2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

th

OPTIMIZATION OF PERFUME COMPOSITIONS

1

Vera G. Mata1*, Paula B. Gomes1, Alrio E. Rodrigues1

LSRE Laboratory of Separation and Reaction Engineering, Faculdade de Engenharia da Universidade do Porto, Portugal

Abstract. Perfumes are complex mixtures of substances having very different chemical properties. Considering one of the most important properties, from the perfumer point of view the volatility there are a great range of materials, classified into three classes: Top notes (the more volatile compounds; they can be in a perfume only a few minutes); Middle notes (they give the character to a perfume and can last a few hours); Base notes (the less volatile compounds; can last more than eight hours and are used as fixatives of the top and middle notes). Carles (1962) represented the structure of a perfume as a triangle divided horizontally in three parts, each one representing top, middle and base notes. The suggested proportions for each type of fragrant notes are: top - 15-25%; middle 30-40% and base: 45-55%. However, one of the difficulties that a perfumer feels is the right choice of each fragrant compound, not in a pure solution, but taking into account its behavior in a perfume mixture. Generally, this choice is based on trial and error experiments and subjective judgments. In this work, a new theoretical methodology to preview the behavior of ternary perfume mixtures in a solvent base is described. It is based on the new strategy called Perfumery Ternary Diagrams (PTD) that results from an analogy between Perfume Pyramid Structure and Engineering Ternary Diagrams. These diagrams allow a fast evaluation of the odor in the gas above a perfume, for all possible liquid mixture combinations of any perfume constituted by three fragrant components and one or more solvents. This methodology can help perfumers in the in the optimization of perfume compositions, eliminating this way some trial and error time and chemical wastes. Keywords: Perfumery Ternary Diagram, chemical product design, engineering perfumes.

Introduction
The traditional classifications for single flavors and fragrances were made in terms of their volatilities (Poucher, 1955; Carles, 1962). However, the behavior of one fragrant chemical or essential oil, and consequently its odor value, OV, depends largely on the medium or solvent and on the remaining fragrant components present in the perfume mixture (Behan and Perring, 1987). This is due to molecular size and in great extent to physical interactions at the molecular level, like polarity forces (ion-dipole, dipole-dipole, London forces and hydrogen bonding forces). Generally, the affinity of a molecule to its surrounding medium is measured using the concept of activity coefficient,

. High

value of

implies an increased tendency for a implies a low concentration in the

given substance to be pushed-out of the mixture and low value of

headspace and higher retention of that substance on the liquid mixture due to the higher affinity to the medium (Behan and Perring, 1987). This means that one smell can change if diluted in different solvents, since the volatility of the single compound and its odor value, OV, will be affected in different ways. This is very important in perfume formulations once the amount of alcohol, water or other solvent/cosmetic base can influence largely the headspace above the liquid and, therefore, its smell. In this work, the theoretical methodology named by Perfumery Ternary Diagram (Mata et al., 2005a,b,c) is described and applied to a test case perfume mixture constituted by limonene, linalool and geraniol. Then, the effect of the addition of ethanol to the test case perfume mixture is shown, as well as the addition of a fixative to

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

the same test sample. It can be seen that this methodology allows a fast theoretical headspace evaluation of all combinations of the three fragrant components and for a fixed amount of one solvent and/or fixative.

Gas-Liquid Equilibrium theory

A perfume system consists in a complex fragrant liquid mixture and a corresponding air phase above it, the headspace. In this work, it is considered that: (i) For every component (Reid et al, 1987):

i in the mixture, the condition of thermodynamic equilibrium is given by

f i g = f il

[1]

where f i is the fugacity of component i , superscripts g and l are gas and liquid phases, respectively. For a mixture of ideal gases, the fugacity value for the gas phases is equal to:

f i g = yi P

[2]

where y i is the gas phase molar fraction of component i and P is the total gas phase pressure. The fugacity of component i in the liquid phase is related to the composition of that phase through the activity coefficient,

i . For subcritical conditions,

[3]

f i l = i xi Pi sat
where

xi is the liquid phase molar fraction of component i in equilibrium with the headspace y i and Pi sat is the saturated vapor pressure of component i .

gas molar fraction,

Substituting Equation 2 and 3 in Equation 1, it results in:

yi =

i xi Pi sat
P

[4]

In this work, values of

i were calculated using the UNIFAC method, which is based on

- and the

molecular group contribution (Fredenslund et al., 1975). This method takes into account the contribution due to differences in molecular size and shape combinatorial part contribution due to molecular interactions residual part. (ii) Once it is considered that the headspace is an ideal solution, the concentration of component the headspace,

i in

cig , is given by:

[5]

cig =

yi M i P RT

To whom all correspondence should be addressed. Address: LSRE FEUP, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal. Tel: +351225081669, Fax: +351225081674, E-mail: vmata@fe.up.pt

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

where

M i is the molecular weight of i , R is the ideal gas constant and T is the absolute

temperature (Kelvin).

The Odor Value Concept

The intensity of a fragrant compound i can be expressed in terms of its odor value, (Calkin and Jellinek, 1994):

OVi , which is defined as

OVi =

cig Thri

[6]

where Thri is the threshold concentration of 2004). Combining Equations 4-6, then we get:

i in air (Devos et al., 1990; Leffingwell and Leffingwell,

P sat M i OVi = i xi i Thr i

1 RT

[7]

This equation relates the liquid perfume composition, xi with the human sensory response of the evaporated perfume. In this work it is considered that one component i , present in a liquid mixture of perceived strongly by human nose when its odor value, those of the others components.

N fragrant components, is

OVi , in the headspace above the liquid is higher than

Perfumery ternary diagram (PTD) concept

Now, let us consider a ternary liquid mixture, A+B+C, constituted by three chemical compounds having different volatilities, each one belonging to one type of fragrant notes, i.e., A represents a top note, B a middle note and C a base note. In some cases, it is also considered that a certain amount of solvent, S, in this case ethanol is added to the mixture, resulting this way in a quaternary mixture, that will be called test perfume. The amount of each component in the test perfume is given by its molar fraction: x A , x B , xC and x S and we will call this mixture X. To simplify, this quaternary system is transformed into a ternary system, by recalculating the molar fractions for A, B and C in a solvent free basis, resulting in the new fractions:

xA '=

xC xA xB , xB ' = and xC ' = . x A + x B + xC x A + x B + xC x A + x B + xC

These three quantities are then represented in a ternary diagram (Mata et al., 2005) as it is shown in Figure 1. Considering a general liquid mixture, which composition is represented by point X ' , the composition (solvent-free basis) of component A, x A ' , is given by the distance X ' A' , the composition of B, x B ' , is given 3

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

by the distance X ' B ' and the composition of C, xC ' , is given by the distance X 'C ' . It is considered that the mole fraction of solvent, x S , is known and constant for each ternary diagram. As it can be seen, the ternary diagram represents all liquid compositions that could be made experimentally using different quantities of A, B and C, for a given amount of ethanol, S. Now, for each liquid perfume mixture, X ' (i.e., each point in the ternary diagram) the odor value of each component in that mixture, OVi , is calculated using Equation 7. After the calculation of all the individual odor values, for a given mixture X ' (represented in the ternary diagram) the maximum odor value is determined:

OVmax = max(OV A , OVB , OVC , OVS ) for point X '

[8]

After the determination of the component that has the OVmax for a large number of points of the ternary diagram, we obtain as a result the Perfumery Ternary Diagram (PTD), as shown in Figure 1. In this diagram three different areas can be distinguished, giving some highlight to the following typical composition points: Pt :

OVMAX = OV A (dark shaded area); Pm : OVMAX = OV B (light shaded area) and Pb : OVMAX = OVC
(dotted area).

Pt
C
' xC

B X
' xB

x 'A

Pm

Pb
A

B
FIGURE 1.

Scheme of a Perfumery Ternary Diagram (PTD) and representation of ternary mixture compositions.

There can exist also points for which solvent has the highest odor value: OVMAX = OVS . These points are referred as Ps . This situation happens generally for mixtures having a high content of solvent, as it will be seen below.

Test case
The Perfumery Ternary Diagram (PTD) strategy was tested using the following ternary system: limonene (A), linalool (B) and geraniol (C). Limonene (A) represents a top note: it is present in citrus peel oil at a concentration of over 90% and having a lemon-like odor; Linalool (B) represents a middle note: it is present in many essential oils, like rosewood oils and lavender oils; it has a flowery-fresh-odor reminiscent of lily of the 4

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

valley; Geraniol (C) represents other middle note: it is present in geranium and rose oils; it has a floweryroselike odor. In this first example, the presence of solvents and base notes are not considered, meaning that xi = xi ,
'

i = A, B, C . This mixture can be a representation of a perfume concentrate with A+B+C or a perfume some
seconds after application, the time that is needed to evaporate the ethanol. Figure 2 shows the resulting Perfumery Ternary Diagram (PTD) for this system.
A

FIGURE 2.

PTD for the test case system: limonene:

; linalool:

; geraniol

The effect of one solvent ethanol (S) was then studied and the result can be easily visualized using the PTD concept. Figure 3 shows the change in the odor value for all possible compositions of A+B+C+S. The percent of ethanol is fixed in each PTD and equal to: (a) 30% S; (b) 50% S and (c) 70% S. Comparing Figure 2 with Figure 3a, it can be seen that there was an increase in the percent of points where the maximum odor value, OV, is that of limonene and a corresponding decrease in the percent of points where both linalool and geraniol has a maximum OV. This happens because ethanol is a strong polar solvent and limonene is a non-polar compound. When ethanol is added to the solution the non-polar compound tend to leave the solution (that is predominantly polar). Increasing the percent of ethanol (Figures 3b and 3c) this effect continues but when the percent of ethanol is higher or equal to 50% it appears points for which the maximum odor value is that of the ethanol. As it was described above, the case study is constituted by one top note and two middle notes. We will study then the effect of adding one base note to the case study system, having 50 % of ethanol (the example shown in Figure 3b). The base note used is tonalide, a strong non-polar component. It is a synthetic musk fragrance, and is also used as a fixative. Figure 4 shows the effect of adding tonalide at a fixed percent of (a) 10% and (b) 20%. Comparing Figure 3b with Figure 4a, it can be seen that the number of points where the maximum odor value is that of ethanol increased. This happens because tonalide is a very non-polar compound and will push-out the more polar compounds in the mixture, in this case, the ethanol. Now, we can compare the increase in ethanol (Figure 3a-c) to the increase in tonalide. In the first case (Figure 3), we can notice that as the percent of ethanol is increased, the compositions where the OV have a maximum 5

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

start in the middle of the Perfumery Ternary Diagram (PTD), but expands to the points where initially (without ethanol) the odor value maximum was that of linalool (B) and geraniol (C). The reason is because B and C are the less volatiles in the system, and when the concentration of ethanol increases, it tends to push-out the less polar compound, limonene, and retain linalool and geraniol because both are polar compounds, having more affinity to the ethanol and consequently both will be retained to a higher extent in the solution.

(a) 30% ethanol

A

(b) 50% ethanol

A

(c) 70% ethanol FIGURE 3. Effect of Ethanol for the test case system: limonene: ; linalool: ; geraniol: ; ethanol:

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

On the other hand, when tonalide is increased (Figure 4a-b), although again the compositions where the OV have a maximum start in the middle of the Perfumery Ternary Diagram (PTD), when the percent of tonalide increases, the points where the odor value maximum is that of ethanol expands is the direction of the three corners of the PTD. This happens for two reasons. First, because tonalide is a non-polar compound, such as limonene. This means that tonalide tend to retain limonene in solution more strongly than it happen when adding ethanol. Second, because the difference in polarities between ethanol and tonalide increases as more tonalide is added, and then the tendency for ethanol leave the solution increases as the liquid solution becomes less polar.

(a) 50% ethanol +10% tonalide

A

(a) 50% ethanol +20% tonalide

FIGURE 4.

Effect of tonalide for the test case system having 50% ethanol: limonene: geraniol: ; ethanol:

; linalool:

Conclusions
A methodology to predict the behavior of non-ideal ternary perfume mixtures in a given solvent was developed, using the concept of Perfumery Ternary Diagrams (PTD). This method allows the prediction of the characteristic smell in the headspace above a ternary perfume mixture in a solvent, described as the odor value, OV, for any combination of ternary fragrant mixtures in that solvent. The liquid is considered to be non-ideal 7

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

and the non-idealities due to molecular interactions are accounted for by the activity coefficients for each component in mixture, calculated by the UNIFAC method. As illustration, one test case perfume mixture was studied: (A) limonene, (B) linalool and (C) geraniol. The effect of the addition of ethanol (S) to the test case perfume mixture was shown, as well as the addition of a fixative (tonalide) to the same test sample. It can be concluded that these Perfumery Ternary Diagrams (PTD) are important tools for perfumers, to take important decisions, such as: to control the intensity of the top note, to find automatically which compositions will have an intense smell to the top, middle or base note and to find the maximum quantity of ethanol for a given mixture.

References
Behan, J.M.; Perring, K.D.; Perfume interactions with sodium dodecylsulphate solutions. Int. J. Cosmet. Sci. 1987, 9, 261. Calkin, R.R.; Jellinek, J.S.; Perfumery: Practice and Principles, John Wiley and Sons: NY , 1994. Carles J.; A method of creation in perfumery. Soap Perf. Cosm 1962, 35, 328. Devos, M.; Patte, F.; Rouault, J.; Laffort, P.; van Gemert, L.J.; Standardized Human Olfactory Thresholds. IRL Press: Oxford, 1990. Fredenslund, A.; Jones, R.L.; Prausnitz, J.M.; Group-Contribution Estimation of Activity Coefficients in Nonideal Liquid Mixtures, AIChE J., 1975, 21(6):1086-1099. Leffingwell, D.; Leffingwell, J.C.; Odor & Flavor Detection Thresholds in Water (In Parts per Billion). http://www.leffingwell.com/odorthre.htm, 2004. Mata, V.G.; Gomes, P.B.; Rodrigues, A.E.; Perfumery Ternary Diagram (PTD): a new concept applied to the optimization of perfume compositions. Flavour and fragrance J., 2005. In press. Mata, V.G.; Gomes, P.B.; Rodrigues, A.E., Effect of non-idealities in perfume mixtures using the Perfumery Ternary Diagrams (PTD) concept. Ind. Eng. Chem. Res. 2005. Accepted for publication. Mata, V.G.; Gomes, P.B.; Rodrigues, A.E., Engineering Perfumes. AIChE J. 2005. Accepted for publication. Poucher, W.A.; A classification of odors and its uses. The American Perfumer & Essential Oil Review, 1955, July, 17. Reid, R.C.; Prausnitz, J.M.; Poling, B.E.; The properties of Gases and Liquids. McGraw-Hill, NY, 1987.

2nd Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering
th

Acknowledgments
We gratefully acknowledge the financial support of FCT - Fundao para a Cincia e Tecnologia and FEDER, project grant POCTI/EQU/39990/2001, pos-doc grant SFRH/BPD/1576/2000 and PhD grant SFRH/BD/968/2000.