EVAPORATION AND THE ODOR QUALITY OF PERFUMES 177 EFFECT OF THE MEDIUM ON EVAPORATION RATES In the last experiment discussed, we called water a blending agent be- cause it lessened the disparity between the vapor pressures of the various alcohols, and because it was present in a concentration comparable to that of the alcohols. Had water (as is often the case) been present in large excess, we would have called it a solvent. Yet it would have had the same influence on the ratio of vapor pressures. A "DRY" B "WET" EFFECT OF SOLUTION o._• VAPOR COMPOSITION A = 0.5 cc of vapor from equal volmaes of ethanol, prop-nol, isobutanol, and butano 1 B = one cc of vapor from same mixture in an equal volume of water ArGon Carrier Gas 8 psi 10% Ca.t-bowaX 1500 Column 75øc 1000 Volts Dr. D. A.M. Mackay Evans Research & Development Corp. Figure 5.--Effect of solution on vapor composition. (Courtesy Dr. D. A.M. Mackay, Evans Research and Development Corp.) Two years ago we demonstrated in a series of experiments that the solvent (or more generally, the medium in which the perfume is incorporated) can influence the odor balance of a perfume blend (4). We found that a given mixture of any two odorants usually smelled differently depending on whether it was dissolved in water or in mineral oil, and that in order to

178 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS get a comparable odor effect in the two solvents different ratios of the odorants had to be used. In that investigation we used the human nose as the measuring device. We can now check the validity of our findings by use of the more objective gas chromatograph. As an example, let us look at linalool and phenyl ethyl alcohol. In our "smelling" experiments we took a mixture of equal parts of these two odorants and dissolved a portion of it in mineral oil, another portion in water. We found that the odor of linalool dominated much more strongly in the water solution than it did in mineral oil. Repeating this experi- ment, but "smelling" with the gas chromatograph rather than with the nose, we found the following: VAPOR OVEE S'OLUTION OF LINALOOL (o. zz) n,, PHEINTL ETHYL ,qL60HOL ß IN MINERAL OIL. i . {oza) VAPOI? OV[• SOLUTION OF VAF:O• OVF'R I'1 IXTI/R•E ß LINALOOL (oozyO •~D f'HENYL OF EQUAL WEIGHTS , . . ETHYL ALCOHOL (aozsx) LINALOOL • PHENYL ETHYL ALCOHOL .•-. Figure 6.--Effect of solvent on vapor composition of a two-component mixture: 5% Carbowax 1540 on Chromsorb W 50 ml. A/min. (8 psi) Column and Der. temp. 130ø Der. voltage 1250 V. Chart speed 15"/hr. a. 10 ml. vapor Peak ratio b. 10 ml. vapor c. 4.5 ml. vapor Peak ratio -- Linalool 10.2 - - 2.0 Phenyl Ethyl Alc. 5.2 Linalool 6.1 - 5.8 -- Phenyl Ethyl Alc. 10.5 The phenyl ethyl alcohol peak, while quite pronounced in the vapor over the mineral oil solution (Fig. 6a), is hardly discernible in the vapor of the water solution (Fig. 6b). This is in complete agreement with our earlier organoleptic findings. For the sake of comparison, we also prepared a gas chromatogram of the vapor over a mixture of equal parts of linalool and phenyl ethyl alcohol in the absence of any solvent (Fig. 6c). A com- parison of Fig. 6c with Fig. 6a indicates that mineral oil does not act like a "neutral" solvent here, but that it holds back linalool more strongly than phenyl ethyl alcohol.