Magnetic Channeling and Confinement in the Hypersonic Winds of Massive, Luminous Stars

43
Magnetic Channeling and Confinement in the Hypersonic Winds of Massive, Luminous Stars Rich Townsend University of Delaware Stan Owocki Asif ud-Doula Mary Oksala David Cohen Marc Gagné Steve Skinner ESO Chile ̶ November 2006

description

Rich Townsend University of Delaware Stan Owocki Asif ud-Doula Mary Oksala David Cohen Marc Gagné Steve Skinner. Magnetic Channeling and Confinement in the Hypersonic Winds of Massive, Luminous Stars. ESO Chile ̶ November 2006. Overview. - σ Ori E. - θ 1 Ori C. Luminous. - PowerPoint PPT Presentation

Transcript of Magnetic Channeling and Confinement in the Hypersonic Winds of Massive, Luminous Stars

Magnetic Channeling and Confinement in the Hypersonic

Winds of Massive, Luminous Stars

Rich TownsendUniversity of Delaware

Stan OwockiAsif ud-DoulaMary OksalaDavid CohenMarc GagnéSteve Skinner

ESO Chile ) November 2006

Overview

- σ Ori E

- θ1 Ori C

Luminous

Radiation-driven wind

CNO cycle

Convective core

Radiative envelope

197820022005

Nomagnetic field*?

Plasma Physics Stellar Evolution

Compact Objects

Circumstellar Environment

θ1 Orionis C

Feigelson et al. (2002)

θ Ori - Chandra ACIS-S

- θ1 Ori C (O5.5)

Wojdowski & Schulz (2005)

107 K / 1 keV / 1,000 km s-1

θ1 Ori C

ζ Pup

Babel & Montmerle (1997)

Magnetically Confined Wind Shock

ud-Doula & Owocki (unpub.)

ZEUS - Density

ud-Doula & Owocki (unpub.)

ZEUS - Temperature

Gagné et al. (2005)

107 K / 1 keV / 1,000 km s-1

ZEUS - Density ZEUS - Temperature

ud-Doula & Owocki (unpub.)

Magnetic Confinement

ηη = B = B22/8/8ππ / / ρρvv22/2/2

ηη = [B = [B**RR**/Mv/Mv∞∞] (r/R] (r/R**)) 2 2 -4.

η*

rrAA/R/R** = = ηη**1/4

ZEUS - Density ZEUS - Temperature

ud-Doula & Owocki (unpub.)

!Field Strength

Rotation RateMass-Loss Rate

σ Orionis E

Skinner et al. (2004)

- σ Ori AB (O9.5 + B0.5)

- σ Ori E (B2)

σ Ori D (B2) -

σ Ori - Chandra ACIS-S

Chandra X-ray Imaging and Spectroscopy of the YoungSigma Orionis Cluster and its Massive Central OB Stars

Ste ve Skinner 1 , David Cohen 2 Marc Gagné 3 , Stan Owoc ki 4, Rich To wnsend 4

1Univ. of Colorado(USA)2Swarthmore College (USA)3WestChester Univ. (USA)

4Bartol Research Inst./Univ. of Delaware (USA)

AbstractWe present an overvie w of a recent Chandra X-ray observation of the young Orionis cluster (d 350 pc; age 2 - 5 Myr), including high-resolution gratingspectra of the massive central O-type star Ori A. Chandra detected soft emission from Ori A that is lik ely of wind origin, but much harder emission indicativ e ofmagnetic processes was detected in the unusual B2Vp star Ori E. We use these results toplace constraints on the or igin of X-rayemission in hot stars,which iscurrently not well-understood.

1 The OrionisClusterROSAT observations near the belt of Or ion led to the dis-

covery of a remarkable X-ray clustercentered on the multiplestar Or ionis (Walter et al. 1997). Thismassive stellarsystemis a quintupletdominated by the visual binary Ori AB, con-sistingof a O9.5V star with a B-type companionat an offsetof 0.2500. The Hipparcos distanceof Or i A is 352 pc. The

Or i ABsystem is a luminous X-ray sourcebut the close paircannotbe resolved with existingX-ray telescopes(Fig. 1). TheB-type stars Or i C and D lie 11.200 and 12.900 away from Or iAB, but they emit little or no X-ray emission. In contrast, theunusualmagnetic B star Or i E lying4200 away is a hard var i-able X-ray source(Fig 1.). Its distanceis uncertain and it mayliein the background at d = 640 pc (Hunger et al. 1989). Sur-rounding this massive quintuplet is a large populationof X-rayemittinglow-mass pre-main sequencestars and several browndwarf candidates(Adams-Wolk et al. 2003).

The Or ionis clusterprovidesan excellenttarget for study-ingX-ra y proper tiesof a diverse population of young stars withages of a few millionyears across the entire stellar massspec-trum. Most importantly, the dust extinctiontoward the regionis quite low (AV 0.2 mag; Lee 1968), providingaccess tosofter X-rayswhich may originatein the shocked stellar windsof massive O stars.

2 ChandraObservationChandra observed the Or ionis clusterfor 91 ksecon 12-

13 August 2003 using the ACIS-S detector and the High En-ergy TransmissionGrating (HETG). Pointingwas centered on

Or i AB. Data were analyzed using standard CIAO threads .Our pr imary objective was to obtain high qualityspectraof themassive stars in the Or i AB region needed to test hot starX-ray emissionmodels. We obtaineda grating spectrum of thebrightestX-ra y source Or i AB, and 0th order spectraof sev-eral other nearby sources includingthe magnetic B star Or iE and low-mass pre-main sequence stars in the surroundingcluster. Figure 1 shows the inner portion of the Chandra im-age.

FIG URE 1: Chandra ACIS-S 0th order image of the core of the Orioniscluster. Boxes enclose the optical positions of the B stars Ori C,D.

Or i AB Ori E

Sp. Type O9.5V B2VpV (mag) 3.80 6.66Dist. (pc) 352 370 - 640 (uncer tain)v1 (km/s) 1100 420Binary yes (0.2500) ?log LX (ergs/s) 31.6 30.7 - 31.2

3 X-ray VariabilityFigure 2 shows the 0th order Chandra light curves of Or i

ABand Or i E. X-ray flares have previouslybeen detectedinOr i E (Groote & Schmitt2004; Sanz-Forcadaet al. 2004), but

no large-amplitudevar iabilitywas detected by Chandra. How-ever, low-level var iability is likely and a Kolmogorov-Smirnovtest gives a variabilityprobabilityP(var) = 0.995 for Or i E. In-terestingly, the minimumin the Or i E light curve (arrow in Fig.2) occursat phase = 0.75 (Reiners et al. 2000 ephemeris),close to = 0.8 where the HeI 4713 line reaches maximumequivalentwidth.

FIG URE 2: Chandra 0th order light curves of Ori AB and Ori E. Ar-row marks the minimum at phase = 0.75 in Ori E, which occurs near thephase where HeI equivalent width is maximum.

4 Spectraof OriAB and Ori EFigure 3 compares the 0th order X-ra y spectra of Or i AB

and Or i E. Although Or i AB is brighter in X-rays, itsspec-trum is much softer with little or no emissionseen above 2keV. Spectral fits with thermal plasmamodels give characteris-tic temperatures of kT 0.2 - 0.3 keV. In contrast, theemissionof Or i E is fainter, but much harder. Acceptable spectralfitsrequire twotemperature componentsat kTcool 0.7keV andkThot 2.2 - 3.8 keV.

FIG URE 3: Chandra 0th order spectra of Ori AB and Ori E.

The grating spectrum of Or i AB (Fig. 4) shows very lit-tle absorption and is dominated by low-temperature lines withmaximum power temperatures log Tmax 7.0. The Mg XII(Ly ) line is absent, confirming that the plasma is cool. Theweaker lines are narrow, but excessbroadeningis detected insome stronger lines suchas O VIII Ly (Fig. 5). We find nosignificantline Dopplershifts in the MEG spectrum. Forbiddenlinesin He-liketriplets are weak, which is mostlikely due to thestrong UV radiationfield of the O star.

FIG URE 4: Chandra 1st order MEGspectrum of Ori AB.

5 ConstraintsonWind Shock ModelsTraditionally, the X-ray emissionfrom hot stars has been at-

tributed to shocks that form as a result of instabilitiesin theirpowerful winds. This standard model predicts soft X-rays (kT<< 1 keV) and wind-broadened emission lines, which shouldbe asymmetric if the wind is optically thick (Owoc ki & Cohen2001). However, these predictions have generally not beenconfirmed, with the O4 supergiant Pup being a notable ex-ception. Specifically, we detect narrow emission lines in Or iAB, with onlyslight broadening seenin the br ightestlines (Fig.5), and no strong asymmetries or significantcentroid shifts.The weak broadeningand suppressed forbidden lines suggestthat the lines of Ori AB are formed in the wind accelera-tion zone within a few radii of the photosphere where the windspeed is well belowv1 and the stellar UV fieldis strong.

Although Or i E was not br ight enough for a grat-ing spectrum to be obtained, the hot plasma in its 0th or-der spectrum is clearly incompatible with the standard windshock model. Its emissioncould be intrinsic to the B-staritself, or due to an as yet undetected late-type compan-ion. Collaborators S. Owocki, R. Townsend, and A. ud-Doulaare actively developing rigidly rotating magnetosphere mod-els for Or i E and are investigating the hypothesis thatcentrifugal breakout of a magnetically confined wind maybe responsible for the X-ray flares detected previously (seehttp://www.star.ucl.ac.uk/ rhdt/research/magnetic).

0.01

0.02

0.03

0.04

Cou

nt R

ate

(cou

nts

s-1

Å-1

)

18.80 18.85 18.90 18.95 19.00 19.05 19.10Wavelength (Å)

-0.02

0.00

0.02

FIG URE 5: Chandra MEG1profile of the O VIII line. Over lays are a best-fitGaussian of half-width (HW) 250 km/s (red dotted) - which is v1 /4, aline with no excess broadening (green dashed), and a line with HW = 1000km/s (b lue dash-dot) as might be expected from a wind-shoc k source suchas the O4 supergiant Pup. Bottom: best-fit residuals .

Ref erences & Acknowledgment

A dams-W olk, N. et al. 2003, B A A S, 35(5), 1324Groote, D. & S chmitt, J . 2004, A & A , 418, 235Hunger , K ., Heber , U .,& Groote, D . 1989, A & A , 224, 57Lee, T.A . 1968, A pJ, 152, 1968Owoc ki, S. & Cohen, D . 2001, A pJ, 559, 1108Reiners, A . et al. 2000, A & A , 363, 585S anz-F orcada, J ., Franciosini, E ., & Pallavicini, R. 2004, A & A , in pressW alter, F et al. 1997, Mem. S.A . It., 68, 1081

This work was suppor ted by S AO grant GO3-4007A .

σ Ori AB

σ Ori E

107 K / 1 keV / 1,000 km s-1

Skinner et al. (2004)

σ Ori - Chandra ACIS-S

Hesser et al. (1977)Landstreet & Borra (1978)

Bohlender et al. (1987)

Kaufer et al. (1999)

“ two more or less isolated clouds … near the

intersection of the magnetic and rotational

equator ”

Groote & Hunger (1982)

Field lines are almost rigidField lines are almost rigid

Rigidly Rotating Magnetosphere

Townsend & Owocki (2005)

Townsend, Owocki & Groote (2005)

Townsend, Owocki & Groote (2005)

Townsend, Owocki & ud-Doula (in prep.)

Rigid Field Hydrodynamics

Townsend, Owocki & ud-Doula (in prep.)

Townsend, Owocki & ud-Doula (in prep.)

Summary

• New approaches for modeling magnetic confinement in massive stars – MHD– RRM– RF-HD

• Rewarding success in reproducing optical & X-ray properties

Ongoing & Future Work

• Confrontation between RF-HD and UV, X-ray (& interferometric) observations

• Hybrid RF-HD/MHD models to simulate centrifugal breakout

• Laboratories for understanding magnetic-radiation-rotation interaction

• Synergies with other classes of magnetic object – Polar, AB Dor, Magnetar

http://shayol.bartol.udel.edu/massivewiki/

Sanz-Forcada et al. (2004)

Centrifugal Breakout

ud-Doula, Townsend & Owocki (2006)

ZEUS - Density

ud-Doula, Townsend & Owocki (2006)

ZEUS - Temperature