Chemical study of intermediate-mass (IM) Class 0 protostars. CO depletion and N2H+ deuteration
T. Alonso-Albi, A. Fuente, N. Crimier, P. Caselli, C. Ceccarelli, D. Johnstone, P. Planesas, J. R. Rizzo, F. Wyrowski, M. Tafalla, B. Lefloch, S. Maret, C. Dominik
aa r X i v : . [ a s t r o - ph . S R ] J u l Astronomy&Astrophysicsmanuscript no. im˙v11 c (cid:13)
ESO 2018November 2, 2018
Chemical study of intermediate-mass (IM) Class 0 protostars.
CO depletion and N H + deuteration T. Alonso-Albi , A. Fuente , N. Crimier , P. Caselli , C. Ceccarelli , D. Johnstone , , P. Planesas , , J.R. Rizzo , F.Wyrowski , M. Tafalla , B. Lefloch , S. Maret , and C. Dominik Observatorio Astron´omico Nacional (OAN,IGN), Apdo 112, E-28803 Alcal´a de Henares, Spain e-mail: [email protected] Laboratoire d’Astrophysique Observatoire de Grenoble, BP 53, F-38041 Grenoble C´edex 9, France School of Physics & Astronomy, E.C. Stoner Building, The University of Leeds, Leeds LS2 9JT, UK Department of Physics & Astronomy, University of Victoria, Victoria, BC, V8P 1A1, Canada National Research Council of Canada, Herzberg Institute of Astrophysics, 5071 West Saanich Road, Victoria, BC, V9E 2E7, Canada Joint ALMA Observatory, El Golf 40, Las Condes, Santiago, Chile Centro de Astrobiolog´ıa (CSIC / INTA), Laboratory of Molecular Astrophysics, Ctra. Ajalvir km. 4, E-28850, Torrej´on de Ardoz,Spain Max-Planck-Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, D-53121 Bonn, Germany Anton Pannekoek Astronomical Institute, University of Amsterdam, P.O. Box 94249, 1090 GE Amsterdam, The NetherlandsReceived February 2, 2010; accepted April 4, 2010
ABSTRACT
Aims.
We are carrying out a physical and chemical study of the protostellar envelopes in a representative sample of IM Class 0protostars. In our first paper we determined the physical structure (density-temperature radial profiles) of the protostellar envelopes.Here, we study the CO depletion and N H + deuteration. Methods.
We observed the millimeter lines of C O, C O, N H + and N D + towards the protostars using the IRAM 30m telescope.Based on these observations, we derived the C O, N H + and N D + radial abundance profiles across their envelopes using a radiativetransfer code. In addition, we modeled the chemistry of the protostellar envelopes. Results.
All the C O 1 → O abundance decreases inwards within the protostellarenvelope until the gas and dust reach the CO evaporation temperature, ≈ H + deuterium fractionation in Class 0 IMs is [N D + ] / [N H + ] = / H] value in the interstellar medium, but a factor of 10 lower than in prestellar clumps. Chemical models account for the C O andN H + observations if we assume the CO abundance is a factor of ∼ OH on the grain surfaces prior to the evaporation and / or the photodissociationof CO by the stellar UV radiation. The deuterium fractionation is not fitted by chemical models. This discrepancy is very likely causedby the simplicity of our model that assumes spherical geometry and neglects important phenomena like the e ff ect of bipolar outflowsand UV radiation from the star. More important, the deuterium fractionation is dependent on the ortho-to-para H ratio, which is notlikely to reach the steady-state value in the dynamical time scales of these protostars. Key words.
Stars: formation – Stars: Emission-Line, Be – Stars: pre-main sequence – Stars: individual: Serpens-FIRS 1 – Stars:individual: Cep E-mm – Stars: individual: L1641 S3 MMS 1 – Stars: individual: IC1396N – Stars: individual: CB3 – Stars: individual:OMC2-FIR4 – Stars: individual: NGC 7129 FIRS 2 – Stars: individual: S140 – Stars: individual: LkHa 234
1. Introduction
Intermediate-mass young stellar objects (IMs) share many char-acteristics with high-mass stars (clustering, PDRs) but theirstudy presents an important advantage: many are located closerto the Sun (d ≤ α CO + depletion. Moreover, the deuterium fractiona-tion, measured as the DCO + / H CO + ratio, decreases by a factorof 4 from the Class 0 to the Herbig Be star, very likely owingto the increase in the kinetic temperature. Regarding the abun-dance of complex molecules, the beam-averaged abundances ofCH OH and H CO increase from the Class 0 to the Herbig Bestar. A hot core was also detected in NGC 7129–FIRS 2 (Fuenteet al. 2005a,b). Although two objects are not enough to establishfirm conclusions, these pioneering results suggest that chemistryis also a good indicator of the evolution of IMs.
T. Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars.
We are carrying out a chemical study of a representative sam-ple of IM Class 0 YSOs This is the first systematic chemicalstudy of IM Class 0 objects that has been carried out so far.Some properties like the temperature of the protostellar enve-lope and the clustering degree depend on the final stellar mass,so the results for low-mass stars cannot be directly extrapo-lated to intermediate-mass objects. In the first paper (Crimieret al. 2010, hereafter C10), we determined the physical struc-ture (density-temperature radial profiles) by modeling the dustcontinuum emission. We now present the observations of themillimeter lines of C O, C O, N H + , and N D + in the samesample. Our goal is to investigate the CO depletion and N H + deuteration in these Class 0 YSOs. For comparison, we also in-clude 2 Class I objects, LkH α
2. Observational strategy
Our selection was made to have a representative sample of Class0 IM YSOs, including targets with di ff erent luminosities (40 –10 L ⊙ ) and evolutionary stages. An important complication inthe study of massive stars is that they are located in complexregions and are therefore di ffi cult to model. The targets in thissample were chosen to lie preferentially in isolated areas withrespect to the 30m telescope beam. We also selected sources forwhich continuum maps at submillimeter and / or millimeter wave-lengths are available in order to be able to model their envelopes.The list of sources and their coordinates are shown in Table 1.To provide a comparison with Class I sources, we added S140and LkH α
234 to the sample.Most of the observations reported here were carried out withthe IRAM 30m telescope at Pico de Veleta (Spain) during threedi ff erent observing periods in June 2004 in position switchingmode. Our strategy was to first make long integration single-pointing observations towards the star position and then to make96 ′′ × ′′ maps around the center position. The maps were sam-pled with a spacing of 12 ′′ in the inner 48 ′′ regions and 24 ′′ outside. The only exceptions were L1641 S3 MMS1 and S140.In L1641 S3 MMS1, we only observed a radial strip in C Oand C O. We did not observe C O and C O maps in S140. Asummary of the observations for each source is shown in Table1, and the list of observed lines and the telescope characteris-tics is shown in Table 2. During the observations, lines of thesame species were observed simultaneously using the multire-ceiver capability of the 30m telescope. In this way, we mini-mized relative pointing and calibration errors. As backends weused in parallel an autocorrelator split into several parts provid-ing a spectral resolution that was always better than ∼
78 kHz anda 1 MHz-channel-filter-bank. Examples of the single-pointingobservations are shown in Figs. 1 and 2. The intensity scale isthe main brightness temperature.Observations of the N H + J = → = ′′ × ′′ with a spacing of 25 ′′ , but the emission wasonly detected towards the star position (see Fig. 3). All the lineswere observed with a spectral resolution of 0.488 MHz. Thespectra of the J = → O, N H + and N D + datatowards NGC 7129-FIRS 2 and LkH α Table 2.
Description of the observations
Line Freq. (MHz) HPBW η MB TelC O 1 → ′′ O 2 → ′′ O 1 → ′′ O 2 → ′′ H + → ′′ H + → ′′ D + → ′′ D + → ′′
3. Results
The spectra of the C O 1 →
0, C O 1 →
0, N H + → D + → O 1 →
0, N H + →
0, andN D + → H + → H + emission are o ff set from the far-IR source.For the Herbig Be star LkH α H + → ff set (-6”,18”) from the star position(see Fuente et al. 2005a and Fig. 5). In S140, the N H + emissionpeaks in an arc-shaped feature surrounding the sources IRS 1,2 and 3. In Class I sources, either the N H + is destroyed by theevaporated CO or the outflow, or the UV radiation from the starhas already disrupted the parent core.The emission in the C O 1 → H + → O line has an elongatedshape, much more extended than that of N H + (see Fig. 4).In OMC2 FIR 4 and NGC 7129-FIRS 2, the emission of theC O line surrounds the star position instead of having a max-imum towards it (see also Fuente et al. 2005a). Only in thelow-luminosity sources Serpens-FIRS 1 and Cep E-mm does theemission from the C O and C O lines peak at the star positionand present a morphology similar to that of the N H + emission.For the sources in which the signal-to-noise ratio of theN D + map is high enough, Serpens-FIRS 1 and NGC 7129FIRS 2, we compared the N D + → H + → D + → O 1 →
0, N H + →
0, and N D + → H + → O 1 → ff erfrom one source to the next. The presumably youngest sources,OMC2 FIR 4 and NGC 7129–FIRS 2, show a flat profile. This isthe expected picture when the CO abundance decreases towardsthe center, mainly because the molecules are frozen onto thegrain surfaces. In Serpens-FIRS 1, Cep E-mm, IC 1396 N, andCB3, the C O 1 → H + but with a less steep profile. Intenseemission of the C O 1 → O1 → ff erent radial profiles of the C O 1 → . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 3 Table 1.
Selected sample
Object RA(2000) Dec(2000) Lum. (L ⊙ ) dSerpens-FIRS 1 18:29:49.6 + O,C O,N H + and N D + Cep E-mm 23:03:13.1 + O,C O,N H + and N D + L1641 S3 MMS 1 05:39:55.9 -07:30:28.0 67 500 Strip in C O and C O, maps in N H + and N D + IC1396N 21:40:41.7 + O,C O,N H + and N D + CB3 00:28:42.7 + O,C O,N H + and N D + OMC2-FIR4 05:35:26.7 -05:10:00.5 1000 450 Maps in C O,C O,N H + and N D + NGC 7129–FIRS 2 21:43:01.7 + O, maps in C O,N H + and N D + S140 22:19:18. 1 +
910 N H + and N D + mapsLkH α
234 21:43:06.8 + O,N H + and N D + Fig. 1.
Examples of the single-pointing observations towards the Serpens–FIRS 1, Cep E–mm, L1641 S3 MMS1, and IC 1396 N.All the spectra were observed with the 30m telescope. The intensity scale is the main brightness temperature.
Fig. 2.
The same as Fig. 1 for CB3, OMC2 FIR 4, NGC 7129–FIRS 2, and S140.
T. Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars.
Fig. 3.
Spectra of the N H + → O emission towards the star positioncould come from CO depletion in the case of a very young objector to photodissociation for a borderline Class 0 / I. The radial pro-files of the N D + → → / N ratio of the maps.
4. LTE column densities
We have derived the C O column densities using the rotationdiagram method. This method gives an accurate estimate of thecolumn density provided that the emission is optically thin, isthermalized, and arises from a homogeneous and isothermalslab. In the case of a density and temperature distribution, thederived column density represents an average value over the ob-servational beam.The excitation of C O and C O are very similar, and wehave better signal-to-noise spectra for the more intense C Olines. For this reason, we derived the C O column density byassuming optically thin emission and the rotation temperaturederived from the C O data. The molecule N H + presents hy-perfine splitting. This allows us to estimate the line opacity di-rectly from the hyperfine line ratios. We derived the total N H + column density from the opacity of the N H + → D + data whenthere was no N H + → H + and N D + . We assumeda beam filling factor of 1 for all the lines regardless of the ob-servational beam size. This approach overestimates the rotationtemperature when the emission is centrally peaked. In Table 3 wepresent the derived C O, C O, N H + , and N D + column densi-ties. We also show the N(C O) / N(C O) (hereafter R ) and theN(N D + ) / N(N H + ) (hereafter R ) ratios. For the whole sample,we obtain a value for R around 3.3 +/ -0.3, the expected value foroptically thin emission. In the worst case, R = O column density because of the opacity is only a fac-tor ∼ ff ects are not important in our C Ocolumn density estimates. D + and N H + The deuterium fractionation of N H + (R ) strongly depends onthe CO depletion factor and the gas temperature (Caselli et al.2002, Ceccarelli & Dominik 2005, Daniel et al. 2007). In oursample of IM Class 0 protostars, we derive values of R rang-ing from 0.005 to 0.014. These values are 3 orders of magnitudehigher than 10 − , the elemental value in the interstellar medium(Oliveira et al. 2003). Nevertheless, the R values are a factorof 10 lower than those found in prestellar clumps by Crapsi etal. (2005). According to the values of R , we can classify oursources into two groups: (i) highly deuterated sources that havevalues of R > < D + ) / N(N H + ) ratio is agood gas temperature indicator, these protostars should be thecoldest and very likely the youngest of our sample. We have to becautious with CB3, however. Since this source is the most distant(d = / I YSO. The results are thus consistent with our interpretationof the objects in this group as being more evolved. In Serpens–FIRS 1, however, the N D + → ffi cult to classify. The deuterium fractionation suggestsan evolved object but the morphology of the C O map is moreconsistent with a young Class 0 star. The peculiarity of this pro-tostellar envelope has already been pointed out by Crimier et al.(2009). They find that the envelope of OMC2 FIR 4 is peculiarlyflat and warm with a radial density power-law index of 0.6.We present the N H + deuterium fractionation as a func-tion of the N(C O) / N(N H + ) ratio in Fig. 7. An excellentcorrelation between these two quantities is found in prestellarcores with the deuterium fractionation decreasing with increas-ing N(C O) / N(N H + ) ratio (see Crapsi et al. 2005). This cor-relation is not valid for IM Class 0 protostars. As discussed inthe following sections, this is due to the complexity of theseintermediate mass star-forming regions. In prestellar cores thechemistry of these species is only driven by the CO depletion.In IMs, other phenomena like photodissociation and shocks arealso playing important roles.
5. Radiative transfer model
We utilized a general ray-tracing radiative transfer code to de-rive the fractional abundance profiles of C O, N H + and N D + across the envelopes. Assuming appropriate radial profiles forthe temperature, density, molecular abundance, and turbulencevelocity, this model calculated the brightness temperature dis-tribution on the sky. The model map was then convolved withthe telescope beam profile. The underlying source geometry wasassumed to be a sphere, with the inner and outer radii, andtemperature-density (T-n) profiles derived by C10. The size ofthe grid was set to 32x32 cells. The cells have di ff erent sizesalong the line of sight to account for the di ff erent slopes in thedensity and temperature profiles. We used very small cells ( . The model is called DataCube and a link to install it is availableupon request.. Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 5
Fig. 4.
Continuum maps at 850 µ m from C10 (left column), and integrated intensity maps of the C O 1 →
0, N H + → D + → → ff erent marks.Contour levels starts at 3 σ level and are a. 0.5, 1.6, 2.7, 3.8, 4.9, and 5.4 Jy / beam; b. 1 to 9 K km s − by steps of 1 K km s − ; c. 3 to9 K km s − by steps of 3 K km s − ; d. 0.5 to 1.5 K km s − by steps of 0.5 K km s − ; e. 0.16, 0.49, 0.81, 1.14, and 1.46 Jy / beam; f.0.5 to 3 K km s − by steps of 0.5 K km s − ; g. 0.5 to 2.5 K km s − by steps of 0.5 K km s − ; h. 0.3, 0.9, 1.5, 2.1, and 2.8 Jy / beam;i. 2.0 to 8.0 K km s − by steps of 2.0 K km s − ; j. 3.0 to 9.0 K km s − by steps of 3.0 K km s − ; k. 0.07, 0.20, 0.33, 0.46, and 0.59Jy / beam; l. 2.0, 4.0 K km s − ; m. 1.0 to 3.0 K km s − by steps of 1.0 K km s − ; n. 1.0, 2.5, 4.0, 5.5, and 7.0 Jy / beam; o. 0.5 to3.5 K km s − by steps of 0.5 K km s − ; p. 4.0 to 24.0 K km s − by steps of 4.0 K km s − ; q. 2.1, 6.3, 10.6, 14.8, and 19.0 Jy / beam;r. 2.0 K km s − ; s. 3.0, 6.0 K km s − ; t. 0.5 K km s − . T. Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars.
Fig. 5.
Integrated intensity maps of the C O 1 → H + → α
234 and S140. Contour levels are: a. / beam; b. − by steps of 1 K km s − ; c. − in steps of1 K km s − ; d. / beam in steps of 1 Jy / beam ; e. − in steps of 3 K km s − . Table 3.
LTE column densities
Source T r N(C O) N(C O) R ∗ T r (K) N(N H + ) N(N D + ) R ∗ (K) (cm − ) (cm − ) (K) (cm − ) (cm − )Serpens-FIRS 1 11 8.2 × × a × × × × b × × × × b × < × < × × b × < × < × × a × × × × a × × × × c × × c × × d × < × < α b
22 1.0 × < × < ∗ R = N(C O) / N(C O), R = N(N D + ) / N(N H + ). a Rotation temperature derived from our N D + observations. b Rotation temperature derived from our N H + observations. c From Fuente et al. (2005a). d From Minchin et al. (1995).
AU in size. The turbulent velocity was assumed to be fixed at1.5 km s − , consistent with the linewidths of the observed lines.Finally, in each cell, the excitation temperature was calculatedwith the RADEX code (van der Tak et al. 2007), which uses theslab LVG approximation at each shell. We used the collisionalrates provided by the LAMDA database (Sch¨oier et al. 2005).For N D + , we used the same collisional rates as for N H + . LAMDA database is available athttp: // / ∼ moldata / . With these assumptions, we searched for the best fit on everysource and molecule observed. The fitting process consisted ofaveraging the observed and modeled fluxes in concentric ringsaround the center position using GILDAS software. The cen-ter position was selected to be the continuum emission peak at850 µ m. The radius of the first ring was set to HPBW /
4, and in-cremented by the same amount in consecutive rings. As a firststep we searched for the best fit using a constant abundance.Since this approach seldom produced good fits, we next searchedfor the best fit using radial power functions and step functions.These functions were not selected arbitrarily but were selected . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 7
Fig. 6.
Radially integrated intensity profiles of the C O 1 →
0, N H + →
0, and N D + → → / Table 4.
Temperature-density (T-n) profiles from Crimier et al. 2010
Source CB3 Cep- E IC 1396 N NGC 7129 Serpensmm mm BIMA2 FIRS 2 FIRS 1RA(J2000) 00:28:42.1 23:03:12.7 21:40:41.8 21:43:01.5 18:29:49.8Dec(J2000) 56:41:59.4 61:42:27.4 58:16:13.5 66:03:25.0 01:15:18.4Dust optical depth at 100 µ m, τ α r out / r i
400 500 630 180 200Inner envelope radius, r in , (AU) 260 70 50 100 30Outer envelope radius, r out , (AU) 103000 35800 29600 18600 5900Radius at T dust =
100 K, r , (AU) 700 223 180 373 102H density at r , n , (cm − ) 7.5 × × × × × Envelope mass, M env , (M ⊙ ) 120 35 90 50 5.0 to mimic the predictions of chemical models. The step func-tion accounts for the abrupt sublimation of the CO ices thanksto the increase in the dust temperature going inwards, whereasthe power-law profile accounts for a smoother change in theabundance of the species. The angular resolution of our observa-tions ( ∼ ∼ < a few 1000 AU) of the protostellar envelope.For this reason, we assume a constant abundance in the innerpart.We fit the integrated intensity maps of the C O 1 →
0, N H + →
0, N H + →
3, and N D + → O 2 → O 1 → → α
6. Limitations of our model: CO evaporationtemperature
The model used here is not a reliable predictor of the chemicaland physical properties within the inner envelope for several rea-sons. First of all we only considered the low-J rotational molec-
T. Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars.
Table 5.
Abundance profiles derived from the observations
Serpens-FIRS 1 X R (AU) X R (AU) X RMS (K km s − ) Diameter(”) / HPBW(”)C O 1.4 × − < × − × − H + × − < × − × − D + < × − × − ∗ ev (K) 20n ∗ ev (cm − ) 2.4 × Cep E-mm X R (AU) X R (AU) X RMS (K km s − )C O 6.0 × − × − × (r / H + × − × − × (r / D + × − ev (K) 20n ev (cm − ) 1.0 × IC 1396 N X R (AU) X R (AU) X RMS (K km s − )C O 6.0 × − × − × (r / H + × − × − × − D + < × − ev (K) 19n ev (cm − ) 7.0 × CB3 X R (AU) X R (AU) X RMS (K km s − )C O 1.3 × − < × − × − H + × − D + × − ev (K) 15n ev (cm − ) 3.0 × NGC 7129–FIRS 2 X R (AU) X R (AU) X RMS (K km s − )C O 4.0 × − H + × − D + < × − × − < × − ∗ Temperature and density at the CO evaporation radius. ular lines. The emission of these transitions arises mainly fromthe outer envelope. This fact, together with the limited spatialresolution of our single dish observations, prevent us from de-termining the variation in the abundance in the inner envelope.Additionally, in our modeling we use the n-T profiles derivedby C10. These profiles are a reasonable approximation for theouter envelope but are relatively unconstrained in the inner re-gion. Some parameters, like the inner radius of the envelope andthe dust temperature at this radius, are thus poorly determined.Another important source of uncertainty is the degeneracyof the solutions. For instance, in the case of C O we can obtainsimilar results by varying the molecular abundance in the innerregion (X ) or the CO evaporation radius (R ), as long as theline is optically thin and the number of molecules in the beamremains constant. Here we discuss the impact of this degeneracyin the derived CO evaporation radius.We ran a grid of models for Cep E-mm, IC 1396 N, andCB3. These are the sources in which the ratio between the angu-lar diameter of the envelope and the HPBW of the telescope isgreatest, so we are more sensitive to the spatial variations of thechemical abundances. In all these models, we assume that theC O abundance profile is X = X for radii less than the evapo-ration radius, R , and X = X × (R / R out ) α for radii larger than R .For each value of X , we fit R and α . X is the un-depletedC O abundance, R the evaporation radius, and α measures thegradient in the C O abundance due to depletion. The canonicalabundance of C O in principle depends on the Galactocentricdistance (DGC). Following Wilson & Matteucci (1992), the COabundance is given by X ( CO ) = . × exp (1 . − (0 . × DGC ( kpc ))) . (1) Following Wilson & Rood (1994), the oxygen isotope ra-tio O / O depends on DGC according to the relationship O / O = × DGC (kpc) + C O = × − . This is an average value, sothere may be local e ff ects. In our models we varied X from1 × − to 4 × − , the range of X values that we considerreasonable. For each value of X , we varied R and α to findthe best fit. In Fig. 8 we show the results of our models forCep E. The best fit is X = × − , R = α = = − . We consider that the models with RMSvalues departing by a factor of less than 1.3 from the mini-mum value are still acceptable. Following this criterion, we havetwo di ff erent families of solutions. Assuming X = × − , thebest fit is obtained with α ∼ ∼ = × − , we have reasonably good solutions with α ∼ ∼ . The evaporation radius of C O is thus2600 ±
600 AU. These radii correspond to dust temperatures of20-25 K. Interestingly, we would have a better fit to the obser-vations (RMS = − ) if we allowed the C O abun-dance in the inner region to fall below 1 × − . This is the solu-tion we show in Table 5. We do not reject this solution becausethere are several physical reasons that the CO abundance couldbe lower in the inner envelope. First of all, C O could havebeen photodissociated in the regions very close to the recentlyformed star. Alternatively, part of the CO could have been trans-formed into CH OH on the icy mantles, and the CO abundanceonce evaporated would then be di ff erent from the initial value.However, given the large uncertainty in the physical structure ofthe inner parts of the envelope, we cannot draw any conclusion. . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 9 In Fig. 9 we show the results for IC 1396 N. Thebest fit is X = × − , R = α = = − . We consider that the models with values ofRMS < = × − and R < >
23 K. Again, the fit can be sig-nificantly improved if values of X lower than 1 × − are con-sidered. This is similar to the case of Cep E. Finally, in Fig. 10,we show the same grid of models for CB3. The best fit is for X = × − , R = α = = − ,which corresponds to a CO evaporation temperature of ∼
30 K.However, this is not a good fit. In fact, we obtain a much bettersolution when assuming a step function and T ev ∼
15 K (see Table5). For Serpens-FIRS 1 and NGC 7129 FIRS2, the protostarsare barely resolved by our single-dish observations. In addition,the contribution of the surrounding cloud is very significant. Forthese reasons, we did not carry out a study similar to what isdescribed above. We looked only for abundance profiles that fitour observations and are consistent with the behavior predictedby chemical models (see Table 5). In these two cases, we canfit the observations when assuming that the CO is evaporated attemperatures around 20–25 K. We conclude, therefore, that ourC O observations towards the Class 0 stars are better fit assum-ing CO evaporation temperatures of 20–25 K that are consistentwith the CO being bound in a CO-CO matrix.The fits to the N H + and N D + emission profiles have beendone by hand. For N H + we tried two options:(i) constant N H + abundance and (ii) an abundance profile with a radial variationsimilar to that of C O. Option (ii) follows the expectation thatthe N H + abundance is strongly dependent on the CO abun-dance. In the case of CB3 and NGC 7129 FIRS 2, we are ableto fit the N H + observations by assuming a constant abundance.For the rest of sources, the N H + observations are better fit byassuming that the N H + is depleted in the cold regions similarlyto CO.One expects an annular distribution for the N D + abundance,with the maximum N D + abundance in the region with the great-est CO depletion. The morphology observed in the integratedintensity map of the N D + → D + emission is dominated by the foregroundmolecular cloud. For this reason we selected a one-step func-tion with a constant abundance inside and outside the protostel-lar core. In the case of NGC 7129–FIRS 2, we detected a clumpof N D + → D + emission assuming an annular abundance distribution.
7. Chemical model
The chemical composition was modeled with the simple chem-ical code originally described in Caselli et al. (2002), and up-dated following Caselli et al. (2008) with new measurements ofthe CO and N binding energies (Collings et al. 2003; ¨Oberg etal. 2005) and sticking coe ffi cients (Bisschop et al. 2006), ther-mal desorption, and the detailed physical structure. The cloudsare assumed to be spherically symmetric, with the density andtemperature profiles derived from the dust continuum emission.An interpolation procedure has been included in the code inorder to have smooth profiles. The chemical network containsthe neutral species CO and N , which can freeze out onto dustgrains and desorb owing to cosmic ray impulsive heating (asin Hasegawa & Herbst 1993) and by thermal evaporation (fol- Fig. 7.
Deuterium fractionation of N H + (R ) vs theN(C O) / N(N H + ) ratio in the studied YSOs. Inverted emptytriangles are upper limits for L1641 S3 MMS1, IC 1396 N, andS140, where we did not detect the N D + → Fig. 8.
Plots of the RMS, defined as Σ ( I model − I obs ) ) for the gridof models run to reproduce the C O emission in Cep E-mm.The C O abundance is assumed to be X = X for radii less thanR , and X = X × (R / R out ) α for radii larger than R . The values ofX are 1.0 × − (panel a.), 2.0 × − (panel b.), 3.0 × − (panelc.), and 4.0 × − (panel d.). The solution with the lowest RMS,0.37, is indicated by a cross. We consider acceptable solutionsthose with RMS less than 0.5. Contour levels are 0.5, 1, 2, and 5K km s − .lowing Hasegawa et al. 1992). The initial abundances of COand N have been fixed to 9.5 × − (Frerking et al. 1982) and2 × − , respectively. The abundances of molecular and atomicnitrogen are di ffi cult to determine in dense cores, but recentworks suggest low values in low-mass, star-forming regions:X(N) = n(N) / n(H ) . × − (Hily-Blant et al. 2010), andX(N ) ∼ − (Maret et al. 2006). Our adopted value is consistentwith 25% of the total abundance of N (n(N) / n(H) = × − ,see Anders & Grevesse 1989) locked in N and with the pseudo-time dependent models of Lee et al. (1996), after the chem-istry reaches steady state. Although atomic oxygen can a ff ectthe amount of deuterium fractionation (see discussion in Caselli Table 6.
Model results
Source Model CO Binding energy χ C O χ N H + χ N D + Serpens–FIRS 1 Model 3 1100 4.37 3.89 4.52 only 10% of CO evaporated
Model 4 1100 2.55 3.77 4.46 only 10% survives for T >
100 K
Cep E-mm Model 1 1100 1.48 4.12 0.71Model 2 5000 1.51 0.86 0.45
Model 3 1100 0.91 2.26 1.19 only 10% of CO evaporated
Model 4 1100 1.42 2.19 1.12 only 10% survives for T >
100 KIC 1396 N Model 3 1100 2.95 4.45 > Model 4 1100 1.79 4.35 > >
100 K
CB3
Model 3 1100 1.63 0.76 0.31 only 10% of CO evaporated
Model 4 1100 1.24 0.86 0.51 only 10% survives for T >
100 KNGC 7129–FIRS 2
Model 3 1100 0.97 1.29 2.05 only 10% of CO evaporated
Model 4 1100 1.39 1.48 1.73 only 10% survives for T >
100 K χ is defined as χ = p [ Σ ( I model − I obs ) ] / N . Note that this is an absolute error and the comparison among values in di ff erent sources is notstraightforward. See Figs. 11 to 18. Fig. 9.
The same as Fig. 8 for IC 1396 N. The solution with thelowest RMS, 0.36, is indicated by a cross. We consider accept-able solutions those with RMS less than 0.5. Contour levels are0.5, 1, 2, and 5 K km s − .et al. 2002), no atomic oxygen is included in the code becauseof the large uncertainties associated with its value (see e.g. Cauxet al. 1999 and Melnick & Bergin 2005). This issue is discussedfurther at the end of this section.The abundances of the molecular ions (HCO + , N H + , H + ,and all their deuterated forms) were calculated in terms of theinstantaneous abundances of neutral species, assuming that thetimescale for ion chemistry is much shorter than for freeze-out(Caselli et al. 2002). The rate coe ffi cients are adopted from theUMIST database (http: // + + HD → H D + + H ), we usedthe rates measured by Gerlich et al. (2002), which better fit thedeuterium fractionation in low-mass Class 0 sources, as recentlyfound by Emprechtinger et al. (2009). Hugo et al. (2009) haverecently measured the proton-deuteron rate coe ffi cients again, Fig. 10.
The same as Fig. 8 for CB 3. The solution with the low-est RMS, 2.5, is indicated by a cross. Contour levels are 2.5, 4,6, and 10 K km s − .finding average values of total rates (for H + and its deuteratedisotopologues, the total rate refers to the average of multiplerates weighted according to the fraction of ortho and para forms;Sipil¨a et al. 2010) about 4-5 times more than those derived byGerlich et al. (2002, see also Sipil¨a et al. 2010). Although thisdi ff erence could a ff ect our results by increasing the deuteriumfractionation by a similar factor (thus worsening the compari-son with observations), we decided not to include the new val-ues because the nuclear spin variants of all H + isotopologuesand H have not been distinguished in the current model. Infact, as Pagani et al. (1992) and Flower et al. (2004) showed,small temperature variations significantly alter the ortho-to-pararatio of H , which in turn strongly a ff ects the D-fractionation(ortho-H being more e ffi cient than para-H in driving backthe proton-deuteron exchange reactions thus decreasing the D-fractionation), even in the temperature regime between 9 and . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 11
20 K, where CO is mostly frozen onto dust grains (see also thediscussion about the drop in ortho-H D + column density at tem-peratures above 10 K in Caselli et al. 2008). Such temperaturevariations are definitely present in the envelopes of intermediate-mass Class 0 protostars (see also Emprechtinger et al. 2009), sothat a simple increase in the rate coe ffi cients without account-ing for nuclear spin variants and, in particular, the possible in-crease in the H ortho-to-para ratio may overestimate the D-fractionation calculated by our models. A more detailed chem-ical network is currently under development. The electron frac-tion has been computed using a simplified version of the reac-tion scheme of Umebayashi & Nakano (1990), where a genericmolecular ion mH + is formed via proton transfer with H + , and itis destroyed by dissociative recombination with electrons and re-combination on grain surfaces (using rates from Draine & Sutin1987). Dust grains follow a Mathis et al. (1977; MRN) size dis-tribution, but the minimum size has been increased to 5 × − cm to simulate possible dust coagulation, as in the best-fit modelof the prestellar core L1544 shown by Vastel et al. 2006 (see alsoFlower et al. 2005 and Bergin et al. 2006). The initial abundanceof metals (assumed to freeze out with a rate similar to that ofCO) is 10 − (see McKee 1989). The cosmic ray ionization rateis fixed at ζ = × − s − (van der Tak & van Dishoeck 2000).The adopted CO binding energy is 1100 K, a weighted mean ofthe CO binding energy on icy mantles (1180 K; Collings et al.2003) and CO mantles (885 K; ¨Oberg et al. 2005), assuming thatsolid water is about four times more abundant than solid CO. Themodels are computed for envelope lifetimes of 10 yr, althoughsolutions do not appreciably change after ≃ yr.In Figs. 11 to 14, we show the fits for Cep E-mm. We adoptthis protostar as a fiducial example because it is the one with thebest spatial sampling of the envelope, its geometry looks spher-ical, and the contribution of the surrounding molecular cloud isnegligible. It is impossible to reproduce the observations towardsCep E-mm using the standard chemical model. With the standardmodel, i.e., assuming a CO binding energy of ∼ O and N H + emission (see Fig. 11), but the model reproduces the line inte-grated intensities poorly. The line intensity predictions are a fac-tor of 2–4 higher than the observations (see Fig. 11). Increasingthe CO binding energy (e.g. assuming that a large fraction of COis trapped in water ice) does not improve the fit (see model 2 inTable 6 and Fig. 12).The abundance profiles given in Table 5 suggest that the fitto the C O emission would improve by lowering the CO abun-dance in the inner part of the core. As pointed out there, this lowCO abundance has some physical justification. One possibility isthat a significant fraction of CO is converted into CH OH on thegrain surfaces before evaporation, in agreement with observa-tions of solid methanol along the line of sight towards embeddedyoung stellar objects (e.g. Boogert et al. 2008). We mimic thissituation with model 3. In this model, only 10% of the CO is re-leased back to the gas phase at the CO evaporation temperature.Another possibility is that CO is either destroyed by the stellarUV radiation and / or X-rays close to the star or transformed intomore complex molecules via hot core chemistry. This case corre-sponds to our model 4, where only 10% of the CO survives whenthe temperature is higher than 100 K. The two models, 3 and 4,fit the C O and N H + emission better than the standard model(see Figs. 13 and 14), with model 3 being a slightly better fit tothe Cep E-mm observations. Thus, we applied these most suc-cessful models, model 3 and 4, to all the other sources and haveobtained reasonable fits (line integrated intensities fitted withina factor of 2) to the C O and N H + emission in all of them. For both IC 1396 N and Serpens–FIRS 1, model 4 gives a better fit,while model 3 is better for the rest.While we obtain reasonable fits for C O and N H + , ourmodels do not succeed in reproducing the N D + data towardsSerpens–FIRS 1 and IC 1396 N. In fact, the chemical modelspredict intensities higher by a factor of ∼
10 than the observedintensities for these sources. All the considered models predictthat the spatial distribution of N D + is similar to that of N H + and that the deuteration fraction [N D + ] / [N H + ] ∼ D + ] / [N H + ]ratio is a few 0.001. This discrepancy could have di ff erent ori-gins. First of all, the models assume a spherical geometry. It isclear that bipolar outflows have excavated large cavities in theseprotostellar envelopes (see e.g. Fuente et al. 2009). The walls ofthese cavities are warmed by the stellar UV radiation and shocks.Moderate temperatures, UV radiation, and shocks would lowerthe abundance of N H + and change the [N D + ] / [N H + ] ratio.In fact, the large linewidths observed ( ∼ − ) are hard tomaintain without shocks and dissipation. The role of UV radia-tion and shocks is expected to be greater in IM Class 0 objectsthan in low-mass ones.This simple model ignores many other important parame-ters that could decrease the deuterium fractionation in warmersources. These include (i) an increased abundance of atomicO in the gas phase, possibly coming from the release of waterfrom the icy dust mantles. Atomic oxygen is an e ffi cient de-struction partner for all the H + isotopologues, and lowers theD-fractionation, as discussed by Caselli et al. (2002). (ii) Theortho-to-para H ratio, which in systems out of equilibrium (suchas free-falling) can exceed the steady-state value by more thanan order of magnitude (Flower et al. 2006, Pagani et al. 2009).A larger fraction of ortho-H leads to a lower D-fractionation,given that the more energetic ortho-H can more easily drive theproton-deuteron exchange reactions (e.g. H + + HD → H D + + H ) backward and reduce the H D + / H + abundance ratio (Gerlichet al. 2002). (iii) A higher ionization rate, ζ , which may be dueto the presence of X-rays. A larger ζ implies a larger electronfraction and thus a higher dissociative recombination rates formolecular ions, including H + isotopologues (e.g. Caselli et al.2008). (iv) The presence of small grains (in particular PAHs; seediscussion in Caselli et al. 2008, Sect. 5.1) may arise from theinteraction of outflow lobes and UV radiation with the molec-ular envelope. The associated shocks will partially destroy dustgrains along the way (e.g. Jones et al. 1996; Caselli et al. 1997;Guillet et al. 2009) and increase the surface area for recombina-tion of molecular ions onto dust grains.Because of the low angular resolution of the present obser-vations, it is hard to disentangle the influences of the above pa-rameters. Both more detailed observations and more comprehen-sive models are needed for a deeper understanding of the chem-ical and physical evolution of the envelopes surrounding youngintermediate-mass stars.
8. Summary and conclusions
We carried out a study of the CO depletion and N H + deuterationin a sample of representative IM Class 0 protostars. Our resultscan be summarized as follows. – We observed the millimeter lines of C O, C O, N H + , andN D + using the IRAM 30m telescope in a sample of 7 Class0 and 2 Class I IM stars. We have found a clear evolutionarytrend that di ff erentiates Class 0 from Class I sources. While Fig. 11.
Left: results of our chemical model for Cep E-mm assuming a binding energy of 1100 K (standard value, model 1). TheC O abundance is shown in black, N H + in red, and N D + in blue. Center:
Comparison between the predicted C O 1 → Right:
Same for N H + → Fig. 12.
The same as Fig. 11 for model 2.
Fig. 13.
The same as Fig. 11 for model 3.the emission of the N H + → , is low, be-low a few 0.001 in all Class I sources. There is, however, awide dispersal in the values of R in Class 0 sources rangingfrom a few 0.001 to a few 0.01. This at least two orders ofmagnitude greater than the elemental value in the interstellarmedium, although a factor of 10-100 lower than in prestellarclumps.It is impossible, however, to establish an evolutionary trendamong Class 0 sources based on simple parameters such as the average CO depletion and the average N H + deuteriumfractionation. This stems from the complexity of these re-gions (multiplicity) and the limited angular resolution of ourobservations, which prevents us from tracing the inner re-gions of the envelope. Interferometric observations are re-quired to provide a more precise picture of the evolutionarystage of these objects. – We used a radiative transfer code to derive the C O, N H + ,and N D + radial abundance profiles in 5 IM Class 0 stars. Inparticular, we fit the C O 1 → O abundance decreases inwards within the protostellarenvelope until the gas and dust reach the CO evaporationtemperature, T ev . Our observational data are better fit with . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 13 Fig. 14.
The same as Fig. 11 for model 4.
Fig. 15.
The same as Fig. 11 for IC 1396 N and model 4.
Fig. 16.
The same as Fig. 11 for CB 3 and model 3.
Fig. 17.
The same as Fig. 11 for NGC 7129–FIRS 2 and model 3.
Fig. 18.
The same as Fig. 11 for Serpens–FIRS 1 and model 3.values of T ev ∼ – We determined the chemistry of the protostellar envelopesusing the model by Caselli et al. (2002). A spherical envelopeand steady-state chemical model cannot account for the ob-servations. We had to introduce modifications to better fit theC O and N H + maps. In particular the CO abundance in theinner envelope seems to be lower than the canonical value.This could be due to the conversion of CO into CH OH onthe grain surfaces, the photodissociation of CO by the stel-lar UV radiation, or even geometrical e ff ects. Likewise, wehave problems fitting the low values of the deuterium frac-tionation ( ∼ a few 0.001) measured for some Class 0 IMs.Several explanations have been proposed to account for thisdiscrepancy. Acknowledgements.
This Paper was partially supported by MICINN, within theprogram CONSOLIDER INGENIO 2010, under grant ”Molecular Astrophysics:The Herschel and ALMA Era – ASTROMOL” (ref.: CSD2009-00038)
References
Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197Bechis, K. P., Harvey, P. M., Campbell, M. F., & Ho ff mann, W. F. 1978, ApJ,226, 439Beltr´an, M. T., Girart, J. M., Estalella, R., Ho, P. T. P., & Palau, A. 2002, ApJ,573, 246Beltr´an, M. T., Girart, J. M., Estalella, R., & Ho, P. T. P. 2004, A&A, 426, 941Bergin, E. A., Maret, S., van der Tak, F. F. S., Alves, J., Carmody, S. M., & Lada,C. J. 2006, ApJ, 645, 369Bisschop, S. E., Fraser, H. J., ¨Oberg, K. I., van Dishoeck, E. F., & Schlemmer,S. 2006, A&A, 449, 1297Boogert, A. C. A., et al. 2008, ApJ, 678, 985Casali, M. M., Eiroa, C., & Duncan, W. D. 1993, A&A, 275, 195Caselli, P., Hartquist, T. W., & Havnes, O. 1997, A&A, 322, 296Caselli, P., Vastel, C., Ceccarelli, C., van der Tak, F. F. S., Crapsi, A., &Bacmann, A. 2008, A&A, 492, 703Caselli, P., Walmsley, C. M., Zucconi, A., Tafalla, M., Dore, L., & Myers, P. C.2002, ApJ, 565, 344Caux, E., et al. 1999, A&A, 347, L1Ceccarelli, C., & Dominik, C. 2005, A&A, 440, 583Cesaroni, R., Felli, M., Jenness, T., Neri, R., Olmi, L., Robberto, M., Testi, L.,& Walmsley, C. M. 1999, A&A, 345, 949Chini, R., Ward-Thompson, D., Kirk, J. M., Nielbock, M., Reipurth, B., &Sievers, A. 2001, A&A, 369, 155Codella, C., & Bachiller, R. 1999, A&A, 350, 659Codella, C., Bachiller, R., Nisini, B., Saraceno, P., & Testi, L. 2001, A&A, 376,271Collings, M. P., Dever, J. W., Fraser, H. J., McCoustra, M. R. S., & Williams,D. A. 2003, ApJ, 583, 1058Crapsi, A., Caselli, P., Walmsley, C. M., Myers, P. C., Tafalla, M., Lee, C. W., &Bourke, T. L. 2005, ApJ, 619, 379 Crimier, N., Ceccarelli, C., Lefloch, B., Faure, A. 2009, A&A, 506, 1229Crimier et al. 2010, SubmittedDaniel, F., Cernicharo, J., Roue ff , E., Gerin, M., & Dubernet, M. L. 2007, ApJ,667, 980de Gregorio-Monsalvo, I., G´omez, J. F., Su´arez, O., Kuiper, T. B. H., Anglada,G., Patel, N. A., & Torrelles, J. M. 2006, AJ, 132, 2584Draine, B. T., & Sutin, B. 1987, ApJ, 320, 803Edwards, S., & Snell, R. L. 1983, ApJ, 270, 605Eiroa, C., Palacios, J., & Casali, M. M. 1998, A&A, 335, 243Emprechtinger, M., Caselli, P., Volgenau, N. H., Stutzki, J., & Wiedner, M. C.2009, A&A, 493, 89Felli, M., Palagi, F., & Tofani, G. 1992, A&A, 255, 293Flower, 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, 933Flower, D. R., Pineau Des Forˆets, G., & Walmsley, C. M. 2006, A&A, 449, 621Froebrich, D., Smith, M. D., Hodapp, K.-W., & Eisl¨o ff el, J. 2003, MNRAS, 346,163Fuente, A., Martin-Pintado, J., Bachiller, R., Neri, R., & Palla, F. 1998, A&A,334, 253Fuente, A., Neri, R., Mart´ın-Pintado, J., Bachiller, R., Rodr´ıguez-Franco, A., &Palla, F. 2001, A&A, 366, 873Fuente, A., Martın-Pintado, J., Bachiller, R., Rodrıguez-Franco, A., & Palla, F.2002, A&A, 387, 977Fuente, A., Neri, R., & Caselli, P. 2005a, A&A, 444, 481Fuente, A., Rizzo, J. R., Caselli, P., Bachiller, R., & Henkel, C. 2005b, A&A,433, 535Fuente, A., Ceccarelli, C., Neri, R., Alonso-Albi, T., Caselli, P., Johnstone, D.,van Dishoeck, E. F., & Wyrowski, F. 2007, A&A, 468, L37Fuente, A., et al. 2009, arXiv:0909.2267Frerking, M. A., Langer, W. D., & Wilson, R. W. 1982, ApJ, 262, 590Fukui, Y., Iwata, T., Mizuno, A., Ogawa, H., & Takaba, H. 1989, Nature, 342,161Gerlich, D., Herbst, E., & Roue ff , E. 2002, Planet. Space Sci., 50, 1275Guesten, R., & Marcaide, J. M. 1986, A&A, 164, 342Guillet, V., Jones, A. P., & Pineau Des Forˆets, G. 2009, A&A, 497, 145Hartigan, P., & Lada, C. J. 1985, ApJS, 59, 383Harvey, P. M., Wilking, B. A., & Joy, M. 1984, ApJ, 278, 156Hasegawa, T. I., Herbst, E., & Leung, C. M. 1992, ApJS, 82, 167Hasegawa, T. I., & Herbst, E. 1993, MNRAS, 263, 589Henning, T., Burkert, A., Launhardt, R., Leinert, C., & Stecklum, B. 1998, A&A,336, 565Hily-Blant, P., Walmsley, M., Pineau Des forˆets, G., & Flower, D. 2010, A&Ain press, arXiv:1001.3930Huard, T. L., Weintraub, D. A., & Sandell, G. 2000, A&A, 362, 635Hugo, E., Asvany, O., & Schlemmer, S. 2009, J. Chem. Phys., 130, 164302Hurt, R. L., & Barsony, M. 1996, ApJL, 460, L45Jones, A. P., Tielens, A. G. G. M., & Hollenbach, D. J. 1996, ApJ, 469, 740Jørgensen, J. K., Sch¨oier, F. L., & van Dishoeck, E. F. 2005, A&A, 437, 501Larsson, B., et al. 2000, A&A, 363, 253Launhardt, R., & Henning, T. 1997, A&A, 326, 329Lee, H.-H., Bettens, R. P. A., & Herbst, E. 1996, A&ASS, 119, 111Lefloch, B., Eisloe ff el, J., & Lazare ff , B. 1996, A&A, 313, L17Maret, S., et al. 2004, A&A, 416, 577Maret, S., Bergin, E. A., & Lada, C. J. 2006, Nature, 442, 425Mathis, J. S., Rumpl, W., & Nordsieck, K. H. 1977, ApJ, 217, 425McKee, C. F. 1989, ApJ, 345, 782Melnick, G. J., & Bergin, E. A. 2005, Advances in Space Research, 36, 1027 . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 15 Minchin, N. R., White, G. J., & Ward-Thompson, D. 1995, A&A, 301, 894Miranda, L. F., Eiroa, C., & Gomez de Castro, A. I. 1993, A&A, 271, 564Neri, R., et al. 2007, A&A, 468, L33Noriega-Crespo, A., Moro-Martin, A., Carey, S., Morris, P. W., Padgett, D. L.,Latter, W. B., & Muzerolle, J. 2005, ESA Special Publication, 577, 453Oliveira, C. M., H´ebrard, G., Howk, J. C., Kruk, J. W., Chayer, P., Moos, H. W.,2003, ApJ, 587, 235Pagani, L., Salez, M., & Wannier, P. G. 1992, A&A, 258, 479Pagani, L., Vastel, C., Hugo, E., Kokoouline, V., Greene, C. H., Bacmann, A.,Bayet, E., Ceccarelli, C., Peng, R., & Schlemmer, S. 2009, A&A, 494, 623Patel, N. A., Greenhill, L. J., Herrnstein, J., Zhang, Q., Moran, J. M., Ho, P. T. P.,& Goldsmith, P. F. 2000, ApJ, 538, 268¨Oberg, K. I., van Broekhuizen, F., Fraser, H. J., Bisschop, S. E., van Dishoeck,E. F., & Schlemmer, S. 2005, ApJL, 621, L33Rodriguez, L. F., Moran, J. M., Ho, P. T. P., & Gottlieb, E. W. 1980, ApJ, 235,845Rodr´ıguez, L. F., Curiel, S., Moran, J. M., Mirabel, I. F., Roth, M., & Garay, G.1989, ApJ, 346, L85Sch¨oier, F.L., van der Tak, F.F.S., van Dishoeck E.F., Black, J.H. 2005, A&A432, 369Serabyn, E., Guesten, R., & Mundy, L. 1993, ApJ, 404, 247Sipil¨a, O., Hugo, E., Harju, J., Asvany, O., Juvela, M., & Schlemmer, S. 2010,A&A, 509, A98Sugitani, K., Matsuo, H., Nakano, M., Tamura, M., & Ogura, K. 2000, AJ, 119,323Tofani, G., Felli, M., Taylor, G. B., & Hunter, T. R. 1995, A&A Sup, 112, 299Umebayashi, T., & Nakano, T. 1990, MNRAS, 243, 103Van der Tak, F. F. S., & van Dishoeck, E. F. 2000, A&A, 358, L79Van der Tak, F.F.S., Black, J.H., Schier, F.L., Jansen, D.J., van Dishoeck, E.F.2007, A&A, 468, 627Vastel, C., Caselli, P., Ceccarelli, C., Phillips, T., Wiedner, M. C., Peng, R.,Houde, M., & Dominik, C. 2006, ApJ, 645, 1198Wang, Y., Evans, N. J., II, Zhou, S., & Clemens, D. P. 1995, ApJ, 454, 217Wilking, B., Mundy, L., McMullin, J., Hezel, T., & Keene, J. 1993, AJ, 106, 250Wilson, T. L., & Matteucci, F. 1992, Astronomy and Astrophysics Review, 4, 1Wilson, T. L., & Rood, R. 1994, Astronomy and Astrophysics Review, 32, 191Yun, J. L., & Clemens, D. P. 1994, AJ, 108, 612
Appendix A:
Below we discuss the details of the modeling for each individualsource.
A.1. Serpens-FIRS 1
Serpens-FIRS 1, located near the center of the Serpens maincore, is the most luminous object embedded in the cloud. Severalcontinuum studies lead to its classification as a Class 0 sourcewith a bolometric luminosity estimated to range from 46 L ⊙ to84 L ⊙ (Harvey et al. 1984; Casali et al. 1993; Hurt & Barsony1996; Larsson et al. 2000). The latest estimate of its luminos-ity (see Table 4) suggests that Serpens-FIRS 1 is on the lowmass / IM borderline. Serpens-FIRS 1 drives a molecular out-flow that is oriented at a position angle of 50 ◦ (Rodr´ıguez et al.1989). C10 modeled this source as a sphere with an outer radiusof 5900 AU, 25.5” at the distance of Serpens. Convolving thissize with the observational beams, one would expect a sourcediameter of 55”, 57”, and 53” for the C O 1 →
0, N H + → D + → > = = × cm − ,T d =
13 K) and the values assumed at r = = × cm − and T d =
10 K. This profile was used for our fit-ting.The integrated intensity map of C O cannot be fit witha constant abundance profile. Thus, we decided to assume astep function for the C O abundance: (i) a constant abun-dance, X , that is expected to be close to the canonical value(X = ± × − ) for radii lower than a given radius, R and (ii)an abundance, X , that is expected to be < X for larger radii.Thus defined R , is the evaporation radius of CO. The values ofX , R , and X were fit by the model. This approximation wasstill not su ffi cient to produce a good fit. We had to add anotherstep, and two new variables, R and X , in the C O abundanceprofile. The best fit is shown in Table 5. We have (i) a warmregion (R < O abundance ∼ × − , (ii)an intermediate layer with a high value for the C O depletion,f D ∼
20, and (iii) an external layer (R > O abun-dance is close to the canonical value again.We followed the same procedure for N H + . In this case theemission extends to the NE and greatly di ff ers from sphericalsymmetry. For this reason we masked the NE quadrant in ourfitting (see Fig. 4). Again, we conclude that the N H + abun-dance has a standard value of a few 10 − in the inner region(R < D + → D + emission is dominatedby the molecular cloud component. Similar to the case ofN H + , we masked the NE quadrant in our fitting. We foundX(N D + ) = × − in the molecular cloud, and obtained anupper limit of X(N D + ) < × − for the protostellar core.Thus we have a deuterium fractionation of ∼ A.2. Cep E
Cep E-mm was cataloged as a Class 0 protostar by Lefloch etal. (1996). Cep E-mm was observed with IRAM 30m (Leflochet al. 1996; Chini et al. 2001), SCUBA (Chini et al. 2001), ISO(Froebrich et al. 2003), and Spitzer (Noriega-Crespo et al. 2005).All these studies confirm the Class 0 status of Cep E-mm andconstrain the source total mass and bolometric luminosity inthe range of 7-25 M ⊙ and 80-120 L ⊙ , respectively. A bipolarmolecular outflow, first reported by Fukui et al. (1989), is asso-ciated with Cep E-mm. The H and [FeII] study by Eislo ff el etal. (1996) shows a quadrupolar outflow morphology suggestingthat the driving source is a binary.C10 modeled this source as a sphere with an outer radiusof 35800 AU, 49” at the distance of Cepheus (see Table 4).Convolving this size with the observational beams, one wouldexpect a source diameter of ∼ ∼ O 1 →
0, N H + →
0, and N D + → × − for radii below 3500 AU, anda power-law variation of the C O abundance for larger radii(see Table 5). The high value of f D , ∼
10, would occur closeto R = H + abundance. In this case, the quality of the N D + → ∼ × − by fitting thespectrum observed towards the center position. A.3. IC 1396N
IC 1396 N is the globule associated with IRAS 21391 + O1 → H + → ff ect our results significantly (more than a factor of 2 in theabundances).Similarly to Cep-E, the C O emission was better fit with aconstant and close-to-standard abundance of 6 × − in the in-ner region (R < O . Alonso-Albi et al.: Chemical study of intermediate-mass (IM) Class 0 protostars. 17 abundance for larger radii (see Table 5). The highest value of f D is ∼
5, and it would occur close to R = H + emis-sion was better fit with a two-step function, X N H + = × − for R < × − for R > × − in between. The last step could be due to the vicinity of thesources BIMA 3 and BIMA 2 that heat the outer part of the en-velope. These sources are not considered in the (n-T) fit by C10.In IC 1396 N, we have not detected the N D + → D + abundance of < × − . Thelower value of f D and the non-detection of N D + is consistentwith this source being a warmer and more evolved object. A.4. CB3
The CB3 Bok globule is located at ∼ O masers (de Gregorio-Monsalvo et al. 2006). This outflow has been mapped in variousmolecular lines by Codella & Bachiller (1999), who concludedthat it originates from CB3-mm. The same authors concludedthat CB3-mm is probably a Class 0 source.CB3 is di ff erent from all the other sources in our sample. Inthis object, the 24 µ m sources are not spatially coincident withthe 850 µ m emission peak. CB3-mm hosts two 24 µ m sourcesseparated by ∼
12” and neither of them is spatially coincidentwith the column density peak. The column density peak, bettertraced by the 850 µ m continuum emission, is located in betweenand almost equidistant from the two 24 µ m sources (see Fig. 4).C10 modeled the SED of this source by adding up the flux ofthe two 24 µ m sources. They fit the spatial distribution of the450 µ m and 850 µ m maps as a sphere with an outer radius of103000 AU (i.e. 41”) (see Table 4). The radial density powerlaw in this source is steeper, ∼
2, than in the others, with lowdensities ( ∼ a few 10 cm − ) in the outer envelope, > O emission with a step function. The abun-dance is constant at 1.3 × − for radii less than 25000 AU,decreases to < × − ( f D > × − for R > O abundanceat large radii is very likely caused by the modeled extendedlow-density envelope providing a poor approximation. Morelikely, there are dense clumps immersed in a lower densitycloud. The N H + emission is fit with a constant abundanceof X N H + = × − . The N D + emission is fit with a con-stant abundance of X N D + = × − , implying an average[N D + ] / [N H + ] = A.5. NGC 7129–FIRS2
NGC 7129 is a reflection nebula located in a complex and activestar-forming site at a distance of 1250 pc (Hartigan & Lada 1985;Miranda et al. 1993). NGC 7129 FIRS 2 is not detected at opti-cal or near-infrared wavelengths. Its position is spatially coinci-dent with a CO column density peak (Bechis et al. 1978) anda high-density NH cloudlet (Guesten & Marcaide 1986), and itis close to an H O maser (Rodr´ıguez et al. 1980). NGC7129–FIRS 2 has been classified as a Class 0 IM protostar by Eiroa etal. (1998), who carried out a multi-wavelength study of the con- tinuum emission from 25 to 2000 µ m. Edwards & Snell (1983)detected a bipolar CO outflow associated with FIRS 2. The inter-ferometric study by Fuente et al. (2001) reveals that this outflowpresents a quadrupolar morphology. In fact, the outflow seemsto be the superposition of two flows, FIRS 2-out 1 and FIRS 2-out 2, likely associated with FIRS 2 and a more evolved star(FIRS 2-IR), respectively. Fuente et al. (2005a,b) carried out acomplete chemical study of FIRS 2 providing the first detectionof hot core in an IM Class 0. Based on all these studies, FIRS 2is considered the youngest IM object known at present.C10 modeled NGC 7129–FIRS 2 as a sphere with an outerradius of 18600 AU, 15” at the distance of NGC 7129 (see Table4). Since the source is very compact and barely resolved in themaps of the C O 1 → H + → D =
3, consistent with the depletion factor foundby Fuente et al. (2005a) based on H CO + observations. Theprotostellar envelope is, however, resolved at the frequency ofthe N D + → D + emissionis absent towards the hot core, we fitted the N D + → D + abundance is ∼ × − , and the deuteriumfractionation, ∼∼