RK and NBD-H hydrazine were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Fragrance mist was obtained via the internet. Acetonitrile was purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). Trifluoroacetic acid was obtained from Wako Pure Chemical Industries (Osaka, Japan).

The HPLC system consisted of a model L-2130 pump (Hitachi, Tokyo, Japan), a Rheodyne injection valve (Cotati, CA, U.S.A.) with a 50-μL loop, and a model F-1050 fluorescence spectrophotometer(Hitachi) with excitation at 470 nm and emission at 550 nm. A 150 mm × 4.6 mm i.d. HPLC column (C₁₈-MS-II, Nacalai Tesque, Kyoto, Japan) containing 5 μm particles of C₁₈ packing material was used. Peak quantification was performed using a model D-2500 Chromato-Integrator (Hitachi). The mobile phase was prepared by the addition of acetonitrile (460 mL) to 540 mL of Milli-Q water containing trifluoroacetic acid (0.1% v/v). Samples were eluted from the column at room temperature at a flow rate of 1.0 mL/min.

Ultrapure water was obtained from a Milli-Q water purification system (Simplicity^® UV, Millipore Corporation, Bedford, MA, USA). Standard solution was prepared by dissolving RK (50 mg) in acetonitrile (200 mL) and stored at 4˚C. Working standard solutions (0.2, 0.5, 1, 2, 5 and 10 μg/mL) were prepared by dilution with acetonitrile.

A previous method using NBD-H required 1 h reaction time at room temperature in trifluoroacetic acid solution (0.0025 v/v%) in acetonitrile for derivatization [15] . This method was used herewith minor modifications. Briefly, trifluoroacetic acid solution in acetonitrile (0.0075 v/v%, 100 μL) was added to diluted standard sample (100 μL), and then NBD-H hydrazine solution in acetonitrile (2 mg/mL, 100 μL) was added in a stoppered test tube(the final concentration of trifluoroacetic acid is the same as in the reported method [15] ), and the mixture was vortexed. The top of the test tube was covered using a paper (PROWIPE, DAIO Paper Corp., Tokyo) wetted with ice-cold water, and then reaction at 80˚C was performed in a water bath. The test tube was placed on ice for 2 min. An aliquot (50 μL) was injected into the HPLC system.

Fragrance mist (200 μL) was dissolved in acetontrile (200 mL). The solution was diluted 1000-fold, and then analyzed as described above.

Fragrance mist (200 μL) was spiked with 0.25 mg of RK and diluted as described above (Application to fragrance mist sample). The resulting sample was analyzed to determine recovery of the added standard in order to assess the accuracy of the method.

For the time course study, the reaction time was set at 10, 20, 30 or 40 min at 80˚C. RK standard solution (1 μg/mL, 100 μL), trifluoroacetic acid solution (0.0075 v/v%, 100 μL) and NBD-Hhydrazine (2 mg/mL, 100 μL) were mixed appropriately and analysis was carried outas described in Materials and Methods. The peak area of NBD-RK reached maximum at 20 min (Figure 2). In a preliminary study, the peak area of NBD-RK at a reaction time of 20 min at 80˚C was higher than those at 60 and 70˚C. Thus, we selected 20 min at 80˚C as standard conditions.

Figure 3 shows typical chromatograms obtained from (a) standard sample (5 μg/mL) and (b) test sample of a fragrance mist (1 μL/mL). The retention time of NBD-RK derivative was 10.3 min. A peak of NBD-RK derivative was observed in the test sample. The running time was set at 18 min.

The standard curve of RK was constructed by plotting integrated peak area vs. concentration. The calibration plot was linear (slope, 317.4; intercept, −130.4) in the range of 0.2 to 10 μg/mL with a correlation coefficient (r²) value of 0.9980. The lower limit of detection for RK was estimated as the concentration giving a detectable peak (signal-to-noise ratio of 3 or more), since the peak was located close to a medium blank peak at 9.1 min. The value of the lower limit of detection was 0.018 μg/mL (absolute amount of 1.8 pmol, 0.3 ng).

RK is a volatile compound, and has been analyzed by GC-mass spectrometry [10] and GC with a flame ionization detector [11] , but those reports did not clearly describe the sensitivity of their methods. Also, the lower limit of detection of RK with the previous HPLC methods was not established [8] [9] [13] . However, HPLC analysis of carbonyl compounds (benzaldehyde, propionaldehyde and heptan-4-one) using NBD-H showed detection limits of 1.9 pmol, 35 fmol and 0.67 pmol, respectively [14] [15] . Therefore the sensitivity of the present method (1.8 pmol RK) appears to be sufficient for the purpose.

Precision and accuracy in intra-day and inter-day assays of RK are shown in Table 1. In the intra-day assay, the range of standard deviation was within 2.9% to 5.7% of the mean and recovery was within the range of 96.0% to 102.0%. In the inter-day assay, the range of standard deviation was within 4.6% to 8.1% of the mean and recovery was within the range of 93.0% to 104.0%.

The developed method was used to determine RK in fragrance mist and in fragrance mist spiked with authentic standard. As shown in Table 2, the concentration of RK in fragrance mist was 1.18 ± 0.07 mg/mL (average ± S.D., range, 1.12 to 1.28 mg/mL, n = 5). Recovery of RK from spiked fragrance mist was 83.9% ± 3.9% (average ± S.D., range, 79.6% to 88.8%, n = 5).