Accepted for Publication in ApJL

Preprint typeset using LATEX style emulateapj v. 12/16/11

STELLAR METALLICITIES AND ELEMENTAL ABUNDANCE RATIOS OF z ∼ 1.4 MASSIVE QUIESCENT

GALAXIES*
Mariska Kriek1 , Sedona H. Price2 , Charlie Conroy3 , Katherine Suess1 , Lamiya Mowla4 , Imad Pasha4 , Rachel
Bezanson5 , Pieter van Dokkum4 & Guillermo Barro6
Accepted for Publication in ApJL

ABSTRACT
arXiv:1907.04327v1 [astro-ph.GA] 9 Jul 2019

The chemical composition of galaxies has been measured out to z ∼ 4. However, nearly all studies
beyond z ∼ 0.7 are based on strong-line emission from HII regions within star-forming galaxies.
Measuring the chemical composition of distant quiescent galaxies is extremely challenging, as the
required stellar absorption features are faint and shifted to near-infrared wavelengths. Here, we
present ultra-deep rest-frame optical spectra of five massive quiescent galaxies at z ∼ 1.4, all of which
show numerous stellar absorption lines. We derive the abundance ratios [Mg/Fe] and [Fe/H] for three
out of five galaxies; the remaining two galaxies have too young luminosity-weighted ages to yield
robust measurements. Similar to lower-redshift findings, [Mg/Fe] appears positively correlated with
stellar mass, while [Fe/H] is approximately constant with mass. These results may imply that the
stellar mass-metallicity relation was already in place at z ∼ 1.4. While the [Mg/Fe]-mass relation at
z ∼ 1.4 is consistent with the z < 0.7 relation, [Fe/H] at z ∼ 1.4 is ∼ 0.2 dex lower than at z < 0.7.
With a [Mg/Fe] of 0.44+0.08
−0.07 the most massive galaxy may be more α-enhanced than similar-mass
galaxies at lower redshift, but the offset is less significant than the [Mg/Fe] of 0.6 previously found
for a massive galaxy at z = 2.1. Nonetheless, these results combined may suggest that [Mg/Fe] in the
most massive galaxies decreases over time, possibly by accreting low-mass, less α-enhanced galaxies.
A larger galaxy sample is needed to confirm this scenario. Finally, the abundance ratios indicate short
star-formation timescales of 0.2 − 1.0 Gyr.
Keywords: Galaxies: evolution — Galaxies: formation

1. INTRODUCTION chemical composition, as stars in accreted galaxies have

The chemical composition of a galaxy reflects the their own chemical footprint (see Maiolino & Mannucci
interplay of several fundamental physical processes in 2019, and references therein).
galaxy formation, including star formation, metal pro- Metallicities of low-redshift galaxies have been stud-
duction, feedback and gas exchange with the surround- ied extensively for both the interstellar gas (e.g.,
ing medium, and galaxy merging. Subsequent star- Tremonti et al. 2004) as well as the gas locked up in stars
formation episodes and the recycling of enriched gas (e.g., Gallazzi et al. 2005). To disentangle the effects of
result in an increase of metallicity with time, while star formation, feedback/gas exchange, and galaxy merg-
the duration of the star-forming phase sets the relative ers on the chemical abundance patterns of galaxies, it is
abundances of different metals. Feedback processes fur- crucial to extend these studies to higher redshifts. Metal-
ther impact the chemical enrichment history, as they licities of star-forming galaxies have been studied out
may expel enriched gas from galaxies. In combina- to z ∼ 4 (e.g., Erb et al. 2006; Mannucci et al. 2009;
tion with inflow of lower metallicity gas from the in- Shapley et al. 2017). However, the chemical composi-
ter/circumgalactic medium, they alter the metal con- tion of quiescent galaxies has only been routinely mea-
tent of the gas supply. Finally, galaxy mergers affect the sured out to z ∼ 0.7 (e.g., Gallazzi et al. 2014; Choi et al.
2014). Because quiescent galaxies dominate the massive
* Based on data obtained at the W.M. Keck Observatory and galaxy population out to z ∼ 2 (Muzzin et al. 2013b;
with the NASA/ESA Hubble Space Telescope (HST). W.M. Tomczak et al. 2014), our current understanding of the
Keck Observatory is operated as a scientific partnership among chemical composition of galaxies through cosmic time is
the California Institute of Technology, the University of Califor- thus incomplete.
nia and NASA, and was made possible by the generous financial
support of the W.M. Keck Foundation. HST is operated by the Measuring chemical compositions of z > 0.7 quiescent
Association of Universities for Research in Astronomy, Inc., un- galaxies is extremely challenging as the required stellar
der NASA contract NAS 5-26555. absorption lines are faint and shifted to near-infrared
1 Astronomy Department, University of California, Berkeley,
CA 94720, USA
wavelengths. Furthermore, because of the younger stel-
2 Max Planck Institute for Extraterrestrial Physics, Giessen- lar ages, the metal absorption lines for distant quies-
bachstrasse 1, D-85741 Garching, Germany cent galaxies are weaker than for their local analogs.
3 Harvard-Smithsonian Center for Astrophysics, 60 Garden
The few available measurements at z > 0.7 are ei-
St., Cambridge, MA, USA
4 Department of Astronomy, Yale University, New Haven, CT ther based on a stacked spectrum of quiescent galax-
06511, USA ies (e.g., Onodera et al. 2015), or on an ultradeep spec-
5 Department of Physics and Astronomy and PITT PACC,
trum of a single quiescent galaxy (e.g., Toft et al. 2012;
University of Pittsburgh, Pittsburgh, PA, 15260, USA Lonoce et al. 2015; Kriek et al. 2016). There are also
6 Department of Physics, University of the Pacific, 3601 Pacific
Avenue, Stockton, CA 95211, USA several studies that use low-resolution grism spectra
2 Kriek et al.
1.5
217249
2.0 213947
214340
213931 1.0
214695

log (Re/kpc)
1.5
U-V

0.5

1.0
0.0

0.5 -0.5
0.5 1.0 1.5 10.0 10.5 11.0 11.5
V-J log (M / MO• )
Figure 1. Left: Rest-frame U − V vs. V − J of the five quiescent galaxies observed with MOSFIRE and LRIS in comparison to all
galaxies with 1.3 < z < 1.5 and log (M/M⊙ ) > 10 in the UltraVISTA field. Right: Rest-frame optical size (at 5000 Å) vs stellar mass of
the same five galaxies in comparison to all galaxies with 1.3 < z < 1.5 and log (M/M⊙ ) > 10 from Mowla et al. (2018). The blue and
red lines represent the best-fit relations for star-forming and quiescent galaxies, respectively, at z ∼ 1.4 (Mowla et al. 2018). The most
massive galaxy, 213931, consists of three clumps. In this panel we show the size and mass of the brightest clump, where the mass has been
estimated using the magnitude ratios of the clumps.
obtained with the Hubble Space Telescope (HST) to lines in atmospheric windows. Galaxies were classified as
derive stellar metallicities (e.g., Morishita et al. 2018; quiescent based on their rest-frame U − V and V − J col-
Estrada-Carpenter et al. 2018). However, because these ors (e.g., Wuyts et al. 2007). Furthermore, we required
studies primarily rely on the continuum shape, they are that the pointing overlaps with the COSMOS-DASH sur-
especially susceptible to modeling degeneracies and other vey (Momcheva et al. 2017; Mowla et al. 2018), which
systematic errors. provides shallow F160W imaging and thus allows the
Early results suggest that z > 0.7 quiescent galax- measurement of rest-frame optical sizes. Figure 1 shows
ies have super-solar metallicities (Onodera et al. 2015; the location of the five targets in the rest-frame U − V
Lonoce et al. 2015; Kriek et al. 2016) and are α- vs. V − J (UVJ) diagram, as well as in Re vs. stel-
enhanced, with [Mg/Fe] that are similar (Onodera et al. lar mass space, compared to the full galaxy distribution
2015) or significantly higher (Lonoce et al. 2015; with M > 1010 M⊙ and 1.3 < z < 1.5 (Muzzin et al.
Kriek et al. 2016) than those of similar-mass low-redshift 2013b; Mowla et al. 2018). The galaxies span a range in
galaxies. These high [Mg/Fe] imply very short star- colors along the quiescent sequence in the UVJ diagram,
formation timescales, though the stellar initial mass func- as well as a range in sizes. On average, they are slightly
tion (IMF) may also affect this abundance ratio (e.g., bluer and smaller than the typical quiescent galaxy at
Fontanot et al. 2017). Furthermore, these measurements this redshift. This bias may be expected; our magni-
raise the question of how high-redshift quiescent galax- tude selection favors post-starburst galaxies, which are
ies with high [Mg/Fe] evolve into the early-type galaxy brighter, bluer, and presumably smaller than older quies-
population with lower [Mg/Fe] seen today. Larger galaxy cent galaxies of similar mass (e.g., Whitaker et al. 2012;
samples are needed to confirm these results Yano et al. 2016; Almaini et al. 2017).
In this Letter we present elemental abundance ratios The LRIS mask containing our five primary targets was
for five massive galaxies at z ∼ 1.4, derived from deep observed for ∼4.5 hrs on 2017/01/04 using the 600/10000
spectra obtained with LRIS and MOSFIRE on the Keck red grating, 1′′ -wide slits, and an ABC dither pattern.
I Telescope. This study is enabled by the large Ultra- The sky was clear with an image quality of ∼ 0.′′ 8 − 1.′′ 0.
VISTA (McCracken et al. 2012) and COSMOS-DASH The same five galaxies were observed within one MOS-
(Momcheva et al. 2017; Mowla et al. 2018) field, which FIRE mask on 2017/03/15 and 2017/04/5-6 for ∼12 hrs
facilitated the identification of pointings for which we can in J, using 0.′′ 7-wide slits and an ABA’B’ dither pattern
observe several bright targets (with HST/F160W imag- (Kriek et al. 2015). The sky was clear and the seeing var-
ing) simultaneously. Throughout this work we assume ied between 0.′′ 5-1.′′ 0. For both instruments we assigned
a ΛCDM cosmology with Ωm = 0.3, ΩΛ = 0.7, and a star to one of the slits to monitor weather conditions
H0 = 70 km s−1 Mpc−1 . and aid the data reduction.
The LRIS and MOSFIRE 2D spectra were reduced us-
2. GALAXY SAMPLE AND DATA ing custom software. Initial sky subtraction was per-
formed using the average of the surrounding frames with
The observed galaxies were identified using the Ultra-
the same integration times. Cosmic rays were identi-
VISTA K-band selected catalog (v4.1) by Muzzin et al.
fied using L.A. Cosmic (van Dokkum 2001) and com-
(2013a). We selected the pointing for which we could
bined with a bad pixel map. Next, the individual sky-
observe the most J < 21.6 quiescent galaxies at 1.3 <
subtracted frames and corresponding masks were resam-
z < 1.5 in one MOSFIRE/LRIS mask. For this redshift
pled onto a common grid to account for the wavelength
range, we catch prominent metal and Balmer absorption
 Abundance Patterns of z ∼ 1.4 Galaxies 3
LRIS-RED MOSFIRE-J
fλ (10-18 erg-1 s-1 cm-2 Angstrom-1) 3
CaIIH MgI FeI FeI FeI UltraVISTA(v4.1)

LRIS-RED
HH H
217249
z = 1.377
2 2
2

MOSFIRE-J
1 1

0
3000 3200 3400 3600 3800 4000 5000 5200 5400 1
λobserved (µm)
fλ (10-18 erg-1 s-1 cm-2 Angstrom-1)

LRIS-RED
5 CaII H 4 H MgI FeI FeI FeI 5 UltraVISTA(v4.1)
H H H
213947
4 4 z = 1.397

3 3
3
2 2

MOSFIRE-J
1
1
2 0
3400 3600 3800 4000 4200 5000 5200 5400 5600 1
λobserved (µm)
fλ (10-18 erg-1 s-1 cm-2 Angstrom-1)

LRIS-RED
CaII H
3
H MgI FeI FeI FeI UltraVISTA(v4.1)
3 H H H 3 214340
z = 1.418

2 2

MOSFIRE-J
2 1
1

0
3400 3600 3800 4000 4200 4800 5000 5200 5400 1
λobserved (µm)
fλ (10-18 erg-1 s-1 cm-2 Angstrom-1)

15 CaII H H MgI FeI FeI FeI UltraVISTA(v4.1)

H H H
LRIS-RED
15 213931
15 z = 1.399

10
MOSFIRE-J
10

5
5
0
3600 3800 4000 4200 5000 5200 5400 5600 1
λobserved (µm)
fλ (10-18 erg-1 s-1 cm-2 Angstrom-1)

LRIS-RED

CaII H H MgI FeI FeI FeI UltraVISTA(v4.1)

3 H H H 3
3 214695
z = 1.396

2 2
MOSFIRE-J

2 1
1

0
3400 3600 3800 4000 4200 5000 5200 5400 5600 1
λrest-frame (Angstrom) λrest-frame (Angstrom) λobserved (µm)
Figure 2. Left: LRIS-RED and MOSFIRE/J-band spectra of 5 massive quiescent galaxies at z ∼ 1.4 (black). The spectra are binned by
10 pixels, corresponding to ∼3.3 and ∼5.4 rest-frame Å per bin for the LRIS and MOSFIRE spectra, respectively. The gray areas represent
the 1 σ uncertainty for the binned spectra. The best-fit alf models are shown in red. The F814W image (4.′′ 5×4.′′ 5) is shown in the inset,
with the MOSFIRE slit overplotted (91o , 0.′′ 7 width). The LRIS slit had a similar orientation (93o ) and width (1′′ ). Right: UltraVISTA
photometry (circles) and best-fit FSPS model (grey) for the same galaxies.
4 Kriek et al.

Table 1 a high order polynomial in order to avoid potential issues

Parameters of z ∼ 1.4 Quiescent Galaxies with the flux calibration of the data. The fitting is done
using a Markov chain Monte Carlo (MCMC) algorithm
IDa FAST alf (Foreman-Mackey et al. 2013). This spectral modeling
log M σv [Fe/H] [Mg/Fe] t tSSP approach is strongly preferred over the use of integrated
M⊙ km/s Gyr Gyr absorption line measurements for distant galaxies (see
217249 10.61 127+58
−29 −1.03+0.47
−0.27 0.52+0.28
−0.35 3.0+0.4
−0.6 0.9+0.8
−0.4 discussion in Kriek et al. 2016).
213947 10.87 170+25
−19 −0.89+0.68
−0.24 0.57+0.24
−0.32 3.0+0.1
−1.0 0.8+0.2
−0.3 For our z ∼ 1.4 galaxies, we assume the Kroupa
214340 10.80 79+24
−27 −0.42+0.16
−0.24 0.22+0.19
−0.14 3.8+5.3
−0.8 1.4+0.6
−0.5 (2001) IMF and allow for a double-component stellar
213931b 11.73 342+12
−11 −0.27+0.07
−0.07 0.44+0.08
−0.07 3.1+0.2
−0.1 2.0+0.8
−0.5 population with two different ages. This model is pre-
214695 11.18 209+30
−33 −0.20+0.17
−0.22 0.28+0.15
−0.14 4.5+2.9
−1.2 3.8+7.3
−1.6 ferred over a single-age model (i.e., SSP), because it in-
a UltraVISTA catalog v4.1 by Muzzin et al. (2013a) directly separates the younger and older stellar popula-
b The mass for this galaxy represents the total mass, including all tions and results in mass-weighted abundance measure-
three clumps. ments. In case we assume an SSP, the younger stars will
disproportionally dominate the results.In Table 1 we list
calibration, dither position, distortions, and drifts. An mass-weighted abundance ratios and ages for the double-
additional sky subtraction was performed by subtract- component model, as well as the ages for an SSP. Due
ing the median background at each wavelength. For the to the mass weighing, the double-component ages are
MOSFIRE spectra, the slit stars were used to derive older than the SSP ages (which are closer to luminosity-
a weighing factor for each science exposure. All LRIS weighted ages). While all abundances are free parame-
frames were weighted equally, since the weather condi- ters, only few can be accurately determined. We also
tions were stable. determine the total stellar metallicity using [Z/H] =
The individual spectra were average combined, taking [Fe/H] + 0.94[Mg/Fe] (Thomas et al. 2003), and derive
into account both the weighing factors and the corre- the 16% and 84% confidence intervals from the [Z/H]
sponding rectified masks. The relative flux calibration distribution following from the MCMC simulations.
was performed using a response spectrum. For MOS- Stellar masses are derived by fitting the combined
FIRE the response spectrum was derived from the spec- continuum spectra and UltraVISTA photometry with
tra of A0 V stars. For LRIS we used the theoretical at- the Flexible Stellar Population Synthesis (FSPS) mod-
mospheric absorption spectrum, adjusted to match the els (Conroy et al. 2009), using the FAST fitting code
atmospheric features in the spectrum of the slit star, (Kriek et al. 2009). We assume solar metallicity, a de-
combined with the intrinsic shape of the slit star and layed exponential star-formation history, the Chabrier
other bright objects in the mask. More details on the (2003) IMF, and the average attenuation law by
MOSFIRE reduction software, which was developed for Kriek & Conroy (2013). We caution though that these
the MOSDEF survey, are given in Kriek et al. (2015). stellar masses are likely underestimated, and more com-
One-dimensional (1D) spectra were extracted using an plicated star formation histories, like the one assumed
optimal weighing procedure. The combined LRIS and in the alf fitting, would lead to masses that are ∼ 50%
MOSFIRE spectra for each galaxy, together with the larger.
HST F814W image (Scoville et al. 2007) and the photo- Rest-frame optical sizes are derived from the combi-
metric spectral energy distribution (Muzzin et al. 2013b) nation of F814W data from Scoville et al. (2007) and
are shown in Figure 2. We detect numerous stellar ab- F160W data from COSMOS-DASH, following the pro-
sorption lines for all five galaxies. cedure described in Mowla et al. (2018). This proce-
The most massive galaxy in the sample, 213931, con- dure relies on Sérsic fits and the GALFIT modeling code
sists of three separate clumps, which were blended to- (Peng et al. 2002). We derive the (non-circularized) size
gether as one system in the UltraVISTA images and cat- at 5000 Å using a linear interpolation between the F814W
alog. As illustrated in Figure 2, the slit was aligned and F160W sizes. For 213931, we measure the size of the
along the two most massive clumps. With MOSFIRE brightest clump.
we detect two blended traces, though one of the traces
is significantly fainter, and no separate analysis could be 4. RESULTS
performed. Using LRIS, we do not detect two separate In Figure 3, we show the abundance ratios [Mg/Fe]
traces, and thus we treat 213931 as one system in our and [Fe/H] of the z ∼ 1.4 galaxies in compari-
analysis. son with galaxies at lower redshift (Choi et al. 2014;
Conroy & van Dokkum 2012) and a single massive qui-
3. ANALYSIS escent galaxy at z = 2.1 (Kriek et al. 2016). Although
We use the absorption line fitter (alf) code to es- metal and Balmer absorption lines are detected for all
timate parameters from the combined 1D LRIS and five galaxies, the abundance ratios for 217249 and 213947
MOSFIRE spectra. The code is described in detail are poorly constrained. Figures 1 and 2 show that
in Conroy & van Dokkum (2012); Conroy et al. (2014); these galaxies have blue rest-frame colors, strong Balmer
Choi et al. (2014). In summery, alf combines libraries lines, and are dominated by A-type stars. Hence, their
of isochrones and empirical stellar spectra with synthetic luminosity-weighted ages are relatively young, and thus
stellar spectra covering a wide range of elemental abun- their metal lines are weak. Therefore, in Figure 3 we
dance patterns. The code fits for C, N, O, Na, Mg, Ca, only show the three redder quiescent galaxies at z ∼ 1.4.
Ti, V, Cr, Mn, Fe, Co, and Ni abundances, redshift, For these galaxies we find similar abundance ratios when
velocity dispersion, stellar population age, and several adopting a single-age model.
emission lines. The ratio of the model and data is fit by In the left panels of Figure 3, we show the abundance
 Abundance Patterns of z ∼ 1.4 Galaxies 5

0.07 < z < 0.09 Nearby galaxies

0.10 < z < 0.20 z = 1.4
0.6 0.6
0.20 < z < 0.30 z = 2.1
0.30 < z < 0.40
0.40 < z < 0.55
0.55 < z < 0.70
0.4 0.4
z = 1.4
[Mg/Fe]

[Mg/Fe]
z = 2.1

0.2 0.2

0.0 0.0

1010 1011 1012 100 200 300 400

M ( MO• ) σv (km/s)

0.0 0.0

-0.2 -0.2
[Fe/H]

[Fe/H]

-0.4 -0.4

-0.6 -0.6

1010 1011 1012 100 200 300 400

M ( MO• ) σv (km/s)
0.6 0.6

0.4 0.4

0.2 0.2
[Z/H]

[Z/H]

0.0 0.0

-0.2 -0.2

-0.4 -0.4

1010 1011 1012 100 200 300 400

M ( MO• ) σv (km/s)
Figure 3. [Mg/Fe] (top), [Fe/H] (middle) and [Z/H] (bottom) vs stellar mass (left) and velocity dispersion (right) of the z ∼ 1.4 quiescent
galaxies 214340 (yellow), 213931 (orange), and 214695 (red), in comparison to nearby early-type galaxies (plusses; Conroy & van Dokkum
2012), stacks of quiescent galaxies at 0.07 < z < 0.70 (squares; Choi et al. 2014), and a quiescent galaxy at z = 2.1 (black filled circle;
Kriek et al. 2016). For all (stacks of) galaxies, the abundance ratios and velocity dispersions were derived using alf. The relations at
z ∼ 1.4 are similar to those at z < 0.7, though [Fe/H] seems offset to lower values. The bottom panels show that a tentative stellar
mass-metallicity relation may already be in place at z ∼ 1.4, possibly offset to lower metallicities.
6 Kriek et al.

their stars over a brief time period of ∼ 0.2 − 1 Gyr.

As discussed in Kriek et al. (2016), by adopting differ-
0.6 ent core-collapse supernova yields or a different Type Ia
0.1 delay time distribution, we can change this model sub-
0.2 stantially.
0.4 0.5
5. DISCUSSION
[Mg/Fe]

In this Letter, we present ultradeep rest-frame opti-

cal spectra of five massive quiescent galaxies at z ∼ 1.4,
0.2 5.0
all of which show multiple stellar absorption lines. For
three galaxies we derive the abundance ratios [Mg/Fe]
and [Fe/H], but the remaining two galaxies have too
young luminosity-weighted ages to yield robust measure-
0.0 ments. Similar to z < 0.7 studies, we find a tentative
positive relation between [Mg/Fe] and stellar mass (or
velocity dispersion). Also similar to z < 0.7, we find no
-0.6 -0.4 -0.2 0.0 correlation between [Fe/H] and stellar mass (or veloc-
[Fe/H] ity dispersion). Our results may imply that the stellar
Figure 4. [Mg/Fe] vs [Fe/H] for massive quiescent galaxies at z ∼
mass-metallicity relation was already in place at z ∼ 1.4
1.4 (colored symbols), in comparison to nearby early-type galaxies (bottom panels, Fig. 3).
(plusses), stacks of quiescent galaxies at 0.07 < z < 0.70 (gray While the [Mg/Fe]-mass relation at z ∼ 1.4 is con-
squares), and a quiescent galaxy at z = 2.1 (black filled circle). sistent with the z . 0.7 relation, [Fe/H] at z ∼ 1.4 is
Symbols are the same as in Figure 3. The dashed line represents a
chemical evolution model, with different star-formation timescales ∼0.2 dex lower than at z < 0.7. We also found a low
in Gyr indicated by the small black squares. The abundance ratios [Fe/H] for a single massive quiescent galaxy at z = 2.1
of the two most massive galaxies (orange and black circles) at z & (Kriek et al. 2016). In addition to the low [Fe/H], this
1.4 suggest that their star-formation timescales are shorter than z = 2.1 galaxy was more α-enhanced than low-redshift
for z < 0.7 galaxies.
galaxies of similar mass, with a [Mg/Fe] of 0.6. The most
ratios [Mg/Fe] and [Fe/H] versus stellar mass. We com- massive galaxy in the current sample has a [Mg/Fe] of
pare our results to the work by Choi et al. (2014), based 0.44+0.08
−0.07 , and thus may also be more α-enhanced than
on stacked spectra of quiescent galaxies at 0.07 < z < lower-redshift analogs. Combined, these results may sug-
0.7. In the right panels we show the abundance ratios gest that [Mg/Fe] of the most massive galaxies decreases
as a function of the observed integrated stellar veloc- over cosmic time, possibly by accretion of low-mass, less
ity dispersion (σv ), in comparison to the nearby early- α-enhanced galaxies.
type galaxy sample by Conroy & van Dokkum (2012). A similar scenario has been proposed to explain the
All abundance ratios and velocity dispersions have been evolution in the mass-size relation of quiescent galax-
derived using the alf code. However, stellar masses and ies between z ∼ 2 and z ∼ 0 (e.g., van Dokkum et al.
velocity dispersions were not consistently derived for all 2008; Bezanson et al. 2009; Naab et al. 2009). In this
samples, and thus we only show the comparison samples context, we note that all three z ∼ 1.4 galaxies as well
if the measurements are available. as the z ∼ 2.1 galaxy have close neighbors, and thus
Similar to z < 0.7, [Mg/Fe] appears positively corre- may be in the process of merging with smaller galaxies
lated with both stellar mass and σv at z ∼ 1.4. These re- (see also Gu et al. 2018). However, as [Fe/H] is constant
sults, however, are based on only three galaxies; a larger with mass, it is not obvious how this scenario could in-
sample is needed to confirm these trends. The two lower- crease [Fe/H], and a larger galaxy sample is needed to
mass galaxies (214340 & 214695) have [Mg/Fe] similar understand the evolutionary scenario (Choi et al. 2014).
to their lower-redshift analogs. With a [Mg/Fe] of 0.44, Other possible scenarios include late-time star forma-
the most massive galaxy (213931) is slightly offset from tion and the growth of the quiescent galaxy population
the low-redshift [Mg/Fe]-mass and [Mg/Fe]-σv relations. over time. Galaxies that stop forming stars at later
This offset, however, is not significant, and larger galaxy times will have lower [Mg/Fe], higher [Fe/H], and larger
samples are needed to assess whether the most massive sizes (Khochfar & Silk 2006). Thus, once these galaxies
galaxies are indeed more α-enhanced at earlier times. join the quiescent population, they will alter the aver-
Similar to the low-redshift samples, we find no correla- age size and abundance ratios (e.g., Carollo et al. 2013;
tion between [Fe/H] and mass or σV . We do find an offset Choi et al. 2014).
in [Fe/H], such that the z ∼ 1.4 galaxies are deficient in Our metallicities are higher than the sub-solar metal-
iron compared to similar-mass low-redshift galaxies. licities found by Morishita et al. (2018) for two 1011 M⊙
In Figure 4 we compare the combined abundance ra- quiescent galaxies at z ∼ 2.2, but lower than the
tios [Mg/Fe] and [Fe/H] to a closed-box chemical evolu- supersolar metallicities found for a stack of massive
tion model for different star-formation timescales. This (log M/M⊙ ∼ 11.4) galaxies at z ∼ 1.6 ([Z/H]=
model assumes a Salpeter (1955) IMF, a constant SFR, 0.24+0.20
−0.14 ; Onodera et al. 2015), our z = 2.1 galaxy
the core-collapse and Type Ia supernova yield models by ([Z/H]= 0.30 ± 0.07; Kriek et al. 2016), as well as the
Kobayashi et al. (2006) and Nomoto et al. (1984), and a z ∼ 1.4 galaxy by Lonoce et al. (2015, [Z/H]>0.5).
Type Ia delay time distribution of the form t−1 between Our results are more similar to the solar metallicities
0.1 and 13 Gyr (Maoz et al. 2012). Comparison with found by Estrada-Carpenter et al. (2018) for stacks of
this model indicates that the z ∼ 1.4 galaxies formed 1.0 < z < 1.8 galaxies (logM/M⊙ ∼ 10.8). We do cau-
 Abundance Patterns of z ∼ 1.4 Galaxies 7

tion though that these studies use varying datasets and Fontanot, F., De Lucia, G., Hirschmann, M., et al. 2017,
techniques to measure stellar metallicities, and thus dif- MNRAS, 464, 3812
ferent assumptions and other systematics complicate the Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J.
2013, PASP, 125, 306
comparison. Gallazzi, A., Bell, E. F., Zibetti, S., Brinchmann, J., & Kelson,
Larger samples are needed to obtain a full census of D. D. 2014, ApJ, 788, 72
the relation between the chemical composition, stellar Gallazzi, A., Charlot, S., Brinchmann, J., White, S. D. M., &
mass, and structures of distant quiescent galaxies. This Tremonti, C. A. 2005, MNRAS, 362, 41
work is part of a survey to obtain deep rest-frame opti- Greene, J. E., Janish, R., Ma, C.-P., et al. 2015, ApJ, 807, 11
Gu, M., Conroy, C., & Brammer, G. 2018, ApJ, 862, L18
cal spectra and elemental abundance measurements for Khochfar, S., & Silk, J. 2006, ApJ, 648, L21
a sample of 20 distant quiescent galaxies; 10 galaxies at Kobayashi, C., Umeda, H., Nomoto, K., Tominaga, N., &
z ∼ 1.4 and 10 galaxies at z ∼ 2.1. The current study Ohkubo, T. 2006, ApJ, 653, 1145
demonstrates that such measurements are only possible Kriek, M., & Conroy, C. 2013, ApJ, 775, L16
for individual quiescent galaxies that are dominated by Kriek, M., van Dokkum, P. G., Labbé, I., et al. 2009, ApJ, 700,
221
an older stellar population, which have more pronounced Kriek, M., Shapley, A. E., Reddy, N. A., et al. 2015, ApJS, 218,
metal lines. For quiescent galaxies that are dominated by 15
a younger stellar population we have to rely on stacking Kriek, M., Conroy, C., van Dokkum, P. G., et al. 2016, Nature,
techniques to obtain more robust measurements. In the 540, 248
near future, we will use our full sample to measure the Kroupa, P. 2001, MNRAS, 322, 231
Lonoce, I., Longhetti, M., Maraston, C., et al. 2015, MNRAS,
relations between the chemical composition, age, stellar 454, 3912
mass/velocity dispersion, and galaxy sizes/structures at Maiolino, R., & Mannucci, F. 2019, A&A Rev., 27, 3
z ∼ 1.4 and z ∼ 2.1, and compare them to the lower Mannucci, F., Cresci, G., Maiolino, R., et al. 2009, MNRAS, 398,
redshift relations. In the more distant future, NIRSpec 1915
on JWST will enable the first resolved stellar abundance Maoz, D., Mannucci, F., & Brandt, T. D. 2012, MNRAS, 426,
3282
studies at these redshifts. These two studies together, in McCracken, H. J., Milvang-Jensen, B., Dunlop, J., et al. 2012,
combination with the resolved stellar abundance stud- A&A, 544, A156
ies at low redshift (e.g., Greene et al. 2015), will open a Momcheva, I. G., van Dokkum, P. G., van der Wel, A., et al.
new window into the chemical enrichment, star forma- 2017, PASP, 129, 015004
tion, and assembly histories of massive quiescent galax- Morishita, T., Abramson, L. E., Treu, T., et al. 2018, ApJ, 856,
L4
ies. Mowla, L., van Dokkum, P., Brammer, G., et al. 2018, ArXiv
e-prints, arXiv:1808.04379
Muzzin, A., Marchesini, D., Stefanon, M., et al. 2013a, ApJS,
We thank the referee for a very thoughtful and con- 206, 8
structive report, Ryan Trainor for his help with the LRIS —. 2013b, ApJ, 777, 18
observations, and Jenny Greene and Marijn Franx for Naab, T., Johansson, P. H., & Ostriker, J. P. 2009, ApJ, 699,
valuable discussions. We acknowledge support from NSF L178
Nomoto, K., Thielemann, F.-K., & Yokoi, K. 1984, ApJ, 286, 644
AAG grants AST-1908748 and 1909942. The authors Onodera, M., Carollo, C. M., Renzini, A., et al. 2015, ApJ, 808,
wish to recognize and acknowledge the very significant 161
cultural role and reverence that the summit of Mauna Peng, C. Y., Ho, L. C., Impey, C. D., & Rix, H.-W. 2002, AJ,
Kea has always had within the indigenous Hawaiian com- 124, 266
munity. We are most fortunate to have the opportunity Salpeter, E. E. 1955, ApJ, 121, 161
Scoville, N., Aussel, H., Brusa, M., et al. 2007, ApJS, 172, 1
to conduct observations from this mountain. Shapley, A. E., Sanders, R. L., Reddy, N. A., et al. 2017, ApJ,
846, L30
REFERENCES Thomas, D., Maraston, C., & Bender, R. 2003, MNRAS, 343, 279
Toft, S., Gallazzi, A., Zirm, A., et al. 2012, ApJ, 754, 3
Almaini, O., Wild, V., Maltby, D. T., et al. 2017, MNRAS, 472, Tomczak, A. R., Quadri, R. F., Tran, K.-V. H., et al. 2014, ApJ,
1401 783, 85
Bezanson, R., van Dokkum, P. G., Tal, T., et al. 2009, ApJ, 697, Tremonti, C. A., Heckman, T. M., Kauffmann, G., et al. 2004,
1290 ApJ, 613, 898
Carollo, C. M., Bschorr, T. J., Renzini, A., et al. 2013, ApJ, 773, van Dokkum, P. G. 2001, PASP, 113, 1420
112 van Dokkum, P. G., Franx, M., Kriek, M., et al. 2008, ApJ, 677,
Chabrier, G. 2003, PASP, 115, 763 L5
Choi, J., Conroy, C., Moustakas, J., et al. 2014, ApJ, 792, 95 Whitaker, K. E., Kriek, M., van Dokkum, P. G., et al. 2012, ApJ,
Conroy, C., Graves, G. J., & van Dokkum, P. G. 2014, ApJ, 780, 745, 179
33 Wuyts, S., Labbé, I., Franx, M., et al. 2007, ApJ, 655, 51
Conroy, C., Gunn, J. E., & White, M. 2009, ApJ, 699, 486 Yano, M., Kriek, M., van der Wel, A., & Whitaker, K. E. 2016,
Conroy, C., & van Dokkum, P. G. 2012, ApJ, 760, 71 ApJ, 817, L21
Erb, D. K., Shapley, A. E., Pettini, M., et al. 2006, ApJ, 644, 813
Estrada-Carpenter, V., Papovich, C., Momcheva, I., et al. 2018,
ArXiv e-prints, arXiv:1810.02824