1. Introduction
There has been a significant increase in public concern for odors from industrial, agricultural, and wastewater treatment facilities (Yun and Cho, 2018). To cope with these odor problems, many governments have implemented standards and regulations for odor emission control. Many of the regulations require the measurement of odors, either to prove compliance or to measure and monitor odors (McGinley and McGinley, 2001).
Odor can be scientifically measured like most environmental quality indexes. Though the odor testing techniques have progressed over time with the development of terminology, methods, and instrumentation (Kim et al., 2017), olfactometry is the mostly used method in odor study. It involves a panel to smell the odor sample and report the human responses using objective and subjective descriptors. The objective descriptors reflect odor concentration, odor intensity, odor persistence, and odor character. On the other hand, subjective descriptors profile odor effect, such as hedonic tone, annoyance, objectionable, and strength. However, in most cases, a suitable odor reference is needed to help the assessors judge the evaluation items.
n-Butanol was recommended as a possible odor reference compound for odor intensity measurement by the Committee E18 of the American Society of Testing and Materials (ASTM, 2010). The odor of n-butanol is particularly useful as odor reference because the chemical is relatively nontoxic and easily obtained, possesses a noncharacteristic odor, and its concentration can be varied to produce a large range of odor intensities (Moskowitz et al., 1974;Zhang et al., 2002). However, Saad et al. (2005) reported that it would produce evaluation error by a factor of 100 with a n-butanol reference. Segura and Feddes's work showed that n-butanol does not appear to be a representative reference odor for livestock odors (Segura and Feddes, 2010). In European standard for odor concentration measurement, a certified reference material of n-butanol (40 ppm and 200 ppm) in nitrogen was used as the reference and main odorant to be used during olfactometry (British Standard, 2003). The reference value for the European odor unit was set at 1 ouE / m3 = 40 ppb/v n-butanol. This value is obtained by a strict performance protocol for panel selection, which has been the major promoting factor in enhancing the reproducibility of olfactometry (van Harreveld et al., 1999). Some US university odor laboratories also committed to implement the butanol screening assessor selection criteria described in European odor testing standard (McGinley and McGinley, 2001).
Odor intensity and hedonic tone of butanol was studied some decades ago (Moskowitz et al., 1974). The threshold for butanol is approximately 62 ppm, based upon an average between 45 ppm and 79 ppm, and the level of 300-ppm butanol appears to be hedonically neutral to strength, being judged neither pleasant nor unpleasant according to the functions obtained from the average estimates. In the study, the odor intensity and the odor pleasantness (or un-pleasantness) with n-butanol concentration were fitted by power functions when the estimates were obtained from ratio scaling procedures, while the results of category scaling were fitted by logarithmic function. There are substantial changes in the odor quality profile of n-butanol at different concentrations. The profiles at below 270 ppm was predominately described as “sweet” (Laing et al., 1978), while diverse descriptors including “alcoholic”, “medicinal”, “chemical”, and “paint” were most used at 160 mL/L (160,000 ppm) (Dravnieks, 1985). These reported n-butanol odor character profiles do not present data points between 270 ppm and 160, 000 ppm. The quantitative assessment of the pleasantness of odorants is an even more complex task since little is known about the physiological basis of pleasant and unpleasant sensations. Unlike intensity, which increases as the concentration of odor is raised, the relationship between pleasantness and concentration appears to depend upon the odorant. At present, no single mathematical expression has been reported which satisfactorily describes the relation for all odorants (Laing et al., 1978).
Comparing to the objective odor sensations, the subjective odor effects of n-butanol is always ignored. The odor effect is reflection of physiological and psychological dimensions underlying odor perception space (Jellinek, 1997;Zarzo and Stanton, 2009). The semantic differential (SD) method is frequently used to describe olfactory perception space. The technique is applied to measure people's affective responses to stimulus in terms of ratings of bipolar scales defined with adjectives on each end, such as “cool-warm”, “heavy-light” and so on. Following principal component and factor analysis (PCA and FA), the SD technique appears to yield reduced factors to profile the olfactory perception space. The SD technique may be useful in evaluating the n-butanol odor characters at different concentrations, especially in a range between 100 ppm and 30,000 ppm, which are adopted as odor reference.
The present work is to study the relationship between n-butanol odor intensities and concentrations, and their odor characters evaluated by Korean panelists. A good odor reference should have no characteristic smell similar to other odorants, no perception interference by similar odor for sensory evaluation. Since culture affects the odor perception and odor classification (Chrea et al., 2004), we present Korean's n-butanol odor perception results using SD technique following PCA and investigate its suitability as Korea's odor intensity reference.
2. Materials and methods
2.1 Materials and panel
The odorant for this study is n-butanol, with a purity of 99.5% and free of strong odorous impurities, which has been used as odor reference in many countries for decades. The n-butanol 5-point Odor Intensity Referencing Scale (ORIS) is adopted, and the concentrations are 100, 400, 1,500, 7,000 and 30,000 ppm, respectively, according to the Korea Standard Industrial Odor Test Method (Ministry of Environment, 2014). The olfactometry panel comprises18 unexperienced and non-smoking college students (9 men and 9 women), with an average age of 19.6 ± 0.7 (woman) and 21.2 ± 1.7 (man) years old.
2.2 Odor character and sensibility evaluation
About 20 μL of the certain concentration of n-butanol solution was applied to one end of the odorless paper sheet (140 × 7 mm). The panelist smelled the paper sheet for 20 seconds, with a distance between the nose and the paper kept at about 5 cm. Then the panelist recorded odor intensity, odor type, and adjective descriptors.
The panelists were asked to finish the perception evaluation as soon as possible after smelling sample, in order to minimize the subjective influence by the surrounding environment and the previous experiences. The evaluation time for each sample was limited to about 5 minutes, and an interval of 15 minutes was required for the next evaluation. Ascending concentration series method was used when presenting an odor sample to odor panelists without prior notice.
In the sensory evaluation, the odor intensity reference scale is defined as 6 points of intensity according to the Korea Standard Industrial Odor Test Method (Ministry of Environment, 2014). The intensity was rated on a 6- point scale with 1=no odor, 2=perceptible, 3=moderate, 4=strong, 5=very strong, 6=overwhelming. When the odor intensity was judged to be higher than level 2, subjects were asked to describe the odor quality. The odor qualities were freely described without restriction of specific objects, nouns, vocabulary, and phrases that was instantaneously perceived.
2.3 Semantic differential method
The semantic differential was chosen as the measuring device for the subjective evaluation. The bi-polar adjectives for the Korean’s sensory recognition were selected on the basis of their highest frequency of occurrence in the literature (Sohn et al., 2002;Chong et al., 2011;Jung et al., 2015). As a result, the following bipolar adjectives were used: unpleasant-pleasant (A1) (Carrasco and Ridout, 1993;Chrea et al., 2004), masculine- feminine (A2) (Jung et al., 2015), sultry-fresh (A3) (Zarzo and Stanton, 2009;Jung et al., 2015), excitingcalming (A4) (Jung et al., 2015), greasy-refresh (A5) (Jung et al., 2015), cool-warm (A6) (Carrasco and Ridout, 1993;Jung et al., 2015), heavy-light (A7) (Jung et al., 2015), bitter-sweet (A8) (Zarzo and Stanton, 2009;Jung et al., 2015), pungent-soft (A9) (Yuwono and Lammers, 2004;Zarzo and Stanton, 2009), musty-clean (A10) (Zarzo and Stanton, 2009;Porcherot et al., 2010). An eleven points of semantic differential scale was used on a -5 ~ +5 scale, e.g., very unpleasant (-5), quite unpleasant (-4), unpleasant (-3), slightly unpleasant (-2), slightly unpleasant (-1), neutral (0), slightly pleasant (+1), slightly pleasant (+2), pleasant (+3), fairly pleasant (+4), and very pleasant (+5).
To categorize and name the elements extracted from the factor analysis, the results of previous studies on sensibility structure were summarized (Table 1). Sohn et al. (2002) categorized 50 sensibility words for cedarwood oil, peppermint oil, grapefruit oil, and teebaum oil scents into 10 factors (intensity, personality, romanticism, esthetics, femininity, naturality, ruralness, passion, fragrance, flower scent). Kweon et al. (2005) categorized 50 sensibility words for freesia and peach into 8 factors (antiquity, intensity, romanticism, esthetics, naturality, passion, uniqueness, excitement); and Min et al. (2007) classified 50 sensibility words for methyl cyclopentenolaone, isovaleric acid, and γ-undecalactone into 5 factors (intensity, complexity, esthetics, gentleness, and activity). Among a total of 17 factors, 5 factors (intensity, romanticism, esthetics, naturality, and passion) are commonly applied in each study, some sensibility vocabulary only belong to a single factor such as ‘deep’ and ‘thick’, and others belong to 2 or more factors such as ‘intense’ and ‘strong’.
Among the 17 factors, the following factors: esthetics (42 words), intensity (20 words) and naturality (13 words) include the most sensibility vocabulary. The amount of sensibility vocabulary in esthetics is more than twice that of other factors, including a number of words that also belong to other factors such as ‘light’ and ‘elegant’, as well as those that belonged only to esthetics such as ‘aggressive’ and ‘bad feelings’. The standard Korean language dictionary of the national institute of Korean language defines ‘esthetics’ as ‘the study and pursuit of beauty’, and it would be difficult to directly relate this to sensibility vocabulary such as ‘light’, ‘rough’, and ‘aggressive’, which were applied in previous studies. Therefore, esthetics in previous studies is thought to include diverse vocabulary rather than dictionary definitions, and indicate a characterless state.
Principal component analysis was carried out using IBM™ SPSS™ (Version 22) with varimax rotation on the SD adjective pairs in order to extract the number of factors present in the data and to identify which descriptors loaded most highly on each factor. In general, the criteria for selecting factors and questions in social science are considered significant variables when their eigenvalue is 1.0 or higher, and factor loading (commonality) is 0.4 or higher; therefore, they are chosen as analysis standards for this study.
3. Results and discussion
3.1 n-Butanol odor intensities and odor character profiles at different concentrations
The results of the sensory evaluation of odor intensities and self-report types according to five different nbutanol concentrations, are shown in Table 2. Among 18 subjects (9 women/9 men), four of them (2/2, women/ men) evaluated the odor intensity at 100 ppm as 0; 4 (3/ 1) as 1; 7 (3/4) as 2; and 3 (1/2) as 3, with the average being 1.5 ± 1.0 (1.3 ± 1.0/1.7 ± 1.1). Perceived odor types are diverse. Women perceive the odor as medicine, adhesive, pain relieving patch, and chocolate, while men per- ceive the odor as feces, feet, expired milk, urine, and alcohol. The odor descriptor wheels (McGinley et al., 2000) of n-butanol are shown in Fig 1. All of the charts stretch upwards, corresponding to characters as “odorless”, “offensive”, “chemical”, and “medicinal”. Odor profiles at low concentration are mostly presented as “odorless” or “offensive”, while described as “chemical” or “medicinal” frequently at high concentration.
As the concentration increased, the intensity perceived by both men and women tended to increase, and the correlation coefficients of n-butanol concentration and odor intensity are 0.97 and 0.90, for women and men, respectively. At concentration levels of 100 ppm, 7,000 ppm, and 30,000 ppm, women demonstrate an intensity level approximately 0.9 times lower than that of men; at 15,000 ppm, it is approximately 1.2 times higher than that of men; and at 400 ppm, both men and women evaluate the same intensity level. Therefore, there are differences between men and women according to concentration level, but the overall difference is not great. As for perceived odor type, subjects perceive the odor very differently from one another. However, the ratio of odor types perceived as a bad smell, such as medicine, feces, feet, expired milk, and mold, is higher than smells perceived as a fragrance, such as the scent of flowers or chocolate, and the variety is great.
As for the conditions of standard odorant, it should be linearity in odor intensity according to concentration change, should be possible to apply it to both men and women, and should not have a characteristic smell because it should be a standard for varying degrees of intensity of diverse odors. The results of the general sensory evaluation show that the correlations between the concentration of n-butanol and odor intensity are over 0.9 for men and women. The good linearity between intensity and concentration, almost no gender difference, diverse odor types rather than a specific one, easy to make a wide range of concentrations, make n-butanol could be considered as a suitable standard odorant.
3.2 Sensory descriptors evaluation according to nbutanol concentration
The scale of sensibility of the five n-butanol concentrations were evaluated in a range between -5 and +5. The results of the sensibility evaluation, limiting the sensibility vocabulary to 10 pairs (20 adjectives) for evaluation, are shown in Table 3. As for opposite words, the (-) value is assigned to the left-side vocabulary and the (+) value to the right-side vocabulary for the odor in question. A greater absolute value indicates a greater sensibility for the corresponding vocabulary. For 100-ppm nbutanol, the average value determined by 18 subjects for the item ‘unpleasant-pleasant (A1)’ on the -5 to +5 scale is -0.33. The sensibility is therefore ‘unpleasant’ on the ‘unpleasant-pleasant’ emotion scale. The average value of nine women is 0.56, evaluating it predominantly as a ‘pleasant’ emotion, and that of nine men is -1.22, evaluating it predominantly as an ‘unpleasant’ emotion.
There are differences in the sensibility according to the 10 evaluation items, but it generally tends to increase in accordance with increase of n-butanol concentration. Among the 10 sensory evaluation items, the coefficients of correlation are 0.9 or higher, showing a good correlation with ‘unpleasant-pleasant (A1)’, ‘masculine-feminine (A2)’, ‘bitter-sweet (A8)’, and ‘pungent-soft (A9)’ for women, and with ‘exciting-calming (A4)’, ‘coolwarm (A6)’, ‘heavy-light (A8)’, ‘pungent-soft (A9)’, and ‘musty-clean (A10)’ for men. Some items with low correlation, such as ‘sultry-fresh (A3)’, have a sensibility of generally less than 1 and approach neutrality, indicating that they do not belong to neither vocabulary. This is probably because the subjects did not perceive the corresponding sensibility from the tested n-butanol odor. Sensibility vocabulary types with low correlation, therefore, are not suitable for evaluating the subjective characteristics of n-butanol odor.
A paired T-test was conducted to assess whether there was any difference in the scale of sensibility for nbutanol between men and women. As the p-values of all concentrations and sensibility evaluation items are greater than 0.05, there are no significant difference in the scales of sensibility felt by men and women. As for the sensibility value, the ratio of left-side strong negative vocabulary, such as ‘unpleasant’, ‘bitter’, and ‘musty’, is greater than that of right-side positive such ‘pleasant’, ‘sweet’, and ‘clean’. Though few positive values are observed at low concentration (100 ppm), and for low relevancy sensibility evaluation items such as ‘cool-warm (A6)’, most of the sensibility values are negative overall. Therefore, n-butanol odor is considered as a negative emotional odor rather than a positive one.
As a result of emotional assessment by limiting sensibility vocabulary, the linearity of the sensibility scales and concentration of n-butanol is good. There is no emotional difference between men and women, and nbutanol is assessed as a negative emotional odor. Therefore, it can be concluded that n-butanol is suitable as odor intensity reference odor.
3.3 Characteristics of sensibility structure according to n-butanol concentrations
Factor analysis was performed using the results of the sensibility evaluation in order to identify the sensibility structure characeristics according to the 5 n-butanol concentrations. Principal Component Analysis (PCA), which extracts factors based on total variance, was used to extract components, and Varimax was adopted to simplify factor loadings.
The results of sensibility structure characteristics according to n-butanol concentration based on the factor analysis are shown in Table 4. Factors are named using the results of previous studies, for example, a factor including ‘unpleasant/pleasant’ is classed as an ‘esthetics’ factor, which encompasses ‘unpleasant’ and ‘bad feelings’, and a factor including ‘sultry/fresh’ and ‘heavy/ light’ is classed as an ‘excitement’ factor, which encompasses ‘light’, ‘frivolous’, and ‘exciting’. The total sensibility structure of 100 ppm of n-butanol is identified by two factors: ‘esthetics’ and ‘passion’ (explained variance: approximately 95.5%). For women, the two factors are ‘esthetics’ and ‘passion’ (approximately 91.7%), while for men, the two factors are ‘esthetics’ and ‘gentleness’ (approximately 81.9%).
The overall sensibility structures according to nbutanol concentration are categorized into 7 factors ‘romanticism’, ‘esthetics’, ‘gentleness’, ‘naturality’, ‘passion’, ‘activity’, and ‘excitement’; and the components are diverse. ‘Esthetics’ is the factor with the highest frequency (15 times), followed by ‘naturality’ (7 times), ‘passion’ (6 times), and ‘gentleness’ (4 times). ‘Esthetics’ is thought to be the primary factor as the level of its explained variance is highest, which is similar to the result of Sohn et al. (Sohn et al., 2002) , indicating that the four factors ‘esthetics’, ‘naturality’, ‘intensity’, and ‘personality’ are stable factors in the fragrance sensibility structure. In terms of men and women, the five factors ‘romanticism’, ‘esthetics’, ‘gentleness’, ‘naturality’, and ‘passion’ are observed by both genders, with the addition of ‘activity’ for women and ‘excitement’ for men.
The representative factors of the n-butanol sensibility structure according to the sensibility evaluation are expressed with diverse sensibility vocabulary, and it is concluded that ‘esthetics’ represent its characterless nature. Among the respective five factors of men and women, the four factors ‘romanticism’, ‘esthetics’, ‘gentleness’, and ‘naturality’ are commonly observed. This result is considered to satisfy the conditions for a standard odorant i.e. that it should not have a characteristic smell and can be commonly applied to both men and women.
4. Conclusion
The odor quality of n-butanol, which generally used as an odor intensity referencing odor both at home and abroad, was evaluated by sensory evaluation as a selfreport vocabulary. The suitability of n-butanol as odor reference was further assessed for its odor effect with a limit of 20 emotional vocabularies.
As a result of sensory and emotional evaluations, it was found that there was a good correlation (r > 0.9) for odor intensity and emotional measure scale with respect to concentration of n-butanol, and the correlation between emotional measure and smell intensity was low for both gender. As a result of the sample T-test, it was confirmed that there was no difference in the emotional measure according to the concentration of n-butanol for both gender. In addition, the perceived odor types were various and no typical odor smell. The evaluated sensory vocabulary were negative. As a result of emotional evaluation, the n-butanol's odor effects are expressed in a variety of emotional terms, and the featureless "aesthetic" is a representative factor.
n-Butanol has no typical odor smell, and its odor effect is featureless aesthetic. It is easy to prepare various solution of wide range concentrations. It also has very good linear relation between odor intensity and concentration. There is no significant difference in odor sense and odor effect for both gender. Therefore, nbutanol is very suitable to be used as odor intensity reference.
