Synthesis and 3D Interconnected Nanostructured h-BN-Based Biocomposites by Low-Temperature Plasma Sintering: Bone Regeneration Applications.

Recent advances and demands in biomedical applications drive a large amount of research to synthesize easily scalable low-density, high-strength, and wear-resistant biomaterials. The chemical inertness with low density combined with high strength makes h-BN one of the promising materials for such application. In this work, three-dimensional hexagonal boron nitride (h-BN) interconnected with boron trioxide (B2O3) was prepared by easily scalable and energy efficient spark plasma sintering (SPS) process. The composite structure shows significant densification (1.6-1.9 g/cm3) and high surface area (0.97-14.5 m2/g) at an extremely low SPS temperature of 250 °C. A high compressive strength of 291 MPa with a reasonably good wear resistance was obtained for the composite structure. The formation of strong covalent bonds between h-BN and B2O3 was formulated and established by molecular dynamics simulation. The composite showed significant effect on cell viability/proliferation. It shows a high mineralized nodule formation over the control, which suggests its use as a possible osteogenic agent in bone formation.

Introduction

Analogous to graphene, hexagonal boron nitride (h-BN) possesses high oxidation resistance and chemical stability, along with high surface area and thermal conductivity.[1−4] Therefore, developing nanostructured three-dimensional (3D) h-BN is expected to increase its significance in applications.[5−7] Various routes (chemical, mechanical, chemical vapor deposition (CVD), and microwave irradiation) are employed for synthesizing porous nanostructured BN-based materials.[8−11] It is extensively prepared by reaction of boric acid, B2O3 or borax with C and N/NH3 and urea at high temperatures around 2000 °C. For nanostructured BN, laser ablation, CVS, and carbothermal reduction of B4C and B2O3 are reported.[12] A transient BO-based liquid phase helps densification at <1000 °C, although the h-BN and the matrix phases develop at ∼1600 °C depending on the low free energy of formation of h-BN.[13] The low melting point of B2O3 (450 °C) and low viscosity of melts of B2O3 help low-temperature consolidation of these compositions. For instance, a one-step synthesis of porous template-free BN microsponges for hydrogen storage application[14] and a template-assisted continuous 3D network of BN aerogels having borazine precursor for oil/organic solvent absorption were proposed recently.[15] However, compared to graphene, the synthesizing routes for few-layered BN nanosheets are less effective due to the strong lip–lip interactions between the adjacent concentric shells of BN basal planes.[16−18] The mechanical cleavage method for preparing h-BN is not suitable for scale up.[19]

Various forms of 3D porous nanostructured materials have been developed by different synthesis methods: bottom-up approaches like solution processing and chemical vapor deposition (CVD) and microwave irradiation. However, these processes yield products having lower modulus (<10 MPa) and mechanical strength values. Spark plasma sintering (SPS) technique allows densification of various ceramic powders at lower temperature and shorter dwell time due to rapid heating and cooling rates. The controlled grain growth in this process leads to desired material properties.[20] Recently, Chakravarty et. al. reported 3D porous graphene by SPS as the implants. They achieved significant mechanical strength of 40 ± 3 MPa, stiffness of 8 ± 1 kN/m, and Young’s modulus of 4.1 ± 0.5 GPa.[21] More recently, Gautam et. al. reported a detailed investigation on synthesis of porous h-BN 3D architecture for effective humidity and gas sensor applications using B2O3 as an additive, which is responsible for the porous structure.[22]

In the present work, attempts have been made to synthesize 3D hBN–B2O3 composite using the SPS technique. The synthesized 3D structures are demonstrated to be structurally stable, mechanically strong, and biologically viable and sufficiently biocompatible to be used as bone implant. The MD simulation has been performed to understand the role of B2O3 in interconnecting h-BN to build the 3D structure and its effect on mechanical properties.

Results and Discussion

Route, Sample Nomenclature, and Mechanism

A typical structure having differently oriented h-BN sheets intercalated by the second-phase B2O3 particles and forming strong covalent bonds under application of current and pressure in SPS has been proposed, as shown in the schematic Figure a. The nomenclature xBNBO was used for the samples synthesized in the current work, where “x” stands for the weight percent of h-BN. This obviously implies that the B2O3 percentage for that sample is (100 – x).

Figure 1

Surface morphology of SPS-sintered BN–B2O3 composites. (a) Schematic of the proposed composite structure. (b) Scanning electron microscope (SEM) image of pure h-BN showing the tiny interconnected sheets of h-BN. (c, d) Low and high-magnification SEM images of sample BN–B2O3 (BNBO) showing interconnected big h-BN sheets within the residual matrix of B2O3 and indicating the fibrous morphology of B2O3. (e) SEM image of high concentration h-BN sample shows fully developed and well interconnected nanosheets of h-BN with minor porosity dispersed in glassy matrix of B2O3.

Morphology Evolution

The influence of B2O3 concentration on grain growth, morphology, and general microstructure of BNBO composites was microscopically analyzed. Figure b shows a typical SEM image of pure h-BN consisting of tiny sheets sintered to each other. The digital image of the sample sintered at 250 °C is shown in the inset. A SEM image of a typical BNBO sample reveals large interconnected h-BN sheets (marked by arrow in Figure c). The high-magnification image (Figure d) clearly reveals the 3D porous, interconnected architecture of the hBN–B2O3 composite with the fibrous morphology of B2O3, as indicated by the dotted region. The high-magnification SEM micrograph shown in Figure e, corresponds to a high concentration h-BN sample and shows interconnected submicron (∼0.4 μm) h-BN sheets, represented by arrows, with large pores at a few regions created by partial evaporation of B2O3 during sintering. It is well known that B2O3 forms liquid phase during sintering; this might explain the high densification rate in the samples even at an extremely low SPS temperature. Formation of liquid phase may also promote orientation of the h-BN grains by processes such as grain boundary rotation.[23]

The transmission electron microscopy (TEM) images, as shown in Figure a–d, correspond to those for the composition 90BNBO and are consistent with the observations made in the SEM images. The general macrostructure of the B2O3 particles dispersed in the h-BN nanosheets is shown in Figure a. The h-BN sheets appear to be of 100–250 nm in diameter. The high-resolution transmission electron microscopy image in Figure b reveals a layered h-BN structure with an interlayer distance of 0.34 nm corresponding to the inter-plane distance (002), in agreement with a previous report.[24] This image also reveals the interface between h-BN and B2O3. The inset of Figure b shows the selected area electron diffraction (SAED) pattern corresponding to (002), (101), (004), and (110) lattice planes of h-BN and reveals the crystalline phase of h-BN, consistent with the X-ray diffraction (XRD) results (Figure S1a). The high-magnification TEM image in Figure c shows randomly oriented layers of h-BN and indicates the B2O3 matrix in the darker regions. The fast Fourier transform patterns (in Figure c) display hexagonal spots with rotation angle of about 11°, which is also corroborated by Figure d, which shows an array of h-BN atoms arranged periodically in a hexagonal symmetry with the lattice constant like the values of bulk h-BN, as previously reported.[25]

Figure 2

Microstructural characterization BNBO (a) low magnification bright field TEM image showing large h-BN sheets along with matrix of B2O3, (b) TEM image of synthesized BN–B2O3 composites showing h-BN nanosheets; resulting in geometrically defined edges of both h-BN and B2O3 and inset showing the selected area electron diffraction (SAED) pattern having hexagonal crystalline patterns of h-BN, (c) high-magnification image at the edge of the sheet showing the number of layers of h-BN across to the (002) plane and also the interconnectivity of these sheets, (d) high-resolution TEM image showing hexagonal lattice of synthesized BN–B2O3 composites, and (e) X-ray photoelectron spectroscopy (XPS) represents the chemical states of B and N elements and insets show the typical B 1s and N 1s spectra with corresponding binding energies of 194.21 and 398.4 eV, respectively.

Structural Analysis

XPS shows the distribution of elemental boron (B), nitrogen (N), oxygen (O), and carbon (C) in the samples, Figure e. The major peaks of these elements were present at 194.2, 398.1, 535, and 287 eV. The observation is in close agreement with previous reports.[25,26] Boron, nitrogen, and oxygen 1s-core levels bear the signature of the formation of hBN–B2O3 composites.[27,28] Carbon pick up is from the SPS graphite mold.

The XRD pattern of all hBN–B2O3 composites shown in the Supporting Information Figure S1a, indicates hexagonal h-BN structure with the two major diffraction peaks at 26.56° (002) and 41.46° (100). With increase in B2O3 content peak, broadening is observed and a low intensity peak of B2O3 appears as a shoulder adjacent to the high intensity h-BN peak at 26.56°. The broadening of the peaks confirms a fine-grained microstructure. The intensity of this peak increases with the doping concentration of B2O3. A low intensity B2O3 peak was also observed at 14.34°. The Raman spectrum in Supporting Information Figure S1b shows peaks associated with the B–N and B–O bonds at wave numbers 1417, 1730, and 1961 cm–1, corresponding to B–N stretching vibrations.[19] Differential thermal analysis (DTA) results from Supporting Information Figure S1c indicate the presence of various endothermic as well as exothermic peaks. The peak at 449 °C occurs due to diffusion of B2O3 into h-BN,[29] whereas the other peaks are due to impurity phases and moisture. The endothermic peaks at 164 and 441 °C appear due to boron trioxide and h-BN in the initial mixture. The thermogravimetric (TG) curve indicates weight loss with temperature in the sample. The Supporting Information Figure S2a shows a decrease in surface area with h-BN content, and the values lie between 0.97 and 14.43 m2/g. The evaporation of B2O3 is responsible for the formation of pores and increase in surface area during sintering at high temperatures. The high surface area of ∼14.43 m2/g for the 10BNBO sample was possibly due to the amorphous nature of B2O3.[30] The Supporting Information Figure S2b shows the variation in density of the composites with B2O3 content.

Contact Angle Study

The hydrophilicity of the composites is critical to explain biocompatibility and degradability of the samples, as it controls cell suspension and transfer of body fluids. The contact angles of all of the composites were measured and are shown in Supporting Information Figure S3a–d. It was observed from the figure that the contact angle values increase with h-BN concentration from 53.44° for 10BNBO to 75.65° for 90BNBO.

Mechanical Analysis

The compressive stress–strain curves for pure h-BN, 50BNBO, and 90BNBO samples, indicated by curves 1–3 respectively, are shown in Figure a. The tests were conducted using an Instron universal testing machine (UTM) (Figure a). Digital images of samples before and after fracture are also shown in the inset. The maximum compressive strength of 291 MPa was observed for the sample with composition 90BNBO having a density 1.9 g/cm3; thereby yielding a bone material index of 153. This value is superior to trabecular human bone having density 1.8 g/cm3 and compressive strength of 221 MPa having a material index of 123.[31] This observation fulfills the mechanical criteria for this material as a possible substitute to commercial bone implant materials in use. A comparison of the synthesis method, mechanical properties, porosity, and Brunauer–Emmett–Teller (BET) surface area for porous nanostructured 3D graphene and h-BN synthesized by SPS is listed in Table . It is observed from the table that the 3D h-BN exhibits better mechanical properties in comparison to other reported ceramic biomaterials, viz. HAp,[32,33] tricalcium phosphate,[34] and HAp-based composites.[35] These biomaterials possess lower strength in comparison to the metallic implants produced by the laser engineered net shaping method.[36] This may be due to randomly oriented large h-BN sheets along different crystallographic planes.

Figure 3

Mechanical characterizations of the SPS-sintered BN–B2O3 composites. (a) Variation of σ vs ε. The inset shows the samples before and after fracture test. (b) Load bearing capacity authenticated by placing a 2 kg weight on the sample without damage or fracture. (c, d) Low magnification, SEM image revealing the randomly oriented crack propagation. (e) High-magnification SEM image of fractured BNBO sample reveal in the h-BN sheets interlocking the crack propagation. The inset shows elongated h-BN grains. (f) Variation of specific wear rate with load for 90BNBO composite.

Table 1

Porosity and Mechanical Properties of Different Bone Implants

bone implant materials	density (g/cm3)	porosity (%)	BET surface area (m/g)	compressive strength (MPa)	Young’s modulus (MPa)	references
titanium (porous)	4.5	48		54 ± 5	7.7 ± 2	(29, 36)
HAp	2.1	41	15	34 ± 2		(32, 33)
0.5 MgO–HAp (microwave irradiation)	3.04	∼6	21.87	21 ± 1	126.31	(35)
50BN–50B2O3 (thin film)	1.3	∼36.4	2.500	37 ± 5		(22)
tricalcium phosphate	3.14	50	12 ± 2	11 ± 1.3		(34)
three-dimensional graphene (SPS)	∼1.42	42		40 ± 3	4.1 ± 0.5	(21)
three-dimensional h-BN (SPS)	1.90	∼10	1.326	281 ± 1	63.65 ± 0.5	current work

The microstructural behavior of the fractured h-BN and BNBO samples after compressive loading was also studied. Figure b shows the real-time load bearing capacity of the samples, with a 2 kg block being placed over the sample. The fracture image of the h-BN sample, as shown in Figure c, shows big h-BN sheets along with high amount of porosity with minimal strength and load bearing capacity. On the contrary, the fracture image of the BNBO samples in Figure d,e yields an interconnected microstructure, as also observed from the subsequent simulations studies, leading to their higher strength; in concurrence with the compression plot in Figure a. The predominance of elongated h-BN grains in the fractured microstructure, as seen in the inset of the Figure a, is attributed to the grain growth induced in presence of B2O3 and may lead to an improved toughness in the composites. To validate the mechanical behavior of the synthesized BNBO composites, the wear rate “Ws” was also evaluated, as shown in Figure f. From the figure, the specific wear rate was observed to decrease exponentially with load. The higher wear rate during the initial loading is due to breaking of weaker van der Waals bonds between adjacent h-BN sheets in the h-BN-dominated regions of the matrix; the subsequent decrease in wear rate is due to presence of strong covalent bonding between h-BN and B2O3 in the remaining part of the matrix.

Molecular Dynamics Simulation of 3D Composites

To establish the effect of SPS processing on mechanical property enhancement in the BNBO composites, a molecular dynamics (MD) simulation was performed. The MD was separated into two parts; the first part tries to mimic the SPS process where B2O3 is mixed with h-BN nanosheets and a compressive strain (load) was applied at constant velocity until the structure reduces its original size up to 60%, as shown in Figure a (state 1), followed by the compressive force being slowly released (unload) till the applied force is 0 (state 2). In the second part, the resulting BNBO composite is subjected to a mechanical test through a tensile strain, as shown in the inset of Figure a.

Figure 4

MD simulation of BNBO: (a) compressive strain curve for the BNBO composite formation. The symbols “1” and “2” represent the BNBO mixture at maximum compression (∼60%) and the formed BNBO composite, respectively. The inset shows the required force (tensile strain) to stretch the BNBO composite compared to a pure h-BN structure. (b) Snapshots of the entire simulation process. (1) “Initial”: the initial configuration, a mixture of h-BN/B2O3; (2) “sintered”: the structure at maximum compression (60%); and (3, 4) “strained”: the BNBO composite under tensile strain in two different stages.

The snapshots in Figure a clearly show that the B2O3 and h-BN nanosheets react covalently leading to a 3D interconnected network in the composite. The snapshot “state 1” shows the composite at its maximum strain and “state 2” represents the 3D structure in its final relaxed state. The hysteresis in the compressive strain curve represents the permanent deformation in the BNBO structure. The tensile strain curves in the inset of the Figure a show the force necessary to stretch up the BNBO composite compared to a pure h-BN structure. In the composite case, a larger force is required to pull apart the h-BN nanosheets due to the covalent bonds formed between B2O3 and h-BN during the simulated SPS process, in contrast to the pure h-BN nanosheets, where only van der Waals forces are present along the pulling direction. The snapshots showed in the Figure b represent four different stages of the compressive/tensile strain of the BNBO composite.

In Vitro Cytotoxicity to Mice Calvarial Osteoblast (MCO) Cells

The toxicity study of the different compositions was performed on mice calvarial osteoblast (MCO) cells. The samples showed no significant toxicity on MCO cells at different concentrations ranging from 1 to 100 μg/mL. In Figure a, treatment with 25 μg/mL (∼58%, p < 0.01), 50 μg/mL (∼77%, p < 0.01), and 100 μg/mL (∼69%, p < 0.01) of 90BNBO showed significant proliferation of osteoblast cells as compared to control. However, MCO cells treated with 1 and 5 μg/mL of 90BNBO composite powder showed a fall in cell viability as compared to control. There were no statistically significant changes in proliferation/viability of osteoblast cells with the other samples having higher B2O3 content.

Figure 5

(a) 90BNBO composition increases cell proliferation of mouse calvarial osteoblast (MCO) cells. Cells were cultured in differentiation medium and treated with various concentrations of the 90BNBO samples ranging from 1 to 100 μg/mL for 48 h, and cell viability was assessed by MTT assay. The percent viable cells were calculated compared to untreated cells taken as control. Data represent the mean ± SEM (**p < 0.01 compared to control). (b) Mice calvarial osteoblasts cells were grown in osteoblast differentiation medium. Cultures were maintained for 21 days. At the end of the experiments, cells were stained with alizarin red-S. Upper panel: representative photomicrographs showing mineralized nodules in various groups with or without treatment. Lower panel: alizarin red-S dye was extracted and mineralization quantified spectrophotometrically. Data represent the mean ± SEM (*p < 0.05 and **p < 0.01 compared to control).

Osteogenic Activity Analysis

The osteogenic activity of the composites was investigated by MTT assay.[37,38] MCO cells were cultured for 21 days with and without the sample, fixed in 4% paraformaldehyde, and drops of alizarin red-S was added, which stains the calcium deposition. The 90BNBO sample revealed significant increase in mineralization. Quantification of mineralization yields significant increase in mineralized nodule formation in the presence of 90BNBO. Data from this study suggest that the 90BNBO sample at concentrations 10 nM (p < 0.05), 100 pM (p < 0.05), and 1 pM (p < 0.01) was found to be potent, and showed increase in mineralization by ∼34% (10 nM), ∼53% (100 pM), and ∼91% (1 pM) over the control (untreated cells), as shown in Figure b. However, there were no statistically significant changes in mineralization of osteoblast cells with the other samples having higher B2O3 content. These investigations clearly point out the potency of the current developed composites to be used clinically as bone implant materials.

Conclusions

Three-dimensional nanostructured high strength hBN–B2O3 composites were synthesized using SPS and their microstructure and properties were correlated. A highly interconnected network was observed in the microstructural observations. B2O3 enhances densification and influences the grain orientation in h-BN. Samples could be sintered to high density at an abnormally low temperature of only 250 °C in 2 min possessing a high compressive strength of ∼291 MPa. For investigating the suitability of these materials as possible implant substitutes, cell viability and proliferation tests were investigated. The results show significant effect of 90BNBO composite on cell viability/proliferation. The same composition also shows good mineralized nodule formation over control and suggests its use as an osteogenic agent in bone formation. The substantially high material index (compared to human bone), extremely good mechanical properties, and impressively high cell culture performance makes this composite material a promising candidate for use in knee and bone implants.

Methods and Materials

Synthesis of h-BN Composites

In the present study, a series of composites in the system x[h-BN]–(100 – x)[B2O3] (x = 10, 50, 70, 90 mol %) were prepared using analytical reagent grade chemicals comprising h-BN (Sigma-Aldrich, 99.99%) and B2O3 (Sigma-Aldrich, 99.99%). The appropriate amounts of these chemicals were calculated based on the composition of the composite and they were weighed using an electronic balance (Shimadzu make), mixed in a mortar and pestle using an acetone medium for 6 h, dried, and calcined at 1000 °C in an electric furnace to remove the volatile impurities. Subsequently, these powders were processed in SPS at different sintering conditions. SPS was performed in a Dr Sinter 1050 apparatus (SPS Syntex Inc., Japan). The sintering conditions are as follows: the temperature was between 150 and 250 °C, applied pressure was 20 and 30 MPa, ramp rate was 75 °C/min with 2 min dwell. Samples of 15 mm diameter and thickness 5.95–8.02 mm were prepared.

Material Characterization

XRD was done using Cu Kα radiation in a Rigaku D/Max Ultima II equipment. Elemental analysis for different chemical compositions was carried out using X-ray photoelectron spectroscopy (XPS, PHI Quantera XPS) on a PHI-5000C ESCA system. Thermogravimetric analysis (TGA) and differential thermal analysis (DTA) were carried out in a SDTQ-600 thermal analyzer from room temperature to 400 °C at 5 °C/min under flowing N2. The BET surface area analysis was carried out in a Quanta Chrome Autosorb-3b BET Surface Analyzer. An FEI Quanta 400 SEM was used at 10 kV for microstructure observations. The sample was sputtered with gold film to prevent charging. The TEM images and SAED patterns were obtained in a JEOL 2100 field emission gun transmission electron microscope. Instron universal testing machine (UTM) was used to obtain the load versus displacement curves from the samples at room temperature. For wear test a pin-on-disc-type machine (DUCOM model TR-20LE) was used in the dry condition. Tangential frictional force and wear were monitored and recorded. Specific wear rate was obtained fromwhere, ΔV: volume loss (mm3), ρ: density of sample (g/cm3), P: applied load (N), L: sliding distance (m), m1 and m2 are the weights before and after test.

Contact angle was measured using an Advanced goniometer (500, Rame-Hart, Inc.) under equilibrium conditions.

Cell Culture and MTT Assay

For each experiment of MTT and mineralization, about 5–10 calvaria were collected from new born Bal b/c mice pups. They were removed from the skull surgically; sutures were separated and soft tissues were removed by scrapping. The calvariae were kept together for repeated digestion in dispase and collagenase P enzymes (0.1% each) for release of cells, and the supernatant was discarded. Cells from subsequent three to four digestions were pooled and cultured in α-MEM supplemented with 10% FCS and 1% penicillin/streptomycin at 37 °C in a humidified atmosphere.[39]

The toxicity experiments for all of the synthesized compositions were performed on mice calvarial osteoblast (MCO) cells. The seed density was 2.0 × 103 cells/well in a 96-well plate and cultured as before with and without the samples at different concentrations ranging from 1 to 100 μg/mL. MTT assay was done to check the cell viability.[40] The cells were incubated for 3 h with MTT reagent at a concentration of 0.5 mg/mL, washed with phosphate-buffered saline (PBS), and dimethyl sulphoxide was added for dissolving the formazan.

MCO cells with seed density 10.0 × 103 cells/well in 12-well plates were cultured with and without samples and treated at concentrations ranging from 100 μM to 1 pM for 21 days; the media was refreshed once every 48 h. After experimentation, the cells were washed with PBS, fixed with 4% paraformaldehyde, stained with alizarin red-S (40 mM, pH 4.5) for 30 min, and washed with water to remove the unbound stain.[41] Photographs of stained cells were captured and alizarin red-S stain was extracted using 10% cetylpyridiniumchloride (CPC), which was used to quantify the bound stain. The extracted stain was solubilized in 10% CPC with gentle shaking. The absorbance of the solubilized stain was read at 562 nm using a microplate spectrophotometer. The data were analyzed using one-way ANOVA followed by the Newman–Keuls test of significance with software Graph Pad Prism version 5.0. Probability p < 0.05 was considered statistically significant.

Simulation

MD simulations were carried out using the open source software LAMMPS[42] with the fully atomistic reactive force field called ReaxFF.[43] ReaxFF allows the chemical reactions as bond formation and/or breaking to be accurately estimated by parameterization through experimental data and the density functional theory calculations. The parameterization used in this work is described elsewhere.[44] The structures subjected to the compressive strain were initially equilibrated during 200 ps under NPT at 300 K and no pressure using Nose–Hoover integrator scheme.[45] The simulation box was kept within the periodic boundary conditions along all directions to avoid any border effect. The tensile strain case the periodicity along z direction was removed to allow the outer h-BN sheets to be pulled up through their center of mass. The strain velocity in all cases was kept 0.002 Å/fs and the time step 0.1 fs.