Reliable’Size’Measurement’of’Massive’Galaxies’at’ 0.53 ...MODS-CANDELS sample K...
1
2 4 6 8 10 12 SFR -1 -3 Age (Gyr) n 3 2 1.5 1 0.75 0.5 z Sérsic n >4 SF Quenching? AGN Feedback? Dry Merger Major/Minor Merger Wet Merger? Local Massive Elliptical Cold Gas Accretion for clumpy-galaxies? Ex#Situ 2 nd -phase 1 st -phase <2 0.1 1 10 0 0.5 1 1.5 2 2.5 3 r e (kpc) z 10 10 M ⊙ )M * cor <10 10.5 M ⊙ _ QGs =-0.09±0.43 (0.5)z)1.5) _ SFGs = 0.57±0.07 (0.5)z)2.5) 0.1 1 10 r e (kpc) M * cor *10 10.5 M ⊙ _ QGs = 1.06±0.19 (0.5)z)2.5) _ SFGs = 0.56±0.09 (0.5)z)3.0) Reliable Size Measurement of Massive Galaxies at z~0.53.0 in the GOODSN region with HST/WFC3 Data Probing the EvoluCon of Sérsic Index Takahiro MORISHITA 1 、Takashi ICHIKAWA 1 、Masaru KAJISAWA 2 1.Astronomical Ins0tute, Tohoku University, Aramaki, Aoba, Sendai 9808578, Japan ; [email protected] ２．Research Center for Space and Cosmic Evolu0on, Ehime University, Bunkyocho, Matsuyama, Japan 1.Brief IntroducCon Morphological (Size) Evolu0on of galaxies at z~1-3 is one of the most interes0ng topics of the universe. Because the starforma0on ac0vity in the universe has peaked at the redshiS, the morphological proper0es of galaxies there are found, with improvement of instruments, to evolve drama0cally (Trujillo+07; van Dokkum+09; and many others). With accurate light profile of those galaxies, the evolu0on of the galaxies over the redshiS range has been ac0vely discussed ! Abstract : We analyze the recent released HST/WFC3 IR images in the GOODSN region to study the evolu0on of the size and Sérsic index of galaxies. We obtain the morphological parameters of 1,069 massive (≥ 10 10 M ⊙ ) galaxies at z ∼ 0.5–3.0 with GALFIT aSer examining the reliability with ar0ficial galaxies. With a careful study of the bias and error in the effec0ve radius of GALFIT, r e , we derive the sizestellar mass rela0ons for 270 quiescent galaxies (QGs) and 799 starforming galaxies (SFGs). The median sizes of QGs and SFGs increase from z ∼ 3.0 to ∼ 0.5 by a factor ∼ 2.0–2.2, following r e ∝ (1 + z) −α with α = 0.81 ± 0.17 and α = 0.70 ± 0.07, respec0vely. The result of massive QGs is consistent with the general picture of the significant size growth, while the evolu0on is much slow at 2.0 < z < 3.0 (α ∼ 0.6). For the further understanding of the evolu0on scenario, we study the evolu0on of Sérsic index n, which is oSen referred to as a shape of galaxies, and find that of massive QGs to significantly evolve as n ∝ (1 + z) −α with α = 1.46 ± 0.18, while those of less massive QGs and SFGs are unchanged (n ∼ 1) over the redshiS range. We discuss the evolu0on in size and light profile of highz galaxies, combining both insitu and exsitu process. 2.Data (Fig.1) 2.1.HST/WFC3 IR From CANDELS imaging data, we use F160W(H) in GOODSN region. FWHM(F160W)~0.180.20. H AB <25 for reliable morphological fit (See §3). 2.2.MOIRCS/SUBARU MOIRCS Deep Survey（K AB ~25 mag; 5σ） for phot-z, M * , SFR. By using RestUVJ, we extract Quiescent galaxies (QGs) from the whole sample, while the other galaxies are classified as Starforming galaxies (SFGs). 3.Analysis Size Measurement with GALFIT Sérsic Fit by GALFIT(Peng+02) package, Σ(r)=Σ 0 exp[-b(n)((r/r e ) 1/n -1)] (1) To study the possible bias depending on the apparent magnitude of galaxies, we test GALFIT using 1,000 mock galaxies buried in the F160W image, finding that GALFIT recovers the original profiles up to m H ~25. We also compare the results with two PSFs; median stacked star and DrizzledTiny Tim (Fig.3). We find significant differences in the derived parameters of galaxies with smaller r e and larger n. We conclude that this difference comes from a 0ny difference of the central few pixels of the PSF profiles (Fig.4), and adopt the median stacked star since it represents the stars in the mosaic image. 4.Results and Discussion – How did the local Red EllipCcal Galaxies evolve with redshi^ ? (See Fig.7 ) ASer deriving Sérsic parameters for the sample, we see the evolu0on in r e and n over the redshiS range for QGs and SFGs in Figs.5 and 6. Best fit slopes are derived with r e ,n∝(1+z) -α , where α for each slope is shown at the leS bopom in the panel. We find size evolu0on in massive QGs with α~1.06,a factor of 2.5 from z~2.5 to 0.5, while no or weak size evolu0on in Less Massive QGs and SFGs. The evolu0on in n also shows similar trend; evolu0on in Massive QGs (α~0.74) but weak in Less Massive QGs and SFGs. From the two results, we propose 2phase evoluCon for the galaxies from z~3 to 0.5, in the view of the forma0on of QGs (though large error dominates QGs at z>2.0). In the 1 st phase (1.5≦z≦3) , SFGs transi0on into QGs (quenching) mainly by InSitu process such as starforma0on or AGN feedback. Gasrich (wet) mergers are also favorable because they hardly enlarge the Sérsic n. ASer that, the 2 nd phase (z≦1.5) , where gaspoor (dry) merger of QGs dominates (ExSitu), contributes an important role for the forma0on of the ellip0cal galaxies with larger (e.g., Naab +09). The schema0c view for the whole scenario is shown in Fig.7. The above, however, dips only one possible scenario out of the hundreds of complicated forma0on models. For a further understanding, we need to carefully stydy each “piece of the puzzle of the GALAXY EVOLUTION”. Fig.1(le^)：Sample from MODS catalog and CANDELS. The black cross represents the sample with K<25 but fail to be detect or H>25 by SExtractor. The lower panel shows the fac0on of the sample with M * >10 10 M ⊙ (MODS limit) and H<25 (CANDELS limit) over the whole sample. Fig.2(right)：QGs selec0on following restframe UVJ color of Williams+09. The value denoted in each panel is median specific starforma0on rate (sSFR) for the sample there. The lower panels show the two popula0ons with sSFR as xaxis, to show clear bimodality. Fig.3：Comparison of the results of real galaxies with two PSFs; median stacked star and DrizzledTiny Tim. Fig.4：Comparison of the radial light profiles of the PSFs. 0 0.1 0.2 0.3 -2 -1 0 1 q log sSFR (yr -1 ) 0.5)z)1.0 -2 -1 0 1 1.0<z)2.0 -2 -1 0 1 2.0<z)3.0 0.5)z)1.0 -0.5 0 0.5 1 1.5 2 V-J (AB mag) 0.5 1 1.5 2 U-V (AB mag) 1.0<z)2.0 0 0.5 1 1.5 2 2.0<z)3.0 0 0.5 1 1.5 2 0.03±0.02 0.08±0.08 0.16±0.16 0.57±0.26 1.88±1.41 2.72±2.74 0.1 1 10 18 20 22 24 26 #n m MedianStar1 (mag) 0.1 1 10 #r e -2 -1 0 1 2 6m (mag) 6 > TinyTimDri - MedianStar1, # > TinyTimDri / MedianStar1 1 10 r e, MedianStar1 (pixel) 1 10 n MedianStar1 2 4 6 8 10 0.5 1 1.5 2 2.5 3 n z 10 10 M ⊙ )M * cor <10 10.5 M ⊙ _ QGs = 0.17±0.21 (0.5)z)1.5) _ SFGs =-0.04±0.09 (0.5)z)2.5) 2 4 6 8 10 n M * cor *10 10.5 M ⊙ _ QGs = 0.74±0.23 (0.5)z)2.5) _ SFGs = 0.10±0.13 (0.5)z)3.0) SFGs QGs SFGs Fit QGs Fit 7 8 9 10 11 12 logM * (M ⊙ ) MODS-CANDELS sample K AUTO,MODS )25 H AUTO,CAND )25 40 60 80 100 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Completeness (%) z 10 -7 10 -6 10 -5 10 -4 10 -3 10 -2 10 -1 0.1 1 Surface brightness r (arcsec) TinyTim TinyTimDri No envelope Envelope 1 1E-04 2E-04 Fig.5(le^)/Fig.6(right)：Evolu0on of r e (leS) and n (right) with redshiS. The regressions with r e (n) ∝ (1+z) -α are shown with thick curves. Fig.7：Schema0c view for the evolu0on of galaxies with redshiS as xaxis. The ver0cal align and sizes for objects roughly represent Sérsic n and r e derived in the present study, respec0vely. In the present study, we mainly refer to the evolu0on of Red QGs, and there s0ll remain many to be studied to understand the evolu0on of Blue SFGs.
Transcript of Reliable’Size’Measurement’of’Massive’Galaxies’at’ 0.53 ...MODS-CANDELS sample K...
1
10
2 4 6 8 10 12SFR
(M⊙
yr-1
Gpc
-3)
Age (Gyr)
-1
0
1
q (/G
pc3 )
0.1
1
10
n
10
11
12
321.510.750.5
M*co
r (M⊙
)
z
QGsSFGs
Sérsic n
>4
SF Quenching? AGN Feedback?
Dry Merger
Major/Minor Merger
Wet Merger?
Local Massive Elliptical
Cold Gas Accretion for clumpy-galaxies?
Ex#Situ(
2nd-phase
1st-phase
<2
0.1
1
10
0 0.5 1 1.5 2 2.5 3
r e (
kpc)
z
1010M⊙)M*cor<1010.5M⊙
_QGs =-0.09±0.43 (0.5)z)1.5)_SFGs= 0.57±0.07 (0.5)z)2.5)
SFGsQGs
SFGs FitQGs Fit
<G09,E><G09,L>
0.1
1
10
r e (
kpc)
M*cor*1010.5M⊙
_QGs = 1.06±0.19 (0.5)z)2.5)_SFGs= 0.56±0.09 (0.5)z)3.0)
Reliable Size Measurement of Massive Galaxies at z~0.5-‐3.0 in the GOODS-‐N region with HST/WFC3 Data
-‐ Probing the EvoluCon of Sérsic Index -‐ Takahiro MORISHITA1、Takashi ICHIKAWA1、Masaru KAJISAWA2
1.Astronomical Ins0tute, Tohoku University, Aramaki, Aoba, Sendai 980-‐8578, Japan ; [email protected] ２．Research Center for Space and Cosmic Evolu0on, Ehime University, Bunkyo-‐cho, Matsuyama, Japan
1.Brief IntroducCon Morphological (Size) Evolu0on of galaxies at z~1-3 is one of the most interes0ng topics of the universe. Because the star-‐forma0on ac0vity in the universe has peaked at the redshiS, the morphological proper0es of galaxies there are found, with improvement of instruments, to evolve drama0cally (Trujillo+07; van Dokkum+09; and many others). With accurate light profile of those galaxies, the evolu0on of the galaxies over the redshiS range has been ac0vely discussed !
Abstract : We analyze the recent released HST/WFC3 IR images in the GOODS-‐N region to study the evolu0on of the size and Sérsic index of galaxies. We obtain the morphological parameters of 1,069 massive (≥ 1010M⊙) galaxies at z ∼ 0.5–3.0 with GALFIT aSer examining the reliability with ar0ficial galaxies. With a careful study of the bias and error in the effec0ve radius of GALFIT, re, we derive the size-‐stellar mass rela0ons for 270 quiescent galaxies (QGs) and 799 star-‐forming galaxies (SFGs). The median sizes of QGs and SFGs increase from z ∼ 3.0 to ∼ 0.5 by a factor ∼ 2.0–2.2, following re ∝ (1 + z)−α with α = 0.81 ± 0.17 and α = 0.70 ± 0.07, respec0vely. The result of massive QGs is consistent with the general picture of the significant size growth, while the evolu0on is much slow at 2.0 < z < 3.0 (α ∼ 0.6). For the further understanding of the evolu0on scenario, we study the evolu0on of Sérsic index n, which is oSen referred to as a shape of galaxies, and find that of massive QGs to significantly evolve as n ∝ (1 + z)−α with α = 1.46 ± 0.18, while those of less massive QGs and SFGs are unchanged (n ∼ 1) over the redshiS range. We discuss the evolu0on in size and light profile of high-‐z galaxies, combining both in-‐situ and ex-‐situ process.
2.Data (Fig.1) 2.1.HST/WFC3 IR From CANDELS imaging data, we use F160W(H) in GOODS-‐N region. FWHM(F160W)~0″₺.18-‐0″₺.20. HAB<25 for reliable morphological fit (See §3). 2.2.MOIRCS/SUBARU MOIRCS Deep Survey（KAB~25 mag; 5σ） for phot-z, M*, SFR. By using Rest-‐UVJ, we extract Quiescent galaxies (QGs) from the whole sample, while the other galaxies are classified as Star-‐forming galaxies (SFGs).
3.Analysis -‐ Size Measurement with GALFIT Sérsic Fit by GALFIT(Peng+02) package,
Σ(r)=Σ0exp[-b(n)((r/re)1/n-1)] (1)
To study the possible bias depending on the apparent magnitude of galaxies, we test GALFIT using 1,000 mock galaxies buried in the F160W image, finding that GALFIT recovers the original profiles up to mH~25. We also compare the results with two PSFs; median stacked star and Drizzled-‐Tiny Tim (Fig.3). We find significant differences in the derived parameters of galaxies with smaller re and larger n. We conclude that this difference comes from a 0ny difference of the central few pixels of the PSF profiles (Fig.4), and adopt the median stacked star since it represents the stars in the mosaic image.
4.Results and Discussion – How did the local Red EllipCcal Galaxies evolve with redshi^ ? (See Fig.7) ASer deriving Sérsic parameters for the sample, we see the evolu0on in re and n over the redshiS range for QGs and SFGs in Figs.5 and 6. Best fit slopes are derived with re,n∝(1+z)-α, where α for each slope is shown at the leS bopom in the panel. We find size evolu0on in massive QGs with α~1.06, a factor of 2.5 from z~2.5 to 0.5, while no or weak size evolu0on in Less Massive QGs and SFGs. The evolu0on in n also shows similar trend; evolu0on in Massive QGs (α~0.74) but weak in Less Massive QGs and SFGs. From the two results, we propose 2-‐phase evoluCon for the galaxies from z~3 to 0.5, in the view of the forma0on of QGs (though large error dominates QGs at z>2.0). In the 1st phase (1.5≦z≦3), SFGs transi0on into QGs (quenching) mainly by In-‐Situ process such as star-‐forma0on or AGN feedback. Gas-‐rich (wet) mergers are also favorable because they hardly enlarge the Sérsic n. ASer that, the 2nd phase (z≦1.5), where gas-‐poor (dry) merger of QGs dominates (Ex-‐Situ), contributes an important role for the forma0on of the ellip0cal galaxies with larger (e.g., Naab+09). The schema0c view for the whole scenario is shown in Fig.7. The above, however, dips only one possible scenario out of the hundreds of complicated forma0on models. For a further understanding, we need to carefully stydy each “piece of the puzzle of the GALAXY EVOLUTION”.
Fig.1(le^)：Sample from MODS catalog and CANDELS. The black cross represents the sample with K<25 but fail to be detect or H>25 by SExtractor. The lower panel shows the fac0on of the sample with M*>1010 M⊙ (MODS limit) and H<25 (CANDELS limit) over the whole sample. Fig.2(right)：QGs selec0on following rest-‐frame UVJ color of Williams+09. The value denoted in each panel is median specific star-‐forma0on rate (sSFR) for the sample there. The lower panels show the two popula0ons with sSFR as x-‐axis, to show clear bimodality.
Fig.3：Comparison of the results of real galaxies with two PSFs; median stacked star and Drizzled-‐Tiny Tim. Fig.4：Comparison of the radial light profiles of the PSFs.
0
0.1
0.2
0.3
-2 -1 0 1
q
log sSFR (yr-1)
0.5)z)1.0
-2 -1 0 1
1.0<z)2.0
-2 -1 0 1
2.0<z)3.0
0.5)z)1.0
-0.5 0 0.5 1 1.5 2
V-J (AB mag)
0.5
1
1.5
2
U-V
(AB
mag
)
1.0<z)2.0
0 0.5 1 1.5 2
2.0<z)3.0
0 0.5 1 1.5 2
0.03±0.02 0.08±0.08 0.16±0.16
0.57±0.26 1.88±1.41 2.72±2.74
0.1
1
10
18 20 22 24 26
#n
mMedianStar1 (mag)
0.1
1
10
#re
-2-1 0 1 2
6m
(mag
)
6 > TinyTimDri - MedianStar1, # > TinyTimDri / MedianStar1
1 10re, MedianStar1 (pixel)
1 10nMedianStar1
2468
10
0.5 1 1.5 2 2.5 3
n
z
1010M⊙)M*cor<1010.5M⊙_QGs = 0.17±0.21 (0.5)z)1.5)
_SFGs=-0.04±0.09 (0.5)z)2.5)
2468
10
n
M*cor*1010.5M⊙_QGs = 0.74±0.23 (0.5)z)2.5)
_SFGs= 0.10±0.13 (0.5)z)3.0)
SFGsQGs
SFGs FitQGs Fit
7
8
9
10
11
12
logM
* (M
⊙)
MODS-CANDELS sample
KAUTO,MODS)25HAUTO,CAND)25
40 60 80
100
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Com
plet
enes
s (%
)
z
10-7
10-6
10-5
10-4
10-3
10-2
10-1
0.1 1
Surfa
ce br
ightne
ss
r (arcsec)
TinyTimTinyTimDriNo envelope
Envelope
1
1E-04
2E-04
Fig.5(le^)/Fig.6(right)：Evolu0on of re (leS) and n (right) with redshiS. The regressions with re (n) ∝ (1+z)-α are shown with thick curves.
Fig.7：Schema0c view for the evolu0on of galaxies with redshiS as x-‐axis. The ver0cal align and sizes for objects roughly represent Sérsic n and re derived in the present study, respec0vely. In the present study, we mainly refer to the evolu0on of Red QGs, and there s0ll remain many to be studied to understand the evolu0on of Blue SFGs.
