RECTIFICATION

What is the rectification of an essential oil?

In a 2025 opinion on spearmint oil (opinion 9785), the EFSA FEEDAP Panel reports that the oil “may” also be subject to rectification (fractional distillation), the fractions obtained being mixed with the essential oil to reach the specified profile. In a 2023 opinion (opinion 8178), the Panel reports that the eucalyptus essential oil it assessed “may” be rectified by distillation to obtain a 1,8-cineole content higher than 70 %.

What is a “folded” oil?

According to a 2021 EFSA FEEDAP opinion (opinion 6891), folded oils are obtained by fractional distillation of the expressed oil, to reduce the content of volatiles (terpenes) and consequently increase the concentration of any non-volatiles extracted by cold pressing.

In the study by Xu et al. (2017), Hamlin and Valencia oranges, Citrus × aurantium f. aurantium (Citrus sinensis (L.) Osbeck in the study), grown in Florida were each harvested twice (“twice per year” in the source), from established groves of the University of Florida Citrus Research and Education Center in Lake Alfred, Florida, early and late in each variety’s season. The oranges were harvested by hand for the 2015-2016 season: the Hamlin oranges on November 15, 2015, and January 14, 2016, the Valencia oranges on February 11, 2016, and March 31, 2016; approximately 30 trees per variety were harvested. After harvest, the fruit were stored in a cold room at 4 °C until processing. Before processing, the fruit were sanitized and surface-dried. Juice and an oil emulsion were extracted simultaneously with a commercial extractor, whose main characteristic is to extract the juice and the oil simultaneously; the extractor was cleaned before and after use. The oil emulsion was collected in 190-litre tanks and put into cold storage at 4 °C for approximately 40 hours. The crude emulsion was brought to room temperature over the course of 3 hours before being refined into cold pressed orange oil. A table-top centrifuge specifically for oil, equipped with a motor operating at 1000 ×g, was used to produce the cold pressed orange oil. Because of the large quantity of emulsion produced (~110 litres), only the top part of the emulsion, which contains the most oil, was fed very slowly by hand into the centrifuge. The oil was aliquoted into vials and stored at −20 °C until use.

According to EFSA opinion 6891, folded oils “may” be misleadingly described as “terpeneless”, a term used by the applicant and in the European Union Register of Feed Additives, since most folded oils still retain a large volatile fraction.

What did one study measure in deterpenated caraway and marjoram oils?

In the study by Ben Salha et al. (2019), fresh plant material of marjoram, Origanum majorana L. (twigs, leaves and flowers), was collected in July 2015 in Sidi Bouzid, Tunisia, and seeds of caraway, Carum carvi L., in July 2017 from cultivated plants in the region of Nabeul, Tunisia.

What did the crude caraway oil contain?

According to the authors, caraway essential oil (CEO) was dominated by oxygenated monoterpenes (77.43 %), followed by monoterpene hydrocarbons (22.48 %), without any trace of sesquiterpenes. Within the oxygenated monoterpenes, (−)-carvone (74.25 %) was the main compound; among the monoterpene hydrocarbons, D-limonene (22.22 %) was dominant.

What did the three fractions of each oil contain?

Nine compounds were identified in the first caraway fraction (CF1), in which D-limonene (77.35 %), (−)-carvone (20.18 %) and β-myrcene (1.54 %) were the major components. The second fraction (CF2) displayed the same chemotype as the caraway oil (CEO), with slight proportional differences. The third fraction (CF3) was composed of 13 constituents and was mainly dominated by (−)-carvone (97.15 %). In the marjoram oil (MEO), fifteen compounds were identified in the first fraction (MF1), whose major components were γ-terpinene (27.53 %), α-terpinene (21.15 %) and 4-(10)-thujene (11.17 %). In the second fraction (MF2), 17 constituents were identified, the main ones being terpinen-4-ol (54.39 %), cis-sabinene hydrate (10.00 %) and β-fenchyl alcohol (9.63 %). In the third fraction (MF3), 21 compounds were found, mainly terpinen-4-ol (48.60 %), cis-β-fenchyl alcohol (23.84 %) and cis-sabinene hydrate (5.79 %).

How do the totals of terpene hydrocarbons and oxygenated terpenes read in Table 1?

In Table 1 of the study (composition identified by GC-MS), total terpene hydrocarbons account for 22.48 % of the caraway oil, 79.06 % of CF1, 19.47 % of CF2 and 0.56 % of CF3. For the marjoram oil and its fractions MF1, MF2 and MF3, the same row reads 50.70 %, 87.92 %, 12.90 % and 8.98 %. Total oxygenated terpenes read, in the same order, 77.43 %, 21.62 %, 79.21 % and 98.56 % for caraway, and 47.36 %, 12.05 %, 85.06 % and 88.86 % for marjoram.

How does fractional distillation change spearmint oil?

In EFSA opinion 9785 (2025), the additive is extracted from the fresh or dried aerial parts of Mentha spicata L. at flowering by steam distillation; the volatile constituents are condensed and then separated from the aqueous phase by decantation. For this assessment, the applicant submitted analytical data on spearmint oil obtained either by steam distillation or by fractional distillation (rectification) and subsequent blending to achieve the intended specifications.

What do the analyses of six batches show?

According to the opinion, the applicant provided a full analysis by GC–MS of six batches of spearmint oil, three obtained by steam distillation and three by fractional distillation. The oil obtained by fractional distillation showed an enrichment in l-carvone ("+26%") and in other carvyl derivatives (trans-dihydrocarvyl acetate and dihydrocarveol) compared with the oil obtained by steam distillation. The opinion states that this increase in the % of l-carvone (76.5 % against 49.5 %) is balanced by the reduction in the % of other terpenes, for example d-limonene (4.5 % against 19.3 %), β-bourbonene (0.07 % against 2.02 %), myrcene (0.04 % against 2.52 %), 1,8-cineole (0.08 % against 1.72 %), β-caryophyllene (0.18 % against 1.39 %), β-pinene (0.05 % against 1.39 %) and α-pinene (0.05 % against 1.37 %). Table 1 of the opinion, which compares the three batches obtained by each process (GC-MS), gives, in % GC area (the areas of all detected peaks summing to 100 %), for carvone a mean of 49.46 (range 48.45–50.78) after steam distillation and of 76.47 (76.05–76.75) after fractional distillation, and for limonene 19.26 (16.70–20.78) against 4.52 (4.40–4.60). The FEEDAP Panel noted that fractional distillation and subsequent blending do not result in an increase in the content of substances of concern in spearmint oil.

Which specifications does the opinion apply?

The product specifications used by the applicant are based on the ISO standards that the opinion names as ISO 3033-1:2005 (“Oil of spearmint”, Part 1: Native type) and ISO 3033-3:2005 (“Oil of spearmint”, Part 3: Indian type, “redistilled oil”), adapted to reflect the concentrations of selected volatile components in spearmint oils obtained by either steam distillation or fractional distillation. Analysis of six batches of the additive, three per process, showed compliance with the specifications when analysed by GC-FID and expressed as a percentage of gas chromatographic peak area (% GC area). In Table 2 of the opinion (six batches, GC-FID, % GC area), l-carvone has a specification of 55–87, a mean of 68.5 and a range of 55.4–81.1, and d-limonene a specification of 2–25, a mean of 12.3 and a range of 4.19–22.1. The opinion’s conclusions apply to essential oils obtained by steam distillation from the aerial parts of Mentha spicata L., including oils produced by fractional distillation and subsequent blending of fractions to meet the intended specifications.

What does EFSA say about rectified eucalyptus oil?

In EFSA opinion 8178 (2023), the additive assessed is extracted by steam distillation from the leaves (either fresh or dried) and twigs of Eucalyptus globulus Labill. sourced from China and India. The volatile constituents are condensed and then separated from the aqueous phase by decantation, and the essential oil “may” be rectified by distillation to obtain a 1,8-cineole content higher than 70 %. The product specifications are based on the standard that the opinion names as ISO 770:2002(E), for crude or rectified oils of Eucalyptus globulus, adapted to reflect the concentrations of selected volatile components of the essential oil. In Table 2 of the opinion, seven batches of the oil of Eucalyptus globulus Labill. analysed by GC-MS give a mean of 79.53 % of GC area for 1,8-cineole (range 72.5–86.9), 8.03 % for d-limonene (4.16–11.12) and 2.45 % for α-pinene (1.28–3.28). According to the opinion’s conclusions, eucalyptus oil from Eucalyptus globulus “may” be produced from plants of different geographical origins and by various processes, resulting in preparations with different composition and toxicological profiles. These conclusions therefore apply only to eucalyptus oil in which methyleugenol was not detected (limit of detection 0.005 %) and which is produced by steam distillation from the leaves and twigs of E. globulus and rectified by distillation to obtain a 1,8-cineole content higher than 70 %.

What does EFSA opinion 6891 report on folded orange oils?

In EFSA opinion 6891 (2021), the applicant provided data on three folded oils produced by the fractional distillation of an expressed oil. For each, the percentage remaining after drying at 100 (the unit is redacted in the published opinion) was determined on three representative batches. According to the opinion, it “can be assumed” that the volatiles were removed during drying and that the dry matter values obtained provide a measure of the non-volatile components of the folded oils.

What are folded oils A, B and C?

Folded oil A, described by the applicant as “orange oil terpeneless *4”, comes from the fractional distillation of an expressed oil; its mean dry matter content is 10.5 % (range 9.6–11.9 %); based on these values, an enrichment factor for the non-volatiles of about 4 is given (the subject of the sentence is redacted in the published opinion). Folded oil B, described by the applicant as “orange oil terpeneless *10”, comes from the same process; its mean dry matter content is 20.9 % (range 20.2–21.5 %); based on these values, an enrichment factor for the non-volatiles of about 10 is given (the subject of the sentence is redacted in the published opinion).

In five batches of each type, the relative density at 20 °C ranged from 856 to 857 kg/m3 (specification: 852–862 kg/m3) for folded oil A and from 872 to 876 kg/m3 (specification: 867–889 kg/m3) for folded oil B. The refractive index at 20 °C ranged from 1.475 to 1.477 (specification: 1.473–1.479) for A and from 1.479 to 1.480 (specification: 1.475–1.485) for B.

What do their volatile fractions contain?

The volatile fraction of folded oil A is specified to contain d-limonene (89–96 %, the phytochemical marker), decanal (0.5–2.0 %), linalool (0.7–1.7 %), myrcene (0.1–1.0 %) and geranial (0.1–1.0 %). That of folded oil B is specified to contain d-limonene (79–89 %, the phytochemical marker), decanal (3.0–5.0 %), linalool (2.0–5.0 %), myrcene (0.1–1.0 %) and geranial (0.5–1.8 %); that of folded oil C, d-limonene (85–95 %, the phytochemical marker) and linalool (0.5–4.0 %). Table 9 of the opinion concerns the folded oils obtained from the fruit peel of Citrus × aurantium f. aurantium (Citrus sinensis (L.) Osbeck in the opinion), five batches of each type. It gives a mean of 87.2 % of GC area of d-limonene (range 83.3–89.0) for A and 68.4 % (64.6–71.6) for B, and of decanal 1.53 % (0.87–1.92) for A and 5.56 % (4.75–6.44) for B. A note to the table states that the GC-MS method underestimates d-limonene, the major component. Among the volatile components, α-sinensal (average 0.13 %; range 0.03–0.19 %), β-sinensal (0.16 %; 0.06–0.29 %), p-mentha-1,8(10)-dien-9-ol (0.015 %; 0.005–0.025 %) and perillaldehyde (0.18 %; 0.15–0.26 %) were detected and quantified in all batches of folded oil A. The corresponding concentrations in folded oil B were 0.38 % (0.24–0.58 %), 0.53 % (0.18–0.90 %), 0.028 % (0.015–0.037 %) and 0.44 % (0.39–0.52 %); for folded oil C, the passage is redacted in the published opinion.

How do folded oils compare with the expressed oil?

The FEEDAP Panel notes a high qualitative and quantitative similarity between the composition of the volatile fraction in expressed orange oil and that of the volatile fraction in the folded oils. The same 55 compounds that account for 99 % of expressed orange oil (as % GC area) also account for 97.9 % (96.5–99 %), 92.6 % (91.3–94.0 %) and 98.8 % (98.7–98.9 %) of the volatiles in folded oils A, B and C respectively. From the quantitative point of view, the main difference among the volatiles in expressed orange oil and in the folded oils is due to the different % of d-limonene. The Panel notes that these percentages refer only to the volatile part as 100 %; related to the whole additive as 100 %, these figures “may be reduced of about 12% to 21%” to account for the presence of non-volatiles in the folded oils. The Panel also notes that the percentages of non-volatile components in the residual fractions of expressed orange oil under assessment “will by far exceed” the levels reported in the literature for citrus oils which did not undergo a concentrating process.

Which furocoumarins were sought?

Furocoumarins (phellopterin, 8-geranoxypsoralen, psoralen, 8-methoxypsoralen or xanthotoxin, dimethoxypsoralen or isopimpellin, imperatorin, oxypeucedanin and bergamottin or 5-geranoxypsoralen) and p-synephrine were below the limit of detection of the HPLC method in ten batches (folded oils A and B, five batches each). Fourteen furocoumarins, including also trioxalene, herniarine, citroptene, toncarine, epoxybergamotine and byakangelicol, were below the limit of quantification of 1 mg/kg in three additional batches of “orange oil terpeneless folded”.

What are the distilled fractions of lemon oil?

According to EFSA opinion 6548 (2021), in addition to the direct use of cold-pressed lemon oil, the applicant also distils the expressed oil to obtain two further preparations: a volatile fraction collected after condensation and a residual fraction. The FEEDAP Panel considers that both fall into the definition “lemon oil distilled”: the residual fraction, that is the residues of the distillation of expressed lemon oil, and the volatile fraction, that is the distillate of that distillation.

What is the residual fraction?

The opinion describes the residual fraction as a greenish yellow, clear, mobile liquid with a characteristic odour; in five batches of the additive, its refractive index ranged between 1.4774 and 1.4788 (specification: 1.4770–1.4815). Its product specification is based on the concentrations, as % of GC area, of its main volatile components: d-limonene (51–63 %, the phytochemical marker), γ-terpinene (8–17 %), geranial (6–12 %), neral (5–9 %), pin-2(10)-ene (0.3–5.5 %) and β-bisabolene (0.3–4.0 %). According to the opinion, HPLC analysis of the 10 batches of the residual fraction of expressed lemon oil showed detectable concentrations of all the substances except byakangelicin and 5-geranyloxy-7-methoxycoumarin, and phellopterin was detected in one batch only. Table 11 of the opinion gives the furocoumarins and methoxycoumarins in the residual fraction of expressed lemon oil obtained from the fruit peels of Citrus × limon (L.) Osbeck (Citrus limon (L.) Osbeck in the opinion), based on the analysis of five batches. The text of the opinion thus speaks of 10 batches, and the caption of Table 11 of five. The table gives, as averages in mg/kg, 71.1 of 5-methoxypsoralen (bergapten; range 31.9–144), 2,825 of oxypeucedanin (range 2,061–4,183), 30.2 of 8-methoxypsoralen (xanthotoxin; range 17–56.1), 1,213 of imperatorin (range 928–1,489) and 690 of citropten (range 220–970; by GC-MS, according to the table’s note).

What is the distilled fraction?

The product specification of the distilled fraction is based on the concentrations, as % of GC area, of its main components: d-limonene (66–78 %, the phytochemical marker), pin-2(10)-ene (5–20 %), α-terpinene (1.5–9.5 %), pin-2(3)-ene (0.5–3.0 %) and “and or” sabinene (0.5–3.0 %). HPLC analysis of five batches of the distilled fraction of expressed lemon oil showed that furocoumarins and methoxycoumarins are below the corresponding limit of quantification.

Under which names do flavour references list these oils?

According to the opinion, C. limon and its extracts are listed in Fenaroli’s Handbook of Flavour Ingredients (Burdock, 2009) and by FEMA, under the reference numbers 2625 (lemon oil obtained by cold expression of the peel, with or without previous separation of the pulp and the peel) and 2626 (“lemon oil terpeneless”). The same opinion lists Citrus × aurantiifolia (Christm.) Swingle (C. aurantiifolia in the opinion) and its extracts in the same handbook and by FEMA under the numbers 2631 (“lime oil distilled”, prepared from the sour juice obtained by expression of the whole fruit) and 2632 (“lime oil terpeneless”).

Which hazards do the EFSA opinions document for these preparations?

In opinion 6891, the applicant produced a safety data sheet for folded orange oils A and B in which hazards for users have been identified; the opinion states that the additives assessed “should be considered” as irritants to skin and eyes, and as skin and respiratory sensitisers. In opinion 6548, since expressed lemon oil and its residual fraction contain furocoumarins, the opinion states that they “may” cause phototoxicity.

References

• Ben Salha G., Herrera Díaz R., Lengliz O., Abderrabba M., Labidi J., « Effect of the Chemical Composition of Free-Terpene Hydrocarbons Essential Oils on Antifungal Activity », Molecules 2019;24(19):3532, doi:10.3390/molecules24193532 — https://www.ebi.ac.uk/europepmc/webservices/rest/PMC6804076/fullTextXML
• Torres-Alvarez C., Castillo S., Sánchez-García E., Aguilera González C., Galindo-Rodríguez S.A. et al., « Inclusion Complexes of Concentrated Orange Oils and β-Cyclodextrin: Physicochemical and Biological Characterizations », Molecules 2020;25(21):5109, doi:10.3390/molecules25215109 — https://www.ebi.ac.uk/europepmc/webservices/rest/PMC7662335/fullTextXML
• Xu B.M., Baker G.L., Sarnoski P.J., Goodrich-Schneider R.M., « A Comparison of the Volatile Components of Cold Pressed Hamlin and Valencia (Citrus sinensis (L.) Osbeck) Orange Oils Affected by Huanglongbing », Journal of Food Quality 2017;2017:6793986, doi:10.1155/2017/6793986 — https://web.archive.org/web/20210613141700id_/https://www.hindawi.com/journals/jfq/2017/6793986/
• EFSA Panel on Additives and Products or Substances used in Animal Feed (FEEDAP), « Safety and efficacy of a feed additive consisting of an essential oil derived from the aerial parts of Mentha spicata L. (spearmint oil) for use in all animal species (FEFANA asbl) », EFSA Journal 2025;23(12):e9785, doi:10.2903/j.efsa.2025.9785 — https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12712543/fullTextXML
• EFSA Panel on Additives and Products or Substances used in Animal Feed (FEEDAP), « Safety and efficacy of a feed additive consisting of an essential oil derived from Eucalyptus globulus Labill. (eucalyptus oil) for all animal species (FEFANA asbl) », EFSA Journal 2023;21(7):e08178, doi:10.2903/j.efsa.2023.8178 — https://www.ebi.ac.uk/europepmc/webservices/rest/PMC10375361/fullTextXML
• EFSA Panel on Additives and Products or Substances used in Animal Feed (FEEDAP), « Safety and efficacy of feed additives consisting of expressed sweet orange peel oil and its fractions from Citrus sinensis (L.) Osbeck for use in all animal species (FEFANA asbl) », EFSA Journal 2021;19(11):e06891, doi:10.2903/j.efsa.2021.6891 — https://www.ebi.ac.uk/europepmc/webservices/rest/PMC8603005/fullTextXML
• EFSA Panel on Additives and Products or Substances used in Animal Feed (FEEDAP), « Safety and efficacy of feed additives consisting of expressed lemon oil and its fractions from Citrus limon (L.) Osbeck and of lime oil from Citrus aurantiifolia (Christm.) Swingle for use in all animal species (FEFANA asbl) », EFSA Journal 2021;19(4):e06548, doi:10.2903/j.efsa.2021.6548 — https://www.ebi.ac.uk/europepmc/webservices/rest/PMC8085978/fullTextXML
• Govaerts R. (ed.), World Checklist of Vascular Plants (WCVP), Royal Botanic Gardens, Kew — https://sftp.kew.org/pub/data-repositories/WCVP/wcvp.zip