EDP Sciences
Free Access
Issue
A&A
Volume 615, July 2018
Article Number L16
Number of page(s) 5
Section Letters to the Editor
DOI https://doi.org/10.1051/0004-6361/201833607
Published online 26 July 2018

© ESO 2018

1. Introduction

Nitrogen is the fifth most abundant element in the universe. Our understanding of nitrogen chemistry in star-forming regions is limited compared to other volatile elements, such as carbon and oxygen. The dominant form of gaseous nitrogen in star-forming regions is unclear (e.g., Bergin & Tafalla 2007), while it is theoretically expected to be either atomic N or N2 (e.g., Aikawa et al. 2005; Le Gal et al. 2014). One of the main limiting factors is that neither atomic N nor N2 is directly observable in the cold and dense gas of star-forming regions. However, the N2 abundance can be constrained indirectly by observing a proxy molecule N2H+ (e.g., Maret et al. 2006).

The observational and theoretical studies of nitrogen isotope fractionation in star-forming regions can help to constrain nitrogen chemistry. Nitrogen has two stable isotopes, 14N and 15N. The elemental abundance ratio [14N/15N]elem in the local interstellar medium (ISM) has been estimated to be ~200–300 from the absorption line observations of N-bearing molecules toward diffuse clouds (Lucas & Liszt 1998; Ritchey et al. 2015). L1544 is a prototypical prestellar core located in the Taurus molecular cloud complex. In L1544, the 14N/15N ratio of several different molecules has been measured: (Bizzocchi et al. 2013; Redaelli et al. 2018), NH2D/15NH2D > 700 (Gérin et al. 2009), CN/C15N = 500 ± 75 (Hily-Blant et al. 2013b), and HCN/HC15N = 257 (Hily-Blant et al. 2013a). Among the measurements, the significant depletion of 15N in N2H+ is the most challenging for the theory of 15N fractionation. In general, molecules formed at low temperatures are enriched in 15N through gas-phase chemistry triggered by isotope exchange reactions (e.g., Terzieva & Herbst 2000). A 15N-bearing molecule has a slightly lower zero-point energy than the corresponding 14N isotopolog. This results in endothermicity for the exchange of 15N for 14N, which inhibits this exchange at low temperature enabling the concentration of 15N in molecules. Astrochemical models for prestellar cores, which consider a set of nitrogen isotope exchange reactions, have indeed predicted that atomic N is depleted in 15N, while N2 (and thus N2H+) is enriched in 15N (e.g., Charnley & Rodgers 2002). The model prediction clearly contradicts the observation of the N2H+ isotopologs in L1544. The 15N depletion in N2H+ was recently found in other prestellar cores as well, such as L183, L429, and L694-2 (Redaelli et al. 2018). Furthermore, Roueff et al. (2015) recently pointed out the presence of activation barriers for some key nitrogen isotope exchange reactions, based on their quantum chemical calculations. Then 15N fractionation triggered by isotope exchange reactions may be much less efficient than previously thought (Roueff et al. 2015, but see also Wirström & Charnley 2018).

Another mechanism that can cause 15N fractionation is photodissociation of N2 (Heays et al. 2014). N2 photodissociation is prone to self-shielding. Because 14N15N is much less abundant than 14N2, 14N15N needs a higher column density of the ISM gas for self-shielding. This makes N2 photodissociation an isotope-selective process. As a result, N2 is depleted in 15N, which is consistent with the observation of the N2H+ isotopologs in L1544. Isotope-selective photodissociation of N2 is efficient only for limited regions where the interstellar UV radiation field is not significantly attenuated, however (Heays et al. 2014; Furuya & Aikawa 2018). The prestellar core L1544 has high density and AV (>10 mag for a millimeter-dust continuum peak). Detection of carbon chain species, such as C3H2, however, may indicate that the interstellar UV radiation penetrates to moderate depth in L1544 (Spezzano et al. 2016). Then it is unclear whether the isotope-selective photodissociation of N2 is at work in L1544, and how it affects the measurement of the 14N/15N ratio of N2H+.

In order to test the selective photodissociation scenario, we observed 15N isotopologs of N2D+ toward prestellar core L1544. Compared with N2H+, N2D+ selectively traces colder and denser regions (i.e., core center) (Caselli et al. 2002a), where the attenuation of the interstellar UV radiation field is more significant. If the isotope-selective photodissociation of N2 by the penetrating UV radiation is the cause of the 15N depletion in N2H+, the 14N/15N ratio of N2D+ should be lower than that of N2H+ and be close to [14N/15N]elem. Moreover, N2D+ is less optically thick than N2H+, which allows us to accurately evaluate the column density of the 14N isotopolog and thus the 14N/15N ratio, although more sensitive observations are required for the detection of the 15N isotopologs of N2D+ than those of N2H+.

2. Observations

We observed the N15ND+(1–0), 15NND+(1–0), and N2D+(1–0) transitions toward the prestellar core L1544 with the IRAM 30 m telescope at Pico Veleta on 2017 December 22–24. We tracked the L1544 continuum dust emission peak at 1.3 mm, where 15N isotopologs of N2H+ were previously detected (Bizzocchi et al. 2010, 2013). The observed position is (αJ2000, σJ2000) = (05h04m17ṣ21, 25°1042ʺ̣8) (Caselli et al. 2002a). The telescope pointing was checked every two hours by observing the continuum source 0439+360 near the target position and was ensured to be better than ±3. The half-beam power width was 32−33.

We employed Eight Mixer Receiver (EMIR) E090 with dual polarization mode. The system noise temperatures were typically from 70 K to 130 K during the observation run. The N15ND+(1–0) and 15NND+(1–0) transitions were observed simultaneously (Set 1), while the N2D+(1–0) transition was observed with a different frequency setting (Set 2). The hyperfine components and their relative intensities of the N15ND+(1–0) and 15NND+(1–0) transitions were experimentally studied by Dore et al. (2009), and they are listed in Table A.1. A frequencyswitching mode was employed with a frequency offset of 7.35 MHz. We used eight Fourier transform spectrometer (FTS) autocorrelators with a bandwidth of 1820 MHz. The frequency resolution of 50 kHz corresponds to 0.2 km s1 at 75 GHz. We integrated the spectrum for a total on-source time of 5.2 h for Set 1 and 0.4 h for Set 2. Two orthogonal polarizations were simultaneously observed, and are averaged together to produce the final spectrum. The main-beam temperature (T MB) is derived by , where is the antenna temperature, Feff is the forward efficiency (95%), and Beff is the main-beam efficiency (74%). The final rms noise is 2.3 mK in TMB for N15ND+(1–0), 2.1 mK for 15NND+(1–0), and 11 mK for N2D+(1–0). The N15NH+(1–0) and 15NNH+(1–0) transitions were also observed in Set 1. Both transitions were detected, and the obtained spectra are similar to those obtained in the framework of ASAI IRAM 30 m large program (De Simone et al. 2018; Lefloch et al. 2018), which observed the same object and the same position with the same velocity resolution, but employing wobbler-switching mode. We do not discuss the observations of the 15N isotopologs of N2H+ in this work because they have been studied in detail in previous work (Bizzocchi et al. 2010, 2013; De Simone et al. 2018).

Table 1.

Derived column density and 14N/15N ratio.

3. Results

The data were processed using the GILDAS software (Pety et al. 2005). The N2D+(1–0) transition was clearly detected, while N15ND+(1–0) and 15NND+(1–0) were not detected, as shown in Fig. 1. Line parameters for N2D+(1–0) were estimated using the HFS routine implemented in CLASS. The derived total optical depth of the lines and the FWHM line width are 3.08 ± 0.19 and 0.406 ± 0.003 km s−1, respectively. The main component of the N2D+(1–0) transition (77.1096162 GHz), which has a fraction of 7/27 of the total line strength, is marginally optically thick (~0.8). We derived the total column density of N2D+ using Eq. (A1) of Caselli et al. (2002b), which is valid for optically thick emission. For the column density calculation, the excitation temperature (Tex) was set to be 5 K, which was previously derived from N2H+(1–0) and N2D+(2–1) observations toward the same object and the same position (Caselli et al. 2002a; Crapsi et al. 2005). The parameters of the observed transitions were taken from the Cologne Database for Molecular Spectroscopy (Müller et al. 2001, 2005). The total column density of N2D+ (Ntot(N2D+)) is evaluated to be (5.4 ± 0.3) × 1012 cm−2. The error on Ntot(N2D+) is given by propagating the errors on the total optical depth and the FWHM line width. Our Ntot(N2D+) is very close to that obtained by Crapsi et al. (2005) ((4.3 ± 0.6) × 1012 cm−2), who derived it from the N2D+(2–1) data.

thumbnail Fig. 1.

Spectra of the N2D+(1–0) transition (top panel), the N15ND+(1–0) transition (middle panel), and the 15NND+(1–0) transition (bottom panel) observed toward L1544. The intensity scale is the main-beam temperature. In the top panel, the red curve depicts the result of the HFS fit. The N15ND+(1–0) transition and the 15NND+(1–0) transition were not detected down to 3σ levels in 0.2 km s−1 channels of 6.9 mK and 6.3 mK, respectively. The vertical blue lines indicate the positions of the expected hyperfine components.

Open with DEXTER

Upper limits to Ntot(N15ND+) and Ntot(15NND+) were obtained from the 3σ upper limits to the integrated intensity of the transitions, where σ is the rms noise of the spectra, Δν is the FWHM line width of the spectra, assumed to be the same as that of N2D+(1–0), and δν is the velocity resolution. Assuming local thermal equilibrium, the 3σ intensity upper limits were converted into the column density upper limits using Eq. (A4) of Caselli et al. (2002b), which is valid for optically thin emission. Tex is assumed to be 5 K. We obtain Ntot(N15ND+) < 7.0 × 109 cm−2 and Ntot(15NND+) < 6.5 × 109 cm−2.

4. Discussion and conclusion

From the column densities of the N2D+ isotopologs, we obtain the lower limits of the 14N/15N ratio of 730 for N15ND+ and 780 for 15NND+. These lower limits are significantly higher than [14N/15N]elem in the local ISM (~200–300; Lucas & Liszt 1998; Ritchey et al. 2015). It is reasonable to consider the 14N/15N ratios of N2D+ as the ratio of N2, because N2D+ primary forms by N2 + X2D+, where X is H or D. Then our observations indicate that N2 is significantly depleted in 15N in the central part of L1544. If we assume Tex of 4.5 K (5.5 K) in the column density evaluation, the lower limits of the 14N/15N ratios become 600 (950) for N15ND+ and 640 (1010) for 15NND+. Our qualitative conclusion is thus robust against the assumed value of Tex. Colzi et al. (2018) recently derived the [14N/15N]elem ratio in the local ISM of ~400 from the observations of HCN isotopologs toward a sample of 66 cores in massive star-forming regions. Even when this higher elemental abundance ratio is adopted, our qualitative conclusion does not change.

As described in Sect. 1, nitrogen isotope exchange reactions make N2 enriched in 15N; they are thus not relevant to the observed fractionation. Isotope-selective photodissociation of N2 by penetrating interstellar FUV photons is also ruled out as the cause of the 15N depletion because the penetration would be negligible in the central part of L1544, where N2D+ emission arises. Our lower limits of the 14N/15N ratios for N2D+ are consistent with those of N2H+ (~1000) obtained by Bizzocchi et al. (2013), which also supports that isotope-selective photodissociation of N2 is not responsible for the 15N depletion, as discussed in Sect. 1. We note that cosmic-ray induced photodissociation of N2 does not cause 15N fractionation because the destruction of N2 by He+ is much faster (Heays et al. 2014; Furuya & Aikawa 2018). Therefore, in situ chemistry is probably not responsible for the 15N depletion in N2 in the central part of L1544.

The most likely explanation is that the 15N depletion is inherited from more diffuse gas, as recently proposed by Furuya & Aikawa (2018), based on their astrochemical models in forming and evolving molecular clouds. They found that during the evolution of molecular clouds, the nitrogen isotopes can be differentially partitioned between gas and ice, making 15N-depleted gas and 15N-enriched ice. In the molecular cloud, where the external UV radiation field is not fully shielded, 14N15N is selectively photodissociated with respect to 14N2, which results in the enrichment of 15N in the photodissociation product, atomic N. Atomic N is adsorbed onto grain surfaces and converted into NH3 ice by surface reactions, while adsorbed N2 does not react with other species, including atomic H. As long as the nonthermal desorption (especially photodesorption in their models) of NH3 ice is less efficient than that of N2 ice, the net effect is the loss of 15N from the gas phase, producing 15N-depleted gas and 15N-enriched ice. When the external UV radiation field is sufficiently shielded, 15N depletion does no longer proceed, but is largely conserved unless a significant amount of NH3 ice is sublimated.

As noted by Furuya & Aikawa (2018), the mechanism is the most efficient around the chemical transition from atomic N to N2, where the self-shielding of 14N2 becomes important. Before the transition, both 14N2 and 14N15N are efficiently photodissociated, while after the transition, the abundance of atomic N is too low to affect the bulk gas isotopic composition. Therefore, if the mechanism proposed by Furuya & Aikawa (2018) was at work in the parent cloud of L1544 or the outer regions of L1544, it means that the transition from atomic to molecular nitrogen should have occurred there as well.

Nitrogen chemistry mainly consists of three competing processes: (i) the conversion of atomic N to N2 in the gas phase, (ii) destruction of N2, for instance, via photodissociation and reaction with He+, and (iii) freeze-out of atomic N and N2 onto dust grains followed by surface reactions (e.g., Daranlot et al. 2012; Li et al. 2013). The conversion of atomic N into N2 has been proposed to occur by slow neutral-neutral reactions, such as NO + N and CN + N (Herbst & Klemperer 1973; Daranlot et al. 2012). According to the pseudo-time-dependent gas-phase astrochemical model under dense cloud conditions (104 cm−3, 10 K, 10 mag) by Le Gal et al. (2014), the conversion of atomic N into N2 takes an order of Myr, depending on assumed elemental abundances. In the gas-ice model of Daranlot et al. (2012), under the similar physical conditions, the conversion of atomic N to N2 takes ~5 × 105 yr, and it occurs after the significant fraction of nitrogen is frozen out. On the other hand, N2 mainly forms via the reactions NH2 + N and NH + N around the transition from atomic to molecular nitrogen in the models of Furuya & Aikawa (2018) and Furuya & Persson (2018), in which the dynamical evolution of molecular clouds is considered. NH2 and NH are mainly formed via photodesorption of NH3 ice, followed by photodissociation in the gas phase. In this case, the formation rate of N2 from atomic N is, roughly speaking, similar to the freeze-out rate of atomic N. Considering that interstellar ices, at least water ice, are already abundant in molecular clouds with relatively low line-of-sight visual extinction (e.g., ~3 mag for the Taurus dark clouds, Whittet 1993), it may not be surprising that the transition from atomic to molecular nitrogen occurs in the parent cloud of L1544 or in the outer regions of L1544. It should be noted that the N2-dominant region could be larger than the regions traced by N2H+ and NH3 emission, since their abundances are controlled not only by N2, but also by CO; the catastrophic CO freeze-out, which occurs in the late stage of the interstellar ice formation at high densities (≳105 cm−3; e.g., Pontoppidan 2006), causes their abundances to be enhanced (e.g., Aikawa et al. 2005).

It may be interesting to estimate the partitioning of elemental nitrogen between gas and ice. The abundance of gaseous N2 in dense prestellar cores was previously estimated from the comparison of N-chemistry models with the observations of N2H+ and other relevant species. Maret et al. (2006) inferred that gaseous N2 contains only a few percent of the overall elemental nitrogen (N/H = 6 × 10−5 in the local ISM; Przybilla et al. 2008) in the dense cloud B68. Pagani et al. (2012) also inferred that gaseous N2 contains ≲1% of elemental nitrogen in the prestellar core L183. While Maret et al. (2006) suggested that atomic N is the primary form of elemental nitrogen in B68 to account for this low gaseous N2 abundance, their model predicts that NH3 ice is the primary nitrogen reserver (see also Daranlot et al. 2012; Furuya & Persson 2018). The gas-ice astrochemical model by Ruaud et al. (2016), on the other hand, predicts that the HCN ice is more abundant than NH3 ice. Figure 2 shows the fraction of elemental nitrogen in the form of ice as functions of the 14N/15N ratio of the bulk ice. In the figure, the 14N/15N ratio of the bulk gas (i.e., that of N2) is assumed to be 1000. If NH3 and HCN ices are the primary forms of elemental nitrogen in L1544, as predicted by the astrochemical models, the 14N/15N ratio of the icy species should be close to but slightly lower than [14N/15N]elem.

thumbnail Fig. 2.

Estimated fraction of elemental nitrogen in ices as functions of the 14N/15N ratio of the ices. The estimated fractions with different [14N/15N]elem are shown by different colors. In all cases, the 14N/15N ratio of the bulk gas is assumed to be 1000.

Open with DEXTER

Gérin et al. (2009) found that the 14N/15N ratio of gaseous NH2D is >700 in L1544. The 14N/15N ratios of gaseous NH3 in the cold gas of dense molecular clouds were derived to be 334 ± 50 in Barnard 1 and 340 ± 150 in NGC 1333 (Lis et al. 2010). These measurements indicate that gaseous NH3 in the cold gas is not significantly enriched in 15N. It should be noted, however, that the origin of gaseous NH3 in the cold gas, that is, whether it is formed by gas-phase reactions or released from ices via nonthermal desorption, remains unclear. The 14N/15N ratio of icy species in star-forming regions could be constrained by measuring the molecular 14N/15N ratios in the warm (≳100 K) gas surrounding protostars. This type of observations is crucial for better understanding the nitrogen partitioning.

Finally, observations of comets have found that cometary NH3 and HCN are enriched in 15N by a factor of around three (Mumma & Charnley 2011; Shinnaka et al. 2016) compared to the Sun (~150 versus 441; Marty et al. 2011). The ammonia abundance with respect to water in cometary ices (0.4–1.4%) is lower than that in interstellar ices (typically ~5%) (Mumma & Charnley 2011; Öberg et al. 2011). These (possible) differences between the N-bearing species in cometary and interstellar ices might indicate a primordial variation in the ice formation environments or ice processing in the solar nebula (e.g., Lyons et al. 2009; Furuya & Aikawa 2014).

Acknowledgments

We are grateful to the IRAM staff for excellent support. We thank Yuto Sato for his help in preparing for the IRAM proposal, and we also thank the anonymous referee for useful comments that helped to improve this paper. This work is partly supported by JSPS KAKENHI Grant Numbers 16K17657 and 17K14245.

References

Appendix A: Additional table

Table A.1.

Hyperfine frequencies for J = 1–0 transitions of N15ND+ and 15NND+ taken from Dore et al. (2009).

All Tables

Table 1.

Derived column density and 14N/15N ratio.

Table A.1.

Hyperfine frequencies for J = 1–0 transitions of N15ND+ and 15NND+ taken from Dore et al. (2009).

All Figures

thumbnail Fig. 1.

Spectra of the N2D+(1–0) transition (top panel), the N15ND+(1–0) transition (middle panel), and the 15NND+(1–0) transition (bottom panel) observed toward L1544. The intensity scale is the main-beam temperature. In the top panel, the red curve depicts the result of the HFS fit. The N15ND+(1–0) transition and the 15NND+(1–0) transition were not detected down to 3σ levels in 0.2 km s−1 channels of 6.9 mK and 6.3 mK, respectively. The vertical blue lines indicate the positions of the expected hyperfine components.

Open with DEXTER
In the text
thumbnail Fig. 2.

Estimated fraction of elemental nitrogen in ices as functions of the 14N/15N ratio of the ices. The estimated fractions with different [14N/15N]elem are shown by different colors. In all cases, the 14N/15N ratio of the bulk gas is assumed to be 1000.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.