Production of mango powder by spray drying and cast ...

Powder Technology 305 (2017) 447–454

Contents lists available at ScienceDirect

Powder Technology journal homepage: www.elsevier.com/locate/powtec

Production of mango powder by spray drying and cast-tape drying Marta Fernanda Zotarelli a,c,⁎, Vanessa Martins da Silva b, Angelise Durigon a, Miriam Dupas Hubinger b, João Borges Laurindo a a b c

Department of Chemical and Food Engineering, Federal University of Santa Catarina, EQA/CTC/UFSC, 88040-900 Florianópolis, SC, Brazil Faculty of Food Engineering, State University of Campinas, P.O. Box 6121, 13083-862 Campinas, SP, Brazil Faculty of Chemical Engineering, Federal University of Uberlândia, 38703-000 Patos de Minas, MG, Brazil

a r t i c l e

i n f o

Article history: Received 3 June 2016 Received in revised form 9 October 2016 Accepted 13 October 2016 Available online 14 October 2016 Keywords: Mango Powder Properties Spray drying Cast-tape drying

a b s t r a c t The production of mango powder by spray drying and cast-tape drying, with and without the addition of maltodextrin was investigated. Moisture, particle size distribution, bulk density, particle density, porosity, morphology, total carotenoids content, water sorption isotherms, glass transition temperature and color of mango powders from both drying processes were compared. Powders resulting from cast-tape drying had irregular structure, different from the spherical structures showed by powders produced by spray drying. Cast-tape drying process resulted in powders with bulk densities of 0.8 g cm−3 (with maltodextrin) and 0.7 g cm−3 (without maltodextrin), higher than the observed for analogous powders produced by spray drying (bulk densities of 0.45 and 0.5 g cm−3). Also, porosity of powders from cast-tape drying (below 60%) was lower than that of powders produced by spray drying. Mango powders produced by spray drying without maltodextrin showed the highest carotenoid concentration (113 μm of carotenoid g−1 of dry mass). The state diagrams show that mango powders produced by spray drying exhibit slightly lower stability than those produced by cast-tape drying. Cast-tape drying is a suitable procedure for the production of mango powders and allows producing powders from whole fruit pulp, without the addition of maltodextrin. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Mangoes are appreciated worldwide for its flavor and color, besides its nutritional value. According to estimates of the Food and Agriculture Organization of the United Nations, the world production of this fruit was approximately 42.1 million tons in 2012, concentrated in tropical regions of Asia and Latin America [1]. Mangoes are quite perishable and susceptible to injuries, requiring care during storage and commercialization [2]. Industrial processing of mango fruits can improve commercialization and consumption of this fruit as puree (or fruit pulp) and powdered products. The development of processes that result in new products, preserving nutritional and some fruit sensory characteristics is important to create alternatives for adding value and helping in reducing post-harvest losses. The production of mango dehydrated powder from fruit pulp is an alternative still underused, but with potential for generating intermediate processing products (business to business products), and even products to be marketed in retail stores to regular consumers. Fruit powder ingredients are convenient for the development of other industrial products and have lower transport and storage costs. ⁎ Corresponding author at: Department of Chemical and Food Engineering, Federal University of Santa Catarina, EQA/CTC/UFSC, 88040-900 Florianópolis, SC, Brazil. E-mail address: [email protected] (M.F. Zotarelli).

http://dx.doi.org/10.1016/j.powtec.2016.10.027 0032-5910/© 2016 Elsevier B.V. All rights reserved.

However, the production of high quality fruit powder must overcome some difficulties, as the hygroscopicity and the sticking characteristic of this high sugar content material. Literature reports the problems faced to produce fruit powders by spray drying (SD). Ripe fruits are rich in sugars and organic acids of low molecular weight that have low glass transition temperatures (Tg), at which the amorphous polymer undergoes a phase transition from the glass state to a rubbery state [3]. When they are dehydrated at temperatures above their Tg, the stickiness characteristic appears and causes their adhesion to the dryer walls, reducing yield and quality of the final product [4,5]. The addition of carrier agents to fruit juices (compounds with higher molecular weight) prior to drying process increases the Tg of the mixture and has been widely used to reduce fruit pulp stickiness [6,3,7,8]. The literature reports studies on the production of mango powder by spray-drying, using carrier agents added to the mango pulp to be dried. Cano-Chauca et al. [7] reported the use of maltodextrin, Arabic gum and waxy starch at concentration of 12%. The authors also evaluated the effect of cellulose solution addition at different concentrations (0, 3, 6 and 9%) on the drying kinetics and powder characteristics. The addition of cellulose led to the formation of less sticky and less soluble crystalline particles. Cano-Higuita et al. [8] evaluated the influence of the addition of maltodextrin, skim milk and of a mixture thereof in the sorption isotherms of the mango powders produced by SD. They used ratios of 1:8,

448

M.F. Zotarelli et al. / Powder Technology 305 (2017) 447–454

3:6 (dry basis) of mango pulp for maltodextrin and 1:4:4 of mango pulp for maltodextrin and skimmed milk. From the results of water sorption isotherm at 20 °C they reported that the formulation 1:8 was the less hygroscopic combination. Alternative drying processes have been investigated for fruit powder production. Cast-Tape Drying (CTD) is a method suitable for food powder production. In fact, the drying process that has been reported as refractance window (RW) is a particular case of CTD. In this process, the solution to be dried is spread on a transparent polyester film, commercially known as Mylar (DuPont®), while its lower surface is in contact with hot water, which provides the heat for the product drying [9]. CTD applied to fruit and vegetable pulps results in films or flakes, but powdered products can be produced by grinding. An advantage of this drying process is the possibility of using moderate drying temperatures (60–80 °C) and relatively short drying times (some minutes) by selecting the thickness of the pulp layer to be dried. The use of this process for drying pulps of strawberry and carrot [10], pumpkin [11], açaí [12], mango [2,13,14] and tomato [15] has been reported in the literature. Caparino et al. [2] evaluated the production of mango powder by different drying processes, i.e. refractance window (RW), freeze-drying (FD) and drum drying (DD), without the addition of carrier agents, and spray-drying (SD) with the addition of 0.25 kg of maltodextrin (10DE) per kg of dried mango solids. The physical properties of the resulted powders were compared. The authors reported that mango powders produced by RW without the addition of carrier agents showed better quality than those produced by SD and DD. Caparino et al. [13] reported moisture sorption, glass transition temperatures and microstructure data of mango particles, and their influence on the stability of powders produced by RW and FD. They observed similar water sorption properties from powders produced by both processes, with type III isotherms [16]. CTD process can be applied to fruit pulps without carrier agents. A fair comparison between CTD and SD processes and resulting products could be better if performed from processes without addition of carrier agents. However, studies on the SD of mango pulp without the addition of carrier agents were not found in the literature. The aim of this study was to investigate the use of spray drying and of cast-tape drying for producing powders from mango pulp with and without maltodextrin. The physicochemical properties of powders from both processes were compared.

2.3. Drying processes Drying of mango pulp without maltodextrin was carried out using spray drying (SD) and cast-tape drying (CTD) processes. Each kilogram of mango pulp with moisture of 5.3 g g−1 was mixed with 0.06 kg of maltodextrin (moisture 16.3 g g−1) and 0.187 kg of water, in order to have a mixture with moisture of 4.7 g g−1 before drying. The drying processes performed with mango pulp were named CTD-P and SD-P, while those processes with pulp added of maltodextrin were named as SD-M and CTD-M. 2.3.1. Spray drying A lab scale spray dryer (Buchi, B-290, Switzerland) with dehumidifier (B-296, Switzerland), was used for the production of mango powder with maltodextrin (SD-M) and without this carrier agent (SD-P). The equipment has a drying chamber (0.65 m × 1.10 m × 0.70 m) and a dual fluid nozzle atomizer type with orifice 0.7 mm in diameter and evaporation capacity of 1 L h−1. A peristaltic pump was used for feeding the dryer with the mango pulp, at a flow rate of 0.42 L h−1. The drying air flow was 35 m3 h−1 with inlet temperature of 150 °C. The outlet air temperature was monitored to observe its variation as a function of the pulp feeding. 2.3.2. Cast-tape drying A CTD lab-scale dryer was made using the same principle present in industrial equipment [9,11,14,17]. It consisted of a hot water reservoir (0.8 m × 0.4 m × 0.05 m) and a thermostatic bath (Quimis, model Q214S, Diadema, SP, Brazil) operating in a closed system. A polyester film (Mylar® type D, DuPont, Wilmington, DE, USA), with 0.25 mm thickness, was fixed on the top of the reservoir frame, ensuring that the whole surface of film's bottom touched the hot water, while its top face was the support onto which the mango pulp was spread. Two exhaust fans provided the airflow over the pulp. The thickness of the Mylar film was chosen based on literature reports [11,14]. The mango pulp was spread over the Mylar film with the aid of a doctor blade, which allows the thickness of the spread pulp to be adjusted. Mango pulp and hot water temperatures were determined by T-type thermocouples (IOPE, model TF-TX-AR-30AWG, Brazil) connected to a data acquisition system (Agilent model 34970A, Malaysia). Water temperature was adjusted to 95 °C. The dried mango pulp was removed as films and flakes and milled in a Wiley type mill (TECNAL, ET 631/2 model, Brazil) during 2 min at 15,000 rpm.

2. Material and methods

2.4. Powder characterization

2.1. Mango fruits

2.4.1. Moisture and water activity During drying, the evolution of the pulp moisture was determined by gravimetric method, using a vacuum oven (TECNAL, TE-395 model, Brazil) at 70 °C [18], from samples taken from the drier. The water activity (aw) was determined in triplicate, using a water activity meter (Aqualab, Decagon Devices, USA).

Tommy Atkins mangoes used in this study were purchased in the local market, in Campinas-SP, Brazil. Fruits were selected according to their degree of ripeness, assessed from visual inspection and soluble solids concentration, determined with a manual refractometer with a 0–90°Brix range and 0.2°Brix resolution (Reichert, Model AR200, USA). The fruits were washed, peeled by hand and ground in a household blender (Arno, São Paulo, Brazil) for obtaining the mango pulp. Before the drying process, the pulp was sieved using a 16 mesh sieve, in order to retain large particles and fibers. Pulp soluble solids content was 17°Brix. Mango pulp was fractionated in 1 kg polyethylene containers and frozen until use.

2.4.2. Hygroscopicity of the mango powders The hygroscopicity of the mango powders was determined according to the methodology proposed by Cai and Corke [19], with modifications [20]. Approximately 1 g of dried sample was placed in a hermetic container at 25 °C, with a saturated NaCl solution, for creating a 75.3% relative humidity atmosphere. The samples were weighed after seven days and the hygroscopicity was determined and expressed in g of adsorbed water per 100 g of dry solid (100 g g−1).

2.2. Carrier agent Maltodextrin MOR-Rex® 1910 (10DE) from Ingredion (Mogi GuaçuSP, Brazil) was used as carrier agent, which is widely used for drying fruit juices by spray drying. Maltodextrin was added to the filtered pulp until complete dissolution.

2.4.3. Total carotenoids Total carotenoids of fresh mango pulps and mango powders were determined according to the methodology described by RodriguezAmaya [21]. This methodology comprehends the extraction of carotenoids from the samples with acetone, followed by separation with


petroleum ether. The diluted solution is submitted to spectrophotometric measurements (UNICO, SQ-2800 UV/VIS, United Products & Instruments Inc., USA) in wavelength of 450 nm, using petroleum ether as the control. The result, expressed in terms of total carotenoids (μg g−1), excluding the mass of the carrier agent, was calculated by Eq. (1): CAR ¼

Abs V 106 A1% 1cm m 100

ð1Þ

in which V is the volume of dilution (mL), Abs is the maximum measured absorbance, A1% 1cm is the absorptivity for the predominant carotenoid in the petroleum ether (β-carotene, in the present case) and m is the sample weight (g). 2.4.4. Particle size distribution The particle size distribution was determined by a particle analyzer based on laser diffraction (Mastersizer, Mastersizer 2000, Malvern Instruments, UK). The average diameter was determined from the diameter of a sphere with equivalent volume, i.e. the Brouckere diameter, given by Eq. (2). In this equation, D[4,3] is the particle diameter and n is the number of particles. Samples were analyzed in sextuplicate with dispersion in 99.5% ethanol. n X

D½4;3 ¼

ð2Þ 3 ni di

i¼1

2.4.5. Bulk density, particle density and porosity The apparent (bulk) density of mango particles in a packed bed was determined for powders from both, SD and CTD drying processes, using the method described by Barbosa-Cánovas et al. [22] and Goula et al. [23]. Five grams of powder was placed in a graduated beaker (1 mL), which was repeatedly hit on a flat surface, until negligible differences of the bed height between successive beats were observed. From the sample weight and the bed volume in the tube, the bulk density was calculated and expressed in kg m−3. The measurements were performed in triplicate. The absolute volume of each sample powder was determined using an air-based pycnometer, according to the methodology described by Sereno, Silva and Mayor [24]. The air pycnometer is based on the measurement of the pressure change experienced by a pressurized gas that fills a reservoir with known volume, when an expansion occurs to a second container (known volume) containing the sample. From the sample weight and the container volumes, the particles volume is determined and used for calculating the particles density. The porosity of the particles bed was calculated from the particle density (ρp) and the bulk density (ρb), as given by Eq. (3), ε¼

ρp −ρb ρp

10° (RSEX mode). Color measurements were expressed in terms of the brightness L* (L* = 0 for black and L* = 100 for white) and chromaticity, defined by a* (+a* for red and − a* for green) and b* (+b* for yellow and −b* for blue). 2.4.8. Moisture sorption isotherms Moisture sorption isotherms were determined using the static method. Nine saturated salt solutions were prepared (LiCl, CH3COOK, K2CO3, Mg(NO3)2, KI, NaCl, KCl e BaCl2) for relative humidity values of 11.3%, 22.6%, 43.2%, 52.9%, 68.9%, 75.3%, 84.3% and 90.2%, respectively [25]. Samples of approximately 2 g of mango powder were weighed in plastic capsules and inserted in the desiccators at 25 °C, with salt solutions for every desired relative humidity. The samples were weighed by analytical balance at regular intervals until they reached the equilibrium. The GAB model (Guggenheim-Anderson-de Boer, Eq. (4)) was fitted to the sorption data of mango powder samples. X¼

ðC−1ÞKaw X m Kaw X m þ 1 þ ðC−1ÞKaw 1−Kaw

ð4Þ

in which Xm is the dry basis moisture of the monolayer, and C and K are the other GAB model constants. These parameters were estimated by nonlinear regression using MATLAB software (R2010a).

4

ni di

i¼1

n X

449

ð3Þ

2.4.6. Optical micrographs Powder samples were placed on thin glass plates and dispersed in glycerol. The suspensions were covered with glass slides and observed in an optical microscope (Carl Zeiss, model Axio Scope A1, Gottingen, Germany) with magnification of 400 times. 2.4.7. Samples color The color parameters of the powder samples were determined by reflectance, using a colorimeter Ultra Scan Vis 1043 (Hunter Lab, Reston, Virginia, USA) with CIELab scale (L*, a*, b*). The measurements were performed at 25 °C, using D-65 illuminant and an observation angle of

2.4.9. Glass transition temperature For the determination of the glass transition temperature, the powder samples were placed in aluminum capsules and inserted in the same desiccators with saturated salt solutions used for the sorption isotherms, kept at 25 °C. After reaching equilibrium, the capsules were hermetically sealed, weighed and used for the DSC analyses. A TA-MDSC2920 calorimeter (TA Instruments, New Castle, EUA) (TA Instruments, New Castle, USA), cooled by a chiller (RCS Refrigerated Cooling Accessory) operating with nitrogen gas at 150 mL min−1 was used. The calibration of the equipment was performed with Indium (Tf = 156.6 °C), using helium as the purge gas, with a constant flow of 25 mL min−1. The samples were cooled to −70 °C and then heated to 120 °C at a constant rate of 10 °C min−1. Two runs were performed to reduce the enthalpy relaxation of amorphous powders. All the runs were performed in triplicate and the experimental data were treated using the software Universal Analysis 2.6 (TA Instruments, New Castle, USA). The plasticizing effect of water and its influence on the glass transition temperature was described using the Gordon-Taylor model [26], Eq. (5): Tg ¼

ws T gs þ kww T gw ws þ kww

ð5Þ

in which Tgs and Tgw are the glass transition temperatures of the solid mixture and water (°C), respectively, while ws and ww are the mass fractions of solids and water in the mixture (g g−1), and k is the constant of the Gordon-Taylor model. The Tgw was assumed to be −135 °C [27]. The parameters of Eq. (5) were estimated by non-linear regression of the experimental data, using the Curvefit Toolbox, from MATLAB (R2010a). The goodness of fit of the Gordon-Taylor model to the experimental data was evaluated using the coefficient of determination (R2) and the root mean squared error (RMSE). 2.5. Statistical analysis The software Statistica 7.1 (Statsoft Inc., Tulsa, USA) was used to perform the statistical analyses of experimental results. A Chi-Squared test was used to determine whether there was a significant difference between the drying methods. As significant differences between drying processes were observed, the data were compared using a Student ttest at 90% confidence level, (p b 0.10), decision justified by the raw material variability.

450


3. Results and discussion 3.1. Drying curves, temperature, water activity and hygroscopicity of mango powders Table 1 shows the results of prevailing temperature during pulp drying, along with moisture (g g−1), water activity and hygroscopicity of powders resulted from both drying methods, with and without addition of maltodextrin to the mango pulp. Data displayed in Table 1 show that mango powders produced by the different drying processes presented moisture lower than 5% and water activity b0.25. Mango powders produced by CTD-P had significantly lower hygroscopicity than those produced by SD-P. It is reported in the literature that the main factors that should influence the absorption of water from the surrounding are: higher temperatures usually produce powders with high hygroscopicity, and a small variation of moisture content of the samples also can influence in the hygroscopicity analysis [2,20]. During the spray drying of mango pulp without or with maltodextrin, part of the mango pulp adhered to the dryer wall, and this fact is related to the high temperature. Nevertheless, the data from Table 1 show that maltodextrin addition reduced the mango powders hygroscopicity. Caparino et al. [2] reported hygroscopicity values slightly lower than those observed in this work. These authors reported hygroscopicity of 16.5% for mango powder produced by spray drying, with 25% maltodextrin. They also reported that mango pulps dehydrated by refractance window, freeze-drying and drum drying resulted in powders with hygroscopicities of 18.0, 18.0 e 20.1%, respectively. According to Schuck, Jeantet and Dolivet [28], powders with hygroscopicity between 15.1 and 20% (at 75% of RH) are highly hygroscopic. The differences between the results of Caparino et al. [2] and the observed in the present study can be due to differences of particle size and pulp sugar content (14–15°Brix), which can greatly influence powder hygroscopicity [28]. 3.2. Size distribution of mango powders The size distributions of the particles from mango powders produced by CTD-P and SD-P, and by CTD-M and SD-M varied widely, as shown in Fig. 1. Powders produced with maltodextrin by SD-M ranged from 0.47 to 549 μm, while those produced by CTD-M (and ground) ranged from 15 to 2188 μm. Mango powders produced without maltodextrin by SD-P presented sizes ranging between 1.9 and 955 μm, while those produced by CTD-P showed sizes between 1.25 and 831 μm. The size distributions of powders produced by spray drying, with and without maltodextrin, exhibited bimodal behavior. Similar results were reported by Tonon, Brabet and Hubinger [20] for açaí powder produced by spray drying with different concentrations of maltodextrin. Ferrari et al. [29] also reported bimodal sizes distribution for blackberry powders produced by spray drying with maltodextrin (20DE), gum Arabic and a mix of these two carrier agents. This can be attributed to the formation of bridges between the smallest particles, resulting in agglomeration that formed larger particles. Powders obtained from the CTD process (with or without maltodextrin) were subjected to a grinding process, whose operational conditions, rotation and grinding time, strongly influenced particle size distribution. In this study, grinding always resulted in powders with

Fig. 1. Particles size distribution of mango powders produced by the different drying processes: (\ \) cast-tape drying (CTD-P), (—) cast-tape drying with addition of maltodextrin (10DE) - (CTD-M), (\ \) spray drying (SD-P) and (—) spray drying with addition of maltodextrin (10DE) - (SD-M).

unimodal size distribution. Pavan, Schmidt and Feng [12] reported particle size distribution of açaí pulp dried by refractance window and by hot air, and processed in a coffee grinder. After, grinding, the powders produced by both processes had unimodal distribution, with sizes ranging from 425 to 600 μm. Table 2 shows the values of the average diameter D[4,3] of the particles resulted from the different processes. The average size of particles of mango powder without maltodextrin produced by cast-tape drying and spray drying showed no significant difference, contrary to that observed for the powders produced with maltodextrin. In this case, the average size of particles of mango powders with maltodextrin produced by CTD-M was four times bigger than the observed to SD-M. It is important to remark that the values of the average sizes of the particles shown in Table 1 may be influenced by particle agglomeration [20,30]. The grinder type, its operational conditions and the grinding time can also change particle size distribution. 3.3. Bulk density, bulk porosity and particle density Particle density and porosity are important physical properties for the stability of food powders during storage. Powders with high porosity and low particle density have many empty spaces with greater presence of oxygen, which can trigger oxidation reactions. Besides, powders of high particle density and low porosity have smaller volume. Mango powders produced by cast-tape drying (CTD-P and CTD-M processes) had higher bulk densities and lower porosities, when compared to mango powders produced by spray drying (SD-P and SD-M processes) (Table 2). Caparino et al. [2] have also reported that smaller particle sizes of mango powder resulted in larger bulk densities and smaller porosities. A part from particle sizes, CTD, refractance window and drumdrying can be categorized as direct drying techniques and this may have caused collapse which resulted in more compact and rigid product reflecting in lower bulk densities [2,9]. In direct drying processes, in which the product remains in contact with air at high temperatures, for long times, collapse of the solid structure may occur, resulting in compact particles (drying shrinking). Caparino et al. [2] reported that

Table 1 Drying temperature, moisture content, water activity and hygroscopicity of mango powders. Drying process

Temperature (°C)

Moisture content (%)

Water activity

Hygroscopicity (%)

CTD-P CTD-M SD-P SD-M

76 ± 3 75 ± 2 150 ± 2 150 ± 1

3.5 4.7 1.2 1.5

0.24 0.22 0.18 0.18

25.4 18.8 26.9 23.9

± ± ± ±

0.1 0.1 0.1 0.1

± ± ± ±

0.01 0.00 0.01 0.01

Values followed by different lowercase letters indicate significant difference (p b 0.10) between hygroscopicities observed for CTD-P, CTD-M, SD-P and SD-M.

± ± ± ±

0.4d 0.3a 0.5c 0.3b

M.F. Zotarelli et al. / Powder Technology 305 (2017) 447–454 Table 2 Mean diameter D[4,3], bulk density, particle density and porosity mango powders produced by the different processes. Drying process

Mean diameter D[4,3] (μm)

Bulk density (g cm−3)

CTD-P CTD-M SD-P SD-M

196 ± 3b 322 ± 5a 198 ± 1b 81 ± 1c

0.80 0.70 0.45 0.50

± 0.01a ± 0.01b ± 0.01d = ±0.01c

Particle density (g cm−3) 1.28 1.59 1.80 1.64

± ± ± ±

0.01d 0.03c 0.01a 0.02b

Porosity (%) 38% 55% 75% 70%

Values followed by different lowercase letters indicate significant difference (p b 0.10) between bulk density and particle density, observed for CTD-P, CTD-M, SD-P and SD-M.

mango powders produced by refractance window and drum drying (followed by a grinding process) showed higher densities and lower porosities, when compared to powders produced by freeze-drying (followed by a grinding process) and spray drying, with similar values to those found in this study. They reported porosity values ranging from 10 to 30% for powders produced by refractance window, and b10% for powders produced by drum drying. The powders produced by freeze-drying and spray-drying showed porosity values of approximately 50%, similar to the results found in the present study. Again, it is important to remark that particle size distribution depends on the grinding process. 3.4. Powder morphology Fig. 2 shows optical micrographs of mango powder samples produced by the different processes. Particles produced by spray drying are spherical, characteristic of this process (Fig. 2c and d), while images of particles of CTD (ground) show more irregular shapes (Fig. 2a and b). Moreover, some filaments on the surfaces of these particles (indicated by arrows) are evident. These filaments suggest the presence of pulp fibers, showing that cast-tape drying is a suitable process for drying mango pulp with its fibers. It is well known that high amounts of fibers in the pulp can lead to the clogging of the spray nozzle.

451

3.5. Color parameters of mango powders Table 3 lists the color parameters L*, a* and b* of mango powder samples produced by the different drying processes. Powders produced by cast-tape drying, had a lower brightness (lower L*), and higher values of a*, indicating the prevalence of red color. For mango, whose predominant color is yellow, the parameter b* should be used to distinguish the color changes resulting from the drying process [2]. In this case, the most intense b* value was observed for mango powders produced by spray drying without maltodextrin. No significant differences were observed (p ≤ 0.10) for the parameter b*, for mango powders produced by cast-tape drying, with and without maltodextrin (CTD-P and CTD-M processes). Mango powder with maltodextrin, dried by spray drying, showed the lowest value of b*, indeed related to the process conditions and to the addition of maltodextrin, which dilutes the powder color. In the case of mango powder added of maltodextrin and dried by cast-tape drying (CTD-M), the effect of the carrier agent is not so evident, maybe due to the lower drying temperature. Exposing the product to high temperatures can trigger darkening caused by chemical reactions between sugars [31]. Caparino et al. [2] evaluated the color parameters of mango powders (var. Carabao) produced by refractance window, freeze drying, drum drying and spray drying. They reported that particles produced by refractance window and freeze drying, with sizes between 180 and 500 μm, had b* values between 40 and 50, slightly lower than the values observed in this study. The powders from spray drying were “less yellow”, with b* values below 40. Regarding the brightness, the authors pointed out that drum drying resulted in the darkest powders, while spray drying produced the lightest powders. The powders produced by refractance window had L* values near 60, but this value is affected by the particle sizes. Color differences between the reports of Caparino et al. [2] and the observed in the present study could be explained by the different varieties of mango used in the both investigations. Moreover, one should consider the influence of process conditions on the characteristics of the resulting particles. The aforementioned authors

Fig. 2. Optical micrographs (400×) of mango powders produced by the different processes: (a) cast-tape drying with addition of maltodextrin (10DE) - (CTD-M), (b) cast-tape drying (CTD-P), (c) spray drying with addition of maltodextrin (10DE) - (SD-M), (d) spray drying (SD-P).

452


Table 3 Color parameters, total carotenoid content (CAR) and total carotenoid content loss (CAR_LOSS) of mango powders produced by the different processes. Drying process

L*

Mango pulp CTD-P CTD-M SD-P SD-M

42.37 73.03 75.13 80.33 87.71

a* ± ± ± ± ±

0.21 0.72d 0.44c 0.56b 0.22a

2.71 9.55 8.71 5.99 1.51

b* ± ± ± ± ±

0.07 0.61a 0.40b 0.33c 0.15d

44.12 47.58 47.85 50.34 32.24

± 1.02 ± 1.72b ± 1.02b ± 0.34a ± 0.65c

CAR (μg g−1)

CAR_LOSS (%)

281 ± 4 94 ± 1c 81 ± 2d 113 ± 1a 98.77 ± 2

– 66 71 60 65

Values followed by different lowercase letters indicate significant difference (p b 0.10) between color parameters, observed for CTD-P, CTD-M, SD-P and SD-M.

reported drying times of about 3 min for mango pulp processed by RW, whereas in the present study the drying time was approximately 12 min, due to the higher pulp thickness. 3.6. Total carotenoids in the mango powders Table 3 shows the carotenoids concentrations and the loss of carotenoids occurred in each drying process, compared to the concentration in the mango pulp, which was 280.94 μg g−1 dry solid or 44.38 μg of carotenoids g−1 of pulp. Mercadante, Rodriguez-Amaya and Britton [32] reported total carotenoid values of mango pulp (Keitt variety) ranging from 49.9 to 55.0 μg of carotenoids g−1 of pulp, which are close to the observed in this study. The literature reports that the loss of β-carotene in mango puree processing is approximately 13% [21]. Mango powder produced by SD-P showed the highest carotenoid concentration (113.2 μg g−1), followed by powders produced by SDM (98.7 μg g− 1). As carotenoids impart a yellow color to the mango pulp, it is interesting to assess the relationship between its concentration and the values of the parameter b* (Table 3). Pulp samples without maltodextrin (no effect of color dilution), dried by spray drying, showed larger values of b* and presented the largest amount of carotenoids. The higher carotenoids degradation during CTD-P drying can be a consequence of the longer drying time. Residence time in cast-tape drying was about 12 min while in spray drying it was much smaller, around 3 s, as reported by Caparino et al. [2]. The high sensitivity of carotenoids to oxidation, due to their number of conjugated double bonds [33,34], suggests that it is important to take care of the processing temperature and the residence time. The reduction of the pulp drying time in CTD-P processes is achieved by reducing the pulp thickness. 3.7. Water sorption isotherms The GAB model fitted well the experimental data of moisture sorption of all the mango powders (R2 N0.998 and RMSE b 0.01). Table 4 shows the estimated parameters. The GAB model allows estimating the molecular moisture monolayer (Xm), which can be used to estimate the moisture for adequate storing of dehydrated foods. The Xm estimated for mango powders were approximately 0.11 g g−1 for mango

powders with maltodextrin (CTD-M and SD-M) and 0.09 g g−1for powders without maltodextrin (CDT-P and SD-P). These values are in agreement with those reported by Caparino et al. [13], who reported monolayer moisture values of 0.078 g g−1 for mango powder produced by refractance window and 0.045 g g−1 for mango powder produced by freeze-drying. Fig. 3 depicts the moisture sorption isotherms (mean of three repetitions) of mango powders, fitted by the GAB model. The curves showed a similar behavior and are type III isotherms [16]. Mango powders with maltodextrin, produced by SD-M and CTD-M processes, showed slight lower moisture sorption, due to the chemical structure of maltodextrin, which has fewer hydrophilic groups [36]. Caparino et al. [13] also reported type III isotherms for mango powders (without carrier agents) dried by both refractance window and freeze-drying processes. Castoldi et al. [15] reported the same behavior for tomato powder produced by cast-tape drying (without carrier agents). Fig. 3 shows that all the powder samples showed similar behavior at relative humidities lower than 43%. At this condition, there was little particle agglomeration. At intermediate humidity, particle agglomeration was observed, forming a more compact and rigid structure, with darker color. At relative humidities of 84%–90%, the formation of exudate with viscous appearance was observed. Under these conditions, the bridges that connect the particles come apart as result of solubilization of low molecular weight fractions [37]. Similar behavior during the storage of açaí powder at different relative humidities was also reported by Tonon et al. [36]. 3.8. Glass transition temperature The relation between glass transition temperature and solids content of mango powders produced by the different drying processes is presented in Fig. 4. The Gordon-Taylor model fitted the experimental data and is represented in the same figure, while the model parameters are shown in Table 4. The values of the parameter K ranged between 3.47 and 4.68, similar to values reported by literature to açaí powder [36] and camu-camu

Table 4 GAB parameters estimated from moisture sorption data and Gordon-Taylor parameters estimated from experimental glass transition temperatures determined for mango powder produced by the different processes. Drying process Parameters C K Xm (d.b.) RMSE R2 Tgs (°C) K R2 RMSE

CTD-P

CTD-M

SD-P

SD-M

0.968 0.966 0.099 0.009 0.998 28.26 3.88 0.966 7.174

0.994 0.907 0.115 0.007 0.998 46.09 4.68 0.979 5.971

1.287 0.979 0.091 0.008 0.999 26.74 3.47 0.937 9.136

0.721 0.920 0.116 0.008 0.998 32.38 3.79 0.977 5.454

d.b. - dry basis (g water g−1 dry solids).

Fig. 3. Water sorption isotherms of mango powders produced by the different processes: (□) cast-tape drying (CTD-P), (■) cast-tape drying with maltodextrin (CTD-M), (○) spray drying (SD-P) and (●) spray drying with maltodextrin (SD-M).


Fig. 4. Influence of solids content on the glass transition temperature of mango powders produced by the different processes: (□) cast-tape drying (CTD-P), (▲) cast-tape drying with maltodextrin (CTD-M), (○) spray drying (SD-P) and (*) spray drying with maltodextrin (SD-M).

[38]. In a binary system, this parameter relates the glass transition temperature to the food moisture or water activity [26]. The Tg values determined in this study were lower than those reported in the literature. The powders produced by SD-P and by CTD-P showed Tg values of 26.74 °C and 28.26 °C respectively, while powders with maltodextrin showed values between 32.4 °C (SD-M) and 46.1 °C (CTD-M). These low Tg values are explained from the large concentration of sugars in mango pulp, particularly sucrose, glucose and fructose, besides organic acids, which have low Tg values [13,39]. According to data reported by Roos and Karel [40] the glass transition temperature of glucose is 31 °C, fructose is 5 °C and sucrose is 62 °C. The low Tg values observed for mango pulp in the preset study reinforce the well-known difficulty of drying this type of material by spray drying without addition of carrier agents [36]. The increase of pulp's Tg, promoted by carrier agents of high molecular mass, is crucial for applying spray drying to high sugar content pulps and juices. To determine the critical moisture and water activity related to the glass transition temperature of the powder samples, data of Tg, aw and moisture sorption isotherms (GAB) are shown along with the Gordon-

453

Taylor model, in Fig. 5. This state diagram is useful for the prediction of the moisture value and the critical water activity below which the product remains in the glassy state [3]. Low critical moistures were found for powdered mango. At 20 °C the mango powders from all the drying processes should have moisture content inferior to 0.01 g g−1 to be stable. Caparino et al. [13] reported that mango powder produced by refractance window, with moisture of 0.017 g g−1 (wet basis) is stable when stored at temperatures up to 23 °C. Regarding the water activity, powders produced by cast-tape drying, with and without maltodextrin, presented critical water activity values ranging from 0.25 to 0.30, which are slightly higher than the values estimated for the powders produced by spray drying. These powders had critical moisture close to 0.19 g g−1 when produced with maltodextrin, and 0.12 g g−1 for the sample without maltodextrin. These results are in agreement with the hygroscopicity values presented in Table 1, wherein the powders produced by tape-cast drying were less hygroscopic. Similar results were observed by Mosquera, Moraga and MartínezNavarrete [41] for pulp and freeze-dried strawberries without carrier agents. At room temperature of 20 °C, the critical water activity was 0.094, indicating that the maximum relative humidity to which the product can be exposed to stay in the glassy state is 9.4%. Moraga, Martínez-Navarrete and Chiralt [42,43] also reported low critical moisture values for freeze-dried kiwi and strawberry, without using carrier agents. The authors reported that, at 30 °C, the critical moisture was 1.4 g g−1, while the water activity was 0.031, indicating the poor powders stability. 4. Conclusions The production of dehydrated mango powder by spray drying should be performed with addition of a carrier agent due to the sticking characteristic of this fruit pulp, which tends to adhere to the dryer wall. Cast-tape dryer allows drying mango pulp without the addition of carrier agents, but the dehydrated product looks like films or flakes that must be ground to produce powder. The use of CTD for the production of mango powders presents advantages, because these powders can be produced from the whole fruit pulp, with its fibers and without the addition of maltodextrin. In terms of cost processing, CTD presents almost 100% of yield. CTD powders present better critical storage conditions than SD powders, but maltodextrin addition is important to improve product stability of powders from both processes. In this way,

Fig. 5. Variation of the glass transition temperature (dotted line) and sorption isotherms (solid line) vs. water activity of mango powders produced by the different processes: (a) cast-tape drying, (b) cast-tape drying with maltodextrin, (c) spray dryer and (d) spray dryer with maltodextrin.

454


the results reported in the present work indicate that CTD is a competitive process for the drying of fruit and vegetable pulps. Acknowledgements The authors thank CNPq (402481/2013-5, 304475/2013-0), FAPESP (2009/54137-1) and Capes (552185/2011-6) for the financial support. References [1] Food and Agriculture Organization of the United Nations (FAO), FaoStat - World Production of Mango (Mangoes, Mangosteens, Guavas), http://faostat.fao.org/site/ 567/DesktopDefault.aspx?PageID=567#ancor2016 accessed 24.04.16. [2] O.A. Caparino, J. Tang, C.I. Nindo, S.S. Sablani, J.R. Powers, J.K. Fellman, Effect of drying methods on the physical properties and microstructures of mango (Philippine ‘Carabao’ var.) powder, J. Food Eng. 111 (2012) 135–148. [3] Y.H. Ross, Phase Transitions in Foods, first ed Academic Press, San Diego, 1995. [4] K. Masters, Spray Drying Handbook, fifth ed Longman Scientific and Technical, London, 1991. [5] B.R. Bhandari, N. Datta, T. Howes, Problems associated with spray drying of sugarrich foods, Dry. Technol. 15 (2) (1997) 671–684. [6] B.R. Bhandari, A. Senoussi, E.D. Dumoulin, A. Lebert, Spray drying of concentrated fruit juices, Dry. Technol. 11 (5) (1993) 1081–1092. [7] M. Cano-Chauca, P.C. Stringheta, A.M. Ramos, J. Cal-Vidal, Effect of the carriers on the microstructure of mango powder obtained by spray drying and its functional characterization, Innovative Food Sci. Emerg. Technol. 5 (4) (2005) 420–428. [8] D.M. Cano-Higuita, H.A. Villa-V'elez, J. Telis-Romero, H.A. V'aquiro, V.R.N. Telis, Influence of alternative drying aids on water sorption of spray dried mango mix powders: a thermodynamic approach, Food Bioprod. Process. 93 (2015) 19–28. [9] C.I. Nindo, J. Tang, J, refractance window dehydration technology: a novel contact drying method, Dry. Technol. 25 (2007) 37–48. [10] B.I. Abonyi, H. Feng, C.G. Tang, B.P. Edwards, D.S. Mattinson, J.K. Fellman, Quality retention in strawberry and carrot purees dried with refractance window system, J. Food Sci. 67 (2002) 1051–1056. [11] C.I. Nindo, H. Feng, G.Q. Shen, D.H. Kang, Energy utilization and microbial reduction in a new film drying system, J. Food Process. Preserv. 27 (2) (2003) 117–136. [12] M.A. Pavan, S.J. Schmidt, H. Feng, Water sorption behavior and thermal analysis of freeze-dried, refractance window-dried and hot-air dried açaí (Euterpe oleracea Martius) juice, Food Sci. Technol. 48 (2012) 75–81. [13] O.A. Caparino, S.S. Sablani, J. Tang, R.M. Syamaladevi, C.I. Nindo, Water sorption, glass transition, and microstructures of refractance window - and freeze-dried mango (Philippine “Carabao” var.) powder, Dry. Technol. 31 (16) (2013) 1969–1978. [14] M.F. Zotarelli, B.A.M. Carciofi, J.B. Laurindo, Effect of process variables on the drying rate of mango pulp by refractance window, Food Res. Int. 69 (2015) 410–417. [15] M. Castoldi, M.F. Zotarelli, A. Durigon, B.A.M. Carciofi, J.B. Laurindo, Production of tomato powder by refractance window drying, Dry. Technol. 33 (12) (2015) 1463–1473. [16] S.S.H. Rizvi, Thermodynamic Properties of Food in Dehydration, in: M.A. Rao, S.S.H. Rizvi, A.K. Datta (Eds.), Engineering Properties of Foods, third edCRC Press, Boca Ratton Fl 1995, pp. 359–436. [17] P.T. Clarke, Refractance WindowTM - “down under”, Proceedings of the 14th International Drying Symposium (2004) 813–820 , IDS 2004 v. B. [18] A.O.A.C. Association of Official Analytical Chemists, Official Methods of Analysis, 18th ed. Gaithersburg, Maryland, 2005.

[19] Z.Y. Cai, H. Corke, Production and properties of spray-dried Amaranthus Betacyanin pigments, J. Food Sci. 65 (6) (2000) 1248–1252. [20] R.V. Tonon, C. Brabet, M.D. Hubinger, Influence of process conditions on the physicochemical properties of açaí (Euterpe oleraceae Mart.) powder produced by spray drying, J. Food Eng. 88 (3) (2008) 411–418. [21] D.B. Rodriguez-Amaya, A Guide to Carotenoid Analysis in Foods, first ed ILSI Press, Washington, D. C, 2001. [22] G.V. Barbosa-Cánovas, E. Ortega-Rivas, P. Juliano, H. Yan, Food Powders: Physical Properties, Processing, and Functionality, first ed Kluwer Academic/Plenum Publishers, New York, 2005. [23] A.M. Goula, T.D. Karapantsios, D.S. Achilias, K.G. Adamopoulos, Water sorption isotherms and glass transition temperature of spray dried tomato pulp, J. Food Eng. 85 (2008) 73–83. [24] A.M. Sereno, M.A. Silva, L. Mayor, Determination of particle density and porosity in foods and porous materials with high moisture content, Int. J. Food Prop. 10 (3) (2007) 455–469. [25] L. Greenspan, Humidity fixed points of binary saturated aqueous solutions, Journal of Research of the National Bureau of Standards – Physics and Chemistry 81 (1) (1977) 89–96. [26] M. Gordon, J.S. Taylor, Ideal copolymers and the second-order transitions of synthetic rubbers. I. Non-crystalline copolymers, J. Appl. Chem. 2 (9) (1952) 493–500. [27] G.P. Johari, A. Hallbrucker, E. Mayer, The glass–liquid transition of hyperquenched water, Nature 330 (10) (1987) 552–553. [28] P. Schuck, A. Dolivet, R. Jeantet, Analytical methods for food and dairy powders, 186187, Wiley-Blackwell, Iowa - USA, 2012. [29] C.C. Ferrari, S.P.M. Germer, I.D. Alvim, F.Z. Vissotto, J.M. Aguirre, Influence of carrier agents on the physicochemical properties of blackberry powder produced by spray drying, Int. J. Food Sci. Technol. 47 (2012) 1237–1245. [30] X.D. Chen, N. Özkan, Stickness, functionality, and microstructure of food powders, Dry. Technol. 25 (2007) 969–979. [31] N.N. Potter, J.H. Hotchkiss, Food Science, fifth ed Springer, New York, 1995. [32] A.Z. Mercadante, D.B. Rodriguez-Amaya, G. Britton, HPLC and mass spectrometric analysis of carotenoids from mango, J. Agric. Food Chem. 45 (1) (1997) 120–123. [33] O. Fennema, Food Chemistry, third ed Marcel Dekker, New York, NY, 1996. [34] S.A. Desobry, F.M. Netto, T.P. Labuza, Comparison of spray-drying, drum-drying and freeze-drying for β-carotene encapsulation and preservation, J. Food Sci. 62 (6) (1997) 1158–1162. [36] R.V. Tonon, A.F. Baroni, C. Brabet, O. Gibert, D. Pallet, M.D. Hubinger, Water sorption and glass transition temperature of spray dried açaí (Euterpe oleracea Mart.) juice, J. Food Eng. 94 (2009) 215–221. [37] M.J. Aguilera, J.M. Del Valle, M. Karel, Caking phenomena in amorphous food powder, Trends Food Sci. Technol. 6 (5) (1995) 149–155. [38] M.A. Silva, P.J.A. Sobral, T.G. Kieckbusch, State diagrams of freeze-dried camu-camu (Myrciaria dúbia (HBK) Mc Vaugh) pulp with and without maltodextrin addition, J. Food Eng. 77 (3) (2006) 426–432. [39] S. Jaya, H. Das, Glass transition and sticky point temperatures and stability/mobility diagram of fruit powders, Food Bioprocess Technol. 2 (2009) 89–95. [40] Y. Roos, M. Karel, Water and molecular weight effects on glass transitions on amorphous carbohydrates and carbohydrate solutions, J. Food Sci. 56 (1991) 1676–1681. [41] L.H. Mosquera, G. Moraga, N. Martínez-Navarrete, Critical water activity and critical water content of freeze-dried strawberry powder as affected by maltodextrin and arabic gum, Food Res. Int. 47 (2) (2012) 201–206. [42] G. Moraga, N. Martínez-Navarrete, A. Chiralt, Water sorption isotherms and glass transition in strawberry, J. Food Eng. 62 (4) (2004) 315–321. [43] G. Moraga, N. Martínez-Navarrete, A. Chiralt, Water sorption isotherms and phase transitions in kiwifruit, J. Food Eng. 72 (2) (2006) 147–156.