Multiscale monitoring of oleogels during thermal transition

2.1 Materials

Rapeseed oil was obtained from a local food retail market (MEGA eG, Stuttgart, Germany). Butylated hydroxytoluene (W218405-1KG-K) was supplied by Sigma-Aldrich (St. Louis, MI, USA) and EC (Ethocel Standard 10 Premium) by DOW Wolf Cellulosics GmbH (Bomlitz, Germany). Koster Keunen Holland B.V. (Aalst, Netherlands) kindly provided RBX (E00130).

2.2 Sample preparation

RBX and EC were mixed with 20 wt% total oleogelator concentration in the following ratios (RBX:EC): 20:0, 10:10, and 0:20. Each mixture was added to a beaker with 80 wt% rapeseed oil with an addition of 0.5 ppm butylated hydroxytoluene and transferred into an oil bath. The suspension was mixed with a propeller stirrer (100 rpm), heated from room temperature to 150 ± 1°C, and held for 10 min to fully dissolve the components. Then, the liquid oleogel was poured into a plastic mold with dimensions of 120 mm (length) × 120 mm (width) × 4 mm (height) and subsequently cooled for 25 min at 7°C. Finally, the oleogel was further processed and stored according to the experiments.

2.3 DSC

A DSC 8000 (Perkin Elmer, Waltham, Massachusetts, USA) was used for DSC measurements. Pieces of the oleogel were stored at 4°C for 1 day. Approximately 2–4 mg oleogel was weighed into an aluminum pan and sealed hermetically. The samples were heated from 20 to 140°C at a rate of 5 K/min with a holding time of 1 min at each start and end temperature and then subjected to a re-cooling process under the same conditions. Thermo-event temperatures were determined using the Pyris software from Perkin Elmer. Each oleogel was measured in triplicate.

2.4 Simultaneous DMA, dielectric, and Raman spectroscopy

The simultaneous dielectric, Raman, and dynamic mechanical characterization of the material was conducted by coupling a rheometer (MCR302e, Anton Paar, Graz, Austria) with an LCR meter (IM35536, Hioki, Ueda, Japan) and a Raman spectrometer (Cora 5001, Anton Paar, Graz, Austria). The rheometer was equipped with a glass-Peltier temperature unit for plate-plate and cone-plate measuring systems, paired with a Peltier hood. The glass plate inserted into the Peltier temperature device was coated on the upper side with an indium tin oxide coating, providing conductive properties while maintaining transparency for Raman spectroscopy. The Raman probe was placed right below the glass plate, measuring at one fixed position of the plate, which can be adjusted by an xy-adjustment screw. The contact between the LCR meter and measuring cell was established by a silver-wire contact to the upper measuring system, enabling it to oscillate with almost no friction due to the contact. The second contact was established by placing a copper ring on top of the indium tin oxide-coated glass plate, which was pressed on the surface by the mounting screw of the glass plate. The measurement was synchronized by the rheometer, submitting trigger signals to the Raman spectrometer and LCR meter. Figure 1 summarizes the experimental setup.

Figure 1

Schematic experimental setup of the multi-scale experiment. An indium tin oxide-coated glass temperature unit mounted to a rheometer served as a measuring cell for dielectric spectroscopy, Raman spectroscopy, and DMA from 150 to −10°C.

The aim of this work is to demonstrate the benefits of a novel combination of dielectric spectroscopy, Raman spectroscopy, and DMA. Accordingly, the data are discussed qualitatively. Reproducibility was assessed for one of the presented samples (data not shown) and showed a standard deviation in the range of 5–7% across all tested methods.

3.1 DSC

Representative thermograms for all oleogels are shown in Figure 2. During heating, a single endothermic melting event was determined for both wax-containing oleogels. The RBX oleogels show a higher melting peak at 75.3 ± 0.6°C compared to the RBX:EC oleogels, which have a melting peak at 72.3 ± 0.9°C. This shift can be mainly attributed to the differences in the wax concentration as previously described for other waxes [19].

Figure 2

Representative heating (a) and cooling (b) thermograms of oleogels with three different ratios of RBX and EC. Measurements were performed from 20 to 140°C and back to 20°C at a rate of 5 K/min. The curves are labeled as RBX:EC ratios (w/w, relative to the total sample).

Similar behavior was found in the cooling cycle with one crystallization event and a higher crystallization peak for the RBX oleogels, relative to the multicomponent oleogels. RBX oleogels exhibit a peak crystallization temperature at 72.2 ± 0.2°C, and RBX:EC oleogels at 68.7 ± 0.4°C. Single melting and crystallization behavior was observed in previous studies with wax and EC containing oleogels [19,23]. Furthermore, RBX is considered a chemically homogeneous wax, with no more than one crystallization and melting peak per sample expected [9,11].

These thermodynamic events originate from a phase transition of RBX, involving combinations of either melting, dissolution, or crystallization, as extensively described in the literature [6,8,24,25]. The melting point for pure RBX has been determined to be in the range of 76–84°C [9–11]. Oleogels’ melting points are lower compared to those of the virgin waxes due to the solubility of the wax components in oil [8,25]. Thermal hysteresis between the recorded melting and crystallization events is a common crystallographic phenomenon explained by the energy barrier of nucleation. Therefore, recorded crystallization temperatures of RBX ranged between 76 and 70°C [8,9].

EC oleogels display no melting or crystallization events in the recorded temperature range, which is in accordance with the literature since semi-crystalline EC starts to melt at ∼180°C [15,26]. EC typically shows a glass transition at 120–135°C [17,26]. This is visible within the cooling cycle of EC oleogels at 130.3 ± 1.6°C. On the other hand, no glass transition is apparent in the heating cycle, which can be linked to thermal hysteresis, which has been described for EC oleogels before [26]. As seen in previous studies as well, no thermal events are observed at or below the glass transition temperature of EC [19,23,26–28]. This has been suggested as an indicator for no secondary structure formation of EC in oleogels [26]. These thermodynamic events are necessary to induce gelation, forming self-stable oleogels, which are necessary for plant-based fat alternatives [29]. Their analysis gives a deeper understanding of these processes and the thermal parameters that are crucial for food production.

3.2 DMA

Figure 3 shows the complex shear modulus | G * | (a) and the loss factor tan δ (b) over temperature as measured from the dynamic mechanical test performed at 1 Hz. Since the experiments were conducted from 150 to −10°C, the results are read from right to left. At 150°C, the complex shear modulus starts at low values of around 10⁻¹ to 10¹ Pa. Samples with RBX start at significantly lower values than the EC oleogel. The increased modulus can be explained by the polymeric structure of EC, acting as a thickener. With decreasing temperature, the RBX oleogel shows a sudden stepwise increase in | G * | at a temperature of about 65°C up to a plateau of about 10⁶ Pa. The EC oleogel shows a gradual increase starting at about 120°C and achieves a much lower maximum value of about 4–10⁴ Pa at a temperature of approximately 50°C. The RBX:EC oleogel shows both effects: First, a gradual increase in the complex shear modulus at a temperature of approximately 105°C, followed by a sudden step in | G * | to ∼10⁶ Pa at a temperature of about 65°C. While the gradual increase in the modulus is related to the main glass transition as the EC goes from amorphous into the glassy state, T G , EC , the sudden stepwise increase is due to crystallization from the liquid phase of RBX. The DMA results are in good qualitative agreement with those obtained from the reference DSC experiments.

Figure 3

Complex shear modulus | G * | (a) and loss factor tan δ (b) as measured during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The curves are labeled as RBX:EC ratios (w/w, relative to the total sample).

The main glass transition can also be read from Figure 3b as a clear maximum in the loss factor for samples containing EC. The EC oleogel shows a maximum at a temperature of about 136°C, while the RBX:EC oleogel shows the maximum at a temperature of about 117°C. In addition, the RBX:EC oleogel shows the maximum in tan δ at a much smaller level than the EC oleogel. This drop can be attributed to RBX acting as a crystalline plasticizer in a multicomponent oleogel, as reported before with lauric acid [30]. Below 100°C, a gradual increase in tan δ is observed for all three samples. However, from about 60°C, the mechanical loss factor drops to 0 for the RBX containing samples as a result of an abrupt physical change. Once the samples are in a solid state, little information can be obtained from tan δ in the dynamic mechanical measurement performed with the parallel-plate geometry.

3.3 Dielectric spectroscopy

In the dielectric spectroscopy, an electric current flowing through the sample in response to alternating electric fields generated by the LCR meter was measured as a function of the field frequency, which varied between 10 Hz and 1 MHz. Thus, the dielectric material was internally polarized. This interaction with the applied electric field results in a redistribution of charge within the atoms and molecules of the dielectric [20]. Consequently, two main raw dielectric data can be obtained from the experiment: the impedance | Z | given in Ω and a dielectric loss factor, tan δ e . The dielectric loss factor is a direct indicator of viscoelasticity and is measured as the delay between the preset voltage and the measured current. Figure 4 shows both | Z | and tan δ e as a function of the dielectric frequency. In all samples, the complex impedance decreases with increasing frequency, following a typical capacitive response. At low frequencies and high temperatures, all three oleogels exhibit a plateau, suggesting the presence of direct current conduction due to polarization of the dielectric. However, other factors are known to influence the polarization behavior of heterogeneous food samples, such as water or trace elements [31]. Lizhi et al. [32] showed similar dielectric loss behavior for edible plant oils. As the frequency increases, the impedance transitions to an inverse proportional decrease, characteristic of capacitive materials.

Figure 4

Frequency-dependent complex impedance | Z | (a), (c), and (e); and dielectric loss factor tan δ e (b), (d), and (f) as measured during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The graphs are labeled as RBX:EC ratios (w/w, relative to the total sample). Every oleogel is drawn in separate plots as a function of frequency: (a) and (b) 0:20; (c) and (d) 10:10; and (e) and (f) 20:0.

Samples with EC exhibit a stronger resistive response at low frequencies, with a transition toward capacitive behavior at higher frequencies. Consequently, the dielectric loss factor follows a decreasing trend with frequency from 10 to 1,000 Hz. At higher temperatures, an even stronger effect can be observed. Thus, the dielectric loss increases at low frequencies with more EC content and with increasing temperature (compare Figure 4b, d, and f). This behavior is typical for high-loss dielectric polymers [33,34]. If the frequency range is extended to lower values, a peak in tan δ e would be expected at the characteristic relaxation frequency. At this frequency, dipoles begin to lag behind the alternating electric field, leading to increased energy dissipation and high losses. This peak in tan δ e can also be interpreted as a relaxation process related to molecular motion, similar to how the maximum in the mechanical loss factor is related to the main motion of the EC molecules (Figure 3). Since EC is semicrystalline and the studied material is a heterogenous mixture of different substances, Maxwell–Wagner–Sillars interfacial relaxation could also be a reason for enhanced conductivity [35,36].

Figure 5 shows the impedance at 50 Hz over temperature with distinct behaviors for all three samples. From 150 to 50°C, oleogels containing EC show a similar behavior with decreasing temperature. At 150°C, the impedance starts at a value of ∼20 MΩ, increasing rapidly to >50 MΩ. This increase can be associated with the glass transition of EC and reflects the dynamic mechanical behavior discussed earlier (Section 3.2, DMA). With higher EC content, the T _G,EC shifts to higher temperatures in good accordance with Figure 3b. The RBX oleogel already starts at 50 MΩ and shows a broad peak around ∼85°C. Samples containing RBX present a similar curve with decreasing temperature from 50 to −10°C, showing a slight decrease in impedance. The EC oleogel impedance slightly increases with decreasing temperature.

Figure 5

Complex impedance | Z | measured at 50 Hz during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The curves are labeled as RBX:EC ratios (w/w, relative to the total sample).

Figure 6 compares the dielectric loss factor, tan δ e as a function of temperature at dielectric frequencies of 50 Hz (a), 500 Hz (b), 5,000 Hz (c), and 50,000 Hz (d). At low temperatures, the dielectric loss factor remains close to zero independent of the sample and the frequency. This indicates dominant capacitive behavior. At 50 Hz we can see two different trends in the curves of the samples containing EC (Figure 6a). In both cases, there is a clear bending point in tan δ e which occurs at around 100°C (see black circles and blue squares in Figure 6a). This change in slope does not occur in the RBX oleogel. This bending point is also observed for the other frequencies, 500, 5,000, and 50,000 Hz, and is thus probably related to the glass transition of EC. In the case of the RBX:EC sample, there is a maximum, which is not observed in the EC oleogel because it is probably shifted to higher temperatures at the measured dielectric frequencies. This is in good agreement with the data obtained in DMA, where a maximum in the mechanical loss factor is shifted to lower temperatures with decreasing EC content (compare black and blue triangles in Figure 3b). A similar agreement is observed in Figure 5, where the decrease in | Z | , related to the glass transition of EC, is shifted to higher temperatures with increasing EC content in the samples.

Figure 6

Dielectric loss factor tan δ e measured at (a) 50, (b) 500, (c) 5,000, and (d) 50,000 Hz during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The curves are labeled as RBX:EC ratios (w/w, relative to the total sample).

In the kHz-frequency range, samples containing RBX exhibit an additional peak in the dielectric loss factor (see red triangles and blue squares in Figure 6c and d). This peak shifts toward higher temperatures as frequency increases, appearing at ∼25°C at 5 kHz and shifting to ∼45°C at 50 kHz. At 0.5 kHz, a slight bend in the wax-containing curves indicates a peak around 5°C. This behavior suggests a temperature-activated relaxation process linked to the crystallization of wax esters present in the RBX (T _c,RBX). Since primary crystallization processes are generally frequency independent, the observed peak shift suggests a structural reorganization in the crystalline phase rather than primary crystallization [37]. Crystal nucleation results in many dispersed needle-like crystals, interacting with high frequencies due to short length scales. As the temperature decreases, crystal growth proceeds, resulting in the formation of a superstructure. Larger structures show higher relaxation times; thus, the peak in tan δ e is shifted toward lower frequencies. Another hypothesis would link this maximum to post-crystallization transitions such as α → β′ → β polymorphic transformations, which are known to occur in lipid-based crystalline systems [7,12,38]. These transitions can influence the dielectric loss factor through dynamic reorientation of molecular dipoles within the crystalline structure. Contrary to this theory, no polymorphic transition has been recorded for RBX or wax esters over the given temperature ranges [24,25,39].

3.4 Raman spectroscopy

Representative Raman spectra with wavenumbers from 900 to 1,700 cm⁻¹ for the cooling samples are shown in Figure 7. The focus of this analysis will remain on bands around 1,061, 1,080, 1,130, 1,296, and 1,440 cm⁻¹, as they offer interpretations concerning the oleogel properties over the given temperature range. Less insightful bands in this context include 975 cm⁻¹ ν(C–O–C), 1,275 cm⁻¹ δ(C–H), and 1,630 cm⁻¹ ν(C═C). These bands are less informative due to their rapeseed oil association, which is not actively involved in the oleogels’ property changes [40–42]. The prominent bands at 1,061 and 1,130 cm⁻¹ are attributed to symmetrical and asymmetrical ν(C–C) stretching vibrations, respectively [41,43]. These indicate trans sequences longer than 10 bonds, while a broad band around 1,080 cm⁻¹ is associated with a mixture of gauche and trans bonds over a long range [44]. This leads to a direct correlation between the intensity of the 1,061 and 1,130 cm⁻¹ bands and the samples’ crystallinity due to a low probability of random trans bond sequences >10 [44]. The bands at 1,296 and 1,440 cm⁻¹ can be attributed to δ(CH₂) bending vibrations, while 1,296 cm⁻¹ is subcategorized into twisting vibrations [43,45]. The bands at around 1,061, 1,080, and 1,130 cm⁻¹ are overlapping as well as the 1,275 and 1,296 cm⁻¹ bands. This complicates the analysis and quantification of these bands, but qualitative conclusions can still be drawn.

Figure 7

Raman spectra between 900 and 1,700 cm⁻¹ obtained during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The intensities are normalized over the stable 1,440 cm⁻¹ band intensity. The curves are labeled as RBX:EC ratios (w/w, relative to the total sample).

Next, the Raman spectra for the two RBX-containing oleogels are discussed, as they show the same chemical trends, albeit to a different extent. The band at 1,296 cm⁻¹ shows a clear peak at high temperatures, which becomes narrower and increases in intensity with lower temperatures. This can be linked to a growth in the samples’ crystalline content and, therefore, a direct indicator for a phase transition. As RBX transitions from a liquid to a crystalline state, its bands become narrower. This is due to a more defined force constant of the vibrating chemical bonds. The Raman band’s wavenumber is dependent on the force constant of the vibrating bonds. In a less defined liquid chemical environment, they have a broader distribution, while crystals with high order over long ranges lead to narrower bands [41]. The bands at 1,061 and 1,130 cm⁻¹ are barely visible at higher temperatures, becoming more apparent as the temperature decreases. As previously described, these bands increase in intensity during crystallization due to their trans bond sequential origin. Accordingly, the band at 1,080 cm⁻¹ decreases in intensity. Additionally, a slight shift toward a lower wavenumber is visible (e.g., band 1,296 cm⁻¹). This might be attributed to a shift in the polymorphic form of the RBX crystals, comparable to cocoa butter polymorphs [46]. The EC oleogel shows no signs of crystallization as most bands remain constant over the temperature range. The bands at 1,338 and 1,600 cm⁻¹ have been linked to pure EC in previous work and are recognizable in the EC oleogel spectra, although not clearly [47]. In the RBX:EC oleogel spectra, these bands are probably not discernable because of the low concentration of EC.

Due to their high variability, the normalized intensities of bands 1,061, 1,130, and 1,296 cm⁻¹ are plotted over the cooling range from 150 to −10°C in Figure 8. The EC oleogel shows no temperature-dependent change in any relative band intensity. The RBX containing oleogels also show no temperature dependency until ∼60°C when a sudden increase in the relative band intensity can be seen. This leads to the conclusion of a phase transition from a liquid to a crystalline state around this temperature (T _c,RBX). After this initial step, a gradual increase is visible until ∼5°C, which could suggest continuous crystal growth or a crystallization phenomenon after the first-order crystallization. The RBX oleogel shows a steeper slope and a higher plateau at low temperatures compared to the RBX:EC oleogel. This can be attributed to a higher total crystalline content coming from the higher wax content. The EC oleogel shows no phase transition over the whole temperature range, as the relative band intensities remain constant. As there was no thermodynamic phase transition visible in the thermograms, no change was expected for these relative intensities. The Raman spectroscopy results are in good qualitative agreement with DMA and DSC, showing a higher crystallization temperature for the RBX compared to the RBX:EC oleogel.

Figure 8

Raman spectra of the peak intensities at 1,061, 1,130, and 1,296 cm⁻¹ over the intensity at 1,440 cm⁻¹ obtained during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The curves are labeled as RBX:EC ratios (w/w, relative to the total sample).

3.5 Insights from the multiscale analysis of oleogels

DMA directly provides mechanical properties and their temperature dependence, which are critical to the texture and stability of oleogels in food applications. Here, the measurement at 1 Hz is of special interest since this frequency correlates well with chewing conditions [48]. From Figure 3a, for example, it can be seen that at temperatures between 130 and about 60°C, samples with EC show a much higher mechanical stability than samples without EC. Besides, the dynamic mechanical measurements are valuable in providing information about the glass transition in the samples containing EC (see black circles and blue squares in Figure 3b). Within this context, DMA is a good complement to the DSC data, where the glass transition is only visible within the cooling cycle of EC oleogels as a small bending in the curves at temperatures of about 130°C (Figure 2).

DMA is also helpful in describing the onset of crystallization from a macroscopic point of view, but does not provide enough information about crystallization kinetics since there is not enough data resolution between the onset of crystallization, say 65°C, and the observed plateau in | G * | , at about 50°C. In fact, it may not accurately represent the true shear modulus in the crystalline state. Instead, it shows a plateau at the same level of about 10⁶ Pa for both samples crystallizing during the measurements (compare red triangles and blue squares in Figure 3). At these levels, torsional compliance effects in the rheometer may be at play when the 40 mm diameter plate is used and the sample suddenly becomes too stiff (see Müller-Pabel et al. [49] or Hossain et al. [50] for more information on compliance effects). Note that although decreasing the diameter may help in decreasing torsional effects, we need to use the 40 mm diameter to improve the torque signal when the sample is in the liquid state and to improve the dielectric contact. Based on the DMA data, we cannot quantitatively distinguish | G * | in the crystalline state between the RBX and RBX:EC samples. However, the EC sample shows qualitatively much lower | G * | in the same temperature range. This clearly has implications for sensory perception, and the RBX-containing samples would likely feel more brittle compared to the EC sample [51,52].

Additional insights into the crystallization kinetics can be obtained by analyzing the intensities of the 1,061, 1,130, and 1,296 cm⁻¹ bands from the simultaneously measured Raman spectra. As shown in Figure 8, the Raman data show higher resolution between 65 and 50°C compared to the mechanical data in the same temperature range, especially in the peak at 1,130 cm⁻¹ corresponding to asymmetric ν(C–C) stretching vibrations. In fact, while | G * | shows a constant plateau at 50°C for the RBX oleogel, Figure 8b shows a gradual increase in the peak intensity in the same sample until the plateau is reached at a temperature of around 5°C. In other words, while the dynamic mechanical data would show that the crystallization process is completed at about 50°C, the Raman data show that it is not complete until ∼5°C when the plateau is achieved. This could be due to an initial crystallization, with a following transition into a more stable crystal structure. While DMA primarily reflects the glass transition due to the amorphous EC, Raman spectroscopy provides additional information on the crystallization process occurring during the experiment.

DMA is indeed a highly sensitive method for studying glass transitions, with the added advantage of providing insights into the time dependence of these phenomena. However, in the experiment presented here, measurements were performed at a single frequency (1 Hz), but this is where simultaneous dielectric spectroscopy becomes particularly valuable. Dielectric spectroscopy is interesting, first because the dielectric frequency range selected complements the mechanical range. Recall that the dielectric loss factor holds a similar physical meaning to the mechanical loss factor, as both correlate with the energy dissipated internally by the sample. Although in the dielectric case, the excitation is electrical rather than mechanical, it is still interesting to have a complementary frequency range. For example, at a mechanical frequency of 1 Hz, the maximum in tan δ related to the glass transition of EC is observed in the EC-containing samples (Figure 3b). At a dielectric frequency of 50 Hz, the maximum for the RBX:EC oleogel is still visible at the temperatures measured, but no longer for the EC oleogel (Figure 6a–d), where the maximum is probably shifted to higher temperatures than 150°C.

The second complementary advantage is that the dielectric excitation occurs significantly faster than mechanical excitation at 1 Hz. In fact, for each individual data point obtained in the DMA temperature sweep, a complete dielectric spectrum was acquired. In other words, we can obtain time-dependent information about the relaxation processes in the time it takes to perform the DMA. This can be clearly seen from Figure 9, where, for the sake of comparison, the dielectric loss factor is plotted as a function of temperature for different frequencies on a logarithmic scale, for each of the samples: EC oleogel (a), RBX:EC oleogel (b), and RBX oleogel (c). Although the maximum in the tan δ e is not appreciated in the case of the EC oleogel, it can be seen that the bending point describing the onset of the glass transition process is slightly shifted to higher temperatures at high frequencies (Figure 9a). Something similar is seen for the RBX:EC sample, where, besides a second local maximum in tan δ e , is also observed from the figure at lower temperatures (Figure 9b). The appearance of a second local maximum is connected to the third and main advantage of the dielectric excitation: Dielectric spectroscopy extends to much higher frequencies, providing insights into the relaxation dynamics on much shorter time scales. This allows us to probe molecular motions at a smaller scale than what is accessible through DMA. This is evidenced from Figure 9a and b, where at frequencies above 500 Hz, a new peak in the dielectric loss factor emerges, indicating localized molecular motion, particularly in wax-containing samples. These molecular motions would be undetectable by DMA at the temperatures studied, where the maximum accessible frequencies are around 100 Hz for commercial devices.

Figure 9

Dielectric loss factor tan δ e as a function of temperature and frequency during the cooling process from 150 to −10°C of oleogels with three different ratios of RBX and EC. The graphs are labeled as RBX:EC ratios (w/w, relative to the total sample). Every oleogel is drawn in separate plots as a function of temperature: (a) 0:20, (b) 10:10, and (c) 20:0.

We hypothesize that this maximum is related to a relaxation process associated with the crystallization of wax esters present in RBX (T _c,RBX). The shift of tan δ e to higher temperatures with frequency (Figure 9b) suggests a structural reorganization in the crystalline phase. In other words, the higher the frequency, the smaller the excited crystals. Thus, at a frequency of about 100 kHz, the secondary local maximum (T _c,RBX) appears at a temperature of about 55°C for RBX and RBX:EC oleogels. This temperature range is in good agreement with the onset of crystallization that we observed from the stepwise increase in | G * | in DMA (Figure 3a). This is reasonable in that the appearance of small crystals would have already produced a very noticeable increase in the mechanical properties measured in shear during DMA.

The agreement with Raman spectroscopy is also interesting based on the study of the peaks at 1,061, 1,130, and 1,296 cm⁻¹ (Figure 8). As the temperature decreases, the crystallization process continues (as seen in the gradual increase of the peak intensity in Figure 8), probably leading to the formation of larger structures. Larger structures should have longer relaxation times, so that the local maximum in tan δ e shifts to lower frequencies, as shown in Figure 9b and c. The final intensities at 1,061, 1,130, and 1,296 cm⁻¹ are larger for the RBX sample with the highest wax content, as compared to the RBX:EC samples (Figure 8). This may correlate with the more pronounced peak in tan δ e for the RBX sample as compared to the RBX:EC sample peak (compare Figure 9b and c). Besides, at same frequencies, the peaks in tan δ e appears at lower temperatures.