Optimization of formulation and process conditions of chestnut-based functional beverage using response surface methodology

Abstract

Response surface methodology was used to optimize the chestnut beverage production under the effect of the independent variables including dilution rate (x₁), dilution temperature (x₂), pasteurization time (x₃) and pasteurization temperature (x₄). The experiments were based on a central composite design with linear and quadratic models employed to study the combined effects of four independent variables. The responses were selected with functional properties such as antioxidative attributes and total phenolic content. The optimal conditions (x₁, x₂, x₃ and x₄) determined for development of chestnut based functional beverage were a dilution rate of 25.19 g/100 mL, a dilution temperature of 37.562 °C, a pasteurization time of 24.996 min. and a pasteurization temperature of 84.433 °C. After comparing the predicted and experimental results, the multi-response surface methodology was more stable with a good correlation for a functional chestnut beverage.

Introduction

Increased attention towards functional foods has been supported by recent research findings regarding their nutritional value and beneficial health effects, as well as improved knowledge and awareness of consumers. Chestnuts are among functional foods that are gaining popularity, as more scientific manuscripts involving chestnuts support their nutritional importance, and role as part of a daily diet that is important for human well-being (Schlörmann et al. 2015; Santos Rosa et al. 2019; Chang et al. 2020; Nguyen 2020).

The chestnut belongs to the Fagaceae family and the Castaneae genus. C. sativa, C. molissima, C. crenata and C. dentata are a widespread species and known as European, Chinese, Japanese, and American chestnuts, respectively. The nutritional composition of chestnuts can change according to genetic factors, geographic locations (climate and/growing conditions), harvesting time and processing. Chestnuts are widely recognized for their nutritive value, distinctive flavour and taste. These nuts consist of moisture (34.2–63.6%), carbohydrates (20–32%), protein (4–7%), fat (2–4%), ash (0.8–4.4 g/100 g DW) and dietary fiber (4–10%). Despite the low fat content, it contains healthy unsaturated fatty acids like linoleic and linolenic acids. It is considered to be a high-quality vegetable protein due to the presence of important amino acids like aspartic acid, glutamic acid, arginine and γ-aminobutyric acid (GABA). It also is an excellent source of vitamins (particularly vitamins E and C) and important minerals (Ca, P, K, Mg and S). Chestnuts have also been described as containing significant levels of antioxidant compounds (L-ascorbic acid, vitamin E, carotenoids and phenolic compounds) and organic acids (oxalic, citric, glutamic, tartaric, pyruvic, ascorbic, malic, quinic, succinic, shikimic, and fumaric acids). As chestnut does not contain gluten, thus, its flour can be used instead of cereal-based foods (Mert and Erturk 2017; Yang et al. 2018; Singh and Rehman 2019; Xu et al. 2020).

The possible health-promoting effects of chestnut consumption include cardioprotective, neuroprotective, antiinflammatory, antidiabetic, antioxidant, antimicrobial, anticarcinogenic and prebiotic properties, due to its bioactive compounds such as essential fatty acids, γ-aminobutyric acid (GABA), vitamin E, phytosterols, dietary fibre, gallic and ellagic acids (Rosa et al. 2019). The chestnut can be consumed fresh or roasted, boiled, fryed, steamed and pureed, it can be used as a funtional ingredient in jams, confectionary, purees, syrup, puddings, canned, icecream, salad, dessert, stuffing, soup, bread, cake, alcoholic beverages, frozen and dried products. Chestnut consumption is gaining more popularity in daily diets due to factors such as nutrient richness, absence of glüten, cholesterol-free and health benefits, thus further studies in food technology and health status have beeen aimed at developing novel chestnut-based products and new market opportunities (Mert and Erturk 2017; Ozcan et al. 2017; Rosa et al. 2019; Singh and Rehman 2019).

As foods are very complex systems, the food industry must develop novel products or new recipes from traditional products and, formulation and processing have a critical importance for obtaining desired quality criteria and properties. An efficient production process should minimize the number of experimental trials and production costs, experiments and maximize the desired techno-functional properties of food. To achieve this, and for optimization of complex processes mathematical and statistical techniques are needed. Response surface methodology (RSM) among mathematical and statistical techniques is widely used in the best combinations of investigated factor levels or process variables (dependent and independent variables) and for developing, and optimizing formulation and process of novel foods (Murevanhema and Jiedani 2015; Mang et al. 2016). Since this technique reduces the number of experimental trials needed to evaluate multiple parameters and their interactions, it results in lower production costs and time spent (Yildiz and Dogan 2014).

Currently, researchers have focused on a positive relationship between consuming the naturally antioxidant components of foods and health. Thus, food manufacturers and consumers are interested in different foods with antioxidant potential. It has been shown that chestnuts possess antioxidant capacities and are able to scavenge radicals, capture reactive oxygen species, decompose peroxides, prevent chain initiation, bind transitional metal ions as well as reducing the power or chelating activity (Barros et al. 2011; Dinis et al. 2012; Otles and Selek 2012; Wani et al. 2017; Chang et al. 2020; Nguyen 2020; Xu et al. 2020). In light of the above information, the aims of this study were (i) to develop chestnut-based beverage as novel functional food, and (ii) to evaluate the effect of formulation and production parameters (dilution rate, dilution temperature, pasteurization time and temperature) on the increase of the antioxidative capacity and total phenolic content of the product and the optimization of the process conditions using response surface methodology.

Materials and methods

Materials

Unshelled and deskinned chestnuts were supplied from Ilka Confectionery Products and Food Industry (Bursa, Turkiye).

Methods

Preparation of functional chestnut beverage

Unshelled and deskinned chestnuts were mixed with distilled water at different ratios/temperatures (Table 1) and ground using a laboratory blender (WARING Commercial Blender 8011G, Stamford CT, USA). Then the liquid was sieved through a muslin cloth to reduce the size of the solid particles before pasteurization.

Table 1.

Experimental design for chesnut-based functional beverage

Experiment number	Dilution rate (g/100 mL) (x1)	Dilution temperature (°C) (x2)	Pasteurization time (min) (x3)	Pasteurization temperature (°C) (x4)
1	25	37.5	15	75
2	17	37.5	15	75
3	25	72.5	15	75
4	17	72.5	15	75
5	25	37.5	25	75
6	17	37.5	25	75
7	25	72.5	25	75
8	17	72.5	25	75
9	25	37.5	15	85
10	17	37.5	15	85
11	25	72.5	15	85
12	17	72.5	15	85
13	25	37.5	25	85
14	17	37.5	25	85
15	25	72.5	25	85
16	17	72.5	25	85
17	33	55	20	80
18	14	55	20	80
19	20	20	20	80
20	20	90	20	80
21	20	55	10	80
22	20	55	30	80
23	20	55	20	70
24	20	55	20	90
25	20	55	20	80
26	20	55	20	80
27	20	55	20	80
28	20	55	20	80
29	20	55	20	80
30	20	55	20	80

Experimental design

Response surface methodology (RSM) was used to optimize the chestnut beverage production under the effect of the independent variables such as dilution rate (x₁: 33; 25; 20; 17 and  14g/ 100 mL water), dilution temperature (x₂: 20 °C; 37,5 °C; 55 °C; 72,5 °C and 90 °C), pasteurization time (x₃: 10 min.; 15 min.; 20 min.; 25 min. and 30 min) and pasteurization temperature (x₄: 70 °C; 75 °C; 80 °C; 85 °C and 90 °C). During chesnut beverage production, 30 models of experiments (number of experiment = 2 ^k + 2 k + 1; k = number of factors) were determined using the Design Expert program and productions were carried out according to the experimental design table (Table 1). Central composite design (CCD) which is a five-level design, was used to determine the effects of these variables.

The quadratic equation providing the most accurate description of the relationships between dependent and independent variables was derived using the RSM and is shown as follows:

$$y=b_0+\sum _{i=1}^4{b}_i{x}_i+\sum _{i=1}^4{b}_{\mathit{ii}}{x}_i^2+\sum _{i=1}^3\sum _{j=i+1}^4{b}_{\mathit{ij}}{x}_i{x}_j$$

where y = Response (antioxidant capacity and total phenolic content), b_o = Regression coefficients for intercept (a constant), b_i = Linear coefficient, b_ii = Quadratic coefficient (square), b_ij = Interaction coefficient, x_i and x_j = Independent variables.

Antioxidant capacity assay

The samples were characterized by antioxidant capacity properties. 2,2′-azino-bis-(3-ethylbenzthiazoline-6-sulphonic acid) (ABTS) radical scavenging activity (Sahin et al. 2012) and ferric reducing antioxidant power (FRAP) of samples (Benzie and Strain 1996) were determined and the results were expressed as mg trolox equivalent (TE) per kg of sample.

Total phenolic content assay

Total phenolic content, using the Folin–Ciocalteau colorimetric method, was carried out and calculated from a standard curve of gallic acid and results were expressed as mg of gallic acid equivalents (GAE) per kg of sample (Sahin et al. 2012).

Analysis of data

The experiments were based on a central composite design with the quadratic model employed to study the combined effects of four independent variables. The dependent variables measured were antioxidative properties and total phenolic content of the chestnut beverage. The responses were selected as functional properties maximising the antioxidant capacity and the total phenolic content. Design Expert 7.0.0 software (Stat-Ease Inc., USA) was used for the statistical analysis.

Results and discussion

The functional chestnut beverage mentioned in the method section was manufactured according to the procedure based on preliminary experiments including different dilution rates. The pH of chestnut beverage ranged between 5.05 and 6.68, dry matter between 6.65 and 13.38% and ash between 0.08 and 0.26%. Protein and fat content in the dry weight of beverage ranged between 0.28–0.61% and 0.42–1.86%, respectively. Recently, health authorities and researchers have demonstrated a positive relationship between dietary intake of foods including antioxidant components and health. Therefore, consumers have tended to favor different foods with antioxidant potential (Wani et al. 2017; Sahin 2018; Yilmaz-Ersan et al. 2018; Xu et al. 2020). Xu et al. (2020) reported that chestnuts have higher antioxidant content (4.7 mmol Fe²⁺/100 g) when compared to fruits (0.4–2.4 mmol Fe²⁺/100 g), many legumes (0.11–1.97 mmol Fe²⁺/100 g), and cereal products (0.5–1.3 mmol Fe²⁺/100 g). During this research, it has been aimed that the antioxidative properties of functional chestnut-based beverage have a crucial role in the intake of antioxidant components. The experimental and predicted values are given in Table 2. The ABTS values of beverages ranged between 56.10 and 62.98 mg TE/kg sample, FRAP values between 76.31 and 228.10 mg TE/kg sample and total phenolic content between 307.80 and 642.61 mg GAE/kg sample.

Table 2.

Central composite design of factors with experimental and predicted values

Treatment	ABTS (mg TE/kg)		FRAP (mg TE/kg)		TPC (mg GAE/kg)
	Experimental	Predicted	Experimental	Predicted	Experimental	Predicted
1	62.09	62.13	183.70	178.69	558.20	525.62
2	56.97	57.74	121.15	129.95	420.15	427.78
3	61.96	62.49	218.95	221.25	524.16	537.65
4	61.54	61.29	195.84	187.54	543.24	503.75
5	62.09	61.38	197.57	201.78	614.16	599.09
6	56.10	56.71	123.66	127.30	433.48	422.58
7	60.29	60.91	221.15	228.10	542.42	531.10
8	58.51	59.43	156.83	168.65	425.46	418.53
9	62.37	61.66	214.05	213.64	574.55	571.61
10	61.52	61.59	146.98	133.42	429.58	411.39
11	56.20	56.26	201.87	191.63	513.48	494.86
12	58.45	59.38	119.24	126.44	393.37	398.57
13	62.06	62.98	225.44	227.13	632.64	642.61
14	62.94	62.63	112.06	121.17	427.08	403.71
15	57.30	56.75	186.26	188.88	503.32	485.82
16	58.96	59.60	99.56	97.96	307.80	310.87
17	60.66	61.02	214.63	215.97	624.53	642.13
18	60.71	59.47	82.46	76.31	347.53	369.33
19	60.48	60.59	180.71	178.88	488.19	511.23
20	58.91	57.91	201.19	198.22	414.07	430.42
21	62.60	62.33	165.45	177.46	436.57	459.63
22	62.42	61.80	188.88	172.07	429.07	445.40
23	60.15	59.33	204.11	194.30	495.25	523.14
24	59.11	59.03	153.55	158.55	449.97	461.48
25	59.69	60.35	215.96	200.90	413.62	443.45
26	60.76	60.35	199.81	200.90	445.58	443.45
27	58.78	60.35	187.13	200.90	449.77	443.45
28	61.27	60.35	200.75	200.90	461.61	443.45
29	61.21	60.35	208.92	200.90	448.38	443.45
30	60.40	60.35	192.82	200.90	441.72	443.45

Model fitting

In this study, four independent variables (dilution rate, dilution temperature, pasteurization time and temperature) were selected for optimization according to the results of scientific studies and preliminary experiments. In particular, dependent variables were determined as having antioxidant capacity and total phenolic content that can significantly affect nutritional value and functional properties of foods (Murevanhema and Jideani 2015; Mang et al. 2016).

ANOVA including linear, quadratic and dual effects revealed that the antioxidant capacity values and total phenolic contents were described by quadratic polynomial models (Table 3). R², regression coefficient, is described as the amount of the variation in the dependent variable referred to in the model. Generally, it has been recomended that R² should be more than 80% for a good fitting model. The antioxidant capacity showed good correlation namely R² = 0.8647 for ABTS and 0.9572 for FRAP (Table 3). The R² value for TPC was determined as 0.9450. The lack of fit tests were insignificant in all the cases which indicates that the models are adequately accurate to predict the ABTS, FRAP values and total phenolic content for any combination of independent factors in the ranges studied. All the terms in the models were significant as shown by F-ratio and p-value (p < 0.01). After comparing the predicted and experimental results, the RSM was more stable with good correlation (R² > 0.95) for chestnut beverage. This suggested that the obtained models can be used to determine the relative effect of the studied factors in order to find out the optimum parameter combinations for desirable responses and to predict the results for other conditions.

Table 3.

Analysis of variance (ANOVA) for the fitted quadratic polynomial model for optimization of production parameters and optimum conditions

ABTS (R2 = 0.8647)						FRAP (R2 = 0.9572)				TPC (R2 = 0.9450)
Source	DF	SS	MS	F	P	SS	MS	F	P	SS	MS	F	P
Model	14	93.01	6.64	6.85	< 0.0003	44,281.15	3162.94	23.97	< 0.0001	165,700	11,838.04	18.42	< 0.0001
Lack of fit	10	9.90	0.99	1.06	0.5045	1432.14	143.21	1.31	0.1720	8350.11	835.01	3.23	0.1037
Pure error	5	4.65	0.93			547.06	109.41			1291.53	258.31

Optimum conditions, predicted and experimental values of responses
Variables	Response surface optimization			A multi-response surface optimization	Predicted (experimental) values of multi-response surface analysis
	ABTS (max.)	FRAP (max.)	TPC (max.)		ABTS	FRAP	TPC
Dilution rate; x1	25.18	25.47	25.26	25.19	62.94 (62.91 ± 2.22)	226.21 (226.10 ± 1.07)	632.70 (632.02 ± 10.25)
Dilution temperature;x2	37.56	38.20	37.52	37.562
Pasteurization time; x3	24.94	21.82	24.86	24.996
Pasteurization temperature;x4	84.98	84.38	84.91	84.433

DF degrees of freedom, SS sum of squares, MS mean squares

RSM model for ABTS antioxidant capacity

In ABTS assay named as TEAC, 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt radical cation is oxidized by peroxyl radicals or other oxidants to the long-life radical anion ABTS^·+, which results in as the ability of test compounds to decrease the color reacting directly with the ABTS^·+ radical (Dong et al. 2015).

The F value for quadratic model of ABTS was significant (p ≤ 0.01). The four variables (dilution rate, dilution temperature, pasteurization time and temperature) have significant (p ≤ 0.01) effect on ABTS value. The quadratic model for ABTS can be expressed by the following equation.

$$\begin{array}{cc}\hfill \text{ABTS}=& 443.45-68.20x_1-20.20x_2-15.42x_4+15.98x_1{x}_2-19.67x_1{x}_3\hfill \\ \hfill & -15.60x_1{x}_4-20.01x_2{x}_3-22.20x_2{x}_4+15.57x_2^2+12.22x_4^2\hfill \\ \hfill \end{array}$$

where x₁ is the coded independent factor dilution rate, x₂ is a coded independent factor dilution temperature, x₃ is a coded independent factor pasteurization time and ×4 is a coded independent factor pasteurization temperature. The results have shown that the ABTS value was significantly influenced by the dilution rate, dilution temperature and pasteurization temperature as linear variables, as well as by the dilution temperature and pasteurization temperature as quadratic terms.

The effects of the factors and antioxidant capacity values determined with the ABTS method were analyzed using the three-dimensional images of response surface plots from the RSM in Fig. 1. At low dilution temperature and low dilution rate, the antioxidant capacity value is at its highest value. At a constant dilution temperature of 90 °C, while the dilution rate increases, the antioxidant capacity value also increases. When experiments 9 and 11 were compared (Table 2), the total antioxidant capacity by the ABTS method increased from 56.20 to 62.37 mg TE/kg chestnut beverage at a lower dilution temperature. And also, when experiments 9 and 12 were compared, the total antioxidant capacity by the ABTS method increased from 58.45 to 62.37 mg TE/kg chestnut beverage at a lower dilution rate. In particular, lower dilution rate and dilution temperature are more effective on the higher ABTS values. Sahin (2018) reported that the extraction temperature used in antioxidant capacity assay was an important parameter and caused both the degradation and loss of the antioxidant compounds and reactions with other compounds especially at higher temperature. As the dilution rate increases, the chestnut content per unit volume will increase (carbohydrate, protein, fat), so an interference effect may occur in antioxidant capacity measurements. Components such as protein, carbohydrate and fat can have an antagonistic effect in antioxidant capacity measurements.

Fig. 1.

Response surface plots showing the effects of formulation and production parameters A Dilution rate and temperatue, B Dilution rate and pasteurization time, C Dilution rate and pastaurization temperature, D Dilution temeperature and pasteurization time, E Dilution temperature and pasteurization time on ABTS

For chestnuts, various researchers reported that the ABTS values were 4.77–8.15 µmol TE/g DM (Neri et al. 2010), 0.564–1.048 mmol TE/kg (Barros et al. 2011) 5.2–14.1 mg/g (EC₅₀) (Dinis et al. 2012) and 6.53–24.83 µmol TE/g (Xu et al. 2020). Various studies demonstrate different ABTS values due to the extraction methods (extraction solvent, time or temperature), component of foods, processing factors, packaging and storage conditions.

Suphamityotin (2011) reported that the response surface methodology (RSM) was utilized to optimize cereal milk. Incubation temperature, hydrolyzing time and concentration of enzyme were chosen as independent variables. Under the optimum condition (incubation temperature of 35 °C, hydrolyzing time 120 min, and 2% (v/w) pectinase), ABTS antioxidant activity of cereal milk increased 1.03-fold.

RSM model for FRAP antioxidant capacity

FRAP method used in the evaluation of antioxidant capacity of the samples is based on the reduction of a ferric tripyridyl triazine (TPTZ) complex to its ferrous form (Fe²⁺), in which it forms an intense blue color related to the amount of antioxidant reductants in the samples. In this method, the antioxidant components of samples act as the oxidants (Yilmaz-Ersan et al. 2018).

Quadratic model fitted to FRAP was significant (p ≤ 0.01) with high F value of 23.97. Lack of fit was insignificant relative to pure error. FRAP response was negatively influenced by the dilution rate, pasteurization temperature and time as a linear term.

$$\begin{array}{cc}\hfill \text{FRAP}=& 200.90-34.92x_1-1.35x_3-8.94x_4-6.43x_1{x}_3-7.87x_1{x}_4\hfill \\ \hfill & -16.14x_2{x}_4-16.69x_1^2-6.53x{2}_3^2-6.12x_4^2\hfill \\ \hfill \end{array}$$

Dilution rate, pasteurization time and pasteurization temperature significantly affected the FRAP (Fig. 2). When experiments 13 and 14 were compared at the same dilution temperature, pasteurization time/temperature, the antioxidant capacity value determined by FRAP method was decreased with increasing dilution rate as antioxidant capacity value determined by ABTS method. At low dilution temperature and low dilution rate, the antioxidant capacity value was at its highest value. At a constant pasteurization temperature of 90 °C, while the dilution rate decreases, the antioxidant capacity value also increases. In particular, lower dilution rate and higher pasteurization temperature were more effective on the higher FRAP values. The increase of FRAP values could have been based on the biosynthesis of antioxidant components, breakdown of complex antioxidants, processing conditions, hydrolysis of proteins, polarity, and ability to donate atoms and electrons of antioxidant biofactors.

Fig. 2.

Response surface plots showing the effects of formulation and production parameters A Dilution rate and pasteurizaton time, B Dilution rate and pasteurization temperature, C Dilution and pasteurization temperature on FRAP

Some studies determined that the FRAP values of chestnuts ranged from 6.6 to 14.6 mg/g (EC₅₀) (Dinis et al. 2012), from 9.08 to 14.15 mM FeSO₄/g DM (Otles and Selek 2012) and from 0.66 to 2.37 mmol FE/100 g (Xu et al. 2020).

RSM model for total phenolic content

Dietary phenolics known as secondary metabolites found in foods have a critical importance due to both antioxidative properties and beneficial health effects such as decreased of cardiovascular diseases, type 2 diabetes, anticancer mechanisms and antiinflammation (Sahin et al. 2012; Yilmaz-Ersan et al. 2018). The fresh chestnut fruits contains gallic and ellagic acids as the main dietary phenolics, in addition to including syringic, caffeic, vanillic acids, catechin, chlorogenic acid, p-coumaric acid, ferulic acid vescalagin, castalagin, and tannins (De Vasconcelos et al. 2010; Otles and Selek 2012).

The quadratic model significantly (p ≤ 0.01) fitted to the total phenolic content with a high F value of 18.42. The quadratic model for the total phenolic content can be expressed by the following equation. Total phenolic content of the beverages was significantly influenced by the dilution temperature and pasteurization time.

$\begin{array}{cc}\hfill \text{TPC}=& 60.35-0.67x_2-0.13x_4+0.80x_1{x}_2\hfill \\ \hfill & +1.08x_1{x}_4-1.44x_2{x}_4+0.43x_4^2\hfill \\ \hfill \end{array}$

Total phenolic content increased as function of lower dilution rate/temperature and higher pasteurization temperature (Fig. 3). When experiments 5 and 13 were compared at the same dilution rate/temperature and pasteurization time, the total phenolic content increased from 599.09 to 642.61 mg GAE/kg chestnut at higher pasteurization temperature. Pasteurization temperature is an important parameter in terms of total phenolic content. At this temperature, the contribution to the total phenolic content may have increased as a result of the hydrolysis of some phenolic substances.

Fig. 3.

Response surface plots showing the effects of formulation and production parameters A Dilution rate and temperature, B Pasteurization temperature and dilution rate, C Pasteurization and dilution temperature on TPC

Fumić et al. (2019) reported that high temperature may also lead to the degradation of sensitive phytochemicals such as phenolics. Nguyen (2020) stated that raw chestnut (Castanea sativa) contained 49.28 mg/g total phenolic, while chestnut roasted at 135 °C for 10 min had 27.81 mg/g total phenolic. Also, the different storage temperature (4 °C and 30 °C) affected the total phenolic contet dried chestnut. However, Gonçalves et al. (2010) observed that boiled chestnuts showed higher fat, soluble fibre, gallic and ellagic acids and total phenolic contents than raw chestnuts.

Related studies reported that the total phenols values of chestnuts were 15.80–22.69 mg GAE/g FW (De Vasconcelos et al. 2007); 7.66–18.30 mg GAE/g FW (De Vasconcelos et al. 2010); 13.6–18.8 mg GAE/g DM (Gonçalves et al. 2010); 0.072–0.517 mg GAE/g DM (Neri et al. 2010); 3,61–3.63 mg GAE/g extract (Carocho et al. 2012); 9.6–19.4 mg GAE/g DM (Dinis et al. 2012); 5.00–32.82 GAE/g DM (Otles and Selek 2012); 1.60–3.30 g GAE/100 g (Wani et al. 2017); 42.8–58.6 mg GAE/100 g FW (Chang et al. 2020) and 1.03–2.19 mg GAE/g (Xu et al. 2020). Researchers reported that the total phenolic content of chestnut may vary depending on the several factors, such as the geographical factors (temperature and sunlight exposure etc.), soil type, precipitation, altitude, processing (raw, boiled and roasted etc.), storage, packaging and assay system.

Optimization of level of independent variables

In order to optimize the level of independent variables, the responses including ABTS, FRAP and total phenolic content were assigned equal importance on the basis of their effect on nutritional quality of the chestnut beverage. Validation tests were used to verify the reliability of the model by comparing the experimental and the predicted values for the the multiresponse surface (MRS) methodology. The optimum production parameters were presented in Table 3. The dilution rate of 25.19 g/100 mL, dilution temperature of 37.562 °C, pasteurization time 24.996 min. and pasteurization temperature of 84.433ºC produced the maximum antioxidant capacities (62.94 mg TE/g for ABTS and 226.209 mg TE/g for FRAP) and the total phenolic content (632.699 mg GAE/g) for chestnut beverage. The experimental values antioxidant capacities (62.91 ± 2.22 and 226.10 ± 1.07 mg TE/kg for ABTS, FRAP respectively) and the total phenolic content (632.02 ± 10.25 mg GAE/kg) did not differ from the predicted value. The excellent correlation (within 95% confidence level) between the predicted and experimental values verifies the model validation. This model could be used to find out an optimised condition of independent variables and large scale production of chestnut beverage with desirable nutritional quality. Additonally, by using this model the processing difficulties and increased cost of raw material and energy could have been prevented, which may result from higher dilution rate, dilution temperature, and pasteurization time.

Mang et al. (2016) studied the optimization of vegetable milk extraction from Mucuna pruriens flours using central composite design with temperature (25 to 95 °C), extraction time (6 to 74 min.) and water to flour ratio (6 to 24 mL/g). The protein and sugar contents, extraction yield of protein and dry matter were choosen as the responses. The optimization conditions for vegetable milk were temperatures between 63 and 66 °C, a water to flour of ratio 12–13 mL/g and an extraction time of 57–67 min. Milk prepared by hydrating Bambara ground nut flour using the optimal hydration time (2–6 h) and temperature (25–35 °C) was determined using a central composite rotatable design for the effect of variety on pH, color and antioxidative activity (DPPH), and consumer acceptability (Murevanhema and Jiedani 2015). They reported that the optimal hydration time and temperature were estimated to be 2 h at 25 °C. Oyedeji et al. (2018) studied functional soymilk by optimizing the sprouting conditions of soybeans using response surface methodology. Soaking (12–24 h) and germination times (48–96 h) were choosen as independent variables. Optimized conditions (12 h soaking and 52 h germination) obtained are adequate in the production of soymilk with improved nutritional and quality attributes such as total proteins, phenolics and color change.

Conclusion

In the present study, the optimum formulation and process conditions (dilution rate, dilution temperature, pasteurization temperature and pasteurization time) were determined by the maximum antioxidant capacity and total phenolic content for manufacturing a functional chestnut beverage. The optimum parameters  25.19 g/100 mL for dilution rate, 37.562 °C for dilution temperature, 24.996 min for pasteurization time and 84.433 °C for pasteurization temperature were predicted by multi-response surface methodology. The results have shown that the optimization parameters (dilution rate, dilution temperature, pasteurization time and temperature) strongly influenced the process conditions of chestnut-based functional beverage. The multi-response surface methodology was used successfully for the optimization conditions of functional chestnut beverage production, which may provide an innovative approach for functional food at commercial scale. Further research should be carried out on the micro and macronutrient components of the chestnut beverage representing a novel food for functional food segment.

Abbreviations

RSM

Response surface methodology

CCD

Central composite design

MRS

Multi response surface

ABTS

2,2′-Azino-bis-(3-ethylbenzthiazoline-6-sulphonic acid

FRAP

Ferric reducing antioxidant power

TPC

Total phenolic content

TE

Trolox equivalent

GAE

Gallic acid equivalent

Author contributions

L.Y.E., B.U.G. and S.S. planned the main ideas and designed this study. Laboratory studies were carried out by B.U.G. All of results obtained were evaluated equally by all authors. L.Y.E., B.U.G. and S.S. contributed to prepare to the published version of the manuscript.

Funding

The authors received no financial support for the research, and publication of this article.

Data availability

The authors confirm that the data supporting. Data and material are available within the article.

Code availability

This research does not include code availability.

Declarations

Conflict of interest

We declare that there are no conflicts of interest.

Consent to participate

This research does not include consent statements.

Consent for publication

This research does not include consent for publication.

Ethical approval

This research does not include ethics approval.