Observing Newtrinos


The field of particle physics spent much of the last century converging on the "standard model" that describes subatomic particles and the forces by which they interact. Ambitious research that began in the early 1900s with hot-air balloon experiments aimed at catching cosmic rays, and followed later in the century with accelerator-based experiments, led to the extraordinary success of the standard model. Nonetheless, the excitement in particle physics often lies not with the vast body of solidly established textbook knowledge, but rather on the fringe, where researchers seek to identify new physics beyond the standard model. Some even hope to find a problem with the model, in order to spur an intellectual expedition into the unknown. And mounting evidence indicates they may finally get that chance.

Exploring the fringe is hard to do, often requiring more money for bigger accelerators in order to probe energy scales that would otherwise remain just out of reach. But sometimes you get lucky: an unexpected quirk could reveal itself in a more mainstream experiment. That's what happened in 1995 when results from Los Alamos's Liquid Scintillator Neutrino Detector (LSND) were released. The results hinted at the existence of a new particle, and in so doing, defied elements of the standard model that had already seemed established through other experiments. As the scientific method demands, the unexpected results were intensively scrutinized—even criticized—and now duplicated. The MiniBooNE collaboration (Mini Booster Neutrino Experiment), located at the Fermi National Accelerator Laboratory in Illinois, has accumulated enough data to shed some light on the anomaly. And their results are consistent with LSND.

Both LSND and MiniBooNE are detectors for subatomic particles called neutrinos. To appreciate their experimental results, it is useful to examine how neutrinos fit in among the other elementary particles. The standard model identifies two categories of matter particles: quarks and leptons. Protons and neutrons, which live inside atomic nuclei, are made from quarks and therefore interact with other matter particles the way quarks do; that is, they interact via electromagnetic forces, strong nuclear forces, and weak nuclear forces. (They also interact gravitationally, as all matter and energy does, but gravity is far too weak at the individual particle level to have any measurable effect in particle physics experiments.)

Leptons behave somewhat differently. They do not experience the strong nuclear force and therefore cannot be bound inside atomic nuclei. Some leptons, like the electron, are electrically charged and therefore interact electromagnetically. But neutrinos are uncharged leptons. They experience only the weak nuclear force, and that force is, as its name indicates, quite weak. As a result, neutrinos rarely interact with other particles. Nearly 100 trillion neutrinos originating in the Sun pass straight though each person on Earth every second, and statistically, only one of these solar neutrinos will interact with any subatomic particle in that person's body during his or her entire life. And it's not just the human body; neutrinos rarely interact with anything, including neutrino detectors, making neutrino research very challenging. (Neutrinos might never have been discovered at all, except that some energy seemed to go missing whenever a neutrons decayed into lighter particles, and to conserve total energy, it seemed reasonable to hypothesize the existence of phantom particles that take the missing energy away with them in order to ensure that total energy is conserved. Neutrinos remained hypothetical for two decades until they were detected experimentally in the 1950s by Los Alamos researchers Fred Reines and Clyde Cowan; Reines lived to receive the 1995 Nobel Prize as a result.)

Photomultiplier tube

Photomultiplier tube.

To Catch a Neutrino

Catching a neutrino interacting with some other particle by the weak nuclear force is difficult, but not impossible. One needs a very large number of neutrinos in a very large detector in order to catch one neutrino interacting amidst countless trillions that cross the detector unnoticed. In addition, the neutrino interaction must be inferred. When a neutrino collides with another matter particle, the collision can create a charged particle—usually either an electron or its antimatter twin, a positron. That electron (or positron), not the neutrino itself, is what produces a detectable signal.

MiniBooNE detector

The 12-meter-diameter MiniBooNE detector (seen here through a fish-eye lens) is located underground to shield it from cosmic rays, which could create false detection signals.

The MiniBooNE detector is a massive, spherical tank filled with 250,000 gallons of clear mineral oil. Nearly 1300 extremely sensitive light sensors, called photomultiplier tubes (PMTs), line the interior wall. When a neutrino interaction within the mineral oil causes an electron to zoom through the detector, that electron travels faster through the mineral oil than light would. (Nothing travels faster than light in vacuum—but through a medium like air, water, or mineral oil, this is possible.) Crossing the light-speed barrier in a medium is like crossing the sound-speed barrier, triggering the optical equivalent of a sonic boom. This is known as the Cherenkov effect. A forward-directed flash of visible light spreads out in an expanding cone until it reaches the PMTs, where it is recorded. In this way, MiniBooNE detects a neutrino.

Bill Louis is a physicist at the Los Alamos National Laboratory and a pioneer in the search for neutrino signatures. He worked on LSND and is currently collaborating on MiniBooNE. He has spent much of his career trying to isolate electron (or positron) signals that originated with a neutrino from false signals that didn't. This is a challenge because electromagnetic signals from other electrons, or even from photons (particles of light), arise frequently and can easily mimic a neutrino-induced electron.

"The trick is to identify the distinguishing characteristics of the electrons and positrons created by neutrinos, and develop methods to reject all the imposters," says Louis. He points out that when a neutrino interacts with the mineral oil, there are side effects to look for. For example, in addition to spawning a Cherenkov cone of detectable light, the neutrino-induced electron or positron bleeds energy into the surrounding oil, causing another glow called scintillation, which spreads out spherically in all directions. "So you've got the Cherenkov cone and the scintillation glow together: that's your neutrino signature," says Louis. If either of these components is missing, the event is disregarded.

But despite this multiple-component recipe for detection, rejecting all the imposters remains a challenge. MiniBooNE employs a variety of methods to ensure that its detections are genuine. It lies underground beneath 10 feet of earth and has a special "veto shield" in order to prevent false signals that could originate with cosmic rays raining down from space. Even the neutrino beam that is deliberately aimed at the MiniBooNE tank is unavoidably accompanied by imposters that must be handled.

The MiniBooNE experiment begins when a Fermilab accelerator sends a beam of high-energy protons to smack into a target, thereby generating a spray of particles. Most of these particles are deliberately blocked by an absorber medium placed in the beam path. But some of these particles are pions, which decay into muons (leptons similar to electrons but 200 times more massive) and the neutrinos (and antineutrinos) the experiment seeks to measure. The trouble is, not all imposters are blocked by the absorber. Muons, and even some of their associated neutrinos, constitute imposters in the eyes of the detector, as do stray gamma rays. All of these can be controlled, or at least accounted for, but this remains a tricky business: anyone hoping to observe new physics must be prepared to observe new imposter physics instead. Thus, if the experiment's results turn out as expected—perhaps matching the majority of neutrino experimental results and leaving LSND as the sole outlier—then the experimenters might feel satisfied they eliminated all the imposters. But if the results differ from expectations, how can experimenters ever be certain they eliminated them all?

One way to help expose the imposters is to alter the experiment, so that the signal from the neutrinos in the beam is the same but the signature from the imposters changes. "In order to make MiniBooNE sensitive to the same effects we saw at LSND, we had to design MiniBooNE to similar specifications, but we didn't want to make an exact copy of LSND because what would we learn from that?" says Richard Van de Water, a MiniBooNE colleague of Louis's at Los Alamos. "If you do the math, you find that it's the ratio of the neutrinos' travel distance to their detected energy that is the critical factor. So with MiniBooNE, we changed the distance and the energy from LSND, but we kept their ratio the same." Indeed, the neutrino signal MiniBooNE attempts to isolate has a different dependence on distance and energy than that of its most troubling imposters.

The Fermilab Booster accelerator generates a beam of protons.

Left to Right:

The Fermilab Booster accelerator generates a beam of protons.

The proton beam hits a beryllium target (center), causing a spray of particles, including positive and negative pions. These pions are steered toward the MiniBooNE detector by a magnetic focusing horn (surrounding the target).

Pions enter a 50-meter-long air-filled pipe where they decay primarily into muons (or antimuons, depending on the charge of the pion) and muon-flavored neutrinos (or antineutrinos).

A steel and concrete barrier allows neutrinos (and antineutrinos) to pass but filters out other particles.

As muon-flavored neutrinos (or antineutrinos) travel from the beam source to the MiniBooNE detector through 480 meters of earth, the flavor states "oscillate" as a probability wave: the probability of measuring the neutrino in each flavor state goes up and down. The wave shown is simplified for two-state mixing; three (or more) flavors would generate a more complicated oscillation pattern. For the wave shown, the troughs represent a 0 percent chance of transition (still muon flavor) and the peaks represent 100 percent chance of transition (to electron flavor). A point at the wave's midpoint between peak and trough represents a 50-50 chance between the muon and electron flavor states.

An electron-flavored neutrino (or antineutrino) may be detected if it happens to collide with the nucleus of a hydrogen atom (a proton) in the mineral oil that floods the detector. Here, an electron-flavored antineutrino (black) collides with a proton (red), causing the production of a neutron (white) and a positron (green). The collision imparts enough energy onto the positron to create the light-equivalent of a sonic boom, sending a cone of visible light forward until it is picked up by the detector's photomultiplier tubes.



Flavor Physics

Quarks and leptons, including neutrinos, come in different varieties known as flavors (sometimes called families or generations) which determine how they behave in interactions involving the weak nuclear force. For example, when the pions in the MiniBooNE beam decay into muons and neutrinos—a process governed by the weak force—the neutrinos are always muon-flavored neutrinos (or muon-flavored antineutrinos, depending on the positive or negative charge of the pion). The reason this is so under the standard model is that weak flavors are always conserved in reactions and decays. So when a flavorless pion decays into a muon, it must also produce a muon-flavored antineutrino. Since we started with no net flavor, the decay must produce none: the flavors of a muon and a muon-flavored antineutrino cancel each other out.

However, such weak flavor states are not defined in a simple way, nor are they necessarily permanent. An electron-flavored neutrino, for example, exists in a blend of multiple mass states—states with different masses—meaning that its mass is not precisely defined. Over time, the mass states interfere with one another in such a way that the flavor state fluctuates as well. Thus, an electron-flavored neutrino can change to muon flavor, and to a third flavor called tau, and back again repeatedly. Indeed, until the beginning of the twenty-first century, this very phenomenon stymied researchers trying to observe the neutrinos emitted by the Sun. Their detectors were only sensitive to electron-flavored neutrinos because those are what the Sun produces, but the neutrinos spontaneously changed to the muon and tau flavors during their journey here and thereby evaded detection. Such flavor changes are known as neutrino oscillations, the phenomenon LSND and MiniBooNE were designed to study.

The Guitar Nebula

The Guitar Nebula was created when a massive star underwent a supernova explosion. The explosion left behind a neutron star, moving rapidly up and to the left in this image, elongating the surrounding material into a guitar-shaped cloud in its wake. Such powerfully-kicked neutron stars are common after supernovae, but it is unclear what causes the kick. One popular explanation involves a recoil from an asymmetrical release of sterile neutrinos. This particular neutron star is moving so quickly that it will eventually escape from the gravity of our Galaxy.

Similar flavor oscillations had already been observed in the decay of some quark-based particles and had been theorized for neutrinos. In 2001, Canada's Sudbury Neutrino Observatory (where Los Alamos's Van de Water was detector manager at the time) was able to observe muon- and tau-flavored neutrinos from the Sun, thus explaining the observed deficit of (electron-flavored) solar neutrinos. With that problem solved, it seemed that the only thing left for neutrino physicists to do was perform experiments to nail down the exact parameters for neutrino oscillations—how different are the masses of the mass states? how far must a neutrino travel before it oscillates? and how does that distance depend on the neutrino's surroundings? Louis and others working on LSND, however, discovered there was much more to be done.

The LSND experiment happened to employ a beam of antineutrinos rather than neutrinos. Most of the physics community assumed that this choice made no difference; an electron-flavored antineutrino would oscillate into a muon-flavored antineutrino just as a regular electron-flavored neutrino would. But when the LSND team measured the parameters associated with their antineutrino oscillations, they got wildly different results than other experiments had obtained for neutrino oscillations. "Since all the other experiments were consistent with each other—neutrinos from the Sun, from cosmic rays, from nuclear reactors, and from beams—a lot of people thought LSND was somehow mistaken," remembers Louis. Today, however, MiniBooNE appears to be confirming the LSND results: antineutrinos do not oscillate the same way as neutrinos do. That statement by itself is virtually revolutionary in the world of particle physics, but the novelty of the MiniBooNE results, and the LSND results they corroborate, runs much deeper.

Los Alamos physicists

Left to right, Los Alamos physicists Richard Van de Water, Bill Louis, and Geoff Mills bravely stand up to the hundred trillion neutrino particles the Sun sends through their bodies (and yours) every second.

Oscillating Interpretations

The equations governing neutrino oscillations depend on the mass states involved in a curious way. If we assume for simplicity that only two states are involved in the oscillation, then those two masses don't show up in the math explicitly; rather, it is the difference between the squares of the masses, Δm2, that appears in the relevant equations. Data from solar neutrinos, oscillating from electron to the muon and tau flavors, peg Δm2 for the associated mass states to about 0.00008 squared electronvolts, or eV2 (mass quoted in units of energy as justified by Einstein's equivalency E=mc2). And data from neutrinos spawned by cosmic ray interactions in the atmosphere, oscillating from muon to tau flavor, reveal Δm2 for that oscillation to be about 0.002 eV2. But LSND's oscillation data, from muon to electron flavor, is best fit with a much larger Δm2 of around 1 eV2.

If that 1-eV2 measurement is correct, then it begs the following question: If there are only three flavors—electron, muon, and tau—then how can the difference between any two of the masses, squared or otherwise, be greater than the total of the other two? To see this, consider an example: If the squared masses were 2, 5, and 9 in some units, then the differences between pairings would be 5 − 2 = 3, 9 − 5 = 4, and 9 − 2 = 7. The largest of these differences, 7, equals the other two added together (3 + 4), as it must. So even if all three states were involved in the LSND oscillations, how could the 1-eV2 measurement, which is far greater than the 0.002- and 0.00008-eV2 measurements for the other two flavor changes, be correct? This question has perplexed neutrino physicists since 1995. And the fact that LSND used antineutrinos rather than neutrinos does not resolve the issue, since antineutrino masses, and antiparticle masses in general, never differ from the corresponding particle masses when measured in other, non-oscillation phenomena.

MiniBooNE was designed to either confirm or refute LSND's too-large Δm2. As it turns out, it did both. First, using a beam of neutrinos (not antineutrinos), MiniBooNE found no oscillation from muon to electron flavor in the energy range of interest. This lack of oscillation agrees with the Δm2 measured for solar neutrinos, which only oscillate after much greater travel distances. That result was reported in 2007, but by 2010, after switching to a beam of antineutrinos, MiniBooNE obtained essentially the same result as LSND, with Δm2 between 0.1 to 1.0 eV2 (this range will tighten over time as more data is collected). This is still too large to be explained with only the three known flavors since it exceeds the sum of the other mass-square differences.

So what about proposing the existence of a fourth flavor? That would be bold. Since all the other quarks and leptons exist in only three known flavors, adding a new flavor, like declaring an observed asymmetry between neutrinos and antineutrinos, is not exactly a minor tweak to the standard model of particle physics. But the MiniBooNE results, taken in context with data from other particle colliders, are even more revolutionary than that.

Geoff Mills is a Los Alamos scientist working on the MiniBooNE collaboration with Louis and Van de Water. Prior to coming to Los Alamos, Mills researched properties of the weak nuclear force at the European CERN collider laboratory. One of these experiments measured how long the Z0 particle—a particle involved in communicating the weak force—could exist before it underwent radioactive decay. Because that particle lifetime depends on how many different sets of particles it can decay into, including neutrinos and antineutrinos, the CERN measurement revealed how many flavors of neutrinos exist. The result was exactly three: electron, muon, and tau.

The CERN experiment pertained to decays mediated by the weak force. Therefore it would be more accurate to say there can be only three "active" flavors of neutrino—ones that interact via the weak force. But taken together with the LSND and MiniBooNE results implying the need for more (and more massive) flavors to account for the large measured Δm2, neutrino physicists have been led to an unexpected conclusion.

cluster of galaxies photo

Most of the mass in this cluster of galaxies takes the form of dark matter, which is difficult to observe because it undergoes no electromagnetic or strong nuclear interactions. Weakly-interacting neutrinos could act like dark matter, but the known types of neutrino are too light to be responsible for the gravitational fields of galaxies and galaxy clusters. Sterile neutrinos, however, could be massive enough to account for the universe's dark matter. In this image, gravity is warping the appearance of background galaxies and causing them to appear like distorted arcs. The blue-purple region indicates where most of the dark matter is located in order to produce this effect.

As Mills explains, "If our antineutrino oscillation experiments can only be explained with one or more additional flavors, and we already know they can't be active, then evidently we need inactive, or sterile, neutrinos." Thus MiniBooNE, confirming results from LSND, points to three major advances in particle physics. First, it suggests a matter-antimatter asymmetry that had never been observed with leptons before (and had been observed only exceedingly rarely with quarks). Second, it implies the existence of a fourth flavor of neutrino. And third, it provides evidence for a new entity: a sterile neutrino. In fact, in order to accommodate MiniBooNE's different neutrino and antineutrino oscillation data, many researchers, including the Los Alamos team, conclude there must be at least two sterile neutrino flavors.

Murmurs of Approval

For now, the evidence for sterile neutrinos remains incomplete. The trouble isn't just the possibility of imposter particles creating false signals in the MiniBooNE detector; as with any new physics, there's also the possibility that other new phenomena are distorting the scientific inferences. For example, sterile neutrinos might in fact exist, but they might be unstable and decay in a way that tricks the detector. Or perhaps there's a new type of force altogether, rather than a new type of neutrino. Time will tell, as MiniBooNE continues to run antineutrino experiments for at least another year to improve the statistical significance of the results. The Los Alamos team also hopes to rearrange the MiniBooNE system to change the distance between the beam source and the detector, because if the number of electron-flavored antineutrinos observed varies with distance, that would be strong evidence that the effect is genuinely the result of an antineutrino oscillation with a Δm2 that requires one or more sterile flavors.

In the meantime, MiniBooNE and LSND are not alone. Among several experiments that observe neutrinos emerging from nuclear reactors in power plants, about 6 percent of the expected antineutrinos appear to be missing. This implies that they are oscillating into unseen flavors over relatively short distances, which should only happen with a large Δm2. These results from nuclear reactor experiments are consistent with MiniBooNE and LSND.

What if the new evidence for sterile neutrinos holds up? There are no obvious technological advances expected to stem from this new science, even though it's always possible that someday there could be. The rarely-interacting nature of neutrinos (and the never-interacting nature of sterile neutrinos!) limits their practical value because machines need to be enormous to capture just a few of them. Even the 12-meter-diameter MiniBooNE detector catches only one out of every trillion active neutrinos that are incident upon it, thus requiring years before it accumulates enough detections for a statistically significant result. But the new neutrino physics would make a real difference in certain astrophysical settings, where neutrinos' gravitational and inertial effects could be observable.

For example, sufficiently massive, noninteracting neutrinos could account for the universe's dark matter—invisible matter whose gravitational influence in galaxies and clusters of galaxies dwarfs that of the observable "normal" matter. And in the supernova that marks the death of a massive star and happens to involve a huge outpouring of neutrinos, any deviation from a perfect spherical explosion could send an excess of neutrinos, including sterile neutrinos, in one direction. If massive enough, sterile neutrinos could generate a recoil that sends the remaining stellar core—usually a neutron star—hurtling through space at high speeds, which has indeed been observed. Additionally, early in the big bang, when the universe was only a tiny fraction of a second old, extreme temperature and density conditions would have made neutrino interactions much more common than they are now. New physics describing how those interactions proceed could improve our understanding of how our universe evolved. The asymmetry between neutrino and antineutrino oscillations, for instance, could shed light on how our universe managed to allow matter, but not antimatter, to endure.

Back on Earth, Louis reminisces about the two decades he has spent investigating neutrino and antineutrino oscillations at Los Alamos National Laboratory. "If the evidence for sterile neutrinos holds up, I will count myself fortunate to have been involved in the discovery of such exotic physics." He pauses, then adds, "You know, even if it's not a sterile neutrino, we still know we've found something new which will be worth the effort to understand."

—Craig Tyler


In this issue...

Contact Us | Careers | Bradbury Science Museum | Emergencies | Inside LANL | Maps | Site Feedback | SSL Portal | Training

Operated by Los Alamos National Security, LLC for the U.S. Department of Energy's NNSA © Copyright 2014 LANS, LLC All rights reserved | Terms of Use | Privacy Policy