Radiopaque Fully Degradable Nanocomposites for Coronary Stents

Bioresorbable scaffolds (BRS) were introduced to overcome limitations of current metallic drug-eluting stents and poly-L-lactide (PLLA) has been used in the fabrication of BRS due to its biodegradability and biocompatibility. However, such polymers have weaker mechanical properties as compared to metals, limiting their use in BRS. We hypothesized that nanofillers can be used to enhance the mechanical properties considerably in PLLA. To this end, polymer-matrix composites consisting of PLLA reinforced with 5–20 wt% barium sulfate (BaSO4) nanofillers as a potential BRS material was evaluated. Stearic-acid (SA) modified BaSO4 nanofillers were used to examine the effect of functionalization. Rigid nanofillers improved the tensile modulus and strength of PLLA (60% and 110% respectively), while the use of SA-BaSO4 caused a significant increase (~110%) in the elongation at break. Enhancement in mechanical properties is attributed to functionalization which decreased the agglomeration of the nanofillers and improved dispersion. The nanocomposites were also radiopaque. Finite element analysis (FEA) showed that scaffold fabricated from the novel nanocomposite material has improved scaffolding ability, specifically that the strut thickness could be decreased compared to the conventional PLLA scaffold. In conclusion, BaSO4/PLLA-based nanocomposites could potentially be used as materials for BRS with improved mechanical and radiopaque properties.

PLLA has been widely studied as a biomaterial for several biomedical engineering applications due to its biodegradability and biocompatibility [7][8][9] . Particularly, PLLA has been investigated widely as a choice of material in BRS fabrication 10 . Unprocessed PLLA will typically exhibit approximately 100-fold lower tensile modulus than cobalt or stainless steel which are the conventional materials used in DES. As a consequence of the lower modulus (both tensile and compressive), BRS fabricated from these bioresorbable materials may require up to 240% thicker struts in order to match the radial strength of the current metallic DES, thereby affecting the device's deliverability 11 . Polymeric devices also have a limit of expansion and can fracture due to over-dilation. It is important to improve the expandability of the BRS while maintaining radial strength 12 . With lesser strength, BRS requires extensive vessel preparation and may achieve lesser acute gain (defined as the difference between pre-procedural minimal luminal diameter, MLD, and immediate post-procedural MLD) than a metallic stent. Current PLLA devices also lack radiopacity, making the visualization and assessment of scaffold expansion difficult 1,13 .
Nanocomposite polymeric materials present a novel class of materials with important properties in several engineering and biomedical applications. A polymeric nanocomposite comprises nanofillers dispersed within a polymer matrix. The concept of a nanocomposite material capitalizes on the inherent properties of the base polymer while enhancing the functionality of the composite device by the addition of nanofillers 14,15 . The nanofillers are able to render additional features to the polymer that are usually not available in polymeric materials such as optical, electrical and mechanical properties [16][17][18] . In the field of biomaterials, the reinforcement effect of rigid nanofillers on polymers' mechanical properties is of particular interest. Nanofillers possess a large surface to volume ratio that increases the number of particle-matrix interactions when dispersed in a polymer and can improve the overall material properties [19][20][21][22] . The presence of the filler can reinforce the mechanical strength of the polymer by: (i) substituting the softer polymer matrix by a stiffer filler, (ii) immobilizing the polymer molecules on filler particle surfaces as a result of filler-polymer interaction, and (iii) stress transfer from the matrix to the filler 23 . The nanofillers can help to absorb energy from the applied stress and disperse it about a larger volume of the nanocomposite material, thereby improving the composite's properties 24,25 .
The reinforcement effect of the nanofillers on the nanocomposite properties depends not just on the nature of fillers and polymer but also on the filler size, filler-polymer interfacial interaction and particle loading. It is well documented that the major challenge in fabricating nanocomposites is achieving a uniform dispersion and good integration of the nanofillers in the polymer matrix [26][27][28] . Most of the polymers are hydrophobic and incompatible with the hydrophilic nanofillers. The issue of nanofiller agglomeration and weak particle-polymer affinity often lead to detrimental effects on the mechanical strength [29][30][31] . The interactions between the polymer and filler at the interface significantly influence the composite's mechanical properties. Interfacial debonding is commonly the first step of failure and a lack of adhesion between the two phases will lead to phase separation and early material failure 32 . In the same way, a strong interfacial bonding between the nanofillers and polymer leads to an effective transfer of load from the matrix to the nanofillers, enhancing the composite's strength.
To achieve good interfacial adhesion between the fillers and polymer matrix, functionalization is often utilized to improve the particle-polymer affinity. Functionalization of the nanofillers can help to establish favorable interactions that prevent agglomeration of the fillers and improve the overall distribution of the fillers within the polymer matrix [33][34][35] . One method to functionalize the nanofillers is to covalently graft another suitable compound such as a surfactant to the surface of the fillers before addition to the polymer and this has yielded positive mechanical outcomes for bone cement applications 36,37 .
In this study, we hypothesize that the use of PLLA reinforced by the addition of inorganic nanofillers such as barium sulfate (BaSO 4 ) can potentially overcome the drawbacks of current PLLA as a load-bearing biomaterial. BaSO 4 has been used extensively as a radiocontrast agent in X-ray imaging and other diagnostic procedures. Several studies have demonstrated the reinforcement and radiopaque properties of BaSO 4 in polymeric materials. Stearic acid was selected as a functionalizing agent due to its surfactant properties, enabling it to be conjugated to the inorganic fillers while forming targeted hydrophobic interaction with PLLA 38,39 . Hence in this study, we have formulated a nanocomposite material based on inorganic BaSO 4 nanofillers and PLLA, and studied the effect of loading and functionalization on the mechanical properties of PLLA. Finite Element Analysis (FEA) was also carried out to evaluate the effect of the materials using a generic stent design, and the nanocomposite material was employed to improve the scaffolding while reducing the stent thickness potentially.

Results
Effect of filler loading on nanocomposite mechanical properties. The addition of non-functionalized BaSO 4 nanofillers in PLLA has a significant effect on the polymer tensile properties (Fig. 1). As seen from Fig. 1a,b, as the amount of nanofillers increased from 0% to 15%, the modulus and ultimate strength of the material increased, thereby demonstrating the mechanical reinforcement effect of the rigid nanofillers. However, further increment in BaSO 4 loading to 20% caused a significant decrease in the modulus and strength (p < 0.05). The optimal reinforcement results occurred at 15% BaSO 4 loading with a 62% and 300% increase in tensile modulus and strength respectively. The presence of the BaSO 4 nanoparticles caused a significant decrease in the material's elongation at break across all filler loadings.
Tensile Modulus. It has been reported that the tensile modulus of polymers can be improved by adding nanofillers since rigid inorganic particles are much stiffer than polymer matrices. Two models that predict the reinforcing effect of fillers on the matrix have been used in this study to correlate the experimental observations and the reinforcement effect of the nanofillers. The Guth's equation (Equation 1), which is a modified version of Einstein's equation was employed 40 : where E c and E m are the tensile modulus of composite and matrix and V p is the particle volume fraction (obtained based on the method described in Section 2.2). This equation is often used to estimate the effect of fillers on composite modulus as it incorporates: (1) a linear term reflects the stiffening effect of the fillers and (2) the second power term is the contribution of filler interaction, which is higher at increased filler loading 40,41 . According to Equation 1, increasing the particle volume fraction is expected to increase the modulus of the composite. Another model used to predict the effect of spherical nanofillers on composite material's modulus is the Kerner's equation (Equation 2) 42 : where v m is the matrix Poisson ratio and taken to be 0.475 according to Soares et al. 43 . Figure 1c shows the estimated effect of nanofillers loading on the composite's modulus based on the two equations as compared to the experimental data. From Fig. 1c, it is observed that BaSO 4 nanofillers have a reinforcement effect on the PLLA matrix, as increasing the filler loading (up till 15% by weight) increased the modulus. This improvement in modulus is in agreement with the modelling equations used. The results demonstrated that the Guth's model for predicting fillers effect on modulus is closer to the observed experimental data as compared to the Kerner's equation. However, both models do not predict a maximum in the modulus vs. volume fraction curve, which was observed experimentally here. The lesser fit of the Kerner's equation to the data can be attributed to its assumption that there is no interaction between particle-particle or matrix-particle, however, this will not be the case in the BaSO 4 /PLLA system as seen in the agglomeration of the nanofillers 44 .
Tensile Strength. Stress transfer between the fillers and the polymer affects the strength of the material. For nanocomposites with good interfacial adhesion, the applied stress can be transferred effectively from the polymer matrix to the fillers, which helps to improve the strength 41 . Nanofillers may also adversely affect the tensile strength of the composite material by acting as stress concentrators, due primarily to poor interfacial adhesion that results in poor or no stress transfer. There are many different equations and empirical models proposed to predict the effect of fillers on the composite tensile strength, but most models assume poor or no adhesion of nanofillers to matrix, which then predict a monotonic decrease in the composite strength with the addition of the fillers [45][46][47] . Interestingly, none of these models are applicable to this BaSO 4 /PLLA nanocomposite system as the experimental data led to an increase in ultimate tensile strength with the addition of the fillers, passing through a maximum at 15% filler loading (similar to tensile modulus). The increase in tensile strength of the composites upon the addition of the rigid fillers affirmed some form of adhesion between the nanofillers and the matrix. However, increasing the filler loading also led to an increased agglomeration of the fillers in the matrix, which could explain the decreased mechanical properties at 20% BaSO 4 loading. Nanosized fillers such as BaSO 4 have high surface energies, causing them to aggregate in order to lower the surface energies 48 . Such agglomeration is observed in the TEM images of the nanocomposite fibers with different BaSO 4 loading ( Fig. 2). At the highest filler loading (20%), agglomeration is more prevalent as seen in Fig. 2c, leading to uneven distribution of fillers, which can result in phase separation, decreasing the mechanical properties. The size distribution of BaSO4 nanoparticles/clusters in the 20% BaSO 4 /PLLA system is shown in Fig. 2d, more than 50% of the agglomerated the fillers were in the 200-300 µm size range. This observation has been reported in several other nanofillers/polymer composite studies [49][50][51][52] .
Elongation at Break. The elongation at break of the nanocomposite material was significantly reduced when BaSO 4 nanofillers were added to PLLA across all filler loading, indicating a decrease in material ductility (Fig. 1). This significant decrease in elongation at break of the nanocomposite demonstrated that the fillers caused a reduction in polymer matrix deformation due to an introduction of mechanical restraints. It has been reported that at the microstructural level, the volume of the ductile polymer phase may be confined by the surrounding stiff nanofillers phase, which may constrain the local deformation under stress, thereby impairing the elongation at break 53,54 . This restraining effect has been observed in this BaSO 4 /PLLA system. Hence, there appears to be a trade-off between the reinforcement effect and reduction in polymer's ductility when rigid BaSO 4 nanofillers were being introduced into PLLA.
Radiopacity. The results for radiopacity evaluation of the nanocomposites are shown in Fig. 3. It can be seen that PLLA (control) is completely radiolucent and not visible under x-ray imaging. The addition of both non-functionalized and functionalized BaSO 4 conferred radiopacity to the nanocomposite material (Fig. 3a) and the higher the filler loading, the higher the radiopacity value of the sample. From 5-15% filler loading, there were . This result demonstrated that the nanocomposite formulation can be to fabricate implants that are radiopaque, making visualization (using radiography) during procedure possible. Current BRS (with the exception of REVA Medical's FANTOM) have metallic (e.g. gold, platinum) radiopaque markers affixed onto the scaffold for visibility under x-ray. During scaffold deployment, the portion of a BRS with a marker may crack or stretch when stress is being applied, causing the markers to be dislodged in the process. A limitation of the current marker technology is that when viewed under fluoroscopy, the radiopaque markers do not provide good indication of scaffold expansion as they are usually only placed at the distal and proximal ends. During and after the procedure, the operator will not be able to assess scaffold expansion and lesion coverage accurately 1,55 . This also complicate retrieval in case of dislodgment of the scaffold from the delivery catheter. Hence, a material with adequate radiopacity can aid in addressing some of the imaging limitations of current polymeric BRS.
Effect of functionalization on nanocomposite mechanical properties. In this study, conjugation of stearic acid (SA) (less than 1 wt% of the functionalized BaSO 4 ) to the surface of the nanoparticles had significant effect on the nanocomposite material as seen in the stress-strain curves (Fig. 4). In terms of modulus, functionalization did not affect the value significantly except at 15% loading (Fig. 4b). 15% filler loading gave the highest modulus for all the fillers-loaded PLLA, though the SA-BaSO 4 /PLLA had a lower modulus compared to non-functionalized ones. As for tensile strength, using the SA-BaSO 4 resulted in higher strength at lower loading (<10%), after which the non-functionalized BaSO 4 /PLLA gave significantly higher tensile strength (Fig. 4c). One interesting observation is that the SA-BaSO 4 /PLLA nanocomposite material has significantly higher elongation at break across all the loading percentage, indicating an improvement in ductility compared to the non-functionalized system. SA-BaSO 4 /PLLA nanocomposite at filler loading of 15% onwards has a marked improvement in elongation at break compared to pristine PLLA material (Fig. 4d).
Tensile Modulus. The slight decrease of SA-BaSO 4 /PLLA modulus as compared to the non-functionalized BaSO 4 /PLLA can be ascribed to the effect of stearic acid on the filler/matrix interlayer, though the effect was only relatively significant at 15% filler loading. This observation is in agreement with literature, demonstrating that the interaction of fillers with matrix do not influence tensile modulus significantly. Since tensile modulus was determined at low stress, in the linear part of the stress-strain curve, filler de-bonding has not yet occurred and hence should not be significantly affected by functionalization [56][57][58] .
Tensile Strength. On the other hand, mechanical properties such as composite strength are more dependent on the interfacial interaction between fillers and matrix, thus warrants more discussion. The use of SA-BaSO 4 led to better tensile strength at lower loading (<10%) compared to non-functionalized fillers. As the content of filler increases, the amount of lubricating stearic acid is increased accordingly in the interlayer, which may lead to lower interfacial stress transfer efficiency. The presence of stearic acid can lead to a plasticizer effect on the composite system, thereby decreasing the modulus compared to non-functionalized nanocomposites. Stearic acid has been observed to act as a plasticizer during the melt compounding process, which decrease the modulus slightly but improved the elongation at break 59,60 .
This plasticizer effect can be affirmed by the decrease in Tg and increased crystallinity of the functionalized SA-BaSO 4 /PLLA system (Fig. 5). The DSC analysis showed that with the addition of the functionalized SA-BaSO 4 nanofillers, the decrease in Tg of the composite is significantly higher than that of the non-functionalized BaSO 4 / PLLA (Fig. 5a). The decrease of Tg of a polymer is attributed to the additional free volume caused by the plasticizer, which aids in facilitating segmental movement 61,62 . Figure 5b shows the crystallinity of the composites at The results agree with other mechanical studies on stearic acid functionalized nanofillers, where results have showed an improvement in filler dispersion due to lesser agglomeration but a reduced reinforcing effect of the fillers 31,38,39,68 . Experimental observations have demonstrated that the stearic acid modified filler did not improve the tensile modulus and strength of the polymers 69 .
Elongation at Break. The significant improvement in elongation at break of the SA-BaSO 4 /PLLA could be explained in terms of interfacial viscoelastic deformation and matrix yielding (Fig. 4d) 70 . Stearic-acid functionalized nanofillers have been reported to achieve improved dispersion state due to lesser agglomeration. Without functionalization, the nanofillers aggregate due to their van der Waals's bonding alignment and are expected to be in large clusters (Fig. 6a) within the polymer matrix. The interactions between the fillers cause a decrease in the inter-particulate distance [d] and result in clustering of the fillers. This leads to stronger inter-particulate   system suggested that matrix deformation is influenced by both interfacial interaction and dispersion state of the fillers. The decrease in ductility is found to be more pronounced at higher filler loading 20 .
On the other hand, stearic acid functionalization caused weakened filler/filler interactions, favoring a dispersion of smaller aggregates within the polymer matrix, increased the inter-particulate distance (d) between fillers as shown in the schematic in Fig. 6b 56,71 . Weakened interactions amongst the fillers and a smaller aggregate size leads to an increase in the number of fillers taking part in the deformation of the material. These explanations can be supported by the TEM images in Fig. 7 whereby the SA-BaSO 4 /PLLA had smaller clusters of fillers within the nanocomposites as compared to BaSO 4 /PLLA at the same loading percentage (Fig. 2). It can be observed that the SA-BaSO 4 also exhibited aggregation at higher loading, reinforcing the detrimental effect of high filler loading on aggregation and mechanical properties. However, for the functionalized system, the resultant clusters are smaller in size and the fillers still maintained their distinct shape unlike the non-functionalized BaSO 4 , which appeared in larger agglomerated clusters.
Functionalization of nanofillers can be done to: (i) increase the hydrophobicity of the hydrophilic fillers in order to facilitate filler/matrix miscibility due to increased interaction between the two, (ii) prevent agglomeration of the fillers by introducing repulsive forces and (iii) improve the interfacial adhesion between the filler and matrix, thereby promoting a more effective transfer of stress, increasing the strength of the composite. Based on the results in this section, it can be concluded that the use of stearic acid to functionalize BaSO 4 led to decreased agglomeration of the fillers but did not improve the interfacial adhesion between the filler and the polymer, which can be observed in the decrease in composite strength compared to the non-functionalized BaSO 4 . This could be due to the short chain length of stearic acid, which is considered too short to be effectively entangled within the polymer matrix 47,72 . Furthermore, it has been reported that the use of surfactant such as stearic acid has the plasticizer effect which weakens interactions between the fillers and polymer thus facilitating interface debonding 73 . Finite Element Analysis (FEA). Simulation work was done to compare the scaffolding ability of the scaffolds fabricated using different materials. The strut thickness with the SA-BaSO 4 /PLLA system was further modified with decreased thickness to evaluate the potential outcomes of using the nanocomposite material. The expansion and recoil of the 15% SA-BaSO 4 /PLLA scaffold with the tube is shown in Fig. 8a,b, and the PLLA scaffold displayed similar deformation. The 15% BaSO 4 /PLLA scaffold had several locations where the strain reached the elongation limit of the material during expansion, causing scaffold fractures as seen in Fig. 8c,d. The averaged displacement of inner tube surface contacted with the scaffold is 0.16 mm for 15% SA-BaSO 4 / PLLA and 0.08 mm for PLLA stent respectively. The peak MP strain that the 15% SA-BaSO 4 /PLLA BRS experienced during expansion was 15.3% (Fig. 8d), which was far lesser than its elongation limit (Fig. 4d). 15% BaSO 4 / PLLA BRS had lesser peak MP stain (13.9%) compared to the functionalized BRS and a similar strain distribution, but its peak strain is near to its the elongation limit. The expansion and scaffolding of the 15% SA-BaSO 4 / PLLA BRS with reduced strut thickness (100 μm) is shown in Fig. 9a,b and the averaged tube displacement was 0.37 mm. During expansion, it had a peak MP stain of 24.9% (Fig. 9c), which is still below the elongation limit of the material.
To compare the influence of material properties on the BRS function, the FEA work evaluated three scenarios based on different stent materials and the results were aligned with the earlier mechanical testing outcomes. 15% BaSO 4 /PLLA BRS experienced fracture due to the low ductility of the material. On the other hand, the PLLA BRS barely survived the scaffold expansion with a scaffolding ability just half of the 15%SA-BaSO 4 /PLLA (considering the averaged displacement of inner tube surface contacting the scaffold).
Based on this simulation result, the 15% SA-BaSO 4 /PLLA BRS demonstrated reliable structure integrity during expansion and better scaffolding ability than PLLA and 15% BaSO 4 /PLLA. With enhanced material properties, the FEA data also showed that BRS fabricated from the functionalized nanocomposite material could potentially increase scaffolding by 130% (0.37 vs 0.16 mm) while reducing the thickness by 33% (0.1 vs 0.15 mm). Although the peak strain was increased by more than 60% (0.249 vs 0.153), the BRS still maintained its integrity and this could be attributed to the improved elongation limit of the SA-BaSO 4 /PLLA system. Following the clinical lessons learnt from the BVS studies, there has been a concerted drive towards minimizing strut thickness. The newer generation of BRS has moved away from the original 150 μm design and has gone to lower than 100 μm. Thinner struts decrease protrusion and improve embedment of the struts, hence is expected theoretically to contribute to less flow disturbance compared to the thick struts of the current BVS, and possibly better endothelialization. Studies have shown that thinner strut BRS has better embedment of struts and lesser  alteration of physiologic shear stress, thereby enhancing re-endothelialization 74,75 . The FEA work has provided important insights into the stress distribution and potential strut fractures of the BRS when deployed, and shown some of the potential mechanical advantages of the nanocomposite material in these aspects. This simulation can also aid in understanding the effect of strut thickness reduction on the expansion limit of the proposed nanocomposite materials.

Conclusion
The mechanical properties of a nanocomposite BaSO 4 /PLLA material were modified using functionalization. The addition of rigid fillers such as nanosized BaSO 4 had a reinforcement effect on the polymer as evidenced by the increase in tensile modulus and strength of the material, but ductility is decreased. The optimized formulation (15% BaSO 4 /PLLA) had an approximate 60% and 110% increase in tensile modulus and strength respectively, followed by 45% decrease in ductility. Functionalization of BaSO 4 with stearic acid was shown to have decrease the agglomeration of the nanofillers, thereby improving the elongation at break of the composite significantly. Functionalization did not increase the modulus and strength of SA-BaSO 4 /PLLA as compared to BaSO 4 /PLLA (p > 0.05). Stearic acid as functionalizing agent was found to exert a plasticizer effect on PLLA, as evidenced by the DSC data and mechanical characterization. The BaSO 4 -filled composites were also radiopaque and could be visualized under x-ray. FEA demonstrated the potential benefit of using the nanocomposite material as a BRS candidate since the improved mechanical properties allowed for reduced strut thickness while maintaining structural support. Future work will address some of the limitations of the current study in evaluating the material as a potential BRS candidate. For example, the nanocomposite material will be extruded into different thicknesses and laser-cut into a BRS prototype for further characterization such as radial testing and overexpansion evaluation. More FEA work will have to be conducted to evaluate other parameters such as different scaffold designs. The degradation rate and biocompatibility of the BRS will also need to be studied in order to understand its long-term behavior.
The use of polymeric composites reinforced with rigid fillers present an effective way to enhance the mechanical properties of the material for load-bearing devices. The fate and clearance of the fillers in the body after polymer degradation remains an area of scrutiny. BaSO 4 is considered a poorly soluble low toxicity particle and its "particokinetics" is influenced by the particle size and route of exposure 76 . While several studies have been conducted on the clearance and toxicity of inhaled and intravenous (IV) injected BaSO 4 nanoparticles, little data is available regarding the fate of implanted BaSO 4 77-79 . It was found that for injected BaSO 4 particles (~300 nm), fecal excretion was the dominant elimination pathway in a rat model. IV-injected BaSO 4 was studied to understand the fate of circulating BaSO 4 nanoparticles in the organs and it was demonstrated that a low concentration in the organs was achieved after 7 days in the animal model 77 . Previous work by Lämsä et al. had also explored the toxicity of a BaSO 4 /PLA stent in a rat model, reporting no adverse effect after 21 days 80 . More work has to be done in this area to better understanding on the clearance of the BaSO 4 fillers after the degradation of the nanocomposite.
Presently, both the polymer-based and magnesium-based BRS platforms remain limited by their large profile and strut thickness, as compared to metallic DES. Research is being done to improve the properties of bioresorbable materials in order to reduce the strut thickness as seen from newer generation of BRS (e.g. Amaranth Medical's FORTITUDE and APTITUDE). In conclusion, BaSO 4 /PLLA-based nanocomposites are good potential candidate materials for BRS with more desirable mechanical and radiopaque properties compared to PLLA alone.

Materials and Methods
Fabrication of nanocomposite polymeric material. PLLA (PURASORB PL inherent viscosity midpoint = 8 dl/g, Lot number: 0404002128) was purchased from Corbion (Netherlands) and employed as the base polymer. Non-functionalized BaSO 4 nanoparticles is a product of Nanoshel (USA). The formulation of different nanocomposite samples fabricated in this study is shown in Table 1. The samples were fabricated using a twinscrew (Xplore Micro 5cc) microcompounder (Netherlands) and the polymer resin and nanoparticles were compounded at a temperature of 190 °C for 7 minutes before extrusion. The extrudates were cut into granules and fed back to the compounder for a second round of mixing for 5 minutes before being extruded. The nanocomposite fibers obtained from the extrusion were 180 ± 10 µm and were used for further testing and characterization.
Transmission electron microscopy (TEM). TEM (Libra 120 Plus, Carl Zeiss, Germany) was employed to examine the dispersion of the nanoparticles within the nanocomposite material. The extruded composite fibers were embedded in araldite epoxy resin (Ted Pella, USA), cured at 60 °C for 24 hours and sectioned into 100 nm slices using ultramicrotomy. The slices were floated on a Formvar-coated copper TEM grid from deionized water and visualized using the TEM. The volume fraction of the nanocomposite material was determined from the automatic threshold method using Image J software. In this method, color-based thresholding was used to segment the fillers from the polymer matrix. The "Analyze" tool from the software will calculate the integrated densities (Area (No. of pixels) x Total Intensity of the pixels) of the different regions. This provides an estimation for the volume fraction of the fillers in the nanocomposite material. The volume fraction of the fillers is presented as an average of 10 TEM images per sample and 3 samples were used for each formulation. Image analysis was also performed using the same software to determine the size distribution of the agglomerated BaSO 4 nanoparticles within the PLLA matrix. A total of 120 nanoparticles were measured.

Differential scanning calorimetry (DSC).
A Perkin-Elmer DSC 8000 was used to study the thermal properties at a heating rate of 10 °C/min. The samples were cut into smaller pieces to improve contact with the sample pan. The Tg was determined as the point of half heat capacity extrapolated and the melting temperature (Tm) was determined as the peak temperature. The crystallinity of the material was calculated from the fusion enthalpy of melting (ΔH m ), which was obtained from the DSC melting curve. Based on literature, the theoretical heat fusion of 100% crystalline PLLA was employed to be 93.1 J/g. The crystallinity of the PLLA in this work was calculated by 81 : where ΔH m is the enthalpy of fusion for the sample in J/g. an W p is the weight fraction of PLLA in the composites. Triplicates were done for each sample to ascertain the reproducibility of the results.
Mechanical testing of nanocomposite polymeric material. All nanocomposite materials were subjected to tensile testing using an MTS C42 Test System (MTS Systems, MN, USA). A 50 N load cell was used for the testing and all samples were pulled at a constant crosshead speed until failure. The extension rate was set at 0.1 mm/min and the stress-strain curves were obtained using the MTS software. The tensile modulus, the ultimate tensile strength (UTS) and elongation at break were obtained from the stress-strain curves. The values were averaged from 10 samples for each sample. The test specimens (nanocomposite fibers) were all standardized to be 180 ± 10 µm in diameter and 40 mm in length (for the testing length) after being mounted onto the MTS.
Radiopacity measurement. For radiopacity evaluation, five samples per concentration (extruded fibers with 1.00 ± 0.05 mm in diameter and 20 ± 2 mm in length) were used. The samples were placed adjacent to a calibrated aluminium step wedge (Biomedia, Singapore) with 3.2-mm increments and imaged according to the adapted protocol 82 . A standard X-ray machine (Philips Clarity FD20) was used to irradiate X-rays onto the specimens using an exposure time of 4 ms at 76 mA and a cathode-target film distance of 100 cm. The tube voltage was set at 50 ± 5 kV. The radiographs were processed (AGFA Enterprise Imaging System), and a digital image of the radiograph was obtained. The grey pixel value on the radiograph, of each step in the step wedge was determined using an imaging programme, ImageJ (NIH, USA). Numbers between 0 and 255 with 0 representing pure black and 255 pure white were assigned accordingly.  Statistical analysis. All mechanical tests were done in 10 repetitions for each type of nanocomposite samples and numerical data were analyzed using standard analysis of variance (ANOVA) technique and statistical significance was considered at p < 0.05.

Finite Element Analysis.
In this study, the simulation involved the implantation of a scaffold into a silicon rubber tube to examine if the device can maintain its integrity during expansion. The whole model employed in this simulation is shown in Fig. 10a,b. The generic scaffold design has a length of 8 mm, an outer diameter of 1 mm, a thickness of 0.15 mm (150 μm) and strut width of 0.9 mm with six rings connected by two links in-between. The material properties of the three materials (PLLA, 15% BaSO 4 /PLLA and 15% SA-BaSO 4 /PLLA) were assigned to the scaffold model for three simulation scenarios. The silicon rubber tube has a thickness of 0.1 mm, an inner diameter of 1.8 mm and a length of 20 cm (E = 1.5 MPa and v = 0.49). The rigid balloon has a diameter of 0.7 mm and a length of 10 mm. The modified scaffold design (Fig. 10c) has a thickness of 0.1 mm (100 μm) and width of 0.15 mm (the corresponding balloon has an increased diameter of 0.8 mm). The scaffold and tube model were meshed with hexahedral solid elements with rigid balloon surface elements. A sensitive analysis has been applied to ensure the elements density is enough. Two interactions, between the inner scaffold surface and rigid balloon, and between outer scaffold surface and inner tube surface, were established to simulate the contacts between scaffold and rigid balloon, and between scaffold and tube.
Quasi-static simulation was carried on using the finite element code ABAQUS/Explicit (Dassault Systèmes Simulia Corp., Providence, RI). In each of the four scenarios, the rigid balloon was expanded until the outer stent diameter reached 3 mm, followed by the recoiling of the balloon to its original shape, leaving the BRS supporting the tube by itself. For boundary conditions, the tube was fixed in all directions at its two ends and the rigid balloon can only expand and recoil in radial direction.
A material fracture mechanism was considered in the simulation to evaluate if the materials could endure the severe deformation during scaffold expansion. When the strain of a scaffold element reaches the elongation limit of the material assigned, the element will be considered as having failed, and therefore deleted from the model. For the scenarios without scaffold element failure, the averaged displacement of the inner tube surface contacted with the scaffold was calculated to evaluate the scaffolding ability of the scaffold; meanwhile, the maximum principal (MP) strain of the scaffold during expansion was also evaluated. As required by the FEA code, the strains applied in the simulation and shown in FEA results were true strains (logarithmic strains, LE).

Data Availability
The data that support the findings of this study are available within the paper.