Cover Page/Proposal Summary
The XO Planet Finding System
Peter R. McCullough Jeff Valenti Chris Johns-Krull Kenneth Janes James N. Heasley
We describe a three year plan to continue discovering and characterizing planets that transit bright stars. To date, astronomers have reported more than 25 such transiting planets orbiting bright stars, and their numbers are increasing rapidly (both the number of planets and the number of astronomers studying them). The brightest typically are discovered by oscillating radial velocities, and somewhat fainter ones in much larger numbers are discovered by transits. The XO observatory has recently been expanded to six 0.1-meter diameter cameras, two of which have been operating robotically on Haleakala, Maui, HI since Sep 2003 and with which we have discovered five transiting planets. We evaluate transiting-planet candidates produced by the XO survey cameras with precise photometric observations, which significantly reduces the number of candidates we observe spectroscopically. For the follow-up photometry, we have recruited a global network of privately-funded and operated observatories, in addition to our team’s institutional access to observatories in Hawaii, Texas, and Arizona. Dozens of candidates per year will be scrutinized with moderate-precision (∼1 km/s) radial velocity observations and a few of those will be confirmed to be planets with (∼10 m/s) precision. We will pursue precise photometric and spectroscopic follow up observations of transiting planets discovered by XO, HAT, WASP, etc. Our team has prior experience with scintillation- limited ground-based photometry, time-series photometry with SST IRAC, spectrophotometry with HST instruments NICMOS, STIS, ACS, and WFC3 (in ground-tests), and is developing spectroscopic analysis techniques to improve the precision of radial velocities from Keck and HET. We also propose to increase the number of amateur astronomers observing such transits and to motivate their training to continuously improve their observational techniques, which in some cases already rival that of professionals. For the period 2003-2009, the XO team has been supported by two NASA Origins awards. Follow-up observations with HST and SST have been supported by those NASA programs. The tripling of XO’s cameras and associated electronics has been supported by peer-reviewed competition for institutional funds of the STScI Director.
1 Contents
1 Objectives and Significance 5 1.1 Scientific Objectives: find more transiting planets ...... 5 1.2 Brighter is Better ...... 5 1.3 Significance of Transits: more than P, m sin i, and e ...... 5 1.4 Longer periods: satellites as habitable worlds ...... 7 1.5 Multiple planets: spectroscopic and photometric searches ...... 7 1.6 Devote more spectroscopy to some Duds ...... 8 1.7 Better characterization: improved photometry and spectroscopy ...... 8 1.8 Additional Science Enabled ...... 8
2 Technical Approach and Methodology 9 2.1 Summary of XO Operations and Results, 2002-2008 ...... 9 2.2 Comparison to other Searches for TEPs ...... 10 2.3 How to find longer-period TEPs ...... 11 2.4 How to find multi-TEP systems ...... 14 2.5 Strategy for Followup ...... 15 2.6 Comparison to other Searches for TEPs ...... 15
3 Impact of Proposed Work 16
4 Relevance to other NASA Programs and Missions 16
5 Management Plan 17
6 Facilities, Equipment and Other Resources 24 6.1 The Robotic Camera ...... 24 6.2 The Extended Team of Amateur Astronomers ...... 24 6.3 Space Telescope Science Institute (McCullough & Valenti) ...... 24 6.4 University of Hawaii (Heasley) ...... 24 6.5 Boston University Astronomy Department and Institute for Astrophysical Re- search(Janes) ...... 24 6.6 Rice University (Johns-Krull) ...... 25
7 Vitae 27
8 Current and Pending Support 32 8.1 Peter McCullough ...... 32 8.2 Jeff Valenti ...... 32 8.3 Ken Janes ...... 32 8.4 James Heasley ...... 32
2 9 Letters of Commitment 33 9.1 Peter McCullough ...... 33 9.2 Jeff Valenti ...... 33 9.3 Ken Janes ...... 33 9.4 James Heasley ...... 34 9.5 You-Hua Chu ...... 35 9.6 Debra Fischer ...... 36 9.7 Ron Gilliland ...... 37 9.8 Frank Summers ...... 37 9.9 ET member 1, Tonny Vanmunster ...... 38 9.10 ET member 2, Paul Howell ...... 39 9.11 ET member 3, Ron Bissinger ...... 39 9.12 ET member 4, Bruce Gary ...... 40
10 Budget Narrative 41 10.1 Personnel ...... 41 10.2 Supplies ...... 41 10.3 Supplies ...... 41 10.4 Travel Costs ...... 41 10.5 Publication Costs ...... 42 10.6 Indirect Costs ...... 42
3 Summary of Personnel and Work Efforts
We request salary support for three persons, Peter McCullough, his data analyst, and his post doctoral researcher. Their work effort and schedule are as follows. In order to apportion costs appropriately, we assume that their additional efforts to study transiting hot Jupiters with HST, Warm Spitzer if it is funded, etc will be funded with grants accompanying observing time on those observatories. That is why, e.g., this grant allocates 1.5 years, not 3 years, to the Postdoc.
Name of PI/Co-I FTE1 devoted to project Years Period
McCullough 25% 3 03/15/2009-3/15/2012 Data Analyst 37.5% 3 09/01/2009-3/01/2012 Postdoc 100% 1.5 09/01/2009-3/01/2012
Valenti * 3 03/15/2009-3/15/2012 Janes * 3 03/15/2009-3/15/2012 Johns-Krull * 3 03/15/2009-3/15/2012 Heasley * 3 03/15/2009-3/15/2012
1 FTE = Full Time Equivalent * Valenti, Janes, Johns-Krull, and Heasley will each contribute approximately 0.05 FTE per year at no cost to this award. (None of them are civil servants – they can work for free without violating full cost accounting practices.) Origins will have to stretch to pay for McCullough, DA, and PD for 3 years. We need to decide how to economize. One thought is to not include a PD in this, but rely on Valenti, Janes, Johns-Krull, Heasley to write ApJ papers. I anticipate that XO, HAT, WASP will begin to publish discoveries in groups instead of one at a time. This is natural if the final RV follow up occurs on a traditionally allocated telescope like Keck as compared to the queued HET. Also, one needs multiple planets in a year; if XO’s output triples, then we can anticipate 6 planets per year, but we may choose to throttle back XO’s rate of 3-day planets in order to increase its rate of 10-day planets by pointing all cameras in the same direction.
4 1 Objectives and Significance
This proposal describes continued progress and future goals of the XO Project, one of the leading teams discovering and characterizing Transiting Extrasolar Planets, or TEPs.
1.1 Scientific Objectives: find more transiting planets 1.2 Brighter is Better The mantra “Brighter is Better” largely has been assimilated into the community studying TEPs. For studies of TEPs, brighter stars are better because they tend to allow a larger variety of possible follow-up observations. Very precise work is photon-limited almost by definition, but in detail there are trades even for this simple principle. For space-based follow up, e.g. with HST NICMOS or SST IRAC, in the stellar magnitude range 8 < V < 12, overheads associated with reading out the detector before saturation tend to make observations of stars at the fainter end (V=12) record much more than 40-times-fewer photons than those collected from stars at the brighter end (V=8). For instance, in a recent HST NICMOS proposal1, to detect water vapor in the atmosphere of a TEP, we showed that a single transit of XO-4b (V=10.8) should produce the same S/N ratio as that of HD 189733b (V=8), after one accounts for readout and buffer dump overheads, the duration of the transits, and the occultation of the Earth (for HD 189733b but not XO-4b). For space-based and ground-based observations, stars near the fainter end (10 < V < 12) tend to have more comparison stars that also are closer in angle and more similar in color to the target star than the 8 < V < 9 stars. For all stars in the range 8 < V < 12, ground-based observations in broad spectral bandpasses are scintillation limited, so the fainter stars and the brighter stars often give similar S/N ratios. Example light curves are in Figure ??. In Section XX, we describe a technique that should permit us to cancel scintillation for observations of some TEPs with the world’s largest ground-based telescopes, preparing the community for the JWST era of photon-limited2 large-aperture observations of TEPs and the scientific opportunities they provide. We need to reconcile the following leftover from an old proposal with our current belief that these stars should be scintillation limited in broad bandpasses. Also, KeplerCam (Holman, XO- 1b) claims shot noise to be dominant - maybe due to poor throughput at z band? With a 0.3-m telescope and a goal of 1 millimag photometry on V=12 stars, an exposure time of 5 minutes is required with no filter, or 15 minutes with a broadband filter such as V, R, or I. For such exposures, scintillation is less than shot noise, except for airmass ∼> 2.
1.3 Significance of Transits: more than P, m sin i, and e Greg Laughlin has noted that 2008 may be the first year that more planets are discovered by transits than by radial velocities, and Kepler has not even been launched yet.3 With hundreds
1McCullough is the P.I. of this GO proposal under review by the HST TAC, May 2008. 2Photon-limited means limited by the Poisson statistics of photons, not by scintillation or other sources of systematic noise that in principle can be removed by calibration. 3Laughlin’s blog entry for April XX, 2008 at http://XXX.
5 of planets discovered and scores of TEPs discovered, Laughlin (and others including us) have described the transition from “discovery” to “commoditization” of first the RV-discovered planets and recently the transit-discovered planets. As the discoveries mount in numbers, the community reaps the rewards of those discoveries in the follow-up characterization phase. XX add to this laundry list more possibilities and citations. It is worth noting that the only things we know about most exoplanets are the orbital periods, P, the minimum values for their masses, i.e. m sin i, and their eccentricities, e. The subset of planets that transit their stars provide many additional possibilities for characterization, first and foremost its radius, mass (not m sin i), density and “surface” gravity. Transiting exoplanets have been the subject of literally scores of programs on SST and HST. Empirical studies of exoplanet atmospheres fall into two broad categories: (1) searches for absorption of starlight by the annulus of planetary atmosphere as the planet transits in front of the star, and (2) searches for diminution of reflected starlight or planetary emission as the planet passes behind the star. In both cases, differential measurements compare flux in and out of eclipse, which typically lasts a couple of hours. Some example results are observation of atmospheric composition (Charbonneau et al. 2002; Vidal-Madjar et al. 2004), exosphere extent (Vidal-Madjar et al. 2003), and infrared emission (Charbonneau et al. 2005; Deming et al. 2005; and many more). The transit permits us to search for satellites and Saturn-like rings orbiting a transiting planet (Brown et al. 2001), although only planets with orbital periods greater than XX days are expected to have stable satellites (Barnes & OBrien 2002). Soon astronomers will measure a planet’s albedo by observing the visible light reflected from the planet before it enters secondary eclipse Comparison of planetary systems teaches us empirically how nature works, i.e. it can constrain theoretical models, but even with the few-dozen examples studied to date, the diversity of planet characteristics continues unabated. Remember that hardly any astronomers expected to find hot Jupiters at all, until 51 Peg was discovered in 1995. Then we expected hot TEPs’ orbits to have been circularized, until in 2007 TEPs with large eccentricities, HAT- P-2b (e=0.50), XO-3b (e=0.25), GJ 436b (e=0.17), and HD 17156b (e=0.XX) proved that was incorrect. Until 2007, no transiting planet was known with a mass larger than 2 Jupiters, and in the statistically “pure” RV-selected sample of 475 F, G, and K stars of Cumming et al. (2008), zero such planets had been found, transiting or not. But then HAT-P-2b, XO-3b, and WASP-XX were discovered, with masses 8, 13, and XX times that of Jupiter. We are beginning to gather enough examples to make correlation plots (e.g. Torres et al 2008) that may guide our imaginations and theories as to how these planetary systems “work.” A likely important advance to come (in 2008?) will be a multi-planet transiting system, the timing of the transits of which will permit more precise tests of orbital dynamics of these systems than have been feasible with radial velocities alone. A few questions XX. Interior to a hot Jupiter, are there planets that were sheperded by its migration? How much of the dispersion of the mass-radius relation for TEPs is due to observational error? Will it tighten up with additional careful characterization? What characteristic(s) are correlated with puffed-up planets (HD 209458, XX) that are bigger than (some) models predict? Why is there an over-density of TEPs at a period of ∼3 days? Can we update this statistic to TEPs in general compared to RV-planets? For instance, can we count papers related to TEPs and RV-planets, and make a Venn diagram showing
6 the relative sizes of each and the set in common? If that’s easily obtained by the exoplanet encyclopedia, let’s make it and then evaluate whether to include it. As a measure of the scientific importance of transits, we note that refereed papers discuss HD 209458 more (250 papers in 5 years) than the famous 51 Peg (360/200 papers in 10/5 yr). The initial discovery paper (Charbonneau et al. 2002) of HD 209458 transits has 300 citations in only 5 years. Hundreds of hours of ground-based and space-based observatory time have been dedicated to the study of HD 209458 alone. By almost any measure, planets that transit bright stars are the gold-standard of exoplanets.
1.4 Longer periods: satellites as habitable worlds If a Jovian planet, located ∼1 AU from a solar type star, has a satellite, then it could be that there are “habitable worlds” outside our solar system that are satellites, not planets, and these worlds may be detected by timing of or precise photometry of a Jovian planet’s transits. Such a satellite’s orbit is not dynamically unstable in 4.6 Gyr if its host planet’s semimajor axis is ∼> 0.3 AU, i.e. with a period ∼> 2 months for a solar-type host star (Barnes & O’Brien 2002). Adapt this sci-justification verbatim text from Fleming 2008. Due to the close proximity of hot Jupiters to their parent stars, large natural satellites are not dynamically stable over significant periods of time (Barnes & OBrien 2002). However, if Temperate Jupiters have systems of natural satellites then they could represent potentially habitable worlds outside our solar system, in addition to HZ planets. A necessary first step in studying HZ natural satellites is to observe howmanyHZ gas/ice giants exist and what their properties are like. While radial velocity surveys have already detected extrasolar planets potentially in their parent stars HZ, transiting HZ planets offer the capability of directly measuring the planets radius, albedo and chemical composition. Depending on the natural satellites orbital parameters, it is possible to measure its radius and mass, as well as estimate its structural composition. It could be that some of the first terrestrial worlds discovered in the HZ of a main-sequence star are found as natural satellites of Temperate Jupiters rather than as HZ planets. Indeed, transiting HZ natural satellites of gas and ice giants might be more easily detectable and more readily studied in the near future because the parent planets act as signposts for when and where to search.
1.5 Multiple planets: spectroscopic and photometric searches Referring to recent (out of transit) observations of XO-1 by the XO Project’s original Extended Team member Bruce Gary, Greg Laughlin blogged on Apr 28, 2008, “This photometry is potentially good enough to confirm a Neptune-sized planet in transit across a Solar-type star, which is absolutely amazing.” Professionals searched for and didn’t see the transit of GJ 436b (cite XX), which we consider support for the concept that currently-known TEP-stars may harbor second TEPs lurking at the edge of detectability. The hot Neptune GJ436b has a radius 0.374 RJ , so it would have a transit depth of 1.4 mmag for a solar-sized star. The nominal prediction for scintillation (Dravins et al XX, Eq. 10) for a 36-cm (14-inch) telescope over a 1-hour averaging time is σ = 0.10 mmag, so it
7 is indeed feasible for skilled amateurs to detect a hot Neptune passing in front of a solar- type star in a single transit (but in the real world would need confirmation by observations of additional transits). For a Neptune-sized planet, a global campaign of expert amateurs observing a specific star known to have one TEP may be a practical means to find a second, smaller TEP orbiting that same star, and/or to prove that such second planets do not transit, within a specific orbital period limit set by the duration of the campaign. The LCO-GT array of six telescope stations around the globe, when operational, will also have that capability. XO’s current post doctoral researcher, Chris Burke, has predicted that hot Earths tran- siting solar-type stars are detectable from the ground with 4-m telescopes. A limitation is the availability of 4-m telesopes for such searches; we expect a TAC will find such a search more palatable if there was some clues that a particular system has an additional planet. Potential clues could be found in RVs or TTVs, and either may provide predictions of when to look for the transits of the second planet.
1.6 Devote more spectroscopy to some Duds Because the radius of a TEP is equal to that of a late M dwarf star, RV confirmation is necessary to be confident a transiting object is a planet. There’s a curious paradox in the TEP field. Because of the historical development of the priority of the RV-technique and because of a statistical argument related to inclinations, a lower-limit on an (RV-detected) object’s mass is adequate to deem it a planet, but an upper-limit on the mass (for a TEP) is inadequate. Paradoxically, it should be the other way around. One implication of the paradox is that a small fraction of RV-discovered (and non-transiting) objects could be planets or stars on very inclined orbits (e.g. Fritz Benedict’s example from AAS in Hawaii XX). If a transiting object can be shown to have substellar mass, i.e. it does not induce a stellar-amplitude RV oscillation, but only an upper limit on the mass can be determined because the stellar lines are too weak or too broad (e.g. an early F-type star) to permit an actual mass determination, the candidate sits in the proprietary lists of XO, HAT, or WASP. XO has a few of these for which we have not (yet) devoted much spectroscopic observing time, because we have been conserving that for the easier ones.4
1.7 Better characterization: improved photometry and spectroscopy Discuss potential for characterization of the flood of TEPs. Point out our transit light curve of XO-4 with Perkins and PRISM is the best ever done from the ground, and we have the team to do studies from space as appropriate.
1.8 Additional Science Enabled it XX I think we can delete this entire section if we need the space or if it deemed unwor- thy. Our focus will be the previously discussed science of TEPs. XO’s observations enable
4HAT and TrES do too (Bakos, p.c.; Charbonneau, p.c.).
8 additional science that is beyond the scope of this proposal, but which we briefly mention here. Starspots sometimes are observable as bumps in transit light curves (e.g. ??). Nonuni- formities in the stellar radiance are a source of noise (or signal depending on one’s interests) for precise transit light curves. Starspots could affect timing of transits, especially when they are near the star’s limb. They could affect inferred planetary properties, at the next level of detail not yet explored, e.g. attempts to measure planetary oblateness by the shape of ingress and egress (e.g. cite XX) will be challenged by nonuniform irradiance at the stellar limb. Starspots especially compromise observations of non-transiting planets intended to de- tect the reflected light (albedo) from the phase-of-venus effect with orbital phase (cite XX). Starspots can mimic a non-transiting RV-discovered planet (e.g. cite XX’s interpretation of a planet orbiting the 10XX- Myr old star TW Tau could be tested with IR RV-observations: if the RV amplitude fades in the IR, the cause is star spots not a planet (Prato, Johns-Krull, et al in prep). Transits provide an empirical determination of stellar limb darkening for stars that are well characterized (because of the interest generated by their TEPs) and of a variety of stellar types. Limb darkening is an important parameter in modeling Michelson interfer- ometry of stellar diameters, so there is synergy between TEP-studies and stellar-diameter studies. Stellar structure models are poorly constrained by the handful of accurate stellar masses less than 1 Msun. Searches for transiting planets are discovering several low mass stellar companions in eclipsing binaries (Bouchy et al. 2005). Gravitational microlensing events of large amplification are the best sort, because the chance of seeing each planet in the lensed system approaches unity for such events, improving interpretations of planetary statistics from microlensing by removing a large fudge factor (for moderate and low amplification events). The analogy for transit studies is that observing with no gaps (e.g. Corot, Kepler, SWEEPS) aids statistical interpretation. Transit surveys observe large areas of sky at high cadence, just what is needed to detect rare high-amplification events. XO would have had the finest coverage of the the factor-of-XX event of Oct 2006 (cite XX) had we not finished observing that region of sky before the event! The XO Constellation (CCD and computer upgrade) permits alerting for such high amplification events in real time and the LCO-GT Maui telescope is across the street. Near Earth Asteroids show up in XO data as lines. Due to XO’s small aperture and large field of view, the sweet spot of detectability is boulder-sized objects a few meters across at the distance of geostationary satellites. In the pre-planet days of XO, we searched XO images that happened to be aimed at the edge of the the Earth’s conical shadow and found one or a few of these, distinguishable because they ingress or egress (the many-degree-long line fades out deeper into the shadow).
2 Technical Approach and Methodology
2.1 Summary of XO Operations and Results, 2002-2008 The XO survey was designed to maximize sky coverage with enough sensitivity to detect planetary transits around stars bright enough, V< 12, to allow the photon-limited studies
9 listed earlier. Proof of concept for XO is in the ApJ papers reporting the discoveries of XO-1b to XO-5b. XO is on pace to deliver what it proposed in 2005:
For the half dozen transiting hot Jupiters that we expect to discover... we can say with absolute certainty that such (follow-up HST and SST) observations are both scientifically compelling and possible, so long as we find the targets while the great observatories are still in orbit and operational. – from the abstract of the 2005 Origins proposal of McCullough et al.
XO’s design and implementation are described in McCullough et al. (2005). XO observes 54 7.2◦X7.2◦ fields-of-view each year, and it monitors ∼100,000 stars with ∼1% photometry per measurement. Each star is observed twice every 10 minutes for up to 8 hours per night (i.e. 96 separate 54-sec exposures). The 54 fields represent 2800 square degrees, or 14% of the northern sky, containing an estimated 2 TEPs brighter than V=11 and 8 brighter than V=12 (McCullough et al. 2005). The McCullough et al. (2005) estimate is 120 TEPs total on the sky orbiting stars with V < 12; so it is conservative, perhaps by a factor of 2 (cite XX), in part because transits permit discovery around stars with larger rotational velocities, Vsini, than the RV-selected samples used for the estimate, and because many of the TEPs discovered orbit larger-than-main-sequence stars, because the probability that a planet transits is proportionally larger for larger radii stars. (Even though the transit depth is proportional 2 to (rp/r∗) , i.e. is smaller for larger stars, if the period is short enough, we find the TEP anyway due to the many transits observed.) More technical details and the unique feature of the Extended Team are discussed in McCullough & Burke (2007). An irony of the latter paper is that we estimated how unlikely it would be for XO to detect Jovian-planet transits of M dwarfs, but we should’ve seen the “glass half-full” rather than “half-empty.” Soon thereafter transits of the RV-discovered planet GJ 436b were announced, and Frederic Pont and McCullough (unpublished) demonstrated that transits were visible in the pre-discovery XO data of GJ 436b! The transits would not have been detected without a priori knowledge of the planet’s ephemeris, in part because of the ∼ 1% sinusoidal variability in stellar brightness also evident in the spectra. Certainly the XO E.T. could have detected transits of GJ 436b had we been thinking of that at the time (such RV-planet follow up is the modus operandi of the transitsearch.org project). There are two advantages of a second season of observation of the same stars: it improves the visibility of transits, especially for longer orbital periods (Figure 2), and the longer time baseline permits more precise period determination, which is important for predictions of future transits and for accurately estimating the orbital phase of radial velocities.
2.2 Comparison to other Searches for TEPs As of May 6, 2008, the list maintained by Frederic Pont5 contains 33 TEPs with prefixes HD (5), GJ (1), TrES (4), XO (3), HAT (6), WASP (5), COROT (2), and OGLE (7). To appreciate the rapid growth in the discovery rate, the same list one year ago contained 17 TEPS: HD (4), GJ (0), TrES (3), XO (1), HAT (2), WASP (2), COROT (0), and OGLE (5).
5http://www.inscience.ch/transits/
10 In 2004, for stars brighter than V=12 the scorecard was 2 TEPs: HD 209458 and TrES-1. So the total number has been doubling each year, ∼ 2year−2003, which rivals Moore’s Law for transistor density on semiconductors. Like Moore’s Law, physical limits will eventually quench that growth, but not in the period of performance of this proposal! There are some caveats to that summary. WASP issued a press release on April 2, 2008 announcing 10 TEPs in preparation for publication. XO planets XO-4b and XO-5b are sched- uled to be submitted to ApJ in May 2008. HAT co-discovered TrES-3 (? check that XX)) and HD 147506b is a.k.a. HAT-P-XXb, and HAT is expected to announce many more planets in 2008. TrES is ramping down by 2009 as its team concentrates on other projects (Kepler, LCO-GT, SOFIA, EPOXI, etc). COROT concentrates on stars typically fainter than V = 12, with the goal of finding transiting planets in longer-period orbits and of smaller sizes than typically have been discovered from the ground. OGLE is expected to discontinue production because the follow up of such faint (V ∼ 15) stars requires too much time on very large tele- scopes. SWEEPS (Sahu et al 2006XX) and Lupus (Lupus-XX, cite) planets orbit very faint stars that limit potential follow up observations. Other projects have not yet announced their first crop of planet(s): e.g. Kepler (cite XX), KELT (cite XX), Pan-Planets (a sub-project of Pan-STARRS, cite XX), and HAT South (Bakos 2008 p.c.). An important “tipping point” recently occured, namely that the number of TEPs grew large enough that on any given night with a telescope in the northern hemisphere, a transit of a TEP probably can be observed.6 That is useful in proposing and planning observations with traditionally scheduled telescopes and instruments scheduled in blocks of time. The TEP distribution on the sky is given in Figure XX; this illustrates how the 5 XO planets fill in half of the northern sky (0h < RA < 12h) that would otherwise be rather sparse (with WASP-1b being the only other TEP found by a transit survey and HD 17156b and GJ 436b being found by RV surveys). That fact is not entirely coincidental: XO was designed to survey the sky widely and the longer winter nights may have skewed XO’s discoveries to those right ascensions.
2.3 How to find longer-period TEPs It is increasingly more challenging to find longer-period TEPs, because the probability of a transit equals the ratio of the star’s radius to the semimajor axis of the planet’s orbit, −2/3 rstar/a, which scales as rstar P , where P is the period. So for example, 2% of planets circling solar-type stars at a ≈ 0.25AU will transit. To find longer-period TEPs, we need ways to overcome some of the diminishing return inherent in the geometric probability of a transit with longer periods. As XO postdoc Chris Burke (2007) has worked out in detail, if one accounts for the distribution of eccentricities of giant planets discovered in RV surveys, and if one computes the net effect of a planet being more likely to transit at periastron than aperastron, then the probability of transits increases by a factor of 1.25 over the prediction for circular orbits. That’s interesting (especially for Kepler) but not large enough to significantly reduce the challenge. One way to significantly improve the ability to find the rare long-period TEPs is to reduce the number of transits required to identify a TEP candidate, which is one
6For nightly predictions, see http://www-int.stsci.edu/˜pmcc/xo/ephemerides/ .
11 of XO’s goals. A single transit would be enough, if it were detected convincingly, because subsequent RV-analysis would reveal the presence of the planet and determine its orbital ephemeris in order to predict future transits. A way to detect an individual transit convincingly is to observe it simultaneously with multiple independent cameras, so from the point of view of the backend analysis pipeline, it is detected multiple times. XO had been doing this “single- transit” detection with two cameras since 2003, when the field was young and astronomers didn’t know how much deeper transits might be than the 2% transit depth of the only TEP then known (HD 209458b), so we occasionally took spectra of stars exhibiting 2% to 4% deep transits that we generally wouldn’t take spectra of today (and the candidates all turned out to be eclipsing binary stars). Other TEP surveys, TrES and HAT, have advocated “networked” operation of S observing sites dispersed in longitude (with S > 1). The argument is that in a single observing season, the probability p(P, N, S) of witnessing N (or more) transits of any particular star by a planet of a given orbital period P is enhanced for a “network” (S > 1) compared to a single station (S = 1). That’s obviously true. However, the rate of TEPs discovered per year per camera, i.e. the efficiency of finding TEPs, is not equal to the aforementioned probability. Logically, once the probability p(P, N, S) approaches unity, further observation is unnecessary for that star and that period. Because the intrinsic distribution of hot Jupiters peaks at P=3 days, the original XO survey was designed such that p(P = 3days, N = 3transits, S = 1) ≈ 0.5, because increasing p higher for a given set of stars is less efficient (at finding TEPs per unit time per lens) than surveying new stars (new fields of view).7 Imagine two pairs of cameras: one pair is deployed at two sites and operated as a network observing the same stars, whereas the other pair is operated independently and observing distinct sets of stars. Like any probability p(P = 3days, N = 3transits, S = 2) cannot exceed unity, so if the cameras of the second pair are each operated for a season sufficiently long that p(P = 3days, N = 3transits, S = 1) ≈ 0.5, then both pairs of cameras should find the same number of TEPs for this example (simplified for discussion) of a delta function at P=3 days. The extended argument for networks states that the network permits each field to be “retired” faster than without the network, i.e. the network can stop observing a given set of stars and move on to new ones, as soon as p(P, N, S) approaches unity for whatever values of P, N, and S are appropriate. The latter argument is of limited validity unless the observing duty cycle is near unity (i.e. no gaps in the observing, such as was the strategy for the SWEEPS observations for seven continuous 24-hour-periods with HST) and the duration of a season is much greater than N P , so that fields are retired by the network rather than simply by the advance of the observing seasons. Another example may be helpful. Imagine one seeks a TEP with a period of ∼>1-month; if a transit is observed on Aug 1, 2008 then observing that star on any other night in August 2008 is futile, but after a year has passed and “randomized” the situation, then each night in August 2009 has an equal (small) chance of a transit of that TEP. The latter example often is the one that helps people understand why networks are not the panacea they sometimes are thought to
7An analogy may help. In fishing a stream, one can spend all day at one spot and try to catch every fish that inhabits that spot, or one can walk the bank and spend only so much time is prudent at each spot, so as to increase the total catch beyond that available from a particular spot.
12 be for finding TEPs. Their utility is greatest when the number of transits required for a definitive detection, N, is large. One reason XO works well with one site (S=1) is that XO can turn less-than-definitive survey detections with small N into definitive detections made by the global Extended Team’s follow-up photometry. With HAT and WASP pursuing the P=3 day planets, XO will play into its strengths by pursuing longer-period TEPs that naturally present any survey with small N. The detection of transits can be modeled by simulation (Figures XX and Table XX for characteristics similar to those for actual XO observing windows; also e.g. Borucki et al XX Vulcan paper; Beatty & Gaudi 2008; Fleming et al. 2008), but for the purposes of designing instruments and operational strategies, a Poisson model provides insight that the simulations lack and approximates well the simulations and actual experience: the probability discussed above, N−1 −λ i − e λ − Γ(N, λ) p(P, N, S) = 1 X = 1 − , (1) i=0 i! (N 1)! where the sum is the cumulative distribution function of the Poisson random variable with minimum number of transits N, and expectation value, N N f T λ = s n g n , (2) P where Ns is the number of seasons, Nn is the number of nights in a season allocated to observing a particular star, fg is the fraction of those nights that are good (not cloudy, etc), and Tn is the number of hours per night observing the star. (The Γ(N, λ) is the incomplete gamma function.) We also define a Lens Equivalent Season, LES, which is the product of the number of lenses Nl and seasons Ns allocated to a particular star,
LES = Nl Ns. (3)
For XO, we use Nn = 60 nights, fg = 2/3, and Tn = 6 hours. The latter are representative for XO, and as Figure XX demonstrates, they statistically match XO’s actual observing window well. For the data obtained already with XO, the LES = 2 × 2 = 4. Note that a network with S=2 sites providing an effective continuous observing duration of twice that of a single station, i.e. Tn(S = 2) = 2Tn(S = 1), is exactly equivalent to NS = 2 seasons from a single site in the Poisson approximation. The Poisson model overestimates the (zero) probability of finding multiple transits in a single season for periods longer than the season, i.e. P > Nn. However, the inaccuracy for probability, p ≈ 0, is not a problem for strategic planning, because those cases can be accounted for separately. Likewise, at the other extreme, p ≈ 1, we know that because of variable stars, crowding, etc guarantee the p is always less than unity regardless of the model predictions. Two useful predictions of the Poisson model, i.e. the equations above, are the shape of p(P, N, S), which rolls off more gradually for smaller N (N = the number of transits required for detection), and the longest period for which p ∼> 0.5, which we will denote Phalfmax(λ, N). To a good approximation, Phalfmax(λ, N) is inversely proportional to N, for 4 ∼> N > 1 (Figure XX).
13 2.4 How to find multi-TEP systems CMJ to investigate eccentricity-resonant-TEP trade off mentioned above. The XO and N2K (Fischer et al. 2005) surveys have complementary samples with minimal overlap. The XO survey does not have a metallicity selection effect and probes two magnitudes fainter. It is very important to have a comparison sample of stars without N2K’s selection bias, in order to understand the role of metallicity in the formation, migration, and physical characteristics of TEPs. Combining results from XO and N2K will allow us to test whether the range of radii, densities, and inferred core sizes for TEPs is linked to the metallicity of the stellar hosts. While most of our candidates have 3 or more transits observed by the survey cameras, our experience is that we correctly find real transits and simulated transits superposed on real data with only two or even a single transit observed. Our agorithms will report an incorrect period, but it will trigger if the S/N is sufficient to detect a single transit. The S/N required implies the star must be bright or small in radius (i.e. low mass), or the planet must be large (e.g. HD 209458). If our analysis triggers on only two transits, follow up photometry can be targeted at likely harmonics in order to observe a 3rd, 4th, etc transit. For example, if XO observes a transit on July 1st and 31st, then we know that XO’s Extended Team (ET) can follow up photometrically on Aug 30, but the ET can also hunt for transits on Aug 15, Aug 7.5, etc, in order to resolve which of the permitted harmonics of 30 days is the true period. If our analysis triggers on only one transit, follow up spectroscopy can measure the mass of the companion that cause the transit, either the planet or the eclipsing binary star. XO has had a few examples of that sort, but so far they have been eclipsing binary stars. Currently TEP-survey groups covet their potential discoveries. HAT and WASP, while not ready to collaborate with XO in 2008, may be in 2011, when each survey has years of observation and is facing a similar number of years of observation to reap diminishing returns for the rare 1-2 month period TEPs, but can reap them in 2011 only by collaboration. XO has thought of a mechanism to accelerate that process, because the community could and should benefit from the teams collaborating. It is mathematically possible for two groups to know whether their lists of “stars of interest” have any elements in common without sharing the lists themselves.8 Or the teams can simply share their lists with a trusted third party (HAT and TrES did that tacitly with David Latham, who provided candidate-culling services to both projects). We propose this as a way to break through to identify the win-win circumstances required for collaboration between competitive TEP-seek teams. At regular intervals, the teams are invited to submit to the trusted third party their lists of up to N stars of some interest, but not promising enough with their own internal analysis to warrant spectroscopic follow up. (These would generally be transits observed only once or twice, i.e. the longest- period TEPs). The third party identifies matches to those that submitted them, and all lists are “retired.” The “retirement” principle allows a group to choose to continue to study unmatched stars without risking a match in the future with a team that notices the same star later. 8Using “zero knowledge proofs” or bit XX.
14 2.5 Strategy for Followup As an initial requirement, planet transit searches must have enough sensitivity to detect transits with a depth of 1% and a duration of a couple of hours. But transits with these char- acteristics are also caused by other types of systems: (i) grazing stellar eclipses, (ii) eclipsing binaries diluted by one or more other star(s), and (iii) transiting M dwarfs or brown dwarfs (e.g., Torres et al. 2005). The XO system uses an efficient sequence of followup photometry and spectroscopy to distinguish planetary transits from these other types of transits. For selecting candidates, all teams have independently invented or adopted similar procedures. Our own version is described in McCullough & Burke (2005). XO benefits from unpaid volunteers, its Extended Team, for global longitude coverage for photometric follow-up (but not for the survey task). The E. T. obtains seeing-limited pho- tometry of transits at high time resolution. These followup observations: (i) confirm orbital periods, (ii) distinguish and reject eclipsing binaries that are diluted by resolved companions, (iii) reject grazing eclipses (V-shaped light curves, rather than U-shaped, or unevenly-spaced timing of eclipses from eccentric grazing orbits) and (iv) reject light curves with secondary eclipses that are too shallow to have been detected with the survey data. The ET consists of professional and talented amateur astronomers in California (2), Ari- zona, Utah, Maine, Tennessee, Belgium, Portugal, Spain, and Italy. Additional nodes in the far east will improve our longitudinal coverage in order to achieve 24/7 capability. ET is an experiment in itself, made possible by the brightness of our targets. A password-protected website hosted at XO HQ (STScI) serves as an archive of ET results until publication. We then obtain multiple spectra of the surviving candidates to identify and reject: (i) eclipsing binaries that are diluted by an unresolved third companion, (ii) M-type and brown dwarfs around stars slightly larger than the Sun, and (iii) grazing binaries that may have survived the photometric filters. XXX Need to renew this relationship. XXX Finally, we will forward to Debra Fischer stars with U-shaped transit light curves and no velocity variations above 100 m/s. Dr. Fischer has committed to following up with 3 m/s precision those candidates that pass all of the filters described above (see attached letter). We have fully implemented the comprehensive followup plan described above, though the last precision velocity step has unfortunately not yet been necessary.
2.6 Comparison to other Searches for TEPs At the present time, transit surveys show great promise for extending the very successful RV technique to large numbers of TEPs. The direct imaging approach is likely to be limited by technology to a few special cases for the foreseeable future and the microlensing method suffers an “Achilles heel” in that the events never repeat, so follow up studies are impossible. The converse of the transit survey approach, the transitsearch.org approach is to hunt for transits of planets discovered by RV, e.g. GJ 432b (P=2.XX days) and HD 17156b (P=21 days). Although small number statistics are problematic, the current yields of TEPs by HAT, WASP, and XO are 7, 5(or 15 as of the April 2 press release), and 5, while the numbers of
15 survey lenses (all Canon f/1.8 200-mm lenses) used to make those discoveries are 6, 16, and 2, so the ratio of TEPs to lenses is ∼1, ∼1, and ∼2. XX ask the IAU 253 reviewer to send you his table. One should not take the statistic too seriously, because each survey has had its own somewhat different scientific goals, programmatic schedules, and funding profiles. Also, XO’s TEP-to-lens ratio was ∼1 in 2007, and we expect HAT to equal XO’s 2:1 ratio by the time you review this proposal. We believe it is a historical fact that with this proposal, the XO team requests grant renewal with the largest number (5) of TEPs submitted to the refereed literature than any previous proposal of any team. We hope the committee agrees that the XO system should be funded for three more years now that it is operational and has triple its original, planet-proven capability.
3 Impact of Proposed Work 4 Relevance to other NASA Programs and Missions
Finding several TEPs around other bright stars will be useful in its own right, but also as a complementary set of star-planet systems to the fainter (more distant) star-planet systems targeted by the Kepler mission. The interpretation of Kepler light curves will be more sophis- ticated and grounded in solid observations if we have numerous TEPs studied around bright (solar metalicity) stars before Kepler is launched in 2008. (Presumably Kepler will not apply a metalicity selection effect to its target stars.) In earlier sections of this proposal we have detailed how TEPs have been and will continue to be high priority science targets on HST and Spitzer. We won’t repeat those here, except to note that the Spitzer observations are a historical landmark, as the first detections of light emitted by exoplanet(s). When it is ready, SOFIA will have some unique features for transits. For one, it can observe continuously for hours without gaps, whereas HST is typically occulted by the Earth for half of each 96-minute low-earth orbit. The HST light curves of HD 209458 and TrES-1 require assembling one continuous transit light curve from segments observed many days apart. The basic assumption that the light curves can be assembled because they are the same from one transit to the next is clearly invalid for TrES-1, even if it seems to work anyway. SOFIA also has infrared capability that can span the gap in years between Spitzer and JWST. There are several species in the near IR (H2O, CO, CO2, etc) that may be detectable. SOFIA instrument HIPO, High-speed Imaging Photometer for Occultations, is perhaps the premier instrument to study time-of-arrival variations of TEPs and eclipsing binary stars, especially if HST is over subscribed more than SOFIA. HIPO benefits from SOFIA’s near- absence of scintillation noise and clouds. SOFIA’s aperture is the same as HST’s so their shot noise limits should be similar. While HST presumably will be more stable than SOFIA, because of the gaps in HST’s observing, for some science the appropriate comparison will be stability on hour timescales for SOFIA compared to 3 or more days with HST. HIPO will permit much higher cadence of observations than any instrument on HST, due to readout and data transfers, except the HST’s FGSs.
16 5 Management Plan
The postdoc will assume some of McCullough’s duties listed above, and in collaboration they will perform the various screens of TEP candidates. Screens are defined are defined as follows: Screen 1 = initial search with XO cameras; Screen 2 = precision photometry by ET or with meter-class facilities; Screen 3 = km/s precision RV spectroscopy; Screen 4 = few m/s precision spectroscopy. Heasley represents the XO project at the University of Hawaii, providing in person com- munications with UH staff and management when McCullough cannot. Heasley also provides the P.I. the local perspective on his and other projects. Heasley’s contribution is critical at times when resources and time are limited, because staff might naturally tend to solve prob- lems for people they see often, whereas the P.I. is hardly ever seen in Hawaii, and the P.I.’s workday is 5 or 6 hours east of the Hawaiian work day. Heasley will assist with observations to be made with UH telescopes. Heasley and Janes will provide expert consultation to McCullough on ground based pho- tometry and calibration. Janes will act as McCullough’s deputy liason to the Extended Team, especially providing mentoring on photometric calibration, and recruiting and training new ET members. Janes will assist with observations to be made with Lowell telescopes, especially with his NSF-funded PRISM CCD photometer. Valenti will provide expert consultation to McCullough and his postdoc on high resolution spectroscopy, including operation and improvements to existing IDL code for measuring radial velocities. McCullough and Long will re-run all the XO photometry with various sized photometric apertures and for each star select that aperture that empirically gives the best photometry. They will re-run the BLS search algorithm on the improved photometry, including longer periods and second- and third-best solutions instead of simply the best-fitting transit. The latter allows a compromise between filtering too narrowly (e.g. missing transits with longer duty cycles than expected) and having to deal with too much human inspection, which we will reduce by greater automation of the grading of candidates and their inclusion in the pool for E. T. follow up. For example, we currently evaluate candidates and THEN estimate their reduce proper motion indicator (Gould & Morgan XX) but we will do better by computing that indicator for every star and automatically adjust thresholds for BLS to report a candidate accordingly. Valenti, McCullough, and his postdoc will collaborate on observing proposals to HST and Spitzer of all TEPs discovered by XO, or by other projects.
17 Roll-off Roof Shelter on Haleakala, Maui, HI
Figure 1: The XO Constellation observatory: six 200 mm f/1.8 lenses and CCDs attached to 3 German equatorial mounts, deployed under a roll-off roof at a clear and dark site, the 10,000-foot summit of Haleakala, Maui (latitude = 21◦ N). Each pair of cameras observe the same field of view simultaneously through a 0.4 µm to 0.7 µm filter. The cameras are operated in drift scan mode to produce images that are 7.2◦ wide and 63◦ tall. Subimages (left to right) centered at 10%, 30%, 50%, 70%, and 90% of the width of the CCD displayed for camera 1 (top) and camera 2 (bottom) show that the 1.8-pixel (FWHM) PSF is nearly constant across the field of view.
18 Figure 2: (Top left) The visibility of transits of a given period are indicated for observations of at least 1 transit (solid), at least 2 transits (dotted), and at least 3 transits (dashed). The plot shows the fraction that could be seen with perfect detection sensitivity, given the actual times of a single season of observation of a given star by XO. During each year, six 7◦-by-63◦ scans are observed repeatedly. The window for a given star is narrow at the beginning of the season, widens to its maximum of 8 hours per night, and then tapers off as the season closes. XO observes the same fields again one year later; that has two benefits: 1) the chance of seeing a second transit (i.e. one or more per season) is greater than 60% for even the longest periods plotted, 10 days, and 2) it facilitates scheduling of future photometric confirmation and radial velocity measurements because we measure the period precisely with the resulting ∼1-year baseline. (Other diagrams) Poisson models of transit visibility for 1, 2, and 3 seasons of observing. One site (6-hour nights, blue) and a 2-site network (12-hour nights, red) visibilities are compared as functions of log10(period) for minimum number of transits seen equal to 1 (solid), 2 (dotted), and 3 (dashed). 19 Figure 3: Transit detection of GJ 436b from pre-discovery XO data analyzed by Pont and McCullough (2007, unpublished).
20 Figure 4: Transiting Planets in the Northern Hemisphere. XO planets provide targets at right ascensions 0h < RA < 12h where there are few others currently available. This figure and associated ephemerides for each planet are maintained at http://www- int.stsci.edu/˜pmcc/xo/ephemerides/ .
21 Figure 5: A scintillation-limited transit light curve of XO-4b (top left) made with the Perkins 1.8-m in R band. For comparison, the light curve of TrES-1 made with HST ACS (top right). We know of no transit light curve made from Earth’s surface with a lower r.m.s. noise per 1-minute interval (solid black line) than this one. In 1-min averages, the r.m.s. of residuals from the best-fitting model are 5 × 10−4 at low airmass (left side) and increase to 2 × 10−3 at airmass > 2 (right side).
22 P1 P2 N1 N2 N3 N4 NXO1 NXO2 NXO3 V<12 V<13 1 2 1 1 9 36 1 0.5 0 2 4 17 16 90 360 10 10.0 3 4 8 9 8 30 120 3 1.5 2 8 16 6 8 19 76 2 0.7 0 16 32 8 8 12 48 1 0 0 32 64 6 8 8 32 1 0 0 64 128 10 8 5 20 0.5 0 0
P1-P2 is a bin of orbital periods, in days.
N1 is the number of RV-discovered planets (M>Saturn) in the bin.
N2 is a simple model of N1: flat in each (logarithmic) bin for P1>4 days.
N3 equals the number of TEPs (M>Saturn) in each bin for stars V<12, over the entire sky. N3 is normalized to the estimate (McCullough et al. 2005) of 120 hot Jupiters with periods less than 7.5 days (and excluding the shortest-period bin which was unpopulated at the time), i.e. 120 = 90+30. The ratio 90:30 comes from the N2 ratio 16:8 multiplied by the scaling of transit probability with period P to the -2/3 power, i.e. the 2:1 ratio of the 2nd and 3rd rows of N2 requires a 3:1 ratio in N3’s 2nd and 3rd rows.
N4 is the same as N3, but 4 times larger, due to the larger number of fainter stars.
NXO1 is 1/9 of N3, because XO observed 1/9 of the sky for two seasons in its first 4 years.
NXO2 is an approximate yield from NXO1 that accounts for the fraction of those that do transit that would be seen (i.e. and detected).
NXO3 is the actual TEP count to date from those first 4 years. The analysis is still improving and we still are following up candidates. So we have another ~half dozen hiding in our existing XO data.
23 6 Facilities, Equipment and Other Resources
6.1 The Robotic Camera The XO Constellation observatory is commercial-off-the-shelf (COTS) technology. It consists of six 200 mm f/1.8 lenses and CCDs attached to 3 robotic German equatorial mounts, de- ployed under a roll-off roof at a clear and dark site, the 10,000-foot summit of Haleakala, Maui (latitude = 21◦ N). The system has operated since Sep 2003 with no significant maintenance until in 2007 when both of the old CCDs failed (one is sick; the other is dead). The two oldest CCDs will be replaced with funds from year 1 of this proposed grant. The XO Constellation is operated by a rack of linux and MS Windows computers contained in a Faraday-cage room adjacent to the telescopes.
6.2 The Extended Team of Amateur Astronomers XX PRM add paragraph here.
6.3 Space Telescope Science Institute (McCullough & Valenti) The STScI currently operates the Hubble Space Telescope and will be the operations center for the James Webb Space Telescope and the data center for Kepler. It is one of the largest employers of astronomers in the world, and accordingly has one of the best technical and intellectual infrastuctures for astronomical research. Amongst the ∼100 PhDs at STScI, there are experts on optics, detectors, and instrumentation; numerical simulation and visualization; data analysis and archiving; and nearly all fields of astronomy and astrophysics. All STScI scientists get excellent (desktop) computer support, including high-end worksta- tions and software licenses (e.g. IDL) as part of their normal employment. The XO Project has used and will continue to use the STScI “Royal” XX-processor Beowulf cluster for searching for transit-like light curves.
6.4 University of Hawaii (Heasley) XO is housed within a roll-off roof building on Haleakala. Internet access is provided through the Institute for Astronomy’s Haleakala computer network. The IfA also providers technical support for the XO project. The Mees Solar Observatory, located near the Baker-Nunn building, has a small machine shop facility. A weather monitoring network exists on site too. Heasley has institutional access to the University of Hawaii’s telescopes on Mauna Kea and Haleakala.
6.5 Boston University Astronomy Department and Institute for Astrophysical Research(Janes) Revised text from Janes here. The Perkins 1.83-meter telescope near Flagstaff, Arizona is owned by Lowell Observatory and operated jointly with Boston University. The Perkins tele-
24 scope has recently been re-instrumented with an optical imager, polarimeter and spectrometer called PRISM (Janes, PI), and Mimir, an IR equivalent that Dan Clemens (BU) and Marc Buie (Lowell) constructed. These instruments are used by BU and Lowell staff and are avail- able to the astronomical community through an NSF “PREST” (Program for Research and Education with Small Telescopes) grant (see www.lowell.edu/VisitingObservers). The PRISM instrument is a re-imager, designed to reduce the f/17.5 beam of the Perkins telescope to f/4.45 with a corresponding increase in the field of view to 13.6 arc-minutes, and 0.4 arc-second pixels. PRISM also has capabilities for imaging polarimetry and low-resolution grism spectroscopy (www.bu.edu/prism). The Mimir instrument has a 10.5 arcminute field of view and operates between 1-5 microns. It also has full polarimetric and spectroscopic capa- bility (people.bu.edu/clemens). Either instrument will be available for follow up photometry of XO targets. Boston University staff share about 180 nights per year; we expect as many as 40 per year could be allocated to TEP photometry. Observing time on the Perkins is scheduled on a quarterly basis. Three key projects will take up the bulk of the BU time over the next several years. Janes’ “Old Open Cluster Study,” a photometric investigation of the old clusters in the outer parts of the galactic disk will be scheduled on about 35 nights per year over the next several years. The remaining nights (about 50) not scheduled for the key projects are available for short-term projects, including XO observing support. Janes’ Old Open Cluster study does not have any time-critical observations, so on any night when the cluster program is on the schedule and there is an XO event, it will be possible to observe the event with PRISM, weather permitting. It will also be possible to request additional single night or even half night observing runs for XO events, depending on other demands for those same nights. There are three possible observers for such situations. As part of the NSF PREST program grant, there is now a graduate student in residence at Lowell. The graduate students helps out the outside visitors and other observers, and in return gets to work on his or her thesis observations. There is also a resident postdoctoral associate who can be called on from time to time. Finally, Boston University pays part of the salary of a Lowell staff person for some of the observing.
6.6 Rice University (Johns-Krull) The NSF-funded robotic upgrade to the 0.8-m telescope at McDonald. XX More words here provided by Johns-Krull.
25 The remaining pages are leftovers from 2005 and need to be updated.
26 7 Vitae
27 Replace this page and the previous one with McCullough’s 2-page CV
28 CURRICULUM VITÆ Dr. Jeff A. Valenti
Address: Space Telescope Science Institute (STScI) 3700 San Martin Dr. Baltimore, MD 21218 1-410-338-2622 [email protected]
Education: 1994 Ph. D. UC Berkeley 1987 B. S. (Astronomy) Caltech
Employment/Research: 2005–present Associate Astronomer STScI 1999–2004 Assistant Astronomer STScI 1997–1999 Postdoctoral Fellow Kitt Peak 1994–1997 Postdoctoral Fellow JILA 1987–1994 Research Assistant UC Berkeley
Selected Bibliography (from 35 refereed and 75 conference): Valenti, J. A., & Fischer, D. A. 2005, “Spectroscopic Properties of Cool Stars (SPOCS). I. 1040 F, G, and K Dwarfs from Keck, Lick, and AAT Planet Search Programs,” ApJS, in press Fischer, D. A., & Valenti, J. A. 2005, “The Planet–Metallicity Correlation,” ApJ, 622, 1102 Valenti, J. A., & Johns–Krull, C. M. 2004, “Observations of Magnetic Fields on T Tauri Stars,” Ap&SS, 292, 619 Valenti, J. A., & Piskunov, N. E. 1996, “Spectroscopy Made Easy: A New Tool for Fitting Observations with Synthetic Spectra,” A&AS, 118, 595 Valenti, J. A., Marcy, G. W. & Basri, G. 1995, “Infrared Zeeman Analysis of Eridani,” ApJ, 439, 939 Valenti, J. A., Basri, G. & Johns, C. M. 1993, “T Tauri Stars in Blue,” AJ, 106, 2024
29 JAMES NORTON HEASLEY
MAILING ADDRESS: Institute for Astronomy 2680 Woodlawn Drive Honolulu, HI 96822 DEGREES: B.A. Honors in Physics, 1969, University of Chicago Ph.D., 1973, Astronomy, 1773, Yale University
CURRENT APPOINTMENT: Professor and Astronomer, Department of Physics and Astronomy & Institute for Astronomy, University of Hawaii, 1991-present
SELECTED PUBLICATIONS:
Hubble Space Telescope Photometry of the Metal-Rich Globular Clusters NGC 6624 and NGC 6637. 2001, Heasley, J. N., Janes, K. A., Zinn, R., Demarque, P., DaCosta, G. S., & Christian, C. Astron J. 120, 879.
Point-Spread Function Fitting Photometry. 1999, Heasley, J. N. in Precision CCD Photometry, ASP Conference Series, Vol 189, p 56.
Stellar Photometry Software. 1993, Janes, K. A. & Heasley, J. N. Pub. Astron. Soc. Pacific, 105, 527.
NGC 6293 and NGC 6333: Photometry of two Clusters in the Central Bulge of the Galaxy. 1991 Janes, K. A. & Heasley, J. N. Astron. J. 101, 2097.
Photometry of the Outer Halo Globular Clusters NGC 5024 and NGC 5053. 1991, Heasley, J. N. & Christian, C. A. Astron J. 101, 967.
GRADUATE AND POSTDOCTORAL ADVISORS: Dr. L. H. Auer (Los Alamos National Laboratory) Dr. Dimitri Mihalas (University of Illinois) Dr. Robert Milkey (Space Telescope Science Institute)
GRADUATE ADVISEES: Dr. Stephen Walton (California State Northridge) Dr. John Varsik (Big Bear Solar Observatory)
30 Replace this page with Ken Janes’ 1-page vitae
31 8 Current and Pending Support
8.1 Peter McCullough Grant: NASA NAG5-13130 Title: A Photometric Search for Jovian Planets Transiting Very Bright Stars PI : P. R. McCullough Period: 03/15/2003 - 03/15/2006 Amount: $336K
8.2 Jeff Valenti Grant: HST/GO-09093.03-A Title: Resolving Molecular Hydrogen Disks Around T Tauri Stars PI : C. M. Johns-Krull (Rice) Period: 07/01/2002 - 06/30/2005 Amount: $8K to Valenti
8.3 Ken Janes NSF PI Development of the Perkins Re-Imaging $280,819 9/01/02 - 6/30/05 System (PRISM) at Boston University for Astronomical Studies
NSF PI Collaborative Research: The Boston $185,880 1/01/05 - 12/31/07 University - Lowell Observatory Partnership: Bringing the Perkins Telescope into the 21st Century Pending:
NSF PI Old Star Clusters: Stellar Activity $217,841 5/01/05 - 4/30/08 and Galactic Structure
8.4 James Heasley HST-GO-09034.04-A "The Masses and Luminosities of Population II Stars" Period: 08/01/01 - 10/31/05 Amount: $92,226
HST-GO-10197.04-A "The Astrophysical Parameters of Very Metal-Poor Halo Binaries" Period: 02/01/05 - 01/31/07 Amount: $18,538
32 9 Letters of Commitment
9.1 Peter McCullough From [email protected] Mon May 23 10:21:24 2005 To: [email protected]
I acknowledge that I am identified as an investigator on this Origins proposal entitled, "The XO Planet Finding System." I will perform all of the duties identified for me in the proposal.
Peter McCullough
9.2 Jeff Valenti Date: Fri, 20 May 2005 13:24:05 -0400 (EDT) From: Jeff Valenti
Dear Peter,
I acknowledge that I am identified as a coinvestigator on your Origins proposal entitled, "The XO Planet Finding System." I will perform all of the duties identified for me in the proposal at no cost to the project.
Sincerely,
Jeff Valenti
9.3 Ken Janes Date: Fri, 20 May 2005 10:06:21 -0400 (EDT) Subject: Re: [Fwd: Origins letter of commitment] From: [email protected] To: "Peter McCullough"
Dear Peter,
I acknowledge that I am identified as a coinvestigator on your Origins proposal entitled, "The XO Planet Finding
33 System." I will perform all of the duties identified for me in the proposal at no cost to the project.
Sincerely,
Kenneth Janes Boston University
9.4 James Heasley Date: Fri, 20 May 2005 11:33:21 From: Jim Heasley
Dear Peter,
I acknowledge that I am identified as a coinvestigator on your Origins proposal entitled, "The XO Planet Finding System." I will perform all of the duties identified for me in the proposal at no cost to the project.
Sincerely,
James N. Heasley Professor of Astronomy
Institute for Astronomy [email protected] University of Hawaii phone: 808-956-6826 2680 Woodlawn Drive fax: 808-956-9580 Honolulu, HI 96822
34 9.5 You-Hua Chu Date: Fri, 9 May 2008 11:43:14 -0500 (CDT) From: You-Hua Chu
Dr. Peter McCullough STScI 3700 San Martin Drive Baltimore MD 21218
Dear Peter,
This e-mail certifies that the UIUC equipment you have had at Haleakala, Maui to search for planet transits will be available for your continued use for the foreseeable future, in particular for the proposed period of 2009-2012. The equipment includes 3 CCD cameras, 3 lenses, an equatorial mount, and 3 computers.
Best regards, You-Hua Chu Professor and Chair Department of Astronomy University of Illinois
35 9.6 Debra Fischer To: Peter McCullough
Dear Peter, 18 May 2005
I am pleased to email you this letter of endorsement of the XO project’s plan to supply me with candidates of transiting hot Jupiters. Our collaboration could speed the process of measuring 5 m/s precision radial velocities, in some cases by nearly a year due to the glare of the Sun.
You and I have discussed me observing fully-vetted XO candidates, i.e. those with accurate periods and phases known from photometry and with preliminary masses or upper limits from 1 km/s radial velocity measurements that you will obtain and analyze.
I appreciate your assistance obtaining follow up photometry of one of the hot Jupiters that we recently discovered as part of our N2K project. The synergy of resources and specialized talent works both ways: we find planets that you can test for transits, and you find transiting candidates for which we can measure the mass.
I wish you well, and your NASA Origins proposal for the period 2006-2008 too,
-Debra
Dr. Debra Fischer Asst Prof. San Francisco State University Dept. Physics & Astronomy 1600 Holloway San Francisco, CA 94132 415-338-1697 FAX: 415-338-2178
36 9.7 Ron Gilliland Date: Sat, 21 May 2005 17:50:17 -0400 (EDT) From: Ronald Gilliland
Dear Dr. McCullough,
When suitable transiting planet detections result from the XO program I would intend, at your invitation, to participate in proposing for follow up HST observations, and collaborate actively in the data analyses. The HST observations are invaluable for refining fundamental parameters for the planet, and in favorable cases may hold promise for exoplanet atmospheric studies.
Best regards, Ronald L. Gilliland
Astronomer, STScI 21 May 2005
9.8 Frank Summers From: Frank Summers
To Whom It May Concern,
I am writing in support of Peter McCullough’s proposal. I am the administrator of a 36 node beowulf cluster at the Space Telescope Science Institute. The capacity of this cluster includes 62 32-bit CPUs, 10 64-bit CPUs, 80 GB of aggregate RAM, and over 8 TB of aggregate disk storage.
This cluster is available to all STScI scientists who have large computational requirements. Peter’s project has been using the cluster for over two years now, and will continue to have access to this hardware and its upgrades in the future.
37 Sincerely,
Frank Summers
Space Telescope Science Institute 410-338-4749 3700 San Martin Drive 410-338-4767 (FAX) Baltimore, MD 21218 [email protected] http://terpsichore.stsci.edu/~summers/
9.9 ET member 1, Tonny Vanmunster From: "Tonny Vanmunster"
Dear Peter,
It has been a true pleasure for me cooperating with you over the past several months in the X0 project, and providing you with photmetric observations of exoplanet targets. I have been using for this one of the CCD-equipped 0.35-m f/6.3 telescopes of CBA Belgium Observatory (Landen, Belgium).
Evidently, I would like to continue participating in your interesting project.
With kind regards, Tonny Vanmunster Landen, BELGIUM
Tonny Vanmunster CBA Belgium Observatory http://www.cbabelgium.com ______PERANSO : The Final Frontier of Period Analysis Software http://www.peranso.com ______
38 9.10 ET member 2, Paul Howell From: "Paul Howell"
Hi Peter -
My name is Paul Howell and I am very excited to be a new ET member. I am also an astronomy graduate student so I am familiar with astrophysical data reduction in general and photometry in particular. I am equipped with a variety of personal scopes, filters and CCD cameras to acquire photometric data and have begun to do so for the XO project. In the past I have also collected and analyzed photometric data for academic purposes in my studies.
-P
9.11 ET member 3, Ron Bissinger From: [email protected] To: [email protected] Subject: XO Project Participation Date: Wed, 18 May 2005 18:05:43 +0000
Peter,
I just wanted to thank you for allowing me to participate in the XO Project. As a serious amateur astronomer interested in assisting professional astronomers such as yourself do real science, the XO Project gives me a great opportunity to help our efforts to find new exoplanets. The advances in off-the-shelf technology now available give many amateur astronomers such as myself the capability to perform in our backyards precision photometry that until recently required the use of major observatories.
As additional potential transiting exoplanet targets are identified under the XO Project, I look forward to continuing to provide you with photometric observations and any other assistance I can offer.
Thanks again!
39 Best Regards,
Ron Bissinger Pleasanton, CA
9.12 ET member 4, Bruce Gary Date: Mon, 16 May 2005 08:53:01 -0700 To: Peter McCullough
Dr. McCullough,
I look forward to continued collaboration with you on any exoplanet transit project that you direct. As an amateur I am able to monitor likely candidates at no cost to the taxpayer, so any NASA-funded program that you secure can be viewed as cost-effective as well as having an outreach component.
Bruce L. Gary Hereford Arizona Observatory (G05)
40 10 Budget Narrative
10.1 Personnel XX Need to update this paragraph A post doctoral research assistant (PD) will be supervised by Dr. McCullough. The PD’s duties will be to assist Dr. McCullough in selecting stars from XO data for follow up. The PD will take and/or analyze some of the photometric observations of the selected stars made primarily with the Perkins 1.83-m (Boston U). The PD will assist Dr. McCullough with follow up spectroscopy via NOAO, etc, by preparing for observing, observing, analyzing, and publishing results. Each time XO finds a TEP, the PD will assist Dr. McCullough to prepare observing programs for HST, Spitzer, etc, and to analyze those data and publish the results. It is expected that the PD will derive his salary support in the latter years of the project from such follow up programs. On-site support will be provided to our robot on Haleakala by the resident staff of the U. of Hawaii and Mees Solar Observatory. This support will include mailing disks full of data to the PI, inserting empty disks into the computers, rebooting “frozen” computers, routine network trouble shooting, keeping the roll-off roof and robotic mount operating reliably, and other small tasks that benefit from an on-site human. For example, in January 2003, the roof warped and failed open; the resident staff returned it to nominal operation by replacing the motor and coupling to the opening mechanism.
10.2 Supplies Two cameras will be purchased in year 1 to replace the two original CCD cameras purchased in 2002 with Research Corporation funds of the University of Illinois The new CCDs will permit identical operation of all three XO units (6 CCDs). Quotation 190A9 for two CCD cameras from Apogee Inc is $8,293 each, or $16,586 total, delivered 8 weeks ARO.
10.3 Supplies We require geographically isolated redundant backups of XO data. XO needs two 500 GB USB disk drives each year. At $1 per GB, 1000 GB/year costs $1000/year.
10.4 Travel Costs One round trip per year from Baltimore, MD to Maui, HI, is budgeted for the purpose of maintaining and upgrading the robotic telescope. These trips will be made by the PI and his assistant, as schedule permits. Each trip will nominally be 8 days long. The one trip per year requirement is based upon our experience since Sep 2003. One round trip per year to attend a meeting in the USA to present results (5 days duration).
41 10.5 Publication Costs Publication of 15 pages per year in professional journals is anticipated with current page charges for relevant journals (e.g. Astrophysical Journal e-manuscripts = $110/page).
10.6 Indirect Costs XX Need to update this paragraph On May 10, 2002, NASA and STScI reached agreement on the following rates which will support the Overhead and General and Administrative (G&A) costs of the Space Telescope Science Institute: 63.60% Overhead, 26.52% Full Benefits and 9.16% Partial Benefits on Direct Labor, 1.39% on Subcontractor Overhead, and 14.36% G&A on Total Costs. These rates are based upon the most cost-effective and equitable manner of administering our programs and charging the appropriate share to all projects carried out at the Space Telescope Science Institute.
42