ALMA最長基線による
VLBIスケールでの大質量原始星観測
元木業人
(山口大学)
2018.12. 1 Vコン@熊本大学Co-I: 廣田朋也(NAOJ), 町田正博 (九州大学), 米倉覚則 (茨城大学)、
本間希樹
(NAOJ)、高桑繁久 (鹿児島大学), 松下聡樹 (ASIAA)
絶好調、
二日酔いです。
世は
ALMA時代
• 星形成の高分解能観測は
ALMA抜きには語れ
ALMA長基線
• 空間分解能
50 – 5 mas (基線長 > 10 km)
→8 GHzの中基線VLBIに十分匹敵
→メーザーの分布スケール(10
2– 10
3AU)を分解可能
• イメージ感度
例
: 50 mas分解能、数分積分
→連続波〜1 K ライン〜10 K
→見えすぎて困るぐらい何でも見える
ALMA時代のVLBI星形成の役割は？
•
100 – 1000 AUスケールでの3次元ガス運動
→Localな運動が何にどう役立つ？
→「アウトフロー」、「円盤」の何がわかる？
•
VLBIによる大規模探査
→特異な天体の事前探査
(JVN少数基線/KaVA SFR WG)
• 時変動モニター
→星表面/円盤内縁スケールでの変動現象
(e.g., CH3OH周期変動)
VLBI→ALMA
研究発展例
観測天体
: G353.273+0.641
• 大質量原始星候補天体
(NGC6357)
(Motogi et al. 2013; 2016)
• 南天
(Dec 〜-34°)
• 比較的近傍
Photometric distance
1.74 ± 0.31kpc
(Neckel 1978)
Parallax
1.70 ± 0.19 kpc
(Motogi et al. 2016)
5’ = 2.55 pcHerschel 遠赤外線イメージNGC6357
赤外線
SED
• 赤外線光度
〜
5x10
3
L
sun
(減光補正
なし
)
おおよその星質量
〜
10 M
sun• 比較的平坦なSED
→Face-on天体の兆候
→アウトフローの穴から
近
-中間赤外の漏れ出し
(e.g., Zhang & Tan 2011)
◯
G353の赤外線SED (Motgi et al. 2017)
IR data: 2MASS, GLIMPSE, MSX, Hi-GAL, ATLASGAL苫小牧
11mによるモニター
• そもそも
“変な水メーザー”として見つかる
(e.g., Caswell et al. 2008, Motogi et al. 2011)
(1) 異常な青方偏移卓越
→光学的に厚い連続波によるマスキング?
(Motogi et al. 2013)
(2) 非常に激しい時間変動
→力学時間の短いジェットによる衝撃波
微弱な 赤方偏移成分±
100 km s
-1ATCA Spectrum of G353.273+0.641 from Caswell & Phillps (2008)
◯メーザー動画
by TRO11m (Motogi et al. 2011; 2016)
VLBI観測 (Motogi et al. 2016)
Coordinate origin (RA,DEC) = (17h 26m 1s.59, -34° 15’ 14”.9)J2000.• 東西方向のジェット
(最大130 km/s)
• ジェットの見込み角
〜
8°
→ほぼ視線方向
に沿っている
•
Face-on円盤が付随し
ているはず
...
◯
VERAで得られた水メーザーの分布と固有運動
170 au 1 mas yr-1 = 32 km s-1 @ 1.7 kpcALMA Cycle 4
• アンテナ配列
: C40-9 (12m x 45、最長12 km)
• 空間分解能
: Band 4 (141.75 – 157.43 GHz) = 2mm
(→大雪でBand 6はキャンセル)
• 積分時間
: 5 min
• 空間分解能
80 x 50 mas (Cont), 100 x 80 mas (Line)
50 x 50 mas (Super-res for cont) for comparison with J-VLA
• 速度分解能
: 〜1 km/s
• イメージ感度
:
0.1 mJy/b (Cont @ 8GHz BW) = 2 K
18
Figure 1. | The bird’s-eye view of G353.273+0.641. a, Original dust continuum map. White circle represents the synthesized beam for continuum. b, Residual image after the subtraction of the compact emission. Contours show the continuum at 45 GHz4, which are every 25 K starting from 100 K (5σ). c, Integrated flux map of the CH3OH emission (Methods). Contours show the dust emission in Fig. 1b, which are every 15 K starting from 40 K (15 σ). Black ellipse represents the synthesised beam for line. d, Distribution of Toomre’s Q parameter. Blue, green and red contours indicate Q = 1.0, 1.5 and 2.0, respectively.
Original Image
• 非常に明るい点源
Peak T
b〜
480 K
直径 〜
170 AU
• 点源周りに薄く広がっ
た放射が存在
• 中心が明るいせいで周
囲がよく見えない
...
秘密
輝度の空間プロファイル
• 明らかに広がった構造
→中心の点源をガウシアンフィットで差し引き
2
Figure 2. Fig. 2
Figure 3. Extended Data Fig. 1
残差イメージ
• 非対称なリング状の分布
→輝度コントラスト E:W = 2:1
•
Peak T
b〜
126 K
• 外半径 〜250 AU
• 幅
〜
150 AU
• 特徴的な構造
アーク状構造
(東)
クランプ状構造
(西)
• 大質量星形成における降着
円盤の
2次元構造を空間分解
した初めての例
秘密
水メーザーとの比較
5Figure 6. Extended Data Fig. 4
• リングの中心付近から水メー
ザージェットの吹き出し
→絵に描いたような
“disk – jet 系”
→7 年越しで想像図が
電波写真に
•
Thermal メタノールで見えて
いる低速のアウトフロー吸収
線
(カラー)と概ね一致
→ジェットを取り巻く
低速の分子アウトフロー？
秘密
1 Figure 1. Fig. 1
J-VLA 7-mm連続波との比較
(Motogi et al. 2017)
• 完全同じビームサイズで
の重ね合わせ
• 中心の穴にすっぽりと
7-mmの放射がハマる
→J-VLAでは中心の
高温成分のみを検出
秘密
コンパクト成分の
SED
(ALMA + VLA K – Ka – Q)
•
VLA Ka/Kデータはビームサイ
ズが倍以上のため
upper limit
•
スペクトル指数
α〜2.5
(ALMA – VLA Q)
• 典型的なダストパラメータを考
えると
2mmでは光学的に厚い
• 光学的厚みと質量の下限値
τ
〜
2 @ 2 mm
mass 〜 0.8 Msun
→中心星の8 %
α〜2.0 102 ν [GHz] 10−1 100 101 102 Sν [mJy] α = 2.0 α = 2.5秘密
面密度分布
• 面密度分布の導出
(1) 温度Profile
(2) Dust mass opacity (β = 1)
→観測されたT
bを
N
gasへ変換
(ガス/ダスト比 = 100)
• 面密度のコントラストは最大
4 - 5
•
‘リング’の質量 〜 1 M
sun→合計質量 〜2 M
sun• 星質量のほぼ
20%
→円盤は自己重力的
On the other hand, physical parameters in the extended emission were estimated by
fol-159
lowing procedure. We first assumed the power-law profile of the dust temperature as, Tdust(r) =
160
360×(r/80 AU)−0.4 K, where r indicates a radius from the centre. The assumed power-law index
161
of -0.4 is typical value for the embedded disk23. The relation was scaled by the observed TB and
162
estimated τν of the compact component. We, then, calculated τν from the ratio between Tdust and
163
observed TB in the extended structure pixel-by-pixel as follows,
164 τν = ln !T dust TB " .
Finally, obtained spatial profile of τν was converted to the surface density profile and total mass by
165
using κ150GHz and Rgd again.
166
CH3OH lines:
167
Since all the detected CH3OH lines showed similar spatial and kinematic profile, we stacked
168
multiple lines for better sensitivity. We divided detected lines into two categories based on the
169
upper-state energy, i.e., ”hot” transitions (> 70 K) and ”cold” transitions (< 70 K). The stacking
170
image was made for each of hot (Fig. 1d) and cold (Extended Data Fig. 4) transitions separately,
171
because only the latter showed the significant absorption feature. Only the negative emission was
172
integrated in the Extended Data Fig.4. The final image noise level (1 σ) was 1.1 mJy beam−1 or 7.5
173
K. We adopted the cutoff signal-to-noise ratio of 10σ (± 75 K) for the integrated flux image. The
174
regions outside this criteria were masked and shown by white colour in Fig 1c. The PV-diagram
175
(Fig. 2a) was made along the right ascension offset = 0, from north to south.
176
We used the kinematic model of the infalling rotating envelope developed for the low-mass
177
9
On the other hand, physical parameters in the extended emission were estimated by
fol-159
lowing procedure. We first assumed the power-law profile of the dust temperature as, Tdust(r) =
160
360×(r/80 AU)−0.4 K, where r indicates a radius from the centre. The assumed power-law index
161
of -0.4 is typical value for the embedded disk23. The relation was scaled by the observed TB and
162
estimated τν of the compact component. We, then, calculated τν from the ratio between Tdust and
163
observed TB in the extended structure pixel-by-pixel as follows,
164 τν = ln !T dust TB " .
Finally, obtained spatial profile of τν was converted to the surface density profile and total mass by
165
using κ150GHz and Rgd again.
166
CH3OH lines:
167
Since all the detected CH3OH lines showed similar spatial and kinematic profile, we stacked
168
multiple lines for better sensitivity. We divided detected lines into two categories based on the
169
upper-state energy, i.e., ”hot” transitions (> 70 K) and ”cold” transitions (< 70 K). The stacking
170
image was made for each of hot (Fig. 1d) and cold (Extended Data Fig. 4) transitions separately,
171
because only the latter showed the significant absorption feature. Only the negative emission was
172
integrated in the Extended Data Fig.4. The final image noise level (1 σ) was 1.1 mJy beam−1 or 7.5
173
K. We adopted the cutoff signal-to-noise ratio of 10σ (± 75 K) for the integrated flux image. The
174
regions outside this criteria were masked and shown by white colour in Fig 1c. The PV-diagram
175
(Fig. 2a) was made along the right ascension offset = 0, from north to south.
176
We used the kinematic model of the infalling rotating envelope developed for the low-mass
177
9
tended Data Fig. 1), we performed the elliptical Gaussian fitting for the compact dust continuum
141
emission, and then, the best-fit Gaussian was subtracted from the original image (Extended Data
142
Table 1). The effect of this subtraction is less than 20 K at the mid point of the residual ring in
143
Fig. 1b. This does not change any conclusions in this paper. The low brightness regions (< 40 K)
144
were masked and shown by white colour in Fig 1c. The spectral index was determined by the total
145
fluxes in Extended Data Table 1.
146
Most of the physical parameters of the dust continuum were estimated using the dust mass
147
opacity at frequency ν (κν), dust opacity index β and the gas-to-dust ratio (Rgd). In this paper, we
148
adopted κν = 0.90×(ν/230 GHz)β cm2g−1, where we assumed the dust model coagulated in the
149
dense gas (106 cm−3) with the thin ice mantle21. Unless typical size of the dust particle is very
150
large as in the evolved protoplanetary disks22, β usually lies between 1 – 25, 10. We adopted β
151
= 1.0 in this work, and hence, κ150GHz is to be 0.59 cm2g−1. This is a conservative choice, i.e.,
152
higher β results in more massive disk with lower Q value. It should be noted that β < 1.0 may be
153
possible if the compact emission is optically thin, but this is clearly not the case, considering the
154
absence of the CH3OH lines towards the centre. The optical depth (τν) of the compact component
155
was determined by fitting the flux ratio between 45 and 150 GHz. We converted τν to the surface
156
density of the dusty gas as,
157
Σ = τν κν
Rgd.
Total mass was, then, acquired by integrating Σ
158
8
