THE DETECTION OF A HOT MOLECULAR CORE IN THE LARGE MAGELLANIC CLOUD WITH ALMA

A source size of ST11, which is measured by detected emission lines, is consistent with those of Galactic hot cores. The molecular lines of NO, SiO, ³³SO, SO₂, ³⁴SO₂, and ³³SO₂ as well as the dust continuum show a compact source size, which is equal to or lower than the beam size of ∼0.1 pc. The lines of CO, C¹⁷O, HCO⁺, H¹³CO⁺, and H₂CO show somewhat extended distributions, but the dominant emission come from the compact region associated with a high-mass YSO. The extended components of these emission lines may arise from their relatively low critical densities and/or efficient formation routes in gas-phase reactions. Protostellar outflows, which are discussed in Section 5.3, could also affect the spatial distribution of these molecular gas.

The spatial association of the molecular emission lines to the infrared source is another key to the nature of the detected warm and dense molecular gas. The separation between a hot core region and an infrared source is reported to be smaller than ∼0.03 pc for typical Galactic sources (De Buizer et al. 2003). With the present spatial resolution, the distribution of molecular lines and the dust continuum should coincide if they arise from a hot core region. This is clearly observed in the present data as shown in Figure 3.

The H₂ gas density around ST11 is estimated to be >2 × 10⁶ cm⁻³ in Section 4.4. The value is a lower limit and actually clear detections of molecular lines with high critical densities such as those of SO₂ (n_cr ∼ 10⁷ cm⁻³; Williams & Viti 2013) imply an even higher gas density around ST11. Galactic hot cores typically show gas densities between 10⁶–10⁸ cm⁻³ (estimated from Table 1 presented in Kurtz et al. 2000). The gas density around ST11 is thus consistent with those of known hot cores.

Rotational temperatures of molecular gas around ST11 are compared with those of Galactic hot cores in Table 3. The temperature of SO₂, T_rot = 195 ± 5 K, is in good agreement with the temperatures of the three Galactic hot cores in the table, which are between 108 K and 284 K. The temperature of ³⁴SO₂, T_rot = 95 ± 7 K, is somewhat lower than that of SO₂, but this may be attributed to the combined effect of the lack of available higher excitation lines of ³⁴SO₂ and the slightly larger optical thickness of the SO₂ lines. Note that the temperature of ³³SO₂ is even lower, but this is probably because the detected lines are biased to those with low excitation temperatures (E_u < 200 K) due to the limited frequency coverage of the data. The rotation analysis of ³⁴SO₂ and ³³SO₂ for the Orion data using a similar set of lines to those used in the analysis of ST11 results in T_rot = 114 ± 6 K for ³⁴SO₂ and T_rot = 73 ± 5 K for ³³SO₂, which are lower than the values derived by Schilke et al. (1997) using a larger number of lines with a wide range of upper state energies (Table 3). Therefore we conclude that the actual temperature of SO₂ gas is in the order of ∼100 K, which can trigger hot core chemistry via sublimation of ice mantles. Sublimation of ice mantles around ST11 is also suggested from the infrared observations of ices in Shimonishi et al. (2010), which reported that ST11 shows the second weakest ice absorption bands among 12 high-mass YSOs in the LMC.

As estimated in Section 4.5, the bolometric luminosity (5 × 10⁵ L_☉) and stellar mass (50 M_☉) of ST11 are consistent with the properties of high-mass YSOs. This indicates that ST11 can form a hot core region with the help of its intense radiation. In addition, the red SED of ST11 (Figure 9) indicates that it is still in an early evolutionary stage, which is consistent with the properties of hot core sources.

The velocity widths of emission lines from ST11 are typically 4–7 km s⁻¹ except for CO and HCO⁺. This is in good agreement with the velocity widths of emission lines from Galactic hot core regions (typically 4–10 km s⁻¹, e.g., Helmich & van Dishoeck 1997). The systemic velocities of the lines are in a very narrow range, typically 250–251 km s⁻¹, which suggests that the detected molecular species spatially coexist in a small region around a high-mass YSO.

SO₂ is often detected in hot core sources and is one of the useful tracers of warm and dense gas around a high-mass YSO (e.g., Beuther et al. 2009). However, sometimes warm SO₂ gas is also prominently detected in deeply embedded and even younger protostellar objects in which hot cores are not yet formed (e.g., W3 IRS5 in Helmich et al. 1994). The reason for such SO₂ enhancement is suggested to be due to shock dominated chemistry triggered by protostellar outflows. These deeply embedded sources often show a deep self-absorption profile in their CO or HCO⁺ emission lines due to the presence of a significant amount of cold gas in the envelope. As shown in Figure 4, emission from ST11 does not show such a deep self-absorption profile. This suggests that emission from ST11 is mostly dominated by warm gas and that ST11 has already reached the hot core phase.

On the other hand, the relatively weak strengths of the hydrogen recombination lines seen in the near-infrared spectrum of ST11 suggests that a prominent H ii region is not yet formed around the source (see Figure 1 in Shimonishi et al. 2010). Such a transitional evolutionary phase is consistent with the properties of hot cores.

In summary, the compact size of the emitting source, the presence of warm and dense molecular gas around a high-mass protostar, the sublimation of ice mantles, and the detections of rich molecular lines suggest that ST11 is associated with a hot molecular core. This is the first detection of an extragalactic hot molecular core. Note that CH₃OH and complex organic molecules, which are often detected in Galactic hot cores, are not detected in ST11. The reason for the lack of these molecular species is discussed in the next section in the context of the characteristic hot core chemistry in low metallicity.

Carbon monoxide is detected by the CO(3–2) and C¹⁷O(3–2) lines in ST11. The line intensity ratio of CO(3–2)/C¹⁷O(3–2) is about 28, and using ¹⁶O/¹⁷O = 3400 in the LMC (Wang et al. 2009) we estimate optical depths of ∼120 and ∼0.04 for CO(3–2) and C¹⁷O(3–2), respectively. The CO(3–2) line is completely optically thick, while the C¹⁷O(3–2) line is optically thin as in most of Galactic hot cores (e.g., Helmich & van Dishoeck 1997). The abundance of CO in ST11 estimated from C¹⁷O is 9.1 × 10⁻⁵, which is very similar to the average CO abundance of 1.0 × 10⁻⁴ for three Galactic hot cores in Table 5. Elemental carbon and oxygen are about three times less abundant in the LMC; [C/H]_LMC = 1.2 × 10⁻⁴ and [O/H]_LMC = 2.3 × 10⁻⁴ (Korn et al. 2002), while [C/H]_☉ = 3.3 × 10⁻⁴ and [O/H]_☉ = 6.8 × 10⁻⁴ (Grevesse & Sauval 1998). Given the low abundances of carbon and oxygen in the LMC, the CO around ST11 is slightly overproduced as compared with Galactic hot cores.

Formyl ion is detected via the HCO⁺(4–3) and H¹³CO⁺(4–3) lines. The line intensity ratio of HCO⁺(4–3)/H¹³CO⁺(4–3) is about 19, and using ¹²C/¹³C = 49 in the LMC (Wang et al. 2009) we estimate optical depths of ∼2 and ∼0.05 for HCO⁺(4–3) and H¹³CO⁺(4–3), respectively. The HCO⁺(4–3) line is moderately optically thick, while the H¹³CO⁺(4–3) line is optically thin. The abundance of HCO⁺ in ST11 estimated from H¹³CO⁺ is 7.5 × 10⁻¹⁰, which is lower by a factor of ∼3 compared to the average HCO⁺ abundance of 2.0 × 10⁻⁹ for three Galactic hot cores in Table 5.

We here discuss the connection between the observed HCO⁺ abundance and the cosmic-ray ionization rate in the LMC. One of the possible pathways to form HCO⁺ in molecular clouds is the gas-phase reaction:

$$\begin{eqnarray}&&{{\rm{H}}}_{3}^{+}+{\rm{C}}{\rm{O}}\to {{\rm{HCO}}}^{+}+{{\rm{H}}}_{2},\end{eqnarray} \tag{ 8 }$$

in which ${{\rm{H}}}_{3}^{+}$ is formed by ionization of H₂ via cosmic rays and subsequent reaction with H₂ (e.g., Caselli et al. 1998).

We here simply assume that HCO⁺ is a major cation and the electron is a major anion in dense clouds. Given that the production of cations and electrons by the cosmic-ray ionization balances with their recombination, the equilibrium is formulated as

$$\begin{eqnarray}&&\zeta {n}_{{{\rm{H}}}_{2}}={k}_{\mathrm{rec}}{n}_{{\mathrm{HCO}}^{+}}{n}_{e},\end{eqnarray} \tag{ 9 }$$

where ζ is a cosmic-ray ionization rate, k_rec = 2.4 × 10⁻⁷ (300/T)^0.69 cm³ s⁻¹ (5.1 × 10⁻⁷ at T = 100 K) is a dissociative recombination rate of HCO⁺ and electrons (Mitchell 1990), and n indicates the number density of each species. With the assumption that cations and electrons have the the same abundance in neutral clouds, this equation is reformatted as

$$\begin{eqnarray}&&{X}_{{\mathrm{HCO}}^{+}}={X}_{e}=\sqrt{\zeta /{n}_{{{\rm{H}}}_{2}}{k}_{\mathrm{rec}}},\end{eqnarray} \tag{ 10 }$$

where X indicates a fractional abundance relative to H₂. This equation indicates that the abundance of HCO⁺ is proportional to ${\zeta }^{0.5}$ under the above simple assumption.

The cosmic-ray density in the LMC is reported to be 20%–30% of the typical Galactic value of ζ = 3 × 10⁻¹⁷ s⁻¹ based on gamma-ray observations (Abdo et al. 2010). The low cosmic-ray density leads to the low cosmic-ray ionization rate (e.g., Spitzer & Tomasko 1968). According to Equation (10), the HCO⁺ abundance in the LMC is expected to be lower than the Galactic value by a factor of ∼2 for the same H₂ density and gas temperature. This factor is close to the observed difference of the HCO⁺ abundance between ST11 and Galactic sources. Therefore, we speculate that the low cosmic-ray ionization rate in the LMC is one of the factors that contribute to the slightly low abundance of HCO⁺ in ST11.

Note that the above assumption is a simplified picture, i.e., the cosmic-ray ionization rate may vary within the LMC and the electron abundance may differ from that of HCO⁺ due to the presence of other ions. Furthermore, in the above discussion, we consider only one of the possible mechanisms to form HCO⁺. In the circumstellar environment of hot cores, however, outflows and stellar radiation would affect the ionization condition. Previous studies actually argue that molecular outflows and shocks contribute to the enhancement of HCO⁺ in star-forming cores (e.g., Girart et al. 1999; Rawlings et al. 2004; Arce & Sargent 2006). Future observations of other molecular ions and shock tracers are required to obtain a comprehensive understanding of ionization conditions in the LMC.

Formaldehyde (H₂CO) is detected by the ${5}_{\mathrm{1,5}}$–4_1,4 transition at 351.7686 GHz (E_u = 62 K). The abundance of H₂CO in ST11 is estimated to be 2.2 × 10⁻¹⁰, while Galactic hot cores show the abundances between 10⁻⁹ and 10⁻⁸ as in Table 5, except for G34.3+0.15 which only has a lower limit. The H₂CO is less abundant in ST11 than Galactic sources by 1–2 orders of magnitude.

Both gas-phase and grain surface reaction pathways are suggested for the formation of H₂CO in dense ISM. The grain surface pathway for the H₂CO formation requires the hydrogenation of CO (e.g., Hama & Watanabe 2013, and references therein). However, this pathways is suppressed in the LMC due to higher dust temperatures (see the detailed discussion presented in Section 5.2.4). Therefore we speculate that the suppressed formation of the H₂CO ice on grain surfaces contributes to the low abundance of H₂CO gas in ST11.

No methanol (CH₃OH) lines are detected in ST11. We estimate an upper limit on the fractional abundance to be <8 × 10⁻¹⁰ using the spectrum at the frequency of the CH₃OH(7₀–6₀ A⁺) transition at 338.4087 GHz (E_u = 65 K). Abundances of CH₃OH in Galactic hot cores are typically between ∼10⁻⁸ and ∼10⁻⁷ as in Table 5, but sometimes the abundance reaches ∼10⁻⁶ (e.g., G5.89-0.39 in Thompson & MacDonald 1999). In ST11, CH₃OH is depleted by at least 2–3 orders of magnitude as compared with Galactic hot cores.

The low abundance of CH₃OH in the LMC has been suggested in previous studies. Nishimura et al. (2016) reported, based on spectral line surveys toward molecular clouds in the LMC, that thermal emission lines of CH₃OH gas are significantly weak in the LMC compared to our Galaxy. Searches for maser emission in the LMC reported the underabundance of CH₃OH masers in the LMC (e.g., Ellingsen et al. 2010). Furthermore, Shimonishi et al. (2016) reported that the CH₃OH ice is less abundant in the LMC than in our Galaxy based on infrared observations. They reported that all of the observed ten high-mass YSOs in the LMC show CH₃OH ice abundances less than 5%–8% relative to water ice, while about one-third of Galactic high-mass YSOs show CH₃OH ice abundances between 10% and 40%. The authors suggested that warm ice chemistry is responsible for the low abundance of solid CH₃OH in the LMC, i.e., high dust temperatures in the LMC suppress the hydrogenation of CO on the grain surface, which leads to inefficient production of the CH₃OH ice. A decrease in the efficiency of CO hydrogenation at an elevated temperature is measured by laboratory experiments (e.g., Watanabe et al. 2003). In addition, the reduced formation of CH₃OH in relatively warm molecular clouds is confirmed by numerical simulations of grain surface chemistry dedicated to the LMC environment (Acharyya & Herbst 2015).

The warm gas temperature around ST11 suggests that ice mantles are mostly sublimated. The low elemental abundances of carbon and oxygen in the LMC should partly contribute to the observed low CH₃OH abundance. However, we need an additional explanation to account for the CH₃OH deficiency by several orders of magnitude, and this should be related to the grain surface chemistry by which CH₃OH is mainly formed. We suggest that suppressed production of the CH₃OH ice due to warm ice chemistry at the molecular cloud stage or the deeply embedded YSO stage is responsible for the deficiency of CH₃OH gas in this LMC hot core.

The above interpretation on the low CH₃OH abundance can also be applied to the low abundance of H₂CO in ST11 since a major grain surface pathway of both species requires the CO hydrogenation. However, the difference is that H₂CO has possible formation routes both in the gas-phase and in the solid-phase, while CH₃OH does not have an efficient formation route in the gas-phase under typical molecular cloud conditions. We speculate that the H₂CO observed in ST11 is mainly produced by gas-phase reactions. A slightly extended distribution of the H₂CO emission (Figure 3) supports this idea, because the species sublimated from ice mantles often show a compact distribution compared to those produced in the gas-phase. Owing to the lack of a possible gas-phase formation route, the degree of depletion around ST11 is actually larger in CH₃OH as compared to H₂CO.

Nitric oxide (NO) is one of the interesting molecules that show a peculiar abundance in ST11. The abundance of NO in ST11 is estimated to be 2.0 × 10⁻⁸. On the other hand, the average abundance and standard deviation of NO in six Galactic sources in Ziurys et al. (1991) including hot cores, high-mass protostellar objects, and Galactic center objects is 8.2 ± 2.9 × 10⁻⁹. In the LMC, nitrogen is even more depleted than other major elements such as carbon and oxygen; the elemental abundance of nitrogen in the LMC is [N/H]_LMC = 1.0 × 10⁻⁵ (Korn et al. 2002), while [N/H]_☉ = 8.3 × 10⁻⁵ for the Sun (Grevesse & Sauval 1998). Despite the lower nitrogen abundance by a factor of 8 in the LMC, the abundance of NO in ST11 is higher than in Galactic star-forming regions by a factor of 2–3.

Ammonia (NH₃) gas, which is often the most abundant nitrogen-bearing species in star-forming regions, is reported to be deficient in the LMC by 1.5–2 orders of magnitude as compared with Galactic star-forming regions (Ott et al. 2010). A previous infrared study suggests that the NH₃ ice around an embedded high-mass YSO in the LMC is possibly less abundant as compared with Galactic high-mass YSOs (Shimonishi et al. 2016). Although the NH₃ abundance in ST11 is unknown, the relatively high abundance of NO implies that NO could play an important role in nitrogen chemistry around ST11.

It is suggested that NO is formed by a neutral–neutral reaction in the gas-phase (e.g., Herbst & Klemperer 1973; Pineau des Forets et al. 1990):

$$\begin{eqnarray}&&{\rm{N}}+\mathrm{OH}\to \mathrm{NO}+{\rm{H}},\end{eqnarray} \tag{ 11 }$$

and destroyed by

$$\begin{eqnarray}&&\mathrm{NO}+{\rm{N}}\to {{\rm{N}}}_{2}+{\rm{O}}.\end{eqnarray} \tag{ 12 }$$

A numerical simulation of nitrogen chemistry in warm molecular gas suggests that the resultant abundance of NO through the above reactions increases as the temperature of molecular gas increases (Pineau des Forets et al. 1990). However, the enhancement of the NO abundance is suggested to be only a factor of 1.5 as the gas temperature increases from 70 to 250 K. Thus we speculate that pure gas-phase chemistry makes only a limited contribution to the enhanced abundance of NO in ST11.

The compact distribution of the NO emission observed in ST11 may hint at a possible origin in ice sublimation. Laboratory experiments argue that the bombardment of energetic ions on interstellar ice analogs containing H₂O, O₂, and N₂, or CO and N₂, produce NO in ice mixtures (Boduch et al. 2012; Sicilia et al. 2012). However, the low cosmic-ray density in the LMC (see Section 5.2.2) implies that the formation of NO by such energetic processing is less efficient around ST11. A reaction pathway though diffusive grain surface chemistry, N + O $\to$ NO, is also suggested in the literature (Tielens & Hagen 1982), but uncertainty remain regarding the efficiency of the reaction in actual grain surfaces.

The reason for the enhanced abundance of NO in ST11 despite the low elemental abundance of nitrogen in the LMC remains to be investigated. Detailed modeling of hot core chemistry in metal-poor environments is obviously necessary to interpret the enhanced abundance of NO in ST11.

Isocyanic acid (HNCO) is not detected in ST11. We estimate an upper limit on the fractional abundance of HNCO to be <1 × 10⁻¹⁰ based on the non-detection of the 16_0,16–15_0,15 transition at 351.6333 GHz (E_u = 143 K). There seems to be an emission-like feature at the position of the HNCO(16_3,14–15_3,13 and 16_3,13–15_3,12, 351.4168 GHz) line with V_LSR = 250.8 km s⁻¹. However, the high upper state energy of this transition (E_u = 518 K) and the non-detection of the 16_0,16–15_0,15 transition suggest that this line is likely a spurious signal. Galactic hot cores and high-mass protostellar objects typically show HNCO abundances between 10⁻⁹ and 10⁻⁸ (Bisschop et al. 2007), which is consistent with the average abundance of 2.3 × 10⁻⁹ for three Galactic hot cores in Table 5. Thus the HNCO abundance in ST11 is lower than Galactic counterparts by at least 1–2 orders of magnitude.

Both gas-phase and grain surface reaction pathways are suggested for the formation of HNCO in dense ISM (e.g., Tielens & Hagen 1982; Turner et al. 1999). The grain surface pathway requires the hydrogenation of CO to form HCO and subsequent reaction of HCO + N $\to$ HNCO. However, as discussed in Section 5.2.4, the hydrogenation of CO is suggested to be less efficient in the LMC due to high dust temperatures, which possibly suppress the grain surface formation of HNCO. This may contribute to the low abundance of HNCO in ST11 along with the low elemental abundance of nitrogen in the LMC.

Cyanoacetylene (HC₃N), an unsaturated carbon-chain molecule often detected in star-forming regions, is not detected in ST11. We estimate an upper limit on the fractional abundance of HC₃N to be <4 × 10⁻¹¹ based on the non-detection of the 38–37 transition at 345.6090 GHz (E_u = 323 K). Abundances of HC₃N in Galactic hot cores vary from ∼10⁻¹¹ to ∼10⁻⁹ as in Table 5. The estimated upper limit on the HC₃N abundance in ST11 is thus at the lower end of the Galactic HC₃N abundance.

Silicon monoxide (SiO) is detected by the J = 8–7 transition at 347.3306 GHz (E_u = 75 K). The abundance of SiO for ST11 is 3.3 × 10⁻¹¹, which seems to be low compared to Galactic sources, but the dispersion in the Galactic SiO abundances is significantly large as shown in Figure 10. The average SiO abundance for the three Galactic hot cores in Table 5 is 1.8 × 10⁻⁹, but their abundances range from ∼10⁻¹¹ to ∼10⁻⁹.

The elemental abundance of silicon in the LMC is about three times lower than the solar abundance; [Si/H]_LMC = 1.3 × 10⁻⁵ (Korn et al. 2002), while [Si/H]_☉ = 3.6 × 10⁻⁵ (Grevesse & Sauval 1998). Thus the elemental abundance in the LMC partly contributes to the observed low SiO abundance, but additional processes may be necessary to account for the 1–2 orders of magnitude lower SiO abundance.

SiO is often linked with the presence of energetic protostellar outflows since destruction of silicon-bearing dust (e.g., MgSiO₃) by shock is believed to be the origin of SiO (e.g., Blake et al. 1996; Wright et al. 1996; Tercero et al. 2011, and references therein). The destruction of dust grains and release of silicon can occur around ST11 since it shows a sign of outflows as described in Section 5.3. On the other hand, the low abundance of SiO would imply that shock chemistry is not very dominant in the present chemical compositions around ST11. Further quantitative discussions require detailed modeling of SiO production in low metallicity environments. Various factors including the degree of shock, elemental abundances, and compositions of dust grains may affect the SiO abundance.

The C³⁴S(7–6) transition at 337.3965 GHz (E_u = 65 K) is not detected toward ST11. We estimate an upper limit on the abundance of carbon monosulfide (CS) to be <3 × 10⁻¹⁰, while Galactic hot cores and high-mass protostellar objects typically show the CS abundances between 10⁻⁹ and 10⁻⁸ (Table 5; see also van der Tak et al. 2003). CS is significantly less abundant by at least 1–2 orders of magnitude in ST11 compared to Galactic sources, and the low elemental abundance of carbon and sulfur in the LMC cannot by itself explain the depletion of CS. Given the relatively high abundance of SO₂ in ST11 as discussed in Section 5.2.12, a significant amount of gas-phase sulfur is possibly incorporated in SO₂ due to different circumstellar chemistry around ST11.

We detect the emission line of thioformaldehyde (H₂CS) at 338.0832 GHz (10_1,10–9_1,9, E_u = 65 K), but the S/N of the line is relatively poor. The H₂CS abundance is estimated to be 6.2 × 10⁻¹¹ in ST11, while Galactic hot cores show the abundance of ∼2 × 10⁻⁹ on average with a relatively low dispersion (Table 5). H₂CS seems to be less abundant in the LMC hot core than in Galactic hot cores as well as CS, and only a small fraction of gas-phase sulfur is incorporated into H₂CS in ST11. Since our H₂CS abundance is estimated using a single line with poor S/N, further multiline observations of H₂CS transitions are necessary for accurate determination of the H₂CS abundance in the LMC.

The abundance of sulfur monoxide (SO) is estimated to be 2.4 × 10⁻⁸ in ST11, while Galactic hot cores and high-mass protostellar objects typically show SO abundances between 10⁻⁹ and 10⁻⁸ (Table 5; see also van der Tak et al. 2003). The Orion hot core shows an exceptionally high SO abundance of 2.0 × 10⁻⁷. The results suggest that ST11 shows a slightly higher abundance of SO than typical Galactic counterparts despite the low metallicity of the LMC. If we use the solar ³²S/³³S ratio of 127, then the SO abundance in ST11 is 7.6 × 10⁻⁸, which is even higher than typical Galactic abundances.

We emphasize, however, that our SO abundance entails considerable uncertainty. The abundance is estimated using a single ³³SO line at 337.1986 GHz (E_u = 81 K), which contains a number of unresolved hyperfine structures, while Galactic SO abundances are estimated from ³⁴SO lines. Multiline observations of SO and its isotopologues are necessary for further discussion.

Sulfur dioxide (SO₂) is in this study a key molecule for which we detect the largest number of transitions; we detect nine SO₂ lines, one SO₂ (ν₂ = 1) line, ten ³⁴SO₂ lines, and five ³³SO₂ lines. The fractional abundance of SO₂ is estimated to be 2.1 × 10⁻⁸ in ST11 based on rotation diagram analysis of ³⁴SO₂ lines. The column density ratio of SO₂ and ³⁴SO₂ is about 14, suggesting that the optically thin assumption is mostly valid for the observed SO₂ lines because the ³²S/³⁴S ratio is reported to be 15 in the LMC. If we assume the solar isotope ratio of ³²S/³⁴S = 22, the SO₂ lines could be moderately optically thick. In either case, we can reasonably assume that ³⁴SO₂ and ³³SO₂ lines are optically thin because these isotopologues are much less abundant than ³²SO₂. The isotope abundance of ³³S in ST11 based on the present results is ³²S/³³S = 40, while the solar value is ³²S/³³S = 127.

Galactic hot cores typically show the SO₂ abundance from ∼10⁻⁸ to ∼10⁻⁷, while even younger embedded high-mass protostellar objects show the abundance of ∼10⁻⁹ (van der Tak et al. 2003). ST11 shows a factor of ∼3 lower SO₂ abundance as compared with the average abundance of 7.0 × 10⁻⁸ for the three Galactic hot cores in Table 5. The elemental abundance of sulfur in the LMC is reported to be [S/H]_LMC = 5.0 × 10⁻⁶, while the solar abundance is [S/H]_☉ = 1.9 × 10⁻⁵ (Russell & Dopita 1992). Sulfur is less abundant by a factor of ∼4 in the LMC compared with the solar abundance. Thus, the low SO₂ abundance in ST11 is well explained by the elemental abundance of sulfur in the LMC. This would suggest that hot core chemistry of SO₂ is dependent on elemental abundances of the host galaxy. A multitude of SO₂ and its isotopologue line detections in ST11 imply that SO₂ can be a key molecular tracer to test hot core chemistry in metal-poor environments.

It should be noted that sulfur chemistry in hot core regions is highly time-dependent (e.g., Charnley 1997). Hence, both the age and the interstellar environment should be taken into account to interpret the chemical compositions of sulfur-bearing species around ST11. Numerical simulations of hot core chemistry dedicated to low metallicity environments are thus highly required.

Spectral lines from complex organic molecules are not detected in this work. We estimate upper limits on fractional abundances of three molecules whose relatively strong transitions are covered in the present data: [CH₃OCH₃/H₂] < 3 × 10⁻⁹, [HCOOCH₃/H₂] < 2 × 10⁻⁸, and [C₂H₅OH/H₂] < 5 × 10⁻⁹. The average abundances of these molecules for Galactic hot cores in Table 5 are [CH₃OCH₃/H₂] = 1.5 × 10⁻⁸, [HCOOCH₃/H₂] = 1.8 × 10⁻⁸, and [C₂H₅OH/H₂] = 3.7 × 10⁻⁹. These estimates suggest that CH₃OCH₃ is less abundant by at least an order of magnitude in ST11 compared with Galactic hot cores, while the upper limit abundances of HCOOCH₃ and C₂H₅OH are comparable to the average abundances of Galactic sources. Although the present data do not provide conclusive upper limits on the abundances of the majority of complex organic molecules in the LMC, the possibly lower abundance of CH₃OCH₃ in ST11 suggests that the formation of large molecules may be less efficient in the LMC.

It is suggested by theoretical studies that CH₃OH in ice mantles plays an important role in the formation of complex organic molecules (e.g., Nomura & Millar 2004; Garrod 2008; Herbst & van Dishoeck 2009). Shimonishi et al. (2016) argue that the formation of complex organic molecules from methanol-derived species could be less efficient in the LMC due to the low abundance of the CH₃OH ice around high-mass YSOs. We confirm the significant deficiency of CH₃OH gas around ST11 in this work. We thus speculate that the low abundance of CH₃OH contributes to the low production efficiency of CH₃OCH₃ and possibly other complex organic molecules in the LMC.

Note that unknown rotational temperatures and limited frequency coverages produce considerable uncertainties on the abundance estimate. Furthermore, the formation of large molecules from other parent species such as H₂CO, C₂H, c-C₃H₂, and NO, which are detected in the LMC, should be taken into account for a comprehensive understanding of the complex chemistry around protostars. Hence, further observations with broader frequency coverage and higher sensitivity are critically needed in conjunction with theoretical and experimental efforts to understand complex organic chemistry in metal-poor environments.