Journal of Food Engineering 82 (2007) 284–291 www.elsevier.com/locate/jfoodeng Kinetics of osmotic dehydration and air-drying of pumpkins (Cucurbita moschata) Carolina Castilho Garcia, Maria Aparecida Mauro *, Mieko Kimura Department of Food Engineering and Technology, Institute of Biosciences, Language and Physical Sciences (IBILCE), UNESP – São Paulo State University Rua Cristóvão Colombo 2265, 15054-000 - São José do Rio Preto, SP, Brazil Received 13 October 2006; received in revised form 31 January 2007; accepted 1 February 2007 Available online 16 February 2007 Abstract Kinetics of osmotic dehydration (OD) and effects of sucrose impregnation on thermal air-drying of pumpkin slices were investigated. A simplified model based on the solution of Fick’s Law was used to estimate effective diffusion coefficients during OD and air-drying. In order to take into account shrinkage, average and variable thicknesses were considered. Pumpkin slices were dehydrated in sucrose solutions (40%, 50% and 60%, w/w, 27 °C). The effective water diffusion coefficients were higher than the sucrose, and low diffusivity dependence with solution concentration was observed. Samples non-treated and pre-treated in 60% osmotic solutions during one hour were dried in a hot-air-dryer at 50 and 70 °C (2 m/s) until equilibrium was achieved. Pre-treatment enhanced mass transfer during air-drying. Great volume reduction was observed in pre and non-treated dried samples. Using variable thickness in the model diminished the relative deviations between predicted and experimental OD and drying data. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Diffusion coefficients; Osmotic dehydration; Convective drying; Shrinkage; Pumpkin 1. Introduction Due to physical, chemical and biochemical changes during drying, quality degradation is a major concern in the selection, design and operation of a food drier (Mujumdar, 1997). Dehydration of foodstuffs by immersion in osmotic solutions previous to convective air-drying seems to improve the quality of the final product since it prevents oxidative browning and/or loss of volatile flavoring constituents, reduces the fruit acidity (Ponting, 1973), and can decrease structural collapse during air-drying (Del Valle, Cuadros, & Aguilera, 1998; Lenart, 1996; Simal, Deya, Frau, & Rossello, 1997). This pre-treatment can also minimize drying color losses (Nsonzi & Ramaswamy, 1998), as well as reduce nutrient losses, e.g. lycopene in vac- * Corresponding author. Tel.: +55 17 3221 2253; fax: +55 3221 2299. E-mail address: cidam@ibilce.unesp.br (M.A. Mauro). 0260-8774/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2007.02.004 uum-dried tomatoes (Shi, Le Maguer, Kakuda, Liptay, & Niekamp, 1999). Osmotic dehydration (OD) of fruits and vegetables is based on their immersion in an aqueous concentrated solution containing one or more solutes. This process involves the simultaneous flow of water and solutes. Water and food solutes diffuse from the food to the concentrated solution, and solution solutes from the osmotic solution into the food. Solute transfer is usually limited due to differential permeability of cellular membranes (Bildweel, 1979). Consequently, more water transfer than solute transfer characterizes this process. The effects of osmotic pre-treatment on drying rates have been investigated by several authors (Karathanos, Kostrapoulos, & Saravacos, 1995; Lenart, 1996; Park, Bin, & Brod, 2002; Rahman & Lamb, 1991; Simal et al., 1997) and vary according to the raw material used and the drying conditions. Sucrose is considered one of the best osmotic substances, especially when the OD is employed before C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 285 Nomenclature effective diffusion coefficients of k species (m2/s) effective diffusion coefficients of moisture (m2/s) total mass variation in relation to initial mass (dimensionless) M mass (kg) N number of observations or residuals P pressure (kg/(ms2)) R universal gas constant (kJ/(kg-mol K)) R2 correlation coefficient (dimensionless) RRMS mean relative error root square (dimensionless) SG sugar gain in relation to initial mass (dimensionless) t time (s) T temperature (K) V volume (m3) a V volume at a specific state a (=OD or D) (m3) WL water loss in relation to initial mass (dimensionless) wk fractional or residual content of species k, dry basis (dimensionless) mass fraction of species k, wet basis (dimensionwk less)  k ðtÞ average content of species k at time t, wet basis w (dimensionless) Dk Dm DM drying. The presence of this sugar on the surface of the dehydrated sample is an obstacle for the contact with oxygen (Lenart, 1996) thus reducing the oxidative reactions. Pumpkins are good sources of carotenoids, and some varieties are rich in provitamins A, mainly a-carotene and b-carotene (Arima & Rodriguez-Amaya, 1988; Azevedo-Meleiro, 2003; Dutta, Raychaudhuri, & Chakraborty, 2006; Murkovic, Mülleder, & Neunteufl, 2002). Carotenoids are among the phytochemical components believed to reduce the risk of developing some degenerative diseases, and are responsible for the attractive color of many fruits and vegetables. Being highly unsaturated, they are prone to isomerization and oxidation during processing because of the contact with acids, heat treatment and exposure to light, resulting in some loss of color and biological activity alteration (RodriguezAmaya, 2002). Hence, improvement of nutritional and sensory properties of air-dried pumpkins could be achieved by applying a suitable pre-osmotic treatment. Nevertheless, as observed by Mayor, Moreira, Chenlo, and Sereno (2006), even though OD of pumpkin could be a useful technique to obtain new processed products of interest to the consumer, few works have been found in the literature. The mentioned authors present a broad study about OD of Cucurbita pepo L. in salt solution. X ðtÞ X y z average moisture content at drying time t, dry basis (dimensionless) fractional or residual moisture content, dry basis (dimensionless) experimental or calculated value (dimensionless) length scale (m) Greek letters q density or volumetric concentration (kg/m3) qk volumetric concentration of species k (kg/m3) Subscripts and Superscripts 0 initial state a sample state, osmotically dehydrated (OD) or dried (D) cal calculated D dried eq equilibrium exp experimental OD osmotically dehydrated rs reducing sugars s sucrose ts total solids w water Kowalska and Lenart (2001) also report the high OD efficiency of pumpkin (Cucurbita maxima vs. Melonowa) in sucrose solution, when compared to carrot and apple tissues. Pan, Zhao, Zhang, Chen, and Mujumdar (2003) studied OD of several plants, including pumpkins in sugar solutions, and subsequent drying. They obtained a significant reduction in the thermal drying time of the impregnated materials in comparison to the fresh material. There are some works that do not deal with OD but apply different techniques aiming to improve dried pumpkin quality such as combination-drying, e.g. initial partial freeze-drying followed by terminal hot air-drying of the Cucurbita maxima species (Kumar, Radhakrishna, Nagaraju, & Rao, 2001). Another work applied freezing or blanching to the same species (Cucurbita maxima) before vacuum drying, which increased the moisture diffusivity (Arévalo-Pinedo & Murr, 2007). The purpose of this work is to evaluate OD kinetics of pumpkins in sucrose solutions and the effects of solute impregnation on the kinetics of air-drying using a simplified model based on Fick’s Law that takes into account shrinkage. Effective diffusion coefficients and methodology to consider shrinkage in the model can be useful for projects and to control the processes. 286 C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 2. Materials and methods sugar contents were determined in triplicate by oxy-reduction titration (William, 1970). 2.1. Raw material 2.5. Calculations Mature pumpkins (Cucurbita moschata) obtained on the local market were cut in three portions in a transversal direction to their axis and each portion was cut in four longitudinal pieces. The pieces were peeled, seeded and sliced (3.97 ± 0.15 mm thickness). The transversal area of the slice was approximately 20–25 cm2. 2.5.1. Mass and volume variation The total mass variation in relation to initial mass, during osmotic dehydration, was calculated from experimental data following Eq. (1): DM ¼ 2.2. Osmotic dehydration Pumpkin slices were weighed, placed in four mesh baskets and immersed in sucrose (commercial sugar) aqueous solutions (40, 50 and 60%, w/w). The OD system consisted of a jacketed stainless steel vessel containing 15 kg of aqueous sucrose solution continuously stirred and maintained at 27 °C by circulation of thermostatically controlled water in the jacket. Each basket containing approximately 14 slices corresponded to a single OD time: 0.5, 1, 2 and 3 h. Two baskets were prepared for each process time. Syrupto-fruit ratio was approximately 15:1. After the pre-established contacting period, the samples were removed, their surfaces cleaned with wet tissue, blotted with absorbing paper, and weighed. Equilibrium data was experimentally determined by immersion of slices (2 mm thickness) during 48 h in 40, 50 and 60% sucrose solutions. Preliminary tests showed that 48 h were enough for equilibrium to be achieved. Solids, total and reducing sugars contents, were determined in fresh and osmotically treated samples. 2.3. Hot-air-drying Samples non-treated and pre-treated in osmotic solutions at 60% w/w during 1 h, were dried at 50 and 70 °C. Drying experiments were carried out in a laboratory scale drier operating with air-velocity of 2 m/s. The drier was equipped with an electronic balance with an accuracy of 0.01 g. The weight was continuously registered in a microcomputer using a RS232 interface. The air-flowed parallel to the bed that consisted of three wire nets supported by a structure, which substituted the balance plate. Approximately 0.3 kg of osmotically treated and fresh samples was dried until equilibrium moisture was achieved. Solid content was determined in fresh and osmotically treated samples. Specific volumes of fresh, osmotically and dried samples were measured and shrinkage was calculated. 2.4. Analytical methods Solid content of fresh and osmotically treated samples was determined in triplicate, gravimetrically, by drying in a vacuum oven at 60 °C, 10 kPa, until constant weight was achieved. The density was determined in duplicate by volume dislocation technique using 50 ml pycnometers and toluene as a displacing fluid. The reducing and total ðM  M 0 Þ M0 ð1Þ The water loss (WL) and the sugar gain (SG) in relation to initial mass was calculated for the OD, through the mass balances shown in Eqs. (2) and (3): ðww MÞ  ðw0w M 0 Þ M0 ðws MÞ  ðw0s M 0 Þ SG ¼ M0 WL ¼ ð2Þ ð3Þ Sugar gain was also calculated by difference: SG ¼ DM  WL ð4Þ To calculate the shrinkage during each step of the processes, the volume variation was expressed by Eq. (5): VaV0 ¼ V0  a  0 1 M M0 M  qa q0 q0 ð5Þ where superscript a indicates the sample condition, osmotically dehydrated (OD) or dried (D). 2.5.2. Effective diffusion coefficients 2.5.2.1. Osmotic dehydration. The effective diffusion coefficients of water and sucrose (Dk) were determined according to Fick’s Second Law applied to a plane sheet. The analytical solution in terms of the mean water or  k ðtÞ, is (Crank, sucrose content in the slab at OD time t, w 1975): wk ¼ ¼  k ðtÞ  weq w k w0k  weq k 1 8 X 1 p2 k ¼ w; s n¼1 p2 Dk t exp ð2n  1Þ 2 z2 ð2n  1Þ  2  ð6Þ In an approximate manner, the thickness of the slab, z, was assumed to be a linear function of ww, since the shrinkage incorporated in the analytical solution of Fick’s Law proved to be adequate to fit experimental data of pumpkins osmotically dehydrated in sodium chloride solutions (Mayor et al., 2006) and, in some specific conditions, of pineapple osmotically dehydrated in sucrose solutions, (Ramallo, Schvezov, & Mascheroni, 2004). The coefficients of Eq. (6) were also determined considering z as an average thickness calculated between initial and final measurements. 287 C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 2.5.2.2. Air-drying. The effective diffusion coefficients of moisture (Dm) were determined according to Fick’s Law applied to a plane sheet. The diffusion model has been applied to the drying of biological materials (Daudin, 1983). However, in Eq. (7), the fractional content changes to express the moisture on a dry basis. X ðtÞ  X eq X 0  X eq   1 2 8 X 1 2 p Dm t ¼ 2 exp ð2n  1Þ p n¼1 ð2n  1Þ2 z2 X¼ ð7Þ In Eq. (7), thickness z was considered both variable (a linear function of X) and invariable (an average value between initial and final dimension) during the processes. 2.5.2.3. Fitting. The diffusion coefficients in Eqs. (6) and (7) were determined from the experimental data by minimizing the squares of the deviations between predicted and observed values. These series converged quickly so four terms were enough. The theoretical model was fitted to experimental data by non-linear regression using the steepest–descent method (Marquard, 1959). The fitting efficiency was evaluated by the correlation coefficient R2 and the mean relative error root square (RRMS); the latter according to Eq. (8): ( )1=2 N 1 X 2 RRMSð%Þ ¼ 100 ½ðy exp  y cal Þ=y exp  ð8Þ N n¼1 3. Results and discussion Total solids, sucrose and reducing sugar contents as well as mass variation Eq. (1), water loss Eq. (2) and sucrose gain (Eq. (3)) of the samples treated osmotically in 40, 50 and 60% sucrose solution are reported in Tables 1–3, respectively. After equal time process, the higher the osmotic concentration, the higher were the total solid and sugar contents measured in the samples. Except for reducing sugars, equilibrium contents also increased with the osmotic solution concentration. Part of these sugars was lost after a long process time (48 h), probably due to damages in the cellular tissue. Water loss and sugar gain did not follow a pattern. This was attributed to differences between the pumpkins, since small differences in initial composition can cause distinct mass variations during OD. For high moisture materials such as pumpkins, a small change in the water content corresponds to a great variation of water loss in relation to the initial mass. Table 4 shows the effective diffusion coefficients of water and sucrose determined from Eq. (6), considering an average thickness (between initial and final state) and a variable thickness (a linear function of water content). In both diffusivity determinations the thickness z was estimated considering similar shrinkage in all dimensions. To do this, volume variation was determined from the water loss and sucrose gain (Tables 1–3) and the corresponding specific volume of the water and sucrose (Perry & Chilton, 1973). Eq. (6) better predicted the water and sugar contents when variable thickness was considered. Reasonable fitting was obtained with this assumption since the values of R2 were always higher than 0.94 and almost all RRMS values were lower than 20% (Table 4). Even though variable thickness consideration has enhanced the fitting, the coefficient values did not differ more than 10% between the different calculation methods. Effective diffusion coefficients of sucrose were lower than water in all treatments. Both water and sucrose diffusivities were independent of osmotic concentration at 50% and 60% sucrose solutions; this was probably due to a conjunction of opposite factors. If, on the one hand, diffusivity decreases as sucrose concentration increases in binary sucrose-water solutions (Henrion, 1964), on the other hand, the higher sucrose concentration the higher damage tissue, which makes sucrose diffusion easier. Diffusivities determined with 40% solution were slightly higher than the others, as would be expected for more diluted solution. The mass transfer efficiency in OD is generally estimated as the ratio water loss/sugar gain. The most efficient treatments were with 40% and 60% solution for 1/2 and 1 h. On taking into account efficiency, short process time and low water content, a 60% sucrose solution during 1 h treatment was selected to be employed in convective drying. Samples non-treated and osmotically treated in this condition were dried at 50 and 70 °C. Relative humidity was approximately 22% and 9% at 50 °C and 70 °C, respectively. Water content measured in fresh and osmotically treated samples, and that obtained in dried samples by mass balance, as well as the mass variation (DM), the water loss (WL) and the sugar gain (SG) (the latter calculated by difference as in Table 1 Total solids (wts), sucrose (ws ) and reducing sugars (wrs) content; variation in mass (DM), water loss (WL) and sucrose gain (SG) in relation to the initial mass (M0), during osmotic dehydration in 40% sucrose solution Time (h) wts (kg/100 kg) ws (kg/100 kg) wrs (kg/100 kg) DM (kg/100 kg) WL (kg/100 kg) SG (kg/100 kg) 0 0.5 1 2 3 48 6.84 ± 0.03 18.04 ± 0.23 23.46 ± 0.17 28.60 ± 0.06 36.63 ± 0.09 39.77 ± 0.05 1.67 ± 0.05 8.62 ± 0.03 13.05 ± 0.18 16.43 ± 0.97 22.15 ± 0.35 25.94 ± 0.54 2.43 ± 0.01 4.21 ± 0.04 4.51 ± 0.04 4.28 ± 0.24 5.92 ± 0.09 3.54 ± 0.01 0 39.64 ± 0.04 48.26 ± 1.39 54.21 ± 0.08 61.95 ± 0.60 – 0 43.71 ± 0.03 53.58 ± 1.07 60.49 ± 0.06 69.06 ± 0.37 – 0 3.53 ± 0.01 5.08 ± 0.18 5.52 ± 0.01 6.76 ± 0.13 – 288 C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 Table 2 Total solids (wts), sucrose (ws ) and reducing sugars (wrs) content; variation in mass (DM), water loss (WL) and sucrose gain (SG) in relation to the initial mass (M0), during osmotic dehydration in 50% sucrose solution Time (h) wts(kg/100 kg) ws (kg/100 kg) wrs(kg/100 kg) DM (kg/100 kg) WL(kg/100 kg) SG(kg/100 kg) 0 0.5 1 2 3 48 8.63 ± 0.03 19.96 ± 0.09 28.53 ± 0.14 31.67 ± 0.14 43.77 ± 0.12 49.54 ± 0.16 1.99 ± 0.26 10.38 ± 0.02 15.99 ± 0.43 18.70 ± 0.85 25.22 ± 1.03 35.72 ± 0.95 2.93 ± 0.13 3.96 ± 0.13 4.99 ± 0.12 4.96 ± 0.07 6.57 ± 0.51 3.40 ± 0.11 0 29.11 ± 3.03 46.48 ± 1.08 49.76 ± 1.05 59.56 ± 1.38 – 0 34.63 ± 2.43 53.12 ± 0.78 57.04 ± 0.72 68.63 ± 0.78 – 0 5.37 ± 0.31 6.57 ± 0.17 7.41 ± 0.20 8.21 ± 0.35 – Table 3 Total solids (wts), sucrose (ws ) and reducing sugars (wrs) content; variation in mass (DM), water loss (WL) and sucrose gain (SG) in relation to the initial mass (M0), during osmotic dehydration in 60% sucrose solution Time (h) wts (kg/100 kg) ws (kg/100 kg) wrs (kg/100 kg) DM (kg/100 kg) WL (kg/100 kg) SG (kg/100 kg) 0 0.5 1 2 3 48 7.23 ± 0.02 23.10 ± 0.12 35.81 ± 0.09 39.08 ± 0.10 51.07 ± 0.22 59.18 ± 0.04 1.62 ± 0.18 12.67 ± 0.21 20.73 ± 0.08 23.13 ± 0.67 31.29 ± 0.72 44.46 ± 1.48 3.27 ± 0.11 5.65 ± 0.26 7.01 ± 0.71 7.52 ± 0.34 8.63 ± 0.31 2.99 ± 0.04 0 47.56 ± 3.55 60.59 ± 1.81 59.59 ± 0.03 68.26 ± 0.18 – 0 52.45 ± 2.72 67.47 ± 1.16 68.15 ± 0.01 77.24 ± 0.09 – 0 4.94 ± 0.44 6.55 ± 0.38 7.73 ± 0.01 8.84 ± 0.06 – Table 4 Effective diffusion coefficients of water and sucrose according to Eq. (6), calculated with average and variable thickness consideration Osmotic solution, w/w 40% 50% 60% Water Sucrose Average thickness Variable thickness Average thickness Variable thickness Dw  1010 (m2/s) RRMS (%) R2 Dw  1010 (m2/s) RRMS (%) R2 Ds  1010 (m2/s) RRMS (%) R2 Ds  1010 (m2/s) RRMS (%) R2 1.66 1.45 1.44 43.1 32.1 18.3 0.927 0.881 0.911 1.69 1.34 1.36 26.8 17.4 11.6 0.970 0.945 0.944 1.35 0.97 0.98 26.9 7.4 11.7 0.903 0.952 0.910 1.37 0.88 0.90 17.2 5.9 7.8 0.954 0.967 0.951 Eq. (3)), are shown in Table 5. The impregnation of sucrose in the pre-treated tissue contributed to a higher water content in the dried samples. This was due to the higher water retention capacity by the sucrose (Chirife, Fontan, & Benmergui, 1980), in comparison to cellulosic (Papadakis, Bahu, Mckenzie, & Kemp, 1993) or proteic (Bull, 1944) compounds at low humidity levels. Despite the initial water content of the pumpkins osmotically dehydrated prior drying at 50 °C and 70 °C being similar, small differences in composition probably influenced the mass transfer behavior that was more effective (ratio water loss/sugar gain) in the second experiment. The densities of these pumpkins were different, as shown in Table 6, where experimental results for densities of fresh, osmotically dehydrated and dried pumpkins, and volume variation calculated according to Eq. (5), are presented. Great volume reduction and density increase was observed after air-drying of both the non-treated and osmotically treated samples. However, some shrinkage prevention was obtained in pre-treated pumpkins (Table 6). The volume reduction, from initial (fresh) to dried state, calculated according to Eq. (5), was compared with the water volume loss, based on sucrose and water changes during OD, water changes during airdrying and corresponding specific volumes of the water and sucrose (Perry & Chilton, 1973). Volume reduction was very similar to the water volume loss, showing that Table 5 Water content (ww) of fresh, osmotically dehydrated and dried pumpkin (50 and 70 °C); variation in mass (DM), water loss (WL) and sugar gain (SG) in relation to the initial mass (M0), during osmotic treatment (60% sucrose solution, 1 h) Drying temperature Fresh Osmotically dehydrated Mass variations during osmotic treatment ww (kg/100 kg) ww (kg/100 kg) DM (kg/100 kg) 50 °C–non-treated 50 °C–pre-treated (60%, 1 h) 70 °C–non-treated 70 °C–pre-treated (60%, 1 h) 92.18 ± 0.03 92.17 ± 0.02 92.74 ± 0.05 93.27 ± 0.53 WL (kg/100 kg) SG (kg/100 kg) 75.56 ± 0.04 41.88 ± 0.75 48.25 ± 0.56 6.37 ± 0.18 74.43 ± 0.10 55.59 ± 0.87 60.22 ± 0.65 4.62 ± 0.22 Dried ww (kg/100 kg) 7.73 9.14 6.01 9.26 289 C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 Table 6 Densities of fresh (q0), osmotically dehydrated (qOD) and dried pumpkin (qD), and volume variation from fresh sample (V0) to the osmotically dehydrated (VOD) or dried (VD) samples, according to Eq. (5) Drying temperature Densities Volume variation q0 (kg/m3) 50 °C–non-treated 50 °C–pre-treated (60%, 1 h) 70 °C–non-treated 70 °C–pre-treated (60%, 1 h) qOD (kg/m3) 1016.5 ± 1.2 1016.5 ± 1.2 1005.6 ± 5.1 998.8 ± 1.9 1262.5 ± 24.2 1361.4 ± 7.5 1307.6 ± 5.0 1302.4 ± 22.9 1104.4 ± 3.9 1074.1 ± 7.2 the osmotic pre-treatment diminished shrinkage due to the volume occupied by sucrose impregnated in the tissue. The density experimental results of fresh pumpkin, close to water density, suggest that this tissue has low porosity. Microscopic observations of the fresh pumpkin tissue, according to a procedure described by Mauro, Tavares, and Menegalli (2003), also confirmed its limited intercellular air-spaces. The Dm values determined from Eq. (7) with average and variable thickness consideration as well as the correspondent R2 and RRMS can be observed in Table 7. Eq. (7) better predicted the moisture content of treated and non-treated samples when variable thickness was considered. In this case the R2 values were always higher than 0.98 and the RRMS values resulted in lower than 20% for non-treated samples. On the other hand, for treated samples, the RRMS values were around 30%, but these higher values are due to small X values at the last drying stages that increase relative deviations in Eq. (8). A comparison between experimental and calculated air-drying curves according to Fick’s model Eq. (7), considering variable thickness of the non-treated and osmotically treated sliced pumpkin, is shown in Fig. 1. The diffusion coefficients found in this work, for nontreated pumpkin, can be compared to those predicted by Rovedo, Suarez, and Viollaz (1997). The authors measured moisture content, temperature and surface area variations during drying of Cucurbita pepo L. at 40, 50 and 60 °C, and applied a rigorous approach that considered mass and heat transfer in a three dimensional shrinking solid slab. They determined fitting parameters according to Arrhenius equation that, if applied at 50 and 70 °C, will result in diffusion coefficients equal to 0.70  1010 m2/s at 50 °C and 1.59  1010 m2/s at 70 °C. These values are very similar to the diffusivities estimated in our work when V OD  V 0 (%) V0 qD (kg/m3) VD V0 (%) V0 93.18 88.33 94.06 90.40 46.51 58.70 variable thickness was considered (Table 7), showing that this shrinkage consideration can be useful to predict diffusivities within reasonable accuracy. Nevertheless, most of the drying diffusion coefficients reported in the literature are calculated using the initial thickness of the slabs (Vaccarezza & Chirife, 1975). In the present work, if Eq. (7) is fitted to the experimental data of non-treated pumpkin considering the initial thickness (4 mm), diffusivities will result equal to 3.25  1010 m2/s at 50 °C and 7.24  1010 m2/s at 70 °C, which are comparable to values reported in the literature. Doymaz (2007) dried pumpkin Cucurbita pepo L. at 50, 55 and 60 °C and so obtained diffusivity values of 3.88  1010, 6.58  1010 and 9.38  1010 m2/s, respectively. Akpinar, Midilli, and Bicer (2003) also investigated pumpkin slices drying using a cyclone type dryer. At 70 °C, moisture diffusion coefficients were obtained between 4.0 and 7.3  1010 m2/s, varying as function of the tray position and air-velocity. Pre-treatment of fruits in sugar solution usually reduces the convective drying rates (Simal et al., 1997; Karathanos et al., 1995; Rahman & Lamb, 1991). However, Park et al. (2002) found that the water diffusion coefficients in osmotically dehydrated pears were greater than in the fresh fruit while the air-drying velocity was 2 m/s, but not for 1 m/s. This was explained by the reduction in the effect of shrinkage and surface hardening due to the osmotic treatment. In this work, the effect of the pre-treatment on convective drying enhanced the water transfer (Fig. 1). The effective diffusion coefficients, calculated with shrinkage consideration (average or variable thickness), were higher for pre-treated samples than non-treated ones (Table 7). This unusual behavior is probably related to the fast drying of the fresh pumpkin surface due to the high air-velocity (2 m/s) and its high water content. Table 7 Effective moisture diffusion coefficients according to Eq. (7) calculated with average and variable thickness consideration 50 °C 70 °C Average thickness 10 Non-treated Pre-treated Variable thickness 2 Dm  10 (m2/s) RRMS (%) R 1.59 2.06 89.3 58.0 0.808 0.911 10 Average thickness 2 Dm  10 (m2/s) RRMS (%) R 0.78 1.34 13.6 29.8 0.991 0.983 10 Variable thickness 2 Dm  10 (m2/s) RRMS (%) R 3.55 4.16 97.6 31.3 0.865 0.938 Dm  1010 (m2/s) RRMS (%) R2 1.51 2.83 11.8 30.2 0.990 0.986 290 C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 1 0.9 non-treated, 50˚C, exp 0.8 treated, 50˚C, exp 0.7 non-treated, 70˚C, exp treated, 70˚C, exp 0.6 non-treated, cal X (t ) − X eq 0.5 X 0 − X eq 0.4 treated, cal 0.3 0.2 0.1 0 0 1 2 Time (h) 3 4 Fig. 1. Comparison between experimental and calculated residual water content according to Eq. (7) with variable thickness consideration, in fresh and osmotically treated pumpkin during drying at 50 °C and 70 °C. 4. Conclusions OD kinetics of pumpkin in sucrose solutions was obtained for 40, 50 and 60% concentration. Effective diffusion coefficients of sucrose were lower than water. Comparison between the treatments showed no dependence of the diffusion coefficients on concentrations at 50% and 60%. Slightly higher values for 40% osmotic treatment were obtained. Drying kinetics was determined for pre-treated (60%, 1 h) and non-treated pumpkin, at 50 and 70 °C. The water effective diffusion coefficients increased with the drying temperature. Pre-treatments enhanced the water transfer during drying. The moisture diffusion coefficients resulted to be higher than those for the non-treated ones. This unusual behavior was related to the fast surface drying of the fresh samples, forming areas of hardness on the surface and reducing drying rates. Great volume reduction and density increase was observed after air-drying of the non-treated and osmotically treated samples. The use of shrinkage consideration in the simplified model based on the analytical solution of Fick’s Law proved to be adequate when the variable thickness was applied, and this procedure better predicted OD and drying experimental data. Acknowledgements The authors thank the FAPESP (Fundacßão de Amparo à Pesquisa do Estado de São Paulo) for the fellowship (proc 02/10806-8) and research financial support (proc 03/10151-4). References Akpinar, E. K., Midilli, A., & Bicer, Y. (2003). Experimental investigation of drying behaviour and conditions of pumpkin slices via a cyclone- type dryer. Journal of the Science of Food and Agriculture, 83, 1480–1486. Arévalo-Pinedo, A., & Murr, F. E. X. (2007). Influence of pre-treatments on the drying kinetics during vaccum drying of carrot and pumpkin. Journal of Food Engineering, 80(1), 152–156. Arima, H. K., & Rodriguez-Amaya, D. B. (1988). Carotenoid composition and vitamin A value of commercial Brazilian squashes and pumpkins. Journal of Micronutrient Analysis, 4, 177–191. Azevedo-Meleiro, C. H. (2003). Análise de Carotenóides em Alimentos Brasileiros por Cromatografia Lı́quida de Alta Eficiência e Espectrometria de Massas. Doctoral Thesis, State University of Campinas (UNICAMP), School of Food Engineering, Department of Food Science, Brazil. Bildweel, R. G. S. (1979). Plant Physiology (2nd ed.). New York: Macmillan Publishing. Bull, H. B. (1944). Adsorption of water vapor by proteins. Journal of the American Chemical Society, 66, 1499–1507. Chirife, J., Fontan, C. F., & Benmergui, E. A. (1980). The prediction of water activity in aqueous solutions in connection with intermediate moisture foods. IV. aw prediction in aqueous non electrolyte solutions. Journal of Food Technology, 15, 59–70. Crank, J. (1975). The Mathematics of Diffusion (2nd ed.). London: Clarendon Press-Oxford. Daudin, J. D. (1983). Calcul des cinétiques de séchage par l’air chaud des produits biologiques solides. Sciences des Aliments, 3, 1–36. Del Valle, J. M., Cuadros, T. R. M., & Aguilera, J. M. (1998). Glass transition and shrinkage during drying and storage of osmosed apple pieces. Food Research International, 31(3), 191–204. Doymaz, I. (2007). The kinetics of forced convective air-drying of pumpkins slices. Journal of Food Engineering, 79(1), 243–248. Dutta, A., Raychaudhuri, U., & Chakraborty, R. (2006). Rheological characteristics and thermal degradation kinetics of beta-carotene in pumpkin puree. Journal of Food Engineering, 76, 538–546. Henrion, P. N. (1964). Diffusion in the sucrose + water system. Transaction Faraday Society, 60, 72–74. Karathanos, V. T., Kostrapoulos, A. E., & Saravacos, G. D. (1995). Airdrying of osmotically dehydrated fruits. Drying Technology, 13(5-7), 1503–1521. Kowalska, H., & Lenart, A. (2001). Mass exchange during osmotic pretreatment of vegetables. Journal of Food Engineering, 49(2-3), 137–140. Kumar, H. S. P., Radhakrishna, K., Nagaraju, P. K., & Rao, D. V. (2001). Effect of combination drying on the physico-chemical characteristics of carrot and pumpkin. Journal of Food Processing Preservation, 25, 447–460. Lenart, A. (1996). Osmo-convective drying of fruits and vegetables: technology and application. Drying Technology, 14(2), 391–413. Marquard, D. W. (1959). Solution of nonlinear chemical engineering models. Chemical Engineering Progress, 55(6), 65–70. Mauro, M. A., Tavares, D. Q., & Menegalli, F. C. (2003). Behavior of plant tissue in osmotic solutions. Journal of Food Engineering, 56, 1–15. Mayor, L., Moreira, R., Chenlo, F., & Sereno, A. M. (2006). Kinetics of osmotic dehydration of pumpkin with sodium chloride solutions. Journal of Food Engineering, 74(2), 253–262. Mujumdar, A. S. (1997). Drying fundamentals. In C. G. J. Baker (Ed.), Industrial Drying of Foods (pp. 7–30). London: Blackie Academic and Professional. Murkovic, M., Mülleder, U., & Neunteufl, H. (2002). Carotenoid content in different varieties of pumpkins. Journal of Food Composition and Analysis, 15, 633–638. Nsonzi, F., & Ramaswamy, S. (1998). Quality evaluation of osmoconvective dried blueberries. Drying Technology, 16(3-5), 705–723. Pan, Y. K., Zhao, L. J., Zhang, Y., Chen, G., & Mujumdar, A. S. (2003). Osmotic dehydration pretreatment in drying of fruits and vegetables. Drying Technology, 21(6), 1101–1114. Papadakis, S. E., Bahu, R. E., Mckenzie, K. A., & Kemp, I. C. (1993). Correlations for the equilibrium moisture content of solids. Drying Technology, 11(3), 543–553. C.C. Garcia et al. / Journal of Food Engineering 82 (2007) 284–291 Park, K. J., Bin, A., & Brod, F. P. R. (2002). Drying of pear d’Anjou with and without osmotic dehydration. Journal of Food Engineering, 56, 97–103. Perry, R. H., & Chilton, C. H. (1973). Chemical Engineers’ Handbook (5th ed.). New York: McGraw Hill. Ponting, J. D. (1973). Osmotic dehydration of fruits – recent modification and applications. Process Biochemistry, 8, 18–20. Rahman, S., & Lamb, J. (1991). Air drying behavior of fresh and osmotically dehydrated pineapple. Journal of Food Process Engineering, 14, 163–171. Ramallo, L. A., Schvezov, C., & Mascheroni, R. H. (2004). Mass transfer during osmotic dehydration of pineapple. Food Science Technology International, 10(5), 323–332. Rodriguez-Amaya, D. B. (2002). Effects of processing and storage on food carotenoids. Sight and Life Newsletter, 3, 25–35. 291 Rovedo, C. O., Suarez, C., & Viollaz, P. E. (1997). Kinetics of forced convective air drying of potato and squash slabs. Food Science and Technology International, 3, 251–261. Shi, J., Le Maguer, M., Kakuda, Y., Liptay, A., & Niekamp, F. (1999). Lycopene degradation and isomerization in tomato dehydration. Food Research International, 32, 15–21. Simal, S., Deya, E., Frau, M., & Rossello, C. (1997). Simple modelling of air drying curves of fresh and osmotically pre-dehydrated apple cubes. Journal of Food Engineering, 33, 139–150. Vaccarezza, L. M., & Chirife, J. (1975). On the mechanism of moisture transport during air drying of sugar beet root. Journal of Food Science, 40, 1286–1289. William, H. (1970). Official Methods of Analysis (11th ed.). Washington DC: Association of Official Analytical Chemists Inc.