300 g of fresh lotus leaves were hot-air dried at 60 °C, 80 °C and 100 °C, respectively, (DHG-9070A, Shanghai Yiheng Scientific Instruments Co., Ltd), with the drying endpoint determined when a 10 g sample maintained constant mass after 5 min at room temperature. The drying durations were 6.55 h at 60 °C, 3.85 h at 80 °C, and 2.12 h at 100 °C. The final moisture content measured by oven-drying gravimetric method was 3.80%, with an average sample loading rate of 12.30%. Then a grinder (1180B, Annos Electric Factory in Dongfeng Town, Zhongshan City) was used to grind. Furthermore, LLP was sieved to three particle sizes: large particle sizes: >500 μm, medium particle sizes: 150–500 μm and small particle sizes: <150 μm and stored in the refrigerator (YC-330, Aucma) at −18 °C.

Green wheat leaf powder (1%, w/v), sodium benzoate (0.02%, w/v), sodium bicarbonate (0.02%, w/v), xanthan gum (0.05%, w/v), and D-sodium ascorbate (0.05%, w/v) were put into four tea bags based on the total volume of the beverage (1,000 mL). Then 0, 0.25, 0.5 and 1.0% (w/v) of LLP (<150 μm) dried at 60 °C (T1) were mixed into each tea bags, respectively. Four tea bags were individually placed in bottles containing 1,000 mL of drinking water and placed in a 75 °C water bath (HH-420, Licheng Technology) for 30 min. Then, each supernatant was bottled after centrifuging (TDL-40B, Bai Ou) at 3,000 rpm for 2 min. Each bottled T1LLPSGWLB was refrigerated at 4 °C before using.

The method of total phenolic compounds (TPCs), DPPH free radical scavenging rate (DPPH), ferrous iron chelating ability (FICA) and ferric reducing power ability (FRPA) were determined by Kim and Chin (2016) and modified slightly.

The determination of TPCs for LLP was conducted as follows. 0.1 g of each LLP was weighed and placed in 10 mL of deionized distilled water (dd-water). After thorough mixing, it was centrifuged. The supernatant was then diluted to prepare 1, 0.5, 0.25, and 0.1% (w/v) sample solutions.

Then, 100 μL of the sample supernatant was pipetted and added with 2.8 mL of dd-water, 2 mL of 2% sodium carbonate, and 100 μL of 50% Folin–Ciocalteu reagent. After standing for 30 min, the absorbance was measured at a wavelength of 750 nm, dd-water was used instead of the sample as the blank control. The experiment was repeated in triplicate, and absorbance values were recorded. Gallic acid was used as the standard (with a concentration range of 0–200 mg/L). The standard curve equation was y = 0.001134x + 0.008350, with an R² = 0.9969.

The determination of DPPH was conducted with ascorbic acid (AA) as a reference. 1 mL of the diluted LLP solution was taken and added with 2 mL of DPPH (0.1 mmol/L). The sample was placed in the dark for 30 min, and then the absorbance was measured at a wavelength of 517 nm. Meanwhile, 0.5 mL of methanol was used to substitute the sample and added to 2 mL of DPPH, with the absorbance measured under the same conditions and repeated three times in parallel. The scavenging ability (%) of the LLP extract against DPPH was calculated according to the following Equation (1):

For the study of FICA, 0.5 mL of the above diluted solution was taken, and then added 0.5 mL of dd-water, 100 μL of ferrous chloride (0.6 mmol/mL), 0.9 mL of methanol, and 100 μL of ferrozine (5 mmol L⁻¹ in methanol). Standing for 10 min, the absorbance was measured at a wavelength of 562 nm. Ethylenediaminetetraacetic acid (EDTA) was used as the reference. In the control group, the ferrozine was replaced with methanol, and the absorbance value was measured under the same conditions. All the operations were carried out in parallel three times, and a blank group without the sample was set up. The calculation method of the chelating activity (%) was as follows, according to Equation (2):

For the determination of FRPA for LLP, similarly, 2.5 mL of the sample diluted solution (with AA as the reference) was put into test tube. Then 2.5 mL of 200 mmol/L phosphate buffer (pH = 6.6) was added and mixed well, followed by 2.5 mL of 10% potassium ferricyanide. The mixture was incubated in a constant temperature incubator at 50 °C for 20 min, then centrifuged at 2,000 rpm for 10 min. 2.5 mL of each supernatant was taken into a test tube. In the experimental group, 2.5 mL of dd-water and 0.5 mL of ferric chloride (1 mg mL⁻¹) were added, while in the control group, 3 mL of dd-water was added. After 10 min, the absorbance was measured at 700 nm. The reducing capacity of the LLP was calculated by the following Equation (3):

For the detection of the T1LLPSGWLB, the method in 2.1 was referred to. 10 mL of T1LLPSGWLB was taken, centrifuged at 3,000 rpm for 2 min, and the supernatant was collected as a sample. Subsequently, it was diluted with dd-water to prepare solutions with concentrations of 1, 0.5, 0.25, and 0.1% (v/v) and the experimental procedures were carried out according to the above-mentioned methods for determining TPCs, DPPH, FICA, and FRPA. And all measured values must be subtracted by the values of the blank group.

The pH values of the T1LLPSGWLB were randomly measured using a pH meter (pHS-3C, Owike), repeated 5 times and recorded. The color values consisting of the lightness (L*), redness (a*) and yellowness (b*) was measured by color reader (CR-10plus) and repeated six times. The changes in pH and color values of the T1LLPSGWLB on days 0, 3, 7, 14, 21, 28, and 35 were also recorded. Seven sets of separate samples representing different storage periods were prepared on day 0 and measured on their respective designated storage days.

The SSC of the T1LLPSGWLB on days 0, 3, 7, 14, 21, 28, and 35 were determined and recorded with a handheld refractometer, referring to the method of Ranganna (1986).

According to the SS measurement procedure of Muhammad et al. (2021) with slight modifications, the prepared T1LLPSGWLB of different formulations were diluted 5 times. Their absorbance values M0 were measured at 550 nm. After centrifuging at 3,000 rpm for 2 min, they were diluted by the same multiple, and their absorbance values were measured under the same conditions, recorded as M1. The results were calculated according to the following Equation (4):

According to the method of Qiu and Chin (2022) and modify it slightly. The 10 mL sample was mixed with 90 mL of dd-water, and then 0.1 mL of the mixed sample was evenly spread on total bacterial counts (TBCs) and enterobacteriaceae counts (EBCs) plates, respectively, the total number of bacteria and the number of EBCs was determined. The TBCs were prepared using plate count agar medium (AOBOXING, Beijing), while the EBCs were prepared using violet red bile agar (AOBOXING, Beijing). The TBCs and EBCs were determined after cultivation at 37 °C for 24 h, and the colony count unit is Log CFU/mL.

Referring to the method of Ozarda et al. (2015) and slightly modifying it. Sensory evaluation was conducted on GWLB samples supplemented with 0, 0.25, 0.5, and 1% T1LLPS, respectively. The appearance, color, palate, taste, smell and overall acceptability were used as the evaluation criteria, with a total score of 9 points (9—Excellent, 8—Extremely good, 7—Very good, 6—Moderately good, 5—Good, 4—Fair, 3—Very fair, 2—Poor, 1—Very poor). A sensory evaluation team consisting of 60 food students (30 men and 30 women) with experience in tasting will conduct sensory evaluations. The panelists consisted exclusively of food science students with no sensory impairments, all of whom underwent professional taste and olfactory training. Samples were presented in identical containers with a fully randomized serving order. Each sample was evaluated in triplicate, and the mean values were calculated for statistical analysis.

The variance analysis of the experimental data was used the SPSS 27.0 software package. Among them, the one-way analysis of variance (ANOVA) was used for the analysis of the TPCs of the LLP and the T1LLPSGWLB, with the different treatments of LLP as the main factor. The two-way ANOVA was applied to the analysis of DPPH, FICA and FRPA of LLP and T1LLPSGWLB, taking the treatments and concentrations as the main factors. In addition, the two-way ANOVA was also adopted for the analysis of pH, color, SS, SSC, TBCs and EBCs of T1LLPSGWLB, with the addition of T1LLPS and storage as the main factors. The Duncan’s multiple range test was used to determine the significance of differences at the 5% significance level.

As can be seen from Figure 1, The TPCs of LLP dried at 60 °C were significantly higher than those at 80 °C and 100 °C (p < 0.05), under the condition of consistent small particle size (SPS), the TPCs of LLP dried with 60 °C hot air was (7.57 ± 0.12 g/100 g), while that dried at 80 °C was (6.79 ± 0.13 g/100 g), and that dried at 100 °C was (5.80 ± 0.18 g/100 g). The observed result was primarily linked to the Maillard reaction under high temperatures. Elevated temperatures promoted the generation of Maillard reaction products, which likely modified the composition and concentration of extracted polyphenols (Antony and Farid, 2022). In this study, the highest TPCs were recorded in the 60 °C drying group, as the Maillard reaction was minimized at this temperature, resulting in less alteration to the polyphenolic profile and greater retention of phenolic compounds. The effect of heat treatment on the leaching of polyphenols in apple parenchyma, too high temperature will lead to the destruction of polyphenol components in apple parenchyma and reduce the phenolic content (Kebe et al., 2015). At the same drying temperature, the TPCs increased with the decrease of LLP particles. Pairwise comparisons among all particle size showed statistically significant differences, with the SPS exhibiting higher TPCs than both the medium particle size (MPS) and large particle size (LPS) (p < 0.05). Under the same 60 °C hot-air drying temperature, the TPCs of LLP with LPS was (4.64 ± 0.19 g/100 g), that of MPS was (6.13 ± 0.13 g/100 g), and that of SPS was (7.57 ± 0.12 g/100 g). The SPS exhibited the highest values, likely due to the increased surface area for molecular transport, which facilitates more extensive diffusion of solutes (Norra et al., 2016). Compared to LLP with LPS dried at 100 °C, the TPCs of SPS dried at 60 °C increased by 23.38%. Furthermore, the TPCs of 60 °C-dried SPS were 38.71% higher than those of 60 °C-dried LPS. These results demonstrate an economically efficient method to enhance the functional value of the product without increasing raw material usage.

DPPH of LLP also tended to increase with increasing concentrations and decreasing particle sizes (Figure 2A). Moreover, statistically significant differences were observed across different temperatures and particle sizes. For instance, the 1% concentration of 60 °C hot-air-dried LLP exhibited a DPPH of 21.99 ± 3.39% for large particles, 29.69 ± 2.84% for medium particles, and 44.29 ± 3.28% for small particles (p < 0.05). For the same small particle size at 1% concentration, the DPPH value of the 100 °C was 25.12 ± 2.96%, representing a reduction of 43.28% compared to the 60 °C. At the concentration of 0.1%, the DPPH of each treatment was 7.30% ± 2.88% to 18.76 ± 3.62%. At 1% concentration, the DPPH reached 11.28 ± 3.48% to 44.29 ± 3.28%, increased 74.53%. The particle size can be effectively reduced and the specific surface area increased after ultra-fine grinding treatment. When evaluating the DPPH of black kidney bean powder using 2,2-diphenyl-1-picrylhydrazyl as the radical, the scavenging ability of the black kidney bean powder was enhanced as the particle size decreased (Sun et al., 2019). In addition, TPCs and DPPH showed a consistent trend. Due to the synergistic effect between phenolic compounds and the high hydrogen atom donating capacity, a higher TPCs may lead to an increased DPPH radical scavenging activity (Das et al., 2012).

In Figure 2B, as the concentration increased from 0.1 to 1%, the FICA of all treatments was observed to increase with the rise in concentration. FICA also tended to increase with decreasing temperature and particle size. Among all particles, the small particle size (<150 μm) dried at 60 °C exhibited the best FICA, with a value of 57.26 ± 4.85% to 79.03 ± 4.62%. Phenolic compounds possessed metal-binding sites, which could prevent the redox cycling of metals through metal chelation, thereby delaying oxidative activity (Andjelković et al., 2006). Consequently, the FICA of LLP (<150 μm) at 60 °C was the highest among treatments. In the study of the effect of particle size on the antioxidant properties of cactus powder, the highest contents of polyphenols and flavonoids were found in the particle size grade of 80–100 μm and antioxidant activities were significantly enhanced with the decrease of particle size (Nabil et al., 2020).

As shown in Figure 2C, the FRPA values were influenced by particle size and oven drying temperature. Except for the reference control (AA), the 1% concentration of T1LLPS still exhibited the highest FRAP value, which was 1.33 ± 0.00 O.D. It increased by 34.59% compared to T3LLPS at the same concentration. As the concentration increased from 0.0 to 1%, the FRPA of all treatments was observed to increase. FRPA serves as a potential marker for antioxidant activity by converting the iron/ferricyanide complex into its ferrous state. All extracts from date plum persimmon exhibited dose-dependent DPPH, FICA and FRPA capabilities, and they showed a good correlation with total phenolic compounds and total flavonoid contents, indicating that phenolic compounds directly contribute to these activities (Gao et al., 2014). In the present manuscript, the FRPA of LLP with SPS was greater than MPS and LPS. In a previous study, the contents of total phenols, total flavonoids and total carotenoids were higher in the red chili powder with smaller particle sizes (Gao et al., 2014).

In present study, when the temperature was decreased from 100 °C to 60 °C and the particle size was reduced from >500 μm to <150 μm, the magnitude of improvement in antioxidant activity was significantly enhanced. The DPPH, FICA, and FRPA of LLP with small particle size (<150 μm) dried at 60 °C (T1LLPS) were the highest among all treatments, and were all influenced by temperature and particle size.

Figure 3 shows the differences in TPCs among GWLB with different concentrations of T1LLPS added. The TPCs of GWLB increased with the increased addition of T1LLPS (p < 0.05). According to Section 3.1, the TPCs of T1LLPS was (7.57 ± 0.12 g/100 g), while the TPCs in GWLB with 1% T1LLPS was (3.19 ± 0.23 g/100 g), resulting in a retention rate of 42.14%. In a previous study, it was found that adding date flour to wheat flour increased the amounts of polyphenols and flavonoids in sponge cakes, as well as antioxidant potential (Najjaa et al., 2020). And Kolonas et al. (2023) mentioned that the antioxidant activity of phenols may be due to their ability to act as reducing agents and hydrogen donors.

From Figure 4, the values of DPPH, FICA and FRPA were found to decrease with the increase in the dilution gradient of T1LLPS. The GWLB with 1% T1LLPS exhibited a DPPH range of 55.69 ± 3.70% to 66.72 ± 1.88%, a FICA range of 61.10 ± 1.97% to 75.05 ± 0.99%, and a FRPA range of 0.41 ± 0.00 O.D to 0.69 ± 0.00 OD. In contrast, without T1LLPS showed a DPPH range of 42.01 ± 1.13% to 46.09 ± 1.20%, a FICA range of 28.19 ± 3.35% to 50.74 ± 2.70%, and a FRPA range of 0.12 ± 0.00 O.D to 0.45 ± 0.01 O.D. A previous study found that adding LLP to bread led to a notable rise in its DPPH and total polyphenol levels (Park, 2017). A strong positive correlation was found between the TPCs and the antioxidant activity as assessed by DPPH and FRPA tests (Feng et al., 2016). In addition to the standard sample group AA, the GWLB with the addition of 1% concentration of T1LLPS still exhibits the highest values of DPPH, FICA and FRPA, which was consistent with the trend observed in the TPCs of GWLB with different concentrations of T1LLPS added. It can be seen from this that the addition of 1% T1LLPS significantly improved the antioxidant activity of GWLB, which provides a specific reference for the formula optimization of functional beverages.

Table 1 illustrates the changes in pH value and color of the GWLB, supplemented with varying concentrations of T1LLPS, over a 35-day storage period. The pH of all treatments decreased over storage time (p < 0.05). Additionally, during the same storage period, the pH of the GWLB decreased with an increase of T1LLPS added (p < 0.05). During day 0 to day 35, the pH of GWLB added with 1% T1LLPS decreased from 7.23 ± 0.01 to 7.11 ± 0.02, representing a change of 1.08%. In contrast, without T1LLPS declined from 7.69 ± 0.02 to 7.52 ± 0.02, corresponding to a change of 2.21%. Both ranges of pH variation were within acceptable limits. In another study, it was proven that with the increase of storage time, some ingredients in the beverages may undergo oxidation or decomposition reactions to generate acidic substances such as oxidation of vitamin C producing acidic products that lower pH (Nicolas et al., 1994). Studies have shown that this may be due to the fermentation of organic acids (such as lactic acid, acetic acid, etc.) due to microbial activity (Kelbore et al., 2022). A study examining the influence of LLP on the quality of chicken patties under refrigeration conditions discovered that the pH values were lower in the samples with LLP than in the control group. The present research has shown that the organic acids in lotus leaves have an influence on the acidity of LLP (Mukherjee et al., 2009).

Due to the significant interaction between treatments and storage days (p < 0.05), it could be demonstrated that the more T1LLPS added to the GWLB, the greater L* values, and as the storage period increased, the L* values of all groups were decreased. From day 0 to day 35, the L* value of GWLB with 1% T1LLPS decreased from 55.14 ± 0.02 to 54.98 ± 0.03, representing a reduction of 0.29%, while that of the control group declined from 49.77 ± 0.02 to 48.69 ± 0.04, corresponding to a decrease of 2.17%. Polyphenolic compounds were contained in lotus leaves, which may influence the absorption and reflection of light, thereby further enhancing lightness (Manach et al., 2004). The longer the beverages were stored, the lower its lightness becomes (Septiyani et al., 2024). Under the same storage period, the higher the content of T1LLPS added, the more a* showed a decreasing trend, while b* exhibited an increasing trend. On day 0, the a* value of GWLB added with 1% T1LLPS was −1.96 ± 0.01, and the b* value was 24.73 ± 0.02, while the a* and b* values of the control group were −1.33 ± 0.02 and 20.17 ± 0.01, respectively and all observed variations fell within sensorially acceptable ranges. During the preservation of yogurt, it was observed that significant changes occurred in the a* and b* values of yogurt supplemented with LLP, which may be attributed to the degradation and oxidation of lotus leaf pigments (Kim et al., 2019). Meanwhile, the presence of chlorophyll would also lead to an increase in the b* value of the beverage (Vebrianty et al., 2021).

It can be seen from Table 2 that under the same storage period, SSC increased with the increase of the concentration of T1LLPS in GWLB, and SS decreased with the increase of the concentration of T1LLPS in GWLB (p < 0.05). On day 0, the SSC and SS of without T1LLPS were 2.16 ± 0.02 and 85.47 ± 0.02, respectively, while those of the GWLB added with 1% T1LLPS were 3.48 ± 0.01 and 81.59 ± 0.06, respectively. Substances rich in minerals and bioactive compounds can enhance the SSC of fruit juices and were shown to possess high antioxidant potential (Prakash et al., 2021). In present study, we found T1LLPS with high antioxidant activity and high amount of TPCs, which might improve the SSC of T1LLPSGWLB. Furthermore, different components in the T1LLPSGWLB (such as polysaccharides and proteins) may underwent phase separation, leading to uneven distribution of suspended particles and a decrease in stability (Norton et al., 2010). The SSC of all groups increased with the increase of storage period, while the SS decreased (p < 0.05). The SSC of GWLB added with 1% T1LLPS increased from 3.48 ± 0.01 on day 0 to 3.57 ± 0.01 on day 35. Conversely, the SS decreased from 81.59 ± 0.06 on day 0 to 80.12 ± 0.02 on day 35. Milk beverages with insufficient pectin content may undergo clumping during storage, leading to the evaporation of their water content (Lucey et al., 1999). As storage time increases, the evaporation of water in the T1LLPSGWLB may result in an increase in sedimentation. Throughout the duration of storage, the T1LLPSGWLB may undergo reactions such as oxidation and hydrolysis, which can alter the surface properties of the particles and weaken their stability (Chanamai and McClements, 2001). Therefore, in present research, the GWLB with the addition of 1% T1LLPS exhibited the highest SSC but the lowest SS among T1LLPSGWLB.

The total bacterial counts (TBCs) and enterobacteriaceae counts (EBCs) of the GWLB with the addition of T1LLPS at different concentrations during the storage period are shown in Table 3. Both the TBCs and EBCs decreased with the addition of T1LLPS in the GWLB during the storage period. The GWLB with 1% T1LLPS added, no total bacteria (TB) were detected before day 3, and no enterobacteriaceae (EB) were detected before day 7. Moreover, on day 7, the GWLB added with 1% T1LLPS exhibited reductions of 0.63 Log CFU/mL in TBCs and 0.47 Log CFU/mL in EBCs compared to the control group. This might be related to the antibacterial components such as tannins, alkaloids and terpenoids contained in T1LLPS (Benslama et al., 2017). Some research has demonstrated that the composite coating containing lotus leaf extract can inhibit the growth of microorganisms in wolfberries during storage and reduce their weight loss, making it an effective natural preservative (Fan et al., 2019). Therefore, adding T1LLPS to GWLB has a certain inhibitory effect on the growth of microorganisms, and among them, adding 1% T1LLPS shows the best effect.

The sensory evaluations for the four types of GWLB with different concentrations of T1LLPS are presented in Table 4. The sensory evaluation results were the best for the GWLB with 0.5% T1LLPS added. However, the sensory evaluation results were not ideal for the GWLB with 1% T1LLPS added. The overall acceptability scores of GWLB added with 0.25 and 0.5% T1LLPS were both 8.4, which were higher than that of the control group as 7.6. In contrast, the score for GWLB with 1% T1LLPS was 7.2. This phenomenon may be attributed to the concentration-dependent effects of specific bioactive compounds present in T1LLPS (Shahrajabian et al., 2022; Zheng et al., 2022).