Medium viscosity (≥2,000 cp) sodium alginate isolated from brown algae, with a molecular weight between 80 and 120 kDa and a mannuronic to guluronic acid ratio of 1.56 was purchased from Sigma-Aldrich, St. Louis, MO, USA. The polymer chitosan (molecular weight: 190 to 310 kDa, deacetylation degree 75% to 85%, and viscosity of 200 to 800 cp) was also purchased from Sigma-Aldrich, USA. High purity grade glycerol (99.8%) and calcium chloride, as a cross linking agent, were purchased from Fisher Scientific, Pittsburgh, PA, USA. All other chemical reagents were of analytical grade.

Glycerol was added at 2.5%, 5%, 7.5%, or 10% to each of two bead solutions. One bead solution contained 2.5% (w/v) sodium alginate in distilled water. The other bead solution contained a chitosan solution (1.25% w/v) in dilute acetic acid (1%) mixed with sodium alginate (2.5% w/v). The alginate-glycerol (AG) and alginate-chitosan glycerol (ACG) beads were prepared by adding the different levels of glycerol into a 500 mL wide mouth beaker containing the two bead solutions, respectively. The bead solution mixture was allowed to stir at 500 rpm for an hour, then transferred into a separating funnel and dropped through an 18-gauge needle (38 mm length, 1.27 mm diameter) from a 100 mL glass syringe into a wide mouth beaker containing a 3% (w/v) calcium chloride solution. The beads were allowed to mix under gentle stirring (10 rpm) for an hour to harden. After one hour, the beads were filtered through a cheesecloth and rinsed under distilled water. Following adequate rinsing, the beads were air dried at 23°C for 12 hours and stored in air tight containers for subsequent experiments.

Representative samples of AG and ACG beads at each level of glycerol inclusion i.e. 2.5%, 5%, 7.5%, or 10% were prepared, in triplicate, to determine free glycerol content according to the method described by Bondioli and Della Bella [11] at 412 nm using a UV spectrophotometer (Shimadzu UV1601, Tokyo, Japan). A calibration curve (R² = 0.996) was generated from glycerol standard solutions at known concentrations. The loading capacity (LC) and encapsulation efficiency (EE) of the AG and ACG glycerol beads were calculated based on the equations below:

The surface morphology of the beads was observed using a field emission scanning electron microscopy (SEM) (FE-SEM, JEOL, 6400F, Japan). The beads were mounted on an appropriate stub, coated with carbon and gold, and photographs were taken at a magnification of 50×, 1,000×, and 10,000×. The working distance was maintained at 20 mm and the acceleration voltage was 5 kV.

Fourier transform infrared (FTIR) spectrum of glycerol, alginate, chitosan, AG, and ACG beads were recorded using a transmittance mode iS50 Thermo Nicolet Nexus 670 FTIR (Thermo Scientific, Waltham, MA, USA) with a built-in diamond crystal. The absorbance pattern of diamond typically seen between 1,900 and 2,400 cm⁻¹ was erased. No bonds were observed in that area. Analysis was performed within the spectral region of 400 to 4,000 cm⁻¹ with 32 scans recorded at 4 cm⁻¹ resolution. FTIR spectra from the different bead preparations were compared to evaluate chemical modifications.

Glycerol release experiments were performed at pH 6, 2, and 8 to simulate rumen, gastric, and intestinal conditions, respectively. The gastric buffer (pH 2) was prepared by mixing 50 mL of a 0.2 M KCl with 13 mL of 0.2 M HCl and brought to a total volume of 200 mL with distilled water. The intestinal buffer (pH 8) was prepared by mixing 100 mL of a 0.1 M KH₂PO₄ with 93.4 mL of 0.1 M NaOH and brought to a total volume of 200 mL with distilled water and the rumen buffer (pH 6) was prepared by mixing 100 mL of 0.1 M KH₂PO₄ with 11.2 mL of 0.1 M NaOH and brought to a total volume of 200 mL with distilled water.

The ability of the encapsulating polymers to protect glycerol from acidic and alkaline hydrolysis was assessed in two separate in vitro experiments. In experiment 1, acidic and alkaline phases were studied sequentially by placing 0.5 g samples of AG and ACG beads inside conical flasks (200 mL), containing 100 mL of pH 2 buffer. Four flasks were prepared per treatment to provide for additional replications. All flasks were incubated at 39°C for 2 h to simulate gastric digestion. After 2 h, a sample was taken and stored for subsequent glycerol analysis. The beads were rinsed with distilled water and transferred to pH 8 buffer at 39°C and incubated until 24 h to simulate intestinal conditions. In experiment 2, 0.5 g samples of AG and ACG beads were incubated in 200 mL of pH 6 buffer at 39°C for 8 h to simulate rumen conditions. Each treatment was repeated four times to provide for additional replications. Samples were withdrawn at 2 h intervals and stored for subsequent glycerol analysis. Free glycerol content was determined at 412 nm using a spectrophotometer [11].

In vitro fermentation was conducted using batch cultures of mixed rumen microorganisms. Diets consisted of pelleted alfalfa and corn (60:40). Free (unprotected) glycerol and encapsulated glycerol beads were included to provide glycerol at 25% of dry matter. Glycerol, AG and ACG were substituted for corn. Treatments included: i) Con (no glycerol), ii) G (free unprotected glycerol), iii) AG (alginate beads), and iv) ACG (alginate-chitosan beads). Donor animal for rumen contents was a non-lactating cannulated Holstein cow fed a predominantly forage diet. Rumen contents were collected 2 h after the morning feeding and immediately transported to the laboratory in a thermal flask. In the lab, rumen contents were strained through four layers of cheesecloth and mixed with a buffer solution [12] in a 1:2 ratio (vol/vol). The rumen contents and buffer were maintained at 39°C and CO₂ was purged through the flasks at all times to minimize exposure to O₂. Individual treatment substrates were quantitatively weighed into 120 mL bottles, in triplicate, for each time period and allowed to ferment for 0 and 24 hours. Bottles containing only the rumen inoculum and no substrate served as blanks. Forty mL of the buffered rumen inoculum were added into each bottle under constant flow of CO₂. Bottles were immediately sealed with rubber lined screw caps and placed in a 39°C water bath. The 0 h samples were sealed and immediately placed on ice to stop fermentation. After 24 h, samples were removed from the water bath, placed on ice and culture pH was recorded prior to storing bottles in a walk in freezer at −20°C for further chemical analysis. Short chain fatty acids (SCFAs) were quantified using a gas chromatograph (Model 3380, Varian, Walnut Creek, CA, USA) using a fused silica capillary column (Nukol; Supelco Inc., Bellefonte, PA, USA). The amount of SCFA in each batch culture after 24 h of incubation was corrected for the amount of SCFA in the rumen fluid used as inoculum.

Data from the encapsulation experiment were analyzed according to a complete block design using the MIXED procedure of SAS (SAS Institute, Cary, NC, USA). The statistical model included fixed effect of polymers, amount of glycerol and the polymer×glycerol interaction. Replicate was included as the random term. Data from the glycerol release experiments were analyzed according to a complete block design using the MIXED procedure of SAS. The statistical model included the fixed effects of polymer, pH and incubation time and the respective interactions. Data from the in vitro fermentation with mixed rumen microbes was analyzed by hour (0 vs 24) as a complete block design. The MIXED procedure of SAS was used to separate means within the hour. The statistical model included the fixed effect of treatment. Replicate was included as random effect. Differences among treatment means in all experiments were determined by the LSmeans/pdiff option. Differences were considered significant at p<0.05 with trends defined as 0.05<p<0.10.

Images of wet and dry beads are presented in Figure 1. Fresh beads (Figure 1A) were round in shape with a smooth and homogeneous surface. The drying process collapsed the beads and reduced their dimensions (Figure 1B). The loss of water made the beads slightly darker in color and more ellipsoidal with a rougher surface. Both, the AG and ACG beads were similar in color and size. The ionotropic gelation method used in this study to encapsulate glycerol through an 18-gauge needle resulted in fresh bead size of approximately 3 mm in diameter with a spherical appearance and a smooth outer matrix (Figure 1).

There was a significant effect of the concentration of glycerol and polymer treatment on glycerol LC (Figure 2). Increasing the concentration of glycerol increased (p<0.001) the amount of glycerol loaded irrespective of the polymer treatment. The greatest increase in LC, irrespective of polymer, occurred when the level of glycerol was raised from 2.5% to 5%. Further increases in the level of glycerol concentrations (7.5% and 10%) did not have an appreciable increase in glycerol LC for either AG or ACG treatments. Compared with AG, the ACG treatment significantly increased LC at all glycerol concentrations (Figure 2). The maximum LC was 73.6% and 66.7% for ACG and AG beads, respectively, when glycerol was included at 10%.

The EE for the AG and ACG beads is illustrated in Figure 3. The ACG beads resulted in the highest efficiency (78.5%) compared to the AG beads (72.2%) at the 2.5% of glycerol addition. Increasing the glycerol LC resulted in a significant and linear decrease in EE with the ACG treatment; the effect of AG was variable but significantly lower than the ACG at all glycerol levels.

SEM of AG (A, B, C) and ACG (E, F, G) beads are illustrated in Figure 4. To better depict the molecular structure beads were scanned at three intensities (50×, 1,000×, and 10,000×) as illustrated in A, B, C for the AG beads and E, F, G for the ACG beads. At lower magnification the AG beads (Figure 4A) were smoother and more spherical compared to ACG beads (Figure 4E) that had a more rough and undulating surface. At higher magnification however, we can clearly observe the presence of cracks present in the AG beads (Figure 4C) but not in ACG beads (Figure 4G).

The FTIR spectra of glycerol, alginate, chitosan powder and AG and ACG beads are shown in Figures 5 and 6, respectively. The large peak in glycerol at 3,331.65 cm⁻¹ is due to the characteristic OH stretching modes. The two peaks at 2,935 and 2,880 cm⁻¹ are reflective of the CH stretching vibrations of glycerol whereas the peaks between 1,500 and 1,200 cm⁻¹ are due to the CH bending. The C-O stretching is shown in peaks between 1,200 and 900 cm⁻¹. Alginate powder spectrum also shows a broad band at 3,266 cm⁻¹ related to OH stretching with strong intra and/or inter hydrogen bonding. The peak at 2,929 cm⁻¹ is ascribed to the overlapping symmetrical and asymmetrical C-H stretching vibration of aliphatic chains (-CH₂-, -CH₃). The asymmetric and symmetric vibrational modes of carboxylate ions (COO⁻) were recorded at 1,600 and 1,408 cm⁻¹, respectively. The vibrational mode at 1,081 cm⁻¹ was attributable to the C–O stretching vibration of pyranosyl ring. Due to its polysaccharide structure the (C-O-C) stretching vibration of sodium alginate was manifested at 1,030 cm⁻¹. The spectrum of chitosan is characterized by broad and intense dimeric bands at 3,353 to 3,288 cm⁻¹ (hydrogen bonded OH stretching overlapped with N-H stretching bands). Chitosan also showed bands at 2,919 and 2,872 cm⁻¹ due to asymmetric stretching of -CH₂- and -CH₃ groups. The N-H deformation band of chitosan was found at 1,559 cm⁻¹. The strong bands at 1,648 and 1,575 cm⁻¹ are ascribed to the amide I (C=O stretching) and amide II (N-H bending modes) and C-N stretching vibrations of chitosan [13,14]. A characteristic peak at 898 cm⁻¹ was observed due to the β-1–4 glycosidic linkage bond in chitosan.

Glycerol release from AG and ACG beads when incubated in rumen buffer at pH 6 and 39°C for 8 h is shown in Figure 7. Irrespective of polymer treatment, exposing beads to rumen pH for 8 h resulted in a steady release of glycerol. However, the amount of glycerol that escaped from the AG beads was significantly greater when compared to ACG. At end of 8 h, 36.1% of glycerol was released from the AG beads compared with 29.8% for the ACG beads.

Glycerol release from AG and ACG beads when incubated at 39°C in gastric buffer at pH 2 for 2 h and then in intestinal buffer at pH 8 until 24 h is shown in Figure 8. At the end of 2 h, the amount of glycerol released in gastric buffer tended (p<0.08) to be similar between AG and ACG beads and averaged 15.4% and 12.6%, respectively. Following gastric conditions, both beads were exposed to intestinal buffer for an additional 22 h. The period from 2 to 6 h simulated retention time in the small intestine and 12 to 24 h simulated retention in the large intestine including colon and cecum of herbivores. During the initial 2 to 6 h corresponding to the small intestinal retention time, 69% of the glycerol was released (p<0.05) from AG compared to 56% from the ACG beads. However, by 24 h, more than 90% of the glycerol was released irrespective of the polymer treatment (p>0.10). At the higher pH of the intestinal buffer simulating the small and large intestines substantial amounts of glycerol escaped from both the AG and ACG beads and polymer treatment did not have an effect (p>0.10). More than 90% of the glycerol was released from both AG and ACG beads at the end of 24 h of incubation in the intestinal buffer. During the initial 6 h of incubation glycerol release seemed more rapid from the AG beads compared to the ACG beads. This suggests that ACG would result in a gradual supply of glycerol within the small intestine compared to a rapid release rat with the AG beads.

There was no difference (p>0.10) in culture pH across all treatments at 0 h (Table 1). As expected, pH dropped at 24 h and was similar for all treatments. After 24 h of incubation, total SCFA increased (p<0.05) in Con and G. Acetate was significantly reduced in cultures receiving G (free glycerol) when compared with either the Con or AG and ACG treatments. In contrast, molar proportion of propionate were much higher (p<0.05) in cultures receiving G compared to either Con or AG and ACG. The lower acetate and higher propionate resulted in a significantly lower acetate:propionate ratio in cultures receiving free glycerol (G). Visual observations indicated that the beads behaved similar to the particulate fraction and stayed in suspension during the initial stages of fermentation. At the end of 24 h of incubation some beads remained in suspension while some sank to the bottom with the more dense material. In this study, we did not measure the density of the beads. However, based on the observed behavior of the beads in rumen cultures, we can speculate that the specific gravity of the beads was greater than 1 and less than 1.5. Beads used in this study were on the average 3 mm in diameter. Particles 3 mm in size with specific gravities ranging between 1 and 1.3 are common within the rumen environment [15].