After his stay in Nernst's lab, Lewis returned to Harvard as an instructor for three more years, and in 1904 left to become Superintendent of Weights and Measures for the Bureau of Science of the Philippine Islands in Manila. The next year he returned to Cambridge when the Massachusetts Institute of Technology (MIT) appointed him to a faculty position, in which he had a chance to join a group of outstanding physical chemists under the direction of Arthur Amos Noyes. He quickly rose in rank, becoming assistant professor in 1907, associate professor on 1908, and full professor in 1911. He left MIT to become professor of physical chemistry and dean of the College of Chemistry at the University of California, Berkeley in 1912. Lewis Hall at Berkeley, built in 1948, is named in his honor.
About 1902 Lewis started to use unpublished drawings of cubical atoms in his lecture notes, in which the corners of the cube represented possible electron positions. Lewis later cited these notes in his classic 1916 paper on chemical bonding, as being the first expression of his ideas.
In 1908 he published the first of several papers on relativity, in which he derived the mass-energy relationship in a different way from Albert Einstein's derivation. He also introduced the thermodynamic concept of activity in a paper, "The osmotic pressure of concentrated solutions, and the laws of the perfect solution," J. Am. Chem. Soc. 30, 668-683 (1908).
In 1913, he was elected to the National Academy of Sciences, but in 1934 he resigned. He refused to state the cause for his resignation; it has been speculated that it was due to a dispute over the internal politics of that institution or to the failure of those he had nominated to be elected. His decision to resign may have been sparked by resentment over the award of the 1934 Nobel Prize for chemistry to his student, Harold Urey, for the discovery of deuterium, a prize Lewis almost certainly felt he should have shared for his work on purification and characterization of heavy water.
In 1916, he published his classic paper on chemical bonding , in which he formulated the idea of what would become known as the covalent bond, consisting of a shared pair of electrons, and he defined the term odd molecule (the modern term is free radical) when an electron is not shared. He included what became known as Lewis dot structures as well as the cubical atom model. These ideas on chemical bonding were expanded upon by Irving Langmuir and became the inspiration for the studies on the nature of the chemical bond by Linus Pauling.
In 1921, Lewis was the first to propose an empirical equation describing the failure of strong electrolytes to obey the law of mass action, a problem that had perplexed physical chemists for twenty years. His empirical equations for what he called ionic strength were later confirmed to be in accord with the Debye-Hückel equation for strong electrolytes, published in 1923.
In 1923, he formulated the electron-pair theory of acid-base reactions. In the so-called Lewis theory of acids and bases, a "Lewis acid" is an electron-pair acceptor and a "Lewis base" is an electron-pair donor. This year he also published a monograph on his theories of the chemical bond
Based on work by J. Willard Gibbs, it was known that chemical reactions proceeded to an equilibrium determined by the free energy of the substances taking part. Lewis spent 25 years determining free energies of various substances. In 1923 he and Merle Randall published the results of this study, which helped formalize modern chemical thermodynamics.
In 1926, he coined the term "photon" for the smallest unit of radiant energy, although his photon differed from the photon of modern quantum physics.
Lewis was the first to produce a pure sample of deuterium oxide (heavy water) in 1933. By accelerating deuterons (deuterium nuclei) in Ernest O. Lawrence's cyclotron, he was able to study many of the properties of atomic nuclei. During the 1930s, he was mentor to Glenn T. Seaborg, who was retained for post-doctoral work as Lewis' personal research assistant. Seaborg went on to win the 1951 Nobel Prize in Chemistry and have the element Seaborgium named in his honor while he was still alive.
He died at age 70 of a heart attack while working in his laboratory in Berkeley. He had been working on an experiment with liquid hydrogen cyanide, and deadly fumes from a broken line were leaking into the laboratory when a graduate student found the professor's lifeless body under a workbench. The coroner said Gilbert N. Lewis died of coronary artery disease; however, some believe that the death may have been a suicide. UC Berkeley Professor Emeritus William Jolly, who reported the various views on Gilbert N. Lewis' death in his 1987 history of the University of California, Berkeley’s College of Chemistry, From Retorts to Lasers, said one higher-up in the department believed the suicide theory.
A possible explanation for the suicide theories was depression following a lunch with Irving Langmuir. Langmuir and Lewis had had a long rivalry, dating back to Langmuir's extensions of Lewis's theories on the chemical bond, and Langmuir had been awarded the 1932 Nobel Prize in chemistry for his theories of surface chemistry, while Lewis had not received the Nobel Prize despite 35 nominations. On the day of Lewis' death, Langmuir and Lewis met for lunch at the University of California, Berkeley--a meeting that was recalled by Lewis's last research associate, Michael Kasha, only years later. It was reported by associates that Gilbert N. Lewis came back from the meeting in a dark mood. He reportedly sat down for a morose game of bridge with some colleagues, and then went back to work in his lab. An hour later, Gilbert N. Lewis was dead. Irving Langmuir's papers at the Library of Congress confirm that Irving Langmuir was on the University of California, Berkeley campus that day. Irving Langmuir had gone to the University of California, Berkeley to receive an honorary degree.