Radiation-driven acceleration in the expanding WR140 dust shell

2025-04-24 0 0 8.83MB 14 页 10玖币

侵权投诉

Nature | Vol 610 | 13 October 2022 | 269

Article

Radiation-driven acceleration in the

expanding WR140 dust shell

Yinuo Han1,2 ✉, Peter G. Tuthill2, Ryan M. Lau3,4 & Anthony Soulain2,5

The Wolf–Rayet (WR) binary system WR140 is a close (0.9–16.7mas; ref. 1) binary star

consisting of an O5 primary and WC7 companion2 and is known as the archetype of

episodic dust-producing WRs. Dust in WR binaries is known to form in a conned

stream originating from the collision of the two stellar winds, with orbital motion of

the binary sculpting the large-scale dust structure into arcs as dust is swept radially

outwards. It is understood that sensitive conditions required for dust production in

WR140 are only met around periastron when the two stars are suciently close2–4.

Here we present multiepoch imagery of the circumstellar dust shell of WR140. We

constructed geometric models that closely trace the expansion of the intricately

structured dust plume, showing that complex eects induced by orbital modulation

may result in a ‘Goldilocks zone’ for dust production. We nd that the expansion of the

dust plume cannot be reproduced under the assumption of a simple uniform-speed

outow, nding instead the dust to be accelerating. This constitutes a direct

kinematic record of dust motion under acceleration by radiation pressure and further

highlights the complexity of the physical conditions in colliding-wind binaries.

The WR140 was observed on six occasions between 2001 and 2017 with

the near-infrared cameras NIRC and NIRC2 at the Keck Observatory.

The near-infrared imagery, displayed in Fig.1, spans orbital phases

between ϕ=0.043 and 0.592, clearly showing an expanding dust shell

with an evolving apparent morphology. Although the images span

two orbital cycles, features appear to be consistent, implying a high

degree of cycle-to-cycle replication of the same underlying morphol-

ogy. Prominent structures at earlier orbital phases (≤0.111) include an

eastern and western dust arc. As the dust shell expands, these structures

give rise to an ‘eastern arm’ and a prominent ‘southern bar’ at later

orbital phases, following the nomenclature of structures identified

in previous imaging4.

The well-resolved detailed form of the plume motivates the con-

struction of a geometric model to explain the structural variation and

expansion over time. We modelled the geometry of the dust plume

as a linearly expanding spiral based on the ‘pinwheel mechanism’

5,6

We assumed that dust production turns on and off episodically as

the binary approaches, and departs the region near periastron. The

wind-collision region can be approximated as the surface of a cone at

large distances from the stars where the velocity of the compressed

wind has reached its asymptotic value7. Dust assumed to form on this

conical interface between the WR star and the O-star primary subse-

quently expands radially outwards as the binary continues in its orbit.

Originally modelled on WR104 (ref.

), such pinwheel models have been

shown to accurately reproduce dust structures in several Wolf–Rayet

(WR) binaries including Apep8 and WR112 (ref. 9).

The orbital parameters of WR140 are well constrained

1,10

and form

the basis for generation of the geometric dust plume model. By fitting

to the location and geometry of dust structures across the multiple

epochs of observations, our model suggests that the half-opening

angle of the conical shock front is θ

=40±5°. This value is in close

agreement with that estimated by Fahed etal.11 (42±3°), which is

derived by fitting a cone model12 to the 5696A C  emission line, and

Williams etal.13 (34±1°) by modelling the 10830 He  subpeak. The θw

value estimated in this study implies a momentum ratio between the

O and WR star of

=0.043

−0.015

+0.021

assuming radiative postshock condi-

tions7,14, which is a larger value than that expected from the mass-loss

rate and wind speed of the stars (Extended Data Table2). Upon fitting

to the turn-on and turn-off values independently, we find that dust

production occurs over a period of

−0.1

+0.3

yr centred at the periastron

passage, with key parameters of the dust production thresholds dis-

played in Extended Data Table3.

An image simulated under this geometric model at ϕ=0.592 is dis-

played in Fig.2a. When comparing with the corresponding observations

(Fig.1), this model successfully reproduces a number of prominent

features, with very accurate registration between the structural edges

in the model and the data. The eastern arm represents a segment of

the ellipse that corresponds to the earliest dust produced in the dust

production episode (when dust production turned on), whereas the

southern bar corresponds to the most recently produced dust in the

episode (just before dust production turned off).

However, not all predicted features are apparent in the data. Although

the geometric model is primarily designed to reproduce structural fea-

tures, its physical interpretation implies a pinwheel system producing

isothermal, optically thin dust at a constant rate. The clear absence of

structures predicted by the geometric model, most noticeably dust

features to the north and west (an outer-western arc) and below the

southern bar, cannot be explained by simple density variations due

https://doi.org/10.1038/s41586-022-05155-5

Received: 15 December 2021

Accepted: 27 July 2022

Published online: 12 October 2022

Open access

Check for updates

1Institute of Astronomy, University of Cambridge, Cambridge, UK. 2Sydney Institute for Astronomy, School of Physics, The University of Sydney, Sydney, New South Wales, Australia. 3NSF’s NOIR

Lab, Tucson, AZ, USA. 4Institute of Space & Astronautical Science, Japan Aerospace Exploration Agency, Sagamihara, Japan. 5Université Grenoble Alpes, CNRS, IPAG, Grenoble, France.

✉e-mail: yinuo.han@ast.cam.ac.uk

270 | Nature | Vol 610 | 13 October 2022

Article

to orbital velocity, and deeper astrophysical insight into the plume

generative model is required.

We found that two modifications to the geometric model enable it to

account for the lack of the north-western outer dust arcs, both of which

have physical motivation in the mechanics of the system. First, as dust

nucleation and condensation sensitively depend on the physical condi-

tions mediated by stellar winds

, a change in the shock structure will occur

over the orbit, resulting in a dust production rate that varies continuously

within the dust-producing segment of the orbit. We found that when the

dust production rate is smoothly reduced to a local minimum at perias-

tron in the model, we obtain a significantly improved fit that removes

the western dust arc from the image, in accordance with observation.

Second, asymmetries introduced at the wind base can be amplified

into asymmetric large-scale structures. Williams etal.

found that an

increased dust density at the trailing edge of the shock cone relative to

the orbital direction produced a better fit than a uniform model. This

asymmetry could be introduced by the fast orbital motion near perias-

tron, which results in an intrinsic ‘headwind’ against the instantaneous

direction of orbit. By allowing dust to preferentially form on the trail-

ing edge of the conical shock front and smoothly decrease in density

towards the leading edge, we again obtained an improved fit with the

relative brightness of the northern dust structures reduced. A sche-

matic diagram illustrating the two effects is displayed in Fig.3a.

Figure2b,c displays model images under the two modifications to

the original model (see also Extended Data Fig.1). We find that a very

close fit to the data is achieved when both the orbital and azimuthal

modulation of dust production are combined (Fig.2d–f). Overlaying

the outline of the model on the corresponding ϕ=0.592 Lp-band obser-

vation shows that the geometry predicted by the model excellently

reproduces the dust structures and their location (Fig.2e).

Previous interpretations of dusty WR binaries have focused on iden-

tifying the upper limit on the binary separation that enables sufficiently

high densities in the colliding winds to form dust. As gas compression

is derived from wind–wind collision, it is natural to expect an upper

distance threshold above which a sufficiently high density cannot be

reached to facilitate dust nucleation. However, our model appears to

I= 0.043

CH4 (mask) NIRC K (mask) NIRC K (mask) NIRC

PAHCS NIRC Kp NIRC2 Lp NIRC2

Ms NIRC2 HNIRC2 Kp NIRC2

Kp (cor.) NIRC2 Lp NIRC2PAHCS NIRC

Lp NIRC2

Ms NIRC2

Southern Bar

Eastern Arm

I= 0.043 I= 0.059

I= 0.592I= 0.059 I= 0.077 I= 0.077

I= 0.111I= 0.077 I= 0.111

I= 0.111 I= 0.111 I= 0.183 I= 0.592

Fig. 1 | Near-infrared imagery of WR140’s expanding circumstellar dust

structure. The observing bands (top left), instrument (top right) and orbital

phase ϕ (bottom right) are labelled in each panel. Coronagraphic (cor.) and

aperture masking (mask) images are labelled in parentheses after the

observing bands. All other images were observed with the full aperture.

The scale bar in each panel indicates 0.3arcsec (corresponding to 501AU at

a distance of 1.67 kpc). The images are stretched linearly by the following

amounts to accentuate the dust structures that are otherwise faint relative to

the bright stellar core: ϕ=0.077 Kp-band image stretched between 0 and 0.8

relative to the brightest pixel; ϕ=0.111 Kp-band image between 0 and 0.5;

ϕ=0.111 H-band image between 0 and 0.8; ϕ=0.592 Lp-band image between 0

and 1.5×10−3; ϕ=0.592 Ms-band image between 1×10−4 and 5×10−3. All other

images are not stretched. An observing log is presented in Extended Data

Table1.

Nature | Vol 610 | 13 October 2022 | 271

suggest that there may also exist a lower limit, implying that dust is

only formed within a Goldilocks zone where just the right conditions

of density and temperature are met.

Observationally, several other WR binaries with infrared-imaged

dust plumes, such as WR112 (ref.

), WR98a

15,16

and WR104 (refs.

5,6

appear to be continuous dust producers that do not show apparent

signs of such a threshold effect. The Apep system appears to oscillate

in and out of a single threshold, with dust production occurring near

periastron8. If dust production in WR140 indeed reaches two maxima

over a single orbit as suggested by our model, it would be the first WR

binary known to exhibit the full range of this Goldilocks effect.

Previous work by Usov3 has shown that the degree of WR wind com-

pression and cooling is sensitively dependent on the wind velocity, and

so it is plausible that, at very short distances, the WR wind is signifi-

cantly slowed by its binary companion via radiative braking, thereby

reducing the rate of dust production. Lau etal.

proposed that this

mechanism may be responsible for impeding dust formation in the

Gamma Velorum system. Alternatively, near periastron, the wind of

the O star may not have accelerated to its terminal velocity before

reaching the shock front.

With all other parameters fixed, the final parameter of interest in the

model is the dust expansion speed. Dust grains are expected to form

in a population of postshock gas where the two stellar winds of the

binary collide. This dust may then be accelerated by stellar radiation

pressure until reaching the terminal dust drift velocity. The assump-

tion of a constant expansion speed similar to the stellar wind speed

has been shown to accurately reproduce the expanding dust shell in

similar systems such as WR112 (ref. 9).

However, the multiepoch observations of WR140 suggest that a uni-

formly expanding dust shell cannot simultaneously reproduce the

time-evolving spatial extent of the dust shell observed in all epochs.

Fitting to the ϕ=0.592 epoch, which shows the best-resolved dust struc-

tures, the model suggests a dust expansion speed of 2,400±100kms

−1

which is broadly consistent with a streaming velocity of 2,170±

100 kms

−1

along the shock cone determined by Fahed etal.

. However,

extending this model to earlier epochs shows a clear misfit to the loca-

tion of the dust shell. Figure2h shows the outline of such a uniform

expansion model overlaid on the earliest epoch of imagery at ϕ=0.043.

The dust shell predicted by uniform expansion fails to fit, being signifi-

cantly larger than that observed, implying that the average expansion

speed up to ϕ=0.043 must be lower than in subsequent epochs.

Dropping the assumption of uniform expansion, the locations of

the dust structures were fitted independently to each epoch. The result-

ing expansion speeds yielded a clear accelerating trend, with most of

the impulse imparted onto the dust by the first two epochs. Given the

uncertainties associated with fitting the location of dust features, direct

derivation of the magnitude of acceleration as a function of distance

from the star was not possible. To constrain the basic physics, we there-

fore posited a simple model based on expectations for radiatively

accelerated dust as illustrated in Fig.3b. Dust is not expected to form

very close to the star owing to high temperatures and harsh ultraviolet

radiation, and so the postshock gas is assumed to originate at a constant

drift velocity, v0. At a distance of rnuc, dust nucleates and condenses to

form an optically thick sheet, which experiences maximal radiation

pressure with a constant acceleration,

amax

. The dust continues to

expand and eventually becomes optically thin at rt, at which stage the

acceleration decreases as 1/r2.

Under these working assumptions, we find that dust grains form at a

speed of

=1,810

0−170

+140

kms

−1

, which is then accelerated in the optically

thick regime by

=900

max−400

+700

kms−1yr−1 up to

=220

t−80

+150

AU before

becoming optically thin. The dust nucleation radius is fitted to be

rnuc=50±30AU, although this value is not well constrained.

Figure3c shows the acceleration and velocity as a function of distance

resulting from the best-fit model. Radiation pressure has long been

suggested to play a major role in accelerating material near WR stars

We find that the best-fit acceleration in the optically thin regime may be

aOriginal bOrbital cAzimuthal dOrbital + azimuthal

eNIRC2/Lp

Acceleration

fNIRC2/Lp

Acceleration

gNIRC/K (mask)

Acceleration

hNIRC/K (mask)

No acceleration

I= 0.592 I= 0.592 I= 0.592 I= 0.592

I= 0.592 I= 0.111 I= 0.043 I= 0.043

Fig. 2 | Geometric model of WR140’s circumstellar dust structure.

a–d, Model images at ϕ=0.592 simulated under four different variants of the

geometric model. The four models assume that the dust production rate is

uniform (a), orbitally modulated (b), azimuthally asymmetric (c) and both

orbitally modulated and azimuthally asymmetric (d). An animation showing

the evolution of the model in time is available in Supplementary Video 1.

e–h, The outline of the ‘orbital + azimuthal’ model overlaid on the observations

at three different epochs, in which e–g show models at orbital phases of 0.592,

0.111 and 0.043adjusted for dust acceleration, whereas h shows a model at an

orbital phase of 0.043 thatis notadjusted for acceleration, resulting in a clear

misfit to the data.The scale bar in each panel indicates 0.3 arcsec.

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

Nature|Vol610|13October2022|269Radiation-drivenaccelerationintheexpandingWR140dustshellYinuoHan1,2 ✉,PeterG.Tuthill2,RyanM.Lau3,4&AnthonySoulain2,5TheWolf–Rayet(WR)binarysystemWR140isaclose(0.9–16.7mas;ref.1)binarystarconsistingofanO5primaryandWC7companion2andisknownasthearchetypeofepisodicdust-pro...

展开>> 收起<<

Radiation-driven acceleration in the expanding WR140 dust shell.pdf

共14页,预览3页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Radiation-driven acceleration in the expanding WR140 dust shell

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: