Discovery of new roAp pulsators in the UVES survey of cool magnetic Ap stars
O. Kochukhov, D. Alentiev, T. Ryabchikova, S. Boyko, M. Cunha, V. Tsymbal, W. Weiss
aa r X i v : . [ a s t r o - ph . S R ] F e b Mon. Not. R. Astron. Soc. , 1–13 (2012) Printed 8 August 2018 (MN L A TEX style file v2.2)
Discovery of new roAp pulsators in the UVES survey of coolmagnetic Ap stars ⋆ O. Kochukhov , D. Alentiev , , T. Ryabchikova , S. Boyko , M. Cunha , V. Tsymbal ,W. Weiss Department of Physics and Astronomy, Uppsala University Box 516, 751 20 Uppsala, Sweden Department of Physics, Tavrian National University, Vernadskiy’s Avenue 4, 95007 Simferopol, Ukraine Centro de Astrofisica da Universidade do Porto, Rua das Estrelas, 4150-762 Porto, Portugal Institute of Astronomy, Russian Academy of Sciences, Pyatnitskaya 48, 119017 Moscow, Russia Department of Physics, M.V.Lomonosov Moscow State University, GSP-1, 1-2 Leninskye Gory, 119991 Moscow, Russia Department of Astronomy, University of Vienna, T¨urkenschanzstrasse 17, 1180 Wien, Austria
Accepted 2013 February 26. Received 2013 February 7; in original form 2012 December 19
ABSTRACT
We have carried out a survey of short-period pulsations among a sample of carefully chosencool Ap stars using time-resolved observations with the UVES spectrometer at the ESO 8-mVLT telescope. Here we report the discovery of pulsations with amplitudes 50–100 m s − and periods 7–12 min in HD 132205, HD 148593 and HD 151860. These objects are thereforeestablished as new rapidly oscillating Ap (roAp) stars. In addition, we independently confirmthe presence of pulsations in HD 69013, HD 96237 and HD 143487 and detect, for the firsttime, radial velocity oscillations in two previously known photometric roAp stars HD 119027and HD 185256. At the same time, no pulsation variability is found for HD 5823, HD 178892and HD 185204. All of the newly discovered roAp stars were previously classified as non-pulsating based on the low-precision ground-based photometric surveys. This shows that suchobservations cannot be used to reliably distinguish between pulsating and non-pulsating starsand that all cool Ap stars may harbor p -mode pulsations of different amplitudes. Key words: stars: chemically peculiar – stars: magnetic fields – stars: oscillations – stars: in-dividual: HD 5823, HD 69013, HD 96237, HD 132205, HD 143487, HD 148593, HD 151860,HD 178892, HD 185204
Rapidly oscillating Ap (roAp) stars are unique astrophysical lab-oratories allowing the study of the effects that strong organisedmagnetic fields have on stellar rotation, convection, pulsations andchemical element transport in the stellar interiors and atmospheres.These stars belong to the group of chemically peculiar, magneticlate-A and early-F objects, commonly known as SrCrEu Ap stars.The roAp stars exhibit high-overtone, non-radial p- mode pulsationswith periods around 10 min and low amplitudes both in photometryand spectroscopy (Kurtz & Martinez 2000; Kochukhov 2008). Thepresence of multi-periodic pulsations in many roAp stars makesthem interesting targets for a classical asteroseismic analysis con-cerned with determining global stellar properties (e.g., Saio et al.2010). In addition, spectroscopic observations of the variabilityin rare-earth spectral lines formed in the outer atmospheric layersof these stars offer unique possibilities for tomographic mappingof the vertical structure of pulsation modes (Ryabchikova et al. ⋆ Based on observations collected at the European Southern Observatory,Paranal, Chile (program 085.D-0124). p- mode oscillations in magnetic Ap stars. The most plausible the-ory (Balmforth et al. 2001) predicts that suppression of convectionin the outer stellar layers allows excitation of the roAp pulsationsdue to the κ -mechanism operating in the hydrogen ionization zone.Since the driving of the oscillations results from a subtle energy bal-ance that depends directly on the interaction between the magneticfield, convection, pulsations, and atomic diffusion, pulsational anal-ysis provides a unique tool for studying these physical processesand their coupling.Modern theoretical pulsation models are fairly successful inmatching the observed pulsation frequencies (Kurtz et al. 2002;Mkrtichian et al. 2008) and even in assessing a complex geom-etry of pulsation modes distorted by rotation and magnetic field(Saio & Gautschy 2004; Kochukhov 2004; Bigot & Kurtz 2011),but are less predictive when it comes to explaining distribution ofthe roAp stars in the H-R diagram. Compared to observations, exci-tation models predict pulsations in systematically hotter and more c (cid:13) O. Kochukhov et al. luminous Ap stars (Cunha 2002; Th´eado et al. 2009). In fact, a sig-nificant fraction of more than 40 currently known roAp stars arelocated beyond the red boundary of the theoretical instability strip.A related observational difficulty is the co-existence of pulsat-ing and apparently constant Ap stars in the same region of the H-Rdiagram. The separation between the roAp and non-pulsating Ap(noAp) stars is largely reliant on the historic ground-based surveys(e.g., Nelson & Kreidl 1993; Martinez & Kurtz 1994), which mightnot be sensitive enough to reveal low-amplitude photometric vari-ability. In fact, observations with the Kepler satellite have shownthat in some roAp stars the amplitude of the oscillations does notexceed a few tens of µ mag (Balona et al. 2011). Clearly such roApstars would have been identified as constant from ground-based ob-servations.Recent spectroscopic detections of pulsations in some of theprototypical “photometric noAp” stars (Hatzes & Mkrtichian 2004;Elkin et al. 2005b; Kochukhov et al. 2009) demonstrated the clearadvantages of the spectroscopic observations over ground-basedphotometry in discovering and characterizing pulsations in cool Apstars. In particular, high-resolution spectroscopy allows one to iso-late rare-earth lines, which often show 10–100 times higher pulsa-tional amplitudes than the lines of light and iron-peak elements.Aiming to establish an unbiased incidence of rapid oscilla-tions among cool Ap stars, we have carried out a survey of pulsa-tions in a small sample of roAp-candidates using the most powerfulinstrumentation currently available for high-resolution stellar spec-troscopy. Our observations, performed with the UVES spectrome-ter at the ESO VLT telescope, turned out to be remarkably success-ful as we were able to demonstrate the presence of pulsations in 9out of 12 observed stars. The discovery of the longest-period roApstar, HD 177765, was reported in a separate paper (Alentiev et al.2012). Here we present further discoveries of new roAp stars, con-firmations of previously known ones and report the sensitive upperlimits on the radial velocity pulsations for a few objects in whichwe could not detect variability.This paper is structured as follows. Target selection is dis-cussed in Sect. 2.1. Observations and data reduction are outlined inSect. 2.2. Details of the methods employed for radial velocity anal-ysis and atmospheric parameter determination of the target stars aregiven in Sect. 2.3 and 2.4 respectively. Results for individual starsare presented in Sect. 3. The outcome of our survey is summarisedand discussed in the context of other recent studies of the roAp starsin Sect. 4. The primary list of roAp candidates was drawn from the catalogueby Renson & Manfroid (2009), taking into account a number ofrecent publications on individual Ap stars. We have selected Apstars falling below the T eff = 8000 K threshold beyond which fewroAp stars are observed. We have also given preference to observeobjects with reasonably sharp spectral lines, for which best accu-racy in radial velocity measurements can be expected. ESO archivalspectra, such as those described by Freyhammer et al. (2008), wereextensively used to confirm the cool Ap-star nature of the tar-gets. We found that many of the late-A objects classified as, e.g.“Ap Sr” by Renson & Manfroid (2009) and included in previousroAp photometric surveys, are either Am stars or very rapid ro-tators for which spectral classification is ambiguous. Finally, we compiled a list of 14 stars, 12 of which were eventually observedduring ESO Period 85. This sample included two known photo-metric roAp stars, HD 119027 and HD 185256, for which no time-resolved spectroscopic studies have been previously carried out. The candidate and known roAp stars were observed in the periodfrom April to July 2010 using the Ultraviolet and Visual EchelleSpectrograph (UVES) installed at one of the ESO 8-m VLT tele-scopes. The spectrograph was configured to use the 600 nm redsetting with an image slicer. This setup provided resolution R ≈
110 000 and the wavelength coverage from 4980 ˚A to 7010 ˚A witha 100 ˚A gap in the region centered at 5990 ˚A.Each stellar observation consisted of 50–67 exposures. Indi-vidual exposure time varied between 60 and 90 s, depending on thestellar brightness. To optimise these time-series observations weemployed the ultra-fast (4-port, 625 kpix s − ) readout mode of theUVES CCDs, which allowed us to reduce the overhead betweenconsecutive exposures to 21 s.In total, we obtained 716 spectra with a signal-to-noise ratio of40–160 over 10 observing nights. The known roAp star HD 119027was observed on two occasions, with 50 exposures each. For an-other target, HD 185204, an incomplete time-series of 38 exposureswas obtained in addition to the standard 50-exposure sequence.We analysed these multiple datasets available for HD 119027 andHD 185204 individually.Reduction of the echelle spectra was carried out with an im-proved version of our UVES pipeline (Alentiev et al. 2012). Thiscode preforms common reduction steps, such as bias and scatteredlight subtraction, order position determination, extraction of one-dimensional spectra, wavelength calibration and continuum nor-malization. A barycentric radial velocity corrections was taken intoaccount in the final reduction step.Detailed information on the adopted exposure times, the num-ber of spectra obtained, the Julian date of the start and end of ob-servations, and typical signal-to-noise ratios is given in Table 1. Reduced, one-dimensional spectra were used to determine radialvelocity (RV) variation in individual lines and groups of lines. Theatomic data necessary for the identification of spectral lines wereobtained from the VALD database (Kupka et al. 1999) and adoptedfrom our previous studies of the roAp stars (e.g., Ryabchikova et al.2007b). We made an effort to select only spectral lines not signifi-cantly affected by blends.The RV analysis started with determinations of the linecenters using the center-of-gravity technique described byKochukhov & Ryabchikova (2001). With our data quality, individ-ual spectral lines seldomly provided accurate enough RVs for anunambiguous detection of pulsations. Therefore, we constructedaverage RV curves for all suitable lines of a given ion by remov-ing occasional linear trends from the RV data of individual linesand averaging the resulting curves. Then, discrete Fourier trans-form was used to obtain an amplitude spectrum and derive an ini-tial guess for the pulsation period from the position of the highestpeak. The False Alarm Probability (FAP) of this variable signal wasestimated as
F AP = 1 − [1 − exp ( − z )] N , where z is the heightof the peak in the variance-normalised Lomb-Scargle periodogram(Scargle 1982) and the number of independent frequencies N was c (cid:13) , 1–13 ew roAp stars Figure 1.
Average spectra of program stars in the 6140–6164 ˚A region. For completeness we also show the mean spectrum of HD 177765 studied byAlentiev et al. (2012). Identification is provided for most important spectral features.
Table 1.
The journal of UVES observations of roAp stars and roAp candidates. The columns give the star name, the magnitude in Johnson V band, the numberof spectra, individual exposure times, the start and end heliocentric Julian dates of time-series observations, and the typical signal-to noise ratio. Superscriptsdenote data sets obtained on different observing nights for HD 119027 and HD 185204.Star V N T exp (s) HJD start − HJD end − S/N
HD 5823 9.98 50 80 372.88456 372.94191 60–100HD 69013 9.56 62 60 300.46474 300.52203 40–70HD 96237 9.45 50 60 300.53081 300.57686 70–115HD 119027 computed according to the prescription given by Horne & Baliunas(1986). The mean pulsation period was determined by averagingperiod estimates for all ions with definite detection of variability(FAP − ). In the final step, RV curves of individual ions werefitted with a linear least-squares algorithm to obtain an estimate ofthe pulsation phase and amplitude.Fitting a cosine curve to the RV data allows us to obtain aproxy of the relative formation height of variable spectral lines.As shown by Ryabchikova et al. (2007a), many roAp stars exhibita regular progression in the phase of pulsation maximum from oneion to the next. This picture reflects the outward propagation of pul-sation waves through the chemically stratified stellar atmosphere.As a first approximation, one can assume that a smaller phase inthe cosine function corresponds to a later pulsation maximum and hence to the line formation in the higher atmospheric layers andvice versa.In addition to looking at the lines of rare-earth ions typicallyshowing the largest amplitudes for the roAp stars, we also carriedout a frequency analysis of mean RV curves of a non-variable ion,typically Feı, to assess intrinsic stability of the spectrograph in thecourse of our observations and detect possible spurious variabilitydue to instrumental artifacts. Effective temperatures of the target stars were estimated using theStr¨omgren photometric data from Martinez (1993) and applyingcalibrations by Moon & Dworetsky (1985) and Napiwotzki et al. c (cid:13) , 1–13 O. Kochukhov et al.
Table 2.
Basic parameters of the program stars. The columns give the starname, spectral classification according to Renson & Manfroid (2009) andestimates of T eff , log g , v e sin i and mean magnetic field modulus h B i .Asterisk marks magnetic field modulus determined with synthetic spectrummodeling.Star Spectral T eff log g v e sin i h B i type (K) (km s − ) (kG)HD 5823 F2 SrEuCr 7300 4.3 13.5 8.5 ∗ HD 69013 A2 SrEu 7600 4.5 4.0 4.8HD 96237 A4 SrEuCr 7800 4.3 6.0 2.9 ∗ HD 119027 A3 SrEu 7050 4.4 4.0 3.1HD 132205 A2 EuSrCr 7800 4.4 9.5 5.2 ∗ HD 143487 A3 SrEuCr 7000 5.0 1.5 4.7HD 148593 A2 Sr 7850 4.4 5.0 3.0 ∗ HD 151860 A2 SrEu 7050 4.5 4.5 2.5HD 178892 Ap SrCrEu 7700 4.0 10.0 18.5 ∗ HD 185204 A2 SrEuCr 7750 4.4 4.5 5.4HD 185256 F0 SrEu 7150 4.3 5.5 ∗ (1993) as implemented in the TEMPLOGG code (Kaiser 2006). ForHD 178892, which lacks the Str¨omgren photometry, we adoptedatmospheric parameters from Ryabchikova et al. (2006).Measurements of the mean magnetic field modulus were ac-complished by Gaussian fitting the resolved components of theZeeman split Fe II YNTH M AG polarised spectrumsynthesis code (Kochukhov 2007) to model magnetic broadeningof magnetically sensitive spectral lines. The latter procedure wasalso necessary when the lines were dominated by rotational broad-ening. The accuracy of the magnetic field estimated by fitting well-resolved Zeeman split lines is about 0.1 kG. On the other hand, thespectrum synthesis determination of the magnetic field modulus isaccurate to within ∼ YNTH M AG was also used to deter-mine v e sin i from the magnetically insensitive lines, such as Feı5434.52 ˚A.Table 2 summarises basic parameters of all program stars.In this table we provide an average T eff obtained with the twoStr¨omgren photometric calibrations and the surface gravity ob-tained with the calibration by Moon & Dworetsky (1985). It isknown that photometric determination of the latter parameter of-ten yields an overestimated surface gravity. Table 2 also providesan estimate of the projected rotational velocity and mean field mod-ulus using the methods described above.Time-averaged spectra for all program stars, includingHD 177765 (Alentiev et al. 2012), are presented in Fig. 1. Most ob-jects are slow rotators, exhibiting rich spectra and, in particular, aprominent Nd III II Here we report results of the search of short-period oscillations inthe target stars. The following sections present details on the anal-ysis of the newly discovered roAp stars HD 132205, HD 148593and HD 151860 (Sect. 3.1), the roAp stars known from previousphotometric observations HD 119027 and HD 185256 (Sect. 3.2),and confirmation of recent spectroscopic detections of pulsationsin HD 69013, HD 96237, HD 143487 (Sect. 3.3). Null results for
Table 3.
Results of the frequency analysis for the newly discovered roApstars. The columns give the stellar name, ion identification, the number ofstudied spectral lines and the amplitude A and phase ϕ derived by the least-squares cosine fit. The phase is given as a fraction of the averaged pulsationperiod indicated in the table. The false alarm probability for the variabilitydetected in each ion is given in the last column.Star Ion N A (m s − ) ϕ FAP P = 7 . ± . minHD 132205 Ce II
10 95.7 ± ± III ± ± III ± ± II
13 45.8 ± ± II ± ± III ± ± III ± ± α ± ± II
20 37.8 ± ± II ± ± P = 10 . ± . minHD 148593 Nd II
36 45.3 ± ± II
16 47.2 ± ± III
12 23.8 ± ± II
23 27.8 ± ± P = 12 . ± . minHD 151860 Tb III ± ± II ± ± II
25 23.5 ± ± III ± ± HD 5823, HD 178892 and HD 185204 are presented in Sect. 3.4.The outcome of the frequency analysis of individual ions is pre-sented in Tables 3–5.
The star HD 132205 is classified as A2 EuSrCr in the catalogueby Renson & Manfroid (2009). Martinez & Kurtz (1994) observedthis star with time-resolved photometry in Johnson B filter duringone hour on a single night, finding no pulsational variability above0.5 mmag in the typical roAp frequency range. No other reports onthis object are available in the literature.We have observed HD 132205 with a time series of 50 UVESspectra. Analysis of these data immediately showed RV variabilityin many rare-earth element (REE) ions and in the H α core. Thisobservation establishes HD 132205 as a new roAp star. Fig. 2 il-lustrates the amplitude spectra of Ce II , Nd II and H α . This figurealso shows the absence of significant variability in the Feı lines,confirming that oscillations detected in the REE lines are not dueto an instrumental artifact. The amplitudes and phases of the meanRV curves of different ions are reported in Table 3. Using mea-surements derived from the Ce II , Gd II , Nd III and Pr
III lines, weestimated an average pulsation period of P = 7 . ± . min.The amplitude of oscillations does not exceed 100 m s − for allions. Pulsation phases of different groups of variable spectral linesdiffer marginally.We determined v e sin i = 9.5 km s − from a spectrum synthe-sis fit to the magnetically insensitive spectral lines. The projectedrotational velocity of HD 132205 is too large to directly detect mag- c (cid:13) , 1–13 ew roAp stars A m p li ude , m s - Ce II
Nd III
H core , mHzFe I
Figure 2.
Amplitude spectra of Ce II , Nd III , H α core and Feı forHD 132205. netically split spectral lines in its spectrum. Nevertheless, spectrumsynthesis analysis of the Feı 6336.8 ˚A line gives evidence of a fairlystrong magnetic field with h B i = 5.2 kG. HD 148593 is another cool Ap star for which practically no in-formation is available in the literature. Renson & Manfroid (2009)give A2 Sr spectral classification for this object. Martinez & Kurtz(1994) failed to detect pulsational variability with 0.65 h photo-metric monitoring on a single night. Wraight et al. (2012) found norotational photometric modulation based on observations with theSTEREO satellites.Our UVES observations of this star consisted of 50 high-quality spectra. Radial velocity measurements show convincing ev-idence of relatively weak pulsations in several REE ions. Thus,HD 148593 is definitely a roAp star. The most significant variabilityoccurs in Nd II , Sm II , Nd III and Gd II . Using the two former ionswe established a mean period of P = 10 . ± . min. Theamplitude spectra of Nd II and Sm II are compared with the mea-surements for constant Feı lines in Fig. 3. The complex shape of theREE amplitude spectra, especially for Sm II , suggests the presenceof multiperiodic oscillations which are not resolved by our shorttime series data. The characteristics of the mean RV curves are re-ported in Table 3. All RV amplitudes are within 20–50 m s − range.We do not find a significant difference in the pulsation phases forvariable spectral lines.We used the magnetically insensitive Feı line to estimate A m p li ude , m s - Nd II
Sm II , mHz
Fe I
Figure 3.
Amplitude spectra of Nd II , Sm II and Feı for HD 148593. v e sin i = 5 km s − . There are no resolved Zeeman split lines inthe spectrum of HD 148593, but the magnetic field is likely to bepresent. We estimated h B i = 3.0 kG from the spectrum synthesis fitusing Fe lines with large Land´e factors. Similar to the two previous objects, not much is known about theA2 SrEu star (Renson & Manfroid 2009) HD 151860. It was ob-served in several photometric surveys of Ap stars (Hauck & North1982; Maitzen & Vogt 1983; Maitzen et al. 2000) and was inves-tigated for pulsations by Martinez & Kurtz (1994). These authorscould not find photometric variability stronger than about 0.5 mmagfrom the 0.73 h observing run on a single night and hence classifiedthis star as noAp.Our set of 50 UVES spectra of HD 151860 shows definite pul-sational variability in several REE ions, establishing this object asa new roAp star. The amplitude spectra of La II , Tb III and Eu II arepresented in Fig. 4. In comparison, Feı lines show no significant os-cillations. The highest amplitude peak occurs around 1.35 mHz forall REE ions. At the same time, several other peaks are present atlower frequencies. It is very likely that this object is a multiperiodicroAp star. An average pulsation period of P = 12 . ± . minwas established from the lines of La II , Tb III and Eu II . All RVamplitudes are below 50 m s − except Tb III , for which oscilla-tions reach 84 m s − . Pulsational phases differ significantly be-tween REE ions. The mean RV curve of Dy III shows the mostdeviating behavior, lagging by ∼ v e sin i = 4.5 km s − was de-termined from the Feı 5434.5 ˚A line. Spectrum synthesis analy-sis of the magnetically sensitive spectral lines indicates a moderate h B i of around 2.5 kG. c (cid:13) , 1–13 O. Kochukhov et al. A m p li ude , m s - La II
Tb III
Eu II , mHzFe I
Figure 4.
Amplitude spectra of La II , Tb III , Eu II and Feı for HD 151860. Photometric variability in A3 SrEu (Renson & Manfroid 2009) starHD 119027 was discovered by Martinez et al. (1993). They re-ported five frequencies with a spacing of 26 µ Hz clustered around8.8 min pulsation period. A follow-up study by Martinez et al.(1998a) found evidence for two additional frequencies but couldnot unambiguously confirm the frequency spacing established inthe previous study. Instead, the authors suggested that a more likelyfrequency spacing is 52 µ Hz. Using mean light observations ob-tained during 45 individual nights, Martinez et al. (1998b) foundno evidence of the rotational photometric variability of HD 119027.They suggested that the rotational period may exceed 6 monthsor that the star is visible pole-on. Mathys et al. (1997) discoveredresolved magnetically split Fe II h B i on the time scale ofseveral weeks.We obtained two UVES data sets for HD 119027, each con-sisting of 50 spectra. Observations were separated by 11 nights. Inboth data sets a clear RV variation of the REE lines was found witha period similar to the previously known photometric one. This isthe first report of the pulsational RV variability in this roAp star.The most secure identification of pulsations is found for Nd III ,Nd II , Ce II , Pr III and Sm II . For all these ions pulsation amplitudesdo not exceed 140 m s − . The amplitude spectra of Nd III , Nd II and Ce II are illustrated in Fig. 5 and Fig. 6 for the first and seconddata set respectively. The corresponding mean pulsation periods are Table 4.
Results of the frequency analysis for the previously known pho-tometric roAp stars. The columns are the same as in Table 3. Superscriptsdenote data sets obtained on different observing nights for HD 119027.Star Ion N A (m s − ) ϕ FAP P = 8 . ± . minHD 119027 Nd II
37 100.8 ± ± II
30 97.6 ± ± III
16 70.0 ± ± III ± ± II ± ± II
28 33.6 ± ± III ± ± II
11 74.4 ± ± P = 8 . ± . minHD 119027 Nd III
16 97.8 ± ± II
32 118.7 ± ± II
31 74.7 ± ± III ± ± II ± ± III ± ± II
10 66.1 ± ± II
26 29.7 ± ± P = 10 . ± . minHD 185256 Ce II ± ± III ± ± II ± ± III ± ± III ± ± III ± ± α ± ± II ± ± P = 8 . ± . and P = 8 . ± . min. The discrepancybetween the period determinations for the two nights and a slightlydifferent shape of the amplitude spectra are evident in Figs. 5 and 6.This is likely explained by the beating of close frequencies, whichare not resolved in our short time series observations. HD 119027shows a large range of pulsation phases for different REE ions. Es-pecially interesting is the behavior of Pr III , which exhibits a largephase lag relative to other ions. This may be an indication of pul-sation node, similar to the one found in 33 Lib (Kurtz et al. 2005)and 10 Aql (Sachkov et al. 2008).For both observing nights we measured an identical field mod-ulus of h B i = 3.14 kG. There is no evidence of magnetic or spec-troscopic variability between our two mean spectra. The projectedrotational velocity of the star is estimated to be v e sin i = 4 km s − . HD 185256 is the second known roAp star in our sample. Classifiedas F0 SrEu (Renson & Manfroid 2009), it was discovered as pulsat-ing by Kurtz & Martinez (1995). Based on the photometric moni-toring in the Johnson B filter, these authors found pulsations with aperiod of 10.2 min and an amplitude of 1.6 mmag. Apart from thebrief communication by Kurtz & Martinez (1995), no pulsationalanalysis of HD 185256 has ever been published. The longitudinalmagnetic field of h B z i = − ± G was found in this star byHubrig et al. (2004).Analysis of our UVES spectra of HD 185256 showed clearrapid RV variations in REE lines (Fig. 7). At the same time, the c (cid:13) , 1–13 ew roAp stars A m p li ude , m s - Nd III
Nd II
Ce II , mHzFe I
Figure 5.
Amplitude spectra of Nd
III , Nd II , Ce II and Feı for HD 119027on the first observing night (data set 1). maximum pulsation amplitude in the Feı lines does not exceed15 m s − . A marginal pulsation signal was detected in the H α coreand in Nd II lines. Radial velocity pulsations in HD 185256 aretypical of most roAp stars. The average period P = 10 . ± . min was determined from the Ce II , Pr III , Nd
III , Sm II ,Tb III , and Dy
III lines. The amplitudes of the oscillations vary be-tween 80 and 250 m s − . The maximum amplitude is found forDy III and Tb
III lines. Pulsational phase shifts are also typical ofthe roAp stars, with the maximum RV first observed in singly-ionised REEs and H α core, followed by doubly-ionised REEs. Theamplitude-phase relation for HD 185256 is similar to the one ob-served for another roAp star, 10 Aql (Sachkov et al. 2008).The projected rotational velocity of HD 185256, v e sin i = 5.5 km s − , was estimated from the Feı λ h B i II λ The presence of oscillations in the A2 SrEu (Renson & Manfroid2009) star HD 69013 was tested by Nelson & Kreidl (1993) andMartinez & Kurtz (1994). Both studies failed to detect pulsationsstronger than 0.6–1 mmag based on a few 1–2 h-long photometrictime series. The high-resolution spectrum of HD 69013 obtained by A m p li ude , m s - Nd III
Nd II
Ce II , mHzFe I
Figure 6.
Amplitude spectra of Nd
III , Nd II , Ce II and Feı for HD 119027on the second observing night (data set 2). Freyhammer et al. (2008) showed resolved Zeeman split lines cor-responding to h B i = 4.8 kG. The spectral appearance of HD 69013is typical of a roAp star. Using two short UVES spectroscopic timeseries data sets Elkin et al. (2011) found a RV variability with am-plitudes of up to 200 m s − in individual REE spectral lines and aperiod of 11.4 min. Follow-up photometric observations, also pub-lished by Elkin et al. (2011), suggested the presence of light vari-ability with a similar period.Our set of 62 spectra of HD 69013 reveals unambiguous ev-idence of pulsations in several REE ions (Fig. 8). The RV ampli-tudes are typically 100 m s − or less. The most significant oscil-lations signals are found for the lines of Nd II , Nd III and Pr
III .The signal-to-noise ratio of the corresponding peaks in the ampli-tude spectra is much higher than in the study by Elkin et al. (2011).Using the mean RV curves of these REE ions, we determined theaverage pulsation period of P = 11 . ± . min. The RVvariations of these ions show a non-negligible phase shift. The RVmaximum is first reached for Nd II , followed by Nd III and thenPr
III . This order presumably reflects the relative formation heightsof the lines of these ions.Using a magnetically insensitive Feı line we measure v e sin i = 3.5 km s − . On the other hand, the splitting of Fe II h B i = 4.8 kG. Bothparameters are in excellent agreement with the results ofFreyhammer et al. (2008). c (cid:13) , 1–13 O. Kochukhov et al. A m p li ude , m s - Ce II
Nd III
Dy III , mHzFe I
Figure 7.
Amplitude spectra of Ce II , Nd III , Dy
III and Feı for HD 185256.
A photometric search of pulsations in this A4 SrEuCr star(Renson & Manfroid 2009) was carried out by Nelson & Kreidl(1993). Based on 2-hour observations, they established an up-per limit of 0.5 mmag on possible rapid oscillations. Hubrig et al.(2000) treated this object as a non-pulsating star in their studyof the H-R diagram position and kinematics of cool Ap stars.Freyhammer et al. (2008) detected Zeeman split lines indicatingmean magnetic field h B i = 2–3 kG and noted an unusually strongspectral variability in this star. They determined a rotation periodof 20.91 d from the archival ground-based and space photometry.From a set of 34 time-resolved UVES spectra Elkin et al. (2011)found a marginal evidence of the RV variability with a period of ≈
14 min and amplitudes up to 100 m s − in REE spectral lines.Our 50 UVES spectra of HD 96237 show very clear pulsa-tional signatures in different REE lines and in the core of H α (Fig. 9). The RV amplitudes exceed 100 m s − for many REEions and reach 290 m s − in the core of H α . The mean period, P = 13 . ± . min, inferred from the measurements of Ce II ,Tm II , Nd II , Nd III , Tb
III and Dy
III lines, is consistent with theresults of Elkin et al. (2011), yet our data yields detection of oscil-lations with 2–3 times higher signal-to-noise ratio than in the pre-vious study. The least-squares analysis reveals a significant differ-ence in phases of the RV curves of different groups of lines. Amongthe species with FAP − a larger phase (which corresponds toan earlier pulsation maximum) is found for H α and singly ionisedREEs. On the other hand, Nd III and Tb
III show a later maximum,suggesting formation of these lines in higher atmospheric layers.
Table 5.
Results of the frequency analysis for the confirmed new roAp stars.The columns are the same as in Table 3.Star Ion N A (m s − ) ϕ FAP P = 11 . ± . minHD 69013 Pr III ± ± III
13 97.3 ± ± II
19 118.9 ± ± II ± ± II ± ± P = 13 . ± . minHD 96237 Nd III
19 142.0 ± ± III ± ± II ± ± II
25 61.8 ± ± II ± ± II
10 91.1 ± ± α ± ± III ± ± III ± ± II ± ± II ± ± P = 9 . ± . minHD 143487 Nd II
30 29.8 ± ± III ± ± III
12 25.4 ± ± α ± ± II
10 22.18 ± ± II ± ± Using spectrum synthesis modeling we found v e sin i = 6 km s − and h B i = 2.9 kG. Both parameters agreewith the findings by Freyhammer et al. (2008). This object is the coolest star in our sample. It is classified as A2SrEuCr (Renson & Manfroid 2009), but has a T eff around 7000 Kaccording to our analysis. Martinez & Kurtz (1994) could not findp-mode pulsations in this star stronger than ≈ − and periods 8.8–10.0 min.This short time-series data did not allow Elkin et al. (2010) to de-termine the main pulsation frequency with any more precision andto look for other modes.Our UVES data set for HD 143487 comprises 62 spectra col-lected over a time span of 1.4 h. Analysis of these spectra yieldsa highly significant detection of pulsations in the lines of Nd II ,Nd III , Pr
III , and in the core of H α . Metal lines show amplitudes of20–30 m s − , while pulsations reach 100 m s − in H α . The averageperiod, P = 9 . ± . min, was determined from the lines ofNd II , Nd III and Pr
III . Within the uncertainties of our analysis allgroups of lines show the same pulsation phase.The representative amplitude spectra for HD 143487 are c (cid:13) , 1–13 ew roAp stars A m p li ude , m s - Nd II
Nd III
Pr III , mHzFe I
Figure 8.
Amplitude spectra of Nd II , Nd III , Pr
III and Feı for HD 69013. shown in Fig. 10. The main pulsation peak is detected with a highersignal-to-noise ratio than in the study by Elkin et al. (2010). Thisfigure also reveals a significant amplitude excess at low frequen-cies in all variable ions and even in the average RV curve of Feı.The corresponding variability with an amplitude of ≈
20 m s − anda period of ≈
42 min is not highly significant for any given ion butis reproduced for each group of lines. On the other hand, it is alsopresent, with a somewhat lower amplitude, in the RV measurementsof sharp telluric features around λ = 6300 and 6900 ˚A. Therefore,we tentatively attribute this low-frequency signal to an instrumen-tal effect. HD 143487 is the only star in our sample showing thisartifact.Using several REE lines with the triplet-like Zeeman splittingpatterns, we determined h B i = . ± . kG, which noticeablyexceeds h B i = 4.2–4.3 kG found by Freyhammer et al. (2008). Wenote that the Fe II h B i . On the other hand, our v e sin i = 1.5 km s − is inreasonable agreement with their estimate of 2 km s − . HD 5823 is classified as a F2 SrEuCr star (Renson & Manfroid2009) and falls in the temperature range occupied by the roAp stars.Pulsations in this object have been repeatedly searched for in photo-metric ground-based surveys. But neither Nelson & Kreidl (1993) A m p li ude , m s - Nd III
Tb III
H core , mHzFe I
Figure 9.
Amplitude spectra of Nd
III , Tb
III , H α core and Feı forHD 96237. nor Martinez & Kurtz (1994) could detect pulsational variability.The latter authors observed HD 5823 on three different nights, find-ing an upper photometric variability limit of 0.5–1.0 mmag.As evident from Fig. 1, HD 5823 shows relatively broad spec-tral lines compared to other stars in our sample. We measured v e sin i = 13.5 km s − from the magneticaly insensitive Feı line.Interestingly, this star also shows evidence of a fairly strong mag-netic field. We find h B i = 8.5 kG from the spectrum synthesis cal-culations. The Fe II III Nd II and Ce II are shown in Fig. 11. The upperlimit of RV oscillations is 100–200 m s − , which does not mean-ingfully constrain the roAp nature of this star. Weak pulsations,similar to those found for HD 132205, HD 148593, HD 151860,could have been easily missed in our data for HD 5823. This Ap SrCrEu star possesses one of the strongest magneticfields among cool Ap stars. A strong longitudinal magnetic field,reaching 7.5 kG in maximum, was measured for HD 178892 byKudryavtsev et al. (2006). Ryabchikova et al. (2006) presented adetailed magnetic field, model atmosphere and abundance analy-sis of this star based on high-resolution spectra. They measured h B i = 17.1–18.0 kG from the resolved magnetically split spectral c (cid:13) , 1–13 O. Kochukhov et al. A m p li ude , m s - Nd III
Pr III
H core , mHzFe I
Figure 10.
Amplitude spectra of Nd
III , Pr
III , H α core and Feı forHD 143487. lines and determined a rotational period of 8.2478 d from the ASASphotometry. A dipolar model fit suggested an inclination angle i = 37 ◦ and a polar field strength B p ≈
23 kG. Abundance anal-ysis of this star performed by Ryabchikova et al. (2006) revealed atypical roAp pattern. Based on these results, the authors suggestedthis star as an interesting target for the search of rapid oscillations.However, prior to our study, no time-resolved photometric or spec-troscopic monitoring of this star was ever carried out.We have acquired 67 time-resolved UVES spectra forHD 178892. The analysis of rapid spectral line variability is com-plicated by a strong magnetic splitting of the majority of the lines.We have focused our RV measurements on the lines with smallLand´e factors and on the spectral features showing well-resolvedZeeman components. Despite a relatively high precision achievedfor the mean RV curves of some REE ions, no oscillations werefound. The amplitude spectra of Nd II , Nd III , Ce II and Cr II areillustrated in Fig. 12. The two former ions yield the upper limit of ∼
10 m s − for the possible RV variability in the typical roAp fre-quency domain. This rules out the presence of pulsations with theamplitudes comparable to those found in other roAp stars in oursurvey.The projected rotational velocity, v e sin i = 10 km s − , esti-mated in our study is in a good agreement with v e sin i = ± km s − reported by Ryabchikova et al. (2006). We also deter-mined a consistent mean field modulus of h B i = 18.5 kG. Nd II A m p li ude , m s - Nd III
Fe ICe II , mHz
Figure 11.
Amplitude spectra of Nd
III , Nd II , Ce II and Feı for HD 5823. HD 185204 is another poorly studied cool Ap star with prominentchemical peculiarities and a strong magnetic field. It is classifiedas A2 SrEuCr (Renson & Manfroid 2009) and was investigated forpulsational variability by Martinez & Kurtz (1994). These authorsobtained photometric observations on two nights, finding no oscil-lations exceeding ∼ h B i = . ± . kG and v e sin i = 4 km s − .For this star we have two UVES data sets, one with 38 andanother with 50 observations. We have focused RV analysis on thelatter one. Due to the sharpness of its spectral lines, many REEabsorption features can be measured in this star, resulting in accu-rate mean RV curves. In Fig. 13 we illustrate the amplitude spectrafor Nd III , Nd II , Ce II and Caı. There is no evidence of pulsationalvariability, with the upper RV amplitude limits of 20–40 m s − de-pending on the ion. Analysis of the first data set gave similar results,although the precision was a bit worse due to a smaller number ofexposures. Thus, we spectroscopically confirm the noAp status ofHD 185204 to a good level of precision.Our determination of the mean field modulus from theZeeman-split lines yields h B i = 5.4 kG. At the same time, the pro-jected rotational velocity is found to be v e sin i = 4.5 km s − . Giventhe precision of ∼ h B i measurement by Elkin et al. (2012) is signif- c (cid:13) , 1–13 ew roAp stars A m p li ude , m s - Nd III
Nd II
Ce II , mHzCr II
Figure 12.
Amplitude spectra of Nd
III , Nd II , Ce II and Cr II forHD 178892. icant and possibly indicates a real variability corresponding to along rotational period. In this paper we presented a spectroscopic search for high-overtone p -mode oscillations in cool magnetic Ap stars using the UVESspectrometer at the ESO 8-m VLT telescope. Our sample, compris-ing 12 objects, with 11 investigated here and another one studiedby Alentiev et al. (2012), was carefully chosen to satisfy severalcriteria. First, we have used the spectral classification information(e.g. Renson & Manfroid 2009) to select Ap stars with anomalousSr, Cr and Eu line strengths. Second, we have restricted the sam-ple to have photometric T eff below ∼ A m p li ude , m s - Nd III
Nd II
Ce II , mHzCa I
Figure 13.
Amplitude spectra of Nd
III , Nd II , Ce II and Caı for HD 185204. Martinez & Kurtz 1994). For three stars, HD 69013, HD 96237,HD 143487, spectroscopic evidence of pulsations was indepen-dently presented by Elkin et al. (2010, 2011) using a lower qual-ity spectroscopic material compared to our observations. We wereable to confirm the presence of pulsational variability in these starswith much higher confidence levels. In addition, we discovered thatfour other cool Ap stars HD 132205, HD 148593, HD 151860 andHD 177765 are also roAp stars. The latter star, studied in detail byAlentiev et al. (2012), is particularly interesting because it showspulsations with a 24 min period – longer than for any previouslyknown roAp star.Among the 9 objects for which we were able to detect spectro-scopic pulsational variability, all stars showed oscillations in REEspectral lines. In addition, HD 96237, HD 132205, HD 143487,HD 177765 and HD 185256 showed variability in the H α core.Typical pulsation amplitudes of new roAp stars are relatively low.The majority of stars exhibit RV oscillations below 100 m s − for most ions. HD 132205 pulsates with the amplitudes below50 m s − . Pulsation periods of newly discovered pulsators stars liein the 7.1–13.9 min range, which is typical of roAp stars.For three stars, HD 5823, HD 178892, HD 185204, we werenot able to detect spectroscopic oscillations. The relatively rapidrotation of HD 5823 prevented us from obtaining precise RV mea-surements. Therefore, our upper limit on the possible pulsations inthis star is not particularly stringent. For the other two stars wecan rule out pulsations stronger than 10–20 m s − in the Nd II and Nd III lines. The extremely strong, 18.5 kG, magnetic fieldHD 178892 makes it unusual among other roAp candidates. Onlyone roAp star, HD 154708 (Kurtz et al. 2006), with comparable c (cid:13) , 1–13 O. Kochukhov et al. magnetic field strength is known. Magnetic splitting of spectrallines in HD 178892 and HD 154708 complicates accurate RV mea-surements. Furthermore, the cool Ap stars with stronger fields seemto have systematically lower pulsation amplitudes compared to theweak-field roAp stars. We showed that the third non-pulsating starin our study, HD 185204, possesses magnetic field of 5.4 kG, whichis again significantly stronger than that found in a typical roAp star.One of the goals of our survey was to answer the question ofwhether there is a substantial difference in the pulsational prop-erties of the roAp stars and the cool Ap stars previously classi-fied as non-pulsating based on photometric observations. We foundweak oscillations in 7 objects previously classified as “noAp”. Itappears that the range of pulsational amplitudes of the roAp starsdoes not have a well-defined lower threshold but spans all the wayfrom a few km s − in stars like HD 83368 (Kochukhov 2006) andHD 99563 (Elkin et al. 2005a) to our detection limit of ∼
20 m s − .Bright roAp stars pulsating with amplitudes close to these limitsare already known. For example, Kochukhov et al. (2009) detected20–30 m s − multi-periodic pulsations in the Ap star HD 75445.Similarly, Kurtz et al. (2007) and Kochukhov et al. (2008) reportedpulsations at the level of ∼
20 m s − for β CrB. Furthermore, ob-servations with the Kepler satellite revealed roAp pulsations withamplitudes as low as few tens of µ mag (Balona et al. 2011). Allthese results suggest that there is no real physical difference be-tween the large-amplitude roAp stars and the so-called noAp stars.It is quite likely that all cool Ap stars are pulsationally unstable, andobjects in which low-amplitude oscillations are currently being de-tected with spectroscopic monitoring represent the low-amplitudetail of the overall amplitude distribution. ACKNOWLEDGMENTS
We thank James Silvester for improving the language of our paper.OK is a Royal Swedish Academy of Sciences Research Fellow,supported by grants from Knut and Alice Wallenberg Foundationand Swedish Research Council. DA and MC acknowledge financialsupport of FCT/MCTES, Portugal, through the project PTDC/CTE-AST/098754/2008. MC is partially funded by POPH/FSE (EC). TRacknowledges support from Basic Research Program of the Rus-sian Academy of Sciences “Non-stationary phenomena in the Uni-verse”. WW was supported by the Austrian Science Fund (projectP22691-N16).
REFERENCES
Alentiev D., Kochukhov O., Ryabchikova T., Cunha M., TsymbalV., Weiss W., 2012, MNRAS, 421, L82Balmforth N. J., Cunha M. S., Dolez N., Gough D. O., VauclairS., 2001, MNRAS, 323, 362Balona L. A. et al., 2011, MNRAS, 410, 517Bigot L., Kurtz D. W., 2011, A&A, 536, A73Cunha M. S., 2002, MNRAS, 333, 47Elkin V. G., Kurtz D. W., Mathys G., 2005a, MNRAS, 364, 864Elkin V. G., Kurtz D. W., Mathys G., Freyhammer L. M., 2010,MNRAS, 404, L104Elkin V. G., Kurtz D. W., Nitschelm C., 2012, MNRAS, 420, 2727Elkin V. G., Kurtz D. W., Worters H. L., Mathys G., Smalley B.,van Wyk F., Smith A. M. S., 2011, MNRAS, 411, 978Elkin V. G., Riley J. D., Cunha M. S., Kurtz D. W., Mathys G.,2005b, MNRAS, 358, 665 Freyhammer L. M., Elkin V. G., Kurtz D. W., Mathys G., MartinezP., 2008, MNRAS, 389, 441Hatzes A. P., Mkrtichian D. E., 2004, MNRAS, 351, 663Hauck B., North P., 1982, A&A, 114, 23Horne J. H., Baliunas S. L., 1986, ApJ, 302, 757Hubrig S., Kharchenko N., Mathys G., North P., 2000, A&A, 355,1031Hubrig S., Szeifert T., Sch¨oller M., Mathys G., Kurtz D. W., 2004,A&A, 415, 685Kaiser A., 2006, in Astronomical Society of the Pacific Confer-ence Series, Vol. 349, Astrophysics of Variable Stars, Aerts C.,Sterken C., eds., p. 257Khomenko E., Kochukhov O., 2009, ApJ, 704, 1218Kochukhov O., 2004, ApJ, 615, L149Kochukhov O., 2006, A&A, 446, 1051Kochukhov O., 2007, in Physics of Magnetic Stars, RomanyukI. I., Kudryavtsev D. O., eds., pp. 109–118Kochukhov O., 2008, Communications in Asteroseismology, 157,228Kochukhov O., Bagnulo S., Lo Curto G., Ryabchikova T., 2009,A&A, 493, L45Kochukhov O., Ryabchikova T., 2001, A&A, 374, 615Kochukhov O., Ryabchikova T., Bagnulo S., Lo Curto G., 2008,Contributions of the Astronomical Observatory Skalnate Pleso,38, 423Kudryavtsev D. O., Romanyuk I. I., Elkin V. G., Paunzen E., 2006,MNRAS, 372, 1804Kupka F., Piskunov N., Ryabchikova T. A., Stempels H. C., WeissW. W., 1999, A&AS, 138, 119Kurtz D. W., Elkin V. G., Cunha M. S., Mathys G., Hubrig S.,Wolff B., Savanov I., 2006, MNRAS, 372, 286Kurtz D. W., Elkin V. G., Mathys G., 2005, MNRAS, 358, L6Kurtz D. W., Elkin V. G., Mathys G., 2007, MNRAS, 380, 741Kurtz D. W. et al., 2002, MNRAS, 330, L57Kurtz D. W., Martinez P., 1995, Information Bulletin on VariableStars, 4209, 1Kurtz D. W., Martinez P., 2000, Baltic Astronomy, 9, 253Maitzen H. M., Paunzen E., Vogt N., Weiss W. W., 2000, A&A,355, 1003Maitzen H. M., Vogt N., 1983, A&A, 123, 48Martinez P., 1993, PhD thesis, University of Cape Town, SouthAfricaMartinez P., Koen C., Sullivan D. J., 1998a, MNRAS, 300, 188Martinez P., Kurtz D. W., 1994, MNRAS, 271, 129Martinez P., Kurtz D. W., Meintjes P. J., 1993, MNRAS, 260, 9Martinez P., Marang F., van Wyk F., Roberts G. R., 1998b, TheObservatory, 118, 153Mathys G., Hubrig S., Landstreet J. D., Lanz T., Manfroid J.,1997, A&AS, 123, 353Mkrtichian D. E., Hatzes A. P., Saio H., Shobbrook R. R., 2008,A&A, 490, 1109Moon T. T., Dworetsky M. M., 1985, MNRAS, 217, 305Napiwotzki R., Schoenberner D., Wenske V., 1993, A&A, 268,653Nelson M. J., Kreidl T. J., 1993, AJ, 105, 1903Renson P., Manfroid J., 2009, A&A, 498, 961Ryabchikova T. et al., 2006, A&A, 445, L47Ryabchikova T., Nesvacil N., Weiss W. W., Kochukhov O., St¨utzC., 2004, A&A, 423, 705Ryabchikova T., Sachkov M., Kochukhov O., Lyashko D., 2007a,A&A, 473, 907Ryabchikova T. et al., 2007b, A&A, 462, 1103 c (cid:13) , 1–13 ew roAp stars Sachkov M., Kochukhov O., Ryabchikova T., Huber D., Leone F.,Bagnulo S., Weiss W. W., 2008, MNRAS, 389, 903Saio H., Gautschy A., 2004, MNRAS, 350, 485Saio H., Ryabchikova T., Sachkov M., 2010, MNRAS, 403, 1729Scargle J. D., 1982, ApJ, 263, 835Th´eado S., Dupret M.-A., Noels A., Ferguson J. W., 2009, A&A,493, 159Wraight K. T., Fossati L., Netopil M., Paunzen E., Rode-PaunzenM., Bewsher D., Norton A. J., White G. J., 2012, MNRAS, 420,757 c (cid:13)000