The suggestion that Mullis was solely responsible for the idea of using Taq polymerase in the PCR process has been contested by his co-workers at the time, who were embittered by his abrupt departure from Cetus. However, other Scientists have written that "the full potential [of PCR] was not realized" until Mullis' work in 1983, and that Mullis' colleagues failed to see the potential of the technique when he presented it to them. As a result, some controversy surrounds the balance of credit that should be given to Mullis versus the team at Cetus. In practice, credit has accrued to both the Inventor and the company (although not its individual workers) in the form of a Nobel Prize and a $10,000 Cetus bonus for Mullis and $300 million for Cetus when the company sold the patent to Roche Molecular Systems. After DuPont lost out to Roche on that sale, the company unsuccessfully disputed Mullis's patent on the alleged grounds that PCR had been previously described in 1971. Mullis and Erlich took Cetus' side in the case, and Khorana refused to testify for DuPont; the jury upheld Mullis's patent in 1991.. However later, in February 1999, the patent of Hoffman-La Roche (United States Patent No. 4,889,818) was found by the courts to be unenforceable, after Dr. Thomas Kunkel testified in the case "Hoffman-La Roche v. Promega Corporation" on behalf of the defendants (Promega Corporation) that "prior art" (i.e. articles on the subject of polymerase chain reaction, published by other groups, prior to the work of Mullis and the patent application) existed, in the form of two articles, published by Alice Chien et al. in 1976 , and A. S. Kaledin et al. in 1980 .