Actinide Production in Neutron-Star Mergers

Erika M. Holmbeck

25 October 2018

Good evening everyone, and thank you for staying here this late. Today I am excited to talk about some new progress we have made towards understanding actinide production using both observational and theoretical methods.

The r-process

Holmbeck with PRISM (T. Sprouse and M. Mumpower)

So we've heard a lot about the r-process in this session. Just as another reminder, about half the elements heavier than iron in the Solar system were made by the rapid-neutron capture process. The actinides--Thorium and Uranium--can only be made by rapid neutron capture. Slow neutron capture cannot cross this instability gap between Bizmuth and Thorium. [START MOVIE] The system starts with some distribution of seed nuclei, then captures neutrons--in this case, all the way out to the neutron-drip line. Eventually, the neutron flux decreases significantly, and the system beta-decays back to stability. As you can see, material can end up in the heavy region where it may populate Thorium and Uranium.

R-II stars

Sneden+ (2003)

One way we can learn about the r-process and its astrophysical site is by looking for signatures that the r-process imprints on the Universe through r-II stars. An r-II star is a very metal poor star that shows heavy enhancement of r-process elements. Their metal-poor nature and enhanced quality indicates that ...some strong r-process event or events occured before these stars were made and enriched the gas from which they were created. These r-II stars are nearly pure r-process patterns from one to a few r-process events

The actinide boost

Holmbeck+ (2018)

If you heard Tim Beers's talk this morning, you might have heard about a huge observational effort we are conducting to find more r-II stars. During this survey, we found an actinide boost star. The rest of the r-process pattern looks like every other r-II star, and the actinide abundances are much greater than what's in the Solar System. Many literature studies can reproduce r-II stellar abundance patterns, but fail to account for the actinide boost.

Are NSMs the source of the actinide boost?

Do varying levels of neutron richness account for the actinide boost?

Does the presence of the actinide boost indicate a separate r-process site? Compared to a traditional CCSN, NSMs are predicted to be more neutron rich, which may encorage more actinide production. Can we turn a knob to change the neutron richness and then get different amounts of Thorium?

T. Sprouse and M. Mumpower

Low-entropy dynamical (tidal) ejecta of a NSM
(Korobkin+ 2012; Rosswog+ 2013)

Vary the initial electron fraction: $Y_e=0.005 - 0.250$

---

$Y_e$ = 0: all neutrons

$Y_e$ = 1: all protons

To investigate this, we use PRISM: an r-process network solver written by Trevor Sprouse and Matt Mumpower. Prism allows for complete flexibility of the nuclear and astrophysical input. For the actinide boost problem, we choose a low-entropy, NSM tidal ejecta trajectory and vary the initial electron fraction. From here on out, I'll be talking about this Y_e, the electron fraction. Y_e essentially measures the fraction of nucleons that are protons. So just remember that a lower Y_e means more neutrons, and a higher Y_e is more protons. We ran 50 separate, full r-process calculations, changing the initial Y_e between 0 and 0.25. We run each of these to 1 Gyr, so every abundance pattern I'll be showing are the final abundances at 1 Gyr.

Actinide and lanthanide production

Holmbeck with PRISM (T. Sprouse and M. Mumpower)

So, starting at the highest Y_e of 0.25, we only make up to the second r-process peak. We make the second r-process peak, but nothing beyond it in the rare-earths. Decreasing the Y_e, we start to get some lanthanides. Decreasing further, we are starting to robustly fill out the rare earth region and the third peak. And finally, we get the actinides...and then we keep getting actinides. We're producing lots of Thorium, which is what we wanted!

Ages and cosmochronometry

232-Th and 238-U are radioactive

Allows radioactive decay dating

$$ t = 46.67~\text{Gyr} \left[\log\epsilon\left(\text{Th/Eu}\right)_0 - \log\epsilon\left(\text{Th/Eu}\right)_{\text{obs}}\right] $$
$$ t = 14.84~\text{Gyr} \left[\log\epsilon\left(\text{U/Eu}\right)_0 - \log\epsilon\left(\text{U/Eu}\right)_{\text{obs}}\right] $$
$$ t = 21.80~\text{Gyr} \left[\log\epsilon\left(\text{U/Th}\right)_0 - \log\epsilon\left(\text{U/Th}\right)_{\text{obs}}\right] $$

Thorium and Uranium are radioactive. Instead of comparing our simulations results directly to observational abundnace patterns, we use radioactive decay dating to estimate the age of the r-process material for stars with actinides. The only things we need are the measured abundance ratios of the actinides to some stable r-process element, and the initial amount the star starts with. These production ratios have to be taken from r-process simulations since we don't know what the star was born with. If the production ratios are a good estimate for what the star began with, then all three of these ages should agree, and our simulation is therefore a good descrption of our star.

The age of actinide-boost star J0954+5246

Holmbeck+ (arXiv:1807.06662)

This plot shows those three age equations applied to the actinide boost star across the whole Y_e range we considered. What we look for is consistency between all three ages, which would imply that our r-process simulations accurately describe the star. This only happens at one point. If the Y_e is lower than about 0.17, then the actinides are always way OVER-produced in these simulations.

Actinides are currently not observed at such high levels

Need a method to dilute the actinides

This star is currently the most actinide-enhanced r-II star known. We don't observe the actinides any higher, so we need a way of diluting the actinides, but keeping the lanthanides so that the Th to Eu ratio is lower.

Distribution of $Y_e$

Tidal ejecta centered at $Y_e=0.16$
(Bovard+ 2017)

Wind centered at $Y_e=0.22$
(Lippuner+ 2017)

$m_{\rm wind}/m_{\rm dyn}= 3$
(Rosswog+ 2017; Tanaka+ 2017)

Since nearly every single Y_e value failed to explain the actinide boost star, we instead develop a distribution where we combine all the individual abundance patterns based on a distribution of Y_e. We chose a distribution of Y_e based on what we found in the literature for NSM ejecta components. Now we have a distribution of Y_e that characterizes the NSM ejecta, and we apply that distribution to our final abundances using all the Y_e calculations we ran.

Distribution of $Y_e$

Holmbeck+ (arXiv:1807.06662)

Using a mass distribution motivated by literature, we get this final combined abundance pattern in red. Each of these blue and green lines are the individual patterns from just choosing one Y_e. This method indeed diluted the actinides down to observational levels, and maintained the classic r-process signature in the lanthanides. The pattern looks good, but we don't really know if we've succeeded until we see the ages. The three ages have to agree in order for our simulated abundances to be a possible explanation of actinide boost stars.

Ages

Holmbeck+ (arXiv:1807.06662)

In blue here are the ages from using just one low Y_e trajectory that overproduced the actinides. You can see that not only are the actinides over produved, but the implied ages are very inconsistent with each other. After applying some actinide diluition, the ages we found are much more realistic than the previous result from just one, low-Y_e trajectory. 15 Gyr is much more successful than 50. This is only a first test. The mass distribution we chose was just from literature, so it could be that a different choice of mass distribution or nuclear physics could make these ages agree perfectly. We at least need some type of dilution scenario if we want NSMs to be the source of the actinide boost.

Conclusions

We are conducting a survey to find more r-II stars

During this survey, we found the most actinide-enhanced r-II star to date

Low-$Y_e$, low-entropy NSM conditions (e.g., in the tidal ejecta) over-produce the actinides

NSMs could be a source of the actinide boost if we add another r-process from a higher-$Y_e$ environment (e.g., the accretion disk wind)

In conclusion, the R-Process Alliance is searching for more r-II stars so we have more observational data to study. During our survey, we've had many successes already, including identifying the most actinide-enhanced r-II star currently known. Which brings up the question of how the actinides are over-produced in nature. We investigated whether NSM can produce this actinide boost and found that the very neutron-rich dynamical ejecta not only produces the actinides, but produces them in amounts that are much too high compared to observations. NSMs could still be a source of the actinide boost if we account for r-process material coming from higher-Y_e parts of the merger, perhaps in the accretion disk wind. I will stop there, and thank you again for tuning in.