- Split View
-
Views
-
Cite
Cite
Jae-Joon Lee, Bon-Chul Koo, Yong-Hyun Lee, Ho-Gyu Lee, Jong-Ho Shinn, Hyun-Jeong Kim, Yesol Kim, Tae-Soo Pyo, Dae-Sik Moon, Sung-Chul Yoon, Moo-Young Chun, Dirk Froebrich, Chris J. Davis, Watson P. Varricatt, Jaemann Kyeong, Narae Hwang, Byeong-Gon Park, Myung Gyoon Lee, Hyung Mok Lee, Masateru Ishiguro, UKIRT Widefield Infrared Survey for Fe+, Monthly Notices of the Royal Astronomical Society, Volume 443, Issue 3, 21 September 2014, Pages 2650–2660, https://doi.org/10.1093/mnras/stu1146
- Share Icon Share
ABSTRACT
The United Kingdom Infrared Telescope (UKIRT) Widefield Infrared Survey for Fe+ (UWIFE) is a 180 deg2 imaging survey of the first Galactic quadrant (7° < l < 62°; |b| ≲ 1| $_{.}^{\circ}$|5) that uses a narrow-band filter centred on the [Fe ii] 1.644-μm emission line. The [Fe ii] 1.644-μm emission is a good tracer of dense, shock-excited gas, and the survey will probe violent environments around stars: star-forming regions, evolved stars, and supernova remnants, among others. The UWIFE survey is designed to complement the existing UKIRT Widefield Infrared Survey for H2 (UWISH2). The survey will also complement existing broad-band surveys. The observed images have a nominal 5σ detection limit of 18.7 mag for point sources, with a median seeing of 0.83 arcsec. For extended sources, we estimate a surface brightness limit of 8.1 × 10−20 W m−2 arcsec−2. In this paper, we present an overview and some preliminary results of this survey.
1 INTRODUCTION
Stars, especially massive ones, heavily influence the surrounding interstellar medium (ISM), both through stellar radiation and often through dynamical interaction, for example through jets and outflows from protostars and mass loss from evolved stars. One of the most extreme cases is the explosion of supernovae (SNe). They heat and excite the interstellar gas, and cause turbulent motion in the ISM. ‘Mass return’ from stars to the ISM is particularly important, as it contributes to the metal enrichment of the Universe. The study of how stars interact with their surroundings is thus of crucial importance for an understanding not only of the formation and evolution of stars but also of the evolution of galaxies.
Emission lines are informative probes of processes associated with the birth and death of stars. The most popular are hydrogen Balmer lines in the optical. Balmer lines are often emitted by gas ionized by young hot stars, and thus are used as a measure of the current star formation rates of galaxies (Kennicutt 1998). They have played a critical role in the study of Galactic objects, and large-scale imaging surveys of the Galactic plane in the Hα line (e.g. Parker et al. 2005; Drew et al. 2005) have constituted a major resource for various studies. The use of Balmer lines is limited for distant or embedded objects, however, owing to the relatively high extinction in the visual band.
With the advent of large-format infrared (IR) detectors, various imaging surveys of lines in near-infrared (NIR) bands have been carried out. The United Kingdom Infrared Telescope (UKIRT) Widefield Infrared Survey for H2 (UWISH2; Froebrich et al. 2011, F2011 hereafter) is one such survey that probes H2 1–0 S(1) emission at 2.122 μm. Using the Wide-Field Camera onboard the 3.8-m UKIRT, the survey covered the northern portion of the area covered in the GLIMPSE survey (Churchwell et al. 2009). H2 emission often traces outflows and jets from embedded young stars and the regions around massive stars that are radiatively excited.
Other prominent emission lines in the NIR are those from Fe+. The Fe+ ion has four ground terms, each of which has 3–5 closely spaced levels, forming a 16-level system (Pradhan & Nahar 2011). The energy gap between the ground level and the excited levels is less than 1.3 × 104 K, making these levels easily excited in the post-shock cooling region. The emission lines resulting from the transitions among the different levels appear in the optical to far-infrared bands. In the NIR JHK bands, 10–20 [Fe ii] lines are visible, with the two strongest lines at 1.257 and 1.644 μm (Koo 2014).
The [Fe ii] lines are stronger in shocked gas than in photoionized gas because Fe atoms in photoionzed gas are probably in higher ionization stages, and also because the Fe abundance in shocked gas is probably enhanced by dust destruction (Koo 2014). Behind radiative atomic shocks, [Fe ii] emissions originate mostly from the cooling region. For example, the ratio of [Fe ii] 1.644 μm to the hydrogen Paβ line in supernova remnants (SNRs) is between 2 and 7, while it is as low as 0.013 in star-forming regions in Orion (Oliva, Moorwood & Danziger 1989; Mouri, Kawara & Taniguchi 2000). Therefore, [Fe ii] forbidden lines could be used as tracers of fast radiative atomic shocks, although they can be emitted from photoionized gas when ionized by high-energy photons such as X-rays, for example in active galactic nuclei (Mouri et al. 2000).
Here we present an unbiased survey of a portion of the Galactic plane in [Fe ii] 1.644-μm emission, known as the UKIRT Widefield Infrared Survey for Fe+, or UWIFE. The aim of the survey was to detect ionized Fe objects (IFOs), where the [Fe ii] line highlights regions of predominantly shock-excited dense gas. [Fe ii] emission is an excellent tracer of dense outflows in young and massive star-forming regions, of eruptive mass loss from evolved stars, and of dense media interacting with SNRs. The survey was designed to complement and cover the same region as the UWISH2 survey. The UWIFE survey was conducted with the same telescope and detector combination as in the UWISH2 survey, but with a different filter. Furthermore, the survey complements the UKIDSS Galactic Plane JHK survey (GPS; Lucas et al. 2008), and various other surveys such as the IPHAS Hα survey (Drew et al. 2005) and the VLA 5-Ghz CORNISH survey (Hoare et al. 2012), etc.
The UWIFE survey made initial observations in the summer of 2012 and was completed in September 2013. In this paper, we define the survey in Section 2 and describe its characteristics in Section 3. Our survey is briefly compared with Hα surveys in Section 4. In Section 5 we outline the scientific objectives of the survey, including some preliminary results when available. Section 6 concludes the paper.
2 TARGET AREA AND OBSERVATIONS
The UWIFE survey was conducted using the Wide-Field Camera (WFCAM, Casali et al. 2007) at UKIRT. The camera has four Rockwell Hawaii-II HgCdTe arrays. Each array has 2048 × 2048 pixels, corresponding to a 13.65 arcmin × 13.65 arcmin field-of-view with a pixel scale of 0.4 arcsec. The arrays are arranged in a square pattern, with gaps of 12.83 arcmin between adjacent arrays. With this layout, observing at four discrete positions results in a contiguous area covering 0.75 deg2 on the sky (a WFCAM tile). For each pointing, images were obtained at three jitter positions. The jitter offsets were (0 arcsec, 0 arcsec), (6.4 arcsec, 0 arcsec) and (6.4 arcsec, 6.4 arcsec). We performed a 2 × 2 microstep about each jitter position, with an offset size of 4.62 arcsec, in order to fully sample the point spread function. Three jitter positions with 2 × 2 micro-stepping give a total of 12 images. An exposure time of 60 s was used, giving a total per-pixel integration time of 720 s. The final stacked images were resampled to 0.2 arcsec.
The survey covered a region within the First Galactic Quadrant (7° ≲ l ≲ 62°; |b| ≲ 1.5°). The coverage is nearly identical to that of the UWISH2 survey, except that the UWISH2 covered up to l ≃ 65°. Fig. 1 shows an overview of the UWIFE survey area in the Galactic plane and compares it with the UWISH2 survey area. The survey area consists of 220 tiles, where a single tile is a square of 54 arcmin×54 arcmin, aligned in equatorial coordinates. The tiles are arranged as 55 stripes of four consequent tiles along lines of constant declination (see Fig. 2). The positions of tiles are identical to those of UWISH2, except for 20 tiles in 43° < l < 48°. For these tiles, the coordinates of UWISH2 differ from those of the GPS by a few arcminutes (they are identical otherwise), and we adopted the coordinates of the GPS. There are small gaps around l ∼ 16° not covered by the survey. The gaps are inherited from the survey area of UWISH2. However, we consider the impact of the gaps on the survey to be minimal.
3 RESULTS
3.1 [Fe ii] filter
The [Fe ii] 1.644-μm filters were not available on WFCAM, so a set of new filters were procured from JDS Uniphase Corporation and installed during the summer of 2012. The filters have a central wavelength of 1.644 μm with a transmittance of 85 per cent and effective bandwidth of 0.026 μm. Fig. 3 plots the filter response curve of the [Fe ii] filter provided by the manufacturer, together with that of H-band filter. Fig. 4 shows the first light image obtained using the newly installed filter, on the SNR G11.2−0.3. The image is continuum-subtracted, as described in Section 3.3. The figure shows a comparison of our image with the [Fe ii] image of the same target obtained with the Wide-field Infrared Camera (WIFC) onboard the Palomar Hale telescope (Koo et al. 2007). Our UKIRT WFCAM image is basically identical to that of the WIRC, but shows details that were not available in the WIRC image.
3.2 Survey status and data quality
The UWIFE observations were conducted during 2012 and 2013, and we observed a total of 220 tiles. The images were reduced by the Cambridge Astronomical Survey Unit (CASU), using the same pipeline as was used to reduce the images of UWISH2 and GPS. The CASU reduction steps are described in detail by Dye et al. (2006); astrometric and photometric calibrations (Hodgkin et al. 2009) were carried out using the 2MASS catalogue.
Fig. Fig. 5 displays the airmass, seeing and limiting magnitudes for 220 tiles (one tile consists of 16 images). The values are shown for each image of a single chip. For airmass values, four images in the same sequence share the same value, resulting in a particular pattern in the plot. The seeing values were obtained from the CASU pipeline products, and measured from the co-added frames. The statistics of these values are shown in Fig. 6 as histograms. The majority of the data have a seeing of between 0.6 and 1.0 arcsec. The median seeing is 0.82 arcsec, which is slightly worse than that of the UWISH2 survey (0.73 arcsec). Virtually all the images have seeing better than 1.5 arcsec.
3.3 Continuum subtraction
The observed [Fe ii] images were continuum-subtracted using the H-band images from the GPS. The H-band images were regridded and rescaled to match the astrometry and the flux scale of the corresponding [Fe ii] images. The point spread functions (PSFs) of [Fe ii] images and corresponding H images often differ significantly. To compensate for the effect of the different PSFs, the image of the better seeing can be convoluted to match the poorer seeing of the other image. However, this approach is not optimal, as the resulting difference image have the PSF of the poorer seeing. This is an important issue for [Fe ii] emission, as [Fe ii] emission is often knotty and compact. To keep the original seeing of the [Fe ii] emission, we adopted an alternative continuum-subtraction method in which we subtracted point-like sources from the images by modelling the PSF. Below, we describe this method in greater detail.
We first re-project the H-band image onto the corresponding [Fe ii] image so that the two images are astrometrically aligned.
We then extract 51 × 51 pixel areas around well-isolated stars whose fluxes are greater than 30σ above the background. They are normalized and median-averaged to form the PSF image. To compensate for the spatial variation of PSF across the focal plane, a single chip area (4096 × 4096 pixels) is divided into subregions and PSFs are estimated in each subregion. It was found that 4 × 4 subregions are usually adequate.
We perform PSF photometry. Among the detected sources, we reject sources that are not point-like (sources with a PSF correlation coefficient of less than 0.7). The source catalogues of [Fe ii] and H-band images are matched and we further reject sources that have a brightness difference greater than 20 per cent and/or a position difference larger than 0.5 pixels. With the second step we distinguish between continuum point sources and unresolved line-emission features.
We reconstruct synthetic images using the source catalogue and the PSFs. These synthetic images are subtracted from the original images. This step is carried out for both [Fe ii] and H-band images, and it removes most of the point-like continuum sources from both images.
We then subtract the point-sources-removed H-band images from the point-sources-removed [Fe ii] images. This is to remove other extended continuum sources that escaped detection in previous steps.
Fig. 7 shows the flowchart of our continuum-subtraction method. We primarily used Starfinder (Diolaiti et al. 2000) for the PSF photometry, together with Scamp (Bertin 2006) and SWarp (Bertin et al. 2002) for the astrometry and the image reprojection.
We note that the difference image between the [Fe ii] and H-band images may have contamination other than [Fe ii] emission. The GPS images were taken during 2005–2008, and the difference may be caused by source variability over a few years or more, in particular for point sources. It is also possible that there are some, but not significant, contributions from lines other than [Fe ii] 1.644 μm, or a differing continuum slope. For example, the hydrogen Br 12 line is within the bandwidth of our [Fe ii] filter. However, we believe its contribution in the continuum-subtracted images is not significant as the H-band images contain Br 10–18 lines and a good fraction of the Br 12 line is subtracted when the H-band images are subtracted.
3.4 Data release
All the images from the survey and the continuum-subtracted images will eventually be made public through our survey web page (http://gems0.kasi.re.kr/uwife). At the time of writing, all the CASU pipeline products are available, together with most of the continuum-subtracted images. The web page also provides pre-generated RGB composite images (UWISH2, UWIFE and GPS J for red, green and blue, respectively) of the entire survey region that can be zoomed down to a resolution of 0.4 arcsec.
4 COMPARISON WITH Hα SURVEYS
The [Fe ii] forbidden lines, in particular the line at 1.644 μm, are good tracers of fast radiative atomic shocks. The intrinsic emissivity of the [Fe ii] line at 1.644 μm from atomic shocks could be fainter than other tracers in optical bands (e.g. Hα) in general. However, this can easily be compensated for by the lower extinction in the infrared for embedded objects, or for those at large distances from us.
Fig. 8 shows model-calculated intrinsic line ratios of Hα to [Fe ii] 1.644 μm for various shock conditions, assuming solar abundances adopted from the shock grids of Allen et al. (2008). The models suggest that the Hα emission is intrinsically brighter than the [Fe ii] emission by an order of magnitude for interstellar shocks of a given shock velocity range of 100 to 500 km s−1. The ratio can become as high as ∼50 for slower shocks and as low as ∼5 for faster shocks.
The line ratios given above do not take the interstellar extinction into account. Owing to the lower extinction in the infrared, compared with that in the optical, the [Fe ii] lines can appear brighter than Hα for objects with high foreground absorption. We can define the amount of differential extinction between [Fe ii] and Hα for a given Av as |$10^{-2.5 \, [\mathrm{A}(1.644\,\mu {\rm m}) - \mathrm{A}(0.656\,\mu {\rm m})]}$|. For example, we obtain ∼30 for an Av of 6 mag. This means that, although [Fe ii] is intrinsically fainter by a factor of 30, [Fe ii] look as bright as Hα for an Av of 6 mag. According to Fig. 8, the intrinsic line ratios of Hα to [Fe ii] 1.644 μm range from 5 to 50, and it is expected that [Fe ii] will be apparently brighter than Hα for objects with Av higher than several magnitudes. A visual extinction of 6 mag is expected for a nominal hydrogen column density of NH = 1022 cm−2; thus, the [Fe ii] 1.644-μm line is often brighter than Hα, particularly for objects in the inner galaxy.
The above comparison does not take into account the sensitivity of the underlying detector system. Because of the high sky background in the NIR compared with that in the optical, a better sensitivity is usually obtained at optical wavelengths, although this depends on various other factors such as the telescope diameter. One of the surveys that complements our [Fe ii] survey is the INT/WFC Photometric Hα Survey of the Northern Galactic Plane (IPHAS), which is an Hα imaging survey being carried out using the 2.5-m Isaac Newton Telescope. The IPHAS survey has an Hα surface brightness limit of 10−20 W m−2 arcsec2 (Sabin et al. 2013), which is comparable to the surface brightness limit of our survey. Therefore, the UWIFE survey will have an advantage over the IPHAS survey for sources with Av greater than several magnitudes. Fig. 9 shows one example at AV ∼ 10 mag where we see bright [Fe ii] emission but an absence of Hα emission.
5 DISCUSSION
5.1 Jets/outflows from young stars
Outflows are ubiquitous in low- to high-mass star formation (Shepherd & Churchwell 1996), so they are significant signposts in searches for star-forming regions. Since the discovery of Herbig–Haro (HH) objects (Herbig 1950; Haro 1952) in the optical, many different types of outflows have been reported over a wide range of wavelengths, from X-ray to radio (Lada 1985; Bachiller 1996; Reipurth & Bally 2001; Shepherd 2005; Bally, Reipurth & Davis 2007). Each emission line in the various wavelength ranges from X-ray to radio shows different aspects of the physical and chemical conditions in the shocks. In the near-infrared, [Fe ii] and H2 emission lines are good tracers for outflows and jets and have often been observed in low-mass star-forming regions (Davis & Eisloeffel 1995; Reipurth et al. 2000; Davis et al. 2003; Hayashi & Pyo 2009). Generally, [Fe ii] emission traces partially ionized, fast, and dissociative J-shocks, while H2 emission traces neutral, slow, C-shocks or non-dissociative shocks (Hollenbach & McKee 1989; Smith 1994). These two emission lines show different spatial distributions and properties, as demonstrated in Fig. 10. While the [Fe ii] emission is confined in narrow jets or relatively compact knots at the tips of jets, H2 emission can be more extended. The fast, well-collimated [Fe ii] jets close to the driving sources provide a clue regarding the launching mechanism (Pyo et al. 2003, 2006, 2009).
In the UWIFE survey region there are thousands of small star-forming regions (Avedisova 2002). There are many objects related to the outflow phenomena detected at various wavelengths: 202 MHOs (molecular hydrogen emission-line objects: Davis et al. 2010), 106 EGOs (extended green objects in the GLIMPSE survey: Cyganowski et al. 2008), 55 CO outflows (Wu et al. 2004), 239 6.7-Ghz methanol masers (Pestalozzi, Minier & Booth 2005), 57 Class I methanol masers (Val'tts & Larionov 2007) and 170 OH masers (Szymczak & Gérard 2004). On the other hand, the number of HH and similar objects is relatively small in this region. Most of the HH objects are concentrated in the region 100°< l < 220°, and not many sources are reported in our survey region: 7 HH objects (Reipurth 2000) and 4 HBC sources (Herbig-Bell Catalog; Herbig & Bell 1988). UWIFE–SYSOP (UWIFE–Searching for the Young Stellar Outflows Project) is a project to find IFOs associated with young stellar objects (YSOs). In this project, we not only will compile a general catalogue of the IFOs but also will study the statistics of outflows and their physical quantities associated with the evolutionary stages of YSOs.
5.2 Massive-star-forming regions
Many aspects of the formation of massive stars (M ≳ 8 Mȯ) are still unclear (Zinnecker & Yorke 2007). One unanswered question is how massive stars obtain their mass. It has been increasingly reported that, as in low-mass stars, the disc-mediated accretion process seems to be occurring in the formation of massive stars (e.g. Beuther et al. 2002; Wu et al. 2004; San José-García et al. 2013; Cooper et al. 2013). However, it is still uncertain if such disc accretion works in the mass range M ≳ 25 Mȯ (Zinnecker & Yorke 2007). One way to understand the accretion process is to trace the outflow features, because the outflow process is closely related to the disc accretion process (cf. Section 5.1). In the following sections, we describe how UWIFE will contribute to the elucidation of massive star formation through the investigation of jet and outflow phenomena.
5.2.1 Outflow features in infrared dark clouds
Infrared dark clouds (IRDCs), seen silhouetted against the bright Galactic background in the mid-infrared (MIR), are cold (<25 K) and very dense (|$n_{\rm H_2} > 10^5\, {\rm cm}^{-3}$|) interstellar clouds with high column densities (∼1023–1025 cm−2; Simon et al. (2006a, and references therein)), and are thus believed to be probable sites for the formation of massive stars. Owing to their high extinction, IRDCs have usually been studied at longer wavelengths (e.g. far-infrared or millimetre) to investigate their pre-stellar or starless cores and associated filamentary structures (e.g. Rathborne, Jackson & Simon 2006; Jackson et al. 2010; Wilcock et al. 2012). Although most IRDCs or their cores (>50 per cent) do not show a signature of ongoing star formation in the MIR, there are some cores with bright MIR stellar sources indicating star formation activity (Chambers et al. 2009). These star-forming cores are, however, still deeply embedded in a cloud with high extinction, and few studies are made towards IRDCs to search for outflow features in the NIR. Therefore, it is worthwhile to examine the data from an unbiased [Fe ii] survey data to detect outflows in IRDCs, particularly in evolved IRDCs that show active star formation in order to understand the process of high-mass star formation in an early evolutionary phase.
One such cloud is the IRDC located at(l, b) ∼ (53°2, 0°0), which shows a number of bright MIR stellar sources, probably YSOs, along its long, filamentary structure; here, we refer to it as IRDC G53.2. The UWISH2 survey data reveal ubiquitous outflow features around YSO candidates in IRDC G53.2 (Fig. 11, left). As noted in other recent studies (e.g. Davis et al. 2009), the H2 outflows are particularly identified around earlier classes of YSOs (e.g. Class I and Flat Class), which contribute ∼50 per cent of the YSO candidates in IRDC G53.2, based on the analysis of Spitzer data (Kim et al. in preparation). In the UWIFE image of IRDC G53.2, only a few [Fe ii] emissions have been detected so far. The inset in Fig. 11 (left) shows [Fe ii] emission identified around a YSO candidate in the IRDC. A point-like knot shows H2 emission as well, but with a small shifted peak between the [Fe ii] and H2 emissions. A preliminary analysis suggests that the number of identified [Fe ii] emissions is much smaller than that of H2 emissions in the same area.
An investigation of the characteristics of [Fe ii] emissions associated with YSOs in IRDCs in the entire UWIFE survey area, which includes ∼300 IRDCs with GRS counterparts (Simon et al. 2006b), will provide much insight into the nature of jets and outflows of YSOs. In particular, comparative studies between the IRDCs and low-mass star-forming regions will be important. We note that there are already ongoing or planned UWISH2 projects to search for outflows/jets towards massive-star-forming regions or filamentary clouds in H2 (Froebrich et al. 2011; Ioannidis & Froebrich 2012; Lee et al. 2012). The UWIFE survey, therefore, will provide good complementary data to deepen our understanding of the physical structure of the outflows of high-mass protostars.
5.2.2 Outflow features around ultracompact H ii regions
Ultracompact H ii regions (UCHII) are very compact H ii regions, whose typical size, density and emission measures are ≲ 1017 cm, ≳ 104 cm−3 and 107 pc cm−6, respectively (Churchwell 2002). They are thought to be ‘late’ stages of a massive young stellar object (MYSO), no longer accreting significant mass (Churchwell 2002; Zinnecker & Yorke 2007). More specifically, they are the stage between the rapid accretion phase, when the central object is being formed, and the ionized phase, when the larger, more diffuse, less obscured H ii region is being produced. The unique position of UCHIIs on the MYSO evolutionary track enables us to study the history of accretion processes as follows.
Because UCHII natal clumps are not yet completely destroyed, one might expect that the materials ejected during the prior, active accretion phase produced shocked outflow features around UCHIIs. By tracing these ‘footprint’ outflow features around UCHIIs, we can study the history of the accretion process, because the outflow process is closely related to the accretion process (cf. Section 5.1). These features can be observed through radiative cooling lines, such as [Fe ii] 1.64 μm, H2 2.12 μm, and CO radio lines (Hollenbach & McKee 1989; Neufeld & Dalgarno 1989; Kaufman & Neufeld 1996; Wilgenbus et al. 2000; Flower & Pineau Des Forêts 2010). Indeed, CO outflow features have been observed around UCHII regions, such as G5.89-0.39 (Watson et al. 2007; Wood & Churchwell 1989) and G18.67+0.03 (Cyganowski et al. 2012). The dynamical time-scale of these CO outflow features is >103 yr, which is comparable to or greater than the typical lifetime of UCHIIs ( ≳ 5 × 104 yr, Wood & Churchwell 1989; González-Avilés, Lizano & Raga 2005) and the MYSO jet phase (∼4 × 104 yr, Guzmán et al. 2012).
A search for [Fe ii] features around UCHIIs is underway using a UCHII catalogue. The catalogue is from the CORNISH survey (Hoare et al. 2012; Purcell et al. 2013), which covers the same part of the Galactic plane as UWIFE. Shinn et al. (in preparation) have found [Fe ii] outflow features around several UCHIIs. An example region, where two UCHII candidates G025.3824–00.1812 and G025.3809–00.1815 are located, is shown in Fig. 12, where the south-westward [Fe ii] outflow/jet feature is evident. The outflow feature seems to emerge from the direction towards the two UCHIIs and has a well-collimated appearance. The [Fe ii] features around UCHIIs detected from the unbiased UWIFE survey enable us to extract some general properties of the MYSO accretion process. The outflow morphology, the outflow mass-loss rate, and their relation with the MYSO physical parameters would be the initial outcomes (cf. Shinn et al. 2013). Based on those results, we can investigate if the outflow features support the disc accretion process, and may be able to infer the final stellar mass in cases where the disc accretion seems to operate.
5.3 Nebulae around evolved stars
Circumstellar nebulae around evolved stars are fossil records of their mass-loss history. They often represent dense structures resulting from sudden changes in wind characteristics. These circumstellar nebulae are often known from their optical emission lines, but infrared emission from the cold and dense circumstellar medium is becoming more important. In principle, emission in different regimes provides us with a more complete view of the progenitor star system and its evolution.
5.3.1 Planetary nebulae
Planetary nebulae (PNe) are the most frequent type of nebula around evolved stars. They consist of material ejected during the asymptotic giant branch (AGB) phase that is ionized by the UV radiation from the newly formed hot central star (Frew & Parker 2010). More than a thousand PNe are currently known in our Galaxy (e.g. Acker et al. 1992). Most of them are high above the Galactic plane, however, and the number of PNe along the Galactic plane is relatively small. With the advent of sensitive Hα surveys of the Galactic plane such as IPHAS (Drew et al. 2005) and VPHAS+, a large number of PNe in the plane are being revealed (Viironen et al. 2009a, 2009b; Sabin et al. 2010). Furthermore, NIR line surveys such as UWISH2 and UWIFE have the potential to uncover a significant population of PNe along the Galactic plane. These surveys will help us to constrain their space density and any variation along the plane that is related to various routes of PN formation. In addition, the findings of numerous new PNe will enable more rigorous statistical analysis.
PNe are often bright in the H2 2.122-μm line. Kastner et al. (1996) detected H2 emission from 23 out of 60 PNe in their narrow-band imaging survey of 2.122-μm emission. The 2.122-μm H2 emission is most prominent in PNe of bipolar morphology, and Kastner et al. (1996) suggested a physical connection. On the other hand, surveys of PNe in [Fe ii] 1.644 μm have been rare. Hora, Latter & Deutsch (1999) observed a sample of PNe with medium-resolution (R ∼ 700) NIR (λ = 1–2.5 μm) spectra. Among the 41 they observed, 16 showed H2 2.122-μm emission but only three showed [Fe ii] 1.644-μm emission. While the fraction of PNe with detected [Fe ii] 1.644-μm emission was small, the emission has been useful in studying the nature of mass loss, as demonstrated in the case of M2-9 (Smith, Balick & Gehrz 2005).
The initial search for [Fe ii] emission from the known PNe in the UWIFE survey resulted in six detections from around 29 known PNe (25 from Acker et al. (1992) and four from F2011). Most of these PNe with [Fe ii] emission have accompanying H2 emission. However, three of them are not detected in H2 (e.g. PN M 1-51, see Fig. 13). For comparison, Kastner et al. (1996) found that all three PNe detected in [Fe ii] show accompanying H2 emission. The study of PNe in terms of connecting their emission characteristics to their morphology will be important, as it may reveal how the mass-loss history of progenitors affects the shaping of PNe.
5.3.2 Nebulae around evolved massive stars
Mass loss in massive stars plays a more critical role than in less massive ones. An important class of evolved massive stars is that of luminous blue variables (LBVs). LBVs are among the most luminous stars in our Galaxy (and thus have very large initial masses) and are typified by their irregular variabilities, which are sometimes associated with eruptive mass loss. The best example of an LBV star is η Car, which is surrounded by an extremely massive circumstellar shell (M ∼ 15 Mȯ), believed to be a product of the great ejection in the 1840s. The line emission from circumstellar shells around some LBVs has been a valuable tool for understanding not only the physical process responsible for LBV mass loss but also the nature of the central star. NIR lines, such as [Fe ii], are well suited to studying the shells around evolved massive stars as they are less affected by interstellar extinction. [Fe ii] emission has so far been found in six LBVs (Smith 2002), and detailed studies of η Car have demonstrated the usefulness of the [Fe ii] emission (Smith 2006).
In our preliminary study, we searched for [Fe ii] emission associated with known LBVs and candidate LBVs from Clark, Larionov & Arkharov (2005). Among several sources covered by the UWIFE survey, we detected [Fe ii] emission from a few of them. An example is HD 168625, shown in Fig. 14. HD 168625 is an intriguing system showing a bipolar nebula several times larger than its equatorial dust torus (Smith 2007), similar to what is found around SN 1987A. We found both H2 and [Fe ii] emission around the equatorial torus, but with different morphologies. The morphology of H2 emission is similar to that of Hα and the thermal emission from the dust (Pasquali et al. 2002) and is sharply peaked along the torus on the south-western side. In contrast, the [Fe ii] emission is more diffuse and more extended towards the north-eastern side. The only other LBV that shows both H2 and [Fe ii] emission is η Car. Follow-up spectroscopic observations of HD 168625 will reveal a detailed mass-loss history of the central star, as in the case of η Car (Smith 2006).
5.4 Supernova remnants
When a SN explodes, a strong shock with a typical expansion velocity of ∼10 000 km s−1 is produced, which later decelerates and becomes radiative at the speed of a few hundreds of kilometres per second. Along with this front shock expanding into the surrounding medium, there exists a reverse shock that propagates back into the SN material. Strong NIR [Fe ii] lines are emitted when these shocks interact with dense material, which could be either the SN or the circumstellar material for young SNRs, and is usually interstellar material for middle-aged SNRs (e.g. Oliva, Moorwood & Danziger 1989, 1990; Koo et al. 2007; Lee et al. 2009; Moon et al. 2009; Lee et al. 2013, see also Koo 2014 and references therein). Moreover, Fe, being the end-product of the stellar nucleosynthesis process, forms the main component of the supernova ejecta; the materials ejected from the deep layers of the progenitor are often enriched in Fe. The Fe abundance can also be significantly enhanced in a shocked region if the interstellar dust is destroyed.
In the survey area, there are over 70 known SNRs, mostly identified by radio and X-ray observations (Green 2009). We searched for [Fe ii] emission around the positions of the known SNRs. Our preliminary results reveal that 20–30 per cent of known SNRs show [Fe ii] emission, including previously observed [Fe ii]-emitting SNRs such as G11.2−0.3 (Koo et al. 2007), W 28 (Reach, Rho & Jarrett 2005), G21.5−0.9 (Zajczyk et al. 2012), 3C 391 (Reach et al. 2002), 3C 396 (Lee et al. 2009), W 44 (Reach et al. 2005) and W 49B (Keohane et al. 2007). In most cases, our [Fe ii] images provide high-quality, wide-area views of detected SNRs, for which sometimes only parts of the area were covered by previous observations. In addition, we found new [Fe ii]-emitting SNRs. Many of them show near-infrared H2 emission identified by the UWISH2 survey and/or thermal X-ray emission in the centre, implying that they are located within dense environments. For example, Fig. 15 shows a colour image of 3C 391, which is one of the bright [Fe ii]-emitting SNRs in the survey area.
5.5 Other sources and unbiased search
The UWIFE survey is an unbiased survey, so we have an opportunity to detect ‘every’ region of dense, shocked gas with bright [Fe ii] line emission in the inner Galaxy, some of which would be serendipitous discoveries. In order to identify IFOs in the UWIFE survey area, we searched for IFOs in the continuum-subtracted [Fe ii]-line images visually using the naked eye. At the time of writing, we have detected 111 IFOs in 10.5 deg2 from l = 53° to 60°. The detected sources include YSO/YSO candidates, AGB/RCB (R Coronae Borealis) stars and their candidates, EGOs and star-forming regions. Most (∼85 per cent) of them, however, do not have counterparts in SIMBAD. The detected IFOs are classified as point-like or diffuse based on their morphology. The diffuse sources can be further divided into amorphous, cometary, jet/filamentary, or shell-like objects. Fig. 16 shows examples of each category. To better systematize the unbiased search of IFOs, an automatic computer-aided method is being developed. We note, however, that a significant number of the point-like sources that we detected above might be variables instead of [Fe ii]-line-emitting sources.
6 CONCLUDING REMARKS
In this paper we have presented an overview and preliminary results of the UKIRT Widefield Infrared Survey for [Fe ii] (UWIFE). The UWIFE survey is an imaging survey of the first Galactic quadrant (7° < l < 62°; |b| ≲ 1| $_{.}^{\circ}$|5) that uses a narrow-band filter centred on the [Fe ii] 1.644-μm emission line, which is a good tracer of dense shock-excited gas. The survey started in the summer of 2012 and was completed in 2013. The resulting images have a median seeing of 0.8 arcsec, with a 5σ limiting magnitude of ∼19 mag, and will provide valuable resources with which to study the Galactic plane.
Our own Galaxy and its inner workings such as how new stars are born and die are still mysterious. Unbiased imaging surveys of the Galactic plane in the infrared broad-band such as GLIMPSE and MIPSGAL have greatly increased our understanding of the Galaxy. A significant advantage of our UWIFE survey is its complementarity with these existing and/or upcoming surveys, in particular with the UWISH2 survey. Line emission such as the [Fe ii] 1.644-μm line and the molecular hydrogen 1–0 S(1) emission line at 2.122 μm will complement the broad-band surveys and probe dynamically active components of the interstellar medium. The high-spatial-resolution narrow-band WFCAM images from the UWIFE and UWISH2 surveys will pinpoint regions of active interaction in complex environments. Furthermore, these imaging survey will provide a wealth of potential targets for follow-up spectroscopic observations.
ACKNOWLEDGMENTS
This work was supported by the K-GMT Science Program funded through the Korea GMT Project operated by the Korea Astronomy and Space Science Institute (KASI). H.-G. Lee acknowledges support from a Grant-in-Aid from the Japan Society for the Promotion of Science (JSPS) fellows (no. 23·01322). H.-J. Kim was supported by a NRF(National Research Foundation of Korea) grant funded by the Korean Government (NRF-2012-Fostering Core Leaders of the Future Basic Science Program). D. Moon was supported by the Korean Federation of Science and Technology Societies (KOFST). M.-G. Lee, B.-C. Koo, H.-M. Lee and M. Ishiguro were supported by a National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIP) (no. 2012R1A4A1028713).