Lyophilized Powder of Hibiscus sabdariffa (Roselle) Extracts using Gum Arabic and Maltodextrin as Carrier Agents

Freeze-drying is a process for drying foods without heat application. The physical, chemical and sensory properties of the food remain without significant changes. In this work, maltodextrin (MD), gum arabic (GA), and a blend of MD:GA (60:40) were used as encapsulating agents of Roselle (Hibiscus sabdariffa) calyces extracts. Lyophilized powders were obtained at different concentrations of encapsulating agent (0, 3, 5, and 10%, w/w). Powders were analyzed in yield and physicochemical (average size diameter (d50), moisture content, water activity (aw), bulk and compacted densities, and color), and antioxidants (anthocyanins content, total phenolic compounds, antioxidant capacity) characteristics. The yields of freeze-drying powders from different encapsulating agents ranged 82 to 95%. The average diameter (d50) was higher for powders without gum (139.5±25.6 μm) than for powders with encapsulating agents (35 to 89 μm). The moisture content and aw of the powders were in the ranges 5.3-11.2% and 0.20-0.29, respectively. The value of the red (a*) color parameter of all powders was 37.0±2.8, decreasing as increasing the gums concentration. Powders with 3% GA and MD showed the highest amount of anthocyanins: 560.93±10.13 and 543.46±15.68 mg/100 g of powder, respectively. The highest total phenolic compounds content was observed in the powder with the 3% MD:GA blend (4,705.70±140.54 mg/100 g of powder). Powder with 3% MD showed the highest antioxidant capacity (1,766.30±31.15 mg of Trolox equivalents/100 g powder).

al., 2017), the equipment needed is readily available, options for encapsulating materials are many, the encapsulation process is efficient, the final product is stable, and there is the potential for continuous large-scale production (Santos and Meireles, 2010). The parameters that have great influence in the spray drying process are nozzle geometry, viscosity of the feeding solution, and the inlet and outlet air temperatures (Munin and Edwards-Lé vy, 2011;Gharsallaoui et al., 2007). Commercially, this technique has been used to encapsulate numerous materials, including flavor agents, fats, oils, vitamins, minerals, microorganisms, enzymes, sweeteners, and colorants (Wijaya et al., 2011).
Freeze drying (lyophilization) is a preservation method where water is evaporated or eliminated by the application of vacuum and low temperatures in the processing system. It consists of evaporating water from a frozen material (sublimation) without passing through the liquid state. Lyofilization is a way to dry heat labile chemicals or food products. This preservation procedure is used for the production of milk for infants, soups, coffee, infusions, the commercial preparation of antibiotics, some vaccines, many foods and flavor products. The main advantage of this method is that products being dehydrated in a frozen state retain their shape, aroma, flavor, vitamins, nutritional value, and active ingredients. In addition, it could be appliable to most foods, facilitating their productions on site with minimal transport costs. However, it is a prolonged and expensive method (Santos and Meireles, 2010;Özkan and Bilek, 2014). Nowadays, in addition to the products above mentioned, the process is being used to obtain powders of extracts from parts of plants in order to evaluate their antioxidant, antimicrobial and coloring properties (Rajarajan et al., 2010;Viloria-Matos et al., 2002;Muro et al., 1997).
There is little research on this technique for obtaining powders from Hibiscus sabdariffa extracts. Some studies have shown that powders obtained by freeze drying, using maltodextrin (mainly) as encapsulating agent, have great stability maintaining their antioxidant properties; therefore, being these powders an option for use them as colorants and flavorings for foods (Duangmal et al., 2004;Selim et al., 2008).
The aim of this work was to evaluate the effect of maltodextrin and gum arabic as encapsulating agents for obtaining lyophilized powders of extracts of Hibiscus sabdariffa calyces.

Material
Calyces from creole Hibiscus sabdariffa long red variety, grown in Chiautla de Tapia, Puebla, Mexico, were used. The Roselle calyces powder (RCP) was obtained using a Veyco stainless steel mill model MPV 100 (Mexico) with a mesh of 0.5 mm.

Average Particle Size
The average particle size was carried out using a Microtac S3500 particle size analyzer (Microtac Inc., Largo, FL, USA) in a range of 0.25 to 2,800 μm. The analysis was carried out in triplicate using approximately 60 mg of Roselle calyces powder (RCP) or 40 mg of lyophilized powder (LP). Curves of granulometry, accumulative retained weight, and average diameter (d 50 ) were obtained (O'Hagan et al., 2005; Cid-Ortega and Guerrero-Beltrá n, 2020).

Roselle Extracts Concentrates-gums (RECG)
A 3x3 factorial design was used in this study. Three types of gums or blends (gum arabic powder from Roller Dry (Central de Drogas S.A. de C.V., State of Mexico, Mexico), maltodextrin ED: 9-14 (CP Ingredientes S.A. de C.V., Guadalajara, Mexico), and a blend of maltodextrin:gum arabic in a proportion of 60:40)) and three concentrations of gums (3, 5, and 10% w/w) were used. The gum was added to each free-ethanol extract and stirred for 15 min at room temperature (22 ± 2°C) to obtain the RECG. The RECGs were placed into 250 mL flasks, covered with aluminum foil and stored in refrigeration (4°C) until drying. A control was also prepared (extract without ethanol and gum). Total soluble solids, density and viscosity were analyzed.

Lyophilization
The RECGs were placed in Petri dishes (13.76 ± 0.07 cm in diameter), covered with aluminum foil and frozen for a minimum of 72 hours in a CHTC-16E horizontal freezer (Torrey, Mexico) at -26 ± 0.5º C. The frozen RECG were lyophilized in a LabConco freeze drier (LabConco Corp. Kansas, City, USA) at 20°C and 0.005 to 0.01 mmHg of vacuum for 72 hours. The lyophilized RECG (Roselle powders = RP) were weighed, pulverized, placed in amber pharmaceutical jars, sealed with plastic wrap and capped. RP were stored in a desiccator with silica at room temperature (22 ± 2º C).

Physicochemical Properties of Extracts
Total soluble solids (TSS) They were measured according to the 932.14C AOAC (1995) method. A manual Atago Master-M model refractometer (Atago Co. LTD., Tokyo, Japan) with a scale of 0-32°Bx was used. The readings were corrected at 20°C using the values established in the reference tables of the AOAC (1995).

Absolut viscosity (μ)
A 350-159I Cannon Fenske capillary viscometer (Cannon Instrument Co., State College, PA, USA) was used. The kinematic viscosity was obtained by multiplying the flow time (seconds) of 6.6 mL of extract at 40°C by the constant of the viscometer (0.4754 mm 2 /s 2 ) at the same temperature. For calculating the absolute viscosity (), Eq.
(2) was used (Cannon Instrument Company, 2000): (2) where  s (g/mL) is the density of the extract and  c (mm 2 /s = cSt) is the kinematic viscosity.

Antioxidant Characteristics
Total monomeric anthocyanins (TMA) The TMAs determination was carried out according to the Lee et al. (2005) method with some modifications according to Cid-Ortega and Guerrero-Beltrá n (2020). Briefly, 0.5 mL of extract or 100 mg of powder were used, made up to 10 mL with distilled water and totally homogenized with a Vortex (2900 to 3000 rpm) for 5 min; these are the Roselle extract solution (RES) or the powder solution (PS). Then, 1 mL of solution was taken and mixed with 4 mL of buffer pH 1 or pH 4.5. The solutions were allowed to stand for 30 minutes at room temperature (22 ± 2°C). The absorbances were measured at 520 and 700 nm in a Cary 100 UV-visible spectrophotometer (Varian Inc., Palo Alto, CA, USA). A blank with distilled water was used for standardizing the equipment. Results were reported as equivalents of cyanidin-3-glucoside (C-3-G) per 100 mL of extract or per 100 g of powder according to Eq. (3). ( where TMA is the total monomeric anthocyanins content (mg/100 mL or mg/100 g); A = (A 520nm -A 700nm ) pH=1.0 -(A 520nm -A 700nm ) pH=4.5 ; MW is the molecular weight of cyanidin-3-glucoside (449.2 g/mol); FD is the dilution factor; L is the cell pathway (1 cm);  is the molar extinction coefficient of cyanidin-3-glucoside (26,900 L/mol*cm); 100 is the conversion factor of mg/mL or mg/g to obtain mg/100 mL or mg/100 g, respectively.
Total phenolic compounds (TPC) They were determined by the method reported by Singleton and Rossi (1965) method with some modifications according to Cid-Ortega and Guerrero-Beltrá n (2020). Briefly, 3 mL of distilled water, 150 μL of Roselle extract solution (RES) or 100 μL of powder solution (PS) were placed in test tubes (covered with aluminum foil), separately, mixed with a Vortex (2,900 at 3,000 rpm) for 5 min. Then, 250 μL of Folin-Ciocalteu reagent were added, completely mixed and left for 8 minutes in the dark. Finally, 750 μL of 20% Na 2 CO 3 was added and made up to 5 mL with distilled water. Mixtures were left for 2 hours at room temperature (21 ± 1°C) in the dark. The absorbances were then measured at 765 nm in a Cary 100 UV-visible spectrophotometer (Varian Inc., Palo Alto, CA, USA). A standard curve was prepared in duplicate at different concentrations of gallic acid (98.5%, Sigma): 0 -66.4 μg. Abs = 17.50 ± 2.36 (1/mg gallic acid) X (mg gallic acid) + 0.018 ± 0.011 (R 2 = 0.998 ± 0.001). Total phenolic compounds (TPC) were reported as gallic acid equivalents (GAE) per 100 mL of extract or 100 g of powder according to Eq. (4). (4) where A is the absorbance of the sample, b is the intercept, m is the slope and DF is the dilution factor.

Antioxidant capacity (AC)
The DPPH (1,1-diphenyl-1-picrylhydrazyl) method (Brand-Williams et al., 1995) was used with some modifications (Cid-Ortega and Guerrero-Beltrá n, 2020). Briefly, 2 mL of RES or 1.6 mL PS were taken and diluted with ethanol (99.5%) to make up 10 mL in a volumetric flask, mixed with a Vortex (2,900 to 3,000 rpm) for 5 min and then filtered (twice) through Whatman paper No. 5. From the filtrates, 1 mL was taken and mixed, in a test tube (covered with aluminum foil), with 1 mL of ethanol (99.5%) and 2 mL of DPPH solution (7.8 ± 0.2 mg in 200 mL of 99.5% ethanol), perfectly mixed and allowed to stand for 45 min at room temperature (21 ± 2°C) in the dark. Absorbances were measured at 517 nm using a Cary 100 UV-visible spectrophotometer (Varian Inc., Palo Alto, CA, USA). The antioxidant capacity (AC) was calculated as percentage of inhibition according to Eq. (5).

(5)
where Ac is the absorbance of the control and As is the absorbance of the sample. A standard curve was prepared in duplicate with different concentrations of trolox (6-hydroxy-2, 5, 7, 8 tetramethylchrome-2, 97% carboxylic acid, Aldrich) (T): 0-28.7 g. I (%) = 3278.9 ± 195.8 (1/mg T) X (mg T) + 2.9 ± 1.3. R 2 = 0.989 ± 0.007. Results were expressed as trolox equivalents (TE) per 100 mL of extract or per 100 g of powder, according to Eq. (6). (6) where A is the absorbance of the sample, b is the intercept, m is the slope and DF is the dilution factor.

Physicochemical Properties of Powders
Yield (Y).
It was calculated according to the amount of total soluble solids (TSS) in the encapsulated extract and the amount of powder obtained (Fazaeli et al., 2012), according to Eq. (7).

Moisture content
It was determined according to the 934.06 AOAC (2000) method. A Cole Parmer vacuum oven (Chicago, Illinois, USA) was used. The sample (1.0 ± 0.003 g) was dried for 8 hours at 70±1°C at a vacuum pressure of 200 to 220 mmHg.
Water activity (a w ) It was measured using an AQUA-LAB hygrometer model 3TE (Decagon Devices Inc. Pullman, Washington, USA) with internal control of temperature. The equipment was calibrated with distilled water and charcoal (Decagon Devices Inc., 2008). The a w of powder was measured at 25.1 ± 0.06°C.
Bulk density ( b ). It was determined according to the Jumah et al. (2000) method. One gram of powder was placed in a 10 mL graduated cylinder. The sample was tapped ten times (on a polystyrene base) from a height of 15 cm. The bulk density was calculated according to Eq. (9). (8) where W (g) is the weight of the powder and V a (mL) is the volume occupied by the powder.
Tap density ( t ) It was carried out according to the method reported in the Official Mexican Norm NOM-104- STPS-2001(NOM, 2001, with some modifications. One gram of powder was placed in a 10 mL graduated cylinder and topped with a rubber stopper. Then, the sample was tapped from bottom to top for 8 min (time in which volume was no longer changed). The tap density was calculated according to Eq. (9). (9) where W (g) is the weight of the powder and V t (mL) is the compacted volume occupied by the powder.  Silva et al., 2013). The L* (luminosity: black = 0; white = 100), a* (green to red) and b* (yellow to blue) color parameters were measured in the CIELab scale. From these data the purity (color saturation, C = [a 2 + b 2 ] 1/2 ) and hue (H = tan -1 [b/a]) were calculated.

Statistical Analysis
It was carried out by ANOVA with a level of significance of 0.05 using the software MINITAB® version 14.1 (Minitab Inc., 2003). To establish differences between the treatments, a Tukey-Krammer multiple comparison test was used with a P value of 0.05.   Table 3 shows the content of total soluble solids, density and viscosity for all RECG.
Viscosity. The viscosity of the RECG (Table 3) added with GA had the highest viscosities (4.18 ± 2.28 mPa.s) than the extracts with MD (2.15 ± 0.34 mPa.s) or a blend of MD:GA (2.87 ± 0.04 mPa.s, global averages. Regarding the concentration of gum, the RECG with 10% had the highest viscosity compared with 0, 3 and 5% of gums (4.58 ± 2.04, 1.72 ± 0.03, 2.10 ± 0.19 and 2.50 ± 0.47 mPa s, respectively, global averages). Gharsallaoui et al., 2007 had reported that GA showed excellent emulsifying properties due to the fact that its structure contains a fraction of protein (approximately 2%), which gives this property.
Average diameter (d 50 Figure 2 shows some examples of the particle size distribution of powders. All powders had a bimodal behavior. The powders showed heterogeneous particles; therefore, large particles were shaped through the process of agglomeration (Tonon et al., 2011). Comunian et al. (2011) reported a bimodal particle size distribution in powders, obtained from a 5% chlorophyllide solution, with d 50 values between 11.2 to 19.04 μm. The powders were obtained as already mentioned above. On the other hand, Janiszewska and Witrowa-Rajchert (2009) reported d 50 values of 55 and 29 μm for powders obtained with 25% of maltodextrin or 30% of gum arabic for encapsulated rosemary aromas obtained by spray drying.  Table 4. According to the type of gum, the powder of REC had the highest density (0.670 ± 0.017 g/mL) in comparison with the densities of GA (0.562 ± 0.044 g/mL), MD (0.559 ± 0.040 g/mL) and MD:GA (60:40) (0.592 ± 0.045 g/mL) powders (global averages for all powders). About the gum concentration, the REC powder had the highest density compared with the densities of GA (0.567 ± 0.064 g/mL), MD (0.552 ± 0.012 g/mL) and MD:GA (60:40) (0.594 ± 0.033 g/mL) powders. Different results were observed by Fazaeli et al. (2012); they reported decreasing bulk densities (from 0.55 to 0.35 g/mL) when increasing the concentration of maltodextrin (DE 9) (8, 12, and 16%) in powders of blackberry juice obtained by spray drying at different temperatures (110, 130, and 150º C). Tonon et al. (2011) pointed out that the smaller the particle size, the greater the apparent density in powders of aç aí (Euterpe oleracea) juice obtained by spray drying using gum arabic and maltodextrin (10 and 20 DE). Results obtained in this work could be due to the drying conditions and types and concentrations of gums.

Color of Powders
The color properties of the Roselle powders are shown in Table 5. All powders had a pale pink color.

Lightness (L*):
The REC powder had the lowest lightness (28.12 ± 0.48) (therefore, the darkest one) in comparison with the GA (36.39 ± 2.71, global average), MD (36.40 ± 2.94, global average) and MD:GA (60:40) (35.30 ± 4.66, global average) powders. It was observed that, increasing the concentration of gum, powders with gum became clearer: 33.24 ± 2.58, 35.30 ± 1.22 and 39.56 ± 2.65 for 3, 5 and 10% of gum, respectively (global averages). The lowest L* value (29.90 ± 0.060) for powders with gum was observed in the MD:GA (60:40) powder at a concentration of 3% of gum. Ersus and Yurdagel (2007) reported an increase in lightness (L*) of microencapsulated anthocyanin pigments from Daucus carota L. obtained by spray drying when decreasing the dextrose equivalents (10, 20-23, 28-31 DE) of maltodextrin. They also pointed out that the hue was higher for the powders obtained with maltodextrin of 28-31 DE. The authors concluded that the color of the powders became paler when increasing the DE of maltodextrin. Idham et al. (2012) reported L*, a*, and b* color values for anthocyanins from Roselle extracts encapsulated by spray drying using the same gums as in this study. They reported values of 39.3, 43.1, and -0.8 for MD, 45.9, 34.8, and -4.3 for MD:GA (60:40), and 44.9, 30.3, and -6.3 for GA for L*, a*, and b* color parameters, respectively. These values are different than those obtained in this study.

Purity (C):
The REC powder had the lowest purity (32.27 ± 1.41) than that of powders with GA, MD and MD:GA (60:40): 39.65 ± 2.17,39.69 ± 0.39 and 39.57 ± 1.53,respectively (global average). Regarding the concentration of gum, the REC powder showed the lowest purity (p ≤ 0.05) than that of powders with 3, 5 and 10% of gum: 39.52 ± 1.32, 40.80 ± 0.68 and 38.58 ± 1.48, respectively (global average). The value of purity or Chroma is proportional to the amount of color or hue. This could be observed by correlating the color data of the a* color parameter and purity of all powders (Figure 3). Salazar-Gonzá lez et al. (2009) obtained microencapsulated powders of Roselle extracts and mesquite gum at different concentrations (1, 2, 3, 4, and 5% w/v). They reported average values, on the Hunter scale, of 40.3 ± 0.71, 31.93 ± 0.29, 0.28 ± 0.00, and 33.19 ± 0.3 for L, a, H, and C, respectively. The authors concluded that the gum concentration did not have a significant effect on the color parameters.

Color of Solutions of Powders
The color characteristics for the reconstituted Roselle powders are shown in Table 5. All solutions had a transparent red-purple color.
Lightness (L*): It can be seen that, increasing the concentration of gum, the solutions were clearer: 71.26 ± 0. 16, 73.96 ± 2.22, 75.42 ± 0.56 and 79.99 ± 0.79 for 0, 3, 5 and 10% of gum, respectively (global averages). The solutions of powers of REC and 3% MD were the darkest. Salazar-Gonzá lez et al. (2009) reported similar values to this research for the L* and a* color parameters of solutions from reconstituted powders (100 mg/7.5 mL distilled water) of Roselle extracts added with mesquite gum at different concentrations (1, 2, 3, 4 and 5% w/v). Then, the type of gum barely affected lightness; however, as the concentration of gum increases, the solutions became lighter; therefore, its purity decreases (Chroma) and consequently, its red coloration.

Green-red color (a*):
The red color of the solutions did not show significant differences (p > 0.05) about the type of gum; however, the solutions with 10% of gum had the lower red color (20.97 ± 0.98, global average) than powder solutions with 0, 3 and 5% of gum (24.69 ± 0.86, 27.10 ± 4.04 and 26.79 ± 0.46, respectively, global averages). The yellow color parameter of the solutions of powder also decreased as the gum concentration increased (12.58 ± 0.12, 11.84 ± 0.85, 10.74 ± 0.72 and 8.29 ± 0.81 for 0, 3, 5 and 10%, respectively, global averages). Purity (C). The purity (C) of the solutions from different types of gum did not show any significant difference (p > 0.05). About the gum concentration, the lowest purity was observed for the solutions of powders with 10% of gum. No significant differences (p > 0.05) were observed within the solutions from 0, 3 and 5% of gum: 27.71 ± 0.13, 29.61 ± 3.74 and 28.87 ± 0.56, respectively (global averages). .31 mg C-3-G/100 g of dry powder, respectively, global average). TMAs of powders from 10% of gum had lower content of anthocyanins than powders with 3 and 5% of gum. It can be also observed that, when increasing the concentration of gums, the content of anthocyanins decreases which is very probably due to the amount of gum in the powder. Comparable results were observed by Cid-Ortega and Guerrero-Beltrá n (2020) for powders of RECG of H. sabdariffa obtained by spray drying. Total phenolic compounds (TPC). The content of TPCs (Table 6) regarding the type of gum was higher for the REC and MD powders (4,289.40 ± 207.07 and 4,434.20 ± 391.60 mg GAE/100 g powder, respectively) (global averages) in comparison to the TPCs of powders with GA and MD (3442.50 ± 715.63 and 3430.60 ± 540.10 mg GAE/100 g, respectively) (global averages) gums. Regarding the gums concentration, a decrease in the content of TPCs was observed as the gum concentration increased (4,296.40 ± 323.44, 3,935.70 ± 549.70 and 3,075.20 ± 647.55 mg GAE/100 g for 3, 5 and 10% of gum, respectively) (global averages). The content of TPCs for the REC powder was similar than for the concentrations of 3 and 5% of gum, and higher than that of the 10% of concentration of gum.

Antioxidant Characteristics
Antioxidant activity (AC). The ACs (Table 6) concerning the type of gum was higher for the REC and MD powders (1,722.40 ± 40.54 and 1,687.20 ± 96.21 mg TE/100 g, respectively) (global averages) compared to the powders with GA and MD:GA (60:40) (1,556.00 ± 61.28 and 1,613.70 ± 80.24 mg TE/100 g, respectively) (global averages). For the gums concentration, a decrease in the ACs was observed as the gum concentration increases (1,689.40 ± 68.33, 1,637.80 ± 76.43 and 1,529.70 ± 63.49 mg TE/100 g of powder for 3, 5 and 10% of gum, respectively) (global averages). The ACs of the REC powders was similar to that of the concentration of 3% and higher than that of the concentrations of 5 and 10% of gum.

Conclusions
The microencapsulation of extracts from Roselle calyces by lyophilization provided high yields, as well as powders with good antioxidant and color properties. The concentration of gum is an important aspect to consider in the encapsulation of extracts from Roselle. According to the results obtained, the extracts of Roselle microencapsulated with maltodextrin and gum arabic at a concentration of 3%, allowed to obtain powders with the best antioxidant and color properties. Therefore, the use of microencapsulated powders obtained with these conditions signifies a viable option in the development of functional foods. However, it is recommended to carry out a stability study of the powders to determine the efficiency of the encapsulation process as well as the fading characteristics due to physical phenomena such as heat, oxygen, and light among others.
(Universidad Tecnológica de Izúcar de Matamoros) for the scholarship granted for the completion of his doctoral studies.