A burst of outflows from the Serpens cloud core: wide-field submillimetre continuum, CO J=2–1 and optical observations

Wide-field mapping of Serpens in submillimetre continuum emission and CO J =2–1 line emission is here complemented by optical imaging in [S  ii ] λλ 6716, 6731 line emission. Analysis of the 450- and 850-μm continuum data shows at least 10 separate sources, along with fainter diffuse background emission and filaments extending to the south and east of the core. These filaments describe 'cavity-like' structures that may have been shaped by the numerous outflows in the region. The dust opacity index, β , derived for the identifiable compact sources is of the order of 1.0±0.2, with dust temperatures in excess of 20 K. This value of β is somewhat lower than for typical class I YSOs; we suggest that the Serpens sources may be 'warm', late class 0 or early class I objects.
With the combined CO and optical data we also examine, on large scales, the outflows driven by the embedded sources in Serpens. In addition to a number of new Herbig–Haro flows (here denoted HH 455–460), a number of high-velocity CO lobes are observed; these extend radially outwards from the cluster of submillimetre sources in the core. A close association between the optical and molecular flows is also identified. The data suggest that many of the submillimetre sources power outflows. Collectively, the outflows traced in CO support the widely recognized correlation between source bolometric luminosity and outflow power, and imply a dynamical age for the whole protostellar cluster of ∼3×10⁴ yr. Notably, this is roughly equal to the proposed duration of the 'class 0' stage in protostellar evolution.     

Application of a clumpy core model to NGC 2071

The multi-transitional observations of CS molecules towards the NGC 2071 core have been re-analysed by using a tri-dimensional Monte Carlo radiative transfer code. Better agreement with the observations is made by an introduction of clumpiness to this model than by smoothly varying density to the 1D microturbulent one. The best-fitting model shows that, when a unique density is assumed for clumps, the volume filling factor of the clumps varies as r ⁻² with an average of ∼5 per cent over the entire core, and that the H₂ number density and the CS abundance of the clump relative to H₂ are ∼ 2 × 10⁶ cm⁻³ and ∼ 6 × 10⁻¹⁰, respectively. The radial density gradient ∝ r ⁻² obtained from our clumpy core model is steeper than that (∝ r ^−1.3) obtained from the microturbulent model. Since all clumps are subject to random bulk motions in this 3D clumpy macroturbulent model, synthesized line profiles do not show self-absorption dips even for opaque transitions and the resulting linewidth is in good accordance with the observations.     

The high-velocity molecular stream in W51B: a ring structure with compact H ii regions

We present ¹³ CO J  = 1 − 0 line observations of the H  ii region complex W51B located in the high-velocity (HV) stream. These observations reveal a filamentary and clumpy structure in the molecular gas. The mean local standard of rest (LSR) velocity ∼ + 65 km s⁻¹ of the molecular gas in this region is greater than the maximum velocities allowed by kinematic Galactic rotation curves. The size and mass of the molecular clouds are ∼ 48 × 17 pc² and ∼ 2.4 × 10⁵ M⊙ respectively. In a position–velocity diagram, molecular gas in the southern part comprises a redshifted ring structure with v _LSR=+ 60 to +73 km s⁻¹. The velocity gradient of this ring is ∼ 0.5 km s⁻¹ pc⁻¹, and the mass is ∼ 6.2 × 10⁴ M⊙. If we assume that the ring is expanding with a uniform velocity, the expansion velocity, radius and kinetic energy are ∼ 7 km s⁻¹, ∼ 13 pc and ∼ 3.0 × 10 ⁴⁹ erg respectively. The kinetic energy and mass spectrum of the ring could be explained by an expanding cylindrical cloud with a centrally condensed mass distribution. The locations of two compact H  ii regions, G49.0−0.3 and G48.9−0.3, coincide with the two molecular clumps in this ring. We discuss star formation, and the mechanism that produced the ring structure.     

A shallow though extensive H2 2.122-μm imaging survey of Taurus–Auriga–Perseus – I. NGC 1333, L1455, L1448 and B1

We discuss wide-field near-infrared (near-IR) imaging of the NGC 1333, L1448, L1455 and B1 star-forming regions in Perseus. The observations have been extracted from a much larger narrow-band imaging survey of the Taurus–Auriga–Perseus complex. These H₂ 2.122-μm observations are complemented by broad-band K imaging, mid-IR imaging and photometry from the Spitzer Space Telescope , and published submillimetre CO   J = 3–2  maps of high-velocity molecular outflows. We detect and label 85 H₂ features and associate these with 26 molecular outflows. Three are parsec-scale flows, with a mean flow lobe length exceeding 11.5 arcmin. 37 (44 per cent) of the detected H₂ features are associated with a known Herbig–Haro object, while 72 (46 per cent) of catalogued HH objects are detected in H₂ emission. Embedded Spitzer sources are identified for all but two of the 26 molecular outflows. These candidate outflow sources all have high near-to-mid-IR spectral indices (mean value of  α∼ 1.4  ) as well as red IRAC 3.6–4.5 μm and IRAC/MIPS 4.5–24.0 μm colours: 80 per cent have [3.6]–[4.5] > 1.0 and [4.5]–[24] > 1.5. These criteria – high α and red [4.5]–[24] and [3.6]–[4.5] colours – are powerful discriminants when searching for molecular outflow sources. However, we find no correlation between α and flow length or opening angle, and the outflows appear randomly orientated in each region. The more massive clouds are associated with a greater number of outflows, which suggests that the star formation efficiency is roughly the same in each region.     

Wind-swept clouds and possible triggered star formation associated with the supernova remnant G357.7+0.3

We present evidence for interaction between the supernova remnant (SNR) G357.7+0.3 and nearby molecular clouds, leading to the formation of wind-swept structures and bright emission rims. These features are not observed at visual wavelengths, but are clearly visible in mid-infrared mapping undertaken using the Spitzer Space Telescope . Analysis of one of these clouds, the bright cometary structure G357.46+0.60, suggests that it contains strong polycyclic aromatic hydrocarbon emission features in the 5.8 and 8.0 μm photometric bands, and that these are highly variable over relatively small spatial scales. The source is also associated with strong variations in electron density; a far-infrared continuum peak associated with dust temperatures of ∼30 K; and has previously been observed in the 1720 MHz maser transition of OH, known to be associated with SNR shock excitation of interstellar clouds. This source also appears to contain a young stellar object (YSO) within the bright rim structure, with a steeply rising spectrum between 1.25 and 24 μm. If the formation of this star has been triggered recently by the SNR, then YSO modelling suggests a stellar mass  ∼5–10 M_⊙  , and luminosity   L _YSO∼10²–2 × 10³ L_⊙  .
Finally, it is noted that a further, conical emission region appears to be associated with the Mira V1139 Sco, and it is suggested that this may represent the case of a Mira outflow interacting with a SNR. If this is the case, however, then the distance to the SNR must be ∼half of that determined from CS   J = 2–1  and 3–2 line radial velocities.     

The star-forming content of the W3 giant molecular cloud

We have surveyed a ∼0.9 square degree area of the W3 giant molecular cloud (GMC) and star-forming region in the 850-μm continuum, using the Submillimetre Common-User Bolometer Array on the James Clerk Maxwell Telescope. A complete sample of 316 dense clumps were detected with a mass range from around 13 to  2500 M_⊙  . Part of the W3 GMC is subject to an interaction with the H  ii region and fast stellar winds generated by the nearby W4 OB association. We find that the fraction of total gas mass in dense, 850-μm traced structures is significantly altered by this interaction, being around 5–13 per cent in the undisturbed cloud but ∼25–37 per cent in the feedback-affected region. The mass distribution in the detected clump sample depends somewhat on assumptions of dust temperature and is not a simple, single power law but contains significant structure at intermediate masses. This structure is likely to be due to crowding of sources near or below the spatial resolution of the observations. There is little evidence of any difference between the index of the high-mass end of the clump mass function in the compressed region and in the unaffected cloud. The consequences of these results are discussed in terms of current models of triggered star formation.     

A numerical simulation of a wind/molecular clump interaction

It has been pointed out in the past that it is impossible to accelerate molecular material to velocities ≥ 25 km s⁻¹ with gasdynamic shocks without dissociating the gas. Because of this, it has been argued that observations of molecular emission with radial velocities ∼ 20–100 km s⁻¹ imply the presence of 'C-shocks' (which have much lower post-shock temperatures, and therefore do not dissociate the gas) and the existence of strong (∼ 10–100 μG) magnetic fields.   In this paper, we discuss an alternative mechanism for accelerating molecular material to high velocities: a high-velocity, low-density wind drives a non-dissociative shock (with shock velocity v _cs ≤ 25 km s⁻¹) into a high-density, molecular clump. Once this shock wave has gone through the clump, the molecular material is moving at a velocity ∼  v _cs and has a gas pressure approximately equal to the ram pressure of the impinging wind. The compressed molecular clump can now be accelerated directly by the ram pressure of the wind (without the passage of further shocks through the molecular material), and will eventually move at the wind velocity.   This mechanism has been previously invoked to explain high-velocity molecular emission. However, numerical simulations have shown that a wind/clump interaction leads to the fragmentation of the clump before it can be accelerated to large velocities. In our numerical simulation (which includes an approximate treatment of the relevant microphysics) we find that the fragments that are produced are still largely molecular, and that they are rapidly accelerated to velocities comparable to the wind velocity. We therefore conclude that a wind/molecular clump interaction is indeed a valid mechanism for producing high-velocity molecular features.     

Chemical evolution ahead of Herbig–Haro objects

Compact regions of enhanced HCO⁺ and NH₃ emission have been detected close to a number of Herbig–Haro objects. An interpretation of these detections is the following: a transient clump within the molecular cloud has been irradiated by the shock that generates the Herbig–Haro object. The irradiation releases icy mantles from the grains within the transient clump and initiates a photochemistry. On the basis of this picture, we have developed an extensive chemical model which predicts that a wide range of species, other than NH₃ and HCO⁺, should also be detectable. These include CH₃OH, H₂S, C₃H₄, H₂CO, SO, SO₂, H₂CS and NS. The chemical effects should last ∼  10⁴ yr  .     

Shocked gas around Cepheus A: evidence for multiple outflows from H2S and SO2 observations

The Cepheus A star-forming region has been investigated through a multiline H₂S and SO₂ survey at millimetre wavelengths. Large-scale maps and high-resolution line profiles reveal the occurrence of several outflows. Cep A East is associated with multiple mass-loss processes: in particular, we detect a 0.6-pc jet-like structure which shows for the first time that the Cep A East young stellar objects are driving a collimated outflow moving towards the south.
The observed outflows show different clumps associated with definitely different H₂S/SO₂ integrated emission ratios, indicating that the gas chemistry in Cepheus A has been altered by the passage of shocks. H₂S appears to be more abundant than SO₂ in high-velocity clumps, in agreement with chemical models. However, we also find quite small H₂S linewidths, suggestive of regions where the evaporated H₂S molecules had enough time to slow down but not to freeze out on to dust grains. Finally, comparison between the line profiles indicates that the excitation conditions increase with the velocity, as expected for a propagation of collimated bow shocks.     

Stellar contents and star formation in the young star cluster Be 59

We present   UBV  I _c   CCD photometry of the young open cluster Be 59 with the aim to study the star formation scenario in the cluster. The radial extent of the cluster is found to be ∼10 arcmin (2.9 pc). The interstellar extinction in the cluster region varies between   E ( B − V ) ≃ 1.4  to 1.8 mag. The ratio of total-to-selective extinction in the cluster region is estimated as  3.7 ± 0.3  . The distance of the cluster is found to be  1.00 ± 0.05 kpc  . Using near-infrared (NIR) colours and slitless spectroscopy, we have identified young stellar objects (YSOs) in the open cluster Be 59 region. The ages of these YSOs range between <1 and ∼2 Myr, whereas the mean age of the massive stars in the cluster region is found to be ∼2 Myr. There is evidence for second-generation star formation outside the boundary of the cluster, which may be triggered by massive stars in the cluster. The slope of the initial mass function, Γ, in the mass range  2.5 < M /M_⊙≤ 28  is found to be  −1.01 ± 0.11  which is shallower than the Salpeter value (−1.35), whereas in the mass range  1.5 < M /M_⊙≤ 2.5  the slope is almost flat. The slope of the K -band luminosity function is estimated as  0.27 ± 0.02  , which is smaller than the average value (∼0.4) reported for young embedded clusters. Approximately 32 per cent of Hα emission stars of Be 59 exhibit NIR excess indicating that inner discs of the T Tauri star (TTS) population have not dissipated. The Midcourse Space Experiment (MSX) and IRAS-HIRES images around the cluster region are also used to study the emission from unidentified infrared bands and to estimate the spatial distribution of optical depth of warm and cold interstellar dust.     

Photodissociation regions and star formation in the Carina nebula

We have obtained wide-field thermal infrared (IR) images of the Carina nebula, using the SPIREX/Abu telescope at the South Pole. Emission from polycyclic aromatic hydrocarbons (PAHs) at 3.29 μm, a tracer of photodissociation regions (PDRs), reveals many interesting well-defined clumps and diffuse regions throughout the complex. Near-IR images  (1–2 μm)  , along with images from the Midcourse Space Experiment ( MSX ) satellite  (8–21 μm)  have been incorporated to study the interactions between the young stars and the surrounding molecular cloud in more detail. Two new PAH emission clumps have been identified in the Keyhole nebula, and have been mapped in  ¹²CO(2–1)  and  (1–0)  using the Swedish–ESO Submillimetre Telescope (SEST). Analysis of their physical properties reveals that they are dense molecular clumps, externally heated with PDRs on their surfaces and supported by external pressure in a similar manner to the other clumps in the region. A previously identified externally heated globule containing IRAS 10430−5931 in the southern molecular cloud shows strong 3.29-, 8- and 21-μm emission, the spectral energy distribution (SED) revealing the location of an ultracompact (UC) H  ii region. The northern part of the nebula is complicated, with PAH emission intermixed with mid-IR dust continuum emission. Several point sources are located here, and through a two-component blackbody fit to their SEDs we have identified three possible UC H  ii regions as well as a young star surrounded by a circumstellar disc. This implies that star formation in this region is ongoing and not halted by the intense radiation from the surrounding young massive stars.     

A silhouette envelope around GGD30IR detected by Spitzer

This paper presents a study of the envelope of the young stellar object (YSO) GGD30IR. What distinguishes this from most other YSOs is the elongated absorption feature seen in silhouette against the background emission in the Spitzer Galactic Legacy Infrared Midplane Survey Extraordinaire 8 μm Infrared Array Camera image of the region. The size and the symmetrical placement of GGD30IR in the centre of this feature suggest that it is an extended envelope, perhaps the remnant of the collapse of the GGD30 core. We have used the extinction in the envelope measured from (i) the reduction in the 8 μm background intensity and (ii) field star colour excesses, to estimate the envelope mass, obtaining values of 0.6 ± 0.2 and  0.5 ± 0.3 M_⊙  , respectively. To investigate the envelope further, we have obtained Australia Telescope Compact Array 3 mm continuum and HCO+ line observations of the region. The continuum emission at 3 mm arises from both a compact (unresolved; ≤730 au) core embedded in an extended envelope ∼18 000 au × 38 000 au in extent. We estimate the core mass to be 0.11 ± 0.02  M_⊙  . The HCO+ emission is extended in a direction perpendicular to the long axis of the envelope, suggesting it comes from an outflow. The spectral energy distribution (SED) provides a 2–24 μm spectral index, α= 1.0, which places GGD30IR in the Class I YSO category. Integrating the SED provides a luminosity of   L _*≃ 25 ± 5 L_⊙  .     

Discovery of a dense bipolar outflow from a new class 0 protostar in NGC 2068/LBS 17

We report the discovery of high-velocity dense gas from a bipolar outflow source near NGC 2068 in the L1630 giant molecular cloud. CO and HCO⁺ J =3→2 line wings have a bipolar distribution in the vicinity of LBS 17-H with the flow orientated roughly east–west and perpendicular to the elongation of the submillimetre dust continuum emission. The flow is compact (total extent ∼0.2 pc) and contains of the order of 0.1 M_⊙ of swept-up gas. The high-velocity HCO⁺ emission is distributed over a somewhat smaller area <0.1 pc in extent.
A map of C¹⁸O J =2→1 emission traces the LBS 17 core and follows the ambient HCO⁺ emission reasonably well, with the exception of the direction towards LBS 17-H where there is a significant anticorrelation between the C¹⁸O and HCO⁺. A comparison of beam-matched C¹⁸O and dust-derived H₂ column densities suggests that CO is depleted by up to a factor of ∼50 at this position if the temperature is as low as 9 K, although the difference is substantially reduced if the temperature is as high as 20 K. Chemical models of collapsing clouds can account for this discrepancy in terms of different rates of depletion on to dust grains for CO and HCO⁺.
LBS 17-H has a previously known water maser coincident with it but there are no known near-infrared, IRAS or radio continuum sources associated with this object, leading to the conclusion that it is probably very young. A greybody fit to the continuum data gives a luminosity of only 1.7 L_⊙ and a submillimetre-to-bolometric luminosity ratio of 0.1, comfortably satisfying the criteria for classification as a class 0 protostar candidate.     

A possible EGRET source with a two-sided radio jet structure on a parsec scale – 3EG J0808+5114

Recent observations have revealed that damped Lyα clouds (DLAs) host star formation activity. In order to examine if such star formation activity can be triggered by ionization fronts, we perform high-resolution hydrodynamics and radiative transfer simulations of the effect of radiative feedback from propagating ionization fronts on high-density clumps. We examine two sources of ultraviolet (UV) radiation field to which high-redshift ( z ∼ 3) galaxies could be exposed: one corresponding to the UV radiation originating from stars within the DLA, itself, and the other corresponding to the UV background radiation. We find that, for larger clouds, the propagating I-fronts created by local stellar sources can trigger cooling instability and collapse of significant part, up to 85 per cent, of the cloud, creating conditions for star formation in a time-scale of a few Myr. The passage of the I-front also triggers collapse of smaller clumps (with radii below ∼4 pc), but in these cases the resulting cold and dense gas does not reach conditions conducive to star formation. Assuming that 85 per cent of the gas initially in the clump is converted into stars, we obtain a star formation rate of  ∼0.25 M_⊙ yr⁻¹ kpc⁻²  . This is somewhat higher than the value derived from recent observations. On the other hand, the background UV radiation which has harder spectrum fails to trigger cooling and collapse. Instead, the hard photons which have long mean free-path heat the dense clumps, which as a result expand and essentially dissolve in the ambient medium. Therefore, the star formation activity in DLAs is strongly regulated by the radiative feedback, both from the external UV background and internal stellar sources and we predict quiescent evolution of DLAs (not starburst-like evolution).     

HST/STIS observations of the high-velocity interstellar cloud HVC 291.2−41.2+80: a warm, mainly ionized high-velocity cloud

We present intermediate-resolution HST /STIS spectra of a high-velocity interstellar cloud ( v _LSR=+80 km s⁻¹) towards DI 1388, a young star in the Magellanic Bridge located between the Small and Large Magellanic Clouds. The STIS data have a signal-to-noise ratio (S/N) of 20–45 and a spectral resolution of about 6.5 km s⁻¹ (FWHM). The high-velocity cloud absorption is observed in the lines of C  ii , O  i , Si  ii , Si  iii , Si  iv and S  iii . Limits can be placed on the amount of S  ii and Fe  ii absorption that is present. An analysis of the relative abundances derived from the observed species, particularly C  ii and O  i , suggests that this high-velocity gas is warm ( T _k∼10³–10⁴ K) and predominantly ionized. This hypothesis is supported by the presence of absorption produced by highly ionized species, such as Si  iv . This sightline also intercepts two other high-velocity clouds that produce weak absorption features at v _LSR=+113 and +130 km s⁻¹ in the STIS spectra.     

An analysis of infrared emission spectra from the regions near the Galactic Centre

We present consistent modelling of line and continuum infrared (IR) spectra in the region close to the Galactic Centre. The models account for the coupled effect of shocks and photoionization from an external source. The results show that the shock velocities range between ∼65 and 80 km s⁻¹ and the pre-shock densities between 1 cm⁻³ in the interstellar medium (ISM) to 200 cm⁻³ in the filamentary structures. The pre-shock magnetic field increases from 5 × 10⁻⁶ G in the surrounding ISM to ∼8 × 10⁻⁵ G in the arched filaments. The stellar temperatures are ∼38 000 K in the Quintuplet cluster and ∼27 000 K in the Arches Cluster. The ionization parameter is relatively low (<0.01) with the highest values near the clusters, reaching a maximum >0.01 near the Arches Cluster. Depletion from the gaseous phase of Si is found throughout the whole observed region, indicating the presence of silicate dust. Grains including iron are concentrated throughout the arched filaments. The modelling of the continuum spectral energy distribution in the IR range indicates that a component of dust at temperatures of ∼100–200 K is present in the central region of the Galaxy. Radio emission appears to be thermal bremsstrahlung in the E2–W1 filaments crossing strip; however, a synchrotron component is not excluded. More data are necessary to resolve these questions.     

Fragmentation in molecular clouds and its connection to the IMF

We present an analysis of star-forming gas cores in a smooth particle hydrodynamics simulation of a giant molecular cloud. We identify cores using their deep potential wells. This yields a smoother distribution with clearer boundaries than density. Additionally, this gives an indication of future collapse, as bound potential cores (p-cores) represent the earliest stages of fragmentation in molecular clouds. We find that the mass function of the p-cores resembles the stellar initial mass function and the observed clump mass function, although p-core masses  (∼0.7 M_⊙)  are smaller than typical density clumps. The bound p-cores are generally subsonic, have internal substructure and are only quasi-spherical. We see no evidence of massive bound cores supported by turbulence. We trace the evolution of the p-cores forward in time, and investigate the connection between the original p-core mass and the stellar mass that formed from it. We find that there is a poor correlation, with considerable scatter suggesting accretion on to the core is dependent on more factors than just the initial core mass. During the accretion process the p-cores accrete from beyond the region first bound, highlighting the importance of the core environment to its subsequent evolution.     

Survival and disruption of subsystems during a cold collapse

Cold collapse of a cluster composed of small identical clumps, each of which is in virial equilibrium, is considered. Since the clumps have no relative motion with respect to each other initially, the cluster collapses under its own gravity. At the first collapse of the cluster, most of the clumps are destroyed, but some survive. In order to find the condition for the clumps to survive, we made a systematic study in two-parameter space: the number of the clumps N _c and the size of the clump r _v . We obtained the condition N _c ≫ 1 and n _k  ≥ 1, where n _k is related to r _v and the initial radius of the cluster R _ini through the relation R _ini/ r _v  = 2 N ^{( n k +5)/6}_c. A simple analytical argument supports the numerical result. This n _k corresponds to the index of the power spectrum of the density fluctuation in the cosmological hierarchical clustering, and thus our result may suggest that in the systems smaller than 2/Ω h ²)Mpc, the first violent collapse is strong enough to sweep away all the substructures that exist before the collapse.     

The 'Carina Flare' supershell: probing the atomic and molecular ISM in a Galactic chimney

The 'Carina Flare' supershell, GSH 287+04−17, is a molecular supershell originally discovered in  ¹²CO( J = 1–0)  with the NANTEN 4 m telescope. We present the first study of the shell's atomic ISM, using H  i 21-cm line data from the Parkes 64-m telescope Southern Galactic Plane Survey. The data reveal a gently expanding,  ∼230 × 360  pc H  i supershell that shows strong evidence of Galactic Plane blowout, with a break in its main body at   z ∼ 280  pc and a capped high-latitude extension reaching   z ∼ 450  pc. The molecular clouds form comoving parts of the atomic shell, and the morphology of the two phases reflects the supershell's influence on the structure of the ISM. We also report the first discovery of an ionized component of the supershell, in the form of delicate, streamer-like filaments aligned with the proposed direction of blowout. The distance estimate to the shell is re-examined, and we find strong evidence to support the original suggestion that it is located in the Carina Arm at a distance of  2.6 ± 0.4 kpc  . Associated H  i and H₂ masses are estimated as   M _H I≈ 7 ± 3 × 10⁵ M_⊙  and     , and the kinetic energy of the expanding shell as   E _K ∼ 1 × 10⁵¹  erg. We examine the results of analytical and numerical models to estimate a required formation energy of several 10⁵¹ to  ∼10⁵²  erg, and an age of  ∼10⁷ yr  . This age is compatible with molecular cloud formation time-scales, and we briefly consider the viability of a supershell-triggered origin for the molecular component.     

Iron line profiles including emission from within the innermost stable orbit of a black hole accretion disc

We propose a model for the source of the X-ray background (XRB) in which low-luminosity active nuclei ( L  ∼ 10⁴³ erg s⁻¹) are obscured ( N  ∼ 10²³ cm⁻²) by nuclear starbursts within the inner ∼ 100 pc. The obscuring material covers most of the sky as seen from the central source, rather than being distributed in a toroidal structure, and hardens the averaged X-ray spectrum by photoelectric absorption. The gas is turbulent with velocity dispersion ∼ few × 100 km s⁻¹ and cloud–cloud collisions lead to copious star formation. Although supernovae tend to produce outflows, most of the gas is trapped in the gravity field of the star-forming cluster itself and the central black hole. A hot ( T  ∼ 10⁶ − 10⁷ K) virialized phase of this gas, comprising a few per cent of the total obscuring material, feeds the central engine of ∼ 10⁷ M⊙ through Bondi accretion, at a sub-Eddington rate appropriate for the luminosity of these objects. If starburst-obscured objects give rise to the residual XRB, then only 10 per cent of the accretion in active galaxies occurs close to the Eddington limit in unabsorbed objects.