Carbohydrate-based fat replacers in the modification of the rheological, textural and sensory quality of yoghurt: comparative study of the utilisation of barley beta-glucan, guar gum and inulin


*Fax: +64 6 350 5657;


Barley beta-glucan, partially hydrolysed guar gum and inulin were used in the processing of low-fat yoghurts. The possible beneficial effects of carbohydrate fat replacers on the rheological, textural and sensory quality of low-fat yoghurt-based products were determined. Comparisons were made between the sample yoghurts made from a low-fat milk base, and full-fat and low-fat yoghurt controls. The inclusion of the carbohydrate components reduced product syneresis and improved the texture and rheological properties of the low-fat-based products so that their quality characteristics were similar to yoghurt made with full-fat milk. Both the type and also the amount of carbohydrate component altered product characteristics. Beta-glucan addition at low level (0.5%) was effective in improving serum retention of the yoghurt and its viscoelastic nature (G′, G′ and tan δ). In contrast, higher levels (above 2%) of inulin and guar gum were needed to exert significant improvements in the textural characteristics of yoghurt. Sensory analysis conducted on the samples illustrated that the inclusion of carbohydrate-based fat replacers could be successfully utilised to mimic full-fat products.


Many health organisations consider the level of fat consumption as too high. A recent World Health Organisation (WHO) report recommended that the level of total fat intake should be between 15% and 30% of the energy, of which saturated fatty acids should account for less than 10% energy (WHO, 2003). There is continued market interest in the use of low-fat dairy products. However, producing such foods is not a straightforward task. The presence of fat in dairy products has considerable impact on their physical properties, rheological and textural characteristics and microbiological stability. In addition, fat influences other product characteristics such as handling, stability, appearance, flavour and mouthfeel. For example, low-fat yoghurts may be poorly acceptable mainly because of the rheological characteristics and increased syneresis, while the main problem associated with low-fat cheese is related to changes in texture (becoming rubbery and hard) and flavour.

While most consumers are aware of the health benefits of low-fat diets, they are not prepared to sacrifice the taste, texture and aroma they enjoy in the dairy products in exchange for healthier products (McIlveen & Armstrong, 1995). Thus, the goal of the food industry is to respond to consumer demand and to offer an increasing variety of low fat choices, in which the attributes that the consumers desire are not impaired.

A reduction in fat content can be achieved by replacing it with water and several ingredients to control this water and to provide the functionality of the missing fat. Carbohydrate-based fat replacers, such as dietary fibres (DF), have been used safely as thickeners and stabilisers especially in sauces and dressing formulations. They can be effective fat replacers in various foods. Several papers have focussed on the use of specific soluble DF in dairy products, such as the use of carrageenan, gellan gum or guar gum in cheese production (Kailasapathy, 1996, 1998; McMahon et al., 1996) and yoghurt (Tamime et al., 1994;Fernández-Garcia & McGregor, 1997; Keogh & O’kennedy, 1998).

This study aimed to contribute to the understanding of the behaviour of certain soluble DF in milk formulations designated to be used in yoghurt production. Barley beta-glucan, inulin and partially hydrolysed guar gum (PHGG) were chosen to be investigated for their effects on products texture and rheological properties.

Materials and methods

Food-grade barley beta-glucan (Glucagel™, 86% beta-glucan; PolyCell Technologies LLC, Crookston, MN, USA) inulin (Frutafit-HD; Seneuiker Unie & Sensus, Dinteloord, The Netherlands), and PHGG (Sunfibre, Selectchemie AG, Switzerland). Concentrated frozen starter culture used consisted of Lactobacillus delbrueckii subsp. bulgaricus and Streptococcus thermophillus (1:1) (CH-3 Chr Hansen A/S, Hørsholm, Denmark). Pasteurised, standardised and homogenised milk batches (3.2% fat and 0.1% fat) were used for the tests.

Manufacture of yoghurt

Yoghurts were manufactured using standardised, homogenised milk. The DF were dispersed in milk at 60 °C for 5 min using a high-speed mixer (model SL2T, Silverson Machines Ltd., Bucks, UK) operating at 8000 r.p.m. DF were incorporated into a skimmed milk base at levels of 0.5%, 1.0%, 1.5%, 2.0% and 2.5% for beta-glucan, and 2.0%, 4.0% and 6.0% for inulin and PHGG (based on preliminary rheological data, not shown). The mixtures were heated at 95 °C for 10 min, cooled to 44 °C, inoculated with 0.04% of the frozen starter culture and incubated at 44 °C to a final pH of 4.2. The coagulum was then broken and stirred yoghurt was stored at 5 °C for 2 days prior to testing. Full fat and skimmed milk yoghurt samples were used as controls (without any addition of DF). Additionally, a skimmed milk yoghurt sample with an addition of 2.0% skimmed milk powder (SMP) was used as a comparison with the fibre-enriched samples in order to determine the effect of increasing the solid content of the yoghurts by 2%. Duplicate trials were conducted.

Yoghurt syneresis

Yoghurt samples (30 g) were centrifuged at 222 g for 10 min at 4 °C. After centrifugation, the supernatant was poured off, weighed and recorded as percentage of syneresis (Keogh & O’kennedy, 1998).

Rheological properties of the yoghurt

Apparent viscosity of the samples was measured using a Brookfield DVIII Rheometer (Brookfield Engineering Laboratories, Inc., MA, Brookfield, USA) and a cylinder in cylinder geometry (probe SC4-21). All the measurements were conducted at 5 °C and 0.5 r.p.m. Yoghurt was gently stirred for 20 s before testing. Triplicate measurements were conducted.

Rheological properties of yoghurt samples were investigated using a controlled stress rheometer (AR1000, TA Instruments, Crawley, UK). Measurements were carried out on shear mode at 5 °C, using a cone and plate geometry (the gap between the plate and the base was of 50 μm). A shear rate sweep test was used with the shear rate ranging from 10−2 to 20 s−1. A frequency sweep test was also performed (with the frequency ranging from 1 to 20 Hz at a maximum strain of 4.06E-03, and an amplitude of 1.42E-04). Triplicate measurements were performed for each sample.

Textural properties of the yoghurt

Textural characteristics of the yoghurt were determined using a TA.XT2 Texture Analyser (Stable Micro Systems, Surry, UK) using a load cell of 5 kg. Back extrusion was conducted using an A/BE back extrusion cell, and the tests were run at the following settings: test speed: 1 mm s−1, post-test speed: 1 mm s−1, distance: 25 mm, rate for data acquisition: 200 pps. Peak positive force represented product firmness, and the negative region of the graph represented the product consistency/resistance to flow. Five replicates were performed for each sample.

Sensory analysis of yoghurt

Descriptive sensory analysis was performed under normal light, in the sensory laboratory at the University of Plymouth. Samples were placed in clear plastic pots. A panel consisting of fourteen semi-trained panellists was used for the evaluation. Five training sessions were held prior to testing. In these sessions, the panellists were trained in the products and descriptors were chosen on the basis of consensus among panellists, using yoghurt products available on the market to cover a range of consistencies (from full-fat Greek yoghurt to low-fat stirred yoghurt). A total of fourteen descriptors were used for the assessment of the product’s appearance, texture, taste and overall acceptability. Test samples, identified by a three-digit code, were presented to the panellists in a randomised order, immediately after being removed from the fridge (4 °C). Testing was conducted on duplicate samples, and each panellist was asked to assess them for each attribute on a seven-point scale. The form, which was filled in by each panellist, is presented in Table 3.

Table 3.   Mean values for texture sensory attributes of dietary fibre-enriched yoghurt
 FF_coSM_co0.5% beta-glucan1.5% beta-glucan2% inulin6% inulin2% PHGG6% PHGG2% SMP6% SMPSignificance
  1. Within the same column, the values with the same letter are not significantly different.

  2. ***P < 0.001; **P < 0.01; *P < 0.05; NS, not significant; PHGG, partially hydrolysed guar gum; SMP, skimmed milk powder; FF_co, full-fat control; SM_co, skimmed milk control.

Texture/body (as perceived by scooping)
 Whey separation4.2a3.8a,b2.4a,b,c1.7c4.0a1.6c2.1a,b,c1.7b,c2.8a,b,c1.2c***
Texture/body (as perceived in the mouth)
 Untypical Taste/flavour2.
 Overall acceptability3.

Statistical analysis

Results from all the tests were calculated as means SD. Analysis of variance (one-way anova and General Linear Model (GLM)) followed by Tukey’s test of Minitab 13.1 software (Minitab Inc., State College, PA, USA) was used for statistical analysis. Data from the experiments involving beta-glucan were analysed using one-way anova, while GLM was used to analyse the data from the experiments involving inulin and PHGG.

Results and discussions

The effects of dietary fibre on yoghurt characteristics

The syneresis results (%) of the yoghurts are illustrated in Fig. 1 and summarised in the anovaTable 1 (for samples containing inulin and PHGG). Yoghurt made from full-fat milk retained significantly higher percentage of serum within its structure, thus being characterised by decreased syneresis in comparison to the yoghurt made from skimmed milk. These results are in agreement with previous studies (Keogh & O’kennedy, 1998) and may be explained by the presence of fat globules which may limit casein aggregation, preventing the shrinkage and rearrangement of the three-dimensional network into a more compact structure (Fox et al., 2000).

Figure 1.

 Syneresis of yoghurt: (a) samples containing beta-glucan; (b) samples containing inulin, partially hydrolysed guar gum and skimmed milk powder.

Table 1. anova table summarising the rheological attributes and syneresis behaviour of yoghurt containing inulin and partially hydrolysed guar gum (PHGG)
SampleSyneresisApparent viscosity (Pa*s)Apparent viscosity (mPas)Storage modulus (Pa)tan δ
  1. Within the same column, the values with the same letter are not significantly different.

  2. ***P < 0.001; **P < 0.01; *P < 0.05; NS, not significant; FF_co, full-fat control; SM_co, skimmed milk control; SEM, standard error of mean.

FF_co49.31.819 967b824b0.280c
SM_co58.61.216 900c557c0.280c
Effect of the type of dietary fibre (DF)
 Inulin50.71.619 811b783b0.285b
 PHGG47.92.925 856a1611a0.294a
Effect of the level of DF addition
 2%52.82.318 933b850b0.284b
 4%47.22.0424 683a1371a0.289b
 6%46.82.4226 721a1370a0.297a

Yoghurt samples containing beta-glucan showed a significantly reduced syneresis in comparison with the low-fat control yoghurt (P < 0.001), and increasing levels of beta-glucan in the formulations led to decreasing values for syneresis (Fig. 1a). Yoghurt containing 0.5% beta-glucan had a syneresis level comparable with the full-fat control yoghurt while addition of higher percentages of beta-glucan significantly improved the ability of the yoghurt to retain larger amounts of serum within the structure. This low susceptibility to syneresis was previously reported for yoghurt-like fermented oat products and it was attributed to the ability of the beta-glucan to entrap water within the three-dimensional network of the product (Martensson et al., 2001).

A similar trend could be observed for the yoghurts containing inulin or PHGG (Fig. 1b). Their effect is not as strong as seen for beta-glucan; the mean values presented in Fig. 1b and Table 1 indicate that both inulin and PHGG reduce the syneresis of low-fat yoghurt, bringing it at levels comparable with full-fat control yoghurt especially at levels of addition above 2%.

Yoghurt rheological and textural characteristics

Rheological measurements were performed using both a controlled stress rheometer and a Brookfield rheometer and different geometries (cone and plate and coaxial cylinders, respectively). Apparent viscosity data is presented as values at specific testing points: at a shear rate of 10 s−1 for the tests performed on the controlled stress rheometer, and at 0.5 r.p.m. and a shear rate of 0.47 s−1 for the tests performed on the Brookfield rheometer. The effect of beta-glucan addition on the apparent viscosity of low-fat yoghurt is illustrated in Fig. 2. The viscosity of low-fat yoghurt is significantly lower than that of its full-fat counterpart. The use of beta-glucan in low-fat formulations significantly increased the viscosity of the product as suggested by Fig. 2a,b. Thus, when added at levels of 1% (w/w) or higher, the viscosity of the product was significantly improved in comparison with low-fat and full-fat control samples. This improvement in product viscosity is likely to be related to increased total solids present in the formulations, as yoghurt manufacturers use this method to improve the texture and reduce the syneresis of their products. To test this hypothesis, rheological measurements were also performed on low-fat yoghurt containing 2% added SMP. However, Fig. 2 illustrates that beta-glucan additions of 1.5% (using the controlled stress rheometer) and 0.5% addition (using the Brookfield rheometer) increased apparent viscosity of the yoghurts more than an addition of 2% SMP.

Figure 2.

 Apparent viscosity of yoghurt samples containing beta-glucan as determined with: (a) a controlled stress rheometer – at a shear rate of 10 s−1; (b) Brookfield rheometer – at 0.5 r.p.m. and a shear rate of 0.47 s−1.

The effects of inulin and PHGG and various levels of addition on the apparent viscosity of low-fat yoghurt are summarised in the anovaTable 1 and illustrated in Fig. 3. Regardless of the instrument used for rheological measurements, statistical analysis suggests that yoghurts containing PHGG were significantly more viscous than those containing inulin (and those containing 2% SMP addition). Both inulin and PHGG appear to improve the viscosity of the products in comparison with full-fat control sample. However, while inulin brings the viscosity of the low-fat product towards values closer to those of full-fat control yoghurt, PHGG significantly increased the viscosity of the products in comparison with both the controls (Fig. 3 and Table 1).

Figure 3.

 Apparent viscosity of yoghurt samples containing inulin and partially hydrolysed guar gum (PHGG) as determined with: (a) a controlled stress rheometer – at a shear rate of 10 s−1; (b) Brookfield rheometer – at 0.5 r.p.m. and a shear rate of 0.47 s−1.

Dynamic/oscillatory rheological tests were performed to complement the large deformation tests by providing useful information on the product’s viscoelastic behaviour. These tests gave information on the elastic modulus (G′), viscous modulus (G′) and the loss tangent (tan δ) presented in Table 1. Values of G′ for full-fat control samples were higher than low-fat control yoghurt samples, suggesting a stronger gel structure. Figure 4a indicates that yoghurts containing beta-glucan are characterised by increased G′ values in comparison with low-fat control yoghurt samples (except for 1% beta-glucan). This may be because of the increased levels of total solids in samples containing beta-glucan, and also potential thermodynamic incompatibility between casein and beta-glucan. It is possible that the beta-glucan promotes self-association of casein resulting in increased gel strength (higher G′).

Figure 4.

 Storage modulus (G′) and tan δ (at 59.9 rad s−1) of yoghurt samples containing dietary fibre as evaluated with the controlled stress rheometer: (a) G′ for samples containing beta-glucan; (b) tan δ for samples containing beta-glucan; (c) G′ for samples containing inulin and partially hydrolysed guar gum (PHGG); (d) tan δ for samples containing inulin and PHGG.

The tan δ values of yoghurts containing beta-glucan (above 1%) also support this idea as they are lower in comparison with both control samples and also the sample with 2% SMP (Fig. 4b), in that the higher the levels of beta-glucan, the lower the tan δ values. Thus, the presence of beta-glucan in yoghurt formulation modifies its viscoelastic behaviour, intensifying its gel-like characteristics. The viscoelastic characteristics (G′ and tan δ) of the yoghurts containing inulin and PHGG are summarised in Table 1 and presented in Figs 4c,d and 5c,d. Statistical analysis indicate that yoghurts containing PHGG were characterised by significantly higher G′′ and tan δ values than those containing inulin (P < 0.001). Moreover, increasing levels of inulin/PHGG resulted in increasing values for both G′ and tan δ (P < 0.001). When compared with the controls (and the sample with 2% SMP), the averaged values for G′ and tan δ of yoghurts enriched with PHGG were higher than those corresponding to both control samples. For yoghurts containing inulin, the overall average tan δ and G′ values indicated also a slight increase in comparison with those of the low-fat control sample.

Figure 5.

 Tanδ of yoghurt samples containing dietary fibre as evaluated with the controlled stress rheometer: (a) control samples; (b) samples containing beta-glucan; (c) samples containing inulin; (d) samples containing partially hydrolysed guar gum (PHGG).

The rheological characteristics of yoghurts containing either PHGG or inulin suggest that the behaviour of these DF differs from the behaviour of beta-glucan. Both inulin and especially PHGG lead to an increase in tan δ of the low-fat yoghurt samples (Fig. 5c,d). This indicates a shift in the viscoelastic behaviour of these yoghurts towards more viscous-like materials although the yoghurt formulations of both PHGG and inulin samples appear to form gel-like structures alongside the casein matrix.

Textural attributes of DF-enriched yoghurts – gel strength/firmness and consistency – were evaluated following a back-extrusion test performed on a Texture Analyser. The results for the samples containing beta-glucan and inulin are presented in Fig. 6 (texture analysis was not performed on samples containing PHGG). Full-fat and low-fat control samples were not significantly different from each other in terms of their firmness or consistency. This supports previous research results (Schmidt & Bledsoe, 1995). The incorporation of either beta-glucan or inulin in yoghurt formulation resulted in an increase in product firmness and consistency in comparison with both the control samples. The highest firmness and consistency of beta-glucan products was obtained for formulations containing 2.5% addition level. The texture results along with the rheological results are in agreement with the trends observed for yoghurts syneresis; increased gel strength (G′ and firmness) would make yoghurt less susceptible to rearrangements within its network, and consequently less susceptible to shrinkage and serum (whey) expulsion.

Figure 6.

 Textural attributes of yoghurt samples containing dietary fibre: (a) firmness for samples containing beta-glucan; (b) firmness for samples containing inulin; (c) consistency for samples containing beta-glucan; (d) consistency for samples containing inulin.

Sensory evaluation of yoghurts

Eight types of yoghurt, out of the range of samples, were selected for sensory evaluation on the basis of the results obtained from rheological and microstructural investigations. The samples subjected to sensory evaluation were: full-fat control (FF_co), low-fat control (SM_co), beta-glucan inclusion at 0.5%, 1.5%, inulin inclusion at 2%, 6%, guar gum at 2%, 6% and SMP at 2% and 6%, respectively.

Yoghurt samples were significantly different only in relation to whey separation (Table 2). Control yoghurts and the formulation containing 2% inulin received the highest scores for whey separation. The lack of whey retention exhibited by the 2% inulin samples may be because of the solubility and relatively low molecular weight of inulin compared with either beta-glucan or guar gum. All the other samples were given significantly lower scores indicating significant decrease in tendency for syneresis. Similar trends were indicated by the whey separation scores as perceived by scooping; the highest scores were again assigned to control yoghurts, but also to yoghurts containing low levels of DF or SMP (Table 2). At high levels of addition, all DF used appeared to significantly reduce yoghurt susceptibility to syneresis in comparison with the control products (Fig. 7 and Table 2), the level of whey separation being comparable with products in which 6% SMP was added.

Table 2.   Mean values for appearance sensory attributes of dietary fibre-enriched yoghurt
 FF_coSM_co0.5% beta-glucan1.5% beta-glucan2% inulin6% inulin2% PHGG6% PHGG2% SMP6% SMPSignificance
  1. Within the same column, the values with the same letter are not significantly different.

  2. ***P < 0.001; **P < 0.01; *P < 0.05; NS, not significant; PHGG, partially hydrolysed guar gum; SMP, skimmed milk powder; FF_co, full-fat control; SM_co, skimmed milk control.

Appearance (as perceived by visual inspection)
 Whey separation4.4a4.7a1.6b1.7b5.0a1.4b1.9b1.3b2.4b1.5b***
 Aspect of the surface5.
Figure 7.

 Sensory attributes of dietary fibre-enriched yoghurts: (a, b) beta-glucan; (c, d) inulin; (e, f) PHGG.

The sensory scores presented in Table 3 indicate that the panellists found the samples tested to have similar firmness, except for the yoghurt containing 2% SMP, which received significantly lower scores. Panellists found no significant differences in the viscosity of the samples tested. This may indicate that the sensory analysis performed was not sensitive enough to highlight the differences that were previously observed instrumentally.

In terms of creaminess, the lowest scores (related to watery texture of the products) were given by the panellists to low-fat control yoghurt and also to low-fat yoghurts containing 2% PHGG or 2% SMP (Table 3 and Fig. 7). Incorporation of β-glucan into formulations or inulin or PHGG at high levels (6%), significantly improved the perceived creaminess of the product (P < 0.05); equally in these formulations the mouthfeel of the products was also improved in comparison with low-fat formulations, the resulting texture being perceived as smoother (P < 0.05; Table 3). Sensory attributes such as colour, glossiness, aspect of the surface and untypical taste/flavour were not affected by the incorporation of DF into the formulation. Overall acceptability of the products was good, the panellists showing no particular preference for a type of product (P > 0.05), although the samples containing inulin at 6% addition, and the sample with 2% SMP, received the highest rankings for acceptability.


The results from these experiments suggests that the DF selected for this study could be used successfully as carbohydrate-based fat replacers in low-fat yoghurt formulations to simulate high-fat products. Interactions between DF and milk protein were seen to promote changes in product structure; these could be directly related to changes in rheological properties. Thus, the increased G′ associated with increasing DF content could be related to the increased firmness of the product as determined via textural analysis. This observation would explain the improved yield of yoghurt, decreased syneresis and overall improved acceptability in terms of sensory attributes. Both guar gum and inulin are already used in some yoghurt formulations as stabilisers; however, this work extends the potential utilisation of beta-glucan into the dairy industry situation in terms of both human nutrition and product texture.