Rheological and structural characterization of carrageenan emulsion gels

Contents lists available at ScienceDirect Algal Research journal homepage: www.elsevier.com/locate/algal Rheological and structural characterization of carrageenan emulsion gels Cynthia Fontes-Candia a , Anna Ström b , Patricia Lopez-Sanchez c , Amparo López-Rubio a , Marta Martínez-Sanz a,⁎ a Food Safety and Preservation Department, IATA-CSIC, Avda Agustín Escardino 7, 46980 Paterna, Valencia, Spain b Applied Chemistry, Chemistry and Chemical Engineering, Chalmers University of Technology, Gothenburg, Sweden c Agrifood and Bioscience, RISE Research Institutes of Sweden, Frans Perssons väg 6, SE-412 76 Gothenburg, Sweden ARTICLE INFO Keywords: Polysaccharide Seaweed Gelation Scattering Fat replacement ABSTRACT Carrageenan emulsion gels containing sunflower oil were prepared using three different commercial carrageenan grades (κ-C, ι-C and λ-C). The effect of the carrageenan and salt content, as well as the oil:water ratio, on the emulsion gel strength was evaluated through a response surface methodology. Moreover, the rheological properties and the microand nanostructure from the stronger emulsion gel formulations were investigated and compared to their analogous hydrogel formulations. Interestingly, emulsion gels formed stronger and more thermally stable networks than the hydrogels, being this effect more evident in ι-C and λ-C. The results indicate that this was mainly due to a polysaccharide concentration effect, as no evidence of interactions between the carrageenan and the oil phase was found. Consequently, the rheological behaviour of the emulsion gels was mostly determined by the type of carrageenan. The association of carrageenan molecular chains was favoured in κ-C and λ-C (due to the presence of κ-carrageenan in the latter) and promoted by the addition of KCl. In contrast, a lower degree of chain association, mostly driven by ionic cross-linking, took place in ι-C. These results evidence the relevance of the gelation mechanism on the properties of emulsion gels and provide the basis for the design of these systems for targeted applications within the food industry. 1. Introduction Emulsion gels are soft-solid materials composed of a three-dimensional network structure formed by a continuous phase consisting of a gel matrix and a dispersed phase consisting of an emulsion [1]. These materials present a great potential within the food industry as functional ingredients to modify food texture [2,3], for solid fat replacement [4,5] or even as templates for the controlled release of bioactive compounds [6]. Fats play a very important role in food processing; however, solid fats contain high amounts of saturated fatty acids which have been associated with several health conditions such as obesity, cancer and cardiovascular diseases [7]. Consequently, the food industry is actively looking for alternative ingredients to replace, at least partially, saturated and trans fats in products such as dairy and processed meats, while preserving their original organoleptic attributes as much as possible to ensure consumers' acceptance. A plausible approach consists in the replacement of solid fats with emulsion gels where an oil is incorporated into the dispersed phase. However, mimicking the rheological and textural properties of solid fats using emulsion gels is not trivial and many factors such as the physicochemical properties of the gelling matrix, the interactions between matrix and oil phase, oil droplet size, volume fraction and distribution, etc., have an impact on the textural and rheological properties of the product [8–10]. Depending on the interactions between the gel and the emulsified phase, the oil droplets can be classified as active (bound to the gel matrix) or inactive (unbound) fillers [11]. Active fillers can increase or decrease the stiffness of the gel depending on the ratio between the modulus of the oil droplets and that of the gel matrix, whereas inactive fillers are thought to cause a reduction on the gel stiffness [12]. Therefore, the properties of emulsion gels can be optimized by means of an adequate selection of the gelling matrices and fillers and a rational structural design of the produced materials. Ethyl cellulose has been widely used as solid fat replacement in meat and bakery products [13,14], since it can be directly dispersed in edible oils. However, the high temperatures required for its processing promote the oxidation of the oil phase. The use of synthetic polymers such a polyvinylalcohol and polysiloxanes as oil structuring agents has also been reported, but due to their synthetic nature, these compounds are not accepted as edible matrices [15]. As an alternative, natural compounds such as proteins [7,16–18] and polysaccharides [19–23] are https://doi.org/10.1016/j.algal.2020.101873 Received 20 December 2019; Received in revised form 17 February 2020; Accepted 9 March 2020 ⁎ Corresponding author. E-mail address: [email protected] (M. Martínez-Sanz). Algal Research 47 (2020) 101873 2211-9264/ © 2020 Elsevier B.V. All rights reserved. T being studied as gelling agents for food-grade emulsions gels. Sulphated polysaccharides are particularly interesting due to their excellent gelling capacities. Carrageenans are sulphated polysaccharides obtained from red seaweeds and are widely used within the food industry as gelling, thickening and stabilising agents [24]. Depending on the amount and distribution of sulphate groups (SO 3 − ) carrageenans can be classified into different groups. The k-, ιand λ-carrageenan (containing one, two and three sulphate groups per disaccharide unit, respectively) are the most commonly used in the food industry, since they are considered as generally recognized as safe (GRAS) [25]. Although carrageenans are not directly dispersible in oils, it is possible to obtain stable emulsion gels by combining a carrageenan solution with a proportion of oil by a simple emulsification process [19,26,27]. However, a detailed structural characterization of carrageenan emulsion gels is not available to date and the underlying gelation mechanism is not completely understood. A recent work [28] demonstrated that the gelation mechanism and type of gel networks originated by different commercial carrageenan grades are significantly different and can be further altered by the inclusion of cations. Specially, the addition of K + was seen to induce the formation of stronger κ-C and λ-C hydrogels. On the other hand, although Ca 2+ promoted more efficiently the formation of ι-C hydrogels than K + (requiring lower salt concentrations), the hydrogel strength was not significantly affected by the type of salt, suggesting that similar supramolecular structures may have been formed. Accordingly, the utilization of these different carrageenan grades to produce emulsion gels is expected to provide materials with distinct structure and rheological behaviour. The presence of oil may also have adifferent impact depending on the type of carrageenan used. A previous study found that ι-carrageenan was more effective than κand λcarrageenan at creating highly charged interfacial membranes to stabilize soybean oil emulsions, which was attributed to its most densely charged helical structure [29]. In this context, the objective of the present work was to investigate the nanostructure, rheological and textural properties of emulsion gels prepared using three different commercial carrageenan grades. The influence of the concentration of the commercial carrageenan and salt, as well as the oil:water ratio (OWR) in the obtained emulsion gels were evaluated through a response surface design and the optimal formulations in terms of gel strength were determined. Moreover, the rheological properties and the nanostructure of the optimal emulsion gels were evaluated to determine the impact of the different carrageenan grades. These results provide the basis for the formulation of emulsion gels with target properties for their application in the food industry. 2. Materials and methods 2.1. Materials Different commercial carrageenan grades, in the form of powders were kindly donated by CEAMSA (Pontevedra, Spain): κ-carrageenan (extracted from Euchema cottonii), composed by 92% κ-carrageenan and 8% ι-carrageenan; ι-carrageenan (extracted from Euchema spinosum), composed by 82% ι-carrageenan and 18% floridean starch; and λ-carrageenan (high viscosity grade) (extracted from Gigartina pistillata), composed by 29% λ-carrageenan, 27% κ-carrageenan, 2% μ-carrageenan, 20% θ-carrageenan and 19% starch [28]. KCl was obtained from Sigma-Aldrich (Spain) and sunflower oil was purchased in a local market (Valencia, Spain). The commercial carrageenan powders were dried in an oven at 60 °C overnight and then stored in a 0%RH cabinet until further use. 2.2. Preparation of carrageenan emulsion gels Emulsion gels were produced from κ-carrageenan (κ-C), ι-carrageenan (ι-C) and λ-carrageenan (λ-C), with the addition of salt (KCl). The range of carrageenan and salt concentration, as well as the oil:water volume ratio (OWR) (cf. Table 1) were established by preliminary experiments and a central composite design was carried out for each carrageenan type. The required amount of carrageenan powder was dispersed in hot water (90 °C) for 30 min. After that, the required volume of oil was added and the samples were homogenized with an Ultraturrax for 2 min, producing an emulsion. The salt was then added to the emulsion, which was gently stirred until the salt was completely dissolved. Aliquots of the prepared emulsions were directly transferred to quartz capillaries for SAXS analyses or to the rheometer plate. For the penetration tests, the hot emulsions were transferred to cylindrical containers (30 mm diameter, 40 mm height) and were cooled down to room temperature and subsequently stored at 4 °C for 24 h prior to the analyses. For selected formulations, the corresponding hydrogel systems (the same formulations but without the addition of oil) were also prepared and subjected to rheological and small angle scattering characterization. 2.3. Emulsion gel strength The emulsion gel strength values were determined from penetration assays using a texture analyser (Stable Micro Systems model TA-XT2, Surrey, UK) equipped with a cylindrical Teflon probe (1 cm diameter). The assays were carried out at room temperature (20–25 °C) and consisted in one penetration cycle at a penetration rate of 1 mm/s, to a depth of 5 mm. All measurements were performed, at least, in triplicate. 2.4. Experimental design A statistical central composite design was carried out to evaluate the influence of the carrageenan concentration (C), salt concentration (KCl) and oil:water volume ratio (OWR) in the emulsion gel strength (GS). The considered ranges for each variable in this model are shown in Table 1 and were established based on preliminary experiments, ensuring that emulsion gels with good homogeneity and sufficient mechanical integrity to be manipulated could be obtained. According to the central composite design, 15 runs for each carrageenan type were carried out. The matrix designs and the graphical analysis were performed using IBM SPSS Statistic software (v.24) (IBM corp., USA) and Matlab (Mathworks, Inc., USA), respectively. The response values were predicted by the quadratic polynomial equation as follows: ∑∑ ∑ =+ + + == ≠ γβ β βX βXX0 i k i i k ii i ij k ij ij 11 2 (1) where γcorrespond to the predicted response: β0, βi, βii and βij were the constant regression coefficients of the model; Xi and Xj represent the independent variables. The results obtained were analyzed by ANOVA to investigate the significance and fitness of the model, as well as the effect of significant individual terms and the interactions on the responses. The optimization of the response was obtained using a response surface methodology. Table 1 Minimum and maximum levels considered in the statistical central composite design for commercial carrageenan (C) and added salt concentration (S) expressed in % m/v, and the oil:water volume ratio (OWR). Level κ-C ι-C λ-C C S OWR C S OWR C S OWR Min 0.50 0.25 0.00 1.00 0.30 0.00 1.00 0.30 0.00 Max 2.00 1.25 1.25 2.00 0.50 0.80 1.50 0.50 0.80 C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 2 2.5. Oscillatory rheological measurements The rheological measurements were performed on DHR-3 rheometer from TA instruments (USA) using a cone-plate geometry with an angle of 1°, a diameter of 40 mm and 26 μm of gap. Temperature was controlled using a Peltier plate. The cone was equipped with a solvent trap and an evaporation blocker from TA Instruments. In addition, the edge of the sample was covered with a layer of paraffin oil. Freshly prepared hot emulsions were loaded onto the rheometer, which plate was pre-heated to 80 °C. After an equilibration time of 5 min, a cooling step from 80 to 20 °C was performed at a constant rate of 1 °C/min and with a constant strain of 1% (selected from the linear viscoelastic region) and a frequency of 6.28 rad/s. The samples were then kept at 20 °C during 10 min. Finally, a heating step from 20 °C to 85 °C was carried out at a rate of 1 °C/min, 1% strain and a fixed frequency of 6.28 rad/s. Frequency sweeps were also performed at different selected temperatures. The samples, initially equilibrated at 85 °C, were cooled down to the desired temperature at a rate of 1 °C/min and frequency sweeps within the range of 0.05–100 rad/s were carried out, at 1% strain amplitude. Additionally, to determine the linear visco-elastic region, strain amplitude sweeps were performed at 20 °C by applying a constant frequency of 6.28 rad/s and varying the strain amplitude within the range of 0.01–100% (see Fig. S1). 2.6. Fourier transform infrared spectroscopy (FT-IR) FT-IR spectra from the raw oil and the commercial carrageenan powders, the emulsion gel samples and the corresponding hydrogels were recorded in attenuated total reflectance (ATR) mode in a controlled chamber at 21 °C and CO 2 using a Thermo Nicolet Nexus (GMI, USA) equipment. The spectra were taken at 4 cm −1 resolution in a wavelength range between 650 and 4000 cm −1 and averaging a minimum of 32 scans. 2.7. Confocal laser scanning microscopy Confocal scanning laser microscopy (CSLM) was used to visualize the microstructure of the emulsion gels and to determine oil droplet size distribution. Imaging was performed using a Leica TCS SP2 (Leica Microsystems, Germany). From each emulsion gel two thick sections (ca. 500 μm thickness) were carefully cut with the help of a scalpel and placed onto a glass slide with the cross section facing upwards. One drop of Nile Red (0.2% wt. in acetone), which is able to stain fat, was added on top of the sample and covered with a glass cover slip. The light source used was an argon laser with an emission wavelength of 488 nm. An oil immersion objective with a magnification of 40 and a numerical aperture (NA) of 0.50 was used throughout the study. A total of six images were taken for each sample. Oil droplet size was determined by image analysis using the ImageJ software. Measurements were performed from at least 120 droplets on 5 different images and mean and standard deviation values were calculated. 2.8. Small angle X-ray scattering (SAXS) SAXS experiments were carried out in the Non Crystalline Diffraction beamline, BL-11, at ALBA synchrotron light source (www. albasynchrotron.es). Aliquots of selected emulsions and their corresponding hydrogel systems (i.e. the same formulations but without the addition of oil) were placed in sealed 2 mm quartz capillaries (Hilgenburg Gmbh, Germany) and were left to cool down at 25 °C for 24 h to form emulsion gels prior to the experiments. The energy of the incident photons was 12.4 KeV or equivalently a wavelength, λ,of1Å. The SAXS diffraction patterns were collected by means of a photon counting detector, Pilatus 1 M, with an active area of 168.7 × 179.4 mm 2 , an effective pixel size of 172 × 172 μm 2 and a dynamic range of 20 bits. The sample-to-detector distance was set to 6425 mm, resulting in a q range with a maximum value of q = 0.23 Å −1 . An exposure time of 0.5 s was selected based on preliminary trials. The data reduction was treated by pyFAI python code (ESRF) [50] , modified by ALBA beamline staff, to do on-line azimuthal integrations from a previously calibrated file. The calibration files were created from a silver behenate (AgBh) standard. The intensity profiles were then represented as a function of q using the IRENA macro suite [51] within the Igor software package (Wavemetrics, Lake Oswego, Oregon). A scattering background (corresponding to a quartz capillary filled with distilled water) was subtracted from all the samples. Different models were used to fit the experimental data, depending on the type of gel structure formed. The scattering patterns from the pure carrageenan hydrogels (with no added salt) were properly described by a mathematical function consisting of a power-law term plus one Gaussian peak: =+ ⎡ ⎣ ⎢−⎛ ⎝ −⎞ ⎠⎤ ⎦ ⎥+Iq A qIqq Bbkg() ·exp 1 2· n002 (2) The first term in Eq. (2) corresponds to the power-law function (where Ais a prefactor and nis the power-law exponent) to account for the underlying diffuse scattering, the second term corresponds to the Gaussian function (where q 0 is the peak position, I 0 is the intensity of the peak and B is the standard deviation of the peak position) and the fourth term accounts for the incoherent background. The scattering patterns from the carrageenan hydrogels and emulsion gels (with added salt) were significantly different and could not be described by the same function. In the case of κ-C, a Gauss-Lorentz Gel model was required to fit the experimental data from the gels. This model, described by the scattering function (Eq. (3)), calculates the scattering from a gel structure (typically physical networks) [30–32]as the sum of a low-q exponential decay plus a Lorentzian at higher qvalues: <= ⎡ ⎣ ⎢−⎤ ⎦ ⎥+++Iq I qIqξbkg() (0)·exp Ξ 2(0)/(1 ) GL 22 22 (3) where I L (0) and I G (0) are, respectively, the linear coefficients of the Lorentzian and Gaussian terms, Ξis the characteristic mean size of the static heterogeneities in the system under study and ξis the correlation length of polymer–polymer interactions between the fluctuating chains of polymer. The ι-C and λ-C gels could be properly fitted using a correlation length model. This model contains a first term, described by a powerlaw function, which accounts for the scattering from large clusters in the low q region and a second term, consisting of a Lorentzian function, which describes scattering from polymer chains in the high q region: =+ ++Iq A q C qξ bkg() 1() nLm (4) where n is the power-law exponent, A is the power-law coefficient, m is the Lorentzian exponent, C is the Lorentzian coefficient and ξ L is the correlation length for the polymer chains (which gives an indication of the gel's mesh size). 3. Results and discussion 3.1. Optimization of emulsion gels Three types of commercial carrageenans (κ-C, ι-C, λ-C) were used to produce emulsion gels in presence of salt. The effect of the carrageenan type and concentration (C), the salt concentration (S) and the oil:water volume ratio (OWR) on the emulsion gels' strength were evaluated through a central composite design. The measured emulsion gel strength (GS) values for each type of carrageenan are compiled in Tables S1–S3 (in the Supplementary material). The results obtained for each type of carrageenan were properly fitted using quadratic models C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 3 (R 2 > 0.97) (cf. Table S4 for the associated statistical parameters), described by the following equations: −=− + ∗+ ∗+ ∗ − ∗ ∗∗ ∗−∗+ ∗ GSκ C 21.44 18.73 C 55.06 S 12.44 OWR 5.72 C –48.96 S –8.78 OWR –3.9 CS 2.75 COWR 0.74 SOWR 2 22 (5) =−∗−∗+∗ +∗+ ∗− ∗ − ∗ − ∗ + ∗ − GS 0.34 0.15 C 1.27 S 0.13 OWR 0.12 C 2.0 1 S 0.20 OWR 0.16 CS 0.04 COWR 0.30 SOWR ιC 2 22 (6) =−+∗+∗+∗ −∗− ∗− ∗ − ∗ + ∗ − ∗ − GS 0.38 0.57 C 0.71 S 0.44 OWR 0.21 C 0.5 8 S 0.30 OWR 0.01 CS 0.03 COWR 0.32 SOWR λC 2 22 (7) To investigate the effect of the studied parameters (commercial carrageenan concentration, salt concentration and OWR) on the emulsion gel strength, the response surface plots were also obtained for each type of carrageenan (cf. Figs. 1–3). In general, the results showed that the κ-C emulsion gels were the strongest compared with those obtained from ι-C and λ-C. This was already anticipated since the lower amount of sulphate substitution in κ-carrageenan (the main component in the κC grade) is known to facilitate the formation of stronger hydrogenbonded gel networks [28]. For the κ-C emulsion gels, the statistical analysis showed that the linear term of OWR, all the quadratic terms and the interaction between κ-C and OWR presented significant effect on the gel strength (P< 0.05). The obtained response surface plots, shown in Fig. 1, illustrate that the highest gel strength values were obtained within the middle range of all the studied variables. This is in contrast with what has been previously reported for κ-C hydrogels, where higher strength values were obtained when increasing the κ-C and KCl concentrations [28]. Whereas the creation of stronger carrageenan networks is favourable for the development of hydrogels, these strong interactions between the polysaccharide chains may not be completely desirable in the emulsion gels since it could impede a proper dispersion of the oil droplets within the polysaccharide matrix. According to Eq. (5), the highest gel strength (8.42 N) would be obtained for a formulation of 1.3% κ-C + 0.5% KCl and 0.5 OWR. For the ι-C emulsion gels, the linear terms of ι-C and KCl concentration, as well as the quadratic terms of ι-C and OWR, showed a significant effect on the gel strength (P< 0.05). In particular, as shown in Fig. 2A and B, the ι-C content had the strongest effect on the emulsion gel strength, with the greatest strength values corresponding to the highest ι-C contents. Similarly, as deduced from Fig. 2C, greater salt contents favoured the formation of stronger emulsion gels. The maximum hydrogel strength (0.30 N), calculated from Eq. (6), corresponded to a formulation of 2.0% ι-C + 0.5% KCl and 0.5 OWR. The same behaviour has been reported for ι-C hydrogels, where the maximum ι-C and KCl concentrations produced the strongest hydrogel [28]. The reduced strength of ι-C emulsion gels as compared with κ-C can be attributed to the electrostatic repulsive forces exerted by the greater amount of sulphate groups in the ι-carrageenan chains (the main component in the ι-C grade), hence producing weaker interactions between the polysaccharide chains. Moreover, although these repulsive forces can be counteracted with the addition of salts, weaker networks Fig. 1. Response surface plots of the emulsion gel strength for κ-C. Effect of (A) κ-C and KCl concentration (OWR fixed at central point), (B) κ-C concentration and OWR (KCl concentration fixed at central point) and (C) KCl concentration and OWR (κ-C concentration fixed at central point). C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 4 held by ionic cross-linking have described in the case of ι-C [28]. The emulsion gels obtained from λ-C presented similar strength values to those from ι-C. Even though λ-carrageenan is thought to be a non-gelling polysaccharide, the presence of 27% κ-carrageenan in the λ-C grade seemed to be sufficient to guarantee its gelation capacity. The statistical analysis showed that the linear terms of the λ-C concentration, the OWR, the interaction between the KCl concentration and the OWR, and all the quadratic terms presented a significant effect in the gel strength. Fig. 3 shows that greater gel strength values could be attained by increasing the λ-C concentration and the OWR. The formulation consisting of 1.4% λ-C + 0.4% KCl and 0.6 OWR was the optimal in terms of gel strength (0.28 N), according to Eq. (7). This formulation is very similar to the κ-C optimum, although the strength was significantly greater in κ-C as compared with λ-C; thus, it seems that the behaviour of the λ-C grade was mostly governed by the small fraction of κ-carrageenan. The three formulations corresponding to the maximum gel strength (calculated from the theoretical models) for each type of carrageenan were prepared and their strength values were measured. Table 2 shows the emulsion gel strength values predicted by the models and their corresponding experimental results, demonstrating the accuracy of the models to predict the behaviour of the carrageenan emulsion gels. As deduced from the results, the carrageenan concentration needed to obtain the maximum strength was higher in the case of ι-C, which can be related to the inherent weaker intermolecular interactions of the ιcarrageenan present in this commercial grade. As previously described [28], while κ-carrageenan chains seem to be able to associate by hydrogen bonding, the greater amount of sulphate groups in ι-carrageenan prevents the formation of strong intermolecular association. The addition of salts is therefore required in ι-C, promoting the gelation through ionic cross-linking. On the other hand, the salt concentration and OWR needed to form stronger emulsion gels where very similar for the three carrageenan grades. The optimum salt content was intermediate for the strong κ-C emulsion gels (to facilitate a proper dispersion of the oil droplets within the strongly bound aqueous carrageenan phase), while greater concentrations were required for the ι-C and λ-C grades (to promote intermolecular chain association). With regards to the OWR, an increase up to intermediate oil contents enhanced the emulsion gel strength values. Further increasing the amount of oil reduced the integrity of the materials leading to softer gels. This may be attributed to an excessive proportion of oil limiting the cross-linking between the carrageenan chains [1]. 3.2. Rheological characterization of optimum formulations The rheological behaviour of the optimum formulations providing the highest emulsion gel strength values for the κ-C, ι-C and λ-C grades was evaluated by means of oscillatory rheological measurements. Temperature and frequency sweeps were carried out to investigate the gelation mechanism of the three produced emulsion gel systems. Their respective hydrogel systems (i.e. the same formulations with no added oil) were also characterized to evaluate the effect of the incorporated oil Fig. 2. Response surface plots of the emulsion gel strength for ι-C. Effect of (A) ι-C and KCl concentration (OWR fixed at central point), (B) ι-C concentration and OWR (KCl concentration fixed at central point) and (C) KCl concentration and OWR (ι-C concentration fixed at central point). C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 5 on the rheological properties of the structured systems. The elastic (G′) and viscous moduli (G″) were firstly recorded during cooling and heating ramps and representative results are illustrated in Fig. 4.As observed, all the samples presented a clear sol-gel transition, showing an initial stage at which G′and G″remained almost constant, followed by a sharp increase in both moduli producing the cross-over between G′ and G″(here defined as the gelling point) and a final step where both moduli reached a plateau or continued increasing slowly. The corresponding gelling temperatures (T gel ), as well as the G′and G″values of the emulsion gels and the control hydrogels at 20 °C were determined and the results are summarized in Table 3. In general, the emulsion gel samples presented the same (κ-C) or even higher (ι-C and λ-C) T gel than the corresponding hydrogels, indicating that the carrageenan coil-helix transition, responsible for the gelation, was somehow facilitated in the case of the emulsions. Similarly, [33] reported a higher gelling temperature (T gel ~35 °C) for an emulsion gel prepared from a mixture of locus bean gum and carrageenan as compared with the hydrogel from the same mixture (T gel ~13 °C). Moreover, the G′values at 20 °C were higher for the emulsion gels than for the control hydrogels (ca. 2-fold increase), suggesting the formation of a stronger gel network in the former ones. This is in agreement with the results from the optimization, where it was observed that maximum gel strength values were attained at intermediate OWR values (cf. Table 2). Several authors have also reported higher G′and G″values for emulsion gels containing κ-C as compared with the corresponding aqueous solutions [19,33]. This observation has been consistently attributed to systems where the oil droplets acted as active fillers [34–36], i.e. the droplets were somehow interacting with the gel matrix and thus, contributed to gel strength. However, it should be mentioned that no direct evidence for the interaction between the matrix and the filler was provided in these studies. On the contrary, a negative effect has been observed in systems where no interactions exist between the filler and the gel matrix [10,37]. Thus, one possible explanation for the increased strength of the emulsion gels would be the existence of interactions between the oil phase and the carrageenan-rich aqueous phase. Another possible explanation is simply that the carrageenan concentration in the aqueous phase of an emulsion gel is greater than that of the corresponding hydrogel formulation. At higher carrageenan concentrations stronger and more stable hydrogels (with higher T gel and T m ) are formed [28,38]. In other words, the emulsion gels would consist of more concentrated and stronger carrageenan hydrogel networks containing homogeneously distributed oil droplets. To investigate the possible interactions established between the carrageenans and the oil phase, FT-IR spectra of the Fig. 3. Response surface plot of the emulsion gel strength for λ-C. Effect of (A) λ-C and KCl concentration (OWR fixed at central point), (B) λ-C concentration and OWR (KCl concentration fixed at central point) and (C) KCl concentration and OWR (λ-C concentration fixed at central point). Table 2 Model-predicted and experimental strength values for the optimum emulsion gel formulations, being C (%) the carrageenan concentration, S (%) the salt concentration, OWR oil:water volume ratio and GS (N) the theoretical emulsion gel strength. Data shown as mean ± SD, n=3. C (%) S (%) OWR GS (N) Experimental strength (N) κ-C 1.33 0.51 0.50 8.42 7.92 ± 0.30 ι-C 2.00 0.50 0.50 0.30 0.28 ± 0.05 λ-C 1.37 0.41 0.58 0.28 0.27 ± 0.01 C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 6 20 30 40 50 60 70 80 0.1 1 10 100 1000 10000 aP /' ' G d n a' G Temperature /°C A 20 30 40 50 60 70 80 0.1 1 10 100 1000 10000 aP/''Gdna'G Temperature /°C C 20 30 40 50 60 70 80 0.1 1 10 100 1000 10000 aP /' 'G d n a ' G Temperature /°C D 20 30 40 50 60 70 80 0.1 1 10 100 1000 10000 aP/ ''Gdn a'G Temperature /°C E 20 30 40 50 60 70 80 0.1 1 10 100 1000 10000 a P/''Gdna'G Temperature /C F 20 30 40 50 60 70 80 0.1 1 10 100 1000 10000 aP/''Gdna'G Temperature /°C B Fig. 4. Temperature dependence of G′and G″moduli of carrageenan emulsion gels during cooling (solid markers) and heating (open markers) ramps. (A) κ-C hydrogel; (B) κ-C emulsion gel; (C) ι-C hydrogel; (D) ι-C emulsion gel; (E) λ-C hydrogel and (F) λ-C emulsion gel. C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 7 emulsion gels and the corresponding hydrogels were recorded and the bands characteristic from the sunflower oil were compared in the raw oil and the emulsion gel to identify band shifts indicative of oil-carrageenan interactions. Fig. S2 shows, as an example, the spectra obtained for the ι-C gels. As observed, the spectra from the ι-C hydrogel corresponded mainly to that of water (due to the high hydration level in the sample), while the ι-C emulsion gel showed several bands attributed to the presence of the sunflower oil, being the most intense ones those located at ca. 3010 cm −1 , 2960 cm −1 , 1740 cm −1 , 1460 cm −1 , 1380 cm −1 and 1160 cm −1 . The position of these bands was not shifted with regards to the raw sunflower oil, hence supporting the fact that no interactions were being established between the carrageenan and the oil phase. Assuming that the commercial carrageenans were only dispersed within the aqueous phase, the commercial carrageenan concentrations in the aqueous phase would be 1.94% for the κ-C emulsion gel, 2.99% for ι-C and 2.17% for λ-C. Interestingly, the increase in the G′values of the emulsion gels with respect to their corresponding hydrogels is very well correlated to the increase in the real concentration of the commercial carrageenans within the aqueous phase in the emulsion gels with respect to their concentrations in the hydrogels. This supports the fact that the changes observed in the properties of the emulsion gels with regards to the hydrogels are due to a polysaccharide concentration effect. When comparing the three different emulsion gels, the T gel values were very similar for κ-C and ι-C, whereas λ-C presented a significantly lower T gel . This could be explained by the heterogeneous composition of the λ-C grade, mostly composed of non-gelling carrageenans and containing only a small fraction of gelling κ-carrageenan. The high fraction of non-gelling carrageenans obstructs the formation of helical strands, hindering the gelation process and therefore, lower temperatures are required for the formation of gel networks. As expected, the κC emulsion formed stronger gels (with a strength value very similar to that previously reported for other κ-C emulsion systems [19]) than the ι-C and λ-C, as deduced by the higher G′values of the former. This can be explained by the lower sulphate substitution in κ-carrageenan, which facilitates the association of the carrageenan double helices due to reduced electrostatic repulsive forces [28]. Furthermore, the addition of KCl is known to promote the intermolecular association of κ-carrageenan double helices by hydrogen bonding through neutralization of the negative charges from sulphate groups [28,39], giving rise to the formation of very strong gel networks. After equilibration at 20 °C to set the formed gels, the samples were heated up to 85 °C. A decrease in G′and G″was observed when increasing the temperature, followed by the cross-over between G′and G″ indicating the gel-sol transition, i.e. the disruption of the carrageenan double helices. The melting temperatures (T m ) of the ι-C and λ-C emulsion gels were higher than those from the corresponding hydrogel formulations, once again indicating that stronger networks were created in the case of the emulsion gels. Interestingly, the κ-C and λ-C gels presented large hysteresis, i.e. a marked difference was observed between the cooling and heating behaviour, being T m higher than T gel . This phenomenon has been previously observed for carrageenan and agar gels and was ascribed to the formation of larger aggregates of polysaccharide helices [40,41]. The formation of these aggregates originates strong gel networks, which are more thermally stable and thus, higher temperatures are required to disrupt the carrageenan helical Table 3 Rheological properties of carrageenan optimal emulsion gels. Gelation temperature (T gel ), melting temperature (T m ) and elastic modulus (G′ 20°C ) and viscous modulus (G″ 20°C ) at 20 °C. T gel (°C) G′ 20°C (kPa) G″ 20°C (kPa) T m (°C) κ-C hydrogel 62 12.5 0.60 79 κ-C emulsion gel 61 25.0 1.40 76 ι-C hydrogel 48 0.50 0.02 49 ι-C emulsion gel 61 1.0 0.03 61 λ-C hydrogel 46 0.8 0.100 61 λ-C emulsion gel 53 1.7 0.100 72 0.1 1 10 100 0.1 1 10 100 1000 10000 100000 aP/''Gdna'G Frequency /rad s-1 A 0.1 1 10 100 0.1 1 10 100 1000 10000 100000 aP/''Gdna'G Frequency /rad s-1 B 0.1 1 10 100 0.1 1 10 100 1000 10000 100000 aP/''Gdna'G Frequency /rad s-1 C Fig. 5. G′and G″as a function of frequency for the optimal emulsion gels, measured at temperature of 70 °C (solid markers) and 20 °C (open markers). (A) κ-C emulsion gel; (B) ι-C emulsion gel; (C) λ-C emulsion gel. C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 8 aggregates upon heating. In contrast, the ι-C gels did not present thermal hysteresis, suggesting that a distinct type of gel structures were formed [24,42]. In fact, it has been demonstrated that gelation of ι-C with the addition of salts occurs mostly through ionic cross-linking, leading to the formation of much weaker gel structures [28]. All the emulsion gels were subjected to frequency sweeps at 70 and 20 °C and the frequency dependence of G′and G″was determined at a fixed strain of 1%. As shown in Fig. 5, all the emulsion gels presented a solid-like behaviour (G′>G″) at 20 °C, which was frequency independent. Similar results were reported by [43] for κ-C emulsions using KCl (G~10 kPa) at 25 °C. At a higher temperature of 70 °C (before the gelling transition of the three carrageenan grades) all the emulsion systems presented a frequency dependent behaviour. While G′was higher than G″at low frequencies, a cross-over point was observed when further increasing the frequency indicating a fluid-like behaviour. 3.3. Microand nanostructural characterization of optimum formulations As shown in Fig. 6, the optimized κ-C, ι-C and λ-C emulsion gels and the corresponding hydrogel formulations presented a good integrity and homogeneity. The microstructure of the emulsion gels was further investigated by means of confocal microscopy by staining the sunflower oil with Nile red. Representative confocal images (cf. Fig. 6G–I) show the distribution of the oil droplets within the carrageen matrix and evidence a great heterogeneity of droplet sizes. Looking at the average droplet sizes (35.8 ± 21.1 μm for κ-C, 21.9 ± 16.4 μm for ι-C and 27.6 ± 18.1 μm for λ-C), it seems that κ-C contained a greater population of large droplets; however, due to the large standard deviation values, the differences between samples were not statistically significant. Previous studies have reported that oil coalescence, induced by the addition of KCl, took place in κ-carrageenan emulsion-filled gels, giving rise to oil droplets much larger than those occurring in the original emulsions [18,44]. In addition, Wu et al. [29]reported that oil body emulsions stabilized with ι-carrageenan presented a more stable structure (less coalescence took place) than those stabilized with κand λ-carrageenan and ascribed such behaviour to the higher charge density of the ι-carrageenan helical structures, forming more charged interfacial membranes and thus reducing the aggregation of oil droplets. This, together with the inherent difficulty of dispersing the oil into the more viscous κ-C solutions during the homogenization step to prepare the emulsions, may explain the larger average droplet size in the κ-C emulsion gel. The nanostructure of the stronger emulsion gels was also investigated by means of SAXS experiments. To understand the structural changes induced by the presence of the oil and by the addition of KCl, the same formulations used for the emulsion gels were also used to produce hydrogels with and without the addition of the salt and the obtained materials were characterized. The experimental scattering curves, shown in Fig. 7, evidenced clear structural differences between the carrageenan grades. Interestingly, while a scattering peak was clearly detected in the pure carrageenan hydrogels, such feature was absent when KCl was incorporated into the formulations. Comparable correlation peaks have been detected in the SAXS patterns from κ-carrageenan solutions and gels and were ascribed to the existence of bundles of parallel carrageenan double helices [45]. To extract more information from the SAXS results, a mathematical function consisting of a power-law term plus a Gaussian peak was applied to fit the experimental data from the pure carrageenan hydrogels. The obtained fitting parameters, gathered in Table 4a, indicate structural differences between the different carrageenan grades. The power-law exponents from κ-C and ι-C are characteristic of non-aggregated polymeric chains, whereas the exponent for the λ-C hydrogel is indicative of a clustered network. This may be a consequence of the more heterogeneous composition of λ-C, giving rise to more randomly distributed structures within the corresponding size range (> 60 nm). On the other hand, the characteristic size of the aggregates of double helices decreased from ca. 15 nm for κ-C, to 12 nm for λ-C and 9 nm for ι-C. The formation of larger aggregates in κ-C is most likely related to the lower amount of sulphate groups in κ-carrageenan, enabling a greater degree of intermolecular chain association. The gel systems produced with the addition of KCl showed distinct scattering patterns, where the Gaussian peak was not visible or replaced by very broad shoulder-like features. Thus, the power-law plus Gaussian peak model was not suitable for these data. According to previous works [28,30], a Gauss-Lorentz gel model is appropriate to fit the scattering data from carrageenan hydrogels. In the present work, ABCEDF G H I Fig. 6. Visual appearance of κ-C hydrogel (A) and emulsion gel (B); ι-C hydrogel (C) and emulsion gel (D); λ-C hydrogel (E) and emulsion gel (F). Confocal microscopy images of (G) κ-C, (H) ι-C and (I) λ-C emulsion gel samples. C. Fontes-Candia, et al. Algal Research 47 (2020) 101873 9