Time Variability of Molecular Line Emission in

IRC+10216

David Teyssier, ESA-ESAC, Spain

J. Cernicharo, G. Quintana-Lacaci, M. Agúndez, M. Barlow, E. de Beck, F. Daniel, L. Decin, P. Garcia-Lario, M. Groenewegen, D. Neufeld, J. Pearson

IRC+10216 •  IRC+10216 (CW Leo) is the nearest carbon AGB star (123 pc –

Groenewegen et al. 2012) •  Highly obscured at low wavelengths but uniquely bright in the

FIR and millimetre lines, it has become the archetype star for study of mass-loss and envelope evolution on the AGB

•  Presents a complex and multiple shell structure (see Homan’s talk)

•  Its circumstellar shell has revealed a remarquable chemical laboratory with more than 70 molecules discovered to date, (e.g. Cernicharo et al. 2000)

•  Virtually observed at any available wavelength and spatial/spectral resolution

Decin et al. 2011

Blue: VLT FORS1 V-band (Leão et al. 2006), Green: PACS 70 µm, Red: PACS 100 µm

3.4’ × 3.4’ FoV

See also posters S1-10, S1-14 for more recent results

IRC+10216 as a Mira •  IRC+10216, like most luminous carbon stars, is strongly variable

in the optical and NIR •  Fitted light-curves from various studies indicative of periods in

the range 630-650 days •  Most recent NIR fit places max.

of light on JD = 2 454 554 (23 Mar 2008) – Menten et al. 2012

•  FIR photometry monitoring gives similar period, with a certain time lag (Groenewegen et al. 2012)

Alksnis et al. 1989

Menten et al. 2012

P = 630 ± 4 d P = 639 ± 4 d

Groenewegen et al. 2012

SPIRE 250 µm

I(0.81)

Thermal line variability in IRC+10216 •  Only limited attempts to study line intensity variation over time •  Some lines known to be masering have significant modulation

(e.g. H2O, OH esp. in the cm, SiS, SiO in the mm) •  Carlström et al. (1989) reported

variation of two SiS v=0 rotational lines around 3 mm

•  They however note the large impact of calibration uncertainty

•  Cernicharo et al. (2000): no noticeable variability in their 2 mm survey within the 20% calibration accuracy (IRAM 30m)

•  Overall this lead to the general assumption that most thermal lines in AGB are non-variable, and their use as secondary calibrators in many mm/submm facilities

Carlström et al. 1989

Thermal line monitoring with Herschel (2) •  Project started almost “by accident” •  Overlapping spectral regions taken ½ yr apart in two projects

revealed unexpectedly “anomalous” intensities in some lines

Black: spectral survey (May 2010) Cernicharo et al.

Red: Hydrides search (Dec 2010) Agúndez et al.

Thermal line monitoring with Herschel (1) •  A total of five dedicated observing slots were conducted with

HIFI/SPIRE/PACS, typically separated by 6 months •  Combined to additional data (e.g. calibration) this leads to 7, 8

and 7 epochs resp. for HIFI/SPIRE and PACS

Analysis method •  Line and continuum signals are separated for each instrument •  Integrated line intensities computed for each available HIFI lines

(~90), and selected ladders in SPIRE and PACS (12CO, 13CO, HCN v=0 and v2, CS, H2O, etc) – unresolved lines can suffer blending

•  Light-curves are fitted as cosine with four free parameters (amplitude, mean level, period, time of maximum light)

•  Intensity series within instrumental uncertainty considered “flat”

Examples SPIRE – CS lines

PACS – HCN v=0 + v2=1 blend

Some highlights - continuum •  Continuum measured from the spectrometers shows clear

modulation, similar to that of photometer (Groenewegen+2012) •  Periods typically in the range 620-700 days •  Modulation amplitude (Fmax/Fmin) tends to increase with freq.

PACS

SPIRE

Some highlights – HCN vibrational modes •  HCN intensity variations observed in v=0, v1=1, v2=1 and v2=2

(overtone) and v3=1 states in several rotational levels •  Amplitude of the modulation tends to scale with energy levels

and is particularly enhanced in the vibrationally excited states

ν1=1: 3311 cm-1 (3 µm)

ν2=1: 713 cm-1 (14 µm) ν2=2: 1426 cm-1 (7 µm)

ν3=1: 2097 cm-1 (4.8 µm)

Maser !

Some highlights – IR pumping

•  de Beck et al. (2012) showed that the inclusion of the first three vibrational modes was crucial to model the HIFI lines intensities

•  This radiative pumping occurs in Infrared bands (3-27 µm), where the star flux already experiences large variations

•  Interestingly the 1st stretching mode v1 is coupled to the 1st electronically excited state (A2Π), lying at 2.5 µm, where the star flux could have its largest amplitude modulation

de Beck et al. 2012

•  CCH shows the largest amplitude modulation of the line sample (N=7-6 and 8-7, in excess of 10)

Some highlights – CO variability (1) •  12CO and 13CO are cases where a transition between a non-variable

and variable regime takes place •  The switch occurs in a smooth transition layer that spreads over

several energy levels – no regime “jump” from one level to another •  From the HIFI perspective, J=10-9 appears “flat” while J=16-15

seems variable (no transition measured in-between with HIFI) •  From the SPIRE/PACS perspective, transitions above J=13-12

could be already modulated but calibration uncertainties are similar to modulation amplitude

•  Impact on mass-loss rate estimates based on 12CO should therefore be limited if highly excited transitions are not included in the modelling

•  PACS lines from J=28 onwards (no data between J=19 and 27) vary with modulation factors increasing from ~1.8 to ~3.2

Some highlights – CO variability (2) H

IFI –

12C

O li

nes

12CO J=10-9

12CO J=16-15

12CO J=5-4

HCN v1=1 blend ?

PACS – 12CO lines

SPIRE – 12CO lines

HCN v=0 blend

Conclusions •  IRC+10216 shows strong variation (10-15% up to >1000%) in

molecular emission of lines such as CCH, HCN, HNC, CS, SiO, SiS or H2O. No variation observed in e.g. SiC2

•  12CO/13CO is non-variable (within calibration uncertainty) at low-to-intermediate transitions, but strongly modulated (factor up to 3) at high energy levels

•  Radiative transfer in molecular lines affected by IR-pumping (typically any high-J line) needs to account for time dependency. However assessment of phase per molecule not trivial

•  Coming up soon: •  Study of variation time lag between transitions and between

blue- and red-shifted portions of the line profiles (HIFI) •  Constrast Herschel results with monitoring currently on-going

at IRAM 30-m in similar species •  How unique is IRC+10216 in that respect: can we detect such

variation in other sources ?

Stay tuned !

Summary Species Variability Amplitude factor

12CO Only for high-J 2-3 13CO Only for high-J 5 HCN v=0 Any transition 1.2-4 HCN v1,v2,v3 Any transition 1.5-5

H13CN v=0 Any transition 1.2-1.5 CS Amy transition 1.2-1.5 SiO Any transition 1.2-1.6 SiS Most but not all ! 1.2-1.6 CCH Any transition 4-14 H2O Any transition 1.5-2 HNC Any transition 1.5-1.7 SiC2 NO N/A