Astronomy Lunch Seminar on Fridays


Seminars start at 1:30 pm and take place in Room 310 in the Collaborative Learning and Innovation Complex (CLIC).

 

SPRING 2017 semester:


Jan 27: Dr. Or Graur (CfA - Harvard)


TITLE: Using Supernova Rates to Reveal their Progenitors


ABSTRACT: We still do not know what types of stellar systems end up exploding as most types of supernovae (SNe). In my talk, I will show how we can use observed correlations between the SN explosion rates and various host-galaxy properties to constrain the progenitor scenarios of different types of SNe. I will focus on Type Ia supernovae, as well as stripped-envelope supernovae (i.e., SNe Ib, Ic, and IIb). The rates, which come from a variety of surveys (of both nearby and high-redshift galaxies, as well as both imaging and spectroscopic surveys), consistently point in the same direction: most Type Ia supernovae probably come from the explosion of two white dwarfs. Stripped-envelope supernovae probably emerge from binary systems as well, and not single, massive stars.



Feb 3: Dr. James Guillochon (CfA - Harvard)


TITLE: Tidal disruptions of stars by supermassive black holes: dynamics, light, and relics


ABSTRACT: Most supermassive black holes in the local universe lie dormant, with only one in a hundred accreting at their Eddington limits. Aside from this active minority, and the black holes in nearby galaxies that we can observe to influence the dynamics of stars and gas, most remain difficult to study. Tidal disruptions of stars by supermassive black holes give these dormant black holes a chance to be seen once every ~10,000 years, and each tidal disruption brings along with it a host of observable signatures that can be studied from gigaparsecs away, from the moment of the disruption to millennia after a disruption has occurred. In my talk I will present work I have done on tidal disruptions of stars, and describe their dynamics, observational signatures from real-time monitoring, and relics of disruption that may exist in plain sight.



Feb 10: CANCELLED DUE TO SNOW STORM...



Feb 17: Dr. Anna Pancoast (CfA - Harvard)


TITLE: Measuring Black Hole Masses in AGN by Revealing the Hidden Broad Line Region


ABSTRACT: Echoes from the broad line region in active galactic nuclei (AGN) allow for the measurement of supermassive black hole masses outside the local Universe. However, the detailed structure of the broad line region is difficult to constrain due to the very small scales involved. With a new generation of high-quality reverberation mapping datasets, we can substitute time resolution for spatial resolution and begin to model echoes from the broad line region directly. I will discuss the development of a direct modeling approach for reverberation mapping data capable of measuring the absolute black hole mass and the geometry and dynamics of the broad line region.  Applying this approach to recent datasets we find that the broad line region emission comes from a thick disk from gas that is generally inflowing towards the black hole.  Finally, I will discuss how this work can improve AGN black hole mass measurements at all redshifts.



Feb 24: Dr. Joel Leja (CfA - Harvard)


TITLE: The Next Generation of Galaxy SED Fitting: Accurate Physical Properties from Broadband Photometry with Prospector


ABSTRACT: Broadband photometry of galaxies measures an unresolved mix of complex stellar populations, gas, and dust. Interpreting these data is a challenge for models: I will show several studies demonstrating that properties derived from modeling galaxy photometry are uncertain by a factor of two or more. Yet, answering key questions in the field now requires higher accuracy than this. Here, we present a new model framework specifically designed for these complexities. Our model, Prospector-α, includes dust attenuation and re-radiation, a flexible attenuation curve, nebular emission, stellar metallicity, and a 6-component nonparametric star formation history. The flexibility and range of the parameter space, coupled with MCMC sampling within the Prospector inference framework, is designed to provide unbiased parameters and realistic error bars. We assess the accuracy of the model with aperture-matched optical spectroscopy, which was excluded from the fits. We compare spectral features predicted solely from fits to the broadband photometry to the observed spectral features. Our model predicts Hα luminosities with a scatter of ∼0.16 dex and an offset of 0.08 dex across a wide range of morphological types and stellar masses. This agreement is remarkable, as the Hα luminosity is dependent on accurate star formation rates, dust attenuation, and stellar metallicities. The model also accurately predicts dust-sensitive Balmer decrements, spectroscopic stellar metallicities, PAH mass fractions, and the age- and metallicity-sensitive features Dn4000 and Hδ.



Mar 3: Mackenzie Jones (Dartmouth)


TITLE: Exploring the consequences of selection effects in black hole-galaxy evolution studies 


ABSTRACT: In probing the connection between active galactic nuclei (AGN) and their host galaxies, an understanding of the full underlying AGN population is presently limited by complex observational biases that can be challenging to model theoretically. Determining the intrinsic Eddington ratio distribution in the optical, in particular, can be strongly influenced by selection effects and dilution from the host galaxy star formation. Using spectroscopic galaxies from SDSS, we show that an optically selected AGN population is consistent with being drawn from a universal AGN Eddington ratio distribution in the form of the broad Schechter function that is observed in the X-ray surveys. To more completely explore selection effects in the context of galaxy formation, we use a semi-numerical galaxy evolution simulation along with this universal Eddington ratio distribution to model the AGN population. We explicitly model selection effects to produce the “observed” AGN population for comparison with both theoretical and observational X-ray data. We investigate the impact on the “observed” population of selecting AGN in the X-rays based on a thresholds in luminosity (as they are selected in most surveys) as well as Eddington ratio. We find that we can broadly reproduce the host galaxies and halos of the X-ray AGN population, and that selecting AGN using these luminosity or Eddington ratio limits yield samples with very different host galaxy properties.



Mar 31: Dr. Francesca Civano (CfA - Harvard)


TITLE: X-rays from the “COSMOS” to understand black hole and galaxy co-evolution


ABSTRACT: Observations indicate that supermassive black holes (SMBHs) ordinarily dwell at the centers of local galaxies. Scaling relations have been identified between SMBHs and many large scale properties of the host galaxies that point to a joint galaxy and black hole cosmic evolution. Is this evolution regulated by the galaxy or by the black hole? Is star-formation triggering black hole activity or vice versa? To address these questions, in this talk I will present my work testing this co-evolution scenario, focusing on SMBH growth mechanisms, accretion and mergers. I will use the extraordinarily rich multiwavelength dataset of the Cosmic Evolutionary Survey (COSMOS). I will concentrate on the highest energy data available, the X-ray ones, from the surveys I have led using both the Chandra and NuSTAR NASA satellites. These data provide us with a unique and powerful tool to find and study accreting SMBHs in the distant Universe.



Apr 14: Dr. Duncan Farrah (Virginia Tech)


TITLE: Insights into galaxy assembly from luminous AGN in the high redshift Universe


ABSTRACT: There is a deep connection between star formation and AGN activity which profoundly impacts the mass assembly history of galaxies. The nature of the connection however remains controversial, due to, for example, evolution in the AGN and starburst duty cycles, and the obscuring effect of dust. In this talk I will briefly review our current knowledge of galaxy assembly across cosmic history, and then discuss two recent results. First is an observed scaling relation between star formation rates and AGN luminosities in luminous type 1 quasars at high redshift, which offers insights into how star formation may be triggered and quenched in these systems. Second is the morphological properties of the most luminous obscured AGN in the Universe, whose nature offers extremely stringent tests for galaxy evolution models and which may be signposting brief but critical phases in massive galaxy assembly. 






FALL 2016 semester:


Sep 30: Dr. Matteo Bonato (Tufts University)


TITLE: Exploring the evolution of star formation, black hole accretion and dwarf galaxy properties with JWST/MIRI serendipitous spectroscopic surveys


ABSTRACT: I will show predictions for spectroscopic observations with the MIRI Medium Resolution Spectrometer on board JWST. Specifically, pointed observations of Herschel sources will require only a few minutes for detections of several star-forming and AGN lines (out to z=3 and beyond), allowing us to efficiently investigate early phases of the star formation rate history and of the galaxy/AGN (co-)evolution. But the same data will also include tens of serendipitous 0<z<4.5 galaxies per field with IR luminosities down to about 10^6 L_sun. Therefore, for the first time and for free, we will be able to detect very low-luminosity galaxies (L_IR<10^9 L_sun) at high redshifts, and with good statistics. This will allow us to study the properties of these extreme galaxies and to test galaxy evolution models. Such surveys will achieve an improvement over Herschel of about three orders of magnitude in the study of the evolution of the cosmic star formation rate and about one order in black hole accretion rate. These results hold for a wide range in the modeled low-L end of the IR luminosity function, and accounting for the PAH deficit in low-L, low-metallicity galaxies.



Oct 14: Mr. Jeremy Bradford (Yale University)


TITLE: Isolated Low-Mass Galaxies: A Control Sample for Galaxy Formation and Evolution


ABSTRACT: Low-mass galaxies with stellar masses between 10^7 and 10^9 M_sun represent an interesting regime of galaxy formation. These galaxies are more susceptible to environmental processes, less efficient at forming stars, and often inconsistent with simulations. When selected with a strict isolation criterion, all low-mass galaxies are found to be star-forming. This implies that feedback processes are insufficient for shutting down star formation at these masses. Therefore, isolated galaxies are an excellent control sample for comparisons to predictions of galaxy formation. I have selected a sample of isolated, low-mass galaxies from the Sloan Digital Sky Survey and obtained neutral hydrogen observations for these galaxies. To compliment this study with high-mass galaxies, I have also re-measured data from the ALFALFA survey. I will present this sample and explore the neutral gas content and the scaling relations of isolated, gas-rich, late-type galaxies. Because it is a commonly used test of galaxy formation theory, I will focus on the baryonic Tully-Fisher scaling relation and the effect of systematic uncertainties on the relation when comparing observations to predictions.



Oct 28: Ms. Kathryn Grasha (UMass - Amherst)


TITLE: The Hierarchical Distribution of Young Stellar Clusters in Nearby Galaxies


ABSTRACT: A turbulent interstellar medium will drive the hierarchical nature of star formation, resulting in a smoothly varying distribution of substructure, where bound star clusters occupy the smallest, densest regions.  We use young stellar clusters to trace the unbound hierarchical star-forming structures for several nearby galaxies drawn from the Legacy ExtraGalactic UV Survey (LEGUS).  We implement the angular two point correlation function to quantify the clustering among stellar clusters as a function of spatial scale and age to establish whether the clustering strength and the survival timescale of the clustered substructure depends on the properties of the stellar populations.  We separate the clusters into different classes, compact (bound) clusters and associations and compare the clustering among the different classes.  We find that younger star clusters are more strongly clustered over small spatial scales and that the clustering disappears rapidly for ages as young as a few tens of Myr. We also find gravitationally bound clusters exhibit less clustering compared to stellar associations, and are no longer spatially correlated after ~40-50 Myr, and beyond 100-300 pc.  Thus, clusters are more evolved and have travel significantly from their birth site within a few tens of Myr, and associations show evidence of disruption occurring very quickly after formation, dispersing over the same timescale.



Nov 18: Dr. Rafael Martinez-Galarza (CfA - Harvard)


TITLE: Bayesian SED-fitting in the Era of Multi-wavelength Astronomy: Applications to Star-Forming Systems.


ABSTRACT: In the study of galactic and extragalactic star-forming systems, spectral energy distribution (SED) fitting is a powerful tool, boosted in recent years by a plethora of multi-wavelength observations covering the full (UV to sub-millimeter) thermal spectrum and by the development of new computational techniques that allow scientist to formulate and solve the problem in a Bayesian framework. Yet, the fundamentals of such Bayesian interpretation are not uniformly understood and applied across the many publications dealing with SED analysis in star formation and this is in part due to a lack of  sufficient statistics background among astronomers. In this talk I intend to provide an overview of a particular approach to the problem of Bayesian SED fitting. I discuss the basic elements of  Bayesian analysis in this context,  describe why Monte Carlo Markov Chains (MCMC) methods appear naturally in the process, and show two applications of Bayesian parameter estimation in star formation: the evolution of star formation in merging galaxies as a function of their interaction stage, and the physical properties of clustered Young Stellar Objects (YSOs) that appear blended together within the beam of infrared telescopes. The latter (SED-fitting of confused sources) is a common problem at all wavelengths, and I describe an algorithmic path to solve it by combining spatial and spectral information.






SPRING 2016 semester:


Feb 12: Mr. Vicente Rodriguez-Gomez (CfA - Harvard)


TITLE: The stellar mass assembly of galaxies in the Illustris simulation


ABSTRACT: In the Lambda-CDM cosmological model, structure grows hierarchically, potentially making mergers one of the primary drivers of galaxy growth. We use the Illustris cosmological simulation to study the relative contributions of in situ star formation and stellar accretion to the build-up of galaxies over an unprecedentedly wide range of masses (Mstar = 10^9−10^12 Msun), galaxy types, environments, and assembly histories. We find that the ‘two-phase’ picture of galaxy formation predicted by some models is a good approximation only for the most massive galaxies in our simulation – namely, the stellar mass growth of galaxies below a few times 10^11 Msun is dominated by in situ star formation at all redshifts. The fraction of the total stellar mass of galaxies at z = 0 contributed by accreted stars shows a strong dependence on galaxy stellar mass, ranging from about 10 per cent for Milky Way-sized galaxies to over 80 per cent for Mstar ~ 10^12 Msun objects, yet with a large galaxy-to-galaxy variation. At a fixed stellar mass, elliptical galaxies and those formed at the centers of younger halos exhibit larger fractions of ex situ stars than disk-like galaxies and those formed in older halos. On average, ~50 per cent of the ex situ stellar mass comes from major mergers (stellar mass ratio \mu > 1/4), ~20 per cent from minor mergers (1/10 < \mu < 1/4), ~20 per cent from very minor mergers (\mu < 1/10), and ~10 per cent from stars that were stripped from surviving galaxies (e.g. flybys or ongoing mergers). These components are spatially segregated, with in situ stars dominating the innermost regions of galaxies, and ex situ stars being deposited at larger galactocentric distances in order of decreasing merger mass ratio.



Feb 19: Dr. Dan Milisavljevic (CfA - Harvard)


TITLE: Reverse Engineering Supernovae


ABSTRACT: Core-collapse supernovae mark the catastrophic deaths of massive stars. They are among the most powerful explosions in the universe that affect many aspects of their host galaxies, and produce a variety of exotic objects including neutron stars, black holes, and some gamma-ray bursts.  Somewhat unsettling, however, is the fact that fundamental properties of supernovae with regard to their progenitor systems and explosion mechanisms remain poorly understood. Indeed, questions such as "Which stars explode?" and "How do stars explode?" do not have clear answers. I will review recent radio-through-X-ray observations that are revolutionizing our understanding of supernova explosions, with some emphasis on my own research that utilizes investigations of the entire supernova life cycle - from progenitor star, to explosion, to remnant - in order to "reverse engineer" key aspects of their dramatic and turbulent nature.



Mar 4: Dr. Brendan Griffen (MIT)


TITLE: There's no place like home: the assembly history of Milky Way-sized halos, from z=100 to z=0


ABSTRACT: In my talk I will present the largest number of Milky Way sized dark matter halos simulated at extremely high mass and temporal resolution done to date, more than quadrupling what is currently found in the literature (www.caterpillarproject.org). Our current set of 30+ halos of the Caterpillar Project whose project goal of 70 halos will eventually be made public. Our detailed contamination study of over thousands of low resolution halos has resulted in obtaining very large and unprecedented, high-resolution regions around our host halos (sphere of radius 2 Mpc) for our flagship resolution allowing for accurate studies of the assembly of the stellar halo, low mass dwarf galaxies at large galactocentric radii and the very first stellar systems at high redshift. In my talk I will discuss the prospects of carrying out "stellar archeology" by using semi-analytic models to identify Pop. III/II star formation sites and tracing their descendants down to the present day to probe the high-z Universe without the need for expensive high-z observations. [http://adsabs.harvard.edu/cgi-bin/bib_query?arXiv:1509.01255]. 



Apr 8: Alex Ji (MIT)


TITLE: A rare and prolific r-process event in the ultra-faint dwarf galaxy Reticulum II


ABSTRACT: Ultra-faint dwarf galaxies contain a simple fossil record of early chemical enrichment that can probe the astrophysical site of r-process nucleosynthesis. Previous measurements found very low levels of neutron-capture elements in ultra-faint dwarfs, possibly supporting supernovae as the primary r-process site. I present high-resolution chemical abundances of nine stars in the recently discovered ultra-faint dwarf Reticulum II, seven of which display extremely enhanced r-process abundances 2-3 orders of magnitude higher than the other ultra-faint dwarfs. Stars with such extreme r-process enhancements are only rarely found in the Milky Way halo. The r-process abundances imply that the neutron-capture material in Reticulum II was synthesized in a single prolific event that is incompatible with r-process yields from ordinary core-collapse supernovae. Reticulum II provides an opportunity to discriminate between different candidates for rare and prolific r-process production, such as neutron star mergers or magnetorotationally driven supernovae. The single event is also a uniquely stringent constraint on the metal mixing and star formation history of this ultra-faint dwarf galaxy.



May 13: Dr. Heath Shipley (Tufts University)


TITLE: The Luminous Polycyclic Aromatic Hydrocarbon Emission Features: Applications to High Redshift Galaxies and Active Galactic Nuclei


ABSTRACT: For decades, significant work has been applied to calibrating emission from the ultra-violet, nebular emission lines, far-infrared, X-ray and radio as tracers of the star-formation rate (SFR) in distant galaxies.  Understanding the exact rate of star-formation and how it evolves with time and galaxy mass has deep implications for how galaxies form.  The co-evolution of star-formation and supermassive black hole (SMBH) accretion is one of the key problems in galaxy formation theory.  But, many of these SFR indicators are influenced by SMBH accretion in galaxies and result in unreliable SFRs.  Utilizing the luminous polycyclic aromatic hydrocarbon (PAH) emission features, I provide a new robust SFR calibration using the luminosity emitted from the PAHs at 6.2μm, 7.7μm and 11.3μm to solve this.  The PAH features emit strongly in the mid-infrared (mid-IR; 5-25μm) mitigating dust extinction, containing on average 5-10% of the total IR luminosity in galaxies.  The PAH luminosity correlates linearly with the SFR as measured by the extinction-corrected Hα luminosity, with a tight scatter of <0.15 dex.  The scatter is comparable to that between SFRs derived from extinction-corrected Paα and Hα emission lines. This new SFR indicator will be useful for probing the peak of the SFR density in the universe (1 < z < 3) and for studying the co-evolution of star-formation and SMBH accretion contemporaneously in a galaxy. 






FALL 2015 semester:



Sep 18: Dr. Annalisa Pillepich (CfA-Harvard)


TITLE: Building Galaxies and their Stellar Haloes by In-Situ and Ex-Situ Star Formation: insights from Eris and the Illustris Simulations


ABSTRACT: In the standard LambdaCDM scenario, hierarchical clustering leads to complex galaxy assembly histories, and the paths leading to the build up of each major stellar component may be different. In particular, galactic stellar haloes are thought to be the direct evidence of the hierarchical growth of structure in the CDM paradigm and numerical simulations have been able to reproduce their broad features from the debris of accreted and disrupted satellite galaxies. In the talk, I will adopt two complementary techniques with the specific purpose of understanding how baryonic physics and the hierarchical clustering leave imprints on the stellar assembly and the stellar distributions within haloes. On the one side, I will show the results of the detailed analysis of the Eris simulation, the first N-body+SPH simulated galaxy to result in a close analog of the Milky Way: I will give an account of the relative contributions of “in-situ” (within the main host) and “ex-situ” (within satellite galaxies) star formation to each major Galactic component of such close Milky Way analog. On the other side, I will support such findings by tackling a statistically-meaningful analysis of a large sample of galaxies from the Illustris Simulation. Illustris is a state-of-the-art simulation which combines the statistical power of a ∼106 Mpc-side cosmological volume with gasdynamics, prescriptions for star formation, feedback, and kpc resolution. For example, in the first work of a series (Pillepich et al. 2014), we show that by solely measuring the power-law slope of the stellar density profile of the stellar haloes, quantitative information can be obtained regarding the total mass, density profile, formation time, and merger history of the underlying DM haloes.



Oct 23: Dr. Matteo Bonato (Tufts University)


TITLE: Predictions for imaging and spectroscopic surveys of galaxies and Active Galactic Nuclei in the mid-/far-Infrared


ABSTRACT: While continuum imaging data at far-infrared to sub-millimeter wavelengths have provided tight constraints on the population properties of dusty star forming galaxies up to high redshifts, future space missions like the Space Infra-Red Telescope for Cosmology and Astrophysics (SPICA) and ground based facilities like the Cerro Chajnantor Atacama Telescope (CCAT) will allow detailed investigations of their physical properties via their mid-/far-infrared line emission. In this talk I present my predictions for SPICA imaging and spectroscopic surveys, obtained using both a phenomenological approach and physically grounded models. These predictions are useful to optimize the planning of the surveys and exploit a recent upgrade of evolutionary models, that include the effect of strong gravitational lensing, in the light of the most recent Herschel and South Pole Telescope data. The relations between line and continuum infrared luminosity were worked out considering differences among source populations, with the support of extensive simulations that take into account dust obscuration. My reference model for the redshift dependent IR luminosity functions was the one worked out by Cai et al. (2013), based on a comprehensive hybrid approach combining a physical model for the progenitors of early-type galaxies with a phenomenological one for late-type galaxies. I present new estimates of redshift-dependent luminosity functions of IR lines detectable by SPICA and excited both by star formation and by AGN activity. To this end I worked out a major upgrade of the Cai et al. (2013) model, by dealing in a self consistent way with emission of galaxies as a whole, including both the starburst and the AGN component. These spectroscopic observations will allow us to probe all phases of the interstellar medium (ionized, atomic and molecular). Measurements of these lines will provide redshifts and key insight on physical conditions of dust obscured regions and on the energy sources controlling their temperature and pressure. This information is critically important for investigating the complex physics ruling the dust-enshrouded active star forming phase of galaxy evolution and the relationship with nuclear activity. Observations of strongly gravitationally lensed galaxies, described in detail by my reference model, will be of special interest, because strong lensing allows us to measure their total mass distribution up to very large distances and to gain information on sources too faint to be detected with current instrument sensitivities, thus testing models for galaxy formation and dark matter.



Oct 30: Dr. Aleksander Sadowski (MIT-Kavli)


TITLE: Simulations of radiative black hole accretion in general relativity


ABSTRACT: In this talk I will summarize the recent progress in simulating optically thick black hole accretion. I will discuss the methods used to evolve gas together with radiation field, their power and limitations. In detail, I will describe simulations of super-critical black hole accretion disks applicable to ultra luminous X-ray sources and tidal disruption events. Special attention will be paid to their luminosities and mildly relativistic, radiative jets formed in such systems. In the end I will discuss challenges related to simulating thin disks and show preliminary simulations of sub-critical accretion disks.



Nov 19: Dr. Remco van der Burg (Paris - CEA Saclay)


TITLE: The distribution of stellar mass in galaxy clusters over cosmic time


ABSTRACT: Galaxy clusters are the most massive gravitationally collapsed structures in the universe, and they have important cosmological and astrophysical applications. Given their extreme overdensities, they allow us to study how the evolution of galaxies depends on their environment. I will present results on the properties of galaxies in clusters around redshift 1, at a time where a significant fraction of the cluster galaxies is still forming stars. These measurements provide constraints on the times when, and the locations in the clusters where, their star formation is being quenched. I will also present measurements on the radial distribution of galaxies in two cluster samples, which span about 8 Gyrs of lookback time. By matching local galaxy clusters to their progenitors at high redshift, we study how clusters grow their stellar mass content. Interestingly, this suggests that the central part of the stellar mass distribution of local galaxy clusters is already in place at redshift, and any further growth seems to happen in an inside-out fashion. I will put these findings into context by comparing them to the results from dark matter simulations. I will also discuss future directions, both on the observational side as on the interpretation of these results.



Dec 4: Andrew Battisti (Umass - Amherst)


TITLE: Characterizing Dust Attenuation in Local Star Forming Galaxies


ABSTRACT: The presence of dust in galaxies causes their spectral energy distribution to experience reddening, a consequence of the highest attenuation occurring in the ultraviolet and decreasing towards longer wavelengths out to the infrared. If not properly accounted for, this effect can lead to incorrect values of derived physical quantities such as the star formation rate, stellar mass, and photometric redshift. Virtually all studies of star forming galaxies, both local and distant, make use the attenuation curve derived for local starburst galaxies (Calzetti et al. 2000) to correct for the effects of dust. However, it is not clear how reasonable the general application of this curve is for more typical star forming galaxies. In this talk, I will present results from a study utilizing ~10000 local (z ≤ 0.1) star forming galaxies to identify the extent to which attenuation varies as a function of galactic physical properties.



Dec 11: Allison Kirkpatrick (Umass - Amherst)


TITLE: What Lies Beneath: Dust Obscured Star Formation and Black Hole Growth


ABSTRACT: The evolution of massive dusty galaxies is driven internally by AGN growth and star formation, but due to the copious amounts of dust, it is often impossible to observe these processes directly. Instead, we must disentangle information about the inner workings of galaxies from the observed dust emission. Using a large sample of z=0.5-2.8 Ultra Luminous Infrared Galaxies, I quantify the contributions to the dust heating from AGN activity and star formation using deep Spitzer mid-IR spectroscopy. I also explore composite galaxies--those galaxies which show both strong mid-IR star formation and AGN signatures. These galaxies exhibit a clear increase of dust temperature in the far-IR as the AGN grows stronger, indicating that a growing AGN is significantly contributing to LIR, although the galaxy is still strongly star-forming. Furthermore, these composite systems are prevalent even at faint 24 micron flux thresholds, and could be an important source of AGN contamination in any high redshift sample of star forming galaxies, if care is not taken to identify them. I will discuss ways to identify composite galaxies and how to correct infrared luminosities and star formation rates for dust emission heated by an obscured AGN.