Highlight
Free Access
Issue
A&A
Volume 522, November 2010
Article Number A51
Number of page(s) 7
Section Galactic structure, stellar clusters and populations
DOI https://doi.org/10.1051/0004-6361/201014302
Published online 01 November 2010

© ESO, 2010

1. Introduction

Massive stars are key agents shaping the interstellar medium, through ionizing starlight, stellar winds and supernovae. This has been recognized throughout galaxies (Elmegreen & Scalo 2004), and also locally in the surroundings of our Sun (Frisch 1995; de Geus 1992). Massive stars tend to be located in clusters (Lada & Lada 2003; Zinnecker & Yorke 2007). Stellar winds and explosions are expected to create interacting gas shells (Weaver et al. 1977), merging of hot cavities may form superbubbles (Tenorio-Tagle & Bodenheimer 1988; Mac Low & McCray 1988; Cho & Kang 2008). The environment of the Sun shows clear signatures of massive-star action, though located in a seemingly quiet region of the Galaxy, about 8 kpc from its center and about 4–5 kpc from the molecular ring identified as the region of the Galaxy’s highest star-forming activity (Scoville & Young 1984). The interstellar medium near the Sun is characterized by a rather complex morphology (Frisch 1995). The Local Bubble, a rather tenuous cavity, surrounds the Sun (Breitschwerdt et al. 1998). Several loops/shells reminiscent of supernova remnants have been identified (de Geus 1992). The local bubble is the most-nearby of such cavities, and finding its origins is a prominent current science topic (Breitschwerdt & de Avillez 2006; Welsh & Shelton 2009). Groups of massive stars are found within a distance of several hundred parsec (de Zeeuw et al. 1999), and also nearby clouds with current star forming activity as evident in T Tauri stars, HII regions, and embedded IR sources. The stellar association of Scorpius-Centaurus (Sco OB2) constitutes a prominent group of most-nearby stars (see Fig. 1). It is located at a distance of about 100–150 pc (de Geus 1992; de Zeeuw et al. 1999; Preibisch & Zinnecker 1999). Apparently it plays a major role for the nearby interstellar medium and accounts for a significant fraction of the massive stars in the solar vicinity (de Zeeuw et al. 1999). This association shows several subgroups of different ages (estimated as 5, 16, and 17 My, with typical age uncertainties of 2 My (de Geus et al. 1989; see however Slesnick et al. 2008; Fuchs et al. 2009). Its location in the sky is centered  ~ 20° above the plane of the Galaxy and hence sufficiently-distinct, unlike other Galactic source groups which would be superimposed along the line of sight towards the Galactic ridge or molecular ring, for example. Our target is 26Al radioactivity from this region.

thumbnail Fig. 1

OB associations near the Sun, with Sco-Cen stars marked (from Preibisch & Mamajek 2008). This top view onto the Galaxy in the local vicinity shows that several nearby groups of massive stars can be identified, with the Scorpius-Centaurus group as the closest to the Sun. This compilation is based on Hipparcos position measurements of nearby stars (for details see Preibisch & Mamajek 2008).

Open with DEXTER

Massive stars are believed to form from dense parts of giant molecular clouds, and have typical evolution time scales of 1–10 My (see reviews by McKee & Ostriker 2007; Zinnecker & Yorke 2007). One of the most important questions in models of the star formation process is the onset of release of energy and matter by the most-massive stars and their effects on the surrounding clouds (Oey & Clarke 2007; Clark et al. 2008). Such feedback on star formation in dense clouds can be either negative (termination of star formation through dispersal of the natal cloud) or positive (triggering of further star formation by compression of cloud material). The massive-star clusters are believed to be dispersed within a typical time scale of 10 My (e.g. Pfalzner 2009).

Within a group of massive stars, Wolf-Rayet-phase winds carry the products of core hydrogen burning including 26Al and lead to a steeply-rising 26Al content in surrounding ISM about 3 My after those stars have been formed (Voss et al. 2009). Interstellar 26Al content then decreases again within 1–2 My, from radioactive decay and because the less-massive, more slowly-evolving stars do not develop the Wolf-Rayet phase below Mi ~ 25 M. Trailing this Wolf-Rayet-dominated initial 26Al production, contributions from core-collapse supernovae take over to enrich the ISM with 26Al. This production smoothly declines as less and less-massive stars terminate their evolution in a supernova, and the interstellar 26Al content fades after  ~ 7–8 My with a tail extending to  ~ 20 My. In general, stellar groups with an age in the range of 3–20 My may be considered plausible sources of interstellar 26Al (see Voss et al. 2009, and references therein). The youngest stellar subgroup of the Sco-Cen association is the Upper Scorpius group at an age of about 5 My. This should correspond to the peak of the expected 26Al content (see Fig. 2 of Voss et al. 2009).

26Al is ejected through stellar winds and supernova explosions at typical velocities around 1000 km s-1 (Niedzielski & Skorzynski 2002; Kifonidis et al. 2000), which will be slowed down by interactions with circumstellar material, and by the reverse shock which results from the discontinuity between the expanding ejecta and ambient ISM. 26Al decays with an exponential lifetime of 1 My. Hence 26Al decay occurs along the ejecta trajectory in the surrounding ISM. Most likely that will be hot and probably turbulent, from the injected radiation, wind and supernova energies. Upon β+-decay, a γ-ray line at 1808.63 keV is emitted, together with a positron (Endt 1990). High-resolution spectroscopy of the 26Al γ-ray line contributes to test massive-star nucleosynthesis, and also the general scenario of their formation in groups and sequences in an independent way, through such a radioactivity clock. Also, the 26Al γ-ray line centroid and width reflect the kinematics of massive-star ejecta at late times: The Doppler effect will encode bulk motion and turbulence in the observed 26Al γ-ray line profile. Here we report our search for 26Al γ-ray emission from the nearby Sco-Cen association and its subgroups with the Compton Observatory and INTEGRAL γ-ray instruments.

thumbnail Fig. 2

All-sky image (top) and expanded view towards Sco-Cen (bottom) of 26Al γ-rays as derived from COMPTEL observations (Plüschke et al. 2001).

Open with DEXTER

2. Earlier data and hints

The COMPTEL instrument was operated aboard the Compton Gamma-Ray Observatory mission of NASA through 1991–2000 (Gehrels et al. 1993; Schoenfelder et al. 1993). It was a Compton telescope, detecting γ-rays in the MeV range through a correlated detection of a Compton scatter process in its upper detector plane and the secondary photon from this scatter process being absorbed in its lower detector plane (see Schoenfelder et al. 1993, for a detailed instrument description). With its large field of view of about 30°, COMPTEL performed a sky survey for MeV γ-ray lines. It thus provided the first all-sky maps of 26Al emission (Diehl et al. 1995), which showed the 26Al emission extending along the entire plane of the Galaxy. From these data, it was concluded that massive stars are the dominating sources of 26Al ejection into the Galaxy’s interstellar medium, through Wolf-Rayet winds and core-collapse supernovae (see Prantzos & Diehl 1996, for a review).

Tantalizing hints for 26Al emission from a region above the plane of the Galaxy and reminiscent of the Sco-Cen region was noticed in COMPTEL’s 26Al all-sky image as derived from maximum-entropy deconvolution of the data (Fig. 2, Diehl et al. 1995; Knödlseder et al. 1998; Plüschke et al. 2001). The significance of this result is low, however, and depends upon the specific formulation of the hypothesis. A maximum-likelihood test for the emission in the Sco-Cen region obtained 2.1σ significance only, from separately modeling the data with the deconvolved map parts of the region of interest (defined as a circle of 15 degrees radius around (l,b) = (340°, 20°)), this representation of the Sco-Cen region led to only marginal improvement in the maximum-likelihood fit to 9 years of observation above the remainder of the MaximumEntropy map. Similar values could be obtained for other intermediate-latitude regions of the sky of similar extent, and thus the signal was not pursued any further.

INTEGRAL, ESA’s INTernational Gamma-Ray Astrophysics Laboratory, has been operating in space since November 2002 (Winkler et al. 2003). Its γ-ray spectrometer SPI (Vedrenne et al. 2003; Roques et al. 2003) features an excellent spectral resolution of 3 keV at  ~ 1800 keV, which compares to about 200  keV spectral resolution for COMPTEL. Spatial resolutions of these instruments are similar, 2.7° for SPI and 3.4° for COMPTEL, with COMPTEL imaging being achieved through kinematical constraints from measuring the Compton scattering within the instrument, while SPI’s imaging approach is through a coded mask casting a shadow onto its pixelated Ge-detector plane. INTEGRAL’s sky exposure has emphasis in the inner Galactic plane and some special extragalactic source regions. 26Al emission from the sky was detected at the intensities expected from COMPTEL’s earlier survey (Diehl et al. 2003). Making use of the advance in spectral resolution, the line intensity and shape were determined for the inner Galaxy region, and could be used to determine the global properties of 26Al in the Galaxy’s interstellar medium (Diehl et al. 2006b,a). The γ-ray line turned out to be rather narrow and not significantly Doppler-broadened, consistent with the 26Al-enriched interstellar medium being moderately turbulent at velocities in or below the 100 km s-1 range. The Doppler shift of the γ-ray line centroid from large-scale rotation of the Galaxy also was indicated in these high-resolution data. This confirmed the presence of 26Al throughout the plane of the Galaxy. The Galactic 26Al content is determined from the integrated intensity of the γ-ray line to range between 2 and 3 M, when plausible models for the spatial distribution of 26Al in the Galaxy are employed to resolve the distance ambiguities along the lines of sight (Diehl et al. 2006a).

The spatial source distribution models used in these determinations all assume that the 26Al sources are adequately represented to first order by spiral-arm or exponential-disk descriptions of the Galaxy. The overall symmetry of the Galactic 26Al emission as observed in the bright inner ridge in the first and fourth quadrants of the Galaxy is reassuring for such large-scale interpretation (Wang et al. 2009). Only the Cygnus region’s 26Al emission appeared as a clear and prominent deviation from this view (del Rio et al. 1996; Knödlseder et al. 2002), while variations of 26Al emission along the inner ridge of the Galaxy as seen in the COMPTEL image were assumed to be due to the clumpiness of massive-star locations throughout the Galaxy (i.e., the “beads-on-a-string” seen in UV images of nearby galaxies when viewed face-on). Tests for local emission from nearby and special groups of OB associations such as attributed to the Gould Belt (Pöppel 1997; Elias et al. 2009) had been made on COMPTEL data (Knödlseder et al. 1998). Only marginal hints were found, and thus the large-scale interpretation of observed 26Al was held up.

But when nearby 26Al sources, such as may be expected from the Sco-Cen groups of stars as close as 100–150 pc (see Fig. 1, and Blaauw 1964) are superimposed on our viewing direction towards the Galaxy’s inner ridge, a significant bias might occur. This could lead to an over-estimate of the Galaxy’s 26Al content, and of the scale height of the 26Al sources in the Galactic plane, which is determined from 26Al γ-ray data as about 130 pc (Wang et al. 2009). It has been shown that the Galactic 26Al content (and hence the nucleosynthesis yields, or supernova and star formation rates derived from it) could be lower by 30% or more, if foreground emission from specific massive-star clusters would bias our inferred scale height towards high values (Martin et al. 2009). Thus the real 26Al content may be on the low side of the uncertainty range quoted before (Diehl et al. 2006a). We exploit deeper observations made with INTEGRAL to follow up on these questions.

3. SPI observations and data analysis

Continued SPI observations of the inner ridge of the Galaxy have reached a depth where spatial separation of the 26Al emission begins to be possible. This allows spatially-resolved spectroscopy of the 26Al γ-ray line, towards determination of 26Al-enriched ISM characteristics for different regions of the Galaxy (see Wang et al. 2009). We use the same data as analyzed by Wang et al. (2009) to determine possible 26Al emission from massive star nucleosynthesis in the Scorpius-Centaurus region. These data include orbits 43–623 from 5 years of INTEGRAL observations, with a total exposure of 61 Ms. We analyze single-detector hits (SE) in an energy range embracing the 26Al line at 1808.63 keV, specifically 1785–1826 keV, in 0.5 keV bins. Results are derived through fitting the set of binned event spectra by the composite of forward-folded sky signal and background models, allowing for intensity adjustments of each of these components in each energy bin. Spectra are obtained for different regions of the sky, as the adopted spatial distributions for each region are fitted separately, though simultaneously, to the complete ensemble of observational data. The subdivision of the sky into different sky regions of interest is limited by the additional parameters introduced into the fit. We minimize these here by representing the sky with two components only, one for the large-scale Galactic emission, and one for the region of interest. Varying the region of interest, we analyze the quality of fits obtained, possible systematics and background artifacts, and the consistency of our results.

thumbnail Fig. 3

The γ-ray spectrum obtained from fitting a homogeneously-bright sky to to the first (top) and fourth (bottom) quadrants of the Galaxy. The lefthand figures show emission at intermediate positive latitudes 5°–20°, the righthand figures show emission from the bright Galactic disk in the latitude interval [−5°, 5°] (Wang 2007).

Open with DEXTER

As a guide towards our study of Sco-Cen emission, we first study separating Galactic-plane 26Al emission from emission at intermediate latitudes. We find significant excess of 26Al emission in the northern part of the Galactic plane (Fig. 3, lower-left graph). Here, an exponential disk has been divided in latitude at 5° and separately fitted in intensity. This indicates that in the fourth quadrant of the Galaxy, there is excess line emission at intermediate latitudes, if compared to the first Galactic quadrant. We refine this through restricting the extra emission component to the Sco-Cen region only, while the second sky emission model component includes the remainder of the sky.

As we include all data for best constraints on instrumental background, we also require an adequate model of the full sky in 26Al. We plausibly employ several alternatives: the 26Al sky as observed with COMPTEL and determined with Maximum-Entropy or Multi-Resolution Expectation Maximization algorithms for imaging deconvolution (Knödlseder et al. 1999b), but also geometrical models which had been found to adequately represent 26Al emission, from COMPTEL and SPI analysis (e.g. Knödlseder et al. 1999a; Diehl et al. 2006b). Examples of the latter are double-exponential disks with Galactocentric scale radius 3.5 to 4 kpc and scale heights 130 to 300 pc, the free-electron model derived from pulsar dispersion measurements by Taylor & Cordes (1993) (we use the NE2001 model, Cordes & Lazio (2002), with an imprinted vertical scale height of the exponential distribution above the plane of 180 or 140 pc), or the FIR emission from warm dust as measured with COBE/DIRBE across the sky (Bennett et al. 1992).

The above models based on non-γ-ray observations may not represent nearby 26Al from Sco-Cen: they characterize the large-scale Galactic 26Al contributions, because they trace different massive-star related phenomena with indirect connection to 26Al per invividual source. On the other hand, the COMPTEL map would obviously include 26Al from Sco-Cen, yet is mediated by COMPTEL’s instrumental response and the deconvolution algorithm to derive the map, and thus not reliable in this detail. We therefore prefer to use the COMPTEL 26Al map for the general large-scale 26Al emission, and here adopt its MREM variant (Knödlseder et al. 1999b). We cut out from it a region which lies above the plane of the Galaxy and towards the Sco-Cen sources here, defined as a circle centered at (l,b) = (350°, 20°) with a radius of 8°. 26Al is ejected at velocities of  ~1000–1500 km s-1 and slows down in interstellar medium from interactions with the gas; in our large-scale Galactic average, but also in the Cygnus region, we find velocities are certainly below 200 km s-1 from Doppler broadening limits on the observed line width (Diehl et al. 2006b; Martin et al. 2009). The size of our region of choice corresponds to 26Al coasting with interstellar velocities of 10 km s-1, hence will not include faster 26Al and miss part of the production in this region. We chose this conservatively-small region to avoid misinterpreting 26Al γ-rays overlapping directions towards the Galaxy’s disk. We test spatially-unstructured models representing the Sco-Cen region, such as a homogeneous-intensity circle centered at this location, or an exponential disk split into a part below 5° latitude containing the bulk emission and above 5° latitude containing emission from Sco-Cen, or the extracted part of a COMPTEL 26Al map suppressing small-scale structure. Detection significances vary between 5 and 7σ for the cases studied, derived fluxes typically vary by  ≤10% for different spatial models of the Sco-Cen region itself; this is well within statistical uncertainties.

The instrumental background created by cosmic-ray interactions with the spacecraft material dominates the measured counts, and only about 1% of the total counts is contributed by the signal from the sky. Therefore, background modeling is crucial, and artifacts from inadequacies of a background model may occur. We therefore also test for 26Al emission in other regions of the sky, where no such emission is plausibly expected. For most-representative tests, we chose regions of similar exposure and similar offset from the bright 26Al emission along the inner Galactic ridge, i.e., the region south of the plane of the Galaxy at longitude 350° and latitude −20°, and a region in the first quadrant of the Galaxy at (l,b) = (10°, 20°).

The characteristics of celestial 26Al emission, i.e. bulk motion and turbulence within the source region, translate into Doppler broadening and shift of the observed γ-ray line. SPI Ge detectors show an intrinsic energy resolution at 1809 keV of 3 keV (FWHM), variations of the energy calibration remain below 0.05 keV (Lonjou et al. 2004). Effects degrading the detector resolution with time due to cosmic-ray irradiation in orbit, and the periodic annealings which successfully restore spectral resolution, are all relevant for strong signal where line shape details matter, and can be neglected for our rather weak target. Here we determine line centroids and widths from Gaussian fits, and find using our more-accurate asymmetric response including degradation effects does not affects our findings. Our spectral result parameters are limited by statistical precision, and overall systematic uncertainties (from energy calibration stability and from using a Gaussian spectral response) are estimated as 0.1 keV.

4. Results

thumbnail Fig. 4

The γ-ray spectrum of the Sco-Cen region. This was obtained from fitting a 10° diameter patch of the sky representing candidate 26Al sources of this region together with a large-scale Galactic-emission model (here: MREM 26Al image as derived from COMPTEL observations) in narrow energy bins to SPI observation data.

Open with DEXTER

Using the COMPTEL 26Al map for this major portion of the sky, and an un-structured circular patch of the sky in the Sco-Cen region ((l,b) = (350°, 20°) with a radius of 10°), we obtain the Sco-Cen region spectrum shown in Fig. 4. The 26Al line appears in the spectrum, now much clearer and at a statistical significance of 6σ (from the likelihood ratio of the fit). The Sco-Cen signal robustly is found in all above-mentioned alternative spatial-distribution models for the large-scale emission.

This 26Al flux attributed to the Scorpius-Centaurus stars was estimated through variations of the candidate source region: we increase the circular region-of-interest radius from 4° to 20°, deriving 26Al line fluxes through 2-component spatial model fits. Figure 5 shows that the detected signal increases up to 10° radius, remaining constant for larger radii until the Galactic-plane emission is included (for the 20° case). None of our three reference regions shows such accumulation of signal with region size, their flux values fluctuate within uncertainties.

thumbnail Fig. 5

26Al flux increase with radius of our region of interest, centered on l =  −10°/b = 20°.

Open with DEXTER

thumbnail Fig. 6

26Al flux determined for different plausible large-scale model distributions (details see text).

Open with DEXTER

Including an emission component for this region attributed to nearby Scorpius-Centaurus stars, the fitted large-scale 26Al flux from the bright ridge along the inner Galaxy obtains lower values. In Fig. 6 we show a set of plausible models for the distribution of large-scale emission as discussed in Diehl et al. (2006a), as its flux values change through inclusion of the Scorpius-Centaurus component. We compare models based on free electrons, on infrared emission from dust, two analytical descriptions of the disk of the Galaxy, and the COMPTEL 26Al gamma-ray image (see details and references as described in Diehl et al. 2006a, Appendix). Separating the Scorpius-Centaurus emission component (shown as thin symbols), the inner-Galaxy flux values (thick symbols; we use [−30°  < l <  30°], [−10°  < b <  10°] approximating the region often referred to as “the inner radian”) vary hardly among models. Accounting for observed data with these same large-scale model alone (in an earlier analysis, Diehl et al. 2006a, Appendix), variations among models were evident. Separating 26Al emission from this foreground region, the fitted large-scale Galactic 26Al intensity decreases by 19% to 2.63 × 10-4 ph cm-2 s-1 rad-1, with a statistical uncertainty of  ± 0.2 × 10-4 ph cm-2 s-1 rad-1 from this analysis using about twice as much data. Also shown in Fig. 6 are the fluxes we derive for the 8° Scorpius-Centaurus region source together with each of these alternate models. Variations are small and within statistical uncertainties.

We thus derive an 26Al flux from our Scorpius-Centaurus region source of (6.1 ± 0.11) 10-5 ph cm-2 s-1. Flux uncertainties are 1.2 × 10-5 ph cm-2 s-1 (statistical) and 1.0 × 10-5 ph cm-2 s-1 (estimated systematics). (The variance of 0.2% for the different large-scale models provides an estimate of systematic uncertainties from this aspect. Main systematic uncertainty is derived from our tests for a signal in the mirror-symmetric regions at (l,b) = (−10°,  −20°) and (l,b) = (+10°,  +20°). No significant celestial intensity is found for any such alternative source, but the scatter of derived fluxes with different candidate source-region sizes results in this estimate of 1.0 × 10-5 ph cm-2 s-1. In these test we note that as now the Sco-Cen region is not part of the sky model, our extra component south of the plane obtains a negative-amplitude line at the position of the 26Al line, which probably compensates for the now-too-intense large-scale model as fitted to our data.)

This 26Al line flux from our Scorpius-Centaurus region source can be converted to an observed amount of 26Al, adopting a specific distance for the source; for a distance of 150 pc we obtain 1.1 × 10-4 M of 26Al.

We note that an independent all-sky analysis on a somewhat smaller dataset represented the spatial distribution of Sco-Cen emission through a 2-dimensional Gaussian and obtained (6.2 ± 1.6) × 10-5 ph cm-2 s-1 (Martin et al. 2009).

The 26Al line from our Scorpius-Centaurus region source is found with a slightly blue-shifted energy: From fitting the line shape with an asymmetric line shape as expected from variable spectral-response degradation between annealings, we determine a centroid energy of 1809.46 keV (±0.48 keV). When we compare to the 26Al-decay laboratory value for the γ-ray energy of 1808.63 keV, this implies a blue shift of  ~ 0.8 keV. If interpreted as a kinematic Doppler effect from bulk motion, this corresponds to about (137 ± 75) km s-1. Our uncertainties on the line centroid are  ± 0.4 keV statistical (for fixed-width Gaussians) and 0.2 keV systematic uncertainty; here systematic uncertainty is estimated from the line centroid difference between Gaussian and instrumental line-shape fits to the Sco-Cen signal. This translates into the quoted bulk velocity uncertainty of 75 km s-1.

In Fig. 4 and most of our analysis for weak 26Al signals, we rather use a Gaussian shape with fixed instrumental resolution to represent the line shape, in order to stabilize the fit from using fewer free parameters. Fitting the width of the observed line in such a Gaussian, we find a value of 2.1  ±  0.5 keV (FWHM), formally below but consistent with the instrumental line width of 3.0 keV (Roques et al. 2003) (which was used in the Gaussian line fitted in Fig. 4).

thumbnail Fig. 7

The 26Al content of the ISM around a group of coeval massive-star groups, as it evolves with time (Voss et al. 2009).

Open with DEXTER

5. Discussion

The “Scorpius-Centaurus” complex of massive stars is the most-nearby region showing star formation within the last 10–20 My, with typical distances below 150 pc from the Sun(de Geus 1992; de Zeeuw et al. 1999; Preibisch & Zinnecker 1999). Our detection of 26Al γ-rays confirms such recent and nearby massive-star activity. Since 26Al decay occurs with a time scale of 1 My, this implies that the massive stars providing the 26Al sources were born less than  ~ 10 My ago (see Fig. 7). We observe an intensity in the γ-ray line at 1808.63 keV from 26Al decay corresponding to a present 26Al mass of 1.1 × 10-4 M at 150 pc distance. The typical individual source-yield is about 1.1 × 10-4 M for massive stars in the 8–40 M range (e.g. Limongi & Chieffi 2006; Woosley & Heger 2007) (see summary in appendix of Diehl et al. 2006a), formally consistent with our derived 26Al mass. The most massive star in Upper Sco was presumably a  ~ 50M O5–O6 star, which may have exploded as a supernova about 1.5 Myr ago; the pulsar PSR J1932+1059 may be its compact remnant (Hoogerwerf et al. 2000; Chatterjee et al. 2004). But because of uncertainty of this event and no obvious supernova activity being evident in Sco-Cen less than 1 My ago (which would be seen e.g. as a supernova remnant), we are led to consider the observed 26Al being a product of massive-star nucleosynthesis from a number of sources during an earlier epoch, rather than from a single event.

The 26Al amounts ejected into the interstellar medium by a group of coeval massive stars evolves with stars developing Wolf-Rayet winds and finally exploding as supernovae (see Fig. 7, from our population synthesis model; Voss et al. 2009). From the number of stars in the Upper Sco group of stars (using 16 stars presently at masses M > 8 M, Preibisch et al. 2002) and an adopted age of 5 My, this population synthesis model predicts an intensity in the 26Al γ-ray line of 7 × 10-5 ph cm-2 s-1 from 1.2 × 10-4 M of 26Al, which is consistent with our measurement. This would support age estimates derived from isochrone fitting to stellar groups (which are somewhat uncertain mainly from the calibration of pre-main-sequence models), and thus support Upper Sco as a region of most recent star formation, with a group age of about 5 My.

When we compare our observed 26Al line centroid to the 26Al-decay laboratory value for the γ-ray energy of 1808.63 keV, the evident blue shift of  ~ 0.8 keV corresponds to a kinematic Doppler effect from bulk motion at about (137 ± 75) km s-1. From our line width analysis we regard as unlikely additional astrophysical Doppler broadening from turbulent or isotropic motion at higher than normal interstellar gas velocities (i.e.,  < 100 km s-1, which would correspond to 30% astrophysical line broadening).

The cavity surrounding Upper Sco appears to expand with a velocity of about 10 ( ± 2) km s-1, which, together with the relative motion of Sco-Cen stars with respect to the Sun of about 5 km s-1 implies an approaching relative velocity of the entire complex of few tens of km s-1 at most. Our measurement of a substantially-higher approaching velocity of 26Al nuclei is incompatible with this. We would thus exclude 26Al to decay while being associated with cavity walls: Decaying 26Al moves at higher velocity. At the time of ejection, velocities of stellar wind or supernova material containing 26Al would be about 1000–1500 km s-1. This velocity, if sustained over a 26Al lifetime, would imply a propagation distance from the source of 1.3 kpc, and source ejecta would fill a sphere of that size, appearing isotropic from our position of the Sun (offset from the center of the sphere by only 0.1 its radius). But, with our analysis circle of 10° diameter around the source, we also impose a bias for 26Al being near its source while decaying, sampling its effective streaming velocity over this volume sampled by our analysis.

Thus, our measured bulk velocity appears consistent with the small or absent 26Al line broadening on one hand, and also with 26Al being slowed down significantly from its high velocity at ejection, on a scale of 10 pc. This appears consistent with our adopted scenario of observing stellar ejecta originating in the Sco-Cen stars and streaming into the pre-blown cavity around it, decelerating through turbulence and interactions with the cavity walls. Other inferences of the morphology of interstellar gas around the Sun have independently suggested and supported this view (Frisch 1995). In particular, a streaming motion of interstellar gas from the general direction of Sco-Cen has been identified (also referred to as the local interstellar wind (LISW); see, e.g., Lallement et al. 2003; Frisch 1995). It also implies that our choice of a 10°-sized region to identify 26Al from Sco-Cen may miss a sigificant part of the emission, as the region filled with 26Al by Sco-Cen should be more extended on the sky. We cannot detect such more extended emission, both because it may be weaker than the main feature we report here, and because other locations of Scorpius-Centaurus stars overlap with Galactic-disk viewing directions. We will pursue this aspect in a refined modeling approach based on details of the observed stellar population nearby (Ohlendorf et al., in prep.).

Further exploring this question of how the nearby regions of the Galaxy have evolved, we may study the broader history of the Scorpius-Centaurus region: Stellar subgroups of different ages would result from a star forming region within a giant molecular cloud if the environmental effects of massive-star action of a first generation of stars (specifically shocks from winds and supernovae) would interact with nearby dense interstellar medium, leading to propagating or triggered star formation. Later-generation ejecta would find the ISM pre-shaped by previous stellar generations. Such a scenario was proposed for the Scorpius-Centaurus Association, based on its different subgroups (de Geus et al. 1989; Preibisch et al. 2002) and several stellar groups around it (e.g. Preibisch & Mamajek 2008; Fernández et al. 2008). Indications of recent star formation have been found in the L1688 cloud as part of the ρ Oph molecular cloud, and may have been triggered by the winds and supernovae which produced the 26Al we observe. The young ρ Oph stars then could be interpreted as the latest signs of propagating star formation originally initiated from the oldest Sco-Cen subgroup in Upper Centaurus Lupus (Wilking et al. 2008). The origin of runaway stars and nearby pulsars also provides hints towards a recent supernova in the solar vicinity (see, e.g. Hoogerwerf et al. 2000; Preibisch et al. 2002).

Many proposed scenarios of triggered star formation are only based on relatively weak evidence, such as the presence of Young Stellar Objects (YSOs) near shocks caused by massive stars. Positional evidence alone is not unequivocally considered to prove triggered star formation. Much more reliable conclusions can be drawn if the ages of the young stellar populations can be determined and compared to the moment in time at which an external shock from another star formation site arrived. Agreement of these timings would add important arguments. Radioactive clocks such as 26Al can provide such information in this context in a most-direct way, being driven by physics which is rather independent of the environmental conditions of the dynamic ISM. Therefore, we aim to constrain 26Al emission from the presumably-older subgroups of the Sco-Cen association in a subsequent study (Ohlendorf et al., in prep.).

Supernovae are also expected to have far-reaching implications such as a biologically-relevant UV flash and enhanced cosmic-ray intensity which may be relevant for climatic changes (as discussed, e.g., in Lallement 2007; Svensmark 2007). It is interesting that the discovery of 60Fe isotopes in Pacific ocean crust material on Earth (Knie et al. 2004) provides strong evidence of a nearby supernova event about 2–3 My ago, which may be attributed to Sco-Cen stars as well. It will remain exciting to obtain refinement in correlated investigations of observables from such nearby cosmic violence.

We note that inclusion of nearby 26Al from Scorpius-Centaurus stars in the determination of the large-scale 26Al emission from the Galaxy reduces the large-scale flux value by 19%. This translates directly into a corresponding reduction of the Galaxy’s total 26Al mass from 2.8 to 2.25 M, propagating to inferences based on this 26Al mass such as the core-collapse supernova rate and the interstellar isotopic ratio 26Al/27Al (see Diehl et al. 2006a). Note that our earlier analysis accounted for this aspect of large-scale source distribution uncertainty through quoting a systematic uncertainty of 30%, and specifically for the Galaxy’s 26Al mass a value of (2.8 ± 0.8) M. As shown recently by Martin et al. (2009) with a ring model of 90 pc scale height for Galactic sources and specific model components for the Cygnus and Sco-Cen regions, the 26Al mass in the Galaxy may well be even below 2 M (they determine (1.7 ± 0.2) M) for this geometry; note however that Wang et al. (2009) find an 26Al emission scale height of 130 pc). With now a clear detection of foreground 26Al emission from the Scorpius-Centaurus stellar group(s), we are in a better position to establish improved 3-dimensional 26Al source distribution models for our Galaxy to reduce the line-of-sight uncertainties in source distances, accounting for the roles of candidate stellar groups such as Cygnus and Scorpius-Centaurus, but also Gould Belt associations and associations in spiral-arm regions such as Carina. Together with reduced statistical uncertainty from more observational data, this will result into a more realistic estimate of the Galaxy’s 26Al mass and derived quantities. Such a study is beyond the scope of this paper.

In conclusion, our observation of 26Al from the Sco-Cen region strengthens evidence for nearby massive star activity in the recent past (about 20 My or less). With improved observations, we aim at spatially-resolved study of the 26Al emission from this region, to discriminate among Sco-Cen subgroups and their 26Al contributions, and to add diagnostics of massive-star activity in the solar vicinity.

Acknowledgments

Part of this work was supported by the Munich Excellence Cluster “Origins and Evolution of the Universe”. The SPI project has been completed under the responsibility and leadership of CNES/France.We are grateful to ASI, CEA, CNES, DLR, ESA, INTA, NASA and OSTC for support. The SPI anticoincidence system is supported by the German government through DLR grant 50.0G.9503.0. We thank the anonymous referee for constructive suggestions which helped to improve this paper.

References

All Figures

thumbnail Fig. 1

OB associations near the Sun, with Sco-Cen stars marked (from Preibisch & Mamajek 2008). This top view onto the Galaxy in the local vicinity shows that several nearby groups of massive stars can be identified, with the Scorpius-Centaurus group as the closest to the Sun. This compilation is based on Hipparcos position measurements of nearby stars (for details see Preibisch & Mamajek 2008).

Open with DEXTER
In the text
thumbnail Fig. 2

All-sky image (top) and expanded view towards Sco-Cen (bottom) of 26Al γ-rays as derived from COMPTEL observations (Plüschke et al. 2001).

Open with DEXTER
In the text
thumbnail Fig. 3

The γ-ray spectrum obtained from fitting a homogeneously-bright sky to to the first (top) and fourth (bottom) quadrants of the Galaxy. The lefthand figures show emission at intermediate positive latitudes 5°–20°, the righthand figures show emission from the bright Galactic disk in the latitude interval [−5°, 5°] (Wang 2007).

Open with DEXTER
In the text
thumbnail Fig. 4

The γ-ray spectrum of the Sco-Cen region. This was obtained from fitting a 10° diameter patch of the sky representing candidate 26Al sources of this region together with a large-scale Galactic-emission model (here: MREM 26Al image as derived from COMPTEL observations) in narrow energy bins to SPI observation data.

Open with DEXTER
In the text
thumbnail Fig. 5

26Al flux increase with radius of our region of interest, centered on l =  −10°/b = 20°.

Open with DEXTER
In the text
thumbnail Fig. 6

26Al flux determined for different plausible large-scale model distributions (details see text).

Open with DEXTER
In the text
thumbnail Fig. 7

The 26Al content of the ISM around a group of coeval massive-star groups, as it evolves with time (Voss et al. 2009).

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.