Optimization of the formula and processing of a sweet potato leaf powder‐based beverage

Abstract For the development of a sweet potato leaf powder (SPLP)‐based beverage, we investigated the effects of blanching methods on SPLP quality (including color, nutritional and functional compositions and antioxidant activity), and the effects of particle size and stabilizers on suspension stability of final product. The total polyphenol and antioxidant activity of SPLP of uncut group were 1.69 and 1.91 times those of cut group, respectively, and the indices of nutritional quality of copper, manganese and vitamin E of uncut group were significantly greater than cut group. The ultrafine SPLP‐produced lowest gravitational sedimentation ratio (49%), indicating it had greatest suspension stability. The optimized formula of SPLP‐based beverage was as follows: ultrafine SPLP of uncut group was mixed with 2.5% (w/w, powder basis) xanthan gum, 1% calcium lactate, 2% ascorbic acid, 12% maltodextrin, 20% xylitol, and 0.9% apple essence. The final product had high nutritional value along with consumer‐acceptable flavor and texture.

reduce transportation and storage costs, but would also have extended shelf lives. Undesirable browning during the process effects mainly caused by polyphenol oxidase (PPO) and peroxidase (POD) on vegetables greatly affects the sensory appeal and flavor of the final products. Blanching is necessary before drying for which not only inactivates PPO and POD to inhibit the browning effect, but also enhances drying rate and removes pesticide residues (Pandey, Mishra, & Misra, 2019;Xiao et al., 2017). In our previous study, we found that after the SPL were blanched, the PPO and POD activities decreased by 87.64% and 85.7%, respectively (Song, Mu, Sun, & Li-Sha, 2015). The traditional blanching method often divides the green leafy vegetables (such as mustard, cabbage, and lettuce) into pieces before blanching, resulting in significant color preservation, but the contents of water-soluble components, such as vitamin C and antioxidant compounds, dropped more than 40% and 50%, respectively (Sengkhamparn, Chanshotikul, Assawajitpukdee, & Khamjae, 2013).
In addition, our pre-experiment revealed that SPLP easily precipitated after mixing with water. The particle size is a key parameter associated with the suspension stability of SPLP. Owing to the lack of an established formula and processing method, the commercialized production of SPLP products is still difficult, which seriously limits the effective development and utilization of SPL.
Therefore, in the present study, the effects of different blanching methods and particle sizes on the quality of SPLP (including color, nutritional and functional composition, antioxidant activity levels, and suspension stability) were investigated. Furthermore, single factor and orthogonal tests were performed to optimize the formula (including stabilizers and flavor agents) of a SPLP-based beverage.
The purpose of this study was to develop a SPLP-based beverage, with a high nutritional value and acceptable sensory qualities, to improve the effective development and utilization of SPL and further promote the sustainable development of the sweet potato processing industry in China.

| Plant materials
The fresh SPL from cultivar Simon No. 1 were provided by Haileda Food Co., Ltd., Beijing, China, in the middle of August 2018. After harvesting in the morning, the fresh SPL were immediately sent to our laboratory.

| Preparation of SPLP
The SPL were washed with tap water, then divided into the cut (leaves cut into pieces) and uncut (whole leaves) groups, blanched in hot water at 96 ± 2°C for 1 min and then cooled in cold water (15°C).
SPL that were cut and not blanched served as the control group, and the three groups were all dried by vacuum freeze dryer. Afterward, the dried SPL were ground and sifted through 80-mesh, 100-mesh, and 200-mesh sieves. The superfine powder was obtained by superfine pulverizer (RT-04, Beijing Kaichuangtonghe Technology Development Co., Ltd.).

| Proximate composition and gross energy
The moisture content was determined using ASAE standards (ASAE, 1983). Crude protein, crude fat, and ash were determined using Association of Analytical Chemists (AOAC) methods (AOAC, 2000).
The crude protein content [g/100 g dry weight (DW)] was measured using the micro-Kjeldahl method, with a nitrogen to protein conversion factor of 6.25 (AOAC method 976.05). The crude fat content (g/100 g DW) was measured using AOAC method 960.39. The ash content (g/100 g DW) was measured by weighing powdered samples before and after a heat treatment (550°C for 12 hr). The crude (g/100 g DW), soluble, and insoluble fiber contents were measured using International Standards Organization (ISO) method 5498 (1981). The carbohydrate content (g/100 g DW) was measured by subtracting the sum of percent ash, crude fat, crude protein, and crude fiber contents from 100. The gross energy (kcal/100 g DW) was measured using a bomb calorimeter with the ISO method 9831 (ISO, 1998).

| Mineral content
SPLP was digested in concentrated HNO 3 (AOAC, 1995). The digest was raised to 25 ml with deionized water. The mineral content was determined by inductively coupled plasma atomic emission spectrometry, expressed in mg minerals/100 g DW.
The vitamin contents were expressed as mg of β-carotene, and the vitamins C, B1, B2, and B3, E and folic acid per 100 g SPLP were expressed on a DW basis.

| TPC and antioxidant activity levels
The extraction of crude polyphenols was performed using ultrasound-assisted ethanol solvent extraction as reported by Sun, Mu, Xi, Zhang, et al. (2014). Briefly, 1 g of sample was extracted with 20 ml of 70% (v/v) ethanol for 30 min at 50°C and subjected to ultrasonic wave treatment. The mixture was centrifuged at 6,000 g for 10 min at 4°C, and the residue was re-extracted two more times as described above. The supernatants were pooled, concentrated, and freeze-dried to obtain a crude extract. TPC was determined using the Folin-Ciocalteu method with a modification (Sun, Mu, Xi, Zhang, et al., 2014). The quantification was carried out using a calibration curve of chlorogenic acid standards, and the linear regression equation was y = 8.7671x + 0.0068, R 2 = 0.9994.
The TPC was expressed as chlorogenic acid equivalents (CAE) on a DW basis.
The antioxidant activity of SPLP was determined using the ferric reducing antioxidant power (FRAP) assay described by Lu, Ross, Powers, Aston, and Rasco (2011) with some modifications. The quantification was carried out using a calibration curve of Trolox standards, and the linear regression equation was y = 0.0029x + 0.017, R 2 = 0.9918. The antioxidant activity of the SPLP was calculated based on a calibration curve of Trolox and expressed as mg Trolox equivalents (TE) per 100 g on a DW basis.

| Amino acid composition
The amino acid compositions of SPLP were measured using a

| Index of nutritional quality
The INQ was calculated according to the following formula (Sun, Mu, Xi, Zhang, et al., 2014): where the Chinese NRVs for protein, fat, carbohydrate, and fiber are 60 g, ≤60 g, 300 g and 25 g, respectively.

| Particle size distribution
The particle size distribution was measured from 0.04 μm to 2,000 μm using a laser diffraction particle size analyzer (BT-9300, Dandong Baite Instrument Co., Ltd) with ultrapure water as the solvent. Briefly, 0.2 g sweet potato leaf powder was mixed with 400 ml ultrapure water, and the particle size was analyzed after circulating for 5 min at room temperature.

| Suspension stability
Suspension stability was measured using the method of Kosin et al. (2010), with a slight modification. Briefly, SPLP was dispersed in boiled water (60°C) at a powder:water ratio of 1:50 (w/w) and mixed to form a suspended mixture. The suspension stability of SPLP was measured using the gravitational sediment ratio, which was calculated according to the following formula:

| Formula optimization
To further improve the dispersion, as well as the sensory properties of SPLP, several downstream processes to prepare a SPLP-based beverage were investigated, including the addition of stabilizers and flavor agents.
sodium, sodium alginate, and konjac gum were investigated. Briefly, the stabilizers were added individually to the powder-water mixture that was mentioned above at concentrations of 2.5% (w/w), which was determined as the optimal concentration in a preliminary experiment (data not shown).

| Orthogonal test
Based on the results of the single factor preliminary experimental analysis (date not shown), the additions of xylitol, ascorbic acid, apple essence, and maltodextrin greatly influenced the color, flavor, and other sensory properties of the SPLP-based beverage, while the addition of calcium lactate had little influence. Therefore, an orthogonal test with four factors and three levels was designed to optimize the formula for the SPLP-based beverage, including the additions of xylitol, ascorbic acid, apple essence, and maltodextrin. The addition of calcium lactate was determined to be 1% according to the Chinese National Standard GB/T 2760 (2014) for food additives. The parameters of the L 9 (3 4 ) orthogonal test are shown in Table 1, and the sensory evaluation standard of the SPLP-based beverage was the response value.
A panel, consisting of 15 trained members, was used to evaluate the color, smell, texture, and taste of SPLP-based beverages brewed with boiled water (60°C). Details of sensory evaluation are shown in Table S1. The evaluation was carried out in a descriptive manner using scores for each attribute. The total score was 100, and the average score was used to evaluate the sensory appeal of the product after removing the highest and lowest scores.

| Statistical analyses
All the experiments were performed in triplicate. Statistical analyses were performed using the SAS version 9.4 software (SAS Institute Inc., Cary, NC, USA). Statistical significance was set to p < .05.

| Color
The color characteristics are listed in Table 2, compared with the control group, whose chromatic L* value was 43.25 ± 0.045, the blanched groups (cut and uncut groups) showed decreasing L* values, resulting in a darker appearance. Similar color changes were reported by Latorre, de Escalada Plá, Rojas, and Gerschenson (2013), in which the changes were correlated with the pigment leakage as consequence of cell membrane thermal denaturalization. In addition, the a* and b* values of the blanched groups were significantly lower than those of the control group, indicating that the blanched groups showed less greenness (lower a*) and less yellowness (lower b*) than the control group, which was probably owing to the loss of water-soluble components during the blanching. The total difference in color (ΔE) indicates the magnitude of the color difference between processed and unprocessed samples.  (Lopriore et al., 2001). The SPLP made from the blanched groups maintained its color for a long time (>5 hr) after brewing, indicating that the blanching operation had a protective effect on the green coloration.
The SPLP made from the cut group had the lowest moisture content (5.74 ± 0.04 g/100 g PB), which could not only reduce the costs of transportation and storage, but also extend the shelf life. Blanching could dissolve the hydrophobic waxy layer and weaken cell walls and membranes. However, the degree of increasing skin permeability was depended on the blanching method, which resulted in the different moisture content in cut group and uncut group (Giovanelli, Brambilla, Rizzolo, & Sinelli, 2012

TA B L E 2
The effect of different blanching method on the color, proximate composition and gross energy, mineral, vitamin, total polyphenol content (TPC), and antioxidant activity of sweet potato leaf powder difference was related to the content of crude fiber content in different blanching treatment group.
There was no significant difference between the crude protein contents of the control and blanched groups, indicating that the blanching operation did not cause a loss of protein. Furthermore, the average crude protein content in the SPLP (29.81 g/100 g DW) was 1.64 to 6.62 times greater than that of commercial green tea powder (18.10 g/100 g DW) and commercial barley leaf powder (4.50 g/100 g DW).
The crude fat content of the control group was 4.12 ± 0.01 g/100 g DW, which was greater than that of the cut group (3.96 ± 0.01 g/100 g DW) and that of the uncut group (4.04 ± 0.02 g/100 g DW). Our results were similar to those reported by Sun, Mu, Xi, and Song (2014), in which boiling decreased the crude fat content of SPL significantly. During the blanching process, the volatiles and water-soluble fatty acids in the SPLP may be partially lost or destroyed, causing a decrease in the crude fat content, and the lowest crude fat content of cut group could be attributed to a higher degree of leaching during the blanching step. Additionally, the use of blanched SPLP with low fat contents as raw materials for our product is in accordance with low-fat diets.
The carbohydrate content was the highest in the uncut group (17.50 ± 0.00 g/100 g DW), which was significantly greater than that of the control and the cut groups. This result was different from that obtained by Mepba, Eboh, and Banigo (2016) The average crude fiber content of SPLP was 36.06 g/100 g DW, which was similar to that of commercial tea powder (40.27 g/100 g DW) and that of commercial barley leaf powder (41.33 g/100 g DW). The average soluble and insoluble fiber contents were 8.58 g/100 g DW and 27.48 g/100 g DW, respectively, indicating that SPLP is rich in insoluble fiber. Dietary fiber, especially insoluble dietary fiber, is good for controlling weight and bowel detoxification owing to its high water-binding and swelling capacities (Tan, Wei, Zhao, Xu, & Peng, 2017).
The ash contents of SPLP ranged from 10.08 ± 0.01 g/100 g to 10.41 ± 0.01 g/100 g DW and that of the blanched group was lower than that of the control group, indicating that blanching decreased the ash content significantly. Our results were similar to those reported by Zoro, Zoué, Adom, and Niamké (2016), who investigated the nutritional and antioxidative properties of blanched leafy vegetables and revealed that blanching had a negative impact on the ash content. This may result from the large amount of cytochylema and that some of the water-soluble nutrients could be lost during blanching process because the SPL are exposed to water, especially the lowest ash content of cut group.
The gross energy of the uncut group was 1,214.00 ± 00 kJ/100 g DW, which was greater than those of control group (1,184.00 ± 00 kJ/100 g DW) and cut group (1,131.00 ± 0.00 kJ/100 g DW). This could result from the increased carbohydrate level in the uncut group, which is the main energy producing substance.

| Minerals
Minerals are classified into two groups: macroelements (Na, P, K, Ca, and Mg) and microelements (Fe, Zn, Cu, Mn, Se, Pb, As, and Hg). As shown in Table 2, the Na content of the control group was 47.36 ± 0.01 mg/100 g DW, and those of the cut and uncut groups increased by 2.14 and 1.90 times, respectively. The aver-  (2007), in which the loss of nutrients during hot water blanching was caused mainly by leaching or diffusion. However, the K/Na ratios determined in the cut (28.14) and uncut (28.23) groups were greater than in spinach (18.10) and water-spinach (11.56) (Taira, Taira, Ohmine, & Nagata, 2013). A low K/Na ratio is correlated with diabetes, high blood pressure and hypokalemia; therefore, SPL powder may be a healthy option for individuals with these conditions.
As for some microelements that are benefit to health, the Fe content of the control group was 12.54 ± 0.10 mg/100 g DW, while those of the cut and uncut groups were reduced by 19.62% and 33.57%, respectively. Zn is an important functional metal in the human body, which has over 300 Zn-containing enzymes. The average Zn content of SPLP was 2.14 mg/100 g DW, which was 8.03 times greater than in commercial tea powder (0.33 ± 0.00 mg/100 g DW). Cu is an essential dietary microelement that is a cofactor of many redox enzymes and works on many biological processes, including antioxidant defense, neuropeptide synthesis, and immune function. The average Cu content of SPLP was 1.15 mg/100 g DW. The most abundant microelement was Mn, with an average content of 12.17 mg/100 g DW.
In addition, Se is an essential antioxidant mineral that performs many functions in the body, including the preservation of normal liver functions and maintaining resistance to disease, and the average Se content of SPLP was 5.68 μg/100 g DW, which was greater than that of cabbage (3.63 μg/100 g DW) but similar to the content of broccoli (Sirichakwal, Puwastien, Polngam, & Kongkachuichai, 2005).
However, the minerals content is considered to be a result of the loss of water-soluble minerals and interaction among other compositions, which resulted in the different minerals content between cut group and uncut group (Ngobese & Workneh, 2018).
There are some microelements that are harmful and toxic, such as Pb, As, and Hg. Neither Pb nor Hg was detected in any of the SPLP, and as content was lower than the limit (0.5 mg/100 g DW) of Chinese National Standard GB/T 29602 (2013) for solid beverages.
Thus, SPLP is a safe food product.

| Vitamins
SPL are a good source of various vitamins. As shown in Table 2, the most abundant vitamin was β-carotene, which had an average content of 110.03 mg/100 g DW, which was 32.69 times greater than that of spinach. In addition, the β-carotene content of the cut (126.50 ± 1.60 mg/100 g DW) and uncut (121.5 ± 5.00 mg/100 g DW) groups was greater than that of the control group, indicating that blanching increased the β-carotene significantly. This could be attributed to β-carotene being a fat-soluble compound, while watersoluble compounds are lost during the blanching process (Kourouma, Mu, Zhang, & Sun, 2019). Similarly, the vitamin E content of the uncut group was 18.90 ± 0.30 mg/100 g DW, which was 2.36 times greater than that of the control group. The average vitamin C content of SPLP was 74.55 mg/100 g DW, which was greater than those of cucumber (10.60 mg/100 g DW), white radish (22.50 mg/100 g DW), and spinach (39.00 mg/100 g DW) (Isabelle et al., 2010). Vitamin B is an essential nutrient for health, growth, and reproduction because it stimulates metabolism and promotes the conversion of proteins, fats, and carbohydrates into energy. The average vitamin B2 and B3 contents of SPLP were 1.05 mg/100 g DW and 0.56 mg/100 g DW, respectively, which was greater than those of wheat (0.11 mg/100 g DW and 0.16 mg/100 g DW, respectively) (Shewry et al., 2011). There was no significant difference between the folic contents of the control and blanched groups, and the average folic content was 55.61 mg/100 g DW. Interestingly, the retention of vitamin C of cut group was higher than that of uncut group, but the change of fat-soluble vitamins retention was the opposite. The vitamin C loss in the uncut group probably was related to the mass transfer phenomena rather than to the leakage of nutrients in the blanching processing, but the fat-soluble vitamins increase in the uncut group probably due to the leakage of the water-soluble nutrients (Epameinondas, Eleni, Petros, & Ahrné, 2018).

| TPC and antioxidant activity
As shown Table 2, the SPLP of the control group had the greatest TPC (6.42 ± 0.25 g CHAE/100 g DW), which was 2.2 and 1.3 times greater than those of cut and uncut groups. The decrease in the TPC agreed with the findings of Sun, Mu, Xi, and Song (2014), in which the effects of domestic cooking methods on the polyphenol of SPL were investigated. They found that boiling reduced the TPC of SPL by 30%. There was a significant difference between the TPC of the cut (2.90 ± 0.10 g CAE/100 g DW) and uncut (4.91 ± 0.20 g CAE/100 g DW) groups, indicating that different blanching methods had significant effects on the TPC of SPLP. This could result from the cut SPL having a greater area that was exposed to water.
As shown in Table 3, the antioxidant activity level of the control group was 14.76 ± 0.09 g TE/100 g DW, which was greater than those of the cut and uncut groups. Decreases of 20.26% and 56.41% in antioxidant activity level were noted in the uncut and cut groups, respectively, which was in agreement with the TPC loss in SPLP. Sun, Mu, Xi, and Song (2014) found that the antioxidant activity of SPL cooked by boiling decreased by 63.82%, and there was a positive correlation between antioxidant activity and polyphenol content in SPL.

| Amino acid composition
As shown in Table 3, the total amino acid contents of the cut and uncut groups, with values of 23.80 ± 0.01 g/100 g DW and 23.49 ± 0.25 g/100 g DW, respectively, were significantly greater than control group value of 22.69 ± 0.01 g/100 g DW. The percentage ratio of EAA to total amino acid ranged from 41.70% (control group) to 42.55% (cut group), and a food with a ratio greater than 36% is considered to be an ideal source of protein (Consultation., 2007). Glutamic acid was the most predominant amino acid in all the SPLP, which was different from the results of Lu et al. (2011), in which the main amino acid of SPL was aspartic acid. Several factors contribute to differences in glutamic and aspartic acid contents, including genotype, maturity, and cultivar. The capacity to satisfy the need for nitrogen and EAAs is primarily evaluated by the nutritional values of proteins (Olaofe, Faleye, Adeniji, & Akinsola, 2009). SPLP contained a high concentration of lysine, with an average content of 1.86 g/100 g DW. Lysine is an EAA and a nutritionally critical amino acid that is lacking in the main cereal crops. The total EAAs of the cut and uncut groups were 10.35 g/100 g DW and 9.88 g/100 g DW, respectively, and these values were greater than those of spinach, fragrant green tea and eucalyptus tea (He et al., 2018). Table 3 also shows the total acidic amino acid content, which was greater than the total basic amino acid content in all the groups, indicating that the proteins in SPLP are mainly acidic in nature.
Furthermore, compared with the amino acid reference patterns an adult diet and whole egg protein (Table S2), the amino acid patterns of the SPLP of the control, cut, and uncut groups were similar, but all the amino acid scores were very low (a score of 100 is considered ideal).

| Index of nutritional quality
The relationship between the nutrient content in single foods, meals, and diets and the NRV is measured by INQ. A food with an INQ value of 2-6 is a good source of that specific nutrient, and a food with an INQ value greater than 6 is an excellent source of that specific nutrient (Sun, Mu, Xi, Zhang, et al., 2014). As shown in

| Suspension stability
Particle size influences the sedimentation behavior and cloudiness of beverages. According to the Stokes' Law and the Derjaguin-Landau-Verwey-Overbeek theory, the smaller the particle sizes and diameters of the solid particles in the suspension system, the more stable the suspension system, because the smaller particle size may result in a slower floating rate (Giri, Mangaraj, Sinha, & Tripathi, 2017). The effects of different particle sizes on the gravitational sedimentation ratio and the sensory phenomenon of SPLP suspension are shown in Figure 1a,b, respectively. In 10 min, the gravitational sedimentation ratio increased sharply from 37% to 60%, as the particle size decreased from 80-Mesh to ultrafine power. From 10 to 30 min, the gravitational sedimentation ratio increased slightly and the ratio values (in the 30th min) of ultrafine TA B L E 3 The effect of different blanching method on the amino acid of sweet potato leaf powder (g/100 g, DW)

| Stabilizer
The effects of different stabilizers on the gravitational sedimentation ratio and the sensory phenomena of SPLP suspension are shown in Figure 1c,d, respectively. The smaller the gravitational sedimentation ratio value, the greater the suspension stability. The gravitational sedimentation ratio of the suspended SPLP mixture with xanthan gum was the smallest, and the suspension mixture could be kept uniform and delicate for an hour. The mixture containing arabic gum had a gravitational sedimentation ratio of 32.0%, and the suspension mixture was still uniform and delicate. The effects of guar gum, hydroxypropyl methylcellulose, carboxymethylcellulose sodium, sodium alginate, and konjac gum were less desirable.
Additionally, xanthan gum can effectively improve suspension stability of SPLP-based beverage owing to the ability to create a strong physical network in the system to resist the gravitational force to prevent particles from sinking (Muhammad et al., 2019).

| Flavor agents
As shown in Table 5. The order of influence for the various factors on the flavor agents of SPLP-based beverages was determined using a range analysis as follows: B (ascorbic acid) > D (maltodextrin) > A (xylitol) > C (apple essence). According to the variance analysis (Table S4), there were significant differences in the scores among different levels of each factor. When ascorbic acid was gradually added, the score decreased, while when maltodextrin was gradually added, the score increased. The fourth group had the greatest score (93.1 ± 2.09), indicating that the best formula for flavor agents was 2% ascorbic acid, 12% maltodextrin, 20% xylitol and 0.9% apple essence. The sweet potato leaf powder-based beverage obtained using this formula had a high nutritional value and desirable flavor and texture.

| CON CLUS ION
The SPLP made from the blanched groups had color-protective and high nutritional value. In addition, the powder of the uncut group had a greater TPC and antioxidant activity level as compared with those of the cut group. The ultrafine SPLP showed a markedly narrower particle size distribution and high suspension stability. We developed and optimized a formula for a SPLP-based beverage that further improved its suspension stability and flavor by adding a stabilizer and flavor agents, respectively. The optimized formula was as follows: the ultrafine SPLP of the uncut group was mixed with 2.5% xanthan gum, 1% calcium lactate, 2% ascorbic acid, 12% maltodextrin, 20% xylitol, and 0.9% apple essence. The SPLP-based beverage obtained by this formula had a R was represented the range value of each factor.
Data are means ± SD (n ≥ 2). Values within columns with different letters are significantly different (p < .05).
high nutritional value and desirable flavor and texture. A SPLPbased beverage may improve the effective development and utilization of SPL and further promote the development of the sweet potato processing industry in China.

ACK N OWLED G EM ENTS
The authors gratefully acknowledge the National Key R&D Program of China (2016YFE0133600 and 2017YFD0400401), the Natural Science Funding of China (31701614), and the Earmarked Fund for China Agriculture Research System (CARS-10-B21).

CO N FLI C T O F I NTE R E S T
The authors declare no competing financial interest.

E TH I C A L A PPROVA L
Ethics approval was not required for this research.