Effects of particle size distribution on some physical, chemical and functional properties of unripe banana flour

Food Chemistry 213 (2016) 180–186

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Effects of particle size distribution on some physical, chemical and functional properties of unripe banana flour Nazlı Savlak ⇑, Burcu Türker, Nazlıcan Yesßilkanat Celal Bayar University, Faculty of Engineering, Department of Food Engineering, 45140 Manisa, Turkey

a r t i c l e

i n f o

Article history: Received 5 January 2016 Received in revised form 20 June 2016 Accepted 20 June 2016 Available online 21 June 2016 Keywords: Unripe banana flour Particle size Physical properties Functional properties

a b s t r a c t The objective of this study was to examine the effect of particle size distribution on physical, chemical and functional properties of unripe banana flour for the first time. A pure triploid (AAA group) of Musa acuminata subgroup Cavendish (°Brix;0.2, pH;4.73, titratable acidity; 0.56 g/100 g malic acid, total solids; 27.42%) which was supplied from Gazipasßa, Antalya, Turkey from October 2014 to October 2015 was used. Size fractions of <212, 212–315, 316–500 and 501–700 lm were characterized for their physical, functional and antioxidant properties. Particle size significantly effected color, water absorbtion index and wettability. L⁄ value decreased, a⁄ and b⁄ values decreased by increasing particle size (r2 =
0.94, r2 = 0.72, r2 = 0.73 respectively). Particles under 212 lm had the lowest rate of wettability (83.40 s). A negative correlation between particle size and wettability (r2 =
0.75) and positive correlation between particle size and water absorption index (r2 = 0.94) was observed. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Banana fruit (Musa spp.) is an important staple food crop and its production is increasing in the tropical and subtropical regions due to high export potential. Banana is easily transported and can be stored longer when unripe. On the other hand, large amounts of unripe banana rejection or post harvest loss are used as raw materials for domestic artisanal flour production (Aurore, Parfait, & Fahrasmane, 2009). Unripe banana is a promising ingredient and many researchers studied applications of unripe banana flour in various food products (Ho, Aziz, & Azahari, 2013; Karim & Sultan, 2015; Agama-Acevedo, Islas-Hernández, Pacheco-Vargas, Osorio-Díaz, & Bello-Pérez, 2012). Some researchers studied the effects of ripening stages on some physical, chemical and functional properties of foods (Onwuka & Onwuka, 2005). There are also studies on flour production from unripe banana which include comparison of physicochemical properties of banana pulp and peel flours (Alkarkhi, Bin Ramli, Yong, & Easa, 2011), production of instant unripe banana flour (Rayo et al., 2015), effect of organic acid pretreatment on some properties of unripe banana flour (Anyasi, Jideani, & Mchau, 2015) and effect of maturation steps on chemical composition of banana and plantain peels (Emaga, Andrianaivo, Wathelet, Tchango & Paquot, 2007). However, to the best of authors’ knowledge, effect of particle size distribution of ⇑ Corresponding author. E-mail addresses: [email protected] (N. Savlak), burcupekmez07@gmail. com (B. Türker), [email protected] (N. Yesßilkanat). http://dx.doi.org/10.1016/j.foodchem.2016.06.064 0308-8146/Ó 2016 Elsevier Ltd. All rights reserved.

unripe banana flour is not studied. As a raw material for different food products, particle size of unripe banana flour is very important for quality attributes such as water holding, oil holding, water solubility, powder dispersion, wettability, bulk density and tapped density. Our former studies on product development by using UBF showed that (data not shown) batter viscosity and end product texture were highly affected by particle size as well, which made the study worthwhile to investigate the physical, chemical and functional data of UBF having different particle size distribution. Since differences in particle size are key parameters which affect the above mentioned properties, the objective of this study was to determine the effects of particle size on some physical, chemical and functional properties of unripe banana flour. Four different size fractions (<212 lm, 212–315 lm, 316–500 lm, 501–700 lm) were chosen for comparison. 2. Materials and methods 2.1. Materials 2.1.1. Unripe banana ‘Dwarf Cavendish’ banana (Musa spp. AAA) cultivated in openfields in the central south coast region (latitude 36°150 N, longitude 32° 170 E) of Gazipasßa, Antalya, Turkey was used in the study. Bananas (°Brix;0.2, pH;4.73, titratable acidity; 0.56 g/100 g malic acid, total solids; 27.42%) were supplied from October 2014–October 2015 in three different time intervals and characterized as unripe banana (Colour index; 2 = entirely green) according to the com-

N. Savlak et al. / Food Chemistry 213 (2016) 180–186

mercial peel color scale described by Aurore et al. (2009). They were harvested at the commercial stage, approximately 110 days after anthesis. Drip irrigation system was used. Base dressing was done using mineral fertilizer TOROS GübreÒ-NPK 15-15-15 (300 g) and animal manure (40–50 kg) in the 8th month of cultivation. Fungicidal agricultural spraying is conducted once a year with Best Captan 50 WPÒ after fertilization and hoeing fields. 2.1.2. Unripe banana flour (UBF) production Unripe banana fruits were hand peeled and immediately rinsed in citric acid solution(1 g/L; Rayo et al., 2015). They were cut into 2 mm slices, again rinsed in the same solution for 1 min and stored at
18 °C. The frozen slices were dried in a freeze dryer (CRYST-Alpha 2–4 LD Plus Freeze Dryer, Newtown Wem, Shropshire, UK) for 48 h, and then grounded by using a blade grinder (Retsch, GRINDOMIX GM200, Germany) to obtain UBF. UBF was passed through a sieve shaker (Sßimsßek Laborteknik, ES-608, Turkey) and size fractions of <212, 212–315, 316–500 and 501–700 lm were collected. They were stored at +4 °C in sealed glass containers for further use.

2.2.1. Moisture content Moisture content of UBF was determined gravimetrically by vacuum oven drying at 70 °C until constant weight (Zenebon & Pascuet, 2005). Results were expressed as percentage. 2.2.2. Water activity Water activity (aw) was measured with a water activity measurement device (HygropPalm HP23-AW, ROTRONIC AG, Bassersdorf, Switzerland). 2.2.3. pH and total soluble solids (TSS) The pH of the UBF was measured according to Suntharalingam and Ravindran (1993) with some modifications. 4% (w:v) UBF suspension was stirred for 5 min, allowed to stand for 30 min, the supernatant was transferred to a beaker and the pH of supernatant was measured by using a WTW inoLabÒ pH meter model pH 7110. Total soluble solids (TSS) in the same flour slurries were measured using a Refractometer (Digital Abbe Refractometer, WAY-2S, Jiangsu Zhengji Instruments Co., Ltd., China). Results were given as Brix. 2.2.4. Color profile of unripe banana flour Color of UBF samples were measured by a CR-5 Konica Minolta Chroma Meter (Osaka, Japan) using CIELAB and CIELCH color scale with the parameters L⁄a⁄b⁄ and L⁄C⁄H⁄. C⁄ coordinate is the chroma (Eq. (1)) and H⁄ coordinate is the hue (Eq. (2)) (Wrolstad & Smith, 2010). The L⁄a⁄b⁄ values were used for determination of Chroma C⁄, Hue angle, whiteness index (WI) and yellowness index (YI) using the following equations;

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2 ða Þ2 þ ðb Þ



Chroma C ¼ 

Hue angle H ¼ tan
1

  b a

ð1Þ ð2Þ

Whiteness index (WI) and yellowness index (YI) of UBF samples were calculated according to Pathare, Opara, and Al-Said (2013) by the following equations:

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi WI ¼ 100
ð100
LÞ  þa2 þ b2

ð3Þ



YI ¼

2.2.5. Bulk density and tapped density The bulk density (qbulk ) of UBF was determined by measuring the weight of the UBF and corresponding volume. Approximately 1 g of UBF was transferred to 10 mL graduated cylinder. The bulk density was calculated by dividing the mass of the UBF by the volume occupied in the cylinder. For the tapped density (qtapped), graduated cylinder was tapped to constant volume by a glass rod. The volume of UBF was used in mass/volume calculation (Jinapong, Suphantharika, & Jamnong, 2008). 2.2.6. Carr index (flowability) and Hausner ratio (cohesiveness) Flowability and cohesiveness of UBF were expressed in terms of Carr index (CI) (Carr, 1965) and Hausner ratio (HR) (Hausner, 1967), respectively. Both CI and HR were figured out from the bulk (qbulk) and tapped (qtapped) densities of the UBF as shown in the equations below:

CI ¼

qtapped
qbulk  100 qtapped

HR ¼

2.2. Methods

142; 86b L

ð4Þ

181

qtapped qbulk

ð5Þ

ð6Þ

Powders with HR under 1.2 are classified as low cohesive group; HR between 1.2 and 1.4 are considered as intermediate cohesive and HR over 1.4 are stated as high cohesive group (Hausner, 1967). Flowability of powders with CI < 15 are classified as very good; 15 < CI < 20 as good; 20 < CI < 35 as fair; 35 < CI < 45 bad and CI > 45 as very bad (Carr, 1965). 2.2.7. Wettability Wettability of the UBF samples were determined according to A/S Niro Atomizer (1978) with some modifications. 100 mL distilled water was poured into 250 mL beaker. A glass funnel held was set over the beaker on a ring stand with the height of 10 cm between the bottom of the funnel and the water surface. A cylinder glass rod was placed inside the funnel to block the opening of the funnel. UBF sample (0.1 ± 0.0001 g) was placed around the rod and the rod was lifted while the stop watch was started simultaneously. Finally, the time for UBF to become completely wet was recorded. 2.2.8. Water solubility index (WSI) and water absorbtion index (WAI) WSI and WAI of UBF was performed according to RodriguezAmbriz, Islas-Hernández, Agama-Acevedo, Tovar, and Bello-Pérez (2008) with some modifications. 1 g UBF was transferred to 35 mL of distilled water. Mixture was homogenized with a mechanical homogenizer (JANKE & KUNKEL IKAÒ-Labortechnick RW20, Staufen, Germany) at level 10 for 5 min, and the solution was transferred in a 50 mL pre-weighted centrifuge tube, tube was incubated at room temperature for 1 h and centrifuged at 3000g for 20 min in a centrifuge device (Universal 32R/Hettich-Zentrifungen D-78532 Tuttlingen, Germany). The tube was drained into a pre-weighted petri dish at a 45° angle for 10 min and dried for 5 h at 105 °C to constant weight. The percentage of residue (mass of soluble flour, g) with respect to the amount of dry UBF sample (mass of total flour, g) used in the test was taken as WSI. WAI was calculated as the mass of centrifuged precipitate (g) per UBF sample used in the test (g). 2.2.9. Oil holding capacity (OHC) The method of (Rodríguez-Ambriz et al., 2008) was used to determine OHC. 25 mL olive-oil were added to 1 g of UBF, vortexed with mixer (WiseMix VM-10 Vortex Mixer, New Zealand) for 2 min and incubated at room temperature for 1 h. Tube was centrifuged at 3000 g for 20 min. The supernatant was decanted and the tube

182

N. Savlak et al. / Food Chemistry 213 (2016) 180–186

was allowed to drain for 10 min at a 45° angle. The centrifuged precipitate was weighted and OHC was calculated as g oil per g dry UBF sample. 2.2.10. Antioxidant activity 2.2.10.1. Extraction. 1 g UBPF was mixed with 25 mL methanol: water mixture (v/v, 1:1) incubated at 50 °C for 15 min in oscillating water bath. The mixture was then centrifugated at 4000 rpm and 20 °C for 10 min. Supernatant was transferred into a 100 mL volumetric flask. The same procedure was applied to the residue two times more. The volume of the flask was completed to 100 mL by methanol:water mixture (v/v, 1:1). The content of the volumetric flask was filtered through Whatman No: 1 and 0.45 lm respectively and kept in light proof 50 mL centrifuge tubes at
86 °C until the antioxidant analysis were performed (Aksoylu, Cagindi & Köse 2015). 2.2.10.2. 1,1-diphenyl-2-picrylhydrazyl (DPPH) scavenging activity. DPPH assay was performed according to the method of Brand-Williams, Cuvelier, and Berset (1995) with some modifications. 0.0154 g DPPH was dissolved in methanol and completed to 500 mL final volume to obtain 78,1 lM DPPH solution. UBF extracts (50, 100,150 and 200 lL) were completed with methanol to 200 lL final volume and were mixed with 3.8 mL DPPH solution to reach final volume of 4 mL. Mixtures were allowed to stand in the dark for 60 min at room temperature. Then, the absorbance was measured at 515 nm.% inhibition values against the amount of sample in extracts (lg) and% inhibition values against the amount of Trolox (lg) in different concentrations of Trolox (5–600 lM) were represented in the graph to obtain two different linear equations. Then, the slope of the sample regression equation was divided by the slope of Trolox regression equation. Results were expressed in mg TE/g dry UBF. Data presented were the average of three parallel readings and three replicates. 2.2.10.3. Ferric reducing antioxidant power (FRAP) assay. Ferricreducing antioxidant power (FRAP) was conducted according to the method described by Liu, Qiu, Ding, and Yao (2008) and modified by Wang, Zhang, and Mujumdar (2012). Briefly, 100 lL extract was mixed with 6 mL freshly prepared FRAP reagent and the mixture was incubated at 37 °C for 30 min. Absorbance against distilled water blank at 593 nm was measured using Shimadzu UV-601 (Japan) spectrofotometer. Different concentrations of FeSO47H2O (0.05–2 mmol/L) were used for the calibration curve. Results of UBF were expressed as mmol of Fe (II)/g oil-free dry matter. Data presented were the average of three parallel readings and three replicates. 2.2.10.4. Total polyphenol content. Total polyphenol concentration in the extracts was determined according to Li et al. (2006). 200 lL extract was mixed with 2.5 mL Folin-Ciocalteu reagent (10%, v/v), vortexed for 15 s. After allowing to stand for 5 min in the dark, 5 mL sodium carbonate solution (7.5%, w/v) was added and the mixture was again allowed to stand in the dark for

60 min at room temperature. Absorbance of the samples were measured at 760 nm using Shimadzu UV-601 (Japan) spectrofotometer. Total polyphenol content of extracts obtained from UBF having different particle size distributions was determined by manipulating the regression equation of gallic acid calibration curve (5–300 mg/L, y = 0.014x + 0.003, r2 = 0.997). Total polyphenol content of UBF belonging to different particle sizes were expressed as mg GAE/g oil-free dry matter. Data presented were the average of three parallel readings and three replications. 2.2.11. Statistical analysis All measurements and analysis were done in three replications and at least three parallels. Results were expressed as mean ± standard deviations in the Tables. The one-way analysis of variance (ANOVA) and the Duncan’s Multiple Comparison Test (p < 0.05) were used to determine least significant differences among the mean values of physical and functional properties of UBF for different particle size distributions. Data were analyzed using Statistical Analysis Systems version 8.2 (1999–2001) software, SAS Institue Inc., Cary NC (SAS, 2001). 3. Results and discussion 3.1. Physico-chemical properties Physico-chemical properties (pH, soluble solids, water activity, moisture) of UBF were given in Table 1. Moisture content and water activity of powder products are critical properties that can affect other physical and chemical properties of foods. They are also critical factors for shelf life and food stability. Moisture content of size fractions did not change among each other statistically (p > 0.05). The effect of particle size distribution on water activity was statistically important (p < 0.05). All the size fractions were within the limits of safe storage in terms of moisture content and water activity. Rodriguez-Ambriz et al. (2008) reported 6.0% moisture for unripe banana flour and 6.8% for fiber rich unripe banana flour. Rayo et al. (2015) found 3.97% moisture content and 0.35 water activity for green banana flour. Particle size distribution did not effect pH values of UBF statistically (p > 0.05). Our results were agreeable with Rayo et al. (2015) and Alkarkhi et al. (2011) who reported mean pH values of 5.94 and 5.06 respectively. Soluble solids (°Brix) of UBF changed between 0.52 and 0.74 and the effect of particle size was statistically important (p < 0.05). Particles in the range of 316–500 lm were different from other fractions statistically. Similarly, Alkarkhi et al. (2011) reported °Brix values between 1.03 and 1.43 in their study. 3.2. Color Profile L⁄ value varied significantly for all size fractions (Table 2.). L⁄ value decreased by increasing particle size (r2 =
0.94) and all the size fractions were different from each other statistically (p < 0.0001). The increase of L⁄ values with reduced particle size

Table 1 Physico-chemical properties (moisture content, water activity, pH, total soluble solid) of Unripe Banana Flour (UBF) having different particle size distributions. Particle size of UBF (lm)

Moisture content (%)

aw

pH

TSS (°Brix)

<212 212-315 316-500 501-700

9.07 ± 0.347a 8.98 ± 0.234a 9.19 ± 0.468a 8.75 ± 0.525a

0.424 ± 0.015bc 0.454 ± 0.019a 0.441 ± 0.012ab 0.407 ± 0.006c

5.665 ± 0.011a 5.586 ± 0.038b 5.582 ± 0.039b 5.629 ± 0.048ab

0.58 ± 0.035b 0.52 ± 0.052b 0.74 ± 0.089a 0.60 ± 0.005b

TSS: Total soluble solids. Values are means ± standard deviation (n = 3). Means with the same letter in a column are not statistically different from each other (p < 0.05).

183

N. Savlak et al. / Food Chemistry 213 (2016) 180–186 Table 2 Color profile of Unripe Banana Flour (UBF) having different particle size distributions. Particle size of UBF (lm) <212 212–315 316–500 501–700

L⁄

a⁄ a

85.00 ± 0.346 82.26 ± 0.225b 81.23 ± 0.211c 80.18 ± 0.675d

b⁄ c

1.83 ± 0.059 2.29 ± 0.044ab 2.37 ± 0.030a 2.23 ± 0.039b

C⁄ b

11.81 ± 0.125 13.37 ± 0.107a 13.73 ± 0.186a 13.47 ± 0.550a

H° b

11.95 ± 0.131 13.57 ± 0.098a 13.93 ± 0.178a 13.66 ± 0.549a

WI a

81.18 ± 0.216 80.24 ± 0.259c 80.17 ± 0.250c 80.56 ± 0.020b

YI a

80.82 ± 0.203 77.67 ± 0.238b 76.61 ± 0.275c 75.43 ± 0.070d

19.85 ± 0.149d 23.22 ± 0.249c 24.20 ± 0.299b 24.69 ± 0.120a

Values are means ± standard deviation (n = 3). Means with the same letter in a column are not statistically different from each other (p < 0.05).

Table 3 Bulk density (qbulk), Tapped Density(qtapped), Carr Index (CI) and Hausner Ratio (HR) of unripe banana flour (UBF) for different particle size distributions. Particle size of UBF (lm) <212 212–315 316–500 501–700

qbulk (kg m
3) a

251.59 ± 0.014 146.20 ± 0.003d 163.04 ± 0.000c 192.28 ± 0.003b

qtapped (kg m
3)

CI (%)

HR

660.38 ± 0.016a 538.24 ± 0.038b 573.08 ± 0.045b 588.21 ± 0.000b

61.92 ± 1.502c 72.93 ± 2.074a 71.65 ± 1.953a 66.33 ± 0.687b

2.63 ± 0.098c 3.68 ± 0.322a 3.63 ± 0.171a 3.03 ± 0.030b

Values are means ± standard deviation (n = 3). Means with the same letter in a column are not statistically different from each other (p < 0.05).

was due to increase in surface area that allows more reflection of light (Ahmed, Al-Jassar, Thomas, 2015). In our study, this was confirmed by the increase in whiteness index from 75.43 to 80.12 by decreasing particle size. When a⁄ and b⁄ values in our study were examined, it was found that a⁄ and b⁄ values decreased by increasing particle size (r2 = 0.72, r2 = 0.73 respectively) and only the finest particles (<212 lm) were different from other size fractions statistically (p < 0.01). Similarly, Ahmed, Taher, Mulla, Al-Hazza, and Luciano (2016a) reported increase in L⁄ value and decrease in a⁄ and b⁄ values of lentil flour between 74 and 210 lm. They explained the significant decrease in color a⁄ and b⁄ values of smaller particles with the loss of pigment during milling process. According to the researchers, the decrease in yellowness of finer particle sizes may be associated with a decrease in protein content. Our color results were in the line with Anyasi et al. (2015) who studied the effect of organic acid pretreatment on unripe banana flour and reported L⁄ values in the range of 79.39–85.29. On the other hand, Alkarkhi et al. (2011) stated that L⁄ value changed between 64.37 and 79.25 in their study. Lower L⁄ values in their study can be attributed to drying process which was done in oven. Wrolstad and Smith (2010) stated that C⁄ value of a food increases with increment of pigment concentration and decreases remarkably as the samples go darker. Similar to a⁄ and b⁄ values, the finest particles under 212 lm size were different from other size fractions in terms of C⁄ and H⁄ (p < 0.01). Our results were comparable with Anyasi et al. (2015). Results of color analysis showed that there were significant differences in WI and YI of all size fractions (Table 2.). WI decreased and YI increased by increasing particle size (r2 =
0.95 and r2 = 0.89 respectively) and all the size fractions were different from each other statistically (p < 0.0001). WI results in this study were agreeable with Anyasi et al. (2015). Increase in whiteness index is associated with increasing surface area that reflects more light which was formerly explained by Ahmed, Al-Jassar, and Thomas (2015). 3.3. Bulked and tapped density Bulk density, tapped density, Hausner ratio and Carr index of UBF were given in Table 3. Ahmed, Al-Foudari, Al-Salman, and Almusallam (2014) also stated that particle size had a significant impact on bulk density and generally, bulk density increased as particle size decreased. Goula, Adamopoulos, and Kazakis (2004) explained the decrease by the stickiness of particles during dehy-

dration and also by product agglomeration. In our study, the effect of particle size on bulk density of UBF was statistically important (p < 0.0001). In line with the literature, particles under 212 lm had the highest bulk density which indicates particle mass of smaller particles was denser as stated by Ahmed et al. (2016a) and it first decreased than increased by increasing particle size. Subba and Katawal (2013) reported increasing bulk density values for increasing particle size of rice flour in the range of 359–746 lm. Similarly, in our study, bulk density of particles between 212 and 700 lm increased by increasing particle size. Tapped density of the particles under 212 lm was the highest very similar to bulk density and was different from other particle ranges statistically (p < 0.01). This can be explained by Abdullah and Geldart (1999) who propounded that free flowing powders had lower consolidation properties while a fine and cohesive powder collapsed quickly due to tapping. Rayo et al. (2015) reported bulk density values of 514.76 kg m
3 for unripe banana flour and 329.19 kg m
3 for agglomerated unripe banana flour. In the same study tapped densities were 652.06 and 403.07 kg m
3 respectively. Our results were comparable with Rayo et al. (2015). 3.4. Carr index (flowability) and Hausner ratio (cohesiveness) In terms of handling properties, all the sizes had similar flow characteristics and were considered very cohesive powders by their Hausner ratio given in Table 2 as they were over 1.4. This is in accordance with their high Carr index which is over 45 indicating that their flowability was very poor. Particle size distribution affects powder properties such as bulk density, tapped density, flowability, hydration properties (Muttakin, Kim, & Lee, 2015). The relation of poor flowability with small particle sizes can be explained by large surface area per unit mass of powder. There is more contact surface area between powder particles available for cohesive forces and frictional forces to resist flow (Fitzpatrick, 2005). Thus, the intermolecular forces strengthens, reducing the ease of flow of the powder. Rayo et al. (2015) found that Carr index changed between 18.3 and 20.95 and Hausner ratio between 1.22 and 1.27 in unripe banana flour presenting intermediate flowability. 3.5. Wettability Reconstitution properties of UBF in terms of wettability was given in Table 4. All the size fractions had different wettability and were different from each other statistically (p < 0.0001). A

184

N. Savlak et al. / Food Chemistry 213 (2016) 180–186

Table 4 Water solubility index (WSI), water holding capacity (WHC), oil holding capacity (OHC) and wettability of unripe banana flour (UBF) for different particle size distributions. Particle size of UBF (lm) <212 212–315 316–500 501–700

WSI (g soluble UBF/g dry UBF) a

WAI (g water/g dry UBF) c

0.074 ± 0.000 0.078 ± 0.001a 0.075 ± 0.000a 0.091 ± 0.019a

OHC (g oil/g dry UBF) c

2.922 ± 0.004 3.580 ± 0.020b 3.702 ± 0.061b 4.604 ± 0.418a

Wettability (s) 83.40 ± 9.828a 1.318 ± 0.039b 1.437 ± 0.122b 1.308 ± 0.090b

1.804 ± 0.002 3.680 ± 0.001a 3.548 ± 0.002a 3.115 ± 0.003b

Values are means ± standard deviation (n = 3). Means with the same letter in a column are not statistically different from each other (p < 0.05).

Table 5 Total phenolic content and antioxidant activity of Unripe Banana Flour (UBF) for different particle size distributions. Particle size of UBF (lm)

Total phenolic (mg GA/g UBF)

FRAP value (Fe(II) mmol/g UBF)

DPPH scavenging capacity (% inhibition)

DPPH scavenging capacity (mg Trolox equivalents/g UBF)

<212 212–315 316–500 501–700

0.23 ± 0.003c 0.31 ± 0.005a 0.19 ± 0.003d 0.27 ± 0.005b

27.91 ± 0.416c 36.83 ± 0.624a 23.15 ± 0.208d 32.37 ± 0.000b

11.37 ± 0.578b 13.45 ± 0.462a 8.700 ± 0.594c 12.00 ± 0.854b

2.63 ± 0.493a 2.99 ± 0.174a 1.74 ± 0.477b 2.82 ± 0.199a

Values are means ± standard deviation (n = 3). Means with the same letter in a column are not statistically different from each other (p < 0.05).

negative correlation between particle size and wettability (wetting time) (r2 =
0.75) was observed. Wetting time increased with decreasing particle size. Similarly, Jinapong et al. (2008) reported the same relation in spray-dried powders which had very poor wettability probably due to their very small size (<25 mm). In this study, particles under 212 lm had the lowest rate of wettability (83.40 s) with the highest wetting time scores. Other particle ranges had very high wettability (1.31–1.44 s) showing that reconstitution and instant properties were good. Schubert (1987) reported that powder flowability, wettability and dispersability increased with the increase in particle size and fine particles had tendency to form lumps on the liquid surface as a result of strong adhesion forces. This situation explains higher wetting time for smallest particle range in our study. 3.6. Water solubility index (WSI), water absorbtion index (WAI), oil holding capacity (OHC) WAI measures the volume occupied by the starch granule after swelling in excess of water, while WSI determines the amount of free molecules leached out from the starch granule in addition to excess water (Ortiz, Carvalho, Ascheri, Ascheri, & Andrade, 2010). It is reported by many researchers that WAI is influenced by the extent of disintegration of native starch granules and is linked to the physical state of starch, dietary fiber and proteins in the unripe banana flour (Alkarkhi et al., 2011). Table 4 gives water holding and oil holding capacity, as well as water solubility. The effect of particle size on WAI was statistically important (p < 0.0001). WAI of particles increased by increasing particle size (r2 = 0.94). This relation was also supported by Ahmed, Al-Attar, and Arfat (2016b) who studied effects of particle size on some properties of chestnut flour. Particles under 212 lm had the lowest WAI, while 501–700 lm had the highest. Results of WAI obtained in the present study were comparable with Anyasi et al. (2015). Cadden (2006) stated that reducing the particle size of wheat bran decreased water-holding capacity, due to the collapse of its fiber matrix. Lower WAI results belonging to finer particles can be explained by the collapse of fiber matrix as a result of size reduction. WSI of different size fractions did not differ from each other statistically (p > 0.05). It was observed that solubility of unripe banana flour in water was very low. Another functional property of banana flour is OHC. Similar to WAI, particle size was statistically important (p < 0.0001) and particles under 212 lm had the lowest OHC. However, there was not a

significant correlation between particle size and OHC. Results belonging to finest particles (<212 lm) were similar to that of Anyasi et al. (2015) with OHC in the range of 1.07–2.00 g oil/g. Rodriguez-Ambriz et al. (2008) reported that OHC is related to the hydrophilic nature of starch existing in the unripe banana flour and it is primarily due to the physical trapping of oil within the starch structure through non-covalent bonds. 3.7. Antioxidant activity Antioxidant activity of UBF extracts determined by two different methods (DPPH and FRAP) and total polyphenol content (TPC) were presented in Table 5. Total phenolic content changed between 0.19 and 0.31 mg GAE/g dry UBF. Effect of particle size on TPC was statistically important (p < 0.0001). Several researchers studied TPC of unripe banana flour. Menezes et al. (2011) reported 50.65 mg GAE/100 g dry weight total polyphenol content for unripe banana flour (Musa acuminata, var. Nanicão). Sarawong, Schoenlechner, Sekiguchi, Berghofer, and Ng (2014) stated that native unripe banana flour had total phenolic content of 220.3 mg GAE/100 g dry weight. Anyasi et al. (2015) reported total phenolic content of three unripe banana flour between 841.59 and 1130.39 mg GAE/100 g dry weight. Total phenolic content of unripe banana flour changed in a wide range in the literature. For the FRAP assay, effect of particle size on results of antioxidant activity was statistically important (p < 0.0001) and all the size fractions were different from each other. FRAP value of UBF changed between 23.15 and 36.83 mmol Fe (II)/g UBF. Particles between 212 and 315 lm had the highest antioxidant activity (36.83 mmol Fe(II)/g dry UBF) whereas particles between 316 and 500 lm had the lowest (23.15 mmol Fe(II)/g dry UBF). Results from the literature include 358.67 lmol of TE/100 g dry weight (Menezes et al., 2011) and 6.56 mmol Fe (II)/100 g dry weight (Sarawong et al., 2014). Our results were higher than the literature. Similar to FRAP assay, DPPH radical scavenging activity of UBF extracts was significantly affected by particle size distribution (p < 0.0001) and ranged between 1.74 and 2.99 mg Trolox equivalent/g dry UBF. DPPH inhibition percentage was between 8.7 and 13.45%. Anyasi et al. (2015) reported DPPH activity of three unripe banana flour between 0.12 and 1.02 mg GAE/g dry weight. Sarawong et al. (2014) reported that DPPH value of unripe banana flour was 2.48 mmol TE/100 g dry weight. The results of all the antioxidant analysis in this study showed that particles between 212 and 315 lm had the highest and particles between 316 and

N. Savlak et al. / Food Chemistry 213 (2016) 180–186

500 lm had the lowest antioxidant activity as seen in Table 5. There was not a definite increasing or decreasing trend in terms of antioxidant analysis among the samples. However, Liu et al. (2015) reported a decrease in total phenolic content of rice up to 55.6% and antioxidant activity up to 92.8% by increasing degree of milling. Some researchers in the literature reported the positive correlation of total phenolic content with DPPH and FRAP assays (Alothman, Bhat, & Karim, 2009; Ribeiro, Barbosa, Queiroz, Knodler & Schieber, 2008). In this research, phenolic content was strongly correlated with FRAP (r2 = 0.98) and moderately correlated with DPPH values (r2 = 0.74). Correlation coefficient between total phenolic content and DPPH inhibition% (r2 = 0.89) was also high. DPPH assay was less correlated with phenolic content in comparison to FRAP. The reason of this was explained by Kaur and Kapoor (2001) who stated that DPPH assay may not give a true picture of total antioxidant capacity as compared to FRAP due to its less sensitiveness towards hydrophilic antioxidants and the interaction of antioxidant compound with DPPH depends on its structural conformation.

4. Conclusion The particle size distribution of flour has an important role on its functional properties and the quality of end products. This study showed that particle size significantly effected color, water absorbtion index and wettability of unripe banana flour. L⁄ value decreased, a⁄ and b⁄ values decreased by increasing particle size (r2 =
0.94, r2 = 0.72, r2 = 0.73 respectively). Particles under 212 lm had the lowest rate of wettability (83.40 s). Other particle ranges had very high wettability (1.31–1.44 s) showing that reconstitution properties were better. WSI of all particle ranges were very low showing that UBF was not water soluble. A negative correlation between particle size and wettability (r2 =
0.75) and positive correlation between particle size and water absorption index (r2 = 0.94) was observed. Carr Index, indicating flowability and Hausner Ratio, indicating cohesiveness were over the upper limits for very bad flowability and high cohesiveness showing that handling properties can be improved for all particle size ranges. As a result, handling properties of UBF is an aspect requiring further improvement.

Chemicals Methanol solvent (Pubchem CID:887) was of spectroscopy grade and purchased from Merck KGaA (Darmstadt, Germany). Folin–Ciocalteu’s phenol reagent (Pubchem CID:516996), Sodium hydrogen carbonate (Pubchem CID:516892), Glacial acetic acid (Pubchem CID:176), Hydrochloric acid 37% (Pubchem CID:313); Iron(III) chloride (Pubchem CID:166033) were obtained from Merck KGaA (Darmstadt, Germany). 2,2-diphenyl-1picrylhydrazyl (DPPH) (Pubchem CID:2735032), Iron(II) sulfate heptahydrate (Pubchem CID:62662), 2,4,6-Tris(2-pyridyl)-striazine (TPTZ) (P98%) (Pubchem CID:77258) were obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). 6-hydroxy-2,5,7,8-tetrame thylchroman-2-carboxylic acid (Trolox) (Pubchem CID: 40634), Sigma-Aldrich Co. (St. Louis, MO, USA) was kindly supplied by Prof. Dr. Neriman Bag˘datlıog˘lu (Department of Food Engineering, Celal Bayar University, Manisa, Turkey).

Conflict of interest statement The authors would like to indicate that there is no conflict of interest for this journal.

185

References A/S Niro Atomizer, Copenhagen, Denmark (1978). Determination of wettability. In I. H. Sørensen, J. Krag, J. Pisecky, & V. Westergaard (Eds.), Analytical methods for dry milk products (pp. 26–27). Copenhagen: De Forenede Trykkerier A/S. Abdullah, E. C., & Geldart, D. (1999). The use of bulk density measurements as flowability indicators. Powder Technology, 102, 151–165. Agama-Acevedo, E., Islas-Hernández, J. J., Pacheco-Vargas, G., Osorio-Díaz, P., & Bello-Pérez, L. A. (2012). Starch digestibility and glycemic index of cookies partially substituted with unripe banana flour. LWT – Food Science and Technology, 46, 177–182. Ahmed, J., Al-Attar, H., & Arfat, Y. A. (2016b). Effect of particle size on compositional, functional, pasting and rheological properties of commercial water chestnut flour. Food Hydrocolloids, 52, 888–895. Ahmed, J., Al-Foudari, M., Al-Salman, F., & Almusallam, A. S. (2014). Effect of particle size and temperature on rheological, thermal, and structural properties of pumpkin flour dispersion. Journal of Food Engineering, 124, 43–53. Ahmed, J., Al-Jassar, S., & Thomas, L. (2015). A comparison in rheological, thermal, and structural properties between Indian Basmati and Egyptian Giza rice flour dispersions as influenced by particle size. Food Hydrocolloids, 48, 72–83. Ahmed, J., Taher, A., Mulla, M. Z., Al-Hazza, A., & Luciano, G. (2016a). Effect of sieve particle size on functional, thermal, rheological and pasting properties of Indian and Turkish lentil flour. Journal of Food Engineering, 186, 34–41. Alkarkhi, A. F. M., Bin Ramli, S., Yong, Y. S., & Easa, A. M. (2011). Comparing physicochemical properties of banana pulp and peel flours prepared from green and ripe fruits. Food Chemistry, 129, 312–318. Alothman, M., Bhat, R., & Karim, A. A. (2009). Antioxidant capacity and phenolic content of selected tropical fruits from Malaysia, extracted with different solvents. Food Chemistry, 115(3), 785–788. Anyasi, T. A., Jideani, A. I. O., & Mchau, G. R. A. (2015). Effect of organic acid pretreatment on some physical, functional and antioxidant of flour obtained from three unripe banana cultivars. Food Chemistry, 172, 515–522. Aurore, G., Parfait, B., & Fahrasmane, L. (2009). Bananas, raw materials for making processed food products. Trends in Food Science & Technology, 20(2), 78–91. Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free radical method to evaluate antioxidant activity. LWT – Food Science and Technology, 28(1), 25–30. Cadden, M. A. (2006). Comparative effects of particle size reduction on physical structure and water binding properties of several plant fibers. Journal of Food Science, 52(6), 1595–1599. Carr, R. L. (1965). Evaluating flow properties of solids. Chemical Engineering, 72, 163–168. Emaga, T. H., Andrianaivo, R. H., Wathelet, B., Tchango, J. T., & Paquot, M. (2007). Effects of the stage of maturation and varieties on the chemical composition of banana and plantain peels. Food Chemistry, 103, 590–600. Fitzpatrick, J. J. (2005). Food powder flowability. In C. Onwulata (Ed.), Encapsulated and powdered foods (pp. 247–260). Boca Raton, FL: CRC Press. Goula, A. M., Adamopoulos, K. G., & Kazakis, N. A. (2004). Influence of spray drying conditions on tomato powder properties. Drying Technology, 22, 1129–1151. Hausner, H. H. (1967). Friction conditions in a mass of metal powder. International Journal of Powder Metallurgy, 3, 7–13. Ho, L. H., Aziz, N. A. A. A., & Azahari, B. (2013). Physico-chemical characteristics and sensory evaluation of wheat bread partially substituted with banana (Musa acuminata X balbisiana cv. Awak) pseudo-stem flour. Food Chemistry, 139, 532–539. Jinapong, N., Suphantharika, M., & Jamnong, P. (2008). Production of instant soymilk powders by ultrafiltration, spray drying and fluidized bed agglomeration. Journal of Food Engineering, 84(2), 194–205. Karim, R., & Sultan, M. T. (2015). Yellow alkaline noodles: An introduction. SpringerBriefs in Food, Health, and Nutrition, 19–33. Kaur, C., & Kapoor, H. C. (2001). Review Antioxidants in fruits and vegetables – The millennium’s health. International Journal of Food Science & Technology, 36(7), 703–725. Li, Y., Guo, C., Yang, J., Wei, J., Xu, J., & Cheng, S. (2006). Evaluation of antioxidant properties of pomegranate peel extract in comparison with pomegranate pulp extract. Food Chemistry, 96, 254–260. Liu, L., Guo, J., Zhang, R., Wei, Z., Deng, Y., Guo, J., & Zhang, M. (2015). Effect of degree of milling on phenolic profiles and cellular antioxidant activity of whole brown rice. Food Chemistry, 185, 318–325. Liu, H., Qiu, N., Ding, H., & Yao, R. (2008). Polyphenols contents and antioxidant capacity of 68 Chinese herbals suitable for medical or food uses. Food Research International, 41(4), 363–370. Menezes, E. W., Tadini, C. C., Tribess, T. B., Zuleta, A., Binaghi, J., Pak, N., & Lajolo, F. M. (2011). Chemical Composition and Nutritional Value of Unripe Banana Flour (Musa acuminata, var. Nanicão). Plant Foods for Human Nutrition, 66(3), 231–237. Muttakin, S., Kim, M. S., & Lee, D. U. (2015). Tailoring physicochemical and sensorial properties of defatted soybean flour using jet-milling technology. Food Chemistry, 187, 106–111. Onwuka, G. I., & Onwuka, N. D. (2005). The effects of ripening on the functional properties of plantain and plantain based cake. International Journal of Food Properties, 8(2), 347–353. Ortiz, J. A. R., De Carvalho, C. W. P., Ascheri, D. P. R., Ascheri, J. L. R., & De Andrade, C. T. (2010). Effect of sugar and water contents on non-expanded cassava flour extrudates. Ciência E Tecnologia de Alimentos, 30(1), 205–212.

186

N. Savlak et al. / Food Chemistry 213 (2016) 180–186

Pathare, P. B., Opara, U. L., & Al-Said, F. A. J. (2013). Colour measurement and analysis in fresh and processed foods: A review. Food and Bioprocess Technology, 6(1), 36–60. Rayo, L. M., Carvalho, L. C., Sardá, F. A. H., Dacanal, G. C., Menezes, E. W., & Tadini, C. C. (2015). Production of instant green banana flour (Musa cavendischii, var. Nanicão) by a pulsed-fluidized bed agglomeration. LWT – Food Science and Technology, 63, 461–469. Rodriguez-Ambriz, S. L., Islas-Hernández, J. J., Agama-Acevedo, E., Tovar, J., & BelloPérez, L. A. (2008). Characterization of a fiber-rich powder prepared by liquefaction of unripe banana flour. Food Chemistry, 107, 1515–1521. Sarawong, C., Schoenlechner, R., Sekiguchi, K., Berghofer, E., & Ng, P. K. W. (2014). Effect of extrusion cooking on the physicochemical properties, resistant starch, phenolic content and antioxidant capacities of green banana flour. Food Chemistry, 143, 33–39. SAS (2001). Statistical analysis software (SAS) systems for windows. Cary, NC, USA: SAS Institue Inc..

Schubert, H. (1987). Food particle technology. Part 1: properties of particles and particulate food systems. Journal of Food Engineering, 6, 1–32. Subba, D., & Katawal, S. B. (2013). Effect of particle size of rice flour on physical and sensory properties of Sel-roti. Journal of Food Science and Technology, 50(1), 181–185. Suntharalingam, S., & Ravindran, G. (1993). Physical and biochemical properties of green banana flour. Plant Foods for Human Nutrition, 43, 19–27. Wang, Y., Zhang, M., & Mujumdar, A. S. (2012). Influence of green banana flour substitution for cassava starch on the nutrition, color, texture and sensory quality in two types of snacks. LWT – Food Science and Technology, 47(1), 175–182. Wrolstad, R. E., & Smith, D. E. (2010). Color analysis. In S. S. Nielsen (Ed.), Food analysis (pp. 573–586). Zenebon, O., & Pascuet, N.S. (2005). Me´todos fı´sico-quı´micos para ana´ lises de alimentos do Instituto Adolfo Lutz. 4a. Ediça~o, Brası´lia, p. 1018.