Synthesis and Characterization of Hierarchical Mesoporous- Macroporous TiO2-ZrO2 Nanocomposite Scaffolds for Cancellous Bone Tissue Engineering Applications

Bone tissue engineering has been introduced several decades ago as a substitute for traditional grafting techniques to treat bone defects using engineered materials. The main goal in bone tissue engineering is to introduce materials and structures which can mimic the function of bone to restore the damaged tissue and promote cell restoration and proliferation. Titania and zirconia are well-known bioceramics which have been widely used in tissue engineering applications due to their unsurpassed characteristics. In this study, hierarchical meso/macroporous titania-zirconia (TiO2-ZrO2) nanocomposite scaffolds have been synthesized and evaluated for bone tissue engineering applications. The scaffolds were produced using the evaporationinduced self-assembly (EISA) technique along with the foamy method. To characterize the samples, X-ray diffraction (XRD), scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), simultaneous thermal analysis (STA), and Brunauer-Emmett-Teller (BET) analysis were performed. The results showed that TiO2-ZrO2 scaffolds can be produced after sintering the samples at 550C for 2 h. Among samples with different weight percentages of zirconia and titania, the sample containing 13wt.% zirconia was considered as the optimum sample due to its structural integrity. This scaffold had pore size, pore wall size, and mesopores in the range of 185 ± 66 μm, 15 ± 4μm, and 7-13 nm, respectively. The specific surface area obtained from the BET theory, total volume, and mean diameter of pores of this sample was 13.627mg, 0.03788 cmg, and 11 nm, respectively. The results showed that the produced scaffolds can be considered as the promising candidates for cancellous bone regeneration.


Introduction
In recent decades, a great deal of effort and attention has been focused on designing scaffolds which are biologically, chemically, physically, mechanically, and structurally carefully matched to that of natural bone . Although each type of bone tissue has specific requirements which must be met, there are some features in common that should be considered in designing bone scaffolds to present the exact function of host tissue [24,25]. Generally, bone has a complex meso/macroporous hierarchical structure [24][25][26][27][28][29] with multisized inter-connected pores [30,31]. One of the main issues is to synthesize scaffolds with a hierarchical porous structure with pore interconnectivity to mimic the ECM of the host tissue. Different pore sizes play a remarkable role in enhancing cell viability and osteogenesis and finally in therapeutic effects for bone tissue engineering applications [32][33][34][35][36]. The macropore size of larger than 150 μm is required for cell accommodation and proliferation and also for vascularization [32,[37][38][39][40][41][42]. Mesopores with the size of 2-50 nm are also required to insert the nutrients and release the biological agents, to remove waste materials, and also to enhance the surface activity and bioactivity [37,[40][41][42].
According to the literature, the sol-gel method is a facile way to synthesize metal oxide nanoparticles with high quality and purity at low temperatures. In this technique, precursors are dispersed in an aquatic or alcoholic environment at temperatures lower than 100°C. Through the hydrolysis and condensation, aggregation in the obtained colloidal solution which is called "sol" causes to make "gel" with a 3D network of M-O-M or M-OH-M units. Unlike other techniques such as the solid-state method, the sol-gel procedure requires low temperatures and very fine powders in which a homogeneous composition can be obtained. However, there are some disadvantages as well which should be considered such as a relatively high cost of initial precursors and the formation of cracks during the heat treatment cycles to remove the organic materials [50].
On the other hand, the evaporation-induced selfassembly technique (EISA) is a successful, simple, fast, and efficient method for the preparation of mesoporous metal oxides [51,52]. The existence of hydrophobic and hydrophilic components in the surfactants causes the formation of micelle-like structures that are removed from the system by burning the hydrocarbon chains which leaves pores behind. This technique was first used to synthesize the silica mesoporous structure [53][54][55] and gradually was incorporated with the sol-gel method by using surfactants for other metal oxides [56]. Among many kinds of surfactants, the nonionic Pluronic F127 formed a well-organized mesopore structure [57].
TiO 2 -ZrO 2 nanocomposite has been used in different applications. For example, Fu et al. [58] and Yu et al. [59] worked on its photocatalytic properties, individually. They reported that this nanocomposite shows better photocatalytic behavior compared to titania or zirconia, separately. The bioactivity of titania-zirconia nanocomposite has been studied by Marchi et al. [60]. The results showed an increase in apatite formation ability which indicates its good bioactivity in the biological environment. Also, this composite has been used as a 3D scaffold and coating. Tiainen et al. [61] investigated the mechanical properties of titania-zirconia nanocomposite scaffolds with different weight percentages of zirconia. The results showed that increasing the amount of zirconia in the composite decreased the mechanical strength of the scaffolds. Titania coating on its alloys is widely used in dental implants and hip prosthesis due to their excellent biocompatibility. According to the literature, TiO 2 has excellent biocompatibility, low toxicity, good corrosion resistance, and low density [62][63][64]. On the other hand, zirconia has several upsides over other bioceramics. Zirconia is bioinert in vivo and in vitro [65][66][67]. In vitro tests have shown that zirconia has lower toxicity than titanium oxide [68]. Additionally, cytotoxicity and carcinogenicity have not been reported [68]. Besides, it has good mechanical properties such as corrosion and abrasion resistance. Furthermore, it increases the crack self-healing potential of scaffolds by its transformation toughening mechanism [69,70]. The strength of interfacial bonding between substrate and coating will be improved by the help of adding ZrO 2 . Also, zirconia enhances apatite formation due to increasing the hydrophilicity of the substrate [71].
In the present study, for the first time, nanocomposite ZrO 2 /TiO 2 scaffolds were produced using the EISA technique combined with the foamy method. Interconnected macropores were prepared by polyurethane (PU) (60 pores per inch (ppi)) sponge, and mesopores were made using Pluronic F127 surfactant by the help of a self-assembly technique in order to simulate the natural bone hierarchical porous structure. Structural features, chemical composition, and porosity features of the ZrO 2 /TiO 2 nanocomposite scaffolds were evaluated. Furthermore, the formation mechanism of the produced scaffolds was evaluated as well.

Sample Preparation.
In this study, the sol-gel technique was used to fabricate TiO 2 -ZrO 2 nanocomposite scaffolds. A schematic representation of the sol-gel method is shown in Figure 1. To prepare the samples, titania and zirconia solutions were produced. Titania solution was prepared according to our previous study [72]. Briefly, 5 ml titanium (IV) butoxide (C 16   2

Journal of Nanomaterials
Based on the previous study [73], to prepare zirconia solution, 3.65 ml zirconium (IV) butoxide (Zr(OC 4 H 2 ) 4 , 80 wt.%, Sigma-Aldrich) and 0.63 gr citric acid (C 6 H 8 O 7 , Merck) were dissolved in 125 ml absolute ethanol and stirred by a magnetic stirrer until a clear solution was obtained. Then, a combination of 2 gr PF127, 125 ml absolute ethanol, and 6.67 ml hydrochloric acid was gently added to the zirconia solution, stirred for 24 h with the speed of 1000 rpm at ambient temperature, and then aged for 48 h at a location with 40% humidity. Samples with different weight ratios of titania and zirconia solutions were prepared to investigate their effect on the sample integrity and microstructure. Table 1 shows the designation and specification of different samples.
PU sponge (60 PPI) with a mean pore size of 300-700 μm was used to make blocks with a dimension of 10 mm × 10 mm × 10 mm. The blocks were dipped in different solutions with various ZrO 2 to TiO 2 weight ratios for 1 h. The saturated PU blocks with the solution were squeezed by a hand roller. These blocks were then dried at room temperature for 72 h. The dried foams were sintered at 550°C for 2 h with the heating/cooling rate of 5°C min -1 . Figure 2 illustrates the preparation steps of different samples.

Sample Characterization.
In order to assess the thermal decomposition temperature of PU foam and volatile materials and to optimize the calcination temperature of materials used to make the solutions, a simultaneous thermal analysis (STA) test was performed up to 550°C with the heating rate of 5°C/min, using the NETZSCH STA 449F3 machine.
To investigate the morphology and to measure the pore size of the scaffolds, scanning electron microscopy (SEM) was applied using SEM/EDS, Zeiss, Germany machine (FEI Quanta 200 SEM) with the working distance in the range of 9.5-13.8 μm and the voltage in the range of 15 to 25 kV. To evaluate the elemental compositions of samples, energydispersive spectroscopy (EDS) was utilized. Map scan analyses were performed to determine the elements formed in randomly selected locations of the scaffolds using the same machine. Small-angle X-ray scattering (SAXS) was performed using the Asenware AW-DX300 diffractometer to approve the formation of a mesoporous structure. SAXS traces were recorded with 2θ in the range of 0.5-10 degrees with a Cu Kα radiation source with λ = 0:154184 nm. Wide-angle Xray diffraction (WAXRD) analysis with 2θ in the range of 10-90 degrees was used to study the crystalline phase of samples using the same machine. The nitrogen adsorption-desorption isotherms were used at 77 K using the BELSORPmini II analyzer to determine the pore structure and to assess the specific surface area.

Results and Discussions
3.1. STA Evaluation. To assess the weight loss and to determine the temperature of reactions that occurred during the sintering procedure of TiO 2 -ZrO 2 scaffolds, simultaneous thermal analysis (STA) was utilized. Different samples ( Table 1, SS which contains only solution, SF which contains only PU foam, and SSF which contains a solution with PU foam together), 3 samples for each, were used to accurately determine the effect of each component on the STA traces. Figure 3 shows the results of the thermogravimetrydifferential scanning calorimeter (TG-DSC) curves of different samples. The TG curve of the SF sample (contains only PU foam) showed that foam is burned out at a temperature of around 310 ± 2°C (Figure 3(a)) [74]. The results proved that at the sintering temperature, no sponge was left in the structure. Figure 3(b) shows the STA analysis of the SS sample which only contains solution A (see Table 1). The results showed that there is an exothermic peak at 301.6°C. The TGA result also showed that this exothermic process accounted for a substantial amount of the total weight loss on the whole decomposition process from about 50°C to around 300°C. The exothermic peak and the associated weight loss were due to the decomposition of organic materials and their removal from the system, as proved by the EDS analysis of the sintered scaffolds, at 550°C in which no organic material was detected ( Figure 4).
The exothermic peak and subsequent mass change at around 300°C were in good agreement with the reported results of previous studies. In other words, the range of around 171.40-301.37°C has been considered as the decomposition temperature of Pluronic F127. The decomposition temperature of zirconium butoxide, acetylacetone, titanium butoxide, and citric acid was reported to be around 176.85°C, 280°C, 122°C, and 19.85-399.85°C, respectively [72,[75][76][77][78][79][80].
Also, Figure 3(c) pertains to the SSF sample, the one with both foam and solution. It can be seen that foam was removed from the system at 283:6 ± 2°C. The exothermic peak occurred at 434:9 ± 1°C could be ascribed to the oxidation of titanium and zirconium. Based on the results of Figure 3, the calcination temperature was adjusted to 550°C to ensure that at the end of the heat treatment cycle, no foam Journal of Nanomaterials and organic additives were left in the scaffolds and only titanium and zirconium oxides were present in the structure.

SEM/EDS Evaluation.
The morphology and pore size of the scaffolds are extremely important to achieve desirable mechanical properties and to have effective cell adhesion, cell proliferation, and strong cell growth into the scaffolds. The optimized sample in this study was known as the one with structural integrity without any cracks. To investigate the morphology of pores that have been formed through the foamy method after heat treatment and their size, samples were studied with SEM. Figure 5 shows the structure, pore morphology, pore size, pore distribution, mesopores, and finally pore wall of a random macropore of S 13Zr sample which was annealed at 550°C for 2 h. On the other hand, Figure 6 shows the structure of other samples with different amounts of zirconia that were annealed in the same condition. As can be seen in Figure 6, all samples have some structural defects such as cracks on the wall of pores (shown by arrows). The high amount of zirconia can be the reason for the existence of cracks in the scaffolds which makes them completely useless for bone tissue engineering applications [61]. However, as can be seen in Figure 5, S 13Zr sample (containing 13 wt.% zirconia and 87 wt.% titania) exhibits an integrated structure with no crack, appropriate pore size for cell accommodation, cell growth, and cell proliferation. Therefore, S 13Zr sample was considered as the optimized sample for further studies. Figure 7 shows the histogram of the macropore size in the S 13Zr sample. As it is obvious, the formation of appropriate macroporosity with a desirable size (more frequency in pore sizes measured greater than 150 μm that is beneficial for cell accommodation, penetration, and differentiation) in the produced scaffold can provide a suitable environment for the vascularization which plays a vital role in the success of the scaffold. EDS was utilized to reveal the average local chemical composition in the produced scaffold. Figure 4 shows the EDS results of the S 13Zr sample. A homogeneous distribution of Ti and Zr elements was observed in the EDS results of the optimized scaffold. Figure 8 shows the results of the EDS elemental mapping of the S 13Zr sample. The results showed a homogeneous distribution of Ti and Zr in this sample which demonstrate the effectiveness of the fabrication technique to form a uniform structure.

XRD Evaluation.
The phase transformation of the S 13Zr sample was investigated by XRD. Figure 9(a) illustrates the wide-angle X-ray scattering (WAXRD) pattern. The presence of metal oxide formed at the low temperature was successfully proved by the XRD patterns. As can be seen in Figure 9(a), the sample has not been completely crystallized due to the low temperature used for heat treatment. The low annealing temperature was used to protect and retain the mesopores of the sample because the mesopores will collapse    Journal of Nanomaterials during a high temperature of sintering [81]. On the other hand, the mesoporous structure transforms the long-range order of crystals into the short-range order. So the XRD pattern of mesoporous structures is similar to semicrystalline structures [81]. An intensive peak can be seen in Figure 9(a) in 2θ equals to 25 degrees that shows the formation of the anatase phase (XRD JCPDS data file No. 01-071-1169) in the structure. Titania and zirconia can form different crystal polymorphs at different temperatures [82,83]. According to their phase diagram and the calcination temperature, the anatase phase for titania and monoclinic phase for zirconia should be stable at room temperature [84].
Based on the literature, the monoclinic phase of zirconia is thermodynamically its most stable phase; however, the less stable phases of zirconia (cubic or tetragonal) could form at higher temperatures. In other words, according to the zirco-nia phase diagram, the monoclinic phase of zirconia is stable from ambient temperature to 1170°C. The tetragonal phase is stable from 1170°C to 2370°C, and from 2370°C to the melting point cubic phase of zirconia will be stable [83].
On the other hand, the most stable phase of titania is rutile; however, according to previous studies, the formation of short-range order TiO 6 in anatase is easier than the formation of long-range order TiO 6 in rutile [82]. Also, thermodynamically, the level of surface free energy of anatase is lower than that of the rutile phase which causes that TiO 2 -based nanocomposites and nanofilms contain anatase phase at room temperature [82].
Some elements such as Ca, scandium, strontium, yttrium, niobium, barium, lanthanum, aurum (gold), boron, aluminum, silicon, phosphorus, sulfur, chlorine, cerium, neodymium, samarium, europium, gadolinium, terbium, dysprosium,  Table 1); (c) simultaneous thermal analysis of SFF sample.   Figure 5: Scanning electron microscopy (SEM) images of sample S 13Zr as the optimized sample (S 13Zr ) sintered at 550°C for 2 h: (a) the general shape of the scaffold which shows the favorable structural integrity, (b) acceptable pore morphology and pore size (no crack can be observed), (c) desirable size of mesopores which are the best space to insert nutrients to be used by cells in order to proliferate and grow into the scaffold, and (d) strong pore wall which exhibits nice interconnectivity and strong structure for each macropore to mimic the natural bone tissue structure. 6 Journal of Nanomaterials holmium, erbium, thulium, ytterbium, and specifically zirconia have an inhibitor role against phase transformation from anatase to rutile in titania [4,82,85]. In this study, Zr has been used as the second phase of the nanocomposite. In fact, the addition of zirconia to titania stabilizes the anatase phase of titania and inhibits its grain growth [86]. Figure 9(b) shows the small-angle X-ray spectroscopy (SAXS) pattern of sample S 13Zr which was sintered at 550°f or 2 h. In the SAXS analysis, the intensity and sharpness of the peak is a sign of the formation of a well-ordered mesoporous structure [87,88]. SAXS was performed in the 2θ range of 0.5-10 degrees. As can be seen in Figure 9(b), a sharp peak in 2θ equals to 0.7 degrees can be observed which indicates the formation of well-ordered mesoporous structures in this sample. The results obtained from the SAXS analysis were in good agreement with those obtained from the SEM observation ( Figure 5(c)).
In general, TiO 2 formation occurs in three different steps. As mentioned, the sol-gel method through the alkoxide route was used to prepare TiO 2 and ZrO 2 solutions. At the first step, the aqueous environment induces alkoxide hydrolysis and produces titanium hydroxide and zirconium hydroxide and alcohol (alkoxide group hydroxide). The second step can occur through the interaction between either two hydroxide metals or a hydroxide metal and an alkoxide metal that finally produce A-O-A, in which A refers to the metal (Zr or Ti) and O refers to the oxygen, and water or alcohol in an aqueous and alcoholic environment [89][90][91]. TiO 2 and ZrO 2 particles act as the nuclei to expand metal oxides formed during the synthesis procedure. The steps of the formation of TiO 2 can be expressed as follows [90]: Figure 6: Scanning electron microscopy (SEM) images of (a) sample S 26Zr , (b) sample S 50Zr , (c) sample S 76Zr , and (d) sample S 86Zr which are completely inappropriate for cells to attach and grow in. It seems that the different amount of zirconia and titania has made heterogeneity in their structure which results in forming cracks and weak pores and pore walls and also the high amount of zirconia in the samples has a detrimental role in structural integrity.
(2) Condensation of hydrolyzed species: Equation (2) can also be expressed as follows: (3) Growth of ZrO 2 sol particles Figure 10 shows the schematic representation of the aforementioned steps clearly in detail.
The nuclei growth phenomenon stems from the incredible high surface area of particles and consequently their high   Journal of Nanomaterials amount of energy which makes them thermodynamically unstable. For this purpose, metal oxide particles (as the nuclei) start to grow through Ostwald ripening or oriented attachment mechanism or even both, to decrease their energy level. According to Ostwald ripening phenomenon, small particles dissolve in the solution because of their high energy level and bigger particles get bigger and bigger in regular spherical shapes by attracting small particles. Based on the oriented attachment mechanism, small particles get together and aggregate at the first step; then, they will be considered as a specific place for other small particles to attach and consequently form crystals [92]. Linear F127 polymer that was used as the template for mesopores has hydrophilic and hydrophobic parts. After dissolving in ethanol, F127 forms special structures that usually transforms into regular spherical micelles in solution. Finally, drying and calcination cause the formation of mesostructure [93]. Figure 11 shows the mechanism of mesopore formation. Figure 12 demonstrates the nitrogen adsorption-desorption isotherms of the S 13Zr sample after sintering at 550°C for 2 h. The specific surface area of the S 13Zr sample measured by the Brunauer-Emmett-Teller (BET) theory was 13.627 m 2 g -1 ; the total porosity volume obtained from the adsorption curve of the isotherm diagram and mean pore diameter were 0.03788 cm 3 g -1 and 11.119 nm, respectively.

Conclusions
In this study, hierarchical meso/macroporous hierarchical titania-zirconia nanocomposite was fabricated for tissue engineering applications. Different samples with different weight percentages of titania and zirconia were prepared by the help of the EISA technique coupled with the foamy method. The foamy method was applied to provide the macropore template, and the EISA method was utilized to  Figure 9: (a) Wide-angle XRD of the optimized sample contains 13 wt.% ZrO 2 and 87 wt.% TiO 2 heated at 550°C for 2 h with the heating/cooling rate of 5°C min -1 . (b) Small-angle XRD of optimized meso/macroporous titania-zirconia sample heated at 550°C for 2 h. 9 Journal of Nanomaterials provide a mesopore structure. The results showed that scaffolds with a lower amount of zirconia had less structural defects. This scaffold had pore size, pore wall size, and mesopores in the range of 185 ± 66 μm, 15 ± 4 μm, and 7-13 nm, respectively. The specific surface area obtained from the BET theory, total volume, and mean diameter of pores of this sample was 13.627 m 2 g -1 , 0.03788 cm 3 g -1 , and 11 nm, respectively.

Data Availability
The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study. If you need to have access to some of the raw data, please contact the corresponding author.

Additional Points
Highlights. (1) ZrO 2 /TiO 2 nanocomposite scaffolds were synthesized using the EISA and foamy method. (2) The formation mechanism of the scaffolds was scrutinized. (3) The influence of the ZrO 2 on the integrity of the scaffolds was evaluated. (4) Nanocomposite scaffolds consisting of 13 wt.% ZrO 2 showed an integrated structure    Figure 12: BET analysis of the S 13Zr scaffold as the optimized sample. 10 Journal of Nanomaterials