Multi-band polarimetry of post-asymptotic giant branch stars stars. Optical measurements

0.99 MB·

by  S. Akras

Mon.Not.R.Astron.Soc.000,1–??(0000) Printed17January2017 (MNLATEXstylefilev2.2)
Multi–band polarimetry of post-asymptotic giant branch
stars stars - I. Optical measurements.
7
1 S. Akras1,2⋆, J. C. Ram´ırez V´elez3, N. Nanouris4, G. Ramos–Larios5, J. M. L´opez6,
0
D. Hiriart6, D. Panoglou7
2
1Observat´orio Nacional/MCTIC, Rua Gen. Jos´e Cristino, 77, 20921-400, Rio de Janeiro, Brazil
n
2Observat´orio do Valongo, Universidade Federal do Rio de Janeiro, Ladeira Pedro Antonio 43, 20080-090, Rio de Janeiro, Brazil
a
J 3Instituto de Astronom´ıa, Universidad Nacional Auto´noma de M´exico, Apartado Postal 70-264, M´exico, D.F. 04510, Mexico
4IAASARS, National Observatory of Athens, I. Metaxa & V. Pavlou, Penteli, GR–15236, Athens, Greece
3
5Instituto de Astronom´ıa y Meteorolog´ıa, Av. Vallarta No. 2602, Col. Arcos Vallarta, CP, 44130, Guadalajara, Jalisco, Mexico
1
6Instituto de Astronom´ıa, Universidad Nacional Auto´noma de M´exico, Ensenada 22800, Baja California, Mexico

7Instituto de Astronomia, Geof´sica e Ciˆencias Atmosf´ericas, Universidade de S˜ao Paulo, SP 05508-900, Brazil
]
R
S
. Received**insert**;Accepted **insert**
h
p
-
o ABSTRACT
r
t We present new optical broad–band(UBVRI) aperture polarimetric observations
s
a of 53 post-asymptotic giant branch (AGB) stars selected to exhibit a large near–
[ infrared excess. 24 out of the 53 stars (45% of our sample) are presented for the

first time. A statistical analysis shows four distinctive groups of polarized post-AGB
1
v stars: unpolarized or very lowly polarized (degree of polarization or DoP<1%), lowly
9 polarized (1% < DoP< 4%), moderately polarized (4% < DoP < 8%) and highly
0 polarized (DoP > 8%). 23 out of the 53 (66%) belong to the first group, 10 (19%)
8 to the second, five (9%) to the third and only three (6%) to the last group. Approx-
3 imately, 34% of our sample was found to be unpolarized objects, which is close to
0 the percentageof roundplanetarynebulae. Onaverage,the lowand moderate groups
1. showa wavelength-dependentpolarizationthatincreasestowardsshorterwavelength,
0 implying an intrinsic origin of the polarization, which signifies a Rayleigh-like scat-
7 tering spectrum typical for non-symmetrical envelopes composed principally of small
1 dust grains. The moderately polarized stars exhibit higher K-W3 and W1-W3 colour
: indices compared with the group of lowly polarized stars suggesting a possible rela-
v
i tion between DoP and mass-loss rate. Moreover, they are found to be systematically
X colder (redder in B-V), which may be associated with the condensation process close
r to these stars that results in higher degree of polarization. We also provide evidence
a
that multiple scattering in optically thin polar outflows is the mechanism that gives
high DoP in post-AGB stars with a bipolar or multi-polar envelopes.
Key words: Polarization–stars:circumstellarmatter–stars:binaries–stars:AGB
and post-AGB
1 INTRODUCTION 2006; Boumis et al. 2003, 2006; Sahai et al. 2011; Sabin et
al. 2014).
One of the most challenging problems in the field of post-
asymptotic giant branch (AGB) stars and Planetary Nebu- Post-AGBstarsrepresentthephaseofstellarevolution
lae (PNe), which still continues to intrigue and perplex as- where the circumstellar envelope (CSE) has not expanded
tronomers, is the mechanism responsible for the formation enough to become optically thin and, thus, observations in
ofcomplexmorphologies.Itisknownfromopticalandnear– the optical are impeded. These CSEs are very thick due to
infrared(IR)imagingsurveysthatthemajorityofpost-AGB the large amount of mass ejected during the AGB phase
starsandPNeshowanasphericalgeometry(elliptical,bipo- (10−4–10−5M⊙ yr−1; Bujarrabal et al. 2001). The central
lar, multi–polar; Manchado et al. 1996, 2011; Parker et al. starsevolvetowardshighereffectivetemperatures,reappear-
ing again in the optical range as a consequence of the dilu-
tionoftheCSEonatime-scalethatismainlydependenton
⋆ e-mail:[email protected] theinitial mass of theprogenitor star (Bl¨ocker 1995). How-
2 Akras et al.
ever,therapidevolutionofthemostmassivepost-AGBstars decreasesagainatthePNephase.Thisresultimpliesthata
does not provide enough time for the CSE to expand and sphericallysymmetricAGBwindmaybreaksomewherebe-
becomeopticallythin.Evidenceofthisevolutionarypathis tween the late AGB mass–loss and early post-AGB phases,
provided by young PNe whose very thick envelopes result whereas the dust formation and mass–loss rates are rela-
inextremelyhighinternalextinctionsthatpreventtheirde- tively high (Soker 2000). Moreover, the expansion of CSEs
tection in theoptical range (Ramos-Larios et al. 2012). over time leads to a gradually decrease of dust density and
The PN phase, on the other hand, is the consequent results in a lower degree of polarization. L´opez & Hiriart
result of the interaction between stellar winds from low– (2011a & 2011b) reported a linear relation between the de-
intermediatemassstars(1–8M⊙)duringtheAGBandpost- greeofpolarizationandthemass–lossrateinevolvedcarbon
AGBphases.Thestarishotandluminousenoughtoionize stars.
the gas, resulting in the formation of a colourful nebula.
This complex structure around the central star undergoes Based on polarimetric images of post-AGB and PPNe
spectacular changes in the structure and geometry during withNICMOSontheHubble Space Telescope (HST),Gled-
thetransition from theAGBtothepost-AGB,and, finally, hill et al. (2001) claimed that the degree of linear polar-
to thePNe phase. ization is strongly related to the light scattered from disk–
OpticalandIRimagingsurveysofpost-AGBstarshave likestructures,orientedatdifferentinclinationangles.Later,
also revealed very complex CSEs, with a wide variety of Ueta et al. (2005, 2007) proposed that besides the effect of
shapes and forms (e.g. Trammell et al. 1994; Meixner et orientation,theopticaldepthisalsoacrucialparameterthat
al. 1999; Ueta et al. 2000; Gledhill et al. 2001; Gledhill reflects the polarization characteristics of DUPLEX (dust–
2005; Parthasarathy et al. 2005; Lagadec et al. 2011). It prominent longitudinally extended; elliptical PPNe) and
issuggestedthatthesphericalsymmetryoftheAGBstellar SOLE (star–obvious low–level elongated; toroidal PPNe)
wind breaks somewhere between the very late AGB mass– nebulae, in agreement with the theoretical predictions of
loss phase (Trammell et al. 1994) and the early post-AGB radiative transfer models (Ueta& Meixner 2003).
phase(Gledhill et al. 2001). However,theexact mechanism
responsible for this deviation is still poorly known (see e.g.
Brown et al. (1978) derived a general expression for
Balick & Frank 2002 and reference therein).
the linear polarization in optically thin CSEs around bi-
The presence of an equatorial density enhancement,
nary systems, providing a method that allows us to de-
such as a disk or torus around these stars, has been cred-
termine the geometric characteristics of these CSEs from
ibly proposed by several authors to be the mechanism re-
the observed Stokes parameters. Posterior investigation of
sponsible for the formation of aspherical PNe (e.g. Livio &
the effect of multi-scattering polarization in axisymmetric
Soker1988;Soker&Livio1994;Soker2002;Frank&Black-
geometries, such as equatorial disks, ellipsoidal envelopes
man2004amongothers).Toenabletheformation ofhighly
and polar jets, has shown that multi–scattering results in a
collimated, fast–moving jets or outflows, a large amount of
higherdegreeofpolarization thansinglescatteringin CSEs
orbiting mass is needed, in accordance to the magnetohy-
with high optical depths (Wood et al. 1996a,b). Finally, a
drodynamicmodel ofFrank& Blackman (2004). Thismass
more comprehensive study on the effect of multi–scattering
can be provided through the mass–exchange interaction of
incircumstellardisksaroundbinarystarswasperformedby
closebinarysystems(e.g.Soker&Livio1994;Rasio&Livio
Hoffman et al. (2003).
1996;seealsothereviewbyDeMarco2009).Thepresenceof
these dusty circumstellar disks has been confirmed in some
proto–PNe (PPNe) with highly collimated outflows such as In this paper, we present new optical broad–band and
M2–9(Lykouetal.2011),Mz3(Chesneauetal.2007),Red multi–band (U, V, B, R and I) linear aperture polarimet-
Rectangle and 89 Her(Bujarrabal et al. 2005, 2007). ric measurements of a sample of 53 post-AGB stars, out of
The presence of such a large amount of dust in CSEs which 24 (or 45%) are presented for the first time. The re-
results in: (i) a large near–IR excess in the spectral energy maining29sources(55%)havebeenalreadyobservedeither
distribution (SED) (e.g. De Ruyter et al. 2006), and (ii) in the optical or near-IR wavelengths. Based on the degree
the production of significant amount of linear polarization of polarization in the V-band, we classify the programme
due to stellar light scattering by large (>1 µm) or small stars from unpolarized to highly polarized. This study fol-
(∼0.2 µm) dust grains, and free electrons. The grains size lows thesametechniqueconductedbyBieging et al. (2006)
mayberesponsibleforthedifferentwavelengthdependences andLopez&Hiriart(2011a)toidentifypossiblecorrelations
of polarization (Gledhill & Yates 2003). CSEs have been between the degree of polarization with near-IR colour in-
found to be richer in small grains. dices.
Optical and near–IR polarization imaging along with
spectro–polarimetric observationshavebecomeverypower- InSection2,wepresenttheselectioncriteriaofoursam-
ful techniques for investigating dust properties (i.e. distri- ple, while the observations and data analysis are described
bution and composition), as well as the geometry of CSEs in Section 3. Our results and distinct comments on each
around AGB and post-AGB stars (e.g. Johnson & Jones programme star are presented in Section 4. We discuss our
1991;Parthasarathyetal.2005;Uetaetal.2000,2005,2007; resultsinSection 5andwewrap upwiththeconclusionsin
Meixner et al. 2000; Gledhill et al. 2001; Gledhill 2005). Section6.Itisnoteworthythatthecurrentpaperisthefirst
Studyingthepolarizationcharacteristicsofseveralstars outof asequencefocusing on theoptical polarimetric data.
fromtheredgianttothePNephase,Johnson&Jones(1991) Furthermore, new near–IR (J, H, K) polarimetric observa-
concluded that thedegree of polarization increases through tions will also be obtained and presented in a forthcoming
the transition from the AGB to the post-AGB phase, and paper.
Optical multi–band polarimetry of post-AGB stars 3
2 SAMPLE SELECTION highinstrumentalpolarizationintheUandIfilters,therel-
ative difference between our measurements and those from
Asampleof53post-AGBstarswasselectedfromDeRuyter
the literature of the polarized standard stars is not higher
etal.(2006)andtheTorun´ catalogueofGalacticpost-AGB
than 8%.
and relative objects (Szczerba et al. 2007, 2012) by apply-
The data reduction was performed with the POLIMA
ingthefollowingcriteriatoalltheprogrammestars:(i)they
pipelinedevelopedbyHiriartetal.(2005).Thecodeapplies
exhibit a near–IR excess in their SED profiles, which may
a standard reduction technique which includes removal of
be associated with the thermal emission of hot dust grains
cosmic rays, subtraction of the dark current and bias, as
(Evans 1985) and provides a signature of a dusty circum-
well as flat–field normalization. Multiple flat–field frames
stellar disk/torus (De Ruyter et al. 2006); and ii) they are
weretakenateachofthefourpositionsofthepolarizerand
bright enough at optical wavelengths to perform accurate
for all the filters at dusk and dawn. The flat–field frames
linear polarimetry from the 0.84 m telescope at the Obser-
were combined to form a normalized frame for each filter
vatorioAstron´omicoNacionalinSierradeSanPedroM´artir,
andprismposition.Finally,allindividualscienceframesfor
Mexico.InTable1,welist forthestarsofoursample,their
eachfilterandprismpositionwerebiassubtractedandflat–
coordinates, theirclassification typeand information about
fielded.
their binarity.
Thenormalized Stokesparametersq anduforeachob-
Our sample consists of 29 post-AGB or likely post-
ject were calculated as
AGB stars (three PPNe: Red Rectangle, Frosty Leo and
MinkowskiFootprint),23RVTauristarsandoneyellowhy-
pergiantstar(YHG).Inparticular,theRVTauristarsdefine F(0◦)−F(90◦)
q = , (1)
a subgroup of pulsating post-AGB stars whose IRASfluxes F(0◦)+F(90◦)
are located in a specific area of the IRAS colour–colour di-
and
agram, knownas theRVTauribox(1.0<[12]–[25]<1.55 &
0.2 < [25]–[60] < 1.0; Evans 1985). Among the 53 objects, F(45◦)−F(135◦)
u = , (2)
seven are known binary systems with a range of orbital pe- F(45◦)+F(135◦)
riodsbetween100and1000d(seeDeRuyteretal.2006for
more details), 15 are candidate binary systems, and 6 are where F(φ) is the flux at polarizer position φ. The q
confirmed single stars. and u Stokes parameters are related to the total linear po-
larization pL and thePA of polarization θ by
3 OBSERVATIONS pL =pq2 +u2 , (3)
3.1 Reduction process
and
Linear polarization measurements were performed with the
1 u
0.84m,f/15,telescopeoftheObservatorioAstron´omicoNa- θ= arctan . (4)
2 q
cional in theSierradeSan PedroM´artir, Mexico, using the
POLIMA polarimeter (Hiriart et al. 2005) in conjunction Since we did not perform measurements of thecircular
witha2048×2048 CCDcamera(13×13 µmpixels),which polarization,wewillhereafterrefertothetotallinearpolar-
results in a scale of 0.44 arcsec pixel−1. The polarimeter ization, given byequation (3), as thedegree of polarization
consists of a rotating Glan–Taylor prism driven by a step- (DoP).
permotor. Asthepolarimeter isa single–beam devicewith Given thatPOLIMAis asingle–beam polarimeter, sky
averyslowmodulation,goodphotometricconditionsarere- variations during the exposures may induce false polar-
quiredforaccuratepolarimetry.Theobservationswereper- ized signals. In general, the total flux F is equal to either
formedbetween2012JanuaryandDecemberat/ornearthe F1=F(0◦)+F(90◦) and F2=F(45◦)+F(135◦) (e.g. Shahza-
new Moon phase. The broad-band filters U, B, V, R and I manian et al. 2015). Hence, the difference between the two
were used with exposuretimes from 5 to 180 s. (parallel and perpendicular) fluxes F1 and F2 indicates the
Separateimageswereobtainedforallbandsandatrel- skystabilityduringtheobservations.Thismethodallowsus
ative position angles of the prism of 0◦, 90◦ and 45◦, 135◦, to reject measurements with a high fluxdifference.
in order to reduce the time between the two measurements Fig. 2 shows the DoP as a function of flux difference.
when evaluating each Stokes parameter. The instrumental The upper panel shows the 447 individual measurements,
polarization as well as the rotation angle of the polarime- while the lower panel displays only the measurements with
ter were estimated by observing a number of unpolarized flux difference <5%. The vertical lines indicate flux differ-
and polarized standard stars (Schmidt et al. 1992) during ences from 1% to 3% in steps of 0.5%. Most of the high
our campaigns. In Fig. 1 we show the observations of the DoP measurements (>10%) show a flux difference lower
polarized and unpolarized standard stars. Five unpolarized than 1.0% and only four of them exhibit a flux difference
and six polarized standard stars were observed. The total between 1.0% and 1.5%. Apparently, high DoP values may
number of observations in all bands for the standars stars, be recorded because of a significantly high flux difference
including both polarized and unpolarized ones, is 168. In due to sky variability and mistakenly be considered as real
Table 2, we present the instrumental degree of polarization (Fig. 1, upper panel). Hence, we decide to accept and pub-
and rotation angle for each filter. Our values in the V fil- lish only those measurements with a flux difference lower
ter are in very good agreement with the values presented than 3%, thus excluding 18 of the total 447 measurements
by Lopez & Hiriart (2011a, 2011b). Despite the relatively (∼4% of thewhole sample).
4 Akras et al.
Table1.Logfileoftheobservations.TheIRASnumber,thename,theGalacticcoordinateslandb,theequatorialcoordinatesRAand
Dec,thetype, anoteofbinarityandthereferencefromoursamplearetabulatedbelowinthesameorder.
IRAS Name lll.l±bb.b RA(J2000) Dec(J2000) Type Binarity Reference
Z02229+6208 − 133.7+01.5 022641.7 +622122 post-AGB − −
04166+5719 TWCam 148.3+05.3 042047.6 +572628.5 RVTauri candidate 1,6
04296+3429 GLMP74 166.2-09.1 043256.9 +343612.4 post-AGB no 12
04440+2605 RVTau 174.8-12.2 044706.7 +261045.6 RVTauri candidate 1
05040+4820 LSV+4826 159.8+04.8 050750.3 +482409.4 post-AGB − −
05113+1347 GLMP88 188.9-14.3 051407.7 +135028.3 post-AGB − −
05208-2035 BD-201073 222.8-28.3 052259.4 -203253.0 post-AGB candidate 1,6
05280+3817 V428Aur 170.5+02.5 053126.7 +381910.5 RVTauri − −
05381+1012 GLMP177 195.5-10.6 054057.0 +101425 post-AGB − −
06072+0953 CTOri 199.4-04.6 060957.9 +095231.8 RVTauri candidate 1
06108+2743 SUGem 184.2+04.8 061400.8 +274212.2 RVTauri candidate 1,6
06160-1701 UYCma 224.8-14.9 061816.4 -170234.7 RVTauri candidate 1
06176-1036 RecRectangle 218.9-11.8 061958 -103815 post-AGB yes 2
06338+5333 V382Aur 161.9+19.6 063752.4 +533102 post-AGB yes 3
07008+1050 PSGem 204.7+07.6 070339.6 +104613 post-AGB yes 6
07134+1005 LSV+1015 206.7+10.0 071610.3 +095947.9 post-AGB no 4,12
07331+0021 ALCmi 217.8+09.9 073541.2 +001458 RVTauri − −
07430+1115 GLMP192 208.9+17.1 074551.4 +110819.6 post-AGB − −
08187-1905 V552Pup 240.6+09.8 082057.1 -191503.4 post-AGB − −
09371+1212 FrostyLeo 221.9+42.7 093953.9 115852.3 post-AGB − −
11157+5257 DZUma 150.7+59.1 111833.6 +524054.6 RVTauri − −
11472-0800 AFCrt 277.9+51.6 114948 -081720.5 RvTauri candidate 5
13467-0141 CEVir 330.8+57.8 134917.1 -015544.9 RVTauri − −
17436+5003 V814Her 077.1+30.9 174455.5 +500239.5 post-AGB no 4
F17495+0757 TXOph 033.3+16.8 175201.1 +075629.3 RVTauri − −
17534+2603 89Her 051.4+23.2 175525.2 +260259.9 post-AGB yes 7
18095+2704 V887Her 053.8+20.2 181130.6 +270515.6 post-AGB − −
18123+0511 − 033.4+10.5 181449.4 +051255.7 RVTauri candidate 1
18281+2149 ACHer 050.5+14.2 183016.2 215200.6 RVTauri yes 10
18564-0814 ADAql 026.5-05.4 185908.7 -081014.1 RVTauri candidate 1
19090+3829 EGLyr 069.8+13.0 191048.7 +383425 RVTauri − −
19114+0002 V1427Aql 035.6-04.9 191358.6 +000731.9 YHG − −
19125+0343 − 039.0-03.5 191501.2 +034842.7 RVTauri candidate 1,6
19163+2745 EPLyr 060.7+06.9 191819.5 +275103.2 RVTauri candidate 1,6
19199+3950 HPLyr 072.0+11.7 192139.1 +395608.0 RVTauri yes 8
19343+2926 MinFootprint 064.1+04.3 193618.9 +293250 post-AGB yes 9
19386+0155 V1648Aql 040.5-10.1 194108.3 +020231.2 post-AGB − −
19475+3119 LSII+319 067.2+02.7 194929.6 +312716 post-AGB no 5
19486+1350 TWAql 052.2-06.4 195100.9 +135815.6 RVTauri − −
19500-1709 V5112Sqr 023.9-21.0 195252.7 -170150.3 post-AGB no 5,12
20000+3239 GLMP963 069.8+01.2 200152.5 +324732.9 post-AGB − −
20004+2955 V1027Cyg 067.4-00.4 200227.4 +300425.5 post-AGB − −
20056+1834 QYSqe 058.4-07.5 200754.6 +184254.5 RVTauri candidate 1,6
20117+1634 RSge 057.5-09.8 201403.7 +164335.1 RVTauri candidate 1
20160+2734 AUVul 067.3-04.5 201805.9 +274404 post-AGB − −
20343+2625 VVul 068.8-08.5 203632.0 +263614.5 RVTauri candidate 1
21546+4721 GLMP1047 095.0-05.6 215632.9 +473612.8 post-AGB − −
22023+5249 LSIII+5224 099.3-02.0 220412.3 +530401.4 post-AGB − −
22223+4327 BD+424388 096.8-11.6 222431.4 +434310.9 post-AGB no 5,12
22223+5556 TXOph 103.4-01.0 222413.7 +561133.3 RVTauri − −
22272+5435 V354Lac 103.3-02.5 222910.4 +545106 post-AGB candidate 5
22327-1731 HD213985 043.2-57.1 223527.5 -171526.9 post-AGB yes 2,6
− BD+39◦4926 098.4-16.7 224611.2 +400629.3 post-AGB yes 11
References:(1)DeRuyter etal.(2006), (2)VanWinckeletal.(1995), (3)Hrivnaketal.(2008), (4)Hrivnaketal.(2011), (5)van
Winckeletal.(2012), (6)Gielenetal.(2008), (7)Watersetal.(1993), (8)Kreiner(2004), (9)Castro–Carrizoetal.(2012), (10)Van
Winckeletal.(1998), (11)Kodairaetal.(1970), (12)Reyniers(2002)
Optical multi–band polarimetry of post-AGB stars 5
Table 2. Instrumental degree of polarization, rotation angle of the polarimeter with respect to north and the relative difference in
percentage betweenourmeasurementsofthepolarizedstandardstarsandtheirvaluesintheliterature.
Filter U B V R I
Pinstrumental (%) 2.58±0.14 1.24±0.07 0.51±0.05 0.57±0.05 2.81±0.11
Rotationangle(◦) -87.44±2.2 -93.04±0.62 -92.3±0.67 -91.58±0.69 -91.74±1.06
Rel.dif.(%) 8.07±3.38 6.64±2.21 6.8±1.86 5.87±1.6 7.96±2.15
Filter U Filter B Filter V Filter R Filter I
4.0 4.0 4.0 4.0 4.0
3.5 3.5 3.5 3.5 3.5
%)3.0 3.0 3.0 3.0 3.0
ol. (2.5 2.5 2.5 2.5 2.5
u. p2.0 2.0 2.0 2.0 2.0
nstr1.5 1.5 1.5 1.5 1.5
I1.0 1.0 1.0 1.0 1.0
0.5 0.5 0.5 0.5 0.5
0.0 0.0 0.0 0.0 0.0
−80 −80 −80 −80 −80
g) −85 −85 −85 −85 −85
e
d
hift ( −90 −90 −90 −90 −90
e s
gl
n
A −95 −95 −95 −95 −95
−100 −100 −100 −100 −100
2.0 2.0 2.0 2.0 2.0
1.5 1.5 1.5 1.5 1.5
oP) 1.0 1.0 1.0 1.0 1.0
D
ef ( 0.5 0.5 0.5 0.5 0.5
bs - P_r−00..50 −00..50 −00..50 −00..50 −00..50
P_o−1.0 −1.0 −1.0 −1.0 −1.0
−1.5 −1.5 −1.5 −1.5 −1.5
−2.0 −2.0 −2.0 −2.0 −2.0
60006100620063006400 60006100620063006400 60006100620063006400 60006100620063006400 60006100620063006400
Julian Date Julian Date Julian Date Julian Date Julian Date
Figure 1. The upper panels correspond to the observations of the unpolarized stars, i.e. the instrumental polarization in each band,
whilethemiddlepanelscorrespondtotheangleshiftofthepolarimeterwithrespecttonorth.Inthelowerpanels,weshowthedifference
between our polarization measurements and the values reported by other authors for the standard polarized stars. The error bars in
the upper two panels represent only the obervational error (in blue), while in the lower panels the error bars (in red) represent the
observational plustheinstrumentalerrors(seeTable1).Intheχ-axis,theobservational dates aregiveninJD-2450000.
3.2 Interstellar polarization
q(λ)=Q⋆(λ)+qISM(λ) (5)
As the stellar light propagates through the interstellar
medium (ISM), it is subject to emission, absorption and u(λ)=U⋆(λ)+uISM(λ) (6)
scatteringmechanismsproducedbythemagneticallyaligned
Todisentanglestarlightpolarization fromtheISMcon-
dust grains that further polarize thelight at all wavelength
bands. In the visible (e.g. Reiz & Franco 1998; Fosalba et tamination, the qISM(λ) and uISM(λ) values must be esti-
mated along the line of sight of each programme star. To
al. 2002) and near-IR regimes (e.g. Jones 1989), the inter-
dealwiththeproperestimationforeachofthefiveavailable
stellarpolarization (ISP)showsastrongdependenceonthe
photometricbands,weadopttheempiricallawofSerkowski
interstellarextinction.TheISPdegree-extinctionrelationis
(1971) which models theISPas a function of wavelength:
foundtobealmostlinearinthenearby(d<2kpc)andclose
to the Galactic plane (|b| <10◦) area. As a result, the ob-
λ
saerrevaedcoSmtobkiensaptiaornamofettehresianttrainwsaicveSlteonkgetshpλa,rqa(mλe)taernsdouf(tλh)e, ISP(λ)=pmax,ISM×exp[−Kln2 mλax] ; (7)
object, Q⋆(λ)andU⋆(λ),andthoseinducedfrom theinter- where λmax is the wavelength at which ISP reaches its
stellar medium, qISM(λ) and uISM(λ). Thus, we have: maximum value pmax,ISM, while K is a constant parameter
6 Akras et al.
valuesdeviatingbyatleast4σfromthemedianareexcluded
25 fromtheprocedure.Adistance-dependentweightwj isthen
assigned to the selected stars following the suggestion of
Bastien (1985):
%)
n ( 20
atio wj =1− rj, (8)
ariz 15 rc
Pol with rj being the angular distance from the pro-
of 10 grammestartothejthcatalogueneighbourstar.Finally,the
e distance-weightedaverageoftheStokesparametersaccount-
e
gr ing for all neighbours are calculated through theformulae:
De 5
q = Pnj=1qj· wj (9)
0 10 20 30 40 ISM Pnj=1wj
Fluxes difference (%)
and
uISM= PPnj=1nj=u1jw·jwj , (10)
25
%) where n is the number of neighbour stars that do not
n ( fulfil our statistical criterion for outlier detection. The ISP
atio 20 and the PA towards each of our programme stars are ob-
z tained by employing equations (3) and (4). The intrinsic
olari 15 Stokes parameters, Q⋆ and U⋆, are finally calculated by
of P subtractingtheISPcontribution by meansof equations(5)
e 10 and(6).Insixcases(IRAS06338+5333,IRAS07134+1005,
e
gr IRAS 07430+1115, IRAS 09371+1212, IRAS 11472+0800
De 5 and IRAS 17436+5003), a circle with a radius of 10◦ was
selected, since nostars were found within the circle of 5◦.
0 Oncepmax,ISMandθISMaredefined,thentheISPvalues
0 1 2 3 4 5 intheremainingU,B,RandIbandsarecalculatedthrough
Fluxes difference (%) equation(7)bysettingaseffectivewavelength thevaluesof
360,430,640,and890nm,respectively.Notethatthesame
Figure2.Fluxdifferencebetweentheparallelandperpendicular methodforestimatingtheISPhasbeenusedinmanyother
polarizer angles versus DoP. Upper panel: DOP values for all similarworksaswell(e.g.Yudinetal.2003;Lopez&Hiriart
availablemeasurements(447).Lowerpanel:DoPvalueswithflux 2011a, 2011b; Malatesta et al. 2013).
differencelessthan5%.Theverticallinesindicatefluxdifferences The upper panel in Fig. 3 displays the distribution of
from1%to3%insteps of0.5%(seetextfordetails). ISPintheVband,andrevealsthat90%oftheestimatedval-
ues are less than 1.5%. Additionally, the lower panel shows
the same distribution depicted in a polar diagram with the
that remains constant irrespective of the examined wave- PA set as the angular coordinate and the radial axis repre-
length(e.g.Serkowski1971;Coyneetal.1974; Serkowskiet senting the ISP degree. The majority of the estimated PAs
al. 1975; Codina-Landaberry & Magalhaes 1976). are found to be distributed around 90◦(±15◦), as expected
Therefore,thepropertiesofalinearISPcanbefullyde- fromlarge-scaleISPstatistics(seeFosalbaetal.2002).Only
termined for any wavelength provided the quantities λmax, 16 measurements are found to be out of this range of PAs,
pmax,ISM andθISM are available, and theISPvectorneither most of which have an ISP degree lower than 0.5%. Only
rotateswiththewavelengthnorvariesthroughouttime(PA 10%ofthesemeasurementsexhibitDoPhigherthan1%,in-
may vary if measurements are obtained at different times; dicatinga significant contribution of theISM(upperpanel,
Coyne 1974). We set λmax=544 nm as the average wave- Fig. 3). Moreover, it is worth mentioning that most of the
length at which the ISP is maximized, while K = 1.15 is ISP measurements with DoP higher than 1% are found to
adopted as theoptimal value (Serkowski et al. 1975). exhibit the same PA∼100◦, and ∼60◦. This indicates that
Inanattempttoestimatetheparameterspmax,ISM and the alignment of dust grains in those directions is strong,
θISM towards the programme stars, we apply the star field probably due to the Galactic magnetic field. Conclusively,
methodusing thecatalogue of Heiles(2000). Thecatalogue this analysis provides a boost of confidence that our ISP
lists only stellar polarization measurements in the V band estimates are reliable within theiruncertainty levels.
that are considered to be a firm representative of the max-
imum ISP. The Stokes parameters qj and uj are then de-
termined for each neighbour star j located inside a circle
4 RESULTS
of angular radius rc = 5◦ centred on our programme star.
Extreme values (outliers) are carefully removed from the In thefollowing sections, the results for each object are de-
analysisbyperformingrobuststatistics.Likewise,starswith scribed and compared to previously published works. The
Optical multi–band polarimetry of post-AGB stars 7
tios(S/N)andfurthercorrection isrequired(see,e.g., Sim-
0.0 0.5 1.0 1.5 2.0 2.5 3.0 mons & Stewart 1985; Quinn 2012 and references therein).
The Modified ASyptotic Estimator (MAS; Plaszczynski et
99.5
%] 98 al. 2014) is used for estimating thebias-noise for each mea-
ncy [ 90 sTuarbelmeeBn1t.. TInhethceorarneacltyesdisDoofPeMacAhSivnadliuveisduaarleoablsjoectgivbeenlowin,
e
u 70 the uncorrected DoP for the low S/N bias-noise are used
q
e 50 sincethiscorrection had not beenapplied in previousstud-
r
e f 30 ies.
v
ati 10
mul 2
4.1 Comments on Individual Stars
u 0.5
C
Angular radius = 5 deg
4.1.1 IRAS Z02229+6208
0.01
24 V band The elliptical nebula in IRAS Z02229+6208 is oriented at
PA=30–35◦ with a size of 2.1 × 1.3 arcsec (Ueta et al.
20
s 2000, 2007). High–resolution polarimetric maps, obtained
r
a
st 16 withNICMOSonboardtheHST,haverevealedhighDoPup
of to 50%–60% in the north-western and south-eastern lobes,
er 12 respectively,whereasitislower(30%)alongthedirectionof
b
m PA∼150◦ (Ueta et al. 2005, 2007). Ground–based mid–IR
Nu 8 images have also shown a seemingly spherically symmetric
envelope with a diameter of 4 arcsec (Meixner et al. 1999).
4
TheshiftinPAbetweentheDoPandmid–IRemissionmaps
is attributed to a possible precession motion of the torus
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 (Uetaet al. 2005, 2007) .
Our multi–band polarimetric observations show that
Degree of ISP [%]
the polarization is wavelength-independent within the er-
rorsin contrast tothespectro-polarimetric data from Beig-
ing et al. (2006). On average, our DoP∼7% and PA∼130◦
agreewiththeaveragevaluesderivedbyBiegingetal.(2006)
and aperturepolarimetry (Uetaet al. 2007). Thedifference
between our measurements and those from Bieging et al.
(2006) may be the correction of bias-noise applied to our
data. Despite the very highly polarized lobes (Ueta et al.
2005, 2007), the net aperture DoP of IRAS Z02229+6208
is significantly lower. The reason for this is twofold: (i) the
cancellation of perpendicular polarization vectors, and (ii)
the dilution of polarized flux due to the unpolarized stellar
Figure 3.Top/center panels:DistributionofISPdegreetoward light. IRAS Z02229+6208 is clearly a moderately polarized
theprogrammestars(top:cumulative frequencyofstars;centre: source.
number of stars). The size of DoP bins is 0.5%. Bottom panel:
Distribution of ISP vector parameters inpolar coordinates. The
blue and green circles correspond to DoP values which differ by 4.1.2 IRAS 04166+5719
0.5%.
IRAS04166+5719 or TW Cam is a variable RV Tauristar,
whoseDoPandPAarewavelength-independent.Adecrease
in DoP towards longer wavelengths has been reported by
intrinsic DoPs as well as the PAs for each star are given in Nook, Cardelli & Nordsieck (1990). Our values are signifi-
Table B1, and the plots of the intrinsic DoP as a function cantlyhigherthanthosequotedbyNooketal.(1990).Using
of wavelength (U, B, V, R and I bands) are shown in Ap- the Julian Dates of the observations and the pulsation pe-
pendixA(Fig.A1).TheeffectivewavelengthsfortheU, B, riodofTWCam(85.6d;Evans1985),weconcludethatour
V, R and I bands filters are 0.36, 0.43, 0.54, 0.64 and 0.89 observations were performed near the minimum magnitude
µm, respectively. phase, in contrast to those obtained by Nook et al. (1990).
In the subsequent analysis, we refer to an object This implies that the polarization measurements by Nook
as unpolarized or very lowly polarized for DoP61%, etal. (1990) are probabledilutedbytheunpolarized stellar
lowly polarized for 1%<DoP64%, moderately polarized for light.
4%<DoP68%, and highly polarized for DoP>8%. Some of The interstellar light may also affect the results. Nook
theverylowlypolarizedobjectcannotbedistinguishedfrom et al. (1990) calculated the ISP in the direction of IRAS
the unpolarized ones because of the relatively high uncer- 04166+5719 close to 2.97% (PA=135.8◦) within a circle of
tainty of DoP dueto the interstellar polarization. 10◦,whereasalowervalueof1.8%andPA=99◦isestimated
It should be noted that polarization measurements in this work based on the recently published catalogue of
(DoP=p(u2+q2)) are biased at low signal-to-noise ra- Heiles (2000).
8 Akras et al.
4.1.3 IRAS 04296+3429 4.1.9 IRAS 05381+1012
HST images of IRAS 04296+3429 have revealed a multi– Thefirstpolarimetric observationsofIRAS05381+1012 in-
polarstructure,withtwopairsofhighlycollimatedoutflows dicate it is an unpolarized object (0.1%<DoP<0.25% and
oriented at PAs of ∼25◦ and ∼100◦, respectively (Sahai a constant PA at 175–180◦). Even after the correction of
1999,Uetaetal.2000,2005,2007).Bothoutflowsarefound ISP,IRAS05381+1012exhibitsawavelength-dependentpo-
to be highly polarized (DoP>30%) in contrast to the lowly larization that resembles the Serkwoski’s law profile, indi-
polarized innerregion. cating a possible interstellar origin. The reddening E(B-
Spectro–polarimetricdatahaveshownthattheDoPin- V)=0.29±0.03 mag of IRAS 05381+1012 (Vickers et al.
creases to theblue part of spectrum (Trammell et al. 1994, 2015) implies an ISP of 1.3%, as inferred from the ISM
Oppenheimer et al. 2005). Our data show a roughly con- extinction–polarization relation by Fosalba et al. (2002).
stant DoP∼5% with a possible increase to the B band (see Thisvalueisalmost5timeshigherthanourvalue,estimated
Table B1). This value of ∼5% is close to the aperture po- with thefield star method,namely 0.26% in theV band.It
larization measurement in the near-IR regime (Ueta et al. should be noted that the former equation is an empirical
2007). The PA measurements are in good agreement with formula that provides the ISP as a function of reddening,
both spectro-polarimetric studies, showing a rotation from but the spread of ISP measurements introduce a high un-
175to135◦.Multi–scatteringintheseopticallythinoutflows certainty in the final result. We classify IRAS 05381+1012
may explain thehigh DoP values (Wood et al. 1996a,b). as an unpolarized object.
4.1.10 IRAS 06072+0953
4.1.4 IRAS 04440+2605
IRAS 06072+0953 is a variable RV Tauri star with a pul-
IRAS04440+2605 shows a strong IRexcess in theSED at-
sation period of 67.3 d (Kiss et al. 2007). Its SED exhibits
tributedtohot–dustemission,owingtoacircumstellardisk
a strong IR excess associated with the emission from hot-
aroundabinarysystem(DeRuyteretal.2006).Wepresent
dust grains in the envelope. Based on the SED, De Ruyter
herethefirstpolarimetricobservationsofthispulsatingRV
et al. (2006) claim that IRAS 06072+0953 is a candidate
Tauri star. We find a moderate time variation in DoP and
binary system, surrounded by a dusty disk. However, there
PA, for which a close binary system may be responsible.
is no observational confirmation of the disk or the binary
IRAS04440+2605 is classified as a lowly polarized object .
companion.
We present here the first polarimetric observations of
thisRV-Tauristar.Itwasobservedduringthreenights,with
4.1.5 IRAS 05040+4820
some evidence of possible time variability. However, a con-
A nearly constant DoP∼2% and PA∼140◦ are found for stantDoP∼0.5%andPA∼140◦ seemsverylikely.Inconclu-
IRAS05040+4820.Noindicationoftimevariabilityofpolar- sion,IRAS06072+0953isclassifiedasaverylowlypolarized
ization is found. Our observed values, before the correction objectdespiteitsaxisymmetricCSE.Thissuggeststhatthe
ofbiasnoiseandISP,areinverygoodagreementwiththose direct unpolarized light from the central star significantly
quotedbyParthasarathyetal.(2005).IRAS05040+4820 is dilutes the polarized light from the CSE, and results in a
apparently a lowly polarized object. verylow net DoP.
4.1.11 IRAS 06108+2743
4.1.6 IRAS 05113+1347
TheSEDofIRAS06108+2743showsastrongIR–excess(De
Despite that we observed IRAS 05113+1347 only in the B, Ruyter et al. 2006). The first polarimetric observations of
R and I bands, we are able to deduce some general results. IRAS06108+2743 revealaverylowly polarizedobject with
Its DoP increases towards shorter wavelengths from ∼1.2% DoP<0.5% and 105<PA<140◦. Despite the correction for
intheIbandto∼2.5%intheBband,whereasitsPAvaries ISP, the polarization of IRAS 06108+2743 displays a Serk-
between ∼100◦ and ∼125◦. Our results agree with those woski profile, implyingan interstellar origin. The large red-
derivedfromspectro–polarimetricdata(Biegingetal.2006). deningE(B-V)=0.8±0.3mag(DeRuyteretal.2006) yields
an ISP degree of 2.92±0.88%, which is almost 3-4 times
higherthanourvalue.Thisfurthersupportstheinterstellar
4.1.7 IRAS 05208–2035 origin for its intrinsic polarization. We, therefore, classify
thisobject as an unpolarized one.
Thefirst polarimetric observations of IRAS05208–2035 are
presentedhere.Despitesomepossibletimevariationsinpo-
larization, IRAS 05208–2035 is most probable an unpolar- 4.1.12 IRAS 06160–1701
ized source.
IRAS 06160–1701 was also observed for the first time. It is
foundtobeaverylowlypolarizedorevenunpolarizedobject
with probably a spherically symmetric CSE. Its strong IR
4.1.8 IRAS 05280+3817
excess implies thepresence of a dustydisk around a binary
Wepresentherethefirst polarimetric observations ofIRAS system (De Ruyter et al. 2006). If IRAS 06160–1701 is a
05280+3817. The DoP is from 0.3 to 0.5%. Because of the binary system, it must be seen face–on because of the very
high uncertainties, it is classified as an unpolarized object. low DoP.
Optical multi–band polarimetry of post-AGB stars 9
4.1.13 IRAS 06176–1036 07008+1050 was observed during three nights with no sig-
nificant evidenceof time variability on this time-scale.
TheIRASsource06176–1036 (alsoknownastheRedRect-
angle),isaPPNwithabi–conicalmorphologyandabinary
system (Waelkens et al. 1996). The first indications of an
4.1.16 IRAS 07134+1005
equatorial disk came from the optical images by Ostebart
et al. (1997) and the high-angular-resolution near–IR im- IRAS07134+1005 is a pulsating supergiant star with a pe-
ages by M´ekarnai et al. (1998), which resolved the central riod of 36.8 d (Barth´es et al. 2000). It displays an ellip-
region and unveiled two lobes divided by a thin dark lane. soidal hollow shell with a size of 5×4 arcsec and the ma-
Interferometric CO observations confirmed the presence of jor axis aligned to PA=25◦ (Ueta et al. 2000, 2005). IRAS
amolecular disk in Keplerian rotation perpendicularto the 07134+1005hashighlypolarizedouteredges(averagevalue
axis of thelobes (Jura 1997, Bujarrabal et al. 2003, 2007). of 55±16%) and significantly low-polarization inner region
Spectro–polarimetric data by Schmidt, Cohen & Mar- (DoP<10%; Ueta et al. 2005).
gon (1980) demonstrated that the DoP of this object in- This star was observed during three nights and all ob-
creases towards shorter wavelengths (from ∼6% to ∼2%) servations indicate a very lowly polarized or most likely an
while it has a nearly constant PA at ∼75%. Cohen et al. unpolarized object with DoP∼0.2%. Our values agree with
(1975) showed, however, that the DoP decreases towards thosereportedbyTrammelletal.(1994)andParthasarathy
shorter wavelengths from 1.30%±0.08% in the I band to et al. (2005) but are slightly lower than the value recently
0.97%±0.03% in the U band, and the PA is significantly reportedbyBiegingetal.(2006).Theverylowaperturenet
lower between 28.3◦±2.2◦and 43.3◦±1.4◦. DoP found in IRAS 07134+1005, despite the highly polar-
Our aperture polarization measurements show that izedouteredges,canbeexplainedbythecancellationofthe
IRAS 06176–1036 is a very lowly polarized object with a polarization vectorsduetothecentrosymmetricpattern,as
DoP nearly constant at 1%, whereas the PA rotates from well as thedilution of scattering light from theCSE.
30◦ to 80◦. Our observed values agree with those derived
by Reese & Sitko (1996) of DoP=2.2% and PA=40◦ in the
optical regime.
4.1.17 IRAS 07331+0021
IRAS 07331+0021 is a metal-poor supergiant star
4.1.14 IRAS 06338+5333
(Klochkova & Panchuk 1998). We find some possible
IRAS06338+5333isaknownbinarysystemwithaperiodof evidence of time variations in DoP consistence with
600±2d(Hrivnaketal.,2008)andastrongIRexcess,asthe Parthasarathy et al. (2005). The DoP varies between 0.1%
resultsofthepresenceofadustycircumstellar/circumbinary and 0.8%, whereas the PA increases from ∼125◦ in the U
disk.TheintrinsicDoPshowsaverylowly polarized object band to ∼150◦in theI band.
(0.32<DoP<0.66%). Our values, before the correction for RecentCOmappingdataofIRAS07331+0021revealed
thebiasnoiseandISP,areingoodagreementwithprevious narrowCOlineslikethoseintheRedRectangleand89Her,
studies (Trammell et al. 1994; Parthasarathy et al. 2005). suggesting a possible emission from a dusty Keplerian disk
(Bujarrabal et al. 2013). The presence of a binary system
may be responsible for thevariation in polarization.
4.1.15 IRAS 07008+1050
IRAS 07008+1050 or PS Gem is a metal–poor ([Fe/H]=-
4.8) RV Tauri star (Waelkens et al. 1991). The binary na- 4.1.18 IRAS 07430+1115
tureofallthemetal–poorstarsincludedinoursample,such
IRAS 07430+1115 was observed only one night in the B,
asIRAS06338+5333, IRAS06167-1036, BD +39◦4926 and
V and R bands. The DoP is found to decrease towards
IRAS 07331+00211, is well established (see Table 1). IRAS
longer wavelength from 6% in the B band to 1.77% in the
07008+1050 is, indeed,abinary star with an orbital period
R band, whereas the PA varies from 135 to 145◦. Near–IR
of 1310±8 d (Van Winckel et al. 1999, Gielen et al. 2008).
polarimetry measurementsin theJ(DoP=1.2%, PA=135◦)
Interferometricobservations haveconfirmedthepresenceof
andKbands(DoP=0.8% andPA=135◦)reported byGled-
a compact circumbinary disk with a radius 60.0055 arcsec
hill (2005) are consistent with our wavelength dependence.
(Deroo et al. 2006).
Thepolarized fluximage intheJbandunveilsthepresence
Recent spectroscopic observations shown a DoP up to
of a ring–like structure with a diameter of 0.4 arcsec along
∼1% with possible variability in time (Harrington & Kuhn
the north-east–south-west direction (Gledhill 2005). IRAS
2009).TheDoPincreasestowardshorterwavelengths(from
07430+1115 is classified as a lowly polarized object.
0.4-0.5% in the I band to ∼1.1-1.2% in the U band). Our
values are very close to those reported by Parthasarathy
et al. (2005). The PA is also found to vary between 5◦
and 60◦. The high DoP towards the blue part of the spec- 4.1.19 IRAS 08187–1905
trum supports an intrinsic origin of the polarization. IRAS We present the first broad-band polarimetric observations
for this pulsating post-AGB star. Both DoP and PA (see
1 The binary nature of this metal–poor star has not been con- TableB1)showtimevariationsthatmaybeassociatedwith
firmedyet.However,therecentdetectionofnarrowCOlines(Bu- the pulsation of the star like IRAS 07331+0021. Further
jarrabal et al. 2013) suggests the presence of a Keplerian disk observations are required. The wavelength dependence of
whichisattributedtoabinarysystem. bothDoPandPAsuggestanintrinsicoriginofpolarization.
10 Akras et al.
4.1.20 IRAS 09371+1212 We find a wavelength-dependent DoP that gradually
increases from 2.8 to 5%, and a PA that rotates from 165◦
IRAS 09371+1212, also known as Frosty Leo, is a well– to135◦.Bothagreewiththenear–IRpolarimetricmeasure-
known and studied PPN. Optical and near-IR images dis-
mentsofGledhill (2005). TheISPof IRAS11472–0800 was
play a complex bipolar or multi-polar structure (Langill, estimatedtobe0.15%withincirclesof5◦ and10◦.Thisfur-
Kwok&Hrivnak1994;Scarrott&Scarrott1994;Roddieret
thersupportsthepreviousestimationofISPbyMathewson
al.1995;Sahaietal.2000)alignedatPA∼110◦,andanequa-
& Ford (1970).
torial density enhancement seen almost edge–on. Roddier
et al. (1995) claimed the detection of a binary companion
at 0.18 arcsec from the central star. However, the identical 4.1.23 IRAS 13467–0141
spectral typeof both stars makes this detection less proba-
blyandthebinarityofFrostyLeostillremainscontroversial. The first aperture polarimetric data of IRAS13467–0141 is
Moleculargashasalsobeendetectedinasmallinnerregion presented here. A constant DoP at 0.5% and PA between
with an angular radius less than 6 arcsec, expanding with 70◦ and 87◦ are found. IRAS13467–0141 is assigned to the
velocitiesupto50kms−1(Sahaietal.2000;Castro–Carrizo group of very lowly polarized sources.
et al. 2005).
Opticalandnear–IRpolarimetricobservationsofFrosty
4.1.24 IRAS 17436+5003
Leo were obtained first by Scarrott & Scarrott (1994) and
more recently by Murakawa et al. (2008). A high DoP up IRAS 17436+5003 has been extensively studied by sev-
to 60% has been found mainly in the outer blobs. Accord- eral groups (Joshi et al. 1987; Trammell et al. 1994; Ueta
ingtotheMiescatteringtheory,Scarrott&Scarrott(1994) et al. 2000, Gledhill et al. 2001; Gledhill & Yates 2003,
argued that dust grains with radii α ∼0.1µm and a steep Parthasarathy et al.2005). HSToptical images (Uetaetal.
grain-sizedistributionarerequiredtoexplaintheirpolariza- 2000)andnear-IRpolarimetricimages(Gledhilletal.2001)
tion results. Although,Murakawa et al. (2008) were ableto displayanellipticalenvelopealignedatPA=10◦.Itisfound
reproducetheobservedDoPconsideringdifferentsizedistri- to be more polarized along the north–south direction than
butionsforthedisk(0.005µm6α62.0µm)andthebipolar along the west–east direction, with a maximum DoP value
lobes (0.005µm 6α60.7µm). of 20% in both J and K bands (Gledhill et al. 2001).
IRAS 09371+1212 is one of the three highly polarized Our observed values, before the interstellar and bias-
objectsinoursample.Wefindawavelength–dependentDoP noise correction, agree with those quoted by Parthasarathy
thatincreasestowardslongerwavelengthsfrom11%to19% et al. (2005). The field star method gives an ISP degree of
in the B and I bands, respectively, which is consistent with 1.72% in the V band, which does not agree with the very
scattering bysmall dustgrains (Murakawaet al 2008). The low reddening of this object (0.03±0.01 mag; Vickers et al.
PA rotates from 55◦ to 75◦, almost perpendicular to the 2015). This reddening value yields ISP= 0.2%. The correc-
symmetricaxisofthelobes.Thisisconsistentwithmultiple tion for the bias noise has resulted in a DoP∼0.4-0.5% (see
scattering of photons in optically thin regions. TableB1).IRAS17436+5003 isclassifiedasanunpolarized
object despite thevalue of 0.67±0.38% in theB band.
4.1.21 IRAS 11157+5257
4.1.25 IRAS F17495+0757
IRAS11157+5257orDZUmaisclassifiedasaRVTauristar
(Szczerba et al. 2007). The first polarization measurements Thefirstpolarimetric measurementsofIRASF17495+0757
ofthisRVTauristarshowanunpolarizedobject(seeTable showaconstantDoPandPAwithintheerrors(DoP∼0.3%,
B1). PA∼80◦). IRAS F17495+0757 is considered as an unpolar-
ized source.
4.1.22 IRAS 11472–0800
4.1.26 IRAS 17534+2603
IRAS 11472–0800 or AF Crt, is a pulsating RV Tauri star
withaperiodof31.16d(Kissetal.2007,VanWinckeletal. IRAS 17534+2603 (or 89 Her) is a pulsating star, exten-
2012). The presence of a binary companion with an orbital sivelystudiedoverthelastdecade.TheCOemissionreveals
periodof14yrorlongerhasbeenproposedbyVanWinckel anhourglassmorphology,seenalmost pole–on(15◦ withre-
etal.(2012),whichisapproximately2timeslongerthanthe spect to the line of sight) expanding with velocities of 5–
maximumorbitalperiodsinbinarypost-AGBstars(2000d; 10 km s−1, and a more compact Keplerian rotating disk
Van Winckelet al. 2009). (Bujarrabal et al. 2007, 2013). This disk is consistent with
Near–IR polarimetric images of IRAS 11472–0800 thestrongIRexcessintheSED(DeRuyteretal.2006)and
present a highly polarized source with an average DoP of thebinary central star (Waters et al. 1993).
∼6% in both J and K bands and maximum values of 10% Its DoP measurements indicate an unpolarized source
and7.7%,respectively(Gledhill2005).Thepolarizationvec- (seeTableB1).Inparticular,weestimatetheDoPbetween
torsarefoundtobeperfectlyalignedalongthenorth-west– 0.1and0.2%.Thisvalueislowerthanthosequotedfrompre-
south-east direction (PA=135◦), perpendicular to the ax- viouspolarimetricandspectro–polarimetricstudies(Joshiet
isymmetric structure seen in the K band image (Gledhill al. 1987; Trammell et al. 1994; Parthasarathy et al. 2005).
2005). This high degree of alignment may be associated However,itshouldbenotedthattheDoPsderivedbyJoshi
with multiple scattering along the polar–axis (Wood et al. et al. (1987) and Parthasarathy et al. (2005) are not cor-
1996a,b; Hoffman et al. 2003). rectedfortheISM,whichwefoundtobe0.33±0.77% inthe