Since his death in 1988, Richard Feynman has become something of an industry. In addition to several biographies, there are published collections of Feynman's essays and lectures, including just about every scrap of paper he ever scribbled on. Much of this presents an image of the man as bongo drummer, player of practical jokes and major-league womanizer, who somehow turned out groundbreaking science in his spare time. But let's face it: when it comes to Feynman's personal life, we really do not know what to believe. Murray Gell-Mann, Feynman's long-time colleague at the California Institute of Technology, once famously grumbled that Feynman "spent a great deal of time and energy generating anecdotes about himself". Most of his adventures do little, if anything, to illuminate Feynman the physicist.

In Quantum Man: Richard Feynman's Life in Science, the scientist, science writer and eminent spokesman for science Lawrence Krauss focuses on Feynman's research, thus providing a much-needed corrective to this caricature. What emerges is a portrait of a man who worked long hours to understand physics. Physics was Feynman's heart and soul, and Krauss has done a superb job of showing this in his book, taking us through Feynman's oeuvre as if teaching a masterclass.

What has always interested me about Feynman was his creative power as a problem solver, which Krauss astutely describes. His keen intuition was honed by considering problems from many different angles, and from making mistakes and learning from them, instead of being discouraged – all hallmarks of high creativity. Again and again, what jumps off the page is Feynman's keen eye for spotting what is a fundamental problem, and his single-minded concentration on solving it. This required great powers of compartmentalization, which involved blocking out of his mind anything and anybody other than the problem he was working on. As his second wife put it during their divorce proceedings, "He begins working calculus problems in his head as soon as he awakens. He did calculus while driving his car, while sitting in the living room and while lying in bed at night."

Krauss argues persuasively for the importance Feynman placed on experimental data at every stage in his theoretical work. However, I must disagree with his claim that Feynman was unmoved by considerations of beauty, or that data were all that mattered. In 1957 Feynman and Gell-Mann worked out a theory of the weak interaction that conflicted with key experimental data. Feynman insisted, along with Gell-Mann, that the data were wrong: "There was a moment when I knew how nature worked. [Our theory] had elegance and beauty." The experiment was redone and the data indeed turned out to have been wrong. This was a bold move with few precedents, although Einstein, with a similar aesthetic bent, had asserted in 1907 that data conflicting with the special theory of theory of relativity were incorrect. He was right too.

This approach was Feynman's route to his first and greatest breakthrough in 1948, when he developed a theory of how light interacts with electrons, known as quantum electrodynamics (QED). This theory agreed with relativity and led to no infinite quantities: it contained only the finite measured experimental values for the electron's charge and mass. Here, Feynman's use of his signature "path-integral formalism" was crucial, and it ran through many of his other important insights like Ariadne's thread.

As a graduate student at Princeton in the early 1940s, Feynman had gleaned the importance of this formalism from a 1933 paper by Paul Dirac, but Dirac drew no pictures. Feynman did, and his famous "Feynman diagrams" subsequently became part of the language of physics. They have achieved iconic status, decorating T-shirts and being emblazoned on Feynman's own van. How Feynman discovered the diagrams that bear his name, though, has not always been clear. Based perhaps on physics "mythology", Krauss writes that the "first Feynman diagram in print was actually [Freeman] Dyson's", referring to a 1948 paper by Dyson. But in fact, something very like a Feynman diagram had appeared five years earlier in a widely read book by Gregor Wentzel, then at the University of Zurich, entitled Einführung in die Quantentheorie der Wellenfelder [Introduction to the Quantum Theory of Wave Fields]. I give the German title because this is the version Feynman referred to in his 1949 paper "The theory of positrons", citing it for a technical point concerning second quantization.

I came across Wentzel's diagram in 1981, while researching the concept of visual imagery in quantum physics. In his book, Wentzel presented it as a didactic device to depict how neutrons and protons might interact by exchanging a charged meson. Also, in 1948, his book had not yet been translated. Yet even if Feynman could not read German (and I wager he could not), he may have seen Wentzel's diagram, become curious and stumbled through Wentzel's description of it. In this way, he may have had an epiphany: the way to give visual imagery to the mathematics of his version of QED. Certainly, his manuscripts indicate that this was what he was searching for. Alas, although I corresponded with Feynman, somehow I never asked him about Wentzel's diagram and so this tantalizing episode remains unresolved.

Krauss writes insightfully about Feynman's dissatisfaction with his renormalized version of QED. Rather than solving basic issues, Feynman believed it merely swept them under the rug by subtracting infinite quantities to produce finite ones. QED turned out to be valid up to a certain scale (distances greater than the electron's Compton wavelength, 2.43 × 10–12 m); owing to this restriction physicists refer to it as an "effective theory". Although string theory avoids the infinities involved in renormalization, Feynman disliked it because of the vague way it interpreted the necessary extra six or seven dimensions (in addition to the four dimensions we experience) and, as one would expect of him, its lack of contact with experimental data.

Feynman's influence went far beyond physics per se, Krauss reminds us, extending into biology, physics education, nanotechnology and computer science; and into ingeniously ferreting out the cause of the Challenger space-shuttle disaster in 1986, two years before he died. He was always on the lookout for fresh fields with new problems to solve. "For a fearless and brilliant adventurer like Richard Feynman," writes Krauss, "this was the reason for living." In his book Krauss brings Feynman's adventures in physics brilliantly to life, with never a bongo drum in sight.