2026 年 19 巻 3 号 p. 119-131
This study aimed to determine the optimal mixture of sucrose, glucose syrup, and Suji leaf extract to produce Suji milk candy using Design Expert 11 with Mixture D-Optimal and evaluated the physical, chemical, and sensory properties of the candy. The optimization process resulted in an optimal formulation with the highest desirability value (0.811), identified in the sample formulated with 14 % sucrose, 5 % glucose syrup, and 3 % Suji leaf extract (A4). Verification of responses value showed that the optimum formulation of Suji leaf milk candy had levels of texture, lightness, redness, antioxidant activity, water content, and reducing sugar of 66.52 g/mm, 33.20, 3.80, 27.86 %, 6.82 %, 4.95 % respectively, and hedonic score (sensory) of color (3.97), taste (4.30), texture (4.23), overall acceptance (4.27).
Candy is a confectionery product primarily composed of corn syrup, sugar, gelling agents (Bulca et al., 2024; Efe et al., 2022; Pocan et al., 2021), and other additives (El-mahrouky et al., 2024; Maringka et al., 2024), that undergoes dissolution in water or milk and subsequent boiling until the desired concentration or caramelization stage is achieved (Onyekwelu et al., 2018; Parra et al., 2023). The characteristics of different candies are based on factors such as heating techniques, gelling agents, water content (Ozel et al., 2024; Pocan et al., 2021), overall ingredients, and boiling duration (Parra et al., 2023; Upadhyaya et al., 2023). Candies are classified into two types, hard and soft, depending on their moisture content. Soft candies, characterized by their soft and adhesive texture, enjoy widespread popularity among consumers of all ages (Gok et al., 2020; Gunes et al., 2022). The texture of soft candies significantly deepens on the gelling agent type and moisture content (Ramírez-Navas et al., 2024), typically ranging between 8 % to 22 %, therefore differentiating them from hard candies (Efe et al., 2022; Pocan et al., 2021). One of the famous variants within the soft candy category is milk candy, characterized as non-crystalline soft confectionery. It comprises ingredients such as sugar (glucose, sucrose, etc.), milk, fat, and other supplementary elements subjected to high-temperature cooking processes (Maringka et al., 2016; Silva et al., 2023).
Sucrose is the primary ingredient in milk candy production, being a common constituent of various confectionery items (Tarahi et al., 2023). Its significance in candy making is indicated by its sweetness and ability to improve texture (Efe et al., 2022; Saraiva et al., 2020). Because sucrose is a non-reducing sugar lacking free formyl and carbonyl groups, it does not actively participate in the Maillard reaction (Efe et al., 2022; Gunes et al., 2022), thus contributing to its distinct properties (Wang et al., 2022). Furthermore, sucrose acts as a preservative by inhibiting microbial growth and reducing water activity in confectionery products (Lara-Cruz et al., 2022; Sahin et al., 2019). However, an overbalance of sucrose can lead to crystallization unless adequate crystallization inhibitors are incorporated into the formulation (Ramírez-Navas et al., 2024). Consequently, supplementary ingredients must be introduced to act as such inhibitors.
As an inhibitor in confectionery production, glucose syrup, also referred to as corn syrup, is a common ingredient (Hobbs, 2009). Its presence as an inhibitor can hinder sugar crystal formation or regulate crystallization towards the desired phase (Owusu-Apenten et al., 2023; Wolf, 2016). Conversely, in the creation of soft candies, where the crystallization of sucrose is undesirable, enough glucose syrup, approximately 50–60 %, is essential. When heated to high temperatures, it tends not to form a crystal core unless the syrup solution reaches the supersaturation zone (Khvorova et al., 2018; Kinugawa et al., 2015). Previous research has also indicated that candies with lower glucose syrup and higher sucrose content maintain sucrose crystallization when stored at high temperatures (Ramírez-Navas et al., 2024).
Besides glucose and sucrose, recently the addition of natural ingredients has become necessary during the making of candy, as consumer preferences in the confectionery industry have shifted towards more natural and healthier products (Kasabova et al., 2022; Nhan et al., 2023). This evolving trend is driven by a demand for healthier and more functional confectionery items (Gok et al., 2020; Karelakis et al., 2020), leading to an increased interest in confectionery enriched with bioactive and natural components (Saraiva et al., 2020). Based on previous research, the development of bioactive and natural ingredients in soft candy products includes the addition of natural colorants (Bouphun et al., 2023; Loñez, 2021), enhancements in functional compounds such as fiber or antioxidants (Delgado et al., 2018; Gok et al., 2020), combination with fruit powders and juices (Cappa et al., 2015; da Silva et al., 2016; Tarahi et al., 2023; Urooj, 2021) and purees (Hariadi, 2020; Nhan et al.,2023; Nurhafsah et al., 2023), and enrichment with marine-derived ingredients (Bartkiene et al., 2023; Lekahena et al., 2020; Matos et al., 2022; Senadheera et al., 2023). Recent studies have shown that combining candy with red dragon fruit (Sulistyowati et al., 2023), grape skin powder (Cappa et al., 2015) or Moringa oleifera leaf powder (Elisanti et al., 2022) has contributed to increasing functional and healthier milk candy products.
To contribute to increased preferences for healthier candy, this study focuses on the addition of Suji (Pleomele angustifolia) leaf as a natural ingredient in the production of milk candy. Suji leaves have antioxidant properties (Indrasti et al., 2018). An earlier analysis demonstrated that Suji leaves contain 3774.9 ppm of chlorophyll, comprising 2524.6 ppm of the chlorophyll a and 1250.3 ppm of the chlorophyll b (Martins et al., 2023). Chlorophyll, with its conjugated double bonds, is believed to contribute to the antioxidant activity observed in Suji leaves (Indrasti et al., 2018; Martins et al., 2023; Yang et al., 2025). Furthermore, the Suji leaves contain organic compounds such as triterpenoids, flavonoids, and tannins, also recognized for their antioxidant capabilities (Paul et al., 2025; Said et al., 2024). Additionally, they serve as a natural coloring agent, giving a green hue color to the presence of chlorophyll compounds in the chloroplast (Aryanti et al., 2017; Indrasti et al., 2018; Murtini et al., 2021; Putri et al., 2021; Rahayuningsih et al., 2018).
Addition of Suji leaves on the soft candy is needed to know for precise formulation to consumer acceptable in the term of the taste, texture, flavor, and appearance of such a healthier products (Gok et al., 2020; Konar et al., 2022). This needs to be considered because it can have a major impact on the properties of soft candy cause of adding each various ingredient and modification process
Due to the complexity involved in material formulation, a mixture design optimization approach using computer-based optimization techniques was taken for the study. Design Expert, a software tool utilized for product or process optimization, serves as the primary method for determining the main responses resulting from multiple variables and the subsequent mapping of the optimized responses. Design Expert has the advantage of higher flexibility in determining the limits of more than two responses, together with a level of numerical accuracy of up to 0.001 in determining a mathematical model suitable for optimization (Chakraborty et al., 2013; Souiy et al., 2023). In this study, Design Expert was employed for mixture formulation, utilizing the D-optimal mixture design method. This methodology seeks to ascertain the ideal combination of components capable of generating a final product which exhibits more desirable properties (Azarbad et al., 2019). Compared to other designs, the D-optimal design requires a smaller number of experimental runs, thereby reducing overall experimentation costs. Moreover, it allows for the incorporation of combined mixture and process variables within the same experimental design (Karoui et al., 2023). The D-Optimal mixture design facilitates variations in the concentration of each component with predefined numbers, simplifying the formulation process for researchers (Şahin et al., 2016).
This research aimed to identify the optimal formulation involving the combination of three components—sucrose, glucose syrup, and Suji leaf extract—to produce a high-quality product. The resulting candy underwent analysis to assess its physical, chemical, and sensory properties. This formulation can be applied to create milk candy with Suji leaf as a natural ingredient and acceptable to consumer preferences.
The initial step involved identifying the variables subject to change, specifically sucrose, glucose syrup, and Suji leaf extract, each defined by its upper and lower limits. The use of interval concentrations of 5–15 % sucrose, 5–15 % glucose syrup, and 1–3 % Suji leaf extract was based on preliminary research on soft candy making recipes from several articles (Abella et al., 2025; Gunes et al., 2022; Ramírez-Navas et al., 2024; Samakradhamrongthai et al., 2021; Sulistyowati et al., 2019). Sucrose ranges from 5 % to 15 %, glucose syrup from 5 % to 15 %, and Suji leaf extract from 1 % to 3 %, provided that the total combination of these three components amounts to 22 %. The chosen responses encompassed physical attributes (texture and color), chemical properties (antioxidant activity, water content, and reducing sugar content), and sensory evaluations (taste, color, texture, and overall perception). These variables, along with constraints and responses, were inputted into the software, resulting in the generation of 18 samples for analysis, as shown in Table 1.
| Run of samples | Sucrose (%) | Glucose syrup (%) | Suji leaf extract (%) |
|---|---|---|---|
| A1 | 5 | 14 | 3 |
| A2 | 15 | 6 | 1 |
| A3 | 10.5 | 10.5 | 1 |
| A4 | 14 | 5 | 3 |
| A5 | 15 | 6 | 1 |
| A6 | 8 | 12.5 | 1.5 |
| A7 | 15 | 5 | 2 |
| A8 | 14 | 5 | 3 |
| A9 | 9.5 | 9.5 | 3 |
| A10 | 10 | 10 | 2 |
| A11 | 12 | 7.5 | 2.5 |
| A12 | 6 | 15 | 1 |
| A13 | 5 | 15 | 2 |
| A14 | 6 | 15 | 1 |
| A15 | 5 | 14 | 3 |
| A16 | 15 | 5 | 2 |
| A17 | 10 | 10 | 2 |
| A18 | 7.5 | 12 | 2.5 |
Conventional solid–liquid extraction was used as extraction method (Fitri et al., 2025) with some modification. Suji leaves are local raw material picked in Jember Regency Indonesia, they should be leafy green and mature enough (the leaves picked after the appearance of five tops of young leaves from the plant). Suji leaves 100 g was extracted with 100 mL distilled water (1:1, w:v), blended using a blender (Philips N.V., Netherlands) at 5,000 rpm for 10 min, then macerated for 30 min at room temperature (25 °C). The sample was filtered to remove solids. The filtrate was evaporated using a rotary evaporator (Heidolph Expert HL63, Heidolph Scientific Products GmbH, Germany) under vacuum at 44.5 °C until 50 mL Suji leaf extract was produced.
2.3. Suji milk candy making processThe outcomes from the Design Expert 11 software resulted in the various formulations of sucrose, glucose syrup and Suji leaf extract to make the milk candy. Moreover, other compositions such as 64 % of fresh cow’s milk, 10 % of full cream milk powder, and 4 % of margarine, while total combination of sucrose, glucose syrup, and Suji leaf extract of 22 % were mixed respectively. The percentage of fresh cow’s milk, cream milk powder, and margarine was based on Sulistyowati et al. (2019). The process involved blending the fresh cow's milk, full cream milk powder, and Suji leaf extract, followed by heating at ±80–85 °C for 7 min. Subsequently, sucrose, glucose syrup, and margarine were added and stirred while heating to a temperature of 118–120 °C for 10 min until the firm ball stage was reached. Finally, the candy was poured onto a baking sheet for cooling and then cut into 2 × 1 × 1 cm3 shapes.
2.4. Physical, chemical, and sensory analysis 2.4.1. Texture profile analysisThe texture was carried at test speed of 1.00 mm/s with a penetration distance of 5.00 mm using a rheotex (SD-700, Sun Scientific Co. Ltd., Japan). The results were reported in terms of mean and standard deviation from five replicates of each Suji milk candy variation.
2.4.2. Color profile analysisColor measurements (L* and a*) were measured using a color reader (CR-10, Minolta Co. Ltd., Japan) (Hutchings, 1999). All color values were measured at six different points for each Suji milk candy formulation.
2.4.3. Water contentThe water content was analyzed with the gravimetric method using oven (UN55, Memmert GmbH, Germany) at 105 °C as described by the AOAC method (2005).
2.4.4. Determination of antioxidant activityDPPH radical scavenging activity was previously measured by Yamaguchi et al. (2000). Suji milk candy 2 g was extracted with ethanol until a volume of filtrate about 25 mL. The 1 mL sample extract was mixed with 2 mL of ethanol and 3 mL of DPPH solution (400 µM), and allowed to stand in the dark for 30 min. Absorbance was immediately measured at 517 nm. The inhibition percentage from the DPPH radical scavenging activity was calculated using the following formula.
| (1) |
Where IP is the inhibition percentage (%), A0 is the absorbance of the control and As is the absorbance of the sample. All samples were analyzed in three replicates.
2.4.5. Reducing sugar contentThe reducing sugar content of the Suji milk candies was determined using the Nelson–Somogyi method. Intensely blue Molybdenum Blue complex from the interaction of Nelson A-B, arsenomolybdate reagent, and each sample was measured spectrophotometrically at λ 540 nm (UV-Vis Spectrophotometer, Shimadzu Corp., Japan) to determine the sugar concentration against a standard curve. All the samples were analyzed in three replicates.
2.4.6. Sensory evaluation of the milk candyAdditionally, sensory evaluations covering taste, color, texture, and overall acceptability were conducted using a hedonic assessment method (likeability), on a scale of 1 to 7 and involving 30 semi-trained panelists. Panelist selection was made through training in basic taste recognition (sweet, sour, salty, and bitter), basic color (there is a color difference to no color difference), texture (soft to hard), and overall acceptance (a combination of taste, color, and texture). Those who passed the training test were selected as panelists. The sensory evaluations were conducted in a sensory booth designated according to BSN (2006). The panelists assessed the samples marked with random three-digit codes and completed a questionnaire based on their level of preference according to the specified score. The scores used in the trial were 1: intensely dislike, 2: dislike, 3: slightly dislike, 4: neutral, 5: slightly like, 6: like, and 7: extremely like (BSN, 2006).
2.5. Statistical analysisThe physical and chemical properties results were reported as mean and standard deviations from three replicates of each milk candy formulation. The statistical analysis was conducted using SPSS 11.0 and Duncan’s multiple range test (DMRT) method. Design Expert 11 was used to determine the optimized mixture of sucrose, glucose syrup, and Suji leaf extract for milk candy using RSM. Significance of p < 0.05 was accepted in all cases.
2.6. Response analysis using Design ExpertThe data input results were analyzed using the DX11 software program for Design Expert 11. The program evaluated the input data results for each response across all formulations and presented ANOVA outcomes. Finally, it illustrated the data via a contour-plot graph. The selected model is the one with the highest order polynomial equation. A response variable is considered significantly different at the 5 % significance level if the ‘prob > F’ value of the analysis result is less than or equal to 0.05. Response variables with significantly different ANOVA analysis results can be used as predictive models because the test variable has a significant influence on the response of the candy formula (Challa et al., 2022).
The following requirements must be met during ANOVA response analysis.
(1) The response model must be significant.
(2) Lack of fit must be insignificant.
(3) The adjusted R2 and predicted R2 must differ by less than 0.2. A difference between the adjusted R2 and predicted R2 of below 0.2 indicates that the predicted R2 value supports the adjusted R2 value.
(4) Adequate precision should be greater than 4, which indicates a high signal-to-noise ratio and that the data can be used in the optimization process.
The next step in the DX11 program was to display the model deemed most suitable in a contour plot, a two-dimensional (2D) or three-dimensional (3D) graph. The Design Expert 11 program also provides a residual normality plot graph (normal plot residual) that indicates whether the residuals (the difference between the actual and predicted responses for each response) follow a normal (straight) line (Sargiotis, 2025).
2.7. Formula optimizationThe process of optimizing formulas is initiated by establishing clear objectives for optimization. Utilizing the DX 11 program, the optimal solution for the candy formula was determined, providing predicted values for each response. The target optimization value that can be achieved is also called the desirability value, indicated by a value of 0 to 1. The higher this value, the higher the suitability of the candy formula obtained to achieve the optimal formula with the desired response variable. At the formula optimization stage, the optimized components are adjusted to the desired target. Furthermore, the optimization target is weighted in line with the desired purpose. This weighting is called importance which can be selected from 1 (+) to 5 (+++++), depending on the importance of the related response variable. A greater number of positive signs given indicates a higher level of importance of the response variable. In the following stage, the Design Expert 11 program provided an optimum candy formula solution and then proceeded to the verification stage to check the correctness of the formula and equations obtained. The optimum formula solution was also equipped with a predicted value of each response so that its suitability could be seen at the verification stage.
2.8. Verification of optimum formulaThe verification stage was conducted by remaking the product according to the optimum formula recommendations given by Design Expert 11. This was done to obtain the actual value of each response from the recommended optimum formula. The tests conducted in the verification stage concerned texture, color, antioxidant activity, water content, reducing sugar, and hedonic rating tests on the four sample attributes (color, taste, texture, and overall acceptability).
The texture of Suji milk candy was analyzed using a rheotex instrument and ranged from 53.6 to 69.8 g/mm (Table 2). Formula A1 (5 % sucrose, 14 % glucose syrup, and 3 % Suji leaf extract) had the lowest texture value at 53.6 g/mm, whereas formula A7 (15 % sucrose, 5 % glucose syrup, and 2 % Suji leaf extract) had the highest at 69.8 g/mm.
| Various of samples (%) | Texture (g/mm) | Color | Antioxidant activity (% inhibition) | Water content (%) | Reducing sugar content (%) | |
|---|---|---|---|---|---|---|
| L* (lightness) | a* (redness) | |||||
| A1 (5:14:3) | 53.60 ± 0.88 a | 30.67 ± 0.23 b | 3.50 ± 0.12 b | 28.26 ± 0.01 hi | 11.03 ± 0.01 i | 5.19 ± 0.11 jk |
| A2 (15:6:1) | 67.12 ± 0.97 i | 38.57 ± 0.12 hi | 4.30 ± 0.23 f | 24.43 ± 0.11 bc | 5.88 ± 0.02 a | 4.96 ± 0.23 ef |
| A3 (10.5:10.5:1) | 60.84 ± 0.96 f | 40.50 ± 0.34 ik | 5.20 ± 0.32 h | 24.23 ± 0.22 b | 6.89 ± 0.03 bc | 4.81 ± 0.21 c |
| A4 (14:5:3) | 62.60 ± 0.53 h | 32.20 ± 0.33 de | 3.27 ± 0.11 ab | 30.56 ± 0.02 k | 6.59 ± 0.02 bc | 4.82 ± 0.12 cd |
| A5 (15:6:1) | 68.00 ± 0.11 j | 39.87 ± 0.45 j | 4.43 ± 0.12 fg | 24.62 ± 0.01 c | 5.88 ± 0.11 a | 4.99 ± 0.33 f |
| A6 (8:12.5:1.5) | 58.28 ± 0.35 cd | 35.20 ± 0.43 f | 4.70 ± 0.13 g | 25.70 ± 0.10 d | 7.51 ± 0.21 de | 5.06 ± 0.35 h |
| A7 (15:5:2) | 69.80 ± 0.46 l | 35.77 ± 0.34 g | 3.63 ± 0.22 bc | 25.96 ± 0.11 de | 6.45 ± 0.22 bc | 4.91 ± 0.34 de |
| A8 (14:5:3) | 62.00 ± 0.81 g | 32.23 ± 0.44 de | 3.13 ± 0.31 a | 29.67 ± 0.03 j | 6.52 ± 0.02 bc | 5.10 ± 0.23 i |
| A9 (9.5:9.5:3) | 59.08 ± 0.44 e | 30.87 ± 0.33 bc | 3.30 ± 0.22 ab | 28.07 ± 0.03 h | 7.24 ± 0.12 d | 4.83 ± 0.41 d |
| A10 (10:10:2) | 58.68 ± 0.38 de | 34.70 ± 0.35 e | 3.83 ± 0.33 cd | 26.85 ± 0.05 fg | 6.87 ± 0.13 bc | 4.77 ± 0.23 b |
| A11 (12:7.5 :2.5) | 61.84 ± 0.80 g | 32.10 ± 0.11 d | 3.70 ± 0.23 bc | 27.11 ± 0.01 fg | 6.84 ± 0.12 bc | 4.95 ± 0.35 ef |
| A12 (6:15:1) | 57.04 ± 1.00 cd | 37.83 ± 0.23 h | 5.23 ± 0.23 hi | 23.79 ± 0.09 a | 8.24 ± 0.13 f | 5.20 ± 0.11 k |
| A13 (5:15:2) | 56.68 ± 0.56 c | 34.70 ± 0.53 e | 3.80 ± 0.33 c | 24.94 ± 0.01 cd | 8.46 ± 0.11 g | 5.03 ± 0.12 g |
| A14 (6:15:1) | 57.52 ± 0.45 d | 38.93 ± 0.44 i | 5.80 ± 0.23 i | 23.91 ± 0.02 ab | 8.27 ± 0.13 fg | 5.22 ± 0.14 l |
| A15 (5:14:3) | 54.08 ± 0.92 b | 29.37 ± 0.25 a | 3.50 ± 0.22 b | 28.33 ± 0.02 i | 9.83 ± 0.15 h | 5.17 ± 0.16 j |
| A16 (15:5:2) | 69.04 ± 0.96 k | 35.17 ± 0.24 f | 4.07 ± 0.21 e | 26.02 ± 0.66 e | 6.34 ± 0.14 b | 4.94 ± 0.70 e |
| A17 (10:10:2) | 60.00 ± 1.00 f | 36.27 ± 0.34 g | 3.93 ± 0.11 cd | 26.73 ± 0.99 f | 6.92 ± 0.16 c | 4.73 ± 0.23 a |
| A18 (7.5:12:2.5) | 57.84 ± 0.80 d | 31.83 ± 0.23 c | 4.30 ± 0.17 f | 27.37 ± 0.09 g | 7.73 ± 0.15 e | 4.78 ± 0.34 bc |
Suji milk candy formulations made from sucrose, glucose syrup, and Suji leaf extract with various concentrations (%).
The different letters in the same column stated a statistical difference (p < 0.05).
Based on the analysis by the Design Expert 11 program, the polynomial model of the texture response was a special quartic model, with the graphical representations shown in Fig. 1. The results of the analysis of variance show that the special quartic model significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0001) and an insignificant lack of fit at 0.1409.
The visual representation highlights the highest response value shown in orange, with the lowest value indicated in dark blue. Variances in surface elevation signify distinct response values corresponding to different combinations of components. Areas with low elevation indicate a lower texture response value, while those with high elevation signify a higher response value. On analyzing the obtained data, the samples formulated with higher sucrose content exhibited elevated texture values, while those with increased glucose syrup content displayed lower values. The texture of the candy tended to become harder with higher texture values, influenced by the inherent characteristics of sucrose and glucose syrup. Excessive sucrose usage prompts easy crystallization within the candy, resulting in a harder texture. This crystallization arises when the sugar content surpasses the solvent, causing a highly concentrated state upon heating.
In contrast, a higher quantity of glucose syrup incorporated into the milk candy production yielded a softer product. The presence of glucose syrup acts as an inhibitor for sucrose crystallization by impeding the bonding of crystal molecules, thereby preventing crystal formation (Willart et al., 2022; Wolf, 2016).
3.2. Color of the Suji milk candy 3.2.1. Lightness (L*)Color measurements were analyzed using a Minolta CR-10 color reader. L* (lightness) color values ranged between 0, signifying the darkest color (black), and approximately 100, indicating the lightest color (white). The L* value of the Suji milk candy formulations ranged from 29.37 to 40.5, as shown in Table 2. Sample A15 (5 % sucrose, 14 % glucose syrup, and 3 % Suji leaf extract) had the lowest L* value (29.37), while sample A3 (10.5 % sucrose, 10.5 % glucose syrup, and 1 % Suji leaf extract) had the highest (40.5).
Based on analysis by the Design Expert 11 program, the polynomial model of the lightness (L*) response was a linear model, as shown in Fig. 2. The results of the analysis of variance show that the linear model is significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0001) and insignificant lack of fit at 0.3128.
The color gradation showed the highest response in orange and the lowest in dark blue, marking distinct surface elevations indicative of varied response values for each component combination. The downward curve identifies lower texture response values, while elevated regions indicate higher ones. Increasing the sugar and Suji leaf extract tended to lower the L* (lightness) value of the candy. The reduction in the lightness of the candy resulted from heightened color concentration due to the presence of the Suji leaf extract containing chlorophyll (Indrasti et al., 2018). Furthermore, sugar influences the L* value, contributing to browning through the Maillard reaction and caramelization processes. Sugar caramelizes when exposed to heat above its melting point, enriching the flavor and causing surface browning (Willart et al., 2014). Glucose syrup also contributes to browning via the Maillard reaction (Wolf, 2016).
3.2.2. Redness (a*)The redness (a*) values ranged from 3.13 to 5.8, as shown in Table 2. Sample A8 (14 % sucrose, 5 % glucose syrup, and 3 % Suji leaf extract) had the lowest value of 3.13, while sample A14 (6 % sucrose, 15 % glucose syrup, and 1 % Suji leaf extract) had the highest (5.8).
Based on analysis by the Design Expert 11 program, the polynomial model of the redness response was a linear model, with graphical representations shown in Fig. 3. The results of the analysis of variance show that the linear model is significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0001) and insignificant lack of fit of 0.1011. Figure 3 shows the highest response in yellow, visualized at the top, while the lowest value appears in dark blue (visualized at the below).
The a* or redness value increased as the proportion of Suji leaf extract decreased, attributed to the leaf extract imparting a green color. In the a* notation, the lower value represents the green color. Additionally, the a* value was influenced by the presence of sugar, causing a browning reaction upon heating (Wang et al., 2022; Wolf, 2016).
3.3. Antioxidant activity of the Suji milk candyThe antioxidant activity of Suji milk candy results ranged between 23.79 % and 30.56 % (Table 2). Sample A4 (14 % sucrose, 5 % glucose syrup, and 5 % Suji leaf extract) had the highest antioxidant activity (30.56 %), while sample A12 (6 % sucrose, 15 % glucose syrup, and 1 % Suji leaf extract) had the lowest (23.79 %).
Based on analysis by the Design Expert 11 program, the polynomial model of the antioxidant activity response was a special quartic model, with graphical representations shown in Fig. 4. The results of the analysis of variance show that the model was significant, with a p-value of “ prob > F” lower than 0.05 (0.0001) and insignificant lack of fit of 0.7742. The highest response value is shown in orange at the top, and the lowest value in dark blue at the bottom (Fig. 4). The figure shows a correlation in that increasing the amount of Suji leaf extract leads to a rise in antioxidant activity. Suji leaves are rich in chlorophyll and are believed to possess antioxidant properties owing to the presence of conjugated double bonds within the chlorophyll molecules (Yang et al., 2025). Suji leaves contain 3774.9 ppm chlorophyll, including 2524.6 ppm chlorophyll a and 1250.3 ppm chlorophyll b (Aryanti, 2017), together with organic compounds such as triterpenoids, flavonoids, and tannins. While the triterpenoids present in Suji leaf extract exhibit modest levels of antioxidant activity, flavonoids, and tannins, categorized as phenolic compounds, display antioxidant properties attributed to their capability to donate hydrogen atoms from hydroxyl groups to radical compounds (Kuljarusnont et al., 2025; Said et al., 2024). The heating process during the making of Suji milk candy is a dependent variable, so each treatment receives the same heating effect, and the impact of the heating process on the antioxidant activity in each treatment is the same.
The water content of Suji milk candy ranged from 5.88 % to 11.03 % (Table 2). Sample A1, formulated with 5 % sucrose, 14 % glucose syrup, and 3 % Suji leaf extract had the highest water content (11.03 %), while sample A5 (15 % sucrose, 6 % glucose syrup, and 1 % Suji leaf extract) had the lowest (5.88 %).
Based on analysis by the Design Expert 11 program, the polynomial model of the water content response was a quadratic model, with graphical representations shown in Fig. 5. The results of the analysis of variance show that the quadratic model significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0001), and insignificant lack of fit of 0.1661. The highest response is shown in yellow, at the top of the figure, and the lowest value in dark blue, visualized at the bottom (Fig. 5). Figure 5 shows that an increased addition of glucose syrup in the candy production correlates with elevated water content. Glucose syrup contributes to enhancing the candy’s viscosity due to its hygroscopic properties. Compared to sucrose, glucose syrup has a higher hygroscopic nature, facilitating greater water absorption. The hygroscopic attributes of a molecule are determined by the presence or absence of available and reactive hydroxyl bonds (Hobbs, 2009; Ramírez-Navas et al., 2024). Glucose syrup contains one aldehyde group and five reactive hydroxyl groups, while sucrose comprises eight bonded hydroxyl groups without any available or reactive hydroxyl groups (Feng et al., 2022; Wolf, 2016).
The reducing sugar content of Suji milk candy ranged from 4.73 % to 5.22 % (Table 2). Sample A14 (6 % sucrose, 15 % glucose syrup, and 1 % Suji leaf extract) had the highest content of reducing sugar (5.22 %), whereas sample A17 (10 % sucrose, 10 % glucose syrup, and 2 % Suji leaf extract) had the lowest at 4.73 %.
Based on analysis by the Design Expert 11 program, the polynomial model of the reducing sugar response was a quadratic model, with graphical representations shown in Fig. 6. The results of the analysis of variance show that the quadratic model significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0041) and insignificant lack of fit of 0.2366. Figure 6 shows the highest response value in orange color, located at the top, while the lowest value appears in dark blue at the bottom. In the graphical representation (Fig. 6), the increased proportion of glucose syrup tends to elevate the levels of reducing sugars. This increase is attributed to the reducing properties inherent in glucose syrup. Despite not being classified as a reducing sugar, when heated sucrose undergoes hydrolysis, breaking down into glucose and fructose. Both possess reducing properties, thereby impacting on the overall reducing sugar content (Lara-Cruz et al., 2022; Wolf, 2016).
Consumer preferences prioritize exceptional attributes in confectionery items, including enticing textures, flavors, and appearances (Periche et al., 2014). Therefore, having an appealing color in candy is crucial to meet what consumers want. The hedonic scores for the color of the Suji leaf milk candy revealed values ranging from 3.4 to 5.7 (Table 3). The candy with the lowest color value was sample A15 (5 % sucrose, 14 % glucose syrup, and 3 % Suji leaf extract), while that with the highest was sample A3 (10.5 % sucrose, 10.5 % glucose syrup, and 1 % Suji leaf extract).
Based on analysis by the Design Expert 11 program, the polynomial model of the hedonic color response was a linear model, with graphical representations shown in Fig. 7 (a). The results of the analysis of variance show that the linear model was significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0041) and insignificant lack of fit of 0.1257. Figure 7 (a) shows that the hedonic color graph highlights the highest response value in orange at the top and the lowest value in dark blue at the bottom. The color of Suji leaf milk candy is significantly influenced by three key components: sucrose, glucose syrup, and Suji leaf extract. While sucrose and glucose syrup contribute to a brown hue due to their change in color when heated, Suji leaf extract lends a green tint. The resulting brown tone stems from the Maillard reaction and caramelization processes. When subjected to heat above its melting point, sugar undergoes caramelization, enhancing the flavor and inducing surface browning (Mariotti et al., 2014). Additionally, glucose syrup participates in the Maillard reaction (Feng et al., 2022; Wolf, 2016). The green color is a product of the Suji leaf extract containing chlorophyll (Yang et al., 2025; Indrasti et al., 2018).
3.6.2. TasteThe sense of taste is evaluated through taste testing, which involves the response or chemical stimulation of the tongue. Taste serves as an integral parameter in sensory evaluations of food items, as products that meet consumer taste preferences will be viable for marketing. The hedonic scores for the taste of the Suji milk candy formulations ranged from 3.6 to 5.7 (Table 3), with sample A9 (9.5 % sucrose, 9.5 % glucose syrup, and 3.5 % Suji leaf extract) having the lowest score and sample A2 (15 % sucrose, 6 % glucose syrup, and 1 % Suji leaf extract) had the highest.
Based on analysis by the Design Expert 11 program, the polynomial model of the hedonic taste response was a special cubic model, with graphical representations shown in Fig. 7 (b). The results of the analysis of variance show that the special cubic model is significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0002) and insignificant lack of fit of 0.1181. The highest response for hedonic taste is shown in orange above, while the lowest response is in dark blue below (Fig. 7 (b)). The average hedonic taste scores ranged from 3.6 to 5.7, indicating panelist ratings ranging from neutral to favorable. The taste profile of the candy is influenced by its constituents; sucrose and glucose contribute a sweet or caramel flavor, while the Suji leaf extract imparts the distinctive taste of Suji leaves.
3.6.3. TextureThe texture of milk candy is one of the significant properties important for its quality. Achieving an ideal texture, neither too hard nor too soft, is crucial for dairy-based candy products. The hedonic texture scores for the Suji milk candy ranged from 3.3 to 5.5 (Table 3), with sample A9 (9.5 % sucrose, 9.5 % glucose syrup, and 3.5 % Suji leaf extract) having the lowest acceptable texture for the consumers, while sample A5 (15 % sucrose, 6 % glucose syrup, and 1 % Suji leaf extract) had the highest score for hedonic texture.
Based on analysis by the Design Expert 11 program, the polynomial model of the hedonic texture response is a linear model, with graphical representations shown in Fig. 7 (c). The results of the analysis of variance show that the linear model is significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0001), and insignificant lack of fit of 0.2200. The highest score for hedonic texture is shown in orange above, with the lowest score in dark blue below. The hedonic texture values ranged from 3.3 to 5.5, showing that the panelists gave ratings ranging from ‘somewhat dislike’ to ‘like’. The candy's texture is impacted by the quantity of sucrose and glucose utilized. Excessive sucrose results in a hardened and crystallized texture, whereas glucose syrup acts as an agent to soften the texture (Ramírez-Navas et al., 2024; Wolf, 2016). Therefore, employing glucose syrup appropriately in candy production is essential.
3.6.4. Overall acceptanceOverall acceptance by the panelists based on the hedonic scores varied from 3.5 to 5.5 (Table 3). Sample A9 (9.5 % sucrose, 9.5 % glucose syrup, and 3.5 % Suji leaf extract) had the lowest hedonic overall score, whereas the highest score was for sample A5 (15 % sucrose, 6 % glucose syrup, and 1 % Suji leaf extract).
Based on analysis by the Design Expert 11 program, the polynomial model of the hedonic overall acceptance response was a quadratic model, with graphical representations shown in Fig. 7 (d). The results of the analysis of variance show that the quadratic model was significant, with a p-value of ‘prob > F’ lower than 0.05 (0.0019), and insignificant lack of fit of 0.1819. Figure 7 (d) shows the graph for the overall hedonic sensory test, with the highest response shown in orange above and the lowest in dark blue below. The average overall acceptance scores ranged from 3.5 to 5.5, indicating that the panelists provided ratings from neutral to favorable.
| Various of samples (%) | Color | Taste | Texture | Overall |
|---|---|---|---|---|
| A1 (5:14:3) | 3.63 ± 0.02 c | 3.93 ± 1.03 b | 3.43 ± 0.62 ab | 3.97 ± 0.80 c |
| A2 (15:6:1) | 5.23 ± 0.24 fg | 5.67 ± 1.00 h | 5.10 ± 1.08 g | 5.37 ± 0.17 gh |
| A3 (10.5:10.5:1) | 5.70 ± 0.76 h | 5.43 ± 1.16 g | 5.03 ± 0.28 fg | 5.37 ± 0.43 gh |
| A4 (14:5:3) | 4.23 ± 0.61 de | 4.43 ± 0.43 cd | 4.80 ± 0.95 e | 4.67 ± 0.88 ef |
| A5 (15:6:1) | 5.30 ± 0.00 fg | 5.57 ± 0.48 gh | 5.47 ± 1.00 h | 5.47 ± 1.00 h |
| A6 (8:12.5:1.5) | 4.23 ± 0.34 de | 4.63 ± 0.75 d | 4.70 ± 0.20 de | 4.57 ± 0.70 de |
| A7 (15:5:2) | 4.47 ± 1.03 e | 4.67 ± 0.79 de | 4.57 ± 0.81 d | 4.60 ± 0.72 e |
| A8 (14:5:3) | 3.83 ± 0.30 b | 4.30 ± 1.01 c | 4.40 ± 0.36 cd | 4.31 ± 0.32 d |
| A9 (9.5:9.5:3) | 3.57 ± 0.12 ab | 3.63 ± 0.80 a | 3.30 ± 0.40 a | 3.50 ± 0.46 a |
| A10 (10:10:2) | 4.27 ± 0.53 de | 4.63 ± 0.61 d | 4.30 ± 0.68 c | 4.37 ± 1.00 de |
| A11 (12:7.5 :2.5) | 4.20 ± 1.04 de | 4.50 ± 1.00 cd | 4.73 ± 0.20 de | 4.70 ± 0.68 ef |
| A12 (6:15:1) | 5.63 ± 0.73 g | 4.63 ± 0.70 de | 4.93 ± 1.03 f | 5.10 ± 0.58 g |
| A13 (5:15:2) | 4.43 ± 0.80 de | 4.47 ± 0.58 cd | 4.30 ± 0.20 c | 4.50 ± 0.85 de |
| A14 (6:15:1) | 5.17 ± 0.55 f | 5.03 ± 0.44 f | 4.53 ± 0.10 cd | 4.73 ± 0.97 ef |
| A15 (5:14:3) | 3.40 ± 1.00 a | 3.77 ± 1.02 ab | 3.67 ± 0.83 b | 3.77 ± 0.55 b |
| A16 (15:5:2) | 4.47 ± 0.50 e | 5.00 ± 0.23 ef | 4.80 ± 0.70 e | 4.97 ± 0.79 f |
| A17 (10:10:2) | 4.40 ± 1.00 de | 4.87 ± 1.00 e | 4.80 ± 0.21 e | 4.90 ± 0.95 ef |
| A18 (7.5:12:2.5) | 4.13 ± 1.21 d | 4.70 ± 0.87 de | 4.63 ± 0.26 de | 4.73 ± 0.96 ef |
Suji milk candy formulations made from sucrose, glucose syrup, and Suji leaf extract with various concentrations (%).
The different letters in the same column stated a statistical difference (p < 0.05).
Fig. 7 3D surface graph response of hedonic score (sensory) such as color (a), taste (b), texture (c), and overall acceptance (d) based on regression model between sucrose (A), glucose syrup (B), and Suji leaf extract (C) in Suji milk candy
The optimization process aims to derive a formula yielding an optimal response value. In Design Expert 11, the optimal formula is determined by a desirability value that is highest or closest to 1. Each component or response targeted for optimization is initially subjected to constraints and weighted interests. These limitations and weighting functions streamline the optimization process to achieve the desired outcomes.
Through the Design Expert 11 program (Table 4), the optimization process resulted in an optimal formula with a desirability value of 0.811. This comprised 14 % sucrose, 5 % glucose syrup, and 3 % Suji leaf extract (A4). Based on Table 4 (at predictions), this formula is expected to exhibit a textural response value of 65.01 g/mm, lightness (L*) color value of 32.76, redness (a*) value of 3.37, antioxidant activity of 28.63 %, water content of 6.42 %, reducing sugar content of 4.95 %, with hedonic sensory values related to color of 4.01, taste of 4.50, texture of 4.49, and overall acceptable of 4.53. The optimum recommended formula (A4) then proceeded to the verification stage for further assessment.
| Response | Predictions | Verification results | 95 % CI low | 95 % CI high | 95 % TI low | 95 % TI high | |
|---|---|---|---|---|---|---|---|
| Texture (g/mm) | 65.01 | 66.52 | 64.11 | 65.91 | 61.26 | 68.76 | |
| Color | L* | 32.76 | 33.20 | 31.95 | 33.56 | 28.76 | 36.76 |
| a* | 3.37 | 3.80 | 3.01 | 3.74 | 1.86 | 4.89 | |
| Antioxidant (%) | 28.63 | 27.86 | 28.33 | 28.94 | 27.37 | 29.90 | |
| Water content (%) | 6.42 | 6.82 | 5.91 | 6.92 | 4.26 | 8.58 | |
| Reducing sugar content (%) | 4.95 | 4.95 | 4.84 | 5.06 | 4.47 | 5.43 | |
| Sensory | Color | 4.01 | 3.97 | 3.80 | 4.31 | 2.80 | 5.23 |
| Taste | 4.50 | 4.30 | 4.20 | 4.80 | 3.30 | 5.70 | |
| Texture | 4.49 | 4.23 | 4.20 | 4.80 | 3.05 | 5.93 | |
| Overall | 4.53 | 4.27 | 4.21 | 4.90 | 3.03 | 6.04 | |
CI: Confidence Interval, T: Tolerance Interval.
The recommended formula selected during the optimization phase (A4) underwent verification. This phase was intended to validate the response values predicted by Design Expert 11. The output from the verification stage displayed actual response values, which were subsequently compared with the predicted values. This resulted in the verification result values shown in Table 4, indicating that the selected or optimum formulation of Suji milk candy had texture of 66.52 g/mm, lightness (L*) of 33.20, redness (a*) of 3.80, antioxidant activity (% inhibition) of 27.86 %, water content of 6.82 %, and reducing sugar content of 4.95 %, with sensory (hedonic) scores for color of 3.97, taste of 4.30, texture of 4.23, and overall acceptance of 4.27.
The actual data obtained during the verification phase show values that deviate from the predicted ones. However, despite not being identical, the actual values remain reasonably close to the predicted ones. The results from the verification are well within the 95 % confidence and tolerance intervals. Therefore, the acquired data remain acceptable, indicating that the selected formula was sufficiently viable as the optimum one. Discrepancies between the predicted and actual values could stem from environmental influences during the sample creation and storage processes. Furthermore, potential inaccuracies in measuring response values within the laboratory might contribute to such variations.
The study outcomes were based on optimizing Design Expert 11 to determine the optimal formulation for Suji leaf milk candy, which was 14 % sucrose, 5 % glucose syrup, and 3 % Suji leaf extract. This formulation achieved an optimal response value following a comparison and verification process between the predicted and actual values. The response values for this optimal formula were a texture of 66.52 g/mm, color lightness (L*) of 33.20, color redness (a*) of 3.80, antioxidant activity (% inhibition) of 27.86 %, water content of 6.82 %, reducing sugar content of 4.95 %, and sensory (hedonic) scores for color of 3.97, taste of 4.30, texture of 4.23, and overall acceptance of 4.27.
Design Expert 11 could be applied as a mixture design optimization approach for making Suji milk candy. The candy produced in our research had physical and chemical properties that were acceptable to consumer preferences based on sensory evaluation. In terms of product diversification, Suji milk candy has functional value, is healthier product, and acceptable to consumer preferences.
In future studies, it will be necessary to evaluate its shelf-life, measure the chlorophyll content in the final product, conduct bioactive compound analysis, and assess storage stability. Those parameters should be studied for better understanding. This could enhance the effectiveness of pilot plant production and maintain the quality of Suji milk candy as a functional candy.
The authors would like to express gratitude towards the Laboratory of Functional Food and Nutraceutical, Laboratory of Innovation of Agricultural Product, and Laboratory of Chemistry and Biochemistry of Agricultural Products, Faculty of Agricultural Technology, University of Jember, Indonesia.
The authors declare no conflicts of interest.
(URLs on references were accessed on 26 May 2026.)
