Searching for Dark Matter with Light

What is the universe made out of? In grade school, most students learn about atoms, then electrons, protons and neutrons. As it turns out, protons and neutrons are made up of even smaller components called quarks. Quarks, however, are only one subset of a whole system of fundamental particles described by the Standard Model of particle physics. The Standard Model is an extremely useful theory that describes the multitude of particles that make up our universe and how they interact with each other. The Standard Model includes particles you’ve probably heard of, like the photon, and some you might not have, like the tau neutrino. The Standard Model does very well at explaining much of what particle physicists have observed over the years, but there are some problems with it—indications that the particles and interactions described in the theory do not account for all of the matter in the universe. This extra matter is known as “dark matter”.

My research with Professor Kerstin Perez this past summer, to be continued as my senior thesis in physics this year, focused on learning how to search for dark matter with a high-energy x-ray satellite telescope called NuSTAR. NuSTAR is the first focusing telescope in its energy range, meaning that it has much better spatial resolution in that range than any other instrument. For this reason, it may be a promising tool for searches for dark matter.

Screen Shot 2015-09-15 at 1.26.19 PM

An illustration from of the NuSTAR satellite telescope, launched in 2012.

Dark matter accounts for over eighty percent of the matter in the universe, but nobody knows exactly what it is. It is so difficult to study because it interacts extremely weakly with light. Direct evidence for its existence comes mostly from its gravitational effects on normal matter, such as the speed of galaxy rotation and gravitational lensing (or bending) of light. Nobody has so far been able to figure out the actual characteristics of the particles that make up dark matter. Particle physicists have put forth a variety of theories, each one attempting to fix a “hole” in the Standard Model. Among the more popular theoretical particles are WIMPs (Weakly Interacting Massive Particles), axions, and sterile neutrinos. My research this summer centers mainly on the sterile neutrino, a theorized type of neutrino that interacts only via gravity, not gravity and the weak force as do other types of neutrinos. Why is the sterile neutrino interesting for my research? As it turns out, even though all theorized dark matter particles interact very weakly with light, some might occasionally decay to produce photons that can be detected with a telescope. When a sterile neutrino decays, it releases a “normal” neutrino along with a photon of energy equal to half the mass of the sterile neutrino (remember Einstein’s E=mc2 equation, which states that energy and mass are essentially interchangeable). The range of possible photon energies from sterile neutrino decay falls within the span energies that NuSTAR can observe, making sterile neutrinos a potentially promising target of study using NuSTAR.

The way this type of dark matter search works is by looking for peaks in spectral data that might correspond to photons from dark matter particle decay. NuSTAR points at some object in space (like a galaxy), collects photons coming from that object, and sorts them by energy. One can observe the output as an image of the light source, or as a graph of energy versus photon counts called a spectrum. Using a spectral analysis tool from NASA called XSPEC, one can fit a model of all the components of the spectrum that are well-understood, i.e. blackbody radiation, atomic transitions lines, etc. If there is a peak in the data above these components that cannot be explained by anything known, there is a chance it can be explained by dark matter.

Of course, despite the multitude of teams of scientists working on such searches, no definitively significant dark matter spectral peaks have been discovered. It is quite likely, then, that despite NuSTAR’s excellent spectral resolution and high energy range, my project will not produce concrete evidence of dark matter particle discovery. But do not despair, for useful results can still be produced from a dark matter search that comes up empty. What a null result can do is rule out certain characteristics of dark matter particles. Particle physicists would like to know the mass and mixing angle—the strength of interaction between a dark matter particle and normal matter that defines how likely a sterile neutrino is to decay into a photon/neutrino pair—of dark matter particles. Using spectral data, I can say, if a sterile neutrino has this combination of mass and mixing angle, I will see it as a peak in the spectrum. If I do not, it must not have these characteristics.

Over the summer, I produced a detailed procedure for finding mixing angle constraints for a given dark matter mass, and used simulations and sample data to find tentative estimates of NuSTAR’s capabilities in this sort of dark matter search for a few promising observation targets. I will try to summarize the procedure without getting overly technical. Once spectral data is obtained—this can be real data or simulated data from a tool like SciSim, a simulator developed for the XMM Newton X-ray observatory—find a model to fit the data using NASA’s XSPEC tool. Next, calculate the total dark matter mass in the field of view of the telescope observation. For my mass/mixing angle estimates looking at the Milky Way center, I did this by integrating a Milky Way dark matter density function found in a 2010 paper from a team led by Alexey Boyarsky. For targets farther away from Earth, I looked up the dark matter density of the object and multiplied by the area of the NuSTAR observation at that object. Next, select a few sterile neutrino masses for which to calculate mixing angle limits. Remember, the photons that correspond to sterile neutrino decay have half the energy of the sterile neutrino. Using the model found for the spectral data, determine the maximum possible flux (related to the photon counts at a certain energy) for each sterile neutrino mass using XSPEC. For the most conservative estimates of mixing angle, just say that the dark matter flux is the flux of the model in XSPEC. Next, calculate the maximum decay rate to photons given each sterile neutrino mass and maximum flux. For this step, I used an equation from the same Boyarsky et. al. paper that contained the dark matter density function. From the decay rate, one can calculate the mixing angle using another formula in a 2006 paper written by Boyarsky et. al. Once each chosen possible sterile neutrino mass is matched with a mixing angle, the results can be compared to existing mass/mixing angle limits from previous experiments.

I ran through this process for a few different potential sources. These sources included the Milky Way center, a galaxy called M31 (also known as Andromeda), and the dwarf spheroidal galaxy Fornax. These sources were chosen for their high dark matter content and well-understood background spectra, among other factors.

Screen Shot 2015-09-15 at 1.26.45 PM

An image of the M31 galaxy from

Over the course of the summer, as I learned more about NuSTAR’s instrumental background spectrum and about the potential dark matter observation targets, I came up with mass/mixing angle limits that were increasingly more accurate to NuSTAR’s capabilities. By the end of my 10 weeks of research, I had this plot:

Screen Shot 2015-09-15 at 1.27.03 PM

My calculated mass/mixing angle limits, overlayed as colored points on a figure taken from Boyarsky et al., 2013, only barely improve on previously discovered limits, if at all. My process for finding these limits, however, was quite conservative, as I explained earlier. A promising next step in my research could be to use maximum dark matter flux values that are, say, related to the error bars on the spectral model’s flux instead of just being the value of the flux from the model. This would make the flux estimates much smaller, and thus make the mixing angle limits smaller as well. In this way, with the help of NuSTAR data, my research could potentially break new ground on mass/mixing angle limits for sterile neutrinos.

Serpentine Soil Summer Research

I spent my summer playing in the mud. And I had a great time.

In the field

In the field

Well, it was a bit more structured than that. For 10 weeks this summer, I completed field work that would become the basis of my senior thesis research project in Byrn Mawr’s geology department. Funded by the KINSC, I conducted a soil survey to analyze different restoration activity at the Unionville Serpentine Barrens in Unionville, PA.

Geologists spend a lot of time thinking in scales, both temporal and spatial. This skill allows us to look at one small fossil and think back 500 million years, to look at a single mountain belt and zoom out to tectonic plate interactions. It also helps us think across academic fields; to understand how the mineral asbestos interacts with public health and why erupting volcanoes affect government security measures. Me personally? I’m interested in how geology and biology interact; how bedrock influences the lives of plants and insects above it. More specifically, I’m fascinated with how the mineral serpentine, formed in the mid-oceanic ridge between 250 and 500 million years ago, facilitates evolution and ecosystem succession in southern Pennsylvania.

Serpentine is an ultramafic mineral, meaning it is igneous, iron- and magnesium-rich, and deficitent of potassium and nitrogen. It forms in the mid-oceanic ridge (MOR) when the dominant rock in the mantle, peridotite, mixes with water and becomes hydrated. The serpentine, now part of the oceanic crust, moves away from the MOR as newer crust is formed. It typically then subducts under the continental crust and rejoins the mantle.


In rare cases, serpentine will be incorporated onto a continent through island arc-continent collisions. If exposed, outcrops are created – which, when weathered, turn into serpentine barrens. Other than their uniqueness, these barrens are interesting because serpentine soil is typically deficient in phosphorus, potassium, and calcium and abundant in heavy metals – all conditions that are highly inhospitable to plants. The mean Ca:Mg ratio of serpentine soil is < 1. Where it is shallow, the soil fails to retain moisture well. In the Northeast, these barrens mark a dramatically different ecological environment from their surrounding temperate forests. The barrens host many disjunct, edge-of-range, state-endangered, state-threatened, and otherwise rare species, as well as two endemics.

In the mainland United States, there are two major areas where serpentine outcrops are found: northern California and the tri-state region of Pennsylvania, Delaware and Maryland. Lucky me! Through a series of fortunate events I met William Ryan, a University of Delaware doctoral candidate whose dissertation is a comprehensive study of the rare plant populations on the shrinking Unionville Serpentine Barrens (USB) in Unionville, PA.

Partially protected by the Natural Lands Trust and partly in private ownership, the Unionville Barrens are of high biodiversity conservation priority. Of the 63 acres of barrens present in 1937, a mere 8 remained in 2008. This mimics a regional trend: across the Northeast, the total acreage of serpentine barrens is shrinking due to mining and quarrying, development, invasion by non-serpentine vegetation and communities’ fear of prescribed burning. This loss of habitat has led to the extirpation of many state-endangered and state-threatened species. Ten plant species of special concern previously found at the Unionville Barrens are no longer present.

Through William, I met Dr. Roger Latham, an ecologist and conservation biologist who created the restoration and adaptive management program of the USB. The management plan includes prescribed burning and a 12-acre invasive tree removal effort (2012) to restore the grasslands to documented conditions that existed 20 to 70 years ago as shown in aerial photos. Fires and tree cuttings (of non-native species like the autumn olive) remove the organic matter that acts as  a buffer between invasive plants and the inhospitable serpentine soil.

By surveying the same 120 square meter quadrats as William’s exhaustive multi-year plant survey, I can analyze the correlation between soil characteristics and vegetation response to restoration activity in a serpentine barren. This type of direct comparison has heretofor never been seen. So, for one and a half months I woke up early to beat the heat, drove to Unionville, unloaded long and heavy soil corers and hiked into the barrens through clouds of mosquitos and painful but scrumptious blackberry bushes. In each study plot I collected organic material and took four soil cores and four depth-to-bedrock measurements. I then categorized, measured and collected the soil horizons based on the cores. I labelled 480 plastic bags, measured 480 soil horizons, took 480 depth-to-bedrock measurements.


1 of 480 soil cores!

For the next portion of my research I moved to Dr. Alain Plante’s lab at the University of Pennsylvania. I had taken Dr. Plante’s soil science course over the spring and he was generous enough to offer up advice, lab space and equipment. In a soil receiving and processing lab, I air-dried, sieved, and grinded all 480 samples. This process took 3 weeks.

Stackin’ up the processed samples in the lab.

The fun is just beginning. This semester, I am continuing the lab portion of the work at Penn. I am running Mehlich 3 extractions on each sample to use an ICP, inductively coupled plasma, machine to measure the extracted nutrients. My soil analysis will investigate soil depth, Ca:Mg ratio, soil organic matter content, and plant-absorbable heavy metal levels. With the help of a lab assistant, this process will hopefully take half the fall semester. Then, I can analyze the data!

I hypothesize occupation by trees for 20 to 70 years deepened the soil and enriched it in organic matter, providing a buffer between colonizing plants and the underlying bedrock chemistry that persists long after tree removal occurs. This buffer is likely to be most pronounced in concave topography and lower slopes and least pronounced on convex topography and upper slopes. Predictions include positive correlations between increasing soil depth, Ca:Mg ratio, soluble nickel concentration and organic matter content, and the prevalence of nonnative invasive plants and common native plants not usually found in serpentine grasslands. I also predict negative correlations between the same soil factors and the prevalence of the characteristic species of mature serpentine grasslands, including the species of conservation concern.

This project is relevant in the world of soil science, plant evolution, and conservation biology. It will provide a yet-to-be-seen direct comparison of serpentine soil and serpentine plants. It will aid serpentine conservationists in implementing restoration practices and protect rare species. The rare plant species, like the aster depauperatus, are host to rare insects, like the barrens buckmoth, which contribute to the pool of biodiversity. In protecting these species from succession of the barren grassland to woodland, we can protect the serpentine ecosystem and help save endangered species.

Aller guten Dinge sind drei*: 22 Prepositions, 264 Sentences, and One Test

—Claire Dinh ’16
*All good things come in threes.

Wie sehr schnell die Zeit vergeht! In German, that means, “How very quickly time flies!”

I can’t believe that summer is already over and that my last year at Haverford has begun. I really do have to thank Dr. Chatterjee and Dr. Jamrozik for making my first experience as a lab intern absolutely wonderful. But while my internship is over, my research experience is not. This fall, I am continuing my work at the Center for Cognitive Neuroscience!

But before I move on to a discussion of what we have completed so far and what we still have to do, here is a brief reminder about what we are trying to study. Let’s talk science!

In my last post, I wrote that Dr. Jamrozik and I plan to create a test that will gauge the abilities of those taking it to use different kinds of prepositions (spatial; temporal; and abstract, metaphorical). The individuals with whom we plan to work are patients who have lesions in the parietal-temporal regions of their brains. We ultimately would like to use the test to answer the following question: Do impairments of spatial, temporal, and abstract prepositions co-occur, or can these uses be independently impaired?

Now let’s move on to a recap of what we achieved over the summer.

Over the course of 12 weeks, we went through a total of 10 iterations of our test—and it still is a work in progress. You may be wondering why the process of writing our test has been so tedious. Well, we need to make sure that it will measure what we want it to measure. The validity of our test determines the validity of our results, which, in turn, will tell us if our hypotheses have been supported or not.

The first step in writing our test was identifying the 22 most commonly used English prepositions. We used two corpora (two large sets of naturally occurring speech and writing) to accomplish this. We then proceeded to write sentences that we eventually will turn into fill-in-the-blank sentences for our final multiple choice test.

So a sentence testing whether or not someone can produce a spatial use of the preposition on would look something like this.

The book is _____ the table.
(a) at
(b) on
(c) through

Correct answer: (b) on.

For each preposition, we drafted four sentences per each use, for a total of 264 sentences (22 prepositions x 3 uses/preposition x 4 sentences/use = 264 sentences).

We did our best to write sentences that included preposition meanings that had been repeated across sources (we took note of such occurrences when conducting our lit review). We wanted to ensure that our sentences were as reflective as possible of everyday speech.

There were a number of other factors we had to consider when writing our sentences. Here are a few.

  • Word order: For the most part, our sentences had the following structure: article + noun + verb + preposition + article + prepositional object (e.g. The cat is in the box). Sometimes, though, it made sense for us to also include a direct object (e.g. The boy threw the ball over the fence) or omit an article (e.g. Monday is before Tuesday).
  • Frequency of verbs: We limited the number of verbs that we used, as they should not be so obscure that our patients will not understand them. We also were consistent in the frequency of our use of the verb “to be” across the different preposition uses.
  • Types of nouns: Prepositional phrases name a relationship between a figure and a ground (e.g. in the sentence, The cat is in the box, the figure is the cat and the ground is the box). In our sentences, we wanted to make sure that our figures were count nouns (i.e. nouns that can be counted), which means we never used mass nouns like coffee or happiness as figures (though coffee does make a lot of us happy). We also wanted to make sure that our grounds were reflective of the preposition use being tested. For example, a sentence demonstrating an abstract use of a preposition would have a ground that is an abstract entity or concept, as demonstrated by the sentence, The café is popular within the community, where the noun community is an abstract ground.

After writing our 264 sentences, we began to norm them. We asked healthy participants to rate how natural each sentence sounded to them, and for each set of sentences for each use of a preposition (e.g. we have four sentences demonstrating a spatial use of the preposition in), we asked our participants to select the two sentences that they believed best represented common uses of that preposition. Ultimately, we would like to cut down the number of sentences we have by about half, so that our patients will not be too overwhelmed by the test.

And that is where I am today! But before I leave you with a Tschüss! for the next several weeks or months, I would like to send a special shout-out to Misha and Melissa, my fellow undergrad interns at the CCN. I really enjoyed spending time with them during my internship and at the other places we visited in the city!


From left to right: Cody, Misha, Melissa, and me enjoying music at Festival Pier.