- Email: [email protected]
Evaluation indicators of explosion puffing Fuji apple chips quality from different Chinese origins
LWT - Food Science and Technology 60 (2015) 1129e1135
Contents lists available at ScienceDirect
LWT - Food Science and Technology journal homepage: www.elsevier.com/locate/lwt
Evaluation indicators of explosion puffing Fuji apple chips quality from different Chinese origins Jin-feng Bi*, 1, Xuan Wang 1, Qin-qin Chen, Xuan Liu, Xin-ye Wu, Qiang Wang, Jian Lv, Ai-jin Yang Institute of Agro-Products Processing Science and Technology, CAAS, Key Laboratory of Agro-Products Processing, Ministry of Agriculture, Beijing 100193, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 May 2013 Received in revised form 11 September 2014 Accepted 1 October 2014 Available online 13 October 2014
In this study, sixteen evaluation indicators of explosion puffing Fuji apple chips from nine different origins in China were analyzed, and the most important evaluation indicators were obtained by analysis of variance (ANOVA), correlation analysis (CA), principal component analysis (PCA) and system cluster analysis (SCA). Results of ANOVA showed that quality of apple chips from different origins showed significant differences (P < 0.05). The coefficient of variation (CV) values of sixteen evaluation indicators for apple chips from nine origins were 5.39 %e56.26%. Through CA results, some indicators were correlated to each other within a certain range, such as crude fat content and crispness. Five principle components were extracted through PCA with eigenvalue of over 1, and the cumulative contribution was achieved at 91.06%. Based on the above results, five characteristic indicators of apple chip quality were obtained by SCA. These were crude fiber content, crispness, titratable acid content, production rate and expansion ratio. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Apple chips Evaluation indicator Principal component analysis Cluster analysis
1. Introduction Apple is one of the most frequently consumed fruit. In China, production of commercial apples amounted to 35 million t/year in recent years, of which most (70%) were used for direct consumption, and others (30%) were processed for jam, juice, and dehydrated products (Liu et al., 2012; Woodroof & Luh, 1975). Recently, fruit and vegetable chips have become very important in the diet of the modern consumer because the chips not only can extend shelf life but maintain the original nutrition as well (Zou, Teng, Huang, Dai, & Wei, 2012). Apple chips occupy big market shares in an increasing puffed food market because of being crispy and refreshing, sweet aroma, high nutrition and ease to carry (Dong et al., 2009). There are several processing technologies for apple chips, such as deep fat frying, vacuum frying, freeze-drying and microwave vacuum drying. Among the apple chips processing technologies, explosion puffing drying (EPD) has its unique advantages, no oxygen reduced oxidation in the processing under
Abbreviations: ANOVA, analysis of variance; CA, correlation analysis; PCA, principal component analysis; SCA, system cluster analysis. * Corresponding author. Tel./fax: þ86 10 62812584. E-mail addresses: [email protected], [email protected] (J.-f. Bi). 1 They contributed equally to the present paper. http://dx.doi.org/10.1016/j.lwt.2014.10.007 0023-6438/© 2014 Elsevier Ltd. All rights reserved.
vacuum drying condition, the processed product is natural and nutritional. EPD means that the material is puffed at higher temperatures between 80 and 130 C with higher vapor pressure of 0.1e0.3 MPa and dried in vacuum conditions with lower temperature between 60 and 90 C. A sudden increase in temperature or a sudden decrease in pressure causes water evaporation and expansion in the cells of the materials (Sullivan & Craig, 1984). A puffing process, which performs as an intermediate stage, involves the release or expansion of vapor or gas within the product, either to create an internal structure or to expand and/or rupture an existing one (Antonio, Alves, Azoubel, Murr, & Park, 2008; Hofsetz, Lopes, Hubinger, Mayor, & Sereno, 2007). Compared with the traditional deep-frying and freeze-drying technologies, the product processed with EPD has lower oil content and longer shelf life. Moreover the EPD machine is less expensive, which may promise a bright future for fruit and vegetable drying technology (Ma, Bi, & Wei, 2005). Parameters optimization of apple chips have been extensive studied (Bi, 2008; Duan & Wang, 2007; DeBelie, DeSmedt, & DeBaerdemaeker, 2000; Dorta & Piotr, 2004; Han, Li, Ma, &Zhao, 2006; Lewicki, Gondek, Witrowa-Rajchert, & Nowak, 2001; Ozilgen, Guvenc, Makaraci, & Tumer, 1995; Sullivan & Craig, 1980). However, the research on quality evaluation of apple chips was mainly about sensory evaluation, instead of quality evaluation
1130
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
models. The quality of chips includes sensory quality, physicochemical nutritional quality and processing quality, which can be evaluated by complex factors, such as product appearance moisture content and Vitamin C (Vc). Different quality factors which were closely related but were relatively independent increase the difficulty of comprehensive evaluation. Therefore, it becomes one of the major tasks to search for a simple method to evaluate the quality of apple chips. Most studies apply principal component analysis (PCA) or cluster analysis to select food evaluation indicators. For example, Bao et al. (2004) studied the selection of navel orange quality evaluation factors using PCA. The result showed that six factors including soluble solids content (or peel redness), titratable acid content, Vc content, juice yield, peel brightness (or peel yellowness), and mean individual fruit weight (or fruit shape index or peel thickness) could reflect the most information on fruit quality of Newhall orange. Xu et al. (2011) adopted cluster analysis to simplify apple quality factors. The result showed that the five most important fruit quality parameters were fruit weight or fruit shape index, fruit firmness, soluble solids content, titratable acid and fruit color. Huang, Yang, and Fang (2003) evaluated the quality of pear varieties using multidimensional value theory, which showed that Vc, soluble solids content and titratable acid were the most important quality indicators. The above studies filter quality factors only using PCA or cluster analysis separately. There is little research combining these two methods. In this study, nine different Fuji apples producing regions in China were selected to manufacture apple chips by explosion puffing drying technology. Meanwhile, sixteen evaluation indicators were measured for apple chip quality. PCA and system cluster analysis (SCA) were adopted to simplify apple chip quality evaluation indicators, which provided a scientific basis for the research on apple chip quality evaluation. 2. Material and methods 2.1. Materials 2.1.1. Fuji apples “No. 2 Changfu” is one of the most widely cultivated Fuji varieties. They were picked from nine different origins, including Aksu (Xinjiang province, 80 290 E, 41150 N), Qixia (Shandong province, 120 830 E, 37 80 N), Lingbao (Henan province, 110 850 E, 34 520 N), Luochuan (Shannxi province, 109 420 E, 35760 N), Hengshui (Hebei province, 115720 E, 37 710 N), Taigu (Shanxi province, 122 530 E, 37420 N), Yingkou (Liaoning province, 122130 E, 40 390 N), Fengxian (Jiangsu province, 116 570 E, 34790 N) and Jingning (Gansu province, 105730 E, 35 510 N). These apples were picked from trees from middle to late October in 2011. One hundred apples were picked in each location and were made into apple chips using explosion puffing drying technology.
equipment system developed by Tianjin Qin-de New Material Scientific Development Co. Ltd. (Tianjin, China). This system can be adjusted to any desired puffing temperature and vacuum pressure. It consists of puffing chamber, vacuum chamber, vacuum pump, decompression valve, air compressor, vapor generator, and manual control panel (Fig. 1). Pre-weighed apple slices (2 kg) were predried at 80 C for 120 min. Then the materials were puffed at 105 C for 10 min with a vapor pressure of 0.3 MPa and dried in vacuum conditions with a lower temperature of 80 C for 120 min (Bi, 2008). The experiment was carried out in triplicate. 2.3. Measurement of evaluation indicators 2.3.1. Color The surface color of the apple chips was measured using colorimeter (D25L, Hunterlab, Virginia, USA). Based on the CIELab color space and after calibration with the white tile and black glass, three equidistant spots were examined on the major axis of each apple chip sample. Since the spot diameter of the instrument was 10 mm, the total area of the slab was 10 cm2. Triplicate experiments were performed. Three colorimetric evaluation parameters, including lightness (L), greenered hue (a) and blueeyellow hue (b) were measured and calculated to observe color changes. (LemusMondaca et al., 2009). 2.3.2. Crispness and hardness Crispness of the apple chips was measured using a TA-XT2 texture analyzer (Stable Micro Systems Ltd., Godalming, UK) fitted with a spherical probe (P/0.25). The pre-test, test, and posttest speed were set at 8.0 mm/s, 5.0 mm/s, and 8.0 mm/s, respectively. The deformation ratio was 80%. A forceetime curve was recorded and analyzed by the software of Texture Exponent 32 (Surrey, UK) to calculate the peak force, which reflects the hardness of the material. The crispness was calculated as (Anon, 1998; Cruzycelis, Rooney, & McDonough, 1996):
C¼
k F
(1)
Where C, k, F refer to crispness of the sample (g1), specific material constant (k ¼ 10000), and hardness of the sample (g), respectively. Triplicates were performed for each batch of samples and the mean value was calculated.
2.1.2. Chemicals Chemicals used in the experiment were all analytical grade. Blue copperas, methylthionine chloride, sodium hydroxide (NaOH), potassium ferrocyanide, oxalic acid, ascorbic acid, sulfuric acid (H2SO4), potassium sulfate (K2SO4) and boric acid were provided by Sinopharm Chemical Reagent Beijing Co., Ltd (Beijing, China). 2.2. Preparation of apple chips Fresh apples were washed, peeled, cut into 5 mm thick slices on a Laboratory Slicer (model FA-200, Nanhai Defeng electrothermal equipment Co., Ltd., Guangdong, China). Apple chips were produced by using the experimental explosion puffing drying
Fig. 1. The diagram of explosion puffing drying equipment.
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
2.3.3. Expansion ratio The bulk of the apple chips were measured using an exclusive method, and the millet was selected as filling material (Li, Zhen, Luo, Fei, & Chen, 2003). The calculating equation was:
Vm ¼ V1 V0
(2)
Where Vm,, V1, V0, refer to volume of material (ml), millet and material (ml), and millet (ml), respectively. The expansion ratio was then calculated using the mean value of volume, and the calculating equation was:
r ¼ Va =Vb
(3)
Where r, Va, Vb refer to expansion ratio, volume of the sample after and before puffing (ml), respectively (Li et al., 2003). 2.3.4. Production rate Production rate (R) was calculated as:
R ¼ m2 =m1
(4)
where m1 and m2 refer to the weight of the pre-puffing and postpuffing samples (g), respectively. 2.3.5. Moisture content Moisture content was determined by a fast microwave moisture tester (LMA200PM-000EU, Sartorius Company, Goettingen, Gemany). The apple chips were crushed with a high-speed tissue stamp mill. About 0.75 g of powder was taken and put on a glass fiber paper for analysis. The measurement for each variety was replicated three times (AOAC., 1995). 2.3.6. Rehydration ratio (RR) Rehydration ratio could reflect the drying damage degree inside the material, which might partly explain the effectiveness of drying. Three grams of apple chips were placed in a beaker with distilled water at room temperature on a 1:50 (m/v) basis. The apple chips were taken out after soaking for 30 min, then drained by dripping and weighed as m3 (g). The measurement was replicated for three times. RR was calculated as (Zou et al., 2012):
RR ¼ m3 =3
(5)
2.3.7. Other evaluation indicators Crude fat content, protein content, crude fiber content, Vc content and reducing sugar content were determined according to standard AOAC Official Method 2003.05 (2003), AOAC Official
1131
Method 991.23 (1994), AOAC Official Method 930.20 (1930), AOAC Official Method 984.26 (1985) and AOAC Official Method 945.66 (1945), respectively; soluble solid was measured by a refractomth et al., 2007), eter (MASTER-a, Atago Co. Ltd., Tokyo, Japan) (Ro and titratable acid was measured with indicator titration since it would greatly impact product flavor (Nie, 2009). 2.4. Statistical analysis All data were processed and analyzed on SPSS12.0 (IBM, Chicago, USA). Single factor variance analysis was applied to analyze sixteen indicators of apple chips from nine different origins, the coefficient of variation for each indicator was calculated, and variances between indicators for apple chips with different origins were observed. The raw data for each indicator were changed into standardized data within 0e1 by using maximum difference normalization. The principle components of quality evaluation factors were determined, based on the cumulative variance contribution rate by using PCA, and characteristic indicators of apple chip quality were finally obtained by SCA. 3. Results and discussion In order to achieve comprehensive quality evaluation of apple chips, three main principles were considered when screening evaluation indicators. Firstly, the evaluation indicators were selected from sensory quality, physical and chemical quality, and processing quality due to their representative and extensive. Secondly, the overlapping and errors of evaluation indicators were removed to gain relative independent indicators through the statistic methods of PCA, CA and SCA. Finally, the data should obtain variation degree between varieties. The difference of evaluation indicators and comprehensive evaluation between varieties was identified by coefficient of variation (CV) and ANOVA. 3.1. Analysis of variance (ANOVA) Measured values of sixteen evaluation indicators of apple chips from nine different origins are shown in Table 1. ANOVA results showed that there were significant differences among the qualities of apple chips from nine different origins (P < 0.05). Among nine different origins, protein content (X1), crude fat (X2), crude fiber (X3), L value (X6), a value (X7) and b value (X8) had highly significant differences (P < 0.01), and reducing sugar content (X4), Vc (X5), titratable acid content (X9), expansion ratio (X10), moisture content (X12), rehydration ratio (X13), crispness (X15) and production rate (X16) also varied significantly. However, soluble solids content (X11)
Table 1 Measured values of sixteen evaluation indicators (X1 to X16) of apple chips from nine different origins (Y1 to Y9). Y*
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
1 2 3 4 5 6 7 8 9
2.79a 1.95b 1.88c 1.81d 1.62e 1.49f 1.47f 1.44fg 1.39g
0.16f 0.28d 0.24e 0.35bc 0.37b 0.44a 0.37b 0.32cd 0.34bc
2.79f 3.59b 3.12d 3.09d 2.94e 2.68f 3.38c 3.86a 3.51bc
59.68b 58.78c 57.66d 53.63e 53.58e 57.58d 53.28f 61.62a 59.92b
0.045bc 0.038c 0.045bc 0.061ab 0.045bc 0.061ab 0.053abc 0.038c 0.038c
41.81h 49.86a 40.13i 44.35e 43.14g 46.86c 43.5f 45.28d 48.87b
6.70d 6.82b 6.25g 6.75c 6.64f 6.90a 6.23h 5.83i 6.68e
17.14f 20.12a 14.97i 17.30e 16.81g 19.2b 16.81g 18.62c 18.48d
2.79bc 3.94ab 2.49bc 4.72a 2.69bc 2.09c 2.29bc 1.99bc 2.79bc
1.08ab 0.97ab 1.05ab 1.14a 1.10ab 0.94b 1.06ab 0.97ab 1.05ab
0.95a 0.75a 0.63a 0.62a 0.93a 0.68a 0.85a 0.90a 0.55a
0.89c 1.25bc 1.29b 1.59b 1.4bc 1.69ab 2.33a 1.34bc 1.56bc
1.69c 1.99bc 1.68c 2.02bc 2.86a 1.85bc 2.27b 1.95bc 2.3ab
381.30a 417.97a 373.18a 512.43a 500.66a 682.11a 512.61a 718.10a 561.07a
0.22b 0.23ac 0.33bc 0.35ac 0.80a 1.05a 0.56ac 0.36abc 0.56abc
16.17a 14.52abc 14.26abc 15.08ab 15.17ab 15.98a 12.62cd 11.14d 13.87bc
* Y1eY9: Aksu, Taigu, Luochuan, Hengshui, Lingbao, Jingning, Qixia, Fengxian, Yingkou; X1eX16: proteins content (g/100 g), crude fat content (g/100 g), crude fiber (g/100 g), reducing sugar content (g/100 g), Vc content (mg/100 g), “L” value, “a” value, “b” value, titratable acid content (g/100 g), expansion ratio, soluble solids content (g/g), moisture content (g/100 g), rehydration ratio, hardness (g), crispness (s), production rate (%); Different lowercase in the same column indicate significant difference (P < 0.05); The experiment was carried out in triplicate.
1132
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
Table 2 Characteristics of sixteen evaluation indicators of apple chips. Quality parameter
Mean
Luffinga
Range
Standard CV (%)b deviation
protein content 1.76 1.39e2.79 1.40 0.44 (X1, g/100 g) 0.32 0.16e0.44 0.28 0.08 crude fat content (X2, g/100 g) crude fiber content 3.22 2.68e3.86 1.18 0.39 (X3, g/100 g) 57.30 53.28e61.62 8.34 3.10 reducing sugar content (X4, g/100 g) 0.05 0.038e0.068 0.03 0.01 Vc content (X5, mg/100 g) “L” value (X6) 44.87 40.13e49.86 9.73 3.20 6.53 5.83e6.9 1.07 0.35 “a” value (X7) “b” value (X8) 17.65 14.97e20.12 5.15 1.59 titratable acid content 2.88 1.99e4.72 2.72 0.91 (X9, g/100 g) 1.04 0.94e1.14 0.20 0.07 expansion ratio (X10) soluble solids content 0.77 0.55e0.95 0.40 0.15 (X11, g/g) 1.49 0.90e2.30 1.40 0.39 moisture content (X12, g/100 g) 2.07 1.68e2.86 1.18 0.37 rehydration ratio (X13) hardness (X14, g) 517.71 373.18e718.1 344.92 121.89 crispness (X15, s) 0.49 0.22e1.05 0.83 0.28 production rate (X16, %) 14.31 11.14e16.17 5.03 1.61
24.84 26.19 12.27 5.41 21.31 7.14 5.39 9.06 31.49 6.41 19.15 26.62 17.81 23.54 56.26 11.25
a “Luffing” means the range of quality parameter between the minimum and the maximum values. b “CV” represents “coefficient of variation”.
Fig. 2. Contributions of the total variance accounted by the principal components.
and hardness (X14) did not show any difference. As the materials were from the same species, the main reason leading to the quality difference was probably due to the differences among raw materials. The differences may be caused by factors such as climate, geographical environment, and cultivation and management techniques. Harvest time, transportation methods and transportation time may have effect on the raw material quality, thus influencing on chip quality. Although the apple chips were manufactured under the same conditions, the layout inside explosion puffing pot may cause quality differences as well. Characteristics of sixteen evaluation indicators of apple chips are shown in Table 2. Among all the evaluation indicators, CVof crispness was the largest, which was 56.26%. Titratable acid content CV values were followed, ranging from 1.99 to 4.72 mg/100 g. While, a values of apple chips from nine different origins showed the smallest CV value (5.39%). The results also provided a valid basis for further screening of the most important evaluation indicators.
3.2. Correlation analysis (CA) To determine the relationships among different evaluation indicators of apple chip quality, correlation analysis (CA) was used. As dimensions of different quality evaluation indicators were inconsistent, the raw data were converted to standardized data ranging between 0 and 1 before CA. CA quantified the relationship between two variables and measured by correlation coefficient (r), which ranges from 1 to þ1. The two variables were highly correlated if r was close to 1 (negative) to þ1 (positive). When r is close to 0, there was almost no relation between the variables. The results of CA are presented in Table 3. According to Table 3, there were significant negative correlations between protein content and soluble solids content, protein content and hardness, b value and titratable acid content (P < 0.05). Positive correlations existed between crude fat and soluble solids content, and crude fat and hardness (P < 0.05).
Table 3 Correlation analysis on 16 evaluation indicators.a
X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
1 0.834** 0.418 0.169 0.343 0.384 0.286 0.175 0.279 0.346 0.692* 0.279 0.474 0.696* 0.561 0.553
1 0.022 0.435 0.475 0.395 0.130 0.266 0.210 0.237 0.713* 0.124 0.514 0.698* 0.815** 0.116
1 0.384 0.291 0.433 0.583 0.35 0.282 0.045 0.129 0.015 0.084 0.245 0.439 0.865**
1 0.248 0.315 0.206 0.46 0.592 0.046 0.569 0.296 0.514 0.200 0.322 0.222
1 0.225 0.431 0.033 0.253 0.654 0.484 0.107 0.132 0.248 0.466 0.264
1 0.378 0.908** 0.521 0.351 0.125 0.240 0.152 0.393 0.150 0.090
1 0.315 0.105 0.305 0.144 0.507 0.072 0.251 0.273 0.868**
1 0.671* 0.115 0.025 0.155 0.107 0.466 0.076 0.060
1 0.052 0.020 0.445 0.334 0.486 0.231 0.238
1 0.208 0.290 0.208 0.022 0.096 0.115
1 0.159 0.354 0.358 0.465 0.355
1 0.001 0.404 .0.440 0.332
1 0.144 0.439 0.103
1 0.552 0.419
1 0.225
1
Note: ** and * mean significant level at 0.01 and 0.05, respectively. a X1eX16: proteins content (g/100 g), crude fat content (g/100 g), crude fiber (g/100 g), reducing sugar content (g/100 g), Vc content (mg/100 g), “L” value, “a” value, “b” value, titratable acid content (g/100 g), expansion ratio, soluble solids content (g/g), moisture content (g/100 g), rehydration ratio, hardness (g), crispness (s), production rate (%).
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135 Table 4 Varimax rotated factor loadings of the first five principal components. Quality parameter
PC1
PC2
PC3
PC4
PC5
protein content (X1, g/100 g) crude fat content (X2, g/100 g) crude fiber content (X3, g/100 g) reducing sugar content (X4, g/100 g) Vc content (X5, mg/100 g) “L” value (X6) “a” value (X7) “b” value (X8) Titratable acid content (X9, g/100 g) Expansion ratio (X10) Soluble solids content (X11, g/g) Moisture content (X12, g/100 g) Rehydration ratio (X13) Hardness (X14, g) Crispness (X15, s) Production rate (X16, %) Characteristic root % of Variance Cumulative %
0.691 0.845 0.014 0.703 0.351 0.180 0.076 0.020 0.025 0.193 0.046 0.724 0.840 0.402 0.664 0.132 4.787 29.90 29.90
0.230 0.275 0.387 0.484 0.008 0.940 0.323 0.976 0.160 0.677 0.164 0.099 0.014 0.439 0.097 0.058 3.727 23.30 53.20
0.439 0.045 0.901 0.234 0.289 0.049 0.847 0.010 0.168 0.127 0.019 0.256 0.019 0.234 0.421 0.969 2.877 17.98 71.18
0.322 0.294 0.159 0.320 0.013 0.132 0.346 0.025 0.901 0.597 0.128 0.140 0.186 0.603 0.578 0.182 1.848 11.55 82.73
0.347 0.278 0.041 0.034 0.801 0.200 0.204 0.021 0.167 0.037 0.927 0.389 0.273 0.126 0.152 0.063 1.333 8.33 91.06
Highly significant negative correlations could be found between protein content and crude fat, and crude fiber and production rate (P < 0.01). Crude fat and crispness, L value and b value, and a value and production rate presented highly positive correlations (P < 0.01). The water loss was beneficial from the crude fiber because the high content of crude fiber in apple would form effective cell support structure, resulting in more efficient water evaporating. As a consequence, the final weights and production ratio might be small, which probably led to negative correlations between crude fiber and production rate. However, there was short
1133
of literature in other significant correlations, the influence each quality indicator has on other quality indicators needs to be studied further. In summary, there were correlations between some indicators within a certain range, indicating that several evaluation indicators were overlapped. To improve the evaluation efficiency and accuracy, it was necessary to further categorize and simplify related indicators. 3.3. Principal component analysis (PCA) PCA is not only a technique to reduce the number of variables by finding a linear combination of variables that explains the variance in the original variables, but also gives prominence to the relationship between the elements as well (Kara, 2009; Wang & Liu, 2010). The entire set of obtained data was analyzed by PCA. The result of decomposition was graphically depicted in the form of a scree plot presented in Fig. 2. Based on Kaiser's rule, only eigenvalues 1 could be viewed as main principle components (Shin, Pegg, Phillips, & Eitenmiller, 2010). In the present study, the eigenvalues of first five principle components were all larger than 1, which explaining 91.06% of the total variance. The first five principle components accounted for 29.90%, 23.30%, 17.98%, 11.55% and 8.33% of the total variation of the data, respectively (Table 4). The first principle component (PC) represents the maximum variation of the data set. To clarify the details of each PC represented, all the evaluation indicator values were further analyzed by varimax rotation. Varimax rotation was a modification of coordinates used in PCA that maximized the sum of the variances of the squared loadings (Brereton, 1990). It sought the most economical basis to represent each individualdso that each individual could be well described by a linear combination in terms of only a few basic
Fig. 3. Varimax rotated principal component loadings. (a) PC loading 1 versus PC loading 2; (b) PC loading 1 versus PC loading 3; (c) PC loading 1 versus PC loading 4; (d) PC loading 1 versus PC loading 5.
1134
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
functions. Varimax rotated factor loadings of the first five PCs were shown in Table 4 and Fig. 3. From Table 4, crude fat content, rehydration ratio and moisture content were the positive correlation with PC1; however, reducing sugar content was negative. The main influencing indicators of the second and third PCs were L and b values, and crude fiber content, a value and production rate, respectively. Titratable acid content, Vc content and soluble solids content were the main influencing indicators in the fourth and fifth PCs, respectively. As shown from Fig. 3, the distances of plots between crude fat content and crispness, “L” value and “b” value, titratable acid content and expansion ratio, Vc content and soluble solids content were very closed, indicating these indicators were overlapped. Thus, they were reasonable to further divide into five categories. And the characteristic indicators of each PC were obtained by SCA.
be the representative indicators (Table 2). In group 2, based on the ANOVA results (Table 2), the largest CV value was crispness (56.26%). It suggested that crispness had the greatest impact on the quality of apple chips and it can be viewed as a characteristic indicator in group 2. In group 3, titratable acid content was regarded as the characteristic indicator due to its highest CV values (31.49%). According to the CV values (Table 2), a value and protein content were less than the production rate (23.54%). At the same time, the measurement of production rate was much easier than the others, so the representative indicator in group 4 was production rate. As expansion ratio stood alone in group 5, which was also the characteristic indicator. In summary, five characteristic indicators of apple chips quality were finally obtained by SCA. These were crude fiber content, crispness, titratable acid content, production rate and expansion ratio.
3.4. System cluster analysis (SCA) 4. Conclusions SCA is the most widely used unsupervised pattern recognition technique in chemometrics. SCA involves hierarchically grouping samples on the basis of similarity without using prior information. In terms of similarity or nearness, different indicators will be grouped in clusters. Dendrogram of system cluster analysis for sixteen evaluation indicators are shown in Fig. 4. Five groups were obtained from SCA when the clustering distance was 14. These groups contained: Group 1: L value, b value, crude fiber content and reducing sugar content Group 2: crude fat content, crispness, hardness, soluble solids content, Vc content and rehydration ration Group 3: titratable acid content and moisture content Group 4: a value, production rate and protein content Group 5: expansion ratio Because there were correlations between evaluation indicators within each group, the representative indicator should be chosen. In group 1, CV values of L value, b value and reducing sugar content were smaller than 10%. Therefore, crude fiber content with CV value of 12.27% had greater impact on apple chips quality, and chosen to
In the present study, ANOVA, CA, PCA and SCA were successfully applied for selecting characteristic evaluation indicators of explosion puffing Fuji apple chips from nine different Chinese origins. Results of ANOVA showed that quality of apple chips from different origins showed significant differences (P < 0.05), and CV values of sixteen evaluation indicators for apple chips from nine origins were within 5.39% (a value) to 56.26% (crispness). Through CA, there were positive or negative correlations between some evaluation indicators, for example, crude fat content had significant positive correlation with crispness. Five characteristic indicators was finally obtained from five PCs through PCA and SCA, which were crude fiber content, crispness, titratable acid content, production rate and expansion ratio. Through screening the five characteristic indicators from 16 indicators, it simplified the quality evaluation process and improved the efficiency, which might be applied in food industry for quality control. Acknowledgments This work was supported by Special Fund for Agro-scientific Research in the Public Interest, Ministry of Agriculture (NO. 201303076, NO. 200903043). References
Fig. 4. Dendrogram of system cluster analysis for 16 evaluation indicators.
Anon. (1998). Comparison of textural qualities of crisp samples by bulk compression using an Ottawa cell. TA-XT2 Application study. Stable Micro Systems Ltd.. ref: CRISP1/OTC. Antonio, G. G., Alves, D. G., Azoubel, P. M. X., Murr, F. E., & Park, K. J. (2008). Influence of osmotic dehydration and high temperature short time processes on dried sweet potato (Ipomoea batatas Lam.). Journal of Food Engineering, 84(3), 375e382. AOAC. (1930). Crude fiber in Cacao products. USA: Official methods of Analysis of AOAC International. AOAC. (1945). Total reducing sugars. USA: Official Methods of Analysis of AOAC International. AOAC. (1985). Vitamin C (Total) in food Semiautomated fluorometric method. USA: Official Methods of Analysis of AOAC International. AOAC. (1994). Protein nitrogen content of milk Kjeldahl method. USA: Official Methods of Analysis of AOAC International. AOAC. (1995). Official methods of analysis of AOAC international (15th ed.) USA. AOAC. (2003). Official methods of analysis of AOAC international (5th ed.) USA. Bao, J. F., Xia, R. X., Deng, X. X., Peng, S. Y., Liu, Y. Z., Ma, X. T., et al. (2004). The quality evaluation factors selection of newhall, orange by the principal component analysis. Journal of Huazhong Agricultural University, 23(6), 663e666 (in Chinese). Bi, J. F. (2008). Optimization of explosion puffing drying process of apple at variable temperature and pressure difference. Journal of Food Science, 29(11), 213e218 (in Chinese). Brereton, R. G. (1990). Chemometrics-applications of mathematics and statistics to laboratory systems. West Sussex, UK: Ellis Horwood Limited.
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135 Cruzycelis, L. P., Rooney, L. W., & McDonough, C. M. (1996). A ready-to-eat breakfast cereal from food-grade sorghum. Journal of Cereal Chemistry, 73(1), 108e114. DeBelie, N., DeSmedt, V., & DeBaerdemaeker, J. (2000). Principal component analysis of chewing sounds to detect differences in apple crispness. Post Harvest Biology and Technology, 8, 109e119. Dong, G. Z., Yuan, B. Q., Gao, H. M., Li, X. L., Liu, Z. Z., & Sun, R. H. (2009). Actuality of production and processing of apple Industry in China. Journal of Innovational Edition of Farm Products Processing, 10, 64 (in Chinese). Dorta, N., & Piotr, P. (2004). Infrared drying of apple slices. Innovative Food Science and Emerging Technologies, 5, 353e360. Duan, J. L., & Wang, X. D. (2007). Study on anti-browning of vacuum freezedrying apple slice. Journal of Food Research and Development, 28(7), 102 (in Chinese). Han, Q. H., Li, S. J., Ma, J. W., & Zhao, D. L. (2006). Microwave vacuum drying and puffing characteristics of apple chips. Journal of Chinese Academy of Agricultural Mechanization Sciences, 37(8), 156e158 (in Chinese). Hofsetz, K., Lopes, C. C., Hubinger, M. D., Mayor, L., & Sereno, A. M. (2007). Changes in the physical properties of bananas on applying HTST pulse during air-drying. Journal of Food Engineering, 83(4), 531e540. Huang, G. S., Yang, G. L., & Fang, C. J. (2003). Evaluating 40 pear fruit varieties by the multidimensional value theory. Journal of Human Forestry Science & Technology, 30(4), 26e27 (in Chinese). Kara, D. (2009). Evaluation of trace metal concentrations in some herbs and herbal teas by principal component analysis. Food Chemistry, 114(1), 347e354. Lemus-Mondaca, R., Miranda, M., Grau, A. A., Briones, V., Villalobos, R., & Vegalvez, A. (2009). Effect of osmotic pretreatment on hot air drying kinetics and Ga quality of Chilean papaya (Carica pubescens). Drying Technology, 27(10), 1105e1115. Lewicki, P. P., Gondek, E., Witrowa-Rajchert, D., & Nowak, D. (2001). Effect of drying on respiration of apple slices. Journal of Food Engineering, 49(4), 333e337. Li, Y. Z., Zhen, S. X., Luo, S. C., Fei, K., & Chen, D. X. (2003). Study on the processing characteristics of potato crispy chips by vacuum microwave. Food and Fermentation Industries, 29(8), 40e43 (in Chinese).
1135
Liu, J., Wei, J. L., Liu, M. L., Song, Y., Feng, S. Q., Wang, C. Z., et al. (2012). The relationships between the enzyme activity of anthocyanin biosynthesis, ethylene release and anthocyanin accumulation in fruits of precocious apple cultivars. Journal of Acta Horticulturae Sinica, 39(7), 1235e1242 (in Chinese). Ma, L. X., Bi, J. F., & Wei, Y. M. (2005). Influence factor of explosion puffing drying process of apple. Science and Technology of Food Industry, 12, 88e93 (in Chinese). Nie, J. Y. (2009). Analytical techniques of fruit quality and safety. Beijing: Chemical Industry Press ( in Chinese). Ozilgen, M., Guvenc, G., Makaraci, M., & Tumer, I. (1995). Color-change and weightloss of apple slices during drying. Zeitschrift für Lebensmittel-Untersuchung und Forschung, 201(1), 40e45. th, E., Berna, A., Beullens, K., Yarramraju, S., Lammertyn, J., Schenk, A., et al. Ro (2007). Postharvest quality of integrated and organically produced apple fruit. Postharvest Biology and Technology, 45(1), 11e19. Shin, E. C., Pegg, R. B., Phillips, R. D., & Eitenmiller, R. R. (2010). Interrelationships among tocopherols of commercial runner market type peanuts grown in the United States. International Journal of Food Science and Technology, 45(12), 2622e2628. Sullivan, J. F., & Craig, J. C. (1980). Continous explosion-puffing of apples. Journal of Food Science, 45, 1550e1555. Sullivan, J. F., & Craig, J. C., Jr. (1984). The development of explosion puffing. Food Technology, 38(2), 52e55. Wang, H. W., & Liu, Y. Q. (2010). Evaluation of trace and toxic element concentrations in Paris polyphylla from China with empirical and chemometric approaches. Food Chemistry, 121(3), 887e892. Woodroof, J. G., & Luh, B. S. (1975). Dehydration of fruits. Commercial fruit processing. Westport, Connecticut: The AVI Publishing Company, Inc. Xu, J. H., Zhao, Z. Y., Wang, L. C., Gao, H., Liu, Z. Z., & Fan, H. K. (2011). Selection of factors for apple fruit quality evaluation. Agriculcure Research in the Arid Area, 29(6), 269e274 (in Chinese). Zou, K., Teng, J., Huang, L., Dai, X., & Wei, B. (2012). Effect of osmotic pretreatment on quality of mango chips by explosion puffing drying. LWT-Food Science and Technology, 51(1), 253e259.
Contents lists available at ScienceDirect
LWT - Food Science and Technology journal homepage: www.elsevier.com/locate/lwt
Evaluation indicators of explosion puffing Fuji apple chips quality from different Chinese origins Jin-feng Bi*, 1, Xuan Wang 1, Qin-qin Chen, Xuan Liu, Xin-ye Wu, Qiang Wang, Jian Lv, Ai-jin Yang Institute of Agro-Products Processing Science and Technology, CAAS, Key Laboratory of Agro-Products Processing, Ministry of Agriculture, Beijing 100193, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 May 2013 Received in revised form 11 September 2014 Accepted 1 October 2014 Available online 13 October 2014
In this study, sixteen evaluation indicators of explosion puffing Fuji apple chips from nine different origins in China were analyzed, and the most important evaluation indicators were obtained by analysis of variance (ANOVA), correlation analysis (CA), principal component analysis (PCA) and system cluster analysis (SCA). Results of ANOVA showed that quality of apple chips from different origins showed significant differences (P < 0.05). The coefficient of variation (CV) values of sixteen evaluation indicators for apple chips from nine origins were 5.39 %e56.26%. Through CA results, some indicators were correlated to each other within a certain range, such as crude fat content and crispness. Five principle components were extracted through PCA with eigenvalue of over 1, and the cumulative contribution was achieved at 91.06%. Based on the above results, five characteristic indicators of apple chip quality were obtained by SCA. These were crude fiber content, crispness, titratable acid content, production rate and expansion ratio. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Apple chips Evaluation indicator Principal component analysis Cluster analysis
1. Introduction Apple is one of the most frequently consumed fruit. In China, production of commercial apples amounted to 35 million t/year in recent years, of which most (70%) were used for direct consumption, and others (30%) were processed for jam, juice, and dehydrated products (Liu et al., 2012; Woodroof & Luh, 1975). Recently, fruit and vegetable chips have become very important in the diet of the modern consumer because the chips not only can extend shelf life but maintain the original nutrition as well (Zou, Teng, Huang, Dai, & Wei, 2012). Apple chips occupy big market shares in an increasing puffed food market because of being crispy and refreshing, sweet aroma, high nutrition and ease to carry (Dong et al., 2009). There are several processing technologies for apple chips, such as deep fat frying, vacuum frying, freeze-drying and microwave vacuum drying. Among the apple chips processing technologies, explosion puffing drying (EPD) has its unique advantages, no oxygen reduced oxidation in the processing under
Abbreviations: ANOVA, analysis of variance; CA, correlation analysis; PCA, principal component analysis; SCA, system cluster analysis. * Corresponding author. Tel./fax: þ86 10 62812584. E-mail addresses: [email protected], [email protected] (J.-f. Bi). 1 They contributed equally to the present paper. http://dx.doi.org/10.1016/j.lwt.2014.10.007 0023-6438/© 2014 Elsevier Ltd. All rights reserved.
vacuum drying condition, the processed product is natural and nutritional. EPD means that the material is puffed at higher temperatures between 80 and 130 C with higher vapor pressure of 0.1e0.3 MPa and dried in vacuum conditions with lower temperature between 60 and 90 C. A sudden increase in temperature or a sudden decrease in pressure causes water evaporation and expansion in the cells of the materials (Sullivan & Craig, 1984). A puffing process, which performs as an intermediate stage, involves the release or expansion of vapor or gas within the product, either to create an internal structure or to expand and/or rupture an existing one (Antonio, Alves, Azoubel, Murr, & Park, 2008; Hofsetz, Lopes, Hubinger, Mayor, & Sereno, 2007). Compared with the traditional deep-frying and freeze-drying technologies, the product processed with EPD has lower oil content and longer shelf life. Moreover the EPD machine is less expensive, which may promise a bright future for fruit and vegetable drying technology (Ma, Bi, & Wei, 2005). Parameters optimization of apple chips have been extensive studied (Bi, 2008; Duan & Wang, 2007; DeBelie, DeSmedt, & DeBaerdemaeker, 2000; Dorta & Piotr, 2004; Han, Li, Ma, &Zhao, 2006; Lewicki, Gondek, Witrowa-Rajchert, & Nowak, 2001; Ozilgen, Guvenc, Makaraci, & Tumer, 1995; Sullivan & Craig, 1980). However, the research on quality evaluation of apple chips was mainly about sensory evaluation, instead of quality evaluation
1130
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
models. The quality of chips includes sensory quality, physicochemical nutritional quality and processing quality, which can be evaluated by complex factors, such as product appearance moisture content and Vitamin C (Vc). Different quality factors which were closely related but were relatively independent increase the difficulty of comprehensive evaluation. Therefore, it becomes one of the major tasks to search for a simple method to evaluate the quality of apple chips. Most studies apply principal component analysis (PCA) or cluster analysis to select food evaluation indicators. For example, Bao et al. (2004) studied the selection of navel orange quality evaluation factors using PCA. The result showed that six factors including soluble solids content (or peel redness), titratable acid content, Vc content, juice yield, peel brightness (or peel yellowness), and mean individual fruit weight (or fruit shape index or peel thickness) could reflect the most information on fruit quality of Newhall orange. Xu et al. (2011) adopted cluster analysis to simplify apple quality factors. The result showed that the five most important fruit quality parameters were fruit weight or fruit shape index, fruit firmness, soluble solids content, titratable acid and fruit color. Huang, Yang, and Fang (2003) evaluated the quality of pear varieties using multidimensional value theory, which showed that Vc, soluble solids content and titratable acid were the most important quality indicators. The above studies filter quality factors only using PCA or cluster analysis separately. There is little research combining these two methods. In this study, nine different Fuji apples producing regions in China were selected to manufacture apple chips by explosion puffing drying technology. Meanwhile, sixteen evaluation indicators were measured for apple chip quality. PCA and system cluster analysis (SCA) were adopted to simplify apple chip quality evaluation indicators, which provided a scientific basis for the research on apple chip quality evaluation. 2. Material and methods 2.1. Materials 2.1.1. Fuji apples “No. 2 Changfu” is one of the most widely cultivated Fuji varieties. They were picked from nine different origins, including Aksu (Xinjiang province, 80 290 E, 41150 N), Qixia (Shandong province, 120 830 E, 37 80 N), Lingbao (Henan province, 110 850 E, 34 520 N), Luochuan (Shannxi province, 109 420 E, 35760 N), Hengshui (Hebei province, 115720 E, 37 710 N), Taigu (Shanxi province, 122 530 E, 37420 N), Yingkou (Liaoning province, 122130 E, 40 390 N), Fengxian (Jiangsu province, 116 570 E, 34790 N) and Jingning (Gansu province, 105730 E, 35 510 N). These apples were picked from trees from middle to late October in 2011. One hundred apples were picked in each location and were made into apple chips using explosion puffing drying technology.
equipment system developed by Tianjin Qin-de New Material Scientific Development Co. Ltd. (Tianjin, China). This system can be adjusted to any desired puffing temperature and vacuum pressure. It consists of puffing chamber, vacuum chamber, vacuum pump, decompression valve, air compressor, vapor generator, and manual control panel (Fig. 1). Pre-weighed apple slices (2 kg) were predried at 80 C for 120 min. Then the materials were puffed at 105 C for 10 min with a vapor pressure of 0.3 MPa and dried in vacuum conditions with a lower temperature of 80 C for 120 min (Bi, 2008). The experiment was carried out in triplicate. 2.3. Measurement of evaluation indicators 2.3.1. Color The surface color of the apple chips was measured using colorimeter (D25L, Hunterlab, Virginia, USA). Based on the CIELab color space and after calibration with the white tile and black glass, three equidistant spots were examined on the major axis of each apple chip sample. Since the spot diameter of the instrument was 10 mm, the total area of the slab was 10 cm2. Triplicate experiments were performed. Three colorimetric evaluation parameters, including lightness (L), greenered hue (a) and blueeyellow hue (b) were measured and calculated to observe color changes. (LemusMondaca et al., 2009). 2.3.2. Crispness and hardness Crispness of the apple chips was measured using a TA-XT2 texture analyzer (Stable Micro Systems Ltd., Godalming, UK) fitted with a spherical probe (P/0.25). The pre-test, test, and posttest speed were set at 8.0 mm/s, 5.0 mm/s, and 8.0 mm/s, respectively. The deformation ratio was 80%. A forceetime curve was recorded and analyzed by the software of Texture Exponent 32 (Surrey, UK) to calculate the peak force, which reflects the hardness of the material. The crispness was calculated as (Anon, 1998; Cruzycelis, Rooney, & McDonough, 1996):
C¼
k F
(1)
Where C, k, F refer to crispness of the sample (g1), specific material constant (k ¼ 10000), and hardness of the sample (g), respectively. Triplicates were performed for each batch of samples and the mean value was calculated.
2.1.2. Chemicals Chemicals used in the experiment were all analytical grade. Blue copperas, methylthionine chloride, sodium hydroxide (NaOH), potassium ferrocyanide, oxalic acid, ascorbic acid, sulfuric acid (H2SO4), potassium sulfate (K2SO4) and boric acid were provided by Sinopharm Chemical Reagent Beijing Co., Ltd (Beijing, China). 2.2. Preparation of apple chips Fresh apples were washed, peeled, cut into 5 mm thick slices on a Laboratory Slicer (model FA-200, Nanhai Defeng electrothermal equipment Co., Ltd., Guangdong, China). Apple chips were produced by using the experimental explosion puffing drying
Fig. 1. The diagram of explosion puffing drying equipment.
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
2.3.3. Expansion ratio The bulk of the apple chips were measured using an exclusive method, and the millet was selected as filling material (Li, Zhen, Luo, Fei, & Chen, 2003). The calculating equation was:
Vm ¼ V1 V0
(2)
Where Vm,, V1, V0, refer to volume of material (ml), millet and material (ml), and millet (ml), respectively. The expansion ratio was then calculated using the mean value of volume, and the calculating equation was:
r ¼ Va =Vb
(3)
Where r, Va, Vb refer to expansion ratio, volume of the sample after and before puffing (ml), respectively (Li et al., 2003). 2.3.4. Production rate Production rate (R) was calculated as:
R ¼ m2 =m1
(4)
where m1 and m2 refer to the weight of the pre-puffing and postpuffing samples (g), respectively. 2.3.5. Moisture content Moisture content was determined by a fast microwave moisture tester (LMA200PM-000EU, Sartorius Company, Goettingen, Gemany). The apple chips were crushed with a high-speed tissue stamp mill. About 0.75 g of powder was taken and put on a glass fiber paper for analysis. The measurement for each variety was replicated three times (AOAC., 1995). 2.3.6. Rehydration ratio (RR) Rehydration ratio could reflect the drying damage degree inside the material, which might partly explain the effectiveness of drying. Three grams of apple chips were placed in a beaker with distilled water at room temperature on a 1:50 (m/v) basis. The apple chips were taken out after soaking for 30 min, then drained by dripping and weighed as m3 (g). The measurement was replicated for three times. RR was calculated as (Zou et al., 2012):
RR ¼ m3 =3
(5)
2.3.7. Other evaluation indicators Crude fat content, protein content, crude fiber content, Vc content and reducing sugar content were determined according to standard AOAC Official Method 2003.05 (2003), AOAC Official
1131
Method 991.23 (1994), AOAC Official Method 930.20 (1930), AOAC Official Method 984.26 (1985) and AOAC Official Method 945.66 (1945), respectively; soluble solid was measured by a refractomth et al., 2007), eter (MASTER-a, Atago Co. Ltd., Tokyo, Japan) (Ro and titratable acid was measured with indicator titration since it would greatly impact product flavor (Nie, 2009). 2.4. Statistical analysis All data were processed and analyzed on SPSS12.0 (IBM, Chicago, USA). Single factor variance analysis was applied to analyze sixteen indicators of apple chips from nine different origins, the coefficient of variation for each indicator was calculated, and variances between indicators for apple chips with different origins were observed. The raw data for each indicator were changed into standardized data within 0e1 by using maximum difference normalization. The principle components of quality evaluation factors were determined, based on the cumulative variance contribution rate by using PCA, and characteristic indicators of apple chip quality were finally obtained by SCA. 3. Results and discussion In order to achieve comprehensive quality evaluation of apple chips, three main principles were considered when screening evaluation indicators. Firstly, the evaluation indicators were selected from sensory quality, physical and chemical quality, and processing quality due to their representative and extensive. Secondly, the overlapping and errors of evaluation indicators were removed to gain relative independent indicators through the statistic methods of PCA, CA and SCA. Finally, the data should obtain variation degree between varieties. The difference of evaluation indicators and comprehensive evaluation between varieties was identified by coefficient of variation (CV) and ANOVA. 3.1. Analysis of variance (ANOVA) Measured values of sixteen evaluation indicators of apple chips from nine different origins are shown in Table 1. ANOVA results showed that there were significant differences among the qualities of apple chips from nine different origins (P < 0.05). Among nine different origins, protein content (X1), crude fat (X2), crude fiber (X3), L value (X6), a value (X7) and b value (X8) had highly significant differences (P < 0.01), and reducing sugar content (X4), Vc (X5), titratable acid content (X9), expansion ratio (X10), moisture content (X12), rehydration ratio (X13), crispness (X15) and production rate (X16) also varied significantly. However, soluble solids content (X11)
Table 1 Measured values of sixteen evaluation indicators (X1 to X16) of apple chips from nine different origins (Y1 to Y9). Y*
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
1 2 3 4 5 6 7 8 9
2.79a 1.95b 1.88c 1.81d 1.62e 1.49f 1.47f 1.44fg 1.39g
0.16f 0.28d 0.24e 0.35bc 0.37b 0.44a 0.37b 0.32cd 0.34bc
2.79f 3.59b 3.12d 3.09d 2.94e 2.68f 3.38c 3.86a 3.51bc
59.68b 58.78c 57.66d 53.63e 53.58e 57.58d 53.28f 61.62a 59.92b
0.045bc 0.038c 0.045bc 0.061ab 0.045bc 0.061ab 0.053abc 0.038c 0.038c
41.81h 49.86a 40.13i 44.35e 43.14g 46.86c 43.5f 45.28d 48.87b
6.70d 6.82b 6.25g 6.75c 6.64f 6.90a 6.23h 5.83i 6.68e
17.14f 20.12a 14.97i 17.30e 16.81g 19.2b 16.81g 18.62c 18.48d
2.79bc 3.94ab 2.49bc 4.72a 2.69bc 2.09c 2.29bc 1.99bc 2.79bc
1.08ab 0.97ab 1.05ab 1.14a 1.10ab 0.94b 1.06ab 0.97ab 1.05ab
0.95a 0.75a 0.63a 0.62a 0.93a 0.68a 0.85a 0.90a 0.55a
0.89c 1.25bc 1.29b 1.59b 1.4bc 1.69ab 2.33a 1.34bc 1.56bc
1.69c 1.99bc 1.68c 2.02bc 2.86a 1.85bc 2.27b 1.95bc 2.3ab
381.30a 417.97a 373.18a 512.43a 500.66a 682.11a 512.61a 718.10a 561.07a
0.22b 0.23ac 0.33bc 0.35ac 0.80a 1.05a 0.56ac 0.36abc 0.56abc
16.17a 14.52abc 14.26abc 15.08ab 15.17ab 15.98a 12.62cd 11.14d 13.87bc
* Y1eY9: Aksu, Taigu, Luochuan, Hengshui, Lingbao, Jingning, Qixia, Fengxian, Yingkou; X1eX16: proteins content (g/100 g), crude fat content (g/100 g), crude fiber (g/100 g), reducing sugar content (g/100 g), Vc content (mg/100 g), “L” value, “a” value, “b” value, titratable acid content (g/100 g), expansion ratio, soluble solids content (g/g), moisture content (g/100 g), rehydration ratio, hardness (g), crispness (s), production rate (%); Different lowercase in the same column indicate significant difference (P < 0.05); The experiment was carried out in triplicate.
1132
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
Table 2 Characteristics of sixteen evaluation indicators of apple chips. Quality parameter
Mean
Luffinga
Range
Standard CV (%)b deviation
protein content 1.76 1.39e2.79 1.40 0.44 (X1, g/100 g) 0.32 0.16e0.44 0.28 0.08 crude fat content (X2, g/100 g) crude fiber content 3.22 2.68e3.86 1.18 0.39 (X3, g/100 g) 57.30 53.28e61.62 8.34 3.10 reducing sugar content (X4, g/100 g) 0.05 0.038e0.068 0.03 0.01 Vc content (X5, mg/100 g) “L” value (X6) 44.87 40.13e49.86 9.73 3.20 6.53 5.83e6.9 1.07 0.35 “a” value (X7) “b” value (X8) 17.65 14.97e20.12 5.15 1.59 titratable acid content 2.88 1.99e4.72 2.72 0.91 (X9, g/100 g) 1.04 0.94e1.14 0.20 0.07 expansion ratio (X10) soluble solids content 0.77 0.55e0.95 0.40 0.15 (X11, g/g) 1.49 0.90e2.30 1.40 0.39 moisture content (X12, g/100 g) 2.07 1.68e2.86 1.18 0.37 rehydration ratio (X13) hardness (X14, g) 517.71 373.18e718.1 344.92 121.89 crispness (X15, s) 0.49 0.22e1.05 0.83 0.28 production rate (X16, %) 14.31 11.14e16.17 5.03 1.61
24.84 26.19 12.27 5.41 21.31 7.14 5.39 9.06 31.49 6.41 19.15 26.62 17.81 23.54 56.26 11.25
a “Luffing” means the range of quality parameter between the minimum and the maximum values. b “CV” represents “coefficient of variation”.
Fig. 2. Contributions of the total variance accounted by the principal components.
and hardness (X14) did not show any difference. As the materials were from the same species, the main reason leading to the quality difference was probably due to the differences among raw materials. The differences may be caused by factors such as climate, geographical environment, and cultivation and management techniques. Harvest time, transportation methods and transportation time may have effect on the raw material quality, thus influencing on chip quality. Although the apple chips were manufactured under the same conditions, the layout inside explosion puffing pot may cause quality differences as well. Characteristics of sixteen evaluation indicators of apple chips are shown in Table 2. Among all the evaluation indicators, CVof crispness was the largest, which was 56.26%. Titratable acid content CV values were followed, ranging from 1.99 to 4.72 mg/100 g. While, a values of apple chips from nine different origins showed the smallest CV value (5.39%). The results also provided a valid basis for further screening of the most important evaluation indicators.
3.2. Correlation analysis (CA) To determine the relationships among different evaluation indicators of apple chip quality, correlation analysis (CA) was used. As dimensions of different quality evaluation indicators were inconsistent, the raw data were converted to standardized data ranging between 0 and 1 before CA. CA quantified the relationship between two variables and measured by correlation coefficient (r), which ranges from 1 to þ1. The two variables were highly correlated if r was close to 1 (negative) to þ1 (positive). When r is close to 0, there was almost no relation between the variables. The results of CA are presented in Table 3. According to Table 3, there were significant negative correlations between protein content and soluble solids content, protein content and hardness, b value and titratable acid content (P < 0.05). Positive correlations existed between crude fat and soluble solids content, and crude fat and hardness (P < 0.05).
Table 3 Correlation analysis on 16 evaluation indicators.a
X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
1 0.834** 0.418 0.169 0.343 0.384 0.286 0.175 0.279 0.346 0.692* 0.279 0.474 0.696* 0.561 0.553
1 0.022 0.435 0.475 0.395 0.130 0.266 0.210 0.237 0.713* 0.124 0.514 0.698* 0.815** 0.116
1 0.384 0.291 0.433 0.583 0.35 0.282 0.045 0.129 0.015 0.084 0.245 0.439 0.865**
1 0.248 0.315 0.206 0.46 0.592 0.046 0.569 0.296 0.514 0.200 0.322 0.222
1 0.225 0.431 0.033 0.253 0.654 0.484 0.107 0.132 0.248 0.466 0.264
1 0.378 0.908** 0.521 0.351 0.125 0.240 0.152 0.393 0.150 0.090
1 0.315 0.105 0.305 0.144 0.507 0.072 0.251 0.273 0.868**
1 0.671* 0.115 0.025 0.155 0.107 0.466 0.076 0.060
1 0.052 0.020 0.445 0.334 0.486 0.231 0.238
1 0.208 0.290 0.208 0.022 0.096 0.115
1 0.159 0.354 0.358 0.465 0.355
1 0.001 0.404 .0.440 0.332
1 0.144 0.439 0.103
1 0.552 0.419
1 0.225
1
Note: ** and * mean significant level at 0.01 and 0.05, respectively. a X1eX16: proteins content (g/100 g), crude fat content (g/100 g), crude fiber (g/100 g), reducing sugar content (g/100 g), Vc content (mg/100 g), “L” value, “a” value, “b” value, titratable acid content (g/100 g), expansion ratio, soluble solids content (g/g), moisture content (g/100 g), rehydration ratio, hardness (g), crispness (s), production rate (%).
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135 Table 4 Varimax rotated factor loadings of the first five principal components. Quality parameter
PC1
PC2
PC3
PC4
PC5
protein content (X1, g/100 g) crude fat content (X2, g/100 g) crude fiber content (X3, g/100 g) reducing sugar content (X4, g/100 g) Vc content (X5, mg/100 g) “L” value (X6) “a” value (X7) “b” value (X8) Titratable acid content (X9, g/100 g) Expansion ratio (X10) Soluble solids content (X11, g/g) Moisture content (X12, g/100 g) Rehydration ratio (X13) Hardness (X14, g) Crispness (X15, s) Production rate (X16, %) Characteristic root % of Variance Cumulative %
0.691 0.845 0.014 0.703 0.351 0.180 0.076 0.020 0.025 0.193 0.046 0.724 0.840 0.402 0.664 0.132 4.787 29.90 29.90
0.230 0.275 0.387 0.484 0.008 0.940 0.323 0.976 0.160 0.677 0.164 0.099 0.014 0.439 0.097 0.058 3.727 23.30 53.20
0.439 0.045 0.901 0.234 0.289 0.049 0.847 0.010 0.168 0.127 0.019 0.256 0.019 0.234 0.421 0.969 2.877 17.98 71.18
0.322 0.294 0.159 0.320 0.013 0.132 0.346 0.025 0.901 0.597 0.128 0.140 0.186 0.603 0.578 0.182 1.848 11.55 82.73
0.347 0.278 0.041 0.034 0.801 0.200 0.204 0.021 0.167 0.037 0.927 0.389 0.273 0.126 0.152 0.063 1.333 8.33 91.06
Highly significant negative correlations could be found between protein content and crude fat, and crude fiber and production rate (P < 0.01). Crude fat and crispness, L value and b value, and a value and production rate presented highly positive correlations (P < 0.01). The water loss was beneficial from the crude fiber because the high content of crude fiber in apple would form effective cell support structure, resulting in more efficient water evaporating. As a consequence, the final weights and production ratio might be small, which probably led to negative correlations between crude fiber and production rate. However, there was short
1133
of literature in other significant correlations, the influence each quality indicator has on other quality indicators needs to be studied further. In summary, there were correlations between some indicators within a certain range, indicating that several evaluation indicators were overlapped. To improve the evaluation efficiency and accuracy, it was necessary to further categorize and simplify related indicators. 3.3. Principal component analysis (PCA) PCA is not only a technique to reduce the number of variables by finding a linear combination of variables that explains the variance in the original variables, but also gives prominence to the relationship between the elements as well (Kara, 2009; Wang & Liu, 2010). The entire set of obtained data was analyzed by PCA. The result of decomposition was graphically depicted in the form of a scree plot presented in Fig. 2. Based on Kaiser's rule, only eigenvalues 1 could be viewed as main principle components (Shin, Pegg, Phillips, & Eitenmiller, 2010). In the present study, the eigenvalues of first five principle components were all larger than 1, which explaining 91.06% of the total variance. The first five principle components accounted for 29.90%, 23.30%, 17.98%, 11.55% and 8.33% of the total variation of the data, respectively (Table 4). The first principle component (PC) represents the maximum variation of the data set. To clarify the details of each PC represented, all the evaluation indicator values were further analyzed by varimax rotation. Varimax rotation was a modification of coordinates used in PCA that maximized the sum of the variances of the squared loadings (Brereton, 1990). It sought the most economical basis to represent each individualdso that each individual could be well described by a linear combination in terms of only a few basic
Fig. 3. Varimax rotated principal component loadings. (a) PC loading 1 versus PC loading 2; (b) PC loading 1 versus PC loading 3; (c) PC loading 1 versus PC loading 4; (d) PC loading 1 versus PC loading 5.
1134
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135
functions. Varimax rotated factor loadings of the first five PCs were shown in Table 4 and Fig. 3. From Table 4, crude fat content, rehydration ratio and moisture content were the positive correlation with PC1; however, reducing sugar content was negative. The main influencing indicators of the second and third PCs were L and b values, and crude fiber content, a value and production rate, respectively. Titratable acid content, Vc content and soluble solids content were the main influencing indicators in the fourth and fifth PCs, respectively. As shown from Fig. 3, the distances of plots between crude fat content and crispness, “L” value and “b” value, titratable acid content and expansion ratio, Vc content and soluble solids content were very closed, indicating these indicators were overlapped. Thus, they were reasonable to further divide into five categories. And the characteristic indicators of each PC were obtained by SCA.
be the representative indicators (Table 2). In group 2, based on the ANOVA results (Table 2), the largest CV value was crispness (56.26%). It suggested that crispness had the greatest impact on the quality of apple chips and it can be viewed as a characteristic indicator in group 2. In group 3, titratable acid content was regarded as the characteristic indicator due to its highest CV values (31.49%). According to the CV values (Table 2), a value and protein content were less than the production rate (23.54%). At the same time, the measurement of production rate was much easier than the others, so the representative indicator in group 4 was production rate. As expansion ratio stood alone in group 5, which was also the characteristic indicator. In summary, five characteristic indicators of apple chips quality were finally obtained by SCA. These were crude fiber content, crispness, titratable acid content, production rate and expansion ratio.
3.4. System cluster analysis (SCA) 4. Conclusions SCA is the most widely used unsupervised pattern recognition technique in chemometrics. SCA involves hierarchically grouping samples on the basis of similarity without using prior information. In terms of similarity or nearness, different indicators will be grouped in clusters. Dendrogram of system cluster analysis for sixteen evaluation indicators are shown in Fig. 4. Five groups were obtained from SCA when the clustering distance was 14. These groups contained: Group 1: L value, b value, crude fiber content and reducing sugar content Group 2: crude fat content, crispness, hardness, soluble solids content, Vc content and rehydration ration Group 3: titratable acid content and moisture content Group 4: a value, production rate and protein content Group 5: expansion ratio Because there were correlations between evaluation indicators within each group, the representative indicator should be chosen. In group 1, CV values of L value, b value and reducing sugar content were smaller than 10%. Therefore, crude fiber content with CV value of 12.27% had greater impact on apple chips quality, and chosen to
In the present study, ANOVA, CA, PCA and SCA were successfully applied for selecting characteristic evaluation indicators of explosion puffing Fuji apple chips from nine different Chinese origins. Results of ANOVA showed that quality of apple chips from different origins showed significant differences (P < 0.05), and CV values of sixteen evaluation indicators for apple chips from nine origins were within 5.39% (a value) to 56.26% (crispness). Through CA, there were positive or negative correlations between some evaluation indicators, for example, crude fat content had significant positive correlation with crispness. Five characteristic indicators was finally obtained from five PCs through PCA and SCA, which were crude fiber content, crispness, titratable acid content, production rate and expansion ratio. Through screening the five characteristic indicators from 16 indicators, it simplified the quality evaluation process and improved the efficiency, which might be applied in food industry for quality control. Acknowledgments This work was supported by Special Fund for Agro-scientific Research in the Public Interest, Ministry of Agriculture (NO. 201303076, NO. 200903043). References
Fig. 4. Dendrogram of system cluster analysis for 16 evaluation indicators.
Anon. (1998). Comparison of textural qualities of crisp samples by bulk compression using an Ottawa cell. TA-XT2 Application study. Stable Micro Systems Ltd.. ref: CRISP1/OTC. Antonio, G. G., Alves, D. G., Azoubel, P. M. X., Murr, F. E., & Park, K. J. (2008). Influence of osmotic dehydration and high temperature short time processes on dried sweet potato (Ipomoea batatas Lam.). Journal of Food Engineering, 84(3), 375e382. AOAC. (1930). Crude fiber in Cacao products. USA: Official methods of Analysis of AOAC International. AOAC. (1945). Total reducing sugars. USA: Official Methods of Analysis of AOAC International. AOAC. (1985). Vitamin C (Total) in food Semiautomated fluorometric method. USA: Official Methods of Analysis of AOAC International. AOAC. (1994). Protein nitrogen content of milk Kjeldahl method. USA: Official Methods of Analysis of AOAC International. AOAC. (1995). Official methods of analysis of AOAC international (15th ed.) USA. AOAC. (2003). Official methods of analysis of AOAC international (5th ed.) USA. Bao, J. F., Xia, R. X., Deng, X. X., Peng, S. Y., Liu, Y. Z., Ma, X. T., et al. (2004). The quality evaluation factors selection of newhall, orange by the principal component analysis. Journal of Huazhong Agricultural University, 23(6), 663e666 (in Chinese). Bi, J. F. (2008). Optimization of explosion puffing drying process of apple at variable temperature and pressure difference. Journal of Food Science, 29(11), 213e218 (in Chinese). Brereton, R. G. (1990). Chemometrics-applications of mathematics and statistics to laboratory systems. West Sussex, UK: Ellis Horwood Limited.
J.-f. Bi et al. / LWT - Food Science and Technology 60 (2015) 1129e1135 Cruzycelis, L. P., Rooney, L. W., & McDonough, C. M. (1996). A ready-to-eat breakfast cereal from food-grade sorghum. Journal of Cereal Chemistry, 73(1), 108e114. DeBelie, N., DeSmedt, V., & DeBaerdemaeker, J. (2000). Principal component analysis of chewing sounds to detect differences in apple crispness. Post Harvest Biology and Technology, 8, 109e119. Dong, G. Z., Yuan, B. Q., Gao, H. M., Li, X. L., Liu, Z. Z., & Sun, R. H. (2009). Actuality of production and processing of apple Industry in China. Journal of Innovational Edition of Farm Products Processing, 10, 64 (in Chinese). Dorta, N., & Piotr, P. (2004). Infrared drying of apple slices. Innovative Food Science and Emerging Technologies, 5, 353e360. Duan, J. L., & Wang, X. D. (2007). Study on anti-browning of vacuum freezedrying apple slice. Journal of Food Research and Development, 28(7), 102 (in Chinese). Han, Q. H., Li, S. J., Ma, J. W., & Zhao, D. L. (2006). Microwave vacuum drying and puffing characteristics of apple chips. Journal of Chinese Academy of Agricultural Mechanization Sciences, 37(8), 156e158 (in Chinese). Hofsetz, K., Lopes, C. C., Hubinger, M. D., Mayor, L., & Sereno, A. M. (2007). Changes in the physical properties of bananas on applying HTST pulse during air-drying. Journal of Food Engineering, 83(4), 531e540. Huang, G. S., Yang, G. L., & Fang, C. J. (2003). Evaluating 40 pear fruit varieties by the multidimensional value theory. Journal of Human Forestry Science & Technology, 30(4), 26e27 (in Chinese). Kara, D. (2009). Evaluation of trace metal concentrations in some herbs and herbal teas by principal component analysis. Food Chemistry, 114(1), 347e354. Lemus-Mondaca, R., Miranda, M., Grau, A. A., Briones, V., Villalobos, R., & Vegalvez, A. (2009). Effect of osmotic pretreatment on hot air drying kinetics and Ga quality of Chilean papaya (Carica pubescens). Drying Technology, 27(10), 1105e1115. Lewicki, P. P., Gondek, E., Witrowa-Rajchert, D., & Nowak, D. (2001). Effect of drying on respiration of apple slices. Journal of Food Engineering, 49(4), 333e337. Li, Y. Z., Zhen, S. X., Luo, S. C., Fei, K., & Chen, D. X. (2003). Study on the processing characteristics of potato crispy chips by vacuum microwave. Food and Fermentation Industries, 29(8), 40e43 (in Chinese).
1135
Liu, J., Wei, J. L., Liu, M. L., Song, Y., Feng, S. Q., Wang, C. Z., et al. (2012). The relationships between the enzyme activity of anthocyanin biosynthesis, ethylene release and anthocyanin accumulation in fruits of precocious apple cultivars. Journal of Acta Horticulturae Sinica, 39(7), 1235e1242 (in Chinese). Ma, L. X., Bi, J. F., & Wei, Y. M. (2005). Influence factor of explosion puffing drying process of apple. Science and Technology of Food Industry, 12, 88e93 (in Chinese). Nie, J. Y. (2009). Analytical techniques of fruit quality and safety. Beijing: Chemical Industry Press ( in Chinese). Ozilgen, M., Guvenc, G., Makaraci, M., & Tumer, I. (1995). Color-change and weightloss of apple slices during drying. Zeitschrift für Lebensmittel-Untersuchung und Forschung, 201(1), 40e45. th, E., Berna, A., Beullens, K., Yarramraju, S., Lammertyn, J., Schenk, A., et al. Ro (2007). Postharvest quality of integrated and organically produced apple fruit. Postharvest Biology and Technology, 45(1), 11e19. Shin, E. C., Pegg, R. B., Phillips, R. D., & Eitenmiller, R. R. (2010). Interrelationships among tocopherols of commercial runner market type peanuts grown in the United States. International Journal of Food Science and Technology, 45(12), 2622e2628. Sullivan, J. F., & Craig, J. C. (1980). Continous explosion-puffing of apples. Journal of Food Science, 45, 1550e1555. Sullivan, J. F., & Craig, J. C., Jr. (1984). The development of explosion puffing. Food Technology, 38(2), 52e55. Wang, H. W., & Liu, Y. Q. (2010). Evaluation of trace and toxic element concentrations in Paris polyphylla from China with empirical and chemometric approaches. Food Chemistry, 121(3), 887e892. Woodroof, J. G., & Luh, B. S. (1975). Dehydration of fruits. Commercial fruit processing. Westport, Connecticut: The AVI Publishing Company, Inc. Xu, J. H., Zhao, Z. Y., Wang, L. C., Gao, H., Liu, Z. Z., & Fan, H. K. (2011). Selection of factors for apple fruit quality evaluation. Agriculcure Research in the Arid Area, 29(6), 269e274 (in Chinese). Zou, K., Teng, J., Huang, L., Dai, X., & Wei, B. (2012). Effect of osmotic pretreatment on quality of mango chips by explosion puffing drying. LWT-Food Science and Technology, 51(1), 253e259.