genovali

Infrared spectroscopy of
Cepheids in the Galactic
Center and more
Kei Fukue
&
Noriyuki Matsunaga
Department of Astronomy
The University of Tokyo
Collaborators
•Japan
• Kobayashi, N.; Yamamoto, R.; Kondo, S.; Ikeda, Y.;
Hamano, S.; Nishiyama, S.; Nagata, T.; Aoki, W.;
Tsujimoto, T.; Baba, J.; Fujii M.
•Italy
• Bono, G.; Inno, L., Genovali, K.
Fukue-san
Outline
• Introduction
• Importance of Cepheids to study the Galaxy
• Our goals and observations
• What we would like to investigate with our Subaru/IRCS
spectroscopic data.
• Analysis of IRCS H-band spectra for metallicity
• Establishing the analysis methods and tools to measure
[Fe/H] and other parameters with near-IR spectra
• Results for target Cepheids
• [Fe/H] of the targets and the implications
• Summary and Future prospects
Introduction
Cepheids as tracers of the Milky Way Galaxy
Classical Cepheids as tracers
• P-L relation → distance (accurate to 10% for each star)
• Ages, kinematics and chemical abundances can be
accurately determined, thus working as good tracers.
Miras
Classical
Cepheid
~2 mag
Type II
Cepheid
Near-IR P-L relations of LMC
variables (Matsunaga, 2013)
Period-age relation
(Bono et al. 2005)
Cepheids as chemical tracers
• Genovali et al. (2013—2015), da Silva et al. (2016)
• Clear and tight metallicity gradient traced by >400
Cepheids and almost no variation in [alpha/Fe]
(Genovali et al. 2015), but significant slopes for
neutron-capture elements (da Silva et al. 2016).
[Na/Fe]
[Fe/H]
[Al/Fe]
RGC [kpc]
Genovali et al. (2014, A&A, 566, 37)
[Si/Fe]
RGC [kpc]
Genovali et al. (2015, A&A, 580, 17)
Our goals and observations
Cepheids we found in the Galactic Center and
what we like to study with H-band spectra
from Subaru/IRCS
IRSF survey of the Galactic Center
• We discovered 4 classical Cepheids (Matsunaga et al.
2011, 2013, 2015) along with other variable stars
(eg >500 Miras).
• IRSF 1.4-m telescope and SIRIUS JHKs-band camera
>500 Miras
30 arcmin
Classical
Cepheids
Type II Cepheids
20 arcmin
The Nuclear Stellar Disk
• A disk-like system where
stars coexist with gas/dust
(Central Molecular Zone)
• Young stars exist in contrast
to the Bulge dominated by
old stars, ~10 Gyr.
• Our 4 Cepheids considered
to be within this disk.
450 pc
Stars
Dust
Gas
Launhardt et al. 2002
Stellar populations in the Nuclear Disk
Current star formation
Eg) Sgr B
Old stars?
representated by
RGB stars
Massive stars (and clusters)
Eg) Arches, Quintuplet
The classical Cepheids
OH/IR stars, SiO masers
Current
1 Myr
10 Myr
100 Myr
1 Gyr
10 Gyr
Our target Cepheids are the only objects aged ~25 Myr
which have been identified in this region.
Goals of IRCS spectroscopy
•Kinematics of the Cepheids
• Confirm their membership to the Nuclear Stellar
Disk by comparing radial velocities with those of
gas and other stars.
• (Proper motions from images separated by a few
years would give a stronger result.)
•Chemical abundances of the Cepheids
• Study chemical evolution of the Nuclear Stellar
Disk and gas supply to the Central Molecular Zone
Observations in 2010—2012
• Subaru/IRCS + AO188
• H- (or K-) band
• R=λ/Δλ~20,000
• 4 Cepheids reported in
Matsunaga et al. (2011,
2013, 2015).
• We also observed 10
standard stars and 2
well-studied Cepheids
(δ Cep and X Cyg).
+0.4
Date
λ
Integration
S/N
Obj.
2010/06/20
K
300 sec×8
~25
(a)
2012/05/25
H
300 sec×12
~20
(a)
2012/05/25
H
300 sec×10
~20
(b)
2012/05/25
H
300 sec×12
~20
(c)
2012/07/26
H
300 sec×14
42
(a)
2012/07/26
H
300 sec×12
55
(b)
2012/07/27
H
300 sec×12
35
(c)
2012/07/27
H
300 sec×8
36
(d)
+0.2
Cep (a)
P=23.54d
3 epochs
-0.2
Cep (b)
P=19.96d
2 epochs
Cep (c)
P=22.76d
2 epochs
-0.4
Cep New
(d)
P=18.87d
1 epoch
Kinematics of the Cepheids
• Matsunaga et al. (2015, ApJ, 799, 46)
• Radial velocities (VLSR) Consistent with the rotation of
the Nuclear Disk.
Cepheid (a)
+129 km/s
Cepheid (b)
—61 km/s
Cepheid (d)
—11 km/s
Sgr A*
12 pc = 5′
5 arcmin corresponds to a projected
distance of 12 pc at 8 kpc.
Cepheid (c)
—80 km/s
In context of chemical evolution
• Chemistry of the targes may tell us the origin(s) of gas in
the Central Molecular Zone.
• 3 possible origins
• Gas in the inner Disk falling through the Bar—metal-rich
• Gas being lost by old evolved stars in the Bulge—around solar
• (High-velocity) clouds in the Halo—metal-poor
Initial condition
Gas
Gas
150 Myr later
Stars
LOS from
the Sun
N-body/SPH simulation by Kim et al. (2011)
Analysis of IRCS H-band
spectra for metallicity
Establishing analysis methods and tools
using spectra of standard stars
Chemical analysis with Hband high-resolution specrtra
• Still state of the art
• line information (identification, strengths)
• how to derive various parameters (Teff, log g, metals)
• APOGEE (Majewski et al. 2015, and a lot more)
• Providing many fundamental results to enable the
accurate chemical analysis with H-band spectra in the
recent couple of years.
• Similar resolution to our H-band data, but slightly
narrower wavelength coverage.
• Our analysis has been developed independently.
Overview: Steps
• For metallicity standard stars
• Reproducing the known metallicities of 10 standards to
verify analysis tool and line information
• Finding a temperature indicator (line-depth ratio method)
• For standard Cepheids
• Reproducing the known temperatures and metallicities of
2 Cepheids (δ Cep and X Cyg) to check the methods
• For 4 target Cepheids
• Deriving the temperatures and metallicities
Tools and basic data
• SPTOOL (by Y. Takeda)
• SPSHOW – synthesizing and displaying model spectra
• MPFIT – measuring abundances
• We combined SPTOOL tasks to effeciently determine
microtubulence and metallicity.
• Stellar atmosphere models (by Kurucz)
• Line lists
• Atomic : Melendez & Barbuy (1999)
• Molecular : Kurucz (http://kurucz.harvard.edu/)
• including wavelengths, excitation potentials, and
strengths
[Fe/H] measurements for
standards
• We selected 95 (relatively unblended) Fe I lines.
• Microturbulences are determined by balancing between
[Fe/H] values from strong and weak lines.
• Derived [Fe/H], standard errors of 0.03 dex, agree very well
with literature values.
0.6
Catalogue
IRCS
0.4
0.60.2
Catalogue
IRCS
[Fe/H] (dex)
[Fe/H] (dex)
0.40.0
0.2
-0.2
0.0
-0.4
-0.2
-0.6
-0.4
-0.8
-0.6
-1.0
-0.8
-1.2
HD018391
HD204155
HD196755
HD223047
mLeo
HD124186
HD012014
-1.2
Arcturus
-1.0
HD137704
HD219978
Comparisons of [Fe/H] of 10
standard stars; red—our results,
and black—literature values.
HD
HD
HD
HD
mL
HD
HD
Ar
HD
HD
Line-depth ratios (LDRs)
as a thermometer
• Fukue et al. (2015, ApJ, 812, 64)
• Combinations of lines with different excitation potential are
sensitive to stellar temperature.
• Low-excitation potential ( < 4eV ) – relatively insensitive to Teff variation
• High-excitation potential ( 5—7 eV ) – strong dependency on Teff
Effective temperature
• Based on IRCS spectra of 8 standard stars with known Teff, we
derived 9 LDR-Teff relations (for the first time in H-band).
• Because of the very large foreground extinction for our targets, this
method is crucial to determine their temperatures.
Line-depth ratio
Analysis for standard Cepheids
6800
6600
6400
6200
6000
5800
5600
5400
5200
Andrievsky05
IRCS
0.30
Reference
IRCS
0.25
0.20
d Cep
0.0
0.2
6200
6000
20
5800
10 Storm04
5600
0
5400
5200
-10
5000
-20
4800
X Cyg
-30
4600
-40 0.0
0.2
-50
0.0
0.2
40
[Fe/H]
Storm04
0.4
0.6
0.8
1.0
Phase
Kovtyukh05
IRCS
[Fe/H] (dex)
Temperature (K)
Temperature (K)
Radial velocity (km/s)
km/s)
Effective temperature
• Teff and [Fe/H] of well-studied Cepheids, δ Cep and X Cyg
• LDR method gives temperatures consistent with the observed phases.
• A typical gravity for Cepheids with P=20 days, log g=1.3±0.5 dex, is
assumed.
• Derived [Fe/H] also agrees with literature values within 0.05 dex.
0.15
0.10
0.05
0.00
−0.05
−0.10
0.4
0.4
Phase
0.6
0.8
1.0
0.6
0.8
1.0
Pulsation
phase
Phase
0.17±0.02
±0.06 dex
0.06±0.02
±0.07 dex
d Cep
X Cyg
δ Cep
X Cyg
Results for target Cepheids
[Fe/H] of Cepheids in the Nuclear Stellar Disk
and their implications
Data to use for the targets
• Subaru/IRCS + AO188
• H-band
• R=λ/Δλ~20,000
• Relatively higher-S/N
spectra takend in 2012
are used for chemical
analysis (but S/N=35~55)
Object
Date
λ
Integration
S/N
GCC-a
2012/07/26
H
300 sec×14
42
GCC-b
2012/07/26
H
300 sec×12
55
GCC-c
2012/07/27
H
300 sec×12
35
GCC-d
2012/07/27
H
300 sec×8
36
Teff and [Fe/H] of the targets
• Although the accuracy is limited by low S/N (35—55),
Teff and [Fe/H] are reasonably constrained.
Object
Teff(K)
[Fe/H] (dex)
GCC-a
5300±290
0.23±0.06+0.14-0.26
GCC-b
5250±250
0.20±0.04+0.27-0.12
GCC-c
4780±160
0.51±0.07+0.14-0.10
GCC-d
5000±120
0.11±0.06+0.13-0.10
Important—this result is preliminary, in particular, for
the case of GCC-c, [Fe/H]~+0.5 dex. In such a highmetal regime, Teff from LDR can be altered due to
metallicity effect on the LDR method.
Comparison with literature
• All the previous measurements indicates [Fe/H] around
the solar, -0.2~+0.2 dex, regardless of ages of tracers.
• 3 Cepheids have [Fe/H] similar to the other tracers, while
1 (GCC-c) have a significantly higher [Fe/H], +0.5 dex.
[Fe/H]
Our Cepheids
50
Most stars including 3 Cepheids
seem to favor <[Fe/H]> of stars in
the Bulge, but gas from the inner
Disk may be also responsible.
Number
10
0
−1.0
10
50
40
4
2
20
10
0
−1.0
15
Number
Number
1.0
80
10
70 NSD 6 < RGC < 9 (kpc)
608
506
40
304
20
2
10
00
−1.0
−0.5
−1.0
−0.5
5 < RGC < 7 (kpc)
0.0
[Fe/H] (dex)
0.5
0.0
1.0
0.0
0.0
[Fe/H]
[Fe/H](dex)
(dex)
9 < RGC < 13
Bulge
0.5
1.0
20
0.5
0.5
1.0
1.0
(Hill+2011)
40
30
30
20
20
10
10
00
−1.0
−1.0
Halo gas is too metal-poor
to produce stars in the NSD.
−0.5
−0.5
[Fe/H] (dex)
Number
Number
0.5
1.0
(Genovali+2014)
0
−1.0
Bulge
30
20
0.0
[Fe/H] (dex)
0.5
6
4
−0.5
0.0
4 < RGC < 6 (kpc)
8
60
50
50
40
0
−1.0
−0.5
[Fe/H] (dex)
−0.5
−0.5
0.0
0.0
[Fe/H]
[Fe/H](dex)
(dex)
0.5
0.5
1.0
1.0
5 < RGC < 7 (kpc)
15
Gas fall from the Halo?
Number
Number
Cepheids in the innermost part of the
Disk, RGC=4~6kpc: <[Fe/H]>= +0.3
6
2
Number
20
Red clump stars in the Bulge:
<[Fe/H]>= +0.0 dex
NSD
8 Cepheids
ber
30
Gas recycling within the Bulge?
Stars in the Nuclear Stellar Disk
<[F/H]>= +0.07 dex
10
Bulge
Gas from the inner Disk?
Number
Metallicities of
relevant systems
40
10
5
High velocity clouds in the Halo
eg) [Fe/H] ~ --0.7dex (Richter+2015)
0
−1.0
−0.5
0.0
[Fe/H] (dex)
80
70
60
7 < RGC < 9 (kpc)
0.5
1.0
Implications and drawbacks
• Implications
• Mass-loss gas from stars in the Bulge may be the main
source of gas and star formation in the Nuclear Stellar Disk
(or the Central Molecular Zone) (see also Cunha+2007).
• One of our Cepheids, GCC-c, has a significantly higher
[Fe/H] than other stars. [Fe/H]=+0.5 favors gas fueling
from the inner Disk.
• Drawbacks
• The S/N of our spectra are low, at around the lower limit of
quality required for reasonable [Fe/H] measurements.
• The number of stars and [Fe/H] measurements are still
limited both for the NSD and for the inner Disk.
Summary
•IRCS results of velocities and [Fe/H] of 4
Cepheids in the Nuclear Stellar Disk
•We have developed chemical abundance
analysis for H-band spectra.
• Line-depth ratio (LDR) method for Teff (Fukue et al. 2015)
• 95 Fe I lines to use for [Fe/H]
• [Fe/H] of Cepheids in the Nuclear Stellar Disk
• 3 Cepheids consistent with enrichment by gas lost by lowmass evolved stars in the Bulge, while 1 may give the first
evidence of metal-rich gas fallen from the inner Disk.
Future prospects
• Abundances of other elements would give further
insights into chemical evolution.
• More Cepheids are being discovered in recent surveys
(eg. KISOGP; VVV—Dekany et al. 2015ab).
• Other near-IR spectrographs, eg WINERED and GIANO,
will be also useful to make similar measurements.
Targets in our S16A
run (2 nights) which
are located in
uncultivated regions
of the Galactic disk.
End
ミクロ乱流ξの決定
• ミクロ乱流ξ
• ライン形成領域における微小な非熱的成分
• サチレーションするような強いラインほどξの影響が大きく、吸収の
浅い弱いラインはξへはほとんど依存しない。
• Blackwell diagram (Saffe & Levato 2004)
• 強いラインと弱いラインのそれぞれで導出された金属量の分散が
最小になるξを探す方法。
• IRCSでの決定精度：~0.5km/s
1.0
9.5
0.8
9.0
loge(Fe) (dex)
Normalized Flux
10.0
0.6
0.4
x = 1.0 kms −1
x = 3.0 kms −1
x = 5.0 kms −1
0.2
0.0
15030
15040
15050
15060
15070
Wavelength(Å)
15080
Blackwell diagram
Arcturus
黒が強いライン
灰が弱いライン
8.5
8.0
7.5
7.0
6.5
15090
6.0
1.0
2.0
3.0
4.0
5.0
-1
x (kms )
6.0
7.0
Comparison with other Cepheids
in the Disk
• 3 Cepheids with [Fe/H] ~ 0.1—0.2 dex are not so metal-rich
as several Cepheids in the innermost part of the Disk, but
the number of the Cepheids in the inner Disk is very limited.
1
Genovali2014
No Cepheids are identified
Martin2015
std Ceps
at 200 pc < RGC<3 kpc. GC Cepheids(This
work)
[Fe/H] (dex)
0.5
0
−0.5
Our 4 Cepeheids
−1
0
Several Cepheids in
the innermost part of
the Disk have [Fe/H]
over +0.2 dex.
5
10
RGC (kpc)
15
20
金属量分布の比較
1
Cepheids
Cepheids(Genovali2014)
Cepheids(Martin2015)
GC Cepheids(This work)
Std Cepheids(This work)
Glimpse9(Yamamoto)
GCC(Cunha07)
GCC(Ryde15)
Quintaplet(Cunha07)
RSGC1(Davies09)
RSGC2(Davies09)
RSGC3(Origlia15)
[Fe/H] (dex)
0.5
0
−0.5
Davies2009
−1
−1
0
1
2
3
4
5
APOGEEのRed giant のgradient
(thin disk)Hayden2014
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20
RGC (kpc)
• Cepheidsのgradientはred giantよりもややslopeはなだらか
• Glimpse9やRSGCなどclusterのある領域にはHII領域が付随
バルジの星の金属量
Ryde2015
Bensby2013
Estimating radial velocities
• Comparing Observed spectra are with Combined
spectra (=Synthesized×Telluric)
• Finding velocities which give the best matches
between the Observed and Combined spectra
Synthesized spectrum (using Kurucz’s ATLAS9/SYNTHE)
Telluric spectrum (from an A-type standard star)
Combined spectrum (=Synthetic×Telluric)
Observed spectrum
Light curve and velocity
curve templates
• H-band light and velocity
curves of 11 Cepheids with
P~20 days are conbined.
• Data compiled by
Groenewegen (2011)
• These templates enable us
to predict the velocity
curve from the light curve.
H-band light curve
Radial velocity curve
(Pulsation-corrected) mean velocities
• Typical uncertainties are roughly ±5km/s.
→ V(LSR)=+129 km/s
→ V(LSR)=—61 km/s
→ V(LSR)=—80 km/s
→ V(LSR)=—11 km/s