A powerful source of light from an electron accelerator, not long ago considered a nuisance by scientists, is emerging as one of the most important research tools since the invention of the microscope.The tool is yielding fundamental new knowledge in such a wide variety of fields that scientists around the world are clamoring for more of the machines that produce the light. Yet the sheer cost and complexity of these electron-whirling behemoths, called synchrotrons, limits how many will be built, at least in the near future.The light is produced by needle-size bunches of electrons shot, riflelike, through a circular tube. The high-energy particles, steered by electromagnets wrapped around the tunnel, hurtle along at nearly the speed of light.They give off intense radiation when traveling in a curved path in a fashion similar to the way car tires squeal when rounding a tight corner. Scientists tap into this radiation through windowlike ''ports'' in the storage ring, the tube the electrons race around.The brilliant beams were not always so coveted. Until recently researchers in high-ener-gy physics who used synchro-trons to probe the mysterious world of sub-atomic particles regarded the radiation as waste.Now, however, it is being harnessed to provide unprecedented looks into the basic structure of matter. The synchrotron doesn't represent a conceptual leap forward in science. It is a dramatic advance in the use of one tool, radiation, which is employed to probe basic structures. ''You could take a shadowy photograph in the 19th century, but photography didn't become a tool until light-sensitive plates were developed,'' says Martin Blume, who heads the National Synchro-tron Light Source at the Brookhaven National Laboratory on Long Island. He says synchrotrons will have a similar impact on radiation research.Synchrotrons perform several tricks that conventional X-rays and ultraviolet light cannot. For one thing, the radiation is far brighter - in some cases on the order of 100 ,000 times as bright as X-rays. The intensity allows scientists to obtain exposures and carry out experiments much more quickly than before. They also can produce radiation in a range of wavelengths, from longwave infrared to shortwave X-ray. Advantage here: Researchers can ''tune'' into the beam they need for the task at hand.Among those tapping the power of this light source are chemists and scientists studying materials. They can use the radiation to explore the arrangement of individual atoms at the surface of materials. Oil companies, for instance, are studying the chemical reactions. The aim: to find better catalysts for refining crude oil.Geologists use the X-rays to probe the conditions at the center of the earth. Minerals are squeezed between diamond bits to simulate the intense pressures inside the planet. Synchroton light then reveals how the atoms react. Other scientists are analyzing molecular changes in substances ranging from food to plastic.Companies are also exploring for better materials to make ''superchips'' for tomorrow's computers. Right now the work by firms such as IBM and AT&T's Bell Laboratories is pure research. But eventually they may use the machines to manufacture the next generation of integrated circuits. IBM, for one , expects to know within two years whether the technique could be used. The Japanese and a consortium of European companies are looking at using them for chipmaking, too.For biochemists, synchrotons are opening a whole new window on cell and other biological behavior that may lead to new agricultural products and a better understanding of pollutants. Many medical uses are envisioned as well.''We don't even understand what we're seeing,'' says Ralph Feder, an IBM scientist working on Brookhaven's machine. ''There is no question that every biology lab will eventually have one of these.'' Maybe not all labs, but certainly plenty of them, biological or not, would like to. Most of the five US facilities that can be used for synchrotron light - at Brookhaven, Cornell, Stanford, the University of Wisconsin, and the National Bureau of Standards - are booked up months in advance.When scientists do get ''beam time,'' they often work around the clock for days at a stint, catnapping on couches and surviving on junk food in between. On a recent routine day at the Cornell High Energy Synchrotron Source, for instance, the lab looked as much like a cheap hostelry as a science shop: Cups were strewn about, papers piled on desks, scientists dozing on chairs.A number of firms - IBM, General Electric, and Bell Labs - have talked about building their own small synchrotrons. But none have moved off the drafting table. Plenty of planning, meanwhile, is going on overseas. Synchrotron labs are under study or construction in Brazil, China, France, West Germany, India, and Japan.There are snags with the big-light machines, though. One is cost. A future facility may cost more than $100 million. A small, single-purpose version could go for under $2 million. A special committee set up by the Department of Energy is looking into what the US science community will need in the way of light sources over the next decade. Proposals for a controversial new synchrotron at Lawrence Berkeley Laboratory in California, as well as plans for expanded facilities at existing centers, are being weighed. The talk will be hard-nosed; no consensus exists on the idea of supporting ''big science'' projects vs. smaller ones, say, a dozen electron microscopes.The machines can also be temperamental. New accelerator designs push the cutting edge of technology. They take a lot of time, money, and expertise to fine-tune. There were long delays in the switching on of the new light sources at Brookhaven and Wisconsin. Europe's most powerful X-ray source, in England, has been hobbled since October. Yet few doubt that the machines will have long-range impact. They are, sums up Peter Eisenberger, director of physical sciences at the Exxon Research & Engineering Company, laying the groundwork for a ''scientific revolution.''