THE CONNECTING MOLECULAR RIDGE IN THE GALACTIC CENTER

Figure 2 shows the integrated intensity (moment 0) maps of the CS(J = 1–0) (Tsuboi et al. 1999), CS(J = 2–1) (Paper II), CS(J = 4–3), and CS(J = 5–4) lines integrated within ±160 km s⁻¹. These maps were smoothed to the same resolution of 40'' for comparisons. Overall, the higher-J CS(J = 4–3) and CS(J = 5–4) lines are less sensitive to the diffuse emission as shown in the lower-J CS(J = 1–0), HCN(J = 1–0), ${\mathrm{HCO}}^{+}$(J = 1–0) (Tsuboi et al. 1999, 2011), and CS(J = 2–1) line (Paper II). The CS(J = 4–3) and CS(J = 5–4) lines are dominated instead by the more compact ridges and dense clouds. The major features of the CND, the 20 km s⁻¹ cloud, the 50 km s⁻¹ cloud, and the high velocity compact cloud (HVCC; CO 0.02-0.02) (Oka et al. 1999) are generally consistent in all of the lines.

Zoom In Zoom Out Reset image size

In contrast to these well-known structures, we found that there is a ridge-like structure shown in the CS(J = 4–3) line map. At the position of this ridge-like structure, compact clouds can be seen in the CS(J = 1–0) and CS(J = 2–1) lines, immersed in the background diffuse emission. In Figure 3, we show the larger scale map of the PA, as measured in the CS(J = 1–0) line, and the central region as measured in the CS(J = 4–3) line. We also overlay the 20 cm continuum emission on these maps. We find that the CO 0.02-0.02 cloud and the ridge-like structure may form a complete "northern molecular loop," which we have highlighted with the yellow dashed ellipse. The northern molecular loop structure may enclose the northern radio halo. A similar structure can also be seen in the low-J CS(J = 2–1) and CS(J = 1–0) lines, but the CS(J = 4–3) line exhibits clearer features.

Zoom In Zoom Out Reset image size

In Figure 4, we show the CS(J = 4–3) line channel maps at the velocity resolution of 10 km s⁻¹. The purpose of showing the channel maps is to demonstrate the kinematics of the individual cloud structures. The coherent kinematics often help to define distinct structures. The CND appears as a high velocity ($-110$ to 110 km s⁻¹) feature (the green ellipse) centered on SgrA* (green cross). The well-studied 20 and 50 km s⁻¹ clouds show a velocity gradient from −30 to 80 km s⁻¹, which is rotating with respect to SgrA* (e.g., Liu et al. 2012; Martín et al. 2012). The CO 0.02-0.02 cloud (Oka et al. 1999, 2008) is seen from 80 to 110 km s⁻¹. The ridge-like structure is detected roughly from 30 to 90 km s⁻¹ (labeled as green stars). Consistent with Figure 1, this structure was only detected with individual clumps in the CS(J = 1–0), the CS(J = 2–1), and the CS(J = 5–4) images. Here we call this ridge-like structure the "connecting ridge" (CR). The width of the CR is $\sim 2^{\prime}$ (4.6 pc). The CR seems not to have been detected before. We searched through the literature. The large-scale survey of ¹³CO (Bally et al. 1988), ¹²CO (Oka et al. 1998, 2001), and NH₃ (Purcell et al. 2012) do not show the CR in their maps. The NH₃ line map of the entire 200 pc TR made with ATCA (Ott et al. 2014) suffers from the missing flux, and there is no detection for the CR. The CR is also absent in the Mopra molecular line maps of the CMZ (Jones et al. 2012, 2013). This is probably because of the lower sensitivity and resolution for the past observations. It is also possible that it is the excitation effect. We will discuss this in Section 4. In the CS(J = 4–3) channel map, the CR seems to extend to the north of the Galactic plane and coincides with the base of the PA (labeled as the green solid squares). Our CS(J = 4–3) line observations did not cover the PA. Hence we show in Figure 5 the channel maps of the CS(J = 1–0) line in the same velocity range of 30–110 km ⁻¹. At the velocity of 30 km s⁻¹, the CR seems to be near the CND and appears to connect to the Galactic plane. We have called this part of the structure in the Galactic plane the "disk ridge" (DR; labeled by the green solid-triangles).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In Figure 6 we show the channel maps of the CS(J = 5–4) line smoothed to the same angular resolution of 38'' for the CS(J = 4–3) line. The CS(J = 5–4) line is generally consistent with the CS(J = 4–3) line, but traces more compact structures. However, this is likely due to a sensitivity that is a factor of 2  in the CS(J = 5–4) line. For the CR, we only detect some faint clumps from 40 to 80 km s⁻¹. We also detected the base of the PA similar to Figure 5.

Zoom In Zoom Out Reset image size

In Figures 7–9 we show the position–velocity diagrams (pv-diagrams) of the DR, the CR, and the PA. We show these diagrams in order to demonstrate the coherent kinematics within each of the structures, and with respect to each other. In Figure 8 we merge the pv-diagrams of the DR, the CR, and the PA in order to see the continuity of the ridges. Our CS(J = 4–3) line data does not cover the PA region, so we also show the pv-diagrams of the CS(J = 1–0) data along the PA for comparison. As indicated in the CS(J = 4–3) line channel maps (Figure 4), the CR appears from ∼50 km s⁻¹ and smoothly connects to the PA at velocity $\geqslant 80$ km s⁻¹. Although the CR becomes fainter in the CS(J = 1–0) line, Figure 8 suggests that the DR, the CR, and the base of the PA have smooth velocity transitions along the structure. We emphasize that this continuous nature of the kinematics is the principal argument in favor of a coherent structure.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In Figure 9 we found that the PA seems to show expanding features. In Paper II, we measured the whole PA with expanding velocity of ∼74 km s⁻¹. The detailed discussion and results are in Paper II. In the DR, there are several clumps that show broadened FWZI line widths of ∼90 km s⁻¹, which might indicate that they are physically near the CND. There are also some clumps in the CR that have broad line widths (FWZI ∼50 km s⁻¹). The line widths are about a factor of two larger than the values in the rest of the region, and can be seen in the other CS lines. This is best seen in the velocity dispersion maps presented below.

From the channel maps and the pv-diagrams, we found that the velocity of the CR is roughly from 30 to 90 km s⁻¹. We made the integrated intensity map and the intensity-weighted mean velocity map of the CS(J = 4–3) line from 30 to 90 km s⁻¹ (Figure 10). The spectra of the DR, CR, and the base of the PA are shown from Figures 11 to 12. The spectra of the CR also suggest that the FWZI velocity range is from 30 to 90 km s⁻¹. The velocity map (Figure 10(b)) shows two kinematic systems: (1) along the 20 km s⁻¹ cloud to the 50 km s⁻¹ cloud, and (2) from the DR to the CR. The velocity patterns are different and this kinematical behavior supports the argument that the DR is a distinct feature within the Galactic plane.

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

In Figure 14, we show the intensity-weighted velocity dispersion map of the CS(J = 4–3) line. The velocity dispersions of the DR, the CND, the CR, the CO 0.02-0.02 cloud, and near the PA, are larger than the surrounding material by about a factor of two or higher. The gas with high velocity dispersions also surrounds the 20 cm radio halo, which is similar to the northern molecular loop structure seen in Figure 2. We note that the gases of the 20 km s⁻¹ cloud and the 50 km s⁻¹ cloud also have higher velocity dispersions near the radio halo. By inspecting the spectra and the channel maps of the data, we find that the line widths of the 20 and 50 km s⁻¹ clouds are mainly intrinsically broad, and the northern molecular loop consists of multiple distinct components at the velocities of $\sim -12.5$ and 47.5 km s⁻¹. We emphasize that it is the enhanced velocity dispersions, together with the continuous kinematics, which indicate that the CR, DR, and PA could be a single coherent feature.

Zoom In Zoom Out Reset image size

In Figure 15 we show the CS(J = 4–3)/CS(J = 2–1), CS(J = 4–3)/CS(J = 1–0), CS(J = 5–4)/CS(J = 4–3), and CS(J = 5–4)/CS(J = 1–0) intensity ratio maps. We integrated the maps from $-160$ to $+160$ km s⁻¹ with a convolved beam of 40''. We utilized data with significance $\geqslant 3\sigma .$ Figure 15 represents the average intensity ratios because multiple components have been integrated along the line of light. The distributions of the CS(J = 4–3)/CS(J = 2–1) and CS(J = 4–3)/CS(J = 1–0) ratios are similar. The CR and the base of the PA have higher ratios than the ambient gas by a factor of ∼2 in the CS(J = 4–3)/CS(J = 2–1) ratio map, and the CS(J = 4–3)/CS(J = 1–0) ratio map shows higher contrasts than the CS(J = 4–3)/CS(J = 2–1) ratio map. The CND has higher ratios than the 50 and 20 km s⁻¹ clouds, and is comparable to the CR in the CS(J = 4–3)/CS(J = 1–0), CS(J = 4–3)/CS(J = 2–1), and CS(J = 5–4)/CS(J = 1–0) ratio maps. We emphasize that the enhanced line ratios are further evidence in support of the CR, DR, and PA as a coherent structure. The high-J line ratios of CS(J = 5–4)/CS(J = 4–3) are higher in the 50 and 20 km s⁻¹ clouds $\geqslant 1.5$ than in the CND ∼1.2. We will discuss the difference of the ratio maps in Section 4.

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

3.6.1. Rotational Diagrams

In Figure 16, we show the rotational diagrams of the major components mentioned above for the CS(J = 5–4), CS(J = 4–3), CS(J = 2–1), and CS(J = 1–0) lines. The usefulness of such a diagram depends on the assumption of thermal equilibrium. We convolved all of the CS line maps to the same angular resolution of 40'' and a velocity resolution of 5 km s⁻¹. For the CR, we measured the rotational temperature of the broad line clump shown in Figure 12. We also present the rotational diagrams of the DR and the base of the PA. For these three regions, the fluxes used for the rotational diagrams are averaged from the spectra shown in Figures 11–13. For the CR, we measured the flux from 30 to 90 km s⁻¹. The fluxes used in the rotational diagrams are presented from Tables 1 to 3. We could estimate the rotational temperature (${T}_{\mathrm{rot}}$) and column densities (${N}_{\mathrm{mol}}$) of the CS lines by assuming optically thin and local thermodynamic equilibrium (LTE). The column density in the upper level (N_u) is expressed as

$$\begin{eqnarray}&&{N}_{{\rm{u}}}=\displaystyle \frac{8\pi {k}_{{\rm{B}}}{\nu }^{2}\int {T}_{{\rm{b}}}{dv}}{{{hc}}^{3}{A}_{\mathrm{ul}}},\end{eqnarray} \tag{ 1 }$$

where ${k}_{{\rm{B}}}$ and h are the Boltzmann and Planck constants, respectively. A_ul is an Einstein A coefficient from upper (u) to lower state (l), ν is the frequency of the line, c is the speed of light, and ${T}_{{\rm{b}}}$ is the brightness temperature of the line. In the LTE condition, N_u is written as

$$\begin{eqnarray}&&{N}_{{\rm{u}}}=\displaystyle \frac{{N}_{\mathrm{total}}}{Q({T}_{\mathrm{ex}})}{g}_{{\rm{u}}}\mathrm{exp}\left(-\displaystyle \frac{{E}_{{\rm{u}}}\;(\mathrm{erg})}{{k}_{{\rm{B}}}{T}_{\mathrm{ex}}}\right),\end{eqnarray} \tag{ 2 }$$

where ${N}_{\mathrm{total}}$ is the total column density of a given molecule, Q(T) is a partition function, E_u is an energy at level u above the ground state, and ${g}_{{\rm{u}}}$ is the statistical weight. Hence we could rewrite the equation as

$$\begin{eqnarray}&&\mathrm{ln}\left(\displaystyle \frac{{N}_{{\rm{u}}}}{{N}_{\mathrm{total}}}\right)=-\displaystyle \frac{1}{{T}_{\mathrm{rot}}}{E}_{{\rm{u}}}({\rm{K}})+\mathrm{ln}\left(\displaystyle \frac{{g}_{{\rm{u}}}}{Q({T}_{\mathrm{rot}})}\right).\end{eqnarray} \tag{ 3 }$$

Therefore, we could determine the total column density and the excitation temperature by fitting a straight line for the data, plotted in logarithmic form, with Equation (3). The ${E}_{{\rm{u}}}$(K) is ${E}_{{\rm{u}}}$(erg)/k_B for multiple lines with different ${E}_{{\rm{u}}}$. This excitation temperature is called the "rotational temperature" (${T}_{\mathrm{rot}}$) because we measured the rotational transitions of CS molecules. The derived ${T}_{\mathrm{rot}}$ is also a lower limit to the kinetic temperature (${T}_{\mathrm{kin}}$), since the gas may not be thermalized. The detailed methods of the rotational diagram were described in Goldsmith & Langer (1999). In Figure 16, we found that the rotational diagrams of all the features could not be described by a single component. The CS(J = 5–4) and CS(J = 4–3) lines define a flatter slope than the CS(J = 1–0) and CS(J = 2–1) lines. This suggests that the molecular gas is a mixture of "high-temperature" and "low-temperature" components along the line of sight. If we assume that the high-J lines (J ≥ 4) represent warmer components, then it indicates that the ${T}_{\mathrm{rot}}$ of the warm gas components are higher than the cold components by a factor of more than three. Our results on the presence of multiple-temperature components are consistent with the NH₃ line observations (Herrnstein & Ho 2005). The fitted results are presented from Tables 1 to 3. The ${T}_{\mathrm{rot}}$ are similar for the DR, the CR, and the base of the PA. However, the base of the PA is very close to the CR and the ${T}_{\mathrm{rot}}$ differences might not be discernible. Also the ${T}_{\mathrm{rot}}$ (High) of the DR is almost twice higher than the CR and the PA. This indicates that the DR is closer to the CND. We note that the fitted rotational temperatures for the CS lines are low relative to the expectations from the NH₃ observations. This is an immediate hint that the assumption of LTE is probably not correct for the CS lines. This means that the density may not be high enough and will have to be taken into account. We discuss this further below. However, regardless of the actual density and temperatures of the gas, we emphasize that the different CS line ratios help us to isolate different cloud components.

Table 1.  The Disk Ridge

Line	CS(J = 1–0)	CS(J = 2–1)	CS(J = 4–3)	CS(J = 5–4)
$\int {T}_{{\rm{b}}}{dv}$ a	69.9 ± 14.1	85.3 ± 11.8	66.1 ± 11.0	52.3 ± 15.4
FWHM dv (km s−1)	17	15	18	13
Rotational Diagrams Method
${N}_{\mathrm{CS}}$ (cm−2)		(4.2 ± 0.6) × 1014
${T}_{\mathrm{rot}}$ (K)		10.0 ± 0.8
${T}_{\mathrm{rot}}$(Low) (K)		4.0 ± 0.8
${T}_{\mathrm{rot}}$(High) (K)		17.0 ± 8.6
Statistical Equilibrium Modeling Results
Minimum ${\chi }^{2}$ b		0.6
${N}_{\mathrm{CS}}$ (cm−2)		4.0 × 1015
${T}_{{\rm{K}}}$ (K)		18.2
$\mathrm{log}({n}_{{{\rm{H}}}_{2}})$ density (cm−2)		5.9
Opacity	0.7	2.8	5.3	4.5
Excitation temperature (K)	20.3	17.1	14.8	12.6

Notes.

^aThe integrated flux is  an average over the nine points shown in Figure 11, where one point is $\sim 40^{\prime\prime} .$ ^bThe degree of freedom of our parameters is 3 and the derived ${\chi }^{2}$ corresponds to a probability of 0.9.

Download table as:  ASCIITypeset image

Table 2.  The Connecting Ridge

Line	CS(J = 1–0)	CS(J = 2–1)	CS(J = 4–3)	CS(J = 5–4)
$\int {T}_{{\rm{b}}}{dv}$ a	32.4 ± 15.6	49.6 ± 19.3	41.9 ± 7.7	20.6 ± 10.9
FWHM dv (km s−1)	17	15	18	13
Rotational Diagrams Method
${N}_{\mathrm{CS}}$ (cm−2)		(2.4 ± 0.9) × 1014
${T}_{\mathrm{rot}}$ (K)		9.0 ± 1.5
${T}_{\mathrm{rot}}$(Low) (K)		5.0 ± 3.2
${T}_{\mathrm{rot}}$(High) (K)		10.0 ± 4.9
Statistical Equilibrium Modeling Results
Minimum ${\chi }^{2}$ b		0.39
${N}_{\mathrm{CS}}$ (cm−2)		2.0 × 1015
${T}_{{\rm{K}}}$ (K)		18.2
$\mathrm{log}({n}_{{{\rm{H}}}_{2}})$ density (cm−2)		5.7
Opacity	0.32	1.56	3.0	2.17
Excitation temperature (K)	24.0	16.0	12.0	9.4

Notes.

^aThe integrated flux is an average over the nine points shown in Figure 12, where one point is $\sim 40^{\prime\prime} .$ ^bThe degree of freedom of our parameters is 3 and the derived ${\chi }^{2}$ corresponds to a probability of 0.9.

Download table as:  ASCIITypeset image

Table 3.  The Base of the Polar Arc

Line	CS(J = 1–0)	CS(J = 2–1)	CS(J = 4–3)	CS(J = 5–4)
$\int {T}_{{\rm{b}}}{dv}$ a	52.4 ± 11.9	63.8 ± 12.4	54.2 ± 9.7	25.6 ± 10.5
Rotational Diagrams Method
${N}_{\mathrm{CS}}$ (cm−2)		(3.4 ± 0.6) × 1014
${T}_{\mathrm{rot}}$ (K)		9.0 ± 0.8
${T}_{\mathrm{rot}}$(Low) (K)		4.0 ± 1.0
${T}_{\mathrm{rot}}$(High) (K)		10.0 ± 3.7
Statistical Equilibrium Modeling Results
Minimum ${\chi }^{2}$ b		1.5
${N}_{\mathrm{CS}}$ (cm−2)		3.2 × 1015
${T}_{{\rm{K}}}$ (K)		18.2
$\mathrm{log}({n}_{{{\rm{H}}}_{2}})$ density (cm−2)		5.6
Opacity	0.61	2.78	5.19	3.89
Excitation temperature (K)	21.8	15.6	12.1	9.5

Notes.

^aThe integrated flux is average over the nine points shown in Figure 13, where one point is $\sim 40^{\prime\prime} .$ ^bThe degree of freedom of our parameters is 3 and the derived ${\chi }^{2}$ corresponds to a probability of 0.7.

Download table as:  ASCIITypeset image

3.6.2. Statistical Equilibrium

In order to derive the physical properties of molecular gas instead of the optically thin assumption, we used the one-dimensional (1D) radiative transfer code Radex (van der Tak et al. 2007) to perform the statistical equilibrium calculations. Radex uses the escape probability formalism (Goldreich & Kwan 1974) to solve the statistical equilibrium equations to model the observed line intensities for given physical conditions. The statistical equilibrium calculations account for the opacity effects as well as subthermal excitation, while the rotation diagram analysis relies on optically thin approximation and LTE conditions.

The Rayleigh–Jeans approximated brightness temperature (T_b (K) or ${\displaystyle \int }_{}^{}{T}_{{\rm{b}}}\;{dv}$ (K km s⁻¹) are calculated for each molecule with varying molecular column density, H₂ density of the main collision partner, and kinetic temperature. The observed main beam brightness temperature (T_mb) is diluted by the beam filling factor f,

$$\begin{eqnarray}&&f=\displaystyle \frac{{\theta }_{{\rm{s}}}}{{\theta }_{{\rm{s}}}+{\theta }_{{\rm{b}}}}=\displaystyle \frac{{T}_{{\rm{b}}}}{{T}_{\mathrm{mb}}},\end{eqnarray} \tag{ 4 }$$

where ${\theta }_{{\rm{s}}}$ is the source size and ${\theta }_{{\rm{b}}}$ is the beam size. The brightness temperature of the source will be diluted if the source size is smaller than the beam size. We adopted line ratios to constrain the statistical equilibrium while canceling beam filling factors. Comparison of the models and the observations was performed with the ${\chi }^{2}$-statistics for each set of parameters. We find the best-fit models by minimizing the ${\chi }^{2}$ value. In this paper, we use the uniform sphere geometry, single collision partner (H₂), and background temperature of 2.73 K. The collisional rate and spectroscopic data of the CS molecule were taken from the Leiden Atomic and Molecular Database (LAMDA; Schöier et al. 2005; Lique et al. 2006).

We construct a 50 × 50 × 50 grid of parameter space for kinetic temperature (${T}_{{\rm{K}}}$), molecular hydrogen density (${n}_{{{\rm{H}}}_{2}}$), and column density of a CS molecule (${N}_{\mathrm{CS}}$).

• 1.  
  ${T}_{{\rm{K}}}$ is from 10 to 650 K.
• 2.  
  ${n}_{{{\rm{H}}}_{2}}$ is from 10⁴ cm⁻³ to 5 × 10⁷ cm⁻³.
• 3.  
  ${N}_{\mathrm{CS}}$ is from 10¹² to 10¹⁶ cm⁻².

We then estimate the best-fitted ${T}_{{\rm{K}}},$ ${n}_{{{\rm{H}}}_{2}},$ and ${N}_{\mathrm{CS}}$ by comparing the ratios and intensities.

The input CS line fluxes are the same as the values used for the rotational diagrams method. From Figures 17 to 19 we show the ${\chi }^{2}$ contours versus ${T}_{{\rm{K}}}$ and ${n}_{{{\rm{H}}}_{2}}.$ From Tables 1 to 3 we show the derived parameters with the best-fitted ${\chi }^{2}.$ The ${T}_{\mathrm{rot}}$ and the ${N}_{\mathrm{CS}}$ derived in the rotational diagram method are lower than the values derived in the statistical equilibrium method. We also present the excitation temperature and the opacity calculated from the Radex code. The excitation temperature of the four CS lines are different. The opacity is optically thin in the CS(J = 1–0) line, and $\geqslant 1$ for the CS(J = 2–1), CS(J = 4–3), and CS(J = 5–4) lines. These might suggest that the CS lines are likely to be subthermal since the excitation temperature is lower than the kinetic temperature. From Figures 17 to 19 we show the CS J-ladder diagram of the data and the model. Our results show that the beam filling factors of this region are from 0.14 to 0.2, which suggest that sources are much smaller than the beam size. This might also explain the lower ${N}_{\mathrm{CS}}$ derived in the rotational diagrams, since the brightness temperatures used for the rotational diagrams are diluted by the beam.

We detected the CR in the new CS(J = 4–3) line map. This component also exists in the CS(J = 1–0), CS(J = 2–1), and CS(J = 5–4) lines, but is much more visible in the CS(J = 4–3) line map. The CS(J = 1–0), CS(J = 2–1), and CS(J = 5–4) lines only detected several compact clouds of the CR. The extended emission was only marginally detected in the CS(J = 1–0) and CS(J = 2–1), but not detected in the CS(J = 5–4) lines. For the CS(J = 5–4) line, insufficient sensitivity is likely the reason that the extended emission of the CR was not detected. The CS(J = 4–3) line intensity of the extended emission of the CR is ∼0.4 K, and the noise level of the CS(J = 5–4) line is 0.17 K (${T}_{\mathrm{mb}}$). Hence the noise will dominate in the CS(J = 5–4) line map if the CS(J = 5–4) and CS(J = 4–3) intensities are comparable. Indeed for the compact clouds detected in both of these lines, the intensities are comparable.

As in previous discussions, it is distinctly possible that the various transitions are sampling different regions with different excitations. We can then set some limits on the possible ranges of conditions by considering the available transitions. We check the excitation conditions under which the CS(J = 4–3) line is most prominent. If the gas is thermalized, i.e., if the density is sufficiently high so that all of the energy levels are in thermal equilibrium, the relative intensity of the different transitions is just a function of temperature. This is the basis of the analysis of Equation (3), and the utility of the rotational diagrams. If the gas is not thermalized, then the relative intensity ratios of the various transitions will be a function of excitation temperature, filling factor, frequency, and opacity. The excitation temperature is of course driven by the density (see van der Tak et al. 2007). The CS line ratios are sensitive to the density of the molecular gas because of its large dipole moment, and small frequency spacing between the transitions. Hence the CS molecule is a useful density tracer (van der Tak et al. 2007). Our statistical equilibrium calculations show that the physical conditions (${n}_{{{\rm{H}}}_{2}},$ ${N}_{\mathrm{CS}},$ and ${T}_{{\rm{K}}}$) of the DR, CR, and the base of the PA are similar. The CS(J = 1–0) line is optically thin, and the other lines are optically thick. The CS(J = 4–3) line has the highest opacity compared to the other lines. For the Rayleigh–Jeans approximation, the brightness temperature is proportional to $f({T}_{\mathrm{ex}}-{T}_{\mathrm{bg}})(1-{e}^{-\tau }),$ where ${T}_{\mathrm{ex}},$ ${T}_{\mathrm{bg}},$ τ, and f are the excitation temperature, background temperature, opacity, frequency, and filling factor of a transition. Considering the modeled parameters and the Rayleigh–Jeans approximation, the CS(J = 4–3) line is intrinsically brighter than the CS(J = 1–0) and CS(J = 5–4) lines. The brightness temperature of the CS(J = 2–1) line is comparable to the CS(J = 4–3) line. Our statistical equilibrium results suggest that for ${n}_{{{\rm{H}}}_{2}}={10}^{5.7-5.9}$ cm⁻², ${T}_{{\rm{K}}}=18\;{\rm{K}}$, and ${N}_{\mathrm{CS}}=(2-4)\times {10}^{15}$ cm⁻², the CS(J = 4–3) will be the brightest line of the CR.

Our resolved ratio maps (Figure 15) might suggest a variation of excitation conditions. As shown in the CS(J = 4–3)/CS(J = 2–1) ratio map (and similarly with CS(J = 4–3)/CS(J = 1–0)), the connecting ridge has higher line ratios for lines with larger energy differences. This indicates higher temperatures, and this is expected, as large energy gaps will be sensitive to higher temperatures. These higher ratios also suggest a higher opacity in the connecting ridge than in the ambient gas.

It is interesting to ask why the CS(J = 4–3)/CS(J = 2–1) and CS(J = 4–3)/CS(J = 1–0) ratios of the 20 and 50 km s⁻¹ clouds are not as significant as in the CS(J = 5–4) ratios. The CS(J = 4–3)/CS(J = 2–1) and CS(J = 4–3)/CS(J = 1–0) ratios show different distributions from the CS(J = 5–4)/CS(J = 1–0), and CS(J = 5–4)/CS(J = 4–3) ratio maps, where the 20 and 50 km s⁻¹ clouds show high ratios. We note that in Figure 16, the dense gas in the GC has structures of multiple-temperature components, where the CS(J = 1–0) and CS(J = 2–1) lines represent cooler components, and the CS(J = 4–3) and CS(J = 5–4) lines represent warmer components. If this implies that the gas is a mixture of different excitation conditions (temperature, density), then in the CS(J = 4–3)/CS(J = 2–1) ratios map, we are sampling the cooler part of the 20 and 50 km s⁻¹clouds, and the CS(J = 5–4) line traces better the warmer components in these clouds. Therefore, the variations of the ratios from high-J to low-J lines indicate the mixture of multiple phases of the molecular gas, and its sensitivity to different tracers. Our results are consistent with the previous results (e.g., Guesten & Henkel 1983; Guesten et al. 1985; Serabyn et al. 1989, 1992; Huettemeister et al. 1993; Herrnstein & Ho 2005).

In Figure 10 we found that there are two kinematic systems: (1) the disk rotation of the 20 and 50 km s⁻¹ clouds, and (2) the velocity gradient along the DR, CR, and the base of the PA (Figure 8). The velocity gradient along the DR, CR, and the base of the PA seems to suggest a coherent extraplanar structure, and the gradient is not aligned with the Galactic plane. Sofue (1996) studied the ¹³CO(J = 1–0) line data from the Bell Laboratory Survey (Bally et al. 1987) (beam = 100''). They pointed out that a pair of molecular spurs are associated with the Galactic center lobe (GCL; Sofue & Handa 1984; Bland-Hawthorn & Cohen 2003). The eastern and western spurs are receding and approaching at V_lsr of ∼100 km s⁻¹ and $\sim -150$ km s⁻¹, respectively. The size scale of the molecular spurs is ∼30 pc. In this paper we confirm that the location of the eastern molecular spur is coincident with the PA with consistent velocity.

In Figure 9, we found that the PA shows expanding motion. We estimate the energy of this feature. The kinetic energy (${E}_{{\rm{k}}}$) of the PA is,

$$\begin{eqnarray}&&{E}_{{\rm{k}}}=\displaystyle \frac{1}{2}{M}_{\mathrm{ridge}}{\rm{\Delta }}{V}^{2},\end{eqnarray} \tag{ 5 }$$

where the ${M}_{\mathrm{ridge}}$ is the molecular hydrogen mass of the ridge, and the ${\rm{\Delta }}V$ is the expanding velocity of 74 km s⁻¹ (paper II). We derive the gas mass (${M}_{{{\rm{H}}}_{2}}$) of the entire PA from the CS(J = 1–0) map. We adopted the same equation from Tsuboi et al. (1999), with ${T}_{\mathrm{ex}}$ of 20 K, and the CS abundance of 10⁻⁸ (Irvine et al. 1987). The corresponding ${M}_{{{\rm{H}}}_{2}}$ is $1.5\times {10}^{5}$ ${M}_{\odot }$ and the kinetic energy of the PA is from 8 × 10⁵¹ erg (Paper II). One of the possible origins of the energy input is from multiple supernova explosions, which are able to release the energy into the ISM with an efficiency up to 20% (Weaver et al. 1977; McCray & Kafatos 1987). Hence we need around 40 supernova explosions to push the gas (PA) away, or 8 supernova explosions for 100% efficiency.

In Figure 20, the 10 GHz continuum emission of the GCL (Handa et al. 1987) is shown as a dome-shaped, limb-brightened bipolar structure in the GC. It has a vertical extension of more than 200 pc. The GCL was detected in the radio continuum emission (Sofue & Handa 1984; Law et al. 2008a, 2008b), the dust emission at 8.3 μm (Bland-Hawthorn & Cohen 2003), and the radio recombination lines (Law et al. 2009). The GCL was proposed to be the entrainment of dust by the large-scale bipolar wind powered by the central starburst, which had taken place within the last 7 Myr with an energy ejection zone of 100 pc (Bland-Hawthorn & Cohen 2003). In Figure 20, the dust emission of the GCL shows the molecular counterparts of the eastern (positive velocity) and western spurs (negative velocity) detected by (Sofue 1996). The eastern spur is the PA shown in the CS(J = 1–0) line (Tsuboi et al. 1999) (Figure 20). The PA shows a vertical extension of ∼43 pc from the central disk. The preliminary idea for this paper is that the molecular gas that was originally in the disk was pushed out by the supernova explosions. Here is some of the indirect evidence for this idea: (1) the continuity of the velocities of the DR, CR, and PA. The CR, DR has a velocity gradient that is orthogonal to the Galactic disk; (2) the expanding motion of the PA; (3) similar physical conditions of the DR, CR, and PA, (4) the high dispersions surrounding the 20 cm radio halo. However, since we only have a morphological association of the GCL and the CR/PA, it is unclear whether the kinetic energy within the CR/PA was converted from the kinetic energy of stellar feedback. A large molecular line map of the entire GCL will provide better constraints. Detailed study of the expanding features in the GC is presented in Paper II. The other possibility is that the DR, CR, ad PA are formed by the magnetic buoyancy caused by the Parker instability (field strength of ∼150 μG; Fukui et al. 2006; Machida et al. 2009; Takahashi et al. 2009). We will discuss this the next paper (Paper II).

The ubiquitous presence of nuclear winds and outflows has recently been discovered in multiple areas, from nearby star-forming galaxies to the high-z galaxies (Veilleux et al. 2005; Chen et al. 2010; Murray et al. 2011; Ho et al. 2014; Rubin et al. 2014). In the case of the IC 342, it is one of the nearest galaxies at a distance of 3.5 Mpc (Wu et al. 2014). Schinnerer et al. (2008) suggested that the giant molecular associations (GMAs; gas mass: ${10}^{7}-{10}^{8}$M${}_{\odot }$) in its nuclear region are pushed out by the violent nuclear starburst. The mechanical energy of the stellar wind is able to push out the GMAs from 12 to 50 pc relative to the GC. On the large scale of kiloparsecs, the intensive studies of NGC 253 (Bolatto et al. 2013) and M82 (Ohyama et al. 2002) also showed that the central starburst activity is able to push large amounts of molecular gas away from disk. This disruption may prevent the gas from forming stars in the immediate future. The Galactic wind can play the role of suppressing star formation in galaxies.