Modeling of Calcium Chloride and Pectin ... - Wiley Online Library

JFS:

Sensory and Nutritive Qualities of Food

Modeling of Calcium Chloride and Pectin Methylesterase Prefreezing Treatments of Strawberries and Jams J. SUUTARINEN, K. HONKAPÄÄ, R.-L. HEINIÖ, A. MUSTRANTA, H. LIUKKONEN-LILJA AND M. MOKKILA

ABSTRACT: Different calcium chloride (CaCl2) and pectin methylesterase (PME) prefreezing treatments in a vacuum were used to clarify the most effective prefreezing factors for strawberries and traditional jams. Fractional factorial design and modeling were used. CaCl2 had the greatest effect on the majority of the 15 studied responses (p < 0.001). The pretreatment time should be short (about 5 to 10 min), the temperature low (less than 20 °C), the vacuum level high (pressure less than 10 kPa), the CaCl2 concentration moderate (about 1%) and the dosage of PME comparatively low (about 50 to 100 nkat/g) in order to yield high quality frozen strawberries for jam making. Keywords: modeling, CaCl2, PME, prefreezing, strawberry jam

Introduction

I

N THE NORDIC COUNTRIES, THE SEASON FOR FRESH STRAWBERRIES IS


short and very few industrially processed jams are made from fresh berries during the harvesting season; instead, frozen berries are commonly used. Unfortunately, the freezing and the heating processes needed for jam making have a negative effect on fruit texture. In order to maintain the original shape of the fruit, it is sometimes necessary to pretreat it to modify its structure. The texture of plant foods is attributed to the structural integrity of the primary cell wall and the middle lamella. The middle lamella is mainly composed of pectic substances partially esterified and cross-linked with Ca2+ divalent cation. According to our earlier studies, the most effective prefreezing treatment for whole strawberries has been the use of calcium chloride (CaCl2) and pectin methylesterase (PME) treatment in a vacuum (Suutarinen and others 2000a, 2000b, 2000c, 2001). This pretreatment 1st stabilized the structure of the vascular tissue and then the cortex and pith. Fourier transform infrared (FT-IR) microscopy and bright-field microscopy studies showed that pectin, protein, and structural carbohydrate components of the pretreated strawberries were more stable than those of the untreated reference samples. The textural properties of the strawberry jams, in particular, were influenced significantly. Food research is often process-oriented, with only a limited knowledge of the underlying mechanisms. Frequently, the relevant backround variables and the nature of their interactions are unknown. Thus, the experimental results must estimate both a functional form and the parameter values for predicting the response. A number of experimental designs have been developed for food research. These designs generally require less research time than the more conventional designs such as full factorial. Experimental designs assume that random variation only exists in the output (response) variable (noise). Food studies frequently have random variation (noise) in the input variables at a level comparable to the noise in the output variable ( Thompson 1980). All models are simplifications of reality. The number of runs required and the model complexity are strongly connected. If the number of factors to be controlled is high, the number of calibration objects in the design can be reduced by suitable ap-

1240

JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002

jfsv67n3p1240-1248ms20010163-SR.p65

1240

plication of fractional factorial designs. The goals of the screening are to determine which factors are the dominating ones and which ranges should be used for these factors, and to clarify whether their ranges are in the correct area. In this way, we can detect which factors are irrelevant for future investigations and which ones must be investigated further (Cochran and Cox 1957; Martens and Næs 1989). Our aim in this study was to clarify which of the studied prefreezing factors are the dominating ones and in what ranges they should be used. Clarification of the most effective factors and defining their favorable ranges then help us to determine their optimal levels in a pilot-scale process for achieving the best prefreezing conditions and for obtaining high quality jam using these factors.

Materials and Methods Materials Strawberries: Fresh Finnish strawberries (Fragaria x ananassa, cv. Jonsok) grown in central Finland in the summer of 2000 were used in the studies. The mature strawberries were harvested and hulled manually and stored in 1-L PET (polyethylene terephtalate) cases at 5 C during cool transport and before being treated. Strawberries (5.0 kg) were randomly chosen for each treatment. Enzyme: A liquid enzyme preparation of a microbial pectin methylesterase (PME, E.C.3.1.1.11) from the microorganism Aspergillus oryzae (Novo Shape, Novo Nordisk Ferment Ltd., Dittingen, Switzerland) was used in the tests. The enzyme hydrolyzes the methyl ester groups of pectic substances, releasing pectic acid. The product is free of polygalacturonase activity. The PME activity was assayed at 23 C by titration of the liberated carboxyl groups of pectin, using an automatic titrator. The reaction mixture consisted of 1.0% (w/v) citrus pectin in 50 mM Na acetate buffer, pH 6 in a total volume of 20 mL. The pH was maintained at 6.0 by adding 50 mM NaOH to the reaction mixture. The consumption of NaOH was measured every 30 s. The initial reaction rate was calculated by plotting the amount of liberated acid (mol) versus time (s) and calculating the reaction rate from the © 2002 Institute of Food Technologists

5/3/2002, 11:31 AM

Prefreezing treatments for strawberries . . . slope of the reaction curve at initial velocity. One unit of PME activity (nkat) was defined as the amount of enzyme that liberated 1 nmol of carboxyl groups per s under the assay conditions. The PME activity of the enzyme preparation was 100 000 nkat/mL.

Table 1—Jam ingredients

Methods

Strawberries Sucrose Water Citric acid solution (50 % w/v) Preservative

randomly distributed among the experiments. Figure 1 shows a flow diagram of the experimental procedure. Strawberries were dipped in CaCl 2 solution (solution made from CaCl2 × 2H 2O, Riedel-de Haën AG, Seelze, Germany and tap water, 90 mmol/L) with PME in a vacuum. The dipping bowl was kept in a temperature-controlled vacuum chamber (W.C. Heraeus GmbH, Hanau, Germany). In all the tests 1.5 kg strawberries were dipped in a 1.8-L dipping solution. A light weight (380 g) was put on the strawberries to ensure that all the berries were submerged in the dipping solution. The total amount of strawberries per test was 5.0 kg. After dipping, the berries were drained for 5 min in a plastic strainer and directly packaged in 0.5-kg portions in 1-L PET cases (length 120 mm, width 85 mm and height 105 mm). Before freezing the strawberries were transferred into shallow perforated 1-L PET cases (length 285 mm, width 220 mm, and height 47 mm) to achieve more effective freezing inside the cases. About 30 min after each dipping, the strawberries were frozen in a liquid nitrogen freezer tunnel (AGA Freeze Mini, model 30-06, Oy AGA Fricoscandia Ab, Espoo, Finland) at –80 C for about 7 min. After freezing, the strawberries were transferred back into the taller 1-L PET cases. PET cases were stored in 2-L plastic bags at –22 C (freezer Gram HF 500-02, Denmark) for about 2 mo before jam making or thawing and analysis of calcium content and instrumental firmness.

Figure 1—Flow diagram of the experimental procedure

Pectin solution:

Pectin Sucrose Water

g/about 5.5 kg jam

%

31.2 124.7 623.7 1925.0 2023.0 729.4 37.2 5.5

0.57 2.27 11.34 35.00 36.78 13.26 0.68 0.10

Jam making: the ingredients for the jam are given in Table 1. Frozen, whole strawberries, sucrose, and water were mixed together. The mixture was heated to boiling. It was then allowed to boil for 10 min, after which citric acid (Riede-de-Haën AG, Seelze, Germany), pectin (Grinsted TM Pectin LA 410, Danisco Ingredients, Brabrand, Denmark) and preservative (Sigma-Aldrich Chemie, Germany, potassium sorbate) solutions were added. The jam was allowed to cool at room temperature for 60 min before filling 0.5-L glass jars. After filling, the jars were closed and stored at 22 C for 1 mo before analysis.

Analysis Methods Soluble solids and pH of jams: the soluble solids content of the jam medium, that is, of the gelatinized agent around the jam berries, was determined by collecting duplicate °Brix readings with an Opton 74016 (West Germany) refractometer. The pH of the jam medium was measured by collecting and averaging 5 replicate readings with an Orion Research Digital Ionanalyser 501 (Orion Research Inc., Cambridge, Mass.U.S.A.) and electrode Orion 8155SC, Orion Research Inc., Boston, Mass., U.S.A.). Calcium analyses: frozen pretreated strawberries were used for calcium analyses. In addition, the amounts of calcium in the jam strawberries and in the medium were determined. Calcium contents were determined about 2 mo after freezing or one month after jam making. The strawberries of each jam sample (230 g) were separated from the jam with a fork. The rest of the medium from the surface of the jam berries was collected with a spoon and put back into the jam sample. Frozen strawberries as well as separated jam berries and medium were weighed and homogenized before the analyses. The homogenized sample was dissolved in dilute nitric acid after dry ashing at 550 °C. Calcium was determined by atomic absorption spectrometry (AAS) after addition of lanthanum using the flame technique (method VTT4289-91, accredited by FINAS, the Finnish Accreditation Service; AOAC. 1995, Method 975.03). The uncertainty of the measurements was ± 10%. Instrumental firmness: after 2 mo of freezer storage, the pretreated strawberries were thawed for analyses by keeping them in their storage cases at 5 ± 1 C for 24 h. Thawed fruits were equilibrated to 17 C by storing them at 20 ± 1 C for 6 h. The thawed strawberries were then dried with tissue paper and weighed into 120-g portions for firmness measurements. The compression force of strawberries was measured with a Texture Analyser (model TA-HDi, Stable Micro Systems, Goladming, England) with a 250-kg load cell using an Ottawa Cell (A/OTC) with a Holed Extrusion Plate (A/HOL). The starting position of the plunger was 50 mm from the base and the final position was 1 mm above the base plate. The plunger speed was 1.5 mm/s. The area of the deformation curve was recorded as the result. The result was an average of 5 to 6 replicates (Suutarinen and others 2000c). Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE

jfsv67n3p1240-1248ms20010163-SR.p65

1241

5/3/2002, 11:31 AM

1241


S

TRAWBERRY PRETREATMENTS: THE FRESH , WHOLE FRUITS WERE

Ingredient

Prefreezing treatments for strawberries . . .


After jam making, jams were stored at 22 C for 1 mo. Strawberries were separated with a spoon from the jam medium onto a plate. For firmness measurements, strawberries or the medium were weighed into 80-g portions. The compression force of strawberries or of the medium was measured using an A/BE/45 Back Extrusion Rig. The starting position of the plunger was 50 mm from the base and the final position was 1 mm above the base plate. The plunger speed was 1.5 mm/s. The area of the deformation curve was recorded as the result. In both cases, the result was an average of 3 to 4 replicates in the case of jam strawberries and 6 in the case of medium (Suutarinen and others 2000c). Instrumental jam color: for instrumental jam color measurements, jam berries were separated from the jam medium with a fork. The rest of the medium from the surface of the jam berries was removed with a spoon. Hunter a (green-red), b (blue-yellow) and L (lightness) parameters of jam berries and medium were determined with a Minolta Chroma Meter CR-200 (Minolta Camera Co., Ltd., Osaka, Japan) in the reflection mode after 1 mo of jam storage. The instrument was standardized with a white ceramic plate (Calibration Plate CR-A43). The tip of the sensor was placed against the surface of berries. Three points on the surface of 6 berries of each test jam were measured. The result was an average of 18 replicates of the berries of each test jam. Medium from each test jam was transferred to the bottom of a transparent plastic box (length 55 mm and width 37 mm) to form a 1-cm layer (about 23 g). The tip of the sensor was positioned against the bottom of the box and 6 measurements from different points of the bottom were carried out. The result was an average of 6 replicates of each test jam medium. Sensory evaluation: sensory evaluations of the jams were carried out after 1 mo of storage. The sensory quality of the strawberry jams was evaluated by a trained panel with proven skills (N = 10 to13) using descriptive analysis (Stone and others 1974). The vocabulary of the sensory attributes was developed by describing differences among several widely varied strawberry samples (of different cultivars and of different shelf life) by a 5member expert panel in a round table session. The assessors were familiarised with the sensory descriptors and the attribute intensities used prior to the evaluations. The multi-product panel was particularly trained for profiling the sensory quality of strawberry jams in several pre-sessions before the actual evaluations, and proper model foodstuffs describing the ends of intensity scales of the attributes were available for the assessors. Attribute intensities were rated on continuous, unstructed 10-unit graphical intensity scales, which were anchored at both ends. The left side of the scale corresponded to the lowest or opposite intensity (value 0) and the right side to the highest intensity (value 10) of the attribute. The sensory attributes evaluated, verbal descriptions of the ends of the scales and the standard model foodstuffs used during the training and evaluation procedures were the following: redness of color (brown-red): syrup-red from an ICI Colour Atlas red-orange-yellow section, wholeness of berries (broken-whole): mashed strawberries-completely whole strawberries, clarity of medium (opaque-clear, transparent): supersaturated confectioner’s sugar solution-water, evenness of medium (uneven-even): cottage cheese-whipped cream, firmness of medium (firm-fluid): marmelade-fool, softness of berries (hard-soft): toffee candy-shortcake and faultlessness of odor and flavor (defective-faultless): water solutions of possible treatment-based defects such as bitter, sour, or salty off-odor or offflavor- fresh, faultless strawberry jam. The samples were presented to the panelists coded and in random order, and water and crackers were provided for cleansing the palate between the 1242

Table 2—Experimental design used in the test Testa 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time (min)

Temperature (°C)

Vacuum (kPa)

CaCl2 (%)

PME (nkat/g)

5 15 5 15 5 15 5 15 5 15 5 15 5 15 5 15 10 10 10 10

10 10 40 40 10 10 40 40 10 10 40 40 10 10 40 40 25 25 25 25

6.67 6.67 6.67 6.67 40 40 40 40 6.67 6.67 6.67 6.67 40 40 40 40 23.3 23.3 23.3 23.3

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 2 2 2 2 2 2 2 2 1.05 1.05 1.05 1.05

500 50 50 500 50 500 500 50 50 500 500 50 500 50 50 500 275 275 275 275

a Does not correspond to order of processing.

jam samples. The samples were analysed on 2 sequential d in 2 sensory replicates; both divided into 2 sessions in which the untreated reference sample was presented to the assessors hidden among the other jam samples. No difference in the results of the untreated reference sample assessed in the 2 sessions was observed, and the mean of these results was used as the reference sample value in the statistical analysis.

Statistics Experimental Design: a fractional factorial screening interaction was used. The resolution of the design was V+, which meant that the main effects were unconfounded with 2-factor interactions and 2-factor interactions were unconfounded with each other. The design contained 16 tests with 4 center points (Tests 17-20 in Table 2). The 3-level factors of the tests were the following: pretreatment time (X1) 5, 10, or 15 min; temperature of the pretreatment solution (X 2) 10, 25, or 40 °C; vacuum level (X3) 6,67, 23.3, or 40 kPa; CaCl2 concentration (X4) 0.1, 1.05, and 2% or PME dosage (X5) 50, 275, or 500 nkat/g (0.50, 2.75, and 50 mL/kg fresh strawberries) (Table 2). The tests were carried out in random order. The approximate estimations of enzyme and CaCl 2 dosages as well as dipping times, temperatures and vacuum strength were based on our earlier experience of the effects of these pretreatments. Before the study, a preliminary fractional factorial design was carried out using the following 3-level factors: pretreatment time 5, 17.5, or 30 min; temperature of the pretreatment solution 20, 35, or 50 °C; vacuum level 6.7, 36.7, or 66.7 kPa; CaCl2 concentration 0.5, 2.5, or 4.5% and PME dosage 50, 275, or 500 nkat/g. The resolution of the preliminary study was IV which meant that the main effects were unconfounded with 2-factor interactions but that 2-factor interactions were confounded with each other. According to the preliminary study, ranges of all other factors except PME could be decreased (data not shown). Furthermore, a CaCl 2 level of 4.5% in the pretreatment solution was too high to make properly gelatinized jam with the pectin used in this study. The use of pectin with a different calcium sensitivity would probably have enabled the use of higher CaCl 2 dosages in this study. However, according to our earlier experiments, a high amount of calcium in jam caused a


jfsv67n3p1240-1248ms20010163-SR.p65

1242

5/3/2002, 11:31 AM

Prefreezing treatments for strawberries . . . Table 3—Overview of significant main effects and 2- factor interactions for the studied responses X1 Ca in frozen berries Ca in jam berries Ca in medium Firmness of jam berries Firmness of jam medium *L (jam) *L (medium) *a (medium) *b (medium) Wholeness of berries Clarity of medium Evenness of medium Softness of berries Firmness of medium Faultlessness of odor and flavor

X2

X3

X4

** * *

X5

*** *** *** *** *** *** *** *** ***

* *

***

X1X5

X2X4

X2X5

X3X4

X4X5

* * * * ***

***

*** * * *

*

**

***

**

*** *** *** *** **

***

* *** **

***

p < 0.001, ** p < 0.01, * p < 0.05. X 1 is the pretreatment time, X 2 the temperature of the pretreatment solution, X 3 the vacuum level, X 4 the CaCl 2 concentration and X 5 the amount of PME.

Calcium of frozen berries (mg/kg)

Calcium of jam berries (mg/kg)

Calcium of jam medium (mg/kg)

N 19 19 20 Range 40.9 – 49.4 3.4 – 3.9 177 – 576 46.1 3.7 330.0 苵x a sxa 2.18 0.14 129.25

19 117 – 316 224.6 56.79

19 56 – 156 109.0 33.93

Parameter

°Brix

pH

a

nent. Raw data were not preprocessed before analysis. Stepwise analysis of screening proceeded as follows: after examination of raw data worksheets and visual histograms, all factors and interaction terms were used in the 1st model calculations. For model qualifying analysis of variance (ANOVA), correlation coefficient R, prediction coefficient Q2, estimate of noise (residual standard deviation, RSD), normal probability plot of residuals and effect plot were used. Multiple correlation squared R2 tells how much the model can explain of the total variation of the response variable (goodness of fit). R2 is calculated from:

*N is the number of samples, 苵x is the mean and s x the standard deviation. The calcium content of the tap water used for analysis was 18 mg/L.

R2 = (SS Total Corrected – SS Residual) / SS Total

bitter off-flavor (data not shown). The experimental design was created and analysed using Modde for Windows 4.0 software (Umetri Ab, Umeå, Sweden). Statistical analysis of experimental design:in the model each of the studied response variables (Yk) occurs as a linear term with the following form: Yk = o + 1X1 + 2X2 + ... +

(1)

where o is constant, 1 and 2 are regression coefficients, X1 and X2 are factors and is noise. In our study, the interaction terms were determined by addition. The resulting model was the following: Yk = o + 1X1 + 2X2 + 12X1X2 + ... +

(2)

Regression coefficients for each linear model were calculated using partial least squares (PLS) regression with 95% confidence levels. PLS is a projection method which models both X- and Ymatrices simultaneously to find the latent variables in X that will best predict the latent variables in Y. These PLS-components are called principal components. Model components are extracted in such a way that most information is conveyed by the 1st principal component, then by the 2nd and so on. PLS has been shown to be effective for eliminating noise due to inaccuracy of measurements (Martens and Næs 1989). PLS dealt with only 1 response variable at a time. Thus, every response was modelled independently of all the others. Results of the PLS fitting were achieved with the 1st principal compo-

Corrected

SS Residual = SS Lack of Fit + SS Pure Error

1243

(4)

where SS is the sum of squares, SS Lack of Fit is the model error and SSPure Error is the replicate error. The lack of fit test indicated whether the model was as good as the repeatability results. The pure error was repeatability expressed in original units. Multiple prediction squared Q2 reveals how good the model really is (prediction strength). It is calculated in the same way as R2 but each predicted point is eliminated alternately from the calculation and then predicted using the remainder of the points (that is, using cross validation). Insignificant terms in the model were reduced and the model was refined until the best model was achieved, that is high R2 ( > 0.7) and Q2 ( > 0.6) values (not separated by more than 0.2 values) with an acceptable probability value (p < 0.05) and without model error (p0.05, not lack of fit) with moderate RSD values. Finally, the robustness of the models was checked by plotting them. For example, normal probability plots of residuals and effect plots. Contour plots were used to indicate the favorable ranges of the studied factors with respect to some of the response variables. Additionally, a linear regression method was performed to analyse the interrelations of the response variables with a Statgraphics Plus software package (Ver. 5.0, Manugistics, Inc., Rockville. Md., U.S.A.).

Results and Discussion

I

N THE PRESENTED REGRESSION EQUATIONS EVERY FACTOR OR ITS IN

teraction term was requested to be below the significance level of 5% (Table 3). Soluble solids and pH of jams: all values of soluble solids were Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE

jfsv67n3p1240-1248ms20010163-SR.p65

(3)

5/3/2002, 11:31 AM

1243


Table 4—Descriptive statistics of measured °Brix (soluble solids), pH and calcium

Prefreezing treatments for strawberries . . . Table 5—Quality parameters of models of calcium contents of frozen strawberries, jam berries and medium Ca

R2 a

Q2 b

pc

Frozen berries Jam berries Jam medium

0.98 0.89 0.94

0.94 0.80 0.87

1.06 3 10–13 1.89 3 10–6 2.14 3 10–8

LoF

d

0.66 0.07 0.34

RSD/sx e 0.13 0.38 0.27

Table 6—Descriptive statistics of instrumental firmness values of jam berries and medium Parameter a

a R 2 is the multiple correlation squared, b Q 2 the multiple prediction squared; c p the probability of a type I error (should be < 0.05 for good models), d LoF the lack of fit (should be 0.05 for good models); an e RSD/s x of about 1 means that the standard deviation cannot be decreased, an RSD/s x of about 0 means that all the standard deviation is explained by the model.

rather similar, about 45 °Brix (Table 4), although the amounts of soluble solids were not adjusted to a certain level in jam cooking but the jams were cooked using similar total amounts of each ingredient. The pH-values of the jams were also very similar, about 3.5 (Table 4). It was not possible to model soluble solids and pH. Thus, the 5 studied factors did not appear to have an effect on the soluble solids or pH of the jams over the ranges investigated. Calcium analyses: descriptive statistics of the calcium content of frozen berries as well as of jam berries and media are shown in Table 4. On the basis of visual examination of the histograms, all the parameters except the calcium contents of the jam berries, were normally distributed. Quality parameters of the final models of calcium contents are shown in Table 5. Regression coefficients were high and RSD/sx values low. The regression equations of calcium in frozen berries (Y5) as well as in jam berries (Y6) and medium (Y7) were the following: Y5 = 157.13 + 140.15 X4 + 0.94 X2

(5)

Y6 = 155.03 + 56.03 X4 + 3.42 X1 - 0.01 X5

(6)

Y7 = 66.44 + 26.10 X4 - 0.11 X3 + 0.44 X3X4

(7)

N Range 苵x a s xa

Firmness of berries (kg s)

Firmness of medium (kg s)

20 51.9 – 238.3 131.5 46.96

20 3.5 – 21.1 12.3 5.95

3

3

a N is the number of samples, 苵 x is the mean and s x the standard deviation.

(R = 0.67, p < 0.01) and medium (R = 0.86, p < 0.001); instrumental firmness of the jam berries (R = 0.89, p < 0.001) and medium (R = 0.87, p < 0.001); instrumental lightness, (R = –0.62, p < 0.01), redness (R = –0.81, p < 0.001), yellowness (R = –0.74, p < 0.001) as well as sensory softness of jam berries (R = –0.81, p < 0.001) and also with clarity (R = –0.73, p < 0.001), evenness (R = –0.80, p < 0.001) and firmness (R = 0.71, p < 0.001) of the jam medium. Thus, raising the CaCl2 concentration in frozen berries increased the firmness of the jam berries and medium but decreased the color coordinates as well as the clarity, and evenness of the jam medium. Instrumental firmness of jam: descriptive statistics of the firmness results of jams after are presented in Table 6. On the basis of visual examination of the histograms, the firmness values of jam berries and medium were not normally distributed. The explanatory power of the models were still good. Quality parameters of the final models of instrumental firmness in jam berries and medium are shown in Table 7. The regression coefficients were high and the RSD/sx values rather low. The regression equations of the instrumental firmness of the jam berries (Y8) and of the medium (Y9) were the following: Y8 = 94.10 + 24.26 X4 + 0.84 X2 - 0.76 X3 - 0.004 X5 + 0.007 X4X5


The factor with the greatest effect on calcium was naturally CaCl 2. Furthermore, in frozen berries, temperature, and in jam berries, pretreatment time favored increasing calcium content. The calcium content of jam medium increased when the vacuum level decreased. According to García and others (1996), a 1% CaCl 2 solution was the most effective treatment for increasing the calcium content of the fruits, for controlling their post-harvest decay, and for maintaining their firmness and soluble solids content. Furthermore, heating to 45 °C enhanced penetration of the calcium into the fruit in 1% CaCl2 treatment. The positive effect of temperature on calcium content in frozen berries was also demonstrated in our study. According to Polesello and Maltini (1970), the best temperature for rapid impregnation of peaches with sucrose and calcium was 15 °C, higher temperatures giving a lower quality product. Contour plots for the effects on the calcium content of frozen and jam berries are shown in Figure 2a-b. CaCl 2 dominated the amount of calcium in frozen berries compared to the temperature in Figure 2a. Maximum calcium content of frozen berries (456 mg/kg) was achieved with a 1.85 to 2% CaCl2 solution in the temperature range of 21 to 40 °C. In Figure 2b, both of the studied factors, CaCl2 and time, affected the calcium contents of jam berries. The highest calcium content (300 mg/kg) was achieved with a 1.75 to 2% CaCl2 solution with a pretreatment time of 11 to 15 min. The dosage of PME was 50 nkat/g. Based on the linear regression results, calcium in frozen berries correlated significantly with the calcium in jam berries 1244

Y9 = 4.66 + 9.07 X4 + 0.05 X3 - 0.11 X3X4

1244

(9)

Instrumental firmness of jam berries increased when CaCl 2, temperature, and vacuum level increased. Firmness of jam medium was affected by CaCl2 alone in the concentration range of 0.1 to 1.0 %. Above this concentration, increasing vacuum level also increased firmness of jam medium. The pretreatment time appeared to have no effect on the firmness of the jam berries or of the medium. The contour plot for the effects on the firmness of jam berries is shown in Figure 2c. Both of the studied factors, CaCl2 and temperature, clearly had an effect on the firmness of jam berries in Figure 2c. The highest firmness value of 186 kgs was achieved with a 1.75 to 2% CaCl2 solution at a pretreatment temperature of 27 to 40 °C. The vacuum level was 6.67 kPa and the PME dosage 275 nkat/g. Use of the linear regression method in comparing the mean values of instrumental firmness of jam berries and medium showed that jams with firmer berries also had firmer medium (R2 = 0.87, p < 0.001). According to Grassin and Fauquembergue (1998), the pretreatment of apple sauce with 30 to 500 PE units/kg pectin esterase (PE) and a CaCl2 concentration of 250 ppm with a reaction time of about 10 min led to considerable firmness. The temperature of the reaction was not critical in the range 10 to 70 °C. Firmness results of the frozen strawberries that were thawed and treated with PE and CaCl2 (PE from 0 to 8600 PE units/kg of treat-


jfsv67n3p1240-1248ms20010163-SR.p65

(8)

5/3/2002, 11:31 AM

Prefreezing treatments for strawberries . . .

Firmness

R2 a

Q2 b

Jam berries Jam medium

0.85 0.87

0.73 0.83

p

c

2.55 10–5 4.16 10–7

LoF

d

0.06 0.09

RSD/sx e 0.45 0.39

a R 2 is the multiple regression squared, b Q 2 the multiple prediction squared; c p the probability of a type I error (should be 0.05 for good models), d LoF the lack of fit (should be 0.05 for good models); an e RSD/s x of about 1 means that the standard deviation cannot be decreased, an RSD/s x of about 0

means that all the standard deviation is explained by the model.

ed strawberries and CaCl2·2H2O 136 mg per kg treated strawberries (0.05% w/w)) following the incubation (at 50 C for 60 min) and heating to 85 C before storing in a cold chamber for 48 h, showed an optimum enzyme effect at a dosage of 2240 PE units/ kg. When studying the effects of PE, holding time, and holding temperature on strawberry firmness with the aid of experimental design, a temperature increase from 40 to 50 C was found to have the greatest effect on firmness improvment. The enzyme amount (6650 to13300 PE units/kg) could not be considered as significant and holding time (5 to 15 min) also had little effect on the process (Coutel and Dale, 1994).

Table 8—Descriptive statistics of instrumental color coordinates of jam berries and medium Parameter Na

Jam berries *L

19 Range 26.9 – 34.6 x苵 a 29.6

*a

Jam medium *b

19 19 17.6 – 27.1 7.4 – 12.9 22.2 9.08

*L

*a

*b

19 19 19 22.7 – 27.1 6.8 – 18.3 1.7 – 8.1 24.6 12.3 4.4

a N is the number of samples, 苵 x is the mean and s x the standard deviation. *L, lightness; *a, redness; and *b, yellowness. Increasing values of ‘L’, ‘a’ and ‘b’ indicate lighter, redder and yellower jam berries or medium, respectively.

Instrumental jam color: descriptive statistics of the color measurements of the jam berries and media are given in Table 8. Color values of jam berries were normally distributed, but those of jam medium were not quite normally distributed. According to the results, jam berries were lighter, redder, and yellower than jam media. The quality parameters of the final color models are shown in Table 9. The regression coefficients were high and the RSD/sx values rather low. The regression equations of lightness of jam berries (Y10) and lightness (Y11), redness (Y12),and yellowness (Y13) of jam medium were the following:

(a)

(b)

(c)

(d)

Figures 2a-b—Contour plot for the effects on the calcium content of frozen berries. Contour plot for the effects on the calcium content of jam berries. Contour plot for the effects on the firmness of jam berries. Contour plot for the effects on the redness of jam medium. Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE

jfsv67n3p1240-1248ms20010163-SR.p65

1245

5/3/2002, 11:31 AM

1245


Table 7—Quality parameters of models of instrumental firmness of jam berries and medium

Prefreezing treatments for strawberries . . . Table 9—Quality parameters of models of instrumental color coordinates of jam berries and medium Color

R2 a

Q2 b

*L *L *a *b

0.94 0.92 0.92 0.89

0.73 0.74 0.74 0.75

(jam berries) (medium) (medium) (medium)

p

c

2.10 10-6 8.76 10–5 4.93 10–5 6.92 10–5

LoFc

d

RSD/sx e

0.78 0.64 0.48 0.37

0.31 0.29 0.35 0.41

a R 2 is the correlation coefficient, b Q 2 the ‘prediction’ coefficient; c p the probability of a type I error (should be 0.05 for good models), d LoF the lack of fit (should be 0.05 for good models); an e RSD/s x of about 1 means that the standard deviation cannot be decreased, an RSD/s x of about 0 means

that all the standard deviation is explained by the model. *L, lightness; *a, redness; and *b, yellowness.

Y10 = 35.89 – 3.33 X4 – 0.11 X2 – 9.59 × 10–4 X5 + 0.04 X2X4 + 5.81 × 10–4 X4X5

(10)

Y11 = 25.95 – 1.25 X4 – 3.94 × 10–3X2 + 1.12 × 10–4 X5 + 0.01 X2X4 – 7.62 × 10–6 X2X5

(11)

Y12 = 15.55 – 4.02 X4 + 0.08 X2 + 5.87 × 10–4 X5 + 0.02 X2X4 – 1.49 × 10–4 X4X5 - 3.08x10–5 X2X5

(12)

Y13 = 6.34 – 2.12 X4 + 0.05 X2 + 1.18 × 10–4 X5 + 0.01 X2X4 – 1.56 × 10–5 X2X5

(13)

Table 10—Descriptive statistics of sensory evaluation (N = 10–13) of jams Attribute

Na

Range

x ·a

sxa

Redness of color Wholeness of berries Clarity of medium Evenness of medium Softness of berries Firmness of medium Faultlessness of odor and flavor

19 19 19 19 19 19 19

6.5-7.6 5.9-8.6 6.0-8.2 4.1-8.3 5.3-7.6 3.6-8.4 7.8-8.6

7.1 7.6 7.4 6.0 6.5 6.9 8.2

0.29 0.65 0.65 1.24 0.74 1.49 0.22

a N is the number of samples, 苵 x is the mean and s x the standard deviation.

It was not possible to model redness or yellowness of jam berries. This meant that the pretreatments did not affect these parameters in the ranges used. By contrast, models of the lightness of jam berries as well as of the lightness, redness and yellowness of the jam medium were good. The pretreatment time had no effect on the studied color responses over the given range. Generally, the lightness of the jam berries and of the medium increased when temperature and CaCl 2 decreased. PME favored increasing lightness of the jam medium. The redness and yellowness of the jam medium increased when CaCl 2 decreased and the temperature increased when a PME dosage of 50 nkat/g or

(a)

(b)

(c)

(d)


Figure 3a-d—Contour plot for the effects on the sensory wholeness of jam berries. Contour plot for the effects on the sensory clarity of medium. Contour plot for the effects on the sensory softness of jam berries. Contour plot for the effects on the sensory faultlessness of odor and flavor. 1246


jfsv67n3p1240-1248ms20010163-SR.p65

1246

5/3/2002, 11:31 AM

Prefreezing treatments for strawberries . . . Table 11—Quality parameters of models of sensory attributes R2a

Q2b

pc

LoF d

Wholeness of berries Clarity of medium Evenness of medium Softness of berries Firmness of medium Faultlessness of odor and flavor

0.74 0.91 0.82 0.93 0.75 0.87

0.58 0.85 0.75 0.84 0.69 0.81

9.81 10–4 5.64 10–6 2.66 10–6 7.96 10–7 2.53 10–4 4.01 10–4

0.57 0.64 0.37 0.26 0.24 0.21

(17)

Y18 = 6.50 + 0.88 X4 – 7.09 × 10–5 X5

(18)

Y19 = 8.24 – 0.01 X2 – 0.01 X4 + 9.92 × 10–5 X5 + 3.51 × 10–3 X2X4 – 4.50 × 10–5 X4X5

(19)

RSD/sxe 0.54 0.27 0.46 0.32 0.33 0.42

a R 2 is the multiple correlation squared, b Q 2 the multiple prediction squared; cp the probability of a type I error (should be 0.05 for good models), d LoF the lack of fit (should be 0.05 for good models); an e RSD/s x of about 1 means that the standard deviation cannot be decreased, an RSD/s x of about 0 means that all the standard deviation is explained by the model.

500 nkat/g was used. This was partly due to the fact that if the temperature is increased, the danger of loss of red color from fresh strawberries and from jam berries into medium is also increased. Furthermore, excess CaCl2 had a negative effect on the lightness of jam berries and on all the other color coordinates of the jam medium. The contour plot for the effects on the redness of jam medium is shown in Figure 2d. Both of the studied factors, CaCl 2 and temperature, had a clear effect on the redness of jam medium in Figure 2d. The highest redness value of jam medium (16.7) was achieved with 0.1 to 0.4% CaCl 2 at a pretreatment temperature of 21 to 40 C. The PME dosage was 50 nkat/g. The use of the linear regression method showed that there was a high positive correlation between redness and yellowness (R = 0.93, p < 0.001) and a relatively high correlation between redness and lightness of (R = 0.74, p < 0.001) of the jam berries. In addition, a very high correlation was found between redness and yellowness (R = 0.99, p < 0.001) as well as redness and lightness (R = 0.91, p < 0.001) of the jam medium. Redness of the jam medium, on the other hand, correlated positively with clarity (R = 0.73, p < 0.001) and evenness (R = 0.84, p < 0.001) but negatively with firmness (R = –0.82, p < 0.001) of the jam medium. Lightness and yellowness of the jam medium had similar effect on the previous variables. Sensory profiles of strawberry jams:descriptive statistics of the sensory results are shown in Table 10. The values of all the sensory attributes except clarity and firmness of medium as well as softness of berries were normally distributed. The ranges of redness and faultlessness of odor and flavor were rather narrow. More intense processing conditions may have made the differences between the jams more pronounced. On the other hand, when studying the effects of 1 pretreatment the effects of other factors should be kept to a minimum. Quality parameters of the final models are shown in Table 11. The regression coefficients were fair and the RSD/sx values rather low. The regression models of the sensory attributes wholeness of berries (Y14), clarity (Y15) and evenness (Y16) of medium, softness of berries (Y17), firmness of medium (Y18), and faultlessness of odor and flavor (Y19) were the following: Y14 = 8.59 – 0.28 X4 – 0.02 X3 – 2.37x10–4 X5 + 1.65 × 10–4 X4X5

(14)

Y15 = 7.64 – 0.49 X4 + 0.01 X2 + 6.81 × 10–5 X5

(15)

Y16 = 7.51 – 1.31 X4 – 4.65 × 10–5 X5

(16)

When modeling sensory data it was important to bear in mind that there were freezing and cooking operations and storage of jams between pretreatments and sensory analyses. Furthermore, sensory data consisted of complex data, which complicated modeling. CaCl2 was the most significant factor for evenness and firmness of medium. According to the effect plots, the 2nd most effective factor was PME, but its effect was almost insignificant compared to CaCl2. A quadratic term of CaCl2 would have been justified in these models. However, in linear models the quadratic term of a factor, in this case CaCl2, cannot be used. Thus, the factor PME was added to the models referred to above despite the fact its effect was not below the significance level of 5 %. Redness of color could not be modelled. This meant that the studied factors did not have an effect on the redness at the ranges investigated. This was in agreement with the instrumental redness results of the jam berries. According to García-Viguera and others (1999), a jam boiling time longer than 15 min had a detrimental effect on jam color. Simulated daylight conditions caused insignificant color and anthocyanin losses during storage compared with jams stored in total darkness. The jam boiling time in our study was less than 15 min. This might be one reason for the sensory redness results of the studied jams. Thus, the pretreatment factors in the given ranges together with the used cooking time had no effect on redness of jam color. CaCl2 had a significant effect on all the other studied sensory parameters. The use of a vacuum, was shown to have a significant effect on the wholeness of the berries, PME and temperature affected the clarity of the medium and also, temperature had an influence on the faultlessness of odor and flavor. The pretreatment time alone did not affect the sensory attributes in the given ranges. However, its interaction term with PME had a significant effect on the softness of berries. Contour plots for the effects on the sensory attributes, the wholeness of berries, the clarity of medium, the softness of berries, and the faultlessness of odor and flavor are shown in Figures 3a-d CaCl2 and vacuum had clear effects on the wholeness of jam berries in Figure 3a. The maximum score value (8.16) was obtained with 0.1 to 0.35 % CaCl 2 in a vacuum level of 6.67 to 12 kPa. The PME dosage was 50 nkat/g. By contrast, in Figure 3b, CaCl 2 dominated the effect on the clarity of jam medium. The maximum score value (7.84) was received with 0.1 to 0.3% CaCl2 at a temperature of 26 to 40 °C. The PME dosage was 50 nkat/g. The highest softness value of jam berries (7.32) was obtained with 0.1 to 0.3% CaCl 2 and a pretreatment time of 11 to15 min and PME dosage of 50 nkat/g (Figure 3c). In Figure 3d, the temperature and PME clearly affected the faultlessness of odor and flavor. The highest value (8.50) was obtained at a temperature of 10 to17 °C and with a PME dosage of 400 to 500 nkat/g. The CaCl2 concentration was 0.1%. Use of the linear regression method in studying the sensory attributes and the other response variables showed the following: the instrumental firmness of jam berries correlated negatively with the softness (R = –0.73, p < 0.001) and positively with Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE

jfsv67n3p1240-1248ms20010163-SR.p65

1247

5/3/2002, 11:31 AM

1247


Attribute

Y17 = 6.64 – 0.75 X4 + 0.07 X1 + 2.90 × 10–4 X5 – 3.06 × 10–5 X1X5

Prefreezing treatments for strawberries . . . the wholeness (R = 0.64, p < 0.05) of jam berries. The instrumental firmness of jam medium, again, had a high negative correlation with the evenness (R = –0.85, p < 0.001) of the jam medium and the softness (R = –0.82, p < 0.001) of jam berries. The instrumental firmness of the jam medium correlated positively with its sensory firmness (R = 0.83, p < 0.001) whereas the evenness of the jam medium correlated highly negatively with its sensory firmness (R = –0.94, p < 0.001). The yellowness of the jam medium recorded a high positive correlation with the evenness of the same (R = 0.81, p < 0.001). The evenness of jam medium had a significant positive correlation with the clarity (R = 0.76, p < 0.001) of medium and the softness (R = 0.80, p < 0.001) of jam berries. Thus, the jams in whch the firmness of the jam berries and medium increased had decreased clarity of the medium. Similarly, the firm medium was on average less even and the clear medium was more even. The pretreatments at the studied levels did not affect °Brix- or pH-values, instrumental or sensory redness values, or yellowness of jam berries. CaCl2 concentration had the greatest effect on most of the studied respones (p < 0.001). On the one hand, with high concentrations of CaCl2 the firmness of jam berries and medium increased but on the other hand the clarity and evenness of the medium decreased. On the basis of the results obtained, PME was a significant factor for calcium uptake in jam berries (p < 0.05), as well as for instrumental lightness of jam berries (p < 0.001), yellowness of jam medium (p < 0.05), and sensory clarity of medium (p < 0.05). The lowest dosage of PME (50 nkat/g) was sufficient to improve the texture of strawberries. Synergy of PME and CaCl2 was needed to increase the instrumental firmness (p < 0.05), lightness (p < 0.001) and sensory wholeness (p < 0.001) of jam berries. The level of vacuum affected calcium in jam medium (p < 0.05), firmness of jam berries (p < 0.05), and sensory wholeness of berries (p < 0.01). Temperature affected absorption of calcium into berries (p < 0.01), instrumental firmness (p < 0.05), and lightness (p < 0.001) of jam berries as well as sensory clarity of jam medium (p < 0.01) and faultlessness of odor and flavor (p < 0.001). The pretreatment time did not play a major role in the study except in the model of calcium in jam berries (p < 0.05). According to Hoover and Miller (1975), over 50% of the absorption of CaCl2 into apple slices was achieved during the first 30 s of vacuum impregnation. Two-factor interaction terms of CaCl2, temperature, PME, or vacuum were significant and they were included in several models of the studied responses.

than Ca contained rather complex information. However, PLS regression appeared to be a good estimation method, partly due to the ease with which it can be extended and modified for different types of data. For achieving high quality and sensorally acceptable jams from pretreated strawberries, the use of high CaCl2 concentrations and high temperatures of the pretreatment solution should be avoided. For further optimization of the quality of pretreated strawberries and jams a response surface methodology (RSM) of the most significant factors (temperature, vacuum, CaCl 2, and PME) should be performed. This would make it possible to identify the processing conditions for optimal quality of the processed strawberries and jams.

Conclusion

This research is a part of a project aimed at enhancement of the industrial use of Finnish strawberries. We thank the National Technology Agency and the participating Finnish Companies for supporting the project. Dr. Juho Rousu from the Department of Computer Science, University of Helsinki, is gratefully acknowledged for his critical comments on the statistics. Special thanks to Anne Ala-Kahrakuusi, Marika Lyly, Pirkko Nousiainen, Ulla Österlund and Eeva-Liisa Peltokorpi from VTT Biotechnology for their skillful technical assistance in carrying out the prefreezing treatments as well as in jam making and sensory evaluation.

O

N THE BASIS OF VISUAL EXAMINATION OF THE HISTOGRAMS, THE

ranges of the data of the studied variables were mostly normally distributed and most of the terms had sufficient variation to find satisfactory models. Modeling of sensory variables of jam was partly complicated by the freezing and cooking operations and the storage after pretreatments and partly by the complexity of the sensory data itself. Furthermore, response variables other


1248

References AOAC. 1995. Official Methods of Analysis. 16th ed. Association of Official Analytical Chemists and AOAC International, Arlington, VA 22201-3301, USA. Cochran WG, Cox GM. 1957. Experimental Designs. New York. John Wiley & Sons Ltd. 611 p. Coutel YAG, Dale RHS, inventors; Gist-Brocades Co., assignee. 1998 Nov 26. A method for fruit processing. WO patent 98/52423. García JM, Herrera S, Morilla A. 1996. Effects of postharvest dips in calcium chloride in strawberry. J Agric Food Chem 44: 30-33. García-Viguera C, Zafrilla P, Tomás-Barberán FA. 1999. Influence of processing and storage conditions in strawberry jam color. Food Sci Tech Int 5: 487-492. Grassin CMT, Fauquembergue PCL, inventors; Gist-Brocades Co., assignee. 1994 June 9. Use of pectinesterase in the treatment of fruit and vegetables. WO patent 94/12055. Hoover MW, Miller NC. 1975. Factors influencing impregnation of apple slices and development of a continuous process. J Food Sci 40: 698-700. Martens H, Næs T. 1989. Multivariate Calibration. New York: John Wiley & Sons Ltd. 419 p. Stone H, Sidel J, Oliver S, Woolsey A, Singleton RC. 1974. Sensory evaluation by quantitative descriptive analysis. Food Tech 28:24-34. Polesello A, Maltini E. 1970. Studies on uptake of calcium salts during pretreatment of peaches for freezing. Industrie-Agrarie 8: 199-205. Suutarinen J, Heiska K, Autio K, Mokkila M 2000a. The effect of CaCl2 and PME prefreezing treatment in a vacuum on the structure of strawberries (abstract). In: Hietaranta T, Linna M-M, editors. 4th International Strawberry Symposium Book of Abstracts; 2000 July 9-14; Tampere, Finland: Kaarinan Tasopaino Oy Ltd. p 228. Suutarinen J, Heiska K, Moss P, Autio K. 2000b. The effect of calcium chloride and sucrose prefreezing treatments on the structure of strawberry tissues. Lebensm Wiss Technol 33: 89-102. Suutarinen J, Honkapää K, Heiniö R-L, Autio K, Mokkila, M. 2000c. The effect of different prefreezing treatments on the structure of strawberries before and after jam making. Lebensm Wiss Technol 33: 188-201. Suutarinen J, Honkapää K, Mokkila M, inventors; Research Centre of Finland, assignee. 2001 May 3. A process for preparing jam. Int Patent Appl No WO patent 01/30178. Thompson DR. 1980. Food research experimental design: effects of variation. Transactions of the ASAE 23 (4): 1034-1039. VTT Biotechnology. 1991. Foodstuffs. The determination of metal contents after dry ashing by atomic absorbtion spectrometry. Method VTT-4289-91 (in-house method). MS 20010163 Submitted 4/2/01, Accepted 8/23/01, Received 10/23/01

The authors are with VTT Biotechnology, P.O. Box 1500, FIN-02044 VTT (Finland). Direct inquiries to author Suutarinen (E-mail: Marjaana. [email protected])


jfsv67n3p1240-1248ms20010163-SR.p65

1248

5/3/2002, 11:31 AM