MNRAS 477, 3390–3396 (2018) doi:10.1093/mnras/sty693Advance Access publication 2018 March 15

Project VeSElkA: results of abundance analysis for HD 53929 and HD63975

M. L. Ndiaye,‹ F. LeBlanc‹ and V. Khalack‹

Departement de Physique et d’Astronomie, Universite de Moncton, Moncton, NB E1A 3E9, Canada

Accepted 2018 February 28. Received 2018 February 22; in original form 2017 September 6

ABSTRACTProject VeSElkA (Vertical Stratification of Element Abundances) has been initiated with theaim to detect and study the vertical stratification of element abundances in the atmosphere ofchemically peculiar stars. Abundance stratification occurs in hydrodynamically stable stellaratmospheres due to the migration of the elements caused by atomic diffusion. Two HgMnstars, HD 53929 and HD 63975, were selected from the VeSElkA sample and analysed withthe aim to detect some abundance peculiarities employing the ZEEMAN2 code. We presentthe results of abundance analysis of HD 53929 and HD 63975 observed recently with thespectropolarimeter ESPaDOnS (Echelle SpectroPolarimetric Device for the Observation ofStars) at Canada–France–Hawaii Telescope. Evidence of phosphorus vertical stratificationwas detected in the atmosphere of these two stars. In both cases, phosphorus abundanceincreases strongly towards the superficial layers. The strong overabundance of Mn found instellar atmosphere of both stars confirms that they are HgMn-type stars.

Key words: atomic processes – line: formation – line: profiles – stars: atmospheres – stars:chemically peculiar – stars: individual: HD 53929, HD 63975.

1 IN T RO D U C T I O N

The HgMn stars constitute a subgroup of the chemically peculiar(CP) stars of the main sequence. They are characterized spectro-scopically by a very large overabundance of mercury and man-ganese. They constitute the third CP star group of the classificationdone by Preston (1974). Indeed, the overabundance of mercury canvary up to 6 dex (e.g. Heacox 1979; Cowley et al. 2006) and upto 3 dex for manganese (e.g. Aller 1970; Morel et al. 2013), bothrelative to their solar value. Chemical elements such as Be, P, Cu,Ga, Sr, Y, Zr, Yb, Pt, and Bi are often overabundant in the atmo-spheres of HgMn stars; in fact, their overabundance is sometimesgreater than 2 dex (e.g. Takada-Hidai 1991; Smith 1993; Castelli &Hubrig 2004). In addition to these overabundant elements, He, N,Mg, Al, Co, Ni, and Zn are generally deficient in HgMn stars buttheir underabundance often does not exceed 0.5 dex (Takada-Hidai1991).

Although several studies have attempted to detect the presenceof a magnetic field (e.g. Mathys & Hubrig 1995; Hubrig, Castelli &Wahlgren 1999; Hubrig & Castelli 2001; Hubrig et al. 2012), thesestars are still considered in the literature as non-magnetic stars or atleast very weakly magnetic stars.

The HgMn stars are characterized by a relatively low rotationalvelocity (V sin i < 100 km s−1; Wolff & Preston 1978) and have

� E-mail: mamadou.lamine.ndiaye@gmail.com (MLN); francis.leblanc@umoncton.ca (FLB); viktor.khalack@umoncton.ca (VK)

effective temperatures between 10 000 and 16 000 K (e.g. Smith1996a). There should not be a hydrogen convective zone near thesurface of HgMn stars for the simple reason that these stars havehigh effective temperatures (e.g. Michaud, Alecian & Richer 2015).Their low rotational velocity means that the atmosphere of thesestars might be hydrodynamically stable, which would allow theprocess of atomic diffusion (Michaud 1970) to take place there anddetermine the chemical properties of the outer layers of HgMn stars.Atomic diffusion can cause a vertical abundance stratification of ele-ments present in the stellar atmosphere of HgMn stars, which in turncan change their physical structure (e.g. Hui-Bon-Hoa, LeBlanc &Hauschildt 2000; LeBlanc et al. 2009).

Several studies have revealed signs of vertical stratification ofmetals in the stellar atmosphere of certain HgMn stars. Verticalstratification of Mn was detected by Alecian (1982) in the atmo-sphere of ν Her and then by Sigut (2001) in the atmospheres of HD186122 (46 Aql) and HD 179761. Vertical stratification of Cr wasdiscovered by Savanov & Hubrig (2003) in a sample of 10 HgMnstars, who also found that its abundance increases towards the up-per atmospheric layers in all studied stars except for HD 49606.Thiam et al. (2010) have found a vertical stratification of Mn in HD178065, while Catanzaro et al. (2016) found evidence of Mg, Si,and P stratification in the HgMn star HD 49606.

In addition to the aforementioned chemical anomalies, HgMnstars also exhibit isotopic anomalies on their surface. For instance,in some HgMn stars (such as χ Lupi), the relative abundances ofthe Hg and Pt isotopes are very different from their solar values(White et al. 1976; Dworetsky, Storey & Jacobs 1984; Leckrone,

C© 2018 The Author(s)Published by Oxford University Press on behalf of the Royal Astronomical Society

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022

Project VeSElkA: HD 53929 and HD 63975 3391

Table 1. Observation data of HD 53929 and HD 63975.

Star Date HJD texp S/N(UTC) (2400000+) (s) Stokes I/V

HD 53929 2013 Jan 1 56293.95962 1160 1070/870HD 63975 2016 Feb 22 57440.81245 348 520/420

Wahlgren & Johansson 1991; Leckrone et al. 1993). These isotopicanomalies are probably due to the physical process of light-induceddrift (see Atutov & Shalagin 1988 and LeBlanc & Michaud 1993for more details).

Recently, a research project named VeSElkA (Vertical Stratifica-tion of Element Abundances) was initiated by Khalack & LeBlanc(2015a,b) to search for vertical stratification of element abun-dances in stellar atmospheres of main-sequence CP stars. Morethan 50 slowly rotating objects with V sin i < 40 km s−1 were se-lected from the catalogue of Ap, HgMn, and Am stars of Ren-son & Manfroid (2009) and observed with the spectropolarimeterESPaDOnS (Echelle SpectroPolarimetric Device for the Observa-tion of Stars) at the Canada–France–Hawaii Telescope (CFHT) inthe frame of this project. Project VeSElkA aims to identify verticalstratification of elements in a large number of CP stars in order toguide the theoretical modelling of their atmosphere and to search forcorrelations between the global stellar parameters and the detectedvertical abundance stratification. Several stars from this sample havealready been analysed, and some of them revealed signatures of ver-tical stratification of metals. HD 22920 shows vertical stratificationof Cr (Khalack & Poitras 2015), HD 95608 and HD 116235 showvertical stratification of Fe and Cr (LeBlanc et al. 2015), while HD41076 and HD 148330 show vertical stratification of several metals(Khalack, Gallant & Thibeault 2017).

The purpose of this paper is to perform a spectral analysis of twostars (HD 53929 and HD 63975) identified as HgMn type by Ren-son & Manfroid (2009). These two stars are part of the VeSElkAsample, each exhibiting a slow axial rotation, thus possessing ahydrodynamically stable stellar atmosphere, where atomic diffu-sion can be effective. Smith & Dworetsky (1993) estimated theabundance of iron-peak elements Cr, Mn, Fe, Co, and Ni in HD53929 using IUE spectra. Using the same spectra, Smith (1993) ob-tained its abundances for Mg, Al, and Si, Smith (1994) estimated itsCu and Zn abundances, Smith (1996b) acquired its Ga abundance,while Smith (1997) evaluated its Hg abundance. For HD 63975, theonly element for which we have found an abundance estimation isHg that was obtained by Woolf & Lambert (1999). In this paper,recently obtained spectra of HD 53929 and HD 63975 are studiedto perform an abundance analysis of their stellar atmospheres for alarge number of elements.

In Section 2, we provide a brief description of the acquired spectraand the reduction procedure. The method used for the abundanceanalysis and the obtained results are presented in Section 3, followedby the discussion in Section 4.

2 O B S E RVAT I O N S A N D DATA R E D U C T I O N

High-resolution (R = λ/(�λ) = 65 000) Stokes IV spectra with ahigh signal-to-noise (S/N) ratio were obtained using ESPaDOnSat CFHT. ESPaDOnS is a bench-mounted high-resolution echellespectropolarimeter designed to obtain a complete optical spectrum(i.e. from 3700 to 10 500 Å) in a single exposure. It is able to measureall polarization components of the stellar light (both circular andlinear) with the same resolving power (Donati 2003). The spectrawere reduced with the help of the data reduction package LIBRE-ESPRIT developed by Donati et al. (1997). In the case of HgMn stars,the Stokes V spectrum shows no detectable signal because they aremost probably non-magnetic stars. Thus, within this framework,we only use the Stokes I spectra for our analysis. Table 1 presentsthe journal of spectral observations, where the first column gives thestar’s name, the second and the third columns provide respectivelythe date and Heliocentric Julian Date of observation, and the fourthand fifth columns present the exposure time and the signal-to-noiseratio for the Stokes I and V spectra.

3 SP E C T R A L A NA LY S I S A N D R E S U LT S

3.1 Analysis method

The abundance analysis of stellar atmospheres of HD 53929 andHD 63975 has been carried out in several steps. The first step was toidentify all absorption lines available in the observed spectra usingthe NIST (Kramida et al. 2015) and VALD3 (Piskunov et al. 1995;Ryabchikova et al. 1997, 2015; Kupka et al. 1999, 2000) data bases.The damping constants are taken from the VALD3 data base, andthe log (gf) values of this data base are used in the majority of oursimulations. When unavailable in the VALD3 data base, the log (gf)values provided by NIST were employed.

For each star, the physical parameters used for our simulationsare presented in Table 2. These parameters were estimated for HD53929 and HD 63975 with the help of FITSB2 code (Napiwotzki et al.2004) by fitting the Balmer line profiles in the non-normalizedspectra of studied stars with non-normalized theoretical fluxes(Khalack & LeBlanc 2015b). For both stars, the best fits are shownin Fig. 1. For each Balmer line, the fit quality is presented at thebottom of each image in the form of differences between the ob-served and synthetic spectra. Stellar atmosphere models with thecorresponding fluxes were computed for the studied stars using ver-sion 15 of PHOENIX code (Hauschildt, Baron & Allard 1997) and thevalues of Teff, log (g), and metallicity listed in Table 2. The uncer-tainties given for the basic parameters of the stars shown in Table 2were estimated similarly to Khalack & LeBlanc (2015b).

For both stars, the values of Teff obtained in this study fromthe fitting of Balmer line profiles are consistent with the re-sults derived from the photometric temperature [c1]-calibration(Napiwotzky et al. 1993) taking into account the correction pro-posed by Netopil et al. (2008) for the HgMn stars (see Table 2).To carry out the photometric temperature [c1]-calibration for HD53929 and HD 63975, we have used the results of uvby photometry

Table 2. Physical parameters found for HD 53929 and HD 63975.

Star Teff (K) log (g) Vr (km s−1) V sin (i) (km s−1) [M/H] χ2

HD 53929 12 764 ± 200 12 552± 380a 3.71 ± 0.20 14.2 ± 2.1 25.8 ± 1.9 −1.0 2.20HD 63975 12 089 ± 200 12 283± 380a 3.27 ± 0.20 32.8 ± 1.6 28.7 ± 0.9 −0.5 2.06

aData obtained from the photometric temperature [c1]-calibration (Napiwotzki, Schonberner & Wenske 1993) and corrected according to Netopil et al.(2008).

MNRAS 477, 3390–3396 (2018)

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022

3392 M. L. Ndiaye, F. LeBlanc and V. Khalack

Figure 1. The Balmer line profiles in the observed spectra (thick line) of HD 53929 (right) and HD 63975 (left) are relatively well adjusted by the syntheticspectra (thin dotted line) resulting in Teff= 12 764 K, log (g)= 3.71 for HD 53929 and Teff= 12 089 K, log (g)= 3.27 for HD 63975. To visualize the quality offits, the differences between the observed and synthetic spectra are shown at the bottom of each image. For the sake of visibility, the Balmer line profiles areshifted by 0.5 and the differences are shifted by 0.1.

published respectively by Hauck & Mermilliod (1998) and Napi-wotzky et al. (1993). Then the obtained values of effective tem-perature are corrected employing the empirical formula derived byNetopil et al. (2008) for the CP3 and CP4 stars. The errors givenin Table 2 are obtained by taking into account the precision ofStromgren photometry given in Hauck & Mermilliod (1998) and inNapiwotzky et al. (1993), and the errors for each component in thecorrection formula (Netopil et al. 2008).

In the case of HD 53929, our estimation of surface gravity is closeto the value of log (g) = 3.60 ± 0.25 derived by Smith & Dworetsky(1993) for this star (see Table 2). Meanwhile, the obtained valueof Teff is significantly smaller than the ones reported by Khalack& LeBlanc (2015b) and by Smith & Dworetsky (1993). This dis-crepancy can be explained by the fact that Smith & Dworetsky(1993) have adopted Teff and log (g) as the mean values of estimatesderived from the analysis of Stromgren photometry data (Hauck& Mermilliod 1980) using the (c0, β) grids of Moon & Dworet-sky (1985) and from the ‘best fit’ of only one Balmer line profileHγ . Meanwhile, to fit the Balmer line profiles in the spectrumof HD 53929, Khalack & LeBlanc (2015b) have mistakenly useda grid of normalized theoretical fluxes, for which the FITSB2 code(Napiwotzki et al. 2004) usually results in overestimated values ofTeff and log (g). The effective temperatures and gravities publishedby Khalack & LeBlanc (2015b) for the other stars in their samplewere obtained using the grids of non-normalized theoretical fluxes.

We then analysed the line profiles of detected ions using the modi-fied ZEEMAN2 code (Khalack & Wade 2006) developed by Landstreet(1988). Knowing that HgMn stars are generally non-magnetic stars,

all parameters related to the magnetic field structure are set to zeroin our model. By fitting the synthetic profile with the observedone, the code determines three free model parameters: the chemicalabundance of the ion, the radial velocity, and the rotational velocityof the star.

For each line, the simulation is repeated several times in orderto reach the absolute minimum in the space of free model param-eters and to get closer to real values of chemical abundance of thestudied ion, radial velocity Vr, and rotational velocity V sin i of thestar. This was done by changing the initial values of the aforemen-tioned parameters at each simulation and by using the downhillsimplex method (Press et al. 1992) to find the best fit (Khalack& Wade 2006). A preliminary analysis of these three parametersallows us to decide which lines are well fitted and which ones maybe misidentified or are blends due to a significant contribution fromother ions.

The ZEEMAN2 code allows one to estimate the optical depth τl atwhich the core of a spectral line is formed (Khalack et al. 2007,2017). For each analysed line profile, the code determines the abun-dance of studied element and the optical depth τl of the line coreformation. Then it finds the respective optical depth τ5000 of standardscale (calculated for λ = 5000 Å). In this method, the standard scaleof optical depth τ5000 is specified according to the PHOENIX model(Hauschildt et al. 1997) for all layers of the stellar atmosphere.

Thus, when a fairly large number of line profiles are detected fora given ion, and assuming that their cores are formed at differentdepths, it is then possible to probe the atmosphere for any variationin abundance as a function of optical depth. In other words, one can

MNRAS 477, 3390–3396 (2018)

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022

Project VeSElkA: HD 53929 and HD 63975 3393

Table 3. Average abundances obtained for HD 53929 and HD 63975.

El. HD 53929 : [X/H] HD 63975 : [X/H]N Here Other N Here Other

studies studies

C II 0 – – 1 0.15 –O I 4 − 1.33 ± 0.64 – 8 − 0.55 ± 0.42 –Ne I 0 – – 4 2.97 ± 0.29 –Mg II 2 − 0.97 ± 0.23 − 0.94 ± 0.18a 6 − 0.65 ± 0.08 –Si II 9 0.13 ± 0.12 0.04 ± 0.07a 6 − 0.42 ± 0.40 –P II 20 2.01 ± 0.23 – 13 1.86 ± 0.31 –S II 4 − 0.18 ± 0.50 – 0 – –Ti I 0 – – 4 − 0.33 ± 0.71 –Ti II 3 0.77 ± 0.31 – 5 0.02 ± 0.13 –Cr I 0 – – 2 0.67 ± 0.52 –Cr II 2 2.13 ± 0.86 − 1.22 ± 0.10b 0 – –Mn I 2 1.01 ± 0.33 – 2 1.15 ± 0.31 –Mn II 7 1.34 ± 0.86 0.73 ± 0.20b 7 1.31 ± 0.63 –Fe II 28 0.73 ± 0.31 0.23 ± 0.05b 80 0.57 ± 0.24 –Hg II 0 ≤ 2.5 0.93 ± 0.10c 1 4.66 2.94 ± 0.14d

Notes.aSmith (1993), bSmith & Dworetsky (1993), cSmith (1997), dWoolf & Lambert (1999).

verify if element’s abundance is vertically stratified in the stellaratmosphere of the studied star. The slope of the abundance changeis considered to be statistically significant if its measured value ex-ceeds its uncertainty more than three times. When this condition issatisfied and the abundance of an element varies strongly (by 0.5 dexor more) in the probed layers, we may then conclude that a verticalstratification is detected for that element. This method of analysiswas first applied to detect the vertical stratification of certain chem-ical elements including iron (Khalack et al. 2007, 2008; Khalack,LeBlanc & Behr 2010) in stellar atmospheres of blue horizontal-branch stars (i.e. helium-burning stars). Theoretical models such asHui-Bon-Hoa et al. (2000) and LeBlanc et al. (2009, 2010) alsopredict the stratification of some elements in these stars. In oursimulations, the estimation errors are caused mainly by the uncer-tainties present in the atomic data, in the observation data and inthe calculation of the theoretical models. Taking into account thatthese uncertainties are not easy to measure, the requirement for thestrong variation of abundance (0.5 dex) is imposed to ensure that thedetected abundance trend does not represent an arbitrary variationdue to these uncertainties.

3.2 Radial and rotational velocities

The averages of the radial and rotational velocities derived fromthe spectral lines selected for each star are shown in Table 2.For HD 53929, the values of the radial velocity 13.3 km s−1

obtained by Zentelis (1983) and 15 ± 1 km s−1 derived by Khalack& LeBlanc (2015b) are in good agreement with our result. The value6.1 km s−1 derived by Evans (1967) and Hube (1970), and the value11.2 ± 1.3 km s−1 obtained by Gontcharov (2006) for HD 53929 arelower than ours and may suggest that this star may be a member of along periodic binary system. The average rotational velocity foundin this study for HD 53929 is also consistent with the values of 30and 25 km s−1 previously found respectively by Smith & Dworetsky(1993) and by Royer et al. (2002) taking into account the simula-tion errors. For HD 63975, the radial velocity value of 32.3 km s−1

obtained by Wilson (1953) is in good agreement with the datapresented in Table 2. The value of 30 km s−1 found by Wolff &Preston (1978) for the rotational velocity of HD 63975 is also con-sistent with our result.

3.3 Abundance analysis

Table 3 presents results derived from the analysis of absorption lineprofiles found in the obtained spectra of HD 53929 and HD 63975.For each ion, we show the derived average abundance relative toits solar value (Grevesse et al. 2010, 2015; Scott et al. 2015a,b),where N represents the number of selected lines of the studied ion.The estimation errors of average abundances are determined bycalculating the mean squared error of the data obtained for all linesof the considered ion. Results of previous studies are also shownin Table 3 for each reported element, which were recalculated withrespect to recent measurements of solar abundances (Grevesse et al.2010, 2015; Scott et al. 2015a,b).

In contrast to studies using the IUE spectra (Smith 1993, 1997;Smith & Dworetsky 1993), we have detected He I, O I, P I, S II, andTi II lines in the visible spectrum of HD 53929, but no Al II, Co II,Ni II, Hg I, or Hg II lines were found in our ESPaDOnS spectra. Forthe average abundances in this star, we found that O and Mg areunderabundant, that Si and S have a nearly solar abundance, whileTi and Fe are slightly overabundant and P, Cr, and Mn are largelyoverabundant. Even though no Hg lines are visible in our spectra,we used its 3984 Å line to estimate its upper limit abundance at+2.5 dex relative to solar. Our results for the Mg and Si abun-dances are consistent with the ones found by Smith (1993). The Mnoverabundance found here is consistent with that found by Smith& Dworetsky (1993), while our results for Cr and Fe are differentfrom theirs, especially Cr that differs greatly. The higher S/N ratioobtained at CFHT as compared to the IUE spectra used by Smith &Dworetsky (1993) could partly explain these discrepancies. The useof different radiative transfer codes and model atmospheres couldalso be contributing factors. On the other hand, the detected discrep-ancies could be an indication of possible binary nature of HD 53929as it was predicted by Zentelis (1983). In a close binary system, aweaker secondary component can contribute to the observed lineprofiles (see e.g. Ryabchikova et al. 1998), and this contributionwill vary with the phase of orbital motion.

For HD 63975, our analysis shows that O and Mg are underabun-dant, that C, Si, and Ti are near their solar abundance (i.e. within0.5 dex), while Cr and Fe are slightly overabundant and Ne, P, Mn,and Hg show large overabundances. The Hg abundance found hereis larger than the one found by Woolf & Lambert (1999). Since our

MNRAS 477, 3390–3396 (2018)

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022

3394 M. L. Ndiaye, F. LeBlanc and V. Khalack

Table 4. The list of spectral lines used for the abundance analysis of O I inthe spectrum of HD 53929. The complete list of spectral lines used for theabundance analysis of both stars can be found online.

Ion λ (Å) log NX/Ntot log (gf) Ei (cm−1)

O I 6158.187 −5.392 ± 0.001 − 0.409 86 631.45O I 7771.940 −3.916 ± 0.007 0.369 73 768.20O I 7775.388 −4.445 ± 0.013 0.002 73 769.08O I 8446.360 −4.171 ± 0.012 0.236 76 794.98

analysis of Hg is solely based on a single line (3984 Å), it is notpossible to estimate an uncertainty for our value, but it could berelatively large.

Generally speaking, the average abundances derived here fortwo stars are consistent with those found in HgMn stars (e.g.Ghazaryan & Alecian 2016). In particular, the underabundance ofMg and the overabundance of P and Mn found in these two ob-jects confirm their classification as HgMn stars as it was suggestedpreviously by Renson & Manfroid (2009).

However, a possible source of uncertainty in our results is relatedto possible non-local thermodynamic equilibrium (NLTE) effects.For example, Takeda & Honda (2016) studied NLTE effects on theoxygen abundance of B-type stars using the O I triplet at 7771 Å.They showed that this triplet is stronger in NLTE, therefore leadingto a smaller abundance ranging from 0.6 to 1.7 dex in B-type stars.The NLTE analysis of the O I triplet at 7771 Å would then result ina higher deficit of oxygen in the atmospheres of the stars studiedhere (see Table 4). This example shows that NLTE effects can beimportant for certain atomic lines.

3.4 Vertical abundance stratification

For some chemical species, we have identified a sufficient numberof line profiles that can be used to analyse a possible variation ofelement’s abundance with optical depth in the stellar atmospheresof studied stars.

As discussed in Section 1, some elements such as Mn and Crhave been observed to be stratified in the stellar atmosphere ofseveral HgMn stars. The list of spectral lines used to derive theaverage abundance of O I in HD 53929 is presented in Table 4 asan example. The first and second columns indicate respectively the

considered ion and the central wavelength of the spectral line. Inthe third, fourth, and fifth columns, we present respectively theabundance found, the oscillator strength, and the associated lowerenergy level for each spectral line. Complete selected line lists forall ions studied are given in the electronic edition of the journal.

The distributions of the chemical abundance of the analysed ionsas a function of the optical depth are shown in Fig. 2 for HD53929 and in Fig. 3 for HD 63975, and the linear regressions ob-tained with the help of the software GNUPLOT (version 5.0 of 2015January; Williams & Kelley 1986) taking into account only the de-rived abundance estimates. In each graph, each point represents anabundance estimate obtained from the analysis of a single line pro-file. We present these graphs only for chemical species for whichwe have obtained abundance estimates from the analysis of at least10 line profiles. The procedure of the determination of the opticaldepth of line core formation is described in detail in Khalack et al.(2007, 2017). The dashed line shows the level of the solar abun-dance of the studied chemical element. The straight line shown ineach graph represents the best linear regression derived using theleast-squares method. For each case, the slope and its uncertaintyhave been calculated using the GNUPLOT software by averaging thequadratic errors between the linear regression and the data points.

Fig. 2 shows that phosphorus appears to have a strong verticalstratification in the outer atmospheric layers of HD 53929 witha statistically significant slope (−0.51 ± 0.08) and a strong abun-dance variation (more than 0.5 dex). As mentioned previously, sinceseveral sources of uncertainties exist in abundance analysis (e.g.atomic data, underlying model atmospheres, NLTE effects, etc.),a strong abundance variation (of at least 0.5 dex or more) in theatmospheric layers where the selected lines are formed is requiredto confidently confirm the presence of stratification. Meanwhile,the slope (0.18 ± 0.07) found for Fe is not statistically significantand its abundance does not vary strongly in HD 53929 (see Fig. 2).Therefore, we cannot conclude that Fe is stratified in the stellaratmosphere of HD 53929.

Phosphorus also shows a clear vertical stratification in the stellaratmosphere of HD 63975 (see Fig. 3), where the slope is found tobe statistically significant (−0.87 ± 0.17) and a strong abundancevariation exists. As in the other star, Fe shows an abundance slopethat is not statistically significant (0.04 ± 0.04) in the atmosphereof HD 63975.

Figure 2. The abundance of P and Fe relative to the total number of atoms obtained from individual lines for HD 53929 under consideration as a functionof optical depth at 5000 Å. The dashed line represents the solar abundance of the chemical element with respect to the Ntot. Linear fits of the abundancestratification are also shown (solid lines).

MNRAS 477, 3390–3396 (2018)

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022

Project VeSElkA: HD 53929 and HD 63975 3395

Figure 3. The abundance of P and Fe relative to the total number of atoms obtained from individual lines for HD 63975.

4 D I S C U S S I O N A N D C O N C L U S I O N S

The aim of this paper was to perform a spectral analysis of twoHgMn-type stars (HD 53929 and HD 63975) selected from thesample of project VeSElkA, in order to verify for the presence ofvertical stratification of element abundances in their atmosphere.We first identified the absorption lines present in each spectrumand then carried out simulations using ZEEMAN2 code (Landstreet1988) modified by Khalack & Wade (2006), which allowed us toestimate the average abundance of each detected ion along with theradial and the rotational velocity of each star. Our simulations werecarried out assuming LTE; therefore, NLTE effects remain a sourceof uncertainty.

We have analysed absorption lines of several ions to deter-mine their average abundance in the atmosphere of HD 53929and HD 63975 (see Table 3). Oxygen and magnesium are foundto be deficient in the atmosphere of the two stars, while phospho-rus, chromium, manganese, and iron are overabundant. Mercury isfound to be strongly overabundant in HD 53929. The strong over-abundance obtained for Mn and the abundances found for the otherelements allow us to conclude that HD 53929 and HD 63975 are in-deed HgMn stars as previously determined by Renson & Manfroid(2009).

For both stars, the values of radial and rotational velocity found inour study are consistent with previous studies (see Section 3.2). ForHD 53929, since a certain dispersion of the radial velocity existsin comparison to the previous studies, it could indicate that HD53929 is a binary star as predicted by Zentelis (1983). More dataare needed for a firm conclusion.

After studying the abundance variation as a function of opticaldepth for elements with a sufficient number of detected line profiles,we found that phosphorus is vertically stratified in the atmosphere ofHD 53929 and HD 63975 and that it exhibits the same behaviour inboth stars. Its abundance increases strongly towards the superficiallayers. Similar results for the stratification of phosphorus abundancewith optical depth have been reported recently by Catanzaro et al.(2016) for another HgMn star HD49606, although the position of thestratification profile is deeper in that star than those studied here. Thestratification detected for phosphorus indicates that atomic diffusioncould play an important role in the atmosphere of these two stars.

AC K N OW L E D G E M E N T S

Authors are thankful to the Faculte des Etudes Superieures et dela Recherche de l’Universite Moncton for the financial support ofthis research. FL and VK acknowledge support from the Natural

Sciences and Engineering Research Council of Canada. Part of thecalculations have been done on the supercomputer briarree of theUniversity of Montreal, under the guidance of Calcul Quebec andCalcul Canada. The use of this supercomputer is funded by theCanadian Foundation for Innovation (CFI), NanoQuebec, RMGAand Research Fund of Quebec – Nature and Technology (FRQNT).This work has made use of the VALD data base operated at UppsalaUniversity, the Institute of Astronomy RAS in Moscow, and theUniversity of Vienna, as well as the NIST data base.

This work is based on observations obtained at Canada–France–Hawaii Telescope (CFHT), which is operated by the National Re-search Council of Canada, the Institut National des Sciences del’Univers of the Centre National de la Recherche Scientifique ofFrance, and the University of Hawaii. The operations at CFHT areconducted with care and respect from the summit of Maunakea,which is a significant cultural and historic site.

R E F E R E N C E S

Alecian G., 1982, A&A, 107, 61Aller M. F., 1970, A&A, 6, 67Atutov S. N., Shalagin A. M., 1988, Sov. Astron. Lett., 14, 284Castelli F., Hubrig S., 2004, A&A, 421, L1Catanzaro G., Giarrusso M., Leone F., Munari M., Scalia C., Sparacello E.,

Scuderi S., 2016, MNRAS, 460, 1999Cowley C. R., Hubrig S., Gonzalez G. F., Nunez N., 2006, A&A, 455, L21Donati J.-F., 2003, in Trujillo-Bueno J., Sanchez Almeida J., eds, ASP Conf.

Ser., Vol. 307, Solar Polarization. Astron. Soc. Pac., San Francisco,p. 41

Donati J.-F., Semel M., Carter B. D., Rees D. E., Cameron A. C., 1997,MNRAS, 291, 658

Dworetsky M. M., Storey P. J., Jacobs J. M., 1984, Phys. Scr., T8, 39Evans D. C., 1967, in Batten A. H., Heard J. F., eds, Proc. IAU Symp. 30,

Determination of Radial Velocities and their Applications. AcademicPress, London, p. 57

Ghazaryan S., Alecian G., 2016, MNRAS, 460, 1912Gontcharov G. A., 2006, Astron. Lett., 32, 759Grevesse N., Asplund M., Suaval A. J., Scott P., 2010, Ap&SS, 328, 179Grevesse N., Scott P., Asplund M., Sauva A. J., 2015, A&A, 573, A27Hauck B., Mermilliod M., 1980, A&AS, 40, 1Hauck B., Mermilliod M., 1998, A&AS, 129, 431Hauschildt P. H., Baron E., Allard F., 1997, ApJ, 483, 390Heacox W. D., 1979, ApJS, 41, 675Hube D. P., 1970, Mem. RAS, 72, 233Hubrig S., Castelli F., 2001, A&A, 375, 963Hubrig S., Castelli F., Wahlgren G. M., 1999, A&A, 346, 139Hubrig S. et al., 2012, A&A, 547, A90Hui-Bon-Hoa A., LeBlanc F., Hauschildt P. H., 2000, ApJ, 535, L43

MNRAS 477, 3390–3396 (2018)

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022

3396 M. L. Ndiaye, F. LeBlanc and V. Khalack

Khalack V., LeBlanc F., 2015a, Adv. Astron. Space Phys., 5, 3Khalack V., LeBlanc F., 2015b, AJ, 150, 2Khalack V., Poitras P., 2015, in Meynet G., Georgy C., Groh J., Stee P.,

eds, Proc. IAU Symp. 307, New Windows on Massive Stars: Astero-seismology, Interferometry, and Spectropolarimetry. Cambridge Univ.Press, Cambridge, p. 383

Khalack V., Wade G., 2006, A&A, 450, 1157Khalack V., LeBlanc F., Bohlender D., Wade G., Behr B. B., 2007, A&A,

466, 667Khalack V., LeBlanc F., Behr B. B., Wade G. A., Bohlender D., 2008, A&A,

477, 641Khalack V., LeBlanc F., Behr B. B., 2010, MNRAS, 407, 1767Khalack V., Gallant G., Thibeault C., 2017, MNRAS, 471, 926NIST ADS TeamKramida A., Ralchenko Yu., Reader J., 2015, NIST Atomic

Spectra Database (version 5.3). National Institute of Standards and Tech-nology, Gaithersburg, MD, Available at http://physics.nist.gov/asd

Kupka F., Piskunov N., Ryabchikova T. A., Stempels H. C., Weiss W. W.,1999, A&AS, 138, 119

Kupka F., Ryabchikova T. A., Piskunov N. E., Stempels H. C., Weiss W. W.,2000, Balt. Astron., 9, 590

Landstreet J. D., 1988, ApJ, 326, 967LeBlanc F., Michaud G., 1993, ApJ, 408, 251LeBlanc F., Monin D., Hui-Bon-Hoa A., Hauschildt P. H., 2009, A&A, 495,

937LeBlanc F., Hui-Bon-Hoa A., Khalack V., 2010, MNRAS, 409, 1606LeBlanc F., Khalack V., Yameogo B., Thibeault C., Gallant I., 2015,

MNRAS, 453, 3766Leckrone D. S., Wahlgren G. M., Johansson S. G., 1991, ApJ, 377, L37Leckrone D. S., Wahlgren G. M., Johansson S., Adelman S. J., 1993, in

Dworetsky M. M., Castelli F., Faraggiana R., eds, ASP Conf. Ser.Vol. 44, Peculiar Versus Normal Phenomena in A-Type and RelatedStars. Astron. Soc. Pac., San Francisco, p. 42

Mathys G., Hubrig S., 1995, A&A, 293, 810Michaud G., 1970, ApJ, 160, 641Michaud G., Alecian G., Richer J., 2015, Atomic Diffusion in Stars.

Springer, BerlinMoon T. T., Dworetsky M. M., 1985, MNRAS, 217, 305Morel T. et al., 2013, A&A, 561, A35Napiwotzki R., Schonberner D., Wenske V., 1993, A&A, 268, 653Napiwotzki R. et al., 2004, in Hilditch R. W., Hensberge H., Pavlovski K.,

eds, ASP Conf. Ser. Vol. 318, Spectroscopically and Spatially Resolv-ing the Components of the Close Binary Stars. Astron. Soc. Pac., SanFrancisco p. 402

Netopil M., Paunzen E., Maitzen H. M., North P., Hubrig S., 2008, A&A,491, 545

Piskunov N. E., Kupka F., Ryabchikova T. A., Weiss W. W., Jeffery C. S.,1995, A&AS, 112, 525

Press W. H., Teukolsky S. A., Vetterling W. T., Flannery B. P., 1992, Nu-merical Recipes in FORTRAN: The Art of Scientific Computing, 2ndedn. Cambridge Univ. Press, Cambridge

Preston G. W., 1974, ARA&A, 12, 257Renson P., Manfroid J., 2009, A&A, 498, 961Royer F., Grenier S., Baylac M.-O., Gomez A. E., Zorec J., 2002, A&A,

393, 897Ryabchikova T. A., Piskunov N. E., Kupka F., Weiss W. W., 1997, Balt.

Astron., 6, 244Ryabchikova T., Kotchoukhov O., Galazutdinov G., Musaev F., Adelman S.

J., 1998, Contrib. Astron. Obs. Skalnate Pleso, 27, 258Ryabchikova T., Piskunov N., Kurucz R. L., Stempels H. C., Heiter U.,

Pakhomov Yu., Barklem P. S., 2015, Phys. Scr., 90, 054005Savanov I., Hubrig S., 2003, A&A, 410, 299Scott P. et al., 2015a, A&A, 573, A25Scott P., Asplund M., Grevesse N., Bergemann M., Sauva A. J., 2015b,

A&A, 573, A26Sigut T. A. A., 2001, A&A, 377, 27Smith K. C., 1993, A&A, 276, 393Smith K. C., 1994, A&A, 291, 521Smith K. C., 1996a, Ap&SS, 237, 77Smith K. C., 1996b, A&A, 305, 902Smith K. C., 1997, A&A, 319, 928Smith K. C., Dworetsky M. M., 1993, A&A, 274, 335Takada-Hidai M., 1991, in Michaud G., Tutukov A., eds, Proc. IAU Symp.

145, Evolution of Stars: The Photospheric Abundance Connection.Kluwer, Dordrecht, p. 137

Takeda Y., Honda S., 2016, PASJ, 68, 32Thiam M., LeBlanc F., Khalack V., Wade G. A., 2010, MNRAS, 405, 1384White R. E., Vaughan A. H., Jr, Preston G. W., Swings J. P., 1976, ApJ, 204,

131Williams T. Kelley C., 1986, Gnuplot Version 5.0: An Interactive Plotting

Program. Available at http://gnuplot.infoWilson R. E., 1953, General Catalogue of Stellar Radial Velocities. Carnegie

Institution of Washington, WashingtonWolff S. C., Preston G. W., 1978, ApJS, 37, 371Woolf V. M., Lambert D. L., 1999, ApJ, 521, 414Zentelis N., 1983, A&AS, 53, 445

S U P P O RT I N G IN F O R M AT I O N

Supplementary data are available at MNRAS online.

Supplementary_data.tar.gz

Please note: Oxford University Press is not responsible for thecontent or functionality of any supporting materials supplied bythe authors. Any queries (other than missing material) should bedirected to the corresponding author for the article.

This paper has been typeset from a TEX/LATEX file prepared by the author.

MNRAS 477, 3390–3396 (2018)

Dow

nloaded from https://academ

ic.oup.com/m

nras/article/477/3/3390/4937812 by guest on 25 August 2022