LIME OIL AS A FACTOR IN DETERIORATION

Transcription

1 258 FLORIDA STATE HORTICULTURAL SOCIETY, 1953 and oxalate-soluble fractions are considered there is an increase in degree of gelation with an increase in these two fractions. If the data in Table 2 were arranged in or der of increasing values for water-soluble plus oxalate-soluble pectin, it is apparent that sam ples having the same amounts of water-solu ble plus oxalate-soluble pectin had less gel stability ^ as the pectinesterase activity in creased. * A high pectinesterase activity and a low water plus oxalate-soluble pectin frac tion was as effective in forming a gel as a high water plus oxalate-soluble fraction with a low pectinesterase activity. The results obtained indicate in a prelimin ary way that gelation of frozen orange concen trate under unfavorable storage conditions is related to the pectinesterase enzyme activity and the amounts of the various pectin fractions present, but these variations are small in com parison to normal variations between samples. Summary The data presented in this paper are the result of analyses of samples of orange con centrate obtained from citrus processing plants during the and canning sea sons. Although data are limited they indicate there is a loss of soluble pectin and divalent ions when the concentrate is stored at temper atures higher than F. (-17.8 C). The degree of gelation increased as the rate of pectin loss increased. Samples containing about the same amounts of water plus oxalatesoluble pectin decreased in stability with an increase in pectinesterase activity. There was some indication that a high pectinesterase ac tivity with low water plus oxalate-soluble pec tin was as effective in producing a gel as a high water plus oxalate-soluble fraction with low pectinesterase activity. Stability to gela tion is related to the pectinesterase activity; as the pectinesterase activity increased the ten dency towards gelation increased. LITERATURE CITED 1. Atkins. C. D., Rouse, A. H., Huggart, R. L., Moore, E. L., and Wenzel, F. W. Gelation and clarifi cation in concentrated citrus juices. III. Effect of heat treatment of Valencia orange and Duncan grape fruit juices prior to concentration. Food Technol., 7, 62 (1953). 2. Baker, G. L., and Goodwin, M. H. Effect of methyl ester content of pectinates upon gel character istics at different concentrations.of sugar. Delaware Agr. Expt. Sta., Bull. 246, 1 (1944). 3. Dietz J. H., and Rouse, A. H.' A rapid method re 11? $%$$** l dt ^ifd 4. Hills, C.' H. Mattern,* H. H., Nutting. G. C, and Speiser, Rudolph. Enzyme demethylated pectinates and their gelation. Food Technol., 3, 9 (1949) 5. Kertesz, Z. I. The pectic substances (1951) Interscience Publishing Co., New York, N Y 6. McCombs, E. A., and McCready. R. *M. Colormetric determination of pectin substances. Anal Chem., 24, 163 (1952). 7. McCready, R. M., and McComb, E. A. Extrac tion and determination of total pectic materials in fruits. Anal. Chem., 24, 1986 (1952). 8. McDonnell, L. R., Jansen, E. F., and Lineweaver, H. The properties of orange pectinesterase. Arch* Biochem., 6, 389 (1945). 9. Olsen, R. W.. Huggart, R. L., and Asbell, D. M. Gelation and clarification in concentrated citrus juices II. Effect of quantity of pulp in concentrates made from seedy varieties of fruit. Food Technol., 5, Roberts, J. A., and Gaddum, L. W. Composi tion of citrus fruit juices. Ind. Eng. Chem., 29, 574 (1937). 11. Rouse, A. H. Gel formation in frozen citrus concentrate thawed and stored at 4 F. Proc. Florida State Hort. Soc, 62, 17 (1949). 12. Wenzel, F. W., Moore, E. L., Rouse, A. H. and Atkins, C. D. Gelation and clarification in concen trated citrus juices. I. Introduction and present status. Food Technol., S, 454 (1951). LIME OIL AS A FACTOR IN DETERIORATION OF FLAVOR OF LIME JUICE George B. Macfie, Jr. Tropical Food Research Laboratories University of Miami Coral Gables The citrus industry as a whole has turned the bulk of its harvest from the fresh market to that of processed products. Since this change was made rather suddenly and re cently, information providing a basis for com plete quality control is not yet available. The citrus fruit that has most recently made its way into the processing industry is the lime. This late arrival was caused mostly by the failure to process an acceptable product under the older methods as well as the limited sup ply of raw material which prevented financing of large scale research. Lime juice is now reaching the market in an excellent, processed form. The industry is becoming a factor in the citrus trade of the country. Even now, however, processing tech niques are based on information provided for the handling of other fruits, as well as a crude system of trial and error. Therefore, research

2 MACFIE: LIME OIL IN LIME JUICE 259 with emphasis on quality control and more specific information on lime products seems to be particularly pertinent. It has been observed by workers in quality control that factors other than oil have an ef fect on flavor. More and more attention is be ing paid to the pulp content of the product. In frozen products, it has been considered that a larger percentage of oil would be desirable, providing a better bouquet or odor. There have also been ideas that perhaps the role of peel oil in the deterioration of flavor has been overemphasized. This idea could be checked if a pulp-free control could be obtained to be compared against other samples with varying pulp content. It has been shown by Kesterton and McDuff ( Season) that peel oil obtained by different methods of extraction vary in their chemical, as well as flavor characteristics. If peel oil were added to frozen products, and to other food stuffs, it might be assumed that these different types of oils would differ in their behavior, and exhibit a different degree of deterioration. This idea could be checked if a series of pulp-free, oil-free samples could be obtained to which could be added specific amounts of different types of oils. The types of oil available for study were: Distilled oil which emphasized the lower boiling and more volatile constituents; Fraser-Brace (rasped) oil (Kesterton and McDuff, Sea son) of which a waxy fraction that had win tered out upon storage was used to emphasize the high boiling and waxy fractions; and fresh, cold pressed lime oil which should include the two fractions above as well as exhibit charac teristics of an important commercial type of oil. These samples could then be tested for ac tual flavor breakdown against a sample possess ing the original flavor and odor. This sample would be a portion of the juice prepared in a specific manner for testing. Immediately after the preparation it would be placed in an ap propriate container and frozen. Application of the Wheeler test could then be verified on controlled samples. Statement of the Problem In the processing of lime juice into single strength, pasteurized, canned limeade a vari ety of off-flavors was encountered which were often, if not always, laid to the oil content of the juice. Since a large part of these offflavors may be described as "terpeney," and it has been shown by Evans and others (1949) that terpeney flavors and oxidation are associ ated in citrus oils, it was natural to conclude that oxidation of the oil content of the juice was the main, or sole offender in the lime products. When a terpeney flavor was formed in juice products shown to be oil free by the Clevenger still method of extraction (Production and Marketing Administration, 1949), it was as sumed that the waxy, nonvolatile constituents of the oil had not been driven off during aera tion and pasteurization and were responsible for the trouble. The main purpose of this problem was two fold. First, an attempt was made to apply the Wheeler test for peroxides to the lime juice as a method of determining the degree of oxida tion, and adding one more test to the quality control scheme of lime juice. Second, the ex periment was devised to show by means of taste tests that other constituents of lime juice may be responsible for off-flavors which may or may not be called terpeney. Preliminary Work The original raw material for this problem was a canned, heat processed, single strength limeade canned under the label FLORI TROPIC LIMEADE, by the Florida Tropical Fruit Growers Association, Goulds, Florida. The cans were 46 oz., tin lined and of ridged construction. The product at the time of ex amination was some three years old, which was too old for market value but it was hoped that due to this age it would constitute good raw material for examination. The first step in the examination was the attempt to extract any residual oil that might have been present and analyze for peroxide formation by use of the Wheeler test (Jacobs, 1951) that had proved successful in the an alysis of citrus oils. Before such an analysis could be carried out on the limeade, however, several adaptations to the test had to be worked out. In the regular test, the raw material was pure lime or orange oil which was dissolved in a solvent composed of chloroform and glacial acetic acid. The oil fraction of the limeade was the portion that was to be examined, but it was dispersed throughout the liquid. In ad dition, sugar had been added as a sweetening agent which made a more complicated dis persing medium. Since the lime juice was a

3 26 FLORIDA STATE HORTICULTURAL SOCIETY, 1953 hot pack product, and canned under a vacuum, it was assumed that the more volatile fractions of the oil had been driven off, leaving the higher boiling and waxy fractions dispersed in the limeade. All of the cans opened exhibited a veiy marked terpeney flavor, and since it is known that peroxide formation in citrus oils parallels the development of terpeney odors, it was assumed that dispersed in the limeade was a high boiling fraction of lime oil which had become oxidized and was presenting its characteristic taste odor. In studies by Markley and others (1937) it was shown that some 7.5% of cold pressed citrus oils was not volatile, and since this was a main part of the suspected oil content, some method other than steam dis tillation had to be employed for the extraction of the oil fraction for analysis. A system of differential solvents was decided upon, and a search for a suitable solvent that would give a satisfactory extraction and fit into the scheme of the Wheeler test was started. Many of the common organic solvents were tried, with the final selection of carbon tetrachloride. In the case of ether, peroxides were apt to be formed in the ether itself. ylene and toluene were prone to give poor separation. After agitation for mixing, a heavy white layer was formed which did not dissipate and give a clear separa tion of the layers. After extractions were started with the carbon tetrachloride, it was found that no similarity of results could be ob tained on extractions of cans from the same case, even though in normal canning procedure they should have been successive samples. To investigate the possibility of a split case, or samples from different mixing tanks, a number of cases were examined, all with the same re sult. Finally, it was decided that perhaps a corre lation could be found from the amount of etch ing of the inner lining of the can and the de gree of flavor deterioration. Upon opening the can, there was found in almost every case a yellow wax-like deposit on the bottom of the can which was insoluble in either carbon tetra chloride or chloroform. It was soluble in gla cial acetic acid, and in the acid-solvent mixture (Jacobs, 1951), and when checked for perox ides showed a definite positive reaction. The lining of all of these cans was very badly etched, some to the point of leakage. This meant that not only tin, but also iron had gone into solution and had combined with ele ments of the liquid to give an undetermined number and variety of compounds, one ol which may have been the wax-like deposit on the bottom of the cans. In view of these facts, in addition to the excessive age of the samples, it was decided to attempt to run the problem under more controlled and uniform conditions, and to try to verify the application of the Wheeler test for determination of peroxide con tent of lime juice, or oil dispersed in lime juice. With a change in the dispersing medium of the oil, a new set of tests was run to determine the best solvent for extraction from pure, un sweetened lime juice. Chloroform was the first of the new solvents to be checked, and was found to give as good separation as any and greater sensitivity. By running blank samples of water to which oxidized oil had been added, it was found that the test had a sensitivity of over 1 to 2,; that is, a trace of peroxide could be detected in oil that was diluted with water 1 to 2,. Since lime oil is virtually insoluble in water, it probably exists in a discrete oil phase. In the presence of fruit pulp, however, it may also be assumed that some of the oil would be absorbed on the par ticle surfaces and that a more incomplete ex traction could be accomplished than in the case of a dispersion in pure water. Sensitivitv tests of both media confirmed this supposition, but it was hoped that the test would still be of practical value. The only way to confirm this supposition was to utilize the test in the analy sis of controlled samples which could be checked against frozen samples preserved to retain their original, or fresh flavor. Experimental Procedure General In the performance of the Wheeler test a solvent for the oil is used composed of 4% chloroform and 6% glacial acetic acid. In the application of the test to juice, the chloroform constituent of the solvent was used for the extracting agent. From two to three hundred ml. of juice are placed in a separatory funnel. To this forty to fifty ml. of chloroform are added. After thorough mixing of the two liquids, the layers are allowed to separate and the chloroform drawn off. To provide uni formity of extraction a method of mixing was devised and used on samples. The funnel is inverted, and swirled five times. This is re peated four times. The chloroform is with drawn into a graduate and at least 9% of the amount placed in the funnel is obtained. To

4 MACFIE: LIME OIL IN LIME JUICE 261 this the correct amount of glacial acetic acid is added and the test performed as follows: One ml. of saturated potassium iodide solution is added. The flask is swirled for exactly one minute. One hundred ml. of distilled water are added. The mixture is then titrated with standard sodium thiosulfate solution of suitable normality. When the yellow color begins to fade, starch solution is added and the mixture is titrated to the end point. Peroxide is cal culated and reported as millimoles per liter of juice. To test for possible peroxide formation in juice to which no oil had been added, the test was first run on distilled water blanks omitting the chloroform element of the test. One or two drops of hydrogen peroxide were placed in the water, the glacial acetic acid added and the test performed as outlined above. Before any series of tests were performed, a set of blank samples was run with fresh solu tions of iodine and starch to check the possibil ity of any contamination. Nonpasteurized food products held at near room temperatures are subject to spoilage by various micro-organisms. To avoid this, a food preservative called "Sodium di-acetate," pro duced by the Union Carbide and Chemical Co., was added to all samples at the rate of.1 gm. per one hundred ml. of juice. To test its ef fect on the titration of the sample, sodium diacetate was added to both water and juice blanks to which oxidized oil had been added. There were no appreciable differences in the results of titrations between blanks with and without sodium di-acetate. Change of flavor due to the preservative was not detectable. In preparation of some of the samples, filtrations were made using distomaceous earth as a filter aid. Since a chemically pure filter aid was not available a technical grade earth called Hyflo-Super-Cel was used. Titration of juice samples before and after filtration showed negligible changes in acidity. After completion of the tests for peroxide formations, one set of samples was arranged in order of their preparation. All reagents for the test were added. Two water blanks, one with and one without Sodium di-acetate, were added to the series. To these samples, several drops of 3% hydrogen peroxide were added. It was noted that the reaction was influenced as much by time as by the amount of peroxide. The water blanks turned black immediately while samples with pulp not only took more peroxide but required a period of several min utes to turn. To test the possibility of a time factor in regular extraction tests, the flasks were allowed to remain untouched for several hours with no positive reaction produced. Preparation of the Sample One group of samples was prepared oil free, but pulp content retained. This group is des ignated Lot A-l, Lot A-2, etc. Preparation was as follows: Lot A-l Control. Limes were peeled lightly through epidermis and oil sac layer with a razor sharp, stainless steel knife. Then with another knife of the same keenness the ex posed albedo was cut away and the fruit sec tions lifted from the section membranes and placed in a small container. At intervals oi several minutes these sections were placed in a covered container filled with carbon dioxide. When a sufficient quantity of these sections had been collected, they were placed in clean cheesecloth and crushed in a fruit juicer. The resulting juice was collected in a carbon diox ide filled container. After all sections had been crushed, the juice was strained through a clean piece of cheesecloth, then filtered through a Buchner funnel using Hyflo-Super- Cel as a filter aid. The filtration was carried out twice. The resulting juice was a light yel low, clear liquid completely free of pulp as well as the pectins and glucosides known to be a part of the albedo and membranous parts of the fruit. A food preservative called "So dium di-acetate" was added at the rate of.1%. The juice was then divided into por tions. One portion was sealed in enamel lined six ounce cans (Moore and others, 1944), one of which was placed into a freezer immediately while the remainder were placed in storage at F. The other portion was placed in clean Erlenmeyer flasks of such size that an equal volume of head space was left above the liquid. The flasks were then closed with soft cotton plugs and placed in storage at F. Lot A-2. Limes were peeled as in Lot A-l. Instead of removing the fruit sections, how ever, the peeled fruit was placed in a fruit juicer and crushed. The juice was collected with the pulp remaining to approximate that found in commercially processed juice. Dur ing the squeezing the juice and fruit were pro tected with carbon dioxide. The juice was then stored as Lot A-l.

5 262 FLORIDA STATE HORTICULTURAL SOCIETY, 1953 Lot A-3. Limes were peeled as in Lot A-l. The peeled fruit was then placed in a Waring Blendor and pureed in order to include all membranous parts of the fruit. The juice was treated and stored as in Lot A-l. Another group of samples, designated Lot B-l, Lot B-2, etc. was prepared in which dif ferent types of lime oil were added to a clear juice base, in order to evaluate the effect of oil without the possibility of pulp interference except in the case of the last lot, wherein limes were crushed with the inclusion of all of the pulp and oil that came with the juice. Lot B-l Control. Same as Lot A-l. Lot B-2. Clear, filtered juice was prepared as in Lot A-l. To this, fresh distilled lime oil (peroxide content zero) was added at a rate of.1 ml. per one hundred ml. of juice. Samples were then treated and stored as in Lot A-l. Lot B-3. Clear filtered juice was prepared as in Lot A-l. To this, the waxy fraction (peroxide content zero) of Fraser-Brace (rasped) oil was added at the rate of.1 gm. per one hundred ml. of juice. Samples were then treated and stored as in Lot A-l. Lot B-4. Clear, filtered juice was prepared as in Lot A-l. Cold expressed lime oil (per oxide content zero) was added at rate of.1 ml. per one hundred ml. of juice. Samples were then treated and stored as in Lot A-l. Lot B-5. Limes were washed, dried, and cut in half. With no further treatment they were placed in a fruit juicer and crushed so as to obtain a normal amount of pulp and a good amount of oil that would come naturally with the juice in such a squeezing or crushing op eration. One sample of the juice was placed in the Clevenger still and the amount of re coverable oil determined as.41%. Samples of the oil-free juice base were also checked in the Clevenger still for oil content. Post-Storage Tests After storage for 19 days at F. the following tests were performed on the samples: Taste test. In preparation for the taste test, a quantity of simple sugar syrup of 12 Brix was prepared. This was added to a quantity of the juice from each of the canned samples at the rate of 7 syrup to 1 juice. The syrup was slightly chilled to present the lime flavors in the best manner possible. The panel was composed of Mr. J. Arthur Lewis, Mr. John S. Lynch, Miss Margaret Jean Mustard, Mrs. Jessica Dendy and Mr. George Macfie. Samples were arranged in order of their preparation with the frozen sample of each lot present for a check as to the original flavor. Terpeniness was scored numerically, with sub jective comments and comparisons made on general flavors. The results are shown as av erages in Table I. After the taste test, all storage samples were opened, extractions made for oxidized oil, and can inspection made. The Wheeler test for peroxide formation was performed on all samples, including a test directly on the juice itself. The results of the can inspection are shown in Table I. The peroxide formation is shown in Table II. Lot No. A-l A-2 A-3 B-l B-2 B-3 B-4 B-5 RESULTS Taste Test Terpeniness Table I Bitterness and general off flavor Can etching The be$ 3t, or least distorted flavor. The worst, or most distorted flavor. Lot No. A-l A-2 A-3 B-l B-2 B-3 B-4 B-5 Peroxide Formation Test Table II Canned, Stored! Samples stored Frozen Sample Samples in Flasks Trace Discussion It is definitely apparent that while pulp con tent is not associated with formation of a terpeney flavor or odor, it is a major contributor to the general off flavor of a stored product. In the case of sample A-3, the limeade made from this juice was almost undrinkable. On the other hand, it was equally apparent that the formation of the terpeney flavor and odor is associated with oil. Some terpeney flavor was associated with sample B-3, which con tained the waxy fraction of Fraser-Brace (rasped) oil. This agrees with the supposition that a deteriorated oil flavor may develop in a product otherwise oil free. The most pro nounced terpeney odor occurred in sample B-5, which contained the most oil. Since the oil was obtained in the juice by the pressure

6 MACFIE: TROPICAL FRUIT WINES 263 of the squeezer, it contained all of the frac tions. Which of these two conditions con tributed most to the flavor was not determined. Peroxide formation of the juice, or oil in juice, did not follow the same pattern as the deterioration of the flavor. It is noteworthy that the only trace of peroxide was found in the sample exhibiting the maximum of terpeney odors However, since it has been shown that the preservative did not initiate a break down of peroxide in an oil-in-water or an oilin-juice dispersion, it may be concluded that either the Wheeler test applied in this experi ment is not sensitive enough to detect peroxide at the level sufficient to produce a decided off flavor, or that during the process of break down something happened to the peroxide so that a sufficient quantity did not accumulate to coincide with the degree of deterioration. The Wheeler test was used by J. A. Lewis and associates in the study of oxidation of pure citrus oil. It was found that peroxide content could be detected by the test before it could be detected organoleptically. Fur thermore, since it has been shown that the test is sensitive enough to pick up measurable amounts of peroxides at dilutions five times greater than the dilutions of the various types of oil in samples, it would seem that the peroxide did not form or that the test was blocked in some way. A close correlation of can etching and ter peney flavors was found. In the case of tin lined cans a metallic flavor would normally be associated with the etched condition. Such a flavor was not noticeable in these samples; however, it could have been over-shadowed by other factors. Enamel lined cans are generally considered superior for citrus prod ucts. Summary A close correlation was found between peel oil content and development of terpeney flav ors. In general off flavors, however, it was found that pulp is a major factor. In both cases, the effect increased with the content of the causative agent. In all preliminary experiments, the Wheeler test proved satisfactory for detection and measurement of peroxide formation in oil-injuice dispersions. However, when this forma tion was initiated in the oil in the dispersed state, detection was not correlated with the formation or development of the terpeney flavor. Until further study can be undertaken, the test cannot be recommended as a quality control measure. LITERATURE CITED Blair, J. S.. Godar, Edith M., Masters, J. E. and Reister, D. W., Exploratory experiments to identify chemical reactions causing flavor deteriora tion during storage of canned orange juice. I. Incompatability of peel-oil constituents with the acid juice. Food Res. 17(3): Evans, Ralph L., Progress report on citrus oil research; stabilization of citrus oil (Kuder). Unpub lished report, Evans Research and Development Corp. New York. Guenther, Ernest S., California citrus oil. The American Perfumer, May, June, July and August The essential oils, III. New York, D. Van Nostrand Company, Inc. Jacobs, Morris B., The chemical analysis of foods and food products. New York, D. Van Nostrand Company, Inc., p Kesterton, J. W., and McDuff O. R., Sea son. Commercial production methods and properties of essential oils. Univ. of Fla. Bull Markley, K. S., Nelson, E. K., and Sherman, Mildred S., Some wax-like constituents from expressed oil from the peel of Florida grapefruit, Citrus Grandis. Jour, of Biol. Chem. 118(2). Moore, Edwin L., Wiederhold, Eunice, and Atkins, Donald C.f Changes occurring in orange and grapefruit juices during commercial processing and subsequent storage of the glass- and tin-packed prod ucts. American Food Manufacturer 23(9): Production and Marketing Administration, United States standards for grades of canned orange juice. CFR 7, Section 52, 488, Washington. WINES FROM TROPICAL FRUITS George B. Macfie, Jr. Tropical Food Research Laboratories University of Miami Coral Gables The problem of making wine from tropical fruits is not new, nor has it proved to be in surmountable in the past. Most tropical re gions of the world have succeeded in making some form of alcoholic drink from the mate rials available which is usually accepted in that particular area, but usually do not find a widespread acceptance throughout the re mainder of the world. Since the universally accepted raw material for wine manufacture is the grape, most of the technical efforts have been aimed at the improvement of this mate rial and the subsequent wine products. Florida is unusual in that it presents an op portunity for study of both tropical and sub tropical fruits. Indeed, a number of the tern-