Structural and morphological properties of ultraluminous infrared galaxies at 1<z<3
aa r X i v : . [ a s t r o - ph . GA ] M a r Baltic Astronomy, vol. 24, 231–241, 2015
Structural and morphological properties of ultraluminous infraredgalaxies at < z < , , Zhongyang Ma , , Yang Chen , , and Xu Kong , Institute for Astronomy and History of Science and Technology, Dali Univer-sity, Dali 671003, China; [email protected] Key Laboratory for Research in Galaxies and Cosmology, CAS, Hefei 230026,China Center for Astrophysics, University of Science and Technology of China, Hefei230026, China; [email protected] SISSA, via Bonomea 265, I-34136 Trieste, Italy
Received: 2015 January 31; accepted: 2015 May 14
Abstract.
Using the Hubble Space Telescope (HST)/Wide Field Camera3 (WFC3) near-infrared high-resolution imaging from the 3D-HST survey, weanalyze the morphology and structure of 502 ultraluminous infrared galaxies(ULIRGs; L IR > L ⊙ ) at 1 < z <
3. Their rest-frame optical morphologiesshow that high-redshift ULIRGs are a mixture of mergers or interacting sys-tems, irregular galaxies, disks, and ellipticals. Most of ULIRGs in our samplecan be roughly divided into merging systems and late-type galaxies (Sb − Ir),with relatively high M ( > − .
7) and small S´ersic index ( n < . M ( < − .
7) andlarger n ( > . r e ) with redshift of ULIRGs at redshift z ∼ − r e ∝ (1 + z ) − (0 . ± . . Key words: galaxies: evolution — galaxies: fundamental parameters —galaxies: structure — galaxies: high-redshift1. INTRODUCTIONULtraluminous InfraRed Galaxies (ULIRGs; L − µ m > L ⊙ ) were firsthinted at by the deep InfraRed Astronomical Satellite (IRAS; Neugebauer et al.1984) surveys. Within the past decade, observations have shown that high-redshiftULIRGs are massive galaxies ( M ∗ > M ⊙ ), with extremely high ratio of in-frared to optical flux density ( F (24 µ m) /F ( R ) > ⊙ yr − ) (Chapman et al. 2003; Houck et al. 2005; Yan et al.2007; Dey et al. 2008; Desai et al. 2009; Huang et al. 2009; Fang et al. 2014).At high redshift there are many pre-selected ULIRGs samples, such as Dusty-Obscured Galaxies (DOGs with ( R − [24]) Vega >
24; Houck et al. 2005), SubMil-limeter Galaxies (SMGs with F (850 µ m) > . Fang et al.
Multiband Imaging Photometer for Spitzer (MIPS) 24 µ m selected samples (Yanet al. 2007), and follow-up analysis is then necessary to single out ULIRGs.Since the discovery, ULIRGs have been suggested to be a feasible evolutionaryphase towards the formation of local massive early-type galaxies (Sanders et al.1988; Veilleux et al. 2009; Hou et al. 2011). But, the existence of a large numberof massive galaxies with M ∗ > M ⊙ at z ∼ − z ∼ < z <
3, Hubble Space Telescope (HST)/Wide Field Camera3 (WFC3) near-infrared (NIR) imaging can provide crucial clues to the rest-frameoptical morphologies ( λ rest ∼ λ rest > < z <
3. By using HST NIR images (NICMOS orWFC3), many groups (Dasyra et al. 2008; Melbourne et al. 2008, 2009; Buss-mann et al. 2009, 2011; Zamojski et al. 2011; Kartaltepe et al. 2012) found themorphologies of ULIRGs are diverse, e.g., disks, bulges , multiple components,and irregulars. This implies that ULIRGs may have different formation processessuch as mergers and secular evolution without mergers.Since the samples of previous research programs are commonly small ( < < z < (Brammer et al. 2012; Skelton et al. 2014). Moreover, comparingwith previous studies based on HST/NICMOS F160W images (0 ′′ .
09 pixel − ), thiswork will utilize HST/WFC3 NIR images (0 ′′ .
06 pixel − ) to investigate the mor-phological diversities of high-redshift ULIRGs, and for the first time we explorethe size evolution with redshift of our sample and calculate nonparametric mor-phological parameters of ULIRGs at 1 < z <
3. Section 2 describes the selectionof ULIRGs and the data (include images and catalogs) from the 3D-HST fields.We present the structural and morphological properties of ULIRGs in Section 3and 4, and summarize our results in Section 5. Throughout this paper, we adopta standard cosmology H = 70 km s − Mpc − , Ω Λ = 0 .
7, and Ω M = 0 .
3. Allmagnitudes use the AB system unless otherwise noted.2. SAMPLE SELECTION AND DATATotal infrared luminosity ( L IR = L − µ m ) is an important measurementin characterizing ULIRGs at 1 < z <
3. Direct measurement of L IR requires http://3dhst.research.yale.edu/Home.html orphology and structure of ULIRGsorphology and structure of ULIRGs
3. Direct measurement of L IR requires http://3dhst.research.yale.edu/Home.html orphology and structure of ULIRGsorphology and structure of ULIRGs L IR ) of 502 ULIRGs at redshift1 < z < M ∗ ) to ULIRGs in our sample.far-infrared photometric data, yet they are not available for most of 24 µ m se-lected sources. This is particularly true for our sample. In our work, we adopta luminosity-independent conversion from the observed Spitzer/MIPS 24 µ m fluxdensity to L IR , based on a single template that is the logarithm mean of Wuytset al. (2008) templates with 1 ≤ α ≤ . . Wuyts et al. (2011) demon-strated that this luminosity-independent conversion from 24 µ m photometry to L IR yields estimates that are in good median agreement with measurement fromHerschel/Photoconductor Array Camera and Spectrometer (PACS) photometry.Finally, we construct a sample of 502 ULIRGs with L IR > L ⊙ at redshift1 < z < µ m from Fang et al. (2014), Muzzinet al. (2013), and Kajisawa et al. (2011), respectively.3D-HST is a NIR spectroscopic survey with the HST, designed to study thephysical processes that shape galaxies in the distant universe. The survey containsa great diversity of objects from high-redshift quasars to brown dwarf stars, but isoptimally designed for the study of galaxy formation over 1 < z < . ′′ .
06 pixel − ) imaging data from the WFC3 on the HST(Grogin et al. 2011; Koekemoer et al. 2011). The 5 σ point-source detection limitis brighter than 27.0 mag in the F160W ( H ) and F125W ( J ) filters. Our studyis performed using the latest data (version 4.1) release of the 3D-HST survey.The stellar mass ( M ∗ ) and photometric redshift ( z , if there is no spectroscopicredshift available) we adopt in our work also come from the 3D-HST photometriccatalogs (AEGIS, COSMOS, and GOODS-N). Further details are in Brammer etal. (2012) and Skelton et al. (2014) for the survey and observational design andthe data products. Figure 1 shows the distributions of L IR and M ∗ of ULIRGswith 1 < z < M ∗ > . M ⊙ . ∼ swuyts/Lir template.html Fang et al.
Figure 2: S´ersic index ( n ) and effective radius ( r e ) histogram of ULIRGs at dif-ferent redshift bins in our sample. The left panel (a) is the distribution for n , andthe right panel (b) is the distribution for r e .3. STRUCTURES OF ULIRGSSince the redshift distribution of ULIRGs is quite broad (1 < z < < z < .
8, we choose WFC3F125W bandpass for structural analysis, it corresponds approximately to V -bandin the rest-frame in this redshift range, but in the redshift range of 1 . < z < V ) from the F160Wimage instead. Finally, 98 ULIRGs in our sample have J -band counterparts (1 3) are detected in H -band image. Thestructural parameters of ULIRGs, S´ersic index ( n ) and effective radius ( r e ), fromthe latest catalog (version 1.0) are provided by van der Wel et al. (2012). Asdescribed above, we use the observed J structures at 1 < z < . H structures at 1 . < z < n and r e distributions of ULIRGs at different redshift binsin our sample. From Figure 2(a), the derived S´ersic indexes ranging from 0.4to 8, indicated that a wide range of structural diversities for these ULIRGs, fromspheroid to diffuse structures, e.g., irregulars in appearance, disk-like systems, andelliptical structures. In total, there are 80% ULIRGs distribute at n < . n > . 5. In addition, we also find the the distribution of sizes of ULIRGsare broad, ranging from 0.5 to 8 kpc, but most (81%) of them distribute at r e < z ∼ . M ∗ > . M ⊙ at 1 < z < r e − M ∗ relation. However, most ofthem have smaller sizes, compared to local LTGs with similar stellar mass. In themeantime, there is also the existence of compact ULIRGs with r e < < z < orphology and structure of ULIRGsorphology and structure of ULIRGs 5. In addition, we also find the the distribution of sizes of ULIRGsare broad, ranging from 0.5 to 8 kpc, but most (81%) of them distribute at r e < z ∼ . M ∗ > . M ⊙ at 1 < z < r e − M ∗ relation. However, most ofthem have smaller sizes, compared to local LTGs with similar stellar mass. In themeantime, there is also the existence of compact ULIRGs with r e < < z < orphology and structure of ULIRGsorphology and structure of ULIRGs M ∗ ) and effective radius ( r e ) for ULIRGsat different redshift bins in our sample. The solid lines with 1 σ standard er-ror are provided by Shen et al. (2003) for local late-type galaxies (LTGs).Black solid circles represent the median sizes of ULIRGs at different M ∗ bins(∆log ( M ∗ /M ⊙ ) = 0 . H -band imaging of 33 z ∼ µ m-selected sample ofthe Spitzer survey, Dasyra et al. (2008) found that their effective radii range from1.4 to 4.9 kpc, with a mean of < r e > = 2 . σ = 0 . z ∼ 2. The median value of sizes of these ULIRGs is 3 . ± . . ± . 37 kpc at R band.In Figure 4, the red solid circles represent the median sizes of our ULIRGssample at different redshift bins (∆ z = 0 . r e ∝ (1 + z ) − (0 . ± . ,corresponds to the best fit for the four median points. The slope ( α = − . 96) ofthe size evolution of ULIRGs is steeper than that of gas-rich LTGs ( α = − . α = − . 48 from van der Welet al. (2014). If the Veilleux et al. (2002) data point of local ULIRG r e wasincluded when fitting a power law to the r e − z relation, we find that the slope( − . ± . 11) closer to the LTG value. A possible explanation is that ULIRGsrepresent a marginally more compact sub-sample of the LTG population. Thisinterpretation supported by a large part of our sample is LTGs (see Section 4).Moreover, we find the sizes of ULIRGs at high redshifts are on average one to twotimes smaller than those of local ULIRGs (from Veilleux et al. 2002) with similarinfrared luminosity.4. MORPHOLOGIES OF ULIRGSMorphologies of galaxies correlate a series of physical properties, such as stellarmass, star formation rate and rest-frame color of galaxies, they can provide direct36 Fang et al. r e ( k p c ) This sampleDasyra et al.(2008) ULIRGsKartaltepe et al.(2012) ULIRGsVeilleux et al.(2002) ULIRGs r e ∝ (1+z) -(0.96 ± Figure 4: Evolution of size with redshift in our ULIRGs sample. Red solid circlesrepresent the median sizes of ULIRGs at different redshift bins (∆ z = 0 . r e ∝ (1+ z ) − (0 . ± . ).The sizes of ULIRGs from the literature are also plotted in this figure (Veilleux etal. 2002; Dasyra et al. 2008; Kartaltepe et al. 2012). Typical error bars (black)are shown in the figure. orphology and structure of ULIRGsorphology and structure of ULIRGs Fang et al. r e ( k p c ) This sampleDasyra et al.(2008) ULIRGsKartaltepe et al.(2012) ULIRGsVeilleux et al.(2002) ULIRGs r e ∝ (1+z) -(0.96 ± Figure 4: Evolution of size with redshift in our ULIRGs sample. Red solid circlesrepresent the median sizes of ULIRGs at different redshift bins (∆ z = 0 . r e ∝ (1+ z ) − (0 . ± . ).The sizes of ULIRGs from the literature are also plotted in this figure (Veilleux etal. 2002; Dasyra et al. 2008; Kartaltepe et al. 2012). Typical error bars (black)are shown in the figure. orphology and structure of ULIRGsorphology and structure of ULIRGs J -band images of ULIRGs at 1 < z < . ′′ × ′′ .information on the formation and evolution history of these objects. Followingthe method we performed in Section 3, we use the observed J morphologies at1 < z < . H morphologies at 1 . < z < J band) and Figure 6 ( H band) show examples of the NIR images forULIRGs in the COSMOS field of the 3D-HST survey. We perform the visualinspection by three of us, and find that galaxies in our sample exhibit very diversemorphologies, covering a wide range of types from interacting systems to compactspheroids. As illustrated in Figure 7, some of the ULIRGs show morphologicalfeatures of early-phase mergers, advanced-phase mergers, or merger remnants.Meanwhile, there are many extended disks and irregular morphologies for high-redshift ULIRGs.In order to quantitatively investigate the morphological features of ULIRGs at1 < z < 3, we also measure nonparametric morphological parameters (Abrahamet al. 1996; Lotz et al. 2004), such as Gini coefficient ( G ; the relative distributionof the galaxy pixel flux values) and high moment ( M ; the second-order momentof the brightest 20% of the galaxy’s flux). Based on the rest-frame optical mor-phologies of galaxies, Lotz et al. (2008) defined G - M criteria to classify ETGs(E/S0/Sa), LTGs (Sb-Ir), and mergers: ETGs (E/S0/Sa): G ≤ − . M + 0 . 33 and G > . M + 0 . LTGs (Sb-Ir): G ≤ − . M + 0 . 33 and G ≤ . M + 0 . Mergers: G > − . M + 0 . G vs. M diagram.For the morphological properties of ULIRGs at 1 < z < 3, the majority ofthem shows mergers and irregular and disk-like structures, with relatively high M ( > − . 7) and small S´ersic index ( n < . h n i = 1 . ± . Fang et al. Figure 6: HST/WFC3 H -band images of ULIRGs at 1 . < z < ′′ × ′′ .Figure 7: Examples of different merging types: early-phase mergers, advanced-phase mergers, and merger remnants. The size of each image is 4 ′′ × ′′ . orphology and structure of ULIRGsorphology and structure of ULIRGs Figure 6: HST/WFC3 H -band images of ULIRGs at 1 . < z < ′′ × ′′ .Figure 7: Examples of different merging types: early-phase mergers, advanced-phase mergers, and merger remnants. The size of each image is 4 ′′ × ′′ . orphology and structure of ULIRGsorphology and structure of ULIRGs M ( < − . 7) and larger n ( > . h n i = 3 . ± . < z < . . < z < M ∗ & at high redshifts challenges the merger scenario for the formation ofmassive galaxies. Current numerical simulations (Narayanan et al. 2009) havefailed to produce as many major mergers as required to explain the observed num-ber of ULIRGs at 1 < z < 3. An alternative formation scenario for ULIRGs: amassive, gas-rich galaxy could have a SFR as high as 180 − M ⊙ yr − withoutany merging process. The diversity of morphologies indicates that ULIRGs mayoccur in different interaction stages of major mergers, in minor mergers, or viasecular evolution not involving mergers at all.For ULIRGs in our sample, the fraction of objects classified as ETGs onlyis small, and remains roughly constant across the full luminosity/redshift range.The fraction of galaxies classified as LTGs decreases dramatically with luminositywhile the fraction of mergers and interactions increases. The fraction of mergersand interactions among the 1 . < z < < z < . . < z < × MS) this relation are considered to be starbursts. For ULIRGs in our sample,about 65% of objects have significantly elevated SFRs relative to the normal MS.This implies that violent starburst play an important role in ULIRGs at z ∼ L IR > L ⊙ at 1 < z < z ∼ 2, we performnonparametric measures of galaxy morphology. In the meantime, we explore thesize ( r e ) evolution with redshift of our sample.We find the rest-frame optical morphologies of high-redshift ULIRGs are amixture of mergers or interacting systems, irregular galaxies, disks, and ellipticals.Most of ULIRGs in our sample can be roughly divided into merging systems andlate-type galaxies (LTGs), with relatively high M ( > − . 7) and small S´ersicindex ( n < . M ( < − . 7) and larger n ( > . < z < r e − M ∗ relation. However, most of them havesmaller sizes, compared to local LTGs with similar stellar mass. Meanwhile, we40 Fang et al. Figure 8: Distribution of the rest-frame optical ( ∼ M vs. Gini coefficient plane. The solid lines represent thedefined criteria of Lotz et al. (2008). Early-type galaxies (ETGs, E/S0/Sa): G ≤ − . M + 0 . 33 and G > . M + 0 . 80. Late-type galaxies (LTGs, Sb − Ir): G ≤ − . M + 0 . 33 and G ≤ . M + 0 . 80. Mergers: G > − . M + 0 . z ∼ − r e ∝ (1 + z ) − (0 . ± . . The slope ( α = − . 96) of the size evolution of ULIRGsis steeper than that of gas-rich LTGs ( α = − . 75) with similar stellar mass, butit’s still far flatter than the massive early-type galaxies (ETGs) with α = − . orphology and structure of ULIRGsorphology and structure of ULIRGs