Synthetic observations of H 2 D + towards high-mass starless cores
Joaquin. Zamponi F., Dominik R. G. Schleicher, Stefano Bovino, Andrea Giannetti, Giovanni Sabatini, Simon Ferrada-Chamorro
BBAAA, Vol. 61A, 2019 Asociaci´on Argentina de Astronom´ıaR. Gamen, N. Padilla, C. Parisi, F. Iglesias & M. Sgr´o, eds. Bolet´ın de art´ıculos cient´ıficos
Synthetic observations of H D + towards high-mass starlesscores J. Zamponi , D. R. G. Schleicher , S. Bovino , A. Giannetti , G. Sabatini & S. Ferrada-Chamorro Departamento de Astronom´ıa, Universidad de Concepci´on, Esteban Iturra s/n Barrio universitario, Casillo160-C, Concepci´on, Chile INAF-Istituto di Radioastronomia - Italian ARC, Via P. Gobetti, 101, I-40129 Bologna, Italy Dipartimento di Fisica e Astronomia, Universit`a degli Studi di Bologna, via Gobetti 93/2, I-40129 Bologna,ItalyContact / [email protected]
Resumen / Las estrellas masivas j´ovenes residen com´unmente en c´umulos moleculares densos y masivos yson conocidas por estar altamente oscurecidas y distantes. Durante su proceso de formaci´on, la deuteraci´on esconsiderada como un buen indicador de la etapa de formaci´on en la que se encuentra un objeto. Observacionesadecuadas de regiones con emisi´on de deuterio son cruciales, aunque dif´ıciles de realizar. En este trabajo se hizoun survey para detectar la transici´on o-H D + (1 -1 ) en mol´eculas deuteradas, utilizando una fuente sint´etica eintentando declarar cu´an diferente es la informaci´on obtenida por un interfer´ometro o un telescopio de disco simple.Con objeto de analizar la detectabilidad de esta transici´on, procesamos la simulaci´on magneto-hidrodin´amica deuna nube colapsante haciendo uso del c´odigo de transferencia radiativa POLARIS. Utilizando los mapas deintensidad resultantes, realizamos observaciones sint´eticas de tipo interferom´etricas (ALMA) y de disco simple(APEX) de una nube en varios estados evolutivos, siempre usando modelos realistas. Finalmente, derivamosdensidades de columna para comparar nuestras simulaciones con observaciones anteriormente realizadas. Lasdensidades de columna obtenidas para el o-H D + concuerdan con valores reportados en la literatura, en el rangode 10 − cm − y 10 − cm − en mediciones interferom´etricas y de disco simple. Abstract / Young massive stars are usually found embedded in dense and massive molecular clumps and areknown for being highly obscured and distant. During their formation process, deuteration is regarded as a po-tentially good indicator of the formation stage. Therefore, proper observations of such deuterated molecules arecrucial, but still, hard to perform. In this work, we test the observability of the transition o-H D + (1 -1 ), usinga synthetic source, to understand how the physical characteristics are reflected in observations through interfer-ometers and single-dish telescopes. In order to perform such tests, we post-processed a magneto-hydrodynamicsimulation of a collapsing magnetized core using the radiative transfer code POLARIS. Using the resulting inten-sity distributions as input, we performed single-dish (APEX) and interferometric (ALMA) synthetic observationsat different evolutionary times, always mimicking realistic configurations. Finally, column densities were derivedto compare our simulations with real observations previously performed. Our derivations for o-H D + are inagreement with values reported in the literature, in the range of 10 − cm − and 10 − cm − for single-dish andinterferometric measurements, respectively. Keywords / ISM: molecules — stars: massive — stars: formation — radio lines: ISM
1. Introduction
Massive star formation takes place in the the densestpart of molecular clouds, which are characterized bylow gas temperatures ( T g < N g ∼ − cm − ) and a large degree of CO-depletion.In the earlier stages, due to the latter process, furtherchemical reactions become much more efficient, such asthe formation of deuterated molecules. This process iscalled deuterium fractionation ( D frac ) and explains theincreased ratio of a deuterated isotopologue column den-sity and its hydrogenated version. The reactions thatlead to deuteration are the following:H +3 + HD (cid:10) H D + + H + 230 K , H D + + HD (cid:10) D H + + H + 180 K , D H + + HD (cid:10) D +3 + H + 230 K , where H D + and D H + are the first two deuterated ionscreated. H D + , in particular, has been suggested to bea reliable chemical clock for star forming regions, due toits sensitivity to environmental conditions (e.g., Caselliet al. (2003); Br¨unken et al. (2014)).Observations of several high-mass clumps haveshown that deuteration increases over time as the col-lapse of the molecular clump proceeds, reaching a max-imum point right before the formation of a protostellarobject (Fontani et al., 2011). At this stage, the radia-tion from the stellar object will lead to an increase oftemperatures, with subsequent evaporation of CO, andan eventual decrease of the deuteration fraction.In this work we traced the abundance of o-H D + ,by performing synthetic observations towards simulatedmassive starless cores, and compare the results withavailable observations. Oral contribution 1 a r X i v : . [ a s t r o - ph . GA ] A p r ynthetic observations of high-mass starless cores FLASH
MHD SIMULATIONS
POLARIS
RTSIMULATIONS
Arrange data
CASA
COLUMN DENSITIES
ALMA simulation
APEX simulation
Python
Figure 1: Flowchart of the pipeline workflow.Table 1: Initial parameters of the core selected. The collapseis isothermal, at T=15 K.Run Radius Mass t ff α vir M (pc) (M (cid:12) ) (kyr)Hmu10M2 0.1 60 67 0.48 2 Fig. 1 shows a schematic view of the workflow fol-lowed here, from the acquisition of the source, to thefinal derived observatory-dependent column densities.
2. Synthetic source
As synthetic source we used the simulated collapsingmolecular cores from K¨ortgen et al. (2017).These cores are turbulent and magnetized ( B ∼ µ G), 60 M (cid:12) in mass, and 0.1 pc in radius. The mainparameters of the cores are listed in Table 1. A snap-shot of their gas surface density distribution after 32.1kyr of evolution is shown in the left panel of Fig. 2.
3. Radiative Transfer
We employed the
POLArized RadIation Simulator (PO-LARIS) radiative transfer code (Reissl et al., 2016) tosimulate the resulting intensity distributions of severalquantities included in the cloud simulation, such as,temperature, dust, gas and magnetic field distribution.POLARIS makes use of a Monte Carlo approach to tracethe path of light rays before reaching the synthetic de-tector. This method significantly reduces the numeri-cal noise as well as the runtime compared to previousmethods (Haworth et al., 2018). To map the deuter-ated molecule distributions, the Line Radiative Transfer(LRT) simulation mode of POLARIS was used for theo-H D + transition (1 -1 ), assuming Local Thermo-dynamical Equilibrium (LTE) conditions. We analyzed11 evolutionary stages of the core up to 1 t ff , placing itat a distance of 1.4 kpc, based on the distance to a sim-ilar source already observed (Pillai et al., 2012). Theresulting intensity distribution of a POLARIS simula-tion is shown in the second panel of Fig. 2.
4. Synthetic observations
To make observations as realistic as possible, we post-processed the ideal synthetic maps from POLARIS, inorder to take into account instrument-related effects.
For the single-dish case, new tasks were written to solveeach step involved in the creation of realistic cubes. TheSimulations presented here are referred to as APEX-likeobservations, as we do not attempt to make statementsof the
Atacama Pathfinder Experiment (APEX) realthroughput and performance. We developed a Pythonmodule that allows to perform common tasks on datacubes and single images, such as the convolution witha Point Spread Function (PSF), conversion between in-tensities, fluxes and brightness temperatures and theaddition of noise. We first convolved the ideal imageswith the beam of the telescope, represented by the PSF,here assumed to be Gaussian. The beam resolution was16.8” at 372.4 GHz (∆ v =0 .
03 km s − ). Then we addednormally-distributed noise to the images. In order toincrease the peak signal-to-noise ratio (SNR), the spec-tra were binned up to 0.5 km s − , resulting in a noiselevel ( T rms ) of 0.02 K. Such a setup returns detectionsat a 5 σ confidence level, for an integration of 6 hr. Thevalues were computed using the APEX Observing TimeCalculator . See the rightmost panel in Fig. 2 for a 16.8”single-pointing synthetic observation.
For the interferometric observations, we used the
Com-mon Astronomy Software Applications (CASA). CASAhas been designed for the
Atacama Large Milimiter-submiliter Array (ALMA) and the
Very Large Array (VLA) data analysis, and provides also a simulationmode. As with the single-dish data, we used the PO-LARIS outcome as the input model and then generatedthe measurement set by calling the simobserve task. simobserve simulates an actual observation by creat-ing a visibility set. We used a synthesized beam of 1”for o-H D + , reached by the Cycle 6 C43-1 array con-figuration in band 7. In this case, we set the ALMAspectral resolution one order of magnitude wider thanAPEX (0 . − ), in order to improve the sensitivity.The noise level was set at 6 mJy for an integration timeof 40 min, obtained by the ALMA Sensitivity Calcula-tor . We imaged the data by deconvolving the visibili-ties using the CLEAN algorithm. The cleaning step wasperformed by the simanalyze task. The outcome of anALMA simulation is shown in the third panel of Fig. 2.
5. Results & Conclusions
Column densities of o-H D + were obtained as in Vastelet al. (2006), via N ( X ) = 8 πν c Q ( T ex ) g u A ul e E u /T ex e hν/kT ex − (cid:90) τ dv, where u and l refer to the upper and lower energy levelof each transition, respectively. All of the parametersrequired here are retrieved from the Leiden Atomic andMolecular Database (LAMDA) (Sch¨oier et al., 2005), x (AU) × y ( AU ) × t = 27.1 kyr 10 -2 -1 S u rf ace d e n s it y ( g c m − ) h m s s h m s R.A. (J2000)20"15"10"05"-40°55"50"45" D e c . ( J ) POLARIS m J y p i x e l h m s s h m s R.A. (J2000)20"15"10"05"-40°55"50"45" D e c . ( J ) C43-1: 1" m J y b e a m h m s s h m s R.A. (J2000)20"15"10"05"-40°55"50"45" D e c . ( J ) APEX: 16.8" m J y b e a m Figure 2:
Left to right:
Gas surface density distribution (10 − -10 g cm − ) of the clump used as synthetic source at27.1 kyr; o-H D + emission after the radiative transfer simulation (0.02-0.14 mJy pixel − ); ALMA synthetic observation (5-45 mJy beam − ) of the clump at 1” resolution and APEX synthetic observation (50-350 mJy beam − ) at 16.8” resolution.All panels share field of view of 40kAU ∼ D + (1 -1 ) tran-sition. The statistical weight g ul is 9 and T ex
15 K.Transition ν (GHz) A ul (s − ) E ul (K) Q( T ex )o-H D + (1 -1 ) 372.421 1.08 · − Time (kyr)
Time (kyr) L o g ( N [ o - H D + ]) ( c m − ) S i n k f o r m a t i o n Model (770AU)ALMA (1" ↔
10 20 30 40 50 60
Time (kyr)
Model (23kAU)APEX (16.8" ↔ Figure 3:
Top panel:
Column densities derived from themodel (solid line) along with interferometric (green pen-tagons) and single-dish (blue hexagons) observations. Col-ored areas and horizontal dashed lines represent ranges ofvalues observed onto similar sources; central and bottom pan-els:
Comparison of the information loss between ALMA(middle) and APEX (bottom) by averaging the model inthe region of each beam. and summarized in Table 2. The partition function ( Q )was recreated by a cubic spline interpolation of the avail-able values in the Cologne Database for Molecular Spec-troscopy (cid:63) (CDMS). Under the LTE assumption, the ex-citation temperature ( T ex ) is assumed to match T gas .In this work we performed synthetic observations of acollapsing molecular magnetized core at different angu-lar resolutions, attempting to understand the main dif-ferences of observing with an interferometer or a single-dish. In the upper panel of Fig. 3 we show the columndensities derived from the model (solid black curve) at (cid:63)
235 AU (highest resolution element in the simulation),from ALMA at 0.55” and from APEX at 16.8”, respec-tively. Column densities at each time were averaged overthe highest resolution element. It can be observed thatretrieved column densities decrease by 10 (ALMA) and10 (APEX) from the model, because, when centeredon the peak emission, the inclusion of lower surround-ing values is more significant for wider beams, then, de-creasing the mean for low resolutions. Therefore, weexpect APEX estimations to be lower than those re-trieved by ALMA. Colored regions in Fig. 3 as well assemi-dashed and dotted lines represent observed columndensities reported in the literature, obtained via single-dish observations (APEX, JCMT and CSO) toward low-mass sources.As a measurement of the information loss, middleand bottom panels of Fig 3 present the ratio betweenboth resolutions and the model averaged over the areacorresponding to each beam. The highest ratios in inter-ferometric observations (middle panel) are up to 10 . .For single-dish observations (bottom panel) the highestratios are around 10 . . Such ratio decrease for widerbeams because the model curve smooths due to the in-clusion of surrounding lower values into the average. Acknowledgements:
JZ and DRGS thank for funding via Fonde-cyt regular 1161247. DRGS and SB also thank for fund-ing via Conicyt Programa de Astronomia Fondo Quimal 2017QUIMAL170001 and BASAL Centro de Astrof´ısica y Tecnolog´ıasAfines (CATA) PFB-06/2007. SB thanks for funding throughFondecyt Iniciacion 11170268. SB and JZ also thanks for fundingthrough the DFG priority program “The Physics of the InterstellarMedium” (project BO 4113/1-2). The simulations were performedwith resources provided by the
Kultrun Astronomy Hybrid Clus-ter at the Department of Astronomy, Universidad de Concepci´on.We also thank Robi Banerjee, Bastian K¨ortgen, Stefan Reissl andSebastian Wolf for important contributions to the presented study.
References
Br¨unken S., et al., 2014, Nature, 516, 219Caselli P., et al., 2003, A&A, 403, L37Fontani F., et al., 2011, A&A, 529, L7Haworth T. J., et al., 2018, NewAR, 82, 1K¨ortgen B., et al., 2017, MNRAS, 469, 2602Pillai T., et al., 2012, ApJ, 751, 135Reissl S., Wolf S., Brauer R., 2016, A&A, 593, A87Sch¨oier F., et al., 2005, A&A, 432, 369Vastel C., et al., 2006, ApJ, 645, 1198