In this study, Comercializadora Tint Café S.A.S, Villeta, Colombia, supplied coffee pulp samples of both colors (red and yellow). β-Carotene of 97.0% (w/w) purity was used from Sigma-Aldrich (product number 7235-40-7). The chemicals 2,2-diphenyl-1-picrylhydrazyl (DPPH), 2,2′-Azinobis (3-ethylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS), and anhydrous caffeine (C₈H₁₀N₄O₂), were also used from Sigma-Aldrich. Absolute ethanol (C₂H₅OH, 99.8% v/v), ethyl acetate (C₄H₈O₂, 99.5% v/v), acetone (C₃H₆O, 99.5% v/v), hydrochloric acid (HCl, 37.0% v/v), and sodium hydroxide (NaOH, 98.0% w/w) were obtained from Panreac (Gmbh). For the determination of total dietary fiber (TDF), insoluble dietary fiber (IDF), and soluble dietary fiber (SDF) heat-stable α-amylase, protease, and amyloglycosidase (Sigma® Life Science) were utilized for enzymatic digestion. Anhydrous disodium hydrogen phosphate (Na₂HPO₄, 99.0% w/w) and anhydrous sodium dihydrogen phosphate (NaH₂PO₄, 99.0% w/w) from Panreac were used for preparing the phosphate buffer in TDF, IDF, and SDF quantification. For the extract dilution a Don Olio (Colombia) sunflower oil was used. As wall materials, arabic gum from CIACOMEQ S.A.S. and maltodextrin from Cimpa S.A.S. were used.

The coffee pulp (red and yellow) was washed, disinfected (100 mg/L sodium hypochlorite), and stored at −18°C. The pulp was divided into two parts, one was dried by forced convection and the other one was lyophilized. The forced convection dehydration process took place in a tray dryer (UOP8-MKIII Armfield, UK) at 40°C for 24 h under atmospheric pressure. For freeze-drying, the samples were frozen at −80°C in an ultra-freezer (Haier Biomedical, UK) prior to the start of the lyophilization process at −40°C for 72 h and 0.086 bar in a freeze-dryer (FreeZone 6 Liter Benchtop, Labconco Corp., KS). Both the lyophilized and forced convection dehydrated samples were ground in a rotary blade mill (Pulverisette 19, Fritsch Comp, Germany). The ground samples were sieved using Tyler series flatbed sieves (Tamiz Standard, Colombia). The particle size fraction for extraction experimental design was between 250 and 425 μm.

The proximate composition analyses were conducted using the techniques described in the AOAC (38) on pulp that was dehydrated by forced convection. Moisture content was determined by oven-drying at 105°C (AOAC 930.15); total fat was determined using the Soxhlet method with hexane (AOAC 991.36); crude protein was quantified using the microKjeldahl procedure (AOAC 960.52); ash content was determined by incineration in a muffle furnace at 550°C (AOAC 942.05), and total dietary fiber (TDF), insoluble dietary fiber (IDF), and soluble dietary fiber (SDF) were quantified using the enzymatic-gravimetric method (AOAC 985.29). The caffeine content was determined by high performance liquid chromatography (HPLC), following the method described by the International Standard ISO 20481:2008 (International Organization for Standardization ISO, 2008). For the calibration curve, anhydrous caffeine solutions with concentrations between 1 and 100 mg/mL (five data points, R² = 0.9999) were prepared. All mentioned determinations were performed for each pulp color in triplicate.

The carotenoid compounds present in the pulp coffee samples were extracted by conventional extraction (CE) following the methodology described by Biswas et al. (39), with some modifications. One gram of each pulp color and each of the two drying methods was weighed and placed in a 50 mL-Falcon tube. Successive extractions (until the extract was colorless) were conducted using 10 mL of cold acetone and a Dispermat^® homogenizer (DISPERMAT^® LC Dissolver, Vma-Getzmann Gmbh, Germany) with a 20 mm light disc set at 10,000 rpm for 5 min. Then, the extracted material was centrifuged (CL30R, Thermo Fisher Scientific, MA, USA) at 4,300 rpm for 8 min at 0°C, filtered with qualitative filter paper (65 g/m²), and collected in another Falcon tube. The extract was then adjusted to a 50 mL with cold acetone and the absorbance was measured at 454 nm using an UV–vis spectrophotometer (Thermo Scientific Genesys 20^®, USA). Total carotenoids content (TCC) was calculated using a calibration curve of all-trans-β-carotene (10–80 μg/mL, eight data points, R² = 0.9975). Acetone was used as a blank. Results were expressed as mg all-trans-β-carotene equivalents/g sample DW. The procedures were performed in triplicate.

Extraction was performed using a mixture of solvents containing butylated hydroxyanisol (BHA). For this, 50 mg of each sample (freeze-dried red or yellow pulp) was extracted with 1 mL acetone:methanol (2:1, v/v) containing 0.5% (w/v) BHA and 0.5 mL hexane during 20 min at 0°C in an ultrasonic bath (Branson 1510, Danbury, CT, USA). The mixture was centrifuged (14,000 rpm; 10 min; 4°C), the supernatant was collected and stored at 4°C. The pellet was re-extracted two times. The three extractions yielded more than 95% carotenoids extraction as evaluated at 450 nm. The three extracts were pulled together, and a liquid–liquid partitioning was done with 4.5 mL aqueous 1 M sodium chloride. The organic layer was dried under nitrogen gas. The dried extract was re-disolved in 1 mL acetone:methanol (2:1, v/v) containing 0.5% (w/v) BHA and analyzed by UHPLC. Chromatographic analyses were performed on a UHPLC 3000 DIONEX (Sunnyvale, CA, USA) equipped with a pump, auto-sampler, and a diode array detector (DAD). The extract (5 μL) was injected onto an Aquity UPLC BEH Shield RP18 column (150 mm × 2.1 mm, 1.7 μm, Waters, Milford, MA, USA). Chromatographic separation was carried out at 20°C. Elution was developed by a gradient of three solvent mixtures, solvent A: acetonitrile/triethylamine (100:0.25, v/v), solvent B: ethyl acetate/triethylamine (100:0.25, v/v), and solvent C: acetonitrile/water/triethylamine (50:50:0.25, v/v/v). The composition was modified during the chromatographic program by changing the proportions of solvents A and B, while solvent C was kept constant at 5%. The elution program was: 0–1 min (0–0% B), 1–13 min (0–30% B), 13–15.5 min (30–30% B), 15.5–17.5 min (30–0% B), 17.5–27 min (0–0% B). The flow rate was set at 0.250 mL/min. Detection was performed at 450, 470, 477, and 503 nm (40).

Identification was carried out by analyzing the order of elution, comparison with absorption spectra (λ_max), and chromatographic characteristics of authentic and similar carotenoids, as well as reference parameters in the literature such as spectral fine structure (%III/II) and maximum cis intensity (II/B) (41). The % III/II ratio represents the correlation between the peak heights of the longest wavelength absorption band (III) and the middle absorption band (II). This calculation is performed by taking the minimum peak between the two bands as the baseline and multiplying it by 100. Additionally, the II/B ratio was determined by dividing the peak heights of the middle absorption band (II) by that of the cis peak, which appears in the UV region. Moreover, authentic standards of all-trans-lutein, all-trans-zeaxanthin, and all-trans-β-carotene were injected separately and co-injected with the extracts for comparative analysis. Each carotenoid was quantified using external analytical curves with six points (in duplicate) for all-trans-β-carotene (1–100 μg/mL, R² = 0.9999). Only peaks with area higher than 2% of the total area as measured at 450 nm and those were resolution did not compromised tentative peak identification and quantification were analyzed. The limits of detection (LOD) and quantification (LOQ) were determined as three and ten times the standard deviation of the noise (blank), respectively. The LOD was 0.3 μg/mL and the LOQ was 1.0 μg/mL. Results were reported as the average of the four replicates performed (±standard deviation) in μg all-trans-β-carotene equivalents/g dry sample. No correction was made for differences in molar absorption coefficients or molecular weight.

A factorial experimental design (2³) was proposed to optimize the UAE of carotenoids from coffee pulp. The effect of color pulp [red (−), yellow (+)], biomass:solvent ratio [1:20 (−), 1:40 (+) w/v], and the solvent composition, ethanol:ethyl acetate ratio [20:80 (−), 80:20 (+) %v/v] on (i) the content of carotenoids extracted from the biomass (CEB) [mg all-trans-β-carotene equivalents/g biomass dry weight], (ii) the extraction yield (EY) [%, g extract/100 g biomass dry weight], and (iii) the concentration of carotenoids in the extract (CCE) [mg all-trans-β-carotene equivalents/g extract dry weight], was investigated. The experimental design was developed in a setup with a tip sonicator (Sonics Materials VCX-750-220, Thermo Fisher Scientific, MA, USA), which included a 13 mm probe, operating at 35% amplitude and a pulse ratio of 20:40 s active-rest time for 30 min. Freeze-dried pulp coffee samples and solvent mixtures were placed in a light-protected Schott container and kept in an ice bath during the procedure. The extraction process was carried out in triplicate for each condition.

After extraction, the raw extract was centrifuged at 4,300 rpm for 8 min at 0°C. The resulting supernatant was filtered with 65 g/m² qualitative filter paper and adjusted to 50 mL or 100 mL according to biomass:solvent ratio. The absorbance of each solution was measured at 454 nm and the CEB was calculated. For the carotenoids quantification, two calibration curves of all-trans-β-carotene were carried out: one for the mixture solvent ethanol:ethyl acetate 80:20% (v/v) (10–80 μg/mL, eight data points, R² = 0.9992) and another one for ethanol:ethyl acetate ratio 20:80% (v/v) (10–80 μg/mL, eight data points, R² = 0.9930). To determine the EY, the residue left after removing the solvent in a rotary evaporator (Yamato, RE801, USA) was weighed from 35 mL of extract. The CCE response variable was calculated as the ratio between the CEB and EY.

The extract with the highest CCE value was evaluated on its in vitro antioxidant activity by the DPPH and the ABTS radical methods.

The concentrated extract with the highest CCE value was encapsulated following the methodology described by López-Bermudez et al. (45). For this purpose, the extract was re-diluted in a known volume (30 mL) of sunflower oil. The polymeric matrix was prepared using gum arabic and maltodextrin in water with a 1:3 (w/w) ratio wall materials:water. Both substances are classified in the Generally Recognized as Safe (GRAS) category by the U.S. Food and Drug Administration (FDA) (46, 47). The wall materials were hydrated separately before mixing using a mechanical stirrer (Hei-Torque, Heildolph, Germany) at 1,200 rpm for 5 min. The emulsion (O/W) was prepared by adding the enriched oil to the polymeric matrix in a ratio 2:1 (w/w) wall material:oil and continuing the stirring for 10 min more. Finally, homogenization was performed in a rotor-stator (SilentCrusher M, Heidolph, Germany) at 18,000 rpm for 5 min. The emulsion was fed to the spray dryer (B290, Buchi AG, Switzerland) with an inlet flow rate of 6.25 mL/min, using an inlet and outlet temperature of 140 and 110°C, respectively. A 0.7 mm diameter co-current spray nozzle with an air inlet of 473 L/h was used. Subsequently, the sample was stored in a desiccator at room temperature avoiding contact with light. The process was done in triplicate.

To prepare the yogurt, 1 L of pasteurized whole milk from Cooperativa Colanta, Colombia was used and maintained at 42 ± 1°C throughout the fermentation process. The milk with a 250 mg capsule with 1 × 10⁹ CFU of starter culture of Bulgarian yogurt (Genesis Laboratories Ltd., Bulgaria) was allowed to reach a pH of 4.5 ± 0.1, and then it was cooled to 4°C. The encapsulated enriched-carotenoids extract was mixed with cooled yogurt in a proportion of 9.5% (w/v) to classify this dairy product as an “excellent source” of vitamin A according to Colombian regulation ‘Resolution 810 of 2021’ (50). This amount of encapsulated is the minimum necessary considering that the standard requires 15% of 800 μg RAE. The mixture was made in a Schoot flask with magnetic stirring for 20 min at 1,200 rpm. The mixing was done in triplicate for a fortified yogurt (FY), and a stirred control (SC) without encapsulated. Additionally, to consider the effect of the stirring on the rheological properties, a sample of the yogurt before agitation was taken as an unstirred control (UC).

The results were analyzed using the Minitab 18^® statistical program with a confidence level of 95%. For proximate analysis, the effect of pulp color was evaluated using the t-student test. An analysis of variance (ANOVA) was performed on the data of total carotenoids extracted by CE and UAE. Statistical significance was determined by applying the Tukey’s test for the CEB response variable from CE and for the three UAE response variables. Similarly, Tukey’s test was used to determine the statistical significance of the response variables related to the rheological and stability properties of yogurt.

Table 1 presents the results of the proximate analysis for red and yellow dehydrated pulp, showing differences in certain parameters between the two colors. Notably, the percentage of TDF was significantly higher in yellow pulp (68.52 ± 3.72%) as compared to red pulp (45.17 ± 1.07%) (p < 0.05). TDF and IDF values differed significantly (p < 0.05) between the two pulp colors. Despite these differences, the fiber values are close to those reported by Mindarti et al. (51), for Arabica brown pulp flour (63.16 ± 0.56% DW, 58.38 ± 0.23% DW, and 4.78 ± 0.33% DW of TDF, IDF and SDF respectively). These values are closer to those reported in this study for yellow pulp (68.52 ± 3.72% DW, 62.31 ± 3.14% DW, and 6.21 ± 0.58% DW of TDF, IDF and SDF respectively). Accordingly, the potential use of the pulp to take advantage of its fiber content is evident, both to produce foods that help the digestive system to function properly and to produce other food additives (51).

The fat content of the red and yellow pulp in this study was found to be independent of the color (p > 0.05), with values of 3.91 ± 0.20% and 3.99 ± 0.14%, respectively. Such values represent differences of less than 20% as compared to 4.28 ± 0.01% and 4.10 ± 0.26% reported in other studies of coffee pulp (20, 52). In terms of ash content, the red and yellow pulp had values of 4.88 ± 0.01% and 4.80 ± 0.01%, respectively. These values are roughly 50% lower than the values reported by Patil et al. (53), for coffee pulp (9.6 ± 2.2%). Patil et al. (53) also revealed that pulp coffee contained high concentrations of minerals such as potassium (410 mg/100 g DW), calcium (97.8 mg/100 g DW), and magnesium (24.5 mg/100 g DW).

The present study determined a caffeine concentration of 8.81 ± 0.18 and 9.06 ± 0.17 mg/g DW for the red and yellow coffee pulps, respectively. Several studies have analyzed the caffeine content in coffee pulp, with varying results depending on the coffee species, the processing method, and the region of cultivation. For example, a study conducted in Mexico (54), reported caffeine levels of 18.60 ± 0.31 mg/g in fresh coffee pulp. Furthermore, another study investigated the caffeine content in six coffee cherry pulps from several Coffea arabica cultivars (Bourbon, Caturra, mixed varieties, Catuai, and Maragogype), origin (Salvador, Honduras, and Congo), and processing methods (wet and semi-dry depulping) (55). The results showed caffeine concentrations ranging from 3.4 to 6.8 mg/g in dried coffee pulp samples. The Caturra sample had a caffeine content of 4.6 mg/g according to Heeger et al. (55), which is half the value found in the present study for the same variety.

Data on TCC obtained by CE and UAE methods using acetone are shown in Table 2. A comparison was made between the two drying methods (forced convention and freeze-dried) to determine the most appropriate approach for ultrasound-assisted extraction. Thus, a TCC of 1.06 ± 0.02 mg all-trans-β-carotene equivalents/g biomass DW for forced convection drying and 1.58 ± 0.08 mg all-trans-β-carotene equivalents/g biomass DW for freeze-drying was obtained for the red pulp; this implies an increase in carotenoids of 33% when the pulp is freeze-dried. For yellow pulp, a CEB of 1.19 ± 0.09 mg all-trans-β-carotene equivalents/g biomass DW was obtained for forced convection drying and 1.646 ± 0.004 mg all-trans-β-carotene equivalents/g sample DW for freeze-drying; this implies an increase in carotenoids of 28% for freeze-drying. Based on these comparisons, it is possible to define that UAE extractions should be carried out using freeze-dried pulp. For example Coelho et al. (56), extracted carotenoids from freeze-dried cashew apple using different solvents, including acetone, and achieved a concentration of 0.102 ± 0.017 mg β-carotene equivalents/g sample. For their part, Sánchez-Camargo et al. (57) performed an extraction of total carotenoids from mango peel dried by forced convection and freeze-drying using cold acetone and vortex agitation. They found a concentration similar in order of magnitude and value (1.22 ± 0.02 mg all-trans-β-carotene/g DW for forced convection and 2.01 ± 0.05 mg all-trans-β-carotene/g DW for freeze-dried) to those found in the present study. Furthermore, López-Bermúdez et al. (45) extracted carotenoids from freeze-dried tomato peel using acetone, which was able to extract 3.06 ± 0.33 mg β-carotene equivalents/g DW. According to these results, coffee pulp has a large potential to be a carotenoid source that can be used in industry and compete against currently established raw materials.

The chromatographic profile at a wavelength of 450 nm of the extracts is shown in Figure 2. Tentatively, 12 carotenoids were identified in the samples, using their UV–Vis spectra, reference parameters [spectral fine structure (III/II) % and maximum cis intensity (II/B)], and in some cases comparison with authentic standards (Table 3). The coffee pulp residue samples showed a wide variety and concentration of xanthophyll esters. These esters have much longer retention times than their precursor xanthophylls because the resulting molecule contains a much larger hydrophobic (nonpolar) part and therefore are more retained in the nonpolar stationary phase. Additionally, based on the UV–Vis spectra of these esters, it was determined that they are all lutein derivatives (16, 21). The studied parameters, however, do not provide structural information beyond their precursor xanthophyll. To completely elucidate the structure, it would be necessary to perform chromatographic studies on alkaline hydrolysis products and analysis of the non-hydrolyzed extract using RP-UHPLC-APCI-MS (58). Yellow coffee pulp residues exhibited higher concentrations of each analyte. Both red and yellow coffee samples were mainly composed by lutein isomers (c1, c3) and lutein esters (c7–c12) at high concentrations (579.1 ± 21.3 and 808.0 ± 18.3 μg all-trans-β-carotene equivalents/g sample for red and yellow coffee, respectively). Additionally, significant amounts of zeaxanthin (c2), β-carotene isomers (c6, c7, c8), and α-carotene (c5) were found, although the presence of violaxanthin and neoxanthin was also expected based on previous reports (21, 59–61).

Table 4 displays the concentrations obtained from the experimental UAE design. The ANOVA is found in Supplementary Material 1. It shows the model fit for the significant terms of each of the response variables. The R² values for EY and CCE were high (90.59 and 89.32%, respectively) and low for CEB (33.36%). In the last case, only one term was significant (term A: BM:S), which could lead to this result. Similarly, the model had no lack-of-fit (p > 0.05), for the three response variables. The Pareto plot (Figure 3) shows with a significance level of 5%, the terms (variables or their interactions) that significantly impacted the CEB, EY, and CCE response variables. Some terms positively or negatively affected the response variables; these effects are also shown in the Pareto plot. The results indicate that, for both red and yellow pulp, the highest level of BM:S ratio (this is 1:40 w/v ratio) produced a higher CEB value. The relationship with the variable E:A ratio is not so clear, in fact, it was not significant according to the Pareto plot. The highest CEB was obtained for the extraction made with yellow pulp, with a BM:S ratio of 1:40 (w/v) and an E:A ratio of 20:80 (v/v) (2.30 ± 0.14 mg all-trans-β-carotene equivalents/g biomass DW); however, the statistical analysis shows that there were no significant differences in the results found for that response variable. For the response variable EY, the highest percent yields were found when the E:A ratio was more concentrated for ethanol; in addition, higher EY values were also obtained when the BM:S ratio level was higher (1:40 w/v). The highest EY (7.30 ± 0.51%) was obtained for the extraction made with red pulp, with a BM:S ratio of 1:40 w/v and an E:A ratio of 80:20 v/v. Lastly, CCE values were higher for the lowest BM:S ratio level (1:20 w/v) and for more concentrated E:A ratios in ethyl acetate (20:80, etanol:ethyl acetate, v/v). The best extraction configuration was taken based on this response variable and was achieved with the yellow pulp, a BM:S ratio of 1:20 w/v and an E:A ratio of 20:80 v/v to obtain 51.28 ± 4.52 mg all-trans-β-carotene equivalents/g extract DW. In comparison, Ordóñez-Santos et al. (62) used sunflower oil in UAE of carotenoids from mandarin peel and obtained 1.40 ± 0.02 mg β-carotene equivalents/g of sample, which is similar to the results obtained in this study. Likewise, Ordóñez-Santos et al. (27) performed carotenoid extractions from peach palm fruit peel using sunflower oil as solvent and obtained a maximum concentration of 1.63 ± 0.04 mg total carotenoids/g peel DW. According to the classification criteria proposed by Britton et al. (63), the coffee pulp used in this study is classified as “very high” for having a carotenoid concentration greater than 2 mg/100 g (230 ± 14 mg/100 g maximum for yellow pulp and 218 ± 2 mg/100 g maximum for red pulp).

Carotenoid extraction is generally higher when obtained by UAE than when obtained by the conventional method. For example, Coelho et al. (56) used UAE to extract carotenoids from red and yellow cashew apple shells using a mixture of acetone and methanol as solvent. They found in 11 of their 15 experiments a higher concentration of carotenoids when extracted by UAE as compared to conventional extraction. Thus, according to the concentrations obtained by UAE, this method proves to be striking for the extraction of carotenoids from plant material. This method has the potential to replace conventional extraction methods and, together with the use of environmentally friendly solvents, decrease the use of environmentally harmful methods.

Coffee pulp is a rich source of antioxidants and other bioactive compounds that have potential health benefits. Some compounds such chlorogenic acid isomers, anthocyanins, and carotenoids have been identified in Arabica coffee varieties with different external fruit color (16). As demonstrated by Geremu et al. (64), nevertheless, the type of solvent extraction exerts a significant influence on the type of bioactive compounds extracted from coffee pulp, and consequently on the antioxidant activity of the extracts. In this study, the optimal extract obtained may contain medium to non-polar compounds (such as β-carotene), since an 80:20 v/v solvent mixture of ethyl acetate:ethanol was used. The concentration (in μg/mL) of the UAE extract able to inhibit 50% of the DPPH free radical (EC₅₀ value) was assessed. The results showed EC₅₀ values of 19.34 ± 1.14 μg/mL. Similar values were found by Manasa et al. (11), in Coffea canephora (Robusta variety grown in India) pulp. Ice cold ethanol was used as solvent because the aim of that work was to extract polyphenols from coffee pulp, having an effective concentration (EC₅₀) value of 12.75 μg/mL. In our study, the EC₅₀ value of enriched-carotenoids extract was determined for the first time, and hence no literature data are available to compare our results with information from other literature. Additionally, the TEAC of the optimal UAE extract was measured using ABTS as radical and a value of 721.08 ± 51.23 μmol Trolox equivalent/g extract was achieved. In the same way, Thaiphanit et al. (20) reported comparable results for Arabica (710.43 ± 32.76 μmol/g) and Robusta (600.46 ± 25.07 μmol/g) coffee pulp extracts, which were obtained using 95% ethanol as the extraction solvent. Meanwhile, Corrêa-Filho et al. (65) measured the TEAC of neat β-carotene (before to encapsulate it) and they found values of 2,350 μmol Trolox equivalent/g β-carotene. This means that 3.26 g of the enriched-carotenoids coffee pulp extract obtained in this study, could have the antioxidant capacity equivalent to 1 g of β-carotene.

The extract obtained by UAE with the highest CCE was chosen for encapsulation. The drying yield and encapsulation efficiency, as well as the powder characterization, are presented in Table 5, which can be directly compared with the study by López-Bermúdez et al. (45). The main difference between the two studies was found in the drying efficiency, which was 93.84 ± 1.08% in the present work, about 30% higher than that one achieved by López-Bermúdez et al. (45) (67.3 ± 0.5%). The main reason for encapsulating the extract is to protect it from degradation due to contact of the carotenoids with oxygen, light, and high temperature, among others. As indicated by Corrêa-Filho et al. (65), however, in the encapsulation process, the higher the air outlet temperature the higher the loss of these compounds. This fact is evidenced in the encapsulation efficiency values. In the present study, the encapsulation efficiency was 45.57 ± 4.03% (w/w), 21.5% lower than the value achieved by López-Bermúdez et al. (45) (58.1 ± 0.8%, w/w) who used an air outlet temperature of 100°C. In the same way, Guadarrama-Lezama et al. (66) achieved even higher encapsulation efficiencies (83.9 ± 1.5%) for carotenoids dissolved in sunflower oil as solvent for the extraction of carotenoids from paprika (Capsicum annuum L.) and used an air outlet temperature of 70°C. In contrast, this reduction in the air outlet temperature increases the moisture percentage of the sample. The moisture percentage achieved in this work (3.03 ± 0.26%) is close to that reported by López-Bermúdez et al. (45) (2.5 ± 0.1%), however Guadarrama-Lezama et al. (66), reports a higher moisture value (5.0 ± 0.1%) which can be explained by the lower outlet temperature that they used. The water activity reported in this study (0.25 ± 0.01) was higher than that achieved in other studies. In the case of Guadarrama-Lezama et al. (66), they found an a_w of 0.18 ± 0.01 for encapsulation with sunflower oil. For their part, Zhang et al. (67) reported values between 0.15 ± 0.01 and 0.22 ± 0.04, which are closer to those reported herein. Thus, these lows a_w values are remarkable since the enzymatic action is reduced and bacterial growth in the powder is avoided. The dissolution rate in this study (153.5 ± 2.8 s) was similar as compared to that achieved by López-Bermúdez et al. (45) (188.68 ± 5.51 s). These values were, however, higher (between 75 and 86%) than those reported by Zhang et al. (67), who obtained values between 25.5 ± 0.7 and 47.5 ± 3.5 s. Similarly, the same authors report tapped density values between 0.628 ± 0.018 and 0.689 ± 0.014 g/mL which were between 20 and 27% higher than that reported in this study (0.50 ± 0.01 g/mL), which could offer a potential advantage in terms of transport efficiency. In relation to provitamin A activity, a value of 12.57 ± 0.60 μg RAE/g encapsulated was calculated. Considering that 9.5 g of encapsulated were added to 100 mL of yogurt, a value of 119 μg RAE/100 mL of yogurt is expected. Such value is close to 120 μg RAE/100 mL of food, which according to the Colombian legislation is needed to declare a food “Excellent source” of vitamin A. Losing about 50% of the carotenoids in the encapsulation process can be considered a high cost. However, encapsulation of carotenoid-rich extracts allows new applications of the extract. One of the most powerful applications is its new ability to dissolve in aqueous matrices because when the extract is only dissolved in oil, this dissolution is more difficult. This ability facilitates the formulation of new foods or the enrichment of existing foods and, therefore, can have a positive impact on consumer health.