Showing 1–37 of 37 results for author: Zick, T

AI Procurement Checklists: Revisiting Implementation in the Age of AI Governance

Authors: Tom Zick, Mason Kortz, David Eaves, Finale Doshi-Velez

Abstract: Public sector use of AI has been quietly on the rise for the past decade, but only recently have efforts to regulate it entered the cultural zeitgeist. While simple to articulate, promoting ethical and effective roll outs of AI systems in government is a notoriously elusive task. On the one hand there are hard-to-address pitfalls associated with AI-based tools, including concerns about bias toward… ▽ More Public sector use of AI has been quietly on the rise for the past decade, but only recently have efforts to regulate it entered the cultural zeitgeist. While simple to articulate, promoting ethical and effective roll outs of AI systems in government is a notoriously elusive task. On the one hand there are hard-to-address pitfalls associated with AI-based tools, including concerns about bias towards marginalized communities, safety, and gameability. On the other, there is pressure not to make it too difficult to adopt AI, especially in the public sector which typically has fewer resources than the private sector$\unicode{x2014}$conserving scarce government resources is often the draw of using AI-based tools in the first place. These tensions create a real risk that procedures built to ensure marginalized groups are not hurt by government use of AI will, in practice, be performative and ineffective. To inform the latest wave of regulatory efforts in the United States, we look to jurisdictions with mature regulations around government AI use. We report on lessons learned by officials in Brazil, Singapore and Canada, who have collectively implemented risk categories, disclosure requirements and assessments into the way they procure AI tools. In particular, we investigate two implemented checklists: the Canadian Directive on Automated Decision-Making (CDADM) and the World Economic Forum's AI Procurement in a Box (WEF). We detail three key pitfalls around expertise, risk frameworks and transparency, that can decrease the efficacy of regulations aimed at government AI use and suggest avenues for improvement. △ Less

Submitted 22 April, 2024; originally announced April 2024.

RewardBench: Evaluating Reward Models for Language Modeling

Authors: Nathan Lambert, Valentina Pyatkin, Jacob Morrison, LJ Miranda, Bill Yuchen Lin, Khyathi Chandu, Nouha Dziri, Sachin Kumar, Tom Zick, Ye** Choi, Noah A. Smith, Hannaneh Hajishirzi

Abstract: Reward models (RMs) are at the crux of successfully using RLHF to align pretrained models to human preferences, yet there has been relatively little study that focuses on evaluation of those models. Evaluating reward models presents an opportunity to understand the opaque technologies used for alignment of language models and which values are embedded in them. Resources for reward model training a… ▽ More Reward models (RMs) are at the crux of successfully using RLHF to align pretrained models to human preferences, yet there has been relatively little study that focuses on evaluation of those models. Evaluating reward models presents an opportunity to understand the opaque technologies used for alignment of language models and which values are embedded in them. Resources for reward model training and understanding are sparse in the nascent open-source community around them. To enhance scientific understanding of reward models, we present RewardBench, a benchmark dataset and code-base for evaluation. The RewardBench dataset is a collection of prompt-chosen-rejected trios spanning chat, reasoning, and safety, to benchmark how reward models perform on challenging, structured and out-of-distribution queries. We create specific comparison datasets for RMs that have subtle, but verifiable reasons (e.g. bugs, incorrect facts) why one answer should be preferred to another. On the RewardBench leaderboard, we evaluate reward models trained with a variety of methods, such as the direct MLE training of classifiers and the implicit reward modeling of Direct Preference Optimization (DPO). We present many findings on propensity for refusals, reasoning limitations, and instruction following shortcomings of various reward models towards a better understanding of the RLHF process. △ Less

Submitted 8 June, 2024; v1 submitted 20 March, 2024; originally announced March 2024.

Comments: 44 pages, 19 figures, 12 tables

The History and Risks of Reinforcement Learning and Human Feedback

Authors: Nathan Lambert, Thomas Krendl Gilbert, Tom Zick

Abstract: Reinforcement learning from human feedback (RLHF) has emerged as a powerful technique to make large language models (LLMs) easier to use and more effective. A core piece of the RLHF process is the training and utilization of a model of human preferences that acts as a reward function for optimization. This approach, which operates at the intersection of many stakeholders and academic disciplines,… ▽ More Reinforcement learning from human feedback (RLHF) has emerged as a powerful technique to make large language models (LLMs) easier to use and more effective. A core piece of the RLHF process is the training and utilization of a model of human preferences that acts as a reward function for optimization. This approach, which operates at the intersection of many stakeholders and academic disciplines, remains poorly understood. RLHF reward models are often cited as being central to achieving performance, yet very few descriptors of capabilities, evaluations, training methods, or open-source models exist. Given this lack of information, further study and transparency is needed for learned RLHF reward models. In this paper, we illustrate the complex history of optimizing preferences, and articulate lines of inquiry to understand the sociotechnical context of reward models. In particular, we highlight the ontological differences between costs, rewards, and preferences at stake in RLHF's foundations, related methodological tensions, and possible research directions to improve general understanding of how reward models function. △ Less

Submitted 28 November, 2023; v1 submitted 20 October, 2023; originally announced October 2023.

Comments: 14 pages, 3 figures

The MOSDEF Survey: Probing Resolved Stellar Populations at $z\sim2$ Using a New Bayesian-defined Morphology Metric Called Patchiness

Authors: Tara Fetherolf, Naveen A. Reddy, Alice E. Shapley, Mariska Kriek, Brian Siana, Alison L. Coil, Bahram Mobasher, William R. Freeman, Sedona H. Price, Ryan L. Sanders, Irene Shivaei, Mojegan Azadi, Laura de Groot, Gene C. K. Leung, Tom O. Zick

Abstract: We define a new morphology metric called "patchiness" ($P$) that is sensitive to deviations from the average of a resolved distribution, does not require the galaxy center to be defined, and can be used on the spatially-resolved distribution of any galaxy property. While the patchiness metric has a broad range of applications, we demonstrate its utility by investigating the distribution of dust in… ▽ More We define a new morphology metric called "patchiness" ($P$) that is sensitive to deviations from the average of a resolved distribution, does not require the galaxy center to be defined, and can be used on the spatially-resolved distribution of any galaxy property. While the patchiness metric has a broad range of applications, we demonstrate its utility by investigating the distribution of dust in the interstellar medium of 310 star-forming galaxies at spectroscopic redshifts $1.36<z<1.66$ observed by the MOSFIRE Deep Evolution Field (MOSDEF) survey. The stellar continuum reddening distribution, derived from high-resolution multi-waveband CANDELS/3D-HST imaging, is quantified using the patchiness, Gini, and $M_{20}$ coefficients. We find that the reddening maps of high-mass galaxies, which are dustier and more metal-rich on average, tend to exhibit patchier distributions (high $P$) with the reddest components concentrated within a single region (low $M_{20}$). Our results support a picture where dust is uniformly distributed in low-mass galaxies ($\lesssim$10$^{10}$ $M_\odot$), implying efficient mixing of dust throughout the interstellar medium. On the other hand, the dust distribution is patchier in high-mass galaxies ($\gtrsim$10$^{10}$ $M_\odot$). Dust is concentrated near regions of active star formation and dust mixing timescales are expected to be longer in high-mass galaxies, such that the outskirt regions of these physically larger galaxies remain relatively unenriched. This study presents direct evidence for patchy dust distributions on scales of a few kpc in high-redshift galaxies, which previously has only been suggested as a possible explanation for the observed differences between nebular and stellar continuum reddening, SFR indicators, and dust attenuation curves. △ Less

Submitted 16 November, 2022; v1 submitted 9 August, 2022; originally announced August 2022.

Comments: 25 pages, 19 figures, accepted to MNRAS

The MOSDEF Survey: Towards a Complete Census of the z ~ 2.3 Star-forming Galaxy Population

Authors: Jordan N. Runco, Alice E. Shapley, Ryan L. Sanders, Mariska Kriek, Naveen A. Reddy, Alison L. Coil, Bahram Mobasher, Brian Siana, Michael W. Top**, William R. Freeman, Irene Shivaei, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We analyze the completeness of the MOSDEF survey, in which z ~ 2 galaxies were selected for rest-optical spectroscopy from well-studied HST extragalactic legacy fields down to a fixed rest-optical magnitude limit (H_AB = 24.5). The subset of z ~ 2 MOSDEF galaxies with high signal-to-noise (S/N) emission-line detections analyzed in previous work represents a small minority (<10%) of possible z ~ 2… ▽ More We analyze the completeness of the MOSDEF survey, in which z ~ 2 galaxies were selected for rest-optical spectroscopy from well-studied HST extragalactic legacy fields down to a fixed rest-optical magnitude limit (H_AB = 24.5). The subset of z ~ 2 MOSDEF galaxies with high signal-to-noise (S/N) emission-line detections analyzed in previous work represents a small minority (<10%) of possible z ~ 2 MOSDEF targets. It is therefore crucial to understand how representative this high S/N subsample is, while also more fully exploiting the MOSDEF spectroscopic sample. Using spectral-energy-distribution (SED) models and rest-optical spectral stacking, we compare the MOSDEF z ~ 2 high S/N subsample with the full MOSDEF sample of z ~ 2 star-forming galaxies with redshifts, the latter representing an increase in sample size of more than a factor of three. We find that both samples have similar emission-line properties, in particular in terms of the magnitude of the offset from the local star-forming sequence on the [N II] BPT diagram. There are small differences in median host galaxy properties, including the stellar mass (M_*), star-formation rate (SFR) and specific SFR (sSFR), and UVJ colors; however, these offsets are minor considering the wide spread of the distributions. Using SED modeling, we also demonstrate that the sample of z ~ 2 star-forming galaxies observed by the MOSDEF survey is representative of the parent catalog of available such targets. We conclude that previous MOSDEF results on the evolution of star-forming galaxy emission-line properties were unbiased relative to the parent z ~ 2 galaxy population. △ Less

Submitted 24 August, 2022; v1 submitted 29 June, 2022; originally announced June 2022.

Comments: 18 pages, 10 figures, 3 tables

The MOSDEF Survey: A New View of a Remarkable z=1.89 Merger

Authors: Jordan N. Runco, Alice E. Shapley, Mariska Kriek, Michele Cappellari, Michael W. Top**, Ryan L. Sanders, Vasily I. Kokorev, Sedona H. Price, Naveen A. Reddy, Alison L. Coil, Bahram Mobasher, Brian Siana, Tom Zick, Georgios E. Magdis, Gabriel Brammer, James Aird

Abstract: We present a detailed study of a galaxy merger taking place at $z=1.89$ in the GOODS-S field. Here we analyze Keck/MOSFIRE spectroscopic observations from the MOSFIRE Deep Evolution Field (MOSDEF) survey along with multi-wavelength photometry assembled by the 3D-HST survey. The combined dataset is modeled to infer the past star-formation histories (SFHs) of both merging galaxies. They are found to… ▽ More We present a detailed study of a galaxy merger taking place at $z=1.89$ in the GOODS-S field. Here we analyze Keck/MOSFIRE spectroscopic observations from the MOSFIRE Deep Evolution Field (MOSDEF) survey along with multi-wavelength photometry assembled by the 3D-HST survey. The combined dataset is modeled to infer the past star-formation histories (SFHs) of both merging galaxies. They are found to be massive, with log$_{10}(M_{\ast}/M_{\odot}) > 11$, with a close mass ratio satisfying the typical major-merger definition. Additionally, in the context of delayed-$τ$ models, GOODS-S 43114 and GOODS-S 43683 have similar SFHs and low star-formation rates (log$_{10}$(SFR(SED)/$M_{\odot}/\rm{yr}^{-1}$) $<$ 1.0) compared to their past averages. The best-fit model SEDs show elevated H$δ_{\rm{A}}$ values for both galaxies, indicating that their stellar spectra are dominated by A-type stars, and that star formation peaked $\sim0.5-1$ Gyr ago and has recently declined. Additionally, based on SED fitting both merging galaxies turned on and shut off star formation within a few hundred Myr of each other, suggesting that their bursts of star formation may be linked. Combining the SFHs and H$δ_{\rm{A}}$ results with recent galaxy merger simulations, we infer that these galaxies have recently completed their first pericentric passage and are moving apart. Finally, the relatively low second velocity moment of GOODS-S 43114 given its stellar mass, suggests a disk-like structure. However, including the geometry of the galaxy in the modeling does not completely resolve the discrepancy between the dynamical and stellar masses. Future work is needed to resolve this inconsistency in mass. △ Less

Submitted 6 October, 2022; v1 submitted 10 June, 2022; originally announced June 2022.

Comments: 12 pages, 7 figures, 1 table, Accepted for publication in MNRAS

Reward Reports for Reinforcement Learning

Authors: Thomas Krendl Gilbert, Nathan Lambert, Sarah Dean, Tom Zick, Aaron Snoswell

Abstract: Building systems that are good for society in the face of complex societal effects requires a dynamic approach. Recent approaches to machine learning (ML) documentation have demonstrated the promise of discursive frameworks for deliberation about these complexities. However, these developments have been grounded in a static ML paradigm, leaving the role of feedback and post-deployment performance… ▽ More Building systems that are good for society in the face of complex societal effects requires a dynamic approach. Recent approaches to machine learning (ML) documentation have demonstrated the promise of discursive frameworks for deliberation about these complexities. However, these developments have been grounded in a static ML paradigm, leaving the role of feedback and post-deployment performance unexamined. Meanwhile, recent work in reinforcement learning has shown that the effects of feedback and optimization objectives on system behavior can be wide-ranging and unpredictable. In this paper we sketch a framework for documenting deployed and iteratively updated learning systems, which we call Reward Reports. Taking inspiration from various contributions to the technical literature on reinforcement learning, we outline Reward Reports as living documents that track updates to design choices and assumptions behind what a particular automated system is optimizing for. They are intended to track dynamic phenomena arising from system deployment, rather than merely static properties of models or data. After presenting the elements of a Reward Report, we discuss a concrete example: Meta's BlenderBot 3 chatbot. Several others for game-playing (DeepMind's MuZero), content recommendation (MovieLens), and traffic control (Project Flow) are included in the appendix. △ Less

Submitted 19 March, 2023; v1 submitted 22 April, 2022; originally announced April 2022.

Choices, Risks, and Reward Reports: Charting Public Policy for Reinforcement Learning Systems

Authors: Thomas Krendl Gilbert, Sarah Dean, Tom Zick, Nathan Lambert

Abstract: In the long term, reinforcement learning (RL) is considered by many AI theorists to be the most promising path to artificial general intelligence. This places RL practitioners in a position to design systems that have never existed before and lack prior documentation in law and policy. Public agencies could intervene on complex dynamics that were previously too opaque to deliberate about, and long… ▽ More In the long term, reinforcement learning (RL) is considered by many AI theorists to be the most promising path to artificial general intelligence. This places RL practitioners in a position to design systems that have never existed before and lack prior documentation in law and policy. Public agencies could intervene on complex dynamics that were previously too opaque to deliberate about, and long-held policy ambitions would finally be made tractable. In this whitepaper we illustrate this potential and how it might be technically enacted in the domains of energy infrastructure, social media recommender systems, and transportation. Alongside these unprecedented interventions come new forms of risk that exacerbate the harms already generated by standard machine learning tools. We correspondingly present a new typology of risks arising from RL design choices, falling under four categories: sco** the horizon, defining rewards, pruning information, and training multiple agents. Rather than allowing RL systems to unilaterally reshape human domains, policymakers need new mechanisms for the rule of reason, foreseeability, and interoperability that match the risks these systems pose. We argue that criteria for these choices may be drawn from emerging subfields within antitrust, tort, and administrative law. It will then be possible for courts, federal and state agencies, and non-governmental organizations to play more active roles in RL specification and evaluation. Building on the "model cards" and "datasheets" frameworks proposed by Mitchell et al. and Gebru et al., we argue the need for Reward Reports for AI systems. Reward Reports are living documents for proposed RL deployments that demarcate design choices. △ Less

Submitted 11 February, 2022; originally announced February 2022.

Comments: 60 pages

Journal ref: Center for Long Term Cybersecurity Whitepaper Series Feb. 2022; see release https://cltc.berkeley.edu/2022/02/08/reward-reports/

Reconciling the Results of the z~2 MOSDEF and KBSS-MOSFIRE Surveys

Authors: Jordan N. Runco, Naveen A. Reddy, Alice E. Shapley, Charles C. Steidel, Ryan L. Sanders, Allison L. Strom, Alison L. Coil, Mariska Kriek, Bahram Mobasher, Max Pettini, Gwen C. Rudie, Brian Siana, Michael W. Top**, Ryan F. Trainor, William R. Freeman, Irene Shivaei, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: The combination of the MOSDEF and KBSS-MOSFIRE surveys represents the largest joint investment of Keck/MOSFIRE time to date, with ~3000 galaxies at 1.4<=z<=3.8, roughly half of which are at z~2. MOSDEF is photometric- and spectroscopic-redshift selected with a rest-optical magnitude limit, while KBSS-MOSFIRE is primarily selected based on rest-UV colors and a rest-UV magnitude limit. Analyzing bot… ▽ More The combination of the MOSDEF and KBSS-MOSFIRE surveys represents the largest joint investment of Keck/MOSFIRE time to date, with ~3000 galaxies at 1.4<=z<=3.8, roughly half of which are at z~2. MOSDEF is photometric- and spectroscopic-redshift selected with a rest-optical magnitude limit, while KBSS-MOSFIRE is primarily selected based on rest-UV colors and a rest-UV magnitude limit. Analyzing both surveys in a uniform manner with consistent spectral-energy-distribution (SED) models, we find that the MOSDEF z~2 targeted sample has a higher median M_* and redder rest U-V color than the KBSS-MOSFIRE z~2 targeted sample, and a smaller median SED-based SFR and sSFR (SFR(SED) and sSFR(SED)). Specifically, MOSDEF targeted a larger population of red galaxies with U-V and V-J >=1.25, while KBSS-MOSFIRE contains more young galaxies with intense star formation. Despite these differences in the z~2 targeted samples, the subsets of the surveys with multiple emission lines detected and analyzed in previously published work are much more similar. All median host-galaxy properties with the exception of stellar population age -- i.e., M_*, SFR(SED), sSFR(SED), A_V, and UVJ colors -- agree within the uncertainties. Additionally, when uniform emission-line fitting and stellar Balmer absorption correction techniques are applied, there is no significant offset between the two samples in the [OIII]$λ$5008/H$β$ vs. [NII]$λ$6585/H$α$ diagnostic diagram, in contrast to previously-reported discrepancies. We can now combine the MOSDEF and KBSS-MOSFIRE surveys to form the largest z~2 sample with moderate-resolution rest-optical spectra and construct the fundamental scaling relations of star-forming galaxies during this important epoch. △ Less

Submitted 9 June, 2022; v1 submitted 17 December, 2021; originally announced December 2021.

Comments: 23 pages, 14 figures, published in MNRAS

Journal ref: MNRAS, 513, 3871 (2022)

The MOSDEF Survey: The Dependence of H$α$-to-UV SFR Ratios on SFR and Size at $z\sim2$

Authors: Tara Fetherolf, Naveen A. Reddy, Alice E. Shapley, Mariska Kriek, Brian Siana, Alison L. Coil, Bahram Mobasher, William R. Freeman, Ryan L. Sanders, Sedona H. Price, Irene Shivaei, Mojegan Azadi, Laura de Groot, Gene C. K. Leung, Tom O. Zick

Abstract: We perform an aperture-matched analysis of dust-corrected H$α$ and UV SFRs using 303 star-forming galaxies with spectroscopic redshifts $1.36<z_\text{spec}<2.66$ from the MOSFIRE Deep Evolution Field (MOSDEF) survey. By combining H$α$ and H$β$ emission line measurements with multi-waveband resolved CANDELS/3D-HST imaging, we directly compare dust-corrected H$α$ and UV SFRs, inferred assuming a fix… ▽ More We perform an aperture-matched analysis of dust-corrected H$α$ and UV SFRs using 303 star-forming galaxies with spectroscopic redshifts $1.36<z_\text{spec}<2.66$ from the MOSFIRE Deep Evolution Field (MOSDEF) survey. By combining H$α$ and H$β$ emission line measurements with multi-waveband resolved CANDELS/3D-HST imaging, we directly compare dust-corrected H$α$ and UV SFRs, inferred assuming a fixed attenuation curve shape and constant SFHs, within the spectroscopic aperture. Previous studies have found that H$α$ and UV SFRs inferred with these assumptions generally agree for typical star-forming galaxies, but become increasingly discrepant for galaxies with higher SFRs ($\gtrsim$100 M$_\odot$ yr$^{-1}$), with H$α$-to-UV SFR ratios being larger for these galaxies. Our analysis shows that this trend persists even after carefully accounting for the apertures over which H$α$ and UV-based SFRs (and the nebular and stellar continuum reddening) are derived. Furthermore, our results imply that H$α$ SFRs may be higher in the centers of large galaxies (i.e., where there is coverage by the spectroscopic aperture) compared to their outskirts, which could be indicative of inside-out galaxy growth. Overall, we suggest that the persistent difference between nebular and stellar continuum reddening and high H$α$-to-UV SFR ratios at the centers of large galaxies may be indicative of a patchier distribution of dust in galaxies with high SFRs. △ Less

Submitted 8 September, 2021; originally announced September 2021.

Comments: 15 pages, 7 figures, accepted to MNRAS

Axes for Sociotechnical Inquiry in AI Research

Authors: Sarah Dean, Thomas Krendl Gilbert, Nathan Lambert, Tom Zick

Abstract: The development of artificial intelligence (AI) technologies has far exceeded the investigation of their relationship with society. Sociotechnical inquiry is needed to mitigate the harms of new technologies whose potential impacts remain poorly understood. To date, subfields of AI research develop primarily individual views on their relationship with sociotechnics, while tools for external investi… ▽ More The development of artificial intelligence (AI) technologies has far exceeded the investigation of their relationship with society. Sociotechnical inquiry is needed to mitigate the harms of new technologies whose potential impacts remain poorly understood. To date, subfields of AI research develop primarily individual views on their relationship with sociotechnics, while tools for external investigation, comparison, and cross-pollination are lacking. In this paper, we propose four directions for inquiry into new and evolving areas of technological development: value--what progress and direction does a field promote, optimization--how the defined system within a problem formulation relates to broader dynamics, consensus--how agreement is achieved and who is included in building it, and failure--what methods are pursued when the problem specification is found wanting. The paper provides a lexicon for sociotechnical inquiry and illustrates it through the example of consumer drone technology. △ Less

Submitted 26 April, 2021; originally announced May 2021.

Comments: 9 pages, 1 figure

The MOSDEF Survey: The Mass-Metallicity relationship and the existence of the FMR at z~1.5

Authors: Michael W. Top**, Alice E. Shapley, Ryan L. Sanders, Mariska Kriek, Naveen A. Reddy, Alison L. Coil, Bahram Mobasher, Brian Siana, William R. Freeman, Irene Shivaei, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro, Jordan N. Runco

Abstract: We analyze the rest-optical emission-line ratios of z~1.5 galaxies drawn from the MOSFIRE Deep Evolution Field (MOSDEF) survey. Using composite spectra we investigate the mass-metallicity relation (MZR) at z~1.5 and measure its evolution to z=0. When using gas-phase metallicities based on the N2 line ratio, we find that the MZR evolution from z~1.5 to z=0 depends on stellar mass, evolving by… ▽ More We analyze the rest-optical emission-line ratios of z~1.5 galaxies drawn from the MOSFIRE Deep Evolution Field (MOSDEF) survey. Using composite spectra we investigate the mass-metallicity relation (MZR) at z~1.5 and measure its evolution to z=0. When using gas-phase metallicities based on the N2 line ratio, we find that the MZR evolution from z~1.5 to z=0 depends on stellar mass, evolving by $Δ\rm log(\rm O/H)\sim0.25$ dex at $M_*<10^{9.75}M_{\odot}$ down to $Δ\rm log(\rm O/H)\sim0.05$ at $M_*>10^{10.5}M_{\odot}$. In contrast, the O3N2-based MZR shows a constant offset of $Δ\rm log(\rm O/H)\sim0.30$ across all masses, consistent with previous MOSDEF results based on independent metallicity indicators, and suggesting that O3N2 provides a more robust metallicity calibration for our z~1.5 sample. We investigated the secondary dependence of the MZR on SFR by measuring correlated scatter about the mean $M_*$-specific SFR and $M_*-\log(\rm O3N2)$ relations. We find an anti-correlation between $\log(\rm O/H)$ and sSFR offsets, indicating the presence of a $M_*$-SFR-Z relation, though with limited significance. Additionally, we find that our z~1.5 stacks lie along the z=0 metallicity sequence at fixed $μ=\log(M_*/M_{\odot})-0.6\times\log(\rm SFR / M_{\odot} yr^{-1})$ suggesting that the z~1.5 stacks can be described by the z=0 fundamental metallicity relation (FMR). However, using different calibrations can shift the calculated metallicities off of the local FMR, indicating that appropriate calibrations are essential for understanding metallicity evolution with redshift. Finally, understanding how [NII]/H$α$ scales with galaxy properties is crucial to accurately describe the effects of blended [NII] and H$α$ on redshift and H$α$ flux measurements in future large surveys utilizing low-resolution spectra such as with Euclid and the Roman Space Telescope. △ Less

Submitted 23 June, 2021; v1 submitted 16 March, 2021; originally announced March 2021.

Comments: 14 Pages, 9 figures, Accepted to MNRAS

AI Development for the Public Interest: From Abstraction Traps to Sociotechnical Risks

Authors: McKane Andrus, Sarah Dean, Thomas Krendl Gilbert, Nathan Lambert, Tom Zick

Abstract: Despite interest in communicating ethical problems and social contexts within the undergraduate curriculum to advance Public Interest Technology (PIT) goals, interventions at the graduate level remain largely unexplored. This may be due to the conflicting ways through which distinct Artificial Intelligence (AI) research tracks conceive of their interface with social contexts. In this paper we trac… ▽ More Despite interest in communicating ethical problems and social contexts within the undergraduate curriculum to advance Public Interest Technology (PIT) goals, interventions at the graduate level remain largely unexplored. This may be due to the conflicting ways through which distinct Artificial Intelligence (AI) research tracks conceive of their interface with social contexts. In this paper we track the historical emergence of sociotechnical inquiry in three distinct subfields of AI research: AI Safety, Fair Machine Learning (Fair ML) and Human-in-the-Loop (HIL) Autonomy. We show that for each subfield, perceptions of PIT stem from the particular dangers faced by past integration of technical systems within a normative social order. We further interrogate how these histories dictate the response of each subfield to conceptual traps, as defined in the Science and Technology Studies literature. Finally, through a comparative analysis of these currently siloed fields, we present a roadmap for a unified approach to sociotechnical graduate pedagogy in AI. △ Less

Submitted 4 February, 2021; originally announced February 2021.

Comments: 8 Pages

The MOSDEF Survey: Environmental dependence of the gas-phase metallicity of galaxies at $1.4 \leq z \leq 2.6$

Authors: Nima Chartab, Bahram Mobasher, Alice E. Shapley, Irene Shivaei, Ryan L. Sanders, Alison L. Coil, Mariska Kriek, Naveen A. Reddy, Brian Siana, William R. Freeman, Mojegan Azadi, Guillermo Barro, Tara Fetherolf, Gene Leung, Sedona H. Price, Tom Zick

Abstract: Using the near-IR spectroscopy of the MOSFIRE Deep Evolution Field (MOSDEF) survey, we investigate the role of local environment in the gas-phase metallicity of galaxies. The local environment measurements are derived from accurate and uniformly calculated photometric redshifts with well-calibrated probability distributions. Based on rest-frame optical emission lines, [NII]$\lambda6584$ and H$α$,… ▽ More Using the near-IR spectroscopy of the MOSFIRE Deep Evolution Field (MOSDEF) survey, we investigate the role of local environment in the gas-phase metallicity of galaxies. The local environment measurements are derived from accurate and uniformly calculated photometric redshifts with well-calibrated probability distributions. Based on rest-frame optical emission lines, [NII]$\lambda6584$ and H$α$, we measure gas-phase oxygen abundance of 167 galaxies at $1.37\leq z\leq1.7$ and 303 galaxies at $2.09\leq z\leq2.61$, located in diverse environments. We find that at $z\sim1.5$, the average metallicity of galaxies in overdensities with $M_*\sim10^{9.8}M_\odot, 10^{10.2}M_\odot$ and $10^{10.8}M_\odot$ is higher relative to their field counterparts by $0.094\pm0.051$, $0.068\pm0.028$ and $0.052\pm0.043$ dex, respectively. However, this metallicity enhancement does not exist at higher redshift, $z\sim2.3$, where, compared to the field galaxies, we find $0.056\pm0.043$, $0.056\pm0.028$ and $0.096\pm 0.034$ dex lower metallicity for galaxies in overdense environments with $M_*\sim10^{9.8}M_\odot, 10^{10.2}M_\odot$ and $10^{10.7}M_\odot$, respectively. Our results suggest that, at $1.37\leq z\leq2.61$, the variation of mass-metallicity relation with local environment is small ($<0.1$dex), and reverses at $z\sim2$. Our results support the hypothesis that, at the early stages of cluster formation, owing to efficient gas cooling, galaxies residing in overdensities host a higher fraction of pristine gas with prominent primordial gas accretion, which lowers their gas-phase metallicity compared to their coeval field galaxies. However, as the Universe evolves to lower redshifts ($z\lesssim2$), shock-heated gas in overdensities cannot cool down efficiently, and galaxies become metal-rich rapidly due to the suppression of pristine gas inflow and re-accretion of metal-enriched outflows in overdensities. △ Less

Submitted 5 January, 2021; originally announced January 2021.

Comments: 18 pages, 8 figures, 3 tables, accepted for publication in ApJ

The MOSDEF Survey: The First Direct Measurements of the Nebular Dust Attenuation Curve at High Redshift

Authors: Naveen A. Reddy, Alice E. Shapley, Mariska Kriek, Charles C. Steidel, Irene Shivaei, Ryan L. Sanders, Bahram Mobasher, Alison L. Coil, Brian Siana, William R. Freeman, Mojegan Azadi, Tara Fetherolf, Gene Leung, Sedona H. Price, Tom Zick

Abstract: We use a sample of 532 star-forming galaxies at redshifts $z\sim 1.4-2.6$ with deep rest-frame optical spectra from the MOSFIRE Deep Evolution Field (MOSDEF) survey to place the first constraints on the nebular attenuation curve at high redshift. Based on the first five low-order Balmer emission lines detected in the composite spectra of these galaxies (${\rm Hα}$ through ${\rm Hε}$), we derive a… ▽ More We use a sample of 532 star-forming galaxies at redshifts $z\sim 1.4-2.6$ with deep rest-frame optical spectra from the MOSFIRE Deep Evolution Field (MOSDEF) survey to place the first constraints on the nebular attenuation curve at high redshift. Based on the first five low-order Balmer emission lines detected in the composite spectra of these galaxies (${\rm Hα}$ through ${\rm Hε}$), we derive a nebular attenuation curve that is similar in shape to that of the Galactic extinction curve, suggesting that the dust covering fraction and absorption/scattering properties along the lines-of-sight to massive stars at high redshift are similar to those of the average Milky Way sightline. The curve derived here implies nebular reddening values that are on average systematically larger than those derived for the stellar continuum. In the context of stellar population synthesis models that include the effects of stellar multiplicity, the difference in reddening of the nebular lines and stellar continuum may imply molecular cloud crossing timescales that are a factor of $\gtrsim 3\times$ longer than those inferred for local molecular clouds, star-formation rates that are constant or increasing with time such that newly-formed and dustier OB associations always dominate the ionizing flux, and/or that the dust responsible for reddening the nebular emission may be associated with non-molecular (i.e., ionized and neutral) phases of the ISM. Our analysis points to a variety of investigations of the nebular attenuation curve that will be enabled with the next generation of ground- and space-based facilities. △ Less

Submitted 21 September, 2020; originally announced September 2020.

Comments: 10 pages, 2 tables, 5 figures; submitted 2020 July 04, accepted 2020 September 06 to the Astrophysical Journal

The MOSDEF Survey: The Evolution of the Mass-Metallicity Relation from $z=0$ to $z\sim3.3$

Authors: Ryan L. Sanders, Alice E. Shapley, Tucker Jones, Naveen A. Reddy, Mariska Kriek, Brian Siana, Alison L. Coil, Bahram Mobasher, Irene Shivaei, Romeel Davé, Mojegan Azadi, Sedona H. Price, Gene Leung, William R. Freeman, Tara Fetherolf, Laura de Groot, Tom Zick, Guillermo Barro

Abstract: We investigate the evolution of galaxy gas-phase metallicity (O/H) over the range $z=0-3.3$ using samples of $\sim300$ galaxies at $z\sim2.3$ and $\sim150$ galaxies at $z\sim3.3$ from the MOSDEF survey. This analysis crucially utilizes different metallicity calibrations at $z\sim0$ and $z>1$ to account for evolving ISM conditions. We find significant correlations between O/H and stellar mass (… ▽ More We investigate the evolution of galaxy gas-phase metallicity (O/H) over the range $z=0-3.3$ using samples of $\sim300$ galaxies at $z\sim2.3$ and $\sim150$ galaxies at $z\sim3.3$ from the MOSDEF survey. This analysis crucially utilizes different metallicity calibrations at $z\sim0$ and $z>1$ to account for evolving ISM conditions. We find significant correlations between O/H and stellar mass ($M_*$) at $z\sim2.3$ and $z\sim3.3$. The low-mass power law slope of the mass-metallicity relation is remarkably invariant over $z=0-3.3$, such that $\textrm{O/H}\propto M_*^{0.30}$ at all redshifts in this range. At fixed $M_*$, O/H decreases with increasing redshift as dlog(O/H)/d$z=-0.11\pm0.02$. We find no evidence that the fundamental metallicity relation between $M_*$, O/H, and star-formation rate (SFR) evolves out to $z\sim3.3$, with galaxies at $z\sim2.3-3.3$ having O/H within 0.04~dex of local galaxies matched in $M_*$ and SFR on average. We employ analytic chemical evolution models to place constraints on the mass and metal loading factors of galactic outflows. The efficiency of metal removal increases toward lower $M_*$ at fixed redshift, and toward higher redshift at fixed $M_*$. These models suggest that the slope of the mass-metallicity relation is set by the scaling of the metal loading factor of outflows with $M_*$, not by the change in gas fraction as a function of $M_*$. The evolution toward lower O/H at fixed $M_*$ with increasing redshift is driven by both higher gas fraction (leading to stronger dilution of ISM metals) and higher metal removal efficiency, with models suggesting that both effects contribute approximately equally to the observed evolution. These results suggest that the processes governing the smooth baryonic growth of galaxies via gas flows and star formation hold in the same form over at least the past 12~Gyr. △ Less

Submitted 2 April, 2021; v1 submitted 15 September, 2020; originally announced September 2020.

Comments: 36 pages, 19 figures, accepted to ApJ. Main results in Figures 8, 9, 12, and 13

The MOSDEF Survey: An Improved Voronoi Binning Technique on Spatially Resolved Stellar Populations at z~2

Authors: Tara Fetherolf, Naveen Reddy, Alice Shapley, Mariska Kriek, Brian Siana, Alison Coil, Bahram Mobasher, William Freeman, Ryan Sanders, Sedona Price, Irene Shivaei, Mojegan Azadi, Laura de Groot, Gene Leung, Tom Zick

Abstract: We use a sample of 350 star-forming galaxies at $1.25<z<2.66$ from the MOSFIRE Deep Evolution Field survey to demonstrate an improved Voronoi binning technique that we use to study the properties of resolved stellar populations in $z\sim2$ galaxies. Stellar population and dust maps are constructed from the high-resolution CANDELS/3D-HST multi-band imaging. Rather than constructing the layout of re… ▽ More We use a sample of 350 star-forming galaxies at $1.25<z<2.66$ from the MOSFIRE Deep Evolution Field survey to demonstrate an improved Voronoi binning technique that we use to study the properties of resolved stellar populations in $z\sim2$ galaxies. Stellar population and dust maps are constructed from the high-resolution CANDELS/3D-HST multi-band imaging. Rather than constructing the layout of resolved elements (i.e., Voronoi bins) from the S/N distribution of the $H_{160}$-band alone, we introduce a modified Voronoi binning method that additionally incorporates the S/N distribution of several resolved filters. The SED-derived resolved E(B-V)$_{\text{stars}}$, stellar population ages, SFRs, and stellar masses that are inferred from the Voronoi bins constructed from multiple filters are generally consistent with the properties inferred from the integrated photometry within the uncertainties, with the exception of the inferred E(B-V)$_{\text{stars}}$ from our $z\sim1.5$ sample due to their UV slopes being unconstrained by the resolved photometry. The results from our multi-filter Voronoi binning technique are compared to those derived from a "traditional" single-filter Voronoi binning approach. We find that single-filter binning produces inferred E(B-V)$_{\text{stars}}$ that are systematically redder by 0.02 mag on average, but could differ by up to 0.20 mag, and could be attributed to poorly constrained resolved photometry covering the UV slope. Overall, we advocate that our methodology produces more reliable SED-derived parameters due to the best-fit resolved SEDs being better constrained at all resolved wavelengths--particularly those covering the UV slope. △ Less

Submitted 8 September, 2020; originally announced September 2020.

Comments: 23 pages, 15 figures, accepted for publication in MNRAS

The MOSDEF Survey: Neon as a Probe of ISM Physical Conditions at High Redshift

Authors: Moon-Seong Jeong, Alice E. Shapley, Ryan L. Sanders, Jordan N. Runco, Michael W. Top**, Naveen A. Reddy, Mariska Kriek, Alison L. Coil, Bahram Mobasher, Brian Siana, Irene Shivaei, William R. Freeman, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We present results on the properties of neon emission in $z\sim2$ star-forming galaxies drawn from the MOSFIRE Deep Evolution Field (MOSDEF) survey. Doubly-ionized neon ([NeIII]3869) is detected at $\geq3σ$ in 61 galaxies, representing $\sim$25% of the MOSDEF sample with H$α$, H$β$, and [OIII]$5007$ detections at similar redshifts. We consider the neon emission-line properties of both individual g… ▽ More We present results on the properties of neon emission in $z\sim2$ star-forming galaxies drawn from the MOSFIRE Deep Evolution Field (MOSDEF) survey. Doubly-ionized neon ([NeIII]3869) is detected at $\geq3σ$ in 61 galaxies, representing $\sim$25% of the MOSDEF sample with H$α$, H$β$, and [OIII]$5007$ detections at similar redshifts. We consider the neon emission-line properties of both individual galaxies with [NeIII]3869 detections and composite $z\sim2$ spectra binned by stellar mass. With no requirement of [NeIII]3869 detection, the latter provide a more representative picture of neon emission-line properties in the MOSDEF sample. The [NeIII]3869/[OII]3727 ratio (Ne3O2) is anti-correlated with stellar mass in $z\sim2$ galaxies, as expected based on the mass-metallicity relation. It is also positively correlated with the [OIII]$5007$/[OII]$3727$ ratio (O32), but $z\sim2$ line ratios are offset towards higher Ne3O2 at fixed O32, compared with both local star-forming galaxies and individual H~II regions. Despite the offset towards higher Ne3O2 at fixed O32 at $z\sim2$, biases in inferred Ne3O2-based metallicity are small. Accordingly, Ne3O2 may serve as an important metallicity indicator deep into the reionization epoch. Analyzing additional rest-optical line ratios including [NeIII]$3869$/[OIII]$5007$ (Ne3O3) and [OIII]$5007$/H$β$ (O3H$β$), we conclude that the nebular emission-line ratios of $z\sim2$ star-forming galaxies suggest a harder ionizing spectrum (lower stellar metallicity, i.e., Fe/H) at fixed gas-phase oxygen abundance, compared to systems at $z\sim0$. These new results based on neon lend support to the physical picture painted by oxygen, nitrogen, hydrogen, and sulfur emission, of an ionized ISM in high-redshift star-forming galaxies irradiated by chemically young, $α$-enhanced massive stars. △ Less

Submitted 24 September, 2020; v1 submitted 20 August, 2020; originally announced August 2020.

Comments: 7 pages, 5 figures, accepted to ApJ Letters

The MOSDEF Survey: A Comprehensive Analysis of the Rest-optical Emission-line Properties of $z\sim 2.3$ Star-forming Galaxies

Authors: Jordan N. Runco, Alice E. Shapley, Ryan L. Sanders, Michael W. Top**, Mariska Kriek, Naveen A. Reddy, Alison L. Coil, Bahram Mobasher, Brian Siana, William R. Freeman, Irene Shivaei, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We analyze the rest-optical emission-line spectra of $z\sim2.3$ star-forming galaxies in the complete MOSFIRE Deep Evolution Field (MOSDEF) survey. In investigating the origin of the well-known offset between the sequences of high-redshift and local galaxies in the [O III]5008/H$β$ vs. [N II]6585/H$α$ ("[N II] BPT") diagram, we define two populations of $z\sim2.3$ MOSDEF galaxies. These include th… ▽ More We analyze the rest-optical emission-line spectra of $z\sim2.3$ star-forming galaxies in the complete MOSFIRE Deep Evolution Field (MOSDEF) survey. In investigating the origin of the well-known offset between the sequences of high-redshift and local galaxies in the [O III]5008/H$β$ vs. [N II]6585/H$α$ ("[N II] BPT") diagram, we define two populations of $z\sim2.3$ MOSDEF galaxies. These include the "high" population that is offset towards higher [O III]5008/H$β$ and/or [N II]6585/H$α$ with respect to the local SDSS sequence and the "low" population that overlaps the SDSS sequence. These two groups are also segregated within the [O III]5008/H$β$ vs. [S II]6718,6733/H$α$ and the [O III]4960,5008/[O II]3727,3730 (O$_{32}$) vs. ([O III]4960,5008+[O II]3727,3730)/H$β$ (R$_{23}$) diagram, which suggests qualitatively that star-forming regions in the more offset galaxies are characterized by harder ionizing spectra at fixed nebular oxygen abundance. We also investigate many galaxy properties of the split sample and find that the "high" sample is on average smaller in size and less massive, but has higher specific star-formation rate and star-formation-rate surface density values and is slightly younger compared to the "low" population. From Cloudy+BPASS photoionization models, we estimate that the "high" population has a lower stellar metallicity (i.e., harder ionizing spectrum) but slightly higher nebular metallicity and higher ionization parameter compared to the "low" population. While the "high" population is more $α$-enhanced (i.e., higher $α$/Fe) than the "low" population, both samples are significantly more $α$-enhanced compared to local star-forming galaxies with similar rest-optical line ratios. These differences must be accounted for in all high-redshift star-forming galaxies -- not only those "offset" from local excitation sequences. △ Less

Submitted 19 January, 2021; v1 submitted 11 August, 2020; originally announced August 2020.

Comments: 16 pages, 7 figures

The MOSDEF Survey: Differences in SFR and Metallicity for Morphologically-Selected Mergers at z~2

Authors: Katelyn Horstman, Alice E. Shapley, Ryan L. Sanders, Bahram Mobasher, Naveen A. Reddy, Mariska Kriek, Alison L. Coil, Brian Siana, Irene Shivaei, William R. Freeman, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We study the properties of 55 morphologically-identified merging galaxy systems at z~2. These systems are flagged as mergers based on features such as tidal tails, double nuclei, and asymmetry. Our sample is drawn from the MOSFIRE Deep Evolution Field (MOSDEF) survey, along with a control sample of isolated galaxies at the same redshift. We consider the relationships between stellar mass, star for… ▽ More We study the properties of 55 morphologically-identified merging galaxy systems at z~2. These systems are flagged as mergers based on features such as tidal tails, double nuclei, and asymmetry. Our sample is drawn from the MOSFIRE Deep Evolution Field (MOSDEF) survey, along with a control sample of isolated galaxies at the same redshift. We consider the relationships between stellar mass, star formation rate (SFR), and gas-phase metallicity for both merging and non-merging systems. In the local universe, merging systems are characterized by an elevated SFR and depressed metallicity compared to isolated systems at a given mass. Our results indicate SFR enhancement and metallicity deficit for merging systems relative to non-merging systems for a fixed stellar mass at z~2, though larger samples are required to establish these preliminary results with higher statistical significance. In future work, it will be important to establish if the enhanced SFR and depressed metallicity in high-redshift mergers deviate from the "fundamental metallicity relation," as is observed in mergers in the local universe, and therefore shed light on gas flows during galaxy interactions. △ Less

Submitted 26 November, 2020; v1 submitted 10 August, 2020; originally announced August 2020.

Comments: 9 pages, 5 figures, 5 figures, accepted to MNRAS

The MOSDEF Survey: the Variation of the Dust Attenuation Curve with Metallicity

Authors: Irene Shivaei, Naveen Reddy, George Rieke, Alice Shapley, Mariska Kriek, Andrew Battisti, Bahram Mobasher, Ryan Sanders, Tara Fetherolf, Mojegan Azadi, Alison L. Coil, William R. Freeman, Laura de Groot, Gene Leung, Sedona H. Price, Brian Siana, Tom Zick

Abstract: We derive the UV-optical stellar dust attenuation curve of galaxies at z=1.4-2.6 as a function of gas-phase metallicity. We use a sample of 218 star-forming galaxies, excluding those with very young or heavily obscured star formation, from the MOSFIRE Deep Evolution Field (MOSDEF) survey with H$α$, H$β$, and [NII]$λ6585$ spectroscopic measurements. We constrain the shape of the attenuation curve b… ▽ More We derive the UV-optical stellar dust attenuation curve of galaxies at z=1.4-2.6 as a function of gas-phase metallicity. We use a sample of 218 star-forming galaxies, excluding those with very young or heavily obscured star formation, from the MOSFIRE Deep Evolution Field (MOSDEF) survey with H$α$, H$β$, and [NII]$λ6585$ spectroscopic measurements. We constrain the shape of the attenuation curve by comparing the average flux densities of galaxies sorted into bins of dust obscuration using Balmer decrements, i.e., H$α$-to-H$β$ luminosities. The average attenuation curve for the high-metallicity sample (12+log(O/H)>8.5, corresponding to $M_*\gtrsim10^{10.4}\,M_{\odot}$) has a shallow slope, identical to that of the Calzetti local starburst curve, and a significant UV 2175A extinction bump that is $\sim 0.5\times$ the strength of the Milky Way bump. On the other hand, the average attenuation curve of the low-metallicity sample (12+log(O/H) $\sim 8.2-8.5$) has a steeper slope similar to that of the SMC curve, only consistent with the Calzetti slope at the $3σ$ level. The UV bump is not detected in the low-metallicity curve, indicating the relative lack of the small dust grains causing the bump at low metallicities. Furthermore, we find that on average the nebular reddening (E(B-V)) is a factor of 2 times larger than that of the stellar continuum for galaxies with low metallicities, while the nebular and stellar reddening are similar for galaxies with higher metallicities. The latter is likely due to a high surface density of dusty clouds embedding the star forming regions but also reddening the continuum in the high-metallicity galaxies. △ Less

Submitted 6 July, 2020; v1 submitted 4 May, 2020; originally announced May 2020.

Comments: 20 pages and 9 figures and 1 appendix, accepted for publication in ApJ

Towards Studying Hierarchical Assembly in Real Time: A Milky Way Progenitor Galaxy at z = 2.36 under the Microscope

Authors: Tom O. Zick, Daniel R. Weisz, Bruno Ribeiro, Mariska T. Kriek, Benjamin D. Johnson, Xiangcheng Ma, Rychard Bouwens

Abstract: We use Hubble Space Telescope (HST) imaging and near-infrared spectroscopy from Keck/MOSFIRE to study the sub-structure around the progenitor of a Milky Way-mass galaxy in the Hubble Frontier Fields (HFF). Specifically, we study an $r_e = 40^{+70}_{-30}$pc, $M_{\star} \sim 10^{8.2} M_{\odot}$ rest-frame ultra-violet luminous "clump" at a projected distance of $\sim$100~pc from a… ▽ More We use Hubble Space Telescope (HST) imaging and near-infrared spectroscopy from Keck/MOSFIRE to study the sub-structure around the progenitor of a Milky Way-mass galaxy in the Hubble Frontier Fields (HFF). Specifically, we study an $r_e = 40^{+70}_{-30}$pc, $M_{\star} \sim 10^{8.2} M_{\odot}$ rest-frame ultra-violet luminous "clump" at a projected distance of $\sim$100~pc from a $M_{\star} \sim 10^{9.8}$M$_{\odot}$ galaxy at $z = 2.36$ with a magnification $μ= 5.21$. We measure the star formation history of the clump and galaxy by jointly modeling the broadband spectral energy distribution from HST photometry and H$α$ from MOSFIRE spectroscopy. Given our inferred properties (e.g., mass, metallicity, dust) of the clump and galaxy, we explore scenarios in which the clump formed \emph{in-situ} (e.g., a star forming complex) or \emph{ex-situ} (e.g., a dwarf galaxy being accreted). If it formed \emph{in-situ}, we conclude that the clump is likely a single entity as opposed to a aggregation of smaller star clusters, making it one of the most dense star clusters cataloged. If it formed \emph{ex-situ}, then we are witnessing an accretion event with a 1:40 stellar mass ratio. However, our data alone are not informative enough to distinguish between \emph{in-situ} and \emph{ex-situ} scenarios to a high level of significance. We posit that the addition of high-fidelity metallicity information, such as [OIII]4363Å, which can be detected at modest S/N with only a few hours of JWST/NIRSpec time, may be a powerful discriminant. We suggest that studying larger samples of moderately lensed sub-structures across cosmic time can provide unique insight into the hierarchical formation of galaxies like the Milky Way. △ Less

Submitted 18 February, 2020; originally announced February 2020.

Comments: Accepted to MNRAS

The MOSDEF Survey: [SIII] as a New Probe of Evolving ISM Conditions

Authors: Ryan L. Sanders, Tucker Jones, Alice E. Shapley, Naveen A. Reddy, Mariska Kriek, Alison L. Coil, Brian Siana, Bahram Mobasher, Irene Shivaei, Sedona H. Price, William R. Freeman, Mojegan Azadi, Gene C. K. Leung, Tara Fetherolf, Tom O. Zick, Laura de Groot, Guillermo Barro, Francesca M. Fornasini

Abstract: We present measurements of [SIII]$λλ$9069,9531 for a sample of $z\sim1.5$ star-forming galaxies, the first sample with measurements of these lines at z>0.1. We employ the line ratio S$_{32}$$\equiv$[SIII]$λλ$9069,9531/[SII]$λλ$6716,6731 as a novel probe of evolving ISM conditions. Since this ratio includes the low-ionization line [SII], it is crucial that the effects of diffuse ionized gas (DIG) o… ▽ More We present measurements of [SIII]$λλ$9069,9531 for a sample of $z\sim1.5$ star-forming galaxies, the first sample with measurements of these lines at z>0.1. We employ the line ratio S$_{32}$$\equiv$[SIII]$λλ$9069,9531/[SII]$λλ$6716,6731 as a novel probe of evolving ISM conditions. Since this ratio includes the low-ionization line [SII], it is crucial that the effects of diffuse ionized gas (DIG) on emission-line ratios be accounted for in $z\sim0$ integrated galaxy spectra, or else that comparisons be made to samples of local HII regions in which DIG emission is not present. We find that S$_{32}$ decreases with increasing stellar mass at both $z\sim1.5$ and $z\sim0$, but that the dependence is weak suggesting S$_{32}$ has a very shallow anticorrelation with metallicity, in contrast with O$_{32}$ that displays a strong metallicity dependence. As a result, S$_{32}$ only mildly evolves with redshift at fixed stellar mass. The $z\sim1.5$ sample is systematicallty offset towards lower S$_{32}$ and higher [SII]/H$α$ at fixed [OIII]/H$β$ relative to $z=0$ HII regions. By comparing to photoionization model grids, we find that such trends can be explained by a scenario in which the ionizing spectrum is harder at fixed O/H with increasing redshift, but are inconsistent with an increase in ionization parameter at fixed O/H. This analysis demonstrates the advantages of expanding beyond the strongest rest-optical lines for evolutionary studies, and the particular utility of [SIII] for characterizing evolving ISM conditions and stellar compositions. These measurements provide a basis for estimating [SIII] line strengths for high-redshift galaxies, a line that the James Webb Space Telescope will measure out to z~5.5. △ Less

Submitted 2 December, 2019; v1 submitted 29 October, 2019; originally announced October 2019.

Comments: 6 pages, 4 figures, accepted for publication in ApJ Letters

The MOSDEF Survey: The Metallicity Dependence of X-ray Binary Populations at $z\sim2$

Authors: Francesca M. Fornasini, Mariska Kriek, Ryan L. Sanders, Irene Shivaei, Francesca Civano, Naveen A. Reddy, Alice E. Shapley, Alison L. Coil, Bahram Mobasher, Brian Siana, James Aird, Mojegan Azadi, William R. Freeman, Gene C. K. Leung, Sedona H. Price, Tara Fetherolf, Tom Zick, Guillermo Barro

Abstract: Population synthesis models predict that high-mass X-ray binary (HMXB) populations produced in low metallicity environments should be more X-ray luminous, a trend supported by studies of nearby galaxies. This trend may be responsible for the observed increase of the X-ray luminosity ($L_{\mathrm{X}}$) per star formation rate (SFR) with redshift due to the decrease of metallicity ($Z$) at fixed ste… ▽ More Population synthesis models predict that high-mass X-ray binary (HMXB) populations produced in low metallicity environments should be more X-ray luminous, a trend supported by studies of nearby galaxies. This trend may be responsible for the observed increase of the X-ray luminosity ($L_{\mathrm{X}}$) per star formation rate (SFR) with redshift due to the decrease of metallicity ($Z$) at fixed stellar mass as a function of redshift. To test this hypothesis, we use a sample of 79 $z\sim2$ star-forming galaxies with oxygen abundance measurements from the MOSDEF survey, which obtained rest-frame optical spectra for $\sim1500$ galaxies in the CANDELS fields at $1.37<z<3.80$. Using Chandra data from the AEGIS-X Deep, Deep Field North, and Deep Field South surveys, we stack the X-ray data at the galaxy locations in bins of redshift and $Z$ because the galaxies are too faint to be individually detected. In agreement with previous studies, the average $L_{\mathrm{X}}$/SFR of our $z\sim2$ galaxy sample is enhanced by $\approx0.4-0.8$ dex relative to local HMXB $L_{\mathrm{X}}$-SFR scaling relations. Splitting our sample by $Z$, we find that $L_{\mathrm{X}}$/SFR and $Z$ are anti-correlated with 97% confidence. This observed $Z$ dependence for HMXB-dominated galaxies is consistent both with the local $L_{\mathrm{X}}$-SFR-$Z$ relation and a subset of population synthesis models. Although the statistical significance of the observed trends is weak due to the low X-ray statistics, these results constitute the first direct evidence connecting the redshift evolution of $L_{\mathrm{X}}$/SFR and the $Z$ dependence of HMXBs. △ Less

Submitted 18 September, 2019; originally announced September 2019.

Comments: 22 pages, 8 figures, 2 tables

The MOSDEF Survey: Sulfur Emission-line Ratios Provide New Insights into Evolving ISM Conditions at High Redshift

Authors: Alice E. Shapley, Ryan L. Sanders, Peng Shao, Naveen A. Reddy, Mariska Kriek, Alison L. Coil, Bahram Mobasher, Brian Siana, Irene Shivaei, William R. Freeman, Mojegan Azadi, Sedona H. Price, Gene C. K. Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We present results on the emission-line properties of 1.3<=z<=2.7 galaxies drawn from the complete MOSFIRE Deep Evolution Field (MOSDEF) survey. Specifically, we use observations of the emission-line diagnostic diagram of [OIII]5007/Hb vs. [SII]6717,6731/Ha, i.e., the "[SII] BPT diagram," to gain insight into the physical properties of high-redshift star-forming regions. High-redshift MOSDEF galax… ▽ More We present results on the emission-line properties of 1.3<=z<=2.7 galaxies drawn from the complete MOSFIRE Deep Evolution Field (MOSDEF) survey. Specifically, we use observations of the emission-line diagnostic diagram of [OIII]5007/Hb vs. [SII]6717,6731/Ha, i.e., the "[SII] BPT diagram," to gain insight into the physical properties of high-redshift star-forming regions. High-redshift MOSDEF galaxies are offset towards lower [SII]6717,6731/Ha at fixed [OIII]5007/Hb, relative to local galaxies from the Sloan Digital Sky Survey (SDSS). Furthermore, at fixed [OIII]5007/Hb, local SDSS galaxies follow a trend of decreasing [SII]6717,6731/Ha as the surface density of star formation (Sigma_SFR) increases. We explain this trend in terms of the decreasing fractional contribution from diffuse ionized gas (f_DIG) as Sigma_SFR increases in galaxies, which causes galaxy-integrated line ratios to shift towards the locus of pure HII-region emission. The z~0 relationship between f_DIG and Sigma_SFR implies that high-redshift galaxies have lower f_DIG values than typical local systems, given their significantly higher typical Sigma_SFR. When an appropriate low-redshift benchmark with zero or minimal f_DIG is used, high-redshift MOSDEF galaxies appear offset towards higher [SII]6717,6731/Ha and/or [OIII]5007/Hb. The joint shifts of high-redshift galaxies in the [SII] and [NII] BPT diagrams are best explained in terms of the harder spectra ionizing their star-forming regions at fixed nebular oxygen abundance (expected for chemically-young galaxies), as opposed to large variations in N/O ratios or higher ionization parameters. The evolving mixture of HII regions and DIG is an essential ingredient to our description of the ISM over cosmic time. △ Less

Submitted 14 August, 2019; v1 submitted 16 July, 2019; originally announced July 2019.

Comments: 6 pages, 4 figures, accepted to ApJ Letters

The MOSDEF Survey: Direct-Method Metallicities and ISM Conditions at $z\sim1.5-3.5$

Authors: Ryan L. Sanders, Alice E. Shapley, Naveen A. Reddy, Mariska Kriek, Brian Siana, Alison L. Coil, Bahram Mobasher, Irene Shivaei, William R. Freeman, Mojegan Azadi, Sedona H. Price, Gene Leung, Tara Fetherolf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We present detections of [OIII]$λ$4363 and direct-method metallicities for star-forming galaxies at $z=1.7-3.6$. We combine new measurements from the MOSFIRE Deep Evolution Field (MOSDEF) survey with literature sources to construct a sample of 18 galaxies with direct-method metallicities at $z>1$, spanning $7.5<1$2+log(O/H$)<8.2$ and log(M$_*$/M$_{\odot})=7-10$. We find that strong-line calibratio… ▽ More We present detections of [OIII]$λ$4363 and direct-method metallicities for star-forming galaxies at $z=1.7-3.6$. We combine new measurements from the MOSFIRE Deep Evolution Field (MOSDEF) survey with literature sources to construct a sample of 18 galaxies with direct-method metallicities at $z>1$, spanning $7.5<1$2+log(O/H$)<8.2$ and log(M$_*$/M$_{\odot})=7-10$. We find that strong-line calibrations based on local analogs of high-redshift galaxies reliably reproduce the metallicity of the $z>1$ sample on average. We construct the first mass-metallicity relation at $z>1$ based purely on direct-method O/H, finding a slope that is consistent with strong-line results. Direct-method O/H evolves by $\lesssim0.1$ dex at fixed M$_*$ and SFR from $z\sim0-2.2$. We employ photoionization models to constrain the ionization parameter and ionizing spectrum in the high-redshift sample. Stellar models with super-solar O/Fe and binary evolution of massive stars are required to reproduce the observed strong-line ratios. We find that the $z>1$ sample falls on the $z\sim0$ relation between ionization parameter and O/H, suggesting no evolution of this relation from $z\sim0$ to $z\sim2$. These results suggest that the offset of the strong-line ratios of this sample from local excitation sequences is driven primarily by a harder ionizing spectrum at fixed nebular metallicity compared to what is typical at $z\sim0$, naturally explained by super-solar O/Fe at high redshift caused by rapid formation timescales. Given the extreme nature of our $z>1$ sample, the implications for representative $z\sim2$ galaxy samples at $\sim10^{10}$ M$_{\odot}$ are unclear, but similarities to $z>6$ galaxies suggest that these conclusions can be extended to galaxies in the epoch of reionization. △ Less

Submitted 25 October, 2019; v1 submitted 28 June, 2019; originally announced July 2019.

Comments: 27 pages + 4 page appendix, 18 figures, accepted for publication in MNRAS

The MOSDEF survey: a census of AGN-driven ionized outflows at $z = 1.4-3.8$

Authors: Gene C. K. Leung, Alison L. Coil, James Aird, Mojegan Azadi, Mariska Kriek, Bahram Mobasher, Naveen Reddy, Alice Shapley, Brian Siana, Tara Fetherolf, Francesca M. Fornasini, William R. Freeman, Sedona H. Price, Ryan L. Sanders, Irene Shivaei, Tom Zick

Abstract: Using data from the MOSFIRE Deep Evolution Field (MOSDEF) survey, we present a census of AGN-driven ionized outflows in a sample of 159 AGNs at $1.4 \le z \le 3.8$. The sample spans AGN bolometric luminosities of $10^{44-47} \mathrm{~erg~s}^{-1}$ and includes both quiescent and star-forming galaxies extending across three orders of magnitude in stellar mass. We identify and characterize outflows f… ▽ More Using data from the MOSFIRE Deep Evolution Field (MOSDEF) survey, we present a census of AGN-driven ionized outflows in a sample of 159 AGNs at $1.4 \le z \le 3.8$. The sample spans AGN bolometric luminosities of $10^{44-47} \mathrm{~erg~s}^{-1}$ and includes both quiescent and star-forming galaxies extending across three orders of magnitude in stellar mass. We identify and characterize outflows from the \hbeta, [OIII], \halpha ~and [NII] emission line spectra. We detect outflows in $17\%$ of the AGNs, seven times more often than in a mass-matched sample of inactive galaxies in MOSDEF. The outflows are fast and galaxy-wide, with velocities of $\sim 400-3500 ~\mathrm{km~s}^{-1}$ and spatial extents of $0.3-11.0$ kpc. The incidence of outflows among AGNs is independent of the stellar mass of the host galaxy, with outflows detected in both star-forming and quiescent galaxies. This suggests that outflows exist across different phases in galaxy evolution. We investigate relations between outflow kinematic, spatial, and energetic properties and both AGN and host galaxy properties. Our results show that AGN-driven outflows are widespread in galaxies along the star-forming main sequence. The mass-loading factors of the outflows are typically $0.1-1$ and increase with AGN luminosity, capable of exceeding unity at $L_\mathrm{AGN} \gtrsim 10^{46.3} \mathrm{~erg~s}^{-1}$. In these more luminous sources the ionized outflow alone is likely sufficient to regulate star formation, and when combined with outflowing neutral and molecular gas may be able to quench star formation in their host galaxies. △ Less

Submitted 25 September, 2019; v1 submitted 30 May, 2019; originally announced May 2019.

Comments: 24 pages, 12 figures, accepted for publication in the The Astrophysical Journal

The MOSDEF Survey: Kinematic and Structural Evolution of Star-Forming Galaxies at $1.4\leq z\leq 3.8$

Authors: Sedona H. Price, Mariska Kriek, Guillermo Barro, Alice E. Shapley, Naveen A. Reddy, William R. Freeman, Alison L. Coil, Irene Shivaei, Mojegan Azadi, Laura de Groot, Brian Siana, Bahram Mobasher, Ryan L. Sanders, Gene C. K. Leung, Tara Fetherolf, Tom O. Zick, Hannah Übler, Natascha M. Förster Schreiber

Abstract: We present ionized gas kinematics for 681 galaxies at $z\sim 1.4-3.8$ from the MOSFIRE Deep Evolution Field survey, measured using models which account for random galaxy-slit misalignments together with structural parameters derived from CANDELS Hubble Space Telescope (HST) imaging. Kinematics and sizes are used to derive dynamical masses. Baryonic masses are estimated from stellar masses and infe… ▽ More We present ionized gas kinematics for 681 galaxies at $z\sim 1.4-3.8$ from the MOSFIRE Deep Evolution Field survey, measured using models which account for random galaxy-slit misalignments together with structural parameters derived from CANDELS Hubble Space Telescope (HST) imaging. Kinematics and sizes are used to derive dynamical masses. Baryonic masses are estimated from stellar masses and inferred gas masses from dust-corrected star formation rates (SFRs) and the Kennicutt-Schmidt relation. We measure resolved rotation for 105 galaxies. For the remaining 576 galaxies we use models based on HST imaging structural parameters together with integrated velocity dispersions and baryonic masses to statistically constrain the median ratio of intrinsic ordered to disordered motion, $V/σ_{V,0}$. We find that $V/σ_{V,0}$ increases with increasing stellar mass and decreasing specific SFR (sSFR). These trends may reflect marginal disk stability, where systems with higher gas fractions have thicker disks. For galaxies with detected rotation we assess trends between their kinematics and mass, sSFR, and baryon surface density ($Σ_{\mathrm{bar},e}$). Intrinsic dispersion correlates most with $Σ_{\mathrm{bar},e}$ and velocity correlates most with mass. By comparing dynamical and baryonic masses, we find that galaxies at $z\sim 1.4-3.8$ are baryon dominated within their effective radii ($R_E$), with Mdyn/Mbaryon increasing over time. The inferred baryon fractions within $R_E$, $f_{\mathrm{bar}}$, decrease over time, even at fixed mass, size, or surface density. At fixed redshift, $f_{\mathrm{bar}}$ does not appear to vary with stellar mass but increases with decreasing $R_E$ and increasing $Σ_{\mathrm{bar},e}$. For galaxies at $z\geq2$, the median inferred baryon fractions generally exceed 100%. We discuss possible explanations and future avenues to resolve this tension. △ Less

Submitted 16 March, 2020; v1 submitted 25 February, 2019; originally announced February 2019.

Comments: Accepted to ApJ. Added Figure 9, corrected sample size (main results unchanged). 28 pages, 13 figures

The MOSDEF Survey: Significant Evolution in the Rest-Frame Optical Emission Line Equivalent Widths of Star-Forming Galaxies at z=1.4-3.8

Authors: Naveen A. Reddy, Alice E. Shapley, Ryan L. Sanders, Mariska Kriek, Alison L. Coil, Irene Shivaei, William R. Freeman, Bahram Mobasher, Brian Siana, Mojegan Azadi, Tara Fetherolf, Francesca M. Fornasini, Gene Leung, Sedona H. Price, Tom Zick, Guillermo Barro

Abstract: We use extensive spectroscopy from the MOSFIRE Deep Evolution Field (MOSDEF) survey to investigate the relationships between rest-frame optical emission line equivalent widths ($W$) and a number of galaxy and ISM characteristics for a sample of $1134$ star-forming galaxies at redshifts $1.4\lesssim z\lesssim 3.8$. We examine how the equivalent widths of [OII]$λλ3727, 3730$, H$β$, [OIII]… ▽ More We use extensive spectroscopy from the MOSFIRE Deep Evolution Field (MOSDEF) survey to investigate the relationships between rest-frame optical emission line equivalent widths ($W$) and a number of galaxy and ISM characteristics for a sample of $1134$ star-forming galaxies at redshifts $1.4\lesssim z\lesssim 3.8$. We examine how the equivalent widths of [OII]$λλ3727, 3730$, H$β$, [OIII]$λλ4960, 5008$, [OIII]$+$H$β$, H$α$, and H$α$+[NII]$λλ6550, 6585$, depend on stellar mass, UV slope, age, star-formation rate (SFR) and specific SFR (sSFR), ionization parameter and excitation conditions (O32 and [OIII]/H$β$), gas-phase metallicity, and ionizing photon production efficiency ($ξ_{\rm ion}$). The trend of increasing $W$ with decreasing stellar mass is strongest for [OIII] (and [OIII]+H$β$). More generally, the equivalent widths of all the lines increase with redshift at a fixed stellar mass or fixed gas-phase metallicity, suggesting that high equivalent width galaxies are common at high redshift. This redshift evolution in equivalent widths can be explained by the increase in SFR and decrease in metallicity with redshift at a fixed stellar mass. Consequently, the dependence of $W$ on sSFR is largely invariant with redshift, particularly when examined for galaxies of a given metallicity. Our results show that high equivalent width galaxies, specifically those with high $W({\rm [OIII]})$, have low stellar masses, blue UV slopes, young ages, high sSFRs, ISM line ratios indicative of high ionization parameters, high $ξ_{\rm ion}$, and low metallicities. As these characteristics are often attributed to galaxies with high ionizing escape fractions, galaxies with high $W$ are likely candidates for the population that dominates cosmic reionization. △ Less

Submitted 28 November, 2018; originally announced November 2018.

Comments: 34 pages, 8 tables, 28 figures; submitted 2018 August 23, accepted 2018 October 29 to the Astrophysical Journal

The MOSDEF Survey: Stellar Continuum Spectra and Star Formation Histories of Active, Transitional, and Quiescent Galaxies at 1.4$<$\lowercase{z}$<$2.6

Authors: Tom O. Zick, Mariska Kriek, Alice E. Shapley, Naveen A. Reddy, William R. Freeman, Brian Siana, Alison L. Coil, Mojegan Azadi, Guillermo Barro, Tara Fetherolf, Francesca M. Fornasini, Laura de Groot, Gene Leung, Bahram Mobasher, Sedona H. Price, Ryan L. Sanders, Irene Shivaei

Abstract: Using the MOSDEF rest-frame optical spectroscopic survey, we investigate the star-formation histories (SFHs) of different galaxy types, ranging from actively star forming to quiescent at $1.4\leq~z\leq2.6$. SFHs are constrained utilizing stellar continuum spectroscopy, specifically through a combination of Balmer absorption lines, the 4000$Å$ break, and the equivalent width of the H$α$ emission li… ▽ More Using the MOSDEF rest-frame optical spectroscopic survey, we investigate the star-formation histories (SFHs) of different galaxy types, ranging from actively star forming to quiescent at $1.4\leq~z\leq2.6$. SFHs are constrained utilizing stellar continuum spectroscopy, specifically through a combination of Balmer absorption lines, the 4000$Å$ break, and the equivalent width of the H$α$ emission line. To attain a sufficiently high signal-to-noise ratio (S/N) to conduct these measurements we stack spectra of galaxies with similar spectral types, as determined from their rest-frame $U-V$ and $V-J$ colors. We bin the MOSDEF sample into five spectral types, subdividing the quiescent and star-forming bins to better explore galaxies transitioning between the two. We constrain the average SFHs for each type, finding that quiescent and transitional galaxies in the MOSDEF sample are dominated by an SFH with an average star-formation timescale of $τ\sim0.1-0.2$~Gyr. These findings contrast with measurements from the low-redshift Universe where, on average, galaxies form their stars over a more extended time period ($τ>1$~Gyr). Furthermore, our spectral index measurements correlate with mass surface density for all spectral types. Finally, we compare the average properties of the galaxies in our transitional bins to investigate possible paths to quiescence, and speculate on the viability of a dusty post-starburst phase. △ Less

Submitted 17 October, 2018; v1 submitted 16 October, 2018; originally announced October 2018.

Comments: Accepted to ApJ Letters

The MOSDEF Survey: No Significant Enhancement in Star Formation or Deficit in Metallicity in Merging Galaxy Pairs at 1.5<=z<=3.5

Authors: Timothy J. Wilson, Alice E. Shapley, Ryan L. Sanders, Naveen A. Reddy, William R. Freeman, Mariska Kriek, Irene Shivaei, Alison L. Coil, Brian Siana, Bahram Mobasher, Sedona H. Price, Mojegan Azadi, Guillermo Barro, Laura de Groot, Tara Fetherolf, Francesca M. Fornasini, Gene C. K. Leung, Tom O. Zick

Abstract: We study the properties of 30 spectroscopically-identified pairs of galaxies observed during the peak epoch of star formation in the universe. These systems are drawn from the MOSFIRE Deep Evolution Field (MOSDEF) Survey at $1.4 \leq z \leq 3.8$, and are interpreted as early-stage galaxy mergers. Galaxy pairs in our sample are identified as two objects whose spectra were collected on the same Keck… ▽ More We study the properties of 30 spectroscopically-identified pairs of galaxies observed during the peak epoch of star formation in the universe. These systems are drawn from the MOSFIRE Deep Evolution Field (MOSDEF) Survey at $1.4 \leq z \leq 3.8$, and are interpreted as early-stage galaxy mergers. Galaxy pairs in our sample are identified as two objects whose spectra were collected on the same Keck/MOSFIRE spectroscopic slit. Accordingly, all pairs in the sample have projected separations $R_{\rm proj}\leq 60$ kpc. The velocity separation for pairs was required to be $Δv \leq 500 \mbox{ km s}^{-1}$, which is a standard threshold for defining interacting galaxy pairs at low redshift. Stellar mass ratios in our sample range from 1.1 to 550, with 12 ratios closer than or equal to 3:1, the common definition of a "major merger." Studies of merging pairs in the local universe indicate anenhancement in star-formation activity and deficit in gas-phase oxygen abundance relative to isolated galaxies of the same mass. We compare the MOSDEF pairs sample to a control sample of isolated galaxies at the same redshift, finding no measurable SFR enhancement or metallicity deficit at fixed stellar mass for the pairs sample. The lack of significant difference between the average properties of pairs and control samples appears in contrast to results from low-redshift studies, although the small sample size and lower signal-to-noise of the high-redshift data limit definitive conclusions on redshift evolution. These results are consistent with some theoretical works suggesting a reduced differential effect of pre-coalescence mergers on galaxy properties at high redshift -- specifically that pre-coalescence mergers do not drive strong starbursts. △ Less

Submitted 13 February, 2019; v1 submitted 29 August, 2018; originally announced August 2018.

Comments: 16 pages, 7 figures, accepted to ApJ

THE MOSDEF SURVEY: The Nature of Mid-Infrared Excess Galaxies and a Comparison of IR and UV Star Formation Tracers at z~2

Authors: Mojegan Azadi, Alison Coil, James Aird, Irene Shivaei, Naveen Reddy, Alice Shapley, Mariska Kriek, William R. Freeman, Gene C. K. Leung, Bahram Mobasher, Sedona H. Price, Ryan L. Sanders, Brian Siana, Tom Zick

Abstract: We present an analysis using the MOSFIRE Deep Evolution Field (MOSDEF) survey on the nature of "MIR-excess" galaxies, which have star formation rates (SFR) inferred from mid-infrared (MIR) data that is substantially elevated relative to that estimated from dust-corrected UV data. We use a sample of $\sim$200 galaxies and AGN at $1.40<z<2.61$ with 24 $μ$m detections (rest-frame 8$μ$m) from MIPS/\te… ▽ More We present an analysis using the MOSFIRE Deep Evolution Field (MOSDEF) survey on the nature of "MIR-excess" galaxies, which have star formation rates (SFR) inferred from mid-infrared (MIR) data that is substantially elevated relative to that estimated from dust-corrected UV data. We use a sample of $\sim$200 galaxies and AGN at $1.40<z<2.61$ with 24 $μ$m detections (rest-frame 8$μ$m) from MIPS/\textit{Spitzer}. We find that the identification of MIR-excess galaxies strongly depends on the methodologies used to estimate IR luminosity ($\rm L_{IR}$) and to correct the UV light for dust attenuation. We find that extrapolations of the SFR from the observed 24 $μ$m flux, using luminosity-dependent templates based on local galaxies, substantially overestimate $\rm L_{IR}$ in $z\sim2$ galaxies. By including \textit{Herschel} observations and using a stellar mass-dependent, luminosity-independent $\rm L_{IR}$, we obtain more reliable estimates of the SFR and a lower fraction of MIR-excess galaxies. Once stellar mass selection biases are taken into account, we identify $\sim24\%$ of our galaxies as MIR-excess. However, $\rm SFR_{Hα}$ is not elevated in MIR-excess galaxies compared to MIR-normal galaxies, indicating that the intrinsic fraction of MIR-excess may be lower. Using X-ray, IR, and optically-selected AGN in MOSDEF, we do not find a higher prevalence for AGN in MIR-excess galaxies relative to MIR-normal galaxies. A stacking analysis of X-ray undetected galaxies does not reveal a harder spectrum in MIR-excess galaxies relative to MIR-normal galaxies. Our analysis indicates that AGN activity does not contribute substantially to the MIR excess and instead implies that it is likely due to the enhanced PAH emission. △ Less

Submitted 23 June, 2018; originally announced June 2018.

Globular Clusters in High-Redshift Dwarf Galaxies: A Case Study from the Local Group

Authors: Tom O. Zick, Daniel R. Weisz, Michael Boylan-Kolchin

Abstract: We present the reconstructed evolution of rest-frame ultra-violet (UV) luminosities of the most massive Milky Way dwarf spheroidal satellite galaxy, Fornax, and its five globular clusters (GCs) across redshift, based on analysis of the stellar fossil record and stellar population synthesis modeling. We find that (1) Fornax's (proto-)GCs can generate $10-100$ times more UV flux than the field popul… ▽ More We present the reconstructed evolution of rest-frame ultra-violet (UV) luminosities of the most massive Milky Way dwarf spheroidal satellite galaxy, Fornax, and its five globular clusters (GCs) across redshift, based on analysis of the stellar fossil record and stellar population synthesis modeling. We find that (1) Fornax's (proto-)GCs can generate $10-100$ times more UV flux than the field population, despite comprising $<\sim 5\%$ of the stellar mass at the relevant redshifts; (2) due to their respective surface brightnesses, it is more likely that faint, compact sources in the Hubble Frontier Fields (HFFs) are GCs hosted by faint galaxies, than faint galaxies themselves. This may significantly complicate the construction of a galaxy UV luminosity function at $z>3$. (3) GC formation can introduce order-of-magnitude errors in abundance matching. We also find that some compact HFF objects are consistent with the reconstructed properties of Fornax's GCs at the same redshifts (e.g., surface brightness, star formation rate), suggesting we may already have detected proto-GCs in the early Universe. Finally, we discuss the prospects for improving the connections between local GCs and proto-GCs detected in the early Universe. △ Less

Submitted 19 February, 2018; originally announced February 2018.

Comments: 13 pages, 6 figures, submitted to MNRAS

The MOSDEF survey: a stellar mass-SFR-metallicity relation exists at $z\sim2.3$

Authors: Ryan L. Sanders, Alice E. Shapley, Mariska Kriek, William R. Freeman, Naveen A. Reddy, Brian Siana, Alison L. Coil, Bahram Mobasher, Romeel Davé, Irene Shivaei, Mojegan Azadi, Sedona H. Price, Gene Leung, Tara Fetherholf, Laura de Groot, Tom Zick, Francesca M. Fornasini, Guillermo Barro

Abstract: We investigate the nature of the relation among stellar mass, star-formation rate, and gas-phase metallicity (the M$_*$-SFR-Z relation) at high redshifts using a sample of 260 star-forming galaxies at $z\sim2.3$ from the MOSDEF survey. We present an analysis of the high-redshift M$_*$-SFR-Z relation based on several emission-line ratios for the first time. We show that a M$_*$-SFR-Z relation clear… ▽ More We investigate the nature of the relation among stellar mass, star-formation rate, and gas-phase metallicity (the M$_*$-SFR-Z relation) at high redshifts using a sample of 260 star-forming galaxies at $z\sim2.3$ from the MOSDEF survey. We present an analysis of the high-redshift M$_*$-SFR-Z relation based on several emission-line ratios for the first time. We show that a M$_*$-SFR-Z relation clearly exists at $z\sim2.3$. The strength of this relation is similar to predictions from cosmological hydrodynamical simulations. By performing a direct comparison of stacks of $z\sim0$ and $z\sim2.3$ galaxies, we find that $z\sim2.3$ galaxies have $\sim0.1$ dex lower metallicity at fixed M$_*$ and SFR. In the context of chemical evolution models, this evolution of the M$_*$-SFR-Z relation suggests an increase with redshift of the mass-loading factor at fixed M$_*$, as well as a decrease in the metallicity of infalling gas that is likely due to a lower importance of gas recycling relative to accretion from the intergalactic medium at high redshifts. Performing this analysis simultaneously with multiple metallicity-sensitive line ratios allows us to rule out the evolution in physical conditions (e.g., N/O ratio, ionization parameter, and hardness of the ionizing spectrum) at fixed metallicity as the source of the observed trends with redshift and with SFR at fixed M$_*$ at $z\sim2.3$. While this study highlights the promise of performing high-order tests of chemical evolution models at high redshifts, detailed quantitative comparisons ultimately await a full understanding of the evolution of metallicity calibrations with redshift. △ Less

Submitted 26 June, 2018; v1 submitted 1 November, 2017; originally announced November 2017.

Comments: 19 pages, 8 figures, accepted to ApJ

The MOSDEF Survey: Direct Observational Constraints on the Ionizing Photon Production Efficiency, $ξ_{\rm ion}$, at z~2

Authors: Irene Shivaei, Naveen A. Reddy, Brian Siana, Alice E. Shapley, Mariska Kriek, Bahram Mobasher, William R. Freeman, Ryan L. Sanders, Alison L. Coil, Sedona H. Price, Tara Fetherolf, Mojegan Azadi, Gene Leung, Tom Zick

Abstract: We combine spectroscopic measurements of H$α$ and H$β$ and UV continuum photometry for a sample of 673 galaxies from the MOSFIRE Deep Evolution Field survey to constrain hydrogen ionizing photon production efficiencies ($ξ_{\rm ion}$, xi_ion) at z=1.4-2.6. We find average log(xi_ion/[Hz erg$^{-1}$])=25.06 (25.34), assuming the Calzetti (SMC) curve for the UV dust correction and a scatter of 0.28 d… ▽ More We combine spectroscopic measurements of H$α$ and H$β$ and UV continuum photometry for a sample of 673 galaxies from the MOSFIRE Deep Evolution Field survey to constrain hydrogen ionizing photon production efficiencies ($ξ_{\rm ion}$, xi_ion) at z=1.4-2.6. We find average log(xi_ion/[Hz erg$^{-1}$])=25.06 (25.34), assuming the Calzetti (SMC) curve for the UV dust correction and a scatter of 0.28 dex in xi_ion distribution. After accounting for observational uncertainties and variations in dust attenuation, we conclude that the remaining scatter in xi_ion is likely dominated by galaxy-to-galaxy variations in stellar populations, including the slope and upper-mass cutoff of the initial mass function, stellar metallicity, star-formation burstiness, and stellar evolution (e.g., single/binary star evolution). Moreover, xi_ion is elevated in galaxies with high ionization states (high [OIII]/[OII]) and low oxygen abundances (low [NII]/H$α$ and high [OIII]/H$β$) in the ionized ISM. However, xi_ion does not correlate with the offset from the z~0 star-forming locus in the BPT diagram, suggesting no change in the hardness of ionizing radiation accompanying the offset from the z~0 sequence. We also find that galaxies with blue UV spectral slopes ($\langleβ\rangle$=-2.1) have elevated xi_ion by a factor of ~2 relative to the average xi_ion of the sample ($\langleβ\rangle$=-1.4). If these blue galaxies are similar to those at z > 6, our results suggest that a lower Lyman continuum escape fraction is required for galaxies to maintain reionization, compared to the canonical xi_ion predictions from stellar population models. Furthermore, we demonstrate that even with robustly dust-corrected H$α$, the UV dust attenuation can cause on average a ~0.3dex systematic uncertainty in xi_ion calculations. △ Less

Submitted 3 March, 2018; v1 submitted 31 October, 2017; originally announced November 2017.

Comments: 14 pages, 8 figures, 1 table, ApJ in press

The MOSDEF Survey: Metallicity Dependence of the PAH Emission at High Redshift and Implications for 24 micron-inferred IR Luminosities and Star Formation Rates at z~2

Authors: Irene Shivaei, Naveen Reddy, Alice Shapley, Brian Siana, Mariska Kriek, Bahram Mobasher, Alison Coil, William Freeman, Ryan Sanders, Sedona Price, Mojegan Azadi, Tom Zick

Abstract: We present results on the variation of 7.7 micron Polycyclic Aromatic Hydrocarbon (PAH) emission in galaxies spanning a wide range in metallicity at z ~ 2. For this analysis, we use rest-frame optical spectra of 476 galaxies at 1.37 < z < 2.61 from the MOSFIRE Deep Evolution Field (MOSDEF) survey to infer metallicities and ionization states. Spitzer/MIPS 24 micron and Herschel/PACS 100 and 160 mic… ▽ More We present results on the variation of 7.7 micron Polycyclic Aromatic Hydrocarbon (PAH) emission in galaxies spanning a wide range in metallicity at z ~ 2. For this analysis, we use rest-frame optical spectra of 476 galaxies at 1.37 < z < 2.61 from the MOSFIRE Deep Evolution Field (MOSDEF) survey to infer metallicities and ionization states. Spitzer/MIPS 24 micron and Herschel/PACS 100 and 160 micron observations are used to derive rest-frame 7.7 micron luminosities (L(7.7)) and total IR luminosities (L(IR)), respectively. We find significant trends between the ratio of L(7.7) to L(IR) (and to dust-corrected SFR) and both metallicity and [OIII]/[OII] (O32) emission-line ratio. The latter is an empirical proxy for the ionization parameter. These trends indicate a paucity of PAH emission in low metallicity environments with harder and more intense radiation fields. Additionally, L(7.7)/L(IR) is significantly lower in the youngest quartile of our sample (ages of 500 Myr) compared to older galaxies, which may be a result of the delayed production of PAHs by AGB stars. The relative strength of L(7.7) to L(IR) is also lower by a factor of ~ 2 for galaxies with masses $M_* < 10^{10}M_{\odot}$, compared to the more massive ones. We demonstrate that commonly-used conversions of L(7.7) (or 24 micron flux density; f(24)) to L(IR) underestimate the IR luminosity by more than a factor of 2 at $M_*$ ~ $10^{9.6-10.0} M_{\odot}$. We adopt a mass-dependent conversion of L(7.7) to L(IR) with L(7.7)/L(IR)= 0.09 and 0.22 for $M_* < 10^{10}$ and $> 10^{10} M_{\odot}$, respectively. Based on the new scaling, the SFR-$M_*$ relation has a shallower slope than previously derived. Our results also suggest a higher IR luminosity density at z ~ 2 than previously measured, corresponding to a ~ 30% increase in the SFR density. △ Less

Submitted 17 February, 2017; v1 submitted 15 September, 2016; originally announced September 2016.

Comments: 17 pages (with appendix), 8 figures, 2 tables. Accepted to ApJ

The MOSDEF Survey: Dynamical and Baryonic Masses and Kinematic Structures of Star-Forming Galaxies at $1.4 \leq z \leq 2.6$

Authors: Sedona H. Price, Mariska Kriek, Alice E. Shapley, Naveen A. Reddy, William R. Freeman, Alison L. Coil, Laura de Groot, Irene Shivaei, Brian Siana, Mojegan Azadi, Guillermo Barro, Bahram Mobasher, Ryan L. Sanders, Tom Zick

Abstract: We present H$α$ gas kinematics for 178 star-forming galaxies at z~2 from the MOSFIRE Deep Evolution Field survey. We have developed models to interpret the kinematic measurements from fixed-angle multi-object spectroscopy, using structural parameters derived from CANDELS HST/F160W imaging. For 35 galaxies we measure resolved rotation with a median $(V/σ_{V,0})_{R_E}=2.1$. We derive dynamical masse… ▽ More We present H$α$ gas kinematics for 178 star-forming galaxies at z~2 from the MOSFIRE Deep Evolution Field survey. We have developed models to interpret the kinematic measurements from fixed-angle multi-object spectroscopy, using structural parameters derived from CANDELS HST/F160W imaging. For 35 galaxies we measure resolved rotation with a median $(V/σ_{V,0})_{R_E}=2.1$. We derive dynamical masses from the kinematics and sizes and compare them to baryonic masses, with gas masses estimated from dust-corrected H$α$ star formation rates (SFRs) and the Kennicutt-Schmidt relation. When assuming that galaxies with and without observed rotation have the same median $(V/σ_{V,0})_{R_E}$, we find good agreement between the dynamical and baryonic masses, with a scatter of $σ_{RMS}=0.34$dex and a median offset of $Δ\log_{10}M=0.04$dex. This comparison implies a low dark matter fraction (8% within an effective radius) for a Chabrier initial mass function (IMF), and disfavors a Salpeter IMF. Moreover, the requirement that $M_{dyn}/M_{baryon}$ should be independent of inclination yields a median value of $(V/σ_{V,0})_{R_E}=2.1$ for galaxies without observed rotation. If instead we treat the galaxies without detected rotation as early-type galaxies, the masses are also in reasonable agreement ($Δ\log_{10}M=-0.07$dex, $σ_{RMS}=0.37$dex). The inclusion of gas masses is critical in this comparison; if gas masses are excluded there is an increasing trend of $M_{dyn}/M_{*}$ with higher specific SFR (SSFR). Furthermore, we find indications that $V/σ$ decreases with increasing H$α$ SSFR for our full sample, which may reflect disk settling. We also study the Tully-Fisher relation and find that at fixed stellar mass $S_{0.5}=(0.5V_{2.2}^2+σ_{V,0}^2)^{1/2}$ was higher at earlier times. At fixed baryonic mass, we observe the opposite trend. [abridged] △ Less

Submitted 15 January, 2016; v1 submitted 10 November, 2015; originally announced November 2015.

Comments: 26 pages, 23 figures, accepted to ApJ