Upper Limit on Brackett-gamma Emission from the Immediate Accretion Flow onto the Galactic Black Hole
Anna Ciurlo, Mark R. Morris, Randall D. Campbell, Andrea M. Ghez, Tuan Do, Devin S. Chu
DDraft version February 26, 2021
Typeset using L A TEX preprint2 style in AASTeX62
Upper Limit on Brackett- γ Emission from the Immediate Accretion Flow onto the Galactic BlackHole
Anna Ciurlo, Mark R. Morris, Randall D. Campbell, Andrea M. Ghez, Tuan Do, andDevin S. Chu Department of Physics and Astronomy, University of California, Los Angeles, CA 90095, USA W. M. Keck Observatory, Waimea, HI 96743, US (Revised February 26, 2021)
ABSTRACTWe present the first observational constraint on the Br γ recombination line emissionassociated with the supermassive black hole at the center of our Galaxy, known asSgr A*. By combining 13 years of data with the Adaptive Optics fed integral fieldspectrograph OSIRIS at the W. M. Keck Observatory obtained as part of the GalacticCenter Orbits Initiative, we extract the near-infrared spectrum within ∼ γ flux. The aperture was set to match the sizeof the disk-like structure that was recently reported based on millimeter-wave ALMAobservations of the hydrogen recombination line, H30 α . Our stringent upper limit is atleast a factor of 80 (and up to a factor of 245) below what would be expected from theALMA measurements and strongly constrains possible interpretation of emission fromthis highly under-luminous supermassive black hole. Keywords:
Galactic Center — Near infrared astronomy — interstellar atomic gas —Supermassive black holes – Accretion INTRODUCTIONThe center of our Galaxy hosts a supermas-sive black hole (SMBH) whose position coin-cides with the radio and infrared source, Sgr A*(Ghez et al. 2000; Sch¨odel et al. 2002). Thanksto its proximity, this environment presents aunique opportunity to study in detail the ac-cretion mechanism onto a SMBH. The SMBHis not very active (10 − L Edd , Baganoff et al.2003) even though it was likely much more ac-tive at various times in the past few hundred
Corresponding author: Anna [email protected] years (Clavel et al. 2013; Terrier et al. 2018)and a recent unusual peak in its activity hasbeen reported in the infrared (Do et al. 2019a).The accretion rate of hot gas through the Bondiradius ( ∼ a r X i v : . [ a s t r o - ph . GA ] F e b Ciurlo et al. creted material comes from the winds of nearbyyoung, windy stars, predicts a disk phase in theaccretion flow. However, the theoretical forma-tion of a disk is not universally agreed upon.Ressler et al. (2020) also modeled the accre-tion flow supplied by stellar winds from nearbyWolf-Rayet stars but they find that a disk wouldnot form around Sgr A*. In their modeling,they argue that the magnetic field drives a polaroutflow that removes angular momentum, andthat the fast inflow and outflow of the gas, to-gether with inefficient cooling, prevent the ma-terial from circularizing.Emission from the accretion flow onto Sgr A*has recently been detected with ALMA obser-vations by Murchikova et al. (2019), who re-ports a very broad ( > − ) emissionfeature in the H30 α radio recombination linearising within ∼ α ( n = 31 →
30 transition), one would expect signif-icant emission in the lower-lying Br γ recombi-nation line (7 → µ m.In this paper, we search for the expected Br γ emission based on data gathered by GalacticCenter Orbits Initiative. In Section 2, we brieflydescribe the data used. Section 3, describes theanalysis procedure to extract the Sgr A* NIRspectrum. In Section 4 we report our flux limit.A preliminary version of our limit on Br γ wasused by Murchikova et al. (2019) to argue thatthe H30 α emission must be augmented by weakmaser action. In Section 5, we discuss our limitin comparison to the reported H30 α flux andto accretion flow theories. Our conclusions arereported in Section 6. DATASETThe data for this study were taken as part ofthe Galactic Center Orbits Initiative (GCOI; PIGhez, W. M. Keck Observatory 1995 - present). We used a subset of the deep observations takenbetween 2006 and 2018 with OSIRIS, the inte-gral field spectrograph fed by a laser guide staradaptive optics system at W. M. Keck Obser-vatory (Larkin et al. 2006; Wizinowich et al.2006). All included epochs consist of 900-secondexposures in Kn3 band (2.121–2.229 µ m) witha 35 mas platescale. In this setup, the averagespectral resolution is R=3800, and at the wave-length of Br γ this corresponds to 6 10 − µ m, or2.5 channels.The initial stages of data analysis (includingsky subtraction and flux calibration) have beendescribed in previous publications (Ghez et al.2008; Do et al. 2009, 2019b; Ciurlo et al. 2020).We begin with the flux- and astrometrically-calibrated, mosaiced data-cubes. In this work,we only examined a sub-set of GCOI raw datathat had already been analyzed with a proce-dure optimized for gas detection reported inCiurlo et al. (2020), which differs from that usedfor the measurements of stellar radial veloci-ties . Furthermore, we exclude epochs showingstrong OH line subtraction residuals (visible asextremely strong absorption lines), which leavesa total of 19 epochs that are summarized in Ta-ble 1. These measurements have an average spa-tial resolution of 73 mas, based on a Gaussianfit to S0-2 in the mosaicked data-cubes. Thissubset of data is optimized for extracting thespectrum of the region around Sgr A*. ANALYSISOur goal is to extract a high signal-to-noisespectrum of the immediate surroundings of While GCOI work on stellar spectra have been ex-tracted from individual dark-subtracted exposures, thespectra in this study are extracted from sky-subtractedmosaics (individual exposures are sky-subtracted beforebuilding the mosaic). Using the mosaics allows us thesky coverage necessary to measure the positions of ref-erence stars that we then use to put all observed fieldsin the same reference frame (for more details see Ap-pendix C and Ciurlo et al. 2020). imit on Infared Emission from the Galactic Black Hole Accretion Flow Figure 1. Br γ emission map in grayscale from the 2017 OSIRIS continuum-subtracted cube, obtainedby collapsing the cube between -1100 km/s and 1100 km/s (displayed in arbitrary units). Contours of thereported H30 α red and blue-shifted emissions (Murchikova et al. 2019) are overlaid (the full spectrum isshown in Figure 2). The blue contours are 1.2, 1.6, 2.0 and 2.4 mJy. The red contours are 0.9, 1.1, 1.3 and1.5 mJy. The white dashed circle represents the half-power beam for the H30 α observations. The whitesolid circle represents the average FWHM of the NIR observations. The position of Sgr A* is marked by theblack cross. The aperture used for this analysis is shown in orange. Sgr A*. In particular, we seek to establishwhether there is a counterpart to the disk-likeemission of the reported H30 α recombinationline. We therefore extract a spectrum withina circular aperture centered on the position ofSgr A* and of radius 0.23” (the same apertureused by Murchikova et al. 2019), correspond-ing to a diameter of about 30 typical spatialresolution elements. Figure 1 shows the chosenaperture along with contours of the H30 α emis-sion lines over a map of the Br γ emission in theregion. The gas has complex, multi-componentemission extended in space and velocity. Wecan identify one component to the West, asso-ciated with radio source Epsilon (Yusef-Zadehet al. 1990), two in the northeast and southeastcorners associated with bright Wolf-Rayet stars(IRS16 C and IRS16 SW) and one featuring analmost linear feature immediately adjacent tothe position of Sgr A* along the line of sight(Sch¨odel et al. 2011; Peißker et al. 2020). More details on this extended gas emission are re-ported in Appendix E.Our spectral extraction procedure is designedto reduce the effect of bright stars and extendedgas emission at the Galactic center. First, tominimize the contribution from bright stars, wecreate and apply an adaptive mask that tracksthe brightest stars and that transmits ∼
60% ofthe aperture area (see Appendix A). The ex-tracted spectra are reported in Appendix B.Second, we apply a continuum subtraction rou-tine to each data cube (see Appendix C) priorto combining the data from all epochs. Theindividual epochal spectra are rescaled by di-viding by their corresponding unmasked areawithin the aperture. The resulting spectra havea typical noise per spectral resolution elementof 0.050 mJy (see Table 1). We then combinethe individual spectra using a weighted aver-age where the weight corresponds to the inversesquare of the noise. The noise of the combined
Ciurlo et al.
Date a Frames FWHM Aperture Noisenumber [mas] fraction [mJy]2006-06-18 a a b
11 81 72% 0.0412009-05-05 b
12 70 56% 0.0462009-05-06 b
12 81 71% 0.0472010-05-05 b b
11 79 55% 0.0932011-07-10 b b c
11 73 67% 0.0622017-05-19 c c
12 77 69% 0.0672017-07-27 c
13 89 76% 0.0432017-08-14 c c c
14 91 55% 0.0502018-07-22 c
11 77 68% 0.0342018-07-31 c
11 73 66% 0.0382018-08-11 c a This is a subset the GCOI dataset (e.g. Do et al.2019b) as described in Section 2.
Table 1.
Table of observations obtained withOSIRIS in Kn3 band, 35 mas plate-scale. All in-tegration times are equal to 900 s. The reportedaperture is the fraction of the total 0.23” radiusaperture that is not masked by our noise cut. Thereported noise is per resolution element and it iscalculated on the spectrum extracted after apply-ing the mask but rescaled to the effective area ofour aperture (i. e. dividing by the fraction of un-masked pixels). References: a ) Ghez et al. (2008), b ) Boehle et al. (2016), c ) Do et al. (2019b). spectrum (0.015 mJy) is lower by a factor of ∼ RESULTSFigure 2 (top) shows our resulting combinedspectrum along with the individual spectra.Subsequent analysis only uses the combinedspectrum. Figure 2 (bottom) shows the compar- ison between the observed Br γ spectrum (notcorrected for extinction) and the H30 α spec-trum reported by Murchikova et al. (2019). Theonly Br γ emission that rises above the noise isassociated with the superimposed lower-velocitygas that is also observed at larger radii (seeFigure 1 and Appendix E). There is no qual-itative evidence for the presence of a broadBr γ emission line. We want to underline thatthere is no sign of any broad ( >
800 km/s) linein the whole wavelength range covered by theOSIRIS observations ( ∼ ∼ γ line, withoutmaking any assumptions about the shape or ex-tent of the line profile, is to set the limit at 5times the root-mean-square noise per resolutionelement. The noise per resolution element is cal-culated over ranges devoid of spectral features(see Appendix C). In this way, we find an upperlimit on the flux density of 0.074 mJy in a 0.23”radius aperture When corrected for the extinc-tion reported by Sch¨odel et al. (2010), this limitcorresponds to 0.71 mJy. This limit can be usedwith any hypothetical line profile to derive thetotal line flux expected for the accretion flow.To quantify the non-detection of any broad-band emission ( ∼ α line and cuts out the low veloci-ties at which the superimposed gas is emitting.Our notch filter extends from a Br γ velocity of-1100 to +1100 km/s excluding the low veloci-ties between -400 and +510 km/s (see Figure 2);this filter contains 15 resolution elements.Using the H30 α line profile reported byMurchikova et al. (2019), we proceed next toderive a limit on the Br γ line flux by determin-ing our sensitivity to the detection of a line hav- imit on Infared Emission from the Galactic Black Hole Accretion Flow −2000 −1000 0 1000 20000.00.51.01.52.0 f l u x den s i t y [ m Jy ] velocity [km/s] f l u x den s i t y [ m Jy ] velocity [km/s] −2000 −1000 0 1000 20000.00.51.01.52.02.53.0 velocity [km/s] f l u x den s i t y [ m Jy ] Figure 2.
T op : Continuum-subtracted spectra extracted over a 0.23” radius aperture centered on Sgr A*.The colored lines show the single-epoch spectra normalized over the fraction of aperture used (dates reportedin YYMMDD format). The thick black line represents the combined spectrum. There is no broad (several1000 km/s) line detected.
Bottom : Observed Br γ spectrum (black), compared to the observed H30 α spectrum from Murchikova et al. (2019) (gray). The flux scale applies to both the radio and NIR emissions.At our level of sensitivity, there is no Br γ counterpart to the broad, double-peaked H30 α emission. Thedash-dotted vertical lines represent the velocity range over which the H30 α line has been reported. Thedashed vertical lines shows the much narrower range over which superimposed gas emission is present. Theshaded area represents our notch filter. ing that profile in our spectrum. We do this bymodelling the published H30 α line profile as adouble-peaked Gaussian, and then planting theresult of that fit on top of our NIR spectrum.We only consider the portion of the resultingspectrum in our notch filter. A limit to the Br γ flux is then determined by identifying the totaladded flux for which the signal-to-noise ratio ofthe planted spectrum within the notched filteris greater than 5. The planted-Gaussian spec- trum corresponding to our 5- σ limit is shownin Figure 3. More details can be found in Ap-pendix D.Since we are assuming that the Br γ profilehas the same shape as that of the H30 α line, wedetermine that the notch filter would contain46% of the flux. Hence we find that the up-per limit to the total Br γ flux (calculated overthe whole H30 α range, -1100 to +1100 km/s)corresponding to a signal-to-noise ratio of 5 is Ciurlo et al. −2000 −1000 0 1000 2000−0.050.000.050.100.15 velocity [km/s] f l u x den s i t y [ m Jy ] Figure 3.
Vertically expanded version of the NIR spectrum of Figure 2 (black) together with the spectrumthat would correspond to a 5- σ detection of a line with the same shape as H30 α (only shown in our notchfilter, magenta). The shaded area represents the notch filter. ± − W/m . The de-reddenedvalue corresponds to 23.7 ± − W/m .We note that, if we don’t mask the brighteststars in our spectral extraction, the resulting5 σ flux limit is ∼
15% higher because of higherphoton noise. DISCUSSIONIn order to compare our flux limit to modelsof the accretion flow and to the reported H30 α line, it is necessary to calculate the correspond-ing emissivities. The total flux integrated overthe H30 α line (from -1100 km/s to +1100 km/s)is estimated to be 0.3 10 − W/m (correctedfor extinction). Our flux limit over the samevelocity range for Br γ is 23.7 ± − W/m (corrected for extinction). The correspondingemissivity (cid:15) can be calculated as: (cid:15) = F λ · πd V , (1) Note the large factor in the ratios of fluxes versusflux densities for a given velocity range, because of thelarge difference in the number of Hz per km/s betweenBr γ and H30 α . where F λ is the total flux of a given emissionline, d is the distance to the Galactic Cen-ter (7.97 kpc, Do et al. 2019b) and V is theemitting volume. We consider two geometriesfor the emitting region: 1) a disk of radius0.23” (same as the disk reported by Murchikovaet al. 2019) and height 0.03” (13% of the ra-dius), 2) a sphere of radius 0.23”. The diskvolume occupies roughly a tenth of the spherevolume. Using the disk geometries we obtainemissivities for Br γ and H30 α of ≤ − and 9.8 10 − erg/s/cm respectively. For asphere geometry we obtain ≤ − and 0.910 − erg/s/cm respectively.To investigate what ranges of temperature anddensity can give rise to these emissivities, we usethe tabulations of Hummer & Storey (1987).Figure 4 shows the Hummer & Storey (1987)emissivities as a function of density, where thedifferent colored curves correspond to differenttemperatures. We plot two sets of curves: onefor each of Br γ and H30 α . The emissivities wederive from the total flux of each line are com-patible with a range of densities when consider- imit on Infared Emission from the Galactic Black Hole Accretion Flow α emissivity derived fromobservations, we can project the compatibledensities onto the Br γ set of temperaturecurves, hence determining the expected emis-sivity in Br γ given the observed H30 α flux.The same process can be applied to the Br γ upper limit, which implies an upper limit onthe expected H30 α emission.We can compare, on the same plot, modelsof the accretion flow. For instance, Calder´onet al. (2020) recently reported simulations ofthe interactions of stellar winds from stars in theyoung nuclear cluster that produce a disk on thesame scale as the one reported by Murchikovaet al. (2019). According to this model, the tem-perature and density in the disk are respectively ∼ K and ∼ cm − .Therefore, we find that: 1) in the case ofan emitting disk, our upper limit on the Br γ flux is compatible with the model reported byCalder´on et al. (2020), 2) regardless of themodeled emission geometry, the reported H30 α emissivity measurement is much higher thanthat predicted by Calder´on et al. (2020) and byour Br γ limit. The width and flux reported byYusef-Zadeh et al. (2020) would lead to similarconclusions.Our measured Br γ flux limits, together withthe expected values from reported H30 α lineand the Calder´on et al. 2020 model, are summa-rized in Table 2. The Br γ flux expected fromthe H30 α observations is several orders of mag-nitude greater than our reported limit. Thisdiscrepancy between the reported H30 α and theupper limit on Br γ led Murchikova et al. (2019)to propose a maser effect to amplify the H30 α emission. However, such a population inversionis not expected from the calculations of Hum-mer & Storey (1987), as reported by Scoville& Murchikova (2013) and shown in Figure 4.Although Hummer & Storey (1987) probe only a limited range of temperatures and densitiesthat might not represent the physical conditionsappropriate to the Sgr A* accretion flow, wenote that, if the gas stays very hot, we shouldnot expect to observe any recombination line,since emissivity declines with temperature (therecombination rate is low at high temperaturesbecause protons are moving too fast).If the level populations of the H30 α lineare inverted, maser amplification can occur.Murchikova et al. (2019) have invoked thismechanism in order to reproduce the flux den-sities that they report, given our upper limitto the Br- γ line. However, the nearby pres-ence of the very bright (several Jansky) pointsource, Sgr A*, at the center of the putativedisk is problematical for the maser hypothesisbecause the preferred gain path for the maserwould always be radial, directed away from SgrA*. In the competition for stimulating newphoton emission from the upper state of theH30 α transition, the strong emission from SgrA* would be strongly favored over the relativelyweak (few milli-Jansky), isotropically emitted,spontaneous emission in the H30 α line fromthroughout the accretion disk. Consequently,if that transition were indeed inverted, the re-sulting maser-enhanced line radiation would al-ways be directed along a radial ray from Sgr A*.Then, rather than observing emission extendedalong the projected disk and showing a strongvelocity difference on opposite sides of the pro-jected disk, as was reported by Murchikovaet al. (2019), we should expect to see an un-resolved point source of line emission having avelocity width reflecting the line-of-sight veloc-ity gradient in the accretion flow. Furthermore,the reported 2000 km s − width of the line isfar larger than could be expected for the ra-dial component of gas moving in quasi-circularorbits in an accretion disk. We conclude fromthese arguments that the maser interpretationof the reported H30 α line emission is unlikely. Ciurlo et al. Br γ limitexpected H30 α limitCalderon+20 T=5000 KT=7500 KT=10000 KT=15000 KT=20000 KH30 α measurementexpected Br γ measurementCalderon+20 Br γ H30 α +5 +6 +7 −20 −18 −16 −14 n e [1/cm ] e m i ss i v i t y [ e r g / s / c m ] Figure 4.
Comparison between the Br γ limit, the H30 α measurement and the disk characteristics emergingfrom the model of Calder´on et al. 2020 assuming a disk geometry. This plot show emissivity as a functionof electron density for a range of temperatures (displayed as groups of lines of different colors) for both Br γ (solid lines, upper left group) and H30 α (dashed lines, lower right group) from Hummer & Storey (1987).Small crosses on the lines represent the points where the emissivity was actually calculated by Hummer &Storey (1987). Br γ and H30 α emissivities computed from the Calder´on et al. (2020) disk model densityare displayed as purple dots. The H30 α measurement by Murchikova et al. (2019) leads to an estimatedemissivity compatible with a range of densities and is shown as the a green line. The corresponding, predictedBr γ emissivity is reported as a green line as well. The same can be done with the Br γ flux limits reportedas a lines (with arrows to underline that it is purely an upper limit).Measured Br γ flux limit based on 5 × RMS noise [10 − W/m ] ≤ γ flux limit based on H30 α profile [10 − W/m ] ≤ ± γ flux from H30 α [10 − W/m ] 5900Expected Br γ flux from Calderon et al. model [10 − W/m ] 10 Table 2.
Summary table of Br γ flux limits measured in this work and expected fluxes based on thereported H30 α line (Murchikova et al. 2019) and Calder´on et al. 2020 model. Our measured flux limit fromnoise corresponds to 5 times the root-mean-square of the noise (multiplied by the expected width of theline, 2200 km/s) whereas the second limit is the one obtained by planting the H30 α profile in our spectrum.Both expected fluxes are calculated using a disk geometry as previously described. All values are correctedfor extinction. 6. CONCLUSIONSAnalyzing 13 years of data gathered withthe integral field spectrograph OSIRIS at theW.M. Keck Observatory, we extracted the near-infrared spectrum of the immediate vicinity ofSgr A*. Within a radius of 0.23” we place a 5-sigma upper limit Brackett- γ emission line (seeTable 2 for a summary). Predictions drawn from future modeling of the accretion flow willneed to be constrained by this result.Our limit on Br γ emission is unexpected giventhe reported millimeter-wavelength detection ofthe higher-level transition H30 α (Murchikovaet al. 2019) within the same region. Given ourlimit on Br γ , the reported excitation of H30 α would need to deviate dramatically from current imit on Infared Emission from the Galactic Black Hole Accretion Flow α emission, we would expect to observe it asan unresolved point source rather than a disk.Support for this work was provided by NSFAAG grant AST-1412615, Jim and Lori Keir,the W. M. Keck Observatory Keck VisitingScholar program, the Gordon and Betty MooreFoundation, the Heising-Simons Foundation, and Howard and Astrid Preston. A. M. G.acknowledges support from her Lauren B. Le-ichtman and Arthur E. Levine Endowed As-tronomy Chair. The W. M. Keck Observatoryis operated as a scientific partnership among theCalifornia Institute of Technology, the Univer-sity of California, and the National Aeronauticsand Space Administration. The Observatorywas made possible by the generous financialsupport of the W. M. Keck Foundation. Theauthors wish to recognize and acknowledge thevery significant cultural role and reverence thatthe summit of Maunakea has always had withinthe indigenous Hawaiian community. We aremost fortunate to have the opportunity to con-duct observations from this mountain.REFERENCES Baganoff, F. K., Maeda, Y., Morris, M., et al.2003, ApJ, 591, 891Boehle, A., Ghez, A. M., Sch¨odel, R., et al. 2016,ApJ, 830, 17Calder´on, D., Cuadra, J., Schartmann, M.,Burkert, A., & Russell, C. M. P. 2020, ApJL,888, L2Ciurlo, A., Campbell, R. D., Morris, M. R., et al.2020, Nature, 577, 337Clavel, M., Terrier, R., Goldwurm, A., et al. 2013,A&A, 558, A32Davies, R. I. 2007, MNRAS, 375, 1099Do, T., Ghez, A. M., Morris, M. R., et al. 2009,ApJ, 703, 1323Do, T., Witzel, G., Gautam, A. K., et al. 2019a,ApJL, 882, L27Do, T., Hees, A., Ghez, A., et al. 2019b, Science,365, 664Eckart, A., & Genzel, R. 1997, MNRAS, 284, 576Ghez, A. M., Klein, B. L., Morris, M., & Becklin,E. E. 1998, ApJ, 509, 678Ghez, A. M., Morris, M., Becklin, E. E., Tanner,A., & Kremenek, T. 2000, Nature, 407, 349Ghez, A. M., Salim, S., Weinberg, N. N., et al.2008, ApJ, 689, 1044Hummer, D. G., & Storey, P. J. 1987, MNRAS,224, 801 Larkin, J., Barczys, M., Krabbe, A., et al. 2006, inProc. SPIE, Vol. 6269, Society of Photo-OpticalInstrumentation Engineers (SPIE) ConferenceSeries, 62691AMurchikova, E. M., Phinney, E. S., Pancoast, A.,& Blandford, R. D. 2019, Nature, 570, 83Peißker, F., Eckart, A., Sabha, N. B., Zajaˇcek, M.,& Bhat, H. 2020, ApJ, 897, 28Ressler, S. M., Quataert, E., & Stone, J. M. 2020,MNRAS, 492, 3272Sch¨odel, R., Morris, M. R., Muzic, K., et al. 2011,A&A, 532, A83Sch¨odel, R., Najarro, F., Muzic, K., & Eckart, A.2010, A&A, 511, A18Sch¨odel, R., Ott, T., Genzel, R., et al. 2002,Nature, 419, 694Scoville, N., & Murchikova, L. 2013, ApJ, 779, 75Terrier, R., Clavel, M., Soldi, S., et al. 2018, A&A,612, A102Witzel, G., Eckart, A., Bremer, M., et al. 2012,ApJS, 203, 18Wizinowich, P. L., Le Mignant, D., Bouchez,A. H., et al. 2006, PASP, 118, 297Yuan, F., & Narayan, R. 2014, ARA&A, 52, 529Yusef-Zadeh, F., Morris, M., & Ekers, R. D. 1990,Nature, 348, 45Yusef-Zadeh, F., Royster, M., Wardle, M., et al.2020, MNRAS, 499, 3909 Ciurlo et al. f l u x den s i t y [ m Jy ] wavelegth [ µ m] Figure 5.
Spectra extracted over a 0.23” radius aperture centered on Sgr A* for each epoch (beforecontinuum-subtraction). Each spectrum has been shifted by an arbitrary constant in order to separate andorder them from most recent (top) to oldest (bottom) epoch (dates reported in YYMMDD format). Thespectra with the original continuum levels are displayed in Figure 6. The central emission line is associatedwith gas superimposed along the line of sight.
APPENDIX A. STAR MASKFor each epoch we mask a region around the brightest stars by applying a continuum cut. Specif-ically, if the total flux in the wavelength range between 2.1343 and 2.1447 µ m, a region outside ournotch filter and devoid of emission features, is higher than a threshold (1.5 mJy) the pixel is masked.This continuum cut roughly corresponds to masking the brightest stars (K-magnitude < γ absorption in the stellarphotospheres of luminous young stars. B. SPECTRAL EXTRACTIONFigure 5 shows the spectra extracted over the masked aperture for each epoch. The spectralcharacteristics vary from epoch to epoch, showing different slopes. This is due to the fact that we areconsidering a very small aperture near Sgr A* through which many stars orbit on short timescales; imit on Infared Emission from the Galactic Black Hole Accretion Flow γ absorption which could potentially influence our measurements, although most of the S-stars haveradial velocities within ±
500 km/s, and therefore are not in the range we are considering for ourlimit on the accretion flow. Furthermore, for such stars, which have magnitudes above 16, the fluxdensity in the Br γ line is expected to be well below the noise of our spectra. C. CONTINUUM SUBTRACTION AND NOISE ESTIMATIONIn order to isolate the Br γ line emission, we removed the stellar continuum from each epoch. We dothat by selecting several wavelength ranges devoid of emission features: from 2.1343 to 2.1447 µ m,from 2.1469 to 2.1503 µ m, from 2.1544 to 2.1568 µ m, from 2.1752 to 2.1783 µ m and from 2.1808 to2.1948 µ m (displayed in Figure 6). We then estimate the continuum slope with a linear fit in thoseranges, using the same procedure as in Ciurlo et al. 2020). In the data reduction each individualexposure is sky-subtracted with a procedure (Davies 2007) that scales OH lines to be consistent withthe science exposures. After that we build the mosaics. There could still be small OH lines residualsafter this procedure. Because this imperfect sky subtraction could lead to poor continuum estimation,we therefore avoided spectral channels containing bright telluric OH lines within the selected ranges.The ranges used for the continuum estimation and the continuum fit are shown in Figure 6. Thecontinuum estimation is less reliable near the edges of the Kn3 filter, but that does not affect ouranalysis, as the broad line we are seeking is close to the center of the spectral band.To calculate the noise level σ we calculate the root-mean-square flux density over ranges devoidof spectral features (the same used for the continuum estimation, see Figure 6). According to theOSIRIS manual, the average resolution with the 35 mas platescale for the Kn3 band corresponds to ∼ D. PLANTING H30 α PROFILE INTO THE NIR SPECTRUMWe model the H30 α profile with a double-peaked Gaussian (Figure 7). To account for the errorin fitting the H30 α profile with Gaussians, we draw randomly the width and central wavelengthwithin the uncertainties of the fitted values 1000 times. We vary the flux between 1/100 and 10times the flux limit estimated from the channel-to-channel noise. We calculate the signal-to-noise asthe total flux in our notch filter, divided by the square-root of the number of independent resolutionelements times the noise per resolution element. We identify the total fluxes that fall in the intervalof (5 ± σ and take the median as our final total flux. This way we obtain a limit on the Br γ fluxof 1.13 10 − W/m in our notch filter. We estimate the uncertainty as the root-mean square of thevalues one obtains by running the procedure 19 times, dropping at each time one of the 19 epochs ofdata. E. THE EXTENDED GAS EMISSIONThe OSIRIS data show that there are multiple extended foreground and background gas componentsto the Br γ emission in this region (evident in Figure 5, near the Br γ rest frequency). The gas2 Ciurlo et al. −5000 0 50004.04.55.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.54.04.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.54.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.53.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.53.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.53.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.53.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.22.42.62.8 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.54.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.53.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50002.42.62.83.03.2 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.54.0 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] −5000 0 50003.03.5 velocity [km/s] F l u x den s i t y [ m Jy ] Figure 6.
Individual spectra and continuum fits. For each epoch (dates reported in YYMMDD format)the spectrum is extracted over a 0.23 ” aperture around Sgr A*. The colors of the individual spectra are thesame as in previous figures. We consider the median (blue dots) for each of several ranges devoid of spectralfeatures (red) to estimate a linear continuum (blue line). We avoid spectral channels containing strong OHlines (lavender). The dash-dotted vertical lines show the width of the reported H30 α line. The shaded arearepresents our notch filter. observed toward SgrA * has complex, multi-component emission extended in space and velocity. Wecan identify at least three components (see Figure 8 top and middle): one associated with nearbyradio source Epsilon (Yusef-Zadeh et al. 1990), one associated with the bright Wolf-Rayet stars inthe field (IRS16 C and IRS16 SW) and one featuring an almost linear feature immediately adjacentto the position of Sgr A* along the line of sight (Sch¨odel et al. 2011; Peißker et al. 2020). Some imit on Infared Emission from the Galactic Black Hole Accretion Flow −2000 −1000 0 1000 20000.00.51.01.52.02.53.0 f l u x den s i t y [ m Jy ] velocity [km/s] Figure 7.
Observed H30 α spectrum from Murchikova et al. (2019) modeled with a double-peaked Gaussian(red). of these components move enough from year to year to require a specific fit for each epoch if theyare to be subtracted from the spectrum. However, given that our goal is to measure or limit verybroad Br γ emission ( ∼ < γ emission. We note that by changing the bright star maskfrom epoch to epoch we sample different portions of the confusing background in different epochs.However, the wavelength extent of the gas does not exceed the boundary of the exclusion zone weset and therefore it does not influence our analysis. Figure 8 (bottom) shows how, at larger radii,this superimposed gas emission becomes more and more challenging to deal with. Therefore, for thiswork we only focus on a 0.23” aperture that allows us compare our limits directly to the Murchikovaet al. (2019) observations.4 Ciurlo et al. −2000 −1000 0 1000 20000.0000.0010.0020.0030.004 velocity [km/s] I n t en s i t y [ m Jy ] average spectrum Sgr A*−1 0 1−1.0−0.50.00.51.0 0.010.1
RA [arcsec] D e c [ a r cs e c ] [mJy] blue Sgr A*−1 0 1−1.0−0.50.00.51.0 0.0250.13
RA [arcsec] D e c [ a r cs e c ] [mJy] red Sgr A*−1 0 1−1.0−0.50.00.51.0 0.0010.1
RA [arcsec] D e c [ a r cs e c ] [mJy] green r=0.1’’r=0.23’’r=0.5’’r=1’’ −2000 −1000 0 1000 2000 0 2 4 6 8 combined spectra no r m a li z ed f l u x den s i t y [ m Jy ] velocity [km/s]velocity [km/s]velocity [km/s]velocity [km/s] Sgr A*−1 0 1−1.0−0.50.00.51.0 0.020.25