 |
|
|
|
|
Development of a celery oil and extract industry By Dr Linda Falzari and Professor Robert Menary
September 2005
RIRDC Publication No 05/133 RIRDC Project No UT-35A
Celery, Apium graveolens L. is a member of the family Apiaceae (formerly known as Umbelliferae) along with carrots, parsley, fennel etc. The species under investigation in this research is Apium graveolens var dulce or stem celery.
Celery seed oil is traded with increasing demand on world markets. This oil is produced by steam distillation of the seeds and is used primarily for flavouring of foods. Trace amounts are used in fine perfumes. Both celery seed oil and celery seed extract have FEMA GRAS registration enabling trade as flavour products within the USA. Trade is established in international markets.
The aim of this project was to identify a celery variety, for use in local production of steam-distilled herb oil, for the flavour and fragrance market. Initially the product is aimed at the domestic market but later also for export. In addition, the same or a further variety could provide seed oil for use as a therapeutic good.
The phthalide components and ?-selinene are the critical components imparting both the typical celery flavour and fragrance and the therapeutic value. An initial literature review showed that although many researchers have analysed the composition of celery products, there is still considerable confusion in the nomenclature and description of the phthalide compounds.
Study of two possible gas chromatography columns, HP1 and HP-INNOWax determined that the most suitable was the non-polar HP1 column. Two phthalides not frequently reported in celery oil were detected. The first of these appears to be 3-butyl-5,6-dihydro-4H-isobenzofuran-1-one. The second was not fully characterised, but has a molecular weight of 204 and the molecular formula C12H12O3.
Data that were used to develop laboratory protocols for testing celery field samples are presented in the report. The nature of the duration of distillation curve is such that there is an initial phase of rapid release of oil components followed by an asymptotic release that continues for many hours. For celery herb, the initial phase is concluded within approximately one hour from breakthrough, though there are statistical differences between the increments to three hours. For celery seed, the nature of the distillation curve is influenced by grinding the seed. Grinding the seed greatly reduced the lag phase at the beginning of distillation but allowed volatilisation of the more volatile components while the charge was being packed into the distillation vats. This reduced the yield from ground seed by half and this preparation method was not recommended.
Change in oil composition with duration of distillation was also considered. The more volatile components distil over more rapidly and appear at higher concentrations in the early increments. For celery, the limonene concentration decreases with duration of distillation, while the concentration of - selinene and the phthalides increases.
Based upon both the oil yield and composition data, arbitrary cut-off points for regular examination of celery field samples in the laboratory were set at 2.5 hours for herb and 6 hours for seed. Celery herb samples can be stored frozen at –18°C prior to distillation, without adverse affects on either oil yield or quality. If longer durations of distillation of celery seed are required, these can be conducted with a mid-distillation break, without substantial effect on the final oil yield and composition.
The criterion of primary importance in the selection of varieties for flavour and fragrance products is organoleptic assessment. The researchers made preliminary selections for oils that are reminiscent of fresh, crushed celery, but the final selection must be undertaken by the end-user. Two varieties, Salad Green and Tendercrisp, were selected as being the most promising of the six varieties tested. These were chosen for further analysis in a semi-commercial trial.
xi Both the absolute concentrations of the key aroma components and their relative proportions are potentially important in determining organoleptic character. Preliminary comparisons between the organoleptic assessments and the oil composition data suggest that the ratio of ?-selinene to n-butyl phthalide may be important. Although sedanolide and sedanenolide are sensorially active compounds, they do not have particularly low odour thresholds. Insufficient data were available in this research project for a full assessment of this observation and it is unlikely that the ratio of two components alone will determine the acceptability of a fragrance or flavour product. However, this ratio may prove a useful reference point in later studies. The component n-butyl phthalide does not appear as part of the ISO standard but the concentration of this component should be included in quality assurance assessments. The absolute value of sedanenolide and the ratio of sedanolide to sedanenolide also appear to have a bearing on organoleptic quality.
Organoleptic properties are not relevant to selection of a variety for the therapeutic market, since the oil will be encapsulated prior to consumption. The criteria of importance here are oil yield per unit area and the concentration of physiologically active components i.e. ?-selinene and the phthalides, particularly n-butyl phthalide.
Of the varieties studied, Salad Green was ultimately selected as the most promising variety for both the flavour and fragrance markets and as a therapeutic good.
A key point of interest to potential growers of celery, as an essential oil crop, is that such production is six months out of season with celery for the fresh vegetable market. This can pose some difficulties in the procurement of seed and potential growers for essential oil need to be aware that seed may not be readily available in autumn.
The physio-chemical properties of herb oils from plant material harvested in mid-February were found to fit reasonably well within the ISO standards for seed oil, although the ?-selinene concentrations tended to be low. The physical properties of the two commercially distilled oils were inside the range specified by the ISO standard.
For flavour and fragrance oils, the optimum harvest time will be determined by the development of the appropriate organoleptic properties oil through the season. This will be determined by the concentration of ?-selinene and the phthalide components. High levels of these components are also required in the production of celery as a therapeutic good. Serial harvest indicated that there is a marked change in oil composition as the plants begin to flower, after which, the oil composition becomes relatively stable. When the crop is harvested, even if it has mature seed, the seed should not be cleaned from the stem material before distillation. Analysis of the waste material from the variety trial indicated that the stem material contained high levels of phthalides, particularly sedanenolide.
Removing this material may reduce distillation costs by decreasing the volume of material to be distilled but this must be weighed against reduced oil quality through lost phthalides.
Besides oil quality, oil yield is also important. Harvest must take place after the commencement of flowering to ensure sufficient yield for economic viability. Oil yields of between 30 and 60kg/ha are predicted when the herb is harvested from mid-to late February.
The oil composition resulting from the commercial distillation was very similar to that predicted by the 2.5 hour laboratory distillation of the herb collected from the commercial vat. Although there may be changes in yield not detected by a 2.5 hour distillation of serial harvest samples, the oil composition is a valid assessment tool.
The duration of commercial distillation will depend upon economic comparison of the cost of distillation per unit time against the value of the oil obtained. This must take into account the fact that oil yield decreases with increasing duration of distillation, but the value of the oil obtained increases due to the increasing concentration of the valuable components in each further increment. The varieties in the semi-commercial trial were distilled to exhaustion, taking approximately three hours.
Consideration must be given to breaking the seed coat prior to distillation to allow better penetration of the steam. Laboratory work in this project showed that the risk of loss to evaporation was significant and it is not recommended that the seed be ground. Crushing or rolling the seed are two possible solutions but any method which exposes the oil to the atmosphere will be subject to the risk of the loss of volatiles. The decision to grind the seed or to distil whole seed must be made based upon economic considerations of the cost of grinding or other preparation, the cost of a longer distillation and the value of the oil lost due to volatilisation of exposed oil before the charge can be sealed into the vat.
It is recommended that under commercial conditions, that the operator separate the first 30 minutes of oil from the remainder. This will improve the quality of the remaining fraction by decreasing the concentration of the monoterpenes and increasing the concentration of the valuable components. It may be possible to rectify the first oil fraction by fractionation-distillation and blend the valuable components back into the bulk oil.
It is recommended that distillation commence with the inner skin of the separator empty and that the separator temperature be maintained as high as possible to allow fast, clean separation of the oil from the water. It was shown in this project that a low separator temperature allowed the oil to cling to the sides of the separator, reducing oil recovery rates.
Oleoresin, produced by solvent extraction of the seed, is another potential product from celery.
Guenther (1950) describes the oleoresin as having more of the body of celery than the oil, due to the presence of more of the less volatile components. This product would have potential as either a flavour or fragrance product but perhaps more particularly as a therapeutic good. The quality of the extract depends primarily on having a high concentration of ?-selinene and the phthalides relative to the other components.
The important criteria for selection of the optimum protocol for extraction of the oleoresin are the yield and quality of the extract. Grinding dramatically increased the efficiency of release of solvent extractable components from the seed. Although the polar solvent increased the yield of the extract, the quality of the hexane extract was higher. The costs associated with extract clean-up and solvent recovery of a polar solvent may outweigh the economic value of the additional yield.
Further aspects of solvent extraction are the number and duration of the washes. Many short duration washes appear to be more effective than a few long duration ones.
Solvent extraction of the marc of steam distilled seed is a promising method, proposed to recover the remaining valuable, but less volatile sesquiterpenes (primarily ?-selinene) and phthalides. The use of this method on laboratory distilled, dry seed produced 1.0-1.2% oil from a 6 hour distillation plus a further 0.3-0.4% of extract of valuable volatiles by solvent extraction.
Celery root extracts also have potential as a therapeutic good. There are substantial differences between the composition of extracts from above and below ground parts of celery and there are phthalides present in the roots of celery that are present only at trace levels in the tops of the plants.
The primary phthalide component in the roots is cis-neocnidilide whereas its isomer, sedanolide, is the dominant form in the tops. The other major phthalide present is the mw 204 phthalide. The value of the root extracts as therapeutic agents depends upon the as yet unknown activity of the extracted components, particularly cis-neocnidilide and mw 204 phthalide. Market assessment of this product is necessary to establish its value.
As with the seed extracts, hexane appears to be the best solvent for the production of root extracts.
The use of slightly more polar solvents such as isohexane, pentane or light petroleum should be considered. Roots should be harvested immediately after harvest of the tops for distillation. Care must be taken that volatiles are not lost during air-drying or wilting of the roots. Rough chopping and extraction of the fresh roots should result in yields of extract approximating 10-15mg/g of root extracted or 100ppm total phthalide.
The researchers are confident that this industry is economically viable. Yields of 50kg oil per hectare should be achievable under commercial conditions and at a farm-gate price of $50/kg for the oil; the gross margin for this crop makes it comparable to poppy and pyrethrum production. The inclusion of additional products, in the form of either marc or root extracts or both, should increase this gross margin further.
|