CYANIDE PHOTOCHEMISTRY AND NITROGEN FRACTIONATION IN THE MWC 480 DISK

The HCN J = 4–3 and CN N = 3–2, J = 7/2–5/2 lines were observed toward MWC 480 and DM Tau with ALMA during Cycle 0 (PI: E. Chapillon, proposal 2011.0.00629). The Band 7 observations were carried out in 2012 October with 23 antennas. The total on-source observing time was 34 minute and the baselines lengths ranged between 21 and 384 m. The frequency bandpass was calibrated by observing the bright quasar J0423–013 at the begining of the observing run. The phase and amplitude temporal variations were calibrated by regularly observing the nearby quasars J0423–013, J01510–180 and Callisto. J01510–180 was used to derive the absolute flux scale. Two spectral windows of 468 MHz bandwidth and 122 kHz spectral resolution were centered on the CN and HCN lines.

The data were calibrated by the ALMA staff using standard procedures, and retrieved from the public archive. We further self-calibrated the data in phase and amplitude in CASA, taking advantage of the bright continuum emission in both sources. The solutions of the self-calibration derived for the continuum were applied to each spectral window. The continuum was then subtracted from the visibilities using channels free of line emission to produce the spectral line cubes. The images were obtained by deconvolution of the visibilities using the CLEAN algorithm in CASA, with Briggs weighting and a robust parameter of 0.5, resulting in an angular resolution of 06. A clean mask was created for each channel to select only regions with line emission during the cleaning process. Table 1 summarizes the observational parameters of the lines.

Table 1.  Observation Parameters

Line	Frequency	Eu/k	Beam	PA	Channel Width	Noise
	(GHz)	(K)	(arcsec)	(°)	(km s−1)	(mJy beam−1 channel−1)
CN N = 3–2, J = 7/2–5/2	340.24777	32.6	0.78 × 0.45/0.65 × 0.48	−23.1/−29.2	0.1	12.3/9.2
HCN J = 4–3	354.50548	42.5	0.77 × 0.41/0.64 × 0.40	−24.1/−29.0	0.1	14.0/11.2
H13CN J = 3–2	259.0118	24.9	0.74 × 0.46	−88	0.2	3.2
HC15N J = 3–2	258.1571	24.8	0.74 × 0.46	−90	0.2	4.2

Note. Two values are given for the beam, PA, and noise, corresponding to the parameters in MWC 480 and DM Tau, respectively.

Download table as:  ASCIITypeset image

The H¹³CN J = 3–2 and HC¹⁵N J = 3–2 lines were observed toward MWC 480 with ALMA during Cycle 2 as part of project 2013.1.00226 (PI: K.I. Öberg). These observations, partially described in Öberg et al. (2015), are part of a larger survey that did not include DM Tau. In brief, the Band 6 observations were carried out in 2014 June with 33 antennas with baseline lengths spanning between 18 and 650 m. The total on-source observing time was 24 minutes. The quasar J0510+1800 was observed to calibrate the amplitude and phase temporal variations, as well as the frequency bandpass. The same quasar was used to derive the absolute flux scale, as no planet was available during the observations. Two spectral windows of 59 MHz bandwidth and 61 kHz channel spacing covered the H¹³CN  and HC¹⁵N lines.

A description of the phase and amplitude calibration, as well as the self-calibration, can be found in Öberg et al. (2015). The self-calibrated visibilities were cleaned in CASA using a robust parameter of 1.0, yielding an angular resolution of ∼06 (see Table 1). This selection of robustness parameter allowed us to improve the signal-to-noise ratio of the weak H¹³CN and HC¹⁵N lines, while maintaining a good angular resolution. A clean mask was created for each channel, similarly to CN and HCN, to select only regions with line emission during the cleaning process.

Figure 1 shows the channel maps of the CN N = 3–2, J = 7/2–5/2, and HCN J = 4–3 lines in the MWC 480 and DM Tau disks. The velocity pattern characteristic of Keplerian rotation is seen for both lines and disks. Toward MWC 480, the HCN emission is clearly more compact than the CN emission in all channels. Such a difference in emission patterns is not observed toward the DM Tau disk. Furthermore, DM Tau exhibits a much more complex emission pattern in both the CN and HCN lines. The CN "butterfly" shape in some of the central channels is caused by spectrally resolved hyperfine components. HCN seems to present an outer ring, whose potential origin is discussed in Section 4.

Zoom In Zoom Out Reset image size

Figure 2 shows the continuum, and velocity integrated maps and azimuthally averaged radial profiles of the CN and HCN lines toward MWC 480 and DM Tau. Toward MWC 480, these data representations confirm the noted differences in CN and HCN emission regions in Figure 1. CN emission is detectable at a 2× larger radius compared to HCN. The best-fit half-light radius for the HCN and CN emission toward MWC 480 are 66 and 102 AU, respectively. We further note that the HCN emission has a similar size to that of the dust continuum emission, which is shown by the dashed line in Figure 2. In contrast, toward DM Tau the best-fit half-radius for HCN and CN are similar, 151 and 158 AU, respectively. The HCN outer ring hinted at in the DM Tau channel maps is also confirmed as a "bump" around 300 AU in the radial profile. No such ring is seen toward MWC 480.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

To quantify the HCN and CN radial abundance profiles we modeled the CN and HCN emission using the procedure and disk density and temperature structure described in Öberg et al. (2015). Key parameters adopted in the model are listed in Table 2. In brief, we model molecular abundances by defining them as power-laws X = X₀(r/R₀)^α, with X₀ the abundance with respect to total hydrogen at R₀ = 100 AU. We also included an outer cut-off radius, R_out. We ran grids of models for different outer radius 50 < R_out/AU < 500, power-law indexes −2 < α < 2 and molecular abundances 10⁻¹⁴ < X₁₀₀/cm⁻² < 10⁻⁹. The line emission was calculated using the radiative transfer code LIME (Brinch & Hogerheijde 2010) assuming non-LTE excitation with collision rates from (Dumouchel et al. 2010) for HCN and from Lique et al. (2010) for CN. For the HCN isotopologues, we use the collisional rates of HCN. The model fitting was done in the u−v plane by computing χ², the weighted difference between model and observations, for the real and imaginary parts of the complex visibilities. The best-fit models were obtained by minimizing χ².

Figure 3 shows the resulting contours of the ${\rm{\Delta }}{\chi }^{2}={\chi }^{2}-{\chi }_{\mathrm{min}}^{2}$ distribution for the HCN and CN abundance profiles. The best-fit model of both CN and HCN lines corresponds to a decreasing abundance profile as a function of radius, with α = −1. The best-fit outer-radius of the HCN emission is 110 AU, while the outer-radius of the CN emission is constrained to >300 AU, i.e., >2× larger than HCN. The best-fit abundances at 100 AU of CN and HCN are given in Table 3 but they are highly model dependent, i.e., they depend strongly on the assumed vertical abundance profile, and should not be directly compared with model predictions. Table 3 lists the best-fit parameters together with what are formally 7σ error bars based on the data SNR. In our case the data SNR is high enough that we are dominated by stochastic noise from the radiative transfer code and the 7σ error bar represents the deviation from the best-fit model where there is a clear, visual misfit between model and data.

Zoom In Zoom Out Reset image size

Figure 4 shows the excellent fit between our best-fit model and the data. Except for the low-velocity CN channels there are no significant residuals, and those residuals are due to the second hyperfine component of CN, which is not included in our model. This hyperfine component is weak in MWC 480, but it is clearly seen in DM Tau (see Figure 1). The good fit is also apparent when comparing the disk-integrated spectra (Figure 5), where the spectra are extracted using a Keplerian mask.

Zoom In Zoom Out Reset image size

Figure 6 displays the moment-zero maps of the H¹³CN and HC¹⁵N lines in MWC 480 together with the integrated emission of the blue and red shifted part of the the spectra, illustrating the disk rotation. The emission of the two isotopologues is compact, visually similar to the dust continuum and the HCN line emission.

Zoom In Zoom Out Reset image size

We modeled the emission of the HCN isotopologues following the same procedure as for CN and HCN above. Figure 7 shows the Δχ² distribution for the H¹³CN and HC¹⁵N abundance profiles. The best-fit parameters are listed in Table 3. The good fit of the best-fit model to the data disk integrated spectra (extracted using a Keplerian mask) is shown in Figure 8. The best-fit model abundances at 100 AU are given in Table 3. Similarly to CN and HCN, the absolute abundances are uncertain, due to an unknown vertical abundance profile. Assuming that the HCN isotopologue emission originates in the same disk layer and that the lines are optically thin, their abundance ratio should be robust, however. This abundance ratio is 2.8 ± 1.4 and holds for the entire disk out to 100 AU since both lines are best fit by the same power-law index of −0.5.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In the MWC 480 disk, the integrated H¹³CN/HCN density flux ratio is ∼36, which is lower than the ¹²C/¹³C isotopic ratios observed in the ISM, indicating that the HCN line is partially optically thick. We therefore use H¹³CN as a proxy of HCN to calculate the nitrogen fractionation in HCN. Assuming an isotopic ratio of ¹²C/¹³C = 70, we derive a disk HCN isotopologue ratio HC¹⁴N/HC¹⁵N = 200 ± 110. To estimate this error we included a 30% uncertainty in the C isotopic ratio, as this value has been shown to vary both with time and density (Roueff et al. 2015). The main uncertainty (∼85%), however, arises from the error in the measured H¹³CN/HC¹⁵N ratio.

The observed CN and HCN disk emission profiles in the MWC 480 disk are well-fit by power-law abundance profiles with a radial cut-off that is two times larger for CN compared to HCN. CN is thus significantly more radially extended than HCN, i.e., we can exclude with a high level of confidence that excitation alone is responsible for the different HCN and CN emission profiles. This result is consistent with chemical model predictions of survival of CN and rapid photodissociation of HCN in the outer, low-density disk (Aikawa et al. 2002; Jonkheid et al. 2007). Indeed, these models predict that the CN abundance can remain constant and even increase with radius because radical formation is favored at low densities, and because CN is a major HCN photodissociation product. HCN, on the other hand, should only be present in the high-density parts of the disk where a shielded molecular layer exists.

The same models also predict a different vertical structure of CN and HCN, with the CN abundance peaking close to the disk surface in the inner disk regions. Constraining the vertical abundances of these molecules requires multiple line transitions and/or direct observations of vertical emission patterns in edge-on disks. In the meantime, the derived abundance ratio of the two molecules remains relatively unconstrained.

The emission of CN and HCN in the disk around T tauri star DM Tau exhibit different patterns compared to MWC 480. First, the spatial distribution (the extension of the emission) is similar for CN and HCN in the DM Tau disk. This difference between MWC 480 and DM Tau may be explained by the fact that the UV field around DM Tau should be substantially lower. Less vertical disk column may therefore be required to shield the outer disk regions, resulting in a shielded molecular layer out to many 100 s of AU in the DM Tau disk. Second, HCN exhibits a low SNR outer ring around 300 AU, best identified in the azimuthally averaged emission profile. The gap demarcating the main HCN disk from the outer ring is faintly seen in CN as well. This gap may be the result of micron-sized dust depletion at the same radius due to dust dynamics, which would enable UV radiation to penetrate down toward the midplane, photo-dissociating HCN. Such a dust lane has been observed toward TW Hya in scattered light (Debes et al. 2013). Similar high-quality scattered light observations toward DM Tau could resolve this question. If this scenario is confirmed, these ALMA observations suggest that CN/HCN observations could be developed into a powerful probe of dust distributions and dynamics in the outer regions of disks.

We used observations of HCN isotopologues to determine the level of nitrogen fractionation in the MWC 480 disk. Figure 9 shows the measured ¹⁴N/¹⁵N ratio in the MWC 480 disk compared with different Solar system bodies and the ISM. Observations of the ¹⁴N/¹⁵N ratio in our Solar System reveal a clear difference between gaseous and rocky bodies (Mumma & Charnley 2011). A low N fractionation (i.e., high ¹⁴N/¹⁵N) is observed toward the Sun and Jupiter (Marty et al. 2011), while a high fractionation (i.e., low ¹⁴N/¹⁵N) is observed in the rocky planets, comets and meteorites (Bockelée-Morvan et al. 2008). Isotopic ratios in the coma of comets are determined remotely through high-resolution radio and optical spectroscopy. The ROSETTA space mission will hopefully provide in situ measurements of the ¹⁴N/¹⁵N ratio in the coma of comet 67P/Churyumov–Gerasimenko. Comets display the highest ¹⁵N enhancements among Solar System bodies (Mumma & Charnley 2011). Prior to this study the only other place where such high nitrogen fractionation is found is cold pre-stellar cores (Hily-Blant et al. 2013; Wampfler et al. 2014). The cometary ¹⁵N abundance could thus signify an interstellar inheritance.

Zoom In Zoom Out Reset image size

Interstellar inheritance is not the only possibility, however. Distinguishing between pre-Solar and Solar Nebula origins of nitrogen fractionation requires observations of nitrogen fractionation in protoplanetary disks. In general, HCN (isotopologue) formation begins with atomic nitrogen. At low temperatures ¹⁵N is preferentially incorporated into cyanides resulting in fractionation. Observations of HCN and HNC fractionation toward protostars present a tentative trend of the ¹⁴N/¹⁵N with temperature, supporting this scenario (Wampfler et al. 2014). The second mechanism to produce an enhancement in ¹⁵N in HCN is isotope selective photodissociation of ¹⁴N¹⁵N over ¹⁴N₂, resulting in an enhanced atomic ¹⁵N/¹⁴N ratio. Theoretical models have shown that this mechanism can increase the ¹⁴N/¹⁵N ratio by up to a factor 10 in protoplanetary disks (Heays et al. 2014).

Pre-stellar inheritance, low-temperature chemical fractionation in cold disk regions, and photochemically driven fractionation in disk atmospheres could thus all explain the observed nitrogen fractionation in comets. The presented HCN isotopologue observations in the MWC 480 disk cannot distinguish between these scenarios. Future, higher SNR and/or higher resolution disk studies could, however. The three scenarios should present very different radial and vertical ¹⁴N/¹⁵N profiles. Inheritance should produce an almost uniform ratio across the disk. Low-temperature chemical fractionation should result in a lower ¹⁴N/¹⁵N ratio in HCN in the outer, cold disk that is still protected from radiation and also a decreasing ratio toward the disk midplane. Finally, nitrogen fractionation from selective photodissociation should result in the lowest ¹⁴N/¹⁵N ratio in HCN closer to the disk atmosphere. In addition to detailed studies of individual sources, observations of the averaged ¹⁴N/¹⁵N ratio in disks exposed to different stellar radiation fields will also bring insight into the origin of N fractionation. If low-temperature chemical fractination is the dominant mechanism producing HC¹⁵N, then a multi-source study should result in a trend of increasing ¹⁴N/¹⁵N ratio with stellar radiation field. In contrast, the opposite behavior should be observed if most of the HC¹⁵N is produced by selective photodissociation in the surface layers. These observations will reveal how frequently protoplanetary disks are seeded with ¹⁵N enriched molecules, and whether additional enrichment occurs in the disk phase, changing the fractionation pattern during planet formation. The clear HC¹⁵N detection in MWC 480 after 20 minute integrated demonstrates that such observations will be readily achievable toward a sample of disks in a reasonable time frame.