On the Effects of UV Photons/X-Rays on the Chemistry of the Sgr B2 Cloud

With the IRAM 30 m telescope, we observed the HCO, HOC⁺, and CO⁺ lines (listed in Table 2) in 2014 December toward 13 selected positions (Program 137-14; PI: Armijos-Abendaño) whose offsets relative to the massive hot core Sgr B2M position (with α_J2000 = 17:47:20.40 and δ_J2000 = −28:23:07.25) are indicated in Table 1 and Figure 1. The Eight MIxer Receiver (EMIR) was used during our observations, tuning horizontal and vertical polarizations in the E090 and E230 bands, while the Fast Fourier Transform Spectrometer (FTS) and Wideband Line Multiple Autocorrelator (WILMA) were used as back ends. The FTS back end worked at a resolution of 200 kHz. We used position switching as the observing mode with the emission-free reference position α_J2000 = 17:46:23.0 and δ_J2000 = −28:16:37.3, which was selected from the maps of the CS molecule by Bally et al. (1987). An on-source integration time of 30 minutes was used for each position. Spectra were calibrated by using ambient and cold temperature loads, which provided a calibration accuracy of around 10%. The spatial resolution provided by the IRAM telescope, which allowed us to spatially separate PDRs, XDRs, and quiescent regions (see Figure 1), is listed in Table 2. The output spectra obtained from the IRAM 30 m telescope are calibrated in the antenna temperature scale (T${}_{{\rm{a}}}^{* }$), which we converted to the main-beam temperature (T_mb) using the relation T_mb = (F_eff/B_eff)T${}_{{\rm{a}}}^{* }$, where F_eff is the forward efficiency and B_eff is the main-beam efficiency. The B_eff values were calculated using the Ruze formula B_eff = 1.2 $\epsilon {e}^{-{(4\pi R\sigma /\lambda )}^{2}}$, where  = 0.69 and Rσ = 0.07. We calculated B_eff values of 0.78–0.45 for the observed frequencies of 85–268 GHz. The F_eff values were taken from the online documentation of the IRAM 30 m telescope.¹¹

The calibrated data observed in 2014 were imported into the MADCUBA software (Martín et al. 2019) for further processing. With this software, the baseline (of order 0–2) subtraction and spectrum averaging were applied. Then the spectra were smoothed to a velocity resolution of ∼5 km s⁻¹, appropriate to resolve the typical line widths observed in the GC. The rms noise level of the spectra is ∼10–15 mK in the T_mb scale.

In our study, we also used on-the-fly (OTF) data observed in 2006 August (Program 012–06; PI: Martín-Pintado). With this observing mode, we mapped a region of 15' × 14' within the Sgr B2 complex. The data were calibrated using the standard dual-load system. The A100 and B100 receivers connected to the 1 MHz filter bank were used in our observations, covering the frequencies of the SiO J = 2–1, C¹⁸O J = 1–0, H¹³CO⁺ J = 1–0, and HCO 1_0,1–0_0,0 (J = 3/2–1/2, F = 2–1) transitions. Our OTF observations of HCO 1_0,1–0_0,0 cover a region of 9' × 9', which is smaller than that mapped in the other molecular lines (see Figures 1 and 2). The data reduction was done with the GILDAS software,¹² which was also used to build the SiO J = 2–1, H¹³CO⁺ J = 1–0, and HCO 1_0,1–0_0,0 data cubes with half-power beams of ∼28''. The data cubes of SiO, H¹³CO⁺, C¹⁸O, and HCO have an rms noise level of ∼25 mK in the ${{\rm{T}}}_{{\rm{a}}}^{* }$ scale (∼30 mK in T_mb). Contours tracing the SiO J = 2–1 emission integrated in two velocity ranges are shown in Figure 1. Integrated intensity maps of the HCO 1_0,1–0_0,0 and H¹³CO⁺ J = 1–0 lines are shown in Figure 2. The C¹⁸O J = 1–0 spectra extracted over 22'' regions at selected positions will be used later to derive molecular abundances.

Zoom In Zoom Out Reset image size

We have used observations available in the XMM-Newton archive to study the X-ray emission from Sgr B2. The selected data were observed with the EPIC/MOS instruments in 2004 (ObsID: 0203930101; PI: Decourchelle) and 2012 (ObsID: 0694640601; PI: Terrier). The data taken in 2004 and 2012 have exposure times of 50,918 and 41,921 seconds, respectively. The emchain task of the Science Analysis Software (SAS; version 16.1.0) was used to run the default pipeline processing, as well as to obtain the calibrated event lists, which were cleaned for flaring events. Then, the evselect task of SAS was used to extract the two images in the range of 6.3–6.5 keV shown in Figure 1. This energy range is chosen with the aim of mapping the 6.4 keV Fe line emission and the underlying continuum from Sgr B2.

As mentioned above, Figure 2 displays maps of the HCO 1_0,1–0_0,0 (F = 2–1) and H¹³CO⁺ J = 1–0 lines integrated over the velocity range of 40–80 km s⁻¹. As seen in this figure, the HCO 1_0,1–0_0,0 line emission is extended and detected above 2σ toward positions 3 and 4 and partially in position 6. As mentioned in Section 2.1, the HCO data cube has an rms level of ∼30 mK. This rms value is not enough to detect the HCO 1_0,1–0_0,0 F = 2–1 line emission (above 2σ) toward positions 1, 2, 5, and 7–13 in our map because in Section 4.4 we notice that this HCO line has peak intensities ≲60 mK toward these 10 positions.

Figure 2 shows widespread H¹³CO⁺ J = 1–0 line emission above 3σ toward Sgr B2. Figure 1 displays widespread SiO J = 2–1 line emission toward Sgr B2 as well. As seen in Figures 1 and 2, the molecular emission toward Sgr B2 is extended over the beam of the 30 m IRAM telescope. Thus, in Sections 4.3–4.7, we consider a beam-filling factor of unity in the line modeling.

As seen in Figure 5, we have detected line emission ≳55 mK from the H42α radio recombination line (RRL) toward positions 1–3, 5, and 7–9 in Sgr B2. The H ii regions 1–2, 5, and 7–9 were selected from the 20 cm continuum emission peaks shown in Figure 1. Following the same method as in Section 3.4 from Armijos-Abendaño et al. (2018), we have derived electron densities of 280–550 cm⁻³ for positions 3, 5, and 7–9, as well as electron densities of 1200–1700 cm⁻³ for positions 1 and 2. We have considered the sources in positions 3, 5, and 7–9 as diffuse H ii regions, as their electron densities are at least a factor 2.2 lower than those of positions 1 and 2 considered as dense H ii regions. Position 9 has the highest electron density of 540 cm⁻³ among all positions tracing diffuse H ii gas.

There is no emission of the H42α RRL in positions 11–13, and it is very low in positions 4, 6, and 10, supporting the idea that these sources are quiescent and affected by shocks, and some of them are affected by X-rays (see Section 3). Toward these quiescent regions, we estimated electron densities lower than 180 cm⁻³. Figure 5 reveals that the ionized gas does not have a molecular counterpart in all velocity components in positions 7–9. As will be discussed below, the C¹⁸O gas has two velocity components at ∼54 and 90 km s⁻¹ in positions 8 and 9, while the ionized gas shows only low velocities within ∼55–70 km s⁻¹ in both positions (see Figure 5). In position 7, the C¹⁸O gas shows three velocity components at ∼64, 92, and 117 km s⁻¹, whereas the H42α RRL intensity peaks at 90–120 km s⁻¹.

Figure 6 shows the spectra of the HOC⁺ 1–0 and 3–2 transitions toward the 13 positions studied in this paper. The HOC⁺ 1–0 line is observed in absorption in position 1 and appears slightly absorbed in position 2. The HOC⁺ 3–2 line is blended with hyperfine lines of ³³SO₂(7_2,6–6_1,5). The SO₂ molecule is considered as a good tracer of hot core chemistry (Jiménez-Serra et al. 2007). As seen in Figure 6, the HOC⁺ J = 1–0 emission is detected in positions 3–13, whereas the HOC⁺ J = 3–2 emission above 3σ is detected only in positions 3, 4, 6, and 9.

Zoom In Zoom Out Reset image size

To derive HOC⁺ column densities (N_tot) and excitation temperatures, we have used the AUTOFIT tool of SLIM in the MADCUBA package (Martín et al. 2019). This tool allows us to fit synthetic spectra to the data assuming local thermodynamic equilibrium (LTE) conditions. The FWHM of the HOC⁺ lines was fixed in the line fitting (except for the low-velocity gas of position 9). Multiple velocity components were considered in the line fitting of positions 6–9, 11, and 13. The excitation temperature (T_ex) in the range from 4.6 to 9.9 K is found toward several positions studied in Sgr B2 (see Table 4). The T_ex was fixed to the average value of 6 K when the AUTOFIT tool did not converge.

We considered Sgr B2M and Sgr B2N as blackbody emitters with a size of 2'' and a temperature of 150 K (Belloche et al. 2013) to fit the absorption line of HOC⁺ J = 1–0 in positions 1 and 2, respectively. This continuum was also used in Rivilla et al. (2018) to study both sources. As mentioned above, the HOC⁺ 3–2 emission is blended with the ³³SO₂ emission. Belloche et al. (2013) explained the ³³SO₂ emission in Sgr B2M (our position 1) using two components. One considers a source size of 2'', a T_ex of 200 K, and a column density of 1.8 × 10¹⁷ cm⁻², while the other considers a source size of 60'', a T_ex of 50 K, and a column density of 5.4 × 10¹³ cm⁻². The observed HOC⁺ J = 3–2 line in position 1 was modeled using both the ³³SO₂ emission from the compact component estimated by Belloche et al. (2013) and the contribution of HOC⁺ J = 3–2. In this modeling, we have neglected the ³³SO₂ emission contribution from the extended component, as it represents ∼5% of the observed HOC⁺ J = 3–2 line. An additional component for HOC⁺ with a T_ex of 30 K and a column density of ∼1 × 10¹³ cm⁻² is needed to model the HOC⁺ J = 3–2 line in position 2, in which we have not taken into account a contribution from ³³SO₂, as this species has not been found in position 2 (Belloche et al. 2013). The derived parameters of HOC⁺ are listed in Table 4.

Figure 7 displays spectra at 1 and 3 mm covering four (F = 2–1, 1–0, 1–1, 0–1) and five (F = 4–3, 3–2, 3–2, 2–1, 2–2) hyperfine transitions of HCO, respectively. The HCO F = 2–1 and 1–0 transitions are detected in positions 1 and 2, whereas the HCO F = 1–1 and 0–1 transitions are blended with the strong absorption of the H¹³CO⁺ J = 1–0 transition in both positions. The emission from the HCO F = 2–1, 1–0, and 1–1 transitions is detected in positions 3–13, while that of the HCO F = 0–1 transition is not detected in all positions. The F = 2–1 transition of HCO shows peak intensities ≲60 mK toward positions 1, 2, 5, and 7–13 (see Figure 7). The emission from the 1 mm HCO transitions is not observed above 3σ in positions 3–13, except that of the HCO F = 4–3 hyperfine transition in position 3. Spectral features of other molecules are also seen in Figure 7 and labeled in the 1 and 3 mm spectral windows of positions 1 and 3.

Zoom In Zoom Out Reset image size

As in the case of the HOC⁺ molecule, the AUTOFIT tool was used to estimate HCO column densities and excitation temperatures assuming LTE conditions. The HCO lines in positions 1 and 2 are modeled by considering the same assumptions for the background sources as in the case of HOC⁺. The HCO shows an average T_ex of ∼5 K derived from seven positions of Sgr B2. A warmer component of HCO with a T_ex of 20 K and a column density of 2 × 10¹⁴ cm⁻² is needed to fit the low-intensity lines of HCO in positions 1 and 2. The T_ex, FWHM, and local standard of rest (LSR) velocity were fixed in modeling when the AUTOFIT tool did not converge in our modeling. Two velocity components are included in the line modeling of positions 6, 11, and 13. The LTE best fits to the HCO lines are shown in Figure 7, and the derived parameters are listed in Table 4. Upper limits to the column density are given in this table when there is undetected HCO emission at a given velocity.

The CO⁺ 5/2–3/2 and 3/2–3/2 spectra toward our studied positions of Sgr B2 are shown in Figure 8. Unfortunately, the CO⁺ 5/2–3/2 transition is blended with two hyperfine transitions of ¹³CH₃OH J = 5–4, while the CO⁺ 3/2–3/2 transition is blended with two transitions (12_11,1–11_10,2 and 12_11,2–11_10,1) of (CH₃)₂CO. The CO⁺ lines are detected in positions 1–3 and 9, although without ambiguity only in position 9. The ambiguity in positions 1–3 is due to the line blending (see below).

Zoom In Zoom Out Reset image size

We have used the AUTOFIT tool of MADCUBA to derive the CO⁺ column density. The line fitting of the CO⁺ lines in positions 1 and 2 is complex because of the line blending. The emission contribution from ¹³CH₃OH and (CH₃)₃CO to the CO⁺ lines cannot be ruled out in positions 1 and 2. For this reason, the lines of the three molecular species are fitted simultaneously with the AUTOFIT tool. The T_ex of 10 K for ¹³CH₃OH and (CH₃)₃CO gives the best fits to the observed lines, which yields values of ∼2 × 10¹⁵ and ∼2 × 10¹⁶ cm⁻² for ¹³CH₃OH and (CH₃)₃CO, respectively. Several parameters were fixed in the CO⁺ line fitting in all positions (see Table 4). The LTE best fits to the observed CO⁺ lines are given in Figure 8, where the synthetic spectrum of ¹³CH₃OH (and also of (CH₃)₂CO for positions 1 and 2) is shown for positions 1–4 and 6 to evidence the emission contribution of ¹³CH₃OH to the CO⁺ line. The synthetic spectrum of ¹³CH₃OH is obtained considering a T_ex of 10 K for the five positions. The estimated parameters are given in Table 4, where an upper limit to the column density is listed in the case of undetected CO⁺ lines.

The C¹⁸O J = 1–0 spectra toward the target positions of Sgr B2 are shown in Figure 9. The ground-state line of C¹⁸O is included in our work as a tracer of H₂, allowing us to derive relative abundances of HOC⁺, HCO, and CO⁺. The C¹⁸O J = 1–0 lines were fitted using MADCUBA and considering multiple velocity components in several positions (see Figure 9). Using C¹⁸O data, Martín et al. (2008) derived the average T_ex of 10 K for several positions of Sgr B2, which is considered in the C¹⁸O J = 1–0 line fitting shown in Figure 9. Thus, we have derived the C¹⁸O column density listed in Table 5. The H₂ column density (${{\rm{N}}}_{{{\rm{H}}}_{2}}$) is also given in this table and was calculated from the column density of C¹⁸O using the ¹⁶O/¹⁸O isotopic ratio of 250 (Wilson & Rood 1994) and the relative abundance of CO to H₂ of 10⁻⁴ (Frerking et al. 1982).

Zoom In Zoom Out Reset image size

Table 5.  Derived Parameters from the C¹⁸O J = 1–0 and HC¹⁸O⁺ J = 1–0 Lines toward 13 Positions of Sgr B2

Pos.	C18O				HC18O+
	VLSR	FWHM	Ntot	${{\rm{N}}}_{{{\rm{H}}}_{2}}$	Tex	VLSR	FWHM	Ntot
	(km s−1)	(km s−1)	(×1016 cm−2)	(×1022 cm−2)	(K)	(km s−1)	(km s−1)	(×1012 cm−2)
1	62.4 ± 0.5	18.6 ± 1.2	12.0 ± 0.8	31.0 ± 2.0	1a	66a	12a	10.0 ± 3.2
2	70.9 ± 0.5	21.0 ± 1.0	6.8 ± 0.3	17.0 ± 0.7	1a	65a	12a	7.1 ± 1.6
3	67.8 ± 0.5	18.8 ± 1.1	3.4 ± 0.2	8.5 ± 0.4	6.3 ± 0.3	71.7 ± 0.6	18a	6.0 ± 0.4
4	68.3 ± 0.6	14.9 ± 1.3	3.7 ± 0.3	9.3 ± 0.7	6.7 ± 0.4	67.9 ± 0.7	18.0 ± 1.7	6.0 ± 0.6
5	66.1 ± 0.7	20.6 ± 1.7	2.8 ± 0.2	7.0 ± 0.5	7.3 ± 0.7	67.1 ± 1.5	21.3 ± 3.6	1.8 ± 0.3
6	64.1 ± 0.5	18a	4.2 ± 0.2	10.0 ± 0.5	7.2 ± 0.2	60.6 ± 0.4	18a	8.9 ± 0.5
	85.1 ± 1.2	18a	1.7 ± 0.2	4.2 ± 0.4	7a	85.1 ± 3.1	18a	1.3 ± 0.3
7	64.3 ± 0.9	18a	1.9 ± 0.2	4.8 ± 0.4	7.6 ± 1.7	65.0 ± 3.0	18a	1.0 ± 0.4
	92.3 ± 1.7	18a	1.0 ± 0.2	2.6 ± 0.4	7a	94a	18a	1.2 ± 0.5
	116.7 ± 1.7	18a	1.0 ± 0.2	2.5 ± 0.4	⋯	⋯	⋯	⋯
8	54.4 ± 1.0	19.1 ± 2.4	2.0 ± 0.2	5.1 ± 0.6	8.7 ± 1.1	53.4 ± 1.4	18.8 ± 3.7	1.6 ± 0.3
	92.9 ± 1.7	18.7 ± 4.0	1.2 ± 0.2	2.9 ± 0.5	7a	85a	18a	0.6 ± 0.2
9	53.8 ± 0.7	18.1 ± 1.6	1.8 ± 0.1	4.4 ± 0.3	10.4 ± 1.4	50.5 ± 1.1	11.0 ± 1.6	0.9 ± 0.2
	89.0 ± 0.8	20.7 ± 2.0	1.8 ± 0.1	4.4 ± 0.4	4.6 ± 2.5	87.1 ± 4.6	18a	1.4 ± 1.0
10	51.6 ± 0.9	12.9 ± 2.0	1.5 ± 0.2	3.7 ± 0.5	7a	53a	15a	1.3 ± 0.4
11	15.3 ± 2.1	18a	1.6 ± 0.2	2.5 ± 0.4	7.3 ± 1.2	9.3 ± 2.3	18a	0.9 ± 0.2
	35.2 ± 1.4	18a	1.0 ± 0.2	4.0 ± 0.5	5.8 ± 1.4	36.0 ± 3.0	18a	0.9 ± 0.3
12	53.1 ± 1.7	14.4 ± 3.9	1.0 ± 0.2	2.6 ± 0.6	5.7 ± 0.6	57.1 ± 0.9	13.9 ± 2.2	1.5 ± 0.2
13	−0.5 ± 4.0	18a	0.4 ± 0.1	1.1 ± 0.4	6.6 ± 1.2	11.6 ± 2.5	18a	1.4 ± 0.6
	18.3 ± 2.4	18a	1.0 ± 0.1	2.6 ± 0.4	5.1 ± 1.8	27.8 ± 3.4	18a	1.2 ± 0.3
	39.1 ± 2.9	18a	0.8 ± 0.1	2.0 ± 0.4	5.6 ± 3.6	39a	18a	1.5 ± 0.6
	57.1 ± 2.5	18a	0.7 ± 0.2	1.8 ± 0.4	⋯	⋯	⋯	⋯

Note. For deriving the column density N_tot of C¹⁸O, the excitation temperature T_ex is assumed to be equal to 10 K for all positions and velocity components (see text).

^aParameter fixed in the LTE analysis.

Download table as:  ASCIITypeset image

To estimate ratios of HCO⁺ to HOC⁺, HCO, and CO⁺, we have also analyzed the HC¹⁸O⁺ J = 1–0 and 3–2 spectra shown in Figure 10 toward our target positions in Sgr B2. As seen in this figure, the HC¹⁸O⁺ J = 1–0 line is detected in all positions, but the HC¹⁸O⁺ J = 3–2 line is not. The ground-state transition of HC¹⁸O⁺ is absorbed in positions 1 and 2.

Zoom In Zoom Out Reset image size

As in the case of the other molecules, we have fitted synthetic spectra of HC¹⁸O⁺ J = 1–0 and 3–2 to our data to estimate the HC¹⁸O⁺ column density. The HC¹⁸O⁺ lines of positions 1 and 2 were modeled considering that Sgr B2M and Sgr B2N have continuum background sources emitting as blackbodies with similar conditions to those used for HOC⁺. An additional warmer component of HC¹⁸O⁺ with a T_ex of 21 K and a column density of (0.5–1) × 10¹³ cm⁻² is needed to fit the HC¹⁸O⁺ lines in positions 1 and 2. Multiple velocity components are taken into account in the HC¹⁸O⁺ line fitting of positions 6–9, 11, and 13. The found parameters are listed in Table 5. As seen from this table, the temperature within ∼6–10 K explains the excitation of HC¹⁸O⁺ toward several positions studied in Sgr B2, except the uncertain T_ex of 1 K required to model the absorbed HC¹⁸O⁺ J = 1–0 lines in positions 1 and 2.

In Section 5.3 we use HCO⁺ column densities derived from the HC¹⁸O⁺ column densities given in Table 5 and using the ¹⁶O/¹⁸O isotopic ratio of 250 (Wilson & Rood 1994). From these HCO⁺ column densities, we have also derived relative abundances of HCO⁺ within (0.5–3.2) × 10⁻⁸ (see Figure 11) considering the H₂ column densities listed in Table 5.

Zoom In Zoom Out Reset image size

As mentioned above, the variability in the X-ray emission in Sgr B2 could have been produced by several flares originated in Sgr A* (Terrier et al. 2010, 2018). In this scenario, Sgr B2 and other GC clouds would act as reflection clouds. Figure 1 shows that the gas mainly toward positions 1, 2, 6, and 12 was affected by X-ray emission in 2004, which may have affected the chemistry of the molecular gas. To address the possible effects of the X-rays on the molecular gas, we have calculated the X-ray ionization rate ζ_X for Sgr B2 using Equation (4) given in Maloney et al. (1996),

$$\begin{eqnarray}&&\begin{array}{l}{\zeta }_{{\rm{X}}}\sim 1.4\times {10}^{-11}\,{s}^{-1}\left(\displaystyle \frac{L}{{10}^{44}\,\mathrm{erg}\,{{\rm{s}}}^{-1}}\right)\\ \quad \times \,{\left(\displaystyle \frac{r}{{10}^{2}\mathrm{pc}}\right)}^{-2}{\left(\displaystyle \frac{{N}_{{{\rm{H}}}_{2}}}{{10}^{22}{\mathrm{cm}}^{-2}}\right)}^{-1},\end{array}\end{eqnarray} \tag{ 1 }$$

where L is the X-ray luminosity, r is the distance to the X-ray source, and ${N}_{{{\rm{H}}}_{2}}$ is the H₂ column density. We have derived ζ_X ∼ 10⁻¹⁹ s⁻¹ by considering L ∼ 10³⁷ erg s⁻¹ found for positions 1 and 12 in Section 3, r = 120 pc (the projected distance between Sgr B2 and Sgr A*), and the average column density ${{\rm{N}}}_{{{\rm{H}}}_{2}}$ of 1 × 10²³ cm⁻² estimated for positions 1, 2, 6, and 12 (see Table 5), where X-ray emission is detected in our 2004 map shown in Figure 1. The ζ_X value of ∼10⁻¹⁹ s⁻¹ derived for Sgr B2 agrees with those shown in Figure 4 of Harada et al. (2015) for the H₂ column densities higher than 10²³ cm⁻². The effects of X-rays on the chemistry of the molecular gas are similar to those of cosmic rays (Harada et al. 2015). The cosmic-ray ionization rate ζ is (4–7) × 10⁻¹⁶ s⁻¹ in Sgr B2 (van der Tak et al. 2006; Bonfand et al. 2019), which is 3 orders of magnitude higher than the ζ_X derived for Sgr B2. The very low ζ_X of ∼10⁻¹⁹ s⁻¹ shows that the effects of the X-rays on the Sgr B2 molecular gas might be negligible.

Figure 11 shows the abundances of HOC⁺, HCO, CO⁺, and HCO⁺ estimated for the 13 positions of Sgr B2. There is more than one bar in positions 6–9, 11, and 13 in this figure because of the multiple velocity components. The gas in positions 1–8 and 10–12 was pervaded by X-rays in 2004, as mentioned in Sections 1 and 3, but from now on, the gas in positions 4 and 10, the high-velocity gas in positions 8 and 9, and the gas in positions 10–12 are considered quiescent, as the effects of the X-rays on the Sgr B2 molecular gas might be negligible, as shown in Section 5.1. The HOC⁺ shows the highest abundance of 9 × 10⁻¹¹ toward position 2 tracing dense H ii gas. The low-velocity quiescent gas of position 13 also shows a high abundance of HOC⁺, which is a factor of 3.2 lower than that in position 2. This suggests that HOC⁺ is formed efficiently in regions affected by UV photons but also through other chemical routes discussed below.

Figure 12 shows fractional abundances of HOC⁺, HCO, and CO⁺ obtained with the gas-grain time-dependent code Nautilus (Hersant et al. 2009; Semenov et al. 2010). The initial condition for hydrogen is in the molecular form, while for other elements with a higher ionization potential than 13.6 eV, it is in the atomic form, and for all other elements, it is in the ionic form. We ran the model with constant physical conditions. The modeled abundances are obtained for an H₂ density of 10⁴ cm⁻³ found toward Sgr B2 (Hüttemeister et al. 1995; Armijos-Abendaño et al. 2015), kinetic temperatures of 50 and 100 K, a visual extinction of 10 mag (implying H₂ column densities of ∼10²² cm⁻², values estimated for several Sgr B2 positions as shown in Table 5), and ζ values of 10⁻¹⁷, 10⁻¹⁶, 10⁻¹⁵, and 10⁻¹⁴ s⁻¹; ζ values of (4–7) × 10⁻¹⁶ s⁻¹ are found toward Sgr B2 (van der Tak et al. 2006; Bonfand et al. 2019). Kinetic temperatures of 50 and 100 K are included in our chemical modeling because these temperatures are the approximate lower and upper values of the range of kinetic temperatures found for GC molecular clouds, which may be driven by turbulent dissipation and/or cosmic rays (Ao et al. 2013). However, Ginsburg et al. (2016) proposed that turbulent heating is the main heating mechanism of the GC dense gas. Meijerink et al. (2011) predicted gas temperatures of ∼100–200 K for hydrogen column densities >10²² cm⁻² (as those found for Sgr B2) and ζ values within 5 × 10⁻¹⁷−5 × 10⁻¹³ s⁻¹ using PDR models that mimic the effects of cosmic rays and mechanical heating on the gas, but when mechanical heating is not included in the modeling, gas temperatures >50 K (for hydrogen column densities >10²² cm⁻²) are only reached for the high ζ values of 5 × 10⁻¹⁴ and 5 × 10⁻¹³ s⁻¹, which are at least a factor of 100 higher than that of ∼5 × 10⁻¹⁶ s⁻¹ estimated for Sgr B2 (van der Tak et al. 2006; Bonfand et al. 2019). This contrasts with what was found by Bisbas et al. (2015, 2017), who predicted gas temperatures ≲50 K using models with ζ values ≤10⁻¹⁴ s⁻¹. Taking into account the ζ value of ∼5 × 10⁻¹⁶ s⁻¹ for Sgr B2, mechanical heating may be the mechanism responsible for increasing the gas temperatures to values >50 K.

Zoom In Zoom Out Reset image size

As seen in Figure 12, the HOC⁺ abundance increases as the value of ζ increases. Figure 12 shows differences of more than 1 order of magnitude in the HOC⁺ abundances predicted by the models with the same ζ value of 10⁻¹⁴ s⁻¹ but different kinetic temperatures of 50 and 100 K at timescales >10^5.0 yr, whereas in the cases with the other ζ values, these differences are of less than a factor of ∼3 over the timescales of 10²–10⁶ yr. The HOC⁺ abundances of ∼(0.5–2.8) × 10⁻¹¹ are estimated toward the positions of quiescent gas in Sgr B2. Figure 12 shows that the HOC⁺ abundances of ∼(0.5–2.8) × 10⁻¹¹ are reached at timescales lower than ∼10^2.8 yr for a ζ value of 10⁻¹⁵ s⁻¹ or at timescales within ∼10^3.0–10^3.2 yr for a ζ value of 10⁻¹⁶ s⁻¹. Both ζ values are in agreement with those found for Sgr B2 (van der Tak et al. 2006; Bonfand et al. 2019). The highest HOC⁺ abundance of 2.8 × 10⁻¹¹ found for the quiescent gas in Sgr B2 is very close to that predicted by models with a ζ value of 10⁻¹⁶ s⁻¹ and timescales >10⁵ yr (see Figure 12). Figure 12 also shows HOC⁺ abundances of >5 × 10⁻¹¹ for a ζ value of 10⁻¹⁴ s⁻¹ and different timescales. These HOC⁺ abundances are at least a factor of ∼1.8 higher than the highest HOC⁺ abundance of 2.8 × 10⁻¹¹ estimated for the quiescent gas in position 13. Our derived abundances of HOC⁺ are also consistent with those predicted by models with kinetic temperatures of 50 and 100 K and a ζ value of 10⁻¹⁷ s⁻¹ at timescales within ∼10^3.2–10^5.2 yr; however, the ζ value of 10⁻¹⁷ s⁻¹ is at least a factor of 40 lower than that found in Sgr B2 (van der Tak et al. 2006; Bonfand et al. 2019). This shows that the HOC⁺ abundances observed in the Sgr B2 quiescent regions are well explained by a high-temperature chemistry and the effects of cosmic rays with ζ values of 10⁻¹⁶ and 10⁻¹⁵ s⁻¹ at different timescales. Both ζ values will be considered in our discussion below.

The HOC⁺ abundance of 4 × 10⁻⁹ is found toward the XDRs in the circumnuclear disk of NGC 1068 (Usero et al. 2004), which is a factor of 143 higher than the highest HOC⁺ abundance derived for the quiescent regions of Sgr B2, ruling out the effects of X-rays on the molecular gas and consistent with our discussion in Section 5.1.

As seen in Figure 11, the highest HCO abundances of ∼1 × 10⁻⁹ are found toward position 3, tracing diffuse H ii gas, as well as toward positions 1 and 2, tracing dense H ii gas. The HCO also shows high abundances of ∼(3–7) × 10⁻¹⁰ for the diffuse H ii gas in positions 5, 7, 8, and 9 and ∼(1–7) × 10⁻¹⁰ for the quiescent gas in positions 4, 6, and 10–13. This indicates that HCO forms efficiently in gas affected by UV photons but also in quiescent gas. It is believed that photoprocessing of ice mantles and the subsequent photodesorption of HCO or H₂CO (followed by gas phase photodissociation) are responsible for high HCO abundances of ≃(1–2) × 10⁻⁹ in PDRs (Gerin et al. 2009), values that are consistent with those of ∼1 × 10⁻⁹ in positions 1–3.

Chemical modeling of a molecular cloud affected by nondissociative shocks of 10 km s⁻¹ shows relative abundances of HCO higher than ∼1 × 10⁻⁹ starting 10³ yr after the shock (Mitchell 1984); this is slightly higher than the value of around 7 × 10⁻¹⁰ found for the low-velocity quiescent gas in position 13.

Figure 12 reveals differences of less than a factor of 10 between the HCO abundances predicted by models with the same ζ values of 10⁻¹⁶ and 10⁻¹⁵ s⁻¹ but different kinetic temperatures over timescales of 10²–10⁶ yr. As seen in Figure 12, the highest HCO abundance of ∼8 × 10⁻¹¹ is predicted for a ζ value of 10⁻¹⁵ s⁻¹, a kinetic temperature of 50 K, and timescales of ∼10^3.5 yr. This HCO abundance is only a factor of 1.2 lower than that of 1 × 10⁻¹⁰, but it is a factor of 8.7 lower than the highest abundance of 7 × 10⁻¹⁰ found for the quiescent gas in position 13. This suggests that high-temperature chemistry and cosmic rays are not enough to explain the highest HCO abundances observed in the Sgr B2 quiescent gas and that other mechanisms, such as shocks, are needed to explain those high HCO abundances. It is believed that HCO in hot corinos is formed by the hydrogenation of CO on grains and subsequent thermal desorption in the protostellar phase (Rivilla et al. 2019). This scenario is consistent with what we propose in our paper, with the difference that instead of desorption in the quiescent gas of Sgr B2, the HCO is released to the gas phase through grain sputtering by shocks.

On the other hand, CO⁺ line emission is detected only in positions 1–3 and 9 (at gas velocities of ∼50 km s⁻¹) and without ambiguity only in position 9, as was mentioned before. The CO⁺ abundances in these four positions are within ∼(4–10) × 10⁻¹⁰. We mentioned in Section 4.2 that the H42α RRL has the highest electron density in position 9 between all the studied positions of diffuse H ii gas. This suggests that CO⁺ is produced mainly via reactions where UV photons play an important role. This is also supported by the coincidence between the CO⁺ relative abundances derived in this paper and those of around 10⁻⁹ shown by the PDR models of Martín et al. (2009b). Limits lower than ∼2 × 10⁻⁹ are estimated for diffuse H ii gas in positions 5, 7, and 8 or quiescent gas in positions 4, 6, and 10–13. The upper limits of these nine positions agree with those lower than ∼5 × 10⁻¹¹ obtained for the four values of ζ shown in Figure 12. This comparison points out that high abundances of CO⁺ are expected only in PDRs, as our chemical models including only high-temperature chemistry and cosmic-ray effects show very low CO⁺ abundances.

We found differences of less than a factor of 2 in the CO⁺ abundances predicted by models with the same ζ values of 10⁻¹⁶ and 10⁻¹⁵ s⁻¹ but different kinetic temperatures of 50 and 100 K over timescales of 10²–10⁶ yr (see Figure 12).

Column density ratios of HCO⁺ versus HOC⁺, HCO, and CO⁺ for the studied positions of Sgr B2 are given in Table 6. This table lists several velocity components for positions 6–9, 11, and 13 as multiple velocity components of gas are found (see Section 4). The ratios corresponding to NGC 253, M82, Horsehead, and the Orion bar, considered as prototypical PDRs, are included in this table as well (Apponi et al. 1999; Schilke et al. 2001; Fuente et al. 2003; Savage & Ziurys 2004; Goicoechea et al. 2009; Gerin et al. 2009; Martín et al. 2009b). The HCO⁺/HOC⁺ and HCO⁺/HCO ratios as a function of the H₂ column density, the electron density (derived in Section 4.2), and the fluxes of the 6.4 keV Fe line (derived from the 2004 X-ray data in Section 3) for the positions studied in our paper are displayed in Figure 13. The lowest HCO⁺/HOC⁺ ratios of 117–250 are found in positions 1 and 2, while the other positions tracing diffuse H ii regions or quiescent gas show HCO⁺/HOC⁺ ratios higher than 337. The bottom panels of Figure 13 show that there is a good correlation between the HCO⁺/HCO ratio and both the H₂ column density and the electron density. Figure 13 also shows that there is no correlation between the 6.4 keV Fe Kα line flux and both the HCO⁺/HOC⁺ and HCO⁺/HCO ratios.

Zoom In Zoom Out Reset image size

Figure 13 shows similar values of the HCO⁺/HOC⁺ ratios in both diffuse H ii and quiescent regions. The HCO⁺/HOC⁺ ratios of 117–250 agree with those of <270 of prototypical PDRs (see Table 6). Based on this, the HCO⁺/HOC⁺ ratio seems to be a good tracer of dense H ii regions where UV photons dominate the heating and chemistry of the molecular gas, but it does not allow one to distinguish between quiescent gas and diffuse ionized gas.

The bottom panels of Figure 12 show the HCO⁺/HOC⁺, HCO⁺/HCO, and HCO⁺/CO⁺ ratios obtained with our chemical modeling. We believe that the agreement of the HCO⁺/HOC⁺ ratios of 337–1541 estimated for the Sgr B2 quiescent regions is better for ζ = 10⁻¹⁶ s⁻¹, as this ζ value agrees better with that of 4 × 10⁻¹⁶ s⁻¹ found for the Sgr B2 envelope and timescales >10^4.5 yr. This is consistent with what was discussed in Section 5.2 and implying that the values of ζ = 10⁻¹⁶ s⁻¹ and timescales >10^5.0 yr give a better agreement with the observations. For these values, our model reveals HCO⁺ abundances of ∼1 × 10⁻⁸ that agree with those of (0.5–3.2) × 10⁻⁸ estimated for Sgr B2 (see Figure 12).

As seen in Figure 13, the HCO⁺/HCO ratio also shows the lowest values within 8–9 in positions 1 and 2 tracing dense H ii gas. On the other hand, the HCO⁺/HCO ratios within 11–20 and 24–60 are found in diffuse H ii gas and quiescent gas, respectively. It is clear in Figure 13 that this ratio shows a range of higher values in quiescent regions than in regions of diffuse H ii gas, implying that the HCO⁺/HCO ratio is a good tracer of regions affected by UV photons, as well as the HCO⁺/HOC⁺ ratio, but that this ratio is also sensitive to the electron density of the ionized gas. The HCO⁺/HCO ratios of 24–60 found for quiescent gas agree with those estimated for a ζ value of 10⁻¹⁵ s⁻¹, kinetic temperatures of 50 and 100 K, and timescales <10^3.1 yr or a ζ value of 10⁻¹⁶ s⁻¹, kinetic temperatures of 50 and 100 K, and timescales within 10^3.1–10^3.2 yr. It is important to stress that the agreement of the modeled ratios for the above ζ and timescales with those derived is due to both the modeled HCO⁺ and HCO abundances being the same orders of magnitude lower than those found for the quiescent gas (see Figure 12).

The HCO⁺/HCO ratios within 8–9 derived for positions 1 and 2 are in agreement with those of the starburst galaxies NGC 253 and M82 (Martín et al. 2009b).

On the other hand, we were able to derive HCO⁺/CO⁺ ratios only for positions 1–3 and 9, where CO⁺ emission was detected. We estimated lower limits on the HCO⁺/CO⁺ ratio for the other positions. As we can see in Table 6, the HCO⁺/CO⁺ ratio is also a good tracer of regions affected by UV photons, as this ratio is within 6–14 in positions 1, 2, and 9, where we found the highest electron densities among all of the studied regions (see Section 4.2), while the HCO⁺/CO⁺ ratio of 41 is derived for the diffuse ionized gas in position 3. The HCO⁺/CO⁺ ratio of 41 found for position 3 is reached at timescales lower than ∼10^3.0 yr for ζ values of 10⁻¹⁶ and 10⁻¹⁵ s⁻¹ (see Figure 12). However, this agreement is due to both the modeled HCO⁺ and CO⁺ abundances being the same orders of magnitude lower than the derived values, as in the case of the HCO⁺/HCO ratio.

Martín et al. (2009b) found HCO⁺/CO⁺ ratios of 32 ± 16 and 38 ± 15 for M82 and NGC 253, respectively, which are between the values found for the dense and diffuse H ii regions. The upper limit of <83–140 on the HCO⁺/CO⁺ ratio found for the Orion bar (Savage & Ziurys 2004; Fuente et al. 2006) is also consistent with those within 6–41 estimated toward positions 1–3 and 9.