ATLASGAL -- selected massive clumps in the inner Galaxy. IX. Deuteration of ammonia
M. Wienen, F. Wyrowski, C. M. Walmsley, T. Csengeri, T. Pillai, A. Giannetti, K. M. Menten
aa r X i v : . [ a s t r o - ph . GA ] F e b Astronomy&Astrophysicsmanuscript no. 31208 © ESO 2021February 10, 2021
ATLASGAL-selected massive clumps in the inner Galaxy. IX.Deuteration of ammonia
M. Wienen , F. Wyrowski , C. M. Walmsley , , T. Csengeri , T. Pillai , A. Giannetti and K. M. Menten Max-Planck-Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, 53121 Bonn, Germanye-mail: [email protected] Osservatorio Astrofisico di Arcetri, Largo E. Fermi, 5, I-50125 Firenze, Italy Dublin Institute of Advanced Studies, Fitzwilliam Place 31, Dublin 2, Ireland INAF - Istituto di Radioastronomia, Via Gobetti 101, 40129 Bologna, ItalyReceived
ABSTRACT
Context.
Deuteration has been used as a tracer of the evolutionary phases of low- and high-mass star formation. The APEX TelescopeLarge Area Survey (ATLASGAL) provides an important repository for a detailed statistical study of massive star-forming clumps inthe inner Galactic disc at di ff erent evolutionary phases. Aims.
We study the amount of deuteration using NH D in a representative sample of high-mass clumps discovered by the ATLASGALsurvey covering various evolutionary phases of massive star formation. The deuterium fraction of NH is derived from the NH D1 − ortho transition at ∼
86 GHz and NH D 1 − para line at ∼
110 GHz. This is refined for the first time by measuring theNH D excitation temperature directly with the NH D 2 − para transition at ∼
74 GHz. Any variation of NH deuteration andortho-to-para ratio with the evolutionary sequence is analysed. Methods.
Unbiased spectral line surveys at 3 mm were conducted towards ATLASGAL clumps between 85 and 93 GHz with theMopra telescope and from 84 to 115 GHz using the IRAM 30m telescope. A subsample was followed up in the NH D transitionat 74 GHz with the IRAM 30m telescope. We determined the deuterium fractionation from the column density ratio of NH D andNH and measured the NH D excitation temperature for the first time from the simultaneous modelling of the 74 and 110 GHz lineusing MCWeeds. We searched for trends in NH deuteration with the evolutionary sequence of massive star formation. We derivedthe column density ratio from the 86 and 110 GHz transitions as an estimate of the NH D ortho-to-para ratio.
Results.
We find a large range of the NH D to NH column density ratio up to 1 . ± . deuterationin a subsample of the clumps. Our analysis yields a clear di ff erence between NH and NH D rotational temperatures for a fraction.We therefore advocate observation of the NH D transitions at 74 and 110 GHz simultaneously to determine the NH D temperaturedirectly. We determine a median ortho-to-para column density ratio of 3 . ± . Conclusions.
The high detection rate of NH D confirms a high deuteration previously found in massive star-forming clumps. Usingthe excitation temperature of NH D instead of NH is needed to avoid an overestimation of deuteration. We measure a higher detectionrate of NH D in sources at early evolutionary stages. The deuterium fractionation shows no correlation with evolutionary tracers suchas the NH (1,1) line width, or rotational temperature. Key words.
Surveys — Submillimeter — Radio lines: ISM — ISM: molecules — Stars: massive — Stars: formation
1. Introduction
High-mass stars are known to form in dense clusters. They aremuch rarer than low-mass stars according to the stellar initialmass function (Kroupa et al. 2013) and are therefore located atgreater distances. Massive protostars evolve embedded in densecores ( ∼ − cm − , Garay & Lizano 1999; Kurtz et al.2000) within high-mass star-forming complexes. These are morecrowded than low-mass star-forming regions and have a shortevolutionary timescale of ∼ yr (McKee & Tan 2002). Theseconstraints restrict observations of the early phases of high-massstar formation. However, a key issue preventing a more com-plete understanding of the formation process of massive stars isthe di ffi culty in revealing their initial conditions.The abundance of deuterium bound in molecules is ordersof magnitude higher in cold molecular clouds than the primor-dial D / H ratio ( ∼ − , Oliveira et al. 2003). Rising deutera-tion is expected from chemical models even into the gravita-tional collapse phase of the molecular cloud core (Caselli 2002;K¨ortgen et al. 2017). Deuterated molecules can form through re-actions between gaseous species as well as through the depletion of those onto grains with subsequent deuteration on the surfacesfollowed by the evaporation of icy grain mantles by the radiationfrom protostars back into the gas. At low temperatures ( <
20 K)and for low ortho-to-para H ratios, deuterium fractionation isprimarily regulated by reactive collisionsH + + HD → H D + + H + ∆ E . (1)The production of H D + is essential for the deuterium chemistry,representing the first stage of deuterium enrichment (“deuter-ation”, Roberts & Millar 2000b; Wu & Yang 2005; Pillai et al.2007). Gas-grain models comparing deuteration of H + at gasand dust temperatures of 10 K and 20 K by Sipil¨a et al. (2015a)lead to a decrease in deuteration at the higher temperature, atwhich the reaction given in equation 1 proceeds more e ffi cientlyin the backward direction. This trend is favoured by a highortho-to-para H ratio as well. Moreover, neutral molecules suchas CO and H destroy H D + at temperatures above 25 K andthus reduce the deuterium fraction (Roberts & Millar 2000b).According to the gas-grain models from Sipil¨a et al. (2015a), atotal depletion of ammonia from the gas phase also occurs after
1. Wienen et al.: ATLASGAL - NH deuteration ∼ years at a density of 10 cm − and a temperature of 20K. In addition to ammonia, various forms of deuterated ammo-nia are also depleted onto grain surfaces with deuterium beingtrapped onto the surfaces (Sipil¨a et al. 2015b). This leads to HDdepletion and the decrease of the overall gas-phase deuteratione ffi ciency (Sipil¨a et al. 2015a).It is known from observations and theory that C-bearingmolecules such as CO freeze out onto dust grains in thecold and dense environment of molecular cores (Caselli et al.1999; Kramer et al. 1999; Tafalla et al. 2002; Flower et al. 2005;Bergin & Tafalla 2007; Giannetti et al. 2014) and therefore in-crease the [H D + ] / [H + ] ratio, which results in an enhancedabundance of deuterated species in the very early evolutionaryphase. Csengeri et al. (2014) estimated that 25% of the embed-ded sources in the ATLASGAL (Schuller et al. 2009) samplewith a peak intensity > ortho-to-pararatio as well. If ortho instead of para H is present, the backwardreaction destroying H D + will be faster because of the four-times-larger ortho than para H rate coe ffi cients (Pagani et al.1992). In addition, if the abundance of ortho H is high, it wille ffi ciently destroy H D + at low temperatures. A lower H ortho-to-para ratio, as found in cold cores (Pagani et al. 1992), conse-quently leads to a higher deuterium enrichment.NH D has been detected in low- and high-mass star-formingregions: it was observed in cold dark clouds by Saito et al.(2000), in low-mass protostellar cores by Shah & Wootten(2001), and in low-mass protostars by Hatchell (2003). Theseauthors measured NH deuteration factors between 0.001 and0.3 with similar errors of ∼
25% on average, while interfero-metric observations with high angular resolution of NH D andNH by Crapsi et al. (2007) found an enhanced deuterium frac-tionation of 0 . ± . cm − in the centreof a nearby Taurus core. In massive star-forming regions, NH Dwas observed for example in pre- and protocluster clumps byPillai et al. (2007) with half of the sample exhibiting a high deu-terium fraction of ≥ D + / N H + ratio in low-mass starless coresand protostars shows the predicted relation of a decreasingdeuterium fractionation from the youngest objects immedi-ately after the beginning of collapse to the more advancedevolutionary state of a Class 0 protostar (Crapsi et al. 2005;Emprechtinger et al. 2009). It is suggested that the NH deuter-ation increases in low-mass cores up to 20 K and is con-stant at higher temperatures (Shah & Wootten 2001). However,these authors only observed a small sample with large errorsin the deuteration factors. Moreover, an enhanced N D + / N H + ratio was also measured at the earliest evolutionary stages ofhigh-mass star formation and a decline from high-mass star-less core candidates to high-mass protostellar objects and ultra-compact (UC) HII regions was found by Fontani et al. (2011).Busquet et al. (2010a) were able to use the [NH D] / [NH ] ra-tio as an evolutionary indicator in the environment of an ultra-compact HII region (UCHIIR). Fontani et al. (2015) comparedNH deuteration in high-mass cores with the evolutionary se-quence and found that the [NH D] / [NH ] ratio determined inmassive starless cores, high-mass protostellar objects, and ul-tracompact HII regions does not decrease with the evolutionof the cores. Two high-resolution studies (Busquet et al. 2010a;Pillai et al. 2011) report that very close to high-mass protostarsand UCHII regions (few 1000 AU), there is evidence of removalof deuterated NH . On large scales Fontani et al. (2015) deter- mined a deuterated fraction of NH above 0.1 right up to themost evolved phase of their single-dish sample. This indicatesno evidence of gas-phase removal on the envelope scales (sev-eral 10000 AU) up to the most advanced evolutionary stage.Existing single-dish and high-resolution data therefore suggestthat deuteration is becoming ine ffi cient (very little additionaldeuteration taking place) on large scales and appears to be moreor less fully removed in the immediate vicinity of protostars andHII regions.Previous studies of deuterated ammonia in high-mass starformation have only focused on small samples or one evolution-ary stage. In the present paper, we determine the NH deutera-tion of a representative sample of massive clumps that are de-tected by the ATLASGAL survey and observed in NH D as partof an unbiased spectral line survey. Our analysis focuses on theinfluence of temperature on deuteration. In particular, this sam-ple covers various phases of high-mass star formation and allowsus to analyse any dependence of the deuterium fractionation onevolutionary stage.We present the NH D observations of the ∼
86 and ∼ D column density, excitation, andNH deuteration in Sect. 3. In addition, we measure the columndensity ratio of NH D at ∼
86 and ∼
110 GHz as an estimate ofthe ortho-to-para ratio. We compare the NH and NH D temper-atures and analyse any trend of NH deuteration with evolution-ary tracers in Sect. 4. Moreover, we study the dependence of theortho-to-para ratio on the NH deuteration, line width, and ro-tational temperature. Our NH D analysis is summarised in Sect.5.
2. Observations
The NH D data were observed within two unbiased spectral-line follow-up observations of large ATLASGAL subsamples.The first project covered 8 GHz centred on 89 GHz in the fourthquadrant with the Mopra telescope (Urquhart et al. 2019) locatednear Coonabarabran in Australia at a latitude and longitude of -31.2678 ◦ and 149.0997 ◦ . The second survey covered the whole3 mm band in the first quadrant using the EMIR receiver at theIRAM 30m telescope (Csengeri et al. 2016). We summarise theNH D and NH transitions, that are used in this article, with theirspectroscopic properties in Table 1. The Mopra 22m telescope was used to observe a 3mmmolecular-line survey towards an unbiased ATLASGAL sub-sample of clumps with infrared association and peak fluxesabove 1.75 Jy / beam at 870 µ m as well as cold sources with peakfluxes above 1.2 Jy / beam (Urquhart et al. 2019). We observed567 ATLASGAL sources located between l = ◦ and 359 ◦ and | b | ≤ . ◦ in 2008 and 2009.This article focuses only on observations of the NH D 1 − ortho transition at 85.926 GHz in the fourth quadrant. Weused a 3mm HEMT receiver as frontend. Our measurements inthe 3 mm band range from a frequency of ∼ . ∼ . ∼ IRAM is supported by INSU / CNRS (France), MPG (Germany) andIGN (Spain)2. Wienen et al.: ATLASGAL - NH deuteration Table 1.
Properties of the NH D and NH transitions. Molecule Quantum numbers Frequency upper energy level statistical weight A ij critical density Reference(GHz) (K) of upper / lower level (1 / s) (1 / cm )NH D ortho 1 - 1 . × − . × a CDMS c NH D para 1 - 1 . × − . × a CDMSNH D para 2 - 2 . × − . × a CDMSNH para 1,1 a - 1,1 s 23.694 23.3 6 1 . × − . × b JPL c NH para 2,2 a - 2,2 s 23.722 64.4 10 2 . × − . × b JPL
Notes. ( a ) The de-excitation rate coe ffi cients are taken at 10 K from Daniel et al. (2014). ( b ) The de-excitation rate coe ffi cients at 15 K fromDanby et al. (1988) are used. ( c ) The statistical weight given by CDMS includes a factor three for the rotation quantum number, a factor three forthe spin of the two H atoms, and the N nuclear spin multiplicity of three. On the contrary, JPL does not include the nuclear spin of N. broadband mode, where each 2.2 GHz wide band has a velocityresolution of 0.9 km s − . The Mopra telescope has a beamwidth(FWHM) of 38 ′′ at the frequency of the NH D line at ∼
86 GHz.Pointed observations were conducted in position-switchingmode. We examined the region around each source usingATLASGAL and infrared continuum maps from the MidcourseSpace Experiment (MSX, Price et al. 2001) and chose an o ff -set position that is free of continuum emission at 20 µ m, either ± ′ in longitude or latitude. We observed two polarisations ofthe NH D line at 86 GHz simultaneously. The total integrationtime for each source was ∼
15 min, resulting in an rms noiselevel of 24 mK on average at a velocity resolution of 0.9 km s − .The median system temperature was about 200 K. Pointing wasmeasured each hour with line pointings on SiO masers and areference spectrum of G327 and M17 was obtained each day.We processed the data initially with the ASAP package,which consisted of processing of the on-o ff observing mode, thetime and polarisation averaging, and baseline subtraction. Weconverted the data to the T ∗ A temperature scale and exported thedata to the CLASS software from the GILDAS package for sub-sequent analysis. For the calibration from T ∗ A to T MB we cor-rected for the beam e ffi ciency of 0.49 (Ladd et al. 2005). The NH D 1 − ortho line at 85.926 GHz and the para tran-sition at 110.154 GHz in the first quadrant were measured aspart of the large molecular line survey of ATLASGAL sourcesconducted with the IRAM 30m telescope in 2011 and 2012(Csengeri et al. 2016). The sample covers a range of evolution-ary phases from the quiescent clumps to actively star-formingclumps hosting HII regions. The IRAM sample targeted brightsources, but also includes infrared-quiet clumps meaning with-out a detection at 22 µ m corresponding to roughly half of the tar-geted sources (see more details in Csengeri et al. 2016). Whilethe Mopra sample consists of clumps with and without infraredassociation based on MSX data at 21 micron, the IRAM sampleused the WISE 22 micron point-source catalogue with higherresolution. In a pilot study, 36 sources with the highest submil-limetre (submm) peak flux densities from the ATLASGAL sur-vey were observed in April 2011 and a second large sample wasfollowed up in February, March, and October 2012. Using theIRAM 30m telescope we observed 425 ATLASGAL sources inNH D within l = ◦ − ◦ and | b | ≤ . ◦ .The observations were carried out with the EMIR receiverE090. The frequency range from ∼
84 to 115 GHz was dividedinto 4 GHz blocks for the pilot study, while the sample in 2012was observed with a total bandwidth of 16 GHz and in two se- available at http: // / IRAMFR / GILDAS tups centred on 88 and 96 GHz. The Fast Fourier TransformSpectrometer (FFTS) was used with a spectral resolution of 200kHz resulting in a velocity resolution of 0.68 km s − at ∼
86 GHzand of 0.53 km s − at ∼
110 GHz. The half-power beam width atthe NH D 1 − line frequencies at ∼
86 GHz and ∼
110 GHzis 29 ′′ and 22 ′′ . The spectra were converted to the main beambrightness temperature scale for the beam e ffi ciency of 0.81 at ∼
86 and 110 GHz as in Csengeri et al. (2016).The observations were conducted in position switchingmode with a constant o ff set of 10 ′ in right ascension and decli-nation with a total integration time of ∼ . + D 2 − para transition at 74.156 GHz was ob-served toward a subsample of the 24 brightest clumps in deuter-ated ammonia that was selected from the NH D observationsat ∼
86 GHz covering di ff erent evolutionary phases such as24 µ m dark sources, active clumps in IRDCs, and HII regions.The EMIR receiver with a frequency range between 71 and 79GHz was used for position-switching observations toward thepeaks of the clumps. Typical system temperatures were about180 K. We measure an rms noise level of 15 mK at a velocityresolution of 0.75 km s − , which is similar to the observations at ∼
86 GHz. A total integration time of ∼
60 minutes was spentper source including on and o ff position. The CLASS software was used to reduce the NH D data. To re-move the baseline from the spectra we subtracted a polynomialbaseline of order zero from the spectra, excluding velocity win-dows that were placed around the NH D lines. The hyperfinestructure of the NH D transitions at 86 and 110 GHz was fittedtaking six hyperfine components into account. The fit of the lineat 110 GHz kept the line width as a fixed parameter using theNH D line width at 86 GHz assuming that the two transitionsoriginate from the same gas. This gives the optical depth of themain line, τ , the radial velocity, v LSR , and the line width, ∆ v, atthe full width at half maximum of a Gaussian profile with theirerrors as the formal fit errors from CLASS. As the line widthof the transition at 110 GHz is set to a fixed value, no error isindicated for this parameter in Table 3. The temperature of theNH D line was measured from the peak of the hyperfine struc-ture fit. The minimum optical depth of the hyperfine structurefit in CLASS is 0.1. Because the hyperfine structure of someNH D lines at 86 GHz and of most transitions at 110 GHz aretoo weak to be detected, we cannot determine their optical depth.In these cases, the fit from the CLASS software gives an errorof the optical depth of greater than 50%. The 86 GHz-NH D
3. Wienen et al.: ATLASGAL - NH deuteration Fig. 1.
Examples of reduced and calibrated spectra of observed NH D transitions at 86 GHz; the fit is shown in green. The hyperfinestructure of the spectra in the first and second rows are clearly detected, while that of the spectra in the third row is too weak to bevisible. Frequencies of the hyperfine structure components are indicated by straight lines.
Fig. 2.
Examples of reduced and calibrated spectra of observed NH D lines at 110 GHz; the fit is indicated in green. Frequencies ofthe hyperfine structure components are labelled.line parameters with the NH rotational temperature between the(1,1) and (2,2) inversion transition from Wienen et al. (2012) andWienen et al. (2018) are given in Table 2. We report the molecu-lar line parameters of the para NH D transition in Table 3 whichlists the line-of-sight velocity of para NH D, v
LSR , the linewidth, ∆ v , and the main beam brightness temperature, T MB .
3. Results and analysis of the NH D sample D detectionrates
We observed 992 ATLASGAL clumps in NH D 1 - 1 at 86GHz in the first and fourth quadrant and detected NH D in 390clumps (39%) corresponding to a S / N ratio >
3. We found thatthe NH D velocities of all detected sources are within ∼ . − of the NH velocities. The hyperfine structure of the NH D line at 86 GHz is clearlyvisible in 79 clumps in the first and fourth quadrants (20%).We can determine the optical depth of these sources, whichranges between 0 . ± .
02 for G15.22 − . ± . + D line at86 GHz of 1 . ± .
35. A few spectra of the NH D transition at86 GHz of these sources and of the clumps, for which the hyper-fine structure is too weak to obtain an optical depth, are shownin Fig. 1.The NH D 1 - 1 line at 110 GHz was observed in 373ATLASGAL sources in the first quadrant of which 65 clumpsare detected (17%) with S / N > D and NH ve-locities of all detections are within ∼ . − . The hyperfinestructure components of most clumps are blended and we detectthose in only seven sources (2%) with an optical depth between
4. Wienen et al.: ATLASGAL - NH deuteration Table 2.
Parameters of the NH D line at 86 GHz and NH rotational temperature (Wienen et al. 2018). Errors are given in paren-theses. The full table is available at CDS. RA d Dec d τ (1,1) e v LSR ∆ v T MB T rot Name (J2000) (J2000) (km s − ) (km s − ) (K) (K)G10.62-0.42 18 10 36.92 -19 57 00.86 0.94 ( ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + +
01 14 57.91 0.12 ( ± ± ± ± ± +
14 06 37.09 0.21 ( ± ± ± ± ± +
14 20 14.86 0.15 ( ± ± ± ± ± +
14 30 49.66 0.80 ( ± ± ± ± ± + +
22 46 37.76 0.10 ( ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Notes. ( d ) Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. ( e ) Thesmallest optical depth given by the CLASS software is 0.1. Sources without detected hyperfine structure or no reliable derivation of the opticaldepth due to low S / N (see Sect. 3.3) are marked by a star.
Table 3.
Line parameters of the NH D transition at 110 GHz with errors noted in parentheses. We made the full table available atCDS. RA f Dec f v LSR ∆ v T MB Name (J2000) (J2000) (km s − ) (km s − ) (K)G14.33-0.64 18 18 54.59 -16 47 41.16 22.72( ± ± ± ± ± ± ± ± + ± ± + ± ± +
14 06 37.09 68.56( ± ± +
14 20 14.85 69.60( ± ± + +
22 46 37.76 37.0( ± ± Notes. ( f ) Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds.
Table 4.
Parameters of the NH D line at 74 GHz. Errors are given in parentheses. The full table is available at CDS. RA g Dec g v LSR ∆ v T rot NH2D
Name (J2000) (J2000) (km s − ) (km s − ) (K)G12.50-0.22 18 13 41.49 -18 12 35.51 36.31 - 17.7G14.33-0.64 18 18 54.59 -16 47 41.16 22.75( + . − . ) 2.7( + . − . ) 22.8 ( + . − . )G19.88-0.54 18 29 14.53 -11 50 25.67 41.87( + . − . ) 4.1 ( + . − . ) 20.0 ( + . − . )G23.21-0.38 18 34 55.02 -08 49 16.96 77.87 - 58.9G27.37-0.17 18 41 51.25 -05 01 42.71 90.76( + . − . ) 3.5 ( + . − . ) 18.7 ( + . − . )G30.85-0.08 18 47 55.54 -01 53 33.38 98.27 - 12.0G31.41 + + . − . ) 5.9 ( + . − . ) 36.5 ( + . − . )G34.26 + + . − . ) 5.7 ( + . − . ) 41.6 ( + . − . ) Notes. ( g ) Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. . ± .
02 for G27.37 − . ± . − . ± .
8. Examples of the NH D transition at 110 GHzare presented in Fig. 2.We measured the NH D transition at 74 GHz in 24ATLASGAL sources in the first quadrant and 5 sources haveS / N >
3. These detections possess NH D and NH velocities that again lie close together, within ∼ . − . The results ofthe modelling of five detections are plotted in Fig. 4. D linewidth
We derive the NH D intrinsic line width from hyperfine struc-ture fits to the transition at 86 GHz, and find values between 1.1
5. Wienen et al.: ATLASGAL - NH deuteration Fig. 3.
Line width of the NH D transition at 86 GHz plottedagainst the NH (1,1) line width. ATLASGAL sources that havean error in the NH D optical depth of less than or greater than50% are indicated as red triangles or black points, respectively.The straight line corresponds to equal line widths.km s − and 7.7 km s − . Fitting of NH D lines with a small S / Nand an error in the optical depth of greater than 50% gives op-tical depths varying between 0.1 and 2.5, which causes the linewidth to vary by ∼ < D observations within the NH sample of ATLASGALsources measured in the fourth quadrant (Wienen et al. 2018)using the Parkes telescope. This resulted in an ATLASGALsubsample of 264 clumps detected in NH D and NH withina FWHM beamwidth of the Parkes telescope of 60 ′′ , slightlylarger than the FWHM beamwidth of the IRAM telescope of29 ′′ at ∼
86 GHz. The NH D line widths are compared withthe NH (1,1) line widths obtained from hyperfine structure fitsin Fig. 3, where the sources with an error in the optical depthsmaller than 50% are shown as red triangles and the clumps withan error in the optical depth larger than 50% as black points.The straight line indicates equal line widths. The whole sam-ple is distributed equally around the straight line and this hintsat a correlation between the NH D and NH line width withinthe noise. Figure 3 suggests that the NH D line at 86 GHz andNH (1,1) line therefore trace similar regions within a source.Although the critical density of the NH D transition at 86 GHzis about a factor 50 higher than that of the NH lines (see Table1), studies of high-mass star-forming regions also reveal an ap-proximate spatial correlation between the emission from the twomolecules (Busquet et al. 2010a; Pillai et al. 2011). The medianNH D line width of ∼ − agrees with the average NH (1,1) line width measured for the whole ATLASGAL sample inWienen et al. (2012). The red contour lines of the sources withan error in the optical depth of less than 50% indicate slightlysmaller NH D than NH line width. D columndensity
The total column density of ortho and para NH D is derived as-suming that the energy levels are in LTE, that is, that they are populated according to a Boltzmann distribution. Non-LTE con-ditions are pointed out in Sect. 3.4. To calculate the column den-sity we distinguish between subsamples with and without de-tected hyperfine structure. For clumps without detected hyper-fine structure we mark the optical depth given in Table 2 with astar, while we give the optical depth of sources with detected hy-perfine structure without a star. For the frequency of the NH Dline at ∼
86 GHz we use the Rayleigh-Jeans approximation fora mean kinetic temperature of our sample at 20 K. We calcu-late the source-averaged column density of sources, for whichthe hyperfine structure components are detected and their ratioprovides a measurement of the optical depth. For these clumps,the optical depth is well determined and the column density isderived from the lower energy level by N l = π kc h g l g u A − ν T ex τ ∆ v , (2)with the statistical weight of the upper and lower levels, g u and g l , respectively, the Einstein A coe ffi cient in s − , the frequencyof the NH D transition at 86 GHz, ν , in GHz, the excitationtemperature, T ex , in K, the optical depth of the NH D transi-tion, τ , and the NH D FWHM line width, ∆ v, in km s − . Weuse the NH kinetic temperature given in Wienen et al. (2012)and Wienen et al. (2018) as excitation temperature assuming thatNH D and NH are co-spatial, as indicated in Fig. 3, and there-fore have similar gas temperatures. For sources with detected hy-perfine structure we calculate the total NH D column density incm − from the optical depth and kinetic temperature, which de-pend on line ratios, and we therefore compute a source-averagedquantity by N tot = . × ν A − Q ( T ex ) g u exp E u T ex ! T ex τ ∆ v , (3)where Q is the partition function and E u the upper energy level inK (see Table 1). Clumps without detected hyperfine structure andinsu ffi cient S / N to reliably determine an optical depth have opti-cal depths with an error of greater than 50% in Table 2. We usethe integrated intensity of the NH D line at 86 GHz, R T mb dv,in K km s − , that is a measure derived over the whole beam, tocalculate the beam-averaged column density in the optically thincase, N tot = . × ν A − T ex R T mb dv T ex − . E u T ex ! Q ( T ex ) g ; (4)we derive Q from a fit to the partition functions measured fordi ff erent rotation temperatures in the range between 9 and 300K that was taken from the Cologne Database for MolecularSpectroscopy (CDMS) . The measured values of the partitionfunctions listed on the CDMS are determined from the sum ofthe population of the 86 and 110 GHz transitions and take thespin multiplicity of the N nucleus, g I =
3, into account. The fityields a total partition function for ortho and para NH D of Q = . T . . We obtain the statistical weight from g = g J × g I = =
1, the N and D nuclear spinsof 1, and the H nuclear spin of 1 / D excitation
Furthermore, we test the assumption that the NH rotational tem-perature between the (1,1) and (2,2) inversion transition is equal see https: // / cdms Because we work with the partition function and molecular lineparameters from CDMS, the D nuclear spin is not taken into account.6. Wienen et al.: ATLASGAL - NH deuteration Fig. 4.
Examples of reduced and calibrated spectra of observed NH D lines at 110 GHz and 74 GHz. The bright line of G31.41 + + OCH ; their frequencies are labelledas lines in the spectra. Results of simultaneous modelling of the NH D transitions at 110 and 74 GHz (see Sect. 3.4) are illustratedin red.to the NH D excitation temperature using the NH D 1 − line at 110 GHz and the NH D 2 − transition at 74 GHz thatwas observed towards a subsample of 24 ATLASGAL sources.The rotational temperature between the NH D transitions at 110GHz and 74 GHz is determined from the simultaneous mod-elling of the 74 GHz and 110 GHz lines (see Fig. 4) using MCWeeds (Giannetti et al. 2017). This assumes equal excitationtemperatures of the 1 − and 2 − lines under LTE con-ditions. This package is based on WEEDS (Maret et al. 2011)from the CLASS software and adds Bayesian statistics and fit-ting algorithms to Weeds to automatise the simultaneous fittingof the lines of several species. Furthermore, errors on the NH D
7. Wienen et al.: ATLASGAL - NH deuteration Table 5.
Source-averaged NH and ortho NH D column densities and deuteration. Errors are given in parentheses. The full table isavailable at the CDS. N NH η N NH D (86 GHz) [NH D] / [NH ] R T mb dv (86 GHz)Name (10 cm − ) (10 cm − ) (K km s − )G10.62-0.42 2.33 ( ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± temperature are also estimated by MCWeeds. It determines theNH D rotational temperature for given starting values of the in-put parameters, that is, the NH D column density, temperature,line width, and source size. As starting values we used the re-sults from our hyperfine structure fitting of the NH D transitionat 110 GHz and the NH (1,1) and (2,2) lines. The fit parame-ters, the radial velocity of para NH D at 74 GHz, v
LSR , the linewidth, ∆ v , and the NH D temperature, T rot , are given in Table4. Fig. 5.
Observed NH rotational temperature between the (1,1)and (2,2) inversion transition is shown against modelled NH Drotational temperature between the 1 − line at 110 GHz andthe 2 − transition at 74 GHz. Sources with a detected 74GHz line are marked in red, and non-detections with an upperlimit of the NH D temperature are labelled as black arrows. Thestraight line corresponds to equal temperatures.While modelling of the two NH D lines leads to the NH Dtemperature for sources with a detected NH D transition at 74GHz, we obtained an upper limit to the NH D temperature forthe non-detections. Comparison of the rotational temperature be-tween the NH (1,1) and (2,2) inversion transition (Wienen et al.2012) with the NH D temperature in Fig. 5 yields a di ff erence in the NH D and NH rotational temperatures for a subsampleof the sources. However, a fraction of the clumps exhibit largerNH than NH D rotational temperatures, in some cases sup-ported by upper limits to the NH D temperature. We note thatthe critical density is proportional to the Einstein A-coe ffi cient,which in turn is proportional to ν . A factor 4.6 higher frequencyof NH D at 110 GHz than of the NH (1,1) and (2,2) lines leadsto a much higher critical density of NH D than NH (see Table1). In addition to our assumption of the same excitation tem-perature for the 74 GHz and 110 GHz lines as there wouldbe for LTE, we also run a non-LTE modelling using RADEX(van der Tak et al. 2007) with a fixed kinetic temperature of 20K and a fixed NH D column density of 1 . × cm − corre-sponding to the mean values of our ATLASGAL sample. We usethe recent calculations by Daniel et al. (2014) for the collisionalrates with the H molecule for our non-LTE calculations. Thisresulted in similar excitation temperatures of the 74 GHz and110 GHz NH D lines at high densities ( > cm − ). The fivedetections at 74 GHz from our sample are classified as youngstellar objects (YSOs), hot cores, or HII regions. Urquhart etal. (in prep.) determined volume densities of ∼ cm − basedon the dust continuum from the ATLASGAL survey for thesesources, which are also consistent with the typical density ofYSOs probed in NH D by Pillai et al. (2011).We compared NH D column densities using the NH rota-tional temperature and NH D temperature as excitation temper-ature for sources with detections at 74 and 110 GHz. This showsthat for the subsample with larger NH rotational temperaturesthan NH D temperatures we overestimate the NH deuterationby 47%. Similarly, we underestimate the NH deuteration by ∼
32% for the subsample with lower NH rotational temperaturethan the NH D temperature. We cannot exclude subthermal ex-citation due to densities below the NH D line critical densitiesthat lead to lower temperatures. deuteration We determine the fractionation ratio by the total column den-sity of ortho and para NH D (see Sect. 3.3) to the NH col-umn density ratio assuming that the two molecules originate
8. Wienen et al.: ATLASGAL - NH deuteration Fig. 6. NH D to NH column density ratio compared with theNH rotational temperature, line width, and the MSX 21 µ mflux for the ATLASGAL sources; clumps with and without de-tected hyperfine structure are shown with red triangles and blackpoints, respectively.from the same gas. The source-averaged NH column densitywas calculated for ATLASGAL sources in the first quadrant inWienen et al. (2012) and for clumps in the fourth quadrant inWienen et al. (2018). To account for the di ff erent column den-sity determinations of NH D (see Sect. 3.3 source- vs. beam- averaged) we derive the NH deuteration in two ways. We di-vide the source-averaged NH D and NH column densities forthe NH D sources with detected hyperfine structure. In the op-tically thin case we correct the NH D column density for thebeam dilution using the beam filling factor derived from NH observations (see Sect. 4.5 in Wienen et al. 2012) to estimate thesource-averaged NH D to NH column density ratio. Table 5shows the source-averaged NH column density, the beam fill-ing factor, the NH D column density, the NH D to NH columndensity ratio, and the integrated intensity of the NH D line at 86GHz, R T mb dv, with their errors. We measure [NH D] / [NH ]ratios between 0.007 and 1.6. The distribution of the deuteriumfraction of NH is shown as a function of the rotational tem-perature in the upper panel of Fig. 6. Optically thin fits resultin lower limits, because the NH D optical depth is not knownfor sources without detected hyperfine structure and we cannotexclude a considerable optical depth. We might therefore under-estimate the column density by using equation 4 in the opticallythin case. However, there is still a range of one order of magni-tude in the fractionation ratio. We consider only measurementswith relative errors on the presented NH D line parameters ofless than ∼
50% for the correlation plots. In 109 out of the 264clumps detected in NH D and NH (41%) with reliable hyper-fine structure fits we determine high deuteration ( > D] / [NH ] ratio of greater than50%. D For a subsample of 113 ATLASGAL sources we detect theNH D ortho line at 86 and para line at 110 GHz. We calculate theortho-to-para column density ratio assuming the same beam fill-ing for the two transitions, as expected for unresolved, clumpystructure observed within molecular clouds that fill the beam(Stutzki & Guesten 1990; Perault et al. 1985). We distinguishbetween three di ff erent cases depending on the optical depth ofthe source: We divide the source-averaged column densities de-rived from the transition at 86 and 110 GHz as given in equation3 with the molecular line parameters listed in Table 1 for sourceswith detected hyperfine structure in ortho and para NH D. Here,Q is determined from separate fits to the partition functions ofortho and para NH D at 86 and 110 GHz against rotation tem-peratures from 3 to 300 K published on the CDMS. These resultin Q = . T . for the ortho line and Q = . T . for thepara line. The ortho and para partition functions di ff er by a fac-tor three which results from the spin multiplicity. Our fit of thetwo partition functions yields an exponent of 3 /
2, as expected forslightly asymmetric top molecules (Mangum & Shirley 2015).For sources without detected hyperfine structure in ortho andpara NH D we derive the beam-averaged NH D column densi-ties using equation 4 with the observed intensities and molecularline parameters for the transitions at 86 and 110 GHz given inTable 1 and separate ortho and para partition functions. To cor-rect the NH D column densities for the beam dilution we dividethe two by the beam filling factor; they cancel out in the compu-tation of the ortho-to-para ratio from the column density ratio.As we cannot detect the hyperfine structure of the NH D lineat 110 GHz for the majority of sources, we cannot measure theoptical depth at 110 GHz directly. To derive the ortho-to-para ra-tio for the subsample with detected hyperfine structure in orthoNH D we calculated the column density of para NH D in the op-tically thin approximation from equation 4 using the molecular
9. Wienen et al.: ATLASGAL - NH deuteration line parameters at 110 GHz and the para partition function. Weget an estimate of the optical depth at 110 GHz iteratively fromthe known optical depth at 86 GHz: We start with an ortho-to-para ratio of three as expected from their statistical values, calcu-late the optical depth at 110 GHz by the ratio of the optical depthat 86 GHz to the ortho-to-para ratio, and compute the columndensity ratio determined from the lines at 86 and 110 GHz. Thisis used subsequently as the ortho-to-para ratio in the calculationof the optical depth at 110 GHz until the ortho-to-para ratio con-verges. With the resulting optical depth at 110 GHz we multiplythe column density of para NH D by the factor τ ′ / (1 − exp( − τ ′ ))with τ ′ = . τ . , which corrects for the optical depth ofthe line at 110 GHz (Stutzki et al. 1989), and divide by the beamfilling factor. The ortho-to-para ratio is then determined from thesource-averaged ortho and para column densities. We also testedif the ortho-to-para ratio depends on the initial value chosen forthis ratio. We therefore varied this from 3 to 1 and 5, but the iter-atively determined ortho-to-para ratios of the whole sample didnot change statistically.Using an initial ortho-to-para ratio of 3 we obtain the distributionof the iteratively derived ortho-to-para ratios shown by the blackhistogram in Fig. 7 with a median ortho-to-para ratio of 3.7 and astandard deviation of 1.2. Our ortho-to-para ratio is close to thevalue expected from the nuclear statistical weights of 3 and 1for the ortho and para NH D species, respectively. The columndensity derived from the NH D transition at 110 GHz and theortho-to-para ratio with their errors are given in Table 6.
Fig. 7.
Number distribution of the column density ratio derivedfrom the ortho and para NH D transitions shown in black for thesubsample detected in NH D with a NH counterpart, as dashedred line for ATLASGAL sources with a NH deuteration < . > .
4. Discussion and NH D rotationaltemperature
The NH rotational temperature between the (1,1) and (2,2) in-version transition is compared with the rotational temperaturebetween the NH D transitions at 110 GHz and 74 GHz in Fig.5. This comparison reveals a large range of variation in temper-ature; two clumps, G31.41 + + D temperatures of 37 and 42 K. The high ex-citation temperature points to the hot molecular core as the main origin of the deuterated ammonia in these sources. We foundlittle deuterium fractionation for G31.41 + + D have shown that while NH D shows an excellentcorrelation with dust continuum in high-mass cold cores, deuter-ated ammonia avoids the dense peaks close to very luminousprotostars (Busquet et al. 2010b; Pillai et al. 2011). While thismight be a temperature e ff ect in regions with complex dynam-ics, without further high-angular-resolution observations, we areunable to confirm that the major contributing factor is deutera-tion. deuterationinATLASGALsourceswithother samples We derive a deuterium fraction between 0.007 and 1.6 inSect. 3.5, sources without detected hyperfine structure exhibitlow [NH D] / [NH ] ratios with a median of 0 . ± . , andclumps with detected hyperfine structure in NH D and NH have a higher deuteration with a median of 0 . ± .
05. The[NH D] / [NH ] ratios of approximately 1 are among the high-est given in the literature so far. Other high average deuterationfactors were estimated to be 0.8 for starless cores associated withthe UCHIIR IRAS 20293 + D] / [NH ] ratios best de-termined, and a deuteration of 0.1 was found in dark cloudsby Tin´e et al. (2000). Shah & Wootten (2001) observed NH Din low-mass and quiescent protostellar cores and measureda deuteration factor between 0.003 and 0.13, and Saito et al.(2000) determined a similar range in deuteration in dark cloudcores located mostly in the Taurus and Ophiuchus regions. Acomparison of the deuterium fraction of NH towards the low-mass samples shows that Saito et al. (2000) and Shah & Wootten(2001) estimated lower values on average than Hatchell (2003).This latter author suggested an increase in deuteration fromlarger to smaller scales towards protostellar cores as reasonfor the low [NH D] / [NH ] ratio obtained by Shah & Wootten(2001) and Saito et al. (2000), who measured NH D with alarger beam width than Hatchell (2003) or with an o ff set fromthe dust peak. We summarise the deuterium fractionation of thedi ff erent samples from the literature in Table 7 with the sourcenumber, the NH D detection rate, the NH D to NH columndensity ratio, and reference. Most of these studies calculatedthe NH D column density under the assumption of LTE; onlyCrapsi et al. (2007) and Shah & Wootten (2001) used non-LTEmodels.For comparison of the NH deuteration in the ATLASGALsample other studies with similar beamwidth as our NH andNH D measurements are available in the literature: Fontani et al.(2015) observed the two molecules in dense cores associatedwith di ff erent evolutionary phases of high-mass star forma-tion and determined a deuterium fraction of NH between 0.21and 0.34. Pillai et al. (2007) found a deuterium fractionationfrom 0.004 to 0.7 that is similar for most clumps embedded ininfrared-dark clouds. These [NH D] / [NH ] ratios are consistentwith those of the ATLASGAL sources, although we estimate
10. Wienen et al.: ATLASGAL - NH deuteration Table 6.
Column density derived from the NH D line at 110 GHz and ortho-to-para ratio. Errors are given in parentheses. The fulltable is available at the CDS. N NH D (110 GHz) N ortho (86GHz) / N para (110GHz)Name (10 cm − )G10.21-0.30 0.19 ( ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± + ± ± ± ± + ± ± + ± ± ± ± ± ± ± ± + ± ± an even higher deuteration than Pillai et al. (2007), up to 1.6 ina few ATLASGAL clumps. Recently, Sipil¨a et al. (2019) com-pared two approaches to model deuterium fractionation that dif-fer in their mechanism to describe ion-molecule proton-donationreactions. The full scrambling model comprises multiple inter-changes of atoms including for example proton hop and protonexchange (Oka 2004). The time evolution of their full scram-bling model (Sipil¨a et al. 2015b) leads to a deuteration of theorder of 10 − over one phase, similar to models of earlier stud-ies (Roberts & Millar 2000b; Millar 2002). To reach such a highrate as measured towards the ATLASGAL sample we specu-late that each of several clumps within the beam goes througha cycle of enrichment consisting of freeze-out and evaporationduring a dense and cold phase. The high deuterium fractiona-tion of our sample might then result from the accumulation of[NH D] / [NH ] ratios of individual clumps within the beam. NH deuteration as high as our observed values is obtained by themodel of Sipil¨a et al. (2019) which limits proton-donation reac-tions to proceed only through proton hop (cf. Hily-Blant et al.2018). This model predicts a [NH D] / [NH ] ratio exceeding 1for a density of 10 cm − after one phase that lasts ∼ years.A comparison of the [NH D] / [NH ] ratio of the ATLASGALsources with that of low-mass star-forming samples, for exam-ple 0.02 - 0.1 by Tin´e et al. (2000), 0.025 - 0.18 by Saito et al.(2000), and 10 − - 10 − by Shah & Wootten (2001), indicatesat least similar deuteration in high-mass star-forming regions.Kau ff mann et al. (2010) found that at a certain radius cluster-forming clouds have more mass and therefore a higher den-sity than their counterparts without cluster formation. As thetimescale of deuteration has been found to be shorter with in-creasing density (K¨ortgen et al. 2017), we expect an enhanceddeuterium fractionation of ATLASGAL sources resulting fromthe higher density of this high-mass star-forming sample. trace the evolutionof ATLASGALclumps? Deuteration as an evolutionary tracer of low- and high-massstar formation:
Busquet et al. (2010a) detected NH D emissionin a few starless cores, while it is not associated with YSOsin a high-mass star-forming region. The deuterium fraction ofNH therefore allowed them to distinguish between the pre- protostellar and protostellar phase. Fontani et al. (2011) obtaineddi ff erences in the deuteration of about 30 cores at various evo-lutionary phases of high-mass star formation from the deriva-tion of the [N D + ] / [N H + ] ratio. These latter authors found adecrease in the deuterium fraction from high-mass cores with-out stars to the evolutionary stages after the formation of proto-stars. In addition, they obtained a slight decrease of the fraction-ation with temperature and N H + line width. Kang et al. (2015)also found the narrowest line widths for the youngest Class 0protostars in Orion with the largest amount of deuteration inH CO, which hints at early stages of star formation. However,they did not find a clear correlation of the [HDCO] / [H CO] ra-tio and the mass-to-luminosity ratio as a tracer of the evolu-tionary phase. Shah & Wootten (2001) claimed to have founda trend of increasing NH deuteration with decreasing tempera-ture. However, the sample of these latter authors contains onlya few low-mass cores, which introduces a large statistical error.Emprechtinger et al. (2009) determined the [N D + ] / N H + ] ratiofor 20 protostellar cores in low-mass star-forming regions, whichalso yields an anticorrelation of deuteration with temperature.While previous studies of deuteration focused on low-massstar-forming samples, molecules di ff erent from NH D, such asN D + , and small source samples, in this section we examinewhether or not the deuterium fraction of NH is an indicatorof the evolution in a large sample of high-mass star-formingregions. Because we obtained a statistically significant correla-tion between the NH (1,1) line width and rotational temper-ature of ATLASGAL sources with the evolutionary phase inWienen et al. (2012), we use these properties to investigate a de-pendence of the NH deuteration on the evolution. NH rotational temperature: The [NH D] / [NH ] ratio isplotted against the NH rotational temperature between the (1,1)and (2,2) inversion transitions in the upper panel of Fig. 6. Therotational temperature of the ATLASGAL sources detected inNH D ranges between 10 and 24 K and NH column densitiesbetween 1 . × and 10 cm − (see Wienen et al. (2012)).One expects the largest deuterium fraction of NH at temper-atures lower than 20 K resulting from collisions between H + and HD producing H D + and increasing the [H D + ] / [H + ] ra-tio (Flower et al. 2004; Roberts & Millar 2000a). H D + is animportant molecule in deuterium chemistry in cold clouds andenhances the deuteration of several other molecules. However,
11. Wienen et al.: ATLASGAL - NH deuteration Table 7.
Comparison of the NH deuteration from ATLASGAL with other samples. Sample selection Sample size detection rate in NH D (%) [NH D] / [NH ] ReferenceHigh-mass ATLASGAL clumps 992 39 0.007 - 1.6 this articleUCHIIR starless cores 7 100 < .
06 - 0.8 Busquet et al. (2010a)Pre / protocluster clumps 32 69 0.004 - 0.67 Pillai et al. (2007)Pre-stellar core L 1544 1 100 0.5 Crapsi et al. (2007) ∗ Perseus protostellar cores 7 100 0.17 - 0.33 Hatchell (2003)Dense cores 2 100 0.02, 0.1 Tin´e et al. (2000)Low-mass protostellar cores 32 70 0.003 - 0.13 Shah & Wootten (2001) ∗ Dark molecular cloud cores 16 50 0.025 -0.18 Saito et al. (2000)
Notes. ( ∗ ) Samples using non-LTE models for the determination of the NH D column density. H D + is destroyed by neutral molecules such as CO at temper-atures above ∼
25 K and by ortho-H at temperatures belowthat as well. Freeze-out of CO onto dust grains increases theabundance of deuterated molecules at low temperatures and highdensities. However, the deuteration of the ATLASGAL sampledoes not show an anticorrelation with temperature. This agreeswith previous results from Pillai et al. (2011) and Fontani et al.(2015). A small sample of IRDCs show no trend between[NH D] / [NH ] ratio and temperature (Pillai et al. 2011) and theNH deuteration does not rise with decreasing temperature fortheir sample of about 30 dense cores in di ff erent evolutionaryphases of high-mass star formation (Fontani et al. 2015).Production of NH D begins during the early pre-protostellarphase. During the evolution of the ATLASGAL sample, an in-ternal heating source forms in the innermost part of the clumpand leads to a higher rotational temperature, while NH D is con-stantly accumulated in the still cold outer envelope. We specu-late that this process might have the consequence of a constantdeuteration for a rising rotational temperature. Alternatively, theluminosity of forming protostars within the clumps, for exam-ple G31.41 + + deuteration on temperature. NH line width: We compared the fractionation ratio withthe NH line width as another evolutionary tracer in the mid-dle panel of Fig. 6. The line widths range between 0.8 and 6.4km s − ; there is no anticorrelation between deuteration and linewidth. This is in agreement with the deuteration of dense cores inmassive star-forming regions (Fontani et al. 2015), which doesnot depend on the line width either. MSX 21 µ m flux: ATLASGAL sources without any star for-mation yet are not or only weakly detected at mid-infrared (MIR)or far-infrared (FIR) wavelengths, while clumps in a later evolu-tionary phase associated with a heating source emit at 21 µ m. Wetherefore also compare the [NH D] / [NH ] ratio with the MSX21 µ m flux to investigate any trend of the NH deuteration withevolution. However, the lower panel of Fig. 6 shows a flat distri-bution of the MSX flux with the fractionation ratio.In summary, Fig. 6 illustrates that high-mass star-formingregions might be too complex to show a trend of decreasingNH deuteration with increasing rotational temperature or linewidth. At the large distances of the ATLASGAL sample witha median distance of 4 kpc (Wienen et al. 2015) a clump likelyharbours several cores at di ff erent evolutionary stages. The pres-ence of multiple evolutionary phases within one source was alsofound by Urquhart et al. (2014). While observations of nearby low-mass star-forming samples have a much higher spatial res-olution and therefore focus on individual cores, the temperatureand line width of an ATLASGAL clump results from an aver-age of these properties over the beam width. An ATLASGALsource might therefore host cores with a large amount of NH deuteration, low temperatures, and narrow line widths as well aswarm, turbulent cores with broad line widths. However, averag-ing over the beam width results then in a high deuterium fractionof NH at relatively high temperatures and broad line widths, andan overall constant distribution of the [NH D] / [NH ] ratio overthe range of NH line parameter values. D detections andnon-detections
We divide the ATLASGAL sources observed in NH D into twosubsamples: one that shows NH D emission, and another forwhich NH D is not detected. Figure 8 shows histograms of theNH (1,1) line width, NH column density, and rotational tem-perature for non-detections in black and for detections in red.There is no di ff erence in the line width, column density, or tem-perature between the two subsamples; they have a peak at ∼ . − , ∼ × cm − , and 17 K respectively. The lower panelhas a smaller number of sources detected in NH D than non-detections at rotational temperatures higher than 22 K. BecauseHII regions have an enhancement in the rotational temperaturedistribution around 20 K (see Wienen et al. (2012)), the lowerpanel of Fig. 8 suggests that the NH D detection rate is low inHII regions. deuterationat differentevolutionaryphases To distinguish various evolutionary stages of ATLASGALsources detected in NH D we followed the classification intro-duced in K¨onig et al. (2017) for the TOP100 sample which cov-ers the whole evolutionary sequence. Based on the brightnessof these sources at infrared wavelengths and their radio contin-uum flux, four classes are separated. Adapting this classifica-tion, Urquhart et al. (2018) identified the evolutionary stage ofthe majority of ATLASGAL sources, which we used for associ-ation with our NH D detections. This results in the following: – an ATLASGAL sample of 19 clumps that are µ m weak :These sources have no pointlike or only a weak counterpartin the Hi-GAL data at 70 µ m. This sample is supposed torepresent a starless phase or about the earliest stage of high-mass star formation. – MIR-weak sources: This sample shows compact 70 µ memission, but is not detected at MIR wavelengths or emitsonly a weak 24 µ m flux below the limit of 2.6 Jy that corre-sponds to an 8 M ⊙ star at 4 kpc.
12. Wienen et al.: ATLASGAL - NH deuteration Fig. 8.
Distribution of ATLASGAL sources with the NH (1,1)line width, rotational temperature, and NH column density areplotted for NH D detections in red and non-detections in black. – MIR-bright clumps: These are identified by their brightcompact emission at 8 and 24 µ m. These sources show signsof star formation activity such as infall or outflows. – compact HII regions : These objects are characterized bya strong MIR and radio continuum flux. They are the lat-est evolutionary stage, where high-mass protostars emit ul-traviolet radiation, and thus heat and ionise their remainingmolecular cloud forming compact HII regions. We summarise the identification of the various evolutionaryphases of the ATLASGAL sample detected in NH D in Table8. This gives the fraction of sources in each evolutionary stagefor the whole sample and shows that most NH D detectionsare MIR bright, that the number of MIR-weak sources andcompact HII regions are similar, and that the lowest numberof NH D detections are 70 µ m-weak clumps. Comparison ofthe [NH D] / [NH ] ratio with the NH (1,1) line width androtational temperature (see Fig. A.1) shows that the 70 µ m-weaksample exhibits the narrowest line widths ( < − ) andsmallest rotational temperatures ( < − and mean rotational temperature of17.3 K. Large line widths up to 6.4 km s − and high rotationaltemperatures up to 23 K are found for the compact HII regions.However, the subsamples in the various evolutionary phases donot di ff er with regard to NH deuteration and the [NH D] / [NH ]ratios do not show any trend with the evolutionary sequence.We performed a Kolmogorov-Smirnov (KS) test with thesubsamples at the di ff erent evolutionary stages to analysewhether or not they di ff er significantly in NH deuteration. Thedistributions of the 70 µ m-weak sources, the MIR-weak sources,and compact HII regions do not contradict the idea that they aredrawn from the same parent population. The cumulative distri-bution plot in Fig. 9 yields a higher NH deuteration for clumpswith weak or bright MIR emission. The NH D / NH ratio islow for the 70 µ m-weak phase, rises for the MIR-weak / MIR-bright sources, and decreases again for the compact HII re-gions. Compared to low-mass star formation, where deuter-ation decreases with rising temperature (Crapsi et al. 2005;Emprechtinger et al. 2009), we find a shift of the maximumdeuterated NH of ATLASGAL sources to a later evolution-ary phase. The initial conditions to form NH D and to increasethe deuterium fractionation is sensitive to (low) temperature and(high) density. While the 70 µ m-weak clump might be cold, itis unclear that it is su ffi ciently dense. There is evidence fromsome recent observations that high-mass 70 µ m-weak clumpsare likely to contain modest sub-structure in dense cores, andthat the clump mass reservoir is dominated by low density mate-rial (Pillai et al. 2019). During the MIR-weak phase, the densityis likely to be high enough while the overall temperatures re-main low enough to result in deuterium enhancement. Duringthe MIR-bright phase, the NH D / NH ratio is likely to reachthe peak. Protostellar heating raises the temperature, but a ff ectsdeuteration only in the immediate vicinity of the core so that thedecrease of the deuterium fractionation might be undetectable insingle-dish observations. This might be di ff erent when zoominginto these cores at a resolution of less than a few thousand AU,where the direct heating from the protostar is e ffi cient and wouldreveal the e ff ect of deuteration (Pillai et al. 2011). The continuedstar formation during the compact HII region phase is expectedto result in a decrease in deuteration that makes it undetectable.The exact nature of deuterium enhancement with evolution canonly be constrained by high-resolution observations in a singlestar-forming region that hosts cores at all these various evolu-tionary stages. D In addition to the calculation of the ortho-to-para ratio given inSect. 3.6, we also determined the velocity integrated intensityratio, r , of the ortho and para NH D transition to compare withother studies. The integrated intensity of the ortho line, para line,
13. Wienen et al.: ATLASGAL - NH deuteration Table 8.
Number and fraction of ATLASGAL sources detected in NH D at di ff erent evolutionary stages of high-mass star formation. Sample Number Fraction (%)70 µ m-weak 19 7MIR-weak 32 12MIR-bright 152 58Compact HII regions 37 14 Table 9.
Velocity integrated intensity derived from the NH D line at 86 GHz, from the line at 110 GHz, and the ratio of bothintegrated intensities. Errors are given in parentheses. The full table is available at CDS. R T mb dv (86GHz) R T mb dv (110GHz) R T mb dv (86GHz) / R T mb dv (110GHz)Name (K km s − ) (K km s − ) (K km s − )G10.21-0.30 1.1( ± .
1) 0.2( ± .
1) 4.8( ± . ± .
1) 0.5( ± .
1) 2.3( ± . + ± .
2) 0.9( ± .
1) 2.8( ± . ± .
1) 1.0( ± .
1) 2.2( ± . ± .
2) 0.6( ± .
2) 3.7( ± . ± .
2) 1.0( ± .
2) 2.5( ± . ± .
2) 0.7( ± .
2) 2.4( ± . ± .
1) 0.9( ± .
1) 2.8( ± . ± .
2) 1.1( ± .
2) 1.2( ± . ± .
1) 0.8( ± .
1) 3.1( ± . + ± .
2) 0.6( ± .
2) 3.0( ± . + ± .
2) 0.9( ± .
2) 2.3( ± . ± .
1) 1.9( ± .
2) 2.5( ± . ± .
1) 1.0( ± .
1) 1.8( ± . Fig. 9.
Cumulative distribution functions display [NH D] / [NH ]ratios for the subsamples in Table 8. The distribution of 70 µ m-weak sources is marked as a solid red line, the MIR-weak clumpsare shown as a dashed blue curve, the MIR-bright phase as a dot-ted green line, and the compact HII regions as a dashed-dottedblack curve. A deuterium fraction of NH > deuteration is given by the di ff erencebetween the NH and NH D excitation temperature.and the ratio of the two is reported in Table 9. This resulted ina mean r of 2 . ± . r of 2 . ± . ff erent evolutionary phases, which is in agreement with ourresult. In addition, our values are similar to the integrated lineintensity ratios derived by Pillai et al. (2007) for clumps embed-ded in infrared-dark clouds. Shah & Wootten (2001) give a rangein N tot (86GHz) / N tot (110GHz) for low-mass protostellar cores of between approximately 2 and 6, which is in agreement with ourvalues for r .We investigate whether or not there is a correlation of theortho-to-para ratio computed in Sect. 3.6 with NH deuteration,NH (1,1) line width, and rotational temperature in Fig. 10 and11. ATLASGAL sources, whose hyperfine structure is not de-tected in ortho or para NH D, are indicated as black points.There are only two clumps with detected hyperfine structure inNH D at 86 and 110 GHz, which are shown as red triangles.However, Fig. 10 and 11 do not show a correlation between theortho-to-para ratio and any of the NH parameters. This yieldstherefore no evolutionary trend of the ortho-to-para ratio. Fig. 10. NH deuteration is compared with the ortho-to-para ra-tio. ATLASGAL sources without hyperfine structure in NH Dat 86 or 110 GHz are illustrated as black points and clumps withdetected hyperfine structure in both lines as red triangles.
14. Wienen et al.: ATLASGAL - NH deuteration Fig. 11.
Ortho-to-para ratio against NH (1,1) line width (up-per panel) and against rotational temperature (lower panel).ATLASGAL clumps, which have no detected hyperfine struc-ture in NH D at 86 or 110 GHz are displayed as black pointsand sources with detected hyperfine structure in both transitionsas red triangles.
5. Conclusions
Using the Mopra telescope and the IRAM 30m telescope, the 86GHz NH D lines were observed toward 992 dust condensationsidentified in the ATLASGAL survey. In the first quadrant of theGalaxy, 373 sources were also observed in the 110 GHz paraline with the IRAM 30m telescope. They are located within aGalactic longitude from 8 ◦ to 60 ◦ and between 300 ◦ and 359 ◦ and latitude of ± . ◦ . We summarise our main results in thissection.1. The detection rate of NH D towards the ATLASGAL sampleis high and therefore yields a large NH deuteration of thesesources.2. We calculate the total NH D column density to determinethe NH D-to-NH column density ratio. This results in alarge range of NH deuteration from 0.007 to 1.6. The deu-terium fraction of NH in ATLASGAL clumps is higher thanthe [NH D] / [NH ] ratios derived for low-mass star-formingsamples and in agreement with those obtained in other high-mass star-forming regions. We measure the highest NH deuteration reported in the literature so far. 3. The excitation of NH D was studied using the transitions at74 GHz and 110 GHz for the first time to our knowledge.This shows a clear di ff erence between the NH D and NH rotational temperatures for a subsample of the sources. Incases where NH D temperatures are lower than NH tem-peratures, deuteration would be overestimated, suggestingnon-LTE conditions. To determine the NH D temperature di-rectly, the NH D lines at 74 and 110 GHz should be observedsimultaneously.4. Comparison of NH D detections and non-detections sug-gests that the fraction of sources detected in NH D is higherfor the earlier evolutionary phases.5. We analyse whether or not there is a trend of NH deutera-tion with evolutionary tracers. While the [NH D] / [NH ] ra-tio is expected to decrease with rising rotational tempera-ture (Roberts & Millar 2000b; Sipil¨a et al. 2015a), we ob-tain a flat distribution of the deuterium fractionation with theNH (1,1) line width, rotational temperature, and the MSX21 µ m flux. Observations of ATLASGAL clumps with a high[NH D] / [NH ] ratio within a large beam width might alsoinclude cores of large line widths and temperatures. An av-erage over the beam width would lead to an increase of theseNH properties with an enhanced NH deuteration. Futureinterferometric follow-up observations could resolve this is-sue.6. We divide the ATLASGAL sample into di ff erent evolution-ary phases, but do not find any correlation between these andNH deuteration. The NH D / NH ratio is maximum duringthe MIR bright phase and is therefore reached at a later evo-lutionary stage compared to low-mass star formation.7. We estimate the ratio of the ortho-to-para NH D columndensity ratio. This results in a median ortho-to-para ratio of3.7 close to the expected value of 3. The ortho-to-para col-umn density ratios are in agreement with those of other low-and high-mass star-forming samples.
Acknowledgements.
M. Wienen acknowledges funding from theEuropean Union’s Horizon 2020 research and innovation pro-gramme under the Marie Skłodowska-Curie grant agreement No796461.This paper is dedicated to the memory of MalcolmWalmsley, who passed away before this study could be com-pleted. The present work benefited greatly from his insight andextensive advice. We are grateful for numerous inspiring discus-sions with him about various aspects of ammonia and its deuter-ation.
References
Bergin, E. A. & Tafalla, M. 2007, ARA&A, 45, 339Busquet, G., Palau, A., Estalella, R., et al. 2010a, A&A, 517, L6Busquet, G., Palau, A., Estalella, R., et al. 2010b, A&A, 517, L6Caselli, P. 2002, Planet. Space Sci., 50, 1133Caselli, P., Walmsley, C. M., Tafalla, M., Dore, L., & Myers, P. C. 1999, ApJ,523, L165Crapsi, A., Caselli, P., Walmsley, C. M., et al. 2005, ApJ, 619, 379Crapsi, A., Caselli, P., Walmsley, M. C., & Tafalla, M. 2007, A&A, 470, 221Csengeri, T., Leurini, S., Wyrowski, F., et al. 2016, A&A, 586, A149Csengeri, T., Urquhart, J. S., Schuller, F., et al. 2014, A&A, 565, A75Danby, G., Flower, D. R., Valiron, P., Schilke, P., & Walmsley, C. M. 1988,MNRAS, 235, 229Daniel, F., Faure, A., Wiesenfeld, L., et al. 2014, MNRAS, 444, 2544Emprechtinger, M., Caselli, P., Volgenau, N. H., Stutzki, J., & Wiedner, M. C.2009, A&A, 493, 89Flower, D. R., Pineau des Forˆets, G., & Walmsley, C. M. 2004, A&A, 427, 887Flower, D. R., Pineau Des Forˆets, G., & Walmsley, C. M. 2005, A&A, 436, 933Fontani, F., Busquet, G., Palau, A., et al. 2015, A&A, 575, A87
15. Wienen et al.: ATLASGAL - NH deuteration Fontani, F., Palau, A., Caselli, P., et al. 2011, A&A, 529, L7Garay, G. & Lizano, S. 1999, PASP, 111, 1049Giannetti, A., Leurini, S., Wyrowski, F., et al. 2017, A&A, 603, A33Giannetti, A., Wyrowski, F., Brand, J., et al. 2014, A&A, 570, A65Hatchell, J. 2003, A&A, 403, L25Hily-Blant, P., Faure, A., Rist, C., Pineau des Forˆets, G., & Flower, D. R. 2018,MNRAS, 477, 4454Kang, M., Choi, M., Stutz, A. M., & Tatematsu, K. 2015, ApJ, 814, 31Kau ff mann, J., Pillai, T., Shetty, R., Myers, P. C., & Goodman, A. A. 2010, ApJ,716, 433K¨onig, C., Urquhart, J. S., Csengeri, T., et al. 2017, A&A, 599, A139K¨ortgen, B., Bovino, S., Schleicher, D. R. G., Giannetti, A., & Banerjee, R. 2017,MNRAS, 469, 2602Kramer, C., Alves, J., Lada, C. J., et al. 1999, A&A, 342, 257Kroupa, P., Weidner, C., Pflamm-Altenburg, J., et al. 2013, The Stellar and Sub-Stellar Initial Mass Function of Simple and Composite Populations, ed. T. D.Oswalt & G. Gilmore, Vol. 5, 115Kurtz, S., Cesaroni, R., Churchwell, E., Hofner, P., & Walmsley, C. M. 2000,Protostars and Planets IV, 299Ladd, N., Purcell, C., Wong, T., & Robertson, S. 2005, PASA, 22, 62Mangum, J. G. & Shirley, Y. L. 2015, PASP, 127, 266Maret, S., Hily-Blant, P., Pety, J., Bardeau, S., & Reynier, E. 2011, A&A, 526,A47McKee, C. F. & Tan, J. C. 2002, Nature, 416, 59Millar, T. J. 2002, Planet. Space Sci., 50, 1189Oka, T. 2004, Journal of Molecular Spectroscopy, 228, 635Oliveira, C. M., H´ebrard, G., Howk, J. C., et al. 2003, ApJ, 587, 235Pagani, L., Salez, M., & Wannier, P. G. 1992, A&A, 258, 479Perault, M., Falgarone, E., & Puget, J. L. 1985, A&A, 152, 371Pillai, T., Kau ff mann, J., Wyrowski, F., et al. 2011, A&A, 530, A118Pillai, T., Kau ff mann, J., Zhang, Q., et al. 2019, A&A, 622, A54Pillai, T., Wyrowski, F., Hatchell, J., Gibb, A. G., & Thompson, M. A. 2007,A&A, 467, 207Price, S. D., Egan, M. P., Carey, S. J., Mizuno, D. R., & Kuchar, T. A. 2001, AJ,121, 2819Roberts, H. & Millar, T. J. 2000a, A&A, 364, 780Roberts, H. & Millar, T. J. 2000b, A&A, 361, 388Saito, S., Ozeki, H., Ohishi, M., & Yamamoto, S. 2000, ApJ, 535, 227Schuller, F., Menten, K. M., Contreras, Y., et al. 2009, A&A, 504, 415Shah, R. Y. & Wootten, A. 2001, ApJ, 554, 933Sipil¨a, O., Caselli, P., & Harju, J. 2015a, A&A, 578, A55Sipil¨a, O., Caselli, P., & Harju, J. 2019, A&A, 631, A63Sipil¨a, O., Harju, J., Caselli, P., & Schlemmer, S. 2015b, A&A, 581, A122Stutzki, J., Genzel, R., Graf, U. U., Harris, A. I., & Ja ff e, D. T. 1989, ApJ, 340,L37Stutzki, J. & Guesten, R. 1990, ApJ, 356, 513Tafalla, M., Myers, P. C., Caselli, P., Walmsley, C. M., & Comito, C. 2002, ApJ,569, 815Tin´e, S., Roue ff , E., Falgarone, E., Gerin, M., & Pineau des Forˆets, G. 2000,A&A, 356, 1039Urquhart, J. S., Figura, C., Wyrowski, F., et al. 2019, MNRAS, 484, 4444Urquhart, J. S., K¨onig, C., Giannetti, A., et al. 2018, MNRAS, 473, 1059Urquhart, J. S., Moore, T. J. T., Csengeri, T., et al. 2014, MNRAS, 443, 1555van der Tak, F. F. S., Black, J. H., Sch¨oier, F. L., Jansen, D. J., & van Dishoeck,E. F. 2007, A&A, 468, 627Wienen, M., Wyrowski, F., Menten, K. M., et al. 2015, A&A, 579, A91Wienen, M., Wyrowski, F., Menten, K. M., et al. 2018, A&A, 609, A125Wienen, M., Wyrowski, F., Schuller, F., et al. 2012, A&A, 544, A146Wu, L.-x. & Yang, J. 2005, Chinese Astron. Astrophys., 29, 370 Appendix A: NH deuteration and evolutionarytracers We associated our NH D detections with the ATLASGALsources classified in Urquhart et al. (2018). The [NH D] / [NH ]ratio of the di ff erent subsamples is plotted against the NH rota-tional temperature and (1,1) line width in Fig. A.1. Appendix B: Do NH D clumps with and withoutdetected hyperfine structure differ?
After not seeing any trend in the NH deuteration with the evolu-tion of our ATLASGAL subsample, we searched for di ff erences Fig. A.1. NH deuteration as a function of NH (1,1) line widthand rotational temperature between the (1,1) and (2,2) inversiontransition. ATLASGAL clumps that are 70 µ m weak are illus-trated as black points, MIR-weak sources as red points, MIR-bright clumps as green triangles, and compact HII regions asblue triangles.of ATLASGAL sources with or without detected hyperfine struc-ture in NH D. We cannot distinguish these two categories basedon the rotational temperature, NH line width, or MSX 21 µ mflux of the ATLASGAL clumps as indicated by Fig. 6. We com-pared the [NH D] / [NH ] ratio with the NH column density inFig. B.1 to examine whether or not the largest column densi-ties are related to the high deuteration of clumps with detectedhyperfine structure and the sources without detected hyperfinestructure and with low deuteration exhibit the lowest columndensities. However, we find no di ff erence in the column den-sity of ATLASGAL clumps with or without detected hyperfinestructure. Because Fig. B.1 might present a decreasing trend inthe deuterium fraction of NH of sources with detected hyperfinestructure and rising NH column density, we performed a t-testto examine whether or not the slope of the distribution is equal tozero. Because the t-test yields a p-value of 0.02, which is belowthe assumed significance level of 0.05, we can reject the hypoth-esis that the distribution can be fitted by a function with a slopeof zero. The fractionation ratio of the sources with detected hy-perfine structure will therefore become lower if the NH columndensity increases and can be fitted by [NH D] / [NH ] = -0.39log (N tot (NH )) +
16. Wienen et al.: ATLASGAL - NH deuteration Fig. B.1.