A quality manager at a caprolactam plant in Liaoning once told me that the two months she spent tracking down an unexplained increase in cyclohexanediol byproduct in the oxidation step were among the most frustrating of her career. Her plant was running the established cyclohexane liquid-phase air oxidation process — the Dupont/DSM route — which produces cyclohexanol and cyclohexanone as the primary products for eventual conversion to cyclohexanone oxime and then caprolactam. The cyclohexanediol level had crept from the normal 0.3% up to 0.9% over about three weeks, which didn’t sound alarming until it started affecting the Beckmann rearrangement selectivity downstream. Diol contamination at that level in the oxime feed shifts the acid-catalyzed rearrangement toward unwanted ring-opening byproducts, and suddenly the amide yield numbers they’d been hitting for two years were off by nearly four percentage points.