Peel-inclusive flours from out-of-specification immature Cavendish and Saba bananas: process optimization for improved flour lightness and product characterization

1 Introduction

Bananas are one of the most important fruit commodities worldwide due to their wide variety of uses, high trade value, and large production volume. The Philippines ranked as the third leading global exporter and remained the largest banana exporter in Asia, accounting for approximately 60% of the region’s banana exports on average (Food and Agriculture Organization, 2025). In 2025, the Philippines produced about 8.61 million metric tons of bananas, primarily from Cavendish (Musa acuminata AAA) and Saba (M. acuminata × balbisiana Colla) varieties, with production volumes of 4.30 and 2.56 million metric tons, respectively (Philippine Statistics Authority, 2026).

Despite its large production volume, the Philippine banana industry remains vulnerable to significant postharvest losses, resulting in economic losses and environmental concerns. Due to its geographical location, the country is highly exposed to frequent tropical cyclones and other extreme weather events that pose substantial risks to banana production. Between 2000 and 2021, these natural hazards caused more than 352 billion pesos (approximately USD 5.7 billion) worth of damage to the country’s agricultural sector (Ortiz, 2023). To minimize crop damage and economic losses during typhoons and other natural hazards, banana growers often practice anticipated or early harvesting of mature and other marketable fruits. Although mature bananas continue to ripen and develop acceptable aroma and flavor after harvest, those harvested at immature stages often exhibit poor quality upon ripening and fail to meet fresh market standards (Food and Agriculture Organization, 2018). This may lead to generation of substantial volumes of out-of-specification (OOS) bananas, which refer to prematurely harvested fruits with substandard ripening quality. If not properly utilized, OOS bananas can contribute to environmental pollution and economic losses for both farmers and processors (Kumar et al., 2019). Since Cavendish and Saba banana varieties represent the predominant banana cultivars produced in the Philippines, their large-scale cultivations also contribute significantly to the generation of OOS fruits. This highlights the need to develop sustainable utilization strategies that can convert these underutilized resources into value-added products.

One approach to preserve and utilize OOS Cavendish and Saba bananas is by processing them into flour. Aside from extending shelf life and facilitating storage and transportation, banana flour production expands the utilization of bananas in various food applications (Manalo et al., 2023). Unripe bananas, including both the peel and pulp, possess valuable nutritional and functional properties that make them suitable for flour production. They are rich in resistant starch, dietary fiber, phenolic compounds, vitamins, and minerals (de Angelis-Pereira et al., 2016; Menezes et al., 2011), making them a promising functional food ingredient. However, one of the major challenges in producing green banana flour, particularly from whole fruits, is the development of undesirable dark coloration which can limit consumer acceptance and industrial applicability (Moon et al., 2020). This discoloration is primarily attributed to enzymatic browning, which is an oxidative reaction that occurs when fruit tissues are damaged, allowing the polyphenol oxidase (PPO) enzyme to interact with phenolic compounds in the presence of oxygen (Sarkar et al., 2024). The inclusion of the peel, which is known to contain high levels of phenolic compounds (Someya et al., 2002; Sulaiman et al., 2024), can further intensify browning in flour samples, making color preservation a major challenge in whole green banana flour production. Because visual appearance is a primary factor influencing consumer acceptance and the potential application of whole banana flour, improving color quality represents an essential first step toward increasing the marketability of peel-inclusive banana flour.

Various studies have investigated the production of green banana flour; however, these have primarily focused on the production of pulp flour from mature fruits, product characterization, or determination of the effects of a limited number of processing parameters on flour properties. For instance, Hasmadi et al. (2021) and Kumar et al. (2019) characterized the physicochemical, structural, and thermal properties of green banana flours from different cultivars produced under fixed processing conditions. Similarly, Tribess et al. (2009) examined the effects of drying conditions on the thermal properties and resistant starch content of green Cavendish banana pulp flour. More recently, Manalo et al. (2023) optimized blanching time and antioxidant chemical pre-treatment concentration for pulp flour produced from OOS Cavendish bananas. While these studies provided valuable insights into green banana flour production and the effects of specific processing treatments, they were not able to establish how the combined effects and interactions among multiple processing variables influence color preservation in peel-inclusive flour, particularly that produced from OOS immature bananas. Because color is influenced by the interaction of multiple processing variables, optimizing only one or a few factors may not adequately minimize discoloration. Moreover, the effectiveness of these processing factors may vary depending on the characteristics of the banana cultivar. Differences in peel-to-pulp ratio and physicochemical composition between Cavendish and Saba bananas may influence their response to processing conditions and the resulting flour properties. Therefore, identifying the optimum combination of processing conditions for each cultivar is necessary to consistently produce peel-inclusive banana flour with desirable quality.

This study was conducted to optimize processing conditions, including blanching time and temperature, soaking time, and drying temperature, to improve the color attributes of whole green banana flour produced from OOS immature Cavendish and Saba bananas, and to determine how the optimized conditions affected the physicochemical characteristics, proximate composition, and functional properties of the flours. To assess the potential market competitiveness of the produced banana flours, the treated samples were compared with commercially available banana flours in selected analyses, specifically physicochemical characteristics, proximate composition, and techno-functional properties. This study supports the sustainable valorization of rejected banana resources and contributes to the development of high-quality banana flour with potential applications in diverse food products.

2 Materials and methods

2.1 Raw materials and chemicals

The out-of-specification (OOS) immature Cavendish and Saba bananas were purchased from OMG Farm in Bay, Laguna, Philippines. All samples were obtained from the same harvest batch to ensure uniformity in maturity. The samples consisted of physiologically immature bananas harvested at <12 weeks after flowering, characterized by a dark green peel color and firm texture. The maturity characteristics and initial specifications of the samples are presented in Supplementary Table S1. Meanwhile, the three commercial banana flour samples analyzed were purchased from local supermarkets. The product labels indicated that all three samples were prepared from the pulp of Cavendish bananas. Among the commercial samples, one was manufactured in the Philippines (Commercial 1), while the other two samples were manufactured in Australia (Commercial two and Commercial 3). Food-grade sodium metabisulfite and citric acid used in the production of banana flour were obtained from Kemrad Incorporated (Quezon City, Philippines). All other chemicals and reagents that were used in the various analyses were of analytical grade.

2.2 Production of banana flour

OOS immature Cavendish and Saba bananas were harvested and processed on the same day to minimize postharvest changes. Before processing, the fruits were thoroughly washed with tap water and sanitized by immersion in a 200-ppm sodium hypochlorite solution for 5 min to remove visible dirt and reduce microbial load. In producing the banana flours, the method of Bunyameen et al. (2020) was followed with slight modifications. Initially, whole banana fingers were soaked in 0.1% (w/v) sodium metabisulfite solution for 1 h, followed by water blanching at varying time and temperature combinations. After blanching, the bananas were rinsed with tap water and sliced to approximately 3 mm thickness using a stainless-steel multifunction fruit slicer. The slices were immediately immersed in an anti-browning solution containing 0.20% (w/v) sodium metabisulfite and 0.75% (w/v) citric acid, ensuring that the slices were completely submerged throughout the soaking period. Following pre-treatment, the slices were drained and dehydrated in a forced-air cabinet dryer at the varying drying temperatures for approximately 15 h or until the moisture content reached 15.5% or below, based on preliminary trials. The dried samples were cooled at ambient temperature for 15–30 min before being milled into flour using a pin mill. The resulting flour (treated) was then passed through a 40-mesh sieve, vacuum-packed, and stored at −18 °C until analysis. For comparison, control flours (untreated) from each banana variety were prepared using the same processing steps but without the blanching and chemical pretreatments, and were analyzed alongside the treated samples.

2.3 Process optimization for OOS Cavendish and Saba banana flours

Response surface methodology (RSM) was employed to optimize the processing conditions for producing peel-inclusive flours from OOS immature Cavendish and Saba bananas. The experimental design and statistical analyses were carried out using Design-Expert® software version 12 (Stat-Ease Inc., Minneapolis, MN, USA). A four-factor, three-level Box–Behnken design (BBD) was used to evaluate the effects of blanching time (5, 10, and 15 min), blanching temperature (60, 70, and 80 °C), soaking time (5, 10, and 15 min), and drying temperature (50, 60, and 70 °C). The selected processing conditions were established based on preliminary trials and previous studies on green banana flour processing (Jeet et al., 2015; Oluwalana et al., 2011; Anyasi et al., 2017; Manalo et al., 2023; Cândido et al., 2023; Jaramillo Garcés et al., 2023; Khoozani et al., 2019; Farias et al., 2020; Kamble et al., 2022). The ranges were selected to encompass conditions that could effectively reduce enzymatic browning while avoiding excessive heat exposure that may negatively affect the properties of the final product. These conditions were also within practical operating ranges for banana flour production. The experimental design for each variety consisted of 27 treatment combinations, including three center-point replicates to estimate experimental error and assess model adequacy.

The optimum processing conditions were determined using the numerical optimization function of the software, with flour lightness (L) set as the response to be maximized. Lightness was selected as the optimization response because color is a critical quality attribute of whole green banana flour, particularly with peel inclusion, where enzymatic browning can negatively affect the visual quality and acceptability of the final product. Maximizing L was therefore used as an indicator of improved visual quality, while other quality attributes were evaluated after optimization. The processing conditions corresponding to the highest desirability value were selected as the optimum solution. To validate the predictive capability of the developed model, experiments were conducted under the optimized conditions, and the experimentally obtained response values were compared with the model-predicted values. The agreement between the predicted and experimental results was used to evaluate the adequacy and validity of the fitted model.

2.4 Physicochemical properties

2.4.1 Color

The color of the banana flour samples was measured using a chromameter (CR-400, Konica Minolta, Japan). Color parameters were recorded in the CIELAB color space as L, a, and b, where L represents lightness, a* represents the red–green axis, and b* represents the yellow–blue axis (Imawati et al., 2024). The corresponding chroma (C) and hue angle (H°) values were calculated from the measured L, a, and b coordinates using Equations 1 and 2, respectively.

C h r o m a C = a * 2 + b * 2 (1)

H u e A n g l e H ° = arctan b * / a * (2)

2.4.2 pH, total soluble solids (TSS), and water activity

For pH and TSS analyses, 1 g of flour was accurately weighed and homogenized with 10 mL of distilled water. The pH of the resulting suspension was measured using a benchtop pH meter (Biobase, China). Then, the suspension was centrifuged, and the supernatant was filtered to obtain a clear sample for TSS determination. The TSS was measured using a hand refractometer (ATAGO N-1E, Atago, Japan) by placing a drop of the filtrate on the prism and recording the reading, expressed as °Brix. The water activity of the banana flours was determined using a benchtop water activity meter (LabSwift-aw, Novasina, Lachen, Switzerland). The sample was placed into the sample cup to ensure adequate surface coverage before measurement. The cup was inserted into the instrument chamber, and the water activity value was recorded after the reading reached equilibrium.

2.4.3 Resistant starch content

The resistant starch (RS) contents of the banana flour samples were determined using a resistant starch assay kit (Megazyme Ltd., Bray, Ireland) in accordance with AOAC Official Method 2002.02 and AACC International Method 32–40.01.

2.4.4 Apparent amylose content

The apparent amylose content of the banana flour samples was determined following the iodine colorimetry method described by Tamura et al. (2014) with slight modifications. About 100 mg of the banana flour sample was mixed with 1 mL of 95% ethanol and 9 mL of 1 M NaOH in a 100-mL volumetric flask. The mixture was heated in a boiling water bath for 10 min, cooled, diluted to volume with distilled water, and stored at room temperature (25 °C) for 23 h. Then, an aliquot of 5 mL was transferred to a second 100-mL volumetric flask containing approximately 70 mL of distilled water, followed by the addition of 1 mL of 1 M acetic acid and 2 mL of iodine solution, before diluting to volume with distilled water. After incubation for 20 min, its absorbance was measured at 620 nm against the blank. The apparent amylose content was calculated from a calibration curve prepared using potato amylose as the standard.

2.5 Proximate composition

The proximate composition of the banana flour samples, including moisture, crude ash, crude fat, crude fiber, and crude protein, was determined following the standard procedures of the Association of Official Analytical Chemists (2000). Crude protein content was estimated by multiplying the measured nitrogen content by a conversion factor of 6.25. The carbohydrate fraction expressed as nitrogen-free extract (NFE) was determined by difference after accounting for all other proximate components.

2.6 Functional properties

2.6.1 Techno-functional properties

2.6.1.1 Water holding capacity

Water holding capacity (WHC) of the banana flour samples was determined using the method of Noor Aziah et al. (2012) with slight modifications. Approximately 1 g of flour was dispersed in 10 mL of distilled water and allowed to hydrate at room temperature for 30 min. The hydrated suspension was centrifuged at 1080 g for 20 min, and the supernatant was carefully decanted. The remaining hydrated sediment was weighed, and WHC was expressed as the amount of water retained per gram of flour (Equation 3).

W a t e r h o l d i n g c a p a c i t y W H C , g / g = w t . o f s e d i m e n t − w t . o f s a m p l e , g w t . o f s a m p l e , g (3)

2.6.1.2 Oil absorption capacity

Oil absorption capacity (OAC) of the banana flour samples was determined according to the method of Noor Aziah et al. (2012) with slight modifications. Approximately 1 g of flour was combined with 10 mL of vegetable oil and mixed for 10 min to ensure complete dispersion. The suspension was then allowed to stand at room temperature for 30 min before centrifugation at 1080 g for 20 min. The unabsorbed oil was carefully decanted, and the remaining sediment was weighed. OAC was expressed as the amount of oil retained per gram of flour and was calculated using Equation 4.

O i l a b s o r p t i o n c a p a c i t y O A C , g / g = w t . o f s e d i m e n t − w t . o f s a m p l e , g w t . o f s a m p l e , g (4)

2.6.1.3 Swelling capacity and water solubility index

Swelling capacity (SC) and water solubility index (WSI) of the banana flour samples were determined following the method of Kakar et al. (2022) with slight modifications. About 0.1 g of flour was dispersed in 10 mL of distilled water and homogenized using a vortex mixer. The suspension was heated in a water bath at 85 °C for 30 min with intermittent mixing at 5-min intervals, cooled to room temperature, and centrifuged at 1080 g for 20 min. The supernatant was carefully decanted into a pre-weighed crucible, while the hydrated sediment was weighed to determine SC using Equation 5.

S w e l l i n g c a p a c i t y S C , g / g = w t . o f s e d i m e n t , g w t . o f s a m p l e , g (5)

The collected supernatant was dried in an oven at 100 °C ± 5 °C overnight, followed by cooling in a desiccator for 30 min before weighing. The WSI was calculated from the weight of the dried solids recovered in the crucible using Equation 6.

W a t e r s o l u b i l i t y i n d e x W S I , g / g = w t . o f c r u c i b l e w i t h s o l i d s − w t . o f c r u c i b l e , g w t . o f s a m p l e , g (6)

2.6.1.4 Bulk and tapped densities

Bulk and tapped densities were determined following the methods of Alam et al. (2023). Approximately 10 g of flour was transferred into a 100-mL graduated cylinder, and the initial volume was recorded after gently settling the sample. Bulk density was calculated as the ratio of the sample weight to its occupied volume (Equation 7).

B u l k / t a p p e d d e n s i t y , g / m L = w t . o f f l o u r , g v o l u m e o c c u p i e d , m L (7)

The same flour sample was used to determine tapped density. The graduated cylinder was tapped 350 times, or until the sample volume reached a constant value. The final packed volume was recorded, and tapped density was calculated as the ratio of sample mass to the packed volume (Equation 7).

2.6.1.5 Carr’s index and hausner ratio

Carr’s Index (CI) and Hausner Ratio (HR) were determined to characterize the flowability and cohesiveness of the flour samples, respectively. The indices were calculated from the obtained bulk and tapped densities according to Alam et al. (2023) using Equations 8 and 9.

C a r r i n d e x C I , % = t a p p e d d e n s i t y − b u l k d e n s i t y , g / m L t a p p e d d e n s i t y , g / m L * 100 (8)

H a u s n e r r a t i o H R = t a p p e d d e n s i t y , g / m L b u l k d e n s i t y , g / m L (9)

2.6.2 Pasting profile

The pasting profile of the banana flour samples was obtained using a Rapid Visco Analyzer (RVA 4800, Perten Instruments, Hägersten, Sweden) based on the method of Cheng et al. (2024), with slight modifications. A flour suspension (16% w/w) was prepared by mixing 4.5 g of banana flour sample with 24 g of distilled water. The suspension was then subjected to a programmed heating–cooling cycle comprising equilibration at 50 °C for 1 min, heating to 95 °C at 12 °C/min, holding for 4.5 min, cooling to 50 °C at 12 °C/min, and a final holding period of 2 min. The paddle speed ranged from 160 to 960 rpm during the analysis. Peak, trough, breakdown, final, and setback viscosities, together with pasting temperature and peak time, were obtained from the RVA curve.

2.6.3 Thermal properties

The thermal properties of banana flour samples were analyzed using a Differential Scanning Calorimeter (TA Instruments, New Castle, DE, USA) based on the method of Tiozon et al. (2024) with slight modifications. Approximately 5 mg of sample was hydrated with distilled water at a 1:2 (w/w) ratio, sealed in hermetic aluminum pans, and heated from 25 °C to 120 °C at 10 °C/min against an empty reference pan. The onset (T_o), peak (T_p), and end (T_e) temperatures, as well as the gelatinization enthalpy (ΔH), were determined using TRIOS software (TA Instruments, New Castle, DE, USA).

2.6.4 Safety evaluation

To evaluate the safety of the optimized treated flour samples, a composite sample was prepared for each variety by homogenizing portions from the three individual production batches. The samples were then subjected to residual sulfite content and microbiological analyses.

2.6.4.1 Residual sulfite content

Residual sulfite analysis was submitted to an accredited third-party testing laboratory and analyzed using AOAC Official Method 990.28 (Optimized Monier-Williams Method). Results were expressed as sulfur dioxide (SO₂) content in parts per million (ppm).

2.6.4.2 Microbiological quality

Microbiological analyses were performed according to the method of the US Food and Drug Administration Bacteriological Analytical Manual (BAM). Coliform count was determined by pour plating in violet red bile agar following BAM Chapter 4 (US Food and Drug Administration, 2020) while yeast and mold count was determined using potato dextrose agar following BAM Chapter 18 (US Food and Drug Administration, 2001).

2.7 Statistical analysis

Process optimization data were statistically modeled and analyzed using Design-Expert® software (Stat-Ease Inc., Minneapolis, MN, USA). Statistical analyses for banana flour characterization were performed using GraphPad Prism version 10.1.1 (Dotmatics, Boston, MA, USA). The bananas used in this study were obtained from a single harvest batch, and flour samples were prepared through three independent processing runs (n = 3). Each processing run produced a separate flour batch, which was considered as an independent experimental replicate. Characterization analyses were performed on each flour batch individually, except for residual content and microbiological analyses, and results are expressed as mean ± standard deviation. Data were evaluated for normality using the Shapiro–Wilk test and for homogeneity of variances using Levene’s test. Comparisons between treated and untreated flour samples within each cultivar (Cavendish and Saba) were performed using paired t-tests and comparisons among treated OOS Cavendish, treated OOS Saba, and commercial flour samples were performed using one-way ANOVA followed by Tukey’s multiple comparison test. Statistical significance was established at p < 0.05.

3 Results

3.1 Process optimization for OOS Cavendish and Saba banana flours

The optimization experiments were conducted using a 27-run Box-Behnken experimental design for each cultivar to evaluate the effects of blanching time and temperature, soaking time, and drying temperature on the lightness (L) of peel-inclusive banana flour produced from OOS immature Cavendish and Saba bananas. The experimental design matrix and the corresponding L values are presented in Supplementary Table S2. The observed L* values ranged from 61.156 to 89.398 for OOS Cavendish flour and 78.624 to 88.598 for OOS Saba flour, demonstrating the influence of the processing variables on flour color and providing an adequate response range for model fitting.

The linear model was identified as the suggested model by the software for both varieties. During model selection, two-factor interaction and quadratic models were also evaluated. However, these higher-order models were not selected because their inclusion did not significantly improve model adequacy, as indicated by the insignificant sequential model p-values (p > 0.05) and less favorable agreement between adjusted and predicted R ² values.

Table 1 summarizes the analysis of variance (ANOVA) and model statistics for the linear models fitted to the lightness of OOS Cavendish and Saba banana flours. Both models were significant (p < 0.0001) and exhibited a non-significant lack-of-fit (p > 0.05). For OOS Cavendish flour, blanching time, blanching temperature, soaking time, and drying temperature significantly influenced lightness, with an R ² of 0.6989, an adjusted R ² of 0.6387, and a predicted R ² of 0.5574. In contrast, the OOS Saba model yielded an R ² of 0.7331, an adjusted R ² of 0.6738, and a predicted R ² of 0.5731, with blanching time, blanching temperature, and soaking time identified as significant model terms. Despite the moderate R ² values, the selected models were considered suitable for identifying the effects of processing variables and determining optimum conditions, as supported by the significant model terms, non-significant lack-of-fit, and acceptable agreement between adjusted and predicted R ² values. Furthermore, the diagnostic plots confirmed the adequacy of the regression models, with the normal probability plots showing that the residuals closely followed a normal distribution pattern and the residuals versus predicted plots indicating random distribution of residuals with no evident trends or non-constant variance (Figure 1). The Box- Cox plot also did not recommend any transformation (λ = 1), which further supports the selection of the linear model for describing the relationship between processing variables and the response.

TABLE 1. Analysis of variance (ANOVA) and model summary for the linear model predicting the lightness (L*) of out-of-specification (OOS) immature Cavendish and Saba banana flours.

The fitted linear regression equations describing the effects of the processing variables on the lightness of OOS Cavendish and OOS Saba banana flours are presented in Equations 10 and 11, respectively. The regression coefficients are expressed in actual factor units, where the sign indicates the direction of the effect on lightness, while the magnitude represents the predicted change in lightness per unit change of each processing variable.

O O S C a v e n d i s h F l o u r L i g h t n e s s L * = 102.06317 − 0.153510 b l a n c h i n g t i m e − 0.138409 b l a n c h i n g t e m p e r a t u r e + 0.143200 s o a k i n g t i m e − 0.086569 d r y i n g t e m p e r a t u r e (10)

O O S S a b a F l o u r L i g h t n e s s L * = 98.30766 − 0.315959 b l a n c h i n g t i m e − 0.236403 b l a n c h i n g t e m p e r a t u r e + 0.408717 s o a k i n g t i m e + 0.045685 d r y i n g t e m p e r a t u r e (11)

Equation 10 indicates that soaking time had a positive linear coefficient, suggesting that increasing soaking time within the experimental range increased the predicted lightness (L) of OOS Cavendish banana flour. In contrast, blanching time, blanching temperature, and drying temperature had negative linear coefficients, indicating that increasing these processing variables decreased the predicted lightness. For OOS Saba banana flour, Equation 11 shows that blanching time and blanching temperature had negative linear coefficients, indicating that increasing either factor reduced the predicted lightness of the sample. Meanwhile, soaking time and drying temperature had positive coefficients, which suggests that longer soaking and higher drying temperatures increased its predicted lightness. Among all variables, soaking time exhibited the greatest positive effect while blanching time had the strongest negative influence. These findings are consistent with the model graphs presented in Figure 2. Significant factors are characterized by steeper slopes, indicating a greater influence on the predicted lightness values. In contrast, insignificant factors, such as drying temperature for OOS Saba banana flour, exhibit relatively flat lines, reflecting minimal changes in the predicted response.

Numerical optimization was performed to identify the optimum processing conditions for OOS Cavendish and Saba banana flours based on the specified constraints and response goals. The employed optimization criteria are summarized in Table 2. Blanching time (5–15 min), blanching temperature (60 °C–80 °C), soaking time (5–15 min), and drying temperature (50 °C–70 °C) were constrained within the experimental ranges, while lightness was set to be maximized, with target values of 83.78–89.40 for OOS Cavendish flour and 79.24–88.60 for OOS Saba flour.

TABLE 2. Optimization criteria for determining the optimum processing conditions for out-of-specification (OOS) immature Cavendish and Saba banana flours.

The optimum conditions were determined using the desirability function approach of the software, based on the specified optimization criteria. For OOS Cavendish banana flour, the identified optimum conditions consist of blanching time of 5 min, blanching temperature of 60 °C, soaking time of 11 min, and drying temperature of 59 °C. The corresponding optimum conditions for OOS Saba banana flour were similar and differ only on soaking time (10 min) and drying temperature (60 °C). The selected solutions for both varieties had a desirability value of 1.00, indicating that the specified factor constraints were satisfied and that the predicted lightness values reached the upper desirability criterion defined for the response.

The optimized processing parameters were experimentally validated through five independent trial runs. Results showed that the measured L* values fell within the corresponding prediction intervals for both OOS Cavendish (88.141–90.775) and OOS Saba banana flours (87.192–91.552). The experimentally obtained L* values ranged from 89.058 to 89.826 for OOS Cavendish flour and 87.975 to 88.970 for OOS Saba banana flour, confirming the adequacy and predictive reliability of the developed models for optimizing the processing conditions.

3.2 Physicochemical properties of OOS Cavendish and Saba banana flours

Table 3 shows the physicochemical properties, including color (L, a, b, chroma, and hue angle), pH, water activity, total soluble solids, resistant starch, and apparent amylose of untreated, treated, and commercial banana flour samples. For both OOS Cavendish and OOS Saba banana varieties, the application of the optimized processing conditions significantly affected some of the physicochemical characteristics of the flour. In terms of color, application of the optimized processing conditions significantly increased the lightness (L) of both banana flours by 13.68% (78.74 ± 0.33 to 89.51 ± 0.39) for OOS Cavendish flour and 13.20% (77.14 ± 0.37 to 87.32 ± 0.10) for OOS Saba flour, indicating that the optimized flours exhibited a significantly lighter color than the untreated samples. The treated samples also showed lower a* values, and higher b*, chroma, and hue angle values, reflecting a brighter and more intense yellow color. Application of the optimized processing conditions also significantly reduced the pH of both banana flours by 17.77% (5.46 ± 0.01 to 4.49 ± 0.01) in OOS Cavendish flour and 13.91% (5.39 ± 0.00 to 4.64 ± 0.03) in OOS Saba flour. In terms of water activity, only the OOS Saba flour exhibited a significant increase following the optimized processing conditions, while no significant change was observed for OOS Cavendish flour. Nevertheless, all flour samples exhibited low water activity values, ranging from 0.18 to 0.27. Lastly, no significant differences in TSS were observed between the untreated and treated flours of either banana variety, with values ranging from 13.20 to 14.80 °Brix.

TABLE 3. Physicochemical properties of banana flour samples.

Values are presented as mean ± SD (n = 3). Asterisk () indicates a significant difference between the untreated and treated flour samples from OOS Cavendish and Saba banana varieties. Values with different superscript letters in the same column indicate a significant difference (p* < 0.05) among treated and commercial banana powder samples. OOS- out-of-specification.

Compared with the commercial banana flours, the treated OOS Cavendish and OOS Saba flours exhibited significantly higher L, b, chroma, and hue angle values. Moreover, the treated flours had more negative a* values than the commercial samples, which indicates a shift from reddish tones in the commercial flours to greener hues in the treated samples. The higher b* and chroma values of the treated flours also indicated a more intense yellow color compared with the commercial products. Overall, the treated peel-inclusive banana flours exhibited a lighter, brighter, and more saturated color than both the untreated and commercial banana flours. The flour samples analyzed in this study are shown in Figure 3. In addition, the treated flours had significantly lower pH values than the commercial flours. Water activity also differed significantly among the treated and commercial flours, although all values remained below 0.30. Lastly, both treated OOS Cavendish and OOS Saba flours exhibited higher TSS values than the commercial flour samples, with no significant difference observed between the two treated flours.

In terms of resistant starch (RS), no significant difference was observed between the treated and untreated flours of either OOS banana variety. However, the RS contents of the treated OOS banana flours (39.08–47.43 g/100 g) were significantly higher than those of the commercial banana flours (32.55–36.03 g/100 g). For apparent amylose, only OOS Saba exhibited a significant increase, from 14.87 ± 0.29 g/100 g in the untreated flour to 16.28 ± 0.5 g/100 g in the treated flour (9.48% increase). In contrast, although no significant change was observed after treatment, OOS Cavendish flour exhibited the highest apparent amylose content (18.58 ± 0.19 g/100 g). Nevertheless, the apparent amylose of both treated OOS flours was comparable to those of the commercial samples.

3.3 Proximate composition

Table 4 presents the proximate composition of untreated, treated, and commercial banana flour samples. Application of the optimized processing conditions significantly affected the moisture, crude ash, and crude fat contents of both OOS Cavendish and OOS Saba banana flours. Moisture content increased by 20.23% (5.14 ± 0.05 to 6.18 ± 0.05 g/100g) and 11.82% (5.33 ± 0.11 to 5.96 ± 0.04 g/100g), while crude fat increased by 32.54% (1.26 ± 0.06 to 1.67 ± 0.06 g/100g) and 35.47% (2.34 ± 0.22 to 3.17 ± 0.21 g/100g) in OOS Cavendish and OOS Saba flours, respectively. In contrast, crude ash content decreased by 7.70% (7.66 ± 0.09 to 7.07 ± 0.05 g/100g) in OOS Cavendish flour and 16% (8.25 ± 0.08 to 6.93 ± 0.01 g/100g) in OOS Saba flour. No significant differences were observed in the crude protein, crude fiber, or nitrogen-free extract (NFE) contents of either banana variety following the optimized processing conditions.

TABLE 4. Proximate composition of banana flour samples.

Values are presented as mean ± SD (n = 3) (g/100 g of sample). All the parameters are in dry basis except moisture. Asterisk () indicates a significant difference between the untreated and treated flour samples from OOS Cavendish and Saba banana varieties. Values with different superscript letters in the same column indicate a significant difference (p* < 0.05) among treated and commercial banana powder samples. OOS- out-of-specification.

Among the treated and commercial samples, the treated OOS Cavendish and OOS Saba flours exhibited significantly higher moisture and crude ash contents than all commercial banana flours. The treated OOS Saba flour had the highest crude fiber (3.89 ± 0.11 g/100 g) and crude fat (3.17 ± 0.21 g/100 g) contents, while the optimized OOS Cavendish flour exhibited the highest crude protein content (6.55 ± 0.97 g/100 g). In contrast, the commercial banana flours contained significantly higher NFE (90.83–91.23 g/100 g) than the optimized OOS flours (81.68–82.66 g/100 g). Overall, the optimized peel-inclusive banana flours were characterized by higher ash, protein, fiber, and fat contents but lower NFE than the commercial banana flours.

3.4 Functional properties of OOS Cavendish and Saba banana flours

3.4.1 Techno-functional properties

Table 5 presents the techno-functional properties of the untreated, treated, and commercial banana flour samples. Results showed that the applied treatments significantly affected selected functional properties of both banana varieties. In OOS Cavendish flour, water-holding capacity (WHC) decreased by 2.67% (3.00 ± 0.01 to 2.92 ± 0.03 g/g), while no significant change was observed in OOS Saba flour. Moreover, oil-absorption capacity (OAC) remained unchanged in both varieties following the applied optimized process. In contrast, swelling capacity (SC) increased by 27.81% (11.29 ± 0.61 to 14.43 ± 0.65 g/g) in OOS Cavendish flour and 18% (11.72 ± 0.33 to 13.83 ± 0.15 g/g) in OOS Saba flour, while the water solubility index (WSI) increased by 23.53% (0.17 ± 0.02 to 0.21 ± 0.01 g/g) and 17.65% (0.17 ± 0.02 to 0.20 ± 0.01 g/g), respectively. No significant differences were observed in the bulk density, tapped density, Carr’s index, or Hausner ratio of either banana variety following the optimized processing conditions.

TABLE 5. Techno-functional properties of banana flour samples.

Values are presented as mean ± SD (n = 3). Asterisk () indicates a significant difference between the untreated and treated flour samples from OOS Cavendish and Saba banana varieties. Values with different superscript letters in the same column indicate a significant difference (p* < 0.05) among treated and commercial banana powder samples. OOS- out-of-specification.

Compared with the commercial banana flours, the treated OOS Cavendish and OOS Saba flours generally exhibited significantly higher WHC, WSI, and SC values. In terms of OAC, treated OOS Saba flour had the highest value (1.10 ± 0.03 g/g), while treated OOS Cavendish flour exhibited OAC values comparable to the commercial samples. On the other hand, the treated flours had significantly lower bulk and tapped densities but higher Carr’s index and Hausner ratio values than the commercial flours.

3.4.2 Pasting properties

Table 6 summarizes the pasting properties of untreated and treated OOS Cavendish and Saba banana flours while Figure 4 shows the viscogram of the samples. Results showed that the optimized process did not significantly affect the pasting properties of OOS Cavendish flour including its pasting temperature, peak, trough, breakdown, setback, and final viscosities. In contrast, pretreatment significantly decreased the trough and final viscosities of OOS Saba flour, while its pasting temperature, peak viscosity, breakdown viscosity, and setback viscosity remained unaffected.

TABLE 6. Pasting temperature (°C) and pasting properties (cP) of out-of-specification (OOS) immature Cavendish and Saba banana flour samples.

Values are presented as mean ± SD (n = 3). Asterisk () indicates a significant difference between the untreated and treated flour samples from OOS Cavendish and Saba banana varieties. Values with different superscript letters in the same column indicate a significant difference (p* < 0.05) between the treated samples from OOS Cavendish and Saba banana varieties. OOS- out-of-specification.

Comparison of the treated OOS Cavendish and OOS Saba flours revealed no significant differences in most pasting properties, except for final viscosity, which was significantly higher in the treated OOS Saba flour.

3.4.3 Thermal properties

Table 7 presents the thermal properties of the untreated and treated OOS Cavendish and OOS Saba banana flour samples, while the corresponding thermograms are shown in Figure 5. Significant differences were observed between the untreated and treated flours of both banana varieties. For OOS Cavendish flour, the optimized process significantly increased the onset temperature from 60.82 °C to 73.90 °C. In contrast, the peak and end temperatures slightly decreased from 78.76 °C to 78.42 °C and 86.04 °C to 85.49 °C, respectively. For OOS Saba flour, the optimized process decreased the onset and peak temperatures from 77.31 °C to 76.26 °C and 82.32 °C to 81.66 °C, respectively, while the gelatinization enthalpy slightly increased from 12.15 to 12.84 J/g. No significant change was observed in the end temperature following the optimized process.

TABLE 7. Thermal properties of out-of-specification (OOS) immature Cavendish and Saba banana flour samples.

Values are presented as mean ± SD (n = 3). Asterisk () indicates a significant difference between the untreated and treated flour samples from OOS Cavendish and Saba banana varieties. Values with different superscript letters in the same column indicate a significant difference (p* < 0.05) between the treated samples from OOS Cavendish and Saba banana varieties. OOS- out-of-specification.

Between the treated banana flours, OOS Saba flour exhibited significantly higher onset, peak, and end temperatures than treated OOS Cavendish flour. Meanwhile, no significant difference was observed in gelatinization enthalpy between the treated OOS Cavendish and OOS Saba flours.

3.4.4 Safety evaluation

Table 8 shows the residual sulfite content, expressed as sulfur dioxide, and microbiological quality of the optimized flour samples. The treated OOS Cavendish and Saba banana flours exhibited low sulfur dioxide contents of 6.86 ± 0.00 ppm and 7.49 ± 0.01 ppm, respectively. Both flour samples also had low microbial counts, with coliform and yeast counts of <10 CFU/g, and mold count of <100 CFU/g.

TABLE 8. Sulfur dioxide content and microbial counts of treated out-of-specification (OOS) immature Cavendish and Saba banana flour samples.

4 Discussion

4.1 Effect of optimized processing conditions in the lightness (L*) of OOS Cavendish and Saba banana flours

The identified optimum processing conditions demonstrate that the combination of relatively mild pretreatments (blanching for 5 min at 60 °C and soaking in a sodium metabisulfite-citric acid solution for 10–11 min), and drying temperature (59 °C–60 °C) effectively improved the lightness of peel-inclusive flours produced from OOS Cavendish and Saba bananas. The high lightness (L*) values obtained under these conditions indicate that browning reactions were successfully minimized throughout processing. This improvement may be attributed to the combined effects of enzyme inactivation, inhibition of oxidative reactions, and the use of moderate drying temperatures, which collectively reduced both enzymatic and non-enzymatic browning while limiting excessive thermal degradation in the samples (Fan et al., 2023; Lim and Wong, 2018; Xiao et al., 2017).

Pretreatments are an essential step in banana flour production because they help preserve the quality of fresh banana tissues before drying by minimizing discoloration and maintaining desirable physicochemical characteristics (Rammohan et al., 2025). Fresh banana pulp and peel are highly susceptible to enzymatic browning immediately after peeling and slicing due to the exposure of phenolic compounds to oxidative enzymes. Without appropriate pretreatments, these reactions rapidly produce dark pigments that negatively affect the appearance and marketability of the resulting flour (Pathak et al., 2021). Blanching and the application of chemical anti-browning agents are commonly employed to suppress browning reactions before drying. In this study, these pretreatments were integrated into the optimized process for peel-inclusive flours from OOS Cavendish and Saba bananas to minimize enzymatic browning and preserving flour quality.

Blanching is a widely used mild heat pretreatment that reduces the activity of polyphenol oxidase (PPO) and peroxidase (POD), the enzymes responsible for browning in fruits and vegetables (Xiao et al., 2017). As determined through process optimization, blanching at 60 °C for 5 min was identified as part of the optimum processing conditions for both OOS Cavendish and Saba bananas. Compared with previous studies on banana flour production, this treatment is relatively mild, as most reported blanching conditions generally employed higher blanching temperatures. For example, Khar et al. (2003) reported that hot-water blanching at 97 °C for 5 min resulted in complete inactivation of PPO and POD, which effectively prevented enzymatic browning during the subsequent dehydration process. Similarly, Belkacemi (2022) and Ayo-Omogie et al. (2021) used blanching at 100 °C for 5 min prior to banana flour production. In contrast, the present findings demonstrate that blanching at a lower temperature (60 °C) for the same duration was sufficient to improve the color of peel-inclusive banana flours, as evidenced by the high lightness values obtained. These results suggest that effective color preservation can be achieved under milder blanching conditions, which may also reduce undesirable thermal effects associated with more severe heat treatments.

The effectiveness of the mild blanching treatment was further enhanced by soaking the banana slices in a sodium metabisulfite–citric acid solution for 10–11 min. These compounds are among the most widely used chemical anti-browning agents because they are inexpensive, readily available, and effective in preserving the color of plant-based products. Their combined application provides complementary mechanisms for browning inhibition. Citric acid inhibits enzymatic browning by lowering the pH of plant tissues and chelating metal ions required for polyphenol oxidase (PPO) activity, while the acidic environment also slows non-enzymatic browning reactions, including the Maillard reaction, which proceeds less readily under acidic conditions (Fan et al., 2023). Sodium metabisulfite, on the other hand, reacts with quinones formed during phenolic oxidation, preventing their polymerization into brown melanin pigments (Lim and Wong, 2018). The effectiveness of these anti-browning agents in banana flour production has been demonstrated in several studies (Salazar et al., 2022; Legowo et al., 2024; Manalo et al., 2023; Mushtaq et al., 2026).

The optimization results indicate that soaking in the sodium metabisulfite–citric acid solution for approximately 10–11 min was sufficient to maximize flour lightness, which demonstrates the synergistic effect of the two anti-browning agents. Similar observations were reported by Hanh et al. (2025), who found that treating yam slices with a combination of 0.50% sodium metabisulfite and 0.15% citric acid for 15 min reduced PPO and POD activities by 89% and 90%, respectively, resulting in a substantially lower browning index. Beyond improving product color, the combined use of citric acid and sodium metabisulfite may also reduce the amount of sulfite required to achieve effective browning control. This is particularly advantageous because sulfites are regulated food additives and may trigger adverse reactions in sulfite-sensitive individuals (Vally and Misso, 2012). Despite the repeated application of sodium metabisulfite during the pretreatment process, residual sulfite analysis, expressed as sulfur dioxide, of the optimized whole green banana flours showed minimal sulfite retention in the final product (6.86–7.49 ppm), with values well below the maximum allowable limit of 200 ppm (Codex Alimentarius Commission, 2026). The low residual sulfite levels may be attributed to the combined effects of blanching and washing, which can promote sulfite removal through leaching, as similarly observed by Hajri et al. (2022). In addition, the relatively short soaking duration may have limited sulfite intake into the tissue matrix, resulting in reduced retention after processing. A similar observation was reported by Kumoro and Hidayat (2018), who found that soaking time influenced sulfite retention in durian flour, with shorter exposure periods resulting in lower residual sulfite levels.

In addition to the pretreatments, the optimized drying condition also contributed to the improved color quality of the peel-inclusive banana flours. Drying is a critical processing step because it reduces moisture content to enhance shelf stability while simultaneously influencing the extent of browning reactions that occur during dehydration (Alam et al., 2023). Excessively high drying temperatures can accelerate Maillard reactions, caramelization, and thermal degradation of natural pigments, leading to darker flour. On the other hand, excessively low temperatures may prolong drying time, which allows residual enzymatic activity and oxidative reactions to continue before complete moisture removal. The optimized drying temperature of 59 °C–60 °C for 15 h therefore appears to provide a balance between efficient dehydration and color preservation. These conditions are consistent with those reported by Tribess et al. (2009), Teeta et al. (2023), Manalo et al. (2023), and Salazar et al. (2022), who also employed moderate drying temperatures to produce banana flours with acceptable color characteristics.

The identified optimum conditions in this study were established based on lightness (L*) as the primary optimization response, reflecting the objective of improving the color characteristics of peel-inclusive banana flours. Unlike previous optimization studies on banana flour production that commonly focused on pulp-derived flours or evaluated individual processing parameters, this study demonstrates that the use of mild blanching, chemical pretreatment, and moderate drying conditions can effectively improve the color quality of peel-inclusive flours produced from OOS Cavendish and Saba bananas. This highlights the potential of utilizing whole green banana materials, including the peel component, which remains relatively underexplored despite its potential contribution to nutritional and functional properties. However, because the optimization approach considered a single response, other flour attributes were not incorporated as optimization criteria in defining the optimum conditions. Evaluation of these attributes is therefore necessary to provide a more comprehensive understanding of the overall quality and application potential of the developed peel-inclusive banana flours.

4.2 Effect of the optimized process on the properties of OOS Cavendish and Saba banana flours and comparison with commercial samples

The optimized processing conditions were further evaluated based on their effects on the various properties of OOS Cavendish and Saba banana flours. The optimized flours were also compared with commercially available banana flours to provide a practical reference for their quality characteristics and assess their potential as value-added banana-based ingredients. Although the commercial flours used may differ from the developed whole green banana flours in terms of raw materials used and processing employed, this comparison provides insights into their relative characteristics and potential relevance for future food applications. Beyond improving color quality, this comprehensive characterization provides new knowledge on the nutritional, functional, and processing-related properties of peel-inclusive banana flour produced from OOS green bananas, an area that remains less explored compared with conventional pulp-based banana flours.

4.2.1 Physicochemical properties

The optimized processing conditions significantly influenced several physicochemical characteristics of both OOS Cavendish and OOS Saba banana flours. In addition to increased lightness discussed in the previous section, the treated flours exhibited lower a* values and higher b*, chroma, and hue angle values, indicating reduced reddish discoloration and a more pronounced yellow color. These changes demonstrate that the optimized processing conditions effectively preserved the natural pigments of green banana tissues while minimizing browning during processing (Fan et al., 2023; Lim and Wong, 2018; Xiao et al., 2017).

The treated samples also exhibited significantly higher L, b, chroma, and hue angle values and more negative a* values than all commercial samples, indicating a lighter, brighter, and more saturated yellowish color. The darker color observed in the commercial flours may be associated with differences in banana cultivar, maturity stage, peel removal practices, and pretreatments used (Udomkun et al., 2021). Differences in banana cultivar and maturity stage may influence phenolic composition and oxidative enzyme activity, affecting pigment formation during processing. Peel removal practices may also contribute to color variation, as banana peel contains higher levels of phenolic compounds and pigments that can alter flour appearance when retained. In addition, variations in pretreatment conditions may affect enzyme inactivation and phenolic oxidation, further influencing the final color of banana flours. Since color is one of the primary quality attributes influencing consumer acceptance and product applicability, the lighter color characteristics of the optimized peel-inclusive flours suggest that the developed process successfully addressed the common challenge of browning in whole green banana flour and may support its potential application in food product development. Consumer acceptability studies are needed to determine whether these improved color characteristics translate into greater consumer preference and acceptance of products formulated with the optimized whole green banana flour.

In terms of total soluble solids (TSS), the results differed from the expected trend based on banana maturity. Alkarkhi et al. (2011) reported that TSS increases with fruit maturity due to the accumulation of soluble sugars during ripening. Therefore, the OOS banana flours produced from immature bananas would be expected to have lower TSS than commercial banana flours, which are typically manufactured from mature unripe (Stage 1) bananas. The contrasting results observed in the present study may be attributed to differences in the processing methods employed, which may have caused the leaching out of soluble sugar from the commercial flours.

The optimized processing conditions significantly reduced the pH of both OOS Cavendish and OOS Saba banana flours, resulting in values that were also significantly lower than those of the commercial banana flours. This reduction can be directly attributed to the citric acid soaking step, which allows residual organic acids to remain associated with the flour matrix after drying. Similar decreases in pH following acid pretreatment have been reported for banana flour and other dried fruit products, where citric acid effectively acidified the tissues while simultaneously inhibiting enzymatic browning (Bunyameen et al., 2020; Salazar et al., 2022). Aside from improved color stability, the lower pH values may provide an additional hurdle against microbial growth by creating less favorable conditions for spoilage microorganisms (Lund et al., 2020). This effect was further supported by the low water activity values of the treated flours. Although water activity differed significantly among the treated and commercial samples, all values remained below 0.30 and were under the critical threshold of 0.61, wherein the growth of most bacteria, yeasts, and molds is inhibited (Tapia et al., 2007). Furthermore, the optimized flours exhibited satisfactory microbiological quality, with coliform count, and yeast and mold count values within the established microbiological limits for flour products set by the Philippine Food and Drug Administration (FDA, 2022). The low microbial counts may be attributed to the combined effects of drying, reduced water availability, and acidification from the citric acid treatment. However, these findings only indicate potential storage stability based on physicochemical and microbiological quality indicators. Further storage studies are still needed to validate this observation and evaluate quality changes and microbial stability of the samples over time.

The optimized processing conditions did not significantly affect the resistant starch (RS) contents of either OOS Cavendish or OOS Saba flour, indicating that the mild blanching, soaking, and drying conditions generally preserved the native starch structure. Starch is one of the primary carbohydrates found in bananas (Phillips et al., 2021). In its green stage, bananas are naturally rich in resistant starch, which is the starch fraction that resists digestion but is instead fermented in the colon (Munir et al., 2024). This results in the formation of short-chain fatty acids that have been associated with various health benefits, including improved intestinal health, glycemic balance, lipid metabolism, and body weight management (Bojarczuk et al., 2022). Compared with the commercial banana flours, both treated flours also exhibited significantly higher RS values, indicating a greater proportion of starch resistant to digestion. These findings suggest that the optimized whole green banana flours may offer a competitive advantage over commercial banana flours by providing a naturally higher resistant starch content, highlighting their potential for functional food applications.

In terms of apparent amylose content, no significant change was observed in OOS Cavendish flour following the optimized processing treatment, while OOS Saba exhibited a significant increase in apparent amylose content. Amylose is the predominantly linear component located in the amorphous regions of starch granules that plays an important role in determining starch gelatinization and digestibility (de Barros Mesquita et al., 2016). It is positively associated with resistant starch, as amylose readily forms ordered double-helical structures that are less susceptible to enzymatic hydrolysis (Wang et al., 2023). The mild processing conditions employed may have been insufficient to induce significant structural changes in the starch granules of OOS Cavendish, resulting in no significant changes in either apparent amylose or resistant starch contents. In contrast, the increase in apparent amylose observed in OOS Saba may be attributed to differences in starch composition and structure, which likely influenced its response to the optimized processing conditions. The differences in apparent amylose content observed among treated and commercial flours may be attributed to inherent varietal characteristics and fruit maturity, both of which influence starch composition and the amylose-to-amylopectin ratio in bananas (Yang et al., 2022).

4.2.2 Proximate composition

In terms of proximate composition, the optimized processing conditions caused only minimal changes in both OOS banana flours. Among the proximate components evaluated, moisture, crude ash, and crude fat were significantly affected by the optimized process, while crude protein, crude fiber, and nitrogen-free extract (NFE) remained unchanged. Moisture content increased slightly in both treated banana flours compared with their untreated counterparts and commercial banana flours. The higher moisture content of the treated flour samples may be attributed to the additional blanching and soaking steps applied during processing, which may have influenced water uptake and retention prior to drying. This variation may also be associated with differences in the water-holding capacity of the flour matrix, which is affected by the composition and structure of its major components, including starch, protein, and dietary fiber that contain hydrophilic groups capable of binding water (Alam et al., 2023). Nevertheless, the moisture contents of the treated flours remained below the maximum limit of 15.5% recommended for flour stability and storage (Codex Alimentarius Commission, 2023). In contrast, crude ash content significantly decreased following the optimized processing, which may be attributed to the leaching of water-soluble minerals during the blanching and citric acid soaking steps. This observation was also reported in various studies on fruits and vegetables subjected to these pre-treatments (Kumari et al., 2026; Belkacemi, 2022). Despite this reduction relative to the untreated flours, the treated OOS Cavendish and OOS Saba banana flours still exhibited significantly higher crude ash contents than all commercial banana flours. Aside from differences in processing and other varietal factors, this observation may most likely be attributed to the use of whole bananas, including the peel. Banana peel is naturally richer in minerals than the pulp and contains high amounts of potassium, magnesium, calcium, sodium, and phosphorus (Aboul-Enein et al., 2016). Since commercial banana flours are typically produced from the pulp alone, they inherently contain lower mineral levels. Thus, although some minerals were lost during the optimized processing, the inclusion of the peel enabled the treated whole banana flours to retain higher mineral levels compared with commercial pulp-based banana flours.

Crude protein, crude fiber, and nitrogen-free extract remained unaffected by the optimized processing conditions, which demonstrates that the mild blanching, controlled soaking duration, and relatively low drying temperature were insufficient to cause significant degradation or leaching of these major nutritional constituents. Although the optimized processing conditions did not significantly affect the crude protein and crude fiber contents of the OOS banana flours, both treated OOS Cavendish and OOS Saba flours contained significantly higher levels of these nutrients than the commercial banana flours, which may be attributed to the inclusion of the banana peel during flour production. A similar trend was reported by Gómez-Montaño et al. (2019), who reported that flour produced from banana peel contained higher protein levels than flour prepared from the pulp alone. Moreover, banana peels are naturally rich in dietary fiber, primarily comprising pectin, hemicellulose, cellulose, xylose, and lignin (Magdy et al., 2023), which contributes to the higher fiber content of OOS banana flours.

4.2.3 Techno-functional properties

Similar to the physicochemical properties, the optimized processing conditions caused only minor changes in the techno-functional properties of the OOS banana flours. Techno-functional properties are important indicators of flour functionality because they provide insights into hydration capacity, oil retention, swelling ability, solubility, and powder flow characteristics, which may influence their behavior and performance in food systems.

In OOS Cavendish flour, there was a slight decrease in water-holding capacity (WHC), while no significant change was observed in OOS Saba flour following the optimized processing. Despite these differences, both treated flours still exhibited significantly higher WHC than commercial banana flours, indicating their greater ability to retain water. The higher WHC observed in the OOS banana flours may be related to differences in flour composition, such as their significantly higher protein and crude fiber contents compared with the commercial flours. These components may contribute to greater water retention through their interactions with water within the flour matrix (Alam et al., 2023). The higher WHC of the developed flours may influence moisture retention and hydration-related characteristics of food products, including texture and juiciness (Tarahi et al., 2024).

Oil-absorption capacity (OAC) was not significantly affected by the optimized process in either banana variety, suggesting that the treatments did not significantly affect the nonpolar binding sites responsible for oil retention. Among the samples, treated OOS Saba flour exhibited the highest oil-absorption capacity, indicating its greater capacity for oil retention compared with the other samples. Higher OAC may contribute to the retention of hydrophobic compounds and may influence flavor-related and textural attributes in food systems (Alam et al., 2023; Sharew and Mengistu, 2024).

In contrast, the optimized process significantly increased both swelling capacity (SC) and water solubility index (WSI) in both OOS banana varieties, resulting in values that were also significantly higher than those of the commercial flours. The higher SC of the optimized OOS banana flours indicates a greater ability of starch granules to absorb water and expand during hydration (Babaniyi et al., 2025). Meanwhile, WSI reflects the amount of water-soluble components released from the sample per unit weight (Carvalho et al., 2026). The higher SC and WSI values suggest improved hydration and solubility behavior of the optimized flours, which may indicate potential applications in food systems where these properties are desirable.

The bulk and tapped densities of both OOS banana flours were not significantly affected by the optimized processing conditions but were significantly lower than those of the commercial flours. This suggests that the treated OOS flours had a less compact particle structure, possibly due to the inclusion of banana peel and their higher dietary fiber content. Lower bulk and tapped densities indicate that the flour occupies a larger volume per unit mass, potentially influencing powder handling and packaging characteristics (Padhi and Dwivedi, 2022). The differences observed in bulk and tapped densities were also reflected in the calculated Carr’s index and Hausner ratio values, which are commonly used indicators of powder flowability and compressibility. Although the treated flours exhibited higher values than the commercial flours, they still possessed passable flowability based on the classification of Gorle and Chopade (2020), indicating acceptable powder flow characteristics that may contribute to improved handling and processing of banana flours samples.

Overall, the techno-functional characteristics of the developed OOS banana flours suggest potential use in food applications where hydration, solubility, and powder handling properties are important, including bakery products, thickened food formulations, and reconstituted mixes. Further evaluation through specific food formulations is needed to confirm their actual performance and suitability for intended applications.

4.2.4 Pasting properties

Pasting properties describe the viscosity changes that occur when starch is heated and cooled in the presence of water and are important indicators of flour performance in food and other industrial products. While no significant changes were observed in the pasting properties of OOS Cavendish flour following the optimized process, treated OOS Saba flour exhibited significant reductions in trough and final viscosities. Aside from the minimal differences in their optimized processing conditions, this variation may also be associated with varietal differences, which can influence the composition and structural characteristics of the resulting flours. The optimized OOS Cavendish and Saba flours differed in several compositional properties, including amylose, fiber, protein, and lipid contents, which may contribute to differences in pasting behavior through their effects on starch swelling, water availability, and interactions between starch and non-starch components. In addition, factors such as the amylose-to-amylopectin ratio, starch granule integrity, and molecular organization have been reported to influence pasting properties (Devi et al., 2020; Ajiboye et al., 2025; Li et al., 2020). However, these properties were not evaluated in the present study, and their potential contribution to the differences observed between OOS Cavendish and OOS Saba flours requires further investigation.

Although OOS Cavendish and Saba flours slightly differed in starch composition and physicochemical characteristics, the optimized process produced flours with similar viscosity profiles including pasting temperature, peak viscosity, trough viscosity, breakdown viscosity, and setback viscosity, which suggests similar pasting behavior during heating and cooling. These similarities may indicate comparable resistance of swollen starch granules to thermal and mechanical breakdown, and a similar stability against retrogradation (Daramola and Osanyinlusi, 2006). The pasting temperature of the treated OOS flours (77.05 °C–80.30 °C) were found to be significantly lower than the reported values in previous studies of Adewole et al. (2012), Ajiboye et al. (2025), with obtained temperatures of 94.50 °C and 84.42 °C–85.6 °C, respectively. The lower pasting temperatures observed in the treated OOS banana flours can likely be attributed to variations in starch composition, structural characteristics, and processing conditions across the referenced studies.

4.2.5 Thermal properties

Thermal properties determined by differential scanning calorimetry (DSC) provide information on the temperatures and energy required for starch gelatinization (Agama-Acevedo et al., 2019). The thermal parameters include the onset temperature (T_o), peak temperature (T_p), end temperature (T_e), and gelatinization enthalpy (ΔH), which represent the initiation, maximum rate, completion, and energy requirement of the gelatinization process, respectively.

The optimized processing conditions resulted in significant changes in the gelatinization temperature parameters (T_o, T_p, and T_e) of both OOS banana flours, except for the end temperature (T_e) of OOS Saba flour, which remained unaffected. For treated OOS flours, the observed T_o (73.90 °C–76.26 °C), T_p (78.42 °C–81.66 °C), and T_e (85.49 °C–91.89 °C) values were comparable to those reported by Sánchez-Mesa et al. (2025) and Huang et al. (2019), indicating that the thermal behavior of the optimized banana flours remained within the typical range reported for green banana starch and flour.

Between the treated OOS flours, OOS Saba exhibited higher gelatinization temperature values than OOS Cavendish, which may be attributed to inherent varietal differences and the compositional variations previously discussed in the pasting properties section. The higher gelatinization temperatures observed in OOS Saba indicate that its starch structure required greater thermal energy to initiate and complete gelatinization, suggesting greater resistance to thermal disruption compared with OOS Cavendish. These differences may influence the processing behavior of the flour by affecting the temperature required for starch gelatinization during food applications.

The gelatinization enthalpy (ΔH), which represents the energy required to melt amylopectin crystallites and reflects the degree of molecular order within starch granules (Wang et al., 2020), exhibited a significant increase in OOS Saba flour following the optimized processing treatment. This suggests that more energy is required to disrupt its crystalline structure. Despite these differences, both treated flours exhibited comparable ΔH values, indicating that the optimized processing had minimal influence on the overall energy required for starch gelatinization.

5 Conclusion

This study demonstrated the potential of producing flour from both the pulp and peel of immature out-of-specification (OOS) Cavendish and Saba bananas, highlighting a sustainable approach to value addition within the banana industry. The optimized processing conditions that produced flours with improved color consisted of blanching for 5 min at 60 °C, soaking for 10–11 min in a citric acid–sodium metabisulfite solution, and drying at 59 °C–60 °C for at least 15 h. Beyond improving flour color, the optimized processing generally preserved key nutritional and functional properties of the flours, including their high resistant starch and dietary fiber contents. Comparison with commercial banana flours provided insights into the quality characteristics of the developed flours and highlighted their potential as value-added banana-based ingredients. These findings support the utilization of OOS whole banana flours as sustainable ingredients for potential food applications while contributing to the reduction of underutilized banana resources. To further establish the effects of the optimized process on flour properties and better understand the contribution of peel inclusion, future studies should include the production and analysis of pulp-only flour from OOS bananas alongside the optimized whole banana. Future research may also evaluate flour shelf-life stability and investigate additional physicochemical, nutritional, and functional characteristics to better assess their potential applications. Moreover, incorporating these flours into various food formulations is recommended to evaluate their processing performance, sensory acceptability, and potential for commercial-scale application.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of the manuscript, the author(s) used OpenAI ChatGPT (GPT-5.5) to check the grammar and improve clarity of the text. The author(s) reviewed and edited the content as necessary and take(s) full responsibility for the final content of the manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frfst.2026.1941070/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Soubhagya Tripathy, Indian Institute of Technology Kharagpur, India

Reviewed by: Sreelakshmi Sankara Narayanan, Taylor’s University, Malaysia

Lubna Shaik, D Y Patil International University, India