After acclimation, the plants were grown independently in different light treatment chambers at 20 ± 0.2 &#8451 and 65 ± 2% humidity until the harvest date (35 days after light treatment). The white fluorescent light (70 ± 5 µmol/m²/s) was maintained for 12 h, and then each of the blue, green, red, and white lights was irradiated at 70 ± 5 µmol/m²/s for 4 h using LED arrays (DR LED Networks Co., Seoul, Republic of Korea). The spectral energy distribution of four different LED arrays was measured from 300 to 800 nm with a spectroradiometer (International Light, RPS-900, U.S.). Their maximum spectral wavelengths were 463 (blue), 518 (green), and 632 nm (red); the white LEDs had a broad spectrum. Irradiance was measured using a quantum sensor (LI-COR, LI-191, Lincoln, NE, U.S.). Water was supplied daily with top irrigation and a nutrient solution (Hoagland, pH = 5.9 ± 0.2, electrical conductivity = 1.2 dS/m) every 4 days until harvest.

A quantitation and principal component analysis biplot demonstrated that luteolin-7-O-glucoside (2), luteolin-7-O-glucuronide (3), and quercetagetin-trimethyl ether (8) were the highest polyphenols yielded under green light, and dicaffeoylquinic acid isomer (4), dicaffeoylquinic acid isomer (5), naringenin (7), and apigenin-7-O-glucuronide (6) were greatest under red light. Chlorogenic acid (1) and 1,2,6-trihydroxy-7,8-dimethoxy-3-methylanthraquinone (9) were produced in similar concentrations under both light types. The white and blue light appeared inefficient for polyphenol production.