Fiery Cores: Bursty and Smooth Star Formation Distributions across Galaxy Centers in Cosmological Zoom-in Simulations
Matthew E. Orr, H Perry Hatchfield, Cara Battersby, Christopher C. Hayward, Philip F. Hopkins, Andrew Wetzel, Samantha M. Benincasa, Sarah R. Loebman, Mattia C. Sormani, Ralf S. Klessen
DDraft version January 28, 2021
Typeset using L A TEX twocolumn style in AASTeX63
Fiery Cores: Bursty and Smooth Star Formation Distributions across Galaxy Centers in CosmologicalZoom-in Simulations
Matthew E. Orr ,
1, 2, 3
H Perry Hatchfield , Cara Battersby , Christopher C. Hayward , Philip F. Hopkins , Andrew Wetzel , Samantha M. Benincasa , Sarah R. Loebman , ∗ Mattia C. Sormani , and Ralf S. Klessen
6, 7 TAPIR, Mailcode 350-17, California Institute of Technology, Pasadena, CA 91125, USA Department of Physics and Astronomy, Rutgers University, 136 Frelinghuysen Road, Piscataway, NJ 08854, USA Center for Computational Astrophysics, Flatiron Institute, 162 Fifth Avenue, New York, NY 10010, USA University of Connecticut, Department of Physics, 196A Auditorium Road, Unit 3046, Storrs, CT 06269 USA Department of Physics & Astronomy, University of California, Davis, CA 95616, USA Universität Heidelberg, Zentrum für Astronomie, Institut für Theoretische Astrophysik, Albert-Ueberle-Str. 2, 69120 Heidelberg,Germany Universität Heidelberg, Interdisziplinäres Zentrum für Wissenschaftliches Rechnen, Im Neuenheimer Feld 205, 69120 Heidelberg,Germany (Received December 6, 2020; Revised January 21, 2021; Accepted January 22, 2021)
Submitted to ApJ LettersABSTRACTWe present an analysis of the R (cid:46) . kpc core regions of seven simulated Milky Way mass galaxies,from the FIRE-2 (Feedback in Realistic Environments) cosmological zoom-in simulation suite, for afinely sampled period ( ∆ t = 2 . Myr) of 22 Myr at z ≈ , and compare them with star formation rate(SFR) and gas surface density observations of the Milky Way’s Central Molecular Zone (CMZ). Despitenot being tuned to reproduce the detailed structure of the CMZ, we find that four of these galaxies areconsistent with CMZ observations at some point during this 22 Myr period. The galaxies presentedhere are not homogeneous in their central structures, roughly dividing into two morphological classes;(a) several of the galaxies have very asymmetric gas and SFR distributions, with intense (compact)starbursts occurring over a period of roughly 10 Myr, and structures on highly eccentric orbits throughthe CMZ, whereas (b) others have smoother gas and SFR distributions, with only slowly varying SFRsover the period analyzed. In class (a) centers, the orbital motion of gas and star-forming complexesacross small apertures ( R (cid:46) pc, analogously | l | < ◦ in the CMZ observations) contributes as muchto tracers of star formation/dense gas appearing in those apertures, as the internal evolution of thosestructures does. These asymmetric/bursty galactic centers can simultaneously match CMZ gas and SFR observations, demonstrating that time-varying star formation can explain the CMZ’s low starformation efficiency.
Keywords:
Galaxy: center, star formation, ISM, spiral, ISM: kinematics and dynamics INTRODUCTIONWithin the context of empirical star formation laws,galaxy centers often exhibit particularly extreme and pe-culiar properties. From observations on scales averaging
Corresponding author: Matthew [email protected] ∗ Hubble Fellow over entire galaxies down to those of ∼
100 pc, the starformation rate (SFR) scales with the surface density ofmolecular gas as a power law relationship known as theKennicutt-Schmidt law (KS, Schmidt 1959; Kennicutt,Jr. 1998). Moreover, the presence of high density ( > cm − ) gas seems to strongly predict the star formationrate on the scale of individual molecular clouds (Ladaet al. 2010, 2012). Some galaxy centers host nuclearstarbursts (e.g. NGC253, Leroy et al. 2018), whereas a r X i v : . [ a s t r o - ph . GA ] J a n M. E. Orr et al. the Milky Way’s central region, known as the CentralMolecular Zone (CMZ) appears to be under-producingstars relative to its dense gas content, with surveys find-ing a SFR 10-100 times lower than that predicted bycontemporary theory ( e.g. , Immer et al. 2012; Longmoreet al. 2013). This deficiency has motivated many stud-ies of the star forming properties of molecular clouds inthis extreme environment (Rathborne et al. 2014; Gins-burg et al. 2018; Walker et al. 2018; Barnes et al. 2019;Henshaw et al. 2019).While there is resilient evidence for this star formationdeficiency in the Milky Way’s CMZ (Barnes et al. 2017),there are still signs of ongoing and previously more in-tense past star formation episodes within the past 2-6Myr ( e.g. , Liermann et al. 2012; Lu et al. 2013; Clarket al. 2019). It has been suggested that the CMZ mayhave previously been in a more vigorous state of star for-mation, perhaps similar to other galaxies with nuclearstarbursts (Kruijssen et al. 2014; Krumholz et al. 2017;Sormani & Li 2020). Arguments that the CMZ has a lowstar formation efficiency per dense gas mass often pre-suppose that the CMZ is in equilibrium, and the timeevolution of galactic centers is difficult to study usingMilky Way observations alone.Although limited to probing scales from 10-100 pc,extragalactic galaxy center studies have measured starformation and gas in MW-mass galaxy centers acrossa range of conditions (Casasola et al. 2015; Gallagheret al. 2018, among others). These observations havehighlighted the variety of conditions under which starformation occurs in galaxy centers, further suggestingthat large variations ( ∼ dex) in SFRs and gas surfacedensities naturally arise (Leroy et al. 2013) in these ex-treme environments.Exploring the nature of star formation in galactic cen-ters requires detailed modeling of star formation andfeedback processes ( e.g. , Armillotta et al. 2019), as wellas a self-consistent picture of gas dynamics in the fullcontext of galactic structure and evolution ( e.g. , Tresset al. 2020; Sormani et al. 2020). Cosmological zoom-insimulations have begun to meet these physics require-ments, and now have adequate spatial/mass resolutionto follow the multiphase turbulent ISM, and capture thecosmological context of Milky Way-like galaxies ( e.g. ,Hopkins et al. 2014, 2018). In particular, work withinthe FIRE collaboration has been able to explore star for-mation in the context of galactic disks, cloud lifetimes,and SMBH–gas dynamics connections (Anglés-Alcázaret al. 2017; Orr et al. 2018, 2020; Benincasa et al. 2020;Gurvich et al. 2020). Recent work by Sanderson et al.(2020) has compared several galaxies in the FIRE-2 suitein detail to properties of the MW. In this letter, we compare the centers of seven FIRE-2 galaxies (Wetzel et al. 2016; Hopkins et al. 2018), allapproximately Milky Way-mass spirals, with Milky WayCMZ and extragalactic observations. These simulationshave z = 0 SFRs of − M (cid:12) /yr, which is more typ-ical of L (cid:63) galaxies compared the MW (our Galaxy ap-pears to be an outlier to lower SFR, Longmore et al.2013). Specifically, we map the centers of the simulatedgalaxies at high spatial resolution to understand the ef-fects of dynamical evolution and feedback over a short( ∼ Myr) timescale on proxies for SFR and gas sur-face density tracers, and subsequent interpretations ofstar formation activity in their central regions. METHODSWe analyze the central regions the seven MilkyWay/Andromeda-mass spiral galaxies from the ‘stan-dard physics’ Latte suite of FIRE-2 simulations intro-duced in Wetzel et al. (2016) and Hopkins et al. (2018).The spatially resolved properties of the gas surface den-sities, velocity dispersions, and SFRs across the disks ofthese galaxies have been studied in detail in Orr et al.(2020). This work makes use of 11 snapshots finelyspaced in time ( ∆ t ≈ . Myr) at z ≈ for each ofthe simulations. A brief summary of the z ≈ globalproperties of the galaxy simulations are included in Ta-ble 1 of Orr et al. (2020).The simulations analyzed here all have minimum bary-onic particle masses of m b,min = 7100 M (cid:12) , minimumadaptive force softening lengths < z = 0 is h ∼ − pc (at a n ∼ cm − ), with the dense turbulent disk structures havingnecessarily shorter softening lengths. The aperture sizesconsidered in this work are − pc , and so are wellabove the minimum resolvable scales in the simulations.Importantly, for discussion here: star formation in thesimulations occurs on a free-fall time in gas which isdense ( n > cm − ), molecular (per the Krumholz &Gnedin 2011 prescription), self-gravitating (viral param-eter α vir < ) and Jeans-unstable below the resolutionscale. Once these requirements are met, the SFR at theparticle scale is assumed to be: ˙ ρ (cid:63) = ρ H /t ff ( i.e. , 100%efficiency per free-fall time). Star particles are treated assingle stellar populations, with known age, metallicity,and mass. Feedback from supernovae, stellar mass loss(OB/AGB-star winds), photoionization and photoelec- For comparison with CMZ observations, a physical radial extentof R (cid:46) pc corresponds to | l | (cid:46) ◦ in Galactic longitude,assuming a distance of d ≈ . kpc. tar Formation Distributions in Fiery Cores Figure 1.
Face-on central regions of two FIRE-2 spiral galaxies at five snapshots in time (advancing right, ∆ t ≈ . Myr): m12b (top subfigure) and m12m (bottom subfigure), cold and dense gas (C&D,
T <
K and n H > cm − ) in blues with Myr-averaged SFR (reds) overlaid, with 25 pc pixel sizes. Outermost rows show kpc regions, with inner rows showing . kpc zoom-ins. Zoom-in panels show two apertures, with R = 145 , pc (inner, outer dashed lines respectively). Middle panelshows the time evolution of SFRs within R < pc (dashed lines) and gas surface densities within
R < pc (solid lines).MW CMZ SFR estimate with uncertainty ( | l | < ◦ from Longmore et al. 2013) plotted as horizontal grey-shaded band. Despitehaving similar gas surface densities on R < pc scales (modulo m12b lacking C&D gas within ∼
300 pc for the first ∼ m12b (green lines) having bursty, intense starformation as opposed to the smoother cirrus of star formation seen in m12m (brown lines). The large SFR variation in m12b essentially corresponds to the evolution and physical motion of a single massive star-forming region. M. E. Orr et al. tric heating, and radiation pressure are explicitly mod-eled. These simulations do not include any supermas-sive black holes (SMBHs), and accordingly do not haveany feedback associated with BH accretion, nor do theyinclude cosmic rays or other MHD physics. Detailed de-scriptions of these physics and their implementation canbe found in Hopkins et al. (2018).We produce mock observational maps from the snap-shots using the same methods as Orr et al. (2018) andOrr et al. (2020): we project the galaxies face-on accord-ing to their stellar angular momentum vector (includingstar particles out to 20 kpc) and bin star particles andgas elements into square pixels with side-lengths ( i.e. ,“pixel sizes”) 25 pc. The maps analyzed here are 3 kpcon a side, and integrate gas and stars within ± . kpcof the galactic mid-plane.We generate a proxy for observational measures of re-cent SFRs by calculating the 10 Myr-averaged SFR inthe pixels. We do this by summing the mass of star par-ticles with ages less than 10 Myr, and correcting for massloss from stellar winds and evolutionary effects usingpredictions from STARBURST99 (Leitherer et al. 1999).This time interval was chosen for its approximate cor-respondence with the timescales traced by recombina-tion lines like H α (Kennicutt & Evans 2012) . To com-pare with gas observations, we calculate column den-sities for the “cold and dense” gas ( Σ C&D throughout)with
T <
K and n H > cm − . This gas reservoirtaken as a proxy for the cold molecular gas in the sim-ulations following the methodology of Orr et al. (2020),and ought roughly to correspond with gas traced by colddust or CO observations in the CMZ. RESULTSIn this sample of only seven FIRE-2 Milky Way ana-logues, there is a surprising variety of conditions in theircenters. As an example of properties seen in the galaxycenters, Figure 1 shows 8.8 Myr of two of the FIRE-2 galaxies ( m12b and m12m ), and how their SFRsand gas surface densities evolve within their central ∼ pc (zoomed insets). Despite having similar masses A direct comparison to observations, by post-processing the snap-shots to explicitly model H α , would make for a more accuratemodeling, but accounting for, e.g. , dust and other complexities,is beyond this letter’s scope, where we wish to focus on the “true”SFRs and their spatial distributions. Work by Velázquez et al.(2020) has suggested that shorter ( ∼ α emission. We use a slightly longer averagingwindow such that the simulations well-resolve SFR estimates ofthe CMZ over these timescales within R < pc (where ∼ young star particles corresponds to a measured SFR of ∼ (cid:12) yr − kpc − ), and are more conservative in our sensitivity to SFRvariability. of cold gas in their galactic centers, the two simulationshave morphologically distinct central regions, in terms oftheir cold gas distributions and star-forming complexes.Of the two galaxies in Figure 1, only m12b is able tomatch observational estimates using YSO counts andHII regions of the CMZ SFR within | l | < ◦ (Longmoreet al. 2013), specifically this is seen to occur during aninter-starburst period (at t = 8 . Myr, center-left col-umn). At least four of the FIRE galaxies ( m12b , m12f , m12m , & m12r ) match CMZ SFR and gas surface den-sity properties concurrently at some time in this Myrperiod.3.1.
Two Morphological Classes of FIRE-CMZs
The FIRE galaxies presented here cover a range ofmorphologies in their centers because these simulationswere not designed to match the detailed structural prop-erties of our Galactic center . Within the sample of sevengalaxies, we see two distinct classes of central morphol-ogy in their fiery cores (gas and star formation distribu-tions within R ≈ . kpc):(a) “Asymmetric/Bursty” ( m12b , m12c , m12f , & m12w ): Large, asymmetric gas clouds and star-forming complexes are seen. Star formation isconcentrated in intense starbursts whose feedbackdramatically shapes the local gas environment (see m12b , upper subfigure of Fig. 1). Two simulationsfalling in this category ( m12b & m12f ) simulta-neously match the MW CMZ gas and SFR mea-surements. The two others ( m12c & m12w ) donot simultaneously have SFR and dense gas trac-ers within the central 145 pc at any point in thistime window .(b) “Smooth” ( m12i , m12m , & m12r ): Gas andstar formation is smoothly distributed within thegalactic centers, with clear feeding of gas into cen-ter, and a cirrus of star formation (see m12m ,lower subfigure of Fig. 1). Feedback events donot dramatically alter the local gas environment,as the feedback is relatively dispersed across theircenters.Interestingly, none of the galaxies here exhibit the ringstructures, presumed to be long-lived, seen by studiesof the central regions of other spiral galaxies and theMW CMZ (Kormendy & Kennicutt 2004; Molinari et al.2011). We note the lack of strong bars in the centers of any of these FIRE galaxies at this time ( m12m diddevelop a strong bar around z ≈ . , but it does notsurvive to z = 0 ; Debattista et al. 2019); without thepresence of bars in these simulations at z = 0 , we can- tar Formation Distributions in Fiery Cores − − l og ( Σ M y r S F R [ M (cid:12) y r − k p c − ] ) MW − CMZ SFR ( | l | < . ◦ ) R <
500 pc . . . . l og ( Σ C & D [ M (cid:12) p c − ] ) CMZ Gas Data ( | l | < . ◦ ) t [Myr] . . . . . l og ( Σ C & D / Σ M y r S F R [ M y r ] ) CMZ Depletion Time ( | l | < . ◦ ) MW − CMZSFR ( | l | < ◦ ) R <
145 pc
CMZ Gas Data( | l | < ◦ ) m12bm12cm12f m12im12m m12rm12w t [Myr] CMZ Depletion Time( | l | < ◦ ) Figure 2.
SFRs and cold & dense (C&D) gas surface densities in central regions of seven FIRE-2 spiral galaxies (coloredlines; “class (a)/(b)” plotted with dashed/solid lines, respectively), for
R < (left column) and
R < pc (right column)apertures, as a function of time near z ≈ ( ∆ t ≈ . Myr, rightmost edge being z = 0 ). Shaded bands indicate SFR andgas surface density observations, with uncertainty, of the CMZ from Longmore et al. (2013) and Mills & Battersby (2017),respectively. Depletion times ( Σ C&D / Σ
10 MyrSFR ) are also presented, in the same style; these CMZ depletion times are producedby combining Longmore et al. (2013) and Mills & Battersby (2017) data. SFRs evolve more smoothly in all galaxies in largerapertures (
R < pc), and the variance in SFRs or gas surface density increases with smaller apertures. However, in thesimulations, two central molecular zone classes appear to exist on
R < pc scales: galaxies like m12b and m12c with veryasymmetric gas distributions and dramatic starbursts on ∼ Myr timescales, “class (a)”; and galaxies like m12i and m12m typifying smoother (though still with non-trivial fluctuations) SFR and gas distributions in their centers, “class (b)” (see, m12b and m12m in Fig. 1 as examples of classes (a) and (b), respectively). Despite temporal and spatial variance, many of the FIREgalaxies are consistent with MW CMZ observations at some point in this time-window. not speak to the dynamical importance of bars in pro-ducing central galactic environments similar to the MWCMZ. Other work has highlighted the potential impactsof bars in funneling gas core-ward and forming rings(Sormani et al. 2015, 2018). However, bar-induced ef-fects would likely push these simulated galactic centerstowards more asymmetric states, supporting the ideathat bursty, rather than steady-state, star formation isnecessary to explain MW CMZ observations. Specifi-cally, Sormani et al. (2018) showed that the gas flowin a barred potential naturally becomes turbulent andasymmetric, even in the absence of any type of stellarfeedback. And so, we leave it to future work to investi-gate the gas flows driven in FIRE galaxies by the barsthat form at higher redshift.The structures in the centers of the FIRE galaxies ap-pear to be fairly transient in nature, existing for (cid:46) Myr (similar to the GMC lifetimes seen in these sim-ulations more broadly by Benincasa et al. 2020). Weshould be clear: spiral structures do exist core-ward inthese simulations (see the clear presence of spiral armsin both m12b and m12m in Fig. 1). In the case ofthe class (a) morphologies, the central (
R < pc) gasstructures are on visibly eccentric orbits through theirCMZs (similar to MW CMZ orbital modeling by Krui-jssen et al. 2015), and are disrupted by intense feedbackfollowing starbursts. Previous work with FIRE by Tor-rey et al. (2017) showed that star formation–feedbackinstabilities in galactic centers can arise when the lo-cal dynamical time becomes shorter than the feedbacktimescale. The class (b) morphologies have smoother gasdistributions in their centers, and with the gentler, moredispersed feedback from diffuse star formation, are notas strongly disrupted. Owing perhaps to their smoother
M. E. Orr et al. gas distributions, less of the gas is on very eccentric or-bits ( ∼ − % of gas having v in − plane > √ v c , vs.up to ∼ − % in the class (a) centers) and, to anextent, the structures are clearer continuations of spi-ral arms down to their centers. The difference betweenthe two classes may, in the case of these FIRE galax-ies, arise from more or less violent recent merger histo-ries/interactions with (smaller) galaxies, with the class(b) galaxies having more quiescent recent histories. Towit, as shown in Garrison-Kimmel et al. (2018), m12i and m12m have not experienced any notable head-onmajor mergers.One caveat to the discussion regarding these mor-phologies is the lack of SMBHs in these simulations.Work by Anglés-Alcázar et al. (2017) has investigatedthe influence of SMBHs on their immediate environ-ments, and their ability to disrupt gas structures whilethey are actively accreting, may disallow the smoothcentral gas distributions within ∼
100 pc in class (b).And so, class (a) galactic centers may be the more real-istic central galactic environments.3.2.
Matching FIRE-CMZs with the Milky Way CMZand External Galaxies
Fig. 2 shows the evolution over 22 Myr of the SFRand cold & dense (C&D) gas surface densities, andderived depletion times, within R = 500 and pcapertures (corresponding roughly to CMZ observationswithin | l | (cid:46) . ◦ and ◦ , with SFRs taken from Long-more et al. 2013 and gas from Mills & Battersby 2017)for the FIRE galaxies. Averaging over larger scales re-duces the degree of scatter seen in SFR and gas surfacedensity for each galaxy. However, as discussed in §3.1,the time evolution of Σ SFR and Σ C&D alone do not fullycapture the idiosyncrasies of each galaxy. For example, m12r is relatively less massive ( ∼ × ) and has a smallercold & dense gas reservoir/lower SFRs compared to theother simulations; m12w (and to a less dramatic extent m12c , though with the same result) exhibits a dramaticlack of gas in its center (within 1 kpc) due to a massivestarburst occurring just before the beginning of our anal-ysis, and only near the end of the ∼ Myr period doesthe central gas reservoir begin to recover (as a resultit does not appear on Fig. 3, since it does not simulta-neously have SFR and gas tracers within 145 pc). Thiscase is very similar to the gas compaction and inside-outquenching episodes seen in simulations of blue nuggets(that become red nuggets) at higher redshift (Tacchellaet al. 2016a,b). These episodes, however, need not berestricted to high-redshift, as observations with ALMAand in the MaNGA Survey (Lin et al. 2020; Brownsonet al. 2020) have shown similar variations in central star log(Σ
C&D [M (cid:12) pc − ]) − − l og ( Σ S F R [ M (cid:12) y r − k p c − ] ) R <
145 pc MW − CMZ SFR( | l | < ◦ ) C M Z G a s D a t a m12bm12fm12im12mm12r Blanc et al . / Figure 3.
KS relation in central regions of the five FIRE-2spiral galaxies (colored lines: “asymmetric” centers plottedwith dashed lines, “smooth” centers with solid lines) that simultaneously have SFR and dense gas tracers within the
R < pc aperture, as a function of time near z ≈ ( ∆ t ≈ . Myr). CMZ SFR and gas estimates, with uncer-tainty, (Longmore et al. 2013 and Mills & Battersby 2017)plotted as horizontal and vertical shaded bands, respectively,and spatially resolved KS observations ( ∼ pc & ∼ ,respectively) of M51 (Blanc et al. 2009, their X CO adjustedto be consistent with MW value) and the Antennae Galaxies(NGC 4038/9; Bemis & Wilson 2019) plotted in greyscalecontours and with green ‘+’s, respectively. The central re-gions of some galaxies remain fairly stable in KS-space over Myr ( e.g. , m12m ), whereas others ( e.g. , m12b ) vary byupwards of a dex in both SFR and Σ C&D . Four FIRE galax-ies ( m12b , m12f , m12m , & m12r ) overlap with the CMZSFR estimate at various times. formation efficiency sans mergers in green valley galax-ies in the local universe.Several of the galaxies match the CMZ observationssimultaneously in Σ C&D and Σ SFR at some point in thistime period, with the galaxies generally exhibiting deple-tion times on the shorter end of CMZ estimates. Fig. 3demonstrates this, placing the five galaxies that simulta-neously have tracers of star formation and cold & densegas in their central pc on the KS plane ( i.e. , all but m12c & m12w ), and comparing them with appropri-ate CMZ observations, and spatially resolved observa-tions of M51 (with 170 pc pixels; Blanc et al. 2009) andthe Antennae galaxy nuclei (NGC 4038/9, at ∼
675 pc;Bemis & Wilson 2019).Interestingly, both m12f , class (a), and significantlylower-mass m12r , class (b), strongly overlap with theAntennae galaxies KS data, suggesting similarities be-tween merger-induced starbursts and self-driven burstystar formation. Indeed, simulation work modeling the tar Formation Distributions in Fiery Cores ∼
10 Myr later) central SFRs.Viewed on the KS relation, the simulations exhibit sig-nificantly different tracks over 22 Myr, with, e.g. , m12m stationary with a nearly constant SFR and cold & densegas reservoir, and m12b traveling dramatically acrossthe KS-plane ( ∼ ≈ τ dep ≈ Σ C&D / Σ SFR ∼ α timescales like thoseof the Longmore et al. (2013) observations, in a time-averaged sense ( ∼ ∼
150 pc scales, modulo ob-servational uncertainties (Orr et al. 2018). SUMMARY & CONCLUSIONSIn this letter, we analyzed the central core regions ofseven FIRE-2 Milky Way-mass simulated disk galaxiesby spatially mapping their SFRs and gas surface den-sities, and primarily compared them with comparableobservations of the Milky Way CMZ. Our main resultsare:• There are two fairly distinct morphological classesof fiery cores in this sample, with some galaxies ex-hibiting very asymmetric/clumpy central gas andstar formation distributions (class ‘a’) and otherswith fairly smooth distributions (class ‘b’). Theintense (concentrated) starbursts in the class (a)cores appear to dramatically alter the gas struc-tures in their centers, whereas the smoother feed-back from the star formation cirrus of class (b)cores appear not able to do so.• Even in the absence of tuning the initial propertiesof any of the simulations, we nonetheless find thatfour of the galaxies analyzed here ( m12b , m12f , m12m , & m12r ) are able to match CMZ SFR and gas surface density observations at some point ina 22 Myr period.• Intriguingly, of the simulated galaxies that simul-taneously match MW CMZ gas and SFR obser-vations, half have asymmetric, time-varying gasand SFR distributions ( i.e. , are in class ‘a’), whilstthe other half are fairly smooth class (b) galacticcenters. These results demonstrate that a time-varying model can account for the low star for-mation efficiency (per mass of dense gas) of theCMZ, and that it is not produced solely by somesteady state equilibrium. In fact, the presence ofbars and the influence of SMBHs may make class(b) galactic centers untenable in reality.In all, these simulated galaxies cover a wide range inSFRs and gas surface densities, exhibit marked morpho-logical differences, and some undergo significant changesin the span of only 22 Myr. The simulations lack SMBHsand (strong) bars, and so we cannot comment on thedirect role that either of those would play in shap-ing and/or regulating the core regions of these galax-ies. However, this letter highlights (1) the ability ofthe FIRE-2 zoom-in simulations to reproduce the “largescale” ( i.e. , 145 pc scale) properties of the CMZ; (2) themarked importance of asymmetric, time-varying ( i.e. ,bursty) star formation and feedback in shaping centralgalactic regions; (3) that future work with these simu-lations may help explain how the variety of naturallyoccurring conditions in central galactic environmentsarises. ACKNOWLEDGMENTSThe authors would like to thank Alexander Gurvich,and an anonymous referee, for helpful comments thatimproved the manuscript.CB and HPH gratefully acknowledge support from theNational Science Foundation under Award No. 1816715.HPH thanks the LSSTC Data Science Fellowship Pro-gram, which is funded by LSSTC, NSF Cybertrain-ing Grant M. E. Orr et al. sities for Research in Astronomy, Inc., for NASA, undercontract NAS5-26555. RSK acknowledges financial sup-port from the German Research Foundation (DFG) viathe Collaborative Research Center (SFB 881, Project-ID 138713538) ‘The Milky Way System’ (subprojects A1, B1, B2, and B8). He also thanks for funding fromthe Heidelberg Cluster of Excellence STRUCTURES inthe framework of Germany’s Excellence Strategy (grantEXC-2181/1 - 390900948) and for funding from the Eu-ropean Research Council via the ERC Synergy GrantECOGAL (grant 855130).REFERENCES
Anglés-Alcázar, D., Faucher-Giguère, C.-A., Quataert, E.,et al. 2017, Mon. Not. R. Astron. Soc. Lett., 472, L109,doi: 10.1093/mnrasl/slx161Armillotta, L., Krumholz, M. R., Di Teodoro, E. M., &McClure-Griffiths, N. M. 2019, arXiv:1905.01309[astro-ph]. http://arxiv.org/abs/1905.01309Barnes, A. T., Longmore, S. N., Battersby, C., et al. 2017,Monthly Notices of the Royal Astronomical Society, 469,2263, doi: 10.1093/mnras/stx941Barnes, A. T., Longmore, S. N., Avison, A., et al. 2019,MNRAS, 486, 283, doi: 10.1093/mnras/stz796Bemis, A., & Wilson, C. 2019, arXiv, 157, 131,doi: 10.3847/1538-3881/ab041dBenincasa, S. M., Wadsley, J., Couchman, H. M. P., &Keller, B. W. 2016, Mon. Not. R. Astron. Soc., 462, 3053,doi: 10.1093/mnras/stw1741Benincasa, S. M., Loebman, S. R., Wetzel, A., et al. 2020,Mon. Not. R. Astron. Soc., 497, 3993,doi: 10.1093/mnras/staa2116Blanc, G. A., Heiderman, A., Gebhardt, K., Evans, N. J., &Adams, J. 2009, Astrophys. J., 704, 842,doi: 10.1088/0004-637X/704/1/842Brownson, S., Belfiore, F., Maiolino, R., Lin, L., &Carniani, S. 2020, Mon. Not. R. Astron. Soc. Lett., 498,L66, doi: 10.1093/mnrasl/slaa128Casasola, V., Hunt, L., Combes, F., & García-Burillo, S.2015, Astron. Astrophys., 577, A135,doi: 10.1051/0004-6361/201425313Clark, P. C., Glover, S. C. O., Ragan, S. E., &Duarte-Cabral, A. 2019, Mon. Not. R. Astron. Soc., 486,4622, doi: 10.1093/mnras/stz1119Debattista, V. P., Gonzalez, O. A., Sanderson, R. E., et al.2019, Mon. Not. R. Astron. Soc., 485, 5073,doi: 10.1093/mnras/stz746Gallagher, M. J., Leroy, A. K., Bigiel, F., et al. 2018,Astrophys. J., 858, 90, doi: 10.3847/1538-4357/aabad8Garrison-Kimmel, S., Hopkins, P. F., Wetzel, A., et al.2018, Mon. Not. R. Astron. Soc., 481, 4133,doi: 10.1093/mnras/sty2513Ginsburg, A., Bally, J., Barnes, A., et al. 2018, Astrophys.J., 853, 171, doi: 10.3847/1538-4357/aaa6d4 Gurvich, A. B., Faucher-Giguère, C.-A., Richings, A. J.,et al. 2020, Mon. Not. R. Astron. Soc., 498, 3664,doi: 10.1093/mnras/staa2578Henshaw, J. D., Ginsburg, A., Haworth, T. J., et al. 2019,MNRAS, 485, 2457, doi: 10.1093/mnras/stz471Hopkins, P. F., Kereš, D., Onorbe, J., et al. 2014, Mon. Not.R. Astron. Soc., 445, 581, doi: 10.1093/mnras/stu1738Hopkins, P. F., Wetzel, A., Kereš, D., et al. 2018, Mon. Not.R. Astron. Soc., 480, 800, doi: 10.1093/mnras/sty1690Immer, K., Schuller, F., Omont, A., & Menten, K. M. 2012,Astronomy & Astrophysics, 537, A121,doi: 10.1051/0004-6361/201117857Kennicutt, R. C., & Evans, N. J. 2012, Annu. Rev. Astron.Astrophys., 50, 531,doi: 10.1146/annurev-astro-081811-125610Kennicutt, Jr., R. C. 1998, Astrophys. J., 498, 541,doi: 10.1086/305588Kormendy, J., & Kennicutt, R. C. 2004, Annu. Rev.Astron. Astrophys., 42, 603,doi: 10.1146/annurev.astro.42.053102.134024Kruijssen, J. M. D., Dale, J. E., & Longmore, S. N. 2015,Mon. Not. R. Astron. Soc., 447, 1059,doi: 10.1093/mnras/stu2526Kruijssen, J. M. D., Longmore, S. N., Elmegreen, B. G.,et al. 2014, Mon. Not. R. Astron. Soc., 440, 3370,doi: 10.1093/mnras/stu494Krumholz, M. R., & Gnedin, N. Y. 2011, Astrophys. J.,729, 36, doi: 10.1088/0004-637X/729/1/36Krumholz, M. R., Kruijssen, J. M. D., & Crocker, R. M.2017, MNRAS, 466, 1213, doi: 10.1093/mnras/stw3195Lada, C. J., Forbrich, J., Lombardi, M., & Alves, J. F.2012, ApJ, 745, 190, doi: 10.1088/0004-637X/745/2/190Lada, C. J., Lombardi, M., & Alves, J. F. 2010, Astrophys.J., 724, 687, doi: 10.1088/0004-637X/724/1/687Leitherer, C., Schaerer, D., Goldader, J. D., et al. 1999,Astrophys. J. Suppl. Ser., 123, 3, doi: 10.1086/313233Leroy, A. K., Walter, F., Sandstrom, K., et al. 2013,Astron. J., 146, 19, doi: 10.1088/0004-6256/146/2/19Leroy, A. K., Bolatto, A. D., Ostriker, E. C., et al. 2018,ApJ, 869, 126, doi: 10.3847/1538-4357/aaecd1 tar Formation Distributions in Fiery Cores Liermann, A., Hamann, W.-R., & Oskinova, L. M. 2012,Astron. Astrophys., 540, A14,doi: 10.1051/0004-6361/201117534Lin, L., Ellison, S. L., Pan, H.-A., et al. 2020, Astrophys.J., 903, 145, doi: 10.3847/1538-4357/abba3aLongmore, S. N., Bally, J., Testi, L., et al. 2013, Mon. Not.R. Astron. Soc., 429, 987, doi: 10.1093/mnras/sts376Lu, J. R., Do, T., Ghez, A. M., et al. 2013, Astrophys. J.,764, 155, doi: 10.1088/0004-637X/764/2/155Mills, E. A. C., & Battersby, C. 2017, Astrophys. J., 835,76, doi: 10.3847/1538-4357/835/1/76Molinari, S., Bally, J., Noriega-Crespo, A., et al. 2011,Astrophys. J., 735, L33,doi: 10.1088/2041-8205/735/2/L33Moreno, J., Torrey, P., Ellison, S. L., et al. 2020, Mon. Not.R. Astron. Soc., 000, doi: 10.1093/mnras/staa2952Orr, M. E., Hayward, C. C., & Hopkins, P. F. 2019, Mon.Not. R. Astron. Soc., 486, 4724,doi: 10.1093/mnras/stz1156Orr, M. E., Hayward, C. C., Hopkins, P. F., et al. 2018,Mon. Not. R. Astron. Soc., 478, 3653,doi: 10.1093/mnras/sty1241Orr, M. E., Hayward, C. C., Medling, A. M., et al. 2020,Mon. Not. R. Astron. Soc., 496, 1620,doi: 10.1093/mnras/staa1619Rathborne, J. M., Longmore, S. N., Jackson, J. M., et al.2014, The Astrophysical Journal, 786, 140,doi: 10.1088/0004-637X/786/2/140Renaud, F., Bournaud, F., Agertz, O., et al. 2019, Astron.Astrophys., 625, A65, doi: 10.1051/0004-6361/201935222Sanderson, R. E., Wetzel, A., Loebman, S., et al. 2020,Astrophys. J. Suppl. Ser., 246, 6,doi: 10.3847/1538-4365/ab5b9d Schmidt, M. 1959, Astrophys. J., 129, 243,doi: 10.1086/146614Sormani, M. C., Binney, J., & Magorrian, J. 2015, Mon.Not. R. Astron. Soc., 449, 2421,doi: 10.1093/mnras/stv441Sormani, M. C., & Li, Z. 2020, Mon. Not. R. Astron. Soc.,494, 6030, doi: 10.1093/mnras/staa1139Sormani, M. C., Tress, R. G., Glover, S. C. O., et al. 2020,MNRAS, 497, 5024, doi: 10.1093/mnras/staa1999Sormani, M. C., Treß, R. G., Ridley, M., et al. 2018, Mon.Not. R. Astron. Soc., 475, 2383,doi: 10.1093/mnras/stx3258Tacchella, S., Dekel, A., Carollo, C. M., et al. 2016a, Mon.Not. R. Astron. Soc., 457, 2790,doi: 10.1093/mnras/stw131Tacchella, S., Dekel, A., Marcella Carollo, C., et al. 2016b,Mon. Not. R. Astron. Soc., 458, 242,doi: 10.1093/mnras/stw303Torrey, P., Hopkins, P. F., Faucher-Giguère, C. A., et al.2017, Mon. Not. R. Astron. Soc., 467, 2301,doi: 10.1093/mnras/stx254Tress, R. G., Sormani, M. C., Glover, S. C. O., et al. 2020,arXiv:2004.06724 [astro-ph].https://arxiv.org/abs/2004.06724Velázquez, J. A. F., Gurvich, A. B., Faucher-Giguére,C.-A., et al. 2020, Mon. Not. R. Astron. Soc., 13, 1,doi: 10.1093/mnras/staa3893Walker, D. L., Longmore, S. N., Zhang, Q., et al. 2018,MNRAS, 474, 2373, doi: 10.1093/mnras/stx2898Wetzel, A. R., Hopkins, P. F., Kim, J.-h., et al. 2016,Astrophys. J., 827, L23,doi: 10.3847/2041-8205/827/2/L23Yusef-Zadeh, F., Hewitt, J. W., Arendt, R. G., et al. 2009,Astrophys. J., 702, 178,doi: 10.1088/0004-637X/702/1/178 M. E. Orr et al. . . . . . . . (cid:113) v x + v y /v c . . . . . . N / N p e a k Class (a) Centers . . . . . . . (cid:113) v x + v y /v c . . . . . . Class (b) Centers
Figure 4.
Distribution of in-plane cold & dense gas velocities (relative to the calculated circular velocity) within the central
R < pc of the seven FIRE-2 spiral galaxies analyzed in this letter at the final snapshot, split into their respective ‘classes’(colors and classifications as Fig. 2). Cold & dense gas averaged into pixels 10 pc on a side. Dashed vertical line plotted at √ in each panel, noting the classical ‘escape’ velocity, relative to calculated circular velocity (this is not to say that materialabove √ v c is unbound here, since the galaxy centers are not point masses, but it does indicate highly non-circular orbits). Leftpanel: class (a) “asymmetric” centers, m12b , m12c , m12f , & m12w , these galaxy centers have highly variable gas velocitydistributions snapshot-to-snapshot (here, e.g. , m12c has very little cold & dense gas, and the distribution from m12w is fairlynarrowly centered on v c , but in previous snapshots both more closely resemble the other two broad class (a) distributions shownhere), and generally have a significant fraction of material with (cid:112) v x + v y > √ v c . And so, the in-plane gas orbital dynamics arehighly variable over short periods ( ∆ t ∼ . Myr), exhibiting high amounts of non-circularity.
Right panel: class (b) “smooth”centers, m12i , m12m , & m12r , these galaxies have gas velocity distributions centered close to v c , with very little gas exceeding √ v c . In short, gas is predominantly moving on circular orbits in these galaxies. Further, snapshot-to-snapshot, these galaxycenters exhibit much less variation in the shapes of their velocity distributions: they are fairly stable in their orbital dynamics. This figure does not appear in the ApJ version of this manuscript, and as such is supplementary material.
APPENDIX A. (SUPPLEMENTAL) DISTRIBUTION OF IN-PLANE GAS VELOCITIES IN CLASS (A) & (B) CENTERS This appendix does not appear in the ApJ version of this manuscript, and as such is supplementary material.
Inaddition to our by-eye classification of the seven galaxy centers analyzed here, we briefly investigated the distributionsof in-plane gas velocities. Fig. 4 shows the distributions of gas velocities within
R < pc for the two classes of galaxycenter at the final snapshot of the simulations. As described in the main text, there is a significant difference betweenthe two classes of galaxy centers, and this extends to this analysis of the velocity distributions: class (a) centers exhibita large amount of non-circular in-plane gas motion, and are highly variable snapshot-to-snapshot, whereas the class(b) centers have predominantly circular orbital motions, and the distributions are fairly stable snapshot-to-snapshot.Radial motion distributions (not shown) largely tell the same story, class (a) galaxies show (in some cases very large)asymmetry in the distribution of inward and outward flowing gas, and class (b) centers have roughly stable andsymmetric distributions (little netnet