Chía seed (Salvia hispanica L.) was obtained from a market located in Cd. Obregon, Sonora (Northwest Mexico). The citrus pectin (PC) was provided for Faga Lab^® food grade with ≈65% D-galacturonic acid (FAVELA PRO, SA DE CV, Sinaloa, Mexico). All reagents used were grade or higher quality.

Firstly, chia seeds were washed with distilled water and left in water [1:30 w/v] for two hours at 50°C and non-stirring. Then, chia seeds-water was collocated in constant stirring for three h and then filtered through nylon mesh to remove seeds. The material filtered was centrifuged at 200 g for 15 min to discard seed remains. Finally, CM was taken from the supernatant and divided into aliquots to keep them at −20°C until processing.

Firstly, a PC solution at 3.0% [w/v] was prepared to elaborate the membranes. Then, blends of PC/CM were prepared as follows: 100/0 (control), 80/20%, 60/40%, and 40/60% [w/w]. These blend membrane solutions were placed in a Petri dish (triplicate) and dried in an oven (Binder-FD53) at 40°C for 48 h. The formed membrane was placed in a desiccator with silica gel at 25°C until analysis.

A micrometer (Mitutoyo, Japan, precision = 0.001 in) was used to measure the membranes’ thickness. Randomly, six measurements of thickness per membrane were carried out and expressed in mm.

Transparency was determined by the method of Han et al. [28]. Firstly, rectangles of membranes of 9 × 45 mm were cut and placed inside the spectrophotometer cell (Jenway 7305, UK). Otherwise, the transmittance was calculated at 600 nm, and the transparency of the membranes was calculated using the Eq. (1).

(1) Transparency=log(T600/t)

T600 corresponds to transmittance at 600 nm; t represents the membrane thickness (mm).

In addition, color and opacity were measured using an X-rite SP-64 spectrophotometer, calibrated with white and black standard plates. The membranes were randomly chosen for analysis, and five measurements were performed. All results were based on the CIE model (L* a* y b*). Each membrane’s opacity (%) was analyzed over white and black standards, and the opacity was calculated using the Hunterlab color scale [29].

Firstly, five milligrams of each membrane (by triplicate) were scanned (2 cm⁻¹ of resolution) and monitored the wavenumber region of 400–4000 cm⁻¹ with an ATR-FTIR Perkin-Elmer 1600 spectrophotometer in transmittance mode. This study evaluated the intermolecular interaction sites of the membrane PC-PC, PC-CM components, and CM-CM biopolymers.

SEM studied the microstructure and interaction of the PC/CM membranes. Samples were carefully cut, then coated with gold-palladium to impart conductive properties and viewed on a scanning Electron Microscopy [JEOL JSM-5410LV, Japan] with an INCA system and Energy Dispersive X-Ray (EDS) detector operated at 20 kV [30].

The tensile strength (TS), %elongation (%E), and elastic modulus (EM) were determined using the standard method ASTM-D882-12. PC/CM membranes. Initially, membranes in rectangles of 6.0 × 1.0 cm were cut and conditioned at 52% relative humidity (RH, used NaBr) for 24 h. Then, the membranes were analyzed utilizing a texture meter TAXT-Plus (UK) under 30 kg load, an initial gauge of 4 cm length, and stretched using a crosshead speed of 20 mm•min⁻¹. TS (MPa), %E (%), and EM (MPa) were calculated using Exponent Lite Software. Each membrane was analyzed five times.

The thermal stability of the compounds from PC/CM membranes was determined by a Perkin-Elmer Thermogravimetric Analyzer model Pyris 1. Samples of the membrane (2.4 mg) were weighed and subjected to healing from 30 to 800°C (using a healing rate of 10 °C•min⁻¹) under a nitrogen atmosphere. Thermogravimetric curves of the mass loss rate were obtained.

Firstly, a human lyophilized serum obtained from SPINREACT company (spintrol H: calibrator of multi-components for clinical chemistry assays—SPINREACT®, Girona, Spain) was reconstituted using 5.0 mL distilled water and mixed thoroughly to prevent foam forming. This human serum contained the albumin concentration of well-known reference to be measured by the bromcresol green colorimetric method. After, the mix was used to have an approximation of the capacity of each membrane (6 mm-disk) to catch serum volume, where both maximal and minimal catchment were observed for the control (45 µL) and 40/60 (40 µL), respectively (Fig. 1), to obtain a margin of sufficient volume of sample (adjustment 80 µL) to be used in the albumin adsorption test.

Eighty microliters of serum were added to an Eppendorf tube that contained a 6 mm-disk of the membrane. The tubes were incubated at 37°C for two h (Binder incubator, Germany); and later, they were centrifuged at 11,000 g (SIGMA-1-15P, Germany) for 10 min. Twenty microliters of supernatant were mixed in 2.0 mL bromcresol green [0.12 mmol/L, and pH 4.2] (SPINREACT^® kit), and the mix was incubated at room temperature for 10 min. Subsequently, the absorbance for the mix was determined at 630 nm (Jenway model 7205 spectrophotometer, UK). Finally, results were calculated according to Eq. (2) and expressed as albumin adsorption percentage (%RA).

(2) %RA=CAC−CAMCAC×100 where CAC = Albumin concentration in the no membrane calibrator; CAM = Albumin concentration with added membrane.

Firstly, the erythrocytes were washed three times in saline solution (NaCl, 0.9% w/v) to remove components from the plasma. Afterward, an erythrocyte suspension (5.0% v/v) was prepared using a saline solution. Eppendorf tubes with a membrane of 6 mm disk were placed on the top, and 200 µL erythrocyte suspension was added and incubated at 37°C for two h (BINDER incubator, Germany). Then, 200 µL of saline solution was added and centrifuged at 2700 × g for 10 min. Later, 200 µL supernatant was dissolved in 2.0 mL distilled water and read at 540 nm. One sample of 100% hemolysis–erythrocyte and water–was used as a reference to calculate the hemolysis percentage [hemolysis (%)] using Eq. (3):

(3) Hemolysis (%)=AHSAHC×100   where AHS = Hemolysis absorbance of the sample; AHC = Complete hemolysis absorbance.

The PC/CM membrane concentrations were assigned as independent variables, and the optical, mechanical, TGA, percentage of hemolysis, and percentage of albumin retention as response variables in the variance analysis. The difference among the means was estimated using the Duncan test (α = 0.05). All statistical analyses were carried out using the NCSS software, version 2000.

Transparency partially determines the miscibility or compatibility of biopolymer blends and the stability of microstructure [31]. Additionally, medical devices usually require acceptable transparency to visualize through material [32]. All membranes were transparent (Fig. 2), and although the control had the highest transparency this also had high thickness (Table 1).

Some authors have reported that thickness can act significantly to the detriment of the membrane’s transparency [33]. Conversely, the addition of CM resulted in thin membranes and reduced (p < 0.05) transparency of the PC/CM membranes in 8.27% (80/20), 12.58% (60/40), and 8.63% (40/60), as compared to the control. Other prepared membranes with polysaccharides blend (chitosan and guar gum) were lesser transparent (46.96–53.87 values) than our membranes. This was attributed to guar gum and water molecules, which modified the refractive index of chitosan [34]. López-Mata et al. made a blend of Aloe gel with egg white (protein animal) to obtain membranes with good transparency values (2.04–2.57), and more transparent than those in this study [35].

Usually, transparency and opacity are considered opposite properties in the optical area. Nevertheless, the measurement of transparency in the membranes is calculated with the transmittance, which depends on color, thickness, and interference fringes (slight) [36]. Otherwise, opacity is calculated using a black and white backing from the reflectance measurements of membranes [37]. Therefore, changes in the transparency values are not necessarily related to a significant change in the opacity (Table 1). The addition of CM in the membranes resulted only in transparency changes but not in opacity. In our study, all the membranes showed low opacity (≈10%), similar to those reported in two CM/whey protein membranes (proportions 1/3 and 1/4) [22], and membranes based on PC/Aloe-gel [38].

The ability of a biomaterial to interact with middle-wave ultraviolet radiation (UVB) in the spectral region has been reported as wavelength ranges that may induce skin damage and wound healing delay [39]. Our results suggested that all membranes at 280 nm (Fig. 3), can block UVB light between 52% and 66%, and the CM addition did not necessarily improve this attribute. Similar behavior was also observed at 350 nm. In comparison with other blends of matrices such as chitosan/guar gum (>90%) [34] or PC/Aloe-gel (>72%) [38], our PC/CM had the lowest efficiency in blocking UVB.

Color parameters (a*, b*, and L*) of the membranes are also presented in Table 1. The lightness (L*) in all membranes was good, but the increase in the CM ratio (60/40 and 40/60 membranes) significantly reduced lightness. Similar behavior was observed for the b* value, where an increase of the CM in the composition of the membrane resulted in increased yellow coloration (p < 0.05). These changes may depend on the optical configuration produced by both the molecular organization and the seed compounds contained in the mucilage during the membrane development [16].

SEM analysis is essential to recognize the microstructure and interaction of the membrane components. Fig. 4 shows the scanning electron micrograph of the membranes obtained. It displays the morphological evolution of the PC membrane across the addition of different proportions of CM. It was observed a homogeneous ribbed surface on the control membrane. Previous studies also reported grooves and rivers marks in protein animal membranes [40] similar to our control. The formation of these channels may be associated with a molecular realignment of the polysaccharides during the water loss. Cokaygil et al. analyzed the microstructural surface of membranes of pectin orange peel and corn starch with layered silicates, and grooves on the membrane surface were associated with fractures [41]. The addition of CM in 80/20 proportion resulted in a homogeneous, smooth, non-porous and non-grooved surface of this membrane. However, a significant increase in the proportion of CM (60/40 and 40/60 membranes) resulted in a sphere (≈13 µm), ovoid (≈25 µm) [shadow black (40/60)], and amorphous aggregates (white). Krystyjan et al. reported that the addition of mucilage in starch/Plantago psyllium seed mucilage membrane formed spherical particles similar to those in this study [42], probably associated with microcapsules formation (size 150–200 nm). In addition, the increase of white particles with granular forms may be attributed to a displacement of insoluble material on the surface of PC by an increase of CM amount (high water content) as previously reported [22].

Table 2 shows the mechanical properties, expressed by tensile strength (TS), elongation at break (%E), and tensile modulus of elasticity (ME) of PC/CM membranes. Good resistance to plastic deformation has been reported for the PC membranes, but limitations such as rigidity and brittleness are present [43]. Our control (only PC) showed similar TS and flexibility as compared with high (TS = 20 MPa, %E = 17) and low (TS = 19 MPa, %E = 17) methyl ester PC membranes, although the elastic modulus was lower in our membranes [44]. This behavior may result from a channel formed on the surface of the control (See SEM analysis), that may affect the membrane lability when subjected to an extension upon tension, causing loss of structural integrity of the membrane [45]. The 80/20 blend showed a significant increase of 29% and 19% in the TS and ME, respectively, compared to the control. This increase in resistance and flexibility may be associated with interactions between PC and CM components. A significant proportion of CM (60/40 and 40/60) in the membrane showed no significant improvement in TS and %E. This may result in the displacement of insoluble material (showed in SEM analysis) affecting the degree of interactions of the membrane components. Initially, the 80/20 membrane was more rigid than the control, while the 60/40 membrane showed the highest flexibility (ME decreased). Significant changes were observed in ME, probably associated with the plasticizer used [46], or changes in the interaction and alignments among the PC chains and CM as recorded previously for fish gelatin/Aloe membranes [47].

Fig. 5 shows the FTIR analysis spectra of the blends of PEC/CM. This test analyzed the intermolecular interaction between components of the membrane. It was observed that the transmittance bands and peaks at 3287 cm⁻¹ (-OH), 2990 (-O-CH₃⁻), and 2880 cm⁻¹ (-CH₂⁻) are specific to PEC. Additionally, the band centered in 1731 and 1620 cm⁻¹ is attributed to the stretching vibration of -COO-CH₃ and -COO⁻, respectively. The stretching of the C-O-C bond of the pyranose ring was identified in the adsorption peaks located at 1230 and 1015 cm⁻¹ [46]. The FTIR spectrums of membrane blends of PEC with CM exhibited changes, particularly in the -OH band when CM increased. These changes indicated intermolecular interaction between PEC-PEC, PEC-CM, and CM-CM biopolymers, mediated by -OH and -COOH groups through the formation and disruption of hydrogen bonds [48]. As the amount of hydrogen bonds was proportional to the intensity of the -OH band, it was suggested that the ratio of hydrogen bonds disrupted in PC-PC is greater than the hydrogen bonds formed in the intermolecular association of PC-CM, due to the low number of -OH groups in the chia mucilage. On the contrary, the transmittance % intensity for the -OH band decreased at a higher ratio of CM (40/60), because the great ability of CM to adsorb water molecules [21].

Thermal resistance of the biopolymeric membranes was measured by thermogravimetry characterization. This technique is based on determining a loss or gain mass of the sample concerning temperature and time (controlled temperature increase), indicating the temperature range at which the membrane components start decomposing [49]. Fig. 6 shows the analysis of the TGA thermograms of the membranes.

The control membrane showed a loss weight of 8.29% at 84°C associated with water evaporation. The first loss of weight in all the membranes with CM was 3.37% (80/20), 13.10% (60/40), and 12.05% (40/60) at 89.6°C, 81.3°C, and 82.3°C, respectively. A similar loss weight, around 6.0%–9.0%, was associated with the desorption of loss moisture linked by hydrogen bonds to the polysaccharides of CM, as previously reported [23]. The second thermal event observed at 102°C–199°C corresponds to glycerol degradation used during the membrane development. Similar thermal behavior of the glycerol in the edible membrane of CM has been observed previously [16]. High membranes’ losses were observed at around 200°C–280°C (between 43%–50% loss) in the third thermal event, associated with the thermal degradation of polysaccharides in the PC and CM. Giancone et al. and de Campo et al. reported a similar thermal degradation of polysaccharides in PC and CM. Another thermal event showed maximum peaks around 312.7°C, which meant a weight loss from 19% to 28% in the control membrane [4,50]. Other temperature peak events at 600°C may be attributed to the decomposition of structures produced due to crosslinking reactions during the heating procedure [51].

The low residual mass-loss for 80/20 (4.58%) and 40/60 (2.52%) membranes as compared with control (1.27%), revealed that incorporation of CM in the described proportions, resulted in a slight improvement in thermal stability. Krystyjan et al. found that mucilage addition may fill areas among starch molecules in the microcapsule inside starch/Plantago psyllium mucilage seed membranes, probably resulting in more thermally stable membranes [42]. In previous CM nanocomposite membranes with starch nanocrystals, a similar behavior has also been observed [23]. This study also revealed the presence of CM microencapsulated inside the PC matrix during SEM analysis.

When a wound dressing is placed on damaged tissue, interaction with blood and serum proteins occurs, and they may be adsorbed on the biomaterial surface. The adsorbed serum proteins may change the initial activity of the biomaterial surface probably modulating the direction of the immune response of the biomaterial [52]. Fig. 7 shows the percentage of human albumin retention from the serum of the tested membranes.

The control (11.4%) and the 80/20 (6.5%) membranes showed albumin retention capacity, being the control that showed (p < 0.05) the highest retention.

Both 60/40 and 40/60 membranes did not show albumin retention capacity, indicating that the PC amount mediates the albumin adsorption in the membrane, while the CM had reduced the albumin adsorption. Cacidedo et al. [53] found 10 times higher the incorporation of human serum albumin in hybrid bacterial cellulose-pectin films than in the control (bacterial cellulose) using a fluorescence technique. The PC chemical structure has an important influence on the albumin adsorption [53]. The presence of free carboxyl groups (negatively charge) in the PC may explain the albumin adsorption (positively charge in serum) in the membranes [54]. Other studies based on the bovine serum albumin adsorption capacity of PC gel bead, reported that this property may also be attributed to Van der Waals interactions, hydrogen bonding and hydrophobic interactions of the chemical structure of the albumin and PC [55]. Visalakshan et al. also reported that albumin had a high affinity with carboxyl functional groups from a variety of biomaterials. They observed that the bound surface with albumin did not trigger an immune response [56], which is particularly important in reducing the probability of rejection of biomaterials in the tissues [57]. The biomaterial-albumin bound may also be advantageous as wound dressing because the albumin on the biomaterial surface can influence the reduction of proinflammatory cytokines [58]. On the other hand, it is relevant to mention that the adsorption capacity of other plasmatic proteins, probably occurred in the membranes, although they were not quantified in this study.

The 40/60 membrane showed good biocompatibility with the erythrocytes (2.94%). Increased hemolysis percentage was a response to increased PC in the membrane composition; 60/40 (56.9%), 80/20 (65.6%), and control (75.6%) (Fig. 8). The high hemolysis percentage shown by the PC has not been reported previously, although it has been considered a biopolymer with low hemolytic capacity (less than 5%) [59–61]. We consider that high hemolysis showed by PC, probably is related to the type of PC used (food grade) in our study, which contained a high content of glucose proportions PC = 113 mg/dL; 80/20 = 106 mg/dL; 60/40 = 65 mg/dL; 40/60 mg/dL which was measured by the oxidase/peroxidase method. A high concentration of extracellular glucose likely causes changes in the osmotic pressure by modifying both [Na⁺] and [K⁺] intracellular. Protein glycation may also induce irreversible changes in the erythrocyte membrane [61]. The ≈5% hemolysis has been considered acceptable for biomaterials [62], indicating that the 40/60 membrane has good biocompatibility, influenced mainly by the CM.