scope:

Introduction

Character or quality of odor is a sensory property of many substances that is important in food and beverage sciences, technology of many industrial products, air and water pollution, indoor air quality control, cosmetics and hygiene, botany, psychology of interpersonal relations, medicine, and other unnamed fields.

In odor evaluations and panelist selection for such evaluations, there is a need for odor reference samples with known odor character. In evaluations, unknown odors may be compared to the reference odors; in panelist selection, those who appear to perceive odors in terms similar to the majority of people can be distinguished from those who have different perceptions. Profile data, especially with help from Table 161, can illustrate various descriptors and instruct panelists on specific odor characteristics. Beyond the distinction in the odor perception, causes for differences in odor perception can be explored.

The ASTM Odor Profiling project was designed to develop odor character information on odors of various types, ranging from the very pleasant to the very unpleasant. The present Atlas is a compilation of the collected information and should serve the various tasks of characterization of unknown odors (by similarities or dissimilarities to the odors characterized by the Atlas), panelist selection, panelist training, and as a data base for research in the science of smell.

The method selected to characterize odors for the purpose of the Atlas is based on semantic profiling of odors. A list of odor descriptors is provided to the panelists, who smell the odorous sample and describe its odor by rating the applicability of each of the descriptors on a scale of 0 to 5. A score of 0 means that the descriptor is not applicable to the odor evaluated; a score of 5 means that the descriptor is highly appropriate.

Figure 1 is a reproduction of the odor evaluation form used in the work. Figure 2 depicts the form of sample in which the odorous materials were supplied to the panelists for evaluation. Odorants were dissolved in a low-odor dipropylene glycol to obtain a concentration exhibiting a noticeable but not too strong odor. Balsa wood chips were impregnated with this solution and sealed within an aluminum "boat" and cover packets, and held together with a low-odor silicone self-adhesive tape. Each panelist received a separate sample and evaluation form, with both the sample and form coded. For an evaluation, the packet was opened, the sample smelled, and the applicability of each descriptor scored (except 0 applicability which was left unmarked).

Data from such profiling were pooled for all (120-140) panelists. Two statistical measures can be derived from such data:

(1) percentage of usage or the percentage of people who used the particular descriptor for the particular sample and:

(2) score level for this descriptor. With a maximum score level of 5, the maximum total score sum is equal to the number of panelists multiplied by 5. When the sum of the scores given by all panelists to the particular descriptor is divided by this maximum score sum, the ratio can be expressed as a percentage of the maximum possible score.

ATLAS OF ODOR CHARACTER PROFILES

Experience has shown that the most equitable indicator of the descriptor applicability is the geometric mean of the percent usage and percentage of the maximum possible score. This mean is termed percentage of applicability. The profile tables list values for this index and the percent usage. The percent applicability is graphically illustrated in the tables, with each star equal to four percent of applicability. This increment is approximately equal to the standard deviation of the percent applicability in repeated profiling of the same odor by a large number of people.

In the odor evaluation forms shown in Figure 1, the ordering of descriptors follows the principle used by Harper. Descriptors are arranged so that similar descriptors are not next to each other, and descriptors with an unpleasant connotation are intermixed with those of pleasant connotation. This arrangement should result in a more independent consideration of each descriptor during profiling and should reduce the tendency of subordinating some descriptors to others. When converting to odor profiles, the descriptor order is changed so that the related descriptors are clustered next to each other. This arrangement permits inspection in terms of clusters-different panelists may use similar but not identical descriptors for the same odor. This becomes evident in the profiles shown in Tables 1 through 160.

The Historical and Technical Notes section provides some information on odor profiling development within ASTM and lists a few references where such information can be found.