Food Research International 69 (2015) 410–417

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Effect of process variables on the drying rate of mango pulp by

Refractance Window
Marta Fernanda Zotarelli, Bruno Augusto Mattar Carciofi, João Borges Laurindo ⁎
Department of Chemical and Food Engineering, Federal University of Santa Catarina, EQA/CTC/UFSC, 88040-900, Florianópolis, SC, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Refractance Window (RW) process is considered a novel and promising drying method, which uses hot water in
Received 15 August 2014 contact with a polyester film (Mylar) at its bottom face to heat up and dry out a solution spread on the film
Accepted 23 January 2015 surface. The aim of this study was to investigate the effect of process variables on the RW drying of mango
Available online 31 January 2015
pulp: water temperature (75, 85 and 95 °C), product thickness (2, 3 and 5 mm) and radiant source (transparent
and painted Mylar film). Films' transmissivities to infrared radiation were determined and the drying kinetics of
Keywords:
Refractance Window
mango pulp with the nine possible pairs of layer thickness and water temperature were assessed. Mylar film was
Pulp partially transparent to infrared radiation, while the black film (Mylar painted) blocked all infrared radiation
Fruit emitted from the hot water. RW evaporation capacity was up to 10 kg m−2 h−1 (pulp with 2 mm, water at
Drying 95 °C), indicating a very efficient drying process. RW is more efficient than black film process for pulp thickness
Powder up to 3 mm. In fact, this study clearly established that radiative heat transfer contributes to less than 5% of the
total amount of energy delivery to food during the RW drying process.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction water circulating in a reservoir. A polyester film, used as the mat in

which the drying solution or suspension is spread, is partially transpar-
The development of powdered foods has grown in recent years, and ent to infrared radiation and commercially known as Mylar, from
this trend is mainly due to the convenience offered by these products, DuPont, USA (Nindo & Tang, 2007). The heat required by the drying
their chemical and microbiological stability, and reduced transportation process is supplied by the hot water circulating in the reservoir. One
and storage costs (Forny, Marabi, & Palzer, 2011). The main drying advantage of this drying method is the possibility of using a relatively
processes used to produce powdered foods are spray drying (SD), low drying temperature and controlling the residence time of the drying
drum drying (DD), and freeze-drying. To be more competitive, solution, which enables obtaining flocculated or powdered products
powdered foods should maintain their nutritional, physical and chemi- with good sensory properties and reduced nutritional losses (Abonyi
cal characteristics, and be produced at low cost. However, producing et al., 2002; Nindo & Tang, 2007).
high-quality powdered foods at low-cost is not always possible from In this sense, RW has been used for processing fruit and vegetable
the above-mentioned processes (Ochoa-Martínez, Quintero, Ayala, & pulps in order to obtain dried powders, with results that indicate good
Ortiz, 2012). retention of natural compounds (Clarke, 2004). Abonyi et al. (2002)
An alternative drying process for producing powdered foods is the investigated different drying methods (Refractance Window (RW),
Refractance Window (RW) drying. RW is a method for concentrating spray drying (SD), drum drying (DD) and freeze-drying) for drying
and drying solutions and purees that allows obtaining films and flakes strawberry pulp and carrot puree, evaluating color and retention of
at relatively low processing temperatures, with reasonable costs carotenoids, vitamin C, and volatile compounds of dried products. The
(Nindo, Feng, Shen, & Kang, 2003). According to Nindo and Tang authors reported that DD caused the greatest loss of carotenes in carrot
(2007), for the same drying capacity, the RW equipment requires puree (approximately 56%), followed by RW (8.7%) and freeze-drying
50–70% less capital expenditure and 50% less energy than the freeze- (4%). The retention of ascorbic acid in strawberry powder was quite
drier process. similar in the products dehydrated by RW and freeze-drying. However,
In the RW drying process, a solution or puree is spread onto a trans- losses of volatiles observed in strawberry pulps dried by RW and SD
parent polyester mat, which has its lower surface in contact with hot were higher than in pulps dehydrated by freeze-drying, a result that is
explained by the use of higher temperatures in the first two processes.
Nonetheless, the authors concluded that the product dried by RW
⁎ Corresponding author. Tel.: +55 48 37215229; fax: +55 48 37219687. had overall quality parameters comparable to those dehydrated by
E-mail address: jb.laurindo@ufsc.br (J.B. Laurindo). freeze-drying.

http://dx.doi.org/10.1016/j.foodres.2015.01.013
0963-9969/© 2015 Elsevier Ltd. All rights reserved.
 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417 411

Nindo et al. (2003) conducted experiments with pumpkin puree to 2003). The device, shown in Fig. 1, consists of a container
evaluate the energy efficiency of the RW process, and its effect on (0.8 m × 0.4 m × 0.05 m) with a plastic film (mat) covering its top
microbial reduction. For heating water at 95 °C the authors reported (fixed and sealed at the border by an O-ring and screws), and hot
logarithmic reductions of 4.6 for total aerobic (APC), 6.1 for coliforms, water circulating from a thermostatic bath (DIST model DIST921,
6.0 for Escherichia coli and 5.5 for Listeria innocua. Brazil).
Caparino et al. (2012) investigated the effects of DD, SD, RW, and The plastic mat is a 0.25 mm-thick Mylar film (D type, DuPont,
freeze-drying processes on the physical properties and microstructures USA), which was set at the top of the reservoir, so that the bottom
of mango powders. They reported very similar results from samples was in contact with the circulating water and the upper face served
dehydrated by RW and freeze-drying, and concluded that RW can as support for the spread fruit pulp (fruit-pulp layer) to be dried.
produce mango powder with quality comparable to those obtained by The thickness of the Mylar film was chosen based on literature data
freeze-drying, and superior than those obtained by SD and DD. (Nindo et al., 2003).
As RW drying is a relatively new technology, much of the available
literature is about the comparison of the physicochemical quality of 2.3. Mylar film transmissivity
the products obtained by RW with those produced by traditional
processes (Abonyi et al., 2002; Caparino, Sablani, Tang, Syamaladevi, & The transmissivities to infrared radiation of 0.25 mm-thick Mylar
Nindo, 2013; Ochoa-Martínez et al., 2012; Pavan, Schmidt, & Feng, film and a sample of this film painted black, with heat-resistant spray
2012; Topuz, Feng, & Kushad, 2009; Topuz et al., 2011). Little informa- paint (Colorgin, Sherwin-Williams, Brazil) were determined. The
tion on the contribution of the heat and mass transfer mechanisms average thickness of the painted film was determined at various film
during drying by RW is found in the literature (Nindo & Tang, 2007). regions with a digital micrometer (Mitutoyo, APB-1D, Japan), resulting
Based on this specific lack of information, the aim of this study was in an average value of 0.33 ± 0.02 mm. The transmissivity was
to investigate the influence of drying temperature, origin of the infrared determined with a FTIR spectrometer (Oriel MIR8025TM Modular IR
radiation, and pulp layer thickness on the drying rates of mango pulp Fourier Spectrometer) based on the Michelson interferometer
treated by RW. The choice of using mango as raw material is related to (Nicolau, Scopel, & Possoli, 2009). The FTIR spectrometer operates
the possibility of adding value to this important tropical fruit. with emission spectrum range from 1.7 to 27 μm.

2. Material and methods 2.4. Drying rates of mango pulp

2.1. Preparation of mango pulp Mango powder was obtained from mango pulp dehydrated by
Refractance Window (RW, non-painted film) and on a black film
The Tommy Atkins mangos used in this study were bought in retail (BF, with the Mylar film painted with a heat-resistant black ink).
stores in Florianópolis, SC, Brazil. The fruits were selected according For BF, the painted Mylar film was fixed to the same experimental
to their degree of ripeness, from visual analysis and soluble solids apparatus used in the RW drying process. In both cases, the mango
concentration (°Brix). The selected mangos had soluble solids content pulp was spread over the well stretched Mylar film with the help of
ranging from 12 to 16 °Bx, as determined by a manual refractometer doctor-blade screws, which allows the pulp thickness to be adjusted.
(Reichert, Model AR200, USA) with measurement range of 0–90 °Bx The drying curves of mango pulp treated by RW and BF were
and 0.2 °Bx resolution. The fruits were washed, peeled by hand, and assessed from samples taken from different areas of the fruit pulp
ground in a household blender (Arno, São Paulo, Brazil) to obtain the spread, during the drying process, avoiding taking samples near the
mango pulp. Before the drying process, the pulp was sieved using a 16 edges. At each sample extraction, approximately 0.7 g was taken from
mesh sieve in order to retain large particles and fibers. the mango pulp with a spatula, weighed in an analytical scale
(Shimadzu, model AY220, Japan), vacuum dried at 70 °C (TECNAL,
2.2. Experimental device: batch RW process model TE-395, Brazil) (A.O.A.C., 2005), and weighed again, for moisture
determination. The experimental conditions for both drying processes
For the RW drying of mango pulp a laboratory-scale RW dryer, (RW and BF) were the same, i.e., three different temperatures of the
operating in batch, was built using the same principle as the circulating water (75, 85 and 95 °C) and three different thicknesses of
industrial equipment (Clarke, 2004; Nindo & Tang, 2007; Nindo et al., the pulp layer (2, 3 and 5 mm). Experiments were carried out at

Fig. 1. Sketch of the Refractance Window experimental device.

412 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417

90
ambient air RH ranging between 58 and 68%, and temperature close to
25 °C. All experiments were performed in triplicate. The drying rates 80
RW
were obtained from the slope of the linear equation fitted to the initial
70 BF
part of the drying curves (moisture content vs. time).

Transmission (%)
60
2.5. Temperature measurements
50

For the temperature measurements during drying, thermographs 40

were taken from the pulp surface (thermographic camera approximate- 30
ly 50 cm above the pulp surface — FLIR model T360, Sweden) during the
drying processes (RW and BF). For fixing the pulp film thickness and 20
delimiting the measurement area, a 2 mm-thick metal frame was 10
used. This procedure was important to allow comparisons between
RW and BF processes, at the different process conditions. This frame 0
2 4 6 8 10 12 14 16 18 20 22 24 26
was attached to the Mylar film and the mango pulp was spread inside Wavelength (µm)
it. Thus, the film thickness was controlled and the pulp excess was
removed with an iron scale. To highlight the region under drying on Fig. 2. Transmissivities: ( ) transparent Mylar used in RW drying and ( ) black
the micrographs, a polystyrene frame was placed over the metal frame. film (Mylar painted black) used in BF drying.
For image recording, some parameters were set in the thermograph-
ic camera, such as the distance between the camera and the mango pulp
surface, the emissivity of mango pulp, ambient temperature and relative
humidity. Due to the high moisture content in the pulp (about 86% the wavelength range evaluated. However, the heated black paint
w.b.), its emissivity was considered the same as the water (ε = 0.96) emits radiation. The emissivity of the black film was determined exper-
(Incropera, DeWitt, Bergman, & Lavine, 2007, Chapter 12). The ambient imentally and the value obtained was 0.98 (data not shown), which is
relative humidity (RH) and temperature were measured with a consistent with the data reported by Incropera et al. (2007).
thermohygrometer with an accuracy of ±2.5% RH (Testo, model 610, According to Wien's displacement law, a black body at around 373 K
Germany). The images were analyzed with the software FLIR (100 °C) shows maximum emission (102 W m−2 μm−1) at wavelength
QuickReport 1.2 SP2. This software provides the maximum and of approximately 8 μm. Furthermore, at wavelengths between 4 μm and
minimum temperature at each image area. The evolution of mango 40 μm, the emission corresponds to more than 10% of the maximum
pulp temperatures was also determined by T-type thermocouples emission (101 W m−2 μm−1) (Incropera et al., 2007, Chapter 12) and
(IOPE, model TF-TX-A-R-30AWG, Brazil) connected to a data acquisition the Mylar film shows transmissivity peaks of about 40%. Due to the
system (Agilent, model 34970A, Malaysia). The thermocouples were product high moisture content, it is reasonable to assume that the pulp's
installed in five microregions of the pulp layer and in nine regions of absorption of radiation is similar to that of pure water, which absorbs
the water container. more infrared radiation at wavelengths around 3.0, 4.7, 6.0 μm and
above 12 μm (Sandu, 1986; Sternglanz, 1956). As can be seen in Fig. 2,
2.6. Statistical analysis at these wavelengths Mylar film showed transmissivity peaks to the in-
cident infrared radiation. These data suggest the effectiveness of heat
The software Statistica 7.1 (Statsoft Inc., Tulsa, USA) was used to transfer by radiation in the RW process.
perform the statistical analyses of the results of drying rate and According to Nindo and Tang (2007), when the pulp with high
evaporative capacity. A multiple comparison of means was analyzed moisture content is spread on the Mylar top, refraction at the plastic-
by using the Tukey test at the 95% confidence level (p b 0.05). pulp interface is reduced and the radiant thermal energy is more easily
transmitted from the plastic to the product. In fact, the refractive
3. Results and discussion indexes are 1.64 for Mylar (Tsilingiris, 2003), 1.33 for water, and 1 for
air (Siegel & Howell, 1992). Thus, the reflection loss coefficient of
3.1. Emission, transmission and absorption of infrared radiation during normal radiation in air–Mylar system (rair–Mylar(0) = 0.0587) is almost
drying five times higher than in the water–Mylar system (rwater–Mylar(0) =
0.01073). Product properties change with the decrease of its moisture
Three main aspects should be evaluated in the RW process: the during drying, i.e., pulp reflectivity increases and pulp absorptivity
emission spectrum of the heating source (hot water or black paint), decreases (Clarke, 2004; Nindo & Tang, 2007; Ratti & Mujumdar,
the transmission spectrum of the propagation medium (Mylar), and 2006). In addition, heat transfer in the RW and BF processes decreases
the product (mango pulp) absorption spectrum. due to the increase of the contact resistance between the Mylar film
The transmissivities data of transparent Mylar film (original) and of and drying fruit pulp. This occurs as a consequence of the air and
Mylar film painted black are shown in Fig. 2. The original Mylar film water vapor layer formed between the Mylar film and the dried fruit
showed transmissivity peaks to the incident infrared radiation mainly pulp, which detached from the plastic. This detachment could be
for wavelengths in the range of 2–6 μm and 14–19 μm. This result is in visually observed.
agreement with those obtained by Meinel and Meinel (1976) in a
study on the transmissivity of some polymer films, including 3.2. Drying experiments
0.13 mm-thick Mylar film. However, the transmission peaks reported
by the authors were almost twice the values observed in the present The drying curves (moisture vs. time) of mango pulp treated by RW
study. This difference may be related to the greater thickness of the and BF, for pulp layer thicknesses of 2, 3 and 5 mm and circulating water
film used in the present work (Ratti & Mujumdar, 2006). Thinner at 75, 85 and 95 °C, are presented in Fig. 3. The drying rates for both
films are more transparent to infrared radiation, but they have lower processes are presented in Table 1, at each experimental condition
mechanical strength, which makes them more difficult to use, because investigated. The drying rates were constant during the drying of
hot water pressure on the film can cause its deformation. pulps with thicknesses of 2 and 3 mm, for all the 3 circulating water
On the other hand, it is evident (Fig. 2) that the paint applied to temperatures. This behavior can be explained by the high initial
the Mylar film blocked almost all the infrared radiation from water in moisture of the fruit pulp (86% wet basis) and its small thicknesses.
 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417 413

2 mm 3 mm 5 mm
7 7 7
6 BF 6 BF 6 BF
5 5 5
4 RW 4 RW 4 RW
3 3 3 75 °C
2 2 2
1 1 1
0 0 0

Moisture content (d. b.)

0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120
7 7 7
6 BF 6 BF 6 BF
5 5 5
4 RW 4 RW 4 RW 85 °C
3 3 3
2 2 2
1 1 1
0 0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120
7 7 7
6 BF 6 BF 6 BF
5 5 5
4 RW 4 RW 4 RW 95 °C
3 3 3
2 2 2
1 1 1
0 0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time (min)

Fig. 3. Moisture content vs. time of mango pulp dried by RW (□) and BF (○) with pulp thickness of: (a) 2 mm, (b) 3 mm, and (c) 5 mm; with water temperatures at: (1) 75 °C, (2) 85 °C and
(3) 95 °C.

At these situations the water vapor pressure at the sample surface is 3 mm pulp layer thickness, respectively. However, at the highest
close to the saturation pressure during the drying process, and the temperature (95 °C) and greatest pulp layer thickness (5 mm) no
internal resistance to mass transfer is negligible (small thicknesses). If difference between the drying rates was observed.
the external conditions do not influence the drying rate (convective As expected, warmer circulating water led to shorter drying times in
drying is negligible), one can postulate that the film dehydration occurs both RW and BF processes. In RW with 2 mm-thick pulp layer, the
by classical evaporation and is controlled by the heat transfer from the drying rate increased 1.7 fold when the water temperature increased
circulating water. from 75 to 85 °C and 2 fold when the water temperature increased
For the cases above (2 and 3 mm thick pulps, at 75, 85 and 95 °C), the from 75 to 95 °C. For BF, at the same experimental conditions, the drying
drying rates of pulps treated by RW were greater than those treated by rate increased 2.1 and 2.5 fold as water temperature increased from 75
the BF process. The differences were observed at 75 °C, for which the to 85 °C and from 75 to 95 °C, respectively. For 5 mm-thick pulp layers,
drying rates of samples treated by RW were 2.2 and 2.4 times greater the temperature increase had a smaller impact on the drying time,
than the drying rates of pulps treated by the BF process, for 2 and mainly because the resistance to conduction heat transfer and mass
transfer (pathway for water diffusion) into the pulp becomes greater.
Besides, the incident thermal radiation becomes less important as
pulp layer thickness increases, due to the low penetration capacity of
Table 1 infrared radiation.
Drying rates (mean ± standard deviation) and evaporative capacity (mean ± standard Pulp layer thickness had great impact on drying rates in both RW
deviation) for mango pulp dried by Refractance Window (RW) and black film (BF). and BF. The drying rate in constant rate period decreased by at least
Drying Water Pulp Drying rate Evaporative 40% when the thickness of the drying pulp increased from 2 to 3 mm.
process temperature thickness (g g−1 min−1) capacity Drying rates observed for 2 mm-thick sample treated by RW was 4.9
(°C) (mm) (kg m−2 h−1) (at 75 °C), 5.5 (at 85 °C), and 6.6 (at 95 °C) times higher than the drying
RW 75 2 0.301 ± 0.033ª 5.21 ± 0.58a rates observed for 5 mm-thick pulp, treated by the same drying process.
3 0.177 ± 0.013a 4.62 ± 0.34a Similarly, the drying rates observed for the pulp layers treated by the BF
5 0.061 ± 0.003a 2.67 ± 0.12a process increased by 3.1, 4.1, and 4.0 fold (at 75, 85, and 95 °C, respec-
85 2 0.508 ± 0.087a 8.81 ± 1.51a
3 0.189 ± 0.011a 4.91 ± 0.30a
tively) when the pulp layer thickness decreased from 5 to 2 mm. Thus,
5 0.092 ± 0.003a 4.00 ± 0.16a the influence of the circulating water temperature on the drying rates
95 2 0.603 ± 0.050a 10.75 ± 0.86a is higher for thin pulp layers. The differences between the drying
3 0.331 ± 0.033a 8.61 ± 0.85a rates of 2-mm-thick pulp layer treated by RW and BF ranged from 43
5 0.091 ± 0.005a 3.94 ± 0.22a
to 55%, depending on the water temperature. Similarly, for a 3 mm-
BF 75 2 0.135 ± 0.019b 2.33 ± 0.34b
3 0.075 ± 0.007b 1.95 ± 0.20b thick pulp layer, the differences between the processes ranged from
5 0.043 ± 0.003b 1.86 ± 0.11b 26 to 57%. However, no difference between the drying rates was
85 2 0.282 ± 0.026b 4.89 ± 0.45b observed for 5 mm-thick pulp layer dried by RW and BF. Drying time
3 0.138 ± 0.014b 3.59 ± 0.37b of fruit pulp with 2 and 3 mm thick ranged from 15 to 20 min, while
5 0.068 ± 0.010b 2.95 ± 0.44b
95 2 0.342 ± 0.037b 5.93 ± 0.65b
the characteristic drying time for 5 mm thick pulps ranged from 60 to
3 0.202 ± 0.012b 5.25 ± 0.32b 80 min. These results indicate that thicknesses around 3 mm tend to
5 0.086 ± 0.005a 3.74 ± 0.23a be the better choice for drying mango pulp by RW.
a,b
— values followed by different lowercase letters indicate significant difference (p b 0.05) The black ink blocked the radiation emitted by the hot water (Fig. 2),
between drying rates and evaporative capacities observed for RW and BF performed with but it is heated and emits thermal radiation (infrared radiation). The
the same temperature and pulp thickness. differences between RW and BF drying rates could be explained from
414 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417

the paint's conductive resistance. In fact, the black paint layer increased On the other hand, the total heat transfer from the hot water to the
the Mylar film thickness by approximately 25%, leading to an increase of fruit pulp can be estimated directly from the constant drying rate
the film's heat transfer resistance. However, this is not the only effect observed in the experiments. For 2 mm thick pulp and heating water
present. at 95 °C the evaporation flux was approximately 10.7 kg m−2 h−1, as
When the thermal radiation from water or black paint hits the Mylar showed in Table 1. This evaporation flux implies a heat flux of approxi-
surface, its intensity is initially attenuated by reflection at the water– mately 6.7 kW m− 2 (considering the latent heat of free water at
Mylar or black paint–Mylar interface, according to Fresnel reflectance. atmospheric pressure, 2256 kJ kg−1). The comparison between heat
The remaining radiation that crosses the interface propagates in the transfer fluxes by radiation and estimated from the evaporation flux
Mylar until the Mylar–pulp interface, suffering attenuation according shows that only approximately 3% of the overall heat supplied to the
to the Beer's Law. At the second interface (Mylar–pulp), the radiation pulp during drying at the constant rate period is from the infrared
will suffer a second reflection, and part of it crosses the interface and radiation emitted by the hot water. This result was reached from an
reaches the pulp. Internal reflections at both interfaces are responsible order of magnitude analysis, neglecting ambient heat losses and
for further multiple reflections and absorption of radiation into the considering that all the radiation emitted by the hot water is transmit-
Mylar film. In this way, the total radiation transmitted should be calcu- ted through the Mylar film and reaches the drying pulp. In fact, only a
lated as the contribution of all multiple transmission components at the fraction of this thermal radiation, no greater than 50%, reaches the
second interface (Siegel & Howell, 1992). As the water–Mylar and black fruit pulp, as showed in Fig. 2.
paint–Mylar contact interfaces are different, the reflection phenomena From Fig. 4 one can perform a detailed analysis of the heat transfer
can have different intensities in each case. mechanisms in RW drying of mango pulp during the constant drying
The pulp absorbs the transmitted radiation according to the Beer's rate period. The resistances present in the thermal circuit are:
Law. The penetration depth of electromagnetic waves in a material is
i) The convective thermal resistance between water and Mylar, Rw,
commonly defined as the distance from the interface (radiation inci-
is given by Eq. (2).
dence) in which the radiation energy is reduced to 1/e (36.79%) of the
incident radiation energy, in which e is the Euler number (Singh &
Heldman, 2009). For water, the penetration depth is 0.09 mm for a 1
wavelength between 2 and 200 μm. Thus, it is evident that thicker Rw ¼ ð2Þ
hw
pulp layer is less influenced by the infrared radiation during drying.
The discussion on the phenomena presented above pointed out the in which hw is convective heat transfer coefficient between the
mechanisms that are present in RW and BF drying processes. Neverthe- hot water and the Mylar film.
less, it is important to evaluate the relative importance of each ii) As conductive and radiative heat transfer occurs in parallel, the
mechanism on the drying rate. Some previous studies reported the equivalent resistance through the Mylar, Rm, is given by Eq. (3).
three heat transfer mechanisms present in RW drying, but they did
not present results or a quantitative analysis that show the relative
!−1
importance of each heat transfer mechanism to the drying process 1 1
(Clarke, 2004; Nindo & Tang, 2007; Nindo et al., 2003). Rm ¼ þ ð3Þ
Rm;k Rm;r
Simply, one can estimate the relative importance of heat transfer by
convection, conduction and radiation through an order of magnitude
in which Rm,k = Lm/km is the Mylar conductive resistance (Lm =
analysis. Radiation heat transfer rate is maximum for a black body
2.5 × 10−4 m and km = 0.15 W m−1 K−1 are the Mylar thickness
(emissivity, ε = 1), for which the net heat flux is given by the Stefan–
and thermal conductivity, respectively) and Rm,r = 1/hr is the ra-
Boltzmann equation,
diative resistance in the Mylar, with hr = εσ(Tb + Ti)(T2b + T2i ).
iii) The heat transfer through the mango pulp can occur by conduc-


 h i
4 4 −8 4 4 −2 tion and free convection, in parallel. Therefore, the equivalent
qrad ¼ εσ T w −T p ¼ 1  5:670410  10 368 −348 ¼ 208 W m
thermal resistance through the pulp, Rp, is given by Eq. (4).
ð1Þ

!−1
in which σ = 5.670410 × 10−8 W m−2 K−4 is the Stefan–Boltzmann 1 1 Lp
Rp ¼ þ ¼ ð4Þ
constant, Tw is the water temperature and Tp is the pulp temperature. Rp;k Rp;c ke f
This calculation was performed to simulate the heat transfer by radia-
tion at the constant drying rate period, when the fruit pulp temperature in which Rp,k is the conductive resistance in the mango pulp, Rp,c
was approximately 75 °C. is the convective resistance in the mango pulp, Lp is the pulp

Ta
Tu

Ti

Tb

Fig. 4. Representation of the thermal circuit in a Refractance Window drying system, in which Tw is the water temperature, Tb is the temperature of water–Mylar interface, Ti is the
temperature of Mylar–pulp interface, Tu is the temperature of pulp–air interface, Ta is the air temperature, qw is the heat transfer rate from hot water to the Mylar surface, qm is
the heat transfer rate through the Mylar, qp is the heat transfer rate through the mango pulp, qa is the heat transfer rate between pulp surface and air; Lp is the pulp thickness, Lm is the
Mylar thickness; Rw is the convective thermal resistance between water and Mylar, Rm is the thermal resistance through the Mylar, Rp is the thermal resistance of the mango pulp, Ra
is the convective thermal resistance between mango pulp and air; ṁw is the evaporation rate and ΔHv is latent heat of vaporization.
 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417 415

thickness, kef is an effective coefficient of heat transfer through resistances through the Mylar at steady state. The radiative resistance
the pulp. depends on the temperature difference between hot water–Mylar (Tb)
iv) The convective thermal resistance between pulp and air, Ra, is and Mylar–fruit pulp (Ti) interfaces. At the constant drying rate period,
given by Eq. (5). Tb = 95 °C (neglecting Rw) and Ti = 75 °C (measured by thermocou-
ples), leading to a ratio Rm,k/Rm,r = 0.017. This ratio validates the
previous analysis based on the evaporation rate, ratifying the negligible
1 contribution of radiative heat transfer in RW drying performed at the
Ra ¼ ð5Þ
ha investigated temperatures.
The evaporative capacity of the RW process was calculated from
in which ha is the convective heat transfer coefficient between the reproducible results showed in Table 1. For pulp layer thickness
pulp surface and air. of 2 mm and heating water at 95 °C (10.75 kg m− 2 h− 1) the result
is in agreement with those reported by Nindo et al. (2003) during
the RW drying of pumpkin puree. These authors reported a water
One cannot calculate the global resistance between hot water and evaporative capacity of 10 kg m− 2 h− 1 for puree thickness ranging
air, because the values of Rw, Rp and Ra are unknown. However, one from 0.4 to 0.6 mm, in a pilot dryer with circulating water at 95 °C.
can calculate the ratio between conductive and radiative thermal Nindo et al. (2003) also reported water evaporative capacity in the

Mylar surface without pulp 30 seconds after spreading pulp

Tmín. 89.0 °C – Tmax. 89.3 °C Tmin. 58.5 °C - Tmax 64.4 °C

(a) (b)

5 minutes - Tmin 67.9 °C - Tmax73.2 °C 10 minutes - Tmin.70.5 °C - Tmax 75.0 °C

(c) (d)

15 minutes - Tmin 71.2 °C - Tmax 88.5 °C 20 minutes - Tmin73.6 °C - Tmax 89.4 °C

(e) (f)

Fig. 5. Infrared thermography during Refractance Window drying of mango pulp with 2 mm-thick pulp layer and water at 95 °C.
416 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417

range of 3.1–4.6 kg m− 2 h− 1 in an industrial RW dryer. Evaporative In BF drying the temperature increase started at approximately 20 min
capacity of 6 kg m− 2 h− 1 during drying of 1 mm-thick carrot puree after the process started (Fig. 6d).
and strawberry pulp (water at 95 °C) in a pilot continuous RW Fig. 7 shows pulp time–temperature evolution registered by five
dryer was reported by Abonyi, Tang, and Edwards (1999). Despite thermocouples inserted into the pulp layer, for both, RW and BF
the differences with the literature results, we supported our higher processes. The temperatures measured by the thermocouples were
evaporation rate results with reproducible experiments. slightly higher than those indicated by the thermographs, which are
The time–temperature evolution of drying pulp surface was related to the pulp temperature. Similar to the thermographs, thermo-
determined by thermography during the RW and BF processes couple data showed a rapid temperature increase until it reached a
performed with 2 mm-thick pulp and water at 95 °C, and is shown in plateau close to 75 °C. In the BF process, a small temperature decrease
Figs. 5 and 6, respectively. No relevant temperature differences can be is observed during the constant drying rate, probably due to the
seen on the Mylar surface without the mango pulp (Figs. 5a and 6a). evaporative cooling caused by a longer contact with the ambient air.
However, temperature gradients were observed on the pulp's surface Thermocouples can detach or be uncovered by the pulp during
in the course of the drying process (Figs. 5b–f and 6b–e), which are drying, because pulp layer thickness decreases during the drying
mainly due to the spatial differences on the pulp layer thickness. After (drying shrinking), which modify the temperature results. Pulp temper-
spreading, pulp temperatures quickly increased up to 70 °C and atures measured by the thermocouples differed by up to 5 °C from the
remained practically stable during the constant drying rate period. temperatures assessed from the thermographs. This difference can
From 10 to 15 min after the RW process started (Fig. 5c and d) the tem- also be attributed to the lack of accuracy in the evaluation of pulp
perature began to rise in some areas, due to their low moisture content. emissivity during drying.

Mylar surface without pulp 30 seconds after spreading pulp

Tmin 85.6 °C - Tmax 89.5 °C Tmin 44.8 °C - Tmax 59.6 °C

(a) (b)

10 minutes - Tmin 65.4 °C - Tmax 73.2 °C 20 minutes - Tmin 65.6 °C - Tmax 85.2 °C

(c) (d)

30 minutes - Tmin 83.7 °C - Tmax 86.1 °C

(e)

Fig. 6. Infrared thermography during black film drying of mango pulp with 2 mm-thick pulp layer and water at 95 °C.
 M.F. Zotarelli et al. / Food Research International 69 (2015) 410–417 417

90

85
Pulp temperatures - RW
80

75

Temperature (°C) 70

65 Pulp temperatures - BF

60

55

50

45

40
0 5 10 15 20 25 30 35
Time (min)

Fig. 7. Time–temperature evolution of mango pulp during RW ( ) and BF (●) drying.

4. Conclusions Caparino, O. A., Tang, J., Nindo, C. I., Sablani, S. S., Powers, J. R., & Fellman, J. K. (2012).
Effect of drying methods on the physical properties and microstructures of mango
(Philippine ‘Carabao’ var.) powder. Journal of Food Engineering, 11, 135–148.
As expected, higher water temperatures led to higher drying rates, Clarke, P. T. (2004). Refractance Window™ — “Down under”. Proceedings of the
presenting great improvement on the drying rates of thinner pulp international drying symposium (IDS 2004) (pp. 813–820).
Forny, L., Marabi, A., & Palzer, S. (2011). Wetting, disintegration and dissolution of
layers. Thicker pulp layers are less influenced by temperature variations, agglomerated water soluble powders. Powder Technology, 206, 72–78.
due to their higher heat and mass transfer resistances, which control the Incropera, F. P., DeWitt, D. P., Bergman, T. L., & Lavine, A. S. (2007). Fundamentals of heat
process. RW drying is more efficient in transferring energy than BF and mass transfer (6th ed.). John Wiley & Sons, Inc.
Meinel, A. B., & Meinel, M. P. (1976). Applied solar energy: An introduction (4th ed.).
drying because the additional layer of black paint increased the total
Addison-Wesley Series in Physics.
resistance from increasing conductive resistance. Evaporation capacity Nicolau, V. P., Scopel, D. A. B., & Possoli, K. (2009). Experimental apparatus to deter-
up to 10 kg m−2 h− 1 (pulp with 2 mm, water at 95 °C) showed that mine spectral emissivities of ceramic samples. Cobem [S.I.], Gramado — RS —
Brasil, 2009.
RW is a very efficient drying process, even if the relative importance
Nindo, C. I., Feng, H., Shen, G. Q., & Kang, D. H. (2003). Energy utilization and microbial
of radiation heat transfer is negligible. In fact, this study clearly reduction in a new film drying system. Journal of Food Processing and Preservation,
established that radiative heat transfer contributes to less than 5% of 27(2), 117–136.
the total thermal energy delivery to food during the RW drying process. Nindo, C. I., & Tang, J. (2007). Refractance Window dehydration technology: A novel
contact drying method. Drying Technology, 25, 37–48.
Results indicated thicknesses close to 3 mm as a better condition for Ochoa-Martínez, C. I., Quintero, P. T., Ayala, A. A., & Ortiz, M. J. (2012). Drying character-
mango pulp drying and also showed RW as a competitive drying pro- istics of mango slices using the Refractance Window™ technique. Journal of Food
cess to dehydrate fruit pulps. Engineering, 109, 69–75.
Pavan, A. M., Schmidt, S. J., & Feng, H. (2012). Water sorption behavior and thermal
analysis of freeze-dried, Refractance Window-dried and hot-air dried açaí (Euterpe
oleracea Martius) juice. LWT — Food Science and Technology, 48(1), 75–81.
Acknowledgements
Ratti, C., & Mujumdar, A. S. (2006). Infrared drying. In A. S. Mujumdar (Ed.), Handbook of
industrial drying (pp. 423–437). CRC Press.
The authors thank CNPq, FAPESC-SC, and CAPES/Brazil for the finan- Sandu, C. (1986). Infrared radiative drying in food engineering: A process analysis.
cial support. Biotechnology Progress, 2(3), 109–119.
Siegel, R., & Howell, J. R. (1992). Thermal radiation heat transfer (3th ed.). London: Taylor
& Francis Group, 93–133.
Singh, R. P., & Heldman, D. R. (2009). Introduction to food engineering (4th ed.). London:
References
Elsevier Publishing, 375–376.
Sternglanz, H. (1956). Using water as an infrared solvent. Journal of Applied Spectroscopy,
A.O.A.C. Association of Official Analytical Chemists (2005). Official methods of analysis
2, 77–82.
(18th ed.). Gaithersburg, Maryland: A.O.A.C. INTERNATIONAL.
Topuz, A., Dincer, C., Özdemir, K. S., Sultan, K., Feng, H., & Kushad, M. (2011). Influence of
Abonyi, B. I., Feng, H., Tang, C. G., Edwards, B. P., Mattinson, D. S., & Fellman, J. K. (2002).
different drying methods on carotenoids and capsaicinoids of paprika (Cv., Jalapeno).
Quality retention in strawberry and carrot purees dried with Refractance Window
Food Chemistry, 129, 860–865.
system. Journal of Food Science, 67, 1051–1056.
Topuz, A., Feng, H., & Kushad, M. (2009). The effect of drying method and storage on color
Abonyi, B. I., Tang, J., & Edwards, C. G. (1999). Evaluation of energy efficiency and quality
characteristics of paprika. LWT — Food Science and Technology, 42, 1667–1673.
retention for the Refractance Window™ drying system. Research Report. Pullman,
Tsilingiris, P. T. (2003). The total infrared transmittance of polymerized vinyl fluoride
WA: Washington State University.
films for a wide range of radiant source temperature. Renewable Energy an
Caparino, O. A., Sablani, S. S., Tang, J., Syamaladevi, R. M., & Nindo, C. I. (2013). Water
International Journal, 28(6), 887–900.
sorption, glass transition, and microstructures of Refractance Window- and freeze-
dried mango (Philippine “carabao” var.) powder. Drying Technology, 1969–1978.