Unusually High HCO+/CO Ratios in and outside Supernova Remnant W49B

Galactic supernova remnants (SNRs) and their environments provide the nearest laboratories to study SN feedback. We performed molecular observations toward SNR W49B, the most luminous Galactic SNR in the X-ray band, aiming to explore signs of multiple feedback channels of SNRs on nearby molecular clouds (MCs). We found very broad HCO+ lines with widths of dv ∼ 48–75 km s−1 in the SNR southwest, providing strong evidence that W49B is perturbing MCs at a systemic velocity of V LSR = 61–65 km s−1, and placing the W49B at a distance of 7.9 ± 0.6 kpc. We observed unusually high-intensity ratios of HCO+ J=1–0/CO J=1–0 not only at shocked regions (1.1 ± 0.4 and 0.70 ± 0.16) but also in quiescent clouds over 1 pc away from the SNR’s eastern boundary (≥0.2). By comparing with the magnetohydrodynamics shock models, we interpret that the high ratio in the broad-line regions can result from a cosmic-ray (CR) induced chemistry in shocked MCs, where the CR ionization rate is enhanced to around 10–102 times of the Galactic level. The high HCO+/CO ratio outside the SNR is probably caused by the radiation precursor, while the luminous X-ray emission of W49B can explain a few properties in this region. The above results provide observational evidence that SNRs can strongly influence the molecular chemistry in and outside the shock boundary via their shocks, CRs, and radiation. We propose that the HCO+/CO ratio is a potentially useful tool to probe an SNR’s multichannel influence on MCs.

1. INTRODUCTION Supernovae (SNe) are among the most energetic explosions in galaxies.SNe quickly fade away but leave their energy and metals in their supernova remnants pingzhou@nju.edu.cnzgy0106@gmail.com(SNRs).While expanding in the interstellar medium (ISM), SNRs return a huge amount of energy (typically around the order of 10 51 erg) to the galaxies via shocks, radiation, and cosmic rays (CRs; Vink 2020).
Theoretical studies show that SN feedback strongly regulates star formation and galaxy evolution (e.g., Scannapieco et al. 2008;Ostriker & Shetty 2011).Nevertheless, there are large uncertainties on the level of SN feedback, because it relies on the understanding of SN population and energy (Keller & Kruijssen 2022), and a clear knowledge of how multiple feedback channels (shocks, radiation, CRs) of SNR influence the ISM.Observing SNRs is a direct way to quantify the SN feedback and test those SN parameters applied in numerical simulations.For example, Koo et al. (2020) used HI and molecular observations of seven SNRs and found that their momentum and kinetic energy agree well with those used in numerical simulations for typical SNe.In addition to the mechanical output, SNRs release part of their energy as CRs and radiation, which can influence the preshock area (see e.g., Vaupré et al. 2014).
As a relatively dense phase of the interstellar gas, molecular clouds (MCs) can efficiently slow down the SNR shocks, absorb the radiation, and interact with CRs.While shocks have been the most widely studied heating channel from SNRs, it is also crucial to learn the influence of SNRs' CRs and radiation.CRs are the main heating sources of dense molecular gas where UV emission is shielded (Goldsmith & Langer 1978).Xray emission has been regarded as important ionization and heating source of neutral gas in active galactic nuclei (AGN) and primitive galaxies (Maloney et al. 1996;Lebouteiller et al. 2017;Vallini et al. 2019).The term X-ray dissociation regions or X-ray dominated regions (XDRs) has been used to describe the regions where the heating and chemistry of the gas are controlled mainly by the intense X-ray fluxes (Maloney et al. 1996).Due to the small cross section (decreasing with X-ray energy), the X-ray photons can traverse deep into the MCs and irradiate the gas opaque to UV photons.So far, more than 70 SNRs have been found to interact (or likely interact) with MCs (see Frail et al. 1996;Reach et al. 2006;Chen et al. 2014;Lee et al. 2019, and references therein).Many of them are bright in the X-ray and γ-ray bands (Slane et al. 2015).They also have large infrared-to-X-ray flux ratios suggesting a radiative heating of dust (Koo et al. 2016).These SNRs and their environments provide the nearest laboratories to study the materials exposed to highly enhanced CRs and X-ray radiation.
SNR W49B is the most luminous Galactic SNR in the X-ray band (Immler & Kuntz 2005) and has been subject of substantial X-ray studies on its plasma properties and SNR type (Hwang et al. 2000;Miceli et al. 2006;Keohane et al. 2007;Ozawa et al. 2009;Zhou et al. 2011;Lopez et al. 2013b,a;Zhou & Vink 2018;Yamaguchi et al. 2018;Zhang et al. 2019;Sun & Chen 2020;Holland-Ashford et al. 2020;Siegel et al. 2020).W49B is detected in H.E.S.S. and Ferm-LAT observations, which suggest its GeV-TeV γ-ray emission is produced through the decay of π 0 created by the interaction between accelerated CR protons and nearby dense ma-terials (H.E. S. S. Collaboration et al. 2018).Both the X-ray and γ-ray studies suggest that W49B is evolving in a dense medium.The IR study of W49B shows that the dense gas and dust have multiple temperature components and the SNR is evolving in a cloudy environment (Keohane et al. 2007;Zhu et al. 2014).
Although W49B appears to be a good target to study SNR feedback onto dense gas, there are disputes on the SNR's distance and associated clouds.Previous HI emission and absorption studies have not reached a consensus on the SNR distance issue.While some HI studies claimed that W49B lies 3-4 kpc closer than the starforming region W49A to the west at ∼ 11 kpc (Kazes 1970;Radhakrishnan et al. 1972;Zhang et al. 2013), others preferred a distance of 11-14 kpc (Lockhart & Goss 1978;Brogan & Troland 2001).The 12 CO observation shows a morphological relation between the molecular gas and the SNR at the local-standard-ofrest velocity (V LSR ) at ∼ 40 km s −1 , corresponding to a distance of 9-10 kpc (Chen et al. 2014;Zhu et al. 2014).The first kinetic evidence of W49B-MC association was provided by a recent study by Lee et al. (2020), who found that the near-infrared H 2 line centroid is at V LSR = 64 ± 2 km s −1 and proposed an SNR distance of ∼ 7.5 kpc.This velocity is consistent with those suggested by the earlier HI absorption studies (e.g., Kazes 1970) and radio recombination lines from ionized gas toward W49B (Downes & Wilson 1974;Liu et al. 2019).The dispute on W49B's distance is not ended.Recently, Sano et al. (2021) proposed a distance of 11 kpc for W49B using 12 CO observations, which reveal warm MCs at V LSR ∼ 10 km s −1 .
Motivated by the above issues, we performed dedicated molecular observations toward W49B at the millimeter wavelength, aiming to achieve two goals: (1) to search for MCs perturbed by this SNR, and thus pin down the SNR distance; (2) more importantly, to explore signs of multiple feedback channels of an SNR on nearby MCs.We found unusually high HCO + /CO ratios in and outside W49B, which are signs that the CRs, shocks, and radiation of W49B strongly influence the chemistry of the environmental gas.We present our observations and data in Section 2. The data analysis is shown in Section 3. Section 4 shows our interpretation of the unusual chemical properties, with discussions of relevant issues.The conclusion is summarized in Section 5.

OBSERVATION AND DATA
We performed millimeter observations toward W49B using the IRAM 30 m telescope during 2019 May 15-16 and 2020 July 24-27 (Project IDs: 167-18 and 024-20;  PI: P. Zhou).We used the Eight MIxer Receiver (EMIR) to observe two frequencies (3 mm and 1 mm or 3 mm and 2 mm) simultaneously in the position-switching model.On-the-fly and on-off observing modes were applied for mapping and pointing observations, respectively.The back end of fast Fourier transform spectrometers (FTSs) provided a frequency resolution of 195 kHz, which corresponds to a velocity resolution of ∼ 0.25 km s −1 at 235 GHz.The half-power beamwidth (HPBW) of the telescope was 10. 7 at 230 GHz, 20. 3 at 115 GHz, and 29 at 86 GHz.The main-beam efficiencies at these three frequencies were 59%, 78%, and 81%, respectively 1 .We mapped W49B in the frequency ranges 83.GHz.The mapped region is similar to the field of view (FOV) of the multiband images in Figure 1.The observations are deeper in the east, west, and south regions to probe faint emission near the SNR boundary.
In this paper, we focus on CO and HCO + lines, but we also detected a few other lines such as HCN and HNC lines.The line frequency and rms of a few molecular lines are tabulated in Table 1.All the IRAM 30m data were reduced using the GILDAS software (vers.10oct182 ).
For comparison purposes, the data cubes of the HCO + , 12 CO, 13 CO, and HCN emission were resampled to have the same velocity resolution of 0.7 km s −1 , while a larger velocity bin is sometimes used for broad lines to increase the signal-to-noise ratios.The pixel sizes of 12 CO J=2-1 and 12 CO/ 13 CO J=1-0 are 5 and 10 , respectively.In the ratio maps, the 12 CO J=2-1 data are convolved to 30 and regridded with a pixel size of 10 .

Search for shocked molecular gas
To search for MCs associated with W49B, we first investigated molecular line profiles across the SNR.Finding broadened molecular lines can provide strong evidence for a shock perturbation on the MCs (see e.g., Jiang et al. 2010), and also helps to determine the SNR distance.
We found very broad HCO + J=1-0 emission only in the southwestern regions of W49B at V LSR = 40-150 km s −1 (see the spectra in Figure 2. All the regions correspond to strong H 2 emission in the southwest (see the box regions A, B, and C in Figure 1).
By fitting the broad lines with Gaussian profiles, we obtained line widths (FWHM) of 48 ± 3 km s −1 , 62 ± 3 km s −1 , and 75 ± 7 km s −1 at regions A, B, and C, respectively (see Table 2).These lines are too broad to be attributed to quiescent gas but are strong evidence of MCs accelerated by the SNR shock.The line widths are also distinctly larger than those pre-dicted in C-type shock, where the H 2 molecules survive up to a shock velocity of 45 km s −1 (Draine et al. 1983).At region C, only one prominent narrow HCO + line is seen at V LSR ∼ 61 km s −1 , while the faint line at V LSR ∼ 10 km s −1 is too far away from the broad component to be related.There are also narrow HCO + lines at V LSR = 61-63 km s −1 at regions A and B, and at V LSR = 64-65 km s −1 at regions D-F (see further discussions in § 3.2.1 and § 4).We fitted the narrow HCO + lines with V LSR > 60 km s −1 using Gaussian lines and tabulate the fit results in Table 2.
Our study supports an SNR systemic velocity of V LSR = 61-65 km s −1 , where we found narrow lines (2-6 km s −1 for HCO + emission based on the Gaussian fit; Table 2) nearest to the broad emission and with relatively high HCO + /CO intensity ratios.The systemic velocity is consistent with that obtained from the previous H 2 spectral study (64 ± 2 km s −1 , Lee et al. 2020).These narrow line components can be explained as the preshock gas.This LSR velocity, combined with the measured Galactic rotation curve (Reid et al. 2014), places W49B at a far distance of 7.9 ± 0.6 kpc or a near distance of 4.1 ± 0.6 kpc, where the uncertainty is given 61.8 ± 0.8 47.5 ± 2.9 0.064 ± 0.005 3.07 ± 0.17 1.60 ± 0.15 (narrow) 61.9 ± 0.1 5.7 ± 0.4 0.14 ± 0.  at the 1 σ level based on the Monte Carlo method by Wenger et al. (2018) 3 .Due to the HI absorption up to V LSR = 70 km s −1 (the LSR velocity of the tangent point at ∼ 6 kpc; see e.g.Zhu et al. 2014), W49B should be a bright radio source behind the tangent point.Therefore, W49B is located at a distance of 7.9±0.6kpc.This means that W49B is far away from the star-forming region W49A at a distance of 11.1 +0.8 −0.7 kpc (based on parallax measurements using H 2 O masers; Zhang et al. 2013).
A centroid shift of the HCO + broad line is clearly seen from regions A to C, indicating different viewing angles to the shock.The shift reaches the maximum of ∼ 36 km s −1 at region C.In regions B and C, we only found red-shift HCO + line wings, corresponding to the receding gas that is shocked in the far side of the SNR.The MCs at region A are shocked perpendicular to the line of sight, as its HCO + profile shows both redshift and blue-shift wings, and the broad-line centroid is the same as the systemic velocity.An interaction with dense MCs explains the deformed radio morphology in the southwest.The schematic views of clouds A-F from the observer and the western side are given in Figure 9 in the appendix.
Figure 3 shows the spatial distribution of shocked HCO + in the velocity range V LSR = 70-120 km s −1 .This velocity range corresponds to the red wing of broadened lines.We have not found any broad HCO + or 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.  12 CO lines in other regions of the SNR, except for those visible in Figure 3.This means that the kinetic evidence of SNR-MC interaction has only been established in the southwestern regions of W49B.The distribution of 12 CO J=2-1 and HCO + J=1-0 in the velocity range V LSR = 0-80 km s −1 with a step of 5 km s −1 are shown in Figures 10 and 11, respectively, in the Appendix.

HCO + and 12 CO emission
As the distance of W49B has been determined, we study the properties of the neighboring MCs. Figure 4 shows the distribution of HCO + J=1-0, 12 CO J=1-0, 12 CO J=2-1, and HCO + / 12 CO J=1-0 ratio at around the systemic velocity of the SNR.Although HCO + and 12 CO emissions are found in the interior of W49B, they might arise from background or foreground MCs, given the lack of physical evidence for an association with the SNR.Only the southwestern region has been found to show broadened HCO + emission (see § 3.1).
It is of interest to learn whether SNRs can influence the ISM outside its radio boundary and how the influence is reflected in the chemistry of the MCs.We found enhanced HCO + / 12 CO J=1-0 ratios near the eastern H 2 filament, and two clouds in the northeastern and southwestern regions, while the ratio is 10 −2 in the SNR interior.Hereafter we focus on the 2.12 µm H 2 filament 0. 5 (∼ 1.2 pc) away from the eastern radio boundary of W49B (see Figure 1 and Keohane et al. 2007). 4The near-IR H 2 emission requires hot or irradiated molecular gas in a high excitation state, raising questions about the heating mechanism of this filament (see discussions in Keohane et al. 2007;Lee et al. 2020).
To test whether the gas at the eastern 2.12 µm H 2 filament is shock heated, we extracted HCO + J=1-0 spectra at regions D, E, and F. The line centroid of the 2.12 µm H 2 emission is at V LSR = 64.3-65.5 km s −1 (Lee et al. 2020).At similar velocities, we found HCO + lines with narrow line widths (∆v = 2-4 km s −1 ), which argues against shock perturbation for producing the HCO + emission.The narrow line width is not unexpected, since the SNR's forward shock as traced by the radio emission does not reach the H 2 filament and thus should not directly impact the molecular gas.Another HCO + peak is found at V LSR ∼ 58 km s −1 to the SNR east (regions D, E, and F).There is no clear evidence to support an association between this velocity and the SNR, given the normal HCO + / 12 CO line ratio (∼ 0.02) and line widths.
The eastern HCO + gas with an unusually high HCO + / 12 CO ratio extends even further away from the SNR than the H 2 filament, although an HCO + knot is seen between the filament and the SNR radio boundary.The 12 CO J=2-1 map with a better angular resolution also shows a filamentary structure east of the H 2 filament.The layered distributions of different molecular emissions can be more clearly seen in Figure 5, which is cut along a slice across the eastern SNR shell (see the black dashed box in Figure 1).Both HCO + and H 2 emission peaks lie 50 (1.9pc) away from the radio peak.The H 2 emission is sharply enhanced on the western surface of the HCO + emission, while the 12 CO J=1-0 profile has little resemblance with the H 2 profile.In Section 4.3.3,we will discuss how the distribution is explained by the irradiated gas.

HCN lines and other molecular lines
We found HCN J=1-0 emission toward W49B, which contains a hyperfine triplet due to the nuclear quadrupole moment of N (F = 2-1, F = 1-1, F = 0-1).The separation of the hyperfine lines between highest and lowest frequencies is 3.522 MHz, which corresponds 4 Enhanced HCO + / 12 CO intensity ratios are found in two clouds in the northeast and southwest of W49B.The two clouds are warmer than 20 K and have the largest column densities in the FOV.Our observations cannot tell whether the two clouds are physically associated to the SNR or unassociated star-forming regions, and thus we do not discuss these clouds.Moreover, given that the 12 CO emission is optically thick, a high HCO + /CO intensity ratio does not necessarily mean a high abundance ratio.to 11.8 km s −1 .The brightness pattern among the hyperfine structures is not constant but depends on the excitation condition (e.g., Walmsley et al. 1982;Loughnane et al. 2012).
Due to the difficulty of separating hyperfine structures, we only examine the HCN properties of the six regions where enhanced HCO + are shown (see regions in Figure 1).We integrated the HCN J=1-0 emission at V LSR = 40-150 km s −1 and tabulated the values of the six regions in Table 2.The flux ratio of HCO + /HCN is 1.9-2.6 in the shocked gas and ∼ 1 near the eastern H 2 filament.
The enhancement of HCO + /HCN J=1-0 flux at the shocked regions is an intriguing phenomenon, which was also found in SNR IC 443 (with a ratio of ∼ 3, Denoyer & Frerking 1981).The explanation of the enhancement requires further investigation, which is outside the scope of this paper.
We also detected HNC J=1-0 at ∼ 64 km s −1 at regions A-F.SiO J=2-1 and DCO + J=2-1 lines are not detected in our observations.We found large HCO + /CO intensity ratios of 0.2-1 in both the shock-broadened lines in the southwest and narrow lines outside the SNR eastern boundary (see Figure 2), while the typical intensity ratios in other molecular components along the line of sight are 10 −2 .This is observational evidence to support that SNRs can influence the chemistry not only of the shocked clouds but also of MCs that are over 1 pc far away.
In the shocked regions, the HCO + /CO intensity is enhanced at the broadened line wings.In the velocity range of 75-150 km s −1 (without strong narrow 12 CO lines), the integrated HCO + intensities are 0.66 ± 0.14, 1.28±0.11,and 1.74±0.16K km s −1 , respectively, at regions A, B, and C, while the values for 12 CO J=1-0 lines are 0.49 ± 0.33, 1.84 ± 0.40, and 1.55 ± 0.31 K km s −1 , respectively.The 12 CO emission has only a 4-5σ detection at regions B and C. One cannot claim detection of broad 12 CO emission in region A as it is below 2σ, We estimate the flux ratio of I(HCO + )/I(CO) = 1.1 ± 0.4 at region C and 0.70 ± 0.16 at region B. Given the large velocity gradient, it is reasonable to assume the optically thin case for the HCO + and 12 CO emission in the broad lines.
Near the eastern H 2 filament, the line ratio of HCO + /CO is large at V LSR ∼ 64 km s −1 compared to those at other velocities.At region D, the mainbeam temperatures of HCO + and 12 CO are 0.22 K and ≤ 1.0 K, respectively, corresponding to an HCO + / 12 CO intensity ratio ≥ 0.2.Here the 12 CO J=1-0 mainbeam temperature is regarded as an upper limit, as it is in a line valley and the emission is contaminated by the red wing of the unassociated component at V LSR ∼ 58 km s −1 .The 12 CO J=1-0 emission is probably optically thin, as indicated by a large 12 CO J=2-1/J=1-0 ∼ 1.5 (only 1 in the optically thick case in the local thermodynamic equilibrium case; see Zhou et al. 2018).
The HCO + /CO intensities obtained above are subsequently used to estimate the abundance ratios (see § 4.1.2),which will be further compared with those predicted by the thermochemical models of molecular gas in the SNR environment (see § 4.2).

Column densities and abundance ratios
On the assumption of local thermodynamic equilibrium (LTE) and optically thin emission, the HCO + J=1-0 line can be used to estimate the column density N (HCO + ) (Mangum & Shirley 2015): where h and k are Planck constant and Boltzmann constant, respectively, ν is the frequency of the J=1-0 transition, J(T ) = hν exp(hν/kT )−1 , and the Cosmic Microwave Background temperature T bg = 2.73 K.For a kinetic/excitation temperature T ex T bg , the above equation is simplified as Similarly, we derive the 12 CO column density using the optically thin 12 CO J=1-0 emission for T ex T bg : From equations ( 2) and ( 3), it is convenient to obtain N (HCO + )/N (CO) ∼ 1.3×10 −3 I(HCO + )/I(CO) ∼ 1.3 × 10 −3 T mb (HCO + )/T mb (CO) for a gas excitation temperature T ex 20 K.
We caution that the HCO + /CO abundance ratios are obtained by assuming an LTE condition, considering we only have one HCO + transition.Further observations with mid-J and high-J transitions of HCO + and 12 CO are needed for a precise diagnosis of the excitation conditions.

Model the chemical and physical properties of a shocked MC exposed to CRs
The SNR shock can heat and ionize the gas and change the abundance of molecular species.The GeV-TeV observations of W49B have proved that the SNR is accelerating CR protons (H.E. S. S. Collaboration et al. 2018), which are known as important sources regulating molecular chemistry.We used two codes to compute the HCO + /CO abundance ratios of a shocked MC exposed to CRs.

Constant-temperature chemical model
We first used a chemical network with molecules containing H, C, O, and N (Bovino et al. 2016;Hily-Blant et al. 2018;Le Gal et al. 2014) to simulate the chemical evolution in a shock-heated cloud by employing the Figure 6.The chemical evolution in the shocked molecular region.ζ is the CR ionization rate per H2.The number density of hydrogen nH is contributed by atoms, ions, and molecules.Av is the extinction magnitude of the standard Draine UV photon field to which the MC is exposed.In the upper row, the effect of the UV photon field is assumed to be extremely small.We discard the ratio values when either the number density of HCO + or CO is lower than 10 −8 cm −3 .astrochemistry package KROME (Grassi et al. 2014).The initial condition is assumed to be in equilibrium at a temperature T = 15 K for various densities.We then calculate the evolution of every species after an abrupt increase of the temperature to about 10 2 K, 10 3 K, and 10 4 K (corresponding to a shock velocity of about 1.4 km s −1 , 5 km s −1 , and 15 km s −1 , respectively; see Figure 6).In the upper row of Figure 6, we ignored the effect of photoionization.In the lower panels, we used an extinction level, A V = 1, with a standard Drain UV field (Draine 1978).The exact radiation field and extinction near W49B have been thus far unclear (see § 4.3.3).By setting a moderate extinction level A V to 1, we assume the column density for shielding the radiation field is about 1.9 × 10 21 cm −2 (Bohlin et al. 1978) to show the effect of a moderate UV photoionization condition.We set a lower limit of the molecular number density, 10 −8 cm −3 , to make sure that the molecular emission are observable and prevent a small denominator.
According to our network, at the temperature of 10 3 K, the main channels to produce HCO + are CO + H + 3 → H 2 + HCO + and HOC + + H 2 → HCO + + H 2 .The former reaction converts some CO molecules to HCO + , raising the abundance ratio between HCO + and CO. Figure 6 reveals a higher HCO + /CO ratio for enhanced CRs, which can increase the number of ions in the MCs.When the temperature is raised to 10 4 K, both CO and HCO + molecules can be easily dissociated, but the CR ionization reaction chain (such as the reaction HCO CR − − → HCO + + e − and CO + H + 3 → HCO + + H 2 ) keeps the HCO + dropping slower than CO, resulting in an HCO + /CO abundance ratio > 10 −3 .In a slow shock with a lower temperature of 10 2 K, the HCO + /CO abundance ratios are below 10 −3 for various densities.Although the CR reaction chain still produces HCO + , the recombining reaction of HCO + (e.g., HCO + +e − → CO + H) and the reactions between HCO + and neutral atoms or molecules (e.g., HCO + + C → CO + CH + ) are relatively fast in this low temperature, resulting in a low HCO + /CO abundance ratio.
We note that this constant-temperature model is a simplified model that does not include cooling/heating mechanisms or the the magnetic fields, and the chemical network only contains up to four atom molecules.Nevertheless, this model shows that HCO + /CO abundances are highly sensitive to gas physical condition and CR ionization rate.This model is also useful to learn the main channels to form HCO + in a given condition.

MHD shock model
We further simulated the HCO + /CO ratio in a magnetohydrodynamic (MHD) shock using the Paris-Durham shock code (Godard et al. 2019), which takes the influence of magnetic fields on shocks into account.The MHD shock code calculates the physical and chemical evolution of the gas, by including over 125 species linked by more than 1000 reactions and with cooling and heating processes considered.We considered a plane-parallel shock wave propagating in a molecular and dusty homogeneous medium with a magnetic field strength of B = nH cm −3 1/2 µG, where n H is the preshock gas density of the three-phase gas n H = n(H)+2n(H 2 )+n(H + ).Firstly, we calculated the thermochemical conditions of the preshock gas, adopting the standard initial distribution of elemental fractional abundances among gasphase species, PAHs, icy grain mantles, and grain cores, as adopted by Flower & Pineau des Forêts (2003).We allow the gas to evolve in a quiet and dark condition with given densities and the Galactic level CR ionization rate ζ = 2.5 × 10 −17 s −1 per H 2 (a few 10 −17 s −1 in dense MCs; van der Tak & van Dishoeck 2000).The final static thermochemical properties are used as input parameters for the preshock gas.
The simulations show that the shock types at V sh < 20 km s −1 are nonstationary "continuous" shocks (young C-type shock), which are composed of a magnetic precursor and J-type tail (see Lesaffre et al. 2004, for the temporal evolution of young C-type shocks).The shocks at V sh 20 km s −1 become "jump" shocks (Jtype), where the ions and neutrals are coupled.
Figure 7 shows the modeled abundance HCO + /CO ratios with a shock age of 1 kyr, as a function of shock velocity, preshock gas density, and cosmic-ray ionization rate.The abundance ratio is obtained from the integrated HCO + and 12 CO column densities in the shocked layer.X[HCO + ]/X[CO] tends to decrease with increasing preshock gas density and shock velocity, while the enhanced CR ionization rate can significantly increase the ratio.
A high CR ionization and a moderate density are needed to explain the observed large X(HCO + )/X(CO), while the Galactic level CR (dotted lines) and very dense preshock gas (orange color) cases predict a too-low value.It is reasonable to expect two orders of magnitude higher CR ionization rate ζ near W49B, as found in other middle-age SNRs such as IC443 (ζ = 2×10 −15 s −1 ; Indriolo et al. 2010) and W28 (Vaupré et al. 2014).With ζ = 2.5 × 10 −15 s −1 , the high X(HCO + )/X(CO) can be reached in a wide range of shock velocities (see Figure 7).At V sh ≥ 25 km s −1 , X(HCO + )/X(CO) is less sensitive to the shock velocity but relies more on the CR ionization rate and density.
The broad HCO + lines (∆v = 50-75 km s −1 ) found in W49B hint at a fast shock in the MC.The broad-line profile can be explained by either a single cloud with a large velocity dispersion or multiple shocked structures accelerated to different LSR velocities.In either case, a large shock velocity is needed to explain the large line centroid shifts (from V LSR ∼ 60 km s −1 of the quiescent gas to V LSR ∼ 100 km s −1 for gas in region C).suggest that the high HCO + /CO ratio in the broad-line regions can result from a CR-induced chemistry in shocked MCs with a preshock density n H ∼ 10 3 -10 4 cm −3 .An enhancement of the HCO + /CO ratio induced by CRs has also been found in a low-tointermediate density cloud, IC 348 (Luo et al. in preparation).We do not claim an enhancement of the HCO + abundance, but only an enhancement of the HCO + /CO abundance ratio.Our MHD shock modeling shows that the 12 CO molecules tend to be dissociated with enhanced shock velocities. 12CO abundance can be 1-2 orders of magnitude lower than the level in the preshock gas with V sh 20 km s −1 (see Figure 8).The HCO + abundance also presents a negative trend to the shock velocity, but the high-level CR ionization strongly regulates the HCO + chemistry and maintains it above 0.1 of the preshock level.When shock velocity is low (V sh 15 km s −1 ) and does not effectively destroy HCO + molecules, the enhanced CRs can indeed cause an HCO + abundance enhancement.Nevertheless, the low shock velocities are unlikely to explain the broad HCO + profiles and the weak CO emission.Therefore, the observed high ratio of HCO + /CO does not mean an enhancement of HCO + , but reflects the more efficient destruction of CO molecules than that of the HCO + in the faster shock.
There are a few caveats in our MHD modeling and comparison.We have not evaluated the influence of external radiation on the HCO + /CO abundance ratio, but only show that the enhanced CRs can already explain the observation.The UV spectrum of W49B is unclear, while the X-ray-induced physics/chemistry is not considered in the MHD models.Moreover, the observed X(HCO + )/X(CO) for comparison is calculated under a simplified LTE assumption.

Irradiated gas in the SNR east
An unusually high HCO + /CO line ratio is also found over 1 pc away from the SNR eastern boundary (see Figure 2).Here we discuss three types of precursors that might cause peculiar chemical properties to the MCs ahead of the shock (near the eastern H 2 filament; see Figure 1): 1) CR precursor, 2) magnetic precursor, and 3) radiation precursor.We do not consider the SNR blast wave as the direct source, as it is too far from these regions to cause the HCO + /CO enhancement.The narrow HCO + line widths do not suggest the shock perturbation, either.

CR precursor
CRs can diffuse away from the SNR boundary.The diffusion length of CRs is expressed as (see Vink 2012, and references therein): where η accounts for the deviation from Bohm diffusion ( 10 in young SNRs, Vink 2012), E is the CR particle energy, B is the magnetic field and t is the acceleration age, equal to the SNR age (5-6 kyr).
Low-energy CRs (< 1 GeV) cannot diffuse over 1 pc away from the shock, unless η is unusually large.Highenergy CRs with E ≥ 1 TeV could reach this distance, but their energy losses via pion production are a few orders of magnitude larger than ionization losses, meaning inefficient in MC heating (Mannheim & Schlickeiser 1994).Assuming a hadronic origin of the γ-ray emission from W49B, γ-rays carry about at least 1/3 of the pion production losses, and the spectrum peaks at about 1/10 of the energy of the primary CRs (Hinton & Hofmann 2009).The total γ-ray flux of W49B above 0.1 TeV is F (> 0.1 TeV) ∼ 4 × 10 −12 erg cm −2 s −1 (H.E. S. S. Collaboration et al. 2018), only comparable to the flux of the 2.12 µm H 2 line east to the SNR (F (2.12µm H 2 ) ∼ 5 × 10 −12 erg cm −2 s −1 ; Lee et al. 2019).The comparison suggests that the ionization losses from TeV CRs of W49B are insufficient to power the observed molecular emissions.Therefore, CRs are not the favored heating source for the clouds east to W49B.

Magnetic precursor
A magnetic precursor can appear in weakly ionized MCs, which coexist two distinct fluids in the shocks: a neutral fluid and an ionized fluid.If the magnetosonic speed exceeds the shock speed (V m > V sh ), compressive waves can be generated upstream ahead of the shock.The extent of the magnetic precursor for a stationary C-type shock is expressed as (Draine 2011): where the fractional ionization x i is ∼ 10 −4 in the diffuse MCs.For a dark cloud with a density of 10 2 -10 5 cm −3 , there is a simple approximation x i ≈ 10 −5 (n H / cm −3 ) −1/2 (Draine 2011).The ionization fraction is raised in clouds exposed to enhanced UV/X-ray radiation or CRs, which could happen near W49B.Assuming the magnetic energy conservation, the magnetic field in the cloud is estimated as . A previous X-ray study found a blast wave velocity of V bw ∼ 480 km s −1 in W49B, with a preshock intercloud density of n 0 ∼ 10 cm −3 (Zhou & Vink 2018).Assuming a pressure equilibrium between the blast wave and cloud shock n H V 2 sh ∼ n 0 V 2 bw (McKee & Cowie 1975), the expression of the magnetic precursor length for W49B can be simplified as: l mag ∼ 2×10 16 b 2 0 cm, where b 0 is in the level of ∼ 1 in typical MCs (Crutcher et al. 2010) but could vary with specific environments.It was proposed that the magnetic precursor causes the high-excitation H 2 emission outside the SNR boundary (Keohane et al. 2007).However, our calculation suggests that the magnetic precursor length of W49B is too small to excite the H 2 and HCO + emission 1 pc away from the SNR boundary, while W49B is unlikely to be located in a strongly magnetized medium (Brogan & Troland 2001).
Our MHD models reveal that shock types in the southwestern MCs are not stationary C-type shocks, but are likely J-type with shock velocities 20 km s −1 (see Section 4.2.2).As the ions and neutrals are coupled in the J-type shocks, the magnetic precursor can be ignored.

Radiation precursor
Unlike the CR and magnetic precursors that have short-length scales, radiation can freely escape the SNR until the photons are absorbed/scattered by the ISM.W49B is the most luminous Galactic SNR in the Xray band (Immler & Kuntz 2005) with an intrinsic luminosity of L X ∼ 2-9 × 10 38 d 2 8 erg s −1 in 0.3-10 keV band (the upper limited is recalculated using the XMM-Newton data and model in Sun & Chen 2020), where the uncertainty mainly comes from different absorption models.Its soft X-ray emission is so heavily absorbed that the observed 0.3-10 keV X-ray luminosity is only 5×10 35 erg s −1 .This means that most of the soft X-ray energy is transferred to the environment of W49B and the ISM along the line of sight.
Previous X-ray spectroscopy reveals that the soft Xray emission below 1 keV is contributed by hot plasmas with a temperature of kT ∼ 0.2-0.3keV (∼ 3 × 10 6 K Zhou & Vink 2018;Sun & Chen 2020;Holland-Ashford et al. 2020).If the soft X-ray emission is characterized by an apec model with a temperature of 0.177 keV and a normalization of 86.7 cm −5 (see the best-fit apec + 2 * vvrnei model in Sun & Chen (2020)), we can roughly estimate a UV luminosity of L UV ∼ 6 × 10 39 erg s −1 at 6-300 eV and 7 × 10 37 erg s −1 at 6-13.6 eV.The values are likely lower limits, as there could be undetected gas components cooler than ∼ 0.2 keV.Therefore, W49B is also a luminous UV source.
Given the high X-ray and UV luminosity of W49B, we suggest radiation precursor be a plausible source for the hot MCs to the SNR east.As the UV emission is highly absorbed, we do not clearly know the UV spectrum of W49B and thus do not evaluate its influence on the MCs outside W49B.However, the strong UV emission can explain the existence of strong 8 µm IR emission to the SNR east (see Figure 1), which traces either the polycyclic aromatic hydrocarbons (PAHs) in strong radiation fields or hot dust grains (Draine 2011).There is a good correlation between the near-IR H 2 emission and 8 µm emission, but the latter emission is even more extended (up to 1. 7 or 4 pc) from the radio boundary.
The large X-ray luminosity of W49B makes it a potential laboratory to study MCs in XDRs.It has been proposed that the intense X-ray emission from SNRs could transfer a fraction of their energy to MCs and give rise to strong vibration-rotational H 2 (Draine & Woods 1990, 1991).Indeed, W49B is the most luminous SNRs in both the X-ray band and near-IR H 2 emission based on our current knowledge (Lee et al. 2019).According to XDR models by Maloney et al. (1996), the H 2 emission in XDRs has a relatively low ratio H 2 2-1 S(1)/ 1-0 S(1) ∼ 0.1 ( 0.3), similar to the values in shock excited gas, while the mean ratio for UV fluorescent excitation of H 2 is ∼ 0.5-0.6 (Black & van Dishoeck 1987).Lee et al. (2020) observed a ratio of 0.07-0.10 in the eastern H 2 filament and proposed that the gas could be either shocked or X-ray-irradiated gas.Here we prefer the latter scenario, as the former case is less likely.
We found that the HCO + emission is generally in the outer layer of the H 2 filament to the SNR east (see Figure 5).This distribution can also be explained with XDR models, which predict that H 2 emission arises from the surface layer of the irradiated MCs (Maloney et al. 1996).The layered HCO + and H 2 distribution likely reflects a radiation attenuation into the MCs.The XDR models by Meijerink & Spaans (2005) indeed provide multiple solutions for producing the observed high HCO + /CO intensity ratio (≥ 0.2 to the SNR east).However, the X-ray spectral shape (exp(−E/1keV)) and flux range (1.6, 16, and 160 erg cm −2 s −1 ) in the available XDR models do not cover that in W49B (see the X-ray spectral models in, e.g., Sun & Chen 2020), while the observed molecular transitions are too few to constrain the gas excitation condition.A reliable comparison can be provided in the future using new molecular observations and XDR models with flexible X-ray spectral parameters.

Remarks about HCO + /CO ratios near SNRs
In W49B, optically thin HCO + emission and HCO + /CO ratio are strongly enhanced in the shocked regions and thus appear to be good tracers of SNR shock under these particular conditions.We also note that high HCO + /CO intensity ratios of ∼ 1 are found in the shocked gas in IC 443 (Dickinson et al. 1980;Dickman et al. 1992).It is of interest to explore the HCO + /CO ratios in more SNRs that are impacting MCs, to test whether the high HCO + /CO ratio can be a good tracer of SNR shock in various physical conditions.If it is the case, HCO + lines will be a useful tool to identify new SNR-MC associations on the Galactic plane.
Broadened 12 CO emission has been widely used to probe shocked MCs and many known SNR-MC interaction has been identified in this way (Jiang et al. 2010).However, 12 CO emission is not a shock tracer, but is usually used as a molecular gas tracer since CO is an abundant molecule in the interstellar MCs and easy to be excited.In the V LSR range of 0-80 km s −1 , the 12 CO J=1-0 and J=2-1 emissions toward W49B are subject to the severe line crowding, because multiple quiescent MCs lie along the line of sight.This causes problems to discern any broad profiles.Although the line of sight contamination should be free at V LSR > 80 km s −1 (see Figure 2), we still do not detect strong broad 12 CO emission across the SNR, probably because of the efficient destruction of 12 CO molecules in the shocks (see Figure 8).
Finally, we observed high HCO + /CO intensity ratios outside W49B, which is likely due to the strong Xray/UV radiation.It is also of interest to test whether such high ratios can be found outside other SNRs (or other X-ray/UV luminous sources) in the molecular environment.Nevertheless, the ratio should be the subject to multiple physical parameters (such as temperature, density, and optical depth), in addition to radiation fields.Further modelings of the MC chemistry in strong radiation are needed for comparison.

CONCLUSION
We have performed a molecular line survey toward SNR W49B using the IRAM 30 m telescope.In this paper, we focus on HCO + and 12 CO emissions, to probe the influence of SNR on the MCs through shocks, CRs, and radiation.Our main conclusions are as follows: 1. W49B is interacting with MCs in its southwest, where we found very broad HCO + emissions with widths of 48-75 km s −1 .The observed flux ratios of HCO + /CO in two shocked regions are 1.1 ± 0.4 and 0.70 ± 0.16, which are significantly larger than that in typical MCs.In the LTE condition and optically thin case, the abundance ratios are estimated as X(HCO + )/X(CO) = 7 × 10 −4 -2 × 10 −3 .We have also found an enhanced flux ratio of HCO + /HCN of 1.9-2.6 in the shocked gas, larger than the ratio of ∼ 1 near the eastern H 2 filament.
2. By comparing with the MHD shock models, we suggest that the high HCO + /CO ratio in the broad-line regions can result from a CR-induced chemistry in shocked MCs, where the CR ionization rate is enhanced to 1-2 orders of magnitude larger than the Galactic level.The observed high HCO + /CO ratio in the broad lines does not mean an enhancement of HCO + but reflects a stronger decline of the CO abundance than that of HCO + .
3. Our molecular study supports the systemic velocity of W49B is V LSR ∼ 61-65 km s −1 and the distance is 7.9 ± 0.6 kpc, consistent with that obtained using the near-IR H 2 emission (Lee et al. 2020).Therefore, W49B does not have a physical relation to the star-forming region W49A.
4. SNR W49B also influences the molecular chemistry far ahead of its shocks.We have found an unusually high HCO + /CO intensity ratio (≥ 0.2) over 1 pc east of the SNR's radio boundary, where the 12 CO emission is optically thin.The HCO + lines outside the SNR are narrow (dV = 2-4 km s −1 ), disfavoring an SNR shock perturbation.The characteristic lengths of low-energy CRs and magnetic precursors of W49B are much smaller than 1 pc, and thus are unlikely to cause the enhanced ratio while the radiation precursor (X-ray or UV) is capable to change the molecular chemistry outside the SNR.
5. As the most luminous Galactic SNR in the X-ray band, W49B is a potential laboratory to study the XDRs.The XDR models can explain the low H 2 2-1 S(1)/ 1-0 S(1) ∼ 0.1 ratio in the H 2 filament east to the SNR (Lee et al. 2020), while PDR models predict larger values.Moreover, we found that the HCO + structure is east of the H 2 filament.This layered distribution can be explained under XDR models, which predict that H 2 emission arises from the surface layer of the irradiated MCs.The layered HCO + and H 2 distribution likely reflect a radiation attenuation into the MCs.Although some XDR models can explain the observed HCO + /CO ratio, we caution that future models and observations are needed for a reliable comparison.
6.The HCO + /CO ratio is a potentially useful tool to study the influence of SNR shocks, CRs and radiation on the MC chemistry.The low-J 12 CO emission alone does not seem to be a good probe for shock conditions near W49B on the Galactic plane, as it is weak in the shocked gas, but too strong in unassociated clouds along the line of sight.

Figure 2 .
Figure 2. Spectra of 12 CO and scaled HCO + and HCN at six regions labeled in Figure 1.The dashed vertical line indicates the LSR velocity of 64 km s −1 , where a high HCO + /CO line ratio is found outside the SNR.All data were convolved to reach the same beam size of 30 .

Figure 3 .
Figure 3. Distribution of the shocked HCO + gas in the velocity range VLSR = 70 km s −1 to 120 km s −1 , overlaid with the 327 MHz radio contours.Values with ≥ 5σ detection are shown.

Figure 5 .
Figure5.Brightness distribution of H2 2.12 µm, HCO + , and 12 CO, and 327 MHz radio emission cut cross the SNR eastern shell from the east to the west (angle=30 • ; see the region in Figure1).

Figure 7 .
Figure 7.The abundance ratio of HCO + /CO as a function of entrance shock velocity, preshock gas density nH, and CR ionization rate per H2 ζ in the MHD shock models.The shock age is taken as 1 kyr.The observed values in regions B and C are denoted by the gray area.

Figure 8 .
Figure8.The abundance ratio of 12 CO (left) and HCO + (right) relative to those in the preshock gas for a shock age of 1 kyr.

Figure 9 .
Figure 9. Schematic view of the interpreted positions of the six regions A-F denoted in Figure 1.The underlined letters denote the regions behind W49B through the observer view or western side view.

Figure 10 .Figure 11 .
Figure 10.Channel map of the averaged main-beam temperature of 12 CO J=2-1 emission with a step of 5 km s −1 , overlaid with contours of 327 MHz radio emission.The same contours are used for the maps in the rest of the paper.The top left panel shows the spectrum averaged over the FOV.

Table 1 .
Information for the IRAM 30m observation aThe rms range of the data cube.The minimum rms is taken from the eastern and western regions of the SNRs.

Table 2 .
Parameters of the HCO + and HCN emission in six regions Region This work is based on observations carried out under project numbers 167-18 and 024-20 with the IRAM 30m telescope.IRAM is supported by INSU/CNRS (France), MPG (Germany) and IGN (Spain).Shock models published in this paper have been produced with the Paris-Durham shock code (Flower & Pineau des Forêts 2003; Lesaffre et al. 2013; Godard et al. 2019, http://ism.obspm.fr).P.Z.acknowledges the support from NSFC grant No. 11590781, Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) Veni Fellowship, grant no.639.041.647, and the Nederlandse Onderzoekschool Voor Astronomie (NOVA).M.A. thanks the NWO for support via the Talent Programme Veni grant.The work of B.-C.K. is supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT and Future Planning (2020R1A2B5B01001994).Z.-Y.Z.acknowledges the support of NSFC (grants 12041305, 12173016) and the Program for Innovative Talents and Entrepreneurs in Jiangsu.Y.C. thanks the support of NSFC grants 11773014, 11633007, and 11851305.