A sample of Mauritia flexuosa (buriti) was deposited in the Graziela Barroso Herbarium at Federal University of Piauí (UFPI) (voucher specimen 30567). The sample (#A690444) was registered in SisGen (Sistema Nacional de Gestão do Patrimônio Genético e do Conhecimento Tradicional Associado, i.e. National System of Management of Genetic Heritage and Associated Traditional Knowledge, Glück Informatica, Rio de Janeiro, Brazil) (11) according to the Brazilian biodiversity legislation (Federal Law No. 13.123/2015) (12). Approximately 300 fruits were collected in Água Branca, Piauí, Brazil, in December 2014 (latitude: 5°54’50.5” S; longitude: 42°38’03.4” W) and taken to the Federal Institute of Education, Science and Technology of Piauí, Teresina, Brazil. Fruit selection, cleaning, separation, lyophilization, packaging, refrigeration and pulverization were carried out according to Pereira-Freire et al. (6). Briefly, all fruits were separated into pulp, peel and endocarp and frozen at −70 °C. For the lyophilization, fruits were placed on a stainless-steel tray of lyophilizer model L101 (Liotop, São Carlos, Brazil). Lyophilization conditions (temperature: 40 °C, vacuum pressure: <500 mmHg, lyophilization rate: 1m/h) were controlled during 72 h. Afterwards, the fruits were packaged in plastic bags under refrigeration at 4 °C before processing into powder using a rotor mill (0.08 mm; Retsch, Haan, Germany).

The yield before lyophilization (Y₁) of samples was obtained on the total fruit mass basis. The average mass (g) of fruits was measured on a semi-analytical scale, and longitudinal and transversal diameter was determined using a digital caliper with 0−150 mm capacity and resolution of 0.01 mm (Digmess, São Paulo, Brazil). Sampling for physical measurements: average fruit mass, transverse diameter (cm) and longitudinal diameter (cm) was performed by using 10% of samples, i.e. 30 units of fresh fruits (1.46 kg). All measurements were done in triplicate.

To calculate the yield of all samples after lyophilization (Y₂), the following equation was used:

where Y_i is the yield of samples before lyophilization (in natura) and Y_f is the yield of samples after lyophilization.

To measure pH, a bench potentiometer (model EEQ9002G-2; A. Científica, Santo André, Brazil) was used. Sugar content in °Brix was measured using a refractometer (DR500; New Instruments, Piracicaba, Brazil), total titratable acidity was determined by the volumetric method, water activity (a_w) (Novasina, Aqualab, São José dos Campos, Brazil) and moisture content were determined by drying in an oven at 105 °C to constant mass, ash content was measured by incineration in a muffle oven at 550 °C, protein content was measured by the Kjeldahl method, lipid content was measured by direct hexane-based extraction using the Soxhlet technique, and carbohydrates were calculated by difference method (13). The total dietary fiber, soluble and insoluble contents were quantified by the enzymatic-gravimetric method (14). All analyses were performed in triplicate.

The total energy value (E) was calculated based on the conversion factors for protein (4 kcal/g or 16 736 kJ), lipids (9 kcal/g or 37 156 kJ) and carbohydrates (4 kcal/g or 16 736 kJ), expressed in kilocalories per 100 g of dried sample (15), using the following equation:

The mineral composition of pulp, peel and endocarp was determined by energy dispersive X-ray fluorescence spectrometry using energy dispersive X-ray spectroscopy (EDX-720; Shimadzu, Kyoto, Japan). Parts were placed in their own sample support separately, sealed with thin polypropylene film to avoid extrusion of samples by activating the vacuum and analyzed (16).

Microstructure, morphology and surface evaluations of pulp, peel and endocarp powders were performed by scanning electron microscopy (SEM, SSX-550 Superscan; Shimadzu). The powders were fixed on metal support under vacuum and metallized with a thin gold film. Micrographs were obtained at different magnifications using acceleration voltages of 8 to 15 kV. The metal plate was covered with platinum (model K 550 metallizer; Emitech, Ashford, UK) and operated at 10 kV, providing a coverage of approx. 25 μm. Images were captured and scanned.

All analyses were performed in triplicate, and the results were analyzed by ANOVA followed by Tukey’s test considering a 5% significance level (p<0.05) using SPSS software v. 24.0 for Windows (24). All analyses were performed in triplicate.

In general, most buriti fruits had the typical ellipsoid-oblong shape, in a very similar way to the globular-oblong shape reported by Milanez et al. (1) and Lorenzi et al. (8). The coefficients of variation showed low uniformity of fruits in relation to the analyzed parameters (Table 1) given that cultivation is not planned, as occurs in other countries of South America. Seed and pulp mass, and consequently, yield (34.04 and 22.6%, respectively) were lower than those described by Milanez et al. (1). These results indicate that the production of fruits is strongly influenced by environmental factors such as temperature, solar radiation and rainfall (1). Consequently, such factors alter the yield of fresh and processed fruits.

After lyophilization, peels (26.0%) and endocarp (20.72%) had higher yields (Table 1) and lower a_w (0.25 and 0.38; Table 2) than pulp (p<0.05). It is noteworthy that pulp yield is a very important quality parameter for the production of dehydrated products containing 15-20% moisture, caramels, honey and candies, since fruits with high pulp yield have a higher yield of product after processing, a clear criterion of productivity. Moreover, peels and endocarp have too low a_w values for microbial growth/proliferation, an excellent feature for future uses in food industries and fabrication of cookies, crackers, breakfast cereals, dry pet food, peanut butter, whole milk powder, dried vegetables, corn starch and potato chips (25).

Lyophilized pulp had higher nutritional contents of ash (i.e. minerals) (2.18±0.02) mg/100 g, protein (5.6±0.2) %, carbohydrates (35.7±0.5) mg/100 g, lipids (51.67±0.09) % and energy (630.3±0.7) kcal/100 g than in natura pulp ((2.27±0.05) mg/100 g, (2.47±0.07) %, (15.1±0.3) mg/100 g, (26.6±0.3) % and (310±4) kcal/100 g, respectively, p<0.05). This is obviously related to the reduction in water content after lyophilization. Similarly, lyophilized peels and endocarp were richer in minerals, protein, carbohydrates and energy, but poorer in lipids (p<0.05) (Table 2). These outcomes indicate that pulp is a good source of lipids. Indeed, buriti pulp oil has carotenoids, tocopherols and monounsaturated fatty acids, which makes it a valuable product with functional potentialities (4-6).

Previous investigations have demonstrated that buriti fruits are sources of carbohydrates, whose amount is controlled by the levels of biosynthetic enzymes and gene expression (26). In a similar way, lyophilized peel samples had higher fiber mass fraction ((50.5±0.6) %), higher content of insoluble ((50±1) %) and lower values of soluble fiber ((0.55±0.03) %) than fresh pulp samples ((38.0±0.3), (27.3±0.4) and (10.6±0.3) %, respectively, p<0.05). This can be considered an advantage since insoluble fiber is widely used to increase the content of compounds added to foods to improve the rheological properties of dietary products, increasing satiety, and the volume and faecal mass, which clearly improves digestive system performance (27).

Buriti pulp, peels and endocarp had lower contents of soluble fiber than the by-products from other fruits such as mango (28.2%), passion fruit (35.5%) and guava (11.1%) (28). However, lyophilized buriti peel had a greater quantity of insoluble fiber than mango (41.5%) and passion fruit (46%) by-products (28). Soluble fiber is advantageous because when incorporated into food, it increases viscosity and the ability to form gels and/or act as emulsifier. In addition, consumption of soluble fiber-rich foods may reduce blood glucose and cholesterol levels (29). In this context, 30 g of pulp, peel or endocarp samples can be used to provide more than 15% of the dietary reference intake (DRI) (30), which recommends consumption of 21 to 38 g/day dietary fiber, taking into consideration different human factors, such as age, physiological condition and sex.

It is worth noting that the consumption of fresh buriti fruits has limitations, mainly due to high water loss and susceptibility to chilling injuries when stored under refrigeration (31). Therefore, processing buriti fruits, such as by lyophilization, and its use for the development of new comestible products is effective for nutrient conservation, increased production, shelf-life extension and availability independent of season (32). Therefore, buriti pulp used for processing should derive from by-products if their chemical and physicochemical characterization is acceptable, which supports a sustainable destination and nutritional applications.

Lyophilized pulp had higher mass fraction of potassium (712.0±0.4) and chlorine (72.0±0.2) mg/100 g, peels had more phosphorus (26.0±0.1) and iron (19.00±0.02) mg/100 g, and copper (1.00±0.01) µg/100 g, and endocarp was richer in potassium (713.0±0.3) and calcium (159.0±0.1) mg/100 g, and copper (1.00±0.01) µg/100 g than fresh pulp (Table 3, p<0.05). Interestingly, chromium ((4.00±0.01) µg/100 g) was detected only in lyophilized peels. In general, the reductions of calcium, magnesium and manganese ranging from 18.5 to 22.7% were observed following the lyophilization.

Lyophilization preserves the unique properties of bioproducts such as minerals, vitamins, bioactive compounds, color and flavor (9), although Marques et al. (32) reported a reduction in mineral content after lyophilization and rehydration of freeze-dried fruits. Peels reveal higher contents of potassium, iron and manganese and lower contents of copper and zinc than Amazonian native fruits, such as biribá (Rollinia mucosa), cubiu (Solanum sessiliflorum Dunal), sapota (Quararibea cordata H.B.K.) and umari (Poraqueiba sericea Tulasne) (33).

All analyzed buriti fruit samples are considered suitable sources of potassium and manganese (30, 34). Potassium is essential for blood pressure control and improves cardiovascular function (35), and manganese is considered a structural co-enzymatic component and protects cell membranes against oxidative processes (36).

Fresh or lyophilized pulp had a higher mass fraction of total monounsaturated fatty acids (both 80.11%), especially oleic acid (79.15 and 80.11%, respectively; Table 4). Meanwhile, lyophilized peels had elevated mass fractions of total saturated (23.18%) and polyunsaturated (2.11%) fatty acids, and they were the only by-product containing phytosterols (β-sitosterol 1.37 and stigmasterol 0.37%). Such results are similar to those in the literature, since previous investigations report oleic (75.7%), palmitic (18.9%), linoleic (2.1%), arachidonic (1.7%), palmitoleic (0.3%) and stearic (1.3%) acids in fruits (4, 6).

The identification and quantification of fatty acids in food is necessary given that clinical and epidemiological studies have established that the quantity and type of lipids have a great influence on cardiovascular risk factors and inflammatory processes (37). Essential long-chain fatty acids belonging to the family omega 6 and 3 have healthy effects on physiological processes, including the prevention and treatment of cardiovascular diseases, atherosclerosis, hypertriglyceridaemia, hypertension, cancer diabetes, arthritis and inflammation-related conditions (37, 38). Indeed, we have recently reported the anti-edematogenic effect of epicarp and mesocarp aqueous extracts from buriti fruits against phlogistic agents (carrageenan, compound 48/80, histamine, serotonin and prostaglandin E₂) and reduction of tissue inflammation and the migration of peritoneal leukocytes and TNF-α in mesocarp-treated mice, but only epicarp reduced inflammatory abdominal pain induced by acetic acid (39).

For the first time, this study has described the presence and quantity of phytosterols in lyophilized peels. Phytosterols reduce cholesterol absorption in the intestine by up to 30% by competition due to the structural similarity between molecules and decrease serum LDL cholesterol levels by 8-10% when 1.6-2.0 g of phytosterols are consumed daily (40). Although our study did not find phytosterols in the pulp samples, it should be pointed out that 100 g of buriti pulp often contains brassicasterol, campesterol, stigmasterol, β-sitosterol and sitostanol. Obviously, samples from Brazilian savannas have nutritional profile distinctions when compared with those from the Amazon region, explained, at least in part, by differences in the Amazon biome conditions (5).

In recent decades, thermoanalytical techniques have received increasing attention in most areas of basic and applied chemistry. For natural products, it is worth mentioning that evaluation of their quality depends not only on the chemical composition of the product but also on the quality of the raw material and it reflects processing and storage conditions (41). Therefore, differential scanning calorimetry (DSC) is an additional analysis widely used to detect the thermotropic behavior of complex inorganic and organic materials, such as buriti samples, and variations in melting, boiling and sublimation points or disappearance of these records (42).

Fig. 1 illustrates DSC thermograms of pulp, peel and endocarp powders. The endothermic peak observed at 87.8 °C for peels refers to an endothermic reaction, i.e. melting and loss of residual water. Crystallization after glass transition and subsequent melting were observed, indicating that the sample was in an amorphous state with little crystallization by quenching after heating. Endothermic peaks were not observed in the pulp and endocarp samples, which may indicate that some components present in the peel have a higher affinity for water molecules. Therefore, higher temperatures are necessary for such elimination (which does not occur in lyophilization), or the material may have absorbed some moisture after drying. In addition, three exothermic peaks were observed between 240 and 530 °C, probably due to the degradation of some components in the peel, whose peak at 296 °C was found in thermograms of peel powder only (Fig. 1b).

The DSC curves for pulp and endocarp were similar (Figs. 1a and 1c), where exothermic degradation peaks were observed at 200 °C. A peak temperature of 327 °C found in endocarp samples was displaced to 331 °C in pulp, requiring a higher temperature to start the process. Moreover, it was also observed that for degradation of the pulp less energy was required than for the endocarp. This lower temperature necessary to disorganize the components of freeze-dried pulp suggests that physicochemical differences, such as the shape and distribution of fiber and the presence of lipids, have a pronounced influence on the thermal behavior (42).

The enthalpy (ΔH) was also measured by DSC analysis to determine qualitative parameters based on shape, position and number of peaks during heating or cooling, while the area under the curve supports quantitative examinations to recognize how external factors can affect DSC results (43).

Fig. 2 shows that drying processes generated semi-crystalline powders, characterized by the presence of larger and more intense peaks ranging from 15 to 25° diffraction angles (2Ɵ) and a lower intensity peak close to 90°. These results differ from those of other lyophilized fruits (44) and imply that constituents from buriti samples (minerals, proteins, lipids and crude fiber) alter the crystallinity of granules (45).

To confirm the XRD results, SEM of all samples was performed. The results corroborated XRD studies, since amorphous structures were visualized (Fig. 2) but with a semi-crystalline appearance.

Amorphous structures can be visualized in systems consisting of sugars. Buriti fruits are rich in different types of carbohydrates, and such structures are characterized by a non-crystalline structure in which there is no repetition of geometric unit cell or the presence of well-established flat faces, a finding more common in freeze-dried products (46). Therefore, the presence of sugars such as fructose, mainly in the pulp, as well as the use of lyophilization, are predisposing factors to generate amorphous microstructures. In the freeze-drying process, the glass transition temperature (T_g) is exceeded during freezing, which makes the concentrated amorphous solution less viscous. Subsequently, product collapse may occur when ice is sublimated (47).

As described above, peels and endocarp contained higher mass fractions of insoluble components (fiber) than pulp and lower mass fractions of sugars. This composition was confirmed by comparable microscopic morphological analysis and similar thermograms and XRD curves.

Pulp and endocarp showed pseudoplastic non-Newtonian behavior (Fig. 3), and the flow behavior index values were lower than 1 (n<1) at 25 °C. However, peel had dilatant behavior (n>1), similar to a suspension of insoluble particles (48), probably due to the insoluble/soluble fiber ratio that was about 90-fold higher (Table 2). Therefore, peels showed an inverse behavior since apparent viscosity increased proportionally to the deformation rate γ, which is related to the total dissolution of particles in the sample and direct contact among them, increasing viscosity (49).