Light-curves of symbiotic stars in massive photometric surveys I: D-type systems
aa r X i v : . [ a s t r o - ph . S R ] J un ACTA ASTRONOMICA
Vol. (2008) pp. 0–0 Light-curves of symbiotic stars in massive photometric surveys I:D-type systems
M. G r o m a d z k i , J. M i k o ł a j e w s k a ,P. W h i t e l o c k , , F. M a r a n g N. Copernicus Astronomical Center, Bartycka 18, PL-00-716 Warsaw, Polande-mail: marg,[email protected] South African Astronomical Observatory, P.O. Box 9, Observatory, 7935, South Africae-mail: paw,[email protected] National Astrophysics and Space Science Programme, Department of Astronomy,University of Cape Town, Rondebosch, 7701, South Africa
Received Month Day, Year
ABSTRACTASAS, MACHO, OGLE and SAAO
JHKL light curves of 13 stars, that have at some timebeen classified as D-type symbiotics, are analysed. Most of the near-IR light-curves that have beenmonitored over many years show long-term changes due to variable dust obscuration, in addition tothe stellar pulsation. The distances to these objects are derived from the period-luminosity relationand estimates of the mass-loss rates made from the K − [ ] colour.We reclassify AS 245 as an S-type symbiotic, with a semi-regular cool component with a pulsa-tion period of about one year. The periods of the large amplitude pulsations of SS73 38 (463 days),AS 210 (423 days) and H2-38 (395 days) are estimated for the first time, confirming that they aresymbiotic Miras.A comparison of the symbiotic Miras with normal Miras of similar pulsation period shows thatthe symbiotic stars have on average higher values of K − [ ] . This may indicate that they havehigher mass-loss rates, or more likely that the dust which is being lost by the Mira is trapped withinthe binary system. Key words: binaries: symbiotic – Stars: individual: o Cet, RX Pup, V366 Car, BI Cru, SS73 38,V347 Nor, AS 210, AS 245, H 2-38, RR Tel, R Aqr, StHA 55, V335 Vul – surveys
1. Introduction
Symbiotic stars are long-period interacting binary systems, in which an evolvedred giant transfers material onto its much hotter companion, which in most systemsis a white dwarf. Based on their near-IR characteristics, symbiotic stars divide intotwo main classes (Allen 1982) depending whether the colours are stellar (S-type) orindicate a thick dust shell (D-type). The majority ( ∼ ol. 58 ∼ − ∼ − e.g. Mikołajewska 2008). The near-IR colours of D-type systems indicatethe presence of a dust shell which obscures the star and re-emits at longer wave-lengths. IR photometric monitoring has shown that these D-type systems have largeamplitude variations and that they contain Mira variables with pulsation periods inthe range 300–600 days; they are often called symbiotic Miras (Whitelock 1987).Since they must accommodate the Mira with its dust shell, these D-type systemsshould have much longer orbital periods than the S-types, a few tens of years andmore. The latest review of symbiotic Miras and a comparison with normal Mirascan be found in Whitelock (2003).Light curves of symbiotic stars reflect the very complex behaviour of thesesystems. They show high and low activity stages, flickering, nova-like outburstsoriginating from the hot component (S & D types), eclipses, ellipsoidal variabilityconnected with orbital motion (S-type), radial pulsations (all D-type and some S-type) and semi-regular variation (S-type) of the cool component, long-term dustobscuration (mostly D-type) and other types of variability (Mikołajewska 2001).In this paper we analyse the light curves of 13 objects that have at some timebeen classified as D-type symbiotics. The light-curves were provided by massivephotometry surveys such as ASAS, OGLE, MACHO and near-IR monitoring atSAAO.
2. Data
Belczy´nski et al. (2000) listed coordinates for symbiotic stars, but many ofthese are not sufficiently accurate to identify the symbiotics unambiguously. So wefirst identified the 2MASS counterparts using the existing finding charts and theAladin Java graphics interface running at the CDS in Strasbourg. This works wellbecause symbiotic stars, which have the near-IR colours of late-type giants, areintrinsically bright in
JHK . The 2MASS coordinates were then used to identifysymbiotic stars in the OGLE, MACHO and ASAS databases.For o Cet, RX Pup, V366 Car, BI Cru, SS73 38, AS 210, RR Tel, R Aqr,StHA 55, and V335 Vul, light-curves were taken from the ASAS database (Poj-ma´nski 2002) . These comprise V -band photometry obtained between November2000 and February 2009. The ASAS V light-curves are illustrated in Fig. 1, whileTable 1 contains the ASAS names, the mean mags ( V ) and the full range of thevariations found in V ( D V ). Although the ASAS database also contains SS73 38,there are only a few observations of it.The OGLE-II database (Udalski et al. includes a light-curve for H2-38.This comprises I -band photometry obtained between 1997 and 2000. The OGLE official home page of ASAS project: official home page of OGLE project: http://ogle.astrouw.edu.pl/ A. A. I light-curve of H2-38 is shown in the second panel of Fig. 2, while Table 1 givesthe OGLE name, the mean mag ( I ) and the full range of the variations found in I ( D I ).The MACHO database (Alcock et al. contains observations for H2-38,obtained between 1993 and 1999. The photometry was made through non-standardblue ( B M ) and red ( R M ) filters. The MACHO light-curve is shown in Fig. 2, whileTable 1 gives the MACHO name, the mean mags ( B M , R M ) and the full range ofthe variations found in each band ( D B M , D R M ).We also analyse near-IR JHKL photometry for o Cet, RX Pup, V366 Car,BI Cru, V347 Nor, SS73 38, AS 210, AS 245, RR Tel and R Aqr. This was ob-tained with the 0.75-m, 1.0-m and 1.9-m telescopes at SAAO and is on the systemdescribed by Carter (1990). The photometry is illustrated in Figs. 3-5 and meanmagnitudes, the full range of variations and the amplitudes of pulsation estimatedby fitting sinusoids to detrended light curves (see Section 3.1..1 for period deriva-tion) are listed in Table 2. Some of the early data were published and discussedby Feast et al. et al. et al. et al. et al. .Visual light-curves for o Cet, RX Pup, BI Cru, AS 210, RR Tel and R Aqrwere extracted from the database of the American Association of Variable StarObservers , and used for comparison with the near-IR data.Most of objects discussed here are definitely symbiotic Miras, with four excep-tions; the symbiotic nature of o Cet, StHA 55 and V335 Vul is not certain. Therehave been suggestions (Jura & Helfand 1984, Kastner & Soker 2004, Ireland etal. et al. et al.
3. Analysis and Results
All light-curves were analysed using the program PERIOD ver. 5.0, based on themodified Lomb-Scargle method (Press & Rybicki, 1989). Long-term trends were official home page of MACHO project: official home page of Acta Astronomica: http://acta.astrouw.edu.pl/ official home page of AAVSO: the source program is available on o l . Table 1: D-type symbiotic stars in optical massive surveys: designation, average magnitudes and full range of the variations ( D )during the observation run. The first column lists the identification number of the symbiotic star from Belczy´nski et al. (2000). No. Name Name in Survey V D V I D I B M D B M R M D R M (mag)010 o Cet ASAS 021920-0258.6 6.75 6.28 - - - - - -026 RX Pup ASAS 081412-4142.5 12.27 1.53 - - - - - -030 V366 Car ASAS 095443-2745.5 13.78 1.95 - - - - - -034 BI Cru ASAS 122326-6238.3 11.44 2.90 - - - - - -039 SS73 38 ASAS 125126-6460.0 - - - - - - - -069 AS 210 ASAS 165120-2600.5 12.91 2.04 - - - - - -120 H 2-38 OGLE 180601.19-281704.2 - - 12.21 1.73 - - - -MACHO 105.21287.23 - - - - -9.61 > A . A . Table 2: Near-IR SAAO photometry of the symbiotic stars: average magnitudes in the
JHKL − bands, full range of the variation ( D )of the JHKL magnitude and the amplitudes of pulsation estimated by fitting sinusoid with peak-to-peak amplitudes of D P J etc to theobservations. The first column lists the identification number of the symbiotic star from Belczy´nski et al. (2000). No. Name
J H K L D J D H D K D L D P J D P H D P K D P L (mag)010 o Cet -1.23 -2.10 -2.55 -3.12 1.38 1.40 1.24 1.39 1.00 1.01 0.84 0.68026 RX Pup 5.67 4.18 2.98 2.28 3.72 3.01 2.30 1.59 1.08 1.00 0.81 0.64034 BI Cru 7.62 6.14 4.84 3.30 2.07 1.40 1.11 0.93 0.69 0.46 0.31 0.19030 V366 Car 7.19 5.77 4.78 3.52 3.05 2.58 1.93 1.28 0.53 0.59 0.49 0.39039 SS73 38 9.41 7.47 5.96 4.18 3.18 2.56 1.79 1.06 0.53 0.59 0.49 0.39060 V347 Nor 7.01 5.54 4.78 4.02 1.81 1.32 1.00 0.94 0.88 0.70 0.53 0.44069 AS 210 9.50 7.74 6.28 4.60 2.67 2.48 1.84 1.27 1.06 1.16 1.05 0.85105 AS 245 9.41 8.05 7.44 6.92 0.54 0.57 0.53 0.44 0.36 0.42 0.38 0.30175 RR Tel 6.57 5.37 4.43 3.13 3.47 2.84 2.05 1.29 0.71 0.69 0.60 0.47188 R Aqr 0.80 -0.27 -0.89 -1.72 3.58 3.07 2.42 1.79 1.00 0.91 0.78 0.72 ol. 58
No. Name Type ASAS OGLE MACHO SAAO Other periods(days)010 o Cet O 338 ±
10 - - 332 ±
026 RX Pup O trend - - 575 ±
030 V366 Car O trend - - 432 ±
034 BI Cru O 280 ± ±
039 SS73 38 C - - - 463 ± ±
069 AS 210 C 407 ±
14 - - 423 ± ±
36 395 ± a -175 RR Tel O trend - - 385 ±
188 R Aqr O 395 ±
13 - - 391 ± s03 StHA 55 C 372 ±
15 - - - 395 s26 V335 Vul C 334 ±
14 - - - 347 derived from R M .References: [1] Kholopov 1985; [2] Mikołajewska et al. et al. et al. et al. et al. et al. et al. Table 4: New pulsation ephemerides.
No. Name Ephemeris026 RX Pup Max ( JHKL ) = + × E
030 V366 Car Max ( JHKL ) = + × E
034 BI Cru Max ( JHKL ) = + × E
039 SS73 38 Max ( JHKL ) = + × E
069 AS 210 Max ( JHKL ) = + × E
120 H 2-38 Max ( R M ) = + × E
175 RR Tel Max ( JHKL ) = + × E s03 StHA 55 Max ( V ) = + × E A. A. removed by subtracting a polynomial of appropriate order, and the resultant powerspectra were compared with the window spectra. The periods were derived fromthe inverse of the maximum of the peak in the periodogram ( P = f − max ), whereastheir accuracy was estimated by calculating the half-size of a single frequency bin( D f ), centred on the peak ( f c is the center of the peak ) of the periodogram andthen converted to period units ( D P = f − c · D f ). The results of our period analysisare summarized in Table 3. Examples of our power spectra are shown in Fig. 6-7and Figs. 8 shows near-IR light curves folded with pulsation periods.The highest peak in a typical power spectrum corresponds to the pulsation pe-riod, while the other peaks represent annual aliases, second and third harmonics,long-term variation and some combination thereof. The only exceptions are theASAS light-curve of V335 Vul, where the highest peak corresponds to the secondharmonic, and the near-IR light-curves of RR Tel where it represents the annualalias. In both exceptional cases this is due to gaps in the light-curve of the object.There is practically no difference between the power spectra derived for the JHK or L observations. They all show the same position of peaks with a littledifference in power. The latter is due to differences of amplitudes of the pulsationsand long-term trends in the near-IR photometry.For four systems (AS 210, H2-38, SS73 38, and AS 245) the pulsation periodsare derived for the first time. Pulsations are also detected in other systems withknown periods. In particular, pulsations are always visible in the near-IR light-curves of all of the symbiotics we examined. However, the periods for o Cet andR Aqr (Table 3) are not as accurate as those derived from visual data collectedover a few centuries (Kholopov 1985), whereas for RX Pup, V366 Car, RR Teland BI Cru our new estimates are better than published values (Mikołajewska et al. et al. et al. et al. et al. (2008) estimated a pulsation period of 395 days forStHA 55 in the V -band. The value is very uncertain because they had only obser-vations covering only one pulsation cycle. The pulsation period of StHA 55 derivedfrom ASAS data, which covers 6 pulsation cycles, is 372 days.The pulsations ephemerides for three systems for which periods are derived ol. 58 et al. et al. (2001), andWhitelock et al. (1994, 2000, 2006). These non-symbiotic Miras are henceforth re-ferred to as ‘normal Miras’, but note that they will include widely separated binarysystems; indeed o Ceti itself is included in both the symbiotic and normal group-ings. The normal Miras show distinct period distributions, with peaks at ∼ ∼
530 days for the O- and C-rich objects, respectively. The period distribu-tion for normal Miras is influenced by selection effect that are extremely difficultto quantify. We do know that there are O-rich Miras with periods over 1000 days,the OH/IR stars, but these have progenitors of several solar masses and are quiterare. The symbiotic Miras have a mean period of about 400 days, and with oneexception range from 280 (BI Cru) to 580 (RX Pup). The exception, V407 Cyg,has the period of 763 days (Kolotilov et al. et al. i.e. we expect there to beunidentified white dwarfs in binary systems with short period Miras with very lowlevels of interaction.
The near-IR light-curves of all symbiotic Miras included in this study show, inaddition to the Mira pulsations, significant long-term variations. Such trends arevery common in symbiotic Miras, and are almost certainly caused by variable dustobscuration ( e.g.
Whitelock 1987; Mikołajewska et al. o Cet, R Aqr, BI Cru and V347 Nor, the pulsationlight-curves are asymmetrical and simply subtracting the sine-curve produced a lotof scatter in the residual. Therefore another method was used. We first generatedan average pulsation curve and then subtracted it from the original light-curve.This method produced less scatter for the short period stars, while for the long
A. A. period pulsators both methods give similar results. In the case of BI Cru, databetween JD2 445 400 and JD2 451 400 were used to prepare the average pulsationlight-curves by excluding the dust obscuration events. In the case of R Aqr, databefore JD2 444 000 were omitted from the average pulsation light-curves for thesome reason.The light-curves, prior to removal of the pulsations, are plotted in the upperpanels of Figs. 3 to 5. The secular trend is clearest at J and the behaviour of J − K indicates that it is due to increasing reddening, as expected for increasing opticaldepth of the dust. This general trend is also present in the L light-curves.The near-IR colours of D-type systems indicate the presence of warm dustshells. bf Fig. 10 presents the J − K vs. K − L diagram for the symbiotic Mi-ras in this study together with those for normal Miras. The symbiotic Miras areshown in both their obscured and unobscured states. This demonstrates that mostof the colours can be qualitatively reproduced with a shell of around 800K (see alsoWhitelock 1987). In at least 50 % of the studied systems the reddening toward theMira star was larger than reddening toward the hot component ( e.g. Mikołajewska1999). There appears to be very much less extinction towards the high excitationregions (emission lines and hot UV continuum) than towards the Mira suggestingthat the hot component generally lies outside the dust cocoon associated with theMira. This fact constrains the orbital separations in these systems to be & &
20 yr, because the typical radius of a dust shellis & R Mira , and the Mira radius, R Mira ∼ ∼ − ∼ et al. J is not associated with abrightening at L , as might be anticipated if the dust forms in a uniform shell aroundthe Mira (see Figs. 3 to 5). BI Cru is the one exception, but it has a spectrum quiteunlike that of a normal O-rich Mira in that it does not show the characteristic H O ol. 58 µ m CO-band emission, and inthat respect it has similarities to certain B[e] stars, such as Hen 3-1138 and Hen 3-1359, (Whitelock et al. et al. et al. . 2003; Whitelock et al. et al. et al. o Cet shows cycle-to-cycle changes of the visual am-plitude. Similar behaviour is seen in most Miras with sufficiently good light-curvecoverage. Erratic variations are also evident in the near-IR data, but the coverage ofthe light-curves is not as good. We note, however, that the highest visual maximaare correlated with the maxima in near-IR light-curve with (see left panels of Fig.3) which may suggest changes of the average luminosity of the Mira.AS 245 has colours typical of an oxygen-rich Mira (see Fig. 10). However, theamplitude of the possible pulsation ( . . ∼ d and low K − [ ] . Gromadzki et al. (2007) suggestedthat the red component of MWC 560 is on thermally-pulsating AGB, althoughits pulsation characteristics may be influenced by its nearby companion possiblyreducing the pulsation amplitude and the circumstellar dust. More observations arenecessary to settle the nature of the cool component of AS 245, but we have enoughinformation here to reclassify it as an S-type; its position in Fig. 10 does not showsigns of the dust associated with most symbiotic Miras.0 A. A.
The great advantage of long-period pulsating AGB stars is an opportunity todetermine various physical parameters such as absolute magnitudes, distances andmass-loss rates using near-IR period-luminosity-colour relations derived by Feast,Whitelock and coworkers in a series of papers. In this section we apply thesemethods to estimate the physical parameters for symbiotic systems containing C-rich and O-rich Miras.
The absolute K magnitudes of both O- and C-rich symbiotic Miras are estimatedusing the latest Whitelock et al. (2008) period-luminosity (PL) relationship: M K = r [ log P − . ] + d . (1)The slope r = − . ± .
20 was derived from large amplitude asymptotic gi-ant branch variable in the LMC. We use the zero-point d = − . ± .
07 for O-rich Galactic Miras, estimated using the revised
Hipparcos parallaxes together withpublished VLBI parallaxes for OH Masers and Globular Clusters. That value agreeswith those estimated for O-rich LMC Miras ( d = − . ± .
06) and C-rich Galac-tic ( d = − . ± .
37) and LMC ( d = − . ± .
07) Miras (assumes an LMCdistance modulus of 18 . ± .
05 mag; van Leeuwen et al. V , A D V , is estimated using the Drimmel et al. (2003) 3-D Galactic extinction model, including the rescaling factors that correctthe dust column density to account for small structures observed in the DIRBE data,but not included explicitly by the model. The extinction at JHK is then calculatedfrom the relations given by Glass (1999) for photometry on the SAAO system.Obviously this extinction represents only the interstellar component and tells usnothing about any circumstellar reddening.We can compare the observed J − K with the intrinsic value, as discussed be-low, to get an estimate of the total, interstellar plus circumstellar extinction. Theintrinsic ( J − K ) for O Miras is estimated using the period-colour relation de-rived by Whitelock et al. (2000) for Miras in the solar neighbourhood observed byHipparcos: ( J − K ) = .
71 log P − . . (2)Intrinsic near-IR colours ( J − K ) of C Miras are obtained from the period-colour relations for C Miras from Whitelock et al. (2006): ( J − K ) = .
811 log P − . . (3)Although the period colour relation for O-Miras is quite well defined, at leastat short periods, that for C-rich Miras is not; for example the best estimated rela-tion, ( H − K ) vs. period, has a standard deviation of s = .
48 mag. Therefore, ol. 58
11a reliable total extinction estimate is often impossible for C-rich objects and theuncertainty of the values derived for E ( B − V ) are huge.We assume wherever possible that the extinction derived from the period-colourrelations is the total, circumstellar plus interstellar, value and that subtracting theinterstellar extinction, evaluated as described above, gives the circumstellar value.The values of E ( B − V ) in the unobscured parts of light-curves, that are used forthese estimates, are listed in Table 5.To estimate distances we derive the absolute K mag, M K , from the period andtake the observed K mag during intervals in which the objects did not show obviousdust obscuration, and average over the pulsation cycle. We use A K derived fromthe colour excess to correct for the interstellar plus circumstellar extinction. Theformal error of the distance derived by this approach is 12 – 20 %. The parametersderived as described above are listed in Table 4. There is a good correlation between mass-loss rate and K − [ ] colour for bothC- and O-rich Miras, provided we can assume that the shells are approximatelyspherically symmetric. This arises because to a first approximation the K and [12]mags are measures of the brightness of the star and of the shell, respectively. If,however, the shell is very asymmetric and our line of site to the star is obscured bythick dust (as it would be during an obscuration event) then we will underestimatethe K brightness and overestimate the mass-loss rate. With this caveat in mind wecan estimate the mass-loss rates for the symbiotic Miras from their K − [ ] colours(Table 5).In the case of carbon Miras, Whitelock et al. (2006) estimated mass-loss ratesusing a modification of Jura’s (1987) method, and fitted an analytical formula whichwe use here:log ˙ M = − . + . ( K − [ ]) − . × − ( K − [ ]) (4) + . × − ( K − [ ]) . The mass-loss rates for the symbiotic O-rich Miras are determined using thecorrelation with K − [ ] colour derived for 58 high mass-loss O-rich AGB stars inthe South Galactic Cap (Fig. 21 of Whitelock et al. [ ] magnitudes are calculated from the IRAS et al. [ ] = − . F + .
62. The K magnitudes for the objectswith SAAO near-IR light-curves are the values observed outside of obvious obscu-ration phases and averaged over the pulsation cycle. For HM Sge and V1016 Cygwe derived the average K from their published near-IR light-curves (Taranova &Shenavrin 2000). The use of a K magnitude obtained during a dust obscurationphase could result in an overestimate of the mass-loss rate by an order of magnitude(we assume, as suggested above, that the faint phases are the result of asymmetric2 A. A. obscuration and that the [12] mag remains approximately constant). For examplethe average K mag for V366 Car outside dust obscuration is 4.4 mag, from whichwe derive a mass-loss rate of 6 . × − M ⊙ yr − ; during dust obscuration K isabout 6 mag and the derived mass-loss rate is 5 × − M ⊙ yr − . The difference isnot so large for objects with greater mass-loss rates, e.g. , RX Pup outside of dustobscuration has ˙ M = . × − M ⊙ yr − , which increases to ˙ M = − M ⊙ yr − during obscuration.For the remaining objects we use data from 2MASS (obtained between 1997and 2000) and Munari et al. (1992; obtained in 1990). Whereas the near-IR datain Munari et al. (1992) are from SAAO using the same photometric system asthe light-curves presented here, the K s magnitudes from 2MASS have been trans-formed to the SAAO system using transformations from Carpenter (2001). A sig-nificant fraction of symbiotic Miras show dust obscuration events, so it is possiblethat some of these measurements were made during such events. Typically, for ob-jects with SAAO near-IR light-curves, the difference between the average K andtransformed 2MASS K s is 0.1-0.2 mag, and so we adopted the average of the Mu-nari et al. (1992) and transformed 2MASS values. The K magnitudes have beencorrected for interstellar reddening (section 3.2..1).We should emphasize that mass-loss rates obtained in this way are not accu-rate. It is also important to appreciate that we do not yet understand the obscu-ration events or their link to the Mira mass-losing activity. If the symbiotic starsshow obscuration events for the same reason that Feast (2003) and Whitelock et al. (2006) have suggested that the C-Miras do, then the mass-loss will be in discreetdust puffs in random directions like RCB stars (see also Whitelock 2003). The factthat symbiotic Miras are in binary systems will almost certainly effect what hap-pens to the dust once it leaves the immediate vicinity of the Mira in a non-randomfashion, even if it does not affect the dust production itself.Fig. 11 presents the K − [ ] colour distribution for symbiotic Miras andcompares it to those for normal Miras. The normal O-rich Miras show a distinct K − [ ] colour distribution, which peaks at 1.76 mag, corresponding to a mass-loss rate of ∼ − M ⊙ yr − , and has a tail that extends across the region occupiedby the symbiotic Miras. As can be seen from Fig. 12, a plot of the K − [ ] colourvs. pulsation period, the tail is comprised of long period (mostly P >
400 days)objects. The C-rich Miras show a broad distribution from 1 to 9.5 mag which im-plies mass-loss rates ranging from ∼ − to ∼ × − M ⊙ yr − . The symbioticMiras show a distinct K − [ ] colour distribution with a peak at 4.25 mag, cor-responding to a mass-loss rate of ∼ . × − M ⊙ yr − . These figures show thatthe symbiotic Miras have on average larger K − [ ] than the their normal coun-terparts with similar periods. Fig. 13 shows a plot of K − [ ] vs. ( J − K ) forsymbiotic and normal Miras. The scatter of symbiotic Miras is larger than that ofthe normal Miras and, although the distributions overlap, the symbiotic Miras havelarger K − [ ] than normal Miras of the same period or same J − K . If this is the ol. 58 e.g. Fleischer, Gauger & Sedlmayer1992) is that the Mira pulsation lifts matter above the atmosphere, but does notaccelerate it to escape velocity. As the matter falls it encounters the next pulsation,which gives it another outward impulse. Matter is pushed by pulsations until itstemperature drops enough for dust to condense (at a few stellar radii). Then, radi-ation pressure on the dust can efficiently accelerate the dust, and the dynamicallycoupled gas, to the escape velocity. K − [ ] is proportional to the amount of dustbetween us and the Mira, as described above. We should emphasize that in a binarysystem higher K − [ ] could be the result of the lost mass not leaving the systemrather than of intrinsically higher mass-loss rates. For example, if mass lost fromthe Mira in the binary was trapped within the system ( e.g. near the L and L La-grange points), more cool dust in the system will mean more obscuration and moreflux at 12 µ m and will result in higher values of K − [ ] .An important question is about the influence of the secondary star on the mass-loss rate and its character. Podsiadlowski & Mohamed (2007) proposed a windRoche lobe overflow (RLOF) model for o Cet. In their model, a slow wind fromMira fills its Roche lobe and then the matter streams – via the L1 Lagrangian point– onto an accretion disk around the companion. This models works most efficientlywhen the radius of dust shell is comparable with the Roche lobe radius ( R RL ). Or-bital separations for symbiotic Miras are essentially unknown, with the exceptionof R Aqr. The mean angular rotation rate from spectro-polarimetry suggest an aver-age period of around 150 yr (Schmid & Schild 2002). This estimation is uncertainbecause their observations cover about 10 yr (only about 5% of the implied orbitalperiod). For typical component masses, M hot ∼ . ⊙ and M Mira ∼ . ⊙ , thisperiod corresponds to an orbital separation of ∼
40 AU and R RL ∼
15 AU. Thuswind RLOF should occur in most symbiotic Miras. Podsiadlowski & Mohamed(2007) made SPH simulations, which show that in a case similar to symbiotic Mi-ras most of the material from the wind should be accreted or remain bound to thesystem, and only very small fraction is ejected to infinity, i.e. , the accretion rate is ∼
100 times higher than the Bondi-Hoyle value and the mass loss will tend to beconfined to the orbital plane.
4. Conclusions
The study of symbiotic Miras is a difficult task because of the presence of anactive accreting companion and the long time-scales of the variations (pulsation ∼ ∼
10 years, orbital period ∼
100 yr). Fortunatelydata and relations estimated for normal Miras allowed us to derive some relevantphysical parameters, such as distances, luminosities and order of mass-loss rates A . A . Table 5: The physical parameters of the symbiotic stars discussed here: distance ( d ), mean magnitude in the K − band outside ofdust obscuration phases, absolute magnitude in K − band ( M K ), average J − K colour ( J − K ), unredded J − K colour derived formcolour-period relation ( ( J − K ) ), interstellar reddening estimated from a 3-D extinction model of Galaxy ( E D B − V ), total reddeningestimated from colour excess ( E B − V ), K -[12] colour and logarithm of mass-loss rate (log ˙ M ). No. Name d K M K J − K ( J − K ) E D B − V E B − V K -[12] log ˙ M (kpc) (mag) ( M ⊙ yr − )010 o Cet 0.11 -2.55 -7.75 1.32 1.40 0.01 - 3.04 -6.2026 RX Pup 1.6 2.80 -8.58 2.36 1.57 0.51 1.51 4.90 -5.2030 V366 Car 2.8 4.40 -8.15 2.07 1.48 0.59 1.13 3.06 -6.2034 BI Cru 2.3 5.02 -7.49 2.58 1.34 0.91 2.36 4.25 -5.5039 SS73 38 4.8 5.34 -8.27 3.09 3.40 0.76 - 3.68 -5.6060 V347 Nor 2.9 4.78 -7.93 2.23 1.44 0.49 1.51 2.83 -6.3069 AS 210 5.6 5.62 -8.11 2.72 2.85 0.27 - 3.43 -5.7105 AS 245 10.2 7.44 -7.89 1.99 1.43 1.32 1.03 4.01 -5.6120 H 2-38 7.2 6.60 -8.01 1.90 1.45 0.82 0.85 3.91 -5.7175 RR Tel 2.5 4.18 -7.97 1.71 1.45 0.04 0.51 3.79 -5.8188 R Aqr 0.24 -1.06 -7.99 1.59 1.45 0.02 0.23 3.38 -6.0s03 StHA 55 3.6 5.27 -7.92 2.99 2.13 0.23 1.64 2.58 -6.1s26 V335 Vul 3.7 5.11 -7.87 1.85 1.95 0.50 - 1.97 -6.4 ol. 58 K − [ ] .This may indicate that they have higher mass-loss rates, or more likely that the dustwhich is being lost by the Miras is trapped within the binary system. Our obser-vations do not settle this issue. Undoubtedly, mass-loss rates and mass transfer inthese systems are most interesting subjects, but more observational data, especiallyin UV, IR and radio bands, are needed to understand it. Acknowledgements.
We are grateful to the following people who made ob-servations from SAAO that were included in this analysis: Greg Roberts, RobinCatchpole, Brian Carter, Dave Laney and Hartmut Winkler. We also would liketo thank Magda Borawska and Anna Lednicka for their work during initial phaseof this project. This study made use of the American Association of VariableStar Observer (AAVSO) International Database contributed by observers world-wide and the public domain databases of The We are grateful to Michael Feastfor his comments on an early draft of this paper. All Sky Automated Survey(ASAS) and The Optical Gravitational Lensing Experiment (OGLE), The Two Mi-cron Sky Survey (2MASS), The MACHO Project (MACHO) which we acknowl-edged. This research has made use of the SIMBAD database, operated at CDS,Strasbourg, France. This work was partly supported by the Polish Research GrantsNo. 1P03D 017 27 and N203 395534.REFERENCES
Alcock et al. in "Robotic Telescopes in the 1990s" , ed. A.V. Fillippenko, ASP Conf. Ser. No.34, p.193.Allen, D.A. 1982, in: The nature of symbiotic stars, Reidel (Dordrecht) , , 27.Belczy´nski, K., Mikołajewska, J., Munari, U., Ivison, R. J. and Friedjung, M. 2000,
A&AS , , 407.Carpenter, J.M. 2001, AJ , , 2851.Carter, B.S. 1990, MNRAS , , 1.Drimmel, R., Cabrera-Lavers, A., and López-Corredoira, M. 2003, A&A , , 205.Feast, M.W., Whitelock, P.A., Catchpole, R.M., Roberts, G., and Carter, B.S. 1983a, MNRAS , ,951.Feast, M.W., Catchpole, R.M., Whitelock, P.A., Carter, B.S., and Roberts, G. 1983b, MNRAS , ,373.Feast, M.W., Whitelock, P.A., and Marang, F. 2003, MNRAS , , 878.Feast, M.W., Whitelock, P.A., and Menzies, J.W. 2006, MNRAS , , 751.Fleischer, A., Gauger, A., and Sedlmayer, E. 1992, A&A , , 321.Glass, I.S. 1999, Handbook of Infrared Astronomy , , CUP.Gromadzki, M., and Mikołajewska, J. 2009,
A&A , , 931.Gromadzki, M., Mikołajewska, J., Whitelock, P.A., and Marang, F. 2007, A&A , , 703.Ireland, M.J., Monnier, J.D., Tuthill, P.G., and et al. ApJ , , 651.Jura, M. 1987, ApJ , , 743.Jura, M. and Helfand, D.J. 1984, ApJ , , 785.Kastner, J.H. and Soker, N. 2004, ApJ , , 1188. A. A.
Kholopov, P.N. 1985,
General Catalogue of Variable Stars , fourth edition, Nauka, Moscow.Kolotilov, E.A., Shenavrin, V.I., Shugarov, S.Yu., and Yudin, B.F. 2003,
Astron. Rep. , , 777.Kotnik-Karuza, D., Friedjung, M., Whitelock, P.A., and et al. A&A , , 503.van Leeuwen F., Feast M.W., Whitelock P.A., and Laney C.D. 2007, MNRAS , , 723.Mikołajewska, J. 1999, in "Optical and Infrared Spectroscopy of Circumstellar Matter" , eds. E.Guenther, B. Stecklum, and S. Klose, ASP Conf. Ser. 188, p.291.Mikołajewska, J. 2001, in "Small-Telescope Astronomy on Global Scale" , eds. B. Paczy´nski, W.P.Chen & C. Lemme, ASP Conf. Ser. 246, p. 167.Mikołajewska, J. 2004, in "Compact Binaries in the Galaxy and Beyond", IAU Coll. No. 194 , ed. G.Tovmassian & E. Sion, Revista Mexicana de Astronomía y Astrofísica, Vol. 20, p.33.Mikołajewska, J. 2007, Balt. Astron. , , 1.Mikołajewska, J., Brandi, S., Hack, W., and et al. MNRAS , , 190.Munari, U., Yudin, B.F., Taranova, O.G., and et al. A&AS , , 383.Munari, U., Siviero, A., Graziani, M., and et al. Balt. Astron. , , 301.Olivier, E.A., Whitelock, P., and Marang, F. 2001, MNRAS , , 490.Posiadlowski, P., and Mohamed, S. 2007 2007, in "Evolution and chemistry of symbiotic star, binarypost-AGB and related objects" , eds. J. Mikołajewska & R. Szczerba, Baltic Astronomy, 16,26.Pojma´nski, G. 2002, Acta Astron. , , 397.Press, W.H. and Rybicki, G.B. 1989, ApJ , , 277.Sobotka, P., Pejcha, O., Šmelcer, L., and Marek, P. 2003, in "Symbiotic Stars Probing StellarEvolution", ASP Conference Proceedings , eds. R.L.M. Corradi, J. Mikołajewska and T.J.Mahoney, p.171.Schmid, H.M. and Schild, H., 2002, A&A , , 117.Taranova, O.G, and Shenavrin, V.I. 2000, Astronomy Letter , , 600.Tatarnikova, A. A., Marrese, P. M., Munari, U., and et al. MNRAS , , 1233.Udalski, A., Kubiak, M., and Szyma´nski, M. 1997, Acta Astron. , , 319.Willson, L.A., Garnavich, P. and Mattei, J.A. 1981, IBVS , , .Whitelock, P.A. 1987, PASP , , 573.Whitelock, P.A. 1988, The Symbiotic Phenomenon, Proceedings of IAU Colloq. 103 , eds. J.Mikołajewska et al. , Kluwer Academic Publishers (ASSL, Vol. 145), p.47.Whitelock, P.A. 2003, in "Symbiotic Stars Probing Stellar Evolution", ASP Conference Proceedings ,eds. R.L.M. Corradi, J. Mikołajewska and T.J. Mahoney, p.41.Whitelock, P.A. & Feast, M.W. 2000,
MNRAS , , 759.Whitelock, P.A., Feast, M., Roberts G., Carter, B. and Catchpole R. M. 1983, MNRAS , , 1207.Whitelock, P.A., Menzies, J., Feast, M., and et al. MNRAS , , 711.Whitelock, P.A., Marang, F. and Feast, M.W. 2000, MNRAS , , 728.Whitelock, P.A., Feast, M.W., Marang, F., and Groenewegen, M.A.T. 2006, MNRAS , , 751.Whitelock, P.A., Feast, M.W., and van Leeuwen, F. 2008, MNRAS , , 313. ol. 58 ASAS 021920-0258.6o Cet 010
ASAS 081412-4142.5RX Pup 026 V m agn i t ude ASAS 095443-5718.9V366 Car 030
JD-2450000
ASAS 122326-6238.3BI Cru 034
ASAS 165120-2600.5AS 210 069
ASAS 200419-5543.6RR Tel 175
ASAS 234349-1517.1R Aqr 188
ASAS 054642+0643.8StHA 55 s03
JD-2450000
ASAS 192314+2427.7V355 Vul s26
Figure 1: ASAS light curves of type-D symbiotic stars.8
A. A. -11.5-11.0 49500 50000 50500 51000 51500
MACHO 105.21287.23 120 I R M JD-2400000
OGLE180601.19-281704.2H 2-38
Figure 2: MACHO (filled circle) and OGLE (open circle) light curve of H 2-38. o l . -3.4-3 < L > o Cet -1.4-1 < J > J - K -3.8-3.4-3-2.6 L -1.8-1.4-1-0.6-0.2 J V i s ua l JD-2440000 < L > RX Pup < J > J - K L J V i s ua l JD-2440000 V i s ua l JD-2440000 V i s ua l JD-2440000 V i s ua l JD-2440000 < L > BI Cru < J > J - K L J
10 11 12 13 14 4000 6000 8000 10000 12000 V i s ua l JD-2440000
Figure 3: Near-IR light curves of o Cet (left panels), RX Pup (middle panels) and BI Cru (right panels). In left panels arrows markmaxima. A . A . < L > V366 Car < J > J - K L J JD-2440000 < L > SS73 38 < J > J - K L J JD-2440000 < L > V347 Nor < J > J - K L J JD-2440000
Figure 4: Near-IR light curves of V366 Car (left panels), SS73 38 (middle panels) and V347 Nor (right panels). o l . < L > AS 210 < J > J - K L J
11 12 13 14 6000 8000 10000 12000 V i s ua l JD-2440000 < L > RR Tel < J > J - K L J V i s ua l JD-2440000 -1.6 < L > R Aqr < J > J - K -2-1 L J V i s ua l JD-2440000
Figure 5: Near-IR light curve of AS 210 (left panels), RR Tel (middle panels) and R Aqr (right panels).2
A. A. V AS 210 P=407 d ASAS 165120-2600.5 V AS 210 P=407 d ASAS 165120-2600.5 V AS 210 P=407 d ASAS 165120-2600.5 -11.5-11.0-10.5 R M MACHO 105.21287.23H 2-38 P=395 d -11.5-11.0-10.5 R M MACHO 105.21287.23H 2-38 P=395 d -11.5-11.0-10.5 R M MACHO 105.21287.23H 2-38 P=395 d V f StHA 55 P=372 d ASAS 054642+0643.8 V f StHA 55 P=372 d ASAS 054642+0643.8 V f StHA 55 P=372 d ASAS 054642+0643.8 P o w e r P o w e r P o w e r f [1/d] Figure 6: Light curves of AS 210, H 2-38 and StHA 55 folded with pulsationperiods (left panels) and related power spectra (right panels). Insights in top rightcorners of power spectra panels show power spectrum of windows. ol. 58 -1 0 1 0.0 0.5 1.0 1.5 2.0 J -1 0 1 0.0 0.5 1.0 1.5 2.0 J -1 0 1 H -1 0 1 H -0.6 0 0.6 K -0.6 0 0.6 K -0.6 0 0.6 0.0 0.5 1.0 1.5 2.0 L f -0.6 0 0.6 0.0 0.5 1.0 1.5 2.0 L f P o w e r P o w e r P o w e r P o w e r f [1/d] Figure 7: Near-IR light curves of SS73 38 folded with pulsation periods (left pan-els) and related power spectra (right panels). Insights in top right corners of powerspectra panels show power spectrum of windows.4
A. A. -2-1.5-1-0.5 0.0 0.5 1.0 1.5 2.0 J o Cet -2-1.5-1-0.5 0.0 0.5 1.0 1.5 2.0 J o Cet -2.5-2-1.5-1 H -2.5-2-1.5-1 H -3-2.5-2 K -3-2.5-2 K -3.5-3-2.5 0.0 0.5 1.0 1.5 2.0 L -3.5-3-2.5 0.0 0.5 1.0 1.5 2.0 L -1 0 1 0.0 0.5 1.0 1.5 2.0 RX Pup -1 0 1 0.0 0.5 1.0 1.5 2.0
RX Pup -1 0 1 -1 0 1 -0.6 0 0.6 -0.6 0 0.6 -0.4 0 0.4 0.0 0.5 1.0 1.5 2.0 -0.4 0 0.4 0.0 0.5 1.0 1.5 2.0 -1 0 1 0.0 0.5 1.0 1.5 2.0
V366 Car -1 0 1 0.0 0.5 1.0 1.5 2.0
V366 Car -1 0 1 -1 0 1 -0.6 0 0.6 -0.6 0 0.6 -0.4 0 0.4 0.0 0.5 1.0 1.5 2.0 -0.4 0 0.4 0.0 0.5 1.0 1.5 2.0
BI Cru
BI Cru -1 0 1 J AS 210 -1 0 1 J AS 210 -1 0 1 H -1 0 1 H -0.6 0 0.6 K -0.6 0 0.6 K -0.6 0 0.6 0.0 0.5 1.0 1.5 2.0 L -0.6 0 0.6 0.0 0.5 1.0 1.5 2.0 L AS 245
AS 245
AS 245
AS 245
AS 245
AS 245
AS 245
Phase
Phase
Phase -1 0 1
RR Tel -1 0 1
RR Tel -1 0 1 -1 0 1 -1 0 1 -1 0 1 -0.6 0 0.6 0.0 0.5 1.0 1.5 2.0 -0.6 0 0.6 0.0 0.5 1.0 1.5 2.0
R Aqr
R Aqr -1 0 -1 0 -1.5-1-0.5 -1.5-1-0.5 -2.5-2-1.5 0.0 0.5 1.0 1.5 2.0 -2.5-2-1.5 0.0 0.5 1.0 1.5 2.0
Figure 8: Near-IR light curves of studied objects folded with pulsation periods.Ephemerides from Table 4 are used for most objects with the exception of o Cetand R Aqr. In case of these object ephemerides are taken from Kholopov 1985. Inpanels of AS 245 dots represent SAAO data, triangle 2MASS, squares DENIS. ol. 58
100 200 300 400 500 600 700 8000510152025303540 O-rich MirasC-rich MirasSymbiotic Miras
Figure 9: The pulsation period distribution for symbiotic Miras together with thosefor normal field Miras (Olivier et al. et al.
A. A. ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 ( J - K ) (K-L) A V =0 A V =3.6 A V =14.5A V =36.4A V = 10 Figure 10: The ( J − K ) vs. ( K − L ) colour-colour diagram. Colours are dereddedusing extinction from galactic 3-D model. Open dots represent oxygen rich sym-biotic Miras during dust obscuration, filled dots outside obscuration. Open squaresrepresent carbon rich symbiotic Miras during dust obscuration, whilst filled outsideobscuration. For comparison we also plot colours of oxygen rich Miras (+) and car-bon rich Miras ( × ). For objects with uncertain nature different symbols were used: o Cet ( H ), AS 245 ( G ). Straight line shows black body colours. Simple model ofstar with shell dust is plot as well. The temperature of star is 2750 K (black body),whilst dust shell temperature is 800 K. This model also includes line blanketing. ol. 58 Figure 11: K − [ ] distribution for symbiotic Miras together with those for nor-mal Miras.8 A. A. K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days] 1 2 3 4 5 6 7 200 300 400 500 600 700 800 K - [ ] P [days]