Scholarly article on topic 'Passion fruit juice with different sweeteners: sensory profile by descriptive analysis and acceptance'

Passion fruit juice with different sweeteners: sensory profile by descriptive analysis and acceptance Academic research paper on "Biological sciences"

0
0
Share paper
Academic journal
Food Sci Nutr
OECD Field of science
Keywords
{""}

Academic research paper on topic "Passion fruit juice with different sweeteners: sensory profile by descriptive analysis and acceptance"

Food Science & Nutrition

l.J.lj.PJm-U

ORIGINAL RESEARCH

Passion fruit juice with different sweeteners: sensory profile by descriptive analysis and acceptance

Izabela Furtado de Oliveira Rocha* & Helena Maria André Bolini

Faculty of Food Engineering, Food and Nutrition Department, University of Campinas, R. Monteiro Lobato 80, 6121 Campinas, Brazil

Keywords

Acceptance, passion fruit juice, PLS, QDA, sensory analysis, sweeteners

Correspondence

Izabela Furtado de Oliveira Rocha, Faculty of Food Engineering, Food and Nutrition Department, University of Campinas, R. Monteiro Lobato 80, 6121 Campinas, Brazil. Tel: +55 19 3521-4084; Fax: +55 19 35213617; E-mail: izabelaforocha@yahoo.com.br

Funding Information

The present study was carried out with the support of the Brazilian Science and Technology Research Council (CNPq).

Received: 2 October 2014; Revised: 1 December 2014; Accepted: 2 December 2014

Abstract

This study evaluated the effect of different sweeteners on the sensory profile, acceptance, and drivers of preference of passion fruit juice samples sweetened with sucrose, aspartame, sucralose, stevia, cyclamate/saccharin blend 2:1, and neotame. Sensory profiling was performed by 12 trained assessors using quantitative descriptive analysis (QDA). Acceptance tests (appearance, aroma, flavor, texture and overall impression) were performed with 124 consumers of tropical fruit juice. Samples with sucrose, aspartame and sucralose showed similar sensory profile (P < 0.05), without bitter taste, bitter aftertaste, and metallic taste, and samples with sucrose and sucralose did not differ from each other for the attribute sweet aftertaste. Passion fruit flavor affected positively and sweet aftertaste affected negatively the acceptance of the samples. Samples sweetened with aspartame, sucralose, and sucrose presented higher acceptance scores for the attributes flavor, texture, and overall impression, with no significant (P < 0.05) differences between them. Aspartame and sucralose can be good substitutes for sucrose in passion fruit juice.

doi: 10.1002/fsn3.195

Introduction

Passion fruit is one of the most popular tropical fruits having a floral, estery aroma with an exotic tropical sul-fury note. The yellow passion fruit, one important commercial variety, is more acidic and mainly used for juice preparation (Deliza et al. 2005). In addition, the production of concentrated passion fruit juice increased from approximately 4400 tons in 2005 to approximately 11,200 tons in 2010 in Brazil (IBGE 2010).

Parallel to the production and consume of concentrated passion fruit juice, nowadays, sweeteners have been used in foods, driven primarily by consumer demand for foods with lower carbohydrate content and energy density compared to sugar-containing variants. Sweeteners and products formulated with sucrose replacers increase consumer choice by providing the potential to reduce calories and to enhance nutritional and health benefits. Thereby,

the availability and acceptability of the passion fruit juice in the Brazilian market and the increasing demand for low-calorie and low-sugar products should be evaluated together (De Marchi et al. 2009, Shortt 2014).

This way, sweeteners are added to foods to replace the sweetness normally provided by sucrose without contributing significantly to available energy and are a means for consumers to control caloric or carbohydrate intake (Pin-heiro et al. 2005; Trevisam Moraes and Bolini 2010; O'Mullane et al. 2014). Several sweeteners are permitted for use in diet foods and beverages, which should have low caloric density on a sweetness equivalency basis, be physiologically inert, organoleptically acceptable, commercially viable, besides assisting in weight-loss maintenance and diabetes management, and dental cavities prevention (Malik et al. 2002).

The sweeteners must be studied in low-sugar products, because that sensory characteristics, acceptance and pref-

© 2015 The Authors. Food Science & Nutrition published by Wiley Periodicals, Inc. This is an open access article under the terms of 1

the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

erence of low-calorie food products are highly dispersion matrix dependents. So, it is essential to study the substitution of sucrose by high-intensity sweeteners every time a formulation or concentration are changed or a new product is developed (De Marchi, et al. 2009; Pinhe-iro et al. 2005). And the sweeteners are only successful if they show a perfect match with the sensory profile of sucrose (Portmann and Kilcast 1996; Cadena and Bolini 2012).

Therefore, sensory evaluation is essential for its implementation in juice and blend formulations, and for the consumption of passion fruit, once the juice has intense acidic flavor, therefore water, sugar, or high-intensity sweeteners should be added to provide a palatable juice (Deliza et al. 2005).

The objective of this study was to evaluate the effect of different sweeteners on the sensory profile, acceptance, and drivers of preference of passion fruit juice samples.

Materials and Methods

Passion fruit juice samples were prepared with unsweetened concentrated juice (Da Fruta®, Araguari, Minas Gerais, Brazil). The samples were sweetened with sucrose and five different sweeteners: aspartame (Ajinomoto, Sao Paulo, Sao Paulo, Brazil), stevia extract, sucralose and neotame (Tovani-Benzaquen, Sao Paulo, Brazil), cycla-mate and saccharin (Sweet Mix, Brazil).

Physicochemical analyses

In this study, we evaluated color, pH, and soluble solid. Sample color (L*, a*, b*) was determined in a Hunterlab Colorquest II model colorimeter (Reston, Virginia, USA). The apparatus was calibrated with the D65 illuminant (6900K), the reading being carried out using a 10 mm quartz cuvette, illuminant C and hue of 10°, with Regular Transmission (RTRAN) at the moment of reading and a white reference plate (C6299 Hunter Color Standard). The pH of the samples was determined using an Orion Expandable Ion Analyzer EA 940 pH meter. The total titratable acidity was measured using AOAC (1997) and expressed as % citric acid. The percentage of soluble solids in terms of °Brix was determined using a Carl Zeiss 844976 Jena refractometer with AOAC (1997). And finally, the ratio was calculated as the ratio of total soluble solids (°Brix) to titratable acidity (Sabato et al. 2009).

Descriptive analysis

Approval for the study was obtained from the Ethics Committee of the University of Campinas, and written consent was given by all volunteers.

The sensory profiles were generated by a panel of 12 trained judges between 18 and 35 years of age, undergraduate or postgraduate students and employees, from the University of Campinas, Brazil, who were experienced in food and beverage sensory evaluation using quantitative descriptive analysis (QDA, Stone et al. 1974; Meilg-aard et al. 1999; Stone and Sidel 2004). The panelists were initially screened using the sequential method proposed by WALD (Amerine et al.1965), in which triangle tests are used to select subjects with a good ability to discriminate samples. A series of triangular tests was conducted, in which the candidates were offered two passion fruit juice samples: A (containing 35 g/L sucrose), and B (containing 50 g/L sucrose), with significant difference of 0.1%.

The parameters used in the sequential analysis were: p0 = 0.45 (maximum unacceptable ability), p1 = 0.75 (minimum acceptable ability), a = 0.10 (likelihood of accepting a candidate without sensory acuity) and b = 0.10 (likelihood of rejecting a candidate with sensory acuity). Based on these parameters, the sensory panelists were selected according to the number of triangular tests and the cumulative number of correct judgments.

Equi-sweetness determination

Initially, a study to determine the ideal sweetness of the passion fruit juice samples sweetened with sucrose was carried out. An acceptance test using a Just About Right (JAR) scale (Meilgaard et al. 2004) was performed with 60 consumers of tropical fruit juices. The samples were sweetened with sucrose at five concentrations: 5.0, 7.5, 10.0, 12.5, and 15.0 g/100 g, in order to determine the ideal sweetness according to consumer's acceptance.

After the determination of ideal sweetness, the relative sweetness of the sweeteners was measured using the Magnitude Estimation method (Stone and Oliver 1969), which makes possible a direct quantitative measurement of the subjective intensity of sweetness.

Five concentrations of each sweetener were evaluated. Firstly, the passion fruit juice sample sweetened with sucrose in the ideal concentration (reference sample) was presented, followed by the samples containing five different concentrations of each sweetener, through randomized complete sets. The subjects were served 30 mL of each sample, and 90 mL of reference sample. Each sweetener was tested in different days. Water was provided for palate cleansing.

The reference sample was taken as intensity of 100, followed by a random series of samples with intensities both less and greater than the reference. The subject was asked to estimate the sweetness intensity of the unknown sam-

ples in relation to the reference. For example, if the sample is two times sweeter than the reference, it should receive an intensity of 200, if the sample is half as sweet, the intensity should be 50, and so on. Assessors were instructed not to rate the samples' intensity as zero.

The "ideal sweetness determination" and "equi-sweet-ness determination" were described according to Rocha and Bolini (2014), and all juices were prepared to be equi-sweet. The juices with different sweeteners developed in this study are presented in Table 1.

Training and selection of panelist

Using Kelly's Repertory Grid Method described by Moskowitz (1983), panelists evaluated the samples of passion fruit juice with five different sweeteners (sucrose, aspartame, sucralose, stevia, cyclamate/saccharin blend 2:1, and neotame). The individuals received two passion fruit juice samples and individually described their similarities and differences with respect to appearance, aroma, flavor, and texture. As a group, the panelists then discussed the terms generated by each individual and, with the supervision of a panel leader, consensually defined the terms that adequately described appearance, aroma, flavor and texture similarities and differences amongst the samples, writing down their definitions and suggesting references for training purposes. In subsequent sessions, the suggested references were presented, discussed, and approved or modified by the group. During this process, eighteen sensory descriptors were consensually generated, as well as the written definitions and references for each one (Table 2).

In consensus, the panelists also elaborated a sensory descriptive term (Table 2) for the samples, associating each descriptor with a 9-cm unstructured scale, anchored at its left and right extremes by the terms "none/weak" and "strong", respectively.

After a training period, a final selection of the panelists was carried out, where each one evaluated three fruit juice samples with three replications. Analysis of variance (ANOVA) with sample (pFsampie) and replication

Table 1. Equi-sweet concentration of the sweeteners used this study.

Concentration

Sweeters equivalent (g/100 mL)

Sucrose 9.400

Aspartame 0.05477

Sucralose 0.01593

Stevia 0.09924

Cyclamate/saccharin blend 2:1 0.03584

Neotame 0.00156

(pFreplication) as source of variation was carried out for each panelist and each sample.

The level of significance for the source of variation "sample" (pFsampie < 0.50) was used as the criterion to estimate the discriminative power of each judge, and the level of significance for the source of variation "replication" (pFreplication > 0.05) was used as the criterion to estimate the reproducibility of each judge. Only individuals showing adequate discriminative power (pFsam-

ple < 0.50), reproducibility (pFreplication > °.°5), and consensus

with the rest of the panel for at least 60% of the descriptors were selected to take part in the descriptive panel (Damasio and Costell 1991).

Sensory profile

Samples of passion fruit juice (30 mL) were presented in codified white disposable cups with 3 digits, and sensory analyses were carried out in individual air-conditioned (22°C) booths with white light.

The samples were tested in complete balanced block design with tree repetitions, and the order of presentation of the samples was balanced for the first-order effect (MacFie et al. 1989).

Acceptance test

One hundred and twenty-four consumers of the tropical fruit juice evaluated all the six passion fruit juice samples to determine liking of appearance, aroma, flavor, texture, and overall impression.

Acceptance was determined using a 9-cm linear hedo-nic scale (not structured) (Stone and Sidel 2004), with anchors of "dislike extremely" on the left and "like extremely" on the right.

All samples were presented using a balanced complete block design (MacFie et al. 1989). According to Greene et al. (2006), sensitivity in defining consumer perception is greater with the use of line scales than with the 9-point hedonic scale.

The sensory profile results were performed using the SAS version 9.1.3 (SAS_ Institute, Cary, NC). The sensory descriptive data were evaluated by ANOVA (sources of variation: passion fruit juice, judges, passion fruit juice_judge) followed by Tukey's test for multiple mean comparisons (P < 0.05). The sensory and analytical data were also analyzed by Principal Component Analysis (PCA), and correlated with the analytical data using the Pearson correlation coefficient.

The acceptance results were analyzed by ANOVA, using two factors (consumer and passion fruit juice), and Tukey's test. Descriptive information obtained from

Table 2. Attributes and reference standards generated by the sensory panel.

Attributes

Definitions

References

Appearance Yellow color

Apparent viscosity

Brightness

Passion fruit

Cooked

Flavor Passion Fruit

Cooked

Sweet taste

Bitter taste

Sweet aftertaste

Bitter aftertaste

Sour taste

Sour aftertaste

Adstringency

Metallic

Texture Viscosity

Yellowish orange color characteristic of passion fruit juice

Flow rate of juice in the cup wall

The degree to which the sample reflects light in one direction

Characteristic aroma from natural passion fruit juice

Aroma due to the presence of sucrose and other sugar from passion fruit

Aroma related to the presence of characteristic organic acids from passion fruit

Characteristic aroma from passion fruit submitted to thermal processing (heat)

Characteristic flavor from natural passion fruit juice

Characteristic flavor from passion fruit submitted to thermal processing (heat)

Taste stimulated by the presence of sucrose and other substances, such as sweetner

Characteristic taste from caffeine

Sweet sensation perceived at the back of the throat after swallowing.

Bitter sensation perceived at the back of the throat after swallowing

Taste related to the presence of characteristic organic acids from passion fruit

Sour sensation perceived at the back of the throat after swallowing

Harsh sensation perceived in mouth and tongue characteristic of passion fruit

Flavor associated with "rust"/"metal"

Perceived time during swallowing

Weak: passion fruit juice concentrate (Maguary®) - 1 part pulp/20 parts water Strong: passion fruit juice concentrate (Maguary®) - 1 part pulp/1 part water Weak: passion fruit juice concentrate (Maguary®) - 1 part pulp/20 parts water Strong: passion fruit juice concentrate (Maguary®) - 1 part pulp/1 part water Weak: cooked egg yolk Strong: peach gelatin (Dr Oetker®) - prepared according to manufacturer

Weak: passion fruit juice concentrate (Maguary®) - 1 part pulp/20 parts water Strong: passion fruit juice concentrate (Maguary®) - 1 part pulp/1 part water Weak: passion fruit juice concentrate (Maguary®) - 1 part pulp/6 parts water + 5 g loaf sugar (Caravelas®) Strong: loaf sugar (Caravelas®) Weak: passion fruit juice concentrate (Maguary®) - 1 part pulp/20 parts water Strong: passion fruit juice concentrate (Maguary®) - 1 part pulp/2 part water

Weak: passion fruit juice concentrate (Maguary®) -Strong: pulp of passion fruit juice (DeMarchi®)

Weak: passion fruit juice concentrate (Maguary®) -Strong: passion fruit juice concentrate (Maguary®) Weak: passion fruit juice concentrate (Maguary®) -Strong: pulp of passion fruit juice (DeMarchi®) Weak: passion fruit juice concentrate (Maguary®) -5 g/L loaf sugar (Caravelas®) Strong: passion fruit juice concentrate (Maguary®) + 20 g loaf sugar (Caravelas®) Weak: passion fruit juice concentrate (Maguary®) -Strong: passion fruit juice concentrate (Maguary®) + 1 g/L cafeine

Weak: passion fruit juice concentrate (Maguary®) -Strong: passion fruit juice concentrate (Maguary®) + 15 g/L aspartame (Ajinomoto®) Weak: passion fruit juice concentrate (Maguary®) -Strong: passion fruit juice concentrate (Maguary®) + 2 g/L stevia extrat (Tovani-Benzaquen®) Weak: passion fruit juice concentrate (Maguary®) -Strong: passion fruit juice concentrate (Maguary®)

1 part pulp/20 parts water

1 part pulp/20 parts water

- 1 part pulp/1 part water 1 part pulp/20 parts water

1 part pulp/6 parts water +

- 1 part pulp/6 parts water

1 part pulp/20 parts water

- 1 part pulp/6 parts water

1 part pulp/6 parts water

- 1 part pulp/6 parts water

1 part pulp/6 parts water

- 1 part pulp/6 parts water

1 part pulp/20 parts water

- 1 part pulp/2 parts water

Weak: passion fruit juice concentrate (Maguary®) - 1 Strong: passion fruit juice concentrate (Maguary®) -Weak: passion fruit juice concentrate (Maguary®) - 1 Strong: cashew juice concentrate (Maguary®) Weak: passion fruit juice concentrate (Maguary®) - 1 Strong: passion fruit juice concentrate (Maguary®) -+ 0.5 g/L FeSO4

part pulp/20 parts water 1 part pulp/2 parts water part pulp/20 parts water

part pulp/6 parts water 1 part pulp/6 parts water

Weak: passion fruit juice concentrate (Maguary®) -Strong: cashew juice concentrate (Maguary®)

1 part pulp/20 parts water

the taste panel was related to the consumer preference data using partial least squares (PLS) regression (Melo et al. 2009; Bayarri et al. 2012; Cadena et al. 2012). PLS regression involved the development of a matrix data where the lines were the passion fruit juice samples (6

lines), and the columns were the eighteen attributes used by the consumers to describe the samples. Statistical analyses were carried out using XLSTATfor Windows version 2012.5 (Addinsoft, Paris, France) at a 5% significance level.

Result and Discussion Physicochemical analyses

With respect to the pH values of the passion fruit juice (Table 3), there were no significant differences (P > 0.05) between the different samples. There was a significant difference (P > 0.05) in the titratable acidity values, and the sample with sucrose presented the lowest mean.

According to Etxeberria and Gonzalez (2005) and Cadena et al. (2013), since sucrose is a soluble solid, this showed a significant influence in relation to °Brix, with a higher value in this sample. Therefore, these samples (sweetened with sucrose), showed a much higher ratio than the other samples, due to the increase in soluble solids.

There were significant differences (P > 0.05) in the color parameters (L*, a*, b*) (Table 3). The passion fruit juice samples that were lighter in color, the values for the parameter of luminosity (L*) retracting and the intensity of the yellow color (b*) were lower. The passion fruit juice samples that were darker in color, the values for the parameter of luminosity (L*) retracting and the intensity of the yellow color (b*) were higher, as in the case of the sample with sucrose. According to Brito and Bolini (2010) and Cadena et al. (2013), this darkened color could be associated with nonenzymatic processes with the formation of caramel colored pigments.

Equi-sweetness determination

The ideal sweetness analysis revealed that 9.4/100 g was the ideal sucrose concentration. The relative sweetness analysis showed that neotame presented the highest sweetening power, being 6025.64 times sweeter than sucrose in relation to passion fruit juice containing 9.4/100 g of sucrose, followed by sucralose (590.02), cyclamate/ saccharin blend 2:1 (262.28), aspartame (171.62), and ste-via (94.72). These results were described according to Rocha and Bolini (2014).

Sensory profile

Table 4 shows the results of each sample for all the 18 descriptors generated by the trained panel.

Figure 1 shows the results for the PCA to illustrate the similarities and differences amongst the passion fruit juice samples with respect to their attributes.

In Figure 1, the sensory descriptors are represented as vectors and the passion fruit juice as numbers from 1 to 6. Each sample is represented 4 times, corresponding to the repetitions performed by the descriptive panel. The Principal Components I, II and III explain 55.7% of the total sensory variation amongst the samples. This percentage can be explained because no differences were observed in passion fruit juice with sucrose and sweeteners in many attributes.

As shown in Figure 1, the position of the three samples sweetened with sucrose, aspartame, and sucralose suggests that these samples presented similar characteristics to each other.

In this study, no differences were observed in color and brightness of passion fruit juice samples with sucrose and sweeteners. These findings are consistent with samples sweetened with different sweeteners and sucrose in peach nectar (Cardoso and Bolini 2008), grape nectar (Voorpo-stel et al. 2014), and diabetic/reduced calorie chocolate (Melo et al. 2009).

No significant (P < 0.05) differences were found for all descriptive terms of the attribute aroma (passion fruit, cooked, acid and sweet), suggesting that those sweeteners had little influence on this attribute. Similar results were observed in a study on grape nectar (Voorpostel et al. 2014).

The presence of cooked aroma and cooked flavor may be due to the heat treatment of passion fruit juice (pasteurization), as described by Sandi et al. 2003.

The average scores for the attribute sweetness were significantly (P < 0.05) higher for neotame, aspartame, and stevia, differing in relation to cyclamate/saccharin 2:1 and sucrose, which exhibited the lowest scores for this attri-

Table 3. Physicochemical characteristics of passion fruit juice sample.

pH Titratable acidity (%) °Brix Ratio1 L* a* b*

Sucrose 2.82a 0.4621c 10.00a 21.64 45.6867a 6.1567a 30.5600a

Aspartame 2.85a 0.5206ab 2.00b 3.84 40.3933c 5.8933b 29.4200b

Sucralose 2.79a 0.5206ab 2.00b 3.84 40.4733c 5.8000c 29.2333bc

Stevia 2.80a 0.5262a 1.83bc 3.48 41.2267b 5.7867c 29.2133c

Ciclamate/saccharin blend 2:1 2.80a 0.5150b 1.67c 3.24 40.6700c 5.7867c 29.3433bc

Neotame 2.81a 0.5206ab 1.75bc 3.36 41.1300b 5.8167c 29.3433bc

Means with same letters in a same line each parameter indicate that samples do not have statistical difference at a significance level of 5% by Tukey's means test.

*L = luminosity; +a = red, -a = green; +b = yellow, -b = blue. 1Ratio of °Brix and titratable acidity (%).

Table 4. Attributes of the descriptive sensory evaluation by the trained panel for each passion fruit juice sample (n = 12 judges).

Cyclamate: saccharin

Sucrose Aspartame Sucralose Stevia blend 2:1 Neotame MSD*

Yellow color 5.84a 5.62a 5.82a 5.65a 5.81a 5.58a 0.3852

Apparent viscosity 2.48a 2.22a 2.38a 2.24a 2.29a 2.28a 0.4077

Brightness 7.13s 7.11s 7.15s 7.08s 7.07s 7.06a 0.3159

Passion fruit aroma 5.26s 5.14s 5.18s 5.14s 5.20s 5.13s 0.4892

Sweet aroma 2.99s 3.08s 3.11 s 3.04s 2.98s 3.26s 0.4861

Acid aroma 3.84s 3.65s 3.69s 3.78s 3.73s 3.72s 0.5341

Cooked aroma 1.26s 1.54s 1.51s 1.58s 1.54s 1.58s 0.4452

Passion fruit flavor 5.25s 5.12s 5.18s 4.67b 4.96sb 5.03sb 0.4322

Cooked flavor 1.50s 1.54s 1.41 s 1.65s 1.77s 1.50s 0.5688

Sweet taste 3.80c 4.67ab 4.03bc 4.12abc 3.53c 4.80s 0.7540

Bitter taste 0.46c 0.62c 0.87c 3.64s 1.56b 0.97c 0.5842

Sweet aftertaste 1.05de 2.35bc 1.77 cd 2.61ab 1.15 cd 3.17s 0.6815

Bitter aftertaste 0.33c 0.39c 0.64c 3.56a 1.39b 0.71c 0.5359

Sour taste 3.45ab 3.38b 3.74ab 3.53ab 3.74ab 3.95s 0.5166

Sour aftertaste 2.50sb 2.38b 2.69ab 2.60ab 2.80ab 2.94s 0.5390

Adstringency 3.48ab 3.41b 3.64ab 3.64ab 3.65ab 3.94s 0.5193

Metallic taste 0.43b 0.50b 0.52b 0.93s 0.50b 0.52b 0.3077

Viscosity 2.19s 2.14s 2.27s 2.23s 2.23s 2.13s 0.3545

Means in the same line showing common letter are not significantly different (P = 0.05). MSD, minimum significant difference.

Figure 1. PCA generated with the sensory data for appearance, aroma, flavor, and texture.

bute. The lower sweeteness observed in nectar sweetened with cyclamate/saccharin 2:1 was also reported in mango pulp by Umbelino (2005). Although this result could be inconsistent with those obtained in the sweetness equivalence test, as reported by Umbelino (2005), it may be due to the sweetness equivalence of sweet taste was assessed globally, while QDA assessed the evaluation of initial and residual sweetness separately.

Some undesirable descriptors were mentioned such as bitter taste, bitter aftertaste, sweet aftertaste, and metallic

flavor, probably due to use of sweeteners (Brito and Bolini 2010).

Samples with stevia presented the most bitter taste, bitter aftertaste, and metallic flavor, and samples with neo-tame presented the sweetest aftertaste, as also reported by Cardoso and Bolini (2008). Furthermore, the sample with stevia presented lower scores for passion fruit flavor. The undesirable descriptors bitter taste, bitter aftertaste, and metallic flavor may have hindered the passion fruit flavor (Brito and Bolini 2010).

Samples sweetened with sucrose, aspartame, sucralose, and neotame did not present bitter taste or bitter aftertaste, and sweet aftertaste was not perceived in sucrose and sucralose samples either. These occurrences were reported in other studies on these sweeteness (Cardoso and Bolini 2008; Brito and Bolini 2010; Cadena et al. 2013).

Samples sweetened with aspartame showed intermediate intensity of sweet aftertaste associated with low intensity of sour taste, sour aftertaste, and adstringency. This characteristic can be due the sweetener presented a significantly higher intensity of fruit flavor than the sample sweetened with sucrose (Cloninger and Baldwin 1974). Similar result was reported by Cavallini and Bolini (2003) in mango juices, because aspartame elicited a significantly longer persistence of fruitiness, suggesting an intensification effect on fruitiness of mango juice. Therefore, the low intensity of sour taste, sour aftertaste, and adstringen-cy may be due to the higher intensity of fruit flavor of aspartame.

Similar sensory profiles (P < 0.05) were observed for the samples with sucrose, aspartame, and sucralose, which did not exhibit bitter taste, bitter aftertaste, and metallic taste, with a higher intensity of the attribute passion fruit flavor; in addition, sucrose and sucralose presented similar results for sweet aftertaste (P < 0.05). Therefore, these results demonstrate that aspartame and sucralose are the best sucrose substitutes.

Acceptance test

Amongst the consumers (n = 124), 67.74% were female and 32.26% were male. The participants were between 18 and 30 (90.32%) and 31 or more years old (9.68%). The volunteers were PhD students (38.71%) and graduated students (54.84%).

The 124 consumers evaluated the passion fruit juice samples for appearance, aroma, flavor, texture, and overall impression. The results are presented in Table 5.

According to the results in Table 5, there was no significant difference (P > 0.05) for the attribute appearance between the passion fruit juice samples evaluated by consumers. Similar results were observed in grape nectar (Voorpostel et al. 2014), diabetic/reduced calorie chocolates (Melo et al. 2009), concentrated reconstituted pineapple juice (Marcellini et al. 2005), guava nectar (Brito and Bolini 2008a), and acerola nectar (Dutra and Bolini 2013).

Regarding the attribute aroma, the samples with aspar-tame presented high scores, followed by sucralose and cyclamate/saccharin blend 2:1. Samples with aspartame only differed from the samples sweetened with stevia. However, Marcellini et al.(2005) found no difference among samples of reconstituted pineapple juice, while Brito and Bolini (2008a) reported that the samples of guava nectar sweetened with sucrose, sucralose, and aspartame had the highest scores for this attribute.

For the attributes flavor, texture, and overall impression, the sample with aspartame showed higher sensory acceptance, and it did not differ from the sample sweetened with sucralose and sucrose (P < 0.05). This characteristic may be due to the significantly higher intensity of fruit flavor of aspartame, thus it elicited significantly

longer persistence of fruitiness (Cloninger and Baldwin 1974; Cavallini and Bolini 2003; Brito and Bolini 2008b).

Brito and Bolini (2008a) also reported that the samples of guava nectar with sucrose, sucralose, and aspartame had the highest flavor scores.

The sample sweetened with stevia had the lowest acceptance regarding the attributes aroma, flavor, texture, and overall impression (P < 0.05). Stevia presents undesirable descriptors for beverages such as bitter taste, bitter aftertaste, and metallic flavor (Table 4). Several authors have reported these stevioside characteristics, including Voo-rpostel et al. (2014) (grape nectar), Melo et al. (2009) (chocolate), Dutra and Bolini (2013) (acerola nectar), and Fernandes et al. (2009) (guava nectar). According to these studies, the samples with stevia had lower acceptance for the attribute flavor. Therefore, the lowest acceptance scores may be due to these descriptors that almost covered the sweet taste starting at concentration equi-sweet to 20% sucrose, as also reported by Bolini-Cardello et al. (1999).

The correlation between the overall impression and sensory descriptors data using PLS regression is shown in Figures 2, 3. PLS is one of the modeling approach used when predictive variables are intercorrelated (Tang et al. 2000; Melo et al. 2009). PLS is a multivariate method suitable for the analysis of sensory descriptors and overall impression by consumers. Furthermore, it may be useful to guide the selection of a subset of relevant attributes from the complete set of attributes, and the number of significant components to be evaluated is usually determined by a cross-validation procedure (Rossini et al. 2012).

The purpose of PLS is to establish the sensory attributes that are mainly related to the preference of the passion fruit juice sweetened with different sweeteners, and to determine the attributes that have contributed positively and negatively to consumer's acceptance, verifying its degree of influence (Morais et al. 2014).

According to Figure 2, the columns represent the sensory descriptors. Columns located on the positive portions of the Y axis are considered to be positively correlated with the acceptance of the passion fruit samples, while columns on the negative portion of the Y axis

Table 5. Mean scores obtained by consumers (n = 124) in the acceptance test of passion fruit juice samples.

Sucrose Aspartame Sucralose Stevia Cyclamate/saccharin Neotame MSD

Appearance 6.37a 6.44a 6.29a 6.27a 6.19a 6.16a 0.3615

Aroma 5.79ab 6.07a 5.91ab 5.55b 5.87ab 5.83ab 0.5083

Flavor 5.79a 5.87a 5.77a 2.92c 4.69b 4.24b 0.6570

Texture 6.36abc 6.55a 6.40ab 5.53d 5.96 cd 6.03bc 0.4412

Overall impression 6.09a 6.27a 5.98a 3.77c 5.28b 4.85b 0.5471

Means in the same line showing common letter are not significantly different (P = 0.05). MSD, minimum significant difference. © 2015 The Authors. Food Science & Nutrition published by Wiley Periodicals, Inc.

Figure 2. External preference map (X and Y are horizontal and vertical axes, respectively) obtained by partial least squares regression of descriptive data and respondent's overall liking scores for the sensory attributes of passion fruit juice (square = samples; circle = consumers; triangle = attributes of quantitative descriptive analysis).

represent the attributes that were negatively correlated with the acceptance of the samples (Cadena et al. 2013; Gomes et al. 2014).

The column size represents the effect (positive or negative) of the attribute on the sample acceptance, and the vertical line represents the 95% confidence interval. It

should be noted that when the vertical line crosses the X axis, the correspondent attribute does not have an influence on the drivers of preference (Gomes et al. 2014).

Furthermore, Figure 2 shows that the attribute passion fruit flavor affected positively the acceptance of the passion fruit juice samples, while the attribute sweet after© 2015 The Authors. Food Science & Nutrition published by Wiley Periodicals, Inc.

taste affected negatively. Despite a negative effect of sweet aftertaste was found in mango juice (Cadena et al. 2013), this attribute presented a positive effect on vanilla ice creams with reduced fat and sugar (Cadena et al. 2012). A positive effect of fruit flavor was found by Voorpostel et al. (2014) in grape nectar.

According to the External Preference Map (Fig. 3), the consumers (circles) were close to the samples (squares) with the highest acceptance scores. Most of the consumer groups were near the passion fruit samples with sucrose, aspartame, and sucralose, which were characterized by the attribute passion fruit flavor, whose intensity may have influenced consumers' acceptance. The neotame was associated with sweet aftertaste, sour taste, sour aftertaste, and adstringency, and stevia was associated with bitter taste, bitter aftertaste, and metallic flavor, which are undesirable descriptors for this sample. Probably the association between these attributes and those that contributed to a better acceptance of the product may have influenced the lower mean scores observed for the samples with neotame and stevia, when compared with the other samples. These results demonstrate that the presence of some types of sweeteners can influence the preference of passion fruit juice by the consumers.

Conclusion

According to QDA, the sweeteners aspartame and sucra-lose showed a sensory profile similar to sucrose, once the consumers that participated in the study preferred the samples sweetened with aspartame, sucralose, and sucrose, which received the highest scores for the attributes flavor, texture, and the overall impression.

These results have proved that aspartame and sucralose are the best sucrose substitutes, because these sweeteners presented a high intensity of passion fruit flavor, and did not present bitter taste, bitter aftertaste, and metallic taste.

The occurrence of undesirable descriptors (sweet aftertaste, bitter, bitter aftertaste, and metallic taste) is a constant problem when dealing with sweeteners. Thus, more studies are required for developing new sweeteners without these undesirable descriptors.

Acknowledgments

The present study was carried out with the support of the Brazilian Science and Technology Research Council (CNPq). The sweeteners were kindly donated by Ajinom-oto®, Tovani-Benzaquen® and Sweet Mix®.

Conflict of Interest

None declared.

References

Amerine, M. A., R. M. Pangborn, and E. B. Roessler. 1965. Principles of sensory evaluation of food. Academic Press, New York, NY.

AOAC. Official Methods of Analysis of AOAC International. 1997. Fruits juices, 16th ed. AOAC International, Gaithersburg, MD.

Bayarri, S., M. Martí, I. Carbonell, and E. Costell. 2012. Identifying drivers of liking for commercial spreadable cheeses with different fat content. J. Sens. Stud. 27:1-11.

Bolini-Cardello, H. M. A., M. A. P. A. Da Silva, and M. H. Damasio. 1999. Measurement of the relative sweetness of stevia extract, aspartame and cyclamate/saccharin blend as compared to sucrose at different concentrations. Plant Foods Hum. Nutr. 54:119-130.

Brito, C. A. K., and H. M. A. Bolini. 2008a. Arnlise da

aceitacao de nectar de goiaba por testes afetivos e mapa de preferencia interno. Revista Brasileira de Tecnologia Agroindustrial 2:67-80.

Brito, C. A. K., and H. M. A. Bolini. 2008b. Percepcao

temporal de docura, amargor e sabor de fruta em nectar de goiaba adocado com diferentes adocantes. Revista Brasileira de Tecnologia Agroindustrial 2:49-66.

Brito, C. A. K., and H. M. A. Bolini. 2010. Sensory profile of sweeteners in guava nectar. Alimentos e Nutricao (UNESP. Marilia) 20:561-572.

Cadena, R. S., and H. M. A. Bolini. 2012. Ideal and relative sweetness of high intensity sweeteners in mango nectar. Int. J. Food Sci. Technol. 12:991-996.

Cadena, R. S., A. G. Cruz, J. A. F. Faria, and H. M. A. Bolini. 2012. Reduced fat and sugar of vanilla ice creams: sensory profiling and external preference mapping. J. Dairy Sci. 95:4842-4850.

Cadena, R. S., A. G. Cruz, R. R. Netto, W. F. Castro, J. A. F. Faria, and H. M. A. Bolini. 2013. Sensory profile and physicochemical characteristics of mango níectar sweetened with high intensity sweeteners throughout storage time. Food Res. Int. 54:1670-1679.

Cardoso, J. M. P., and H. M. A. Bolini. 2008. Descriptive profile of peach nectar sweetened with sucrose and different sweeteners. J. Sens. Stud. 23:804-816.

Cavallini, D. C. U., and H. M. A. Bolini. 2003. Comparacao da percepcao temporal de docura, amargor e sabor de fruta em suco de manga reconstituido e adocado com sacarose, mistura ciclamato/sacariona 2:1, aspartame, sucralose e estevia. Boletim Centro de Pesquisa de Processamento de Alimentos 23:361-382.

Cloninger, M. R., and R. E. Baldwin. 1974.

L-aspartyl-L-phenylalanine methyl ester (aspartame) as a sweetener. J. Food Sci. 39:347-349.

Damásio, M. H., and E. Costell. 1991. Amlisis sensorial descriptivo: generacián de descriptors y selección de

catadores. Revista Agroquímica de Tecnología de Alimentos 31:165-178.

De Marchi, R., M. R. McDaniel, and H. M. A. Bolini. 2009. Formulating a new passion fruit juice beverage with different sweetener systems. Journal of Sensory Studies 24:698-711.

Deliza, R., H. MacFie, and D. Hedderley. 2005. The consumer sensory perception of passion-fruit juice using free-choice profiling. J. Sens. Stud. 20:17-27.

Dutra, M. B. L., and H. M. A. Bolini. 2013. Sensory and

physicochemical evaluation of acerola nectar sweetened with sucrose and different sweeteners. Food Sci. Technol. Campinas 33: 612-618.

Etxeberria, E., and P. Gonzalez. 2005. Accumulation of sucrose and other soluble solids in citrus juice cells. Proc. Fl. State Hortic. Soc. 118:43-48.

Fernandes, A. G., P. H. M. Sousa, G. A. Maia, D. S. Silva, and S. M. L. Santos. 2009. Avaliacao sensorial de bebidas de goiaba adocadas com diferentes agentes adocantes Cienc. Tecnol. Aliment. Campinas 29: 358-364.

Gomes, C. L., S. B. Pflanzer, A. G. Cruz, P. E. Felicio, and H. M. A. Bolini. 2014. Sensory descriptive profiling and consumer preferences of beef strip loin steaks. Food Res. Int. 59:76-84.

Greene, J. L., K. J. Bratka, M. A. Drake, and T. H. Sanders. 2006. Effectiveness of category and line scales to characterize consumer perception of fruit fermented flavor in peanuts. J. Sens. Stud. 21:146-154.

IBGE (Instituto Brasileiro de Geografia e Estatistica). 2010. Producao agrícola municipal. Culturas temporarias e permanents, 37.

MacFie, H. J., N. Bratchell, K. Greenhoff, and L. V. Vallis. 1989. Designs to balance the effect of order of presentation and first-order carry-over effect in halls tests. J. Sens. Stud. 4:129-148.

Malik, A., T. Jeyarani, and B. Raghavan. 2002. A comparison of artificial sweeteners' stability in a lime-lemon flavored carbonated beverage. J. Food Q. 25:75-82.

Marcellini, P. S., T. F. Chainho, and H. M. A. Bolini. 2005. Docura ideal e analise de aceitacao de suco de abacaxi concentrado reconstituido adocado com diferentes edulcorantes e sacarose. Alimentos e Nutricao 16:177-182.

Meilgaard, M., G. V. Civille, and B. T. Carr. 1999. Sensory evaluation techniques. CRC Press, Boca Raton, FL.

Meilgaard, M., Civille, G.V., and Carr, B.T. 2004. Sensory Evaluation Techniques. (3 ed). New York: Boca Raton 387p.

Melo, L. L. M. M., H. M. A. Bolini, and P. Efraim. 2009. Sensory profile, acceptability, and their relationship for diabetic/reduced calorie chocolates. Food Qual. Prefer. 20:138-143.

Morais, E. C., A. G. Cruz, J. A. F. Faria, and H. M. A. Bolini. 2014. Prebiotic gluten-free bread: sensory profiling and drivers of liking. LWT Food Sci. Technol. 55:248-254.

Moskowitz, H. R. 1983. Product testing and sensory evaluation of foods: marketing and RandD approaches. Food and Nutrition Press, Westport, CT.

O'Mullane, M., B. Fields, and G. Stanley. 2014. Food Additives: sweeteners. Encyclopedia of Food Safety 2:477-484.

Pinheiro, M. V. S., M. N. Oliveira, A. L. B. Penna, and A. Y. Tamime. 2005. The effect of different sweeteners in low-calorie yogurts - a review. Int. J. Dairy Technol. 58:193-199.

Portmann, M. O., and D. Kilcast. 1996. Psychophysical characterization of new sweeteners of commercial importance for the EC food industry. Food Chem. 56:291-302.

Rocha, I. F. O., and H. M. A. Bolini. 2014. Different sweeteners in passion fruit juice: ideal and equivalent sweetness. LWT Food Sci. Technol. Available at http:// dx.doi.org/10.1016/j.lwt.2014.10.055.

Rossini, K., S. Verdun, V. Cariou, E. M. Qannari, and F. S. Fogliatto. 2012. PLS discriminant analysis applied to conventional sensory profiling data. Food Qual. Prefer. 23:18-24.

Sabato, S. S., J. M. da Silva, J. N. da Cruz, S. Salmieri, P. R. Rela, and M. Lacroix. 2009. Study of physical-chemical and sensorial properties of irradiated Tommy Atkins mangoes (Mangifera indica L.) in an international consignment. Food Control 20:284-289.

Sandi, D., J. B. P. Chaves, J. F. M. Parreiras, A. C. G. Souza, and M. T. C. Silva. 2003. Sensory quality of yellow passion fruit juice (Passiflora edulisvar.flavicarpa) under pasteurization and storage. Food Res.Proc. Centre Bull. 21: 141-158.

SAS System for Windows (Statistical Analysis System). 2008. version 9.2. Cary, USA. SAS Institute Inc.

Shortt, C. 2014. 8 - Authorised EU health claims for intense sweeteners and sugar replacers. Foods Nutr. Food Ingre. Authorised Eu Health Claims, 151-176.

Stone, H., and S. M. Oliver. 1969. Measurement of the relative sweetness of selected sweeteners and sweetener mixtures. J. Food Sci. 34:215-222.

Stone, H., and J. Sidel. 2004. Sensory evaluation practices, 3rd ed. Academic Press, San Diego, CA.

Stone, H. S., J. L. Sidel, S. Oliver, A. Woosley, and R. C. Singleton. 1974. Sensory evaluation by quantitative descriptive analysis. Food Technol. 28:24-34.

Tang, C., H. Heymann, and F. Hsieh. 2000. Alternatives to data averaging of consumer preference data. Food Qual. Prefer. 11:99-104.

Trevisam Moraes, P. C. B., and H. M. A. Bolini. 2010. Different sweeteners in beveranges prepared with instant and roasted ground coffee: Ideal and equivalent sweetness. J. Sens. Stud. 25:215-225.

Umbelino, D. C. 2005. Caracteriza^ao sensorial por analise descritiva quantitativa e analise tempointensidade de suco e

de polpa de manga (mangífera indica l.) adobados com diferentes edulcorantes. Tese de doutorado, Faculdade de Engenharia de Alimentos, Unicamp. Voorpostel, C. R., M. B. L. Dutra, and H. M. A. Bolini. 2014. Sensory profile and drivers of liking for grape nectar among

smoker and nonsmoker consumers. Food Sci. Technol. Campinas 34: 164-173, Jan.-Mar. XLSTAT. 2007. XLStat version 2007. 7. Paris, France: Addinsoft SARL.