The detection of a hot molecular core in the extreme outer Galaxy

The target star-forming region is WB89-789 (Brand & Wouterloot, 1994). The region contains three Class I protostar candidates identified by near-infrared observations (Brand & Wouterloot, 2007), and one of them is a main target of the present ALMA observations. The region observed with ALMA is indicated on a near-infrared two-color image shown in Figure 1. The observed position is notably reddened compared with other parts of WB89-789.

Observations were conducted with ALMA in 2018 and 2019 as a part of the Cycle 5 (2017.1.01002.S) and Cycle 6 (2018.1.00627.S) programs (PI: T. Shimonishi). A summary of the present observations is shown in Table 1. The pointing center of antennas is RA = 06^h17^m23^s and Dec = 14$\arcdeg$54$\arcmin$41$\arcsec$ (ICRS). The total on-source integration time is 115.5 minutes for Band 6 data and 64.1 minutes for Band 7. Flux and bandpass calibrators are J0510+1800, J0854+2006, and J0725-0054 for Band 6, while J0854+2006 and J0510+1800 for Band 7, respectively. Phase calibrators are J0631+2020 and J0613+1708 for Band 6 and J0643+0857 and J0359+1433 for Band 7. Four spectral windows are used to cover the sky frequencies of 241.40–243.31, 243.76-245.66, 256.90–258.81, and 258.76–260.66 GHz for Band 6, while 337.22–339.15, 339.03-340.96, 349.12–351.05, and 350.92–352.85 GHz for Band 7. The channel spacing is 0.98 MHz, which corresponds to 1.2 km s^-1 for Band 6 and 0.85 km s^-1 for Band 7. The total number of antennas is 45–49 for Band 6 and 43–44 for Band 7. The minimum–maximum baseline lengths are 15.1–783.5 m for Band 6 and 15.1–500.2 m for Band 7. A full-width at half-maximum (FWHM) of the primary beam is about 25$\arcsec$ for Band 6 and 18$\arcsec$ for Band 7.

Raw data is processed with the Common Astronomy Software Applications (CASA) package. We use CASA 5.4.0 (Band 6) and 5.1.1 (Band 7) for the calibration and CASA 5.5.0 for the imaging. The synthesized beam sizes of 0$\farcs$39–0$\farcs$42 $\times$ 0$\farcs$49–0$\farcs$52 with a position angle of -36 degree for Band 6 and 0$\farcs$45–0$\farcs$46 $\times$ 0$\farcs$51–0$\farcs$52 with a position angle of - 54 degree for Band 7 are achieved with the Briggs weighting and the robustness parameter of 0.5. In this paper, we use a common circular restoring beam size of 0$\farcs$50, which corresponds to 0.026 pc (5350 au) at the distance of WB89-789. The synthesized images are corrected for the primary beam pattern using the impbcor task in CASA. The continuum image is constructed by selecting line-free channels. Before the spectral extraction, the continuum emission is subtracted from the spectral data using the CASA’s uvcontsub task.

The spectra and continuum flux are extracted from the 0$\farcs$50 diameter circular region centered at RA = 06^h17^m24$\fs$073 and Dec = 14$\arcdeg$54$\arcmin$42$\farcs$27 (ICRS), which corresponds to the submillimeter continuum center of the target and is equivalent to the hot core position. Hereafter, the source is referred to as WB89-789 SMM1.

Figures 2–3 show submillimeter spectra extracted from the continuum center of WB89-789 SMM1. Spectral lines are identified with the aid of the Cologne Database for Molecular Spectroscopy¹¹1https://www.astro.uni-koeln.de/cdms (CDMS, Müller et al., 2001, 2005) and the molecular database of the Jet Propulsion Laboratory²²2http://spec.jpl.nasa.gov (JPL, Pickett et al., 1998).

Line parameters are measured by fitting a Gaussian profile to detected lines. We estimate the peak brightness temperature, the FWHM, the LSR velocity, and the integrated intensity for each line based on the fitting. For spectral lines for which a Gaussian profile does not fit well, their integrated intensities are calculated by directly integrating the spectrum over the frequency region of emission. Full details of the line fitting can be found in Appendix A (Tables of measured line parameters) and Appendix B (Figures of fitted spectra). The table also contains the estimated upper limits on important non-detection lines.

A variety of carbon-, oxygen-, nitrogen-, sulfur-, and silicon-bearing species, including COMs containing up to nine atoms, are detected from WB89-789 SMM1 (see Table 2). Multiple high excitation lines (upper state energy $>$100 K) are detected for many species. Measured line widths are typically 3–6 km s^-1. Most of lines consist of a single velocity component, but SiO has doppler shifted components at $V_{sys}$ $\pm$ 5 km s^-1 as indicated in Figure 14 in Appendix B.

Figures 4–5 show synthesized images of continuum and molecular emission lines observed toward the target region. The images are constructed by integrating spectral data in the velocity range where the emission is detected. Most molecular lines, except for those of molecular radicals CN, CCH, and NO, have their intensity peak at the continuum center, which corresponds to the position of a hot core. Simple molecules such as H¹³CO⁺, H¹³CN, CS, and SO are extended compared to the beam size. Secondary intensity peaks are also seen in those species. Complex molecules and HDO are concentrated at the hot core position. A characteristic symmetric distribution is seen in SiO. Further discussion about the distribution of the observed emission is presented in Section 4.2.

The nature of WB89-789 SMM1 is characterized as (i) the compact distribution of warm gas ($\sim$0.03 pc, see Section 4.2), (ii) the high gas temperature that can trigger the ice sublimation ($\geq$100 K, Section 3.3.1), (iii) the high density (2 $\times$ 10⁷ cm^-3, Section 3.3.3), (iv) the association with a luminous protostar (see below), and (v) the presence of chemically rich molecular gas. Those properties suggest that the source is associated with a hot molecular core.

Figure 7 shows a SED of SMM1, where the data are collected from available databases and literatures (Brand & Wouterloot, 2007; Wright et al., 2010; Yamamura et al., 2010). The bolometric luminosity of the source is estimated to be 8.4 $\times$ 10³ L_☉ based on the SED fitting with the model of Robitaille et al. (2007). This luminosity is equivalent to a stellar mass of about 10 M_☉ according to the mass-luminosity relationship of zero age main sequence (ZAMS) stars (Zinnecker & Yorke, 2007).

Note that far-infrared data, which is important for the luminosity determination of embedded sources, is insufficient for SMM1. Only upper limits are provided due to the low angular resolution of available AKARI FIS all-sky survey data. Thus the derived luminosity (and therefore mass) may be lower than the current estimate. Future high spatial resolution infrared observations in those missing wavelengths are highly required.

Alternatively, we can estimate the luminosity of SMM1 by scaling the luminosity of a low-metallicity LMC hot core, ST16, whose SED is well determined based on a comprehensive infrared dataset from 1 to 1200 $\mu$m (Shimonishi et al., 2020). This LMC hot core has a total luminosity of 3.1 $\times$ 10⁵ L_☉ and a $K_{s}$-band magnitude ([$K_{s}$]) of 13.4 mag at 50 kpc, while SMM1 has [$K_{s}$] = 14.7 mag at 10.7 kpc. Scaling the luminosity of ST16 with the distance and $K_{s}$-band magnitude, we obtain 4.3 $\times$ 10³ L_☉ for SMM1, which is a factor of two lower than the estimate by the SED fitting. In either case, present estimates suggest that the luminosity of SMM1 would correspond to the lower-end of high-mass ZAMS or upper-end of intermediate-mass ZAMS.

The observed emission lines and continuum show different spatial distributions depending on species. Those distributions have important clues to understand their origins. A schematic illustration of the temperature structure and molecular gas distribution in WB89-789 SMM1 are shown in Figure 8 based on the discussion in this section.

We have estimated the spatial extent of observed emission by fitting a two-dimensional Gaussian to the continuum center (Table 3). Compact distributions (FWHM = 0$\farcs$5–0$\farcs$6, 0.026–0.031 pc), that is comparable with the beam size, are seen in HDO, COMs, CH₃CN, HNCO, OCS, and high-excitation SO₂ lines. HC₃N is slightly extended (FWHM = 0$\farcs$65). They are concentrated at the hot core position, suggesting that they are originated from a warm region where ice mantles are sublimated.

SO, ³⁴SO, ³³SO, and low-excitation SO₂ show relatively compact distributions (FWHM = 0$\farcs$5–0$\farcs$7, 0.026–0.036 pc) at the hot core position, but also show a secondary peak at the south side of the hot core. This secondary peak coincides with the peak of the NO emission. Other sulfur-bearing species such as C³⁴S, C³³S, and H₂CS show compact distributions (FWHM = 0$\farcs$6–0$\farcs$0.7, 0.031–0.052 pc) centered at the hot core.

A characteristic distribution that is symmetric to the hot core position is seen in SiO. It shows a compact emission (FWHM = 0$\farcs$65) at the hot core center, but also shows other peaks at the north-east and south-west sides of the hot core. Those secondary peaks are slightly elongated. SiO is a well-known shock tracer. The observed structure would be originated from the shocked gas produced by bipolar protostellar outflows. A driving source of the outflows would be a protostar embedded in a hot core, since the distribution of SiO is symmetric to the hot core position.

Even extended distributions (FWHM $>$ 1$\farcs$0) are seen in CN, CCH, H¹³CO⁺, HC¹⁸O⁺, H¹³CN, HC¹⁵N, NO, CS, H₂CO, and HDCO, D₂CO, and low-excitation CH₃OH. Gas-phase reactions and non-thermal desorption of icy species would have non-negligible contribution to the formation of those species, because they are widely distributed beyond the hot core. We note that dust continuum, H¹³CN, HC¹⁵N have a moderately sharp peak (FWHM $<$ 1$\farcs$0) at the hot core position in addition to the extended component. c-C₃H₂ shows a patchy distribution, whose secondary peak at the south-west of the hot core does not coincide with those of other species.

Molecular radicals (CN, CCH, and NO) do not have their emission peak at the hot core position. This would suggest that the chemistry outside the hot core region largely contributes to their production. CN and CCH are known to be abundant in photodissociation regions (PDRs), because atomic carbon is efficiently provided by photodissociation of CO under moderate UV fields (e.g., Fuente et al., 1993; Sternberg & Dalgarno, 1995; Jansen et al., 1995; Rodriguez-Franco et al., 1998; Pety et al., 2017). In the present source, their emission shows a similar spatial distribution. A similar distribution between CN and CCH has been also observed in a LMC hot core (Shimonishi et al., 2020); they argue that CN and CCH would trace PDR-like outflow cavity structures that are irradiated by the UV light from a protostar associated with a hot core. We speculate that this is also the case for WB89-789 SMM1.

Figure 9 shows velocity maps (moment 1) of CN and CCH lines. CN and CCH emission are elongated in the south-west direction from the hot core (see also Figure 4). The figure also shows a possible direction of protostellar outflows expected from the spatial distribution of SiO. The elongated directions of CN and CCH coincide with the inferred direction of outflows. In addition, the elongated south-west parts of CN and CCH are blue-shifted by $\sim$1–2 km s^-1 compared to the hot core position. This may be due to outflow gas motion, although CN and CCH would trace an outflow cavity wall rather than outflow gas itself. Actually the observed velocity shift is smaller than a typical value of high-velocity wing components in massive protostellar outflows ($\geq$ 5 km s^-1, e.g., Beuther et al., 2002; Maud et al., 2015). Future observations of optically-thick outflow tracers such as CO are necessary to confirm the presence of high-velocity gas associated with protostellar outflows.

Figure 10 shows a comparison of molecular abundances between WB89-789 SMM1 and other known Galactic hot cores. The data for an intermediate-mass hot core, NGC7192 FIRS2, is adopted from Fuente et al. (2014). The abundances are based on the 220 GHz region observations for a 0.009 pc diameter area centered at the hot core. The luminosity of NGC7192 FIRS2 ($\sim$500 L_☉) corresponds to that of a 5 M_☉ ZAMS. The data for a high-mass source, the Orion hot core, is adopted from Sutton et al. (1995), which is based on the 340 GHz region observations for a 0.027 pc diameter area at the hot core. The abundance of HNCO is taken from Schutte & Greenberg (1997).

The molecular abundances in WB89-789 SMM1 is generally lower than those of inner Galactic counterparts. The degree of the abundance decrease is roughly consistent with the lower metallicity of the WB89-789 region as indicated by the scale bar in Figure 10. Particularly, SMM1 and the intermediate-mass hot core NGC7192 FIRS2 show similar molecular abundances after taking into account the four times lower metallicity of the former source. For the comparison with Orion, it seems that HC₃N, C₂H₅CN, and SO₂ are significantly less abundant in SMM1 even taking into account the lower metallicity, while CH₃OH is overabundant in SMM1 despite the low metallicity.

To further focus on chemical complexity at low metallicity, Figure 11 shows a comparison of fractional abundances of COMs normalized by the CH₃OH column density for WB89-789 SMM1 and NGC7192 FIRS2. Such a comparison is useful for investigating chemistry of organic molecules in warm and dense gas around protostars (Herbst & van Dishoeck, 2009; Drozdovskaya et al., 2019), because CH₃OH is believed to be a parental molecule for the formation of even larger COMs (e.g., Nomura & Millar, 2004; Garrod & Herbst, 2006). In addition, CH₃OH is a product of grain surface reaction, thus warm CH₃OH gas mainly arise from a high-temperature region, where ices are sublimated and characteristic hot core chemistry proceeds. Furthermore, the normalization by CH₃OH can cancel the metallicity effect in the abundance comparison.

The $N$(X)/$N$(CH₃OH) ratios are remarkably similar between WB89-789 SMM1 and NGC7192 FIRS2 as shown in Figure 11 (a). The ratios of SMM1 coincide with those of NGC7192 FIRS2 within a factor of 2 for the most molecular species. The correlation coefficient is calculated to be 0.94, while it is 0.96 if CH₃CN is excluded. It seems that CH₃CN deviates from the overall trend, although the uncertainty is large due to the opacity effect (see 3.3.4). C₂H₅OH also slightly deviates from the trend. The reason for their behavior is still unclear, but it may be related to the formation pathway of those molecules.

The above two comparisons suggest that chemical compositions of the hot core in the extreme outer Galaxy scale with the metallicity. In the WB89-789 region, the metallicity is expected to be four times lower compared to the solar neighborhood. The observed abundances of COMs in the SMM1 hot core is lower than the other Galactic hot cores, but the decrease is proportional to this metallicity. Furthermore, similar $N$(COMs)/$N$(CH₃OH) ratios suggest that CH₃OH is an important parental species for the formation of larger COMs in a hot core, as suggested by aforementioned theoretical studies.

CH₃OH ice is believed to form on grain surfaces and several formation processes are proposed by laboratory experiments; i.e., hydrogenation of CO, ultraviolet photolysis and radiolysis of ice mixtures (e.g., Hudson & Moore, 1999; Watanabe et al., 2007). It is known that CH₃OH is already formed in quiescent prestellar cores before star formation occurs (Boogert et al., 2011). Solid CH₃OH will chemically evolve to larger COMs by a combination of photolysis, radiolysis, and grain heating during the warm-up phase that leads to the formation of a hot core (Garrod & Herbst, 2006). High-temperature gas-phase chemistry of sublimated CH₃OH would also contribute to the COMs formation (Nomura & Millar, 2004; Taquet et al., 2016). The present results suggest that various COMs can form even in a low-metallicity environment, if their parental molecule, CH₃OH, is efficiently produced in a star-forming core. We here note that observations of ice mantle compositions are not reported for the outer Galaxy so far . Future infrared observations of ice absorption bands towards embedded sources in the outer Galaxy are important.

It is still unknown if the observed simply-metallicity-scaled chemistry of COMs in the WB89-789 SMM1 hot core is common in other hot core sources in the outer Galaxy. A comparison of the present data with those of hot cores in the LMC would provide a hint for understanding the universality of low-metallicity hot core chemistry. The metallicity of the LMC is reported to be lower than the solar value by a factor of two to three (e.g., Dufour et al., 1982; Westerlund, 1990; Russell & Dopita, 1992; Choudhury et al., 2016), which is in common with the outer Galaxy.

Figure 12 shows a comparison of molecular abundances between WB89-789 SMM1 and three LMC hot cores. The plotted molecular column densities for LMC hot cores are adopted from Shimonishi et al. (2016a) for ST11, Shimonishi et al. (2020) for ST16, and Sewiło et al. (2018) fro N113 A1. Another LMC hot core in Sewiło et al. (2018), N113 B3, have similar molecular abundances with those of N113 A1. The $N_{\mathrm{H_{2}}}$ value of ST11 and N113 A1 is re-estimated using the same dust opacity data and dust temperature ($T_{d}$ = 60 K) as in this work; We obtained $N_{\mathrm{H_{2}}}$ = 1.2 $\times$ 10²⁴ cm^-2 for ST11 and $N_{\mathrm{H_{2}}}$ = 9.2 $\times$ 10²³ cm^-2 for N113 A1. The dust temperature assumed in ST16 is 60 K as described in Section 3.3.3. Molecular column densities are estimated for circular/elliptical regions of 0.12 $\times$ 0.12 pc, 0.10 $\times$ 0.10 pc, and 0.21 $\times$ 0.13 pc for ST11, ST16, and N113 A1, respectively. For a fair comparison, we have re-calculated $N_{\mathrm{H_{2}}}$ and molecular column densities of SMM1 for a 0.1 pc (1$\farcs$93) diameter region. Those abundances are plotted in Figure 12 and summarized in Table 4.

The chemical composition of the outer Galaxy hot core does not resemble those of LMC hot cores as seen in Figure 12. The dissimilarity is also seen in the $N$(X)/$N$(CH₃OH) comparison between SMM1 and ST16 as shown in Figure 11 (b), where the correlation coefficient is calculated to be 0.69.

Shimonishi et al. (2020) argue that SO₂ will be a good tracer of low-metallicity hot core chemistry, because (i) it is commonly detected in LMC hot cores with similar abundances, and (ii) it is originated from a compact hot core region. SO also shows similar abundances within LMC hot cores. In WB89-789 SMM1, however, the abundances of SO₂ and SO relative to H₂ are lower by a factor of 28 and 5 compared with LMC hot cores. The measured rotation temperatures of SO₂ are similar between those hot cores, i.e., 166 K (SO₂) for SMM1, 232 K (SO₂) and 86 K (³⁴SO₂) for ST16, 190 K (SO₂) and 95 K (³⁴SO₂) for ST11. The SO₂ column densities for ST16 and ST11 are estimated from ³⁴SO₂, while that for SMM1 is from SO₂. However, the SO₂ column density of SMM1 increases only by a factor of up to three when it is estimated from ³⁴SO₂ (see Section 3.3.4). Thus the low SO₂ abundance in the outer Galactic hot core would not be due to the optical thickness.

In contrast to the S-O bond bearing species, the C-S bond bearing species such as CS, H₂CS , and OCS do not show significant abundance decrease in WB89-789 SMM1. Thus it is not straightforward to attribute the low abundance of SO₂ (and perhaps SO) to the low elemental abundance ratio of sulfur in the outer Galaxy. Hot core chemistry models suggest that SO₂ is mainly produced in high-temperature gas-phase reactions in warm gas, using H₂S sublimated from ice mantles (Charnley, 1997; Nomura & Millar, 2004). This also applies to the SO₂ formation in low-metallicity sources as shown in astrochemical simulations for LMC hot cores (Shimonishi et al., 2020). We speculate that the different behavior of SO₂ in outer Galaxy and LMC hot cores may be related to differences in the evolutionary stage of hot cores. A different luminosity of host protostars may also contribute to the different sulfur chemistry; i.e., $\sim$8 $\times$ 10³ L_☉ for WB89-789-SMM1, while several $\times$ 10⁵ L_☉ for LMC hot cores. A different cosmic-ray ionization rate between the outer Galaxy and the LMC may also affect the chemical evolution, although the rate is not known for the outer Galaxy.

Among nitrogen-bearing molecules, NO shows interesting behavior in LMC hot cores. After corrected for the metallicity, NO is overabundant in LMC hot cores compared with Galactic counterparts despite the low elemental abundance of nitrogen in the LMC (Shimonishi et al., 2020). Only NO shows such behavior among the nitrogen-bearing molecules observed in LMC hot cores. In WB89-789 SMM1, however, such an overabundance of NO is not observed. The NO abundance of SMM1 is 1.6 $\times$ 10^-9 for a 0.1 pc region data. This is a factor of five lower than a typical NO abundance in Galactic high-mass hot cores (8 $\times$ 10^-9, Ziurys et al., 1991), which is consistent with a factor of four lower metallicity in WB89-789. The present high-spatial resolution data have revealed that NO does not mainly arise from a hot core region, as shown in Figure 4. It has an intensity peak at the south part of the hot core, where low-excitation lines of SO and SO₂ also have a secondary peak (Section 4.2). Thus, shock chemistry or photochemistry, rather than high-temperature chemistry, would contribute to the production of NO in low-metallicity protostellar cores. In that case, a lower luminosity of SMM1 than those of LMC hot cores may contribute to the different behavior of NO.

For other nitrogen-bearing molecules, HNCO and CH₃CN, a clear difference is not identified between outer Galactic and LMC hot cores, although the number of data points is limited and the abundance uncertainty is large. The reason of the unusually low abundance of SiO in SMM1 is unknown. It may be related to different shock conditions or grain compositions, because dust sputtering by shock is mainly responsible for the production of SiO gas.

Formation of COMs is one of the important standpoints for low-metallicity hot core chemistry. It is reported that CH₃OH show a large abundance variation in LMC hot cores (Shimonishi et al., 2020). There are organic-poor hot cores such as ST11 and ST16, while N113 A1 and B3 are organic-rich. The CH₃OH abundance of WB89-789 SMM1 is higher than those of any known LMC hot cores. The abundances of HCOOCH₃ and CH₃OCH₃ in SMM1 are comparable with those of an organic-rich LMC hot core, N113 A1. Th detection of many other COMs in SMM1 suggests the source have experienced rich organic chemistry despite its low-metallicity nature.

Astrochemical simulations for LMC hot cores suggest that dust temperature at the initial ice-forming stage have a major effect on the abundance of CH₃OH gas in the subsequent hot core stage (Acharyya & Herbst, 2018; Shimonishi et al., 2020). Simulations of grain surface chemistry dedicated to the LMC environment also suggest that dust temperature is one of the key parameters for the formation of CH₃OH in dense cores (Acharyya & Herbst, 2015; Pauly & Garrod, 2018). This is because (i) CH₃OH is mainly formed by the grain surface reaction, and (ii) the hydrogenation reaction of CO, which is a dominant pathway for the CH₃OH formation, is sensitive to the dust temperature due to the high volatility of atomic hydrogen. For this reason, it is inferred that organic-rich hot cores had experienced a cold stage ($\lesssim$ 10K) that is sufficient for the CH₃OH formation before the hot core stage, while organic-poor ones might have missed such a condition for some reason. Alternatively, the slight difference in the hot core’s evolutionary stage may contribute to the CH₃OH abundance variation, because the high-temperature gas-phase chemistry is rapid and it can decrease CH₃OH gas at a late stage (e.g., Nomura & Millar, 2004; Garrod & Herbst, 2006; Vasyunin & Herbst, 2013; Balucani et al., 2015).

Low-metallicity hot core chemistry simulations in Shimonishi et al. (2020) argue that the maximum achievable abundances of CH₃OH gas in a hot core stage significantly decrease as the visual extinction of the initial ice-forming stage decreases. On the other hand, the simulations show that the CH₃OH gas abundance is simply metallicity-scaled if the initial ice-forming stage is sufficiently shielded. In a well-shielded initial condition, the grain surface is cold enough to trigger the CO hydrogenation, and the resultant CH₃OH abundance is roughly regulated by the elemental abundances. The observed metallicity-scaled chemistry of COMs in WB89-789 SMM1 implies that the source had experienced such an initial condition before the hot core stage.

Deuterium chemistry is widely used in interpreting chemical and physical history of interstellar molecules (e.g., Caselli & Ceccarelli, 2012; Ceccarelli et al., 2014). The measured CH₂DOH/CH₃OH ratio in WB89-789 SMM1 is 1.1 $\pm$ 0.2 $\%$, which is comparable to the higher end of those ratios observed in high-mass protostars and the lower end of those in low-mass protostars (e.g., see Fig.2 in Drozdovskaya et al., 2021). The ratio is orders of magnitude higher than the deuterium-to-hydrogen ratio in the solar neighborhood (2 $\times$ 10^-5; Linsky et al., 2006; Prodanović et al., 2010) and that in the big-bang nucleosynthesis (3 $\times$ 10^-5; Burles, 2002, references therein). This suggests that the efficient deuterium fractionation occurred upon the formation of CH₃OH in SMM1. The D₂CO/HDCO ratio is 45 $\pm$ 10 $\%$, which is comparable to those observed in low-mass and high-mass protostars (e.g., Zahorecz et al., 2021). This would suggest that physical conditions for deuterium fractionation could be similar between WB89-789 SMM1 and inner Galactic protostars. Note that higher spatial resolution observations and detailed multiline analyses would affect the measured abundance of deuterated species as reported in Persson et al. (2018) for the case of a nearby low-mass protostar. The H₂CO column density derived in this work may be a lower limit because the line is often optically thick, thus we do not discuss the abundance ratio relative to H₂CO.

It is known that the deuterium fractionation efficiently proceeds at low temperature (e.g., Roberts et al., 2003; Caselli & Ceccarelli, 2012; Taquet et al., 2014; Furuya et al., 2016). This is because the key reaction for the trigger of deuterium fractionation, H${}^{+}_{3}$ + HD $\rightarrow$ H₂D⁺ + H₂ + 232 K, is exothermic and its backward reaction cannot proceed below 20 K. In addition, gaseous neutral species such as CO and O efficiently destruct H₂D⁺, thus their depletion at low temperature further enhances the deuterium fractionation (e.g., Caselli & Ceccarelli, 2012). A sign of high deuterium fractionation observed in WB89-789 SMM1 suggests that the source had experienced such a cold environment during its formation. This picture is consistent with the implication obtained from the metallicity-scaled chemistry of COMs, which also suggests the occurrence of a cold and well-shielded initial condition as discussed above.

Although the low metallicity is common between the outer Galaxy and the LMC, their star-forming environments would be different; the LMC has more harsh environments as inferred from active massive star formation over the whole galaxy, while that for the outer Galaxy might be quiescent due to its low star formation activity. Those environmental differences need to be taken into account for further understanding of the chemical evolution of star-forming regions at low metallicity. Future extensive survey of protostellar objects towards the outer Galaxy is thus vitally important for further discussion. Astrochemical simulations dedicated to the environment of the outer Galaxy, and the application to lower-mass protostars, are also important.

We have also detected a compact source associated with high-velocity SiO gas at the east side of WB89-789 SMM1. Hereafter, we refer to this source as WB89-789-SMM2. According to the SiO emission, the source is located at RA = 06^h17^m24$\fs$246 and Dec = 14^∘54$\arcmin$43$\farcs$25 (ICRS), which is 2$\farcs$7 (0.14 pc) away from SMM2. Figure 13(a) shows the SiO(6-5) spectrum extracted from a 0$\farcs$6 diameter region centered at the above position. The SiO line is largely shifted to the blue and red sides relative to the systemic velocity in a symmetric fashion. The peaks of the shifted emission are located at $V_{sys}$ $\pm$ 25 km s^-1.

Figure 13(b) shows a velocity map and integrated intensity distribution of SiO(6-5). In the figure, to focus on SiO in WB89-789-SMM2, the intensity is integrated over much wider velocity range (0–60 km s^-1) compared with that adopted in Figure 4 (31–38 km s^-1). The velocity map clearly indicates that the velocity structure of SiO in SMM2 is spatially symmetric to the SiO center. At this position, a local peak is seen in 1200 $\mu$m continuum as shown in the figure, suggesting the presence of an embedded source. SMM2 does not show any emission lines of COMs, and no alternative molecular lines are identified at the frequencies of doppler-shifted SiO emission. Also taking into account the clear spectral and spatial symmetry, the observed lines must be attributed to high-velocity SiO gas.

The spectral characteristics of the observed high-velocity SiO resemble those of extremely high velocity (EHV) outflows observed in Class 0 protostars (Bachiller et al., 1991; Tafalla et al., 2010, 2015; Tychoniec et al., 2019). The EHV flows are known to appear as a discrete high-velocity ($V$ $\gtrsim$30 km s^-1) peak, and observed in the youngest stage of star formation (Bachiller, 1996; Matsushita et al., 2019, references therein). The EHV flows extends up to several thousands au from the central protostar in SiO, and usually have collimated bipolar structures (e.g., Bachiller et al., 1991; Hirano et al., 2010; Tychoniec et al., 2019; Matsushita et al., 2019). The beam size of the present data is about 5000 au, thus such structures will not be spatially resolved. Actually, a symmetric spatial distribution of blue-/red-shifted SiO is only marginally resolved into two beam size regions (Fig. 13(b)). A spatial extent of SiO emission is about 1$\arcsec$ (0.052 pc). Assuming an outflow velocity of 25 km s^-1, we estimate a dynamical timescale of EHV flows to be at least 2000 years. This is roughly consistent with dynamical timescales of other EHV sources, which range from a few hundred to a few thousand years (Bachiller, 1996, references therein).

A 1200 $\mu$m continuum flux in a 0$\farcs$6 diameter region centered at SMM2 is 0.60 $\pm$ 0.05 mJy/beam. Assuming $T_{d}$ = 20 K, we obtain $N_{\mathrm{H_{2}}}$ = 3.2 $\times$ 10²³ cm^-2. This is equivalent to a gas number density of $n_{\mathrm{H_{2}}}$ = 4.9 $\times$ 10⁶ cm^-3. If we assume a higher $T_{d}$, i.e. 40 K, then the derived column density is 2.5 times lower than the 20 K case. In either case, the continuum data suggests the presence of high-density gas at this position.

Previous single-dish observations of CO detected extended ($\sim$20$\arcsec$) molecular outflows in the WB89-789 region (Brand & Wouterloot, 1994, 2007). The center of the outflow gas coincides with the position of the IRAS source (IRAS 06145+1455; 06^h17^m24$\fs$2, 14$\arcdeg$54$\arcmin$42$\arcsec$, J2000). This position is consistent with those of SMM1 or SMM2, given the large beam size of CO(3-2) observations (14$\arcsec$) in Brand & Wouterloot (2007). The observed CO outflow gas has an extended blue-shifted component (20 $<$ $V_{LSR}$ $<$ 31 km $s^{-1}$) towards the south-east direction from the center, while a red-shifted component (37 $<$ $V_{LSR}$ $<$ 55 km $s^{-1}$) is extended towards the north-west direction (see Figure 9 in Brand & Wouterloot, 2007). This outflow direction coincides with that of the high-velocity SiO outflows observed in this work. The SiO outflows from SMM2 may have a common origin with the large-scale CO outflows.

In summary, it is likely that a compact, high-density, and embedded object is located at the position of WB89-789-SMM2. Presumably, a protostar associated with SMM2 is driving the observed high-velocity SiO gas flows. Its short dynamical timescale and similarity with EHV flows suggest that the object is at the youngest stage of star formation (Class 0/I). Non-detection of warm gas emission also supports its young nature. We note that the detailed structure of high-velocity SiO gas is not spatially resolved, and CO lines, which often trace high-velocity outflows, are not covered in the present data. Future high-spatial resolution observations of CO and other outflow tracers are key to further clarify the nature of WB89-789-SMM2.