Research Interests

Benjamin D. Oppenheimer

oppenheimer at strw dot leidenuniv dot nl
This page summarizes some of my research interests. Additionally, I try to fit my research into the broader context of galaxy formation. I also try to distill what I think is most important and express some opinions you won't find in my papers. A more general explanation focused on visualizations can be found here.

About Me: I am a post-doctoral researcher using the tools of cosmological hydrodynamic simulations to unravel the mysteries of galaxy formation. Right now, I have a VENI Fellowship at Leiden Observatory where I mainly work with Joop Schaye and the OWLS team, but I still am connected to my dissertation adviser, Romeel Davé and our extended collaborators. My thesis concentrated on the enrichment of the intergalactic medium (IGM) with the original intent of finding Z(z), however we found that understanding the chemical enrichment of the IGM likely holds critical clues about how galaxies form and regulate their growth. As a field we may not have "solved" galaxy formation (or are even close), but a significant piece of the puzzle is the galaxy-IGM connection. And it may not be the link we initially expected even just a couple years ago.

A New Paradigm of Galaxy Formation: The era when galaxies were treated as closed boxes is over-- just ask any galaxy formation theorist these days. Galaxies are living, evolving, literally breathing entities whose central characteristics may be set by the regulation of gas via accretion and outflows. Anyone who runs a cosmological simulation can find that gas often accretes onto galaxies via filaments, but it was Dusan Keres's work that really demonstrated this "cold mode" dominated how common galaxies get their gas. Working at Arizona, we started running simulations with theoretically and observationally motivated wind prescriptions, and we made galaxies and an IGM that agreed surprisingly well with observations (see below). We argue the galaxy formation depends more on how galaxies "lose" their gas via outflows as opposed to the star formation law that determines how gas is turned into stars. Joop and his crew find the same thing with their much larger set of simulations, but I bet a lot of astronomers still find this idea heretical. Also, read to the end as to why I put parentheses around lose.

Our Wind Prescription: Norm Murray came and gave a talk at Arizona early 2005 and we thought it would be interesting to add their momentum-driven wind prescription. Little did we know that these star formation-driven winds with higher mass-loading for low-mass galaxies would be so successful. We don't actually "drive" these winds since we cannot hope to resolve the scale of this physics, but we put in the relations. I'm sometimes ambivalent about these winds because I think there are some theoretical challenges in how they work (i.e. a single photon has to scatter off of dozens of dust grains), but I think the high mass-loading at early times and from low-mass galaxies is key. Maybe it is another physical mechanism driving the outflows-- playing around with thermally-driven wind prescriptions I can easily drive highly mass-loaded outflows, but I'll wait for the OWLS team to show how successful such winds are (stay tuned). One other thing I learned is that I initially kept tweaking our wind model in each subsequent paper (Joop constantly reminds me of this) and we generated an overly complicated wind prescription. Sometimes it's nice to explore simpler wind prescriptions to understand what simple changes can do (i.e. halving wind velocities), because often such tweaks make huge changes and you can understand so much more. However, our philosophy remains to put in the wind relations observed in the local Universe at all epochs. Our high-z winds aren't comparatively exotic and in fact have quite moderate energy input.

The High-Redshift (z>5) Universe: Romeel, Kristian Finlator, and I have done a series of papers on the high-z Universe looking at the reionizing galaxies, the physical characteristics of such galaxies, the enrichment of the reionization-era IGM, and the reionization of the Universe using Kristian's radiative transfer code. I see our approach to the reionization-era epoch Universe as a three-pronged approach where we self-consistently model i) early galaxies, ii) IGM metallicity, and iii) and the process of reionization itself. I think we do a surprisingly good job of all three, but it is Kristian who will bring everything together when he gets his rad-hydro version of Gadget going and the interplay of different forms of feedback (galactic winds vs. radiative) is better understood. I find that only ~1% of the z=6 Universe is enriched and that C IV hold great potential for tracing the true volume filling factor of the high-z Universe. I think the lack of weak C IV lines observed by George Becker is a stunning finding and may hint at a mostly pristine IGM, but I'm really excited about new observations with XSHOOTER and FIRE, because these instruments may be able to detect metal-enrichment from Pop III stars potentially covering a much larger volume at the level of Z~0.001 Zsolar. Stay tuned because these new observations could completely turn what we know about early-IGM enrichment on its head-- it has happened before!

The Medium-Redshift (z~2-5) IGM: Much happens during the peak of star formation in our Universe including the continual enrichment of the IGM, which was central to my first paper with Romeel. Momentum-driven winds are just right to prolifically enrich the IGM (due to high mass-loading from small galaxies) without overheating it with excessively fast outflows. We argued that Omega(C IV) stays relatively constant despite the enrichment level of the IGM increasing by a factor of 10, because the ionization correction for C IV also increases by a similar factor due to decreasing physical densities, a potentially increasing ionization background, and hotter metals. The column density distribution of C IV absorbers is the most important thing to consider as Omega(C IV) is highly stochastic, often reliant on only a couple strong absorbers. We do quite well relative to this result, although I have to say I under-estimated the strength of our strongest simulated C IV lines in that paper (long story, but e-mail me as to why). So it was actually a relief to me when d'Odorico et al. found an increasing Omega(C IV) over this range, predominantly due to an increase in strong lines from z=4->1.5 while the density of total lines stayed relatively constant. Early galaxies (z>4) can enrich a greater volume reaching lower overdensities with time, while later galaxies (z<4) re-enrich more overdense regions making stronger absorbers. While Omega(C IV) increases, Omega(C) increases faster, so don't ever use a constant C IV ionization correction with time!

The Low-Redshift (z=0-0.5) IGM: With the advent of the Cosmic Origins Spectrograph on Hubble, plus the fact that my post-doc year in Arizona was funded through a HST Theory Grant, we published our paper on the low-z IGM as traced by O VI. Initially I was afraid to publish anything on this, because most of the O VI in our simulations is photo-ionized and this would upset people who thought O VI was a good tracer of the WHIM (IGM at T>100,000 K). However our O VI did fit the observations surprisingly well, plus a series of observational papers came out, including Todd Tripp's that indicated weaker O VI lines are often photo-ionized. We went ahead and argued that most O VI lines are photo-ionized, trace moderate overdensities within 300 kpc of mostly sub-L* galaxies, and gain much of their linewidth through turbulent broadening. Definitely some O VI has to be collisionally ionized (i.e. the strongest lines), and while I'm tempted to write a paper exploring the potential that most Omega(O VI) is collisionally ionized I'll let someone else do this (i.e. R. Cen, T. Tepper-Garcia, or B. Smith). However the real test is to see how well it aligns with H I lines. COS will give us some new answers surely by the end of 2010!

Wind Recycling: I started tracking wind particles in our simulations and to my surprise I found that they mostly recycle! That is, winds launched at velocities in excess of the escape velocity of the halo potential return to the galaxy's ISM on a shorter timescale than expected from gravitational considerations alone. Romeel and I went into length about this in our feedback over a Hubble time paper where we integrated the amount of mass launched in winds to be 50% of Omega(b), which actually corresponded to about 15-20% of baryons since the average wind particle was launched about three times (i.e. recycled twice). Outflows are a force in galaxy formation when you compare this to only 5-8% of baryons observed to be in stars! Serena Bertone and others found similar results using a semi-analytical approach, and like us found that winds more easily escape low-mass halos. Even though our winds are launched at a constant multiple of the halo escape velocity, winds recycle faster when launched from more massive galaxies living in denser environments because hydrodynamic slowing dominates. Wind recycling dominates especially in the z<1 Universe where our outflows cannot escape from halos M* or greater galaxies. Romeel thought it would be cool to coin a new term for this: "halo fountains."

The History of Chemical Enrichment: Having finished my thesis in 2008, I summarized (pdf- summary only) what I think are the most important points and a self-consistent picture of 13+ Gyrs of chemical enrichment. I don't think it is energetically difficult to enrich the IGM to the observed levels especially because we find more metals close to galaxies, on average traveling ~100 physical kpc at all epochs. Metals are distributed inhomogeneously around galaxies, which is something we get for free in SPH simulations-- one cannot apply the semi-analytical paint brush of smoothly distributed metals to a post-processed simulation (but this is still an extremely valuable model to evaluate). It is easiest to enrich the largest volume at early times, but as of yet I don't see any evidence for large swaths of the IGM enriched and think it is quite possible much of the IGM remains pristine in the present day. Also check out our paper on the enrichment history of baryons, which shows that metals at z=0 live at much higher overdensities than they do at z=3.

Galaxy Assembly via Recycled Wind Accretion: The concept that recycled wind material provides a new avenue of accretion and star formation represents closure for me on my thesis research. By simulating observations of metal enrichment traced by C IV and O VI from z=6->0, we find metals remain in the proximity of galaxies-- left to the forces of gravity and hydrodynamics these wind materials participate in later galaxy growth. Our paper introduced a third channel of accretion: recycled wind accretion, in addition to hot and cold mode accretion. I thought it could be overzealous to introduce a new accretion mode, so I was more than gratified when the six other authors on this paper agreed to not only be on this paper but heavily participated in its production. I have some doubts whether "wind mode" is accreting onto galaxies in the real Universe as it is in our simulations, especially for more massive galaxies, but I hope others consider, critique, and even challenge re-accreting winds as a significant source of galaxy growth. The observational tie-in though is that we can just about reproduce the shallow slope of galaxy stellar mass function (GSMF) below M*, because wind recycle faster onto more massive galaxies. Outflows govern the GSMF not primarily through the ejection of materials, but by how those ejected materials are re-accreted. This paper also explores how Neal Katz's two versions of feedback, ejective and preventative, suppress feedback in simulations, which is itself is valuable theoretical exercise.

Quenching Star Formation in Massive Galaxies: Star formation-driven winds appear incapable of suppressing star formation in massive galaxies, therefore Jared Gabor has a paper exploring quenching mechanisms for massive galaxy growth. We don't have the sharp knee of the observed Schechter function in our simulated GSMFs, nor the characteristic red sequence-blue cloud bimodality of galaxies as observed. Jared's first step was to post-process simulations using various quenching mechanisms. A quenching mechanism based on a sharp cut a critical virial mass fails drastically, but so does any sort of evolving mass cut explored. Quenching based on merger criteria seem to fare better, but produces too many massive galaxies. Quenching massive galaxy growth cannot be implemented using a simple prescription, so Jared is going to implement actual AGN feedback physics into the simulations. Still, it may be even more complicated and require new physics/better resolution to deal with cold dense gas in hot diffuse halos (e.g. cold drizzle).

The Thermal History of Baryons: Simulations consistently show baryons growing hotter with time as the warm-hot IGM (WHIM) forms from larger structures collapsing and shocking. Feedback additionally heats the IGM, especially at early times, but is subdominant to structure-formation shocking at late times usually resulting in 40-50% of z=0 baryons at T>100,000 K. The decline in global star formation is almost certainly related to the formation of hot structures from which accretion onto galaxies is more difficult, but dedicated COS observations should hopefully help us understand this connection between baryons and galaxies better. For now, I just make pretty movies of the temperature evolution (Mp4 41 MB, Mpeg 49 MB), but I intend to do more later regarding the thermal history of baryons.

The Missing Baryon Problem: I don't lose too much sleep over missing baryons, but I'm not nearly as familiar with this issue as someone like Joel Bregman or Stacey McGaugh. I declared we found all the baryons in our simulation box at the Sydney Missing Baryons Conference, but of course this is not that surprising when one includes them all in the ICs. However, Romeel showed that outflows evacuate a significant fraction of their baryons from their halos using our outflows, especially at lower masses. I still think it is going to be really, really hard to find of the baryons any time soon, because so many of them are in a hot, diffuse phase. And don't get me started about using metal lines to trace missing baryons assuming a constant metallicity and ionization correction.

The Mass-Metallicity Relation: Kristian Finlator is the expert on this, and published a really nice paper explaining how the mass-metallicity relation can be understood as the balance between accretion and feedback-- the cornerstone in our view of how galaxy formation is regulated. I can understand this paper at the level of the abstract, but Kristian goes into much more analytical detail for those who are interested. More work will be done on this soon, especially concerning how this relation evolves over a Hubble time.

Sub-MM Galaxies: Romeel recently submitted our sub-mm galaxies paper arguing that such galaxies violate the First Law of Ducks, which I think means these aren't simply z=0 ducks (ULIRGs that are recent major mergers) scaled up to be z=2 ducks (Hyper-LIRGs, also major mergers). They are more ordinary, passively evolving massive star forming galaxies residing in the most massive halos at this redshift, more like fat pigeons I guess. These are the high-mass tail of the "galaxy main sequence" at z=2, which will eventually evolve into BCGs of clusters today. Still, our SFRs are lower by a factor of three than inferred from the far-IR, but this disagreement may be tractable given that observers disagree on stellar masses of sub-mm galaxies by a factor of six!

Downsizing: I'm often surprised that there are two distinct types of downsizings in a LambdaCDM universe. Archaeological downsizing, where massive halos form their stars earlier, is naturally expected from the bottom-up clustering of halos. Specific SFR downsizing, where smaller galaxies are forming stars faster relative to their masses, is a completely different phenomenon and may be difficult to explain within a LambdaCDM cosmology. One can possibly delay the star formation in smaller halos, and I wonder if recycling gas in outflows is a way to do this since recycling takes longer at lower masses. Still, I think we first need to quench more massive galaxy growth in our simulations before we start drawing conclusions, because without quenching we don't get SSFR downsizing in our favored vzw wind model.

Groups and Clusters:

The Galaxy-Absorber Connection:

The Lyman-Alpha Forest:


Scientific Visualization:

Back to Homepage