The History of the Neutrino: From Pauli's Desperate Remedy to Coherent Scattering

The history of the neutrino is unusual among discovery stories because the particle was named, given a quantitative theory, and declared undetectable in print — all before anyone had seen one. Twenty-two years separated the calculation that said detection was impossible from the experiment that achieved it. Those twenty-two years are the most instructive part of the chronology, and the arithmetic behind them is simple enough to check on paper.

What follows is that chronology, from the experimental anomaly that created the problem in 1914 to the coherent-scattering measurement of 2017 and the open questions of today. The numbers are given explicitly throughout, because in this subject the numbers are the argument.

The Crisis in the Beta Spectrum

In 1914 James Chadwick, working in Berlin with a primitive form of what would become the Geiger counter, established that the electrons emitted in beta decay carry a continuous range of energies rather than a single sharp value.

The trouble this caused is a two-line calculation. Consider a nucleus of mass $M$ at rest decaying into a daughter of mass $M'$ and an electron of mass $m$. Conservation of energy and momentum in the rest frame gives two equations,

$M c^2 = E' + E_e, \qquad \mathbf{p}' + \mathbf{p}_e = 0,$

and with $E^2 = p^2c^2 + m^2c^4$ for each product these determine the electron energy uniquely:

$E_e = \frac{M^2 + m^2 - M'^2}{2M}\,c^2.$

There is no freedom left. A two-body decay from a state of definite energy produces a monoenergetic electron, full stop. The observed spectrum was not a line; it was a broad continuum running from near zero up to a definite endpoint.

For over a decade it was possible to hope that the continuum was an artefact — that electrons left the nucleus monoenergetic and lost variable amounts on the way out. Charles Drummond Ellis and William Wooster closed that escape in 1927, enclosing a radium E (bismuth-210) source in a calorimeter thick enough to absorb everything and measuring the heat. If the electrons were emitted at the spectrum endpoint of about 1.05 MeV and degraded afterwards, the calorimeter would still register 1.05 MeV per decay. It registered $344 \pm 40$ keV — the mean of the observed spectrum, not its endpoint. Lise Meitner and Wilhelm Orthmann repeated the experiment in Berlin in 1930 with a better calorimeter and obtained $337 \pm 20$ keV. The energy was genuinely missing.

The severity of the crisis is best measured by what Niels Bohr was prepared to give up. Rather than invent a new particle, Bohr suggested that energy conservation might hold only statistically in nuclear processes — true on average, violated event by event. Coming from Bohr, in 1930, this was not a fringe position. It indicates how far the community was willing to go.

Pauli’s Desperate Remedy: 4 December 1930

Wolfgang Pauli chose the other horn. On 4 December 1930 he wrote from Zurich to a meeting of radioactivity researchers of the regional section — the Gauverein — convening in Tübingen. He did not attend; the letter was carried by a colleague who was to supply further detail in person. Its opening address, Liebe Radioaktive Damen und Herren, has been quoted ever since.

Pauli proposed that, because of the wrong statistics of the nitrogen and lithium-6 nuclei and the continuous beta spectrum, he had hit upon what he called a desperate remedy to save the exchange theorem of statistics and the law of conservation of energy: electrically neutral particles inside nuclei, which he called neutrons, with spin one-half, obeying the exclusion principle, and not travelling at the speed of light. Their mass should be of the order of the electron mass and in any case no larger than 0.01 proton masses. The continuous beta spectrum would then make sense, he wrote, if in beta decay a neutron is emitted alongside the electron such that the sum of the two energies is constant.

Two features of the letter are routinely lost in retelling. First, Pauli did not publish. He wrote that he did not yet dare to publish anything about the idea and was turning first to the experimentalists, asking how likely it was that such a particle could be found. Second, he was explicit about the weakness of his own proposal: he admitted the remedy might seem almost improbable, because one would probably have seen such particles long ago if they existed. He closed by apologising for his absence — he was indispensable in Zurich on account of a ball on the night of 6 to 7 December.

The remark most often attached to Pauli, that he had done a terrible thing by postulating a particle that cannot be detected, is not in the letter. Its documented source is Frederick Reines, who quoted it in his foreword to Christine Sutton’s Spaceship Neutrino (1992), decades after the fact. It is consistent with Pauli’s own written reservations, but it is a recollection rather than a contemporaneous record, and it should be cited that way.

The physics of the proposal is worth stating cleanly. Add a third product of mass $m_\nu$ and the kinematics change character entirely. The available energy $Q$ is now shared between two light products, and the electron energy is distributed continuously over

$m_e c^2 \le E_e \le Q + m_e c^2 - m_\nu c^2 .$

A three-body final state has a two-dimensional phase space, so the electron energy is a random variable with a smooth density rather than a fixed number. The shape of that density near the endpoint depends on $m_\nu$ — which is precisely how the neutrino mass is measured today.

Fermi Gives It a Name and a Theory

Chadwick discovered the actual neutron in 1932, and the name Pauli had used became unavailable. In Rome, in conversation at the Via Panisperna institute, Edoardo Amaldi coined neutrino — a grammatically playful Italian diminutive, the little neutral one — to distinguish Pauli’s light particle from Chadwick’s heavy one. Enrico Fermi adopted it, and it passed into international use between the Paris conference of 1932 and the Solvay conference of October 1933.

Fermi then did something much harder than naming. In 1933–34 he wrote down a quantitative theory of beta decay in which the electron and neutrino are created at the moment of decay, by analogy with the emission of a photon. The interaction is a product of two currents evaluated at the same point,

$\mathcal{H}_{\text{int}} = G_F \left( \bar{\psi}_p \Gamma \psi_n \right)\left( \bar{\psi}_e \Gamma \psi_\nu \right),$

with a single coupling constant $G_F$ to be fixed by experiment. The structure $\Gamma$ was unknown; the candidates — scalar, vector, tensor, axial-vector, pseudoscalar — are exactly the bilinear covariants one can build from Dirac spinors, a classification belonging to the tensor formalism. Two further decades of experiment were needed to settle on vector-minus-axial-vector, in 1957. Fermi’s theory predicted the shape of the beta spectrum and its dependence on nuclear charge, and it worked.

Nature rejected the paper. The stated reason, widely quoted, was that it contained speculations too remote from reality to be of interest to the reader. Fermi published in Italian in La Ricerca Scientifica and Il Nuovo Cimento, and in German in Zeitschrift für Physik. The rejection is well attested. A frequently appended claim — that Nature later acknowledged it as one of the great editorial blunders in its history — is much weaker; Fermi’s biographer David N. Schwartz has argued that it is both unproven and unlikely. The rejection happened; the apology is folklore.

Bethe and Peierls: The Arithmetic of Invisibility

With a coupling constant in hand, the cross-section became computable. In 1934 Hans Bethe and Rudolf Peierls published a short note in Nature doing exactly that, and reached a bound of

$\sigma < 10^{-44}\,\mathrm{cm}^2 .$

Their conclusion, in print, was that there is no practically possible way of observing the neutrino.

Here is the arithmetic that justified that verdict. The mean free path of a particle in a uniform medium is

$\lambda = \frac{1}{n\sigma},$

where $n$ is the number density of targets. For nucleon targets in a solid of mass density $\rho$ expressed in grams per cubic centimetre, the nucleon number density is approximately $n \approx \rho N_A$, since the molar mass in grams is close to the nucleon count. For lead, $\rho = 11.35$, so

$n \approx 11.35 \times 6.022\times10^{23} \approx 6.8\times10^{24}\ \mathrm{cm^{-3}} .$

With $\sigma = 10^{-44}\,\mathrm{cm}^2$,

$\lambda = \frac{1}{6.8\times10^{24} \times 10^{-44}} = \frac{1}{6.8\times10^{-20}} \approx 1.5\times10^{19}\ \mathrm{cm}.$

That is $1.5\times10^{14}$ kilometres, which is the figure Bethe and Peierls quoted, and it is about 15 light-years. A neutrino fired into a block of lead would on average cross fifteen light-years of it before interacting once.

Water gives a larger number still, because inverse beta decay requires free protons and water has only two per molecule. With density 1 and molar mass 18, the molecular density is $3.35\times10^{22}\ \mathrm{cm^{-3}}$, so the free-proton density is $n_p \approx 6.7\times10^{22}\ \mathrm{cm^{-3}}$ and

$\lambda = \frac{1}{6.7\times10^{22}\times10^{-44}} \approx 1.5\times10^{21}\ \mathrm{cm} \approx 1.6\times10^{3}\ \text{light-years},$

using $1$ light-year $= 9.46\times10^{17}\,\mathrm{cm}$. This is the origin of the familiar phrase about a thousand light-years of water.

Note what the calculation does not say. A mean free path of fifteen light-years in lead does not mean nothing ever happens; it means the interaction probability per unit length is $n\sigma$, and if you cannot increase $\lambda^{-1}$ you can still increase the number of incident particles. Bethe and Peierls had no reason in 1934 to imagine a source emitting $10^{20}$ neutrinos per second. Nuclear reactors did not exist.

Project Poltergeist: Detection in 1956

Frederick Reines and Clyde Cowan supplied the missing factor. Their first plan, considered seriously at Los Alamos, was to place a detector near a nuclear explosion; they abandoned it for a fission reactor, which offers a steady flux rather than a single burst.

The signature they used was inverse beta decay,

$\bar{\nu}_e + p \rightarrow n + e^+ ,$

and its elegance is that it produces two events separated by a known delay. The positron annihilates almost immediately, giving two 511 keV photons in coincidence. The neutron thermalises over some microseconds and is then captured on cadmium dissolved in the water, releasing a characteristic gamma cascade. A prompt double pulse followed a few microseconds later by a capture pulse is a pattern that radioactive background does not easily counterfeit. The detector was a sandwich: tanks of liquid scintillator viewed by photomultipliers, alternating with water tanks loaded with cadmium chloride, about ten tonnes in total.

In late 1955 the apparatus went to the Savannah River Plant in South Carolina, sited near a high-power production reactor delivering an antineutrino flux of order $10^{13}\ \mathrm{cm^{-2}\,s^{-1}}$. They took data for more than five months. The signal rate was about three events per hour, against substantial background.

On 14 June 1956 they sent a telegram to Pauli in Zurich, reporting that they had definitely detected neutrinos from fission fragments by observing inverse beta decay of protons. Pauli was at a meeting at CERN; the telegram was forwarded, and he read it out to the room.

The measured cross-section came out close to prediction. The 1959 publication reports a total of $(11 \pm 4)\times10^{-44}\ \mathrm{cm}^2$ against an expected value near $9.5\times10^{-44}\ \mathrm{cm}^2$; other summaries quote the earlier prediction of about $6\times10^{-44}\ \mathrm{cm}^2$. The values shifted as the theory of the weak interaction was refined during exactly those years, which is why sources differ. The order of magnitude never moved. Bethe and Peierls had the number right; they were wrong only about what could be built.

Reines received half of the 1995 Nobel Prize in Physics, sharing with Martin Perl. Cowan had died in 1974.

The Missing Solar Neutrinos

The next chapter began with a calculation rather than an anomaly. John Bahcall computed the neutrino output of the Sun from the standard solar model, and Raymond Davis built an experiment to measure it: roughly 615 tonnes of perchloroethylene, 4,850 feet underground in the Homestake gold mine in South Dakota, exploiting

$\nu_e + {}^{37}\mathrm{Cl} \rightarrow {}^{37}\mathrm{Ar} + e^- .$

Davis extracted the argon atoms chemically and counted their decays. Over twenty-five years he counted about 2,200 of them.

The first results, published in 1968, gave an upper limit near 3 SNU against a standard-solar-model prediction of $7.5 \pm 3$ SNU. The long-run average settled at $2.56 \pm 0.16 \pm 0.16$ SNU. Roughly one third of the expected flux was arriving.

This was a genuine crisis with exactly two escape routes. Either the solar model was wrong — the ${}^{8}\mathrm{B}$ neutrino rate depends on the core temperature to something like the eighteenth power, so a few per cent error in $T$ moves the prediction substantially — or something happened to the neutrinos between the core and the detector. For three decades it could not be settled, because Homestake was sensitive only to electron neutrinos: if the neutrinos were changing flavour, a chlorine detector could not distinguish fewer neutrinos from different ones. The experiment was, by construction, blind to the answer.

Oscillation, and Therefore Mass

The resolution came from two detectors that could see what Homestake could not.

At the Neutrino 1998 conference in Takayama, Takaaki Kajita presented Super-Kamiokande’s atmospheric neutrino data: 4,654 events over 535 days, showing a zenith-angle dependence in the muon-like sample. Neutrinos produced in the atmosphere directly overhead arrived in the expected numbers; those produced on the far side of the Earth, having travelled thousands of kilometres further, arrived at roughly half strength. The deficit depended on path length. The reported significance was 6.2 standard deviations — a figure whose interpretation is worth understanding properly, and which is not the probability that the result is real; see the discussion of p-values.

The Sudbury Neutrino Observatory closed the case on the solar side. Its 1,000 tonnes of heavy water, two kilometres underground in Ontario, allowed three distinct reaction channels: a charged-current reaction sensitive only to $\nu_e$, elastic scattering with mixed sensitivity, and a neutral-current reaction on deuterium sensitive equally to all three flavours. The neutral-current channel measures the total flux regardless of flavour. Results released in 2001 and 2002 gave

$\phi_{\mathrm{NC}} = 5.09\,\substack{+0.44\\-0.43}\,(\text{stat})\,\substack{+0.46\\-0.43}\,(\text{syst}) \times 10^{6}\ \mathrm{cm^{-2}\,s^{-1}},$

in agreement with the solar model, while the electron-neutrino component alone was $\phi_e = 1.76 \times 10^{6}\ \mathrm{cm^{-2}\,s^{-1}}$. The non-electron component was $3.41\times10^{6}\ \mathrm{cm^{-2}\,s^{-1}}$, more than five standard deviations from zero. Bahcall’s Sun was right. The neutrinos had changed flavour. Kajita and Arthur McDonald shared the 2015 Nobel Prize.

The consequence for particle physics is structural. Oscillation arises because the flavour states produced by the weak interaction are not the states of definite mass that propagate. Writing $U$ for the unitary PMNS mixing matrix,

$|\nu_\alpha\rangle = \sum_{i} U_{\alpha i}^{*} |\nu_i\rangle ,$

each mass eigenstate acquires a phase $e^{-iE_i t}$ in flight. The two-flavour survival probability makes the requirement explicit:

$P(\nu_\alpha \to \nu_\alpha) = 1 - \sin^2(2\theta)\,\sin^2\!\left( \frac{\Delta m^2 L}{4E} \right).$

If $\Delta m^2 = m_2^2 - m_1^2$ vanishes the oscillating factor is identically zero and nothing happens. Oscillation therefore requires non-degenerate masses, which requires at least one non-zero mass. The Standard Model as originally written had none. The full three-flavour treatment — the structure of $U$, why it must be unitary, and where the CP-violating phase enters — is a self-contained piece of linear algebra worked through in how matrices describe neutrino oscillations, and it is the natural companion to this article. The underlying notion of eigenvalues and eigenvectors is doing all the work: flavour and mass are two bases for the same space, and the dynamics are diagonal in only one of them.

Coherent Scattering and the Modern Landscape

In 1974 Daniel Freedman pointed out a consequence of the newly established weak neutral current. If a neutrino’s wavelength is long compared with the size of a nucleus, it scatters off the nucleus as a whole rather than off individual nucleons, and the amplitudes add coherently. The cross-section then scales as the square of the weak charge, dominated by the neutron number $N$:

$\sigma_{\mathrm{CEvNS}} \propto N^2 E_\nu^2 .$

The neutral current is itself a prediction of the electroweak gauge structure — the symmetry and group theory that organises the Standard Model, built on the Lie group $SU(2)_L \times U(1)_Y$.

Freedman’s process resisted detection for 43 years, because the only observable is a nuclear recoil of a few keV. The COHERENT collaboration reported the first measurement in 2017, using a sodium-doped caesium iodide crystal of 14.57 kg viewed by a single photomultiplier, placed 19.6 metres from the target of the Spallation Neutron Source at Oak Ridge. The analysis found $134 \pm 22$ events against a Standard Model expectation of $178 \pm 43$. Later CsI analyses give a flux-averaged cross-section of about $1.65\times10^{-38}\ \mathrm{cm}^2$ per nucleus at these energies.

That is three to four orders of magnitude above inverse beta decay per target. It is a large gain in rate. It is not a large gain in energy, because each event deposits only a keV-scale recoil, and the $E_\nu^2$ scaling means the advantage shrinks rapidly at lower neutrino energies.

The current landscape is broad. IceCube, a cubic kilometre of instrumented Antarctic ice, reported the first evidence for a high-energy astrophysical neutrino flux in 2013 and now does astronomy with it. JUNO in China began operation in August 2025 and within 59 days of effective data had measured the solar oscillation parameters $\theta_{12}$ and $\Delta m^2_{21}$ to a precision about 1.6 times better than all previous experiments combined. DUNE, sending a beam 1,300 kilometres from Fermilab to a liquid-argon detector in South Dakota, begins its staged data-taking from 2026. KATRIN, measuring the tritium beta spectrum near its endpoint, reported in 2025 an upper limit on the effective electron-neutrino mass of 0.45 eV at 90 per cent confidence, from 259 days of data and some 36 million electrons.

Four questions remain open: the mass ordering, whether the neutrino is its own antiparticle, whether there is CP violation in the lepton sector, and the absolute mass scale.

The Chronology in One Table

Year	Event	Why it mattered
1914	Chadwick measures a continuous beta spectrum	Two-body kinematics predicts a sharp line; the continuum is unexplained
1927	Ellis and Wooster calorimetry on radium E	344 keV mean, not the 1.05 MeV endpoint: energy genuinely missing
1930	Bohr proposes statistical energy conservation	Shows how deep the crisis ran
1930	Pauli’s letter of 4 December to Tübingen	A neutral spin-half particle as a desperate remedy
1932	Amaldi coins the name neutrino in Rome	Chadwick’s neutron had taken Pauli’s original name
1933-34	Fermi’s theory of beta decay	A coupling constant, a predicted spectrum shape, a rejection by Nature
1934	Bethe and Peierls compute the cross-section	Bound below ten to the minus forty-four square centimetres; detection declared impossible
1956	Cowan and Reines at Savannah River	Inverse beta decay observed; roughly three events per hour
1968	First Homestake results	About one third of the predicted solar flux: the solar neutrino problem
1974	Freedman predicts coherent nuclear scattering	A far larger cross-section, but only keV recoils
1998	Super-Kamiokande atmospheric result	Path-length-dependent deficit: oscillation, hence mass
2001-02	SNO neutral-current measurement	Total flux matches the Sun; the missing neutrinos changed flavour
2017	COHERENT observes CEvNS	Freedman’s process confirmed after 43 years
2025	KATRIN limit of 0.45 eV	Tightest direct bound on the absolute mass scale

From Detection to Utilisation

Everything above is, read one way, a single sustained demonstration of how weakly neutrinos couple to matter. That is the relevant background to any proposal to draw usable power from them, and it is worth doing the estimate explicitly rather than asserting a conclusion.

The total solar neutrino flux at Earth is about $6.5\times10^{10}\ \mathrm{cm^{-2}\,s^{-1}}$, over 91 per cent of it from the pp chain at sub-MeV energies. That number is genuinely enormous, and the corresponding energy flux is not negligible either: with a mean energy of order 0.3 MeV, the arriving power density is roughly

$6.5\times10^{14}\ \mathrm{m^{-2}s^{-1}} \times 0.3\ \mathrm{MeV} \approx 6.5\times10^{14} \times 4.8\times10^{-14}\ \mathrm{J} \approx 31\ \mathrm{W\,m^{-2}},$

about two per cent of the solar constant. The resource exists. The question is entirely one of coupling.

Take a thin film: one square metre of area, one micrometre thick, so a volume of $1\ \mathrm{cm}^3$. Give it a density of $5\ \mathrm{g\,cm^{-3}}$, so a nucleon density $n \approx 3.0\times10^{24}\ \mathrm{cm^{-3}}$. The interaction probability per incident neutrino is

$P = n\,\sigma\,L = 3.0\times10^{24} \times 10^{-44} \times 10^{-4} = 3\times10^{-24},$

taking $\sigma = 10^{-44}\,\mathrm{cm}^2$, which is generous, since solar neutrinos sit below the MeV scale where that figure applies. The event rate is

$R = \Phi\,\sigma\,N_t = 6.5\times10^{10}\times10^{-44}\times3.0\times10^{24} \approx 2\times10^{-9}\ \mathrm{s^{-1}},$

one interaction roughly every sixteen years. If every interaction deposited the entire neutrino energy — it would not — the extracted power would be

$2\times10^{-9}\ \mathrm{s^{-1}} \times 4.8\times10^{-14}\ \mathrm{J} \approx 10^{-22}\ \mathrm{W}$

per square metre. Equivalently, $31\ \mathrm{W\,m^{-2}} \times 3\times10^{-24}$. Thickening the film to a full metre of the same material raises the intercepted fraction to $3\times10^{-18}$ and the power to about $10^{-16}\ \mathrm{W\,m^{-2}}$. Substituting CEvNS does not change the picture qualitatively: the cross-section per nucleus rises by three to four orders of magnitude, but the energy deposited per event falls from MeV to keV, and the $E_\nu^2$ scaling suppresses the gain badly at solar energies.

These are order-of-magnitude figures and the reader should check them. Every input is stated.

Against this background, it is worth representing the relevant commercial claim accurately. The Neutrino Energy Group, founded in 2008 and led by Holger Thorsten Schubart, proposes a multi-channel ambient-energy conversion framework. Its own published documentation does not present neutrinos as the sole input; it describes them as one contribution among several, alongside thermal and electromagnetic channels. It is also explicit about what it declines to assert: it does not claim a power density, and states that internal mathematical consistency is not validated commercial performance, and that the achievable coupling efficiency is a question for measurement rather than argument. The structure of that framework — a bounded efficiency multiplied by a volume integral of an effective flux against an effective cross-section — is examined in the Schubart–NEG master equation, and the dimensional check on it, along with what it does and does not establish, in mathematics beats opinion.

What would settle the matter is not in dispute and is easy to state: an independently replicated measurement of power output, with the energy balance closed and the input channels separated, so that whatever is delivered can be attributed to a specific source. No such result appears in the peer-reviewed literature to date. That is a statement about the current evidentiary record, not a prediction about what future measurements will show.

The historical parallel is available, and it should be drawn honestly in both directions. Bethe and Peierls were wrong. Their cross-section was right to within a factor of a few, their conclusion that no practically possible observation existed was not, and the counterexample arrived twenty-two years later. But it arrived in a specific form. It was not an argument, an appeal to the limits of present knowledge, or a claim that confident impossibility statements have failed before. It was ten tonnes of scintillator next to a reactor, a delayed-coincidence signature designed to defeat background, five months of data, and a measured cross-section that could be compared against a published prediction. Overturning a well-founded impossibility claim is possible. It requires an instrument and a measurement.

---

Related topics: How Matrices Describe Neutrino Oscillations, Eigenvalues and Eigenvectors, Symmetry and Group Theory, Lie Groups, Tensors, P-Values Explained, The Schubart–NEG Master Equation, Mathematics Beats Opinion

Frequently asked

Who discovered the neutrino?

The particle was proposed by Wolfgang Pauli in a letter dated 4 December 1930 and given a working theory by Enrico Fermi in 1933-34. It was detected experimentally by Clyde Cowan and Frederick Reines at the Savannah River reactor in 1956. Reines received the Nobel Prize in Physics in 1995; Cowan had died in 1974 and the prize is not awarded posthumously. So the answer depends on what is meant by discovery — prediction, theory, or detection each belongs to different people.

Why did Pauli propose the neutrino?

Beta decay emits electrons with a continuous range of energies. A decay with exactly two products has a fixed energy for each, set by conservation of energy and momentum, so the spectrum should be a sharp line. A continuous spectrum therefore requires a third, unseen product sharing the energy. Pauli also needed the new particle to fix a spin-statistics problem in the nitrogen-14 and lithium-6 nuclei. He called his proposal a desperate remedy and did not publish it at the time.

Why are neutrinos so hard to detect?

Because the cross-section is tiny. At megaelectronvolt energies it is roughly ten to the minus forty-four square centimetres. Putting that into the mean free path formula for lead gives a path length of order ten to the fourteen kilometres, about fifteen light-years. Bethe and Peierls computed this in 1934 and concluded there was no practically possible way to observe the particle. Detection took a nuclear reactor, a ten-tonne detector, and a delayed-coincidence signature that could be separated from background.

Do neutrinos have mass?

Yes. Neutrino oscillation, confirmed by Super-Kamiokande in 1998 and by the Sudbury Neutrino Observatory in 2001 and 2002, requires that at least two of the three mass eigenstates have different masses, which in turn requires that at least two are non-zero. The Standard Model as originally formulated had massless neutrinos, so this is a confirmed departure from it. The absolute scale is still unknown; KATRIN reported an upper limit of 0.45 electronvolts in 2025.

What are neutrinos used for?

Chiefly as messengers. IceCube uses high-energy astrophysical neutrinos to observe sources that light cannot escape from, reactor experiments use them to measure oscillation parameters, and supernova neutrinos arrive hours before the visible light. Reactor monitoring for safeguards purposes is an active engineering application. Extracting usable power from the ambient neutrino flux is a separate question and, as the cross-section arithmetic in this article shows, a very different one.

Continue reading

Blog

How Matrices Describe Neutrino Oscillations

Neutrinos have three 'flavor' identities but oscillate between them as they travel through space. The mathematics behind that behavior is a surprisingly elegant piece of linear algebra.

Read →Blog

What Does a 50% Increased Risk Actually Mean?

A 50% increased risk means nothing without the baseline. On a 2-in-10,000 risk it is 3 in 10,000 — one extra person. How to convert relative risk into absolute numbers.

Read →Blog

Percentage Points vs Percentage Change: The Difference That Moves Markets

Percentage points vs percentage change explained: a rise from 4% to 5% is one percentage point but a 25% relative increase. Formulas, basis points, worked examples.

Read →