436 1.4 12 1.0 � 0.8 � w 5= 0.6 0.4 0.2 JOURNAL OF COSMETIC SCIENCE ----4o•c -----5o·c __,._60-c 0.0 +----r---r-----r-----r-�-----,-----,,---..----r-----1 50 100 150 200 250 s/f ratio Figure 3. Supercritical extraction yield of red pigment measured at various extraction temperatures of carbon dioxide. The extraction pressure was maintained at a constant 400 bar. with supercritical extraction. The red pigment produced via ethanol extraction was compared to that of supercritical extraction in several ways. SKIN IRRITATION Skin irritation patch tests were conducted to determine the toxicity of the red pigment extracted with supercritical carbon dioxide. When the test was performed with cosmol solution containing 17% red pigment, coloration was clearly observed. No edema, however, was detected. When the concentration of red pigment was reduced to 10% in the lipstick, less coloration was observed than with the 17% sample. In this test, also, no edema was observed. The lipstick containing 10% red pigment via supercritical extraction was acceptable from the standpoint of skin irritation. According to the results of skin irritation tests, the red pigment generated via supercritical extraction was not recommended for use in lipstick, which requires more than 10% of supercritical red pigment from Lithospermum erythrorhizon. COLOR ASSESSMENT Four different lip glosses were prepared for color assessment, as is shown in Table I. Lip glosses with two different sources of red pigment, supercritical carbon dioxide and ethanol extracts, were evaluated and compared to commercial lip glosses containing D&C Red No. 6 (C 18 H 1 4N2 06S·2Na) and D&C Red No. 7 (C 18 H 1 4N2 06S·Ca). Table II shows the CIE tristimulus color values and lightness for the evaluations of chroma­ ticity and color differences (17). The lip gloss prepared with supercritical red pigment evidenced a similar or acceptable lightness and red color value as compared to lip glosses with the synthetic colors, D&C Red Nos. 5, 6, and 7. The total color difference, dE*a6, was also found to be acceptable as compared to the synthetic colors. The lip gloss prepared via ethanol extraction, however, evidenced too low a red color value to be acceptable, despite the fact that it exhibited the highest lightness. The total color difference value was too high to be acceptable when it was compared to synthetic colors,

EXTRACTION OF RED PIGMENT FROM L. ERYTHRORHIZON 437 Table II Evaluation of Chromaticity and Color Difference Sample L* a* b* SFE 60.94 50.78 12.06 Ethanol 85.54 12.72 -0.57 47.04 D&C Red No. 6 (C18H14N206S·2Na) 57.28 67.09 21.73 19.31 D&C Red No. 7 (C18H14N206S·Ca) 50.55 63.76 -3.7 22.91 dE*ab 47.04 65.21 61.96 19.31 65.21 26.52 as well as to the supercritical red pigment. According to the results of our color assessment, the red color produced via ethanol extraction was quite different from that of the supercritical red pigment and the commercial red colors. We surmised that this was because the ethanol-extracted pigment harbors more compounds unrelated to red color properties, as ethanol has amphithetic properties, and thus tends to extract a broader range of compounds than does supercritical carbon dioxide. COLOR STABILITY Color stability is one of the most important factors in the development of new natural pigments. Many natural pigments have been discontinued due to stability issues, despite their obvious advantages. The light-illuminated color stability tests were conducted with lip glosses containing supercritical red pigment and ethanol extract. Table III shows the evaluation of chromaticity and color difference prior to and after five days of light illumination. No significant changes in lightness and color values were observed in the lip gloss prepared with supercritical red pigment. Also, the total color difference was sufficiently small to assure stability. The lip gloss with ethanol extract, however, evidenced significant changes in color values, as well as a total color difference. The reduction in the red color value was also evaluated via a visual test, in which almost no red color properties were detected. The profound stability of the supercritical red pigment could be explained by the results of the DPPH scavenging activity test. Samples harboring various concentrations (1 %, 2%, 3%, 5%, 7%, 10%) of the red pigments obtained by supercritical carbon dioxide extraction, and those obtained by ethanol extraction, were tested. As is shown in Figure 4, the supercritical red pigment resulted in higher antioxidant activities than were seen in the ethanol extract samples at every different concentration. The DPPH test did not, however, completely explain the color stability of the supercritical red pigment. The observed profound antioxidant activity, however, could be correlated with the color stability (18,19). When we assessed the concentration of specific compounds in the red Table III Evaluation of Light-Illuminated Color Stability Sample L* a* b* dE*ab SFE day 1 60.94 50.78 12.06 SFE day 5 61.67 48.48 14.69 3.57 Ethanol day 1 85.03 12.72 -0.51 Ethanol day 5 89.61 0.46 7.02 15.62