SOLVENTS, HUMECTANTS AND BLENDING AGENTS vents formerly unknown in cosmetic formulations. The ideal solvent must possess certain definite properties other than merely being a good solvent for eosin. It should be compatible with the majority of lipstick ingredients and preferably act as a plasticiser for the mass. It must be non-toxic and non-irritant, it should neither evaporate nor unduly modify the consistency of the product. It appears that two main factors are involved in the extent and rapidity of staining of the skin by dyestuffs--the concentration of dis- solved dye being one factor and the skin-penetrating action of the solvent being the other. The concentration factor implies the use of a solvent of sufficient power to maintain the dye present in a lipstick in soluble form and since skin - penetration is primarily a diffusion process the size of the solvent molecules becomes a critical factor. For rapid permeation by the solvent there is thus a limit in molecular weight and a compro- mise has to be made between the rate of penetration and rate of evaporation of the solvent since the latter affects the shelf-life of the stick. The evidence now available to corroborate the allergic reactions due to eosin o.r its associated impurities 2 favours the use of efficient solvents since these permit a reduction in the amount o.f soluble dye required and ensure that it is evenly distributed in a uniform film on application rather than as dispersed particles which involve local areas of high dye concentration. The most common media em- ployed as dye solvents in the past have been castor oil and butyl stearate in combination. The solu- bility of eosin in each liquid is so small that they can hardly be classed as eosin solvents. With the normal amounts of soluble dye used it is evident that the greater part must be present in the dispersed rather than the dissolved state. Although castor oil is incompatible with hydrocarbons and requires a blend- ing agent for successful inclusion it imparts such advantages that it has become a classic component of modem lipsticks. It is almost essen- tial therefore that any solvent employed must either be an efficient blending agent for castor oil and hydrocarbons or must be capable of replacing it entirely. Butyl stearate, although an excellent emollient, could usefully possess a wider range of compatibility. Its low solvent power for eosin and the readiness with which it exudes even under temperate conditions tend to make it a rather doubtful asset. The investigations of Hall 3 led to the conclusion that the solution of the problem of eosin solvents rests "in the field of short chain aliphatic alcohols, ethers or combined groups" and other investigators 4 have suggested glycerol and glycol derivatives containing one or more free hydroxyl groups as suitable media. Evaluation of such products confirms the effectiveness of the alcohols, ethers and glycols but the 109

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS majority must be rejected on the various grounds of incompatibility, high volatility, toxicity, taste, odour, stability and irritant effects. In the form of the fatty esters the solvent power falls very considerably. Other products which have re- ceived specific attention in this field are 'tetrahydrofurfuryl alcohol •, which is too volatile for the prepara- tion of lipsticks of reasonable stability and storage life and is not entirely free from suspicion on the grounds of toxicity • and tetrahydro- furfuryl acetate 7 s which possesses penetrating odour and taste com- bined with definite irritant proper- ties. In contrast to these two pro- ducts, each of which is an excellent eosin so]vent (20ø/ø), tetrahydro- furfuryl stearate 8 dissolves 0.5% eosin at 20øC. The alkyl esters of the higher fatty acids are in general poor solvents for eosin although good dispersing agents for dyes, pigments and micro- crystalline waxes. Their use as eosin solvents seems to rely on their skin penetrating properties. The effect of molecular size is apparent in this series when one recalls that isopropyl myristate is somewhat more effective than butyl stearate in the rapidity with which a permanent stain is produced. Esters of dicar- boxylic acids e.g. diethyl sebacate are rather better solvents and are also suitable blending agents (vide Blending Agents) for vegetable and hydrocarbon oils. Lipsticks incor- porating 15% of isopropyl myristate or diethyl sebacate have proved quite stable after six months storage. Partial esters of glycerol, typified by the monostearate and monolaur- ate possess relatively poor solvent power for eosin but, in contrast, the glycols and their mono-ethers are more effective'. The lower glycols and their ethers, however, suffer from poor compatibility with the usual lip- stick components and the ethers are, in addition, too volatile for safe use. Some of the higher glycols have been examined as eosin solvents 9 and, of these, hexylene glycol appears to have many of the neces- sary characteristics. It is a good eosin mlvent with a suitable range of compatibility. Tests on the chronic toxicity of this product (carried out since publication of the original paper) have confirmed that the chronic oral toxicity is so low as to involve no danger in its applica- tion in cosmetics. Lipsticks con- taining hexylene glycol as eosin sol- vent have been used exclusively by one subject of moderately sensitive skin for a period of eighteen months without any symptoms of irritation, sensitisation or toxic effects. This product is now in use on a com- mercial scale in lipsticks. The mono- acetate of this glycol suffers from defects similar to those noted with tetrahydrofurfuryl acetate and the mono-stearate ester further demon- strates the poor solvent power of the higher fatty acid esters. The manifold functions required of a solvent for the eosin used in lipsticks, and the fact that so many of the investigations so far carried out have been confined to a few rather than all aspects involved in 11o