------------------------------------------------------------------------ ms.tex ApJL, 2009, in press Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-MailScanner-Information: Please contact the postmaster@aoc.nrao.edu for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-4, required 5, autolearn=disabled, RCVD_IN_DNSWL_MED -4.00) X-MailScanner-From: deokkeun@ipac.caltech.edu X-Spam-Status: No %astro-ph/0907.4752 %\documentclass[12pt,preprint]{aastex} \documentclass[apj]{emulateapj} \bibliographystyle{apj} \usepackage{lscape} \usepackage{apjfonts} \usepackage{graphicx} \slugcomment{Revised after referee report} \shorttitle{Massive YSOs in the Galactic Center} \shortauthors{An et~al.} \begin{document} \title{First Spectroscopic Identification of Massive Young Stellar Objects\\ in the Galactic Center} \author{Deokkeun An\altaffilmark{1}, Kris Sellgren\altaffilmark{2}, A.\ C.\ Adwin Boogert\altaffilmark{1},\\ Susan R.\ Stolovy\altaffilmark{7}, Thomas P.\ Robitaille\altaffilmark{9,10}, and Howard A.\ Smith\altaffilmark{9} California Institute of Technology, Mail Stop 100-22, Pasadena, CA 91125; 140 West 18th Avenue, Columbus, OH 43210; sellgren@astronomy.ohio-state.edu.} \altaffiltext{3}{CRESST/UMBC/GSFC, Code 665, NASA/Goddard Space Flight Center, 8800 Greenbelt Road, Greenbelt, MD 20771.} \altaffiltext{4}{Science Systems and Applications, Inc.} \altaffiltext{5}{Observatoire de Besan\c{c}on, 41bis, avenue de l'Observatoire, F-25000 Besan\c{c}on, France.} \altaffiltext{6}{Institut d'Astrophysique de Paris, CNRS, 98bis Bd Arago, F-75014 Paris, France.} \altaffiltext{7}{Spitzer Science Center, California Institute of Technology, Mail Code 220-6, 1200 East California Boulevard, Pasadena, CA 91125.} \altaffiltext{8}{SETI Institute, 515 North Whisman Road, Mountain View, CA 94043.} \altaffiltext{9}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138.} \altaffiltext{10}{Spitzer Postdoctoral Fellow.} \begin{abstract} We report the detection of several molecular gas-phase and ice absorption features in three photometrically-selected young stellar object (YSO) candidates in the central 280~pc of the Milky Way. Our spectra, obtained with the Infrared Spectrograph (IRS) onboard the {\it Spitzer Space Telescope}, reveal gas-phase absorption from CO$_2$ ($15.0\mu$m), C$_2$H$_2$ ($13.7\mu$m) and HCN ($14.0\mu$m). We attribute this absorption to warm, dense gas in massive YSOs. We also detect strong and broad $15\mu$m CO$_2$ ice absorption features, with a remarkable double-peaked structure. The prominent long-wavelength peak is due to CH$_3$OH-rich ice grains, and is similar to those found in other known massive YSOs. Our IRS observations demonstrate the youth of these objects, and provide the first spectroscopic identification of massive YSOs in the Galactic Center. \end{abstract} \keywords{infrared: ISM --- ISM: molecules --- stars: formation} \section{Introduction} The Central Molecular Zone (CMZ) is the innermost $\sim200$~pc region of the Milky Way Galaxy. It is a giant molecular cloud complex delineated by a gradient in the CO column density and temperature. The CMZ contains $\sim10\%$ of the Galaxy's molecular gas, and produces $5\%$--$10\%$ of its infrared and Lyman continuum luminosities \citep[see a review by][and references therein]{morris:96}. Evidence is mounting that conditions for star formation in the CMZ are significantly different from those in the Galactic disk. The gas pressure and temperature are higher in the CMZ than in the average disk, conditions that favor a larger Jeans mass for star formation and an initial mass function biased towards more massive stars. Furthermore, the presence of strong magnetic fields, tidal shear, and turbulence challenges the standard paradigm of slow gravitational collapse of molecular cloud cores. The CMZ provides several signposts of {\it in situ} star formation, such as H$_2$O masers, (ultra-)compact \ion{H}{2} regions, young OB stars, and young supernova remnants. However, young stellar objects (YSOs or protostars), which are the direct tracers of current star formation, have so far eluded detection in the CMZ. They have been inferred to be present based on infrared photometry \citep[e.g.,][]{felli:02, schuller:06,yusefzadeh:09}, but spectroscopic observations are required to confirm their status as a YSO. This is because evolved stars can look like YSOs in broad-band photometry, if they are heavily dust attenuated \citep[e.g.,][]{schultheis:03}, a problem towards the Galactic Center (GC), where $A_V \approx 30$. In this {\it Letter}, we present spectroscopic follow-up observations of YSO candidates in the CMZ, using the Infrared Spectrograph \citep[IRS;][]{houck:04} onboard the {\it Spitzer Space Telescope} \citep{werner:04}. Massive YSO candidates were photometrically selected from the point source catalog \citep{ramirez:08}, which was extracted from images of the CMZ \citep{stolovy:06} made using the Infrared Array Camera \citep[IRAC;][]{fazio:04}. This high sensitivity and high spatial resolution image has led to a better identification of YSO candidates and their follow-up spectroscopic observations. \section{Photometric Sample Selection} The IRAC point source catalog \citep{ramirez:08} contains photometry for more than a million point sources in the entire CMZ ($2\arcdeg \times 1.4\arcdeg$ or $280 \times 200$~pc) in four channels ($3.6\mu$m, $4.5\mu$m, $5.8\mu$m, and $8.0\mu$m). Initially, we selected point sources with ${\rm [3.6] - [8.0]} \geq 2.0$, corresponding to YSOs with $M_* \ga 2.5 M_\odot$ \citep{whitney:03,whitney:04}. We further confined the sample to those within $|b| < 15\arcmin$, resulting in $1207$ objects. When we had photometric measurements in at least 5 bandpasses from IRAC, 2MASS \citep[$JHK_s$;][] {skrutskie:06}, and/or ISOGAL \citep[$7\mu$m and $15\mu$m;][]{omont:03}, we selected YSO candidates by comparing the observed spectral energy distribution (SED) with YSO models \citep{robitaille:06} using a SED fitting tool by \citet{robitaille:07}. Otherwise, we applied additional color constraints from \citet[][${\rm [3.6]-[4.5]} \geq 0.5$, ${\rm [4.5]-[5.8]} \geq 0.5$, and ${\rm [5.8]-[8.0]} \geq 1.0$]{whitney:04} to identify YSO candidates. SED fitting and color selection narrowed down our sample to about $200$ objects. Then, we carefully inspected IRAC three-color images to select objects that are distinct within the IRS slit entrances against the crowded stellar field and bright local background. Finally, a literature search was carried out for the selected objects, and one Wolf-Rayet star and four OH/IR stars were discarded. Our final sample is composed of 107 objects, among which 25 were previously known YSO candidates from ISOGAL \citep{felli:02}. \section{IRS Observations and Data Reduction} \begin{figure*} %\epsscale{1.1} \epsscale{1.0} \plotone{f1.eps} \caption{Composite IRS spectrum of SSTGC~797384. The spectrum is from SL at $\lambda \le 11.2\mu$m, SH at $11.2\mu$m $\leq \lambda \leq 19.3\mu$m, and LL at $\lambda \geq 19.3\mu$m. This composite spectrum is characterized by an extremely red spectral energy distribution, strong and deep silicate absorption, and several molecular gas- and solid-phase absorptions. \label{fig:all}} \end{figure*} \begin{deluxetable*}{lcccc} \tablewidth{0pt} \tablecaption{Properties of the Sample\label{tab:tab1}} \tablehead{ \colhead{Quantities} & \colhead{Units} & \colhead{SSTGC~524665} & \colhead{SSTGC~797384} & \colhead{SSTGC~803187} } \startdata R.A.(J2000.0) & h:m:s & 17:45:39.86 & 17:47:23.68 & 17:47:26.29 \nl Decl.(J2000.0) & d:m:s & -29:23:23.4 & -28:23:34.6 & -28:22:1.5 \nl UKIDSS $J$\tablenotemark{a} & mag & \nodata & $18.23\pm0.06$ & $17.39\pm0.03$ \nl UKIDSS $H$\tablenotemark{a} & mag & \nodata & $14.68\pm0.01$ & $16.60\pm0.05$ \nl UKIDSS $K$\tablenotemark{a} & mag & $15.71\pm0.10$ & $12.92\pm0.01$ & $14.37\pm0.02$ \nl IRAC ${\rm[3.6]}$\tablenotemark{b} & mag & $11.42\pm0.01$ & \nodata & $12.22\pm0.02$ \nl IRAC ${\rm[4.5]}$\tablenotemark{b} & mag & $ 8.63\pm0.01$ & $9.41\pm0.01$ & $ 8.97\pm0.01$ \nl IRAC ${\rm[5.8]}$\tablenotemark{b} & mag & $ 7.08\pm0.01$ & $7.66\pm0.01$ & $ 7.24\pm0.01$ \nl IRAC ${\rm[8.0]}$\tablenotemark{b} & mag & $ 6.13\pm0.01$ & $5.64\pm0.01$ & $ 5.11\pm0.01$ \nl MIPS ${\rm[24]}$\tablenotemark{c} & mag & $ 1.54\pm0.01$ & $0.55\pm0.01$ & \nodata \nl ${\rm T_{\rm ex}~(C_2H_2~gas)}$ & K & $300\pm150$ & $200\pm150$ & $300\pm150$ \nl ${\rm T_{\rm ex}~(HCN~gas)}$ & K & $400\pm50$ & $100\pm50$ & $100\pm50$ \nl ${\rm T_{\rm ex}~(CO_2~gas)}$ & K & $200\pm50$ & $100\pm50$ & $100\pm50$ \nl ${\rm N_{col}~(C_2H_2~gas)}$ & $10^{16}$cm$^{-2}$ & $7.9\pm3.3$ & $1.0\pm0.3$ & $2.0\pm0.8$ \nl ${\rm N_{col}~(HCN~gas)}$ & $10^{16}$cm$^{-2}$ & $15.8\pm4.6$ & $1.0\pm0.4$ & $2.0\pm0.9$ \nl ${\rm N_{col}~(CO_2~gas)}$ & $10^{16}$cm$^{-2}$ & $20.0\pm5.6$ & $5.0\pm1.5$ & $7.9\pm2.8$ \nl ${\rm N_{col}~(CO_2~solid)}$ & $10^{19}$cm$^{-2}$ & $0.11\pm0.01$ & $0.13\pm0.01$ & $0.21\pm0.01$ \nl ${\rm N_{col}~(H_2O~solid, 6\mu m)}$ & $10^{19}$cm$^{-2}$ & $<1.7$ & $<2.3$ & $<4.7$ \nl ${\rm N_{col}~(H_2O~solid, 13\mu m)}$& $10^{19}$cm$^{-2}$ & $0.6\pm0.4$ & $1.3\pm0.4$ & $1.9\pm0.3$ \nl ${\rm N_{col}~(H_2)}$ & $10^{22}$cm$^{-2}$ & $2.3\pm0.2$ & $5.0\pm0.5$ & $5.8\pm0.7$ \nl ${\rm N_{C_2H_2} / N_{H_2}}$ & $10^{-7}$ & $34.3\pm14.7$ & $2.0\pm0.6$ & $3.4\pm1.4$ \nl ${\rm N_{HCN} / N_{H_2}}$ & $10^{-7}$ & $68.7\pm20.9$ & $2.0\pm0.8$ & $3.4\pm1.6$ \nl ${\rm N_{CO_2,gas} / N_{H_2}}$ & $10^{-7}$ & $87.0\pm25.5$ & $10.0\pm3.2$ & $13.6\pm5.1$ \nl ${\rm N_{CO_2,gas} / N_{CO_2,solid}}$& & $0.18\pm0.05$ & $0.04\pm0.01$ & $0.04\pm0.01$ \nl ${\rm N_{CO_2,solid} / N_{H_2O,solid}}$& & $0.18\pm0.12$ & $0.10\pm0.03$ & $0.11\pm0.02$ \nl $A_V~(\tau_{9.7\mu m})$ & mag & $27\pm3$ & $53\pm5$ & $62\pm7$ \nl $A_V$~(color)\tablenotemark{d} & mag & $29.0\pm2.6$ & $32.0\pm2.6$ & $27.0\pm5.4$ \enddata \tablecomments{Random errors are shown in the photometry.} \tablenotetext{a}{Aperture3 magnitudes from UKIDSS DR2 \citep{warren:07}.} \tablenotetext{b}{Photometry from \citet{ramirez:08}.} \tablenotetext{c}{Photometry from MIPSGAL (S.\ Carey, 2008, private communication).} \tablenotetext{d}{Based on the 2MASS and IRAC color-magnitude diagrams of GC red giant branch stars within $2\arcmin$ of the source \citep{schultheis:09}.} \end{deluxetable*} We obtained spectroscopic data for 107 YSO candidates using the four IRS modules in May and October 2008. We observed each target in IRS staring mode with 4 exposures per source (2 cycles). Exposure times were 6~sec--120~sec in SH (short-high; short wavelength, high resolution), 6~sec--60~sec in LH (long-high), 6~sec--14~sec in SL (short-low), and 6~sec in LL (long low) modules, depending on the source's brightness, to achieve a signal-to-noise ratio (S/N) of at least $50$ in SH and SL, and a minimum S/N of $10$ in LH and LL. We reduced the IRS spectra from the basic calibrated data (BCD) products version S17.2.0 and S18.1.0, using the SSC software packages {\tt IRSCLEAN} (to correct for bad pixel values) and {\tt SPICE} (to extract spectra). Because the GC exhibits strong, spatially variable background, we observed multiple off-source measurements (one cycle, $1\times1$ mapping mode) to derive backgrounds near each of our YSO candidates in the four IRS modules. The on-source and the off-source observations were taken consecutively to minimize zodiacal light and instrumental variations. For the high resolution observations, we observed and extracted four background positions ($\sim \pm1\arcmin$ offsets in either R.A.\ or Decl.). For the low resolution observations, we took spectra from two background positions at $\sim \pm1\arcmin$ away in the direction perpendicular to the slit, and extracted two additional background spectra at positions along the on-source slit. In all of the four different IRS modules, we tried to extract the background spectra at the same position as much as possible, to minimize the flux difference from different modules. We made an interpolation of a plane in three dimensional space (positions on the IRAC map and wavelength) to obtain a background spectrum at the source position. We estimated an error in each source's background from the dispersion of four different background spectra, constructed from alternate sets of three out of the four background pointings. A complete analysis of spectra for all of our 107 YSO candidates will be presented elsewhere (D.\ An et al. 2009, in preparation). For the current analysis, we selected three targets (Table~\ref{tab:tab1}) from among those showing characteristic spectral features of massive YSOs, which include gaseous molecular absorptions from C$_2$H$_2$, HCN, CO$_2$ \citep[e.g.,][]{lahuis:00,boonman:03,knez:09}, and a solid-phase absorption from CO$_2$ ice bending mode \citep[e.g.,][]{gerakines:99}. Both SSTGC~797384 and SSTGC~803187 are associated with a relatively weak radio continuum source \citep[SGR~B2(P) and SGR~B2(R), respectively;][]{mehringer:93}. They are on the outskirts of the Sgr~B2 molecular cloud ($\sim2$~pc--$4$~pc from the well-studied radio source SGR~B2(M)), which is one of the most active complexes of compact \ion{H}{2} regions in the Galaxy \citep[e.g.,][]{mehringer:95}. \citet{mehringer:93} derived zero-age main-sequence spectral types of B0 and O6.5 for these compact \ion{H}{2} regions, respectively, from the number of ionizing photons. SSTGC~524665 does not have radio continuum emission associated with it. However, it is coincident with an H$_2$O maser \citep{forster:89}, and is adjacent to a region of $4.5\mu$m excess emission \citep{yusefzadeh:09}, possibly tracing shocked molecular outflows \citep[e.g.,][]{smith:06}. For SSTGC~803187, we used a non-standard extraction aperture in SL, because of a nearby source ($\approx7\arcsec$ south of the target) along the slit. We followed the prescription on the IRS data reduction website\footnote{See http://ssc.spitzer.caltech.edu/IRS/calib.} to calibrate the flux. We trimmed the end of the orders to remove the noisy part of spectra, and spectra from different orders in high-resolution modules were averaged using a linear ramp. After background subtraction, the SH and LH spectra were scaled down in flux to LL over the common wavelength interval for SSTGC~797384 and SSTGC~803187. The SL spectra were then scaled to SH. For these sources, we assumed that the flux mismatch is due to narrower slit entrances in SH and SL. For SSTGC~524665, we used the SL as a basis for the scaling, because our observations in LL and LH were contaminated by extended emission from a nearby ($\approx10\arcsec$ southwest of the target) bright source on the $24\mu$m MIPS image \citep{carey:09,yusefzadeh:09}. The background for this target is likely to be over-subtracted, because the target lies on a dark cloud with high extinction, while background spectra were taken at brighter spots. The potential problem of the background subtraction results in H$_2$ lines (arising from the surrounding sky) appearing in absorption in SSTGC~524665. In the following initial analysis, we did not use LH data for all targets, but focused on the spectral features in other modules. \section{Analysis and Results} Figure~\ref{fig:all} displays background-subtracted spectra of SSTGC~797384, in SL ($\lambda \leq 11.2\mu$m), SH ($11.2\mu$m $\leq \lambda \leq 19.3\mu$m), and LL ($\lambda \geq 19.3\mu$m). The observed spectrum is characterized by an extremely red SED [$\alpha \equiv d\log(\lambda F_\lambda)/d\log(\lambda) \approx 2$], strong and deep silicate absorptions at $9.7\mu$m and $18\mu$m, ice absorption features at $6\mu$m, $6.85\mu$m, $13\mu$m, and $15.2\mu$m. Although the presence of forbidden lines indicates that these objects are likely associated with an (ultra-)compact \ion{H}{2} region, it could be also due to under-subtracted emissions from the background. \begin{figure} \epsscale{1.15} %\epsscale{0.65} \plotone{f2.eps} \caption{Gas-phase molecular absorptions from C$_2$H$_2$ $\nu_5 = 1-0$ ($13.71\mu$m), HCN $\nu_2 = 1-0$ ($14.05\mu$m), and CO$_2$ $\nu_2 = 1-0$ ($14.97\mu$m). Best-fitting models are shown in solid lines. \label{fig:gas}} \end{figure} \begin{figure} \epsscale{1.15} %\epsscale{0.65} \plotone{f3.eps} \caption{Optical depth spectra of solid-phase absorption from the CO$_2$ ice bending mode. Best-fitting CO$_2$ ice models and individual CO$_2$ ice components are shown for each target: polar (dotted line, centered at $\sim15.3\mu$m), apolar (dotted line, centered at $\sim15.1\mu$m), pure (blue shaded), diluted (black solid line), $15.4\mu$m shoulder (orange-shaded), and the sum of these absorption components (green line). The bottom panel shows a comparison of the ice absorption profile between our sources (grey) and massive YSO W33A (blue). The optical depths for our targets were scaled in the bottom panel for comparison. \label{fig:co2}} \end{figure} Figure~\ref{fig:gas} shows gas-phase molecular absorptions at $13.71\mu$m (C$_2$H$_2$ $\nu_5 = 1-0$), $14.05\mu$m (HCN $\nu_2 = 1-0$), and $14.97\mu$m (CO$_2$ $\nu_2 = 1-0$), detected in three YSO candidates. To derive the excitation temperature ($T_{\rm ex}$) and column density ($N_{\rm col}$) for each molecular species, we used model spectra from \citet{spectrafactory} based on {\tt HITRAN04} linelist \citep{hitran} for C$_2$H$_2$ and HCN, and those based on {\tt HITEMP} \citep{hitemp} for CO$_2$. A second order polynomial was used to set a local continuum at $13.30\mu$m $\leq \lambda \leq 14.55\mu$m for C$_2$H$_2$ and HCN, and $14.77\mu$m $\leq \lambda \leq 15.06\mu$m for CO$_2$. We did not include isotopes in the computation because of the limited parameter span in the model grids. However, even a relatively high fraction of isotopes in GC \citep[$^{12}{\rm C}/^{13}{\rm C}\approx23$;] []{wannier:80} has a negligible impact in the model fitting. We first made a fit to C$_2$H$_2$, and subtracted its contribution to the absorption near weaker HCN bands. Best-fitting model $T_{\rm ex}$ and $N_{\rm col}$ were found by searching the minimum $\chi^2$ of the fits over 100~K $\leq T_{\rm ex} \leq$ 1000~K in steps of $\Delta T_{\rm ex} = 100$~K, and $15 \leq \log{N_{\rm col}} \leq 18$ for C$_2$H$_2$, $16 \leq \log{N_{\rm col}} \leq 18$ for HCN, and $16 \leq \log{N_{\rm col}} \leq 22$ for CO$_2$ with intervals of $0.1$~dex. Solid lines in Figure~\ref{fig:gas} show our best-fitting models, and their $T_{\rm ex}$ and $N_{\rm col}$ are listed in Table~\ref{tab:tab1}. Errors in these parameters were estimated from $\Delta \chi^2$, where $1\sigma$ measurement errors were taken from the scatter of flux in the spectra. Systematic errors from background subtraction and nodding differences were then added in quadrature. We tested with varying covering factors, but found that best-fitting case yields its value equal to or close to unity. These gaseous bandheads have been detected in absorption toward YSOs, tracing the warm and dense gas in the circumstellar disk and/or envelopes \citep[e.g.,][]{lahuis:00,boonman:03,knez:09}. They are sometimes detected in the photosphere and/or the circumstellar envelope of carbon-rich asymptotic giant branch stars \citep[e.g.][]{aoki:99}, but carbon stars have not been found in the GC region \citep[e.g.,][]{guglielmo:98}. The above estimates are based on models with a Doppler parameter $b = 3\ {\rm km\ s^{-1}}$. The line width measurements of these molecules for several massive YSOs and that of the strongest H$_2$CO absorption components near SSTGC~803187 are in the range of $b = 1-7\ {\rm km\ s^{-1}}$ \citep[e.g.,][]{mehringer:95,vandertak:00,knez:09}. There are limited model grids at $b = 10\ {\rm km\ s^{-1}}$ for C$_2$H$_2$ and HCN, but $T_{\rm ex}$ and $N_{\rm col}$ were generally found within $2\sigma$ from those at $b = 3\ {\rm km\ s^{-1}}$. \begin{figure} \epsscale{1.15} %\epsscale{0.65} \plotone{f4.eps} \caption{Fit to the H$_2$O ice and silicate absorption for SSTGC~797384. {\it Top:} SL and LL data (grey), with a best-fitting pseudo-continuum (black line). {\it Bottom:} Decomposition of optical depth spectra (grey) with the silicate (red) and the laboratory H$_2$O ice profiles (blue). Black line represents a sum of these two components. \label{fig:continuum}} \end{figure} Figure~\ref{fig:co2} shows optical depth spectra of our sources (grey) at $\sim15.2\mu$m, where the strong and wide CO$_2$ ice absorption is seen. We set a local continuum over $14.5\mu$m $\leq \lambda \leq 16.5\mu$m using a 3rd order polynomial, and followed the prescription in \citet{pontoppidan:08} to decompose the absorption profile with five laboratory spectral components: polar (CO$_2$:H$_2$O $= 14:100$ at 10~K; dotted line, centered at $\sim15.3\mu$m), apolar (CO:CO$_2 = 100:70$ at 10~K; dotted line, centered at $\sim15.1\mu$m), pure CO$_2$ (15~K; blue shaded), diluted CO$_2$ (CO:CO$_2 = 100:4$ at 10~K; black solid line), and $15.4\mu$m shoulder CO$_2$ ice profile (modeled with two Gaussians in wavenumber space; orange shaded). We found a best-fitting set of models from the non-linear least squares fitting routine MPFIT \citep{markwardt:09}. Green solid line represents the sum of all of the ice components, and the CO$_2$ ice column density in Table~\ref{tab:tab1} was estimated from the integrated absorption, adopting the integrated line strength $A = 1.1\times10^{-17} {\rm cm\ molecule^{-1}}$ \citep{gerakines:95}. Unlike the CO$_2$ absorption profiles observed in quiescent molecular clouds \citep[e.g.][]{whittet:09}, the $15.2\mu$m band in Figure~\ref{fig:co2} shows a remarkable double-peaked profile. Double peaked profiles are commonly observed toward YSOs \citep[e.g.,][]{gerakines:99,pontoppidan:08}, and are ascribed to pure CO$_2$ ices resulting from crystallization of heated H$_2$O-rich ices. However, the double peaks toward the GC candidate YSOs are centered at longer wavelengths ($15.15\mu$m and $15.4\mu$m vs.\ $15.10\mu$m and $15.25\mu$m), and result from CO-rich ($15.15\mu$m peak) and CH$_3$OH-rich ices ($15.4\mu$m peak; see Fig.~\ref{fig:co2}). The strength of the $15.4\mu$m peak is similar to that of the well-studied embedded massive YSO W33A \citep[][bottom panel in Fig.~\ref{fig:co2}]{gerakines:99}. It is ascribed to a Lewis acid-base interaction of CO$_2$ (the Lewis acid) with CH$_3$OH \citep{dartois:99a}. Other species could be acting as a base as well, but CH$_3$OH is preferred due to its high abundance toward W33A: $5\%$--$22\%$ relative to solid H$_2$O \citep{dartois:99b}. Two other YSOs (AFGL~7009S, AFGL~2136) show a prominent $15.4\mu$m peak, and indeed these sources have high CH$_3$OH abundances as well \citep{dartois:99b,gibb:04}. This suggests that the GC candidate YSOs have high solid CH$_3$OH abundances as well.\footnote{This needs to be verified by independent L-band spectroscopy of the $3.53\mu$m C-H stretch mode of CH$_3$OH \citep[e.g.,][]{dartois:99b}.} Although the origin of the large quantities of CH$_3$OH in the previously studied massive YSOs is not fully understood \citep{dartois:99a}, so far all lines of sight with high solid CH$_3$OH abundances are associated with star formation, strengthening the idea that the sources studied in this paper are indeed YSOs. To derive abundances of these molecular absorptions with respect to the hydrogen and solid H$_2$O column densities, we followed the procedure in \citet{boogert:08} to fit the H$_2$O ice and silicate absorption profiles to SL and LL spectra. Figure~\ref{fig:continuum} shows an example for SSTGC~797384. We used the silicate absorption profiles in the line of sight to the GC \citep[GCS~3 spectrum;][]{kemper:04} plus a laboratory spectrum of pure amorphous H$_2$O ice at $T = 10$~K \citep{hudgins:93}. We simultaneously fit a second-order polynomial for a pseudo-continuum (i.e., including corrections for the continuous extinction), the silicate profile, and H$_2$O ice absorption to the $5\mu$m $\leq \lambda \leq 32\mu$m spectrum. We masked absorption features at $6\mu$m, $7\mu$m, and $15\mu$m, and all unresolved emission lines, before performing a non-linear least squares fit. Best-fitting parameters are listed in Table~\ref{tab:tab1}. We obtained a total hydrogen column density from the optical depth of the $9.7\mu$m silicate absorption, assuming $A_V / \tau_{9.7} = 9$ \citep{roche:85} and $N_{\rm H} / A_V \approx 1.87 \times 10^{21}$ cm$^{-2} {\rm mag}^{-1}$ \citep{bohlin:78} at $R_V = 3.1$. The H$_2$ column density was then approximated by $N_{\rm H_2} = N_{\rm H}/2$. The ice column density for the $13\mu$m librational H$_2$O absorption was estimated from the integrated absorption of the best-fitting H$_2$O model. The H$_2$O ice column density from the $6\mu$m bending mode, fit separately after fixing the continuum and extinction to previously found values, is an upper limit because the $6\mu$m absorption is not solely due to H$_2$O ice. We adopted the integrated line strengths $A = 1.2\times10^{-17} {\rm cm\ molecule^{-1}}$ for the bending mode and $A = 3.1\times10^{-17} {\rm cm\ molecule^{-1}}$ for the librational mode \citep{gerakines:95}. Errors in these parameters (Table~\ref{tab:tab1}) are formal estimates made by varying the range of wavelengths that we used for the $9.7\mu$m silicate fitting, or by taking a few different ways of setting the continuum. The gas-phase molecular abundances relative to H$_2$ are listed in Table~\ref{tab:tab1}. Our derived abundances of $\sim10^{-7}$--$10^{-6}$ for C$_2$H$_2$ and HCN are comparable to those found for massive YSOs \citep{lahuis:00,knez:09}, although abundances for SSTGC~524665 have large errors. Intervening molecular clouds in the line of sight to the GC are less likely the main cause of these absorptions, because the average HCN abundance of $2.5\times10^{-8}$ \citep{greaves:96} towards Sgr~B2(M) is an order of magnitude lower than our measurements. Our gas-phase CO$_2$ abundances are an order of magnitude larger than those found towards massive YSOs in \citet{boonman:03}, but our gas to solid abundance ratios for CO$_2$ are consistent with their estimates ($10^{-1}$--$10^{-2}$). Our abundance of CO$_2$ ice relative to H$_2$O ice is within the range ($0.10$--$0.23$) found towards massive YSOs \citep{gerakines:99}. Finally, Table~\ref{tab:tab1} lists our estimates on $A_V$ from the $9.7\mu$m silicate absorption and those from \citet{schultheis:09}, based on the 2MASS and IRAC color-magnitude diagrams of GC red giant branch stars within $2\arcmin$ of the source. Both SSTGC~797384 and SSTGC~803187 have higher $A_V$ values than the average for field stars, implying that a significant fraction of the attenuation is intrinsic to the source. SSTGC~524665 has a lower $A_V$, comparable to the average value for surrounding field stars. We also note that SSTGC~524665 is located at $b \approx -0.2\arcdeg$, so it is possible that it is in front of the GC. If we assume a distance of $8$~kpc for all three sources, and adopt the extinction of surrounding field stars as the foreground extinction to each source, then we derive stellar masses of $12\pm3M_\odot$, $14\pm3M_\odot$, $17\pm6M_\odot$ for SSTGC~524665, SSTGC~797384, and SSTGC~803187, respectively, by using a grid of YSO models \citep{robitaille:06,robitaille:07}. More detailed discussion of the model fitting will presented in a future paper. To summarize, we presented the evidence from IRS spectra for the first spectroscopic identification of massive YSOs in the GC. In our next paper (D.\ An, 2009, in preparation), we will present the results for all 107 YSO candidates, together with additional data from millimeter to radio observations, and use them to better understand the nature of these embedded sources. \acknowledgements We thank David Ardila for helpful discussions of the IRS data reduction. We thank Sean Carey for providing us MIPS photometry before publication. D.\ An and S.\ Ram\'irez thank John Stauffer for helpful discussions. This work is based on observations made with the Spitzer Space Telescope, which is operated by the Jet Propulsion Laboratory, California Institute of Technology under a contract with NASA. Support for this work was provided by NASA through an award issued by JPL/Caltech. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France. \begin{thebibliography} \bibitem[Aoki et al.(1999)]{aoki:99} Aoki, W., Tsuji, T., \& Ohnaka, K.\ 1999, \aap, 350, 945 \bibitem[Bohlin et al.(1978)]{bohlin:78} Bohlin, R.~C., Savage, B.~D., \& Drake, J.~F.\ 1978, \apj, 224, 132 \bibitem[Boogert et al.(2008)]{boogert:08} Boogert, A.~C.~A., et al.\ 2008, \apj, 678, 985 \bibitem[Boonman et al.(2003)]{boonman:03} Boonman, A.~M.~S., van Dishoeck, E.~F., Lahuis, F., \& Doty, S.~D.\ 2003, \aap, 399, 1063 \bibitem[Cami et al.(2006)]{spectrafactory} Cami J., Markwick-Kemper A.J., Van Malderen R., 2009, \apjs submitted \bibitem[Carey et al.(2009)]{carey:09} Carey, S.~J., et al.\ 2009, \pasp, 121, 76 \bibitem[Dartois et al.(1999a)]{dartois:99a} Dartois, E., Demyk, K., d'Hendecourt, L., \& Ehrenfreund, P.\ 1999a, \aap, 351, 1066 \bibitem[Dartois et al.(1999b)]{dartois:99b} Dartois, E., Schutte, W., Geballe, T.~R., Demyk, K., Ehrenfreund, P., \& D'Hendecourt, L.\ 1999b, \aap, 342, L32 \bibitem[Fazio et al.(2004)]{fazio:04} Fazio, G.~G., et al.\ 2004, \apjs, 154, 10 \bibitem[Felli et al.(2002)]{felli:02} Felli, M., Testi, L., Schuller, F., \& Omont, A.\ 2002, \aap, 392, 971 \bibitem[Forster \& Caswell(1989)]{forster:89} Forster, J.~R., \& Caswell, J.~L.\ 1989, \aap, 213, 339 \bibitem[Gerakines et al.(1995)]{gerakines:95} Gerakines, P.~A., Schutte, W.~A., Greenberg, J.~M., \& van Dishoeck, E.~F.\ 1995, \aap, 296, 810 \bibitem[Gerakines et al.(1999)]{gerakines:99} Gerakines, P.~A., et al.\ 1999, \apj, 522, 357 \bibitem[Gibb et al.(2004)]{gibb:04} Gibb, E.~L., Whittet, D.~C.~B., Boogert, A.~C.~A., \& Tielens, A.~G.~G.~M.\ 2004, \apjs, 151, 35 \bibitem[Greaves \& Nyman(1996)]{greaves:96} Greaves, J.~S., \& Nyman, L.-A.\ 1996, \aap, 305, 950 \bibitem[Guglielmo et al.(1998)]{guglielmo:98} Guglielmo, F., Le Bertre, T., \& Epchtein, N.\ 1998, \aap, 334, 609 \bibitem[Houck et al.(2004)]{houck:04} Houck, J.~R., et al.\ 2004, \apjs, 154, 18 \bibitem[Hudgins et al.(1993)]{hudgins:93} Hudgins, D.~M., Sandford, S.~A., Allamandola, L.~J., \& Tielens, A.~G.~G.~M.\ 1993, \apjs, 86, 713 \bibitem[Kemper et al.(2004)]{kemper:04} Kemper, F., Vriend, W.~J., \& Tielens, A.~G.~G.~M.\ 2004, \apj, 609, 826 \bibitem[Knez et al.(2009)]{knez:09} Knez, C., Lacy, J.~H., Evans, N.~J., van Dishoeck, E.~F., \& Richter, M.~J.\ 2009, \apj, 696, 471 \bibitem[Lahuis \& van Dishoeck(2000)]{lahuis:00} Lahuis, F., \& van Dishoeck, E.~F.\ 2000, \aap, 355, 699 \bibitem[Markwardt(2009)]{markwardt:09} Markwardt, C.~B.\ 2009, arXiv:0902.2850 \bibitem[Mehringer(1995)]{mehringer:95} Mehringer, D.~M.\ 1995, \apj, 454, 782 \bibitem[Mehringer et al.(1993)]{mehringer:93} Mehringer, D.~M., Palmer, P., Goss, W.~M., \& Yusef-Zadeh, F.\ 1993, \apj, 412, 684 \bibitem[Morris \& Serabyn(1996)]{morris:96} Morris, M., \& Serabyn, E.\ 1996, \araa, 34, 645 \bibitem[Omont et al.(2003)]{omont:03} Omont, A., et al.\ 2003, \aap, 403, 975 \bibitem[Pontoppidan et al.(2008)]{pontoppidan:08} Pontoppidan, K.~M., et al.\ 2008, \apj, 678, 1005 \bibitem[Ram{\'{\i}}rez et al.(2008)]{ramirez:08} Ram{\'{\i}}rez, S.~V., Arendt, R.~G., Sellgren, K., Stolovy, S.~R., Cotera, A., Smith, H.~A., \& Yusef-Zadeh, F.\ 2008, \apjs, 175, 147 \bibitem[Robitaille et al.(2006)]{robitaille:06} Robitaille, T.~P., Whitney, B.~A., Indebetouw, R., Wood, K., \& Denzmore, P.\ 2006, \apjs, 167, 256 \bibitem[Robitaille et al.(2007)]{robitaille:07} Robitaille, T.~P., Whitney, B.~A., Indebetouw, R., \& Wood, K.\ 2007, \apjs, 169, 328 \bibitem[Roche \& Aitken(1985)]{roche:85} Roche, P.~F., \& Aitken, D.~K.\ 1985, \mnras, 215, 425 \bibitem[Rothman et al.(1996)]{hitemp} Rothman, L.~S., et al.\ 1996, Journal of Quantitative Spectroscopy and Radiative Transfer, 60, 665 \bibitem[Rothman et al.(2005)]{hitran} Rothman, L.~S., et al.\ 2005, Journal of Quantitative Spectroscopy and Radiative Transfer, 96, 139 \bibitem[Schuller et al.(2006)]{schuller:06} Schuller, F., Omont, A., Glass, I.~S., Schultheis, M., Egan, M.~P., \& Price, S.~D.\ 2006, \aap, 453, 535 \bibitem[Schultheis et al.(2003)]{schultheis:03} Schultheis, M., Lan{\c c}on, A., Omont, A., Schuller, F., \& Ojha, D.~K.\ 2003, \aap, 405, 531 \bibitem[Schultheis et al.(2009)]{schultheis:09} Schultheis, M., Sellgren, K., Ram{\'{\i}}rez, S., Stolovy, S., Ganesh, S., Glass, I.~S., \& Girardi, L.\ 2009, \aap, 495, 157 \bibitem[Skrutskie et al.(2006)]{skrutskie:06} Skrutskie, M.~F., et al.\ 2006, \aj, 131, 1163 \bibitem[Smith et al.(2006)]{smith:06} Smith, H.~A., Hora, J.~L., Marengo, M., \& Pipher, J.~L.\ 2006, \apj, 645, 1264 \bibitem[Stolovy et al.(2006)]{stolovy:06} Stolovy, S., et al.\ 2006, Journal of Physics Conference Series, 54, 176 \bibitem[van der Tak et al.(2000)]{vandertak:00} van der Tak, F.~F.~S., van Dishoeck, E.~F., Evans, N.~J., II, \& Blake, G.~A.\ 2000, \apj, 537, 283 \bibitem[Wannier(1980)]{wannier:80} Wannier, P.~G.\ 1980, \araa, 18, 399 \bibitem[Warren et al.(2007)]{warren:07} Warren, S.~J., et al.\ 2007, arXiv:astro-ph/0703037 \bibitem[Werner et al.(2004)]{werner:04} Werner, M.~W., et al.\ 2004, \apjs, 154, 1 \bibitem[Whittet et al.(2009)]{whittet:09} Whittet, D.~C.~B., Cook, A.~M., Chiar, J.~E., Pendleton, Y.~J., Shenoy, S.~S., \& Gerakines, P.~A.\ 2009, \apj, 695, 94 \bibitem[Whitney et al.(2003)]{whitney:03} Whitney, B.~A., Wood, K., Bjorkman, J.~E., \& Cohen, M.\ 2003, \apj, 598, 1079 \bibitem[Whitney et al.(2004)]{whitney:04} Whitney, B.~A., Indebetouw, R., Bjorkman, J.~E., \& Wood, K.\ 2004, \apj, 617, 1177 \bibitem[Yusef-Zadeh et al.(2009)]{yusefzadeh:09} Yusef-Zadeh, F., et al.\ 2009, arXiv:0905.2161 \end{thebibliography} \end{document}