Sgr A* and its Environment: Low Mass Star Formation, the Origin of X-ray Gas and Collimated Outflow F. Yusef-Zadeh1, M. Wardle2, R. Sch¨odel3, D. A. Roberts1, W. Cotton4, H. Bushouse5, R. Arendt6, & M. Royster1 1CIERA, Department of Physics and Astronomy Northwestern University, Evanston, IL 60208 2Department of Physics and Astronomy, Macquarie University, Sydney NSW 2109, 6 Australia 1 0 3Instituto de Astrofi(cid:48)sica de Andaluci(cid:48)a (CSIC), Glorieta de la Astronomica, 18008 2 n Granada, Spain a J 4National Radio Astronomy Observatory, Charlottesville, VA 22903 1 5Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218 ] A 6NASA GSFC, Code 665, Greenbelt, MD 20771 G . h p ABSTRACT - o r t We present high-resolution multiwavelength radio continuum images of s a the region within 150(cid:48)(cid:48) of Sgr A*, revealing a number of new extended features [ and stellar sources in this region. First, we detect a continuous 2(cid:48)(cid:48) east-west 1 ridge of radio emission, linking Sgr A* and a cluster of stars associated with v 6 IRS 13N and IRS 13E. The ridge suggests that an outflow of east-west blob- 1 1 like structures is emerging from Sgr A*. In particular, we find arc-like radio 0 structures within the ridge with morphologies suggestive of photoevaporative 0 . protoplanetary disks. We use infrared K and L(cid:48) fluxes to show that the 1 s 0 emissionhassimilarcharacteristicstothoseofaprotoplanetarydiskirradiated 6 1 by the intense radiation field at the Galactic center. This suggests that star v: formationhastakenplacewithintheScluster2(cid:48)(cid:48) fromSgrA*. Wesuggestthat i X the diffuse X-ray emission associated with Sgr A* is due to an expanding hot r wind produced by the mass loss from B-type main sequence stars, and/or the a disks of photoevaporation of low mass young stellar objects (YSOs) at a rate ∼ 10−6M yr−1. The proposed model naturally reduces the inferred accretion (cid:12) rate and is an alternative to the inflow-outflow style models to explain the underluminous nature of Sgr A*. Second, on a scale of 5(cid:48)(cid:48) from Sgr A*, we detect new cometary radio and infrared sources at a position angle PA∼ 50◦ – 2 – which is similar to that of two other cometary sources X3 and X7, all of which face Sgr A*. In addition, we detect a striking tower of radio emission at a PA∼ 50◦−60◦ along the major axis of the Sgr A East SNR shell on a scale of 150(cid:48)(cid:48) fromSgrA*. Wesuggestthatthecometarysourcesandthetowerfeature are tracing interaction sites of a mildly relativistic jet from Sgr A* with the atmosphere of stars and the nonthermal Sgr A East shell at a PA∼ 50−60◦ with M˙ ∼ 1×10−7M yr−1, and opening angle 10 degrees. Lastly, we suggest (cid:12) that the east-west ridge of radio emission traces an outflow that is potentially associated with past flaring activity from Sgr A*. The position angle of the outflow driven by flaring activity is close to −90◦ which is different than the PA∼ 60◦ of the radio ridge. Subject headings: accretion, accretion disks — black hole physics — Galaxy: center 1. Introduction A 4×106 M black hole is coincident with the compact nonthermal radio source Sgr A* at (cid:12) the center of the Galaxy (Ghez et al. 2008; Gillessen et al. 2009; Reid and Brunthaler 2004). The estimated mass accretion rate onto Sgr A* is several orders of magnitude smaller than the rate at which young, windy stars in the innermost 0.5 pc supply mass to the Bondi radius of Sgr A* (Coker & Melia 1997; Cuadra et al. 2006, 2008). Chandra observations have characterized the X-ray emission surrounding Sgr A* to be spatially extended with a radius of ∼1.5(cid:48)(cid:48) (Baganoff et al. 2003; Wang et al. 2013). The X-ray luminosity is interpreted as arising from a radiatively inefficient accretion flow (RIAF, e.g. Yuan et al. 2004; Moscibrokzka et al. 2009). In this model, a fraction of the gaseous material accretes onto Sgr A* and the rest is driven off as an outflow from Sgr A* (e.g., Quataert 2004; Shcherbakov and Baganoff 2010; Wang et al. 2013). Another mechanism that may reduce the accretion rate is interaction with a jet or an outflow limiting the amount of gas falling onto Sgr A* (Yusef-Zadeh et al. 2014a), thus modifying the accretion flow. In this picture, the interaction of the outflow with the surrounding gas or the atmosphere of mass-losing stars can provide an estimate of the power of the outflow. Two different types of activity are associated with Sgr A*. One is flaring on hourly time scale at multiple wavelengths (e.g., Baganoff et al. 2001; Genzel et al. 2003). Observations of Sgr A* have detected a time delay at submm, mm, and radio wavelengths consistent with a scenario in which plasma blobs expand away from the disk, becoming visible at successive longer wavelengths as the optical depths become of order unity effects (Yusef-Zadeh et al. 2006, 2008, 2009; Marrone et al. 2008; Eckart et al. 2008; Brinkerink et al. 2015). The other is a jet-driven outflow (e.g., Falcke & Markoff 2000). Unlike the flare activity, the existenceofajetfromSgrA*hasnotbeenfirmlyestablishedbecauseofthecomplexthermal and nonthermal structures in this confused region of the Galaxy. At least five independent investigations based on X-ray, near-IR, and radio observations have suggested that a jet is emanating from Sgr A*. These studies have found discrepant values for the jet position angle – 3 – (PA) and inclination (Markoff et al. 2007; Broderick et al. 2011; Zamaninasab et al. 2011; Muno et al. 2008; Li et al. 2013; Yusef-Zadeh et al. 2012; Shahzamanian et al. 2015). It is possible that some of the gas approaching Sgr A* is pushed away as part of an expanding hot plasma driven by flaring and jet activity, resulting Sgr A*’s low radiative efficiency. Thus, the presence of collimated structures from Sgr A* is critical in distinguishing between the competing accretion and outflow models. The Galactic center is a challenging region in which to image a radio jet or a flare close to Sgr A*, because of the limited spatial resolution and dynamic range caused by confusing sources, scatter broadening, and intrinsic temporal variability of Sgr A* on hourly time scales (e.g., Bower et al. 2014). Here we present sensitive observations of the Galactic center at multiple radio frequencies obtained using the improved broad-band capability of the VLA, finding new radio structures interpreted to be associated with Sgr A* activity. On a scale of a few arcseconds from Sgr A*, we identify a ridge of east-west radio emission which bends toward the SW in the direction away from Sgr A*. This ridge, which is detected to the west of Sgr A*, shows a number of blobs and arc-like features surrounded by a diffuse plume-likestructure. Weinterpretthattheplume-likefeatureasarisingfromflaringactivity, thus produces an outflow from the direction of Sgr A* on a scale of ∼0.1 pc, with an opening angle of ∼ 35◦. The position angle of the outflow driven by flaring activity of Sgr A* is consistent with the east-west elongation of Sgr A* observed on milliarcsecond (mas) scale (Bower et al. 2015). A large number of stars being members of the so-called “S-cluster”1 also lie along the ridge. We compare infrared and radio images of the ridge and argue that bow-shock structures detected within the inner 2(cid:48)(cid:48) of Sgr A* are proplyd candidates. We show that radio sources with comparable scale size to those associated with proplyds can not be simply blobs of dusty ionized gas but are associated with a reservoir of hot dust surrounded by ionized gas. On a larger scale, within two arcminutes NE of Sgr A*, we find new cometary sources (F1, F2 and F3) pointing toward Sgr A* and a large-scale tower-like structure associated with the Sgr A East supernova remnant (SNR) 150(cid:48)(cid:48) from Sgr A*. The position angle of these new structures is similar to two other IR-identified cometary sources, X3 and X7, found to the SW of Sgr A* (Muzic et al. 2007, 2010). One interpretation that we put forth is that these features could be the result of a jet from Sgr A* interacting with the atmosphere of dusty stars near Sgr A* and with the Sgr A East shell, respectively. In addition, we suggest that the diffuse X-ray emission centered on Sgr A* arises through hot gas created by the collision of stellar winds from B stars in the S-star cluster or young low-mass stars (c.f. Loeb 2004). This nuclear wind created by mass-losing stars near Sgr A* produces hot expanding X-ray gas (c.f. Quataert 2004) that excludes the shocked winds from O and WR stars in the central parsec of the Galaxy and prevents accretion onto Sgr A*. 1A loosely defined term consisting of ∼30 stars within a projected distance of 1” from Sgr A*, 2/3 of which are spectroscopically classified as O/B stars with orbital periods of a dozen to few hundred years and the rest older stars (e.g., Genzel et al. 2003). We point out that the stars in this region have different spectral types and ages and probably heterogeneous origins – 4 – Meanwhile Sgr A* accretes material from the cluster winds at a much lower rate potentially explainingthelowluminosityofSgrA*withouttheejectionofalargefractionoftheaccreted material. 2. Observations and Data Reduction 2.1. Radio Data Multi-wavelength radio continuum observations were carried out with the Karl G. Jansky Very Large Array (VLA)2 in its A-configuration at 44, 34.5, 8.5, 5.5 and 1.4 GHz during March and April 2014. Table 1 gives columns of the date, center frequency, bandwidth, the number of subbands (IF), the number of channels, and the spatial resolution of each observation. In all observations, we used 3C286 to calibrate the flux density scale, 3C286 and J1733-1304 (aka NRAO530) to calibrate the bandpass, and J1744-3116 to calibrate the complex gains. The broad 8 GHz bandwidths at 34 and 44 GHz, 2 GHz bandwidths at 5 and 14 GHz, and 1 GHz bandwidth at 1.4 GHz provide a significant improvement over earlier observations that had only 100 MHz of bandwidth. We observed Sgr A* using the 3-bit sampler system at 34 and 44 GHz, which provided full polarization correlations. 2.2. Infrared Data Details of the near-IR observations and data reduction of the Galactic center at K and L(cid:48) s bands, centralwavelengths2.18, and3.8µm, respectively, wererecentlygiveninYusef-Zadeh et al. (2015). TheseobservationsusedadaptiveopticsandwereacquiredwithVLT/NACO3, with a pixel scale of 0.027” per pixel. L(cid:48)-band observations were obtained in speckle mode using five fields with different pointing and depths. Field1 was centered on Sgr A*, and Fields 2–5 were offset by approximately 20(cid:48)(cid:48) to the northeast, southeast, southwest, and northwest, respectively. Standard near-IR image data reduction was applied followed by combining individual pointings into large mosaics. In the case of the L(cid:48) image, the speckle holography technique, as described in Scho¨del et al. (2013), was applied to the thousands of obtained speckle frames to create final high-Strehl images for each pointing. Finally, we calibratedtheimagesastrometricallybyusingthepositionsandpropermotionsofSiOmaser stars in the Galactic center (Reid et al. 2007). ImagingdatainM(cid:48)-bandweretakenwithNACO/VLTin2003, 2004, and2006. Weretrieved the data from the ESO archive. Chopping was used for background subtraction. The images were flat-fielded and corrected for bad pixels. Since the chop throw was small in all observa- tions (due to technical limitations at the VLT) the images from the individual epochs show 2Karl G. Jansky Very Large Array (VLA) of the National Radio Astronomy Observatory is a facility of the National Science Foundation, operated under a cooperative agreement by Associated Universities, Inc. 3Based on observations made with ESO Telescopes at the La Silla or Paranal Observatories under pro- grams ID 089.B-0503 – 5 – strong negative residuals from stellar and diffuse sources in the off-target chop positions. However, some dithering was applied during the observing runs and the initial pointing as well as the chopping angle were different for the different observing epochs. Therefore, residuals could be effectively removed by averaging the images from all observing epochs; the images were sub-pixel shifted onto a common position of the centroid of the star IRS 16C and median combined. Any remaining artifacts from chopping were effectively removed by rejecting the lowest 20% at each pixel before calculating the median value. Approxi- mately 12000 frames, each with 0.056 s integration time, were combined, corresponding to an accumulated exposure time of about 800 s. The photometry was calibrated by assuming constant extinction between the L and M bands and the same magnitudes at both bands for the sources IRS 16C and IRS 16NW. Their L’-band fluxes were taken from Scho¨del et al. (2010). The uncertainty of the zero point was estimated to be 0.15 mag. 3. Results Because of better sensitivity to detect weak radio emission, we first present newly recognized features within a few arcseconds of Sgr A* and then compare the positions of near-IR iden- tified stellar sources with radio data. Lastly, we present large structures on a scale between a few arcseconds to two arcminutes surrounding Sgr A* and identify new features associated with the Sgr A East SNR shell. 3.1. Small Scale Radio Features 3.1.1. The Sgr A* Radio Ridge Figure 1a,b show images of the inner 10(cid:48)(cid:48) of Sgr A* at mean frequencies (wavelengths) 34.5 GHz (9mm) and 14.1 GHz (2cm). Prominent stellar and ionized features are labeled. These images reveal a ridge of emission to the west of north of Sgr A* extending for about 1.5– 2(cid:48)(cid:48) at position angles ranging between ∼90 and −125◦. This ridge of emission, which was previously detected in low resolution images at 8 GHz (Wardle and Yusef-Zadeh 1992) links Sgr A* to the western edge of the minicavity, an ionized feature with a diameter of 2(cid:48)(cid:48). The western edge of the minicavity coincides with the two stellar clusters IRS 13N and IRS 13E. The Sgr A* radio ridge consists of a number of blob-like and arc-like structures with angular scales of 0.2–0.5(cid:48)(cid:48). The two arc-like structures coincide with source (cid:15) which has been detected in earlier narrow bandwidth images with low spatial resolutions at 15 GHz (Yusef-Zadeh et al. 1990; Zhao et al. 1991). To illustrate the asymmetric nature of the ridge structure, Figure 1c shows a saturated image of the region shown in Figure 1a,b. We interpret that this structure has an opening angle of ∼ 35◦ pointed in the direction toward W and SW from Sgr A*. Figure 1d shows another 34.5 GHz rendition of the ridge of emission in reverse color, but at higher resolution than that of Figure 1a. The images at multiple frequencies confirm the reality of blob and arc-like structures along the ridge. Note the diffuse plume-like structure, as drawn schematically on Figure 1a, surrounding the blob and arc-like structures. – 6 – The widening of the plume-like structure away from Sgr A* suggests that a plume of gaseous material is moving outward and expanding away, suggesting that Sgr A* is responsible for the ridge. Proper motion measurements of the brightest source in the ridge indicate high velocity ionized gas moving away from Sgr A* to the SW (Zhao et al. 2009). This radio ridgeappearsdistinctfromtheionizedgasassociatedwiththemini-spiralHIIregionorbiting Sgr A* and is not confused with numerous dusty stellar sources and diffuse emission found in mid-IR images of this region. Future high resolution proper motion, polarization and spectral index measurements of the ridge will provide additional constraints in the claim that this feature is associated physically to Sgr A*. 3.1.2. Radio Emission from the S Cluster The region within ∼2(cid:48)(cid:48) of Sgr A* where the ridge of radio emission and the plume-like structure are detected, is adjacent to the S-cluster. This cluster consists mostly of young, early type stars with orbital periods of 10 to a few 100 years. The kinematics of the stars in the cluster, which following other workers we refer to as “S stars(cid:48)(cid:48), have been used to measure the mass of Sgr A* (Ghez et al. 2008; Gillessen et al. 2009). Figures 2a,b show contours of 34.5 GHz emission and a grayscale image of a region of 3.5(cid:48)(cid:48) ×2.5(cid:48)(cid:48) centered on Sgr A*. Figure 2c shows contours of 5.5 GHz emission superimposed on a grayscale image. The crosses on Figure 2a,c correspond to eight radio sources RS1–8 found in the plume-like region to the west of Sgr A*. The trajectories of young stars are known from proper motion measurements in the near-IR. To see if these stars have radio continuum counterparts, their positions at the epoch of our 34 GHz observation on March 9, 2014 (2014.19) have been calculated based on proper motions and orbital accelerations derived from near-IR observations (Gillessen et al. 2009; Lu et al. 2006; Yelda et al. 2014). Tables 2, 3 give the positions of S cluster and their corresponding positional uncertainties at the epoch of 2014.19 and from two different catalogs (Gillessen et al. 2006; Lu et al. 2009). Table 4 gives the predicted X/Y offsets (in arcsecond) of stars relative to Sgr A* for the epoch 2014.19. These tables list the positions of the 28, 31 and 117 stars identified by Gillessen et al. (2009), Lu et al. (2006) and Yelda et al. 2014, respectively. Table 5 lists Gaussian-fitted positions of 8 radio sources (RS1–8) embedded within the diffuse extended emission associated with the Sgr A* ridge and the plume-like structure at 34 GHz (see Figs. 1 and 2). Entries in columns 1 to 9 give the name of the source at 34 GHz, alternative names in the literature, the RA and Dec, the angular distance from Sgr A* in increasing order, positional accuracy, the peak intensity and the flux density that are associated with these sources. Figure3a,bsuperimposecontoursofradioemissionona34GHzgrayscaleimage,withcrosses indicating the positions of the stars in the S cluster for which orbits have been determined by Gillessen et al. (2009) and Lu et al. (2009), respectively. Many of the S-stars, as noted in Figure 3a, are projected against the bright, scatter-broadened radio source Sgr A* and the diffuse emission from the Sgr A* ridge, thus they can not be discerned at radio wavelengths. The comparison of radio sources (RSs) listed in Table 5 with stellar sources given in Table – 7 – 2 indicates that RS1, RS2 and RS3 lie within the 1σ position of S33, which is an early type star (Gillessen et al. 2009). These radio sources appear compact but they are embedded within the extended emission from the ridge of emission (see Fig. 2b) so it is not clear if these sources are radio counterparts to S33 unless we measure the proper motion of radio sources. If these radio sources are not randomly coincident with S33 and are associated with stars, we can determine the mass loss rate from the ionized winds. The radio emission is assumed to arise from S33 with a flux density of ∼ 0.2 mJy at 34.5 GHz and is due to a spherically-symmetric, homogeneous wind of fully–ionized gas expanding with a constant terminalvelocity∼700kms−1 (Panagia&Felli1975). ThemasslossrateofS33isestimated to be M˙ = 2×10−6 M yr−1. Clumpiness of the ionized wind would reduce this estimate. (cid:12) Figure 3c shows a grayscale 34 GHz image in negative and labeled with the positions of those stars for which orbits have been determined by Yelda et al. (2014). We note several prominent young stars, IRS 16SW, IRS 16NW, IRS 16C, that are not members of the S-cluster but have radio counterparts (Yusef-Zadeh et al. 2013). Figure 3d shows the distribution of stars over the area shown in Figure 3c at 3.8µm. The crosses correspond to the position of stars (Yelda et al. 2014) at the epoch that radio data were taken. Apart from stars labeled on this figure, the spectral classification of remaining faint sources is unknown. These faint sources are possibly late-type stars associated with the evolved nuclear cluster or young low-mass stars (see §4). Five near-IR stellar sources A–E are labeled on Figure 3d and will be discussed below. 3.1.3. Bow Shock-like Radio Sources RS5 and RS6 Figure 4a,b,c show the relative position of 34.5 GHz sources RS5 – RS8 and 3.8 µm and 2.18 µm sources A – D with respect to each other. Two of the newly detected radio sources in the ridge, RS5 and RS6 located ∼ 1.6(cid:48)(cid:48) from Sgr A* (see Fig. 2), are partially resolved, with size scalesrangingbetween850to1200AU.Aclose-upviewsofthesearc-likestructuresareshown in Figure 4a whereas Figures 4b,c show contours of 3.8 and 2.18µm emission superimposed on a 34.5 GHz image, respectively. A comparison of radio with near-IR images reveals that the arc-like structures RS5 and RS6 have 3.8µm counterparts with a bow shock morphology. There are two near-IR stellar sources, S1-22 (Lu et al. 2009) and a stellar source A, as labeled on Figure 4c. S1-22 lies at the apex of a bow shock-like structure and is an early type star which is projected against the extended radio emission associated with RS6. Star A has no radio counterpart, and is not identified in any catalogs of early type stars near Sgr A*. It is unlikely that star A is associated with RS6 for two reasons. First, proper motion measurements indicate that star A is moving to the NW with a velocity of 117±16 km s−1 (Scho¨del et al.2009). The proper motion of ionized gas associated with the blob (cid:15) which coincides with RS5, RS6 and RS7 is ∼ 338 ± 21 km s−1 to the SW (Zhao et al. 2009). Second, it is unlikely that a late-type star could have a stellar wind strong enough to produce the observed stand-off distance of the bow-shock’s apex, ∼6 milliparsec (mpc). This requires a Wolf-Rayet star (Tanner et al. 2005; Sanchez-Bermudez et al. 2014). Thus, it is unlikely that star A is associated with the bow-shock structure RS5. – 8 – As for S1-22, the proper motion data gives a tangential velocity of ∼326 km s−1 to the SE (Yelda et al. 2014). Radio proper motion measurements have a coarse arcsecond spatial resolution , giving the proper motion of RS5, R6 and RS7, when compared to those of near- IR sources. In spite of the difference in resolution between radio and near-IR proper motion measurements, the magnitude of radio proper motion of RS5, RS6 and RS7 is similar to that of S1-22. Thus, it is possible that S1-22 is physically associated with RS6 with the standoff distance of ∼4 mpc. In addition to star A, three 3.8µm sources B, C and D, labeled on Figures 4b,c, are not found in the catalogs of early type stars. RS7 and RS8 are likely to be radio counterparts to source B, and C, respectively. A radio source with a flux density of 58 µJy at 34 GHz is detected at the position of source D which lies in an extended region associated with the ridge. The bright radio sources B and D are clearly stellar sources since they are detected at L(cid:48) and K bands. The offsets seen in the position of radio and near-IR sources could have s a contribution from proper motion of individual sources. The near-IR images at 2.18 and 3.8µm were taken during June and September 2012 whereas the 34.5 GHz data was taken on March 9, 2014. The L(cid:48) sources associated with RS5, RS7 (or source A), RS8 (or source B), and D, lie to the SSE of their radio counterparts, suggesting either correlated motion, coincidence, or a systematic error in the image registration arising because the radio and infrared images are taken at different epochs. We compared Gaussian fitted positions of B, C and D at 34.5 GHz and L(cid:48) band and found 3.4 and 3 sigma offsets to the north in the positions of stars B and D and their radio counterparts, respectively. We also compared the positional offset for IRS 16C and its radio counterpart and determined that this implies a proper motion of IRS 16 that is roughly twice higher the actual value determined at radio and infrared wavelengths (Lu et al. 2009; Yusef-Zadeh et al. 2015a). Thus, we can not establish that radio sources are counterparts to stellar sources. Given the offsets in positions and the non-detection of source C in K , it is just possible that the radio sources are gas s blobs and are not directly associated with stars. However, it is not clear how gas blobs near Sgr A* could survive the tidal shear of Sgr A* unless they have densities that withstand the tidal shear or that they are transient (Yusef-Zadeh et al. 2015c). 3.2. Large Scale Features 3.2.1. Cometary Sources F1 and F2 There are two known cometary sources, X3 and X7, lying within 3(cid:48)(cid:48) to the SW of Sgr A* at near-IR wavelengths. X3 is located ∼ 3(cid:48)(cid:48) from Sgr A* (Muzic et al. 2007, 2010) showing tail-head structure pointing toward Sgr A*. It has a radio counterpart at 44 GHz (source 16 in Yusef-Zadeh et al. 2014b) and a peak flux density 0.22 mJy beam−1 at 34.5 GHz. The second cometary source X7 (Muzic et al. 2007) is identified in Figure 3d as source E at near-IR. This cometary source also points toward the direction of Sgr A*. A compact 34.5 GHz source with a peak flux density of 100 µJy is detected at the position of source E 0.7(cid:48)(cid:48) from Sgr A*. This source lies too close to Sgr A* where the noise increases near the bright source Sgr A*, thus, structural details of this source are not clear at radio wavelengths. We – 9 – detect a third radio source with a cometary morphology similar to X3 and X7. Unlike X3 and X7, this source, which we denote F1, lies 4.8(cid:48)(cid:48) to the NE of Sgr A*. Figure 5a,b show grayscale contours of this cometary feature at 34.5 GHz and 3.8µm, respectively. F1 has an extent of 0.65(cid:48)(cid:48) × 0.25(cid:48)(cid:48) (length × width) with an integrated flux density of ∼ 2 mJy and background subtracted peak intensity of 416 mJy per (0.12(cid:48)(cid:48))2 at 34.5 GHz and 3.8µm, respectively. Figure 5c shows a large view of the region which includes X3, X7, the radio cometary feature (F1) and the MIR cometary feature (F2) at 4.68 µm in reverse color. We note a gap in the region to the NE of Sgr A* where F1 is detected. This gap appears to be devoid of dust emission at MIR. The MIR gap can also be identified at radio in Figure 1b. However, the lack of short uv spacings may contribute in suppressing the emission from the bright source Sgr A*. A close up view of infrared emission at M(cid:48) band from the inner 3(cid:48)(cid:48) ×3(cid:48)(cid:48) of Sgr A* is shown in the inset to the right of Figure 5c. An additional cometary feature is detected at MIR to the NE of IRS 16C in the inset. This source, which we call F2, is well within the MIR gap and lies along the direction where X3, X7 and F1 are located. Lastly, an additional radio source ∼ 13(cid:48)(cid:48) NE of Sgr A* shows a head-tail structure pointing toward Sgr A*. This radio feature shows a tail feature with an extent of ∼ 0.6(cid:48)(cid:48) at the PS∼ 57◦. The peak flux density of this source which we call F3 is 0.32 mJy beam−1 at 34.5 GHz with a spatial resolution of 89(cid:48)(cid:48)×46(cid:48)(cid:48) mas. This source has been detected in the L(cid:48) band as IRS 5 SE (Perger et al. 2008). This source is interpreted as a stellar bow-shock resulting from the interaction of a masss-losing stars and the minispiral. 3.2.2. Sgr A East Tower The nonthermal radio source Sgr A East is a young shell-type SNR with an angular size of 2.7(cid:48)×3.6(cid:48) and a spectral index α=0.76, where the flux density S ∝ ν−α (Ekers et al. 1983; ν Yusef-Zadeh & Morris 1987; Pedlar et al. 1989). Thermal X-ray emission is concentrated in the interior of the remnant suggesting that Sgr A East is a mixed morphology SNR interacting with the 50 km s−1 molecular cloud (Maeda et al. 2002; Park et al. 2005). The thermal X-ray emitting plasma has two components, characterized by temperatures of 1 and 6 keV and corresponding electron density of 4.7 and 0.6 cm−3, respectively. (Park et al. 2005; Koyama et al. 2007). A candidate neutron star CXOGC J174545.5–285829 (the cannonball) detected in X-ray and radio has also been associated with the remnant (Park et al. 2005; Zhao et al. 2013). Our broad band 1.5 GHz image of Sgr A East provides a wealth of details associated with Sgr A East and the surrounding environment. Figures 6a,b show large scale views of Sgr A East displayed with two different grayscale levels at 1.5 GHz. The major axis of the Sgr A East shell is along the Galactic plane. A number of new features are detected in this complex region. Here we describe three new features. One is a distorted region to the NE of the Sgr A East shell, 100(cid:48)(cid:48) E and 80(cid:48)(cid:48) N of Sgr A*. A striking tower-like structure with an extent of 100(cid:48)(cid:48) appears to emerge from a gap in the brightness distribution of Sgr A East. The base of this tower is about 20(cid:48)(cid:48)−25(cid:48)(cid:48) across with mean flux density of ∼ 2.8 mJy per 1.39(cid:48)(cid:48)×0.6(cid:48)(cid:48) – 10 – beam at 1.5 GHz. The base narrows as it extends to the NE with a position angle of ∼ 50◦. A schematic diagram of these features is shown in Figure 6c. The tower is terminated by two bow-shock-like structures. Grayscale contours of these are displayed in Figure 7a. The second feature is the polarized source P1 which was first identified at 8 GHz (Yusef- Zadeh et al. 2012). At 1.5 GHz, P1 is resolved into two linear structures (see the total intensity image in Figure 6b), that appears to cross each other at right angles. One of the linear features has a PA∼ 50◦, as shown in Figure 7b. Lastly, the region surrounding the polarized source P4 (Yusef-Zadeh et al. 2012) is shown at 1.5 GHz in Figure 7c. This elongated feature extents for 20(cid:48)(cid:48) at 1.5 GHz with a position angle of ∼ 50◦. A number of blob-like structures were previously reported to the NE of this feature (Yusef-Zadeh et al. 2012). 3.3. Other Features 3.3.1. A Semi-linear Feature We identify a striking semi-linear radio continuum feature projected perpendicular to the radioSgrA*ridge4. Figure8ashowsa5.5GHzimagethatrevealsroughlyuniformbrightness ∼ 0.5 mJy per 0.5(cid:48)(cid:48)×0.27(cid:48)(cid:48) beam. The semi-linear feature appears to arise from the ionized bar as it curves concave toward Sgr A*, crosses the ridge at an angular distance of ∼1(cid:48)(cid:48) west of Sgr A*, and continues to the north of Sgr A*. To highlight this structure, white dashed lines are drawn along this radio feature. The semi-linear continuous structure has an extent of 5(cid:48)(cid:48) and width of 0.5(cid:48)(cid:48), becoming wider and more diffuse as it extends to the north of Sgr A*. The Northern arm shows a discontinuity, as it approaches Sgr A*, which is best seen at α,δ(J2000) = 17h45m40.18s,−29◦0(cid:48)29(cid:48)(cid:48) in Figures 1a-c. The elongated features makes a 90◦ change to the south in its direction as it approaches Sgr A* (see the schematic diagram in Fig. 9). The kinematics of ionized gas show that the radial velocity of this elongated feature is close to zero km s−1 but changes by ∼200 km s−1 close to the location of the discontinuity (Zhao et al. 2009). 3.3.2. The Minicavity and the Bar The bar of ionized gas lies a few arcseconds to the south of Sgr A* where the Eastern and Northern arms of Sgr A West cross each other. High spatial resolution images of the ionized bar show a minicavity of ionized gas with high velocity dispersion (Lacy et al. 1991; Roberts et al. 1996; Zhao et al. 2009). Figure 8b shows a larger area of the Eastern arm and the minicavity at 34 GHz. The mini-cavity and the Eastern arm are distinct from the North arm. High resolution observations have shown that the minicavity extends further to the SE (Zhao et al.1991). The new images show that a gap separates the eastern and western halves 4The Sgr A* radio ridge should be distinguished from the mid-IR Sgr A* ridge discussed by Sch¨odel et al. (2011)