Supercritical Carbon Dioxide Pasteurization of Coconut
Water, a Sport Drink with a High Vitaminic and
Nutritional Content
Martina Cappelletti1, Giovanna Ferrentino1, *, Isabella Endrizzi2, Eugenio Aprea2, Emanuela Betta2, Maria Laura Corollaro2, Mathilde
Charles2, Flavia Gasperi2, Sara Spilimbergo1
1Department of Industrial Engineering, University of Trento, via Mesiano 77, 38123 Trento, Italy
2Edmund Mach Foundation, via E. Mach 1, 38010 S. Michele all’Adige, Trento, Italy
Abstract. Supercritical carbon dioxide was applied for the pasteuriza-tion of coconut water to guarantee its microbial stability and preserve its sensory and nutritional quality attributes. A multi-batch apparatus was used to test supercritical carbon dioxide feasibility to inactivate the natural microbial flora of coconut water as a function of pressure (80 and 120 bar), temperature (22, 30, 35, 40 and 45°C) and time (from 5 to 60 min). The study indicated that 120 bar, 40°C, 30 min were the optimal process conditions to induce 5 Log reductions of mesophilic microorganisms, lactic acid bacteria, yeasts and molds and 7 Log of total coliforms. A deep chemical/physical and sensory char-acterization was also performed to investigate the impact of the treat-ment on the quality traits of fresh coconut water. To complete the study, a thermally pasteurized (90°C for 1 minute) coconut water was also considered in order to compare the impact of the thermal and non-thermal processes.
Keywords: coconut water, supercritical carbon dioxide, pasteuriza-tion, natural microbial flora, quality, nutritional and sensory attributes.
1.
Introduction
The hydration and absorption of liquids with a high content of mineral salts, vitamins, sugars and nutritional substances are fundamental for the correct feed of all sportive people. Decades of scientific research clearly demonstrate the benefits derived from the assumption of drinks during and after a physical activity [1]. Coconut water is be-coming more and more popular within the athletes as “energy and healthy drink” rich in vitamin C, magnesium, calcium, potassium, vitamin B, arginine, alanine, lysine,
glutamic acid, enzymes with anti-inflammatory properties, minerals and antioxidants [2]. Coconut water is processed by heat treatment which is able to destroy the natural microbial flora occurred in the product and to prologue its shelf life for 2/3 months [2]. However, as the high process temperatures grossly alter the sensory qualities with changes in the product nutritional contents, different techniques of preservation like filtration, increasing the sugar and total solid contents, adjustment of pH, ultrasonic treatment, concentration by reverse osmosis, spray drying, addition of preservatives, etc. have been investigated either alone or in various combinations [2, 3, 4]. Among these techniques, the use of supercritical carbon dioxide (SC-CO2) seems to constitute one of the most promising alternatives. Since the 1980s, SC-CO2 has been increas-ingly investigated as a technique able to induce the inactivation of the natural micro-bial flora but also pathogens occurring in solid and liquid matrices [5, 6, 7, 8]. CO2 used in this process is relatively inert, inexpensive, nontoxic, nonflammable, recy-clable and readily available in high purity leaving no residues when removed after the process. Furthermore, it is considered a GRAS (Generally Recognized as Safe) sub-stance, which means it can be used for food products.
In this context, the present project aimed to investigate the possibility to apply SC-CO2 for the pasteurization of coconut water in order to guarantee its microbial stabil-ity and preserve its qualstabil-ity. The feasibilstabil-ity of the process was investigated to inacti-vate the natural microbial flora (mesophilic microorganisms, lactic acid bacteria, total coliforms, yeasts and molds) as a function of pressure, temperature and time.
The impact of the SC-CO2 process on the quality traits of coconut water was verified considering both chemical physical and sensory parameters and the results were com-pared with a reference product obtained by the traditional heat pasteurization. A deep chemical/physical characterization (pH, soluble solids, mineral salts, sugars, vitamins, volatile compounds) was performed to investigate the effects on the composition with particular attention to compounds with nutritional importance or sensory impact. A trained panel was also performed to verify whether the treatments induced sensory modifications, potentially perceptible by consumers, and to describe these differences in term of the intensity of sensory attributes related to odor and flavor.
Materials and Methods
1.1 Extraction of coconut water
Seventy young green coconuts (Cocos nucifera cv Nam Hom) from Thailand were bought and sent to Trento where they were aseptically opened, the water extracted and accumulated in a 20 liters plastic pail placed in ice. Once the extraction process ended, the coconut water was homogenized, portioned in sterilized glass jars of 200 and 400 ml and immediately frozen at -20°C to prevent any microbial or enzymatic activity. Before the treatments and the analysis, the samples were thawed (at 4°C for 12 h) and stored again at 4°C.
1.2 SC-CO2 treatment
SC-CO2 treatment was carried out in a multi-batch apparatus. The system consists of 10 identical reactors with an internal volume of 15 ml connected in parallel, so that
each experimental run provides a set of experimental data taken at identical process conditions but different treatment times. Each reactor is connected to an on-off valve that can be used to depressurize it independently from the others. The 10 reactors are submerged in a single temperature-controlled water bath. Liquid CO2 (Messer, Car-bon dioxide 4.0, purity 99.990%) is fed into the reactors by a volumetric pump (LEWA, mod. LCD1/M910s). The apparatus is provided with a pressure transducer while one cover lid of the 10 reactors is equipped with a fixed temperature probe (Pt 100 Ώ). The operating parameters (temperature and pressure) are continuously recorded by a real time acquisition data system (NATIONAL INSTRUMENTS, field point FP-1000 RS 232/RS 485) and monitored by a specific software (LabVIEW TM 5.0). The process conditions tested were: 60 and 120 bar, 22, 30, 35, 40 and 45°C with treatment times ranging from 5 up to 60 minutes. Once the optimal process con-ditions were defined, SC-CO2 treatments were performed in 310 ml reactors in order to obtain a higher amount of coconut water for physico-chemical, nutritional and sen-sory characterizations.
1.3 Heat treatment
Heat pasteurization equipment consisted of a water bath with an agitated platform in which the 200 ml coconut water jars where placed. Pasteurization of coconut water was performed at 72°C for 5 and 10 min and 90°C for 1, 3, 5, 10 and 20 min. The process conditions were chosen based on literature findings [2]. After the treatments, the samples were cooled down in a water bath and subjected to the microbial analysis. Once the optimal process conditions were defined, the treatment was performed in or-der to obtain a higher amount of coconut water for physico-chemical, nutritional and sensory characterizations.
1.4 Microbiological analysis
Coconut water (frozen at -20°C) was thawed at 4°C for 12 hours and then aged at 30°C for 18 hours to increase the initial microbial load. The microbial analyses were performed with the plate count method. The sample was serially diluted in a phos-phate buffer solution (PBS, 0.01 M, pH 7.4) and plated in duplicate onto selective me-dia. Mesophilic microorganisms, total coliforms, yeasts and moulds, and lactic acid bacteria were plated onto plate count agar (Liofilchem, TE, Italy), chromatic coli/col-iform agar (Liofilchem, TE, Italy), malt extract agar (Liofilchem, TE, Italy), and MRS agar (de Man, Rogosa and Sharpe, Oxoid, Milano, Italy), respectively.
The incubation temperatures and times were: 30 °C for 48 h for mesophilic micro-organisms; 30 °C for 24 h for total coliforms; 25 °C for 4 d for yeasts and moulds; and 35 °C for 48 h for lactic acid bacteria.
At the end of the incubation periods, the number of colonies was counted and the in-activation level was determined by evaluating the Log(CFU/ml) (logarithm of colony forming unit per ml of sample) of the microorganisms before and after the treatments. The results were means based on data from at least three experimental runs. Standard deviations were shown by error bars.
1.5 Product characterization
1.5.1 Basic physico-chemical and nutritional parameters
Soluble solids, expressed as °Brix, were measured using a digital bench and portable
refractometer (DBR95, Singapore). The pH of the samples was measured using a dig-ital pH meter (Inolab pH level 1, WTW GmbH, Weilheim). Water content was deter-mined by weight loss during heating at 70 °C for 96 h and dry matter was deterdeter-mined by weight difference. Sugars (sucrose, glucose, and fructose) were quantified by HPLC coupled to a refractive index detector (Agilent 1200). Minerals (Mg, K, Ca, Fe) were measured by Inductively Coupled Plasma Mass Spectrometer (Agilent 7500 ICP-MS). Water-soluble vitamins were analyzed by HPLC-DAD-MS/MS (Acquity/Xevo TQ; Waters).
1.5.2 Volatile compounds profiling
A 5 ml aliquot of coconut water was transferred to a 20 ml vial, and 2 g of NaCl were added together with 100 µl of internal standard (2-octanol, 0.25 mg/l). The vial was crimp-closed with a Teflon-lined silicacap (Supelco) and equilibrated at 50°C for 10 min under constant stirring. Solid Phase Microextraction fiber (DBV/CAR/PDMS, Supelco, Bellefonte, PA, USA) was exposed for 60 min in the vial headspace. The compounds obtained from coconut water by HS-SPME were analyzed on a GC inter-faced with a mass detector which operates in electronionisation mode (EI, internal ionisation source; 70 eV) with a scan range from m/z 35 to 300 (GC Clarus 500, PerkinElmer, Norwalk CT,USA). Separation was achieved on a HP-Innowax fused-silica capillarycolumn (30 m, 0.32 mm ID, 0.5 μm film thickness; Agilent Technolo-gies, Palo Alto, CA, USA). The GC oven temperature programme started at 40°C and arrived to 150°C at 2°C min−1, and then increased at 6°C min−1 till 250 °C which was maintained for 15 min. Helium was used as carrier gas with a constant column flow rate of 1.5 mL min−1.
1.6 Sensory analysis
In order to verify whether SC-CO2 treatment induces overall sensory differences in comparison to the fresh untreated and the pasteurized coconut water, the products were compared by smelling according to the standard triangle test procedure by a panel of 33 trained panelists [9]. Three consecutive triangle tests were performed in one session: Fresh vs. SC-CO2; SC-CO2 vs. Pasteurized; and Fresh vs. Pasteurized. Coconut water samples (10 ml) were presented in 40 ml screw cap vials coded with 3 digit random numbers. Test order was randomly presented and balanced over judges. A comparative descriptive flash profiling was also applied by a panel of 16 trained judges to establish how the treatments affected sensory properties [10]. By mean of a focus group, it was developed a vocabulary of consensus that consists of 14 obliga-tory descriptors: 6 odors, 2 tastes and 6 flavors (odor sensation by retro-olfactive eval-uation). The intensity of each descriptors was rated using a linear scale anchored to 0 (mimimum intensity) and 100 (maximum intensity). The samples were presented un-der red light and in the same conditions of temperature and quantity, in 80 ml plastic
glasses, closed with a lid, coded with random numbers in a random, balanced order for judge and analyzed in triplicate over three consecutive days. All tests were con-ducted in FEM sensory lab that has 22 booths for individual computer assisted evalua-tion. The implementation of the test, the recording of the judges’ responses and the data analysis were performed with FIZZ software 2.46A (Biosystemes, France). 1.7 Statistical analysis
Statistical analysis of triangle test data was based on binomial distribution with p=1/3. A significance level of 95% was assumed if not otherwise stated. A three-way Anova, including product and replicate as fixed factors and judge as random factor, was used to evaluate the performance of the descriptive panel and to identify the sensory at-tributes to discriminate the products (Statistica v.9, Statsoft Italia srl). Principal Com-ponent Analysis (PCA) was used for exploring VOC data by Simca P+ v.12 (Umet-rics, Sweden).
2. Results
3.1 Sample microbial contamination
The initial microbial load of coconut water was measured after the extraction process. The product showed 4 Log(CFU/ml) of mesophilic microorganisms and 2 Log (CFU/ml) of lactic acid bacteria, total coliforms and yeasts and molds. To evaluate the effect of the process variables and obtain reliable inactivation kinetics as a function of pressure, temperature and time, the initial microbial load was increased aging the product at 30°C for 18 hours. The resulting microbial load was about 8.5 Log(CFU/ml) of mesophilic microorganisms, lactic acid bacteria, and total coliforms and about 6 Log(CFU/ml) of yeasts and molds.
3.2 Heat pasteurization
The results of the microbial analysis performed on the heat pasteurized coconut water demonstrated that 90°C, 1 min were the optimal process conditions able to induce in-activation to undetectable levels of the natural microbial flora. At the same time, these process conditions induced fewer changes in the product visual appearance compared to the others conditions tested.
3.3 SC-CO2 microbial inactivation kinetics
The inactivation kinetics as a function of temperature and time are reported in Figs. 1-4 for each microbial strain and two pressures: 80 bar (a) and 120 bar (b).
The results indicated that the temperature induced a higher microbial reduction both at 80 and 120 bar. Several studies demonstrated that higher temperatures can increase CO2 diffusivity and the fluidity of cell membrane facilitating CO2 penetration into the cells [11, 12]. As concerns the effect of the increase of pressure from 80 to 120 bar, an enhanced microbial inactivation was also observed. It has been demonstrated that
higher pressures facilitate the penetration of CO2 into the cells, although this increase is limited by the saturation solubility of CO2 in the treatment medium [13].
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Fig 1. Mesophilic microorganisms inactivation kinetics as function of temperature and
treat-ment time at (a) 80 and (b) 120 bar.
The results also showed that microbial inactivation is highly dependent on the type of microorganisms present in the food matrix due to the distinct microbial cell mi-crostructure: total coliforms, and yeasts and molds (Figs. 3 and 4) were inactivated faster than mesophilic and lactic acid bacteria (Figs. 1 and 2).
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Fig. 2. Lactic acid bacteria inactivation kinetics as function of temperature and treatment time
at (a) 80 and (b) 120 bar.
On the basis of a multifactor Anova analysis performed on the results of the inactiva-tion kinetics, 120 bar, 40°C, 30 min were identified as the optimal process condiinactiva-tions able to achieve 5 Log reductions of all the microbial strains detected in the coconut water as required by the Food and Drug Administration (FDA).
(a)
(a)
(b)
Treatment time (min) 0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Fig. 3. Total coliforms inactivation kinetics as a function of temperature and treatment time at
(a) 80 and (b) 120 bar.
These conditions were also tested in the 310 ml reactors and the microbial analysis confirmed the results obtained in the multi-batch apparatus.
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Treatment time (min)
0 10 20 30 40 50 60 L og (C F U /m l) 0 2 4 6 8 10 22°C 30°C 35°C 40°C 45°C
Fig. 4. Yeasts and molds inactivation kinetics as function of temperature and treatment time at
(a) 80 and (b) 120 bar.
3.4 Product characterization
Both heat and SC-CO2 pasteurization treatments had no effect on dry matter, soluble solids and sugars content of coconut water as shown in Tables 1 and 2.
Table 1. Physico-chemical parameters of untreated (fresh) and treated coconut water.
Fresh product SC-C02 product Heat pasteurized product
pH 6.3 5.9 6.4
Water (%) 93.5 93.4 93.3
Dry matter (%) 6.5 6.6 6.7
Soluble solids (°Brix) 6.8 6.9 7
(a)
(a) (b)
SC-CO2 brought to a product with a pH lower than the fresh or the heat pasteurized (Table 1) while minerals were lower in the heat treated product probably due to floc-culation occurring during the process with subtraction of minerals from the soluble fraction. Only three water-soluble vitamins (B1, B2, B5) were founds in the fresh co-conut water (Table 2). Heat pasteurization brought to a slight decrease of all three vi-tamins as expected while SC-CO2 slightly reduced only vitamin B2 showing a lower detrimental impact on the nutritional values of the final product.
Table 2. Nutritional composition of untreated (fresh) and treated coconut water.
Fresh product SC-CO2 product Heat pasteurized product Sugars (g/l) Glucose 21.4 21.5 21.7 Fructose 19.8 20.1 20.2 Sucrose 13.1 13.3 13.3 Minerals (mg/l) Mg 155 182 111 K 2197 2577 1593 Ca 145 170 100 Fe 0.12 0.17 0.07 Vitamins (µg/l) B1 134.9 136.3 121.4 B2 72.6 67.9 67.1 B5 42.1 42.2 36.3
As concerns the volatile compounds, the 3 products were clearly distinguishable: SC-CO2 reduced most of the volatile fraction of coconut water (Fig. 5) as also observed in the literature results on SC-CO2 treated apple juice [14]. Only alcohols were slightly higher compared to the fresh product.
The main impact of the heat pasteurization was a considerable increase in ketones (6 fold) other than in alcohols (1.4 fold) and esters (1.5 fold). On the contrary, 3,4-dimethyl benzaldehyde (the main benzaldehyde derivative found in fresh coconut wa-ter headspace) was strongly depleted (9 times less) afwa-ter the heat pasteurization process compared to the fresh product. The 2-Acetyl-1-Pyrroline was found at very low concentration in the headspace of fresh and SC-CO2 processed coconut water. This compound that has a sensory impact described as “toasted”, “popcorn”, and “malty”, increased significantly in the heat pasteurized product (6 fold).
-1.0 -0.8 -0.6 -0.4 -0.2 -0.0 0.2 0.4 0.6 0.8 1.0 -1.1 -1.0 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 p(corr)[2] (X) t(corr)[2] (CO2 Past) t(corr)[2] (Fresh) t(corr)[2] (Heat Past)
SIMCA-P+ 12.0.1 - 2013-11-22 09:16:11 (UTC+1)
Fig. 5. Plot of the first 2 components of the PCA explaining 93% of sample variance using the
volatile compounds.
3.5 Sensory evaluations
Table 3 summarizes the results of the triangle tests. SC-CO2 treatment did not induce significant differences in the coconut water sensory quality at least by smelling. SC-CO2 product reported no significant differences compared to the fresh product while both were statistically different from the heat pasteurized one.
Table 3. Triangle test results. Statistically significant p-values are written in bold.
Comparison Total Correct % correct p-value
Fresh vs. SC-CO2 33 12 36% 0.419
SC-CO2 vs. Heat 33 17 52% 0.023
Fresh vs. Heat 33 17 52% 0.023
The descriptive profile showed a good panel performance with a high discriminant ca-pacity and reproducibility. In Fig. 6 the averaged profiles of each product in compari-son with the others are reported: SC-CO2 treated product and the fresh one are very similar for all the sensory attributes, except for “Cardboard” flavor which is less present in samples treated by SC-CO2. The pasteurized product seems to be character-ized by more intense odors and flavors of “Hazelnut” and “Toasted bread” (probably induced by aldehydes and ketones that are higher in the heat pasteurized product). These compounds probably induced the higher sensation of sweetness although no differences were detected for sugar contents (see Table 2). However, this result was in accordance with the increased amount of 2-Acetyl-1-Pyrroline found in the headspace, as previously reported.
Fig. 6. Spider plot for the average profiles of coconut water samples. For each sensory attribute
the names (O: odor; F: flavor) and its significance of ANOVA are reported (ns: not significant at 5%; *: 5%; **: 1%; ***: 0.1%)
Conclusions
The study indicated that SC-CO2 treatment was able to pasteurize coconut water. Process conditions of 120 bar, 40°C, 30 min induced about 5 Log reductions of mesophilic microorganisms, lactic acid bacteria, yeasts and molds and about 7 Log re-ductions of total coliforms. As regard as nutrional parameters very little differences are induced by SC-CO2 treatment whereas VOC analysis highlighted clear differences among fresh, heat and SC-CO2 treated products. Nevertheless, no sensory differences were perceived between the fresh coconut water and the product treated by SC-CO2, both clearly different from the pasteurized one.
References
1. Sawka, M.N., Burke, L.M., E Randy Eichner, E.R: American College of Sports Medicine position stand. Exercise and fluid replacement. Med. Sci. Sports. Exerc. 39, 377–390 (2007)
2. Reddy, K.V.: Processing of green coconut water for bottling. M. Tech. Thesis, Dairy and Food Engineering, Department of Agricultural Engineering, Indian In-stitute of Technology, Kharagpur, India (1995)
3. Bergonia, H.A., Papa, G.M., Fabilitante, J.R., Del Rosario, E.J.: Reverse osmosis of coconut water through cellulose membrane. Philipp. J. Food Sci. Technol. 6, 31–40 (1982)
4. Magda, R.R.: Coco soft drink: Health beverage from coconut water. Food Marketing Technol. 6, 22–23 (1992)
5. Reddy, K.V., Das, M., Das, S.K.: Filtration resistances in nonthermalsteriliza-tion of green coconut water. J. Food Eng. 69, 381–385 (2005)
6. Arreola, A.G., Balaban, M.O., Marshall, M.R., Peplow, A.J., Wei, C.I., Cornell, J.A.: Supercritical carbon dioxide effects on some quality attributes of single strength orange juice. J. Food Sci. 56, 1030–1033 (1991)
7. Zhou, L., Wang, Y., Hu, X., Wu, J., Liao, X.: Effect of high pressure carbon diox-ide on the quality of carrot juice. Innov. Food Sci. & Emerg. Tchnol. 10, 321–327 (2009)
8. Spilimbergo, S., Ciola, L.: Supercritical CO2 and N2O pasteurization of fresh peach and kiwi juice. Int. J. Food Sci. & Technol. 45, 1619-1625 (2010)
9. ISO Standard 4120: Sensory analysis, methodology, triangular test. Geneva, Switzerland: ISO (2004)
10.Delarue, J., Sifferman, J.-M.: Sensory mapping using Flash profile. Comparison with aconventional descriptive method for the evaluation of the flavour of fruit dairy products. Food Qual. and Pref. 15, 383-392 (2004)
11.Lin, H.M., Yang, Z., Chen, L.F.: Inactivation of Leuconostoc dextranicum with carbon dioxide under pressure. Chem.Eng. J. 52, B29–B34 (1993)
12.Erkmen, O.: Predictive modeling of Listeria monocytogenes inactivation under high pressure carbon dioxide. Leben. Wiss. und Technol. 33, 514–519 (2000) 13.Damar, S., Balaban, M.O., Sims, C.A.: Continuous dense-phase CO2 processing of
a coconut waterbeverage. Int. J. Food Sci & Technol. 44, 666-673 (2009)
14.Gasperi, F., Aprea, E., Biasioli, F., Carlin, S., Endrizzi, I., Pirretti, G., Spilim-bergo, S.: Effects of supercritical CO2 and N2O pasteurisation on the quality of fresh apple juice. Food chem. 115, 129-136 (2009).
