Optimization for Brewing Technology of Jujube Brandy Using Response Surface Methodology

In order to obtain a proper brewing method of jujube brandy, one-factor experiment and response surface methodology were applied to get the maximum alcohol content. Using single-yeast GH and fermentate at 28°C for 20d was suggested by one-factor test. The use of a central composite design and the response surface methodology to determine the best conditions allows the optimum combination of analytical variables (yeast strains, fermentation temperature and time) to be identified: single-yeast GH, fermentation temperature of 18°C, fermentation time of 24d and the alcohol content was 38.7%vol, almost accords with the predicted data. The optimized process improved the mellow flavor of jujube brandy, which has great practical values.


INTRODUCTION
Jujube brandy, a unique brandy product in China, has a long history.Jujube brandy is produced by solid fermentation, distillation and aging using Chinese jujube as raw material.However, since mature production technology is lacking, development of jujube brandy in trade market is restricted severely, it cannot be produced as a standardized commodity.
Jujube is one of the characteristic fruit in China.The total cultivating area of jujube in China has reached 3200000 hectares by 2012, with annual output of 4.683 million tons.Hebei is a major produce place of jujube, but the development of processing technology and high value-added products need to be improved.Studies have shown that jujubes are rich in sugar and contain similar components as grapes, which means jujube are proper to produce brandy (Claus and Berglund, 2005;Li et al., 2007).
Fermentation conditions are the decisive factor of quality and flavor of liquor (Jackson, 2002).The main factors influencing the liquor aroma components include yeast strains, fermentation temperature and time (Rapp, 1998).In western countries, brandy is produced with grape juice or hide trimmings and different kinds of yeasts (Jiming and Puchao, 2004;Jijun et al., 2005).Britain liquor brewster think that the best temperature for brewing fruit wine is between 22-25°C, because low fermentation temperature could reduce the generation of higher alcohols (Huafeng et al., 2003).But for the French and German winemaker, 15-18°C is considered the best temperature for fermentation for a long time (Qianwen and Zhengjun, 2000).Daqu and solid-state fermentation are characteristic of Chinese traditional liquor production techniques (Zheng et al., 2011;Berradre et al., 2009;Zhang et al., 2013) and have recently been used in the brewing of fruit wine, bringing unique flavors and improving the quality of production (Chang et al., 2014;Fan and Qian, 2005).Most white wines in China have long fermentation time at the temperature of 25-30°C, maybe as long as 3 months (Fan and Qian, 2006;Zhu et al., 2007;Luo et al., 2008;Fan et al., 2011).
In this study, yeast strains, fermentation temperature and time were selected for one-factor experiment, then the response surface analysis test was performed to get the optimal fermentation parameters, which would obtain higher quality jujube brandy.

Brewing process of jujube brandy:
• Add equal water to shredded jujube, soak 5-6 h.
Alcohol test: Alcohol content is tested with alcohol meter.All of the analyses were performed three times.

SPME-GC-MS parameters:
Jujube brandy was diluted to 10% alcohol content by distilled water. 1 g NaCl was added to 7.5 mL of sample solution in a 20 mL sealed glass vial.Flavor compounds were exacted at 40°C for 40 min with 50/30µm DVB/CAR/PDMS fiber, then used to GC-MS analysis.
Flavor compounds of jujube brandy were detected by GC-MS (Agilent 5975 Mass Spectrometer coupled to an Agilent 7890A Gas Chromatograph, DB-WAX column, 60 m×0.25 mm ID and 0.25 µm film thickness, USA).The injector temperature was 250°C, EI source was 230°C, MS Quad was 150 °C and transfer line was 250°C.The initial temperature was 50 °C for 3 min, which was increased to 80°C at a rate of 3 °C/min.The temperature was further raised to 230°C at 5°C/min and maintained at 230°C for 6 min.The carrier gas had a flow rate of 1.0 mL/min.Samples were injected using the splitless mode.A mass range of 50-550 m/z was recorded at one scan per second.Qualitative and quantitative analysis: Flavor compounds were identified by Nist 2005 library of GC-MS.The contents of flavor compounds were quantified using an internal standard (3-octanol, 99%, Sigma-Aldrich).
m i , m s , m r represent contents of determinand, internal standard, contrast, A i , A s , A r represent peak area or peak height of determinand, internal standard, contrast, f represent correction factor.

Box-Behnken design:
Based on one-factor test, a Box-Behnken Design (BBD) with three independent factors (X 1 , yeast strains; X 2 , fermentation temperature; X 3 , fermentation time) set at three variation levels was implemented (Table 1).And +1, 0, -1 encoded factors represent variables (Ni and Zeng, 2010).The alcohol content of jujube brandy was selected as the responses for the combination of the independent variables (Table 2).

RESULTS AND DISCUSSION
One-factor test results: Yeast strains: Besides ZJ Jiuqu, alcohol of jujube brandy maintain between 33 to 36% vol.Jujube brandy  fermented with single-yeast PH and mixed-yeast GHSX have higher alcohol than others (Fig. 1).Therefore, single-yeast GH, PH and mixed-yeast GHSX are proper yeast strains for brewing jujube brandy.
Fermentation temperature: Significant difference of alcohol appeared with different fermentation temperatures (p<0.05).Jujube brandy got the highest alcohol at 28°C, then at 18°C, the least at 15°C (Fig. 2).Therefore, the proper temperature for brewing jujube brandy is 28°C.
Fermentation time: Significant difference of alcohol also appeared with different fermentation time (p<0.05).Jujube brandy got the highest alcohol at 6d, then decreased gradually, which means jujube brandy got fully fermentation during 6d, then went on flavor generation reaction (Fig. 3).Therefore, although alcohol fermentation finish at 6d, for obtaining highquality-flavor jujube brandy, 20d should be chosen to be the proper fermentation time.

Box-Behnken result:
Statistical analysis and model building: Seventeen tests were complemented as Box-Behnken designing (Table 2).Regression and variance analysis was carried out to determine the coefficient of determination, lack of fit and the significance of the linear, interaction effects and quadratic of the independent variables on the response (Table 3).F-test and p-value were used to determine the significance of each coefficient (Table 3).The p-value represents the significance of the corresponding coefficients in terms of alcohol content, with a smaller p-value indicating more significant impact of the corresponding coefficient.The results of regression coefficient analysis showed that the variable with the largest effect was the quadratic term of fermentation time (X 3 2 ), followed by liner term of fermentation time (X 3 ), which were extremely significant (p<0.01).Also, the quadratic term of fermentation time (X 2 2 ) and the interaction effects of yeast strains and fermentation temperature (X 1 X 2 ) were significant (p<0.05).However, the interaction effects of yeast strains and fermentation time (X 1 X 3 ), fermentation temperature and time (X 2 X 3 ), the quadratic term of fermentation temperature (X 1 2 ), liner term of yeast strains (X 1 ) were not significant (p>0.05).
Design Expert was applied to make regression fitting analysis, the quadratic model was obtained as follows: where, Y is the predicted response (alcohol content of jujube brandy) and X1, X2, X3 are coded values of yeast strains, fermentation temperature and fermentation time, respectively.
From F-test, the low value of CV (1.96)  implies that the sample variation of 90.03% for the alcohol content of jujube brandy is attributed to the independent variables.Meanwhile, the high R 2 (0.9003), adj-R 2 (0.7720) and preR 2 (0.7143) clearly demonstrated that the experiment and the theoretical values predicted by polynomial model had a very close agreement.From the analysis, the F-value of 7.02 and p-value<0.01indicates the response surface quadratic model was significant.Furthermore, results of the ANOVA indicated that the lack of fit of 0.9359 was insignificant.

Analysis of response surface:
Perturbationplot: Perturbation plot could be used to find the most effective factors by the steep slope or   conditions and the maximum alcohol content were obtained desirability function approach was singleyeast GH, fermentation temperature of 18°C, fermentation time of 24d and the maximum alcohol content of jujube brandy was 39.905%vol with a desirability value of 0.399.Triplicate duplicate tests were performed under the optimized conditions with the mean values of 38.70±0.02%vol,which was consistent with the expected value of 39.905%vol, demonstrating that the optimized conditions agree well with the real experiments.

Quality of jujube brandy:
Under the optimum fermentation conditions, the concentration of alcohol, total acid and esters in the final product were 38.7%vol, 0.55 g/L (calculated by the content of acetic acid) and 2.35 g/L, respectively.The product had a typical characteristic of brandy.Harmful by-products of methanol were 0.034 g/100 mL.
Flavor compounds of jujube brandy: Flavor compound of jujube brandy with optimized and normal brewing method have been compared (Fig. 6).The GC-MS results demonstrated that there is a large difference between optimized and normal brewing method (Table 4).It determined that amount and content of flavor compounds in jujube brandy brewed by optimized process were higher than that of normal process, especially the esters.Such results indicated that the optimized process improved the mellow flavor of jujube brandy.

Fig. 1 :
Fig. 1: Influence of yeast and Jiuqu on the alcohol of jujube brandy indicates that the experiments are precise and reliable(Prakash Maran et al., 2013).The determination coefficient (

Fig. 4 :
Fig. 4: Perturbation plot showing the effect of process variables

Fig. 5 :Fig. 6 :
Fig. 5: Surface plots for flavor content of jujube brandy; (a): figure plot to show yeast strains and temperature; (b): figure plot to show yeast strains and time; (c): figure plot to show temperature and time curvature.Arelatively flat line means in sensitive to change in that particular factor.The response (Y) was plotted against the deviation from the reference point by changing only one factor over its entire range while holding all other factors constant (Actual Factors: Ayeast = 2.02703, B-temperature = 24, C-time = 16, Fig. 4).The relationship between the responses and the experimental variables can be clarified graphically by

Table 1 :
Independent variables and their levels used in the response

Table 2 :
Variable levels and responses of flavor content based on yeast, fermentation temperature and time

Table 3 :
Analysis of Variance (ANOVA) for response surface quadratic model for flavor content of jujube brandy and independent variables (X1, X2, X3)

Table 4 :
Flavor compound of jujube brandy with optimized and normal brewing method