Analysis of Techno-Functional Properties of Fermented and Non-Fermented Buttermilk-Containing Ice Creams

Abstract

The aim of this study was to investigate the utilization of buttermilk, a by-product of butter production, in ice cream. Butterfly pea flower, which provides natural coloring and antioxidant properties, was added to buttermilk for investigating its improving effect on the techno-functional and sensory attributes of ice cream. Ice cream mixes were prepared with varying buttermilk concentrations (0%, 20%, 40%, 60%, 80%, 100%) as the first factor of the research. In addition, the effect of fermentation was also investigated as the second factor of the experiment. The ingredients included buttermilk, milk, cream, sucrose, dextrose, locust bean gum, butterfly pea flowers, and vanilla extract. The preparation involved the extraction of the butterfly flowers, fermentation in case of the fermented samples, homogenization, pasteurization, freezing, and hardening. Quality attributes such as dry matter content, pH, color, rheological properties of the ice cream mixes, overrun, melting properties, and ice cream hardness were analyzed to determine the maximal substitution level of milk by buttermilk without compromising ice cream quality. Our results explore the impact of buttermilk content and fermentation on the techno-functional properties of ice cream. As buttermilk concentration increased, dry matter content decreased, ranging from 34.4 g/100 g at 0% buttermilk to 31.9 g/100 g at 100% buttermilk. pH levels were lower in the fermented samples, decreasing from 6.5 in the non-fermented to 4.6 in the fermented samples. L* decreased with higher buttermilk content, while a* and b* values increased slightly. The butterfly pea flower provided a blue hue across all samples; the blue hue increased by 20% with a higher buttermilk content. Increasing the buttermilk concentration led to a 40% decrease in the yield stress and consistency coefficient, indicating a less viscous mix. The flow behavior index slightly increased, suggesting a more Newtonian-like flow at higher buttermilk levels. Overrun decreased with a higher buttermilk content, from 45% at 0% buttermilk to 30% at 100% buttermilk, indicating reduced air incorporation. The meltdown rate increased with a higher buttermilk content, meaning the ice cream melted more rapidly. The hardness of the ice cream decreased as buttermilk concentration increased, from 15 N at 0% buttermilk to 10 N at 100% buttermilk. The fermented sample groups were on average 44% harder than the non-fermented sample groups. The findings suggest that up to 100% of buttermilk can effectively replace milk in ice cream formulations without compromising quality, providing a sustainable and health-beneficial use for this dairy by-product.

1. Introduction

Globally, 3.2 million tons of butter are produced annually [1] and 10.98 thousand tons were produced in Hungary in 2023 [2] leading to an estimated equal volume of its by-product, buttermilk [1,3]. While there is a great tradition of using buttermilk around the world, unfortunately, this valuable by-product is rarely used in Hungary. The dairy industry thus produces a significant amount of wastewater, which has a high biological and chemical oxygen demand. Besides being a clear economic problem for agri-food industries (e.g., high processing costs, low energy efficiency), by-products increase environmental pollution [4,5]. Several options are available to address this problem. The dairy industry has been looking for creative ways to employ buttermilk to make dairy products that are healthier, more affordable, and sustainable. Ice cream is one such product that might be used to repurpose dairy by-products [6].

Buttermilk is the aqueous phase released during the churning of cream in butter manufacture. It contains all the water-soluble components of cream such as milk protein, lactose, and minerals [7]. Sodini et al. (2006) found that the foaming capacity of sweet or cultured buttermilk was lower than that of skimmed milk. However, it also contains materials derived from milk fat globule membrane (MFGM) [8,9], which is disrupted during the churning and mostly migrates to the buttermilk fraction. Therefore, buttermilk contains more phospholipids than milk [10]. It is similar in composition to skimmed milk and the higher MFGM content can enhance its emulsifying properties.

Its emulsifying properties make buttermilk adaptable for use in yogurt [11], kefir [12], and cheeses [13] like cheddar cheese [14], mozzarella cheese [15], and ice creams [16,17]. Because of these qualities, it works well as an ice cream base instead of milk [16].

Fermented dairy products are considered an essential part of daily meals all over the world because of their nutritional value, and they have been vastly recommended for their health benefits [18]. Fermented buttermilk is a dairy by-product, which may have unique functional properties. The fermentation process increases the bioavailability of certain nutrients in buttermilk such as calcium, protein, and B vitamins, allowing the body to absorb them more effectively [19]. Fermentation can increase the antioxidant content and activity of buttermilk, providing additional health benefits [20].

Based on the findings of Szkolnicka et al. (2020), the main textural differences between buttermilk ice cream and conventional ice cream were that buttermilk ice cream was creamier, smoother, and more uniform in texture compared to conventional ice cream, probably due to the emulsifying properties of buttermilk. The sweet buttermilk version, however, had an increased stickiness. Overall, the use of buttermilk seems to improve some textural properties of the ice cream [16].

Butterfly pea (Clitoria ternatea L.) is a native herb from subtropical regions [21] like Malaysia and Thailand [22] that is commonly used to flavor and color desserts and beverages [23]. The flower’s purple color is attributable to anthocyanin [24], which has proven antioxidant properties [25]. The use of butterfly pea flower in ice cream offers several advantages. Butterfly pea flower is a natural colorant that provides natural blue color, which is appealing and can be used as a substitute for artificial food colorings. This makes butterfly pea flower-containing ice cream a healthier option compared to ice creams with artificial additives [26]. Butterfly pea flower extract can also provide functional properties, such as improved texture and increased antioxidant activity in ice creams. This can enhance the overall quality and appeal of the ice cream [22].

The hypothesis of this article is that the fermentation and addition of butterfly pea flower-containing buttermilk can be applied in ice cream production, but some of its properties (color, pH, texture, melting properties, and overrun) might be changed. Therefore, the aim of this study was to compare the quality characteristics of butterfly pea flower-containing ice creams made with different concentrations of fermented and non-fermented buttermilk and to determine the amount of buttermilk that can be used to substitute milk without quality loss. For this purpose, a factorial experimental design was set up, in which the two factors were (1.) the concentration of milk substituting buttermilk (0%, 20%, 40%, 60%, 80%, 100%) and (2.) the fermentation treatment (fermented, non-fermented). Fermentation is not always part of the technology used to produce ice cream mixes, but as previously described, it has several positive effects. For this reason, it was important to investigate the effect of fermentation on samples made with buttermilk. The quality of ice cream was compared through its techno-functional properties: dry matter content, pH, color, rheological properties of ice cream mix, overrun, meltdown, and ice cream hardness.

2. Materials and Methods

2.1. Ice Cream Production Procedure

Ice cream mixes were prepared using the following ingredients: milk (3.5% fat content, Magyar Tej UHT, Alföldi Tej Ltd., Székesfehérvár, Hungary), cream (30% fat content, Parmalat chef UHT, Lactalis Ltd., Zagreb, Croatia), sucrose (Diamant, 1MCM Ltd., Kaposvár, Hungary), dextrose (m-GEL Hungary Ltd., Budapest, Hungary), locust bean gum powder (m-GEL Hungary Ltd., Budapest, Hungary), milk protein isolate 85% (Sole-Mizo Ltd., Csorna, Hungary), vanilla extract (Dr. Oetker Ltd., Jánossomorja, Hungary), dried butterfly pea flowers (Clitoria ternatea) (Manutea Ltd., Opava, Czech Republic), and FD-DVS CHN-22 eXact^® Mesophilic aromatic culture (Chr. Hansen A/S Ltd., Hoersholm, Denmark). The buttermilk was kindly provided by Alföldi Tej Ltd. (Székesfehérvár, Hungary).

Firstly, dried butterfly pea flowers were extracted in buttermilk. Buttermilk was heated to 80 °C and 1.5 g of dried butterfly pea flowers / 100 g buttermilk were added to buttermilk. Flowers were filtered through stainless steel strainer after 10 min, and liquid was stirred for 2 min.

Six different ice cream mixes were prepared, in which the milk content was replaced by a certain amount of buttermilk (0%, 20%, 40%, 60%, 80%, 100%). Half of each sample was non-fermented and the other half was fermented by mesophilic aromatic culture. The ingredients were combined in calculated proportions to obtain 8.5–9.5 g/100 g fat and 2.2–2.3 g/100 g protein in the mixture. The calculated dry matter content was 31.9–34.4 g/100 g. The recipes of the prepared ice creams are presented in

Table 1. Quantity of ingredients in samples of different formulations (g/100 g). Fermented and non-fermented samples containing same amount of buttermilk were prepared based on same recipes.

Ingredients	Buttermilk Content *
	0%	20%	40%	60%	80%	100%
Buttermilk	0	11.14	22.28	33.42	44.56	55.7
Milk	55.7	44.56	33.42	22.28	11.14	0
Cream	25.3	25.2	26.0	26.6	26.5	27.5
Sucrose	12.8	12.8	12.2	11.7	11.7	10.9
Dextrose	6.4	6.4	6.1	5.8	5.8	5.5
Milk protein isolate	0.1	0.15	0.25	0.4	0.5	0.6
Locust bean powder	0.1	0.1	0.1	0.1	0.1	0.1
Vanilla extract	0.1	0.1	0.1	0.1	0.1	0.1

* Buttermilk content means the percentage of buttermilk substituting milk in the recipe.

.

The ice creams were prepared using the following method. The liquid ingredients (milk, buttermilk, and cream) were mixed and heated to 90 °C and the mixture was pasteurized for 2 min. The fermented samples were cooled to room temperature and inoculated with 0.16 g/L of mesophilic lactic acid bacteria starter culture, incubated for 16 h at room temperature (~25 °C). The non-fermented samples were cooled to 55–60 °C. After that, the powdered ingredients were added to the liquids and the mixture was homogenized by Robot Coupe Mini MP 160 V.V. mixer (Robot Coupe, Montceau-en-Bourgogne, France) at speed 3 (speed 1–9; 2000–12,500 rpm) for 2 min. The homogenized ice cream mixes were filled into PA-PE (polyamide-polyethylene) bags (90 μm: 20 μm PA + 70 μm PE; AMCO Ltd., Budapest, Hungary) and sealed using an impulse sealer [27]. The fermented ice cream mixes did not undergo the pasteurization process due to their lower pH and the presence of live lactic acid bacteria. The non-fermented mixtures were pasteurized at 75 °C for 30 min. All the samples were rapidly cooled to 4 °C and aged at this temperature for 24 h. After the aging process, the ice cream was produced by using a horizontal freezing machine (Telme CRM GEL 5; Telme, Codogno, Italy) for 13–15 min. The frozen samples were added to plastic containers and were hardened at −30 °C for 1 h. The samples were stored at −24 °C for 1 day. The processing technology of the samples is presented in Figure 1.

2.2. pH Measurement

The pH was measured before the aging step of the ice cream mix in its liquid state with a Testo 206 portable pH meter (Testo AG., Titisee-Neustadt, Germany). All the samples were analyzed in triplicate. The assumption was that fermentation would lower the pH.

2.3. Color Measurement

The Konica Minolta CHROMA METER CHR-400 tristimulus color measuring system (Konica Minolta Sensing Europe B.V., Nieuwegein, The Netherlands) was used to measure the surface color of the aging liquid samples in CIELAB values (lightness, L*; redness, a* and yellowness, b*) [28] and hue angle (h_ab), which was calculated from a* and b* color factors, with standard illuminant D65 [29]. The measurements were carried out in 12 replicates. Color difference (ΔE*) was calculated according to the following formula [30] (Equation (1)):

$$∆{\mathrm{E}}_{\mathrm{a}\mathrm{b}}^{\mathrm{*}}=\sqrt[2]{{({\mathrm{L}}_1-{\mathrm{L}}_2)}^2+{({\mathrm{a}}_1-{\mathrm{a}}_2)}^2+{(\mathrm{b}-{\mathrm{b}}_2)}^2}.$$

(1)

The hypothesis was that buttermilk would make the color of the samples bluer thanks to the butterfly pea flower enrichment. In addition, the lightening effect of fermentation was expected due to pH reduction, as well as the yellowing effect of the natural carotenoid content of milk in the samples with low buttermilk content.

2.4. Rheological Measurement

The examination of the rheological behavior of the aging ice cream mixes was performed with an MCR 92 rheometer (Anton-Paar, Ostfildern, Germany) in rotational mode equipped with a concentric cylinder (CC27). Anton Paar RheoCompass software (version 1.21.852) was used to control the equipment. The temperature of the rheological measurements was kept constant at 20 °C. Shear stress was measured by increasing and decreasing the shear rate between 10 and 1000 s⁻¹ for 31 measurement points with a period of 3 s, similar to the method of Hidas and her co-researchers [31]. The Herschel–Bulkley model (Equation (2)) was applied to analyze the flow curves (shear rate–shear stress diagrams) since multiple authors [31,32,33,34,35,36,37] determined that this model was sufficient to explain the rheological behavior of ice cream mixes.

$$\mathsf{\tau }={\mathsf{\tau }}_0+\mathrm{C}\times {\dot{\mathsf{\gamma }}}^{\mathrm{n}}$$

(2)

where τ—shear stress (Pa); τ₀—yield stress (Pa); $\dot{\mathsf{\gamma }}$—shear rate (s⁻¹); C—consistency coefficient (Pa sⁿ); and n—flow behavior index (dimensionless).

From the measurement results, a theoretical shear rate was calculated for each measurement point based on the model equation. The sum of the squares of the differences between the calculated and measured shear rates was iterated using the Excel Solver to fit the model to the measurement points and to determine the values of the constants in the equation. After applying the method of least squares, the model was validated by considering the coefficient of determination (R²), which indicates the goodness of fit. All R² values of the fitted model were higher than 0.99 [38].

2.5. Overrun Measurement and Meltdown Test of Ice Creams

Overrun shows the foaming capability of ice creams. This characteristic of ice cream was measured by the method of [39]. The weight of all the 50 mL samples was measured before and after the freezing process. The weight difference was calculated in percentage. All the samples were analyzed in triplicate.

The meltdown tests were carried out at room temperature (21 °C) based on the method of a previous study about ice cream investigation [27]. A portion of 50 g sample was cut into a 30 mm high and 50 mm diameter cylinder and was allowed to melt on a stainless steel grid. A beaker was placed underneath the grid on a scale to collect and weigh the drip losses. The weight of the melted sample was noted every 5 min until the entire sample had melted. The meltdown is defined as the mass of the drip loss (m_D) divided by the total mass of the ice cream sample (m₀) and plotted against time. The maximum meltdown rate (MDR) was determined at the highest gradient in the ascending meltdown curve [40]. All the samples were analyzed in triplicates.

According to our hypothesis, the buttermilk concentration would affect these properties by reducing the fat and protein content.

2.6. Texture Measurement of Frozen Ice Cream

A Stable Micro System (SMS) TA. XT Plus (Stable Micro System, Godalming, UK) texture analyzer was used to analyze the texture of the pre-shaped (5 mm wide and 50 mm long half cylinders) samples. The force (N) required to shear the sample was examined using the Warner–Bratzler (W-B) test. The samples were cut using a measurement speed of 2 mm/s with a blade measurement head (HDP/BS) [38]. Before the measurement, the ice creams were kept at −24 °C for a minimum of one day following manufacturing. Between two measurements, the measuring head and the tray were kept in ice to prevent melting.

The hypothesis was that the texture is influenced by pH, protein content, fat content, and dry matter content, which explains the level of both factors.

2.7. Statistical Analysis

IBM SPSS v27 (IBM, Armonk, New York, NY, USA) and Microsoft Excel 365 version 2010 (build: 13328.20356) were used to analyze the measurement results. Excel was used for mathematical operations, correlation analysis, data representation, and rheological model fitting. Multivariate analysis of variance (MANOVA) was used to determine the impact of the fermentation and various buttermilk concentrations on the rheological behavior (τ₀, C, n), texture (hardness and work of shear), color (L*, a*, b*, h_ab), pH, overrun, and MDR of each ice cream mix. Three MANOVAs were carried out because of the degree of expected correlations that met the MANOVA conditions. One was carried out for general techno-functional properties, one for rheological properties, and one for overrun and MDR properties [41]. The latter was necessary because previous research has shown a relationship between these two properties. The value of the unexplained variance rate (Wilks’s lambda) was evaluated. The Tukey HSD post hoc test (α = 0.05) was used to differentiate the homogeneous groups. The homogeneity of variances was examined by Levene’s test [L*: F(11,132) = 4.560, p < 0.001; a*: F(11,132) = 3.429, p < 0.001; b*: F(11,132) = 3.634, p < 0.001; h_ab: F(11,132) = 6.473, p < 0.001; pH: F(11,132) = 4.187, p < 0.001; hardness: F(11,132) = 2.947, p = 0.002; work of shear: F(11,132) = 2.691, p = 0.004; τ₀: F(11,24) = 5.902, p < 0.001; C: F(11,24) = 4.088, p = 0.002; n: F(11,24) = 3.292, p = 0.007; overrun: F(11, 60) = 2.688, p = 0.007; MDR: F(11,60) = 2.779, p = 0.008]. The Kolmogorov–Smirnov test was used to verify that the residuals were normal. Since the conditions of the Kolmogorov–Smirnov test were not fulfilled in the cases of L*, a*, b*, h_ab, pH, and texture properties, the D’Agostino test was used to determine whether the residuals were normal. The sum of squares of the ratio of skewness and standard error of skewness, and the peak and standard error of peak were compared to a Khi-squared distribution with 2 degrees of freedom (p ≥ 0.001 in the case of all the unstandardized residuals of dependent variants).

3. Results and Discussion

After checking the equality of error variances by Levene’s test and normality of residuals by the D’Agostino test, MANOVA was carried out. The value of the unexplained variance rate suggested that the percentage of buttermilk [Wilks’ lambda < 0.001, p < 0.001] and the treatment (fermentation) [Wilks’ lambda = 0.018, p < 0.001] also had a very strong and significant effect on the techno-functional properties (color, pH, texture) of the samples. The interaction of the two factors [Wilks’ lambda < 0.001, p < 0.001] also had a very strong and significant effect. It can be said that the values of the independent variables were best explained by the buttermilk ratio, but since all the Wilks’ lambda values were below 0.1, the effects of each factor were very strong. The techno-functional properties discussed in the following sub-sections are also included in the analysis of the effects of the factors, but this overall result is only presented here.

3.1. Color

Based on the results of the Tukey post hoc test, significant (p = 0.05) difference was found between the lightness of the sample groups with different buttermilk ratios. All sample groups with different buttermilk ratios were significantly (p = 0.05) different except for the samples with 20% and 40% buttermilk. The difference between the lightness of sample groups was on average 4–5% per 20% of buttermilk level including the difference between sample groups with 20% and 40% buttermilk content. The fact that the difference was not significant (p = 0.05) in the latter case could be explained by the large standard deviation of the samples with 20% buttermilk content, as can be seen in Figure 2.

This difference was explained by the fact that the pigments of the flower in the recipe were added to the ice cream with buttermilk, resulting in a darker color. The difference between the sample groups with and without fermentation was also significant (p = 0.05) based on lightness, which meant an average 11% lightness difference. In all cases, the fermented samples were lighter than the non-fermented samples.

Aksornsri et al. (2023) prepared fermented kombucha-like beverage using butterfly pea flower extract. The color changes in the samples during fermentation were monitored using the CIELab system and all the samples were observed to become lighter during the 4-day-long fermentation period [42].

Setiawati and Kusnadi (2021) have concluded that fermentation had a decreasing effect on the brightness of butterfly pea water kefir due to pH changes. Lower pH has a stabilizing effect on anthocyanins. Long fermentation times degrade tannins because of the presence of acids, which results in color fading [43].

This is partly due to the effect of fermentation on pH and the pH sensitivity of the flower pigment used in the recipe, and partly due to the pigments that decompose during fermentation. Butterfly pea flower has a high anthocyanin content, which can be incorporated into polymer-based films to produce intelligent packaging for real-time food freshness indicators. Some researchers have made a visual intelligent pH indicator film of eggshell membrane, gelatine, pectin, and anthocyanin pigment from butterfly pea flower. The anthocyanin extracted from butterfly pea flower generally undergoes structural changes and displays a variety of colors under different pH conditions [44,45,46,47].

In the case of redness–greenness, significant (p = 0.05) differences were found between the sample groups with different buttermilk ratios according to the Tukey post hoc test, but these statistically significant (p = 0.05) differences were nominally not significant. As can be seen in Figure 3, this trend could be observed only in the case of the fermented samples. Interestingly, there was hardly any difference between the non-fermented sample groups. This means that the green color measured was probably due to the green pigments in the milk raw material from the feed. It was steadily diluted with the added buttermilk, but it was only the change in pH caused by fermentation that made this dilution measurable. This could be explained by the pH dependence of the color of natural pigments. It is important to mention that the change in both color factors was nominally very small and not visible to the naked eye. Only the change in the brightness factor was visible to the naked eye. On average, the redness–greenness factor of the non-fermented samples was 62% lower than that of the fermented samples, which means that the non-fermented samples were on average 162% relatively greener.

According to the results of the Tukey test based on yellowness–blueness, each sample group with a different buttermilk ratio was significantly (p = 0.05) different. The trend was clear: the higher the buttermilk content, the lower the yellowness–blueness factor, and the bluer the product. There was only a very small difference between the average results of the samples with zero buttermilk content. This could be explained by the fact that the blue pigment from the flower was added to the ice cream with the buttermilk. Due to the natural coloring agents (carotenoids) in the milk derived from the feed, the control sample was more yellowish, but became bluer with the addition of buttermilk containing the blue pigment. An interesting observation was that the bluing was much lower for the fermented samples than for the non-fermented samples. This could be explained by the fact that the color of bioflavonoids depends on the standard electrode potential of their environment. Thus, fermentation caused the pigment of the flower to lose its blue coloring effect. The average difference in the yellowness–blueness factor between the fermented and unfermented samples was 167%. The yellowness–blueness results are shown in Figure 4.

The hue angle is used to determine the color of objects. The color of samples varied from green (120) and bluish green (164) to cyan (180–196) and blue (240). On average, there was a 50% difference between the fermented and unfermented samples. Overall, the hue angle of the non-fermented samples was shifted more towards a blue color. The large standard deviation of the 20% buttermilk sample group was cumulative on this factor. There was a noticeable saturation mechanism towards blue in the color trend of the non-fermented samples. The same significant (p = 0.05) difference found for the yellowness–blueness factor was also observed based on the hue angle for each sample group with different buttermilk contents. Saturation mechanism is the most common natural mechanism. As can be seen in Figure 5, from about 60% onwards, the non-fermented samples stopped becoming bluer, and from this point on they took on the color of the flower pigment. For the fermented samples, the color of the samples started to shift towards a bluer color only from the 60% buttermilk content at a lower pH. But above 60% the ice cream started to turn blue in an exponential trend due to the high amount of pigment. The color of the ice cream samples could be changed due to the presence of anthocyanins in the blue pea powder which are sensitive to pH [48].

Color difference, which is presented

Table 2. Color difference (ΔE*) [–] of ice cream sample groups made from different percentage of buttermilk (0, 20, 40, 60, 80, 100) with different treatments (NF = non-fermented, F = fermented).

Buttermilk Content (%)/Fermentation		0		20		40		60		80		100
		NF	F	NF	F	NF	F	NF	F	NF	F	NF	F
0	NF	0.00	4.50	15.49	3.77	17.30	5.79	20.95	8.93	23.40	11.94	28.21	14.57
	F	-	0.00	19.39	4.88	20.93	8.76	24.52	12.05	27.15	15.35	32.09	18.17
20	NF	-	-	0.00	15.04	2.67	11.26	5.97	8.69	7.99	6.22	12.77	4.93
	F	-	-	-	0.00	16.34	4.09	19.87	7.30	22.58	10.66	27.57	13.53
40	NF	-	-	-	-	0.00	12.54	3.67	9.64	6.26	6.88	11.33	5.12
	F	-	-	-	-	-	0.00	16.06	3.30	18.72	6.61	23.65	9.45
60	NF	-	-	-	-	-	-	0.00	13.02	3.03	10.06	8.09	7.81
	F	-	-	-	-	-	-	-	0.00	15.72	3.38	20.63	6.28
80	NF	-	-	-	-	-	-	-	-	0.00	12.65	5.13	10.18
	F	-	-	-	-	-	-	-	-	-	0.00	17.44	2.90
100	NF	-	-	-	-	-	-	-	-	-	-	0.00	14.77
	F	-	-	-	-	-	-	-	-	-	-	-	0.00

, shows the degree of difference between the color of each sample group and the other sample groups in a pairwise comparison. The higher the value, the more the difference is visible.

Similar results to previous studies were obtained for the color of the control ice cream, but thanks to the enrichment, the ice cream made with buttermilk was bluer than the results measured by others [17,49].

3.2. pH

No significant (p = 0.05) difference was observed between the samples with different buttermilk contents. However, there was a significant (p = 0.05) difference between the samples treated in different ways, which is clearly visible in Figure 6. The pH of the non-fermented sample groups was on average 43% higher than the pH of the fermented sample groups. Szkolnicka and her co-researchers (2020) observed similar results [16]. In their study, the control and non-fermented buttermilk ice cream pH values were similar with on average 6.33–6.59 values, and the fermented buttermilk ice cream pH values were lower with on average 5.16–5.66 values. In the case of this research, the fermented pH values were much lower with on average 4.52–4.64 values. Gebreselassie et al. (2016) drew similar conclusions that the pH was reduced to 4.43 ± 0.18 in fermented buttermilk and was not significantly influenced by either buttermilk type or incubation time [20]. Proteins become externally neutral as they approach their isoelectric point, thus losing their solvation shell. This has an effect on the properties of the foods they are made from, such as texture and lightness, as less light passes through material with a higher bound water content.

3.3. Texture

The hardness and work of the shear results are shown in

Table 3. Hardness [N] and work of shear [Nmm] results of ice cream sample groups made from different percentage of buttermilk (0, 20, 40, 60, 80, 100) with different treatments (non-fermented, fermented). Values are means ± standard deviation.

Percent of Buttermilk (%)	Treatment	Hardness [N]	Work of Shear [Nmm]
0	Non-fermented	8.45 ± 3.975	18.93 ± 9.117
	Fermented	21.74 ± 5.546	40.40 ± 20.819
20	Non-fermented	10.58 ± 4.013	13.58 ± 5.594
	Fermented	27.24 ± 9.548	49.27 ± 21.027
40	Non-fermented	10.43 ± 4.128	20.97 ± 8.741
	Fermented	18.24 ± 4.323	31.61 ± 11.068
60	Non-fermented	14.27 ± 5.907	35.31 ± 21.260
	Fermented	7.97 ± 2.232	18.89 ± 5.382
80	Non-fermented	5.77 ± 1.919	11.04 ± 5.434
	Fermented	21.21 ± 6.486	38.87 ± 18.888
100	Non-fermented	13.73 ± 3.997	22.48 ± 8.843
	Fermented	16.47 ± 4.188	26.49 ± 11.346

. No significant differences were found between the different sample groups with different buttermilk contents. Ice cream is a matrix in which crystals, air bubbles, and liquid or semi-liquid substances are present simultaneously. Because of this and the continuous melting during measurement, the standard deviation of the results was very large. Homoscedasticity for the texture results was the least significant among the techno-functional properties, but it was still acceptable. It can be observed that there was a significant (p = 0.05) difference between the samples treated in different ways. The trend is not entirely clear, but the fermented sample groups were on average 44% harder than the non-fermented sample groups, and the work required to shred through the fermented sample groups was on average 15% greater than that required for the non-fermented samples. Therefore, it can be concluded that fermentation hardened the samples. It was expected that the proteins closer to their isoelectric point would lose their water-binding capacity, so the fermented samples with lower pH would melt more easily and remain less hard. However, the opposite was experienced. This observation is in line with previous studies [50,51], as the viscosity of ice cream increases during fermentation to determine gelation.

3.4. Rheology

The results of the rheological properties are presented in Figure 7, Figure 8, Figure 9 and Figure 10. It should be mentioned that although significant (p = 0.05) differences in rheological attributes were observed, the rheological behavior of ice cream mixes did not differ remarkably. They were all described by the same rheological model, the Herschel–Bulkley model, with a coefficient of determination above 99%. The value of the unexplained variance rate suggested that the percentage of buttermilk [Wilks’ lambda = 0.001, p < 0.001] and the treatment (fermentation) [Wilks’ lambda = 0.002, p < 0.001] also had a very strong and significant effect on the rheological behavior of the samples. The interaction of the two factors [Wilks’ lambda = 0.001, p < 0.001] also had a very strong and significant effect. This means that the levels of the independent variables explain the dependent variables, namely, the rheological constants of each sample group, very well. This effect is best seen in yield stress. The higher the ratio of buttermilk in the product, the lower the yield stress becomes, indicating that the minimum stress required to maintain or stop flow is decreasing [52]. The shear stress represents the axial cross-section in the Herschel–Bulkley model, showing the apparent viscosity at the first break of the structure of the material. It also gives a good picture of the cohesivity of the material. It was interesting to observe that the yield stress decreased by increasing buttermilk content only in the case of the non-fermented ice creams. The yield stress value of the fermented ice creams was on average 68% lower than that of the non-fermented ice creams. This can be explained by the fact that the fermented buttermilk ice cream mix required slightly more intensive mixing. According to the results of the Tukey post hoc test, based on the yield stress results, four homogenous groups were significantly (p = 0.05) separated. The sample groups without buttermilk had the highest yield stress. These were followed by the sample groups containing 20% buttermilk, and then by the sample groups containing 40% buttermilk. The sample groups with 60%, 80% and 100% buttermilk were no longer significantly (p = 0.05) separable from each other, but they were separable from the other groups and had the lowest yield stress scores.

In contrast, for flow index, a decrease was observed with increasing buttermilk concentration. The flow index is the exponential component of the Herschel–Bulkley model, which is responsible for the curve’s steepness. Thus, increasing the shear stress above 1 Pa s leads to a smaller increase in viscosity for samples with a lower flow index, and therefore higher buttermilk content. This finding can be an important result for the selection of mixing and pump efficiency for the industrial production of buttermilk ice creams, especially in the case of mixing or pipe flow, in order to achieve favorable energy consumption. In the case of the flow index, it was also observed that the fermented ice-creams showed lower results than the non-fermented ice-creams. This relative difference was on average 9%. According to the Tukey test, four groups were separated significantly (p = 0.05). In order of increasing flow index, these were as follows: (1) a sample group with 100% buttermilk content; (2) sample groups with 40%, 60%, and 80% buttermilk content; (3) a sample group with 20% buttermilk content; and (4) a sample group without buttermilk.

For the consistency index, there was no clear trend. It was clear that the consistency index of the fermented ice creams was on average 20% lower than the consistency index of the non-fermented ice creams. Similar behavior was reported for ice cream mixes in a previous study dealing with the production of ice cream made with the addition of EPS (exopolysaccharide)-producing strains. It was found that the use of EPS-producing strains resulted in a remarkable increase in K values in comparison to the control sample [53].

It was interesting to observe that the consistency index increased up to an average of 0.041 with increasing buttermilk content up to 60%, then decreased at 80% buttermilk content to an average of 0.035 observed at 40% buttermilk content, and then reached an average of 0.041 again at 100% buttermilk content. According to the Tukey test, four groups were separated significantly (p = 0.05). In order of increasing consistency index, these were the following: (1) a sample group with 0% buttermilk content; (2) a sample group with 20%; (3) sample groups with 40% and 80% buttermilk content; (4) sample groups with 60% and 100% buttermilk content.

Our results contradict a previous paper in which increasing the buttermilk ratio increased the yield stress and the consistency index and slightly decreased the flow behavior index [49].

3.5. Overrun

In the case of the third MANOVA, which evaluated the results of the MDR and overrun, the unexplained variance rates were as follows: buttermilk ratio [Wilks’ lambda = 0.001, p < 0.001]; treatment [Wilks’ lambda = 0.004, p < 0.001]; and interaction of buttermilk ratio and treatment [Wilks’ lambda = 0.008, p < 0.001], which suggested a very strong determination and significant effect of both factors and the interaction of the buttermilk ratio and treatment as well. The equality of error variances was checked by Levene’s test. According to Tukey’s test for the buttermilk ratio, the overrun value of all the sample groups differed significantly (p = 0.05) from each other except for the sample groups with 20% and 40% buttermilk concentration. The overrun increased with decreasing buttermilk ratio. The average relative difference between the fermented and non-fermented sample groups was 23%.

The overrun values of the ice cream samples in this study varied between 9.9% and 30.3%. These values are lower than the average overrun values of milk-based ice creams, [54] as the ice cream machine used in this study was a piece of semi-factory equipment with lower overrun values. However, trends are clearly visible (Figure 11). According to the overrun values, as buttermilk content increases, the overrun decreases and the air intake also decreases with buttermilk content. This phenomenon was also reported by Szkolnicka and her co-researchers (2020). This is most likely due to the higher amount of MGFM in ice creams with more buttermilk [16]. Although MGFM helps to create a uniform structure, it also lowers air intake.

3.6. Melting Properties

The ingredients and composition of ice cream affect its melting properties [39]. Based on Figure 12, the first drop occurred in the non-fermented samples after 15 min and in the fermented samples after 9 min, so all the samples had a similar structure. The majority of the samples melted—approximately 90–99%—within 80 min. Two groups were identified from the samples. The ice cream samples made with non-fermented buttermilk formed the first group. Consistent melting was observed in the non-fermented samples. The ice cream samples produced with fermented buttermilk constituted the other group. There seemed to be differences in the melting curves. The fermented curves had a greater slope. Figure 13 presents the maximum meltdown rate (MDR) values. Melting occurred much more slowly in the non-fermented samples due to a lower MDR than in the fermented samples. There were variations in the fermented samples. Faster melting was achieved with a higher buttermilk content. Thus, the ice cream melted more readily when the amount of buttermilk increased. This phenomenon is associated with higher amounts of MGFM in ice creams containing more buttermilk [16]. Wu and his co-researchers (2019) reported that changing emulsifier types and concentrations contributes to different levels of fat destabilization, which in turn influences ice cream meltdown rate [55].

The equality of error variances was checked by Levene’s test. According to the Tukey test, the MDR of each sample group differed significantly (p = 0.05) by buttermilk ratio, except for the 100% and 80% buttermilk sample groups, which did not differ significantly (p = 0.05) from each other, but did differ significantly (p = 0.05) from the other sample groups. However, only a slight trend was observed for the MDR in relation to buttermilk percentage, where the MDR increased slightly with increasing buttermilk concentration. Nevertheless, there was a clear difference between the fermented and non-fermented sample groups, with a relative difference in the MDR values of on average 43%. Favaro-Trindade and his colleagues (2007) obtained the opposite finding that the fermented samples had lower melting rates [56].

4. Conclusions

In this study, the techno-functional properties of ice creams formulated with varying ratios of buttermilk substituting milk and subjected to fermentation were examined. The findings offer significant insights into the impacts of buttermilk content and fermentation on ice cream properties, potentially guiding the development of new and improved dairy-based frozen desserts.

The color results showed that both buttermilk ratio and fermentation significantly affected the color of the ice cream samples. There was a significant difference in lightness between the sample groups with varying buttermilk ratios. Fermented samples were generally lighter than non-fermented ones. In addition, fermented samples had a lower pH, which affected their texture and lightness as well, due to changes in protein solvation and pigment behavior. In the case of yellowness–blueness, a higher buttermilk content led to a decrease in yellowness, shifting towards blue. Fermentation reduced the blue hue. Non-fermented samples showed a saturation trend towards blue, while fermented samples exhibited a bluer hue only above 60% buttermilk content. Fermentation significantly impacted texture, making the samples harder and more difficult to shear.

The rheological behavior of ice cream mixes was best described by the Herschel–Bulkley model. Significant effects of buttermilk ratio and fermentation were observed on yield stress, flow index, and consistency index. Increased buttermilk content reduced yield stress, particularly in non-fermented samples. Flow index decreased with higher buttermilk content. Fermented samples had a lower flow index than non-fermented ones. Overrun decreased with increasing buttermilk content. Fermentation significantly accelerated the melting process, with higher buttermilk content leading to faster meltdown rates.

This study underscores the complex effect of buttermilk content and fermentation in determining the techno-functional properties of ice cream. These insights can guide the development of new ice cream formulations, leveraging buttermilk and fermentation to achieve desired sensory and functional attributes. Utilizing buttermilk in a value-added way enhances sustainability in the food industry.