In the 1980s, the solar neutrino problem had been simmering for two decades. Raymond Davis’s chlorine experiment at Homestake was finding about a third of the predicted solar neutrinos. The Kamiokande water Cherenkov was finding about half. The community had several candidate explanations — wrong solar models, neutrino oscillation, or something more exotic — but no way to distinguish them cleanly.
A critical missing piece was a measurement of the pp neutrinos — the low-energy neutrinos emitted by the primary proton-proton fusion that powers the Sun. These dominate the total solar neutrino flux but have energies below 423 keV, well under the thresholds of all the existing experiments. Without a pp measurement, you could not even rule out the possibility that the Sun was simply burning at a different rate than the models assumed.
Two experiments were built specifically to fill that gap. GALLEX, in the Gran Sasso underground laboratory in Italy, started running in 1991 with about 30 tons of gallium chloride. SAGE, in the Baksan Neutrino Observatory in the Russian Caucasus, started somewhat earlier — in 1990 — with about 60 tons of metallic gallium. Together they pioneered the radiochemical detection of pp solar neutrinos, made some of the cleanest contributions to resolving the solar neutrino problem, and — to everyone’s surprise — produced a calibration anomaly that has lasted for over thirty years.
Why gallium
The chemistry that made these experiments possible is a single reaction. An electron neutrino interacts with a gallium-71 nucleus, converting it into germanium-71 and emitting an electron:
ν_e + ⁷¹Ga → ⁷¹Ge + e⁻
This reaction has a very low energy threshold — only about 233 keV — because gallium-71 to germanium-71 is a near-zero-Q-value nuclear transition. Solar pp neutrinos, with energies up to 423 keV, can comfortably trigger it. Chlorine experiments, by comparison, need about 814 keV. Water Cherenkov detectors need a few MeV to see anything at all. Gallium opened a window onto the dominant low-energy solar flux that no other technique could reach.
The germanium-71 produced is itself unstable, decaying back to gallium-71 with a half-life of 11.4 days, through electron capture. The decay leaves a characteristic signature — an X-ray and Auger-electron cascade in the few-keV range — that can be detected in a small proportional counter.
The experimental challenge is that the rate is roughly one germanium-71 atom produced per day in 30 tons of gallium. The experiment exposes the target for about a month, then has to extract those few dozen germanium atoms from tens of tons of gallium and count them. The extraction efficiency, chemistry, and counter performance have to be characterized to extraordinary precision.
How the extraction worked
GALLEX used liquid gallium chloride as its target. The solution sat in a 70-cubic-meter tank in the Gran Sasso underground laboratory. After about three weeks of exposure, the team blew nitrogen gas through the solution to sweep out the dissolved germanium tetrachloride. The extracted germanium was concentrated, converted to germane gas (GeH₄), and loaded into a miniaturized proportional counter for radioactive counting.
SAGE used metallic gallium, kept liquid at about 30°C. After exposure, the gallium was reacted with hydrochloric acid in a multi-stage chemical procedure to extract the trace germanium, which was then converted to germane and counted similarly to GALLEX.
The counters themselves were heroic engineering. Each contained a few cubic centimeters of germane gas at a precise pressure, with very low background. The detector arrays were located in dedicated underground rooms with extensive shielding from external radioactivity and active vetoes against cosmic rays.
A typical exposure might produce 10–20 germanium-71 atoms in 30 tons of gallium. After extraction efficiency losses, perhaps 60–80% of those would reach the counter. Each was tracked individually as it decayed over the following several months.
The solar neutrino result
Across all their runs, GALLEX and SAGE both measured a solar neutrino flux of roughly 55–65 SNU (Solar Neutrino Units, the rate per 10³⁶ target atoms per second), compared to the standard solar model prediction of about 130 SNU. The deficit was about 45%.
This was significantly less severe than the deficit at Homestake or Kamiokande. The pattern across all experiments — large deficit at high energies, smaller deficit at low energies — was a smoking gun for energy-dependent oscillation, consistent with the MSW effect operating in the dense solar interior. The pp neutrinos, lower in energy, do not cross the MSW resonance and oscillate roughly like in vacuum, producing the milder ~45% conversion that gallium experiments measured.
The eventual resolution of the solar neutrino problem came from SNO’s 2001 result, which directly observed flavor conversion for ⁸B solar neutrinos. But GALLEX and SAGE’s pp measurements were a critical input to the global analysis that pinned down the oscillation parameters.
The chromium calibration runs
To verify their measurement, GALLEX and SAGE both calibrated their detectors using artificial radioactive sources. The calibration relied on chromium-51, a radioactive isotope that emits an electron neutrino in a clean, monoenergetic decay (with energy near 750 keV). A strong chromium-51 source could be lowered into the gallium target, producing a known neutrino flux at the location of the experiment.
GALLEX ran two such calibrations in 1994 and 1995. SAGE ran two of its own in 1995 and 1996 using chromium-51 plus a 2001 calibration using argon-37. The exact strength of each source was carefully characterized through independent measurements.
The results were consistent and troubling. In every calibration run, the measured event rate was about 20% lower than expected from the calibrated source strength. The deficit appeared independent of which source was used, which experiment ran the calibration, and which year the runs took place. This became known as the gallium anomaly.
The simplest explanations available were either a systematic error in the gallium cross-section calculation (which would be a chemistry/nuclear-physics issue) or a real physical effect — possibly short-baseline oscillation into a sterile neutrino, into which a fraction of the chromium-source antineutrinos converted before reaching the gallium target.
The BEST follow-up
In 2019 the Baksan Experiment on Sterile Transitions (BEST) was launched as a direct follow-up. BEST is a SAGE-style gallium experiment with two concentric volumes of gallium, exposed to a single chromium-51 source placed at the center. The two volumes correspond to two different effective baselines (about 0.5 meters and 1 meter), so the experiment can directly compare the rate at two different distances from the source.
If a sterile neutrino exists with parameters that explain the gallium anomaly, the inner and outer volumes should see different rates from each other — a smoking-gun oscillation pattern.
The BEST results, published in 2022, strengthened the gallium anomaly significantly. Both volumes measured deficits, and the inner/outer ratio showed a hint (though not yet a definitive measurement) of distance dependence. The combined gallium-anomaly evidence is now at roughly the 5σ level when interpreted purely as a flavor-changing process.
This sits in tension with MicroBooNE and short-baseline reactor experiments, which have largely excluded the simplest sterile-neutrino interpretations at the parameters needed to explain LSND and MiniBooNE. The gallium anomaly’s preferred parameter space is similar but not identical, leaving theorists with a puzzle: a real and growing anomaly in gallium experiments that is hard to reconcile with the null results elsewhere.
The most-discussed resolutions now involve either updates to the gallium cross-section (which depends on contributions from excited states of germanium-71 that may have been miscalculated) or more elaborate sterile-neutrino scenarios. New cross-section measurements using germanium-71 produced through other means are underway.
After the original experiments
GALLEX completed its runs in 1997 and was succeeded by GNO (Gallium Neutrino Observatory), which ran from 1998 to 2003 using the same target and infrastructure. GNO refined the GALLEX measurement with more statistics.
SAGE continued operating until 2007, accumulating one of the longest-running solar neutrino datasets in physics.
After the experiments wound down, the gallium target was returned to industrial use. (Gallium is expensive and was rented from Russian and Western suppliers.) The radiochemistry techniques developed for GALLEX and SAGE — particularly the extraction of trace germanium from huge volumes of gallium — have informed multiple subsequent low-background experiments.
The legacy of patient chemistry
GALLEX and SAGE were among the most chemistry-heavy experiments in modern physics. They were not flashy. They did not produce big cascades of Cherenkov light. Their detectors did not look like cathedrals of glass and steel. They were warehouse-sized chemical processing plants in underground laboratories, churning through metric tons of gallium to extract a handful of atoms.
But the results they produced were essential. The pp-neutrino measurement closed the energy range that no other experiment could touch. The energy-dependent deficit pattern across all solar experiments — when GALLEX and SAGE numbers were combined with Homestake, Kamiokande, and later SNO and Borexino — was the fingerprint that confirmed MSW-driven flavor conversion. And the calibration anomaly they uncovered is still, three decades later, one of the most actively studied puzzles in neutrino physics.
It is sometimes worth remembering that some of the cleanest results in modern physics came not from large optical detectors but from the patient counting of a few radioactive atoms produced by neutrinos in tons of liquid metal.
For the broader solar neutrino story, see The solar neutrino problem and SNO. For the closely-related Borexino program, see Borexino. For why energy-dependent oscillation is what we see, see The MSW effect. For the sterile-neutrino context, see MicroBooNE and The reactor antineutrino anomaly.
Frequently asked
What were GALLEX and SAGE?
GALLEX (Gran Sasso, Italy) and SAGE (Baksan Neutrino Observatory, Russia) were two solar neutrino experiments based on radiochemistry — specifically the reaction by which an electron neutrino converts a gallium-71 nucleus into germanium-71. Both ran from the early 1990s through the 2000s and were the first detectors sensitive to the low-energy pp solar neutrinos, the dominant component of the solar flux.
Why was gallium the right target?
Because the neutrino-induced conversion of gallium-71 to germanium-71 has a very low energy threshold — about 233 keV — far lower than any other practical neutrino target. This made gallium uniquely sensitive to the pp neutrinos produced in the Sun's primary fusion reaction, whose energies extend up to only about 423 keV. Other historic solar neutrino targets (chlorine, water) had energy thresholds too high to catch pp neutrinos at all.
How does radiochemistry work for neutrino detection?
After a long exposure (typically a month), the experiment chemically extracts the few atoms of germanium-71 that solar neutrinos have produced from the tens of tons of gallium target. The germanium atoms are concentrated, purified, and placed in low-background counters where their characteristic radioactive decay back to gallium-71 is recorded. The total number of decays measures the integrated solar neutrino flux during the exposure.
What did GALLEX and SAGE find?
Both experiments measured solar neutrino rates roughly 55–60% of the standard solar model prediction. The deficit was real but smaller than at higher-energy experiments like Homestake and Kamiokande, consistent with a vacuum-like oscillation at low energies. The energy-dependent deficit pattern across all solar experiments was a key fingerprint of MSW-driven flavor conversion, eventually confirmed by SNO.
What is the gallium anomaly?
Both experiments were calibrated using artificial radioactive chromium-51 sources placed inside the gallium target. The calibration consistently produced about 20% fewer events than expected from the source strength. The discrepancy has lasted for three decades and has been suggested as evidence for short-baseline sterile-neutrino oscillation, though alternative explanations involving uncertainties in the gallium cross-section also remain on the table. Recent gallium-based experiments like BEST have sharpened the anomaly significantly.
Cite this article 5 formats
APA
Neutrino Times Editorial Team. (2025, October 31). GALLEX and SAGE: how chemistry caught the Sun's lowest-energy neutrinos. Neutrino Times. https://neutrino-times.com/articles/gallex-sage-gallium-solar-neutrino-experiments/
Chicago
Neutrino Times Editorial Team. "GALLEX and SAGE: how chemistry caught the Sun's lowest-energy neutrinos." Neutrino Times, October 31, 2025. https://neutrino-times.com/articles/gallex-sage-gallium-solar-neutrino-experiments/.
MLA
Neutrino Times Editorial Team. "GALLEX and SAGE: how chemistry caught the Sun's lowest-energy neutrinos." Neutrino Times, 31 Oct. 2025, https://neutrino-times.com/articles/gallex-sage-gallium-solar-neutrino-experiments/.
BibTeX
@misc{neutrino-times-gallex-sage-gallium-solar-neutrino-experiments,
author = {Neutrino Times Editorial Team},
title = {GALLEX and SAGE: how chemistry caught the Sun's lowest-energy neutrinos},
howpublished = {Neutrino Times},
year = {2025},
month = {oct},
url = {https://neutrino-times.com/articles/gallex-sage-gallium-solar-neutrino-experiments/},
note = {Accessed: 2025-10-31}
} RIS
TY - GEN TI - GALLEX and SAGE: how chemistry caught the Sun's lowest-energy neutrinos AU - Neutrino Times Editorial Team PY - 2025 DA - 2025-10-31 PB - Neutrino Times UR - https://neutrino-times.com/articles/gallex-sage-gallium-solar-neutrino-experiments/ ER -