Optimization of Pinhão Extract Encapsulation by Solid Dispersion and Application to Cookies as a Bioactive Ingredient

Pinhão residues have a wide range of bioactive compounds and encapsulation can be one of the alternatives to increase their bioavailability. Thus, this work aimed to apply pinhão extract, pure and encapsulated by solid dispersion, in the formulation of cookies as a bioactive ingredient. For that, pinhão extract was encapsulated in different biopolymers (sodium caseinate, gelatin, and gum arabic) and with different shear mechanisms (sonication, Ultra-Turrax, and magnetic stirring). The best encapsulation procedure has been defined by a chemometric analysis (hierarchical cluster analysis), considering thermal properties (DSC) of particles and ( +)-catechin encapsulation efficiency (HPLC). The optimized conditions were gelatin as encapsulation agent and Ultra-Turrax as shear mechanism (70.1 ± 2.8 °C maximum endothermic peak temperature and 96.0 ± 2.3% ( +)-catechin encapsulation efficiency). The phenolic profile of the encapsulated extract showed the presence of ( +)-catechin (0.31 ± 0.01 (mg/gparticle), protocatechuic acid (0.29 ± 0.00 mg/gparticle), and ( −)-epicatechin (0.11 ± 0.00 mg/gparticle). Both the pure and encapsulated extracts were incorporated into the cookie formulation, which was characterized in terms of centesimal composition, color parameters, texture, and sensory aspects. It was found that cookies with the pure and the encapsulated extract showed significant differences concerning the centesimal composition, products added with pinhão extract and encapsulated extract presented higher values when compared to the control, probably influenced by the mineral content of the pinhão. In addition, higher hardness values were detected for cookies formulated with the encapsulated extract, which possibly negatively affected the consumer’s sensory perception.


Introduction
The seeds of Araucaria angustifolia (Bertol.) Kuntze are a native product widely consumed in southern Brazil. These seeds, called pinhão, are collected by small producers as an alternative for sustainable income . They are rich in starch, proteins, and flavonoids, presenting high nutritional value (Brandão et al., 2019). The use of pinhão bio-waste as a sustainable alternative has been proposed by several research groups, which have evaluated different strategies, such as, for example, the use of sterile bracts as a natural absorbent to remove methylene blue from aqueous solutions (Matias et al., 2019) and pinhão seed coats as a source of antimicrobial extract (Trojaike et al., 2019), as well as to produce extracts with other bioactive properties (Branco et al., 2019;da Silva et al., 2014;Koehnlein et al., 2012;Peralta et al., 2016;Souza et al., 2014).
The bioactivities related to pinhão extracts are linked to the presence of phenolic compounds (de Souza et al. 2020;Silva et al., 2019). In the case of pinhão seed coat extracts, catechins are the main compounds obtained when hydroalcoholic solvent mixtures are used . Catechins are flavonoid compounds known to have improved antioxidant properties than α-tocopherol, butylated hydroxyanisole, or butylated hydroxytoluene (Ahmad et al., 2019). However, catechins, as other phenolic compounds, present low bioavailability, which has been partly attributed to degradation and metabolism in the gastrointestinal tract, low membrane permeability, and pre-systemic hepatic elimination (Ye & Augustin, 2019).
The encapsulation of bioactive compounds is an effective strategy to improve their stability and improve bioavailability. Other results can be achieved when encapsulation is applied to bioactive extracts or compounds, such as masking an unpleasant taste/aroma and allowing it transformed into a useful physical form, e.g., from a liquid to a powder (Ye & Augustin, 2019). According to Raddatz and Menezes (2021), the encapsulation of active compounds is an effective means of meeting the constant changes of the consumer market, allowing the food industry to create products with a functional and nutritional appeal. The size of the food encapsulation market exceeded USD 34 Billion globally in 2019, being this demand driven by the improvement of end-products shelf life along with reduced nutrient loss (Ahuja & Rawat, 2020).
The encapsulation of pinhão extracts has been proposed by a few authors as follows. Dorneles and Noreña (2020) used hydrolyzed pectin/collagen and partially hydrolyzed polydextrose/ guar gum as encapsulating agents. The authors obtained microparticles by spray-drying and also by lyophilization, resulting in improved stability of antioxidant capacity during storage. Fonseca et al. (2020) encapsulated the pinhão extract in electrospun starch fibers. Improved thermal stability, protection against harmful external elements, and control of the phenolic compound were the outstanding results.
The solid dispersion encapsulation technique is widely used to increase the solubility and bioavailability of poorly water-soluble compounds. This technique allows the formation of solid amorphous solutions, where the encapsulated compound and the carrier (encapsulating agent) are miscible and soluble. Thus, a homogeneous molecular interaction between the components of the formulation is created (Vasconcelos et al., 2007) as miscibility seems to be essential to maintain the long-term physical stability of the solid dispersion (Newman et al., 2012).
The interaction between the bioactive compounds and encapsulating agents during the encapsulation procedure can be improved by the shear mechanism (Phunpee et al., 2018). Shear forces can be generated using several different approaches, including the application of ultrasound generated by sonicators, the use of high-shear mixers (for example, rotor-stator systems), or passing through high-pressure homogenizers (Chandrapala et al., 2014).
Thus, to improve the viability of pinhão extract encapsulation, a systematic study is proposed in the present work, where three low-cost GRAS (Generally Recognized as Safe) encapsulating agents were applied (gelatin, arabic gum, and sodium caseinate), as well as three shear mechanisms (magnetic stirring, ultra-turrax, and ultrasound) to improve the interaction between the compounds of the extract and the encapsulating agents. The technique of encapsulation by solid dispersion was proposed, as an alternative considered easier to apply and less costly than other encapsulation techniques (Leimann et al., 2019). The lyophilized solid dispersions were evaluated concerning their thermal characteristics and encapsulation efficiency in terms of ( +)-catechin. A hierarchical cluster analysis (HCA) was applied to verify similarities between the different encapsulated materials obtained and to select an experimental condition to be applied in the formulation of cookies.

Extraction and Encapsulation of Pinhão Extract
The extract was prepared as described by de Freitas et al. (2018); the pinhão was boiled in water (500 g/L) for 2 h, filtered on qualitative filter paper, and then frozen in an ultrafreezer (Liobrás, Brazil) at − 90 °C. The encapsulating conditions evaluated in the encapsulation procedure are shown in Table 1. The technique applied to encapsulate the pinhão extract was the solid dispersion, described by Karavas et al. (2006) with minor modifications. Briefly, the extract was thawed (50 mL) and then Tween 80 (90 mg), used as a surfactant, was added under magnetic stirring. After that, the encapsulating agent (348 mg of gelatin, sodium caseinate, or arabic gum) was added and the mixture was heated at 50 °C for 15 min to allow the encapsulating agents to solubilize. After this step, each mixture was stirred according to the conditions in Table 1. A Fischer Scientific sonicator (Loughborough, UK, 150 W, 1/8′ tip) was used for the shear acoustic cavitation mechanism, with a 30-s pulse on/10 s off for 15 min. A Fisatom (model 5B, Brazil) magnetic stirrer was used for a low-shear mixture. An Ultra-Turrax (Ika, T25, Germany) was used as the third mechanism for high-shear generated by the rotor-stator system. Finally, all the solutions were lyophilized (Liotop, Liobrás, L101, Brazil).

Thermal Characterization
The thermal characterization of the encapsulated extracts was performed using a differential scanning calorimeter (DSC, Perkin Elmer, 4000, USA). Samples (approximately Table 1 Experimental conditions applied for the production of encapsulated pinhão extract and results obtained by DSC (T (°C)-maximum endothermic transition peak temperature) and encapsulation efficiency in terms of catechin (Catechin EE (%) 1 3 5 mg) were placed in closed aluminum holding pans and analyzed in a temperature range of 0 to 200 °C under a nitrogen flow rate of 20 mL/min and a heating rate of 20 °C/min . The analysis was done in duplicate for each sample.
Finally, the results were used to determine ( +)-catechin encapsulation efficiency (CEE) according to Eq. (1), where C T is the total concentration of ( +)-catechin present in the extract used to prepare the particles according to HPLC-DAD-ESI/MSn analysis (mg/g extract ), and C F is the concentration of free ( +)-catechin extracted from particles determined by HPLC-DAD (mg/g extract ) (Pillai et al., 2012). Samples were analyzed in triplicate.

Cookie Production
The optimized particle formulation was morphologically characterized by transmission electron microscopy (TEM; JEOL model JEM 2100, 200 kV, USA). Samples diluted with ethanol were dripped onto 300 mesh copper grids coated with carbon and submitted to analysis.
Three cookie formulations were prepared as described by , with modifications. A control formulation (C, without adding encapsulated or pure extract) was prepared, one with the addition of 10% of encapsulated extract concerning the total weight of the wheat flour (EE) and the other with the incorporation of the equivalent of pure extract to the weight of extract present in the added particles (PE). To produce the cookies, flour (111 g), salt (1.05 g), sugar (50 g), and baking soda (2.50 g) were initially mixed manually. Then, the particles were added with encapsulated extract (11.1 g), or the pure extract (1.49 g), being mixed for 3 min. After that, butter (33.75 g) was added and mixed for another 3 min. In the sequence, water was incorporated (20 mL for control and 25 mL for EE and PE due to the increase in the solids content in these formulations), and the mixture was homogenized for 3 min. The dough was spread with the aid of a roller, with a standardized thickness (5 mm) and then cut into discs of 50 mm in diameter. Finally, the cookies were baked in an oven (Tedesco, FTT 240E, Caxias do Sul, RS, Brazil) at 170 °C for 15 min. The cookies were cooled to room temperature, then packed in polyethylene bags (20 units per package), and stored at 25 °C.

Weight Loss, Texture Measurements, and Color Determination
Weight loss (WL) of the cookies was calculated using Eq. (2), where W dough is the weight (g) of the cookie dough before baking, and W cookie is the weight (g) of the cookies after baking (Rodríguez-García et al., 2014).
Mechanical testing was performed first in the cookie dough (Texture Profile Analysis test, TPA) and also in baked cookies (hardness test), both using a Stable Micro Systems texturometer (TA-XT Express model, Godalming, Surrey, UK). To assess the cookie dough, TPA was applied to dough cylinders (50 mm in diameter × 5 mm in height) equipped with a 10-kg load cell and using a 34-mm cylindrical probe (P/34), following the method described by Venturini et al. (2018), with minor modifications. Ten samples of cookie dough were analyzed for each treatment (C, EE, and PE), with a compression of 25% of samples height, a speed test of 5 mm/s, and an interval of 5 s between compressions. Parameters evaluated were hardness (N), adhesiveness (N.s), springiness (mm), cohesiveness (dimensionless), and resilience (dimensionless).
The hardness test was applied to evaluate baked cookies also as described by Venturini et al. (2018). Initially, the thickness of baked cookies was measured on a digital caliper (Ford, Brazil), then the cookies (10 samples for each treatment and storage time: 0, 15, and 30 days) were compressed to 50% of their height using a 2-mm cylindrical probe (P2) with 10 (mm/s) of compression speed. The puncture force is the highest force (N) detected in the test.
For color analysis, ten samples of each treatment were evaluated concerning the parameters L* (luminosity), C* (Chroma), and h° (Hue angle) with the Delta Vista 450G colorimeter (Delta Color, Brazil) with standard illuminant D65 and standard observer of 2°. The samples were evaluated at time intervals of 0, 15, and 30 days.

Proximate Composition of the Cookies
Moisture was determined by the gravimetric method, at 105 °C (air circulation oven, Cienlab, CE220/1152, Brazil) until a constant weight was reached. In the analysis of the ashes, the cookies were incinerated in a muffle at a temperature of 550 °C (muffle, Nova Ética NT380, Brazil). The Microkjeldahl method (TE-0364, Tecnal, Brazil) was used to obtain the protein content, and the lipid determination was done by the Soxhlet method (Marconi MA 044/5/50, Brazil) (Lutz, 2008).

Microbiological Quality and Sensory Analysis
The cookies were submitted to microbiological analysis of coliforms at 45 °C, using Salmonella sp. and coagulasepositive Staphylococcus aureus as required by the ANVISA resolution (Brazil, 2001). The sensory tests were performed at the Sensory Analysis Laboratory of the Federal Technological University of Parana (UTFPR), under the approval of the Research Ethics Committee of the same university, under protocol number 13799119.9.0000.5547. The cookie samples were sensorially evaluated by the ranking preference test (ISO, 2006) by 63 untrained assessors (29 male and 34 female), aged between 17 and 49 years old, involving students and employees of the institution. Three samples (C, EE, and PE) were randomly presented, and the assessors were asked to order them in decreasing order of overall preference. Cookies were served on white plates coded with three random digits, in random order. The assessors were asked to drink water before testing the samples. In the analysis of sensory data, a rating of 1 was assigned to the most preferred sample and a rating of 3 to the least preferred cookie.

Statistical Analysis
Hierarchical cluster analysis (HCA) was performed using MATLAB R2008b (MathWorks Inc., Natick, MA). The objective was to identify similarities between the results of the experimental conditions tested. The results obtained with the thermal characterization (maximum endothermic transition peak temperature) and encapsulation efficiency were placed in columns and the experimental runs in rows. Ward's method based on Euclidean distance was used for cluster formation and sample pattern recognition (Granato et al., 2018;Šoronja Simović et al., 2017). The results obtained for extract composition (HPLC-DAD-ESI/MSn), cookie texture, color, and proximate composition were evaluated using analysis of variance (ANOVA), and the averages of the results were compared using Tukey's test at a significance level of 5% (p < 0.05) using the software Statistica 7.0 (Statsoft, USA). In the sensory tests, a Friedman test (p ≤ 0.05) was used to estimate whether there were significant differences, and tables of the maximum distance between the samples (Christensen et al., 2006) were used to compare these rank sums.

Characterization of Extract and Microencapsulated Extracts
The phenolic profile determined for the pinhão extract is shown in Table 2. Three compounds were identified by comparing the retention time, maximum absorption wavelengths in the visible region, and mass spectral data with the available standard compounds. ( −)-Epicatechin was the major compound found in the extract. The total phenolic content of the extract was equal to 2.43 mg/mL, lower than those reported by de Oliveira (2021) (4.50 ± 0.06 mg/ g extract ), de Souza (2020) (6.42 ± 0.01 mg/g extract ), and da Silva et al. (2019) (8.25 ± 0.05 mg/g extract ). With the statistical analysis, it was found that there was a statistical difference between the concentrations of protocatechuic acid and ( −)-Epicatechin (p < 0.05). On the other hand, the concentration of ( +)-Catechin was equivalent to the other compounds (p > 0.05).
The thermograms of pinhão extract microencapsulated in gelatin, sodium caseinate, and arabic gum are shown in Fig. S1(a, b, and c), respectively. The maximum endothermic peak temperatures are presented in Table 1. In the DSC thermograms, the maximum temperature of the endothermic peak was recorded since the melting temperature of the encapsulating agents and the evaporation of water are overlapped in this region. Lower values of maximum endothermic transition temperatures were obtained for all the particles in which the applied shear mechanism was the sonicator and the particles produced with gum Arabic as encapsulating material presented a higher transition temperature, differing significantly from the other matrices (gelatin and sodium caseinate, Table 1). On the other hand, higher temperature results were determined for all samples submitted to magnetic stirring, and there is no significant difference between the encapsulating agents used (Table 1). This result may be associated with the cavitation-induced cleavage of macromolecular chains (Leimann et al., 2013). The sound waves generated during sonication pulses interact with gas bubbles present in the liquid, leading to their coalescence and collapse. This cavitation process results in the generation of high temperatures within these bubbles (Ashokkumar et al., 2007). Samples processed by Ultra-Turrax presented similar transition temperatures to sonicated samples. This may also be attributed to the high shear physical mechanism, which can generate shear force orders of magnitude higher than overhead stirring (Chandrapala et al., 2014).
Encapsulation efficiency results are also shown in Table 1. The results revealed that for all the encapsulating agents, as well as the shear mechanisms used, the encapsulation efficiencies were at least 91.4% and did not differ statistically (p < 0.05) about the encapsulating matrix and agitation mechanism used in the production of particles. Jain et al. (2015) when developing microcapsules using Whey Protein Isolate and acacia gum by the complex coacervation method to encapsulate beta-carotene, achieved an encapsulation efficiency close to 77.3%, a lower value than that achieved by the solid dispersion technique used in the present study. Figure 1 presents the similarities between the experimental conditions tested found by the hierarchical cluster analysis (HCA). It is possible to observe that four groups were formed considering a dissimilarity equal to 0.3. All experiments with sodium caseinate as the encapsulating agent were grouped (C1, C2, and C3). Furthermore, among these samples, the experiments submitted to magnetic stirring and Ultra-Turrax were the most similar among all samples (C1 and C2). This is an indication that sodium caseinate was the least encapsulating agent influenced by the shear conditions in encapsulation efficiency and thermal properties. On the other hand, for gelatin and arabic gum, the dissimilarities were evidenced by the clustering of samples prepared with magnetic stirring in one group (G1 and GA1), and two groups related to each encapsulating agent: G2 and G3; GA2 and GA3. This result evidence that for these encapsulating agents, only the shear mechanism by magnetic stirring affected the evaluated responses.

Selection of Encapsulation Conditions
Based on the hierarchical cluster analysis, the experimental condition with gelatin as encapsulation applying Ultra-Turrax as shear mechanism (G3 sample) was chosen to be applied in the formulation of cookies.
The produced particles were also characterized in terms of their morphology and size by transmission electron microscopy (TEM) and the images are presented in Fig. 2A. It is possible to observe a needle-shaped structure with a size in the range of 50-150 nm. Encapsulated formulations can be presented either with spherical geometry (single-particle structure), with irregular geometry (aggregated structure) (Jain et al., 2016), or even more complex, depending on the encapsulation technique, encapsulating material, shear forces, pH, etc. which are applied to the encapsulating system (Gharieh et al., 2019;Silva et al., 2017).
The gelatin particles containing pinhão extract as well as the crude extract were applied to cookie formulations. The results of the texture profile analysis (TPA) applied to cookie dough are shown in Table 3.
It is possible to observe that the hardness parameter showed a significant difference (p < 0.05) among all treatments. The hardness result of cookies added with the encapsulated extract was approximately twofold higher than the result obtained for the control formulation. For cookies added with unencapsulated pinhão crude extract, this effect was less evidenced. Wang et al. (2007) found that the addition of green tea extract to bread formulations increased hardness and adhesiveness. The interactions between phenolics with proteins and starch affect the physical and rheological properties of wheat doughs . In the present study, there was no significant difference in adhesiveness for the PE sample when compared to the control, and cohesiveness also increased. For the dough added with encapsulated pinhão extract, there was a decrease in adhesiveness and an increase in springiness and cohesiveness. Yu et al. (2019) added pigskin gelatin to bread formulations and noted that gelatin induced the dough to be less stiff and softer. In addition, the authors found that an increase in the level of gelatin gave rise to an increase in dough water absorption, due to more hydrogen bonding interactions between water and hydroxyl groups in the gelatin structure. In the present work, a larger amount of water had to be added to the EE cookie formulation, due to the presence of gelatin.
The only significant differences (p < 0.05) were found among the proximate composition parameters for ash and moisture content. Cookies added with crude pinhão extract and encapsulated extract presented higher values for both parameters when compared to the control. This behavior may be related to the presence of mineral content in the pinhão extract (Biel et al., 2020), as well as to a greater water retention capacity of both, crude and encapsulated extract (due to gelatin). This was confirmed by the weight loss since the results were statistically equal for all treatments (p > 0.05), and the moisture content was higher for cookies added with crude pinhão extract and encapsulated pinhão extract compared to control, suggesting a higher load of water.
The results of puncture force showed that the formulation prepared with the addition of encapsulated pinhão extract (EE) resulted in harder cookies. In addition, during the first 15 days of storage, there was a significant increase (p < 0.05) in the puncture force of 1.6-fold for cookies added with crude pinhão extract (PE) and encapsulated pinhão extract (EE). This result is probably related to moisture loss during storage (Piga et al., 2005). Barros et al. (2020) when analyzing cookies formulated with cocoa shells, also noticed an increase in hardness. The authors stated that water-absorbent compounds, such as fibers and proteins, can contribute to the sticky nature of the dough, reducing its extensibility and increasing hardness, especially after baking. Rocha Parra et al. (2019) realized that the product showed greater hardness by partially replacing wheat flour with apple pomace in the formulation of cookies.  The process of encapsulating the pinhão extract was not able to reduce the effect of the color change on the cookies, as can be seen in the images presented in Fig. 2 and results from Table 4. Regarding the luminosity (L*), it was possible to note that cookie samples added with crude pinhão extract (PE) and encapsulated pinhão extract (EE) are darker than control cookies (p < 0.05). Also, there was a slight decrease in luminosity throughout storage time and the luminosity was constant over time for PE and EE cookies. The pinhão extract has a dark brown color, but even with the encapsulation, which could mask the color of the extract, there was a visible effect on the final color of the cookies.
Chroma results also showed that the control samples have more saturated color than cookies added with crude and encapsulated pinhão extracts (Wrolstad & Smith, 2017). The hue angle (h°) represents the attribute of visual perception according to which the color of the object appears to be similar to red, yellow, green, or blue, or also to a combination of adjacent pairs of these colors considered in the L*C*h color space (Minolta, 2020;Zhang et al., 2019). The red color is represented by a hue angle of 0° and yellow of 90°; thus, it can be seen that the control cookies present a tonality tending to light yellow, in contrast to cookies containing crude extract and encapsulated pinhão extract, which showed lower hue angle values, as well as lower luminosity results, tending to a dark red tonality.

Microbiological and Sensory Analysis
The microbiological evaluation of the produced cookies was performed to assure the safety of panelists' consumption. Results (Table S1) demonstrate that all prepared cookies are below the tolerable limits for Staphylococcus aureus and coliforms at 45 °C, as well as present the absence of Salmonella sp, as described in RDC N° 12 of January 2, 2001, and are therefore safe for human consumption.
The results of total rank sums obtained from the ranking preference test were subjected to Friedman's test (p ≤ 0.05) and the rank of sums was compared to the tables of maximum distance between samples (Christensen et al., 2006). By this method, all possible pairs of tested samples can be compared. Statistical analysis of data from the preference ranking test revealed that all cookie samples differ significantly (p ≤ 0.05) from each other in terms of preference. The control sample (C) was the most preferred and the EE sample was the least preferred cookie. Cookie PE had an intermediate preference. Preference was possibly influenced by the addition of particles because it significantly affected

Sensory analysis
Total rank sums (Σ orders) 91 a 120 b 166 c the textural aspects of the product, and the EE presented a higher hardness compared to PE (Table 4). Indeed, the addition of fruit and vegetable by-products may reduce the sensory acceptability of bakery products (Rocha Parra et al., 2019). Ghoshal and Kaushik (2020) when evaluating the partial substitution of wheat flour with some flour in cookies noticed a decrease in sensory acceptability, being attributed to the darker coloration that products developed, a fact that was also observed in the present study.

Conclusions
The pinhão extract encapsulated in gelatin and obtained with the Ultra-Turrax shearing mechanism was selected with the hierarchical cluster analysis among the evaluated formulations. The particles were incorporated into cookies, as well as the pure pinhão extract. These cookies presented higher values for ash and moisture content when compared to control samples, probably due to the higher mineral content of pinhão extract. Also, concerning the texture parameters, cookies containing the particles showed higher hardness values, possibly influencing the sensory perception of the assessors. The encapsulation of extracts from residues of pinhão consumption may contribute to the development of food products with high added value and bioactive properties that can bring benefits to the health of consumers. Thus, it becomes important to optimize food formulations procedures aiming to improve sensorial perception.
Funding This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior-Brasil (CAPES)-Finance Code 001. Fernanda V. Leimann (process 039/2019) thanks to Fundação Araucária (CP 15/2017-Programa de Bolsas de Produtividade em Pesquisa e Desenvolvimento Tecnológico) and to CNPq (process number 421541/2018-0, Chamada Universal MCTIC/CNPq n.º 28/2018). The authors are also grateful to the Foundation for Science and Technology (FCT, Portugal) for financial support through national funds FCT/MCTES to CIMO (UIDB/00690/2020); national funding by F.C.T. and P.I., through the institutional scientific employment program-contract for M.I.D and L.B. contracts. Also, to FEDER-Interreg España-Portugal programme for financial support through the project TRANSCoLAB 0612_TRANS_CO_LAB_2_P.

Data Availability
The authors declare that all data supporting the findings of this study are available within the article and its supplementary information file.