A Molecular Face-on View of the Galactic Centre Region

0.6 MB·English

by  T. Sawada

Mon.Not.R.Astron.Soc.000,1–13(2004) Printed2February2008 (MNLATEXstylefilev2.2)
A Molecular Face-on View of the Galactic Centre Region
Tsuyoshi Sawada,1,2,3⋆ Tetsuo Hasegawa,2,3 Toshihiro Handa3 and R. J. Cohen4
1Nobeyama Radio Observatory,462-2 Nobeyama, Minamimaki, Minamisaku, Nagano 384-1305, Japan
2National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo181-8588, Japan
3Institute of Astronomy, University of Tokyo, 2-21-1 Osawa, Mitaka, Tokyo181-0015, Japan
4 4Jodrell Bank Observatory, University of Manchester, Macclesfield, Cheshire SK11 9DL
0
0
2
Accepted 2004January12.Received2003October 11

n
a
J
ABSTRACT

5 We present a method to derive positions of molecular clouds along the lines of sight
1 froma quantitative comparisonbetween2.6mmCO emissionlines and18cmOH ab-

sorption lines, and apply it to the central kiloparsecs of the Milky Way. With some
1
simple but justifiable assumptions, we derive a face-on distribution of the CO bright-
v
nessandcorrespondingradialvelocityintheGalacticcentrewithoutanyhelpofkine-
6
maticalmodels.Thederivedface-ondistributionofthegasiselongatedandinclinedso
8
2 thatthe Galactic-eastern(positive longitude)side iscloserto us.The gasdistribution
1 is dominated by a barlike central condensation, whose apparent size is 500×200pc.
0 A ridge feature is seen to stretch from one end of the central condensation, though
4 its elongated morphology might be artificial. The velocity field shows clear signs of
0 noncircular motion in the central condensation. The ‘expanding molecular ring’ fea-
/ ture corresponds to the peripheral region surrounding the central condensation with
h
p the Galactic-easternend being closerto us. These characteristicsagreewith a picture
- in which the kinematics of the gas in the central kiloparsec of the Galaxy is under a
o strong influence of a barredpotential. The face-ondistribution of the in situ pressure
r
of the molecular gas is derived from the CO multiline analysis. The derived pressure
t
s is found to be highest in the central 100pc. In this region, the gas is accumulating
a
and is forming stars.
:
v
i Key words: ISM: molecules – Galaxy: centre – Galaxy: kinematics and dynamics –
X
radio lines: ISM
r
a
1 INTRODUCTION project position-velocity diagrams of molecular/atomic line
data into a face-on view (see, e.g., Cohen & Dent 1983;
The behavior of molecular gas, in particular its physical
Sofue 1995; Nakanishi& Sofue 2003). This is an indirect
conditions and kinematics, in central regions of galaxies
method to investigate the spatial distribution of the gas:
is key information to understand the star forming activity
itwouldbeinvaluableifwecouldderivepositionsofmolec-
which occurs there. The Galactic centre can be observed in
ular clouds independent of kinematical assumptions. The
much greater detail compared with central regions of other
molecular content and star forming regions in the Galactic
galaxies. However, its inevitable edge-on perspective some-
centrearestronglyconfinedinthecentralafewhundredpar-
times complicates the interpretation of the data. In partic-
secs: the region is called the central molecular zone (CMZ;
ular, a face-on image of the Galactic centre is very hard
Morris & Serabyn1996).Thoughthekinematicsandface-on
to construct, though such an image would be very help-
distribution of the gas within the CMZ are essential infor-
ful to understand its kinematics and to make a compari-
mation to study the star forming activity in the Galactic
son with central regions of other galaxies. Attempts have
centre, they are complicated and less well understood than
been made to construct models of gas kinematics (see, e.g.,
the outer region (up to the Galactocentric radius of a few
Liszt & Burton 1980; Binney et al. 1991; Fux 1999). Their
kpc) mainly traced by atomic hydrogen line.
models give more or less reasonable interpretation of some
aspects ofthedistribution andmotion of thegas, inpartic-
ular its asymmetric position-velocity appearance and some Inthispaper,wepresentamethodtoderiveamolecular
outstanding features. Kinematical models can be used to face-onviewoftheGalacticcentrewithoutanyhelpofkine-
matical models. In 2 we describe the basic methodology.
§
Using that, we draw a face-on distribution of the molecu-
⋆ E-mail:[email protected] lar gas from existing data and discuss the resultant face-on
2 T. Sawada et al.
view in 3. Physical conditions of the gas and the validity 10
§
of parameters used in themodel are also discussed.
]
K 5
[
2 THE METHOD e CO
r
The essence of our method lies in comparing emission and atu emission
r
absorption spectratoward theGalactic centre.Wecompare pe 0
m
theCO2.6mmemissionwiththeOH18cmabsorption.Be-
Te OH
cause the Galactic centre region itself is an intense diffuse s absorption
18cm continuum source, strong OH absorption arises pref- es
n -5
erentially from the gas that lies in front of the continuum ht
regions,ratherthangasthatliesbehindthem.Ontheother rig
hand, the CO emission samples the gas both in front and B
backofthecontinuumsourcesequally.ThustheOH/COra- -10
tio carries information on the position of the gas along the
lineofsightrelativetothecontinuumsources.Inthefollow- -300 -200 -100 0 100 200 300
ing, we devise a method to estimate the position quantita-
LSR Velocity [km s-1]
tively by introducing some simple assumptions.
Figure 1. Sample spectra of CO emission (thin line) and OH
absorption(thickline)at(ℓ,b)=(−0◦.2,0◦.0).Shadowedvelocity
2.1 Qualitative inspection on molecular gas
components demonstrate the difference of cloud position along
distribution
thelineofsight(seetext).
We used the Columbia-U. Chile 12CO J = 1 0 data
−
(Bitran et al. 1997) and a large-scale absorption survey of
the OH main lines (1665MHz and 1667MHz) made by
Boyce & Cohen (1994). Both data were binned over suc-
cessive 5kms−1 velocity ranges. The CO data, for which
the original resolution and grid spacing are respectively 9
arcmin and 7.5 arcmin, were smoothed to 10-arcmin reso-
lution and resampled with a 12-arcmin grid to match the
OH data. Hereafter we mainly used the 1667MHz line for
the OH absorption. However, because of the large velocity
spread of the molecular clouds in the Galactic centre, the
positive-velocity end of the 1667MHz line blends together
with the negative-velocity end of the 1665MHz line. Thus
we replaced the vLSR > 100kms−1 part of the 1667MHz
line with the 1665MHz data scaled by 1.37, which is the
meanratiooftheabsorptiondepthbetweenthetwolinesin
thevelocity range 256vLSR6100kms−1.
Figure 1 shows the processed spectra of the OH ab-
sorption and the CO emission at (ℓ,b) = ( 0◦.2,0◦.0). This
−
figure demonstrates that the OH/CO ratio varies signifi-
cantly as a function of the radial velocity, vLSR. For ex-
ample, we may draw attention to the velocity components
near vLSR 130kms−1 and vLSR 160kms−1 (shad-
≃ − ≃
owed in Fig. 1), both of which belong to the so-called ‘ex-
pandingmolecular ring’ (EMR; Kaifu, Kato & Iguchi1972;
Figure 2. The ratio between the OH apparent opacity and the
Scoville 1972). The CO intensities at these components are CO J = 1−0 intensity at b = 0◦.0. Contours are TMB(CO) =
almostthesame.Thissuggeststhattheamountsofmolecu- 1,2,4,7,10, and 14[K]. The ratio is shown in the region where
largasaresimilarineachcomponent.Ontheotherhand,the TMB(CO) exceeds 1[K]. Hatched velocity ranges were excluded
OHabsorptiondepthsarestrikinglydifferent.Thenegative- fromtheanalysisbecauseofcontaminationbycloudswelloutside
velocitycomponentshowsdeepabsorption,whilethereisno theGalacticcentreregion.
distinctabsorptionaroundthepositive-velocitycomponent.
We immediately deduce from this fact that the negative-
velocity component is located in front of strong continuum Tabs/Tcont; where Tabs and Tcont are line absorption and
source surrounding the Galactic centre, while the positive- continuum antenna temperatures, respectively) and the
velocity component is behind it. This logic led Kaifu et al. CO J = 1 0 line intensity. The ratio has a clear trend;
(1972)toconcludethatthisfeatureisexpandingawayfrom itissmaller−inℓ<0◦ andvLSR>0kms−1,whileitislarger
theGalactic centre. in ℓ > 0◦ and vLSR < 0kms−1 and in the feature with
◦
Figure 2 shows the longitude-velocity (ℓ-v) diagram extremely large velocity width at ℓ 3 , which is called
≃
of the ratio between the OH apparent opacity (τapp Bania’s Clump 2 (Bania 1977). The ratio in the velocity
≡
Molecular Face-on View of the Galactic Centre 3
range 60kms−1 . vLSR . 10kms−1 is affected by the
foregro−und gas in the Galactic disc and does not show the Line of Sight y
intrinsic value of the Galactic centre. Thus the hatched ve-
locityrangesinFig.2areexcludedfromthefollowing anal- The
ysis.
Galactic Centre
2.2 Deriving distances to clouds l
x
Based upon the logic shown in the previous section, we ex-
s
tendittoquantitativeestimation ofmoleculargasdistribu- r
tion.Inordertodeterminethedistancestomolecularclouds
quantitatively,we adopt the following four assumptions:
DIFFUSE
(1)Atagivenℓ,emissionobservedateachvelocitybincomes
CLOUDLET RADIO CONTINUUM
from a single position along theline of sight. SOURCE AROUND
(2) The CO line intensity TCO at a given velocity is pro- THE GALACTIC CENTRE
portional to the amount of molecular gas in unit velocity
width; and the OH opacity τOH (not the ‘apparent’ one
but the real one) at a given velocity is also proportional
totheamountofmoleculargasinunitvelocitywidth.Con-
sequently,τOH=ZTCO where Z is a constant. Observer
(3)TheexcitationtemperatureofOH[Tex(OH)]isuniform.
(4) The 18cm continuum emission is optically thin and Figure 3. The schematic relation of geometrical parameters in
arises from a distributed volume emissivity, j(r) [r is theface-onviewoftheGalacticcentreregion.
the Galactocentric radius], that is axisymmetric about the
Galactic centre as modelled in 2.3.
§
Now Z and Tex(OH) are unknown parameters. How to Figure3showstheschematicrelationofgeometricalparam-
determine them is described in 2.4; validity of them and eters.TheGalacticlongitudeℓisindegrees.Herer,s,andσi
theassumption (1) are discussed§in 3.3. areinunitsofprojecteddistancecorrespondingto1◦ atthe
§
When a cloud whose OH opacity is τOH is located at distance to the Galactic centre: i.e., 150pc at 8.5kpc. The
s = s0 (s is the position along the line of sight), the cloud observed longitudinal distribution of the continuum bright-
absorbs the continuum intensity behind it, i.e., the contin- ness Tcont(ℓ) can be written as
uumemissivityintegratedalongthelineofsightbehindthe s⊙
cloud, −s0∞j(r)ds using the assumption of optically thin Tcont(ℓ)= j(r)ds. (4)
continuum emission. The absorption depth is written as Z−∞
f[1−exRp(−τOH)][ −s0∞j(r)ds−Tex(OH)]wheref isthebeam Since the continuum emissivity beyond the solar circle
filling factor of OH absorbing gas. The total continuum in- should be negligible compared with that in the centre,
tensityobservedbRyusis −s⊙∞j(r)ds,andtheapparentopac- −s⊙∞j(r)ds can bereplaced with −∞∞j(r)ds. Thus
ity τapp of thecloud is expressed as
τapp = f[1−exp(−τOH)R]sh⊙R−js(0∞r)jd(sr)ds−Tex(OH)i. (1) TRcont(ℓ) ≃ Z−∞∞"Xi=N1 √2aπiσi exRp(cid:18)−ℓ22σ+i2s2(cid:19)#ds (5)
−∞ N ℓ2
Since τapp and TCO (aRnd thus τOH = ZTCO) are known = aiexp −2σi2 . (6)
through the observations, we can obtain the value of Xi=1 (cid:18) (cid:19)
−s0∞j(r)dsifweknowf,Z,andTex(OH).Thens0isderived Now ai and σi can be drawn so that the Eq. (6) repro-
from the value of s0 j(r)ds using a distribution model of duces the observed longitudinal distribution of the contin-
R −∞
j(r). uumbrightness.Itisfoundthatobserved continuumdistri-
R
bution is well fitted by up to three components (N = 3).
◦
For b = 0.0, we obtain a1 = 98.3, a2 = 34.0, a3 = 11.4;
2.3 Modelling the distribution of continuum σ1 = 0.120, σ2 = 0.677, σ3 = 7.18. Parameters ai are in
emissivity Kelvins, and σi are in degrees. Figure 4 shows the result of
ourfit.Themodelreproduces theobserved dataquitewell.
Weassumethatthe18cmcontinuumemissivityj(r)around
theGalactic centreisdescribed asasum ofseveral axisym- Sinceσ1,σ2 aresmallenough(.100pc)anda1,a2 are
metric Gaussians: large, the continuum distribution due to the first (i = 1)
N a r2 and thesecond (i=2) components is well defined.In some
j(r) = i exp (2) galaxies seen in nearly face-on perspective, such as M83
Xi=1 √2πσi (cid:18)−2σi2(cid:19) (Ondrechen 1985) and IC 342 (Crosthwaite, Turner& Ho
2000), λ 20cm continuum emission in the central kilo-
N a ℓ2+s2 ≃
i exp . (3) parsec consists of strong central source and extended emis-
≃ Xi=1 √2πσi (cid:18)− 2σi2 (cid:19) sion seen in resolutions of hundreds of parsecs. For such
4 T. Sawada et al.
150 5 and 6. Trials have been done for Z = 0.04 to 0.70[K−1]
and Tex(OH)=0 to 10[K] 1.
We have chosen an appropriate set of (Z,Tex(OH)) so
that thefollowing threeconditions are satisfied:
100 (1) The resultant face-on distribution of the CO brightness
]
K is not too asymmetric between the near and far sides with
[
nt respect to thecentre,
co (2) The features extending above and below the Galactic
T
50 planeareplacedinsimilarface-onpositionsatdifferentlat-
itudes, and
(3) Most of theCO emission has a solution for theposition
s0.
0
In order to check the condition (1), the CO emission
6 4 2 0 -2 -4 -6 within ℓ 61◦istreated.ThepercentageoftheCOemission
Galactic Longitude [degrees] | |
located in the near side (s0 > 0) is displayed in Figure 7a.
Thecentralmolecularemissionhasawell-knownasymmetry
Figure 4. Longitudinal distribution of the 18cm continuum in with respect to ℓ = 0◦: Bally et al. (1988) mentioned that
units of antenna temperature at b = 0◦.0 by Boyce&Cohen about three-fourths of molecular (13CO and CS) emission
(1994) (open circles) and a fit by 3 Gaussians (solid line). The
comes from positive longitudes. Therefore asymmetry to a
fittingfunctionisTcont = 3i=1aiexp −ℓ2/(2σi2) wherea1= similar degree along the line of sight is acceptable. Sets of
9σ83.3=[K7◦.],18a.2 = 34.0[K], a3P= 11.4[K];(cid:0)σ1 = 0◦.120,(cid:1)σ2 = 0◦.677, lmarogsetZofatnhdesemmaisllsiToenx(oOnHth)earfearresjiedcet.edsincetheywouldput
Acloudcomplexaround(ℓ,vLSR) (1◦.3,+100kms−1)
◦ ≃
[hereafter the ‘1.3 region’] is chosen to check the condition
(2). At first we have derived the centres of emission along
continuum distribution, axisymmetric emissivity distribu- the lines of sight in b = 0◦.0, 0◦.2, 0◦.4. Then the stan-
tion is a good approximation in the first order. On the darddeviation of thecentres is±calcul±ated: Figure7bshows
other hand, there are galaxies whose central λ 20cm theresults.Thestandarddeviation becomessmallwhenwe
≃
emission is dominated by ring structure, such as NGC adoptZ &0.10[K−1]andTex(OH) 4[K].Thegasathigh
1097 (Hummel,van der Hulst & Keel 1987) and NGC 1365 latitudes goes farther with respect ≃to that at b = 0◦.0 if we
(Sandqvist,J¨ors¨ater & Lindblad1995).Thisisnotthecase adoptedsmallervaluesforTex(OH);thereverseisalsotrue.
for the Milky Way Galaxy because the 18cm continuum Forthecondition(3),wedealwithClump2.Becauseof
brightness in the Galactic centre, our edge-on view, has an its large OH/CO ratio, Clump 2 is apt to haveno solution;
intense central peak with a monotonic decrease away from i.e., the large absorption depth cannot be reproduced even
thecentre(seeFig.4).Thefactthatthecentralpeak,which ifthecloudislocatedjustinfrontofus.Figure7cshowsthe
should originate in the Sgr A region, is seen through the percentage of the CO emission which can be placed within
foregroundmediumsupportsourassumptionthatthe18cm 3750 < s0 < 3750pc for the used sets of (Z,Tex(OH)). If
continuum is optically thin. w−eusesmallZ orlargeTex(OH),significantemissionislost.
On the other hand, because of large σ3 and small a3, By combining these conditions, we have chosen Z =
the third (i = 3) component should suffer from possible 0.15 0.03[K−1]andTex(OH)=4 1[K].Asnotedin 2.3,
non-axisymmetric distribution on the largest scale and/or thed±eriveddistancestothecloudsa±reuncertainin ℓ &§1◦.5.
| |
individualsources;furthermore,smalldTcont/dscauseslarge Thus the constraint on the parameters given by condition
errorsinthederivedpositionsofclouds.Therefore,inlongi- (3) using Clump 2 (ℓ 3◦) is weaker than those given by
≃
tudes where the contribution from the first and the second conditions(1)and(2).However,parameterspaceallowedby
components is negligible (i.e., ℓ & 1◦.5), the positions of condition(2)isinvolvedinthatgivenbycondition(3).Thus
| |
clouds obtained using thismodel are rather uncertain. we can obtain (Z,Tex(OH)) = (0.15 0.03[K−1],4 1[K])
± ±
usingtheconditions(1)and(2) alone;condition (3)isthen
automatically satisfied.
2.4 Choice of the Z and Tex(OH) values
3 RESULTS AND DISCUSSION
Following the procedure, we can draw a face-on (x-y) dis-
tributionofthemoleculargasbyputtingeachdatapointof 3.1 Distribution and kinematics of the gas
the ℓ-v diagram (Fig. 2) on to the x-y plane with a projec-
3.1.1 Overall structure
tion of (ℓ,s) (x,y). However, f, Z, and Tex(OH) are still
→
unknown.Hereweassumef =1:thevalidityofthisassump- Figure 8 shows the resultant molecular face-on view of the
tion is discussed in 3.3.1. There is no bottom-up scheme Galactic centre region at b = 0◦.0 seen from the direction
§
todetermineZ andTex(OH).TodetermineZ andTex(OH), ofthenorth Galactic pole. TheCO brightnessofeach pixel
we employ a trial-and-error scheme, making face-on maps
◦ ◦ ◦ ◦ ◦
at b=+0.4,+0.2,0.0, 0.2, and 0.4 with various values
− −
ofZ andTex(OH).Thedataused,i.e., theOH/CO ℓ-v dia- 1 Tex(OH) is measured as an excess over the cosmic microwave
gramsandthecontinuumdistribution,areshowninFigures background(CMB).
Molecular Face-on View of the Galactic Centre 5
Figure 5.Longitude-velocity distributionoftheOH/COratio,showninthesamewayasFig.2.
in Fig. 2 is projected on to the x-y plane and smoothed by 1996; Gerhard 1999, and references therein). Our face-on
σ = 20pc Gaussian. Figures 8a and 8b show the distribu- map shows, as a whole, thesame trend as thispicture.
tions of CO brightness and corresponding radial velocity, Some barred galaxies, such as NGC 1097
◦
respectively. Dashed lines in the Figure indicate ℓ = 1.5: (Ondrechen& van der Hulst 1983; Hummelet al. 1987)
±
as mentioned in 2.3, the obtained face-on maps are less and M83 (Ondrechen 1985), show λ 20cm continuum
§ ≃
reliable outside them. enhancements in their bars; in particular, in the dust
lanes. If the Milky Way is barred and the 18cm continuum
emissivity is elongated along its bar like these galaxies,
TheresultantCOdistributionmainlyconsistsofacen- thentheactualdistributionofmoleculargaswouldbemore
tral condensation (corresponding to the CMZ) and a thin
inclined than that obtained here. Therefore the trend of
ridge feature at (x,y) (400pc, 1000pc) that stretches
inclination is verycertain.
≃ −
almostalongthelineofsightfromtheGalactic-eastern(pos-
itivexinFig.8)endofthecentralcondensation.Asseenin
Fig.2,highOH/COratioismorewidespreadinthepositive
3.1.2 Central condensation
longitudes. This comes about because the face-on CO dis-
tributionisinclined sothatthegasinthepositive x(corre- ThecentralcondensationdominatestheCOemissioninthe
spondingtopositiveℓ)liesclosertous,asqualitativelysug- Galacticcentreregion.Itiselongated:itsapparentsizeisap-
gestedbyCohen & Few(1976).Ithaslongbeenarguedthat proximately500 200pc,andshouldbecomparedwith‘twin
×
theMilkyWayisabarredgalaxy.Recentstudiesagreewith peaks’ in central regions of barred galaxies (Kenney et al.
a picture in which the Galactic-eastern (positive longitude) 1992). The minor axis length of the condensation might be
end of the bar is closer to us (see, e.g., Morris & Serabyn smallersincetheface-onmapinvolvespositionalerrorsalong
6 T. Sawada et al.
160
112400 b=+0.4° aaa123===–19–0.5–.3,, ,σσ σ31=2==7–0.–8.6–65 b=-0.2° aaa213===311200.7.19.,0, σσ, σ23==1=070..72.1996
100
K]
T [cont 6800
40
20
0
160
112400 b=+0.2° aaa123===519.8.69.,3, ,σσ σ13==2=09..015.2269 b=-0.4° aaa123===111160...300,,, σσσ123===006...484083
100
K]
T [cont 6800
40
20
0
160 8 6 4 2 0 -2 -4 -6 -8
140 b=0.0° a1=98.3, σ1=0.12 Galactic Longitude [degrees]
120 aa23==3141..04,, σσ23==07..6188
K]100
T [cont 6800
40
20
0
8 6 4 2 0 -2 -4 -6 -8
Galactic Longitude [degrees]
Figure6.Longitudinaldistributionofthe18cmcontinuumbrightnessandmodelfitsshowninthesamewayasFig.4.Thefitparameters
ai andσi (i=1,2,3)areinKelvinsanddegrees,respectively.
Figure 7. (a) Percentage of the CO emissionwithin |ℓ|61◦ whichis located inthe near side. The unitis per cent. (b) The standard
deviationofthecentres ofCOemissioninacloudcomplex(1◦.06ℓ61◦.6,406vLSR6130kms−1)atb=0◦.0,±0◦.2,±0◦.4inparsecs.
(c)Percentage oftheCOemissioninClump2(2◦.66ℓ63◦.4,0◦.06b60◦.4,306vLSR6200kms−1)whichcanbeplacedwithinthe
inner3750pc.Theunitispercent. Desirableparameterspaceisshadowed.
thelinesofsightasnotedlaterin 3.1.3.Themajoraxisof shows that the gas motion in the condensation is strongly
§
thecondensationisinclinedwithrespecttothelineofsight noncircular. The gas in the far-side has larger receding ve-
◦
(x=0)by 70 sothattheGalactic-easternsideiscloserto locity, while the gas in the near-side is rather approaching.
≃
us.Thisangledoesnotchangesignificantlyintheacceptable This velocity structure can be explained if the gas orbit is
parameterspaceof(Z,Tex(OH)).Thecondensationincludes elongated along themajor axis of thecentral condensation.
◦ ◦ ◦
Sgr A (l 0.0), Sgr B (ℓ 0.6), and Sgr C (ℓ 0.5) Itisaclearsignofabar:asimilartrendisoftenseeninboth
≃ ≃ ◦ ≃ −
molecular cloud complexes and the 1.3 region. The resul- numericalsimulationsofgaskinematicsinabarredpotential
tant face-on distribution of radial velocity (Fig. 8b) clearly and observations of barred galaxies (see, e.g., Athanassoula
Molecular Face-on View of the Galactic Centre 7
Figure 8. The resultant molecular face-on view of the Galactic centre at b= 0◦.0. The parameter set (Z,Tex(OH)) = (0.15K−1,4K)
is used. Distribution of CO J = 1−0 emission (normalized by the peak value) is shown in (a). Contours are 0.01, 0.02, 0.04, 0.08,
0.16,0.32,and0.64ofthepeakvalue.Correspondingradial velocityisshownin(b).Distributionofthevelocity isshownintheregion
where the CO emission is more intense than 1 per cent of the peak. Contours are −200 to +200kms−1 with a spacing of 50kms−1.
Thick contour is vLSR = 0[kms−1]. The solar system is located at (x,y) =(0,−8500). White crosses in both panels show the centre,
(x,y)=(0,0).Dashedlinesindicateℓ=±1◦.5.
1992; Lindblad,Lindblad & Athanassoula 1996). This fact However, the elongation of the ridge along the line of
supports thearguments that theMilky Way is barred. sight mightbeartificial: theface-on mapsobtained herein-
evitably involve positional errors along the lines of sight.
Thenominalerrorreducesaroundthecentresincethegradi-
3.1.3 Bania’s clump 2 entof continuumbrightness(dTcont/ds,namely,continuum
emissivityj)becomessteeper;andviceversa.The1σ devia-
The ridge, which lies at (x,y) (400pc, 1000pc), cor-
≃ − tionsinTCO andτapp typicallycause 200pcerroraround
responds to the feature called Bania’s Clump 2. Clump 2 ≃
Clump 2, while the error in the central condensation is a
haslongbeennoticedbecauseofitspeculiarcharacteristics:
few tens of parsecs. A localized source of radio continuum
widespread velocity structure and large latitudinal extent.
emission embedded in the molecular cloud could also lead,
Stark & Bania (1986) considered that Clump 2 can be un-
inourmodel,toanapparentlymorewidespreaddistribution
derstood as an end-on projection of a dust lane or an inner
ofthecloudinthederivedface-onmapthantheactualone.
spiral arm. Fux (1999) argued that, however, a dust lane
Thereforeourresultscannotdeterminewhichofthein-
which associates with the Galactic-eastern side of the bar
terpretations[Stark & Bania(1986)andFux(1999)]isrele-
does not correspond to the Clump 2, but to the ‘connect-
vant.Nevertheless,itiscertainthatClump2liesinfrontof
ingarm’feature(Rougoor & Oort1960)basedonnumerical
thecentre,sincetheOH/COratiothereisremarkablyhigh.
simulation. Fux (1999) discussed Clump 2 and interpreted
it as gas clouds which are crossing a dust lane shock.
In our face-on map, the ridge corresponding to Clump
3.1.4 ‘Expanding Molecular Ring’ feature
2resemblesoffsetridges(ordustlanes)oftenseeninbarred
spiralgalaxies.Naivelyspeaking,theentirestructure,acen- Theso-called EMR featurewas considered asan expanding
tral condensation and a possible offset ridge-like feature, is (and rotating) ring in the early works (Kaifu et al. 1972;
similar to those seen in central regions of barred galaxies Scoville1972),asobviousfrom itsname,based on itstilted
(see, e.g., Kenney et al. 1992). It also mimics numerically oval ℓ-v appearance. Later, another interpretation that the
simulateddistributionsofmaterialsorbitinginabarredpo- apparent expansion results from elongated orbit of the gas
tential (see, e.g., Athanassoula 1992). was proposed. Binney et al. (1991) claimed that the ℓ-v lo-
8 T. Sawada et al.
CO Integrated Intensity to (x,y) (90pc, 100pc). That is highly offset from the
≃ −
OH Integrated Apparent Opacity other part of the central condensation toward us, and the
O
350 14 H isovelocity contours are unnaturally curved around there.
m s]-1 300 12 Integ TSghrisBi1sacnadusBe2d.bTyhetrheeforinetwenesreeadsissicgrneteed cthonetcinlouuudmpossoiutirocness
nsity [K k 220500 810 rated App a0to◦.rwe4aparlnaddceℓ0d=◦.o8n0:◦.it6.oe.bt,hydeainpttaoesraipttiooelnaatcoihnfgtvhetelhocecoirvtryeelsobpcinoitnytdofiiwnegaldrvdefrlℓoomc=ityℓ0◦.i=n6
e a
nt 150 6 re the interpolated velocity field. The resultant face-on maps
ated I 100 4 nt Op oFfigCurOesa1n0dacaonrdre1sp0bo,ndreinspgecrtaidvieally.velocity are presented in
ntegr 50 2 acity The velocity field in Fig. 10b can be used to con-
CO I 0 0 [km vlienret ainnteℓn-vsitpylotraotfioasn,yeptca.r)amtoetear (fianctee-nosnityvioefwo.thWerelipnreos-,
3 2 1 0 -1 -2 -3 s]-1 jected the 12CO J = 2 1 and 13CO J = 2 1 data
Galactic Longitude [degrees] (Sawada et al. 2001) at b−= 0◦.0 on to the x-y pl−ane. Fig-
ures 10c and 10d show the face-on distribution of inten-
Figure 9.Longitudinal distributionofCOJ =1−0integrated
liinnteen)sfiotyr n(tehgiantivlien-ev)eloancidtyOcHominptoengernatteodfathpepaEreMntR.opTahceityve(lothciictyk tsiitvyelyra.tOions Ron2e−1h/a1−nd0(,1R2C2−O1)/1a−n0d(1R2C13O/1)2(isJh=igh2−(11.0)–,1r.e2s)peacl--
range for the integration is −300 6 vLSR[kms−1] 6 −65+5ℓ most all over the central condensation. On the other hand,
(ℓ is indegrees): approaching toward us faster than the hatched highR (J =2 1)(i.e.,0.10–0.15)isseenmainlywithin
13/12 −
velocityrangesinFig.2. a radius of 100pc around thecentre.
≃
Using one-zone large velocity gradient (LVG; see, e.g.,
Goldreich & Kwan 1974) analysis, Sawada et al. (2001) de-
cus of the EMR, a parallelogram rather than an ellipse,
rivedthephysicalconditionsofthegas from theseintensity
can be understood as a projection of the so-called x1 or- ratios. Though the derived physical conditions depend on
bit (Contopoulos & Mertzanides 1977) in a barred poten-
the assumed gas kinetic temperature Tk, the thermal pres-
tial. As noted in the present (see 2.1) and past (see, e.g.,
§ surep/k=n(H2)Tk [n(H2)isthederivednumberdensityof
Kaifu et al. 1972) works, the negative-velocity component
molecularhydrogen]wasfoundtobealmostindependentof
of the EMR lies in front of the centre, while the positive-
theassumed Tk.
velocity component is located behind of the centre. Both
Figure 11a shows the distribution of pressure n(H2)Tk
interpretations, an expandingring and an orbit in a barred obtained in the same way. The highest pressure, 104.9–
potential, were constructed to match thisgeometry. 105.1cm−3K, is seen within the central 100pc. It is note-
Intheface-on mapobtained inthiswork,theEMR (in
worthy that the highest pressure is found roughly symmet-
particular negative-velocity component) is not clearly sep-
rically around the centre, whereas the molecular mass dis-
arated from the central condensation. We investigate the
tribution is asymmetric in the way that the larger fraction
origin of the EMR, an energetic expansion or an orbital
of the CO emission comes from the positive longitude (see
crowding, basedontheOH/CO ratio.Wewould payatten-
Fig. 10a).Thecentral region corresponds tothe‘high pres-
tiontothenegative-velocitycomponent oftheEMR,which
sure region’ in the ℓ-v diagram described in Sawada et al.
is prominently seen in ℓ-v diagrams at b = 0◦.0, 0◦.2, 0◦.4
− − (2001). Launhardt,Zylka& Mezger (2002) investigated the
(Fig. 5). Higher ratio is more extensively seen in the pos-
inner hundreds of parsecs of the Galaxy using IRAS and
itive longitudes. For example, we present the longitudinal
COBEdata,anddeducedthatgasanddustwithinthegalac-
distribution of the CO emission and the OH absorption at
tocentric radius of 120pc (referred to as ‘inner warm disk’
◦
b = 0.2 in Figure 9. The Figure shows the trend that
− intheirpaper)arewarmerthanthoseatlargerradii(‘outer
theOHabsorptionisdeeperinthepositivelongitudescom-
coldtorus’).The‘highpressureregion’foundinouranalysis
paredwiththeCOemission.ThissuggeststhattheEMR,at
maycorrespondtothe‘innerwarmdisk’inLaunhardt et al.
leastitsnegative-velocitycomponent,isinclinedsothatthe ◦
(2002).The1.3regionshowsslightlylowerpressurethanthe
positive-longitude end is closer to us. This geometry agrees
averageofthecentralcondensation;andtheEMRhaseven
with the interpretation by Binney et al. (1991) rather than
lower pressure.
theaxisymmetric ‘expanding ring’.
A schematic illustration of the major molecular fea-
tures is shown in Figure 11b. Sofue (1995) identified a pair
of arm-like features in the central condensation from the
3.2 Physical conditions
13COdatatakenbyBally et al.(1987),thoughthese‘arms’
Sawada et al. (2001) have investigated ℓ-v distribution of are not clearly separated in our face-on map. It may be
physical conditions of molecular gas in the Galactic centre due to insufficient spatial resolution of the present anal-
usingtheCOintensityratios12COJ =2 1/12COJ =1 0 ysis and to deviation from the assumed smooth, axisym-
[R2−1/1−0(12CO)] and 13CO J = 2−1−/12CO J = 2−−1 metric continuum emissivity because of embedded discrete
[R (J =2 1)].Inthissubsectionwediscusstheface-on radio continuum sources. These arms and the high pres-
13/12 −
distribution of CO intensity ratios and physical conditions sure region occupy a similar location in the ℓ-v diagram
of the gas derived from theratios. (Sawada et al. 2001). Okaet al. (1996) argued that forma-
In the face-on map derived in 3.1 (Fig. 8), the Sgr B tion of massive stars may have been taking place contin-
cloud complex at (ℓ,vLSR) (0◦.6,5§0kms−1) is mapped on uously or intermittently in this region for more than 108
≃
Molecular Face-on View of the Galactic Centre 9
Figure 10.Face-ondistribution(atb=0◦.0)of(a)12COJ =1−0;(b)Radialvelocity;(c)12COJ =2−1/12COJ =1−0intensity
ratio; and (d) 13COJ =2−1/12COJ =2−1 intensity ratio. Values are shown in the region where the 12COJ =1−0 emission is
more intense than 2 per cent of the peak. Contours show the 12COJ =1−0 distribution: levels are 0.01, 0.02, 0.04, 0.08, 0.16, 0.32,
and0.64ofthepeak.
years based upon comparisons between their CO data and the far side of the central condensation shows larger reced-
the ℓ-v distributions of H109α recombination line emission ingvelocity(see 3.1).Thisgeometryisconsistentwiththat
§
(Pauls & Mezger 1975) and OH/IR stars (Lindqvist et al. suggested by Sofue (1995), therefore Sgr B and C might be
1991),andcalleditthe‘star-formingring’.Weconsiderthat actually at the leads of these arms. The geometry that the
concentration of gas may occur toform high pressure (high sitesofactivestarformationarelocatedattheleadsofinner
densityand/orhightemperature)cloudsinthisregion,from arms may suggest a time lag between the arrival of the gas
which stars form. Mizutani et al. (1994) observed the [Cii] intothecentralorbitandthebeginningofstarformationas
158µm line, which is considered to trace photodissociation discussedbyKohno, Kawabe & Vila-Vilar´o(1999)forNGC
regions, within ℓ 6 0◦.7 and found that the emission has 6951.
| |
peaksatSgrB1andSgrC,inadditiontoSgrAregion.This
◦
The 1.3 region lies at the Galactic-eastern end of the
might suggest that the star-forming activity is enhanced in
centralcondensation.Okaet al.(1998b)notedthatthelarge
the Sgr B and C regions. In the two-arm model by Sofue
velocityextentsandthefluffystructureofthecloudsandfil-
(1995), Sgr B and C are both located on the leads of these
aments in this region may be a sign that these clouds are
arms. In his analysis, ‘Arm I’ (which contains Sgr B) with
formed by large-scale shocks. Hu¨ttemeister et al. (1998) re-
lower(ratherapproaching)radialvelocityisinthenearside;
ported that SiO in this region is exceedingly abundant and
‘Arm II’ with Sgr C is in the far side. In our face-on map,
its emission arises from hot, thin gas; which also indicates
10 T. Sawada et al.
Figure 11. (a) Face-on distribution of gas thermal pressure derived from one-zone LVG analysis at b=0◦.0. Values are shown in the
region where the 12CO J = 1−0 emission is more intense than 2 per cent of the peak. (b) Schematic illustration of some molecular
features.ContoursinbothpanelsarethesameasthoseinFig.10.
Central Condensation (2000)suggestedthat‘episodicfueling’isoccurringinNGC
...A pair of arms? 5005: the1◦.3 region might bearesult ofsuch kindof cloud
infall. Our results are schematically summarized in Figure
12.
EMR
We have found that the major axis of the central con-
◦
1.3 region densation is inclined by an angle of 70 with respect to
Sgr C ≃
the line of sight. Therefore, the position angle between the
major axes of the large-scale stellar bar and the central
◦ ◦
condensation is 40 –50 (central condensation is lead-
≃ ◦ ◦
ing the bar), if our viewing angle of the bar is 20 –30
≃
(see Morris & Serabyn 1996; Gerhard 1999; Deguchi et al.
2002, and references therein). In some barred galaxies,
therearecentralmoleculargasconcentrations(‘twinpeaks’,
Kenneyet al. 1992). In a face-on projection of high resolu-
tion images, someof theseconcentrationstendtoconsist of
Sgr B
a pair of arms and their major axes are inclined with re-
spect to the bar major axes by moderate angles: such as
300 pc IC342byIshizukiet al.(1990),M101etc.byKenney et al.
Observer
(1992),NGC1530byReynaud& Downes(1999),NGC6951
by Kohno et al. (1999), NGC 5383 by Shethet al. (2000),
Figure 12. A picture of molecular gas in the Galactic centre
andNGC4303bySchinnereret al.(2002).Theyaresimilar
regionproposedonthebasisoftheresultsfromthiswork.Linear
to our face-on view of the Galactic centre. Hydrodynami-
scalesareapproximate.
cal simulations of the barred galaxies also produce similar
gas distributions: e.g., simulations for NGC 1365 made by
theexistenceofshocks.Fromtheirargumentsanditsgeom- Lindblad et al.(1996)produceacentralcondensationwhich
etry,wesuppose that thegas in the1◦.3 region has recently is inclined from the bar axis with a similar angle as that
◦ ◦
arrived in the central region and collided with the central wehaveobtained,40 –50 .Binney et al.(1991)interpreted
condensation. The extremely large scale height of this re- the longitude-velocity feature that corresponds to the cen-
gion compared with otherpartsof thecentralcondensation tral condensation as a projection of gaseous materials in x2
supportsouridea:thegaswhichhasjustfallenintothecen- orbits. However the x2 orbits are aligned perpendicular to
tral gravitational field has not relaxed yet. The envelope of thepotential(Fig.3in Binney et al.1991).Asaresult, the
the 1◦.3 region toward the Galactic-east and smaller radial sense of inclination with respect to the line of sight are op-
◦
velocityisplacedclosertousthanthemainbodyofthe1.3 posite between our and theirinterpretation. This difference
region in the face-on map. This may support our idea that may be due to the fact that the x2 orbits perpendicular to
◦
the 1.3 region has just bumped into the central condensa- thebararestellar(collisionless)orbits.Wada(1994)showed
tion: the envelope may be a tail of the cloud that has not from his damped-orbit model that gaseous orbits do not
yet reached the condensation. Sakamoto, Baker & Scoville alignwithrespecttothebarbuttheirmajoraxesgradually