GBT Multiwavelength Survey of the Galactic Center Region
DDraft version October 23, 2018
Preprint typeset using L A TEX style emulateapj v. 08/22/09
GBT MULTIWAVELENGTH SURVEY OF THE GALACTIC CENTER REGION
C. J. Law , F. Yusef-Zadeh , W. D. Cotton , and R. J. Maddalena Draft version October 23, 2018
ABSTRACTWe describe the results of a radio continuum survey of the central 4 ◦ × ◦ with the 100 m GreenBank Telescope (GBT) at wavelengths of 3.5, 6, 20, and 90 cm. The 3.5 and 6 cm surveys are the mostsensitive and highest resolution single dish surveys made of the central degrees of our Galaxy. Wepresent catalogs of compact and extended sources in the central four degrees of our Galaxy, includingdetailed spectral index studies of all sources. The analysis covers star-forming regions such as Sgr Band Sgr C where we find evidence of a mixture of thermal and nonthermal emission. The analysisquantifies the relative contribution of thermal and nonthermal processes to the radio continuum fluxdensity toward the GC region. In the central 4 ◦ × ◦ of the GC, the thermal and nonthermal fluxfractions for all compact and diffuse sources are 28%/72% at 3.5 cm and 19%/81% at 6 cm. Thetotal flux densities from these sources are 783 ±
52 Jy and 1063 ±
93 Jy at 3.5 and 6 cm, respectively,excluding the contribution of Galactic synchrotron emission.
Subject headings:
Galaxy: center — surveys — radio continuum: general INTRODUCTION
The central few hundred parsecs of the Milky Waycomprise a region in the Galaxy unique for its highstellar density, intense ionizing radiation field, massiveblack hole, enhanced density of cosmic rays, and unusualmagnetized structures (Figer et al. 2004; Rodr´ıguez-Fern´andez, Mart´ın-Pintado, & de Vicente 2001; Yusef-Zadeh, Wardle, & Roy 2007; Bally et al. 1987; Yusef-Zadeh, Hewitt, & Cotton 2004). The extent of the GCregion is roughly 400 pc in diameter, defined by a regionwith relatively high gas density ( n H (cid:38) cm − ; Hut-temeister et al. 1998). This region, sometimes called theGalactic nucleus or “central molecular zone”, produces5%–10% of the Galaxy’s infrared and Lyman continuumluminosity and contains 10% of its molecular gas (Ballyet al. 1987; Morris & Serabyn 1996). The density andphysical diversity of objects in the GC region make ithighly complex and require a wide range of observationsto unravel.At the simplest level, it is important to understand howthe basic components of the region interact. Nonthermalemission, in the form of supernova remnants (SNRs) andnonthermal radio filaments (NRFs; Yusef-Zadeh, Morris,& Chance 1984), dominates the cm-wavelength emissionin the region (e.g., Altenhoff et al. 1979; Duncan et al.1995). The relativistic component of the ISM has alsobeen observed at TeV energies as a diffuse source tracingthe molecular gas distribution (Aharonian et al. 2006).However, it is not clear how these energetic electronstraced by the nonthermal emission affect other compo-nents of the GC interstellar medium. For example, en-ergetic electrons may explain the unusually high ambi-ent gas temperature in the region and subsequent low Department of Physics and Astronomy, Northwestern Univer-sity, 2145 Sheriadan Road, Evanston, 60208 IL, USA Astronomical Institute “Anton Pannekoek”, University of Am-sterdam, Kruislaan 403, 1098 SJ Amsterdam, Netherlands National Radio Astronomy Observatory, 520 Edgemont Road,Charlottesville, VA 22903, USA National Radio Astronomy Observatory, P.O. Box 2, Rt.28/92, Green Bank, WV 24944, USA star formation efficiency (Yusef-Zadeh, Wardle, & Roy2007). Thermal emission traces star formation regionsand photoionized clouds. An accurate census of ther-mal emission can constrain the amount of star formationoccurring there or find new star forming regions. Sepa-rating the thermal and nonthermal processes also allowsan estimate of their basic properties ( n e , B ; Heiles et al.1996).Single-dish radio continuum observations are a usefultool for studying the large-scale properties of the GCregion. Several other single-dish surveys of radio contin-uum emission from the GC region have been conducted(Altenhoff et al. 1979; Handa et al. 1987; Reich et al.1990; Haynes et al. 1992; Duncan et al. 1995). Altenhoffet al. (1979) observed the Galactic disk from l = 60 ◦ to the GC region near 6 cm with the 100 m Effelsbergtelescope. This was one of the first single-dish surveysof the GC region with a resolution of a few (2 . (cid:48)
6) ar-cminutes and it discovered many compact and diffusesources. Since that time, the Parkes 64 m and Effelsberg100 m telescopes have been used to create complete sur-veys of the Galactic disk in the northern and southerncelestial skies near 12 cm, which were sensitive to faintemission on large scales (Reich et al. 1990; Duncan et al.1995). These studies revealed dozens of new SNR can-didates, compact H II regions, and Galactic loops andspurs, vividly demonstrating the chaotic structure of theGalactic interstellar medium.Although most of the world’s best radio telescopes havesurveyed the radio continuum in the GC region, therehas been limited study of spectral indices on arcminutescales. We were motivated to extend upon previous ob-servations using the largest, fully-steerable telescope, theGreen Bank Telescope (GBT). The high resolution andsensitivity of the GBT observations allow us to separateand quantify flux from all thermal and nonthermal emit- The GBT is operated by the National Radio Astronomy Ob-servatory, which is a facility of the National Science Foundationoperated under cooperative agreement by Associated Universities,Inc. a r X i v : . [ a s t r o - ph ] J a n ters. Section 2 describes the observations and data re-duction. In §
3, the results of the survey are described, in-cluding the compilation of compact and extended sourcecatalogs at 3.5 and 6 cm and a detailed discussion of theradio continuum properties for each source in the region.We calculate the percentage of flux from sources in thecentral degrees of the Galaxy from thermal/nonthermalprocesses at 3.5 and 6 cm. Finally, § OBSERVATIONS AND DATA REDUCTIONS
Description of Observations
In June 2003, we surveyed the central degrees of theGalaxy with the GBT at 3.5, 6, 20, and 90 cm. TheGBT is located in Green Bank, West Virginia at a lat-itude of 38 ◦ (cid:48) N. The GBT is unique in that it has alarge (100 m by 110 m) unblocked aperature that is fullysteerable, with an elevation range of 5–90 ◦ . Observa-tions were conducted over five days in August of 2002for a total observing time of 30 hours. At 3.5, 6, and 20cm, observations were made with the Digital ContinuumReceiver in “on-the-fly” mapping mode, while at 90 cmthe observations were made with the Spectral Processor.After flagging the 90 cm data for radio frequency inter-ference, the data were averaged in frequency and treatedidentically to the higher frequency data. Observationshad bandwidths of 320, 320, 20, and 40 MHz at 3.5, 6,20, and 90 cm.All observations surveyed at least a 4 ◦ × ◦ area roughlycentered on the GC, which includes the molecular zone inthe central 3 ◦ ( ∼
400 pc); the 6, 20, and 90 cm maps alsosurveyed beyond this region. The 3.5 and 6 cm surveys,shown in Figure 1, have similar coverage, while the 20and 90 cm surveys, shown in Figure 2, cover the centralten degrees. The spatial coverage, resolution, and typical1 σ sensitivity of these maps are shown in Table 1.Flux calibration was done by adding a calibrated noisesource to every other integration. The brightness of thenoise source is estimated from observations of 3C286and gives a conservative absolute flux accuracy of 5%.The final amplitude calibration is made from the medianbrightness of the noise source for each scan.No simultaneous measurement of the sky brightnesswas made during these observations, so a few techniqueswere used to estimate the noncelestial background foreach map. First, an initial esimate of the sky brightnessis done by observing a position far from the Galacticplane, near 3C286. Second, the atmospheric opacity andtemperature are estimated from weather monitoring sta-tions and subtracted from the data. Finally, an iterativescheme was used to separate the constant celestial bright-ness distribution from the time variable atmospheric andspillover contributions. An image of the sky was madewith the initial estimates of the atmospheric and receivercontributions removed from the data. This model skybrightness was then subtracted from the data and a highpass time filtering produced a refined estimation of theatmospheric and spillover contributions. This refined cal-ibration was then used to correct the data, resulting inan improved estimate of the sky distribution. Several it- eration with decreasing time constants of the high passfiltering yielded the results presented. It is important tonote that this technique cannot distinguish between slowvariations in the atmospheric brightness and true celes-tial changes, which effectively introduces a local “zeropoint”. Thus, measurements of brightnesses and fluxdensities require the subtraction of a local background,especially in the 3.5 and 6 cm maps, where atmosphericeffects are stronger.Each of the radio continuum maps was observed withorthogonal, “basket-weaving” patterns in Galactic coor-dinates. The redundancy of observing each position inthe map in two orthogonal scans is used to identify theflux contribution from noncelestial sources. The finalimages were made by convolving the data with a Gaus-sian kernel and resampling onto the image grid. Mapswere made with “Obit”, a group of software packagesdesigned to handle single-dish and interferometric radioastronomy data, such as AIPS data disks or FITS files.“ObitSD” is a low-level addition to the software packageand is designed for making maps from on-the-fly data.The 3.5 and 6 cm surveys have sensitivities of 8 and10 mJy beam − , which makes them one of the highest-resolution and most sensitive single-dish surveys of theregion made at these wavelengths. At wavelengths near3.5 cm, the GC region has previously been best sur-veyed with 45 m Nobeyama Radio Observatory and64 m Parkes Telescope to sensitivities of 15 and 30mJy beam − , respectively, with resolutions of about 2 . (cid:48) . (cid:48) − , or abouttwice as sensitive as the present work. An advantage ofthe present work is that it was planned as a multiwave-length campaign, so the observations at 3.5, 6, 20, and 90cm were calibrated by similar techniques; this makes thesurvey robust and especially suited to studying spectralindices. Basic Source Analysis
Compact sources were identified by eye by searchingfor sources not dominated by extended emission. Sourceproperties were measured by JMFIT of AIPS. JMFITfits a 2-D Gaussian to an image, assuming an initial sizeand shape equal to the beam size. The JMFIT routinealso fits for the absolute background flux and its firstspatial gradient, which is particularly useful for the 6 cmmaps. The output of the source detection is a set of po-sitions, flux densities, sizes for a best-fit Gaussian, plus aresidual image. The residual images were inspected as atest of the fit quality and were generally less than about10% of the source peak brightness. Other source-fittingalgorithms, including the IDL astrolib photometry pro-grams and MIRIAD’s sfind, were tested on our images,but neither of these produced as trustworthy results asJMFIT. In particular, the algorithms did not easily pro-duce source-subtracted residual images to enable visualinspection of the fit quality.All sources detected at 6 and 3.5 cm have a spectralindex calculated. The spectral index is calculated fromthe integrated flux densities assuming S ν ∝ ν α , whichgives: α = log(S / S ) / log( ν /ν ) (1) σ α = (1 / log( ν /ν )) ∗ (cid:112) ( σ / S ) + ( σ / S ) (2)where “1” and “2” refer to the observing frequencies,which are abbreviated as X for 3.5 cm, C for 6 cm, L for20 cm, and P for 90 cm. The images were not convolvedto the same resolution for measuring the compact sourceflux densities, since tests with JMFIT show that the in-tegrated flux density of the 3.5 cm image is the same atits natural resolution and after being convolved to matchthe 6 cm GBT resolution.Extended sources in the GC region, such as H II re-gions, SNRs, and nonthermal radio filaments (NRFs),dominate the radio brightness of the GC region. For amore detailed study of these sources, flux density sliceswere taken at different positions and orientations throughextended sources in order to measure brightnesses andthe corresponding spectral index. Slices were taken fromtwo images convolved to the same resolution; the con-volution size was 2 . (cid:48) (cid:48) forthe 6/20 cm slices. To estimate the spectral index of asource, a background was subtracted prior to taking theratio of the source brightnesses. This is done by a re-duced chi squared fit of a line to a portion of the slicedata. All structure in the slice that does not look like abackground was ignored in fitting the background. Thebackground was fit to each slice independently, althoughthe same background region was used at both frequen-cies. The best-fit line was then subtracted from the dataprior to calculation of the spectral index.Figure 3 demonstrates the slice analysis technique ontwo sources with relatively simple morphologies. Foreach slice shown in the image, the flux and spectral in-dex are shown in a plot. The value of the spectral indexshown in the image is measured at the peak brightness ofthe shorter wavelength image (3.5 cm for 6/3.5 cm com-parison and 6 cm for 20/6 cm comparison). The dashedlines show the best-fit background for the slices, madeby ignoring the source emission, which is shown with adotted line; sometimes the “source” includes parts of theslice that are confused with other sources. The spectralindex measurements were considered trustworthy only ifthey were found not to vary significantly with ∼ α CX = 0 . ± . α CX = − . ± .
04, as hadbeen previously observed (Helfand & Becker 1987). Notethat we refer to the spectral index at the peak flux as thesource’s spectral index, but the spectral index is calcu-lated for every point along the slice. In some extendedsources, the spectral index changes away from the peakflux, and the spectral index distribution is discussed inmore detail. RESULTS
Compact Source Catalog
Here we describe the properties of all compact sourcesfound in the 3.5 and 6 cm maps. A source is considered tobe compact if its best-fit width is less than approximatelytwice the beam FWHM. Only one of the 36 sources inthe 3.5 cm compact catalog has a Gaussian width ap-proaching roughly twice that of the beam (roughly 3 (cid:48) ).Any source that is not compact in the highest-resolutionimage (3.5 cm) is considered “extended” and describedin § § . (cid:48) − , respec-tively). Sources in that catalog that are coinicident withthe present 3.5 cm survey are shown in Table 2 for com-parison. The integrated flux densities observed here aresystematically 24% less than in Haynes et al. (1992), withthe greatest difference seen in sources in dense regionssuch as Sgr B and Sgr C. Since this difference is morethan expected from the calibration uncertainty (6% fortheir work and 5% for the present work), we repeated oursource detection on an image convolved to the resolutionof Haynes et al. (1992). The integrated fluxes measuredin our convolved map is consistent with that of Hayneset al. (1992), which suggests that the larger beam tendsto include more extended emission in the integrated fluxof compact sources. This caveat should be consideredwhen using compact source fluxes, especially in regionswith lots of extended emission.The 6 cm catalog in Table 3 was compared to cat-alogs in the literature to confirm those results, but nosingle-dish survey at this wavelength could be found.In the VLA study of Sgr E by Gray et al. (1993), twosources (J174203–300405 and J174227–295559) are asso-ciated with relatively unconfused sources in the present6 cm catalog. Also, in the 6 cm VLA survey of Beckeret al. (1994), three sources can be compared to the GBTsurvey. In all these cases, the VLA flux densities rangefrom 2–70% of that measured by the GBT 6 cm survey;there is likely to be flux missing from the interferometricobservations.To aid interpretation of the source catalogs, we esti-mated the number of extragalactic sources expected inthe survey region. Huyhn et al. (2005) fit a seventh-orderpolynomial to the flux distribution of 20 cm compactsources from several high Galactic latitude catalogs. Weestimate the number of extragalactic sources by integrat-ing over this distribution and assuming a spectral indexof –0.7. At 3.5 and 6 cm, we estimate 0.4 and 0.2 sources,respectively, in the 4 ◦ × ◦ survey region with a flux den-sity higher than the faintest compact source detected ateach wavelength (0.12 Jy and 0.29 Jy, respectively). Extended Sources
The following section describes the cataloging andanalysis of extended sources observed in the survey re-gion. The spectral index is studied by integrated fluxesand slices of the maps. Section 3.2.1 describes howsources are identified and the integrated spectral indexbetween 3.5 and 6 cm, while § Extended Source Catalog
Here we describe the construction of a catalog of theextended sources. All sources not considered compact(i.e., not satisfying θ . < ∗ FWHM) were includedin the extended source catalog shown in Table 5.Figure 4 shows the 3.5 and 6 cm images with extendedand compact regions overlaid. Regions were defined toenclose all flux from an object (in both the 3.5 and 6cm images) that is believed to have a similar origin.Often this is simple, such as for the supernova rem-nant G359.1–0.5, which has a distinctive ring-like shapeand has been extensively studied (Uchida et al. 1992;Yusef-Zadeh, Uchida, & Roberts 1995). Otherwise, theregions were simply defined by whether it had a ther-mal or nonthermal spectral index between 6 and 3.5 cm(see § R eff = (cid:112) A/π ). Columns(5)-(8) give the 3.5 cm raw source flux density, the rms uncertainty measured in a background region, andthe background-subtracted flux density and its error.Columns (9)-(12) give the same quantities for the sourcesat 6 cm. The source brightness is integrated over all pixelvalues and then scaled by the ratio of the pixel area tothe beam area to get a flux density in Jy. The scale fac-tor is (30 (cid:48)(cid:48) ) / (1 . ∗ F W HM ) (as used in AIPS), with F W HM = 88 (cid:48)(cid:48) and 153 (cid:48)(cid:48) at 3.5 and 6 cm, respectively.The background flux density is measured over a nearbyregion. The rms in the background is scaled to the sourcearea to estimate the uncertainty in the integrated sourceflux density. Often this method overestimates the errorin the flux density, since the rms in the background isdominated by other sources or Galactic emission. Withthe flux density measured at 3.5 and 6 cm, the spectralindex and its statistical error is calculated and shownin Table 6. The sixth column shows our conclusion onthe nature of the radio continuum emission (thermal vs.nonthermal), based on the integrated and slice spectralindex analysis.The spectral index values shown in Table 6 are gener-ally consistent with the spectral indices measured fromslices of the data presented in § σ errors from the slice analysis. This may becaused by imperfect subtraction of Galactic synchrotronemission, which is stronger at 6 cm than at 3.5 cm, andstronger near the Galactic plane. Generally, the sliceanalysis gives a better estimate of the local background,so it gives a more robust estimate of the spectral index.However, the slice analysis is limited because it cannotmeasure the integrated flux from an extended source. Known and Candidate Supernova Remnants
The extended source catalog includes eight sourcesthat have previously been identified as known or can-didate SNRs. The types of SNRs represented in the GCregion vary from traditional shell-types to those havingpulsar-wind neblae to mixed-mophology types. Table 7summarizes the observed characteristics of all supernovaremnants in the region. The results presented here aregenerally consistent with the high-resolution, 843 MHzobservations presented in Gray (1994) and associated pa-pers. Between 3.5 and 6 cm, the spectral indices rangefrom flat to α ≈ −
2. A detailed discussion of the sourcesand comparison to previous work is given in § Thermal/Nonthermal Flux Fractions
Each extended source in Table 6 has been classified aseither thermal or nonthermal according to the spectralindex analysis (see § ± ± ± ± . Table 8 shows the bestestimate of the thermal and nonthermal flux fractionsin the survey is 28%/72% at 3.5 cm and 19%/81% at 6cm. The total flux densities for all compact and extendedsources are 783 ±
52 Jy and 1063 ±
93 Jy at 3.5 and 6cm, respectively. The survey region covers the central4 ◦ × ◦ equivalent to a physical size of 560 pc ×
140 pc ata distance of 8.5 kpc.The first caveat in the interpretation of these resultsis that these flux fractions are measured for discretesources and does not consider the diffuse Galactic emis-sion. Overall, the Galactic emission is dominated by ex-tended synchrotron emission that fills the field observedhere, particularly at low frequencies. Second, the dis-tinction of “thermal” and “nonthermal” sources by themeasured spectral index may not be accurate for sourceswith a mixture of thermal and nonthermal processes. Forexample, it would not be accurate to describe a super-nova remnant embedded in a star forming complex as oneof these two categories. While the source regions used tomake Table 5 were defined to separate sources with dif-ferent structure and emission mechanisms as best as pos-sible, there may be weak emission that is miscategorizedby this analysis.A previous study of the thermal/nonthermal flux dis-tribution toward the GC region was done with the Effels-berg 100 m telescope (Schmidt 1978). That work foundthat the central 350 pc (2 . ◦
5) had roughly equal contri-butions from thermal and nonthermal emission at 6 cm,with a total flux density of about 2000 Jy (Kruegel etal. 1983; Lis & Carlstrom 1994). The total flux densityis twice that observed here for a similar region, so thatwork probably included the extended background emis-sion that this work excluded (the work is only availablein a thesis). Mezger (1996) found a roughly equal frac-tion of thermal and nonthermal emission at 6 cm in thecentral 400 pc ×
350 pc, with a free-free flux density of ∼
580 Jy at 6 cm. The present work finds a similar totalflux as that of Mezger (1996), but we identify more ofthe emission from discrete sources as nonthermal.A number of recent low-frequency radio continuum,X-ray, and molecular line observations suggest an in-crease in the cosmic ray ionization rate in the nucleardisk (Yusef-Zadeh et al. 2006; Oka et al. 2005). The highfraction of nonthermal emission from the central regionshown here is qualitatively consistent with other mea-surements.
Slices of Extended Sources in GBT Images
Figures 5 through 21 show details of the 3.5, 6, and20 cm surveys and the slices used to study the spectralindices of objects between these frequencies. The figuresare ordered with increasing Galactic longitude, startingfrom the western edge of the 3.5 and 6 cm surveys. Each This assumption should not affect the final values much, sincethe flux from these sources is smaller than the uncertainty in thetotal flux figure shows two images convolved to the same resolu-tion with slice positions overlaid. For each figure, onerepresentative slice is plotted below the images.
G357.7–0.1 (The Tornado) and G357.7+0.3
G357.7–0.1, also known as “The Tornado” for its un-usual, twisted morphology, is found near the westernedge of the 3.5 and 6 cm maps. In Figure 5, the 3.5and 6 cm morphology looks like a head-tail source withintegrated flux densities of about 14, 18, and 37 Jy at 3.5,6 cm, and 20 cm, respectively. The Tornado is believedto be a mixed-mophology SNR at a distance of about 12kpc (Gaensler et al. 2003; Brogan & Goss 2003a). Mixed-morphology SNRs are characterized by a radio contin-uum shell filled with thermal, x-ray–emitting gas (e.g.,Yusef-Zadeh et al. 2003). The elongated morphology ofthe Tornado is unsual for a SNR, but can be explained bythe fact that the Tornado is interacting with a molecularcloud (Frail et al. 1996).The spectral index between 6 and 3.5 cm is shown inFigure 5 and between 20 and 6 cm in Figure 8. The6/3.5 cm spectral index for most slices perpendicular tothe long axis are equal within their 3 σ errors except forone slice on the western side (the “head”) of the Tornado.The typical spectral index values are α CX ∼ − .
45 and α LC ∼ − .
63, with systematic uncertainties of about0.03 and 0.01, respectively. The 6/3.5 cm spectral indexfrom the slice analysis is consistent with the integratedspectral index of − . ± .
07, given in Table 6. Note thatcomparing these spectral indices to other works shouldaccount for the flux calibration uncertainty by addinga spectral index error of 0.13 in quadrature. The 20/6cm spectral index is steeper than the 6/3.5 cm index,which suggests that the spectral index steepens at lowerfrequencies. From Green (2004) and Becker & Helfand(1985), the 1 GHz (30 cm) flux of the Tornado is 37Jy and spectral index is –0.4, while Gray (1994) finds S MHz = 49 Jy. Extrapolating from our measured 20cm flux density of 37 Jy and α LC ∼ − .
63, we predict anintegrated flux density of 46 and 51 Jy at 1 GHz and 843MHz (35 cm). Thus, the present observations overpredictthe flux given in Green (2004), but are consistent withthe observations of Gray (1994). The present single-dish–derived flux and spectral index are less likely to be biasedthan the previous interferometric values, which are morelikely to underestimate the flux density.The spectral index and flux density can be used to cal-culate an equipartition magnetic field. Beck & Krause(2005) give a new derivation for the equipartition mag-netic field strength as B eq = (cid:32) π (2 α + 1)( K + 1) I ν E − αp ( ν/ c ) α (2 α − c ( α ) lc ( i ) (cid:33) / ( α +3) (3)where K is the proton to electron number density ra-tio, c i are constants that depend on the spectral indexand magnetic field inclination angle, which hereafter isassumed to be equal to 0 (in the plane of the sky). Al-ternatively, the classical formulation of the equipartitionmagnetic field is B class = (8 πG ( K + 1) L ν /V ) / (4)where G is a function of the energy range considered andspectral index, K is the energy density ratio between pro-tons and electrons, and V is the volume of the emittingregion (Pacholczyk 1970; Beck & Krause 2005).Assuming a proton to electron energy density ratio of40-100 (Beck & Krause 2005) and a path length throughthe Tornado equivalent to its 2 (cid:48) width (10.4 pc assuming D = 12 kpc), the 6 cm peak brightness of the Tornadoof 5 Jy beam − and spectral index –0.48 gives B class =100 − µ G. This calculation integrates over frequenciesfrom 10 MHz (3 m) to 10 GHz (3 cm), but changes byless than 10% for an upper limit of 100 GHz.The spatial dependence of the spectral index is appar-ent in Figure 5, which shows the flux density and spec-tral index for a slice through the elongated portion of theTornado. This slice shows a regular change in α CX from − . ± .
02 near the brightest emission to − . ± .
07 inthe tail. The spectral index is a direct measure of the en-ergy distribution of the electrons and suggests that thatthe electrons in the tail region are more energetic thanin the head region.A new source, G357.7–0.4 is found near the Tornadoin projection as shown in Figure 5. G357.7–0.4 is anelongated, wavy structure with a thermal spectral index.The morphology and difference in spectral index suggeststhat it is unrelated to the Tornado. The thermal emissionfrom G357.7–0.4 should not affect the measurements ofspectral index from the tail of G357.7–0.1, since theyare significantly separated and oriented perpendicular toeach other.At the top of Figure 5 and north of the Tornado isG357.7+0.3, which has long been known as a supernovaremnant from its ring-like morphology, linearly polarizedradio emission, and soft X-ray emission (Reich & Fuerst1984; Leahy 1989; Gray 1994). The slice across the south-east portion of G357.7+0.3 shown in Figure 5 has a flatspectral index with large uncertainty. The spectral in-dex is significantly steeper toward the southwest, sincethe 3.5 cm emission is absent, but the 6 cm brightness issimilar to the southeast; based on the 3.5 cm upper limit,we estimate α CX (cid:46) − . Sgr E Region (G358.7-0.0)
Figures 6 and 7 show the radio continuum emissionfrom the E3 filament (G358.60-0.27) and the Sgr E com-plex (G358.7-0.0), respectively (Yusef-Zadeh, Hewitt, &Cotton 2004). The E3 filament is about 25 (cid:48) long andtakes a twisting path from the Sgr E star-forming com-plex toward the southwest. If the E3 filament is near theGC, it has a length of about 58 pc. The Sgr E complexis made of a collection of compact sources within a 20 (cid:48) region near (358.7,0.0), surrounded by extended sourcesnear (359.0,+0.0) and (358.4,+0.1).The spectral index between 6 and 3.5 cm are shownin Figures 6 and 7 and between 20 and 6 cm in Figure8. High-resolution 20 cm continuum observations withthe VLA show the region filled with compact H II re-gions (Yusef-Zadeh, Hewitt, & Cotton 2004), so much ofthe extended emission here could be unresolved compactsources. Consistent with this expectation, the 6/3.5 cmspectral index for most of this emission is consistent with a thermal origin. The E3 filament also has a thermal in-dex throughout.The easternmost and westernmost slices in Figure 7show nonthermal 6/3.5 cm indices. The easternmostslice passes through G359.0+0.0, which is seen in high-resolution, 20 cm images as extended, filamentary struc-tures (Liszt 1992; Yusef-Zadeh, Hewitt, & Cotton 2004).The morphology is remeniscent of an evolved H II region,but the slice and integrated spectral indices are consis-tent with nonthermal emission. The peak 6 cm bright-ness of 1 Jy beam − and 6/3.5 cm spectral index of –0.63implies a revised equipartition magnetic field strength of65 − µ G, assuming it is in the GC and has an numberdensity ratio of 40–100 (Beck & Krause 2005).
SNR G359.1–0.5 and the Snake NRF (G359.1-0.2)
Figure 9 shows the 6 and 3.5 cm emission and slicesacross the SNR G359.1–0.5 and the G359.1–0.2 (alsoknown as “The Snake”). G359.1–0.5 appears here as aring of radius 9 . (cid:48) . ± . . ± . − ; see Figure 9) and associated 6/3.5 cm spectralindex (–0.84) are consistent with a revised equipartitionmagnetic field of 66 − µ G, assuming it is located nearthe GC and has a number density ratio of 40–100 (Beck& Krause 2005).Studies of molecular gas and tracers of shocked gashave found that G359.1–0.5 is interacting with a molec-ular cloud (Uchida et al. 1992). Interestingly, G359.1–0.5has a statistically significant spatial variation in its spec-tral index. Figure 10 plots its 6/3.5 cm spectral indexas a function of theta, the position on the ring of emis-sion relative to galactic north. The 20 cm image doesnot have the resolution to allow a similar study of the20/6 cm spectral index. The 6/3.5 cm spectral index forG359.1–0.5 is significantly flatter for θ = 250 − ◦ (thegalactic west through north side). The region with flat-ter spectral index is nearly identical to the region withmost intense emission from the surrounding HI, CO,and OH(1720 MHz) maser emission (Uchida et al. 1992;Yusef-Zadeh, Uchida, & Roberts 1995). This is consis-tent with the idea that G359.1–0.5 is interacting with thesurrounding molecular cloud (Uchida et al. 1992).The Snake is a long ( ∼ (cid:48) or ∼
46 pc at 8 kpc),nonthermal filament that runs from the Galactic planeto the G359.1–0.5 SNR (Uchida et al. 1992; Gray et al.1995; Yusef-Zadeh, Hewitt, & Cotton 2004). The Snakeis unusual among NRFs because it has two sharp kinksalong its length (not visible with this data, but see Grayet al. 1995). The integrated flux densities given in Ta-ble 5 imply a 6/3.5 cm spectral index of − . ± . − , and a depth equal to its width of 9 . (cid:48)(cid:48) µ G, for K = 40 − K , which is more consistentwith current estimates (Beck & Krause 2005). Sgr C Region (G359.5-0.0)
Figure 12 shows the 6 and 3.5 cm flux densities, slices,and spectral indices in the Sgr C complex. The Sgr C H II region around (359.5,–0.1) has flux densities of about8 and 7 Jy at 3.5 and 6 cm, respectively, making it thebrightest radio continuum source in the western half ofthe survey. To the north of Sgr C near (359.4,+0.3) andpossibly just south of Sgr C is the radio continuum coun-terpart of the GC lobe (Law et al. 2008). G359.2+0.0 islocated near where the Snake meets the Galactic plane.This extended structure has flux densities of about 3 and8 Jy at 3.5 and 6 cm, respectively.The slice and integrated flux densities show that non-thermal emission dominates the extended emission in theregion, outside of the Sgr C H II region. G359.2+0.0 hasa very steep spectral index between 6 and 3.5 cm. Theslice plotted in the bottom of Figure 12 corresponds tothe vertical slice just west of Sgr C; it shows how theemission surrounding Sgr C is predominately nonther-mal.We note that some of the extended structure seen inthe present survey is resolved into NRFs in high resolu-tion observations (Yusef-Zadeh, Hewitt, & Cotton 2004;Nord et al. 2004). The observed distribution of NRF fluxdensity shows an increasing number of NRFs down to thepresent detection limits (Nord et al. 2004). More sensi-tive, high-resolution observations are likely to find manymore faint NRFs than are currently observed. Near SgrC, the total 20 cm flux density of known NRFs and NRFcandidates is roughly 2 Jy (Yusef-Zadeh, Hewitt, & Cot-ton 2004); this is roughly 10% of the total, background-subtracted, nonthermal 20 cm flux density observed byGBT of about 10–20 Jy. We suggest that extended (butnot Galactic synchrotron), nonthermal sources in thecentral degree of the GC region may have significant fluxfrom NRFs. While many other processes also producenonthermal radio emission (e.g., SNRs), the example ofthe Sgr C region may help direct searches for NRFs; otherregions with similar radio continuum morphologies andspectral indices include G359.0+0.0, G359.2+0.0, andG0.8+0.0. G359.8–0.3
Figure 13 shows the continuum emission and slices inG359.8–0.3. This source has a lumpy, shell-like morphol-ogy in this data, with flux densities of about 19 and 28 Jy at 3.5 and 6 cm, respectively. The slices and inte-grated spectral indices of G359.8–0.3 are consistent witha thermal origin, suggesting that it is an H II region.Low-frequency radio continuum absorption and H α emis-sion are observed from G359.8–0.3, as would be expectedfrom a thermal source in the foreground to the GC region(Brogan et al. 2003b; Gaustad et al. 2001). A Chandra survey of the the region shows soft, X-ray emission asso-ciated with this shell, consistent with a foreground source(Wang, Gotthelf, & Lang 2002).
Sgr A (SNR G0.0+0.0)
Figures 14 and 15 show the Sgr A region. The peakradio brightness on arcminute scales in this region in-cludes emission from Sgr A East and Sgr A West, the twobrightest radio continuum sources in the central severalarcminutes. Sgr A East is roughly 4 (cid:48) × (cid:48) , so no spatialstructure is seen in the present survey and only the fluxfrom the whole Sgr A complex is measured. The peakbrightness in this region is 39 Jy beam − at 3.5 cm and85 Jy beam − at 6 cm. Convolving the 3.5 cm map tothe resolution of the 6 cm map gives a brightness of 66Jy per 2 . (cid:48) ±
10 Jy comes from SgrA East and 21 ± ±
10 Jy (Brown & Johnston 1983;Mezger et al. 1989; Pedlar et al. 1989). The results ofthe present survey are consistent with the interferomet-ric observations, which suggests that they do not resolvemuch flux out (Pedlar et al. 1989).The 6/3.5 cm spectral index for the three slices acrossSgr A, and for the integrated brightnesses given aboveare consistent with α CX = − . ± .
02 for the central ∼ . (cid:48) α LC = − . ± .
02 for the central ∼ (cid:48) . High resolution observationsof Sgr A East measure a spectral index of about –1.1between 20 and 6 cm, which is believed to extend throughthe cm-wavelength regime (Pedlar et al. 1989). Sgr AWest has a flat spectrum at wavelengths between 3.6and 20 cm (Mezger et al. 1989; Pedlar et al. 1989). Thecombination of these two effects can explain the value of α CX = − .
44 observed in the present survey, althoughat longer wavelengths, the spectral index should becomesignificantly steeper, since Sgr A East begins to dominatethe apparent flux. It is likely that the 6/20 cm spectralindex measures a significant amount of the “halo” fluxaround the Sgr A complex, since the 20 cm map has abeam size of 9 (cid:48) . The halo emission has a flat spectralindex between 20 and 90 cm that may extend to shorterwavelengths (Pedlar et al. 1989).
Radio Arc (G0.2-0.0) and Arched Filaments(G0.07+0.04)
The Arched filaments/Radio Arc complex has been ex-tensively studied at high resolution with the VLA (Yusef-Zadeh, Morris, & Chance 1984; Tsuboi et al. 1986; Lang,Goss, & Morris 2002; Yusef-Zadeh et al. 2002). The slicesshown in Figures 15, 16, and 17 confirm the results ofearly observations that found that the Arched filamentsare thermal and the vertical component of the Radio Arcis nonthermal. The slice in Figure 16 shows a slightlynegative index, but this is likely due to a mixture of thethermal Arched filaments and the ambient nonthermalemission near the Radio Arc; using a closer part of theslice to estimate the background of the Arched filamentsgives α CX ∼ − .
1. The integrated arched filament fluxesand spectral index are difficult to evaluate due to confu-sion with surrounding emission, but are consistent witha thermal origin (see Table 5). The Radio Arc spectralindex is generally quite flat, as is shown in Figure 15,where α LC = − . ± . α CX = 0 . ± .
13, similar to the Arched filaments.There is also a coincidence between the brightness ofthe Radio Arc and the extent of the Bubble. As seenin Figure 24, the 3.5 cm emission from the Radio Arc isbrightest inside the ring made by the Arched filamentsand southern Arched filaments, but fades rapidly outsideof that ring. The brightness enhancement of the RadioArc inside the Bubble is even more pronounced whencompared to high Galactic latitutes (in the GC Lobe),where cm-wavelength continuum brightness is more thana factor of ten less than inside the Bubble (Simpson etal. 2007; Cotera et al. 2007).
SNR G0.33+0.04
The SNR G0.33+0.04 is located adjacent to (and par-tially confused with) the Radio Arc in Figure 17. Becauseit is so confused with the Radio Arc emission, an inte-grated spectral analysis is not done. A multiwavelengthstudy of G0.33+0.04 finds a spectral index of –0.56 forfrequencies higher than a GHz (Kassim & Frail 1996).The slice shown in Figure 17 crosses a portion of thisSNR and finds a spectral index α CX = − − .
5, sig-nificantly steeper than previous measurements at 5 and15.5 GHz (6 and 2 cm) (Altenhoff et al. 1979; Gordon1974). However, we note that at these frequencies theemission is dominated by emission from the Radio Arcand a few H II regions and the morphology does not re-semble the clear SNR-like morphology seen at lower fre-quencies (Kassim & Frail 1996). This confusion is likelyto bias the flux densities measured by the present andprevious works at frequencies above 5 GHz. G0.5–0.5
Figure 18 shows G0.5–0.5, an irregular complex ofclumpy extended emission with peak brightnesses around 2 Jy beam − at 6 cm. The 6/3.5 cm spectral index mea-surements shown in Figure 18 are near zero, showingthat G0.5–0.5 is predominately thermal. Like G359.8–0.3, this region is seen in absorption in 74 MHz contin-uum and emission in optical H α , indicating that it is inthe foreground of the GC region (Brogan et al. 2003b;Gaustad et al. 2001). One slice in Figure 18 shows asignificantly nonthermal spectrum, although it appearsmorphologically similar to the rest of the G0.5–0.5 com-plex. The integrated and slice spectral index values areconsistent with nonthermal emission in the southern por-tion of G0.5–0.5. All other 20/6 cm spectral index mea-surements across G0.5–0.5 shown in Figure 19 are con-sistent with a thermal origin. Sgr B (G0.5–0.0 and G0.7–0.0)
As shown in Figures 20 and 21, Sgr B is the brightestradio continuum source in the eastern half of the survey.The emission is brightest in two regions, called Sgr B1(G0.5–0.0) and Sgr B2 (G0.7–0.0), the western and east-ern halves, respectively. The entire Sgr B region has fluxdensities of 90 Jy and 85 Jy at 3.5 and 6 cm, respec-tively, of which roughly 60% comes from Sgr B2. Sgr Bis also surrounded by a few isolated H II regions that aredetected as compact sources in Tables 2 and 3.The 6/3.5 cm and 20/6 cm spectral index measure-ments for most of the Sgr B complex are near zero, whichis consistent with it being thermal. As shown in Table5, the 6 and 3.5 cm spectral index values are near zerowith three exceptions: one significantly larger than zeroand two much less than zero. The positive spectral indexis measured at the brightest portion of Sgr B2, a regionknown to be filled with optically-thick, ultracompact H II regions (Gaume et al. 1995). The nonthermal spectral in-dex measured east of Sgr B2 is in a region with extendedemission that is not clearly associated with the Sgr Bcomplex. The other nonthermal spectral index is mea-sured between Sgr B1 and B2. Sgr B1 has more extendedradio continuum morphology and has fewer compact H II regions that Sgr B2, which suggests that the region isrelatively older than Sgr B2 (Mehringer et al. 1992). Thenonthermal index measured between the two, at (0.55,–0.05), may indicate that a supernova has occurred there,although no clear morphological signatures of a super-nova in the Sgr B complex have ever been reported.Koyama et al. (2007) recently reported an extended re-gion with 6.7 keV thermal line emission at (0.61,–0.01),which they interpret as a young (7 × yr-old) super-nova remnant. The 6.7 keV-emitting region and the non-thermal radio emission are both located between Sgr B1and B2 and only about 4 (cid:48) from each other, so the ra-dio continuum emission is consistent with the supernovahypothesis. The nonthermal radio index measured be-tween Sgr B1 and Sgr B2 is –0.2, flatter than that ofSNRs, but it is likely to be mixed with H II regions. Theobservation of nonthermal continuum emission in the SgrB complex is consistent with previous observations fromthroughout the electromagnetic spectrum (Yusef-Zadeh,Wardle, & Roy 2007; Hollis et al. 2007; Crocker et al.2007). Observations of 90 cm continuum find a total SgrB flux density of ∼ ∼ .
35, perhaps indicative of significantabsorption by thermal gas in the region.
Sgr D (G1.1–0.1), SNR G1.0–0.2, and SNRG0.9+0.1
The eastern edge of the survey has a mixture of ther-mal and nonthermal sources, as seen in the images andslices in Figures 21 and 23. Sgr D (called “G1.1–0.1” inTable 7) appears here as a compact source (at 3.5 cm;see Table 2) surrounded by a shell. Previous work hassuggested that the Sgr D shell is ionized by the compactsource at its center and that the extended radio contin-uum emission toward the east (called “G1.2+0.0” here)may be gas escaping from Sgr D (Liszt 1992). G0.9+0.1and G1.0–0.2 (a.k.a. “G1.05–0.1” Gray 1994) are super-nova remnants with clear shell-like structures and non-thermal 6/3.5 and 20/6 cm spectral indices. G0.9+0.1also has a Crab-like source inside the SNR (with a sim-ilar flux density as the shell at 6 cm), making it one ofthe first SNRs categorized as a “composite” (Helfand &Becker 1987).The integrated spectral index of the Sgr D extendedemission of α CX = − . ± .
14 is consistent with itsidentification as an ionized shell of gas. Although theslice through G1.1–0.1 (the extended Sgr D emission) isnominally nonthermal ( α CX = − . ± . α CX = − . ± .
10, which suggests thatthere is a mixture of thermal and nonthermal emissionin the region. The slice through the compact source at(1.1,–0.1) clearly has a thermal-like index of 0 . ± . II region with radio recombinationline emission with v LSR ≈
25 km s − (source G1.127-1.04 of Liszt 1992). The source is detected as a compactsource at 3.5 cm at (17:48:40.29,–28:01:23.5) with a peakbrightness of 2.3 Jy and flux density of 3.5 Jy, as com-pared to the flux density near 19 cm of 1.7 Jy (Liszt1992).G0.9+0.1 is known to have a flat spectrum pulsar windnebula in its center and a steep-spectrum shell (Helfand& Becker 1987; Gray 1994). The slices shown in Figure23 are generally consistent with this picture, showing flat6/3.5 cm indices in the core ( ∼ − .
16) and a steeper spec-tral index in the shell ( ∼ − . . (cid:48) . ± .
07 Jy (LaRosa et al. 2000) consideringthe flat spectral index of the core. The flux density ofthe shell is difficult to estimate, since it is confused withthe core in the 2 . (cid:48) . (cid:48) − . ± .
03 and − . ± .
04, re-spectively, which is consistent with the values of − . − .
45 found by Helfand & Becker (1987). In theshell, comparing the 6 cm flux density to the 90 cm fluxdensity of LaRosa et al. (2000) implies a slightly flatterindex of ∼ .
2, possibly indicating a spectral turnover. G1.0–0.2 is a supernova remnant with a shell mor-phology and nonthermal radio spectral index (calledG1.05–0.1 in Gray 1994). Unlike G0.9+0.1, there is nocompact source inside G1.0–0.2. Figure 23 shows twoslices through G1.0–0.2 with α CX = − . ± .
05 and − . ± .
08, consistent with previous observations thatfound an index that ranged from − . − . α CX = − . ± .
18, which is roughly similar tothe slice results. Liszt (1992) report that G1.0–0.2 has aflux density of 12.3 Jy from interferometric observationsat 1616 MHz; this gives an upper limit to the spectralindex of α LC < − . CONCLUSIONS
This paper has shown results from a new survey of theradio continuum emission from the central degrees of theGalaxy at 90, 20, 6, and 3.5 cm with the GBT. The 6 and3.5 cm surveys are the most sensitive, highest resolution,single-dish radio surveys of the central degrees of the GCmade at these wavelengths. The primary products of thisstudy are catalogs of all compact and extended sources,including a spectral index analysis of these sources. Wehave shown that the compact sources ( θ (cid:46) (cid:48) ) detectedat 6 and 3.5 cm surveys are most likely to be GalacticH II regions and are mostly found near well-known H II complexes, such as Sgr B and Sgr E. About one quarterof the sources detected at 3.5 cm are also detected at 6cm; most of these sources have thermal spectral indices.Extended, nonthermal emission in the 3.5, 6, and 20cm surveys is found associated with GC star-forming re-gions. The emission is on size scales of tens of arcmin-utes and is particularly found for l = 358 . ◦ − . ◦ ◦ × ◦ of the Galaxy, the thermal to nonthermal flux fractionsfor all discrete emission are 28%/72% at 3.5 cm and19%/81% at 6 cm. This does not include the backgroundsynchrotron contribution from the Galactic plane, whichbegins to dominate the Galaxy’s flux density for wave-lengths longer than 6 cm. Also, some of these sourcesin the field are likely to be in the foreground of the GCregion, although the density of gas and stars is generally much higher in the GC region, so most sources are likelyto truly be in the central few hundred parsecs of theGalaxy. The high fraction of nonthermal emission in theradio continuum emission is consistent with the idea thatthe cosmic ray density is enhanced in the nuclear disk,assuming that the magnetic field is not unusually strong( ∼ Facilities:
GBT ()
REFERENCESAharonian, F., et al. 2006, Nature, 439, 695Altenhoff, W. J., Downes, D., Pauls, T., & Schraml, J. 1979,A&AS, 35, 23Bally, J., Stark, A. A., Wilson, R. W., & Henkel, C. 1987, ApJS,65, 13Beck, R. & Krause, M. 2005, AN, 326, 414Becker, R. H. & Helfand, D. J. 1985, Nature, 313, 115Becker, R. H., White, R. L., Helfand, D. J., & Zoonematkermani,S. 1994 ApJS, 91, 347Brogan, C. & Goss, W. M. 2003, AJ, 125, 272Brogan, C. L., Nord, M., Kassim, N., Lazio, J., &Anantharamaiah, K. 2003, ANS, 324, 17Brown, R. L. & Johnston, K. J. 1983, ApJ, 268, L85Cotera, A., et al. 2007, in preparationCrocker, M. et al. 2007, ApJ, 666, 934De Pree, C. G., Wilner, D. J., Deblasio, J., Mercer, A. J., &Davis, L. E. 2005, ApJ, 624L, 101Duncan, A. R., Stewart, R. T., Haynes, R. F., & Jones, K. L.1995, MNRAS, 277, 36Figer, D. F., Rich, R. M., Kim, S. S., Morris, M., & Serabyn, E.2004, ApJ, 601, 319Frail, D. A., Goss, W. M., Reynoso, E. M., Giacani, E. B., Green,A. J. & Otrupcek, R. 1996, AJ, 111, 1651Gaume, R. A., Claussen, M. J. , De Pree, C. G., Goss. W. M. &Mehringer, D. M. 1995, ApJ, 449, 663Gaensler, B., Fogel, J. K. J., Slane, P. O., Miller, J. M., Wijnands,R., Eikenberry, S. S., & Lewin, W. H. G. 2003, ApJ, 594, L35Gaustad, J. E., McCullough, P. R., Rosing, W. and Van Buren,D. 2001, PASP, 113, 1326Ghez, A., et al. 2005, ApJ, 620, 744Gordon, M. A. 1974, IAUS, 60, 477Gray, A.D., et al. 1993, MNRAS, 264, 678Gray, A. D. 1994, MNRAS, 270, 835Gray, A.D., Nicholls, J., Ekers, R. D., & Cram, L. E. 1995, ApJ,448, 164Green, D. A., 2004, Bulletin of the Astronomical Society of India,32, 335Handa, T., Sofue, Y., Nakai, N., Hirabayashi, H., & Inoue, M.1987, PASJ, 39, 709Haynes, R. F., Stewart, R. T., Gray, A. D., Reich, W., Reich ,P.,& Mebold, U. 1992, A&A, 264, 500Heiles, C., Reach, W. T., & Koo, B.-C. 1996, ApJ, 466, 191Helfand, D. J. & Becker, R. H. 1987, ApJ, 314, 203Heyvaerts, J., Norman, C., & Pudritz, R. E. 1988, ApJ, 330, 718Hollis, J. M., Jewell, P. R., Remijan, A. J., & Lovas, F. J. 2007,ApJ, 660L, 125Huettemeister, S., Dahmen, G., Mauersberger, R., Henkel, C.,Wilson, T. L., & Martin-Pintado, J. 1998, A&A, 334, 646Huynh, M. T., Jackson, C. A., Norris, R. P., & Prandoni, I. 2005,AJ, 130, 1373 Kassim, N. E. & Frail, D. A. 1996, MNRAS, 283, L51Koyama, K. et al. 2007, PASJ, 59S, 221Kruegel, E., Tutukov, A., & Loose, H. 1983, A&A,124, 89Lang, C. C., Goss, W. M., & Morris, M. 2002, AJ, 124, 2677LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D.2000, AJ, 119, 207LaRosa, T. N., Brogan, C. L., Shore, S. N., Lazio, T. J., Kassim,N. E., & Nord, M. E. 2005, ApJ, 626, L23LaRosa, T. N., Lazio, T. J. W., & Kassim, N. E.2001, ApJ, 563,163Law, C., et al. 2008, in preparationLeahy, D. A. 1989, A&A, 216, 193Lerche, I. & Schlickeiser, R. 1982, A&A, 107, 148Lis, D. C. & Carlstrom, J. E. 1994, ApJ, 424, 189Lis, D. C. & Goldsmith, P. F. 1989, ApJ, 337, 704Liszt, H. S. 1992, ApJS, 82, 495Liszt, H. S. & Spiker, R. W. 1995, ApJS, 98, 259Little, A. G. 1974, IAUS, 60, 491Maeda, Y. 2002, ApJ, 570, 671Mehringer, D. M., Yusef-Zadeh, F., Palmer, P., & Goss, W. M.1992, ApJ, 401, 168Melia, F. & Falcke, H. 2001, ARA&A, 39, 309Mezger, P. G., Duschl, W. J., & Zylka, R. 1996, A&ARv, 7, 289Mezger, P. G., Zylka, R., Salter, C. J., Wink, J. E., Chini., R.,Kreysa, E., & Tuffs, R. 1989, A&A, 209, 337Morris, M. & Serabyn, E. 1996, ARA&A, 34, 645Nord, M. E., Lazio, T. J. W., Kassim, N. E., Hyman, S. D.,LaRosa, T. N., Brogan, C. L., & Duric, N. 2004, AJ, 128, 1646Oka, T., Geballe, T. R., Goto, M., Usuda, T., & McCall, B. J.2005, ApJ, 632, 882Pacholczyk, A. G. 1970, Radio Astrophysics, Freeman and Co.,San FranciscoPedlar, A., Anantharamaiah, K. R., Ekers, R. D., Goss, W. M.,van Gorkom, J. H., Schwarz, U. J., & Zhao, J-H. 1989, ApJ,342, 769Platania, et al. 1998, ApJ, 505, 473Pohl, M. & Schlickeiser, R. 1992, A&A, 263, 37Reich, W. & Fuerst, E. 1984, A&AS, 57, 165Reich, P. & Reich, W. 1988, A&AS, 74, 7Reich, P. & Reich, W. 1988, A&A, 196, 211Reich, W., Reich, P., & F¨urst, E. 1990, A&AS, 83, 539Roberts, D. A., Yusef-Zadeh, F., & Goss, W. M. 1996, ApJ, 459,627Rodr´ıguez-Fern´andez, N. J., Mart´ın-Pintado, J., & de Vicente, P.2001, A&A, 377, 631Sato, F. Hasegawa, T., Whiteoak, J. B., & Miyawaki, R. 2000,ApJ, 535, 857Sawada, T., Hasegawa, T., Handa, T., & Cohen, R. J. 2004,MNRAS, 349, 1167Schmidt, J. 1978, Ph.D. thesis, Univ. of Bonn Serabyn, E. & Morris, M. 1994, ApJ, 424, L91Simpson, J. P. et al. 2007, ApJ preprint doi:10.1086/522295Sofue, Y. 2003, PASJ, 55, 445Sofue, Y. & Handa, T. 1984, Nature, 310, 568Tsuboi, M., et al. 1986, AJ, 92, 818Uchida, K. I., Morris, M., Bally, J., Pound, M., & Yusef-Zadeh, F.1992, ApJ, 398, 128Uchida, K. I., et al. 1994, ApJ, 421, 505Wang, Q. D., Gotthelf, E. V., & Lang, C. C. 2002, Nature, 415,148Yusef-Zadeh, F., Hewitt, J., & Cotton, W. 2004, ApJS, 155, 421Yusef-Zadeh, F., Law, C., Wardle, M., Wang, Q. D., Fruscione,A., Lang, C. C., & Cotera, A. 2002, ApJ, 570, 665 Yusef-Zadeh, F. & Morris, M. 1987, ApJ, 320, 545Yusef-Zadeh, F., Morris, M., & Chance, D. 1984, Nature, 310, 557Yusef-Zadeh, F., Muno, M., Wardle, M., & Lis, D. C. 2007, ApJ,656, 847Yusef-Zadeh, F., Uchida, K. I., & Roberts, D. A. 1995, Science,270, 1801Yusef-Zadeh, F., Wardle, M., Rho, J., & Sakano, M. 2003, ApJ,585, 319Yusef-Zadeh, F., Wardle, M., & Roy, S. 2007, ApJ, 665L, 123 TABLE 1Overview of GBT Surveys of GC Region
Band λ ν
Resolution Long. Range Lat. Range Sensitivity b (cm) (GHz) (arcmin) (deg) (deg) (mJy beam − )X 3.5 8.50 1.5 357.5, +1.5 –0.7, +0.35 9C 6.2 4.85 2.5 357.5, +1.5 a –0.7, +0.35 a
20L 21.3 1.42 9.0 355.5, +7.6 –5.1, +3.4 300P 92.3 0.325 38.8 356.4, +7.6 –5.4, +5.4 4000 a Coverage also extends up to b = +0 . ◦ l = 359 . ◦ ◦ b Measured away from obvious sources, but tends to include background synchrotronemission at longer wavelengths.
TABLE 23.5 cm Compact Source Catalog for GC Region S p a σ S p a S i σ S i bmaj bmin bpa S H i (deg) (deg) (J2000) (J2000) (Jy/bm) (Jy/bm) (Jy) (Jy) (arcsec) (arcsec) (deg) (Jy)1 1.128 –0.100 17:48:40.29 –28:01:23.5 2.29 1.7E-2 3.48 4.0E-2 115.2 102.3 24 8 . ± . ± . ± . ± . ± . ± . ± . ± a GBT beam FWHM at 3.5 cm is 88 (cid:48)(cid:48) . TABLE 36 cm Compact Source Catalog for GC Region S p a σ S p a S i σ S i bmaj bmin bpa(deg) (deg) (J2000) (J2000) (Jy/bm) (Jy/bm) (Jy) (Jy) (arcsec) (arcsec) (deg)1 0.862 +0.081 17:47:20.61 –28:09:25.4 3.81 1.5E-1 7.35 4.2E-1 224.0 200.8 1232 0.724 –0.090 17:47:41.07 –28:21:51.9 1.79 1.5E-1 3.81 4.5E-1 223.4 212.8 1383 0.670 –0.034 17:47:20.33 –28:22:51.2 21.73 1.6E-1 32.8 3.6E-1 199.5 176.5 684 0.673 +0.084 17:46:53.31 –28:19:05.4 0.54 1.5E-1 1.49 5.5E-1 266.4 243.0 1135 359.282 –0.258 17:44:54.68 –29:40:59.6 1.34 1.6E-1 1.79 3.3E-1 177.9 175.9 926 358.791 +0.063 17:42:27.76 –29:55:59.2 0.45 1.6E-1 0.52 3.1E-1 192.2 139.0 787 358.630 +0.066 17:42:03.57 –30:04:05.8 0.46 1.6E-1 0.57 3.2E-1 182.0 157.5 848 358.606 –0.061 17:42:30.04 –30:09:20.1 0.86 1.5E-1 1.60 4.1E-1 239.6 180.9 1169 358.004 –0.634 17:43:17.56 –30:58:07.7 0.29 1.6E-1 0.29 2.8E-1 160.5 148.1 5110 357.990 –0.158 17:41:22.48 –30:43:47.2 0.54 1.6E-1 0.77 3.5E-1 197.7 168.3 94 a GBT beam FWHM at 6 cm is 153 (cid:48)(cid:48) . TABLE 4Spectral Index for CompactSources Detected at 3.5 and 6 cm α CX σ α a (deg) (deg)1 0.862 +0.081 –0.93 0.102 0.724 –0.090 –1.42 0.213 0.670 –0.034 –0.28 0.024 0.673 +0.084 –1.88 0.675 359.282 –0.258 –0.08 0.336 358.791 +0.063 0.00 1.077 358.630 +0.066 –0.27 1.018 358.606 –0.061 –0.87 0.469 358.004 –0.634 0.00 1.7310 357.990 –0.158 –0.17 0.82 a Error in spectral index based on statisti-cal errors and do not account for absoluteflux calibration errors. For the expected5% absolute flux errors in the 3.5 and 6cm maps, a spectral index uncertainty of ∼ .
13 should be added in quadrature tothese errors. TABLE 5Extended Source Catalog for 3.5 and 6 cm GBT Observataions of GC Region R eff raw S i,X σ bg,X S i,X σ S i,X raw S i,C σ bg,C S i,C σ S i,C (deg) (deg) (J2000) (J2000) (arcmin) (Jy) (Jy/bm) (Jy) (Jy) (Jy) (Jy/bm) (Jy) (Jy)Tornado 357.66 –0.09 17:40:17.0 –30:58:12 6.5 14.16 9.7E-3 13.81 0.51 18.45 1.6E-2 18.27 0.41G357.7–0.4 357.70 –0.44 17:41:46.6 -31:07:35 7.0 2.09 1.0E-2 1.55 0.71 1.30 1.6E-2 1.05 0.55G358.4+0.1 358.38 +0.12 17:41:15.9 –30:15:01 4.0 1.52 7.8E-3 1.24 0.14 3.41 1.9E-2 1.85 0.16RF E3 358.54 –0.29 17:43:14.1 –30:19:56 6.5 1.99 9.0E-3 1.51 0.45 4.28 2.3E-2 1.99 0.56Sgr E th a a b c c a b d d a The “th” appended to the source name emphasizes that the source region is defined for the thermal-emitting part of the complex. b The “nt” appended to the source name emphasizes that the source region is defined for the nonthermal-emitting part of the complex. c Flux includes Arched filaments extended source, which is subtracted with the background. d Flux includes the Sgr B2 point source, which is subtracted with the background. TABLE 6Spectral Indices for Extended Sources
Name l b α CX σ α a Thermal/Nonthermal(deg) (deg)Tornado 357.66 –0.09 –0.50 0.07 NTG357.7–0.4 357.70 –0.44 +0.69 1.04 TG358.4+0.1 358.38 +0.12 –0.71 0.23 NTSgr E th b b c b c a Error in spectral index based on statistical errors and do not account for ab-solute flux calibration errors. For the expected 5% absolute flux errors in the3.5 and 6 cm maps, a spectral index uncertainty of ∼ .
13 should be added inquadrature to these errors. b The “th” appended to the source name emphasizes that the source region isdefined for the thermal-emitting part of the complex. c The “nt” appended to the source name emphasizes that the source region isdefined for the nonthermal-emitting part of the complex.
TABLE 7Catalog of Known and Candidate Supernova Remnants Observed at 3.5, 6, and 20 cm
Name RA Dec Size S X S C α int CX α slice CX α slice LC Figures a References b (J2000) (J2000) (arcmin) (Jy) (Jy)G357.7–0.1 (Tornado) c × . ± .
51 18 . ± . − . ± . ∼ − . ∼ − .
63 5,8 1,2G357.7+0.3 d ∼ .
04 to ∼ − . . ± .
84 6 . ± . − . ± . − . − . ∼ − . d × − . ± . − . ± .
02 14,15 1,7,8G0.33+0.04 d × − . − . e . ± .
35 9 . ± . − . ± . − .
35 to 0.08 − .
35 23,11 1,5G1.0–0.2 e × . ± .
53 9 . ± . − . ± . − .
92 to − . − .
47 23,11 1,6 a The figure number featuring the SNR is listed here. b References that discuss object — 1: Gray (1994), 2: Becker & Helfand (1985), 3: Reich & Fuerst (1984), 4: Uchida et al. (1992), 5: Helfand & Becker (1987), 6: Liszt (1992),7: Pedlar et al. (1989), 8: Yusef-Zadeh & Morris (1987), 9: Kassim & Frail (1996) c The Tornado is a candidate supernova remnant. d Source is only partially surveyed or confused with other sources, so no integrated characteristics are given. e Flux densities are given for the shell component only. TABLE 8Flux Contributions by Thermal and Nonthermal Sources at 3.5 and 6 cm
Extended Catalog Extended and Compact Catalogs a Type 6 cm 3.5 cm α CX α CX (Jy)/(%) (Jy)/(%) (Jy)/(%) (Jy)/(%)thermal 147 ±
43 (15 ±
4) 173 ±
33 (24 ±
5) 0 . ± .
63 202 ±
48 (19 ±
4) 220 ±
33 (28 ±
4) 0 . ± . ±
44 (85 ±
4) 552 ±
19 (76 ± − . ± .
11 862 ±
46 (81 ±
4) 563 ±
19 (72 ± − . ± . ±
87 (100) 725 ±
52 (100) − . ± .
20 1063 ±
93 (100) 783 ±
52 (100) − . ± . a Includes all sources listed in Tables 2, 3, and 5. See text in § Fig. 1.—
The GBT radio continuum survey of GC region at 3.5 and 6 cm is shown in the top and bottom panels, respectively. Thesurveys cover a similar region of roughly 4 ◦ × ◦ , although the 6 cm survey also covered a region at higher positive latitudes. Contours forthe 3.5 cm survey are at levels of 0 . ∗ n Jy beam − , for n = 0 − (cid:48)(cid:48) . The contours for the 6 cm survey are identical,but for n = 1 − (cid:48)(cid:48) . Galactic coordinates are shown on each image. Fig. 2.—
The GBT radio continuum survey of GC region at 20 and 90 cm is shown in the top and bottom panels, respectively. Thesurveys cover a similar region of roughly 10 ◦ × ◦ . Contours for the 20 cm survey are at 5, 10, 15, 20, 30, 40, 80, 160, and 320 Jy beam − , with a beam size of 9 (cid:48) . Contours for the 90 cm survey are at 80, 160, 320, 640, and 1280 Jy beam − , with a beam size of 38 . (cid:48) Fig. 3.—
Top : 3.5 cm image with the locations of two slices used to demonstrate the spectral index slice analysis method. The slicespass through Sgr B2 and G0.9+0.1, two objects with well-known radio spectral indices. Both slices are labeled with the 6/3.5 cm spectralindex at the brightest part of the slice at 3.5 cm.
Middle and bottom : The flux and spectral index for the slices shown above. Each plotshows slices for two frequencies with the 3.5 cm (“X”) slice below the 6 cm (“C”) slice. The x axis shows the entire length of the slice inunits of arcmin. The slice brightnesses are indicated by the left axis and the spectral index value is shown on the right axis. The best-fitbackground to each slice is shown as a dashed line. The dotted line shows the “source” region, which is ignored in the determination ofthe background. The rms deviation of the background data about the best-fit line is shown as an error bar to the right of each slice. Partsof the slice with spectral index error, σ α <
1, are plotted. Fig. 4.—
Top : GBT map of 3.5 cm continuum emission showing the extended sources listed in Table 5 and the compact sources givenin Table 2. The brightness of the image is shown in a logarithmic scale from 0 to 5 Jy beam − . Bottom : GBT map of 6 cm continuumemission with the same regions overlaid, except the circles show 6 cm compact sources given in Table 3. The brightness of the image isshown in a logarithmic scale from 0 to 10 Jy beam − . The polygon regions define extent of extended sources and are identified with theircommon names. The flux density, location, and other properties of the extended sources are given in Tables 5 and 6. Fig. 5.—
Top, left : 3.5 cm image of the region around G357.7–0.1 (Tornado) with contours at levels of 0 . ∗ n Jy beam − , for n = 0 − Top, right : Grayscale shows the 6 cm image of the same region with slices and corresponding spectral index at the peak brightness of eachslice.
Bottom : Brightnesses and the corresponding spectral index are shown for the slice with α CX = − . ± .
02 in the top right figure(along the long axis of the Tornado). Note that comparing these spectral indices to other works should account for the flux calibrationuncertainty by adding a spectral index error of 0.13 in quadrature. Fig. 6.—
Same as for Fig. 5, but for the E3 radio filament (G358.60-0.27). The plotted slice values correspond to the slice with α CX = 0 . ± .
18 with the origin at the southeast. Fig. 7.—
Same as for Fig. 5, but for the Sgr E complex (G358.7-0.0). The plotted slice values correspond to the slice with α CX = − . ± .
09 with the origin at the northeast. Fig. 8.—
Same as for Fig. 5, but the left and right images show 6 and 20 cm emission in the west of the survey, respectively. Contourson the 6 cm survey are at 0 . ∗ n Jy beam − , for n = 0 −
10. The plotted slice values correspond to the slice with α LC = − . ± . Fig. 9.—
Same as for Fig. 5, but for G359.1–0.5 and the Snake (G359.1-0.2). The plotted slice values correspond to the slice with α CX = − . ± .
13 with origin at the south. Fig. 10.—
Plot of the 6/3.5 cm spectral index values as a function of position on the circular supernova remnant, G359.1–0.5. Positionis represented by an angle measured in degrees, increasing counterclockwise starting at galactic north. Fig. 11.—
Same as for Fig. 8, showing images at 6 and 20 cm for the G359.1–0.5 SNR. The plotted slice values correspond to the slicewith α LC = − . ± .
07 with origin at the east. Fig. 12.—
Same as for Fig. 5, but for the Sgr C complex (G359.5-0.0). The plotted slice values correspond to the slice with α CX = − . ± .
13 with origin at the south. Fig. 13.—
Same as for Fig. 5, but for the G359.8–0.3 complex. The plotted slice values correspond to the slice with α CX = 0 . ± . Fig. 14.—
Same as for Fig. 5, but for the Sgr A complex (G0.07+0.04). The plotted slice values correspond to the slice with α CX = − . ± .
02 and the rightmost label with origin at the southeast. Fig. 15.—
Same as for Fig. 8, showing images at 6 and 20 cm for the Sgr A complex (G0.0+0.0). The plotted slice values correspond tothe slice with α LC = − . ± .
01 with origin at the east. Fig. 16.—
Same as for Fig. 5, but for the Arched filaments complex (G0.07+0.04). The plotted slice values correspond to the slice shownat the top right with origin at the south. Fig. 17.—
Same as for Fig. 5, but for the emission east of the Radio Arc (G0.2-0.0). The plotted slice values correspond to the sliceshown at the top right with origin at the north. At position near 10 (cid:48) , the slice crosses the SNR G0.33+0.04 and has a spectral index α CX of roughly –2.0 to –1.5. Fig. 18.—
Same as for Fig. 5, but for the G0.5–0.5 complex. The plotted slice values correspond to the slice with α CX = − . ± . Fig. 19.—
Same as for Fig. 8, showing images at 6 and 20 cm for the G0.5–0.5 complex. The plotted slice values correspond to the slicewith α LC = − . ± .
11 with origin at the east. Fig. 20.—
Same as for Fig. 5, but for the Sgr B complex (G0.5–0.0 and G0.7–0.0). The plotted slice values correspond to the slice with α CX = 0 . ± .
03 with origin at the south. Fig. 21.—
Same as for Fig. 8, showing images at 6 and 20 cm for the Sgr B (G0.5–0.0 and G0.7–0.0) complex and eastern sources. Theplotted slice values correspond to the slice with α LC = − . ± .
02 with origin at the north. Fig. 22.—
Same as for Fig. 5, but for the G0.8–0.4 complex. The plotted slice values correspond to the slice with α CX = − . ± . Fig. 23.—
Same as for Fig. 5, but for G0.9+0.1 and eastern complex. The plotted slice values correspond to the slice with α CX = − . ± .