EDP Sciences
Free Access
Issue
A&A
Volume 566, June 2014
Article Number A38
Number of page(s) 17
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/201323069
Published online 05 June 2014

© ESO, 2014

1. Introduction

The Lyα emission is one of the most useful tracers of massive star formation in the early Universe, since at high redshift it becomes visible in the optical-infrared range. Moreover, for redshifts above z ~ 4 Lyα becomes the only emission line easily accessible from ground-based telescopes. But the study of Lyα emission in starburst galaxies in the Local Universe has shown that radiation transfer through neutral hydrogen can completely distort its properties, affecting its profile and intensity, and even transforming the emission into a damped absorption when the column density of neutral gas is high enough (Mas-Hesse et al. 2003). We embarked on a program to study Lyα emitting galaxies in the Local Universe in detail (Östlin et al. 2009; Otí-Floranes et al. 2012; Hayes et al. 2013) to further investigate the properties of the Lyα emission and its correlation with other characteristics of its origin starbursts. The final aim of this program is the proper calibration of Lyα as a tracer of star formation, which could be used to derive the star formation intensity along the history of the Universe. In this work, we present a detailed multi-wavelength analysis of the Lyα emitting galaxy IRAS 08339+6517, complementing previous partial results by Mas-Hesse et al. (2003) and Östlin et al. (2009).

Table 1

IRAS 0833 coordinates, redshift, distance, scale, Galactic H I, color excess towards the source, and oxygen abundance.

The source IRAS 08339+6517 (hereinafter IRAS 0833) is a face-on spiral galaxy at 80.2 Mpc, which is undergoing a starburst episode and shows a prominent Lyα emission line. The brightest UV knots are aligned along an internal bar oriented close to the N − S direction (Östlin et al. 2009). López-Sánchez et al. (2006) derived an age of 4−6 Myr for the most recent burst, using Hα spectroscopic data. They also estimated that two older underlying stellar populations of 100−200 Myr and 1−2 Gyr exist. Using data of the Very Large Area (VLA) at the National Radio Astronomy Observatory (NRAO), Cannon et al. (2004) reported that IRAS 0833 harbours 1.1 × 109 M of neutral hydrogen. They also detected a tidal stream between IRAS 0833 and its minor companion 2MASX J08380769+6508579, estimating a mass of 3.8 × 109 M for the stream material. The Lyα emission of IRAS 0833 was first reported by Kunth et al. (1997) from the analysis of HST/Goddard High Resolution Spectrograph (GHRS) spectrum of its nuclear region. They found that the Lyα line shows a P Cyg profile, as well as the presence of a blue-shifted secondary Lyα emission component, on top of the spectral P Cyg trough. From the same data, Kunth et al. (1998) measured the flux of the line, as well as the column density of the neutral hydrogen, which produces the observed P Cyg absorption. Later, HST/Space Telescope Imaging Spectrograph (STIS) long-slit observations with high spectral resolution led Mas-Hesse et al. (2003) to confirm that Lyα photons can actually escape from IRAS 0833 because most of the neutral gas surrounding the source is being pushed out by the starburst activity. The velocity shift causes the red wing to eventually escape, whereas the blue photons of the line are actually resonantly scattered by H I, leading to the observed P Cyg profile. They measured the velocity of expansion of the superbubble deriving ~300 km s-1. They argued that the secondary Lyα emission could be produced at the leading front of the superbubble with no neutral hydrogen ahead, explaining why this minor component is detected blue-shifted by ~300 km s-1 on top of the absorption trough caused by the outflowing medium. Table 1 summarizes the basic properties of the starburst galaxy IRAS 0833.

thumbnail Fig. 1

Left: multispectral HST/ACS image of IRAS 0833. Blue: F300W, green: F435W, red: FR656N (Courtesy of Ángel R. López-Sánchez – AAO). Right: F140LP UV image. Integrated emissions of the starburst (STB) and central super stellar cluster (SSC) are labelled. Intensity scale is logarithmic. Same spatial scale in both images, as marked on the right panel. North is up and east is to the left.

Open with DEXTER

Table 2

Log of HST/GHRS and IUE observations of IRAS 0833 used in this work.

Table 3

Log of HST/STIS observations of IRAS 0833 used in this work.

In this work, we have analysed HST, XMM-Newton, IUE (International Ultraviolet Explorer), and ground-based observations aiming to characterise the properties of the starburst in IRAS 0833, and to check whether the conclusions from the study of Haro 2 (also known as Mrk 33 and IRAS 10293+5439) by Otí-Floranes et al. (2012) about the nature of the diffuse, extended Lyα emission, could be extended to another Lyα emitting starburst galaxy with very different morphology and global properties. We describe the observational data in Sect. 2, the results are presented in Sect. 3, and they are discussed in Sect. 4. Finally, conclusions are outlined in Sect. 5.

2. Observational data and synthesis models

2.1. The HST observations: ultraviolet and optical

We have used HST observations of IRAS 0833 obtained with Advanced Camera for Surveys (ACS), STIS, and GHRS, in the UV and optical. In this section, we describe the data and processing of the observations.

2.1.1. The HST images

The UV, optical, and Hα images used in this work were obtained and processed by Östlin et al. (2009) from HST/ACS observations. The UV image we use to study the distribution of the massive, young stellar clusters corresponds to the observation with the F140LP filter (PI: Kunth). On the other hand, the Hα image was created from the observation with the narrow filter FR656N, estimating the Hα continuum by fitting pixel by pixel UV-optical photometric data points (HST/ACS F140LP, F220W, F330W, F435W, and F550M observations) with SEDs obtained from Starburst99 synthesis models (Leitherer et al. 1999). The resulting continuum was subtracted from the FR656N image. For a more complete description of the methodology, see Östlin et al. (2009). We show in Fig. 1 (left panel) a multi-spectral image created by Ángel R. López-Sánchez (CSIRO/ATNF) combining the HST/ACS images obtained with the F300W, F435W, and FR656N filters. In the right panel, the regions we analyse in the paper are labelled on top of the F140LP image: 1) the spatially integrated emission of the whole starburst (STB), which comprises the total radiation of IRAS 0833; and 2) the central supermassive stellar cluster (SSC).

2.1.2. The HST spectroscopy

We have re-analysed in this work the low-resolution GHRS spectroscopic data of IRAS 0833 described by González Delgado et al. (1998) as well as the high-resolution, long-slit HST/STIS observations obtained by Mas-Hesse et al. (2003). Previous high-resolution HST/GHRS observations by Kunth et al. (1998) were not used, since the HST/STIS data provided a broader wavelength coverage and higher spatial resolution along the long slit. The GHRS observation was performed with a squared aperture 1.7″ × 1.7″ centred on the maximum of the UV emission of IRAS 0833, i.e., its nucleus. Given its limited size, the aperture covered only the region around the central SSC. The spectral coverage of the spectrum was 1100 − 1900 Å, which thus includes the Lyα line as well as the stellar absorption lines Si IV λλ1394, 1403 Å, and C IV λλ1548, 1551 Å. The log of the HST/GHRS observations is shown in Table 2.

thumbnail Fig. 2

F140LP UV image. HST/GHRS (squared) and IUE (oval) apertures are superimposed, together with the location of the STIS slit. The width of the apertures and the slit are represented at real scale. The location of the region LyαS discussed in Sect. 4.3 is marked with a cross within a circle. Intensity scale is logarithmic. North is up and east is to the left.

Open with DEXTER

The G140M and G430L gratings were used in the STIS observations of the source. A 52″ × 0.5″ slit was placed onto the nucleus of IRAS 0833 with a position angle of 171°. High spectral resolution data of the Lyα line and the UV continuum in the spectral range 1200 − 1250 Å were obtained with G140M. The G430L spectral image includes the emission of the nebular lines O [II]3727, Hβ, and O [III]4959, 5007. However, the integration time for G430L was very low, and the observation was severely affected by cosmic rays. The extraction and background removal processes in the STIS spectral images were carried out using the Image Reduction and Analysis Facility (IRAF)1, following the procedure described by Otí-Floranes et al. (2012) for the STIS observation along the minor axis of the starburst galaxy Haro 2. See Mas-Hesse et al. (2003) for a more detailed description. The observation journal is included in Table 3. Figure 2 illustrates the position of the GHRS aperture and the STIS slit onto the ACS F140LP UV image of IRAS 0833.

Table 4

Log of WHT/ISIS observations of IRAS 0833 used in this work.

2.2. The IUE observations

The IUE/Short Wavelength Prime (SWP) and Long Wavelength Prime (LWP) spectra of IRAS 0833 were downloaded from the IUE Newly Extracted Spectra (INES) archive at the Centre of Astrobiology (CAB, Madrid2). They were obtained with the IUE large aperture (20″ × 10″), which enclosed most of the UV emission of the source. The spectral coverage of the cameras are 1150 − 1980 Å (SWP) and 1850 − 3350 Å (LWP). In Fig. 2, the position of its large aperture relative to the ACS F140LP UV image of IRAS 0833 is shown. The log of the IUE observation is shown in Table 2.

2.3. The XMM-Newton: X-rays

The X-ray Multi-mirror Mission (XMM-Newton) observation 0111400101 of π1 Ursae Majoris (PI: Albert Brinkman) included IRAS 0833 in its field of view ~8′ away from the optical axis. Data from XMM-Newton analysed in this work were obtained by the European Photon Imaging Camera (EPIC), which consists of 2 metal oxide semi-conductor (MOS) and 1 pn CCD arrays. Each detector is located in a different X-ray telescope on-board XMM-Newton, and whereas only around half of the radiation entering the corresponding telescopes reaches the MOS cameras, the pn camera receives an unobstructed beam. The EPIC observation consisted in imaging of the source with full frame (MOS1) and small window (MOS2), as well as one timing observation (pn). We focused on the MOS1 and MOS2 imaging observations since we are not interested in the timing analysis of the source. Total exposure time of the observation was 50.1 ks (MOS1) and 45.2 ks (MOS2). Reduction of the data was performed with Scientific Analysis Software 13.0.0 (SAS), following the standard threads available on the website of the XMM-Newton-SAS. Only the spectral range 0.3−10.0 keV was considered, as well as events with patterns3PATTERN4 to maximize the signal-to-noise ratio (S/N) against non-X-rays events. Source regions in the EPIC images were calculated with SAS task eregionanalyse to increase the S/N value over the background, resulting in circular regions with a radius of 31″ (MOS1) and 28″ (MOS2). Re-distribution matrices and ancillary files were calculated for the MOS1 and MOS2 images, assuming the source regions for each image from which spectrum was extracted. Figure 3 shows the MOS1 and MOS2 images with the circular source regions assumed and the UV countours from the HST/ACS F140LP image. The source in MOS2 happened to be located close to a CCD border, and thus a fraction of the total X-ray emission may have been missed in this image. For consistency, when modelling the total spectrum, we checked that similar results were obtained for both MOS spectra. Background was extracted from two circular regions with radius~1′ with different locations in each of the two images but which would be close to IRAS 0833 and would not include any bright source. A net count rate of ~1.5 × 10-2 cts s-1 was measured in the source regions after background subtraction. We used XSPEC v.12.8 (Arnaud 1996) to analyse the spectral data. Source counts were grouped to have at least 25 counts per bin in order to be able to apply χ2 statistics in the fitting analysis. The final binned X-ray spectrum of IRAS 0833 is shown in Fig. 11. Since this is an off-axis observation, the point spread function (PSF) of the instrument is large and distorted. We compared the radial profiles of the source with the EPIC PSF using SAS task eradial at low, medium, and high energy, and concluded that IRAS 0833 is not resolved and can be treated as a point-like source. Therefore, no information on the structure of the X-ray-emitting regions is available in Fig. 3, except for the spectral properties of their integrated emission.

thumbnail Fig. 3

XMM-Newton/EPIC MOS1 (left) and MOS2 (right) X-ray images of IRAS 0833. Intensity scale is logarithmic. Source region from which X-ray spectrum was extracted for each camera is confined within the white circle. Contours from HST/ACS F144LP image are shown in black. The PSF of this off-axis observation is large and asymmetric, and thus the emission must be considered as non-resolved. North is up and east is to the left.

Open with DEXTER

thumbnail Fig. 4

HST/ACS images of IRAS 0833 in the ultraviolet (filter F140LP, green, top left), continuum-subtracted Hα (red, top right), the composite of both (bottom left), and the composite including Lyα emission in blue (bottom right, adapted from Östlin et al. 2009). The aperture of HST/GHRS is marked with a square. The circular region R1 (radius ~0.5 (~200 pc)), as discussed in the text, is also marked. Intensity scale is linear, except for the last image, in which intensity scale is logarithmic. North is up and east is to the left.

Open with DEXTER

2.4. Ground-based observations

Two long-slit observations with the blue and red arms of William Herschel Telescope/Intermediate dispersion Spectrograph and Imaging System (WHT/ISIS) were downloaded from the archive (PI: Ángel R. López-Sánchez). The location and position angle of the 1.0″-wide ISIS slit coincided with that of the HST/STIS slit. The spatial profiles of the Hα and Hβ emission lines were extracted from the ISIS spectral images. The log of these observations is included in Table 4.

2.5. Synthesis models

Evolutionary population synthesis models by Leitherer et al. (1999, Starburst 99, herineafter SB99) and Cerviño et al. (2002, hereinafter CMHK02) have been used to characterize the star-formation episode taking place in IRAS 0833. Both sets of models compute the evolution of a cluster of massive stars for a given initial mass function, metallicity, and star formation regime. Once the structure (number of stars of each spectral type and luminosity class at a given evolutionary time) of the stellar population is derived, the models are used to calculate a set of observable parameters (continuum emission, number of ionizing photons produced, supernova rate, equivalent width (EW), and intensity of nebular lines, etc.). Two different star-formation regimes are considered in these models: 1) stars are produced at a constant rate (extended burst, EB); and 2) prompt formation occurs during a short period of time, after which no more stars are produced (instantaneous burst, IB). For consistency, we assume in both cases the same Salpeter initial mass function (IMF, φ(m) ~ m-2.35) with mass limits of 2−120 M. While the lower mass stars do not produce any observable effect in the UV range, they provide the bulk of the stellar mass and can even dominate the optical-IR continuum after some tens of Myr of evolution. The output magnitudes of the models are scaled by the intensity of star formation, which is characterized by the star-formation rate (SFR, the velocity at which stars are being produced) in EB models, and star formation strength (SFS, total mass converted into stars) in IB models. See the cited references for a more detailed description of the models.

3. Analysis and results

3.1. The HST morphology and photometry

The F140LP UV image in Fig. 1 shows that IRAS 0833 is composed of a large, bright central stellar cluster along an internal bar almost aligned from N to S, surrounded by a rather diffuse UV emission from several spiral arms which enclose smaller, unresolved stellar knots. Whereas the elongated central cluster is comprised within 1″ × 2″ (~400 pc × 800 pc), the spiral arms extend over a circular region of radius ~5.5″ (~2 kpc). As shown in the right panel of Fig. 1, hereinafter we refer to the global starburst in IRAS 0833 as STB, and we characterize it by analyzing the integrated emission of the whole source. On the other hand, we have labelled its central UV-bright stellar cluster as SSC, whose emission is roughly included within the GHRS aperture.

thumbnail Fig. 5

HST/GHRS spectrum of SSC in IRAS 0833 (solid line) fitted with an SB99 IB model for an age of ~5.5 Myr (dashed line). Left: normalized spectrum with labelled absorption lines Si IV λλ1394, 1403 Å, and C IV λλ1548, 1551 Å. Right: regular spectrum reddened to account for Galactic (Cardelli extinction law, E(BV) = 0.094) and intrinsic (LMC law, E(BV) = 0.01) extinctions.

Open with DEXTER

thumbnail Fig. 6

Left: IUE spectrum of STB in IRAS 0833 (black, solid line) fitted with an SB99 spectrum for an age of 5.5 Myr (dashed lines). The model has been reddened to account for the Galactic extinction, as well as for the intrinsic extinction assuming both an LMC law and E(BV) = 0.15 (red), and the Cardelli extinction law with E(BV) = 0.20 (green). Right: IUE spectrum is shown as in the left panel (black, solid line). A composite model (blue, dashed line) was obtained summing the starburst model (red) and an SB99 spectrum of a 1-Gyr-old population reddened similarly as the young burst model (magenta, dashed line). Near-UV and optical photometric fluxes from Östlin et al. (2009) are marked with green crosses.

Open with DEXTER

We measured the total UV flux of the STB region (F1529) at 1500 Å rest frame in two ways: 1) through the integrated emission of HST/ACS F140LP image; and 2) from the flux at 1500(1 + z) Å in the IUE spectrum. Although very large, the IUE aperture might have not included the whole emission from the source. To check whether this was the case, we calculated both the total integrated flux from the HST/ACS F140LP image (FUV integrated), as well as the flux enclosed in the same image by an aperture similar to the IUE one positioned with the same angle (FUV IUE), as shown in Fig. 2. We derived a ratio FUV integrated/FUV IUE = 1.04, indicating that the IUE aperture encloses 96% of the UV emission of IRAS 0833. After correcting for this aperture loss, the value of the IUE UV flux remains 15% lower than the integrated HST/ACS F140LP flux. The disagreement is most likely a combination of the 10% photometric accuracy of IUE (see Fig. 7) and the very broad, skewed shape of the F140LP filter. We, therefore, decided to take the average of both flux measurements as the total UV flux at 1500 Å of STB, F1529= 4.5 × 10-14 erg s-1 cm-2 Å-1. This uncertainty in the measurement of F1529 translates into a systematic uncertainty of ± 10% in the determination of any absolute magnitude of STB based on it, like the total mass of the starburst or the predicted luminosities at different ranges.

Given the size of its aperture, the spectrum obtained by HST/GHRS includes mostly the emission from the central cluster, which we have labelled SSC. The UV flux value found for this region F1529= 8.5 × 10-15 erg s-1 cm-2 Å-1 is 20% of STB, i.e., the total integrated observed UV flux of IRAS 0833.

We show in Fig. 4 a color-coded image of IRAS 0833 combining F140LP UV (green) and Hα (red), along with the single images of each filter, as generated by Östlin et al. (2009), and merged together using ds9 (Joye & Mandel 2003). While in the Hα image a filamentary structure extending largely around the stellar clusters in the spiral arms appears very prominent, the region dominated by the UV-bright SSC shows a very low Hα emission, if any.

3.2. The UV stellar continuum

We fitted the profile of the spectral lines Si IV λλ1394, 1403 Å, and C IV λλ1548, 1551 Å in the normalized HST/GHRS (SSC) and IUE (STB) spectra using high-resolution UV spectra generated with the SB99 models to estimate the evolutionary state of the star formation episode in IRAS 0833. As shown in Fig. 2, and as already discussed above, whereas the HST/GHRS spectrum only covers the central stellar cluster of IRAS 0833, the low-resolution IUE spectrum contains most of the integrated UV emission of IRAS 0833 (STB). Both the Milky Way and the LMC/SMC libraries were used in the SB99 models, assuming a Salpeter IMF (φ(m) ~ m-2.35) with mass limits of 2 − 120 M and Z = 0.008 for the evolutionary tracks and the high-resolution spectra. The metallicity value considered was the closest available to the observed one (López-Sánchez et al. 2006). We started by fitting the GHRS spectrum of the SSC, given that its higher spectral resolution allows for a more detailed analysis than the IUE spectrum. We checked that no acceptable fitting was possible for the C IV lines with the Milky Way library. The model that more accurately reproduces the spectral profile of the lines is an instantaneous burst (IB) model of age ~5.5 Myr obtained with the LMC/SMC library, in agreement with the age range 4 − 6 Myr found by López-Sánchez et al. (2006). Furthermore, these authors detected the presence of WR stars within an aperture similar to the HST/GHRS one, with a ratio WR/(WR + O) ~ 0.03. According to evolutionary synthesis models (Schaerer & Vacca 1998; Cerviño & Mas-Hesse 1994), the presence of WR stars at the metallicty of IRAS 0833 is only possible if the prompt star forming event is younger than around 6 Myr. The measured WR/(WR + O) ratio is consistent with the predictions for a starburst at 5 − 5.5 Myr (Schaerer & Vacca 1998). Even more, the WR/(WR + O) value is incompatible with an extended starburst (EB) which had been producing stars for more than 10 Myr, since then it would show much lower values (WR/(WR + O) < 0.009Cerviño & Mas-Hesse 1994). Both the observational HST/GHRS and the SB99 normalized spectra around the Si IV λλ1394, 1403 Å and C IV λλ1548, 1551 Å lines are shown in Fig. 5 (left panel).

Finally, we checked that the normalized IUE spectrum of STB is consistent with the age obtained from the GHRS data of the SSC region, as far as the lower IUE spectral resolution allows. Though we can not exclude some age spread of the star-forming knots along the spiral arms, we conclude that the global star formation episode in IRAS 0833 is compatible with a short-lived burst which started ~5.5 Myr ago over the whole galaxy.

3.3. Extinction

We derived the average reddening of the integrated UV emission of IRAS 0833 (STB) by fitting the IUE spectrum with the SB99 model computed above. The fit was performed by applying a certain intrinsic E(BV) value to the predicted continuum. The spectrum was then redshifted (z = 0.0191) and finally reddened again, assuming E(BV)= 0.094 and Cardelli extinction law (Cardelli et al. 1989) to account for the Galactic extinction. We repeated the procedure for different intrinsic E(BV) values until we satisfactorily reproduced the observed UV slope and shape. As can be seen in Fig. 6, the IUE spectrum of IRAS 0833 shows a prominent 2175 Å feature, typically associated in the past to graphite in the dust grains, but whose origin might rather be in hydrogenated amorphous carbon materials (Gadallah et al. 2011). The SMC law (Prévot et al. 1984) was not considered when modelling the intrinsic extinction since it does not reproduce the 2175 Å bump, unlike the case of the LMC (Fitzpatrick 1985) and Galactic laws (Cardelli et al. 1989). We assumed RV = 3.1 for these laws. Once the reddening correction was applied, we computed the mass of the integrated source by dividing the observed UV flux at ~1529 Å (~1500 Å rest frame) by the reddened, redshifted value predicted by the SB99 model. The best fit yielded E(BV)= 0.15, assuming an LMC-like law and M ~ 1.4 × 108 M for the initial mass of stars (in the 2120 M range) having formed all over IRAS 0833 around 56 Myr ago. This value of the extinction is in agreement with González Delgado et al. (1998) who obtained E(BV)= 0.17 with an LMC law when fitting the integrated continuum emission observed with Hopkins Ultraviolet Telescope (HUT). When assuming the Cardelli Galactic law for the internal extinction, either the 2175 Å feature or the UV slope at λ< 1800 Å was reproduced, but not both simultaneously. As shown in Fig. 6 (left panel), the best fit is obtained using the LMC law for the internal extinction, although the 2175 Å is not completely reproduced. If the Cardelli law is assumed, E(BV) = 0.20 would be required, but the fits neither improve, nor reproduce the slope of the UV continuum correctly. Leitherer et al. (2013) measured E(BV) = 0.25 when fitting the spectrum of IRAS 0833 obtained with HST/COS (circular aperture with radius =1.25), which only includes the spectral range 1100 − 1450 Å. This fit is significantly worse than that achieved with the LMC law and E(BV) = 0.15 when the whole UV range is considered. We want to stress that, contrary to the optical range, different E(BV) values are required to get the same extinction factor in the UV for different laws.

thumbnail Fig. 7

Observed HST/GHRS (black, solid line, SSC) and IUE (blue, solid, STB; errors in green) spectra with the SB99 model for an age of 5.5 Myr (red, dashed). Vertical scale corresponds to the GHRS spectrum. The other spectra have been arbitrarily scaled for an easier comparison. SB99 model has been corrected only for Galactic extinction.

Open with DEXTER

The reddening affecting the UV continuum of the central stellar cluster in IRAS 0833 (SSC) was also computed by fitting the HST/GHRS spectrum with the predicted SB99 SED, obtaining a very low extinction with E(BV)= 0.01. We assumed an LMC-like extinction law, as derived from the IUE spectrum. The observed UV continuum, together with the best fitting, reddened SB99 model, have been plotted in Fig. 5 (right panel). Assuming this extinction, the UV continuum emitted by the SSC is compatible with a burst at 5.5 Myr with an initial mass of M ~ 7.0 × 106 M, within the assumed mass limits of 2−120 M.

Table 5

Predicted values of the age, mass, stellar and nebular extinctions, fraction of ionizing photons absorbed by dust, emission fluxes and luminosities and EW(Hβ) of the starburst (STB) in IRAS 0833, assuming the UV fluxes from HST/ACS and IUE and the average value of both.

Table 6

Observed values of the UV and Hα emissions and the stellar and nebular extinctions of the integrated source IRAS 0833 (STB) and its central stellar cluster (SSC) with the predicted stellar mass for each of them.

The values for the best fits are given in Tables 5 and 6, including the reddening and mass values for both the STB and SSC regions. Also, the ratios of the UV and Hα emissions and the masses of SSC over STB, i.e., the central region over the integrated IRAS 0833, are included. The compact nature and the mass value of the central stellar cluster in IRAS 0833 indicates that it must be considered an SSC (Adamo et al. 2010), hence its name SSC. As shown in Table 6, whereas the SSC contributes up to 20% to the total observed UV flux, its intrinsic UV emission accounts for only 5% of the integrated unreddened value. This means that around 5% of the mass within STB is associated with the SSC. Finally, in Fig. 7 we show the HST/GHRS spectrum with the IUE one. Also, the predicted SB99 model (age = 5.5 Myr, Z = 0.008) is included, which has been reddened only by the Galactic extinction (E(BV)= 0.094). As can be seen, the stellar continuum from the SSC is hardly affected by dust extinction, showing a slope similar to the unreddened SB99 model, unlike the STB continuum, which actually shows a positive UV slope below 1500 Å.

3.3.1. Differential extinction

Using Nordic Optical Telescope/AndaLucia Faint Object Spectrograph and Camera (NOT/ALFOSC), López-Sánchez et al. (2006) obtained an Hα image of IRAS 0833, as well as a spectral image using ALFOSC in spectrographic mode with a 6.4′ × 1″ slit placed on the centre of the source and with a position angle PA = 138°. Since the spectral range extended from 3200 to 6800 Å, López-Sánchez et al. (2006) could calculate the Balmer decrement and hence the nebular dust extinction. They obtained an integrated value of C(Hβ)= 0.22 for the nebular extinction (see Mazzarella & Boroson 1993, for the definition of C(Hβ)), which corresponds to Hα/Hβ= 3.4 including also the Galactic contribution. This translates into an internal nebular extinction of E(BV)= 0.06. On the other hand, these authors found that the nebular extinction is largest within the central region ~1″ × 1″ (~400 pc ×400 pc), with C(Hβ)= 0.30 (corresponding to E(BV)~ 0.11 after correction from Galactic extinction).

The effect is reversed when we look at the extinction affecting the stellar continuum. As shown in Fig. 7, while the integrated stellar UV continuum from STB is severely affected by dust, with a global E(BV)~ 0.15 (corrected from Galactic extinction), the stellar continuum from SSC shows a very low, almost negligible extinction (E(BV)~ 0.01).

Furthermore, Fig. 4 shows that the Hα emission is concentrated in circumnuclear blobs around the central cluster and in the spiral arms, but is very weak just on top of it. The origin of these apparent discrepancies is related to the cavity apparently devoid of gas and dust which surrounds the central SSC. The mechanical energy released by the SSC in the form of stellar winds and supernova explosions has efficiently swept out the gas and dust, leaving a rather clean region just around the central SSC. This explains the low reddening affecting the UV continuum and the lack of Hα emission in the close vicinity of the SSC. These stellar winds have apparently concentrated the dust and gas at larger distances (but still close to the SSC), so that the Hα emission integrated over this central region appears severely obscured. Outside this central region, the distribution of dust is smoother. In Fig. 8, the spatial profile of Hα/Hβ as derived from the WHT/ISIS observations is shown to be quite uniform over large areas, with just a weak gradient from north to south. This scenario exemplifies the spatial decoupling between the extinction affecting the stellar continuum and that affecting the nebular lines, which is very frequent in starburst galaxies and can yield misleading conclusions when analysed globally (see further discussion below in Sect. 4.1).

3.4. Far-infrared emission

Once the starburst in IRAS 0833 had been characterized and its main properties constrained, we checked whether the FIR and Hα luminosities from STB could also be reproduced. For this task we used the evolutionary synthesis models by CMHK02 models, which allow us to compute the expected FIR flux for a given burst and a specific extinction. López-Sánchez et al. (2006) reported an integrated Hα luminosity L(Hα)= 1.2 × 1042 erg s-1 already corrected for N [II] contamination and internal and Galactic extinctions. Assuming the starburst model described above, the expected intrinsic Hα emission predicted by CMHK02 models for STB in IRAS 0833 is L(Hα)= 1.8 × 1042 erg s-1, which is around 50% higher than the observed value.

This disagreement might be due to the value we have assumed for the fraction f of ionizing photons emitted by the massive stars of STB which are considered to contribute to ionization. Hitherto we have assumed when using the predictions by the CMHK02 models that 30% (1 − f= 0.3) of ionizing photons are absorbed by dust before they ionize any hydrogen atom. Mathis (1971) and Petrosian et al. (1972) obtained analytical expressions of f as a function of the optical depth τ. Assuming the latter expression, Degioia-Eastwood (1992) obtained values 1 − f ~ 0.3 for regions in LMC, similar to those found by Mezger (1978) in the Galaxy. According to these results, CMHK02 models assume as an average value (1 − f) = 0.3 when calculating the intensities of the nebular lines from the predicted number of ionizing photons produced by the massive stars of the burst. However, values measured for τ have uncertainties and might have a spread, which translates into a spread in the parameter 1 − f. Actually, when assuming 1 − f = 0.5, CMHK02 models predict a value for Hα luminosity in IRAS 0833 in agreement with the value reported by López-Sánchez et al. (2006). This value of 1 − f is realistic, and indeed similar to the fraction measured in several Galactic H II regions (Mezger 1978), and consistent with the rather high value of extinction measured in the STB spectrum.

We calculated the infrared emission FIR(40−120 μm) using the relation by Helou et al. (1985) based on the IRAS fluxes F60 μm = 5.81 Jy and F100 μm = 6.48 Jy. Following Calzetti et al. (2000), we converted the value of FIR(40−120 μm) to the estimated total far-infrared emission with the relation FIR(1−1000 μm)/FIR(40−120 μm) ~ 1.75. The resulting value is LFIR= 3.6 × 1044 erg s-1. Assuming E(BV) = 0.15 with the LMC extinction law and 1 − f = 0.5, the CMHK02 models predict LFIR= 4.2 × 1044 erg s-1 for STB. Though the predicted LFIR value is higher than the observed value by 16%, both values are consistent considering the inherent uncertainties of the method and the systematic ± 10% uncertainty in the predicted luminosities discussed above. All the predicted and observed values are shown in Table 5.

The rather good agreement between the predicted and observed values of L(Hα) and LFIR, with a general trend to some overestimation by the models, indicates that the UV continuum (used for normalization) is contributed by the complete population of young stars in the central region of IRAS 0833. Therefore, there is no evidence of additional stellar knots potentially hidden within dense dust clouds, which would be invisible in the UV, but would contribute to L(Hα) and LFIR, as found in other similar galaxies like Haro 2 (Otí-Floranes et al. 2012).

3.5. Underlying stellar population

Leitherer et al. (2013) considered that the HST/COS aperture was large enough (radius = 1.25) to contain short-lived massive bursts at different ages, mimicking a quasi-continuous star formation rate, and adopted a characteristic age of 20 Myr. The detection by Pellerin & Robert (1999) of the Ca II λλ8498, 8542, 8662 Å and Mg2λ5177 Å absorption lines in an IR-optical spectrum of the source, which are signatures of red (super)giants (RSGs) and late-type stars, respectively, supported this assumption in principle, since the presence of both O stars and RSGs requires starbursts extended over at least ~10 Myr. Nevertheless, Pellerin & Robert (1999) attributed the strength of the Ca II triplet and Mg2 absorption lines to an old stellar population of 1 Gyr, rather than to young RSGs. As discussed by López-Sánchez et al. (2006), it is known that IRAS 0833 has experienced previous star-formation episodes, with at least two older stellar populations with ages 100 − 200 Myr and 1−2 Gyr, respectively. Mas-Hesse et al. (2003) showed in their Fig. 11 that the optical continuum peaks ~1.2″ (~470 pc) north of a high-intensity emitting ionized gas within the STIS slit. This region (which they labelled “N”) presents a strong Balmer discontinuity and a lack of ionizing power, and is consequently dominated by older stars. We argue that, as we discuss below, the stars in this and other similar regions are dominating the features detected by Pellerin & Robert (1999), but do not have any relation to the young starburst in the core of IRAS 0833.

Moreover, when we tried to reproduce the different observables assuming synthesis models for extended bursts with the parameters considered by Leitherer et al. (2013), the Si IV and C IV spectral profiles could not be simultaneously fitted, and the stellar mass value which would account for the UV emission from STB overestimated the predicted L(Hα) and LFIR by factors 7 and 3, respectively. As discussed above, the measured WR/(WR + O) ratio is also not consistent with an extended starburst. Therefore, our results reject the hypothesis of a star formation episode extended over several tens of Myr in IRAS 0833. We conclude that the Balmer lines emission, the UV continuum and the far-infrared luminosity are dominated by a starburst, which converted ~1.4 × 108 M of gas into stars ~5.5 Myr ago. As already explained, our data do not allow us to constrain the age of the young, massive stars contributing to the UV continuum outside the central bar SSC, but the good fit with the lower resolution global IUE spectrum indicates that the spread in ages has to be very small. In addition, an underlying, older stellar population which dominates the optical to infrared continuum must have been formed in previous star formation episodes distributed over the last 1 Gyr.

In order to check this assumption, we plotted in Fig. 6 (right panel) the UV-optical SED of IRAS 0833. The plot includes the IUE spectrum already fitted in the left panel and discussed in the text, as well as the photometric fluxes obtained by Östlin et al. (2009) from HST/ACS observations. We noticed that the same model described in Table 5 and which can reproduce the UV slope and the UV, Hα, and FIR emissions from STB, was unable to account for the SED at λ ≳ 4000 Å. Since there is evidence of the presence of an older stellar population, as discussed above, we included a 1-Gyr-old SB99 model with same metallicity and mass limits as for the young starburst. This spectrum was reddened by the same amount as the IUE continuum to account for internal (E(BV) = 0.15, LMC law) and Galactic (E(BV) = 0.094, Cardelli law) extinctions. As shown in Fig. 6, whereas the young starburst population dominates the UV emission, the composite of the old and young populations properly reproduces the optical fluxes reported by Östlin et al. (2009) when Mold = 6 × 1010 M.

We have considered only a test case assuming a model for an age of 1 Gyr. The results do not vary significantly for other configurations, so we can conclude that a stellar population some hundreds of Myr old contributes significantly to the optical-IR continuum observed in IRAS 0833. Indeed, Fig. 6 shows that the continuum at λ ~ 8500 Å is dominated by the older stars, which therefore must account for the Ca II triplet absorption lines, as suggested by Pellerin & Robert (1999). Finally, when considering the contribution by the old stellar population, CMHK02 models predict EW(Hβ)~14 Å for the integrated starburst in IRAS 0833. Bearing in mind the uncertainties in the actual contribution of the underlying stellar population to the stellar continuum, as well as the effect of extinction in its optical emission, this EW(Hβ) estimate is in agreement with the value EW(Hβ) =19 Å reported by López-Sánchez et al. (2006). The EW(Hβ) calculated from the models, as well as the observed value, are included in Table 5.

In brief, the UV, Hα, and FIR emissions in IRAS 0833 are dominated by the young stars produced in a starburst that occurred ~5.5 Myr ago, which took place all along the central region of the galaxy, as well as in the surrounding spiral arms. On the other hand, the optical–IR continuum, as well as other features such as the Ca II triplet, are mostly contributed by old stars formed some hundred millions years before the current starburst took place.

thumbnail Fig. 8

Top: HST/STIS G140M spectral image of IRAS 0833. Spatial extension is ~ 19″ (~7.5 kpc) . Middle: emission profile of the main component of Lyα (blue) and UV continuum (1224.5 Å rest frame, black) from STIS data, and Hα (dashed red) from ACS data. Bottom: spatial profiles of Lyα (blue) and observed Lyα/Hα (solid black) and Hα/Hβ (from WHT/ISIS data, in red), and the expected Lyα/Hα ratio (dashed brown), assuming Case B recombination and internal reddening as derived from the Hα/Hβ ratio, as well as the Galactic extinction. The Lyα escape fraction , as defined in the text, has been plotted in dotted black line. In the middle and bottom panels, the negative x-axis corresponds to north, as indicated by N. To ease the comparison, the dashed red line in the middle panel shows the Hα profile scaled by a factor 0.2. The vertical scale corresponds to the Lyα and non-scaled Hα fluxes. Whereas the vertical scale of the left axis in the bottom panel corresponds to the Lyα/Hα and Hα/Hβ ratios, the scale of the right axis corresponds to . Position of the region LyαS is marked.

Open with DEXTER

3.6. The Lyα emission

A high-resolution spectrum around Lyα of the central region of IRAS 0833 is shown in Fig. 10. This 1D spectrum was extracted by Mas-Hesse et al. (2003) from the HST/STIS G140M spectral image, which is shown in Fig. 8 (top panel). The Lyα emission is very prominent, extending well beyond the stellar continuum to the north direction, whereas the emission towards the south is much weaker. Unlike what Otí-Floranes et al. (2012) found on the major axis of Haro 2, regions of total Lyα absorption are not observed along the slit. The Lyα spatial profile in Fig. 8 (middle panel) shows two compact components at both sides of the central cluster: north peak (position ~−0.7″ ~ −300 pc, width ~1″ ~ 400 pc, F ~ 4 × 10-16 erg s-1 cm-2) , and south peak (position ~1″ ~ 400 pc, width ~ 1.5″ ~ 600 pc, F ~ 6 × 10-16 erg s-1 cm-2). In the region between them there is an abrupt decline, which is co-spatial to the stellar cluster. Also, a somewhat diffuse emission is observed extending to the north (position <−1″ ~ 400 pc) over ~6″ (~2.3 kpc), and with a lower flux (F< 2 × 10-16 erg s-1 cm-2). On the other hand, the emission extending south from the southern peak is much weaker. The total spatial extension of the Lyα emission is ~16″ (~6 kpc), whereas the stellar continuum, dominated by the central stellar cluster, extends over ~5″ (~1.9 kpc). The observed, integrated flux of Lyα along the STIS slit is F(Lyα) ~ 6.4 × 10-14 erg s-1 cm-2 (not corrected for any kind of absorption or extinction), with the diffuse emission contributing up to ~25% of the total one.

Figures 8 and 10 show also a deep, broad Lyα absorption bluewards of the emission. Mas-Hesse et al. (2003) concluded that this P Cyg profile was consistent with scattering by a neutral medium outflowing at ~300 km s-1. A weak secondary Lyα emission component is observed onto the deep absorption trough. This emission is also blue-shifted by ~300 km s-1. Moreover, this secondary emission extends along ~6.0″ (~2.3 kpc), with a total observed flux of F(Lyα)~ 5.5 × 10-15 erg s-1 cm-2. Figure 9 shows the spatial profile of the main and secondary Lyα components, together with the profile of the UV continuum. As observed, although ~10 times weaker, the secondary Lyα emission roughly follows the spatial profile of the primary component, except for the peak ~1″ north of the central cluster, which is not observed in the former. Both emissions are spatially decoupled from the stellar continuum.

thumbnail Fig. 9

Emission profiles of the UV continuum (1224.5 Å rest frame, black), primary Lyα (blue), and secondary Lyα (red) emissions from HST/STIS data. The vertical scale corresponds to the Lyα flux. The secondary emission has been scaled by a factor 10 (dashed red line) to ease the comparison. Spatial resolution of all the profiles was degraded to 0.25 arcsec pixel-1.

Open with DEXTER

thumbnail Fig. 10

Extracted 1D spectrum of the central region of IRAS 0833 (corresponding to 2.̋0). The position of the geocoronal Lyα line has been marked, in the centre of the broad Galactic absorption profile. We have also marked the position of the secondary emission peak discussed in the text. Reproduced from Mas-Hesse et al. (2003).

Open with DEXTER

Mas-Hesse et al. (2003) found the same velocity structure in the P Cyg profile of the Lyα emission detected in IRAS 0833 all along the STIS slit, i.e., over at least 10″ (~4 kpc). They found that the profiles fall to zero at the same velocities within a narrow range (± 50 km s-1), regardless of the distance to the central region. Therefore, these observations indicate that most of the neutral gas responsible for the absorption feature must be approaching us at the same velocity ~300 km s-1 all along the slit.

Table 7

X-ray spectral fitting of IRAS 0833, with the number of degrees of freedom (ν) and the reduced χ2 value (χ2/ν).

Table 8

X-ray fluxes and luminosities of IRAS 0833 with the fractional contribution to the soft X-ray range by the harder spectral components.

3.7. X-rays

As explained above, the analysis of the X-ray data of IRAS 0833 was performed using XSPEC. The complexity of each model was increased once its statistical significance had been checked with respect to the previous more simple model. We considered the changes in the models valid if the probability significance yielded from the F-test was >99%. Galactic neutral absorption was fixed to N(H I)Gal = 4.5 × 1020 cm-2 (see Table 1). The intrinsic neutral hydrogen column density was measured by Kunth et al. (1998) from the Lyα profile, obtaining N(H I)int = 7.9 × 1019 cm-2, although the uncertainties were large and this value only corresponds to the line of sight of the central cluster. We studied models with the parameter N(H I)int free, but no improvement on the fitting was obtained over the models with N(H I)int fixed to the value reported by Kunth et al. (1998), which was therefore assumed. This is due to the low intrinsic column density of IRAS 0833, which is roughly one order of magnitude lower than the Galactic one. Metallicity was fixed to Z = 0.4 Z as measured by López-Sánchez et al. (2006). As explained in Sect. 2.3, models were checked to yield similar values for the parameters for both MOS1 and MOS2 spectra.

thumbnail Fig. 11

XMM-Newton/EPIC X-ray spectrum of IRAS 0833 with fitting Model 1 (left) and Model 2 (right). The different components of the models are shown in dashed and dotted lines. The MOS1 and MOS2 data are shown with the residuals of the fits (lower panels). Model 2 is favoured over Model 1 since it reproduces the hard X-ray emission with a non-thermal component, as predicted by synthesis models.

Open with DEXTER

Models with only one emitting component did not yield satisfactory fittings. Finally, we found two models which fitted well the spectra, and which we labelled as Model 1 and Model 2. The former is a two-temperature hot plasma (HPsoft and HPhard), modelled by mekal in XSPEC (Mewe et al. 1985). On the other hand, Model 2 is a composite of a hot plasma and a power-law (PL) emission . All components from both models are affected by the Galactic and the intrinsic absorptions. The values of the parameters obtained for each of the models are listed in Table 7, where errors were calculated for a confidence level of 90%. The temperature of HPsoft is similar for Model 1 and Model 2 ( keV and keV, respectively), as well as its flux contribution, which indicates that HPsoft accounts for the same emission in both models. Only the modelling of the hard emission varies between them, being another hot plasma (HPhard) with keV in Model 1 and a PL with in Model 2. When considering a Raymond-Smith model (raymond model in XSPEC, Raymond & Smith 1977) instead of mekal, a worse fitting was obtained, and thus raymond was rejected.

In Table 8, the values of the fluxes and luminosities for Models 1 and 2 are shown, along with , the fractional contribution to the total soft X-ray (0.4 − 2.4 keV) luminosity by the component dominating the hard X-ray range, i.e., for Model 1, and / for Model 2. Errors correspond to a confidence level of 90%. The MOS1 and MOS2 spectra of the source are shown in Fig. 11, overlaid with Models 1 and 2. Stevens & Strickland (1998) analysed a ROSAT/Position Sensitive Proportional Counters (PSPC) observation and fitted the resulted spectrum of the source with a model of an absorbed hot plasma. Assuming a Galactic column density of N(H I)Gal = 4.08 × 1020 cm-2, they obtained keV, N(H I cm-2, , and a value for the soft X-ray luminosity erg s-1. Uncertainties are very high since the statistics of the ROSAT spectrum contained only nine points. Our XMM-Newton data correct downwards the luminosity measured by Stevens & Strickland (1998).

4. Discussion

We have analysed UV-optical observational data of the local, face-on spiral galaxy IRAS 0833, which hosts an intense Lyα-emitting starburst. We have concluded that the onset of the present starburst occurred ~5.5 Myr ago, leading to the production of a total stellar mass of 1.4 × 108 M. While the stellar UV continuum of the integrated source (STB) is highly reddened by dust, the emission of the SSC in its core barely shows any dust extinction. This irregular distribution of dust causes SSC to account for 5% of the STB mass, but up to 20% of the integrated observed UV flux. Also, IRAS 0833 shows a prominent Lyα emission, with a complex spectral profile featuring two emission peaks. The brightest peak shows a clear P Cyg profile due to resonant scattering by neutral gas outflowing at ~300 km s-1. The ~10 times weaker secondary Lyα emission is observed on top of the deep saturated absorption trough caused by the neutral hydrogen in the main component. This secondary Lyα emission appears blue-shifted by the same velocity ~300 km s-1 as the outflowing neutral gas. Both compact and diffuse Lyα emissions were detected along the slit. Whereas the former is observed in two regions located between the central massive stellar cluster and the Hα-emitting structures, the latter was detected extending ~6″ (~2.3 kpc) northward from the northern Lyα compact component, with a much weaker diffuse emission extending to the south. Finally, we have analysed the X-ray spectrum of STB, which is contributed by thermal emission in the soft X-ray range and a non-thermal power law, or a hotter thermal component, also extending to hard X-rays. The low spatial resolution of the X-ray image does not allow us to extract any morphological information.

In this section, we will discuss the properties of SSC and its surroundings, the origin of the X-ray luminosity from STB, and the distribution of the Lyα emission over the central area of IRAS 0833.

4.1. Central super stellar cluster

We have seen in Fig. 4 that the central SSC in IRAS 0833 shows a conspicuous UV continuum, but very low local Hα emission. Also, Lyα shows a local minimum at the position of the central SSC, as shown in Fig. 8. To analyse the lack of Hα emission in the very inner regions of the IRAS 0833 core, a smaller circular region with radius =0.5 (~200 pc, hereinafter R1) was considered. This is marked on Fig. 4. We extracted the Hα and UV fluxes of R1 and corrected them for the nebular and stellar dust extinction, respectively, as discussed in Sect. 3.3. For an instantaneous starburst with the properties given in Table 5, and a mass which would account for the UV emission within R1 (M ~ 3.1 × 106 M), CMHK02 models predict a value of L(Hα)~ 1.5 times larger than indeed observed within the aperture, even when assuming 1 − f= 0.5 as discussed above. In brief, around 30% of the ionizing photons emitted by the SSC and not absorbed by dust, are actually escaping from the region.

On the other hand, Fig. 4 and the spatial profiles of the UV continuum and Hα emission in Fig. 8 indicate that the emitting nebular gas is mostly located towards the north and the south of the central SSC. Specifically, we observe in Fig. 8 that, whereas the bulk of the stellar cluster continuum is concentrated within 1″ (400 pc), Hα shows two main components: north peak (position ~−1.5″ ~ − 600 pc, width ~1″ ~ 400 pc, F ~ 2.5 × 10-15 erg s-1 cm-2), and south peak (position ~1″ ~ 400 pc, width ~1.5″ ~ 600 pc, F ~ 5 × 10-15 erg s-1 cm-2). The weak Hα emission observed at the position of the SSC (F ~ 2.5 × 10-16 erg s-1 cm-2) has to be due to the lack of gas within R1. As discussed above, the gas has been swept up by intense starburst activity (stellar winds, SNe, etc.) during the last 5.5 Myr since the onset of the burst, creating a cavity with a diameter ~1.5″ (~600 pc). Outside this cavity, Hα is actually prominent and structured, resulting from ionization by the local massive stars and by the ionizing photons which escaped from the central region. As already mentioned, the low extinction derived from the stellar continuum in the core SSC indicates that the abundance of dust within the cavity is also very low. These empty, dust-free cavities are usual around strong starbursts, as discussed by Maíz-Apellániz et al. (1998) for NGC 4214. Figure 12 shows in high detail how, in these cases, the nebular lines are emitted by the gas swept up by the starburst with the dust, yielding a strong differential extinction with respect to the stellar cluster, whose line of sight remains almost free of dust and gas.

thumbnail Fig. 12

Central massive stellar cluster in the starburst galaxy NGC 4214, showing the cavity formed around the massive stars. Adapted from the Hubble Heritage collection, courtesy of R. O’Connell and the WFC3 Scientific Oversight Committee.

Open with DEXTER

4.2. X-rays

The low spatial resolution of the X-ray XMM-Newton/EPIC observations, with IRAS 0833 ~8′ away from the optical axis, did not allow us to study its morphology, but only to derive the integrated average physical properties of the emitting gas, as discussed above. The spectral analysis of IRAS 0833 unveiled the contribution by a hard component, which may be thermal (Model 1 in Table 7) or non-thermal (Model 2) in nature, and which in any case contributes up to ~60 − 70% to L0.4 − 2.4 keV. The rather evolved state of the starburst in IRAS 0833 indicates that supernova remnants (SNRs) should already be present, and that a significant number of binary systems might have already become active as high-mass X-ray binaries (HMXBs), i.e., might have started to accrete material onto a compact stellar remnant (black hole or neutron star formed after the SN explosion of its massive, short-lived progenitor). The CMHK02 models predict NSN ~ 10-9 M-1 yr-1 (Cerviño & Mas-Hesse 1994), which scales to around 150 SN explosions every 103 yr for IRAS 0833, half of which are expected to be part of a binary system. The fraction that also have a massive secondary, already in its last stages of evolution after 5.5 Myr, should be active as HMXBs. The CMHK02 models predict ~2 × 103 active HMXBs in IRAS 0833 given the mass and evolutionary status of the starburst (Cerviño et al. 1997). Assuming a typical luminosity of LX = 1038 erg s-1 for a single HMXB, a contribution of LX = 2 × 1041 erg s-1 by HMXBs is expected. This value matches the luminosity contribution of the non-thermal component of Model 2 of Table 7. Furthermore, following the method proposed by Mineo et al. (2012a), based on the calibration of HMXB luminosity function in star-forming galaxies using a composite of LFIR and observed LUV as a proxy, we find that ~60 HMXBs with LhardX> 1038 erg s-1 should be currently active in IRAS 0833, providing an integrated hard X-ray luminosity L0.5−8 keV ~ 5 × 1040 erg s-1. While this value is lower than the observed value, L0.5−8 keV ~ 2 × 1041 erg s-1, it is within 1σ of the dispersion in the sample by Mineo et al. (2012a). We conclude, therefore, that the hard X-ray component identified in IRAS 0833 is consistent with the emission from the expected number of HMXBs originated by the evolution of the starburst episode, favouring so the non-thermal nature of the hard X-ray emission, as assumed in Model 2 of Table 7.

On the other hand, the soft X-ray component associated with the emission by a thermal plasma at kT ~ 0.62 keV, as in Model 2, contributes with L0.4−2.4 keV ~ 4.2 × 1040 erg s-1. The soft X-ray emission in starburst galaxies originates from 1) the heating of the diffuse gas surrounding the stellar clusters up to temperatures of millions of Kelvin, due to the stellar winds and supernovae injecting mechanical energy into the medium; and 2) by the emission of SNRs during the adiabatic phase. The contribution due to the direct emission of individual stars is negligible when compared to the previous components (Cerviño et al. 2002). Evolutionary population synthesis models CMHK02 compute the expected soft X-ray emission of a starburst after defining the physical properties of the burst, considering that a free fraction ϵxeff from the mechanical energy injected into the medium is finally converted into soft X-ray luminosity. Typical values of ϵxeff = 0.010.1 are found for star-forming objects (Summers et al. 2004; Mas-Hesse et al. 2008; Mineo et al. 2012b).

The CMHK02 models for a starburst episode as defined for IRAS 0833 in Table 5 could reproduce the total observed soft X-ray luminosity L0.4−2.4 keV ~ 1.47 × 1041 erg s-1, assuming ϵxeff= 0.03 − 0.04. If we remove the contribution to the soft X-ray luminosity by the non-thermal power law component, then the efficiency would be constrained to ϵxeff= 0.01 − 0.02. Nevertheless, since the extrapolation of the hard X-ray component to the soft X-ray range might be quite uncertain, we conclude that between 1% and 4% of the mechanical energy released by the starburst in IRAS 0833 is heating the surrounding medium, being re-emitted in the form of soft X-ray photons. This value is within the usual range in starburst galaxies, as commented above. Indeed, the LsoftX/LFIR ratio measured in IRAS 0833, log (LsoftX/LFIR) = −3.9, is very close to the mean ratio measured by Mas-Hesse et al. (2008) on their complete sample of star-forming galaxies, log (LsoftX/LFIR) = −3.92 (see Fig. 5 in Mas-Hesse et al. 2008). We conclude, therefore, that the integrated soft and hard X-ray emissions of IRAS 0833 are consistent with the predictions for a starburst with the properties we have derived, so that no additional sources of X-rays (e.g., a low-luminosity active galactic nucleus) have to be claimed to explain them.

4.3. The Lyα emission

We show in Fig. 8 the spatial profiles of the UV continuum, Lyα, and Hα emissions for IRAS 0833 with the observed values of the Hα/Hβ and Lyα/Hα ratios. The expected Lyα/Hα has also been included, calculated assuming Case B recombination (Lyα/Hα= 8.7, Hα/Hβ= 2.87, Dopita & Sutherland 2003) and applying the dust extinction derived from the Balmer decrement. In what follows, the effect of the Galactic extinction was removed before the analysis. The values of color excess, line ratios or escape fractions do not include this effect. The profiles were binned spatially, lowering the spatial resolution when needed in order to always get a signal-to-noise ratio larger than 3. Resulting resolutions are within the range 0.1−0.5 arcsec pixel-1. We have also plotted the profile of the Lyα escape fraction in the bottom panel of this figure, as defined by Atek et al. (2008),

Therefore, 1  represents the fraction of Lyα photons which are destroyed internally (or at least scattered out of the line of sight), either by dust absorption or by multiple interaction with the neutral gas. As we discuss below, resonant scattering redistributes the Lyα photons over large areas so that local values may differ significantly from globally integrated values.

The spatial profiles in Fig. 8 show that in addition to the compact, central emitting blobs, a diffuse and extended Lyα component is present along the STIS slit in IRAS 0833, as discussed in Sect. 3.6. Neither of them show any clear spatial correlation with the UV continuum, Balmer decrement, or Hα emission whatsoever, similar to the case of Haro 2 (Otí-Floranes et al. 2012). As we concluded in Sect. 4.1, the nebular gas seems to have been pushed out by the stellar activity, creating a shell around the central, massive super-cluster SSC. This shell is being ionized by SSC and generates most of the Hα emission. Figure 8 shows that the brightest Lyα emitting regions are also located around the central cluster. Nevertheless, there is a clear spatial decoupling between the Lyα and Hα profiles, as evidenced by the profile of the Lyα/Hα ratio and over the nuclear region, with a local maximum within the central 3″ (~1 kpc). On the other hand, the highest values of Lyα/Hα ratio and are detected in regions dominated by diffuse emission northward of the stellar cluster, at distances higher than 2″ (~800 pc), and in a very localized region at ~5.5″ (~2.1 kpc) to the south (LyαS, see Figs. 8 and 13). The location of LyαS at the outskirts of IRAS 0833 nuclear region has been marked in Fig. 2.

thumbnail Fig. 13

Detail of the Lyα (blue) and Hα (red) spatial profiles around the region at ~5.5 (~2.1 kpc) south of the central SSC showing the highest value of the Lyα/Hα ratio (LyαS). The profiles have been plotted at their original resolution, 0.029 arcsec pixel-1 for Lyα and 0.025 arcsec pixel-1 for Hα.

Open with DEXTER

Otí-Floranes et al. (2012) analysed the Lyα emission in Haro 2 and also found both compact and diffuse components, neither of them spatially coupled to UV continuum, Balmer emission, or dust extinction as traced by Hα/Hβ. However, they found that 1) diffuse Lyα emission was co-spatial to the diffuse soft X-ray emission produced by the injection of mechanical energy into the ISM by the starburst; and 2) the intensity of the diffuse Lyα emission was higher than the values predicted from Hα emission, assuming Case B recombination and correcting for dust extinction using the Balmer decrement. Otí-Floranes et al. (2012) argued that this diffuse emission in Haro 2 might be originated in gas ionized by the hot plasma responsible for the soft X-ray radiation, therefore not ionized by the young massive stellar clusters in the central region. Recombination in this environment would take place far from Case B conditions, causing 1) values of Lyα/Hα similar to those of Case B, but with 2) large values of Hα/Hβ, mimicking the presence of dust, and 3) values of the L(Hα)/L0.4 − 2.4 keV ratio much lower than when ionization is driven by massive stars. This model for the diffuse Lyα emission could explain the observed ratios in Haro 2.

But the scenario seems to be quite different in IRAS 0833. While the highest values of the Lyα/Hα ratios in IRAS 0833 also occur in regions dominated by the diffuse emission, Lyα emission marginally over the values expected from Case B recombination is observed only at region LyαS. Unfortunately, no X-ray image with enough spatial resolution is available to study 1) the co-spatiality of the diffuse soft X-ray emission and the components of Lyα; and 2) the local values of the ratio L(Hα)/L0.4 − 2.4 keV. Thus, in this case, we can not analyse whether the hot plasma could contribute significantly to the ionization of the gas in some regions, as done for Haro 2. On the other hand, the bulk of the Lyα emission comes from the region where Hα is also emitted, around the central stellar cluster. The P Cyg profile of the Lyα line along the STIS slit indicates that a large fraction of the Lyα photons are being resonantly scattered out of the line of sight by the outflowing neutral gas. Relatively small variations in the properties of this outflow from one location to another might explain why the resulting Lyα spatial profile differs from the Hα spatial profile in the central 3″ area (~1 kpc). The scattering of a significant fraction of the Lyα photons explains the low values of the Lyα/Hα ratio along the STIS slit.

This scattering by the neutral gas surrounding the SSC could transport the Lyα photons to the outer regions of IRAS 0833, leaking from the galaxy as a weak, diffuse emission halo. The bottom right panel of Figure 4 shows the Lyα image produced by Östlin et al. (2009). While the procedure used to extract the Lyα emission in IRAS 0833 might be quite uncertain where the stellar continuum is strong (and the Lyα line shows a strong P Cyg profile), it provides a fair representation of the diffuse Lyα emission. This diffuse emission is quite homogeneous and does not follow the spatial structure of the Hα emission. A good example is found in region LyαS, as shown in Fig. 13, where Lyα/Hα becomes higher than 1 because of the local lack of Hα photons. Eastwards of the nucleus, where the Hα emission is less prominent, the diffuse Lyα also remains at a similar surface brightness as in other areas. The morphology of the Lyα emission supports our conclusion that the Lyα photons which are missing along the slit are indeed smoothly redistributed over the whole central area of IRAS 0833. Moreover, after refining the method used by Östlin et al. (2009) to remove the UV continuum from the cont. +Lyα images, Hayes et al. (2014) obtained Lyα images of the 14 star-forming galaxies of the LARS sample (Hayes et al. 2013). Half of the sources show a conspicuous Lyα halo which do not present any spatial coupling to UV or Hα emissions whatsoever. Hayes et al. (2014) argued that resonant scattering may have redistributed the Lyα photons over the large areas observed. A similar situation seems to be present in IRAS 0833.

Combining the IUE Lyα line flux and the value of the Hα integrated emission measured by López-Sánchez et al. (2006), we derived a globally integrated ratio Lyα/Hα= 0.4 (corresponding to ~ 0.04). This value is ~10 times lower than the expected value, assuming case B conditions and the average nebular extinction, Lyα/Hα= 4.4. The destruction of Lyα photons is therefore significantly larger than the extinction extrapolated from the Balmer lines. Multiple resonant scattering of Lyα photons by neutral atoms significantly increases their probability of being destroyed by dust grains, yielding a much higher extinction by dust for photons around 1216 Å. As predicted by Lyα photons transfer models (Verhamme et al. 2006), multiple scattering can also end up shifting their energy, becoming diluted at the wings of the emission line; this effect is negligible when dust is abundant, as in the central area of IRAS 0833. This redistribution and spatial smoothing of the Lyα photons are responsible for the spatial decoupling between Lyα and Hα emissions visible in Figs. 4 and 8. This leads to the conclusion that the star formation rate may be severely underestimated when based solely on integrated fluxes of the Lyα line without a proper correction of the aforementioned effects. See Fig. 2 in Hayes et al. (2011), for instance.

In any case, we do not find evidence in IRAS 0833 of any further source of Lyα photons like ionization by hot plasma, as it was the case in Haro 2. If present, its contribution should be negligible when compared to the ionization by massive stars. Comparison with the results on Haro 2 brings us once again to the evidence that the physics of Lyα photon production, transfer, and destruction is extremely dependent on the conditions of each galaxy, making it difficult to find any clear correlation with the macroscopic properties of each object.

Based on the work by Tenorio-Tagle et al. (1999), Mas-Hesse et al. (2003) proposed an evolutionary picture of starbursts to try to explain the presence or lack of Lyα emission in star-forming galaxies, as well as the spectral features of the line. They argued that the starburst activity in IRAS 0833 has accelerated the surrounding material, forming a superbubble which expands at a velocity of ~300 km s-1 in whose interior massive stars have ionized the gas. Nebular emission lines are produced after recombination, Lyα among them. Because of its resonant nature, the bluest photons of the Lyα line are scattered by the neutral hydrogen on the shell of the superbubble, originating the P Cyg profile of the primary Lyα component. Furthermore, they proposed that the shock produced in the leading front may cause the shocked gas to undergo recombination. Since no dense neutral gas layers are ahead of this front, the Lyα photons would not be scattered and would be detected by the observer with a blueshift corresponding to the expansion velocity of the superbubble. Mas-Hesse et al. (2003) proposed that the secondary Lyα component in IRAS 0833 could be produced by this mechanism. This secondary emission does not seem to be affected by neutral hydrogen scattering, which indicates that the H I column density in front of the expanding superbubble must indeed be low. This blue-shifted emission is in any case weaker by an order of magnitude than the primary component, and does not contribute significantly to the integrated Lyα emission. According to the Lyα spectral profiles observed, the starburst in IRAS 0833 seems to be in phase (d) of the model by Tenorio-Tagle et al. (1999), as shown in their Fig. 8, which is characterized by a Lyα emission with a P Cyg profile, together with a blue-shifted emission component.

On the other hand, Leitherer et al. (2013) obtained an HST/COS spectrum of IRAS 0833 which includes the Lyα line. They fitted its spectral profile using the radiation transfer code described by Schaerer et al. (2011), which is an improved version of the models developed by Verhamme et al. (2006). Assuming the simple geometry of an expanding spherical shell, with expansion velocity, Doppler parameter, column density, and dust optical depth as free parameters, Leitherer et al. (2013) obtained a spectral profile which nicely reproduces the primary Lyα spectral component and predicted the presence of a secondary emission. Nevertheless, the model fails to properly reproduce the wavelength of the secondary emission peak, as well as the profile of the blue, P Cyg-like absorption wing. More detailed modelling of the Lyα line in IRAS 0833 would be needed to unveil the origin of this secondary Lyα emission.

5. Conclusions

We have carried out a multiwavelength spectral and photometric analysis of the face-on, spiral, star-forming galaxy IRAS 0833. The properties of its starburst episode and associated X-ray emission have been defined, and the morphology of the Lyα emission has been discussed, together with its relation to the UV continuum from the massive stars, the Hα line, and the Hα/Hβ ratio.

  • The properties of the starburst in IRAS 0833 have been defined by comparison to evolutionary population synthesis models. We conclude that a single stellar population with an age of 5.5 Myr, a mass of M ~ 1.4 × 108 M, and average stellar and nebular extinctions of E(BV) = 0.15 and E(BV) = 0.06, respectively, can account for the integrated UV, Hα, and FIR photometry, as well as for the spectral profile of the photospheric Si IV and C IV lines. The integrated UV spectrum shows the 2175 Å feature, consistent with an LMC-like extinction law.

  • An underlying stellar population, some hundreds of Myr old, dominates the IR-optical continuum in IRAS 0833, as well as the observed Ca II triplet and Mg2 absorption lines.

  • The nucleus of IRAS 0833 shows a massive cluster whose stellar continuum is barely attenuated (E(BV) = 0.01). We have estimated that it accounts for 20% of the observed integrated UV flux and 5% of the total mass of the starburst. The images of the nucleus show a very weak Hα emission co-spatial to the central cluster. In fact, the synthesis models overestimate by a factor 1.5 the Hα luminosity emitted within a central circular region of radius = 0.5″ (~200 pc). We argue that the low Hα emission observed in this core region is due to the lack of nebular gas around the massive central cluster, which has cleaned its surroundings after pushing out the nearby gas, producing an expanding Hα-emitting shell.

  • The X-ray image does not provide any spatial information on the morphology of the emission. The starburst episode can account for the thermal soft X-ray emission detected assuming an efficiency in the conversion of mechanical energy into X-rays of ~1 − 4%, within the typical range for starburst galaxies. The non-thermal component, which dominates the hard X-ray emission, is consistent with the number of active HMXBs expected for an evolved starburst with the properties derived for IRAS 0833.

  • Two spectral Lyα components are observed. The main Lyα is detected along 16″ (~6 kpc) and shows a P Cyg profile originated by resonant scattering by neutral hydrogen outflowing at ~300 km s-1. The velocity structure extends all along the slit. The secondary emission is ~10 times weaker, extends over 6″ (~2.3 kpc), and is detected over the spectral trough caused by the outflowing neutral material. All this indicates that the starburst activity has accelerated the sourrounding neutral gas, which no longer scatters the whole line, but only its bluest photons. In the leading front of the superbubble the gas seems to have undergone recombination, producing the secondary Lyα emission, with probably scarce neutral material ahead, so that it is detected blue-shifted by ~300 km s-1.

  • The Lyα line is observed in emission along 16″ (~6 kpc), without regions of total Lyα absorption. It does not show any spatial matching with the UV continuum, Hα line, or Balmer decrement whatsoever. The Lyα photons escape mostly from two compact areas associated with, but not coincident with, the Hα-emitting shell, with a size of 1″ (~400 pc, north) and 1.5″ (~600 pc, south). In addition, there is a diffuse Lyα component which extends northward 6″ (~2.3 kpc).

  • Similarly, as found in other starburst galaxies, like Haro 2 (Otí-Floranes et al. 2012), the diffuse Lyα emission shows the highest Lyα/Hα ratios, as well as the highest values of the Lyα escape fraction, . The globally integrated Lyα/Hα ratio is well below the Case B predictions (even after dereddening using the E(BV)neb average value derived from the Hα/Hβ ratio), indicating an enhanced destruction of Lyα photons with respect to the extinction extrapolated from the Balmer lines.

  • We conclude that the outflowing neutral gas in front of the IRAS 0833 starburst area is resonantly scattering most of the Lyα photons, producing the observed P Cyg profile. These scattered photons are smoothly redistributed over the whole central area until either they are destroyed by interaction with dust, or escape from an extended region. Multiple scattering enhances the probability of interaction with dust, explaining why the Lyα/Hα ratio values are much lower than expected for Case B conditions and the average extinction. We do not find any evidence in IRAS 0833 of any other further source of Lyα photons, like ionization by a hot plasma as proposed for Haro 2. If present, its contribution should be negligible when compared to the ionization by massive stars.

  • The star formation rate in galaxies, e.g., high-redshift sources, may be severly underestimated when it is only derived from integrated fluxes of the Lyα line without a proper correction of transfer effects in the interstellar medium.


1

IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under cooperative agreement with the National Science Foundation.

3

PATTERN describes how the charge cloud released by the incoming X-ray photon distributes over the pixels. See the SAS manual for a more detailed description.

Acknowledgments

HOF and JMMH are partially funded by Spanish MINECO grants AYA2010-21887-C04-02 (ESTALLIDOS), AYA2011-24780/ESP, and AYA2012-39362-C02-01. HOF is funded by Spanish FPI grant BES-2006-13489, CONACYT grant 129204, and a postdoctoral UNAM grant. G.Ö. is a Swedish Royal Academy of Sciences research fellow supported by a grant from Knut and Alice Wallenberg foundation, and also acknowledges support from the Swedish Research Council (VR) and the Swedish National Space Board (SNSB). M.H. received support from the French Agence Nationale de la Recherche (ANR-09-BLAN-0234-01). H.A. and D.K. are supported by the Centre National d’Études Spatiales (CNES) and the Programme National de Cosmologie et Galaxies (PNCG). This research has made use of the Spanish Virtual Observatory (http://svo.cab.inta-csic.es) supported from the Spanish MICINN/MINECO through grants AyA2008-02156, AyA2011-24052. We want to acknowledge the use of the Starburst99 models and the NASA/IPAC Extragalactic Database (NED). We are very grateful to the ESTALLIDOS collaboration for its scientific support. This paper was based on observations with Hubble Space Telescope, XMM-Newton, International Ultraviolet Explorer, and William Herschel Telescope. We thank both the anonymous referee and the language editor for their comments, which helped to improve the understandability and clarity of the manuscript.

References

All Tables

Table 1

IRAS 0833 coordinates, redshift, distance, scale, Galactic H I, color excess towards the source, and oxygen abundance.

Table 2

Log of HST/GHRS and IUE observations of IRAS 0833 used in this work.

Table 3

Log of HST/STIS observations of IRAS 0833 used in this work.

Table 4

Log of WHT/ISIS observations of IRAS 0833 used in this work.

Table 5

Predicted values of the age, mass, stellar and nebular extinctions, fraction of ionizing photons absorbed by dust, emission fluxes and luminosities and EW(Hβ) of the starburst (STB) in IRAS 0833, assuming the UV fluxes from HST/ACS and IUE and the average value of both.

Table 6

Observed values of the UV and Hα emissions and the stellar and nebular extinctions of the integrated source IRAS 0833 (STB) and its central stellar cluster (SSC) with the predicted stellar mass for each of them.

Table 7

X-ray spectral fitting of IRAS 0833, with the number of degrees of freedom (ν) and the reduced χ2 value (χ2/ν).

Table 8

X-ray fluxes and luminosities of IRAS 0833 with the fractional contribution to the soft X-ray range by the harder spectral components.

All Figures

thumbnail Fig. 1

Left: multispectral HST/ACS image of IRAS 0833. Blue: F300W, green: F435W, red: FR656N (Courtesy of Ángel R. López-Sánchez – AAO). Right: F140LP UV image. Integrated emissions of the starburst (STB) and central super stellar cluster (SSC) are labelled. Intensity scale is logarithmic. Same spatial scale in both images, as marked on the right panel. North is up and east is to the left.

Open with DEXTER
In the text
thumbnail Fig. 2

F140LP UV image. HST/GHRS (squared) and IUE (oval) apertures are superimposed, together with the location of the STIS slit. The width of the apertures and the slit are represented at real scale. The location of the region LyαS discussed in Sect. 4.3 is marked with a cross within a circle. Intensity scale is logarithmic. North is up and east is to the left.

Open with DEXTER
In the text
thumbnail Fig. 3

XMM-Newton/EPIC MOS1 (left) and MOS2 (right) X-ray images of IRAS 0833. Intensity scale is logarithmic. Source region from which X-ray spectrum was extracted for each camera is confined within the white circle. Contours from HST/ACS F144LP image are shown in black. The PSF of this off-axis observation is large and asymmetric, and thus the emission must be considered as non-resolved. North is up and east is to the left.

Open with DEXTER
In the text
thumbnail Fig. 4

HST/ACS images of IRAS 0833 in the ultraviolet (filter F140LP, green, top left), continuum-subtracted Hα (red, top right), the composite of both (bottom left), and the composite including Lyα emission in blue (bottom right, adapted from Östlin et al. 2009). The aperture of HST/GHRS is marked with a square. The circular region R1 (radius ~0.5 (~200 pc)), as discussed in the text, is also marked. Intensity scale is linear, except for the last image, in which intensity scale is logarithmic. North is up and east is to the left.

Open with DEXTER
In the text
thumbnail Fig. 5

HST/GHRS spectrum of SSC in IRAS 0833 (solid line) fitted with an SB99 IB model for an age of ~5.5 Myr (dashed line). Left: normalized spectrum with labelled absorption lines Si IV λλ1394, 1403 Å, and C IV λλ1548, 1551 Å. Right: regular spectrum reddened to account for Galactic (Cardelli extinction law, E(BV) = 0.094) and intrinsic (LMC law, E(BV) = 0.01) extinctions.

Open with DEXTER
In the text
thumbnail Fig. 6

Left: IUE spectrum of STB in IRAS 0833 (black, solid line) fitted with an SB99 spectrum for an age of 5.5 Myr (dashed lines). The model has been reddened to account for the Galactic extinction, as well as for the intrinsic extinction assuming both an LMC law and E(BV) = 0.15 (red), and the Cardelli extinction law with E(BV) = 0.20 (green). Right: IUE spectrum is shown as in the left panel (black, solid line). A composite model (blue, dashed line) was obtained summing the starburst model (red) and an SB99 spectrum of a 1-Gyr-old population reddened similarly as the young burst model (magenta, dashed line). Near-UV and optical photometric fluxes from Östlin et al. (2009) are marked with green crosses.

Open with DEXTER
In the text
thumbnail Fig. 7

Observed HST/GHRS (black, solid line, SSC) and IUE (blue, solid, STB; errors in green) spectra with the SB99 model for an age of 5.5 Myr (red, dashed). Vertical scale corresponds to the GHRS spectrum. The other spectra have been arbitrarily scaled for an easier comparison. SB99 model has been corrected only for Galactic extinction.

Open with DEXTER
In the text
thumbnail Fig. 8

Top: HST/STIS G140M spectral image of IRAS 0833. Spatial extension is ~ 19″ (~7.5 kpc) . Middle: emission profile of the main component of Lyα (blue) and UV continuum (1224.5 Å rest frame, black) from STIS data, and Hα (dashed red) from ACS data. Bottom: spatial profiles of Lyα (blue) and observed Lyα/Hα (solid black) and Hα/Hβ (from WHT/ISIS data, in red), and the expected Lyα/Hα ratio (dashed brown), assuming Case B recombination and internal reddening as derived from the Hα/Hβ ratio, as well as the Galactic extinction. The Lyα escape fraction , as defined in the text, has been plotted in dotted black line. In the middle and bottom panels, the negative x-axis corresponds to north, as indicated by N. To ease the comparison, the dashed red line in the middle panel shows the Hα profile scaled by a factor 0.2. The vertical scale corresponds to the Lyα and non-scaled Hα fluxes. Whereas the vertical scale of the left axis in the bottom panel corresponds to the Lyα/Hα and Hα/Hβ ratios, the scale of the right axis corresponds to . Position of the region LyαS is marked.

Open with DEXTER
In the text
thumbnail Fig. 9

Emission profiles of the UV continuum (1224.5 Å rest frame, black), primary Lyα (blue), and secondary Lyα (red) emissions from HST/STIS data. The vertical scale corresponds to the Lyα flux. The secondary emission has been scaled by a factor 10 (dashed red line) to ease the comparison. Spatial resolution of all the profiles was degraded to 0.25 arcsec pixel-1.

Open with DEXTER
In the text
thumbnail Fig. 10

Extracted 1D spectrum of the central region of IRAS 0833 (corresponding to 2.̋0). The position of the geocoronal Lyα line has been marked, in the centre of the broad Galactic absorption profile. We have also marked the position of the secondary emission peak discussed in the text. Reproduced from Mas-Hesse et al. (2003).

Open with DEXTER
In the text
thumbnail Fig. 11

XMM-Newton/EPIC X-ray spectrum of IRAS 0833 with fitting Model 1 (left) and Model 2 (right). The different components of the models are shown in dashed and dotted lines. The MOS1 and MOS2 data are shown with the residuals of the fits (lower panels). Model 2 is favoured over Model 1 since it reproduces the hard X-ray emission with a non-thermal component, as predicted by synthesis models.

Open with DEXTER
In the text
thumbnail Fig. 12

Central massive stellar cluster in the starburst galaxy NGC 4214, showing the cavity formed around the massive stars. Adapted from the Hubble Heritage collection, courtesy of R. O’Connell and the WFC3 Scientific Oversight Committee.

Open with DEXTER
In the text
thumbnail Fig. 13

Detail of the Lyα (blue) and Hα (red) spatial profiles around the region at ~5.5 (~2.1 kpc) south of the central SSC showing the highest value of the Lyα/Hα ratio (LyαS). The profiles have been plotted at their original resolution, 0.029 arcsec pixel-1 for Lyα and 0.025 arcsec pixel-1 for Hα.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.