Theorists are quite sure that the proton—proton chain operates in the Sun. However, because the gamma-ray energy released by the nuclear reactions just described is transformed into visible and infrared radiation by the time it emerges from the Sun, astronomers have no direct electromagnetic evidence of conditions in the solar core. However, the neutrinos that arise as by-products of the proton—proton chain do escape from the Sun. Interacting with virtually nothing, they leave at or near the speed of light, escaping into space a few seconds after being created in the core. They offer, at least in principle, the possibility of probing directly the conditions at the heart of the Sun.

Of course, the fact that they can pass through the entire Sun without interacting also makes neutrinos very difficult to detect on Earth. Nevertheless, they do interact a little more strongly with some elements—chlorine and gallium, for example—than with others, and this knowledge can be used in the construction of Earth-based neutrino-detection devices. Occasionally, a neutrino from the Sun will encounter a chlorine-37 nucleus, converting it into a nucleus of argon-37, or one will interact with a nucleus of gallium-31, turning it into germanium-31.

Figure 16.27 shows a photograph of the first experiment designed to detect solar neutrinos. Built in the late 1960s by a team of researchers from Brookhaven National Laboratory, it consisted of a large tank filled with 400,000 liters (about 100,000 gallons) of a chlorine-containing chemical—the common cleaning fluid used by dry cleaners. It was sited near the bottom of the Homestake gold mine in South Dakota. At 1.5 km below ground level, the experimenters could be reasonably sure of avoiding interference from other sources, as most subatomic particles are unable to penetrate Earth's crust to such a depth. They left their tank in the mine for months at a time, periodically checking to see if any of the chlorine had been converted into argon, which would signal the absorption of a neutrino.

Figure 16.27 This swimming-pool-sized detector was a "neutrino telescope" of sorts, buried underground in a South Dakota gold mine.

Given the size of the detector and the physical conditions in the Sun's core implied by the Standard Solar Model, about one solar neutrino of the roughly 10¹⁶ that streamed through the tank each day should have been detected. The experiment did succeed in detecting solar neutrinos—in itself a remarkable achievement—but the numbers were not as great as predicted. Over the course of the entire experiment, neutrinos were detected about twice per week, on average, not once per day. The neutrino deficit persisted over two decades of almost continuous monitoring (until the experiment was terminated in 1993), and though scores of technicians examined every facet of the Homestake instrument for instrumental flaws, they found none. This disagreement between theory and observation is known as the solar neutrino problem.

Other, more recent experiments, having quite different detector designs, have also found significantly fewer neutrinos than predicted by theory. Researchers using a detector located in Kamioka, Japan, reported only about 45 percent of the expected number of neutrinos, and two more sensitive experiments—the Soviet—American Gallium Experiment (SAGE, for short) and the U.S.—European GALLEX collaboration—each using the element gallium to capture solar neutrinos, find a roughly 50 percent shortfall. The Homestake and Kamioka experiments were actually sensitive not to neutrinos produced in reaction (I) but to those created by a much less probable sequence of events, occurring only about 0.25 percent of the time, leaving some room for theorists to maneuver in their attempts to explain the observations. However, SAGE and GALLEX can detect neutrinos produced by reaction (I)—the initial step in the proton—proton chain—so they provide a much more direct probe of energy generation in the solar core.

The four neutrino-detection experiments just described disagree somewhat in their measurements of the precise extent of the deficit, but each of them sees significantly fewer than the expected number of solar neutrinos. The inescapable conclusion is that although solar neutrinos are observed (and in fact their measured energies do lie in the range predicted by the Standard Solar Model), there is a real discrepancy between the Sun's theoretical neutrino output and the neutrinos we detect on Earth. How can we explain this contradiction? If, as we think, the detectors are working correctly, there are really only two possibilities. Either neutrinos are not produced as frequently as we think, or not all of them make it to Earth. Let us now consider these alternatives in turn.

If the temperature in the solar core were lower, then the number of neutrinos predicted by theory would also be lower. If the center of the Sun were about 10 percent cooler than in the Standard Solar Model—about 13,500,000 K—the neutrino emission would be reduced to a level consistent with observations. But lowering the temperature would also lower the Sun's luminosity, and most theorists agree that the numerical models could not be in error by as much as 1,500,000 K while remaining consistent with all other solar observations. In addition, observations by the GONG group (discussed earlier) seem to rule out a central temperature below 15,000,000 K. Most astronomers regard it as quite unlikely that the resolution of the solar neutrino problem will be found in the nuclear physics of the Sun's interior.

Instead, the properties of the neutrinos themselves may provide the answer. If neutrinos do have a minute amount of mass, it may be possible for them to change their properties, even to transform into other particles, during their 8-minute flight from the solar core to Earth, through a process generally known as neutrino oscillations. In this picture, neutrinos are produced in the Sun at the rate required by the Standard Solar Model, but some of them turn into something else (they are said to "oscillate" into other particles) on their way to Earth and so go undetected. In June 1998, the Kamioka group reported the first experimental evidence of neutrino oscillations. While this represents a breakthrough in neutrino physics (if confirmed), it is presently unclear if the oscilations are of the right type to explain the solar neutrino deficit.

Where are the missing neutrinos? Is the proton—proton chain operating as we think? Do we really know what processes are at work deep in the hearts of stars? For now, the mystery of the solar neutrinos remains unsolved, although most physicists favor the neutrino-oscillation explanation. Virtually all researchers concur—or at least hope—that the correct interpretation of the solar neutrino problem will not tear apart the theoretical fabric of the proton—proton chain. Most believe that the description of solar fusion we have presented is this chapter is basically right; our understanding of neutrino physics just needs to be fine-tuned. But should drastic measures be needed to solve the solar puzzle, we may yet have to return to the drawing board to answer one of the most fundamental scientific questions of all: How does a star shine?