42 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Fig. 1. Diagram of Instrumentation. A. Helium carrier gas, B. Flow regulator, C. Denitrosation apparatus and sample container, D. Cold trap, E. Vapor scrubber, F. Teflon filter, G. Bendix NOx Analyzer, H. Oxygen, I. Pump, J. Strip chart recorder, K. Electronic integrator. bond to generate nitric oxide. The nitric oxide is swept from solution by an inert carrier gas (helium) through a series of traps to an oxides of nitrogen analyzer. The nitric oxide is determined via its chemiluminescent reaction with ozone. Nitrite is chemically reduced to nitric oxide. REAGENTS AND CHEMICALS N-nitrosodiethanolamine was obtained from the National Cancer Institute chemical repository at the Illinois Institute of Technology Research Institute (Chemical #208). Stock solutions of 10 -2 M were prepared in water and stored in the dark at -5øC. Standard solutions of sodium nitrite (Matheson, Coleman and Bell, #SX 665) were prepared in a similar manner. Reagents used for the denitrosation step were acetic acid (J. T. Baker Chemical Co., ff9507), sulfuric acid (J. T. Baker Chemical Co., ff)681) and sodium iodide, Suprapur © (Matheson, Coleman and Bell, #SX 623). Other chemicals used were hydrochloric acid (J. T. Baker Chemical Co., ff9535) and sulfanilamide (Matheson, Coleman and Bell, #SX 1150). Distilled water was further treated by passing it through two IWT (Illinois Water Treatment Co.) Research Model 1 ion exchange columns. Dowex © l-X8 anion exchange resin (chloride form, 20-50 mesh, #1904) was obtained from the J. T. Baker Chemical Co. Cosmetic samples were purchased locally and supplied by the Food and Drug Administration, Department of Health and Human Services.

NDELA DETERMINATION 43 PROCEDURES A known weight of sample was diluted or suspended in a 50 ml centrifuge tube with 40 mL distilled, double deionized water. The sample and water were then mixed on a Vortex mixer at high speed for approximately 3 minutes. The suspension was centrift•ged at 2500 rpm for 10 minutes using an International Equipment Company (Model SBV) centrifuge. The aqueous portion was carefully removed from the centrifuge tube by decantation or transfer pipets and retained for analysis. NITRITE The nitrite determination was based on methodology introduced by Cox (8). The method involved the reduction of nitrite by iodide ion in a weakly acidic medium. One milliliter of 10% (w/v) sodium iodide and 3 mL glacial acetic acid were added to the 20 mL sample or an aliquot of sample diluted to 20 mL in a reaction vessel. The sample was outgassed until no further NO evolution was observed. Detector response was monitored using a strip chart recorder peak areas were obtained by electronic integration. Nitrite concentrations were determined by comparison to nitrite stan- dards. N-NITROSAMINE Further clean-up of the aqueous suspension was necessary if nitrite was present in excess of 1 ppm NO2-N. Ion exchange and chemical removal were used for these purposes. Ion exchange resin was prepared by slurrying in 2 N HC1. A 5 mL bed was washed with two 25 mL portions of 2 N HCI and two 25 mL portions of distilled, double dionized water prior to sample application. A 10 mL aliquot of sample was applied to the column and eluted at a rate of ! mL per minute. The first 3 mL of eluate were discarded and the next 20 mL collected and retained for analysis. Following ion exchange or if ion exchange was unnecessary, sulfanilamide was used to remove nitrite. A solution of 0.10 M sulfanilamide in 10% (v/v) HCI was prepared and 0.2 mL added to a 5 mL sample. The sample was mixed and allowed to stand two minutes before analysis. Since many of the aqueous cosmetic suspensions are alkaline, the pH of the sample after addition of sulfanilamide reagent should be checked. If the pH was above 3.5, 10% (v/v) HC1 was added. The pH was monitored with pH paper. Denitrosation was achieved by mixing the sample (5 mL) or aliquot of sample diluted to 5 mL, 13 mL of glacial acetic acid, and 3.5 mL of concentrated sulfuric acid in a reaction vessel. One milliliter of a 10% (w/v) sodium iodide solution was then added and the reaction vessel placed immediately on the outgassing apparatus. The sample was outgassed until no futher NO evolution was observed. Peak areas were determined by electronic integration. Nitrosamine concentration was determined by comparison to standard solutions of NDELA. For many samples, excessive foaming occurred. By adding a few drops of 2-octanol to the reaction solution, the foaming was controlled.