SOME SPECIFIC SOURCES OF CONTAMINATION IN PROCESSING

MURDOCK AND BROKAW: FRUIT PREPARATION 231 SOME SPECIFIC SOURCES OF CONTAMINATION IN PROCESSING FROZEN CONCENTRATED ORANGE JUICE - 1. HANDLING AND PREPARING FRUIT FOR EXTRACTION D. I. MUBDOCK Staff Bacteriologist Charles H. Brokaw Chief, Quality Control Department Minute Maid Corporation Orlando Frozen orange concentrate and other citrus concentrates, because their ph values aver age from 3.4 to 4.0 (4, 10), limit the growth of organisms to those capable of tolerating this acid medium. Microorganisms known to grow in the single strength juices of these acid foods (not including lemon or lime juices) are lactic acid and acetic bacteria, yeast, and molds. Of this group, organisms belonging to the genus Lactobacillus are of prime concern to bacteriologists. It is the lactic acid bacteria that have been attributed to the production of abnormal flavors and odors in concentrate, one of which is described as being similar to those characterizing "but termilk" (5, 6, 8). On the other hand, acetic bacteria, yeast, and molds generally do not grow rapidly enough to build up large popu lations under conditions that normally prevail during concentration of the juice (13). Flash heating product, either as it enters the evaporators or early in the process of con centration, was adopted to retard gelation and separation of cloud and pulp, but it also has a beneficial effect in reducing bac terial contamination in evaporators (1, 11, 12, 13). However, lack of sufficient heat treatment to destroy all microorganisms in the product has necessitated new sanitary con cepts which are unique to this industry. To insure a minimum of micro bial activity in stages of processing prior to evaporators requires the maintenance of an efficient sani tation program. In our plants this consists of flushing juice room equipment, including extractors, finishers, and juice holding tanks, with chlorinated water on a routine basis. It also has involved the elimination of focal pockets of contamination such as dead ends, use of sanitary-designed equipment, etc. Of the few studies in literature on sanitary aspects of processing citrus concentrates, Beisel (2) and Brokaw (3) were most informative. These two authors described the manufacturing op erations and discussed sanitary practices em ployed. The purpose of this paper is to point out specific sources of contamination in (1) han dling fruit from bins to extractors, (2) juice after it leaves extractors until it enters the evaporators, and (3) concentrating and packaging. Procedure The fruit surface studies reported here in were made in accordance with the procedure described by D. I. Murdock, et al. (Florida State Horticultural Proceedings, No vember, 1953) (9). Standard plating tech nique was employed throughout the study. Orange serum agar was used as the plating medium, and all plates were counted after 48 hours of incubation at 30 C. (86 F.). Results Fruit Handling At most concentrate plants fruit is received by truck, dumped on to a conveyor system, graded for maturity and soundness, stored in bins, washed, regraded, and sanitized prior to entering the extractors. Each of these op erations has a direct bearing on the micro bial population of the extracted juice. Improperly washed fruit and insanitary fruit handling equipment are both contributing factors. Each fruit handling operation was investi gated to determine its effect on fruit surface contamination. The average results obtained from 14 test runs, which are presented in Figure 1, show the percent increase or de crease in fruit surface microflora after each sampling station. Fruit surface contamination from bins, for example, was based on sam ples collected before and after the bins and, from brush washer, from samples collected before and after this operation. It will be

232 FLORIDA STATE HORTICULTURAL SOCIETY, 1957 FIGURE 1. PERCENT INCREASE OR DECREASE IN FRtJIT SURFACE MICROFLORA WITHIN EACH FRUIT HANDLING OPERATION 150 116* Data - average of ik tests* 100 50 50 39* u 31* 25. &vg. cone. 100 * CI2 * FINAL CI2 BINS WASHER SPRAY ELEVATOR SPRAY noted that fruit is recontaminated in the bins and by the flight elevator. The brush washer employed in conjunction with a chlorinated water rinse (25 p.p.m. average concentration) was very effective in reducing fruit surface contamination. The chlorinated water rinse (25 p.p.m.) used after the final graders and before the extractors also reduced the total fruit surface microflora, but to a much lesser degree. Fruit Grading Type of fruit received by a frozen concen trate plant has a direct bearing on the quality of the finished product. A large percentage of drops, soft deteriorated spots, splits, etc., for example, may result in off-flavors ("stale" or "old fruit") in the final product. To mini mize the number of unsound fruits reaching thie extractors, fruit is graded (Figure 2) be fore entering the bins and again before the extractors. In the initial operation, unsound fruit received at the plant is sorted out; and fruit damaged by the bins removed by the final graders. Unfortunately, the efficiency of grading operations in most plants leaves a lot to be desired, and it is not uncommon for fruit not of concentrate quality to reach the extractors. Fruit grading is one operation where there has been little or no technological advances since the advent of the citrus industry. Grad ing tables are not standardized and are of various sizes. It has been found (7) that on a grading table 4 feet in width, normal fruit should move no faster than about 30 feet per minute and at least 2 graders should be located on each side of the table. Under these conditions, fruit should move over a table at

MURDOCK AND BROKAW: FRUIT PREPARATION 233 Fig. 2. Removal of Defective fruit is essential in maintaining product quality. a rate no greater than 120 boxes (90 pounds each) per grader per hour for satisfactory grading. If fruit is less than normal quality it may be necessary to reduce the ratio of fruit to graders by as much as 50%. Never theless, grading tables are frequently oper ated at speeds far in excess of this rate. Move ment of fruit at 100 to 120 feet per minute are not uncommon in some plants. The abil ity to grade fruit efficiently when it passes over roller conveyors at this speed is ques tionable. Another common practice, which is believed to occur not only in our plants but in the industry as a whole, is when an extra "hand" is needed for some trivial task he is usually obtained from the grading tables, thus creating a shortage of graders. In view of these facts, a study was made to show the effect of fruit surface contamina tion on microbial population of extracted juice. Defective fruits, represented by drops, splits, and deteriorated oranges (soft spots and par tially rotten), were sampled after the germicidal rinse and handled in the same manner as in previous tests. Both fruit surfaces and extracted juice from splits and deteriorated fruit were heavily contaminated with micro organisms, as indicated in data presented in Tables I and II. It is quite obvious from these results, if defective fruit is not removed from the line by the final graders it will readily "seed" juice extracted from sound oranges. For example, if 100 ml. of juice is obtained from 1 orange, and the juice is from a deteriorated fruit containing 9,200,000 mi croorganisms per ml. (based on data in Table II), the juice from the whole orange would be equivalent to 100 times this amount. Now, if juice from this one orange was mixed, by chance, with 10 gallons of fresh juice it would raise the bacterial load of the juice by ap proximately 24,000 microorganisms per ml.

234 FLORIDA STATE HORTICULTURAL SOCIETY, 1957 The foregoing studies were based on a laboratory scale to show contamination re sulting from various types of fruit selected. An attempt was then made to determine the effect of fruit quality on contamination of juice during a commercial operation. For 2 days every effort was made by the final grad ers to remove all defective fruit passing over the line. During this period the fruit was con sidered to be "sound" having no splits, bruises, or soft spots. The next phase of study was made over a 3-day period when a large Type of Fruit Sound Fruit Drop* Splits Deteriorated Fruit TABLE I. 'MICRQFLOBA ON THE SURFACES OF SOUND. AND DEFECTIVE FRUIT Total Microflora per Orange Test No. 1 64,000 113,700,000 3,250,000 Test No. 2 74,000 855,000 21,400,000 44,200,000 percentage of fruit entering the extractors contained splits, soft spots, etc., resulting from inability of the graders to remove all defective fruit. Data in Table III show contamination of juice collected from various sampling sta tions from the extractors to the cutback tank. It will be noted that juice extracted from sound fruit contained fewer microorganisms than it did during the period when juice was extracted from poor fruit. How does one check for grading efficiency? To the authors' knowledge, there is no infor mation in the literature covering this subject. The usual procedure is to make a casual in spection of a fruit line after the graders. The inspector may report that grading is either good or bad but, in so doing, he has no factual data to back up his statements. It would be very desirable if information could be pre sented to management showing an acceptable fruit grading process. With this in mind, sev eral methods were investigated to determine grading efficiency. One procedure which showed some promise was referred to as the "5-gallon bucket method." This consisted of filling a 5-gallon container, as rapidly as pos sible, with fruit from the conveyor belt after the final graders. The fruit was then examined individually and the number of defects (splits, soft spots, etc.) recorded as the percent of "grade-outs" per 5-gallon sample (approxi mately 50 oranges of average size). A study was made using this procedure over a 5-day period. The percentage of defective fruit in 9 samplings ranged from 2 to 10% with an average of 6%. It is realized this procedure is very crude but it might add "food for thought" for additional studies to be made along this line. Recontamination of Fruit Surfaces Recontamination of fruit surfaces occurs after sanitizing if subsequent fruit handling equipment is not kept in a sanitary condition. Data have been presented to show that ele vators are a source of contamination. Fruit surfaces may also become contaminated from slimy fruit sizers, and from conveyor belts, especially those in the juice room. In some plants belts that feed fruit to the extractors, travel over expanded metal which becomes impregnated with pieces of fruit that is al most impossible to remove during the clean ing process. In other cases, conveyor belts move over a series of rollers which frequently stick and then become coated with citrus solids resulting in the rollers contaminating the belt surfaces (Figure 3); as indicated when a sample of this material was plated and found to contain 15,700,000 microorgan isms per gram. TABLE H CONTAMINATION OF EXTRACTED JUICE FROM Type of Fruit Sound Fruif, Drops Splits Deteriorated Fruit SOUND AND DEFECTIVE FRUIT MICROORGANISMS PER ML. Test No. 1 29,000 2,800,000 9,200,000 Test No. 2 500 1,500 620,000 3,ltiO,000 Another source of fruit surface contamina tion is the short strips of canvas or conveyor belt material referred to as "flaps," which, are frequently used to provide a more even distribution of fruit along the fruit line, es pecially ahead of the brush washer and size grader. This material which readily becomes coated with slime is heavily seeded with mi croorganisms resulting in an increase of the

MURDOCK AND BROKAW: FRUIT PREPARATION 235 TABLE IE EFFECT OF FRUIT QUALITY ON CONTAMINATION OF CUTBACK JUICE SOURCE OF SAMPLE POOR No. of Samplings FRUIT (2) Average Org. per ml. SOUND - No. of Samplings FRUIT (2) Average Org. per ml. Juice After Extractors (1) 11 76,363, 6 7,950.Before Finisher 11 164, 363 6 9,583 After Finisher 11 173,181 4 8,825 Cutback Tank 9 150,222 6 15,550 (1) Composite sample of juice from Brown Extractors before finisher. (2) All data obtained over a single 5-day period; viz., "Poor Fruit" for 3 days and "Sound Fruit" for 2 days. 11 Poof Fruit" - Containing a high percentage of soft fruit with numerous splits, bruises and soft spots. nsound.fruit" - Fruit supply was firm with no splits, bruises or soft spots microbial load on any fruit surfaces it con tacts. A sample of slime from a canvas flap before a fruit sizer gave a total viable count of 368,000,000 microorganisms per gram when plated, indicating flaps are a potential source of fruit surface contamination. Flaps should be eliminated wherever possible, but if necessary, a non-porous material such as metal should be used. Recontamination of the fruit surfaces may also occur just before the fruit enters the ex tractors if they are not properly cleaned in their entirety. In the Brown machine this may result from an accumulation of citrus solids in fruit hopper and/or the fruit elevator which conveys individual fruit to the cutting knives; and in the F.M.C. extractor, from the vibrating fruit feeder and/or bar lift. Excess or overflow fruit from the extrac tors frequently is recontaminated before it is returned to the extractor line. In one plant a screw conveyor was installed to elevate the fruit back to the extractor fruit line. The conveyor proved difficult to clean, and also split a large number of fruit. Fruit surface studies showed it to be a serious source of microbial build-up, as the data in Table IV indicate. The overflow fruit, which TABLE IV RECONTAMINATION OF FRUTT SURFACES BY SCREW CONVEYOR USED FOR HANDLING OVERFLOW FRUIT FROM EXTRACTORS Date Feb.. 24 Feb. 25 Feb. 26 Feb. 27 (AM) Feb. 27 (PM) Screw Conveyor Entering Leaving 500,000 15,000,000 520,000 4,600,000 400,000 2,600,000 220,000 8,000,000 280,000 3,100,000 Results expressed as the number of microorganisms per orange.

236 FLORIDA STATE HORTICULTURAL SOCIETY, 1957 Fig. 3. Rollers that do not operate properly become a source of contamination. was recontaminated by the screw conveyor, was fed back into Line No. 2 along with the regular fruit. Line check samples showed juice obtained before finishers from Line No. 2 to be more heavily contaminated than juice sampled from Line No. 1, which contained no overflow fruit (Table V). In most concentrate plants overflow fruit is removed from the conveyor belt by means of a baffle plate which shunts it on to another belt. It was noted, in one operation, that fruit frequently becomes impinged on this plate, splits and lodges between it and the conveyor belt. Fruits, which may accumulate at this point to a depth of 5 or 6 inches, continually slough off and are then conveyed back to the line feeding the extractors. It was believed this condition was responsible for juice sampled from Line No. 2 having a higher plate count than Line No. 1, which received no overflow fruit. To verify this assumption, operating personnel in the juice room was given spe cific instructions to remove fruit from the baffle plate every 4 hours. Subsequent line checks showed a reduction in total viable tablev JUICE EXTRACTED FROM ORANGES ON LINE CONTA2KDft> FRUIT Feb. 3 Feb. 4 Feb. 5 Feb. 6 Feb. 7 RECONTAMINATED BY SCREW CONVEYOR LueNo. 1 LteeNo. 2 NO OVERFLOW FRUIT OVERFLOW PLUS REGULAR FRUIT 47,500 5,250 3,500 13,900 11,260 107,200 13^750 44,000 13,500 31,250 Results expressed is number of microorganisms per ml. Samples collected before ltaiabers/

MURDOCK AND BROKAW: FRUIT PREPARATION 237 count of extracted juice from Line No. 2, indicating fruit accumulating on the baffle plate to be source of contamination (Table VI). Summary A study was made of the total microflora found on surfaces of oranges from fruit bins to extractors to show sources of contamina tion. Several focal points of microbial build up were noted. It was found that the brush washer in conjunction with a chlorinated water rinse was very effective in reducing fruit surface contamination. Recontamination occurred in subsequent fruit handling if im proper sanitary precautions were not ad hered to. Defective fruit that was not removed in the final grading operation was found to be heavily contaminated with microorganisms, both externally and internally. The juice ex tracted from this type of fruit readily seeded juice extracted from sound oranges. Each unit operation of the fruit handling system must be carefully watched to prevent con tamination of the extracted juice. LITERATURE CITED 1. Atkins, C. D., and Rouse, A. H. Time-temperature relationships for heat inactivation of pectinesterase in citrus juice. Food Technol. 7, 489 (1953). 2. Bdisel, C. G. How frozen concentrate biochemists are working out the fruit bugs. Food Eng. 23, 82 (1951). 3. Brokaw, Charles H. The role of sanitation in quality control of frozen citrus concentrates. Food Technol. 6, 344 (1952). 4. Cameron, E. J., and Esty, J. R. Comments on the microbiology of spoilage in canned foods. Food Research. 5:549-557 (1940). 5. Hays, G. L. The isolation, cultivation, and identifi cation of organisms which have caused spoilage in frozen concentrated orange juice. Proc. Fla. State Hort. Soc. (1951). 6. Hays, G. L, and Reister, D. W. The control of "offodor" spoilage in frozen concentrated orange juice. Food Technol. 7, 383 (1952). 7. Minute Maid Corporation. Unpublished laboratory rec ords. 8. AAurdock, D. I., Troy, V. S., and Folinazzo, J. F. De velopment of off-flavor in 20 Brix orange concentrate inoculated with certain strains of lactobacilli and leuconostoc. Food Technol. 6, 127 (1952). 9. AAurdock, D. I., Folinazzo, J. F., and Brokaw, Charles H. Some observations of gum forming organisms found on fruit surfaces. Proc. Fla. State Hort. Soc, 278 (1953). 10. Pilcher, R. W. The canned food reference manual (3rd ed.). New York; American Can Co. (1947). 11. Rouse, A. H., and Atkins, C. D. Heat inactivation of pectinesterase in citrus juices. Food. Technol. 6, 291 (1952). 12. Stevens, J. W., Pritchett, D. E., Baier, W. E. Control of enzymatic flocculation of cloud in citrus juices. Food Technol. 4, 469 (1950). 13. Vaughn, Reese H., and AAurdock, D. I. Sanitary sig nificance of microorganisms in frozen citrus products. Amer. Journ. of Public Health. 46, July, 886 (1956). TABLE VI CONTAMINATION OF JUICE FROM TWO EXTRACTOR LINES SHOWING APPARENT EFFECT OF FRUIT BREAKDOWN ON BAFFLE PLATE USED ON LINE NO. 2 IN CONJUNCTION WITH HANDLING OVERFLOW FRUIT FROM EXTRACTORS Fruit Not Removed From Baffle Plate Fruit Removed From Baffle Plate Date Line No. 1 No Baffle Line No. 2 With Baffle Date Line No. I No Baffle Line No. 2 With Baffle 14 Not Operating 424, 000 18 7,000 15 15,000 80, 000 1.9 16 3,000 212, 000 20 2,000 17 4,000 45,000 23 67,000 6 Results expressed as number of microorganisms per ml. All samples collected before finisher in each line.