next up previous contents index
Next: Some Recent Results Up: QSOsGalaxies, and Previous: Automated Morphological Classification

Science with the Hubble Space Telescope -- II
Book Editors: P. Benvenuti, F. D. Macchetto, and E. J. Schreier
Electronic Editor: H. Payne

Intergalactic Helium Absorption Toward Quasars

Peter Jakobsen

Astrophysics Division, Space Science Department of ESA, ESTEC, 2200 AG Noordwijk, The Netherlands



Strong redshift-smeared HeII absorption from singly ionized intergalactic helium has so far been detected in the far-UV spectra of three quasars, twice with HST at and once with HUT/Astro-2 at . Weak HeI lines of neutral helium have also been detected in several metal line systems with HST. The helium detections confirm that intergalactic gas is highly ionized and provide an important, but qualitative, consistency check of Big Bang nucleosynthesis theory. However, since the existing HeII observations cannot distinguish between absorption due to unresolved lines in the HeII forest and true Gunn-Peterson absorption, it is not clear at present whether a diffuse IGM is detected in the data.

Introduction and Background

Nearly all that is known about the content and physical conditions in intergalactic space at high redshift stems from the study of absorption lines in quasar spectra. A long-standing goal of ultraviolet space astronomy---and HST in particular---has been to extend such studies to cover the redshifted extreme ultraviolet resonance lines of neutral helium at Å and singly ionized helium at Å. The element helium not only plays a central role in modern cosmology, but the singly ionized HeII ion is also an important diagnostic of ionized gas.

The Lyman forest clouds are of particular interest in the context of helium observations. Although it is now known that at least some forest systems contain heavy elements, (Cowie et al. 1995, Tytler et al. 1995), the forest clouds are still the closest approximation to primordial gas currently known. From the so-called proximity effect (Carswell et al. 1982, Murdoch et al. 1986, Bajtlik, Duncan & Ostriker 1988, Lu, Wolfe & Turnshek 1991, Bechtold 1994), the Lyman forest clouds are inferred to be highly photoionized (to of order one part on in HI) by a diffuse meta-galactic background flux believed to be due to the integrated light of quasars and/or young galaxies (Bechtold et al. 1987, Miralda-Escudé & Ostriker 1990,1992, Madau 1992). The helium present in the forest clouds is therefore also expected to be highly ionized, and detectable mainly in the form of singly ionized HeII rather than neutral HeI. In particular, one clear prediction is that the forest should be very weak in HeI584, but extremely strong in HeII304, with the density of residual HeII in the forest clouds exceeding that of HI by one order of magnitude or more, depending on the spectral shape of the ionizing background (Sargent et al. 1980, Miralda-Escudé 1993, Jakobsen et al. 1993).

A second motivation for pursuing observations of helium absorption is the hope of detecting a smoothly distributed intergalactic medium (IGM) through its redshift-smeared line absorption (Gunn & Peterson 1995). Although Gunn-Peterson absorption has not been reliably detected in HI Ly (Steidel & Sargent 1987, Schneider, Schmidt & Gunn 1991, Webb et al. 1992, Giallongo, Cristiani & Trevese 1992) or HeI584 (Beaver et al. 1991, Reimers et al. 1989, Tripp, Bechtold & Green 1994), a highly photoionized IGM could still reveal itself through detectable Gunn-Peterson absorption in the HeII 304 line.

A considerable hurdle standing in the way of helium absorption measurements, however, is presented by the fact that the HeI and HeII resonance lines fall in the extreme ultraviolet. The HeII 304 transition is only accessible above the MgF short wavelength cutoff of the HST optics in the case of the most remote---and therefore faint---quasars at redshifts z>3. Moreover, since the transition lies below the Lyman limit at Å, the vast majority of z>3 quasars are predicted to be completely unobservable at redshifted HeII 304 due to strong photoelectric absorption in the numerous Lyman forest and Lyman limit systems intercepted by the line of sight out to such large redshift (Møller & Jakobsen 1990, Picard & Jakobsen 1993, Zuo & Phinney 1993).

Detections of Helium Absorption

Since the launch of HST, there have been two systematic searches for rare cases of z>3 quasars whose brightness and exceptionally clear line of sight combine to make detection of redshifted HeII absorption feasible: the FOC objective prism survey (Jakobsen et al. 1993) and the FOC ``snapshot'' imaging survey of David Tytler and coworkers. Between them, these two surveys have so far sampled the far-UV fluxes more than 110 bright z>3 quasars selected from the literature. Both surveys have now met with success and each have uncovered an object against which the first detections of strong intergalactic HeII absorption at have been made (Jakobsen et al. 1994, Tytler et al. 1995).

Other successes concern the remarkable UV bright quasar HS1700+6416 discovered by Reimers et al. (1989). Although of too low redshift to enable HeII304 to be reached with HST, FOS observations of HS1700+6416 have led to the first detections of HeI in quasar absorption systems (Reimers et al. 1992, Reimers & Vogel 1993). More recently, shorter wavelength spectroscopic observations of HS1700+6416 obtained with the Hopkins Ultraviolet Telescope flown on the Astro-2 mission have detected HeII absorption at redshifts z<3 which are inaccessible to HST (Davidsen, Kriss & Zheng 1996)

The HeI and HeII detections confirm that HeI absorption is universally very weak, while HeII absorption is strong compared to HI. This provides further confirmation that quasar absorption systems---and the Lyman forest clouds in particular---are indeed highly ionized, and that the observable ions HI, HeI, and possibly HeII, therefore, represent only minority species. One consequence of this high degree ionization is that it precludes being able to derive helium abundances from the measurements at any level of accuracy of interest to Big Bang nucleosynthesis theory because of the large and uncertain ionization corrections involved. Nonetheless, the mere fact that intergalactic gas at these large redshifts has been shown to contain helium still provides an important, albeit qualitative, confirmation of a key prediction of Big Bang cosmology

HeI Absorption

The UV continuum of HS1700+6416 was first detected by Reimers et al. (1989) using IUE. The high intrinsic brightness of HS1700+6416 () and its exceptionally clear line of sight conspire to make this object the most UV bright high redshift quasar presently known (see, however, Reimers et al. 1994). The incredibly rich extreme ultraviolet absorption spectrum of HS1700+6416 has been extensively studied with HST by Reimers et al. (1992) and Vogel & Reimers (1995) using the FOS (and more recently the GHRS). These observations confirm the prediction that HeI lines corresponding to the Ly forest are too weak to be detectable, but weak HeI lines are detected in the case of four denser metal line systems in the spectrum at --2.43.

A detailed analysis of the HeI absorption in HS1700+6416 is given by Reimers & Vogel (1993), who show that the strengths of the HeI lines appear to be a factor of higher than predicted for the single component photoionization models that best fit the ionization of the heavy elements, assuming a cosmological helium abundance. As also discussed by Reimers & Vogel (1993), however, this discrepancy probably does not imply a high helium abundance, but rather reflects inadequacies in accurately modeling highly ionized quasar absorption systems.

HeII Absorption

Singly ionized HeII absorption has now been detected toward three quasars, twice with HST and once with HUT/Astro-2. It is important to appreciate that because of the exceeding faint UV fluxes involved, all these detections push the sensitivity limits of present day instrumentation. Neither the existing HST nor the HUT observations have the spectral resolution nor signal to noise ratio to resolve individual spectral lines. What is detected is the redshift-smeared absorption trough shortward of the HeII line in the quasar rest frame due to the combined HeII opacity of both the HeII forest and the diffuse IGM, and the quantitative strength of this ``Gunn-Peterson'' trough as measured by its wavelength-averaged effective optical depth, .

Figure: Far-ultraviolet spectrum of Q0302-003 obtained with the FOC in objective prism mode. The thin solid line gives the 1 uncertainty per Å wavelength bin. The position of the HeII line in the quasar rest frame is also marked.

The quasar Q0302-003 (, ) was first detected in the far-UV in the FOC prism survey (Jakobsen et al. 1993), and then re-observed with the COSTAR-corrected FOC (Jakobsen et al. 1994). The low resolution ultraviolet spectrum of this quasar obtained with the FOC far-UV objective prism is shown in Figure 1. The spectrum of Q0302-003 is seemingly completely absorbed on the blue side of the redshifted HeII line, with no flux detected below the edge at the sensitivity limit of the 3 hour FOC integration. This places a lower limit on the total effective HeII optical depth at of (90% confidence).

Figure: FOC far-UV prism spectrum of PKS 1935-692 with the position of the HeII line in the quasar rest frame marked. The thin solid line gives the 1 uncertainty per Å wavelength bin. The depressions in the spectrum near 1580Å and 1330Å are due to an intervening damped Ly system at a redshift low enough not to obscure the HeII absorption.

This finding has recently been confirmed by a second HST detection toward the quasar PKS 1935-692 (, ) . This object was first detected in the UV during FOC snapshot survey of David Tytler and coworkers, and was subsequently re-observed spectroscopically using the FOS (Tytler et al. 1995). The FOS spectrum of PKS 1935-692 initially suggested less intense HeII absorption () than seen toward Q0302-003 at slightly higher redshift. However, this has not been confirmed by very recent follow-up observations of PKS 1935-692 taken with the FOC objective prism (Jakobsen & Tytler, Director's Discretionary Program 6156). The FOC prism spectrum of PKS 1935-692 (Figure 2) instead shows no detectable flux below the redshifted HeII line, yielding the limit (90% confidence). Given that the FOC spectrum (which is otherwise in good agreement with the FOS spectrum at wavelengths above the HeII edge) is technically speaking the cleaner of the two in terms of measuring on account of the accurate background subtraction afforded by the two-dimensional FOC detector, it is believed that the discrepancy with the initial suggestion of can be traced to residual background/stray light in the FOS at the painfully faint signal levels involved. The two existing HeII detections thus appear to be mutually consistent and can be combined statistically to yield the measure: (90% confidence) at .

The HST HeII measurement at high redshift can now be compared to the recent successful detection of HeII absorption at lower redshifts toward HS1700+6416 by Davidsen, Kriss, & Zheng (1996) with HUT/Astro-2. The impressive HUT spectrum of HS1700+6416 shows a clear and finite flux in the HeII trough, which implies a lower HeII opacity over the redshift range sampled by the HUT data. The inferred value of at quoted by Davidsen, Kriss, & Zheng (1996) is plotted together with the combined HST measure in Figure 3.

Figure: Comparison of the HST and HUT measurements of and the redshift ranges probed by the data. The two continuous curves show the expected evolution in redshift for a non-evolving ionizing background in the two extreme cases where the absorption is completely dominated by unresolved line absorption in the HeII forest and true Gunn-Peterson absorption in a diffuse IGM.

Uncertainties of Interpretation

It is presently an open question whether the strong HeII absorption detected is dominated by unresolved line absorption or by true Gunn-Peterson absorption from a diffuse IGM. Since it is only the sum of the two contributions that is observed, and none of the existing HeII spectra is capable of resolving the HeII forest, this question can only be addressed indirectly. As a consequence, the parameter space of the problem is still very large, permitting both forest- and IGM-dominated interpretations.

The optical depth of the Gunn-Peterson HeII 304 absorption trough produced by a smooth IGM having HeII volume density in an universe is given by

where km s Mpc is the Hubble constant, cm the integrated cross section of the HeII 304Å transition. The corresponding wavelength-averaged contribution to from HeII line blanketing in the Lyman forest systems can be written

where is the number of forest lines per unit redshift and is the mean HeII 304Å equivalent width of the forest systems averaged over their distribution in column density.

Both contributions to are expected to evolve rapidly with redshift. In the simplest case of an IGM held photoionized by a background flux of constant intensity and spectral shape, it is easy to show that Eqn.(1) predicts that should evolve as . Similarly, since the Lyman forest line density evolves as where (Sargent et al. 1980, Murdoch et al. 1986), the same assumptions lead to the prediction . From Figure 3 it is seen that---at least to first order---the observed relative increase in seen between the HUT and HST measurements is consistent with any combination of these two naive evolution models.

The absolute value of the forest contribution to can be also estimated from Eqn.(2) based on the known characteristics of the forest in HI Ly as measured from the ground. However, the calculated value of depends critically on both the assumed ionization level of the forest gas and its detailed distribution in column density. Of particular importance to this problem is the recent Keck discovery that the column density spectrum of Lyman forest clouds extends uninterrupted down to cm and possibly lower (Songaila, Hu, & Cowie 1995, Tytler et al. 1995). This population of very low column forest clouds is very important to the HeII opacity problem since it is conceivable that the ionization conditions in the forest are such that the residual density of HeII in the forest clouds is much larger than the HI density, in which case the low column population dominates the HeII absorption to the point of being able to explain all the absorption observed without any need for a diffuse IGM (Miralda-Escudé 1993, Jakobsen et al. 1994, Madau & Meiksin 1994, Songaila, Hu, & Cowie 1995, Giroux, Fardal & Shull 1995, Davidsen, Kriss & Zheng 1996).

This basic uncertainty concerning the exact geometry and ionization conditions in the plasma giving rise to the detected HeII absorption also affects the quantitative estimates of the total amount of intergalactic baryonic matter detected. The HeII measurements can only provide a lower limit on the baryonic mass detected since there is no way of directly measuring the amount of unseen mass present in the form of invisible HeIII. Moreover, since it is the net opacity in HeII that is measured, the assumed geometry of the absorbing gas is crucial in determining the amount of absorbing gas detected, since it takes many more HeII ions to produce a given amount of absorption if the gas is clumped in self-shielding clouds rather than uniformly distributed. This last issue is also important for the question of whether a proximity effect should be detectable in HeII (Zheng & Davidsen 1994, Giroux, Fardal, & Shull 1995).

In one extreme, if it is assumed that all the HeII detected is in a diffuse IGM, applying equation (1) to the HST measurement of at leads to the following lower limit on the cosmological density of gas detected:

where is the fraction of helium in singly ionized form, and a cosmological helium to hydrogen abundance of by number has been assumed. The matching calculation in the opposite extreme case of Eqn.(2) where pure forest line absorption is assumed leads to

i.e., a density some three orders of magnitude higher. Very similar estimates result from the equivalent analysis of the HUT results.

In the first case, only an extremely low density IGM having is required to explain the observations. In contrast, the required baryonic density in the second case is comparable to the canonical Big Bang nucleosynthesis density of (Walker et al. 1991). In other words, for the detected HeII absorption to be dominated by unresolved forest absorption, it is necessary that the forest clouds at high redshift contain the bulk of the baryonic mass of the universe. Interestingly, entirely different considerations prompted by other HST observations of the forest clouds suggest that the latter extreme may be closest to the truth (Rauch & Haehnelt 1995).


Bajtlik, S., Duncan, R. C., & Ostriker, J. P. 1988, ApJ, 327, 570

Beaver, E. A, Burbidge, E. M., Cohen, R. D., Junkkarinen, V. T., Lyons, R. W., Rosenblatt, E. I., Hartig, G. F., Margon, B., & Davidsen, A. F. 1991, ApJ, 377, L9

Bechtold, J. 1994, ApJS, 91, 1

Bechtold, J., Czerny, B., Elvis, M., Fabbiano, G., & Green, R. F. 1987, ApJ, 314, 699

Carswell, R. F., Whelan, J. A. J., Smith, M. G., Boksenberg, A., & Tytler, D. 1982, MNRAS, 198,91

Cowie, L. L., Songaila, A., Kim, T. S., Hu, E. M. 1995, AJ, 109, 1522

Davidsen, A. F., Kriss, G. A., & Zheng, W. 1996, Nature (in press)

Giallongo, E., Cristiani, S., & Trevese, D. 1992,ApJ, 398, L9

Giroux, M. L., Fardal, M. A., & Shull, M. J. 1995 ApJ, 451, 477

Gunn, J. E., & Peterson, B. A. 1965, ApJ, 142, 1633

Hu, E. M., Kim, T. S. Cowie, L. L., Songaila, A., & Rauch, M. 1995, AJ, 110, 1526

Jakobsen, P. et al. 1993, ApJ, 417, 528

Jakobsen, P., Boksenberg, A., Deharveng, J. M., Greenfield, P., Jedrzejewski, R., & Paresce, F. 1994, Nature 370,35

Lu, L., Wolfe, A. M., & Turnshek, D. A. 1991, ApJ, 367, 19

Madau, P. & Meiksin, A. 1994, ApJ, 433, L53

Madua, P. 1992, ApJ, 389, L1

Miralda-Escudé, J. 1993, MNRAS, 262, 273

Miralda-Escudé, J. & Ostriker, J. P. 1990, ApJ, 350, 1

Miralda-Escudé, J. & Ostriker, J. P. 1992, ApJ, 392, 15

Murdoch, H. S., Hunstead, R. W., Pettini, M., & Blades, J. C. 1986, ApJ, 309

Møller, P. & Jakobsen, P. 1990, A&A, 228, 299

Picard. A. & Jakobsen, P. 1993 A&A, 276, 331

Rauch, M. & Haehnelt, M. G. 1995, MNRAS, 275, L76

Reimers, D. & Vogel, S. 1993, A&A, 276, L13

Reimers, D., Vogel, S., Hagen, H. J., Engels, D., Groote, D., Wamsteker, W., Clavel, J., & Rosa, M. R. 1992, Nature 360, 561

Reimers, D., Clavel, J., Groote, D., Engels, D., Hagen, H. J., Naylor, T., Wamsteker, W., & Hopp, U. 1989, A&A, 218, 71

Reimers, D., Rodriguez-Pascual, P., Hagen, H. J., Wisotski, L. 1994, A&A, 293, L21

Sargent, W. L. W., Steidel, C. C., & Boksenberg, A. 1989, ApJS, 69, 703

Sargent, W. L. W., Young, P. J., Boksenberg, A., & Tytler, D. 1980, ApJS, 42, 41

Schneider, D. P., Schmidt, M., & Gunn, J. E. 1991, AJ, 101, 2004

Songaila, A., Hu, E. M., & Cowie, L. L. 1995, Nature, 375, 124

Steidel, C. C. & Sargent, W. L. W. 1987, ApJ, 318, L11

Tripp, T. M., Bechtold, J., & Green, R. F. 1994, ApJ, 433, 533

Tytler, D., Fan, X. M., Burles, S., Cottrell, L., Davis, C., Kirkman, D., & Zuo, L. 1995, in QSO Absorption Lines, ed. G. Meylan (Berlin: Springer), p.289

Vogel, S. & Reimers, D. 1995, A&A, 294, 377

Walker, T. P., Steigman, G. Schramm, D. N., Olive, K. A., & Kang, H.-S. 1991, ApJ, 376, 51

Webb, J. K., Barcons, X., Carswell, R. F., & Parnell, H. C. 1992, MNRAS, 255, 319

Zheng, W. & Davidsen, A. 1994, ApJ, 440, L53

Zuo, L. & Phinney, E. S. 1993, ApJ, 418, 28

next up previous contents index
Next: Some Recent Results Up: QSOsGalaxies, and Previous: Automated Morphological Classification