ERRORS IN THE ANALYSING OF PERFUMERY RAW MATERIALS of different polarities in the misfortunes experienced recently by yon Rudloff •. This author announced that gas chromatography showed the presence of camphene amongst the products resulting from the dehydration of terpineol, basing his results particularly on the retention time in a rapeseed oil chro- matographic column. It was in fact menthene-3, as has been shown by others 6 who identified the effluent not only by retention time, but by IR and N3/[R spectra. Temperature is another factor the variation of which is easily used to advantage in achieving separations. The logarithms of retention time are inversely proportional to the absolute temperature of the column. It therefore saves time to work at as high a temperature as possible. On the other hand it is known that the best separations are achieved at low tem- peratures. A compromise is therefore sought between the two require- ments. It should not be forgotten, however, that the constituents of perfumery raw materials belong to classes of products which behave very differently from the point of view of the relationship between retention time and temperature. It follows that at certain temperatures the separation of some constituents is no longer possible. It is therefore important to prepare comparative chromatograms at different temperatures, as illustrated by the following example. J. B. Roberts 7 has demonstrated the influence of temperature by the chromatography of a specimen of hop oil at various temperatures, all other conditions being the same. Below 125 ø the principal constituents appeared in the following order: /%myrcene, caryophyllene, farnesene, humulene. Above 125 ø the order became /%myrcene, farnesene, caryo- phyllene, humulene. At 125 ø caryophyllene and farnesene were not separated. It is evident that in the analysis of perfumery raw materials preference is given to the use of non-destructive detectors, permitting the recovery and identification of constituents, and above all their olfactive examination. The perfumer therefore is the first of the biological detectors, amongst which A. P. J. Martin 8 has suggested cockroaches, gipsy moths, dogs, etc. The most frequently used non-destructive detectors are thermal con- ductivity cells (catharometers) using hot filaments, and radiological detectors operating by ionisation. By using them it is possible to trap efficiently the elute with a view to identifying it by physical or chemical means. Modern instrumentation allows this identification to be rapid, economical and positive, with quantities of recovered substances of the order of milligrams or of micrograms. IR microspectrometry, NMR spectrography, and mass spectrography are used. The use of time-of-flight mass spectrometry, whilst the power of resolution of spectrometers built on this principle (around 200) is relatively low, permits the continuous monitoring of the effluent and the

32 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS consequent detection of unresolved components 9'•ø'u'•. But a time-of-flight spectrometer is very expensive, costing around •5,000. Whilst it is evident that the carrying out of gas-chromatography, with columns of different polarities at different temperatures, facilitates exact characterisation of a complex product and the identification of its con- stituents, it presupposes a knowledge of the retention times. But the relative uncertainty which surrounds the evaluation of retention times maintains the risk of confusion. We have only to consider the complex composition of most perfumery raw materials and the fact that we are acquainted with only some of the substances which could be present. We could be faced with substances previously unknown, and there is therefore the risk of confusing an unknown constituent, with a substance already encountered elsewhere. In this sense, the compositions attributed by many authors to essential oils and fractions of essential oils purely on the basis of retention time must be accepted with the greatest reserve, particularly when they result from a single chromatogram. It is essential for the elute to be identified by other methods such as IR spectrometry, NMR spectrography, and mass spectrography. The recording of peaks obtained by the use of thermal conductivity or ionisation cells is presented in the form of a series of sigmoidal curves of which some are more or less superimposed and therefore deformed. When the support for the liquid phase is totally inert, the peaks are symmetrical. Overloading of the column, i.e. the increase of the quantity of product injected beyond a certain value, lengthening of the column, too long an introduction time, an inadequate vaporisation temperature or the use of too high a proportion of stationary liquid in the charge, will enlarge the peaks. Symmetry of peaks is favourable to the separation of neighbouring substances. The conditions for distinguishing peaks in relation to the differences in retention volumes have been studied by various authors. They are defined as a function of a variable t which is the same as unity of error in the theory of errors, the elution curve itself having the form of a •urve of error. The curve representing the quantity of elute in relation to the volume W .of the mobile phase is, according to Martin & Synge, represented by the equation Q/Qmax:e '-•t' in which e is the elongation of the recorder. The ratio decreases rapidly when t increases, and it is practically nil when t = +3. Two substances eluted successively, and present in almost equal quantities, are therefore