On the evolution of the magnetic field of Ap star α 2 CVn
aa r X i v : . [ a s t r o - ph . S R ] J un On the evolution of the magnetic field ofAp star α CVn
V.D. Bychkov , L.V. Bychkova , J. Madej , G.P. Topilskaya Special Astrophysical Observatory, Russian Academy of Sciences; email: [email protected] Warsaw University Observatory, Poland; email: [email protected] North-Caucasian Federal University, Stavropol, Russia; email: [email protected]
Abstract
New high-precision measurements of the longitudinal magnetic fieldof Ap stars suggest the existence of secular intrinsic variations of theglobal magnetic field in some stars. We argue that such changes areapparent in the Ap star α CVn in the time scale of ∼
10 years, whichresults from the analysis of literature data. Therefore, such an obser-vation implies, that the rate of magnetic field evolution of Ap stars ismuch higher than was previously thought. submitted to
Proc. of the IV Annual Int. Conf.North-Caucasian Federal UniversitySTAVROPOL, (Russia)20-22 April 20161
Introduction
The study of evolution of the global magnetic fields of Ap stars is a veryimportant recent research subject. Theoretical studies of the evolution ofthese objects were carried out by a number of authors and were extensivelysummarized by Moss (1990). According to his results, magnetic fields ofstars also must evolve in the time scale of a general evolution of stars, i.e.in time scale of about 2 . × − . × years. Landstreet et al. (2007)presented the most accurate estimates of the rate of magnetic field evolutionof Ap stars obtained on the basis of observational data.Estimates of the characteristic time of the magnetic field evolution yield2 − × years for massive A stars of mass M higher than 3 solar masses, M > M ⊙ . For stars with masses lower than 3 solar masses, M < M ⊙ , therate of evolution is of the order of 10 years. These numbers basically referto stars which have the global magnetic field of a simple dipole structure andare fully consistent with theoretical estimates by Kochukhov and Bagnulo(2006).According to theoretical studies by Krause and Raedler (1980), if theglobal magnetic field has a more complex structure of the quadrupole, oc-tupole, etc., then its evolution proceeds much more rapidly than in the caseof a simple dipole. Currently, accuracy of determination of the longitudinal magnetic field instars has substantially improved due to the use of modern light detectorsand new methods of processing of observational data. New techniques re-vealed a number of subtle effects in the run of the longitudinal magneticfield variations, which previously were unavailable for research.Two series of projects were recently carried out on exact measurements of B e in stars during past two decades, which enabled us for the constructionof the magnetic phase curves of some Ap stars. Results of the first high-precision measurements of the longitudinal magnetic fields B e using theLeast Squares Deconvolution (LSD) method (Donati et al. 1997) and WLSD(Wade et al. 2000a), obtained at the end of the XXth century, were publishedin Wade et al. (2000b). In the latter study, series of B e measurements wereobtained for 14 Ap stars and for 11 of stars in that set observations coverall phases of the rotational period. Silvester et al. (2012) also published B e series obtained by analogous methods for 7 Ap stars.The latter measurements were carried out in years 2006 - 2010 and forthree stars they satisfactorily cover all phases of the rotational period. Inboth quoted papers the average B e measurement error corresponding to1 σ is lower than 50 G. There are 6 stars studied in both sets of B e mea-1 (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) B e ( G ) (cid:13) Figure 1: Magnetic phase curves of α CVn. Red line denotes phase curvederived from B e points in Wade et al. (2000), whereas black line was derivedfrom Silvester et al. (2012).surements for which authors presented data on two high-precision phasecurves separated by several years of observations and obtained using thesame methodology. Therefore, these phase curves can be directly comparedin order to search for possible changes in the magnetic phase curves of CPstars for 10 years. α CVn
Significant changes of the magnetic phase curve showed up in the well-studied magnetic Ap star α CVn (SiCrEuHg type). Periodic variabilityof the longitudinal (effective) magnetic field B e with the rotational phasehas a complex double-wave curve which corresponds to the structure of aquadrupole magnetic field (at the first approximation). Results of the firsthigh-precision measurements using the LSD method were first publishd byDonati et al. (1997) and Wade et al. (2000b). Authors of the latter paperpresented a series of 18 measurements distributed over 700 days (1.9 years)around the average date JD 2550848.56, see open circles in Fig. 1. It wasfound that the average accuracy of B e measurements equals 27.8 G.The second set of high-precision measurements was obtained by the samemethods and was published in Silvester et al. (2012). Their results form a2 (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) -1sigma(cid:13)+1sigma(cid:13)-3sigma(cid:13)+3sigma(cid:13) (cid:13) (cid:13) B e ( G ) (cid:13) Phase(cid:13)
Figure 2: Difference between both phase curves of α CVn, B e [2000] − B e [2012], is given by the solid line. The amplitude of the residual phasecurve strongly exceeds the 3 σ uncertainty range of the magnetic phase curvedetermination.series of 27 measurements obtained over 1295 days (3.5 years) with theaverage date JD 2554722.51. These estimates are plotted in Fig. 1 as filledcircles with the corresponding magnetic phase curve. Phase curve and itsamplitude was found with the average accuracy of 9.4 G (solid line).Rotational phases of α CVn in Fig. 1 were determined using the ephemerisby Farnsworth (1932),
J D ( EuIImax ) = 2419869 .
720 + 5 . E .Time interval between the centers of these two sets equals 3874 days(10.6 years). Figure 1 clearly shows that during the period of about 10 yearsphase curve markedly changed. Calculated difference between both phasecurves is plotted as the function of the rotational phase in Figure 2, whichalso shows the 1 σ uncertainty of the amplitude of the residual phase curvesequal to 29.4 G.The largest difference of 222 G occurs in the rotational phase 0.96, whichequals 7.5 sigma. In phase of 0.39 the difference between both phase curvesequals -125 G, or 4.3 sigma. Therefore, it is a very significant difference. Itis very likely that this is a real change of the magnetic phase curve occuringduring about 10 years, Therefore, we note that the rate of evolution of themagnetic field in this star by about 5 - 6 orders of magnitude faster than3as previously thought.It should be noted that, according to Krause and Raedler (1980), evo-lution of the global magnetic fields of stars with a complicated structure(quadrupole, octupole, etc.) should take place significantly faster than instars with a simple dipole structure of the magnetic field. But even in sucha case it turns out that the magnetic field of α CVn actually evolves muchfaster (4 - 5 orders of magnitude) than was predicted before (Krause andRaedler 1980).
Naturally the following question arises - how realistic is this effect? To con-firm that this is a real feature we demonstrate lack of such phase curvesdifferences in some other Ap stars in which the longitudinal magnetic fieldwas measured by the same instrument and methods, published by the sameauthors in the same papers. We have not found such significant changes ofthe magnetic phase curves for the other CP stars from the list.As an example, consider magnetic Ap star HD62140. In the paper byWade et al. (2000a), 14 high-precision measurements of the longitudinalmagnetic field B e were obtained during 705 days of observations (1.9 years)centered on JD 24550858.0. Fig. 3 presents two magnetic phase curves, againthe second curve was derived from Silvester et al. (2012). The latter paperpresented 19 measerements obtained during 1171 days (3.2 years) about JD24554666.3.Time span between the centers of both sets of B e measurements equals3808.3 days (10.4 years). Duration of sets, number of B e points and thetime interval between sets is very close to that obtained for α CVn. At therotational phase ψ = 0 .
01 the difference between phase curves amounts to1 . σ , for ψ = 0 .
51 to 4 . σ .For HD62140 long-term changes of the phase curve are much lower andapparently show the opposite sign. I.e. the amplitude value estimates for α CVn increased as compared with the first set (phase change and theshape of the curve), while for the opposite HD62140 slightly decreased inhigh magnetic field. Similar results were obtained for other stars have beeninvestigated in these studies.Fig. 4 shows results of another set of high-precision of B e measurementsof HD71866, derived from the same two papers using the same methods,instrument and similar dates of observation. We noted the probable existence of secular changes of the global magneticfield in Ap star α CVn exceeding the estimated 3 σ uncertainty level. Our4 (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) B e ( G ) (cid:13) Phase(cid:13)
Figure 3: Phase curves of HD62140. Open circles - B e measurements byWade et al. (2000a), filled circles - Silvester et al. (2012).observation was dervived from two published papers which presented seriesof the longitudinal magnetic field B e of a small group of Ap stars obtainedby the precise Least Squares Deconvolution method. We compared here B e rotational phase curves of α CVn separated by a time span of ≈
10 years.At present it is still too early to draw definite conclusions on the promptsecular evolution of the magnetic field of α CVn. That problem requiresobtainin of more high-precision observational data. Actually we prepare newinstrument at SAO in order to verify the complex and subtle problem ofsecular variations of the magnetic field in Ap stars (Valyavin et al. 2014).If the change of the magnetic phase curve of α CVn is confirmed, it willradically change the whole idea of the evolution of global magnetic fields Apstars.
Research by V.D. Bychkov was supported by the Russian Scientific Foun-dation grant N14-50-00043. 5 (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13)
HD71866(cid:13) (cid:13) (cid:13) B e ( G ) (cid:13) Phase(cid:13)
Figure 4: Phase curves of HD71866. Open circles - Wade et al. (2000a), filledcircles - Silvester et al. (2012).