The effect of sugar substitution on model confectionary systems

Excessive consumption of added sugar is associated with many health problems, for example obesity, type 2 diabetes, etc. Hence there is an urgent need for the product reformulation by total replacement or partial reduction of sugar in food industry. The aim of this research was to study the effect of sugar substitution (by stevia and xylitol) on model confectionary systems. We investigated differences in the texture properties, the viscosity and thermal properties of the blends. Based on our results, the sugar substitution affects the physical properties of the measured samples. The apparent viscosity and the texture properties were changed due to the different dry matter content in the samples. In the differential scanning calorimeter (DSC) curves the different melting of the samples were expressed according to the changes in sugar content. Further work is needed in this field to follow up the discovered changes in thermal behaviour of these mixtures.


INTRODUCTION
In the world around 30% (2.1 billion) inhabitants are overweight or obese (Dobbs et al., 2014;Miele et al., 2017).According to the mortality rate, the deaths associated with excessive weight gain are more prevalent than those caused by underweight (Miele et al., 2017;WHO, 2016a).Excessive sugar intake is generally considered to contribute to excessive caloric intake, lower diet quality and adverse health affect such as dental caries, obesity, type 2 diabetes, cardiovascular disease and some cancers (Luo et al., 2019).In modern diet, sugar gives around 25% of the 2,000 kJ recommended daily calorie intake (Lustig et al., 2012), which is far more than the maximum 10% recommended by the WHO (Miele et al., 2017).Approximately 80% of youth (Wang et al., 2008) and 63% of adults (Bleich et al., 2009) consume sugar-sweetened beverages every day.
Due to the above, there is an urgent need for sugar reduction and/or substitution in food products.The key tool for sugar replacing is the use of alternative sweetening agents that provide less energy per gram than sugar (Buttriss, 2017).However, delivering the same texture and sugar-like sweetness with partial or full sugar reduction is a challenging task as sweeteners typically display an unusual sensory characteristic and different texture (Moraes and Bolini, 2010).
Use of alternative sweeteners almost inevitably requires trade-offs and the best choice will probably depend on the matrix and recipe of the food, legal restrictions on certain food categories, and the expectations and preferences of consumers (Buttriss, 2017).
In confectionary products the main ingredients are fats and sugars.Generally, in creams and fillings palm oil is used due to its wide applicability.
Studying the crystallization phenomenon of fats (e.g.palm oil) is essential for food products.However, fats and oils are rarely present as a bulk material in processed foods.Rather, they often act as a continuous matrix that incorporates non-fat ingredients for example sugar.The dispersed ingredients may alter the crystallization pathway and rigidity (Fernandes et al., 2013;Walstra, 2002), with potential consequences on techno-functional and sensory properties of the products (Beckett, 2000;West and Rousseau, 2019).
The processing parameters of fat itself influence the behaviour of a lipid based system, as it was shown by Dhaygude et al. (2018) in case of coconut oil.
The aim of this research was to study the effect of sugar substitution (by stevia and xylitol) on model confectionary systems.Here, we investigated differences in the texture properties, the viscosity and thermal properties of the blends.

MATERIALS AND METHODS
Commercial sugar (sucrose), stevia and xylitol were used in this study.Palm oil was purchased from Palmfood Kft.The oil was creamed with an equal mass fraction (50 wt.%) of sugar manually at room temperature (23 8C).Different sweeteners were added according to their sweetening power, thus xylitol was used in the same proportion as sucrose (50% fat þ 50% sugar/xylitol), whereas stevia was added in a smaller ratio (80% fat þ 20% stevia).

Apparent viscosity
The apparent viscosity of the samples was analyzed with Brookfield DV-E rotation viscometer.The tests were made at 30 8C with increasing speed from 0 to 20 RPM with LV4 cylindrical spindle.

Thermal properties
Thermal properties were analysed by differential scanning calorimeter (DSC) (NETZSCH DSC 3500) previously calibrated with indium.The samples (14.0 ± 1 mg) were hermetically closed in concave aluminium pans and heated in a calorimeter from 25 to 80 8C, cooled to À20 8C and heated again to 200 8C at constant rate 5 8C min À1 .An empty aluminium pan was used as reference.Temperatures (T oonset, T ppeak, T eend) and enthalpy of thermal transitions (ΔH-Area, J g À1 ) were determined with the use of the instrument's software Proteus Analysis.

Texture analyses
Hardness, work of penetration and adhesiveness were measured using a texture analyzer (TA.XT2i, Stable Micro Systems) calibrated with a 2 kg load cell.Plastic cups were filled with the room temperature (23 8C) samples (∼10 g).For analysis, a 458 steel cone was calibrated to a 10 mm height above the sample surface and programmed for approach at 2 mm s À1 .Upon contact with its surface, the cone penetrated the sample at 1 mm s À1 for 80% of the original height, at which point the measured force value was recorded as hardness.The area under the positive curve was determined as work of penetration, which is the energy required to deform the samples to the defined distance.Withdrawal of the cone probe from the sample provides information about adhesive characteristics of the samples.

Statistics
Data were analyzed by one-way analysis of variance (one-way ANOVA) to test the variations in different samples, P < 0.05 was considered statistically significant.If there were significant differences in variance, paired-samples t-test was used.All measurements were done at least triplicate.

Apparent viscosity
In the literature one of the commonly used rheological model is Casson model (Barbosa et al., 2016).This used to describe the flow behaviour of high sugar and fat content matrixes such as e.g.molten chocolate.
ffiffi ffi In our case, the parameters in the Casson equation were calculated from the measured apparent viscosity and speed values.The calculated points of the shear rate and shear stress for all investigated mixtures are shown in Fig. 1.
In every case the shear stress value increased with increasing shear rate.The addition of stevia decreased the shear stress values, while sucrose and xylitol increased it, compared to pure fat.The observed reduction of the viscosity in stevia sample was due to the lower dry matter content of the mixtures.The results of the sucrose and xylitol samples were influenced by the water content of the ingredients, which is related to the hygroscopic behaviour of the materials.

Texture analysis
The presence of dispersed phase influenced the texture properties of the sample (Fig. 2).Hardness and the work of penetration values increased with the addition of sucrose and xylitol compared to pure fat.However, in the case of stevia these values decreased, due to the lower dry matter content of the samples.In every case the adhesiveness of the mixtures decreased compared to pure fat.This effect can be related to the mixing ratios.From the statistical analyses we found that, in work of penetration only stevia, in adhesiveness only palm oil were significantly different compared to other samples.In case of the measured hardness surprisingly there was no significant difference between pure fat and the palm oilþ stevia mix, even though the dry matter content has changed by adding stevia into the oil.Mixing sugar or xylitol to the palm oil in 50-50% resulted in significant harder samples.There was also significant difference between these two mixtures.

Thermal properties
The crystallization and the melting behaviour of the samples were analysed with DSC (Fig. 3).
In Table 1 the onset, peak, area and the end temperature values of melting are shown.Generally, the addition of sugar resulted in a high peak on the melting temperature curves of sugar or sweetener, unlike the fat-connected peaks, which were negligibly small in these cases.The Onset, Peak and End temperature of the samples were shifted according to the melting characteristic of the samples.Xylitol melted first, than stevia and saccharose at the highest temperature around 190 8C.However, the enthalpy of the mixtures showed a reverse result, with the increasing temperature values the enthalpy values decreased.All the measured parameters were significantly different among the three samples.
In Table 2 the onset, peak, area and end temperature values of the crystallization are shown.In every case 2 peaks can be seen in the curves.According to our results, the main crystallization temperatures were almost the same in every sample, which is related to the crystallization characteristic of the fat crystals.These values were not affected by the dispersed sugar or sweeteners particles.The first Onset temperature was slightly influenced by the presence of sucrose or xylitol.However, the enthalpy values showed bigger differences.According to the statistical analyses in the first enthalpy values only between sugar and xylitol were not significantly different.In the second enthalpy values of palm oil showed significant difference to the other samples and xylitol and stevia were also different.In every case a reduction in enthalpy can be seen compared to pure fat.This can be due to the palm oil content of the samples, which is influenced by the mixing ratios.Therefore, it is possible that sugar and xylitol gave almost the same results, because they were mixed with 50% solid matter ratio.Stevia was added in lower amount because of its higher sweetening power, it resulted in a higher enthalpy value.

CONCLUSIONS
The effect of sugar and sugar substitution on a model confectionary system was examined.The objective of this study was achieved through the measurement of apparent viscosity, texture and thermal properties of the samples.The different parameters were influenced by the addition of sucrose, stevia or xylitol to pure palm oil.Shear stress of palm oil with xylitol and with sucrose increased compared to pure fat.Adhesiveness of the mixtures decreased independently from the type of sweetener.Hardness of mixtures increased due to their increased solid matter content.The work of penetration was significantly smaller in the case of xylitol compared to sugar, which could be related to its smaller particle size; however, this was not verified in this research.In the melting of the samples sugar or sweetener resulted in high peak in their melting temperature, pure fat peak was negligibly small in every case.In crystallization all the enthalpy values decreased compared to palm oil due to the mixing ratios of the blends.Most of the changes were due to a different dry matter content of the samples, which was defined by the sweetening power of different ingredients.However, the sweetening power of the samples was not tested by organoleptic evaluation.Therefore, this work will continue towards sensory experiments and more complex rheological modelling.

Fig. 1 .
Fig. 1.The average shear stress and shear rate calculated by the Casson model (n 5 3)

Table 2 .
Crystallisation characteristics of the mixtures (average± standard deviation)