ABUNDANT CH₃OH MASERS BUT NO NEW EVIDENCE FOR STAR FORMATION IN GCM0.253+0.016

GCM0.253+0.016 (also, G0.253+0.016, G0.216+0.016, M0.25+0.01, M0.25+0.11, or "The Brick," as it has been variously referred to in the literature) is one such extremely quiescent CMZ cloud, located ∼45 pc in projection from the dynamical center of the Galaxy (assuming a Galactocentric distance of 8.4 kpc; Ghez et al. 2008; Gillessen et al. 2009; Reid et al. 2014). Although its appearance as a prominent infrared dark cloud indicates that it is occulting most of the infrared emission from the nuclear bulge; its chemistry, kinematics, high temperatures, and large linewidths are all consistent with being located at the CMZ, and it is commonly taken to lie on the near side of the CMZ (Lis & Menten 1998). In total, GCM0.253+0.016 is suggested to have a mass of 1–2 × 10⁵ ${{M}_{\odot }}$, making it one of the five most massive clouds in the CMZ (Lis et al. 1994; Longmore et al. 2012). It is also the only compact >10⁴ solar mass cloud found thus far in the entire Galaxy which does not exhibit advanced stages of star formation (Ginsburg et al. 2012; Tackenberg et al. 2012; Urquhart et al. 2014). The comparably massive Maddalena cloud in the outer Galaxy, which also does not show evidence of active massive star formation, is extended over ∼100 pc. (Maddalena & Thaddeus 1985; Megeath et al. 2009). The large mass and relatively high average density of this cloud (n ∼ 1 × 10⁵ cm⁻³; Longmore et al. 2012; Kauffmann et al. 2013) suggest that it is capable of massive star and perhaps even cluster formation (Longmore et al. 2012). However, there is no clear evidence in this cloud for ongoing massive star formation apart from a single water maser (Lis et al. 1994).

Recent continuum studies of GCM0.253+0.016 at infrared to radio wavelengths have continued to search for signposts of ongoing star formation. Longmore et al. (2012) analyze Herschel observations of the cloud and find no embedded heating sources at wavelengths up to 70 μm. At 280 GHz with the SMA, Kauffmann et al. (2013) find only one strong, compact dust core, which they suggest is indicative of a low potential for star formation (but see also Johnston et al. 2014, who find more extensive dust continuum emission at 230 GHz, also with the SMA). Both Johnston et al. (2014) and Rathborne et al. (2014b) then measure the column density probability distribution function (PDF) from the dust continuum, finding it to be log-normal. Rathborne et al. (2014b) do find a deviation from this form at high column densities (interpreted as self-gravitation), but state that this corresponds to just the single dust core already known to contain a water maser. The only potential indications of more advanced star formation come from high-resolution radio observations by Rodríguez & Zapata (2013) who identify three compact thermal continuum sources which they suggest could be embedded B-stars. However, all of these sources are located outside of the bulk of the gas and dust emission in the cloud, on its periphery.

Although continuum observations show few signs of previously missed star formation and are largely consistent with GCM0.253+0.016 being in a quiescent, pre-star-forming stage, a host of recent observations of the gas reveal many other complexities. Velocity dispersions on spatial scales of 0.07–0.1 pc are observed to range from extremely turbulent (>30 km s⁻¹ as measured in ALMA observations of SO; Higuchi et al. 2014) to clumps with line widths apparently less than 1 km s⁻¹ (from SMA observations of the quiescent gas tracer N₂H⁺, Kauffmann et al. 2013)—the narrowest line widths yet observed in a CMZ cloud. Abundant emission from other shock-tracing molecules SiO and CH₃OH are additionally observed in GCM0.253+0.016 by Rathborne et al. (2015) and Johnston et al. (2014). Johnston et al. (2014) also present the first resolved temperature measurements of the cloud, using H₂CO, which indicate extremely high temperatures in the clumpy gas (T ∼ 300 K), much higher than temperatures measured from single-dish observations of the same lines (Ao et al. 2013), though comparable to temperatures measured in this cloud from single-dish observations of highly excited lines of NH₃ (Mills & Morris 2013). The most surprising new observations are of a series of HCO⁺ absorption filaments, suggested to be tracing the surface magnetic field lines in the cloud, which which have never before been seen in any giant molecular cloud (Bally et al. 2014).

Complementing this existing suite of molecular line observations, we present the first interferometric study of both the molecular and ionized gas in GCM0.253+0.016 at radio wavelengths, using the Karl G. Jansky Very Large Array (hereafter, "VLA"), a facility of the National Radio Astronomy Observatory.⁷ The upgraded VLA offers more sensitive receivers and broad spectral bandwidths up to 8 GHz for both sensitive radio continuum imaging and spectral line surveys. Our new VLA observations of GCM0.253+0.016 exploit both of these capabilities to provide a comprehensive new study of the sub-parsec morphology, kinematics, and physical conditions of gas in this cloud. Our radio continuum observations allow for a more sensitive search for signs of ongoing star formation to determine whether indications of high-mass star formation in CMZ clouds may have been previously missed. With our molecular line data, we can examine kinematics of both low and high density gas in the cloud using NH₃, an abundant tracer of gas having densities greater than a few 10³ cm⁻³. The NH₃ observations are also sensitive to gas over a wide range of temperatures (from tens to hundreds of K), enabling us to investigate the full range of temperatures present in the pre-star-forming gas, and to map the temperature structure of that gas across the cloud. In this paper, we first present an analysis of the continuum emission and the morphology of the detected molecular species in GCM0.253+0.016, with detailed analyses of the kinematic and temperature structure to be presented in subsequent papers.

In Section 2, we describe the VLA observations and the procedures used to calibrate and image these data. We then present an overview of our study of GCM0.253+0.016, beginning with the properties of the continuum emission which are analyzed in Section 3. The morphology and kinematics of the molecular gas are subsequently analyzed in Section 4. In Section 5, we focus on emission from the 36.2 GHz CH₃OH line and present a catalog of more than 70 candidate collisionally excited masers. Finally, in Section 6, we discuss constraints on the amount and nature of ongoing star formation in this cloud.

These observations employed the WIDAR correlator on the VLA in order to simultaneously observe both a wide spectral bandwidth for continuum studies and a large number of spectral lines. Observations in each band (K and Ka) are divided into two separate, continuous subbands of ∼0.84 GHz width which are each subdivided into seven spectral windows. In the K band, the subbands were centered on 24.054 and 25.375 GHz, and for the Ka band the subbands were centered on 27.515 and 36.35 GHz. The typical spectral resolution per spectral window is 250 kHz, with 512 channels per spectral window. However, for three spectral windows, covering (1) the NH₃ (1,1) and (2,2) lines and their hyperfine satellites, (2) the 36.6 GHz CH₃OH maser line, and (3) the CH₃CN K = 2–1 transitions, the resolution was increased (125 kHz, ∼1–1.5 km s⁻¹) to better resolve the line structure.

In order to map the majority of GCM0.253+0.016 in the K and Ka bands we used two pointings (see Table 1) oriented along the major axis of the cloud, which is elongated in declination. The total integration time for each field was ∼25 minutes. The distance between the pointings was 83 5 at K band and 70 7 at Ka band. Primary beam sizes range from ∼110'' at K band to ∼100'' and ∼75'' for the lower and upper frequencies observed at Ka band, respectively.

Table 1.  Observed Fields

Field	R.A.	Decl.	Array	Int.
	(J2000)	(J2000)	ConFigure	Time (minutes)
K-band North	17h46m0895	−28°41'56 8	DnC	24.6
K-band South	17h46m0960	−28°43'24 8	DnC	24.9
Ka-band North	17h46m0844	−28°42'01 0	DnC	24.3
Ka-band South	17h46m1026	−28°43'07 8	DnC	24.4

Download table as:  ASCIITypeset image

The calibration of our K- and Ka-band VLA observations was performed using the CASA software provided by NRAO.⁸ 3C286 was observed as the flux density calibrator, J1733-1304 was used as the bandpass calibrator, and J1744-3116 was used as the gain and phase calibrator. Reference pointing was performed every hour on the phase calibrator, which was observed every ∼25 minutes throughout the observations. In addition, at these higher VLA frequencies, corrections were made for the atmospheric opacity, which is determined from a mix of both actual weather data and a seasonal model using the plotweather task.

The K- and Ka-band continuum and spectral data were imaged using the CLEAN task in CASA. As stated in Section 2.1, two pointings were observed to cover the cloud at both K and Ka bands, which were mosaicked together when imaging. Continuum data were obtained by flagging out the spectral lines and end channels. Briggs weighting with a robust parameter of 0.5 in order to balance the point-source resolution with the sensitivity, giving synthesized beams of 1 59–2 30. The rms noise levels in the continuum images ranges from 29 to 55 μJy beam⁻¹. These values are generally about twice the theoretical rms noise levels of 17–27 μJy beam⁻¹, which may be due to the increased contributing emission coming from the Galactic plane. The largest angular scale to which the data are sensitive is ∼60'' at K band (or ∼2.4 pc) and ∼40'' (∼1.6 pc) at the higher-frequency Ka-band subband.

The K- and Ka-band spectral line data were also imaged using CLEAN. All lines were imaged individually, and continuum emission was subtracted using the CASA task imcontsub. The data were imaged at their intrinsic frequency resolution with no smoothing, and all spectral lines (except for the 36 GHz CH₃OH line) were imaged with natural weighting, resulting in synthesized beams which ranged from 1.58 to 1 77. For the 36 GHz CH₃OH line, which exhibited primarily point-source emission, the data were imaged using Briggs weighting with a robust parameter of 0.5. Beam parameters for all images are given in Table 2.

For the NH₃(1, 1)–(6, 6) and HC₃N (3–2) and (4–3) transitions we used multiscale deconvolution to maximize sensitivity to extended emission in the cloud. The images were cleaned using beams with sizes of 0, 5, 20, and 80 pixels, and the parameter setting the relative weighting between these scales (0–80) was set to 0.6.

In the resulting multiscale images, "negative bowl" features indicative of missing extended flux were minimized, however the observations are still not sensitive to emission on scales larger than ∼1'. We did not use multiscale CLEAN for the NH₃(7, 7)–(9, 9) and CH₃CN (2–1) lines, as it was not found to significantly improve the imaging of these weak lines, or for the 36 GHz CH₃OH line, for which the emission was primarily unresolved or on small spatial scales. The individual imaging parameters for each line can be found in Table 2. In general, the rms noise levels per channel range from 0.7 to 3.0 mJy beam⁻¹ with the exception of CH₃OH (4–3). The typical rms levels for this transition varied, from 2.16 mJy beam⁻¹ in maser-free channels to 116 in the brightest maser channel. These rms values are consistent with the theoretical rms values of 1.4 and 0.9 mJy beam⁻¹ for channel widths of 0.125 and 0.250 MHz, respectively.

For observations of the Ka-band (36 GHz) CH₃OH line, in which there are many strong point sources, we additionally self-calibrated the data. For each pointing, a bright CH₃OH point source with minimal additional emission surrounding the source was chosen for self-calibration. Here, the requirement that the point sources be relatively isolated was more important than that they be the strongest in the cube. Each pointing was first imaged using CLEAN with a small number of iterations, thus producing a map and model of the emission. This model was used by the CASA task gaincal for both phase and amplitude calibrations. To begin with, phase-only self calibration was applied until the signal-to-noise improvement was no longer significant (2–3 iterations). After phase-only calibration, a single iteration of amplitude and phase self-calibration was performed. The self-calibration amplitude and phase solutions for this single channel were then applied to all of the channels in each pointing, and the pointings were jointly imaged using CLEAN as described above to form a final image.

The regions of continuum emission that we will evaluate here are shown in Figure 2. The regions are defined by a contour levels of 6 or 10 times the 24.1 GHz rms noise level of 30 μJy beam⁻¹) m depending on whether the regions are extended or compact, respectively. Table 3 summarizes the properties of these regions of continuum emission. In addition to presenting flux measurements for these regions from each of the VLA continuum images, we give the 90 GHz fluxes from both ALMA 12 m only images and single-dish-corrected ALMA images from Rathborne et al. (2014b).

Zoom In Zoom Out Reset image size

Out of the 10 selected areas of interest, 4 regions (C2, C7, C8 and C9) are large and diffuse (>30''/1 pc across) and are primarily located in the eastern part of the field. Of these, C9, located in the southern part of the field, is the largest and brightest at 24.1 GHz. Contained within the 6σ contours of this region is a large elliptical or "shell"-like region having a long axis extent of ∼30'' as well as a long thin "filament" of emission running from northwest to southeast that lies tangential to the elliptical region, with the two intersecting at α(J2000) = 17^h46^m127, δ(2000) = −28°43'24''. This tangential filament is ∼35'' in length, corresponding to a physical size of 1.4 pc. Although some clumpy emission in the region of the shell and filament is detected above the noise in the 36.4 GHz image, the size of the region is on the order of the largest angular size at this frequency, and likely its emission has been suppressed. Both shell and filament are still faintly seen at 3 mm (Rathborne et al. 2014b), though their structure is not apparent at 230.9 GHz (Johnston et al. 2014).

Regions C7 and C8 have very similar structures and together form an apparent curved ridge of continuum emission across the cloud. While these two regions appear to fall along the same structure, they are separate at the 6σ level, and we treat them as separate regions. We note that like C9, the large-scale diffuse structure of these sources has likely been suppressed, and if there were a true connection between them, the present data would not be sensitive to emission on that spatial scale. At 36.4 GHz, where the suppression of large-scale emission is most severe, only the brightest knots in these sources are detected above the noise at this frequency. The morphologies of these two regions are also reminiscent of that of the adjacent molecular gas in the clouds, as we will discuss further in Section 6.2.

Like C9, the continuum emission in the most northern diffuse region, C2, exhibits filamentary structure. The eastern edge of C2 forms a straight line that is parallel to the tangential filament in C9. The straight edge of C2 also extends northward to an additional clump that is separate in the 6σ contour levels. The total length of this linear feature, including the additional northern clump of emission, is ∼46'' (1.85 pc), larger than the largest angular size scale of these data, suggesting that there could be extended emission connecting the two northern clumps that has been resolved out. This linear structure is apparent from 24.1 to 27.5 GHz, but it is absent at 36.4 GHz, likely because it falls in the less-sensitive outer regions of the primary beam at that frequency. As with the tangential filament C9, the linear structure of C2 is also present at 3 mm (Rathborne et al. 2014b). The morphology of this feature also appears to match that of an adjacent molecular gas filament seen in the NH₃ maps presented in Section 4.1. We discuss this relationship further in Section 6.2.

In addition to the extended continuum emission discussed above, there are several smaller regions of more compact continuum emission located inside of the confines of the cloud, as traced by the molecular gas emission. These continuum regions (C1, C3, C4, and C6) are relatively bright and also have counterparts at higher frequencies: all are detected in the 90 GHz ALMA image of Rathborne et al. (2014b). The remaining compact regions, C5, and 10, are the the brightest sources detected at 24.1 GHz, and lie outside of the boundaries of the cloud traced by the molecular gas emission. These sources are only marginally resolved by the VLA observations, and their peak intensities at 24 GHz are greater than 1.0 mJy beam⁻¹, and are larger than the peak intensities of other sources in the cloud (except C1) by at least a factor of 2. Although they lie near the edge of the Rathborne et al. (2014b) ALMA map, they are still detected at 90 GHz.

A number of the compact (<2'') sources were also identified by Rodríguez & Zapata (2013) using higher resolution VLA data. Their sources JVLA 1, 4, and 6 correspond to the previously discussed sources C5, C1 and C10, respectively. In addition, they detect a source (JVLA 5), that is part of our more extended source C7. As Rodríguez & Zapata (2013) also do not detect our C3, one of the stronger compact sources we detect in GCM0.253+0.016, it seems likely that these higher resolution data are more strongly affected by spatial filtering (the largest angular scale recoverable in the B-configuration VLA data of Rodríguez & Zapata 2013 should be around 10''). It may then be that several of the apparently compact sources they detect are simply the peaks of intrinsically more extended structures. Rodríguez & Zapata (2013) suggested these sources may represent high mass star formation associated with this cloud, in excess of that previously inferred by the presence of a single water maser in the northern part of the cloud. However, we note that in addition to several of their sources being associated with extended emission, all of them also lie outside of the dense molecular gas in the cloud, which is not what is expected if these sources are (proto)stellar in nature. We will discuss which of the compact sources we identify could be most consistent with star formation, based on their spectral indices, fluxes, morphologies, and relation to the molecular gas in the cloud, in Section 6.2.

To determine the nature of the radio continuum sources, and particularly to judge the amount of ongoing star formation in GCM0.253+0.016, the emission mechanism of the observed radio continuum in this cloud must be identified. We have calculated spectral indices for the 10 representative regions identified above. However, the VLA data alone are found to be insufficient for calculating accurate spectral indices, as the small range of radio frequencies probed by these observations does not provide a large lever arm for determining spectral indices, and the combination of the extended nature of much of the emission and the location of many sources near the edge of the primary beam appear to make the fluxes derived in the Ka band systematically low. For this reason, spectral indices are calculated using the fluxes at both K-band frequencies, and the 90 GHz fluxes from the ALMA continuum images of Rathborne et al. (2014b). Separate spectral indices are calculated using the both the images made from just the ALMA 12 m array, and the image that has been additionally corrected for missing extended flux via combination with single dish data. We expect that the true spectral indices likely fall between these two values, as our K-band data are sensitive to larger angular scales than the ALMA data, and should thus recover more flux than the ALMA 12 m only image, but less than the single-dish corrected ALMA image.

We find that the three sources at the center of the cloud (C3, C4, and C6) have consistently rising spectral indices, with α between 0.5 and 1.4, depending on whether the uncorrected or corrected ALMA fluxes are used in this calculation. The two sources outside of the cloud (C5 and C10) have consistently negative spectral indices: the spectral index of C5 is between −0.9 and −1.3, while the spectral index of C10 is between −0.3 and −0.7. This suggests that the emission from these two sources is dominated by nonthermal processes. This is in good agreement with Rodríguez & Zapata (2013), who find that the spectral indices between 1.3 and 5.6 cm for C5 and C10 (their sources JVLA 1 and JVLA 6, respectively) are −0.9 ± 0.1 and −0.3 ± 0.2, respectively.

The remaining sources in GCM0.253+0.016 (C1, C2, C7, C8 and C9), the majority of which are extended, have nearly flat spectral indices (ranging from slightly negative to slightly positive in the uncorrected and corrected ALMA images, respectively), consistent with free–free emission. It has been suggested that this cloud could be a good target for detecting limb-brightened synchrotron emission from cosmic ray interactions, as it previously showed few signs of free–free continuum emission that would confuse the synchrotron signal (Jones 2014). However, our observations show that there is actually significant free–free emission in GCM0.253+0.016, and we do not see any indication of extended synchrotron emission. While we do detect continuum emission from the limb of GCM0.253+0.016, it appears thermal in nature, and is only present on the eastern edge of the cloud. While lower frequency observations might prove more optimal for searching for extended synchrotron emission from this cloud, GCM0.253+0.016 overlaps with a supernova remnant identified at 90 cm (Kassim & Frail 1996), and so ultimately this cloud is likely not an ideal candidate for detecting a synchrotron signal from cosmic ray interactions.

For these sources and the sources with rising spectral indices we then calculate the Lyman-continuum photon rate to further constrain the properties of this thermal emission and its origin. Assuming the continuum emission in GCM0.253+0.016 is thermal in nature, caused by ionization from an external or embedded source, calculating the Lyman-continuum photon rate can give insight into the types of sources required to stimulate this emission. The number of ionizing Lyman-continuum photons needed to produce this emission can be calculated using the formulation of Mezger & Henderson (1967):

$$\begin{eqnarray}\begin{array}{ccccccccccccccc} {{N}_{{\rm Lyc}}} & = & 1.301\times {{10}^{49}}{{\left( \frac{{{T}_{e}}}{{\rm K}} \right)}^{-0.3}}\left( \frac{{{S}_{\nu }}}{{\rm Jy}} \right){{\left( \frac{D}{{\rm kpc}} \right)}^{2}} \\ {} & {} & \times \;{{\left( {\rm ln} \left( \frac{0.0499\;{\rm GHz}}{\nu } \right)+1.5{\rm ln} \left( \frac{{{T}_{e}}}{{\rm K}} \right) \right)}^{-1}} \\ \end{array}\end{eqnarray} \tag{ 1 }$$

where T_e is the electron temperature, assumed to be 10,000 K, S_ν is the flux density at a frequency ν in GHz, and D is the distance to the Galactic center, assumed to be 8.4 kpc. The Lyman-continuum photon rate was calculated for all 10 of the sources in Figure 2 (except for C5 and C10, which have a nonthermal spectral index at these frequencies) using the flux densities at 24.1 GHz, and are presented in column 12 of Table 3. The tabulated values of log N_L range from 45.9 to 47.5 with the largest values (>46.7) corresponding to the large diffuse regions. Assuming the stars producing this ionization are on the zero-age main sequence, which would be expected for a cloud undergoing star formation, this range of log N_L values would correspond to ionization by a single star with a spectral type of B1 to O9.5 (Panagia 1973). The three compact regions (C3, C4, and C6) located toward the center of the cloud would each be ionized with a star of spectral type B0.5. However, as these latter sources have apparently rising spectral indices, it is likely that the inferred Lyman continuum fluxes are either systematically overestimated (if these fluxes are contaminated with dust emission) or systematically underestimated (if these sources are in fact optically thick). As we will discuss further in Section 6.2, it is not possible to determine which is more likely.

The majority of the observed lines in GCM0.253+0.016 (8/12) are from ammonia (NH₃ ). Figure 3 shows the peak intensity of these eight observed transitions of NH₃: (J,K) = (1, 1), (2, 2), (3, 3), (4, 4), (5, 5), (6, 6), (7, 7), and (9, 9) (see Table 2 for image parameters). The first seven lines were imaged simultaneously with a single correlator setting at K band, while the (9, 9) was observed separately at Ka band. As the emission from the (9, 9) line is much weaker than the others, it was smoothed to improve the imaging signal to noise. The strongest observed line is the (3, 3) line (typically about twice as bright as the (1, 1) or (2, 2) lines). We take it to be generally representative of the distribution of NH₃ in this cloud given that, as can be seen in Figure 3, all of the observed NH₃ transitions exhibit very similar structure. In general, the emission is diffuse and filamentary throughout the cloud with many curved features. In the strongest lines (J, K ≤ 4), much of the emission can be seen to be concentrated in a number of compact clumps. These clumps are most prominent in the (3, 3) line, likely in part because this line is the strongest. The clumps have typical brightness temperatures of 10–60 K in the (3, 3) line which is consistent with thermal emission, although (3, 3) masers have been previously suggested to exist in CMZ clouds (Martin-Pintado et al. 1999).

Zoom In Zoom Out Reset image size

In our NH₃ maps, we identify two primary features of the morphology of GCM0.253+0.016. emission. The first is an apparently "C" shaped arc located near the center of the cloud (hereafter "C-arc"). The "C-arc" extends roughly 90'' (or 3.7 pc) in decl. (from −28°42'00'' to −28°43'30'', at an R.A. of ∼17^h46^m08^s). The "C-arc" structure can be seen in recent millimeter spectral line studies of this cloud using ALMA (Higuchi et al. 2014; Rathborne et al. 2015), and the SMA (Johnston et al. 2014). It is suggested by Higuchi et al. (2014) that this feature is the remnant of a recent collision between GCM0.253+0.016 and a smaller cloud. The "C-arc" has roughly the same brightness as emission in other regions of the cloud in the (1, 1) and (2, 2) lines (likely because the bulk of the emission in these lines is optically thick, as will be discussed further in a subsequent paper) but it is prominent in the higher-excitation lines of NH₃. Intriguingly, this arc follows very well the direction of the magnetic vectors inferred from recent polarization studies of this cloud (Pillai et al. 2015).

The second feature is a "tilted bar" below the "C-arc," beginning on the eastern side of the cloud at a declination of −28°435 and spanning nearly the entire width of the cloud in right ascension. In lines of NH₃ (3, 3) and above, this region and the adjacent southern portion of the C-arc are the sources of the most intense NH₃ emission in GCM0.253+0.016, and this tilted bar contains the bulk of the brightest clumps seen in the NH₃ (3, 3) line. In addition to the prominent tilted bar in the southern half of the cloud, there are also a number of weaker linear emission features in both the north and south of the cloud. The easiest to identify in the NH₃ maximum emission maps is a feature on the eastern edge of the cloud above the "C-arc." It is narrow, extending from a decl. of 4125 to 4175, and is best seen in the K ≤ 4 lines. The southern edge of the "C-arc" and the southwestern "tail" of the bar also form nearly linear elongations. We will return to a discussion of the linear emission features in GCM0.253+0.016 and how they relate to the linear radio continuum features and the HCO⁺ absorption filaments identified by Bally et al. (2014) in Section 6.2.

In addition to the NH₃ lines, in the Ka band, we also detect the (4₋₁–3₀) line of CH₃OH and the (3–2) and (4–3) transitions of HC₃N and the (2−1) doublet of CH₃CN, maximum emission maps of which are shown in Figure 4. The CH₃OH line shows the most striking difference in morphology compared to all other lines imaged in this cloud: it is primarily composed of emission from dozens of discrete point sources. More than half of these point sources are located in the southern bar. The majority of the observed point sources have brightness temperatures >400 K and are likely masers; we discuss the nature of the observed 36.2 GHz CH₃OH emission sources in GCM0.253+0.016 further in Section 5. Overall, these molecules trace the same general structure as seen in NH₃: the "C-arc" is clearly visible, and the southern bar can be seen in the CH₃OH and HC₃N lines, though it is less prominent in the faint line CH₃CN line. However, the HC₃N and CH₃CN images both exhibit stronger emission at the center of the cloud (along the "C-arc," between a declination of −28°42'.5 and −28°43') than is seen in either the NH₃ or CH₃OH images. As can be seen in millimeter continuum images (Johnston et al. 2014; Rathborne et al. 2014b), the center of the cloud is also where the dust continuum emission is strongest, suggesting that it is the location of the densest gas.

Zoom In Zoom Out Reset image size

Although GCM0.253+0.016 may be considered quiescent in terms of its (lack of) ongoing star formation activity, its kinematics are much more active. In the left panel of Figure 5, we show a Moment 1 map of the intensity-weighted velocity from the NH₃ (3, 3) line, which is the brightest of the observed NH₃ transitions. The map was made by limiting the emission spatially to the region previously identified during the initial CLEAN, and by using a threshold of 2–3 times the rms noise value for the spectral line. Emission toward the cloud in the (3, 3) line spans velocities from −10 to 90 km s⁻¹ (with weak emission, which does not contribute significantly to the average values in this figure, extending to velocities as low as −40 km s⁻¹). The lowest velocities (−10 to +20 km s⁻¹) in the cloud generally fall in the northern region of the cloud, while the southern region of the cloud is characterized by gas in the range of 20–60 km s⁻¹. The highest velocities (70–90 km s⁻¹) are primarily confined to a region at the southwest edge of GCM0.253+0.016, which has been suggested by Rathborne et al. (2014a) and Johnston et al. (2014) to be a separate cloud which may or may not be related to the main cloud.

Zoom In Zoom Out Reset image size

A similarly constructed map of the intensity-weighted velocity dispersion (Moment 2) in the same line (Figure 5, middle panel) shows dispersions ranging from 2 to 30 km s⁻¹. However, the largest of these velocity dispersions (Δv > 20 km s⁻¹) may be misleading due to the presence of multiple components along some lines of sight. This complicated velocity structure of GCM0.253+0.016 is represented in three example spectra shown in the rightmost panel of Figure 5. There are multiple velocity components in the northern part of the cloud along the same line of sight that confuse this analysis, causing the velocity dispersion to represent a combination of the width of these components and their separation. An example of this is shown in spectrum A, a triple profile spectrum with intensity peaks around −15, 0, and 30 km s⁻¹. Areas with multiple velocity components can be similarly poorly represented in the Moment 1 map: often the average velocity lies between velocity peaks, and is located at a velocity which little or no gas is present. Spectra toward the center of the Brick, along the "C-arc," discussed in Section 4.1, show a double peak profile as can be seen in spectrum B of Figure 5. The peak intensities of the profiles presented in spectrum B fall at velocities of 10–15 and 35–40 km s⁻¹. This kinematic structure is typical for regions along the "C-arc." The southern part of GCM0.253+0.016 has single line profiles (see spectrum C), with velocities greater than 30 km s⁻¹, that typically represent the brightest emission in the cloud, with brightness temperatures at least 2–3 times the emission from other parts of the cloud.

Although we observe the CH₃OH line with similar spatial resolution as Yusef-Zadeh et al. (2013a), our spectral resolution of ∼1 km s⁻¹ is better able to discern whether these sources have nonthermal brightness temperatures. In order to more quantitatively analyze the observed CH₃OH emission in GCM0.253+0.016 and determine whether the observed sources are masers, we have produced an initial catalog listing the properties of the strongest detected sources.

The CH₃OH emission seen in Figure 6 is clustered together both spatially and spectrally. This clustering makes manually distinguishing between individual sources difficult. In order to examine these complicated fields, we adopt a version of the source detection algorithm Clumpfind (Williams et al. 1994) which distinguishes between sources that may partially overlap in position or velocity. Clumpfind identifies local maxima, then examines the emission surrounding the maxima both spatially and spectrally to determine the boundaries of the source. No assumptions about the clump geometry, neither spatially or spectrally, are made during processing by the algorithm. Clumpfind produces a list of maser candidate clumps with uniform criteria. The output of Clumpfind is then used to construct the catalog. Since a significant portion of the maser emission lies near the edge of the observed Ka-band field, a primary beam correction was applied while calculating the properties of the sources found by Clumpfind. Clumpfind searched for emission down to six times the rms noise in each channel.

Our Clumpfind analysis of GCM0.253+0.016 yields 383 CH₃OH clumps with a brightness above six times the rms noise. However, in order to remove the possibility of false detections, we required sources to have a brightness greater then 10 times the rms noise in their spectral channel in order to be included in the catalog; 195 CH₃OH clumps meet this criterion, which is a conservative cut-off that ensures that we are examining masers and not artifacts from several of the extremely bright masers in the field. However, as a result the final catalog of sources is incomplete below a flux of 1.0 Jy (the largest residual in the cube after removing the Clumpfind-detected sources), with the incompleteness being most significant near the velocity range of the brightest masers. Additionally, we make a total flux cut at 0.3 Jy, below which emission structure in the image begins to become significantly compromised by the missing flux on large scales. This removes 47 more sources, leaving a total of 148 detected CH₃OH point sources.

These Clumpfind sources are then divided into two catalogs: masers, and candidate masers. Of the 148 cataloged point sources, 68 have brightness temperatures >400 K, in excess of the highest gas temperatures suggested to exist in this cloud (∼325 K, Mills & Morris 2013; Johnston et al. 2014). This indicates that they are likely nonthermal, and we classify them as masers. Their properties, including the FWHM line width and peak brightness temperature, are given in Table 4. The remaining 80 sources have brightness temperatures that could be thermal, and so cannot yet be confirmed to be masers, given the limited spatial and spectral resolution of these data. Their properties are given in Table 5. Spectra for all of the cataloged CH₃OH sources, both masers and candidates, are presented in Figures 7 and 8. Due to the spatial and spectral clustering of the sources, the spectrum of a candidate may show other peaks from bright masers located nearby. To aid in identification of weak masers near brighter sources, the central velocity of each maser candidate is indicated by a dashed line in each spectrum.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

More than half of both the masers (37; 54%) and maser candidates (43; 54%) are spatially unresolved. These sources are deemed to be spatially resolved if their FWHMs are larger than twice the synthesized beam area. All of the masers and maser candidates also have relatively narrow line widths, with a mean FWHM for all cataloged sources of 3.6 km s⁻¹. More of the candidates (28; 35%) are spectrally unresolved than the masers (10; 15%), with sources deemed to be spectrally resolved if their FWHMs are larger than twice the channel resolution of 1.02 km s⁻¹. As we expect masers to have subthermal (<1 km s⁻¹) linewidths and generally to be spatially unresolved point sources, the large fraction of masers that are both spectrally and spatially resolved suggests that there is still confusion in this catalog, and that our observations are still underestimating the true number of maser sources in this cloud. Higher-resolution VLA observations should be able to confirm this and to determine the clumping properties of the CH₃OH masers in this cloud.

As the maser candidates appear, apart from their lower brightness temperatures, to be quantitatively similar to the masers (with similar fractions of both sources spatially and spectrally unresolved) we expect that the bulk of these sources will also prove to be masers. Possible exceptions to this are some of the candidate sources that have broader lines and more spatially extended emission. These sources (as well as regions of extended emission with ${{T}_{B}}\ll 100$ K that are not included in our catalogs) could represent thermal or "quasithermal" emission, as we discuss further below.

The number of CH₃OH sources detected in GCM0.253+0.016 in our observations (148) is thus far unprecedented for a CMZ cloud. It is larger than the number of 36 GHz masers (10) recently identified by Sjouwerman et al. (2010) in the 50 and 20 km s⁻¹ CMZ clouds, however this difference in the number of sources may be due in part to the fact that our observations are ∼5× more sensitive. It is also more than an order of magnitude greater than the number of 36 GHz sources (8) previously identified in this cloud by Yusef-Zadeh et al. (2013a). However, we should note that the positions for these sources given by Yusef-Zadeh et al. (2013a) do not match the positions of any of our masers, and it appears upon checking the archival data for these observations, that the previously published positions of 36 GHz sources in this cloud are incorrect. In the central part of the cloud, where we compare our data to the archival data from the Yusef-Zadeh et al. (2013a) observations, we detect three sources, at the positions of our brightest masers (M10, M18 and M25), but nothing at the published positions of three sources in this field (catalog numbers 42, 43, 44). We will assume that the number of sources cataloged by Yusef-Zadeh et al. (2013a) in this cloud can still be taken to be order-of-magnitude representative of what can be detected in this cloud at the sensitivity of their survey. Based on the yield of the Yusef-Zadeh et al. (2013a) study—356 individual 36 GHz sources over a surveyed area of 0.33 square degrees—our observations then suggest that clouds in the CMZ could host thousands of these masers, detectable by observing more clouds in the same way as for the GCM0.253+0.016 data presented here, or potentially with a higher spectral-resolution survey than that conducted by Yusef-Zadeh et al. (2013a).

Focusing on the subset of sources that we can confirm to be masers, it can be seen from Figure 6 that these are distributed throughout the entire cloud. However, 44 of the masers (more then 60% of the total) are concentrated in the southern regions of the cloud, south of declination = −28°43'00 0 (see Figure 6 left). These regions correspond to the regions denoted from the ammonia maps as the "C-arc" and the "bar" (see Section 4). Masers in the northern part of the cloud are typically more isolated than masers that fall in these other two regions. In addition to containing the majority of the maser emission, the southern region of the cloud also contains the brightest maser emission. With the exception of M3, all of the brightest masers are found in the southern region of the cloud. These brightest sources (M1 through M8) exhibit brightness temperatures in excess of 4000 K. In addition to the seven brightest masers, 25 additional masers have brightness temperatures in excess of 1000 K. The velocity range of the CH₃OH masers is from −5 to 50 km s⁻¹. Masers with velocities less than 20 km s⁻¹ are seen only in the northern portion of the cloud, while masers at velocities greater than 20 km s⁻¹ are seen throughout the cloud.

In general, the velocity distribution of the masers (−5 to 50 km s⁻¹) follows the same velocity distribution traced by NH₃ (3, 3) and seen in Figure 5. The northern part of the cloud shows maser emission occurring at roughly two velocities, ∼10 and 35 km s⁻¹, while the southern part of the cloud has a single, higher velocity component, ∼35 km s⁻¹. These maser velocities are similar to the NH₃ (3, 3) velocities, discussed in Section 4.2. The brightest masers, typically found in the southern part of the cloud, have velocities between 30 and 40 km s⁻¹, which is also the velocity of the brightest NH₃ (3, 3) emission.

The majority of the CH₃OH masers do not correspond to any continuum features, though a few exceptions are seen. The rising-spectrum continuum source C3 and the shell-like continuum source C9 are both associated with maser emission. The masers located in the vicinity of C3 (M10, M12, M32, M47, M49, M52, M53) are clustered around the continuum source, near a velocity of 40 km s⁻¹. We further discuss possibilities for the nature of this region, which is also associated with a peak in the millimeter dust continuum, in Section 6. The masers located in the vicinity of the shell source (M4, M20, M22, M14, M23), together with M6, M7, M27, M28, M30, and M33, form a nearly straight line at roughly constant declination across the cloud. This linear feature corresponds to the northeastern edge of the "bar" feature seen in NH₃ and discussed in Section 4. Two weaker masers (M44 and M54) are located near the top of the western edge of the shell.

The spatial distribution of the CH₃OH masers appears extremely similar to that of the dense gas traced by NH₃ (especially the NH₃ (3, 3) line) in GCM0.253+0.016. Like the morphology of the NH₃ lines in Figure 3 which are brightest in the southern part of the cloud, the majority of the brightest CH₃OH masers are also observed to be in the southern part of the cloud, and are associated with several bright, compact regions of NH₃. The CH₃OH masers also appear to be coincident with other prominent features traced in NH₃ (3, 3), such as the "C-arc" and, as previously mentioned, the "bar." A close correlation between CH₃OH and NH₃ emission, especially in the (3, 3) line of NH₃, has been previously noted for gas clouds in the GC (e.g., M–0.02–0.07, Sjouwerman et al. 2010). In other star-forming regions in the Galaxy, masers in the (3, 3) line of NH₃ have been observed to arise in the same region as collisionally excited CH₃OH masers (Mangum & Wootten 1994, e.g.,). While NH₃ (3, 3) masers have been suggested to exist in the CMZ cloud Sgr B2 (Martin-Pintado et al. 1999), in GCM0.253+0.016 all of the (3, 3) emission has brightness temperatures <100 K and so cannot be clearly attributed to masers. Finally, no CH₃OH emission is observed at the location of the H₂O maser identified by Lis et al. (1994).

In general, the maser candidates follow the same distribution as the masers: distributed throughout the cloud, with the majority lying in the southern half. There are also two CH₃OH maser candidates associated with the faint, 80 km s⁻¹ component in the south-east portion of the cloud (CM44 and CM78). A number of the candidate sources (e.g., CM8, CM9, CM14, CM15, CM16, CM21, and CM28) also trace out a crescent-like feature the center of the cloud corresponding to the NH₃"C-arc." The typical brightness temperatures of these sources are 200–300 K, and they tend to have somewhat broader than normal measured FWHMs: ∼4.5–5 km s⁻¹. In addition to this main peak, the properties of which are cataloged, many of these spectra also exhibit a weaker superposed component having brightness temperatures of 40–80 K, and linewidths of 6–10 km s⁻¹, which appears as a plateau or "wings" in the spectra of many of these candidates (e.g., CM16). This weak and apparently spatially extended CH₃OH emission appears to be primarily associated with an analogous "c"-shaped feature in the 3 mm ALMA dust continuum map of Rathborne et al. (2014b). The low brightness temperature of the extended emission and its close correspondence with the 3 mm continuum could indicate that this region contains "quasithermal" emission from gas sufficiently dense that the maser in this line is quenched, analogous to that seen in the 44 GHz line in Sgr B2 (Mehringer & Menten 1997). If emission in this region is quasithermal in nature, then it is likely to be seen as a maser in other CH₃OH transitions which quench at higher densities (e.g., 44, 84, or 96 GHz; Cragg et al. 1992; McEwen et al. 2014).

In the interstellar medium, CH₃OH and other "saturated" (hydrogen-rich) molecules are primarily believed to be formed on the surface of dust grains (Tielens & Hagen 1982; Charnley et al. 1992; Watanabe & Kouchi 2002), where CH₃OH and NH₃ are among the most abundant mantle species present relative to H₂O, as measured in both low and high-mass young stellar objects (YSOs) and cold cloud cores (Tielens & Allamandola 1987; Dartois et al. 1999; Pontoppidan et al. 2004; Gibb et al. 2004; Boogert et al. 2008; Öberg et al. 2011). Notably, for CH₃OH, the high abundances measured for maser sources are inconsistent with those predicted by gas-phase formation models (Menten et al. 1986; Hartquist et al. 1995). To get the CH₃OH off of the dust grains and into the gas phase in the observed large quantities then requires a mechanism to liberate the CH₃OH from the grain mantles. Proposed mechanisms include thermal desorption via heating from an embedded protostar, shocks, or cosmic rays (requiring grain temperatures ≳90 K Tielens 1995; Brown & Bolina 2007), or nonthermal desorption processes including photodesorption via far-UV photons from cosmic ray interactions (Prasad & Tarafdar 1983; D'Hendecourt et al. 1985; Öberg et al. 2009a), grain sputtering, wherein the ice mantles of grains are dislodged via collisions (often in shocks) with other grains, neutrals, ions, or cosmic rays (Johnson et al. 1991; Caselli et al. 1997), and finally exothermic chemical reactions on the grain surfaces Duley & Williams (1993), Roberts et al. (2007). In particular, the presence of molecules formed on grains in relatively cool and dense environments requires an efficient nonthermal desorption process (Willacy & Williams 1993; Roberts et al. 2007; Öberg et al. 2009b; Caselli et al. 2012). In the CMZ, the abundance of 36 GHz CH₃OH sources has been suggested to be due to photodesorption from cosmic rays in this region Yusef-Zadeh et al. (2013a). We reconsider this in the light of our new, high-resolution observations.

Our observations reveal that locations of CH₃OH emission are in general an excellent match to the 3 mm ALMA dust continuum map shown in Rathborne et al. (2014b). However, the observed CH₃OH masers are stronger and more numerous in the southern part of the cloud, while the stronger dust emission is found in the northern portion of the cloud, north of declination −28:42:34.2. If high column densities of CH₃OH simply originate from high column densities of dust, and if all of the excitation conditions are uniform, one might expect more masers in the northern parts of the cloud (we also note that no CH₃OH emission—thermal or maser—is detected toward the two strongest millimeter continuum peaks identified in Rathborne et al. 2014b). One possible explanation might be that, for much of the northern part of GCM0.253+0.016, the CH₃OH emission is quasithermal, and the masers are quenched in those regions which correspond to not just high column densities but high volume densities. With future observations, it should be possible to test this with observations of more highly excited masers (e.g., 44 GHz) which are quenched at higher densities (Cragg et al. 1992; Mehringer & Menten 1997; McEwen et al. 2014).

Another possibility is that the differences in the distribution and strengths of the masers are a result of variations in the geometry and kinematics of the cloud. Maser emission requires a velocity-coherent path length of gas for amplification of the emission. Class I CH₃OH masers are, for example, rarely seen in the high-velocity components of outflows, which is suggested to be because the longest gain paths are found perpendicular to the outflow, at velocities near the systemic values (Menten 1991, although as noted by Voronkov et al. 2014 this is also partially a selection effect). This makes it somewhat surprising to see a large quantity of masers in an extremely turbulent environment like that of GCM0.253+0.016. However, the vast majority of the detected masers are relatively weak (having intensities <5 Jy) which could be a result of the short coherent path lengths in this gas. The stronger masers observed in the southern parts of the cloud could be due to a geometrical effect, larger gain lengths can be had perpendicular to the motion of a shock front (Kaufman & Neufeld 1996), so if these masers trace a shock propagating in the plane of the sky, it could explain their enhanced intensity. However, are shocks really the mechanism responsible for generating these masers?

Prior observations would seem to be able to rule out thermal desorption processes for clouds in the CMZ like GCM0.253+0.016 which lack advanced stages of star formation. Measured dust temperatures in GCM0.253+0.016 are <30 K (Molinari et al. 2011; Longmore et al. 2012). Dust heating via cosmic rays should be relatively uniform (though it may be more efficient toward the edges of the cloud), and is further not predicted by models to yield dust temperatures above >40 K in GCM0.253+0.016 (Clark et al. 2013), so thermal desorption via cosmic ray heating can be ruled out. Heating via shocks or embedded sources might lead to discrete regions of higher dust temperatures, however the dust temperature maps of Longmore et al. (2012) show no signs of temperature variation that might indicate unresolved regions of higher temperatures. Although there is one location, C3, where the observed CH₃OH masers cluster around a bright radio continuum source, the masers generally do not correlate with the radio continuum, making the heating of dust from embedded sources an unlikely source for the liberated CH₃OH more globally observed in the widespread masers in this cloud. In general, the distribution and relatively low intensities of the continuum emission do not suggest that there are embedded, ionizing sources within this cloud (see below).

Among the previously listed nonthermal desorption processes, those most likely to be important in the unique environment of dense clouds in the CMZ are then sputtering from shocks and UV photodesorption due to cosmic rays. There is evidence in the CMZ for enhanced rates of both of these processes: both a cosmic ray ionization rate several orders of magnitude greater than the local value in the solar neighborhood (ζ_GC = 10⁻¹⁵–10⁻¹³ s⁻¹; Dalgarno 2006; Goto et al. 2013; Yusef-Zadeh et al. 2013b, 2013c; Harada et al. 2014, but see also van der Tak et al. 2006 who find ζ_GC ∼ 10⁻¹⁶ s⁻¹ in Sgr B2) and strong, widespread shocks (e.g., Martin-Pintado et al. 1997; Martín-Pintado et al. 2001; Rodríguez-Fernández et al. 2004; Mills & Morris 2013, although see also Yusef-Zadeh et al. 2013b, who suggest that a high SiO abundance in CMZ clouds could also be a consequence of a high cosmic ray ionization rate).

For CH₃OH to form via sputtering from shocks requires shocks of velocities sufficient to disrupt grain mantles; as species like CH₃OH are relatively loosely bound to these mantles, velocities ≳10 km s⁻¹ in continuous or C-type shocks are suggested to be sufficient to enhance the observed CH₃OH (and NH₃) abundances, while shock velocities >15 km s⁻¹ will completely release the ice mantles into the gas phase (Caselli et al. 1997). For NH₃ and CH₃OH abundances to both be enhanced is consistent with our observations showing the morphologies of CH₃OH and NH₃ to be extremely similar on the scales probed here. Further, SMA observations by Johnston et al. (2014) illustrate the similar morphologies of CH₃OH and SiO, which suggests that the shocks may be yet stronger (shock velocities of 25–40 km s⁻¹ are suggested for CMZ clouds from the abundances of complex molecules and models for their heating Martín-Pintado et al. 2001; Rodríguez-Fernández et al. 2004).

The question for GCM0.253+0.016 then becomes identifying the origin of these shocks. We suggest that shocks due entirely to protostellar outflows are unlikely, given the observed lack of 6 GHz radiatively excited (Class II) CH₃OH masers in this cloud (Caswell 1996; Caswell et al. 2010), which are typically found to be associated with regions of ongoing massive star formation (Voronkov et al. 2010). Elsewhere in the Galaxy, Class I masers are observed to be clustered around Class II masers (Slysh et al. 1994; Val'tts et al. 2000; Ellingsen 2005; Voronkov et al. 2014); the lack of Class II masers in GCM0.253+0.016 suggests that a different mechanism is responsible for the large CH₃OH abundances implied by the Class I masers here. In lieu of protostellar outflows, cloud–cloud collisions have been posited to enhance CH₃OH abundances in other CMZ clouds leading to Class I CH₃OH masers observed in Sgr A and Sgr B2 (Mehringer & Menten 1997; Sjouwerman et al. 2010). A cloud collision model has been independently suggested for GCM0.253+0.016 by Higuchi et al. (2014). However, more recent analyses favor a collapse model for the observed kinematics and morphology of the cloud instead (Rathborne et al. 2014a, 2015; J. M. D. Kruijssen et al. 2015, in preparation). Another possibility that should be investigated for this cloud is an interaction with a supernova remnant: the GCM0.253+0.016 cloud overlaps in projection with a suggested supernova remnant identified by Kassim & Frail (1996). However, it has not yet been demonstrated that this supernova remnant is indeed located at the Galactic center.

An alternative to a model of shock-enhanced CH₃OH abundances in the CMZ is a high cosmic ray ionization rate. In brief: cosmic rays impact the dense gas, exciting the Lyman and Werner bands of H₂, generating a weak far-UV field in the cloud interior (Prasad & Tarafdar 1983). The UV photons are absorbed by molecules in the mantle exterior, which undergo photochemistry (photodissociation, diffusion and recombination), with some of the reaction products with excess energy being desorbed. Yusef-Zadeh et al. (2013a) favor this mechanism over cloud–cloud shocks for the generation of CH₃OH, as they assert that the effects of such shocks are limited to the cloud surfaces, although they do not consider the effects of smaller-scale turbulent shocks in cloud interiors. Given that the distribution of CH₃OH masers we observe is clustered and predominantly located in the southern part of the cloud, this might suggest that the masers are indeed limited to locations of large-scale and possibly even surface shocks, rather than more uniformly distributed in the cloud interior (though, as noted above, the apparently nonuniform distribution of CH₃OH masers could be an effect of density and/or orientation, and surface shocks could also result in a distribution of masers that, in projection, appears roughly uniform over the entire cloud).

Ultimately, both UV photodesorption from cosmic rays and grain mantle sputtering via shocks appear to be viable mechanisms for generating the observed CH₃OH abundances in GCM0.253+0.016. While our observations of stronger and more numerous masers in the southern part of GCM0.253+0.016 might slightly favor shocks as the primary mechanism for the abundant CH₃OH required to generate these masers, uncertainties as to the role gas density plays in quenching the 36 GHz maser in this cloud make it currently impossible to definitively determine the mechanism responsible. Various future observations could help to distinguish between shocks and cosmic rays as a mechanism for generating the observed CH₃OH abundances in the CMZ. First, as previously mentioned, observing other class I CH₃OH masers in CMZ clouds (e.g., the 44, 85, and 96 GHz masers) would aid in determining whether all class I maser transitions are stronger and more abundant in the southern part of GCM0.253+0.016, or if varying density in the cloud selectively quenches the 36 GHz masers. As these are the first observations of extremely abundant 36 GHz maser emission in a giant molecular cloud, it would also be valuable to search for similar emission outside of the CMZ, to determine whether this is truly a phenomenon unique to the inner few hundred parsecs of the Galaxy. While molecular clouds interacting with supernova remnants are not ideal targets as they could be expected to experience both enhanced shocks and cosmic ray ionization, other turbulent clouds lacking star formation should be searched on large-scales for class I CH₃OH maser emission. Finally, identifying the heating source for the molecular gas in the CMZ will likely also shed some light on the mechanism responsible for generating the observed abundances of CH₃OH and other molecules which are the product of grain-surface chemistry. As both cosmic rays and shocks are suggested to be the most likely source of cloud heating in the CMZ (e.g., Ao et al. 2013), it may be that the mechanism responsible for the heating drives not only the excitation of molecules in this region, but their chemistry as well.

6.2.1. The Ionization Source of the Extended Emission

Much of the continuum emission in GCM0.253+0.016 (e.g., regions C2, C7, and C9) is extended on scales of tens of arc seconds (1–2 pc at the Galactic center), forming rough arcs and filaments. We have argued, based on its roughly flat spectral index in several representative regions and its morphological similarity to the continuum emission seen at 90 GHz, that this emission is likely thermal, due to free–free radiation. Assuming this to be the case, the ionization source of this radiation needs to be determined.

Comparing the spatial distribution of the radio continuum to that of the molecular gas traced by the NH₃ (3, 3) line in Figure 9, it is clear that the continuum emission is primarily outlining the eastern edge of the cloud. The spatially extended regions C2, C7, and C9 all lie to the east of the peak of the molecular gas emission, roughly paralleling structures seen in the NH₃ line. In particular, C2 exactly parallels a similarly linear NH₃ feature 5'' to the west, while the regions C4, C7, and C8 all seem to trace the outer extent of the "C-arc" feature identified in NH₃ maps, with a similar spatial offset between the continuum and molecular line emission. C1 may also be a more northern extension of the same structure seen in C2. All together, this structure suggests that the extended emission in GCM0.253+0.016 is primarily at the cloud's surface, and that its ionization source is external to the cloud and to the east, perhaps a nearby O or B star. There is a known O4-6 supergiant located ∼11 pc away in projection to the east of the cloud, at R.A. = 17^h46^m282, decl. = −28°39'20'' (Mauerhan et al. 2010). Its bolometric luminosity is estimated to be 10⁶ ${{L}_{\odot }}$. Using the parameters of Martins et al. (2005), a supergiant with this luminosity would have a Lyman continuum flux of log Q₀ ∼ 49.8. The sum of the Lyman continuum fluxes for the regions in GCM0.253+0.016 that are cataloged in Table 3 is log Q₀ ∼ 47.8. Assuming that the cloud intercepts a fraction of the Lyman continuum photons from this sources proportional to its surface area over the distance to the source squared, and that the surface area of GCM0.253+0.016 as seen by this source is the same as we see in projection (20.5 pc²) then for the cloud to intercept ∼1% of the Lyman continuum photons, the ionizing source would have to be at a distance of ∼13 pc, which is consistent with its projected distance, making this a plausible source for ionizing the exterior of GCM0.253+0.016.

Zoom In Zoom Out Reset image size

While the filament in C2 appears to have a parallel molecular counterpart, there do not appear to be any molecular counterparts to the linear continuum filament seen in the south of the cloud, next to the shell-shaped C9. This filament is also seen in the 90 GHz maps of Rathborne et al. (2014b). It is then not clear whether this is also an externally ionized surface feature. A comparison of the continuum emission to the enigmatic HCO⁺ absorption filaments of Bally et al. (2014) does not show any correspondence: none of the HCO⁺ filaments have a counterpart in either radio continuum or NH₃ emission. Neither do any of the continuum filaments appear to be oriented parallel to nearby HCO⁺ filaments. Only the compact continuum source C4 appears to lie along the broad-line absorption filaments, near the junction of filaments 1 and 2, and could be just a chance superposition. However, the NH₃ emission just to the west of the bottom of the "C-arc" (at R.A. = 17^h46^m08^s, decl. = −28°43'30'') does have a steep drop in its brightness, tracing a sharp, nearly linear edge. This edge corresponds to the "NLA 3" cluster of narrow-line HCO⁺ absorption features identified by Bally et al. (2014), and it is possible that this edge could represent an analogous absorption feature. However, without the addition of single-dish data to provide a reliable flux zeropoint, it is not clear whether this edge seen in the NH₃ emission is actually an absorption feature, or simply the absence of emission.

6.2.2. Evidence for Ongoing Star Formation