Sustainable Preservation of Opuntia ficus-indica Peel Waste for Resource Recovery Through Pretreatment and Convective-Drying Processes

Abstract

Opuntia ficus-indica peel waste (OPW) accounts for approximately 50% of the fresh fruit weight, a relatively higher rate in comparison to other fruits, making it a particularly abundant and underutilized agricultural by-product. This study investigates Tunisian OPW through convective drying (50, 60, 75, and 85 °C) using four different pretreatments: no pretreatment (Fresh OPW), dipped peels in acetic acid solution (AA OPW), frozen and thawed peels (FZ OPW), and dipped–frozen combined pretreatment (AA-FZ OPW). The drying kinetic evaluation clearly demonstrated an increase in drying rate by lowering the required drying time for the various pretreatments and drying temperatures. This drying time decreased by 8.64–17.39%, 23.46–36.70%, and 28.39–39.45% for AA OPW, FZ OPW, and AA-FZ OPW, respectively, as compared to fresh OPW. The total phenolic and betalain contents, found in dried OPW, ranged from 7.87 ± 0.254 g GAE/100 g dw to 12.29 ± 0.421 g GAE/100 g dw and from 22.74 ± 2.34 mg/100 g dw to 59.86 ± 3.24 mg/100 g dw, depending on the drying temperature and pretreatment applied. This study employs thin-layer and artificial neural network (ANN) modelling, finding that both methods produce highly accurate moisture-ratio prediction during the drying process. The peels from Opuntia ficus-indica have considerable potential to be reused as functional ingredients.

1. Introduction

The prickly pear (Opuntia ficus-indica) is a widely consumed fruit around the world. This generates an abundant resource in by-products, notably the peels, representing about 50% of the fruit’s fresh weight [1]. These peels are particularly interesting because they are extremely rich in bioactive compounds: pigments, minerals, antioxidants, polyphenols, and vitamins [2]. Therefore, their valorization offers an appealing method to recover beneficial substances suitable for the food, pharmaceutical, and cosmetic industries, as well as to reduce agri-food waste [2]. Nevertheless, these by-products are highly perishable due to the peel’s high moisture (from 75% to 85%) [3], organic content, and pH range from (4.4 to 5.6), leading to its rapid spoilage [1,4]. Their improper disposal requires further processing steps to guarantee their safe and sustainable use.

Drying is one of the most common methods for stabilizing this biomass since it reduces moisture content and, thus, the risk of microbial and enzymatic degradation.

Furthermore, the drying processes have to be carefully optimized, especially when applied for extended periods at high drying temperatures to preserve the most sensitive bioactive compounds as well as flavour and colour attributes. Therefore, several pretreatments prior to the drying process have been used to mitigate these drawbacks. Among them, soaking, dipping and spraying in ethanol and organic acids (such as citric acid and acetic acid) solutions have been largely applied as a drying pretreatment for fresh matrices of pumpkin [5], beetroot, peach peels [6], mango peels [7], Opuntia ficus-indica and Opuntia stricta fruit [8], carrots [9], and Opuntia ficus-indica peels [10]. These investigations concluded that drying rates are increased when compared to control samples (without pretreatment). This behaviour was linked to the Marangoni effect, which causes liquid to flow from matrices due to the naturally occurring surface tension gradient between two fluids, such as ethanol and water.

Additionally, freezing is one of the innovative pretreatments for fruit and vegetable drying. Since no chemicals or catalysts are required, it is both economically and environmentally friendly [11]. It induces the formation of ice crystals from free water within the cells, resulting in the breakdown of the cell wall and accelerating the drying process [12,13]. Also, the freezing rate has to be regulated to form intracellular ice crystals with the least amount of damage to the cellular structure in order to preserve the high-quality dried product [14].

Moreover, drying is recognized as one of the most energy-intensive processes in the food industry, accounting for up to 70% of the energy consumption related to fruit and vegetable processing [15]. This reality highlights the need to optimize this process in order to reduce production costs and limit its environmental impact. To tackle these challenges, mathematical modelling is commonly used to evaluate the effect of process parameters, optimize the drying process, and integrate energy sources. Thin-layer models, as well as empirical semi-theoretical ones, are widely used in drying processes for their simplicity and accuracy. Such models have been used to optimize dryers, drying settings, and improve process control [16,17]. Furthermore, an interesting alternative modelling is provided by a physical-based model. As it is based on Fick’s second law, this theoretical model accurately reflects the internal moisture diffusion during drying. To support its use, this model requires, unfortunately, several assumptions, including isothermal drying, low shrinkage, negligible external resistance, homogenous diffusivity, and uniform initial moisture content. In order to address the issues encountered with these latter models, Black-box modelling, such as artificial neural networks (ANN), has emerged in recent years as a novel method for nonlinear computational modelling in food material. This model successfully captures the biological properties of food and has the ability to accurately simulate intricate nonlinear processes in food engineering [18].

In drying processes, modelling has been particularly reported for its high performance, predictive capability and robustness in several studies [19]. Thus, ANN is examined as a potential alternative to predict drying kinetics as well as other physical and chemical properties in dried materials. The ANN framework includes three sections: the input layer, the hidden layer, and the output layer. In the drying process, the input layer incorporates parameters such as drying temperature, drying time, and possible pretreatments of the raw material. The output layer represents what is expected to be predicted, like moisture ratio, nutritional component deterioration, shrinkage level, and so on. The hidden layer will be employed for the intermediate computation from the input to the output layer [20].

Despite all of the drying research conducted on Opuntia ficus-indica by-products, to the knowledge of the authors, no study has been carried out to study the effect of combined organic acid and freezing pretreatment on the quality attributes of dried peels, as well as the modelling of the drying process using thin-layer and ANN during drying. Therefore, this study focuses on the valorisation of Opuntia ficus-indica peels using a drying process at four different temperatures (50 °C, 60 °C, 75 °C, and 85 °C) combined with acetic acid and freezing pretreatment in order to optimize the drying conditions while reducing energy consumption and preserving the quality of the dried peels as a natural source of high-value-added bioactive compounds, particularly betalains and total polyphenols.

2. Materials and Methods

2.1. Material

Opuntia Ficus-indica fruits used in the drying experiments were cultivated in the northern Tunisian province of Bou Argoube (Nabeul). The collecting period was from September to October 2023. The harvesting was carried out on physiologically mature fruits, either yellow or red in colour, showing uniform coloration, adequate firmness, no mechanical damage and homogeneous size. The peels were physically removed from the pulp using a knife and sliced into uniform 4 mm thick sections. They were subsequently divided into two portions; the first one was the fresh peels (Fresh OPW), which were instantly dried. The second portion was considered for pretreatment and drying experiments. The different dried OPW samples were ground to obtain OPW powder and stored in vacuum bags until later analysis.

2.2. Pretreatment Investigation

Two batches, either fresh or frozen peels (after slow thawing for approx. 2.5 h [21]), were cut into rectangular pieces measuring 5 cm in length, 4.5 cm in width, and 4 mm in thickness. A proportion of 1 piece of yellow peel versus 2 pieces of the red one was chosen to simulate the natural distribution of coloured fruits. Peel samples, approximately 100 g, were dipped into 200 mL acetic acid solution 1.5% (v/v) for 5 min. The excess liquid from the peels was drained after they were dipped in the acetic acid solution to prevent misleading measurements of moisture content during the drying process and to correctly compare treated and untreated samples (Figure 1).

As described, four different pretreatments were considered in this study: no pretreatment (Fresh OPW), dipped peels in acetic acid solution (AA OPW), frozen and thawed peels (FZ OPW), and dipped–frozen combined pretreatment (AA-FZ OPW).

2.3. Drying Experiments

The drying experiments were carried out in a Food Dehydrator (LT-91, Zhejiang Lian Teng Intelligent Equipment Co., Ltd., Zhejiang, China); an electric indirect forced-convection dryer, equipped with a circulation fan and a drying chamber. The configuration of its ventilation system ensures a uniform air circulation with a velocity of around 2 m·s⁻¹. Samples (100 g) of the fresh (fresh OPW) or with the applied pretreatments (AA OPW, FZ OPW and AA-FZ OPW) were evenly spread in a thin layer on a holder tray and placed inside the convective dryer. Four different drying temperatures (55, 65, 75, and 85 °C) were applied for peels drying while maintaining constant air velocity and relative humidity. Every 20 min, the sample mass was measured to monitor the weight reduction. When the equilibrium moisture content was reached, which was defined as a 1.5% reduction in the total weight, the drying operation was stopped [22]. Every drying experiment was performed twice on different days.

To investigate the temperature effect on peel quality, drying kinetics were established by expressing the moisture content (X_t: g water (g dry basis)⁻¹) and the dimensionless moisture ratio (MR_t: (−)) during the drying process. They were calculated according to Equations (1) and (2), respectively.

$${\mathrm{X}}_{\mathrm{t}}={\displaystyle \frac{\left({\mathrm{M}}_{\mathrm{t}}-{\mathrm{M}}_{\mathrm{d}}\right)}{{\mathrm{M}}_{\mathrm{d}}}}$$

(1)

where ${\mathrm{M}}_{\mathrm{t}}$ is the weight (g) at time t (min), and M_d is the weight of the dry material.

$${\mathrm{M}\mathrm{R}}_{\mathrm{t}}={\displaystyle \frac{{\mathrm{X}}_{\mathrm{t}}-{\mathrm{X}}_{\mathrm{e}}}{{\mathrm{X}}_0-{\mathrm{X}}_{\mathrm{e}}}}={\displaystyle \frac{{\mathrm{X}}_{\mathrm{t}}}{{\mathrm{X}}_0}}$$

(2)

where X_t, X₀, and X_e are the moisture content at any given drying time, the initial moisture content, and the equilibrium moisture content [kg H₂O/kg d.b.], respectively. Under the experimental conditions, X_e was relatively small compared to X₀; therefore, its value was numerically set to zero [21].

2.4. Mathematical Modelling of Thin-Layer Drying Kinetics

According to

Table 1. Empirical drying models used.

Model Name	Model Equation	References
Newton	$\mathrm{M}\mathrm{R}=\mathrm{e}\mathrm{x}\mathrm{p}(-\mathrm{k}\mathrm{t})$	[25]
Page	$\mathrm{M}\mathrm{R}=\mathrm{e}\mathrm{x}\mathrm{p}(-\mathrm{k}{\mathrm{t}}^{\mathrm{n}})$	[26]
Verma	MR = a exp (−kt) + (1 − a)exp(−gt)	[27]
Aghbaslo	MR = exp (−k1t/1 + k2t)	[28]
Alibas	MR = a exp ((−ktn) + (bt))+ g	[29]
Balbay and Sahin	MR = (1 − a) exp (−ktn) + b	[30]

Abbreviations: MR—moisture ratio (dimensionless); k, a, g, k1, k2, b—drying constants; t—time (min).

, different thin-layer drying models were selected and fitted to the experimental data following the non-linear regression analysis. The determination coefficient (R²), reduced chi-square (χ²) and root mean square error (RMSE) were used to assess the fitness quality. They were determined using Equations (7), (8) and (9), respectively. The model with a high R² value, low χ², and low RMSE values was considered best.

$${\mathrm{R}}^2=1-{\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\left({\mathrm{M}\mathrm{R}}_{\mathrm{e}\mathrm{x}\mathrm{p},\mathrm{i}}-{\mathrm{M}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e},\mathrm{i}}\right)}^2}{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\left({\mathrm{M}\mathrm{R}}_{\mathrm{e}\mathrm{x}\mathrm{p}}-{\mathrm{M}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e},\mathrm{i}}\right)}^2}}$$

(3)

$${\mathsf{\chi }}^2={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\left({\mathrm{M}\mathrm{R}}_{\mathrm{e}\mathrm{x}\mathrm{p},\mathrm{i}}-{\mathrm{M}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e},\mathrm{i}}\right)}^2}{\mathrm{N}-\mathrm{n}}}$$

(4)

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}=\sqrt{{\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}\left({\mathrm{M}\mathrm{R}}_{\mathrm{e}\mathrm{x}\mathrm{p},\mathrm{i}}-{\mathrm{M}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e},\mathrm{i}}\right)2}{\mathrm{N}}}}$$

(5)

where MR_exp,i and MR_pre,i represent the experimental and predicted moisture ratio, respectively, MR_exp is the mean experimentally measured value of MR, N is the number of observations, and n is the numb er of constants [23,24].

2.5. Artificial Neural Network

The configuration of the ANN was set up using drying time, temperature, and with and without the different applied pretreatments as inputs and moisture ratio as the output. The data was pre-processed. Numerical features were then standardized using a StandardScaler. Categorical variables were encoded as numerical values. The model was implemented and trained with Python (3.12.11), tensorflow (2.20.0), keras (3.11.3), and scikit-learn (1.7.2) [31]. The Adam function in TensorFlow is an optimizer used for training deep learning models. The Relu function was selected as the transfer function for the hidden layers. Different configurations were tested by varying the number of layers as well as the number of neurons per layer for creating artificial neural networks. For ANN training and validation, the experimental data were divided into 60% training data, 20% validation data, and 20% testing data. To avoid overfitting, training was terminated when one of the termination conditions was satisfied by reaching the maximum number of epochs (1000) or validation loss stopped improving for a certain number of epochs (patience = 30). To optimize the learning rate management and accelerate the learning process, the learning rate was reduced by a factor of 0.5 if validation loss stopped improving for 8 epochs, with a minimum learning rate of 10⁻⁶ as a limit. The performance of the ANN model was assessed by comparing the estimated data from the model with the experimental data using the RMSE and R² values.

2.6. Methods of Analysis

2.6.1. Moisture Content and Water Activity

The moisture content (MC) of both fresh and dried Opuntia ficus-indica peels with and without different applied pretreatments was determined using an oven at 105 °C for 24 h [32]. Water activity (a_w) was measured using an a_w-meter (model CH-8853, LabSwift-a_w, Novasina AG, Lachen, Switzerland).

2.6.2. Colour Parameters

The colour parameters of the OPW powder with (AA OPW, FZ OPW, and AA-FZ OPW) and without the different applied pretreatments (fresh OPW) were measured using a colorimeter (Minolta, model CR-400, Osaka, Japan). This device calculates the L, a, and b* values, which represent lightness, red–green hue, and yellow–blue hue, respectively. Measurements were taken eight times for each sample.

2.6.3. Betalain Content

The OPW powder obtained after the drying process with and without different applied pretreatments (fresh OPW, AA OPW, FZ OPW, and AA-FZ OPW) at four drying temperatures (55 °C, 65 °C, 75 °C, and 85 °C) was used for betalain content. The peel powder (100 mg) was shaken for 10 min in 20 mL of water, which was used as the extraction solvent. After shaking, the samples were centrifuged at 12,000× g at 15 °C for 15 min using a Hermle Z323K centrifuge (Wehingen, Germany). The supernatants were filtered through a 0.45 µm nylon membrane filter (Millipore Corp., Bedford, MA, USA), and the resulting extracts were analyzed by spectrophotometry. For the photometric quantification of betalains, all measurements were carried out using a UV/visible spectrophotometer (Shimadzu, Sydney, Australia). Pigments were extracted using water. Absorbance was measured at the maximum absorption wavelengths of 535 nm for betacyanins and 483 nm for betaxanthins, respectively.

2.6.4. Total Polyphenol Content

A mass of 2 g of OPW powder was macerated in 20 mL of 80% (v/v) methanol/water solvent at room temperature. The mixture was placed in an airtight container under agitation for 24 h. Next, the mixture was centrifuged at 3000 rpm for 10 min. Finally, the supernatant was collected in a test tube, sealed hermetically, and stored in the dark at −20 °C until further use.

The total polyphenol contents were determined using the Folin–Ciocalteu method [33], but slightly modified [34]. The Folin-Ciocalteau reagent, yellow in colour, consists of a mixture of phosphotungstic acid and phosphomolybdic acid, which are reduced during the oxidation of phenols into a mixture of blue tungsten and molybdenum oxides. The blue colouration produced is proportional to the levels of total phenolic compounds and has a maximum absorption around 765 nm. The polyphenol assay protocol involves mixing 100 μL of appropriately diluted extract with 500 μL of Folin–Ciocalteu reagent (10 times diluted). After 5 min, 400 μL of a Na₂CO₃ (7.5%) solution is added. The absorbance is measured at 765 nm, after an incubation of 90 min in the dark. The polyphenol contents of the different extracts are expressed in mg of gallic acid equivalent per 100 g of dry matter (mg GAE/100 g DM).

2.7. Statistical Analysis

All statistical analyses were performed using SPSS^® Statistics version 26.0 software for Windows (IBM, Armonk, NY, USA). Results were represented as mean ± standard deviation, based on a minimum of three independent assays performed. Results obtained for the variables studied (physical and chemical parameters, betalain content, and TPC) were compared, at a significance level of 0.05, by one-way analysis of variance (ANOVA) and Tukey’s multiple comparison test for normally distributed data and homogenized variances, by Welch’s and the Games–Howell multiple comparison tests for normally distributed data and non-homogenized variances, and with the Kruskal–Wallis test and stepwise multiple comparison test for non-normally distributed data. The analyses of non-linear regression were performed by OriginPro 8.5.0 SR1 (2010) (OriginLab Corporation, Northampton, MA, USA; trial version) software for drying kinetics.

3. Results and Discussions

3.1. Effect of Drying Temperature and Pretreatment on Drying Characteristics

The effect of drying temperatures and pretreatments on the drying behaviour of Opuntia ficus-indica peel waste (OPW) was studied. Four different pretreatments were considered in this study: no pretreatment (Fresh OPW), dipped peels in acetic acid solution (AA OPW), frozen and thawed peels (FZ OPW), and dipped–frozen combined pretreatment (AA-FZ OPW). The drying kinetics were presented in Figure 1 showing the variation in reduced moisture content (MR) versus drying time for different drying temperatures (55, 65, 75, and 85 °C) for OPW with and without the different applied pretreatments (AA OPW, FZ OPW, AA-FZ OPW, and fresh OPW), respectively. The initial moisture content of the fresh peel waste used in this study was 88.298% ± 0.868% (wet basis). The drying process was conducted until the sample mass variation dropped to 1.5%, indicating a stable phase in which no more water could be removed from the dried OPW. As described, four different pretreatments were considered in this study: no pretreatment (Fresh OPW), dipped peels in acetic acid solution (AA OPW), frozen and thawed peels (FZ OPW), and dipped–frozen combined pretreatment (AA-FZ OPW).

The characteristics of the dried OPW (water activity (a_w), moisture content (MC), and drying time (DT)) under different operating conditions were summarized in

Table 2. Effect of drying temperature and pretreatment on drying characteristics of Opuntia ficus-indica peel waste (OPW).

Drying Temperatures Quality OPW		55 °C	65 °C	75 °C	85 °C	F-Value	p-Value
Fresh OPW	aw(-)	0.503 ± 0.003 Aa	0.490 ± 0.004 Ab	0.484 ± 0.004 Ab	0.372 ± 0.008 Ac	20.173	0.000
	MC (%)	18.05± 0.008 Aa	17.50 ± 0.004 Aa	14.97 ± 0.001 Ab	12.68 ± 0.001 Ac	347.800	0.000
	DT (min)	545 ± 10 Aa	405 ± 10 Ab	345 ± 10 Ac	290 ± 11.547 Ad	14.414	0.02
AA OPW	aw(-)	0.493 ± 0.005 Ba	0.480 ± 0.006 Bb	0.471 ± 0.001 Bc	0.354 ± 0.005 Bd	21.742	0.000
	MC (%)	15.46 ± 0.005 Ba	14.74 ± 0.003 Ca	14.68 ± 0.001 Aa	12.12 ± 0.00 Bb	11.801	0.008
	DT (min)	465 ± 10 Ba	370 ± 11.547 Bb	285 ± 10 Bc	240 ± 00 Bd	14.545	0.002
FZ OPW	aw(-)	0.449 ± 0.008 Ca	0.423 ± 0.023 Ca	0.376 ± 0.006 Cb	0.32 ± 0.006 Cc	20.265	0.000
	MC (%)	16.61 ± 0.005 Ba	15.51 ± 0.005 Bab	14.41± 0.006 Ab	11.98 ± 0.0004 Bc	151.954	0.000
	DT (min)	345 ± 10 Ca	310 ± 11.547 Cb	240 ± 00 Cc	220 ± 00 Cd	14.679	0.002
AA-FZ OPW	aw(-)	0.411 ± 0.010 Da	0.382 ± 0.007 Db	0.359 ± 0.007 Dc	0.30 ± 0.006 Dd	21.713	0.000
	MC (%)	15.51 ± 0.005 Ba	14.60 ± 0.004 Ca	13.40 ± 0.003 Bb	11.37 ± 0.002 Cc	141.075	0.000
	DT (min)	330 ± 11.547 Ca	290 ± 11.547 Cb	220 ± 00 Dc	200 ± 00 Dd	14.634	0.002
F-value		21.666(aw) 17.103(MC) 13.848(DT)	21.519(aw) 11.912(MC) 90.333(DT)	21.694(aw) 10.853(MC) 14.724(DT)	138.800(aw) 74.011(MC) 14.815(DT)
p-value		0.000(aw) 0.000(MC) 0.003(DT)	0.000(aw) 0.008(MC) 0.000(DT)	0.000(aw) 0.013(MC) 0.002(DT)	0.000(aw) 0.000(MC) 0.002(DT)

Values are the mean of 3 independent determinations ± standard deviation. Uppercase letters indicate statistically significant differences (0.05) between Fresh OPW, AA OPW, FR OPW, and AA-FR OPW. Lowercase letters indicate statistically significant differences (0.05) between different drying temperatures Abbreviations: a_w—water activity; MC—moisture content; DT—drying time; Fresh OPW—no pretreatment; AA OPW—dipped peels in acetic acid solution; FZ OPW—frozen and thawed peels; AA-FZ OPW—dipped–frozen combined pretreatment.

. The final a_w and MC of dried OPW ranged from 0.3 ± 0.006 to 0.503 ± 0.003 and from 18.05% ± 0.008 to 11.98% ± 0.0004, respectively, depending on the drying temperature and pretreatment applied. This final MC variation in dried Opuntia ficus-indica peels has already been confirmed in previous studies by El-Said et al. (2011) [35] and Ettalibi et al. (2020) [36].

Figure 2 clearly shows that all drying kinetic curves follow a decreasing exponential trend, where water evaporation slows down as the peels become drier [37]. This behaviour is commonly observed in drying processes of plant material [38], where the drying entirely occurs in the second phase, and the drying process is governed by a diffusional water mechanism [39].

In addition, for the various drying temperatures, the drying kinetic curves (Figure 2) display an increase in drying rate by decreasing the necessary DT for the different pretreatments applied. This increase varied from 8.64 to 17.39%, 23.46 to 36.70%, and 28.39 to 39.45% for AA OPW, FZ OPW, and AA-FZ OPW, respectively, in comparison to fresh OPW. The freezing pretreatment used alone or combined with acetic acid appeared to be the most relevant in accelerating the drying process. In fact, freezing causes the formation of ice crystals inside the cells of OPW. These crystals can perforate the cell walls, thereby increasing the porosity of the material once thawed. This increase in porosity facilitates the release of water during the drying process, which explains the accelerated drying when frozen [39]. This result was also observed by Reyes et al. (2008) [40] during the drying of carrot pieces with freezing pretreatments. At higher temperatures, the thermal energy provided is sufficient to break the water bonds more quickly. This is true for both fresh and frozen peels, but the effect is amplified in frozen peels due to their already altered cellular structure.

Furthermore, when compared to FZ OPW, AA OPW exhibited a minor effect on drying kinetics (around 17.5% at 75 °C). This phenomenon is induced by a surface tension gradient between the two fluids: the acetic acid solution at the OPW surface and water contained within the OPW matrix. Based on the Marangoni effect, the liquid is induced to move from a low to a high surface tension region, hence accelerating the drying process [10]. In addition, acid dipping might potentially modify the pectin structure, which improves drying kinetics [41]. This result was similarly demonstrated by Hiranvarachat et al. (2011) [9], who studied the effect of soaking carrots in a citric acid solution before drying. It has been shown that this pretreatment allowed for a significant reduction in drying time compared to the untreated carrots. The acid treatment increased the permeability of the carrot cells, thus facilitating the removal of water.

This acceleration of the drying process was particularly marked during the first hours of the experiment. An initial reduction in MR was more significant for pretreated peels, indicating an accelerated loss of water from the beginning of the drying process. This phenomenon could be explained by the action of the acetic acid present on the cellular structure of the peels, which tends to damage the integrity of cell membranes by inducing structural changes in pectocellulosic cell wall structure (which plays a central role in cell cohesion and water diffusion), thereby facilitating the release of intracellular water.

The effect of pretreatment with acetic acid persists throughout the drying process, although it is less pronounced in the later phases. This trend suggests that acetic acid durably modifies the water transport properties in the peel tissues, allowing for more efficient drying even after the initial rapid evaporation.

The comparison between fresh and frozen peels shows that pretreatment with acetic acid is effective in both cases. This observation indicates that this pretreatment could be successfully applied to different states of the raw material, offering flexibility in the transformation process.

3.2. Thin-Layer Drying Kinetics and ANN Modelling of Opuntia ficus-indica Dried Peel Waste

Two different methods were used in this study for predicting the moisture ratio of the OPW samples during the drying process: thin-layer mathematical modelling and artificial neural network (ANN) as black-box modelling.

Concerning the first method, nonlinear regression analysis was applied to fit the experimental MR of OPW data using six thin-layer models (Newton, Page, Verma, Aghbaslo, Alibas, Balbay, and Sahin). The best model was evaluated using statistical parameters (R², χ², and RMSE) after fitting the models to the experimental data.

Table 3. Comparative thin-layer and ANN models, their model constants, and statistical parameters of OPW samples.

Peel Material	Temperatures	Constants	R2	χ2	RMSE
Newton
Fresh OPW	55	k = 0.00598	0.99377	0.0005234	0.02288
	65	k = 0.00928	0.99413	0.00052624	0.01498
	75	k = 0.01038	0.99297	0.000664	0.02576
	85	k = 0.01132	0.98275	0.00184	0.04288
AA OPW	55	k = 0.00659	0.99867	0.00010345	0.01341
	65	k = 0.00928	0.99413	0.000526	0.03304
	75	k = 0.01344	0.99592	0.000378	0.02213
	85	k = 0.01567	0.99254	0.00078	0.02795
FZ OPW	55	k = 0.0114	0.99291	0.000663	0.02574
	65	k = 0.01269	0.99649	0.000323	0.01798
	75	k = 0.01488	0.9902	0.00104	0.0323
	85	k = 0.01649	0.98524	0.00168	0.04093
AA-FZ OPW	55	k = 0.014	0.99719	0.000249	0.01577
	65	k = 0.01486	0.99629	0.000343	0.01851
	75	k = 0.01912	0.9921	0.000845	0.02907
	85	k = 0.01944	0.98906	0.00125	0.03539
Page
Fresh OPW	55	k = 0.00357; n = 1.09695	0.99647	0.0002961	0.01721
	65	k = 0.00694; n = 1.0496	0.99799	0.00017001	0.01304
	75	k = 0.00497; n = 1.15438	0.99834	0.000157	0.01252
	85	k = 0.00309; n = 1.27823	0.9976	0.000255	0.01598
AA OPW	55	k = 0.00561; n = 1.03063	0.99893	0.000083423	0.01369
	65	k = 0.00549; n = 1.10739	0.99696	0.000272	0.01773
	75	k = 0.00878; n = 1.09395	0.99772	0.000211	0.01425
	85	k = 0.0069; n = 1.18772	0.99901	0.000103	0.01017
FZ OPW	55	k = 0.00515; n = 1.17079	0.99829	0.00016	0.01264
	65	k = 0.00752; n = 1.11486	0.99904	0.0000884	0.0094
	75	k = 0.00554; n = 1.22449	0.99912	0.000094	0.0097
	85	k =0.00452; n = 1.3008	0.99917	0.0000947	0.00973
AA-FZ OPW	55	k = 0.00944; n = 1.08827	0.99853	0.000131	0.01143
	65	k = 0.00858; n = 1.12488	0.99891	0.000101	0.01004
	75	k = 0.00764; n = 1.22074	0.9991	0.0000967	0.00983
	85	k = 0.00652; n = 1.26344	0.99937	0.0000718	0.00847
Verma
Fresh OPW	55	A = 1.02852; k = 0.00614; g = 47.20492	0.9939	0.0005117	0.02262
	65	A= 1.00983; k = 0.00895; g = 34.65301	0.99712	0.000244	0.01562
	75	A = 1.07445; k = 0.01112; g = 35.43818	0.995101	0.0004711	0.02171
	85	A = 1.13784; k = 0.01277; g = 37.27862	0.9895	0.00112	0.03345
AA OPW	55	A = 0.97335; k = 0.00721; g = 35.9345	0.99804	0.00015447	0.01243
	65	A = 1.08104; k = 0.0096; g = 40.81991	0.99124	0.00083914	0.02897
	75	A = 1.06304; k = 0.01496; g = 25.54005	0.99569	0.00041409	0.02035
	85	A = 1.11815; k = 0.01738; g = 26.17809	0.99612	0.00040636	0.02016
FZ OPW	55	A = 1.10596; k = 0.01256; g = 37.18769	0.997	0.000280533	0.01675
	65	A = 1.08762; k = 0.01376; g = 34.06461	0.99914	0.000079350	0.00891
	75	A = 1.14358; k = 0.01686; g = 29.90115	0.99607	0.0004185	0.02046
	85	A = 1.19741; k = 0.01941; g = 28.8663	0.99453	0.00062138	0.02493
AA-FZ OPW	55	A = 1.06989; k = 0.01494; g = 29.23156	0.99858	0.000126039	0.01123
	65	A = 1.10059; k = 0.01628; g = 29.33242	0.9991	0.000082901	0.00911
	75	A = 1.17086; k = 0.02206; g = 23.06273	0.99777	0.000238637	0.01545
	85	A = 1.19742; k = 0.02284; g = 23.21982	0.99646	0.00040568	0.02013
Aghbaslo
Fresh OPW	55	k1 = 0.00517; k2 = −5.714 × 10−4	0.9983	0.000142	0.01193
	65	k1 = 0.00817; k2 = −4.8724 × 10−4	0.99873	0.0001072	0.01036
	75	k1 = 0.00867; k2 = −0.00122	0.99907	0.00088	0.00936
	85	k1 = 0.00828; k2 = −0.00221	0.9994	0.0000602	0.00776
AA OPW	55	k1 = 0.00721; k2 = −1.3821 × 10−4	0.99779	0.0001738	0.01319
	65	k1 = 0.00708; k2 = −0.00132	0.99879	0.000116	0.01077
	75	k1 = 0.01217; k2 = −0.00137	0.99857	0.000137	0.01171
	85	k1 = 0.01275; k2= −0.00207	0.99934	0.00026	0.00834
FZ OPW	55	k1 = 0.00984; k2 = −0.00113	0.99657	0.00032086	0.01791
	65	k1 = 0.01159; k2 = −7.78475 × 10−4	0.99792	0.000190998	0.01382
	75	k1 = 001183; k2 = −0.0022	0.99849	0.000160659	0.01268
	85	k1 = 0.01207; k2 = −0.00321	0.99874	0.0001433	0.01197
AA-FZ OPW	55	k1 = 0.01317; k2 = −5.88254 × 10−4	0.99772	0.000202326	0.01422
	65	k1 = 0.01353; k2 = −9.41222 × 10−4	0.99762	0.000219654	0.01482
	75	k1 = 0.01541; k2 = −0.00267	0.99793	0.0002216	0.01489
	85	k1 = 0.01474; k2 = −0.00336	0.99923	0.00008785	0.00937
Alibas
Fresh OPW	55	A = 0.67752; k = 1.03725; b = 1.08228 g = 0.1393; n = 1.01046	0.83203	0.0141	0.11875
	65	A = 0.91945; k = 1.0489; b = 1.0522 g = 0.03398; n = 1.00249	0.99027	0.000825	0.02872
	75	A = 1.07997; k = 1.03027; b = 0.99764 g = −0.03033; n = 0.9962	0.95576	0.00463	0.01991
	85	A = 1.13884; k = 1.01224; b = 0.99139 g = −0.08807; n = 0.9977	0.95755	0.04526	0.06727
AA OPW	55	A = 0.95741; k = 1.06964; b = 1.05008 g = 0.06252; n = 0.99805	0.9454	0.0043	0.06554
	65	A = 0.84052; k = 0.97314; b = 0.99301 g = 0.06279; n = 1.00633	0.96661	0.0032	0.05657
	75	A = 1.05795; k = 1.01894; b = 0.9991 g = −0.03849; n = 0.99852	0.98938	0.00102	0.03195
	85	A = 1.07997; k = 1.03027; b = 0.99764 g = −0.03033; n = 0.9962	0.95576	0.00463	0.06807
FZ OPW	55	A = 1.10323; k = 1.02389; b = 0.99553 g = −0.03666; n = 0.99644	0.94671	0.00498	0.07059
	65	A = 1.02793; k = 1.01492; b = 1.00087 g = −0.00522; n = 0.99971	0.99551	0.000413086	0.02032
	75	A = 0.98956; k = 1.04479; b = 0.98686 g = 0.8432; n = 0.99178	0.79321	0.02201	0.14835
	85	A = 1.09311; k = 1.01008; b = 0.97552 g = −0.03164; n = 0.99585	0.93592	0.00727	0.08529
AA-FZ OPW		A = 1.04033; k = 1.02428; b = 1.00221 g = −0.00882; n = 0.99825	0.98918	0.000958764	0.03096
	65	A = 1.04663; k = 1.02561; b = 1.00342 g = −0.01164; n = 0.99838	0.98883	0.00103	0.03212
	75	A = 1.06138; k = 1.01326; b = 0.98919 g = −0.02666; n = 0.99798	0.98121	0.00201	0.04483
	85	A = 1.10898; k = 1.02641; b = 0.98633 g = −0.05502; n = 0.99497	0.94467	0.00634	0.07961
Balbay and Sahin
Fresh OPW	55	A = −0.02701; k = 0.00408; n = 1.04862 b = −0.04427	0.99794	0.0001727	0.01315
	65	A = −0.01081; k = 0.0074; n = 1.02262 b = −0.02284	0.99846	0.00013022	0.01141
	75	A = 0.00248; k = 0.0068; n = 1.18729 b = −0.00407	0.99889	0.000117	0.01229
	85	A = −0.00958; k = 0.00326; n = 1.25103 b = −0.02297	0.99809	0.00224	0.01428
AA OPW	55	A = −0.04615; k = 0.00968; n = 0.92108 b = −0.05546	0.99919	0.00006347	0.00797
	65	A = −0.01325; k = 0.00379; n = 1.1562 b = −0.02897	0.99745	0.000244	0.01563
	75	A = 0.00773; k = 0.00773; n = 1.13041 b = −0.00297	0.99766	0.000225	0.015
	85	A = 0.00248; k = 0.0068; n = 1.18729 b = −0.00407	0.99889	0.000117	0.0108
FZ OPW	55	A = 0.03252; k = 0.00366; n = 1.25688 b = 0.02386	0.99937	0.0000590518	0.00768
	65	A = 0.01903; k = 0.00621; n = 1.16875 b = 0.01875	0.99967	0.0000500652	0.00548
	75	A = 0.01282; k = 0.00495; n = 1.2524 b = 0.00626	0.99904	0.00010258	0.01013
	85	A = 0.00819; k = 0.0042; n = 1.31749 b = 0.00167	0.99902	0.0001114	0.01056
AA-FZ OPW	55	A = 0.02503; k = 0.00729; n = 1.16029 b = 0.02219	0.99975	0.0000219945	0.00469
	65	A = 0.02077; k = 0.00686; n = 1.18901 b = 0.02001	0.99994	0.00000581421	0.00241
	75	A = 0.01716; k = 0.00647; n = 1.26956 b = 0.01361	0.99947	0.0000561902	0.0075
	85	A = 0.00424; k = 0.00628; n = 1.2718 b = 4.36784 × 10−4	0.99922	0.0000898175	0.00948

Abbreviations: R²—the determination coefficient; χ²—reduced chi-square; RMSE—root mean square error; Fresh OPW—no pretreatment; AA OPW—dipped peels in acetic acid solution; FZ OPW—frozen and thawed peels; AA-FZ OPW—dipped–frozen combined pretreatment; k, a, g, k1, k2, b—drying constants.

displays the model coefficients and statistical results for the selected models. It has been shown that the R², χ², and RMSE values of these models ranged from 0.79321 to 0.99994, 22.01 × 10⁻³ to 5.81 × 10⁻⁶, and 0.14835 to 0.0024, respectively, indicating that all of the models could demonstrate a satisfactory level of fit between experimental MR and the predicted one. Among the investigated models, the Aghbaslo model performed well in describing the drying process for both fresh OPW and AA OPW, particularly at the 85 °C drying temperature. However, for FZ OPW and AA-FZ OPW, Balbay and Sahin’s model was the best fit for explaining the OPW drying process at 65 °C drying temperature. These models (Aghbaslo, Balbay and Sahin) have also shown satisfactory MR prediction results for the freeze drying of Citrus medica fruit under varying thicknesses and cabin pressures [42].

The ANN modelling was the second method adopted for MR estimation; this model was run using four input parameters: drying time, drying temperature, pretreatment or not, and peel quality (fresh or frozen). The training datasets were employed to evaluate the best neuronal and hidden layer count combination for neural network-based multi-layered modelling and to identify the most accurate prediction approach. The range of the hidden layers and the number of neurons were 1–14 and 4–32 (with a 4-step range), respectively. When a high R² value and a lower RMSE value were combined, an optimum model was identified, signifying stronger prediction accuracy. For the Opuntia ficus-indica PW drying, the ANN model with 24 neurons and 14 hidden layers was determined to be the most effective structure with the lowest RMSE (0.016448) and the best R² (0.996164). Based on this optimal ANN model, Figure 3 presents the comparison between the experimental and predicted moisture ratio (MR). For training, validation, and testing, the ANN predicted R² values were 0.9984, 0.9902, and 0.9893, respectively.

Overall, according to statistical criteria, the R² values were 0.79321–0.99994 and 0.996164 for thin-layer modelling and ANN, respectively. The RMSE values were 0.14835–0.0024 and 0.016448 for thin-layer modelling and ANN, respectively. It is clearly shown that mathematical models can accurately predict MR evolution during the drying process under each of the operating conditions. However, ANN stands out as a superior prediction tool since it is employed for solving more complex problems, analyzing all data simultaneously and generating different outputs (such as drying rate, moisture content, and physicochemical properties) [43,44].

3.3. Effect of Drying Temperature and Pretreatment on Physical Characteristics of Dried Peel Wastes

The fresh and frozen PW was dried under different operating temperature conditions without and with acetic acid pretreatment. Acetic acid (CH₃COOH), the main component of vinegar, is widely used in the food industry for its preservative and acidifying properties [45]. As a preservative, it inhibits the growth of pathogenic microorganisms by lowering the pH of foods, making it an essential additive for extending the shelf life of products, such as pickles, sauces, and condiments. Its acidifying action also helps improve the flavour and stability of foods by balancing their taste profile and enhancing their texture.

In the context of this study, acetic acid was used as a chemical pretreatment to optimize the drying of Opuntia ficus-indica PW. Its application demonstrated a significant improvement in drying efficiency by modifying the cellular structure of the peels, thereby facilitating water evaporation. This approach aligns with the observations of Hiranvarachat et al. (2011) [9], who showed that acid treatments increase the permeability of plant tissues, thereby accelerating dehydration processes. Additionally, Deshmukh and Manyar (2020) [45] confirmed the versatile function of acetic acid in a range of industrial processes, including the food industry.

3.3.1. Effect of Drying Temperature and Pretreatment on Water Activity and Moisture Content of OPW Powder

The water activity (a_w) and moisture content (MC) of the obtained OPW powder at different drying temperatures with and without the different applied pretreatments (AA OPW, FZ OPW, AA-FZ OPW, and fresh OPW), respectively, are presented in Table 2. The results showed a reduction in a_w and MC of around 28% for Fresh OPW, AA OPW, FZ OPW, and AA-FZ OPW with increasing drying temperatures from 55 °C to 85 °C. These results confirmed the fundamental principles of drying, according to which a higher temperature accelerates the vaporization of water [24]. This trend is corroborated by the study of Lahsasni et al. (2002) [46], where equilibrium moisture content decreases with increasing temperatures determined by the sorption isotherms for Opuntia ficus-indica peels. In addition, it has already been shown in the kinetic drying results that the AA pretreatment reduced the drying time for all tested temperatures, even for AA OPW and AA-FZ OPW (Section 3.1). This shortest drying time was associated with a decrease in a_w and MC of the OPW powder in the ranges of 1.99–4.84% and 4.42–15.77%, respectively. These results were consistent with the work of Kouhila et al. (2002) [47], which showed that chemical pretreatments can alter the cellular structure, thereby facilitating water extraction during the drying process. However, a higher decrease was recorded for the a_w and MC in FZ OPW and AA-FZ OPW in comparison to fresh OPW in the range of 10.73–22.31% and 8.92–23.78%, respectively. This difference is explained by the structural damage caused by the formation of ice crystals, which create microchannels facilitating water migration during drying, as shown by Li, Zhu, and Sun (2018) [39]. These results aligned with the conclusions of Junqueira et al. (2017) [21], who demonstrated that high temperatures, combined with chemical pretreatment and prior freezing, optimize the drying kinetics and improve the quality of dehydrated fruits.

The interaction between temperature, chemical pretreatment, and freezing shows that the drying process is multifactorial. A high temperature accelerates evaporation, pretreatment modifies the permeability of the tissues, and freezing creates preferential pathways for water migration. These observations are consistent with the principles described by Mujumdar (2015) [48] in his reference work on drying technologies.

3.3.2. Effect of Drying Temperature and Pretreatment on the CIELab Colour Parameters

Colour is a key quality indicator that influences the commercial value and customer acceptance of food products. Its variation during the drying process can be monitored using the CIELab colour parameters (L*—lightness, a*—redness, and b*—yellowness). These colour parameters for OPW powder with and without the different applied pretreatments (AA OPW, FZ OPW, AA-FZ OPW, and fresh OPW), respectively, at different drying temperatures are summarized in

Table 4. Effect of drying temperature and pretreatment on colour parameters (L*, a*, and b*) of OPW powder.

Drying Temperatures OPW Quality	55 °C	65 °C	75 °C	85 °C
		L*
Fresh OPW	52.917 ± 1.634 Cb	47.57 ± 3.043 Bd	49.901 ± 0.858 Dc	62.535± 1.06 Ba
AA OPW	52.327 ± 1.051 Cc	48.521 ± 1.143 Bd	56.383 ± 3.133 Cb	63.633 ± 0.296 Aa
FZ OPW	62.141 ± 0.871 Bc	63.391 ± 0.883 Ab	64.305 ± 0.344 Ba	58.496 ± 0.658 Dd
AA-FZ OPW	65.274 ± 0.725 Aa	63.391 ± 0.883 Ab	65.621 ± 0.645 Aa	60.637 ± 0.388 Cc
		a*
Fresh OPW	7.227 ±0.570 Ac	11.119 ± 0.796 Aa	11.029 ± 0.694 Aa	9.687 ± 0.213 Ab
AA OPW	5.534 ± 0.463 Bb	7.004 ± 1.016 Ba	7.328 ± 0.698 Ca	7.295± 0.154 Ca
FZ OPW	4.233 ± 1.010 Cc	5.598 ± 0.193 Cb	8.512 ± 0.488 Ba	8.466 ± 0.223 Ba
AA-FZ OPW	3.446 ± 0.285 Cc	6.616 ± 0.791 Bb	8.166 ± 0.767 Ba	7.12 ± 0.218 Db
		b*
Fresh OPW	19.047 ± 1.412 Cc	16.409 ± 3.709 Dc	20.865 ± 0.764 Cb	25.966 ±1.083 Ba
AA OPW	24.292 ± 0.523 Ba	20.077 ± 1.638 Cc	23.418 ± 1.882 Bab	22.818 ± 1.603 Cb
FZ OPW	34.103 ± 1.78 Aa	27.693 ± 0.236 Bc	28.896 ± 0.371 Ab	25.338 ± 0.505 Bd
AA-FZ OPW	34.483 ± 1.100 Aa	29.206 ± 0.636 Ab	29.078 ± 0.303 Ab	27.318 ± 0.251 Ac

Values are the mean of 3 independent determinations ± standard deviation. Uppercase letters indicate statistically significant differences (0.05) between Fresh OPW, AA OPW, FR OPW, and AA-FR OPW. Lowercase letters indicate statistically significant differences (0.05) between different drying temperatures. Abbreviations: Fresh OPW—no pretreatment; AA OPW—dipped peels in acetic acid solution; FZ OPW—frozen and thawed peels; AA-FZ OPW—dipped–frozen combined pretreatment.

. Their evolutions were variable regarding the drying temperatures and pretreatments. In fact, increasing drying temperatures enhanced lightness in Fresh OPW and AA OPW by the range of 15.22% and 17.81%, respectively, between 55 °C and 85 °C. However, this trend was the inverse in FZ OPW and AA-FZ OPW, since the L* values recorded a decrease in the range of 10.72% and 8.77%, respectively, for the same temperature interval. However, the a* and the b* intensities indicated an opposite behaviour by a significant increase in the redness character and a significant reduction in the yellowness by the range of 31.81–106.62% and 6.07–25.70%, respectively, between 55 °C and 85 °C. These outcomes were consistent with the Belkhir et al. (2025) [49] study on the convective drying of Opuntia ficus-indica peels at 40 °C and 60 °C. This observation might be the result of a significant temperature effect that causes enzymatic browning and degrading reactions like the Maillard reaction [49,50].

Furthermore, the lightness parameter showed a significant increase between fresh OPW and FZ OPW. This increase in L* values ranged from 17.49% to 31.03% and from 16.45% to 3.54%, respectively, in FZ OPW and AA–FZ OPW in comparison to fresh OPW for the different temperatures used, except for 85 °C. On the other hand, the colour parameters a* and b*, with inverted behaviour, demonstrated a significant decrease and increase, respectively, between FZ OPW and fresh OPW, and even in AA-FZ OPW and AA OPW. This behaviour might be explained by the pretreatment effect, which augmented the lightness of the dried samples by revealing a yellow aspect.

3.4. Effect of Drying Temperature and Pretreatment on Betalain Content in Dried OPW

Table 5. Effect of drying temperature and pretreatment on betacyanin, betaxanthin, and betalain content of OPW powder.

Drying Temperatures Quality OPW		55 °C	65 °C	75 °C	85 °C	F-Value	p-Value
Fresh OPW	Betacyanin	22.54 ± 1.75 b	26.05 ± 0.99 ab	29.67 ± 2.08 a	25.88 ± 0.69 ab	7.660	0.017
	Betaxanthin	21.57 ± 2.98 bc	23.60 ± 0.06 c	30.19 ± 1.15 a	27.95 ± 0.32 ab	220.042	0.000
	Betalain	44.11 ± 4.73 Abc	49.65 ± 0.93 Ac	59.86 ± 3.24 Aa	53.82 ± 1.01 Aab	19.867	0.01
AA OPW	Betacyanin	14.12 ± 0.02 b	13.46 ± 0.83 b	22.28 ± 1.06 a	23.59 ± 1.46 a	11.786	0.008
	Betaxanthin	19.59 ± 3.51 c	17.08 ± 2.32 c	24.57 ± 1.01 b	26.99 ± 0.04 a	13.214	0.004
	Betalain	33.71 ± 3.53 ABc	30.53 ± 3.15 Bc	46.85 ± 2.07 Bb	50.59 ± 1.41 Ba	13.214	0.004
FZ OPW	Betacyanin	12.17 ± 1.65 d	14.61 ± 0.25 c	18.77 ± 1.78 a	15.93 ± 0.28 b	14.286	0.003
	Betaxanthin	14.10 ± 1.30	15.37 ± 1.42	14.56 ± 0.39	14.26 ± 0.33	0.936	0.479
	Betalain	26.27 ± 2.95 BCb	29.97 ± 1.67 Bab	33.32 ± 1.38 Ca	30.18 ± 0.05 Cab	7.279	0.028
AA-FZ OPW	Betacyanin	10.08 ± 1.31 c	14.31 ± 0.41 b	16.20 ± 0.70 a	13.34 ± 0.57 b	22.349	0.001
	Betaxanthin	12.67 ± 1.03	14.77 ± 1.31	13.42 ± 0.80	12.59 ± 0.57	3.950	0.083
	Betalain	22.74 ± 2.34 BCb	29.08 ± 0.89 Ba	29.62 ± 1.50 Da	25.93 ± 0.73 Db	15.066	0.003
F-value	Betalain	21.716	8.929	14.286	14.286
p-value	Betalain	0.001	0.03	0.003	0.003

Values are the mean of 3 independent determinations ± standard deviation. Uppercase letters indicate statistically significant differences (0.05) between Fresh OPW, AA OPW, FR OPW, and AA-FR OPW. Lowercase letters indicate statistically significant differences (0.05) between different drying temperatures. Abbreviations: Fresh OPW—no pretreatment; AA OPW—dipped peels in acetic acid solution; FZ OPW—frozen and thawed peels; AA-FZ OPW—dipped–frozen combined pretreatment.

reports the betacyanin, betaxanthin, and betalain content in dried OPW obtained from various pretreatments at different drying temperatures (55, 65, 75 and 85 °C). The values, expressed in mg per 100 g dry weight PW (mg/100 g dw), are accompanied by their standard deviations, reflecting the variability of the measurements.

Betalain content in dried OPW ranged from 22.74 ± 2.34 mg/100 g dw to 59.86 ± 3.24 mg/100 g dw, with betacyanin (49%) and betaxanthin (51%) in the differently treated materials (Fresh OPW, AA OPW, FR OPW, and AA-FR OPW) and various drying temperatures (55, 65, 75 and 85 °C) (Table 5). These results were consistent with those reported by Ettalibi et al. (2020) [36] and Belkhir et al. (2025) [49] with 99.1 mg/100 g dry weight and 34.59 ± 0.34 mg/100 g dw, respectively, in Opuntia ficus-indica dried peels at 60 °C.

Furthermore, the betalain content increased significantly as drying temperatures increased, peaking at 75 °C before declining further. This behaviour is clearly linked to betalain thermal degradation (drying temperatures exceeding 75 °C). A number of well-established reactions, including decarboxylation, dehydrogenation, glycosidic bond cleavage, and hydrolysis, can convert betanin, a main fragment of betacyanin, into new derivatives, such as betanidin, neobetanin, betalamic acid, and cyclo-DOPA-5-O-glycoside (colourless) [51,52,53]. The isomerization and decarboxylation reactions, particularly, alter the betalain chromophore structure; these changes are typically followed by changes in absorption maxima, resulting in yellow pigments such as neobetacyanins, betalamic acid, and newly formed betaxanthins [54].

Additionally, among the pretreatments applied, the highest betalain content was recorded using the acetic acid dipping as a pretreatment with 46.85 ± 2.07 mg/100 g dw, at 75 °C drying temperature, followed by FR OPW (33.32 ± 1.38 mg/100 g dw) and AA-FR OPW (29.62 ± 1.50 mg/100 g dw), showing a decreasing rate of about 21.73%, 44.33%, and 50.51%, respectively, in comparison to fresh OPW at the same drying temperature. This result may be explained by different interfering mechanisms, such as polyphenoloxidase inactivation under slight acidification or betalain releasing in the dipped AA solution. Furthermore, the freezing pretreatment showed a significant effect on the reduction in betalain, by inducing intense cell rupture, increasing pigment exposure to oxygen and enzymes, and thus, accelerating betalain degradation. This result has already been demonstrated by Castro-Enríquez et al. (2020) [55], who reported that cell rupture brought on by ice crystal formation resulted in minimal betalain retention compounds.

At 75 °C, the combined pretreatment AA-FR OPW reduced betalain retention even further. In this case, the pretreatment changed the pH and oxygen accessibility of the cellular milieu, promoting specific breakdown pathways, including enzymatic and chemical oxidation.

Even if the AA and freezing pretreatments used in this study were intended to maintain the fresh OPW until the drying process, the results clearly showed that various pretreatments (AA OPW, FZ OPW, and AA-FZ OPW) had a limited effect on improving betalain retention from OPW in comparison to fresh OPW, although the energetic aspects were encouraging (Section 3.1) [56].

3.5. Effect of Drying Temperature and Pretreatment on Total Polyphenol Content in Dried Peel

Table 6. Effect of drying temperature and pretreatment on total phenolic content of OPW powder.

Drying Temperatures Quality OPW	55 °C	65 °C	75 °C	85 °C	F-Value	p-Value
Fresh OPW	10.68 ± 0.18 Ac	9.78 ± 0.815 Abc	12.29 ± 0.421 Aa	10.75 ± 0.647 Aab	20.118	0.000
AA OPW	7.75 ± 0.365 Cb	8.51 ± 0.814 Bab	10.07 ± 1.151 Ba	9.47 ± 0.525 Ba	16.757	0.000
FZ OPW	9.37 ± 0.944 B	10.08 ± 0.743 A	10.75 ± 0.953 B	9.56 ± 0.159 B	3.452	0.066
AA-FZ OPW	8.56 ± 0.129 Bc	7.87 ± 0.254 Bd	9.53 ± 0.207 Ba	9.08 ± 0.283 Bb	59.809	0.000
F-value	19.695	13.461	61.774	10.734
p-value	0.000	0.000	0.000	0.002

Values are the mean of 3 independent determinations ± standard deviation. Uppercase letters indicate statistically significant differences (0.05) between Fresh OPW, AA OPW, FR OPW, and AA-FR OPW. Lowercase letters indicate statistically significant differences (0.05) between different drying temperatures. Abbreviations: Fresh OPW—no pretreatment; AA OPW—dipped peels in acetic acid solution; FZ OPW—frozen and thawed peels; AA-FZ OPW—dipped–frozen combined pretreatment.

reports the total polyphenol content (TPC) determined in the dried OPW across the different applied pretreatments and drying temperatures. The TPC results ranged from 7.87 ± 0.254 g GAE/100 g dw to 12.29 ± 0.421 g GAE/100 g dw depending on the drying temperature and pretreatment applied. These values were similar to those reported by previous works that evaluated TPC in dried prickly-pear by-products, such as El Mannoubi (2021) [57] with 9.223 ± 0.371–86.287 ± 0.741, Scarano et al. (2022) [58] with 36.17 ± 0.003 mg/g, and Belkhir et al.(2025) [49] with 35.54 mg/g in convective drying at 60 °C [59].

The highest TPC (12.29 ± 0.421 g GAE/100 g dw) was found in fresh OPW at 75 °C. According to Shahidi et al. (2025) [56] and Belkhir et al. (2025) [49], this rise in TPC levels throughout the drying process can be attributed to the release of phenolic compounds originally bound to cell wall polymers, which become more extractable when heat weakens the bonds (as ester and glycosidic bonds). Furthermore, the lower TPC values observed in fresh OPW at 55° and 65 °C with extended drying durations can be related to enzymatic oxidation by endogenous enzymes (polyphenol oxidases or peroxidases), which, once released after tissue disruption, can oxidize phenolics into quinones, which may further polymerize into insoluble brown pigments, reducing the measurable TPC. This enzymatic degradation is often observed when drying is slow or at relatively mild temperatures [49,60].

When pretreatments were applied (AA OPW, FZ OPW, and AA-FZ OPW), the TPC values were decreased in comparison to the fresh OPW. The highest TPC value (10.75 ± 0.953 g GAE/100 g dw) was obtained for FR-OPW at 75 °C with no statistical difference between the different pretreatments (AA OPW, FR OPW, and AA-FR OPW), especially at 75 °C and 85 °C. Thus, the freezing pretreatment weakens cell integrity through ice-crystal formation and increases contact between phenolic substrates and oxidative enzymes, thereby accelerating enzymatic or chemical oxidation. Moreover, the leaching phenomenon during thawing may contribute to the additional loss of bioactive molecules [60]. This phenomenon has been observed in various fruits, as noted by Bonat Celli et al. (2015) [61]. As a result, the final TPC measured in dried OPW, whether fresh OPW or with the different pretreatments (AA OPW, FZ OPW, and AA-FZ OPW), reflects a dynamic balance between heat-induced phenolic liberation, which increases extractable polyphenols, and thermal and oxidative degradation mechanisms, which reduce polyphenol stability and extractability (Xiao, 2022 [62]). Given these considerations, preserving polyphenols seems optimal when drying is conducted at 75 °C without freezing or harsh pretreatments, maximizing release while minimizing degradation.

4. Conclusions

Opuntia ficus-indica peel waste (OPW) is a matrix rich in bioactive components with high nutritional and biotechnological benefits. This waste requires an innovative preservative process. Drying process results showed that the increase in temperature significantly accelerates the drying process, particularly at 85 °C. Both freezing and acetic acid pretreatments alter the cellular structure, thus facilitating a faster migration of water and a better drying rate. However, regarding the preservation of bioactive compounds, particularly polyphenols and betalains, the best results were observed at 75 °C without any pretreatment, especially for fresh samples. In general, acetic acid ad freezing pretreatments led to a decrease in these compounds, likely due to enzymatic degradation and oxidation. From a colorimetric perspective, high temperatures and freezing generally improved lightness (L*) and the intensity of the yellow value (b*), but reduced the red component (a*), especially in frozen samples. Thus, the ideal conditions for an effective and time-efficient drying process corresponded to 85 °C with pretreatment to accelerate the process, while the best compromise for preserving bioactive compounds was achieved with drying at 75 °C without pretreatment for fresh peels. Both thin-layer and ANN modelling had good performance in predicting the moisture ratio of OPW. This information is essential for developing strategies to enhance the recovery of this agri-food by-product, rich in functional compounds. These attributes, combined with their applications as natural colourants, fibre-enriched ingredients, or nutritional additives, make them an ideal candidate for the development of functional products in the agri-food, nutraceutical, and cosmetic sectors. Developing such uses for this agri-food by-product is fully aligned with a circular bioeconomy approach and sustainable development.