Fruit Manufacturing

2 . Processing of Fruits 41 Figure 2.14. Scanning electron photomicrograph of an isolated apple starch granule (5 kV Â 4,400). The scanning electron micrograph shows how the apple starch granules collapsed after heat treatment, while dispersion of gel-like starch fragments among the other components of turbidity (pectin, cellular wall, etc.) can be observed. Similar behavior was found when wheat starch was gelatinized by heat in excess water (Lineback and Wongsrikasen, 1980). Figure 2.15. Scanning electron photomicrograph of cloudiness precipitated from a pasteurized (5 min at 908 + 1 C) apple juice (5 kV Â 3,600).

42 Fruit Manufacturing 2.2.5. Filtration Filtration is also a mechanical process designed for clarification by removing insoluble solids from a high-value liquid food, by the passage of most of the fluid through a porous barrier, which retains most of the solid particulates contained in the food. Filtration is performed using a filter medium, which can be a screen, cloth, paper, or bed of solids. Filter acts as a barrier that lets the liquid pass while most of the solids are retained. The liquid that passes through the filter medium is called the filtrate. There are several filtration methods and filters (Lozano, 2003) including: . Driving force: The filtrate is induced to flow through the filter medium by: (a) hydro- static head (gravity), (b) pressure (upstream filter medium), (c) vacuum (downstream filter medium), and (d) centrifugal force across the medium. . Filtration mechanism: (a) cake filtration (solids are retained at the surface of a filter medium and pile upon one another). (b) depth or clarifying filtration (solids are trapped within the pores or body of the filter medium). . Operating cycle: (a) Intermittent (batch) and (b) Continuous rate. These methods of classification are not mutually exclusive. Thus filters usually are divided first into the two groups of cake and clarifying equipment, then into groups of machines using the same kind of driving force, then further into batch and continuous classes. Some filtering devices usually employed in the food industry are described here. 2.2.5.1. Pressure Filters The main advantages of pressure filtration compared to other filtering methods are: Cakes are obtained with very low moisture content, clean filtrates may be produced by recirculating the filtrate or by precoating, and the solution can be polished (finished) to a high degree of clarity. Among disadvantages it must be noted that cloth washing is difficult, and if precoat is required, the operator cannot see the forming cake and is unable to carry out an inspection while the filter is in operation, and the internals are difficult to clean, which can be a problem with food-grade applications. With the exception of the rotary drum pressure filter, this type of filters has a semicontin- uous machine in which wash and cake discharge are performed at the end of the filtration cycle. Since the operation is in batches, intermediary tanks are required. The collection of filtrate depends on the operating mode of the filter, which can be at constant flow rate, constant pressure, or both, with pressure rising and flow rate reducing during filtration. The filtration rate is mainly influenced by the properties of food (particle size and distribution, presence of gelatinous solids like pectin, liquid viscosity, etc.) Although continuous pressure filters are available, they are mechanically complex and expensive, so they are not common in the food industry. 2.2.5.2. Filter Aid and Precoating Filter aid and precoating are often used in pressing and in connection with pressure filtration. Filter aid is used when the pulp or turbid liquid food is low in solids’ content with fine and muddy particles that are difficult to filter. To enhance filtration coarse solids with large surface area that capture and trap the slow-filtering particles from the suspension in its interstices and produces a porous cake matrix are used.

2 . Processing of Fruits 43 On the other hand, precoating is the formation of a defined thick medium of a known permeability on the filter plates. Precoating prior to filtration serves when the particles that are to be separated are gelatinous and sticky, forming a barrier that avoids cloth blinding. The most common filter aids and precoating materials employed in the food industry are: diatomaceous earth (silicaceous skeletal remains of aquatic unicellular plants), perlite (glassy crushed and heat-expanded volcanic rock), cellulose (fibrous light weight and ashless paper), and special groundwood. 2.2.5.3. Types of Pressure Filters Pressure filters usually found in the food industry are: . Filterpress, also called Plate and Frame, consists of a head and a follower that contain in between a pack of vertical rectangular plates that are supported by side or overhead beams (Fig. 2.16). The head serves as a fixed end to which the feed and filtrate pipes are connected, while the follower moves along the beams and presses the plates together during the filtration cycle by a hydraulic or mechanical mechanism. Each plate is covered with filter cloth on both sides and, once pressed together, they form a series of chambers that depend on the number of plates. The plates have generally a centered feed port that passes through the entire length of the filter press, so that all chambers of the plate pack are connected together. . Vertical and horizontal pressure leaf filters consist of a vessel that is fitted with a stack of vertical (Fig. 2.17), or horizontal leaves that serve as filter elements. The leaf is constructed with ribs on both sides to allow free flow of filtrate toward the neck and is covered with coarse mesh screens that support the finer woven metal screens or filter cloth that retain the cake. The space between the leaves varies from 30 to 100 mm depending on the cake formation properties and the ability of the vacuum to hold a thick and heavy cake to the leaf surface. The filters can be used for polishing fruit juices Filtrate out Feed Filtrate Feed Filtrate Figure 2.16. Filterpress operation sketch and details of filtering plate. Reprinted from the Encyclopedia of Food Science and Nutrition, Lozano J.E., Separation and Clarification, pp. 5187–5196. (copyright) 2003, with permission from Elsevier.

44 Fruit Manufacturing Feed in Plate Filtrate Figure 2.17. Horizontal plate filter. Reprinted from the Encyclopedia of Food Science and Nutrition, Lozano J.E., Separation and Clarification, pp. 5187–5196. (copyright) 2003, with permission from Elsevier. with very low solids or for cake filtration with a solids’ concentration of <20–25%. The cloth mesh screens that cover the leaves can be more easily accessed on horizontal tanks than on vertical tanks. . Candle filters are used in applications that require efficient low moisture cake filtration or high degree of polishing (finishing). Candle filters can contain more than 250 filtering elements. They consist basically of three major components (Fig. 2.18): (1) The vessel; (2) The filtering elements (candles) and (3) The cake discharge outlet. The vessel configuration has a conical bottom for cake filtration and polishing, or has a dished bottom for slurry thickening, though it is scarcely used in the food industry. The filtering element generally consists of the filtrate core and the filtering medium. The core helps in filtrate passage and supports the filter medium. The core is a bundle of Filtrate Bunch of Vessel candles Figure 2.18. Candle filter sketch showing major components. Reprinted from the Encyclopedia of Food Science and Nutrition, Lozano J.E., Separation and Clarification, pp. 5187–5196. (copyright) 2003, with permission from Elsevier.

2 . Processing of Fruits 45 perforated stainless steel tubes. The filter medium can be a porous ceramic, woven mesh screen, sintered metal tube, or synthetic filter cloth. The advantages of a candle filter are the excellent cake discharge capacity and its mechanical simplicity. 2.2.5.4. Vacuum Filters Vacuum filters are simple and reliable machines widely used in the fruit industry. Among the different types of vacuum filter (drum, disk, horizontal belt, tilting pan, and table filters) drum filters are most commonly utilized in the food industry. The advantages and disadvan- tages of vacuum filtration compared to other separation methods are: . Advantages: Continuous operation, very effective polishing (finishing) of solutions (on a precoat filter), and easy control of operating parameters such as cake thickness. . Disadvantages: Higher residual moisture in the cake and difficulty in cleaning (as required mainly for food-grade applications). Precoat filters are used when liquid foods (e.g., clarified apple juice) require a very high degree of clarity. To polish the solution the drum deck is precoated with an appropriate medium (See Filter aids and precoating in this chapter). A scraper blade also called ‘‘Doctor Blade,’’ moves slowly toward the drum and shaves off a thin layer of the separated solids and precoating material. This movement exposes continuously a fresh layer of the precoat surface, so that when the drum submerges into the tank it is ready to polish the solution. Precoat filters are used to recover juice from the sediments originating in clarification tanks. In precoat filters the entire drum deck is subjected to vacuum (Fig. 2.19). 2.2.6. Membrane Filtration Filtration of coarse particles down to several microns is done by conventional dead-end filtration, where all influent passes through a filter medium that removes contaminants to produce higher-quality clarified juices. Rough screens, sand filters, multimedia filters, bag filters, and cartridge filters are examples of filtration products that remove 0.1-micron size particles or larger. Once the medium becomes loaded, it can be backwashed as with multi- media filters, or discarded and replaced as with cartridge filters. The method of obtaining clean filtration medium is based on economic and disposal concerns. Particles retained by the filter in dead-end filtration build up with time as a cake layer, which results in increased resistance to filtration. This requires frequent cleaning or replacement of filters. Doctor blade Vacuum pump Receiver Feed Figure 2.19. Vacuum drum precoat filter.

46 Fruit Manufacturing Commercially available coarse filtration devices are effective in separating particles down to about 20 mm. On the other hand, membrane technology involves the separation of particles below this range, extending down to dissolved solutes that are as small as several Angstroms. Membranes are manufactured with a wide variety of materials, including sintered metals, ceramics, and polymers (Zeman and Zydney, 1996). In order to reject substances smaller than 0:1 mm using polymeric membrane is by far the most popular filtration method in the fruit processing industry. Millions of small pores per unit area of membrane allow water and low- molecular weight substances to pass through it while undesired substances are retained on the influent side. The problem is solved by operating polymeric membranes in the crossflow mode. In crossflow membrane filtration, two effluent streams are produced, the permeate and the concentrate. The permeate is the purified fluid that has passed through the semi- permeable membrane. The remaining fluid is the concentrate, which has become enriched with organics and salts that could not permeate the membrane. By doing so, rejected contaminants are continuously carried away from the membrane surface, thereby minimizing contaminant build up, leaving it free to reject incoming material and allow free flow of purified liquid. The size of the polymer’s pores categorizes the membrane into one of the following groups: reverse osmosis, nanofiltration, ultrafiltration, and microfiltration. Reverse osmosis (RO) membranes have the smallest pore size ranging from 5 to 15 angstroms, nanofiltration (NF) covers separations in the 15–30 angstrom size, ultrafiltration (UF) removes organics in the 0:002---0:2 mm range, while microfiltration (MF) effects separation typically in the range of 0:1---10 mm. Figure 2.20 schematically shows the filtering capacity of these ‘‘crossflow’’ membrane systems. RO is both a mechanical and chemical filtration procedure by which the membrane’s surface sieves organic substance and actually repels ions. The dielectric repulsion of ions from the membrane is influenced by the ion’s charge density. Unlike RO membranes that have salt retentions of 80–99%, NF membranes reject 30–60% salts. UF and MF membranes have even larger pores and therefore pass most of the salts. Although membrane cleaning is periodically Suspended Ions Water particles (multivalent) U NR Macromolecules F Ions F (monovalent) O Figure 2.20. Crossflow-type membrane classification by ‘‘rejection’’ capacity.

2 . Processing of Fruits 47 required, the self-cleaning nature of crossflow filtration lengthens membrane life enough to make it economically attractive. The manufacturing processes result in a number of different membrane structures such as microporous, asymmetric, composite, etc. Membranes are assembled as modules that are easily integrated into systems containing hydraulic components. The module allows to accommodate large filtration areas in a small volume and resist the pressures required in filtration. Tubular, hollow fiber, spiral, and flat plate are the common modules (Cheryan, 1986). . Tubular module consists of tubular membranes held inside individual perforated sup- port tubes, assembled onto common headers and permeate into the container to form a module. When several channels are formed in a porous block of material, the tubular system is called ‘‘monolithic.’’ . Hollow fiber module consists of bundles of hollow fibers (0.5–3 mm internal diameter) sealed into plastic headers and assembled in permeate casings (Fig. 2.21). Hollow fibers Figure 2.21. Hollow fiber membrane configuration. (a) Manifold with HF cartridge; (b) SEM micrography of a single fiber (internal diameter 1 mm); (c) SEM magnification of a single fiber, showing filtration surface and support.

48 Fruit Manufacturing are used in low-pressure applications only. These can accommodate moderate levels of suspended particles. . Spiral modules are made by placing a woven plastic mesh, which acts as the permeate channel between two membrane layers and seals three sides. The fourth side of this sandwich is attached to the permeate tube. Another plastic mesh that acts as the feed channel is laid over it and the assembly is wrapped around the central permeate tube. . Flat-platemodules use multiple flat sheet membranes in a sandwich arrangement con- sisting of a support plate, membrane, and channel separator. The membranes are sealed to the plates using gaskets and hydraulically clamped to form a tight fit. Several of these membranes are stacked together and clamped to form a complete module. The main advantages of the flat-plate module design are that they have high membrane- packing densities and low hold-up volumes. This is due to the small channel height of the flat-plate modules. The main application for the flat-plate module is in recovering biological products. The advantages and disadvantages of the different UF configur- ations (Cheryan, 1986) are listed in Table 2.7. Over time, the physical backwash will not remove some membrane fouling. Most membrane systems allow the feed pressure to gradually increase over time to around 30 psi and then perform a clean-in-place (CIP). CIP frequency might vary from around 10 days to several months. Another approach CIP practice is to use a chemically enhanced backwash (CEB), where on a frequent basis (typically every 1–14 days), chemicals are injected with the backwash water to clean the membrane and maintain system performance at low pressure without going offline for a CIP. The application of ultrafiltration (UF) as an alternative to conventional processes for clarification of apple juice was clearly demonstrated (Heatherbell et al., 1977; Short, 1983; Wu et al., 1990). However, the acceptance of UF in the fruit processing industry is not yet complete, because there are problems with the operation and fouling of membranes. During UF two fluid streams are generated: the ultrafiltered solids’ free juice (permeate), and the retentate with variable content of insoluble solids, which, in the case of apple juice, are mainly remains of cellular walls and pectin. Permeate flux (J) results from the difference between a convective flux from the bulk of the juice to the membrane and a counterdiffusive flux or outflow by which the solute is transferred back into the bulk of the fluid. The value of J is strongly dependent on hydro- dynamical conditions, membrane properties, and the operating parameters. The main driving force of UF is the transmembrane pressure (DPTM), which in the case of hollow fiber ultrafiltration systems (HFUF) can be defined as: DPTM ¼ (Pi þ Po) ÀPext (2:1) 2 Table 2.7. Advantages and disadvantages of the different UF configurations. UF MEMBRANE CONFIGURATION Pack density (m2=m3) Flat/press Spiral Tubular Hollow fibers Fouling resistance Cleaning facility 300–500 200–800 30–200 500–9000 Relative cost Good Moderate Very good Poor Good Fair Very good Poor High Low High Low Source: Cheryan, 1986; Zeman and Zydney, 1996.

2 . Processing of Fruits 49 where Pi is the pressure at the inlet of the fiber, Po the outlet pressure, and Pext the pressure on the permeate side. In practice, the J values obtained with apple juice are much less than those obtained with water only. This phenomenon is attributable to various causes, including resistance of gel layer, concentration in polarization boundary layer (defined as a localized increase in concentration of rejected solutes at the membrane surface due to convective transport of solutes (Porter, 1972)), and plugging of pores due to fouling, where some of these phenomena are reversible and disappear after cleaning of the UF membranes while others are definitively irreversible. 2.2.6.1. Stationary Permeate Flux It is well known (Iritani et al., 1991) that the transmembrane pressure-permeate flux charac- teristic for ultrafiltration shows a linear dependence of J with DPTM at lower values of pressure (1st region), while the permeate flux approaches a limiting value (Jlim) independent of further increase in pressure at higher pressures (2nd region). The last situation was assumed to be controlled by mass transfer. Figures 2.22 and 2.23 show the variation of J with DPTM as a function of VCR or volume concentration ratio (defined as the initial volume divided by retentate volume at any time), which is a measurement of the retentate concentration, and recirculating flow rate, Qr, respectively (Constenla and Lozano, 1996). Pressure independence (2nd region) was observed to occur at a higher pressure at higher Qr. The point at which the pressure independence is evident is called optimum transmembrane pressure (DPTMo). 80.00 T = 50؇C, PC = 50,000 ds, VCR = 1 J,L/hm2 60.00 40.00 Qr = 10 L/min 20.00 Qr = 12.5 L/min Qr = 15 L/min 0.00 0.40 0.80 1.20 1.60 0.00 ∆PTM, Kg/cm2 Figure 2.22. Effect of DPTM and Qr on J at 508C. (Constenla and Lozano, 1996) with permission.

50 Fruit Manufacturing 80.00 T = 50؇C, PC = 50,000 ds, Qr = 10 L/min J,L/hm2 60.00 40.00 20.00 VCR = 1 VCR = 2 VCR = 5 0.00 0.00 0.40 0.80 1.20 1.60 DPTM, Kg/cm2 Figure 2.23. Effect of DPTM and VCR on J at 508C (Constenla and Lozano, 1996). Reprinted from Lebensm. Wiss. u. Technol. 27: 7–14, Constenla, D.T. Lozano, J.E., Predicting stationary permeate flux in the ultrafilteration of apple juice. (copyright) 1996, with permission from Elsevier. The reduction of Jlim with Qr may be associated with a reduction in the boundary layer due to an increase in the turbulence. On the other hand, the optimal DPTM values were practically independent of VCR at Qr > 10 L/min. A hysteresis effect in the permeate flux, attributable to the consolidation of the gel layer (Omosaiye et al., 1978), has been observed. The area enclosed by the hysteresis loop was greater at lower Qr and VCR values. Tradition- ally, correlations of J with DPTM and VCR were determined by parameter fitting of the experimental data. Since the polynomial functions have no physical basis, a large number of experimental data are needed for determination of J. Therefore other theoretical and semiempirical approaches should be considered (Constenla and Lozano, 1986). 2.2.6.2. Permeate Flux as a Function of Time Several models are proposed in the literature for representing J, most of them being semi- empirical and practical equations (Table 2.8). Membrane fouling mechanisms may be studied through the classical laws of filtration under constant pressure (Table 2.9). During UF process (Iritani et al., 1991) J behaves as in cake filtration only at the very beginning, attributable to the formation of the gel layer with minor counterdiffusion flux.

2 . Processing of Fruits 51 Table 2.8. Some equations representing permeate flux as a function of time. No. Permeate flux equations (i) J ¼ J0 exp ( À Bt) (ii) J ¼ JF þ (J0 À JF) exp ( À At) (iii) J ¼ (J0À2 þ 2K t)À1=2 (iv) J ¼ J0 À B ln(VCR) (v) J À JF ¼ (J0 À JF)(1 À exp ( ( À t=B) ) Subindex: 0, 1, F are zero, initial, and final time, respectively; Vp ¼ permeate volume; A, B, and K are constants. Sources: Heatherbell et al. (1977), Probstein et al. (1979), Mietton-Peuchot et al. (1984), Koltuniewicz (1992), Con- stenla and Lozano (1996). As previously indicated, pectin and other large solutes like starch, normally found when unripen apples are processed, tend to form a fairly viscous and gelatinous-type layer on the ‘‘skin’’ of the asymmetric fiber. Flux decline, due to this phenomenon, can be reduced by increasing flow velocity on the membrane Traditionally correlations of J with DPTM and VCR have been determined by parameter fitting of the experimental data. It was found that the following exponential equation, proposed in the SRT model (Constenla and Lozano, 1996), fitted appropriately: J ¼ JF þ (JO À JF) exp ( À At) (2:2) Jo, JF, and A values can be obtained at different Qr and constant values of VCR and DPTM. An increase in Qr significantly increases the permeate flux. This behavior was reflected as an extensive increase in the parameter A. 2.2.6.3. Influence of VCR on the Permeate Flux Constenla and Lozano (1996) found that in the case of pseudoplastic fluids, as fruit juice retentates, different operative conditions restrain the VCR up to a maximum of 14. The permeate flux becomes independent of the solute rejection, characteristic of the hollow fibers Table 2.9. Classic filtration models (pseudoplastic fluids). Mechanism Scheme Representative equations (1) Total pore blocking J0 À J ¼ K1Vp ln (J=J0) ¼ ÀK1 (2) Partial pore blocking . ln (J=J0) ¼ ÀK2Vp1=J À 1=J0 ¼ K2t (3) Blocking Progressive pore J ¼ J0(1 À K3Vp=2)(3n þ 1)=2nJ (4) Cake filtration ¼ J0(1 þ ((n þ 1)=n)J0K3t)À(3n þ 1)=(n þ 1) (1=J)n ¼ (1=J0)n þ K4Vp: (1=J)nþ1 ¼ (1=J0)nþ1 þ ((n þ 1)=n)K4t K1, K2, K3, K4: experimental constants; n: flow behavior index. Source: Lozano et al. (2000). (with permission)

52 Fruit Manufacturing after a few minutes of operation. This effect is commonly attributed to the build up of the concentration polarization/gel layer. During the UF of apple juice in the mass transfer region, a 60% increase in DPTM was reflected only as a 5% increase in J. Acceleration of the fruit juice retentate near the membrane surface removes the accumu- lated macromolecules, thereby reducing the effect of concentration polarization gel layer. Due to the low diameter of the hollow fibers, high-tangential velocities can be obtained at laminar rates. Equation iv in Table 2.8, fit reasonably well for these types of membranes: J ¼ Jo À B In (VCR) (2:3) where JO is the initial permeate flux, and B is a constant, which depends on the system, operating conditions, and juice properties. Decrease of flux with concentration is nonlinear, and changes in the rate of permeation were better followed when plotted against ln VCR (Fig. 2.24). The rate of flux decrease J could be divided into three periods. The first period, characterized by a rapid decrease in J, occurred in a few minutes. During the second period (up to VCR ¼ 3 approximately) the variation of J is unstable, depending on fiber cut-off. Then J approached a ‘‘linear’’ steady logarithmic decrease with VCR. This behavior could be explained by considering the resistance to flux as two separate additive resistances in series: (i) the membrane resistance (Rm0 ); and (ii) the concentration polarization/gel layer resistance (Rp). During the first period Rp increases very fast reaching a value equivalent to that of R0m. In the second region, the R0m value is still an important 100.00 PC = 30,000 ds J, L/hm2 PC = 50,000 ds PC = 10,000 ds 80.00 I 60.00 II 40.00 III 20.00 0.00 2 3 45 6 7 8 9 10 1 VCR Figure 2.24. Decrease of permeate flux with ln VCR for hollow fibers with different MWCO. Full line represents Eq. (2.4). Reprinted from Lebensm. Wiss. u. Technol. 27: 7–14, Constenla, D.T. Lozano, J.E., Predicting stationary permeate flux in the ultrafilteration of apple juice. (copyright) 1996, with permission from Elsevier.

2 . Processing of Fruits 53 component of the total resistance and J is not completely independent of the properties of the fiber. Finally, during the last period Rp is dominant and the cut-off of the hollow fiber becomes irrelevant. REFERENCES Alkorta, I., Garbisu, C., Llama, M.J. and Serra, J.L. (1996). Immobilization of pectin lyase from Penicillium Italicum by covalent binding to Nylon. Enzyme Microb. Technol. 18: 141–146. Ates, S. and Pekyardimci, S. (1995). Properties of immobilized Pectinesterase on Nylon. Macromol. Rep. A32: 337–345. Ben-Shalom, N., Levi, A. and Pinto, R. (1986). Pectolytic enzyme studies for peeling of grapefruit segment mem- brane. J. Food Sci. 51: 421–423. Bump, V.L. (1989). Apple pressing and juice extraction. In Processed Apple Products, Downing, D.L. (ed.). AVI Publishing Company, Van Nostrand Reinhold, New York, pp. 53–82. Ceci, L. and Lozano, J.E. (1998). Determination of enzymatic activities of commercial pectinases. Food Chem. 31(1/2): 237–241. Cheryan, M. (1986). Ultrafiltration Handbook. Technomic Publishing Company, Lancaster. Constenla, D.T. and Lozano, J.E. (1996). Predicting stationary permeate flux in the ultrafiltration of apple juice. Lebensm. Wiss. Technol. 27: 7–14. Dietrich, H., Patz, C., Scho¨ pplain, F. and Will, F. (1991). Problems in evaluation and standardization of enzyme preparations. Fruit Process. 1: 131–134. FAOSTAT Data (2005). FAO Statistical Databases. www.fas.usda.gov/htp/Presentations/2005. Felloes, P. (1988). Food Processing Technology: Principles and Practice. Ellis Horwood International Publishers, Chichester, England, pp. 300–310. Grampp, E.A. (1976). New process for hot clarification of apple juice for apple juice concentrate. Fluss. Obst. 43: 382–388. Heatherbell, D.A., Short, J.L. and Stauebi, P. (1977). Apple juice clarification by ultrafiltration. Confructa 22: 157–169. Hui, Y.H. (1991). Data sourcebook for Food Scientists and Technologists. VCH Publishers, New York. Iritani, E., Hayashi, T., and Murase, T. (1991). Analysis of filtration mechanism of crossflow upward and downward ultrafiltration. J. Chem. Eng. Jpn 1: 39–44. Koltuniewicz, A. (1992). Predicting permeate flux in ultrafiltration on the basis of surface renewal concept. J. Membr. Sci. 68: 107–118. Lineback, D.R. and Wongsrikasen, E. (1980). Gelatinization of starch in baked products. J. Food Sci. 45: 71–74. Liu, Y.K. and Luh, B.S. (1978). Purification and characterization of endo-polygalacturonase from Rhizopus arrhizus. J. Food Sci. 43: 721–726. Lozano, J.E. (2003). Separation and clarification. In Encyclopedia of Food Science and Nutrition, Caballero, B., Trugo, L. and Finglas, P. (eds.). AP Editorial, Elsevier, London, UK, pp. 5187–5196. ISBN: 0-12-227055-X. Lozano, J.E., Constenla, D.T. and Carr´ın, M.E. (2000). Ultrafiltration of apple juice. In Trends in Food Engineering, Lozano, J.E., An˜ o´ n, C., Parada-Arias, E. and Barbosa-Ca´novas, G. (eds.). Food Preservation Technol. Series. Technomics Publishing Company, Inc., Lancaster, Basel, pp. 117–134. McLellan, M.R. (1996). Juice processing, Chapter III. In Processing Fruits: Science and Technology. Biology, Principles and Applications, Vol. 1, Somogyi, L.P., Ramaswamy, H.S. and Hui, Y.H. (eds.). Technomic Publishing Company, Inc., Lancaster, Basel. Mietton-Peuchot, M., Milisic, V. and Ben Aim, R. (1984). Microfiltration tangentielle des boissons. Le Lait. 64, 121–128. Nagy, S.; Chen C.S. and Shaw, P.E. (eds.) (1993). Fruit Juice Processing Technology. Agscience, Inc. Auburndale, FL. Omosaiye, O., Cheryan, M. and Mathews, M. (1978). Removal of oligosacharides from soybean water extracts by ultrafiltration. J. Food Sci. 51: 354–358. Porter, M. (1972). Concentration polarization with membrane ultrafiltration. Ind. Eng. Chem.—Prod. Res. Develop. 11(3): 234–248. Probstein, R., Leung, W. and Alliance, Y. (1979). Determination of diffusivity and gel concentration in macromol- ecular solutions by ultrafiltration. J. Phys. Chem. 83(9): 1228–1236.

54 Fruit Manufacturing Ramaswamy, H.S. and Abbatemarco, C. (1996). Thermal processing of fruits. In Processing Fruits: Science and Technology, Vol. I, pp. 25–65. Reed, G. (1975). Enzyme in food processing, 2nd ed. Academic Press, New York. Rombouts, F.M. and Pilnik, W. (1978). Enzymes in fruit and vegetable juice technology. Process Biochem. 13: 9–13. Sakai, T., Sakamoto, T., Hallaert, J. and Vandamme, E.J. (1993). Pectin, pectinase, and protopectinase: production, properties, and applications. Adv. Appl. Microbiol. 39: 213–294. Short, J.L. (1983). Juice clarification by ultrafiltration. Process Biochem. 18(5): VI. Somogyi, L.P., Ramaswamy, H.S. and Hui, Y.H. (eds.) (1996). In Processing Fruits: Science and Technology. V.2 Major Processed Products, Technomic Publishing Company, Inc., Lancaster, PA. Spagna, G., Pifferi, P.G. and Martino, A. (1993). Pectinlyase immobilization on epoxy supports for application in the food processing industry. J. Chem. Tech. Biotechnol. 57: 379–385. Toribio, J.L. and Lozano, J.E. (1984). Non enzymatic browning in apple juice concentrate during storage. J. Food Sci. 49: 889–893. Woodroof, J.G. and Luh., B.S. (1986). Commercial fruit processing, 2nd ed. AVI Publishing Company, Westport, CT. Wu, M.L., Zall, R.R. and Tzeng, W.C. (1990). Microfiltration and ultrafiltration comparison for apple juice clarification. J. Food Sci. 55(4): 1162–1163. Zeman, L.J. and Zydney, A.L. (1996). Microfiltration and Ultrafiltration: Principles and Applications. Marcel Dekker, Inc., New York, NY. Zobel, H.F. (1984). Starch gelatinization and mechanical properties. In Starch: Chemistry and Technology, 2nd ed., Whstler, R.L., BeMiller, J.N. and Paschall, E.F. (eds.), Academic Press, Orlando, FL, pp. 300–302.

CHAPTER 3 PROCESSING OF FRUITS: ELEVATED TEMPERATURE, NONTHERMAL AND MISCELLANEOUS PROCESSING 3.1. PASTEURIZATION The process of ‘‘pasteurization’’ pioneered by Louis Pasteur was aimed at the destruction of bacteria, molds, spores, etc. by exposing them to a certain minimum temperature for a certain minimum time; the higher the temperature, the shorter the time required. The term ‘‘pasteur- ized’’ can be used to refer to products with reduced bacteria. Products with no bacteria are referred to as ‘‘sterile’’ or ‘‘ultrapasteurized.’’ Some products are ‘‘sterilized’’ before they are sold to the public. Most of the fruit juice sold on store shelves is produced this way. These products have relatively unlimited shelf life even without refrigeration. However, the time/temperature combination required to kill 100% of bacteria also destroys some of the flavor components in the juice. There is some dispute over how much flavor degradation actually occurs, and since this is a subjective opinion on the part of the consumer, no definitive data are available. The following methods are commonly accepted for pasteurization. 3.1.1. Batch Pasteurization This is a typical pasteurization process, by heating the product in a batch pan to about 638C for relatively long periods (Table 3.1). This method destroys very common pathogenic bacteria. However, as production demands grow, simply adding more number of pans is usually not feasible. 3.1.2. HTST (Short Time) Pasteurization High-Temperature, Short-Time pasteurization is typically conducted at 728C for 15 s. A hold time of 15 s can easily be achieved in a continuous process by installing a holding tube. The product is then cooled for storage. This method provides the convenience of continuous processing, at a temperature low enough to prevent taste and aroma deterioration. 3.1.3. UHT Pasteurization In UHT pasteurization, the product is brought to over the boiling point (under pressure) for only a fraction of a second. This results in a sterile product that does not require refrigeration 55

56 Fruit Manufacturing Table 3.1. Typical pasteurization methods and conditions. Method Temperature (8C) Hold time Batch 63 30 min HTST (High temperature/short time) 72 15 s UHT (Ultrahigh temperature) >121 0.1 s Source: Nickerson and Sinskey, 1972; US Department of Health and Human Services, 2004. later. However, after being brought to this temperature, a slight ‘‘cooked’’ taste is sometimes said to be detectable. Most apple juice producers are relatively familiar with both batch and UHT pasteuriza- tion. While smaller producers use batch method, UHT systems are commonly employed by large processors. 3.1.4. Nonthermal Pasteurization Much has been written about ‘‘new’’ pasteurization and sterilization technologies such as irradiation, microwave sterilization, and high-pressure processes that have long been avail- able, but, for various reasons, scarcely applied in food processing. Meanwhile, new or improved thermal and nonthermal technologies have emerged that are available now for pasteurizing, sterilizing, or otherwise reducing microbiological contamination of foods. Moreover, some new nonthermal pulsed technologies have been cleared by the FDA for antimicrobial applications (Bolando-Rodriguez et al., 2000). Traditional thermal sources, such as steam, are being engineered into new and improved processes to destroy pathogens with minimal heat, as for example the continuous steam- fusion cookers. This technique involves a variable-speed agitator, which rotates on an axis parallel to product flow through a vertical tube. A row of steam injectors along each side of the tube provides rapid heating, while turbulence created by the agitator fuses steam to evenly heat the product as it flows through the tube. The product goes up to cooking temperature in just a few seconds, and steam is combined with the product to avoid overcooking. Cooker surfaces are kept at practically the same temperature as the product. Pressurized juice should be preserved under chilled conditions to retain its fresh flavor and taste. Low temperature also helps to reduce the development of precipitates, since low- temperature storage keeps pectin esterase activity low; thus, pectin esterase cannot participate in the formation of a precipitate. New nonthermal technologies with potential for processing juices to retain flavor and extend shelf life include ultrahigh pressure (UHP), pulsed electric fields (PEF), ultraviolet (UV) light, electric pulse, and carbon dioxide (CO2) (Ohlsson and Bengtsson, 2002). Over the past several years, intense R&D efforts have aimed at validating and commercializing these technologies. The low-temperature storage is important to other nonthermally treated fruit products. Pulsed electric-fields’ method is a cold-pasteurization process for antimicrobial treatment of liquids and pumpable foods. It is based on the application of short-duration, high-intensity electric-field pulses to kill both spoilage and pathogenic organisms without affecting product taste or color. The use of high-intensity pulsed light to control microorganisms on food surfaces applies nonionizing, high-intensity flashes of broad-spectrum light to reduce microbial populations

3 . Processing of Fruits 57 on foods and packaging materials. Each flash is approximately 20,000 times the intensity of sunlight, with wavelengths ranging from ultraviolet to near infrared (Barbosa-Canovas et al., 1997). Radio frequency (RF) energy has been investigated as a nonthermal alternative to thermal pasteurization (Geveke et al., 2002). Electric-field strengths of 14–30 kV/cm generated with RF power supply systems at frequencies in the range of 20 kHz–27 MHz were applied to suspensions of Saccharomyces cerevisiae in water over a temperature range of 28–558C. The population of S. cerevisiae was reduced by >5 log following 30 exposures to a 100-kHz, 25-kV/cm field at 288C. 3.2. STERILIZATION OF FOOD BY HIGH PRESSURE The basis for applying high pressure to food is to compress the water surrounding the food. A decrease in volume of water with increasing pressure is very minimal compared to gases. The volume decreased for water is approximately 4% at 100 MPa, 7% at 200 MPa, and 11.5% at 400 MPa at 228C. Above 1,000 MPa and at room temperature, however, water changes to a solid (type VI ice), whose compressibility is very small. Usually irreversible effects on biological materials are observed at pressure >100 MPa. Therefore, pressure of 100 –1,000 MPa could be useful in food treatment. For reversible effects, pressure up to 200 MPa may be used. Microbial death at higher pressure is considered to be due to changes in permeability of cell membranes (Farr, 1990). Bacteria, yeasts, and molds in foods, such as meat, fish, and agricultural products, are sterilized by high-pressure treatment at 400–600 MPa. The pressur- ization of mandarin or orange juices at 300–400 MPa for 10 min is enough to sterilize vegeta- tive microorganic cells, although spores of Bacillus sp. are not killed. This retains good taste and flavor of the juice, and allows to store it at room temperature for 5 months. When pressure was applied at 458C, the results were considerably better than that at the room temperature. The major use of the high-pressure sterilization is in partially prepared foods or oven- ready foods. Pressure treatment preserves flavor, taste, and natural nutrients, but bacterial spores are not killed. Hence, these foods require chilled transportation. 3.2.1. High-Pressure Equipment and the System Test equipment for foods have been developed by several equipment industries and are available on the market. A typical equipment has 500-ml capacity, is made of stainless steel, and works at a maximum pressure of 700 MPa. It takes only 90 s to attain the maximum pressure. Temperature of the inside water, used as the pressure-transducing med- ium, is regulated by an electric heater outside the pressure vessel. Thus, the hydrostatic pressure is directly applied to foods placed in the pressure vessel at high speed under regulated temperature without any harmful contaminants. Several food companies and government institutions in Japan have been equipped with high-pressure test machines in recent years and are performing research and development of new food products based on the high-pressure processing (Barbosa-Canovas et al., 2004) Industrial equipment for high-pressure processing of foods are operational in several food industries: a batchwise system of 10–50 l capacity and a semicontinuous system of 1– 4 ton/h treatment. The former is used for the processing and sterilization of packed foods and the latter for the treatment of liquid foods. These machines are as small as an

58 Fruit Manufacturing industrial machine, but a pressure vessel of 50 l is similar to a heating vessel of 200 l in capacity. The cycle time for operating the pressure machine is short, generally being 15 min for food sterilization or food processing, while a large pan takes about 1 h for heating and cooling in conventional processing (Cheftel, 1995; Cole, 1997). An industrial system for the high-pressure processing of foods is similar to the conven- tional heat processing: Raw materials are pretreated, filled in plastic bags, sealed in vacuum, and pressurized. Final products are obtained after drying the bags. Liquid food may be placed directly in the pressure vessel in a semicontinuous way. It was also indicated that the use of pressure processing may save energy and improve sanitary conditions in the use of high- temperature processing. Selection of packaging materials is important for high-pressure food processing. While metal and glass are not suitable for high-temperature processing, plastic films are generally acceptable. Ochiai and Nakagawa (1991) pointed the importance of head space in plastic cups and suggested the use of plastic as a package material because of its heat sealability, hygiene, and safety. Packaging materials, which prevent oxygen permeability and light exposure, should be developed especially for retaining fresh color and flavor of foods. In brief, high pressure similar to high temperature is useful for the purpose of cooking, processing, sterilizing, and preserving food. The advantage of high pressure lies in the fact that it avoids destruction of the covalent bonds and retains natural flavor, taste, color, and nutrients. Thus, high-pressure technology is of great importance to the fruit industry. 3.3. CONCENTRATION BY EVAPORATION Evaporation refers to the process of heating the liquid to boiling point to remove water as vapor. Because fruit products, in particular fruit juices, are heat sensitive, heat damage can be minimized by evaporation under vacuum to reduce the boiling point (Heldman and Singh, 1981). The basic components of this process consist of: (i) heat exchanger, (ii) vacuum system, (iii) vapor separator, and (iv) condenser. The heat exchanger transfers heat from the heating medium, usually low-pressure steam, to the product via indirect contact surfaces. The vacuum system reduces the product tem- perature. The vapor separator removes juice from the vapors, driving juice back to the heat exchanger and vapors out to the condenser, which condenses the vapors from inside the heat exchanger and may act as the vacuum source. The driving force for heat transfer is the difference in temperature between steam and juice. The steam is produced in large boilers, generally tube and chest heat exchangers. The steam temperature is a function of the steam pressure. Water boils at 1008C at 1 atm., but at other pressures the boiling point changes. At boiling point, the steam condenses in the coils and gives out latent heat. If the steam temperature is too high, burn-on/fouling increases, so there are limits to how high steam temperatures can go. The juice is also at its boiling point, with an increase of solids’ concentration. The most important types of single effect evapor- ators are described by Minton (1986) and Perry and Chilton (1973). 3.3.1. Batch Pan It consists of spherical-shaped, steam-jacketed vessels (Fig. 3.1). The heat transfer per unit volume is small, requiring long residence times. The heating is due to natural convection. Heat transfer characteristics are poor.

3 . Processing of Fruits 59 Vapor Steam Condensate Concentrate Figure 3.1. Batch pan or calandria. 3.3.2. Rising Film Evaporator This type of evaporator consists of a heat exchanger isolated from the vapor separator (Fig. 3.2). The heat exchanger consists of 10 –15 m long tubes in a tube chest, which is heated with steam. The liquid rises by percolation from the vapors formed near the bottom of the heating tubes. The thin liquid film moves upward rapidly. The product may be recycled if necessary to arrive at the desired final concentration. Vaccum Feed Steam Feed Concentrate Condensate Figure 3.2. Rising film evaporator (single stage).

60 Fruit Manufacturing 3.3.3. Falling Film Evaporator The falling film evaporators are the most widely used type of evaporators in the food industry. It has similar components to the rising-film type except that the thin liquid film moves downward under gravity in the tubes (Fig. 3.3). Specially designed nozzles or spray distributors at the feed inlet permit it to handle products that are more viscous. The residence time is 20–30 s as opposed to 3– 4 min in the rising film type. The vapor separator is at the bottom, which decreases the product holdup during shutdown. The tubes are 8–12 m long and 30–50 mm in diameter. 3.3.4. Scraped-Surface Evaporator Scraped-surface evaporators are designed for the evaporation of highly viscous and sticky products, which cannot be otherwise evaporated. This type of evaporator has been specially designed to provide a high degree of agitation as well as scraping the walls of the evaporator to prevent deposition and subsequent charring of the product. Scraped-surface heat evapor- ators consist of a cylinder that has an inner tube (heat transfer surface area) and an outer tube. Between the two tubes, there is annular space, where the heating or cooling media flows countercurrent to the product. Inside the inner tube, a bladed shaft rotates and removes the product from the heat transfer wall areas (Fig. 3.4). Scraped-surface heaters improve cooking by allowing better heat transfer to the batch and preventing burn-on. Typical scraped film evaporator application includes processing of fruit pure´es, mashes, pulps, concentrates, and pastes. Feed Steam Vacuum Condensate Concentrate Figure 3.3. Falling film evaporator (single stage).

3 . Processing of Fruits 61 Steam Rotor Fruit product Figure 3.4. Cross section of a scraped-surface evaporator. 3.3.5. Multiple Effect Evaporator Two or more evaporator units can be run in sequence to produce a multiple effect evaporator. Each effect would consist of a heat transfer surface, a vapor separator, as well as a vacuum source and a condenser. The vapors from the preceding effect are used as the heat source in the following effect. There are two advantages to multiple effect evaporators: . Economy: They evaporate more water per kilogram steam by reusing vapors as heat sources in subsequent effects . Improvement in heat transfer: This is due to the viscous effects of the products as they become more concentrated. Each effect operates at a lower pressure and temperature than the effect preceding it so as to maintain a temperature difference and continue the evaporation procedure. The vapors are removed from the preceding effect at the boiling temperature of the product, so that no temperature difference would exist if the vacuum were not increased. The operating costs of evaporation are relative to the number of effects and the temperature at which they operate. As evaporation is a very energy-consuming process, the availability and the relative cost of energy determine the design of the evaporation plant. Normally an evaporation plant is designed to use energy as efficiently as possible by using more than one effect. Therefore the following technical solutions are used in order to keep the temperature of the steam high enough to run the process: . Thermal vapor recompression (TVR), . Mechanical vapor recompression (MVR), or . Combination of both. 3.3.5.1. Thermocompression (TC) This includes the use of a steam-jet booster to recompress part of the exit vapors from the first effect. Through recompression, the pressure and the temperature of the vapors are increased. As the vapors exit from the first effect, they are mixed with very high-pressure steam. The steam entering the first effect is at a slightly less pressure than the supply steam. There are

62 Fruit Manufacturing usually more vapors from the first effect than can be used by the second effect; usually only the first effect is coupled with multiple-effect evaporators. 3.3.5.2. Mechanical Vapor Recompression (MVR) Whereas only part of the vapor is recompressed using TC, the entire vapor is recompressed in an MVR evaporator. Vapor is mechanically compressed by radial compressors or by simple electrical fans. There are several variations: in single effect, all the vapors are recompressed, therefore no condensing water is needed; in multiple effect, MVR is possible on first effect, followed by two or more traditional effects, or recompress vapors from all effects. 3.4. DEHYDRATION Dehydration refers to the nearly complete removal of water from foods to a level of less than 5%. Dehydrated foods are protected from spoilage by lowering the water activity: aw ¼ pv=po (3:1) where pv is the vapor pressure of water in the product; and po the vapor pressure of saturated water at the same temperature. In dry fruits at aw ¼ 0:85 or above, some other form of preservation such as SO2 or potassium sorbate may be used. Some definitions are of interest. . Dried: Refers to all products with reduced moisture content regardless of the method. . Evaporated: Refers to use of sun and forced air driers to evaporate moisture to a fairly stable product. Sun drying in general will not reduce moisture below 15%. Many evaporated fruits will have up to 25% water level. These products have short storage life even when a high level of preservative is added. Evaporated fruits need refrigeration. . Dehydrated: Refers to fruits whose moisture has been reduced to 1–5% under carefully controlled conditions. Dehydrated fruits have more than 2 years’ storage life, particu- larly if stored in modified atmosphere or low-temperature conditions. Drying rate of fruits depends on particle size and mode of heat transfer. Sliced and diced products dry faster. None of the mechanisms involved in drying of fruits are as simple as they might seem, as foods do not usually deal with a single uniform phase. Two important aspects of mass transfer in dehydration are: (a) movement of water to the surface of material being dried, and (b) removal of water from the surface through the (probably) thin immobile boundary layer. Considering some possible modes of heat transfer that can be applied in fruit drying we find: Phases involved in fruit drying Examples Gas–solid Conventional dehydration of fruits Gas–liquid Concentration of syrup in cascade-type drier Liquid–gas Internally during moist fruits’ drying Liquid–solid Frying Solid–gas Internally during moist fruits’ drying Solid–liquid Drum drying Solid–solid Drum drying of fruit pure´e

3 . Processing of Fruits 63 Constant rate period Drying Rate Falling rate period Xe TIME, t Xcr Figure 3.5. Characteristic drying rate curve. Drying curve relates to the amount of water removal with time (Fig. 3.5) and can be divided into different regions (Crapiste and Rotstein, 1997): (1) An initial period in which evaporation occurs on the surface and temperature wet bulb value. This is the constant-rate period. However, due to water conditions on product surface and shrinkage during drying a pseudoconstant-rate period may be observed. (2) At a point usually called the critical moisture content (Xcr) the falling rate period starts. When surface moisture is lost, the rate falls as water must diffuse from inside to surface before evaporation can take place. Dominant factor is availability of water at evaporation surface. The process of dehydration slows greatly at this period, until moisture content asymptotically reaches Xe, the equilibrium value at the relative humidity and temperature of the air. Drying rate is affected by drier loading on tray or belt driers. This can be overcome by fluidizing fruit particles. During the initial constant-rate stage hotter is better until falling rate phase is reached. At Xcr, temperature must be reduced to avoid deteriorative reactions. Table 3.2 lists different dehydration systems’ characteristics. Other driers used in the fruit industry are bin driers, simple devices consisting of bins with perforated bottoms, and fluidized bed driers, used for drying food powders. In fluidized bed driers, air is blown up through a wire mesh belt on porous plate that supports and conveys the product. A slight vibration motion is imparted to the food particles. Fluidization occurs when the air velocity is increased to the point where it just exceeds the velocity of free fall of the particles. The fluidization provides intimate contact of each particle with the air. With products that are particularly difficult to fluidize, a vibrating motion of the drier itself is used to aid fluidization—this is called vibrofluidizer. The fluidized solid particles then behave in a manner analogous to a liquid, i.e., they can be conveyed. Air velocities will vary with particle size and density but are in the range of 0.3–0.75 m/s. They can be used not only for drying but also for cooling. If the velocity is too high, the particles will be carried away in the gas stream, therefore, gravitational forces need to be only slightly exceeded. Minimum air velocity to fluidize 10-mm pulp is about 115 m/min. It can be used to dry uniform-sized products.

64 Fruit Manufacturing Table 3.2. Description and schematic diagram of different driers. Type of drier Description Drier sketch Sun or solar Simplest systems, consisting of trays laid flat on ground or supported slightly above it. Used for drying apricots, raisins, etc. Stacked tray system for very low humidity Wind areas improves dehydration due to natural air movement Cabinet It is a typical batch operation in which air is heated drier and forced to circulate between trays or through the product by using perforated trays. These driers can process from pulps to solid pieces of fruits Hot air Tunnel Tunnel driers have been the most widely used form of Exhaust drier fruit dehydrators. They are set up with parallel flow air air during the constant-rate period, arranged in such Hot Continuous a way that incoming fruit encounters the hottest, air belt or driest air, then counterflows so that the outgoing Trucks’ direction conveyor product encounters the driest air. Air velocities of Hot air driers 200 – 400 m/min are commonly used. Initial temperatures of 1008C may be used. Final Feeder Belt driers temperatures are about 708C or less, depending on fruit products Product moving through these driers is exposed to the same successive sets of drying conditions. A single continuous belt or series of small belts may be used. Conveyor may consist of mesh belts or perforated metal plates. Temperature is reduced in the direction of the outlet of the system Highly efficient device units, occupying a relatively small plant area per ton of product. In these driers high-velocity air (not enough to fluidize the product) passes through the belt. Adequate for dehydro freezing lines. Drying fruits to low moisture is usually attained by a complementary system Drum driers Originally used for dry milk. Can be operated at Air temperature decrease atmospheric pressure or under vacuum. In the fruit downstream industry it is used for making apple sauce flakes. Almost any pure´e can be dried if the fiber content Feeder is adequate. Control factors include sheet thickness, temperature, drum speed, and air flow over drums Source: Karel et al., 1975; Crapiste and Rotstein, 1997.

3 . Processing of Fruits 65 Microwave drying (Van Arsdel et al., 1973), osmotic dehydration (Shi and Fito, 1993), explosion puffing (Saca and Lozano, 1992), and freeze drying (Mujumdar and Menon, 1995) have been also applied for fruit dehydration. 3.4.1. Spray Drying Spray driers are one of the most widely used types of air convection drier. Spray drying involves transforming a pumpable food, i.e., juices, low-viscosity pastes, and pure´es into a dry-powdered or particle form. This is achieved by atomizing the fluid into a drying chamber, where the liquid droplets are passed through a hot-air stream (Heldman and Singh, 1981; Green and Maloney, 1999). The objective is to produce a spray of high surface-to-mass ratio droplets and then to uniformly and quickly evaporate the water. Evaporation keeps product temperature to a minimum, so little high-temperature deterioration occurs. In its simplest form, spray drying consists of four separate process stages: . Atomization of the liquid food feed, . Spray-air contact, . Drying, . Separation of the dried food product from the drying air. Atomization is generally accomplished by: (i) a single-fluid (or pressure) nozzle, (ii) a two-fluid nozzle, or (iii) a rotary atomizer, also known as a spinning disk or a wheel. The single-fluid nozzle allows more versatility in terms of positioning with the spray chamber, so the spray angle and spray direction can be varied. A typical drying-chamber design used for fruit pulps and juices is the cylindrical flat- bottomed drier. A pneumatic powder discharger removes the product, while an air broom cools chamber walls to prevent sticking. This design allows easier access for cleaning. Drying occurs in two phases, and air-temperature control is vital to their control. The first phase is a constant-rate step, in which the moisture rapidly evaporates from the surface, and capillary action takes out the moisture from within the particle. Then, during the ‘‘falling-rate’’ period, diffusion of water to the surface controls the drying rate. As moisture content drops, diffusion rate also decreases. Removing moisture to required values in a single-stage drier is responsible for most of the residence time in the drier. As a rule, the residence time of the air and the particle in a single-stage cocurrent drier is about the same. Since the moisture level is still decreasing toward the end of the process, the outlet temperature must be high enough to continue the drying process. This can be avoided by adding a fluid bed after the drier. The final stage of spray drying is the removing of the dried product from the air. Depending on drier design, the dried product can be separated at the base (as in a flat- bottomed drier). While the heavier product is removed by gravity, the smallest particles are pulled together in some type of collection equipment. Otherwise, the entire product and air can be moved to equipment designed to separate particles from air. Fine particles are removed with cyclones, bag filters, electrostatic precipitators, or scrubbers. Fines are bagged or returned to an agglomeration process; air is returned to the system. Drying takes place within a matter of seconds at temperatures approximately 2008C. Although evaporative cooling maintains low product temperatures, rapid removal of the product is still necessary.

66 Fruit Manufacturing Air blower Feed Nozzle Air outlet Dry product Figure 3.6. Typical spray drier configuration. The liquid food is generally preconcentrated by evaporation. The concentrate is then introduced as a fine spray into a tower or chamber with heated air. As the small droplets are put in contact with the heated air, they flash off their moisture and drop to the bottom of the tower and are removed. Principal spray components are: (1) a high-pressure pump, for introducing liquid into the tower, (2) a nozzle for atomizing the feed stream, (3) a blower with a source of hot air, and (4) a system for removing the dried food (Fig. 3.6). Several billion particles per liter ensure a large surface area for exposure to drying forces. Particle size must be reduced in three ways: (1) using a smaller orifice, (2) increasing atomization pressure, or (3) reducing product viscosity, by increasing feed temperature or redilution. The exit air temperature is an important control parameter, which can be used to adjust feed flow rate and inlet temperature. 3.4.2. Powder Recovery Three systems are available (Walas, 1976; Green and Maloney, 1999) for powder recovery from the air stream: (1) Bag filters: Although very efficient they are not very popular due to labor costs and sanitation problems. They are not recommended for hygroscopic particles. (2) Cyclone collector: In this type of powder collector, air enters at tangent at high velocity into a cylinder or cone, which has a much larger cross section. Air velocity is decreased in the cone, permitting settling of solids by gravity. Several cyclones can be placed in series. High air velocity is needed to separate small diameter and light materials, while centrifugal force is important in removing particles from the air stream. To increase centrifugal force cyclone diameter may be reduced. A rotary airlock is used to remove powder from the cyclone. (3) Wet scrubber: Wet scrubbers are the most economical outlet air cleaner. The principle of a wet scrubber is to dissolve any dust powder left in the airstream into either water or the feed stream by spraying the wash stream through the air. Wet scrubbers also recover approximately 90% of the potential drying energy normally lost in exit air. Cyclone separators are hygienic and easy to operate. However, high losses may occur. Either feed stream or water can be used as scrubbing liquor.

3 . Processing of Fruits 67 Two- and Three-Stage Drying Processes In single-stage spray drying, the rate of evaporation is particularly high in the first part of the process, and it gradually decreases because of the falling water content of the particle surfaces. Therefore a relatively high outlet temperature is required during the final drying phase. The two-stage drying process was introduced to reduce temperatures and cost of production, and increase product quality. The two-stage drier consists of a spray drier with an external vibrating fluid bed placed below the drying chamber. The product can be removed from the drying chamber with a higher moisture content, and the final drying takes place in the external fluid bed where the residence time of the product is longer and the temperature of the drying air is lower than in the spray drier. This principle forms the basis of the development of the three-stage drier. The second stage is a fluid bed built into the cone of the spray drying chamber. This fluid bed is called the integrated fluid bed. The inlet air temperature can be raised, resulting in improved efficiency in the drying process. The exhaust heat from the chamber is used to preheat the feed stream. The third stage is again the external fluid bed, for final drying and/or cooling the powder. As a result, a higher quality powder with much better rehydrating properties is obtained. More- over, lower energy consumption and smaller space requirements are obtained. 3.5. MISCELLANEOUS PROCESSING 3.5.1. Size Enlargement Size enlargement operations are used in the food processing industry for improving handling and flowability, producing structural useful forms, enhancing appearance, etc. Food size enlargement operations are known as compaction, granulation, tabletting (palletizing), en- capsulation, sintering, and agglomeration. The main objective of agglomeration is to control porosity and density of materials in order to manipulate properties like dispersibility and solubility, known as instantizing, because rehydration is an important functional property in food processes. If size enlargement is used with the objective of obtaining definite shapes, extrusion is the selected process to shape and cook at the same time. 3.5.1.1. Instantizing Powdered fruit juice with particle size less than about 10 mm is considered. Powdered fruit juice is the product obtained from fruit juice of one or more kinds by the physical removal of virtually all the water content. The resulting product will be in the powder form and will require the addition of water before use. This product tends to form lumps when dissolved in water and require strong mechan- ical stirring for obtaining a homogeneous dispersion. It was proposed that under those conditions water penetrates into the narrow spaces between the particles by capillarity, and the powder starts to dissolve, forming a thick, gel-like mass, which resists further penetration of water. Therefore, fruit particle agglomeration, with a dried core is formed. Moreover, if enough air is locked into these lumps, they will float on the water surface, resisting further dispersion. To avoid this problem, agglomerated powder needs an open structure, allowing water to penetrate before a tightly packed gel layer is formed. To do this the specific surface of the powder has to be reduced and the liquid needs to penetrate more evenly around the particles. In this way the powder can disperse into the bulk of the liquid, and the following steps represent a complete dissolution (Ortega-Rivas, 2005):

68 Fruit Manufacturing (1) Granular juice particles are wetted, and water penetrates into the pores of the granule structure. (2) The wetted particles sink into the water and granules disintegrate into their original smallest particles. (3) The small, dispersed particles dissolve in the water. The total time required for all these steps should be the criteria used to evaluate a product’s instant properties. 3.5.1.2. Agglomeration Agglomeration can be defined as the process of size enlargement by which particles are joined or bind each other in a random way, finishing with an aggregate of porous structure much larger in size than the original material. Agglomeration is used in food processes mainly to improve properties related to handling and reconstitution, that is to produce a more uniform particle size, increase or decrease particle bulk density, improve solubility and dispersibility, and reduce caking. Dispersibility is defined as the time to dissolve a given weight of material in water. The term agglomeration includes varied unit operations and processing techniques aimed at agglomerating particles (Green and Maloney, 1999). The main attractive forces involved in the agglomeration of food particles are studied by Rumpf (1962) and Rumpf and Schubert (1978): (1) Solid bridges, liquid bridges, and capillary forces. The force exerted by a liquid bridge at rest depends on both the surface tension of the interface and the capillary effects due to the curvature of the bridge. Solvents’ evaporation and treatment of food particles at elevated temperatures form solid bridges, by formation of salt bridges, and partial sintering or melting at the intra-agglomerate contact points. Alternatively chemical bonding may occur with the use of organic binders, like hydrocolloid for hydrophilic materials. If the material has some hydrophobicity, the binder may have to have a wetting agent, e.g., lecithin. Figure 3.7 shows a simplified agglomeration by liquid bridges’ development. After drying soluble solids crystallize out of the liquid bridge, forming solid bridges. (2) Van der Waals’ Forces. These are commonly known as dispersion forces, and are quantum mechanical in origin. These forces of attraction exist between molecules of any kind and constitute a general property of matter. Such attractions will cause particles to stick to each other when they come within a few nm of each other. While

3 . Processing of Fruits 69 Wetting Collision Liquid bridge formation Figure 3.7. Agglomeration by liquid bridge formation. van der Waals’ attractions can be strong at short distances (<10 nm), the attraction becomes negligible for particles that are far apart. (3) Electrostatic forces. These tend to be the weakest force between particles. Particles in close proximity will tend to be held together by the difference in electrostatic potential, U. To enable calculation of the strength of the agglomerate and the manner of the bonding, the number of bonds and the mechanics of the breakup need to be considered (Rumpf, 1962). Figure 3.8 summarizes the mechanisms usually involved during food particles’ agglom- eration, like partial melting, liquid bridges, molecular, interlocking bonds, and electrostatic and capillary forces (Pietsch, 1991). The agglomeration process causes an increase in the amount of air incorporated between powder particles. More incorporated air is replaced with more water when the powder is reconstituted, which immediately wets the powder particles. Figure 3.8 represents in fact a three-dimensional structure containing a large number of particles. Particles in agglomerates could be quite numerous. The points of interaction can be characterized by contact, or by a distance small enough for the development of binder bridges. Alternatively, sufficiently high attraction forces can be caused by one of the short- range force fields. The number of all interaction sites of one particle within the agglomerate Partial Capillary melting force Interlocking +− bonds +− +− Van der Waal and electrostatic forces Liquid bridge Figure 3.8. Mechanisms usually involved during food particles’ agglomeration.

70 Fruit Manufacturing structure is called the coordination number (Green and Maloney, 1999; Pietsch, 1983; Ortega- Rivas, 2005). Indirect measurement of the coordination number can be made as a function of other properties of the agglomerate, namely porosity. 3.5.1.3. Agglomeration Process and Equipment Two particles can be made to agglomerate if they are brought into contact. Food particles are then brought into a sticky state, by wetting with the application of a finely dispersed liquid or steam, by heating (thermoplastic materials), or by the addition of binder media (a food adhesive). Use of binder agents is considered a nonagglomeration method. The steam conden- sation method usually cannot provide enough wetting without adversely heating the material and is used less frequently on newer systems. The food particles’ surface must be uniformly wetted and held wet over a selected period of time to give moisture stability to the clusters formed. The clusters are dried to the desired moisture content and then cooled. Dried clusters are screened and sized to reduce excessively large particles and remove excessively small ones. Subsequently, the particles are placed under such conditions where they can form structures. Successful formation of stable agglomerate structures depends on product solu- bility and surface tension, as well as on the conditions that can be generated in the process equipment. For most products, combinations of moisture and temperature can be established. Generally agglomeration temperature decreases when particle moisture increases. 3.5.1.4. Agglomeration Equipment Agglomeration is a complicated process. However, it is possible to agglomerate powder foods by means of comparatively simple equipment (Pietsch, 1991), which involves the use of a fluidized bed for rewetting and particle contact phase, followed by a belt or a fluid bed for moisture removal (Fig. 3.9). The process must be strictly controlled to avoid deposit formation in the chamber by overwetting, and weak agglomerates due to insufficient liquid or vapor rate. Water/vapor Air outlet Powder in Hot air Agglomerated product Figure 3.9. Typical agglomeration system.

3 . Processing of Fruits 71 Figure 3.10. Gear pelletizer. Deposit formation will always be a concern in agglomeration equipment. Proper implementa- tion of a fluid bed agglomeration system requires detailed knowledge of the fluidization technology, including fluidization velocities, bed heights, air-flow patterns, and residence time distribution. These processes have allowed the manufacturing of powders with better reconstitution properties, such as fruit juice powder. During agglomeration, the powder is wetted, uniformly but not excessively, with water or steam. Other agglomeration methods are compaction, extrusion, melt forming, mixing, tum- bling, and sintering. During pressure agglomeration particles with only slight amounts of moisture are formed into tablets, and briquettes into stamp presses, tablet presses, and roller presses. The principal binding force is van der Waals’ attraction. Figure 3.10 shows a typical gear pelletizer. 3.5.1.5. Selective Agglomeration (Spherical Agglomeration) In the latest agglomeration process, a second immiscible phase is added to the suspension. This wets the solid phase and binds the particles together by means of capillary forces. As a result, rounded flocks or agglomerates form with diameters up to 5 mm. Selective agglomer- ation can be achieved for mixtures of solids. REFERENCES Barbosa-Canovas, G.V., Palou, E., Pothakamury, U.R. and Swanson, B.G. (1997). Application of light pulses in the sterilization of foods and packaging materials. Nonthermal Preservation of Foods. Marcel Dekker, New York, Chapter 6, pp. 139–161. Barbosa-Ca´novas, G.V., Maria Tapia and Pilar Cano, M. (eds.) (2004). Novel Food Processing Technologies. CRC Press, Boca Raton, FL.

72 Fruit Manufacturing Bolando-Rodr´ıguez, S., Go´ ngora-Nieto, M.M., Pothakamury, G.V., Barbosa-Ca´novas, G. and Swanson, B.G. (2000). A review of nonthermal technology. In Trends in Food Engineering. Aspen Publishers, Inc., Maryland, EEUU, pp. 117–134. Cheftel, J.C. (1995). High-pressure, microbial inactivation and food preservation. Food Sci. Technol. Int. 1: 75–90. Cole, R. (1997). High pressure processing: a technology of the future. Food Manuf. 72: 21–26. Crapiste, G.H. and Rotstein, E. (1997). Design and performance evaluation of driers. In Handbook of Food Engineering Practice, Valentas, K.J., Rotstein, E. and Singh, R.P. (eds.). CRC Press, Boca Raton, Chapter 4, pp. 125–165. Farr, D. (1990). High pressure technology in the food industry. Trends Food Sci. Technol. 1: 14–16. Geveke, D., Kozempel, M., Scullen, O.J. and Brunkhorst, C. (2002). Radio frequency energy effects on microorgan- isms in foods. Innov. Food Sci. Emerg. Technol. 3: 133–138. Green, D.W. and Maloney, J.O. (1999). Perry’s Chemical Engineers’ Handbook. McGraw-Hill, New York. Heldman, D.R. and Singh, R.P. (1981). Food Processing Engineering, 2nd ed. AVI Publishing Company, Inc., Westport, USA. Karel, M., Fennema, O.R. and Lund, D.B. (1975). Principles of Food Science. Part. 2. Physical Principles of Food Dehydration. O.R. Fennema Ed. M. Dekker Inc. NY. Minton, P.E. (1986). Evaporator types and applications. In Handbook of Evaporation Technology. William Andrew Publishing/Noyes. Mujumdar, A.S. and Menon, A.S. (1995). Drying of solids: principles, classification and selection of driers. In Handbook of Industrial Drying, Mujumdar, A.S. (ed.). Marcel Dekker, Inc., New York, Chapter 1, pp. 1–40. Nickerson and Sinskey (1972). Microbiology of Foods and Food Processing. American Elsevier Publishing Company, NY. Ochiai S, Nakagawa, Y. (1991). High Pressure Science for Food. Hayashi Edition, Kyoto, Japan. Ohlsson, T. and Bengtsson, N. (2002). Minimal processing of foods with non-thermal methods. In Minimal Process- ing Technologies in the Food Industry, Thomas Ohlsson and Nils Bengtsson (eds.). CHIPS, Texas, USA. Ortega-Rivas, E. (2005). Handling and processing of food powders and particulates. In Encapsulated and Powdered Foods, Onwulata, C.I. and Konstance, R.P. (eds.). Marcel Dekker, New York, in press. Perry, R.H. and Chilton, C.H. (1973). Chemical Engineers’ Handbook, 5th ed. McGraw-Hill Book Company, New York, pp. 11–27. Pietsch, W. (1983). Low-energy production of granular NPK fertilizers by compaction-granulation. Proceedings of Fertilizer’83. British Sulphur Corp., London, UK, pp. 467–479. Pietsch, W. (1991). Size enlargement by agglomeration. John Wiley & Sons Ltd., Chichester, England. Rumpf, H. (1962).The strength of granules and agglomerates. In Agglomeration, Knepper, W.A. (ed.). Interscience, New York, pp. 379–418. Rumpf, H., Schubert, H. (1978). Adhesion forces in agglomeration processes. In Onada & Hench: Ceramic processing before firing. J. Wiley and Sons, Inc., London. Saca, A. and Lozano, J.E. (1992). Explosion puffing of bananas. Int. J. Food Sci. Technol. 27: 419–423. Shi, X.Q. and Fito, P. (1993). Vacuum osmotic dehydration of fruits. Drying Technol. 11: 1429–1442. US Department of Health and Human Services (2004). Juice HACCP Hazards and Controls Guidance. Guidance for Industry, 1st ed. Food and Drug Administration, Center for Food Safety and Applied Nutrition (CFSAN). Van Arsdel, W.B., Copley, M.J. and Morgan Jr., A.I. (1973). Food Dehydration, Vols. 1 and 2. AVI Publishing Company, Inc., Westport, CN. Walas, S.M. (1976) Spray driers. In Encyclopedia of Chemical Processing and Design. Vol. 53. J.J. McKetta Ed. Marcel Dekker Inc. NY. pp. 22–44.

CHAPTER 4 THERMODYNAMICAL, THERMOPHYSICAL, AND RHEOLOGICAL PROPERTIES OF FRUITS AND FRUIT PRODUCTS 4.1. INTRODUCTION Most processed and many freshly consumed fruits receive some type of heating or cooling during handling or manufacturing. Design and operation of processes involving heat transfer needs special attention due to heat sensitivity of fruits. Both theoretical and empirical relationships used when designing, or operating, heat processes need knowledge of the thermal properties of the foods under consideration. Food thermal properties can be defined as those properties controlling the transfer of heat in a specified food. These properties are usually classified (Perry and Green, 1973) into thermodynamical properties, viz, specific volume, specific heat, and enthalpy; and heat transport properties, namely, thermal conduct- ivity and thermal diffusivity. When considering the heating or cooling of foods, some other physical properties must be considered because of their intrinsic relationship with the ‘‘pure’’ thermal properties mentioned, such as density and viscosity. Therefore, a group of thermal and related proper- ties, known as thermophysical properties, provide a powerful tool for design and prediction of heat transfer operation during handling, processing, canning, and distribution of foods (Fig. 4.1). Abundant information on thermophysical properties of food (Polley et al., 1980; Wallapapan et al., 1983; Choi and Okos, 1986; Rahman, 1995) is available to the design engineer. However, finding relevant data is usually the controlling step in the design of a given food operation, and the best solution may be the experimental determination. This chapter provides data and information for thermal process calculation for fruits and fruit products, including a brief description of more commonly used methods for measurement and determination of thermophysical properties. 4.2. THERMOPHYSICAL PROPERTIES’ IDENTIFICATION Thermophysical properties include different types of parameters associated to the heat transfer operations present during fruit processing. It is well known that heat can be trans- ferred by three ways: radiation, conduction, and convection. Radiation is the transfer of heat by electromagnetic waves. The range of wavelength 0.8–400 mm is known as thermal radiation, since this infrared radiation is most readily 73

74 Fruit Manufacturing Thermodynamical Specific volume n (m3/kg) properties Specific heat cp (kJ/kg−1/ Њ C−1) Enthalpy ∆H (kJ/kg−1) “Pure” Thermal conductivity k (W/m−1/K−1) Thermophysical properties Heat transport properties Thermal diffusivity a (m2/s−1) Physical properties Density r (kg/m−3) Porosity e Viscosity m (Pa s) Figure 4.1. Thermophysical properties associated to fruit processing. absorbed and converted to heat energy. A body emitting or absorbing the maximum possible amount of radiant energy is known as a ‘‘black body.’’ Energy emitted by a black body is given by the Stefan–Boltzmann law: Q ¼ sAT 4 (4:1) where s is the Stefan–Boltzmann constant; A the area of transfer, and T the absolute temperature. For no ‘‘perfect’’ black bodies, as real bodies are, Eq. (4.1) is corrected by as factor known a emissivity («): Q ¼ s«AT4 (4:2) Emissivity values of foods are in the range 0.5–0.97 (Karel et al., 1975). Conduction is the movement of heat by direct transfer of molecular energy within solids (for example, heating of a fruit pulp by direct fire through metal containers). Convection is the transfer of heat by groups of molecules that move as a result of a gradient of density or agitation (for example, the stirring of tomato pure´e). Heat transfer may take place: (i) in steady-state way by keeping constant the temperature difference between two materials or (ii) under unsteady-state way when the temperature is constantly changing. Calculation of heat transfer under these conditions is extremely com- plicated but is simplified by making a number of assumptions or giving approximate solutions from prepared graphic or tabulated information. Table 4.1 shows some common simplified equations used for the calculation of heat transfer. Most of the thermophysical properties are required to solve the heat transfer equations by conduction and convection. During processing, the temperature within a fruit, changes continuously depending on the temperature of the heating medium, and two prop- erties of the fruit: the thermal conductivity (k) and the specific heat Cp. On the other hand thermal diffusivity is related to k and Cp: a ¼ k=rcp (4:3)

4 . Thermodynamical, Thermophysical, and Rheological Properties 75 Table 4.1. Simplified equations used for heat conduction and convection calculations (Lozano 2005, with permission). Conduction (unidirectional ) 8 Q ¼ kA(u1 À u2)=x Convection < Steady state: : 8 du=dt ¼ ad2u=dx2 8 Unsteady state: < Q ¼ hsA(DT) >>< Natural : Gr ¼ r2gbl3DT=m2 >:> Forced 8 Nu ¼ K(GrPr) < Nu ¼ f (Re, Pr) ¼ hl=k : Pr ¼ cm=k Re ¼ lnr=m Reproduced with permission from Thermal Properties of Foods. In ‘‘FOOD ENGINEERING 1. Engineering Properties of Foods.’’ Edited by Gustaro V. Barbosa-Ca´novas. Paris UNESCO Publishing. ISBN 92-3-103999-7, pp.45-64. All of which is part of Encyclo- pedia of Life Support Systems (EOLSS). http://www.eols.net (copyright) EOLLS Publishers Co.Ltd. Gr, Nu, Pr, and Re are the Grashoff, Nusselt, Prand, and Reynolds’ numbers, respectively. When a solid piece of food is heated or cooled by a fluid, the resistance to heat transfer, which are the surface heat transfer coefficient and k may be related as follows: Bi ¼ hl=k (4:4) where h (W=mÀ2=KÀ1) is the heat transfer coefficient, f the characteristic half-dimension, and k (W=mÀ1=KÀ1) the thermal conductivity. At small Bi (<0:2) the surface film is the predominant resistance, while for Bi > 0:2 it is the thermal conductivity, which limits the rate of heat transfer (Urbicain and Lozano, 1997). 4.3. FRUITS AND FRUIT PRODUCTS’ PROPERTIES 4.3.1. Fruit and Fruit Products’ Properties During Freezing It must be considered that thermophysical properties of foods change dramatically during the freezing process. One of the characteristics of food freezing is that temperature changes gradually with the phase change, which implies that the fraction of water frozen always changes continuously with temperature below the freezing point. Depression of the initial freezing temperature can be predicted from (Heldman and Singh, 1981): ! l 1 À1 R TAo TA ¼ ln XA (4:5) where l is the molal latent heat of fusion, R the universal gas constant, TAo the freezing temperature of water, TA the freezing temperature of food product, and XA the mole fraction of water. In most cases Eq. (4.5) is used to predict the unfrozen water fraction with the initial freezing temperature, through the determination of apparent molecular weights for the mass fraction of total solids in the product. 4.3.2. Water Content Though water content (Xw) is not a thermophysical property, it significantly influences all thermophysical properties (Lozano 2005). As food is a living commodity, its water content changes with maturity, cultivar, growth, and harvest and storage conditions. Values of most of the thermophysical properties can be calculated directly from the water content. Xw is usually expressed as water mass fraction [kg water per kg food; wet basis].

76 Fruit Manufacturing 4.4. EXPERIMENTAL DATA AND PREDICTION MODELS Fruits and fruit products show extended variability in composition and structure, which must be kept in consideration when modeling their thermal properties. Fruits are generally non- homogeneous, varying in composition and structure not only between products but also within a single product. Some thermophysical properties for fruits were modeled only as a function of the water content (Alvarado, 1991; Gupta, 1990). However, the presence of proteins, fats, and carbo- hydrates, as major components besides water, differs from one fruit to another. These compounds have variable effects on the properties of the complex fruit structure (Sweat, 1995). As a result some proposed thermophysical properties models applied to fruit and fruit products include a combination of the properties of water, fats, proteins, carbohydrates, and/ or ash (Oguntunde and Akintoye, 1991; Rahman, 1995). Literature provides a large volume of experimental thermophysical food properties data (Dickerson, 1968; Mohsenin, 1980; Jowitt et al., 1983; Rahman, 1995; Urbicain and Lozano, 1997). However, as the amount of thermal properties data required for describing any foodstuff under the varied handling, processing, and storage condition is practically infinite, modeling and prediction of such properties is a must. Thermophysical properties of any material control the thermal energy transport and/or storage within it, as well as the transformations undergone by the material under the action of heat. Fruits are no exception; they are dependent on the temperature and the material’s chemical composition, and physical structure. Since fruits are complex materials, relevant information on their properties is given as average or effective value. For this reason the generation of predictive models requires a physical representation of the material under study. Fruits show three different levels of complexity. First, microscopically, fruits may look as a continuous and homogeneous single phase. However, they are composed of different chemical compounds including proteins, carbohyd- rates, fats, fiber, water, and other minor components. For this reason models proposed for the prediction of a given property must consider the individual contribution of such com- pounds. It may be done by ‘‘weighing factors’’ accounting for the proportion in which they are present. In the second place, some fruits or fruit products can be considered as a solid matrix of the continuous described above and a disperse phase of air or water, respectively. This description corresponds typically to porous fruits and fruit powders. In this case both the volumetric fraction and the spatial distribution of each phase are to be considered, which is done by means of distribution factors adequately described. Finally, a third level of complexity is achieved when different food materials, including fruits, are processed together to give composite food. This group includes all kinds of canned and packed foods, pastries, confectioneries, and a wide variety of prepared foods. Once more, modeling requires the information of the mean or effective values of the components together with the representation of the physical structure. As previously mentioned, the value of the thermophysical property will be a function of the temperature, through the dependence of the components, and porosity or water content, for porous or composite foodstuffs. Since water can be either liquid or solid, particular attention is paid to frozen fruit products. Available information may be contradictory, due to the different conditions at which thermophysical properties were gathered, as well as to the differences among fruits of different origin, composition, and structure.

4 . Thermodynamical, Thermophysical, and Rheological Properties 77 4.4.1. Density* Density (r) is the unit mass per unit volume. SI unit for density is [kg=m3]. In particular, when the fruit or fruit product is a porous solid, density plays an important role in heat transfers intrinsically or through the definition of porosity. A few definitions are necessary (Lozano, 2005): . Substance density: rs (or true density), is the density measured when the substance has been broken, milled, or mashed to ensure that no pores remain. . Particle density: rp, is the density of a sample that has not been structurally modified. In the case of pores not externally connected to the surrounding atmosphere, particle density will include these close pores. . Bulk density: rb (or apparent density), is the density measured so as to include the volume of the solid and liquid materials, and all pores, closed or open to the surrounding atmosphere. As other authors (Maroulis and Saravacos, 1990; Farkas and Singh, 1991) have used different terms for the same condition, it is recommended to verify the definition of density before using density data. 4.4.1.1. Porosity* Porosity indicates the volume fraction of air (or void space). On the basis of the given densities, the following definitions of porosity have been proposed (Lozano et al., 1980; Lozano, 2005): . Total porosity («t) is the ratio of air space volume to total volume: (4:6) «t ¼ (rs À rb)=rs . Open pore porosity («a) is the ratio of the volume of pores connected to the outside to the total volume: «a ¼ (rp À rb)=rp (4:7) As in the case of density, it is recommended to verify definitions before using porosity data. 4.4.1.2. Density Measurement Methods Techniques developed for density measurement are basically methods for the measurement of volume, weight being easily measured with different types of precision balances. The principal measurement techniques applied for volume (and density) determination in fruit and fruit products are: . Hydrometric method: The bulk density is calculated from the apparent weight of the sample and the buoyant force E. rb ¼ rliq(Wair=E) (4:8) Fruit tissue must be coated to avoid mass loss by dilution or pore inundation (Lozano et al., 1980). * Reproduced with permission from Thermal Properties of Foods. In ‘‘FOOD ENGINEERING 1. Engineering Properties of Foods’’. Edited by Gustaro V. Barbosa-Ca´novas. Paris UNESCO Publishing. ISBN 92-3-103999- 7pp.45-64. All of which is part of Encyclopedia of Life Support Systems (EOLSS). http://www.eols.net (copyright) EOLLS Publishers Co.Ltd.

78 Fruit Manufacturing . Geometric method: The volume is calculated from the dimensions (L) of the food sample. rb ¼ L3=W (4:9) It is not suitable for soft and irregular solid foods. . Pycnometry: Liquid pycnometry. The pycnometer is a calibrated flask that allows the weighing of an exactly known volume of liquid, which in turn gives the density. The weight is determined by difference with the empty flask. To set the experimental temperature, the pycnometer is immersed in a constant temperature bath and filled with the sample. Gas pycnometry. Using perfect gas law it is possible to determine the volume of open pores (Vpore) in food by determining the gas volume in a chamber (Vch) with and without the sample: PiVi ¼ mRT (4:10) Vpore ¼ Vch(P1 À P2)=P2 P1 and P2 are pressure in empty space and in sample chamber,respectively (Mohsenin, 1980; Lozano et al., 1980). Table 4.2 lists bulk density values of two fruits (apple and pear) and fruit tissue components. These values cover a relatively wide range of density, illustrating the influence of water and air (porosity) in this property. 4.4.1.3. Empirical Equations and Theoretical Density Models Some empirical equations for the calculation of bulk density of selected fruits, in terms of temperature and water content, are listed in Table 4.3. Constenla et al. (1989) proposed a theoretical approach by considering the thermo- dynamic expression for the specific volume of a multicomponent solution (reciprocal of density) in terms of partial specific volumes: X (4:11) V ¼ 1=r ¼ wivi where wi and vi; represent the mass fraction and the partial specific volume of the i-component in solution, respectively. Sugars, organic acids, and different macromolecules interact with a substantial number of water molecules, resulting in a nonideal solution behavior. Therefore, specific volume is not necessarily equal to the specific volume of the pure component. Table 4.2. Bulk density of selected fruits and fruit products or components. Fruit/component Temperature/range (8C) Bulk density (kg=mÀ3) Reference Apple (GS; Xw ¼ 0:86) 25 837 Lozano et al. (1980) Pear 25 990 Lozano et al. (1983) Cellulose — 1,550 Kirk–Othmer (1964) Fat — 900/950 Lewis (1987) Glucose (solid) — 1,560 Kirk–Othmer (1964) Protein — 1,400 Lewis (1987) Water 4 1,000 Perry and Green (1973)

4 . Thermodynamical, Thermophysical, and Rheological Properties 79 Table 4.3. Empirical equations to calculate bulk density of selected fruit products or fruit components. Application range Parameter Xw (kg=kg) T (8C) Equation Product ab c rb ¼ a þ T þ cT2 n. a. 0.05/0.40 — 439 5.003 — rb ¼ a þ bXw þ þcXw2 Pistachio — — 236 440 — Coconut rb ¼ a þ b ln Xw 0.09/0.65 21 0.994 0.307 0.282 rb ¼ a À bxr þ c  10À6e(1:33xX=Xo) (shredded) 0.8/6.6 Ambient 0.636 0.102 a þ þbe(0:01BrixþcT) Orange juice 0.15/7.0 Ambient 1.251 À0.153 — rb ¼ 0:852 À 0:462eÀ0:66X(Brix) Apple (G. Smith) 0.25/1.0 10/90 0.828 0.3471 rb ¼ Pear Ambient 0.852 À0.462 À0.107 Apple juice À5:479À4 Apple From Lozano et al. (1979, 1983, 2002, 2005), Constenla et al. (1989), Moresi and Spinosi (1980), Jindal and Murakami (1984), Hsu et al. (1991). In the dilute limit, vw (water) has contributions mainly from structured free-solvent regions, while vs (solute) is affected by hydration and water–solute interactions. In the concentrated limit, vw is defined by water–solute aggregations, i.e., hydrogen bonded to hydroxyl groups. For these reasons, in sugar solutions both vw and vs are functions of concentration and temperature (Taylor and Rowlinson, 1955; Maxwell et al., 1984). Constenla et al. (1989) also suggested that the thermal effect on density could be significantly reduced by referring the specific volume to that of pure water vwo, so Eq. (4.11) can be written as: V =vwo ¼ Vw=IVwo þ ws(vs À vw)=vwo (4:12) Although according to the above discussion the partial specific volumes depend on concentration, from a practical point of view a linear relationship as suggested in Eq. (4.11) can be used to correlate density data, as Constenla et al. have found for apple juice: r ¼ rw=(0:992417 À 3:7391  10À3X ) (4:13) r2 ¼ 0:9989 Predictions of this equation were also extrapolated to temperatures in the range 10–908C. Perez and Calvelo (1984) proposed the following semiempirical equation for the bulk density calculation of beef muscle, during cooking: rb ¼ h rbo(1 À Xwo) i (4:14) Xw) 1 þ þrbo rbo(1ÀXwo) Xw)v (1 À À u(Xwo À rw rb(1ÀXw) where u and v are empirical parameters. Changes in food density by freezing were predicted by Hsieh et al. (1977) as follows:    1 1 1 1 r ¼ Mu ru þ Ms rs þ MI rI (4:15) where M is the mass fraction of unfrozen water (u), ice (I) and solids (s). The most significant change in density occurs immediately below the initial freezing temperature.

80 Fruit Manufacturing 4.4.2. Specific Heat Specific heat is the amount of heat required to increase the temperature of unit mass by unit degree at a given temperature. SI unit for Cp is [kJ/kg/K]. Specific heat of solids and liquids depends upon temperature but is not sensitive to pressure, as it is incompressible to practical purposes. It is common to use the constant pressure specific heat, Cp, which thermodynam- ically represents the change in enthalpy for a given change in temperature when it occurs at constant pressure: Cp ¼ (dH=dP)p (4:16) where H is enthalpy. 4.4.2.1. Measurement Methods Several methods have been used for the specific heat measurement (Rahman, 1995). Both differential scanning calorimetry (DSC) and the method of mixtures are commonly used techniques (Table 4.4). The advantages of DSC are that measurement is rapid, and a very small sample can yield accurate results for homogeneous products (Wang and Kolbe, 1991). Specific heat of selected fruits is listed in Table 4.5. 4.4.2.2. Prediction Models and Empirical Equations Mohsenin (1980) proposed an equation valid for the calculation of Cp of meats, fruits, vegetables, and other foods, which equals the sum of the specific heat of water (Cpw ) and solid matter (Cpsm ): Cp ¼ Cpsm þ (Cpw À Cpsm )xW (4:17) Table 4.4. Methods for specific heat measurement (Lozano 2005, with permission). Method Principle of operation Governing equation Additional comments Mixture A sample of known mass Cp ¼ Cpref Wref (Tref À Teq) Numerous calorimeters DSC (Ws) and temperature (Ts) Ws(Teq À TS) were developed to reduce is dropped into a calorimeter heat loss, heat generation of known specific heat, containing Cp ¼ Cpref Wref d by mixture and mixing a liquid (usually water) of known Ws dref problems mass (Wref ) and temperature(Tref ). Temperature of the mixture is recorded Reduced sample size and until equilibrium (Teq) escape of water vapor during heating are This method is based on the important limitations of determination of the amount of this technique heat required to raise the temperature of sample of known mass (Ws), at a given rate within a given interval. The measurement requires an external standard of known mass (Wref ) and specific heat (Cpref ) Reproduced with permission from Thermal Properties of Foods. In ‘‘FOOD ENGINEERING 1. Engineering Properties of Foods.’’ Edited by Gustaro V. Barbosa-Ca´novas. Paris UNESCO Publishing. ISBN 92-3-103999-7, pp.45-64. All of which is part of Encyclo- pedia of Life Support Systems (EOLSS). http://www.eols.net (copyright) EOLLS Publishers Co.Ltd.

4 . Thermodynamical, Thermophysical, and Rheological Properties 81 Table 4.5. Specific heat of selected fruits and fruit products and components. Product Xw DT (8C) Cp (kJ=kgÀ1=KÀ1) Apple juice 758Brix 30 2.805 758Brix 90 2.973 Apple, RD whole 108Brix 30 3.894 Apple sauce 0.75/0.85 — 3.95 Banana — — 3.73 Orange juice 74.8 — 3.35 Sugar 0.105 20/40 1.85 0.133 59.1 1.256 From Polley et al. (1980), Constenla et al. (1989), Gupta (1990), and Rahman (1995). A theoretical approach can also be used to predict the specific heat of a solution in terms of partial specific heats of individual components as follows: X (4:18) Cp ¼ Cpi wi A linear relationship of Cp with concentration as suggested by Eq. (4.18) is the basis for most of the existing correlations to evaluate specific heat of liquid foods (Choi and Okos, 1983). However, due to water–solute interactions, Cpi differs from the specific heat of pure com- ponent and usually changes with the concentration of soluble solids. Actually, the resulting values of Cps for sugar solutions are significantly higher than those corresponding to crystalline sugar at the same temperature (Taylor and Rowlinson, 1955; Pancoast and Junk, 1980), while at high water contents Cpw approximates to the heat capacity of pure water Cpwo . In addition, while Cpwo remains almost constant with temperature, the specific heat of the solution increases with this variable following the same pattern as that of crystalline sugar. This behavior was also observed in apple juice (Constenla et al., 1989), so no improvement in the correlation ability of Eq. (4.18) may be obtained by using the ratio Cp=Cpwo . Heldman (1975) proposed an expression for heat-capacity calculation of foods, based on the composition: Cp ¼ 4:180 (0:34Xca þ 0:37Xp þ 0:4XFA þ 0:2Xas þ 1:0Xw (4:19) Although widely used, Eq. (4.19) shows deviation when compared with experimental values, due to, among other conditions, the variation of Cp between bound and free water, and the excess specific heat due to the interaction of the component phases. Rahman (1993) proposed corrections reducing some of the limitations of Eq. (4.19). Below freezing point the calcula- tion is more difficult. Rahman (1995) cited Van Beek equation, which is considered to be applicable to all foods below the freezing temperature: Cp ¼ Cpso (1 À Xwo) þ CpwXwo(Tfreezing=T ) þ CpiceXwo(1 À Tfreezing=T ) (4:20) À LXwo(Tfreezing=T 2) Schwartzberg (1976), Chang and Tao (1981), and Mannapperuma and Singh (1989), among others, presented more complete equations for the prediction of heat capacity of foods below freezing point. Some other models applicable for different foods and conditions are listed in Table 4.6.

82 Fruit Manufacturing Table 4.6. Models for predicting heat capacity of fruit products and components. Application range Parameter Equation Fruit/component Xw (Kg=kg) T (8C) a b c  106 Cp ¼ a þ bT þ cT2 Sugars >0.5 À40/150 1.548 1:962:10À3 À5.934 Ash 1.093 1:890:10À3 À3.682 Cp ¼ a þ bXW Fiber 0.001/0.80 À40/150 1.846 1:831:10À3 À4.651 Cp ¼ a þ bXw þ T Fats 0.012/0.945 1.984 1:473:10À3 À4.801 Cp ¼ 1:56e0:945X w Ice 30/60 2.063 6:077:10À3 Proteins 20/40 2.008 1:209:10À3 — Water 4.176 9:086:10À5 À1.313 Fruits and 1.670 2.500 5.473 vegetables 2.477 2.356 — Food in general — — Fruit pulps 3.791 — * Adapted from Dickerson (1969), Mohsenin (1980), Gupta (1990), Alvarado (1991), Choi and Okos (1986), Lozano (2005). 4.4.3. Thermal Conductivity Thermal conductivity (k) is an intrinsic property of material and represents the quantity of heat that flows in unit time through a plate of unit thickness and unit area having unit temperature difference between faces. SI unit for k is [W/m/K]. Figure 4.2 shows thermal conductivity values measured in selected fruits at ambient temperature, while Fig. 4.3 shows the influence of temperature on k. 4.4.3.1. Measurement Methods Techniques for measurement of thermal conductivity can be classified as: (i) steady state, (ii) quasisteady state, and (iii) transient. 0.7 0.6 87% 88% 100% 78% 85% 0.5 89% k (W/m K) 86% 0.4 0.3 30% 0.2 0.1 0 Plum, dried Applesauce Pineapple Apple Pear Orange juice Strawberry Water Figure 4.2. Measured k of selected fruits and fruit products at ambient temperature.

4 . Thermodynamical, Thermophysical, and Rheological Properties 83 2 Sucrose (Xw=75) 1.5 Apple (GS) k, W/m K 1 0.5 0 60 40 20 0 20 40 60 80 Temperature, ؇C Figure 4.3. Influence of temperature in the thermal conductivity of selected foods (Keppler and Boose, 1970; Kent et al., 1984; Constenla et al., 1989). These techniques were extensively reviewed and compared by Rahman (1995). Table 4.7 describes different methods for k determination. Steady-state techniques (SST) are well- established methods based on the determination of constant heat flux under constant tem- perature gradient and the solution of the unidirectional steady-state heat conduction equation (Fourier’s equation) for k calculation. Although SST are mathematically simple, highly experiment controlled, and precise, they are practically not applicable in foods due to long equilibrium time (biological active samples, e.g., fruits, deteriorate), the need of geometrically defined samples (most foods are amorphous and/or soft), hence convection must be avoided (this excludes liquids or high-moisture foods). 4.4.3.2. Prediction Models and Empirical Equations Several empirical, semiempirical, and theoretical models and equations have been developed to predict thermal conductivity of composed material in general, and foodstuffs, in particular. Tables 4.8 and 4.9 list the most commonly used predictive models and empirical equations, respectively, valid for fruits and fruit products. 4.4.4. Thermal Diffusivity Thermal diffusivity (a) is a combination of three thermophysical properties that result from the derivation of the Laplace equation of heat conduction (Fourier equation in three dimensions): a ¼ k=rcp (4:21) Physically it represents the change in temperature produced in a unit volume of unit surface and unit thickness, containing r[kg] of matter, by heat flowing in the unit time through the unit face under unit temperature difference between opposite faces. Figure 4.4 shows thermal diffusivity values measured in selected fruits at ambient temperature.

Table 4.7. Description of different methods for k determination (Lozano 2005, with permission). Technique Method Principle of operation Governing equation Additional comments Q ¼ kA(Ti À To)=l Steady state 8 Heat source surrounded by sample, and in To calculate k, inner and outer >>>>><>>>>>>>>> turn surrounded by a heat sink. Insulation Q ¼ kAo[(To À Ti)= temperatures (Ti,To), sample Quasisteady Guarded hot is located at ends to avoid heat loss and ro ln (ri=ro)] thickness (l) and heat quantity state plate ensure unidirectional heat conduction (Q) must be measured k ¼ Wl=(Ti À To) >>>>>>>>>>>>>>:Concentric Usually heat source is the outer cylinder and Same as previous heat sink the inner cylinder. Heat absorbed ln [(Ti À Ts)=(T À Ts)] cylinders by coolant is the same as the heat ¼ kAt=lmcocpco W is the heat flux (W=m2). Heat flux conducted through sample. k is calculated from the slope of the plot of T i, Ts, mco, and cpco are initial 8 It is based on temperature-gradient ln [(Ti À Ts)=(T À Ts)]vs:t temperature, source >><Fitch (1935) and determination across sample. temperature, copper mass, and T À T1 ¼ (Q=4kK) ln (t=t1) copper Cp, respectively. >>: further k is evaluated at (Ti À To)=2 To do this, time and modifications Sample is ‘‘sandwiched’’ in between a temperature are correlated by T and T1, and t and t1, are the model equation: temperatures and times Transient 8 constant temperature heat source and a k ¼ Q=4ks corresponding to final and initial >>>>>>>>>< Line source copper plug as heat sinks insulated on all time, Q is heat produced per unit :>>>>>>>>> faces but one emf ffi k1=2 length of probe Thermal Zuritz et al. (1987) and Rahman (1991) Calibration is required with a comparator modified Fitch’s method for small number of materials of known individual food particles and frozen foods thermal conductivity Thermal conductivity probe is a test body made basically of a line source, providing a constant amount of heat and a temperature-measuring device. Alternative designs of the instrument have been discussed by Sweat (1974) and Hayashi et al. (1974) A probe is equilibrated to a higher temperature than the sample. Then the probe is placed in contact with food and it changes temperature at a different rate increasing emf, which is related by calibration to k Reproduced with permission from Thermal Properties of Foods. In ‘‘FOOD ENGINEERING 1. Engineering Properties of Foods.’’ Edited by Gustaro V. Barbosa-Ca´novas. Paris UNESCO Publishing. ISBN 92-3-103999-7, pp.45-64. All of which is part of Encyclopedia of Life Support Systems (EOLSS). http://www.eols.net (copyright) EOLLS Publishers Co.Ltd.

4 . Thermodynamical, Thermophysical, and Rheological Properties 85 Table 4.8. Predictive models for thermal conductivity estimation applicable to fruit products (Lozano 2005, with permission). Model Description Equation Series P In series distribution, layers of components Parallel are thermally in series with respect of heat 1=ke ¼ (fi=ki) flow. It is applied in food gels. Maxwell P Random Layers are considered as thermally in parallel ke ¼ (fiki) with respect to the direction of heat flow. Effective It is Proposed for liquid and powder foods kd þ 2kc À 2fa(kc À ka)! medium theory kd þ 2kc þ 2fd(kc À kd) Based on random distribution of noninteractive km ¼ kc spherical particles in a continuous medium k ¼ Pfi i The simplest is based on the weighed geometric mean of components, using the volume fraction P as the weighing factor fi[(k À ki)=ki À 2k)] ¼ 0 A model of statistical nature that considers a For porouspmffiffiaterials: heterogeneous medium as represented by a virtual k ¼ kp[b À (b2 þ 2m)] homogeneous one with the same properties. kp is b ¼ 3e À 1 þ [3(1 À e) À 1] the thermal conductivity of pores Reproduced with permission from Thermal Properties of Foods. In ‘‘FOOD ENGINEERING 1. Engineering Properties of Foods.’’ Edited by Gustaro V. Barbosa-Ca´novas. Paris UNESCO Publishing. ISBN 92-3-103999-7, pp.45-64. All of which is part of Encyclo- pedia of Life Support Systems (EOLSS). http://www.eols.net (copyright) EOLLS Publishers Co.Ltd. 4.4.4.1. Measurement Techniques The estimation of thermal diffusivity of foods can be done by: (i) direct measurement or (ii) indirect calculation using Eq. (4.21). Several direct methods for a determina- tion were proposed (Rahman, 1995) based on the solution of one-dimensional unsteady state heat transport equation with the appropriate boundary conditions for different geometries. Fourier equation has been solved for numerous conditions, and graphical solutions are also available (Bird et al., 2002). Fourier equation is limited to temperatures above freezing and restricted to homogeneous, isotropic substances. However, analytical and numerical solutions of the one-dimensional heat conduction equation have been used to determine the thermal behavior of foods. These techniques are similar to those used for k determinations, in particular the thermal probe. Table 4.10 lists some of the proposed direct techniques for the determination of a. Indirect method for determination of r, k, and Cp values needs more time and instrumentation. It was indicated that indirect determination yielded statistically more accurate values of a (Drouzas et al., 1991). Andrieu et al. (1989) found roughly 3% difference when determining thermal diffusivity in potato using both pulse and hot wire (probe) methods. 4.4.4.2. Empirical Equations Several empirical equations have been developed to predict thermal conductivity of fruits, and fruit products and components. Table 4.11 lists some empirical equations. 4.4.5. Viscosity Fluid and semisolid foods exhibit a variety of rheological behaviors ranging from Newtonian to time dependent and viscoelastic. Whereas an ideal elastic solid produces an elastic

86 Fruit Manufacturing Table 4.9: Empirical equations for calculating thermal conductivity of fruits and fruit components. Application range Parameters b  103 Equation* Fruits or fruit Xw (kg=kg) T (8C) a c  106 component k ¼ a þ bT Carbohydrates À40/150 0.2014 1.387 À4.331 þcT 2 Ash Fiber À40/150 0.3296 1.401 À2.907 Fats 0.1833 1.250 À3.168 Ice À25/0 0.1807 0.276 À0.177 Proteins 0/25 2.2196 À6.249 101.5 Water 30/70 0.1788 1.196 À2.718 Apple (GD) 0.5712 1.763 Ambient 1.290 À9.50 1.015 k ¼ a þ bXw Starch (gelatinized) 0.394 2.120 — þcXw2 Starch 0.210 0.410 — Pears 0.478 6.90 — k ¼ a þ bXw þcT 0.4875 0.057 XwÀ2 — Tomato paste 0.54/0.71 30 0.029 0.793 0.0227 Apple juice 0/758Brix 0.2793 À3.572 (8Brix) ( ln Xw)XwÀ2 Apple Fresh 10/80 À0.994 15.9 À10.3 1.135 (K) k ¼ kwor=rw Apple juice 0/758Brix T > Tfreezing 0.378 0.133 2.500 [a À bXw(Brix)] T Tfreezing 2.5289 1.376 10.300 Sucrose 0/75 T < Tfreezing 1.0052 0.930 T À2 Glucose 0/75 10/80 Fructose 0/75 — Fruits and meats 0.05/0.88 kef ¼ 1:82À1:66 10/80 2.6122 1.0381 — ko Liquid and solid 10/80 2.5617 0.9725 — foods (more 10/80 2.4153 0.9807 — than 400 data) 20/25 — — — [esp (À0:85) Xw ] X wo keff ¼ 0:58Xw — —— — þ0:155Xpr þ0:25Xcaþ þ0:135Xas þ0:16XFA *Equations must be checked for application ranges and influence of porosity. Adapted from Ramaswamy and Tung (1981, 1984), Drouzas and Saravacos (1985), Mattea et al. (1986), Sweat (1986), Constenla et al. (1989), Marouilis et al. (1991), Rahman (1991), Renaud et al. (1992), Choi and Okos (1996). displacement when shear stress is applied, a fluid produces viscous flow. If a liquid is held between two parallel infinite plates and the top plate moves at a velocity U (length/time) relative to the bottom plate, the force required to maintain this motion will produce a viscous flow and a velocity gradient, which is equivalent to the shear rate (g ¼ dU=dx) developed. Under this condition viscosity can be defined as: g ¼ t=ma (4:22) where ma is called the apparent viscosity and is constant only for this one value of g. If ma is a constant at different values, then: t¼m (4:23) where m is the Newtonian viscosity of the fluid.

4 . Thermodynamical, Thermophysical, and Rheological Properties 87 1.7 Thermal diffusivity (107 m2/s) 1.6 Water 1.5 Sucrose 1.4 1.3 Apple Peach Strawberry 1.2 0.95 1 1.1 0.75 Banana 0.9 0.7 0.8 0.85 Xw Figure 4.4. Thermal diffusivity of selected fruits and fruit components. Newtonian behaviors indicate that the viscosity of the food is shear-independent. Other than water, Newtonian flow is exhibited by sugar solutions and vegetable oils also. Viscosity of Newtonian foods has the unit Pas in the International System. Table 4.12 lists definitions of interest in the rheological study of foodstuffs. Most of the foods show more complicated relationships between shear rate and shear stress. It is no longer feasible to talk in terms of viscosity, since m varies with the rate of shear. Table 4.13 shows types of non-Newtonian fluids: (1) Those whose properties are independent of time of duration of shear, (2) Those whose properties are dependent on time of shear, and (3) Those exhibiting characteristics of a solid. 4.4.5.1. Measurement Techniques Methods of viscometry (measurement of apparent viscosity) are described in Table 4.14. Viscometers are based on the measurement of either the resistance to flow in a capillary tube, Table 4.10. Principle of operation and governing equation for the two most commonly used methods for the determination of thermal diffusivity in fruit products. Method Principle of operation Governing equation Analytical solution Sample is located in a cylinder (L=D > 4) a ¼ VR2=[4(Ts À Tc)] of Eq. (4.9) immersed in a water bath at constant T=¼2:1(lÀ=2gp4=k4)[:2À! þ0:5. .8.=]2gÀ¼lnr=g(2þpgaffiffiffi2t) temperature. Thermocouples located Probe method at the center of the sample (axis) and surface of cylinder measure are T vs. t. After transition, both temperature gradients are time independent Similar to thermal conductivity probe, with additional thermocouple placed at a known distance in the sample Dickerson (1965), Hayakawa (1973), Uno and Hayakawa (1980), Singh (1982), Gordon, Lozano (2002) and Thorne (1990); Nix et al. (1967).

88 Fruit Manufacturing Table 4.11. Empirical equations for the calculation of the thermal diffusivity of selected fruits and fruit products. Equation Product/component Application range a  102 Parameter c  106 T (C) b  104 a ¼ a þ T þ cT2 Ash À40/160 12.46 3.73 1.22 Carbohydrates À40/160 2.32 Fats À40/160 8.08 5.30 0.039 Fiber À40/160 2.22 Ice À40/160 9.88 1.26 95.0 Proteins À40/160 1.46 Water À40/160 7.39 5.19 2.40 Apple À25=À10 À10=Tfreezing 1.17 6.08 — a ¼ a þ bXw þ cXw2 Corn T > Tfreezing — a ¼ a þ bXw þ cT Multiple regression 6.87 4.76 — a ¼ 0:88:10À7 þ awXw Multiple regression 20/90 — 13.17 6.25 0.44 À0:88:10À7:Xw 8/23 À1:22:10À5 À0:187:10À3 0:29:10À3 — À4:37:10À5 À0:437:10À3 — À1:39:10À5 À0:278:10À5 0/80 À1:34:10À6 1:51:10À5 À1:83:10À2 9:56:10À5 0:786:10À5 0:574:10À3 — — Adapted from Lozano (2002), Riedel (1980), Ramaswamy and Tung (1981), Choi and Okos (1986), and Rahman (1995). or the torque produced by the movement of an element through the fluid. There are three main categories of commercially available viscometers applicable to foodstuffs: Capillary, falling-ball, and rotational viscometers. Of late, food scientists and technologists use rhe- ometers available at a relatively low cost, which can measure over wide ranges of shear behavior and perform complete rheograms, including thixotropic recovery, stress relaxation, or oscillatory experiment at programmed temperature sweep. For Newtonian liquid foods it is sufficient to measure m as the ratio t=g. Besides the ratio of shear stress and rate of shear, the properties required to describe a non-Newtonian material can be measured by: (a) compression (force–deformation relationship), (b) creep test (stress versus strain as a function of time), (c) stress relaxation (stress required to maintain a constant strain), and (d) dynamic test (deformation by a time variable stress, generally oscillatory stress). Table 4.12. Definition of different types of viscosities. Name Equation Comments Kinematic viscosity y ¼ m=r In (cm2=s stoke); where r is density Relative vis The reference Dickerson mr ¼ m=mo It is the ratio of solute to solvent viscosity at (1969) is not given in the reference list. msp ¼ mr À 1 equal temperatures Please check.cosity mred ¼ msp=c Specific viscosity [m] ¼ ( ln mr=c)[c!o] Where c is the concentration of solute Reduced viscosity Also called limiting viscosity number, which is Intrinsic viscosity usually correlated with molecular weight

4 . Thermodynamical, Thermophysical, and Rheological Properties 89 Table 4.13. Types and examples of non-Newtonian foods (Lozano 2005, with permission). Rheological Description Descriptive diagram Examples classification Non-Newtonian time-independent foods Most foods Bingham plastic Linear relationship between g and t Cocoa butter foods (ideal plastic t does not go through the origin. material) The t value at g ¼ 0 is the yield ty value or yield stress (ty) γ. Shear-thinning When g increases with t. When a t Most of non-Newtonian behavior food is pseudoplastic above yield γ. foods (fruit pure´es, (pseudoplastic stress is also known as condensed milk, foods) mixed-type plastic ketchup, etc.) Shear-thickening Their rheological behavior is t Starch suspensions and behavior opposite to the pseudoplastic in γ. some chocolate syrups (dilatant foods) which g decreases as t increases exhibit dilatant flow Thixotropic foods Time-dependent foods ma Some fruit pulps t Rate the shear stress value decrease with time at constant shear, while the structure collapses Rheopectic materials Include those few materials that are ma Whey protein polymers able to build up (or set up) while t have strong submitted to a shear stress at rheopectic properties constant. Semisolid foods These foods show both solid The viscoelastic Cheeses (elasticity) and fluid (viscosity) behavior of foodstuffs behavior when they are subjected is commonly to a sudden, instantaneous, explained by two constant shear stress; sufficient basic tests: stress time is allowed for the test; and relaxation and creep the stress is large enough to (increase of strain prevent the food showing pure with time) elasticity. During flow, normal stresses s are built up Reproduced with permission from Thermal Properties of Foods. In ‘‘FOOD ENGINEERING 1. Engineering Properties of Foods.’’ Edited by Gustaro V. Barbosa-Ca´novas. Paris UNESCO Publishing. ISBN 92-3-103999-7, pp.45-64. All of which is part of Encyclo- pedia of Life Support Systems (EOLSS). http://www.eols.net (copyright) EOLLS Publishers Co.Ltd. 4.4.5.2. Newtonian Fruit Products Liquid foods, such as clarified fruit juice, exhibit Newtonian behavior. As an early approxi- mation viscosity of Newtonian foods can be estimated as the viscosity of water (mw) and that of the prevalent soluble solids. Different empirical equations relating liquid food viscosity with both soluble solids and temperature were published (Rao, 1977). The viscosity of water,

90 Fruit Manufacturing Table 4.14. Methods for determination of viscosity. Method Description Governing equation* Capillary tube The time for a standard volume of fluid m ¼ pDPr4t=8Vl viscometers to pass through a length of capillary tube is measured. The driving force is gravity, where Falling-ball gas pressure, a descending piston, or partial DP is the driving force (pressure), viscometers vacuum at the exit. The force used to create r the capillary radius, flow is usually gravity as seen in Ostwald and V the volume, and Rotational Cannon–Fenske viscometers l the capillary length. viscometers For Newtonian liquid foods A standardized tube is filled with the product, n ¼ Kt where and the time under the influence of gravity for K is a constant a ball to pass between two specified points is measured. The falling ball reaches a limiting m ¼ K(rball À rs)=n velocity when the acceleration is exactly compensated by the friction of the fluid on the ball. The rolling-ball where instrument uses a tube inclined at a 10-degree angle, K ¼ (0:374gD(D þ d) sin u, and which allows the ball to remain in contact with the d is the diameter of the ball, inner tube surface. The drawn-ball type uses a ball D the diameter of the tube, mechanically pulled upward through the tube. g the acceleration of gravity, and These viscometers accurately measure low- to u the angle of tilting of the tube medium-viscosity Newtonian fluids. The drawn ball may allow the measurement of opaque fluids Coaxial cylindersþ: These instruments can determine the viscosity m ¼ GT(1=R21 À 1=R22)=4plv of Newtonian and non-Newtonian fluids Cone plate: contained between two coaxial cylinders m ¼ 3GTu=(2pR3v) (bob and cup), or different geometries as the where GT is the torque of the bob, cone and plate geometry, by measuring the drag v the angular velocity, of the fluid on a mobile member (cylinder or cone) R1 the radius of the bob, while the other member (cylinder or plate) remains Ro the radius of the cup, and stationary. They produce precise measurements of R the radius of the plate absolute viscosity for a wide range of viscosities. u the cone angle, Because the shear rate can be varied, it is possible v the angular velocity, to plot the flow curves of non-Newtonian fluids. l the cylinder length. Time effects can be studied either manually or automatically with computerized controls *Adapted from Slattery (1961), Van Waser et al. (1963), Johnson et al. (1975), and Bourne (1982). and salt and sucrose solutions, major solutes in foodstuffs, can be calculated by Eqs. (4.24), (4.25), and (4.26), respectively (Kubota et al., 1980): ln mw ¼ 0:266 À 2:02 Â 10À2T þ 4:4 Â 10À5T2 (4:24) where T is the temperature in 8C. m ¼ a exp (b=Tn) (283:2K < T < 323:2K) (4:25) (0 < X < 40Brix) where log a ¼ 0:00458X 1:15 À 3:05 and b ¼ 9:90 Â 104X 1:51 þ 6:1 Â 107