With this Herschel Key Project we aim to answer the following relevant questions concerning the dust formation in the young universe:


In order to address these points observationally we will specifically focus on the following relevant aspects:

We will investigate the FIR/sub-mm SEDs of the highest redshift QSOs. Currently, 80 QSOs with redshift z > 5 are known and are included in our target list. In order to investigate radio-loud quasars, a dozen of the highest redshift radio quasars and galaxies are included. An accurate measurement of the FIR/Sub-mm spectral energy distribution will be used to derive dust temperatures, dust masses, FIR luminosities, etc. of the QSO host galaxies.

We will investigate the sub-class of Broad Absorption Line QSOs, where outflows from the QSO core itself rather than winds from evolved stars are considered to be an important source of dust. Since very few BAL QSOs are known at z > 4, a sizable sample of source in the lower redshift range 2.0 ≤ z ≤ 2.5 will be studied to establish the differences between the FIR properties of BAL and non-BAL QSOs.

We will investigate the physical conditions of the gas and dust in high redshift QSOs and galaxies in more detail by employing FIR line spectroscopy. The highest redshift objects accessible for FIR spectroscopy are sources that are significantly amplified by lensing foreground galaxies and clusters. Since the sources are nevertheless rather faint, only a small number of sources can be targeted with the PACS spectrometer to detect narrow fine structure emission lines and possibly other, broader spectral features.



Dust in the highest redshift QSOs (95 hours)

PI: Klaus Meisenheimer (MPIA)

We plan to determine the Far-InfraRed (FIR) to sub-mm part of the SED for all known quasars with redshift z > 5. Together with existing Near- and Mid-IR photometry and ground-based observation at wavlengths > 800 μm we will get the complete SED over more than three decades in wavelengths. Thus the three most important components of the quasar spectrum: emission from the hot accretion disk, radiation from host dust close to the AGN core, and dust emission from cooler dust will be covered. A proper determination of the important properties of the dust: total dust luminosity, temperature distribution, and dust mass requires that the peak of the SED around 200 μm is well sampled and measured with high signal-to-noise ratio. To do so we need to achieve a S/N ratio around 10 even for the faintest sources (expected: Speak>= 10mJy). The observed SEDs will be compared with radiative transfer models for dust tori in AGN (Schartmann et al. 2005, 2008, 2009) and star burst models.

Even with the point-source photometry AOT, PACS provides a field of 0.85'x1.7' that will get the full integration time. Due to chopping and nodding a total area of 9 square arcminutes will receive at least half of the integration time. The SPIRE photometry will cover a field of 2'x4' (not fully sampled). We will therefore also be able to detect objects that are closely related to the prime target with the point source sensitivity and other infrared galaxies within a projected area of about 1 Mpc2. Together with the IRAC and MIPS 24 μm images (which will help to identify objects detected in the FIR), we will be able to study the cluster environment of high-z QSOs to unprecedented depth. In addition, the images will add up to a very deep survey at mid-infrared to sub-mm wavelengths covering some 720 square arcminutes, which is unique in the combination of depth and wavelength coverage. We foresee, that this "byproduct" of our targeted photometry will eventually turn out to be of similar legacy value as the pointed observationes themselves. We plan to add ground-based NIR photometry and spectroscopy to derive redshifts of the field objects (at telescopes to which we have priviliged access).



Typical SED of a dust-rich QSO at redshift z=6 (combined measurements of SDSS J1148+5251 (z=6.41, dots) and SDSSJ 1044-0125 (BAL at z=5.71, diamonds), shifted to z=6). The far-infrared to sub-mm wavelengths are essential for determining the dust temperature. The 5σ-limits to be reached by PACS (110/170) and SPIRE (250/350/490) are shown as the lower edge of vertical bars.


Dust in high-redshift Broad Absoprtion Line (BAL) Quasars (20 hours)

PI D. Hutsemekers (U Liege)

Up to 20% of QSOs show broad absorption lines (BALs) in their spectra indicating high-velocity outflows (v ≈0.1 c, Reichard et al. 2003). Most BAL quasars have outflowing absorption only in high ionization lines like CIV ('HiBALs'), but 1 in 7 BAL quasars also display outflows in low-ionization lines like MgII as well ('LoBALs'), and 1 in 50 are 'FeLoBALs' with excited-state FeII absorption. Since BAL outflows have mass loss rates comparable to quasar accretion rates and may be present around every quasar for at least part of its lifetime, understanding the BAL phenomenon is essential for our understanding of quasars in general, as well as for understanding the feedback effects of outflows on theIGM (Furlanetto & Loeb 2001) and on AGN hosts (Appleton et al. 2002).

LoBAL QSOs apparently contain large amounts of dust, as suggested by their optical-UV continuum (Reichard et al. 2003) and by their far-IR colours (Low et al. 1989; but see Willot et al. 2003). These objects could be young quasars just emerging from their dust cocoon. The presence of strong FeII absorption in some of them points towards recent star formation. On the other hand, HiBAL and normal QSOs would show lower extinction after heating and destruction of dust. Thus an evolutionary sequence like "FeLoBAL → LoBAl → HiBAL → normal non-BAL QSO" might exist. Alternatively, the difference between the various types of quasars might be an orientation effect. In this case our line-of-sight through the outflow determines the BAL classification.

Only a handful of BAL QSOs - the brightest or the gravitationally lensed ones - have already been measured in the far-infrared with IRAS or ISO. For a few more (yet unpublished) Spitzer observations exist. For high-redshift (2 ≤ z ≤ 4) BAL QSOs, cold dust emission is expected to peak at 200 - 400 μm with flux densities around 30 - 50 mJy. Herschel's instruments PACS and SPIRE are ideally suited to accurately determine their dust properties (detailed SED, total mass, temperature). We plan to discriminate between different BAL scenarios by correlating the (non-directional) dust emission with the BAL class: If all BAL QSOs are the same objects seen under different orientations, very similar dust properties are expected for the different sub-types. On the other hand, the detection of a correlation between dust properties and BAL QSO sub-types would support the evolutionary scenario. The BAL QSOs sample has been drawn from the SDSS. Since the SDSS provides good spectroscopic data for all objects, sub-types can be accurately determined as well as the "BALcinity" and detachment indices (i.e. roughly the depth and shape) of the line profiles.



PACS Spectroscopy of gravitationally lensed high-redshift QSOs and Galaxies (50 hours)

PI L. Tacconi (MPE)

As in dusty local ultraluminous galaxies, the central active regions of dusty high-z QSOs and sub-millimeter galaxies have very high extinctions that are equivalent to tens of magnitudes in the visual. Use of infrared tracers is mandatory to penetrate this obscuration and obtain a complete view. The spectroscopic properties in the far-infrared are completely unexplored at these redshifts, since before PACS there has not been a spectrometer onboard a space infrared mission that has been sensitive enough to detect them. They are, however, crucial to estimate the relative contributions of AGN activity and star formation to the total energy output of these systems.

From our previous experience at low redshift, we conclude that the best method relies on the fine structure line spectrum (see figure below), which contains tracers of low excitation gas (H II region) and high excitation gas (AGN NLR) as well as partially inized zones (AGNs, shocks). Rest-frame Mid-Infrared fine structure diagnostic diagrams equivalent to the traditional optical ones (e.g. Veilleux & Osterbrock 1987) can be used to classify and quantify activity in obscured objects (Sturm et al. 2002). With PACS onboard Herschel, we will be able to apply this technique for the first time to the high-redshift universe in a small number of highly lensed quasars and dusty galaxies.

We therefore propose to carry out PACS rest-frame mid-infrared spectroscopy for four lensed, obscured QSOs and sub-millimeter galaxies to quantify the contribution of star formation and AGN activity in these objects, and to study their obscuration and physical conditions. Note that we only require unobstructed (NH < 1023 cm-2) views to the narrow-line region (NLR; 100 - 1000 pc scale), not to the central X-ray source where line-of-sight column densities may sometimes exceed 1024 cm-2. We plan to target mainly the fine structure lines of [SIII] (λrest = 33.48 μm) and [OIV] (λrest = 25.98 μm) as the main mid-IR diagnostic lines in the PACS wavelengths bands for low excitation and high excitation gas, respectively, at redshifts ≈ 2-4. With a detection of [SIII] and either a detection or a sensitive limit on [OIV] we will be able, for the first time, to place high redshift QSOs and obscured star forming galaxies on excitation diagnostic diagrams. We will make quantitative assessment of the relative importance of star formation and AGN activity to the energy output of systems that are at their peak of star and black hole formation activity. We will then compare these estimates with those derived from Spitzer low resolution spectra of aromatic "PAH" features found in star forming galaxies, and rest frame ≈6 μm HII region and AGN continuum slopes. Such low resolution observations be available for one of our objects, in particular a large number of z ≈ 2-3 dusty star forming galaxies and QSOs in general.


Examples of fine structure lines and PAH features in AGNs. Note that the [NeII]/[OIV] ratio correlates with the strength of the PAH feature, which is absent in AGN heated dust.