Figure 1: An artist’s impression of a spacecraft navigating in deep space using pulsar signals. Fig. 12 in the paper.

• Title: Autonomous Spacecraft Navigation With Pulsars
• Authors: Werner Becker, Mike G. Bernhardt, Axel Jessner
• First Author’s Institution: Max-Planck-Institut für extraterrestrische Physik and the Max-Planck-Institut für Radioastronomie, Germany

How do we currently navigate spacecraft?
Current navigation systems for spacecraft use a combination of radio tracking stations on Earth, and optical data from on-board cameras as the spacecraft pass by known solar system bodies. Radio tracking from Earth has worked out for us up until now, but it has its limits: this is pretty accurate for measuring distances and velocities of the spacecraft in the direction of the line of sight from Earth, but the spacecraft’s perpendicular position can only be pinned down to within 4km per AU of distance between Earth and spacecraft, due to limits of radio resolution. This might not sound too terrible, but it corresponds to an uncertainty of ~200km at a distance of Pluto’s orbit and ~500km at the current distance of Voyager 1! In addition to that uncertainty, we also have to deal with the weakening signal strength as the spacecraft gets further away, as well as the increasing lag time (nearly 7 hours to Pluto at the furthest point of its orbit!). If we want to be able to send spacecraft beyond this distance, we’ll have to find a better solution for spacecraft navigation.

The suggestion in this paper: Use pulsar timing to set up an autonomous navigation system.

Figure 2: Artist’s conception of a rotation-powered pulsar. The pulses we see as the beam of radiation passes over us, averaged over many rotation periods, produces a very stable pulse profile. Fig. 1 in the paper.

What pulsars should you use?
Pulsars are rotating neutron stars with misaligned spin and magnetic field axes. This causes a beam of electromagnetic radiation to sweep around as the pulsar rotates, and if this beam crosses our line of sight, we see a series of extremely regular pulses from this object.

There are three main kinds of pulsars: accretion-powered pulsars (pulsars in a binary system with a donor star), magnetars (isolated neutron stars with extreme magnetic fields), and rotation-powered pulsars (pulsars that radiate away rotational energy as they spin down). Of these, rotation-powered pulsars (depicted in Figure 2) are the only ones that are both extremely steady and have well-understood long-term timing behavior (see these astrobites for more details). We know of about 2200 rotation-powered pulsars today, and many of them have been regularly timed with very high precision (at the level of 10^-15 accuracy), making these ideal candidates for use in a navigation system.

So how do you build a navigation system around a bunch of pulsars?
Navigation using pulsars centers around recording the arrival times of pulses and comparing them to predicted arrival times based on assumptions about where you are and how fast you’re going. If your assumptions about your speed and location are correct, then the pulses will arrive on time and everything’s great. If your assumptions are wrong, the pulses will arrive out of phase with what you predicted — which means you now need to tweak your position or velocity and check again. You then iterate through this process again and again until the pulses arrive in phase with the prediction.

Figure 3: Schematic illustrating the procedure for determining position and velocity using photons from pulsars. Fig. 5 in the paper.

Figure 3 describes this process in more detail. In (1) position and velocity are assumed based on the planned orbit parameters of the spacecraft. In (2) arrival times of photons from the pulsar are recorded. In (3) those arrival times are corrected by removing dispersion effects from the interstellar medium and correcting for the position and proper motion of the observatory. These corrected times are then coherently folded into a pulse profile in (4), which allow for extraction of a pulse-arrival-time (TOA) measurement in (5). In (6) this is compared to the predicted TOA for the assumed position and velocity. If there’s a difference, the position and velocity of the spacecraft are adjusted in (7) and the circle gets repeated. If there’s no difference, then the assumed position and velocity for the spacecraft are correct.

This process has to be done using at least three different pulsars in order to set a three-dimensional position, and a fourth pulsar is needed if on-board clock calibration is necessary. This system works in much the same way as hand-held GPS units do — the difference, of course, being that the spacecraft receives its signals from pulsars instead of from satellites orbiting the Earth!

How feasible is this plan?
This paper has a pretty great discussion of the different types of radio and X-ray detectors that could be used to pick up pulsar signals — I recommend checking it out for yourself via the title link at the top of the post if you’re interested. The extremely brief summary, however, is: a spacecraft could navigate autonomously using pulsars if it had either 150 m² of radio antenna area (not super feasible), or else compact light-weighted X-ray telescopes and detectors (feasible; these are currently developed for the next generation of X-ray observatories). This technology would allow for navigation accuracy of ±5 km, and that error would improve as we better-measure pulse profiles of known pulsars in the future.

In short, the authors argue that this navigation technique is feasible and would provide significant improvement over how we currently steer spacecraft at large distances. Folks, I think the future is here.

TITLE: Supernovae in the Central Parsec: A Mechanism for Producing Spatially Anisotropic Hypervelocity Stars

AUTHORS: Kastytis Zubovas, Graham A. Wynn, Alessia Gualandris

AUTHORS’ INSTITUTION: Theoretical Astrophysics Group, University of Leicester

Hypervelocity Stars
In 2005, Brown et al. discovered a star with a radial velocity of ~700 km/s, which is more than 3 times the Solar velocity! This star is moving so quickly that its velocity is high enough to escape the Milky Way. The existence of such stars, deemed hypervelocity stars (HVSs), was predicted almost 20 years earlier by Hills (1988).

The so-called Hills mechanism ejects stars at high speeds from the center of the Galaxy after a binary stellar system gravitationally interacts with the supermassive black hole at the center of the Galaxy. In such a three-body interaction, one star can be ejected at very high speeds while the other remains in the central region of the Galaxy on a highly eccentric orbit. A number of highly eccentric short-period stars are observed in the Galactic center, which suggests associated HVSs may exist. Since the initial discovery of an HVS in 2005, many more have been discovered (see this astrobite). A small warning: the exact definition of HVSs can vary throughout the literature. In this post, stars traveling away from the Galactic center with velocities high enough to have become unbound from the central black hole are referred to as HVSs.

While it is very likely that the Hills Mechanism does create HVSs, it is unclear whether all HVSs are created via this method. Curious if there is a supplemental method to produce HVSs, the authors of today’s paper examine the production of HVSs via supernovae explosions.

How many HVSs are produced by supernovae?
Zubovas et al. examine the production rate and spatial distribution of HVSs from the ejection of a binary companion during a supernovae explosion. Most stars, especially most large stars, are part of a binary system. If one of the stars is massive enough to undergo a core collapse supernovae, the ejection of its envelope will push its companion outward. In some cases, this outward force will be strong enough not only to unbind the binary, but to send the second star whizzing out of the Galaxy.

This study expands upon previous studies of supernovae production of HVSs by performing a detailed Monte Carlo analysis. The Monte Carlo analysis determines the production rate and spatial distribution of HVSs by randomly sampling from predicted initial distributions and applying expected physical models. The initial parameters include inputs such as the expected number and spatial orientation of binary systems in the Galactic center, the expected spatial separation of the stars within those systems, and how much of the ejecta energy is imparted onto the companion star.

While the initial parameter distributions are physically motivated, it is important to note that these are underlying assumptions which strongly affect the results. In a simple example, if the binary fraction is actually 2 times lower than the input assumption, a Monte Carlo simulation will predict twice as many HVSs as there actually are. The physical models of supernovae explosions are also rather uncertain, which could strongly affect the Monte Carlo predictions.

Results

Zubovas et al. figure 2a. This figure shows a cumulative distribution of the velocities for stars ejected from the system. Only a small fraction of ejected stars have velocities high enough to escape the Galactic potential (~400 km/s)

The authors find that more than 93% of the time, supernovae explosions disrupt the binary system, ejecting the secondary; however, most ejected secondary stars in the Galactic center remain bound to the central supermassive black hole. Based on different initial model parameters, the ejection rate of binary companions from the Galactic center could be between ~ 4 – 25%. This corresponds to ejecting a star around once every 4 – 22 million years.

Zubovas et al. find that over 100 million years the highest speed HVS produced via supernovae explosions is ~ 500 – 700 km/s, indicating that this ejection method cannot explain the fastest known HVSs (~750 km/s). The predicted spatial distribution of HVSs formed by supernovae, like the observed population of HVSs, is not spherically symmetric. This is because the initial simulated population was anisotropic, and this quantity is preserved, which is also true for the Hills Mechanism. Overall, the predicted production rate of HVSs via supernovae is comparable to the production rate predicted for the Hills mechanism. This suggests the observed population of HVSs may have two progenitor populations.

Title: A Posteriori Transit Probabilities
Authors: Daniel J. Stevens and B. Scott Gaudi
Institution: Department of Astronomy, The Ohio State University

Bonus Video: OSU Astronomy Coffee Brief

Transiting exoplanets are the best! Well, I think so, anyways. A transit is the dip in light we see when a planet passes between us and its host star. Transits turn exoplanets into real worlds—we can characterize their atmospheres, bulk compositions, and much else. That’s why little dips in light curves set my heart aflutter. But discovering a transiting exoplanet requires a bit of luck. After all, we can’t observe exoplanetary systems from every direction; we’re stuck on Earth or close to it, so an exoplanet’s orbit needs to be aligned just so for us to view a transit.

How do we find transiting exoplanets? There are two general methods. First, we can look at a plethora of stars at once. The Kepler mission (RIP) observed >150,000 stars continuously for four years. Because Kepler looked at so many stars, it found thousands of transiting planetary candidates. Follow-up observations are difficult, however, because the Kepler stars are relatively faint. Second, we can target planets discovered using radial velocity (RV), hoping to see a transit. These RV planets are scattered all around the sky, so we can’t observe all target stars simultaneously using a single telescope like Kepler—although a future mission (TESS) should change this! Most RV planets won’t transit, but we can extract a huge amount of information from those that do.

What’s the probability that an RV planet will transit? As they (should) teach in kindergarden, a simple estimate is that the probability of a transit is equal to the radius of the host star divided by the semi-major axis of the companion’s orbit. More compactly, P = R/a. This is because the inclination of planetary orbits should be randomly distributed from our point of view. We fit the RV data to find the semi-major axis, a. We can estimate the stellar radius, R, from the stellar spectra we used to measure the RV. Ergo, we know precisely the probability of a transit and thus how many stars we need to target to find a transiting exoplanets… right? Actually, no. As Stevens and Gaudi discuss in this paper, it’s more complicated than that.

The orientation of planetary orbits, in general, should be random. But not all orientations for planets that we detect are equally likely. RV data can’t constrain the mass of an exoplanet independently from its orbital inclination. We only measure a combination of the two, called the “minimum mass.” In reality, planetary masses aren’t distributed evenly; the abundances of super-Earths, Neptunes, and Jupiters are different. So, certain planetary masses are more likely than others and thus our RV planets are more likely to be in some orientations than others. The authors use Bayesian statistics to examine the effects of possible mass distributions on the probabilities that known RV planets will transit. They conclude that exoplanets less massive than Jupiter have up to a ~20% better chance of transiting than previously assumed. This isn’t a huge effect, but it should buoy our optimism for finding transiting exoplanets among RV targets.

The Problem

Our problem with the simple formula (P = R/a) is that it ignores an additional piece of information. Specifically, we’ve also measured the “minimum mass” of the planet with RV. This minimum mass is equal to the actual mass of the companion times the sine of its orbital inclination. Think of it this way: RV measures the back-and-forth motion of the star that we see (in the radial direction). A bigger companion will exert a bigger pull on the star, but the component of motion in the radial direction depends on the orientation of the object. For example, a companion that orbits exactly in the plane of the sky won’t cause any movement in the radial direction no matter how massive it is. At the other extreme, an orbital inclination of 90° corresponds to a system in which the orbit is precisely aligned with our line of sight—a transiting system! The mass of a transiting exoplanet is approximately equal to the measured minimum mass. For misaligned orbits, the actual mass is larger than the minimum mass.

Let’s consider a simple example. Imagine that one Earth-mass was the strict lower limit to the mass distribution; no less-massive exoplanets exist. Now, say that you discovered a RV planet with a minimum mass of 0.1 Earth-masses. You’d know the real mass must be larger than this minimum mass, so the planetary orbit must be misaligned. The probability that this exoplanet will transit is identically zero, regardless of the R and a of the system. Our a priori knowledge of the mass distribution directly affects our knowledge of the orbital alignment.

Unfortunately, we don’t know the real distribution of exoplanetary masses. In fact, that’s currently one of the major goals of exoplanet science! But that won’t stop us from speculating. Theorists run simulations of planet formation that predict different mass distributions. These distributions disagree in many respects and won’t be observationally verified for a while. However, most people believe that small planets are intrinsically more abundant than large ones, which is consistent with the Kepler results.

Figure 1: Comparison of the actual mass distribution (filled histogram) to the distribution of minimum masses that we’d observe with a RV survey (empty histogram). If we took the minimum mass distribution as the actual mass distribution, we’d be very wrong. Vertical axis is the number of objects; horizontal axis is companion mass in units of Earth’s mass. (Stevens and Gaudi 2013)

The Solution

At this point, you might be confused. How do we deal with all these nested probabilities? How do we really know what’s really real, in reality? Stevens and Gaudi appeal to Bayesian statistics, a theoretical framework that’s becoming increasingly popular in astrophysics. Bayes’ theorem, simply, states that the probability that a model is true given the data (the posterior) is proportional to the probability that the model would produce the data (the likelihood) times the prior probability that the model is true (the prior). If you understand the Bayesian lingo of posteriors and priors, you’ll get the gist of their paper, even if the algebra remains opaque.

Let’s translate our problem into Bayesian-speak. We care about the actual mass of the planet—that’s the posterior. We know the likelihood that a companion of a given mass will be aligned in such a way to produce the measured minimum mass (P = R/a). We need to think carefully about the prior probability that a given mass distribution exists in reality.

Figure 1 shows a sample mass distribution, assembled from simulations of the formation of planet-sized companions and from studies of the occurrence rate of stellar companions around nearby stars. Again, this is probably wrong in detail; the important point is that it’s not flat. Fig. 1 also shows the resulting distribution of minimum masses as measured with an RV survey on Earth. As you can see, the correspondence between the two distributions depends on the slope of the mass distribution. Where the abundance is relatively insensitive to mass (in the Jupiter zone), the observations basically match reality. But many minimum masses will be measured in a range where no real objects exist (brown dwarfs). Where the mass distribution is sloped, the distributions of the real masses and minimum masses are different—not dramatically so, but enough to make a difference when planning follow-up campaigns to target RV candidates.

The authors then searched the Exoplanets Orbit Database for planets discovered using RV. They computed the posterior probability that each planet would transit, given an underlying mass distribution. They compared these probabilities to those predicted from the simple R/a scaling. They tried two different distributions from the literature, both of which yielded similar results. In particular, planets are more likely to transit than the simple scaling suggests. As seen in Figure 2, a number of planets with relatively large orbital periods have probabilities of transits >10%, even though the simple scaling predicts <1% chances. In other words, a small minimum mass is more likely to be a small planet near-transit than a massive super-Jupiter on a very inclined orbit, since small planets are intrinsically more numerous than large ones. This result should make us more optimistic about conducting surveys of RV targets. Perhaps most importantly, however, this paper highlights the importance of using all your available information to plan observations!

Figure 2: Transit probabilities for RV planets in the exoplanets.org database. Bluer colors indicate more massive planets. Plus signs represent planets orbiting stars with very large radii. The solid grey line is the simple P = R/a scaling and the horizontal dotted line represents planets with a >10% chance of transiting. The simple scaling isn’t horrible, in general, but actual transit probabilities are probably a higher than predicted. (Stevens and Gaudi 2013)

You have probably heard that the Kepler Space Telescope, which has discovered various surprising planetary systems as well as small planets in the habitable zones of their stars, suffered a major technical failure last week.  According to the Kepler Mission Manager Update, engineers determined that reaction wheel #4 failed to restart after the spacecraft entered a self-preserving “safe mode.”  In this post, I will discuss what happened to the spacecraft and how this will affect our ability to find Earth-sized planets.  The short story is that the space telescope probably won’t detect any new Earth-sized planets, but we can continue discovering such planets in the wealth of existing data.

What are the reaction wheels anyway, and why does Kepler need them?

One of Kepler’s reaction wheels, manufactured by Goodrich Corporation.  The reaction wheels are gyroscopes that keep the Kepler spacecraft pointing in a fixed direction.  The telescope, which has four reaction wheels, requires three to be operational at a given time to maintain its stable pointing.  Reaction wheel #2 failed in July 2012.  With only two operational reaction wheels, Kepler can no longer point with the accuracy required to find Earth-sized planets.

The reaction wheels, capable of spinning at 4,000 rpm, generate the angular momentum necessary to control and stabilize the telescope pointing.  They are a mechanical type of gyroscope manufactured by Goodrich Corporation and contracted for the Kepler Space Telescope by Ball Aerospace.

The detection of Earth-sized planets requires extremely precise pointing of the telescope.  This is because the transit of an Earth-sized planet across a sun-sized star results in a change in the star’s brightness of 1/10,000 (or 100 parts per million).  Kepler was designed to detect changes in brightness of 20 parts per million, which meant it was capable of finding Earth-sized planets and smaller.  To save on bandwidth, Kepler only downlinks data from the pixels associated with 156,000 target stars out of the millions of stars in the Kepler field.  Data from an “aperture” of pixels around each target star are downlinked to Earth, and computer programs on Earth measure the brightness of the star based on the light that hit the pixels in the aperture.  If the telescope pointing is not good enough to keep the target stars in their respective apertures on the pixels, it is impossible to measure the brightness of those stars with a precision of 20 parts per million.

Can the pointing of the spacecraft be restored?

Not to the precision required to detect changes in brightness of 20 parts per million.  The Mission Manager Update offers the possibility of using the thrusters to help the telescope maintain its pointing, but it seems unrealistic to expect a crude thruster to achieve the precision of a gyroscope.  Although it might be possible to point Kepler well enough to do some new science, it seems unlikely that Kepler will be able to continue discovering Earth-sized planets.

Light curve of several “quarters” of data of the Kepler target star KID 6867155. A planet candidate has been identified from transits in the light curve, but more follow-up is needed to confirm the planet. Data from the NexSci Exoplanet Archive.

Can we still find Earth-sized planets (in the habitable zone and elsewhere)?

Yes!  Kepler has already given us a plethora of data that will feed hungry scientists for years to come.  Buried within the 156,000 light curves of stars spanning four years are many of the tell-tale “dips” indicating the transits of planets–including small ones.  The current challenge is to process the data in a manner that teases these tiny signals out of the noise and unwanted signals of stellar variability and spacecraft systematics, much like finding a needle in a haystack.  The game of detecting Earth-sized planets has moved from the telescope to the computer, and I invite all of you enthusiastic data processors to take up the challenge!  You can find and download the light curves of the Kepler stars from the NexSci Exoplanet Archive.

Other types of observational follow-up present major opportunities for confirming and characterizing Earth-sized planets given the existing Kepler data.  Radial velocity measurements of host stars, stellar spectra, adaptive optics imaging, and speckle interferometry help rule out false positive scenarios and better characterize the planets and their stars.  Additionally, transit timing variations in the Kepler data can help characterize the architectures multi-planet systems.  Describing each of these follow-up methods is outside the scope of this article, but I encourage you to leave a comment if you are curious about them.

What are the major repercussions of having only four years of data instead of eight?

The “extended” Kepler Mission was supposed to run for eight years.  (The original mission design was for 3.5 years, and Kepler exceeded this!)  Eight years would have yielded twice as much data on Earth-sized planets, resulting in a square-root-of-two improvement in our ability to extract their transit signals from the noise.

The length of the mission corresponds to how “far-out” we can probe stellar systems.  For instance, since Earth has a one-year orbital period, we would expect to see three transits of an Earth-analog (enough to identify it as a planet candidate) in three years. However, a transiting Mars-analog (with an orbital period of 1.88 years) would transit two or three times in four years, but would certainly transit three times in eight years.  So the four-year Kepler Mission can find and characterize the occurrence of exoplanets out to the orbital period of Earth, but an eight-year mission could have probed as far out as the orbital period of Mars.

• Title: Herschel Observations of Gas and Dust in the Unusual 49 Ceti Debris Disk
• Authors: A. Roberge, I. Kamp, B. Montesinos, W. R. F. Dent, G. Meeus, J. K. Donaldson, J. Olofsson, A. Moor, J.-C. Augereau, C. Howard, C. Eiroa, W.-F. Thi, D. R. Ardila, G. Sandell, P. Woitke
• First Author’s Institution: NASA Goddard Space Flight Center

Background

Young stars are surrounded by a protoplanetary disk from which planets form. Over time, the gas in the disk dissipates, leaving behind the solid material (planets, dust, and everything in between) in a debris disk. A star is typically 10 million years (Myr) old when the transition from the protoplanetary disk phase to the debris disk phase takes place — although this age varies from star to star.

The star called 49 Ceti does not fit nicely into this picture. Previous observations have found two dust belts around 49 Ceti, an inner warm component and an outer cold component — a pattern common in the debris disk phase. However, carbon monoxide (CO) gas has also been detected in this disk, a feature usually only found in protoplanetary disks. Estimates of 49 Ceti’s age range from 9 to 61 Myr, with the most recent measurement at 40 Myr (in general, determining the age of a star can be difficult business). One explanation for these discrepancies is that 49 Ceti has an unusually long-lived protoplanetary disk. Another theory is that the CO gas is not primordial, but is being continuously generated from evaporating comets or colliding planetary material.

To get a handle on 49 Ceti’s true nature, the authors of this paper (including fellow Astrobites writer Jessica Donaldson) studied it in the far-IR and sub-mm (wavelengths where debris disk dust emission is typically strongest) with the Herschel Space Observatory. They obtained photometry to model the spectral energy distribution (SED) of the disk, one resolved image to constrain the size of the disk, and spectra to search for emission lines from gas.

Photometry and SED

The authors image 49 Ceti at 70 and 160 microns with Herschel’s Photodetecting Array Camera and Spectrometer (PACS) instrument and at 250, 350, and 500 microns with the Spectral and Photometric Imaging Receiver (SPIRE) instrument.  All but one of the images were unresolved spatially, meaning the shape and size of the image are not determined by the true shape and size of the disk, but by the diffraction properties of the telescope. So although the shape and size of the disk itself could not be determined from these data, the total amount of light at each of these wavelengths (the photometry) could be measured.

The authors then create an SED of 49 Ceti using this new photometry plus data from the literature at wavelengths both longer and shorter than their Herschel measurements. The SED is shown in Figure 1. They fit the SED with a model consisting of the star plus one or two blackbody emission profiles representing the dust belts in thermal equilibrium, with temperatures set by their distance from the star. The blackbody curves are modified to suppress emission at very long wavelengths because grains emit radiation inefficiently at wavelengths much larger than their own size. The authors find that one blackbody cannot fit data at all wavelengths consistently, but two blackbodies (at temperaures of 175 and 62 K) fit the data nicely.

Based on these temperatures, the authors estimate that the rings are located at 11 and 84 AU from the star, but these distances are only lower estimates because realistic dust grains can be hotter than their equilibrium temperature, and thus seem closer to the star than they actually are. Nevertheless, this confirms that the dust around 49 Ceti is confined to two discrete rings, a feature of debris — not protoplanetary — disks.

Figure 1. The SED of 49 Ceti. The new data from Herschel are shown along with previous data from the visible to the millimeter regime. The emission from the star and the best-fit two-blackbody model are plotted with the data. From Figure 4 of the paper.

Resolved 70 Micron Image

The image at 70 microns with PACS was spatially resolved. Resolved images are very useful because they allow a disk’s true size to be determined; an SED can only estimate the minimum disk size (as discussed in the previous section).

To verify that their image is truly resolved, the authors also observe the star Alpha Boo, which does not have a disk, at 70 microns to see how the telescope diffracts light from an unresolved source (this is known as the telescope’s point spread function, or PSF). The image of 49 Ceti is shown in the left panel of Figure 2, and it is clearly more extended than the image of Alpha Boo, shown in the insert. To undo the effect of the telescope’s diffraction, the authors deconvolve their image of 49 Ceti using the PSF measured from Alpha Boo, obtaining a truer image of 49 Ceti’s disk, shown in the right panel of Figure 2.

This image traces the cold outer dust component of 49 Ceti’s disk. They find that the disk extends to 200 AU, larger than the estimate from the SED. Previous studies found that the CO gas in this disk was also coming from 200 AU, suggesting that the gas may be arising from the same location as the solid material.

Figure 2. The left panel shows the resolved 70 micron PACS image of the 49 Ceti disk. The PSF observation of Alpah Boo is shown in the insert. The right panel shows 49 Ceti’s disk after deconvolving the image with the PSF. From Figure 1 of the paper.

Gas Emission Features 

The authors search for specific gases in 49 Ceti’s disk by taking spectra with PACS at wavelengths where these gases emit. Specifically, the authors search for CO, H2O, DCO+, O I, and C II. (The roman numerals refer to the ionization state of the atomic gas, I means the atom is neutral, II means the atom has lost one electron).

The only gas they detect is C II, emitting at 158 microns. Protoplanetary disks commonly show emission from O I at 63 microns, so the fact that this gas was not detected in 49 Ceti suggests that the gas in this disk is not simply leftover from the protoplanetary disk phase. The detection of the ionized carbon and the non-detection of the oxygen are shown in Figure 3. 49 Ceti is one of only two known disks in which C II is detected while O I is not (the other is around the star HD 32297).

Figure 3. PACS spectra of the O I emission line (left panel, not detected) and C II line (right panel, detected). From Figure 2 of the paper.

Modelling the Disk

The authors create physical models of the disk in an attempt to explain their observations, and they start with a model that assumes 49 Ceti has a long-lived protoplanetary disk. They use a computer code that includes various heating and cooling mechanisms for dust and gas in the disk, as well as chemical and physical reactions involving dust and numerous gas species. Despite tweaking the model’s many parameters, the authors could not match their data, most importantly the presence of of C II and CO (known from the literature) and the absence of O I.

The authors are not able to create a model of this same complexity to test the second scenario — that the gas is being generated from comets — because not enough is know about all the sources of potential carbon and oxygen gas. However, by estimating the rate at which CO is destroyed in the disk (based on the expected flux of molecule-breaking ultraviolet radiation), they use the known amount of CO to determine how fast new CO is being produced. At this rate, a large comet would be depleted in 0.4 to 32 Myr. So this explanation is plausible, given 49 Ceti’s range of possible ages.

Conclusions

So what is 49 Ceti? Two theories existed originally: it is either a long-lived protoplanetary disk or a gas-generating debris disk. These new Herschel observations point strongly towards the second interpretation. If gas truly comes from evaporating comets or colliding planetary material, these observations could be used to learn more about the composition of pre-planet material and the nature of the planet-formation process.

The Herschel mission has recently come to an end (the all-important coolant is depleted), but the work of analyzing and interpreting the data will continue, so expect new results from Herschel to keep hitting astro-ph. And while Herschel has passed, astronomers are looking to ALMA and JWST (in the sub-mm and infrared, respectively) to further our understanding of circumstellar disks and the formation of planets.

Hi all, and welcome to the return of the undergrad research posts! For those who don’t remember this series: this is where we feature the research that you’re doing. If you’ve missed the previous installments, you can find them under the “Undergraduate Research” category here.

What does this series mean for you? We want to hear from you! Whether you’ve done an REU project, you’re working on your senior thesis, or you’ve recently started a research project in between homework sets — if you’re an undergrad doing research, we’d love to hear about it.

You can share what you’re doing by clicking on the “Your Research” tab above (or by clicking here) and using the form provided to submit a brief (fewer than 200 words) write-up of your work. The target audience is one familiar with astrophysics but not necessarily your specific subfield, so write clearly and try to avoid jargon. Feel free to also include either a visual regarding your research or else a photo of yourself!

We look forward to hearing from you!

************

Halston Lim and Jason Liang
Halston and Jason did this work jointly at the North Carolina School of Science and Mathematics.

Neutrinos, fundamental particles of the Standard Model of particle physics, can provide unique information about the internal processes of opaque high-energy astrophysical events. The ability of neutrinos to travel vast distances through matter is a crucial advantage of neutrino astronomy over optical astronomy. We demonstrated how neutrinos can be used to study the properties of failed supernovae (fSN) and black hole-neutron star mergers (BHNSM), thus providing a valuable contribution to neutrino astronomy. fSN neutrino detection would result in the first observation of black hole formation, while neutrinos from BHNSM could be used to determine if BHNSM are progenitors of short-period gamma-ray bursts, some of the most energetic events in the universe. By calculating the observed neutrino signal in various current and proposed detectors, we determined the detectability of fSN and BHNSM and demonstrated how the observed neutrino signal can provide information about the temperature and average energy of the neutrinos at the source. We also showed how these emission characteristics can then provide further information about the production of heavy metals in fSN and BHNSM. Our results confirm that neutrino observations of galactic fSN and BHNSM are feasible and provide fundamental groundwork for future research on fSN and BHNSM.

Andrew Emerick
Andrew is a graduating senior at the University of Minnesota. He worked on this project for his Honors thesis under Dr. Tom Jones and Dr. David Porter, using the resources of the Minnesota Supercomputing Institute. Andrew will be entering graduate school this fall at Columbia University, pursuing a doctorate in Astronomy with an intended focus in computational astrophysics.

Galaxy clusters are the largest gravitationally bound objects in the Universe, containing hundreds to thousands of individual galaxies. A majority of the baryonic matter in a cluster is contained within the intracluster medium (ICM): a hot, diffuse plasma that is interspersed throughout the galaxy cluster. The ICM is host to many phenomena, some of which can be used as key diagnostics, such as its often strong X-ray and radio emission. By studying the radio emission, we know that the ICM contains weak, cluster wide magnetic fields, but we do not understand well where they came from, or how they grew to the strength that is observed. One means to study the problem is to simulate the detailed microphysics of the interactions between the magnetic field and the “weather” of the ICM. We study the evolution of a weak, non-uniform magnetic field in a turbulent plasma, focusing on the details as to how turbulence amplifies a magnetic field. We focus primarily on the early evolution, and concern ourselves with how various magnetic field conditions can affect how the magnetic field grows over time, while fixing the nature of the turbulence. This study provides insight which can improve the accuracy of cosmological scale models of galaxy clusters. In addition, we know at some point information of the magnetic field conditions will be erased in the course of the ICM’s evolution. This study serves to help pinpoint exactly when that occurs, and thus if it could be possible to extract that information from potential observations.

************

Many thanks to Halston, Jason and Andrew, as well as everyone else who has recently submitted contributions! Look for more undergraduate research posts in the future — this series will continue once a month.

Title: Galaxy pairs in the Sloan Digital Sky Survey – VI. The orbital extent of enhanced star formation in interacting galaxies

Authors: David R. Patton, Paul Torrey, Sara L. Ellison, J. Trevor Mendel, and Jillian M. Scudder

First author’s institution: Department of Physics and Astronomy, Trent University, Canada

Ok, this is no big surprise: enviroment affects star formation in galaxies. Observations have long shown that the star formation rate (SFR) is strongly enhanced when two galaxies merge or simply interact, with strongest enhancements found in the closest galaxy pairs, such as coalescing galaxies, or systems observed near to the first pericentre passage. Enhancements in star formation result in bluer colours and lower metallicities, i.e. characteristic features of young stellar populations, and spectacular objects such as luminous infrared galaxies.

However, a question is still open, as you can guess from the title of today’s astrobite: what is the orbital extent of enhanced star formation in interacting galaxies? At which projected separation of the two galaxies does it disappear? This Letter aims at investigating the enhancement of star formation as a function of the separation in galaxy pairs. The issue is addressed in two complementary ways: from an observational perspective, analyzing galaxy pairs from the Sloan Digital Sky Survey (SDSS), and from a theoretical perspective, studying the outputs of numerical simulations of galaxy mergers.

First, a large sample of ~600,000 galaxies from the SDSS is considered, which have secure spectroscopic redshift between 0.02 and 0.2, and total stellar mass estimated from photometry. For each galaxy, the closest neighbour is singled out, by requiring that it has 1) the smallest projected separation from the galaxy, 2) a rest-frame relative velocity lower than 1000 km/s, and 3) a stellar mass which is not excessively different (a factor of 10) from that of the galaxy.

Then, based on previous measurements of the SFR (see the catalogue in Brinchmann et al 2004), only star-forming galaxies are selected from the sample, without any special requirement on the SFR of their neighbours. In this way, also “mixed” galaxy pairs are included in the resulting sample, which contains ~211,000 star forming galaxies. For each of these galaxies, the authors determine a statistical “control sample” which matches each galaxy in both physical properties (stellar mass, redshift) and environment (local density, isolation), but does not necessarily contain star forming galaxies. The details of the procedure adopted to identify such control samples are deferred to a subsequent paper.

Figure 1 (from Patton et al 2013). Mean SFR enhancement (top panel) and mean SFR (bottom panel) versus projected separation of galaxy pairs. The error bars are the standard error in the mean. Blue is for galaxy pairs from SDSS; red is for their statistical control samples. The dashed horizontal line represents zero enhancement of star formation.

The bottom panel of Figure 1 shows, as a function of projected distance, the mean SFR of all the paired galaxies (blue) and of their statistical control samples (red). The ratio of these two quantities, which is defined as the “enhancement in star formation” is plotted in the top panel, where the inset plot shows its behaviour at even larger values of the projected separation. This figure nicely shows that star formation is enhanced in interacting galaxies, that such enhancement is stronger at the smallest separations, especially less than 20 kpc, and finally that the enhancement in SFR extends to larger separations than what was previously thought, being visible out to projected separations of ~ 150 kpc. In particular, it is found that the 66% of the enhanced star formation in galaxy pairs occurs at separations greater than 30 kpc.

Takeaway message: an enhancement in star formation is not only limited to strongly interacting galaxies with a very close companion, but also to wide galaxy pairs.

Now, are these findings consistent with the predictions from numerical simulations of interacting galaxies? In order to answer to this question, the authors investigate a suite of ad-hoc simulations of galaxy mergers run with the N-body/SPH code GADGET.

The simulated galaxy pairs are simple binary systems, where the stellar masses of the two initial galaxies is set to match the median stellar mass and mass-ratio of the observed SDSS sample. The simulated mergers span a significant set of five values of orbital eccentricities, five values of impact parameters, and three values of merger disc orientation, not limiting the galaxy orbits to low values of eccentricities and to small values of impact parameters. In total, 75 (5 x 5 x 3) orbital configurations for galaxy mergers are explored, and each one can be observed from a random set of viewing angles and at random times during the orbital evolution.

The authors compute the mean SFR over the 75 orbital configurations, observing each orbit from random orientations and at random moments during the merging history. Of course, these random times imply many different values of projected separations. This measurement of SFR is then translated into a measurement of SFR enhancement by normalizing by the SFR of the same galaxy evolved in isolation.

Figure 2 (from Patton et al 2013). Mean SFR enhancement as a function of projected separation in galaxy pairs from SDSS (blue) and numerical simulations of mergers (black).

Figure 2 shows the mean enhancement in star formation rate computed from galaxy merger simulations (black), and the extremely small error bars are due to the average over many orbit orientations. The curve showing the same data derived from galaxy pairs in SDSS is overlaid in blue. Remarkably, the two curves, hence the two different approaches, yield a similar result: an enhancement in SFR is observed out to large projected distances ~150 kpc, though stronger in the SDSS data. In the simulations, the enhancement is a result of starburst activity triggered at the first pericentre passage, which persists as the galaxies move to wider separations.

Hence, the authors can safely conclude that interaction-induced star formation is not only limited to those galaxies which have a close companion, but rather it affects a larger variety of galaxies.

Title: A Shapiro delay detection in the binary system hosting the millisecond pulsar PSRJ1910-5959A
Authors: A. Corongiu, M. Burgay, A. Possenti et al.
First Author’s Institution: INAF (National Institute for Astrophysics in Italy)

Some History
Shapiro time delays are one of the four tests of general relativity possible in the solar system. Because mass curves spacetime, light traveling close to a massive object must take a longer path to reach a target than if spacetime were flat, as this video and animation show. Irwin Shapiro was the first to test this phenomenon by bouncing radar signals off Venus and Mercury in the 1960s. The time delay for these signals was only about 200 microseconds.

This paper measures a Shapiro delay in a binary pulsar system called PSRJ1910-5959A. This pulsar has been previously studied but the results here include more data that allows for a more refined analysis. (See here and here for previous Astrobites posts on pulsars). This pulsar has a spin period, how long it takes the pulsar to spin about its axis, of 3.27 ms. The companion star to the pulsar is a helium white dwarf, determined by independent spectroscopic observations with the ESO Very Large Telescope and the Hubble Space Telescope. The white dwarf orbits the pulsar with an orbital period of 0.84 days. Though this pulsar visually appears to be part of globular cluster NGC 6752, it is a matter of debate whether this is actually true or is just an illusion. If the pulsar is part of the globular cluster, this represents the first time a Shapiro delay has been detected for a pulsar in a globular cluster and offers important insights into the history of the cluster.

As the white dwarf passes between our line of sight and the pulsar, there is a slight delay in the pulses from the pulsar. Since pulsars pulse so regularly, any irregularity is a sign that something interesting is happening. This delay is on the order of microseconds; it took observations spanning 10 years to detect it. Finding a Shapiro delay is exciting because it allows for very tight constraints on the mass of the companion star and the pulsar, as well as the inclination of the system.

How They Did It
To detect this delay, the research team used the 64m Parkes Radio Telescope located in Australia. For over 10 years, they regularly monitored this pulsar to detect times of arrival for the pulses. To accurately time a pulsar, astronomers fold the data on itself at the pulse period to increase the signal to noise. This yielded the team ∼1000 usable pulse timings. Check out this post for more details on the methods radio astronomers use to measure pulsar timing.

The research team used a model called the DD binary model to precisely measure the expected time of arrival for each pulse and the residuals for each detected pulse, the amount each pulse varies from the best fit. The DD binary model includes two parameters called the range and the shape that are related to the companion mass (the white dwarf) and the orbital inclination of the system. Check out Table 1 in the paper to see all the parameters that were measured or derived for this fit, and specifically note how amazing it is that pulsar periods can be measured to thirteen places past the decimal point!
To detect the Shapiro delay, a fit to the residuals is then determined, shown in the figure to the left. By finding the best fit to all the parameters of the model, then setting the companion mass to 0 and the orbital inclination to 90 degrees, a fit is determined and the remaining residuals can be seen in the top left of the figure. The team then binned and averaged the results to find an obvious harmonic in the fit of this data (bottom left). By  fitting again and removing the parameters related to the Shapiro delay,  they form the plots on the right side of the figure. Binning again brings out another harmonic, called the third harmonic, seen in the bottom right. Placing the binary companion in an elliptical orbit can explain this first harmonic, but the third harmonic can only be due to a Shapiro delay present in the data. The solid line in the figure shows the theoretical prediction of the harmonics, which matches the data well.

Once the Shapiro delay was determined, the team used these results to determine the inclination of the system and the mass of the white dwarf. They determined a companion mass of 0.180 ± 0.018M_⊙ and an inclination of at least 88 degrees. Recall that an inclination of 90 degrees is defined as a perfectly edge-on orbit. The mass of the pulsar can then be determined, yielding a mass of 1.33 ± 0.11 M_⊙. It is interesting to compare these results to those presented in the previous papers that used photometric and spectroscopic data to determine the inclination and companion mass. The results are consistent, which give credence to both methods as ways to determine these parameters.

Other Thoughts
The proper motion of this pulsar must be measured to determine once and for all if it is part of the globular cluster. If it is, this system will prove very useful in understanding mass-radius relationships for helium white dwarfs. It is difficult to determine the mass and the radius of white dwarfs using optical observations. Doing so requires white dwarf spectral models to estimate the surface gravity and effective temperature, then infer the mass and radius. Observations that do not rely on these models are needed so we can understand the interaction of these fundamental properties better.

Title: Discrete clouds of neutral gas between the galaxies M31 and M33 Nature May 9, 2013
Authors: Spencer A. Wolfe, D. J. Pisano, Felix J. Lockman, Stacy S. McGaugh & Edward J. Shaya
First author’s institution: Department of Physics, West Virginia University

Figure 1 – Artist’s conception of the region between M31 and M33 with an image of the new high resolution observations of the clouds in between the two galaxies (inside box)

Astronomers recently found seven clouds of neutral hydrogen gas (HI – “H” and roman numeral one) spread out between the galaxies M31 and M33. Could these clouds have condensed around dark matter-rich filaments, or are they leftover gas strewn across intergalactic space from a galaxy interaction event that occurred billions of years ago? Wolfe et al. use new high resolution radio observations from the Green Bank Telescope (GBT) to sort out the origin of these mysterious clouds.

The presence of neutral hydrogen gas in the region between M31 and M33 was confirmed last year with the GBT by Lockman et al. (2012). The velocity of the gas is similar to the systemic velocities of M31 and M33, confirming that it is not Milky Way gas. But the sensitivity of these initial observations was not very high. Longer integration time were needed to get sensitive high resolution images of the gas. The high resolution determines whether the gas is diffuse or clumpy. This is important for determining the origin of the gas – intergalactic filament or debris from tidal interaction between the two galaxies.

Intergalactic filaments between galaxies can serve as a bridge to funnel gas into galaxies. This has been proposed as a mechanism to fuel further star formation in spiral galaxies for a few more billion years from the gas in the intergalactic medium. However the gas seen in the space between the galaxies could have come from a tidal interaction event. When M31 and M33 came much closer together a few billion years ago, the gravitational force of the two galaxies could have stretched gaseous material between them in a tidal tail.

Figure 2 – Illustration of the spin-flip transition that gives rise to the 21 cm line.

The HI observations of the clouds were made using the 21 cm line of neutral hydrogen. This line arises from the spin alignment transition in the ground state. As illustrated in Figure 2, when the spin alignment of the proton and the electron switch from aligned parallel to anti-parallel, the atom emits a photon corresponding to the small change in energy states – one with wavelength of 21 cm. The 21 cm line is extremely useful for mapping hydrogen gas, since the symmetric H₂ molecule does not emit strongly in the radio.

Wolfe et al. observed the region of the HI gas again with the GBT, this time with higher sensitivity and higher resolution. They found the HI gas formed seven distinct dense clumps (see Figure 3). About 50% of the HI in the region is in the clouds. The clouds are about the size of dwarf galaxies. However, there are no stellar overdensities in the region, so they are not thought to be dwarf galaxies.

Figure 3 – Map of the 21 cm emission detected by the GBT in between M31 and M33. Six of the seven clouds are visible in this image (labeled numerically). The seventh is visible when the data is smoothed to a lower resolution. The directions to M31 and M33 are marked by the arrows.

Figure 4 – This position-velocity plot shows the angular distance from M31 (x-axis) vs. the velocity (y-axis) for M31, M33 (blue squares), high velocity clouds (red circles), and the new clouds detected with the GBT (black plus signs). The new clouds occupy a distinct region in position-velocity space compared to other Local Group objects.

There are several factors that make these clouds look distinctly different from the high velocity clouds (HVCs) that surround M31 and M33. First, they are much further away from either galaxy than any of the HVCs. The plot of position vs. velocity (Figure 4) shows the HVCs are clustered around their host galaxies in the position axis, while the gas clouds occupy a space distinctly their own. The velocities of the clouds also distinguish them from HVCs. The new clouds have velocities similar to M31 and M33, while the HVCs tend to have a much larger spread in velocity.

The authors identify several possible origins of the clouds between M31 and M33.
1) The clouds are primordial, gas-rich objects, like dwarf galaxies.

• We already discussed several reasons above why the clouds don’t look like dwarf galaxies or HVCs. This possibility also does not explain why the clouds seem to lie along a connecting line.

2) The gas has accreted onto local overdensities of dark matter.

• If the gas came from a tidal interaction between the two galaxies, the velocity of the gas would be too high to be accreted.

3) The clouds could be tidal dwarf galaxies – a type of irregular dwarf galaxy that form in tidal tails after a tidal galaxy interaction.

• This would explain the position and velocity of the clouds, making some of our previous arguments against them being dwarf galaxies invalid. However, this does not explain the lack of stars or why they have low internal velocity distributions compared to other tidal dwarf galaxies.

4) The clouds are transient objects that condensed from an intergalactic filament.

• This explains the location of the clouds and the lack of stars in the region. This scenario has been shown to be possible from simulations by Fernandez et al. (2012).

The authors prefer this last scenario. If there is a galactic filament connecting the two galaxies, it can funnel gas into the galaxies and fuel star formation for a few more billion years.

Title: Photophoretic separation of metals and silicates: the formation of Mercury like planets and metal depletion in chondrites
Authors: Gerhard Wurm, Mario Trieloff, and Heike Rauer
First author’s institution: University of Duisburg-Essen

Iron-rich Mercury

Mercury’s high density has been a longstanding puzzle in planetary science. Its density means that it must have a significantly higher iron abundance than Venus, Earth, Mars, or the asteroids, probably in the form of a large iron core. A common popular explanation (the one I was taught four years ago in my Intro Planetary Science course) is that a giant impact came along late in the formation process and stripped the silicate mantle off of Mercury, leaving behind an iron-rich planet. However, in the last four years NASA’s MESSENGER mission has challenged that idea: measurements of elements like potassium—which would be extremely depleted after such a giant impact—show no depletion, so the giant impact hypothesis has been largely ruled out. A new hypothesis is required.

In the past four years, we’ve also found some irradiated rocky exoplanets, most of which are even hotter than Mercury, and there is some evidence to suggest that the two hottest rocky exoplanets with measured masses (Corot-7b and Kepler-10b) may also be more iron-rich than the Earth. While still in the realm of small-number statistics, this is suggestive that there is some more universal process shaping the formation and evolution of rocky planets. In this paper, the authors propose an explanation for the formation of both Mercury and these iron-rich exoplanets.

Sorting the particles in a disk

This cartoon illustrates how the general sorting mechanism proposed in this model works. The top panel shows that at the illuminated edge of the disk, silicates are pushed more efficiently than metals by this process. The bottom panel illustrates the resulting composition difference in the protoplanetary disk.

The basic idea proposed here is that there is a mechanism in the disk that sorts metals from silicates, making the inner part of the disk more metal-rich and the outer part of the disk more metal-poor (and thus silicate-(rock-)rich). (I’ll note that here I’m using “metal” not in the astronomer’s sense—anything heavier than H/He—but in the normal Earthling usage where metals are metallic.) If this sorting mechanism were efficient on timescales shorter than the formation of planets (a few million years), then the planets that formed from the inner material would naturally be more iron-rich.

Figure 3 shows a cartoon version of how this process occurs. The disk starts out with uniformly distributed iron and silicates, and then the silicates get preferentially pushed to the outer regions of the disk.

The proposed mechanism: Photophoresis

This cartoon illustrates how the proposed mechanism photophoresis works. In the presence of irradiation, a particle will have a warm side and a cool side. When gas molecules hit the particle, those rebounding from the warm side will have higher energy (and therefore higher velocities) than those rebounding from the cool side. This imparts a net force, away from the irradiating source, on the particle.

The mechanism that they propose to do this sorting is called photophoresis. The idea is that a particle in a protoplanetary disk will be illuminated on one side (by the Sun), and that side will become warmer. The particle will then be hit by the molecules in the disk. These molecules will equilibrate to the surface temperature of the particles before being ejected with the energy (and velocity) corresponding to that temperature. Since the temperature is asymmetric—hotter on the illuminated side—the net momentum imparted onto the particle will also be asymmetric. In general, the particle will be pushed outward, away from the illuminating source.

It’s clear how this mechanism could move particles—but why does it sort them by composition? The reason is that the strength of the photophoretic force is inversely proportional to the thermal conductivity. This is straightforward to understand conceptually: if a particle has high thermal conductivity, the heat travels efficiently through the particle, so even though only one side is illuminated, the whole particle will warm up. The temperature difference between the warm and cold side will be small, and a smaller net momentum will be imparted by the molecules hitting and rebounding from the particle. If instead the material has a low thermal conductivity (an insulator), heat travels slowly through the particle, so the temperature difference between the sides of the particle will be larger, and the corresponding net momentum gained will be larger.

Since iron is a metal, it has a high thermal conductivity, about 50 W/(mK). Silicates have much lower thermal conductivity, about 1 W/(mK). This means that photophoresis will be about 50 times more efficient at transporting silicates away from the Sun than iron particles.

Based on the properties of a protoplanetary disk, the authors calculate the drift times for particles to move 1 AU to be 300,000 years for silicate grains (well within the formation timescale of planets) and 50 times longer, or 15 million years, for iron grains (similar to the timescale to form planets).

Photophoresis would thus naturally create a composition gradient in the disk, with metal-enrichment in the inner parts and silicate-enrichment in the outer parts. This could explain the iron-enrichment of Mercury, as well as the apparent slight iron-enrichment of the rocky exoplanets Corot-7b and Kepler-10b.  Of course, many other physical mechanisms are at play during planet formation, but the authors suggest that exploring this mechanism further could be fruitful in explaining both solar and extrasolar planetary systems.