Ultrasound‐assisted solubilization of calcium from micrometer‐scale ground fish bone particles

Abstract In order to promote the extraction of biological calcium from fish bone, ultrasonication was used to process micrometer‐scale fish bone particles (MFPs) and investigate the mechanism of action in relation to bone structure. With ultrasonication treatment (300 W, 60°C, 2 h), the content of calcium release increased by 25.6%. Calcium release reached 94.0% of total calcium after 24‐h treatment. The surface of the MFPs was significantly damaged by ultrasound‐induced cavitation, resulting in holes and separation of the layered structure. X‐ray diffraction (XRD) and Fourier transform infrared (FT‐IR) analysis demonstrated that the crystalline structure of hydroxyapatite was disrupted, the triple helical structure of mineralized collagen fibrils (MCFs) was loosened, and hydrogen bonding in collagen decreased, facilitating the release of hydroxyapatite crystals. Thus, ultrasonication may be a practical alternative to nanomilling for industrial processing of waste fish bones to produce soluble calcium as an ingredient in calcium supplements and supplemented foods.

| 713 GUO et al. resilience. Since hydroxyapatite crystals are embedded in the collagen matrix, it is extremely difficult to extract calcium from them. The common current method to promote calcium release from fish bones is grinding the material into nanometer-scale particles. For example, high-energy wet ball milling can grind fish bones into particles with an average size of 110 nm, which increased the degree of calcium dissolution from 0.7% to 27.4% . Nanogrinding degrades the collagen-fiber network structure and breaks down the hydroxyapatite crystals by collision and crushing (Eskin et al., 2005).
Thermal treatment is another method to promote calcium release from fish bone through degradation of the collagen fiber matrix and reducing the mechanical strength of the bone. Heating effectively promoted calcium release from previously ground nanoscale fish bone particles (Jiang et al., 2020;Zhang et al., 2016). Therefore, degradation of the collagen matrix is an effective method to release calcium from fish bones. However, nanomilling is costly, and a feasible alternative technology to increase calcium release from microscale fish bone particles is indispensable.
Ultrasound waves in a liquid system generate cavitation bubbles, which grow larger by absorbing gas or vapor, then collapse, releasing the absorbed energy as heat and breaking up any solid material present, which has been reported that can disrupt the triple helical structure of collagen and facilitate the decalcification of bones in the medical field (Amiri et al., 2018). Ultrasonication has been used to extract gelatin from animal skins, by destabilizing the collagen structure and loosening the triple helical structure (Ali et al., 2018b).
Ultrasound treatment has been used to extract collagen-Ⅱ from chicken sternal cartilage; the triple helical structure was partially denatured after longer treatment times (Akram & Zhang, 2020).
Ultrasound treatment has also been used for decalcification of bones, or organs in the medical field, indicating that it should be able to release calcium from fish bones (Chen et al., 2007;Milan & Trachtenberg, 1981). However, to our knowledge, there has been no report on the extraction of soluble calcium from fish bones using ultrasonication.
In this study, fish bone particles in the micrometer size range were made from Mackerel bones by ultrafine friction milling.
Changes in the physicochemical properties (particularly calcium release and micromorphology) of fish bone particles during ultrasound treatment were investigated and compared. The mechanism by which ultrasound treatment promotes calcium release from fish bone particles was analyzed to help provide a theoretical basis for the practical industrial extraction of soluble biological calcium from fish bones.

| Materials
Mackerel (36 cm length, 457 g/fish) were provided by Ruifang Food Co., Ltd., Quanzhou, China. Mackerel were filleted and deboned using a roll-type meat separator (YBYM-6004-B, Yingbo Food Machinery Co., Ltd.), and fish backbones were collected. The fish backbones were cleaned two to three times using deionized water to remove blood and flesh, and then stored at −20°C. All reagents used were of analytical grade from local suppliers.

| Preparation of coarse fishbone particles (CFPs) and micrometer fishbone particles (MFPs)
Frozen fish backbones were thawed at room temperature, cut with a knife into pieces 5-10 cm long, then immersed in 40% aqueous Na 2 CO 3 at a ratio of 1:3 (w/v), and autoclaved at 121°C (100 kPa) for 1 h to remove connective tissue and fat. The cooled, separated vertebrae were cleaned two to three times using deionized water and dried at 105°C for 1 h.
Coarse fishbone particles (CFPs): The fish vertebrae were ground using a high-speed pulverizer (FW100, 50 Hz, Tester Instruments Co. Ltd) at a speed of 4,500 rpm for 60 s.
Micrometer fishbone particles (MFPs): The CFPs were further ground using an ultrafine friction grinder (PX-MFC 90D, WIGGENS) at a speed of 3,000 rpm for 45 s.

| Determination of particle size
The mean particle size and particle size distribution of CFPs and MFPs were determined by laser light-scattering, using a Microtrac S3500 analyzer (Microtrac Inc.). The particle sizes of CFPs and MFPs in the micrometer range (1-1,000 μm) were measured, and the results were analyzed with Microtrac S3500 software using a Mie scattering model. Raw data were drawn with Origin 8.5 software.

| Determination of content of calcium release
The content of calcium release was determined under an in vitrosimulated digestion system to digest samples according to a previously described method (Zhang et al., 2017). Samples (3 g) were dissolved in simulated gastric fluid (SGF, 100 ml). SGF was composed of KCl (6.9 mM), NaCl (42.7 mM), CaCl 2 .2H 2 O (0.15 mM), NaHCO 3 (25 mM), KH 2 PO 3 (0.9 mM), MgCl 2 (H 2 O) 6 (0.12 mM), (NH 4 ) 2 CO 3 (0.5 mM), and HCl (15.6 mM) (Brodkorb et al., 2019). The pH was adjusted to 3.0 ± 0.2 using 6 mol/L HCl and porcine pepsin solution (0.5 ml, 3200 U/mg, Sigma-Aldrich) was added to achieve an activity of 2,000 U/mL in the final digestion mixture. The mixture was incubated in a constant temperature shaker at 37°C and 140 rpm/min for 2 h, and then, the samples were centrifuged at 3996 g for 20 min and the supernatant was filtered and diluted with deionized water. Additionally, the factors in simulated gastric fluid including enzyme (porcine pepsin, other enzymes), pH (pH = 1, 3, 5 and 7) and digested time (0, 1, 2, 4, 6, and 8 h), were further investigated to explore the optimal factors. The kinetics of ultrasonic treatment time was also conducted. The calcium concentration was measured using an atomic absorption spectroscope (A3, Beijing Purkinje General Instrument Co. Ltd) as the method of GB/T 5009.92-2003; the determination parameters were set to wavelength of 422.7/nm, height of burning head 6/mm, lamp current of 3/mA, flame Air-Acetylene, acetylene flow rate of 1500 L/min.

| Determination of chemical components
The contents of moisture, ash, fat, protein, phosphorus, and heavymetal ions including, Pb, As, Hg, Cd, and Cr, were detected according to Jin Zhang and Pham Viet Nam's methods (Nam et al., 2019;Zhang et al., 2016Zhang et al., , 2017. The inductively coupled plasma-optical emission spectrometry (ICP-OES) analysis was performed using a PerkinElmer Optima 4300 DV spectrometer.

| Ultrasound-assisted calcium extraction
Control group: MFPs (3 g) were added to 100 ml simulated gastric fluid as described above without adding a digestive enzyme, and the mixture was incubated as described in Section 2.4. Ultrasoundtreated fish bone particles (U-MFPs): MFPs (3 g) were added to 100 ml simulated gastric fluid as described above without adding a digestive enzyme, and the mixture was prepared using an ultrasonic extractor SCIENTZ-ⅡDM (Ningbo Scientz Biotechnology Co., LTD) equipped with an amplitude transformer-Φ6. To optimize the ultrasound treatment conditions, ultrasound treatment was performed at a fixed frequency (20 kHz), but the power (100, 300, and 500 W) at temperature 37°C for 2 h, temperature (37, 60, and 85°C) at power 300 W for 2 h and time (0, 1, 2, 3, 4, 6, 8, 10, 12, 14 and 24 h) at 300 W, 60°C were varied. The processing mode was tiptype, and the extractor was operated in a pulsed mode, with 5-s sonication and 5-s resting time, in order to avoid over-heating of the reaction system.

| Mathematical fitting of calcium release with ultrasonic treatment
The Higuchi equation (Equation 1) and zero-order kinetic function (Equation 2) were used to curve-fit calcium release after ultrasonic treatment. The mathematic models were as follows: where R 0 is the initial calcium release of MFPs (mg/g MFPs ), k is a calcium release constant, which can be determined by linear regression analysis between R(t) and ultrasonication time 1/2 , t is the ultrasonication time (h), and R(t) is the calcium release after a given ultrasonication time (mg/g U-MFPs ). A higher k value corresponds to a higher overall release rate.

| Morphological observation
The microstructures of CFPs, MFPs and ultrasound treated samples (U-MFPs) were visualized using a scanning electron microscope (SEM; JSM-6380LV, Tokyo, Japan). Before scanning, samples were mounted on a bronze stub and sputter-coated with gold. The instrument settings were as follows: accelerating voltage, 5.0 kV, and magnification, ×50,000.

| Fourier-transform infrared spectroscopy (FTIR)
FTIR was used to analyze the functional groups present in inorganic hydroxyapatite and organic mineralized collagen fibrils (MCFs). The samples were vacuum freeze-dried and recorded using a Nicolet™ 360 FTIR spectrometer (Thermo Fisher Scientific, Waltham, MA). A spectral range from 500 to 4000 cm -1 with a resolution of 2 cm −1 was analyzed.

| X-ray diffraction (XRD)
The samples were vacuum freeze-dried, and crystal structural analysis of hydroxyapatite and MCFs was performed using an X-ray diffractometer (Rigaku TTR-III, Rigaku) with a CuKα X-ray source (λ = 1.54178, generator voltage of 40 kV, incident current of 200 mA). Scanning was carried out from 5 to 50° at a scan rate of 0.02 deg/s.

| Statistical analysis
All experiments were performed in triplicate with a completely randomized design. Analysis of variance (ANOVA) and regression was accomplished using Statistical Analysis System software (SAS Institute). Differences between mean values were determined by Duncan's multiple range tests.

| Particle size and morphology of CFPs and MFPs
The mean diameter of the volume distribution (MV) was 124.6 μm for CFPs, markedly higher than that of MFPs (MV = 68.5 μm) ( Figure 1a). The distribution peak at D 50 = 337.9 μm for CFPs was absent after ultrafine friction treatment and the size distribution of MFPs contained few particles larger than 300 μm. The volume % of particles larger than 200 μm reduced from 27.5% in CFP material (2) R(t) = R 0 + kt to 2.8% in MFP. The ultrafine friction treatment effectively reduced the particle size of ground fish bones. The reduced particle size after ultrafine friction treatment was also apparent from SEM imaging ( Figure 1b). However, the surface morphology of CFPs and MFPs did not appear significantly different, suggesting that ultrafine friction treatment did not disrupt the internal structure of the particles, which is necessary to efficiently release soluble calcium. Compositional analysis of CFPs and MFPs (Table 1) revealed some significant differences. The contents of fat, protein, and phosphorus decreased after ultrafine friction treatment. The contents of moisture (2.7%-3.0%) and calcium (152-175 mg/g) increased in MFPs. What is more, the contents of heavy metal ions were significantly decreased after ultrafine friction treatment, and within the range of Chinese standards. A previous report demonstrated that reduction in fish bone particle size facilitates calcium release, because of larger specific surface area and increased porosity, which facilitates the access of acid to the inorganic phase of the particles (Jeong et al., 2013). For example, nanomilling increases calcium release from fish bone particles compared with micromilling (Li et al., 2020;Yin et al., 2015). In addition, thermal treatment of nanoscale fish bone particles was more effective in promoting calcium release than that on microscale particles . However, nanomilling is costly and a feasible alternative technology to increase calcium release from microscale fish bone particles is indispensable.

| Effect of digest factors
The assessment of calcium release was conducted in a simulated gastric system (Li et al., 2020;Yin et al., 2015); thus, the factors that might affect the calcium release in the simulated gastric system were explored. The pH (pH = 3) and digestive enzyme (pepsin) were the main factors in the gastric system that affected calcium release, so the effects of different enzymes and pH were investigated.
Pepsin and other enzymes (trypsin, neutral protease, and papain) were applied to digest MFPs. However, enzymes other than pepsin, operating at their pH optima, had no effect on calcium release F I G U R E 1 Particle size of CFPs and MFPs. (a) Particle size distribution and mean particle size (μm) of CFPs and MFPs. (b) Scanning electron microscopy images for CFPs and MFPs ( Figure 2a). Moreover, the amount of calcium release was not significantly different with additions of pepsin between 0 and 3,000 U/ mg, indicating that calcium release in the gastric system depends on acid hydrolysis, rather than on pepsin (Figure 2b). Varying the pH ( Figure 2c) indicated that little calcium was released at pH above 3, but that pH 1 was only slightly more effective. The amount of calcium release did not change much after acid hydrolysis between 1 and 8 h (Figure 2d). Therefore, subsequent calcium release measurements were conducted in the simulated gastric system at pH 3, without pepsin present.

| Effect of ultrasound
Since ultrasonication can disrupt the triple helical structure of collagen and facilitate the decalcification of bones, the parameters of ultrasonic power, temperature, and treatment time were systematically varied to optimize ultrasound-assisted acid hydrolysis and promote calcium release from MFPs. The amount of calcium released increased with ultrasonic power, and at 300 W, was about 1.3 times that of the no-ultrasonication control (p < .05, Figure 3a). Ultrasonic treatment temperature also influenced calcium release, reaching a maximum of 107.5 mg/g MFPs (64%) at 60°C (Figure 3b). However, this figure was only slightly higher than that achieved at the 300 W power setting, indicating that temperature has minimal influence on ultrasonication-assisted calcium release. This is in contrast with a report that thermal treatments increase calcium release from fish bone particles during high-energy wet ball milling ). This appears to be because the particles in this study were of micrometer scale, rather than nanometer scale (Jiang et al., 2020). Calcium release significantly, but very gradually, increased with increasing treatment time (Figure 3b), reaching 94.0% after 24 h at 300 W and 60°C, compared with 53.0% after 0.5 h.

| Kinetics of ultrasonic treatment
The relationship between calcium release and ultrasonication time was fitted to the Higuchi function (Section 2.2.5, Equation 1) and the zero-order kinetic function (Equation 2) ( Table 2). The Higuchi function had a better correlation (R 2 > 0.99), similar to that in a previous report on the effect of thermal treatments on calcium release from fish bones during high-energy wet ball milling (Zhang et al., 2017).
These data suggest that the effects of ultrasonic treatment on calcium release may be similar to that of high-energy wet ball milling.

| Surface morphology
The surface morphology of MFPs after different ultrasonic power, temperature, and time treatments is shown in Figure 4. The fish bones used were vertebrae (inset in Figure 4a1). Without ultrasonication, the surface of the MFPs was dense and ordered, with stacked layers of planar mineralized collagen fibrils (MCFs) and much smaller hydroxyapatite crystals. The MCFs were well ordered along the longitudinal axis, and hydroxyapatite crystals filled the intervals between adjacent MCFs in a staggered arrangement ( Figure 4a1). After different ultrasonic power treatments, multilevel hierarchical structures were clearly visible. The MCFs appeared to be arranged in parallel, forming flat sheets, which had been partially separated by the ultrasonic treatment. It also appeared that many of the hydroxyapatite crystals had been removed, making the layered structure more clearly visible (Figure 4a2, a3, a4 in 300 μm and 30 μm magnification). It appears that disruption of the layered structure facilitates the release of hydroxyapatite crystals. Moreover, the MCFs displayed holes and damage to their edges, which became more prominent with increased ultrasonic power (Figure 4a2, a3, a4 in 5 μm magnification). Ultrasound waves cause the creation, growth, and collapse of bubbles in liquid, known as cavitation (Liew et al., 2015). The diffusion of solvent vapor into the bubbles leads to their growth and implosive collapse, producing a localized hotspot by adiabatic compression (Bastami & Entezari, 2012). The hot spots momentarily reach very high temperatures (5000-25,000 K) and can break intramolecular chemical bonds, especially in organic compounds (Neppolian et al., 2008). Therefore, it appears that cavitation disrupts the surface of MFPs, especially the organic collagen, to facilitate the release of hydroxyapatite crystals.
After ultrasonic treatment at different temperatures, the changes in morphology were similar (Figure 4b) to those in Figure 4a, but more pronounced. With treatment at 60°C, the damage was more severe and the holes were bigger (Figure 4b2). However, after treatment at 85°C, the particles appeared much more extensively disrupted, with the layered structure mostly separated and the layers broken into much smaller pieces, although calcium release did not increase further (Figure 4b3, Figure 3b). Increasing treatment time (at 300 W and 60°C) resulted in similar changes to those in Figure 4a, which became gradually more pronounced (Figure 4c).
Therefore, ultrasonic cavitation could disrupt the layered structure of the MFPs resulting in the appearance of holes in the layers and the near disappearance of hydroxyapatite crystals from the outside of the particles and from between the separated layers.

| FT-IR spectra
Ultrasound waves can break down collagen by loosening or denaturing its triple helical structure (Ali et al., 2018a(Ali et al., , 2018b. Ultrasound-assisted extraction of hydroxyapatite from Nile tilapia fish scales disrupted the collagen structure (Sricharoen et al., 2020).
Ultrasonic-assisted extraction of nanocalcium from eggshell, which contains collagen, has also been reported (Liew et al., 2015), but the mechanism of promoting calcium release was not discussed.
Therefore, this was explored in the study. The  bonding is the main stabilizing interaction of the α-helical peptide chains forming the triple helical structure of collagen (Bhattacharjee & Bansal, 2010), and the weakening of these interactions in U-MFPs suggests that the collagen structure had been partially denatured.

| X-ray diffraction analysis
X-ray diffraction was also used to detect the changes to the crystal- The peak intensity and sharpness of the peak at 25.82° decreased and a new peak at 23.41° appeared. The broad peak at 32.01° greatly decreased in intensity and two sharp and intense peaks appeared at 29.28° and 30.61°. These spectral changes indicated that the crystal structure of hydroxyapatite was disrupted and some of the hydroxyapatite was released from the MFPs; thereafter, the collagen appeared to rearrange into its preferred crystalline structure. These findings are different from those for nanomilling, which also attenuated the characteristic peaks of collagen (Jiang et al., 2020).

| CON CLUS ION
Taken together, the findings of this study indicated that ultrasonic treatment promoted calcium release from micrometer-scale fish bone particles (MFPs). Ultrasonication at 300 W, 60°C and for 2 h increased calcium release by 25.6% and calcium release reached 94.0% of total calcium, after 24 h ultrasonication. Overall, these findings demonstrate that ultrasonication can disrupt the crystalline structure of hydroxyapatite, promoting its dissolution, but did not denature the crystalline triple helical structure of collagen in the MCFs, only loosened and opened it up, allowing the hydroxyapatite crystals to escape. Ultrasonication may be a practical alternative to nanomilling for industrial processing of waste fish bones to produce soluble calcium as an ingredient in calcium supplements and supplemented foods.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L A PPROVA L
Informed Consent: Written informed consent was obtained from all study participants.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available on request from the authors.