COLD TREATMENT

Abstract

The present invention relates to a new high-throughput process for reducing impurities in essential oils and extracts (in particular for fragrances, fragrance ingredients, flavours and cosmetic ingredients) under mild conditions. Undesirable natural components such as waxes, but also synthetic materials such as agrochemicals and other environmental pollutants are reduced by using at least one selective nanofiltration membrane. In addition, the present invention relates to a method for reducing coloured components in essential oils to obtain a less coloured or even colourless essential oil, while achieving high re-colouration stability over time. Further, the odour quality is maintained or increased through reduction of undesirable olfactory substances to achieve a purified and higher quality oil.

Claims

1. A nanofiltration process to provide a less coloured essential oil for a fragrance, flavour or cosmetic ingredient, comprising the following steps: (i) providing a selectively permeable thin film composite (TFC) nanofiltration membrane, wherein the membrane consists of a support layer and a top layer, (ii) providing a flowable input of essential oil, optionally with a solvent component; (iii) separating the flowable input by transferring it across the surface of the membrane to form a retentate and a permeate, such that the concentration of one or more components of the permeate is reduced compared to the flowable input; and wherein the permeate is decoloured in comparison to the flowable input, such that the lightness value L* of the permeate is increased in comparison to the flowable input to ΔL* greater than or equal to 1 and the chromaticity C* of the permeate is decreased in comparison to the flowable input to ΔC* less than or equal to −2, according to the CIELAB colour measurement system, as specified by the International Commission on Illumination.

2. The process according to claim 1, wherein the lightness value L* of the permeate is increased in comparison to the flowable input to ΔL* greater than or equal to 1.5, and/or the chromaticity C* of the permeate is decreased in comparison to the flowable input to ΔC* less than or equal to −4.

3. The process according to claim 1, wherein the colour stability of the permeate is such that the re-colouration measured by the change in lightness of the permeate 48 h, preferably 20 days, most preferably 1 year, after nanofiltration in comparison to the permeate just after nanofiltration is only decreased to ΔL* greater than or equal to −1.0, preferably greater than or equal to −0.5 and/or the chromaticity C* of the permeate 48 h, preferably 20 days, most preferably 1 year, after nanofiltration is only increased in comparison to the permeate after nanofiltration to ΔC* less than or equal to 10, preferably less than or equal to 5, most preferred less than or equal to 1.

4. The process according to claim 1, wherein the TFC nanofiltration membrane does not include a nitrogen-containing polymer.

5. The process according to claim 1, wherein the support layer of the TFC nanofiltration membrane comprises a polymer, which includes one or more of the heteroatoms O, N, S, and/or halogen, and/or Si, more preferably a polymer including the heteroatom O and/or S.

6. The process according to claim 1, wherein the support layer comprises a material chosen from the group consisting of: polydimethylsiloxane, polyoctylmethylsiloxane, poly[1-(trimethylsilyl)-1-propyne], polytetrafluoroethylene, polysulfone, polyethersulfone, polyvinylidene fluoride and polyetheretherketone.

7. The process according to claim 1, wherein the top layer comprises a material chosen from the group consisting of: polydimethylsiloxane, polyoctylmethylsiloxane, poly[1-(trimethylsilyl)-1-propyne], poly(2,6-dimethyl-1,4-phenylene oxide), polyacrylacid, polymer of intrinsic microporosity (PIM-1), polystryrene-b-poly(ethylene oxide) diblockcopolymer, poly(sodiumstyrenesulfonate (PSS), and polyvinylsufate (PVS), and mixtures thereof.

8. The process according to claim 1, wherein the top layer is a silicone coated organophilic layer, preferably a cross-linked polydimethylsiloxane.

9. The process according to claim 1, wherein the TFC nanofiltration membrane has a molecular weight cut off between 150 g/mol and 1200 g/mol.

10. The process according to claim 1, wherein the flowable input comprises an essential oil and a solvent, wherein the solvent comprises at least 10 wt. % organic solvent component, and optionally water, and the organic solvent component has a dipole moment of at least 3*10.sup.−30 Cm.

11. The process according to claim 10, wherein the organic solvent used to prepare the flowable input has a dipole moment of at least 4*10.sup.−30 Cm, preferably the organic solvent is chosen from the group consisting of: diethyl ether, ethanol, isopropanol, ethyl acetate, methylethylketone, butylacetat, methyl-tert-butyl-ether, cyclohexanol and acetone, more preferably the organic solvent is methyl-tert-butyl-ether or ethanol.

12. The process according to claim 10, wherein the solvent further comprises a second organic solvent component, which has a dipole moment of less than 2*10.sup.−30 Cm, preferably hexane or heptane.

13. The process according to claim 1, wherein, the flow rate through the first membrane is at least 8 kg [permeate]/h*m.sup.2 [membrane].

14. The process according to claim 1, wherein the essential oil derives from the genus Citrus, such as preferably chosen from the group consisting of: sweet orange, orange, lemon, lime, grapefruit, bergamot, key lime, pomelo, citron, mandarine, tangerine, bitter orange, blood orange and/or wherein the essential oil is selected from the group consisting of: peru balsam oil, benzoin siam oil, patchouli oil, rose oil, ylang ylang oil, clove leaves, lemon oil, oak moss absolute and vanilla extract.

15. The process according to claim 14, wherein the essential oil is chosen from the group consisting of: mandarine oil, peru balsam, tangerine oil, blood orange oil, patchouli oil, vanilla extract and benzoin siam oil.

16. The less coloured permeate obtained by the process of claim 1.

17. A fine fragrance composition, a fragrance formula, a flavour or cosmetic ingredient, or a shower gel comprising the less coloured permeate obtained by the process of claim 1.

18. A system for performing the nanofiltration process for the purification of essential oils according to claim 1, comprising a selectively permeable thin film composite (TFC) nanofiltration membrane with a support layer and a top layer, wherein the TFC nanofiltration membrane does not include a nitrogen-containing polymer; wherein the top layer is a polydimethylsiloxane; a flowable input of essential oil; wherein the flow rate of the flowable input through the first membrane is at least 8 kg [permeate]/h*m.sup.2 [membrane]; and wherein the TFC nanofiltration membrane has a molecular weight cut off between 350 g/mol and 500 g/mol.

19. The process according to claim 1, wherein the flowable input comprises a solvent.

20. The process according to claim 1, wherein the lightness value L* of the permeate is increased in comparison to the flowable input to ΔL* greater than or equal to 2.0, and/or the chromaticity C* of the permeate is decreased in comparison to the flowable input to ΔC* less than or equal to −10.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0139] FIG. 1 is a schematic diagram of the cross-flow nanofiltration system, which explains the terms Input (A), Permeate (P) and Retentate (R) as used in the examples.

[0140] FIG. 2 shows the differences between Input (A), Permeate (P) and Retentate (R) Samples of Mandarine oil before and after nanofiltration according to Example 1.

[0141] FIG. 3 shows the HPLC/MS-analysis of Input (A), Permeate (P) and Retentate (R) Samples of Mandarine oil.

[0142] FIG. 4 shows the differences between Input (A), Permeate (P) and Retentate (R) of Vanilla Extract before and after nanofiltration according to Example 2.

[0143] FIG. 5 shows the differences between Input (A), Permeate (P) and Retentate (R) of Patchouli oil (Example 3a) before and after nanofiltration.

[0144] FIG. 6 shows the differences between Input (A), Permeate (P) and Retentate (R) of Patchouli oil with solvent (Example 3b) before and after nanofiltration with solvent.

[0145] FIG. 7 shows the differences between Input (A), Permeate (P) and Retentate (R) of Orange oil Brasil (Example 4) before and after nanofiltration.

[0146] FIG. 8 shows the differences between Input (A), Permeate (P) and Retentate (R) of Peru Balsam (Example 6) before and after nanofiltration.

[0147] FIG. 9 shows the differences between Input (A), Permeate (P) and Retentate (R) of Blood Orange oil Italian (Example 7) before and after nanofiltration.

[0148] FIG. 10 shows the differences between Input (A), Permeate (P) and Retentate (R) of Tangerine oil Italian (Example 8) before and after nanofiltration.

[0149] FIG. 11 shows the differences between Input (A), Permeate (P) and Retentate (R) of Lemon Oil Italian (Example 9) before and after nanofiltration.

[0150] FIG. 12 shows the differences between Input (A), Permeate (P) and Retentate (R) of Clove leaves (Example 10) before and after nanofiltration.

[0151] FIG. 13 shows the differences between Input (A), Permeate (P) and Retentate (R) of Ylang Ylang (Example 11) before and after nanofiltration.

[0152] FIG. 14 is a schematic diagram (Flow chart) of the cross-flow nanofiltration system describing the purification process of the essential oils, to produce Permeate (P) and Retentate (R), working with external pressure (N.sub.2), containing one flat sheet crossflow filtration cell.

[0153] FIG. 15 is a schematic diagram (Flow chart) of the cross-flow nanofiltration system describing the purification process of essential oils, producing Permeate (P) and Retentate (R), working with or without external pressure (N.sub.2) (in-build pump develops pressure), containing one spiral filtration cell.

METHODS

[0154] I. Colour Determination of Transparent Liquids (L*a*b*) with Spectral Photometer.

[0155] This method serves to determine the colour for transparent liquids using a spectral photometer (Lico 400 compact) from Hach Lange.

[0156] Type: Scanning grating spectrophotometer with reference beam path

[0157] Viewing geometry 0°/180° (transmitted light)

[0158] Spectral range Colour measurement: 380 nm-780 nm/10 nm,

[0159] X (red), Y (green) and Z (blue)-transmission calculated for standard illuminant C and 2° standard observer (DIN 5033)—Colour measurement (also ASTM E 170, ISO 7724)

[0160] In the L* a* b* colour system, L* is the lightness factor, a* and b* are the chromaticity coordinates (shade, saturation). a* is the position on the red-green axis and b* the position on the yellow-blue axis.

[0161] Precision validation: To ensure precision or repeatability the absolute and relative standard deviation of 6 independent measurements were calculated for the example of the 100140 lemon oil ital., lot 82 (Lico 500; QC0155). This analytical method is valid for the SAP methods AFW060 and AFW061.

[0162] II. Pesticide Screening:

[0163] The results of the above mentioned analyses are in accordance with the requirements of regulation (EC) 396/2005 (regulation on maximum residue levels of pesticides in or on food and feed) in its currently valid version.

[0164] The analysed sample can be classified as processed food which is concentrated during processing according to Article 20 of regulation (EC) 396/2005 (regulation on maximum residue levels of pesticides in or on food and feed). Therefore, the corresponding Maximum Residue Levels have to be calculated considering a concentration factor of 200.

[0165] The samples were analysed for pesticide composition by LC-MS-MS liquid chromatography method, GC-MS gas chromatography method, GC-FPD gas chromatography method with flame photometric detector or gas chromatography method with electron capture detector. The results for removal of select pesticides during filtration are shown in Table 2 and Table 5.

[0166] III. Photometric Determination of Iron:

[0167] The iron content was determined by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES). The Inductively Coupled Plasma (ICP) is an analytical technique to determine quantitative bulk elemental composition. The analyte species is detected and quantitated with an optical emission spectrometer (OES), which measures the intensity of radiation emitted at the element-specific, characteristic wavelength from thermally excited analyte atoms or ions. It was executed according to the DIN EN ISO 11885. The apparatus used is a Whatman, Typ Spartan 30.

[0168] The principle to determine the iron content is the reaction of Iron (III) ions with thiocyanate (rhodanide) which build a dark red colour transfer metal complex, which remains in the solution: FeCl.sub.3+6KSCN.fwdarw.K.sub.3Fe(SCN).sub.6+3 KCl

[0169] Procedure: The sample (1.00 ml) is put in a 10 ml measuring flask: 8.00 ml ethanol 96 vol % 1.00 ml water dist. with 0.05 ml HNO3 conc. and 1.00 ml reagent solution (KSCN 0.5 M). Shake well and measure the absorption at 465 nm after phase separation. Reference is the blank value. For the blank value mix 9.00 ml ethanol 96 vol % 1.00 ml water dist. with 0.05 ml HNO3 conc. and 1.00 ml reagent solution (KSCN 0.5 M). If there is a problem with the phase separation, filter the solution gently through a disposable membrane filter (Whatman, Typ Spartan 30, 0.45 μm).

[0170] To ensure precision or repeatability from the extinctions of the standard solutions and the corresponding iron contents (5 ppm-100 ppm) a calibration curve is prepared. This can be done with a photometric software, graphically or with Excel. The content of the samples is determined by reading the values off the calibration curve over the extinction value or calculated by the photometric software. This analytical method is valid for the SAP method AUV235.

[0171] IV. Dry Residue:

[0172] The dry residue is determined according to the European pharmacopoeia by the evaporation of the solvents of a weighing scale. A defined measure or volume (2.0 g) or (2.0 ml) of the extract are placed in the weighing scale and 3 h in a drying cabinet at 100 to 105° C. Then the residue is cooled down via phosphorus pentoxide or silica. The test result is defined in percent (m/m) or gram/liter.

[0173] VI. Peroxide Values:

[0174] The peroxide values were detected before and after the nanofiltration. The detection of peroxide gives initial evidence of rancidity in unsaturated fats and oils. It provides a measure of the extent to which an oil sample has undergone primary oxidation. The peroxide value is defined as the amount of peroxide oxygen per 1 kilogram of fat or oil. The SI units are defined in millimoles per kilogram (N.B. 1 milliequivalents=0.5 millimole; because 1 mEq of O.sub.2=1 mmol/2=0.5 mmol of O.sub.2, where 2 is valence).

EXAMPLES AND RESULTS

Example 1: Mandarine Oil Italian

[0175] Single-fold Mandarine oil (250 g) was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 226 g of the oil was permeated.

[0176] The permeate flux was 91.07 kg/h*m.sup.2.

[0177] FIG. 2 shows the differences between Input (A), Permeate (P), and Retentate (R) samples of Mandarine oil before and after nanofiltration

TABLE-US-00001 TABLE 1 CIELAB-Results of Mandarine Oil before and after nanofiltration CIELAB-Results: L* A* B* Flowable input 90.9 −3.0 133.9 Permeate 93.1 −10.0 128.5

[0178] Table 1 and FIG. 2 show that the colour of the permeate samples was decreased in contrast to the flowable input and the retentate. These results demonstrate the advantages of the inventive process to reduce the colour of essential oils, including citrus oils, especially Mandarine oil.

[0179] According to Tab. 1 the difference of ΔL* between permeate after nanofiltration to the flowable input is 2.2 and ΔC* could be calculated as −5.0.

[0180] The results from the fragrance test showed that the odour of the Mandarine oil is stronger, with sparking and transparent note, especially the aldehyde part is stronger.

[0181] The results from HPLC/MS-analysis of input, permeate and retentate samples of Mandarine oil are shown in FIG. 3. The intensity of impurities in the retentate are higher than in permeate, while they are reduced in the permeate as compared to the input.

[0182] Samples of the feed (flowable input), retentate and permeate solutions were taken to determine the concentration of pesticides (agrochemical residues) present in the solutions. The pesticide screening of Mandarine oil Italian shows the removal of selected pesticides and fungicides through the TFC nanofiltration in Table 2.

TABLE-US-00002 TABLE 2 Concentration of select pesticides in Input, Retentate and Permeate solutions of Mandarine Oil and maximum residue level of the pesticides Maximum Residue Compound Input Retentate Permeate Level mg/kg Organochlorine Pesticides and Pyrethroides p,p-Dicofol 0.093 0.12 — 4 Organophosphorus Pesticides (Citrus Oils) Chlorpyrifos Traces Traces — 60 <0.1 <0.1 Organonitrogen Pesticides Pesticide Screening LC-MS/MS Pyraclostrobin Traces 0.11 — 200 <0.1

[0183] The data shows that with the present invention, fungicides, including strobilurine, especially pyraclostrobin are almost completely removed from the essential oil with the inventive process and the TFC nanofiltration membrane. Further reductions in agrochemical residues (pesticides), including organochlorine, especially p,p-Dicofol (acaricide) and organophosphorus, especially chlorpyrifos could be achieved.

[0184] Also naturally occurring impurities could be removed, preferably furocoumarine, such as oxypeucedanin, heraclenin, byak-angelicol, 8-Geranyloxypsoralen or Imperatorin were removed.

Example 2: Vanilla Extract—Fair Trade

[0185] 10-fold Vanilla extract (200 g) diluted with ethanol/water 95/5 was filtered at a temperature of 40° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. The permeate flux was 8.57 kg/h*m.sup.2. As a result, 312 g of the diluted oil was permeated. The solvent was removed under vacuum.

[0186] FIG. 4 shows the differences between Input (A), Permeate (P), and Retentate (R) samples of Vanilla extract before and after nanofiltration. The colour was reduced in the permeate for the Vanilla extract sample.

[0187] The results from the fragrance test showed the vanilla character to be more recognizable.

Example 3a: Patchouli Oil without Solvent

[0188] Single-fold Patchouli oil (250 g) was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 208 g of the oil was permeated. The permeate flux was 14.14 kg/h*m.sup.2.

[0189] FIG. 5 shows the differences between Input (A), Permeate (P), and Retentate (R) of Patchouli oil before and after nanofiltration.

Example 3b: Patchouli Oil with Solvent

[0190] Single-fold Patchouli oil (180 g) dissolved in 70 g MTBE was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 210 g of the oil was permeated. MTBE was removed under vacuum and with 50 ml ethanol casted out. The permeate flux was 53.57 kg/h*m.sup.2.

[0191] FIG. 6 shows the differences between Input (A), Permeate (P), and Retentate (R) of Patchouli oil with solvent before and after nanofiltration with solvent.

TABLE-US-00003 TABLE 3 CIELAB-Results of Patchouli oil after nanofiltration CIELAB-Results: L* A* B* Flowable Input 93.4 −11.8 49.2 Permeate 96.3 −11.8 35.0

[0192] FIG. 6 and table 3 demonstrate the reduction in colour of essential oil, especially Patchouli oil. Although the flux was three to four times faster, the selective reduction of coloured impurities was just as good as when the flux was slower without solvent. The faster flux lead to a faster production process and, therefore, energy costs can be saved.

[0193] According to Tab. 3 the difference ΔL* of permeate (after nanofiltration) to the flowable Input is 2.9 and ΔC* could be calculated as −13.7.

[0194] The results from the fragrance test showed the patchouli oil character is more recognizable and that the patchouli part has more volume and is stronger. The results from the fragrance test of the purified Patchouli oil without solvent was not as good as the result of the purified Patchouli oil with solvent.

TABLE-US-00004 TABLE 4 Iron content in Patchouli oil before and after nanofiltration Iron-Content in (ppm): Input Permeate Retentate 1. Measurement 267 40 916 2. Measurement 259 23 729

[0195] The results show that the iron content in permeate was reduced about 230 ppm (more than 84%) through the nanofiltration with a solvent.

Example 4: Orange Oil Brasil

[0196] Ten-fold Orange oil (202 g) was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 68 g of the oil was permeated. The permeate flux was slow 19.29 kg/h*m.sup.2.

[0197] FIG. 7 shows the differences between Input (A), Permeate (P), and Retentate (R) of Orange oil Basil before and after nanofiltration.

[0198] These results demonstrate the potential to reduce the colour of essential oils, including citrus oils, especially orange oil.

[0199] The results from the fragrance test showed that the zesty part of orange was kept and that the Bench Mark is different and less good.

[0200] Pesticide Screening: Orange Oil Brasil

TABLE-US-00005 TABLE 5 Concentration of select pesticides in Input, Retentate and Permeate solutions from Orange oil Brasil and maximum residue level of the pesticides Maximum Residue Compound Input Retentate Permeate Level mg/kg Organochlorine Pesticides and Pyrethroides Bifenthrin 0.71 1.4 0.43 20 Cyfluthrin — 0.29 — 4 Cypermethrin — 0.26 — 400 Organophosphorus Pesticides (Citrus Oils) Chlorpyrifos 2.6  3.9 2.1  60 Methidathion — 0.12 — 4 Organonitrogen Pesticides (ON/HT) Propargite 1.8  4.1 1.0  2 Pesticide Screening LC-MS/MS Azoxystrobin 0.17 0.3 — 3000 Diflubenzuron 0.19 0.47 0.11 200 Phosmet 0.14 — 0.11 100 Pyraclostrobin 0.35 0.74 0.22 400 Trifloxystrobin 0.68 1.4 0.43 100

[0201] The data show that with the present invention fungicides, including strobilurine, especially azoxystrobin, pyraclostrobin and trifloxystrobin are almost completely removed from the oil with the inventive process. Further reductions in agrochemical residue (pesticides), including organophosphorus pesticides, especially chlorpyrifos and methidathion, and organonitrogen pesticides, especially propargite, could be achieved. Furthermore, insecticides (pyrethroides), especially bifenthrin, cyfluthrin and cypermethrin, are reduced. This example demonstrates the removal of agrochemical residues from essential oils.

Example 5: Benzoin Siam Resin

[0202] Benzoin siam oil (50 g) diluted in 200 g ethanol was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 58 g of the oil was permeated. The permeate flux was 21.43 kg/h*m.sup.2.

[0203] These results demonstrate the reduction in colour of the essential oil.

[0204] The results from the fragrance test showed that the Benzoin siam character was stronger and more floral.

Example 6: Peru Balsam

[0205] Peru balsam oil (200 g) was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 90 g of the oil was permeated. The permeate flux was 8.57 kg/h*m.sup.2.

[0206] FIG. 8 shows the differences between Input (A), Permeate (P) and Retentate (R) of Peru Balsam (Example 6) before and after nanofiltration

[0207] These results demonstrate the reduction in colour of the essential oil.

[0208] The results from the fragrance test showed that the character was more cinnamic, spicy and more sweet.

Example 7: Crude Blood Orange Oil Italian

[0209] Blood orange oil (20 g) was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 16 g of the oil was permeated.

[0210] The permeate flux was 64.29 kg/h*m.sup.2.

[0211] Naturally occurring impurities could be removed, in particular furocoumarines, such as oxypeucedanin, heraclenin, byak-angelicol, 8-geranyloxypsoralen or imperatorin were removed.

[0212] FIG. 9 shows the differences between Input (A), Permeate (P) and Retentate (R) of blood orange oil before and after nanofiltration

[0213] These results demonstrate the reduction in colour of the essential oil.

[0214] The results from the fragrance test showed that the blood orange character was weaker than before nanofiltration. The retentate showed stronger blood orange character than permeate and flowable input.

Example 8: Tangerine Oil

[0215] Tangerine oil (263 g) was filtered at a temperature of 30° C. and filtration pressure of 30 bar with a spiral or flat sheet PDMS-GMT-NC-1 membrane with a nominal molecular weight cut-off of approximately 350 g/mol. As a result, 51 g of the oil was permeated. The permeate flux was 64.29 kg/h*m.sup.2.

[0216] FIG. 10 shows the differences between Input (A), Permeate (P) and Retentate (R) of Tangerine oil before and after nanofiltration.

[0217] These results demonstrate the reduction in colour of the essential oils, including citrus oils, especially tangerine oil.

TABLE-US-00006 TABLE 6 CIELAB-Results of Tangarine Oil before and after nanofiltration CIELAB-Results: L* A* B* Flowable Input 94.7 −10.9 102.2 Permeate 100.7 −5.1 11.0

[0218] According to Tab. 6 the difference of ΔL* of the permeate after nanofiltration to the flowable input is 6.0 and ΔC* could be calculated as −90.7. There was a considerable reduction in colour.

TABLE-US-00007 TABLE 7 CIELAB-Results of Tangarine Oil before and after nanofiltration as well as the permeate after 48 h exposure to UV-light CIELAB-Results: L* A* B* Flowable Input 98.7 −8.1 27.5 Permeate 100.0 −0.4 1.3 Permeate after 48 h 99.9 −0.4 0.8 exposer to UV- light

[0219] According to Tab. 7 the difference of ΔL* of the permeate after nanofiltration to the flowable input is 1.3 and ΔC* could be calculated as −27.3.

[0220] The difference of ΔL* of permeate after nanofiltration to the permeate after nanofiltration+UV-light is −0.1 and ΔC* could be calculated as 0.47. Evidently, the lightness only dropped slightly as compared to the freshly filtered permeate. In addition, the chromaticity was stable after 48 h UV-light and did not re-colour.

[0221] The results from the fragrance test showed that the character was less terpenic, more natural and dimethylanthranilate is less present. The GC/MS-analysis showed that the amount of decanal was reduced.

[0222] The results from the peroxide value test showed that the tangerine peroxide value dropped from 7.72 mEq of O.sub.2/kg (before TFC nanofiltration) to 5.37 mEq of O.sub.2/kg after TFC nanofiltration. These results demonstrate the potential to reduce the rancidity of essential oils, including citrus oils, especially tangerine oil.

[0223] The dry residue is determined according to the European pharmacopoeia by the evaporation of the solvents of a weighing scale. A defined measure or volume (2.0 g) or (2.0 ml) of the extract are placed in the weighing scale and 3 h in a drying cabinet at 100 to 105° C. Then the residue is cooled down via phosphorus pentoxide or silica. The test result is defined in percent (m/m) or gram/liter.

[0224] The dry residue from Tangerine oil 2.2% (before TFC nanofiltration) to 0.47% after TFC nanofiltration showed that impurities were reduced by the inventive process.

Example 9: Lemon Oil Italian

[0225] Lemon oil was purified according to the method described for example 1 without solvent. The permeate flux was 71.14 kg/h*m.sup.2.

[0226] FIG. 11 shows the differences between Input (A), Permeate (P) and Retentate (R) of Lemon oil before and after nanofiltration.

[0227] These results demonstrate the reduction in colour of the essential oils, including citrus oils, especially lemon oil.

[0228] Also naturally occurring impurities could be removed, preferably furocoumarine, such as oxypeucedanin, 8-geranyloxypsoralen or bergamottin were removed.

[0229] The results from the fragrance test showed that the character was cold smoke, cigarette, and coffee.

Example 10: Clove Leaves Oil

[0230] Clove leaves oil was purified according to the method described for example 1 without solvent. The permeate flux was 12.64 kg/h*m.sup.2.

[0231] FIG. 12 shows the differences between Input (A), Permeate (P) and Retentate (R) of Clove leaves oil before and after nanofiltration.

[0232] These results demonstrate the reduction in colour of the clove leaves oil.

[0233] The results from the fragrance test showed that the character of clove leaves was more recognizable after nanofiltration than before.

Example 11: Ylang Ylang Oil

[0234] Ylang Ylang oil was purified according to the method described for example 1 without solvent. The permeate flux was 24.00 kg/h*m2.

[0235] FIG. 13 shows the differences between Input (A), Permeate (P) and Retentate (R) of Ylang Ylang oil before and after nanofiltration.

[0236] These results demonstrate the reduction in colour of the Ylang Ylang oil.

[0237] The results from the fragrance test showed that the character was more floral and exotic, more salicylate. The difference before and after nanofiltration is obvious.

TABLE-US-00008 TABLE 8 Overview of the calculated CIELAB values ΔL* and ΔC*: Example Compound L* ΔL* C* ΔC* 1 Mandarine Oil 90.9 2.2 133.9 −5.0 (flowable input) 1 Mandarine Oil (permeate) 93.1 128.9 3a Pachtouli Oil (flowable input) 93.4 2.9 50.6 −13.7 Pachtouli Oil (permeate) 96.3 36.9 8a Tangerine Oil (flowable input) 94.7 6.0 102.8 −90.7 8a Tangerine Oil (permeate) 100.7 12.1 8b Tangerine Oil (flowable input) 98.7 1.3 28.7 −27.3 Tangerine Oil (permeate) 100.0 1.4 8b Tangerine Oil (after 99.9 −0.1 0.9 0.47 filtration + 48 h UV-light)

[0238] Table 8 shows the difference of ΔL* and ΔC* of the permeate after nanofiltration to the flowable input. In each example there was an evident increase in the lightness and decrease in the colourfulness after filtration, which demonstrates the efficacy of the purification method. For the UV-stability test of the permeate there was no significant change in the lightness and colourfulness, which means there was no re-colouration.

TABLE-US-00009 TABLE 9 Showing example 9, Mandarine Oil and Orange Oil purified according to the invention in a colourless Fine Fragrance. comparison filtered EO MANDARINE COLD TREATMENT 30 ORANGE BRAS COLD TREATMENT 125 BERGAMOT ECO ESSENCE (E0636) W/O 200 200 MYRCENE LINALYL ACETATE 25 25 LEMON OIL WINTER ITALIE 50 50 ORANGE OIL 125 MANDARIN OIL ITAL. 30 GRAPEFRUIT OIL 50 50 PEPPER OIL BLACK PERF. 10 10 RED BERRY EXTR. 5 5 HELIONAL 20 20 HYDROXY CITRONELLAL 10 10 GERANYL ACETATE PURE 5 5 BENZYL ACETATE 5 5 HEDIONE 100 100 HEDIONE HC/70 50 50 HEXYL CINNAMIC ALDEHYDE ALPHA 130 130 ISORALDEINE 95 15 15 ISO E SUPER 20 20 VETIVEROL 10 10 PATCHOULI OIL DECOL. MD 50 50 ELEMI OIL 10 10 AURELIONE ® 20 20 ETHYLENE BRASSYLATE 30 30 GALAXOLIDE 50% IN IPM 30 30 Parts in g: 1,000.000 1,000.000

[0239] The test preparation of a colourless fine fragrance demonstrated that the final composition including the filtered essential oil was visually colourless and transparent to the observer.

TABLE-US-00010 TABLE 10 Showing example 10, Mandarine Oil and Orange Oil purified according to the invention in a colourless Showergel. comparison filtered EO MANDARINE COLD TREATMENT 25 ORANGE BRAS COLD 300 TREATMENT ALDEHYDE C 8 0.5 0.5 HEXENYL ACETATE CIS-3 4 4 ORANGE OIL 300 MANDARIN OIL ITAL. 25 GRAPEFRUIT OIL 20 20 PEPPER OIL BLACK PERF. 5 5 FIR NEEDLE SIBERIA H 10 10 ISOAMYL ACETATE 0.5 0.5 DECALACTONE GAMMA 5.5 5.5 ETHYL METHYL BUTYRATE-2 1 1 ETHYL MALTOL 1 1 DAMASCONE ALPHA 3 3 DAMASCONE DELTA 0.5 0.5 IONONE BETA 35 35 CLOVE LEAF OIL DECOL. 5 5 AGRUMEX LC 70 70 KOAVONE 80 80 ORYCLON SPECIAL 120 120 CEDARWOOD OIL 60 60 CEDARWOOD OIL CHIN. 15 15 PATCHOULI OIL DECOL. MD 85 85 ISOLONGIFOLANON COEUR 85 85 AMBROCENIDE ® CRYST. 10% IPM 1.5 1.5 MACROLIDE ® SUPRA 5 5 DIPROPYLENE GLYCOL 62.5 62.5 Parts in g: 1,000.000 1,000.000

[0240] The test preparation of a colourless shower gel demonstrated that the final composition including the filtered essential oil was visually colourless and transparent to the observer.

TABLE-US-00011 TABLE 11 Data overview of the results of all purified essential oils: Ex CIE pl. Compound LAB ΔL* ΔC* Photo Impurities Flow rate Extra Test Fragrance 1 Mandarine oil Italian L, A, B 2.2 −5.0 FIG. 3 Pesticide 91.07 HPLC/MS-analysis odour of mandarin (Single-fold) Furocoumarine kg/h*m.sup.2 Fine oil stronger, with Fragrance + sparking transparent Shower-gel note, aldehyde part is stronger 2 Vanilla extract (10-f) FIG. 4 8.57 vanilla character kg/h*m.sup.2 more recognizable  3a Patchouli oil + L, A, B 2.9 −13.7 FIG. 5 Iron 14.14 3b is better than 3a MTBE kg/h*m.sup.2  3b Patchouli oil (−) Solvent FIG. 6 53.57 More patchouli kg/h*m.sup.2 character, more volume, stronger. 4 Orange oil Brasil FIG. 7 Pesticides 19.29 Fine Maintained zesty (10-fold) kg/h*m.sup.2 Fragrance + part of orange; Shower-gel Bench Mark is different and less good 5 Benzoe Siam oil + 21.43 More floral and MTBE kg/h*m.sup.2 more stronger 6 Peru Balsam FIG. 8 8.57 More cinnamic, kg/h*m.sup.2 spicy and more sweet. 7 Blood Orange Oil FIG. 9 Furocoumarine 64.29 Weaker than before kg/h*m.sup.2 nanofiltration. Retentate is more stronger than Permeate and Input  8a Tangerine Oil L, A, B 6.0 −90.7 FIG. 10 64.29 Peroxide Less terpenic, and  8b kg/h*m.sup.2 Dry Residue more natural, dimethylanthranilate is less present. L, A, B 1.3 −27.3 +48 h UV-light −0.1 0.47 9 Italian Lemon Oil FIG. 11 Furocoumarine 71.14 Cold smoke, kg/h*m.sup.2 cigarette, coffee, very interesting. 10  Clove leaves Oil FIG. 12 12.64 After nanofiltration is kg/h*m.sup.2 better than not filtrated 11  Ylang Ylang oil FIG. 13 24.00 More floral and kg/h*m.sup.2 exotic, more salicylate. The difference is obvious

[0241] The data in the tables and figures demonstrate the reduction in colour of the essential oils filtered, preferably citrus oils. Especially the difference of ΔL* and of ΔC* of permeate after nanofiltration as compared to the flowable input shows that the essential oils had less chromaticity. Further, the results from the HPLC/MS-analysis of input, permeate and retentate samples of Mandarine oil showed that the intensity of the impurities could be reduced in the permeate while they were higher in the retentate. The pesticide screening of the essential oils showed the removal of selected pesticides and fungicides through the TFC nanofiltration. Also the iron content could be reduced and the peroxide value dropped. The dry residue of the permeate was less than of the flowable input and high molecular weight components and waxes could be removed. These results show that the inventive process delivers purified essential oils which can be used as fine fragrances, fragrance ingredients or flavours, especially in colourless or transparent compositions. Also the results from fragrance tests showed that the odour of the essential oils was more prominent while bad odours could be removed and suppressed.