the molecular to atomic hydrogen column density ratio is , which is similar to the lower limit estimated from the optical detection of molecular hydrogen in the DLA towards B 0528-2505 (Carilli et al. 1996). That is, although we have no clear detections of any molecular absorption lines, the current limits may be approaching those sufficiently low in order to detect molecular absorption in damped Lyman-alpha systems.

Bearing this in mind, we then analysed the spectra for tentative features within a range corresponding to the uncertainty in the DLA redshift. In each case the continuum level was defined using a polynomial baseline fit and for each channel we generated a 1 error from the rms noise in a window of channels centered on that channel. Using this error array, we identified absorption features as series of n channels over which a deviation from the continuum level was observed with significance > standard deviations. Note that the significance of an absorption feature, , is an overestimate because the flux in adjacent channels is (positively) correlated (each spectral resolution element is sampled at approximately the Nyquist rate, see Murphy et al. (2002) for details).

According to the "expected''    velocity differences in the optical and millimetre wave lines-of-sight5, the only candidate features identified in our analysis are HCO+ and CO towards B 0458-020 as well as the HCO+ line towards B 0834-201. As seen in Table 4, however, increasing the line-of-sight velocity difference to     significantly increases the number of candidates. This suggests that either the authenticity of the candidate features is questionable and/or we have underestimated the uncertainties in the DLA redshifts. It is interesting to note that all but one of the candidate lines occur in DLAs with high column densities ( ). The one exception is B 0834-201 which has the highest visual magnitude of these candidates. We emphasise that Table 4 is only intended as a shortlist for follow up observations, which may or may not yield absorption systems at the listed redshifts.

Of all the candidate systems, the most promising and visually striking example of a possible absorption line is the 151.71 GHz feature towards B 0458-020 (Fig. 1), which has a high visual magnitude as well as the highest neutral hydrogen column density of the sample. Comparing this "detection'' of -8 mK (0.24 Jy at SEST) with the estimated flux density at 152 GHz (0.62 Jy, estimated from the neighbouring values in Table 1) gives an optical depth of . The   FWHM of the line gives a column density estimate of    or   , i.e. 20% of  . This compares well with the column densities and their ratios for the known absorbers, i.e. (Wiklind & Combes 1994, 1995, 1996c, 1999; Carilli et al. 1998), although it is considerably higher than the rest of the DLA sample.

 Figure 1: CO at z=2.039 towards B 0458-020 at the full spectral resolution of 1.38  . The antenna temperature of the peak channel is 8 mK (cf.  mK) and the integrated main-beam intensity of the line is  K  . The expected frequencies of the line for the (approximate) DLA redshift of 2.0399 and H  I  redshift of 2.03945 (Turnshek et al. 1989; Wolfe et al. 1995) are shown. Open with DEXTER

### 4 Summary

We have observed 18 transitions for molecular absorption over the redshift range in 11 damped Lyman-alpha absorption systems which lie along the line-of-sight towards radio-loud quasars. Of these, 9 quasars have flux densities 0.1 Jy in the millimetre band, making our observations sensitive to optical depths 0.1 at a level and a spectral resolution of 10  , thus improving the limits for 11 transitions over the previous results.

From these observations and the previously published results, we estimate an upper limit of 2% for the molecular hydrogen to neutral atomic hydrogen column density ratio. This is the lower limit derived by Carilli et al. (1996) based on a molecular hydrogen detection and a redshifted H  I 21 cm non-detection. This is consistent with the results of Liszt (2002) who suggests that the low metallicity of the Lyman-alpha system is not favourable for the formation of molecules, but our limit perhaps suggests that we are close to the values necessary to detect absorption lines in at least the high column density (   ) DLAs. Subsequently, from an analysis of the spectra we find several significant features located close to the expected redshifts, which do, however, require confirmation. Therefore follow up observations of the candidates given in this article should be the next step in searching for the elusive high redshift molecular absorption systems.

Acknowledgements
We wish to thank the John Templeton Foundation for supporting this work. SJC acknowledges receipt of a UNSW NS Global Fellowship and FTR acknowledges support from the Chilean Centro de Astrofísica FONDAP No. 15010003.

## References

• Briggs, F. H., & Wolfe, A. M. 1983, ApJ, 268, 76 In the text NASA ADS
• Briggs, F. H., Wolfe, A. M., Liszt, H. S., Davis, M. M., & Turner, K. L. 1989, ApJ, 341, 650 In the text NASA ADS
• Carilli, C. L., Lane, W., de Bruyn, A. G., Braun, R., & Miley, G. K. 1996, AJ, 111, 1830 In the text NASA ADS
• Carilli, C. L., Menten, K. L., Reid, M. J., Rupen, M. P., & Yun, M. S. 1998, ApJ, 494, 175 In the text NASA ADS
• Carilli, C. L., et al. 2000, PhRvL, 85, 5511 In the text NASA ADS
• Combes, F., & Wiklind, T. 1998, ESO Messenger, 91, 29
• Cowie, L. L., & Songaila, A. 1995, ApJ, 453, 596 NASA ADS
• Curran, S. J., Webb, J. K., Murphy, M. T., Bandiera, R., Corbelli, E., & Flambaum, V. V. 2002, PASA, 19, 455 In the text
• Douglas, N. G., Radford, S. J. E., Roland, J., & Webb, J. K. 1992, A&A, 262, 8 NASA ADS
• Drinkwater, M. J., Webb, J. K., Barrow, J. D., & Flambaum, V. V. 1998, MNRAS, 295, 457 In the text NASA ADS
• Foltz, C. B., Chaffee, F. H., & Black, J. H. 1988, ApJ, 324, 267 NASA ADS
• Ge, J., & Bechtold, J. 1997, ApJ, 477, L73 NASA ADS
• Ivanchik, A. V., Potekhin, A. Y., & Varshalovich, D. A. 1999, A&A, 343, 439 NASA ADS
• Le Brun, V., Bergeron, J., Boisseé, P., & Deharveng, J. M. 1997, A&A, 321, 733 NASA ADS
• Ledoux, C., Srianand, R., & Petitjean, P. 2002, A&A, 392, 781 NASA ADS
• Levshakov, S. A., Molaro, P., Centurión, M., et al. 2000, A&A, 361, 803 NASA ADS
• Levshakov, S. A., Dessauges-Zavadsky, M., D'Odorico, S., & Molaro, P. 2002, ApJ, 565, 696 NASA ADS
• Liszt, H. 2002, A&A, 389, 393 In the text NASA ADS
• Lovas, F. J. 1992, J. Phys. Chem. Ref. Data, 21(2), 181 NASA ADS
• Murphy, M. T., Webb, J. K., Flambaum, V. V., et al. 2001a, MNRAS, 327, 1244 In the text NASA ADS
• Murphy, M. T., Webb, J. K., Flambaum, V. V., et al. 2001b, MNRAS, 327, 1208 In the text NASA ADS
• Murphy, M. T., Webb, J. K., Flambaum, V. V. Churchill, C. W., & Prochaska, J. X. 2001c, MNRAS, 327, 1223 In the text NASA ADS
• Murphy, M. T., Curran, S. J., & Webb, J. K. 2002, MNRAS, submitted In the text
• Petitjean, P., Srianand, R., & Ledoux, C. 2002, MNRAS, 332, 383 NASA ADS
• Srianand, R., Petitjean, P., & Ledoux, C. 2000, Nature, 408, 931 NASA ADS
• Turnshek, D. A., Wolfe, A. M., Lanzetta, K. M., et al. 1989, ApJ, 344, 567 In the text NASA ADS
• Webb, J. K., Flambaum, V. V., Churchill, C. W., Drinkwater, M. J., & Barrow, J. D. 1999, PhRvL, 82, 884 NASA ADS
• Webb, J. K., Murphy, M. T., Flambaum, V. V., et al. 2001, PhRvL, 87, 091301 In the text
• Wiklind, T., & Combes, F. 1994a, A&A, 286, L9 NASA ADS
• Wiklind, T., & Combes, F. 1994b, A&A, 288, L41 In the text NASA ADS
• Wiklind, T., & Combes, F. 1995, A&A, 299, 382 In the text NASA ADS
• Wiklind, T., & Combes, F. 1996a, in ESO Symposia (Springer), ed. P. Shaver, 86 In the text
• Wiklind, T., & Combes, F. 1996b, A&A, 315, 86 In the text NASA ADS
• Wiklind, T., & Combes, F. 1996c, Nature, 379, 139 In the text NASA ADS
• Wiklind, T., & Combes, F. 1997, A&A, 328, 48 In the text NASA ADS
• Wiklind, T., & Combes, F. 1998, ApJ, 500, 129 In the text NASA ADS
• Wiklind, T., & Combes, F. 1999, in Highly Redshifted Radio Lines, ASP Conf. Ser., 156, 202 In the text
• Wiklind, T., & Combes, F. 2001, in Gravitational Lensing, ed. T. G. Brainerd, & C. S. Kochanek (San Francisco: Astronomical Society of the Pacific), ASP Conf. Proc., 237, 155 In the text
• Wolfe, A. M., Briggs, F. H., Turnshek, D. A., et al. 1985, ApJ, 294, L67 In the text NASA ADS
• Wolfe, A. M., Lanzetta, K. M., Foltz, C. B., & Chaffee, F. H. 1995, ApJ, 454, 698 In the text NASA ADS