Structure Formation in Tailor‐Made Buriti Oil Emulsion During Simulated Digestion

Functional food emulsions enriched in health‐promoting nutrients can help to maintain and improve health and lifestyle. The oil extracted from the Amazonian buriti fruit is renowned for its high levels of carotenoid, vitamin E, and unsaturated fatty acids, which have been linked to improvements in cardiometabolic health. Here, buriti oil in water emulsions are developed and their colloidal transformations are investigated in an advanced digestion model with oral, gastric, and intestinal parts with in situ synchrotron small‐angle X‐ray scattering, cryogenic electron microscopy, and dynamic light scattering under simulated “healthy” and “compromised” digestive conditions. The interior oil phase of whey‐stabilized buriti oil‐in‐water emulsion transforms into highly ordered lyotropic liquid crystalline structures during simulated intestinal digestion at compromised bile conditions. Simulated gastric digestion influences intestinal digestion by slowing it down, resulting in less ordered structures. The digestion‐triggered structure formation is pH‐ and bile salt‐dependent and can be modulated by adding vitamin E to the oil. This tailoring of structures during digestion offers a new pathway to steer digestion kinetics and nutrient bioavailability.


Introduction
Buriti fruit (Mauritia flexuosa) is a valuable natural resource in the Amazonian region.This fruit is a source of several nutrients, and FFAs that are then uptaken by enterocytes. [16]Since lipases are water-soluble enzymes, they can only interact with their hydrophobic substrates at the oil/water interface.Therefore, the composition and colloidal properties of the interface play a crucial role in lipase activity and the bioaccessibility and bioavailability of the nutrients in the gastrointestinal tract. [17]As surfaceactive molecules, lipolysis products can hinder lipase binding to the oil/water interface.20] BS was also found to affect the colloidal structure during lipid digestion.They actively participate in the lipid self-assembly process of monoacylglycerol and FFAs. [18,21]Mixed micelles and vesicles appear mostly at high BS concentrations in intestinal digestion, whereas lyotropic liquid crystalline (LLC) structures appear at compromised BS concentrations during digestion. [18,22]However, the role of BS concentration between the compromised and the fed state (10 mm) [23] has yet to be studied in detail.This holds significant relevance in developing strategies to improve the absorption of lipids and liposoluble nutrients in patients with BS deficiencies.These individuals often encounter challenges absorbing crucial hydrophobic nutrients, such as vitamin E, abundant in buriti oil. [24]This study will shed light on the nanostructural transformations in buriti oil emulsions in the simulated gastrointestinal tract in the presence and absence of BSs.
The high content of oleic acid and other unsaturated fatty acids in buriti oil and its high amount of hydrophobic nutrients such as carotenoids and vitamin E make this oil a promising stimuliresponsive material for studying nutrient bioaccessibility.The simulated intestinal digestions were carried out at different pH values (6.5, 7.0, and 7.5) to address the hypothesis that buriti emulsion's digestion structures are modified with pH from fatty acid protonation/deprotonation.
The lipolysis of TAGs was found to result in the formation of highly geometrically ordered LLC structures in the simulated intestinal digestion of triolein model emulsion, [27] milk, [28][29][30] krill oil emulsion, [31] mayonnaise, [32] and infant formula. [33]The presence of LLC structures during the digestion of foods can impact the absorption of lipids and the bioaccessibility of nutrients, which can be explored for the design of functional food materials. [34]These thermodynamically stable structures arise from interactions between the hydrophobic and hydrophilic domains of the digestion products, mainly long-chain monoacylglycerols and FFAs, and their hydrophobic part's incompatibility with water.They pack into defined oil-continuous packing geometries, such as lamellar, inverse hexagonal, inverse bicontinuous-, and inverse discontinuous cubic phases. [35]The packing geometry can be predicted from the geometry of individual molecules with the critical packing parameter (CPP) theory [36] illustrated in Figure 1.The repeating unit cells of LLC phases leads to diffraction peaks in small-angle X-ray scattering (SAXS). [37]To answer the hypothesis that the LLC structures solubilize the hydrophobic nutrients such as vitamin E, buriti oil was also digested in the presence of different vitamin E concentrations as a model additive.The structure type and dimensions of the domains, typically in the < 30 nm range, are determined with SAXS.The CPP can be calculated from the volume of the hydrophobic tail (V), the effective headgroup area at the lipid-water interface (a 0 ), and the length of the hydrophobic tail (l c ).
[40] Investigating nanostructural transformations during lipid digestion requires characterization techniques capable of probing at the colloidal scale and appropriate simulated digestion models. [22,34]SAXS is a well-established method for the non-invasive characterization of these structures' size, shape, and morphology. [37]The colloidal fate of TAG emulsions in a more realistic digestion model comprising oral, gastric, and intestinal conditions is primarily unknown. Further, the enzymes and pH conditions during gastric digestion can affect emulsion stability by, for instance, modifying electrostatic interactions or digesting protein stabilizers.
To our knowledge, for the first time, this study investigates the nutrient composition of buriti oil (fatty acid profile, vitamin E, and carotenoid content), its emulsification, and the colloidal structure formation during complete in vitro digestion.The structure formation and transformation are analyzed with online SAXS, cryogenic transmission electron microscopy (cryo-TEM), and dynamic light scattering (DLS), see Figure 2. The multi-step (oral, gastric, intestine) static in vitro digestion model with dynamic structural analysis is based on the standardized conditions established by the INFOGEST network to simulate the digestion of carbohydrates, proteins, and lipids in the different digestive stages. [23]The digestion occurs in a temperature and pHcontrolled vessel.At the same time, the digest is pumped through a closed-loop flow-through SAXS cell for the real-time analysis of structure formation and transformation at all stages of digestion.The analysis is complemented with imaging of structures and particle size analysis at defined oral, gastric, and intestinal digestion stages using cryo-TEM and DLS.These results provide unprecedented insights into colloidal structure formation during digestion.They help decipher the effect of digestion variables such as pH, bile salts, and the incorporation of hydrophobic nutrients on the properties of these structures.

Fatty Acid Profile and Vitamin E Composition of Buriti Oil and Its Emulsification
The fatty acid profile of buriti oil is presented in Figure S1 (Supporting Information).While oleic acid makes up 70.8% of the oil, many other fatty acids are present.Palmitic and stearic acids are the most abundant saturated fatty acids in buriti oil, corresponding to 16.8% and 3.1% of the composition.Different polyunsaturated fatty acids, especially linoleic acid but also -linolenic and gondoic acid, can be found in buriti oil.Regarding its vitamin E content, buriti oil was found to have a large concentration of this nutrient, with a total tocopherol concentration of 1750 mg kg −1 as determined by HPLC, of which 42% correspond to -tocopherol, 51% to (+)-tocopherols and the remaining 7% to -tocopherol. Figure S2 (Supporting Information) compares the amount of vitamin E measured in buriti oil to other vegetable oils where values were obtained from the literature. [43]Finally, the total amount of carotenes was determined by HPLC to be 301 mg kg −1 , of which 3% corresponded to -carotene, 77% to -carotene, 8% to 9-cis--carotene, 11% to 13-cis--carotene, and less than 1% to lutein (Table S1, Supporting Information).
Whey-protein stabilized buriti oil-in-water emulsions were prepared by ultrasonication.Images of the emulsions, taken directly and 10 days after preparation, show homogeneous, turbid liquids of yellowish appearance (Figure S3, Supporting Information).The color is primarily a result of the high carotenoid content.The turbidity results from multiple scattering of light between the emulsion particles with diameters in the range of the wavelength of light.DLS analysis confirmed the emulsions' stability over at least 10 days (Figure S4a, Supporting Information).The hydrodynamic diameter (D H ) directly after preparation was 350 nm, with a polydispersity index (PDI) of 0.39 (Figure S4b, Supporting Information).After 10 days D H was 347 nm and PDI 0.36.The autocorrelation curves obtained by DLS (Figure S5a, Supporting Information) exhibited a single decay throughout the monitored storage period, suggesting a monomodal particle size distribution for this sample.

Self-assembled structures formed during simulated digestion of buriti oil emulsion
The in vitro digestion of the buriti oil emulsion combined simulated oral, gastric, and intestinal digestion steps.The SAXS curve of the original emulsion (Figure S6, Supporting Information) displayed a power-law decay of the scattering intensity proportional to q −3.5 in the lower q region, which is expected from large emulsion droplets.The broad correlation peak centered around 2.70 nm −1 can result from the form factor scattering of whey proteins and the scattering from the local arrangements of TAGs in the internal phase of the droplet.This was also reported in previous SAXS studies on triolein emulsions [26] and whey proteins. [44]he SAXS curve (Figure S6, Supporting Information) of the oral phase (pH 7.0) is similar to the curve of the original formulation (pH ≈ 6.5), with additional peaks at q = 1.36 nm −1 and q = 2.72 nm −1 .These two peaks at a q-ratio of 1:2 correspond to a lamellar phase (with the center-to-center distance between two bilayers of d = 2 q = 4.6 nm) formed between FFAs naturally found in the oil and the Ca 2+ ions in the simulated salivary fluid.These peaks disappeared at the start of the simulated gastric digestion step owing to the low pH of 3.0, where the carboxyl groups of the free fatty acids are mostly protonated.From cryo-TEM imaging of the oral phase (Figure S7, Supporting Information), emulsion droplets between 300 and 500 nm were observed.Time-resolved SAXS curves of the simulated gastric digestion of the buriti oil emulsion are shown in Figure 3a.A decrease in the scattering intensity at the lower q values (forward scattering) can be observed during gastric digestion.The intensity of the scattering curve for the lowest measured q value at 0.05 nm −1 was reduced to 0.4x of its initial value after 120 minutes of digestion.Such a decrease is likely caused by a reduction in the concentration of emulsion particles during gastric digestion, probably owing to emulsion instability during the gastric digestion of the whey protein emulsifier.Throughout the gastric digestion, no further significant changes were observed with SAXS (Figure S8, Supporting Information).Representative cryo-TEM images of this sample at the end of the gastric digestion (Figure 3b) show emulsion droplets of around 300 nm diameter with nanoscale patterns on the particle surface.These surface structures are likely whey protein aggregates on the oil droplet surface.The emulsion droplets are found to co-exist with smaller particles of various shapes, ranging between 15 and 40 nm in diameter, likely protein aggregates.
The simulated intestinal digestion followed the gastric digestion of the emulsion.To further study the effect of compromised bile salt concentrations on digestion, the intestinal digestion was performed in the presence and the absence of bile/phospholipid micelles.The pH-stat method was used to analyze the digestion reaction by measuring the required NaOH concentration to maintain the pH at 7.0 throughout the digestion (Figure S9, Supporting Information).This amount is proportional to the released FFAs over time during TAG digestion.It showed that the simulated intestinal digestion of the buriti oil emulsion released around 1.5x more FFA in the presence of the bile/phospholipid micelles compared to the digestion in their absence.
Figure 4a shows the SAXS curves from the simulated intestinal digestion of buriti oil emulsion in the presence of bile/phospholipid micelles.A peak, most likely from the lamellar Ca-fatty acid complexes, is observed in the scattering curve around q = 1.36 nm −1 , corresponding to d = 4.6 nm, similar to the oral phase.This peak does not change position during the 120 minutes of intestinal digestion, indicating that the interlamellar spacing remains constant.Over time, the shape of the SAXS curves gradually changed, with a decrease in the scattering intensity at lower q values and a modification of the power-law scattering regime, observed at q < 0.8 nm −1 (Figure S10, Supporting Information).While a power-law exponent of −3 was found at the beginning of the intestinal digestion, characteristic of the interface scattering from emulsion droplets, it was ≈−2 towards the end, indicating the formation of vesicles. [45]Additional cryo-TEM analysis of the sample, shown in Figure 4b, confirms the transformation of large emulsion droplets to vesicles with various sizes during intestinal digestion.Small vesicles and elongated, coiled structures with a thickness under 10 nm can be observed in the images.
SAXS was also used to investigate the nanostructure of the buriti oil emulsion in the absence of bile/phospholipid micelles during simulated intestinal digestion (Figure 5a).Also in this sample, the reflection from the likely lamellar fatty acid soap structures was observed at q = 1.36 nm −1 .However, additional peaks appeared in the scattering curve after adding pancreatin extract during simulated intestinal digestion under compromised conditions.At 10 min, a reflection at q = 1.32 nm −1 is observed and gradually becomes more intense and sharper over time.After 10 more minutes, very weak reflections at q = 2.27 nm −1 and 2.61 nm −1 are seen.Together, these three peaks at relative q-positions of the peaks at 1, √3, and √4 (Miller indices 10, 11, and 20, respectively) indicate the presence of an inverse hexagonal (H 2 ) structure.Another set of peaks was also observed after 10 minutes of digestion at q = 1.15 nm −1 and 1.35 nm −1 .As time evolves and more peaks are seen, these peaks are attributed to an inverse micellar discontinuous cubic phase (Fd3m t) based on the relative peak positions (√3, √8, √11, √12, √16, √19).Averaging the last minutes of digestion where no further changes were observed (Figure 6a) improved the signal/noise ratio, allowing for a more precise assignment of the peaks.The changes in the lattice constant of the observed LLC structures during intestinal digestion are presented in Figure 6b.While the lattice constant of the lamellar phase remained stable at d = 4.6 nm over 120 min, the lattice constant of the H 2 phase slightly increased from 5.5 nm (time = 10 min) to 5.8 nm (time = 120 min).Likewise, the lattice constant of the Fd3m phase increased from 15.5 nm (time = 10 min) to 16.3 nm (time = 120 min).Cryo-TEM imaging (Figure 5b) at the end of this intestinal digestion indicated the presence of a few vesicles with sizes ≈50-60 nm, coexisting with particles with a diameter between 30 and 60 nm that are internally structured.Although the resolution limit of the method limits the accurate identification of the type of internal structure, they are likely to possess hexagonal or cubic structures based on SAXS data (Figure 6a).
Varying the pH of the sample after the end of the simulated intestinal digestion step (without bile/phospholipid micelles) allowed for different structures to be observed (Figure S11, Supporting Information).At pH 6.5, the H 2 phase was no longer observed, and the Fd3m (lattice constant 14.7 nm) became the dominating structure alongside the L  phase, which was present in all curves.At pH 7.5, the reflections of the Fd3m and the H 2 structures became more pronounced, with their lattice constants increasing to 18.2 nm and 6.6 nm, respectively.A swollen bicontinuous cubic phase with Im3m symmetry was also observed at lower q values, with a calculated lattice parameter of 31.5 nm.Finally, increasing the pH to 8.0 led to the disappearance of the Fd3m and H 2 phases.At this pH, the peaks associated with the Im3m phase became broad, indicating decreasing long-range order and transformation into a sponge phase (L 3 ). [46]he changes in the effective D H and PDI of the buriti oil emulsion at different stages of simulated digestion were studied by DLS (Figure 7) in the presence and absence of bile salts/phospholipids in the intestinal phase.The initial D H of the emulsions does not change significantly at the oral phase or the beginning of gastric digestion at pH 3.0.However, at the end of the gastric digestion step, the D H increased considerably to 650-700 nm (PDI ≈0.3-0.4).At the start of the intestinal digestion at pH 7.0, the D H decreased to values between 250 and 310 nm (PDI ≈0.5-0.6).Adding porcine pancreatin extract resulted in a slight increase in the D H to 300 nm in the sample without bile salt/phospholipid mixed micelles.No modification in the particle size was observed in the presence of the bile salt/phospholipid mixed micelles.This may result from the solubilization of pancreatic extract components in the bile salt micelles.At the end of the simulated intestinal digestion, the sample with bile salts presented a D H of roughly 110 nm (PDI = 0.86).The sample without bile salts exhibited a D H of 170 nm (PDI = 0.47).The higher PDI in the samples with bile salts may correspond to the multiple structures coexisting at this stage (e.g., LLC phases, vesicles, mixed micelles, and emulsion droplets); see Figure S12 (Supporting Information).

Effect of Gastric and Intestinal Conditions
To analyze the effect of gastric digestion on forming the selfassembly structures in intestinal digestion, the simulated intesti- nal digestion of buriti oil emulsion was performed with and without prior gastric digestion (Figure 8).The intestinal digestions were carried out at different pH values (6.5, 7.0, and 7.5).At pH 6.5, only the lamellar phase was observed regardless of the inclusion of the gastric digestion step.At pH 7.0, skipping gastric digestion yielded more intense LLCs reflections during simulated intestinal digestion.Swelling of the LLC structures was observed when the gastric step was skipped, increasing the lattice constant for the H 2 phase from 5.8 nm to 6.0 nm and the Fd3m phase from 16.3 nm to 16.8 nm compared to the complete digestion.Furthermore, at pH 7.0, a second H 2 phase with a lattice constant of 5.9 nm was observed during intestinal digestion without prior gastric digestion (purple arrows in Figure 8b).At pH 7.5, a possible cubic Im3m phase (lattice constant = 24.0nm) was observed at the end of the complete digestion, although the peaks' low intensity made the precise phase assignment difficult.The defined Im3m reflections were not detectable without the gastric digestion step.Instead, a broad peak was observed at similar q values of the cubic phase, which may indicate a loss of order with a transition to a sponge phase in the absence of gastric pre-digestion.
Regarding digestion kinetics, Figure S13a (Supporting Information) shows that skipping the gastric digestion resulted in a release of ≈1.5× more fatty acids during the 120 min of intestinal digestion at pH 7.0 than the complete digestion.This suggests faster TAG hydrolysis when excluding the gastric step.The timeresolved SAXS curves of the intestinal digestion at pH 7.0 (Figure S13b, Supporting Information) revealed that after 10 minutes of intestinal digestion, the peaks of the LLC phases are sharper and more pronounced when skipping gastric digestion.

Role of Stabilizer During Simulated Digestions
The influence of the emulsion stabilizer on the formation of the self-assembled structures and simulated intestinal digestions was also performed on a buriti oil emulsion using polysorbate 80, a synthetic food-grade nonionic surfactant (Figure S14, Supporting Information).Differences occurred in the swelling behavior of the LLC structures over intestinal digestion (Figure S15c, Supporting Information).The lattice constant of the H 2 phase increased for the first 40 minutes (from 6.2 to 6.4 nm) and then slightly decreased to 6.3 nm during the remaining 80 minutes of digestion for the polysorbate 80-stabilized emulsion.For the whey protein-stabilized emulsion, the lattice constant of the H 2 phase showed a more constant increase (from 5.8 to 6.0 nm) throughout the intestinal digestion.At the end of the intestinal digestion (Figure S14b, Supporting Information), the lattice constants of the structures in the polysorbate 80 stabilized emulsions were 17.4 nm for the cubic Fd3m, 6.3 nm for the H 2 phase, and 6.0 nm for the second H 2 phase.In comparison, values found for the whey protein-stabilized emulsion were 16.8, 6.0, and 5.9 nm, respectively, see Figure 8b.This showed a swelling of the self-assembled structures when polysorbate 80 was used as a stabilizer instead of whey protein.The D H of the polysorbate 80-stabilized buriti oil emulsion was 350 nm (PDI = 0.48), see Figure S15a (Supporting Information), similar to the whey protein-stabilized emulsions (Figure S4b, Supporting Information).Overall, the digestion of polysorbate 80 stabilized emulsions resulted in a lower FFA release than whey protein stabilized emulsions during 120 min of simulated intestinal digestion at pH 7.0 (Figure S15b, Supporting Information).

Influence of Vitamin E in the Nanostructure of Digested Emulsions
In vitro digestion with SAXS was used to investigate the impact of the model-hydrophobic nutrient vitamin E on the colloidal structures during digestion.Figure 9 shows the final structures obtained after intestinal digestion without bile salts.Gastric digestion was skipped to minimize the presence of surface-active compounds (fatty acids, monoglycerides) that can solubilize vitamin E. Emulsions containing increasing amounts of vitamin E/buriti oil of 0:100; 2:98; 10:90, and 30:70 w/w were investigated.
The addition of vitamin E/buriti oil at 2:98 resulted in the disappearance of the H 2 phase and, simultaneously, an increase in the intensity of the discontinuous cubic Fd3m phase peaks in Figure 9. Additionally, there was a decrease in the lattice constant of the Fd3m phase from 16.8 nm (without vitamin E) to 16.4 nm.Increasing the vitamin E/buriti to 10:90 further reduced the Fd3m lattice constant to 15.5 nm.At the highest vitamin E/buriti oil ratio of 30:70, sharp reflections of LLC are absent.A broad correlation peak at q around 1.3 nm −1 is present, likely from inverse micellar structures inside the digested oil droplet.
The particle size and stability of vitamin E/buriti oil emulsions were also studied to investigate the effect of vitamin E on the stability of the emulsions.The DLS autocorrelation functions (Figure S5, Supporting Information) exhibit a single decay throughout the monitored storage period of 10 days, suggesting a rather monomodal particle size population for all vitamin E/buriti oil ratios.The resulting D H values were ≈340-360 nm (Figure S4a, Supporting Information).The exception was the emulsion containing a vitamin E/buriti oil ratio of 30:70, which presented a D H of ≈420 nm.Over a storage time of 10 days at room temperature, no significant changes in particle size occurred.There was a slight increase of the PDI over 10 days from 0.32 to 0.39 (vitamin E/buriti oil 2:98), from 0.36 to 0.45 (vitamin E/buriti oil o 10:30), and from 0.34 to 0.54 (vitamin E/buriti oil 30:70).This suggests that the emulsions' size distribution was broadening over time, while the center of the distribution (i.e., the reported D H values) remained relatively constant.Images of the emulsions (Figure S16, Supporting Information) also displayed no visual evidence of phase separation in this 10day storage period.

Buriti Oil Composition and Emulsification
The fatty acid profile of buriti oil (Figure S1, Supporting Information) shows a predominance of unsaturated fatty acids (79.9% of total fatty acids) and is in close agreement with values reported by other authors (78.6%). [47]The higher ratio of monounsaturated to polyunsaturated fatty acids in buriti oil has benefits regarding stability toward oxidation. [48]Another essential feature to protect vegetable oil from oxidation is the presence of antioxidants such as vitamin E, [49] which is present in a high concentration in buriti oil compared to other common vegetable oils (Figure S2, Supporting Information).The 1750 mg kg −1 of total vitamin E content found in the buriti oil in this work is within the range reported by other authors when investigating buriti oil from different sources (artisanal/industrial), in which the vitamin E content ranged from 580 to 2017 mg kg −1 . [50]Likewise, the total carotenes (301 mg kg −1 , Table S1, Supporting Information) is also within the range of the different buriti oil sources, reported in the literature from 252 to 1890 mg kg −1 . [50]hey-protein stabilized buriti oil emulsions were engineered using high-energy emulsification by ultrasonication.The whey protein kinetically stabilized the buriti oil emulsions (Figures S3 and S4, Supporting Information).Given the negative surface charge of the whey protein-stabilized emulsions above the isoelectric point of the protein (pH ≈ 4.8), [51] the electrostatic repulsion between particles inhibited aggregation in the 10-day-long monitored periods.

Digestion of Buriti Oil Emulsions in the Absence and Presence of Bile Salt
For the first time, buriti oil emulsions were digested using an in vitro digestion model combining oral, gastric, and intestinal steps.During the oral stage, SAXS (Figure S6, Supporting Information), cryo-TEM (Figure S7, Supporting Information), and DLS (Figure 7) do not show significant changes.No starch is in the formulation to digest, and the effects from the change in the pH and the ionic strength of the simulated oral fluids compared to the initial emulsion appear negligible.No major changes are observed in the gastric digestion step from SAXS analysis (Figure 3).Only ≈10-30% of the lipolysis occurs in the gastric phase, with the remaining occurring in the intestinal step. [16]he gastric lipase converts TAGs into diacylglycerols and free fatty acids rather than monoacylglycerols. [16]Diacylglycerols are less surface active than monoacylglycerols and were not found to self-assemble into ordered structures.The FFAs are primarily in their protonated form in the simulated gastric digestion's acidic (pH = 3.0) conditions.Self-assembly of oleic acid, the most abundant type of fatty acid in buriti oil, was found at pH values above 6.8. [25]Hence, conditions for the lipid digestion products to form sophisticated self-assembled structures are mostly absent in the gastric environment.However, the low pH and presence of proteases in gastric digestion appear to modify whey protein's ability to stabilize the buriti oil in water emulsions.Whey proteinstabilized emulsions usually present a positive charge below its isoelectric point (pH ≈ 4.8), switching to a negative charge above it. [51]The formation of nanoscale structures in solution and on the surface of the emulsion droplets, most likely from whey protein digestion and aggregation, were seen in the cryo-TEM image, Figure 3b.This leads to decreased emulsion stability, increasing the D H from under 300 nm to ≈650-700 nm at the end of gastric digestion from DLS (Figure 7).Some studies reported flocculation of whey protein-stabilized emulsions during simulated gastric digestion. [52]Nevertheless, this droplet aggregation seemed to be reversible upon pH increase since the particle size is restored to values closer to the initial emulsion size once the digestion goes to the intestinal phase, in which pH is increased to 7.0, see Figure 7.This could be due to released FFAs during gastric digestion becoming much more deprotonated at pH 7.0 (and thus surface active), helping to disperse and stabilize the emulsion together upon stirring.Deprotonated oleic acid can significantly reduce the oil/water interfacial tension, allowing emulsification by simple stirring without needing high-energy methods. [26]The switch of the remaining whey protein from positively to negatively charged once pH changes from 3.0 to 7.0 could also contribute to the FFAs to stabilize the emulsion.
In the subsequent intestinal digestion, the colloidal structure formation in buriti oil emulsions was analyzed in the absence and the presence of a simulated bile solution containing bile salt/phospholipid micelles.The dynamic composition of the system triggers structure formation and transformation.The bile/phospholipid micelles significantly influenced this process.Lamellar and non-lamellar LLC structures are observed without the bile salt/phospholipid micelles.The formation of the lamellar phase attributed to calcium soaps in the presence and absence of the bile salt/phospholipid micelles is consistent with the digestion of other lipid-rich materials such human milk [28] and bovine milk. [29]The calcium soaps' lattice constant depends on the fatty acids' chain length.The determined lattice constant of 4.6 nm for the L  phase during the digestion of buriti oil emulsion closely matches the observed value of 4.54 nm reported for calcium palmitate (C16:0). [53]Indeed, palmitic acid is the second most abundant type of fatty acids in the composition of buriti oil (Figure S1, Supporting Information).Calcium soap formation occurs to a greater extent with saturated fatty acids than unsaturated ones, [54] which would explain why we observe mainly calcium palmitate in these digestions despite oleic acid being the most abundant fatty acid of buriti oil.Calcium presence has opposing effects on the digestion of lipids.While it promotes the lipolysis rate by precipitating FFAs from the lipid/water interface and acting as a cofactor for pancreatic lipase, it can also slow down lipid digestion by destabilizing emulsions via flocculation and aggregation. [55]he formation of non-lamellar LLC structures inside the oil droplet in the intestinal digestion without the bile salt/phospholipid micelles may influence the bioavailability and controlled release of nutrients.Besides SAXS, cryo-TEM also supports the presence of these LLC structures based on the internally structured particles observed in Figure 5b.Their smaller size, with diameters ≈30-60 nm, as seen in the cryo-TEM images, can be connected with the observed decrease in D H from the beginning (300 nm) to the end (170 nm) of intestinal digestion as examined by DLS (Figure 7).LLC structures were reported to solubilize hydrophobic, amphiphilic, and hydrophilic nutrients. [56]While in the simulated intestinal digestion of bovine or human milk, some formed LLC phases tend to be transient, [28][29] in the case of buriti oil emulsions at pH 7.0, the formed phases remain present throughout the monitored period.Comparing the intensity of the peaks between the H 2 and Fd3m phases, the H 2 phase appeared to be the dominant structure.In terms of its lattice constant, the dimensions of the hexagonal phase observed during the digestion of buriti oil (≈5.5 nm at pH 7.0, Figure 6b) were smaller than those observed for human milk (≈6.5 nm at pH 7.0), [28] emulsified krill oil (around 7.2 nm at pH 6.5) [31] or mayonnaise (around 6.0 nm at pH 7.0). [32]The differences related to types and sizes of formed mesophases between different food emulsions lie mostly on the compositional specificities of each food, since the lipid's molecular structure (e.g., chain length and degree of unsaturation), as well as the presence of other components and nutrients, are known to modulate lipid self-assemblies. [57]This was observed in digestion studies comparing emulsions of lipids with different chain length [27] or with different headgroups and degrees of unsaturation. [31]hile the lattice constant of the lamellar phase was constant throughout the simulated intestinal digestion, changes are observed in the lattice constants of the Fd3m and H 2 phases (Figure 6b).The gradual swelling of these mesophases is likely due to changes in the proportion of different lipolysis products during digestion.The difference in the ionization state of the carboxyl group of free fatty acids released during digestion from pH changes affects the effective headgroup area at the interface.The transition Fd3m to H 2, followed by Im3m and L 3 , as pH increases from 6.5 to 8.0 (Figure S11, Supporting Information), aligns with the expectations from the CPP theory. [36]With the carboxyl group becoming gradually more deprotonated upon pH increase, the effective headgroup area at the interface increases owing to the repulsion among the negative charges there.This leads to a gradual decrease in the CPP value, favoring the formation of more polar interfaces.Hence, the gradual increase in the GI tract's pH may trigger colloidal structure formation and transformation during lipolysis.Differences in the self-assembled structures are observed when comparing the full digestions with the "intestinal only" digestions (Figure 8).The changes in the LLC structures suggest that the composition of the emulsion and its interface may be altered in terms of the amounts and types of generated lipolysis products.Thus, simplifying digestion models by removing the gastric step must be taken cautiously.Interestingly, performing the intestinal digestion step at pH 7.0 and then changing the pH to 6.5 or 7.5 (Figure S11, Supporting Information) at the end of the 2 h of simulated digestion results in different structures than performing the intestinal digestion directly at pH 6.5 or 7.5 (Figure 8a).This may arise from the influence of the pH on the enzyme activity and selectivity, which could lead to changes in the amount and the proportion of different lipolysis products at the end of the digestion at each pH value. [58]The Im3m bicontinuous cubic phase (Figure 8a) was not observed at pH 7.0 but only at higher pH (7.5) with a lattice constant of 24 nm.Interestingly, the Im3m phase was also reported for the digestion of human breast milk at pH 7.0 and with a comparable lattice constant of 20.5-23.0 nm. [28]Overall, the effect of the pH of the surrounding solution on the structural features of the emulsions implies that structural transformations occur in the distinct segments of the small intestine, from the duodenum (pH ≈ 6.6) to the terminal ileum (pH ≈ 7.5). [59]Generally, the lipid/water interfaces become more hydrophilic with increasing pH.This may tune the release of loaded molecules and trigger interactions with the digestive tract.
Furthermore, a faster release of fatty acids during intestinal digestion is observed when the gastric digestion is skipped (Figure S13a, Supporting Information).This is a result of the modification of the emulsion interface during gastric digestion.TAG digestion products from gastric digestion can slow down digestion by compromising the adsorption of the pancreatic lipase. [60]urther, the gastric digestion of the whey protein also modifies the oil-water interface, which can further compromise lipase adsorption, and hence its access to the substrate.
Replacing whey protein with polysorbate 80 as a stabilizer for the buriti oil emulsions resulted in some changes.Despite the emulsions stabilized by both stabilizers having the same initial hydrodynamic diameter, whey protein-stabilized emulsion presented a faster intestinal digestion rate than polysorbate 80 (Figure S15b, Supporting Information).This agrees with other authors who also compared intestinal digestion for sunflower oil emulsions stabilized by whey protein and polysorbate 80, pointing out that the faster digestion with whey protein is likely due to easier adsorption of intestinal lipase to the emulsion interface. [61]owever, the LLC formation inside the emulsion droplet during digestion was comparable in terms of structure type and sequence (Figure S14, Supporting Information).The structures appeared more swollen when polysorbate 80 was employed (Figure S15c).Such swelling of mesophases by polysorbate 80 can be attributed to its incorporation in lipid bilayers, producing changes in the lipid/water interface curvature. [62]n the presence of bile/phospholipid micelles, highly ordered LLC structures were absent in SAXS (Figure 4a).The TAG digestion products and the bile/phospholipid mixture selfassemble into the nanoobjects seen in the corresponding cryo-TEM image (Figure 4b).[65] This also underlines the natural role of bile salts in sequestering interface-active lipolysis products from the interface, facilitating lipase binding to the oil/water interface. [18,66]his finding supports our hypothesis that the LLC structures help maintain digestion and nutrient uptake under compromised bile salt conditions, such as in patients with digestive disorders. [24]The lower extent of intestinal digestion in the absence of bile/phospholipid micelles (Figure S9, Supporting Information) may thus result from the increased difficulty for lipase to access the oil/water interface in the presence of interface-active lipolysis products. [60,66]Using more complex digestion models, in which the bile salt concentration increases progressively over time, could lead to different results in lipolysis kinetics and the types of formed self-assemblies. [67]The significant findings are summarized in Figure 10.

Vitamin E Enrichment of Buriti Oil Emulsions
The buriti oil emulsions were further enriched with vitamin E to analyze its impact on emulsion stability and digestion-triggered structure formation.We use vitamin E, also a model for hydrophobic nutrients, to evaluate bioaccessibility by incorporation/solubilization in the oil or colloidal structures.In individuals with BS deficiencies, LLC phases could compensate them, playing the role of nanocarriers for liposoluble nutrients. [34]Since the natural concentration of vitamin E in buriti oil was determined as 1750 mg kg −1 , which corresponds to 0.175% w/w (Figure S2, Supporting Information), larger amounts of vitamin E (vitamin E/buriti oil ratios of 2:98, 10:90, and 30:70 w/w) were added to the formulation to have a clearer understanding on the effect of loading hydrophobic nutrients at different vitamin/oil ratios.Adding moderate amounts of vitamin E (≤ 10% w/w) did not significantly affect the particle size, as shown in Figure S4a.However, at the highest ratio of vitamin E to buriti oil (30:70 w/w), there was a noticeable increase in particle size from 340-360 nm to over 420 nm.This suggests that large quantities of vitamin E may destabilize the emulsions.Additionally, when a high amount of vitamin E (30% w/w) was used, the polydispersity index (PDI) increased over time (Figure S4b, Supporting Information).The PDI rose from 0.34 on day 0 to 0.54 on day 10.These findings align with previous studies indicating that vitamin E reduces emulsion stability, raises the interfacial tension between oil and water, and increases the particle size in whey protein-stabilized flax seed oil emulsions. [68]Moreover, an elevation in polydispersity can result in diminished colloidal stability owing to Ostwald ripening from the greater difference in chemical potential between droplets. [69]he effect of vitamin E solubilization on nanostructural transformations inside the emulsion droplet during in vitro digestion was analyzed using SAXS.The final digestion structure inside the emulsion droplets, presented in Figure 9, shifts from H 2 to Fd3m and, ultimately, a possible inverse micellar phase upon increasing the vitamin E content.This structural transformation agrees well with the CPP theory: the hydrophobic vitamin E increases the CPP value by solubilizing into the hydrophobic regions of the structures and increasing the overall tail volume there.[72][73] Thus, the addition of vitamin E tunes the type and dimensions of the LLC structures during the digestion of the buriti oil emulsions.Given that buriti oil is naturally rich in vitamin E, the weaker Fd3m phase observed even in the absence of supplementary vitamin E (bottom curve in Figure 9) may be marking already the beginning of the transition of the H 2 to discontinuous cubic mesophase due to this innate amount of vitamin E in the oil.In food materials research, tocopherols are generally added in the range of 0.001% to 2% w/w to enhance antioxidant properties. [74]Meanwhile, our system was capable of preserving the formation of LLC phases during digestion (and overall colloidal stability) at much higher vitamin E loadings, up to 10% w/w vitamin E in buriti oil.

Conclusion
Whey-protein stabilized buriti oil-in-water emulsions were engineered, and their colloidal behavior during simulated digestion was analyzed with a novel digestion model with in situ analysis of changes in composition and structure.The digestion model combines oral, gastric, and intestinal digestion with in situ synchrotron SAXS.Additionally, cryogenic electron microscopy and dynamic light scattering were employed to complement the analysis.We found that the formulations remained stable during the oral stage, and reversible aggregation occurred during the gastric digestion step.SAXS showed mostly normal emulsions without significant scattering features in the nanometer region.The formation of nanostructures was then observed during the simulated intestinal digestion step: in the presence of bile salts, a gradual shift from normal emulsions to vesicles was discovered.In the absence of bile salts, the formation and transformation of highly geometrically ordered LLC structures were observed.Over the course of digestion, the normal emulsions from the gastric phase transformed into internally nanostructured particles with, for instance, inverse hexagonal and discontinuous cubic and lamellar structures.The nanostructures were pH-responsive, with structural changes in agreement with the critical packing parameter theory.
The supplementation of buriti oil emulsions with vitamin E significantly influences the self-assembly structure formation inside the oil droplets during intestinal digestion.Increasing the vitamin E content leads to a gradual shift from the inverse hexagonal structure as the dominant phase, observed in the buriti oil (without supplemented vitamin E) digestion, toward the discontinuous cubic phase for the sample at vitamin E/buriti oil 2:98 and 10:90 ratios.From these results, whey-stabilized buriti oil emulsions were found to combine appealing aspects such as the high content of unsaturated fatty acids in its composition, colloidal stability, and the potential to maintain and trigger the formation of non-lamellar LLC phases during (simulated) intestinal digestion.The possibility to tune these LLC structures during digestion is an essential feature toward the development of functional foods since the loading and release rate of drugs and nutrients is connected to the type of the LLC phase.
The results from this study mark a step toward establishing the connection between composition and in situ structure formation during the digestion of food emulsions.The findings can guide the development of advanced, innovative food and pharmaceutical materials.
Fatty Acid Profile of Buriti Oil: Buriti oil (≈20 mg) was weighed.Sodium methoxide solution (1 mL) was added, mixed by vortex, and heated at 50 °C for 10 min (the mixture was shaken at three-minute intervals).Upon cooling, sodium bisulfate (15%) (1 mL) was added and mixed by vortex.Heptane (3 mL) was added and agitated on an electronic mixer for 5 min.The organic phase was washed with water (1 mL), and then an aliquot of the organic phase (1 mL) was dried over sodium sulfate, mixed by vortex, and centrifuged (5 min; 16 000 × g).A volume of 1 μL was injected in split mode (35:1) in an Intuvo 9000 gas chromatographic (GC) system (Agilent, Santa Clara, USA) with a CP-Sil 88 for FAME column, 60 m long, 0.25 mm inner diameter, 0.2 μm column operated with hydrogen as carrier gas at a flow rate of 1.5 mL min −1 for a total run time of 20 min.A flame ionization detector (FID) performed fatty acid detection at 300 °C.Identification was made according to relative retention time and quantification based on peak area.
Vitamin E and Carotenoid Content of Buriti Oil: For Vitamin E quantification, 1 g of buriti oil was weighed and mixed with 3 mL of hexane in a volumetric flask.The volume was then adjusted to 25 mL with ethanol.The mixture was vigorously agitated by hand and decanted for 3 minutes to allow phase separation.The upper organic phase (10 μL) was injected on a Flexar liquid chromatographic (LC) system (Perkin Elmer, Shelton, USA) operated at 30 °C with Nucleodur C18 ec column (Macherey-Nagel, Düren, Germany) at a flow rate of 1 mL min −1 of 0.5% water in methanol (v/v) in isocratic mode and UV detection at 292 nm.The relative amounts of -Tocopherol, +-Tocopherol, and -Tocopherol were calculated from the respective peak areas and corrected by appropriate factors.
For carotenoids quantification, 1 g of buriti oil was weighed and mixed with 50 mg of butylated hydroxytoluene (BHT) in a volumetric flask.Cumulative volumes of dichloromethane (24 mL) and ethanol (20 mL) were added and vigorously agitated by hand for 3 min.The mixture was allowed to stand for 2 h, protected from light.The total volume was adjusted to 50 mL with dichloromethane and remixed.A volume of 25 μL was injected on the same LC system as above, operated at 30 °C with Suplex PKB-100 column (Supelco, Bellefonte, USA) at a flow rate of 0.9 mL min −1 of dedicated solvent (50 mg BHT in 20 mL of propanol, 0.2 mL of Nethyldiisopropylamine, 25 mL of 0.2% ammonium acetate, 455 mL of acetonitrile, and 500 mL of methanol) and UV detection set at 448 nm.The total amount of carotenes and lutein was calculated from their respective peak areas.
Emulsion Preparation: Buriti oil was dispersed in ultra-pure water containing dissolved whey protein by sonication with a microtip using a Lab500 NexTgen Ultrasonic platform (SinapTec, Lezennes, France) at 25% amplitude (max.power 500 W) during 40 cycles (3s ON, 3s OFF each cycle).The oil fraction in the emulsion was set at 20% w/w and the stabilizer (whey protein) at 2% w/w to maintain a stabilizer/oil ratio of 1:10 w/w.Emulsions were stored at room temperature and used within 24 h of preparation.For the emulsions containing a mixture of buriti oil and dl--Tocopherol acetate (vitamin E), the total emulsified content (buriti + vitamin E) were also set at 20% w/w relative to water.For the samples used for the long-term storage stability test at room temperature, 0.02% w/w sodium azide was added as a preservative.For the emulsion prepared with polysorbate 80 instead of whey protein, the polysorbate 80 concentration was maintained at 2% w/w and stabilizer/oil ratio at 1:10 w/w.
Preparation of Simulated Digestive Fluids and Bile Salt Solution: Simulated salivary, gastric, and intestinal fluids (SSF, SGF, and SIF, respectively) were prepared by mixing HCl, NaCl, CaCl 2 , KH 2 PO 4 , MgCl 2 .(H 2 O) 6 , and (NH 4 ) 2 CO 3 in concentrations specified by the INFOGEST 2.0 protocol [23] for each digestive fluid.Using HCl or NaOH solutions, pH was adjusted (SSF to pH 7.0, SGF to pH 3.0, and SIF to pH 7.0).In all simulated digestive fluids, NaHCO 3 was replaced by NaCl (in amounts to keep ionic strength unchanged) as recommended by the INFOGEST protocol in cases of pH drift issues.A bile salt solution (bile/phospholipid micelles) was prepared by combining NaTDC and DOPC at a 4:1 molar ratio.][77] DOPC was dissolved in a small amount of chloroform, which was removed via rotary evaporation under vacuum (4 h, 200 mbar, 40 °C).Afterward, NaTDC was dissolved in SIF, and this solution was added to the dried phospholipid film.The mixture was ultrasonicated with an ultrasonic bath (Emmi H-60, EMAG GmbH, Salach, Germany) for 30 min to yield a bile salt stock solution containing 160 mm NaTDC and 40 mm DOPC.Simulated digestive fluids and bile salt solution were stored at 4 °C until use.
Enzyme Solution Preparation: Rabbit gastric extract was used as a model for human gastric enzymes, and pancreatin from the porcine pancreas as a surrogate for human intestinal enzymes.The enzymes were assayed to determine their lipolytic activity, and the concentrations used were adjusted to appropriate levels.The gastric enzyme solution was prepared by gently stirring rabbit gastric extract in cold water (40 mg in 0.25 mL to achieve 60 U mL −1 of gastric digestion volume) for 1 min.The intestinal enzyme solution was prepared by gently mixing pancreatin in cold SIF (1.8 g in 3.8 mL) followed by centrifugation (10 100 RCF, 15 min, 4 °C) and supernatant collection.Enzyme solutions were produced immediately before use.The lipase activity in the pancreatin porcine extract was evaluated based on published protocols, [23] with one TBU corresponding to the release of one μmol of butyric acid per min.0.5 mL of tributyrin was vortex-mixed with 14.5 mL digestion medium (4 mm NaTDC, 150 mm NaCl, and 1.4 mm CaCl 2 ) to form a fine oil-in-water emulsion.Then, 0.100 mL of pancreatin solution (10, 50, or 100 mg mL −1 ) was added, and the amount of 0.2 m NaOH solution was monitored to keep the pH of the solution at 8.0 for 5 min at 37 °C.Pancreatin solutions were prepared by gently mixing pancreatin in cold water followed by centrifuga-tion (10 100 RCF, 15 min, 4 °C) and collection of the supernatant.Enzyme solutions were produced immediately before use.Pancreatin activity was determined as 8.15 ± 0.03 TBU mg −1 (Figure S17, Supporting Information).
In Vitro Digestion Model: Oral Phase: In a 50 mL conical centrifuge tube, 2.5 g of the emulsion was mixed with 2 mL of SSF.pH was adjusted to 7.0, and water was added to reach a total volume of 5.0 mL.Amylase was not included in the oral phase due to the absence of starch in these formulations.The mixture was incubated while mixing for 2 min inside a 37 °C water bath.
Gastric Phase: The oral bolus was combined with 4 mL of SGF.The pH was adjusted to 3.0.The gastric enzyme solution was added (0.25 mL) via a syringe drive, and the volume inside the tube was completed to 10 mL with water.The enzyme activity in this step is ≥2000 U pepsin mL −1 and ≥60 U lipase mL −1 .The mixture was incubated while mixing for 2 h at 37 °C under gentle mixing on a magnetic stirrer.During the incubation, a pH-stat (Titrando 906 titrator, Metrohm AG, Herisau, Switzerland) was used to monitor and titrate the system with a NaOH 0.2 m solution to maintain the pH at 3.0.
Intestinal Phase: The gastric chyme was combined with 4.25 mL of SIF.The pH was adjusted to 7.0.For the intestinal digestion with bile salt solution, 1.25 mL of the bile salt solution was added (for a final NaTDC concentration of 10 mm).In contrast, 1.25 mL of additional SIF was added instead for the digestion without a bile salt solution.Water was then added to reach a total volume of 17.5 mL.The intestinal enzyme solution was added (2.5 mL) via a syringe drive, resulting in a final volume of 20 mL.The intestinal lipase activity in this step was calculated as 482.6 TBU mL −1 of the final volume (20 mL).The mixture was incubated for 2 h at 37 °C under gentle mixing on a magnetic stirrer.During the incubation, the pHstat was set to keep pH constant at 7.0 by adding NaOH 0.2 m whenever necessary.
Small-Angle X-Ray Scattering: The structure generation during all digestion steps was analyzed dynamically using flow-through small-angle X-ray scattering (SAXS).Therefore, synchrotron SAXS was measured in situ at the Austrian SAXS beamline at Elettra (Trieste, Italy) and the cSAXS beamline at the Swiss Light Source, Paul Scherrer Institute (PSI, Villigen, Switzerland).A flow-through quartz capillary (1.5 mm thickness, Hilgenberg GmbH, Malsfeld, Germany) was mounted in the X-ray beam path.During the digestion reaction, the sample was continuously circulating through the capillary from the reaction vial via silicone tubing, using a peristaltic pump (Alitea MIDI-D U1, Alitea AB, Stockholm, Sweden).The flow rate was roughly 10 mL/min.The coupling of simulated digestion with SAXS is illustrated in Figure 2.
For measurements at the Austrian SAXS beamline at Elettra, an X-ray beam with a wavelength of 1.54 Å (energy of 8 keV) was used.The sampleto-detector distance was 906.18 mm, covering a q-range of 0.08-7.0nm −1 , where q is the length of the scattering vector, defined by q = 4 /  sin(/2),  being the wavelength and  the scattering angle.The 2D SAXS patterns were acquired for 10 s using a 2D Pilatus3×1 M detector (Dectris Ltd, Baden, Switzerland; active area 169×179 mm 2 with a pixel size of 172 × 172 μm 2 ), and the data was radially integrated into 1D curves and plotted as a function of scattering intensity I(q) versus q.For measurements at the Swiss Light Source at PSI, an X-ray beam with a wavelength of 1.0 Å (energy of 12.4 keV) was used.The sample-to-detector distance was 2164 mm, with a q-range of 0.03-6.8nm −1 .The 2D SAXS patterns were acquired using a Pilatus 2 M detector (in-house prototype) with an active area of 254×289 mm 2 and a pixel size of 172 × 172 μm 2 .Exposure time was 1 s with 9 s delay between the frames.
The 2D scattering frames were radially integrated into 1D curves and plotted as a function of scattering intensity I(q) versus q. Background subtraction was performed for all data points using the simulated digestion fluids as background.
SAXS Data Analysis: Lyotropic liquid crystalline structures were identified by the relative position of their Bragg peaks.The lattice constant (a) for each type of structure was calculated by the equations:

Figure 1 .
Figure 1.Schematic illustration of oil-continuous self-assemblies formed during digestion and the critical packing parameter model (CPP).The structure type and dimensions of the domains, typically in the < 30 nm range, are determined with SAXS.The CPP can be calculated from the volume of the hydrophobic tail (V), the effective headgroup area at the lipid-water interface (a 0 ), and the length of the hydrophobic tail (l c ).

Figure 2 .
Figure 2. Experimental setup for the simulated digestion studies within operando SAXS characterization.Oral, gastric, and intestinal digestion were performed in a reaction vessel, and structure formation was followed in real-time using in situ SAXS and pH-stat titration.Samples for cryo-TEM and DLS are taken at defined time-steps during digestion to complement the in situ SAXS data on structure formation during digestion.

Figure 3 .
Figure 3. a) Time-resolved in situ SAXS curves of simulated gastric digestion of whey-stabilized buriti oil emulsion at pH 3.0.b) Cryo-TEM image at the end of gastric digestion.Black arrows indicate emulsion droplets covered by smaller particles, while white arrows indicate smaller particles of various shapes and sizes, likely whey protein aggregates.

Figure 4 .
Figure 4. a) Time-resolved in situ SAXS curves of simulated intestinal digestion (pH 7.0) of buriti oil emulsion in the presence of bile/phospholipid micelles (the curve directly after lipase addition and after 120 min of digestion is compared in Figure S10, Supporting Information).b) Cryo-TEM image at the end of this intestinal digestion.Red arrows indicate vesicles, and blue arrows indicate elongated, coiled structures.

Figure 5 .
Figure 5. a) Time-resolved in situ SAXS curves of simulated intestinal digestion (pH 7.0) of buriti oil emulsion without bile/phospholipid micelles and artistic representations of the observed structures.b) Cryo-TEM image at the end of this intestinal digestion.Red arrows indicate vesicles, and green arrows indicate particles with internal structure.

Figure 6 .
Figure 6.a) Average of the SAXS curves between 105 and 120 min of simulated intestinal digestion of buriti oil emulsion (Figure 5a).Labels on top of arrows indicate (hkl) Miller indices of the peaks for the different LLC phases.b) The corresponding calculated lattice constants versus digestion time for the LLC structures.

Figure 7 .
Figure 7. a) D H and b) PDI from DLS at various stages of simulated digestion of buriti oil emulsion, both in the presence and absence of bile salts/phospholipids mixed micelles.a, Initial emulsion.b, End of the oral phase.c, Addition of simulated gastric medium and pH set to 3. d, Addition of gastric enzymes.e, End (120 min) of gastric digestion.f, Addition of simulated intestinal medium and pH set to 7.0.g, Addition of bile salts/phospholipids. h, Addition of intestinal enzymes.i, End (120 min) of intestinal digestion (n = 3).

Figure 8 .
Figure 8. SAXS curves for the final point of simulated intestinal digestion (without mixed bile/phospholipid micelles) of buriti oil emulsion using different pH values for the intestinal phase and a) performing complete digestion or b) only intestinal digestion (skipping gastric digestion step).Labels on top of arrows indicate Miller indices of the peaks for the different LLC phases.

Figure 9 .
Figure 9. SAXS curves at 120 min of simulated intestinal digestion (skipping gastric digestion step and without mixed bile/phospholipid micelles) of emulsions composed of buriti oil and vitamin E mixtures at different mass ratios.Labels on top of arrows indicate Miller indices of the peaks for the different LLC phases.

Figure 10 .
Figure 10.Scheme summarizing the proposed colloidal transformations during the digestion of buriti oil emulsion when transiting from oral to gastric and intestinal digestion.