Comparison of Methods for Surface Modification of Barium Titanate Nanoparticles for Aqueous Dispersibility: Towards Biomedical Utilization of Perovskite Oxides

Abstract

Colloidal perovskite barium titanate (BaTiO₃, or BT) nanoparticles (NPs), conventionally used for applications in electronics, can also be considered for their potential as biocompatible computed tomography (CT) contrast agents. NPs of BT produced by traditional solid state methods tend to have broad size distributions and poor dispersibility in aqueous media. Furthermore, uncoated BT NPs can be cytotoxic due to leaching of the heavy metal ion, Ba²⁺. Here we present and compare three approaches for surface modification of BT NPs (8 nm) synthesized by the gel collection method to improve their aqueous stability and dispersibility. The first approach produced citrate capped BT NPs that exhibited extremely high aqueous dispersibility (up to 50 mg/mL) and a small hydrodynamic size (11 nm). Although the high dispersibility was found to be pH dependent, such aqueous stability sufficiently enabled a feasibility analysis of BT NPs as CT contrast agents. The second approach, a core/shell design, aimed to encapsulate BT nanoaggregates (BTNA) with a silica layer using a modified Stöber method. A cluster of 7–20 NPs coated with a thick layer (20–100 nm) of SiO₂ was routinely observed, producing larger nanoparticles in the 100–200 nm range. A third approach was developed using a reverse-microemulsion method to encapsulate a single BT core within a thin (10 nm) silica layer, with an overall particle size of 29 nm. The -OH groups on the silica layer readily enabled surface PEGylation, allowing the NPs to remain highly stable in saline solutions. We report that the silica coated BT NPs in both methods exhibited low level of Ba²⁺ leaching (≤3% of total barium in NP) in PBS for 48 hours compared to the unmodified BT NPs (14.4%).

Graphical Abstract

1. INTRODUCTION

Perovskite oxide nanocrystals have been extensively studied due to their tunable chemical and physical properties and wide range of applications in catalysis, photovoltaics and photonics.^1–3 The highly versatile perovskite structure (ABO₃) can be intersubstituted with main group, transition and rare-earth metals, which gives rise to a wide range of properties that might ultimately be utilized in the diagnosis and treatment of disease. Barium titanate (BaTiO₃, labelled here as BT) is probably one of the most studied compounds of the perovskite family and serves as a prototypical example.^4–6 Colloidal BT NPs have recently prompted researchers to explore their biomedical applications owing to their tunable sizes, non-linear optical characteristics and piezoelectric properties.⁷ Ciofani et.al. showed that BT-NPs non-covalently stabilized by glycol chitosan possess cytocompatibility at a concentration up to 100 μg/mL.⁸ Pantazis and coworkers introduced BT NPs as second harmonic generation (SHG) nanoprobes for in vivo imaging in living zebrafish embryos.⁹ Most recently, Luke and coworkers reported a method to prepare antibody conjugated BT NPs which exhibit cell-specific targeting ability.¹⁰ In combination with other materials, BT NPs were also used for the preparation of composite scaffolds for tissue regeneration, especially as bone tissue engineering agents, owing to their piezoelectric properties which stimulate signaling pathways for bone growth in response to mechanical stress.¹¹ These examples demonstrate the potential of BT NPs as biocompatible imaging probes and biosensors, as well as tissue engineering agents.

BT NPs with a wide range of sizes (3–300 nm) and crystalline phases can be produced through a number of synthetic routes, such as hydrothermal^12,13, solvothermal^14,15, sol-gel methods¹⁶, co-precipitation^17,18 and others^19–21. Our group previously developed a simple and scalable sol-gel-like technique (referred to as “gel collection”) to produce relatively monodisperse ligand-free BT NPs in the smaller diameter range, and with some degree of tunability, between 7 to 25 nm.²² This size range of BT NPs is advantageous for further surface modifications so that the final hydrodynamic size can remain small (< 50 nm) to minimize liver-NP interactions.^23–25 To the best of our knowledge, the use of sub-10 nm BT NPs as imaging agents has not yet been explored. The use of microscale BT in colloidal suspension was investigated as an alternative contrast medium in place of barium sulfate for the X-ray imaging of upper gastrointestinal tract.²⁶ However, BT in this format can be potentially cytotoxic due to the leaching of the surface Ba²⁺ ions²⁷ which potentially block the potassium ion channels leading to a disruption of the nervous system.²⁸ Therefore, it is necessary to perform proper surface modification to minimize, ideally inhibit, Ba²⁺ leaching. Surface modification of metal oxide NPs in this size range (sub-10 nm) have been a great challenge due to the hydrophobicity of the NP surface. Although it is one of the most common methods for SiO₂ coating, the efficacy of the Stöber process²⁹ is limited by the size, concentration and hydrophobicity of the NPs. Conditions such as small size and high concentration of the NPs can result in SiO₂ coating of aggregated cores.³⁰ Furthermore, due to the fast hydrolysis and condensation of the silane precursor in the Stöber process, a thickness of the SiO₂ shell below 20 nm is difficult to achieve without the use of a surface primer such as polyvinylpyrrolidone (PVP).^31–33 An alternative SiO₂ coating method for NPs in this size range is reverse microemulsion. Although it involves a more complicated phase transfer step and usually requires a longer synthetic time, the reverse microemulsion method has been used to effectively coat a variety of inorganic nanoparticles with sizes in the 10 nm range, with a single core-shell structure and a SiO₂ layer as thin as 1 nm.^34,35 The -OH group on the SiO₂ surface further enables covalent coupling of non-fouling polymers such as polyethylene glycol (PEG) for prolonged blood circulation of the NPs.

In this study, we develop and compare three different approaches to modify the surface of BT NPs synthesized by the gel collection method. These approaches include citrate adsorption, silica coating by the Stöber process, and silica coating by reverse microemulsion. We aim to enhance the aqueous dispersibility and biocompatibility of sub-10 nm BT NPs. We examine the colloidal stability and the leaching of Ba²⁺ of the post-modified BT NPs under physiological conditions. Furthermore, we contextualize potential applications for these approaches of BT NP surface modifications, aiming to provide the best strategy for their potential biomedical applications. By comparing the contrast ability between our surface modified BT NPs and commercial barium sulfate (BaSO₄) suspension, we demonstrate that sub-10 nm BT NPs are potential candidates for intravenous computed tomography (CT) imaging contrast agents.

2. EXPERIMENTAL SECTION

2.1. Materials.

All the chemicals were purchased from Sigma Aldrich unless otherwise stated. Milli-Q water (18.2 MΩ cm) was used in all the experiments.

2.2. Synthesis of 8 nm BaTiO₃ NPs.

BaTiO₃ NPs with an average diameter of 8 nm (±1 nm) were prepared according to a modified sol-gel method that affords fully crystallized NPs from solution at near room temperature. We refer to the method as Gel-Collection,²² with some variation described here. In a N₂ protected glove box, 5.38 mL of barium ethoxide (8.65 % w/v in ethanol) and 1 mL of titanium isopropoxide were mixed with 33.62 ml of absolute ethanol under mild stirring. After 10 minutes, 10 ml of 3:1 ethanol:H₂O was added to the mixture with continuous stirring. This process was kept for roughly 5 minutes until the mixture turned into a viscous clear solution. The viscous solution was then transferred to a closed container out of the glove box and warmed in an oven at 55 °C for 10–24 h. A self-accumulated white solid monolith product was formed at this point. The monolith was then separated from the solvent, rinsed with ethanol and redispersed in an appropriate amount of ethanol for further modifications. The dispersibility of BaTiO₃ NPs synthesized by this method can reach up to 50 mg/mL in absolute ethanol as a transparent colloidal solution.

2.3. Surface modification of BaTiO₃ NPs by citrate adsorption, BT@Citrate.

To 5 ml of as-synthesized BaTiO₃ NPs dispersed in ethanol at 10 mg/ml, 5 ml of 50 mM citric acid (C₆H₈O₇) was added while stirring. The mixture immediately turned into a cloudy solution. The stirring was stopped after 2 h followed by centrifugation at 7500 rpm for 10 minutes. The pellet was resuspended in 5 ml of citric acid and the centrifugation cycle was repeated two more times to completely remove the ethanol. After last centrifugation, the pellet was redispersed in 5 ml of 50 mM citrate buffer at pH 7.0 to form a transparent colloidal solution of BT@Citrate. The dispersibility of BT@Citrate is dependent on the pH of the solution. A reversible transition between the dispersed NPs and the agglomerated NPs can be achieved by adding HCl or NaOH in the solution (Video S1 in Supporting Information). The concentration of BT@Citrate NPs can reach up to 50 mg/mL in citrate buffer (pH 5.5–10.5) or phosphate buffer (pH 7.4) without the observation of agglomerates.

2.4. Synthesis of silica coated BaTiO₃ nanoaggregates, BTNA@SiO₂.

Silica coated BaTiO₃ aggregates, or BTNA@SiO₂, were prepared according to the Stöber method²⁹ with varied amounts of TEOS and water. Briefly, for the spherical BTNA@SiO₂, 9 ml of 1 mg/ml BT NPs in ethanol was mixed with 1 ml of water and 100 μL of TEOS with stirring. The mixture was stirred for 20 minutes and then 0.2 ml of ammonia solution (28%) was added to trigger the hydrolysis. The reaction was continued for 2 h until the mixture turned slightly turbid. The spherical BTNA@SiO₂ NPs were collected by centrifugation and resuspension in either water or ethanol. For the elongated BTNA@SiO₂, 2 ml of water and 50 ul of TEOS were used while the quantities of all other reagents remained the same.

2.5. Synthesis of single core silica coated BaTiO₃ NPs, BT@SiO₂.

Single core silica coated BaTiO₃ NPs were prepared following a two-step synthesis. First, the BaTiO₃ NPs dispersed in ethanol was treated with oleic acid to create a non-polar surface ligand. One milliliter of BT NPs in ethanol (20 mg/ml) was mixed with 4 ml of oleic acid solution (5% v/v in cyclohexane) with stirring. The mixture was stirred for 12 h to allow the surface of the NPs to be fully covered by the oleic acids. Then the mixture was centrifuged at 13500 rpm for 20 minutes to remove ethanol and excessive oleic acids. The pellet was redispersed in cyclohexane to produce a transparent colloidal solution of BT@OA. The second step followed the reverse microemulsion method for silica coating according to literature³⁵. Briefly, 0.5 ml of Igepal CO-520 was mixed with 10 mL of 0.2 mg/ml BT@OA and stirred for 10 min. Then, 0.1 mL of ammonia solution (28%) was added to the above mixture. Finally, 60 μL of TEOS was added via the equivalently fractionated drop method (adding 10 μL per 30 min). The resulting BT@SiO₂ core/shell products were collected after centrifuging and washing, and then were redispersed in either water or ethanol. By adjusting the initial concentration of BT@OA and the amount of TEOS added in the final step, single core BT@SiO₂ NPs with average shell thicknesses between 2 nm and 20 nm can be achieved.

2.6. Synthesis of BT@SiO₂-PEG.

The PEGylation of BT@SiO₂ was completed via two steps. The BT@ SiO₂ NPs were first functionalized with amine groups. In a typical synthesis, 5 mL ethanol dispersion of 0.2 mg/mL BT@ SiO₂ NPs was treated with 5 μL of APTES and 50 μL of ammonia solution (28%). The mixture was stirred for 2 hours at room temperature. The resulting BT@ SiO₂-NH₂ products were centrifuged and washed three times to remove excessive APTES and ammonia, and finally redispersed in 5 ml of 10 mM MES buffer at pH 5.5. At this point, a significant change of the surface charge from previously negative to positive was observed. The coupling of PEG-COOH to BT@ SiO₂-NH₂ was performed using the EDC/NHS strategy. In a typical synthesis, 5 mL of BT@ SiO₂-NH₂ at 0.2 mg/ml in the above MES buffer was mixed with an aqueous solution of PEG-COOH 5000 (1 mg/mL, 100 μL) and stirred for 10 minutes. To this mixture, aqueous solutions of coupling reagents EDC (5 μL, 100 mM) and NHS (10 μL, 100 mM) were added simultaneously. The resulting mixture was allowed to react under stirring for 24 hours at room temperature. The products were centrifuged and washed for three cycles at 7500 rpm with water, and finally redispersed in 10 mM of phosphate buffer at pH 7.4.

2.7. Transmission Electron Microscopy.

The TEM images were taken on FEI Titan Themis 200kV TEM (USA). For all TEM samples, 5 μL of appropriate BT solution was drop casted on a Formvar/carbon film grid (300 mesh, copper; TED PELLA, USA). The grid was then dried at room temperature overnight before examining.

2.8. Fourier Transform Infrared Spectroscopy.

The FTIR spectra were acquired from 4000 to 500 cm⁻¹ with 64 scans at 4 cm⁻¹ resolution on a Bruker Vertex 70 spectrometer (USA). BT in ethanol samples and BT@Citrate aqueous samples were lyophilized and dried in an oven at 70° C overnight. FTIR pellets were prepared after homogeneous mixing of the dried samples with KBr. The background was corrected by a reference of KBr pellets.

2.9. Hydrodynamic Diameter and ζ-Potential Measurements.

The hydrodynamic diameter and surface charge (ζ-potential) were recorded with Anton Paar Litesizer 500 Particle Analyzer (Austria). For the particle size measurements (n = 3), 1 mL of 0.1 mg/mL BT or surface modified BT samples were placed in quartz cuvettes. The volume weighted size distribution peak values were used to report the hydrodynamic diameters of samples. For the ζ-potential measurements (n = 3), 0.3 mL of 0.1 mg/ml aqueous solutions of BT@Citrate, BT@SiO₂, BT@SiO₂-NH₂ or BT@SiO₂-PEG were injected in Omega cuvettes. The Smoluchowski approximation was used to calculate the ζ-potential values.

2.10. Barium ion leaching study.

Ba²⁺ leaching was tested by dialysis for all the surface-modified BT NPs as well as the unmodified BT NPs. Dialysis tubings (MWCO 3500 Da) were purchased from Spectrum Laboratories, Inc. (Houston, TX). One milliliter of each BT sample in PBS at pH 7.4 was dialyzed against 100 mL of PBS for 48 hours (n = 2). During the dialysis, aliquots in the outside medium (2 mL, duplicated) at appropriate timepoints were subject to elemental analysis by ICP-OES to determine the amount of Ba²⁺ leaching. Also, one milliliter of each BT sample prepared at the same concentration as the mother liquor in the dialysis was subject to elemental analysis by ICP-OES to determine the total amount of barium in the NPs. Details of the operating conditions for ICP-OES are described below in Section 2.14. The amount of Ba²⁺ leaching was calculated as a weight percentage of the total amount of barium in the NPs.

2.11. CT contrast analysis of BT NPs and Readi-Cat^®2 BaSO₄ suspension.

Unmodified BT NPs in ethanol or BT@Citrate water dispersions at different concentrations in the range from 10 to 50 mg/ml and Readi-Cat^®2 (E-Z-EM Canada, Inc.) BaSO₄ suspension at 1× concentration (2.1% w/v, or 21 mg/ml) were filled in 1.5 ml microcentrifuge tubes. CT images were obtained using a MicroCAT-II™ scanner (ImTek, Knoxville, TN). The X-ray tube voltage was fixed at 55kVp and the current at 800 μA. A 360° (one revolution) micro-CT acquisition was performed for each set of the samples shown in Figure 6. The HU value of each sample was calculated by averaging the HU values of 3 ROIs (regions of interest) near the center of the sample.

Figure 6.

Proposed coating mechanism of silica on BT NPs via the reverse microemulsion method.

2.12. Cell culture and tumor xenograft.

All experiments involving animals were performed according to the guidelines approved by the Research Animal Resource Center and Institutional Animal Care and Use Committee at Memorial Sloan Kettering Cancer Center, NY, USA. We adhere to the animal research: reporting of in vivo experiments (ARRIVE) guidelines and to the guidelines for the welfare and use of animals in cancer research. The CT26 cell line was obtained from the American Type Culture Collection (ATCC), Manassas, USA. Cells were cultured in Roswell Park Memorial Institute (RPMI)-1640 growth media, supplemented with 10% fetal bovine serum (FBS). Six-to-eight week old female BALB/c mice (Charles River Laboratories, New York, USA) were injected subcutaneously into the right rear flank with CT26 (2 × 10⁶) cells in 150 μL 1:1 growth media/Matrigel® (BD Biosciences, San Jose, CA). Micro-CT imaging and biodistribution studies were performed within 2–4 weeks after injection.

2.13. In vivo micro-CT imaging.

Healthy mice and CT26 xenografted mice were intravenously injected through tail veins with a fixed amount of 150 μL of 8.33 mg/ml BT@Citrate NPs (1.25 mg dose). Prior to imaging, the mice were anesthetized by inhalation of 1.5% isoflurane. The mice were placed supine on the scanner couch with continuous inhalation of isoflurane to maintain anesthesia. CT images were obtained using the Inveon microPET/CT scanner (Siemens Healthcare GmbH, Erlangen, Germany). The spatial resolution of the CT sub-system of the Inveon microPET/CT scanner is 90 microns. The X-ray tube voltage was fixed at 55kVp and the current at 800 μA. A 360° (one revolution) micro-CT acquisition was performed every 10 minutes in the first hour and after 24 hours of NP injection (2 mice per timepoint). No signs of toxicity were observed from the physical appearance of the mice after 24 hours of NP injection. All mice were euthanized by carbon dioxide asphyxiation after the final micro-CT scan.

2.14. Ex vivo biodistribution study by ICP-OES.

After euthanasia, mouse organs (liver, spleen, and lung) and tumors were collected, washed in PBS buffer, and dried. The dried tissues were weighed, and then digested in a 2:1 mixture of concentrated nitric acid (70%) and hydrogen peroxide solution (30%). The sample digestion was assisted by applying low heat at 40–50°C until the solution turned completely clear. After dilution in DI water, the Ba and Ti contents (2 mice per timepoint at 30 min, 1 hour and 24 hours) were analyzed by ICP-OES on a Perkin-Elmer Optima 7300 DV spectrometer. The instrument was calibrated using a six-point calibration curve between 0.005 and 2.5 ppm and checked by three QC samples at the low, middle, and high points of the curve. The operating conditions employed for ICP-OES determination were as follows: 1300 W RF power, 15 L min⁻¹ plasma flow, 0.5 L min⁻¹ auxiliary flow, 0.8 L min⁻¹ nebulizer flow, and 1 mL min⁻¹ sample uptake rate. The wavelengths (λ) used for Ba and Ti analysis were 223.527 nm and 334.94 nm, respectively. The low limit of quantification for Ba and Ti were determined as 0.05 ppm. The percentage injected dose per tissue mass (% ID/g) was calculated by the following equation:

$$\frac{mass\left(Ba+Ti\right)intissue}{mass\left(Ba+Ti\right)indose}÷tissuemass\times 100$$ (1)

3. RESULTS AND DISCUSSION

As-synthesized BT NPs prepared via gel collection²² were found to be highly dispersible in polar organic solvents such as ethanol and furfuryl alcohol, but generally poorly dispersible in aqueous or non-polar organic solvents. This result can be explained by the BT surface that is rich in a combination of hydroxy and alkoxy groups (the surface will depend on the alkoxide precursors used).²² To enhance the aqueous dispersibility, we developed three different approaches for the surface modification of 8 nm BT NPs (Scheme 1). In the first approach we performed a simple surface modification with a single layer of citrate via molecular adsorption. The as-synthesized ligand-free BT NPs suspended in ethanol are extracted from the organic to aqueous solvent using citric acid followed by the addition of NaOH to increase pH. The second approach was achieved by encapsulating BT nanoaggregates (BTNA) – clusters of NPs within an amorphous silica layer (30–90 nm) using a modified Stöber method.²⁹ The third approach aimed to encapsulate a single BT NP core with a thin silica layer (<10 nm). This multistep process involved first an addition of an intermediate surface ligand to allow NP dispersion in a non-polar solvent, and second, a reverse-microemulsion technique to develop the silica layer. We further silanized the silica surface with an amine group and subsequently coupled it with polyethylene glycol (PEG).

Scheme 1.

Schematics showing the methods for surface modification of BT NPs and in vivo CT imaging using BT@Citrate as a contrast agent.

3.1. Nanoparticle Characterization.

3.1.1. BT@Citrate.

Citrates play an important role as a surface capping agent and stabilizer in many synthetic routes of inorganic NPs,^36,37 and the most well-known systems are gold and silver NPs.^38,39 Herein we used citrate as a post-synthesis surface capping agent to enhance the colloidal stability of BT NPs in water. Figure 1 illustrates the change of surface chemistry on BT before (Figure 1a) and after (Figure 1b) the citrate adsorption. We propose that the citrate molecules adsorb onto the BT surface by stripping the hydroxy/alkoxy terminal groups, ultimately forming ionic bonding between the negatively charged carboxylate and the positively charged barium/titanium. This can explain the observation that the BT NPs immediately precipitated out of the ethanol solution after the addition of citric acid solution: as the hydroxy/alkoxy terminal groups were stripped, the exposed BT surface promoted particle aggregation. After the addition of NaOH solution, the citric acid became deprotonated and the negatively charged carboxylates were adsorbed on the surface of BT to provide repulsive forces and hence the particles became redispersed in the aqueous solution. The grafting density of the citrates on an 8 nm BT NP was estimated to be 402 ligands/NP or 2 ligands/nm², with estimations based upon previously reported citrate densities on gold NPs.⁴⁰

Figure 1.

Schematic illustration of BT@Citrate synthesis showing (a) surface of the BT NPs prepared by the “gel collection” method and (b) surface of the BT@Citrate NPs.

The FTIR spectra in Figure 2a confirm that the citrate molecules were adsorbed on the surface of BT@Citrate, as indicated by the emerging peaks at 1562 cm⁻¹ and 1392 cm⁻¹, the characteristic peaks arising from the asymmetric and symmetric stretching of the carbonyl groups (C=O) on carboxylates (COO^-). The hydrodynamic diameters and ζ-potentials of the BT@Citrate NPs are listed in Table 1. The hydrodynamic diameter of BT NPs increased from 8.3 ± 0.3 nm (Figure S1) to 11.4 ± 0.2 nm (Figure S2) after the citrate adsorption, as anticipated. The ζ-potential of BT@Citrate (−33.1 mV) confirms a negatively charged surface indicating the presence of the deprotonated form COO^-. The photographs in Figure 2b indicate that the BT@Citrate NPs dispersed in water formed a transparent colloidal solution, while the BT NPs in water clearly settled at the bottom of the solution. These results demonstrate that the BT NPs in water are not irreversibly agglomerated when suspended in water and can be surface functionalized to attain a high degree of dispersibility, towards a highly stable colloidal solution. The small size of BT is likely critical in this regard. The TEM images of the BT NPs before and after citrate modification are shown in Figure 2c and 2d, respectively (also Figure S8 and S9 in Supporting Information). The well separated particles of BT@Citrate further confirmed that they exhibit high dispersibility due to their repelling surface. The citric acid solution added to the colloidal BT solution in ethanol was partially deprotonated (pK_{a 2} = 4.76 and pK_{a 3} = 6.4) to yield a mixture of citric acid molecules and citrate ions. As such, the dispersibility of the BT@Citrate NPs was found to be highly pH dependent (Figure 2e). When the pH of the dispersion falls below 5.2, the transparent colloidal solution started to get cloudy as a result of NP agglomeration. These large agglomerates of BT NPs reached thousands of nanometers (Figure 2e). As the pH was increased again to pass 5.5, the cloudiness disappeared, and the solution turned back to transparent, indicating that the BT agglomerates redispersed in the solvent. The reversible transition of pH dependent dispersibility of BT@Citrate is demonstrated in Video S1 in Supporting Information. This observation was anticipated as the highly negative ζ-potential provides repulsive forces on the surface of NPs at high pH, whereas the near-neutral ζ-potential promotes NP agglomeration.

Figure 2.

Characterizations of BT@Citrate NPs. (a) FTIR spectra of BT (top) and BT@Citrate (down). (b) Macroscopic photographs of BT in water (left) and BT@Citrate in water at pH 7.5 (right). (c) and (d) TEM images of dried samples of BT and BT@Citrate NPs. (e) ζ-potential and hydrodynamic size measurements of BT@Citrate NPs in 10 mM phosphate buffer as a function of pH. Results reported are mean ± SD, n = 3.

Table 1.

Particle size and surface charge measurements at pH 7.4.

Nanoparticles	Particle Size (nm)	Zeta Potential (mV)
BT (in ethanol)	8.3 ± 0.3	34.1 ± 5.1
BT@Citrate	11.4 ± 0.2	−33.1 ± 0.7
BT@OA (in cyclohexane)	21.7 ± 0.3	N/A
BT@SiO2	28.7 ± 2.9	−31.1 ± 0.5
BT@SiO2−NH2	29.0*	32.0 ± 0.7
BT@SiO2−PEG	53.5 ± 3.8	−16.5 ± 0.8

^*Estimated based upon BT@SiO₂ data.

3.1.2. BTNA@SiO₂.

BTNA@SiO₂ NPs were synthesized using a modified Stöber method.²⁹ According to our previous study, the surface charge of BT NPs synthesized by the gel collection method is positive (+30 mV),²² which is in agreement with our measurement (Table 1 and Figure S1 in Supporting Information). The cationic nature of the surface readily attracted the negatively charged silanol group on the hydrolyzed TEOS under basic environment (pH 10–11) and promoted nucleation and growth of the silica shell on the surface of BT. This largely avoided the formation of core-free silica NPs without using a surface primer as observed by other studies on the silica coating of inorganic NPs.^30,31,41 Figure 3a illustrates the reaction scheme for the silica coating on BT NPs. It is important to note here that the BT aggregates resulted from the addition of H₂O during the silica coating synthesis. The average thickness of the silica shell of BTNA@SiO₂ is tunable between 20–100 nm by adjusting the amount of TEOS. A large thickness (100 nm) and smooth surface of silica shell is achieved by increasing the amount of TEOS to 100 μL to coat 9 mL of 1 mg/mL BT NPs. Figure 3b (also Figure S10 in Supporting Information) shows the TEM image of spherical BTNA@SiO₂ NPs with an average silica shell thickness of 90 nm. The size of the BT aggregates core and the morphology of the silica coated NPs can be tuned by changing the ethanol/H₂O ratio. As-synthesized BT NPs using the gel collection method are highly dispersible in polar organic solvents such as ethanol and furfuryl alcohol, but poorly dispersible in H₂O. However, H₂O is one of the key reagents in the Stöber process. As such, aggregation of BT NP during prior to the coating is inevitable and the degree of aggregation of the NPs is dependent on the ethanol/H₂O ratio. Elevated concentration of H₂O also leads to inhomogeneous nucleation of SiO₂ on the BT surface, creating rough and nonuniform silica surfaces.⁴² The elongated or irregular shape of BTNA@SiO₂ with large cores was observed by adding a high concentration of H₂O (ethanol/H₂O ratio < 4). Figure 3c shows the TEM image of elongated BTNA@SiO₂ NPs with an average silica shell thickness of 20 nm. As shown, the morphologies of both the BT core and the silica shell structures differ greatly from the spherical BTNA@SiO₂. Elemental composition of the core-shell structure of BTNA@SiO₂ NPs was confirmed using Energy-dispersive X-ray spectroscopy (EDS) (Figure 3d).

Figure 3.

(a) Reaction scheme of silica coating on BT aggregates. The shell thickness and the size of the BT aggregates core can be controlled by changing the amounts of TEOS and H₂O added in the synthesis. (b) and (c) TEM images of spherical BTNA@SiO₂ and elongated BTNA@SiO₂. (d) Energy-dispersive X-ray spectroscopy (EDS) elemental mapping of BTNA@SiO₂.

3.1.3. BT@SiO₂.

A schematic illustration of a multi-step synthesis of BT@SiO₂ from BT NPs is shown in Figure 4a and 4b. It involved first a solvent transfer from ethanol to cyclohexane using oleic acid (OA) and followed by a reverse microemulsion process for silica coating. Figure 5a (also Figure S11 in Supporting Information) shows the TEM image of well-separated BT@OA NPs after the solvent transfer of BT from ethanol to cyclohexane. The bottom right inset in Figure 5a reveals a transparent solution of BT@OA (10 mg/mL) in cyclohexane. The size of the BT@OA NPs increased to 21.7 nm due to the length of OA molecules (Table 1 and Figure S3 in Supporting Information). The OA spacer pre-coating is considered a crucial step to prevent encapsulation of multiple particles per silica shell due to mild agglomeration of BT NPs. With the carboxylate group adsorbed on the BT surface, the long hydrocarbon chain of OA allows the NPs to achieve a high dispersibility in non-polar solvents such as cyclohexane. The reverse microemulsion method is known to be able to achieve a high level of control over the thickness of the silica shell for the coating of sub-10nm NPs.^43,44 Based on the mechanism of silica coating on Fe₃O₄ proposed by Ding et al.,⁴³ we suggest the following coating mechanism for BT NPs (Figure 6). First, the surfactants of Igepal CO-520 form micelles in the cyclohexane solution. Then, the BT@OA NPs, when added to the solution, undergo a ligand exchange on the particle surface between the OA and some of the Igepal CO-520 molecules. The ammonia solution added in the next step forms micro-droplets with the remainder Igepal CO-520 molecules wrapped around the water/oil interface. Subsequently, the TEOS added in the solution is hydrolyzed at the at the water/oil interface. The hydrolyzed TEOS at the interface attracts the BT NPs and drives the NPs to enter the water phase. As a result, the nucleation and silica growth take place preferentially on the BT surface inside the water droplet. Figure 5b shows the TEM image of the resulting BT@SiO₂ NPs with an average shell thickness of 10 nm. As shown, most of the NPs have a single BT core except a few that are encapsulated with dimers or trimers of BT NPs. This result indicates that a small number of BT NPs were still bound together as dimers or trimers when the silica nucleation and growth occurred. DLS measurements showed that the average size of the BT@SiO₂ NPs is 28.7 ± 2.9 nm. This measurement confirms that the thickness of the silica shell is approximately 10 nm given that the size of the bare BT is 8.3 nm on average.

Figure 4.

Multi-step synthetic route of BT@SiO₂-PEG NPs from BT NPs. (a) Transfer of the NPs to a non-polar solvent using oleic acid as surface ligands; (b) silica coating via the reverse microemulsion method; (c) surface amination using APTES; (d) surface PEGylation by EDC/NHS coupling. (e) Chemical structures of different reagents used.

Figure 5.

TEM images of (a) BT@OA, (b) BT@SiO₂, (c) BT@SiO₂-NH₂ and (d) BT@SiO₂-PEG. The top left insets in (a) and (b): magnified TEM images of each sample. Bottom right insets: photographs of each sample in their appropriate solvent. (a) BT@OA in cyclohexane; (b) BT@SiO₂ in ethanol; (c) BT@SiO₂-NH₂ in ethanol; (d) BT@SiO₂-PEG in water.

3.1.4. BT@SiO₂-PEG.

The surface of BT@SiO₂ was further covalently coupled with polyethylene glycols (PEGs). A schematic illustration of a multi-step synthesis of BT@SiO₂-PEG from BT@SiO₂ is shown in Figure 4c and 4d. APTES ((3-Aminopropyl)triethoxysilane) was used to functionalize the silica surface on BT@SiO₂ NPs with amine groups that are readily available for EDC/NHS coupling with PEG-COOH. The EDC/NHS set is one of the most commonly used coupling reagents to catalyze the amide bond formation between carboxyl and amine groups for water-soluble NPs.^45–47 The successful coupling in each step was confirmed by the change in ζ-potential measurements (Table 1 and Figure S4, S5, and S6 in Supporting Information). From BT@SiO₂ to BT@SiO₂-NH₂, a drastic change of surface charge was observed from −31.1 to +32.0 mV, from which we deduced that the -OH group on the surface was changed to -NH₂. Then from BT@SiO₂-NH₂ to BT@SiO₂-PEG, the change of ζ-potential from +32.0 mV to −16.5 mV giving a strong indication of the successful coupling of PEG. The TEM images of the amine functionalized BT@SiO₂ NPs (BT@SiO₂-NH₂) and the PEGylated BT@SiO₂ NPs (BT@SiO₂-PEG) are shown in Figure 5c and 5d, respectively (also Figure S12 in Supporting Information). It can be noticed that both BT@SiO₂ and BT@SiO₂-NH₂ NPs were slightly more agglomerated compared to the BT@SiO₂-PEG NPs. This suggested that PEGylation indeed enhanced the aqueous dispersibility of the BT NPs.

3.2. Comparison of all three methods.

We successfully altered the surface chemistry of the as-synthesized BT NPs through three different approaches, including the citrate coating, the multi-core silica coating, and the single-core/shell silica coating. The citrate coating method allowed the BT NPs to achieve high dispersibility in aqueous media up to 50 mg/mL. This level of dispersibility of colloidal BT NP in aqueous solutions is significantly higher than previously reported.^48–50 The size of the particles also remained small (11 nm) as the coating was via a single layered molecular adsorption. However, the colloidal stability of BT@Citrate is largely affected by the ionic strength in the aqueous solution. At high salt concentration (PBS buffer), the BT@Citrate dispersion turned turbid and settled over time; in contrast, the BT@Citrate NPs remained stable in 10 mM phosphate buffer for 48 hours (Figure S7, Supporting Information). This suggested that the citrate adsorption on the surface of BT is dynamic and can be disrupted or replaced by a high concentration of ions or ligands. In this respect, the citrate-NP interactions on BT@citrate behave in a similar fashion as the widely studied citrate capped gold nanoparticles - that they start to aggregate beyond a threshold of salt concentration.^51,52 It is therefore clear that, while BT@citrate NPs possess high dispersibility in water, the nature of the relatively weak electrostatic interaction between the coating ligand and the surface will eventually fail to maintain colloidal stability in biological fluids. Nonetheless, this post-synthetic citrate treatment can be applied to many other doped derivatives²² of BT NPs prepared by the gel collection method.

The BTNA@SiO₂ NPs were synthesized using the Stöber method, a simple technique that is both time and cost efficient. The thickness of the silica shell can also be tuned in a relatively large range from 20 nm to 100 nm. Generally however, thickness control < 20 nm is difficult to achieve.³¹ Additionally, the BTNA@SiO₂ NPs did not exhibit as high dispersibility in water as did BT@Citrate. These large NPs started to turn turbid in aqueous solution and settle overtime as the concentration increased to approximately 5 mg/mL. Below 5 mg/mL, the NPs maintained homogeneous dispersity in aqueous buffer solution even at saline concentration (Figure S7, Supporting Information). With the overall size of 100 to 200 nm and the highly concentrated piezoelectric BT core, these BTNA@SiO₂ NPs can potentially serve as scaffolds for orthopedic tissue engineering.^53,54

The BT@SiO₂ produced via the reverse microemulsion method is the most promising in terms of particle size control (~30 nm) and structure (single core/shell). The current limitation of this approach is the requirement of a longer synthetic time (up to 48 hrs) and relatively small synthesis scale, with a current upper limit of ~1 mg BT/mL. The colloidal stability remained high without signs of agglomeration at a concentration up to 10 mg/mL and for a period up to 48 hours (Figure S7, Supporting Information). Compared to the other two modification methods presented here, these highly stable PEGylated BT NPs demonstrate the highest potential to serve as biocompatible imaging agents.

3.3. Biocompatibility of the surface-modified BT NPs.

The application of BT NPs for biomedical purposes, especially as therapeutic and diagnostic nanomaterials, is relatively recent. To date, pristine BT NPs produced by most of the synthetic routes,^12–22 including the gel collection method,²² are not biocompatible due to their inability to disperse in aqueous solutions. BT is known to be thermodynamically unstable in aqueous environment due to the leaching of Ba²⁺ at low pH or the formation of BaCO₃ at high pH.^27,55,56 It has been shown that the leaching of Ba²⁺ on BT surface has a high dependency on the pH of the solution, with higher leaching amount and faster leaching rate at lower pH.^56,57 As a result, the surface of BT will become rich in TiO₂ and hence it will promote NP aggregation. In the case of our BT NPs, after the citric acid treatment, immediate precipitation of the NPs was observed. We propose two different scenarios that could cause this observation. First, the surface Ba²⁺ ions were likely to be stripped during the acid treatment, giving rise to a TiO₂-type surface. It has been reported that titanium-based materials tend to undergo a phase transfer due to the exchange of surface metals by protons under acidic conditions.⁵⁸ In our case, the ion exchange would occur between the surface Ba²⁺ and H⁺, creating an electropositive surface. We assume the subsequent addition of NaOH deprotonated the citric acid groups to form citrates, which in turn formed metal complex bonds with Ti (IV). Thus, the NPs were dispersed due to electrostatic repulsion. In the second scenario, the citric acid formed metal complex bonds directly with the surface Ba²⁺ which, contrarily, was stabilized. While the precipitation was likely caused by hydrogen bonding between the carboxylic acids on adjacent NPs, the addition of NaOH deprotonated the acids to form electrostatically repulsed carboxylates on the citrates on the NP surface. A similar study was performed on 50 nm barium strontium titanate (BST) NPs by first using nitric acid to remove the surface Ba²⁺ and subsequent treatment with citric acid for surface adsorption.⁵⁹ In our study, however, treating small BT NPs (< 10 nm) with nitric acid caused a significant size reduction of individual NP and also random aggregation of some NPs due to the TiO₂-rich surface. Thus, direct citric acid treatment was used in our study.

Elemental analysis by inductively coupled plasma optical emission spectrometry (ICP-OES) was used to monitor the behavior of barium in all three systems, and whether leaching of Ba²⁺ from the NP into the suspending media was prevented by the surface coating configuration. Our aforementioned suggested mechanism of citrate adsorption to the NP surface appears to be confirmed by the results. The amount of barium measured by ICP-OES in BT@Citrate was significantly reduced compared to the calculated amount of barium in bare BT before the citric acid treatment, suggesting that a large number of surface Ba²⁺ was stripped by the citric acid (Figure S13, Supporting Information). Next, we performed Ba²⁺ leaching study on all the three variations as well as the unmodified BT NPs over 48 hours. As shown in Figure 7, the unmodified BT NPs exhibited significantly higher initial rate of Ba²⁺ leaching compared to the other surface-modified BT NPs, and higher overall Ba²⁺ leaching compared to BTNA@SiO₂ and BT@SiO₂ NPs. The amount of Ba²⁺ leaching was calculated as a weight percentage of the total amount of barium in the NPs. Almost 5% of the barium leached from the NPs in the first 2 hours and 14.4% leached at 48 hours. In contrast, the BTNA@SiO₂ and BT@SiO₂ NPs showed minimal level of Ba²⁺ leaching, at about 3% for 48 hours. This level of Ba²⁺ leaching would account for ~45 μg of barium if 3 mg of NPs were injected in a 25 g mouse, or a barium exposure of 1.8 mg/kg of mouse. In comparison, the lowest level of barium exposure which leads to increased mortality was reported to be 160 mg/kg/day in a study exposing 120 of mice to barium chloride in drinking water for two years.⁶⁰ Our results suggest that the silica shell coatings on the BT NPs were able to effectively confine the barium within the core, effectively minimizing the cytotoxicity of these NPs and increasing their potential for biomedical research. Interestingly, the BT@Citrate NPs did not show any leaching in the first two hours, possibly because a large amount of the surface Ba²⁺ was already stripped during the citric acid treatment. Then, the rate of Ba²⁺ leaching in BT@Citrate was higher than that observed in the unmodified BT after ~16 hours, eventually reaching 18.7% at 48 hours. This behavior, of more rapid leaching after the first few hours, is likely due to an increased level of undercoordinated surface sites on the BT NPs caused by the citric acid treatment, as well as the instability of the surface citrates.

Figure 7.

Barium ion leaching as weight percentage of total barium in the NPs as a function of time. The unmodified BT NPs showed significantly higher initial rate of Ba²⁺ leaching. In contrast, the BT@Citrate NPs showed no Ba²⁺ leaching in the first two hours because a large amount of surface Ba²⁺ had been removed during the citric acid treatment. Both silica coated BT NPs exhibited substantially lower Ba²⁺ leaching over the 48 hours compared to the unmodified and citrate coated BT. Results shown are mean ± SD, n = 2.

3.4. BT NPs as contrast agents for computed tomography (CT).

Computed tomography (CT) is a non-invasive tissue imaging technique employed in numerous research and clinical settings.^61,62 Barium has long been known to be of great value in highlighting the area of interest in CT imaging. Barium sulfate (BaSO₄) suspension has been the predominant contrast agent for improving the visualization of CT imaging of the gastrointestinal tract. The reason why barium is excellent for CT imaging is multifold. Due to the high atomic number, Z, of barium, the position of its K-shell electron absorption edge (k-edge, ~39 keV), i.e., the photon energy at which barium has the highest attenuation, overlaps with the distribution peak of the X-ray energy produced by clinical CT scanners (typically operate at a voltage between 80–140 kV).⁶³ This means that barium absorbs more photons than most other elements do from the X-rays produced by clinical CT scanners. Moreover, barium is abundant and BaSO₄ is relatively cheap to produce. CT is an extremely useful, low-cost and routine technique, but frequently lacks the desired level of contrast and resolution, when compared to magnetic resonance imaging (MRI) or positron emission tomography (PET), and it has been speculated that improved contrast enhancement would greatly benefit the medical profession.⁶⁴

CT relies on the X-ray attenuation properties of different materials to achieve imaging effectiveness. For example, bones appear much brighter than lungs do in a CT scan because X-ray attenuates (being absorbed) on bones much more than on lungs. In general, tissue with higher density (ρ) or higher atomic number (Z) tend to better absorb X-rays.⁶⁴ The degree of X-ray attenuation on different materials, or X-ray absorption coefficient (μ), and is expressed as

$$\mu =\frac{\rho Z^4}{A{E}^3}$$ (2)

where A is the atomic mass and E is the X-ray energy. Figure S14 in Supporting Information shows a plot of attenuation coefficient of different elements vs. photon energies, illustrating that the k-edge of Barium overlaps with the energy of the main population of the photons emitted by clinical CTs. A CT scan uses a standardized scale to measure the ability of tissues to attenuate X-ray in Hounsfield units (HU). By definition, water is assigned to have a value of 0 HU, so that the CT scanners can be calibrated with a reference to water. For a material with a linear X-ray attenuation coefficient “μ” the corresponding HU value is calculated by

$$\mathrm{H}\mathrm{U}=\frac{\mathrm{\mu }-{\mu }_{water}}{{\mathrm{\mu }}_{water}}\times 1000$$ (3)

where μ_water is the X-ray attention coefficient of water. Many soft tissues, e.g., livers and spleens, share similar HU values as they are made of mostly water. As such, contrast agents with relatively higher HU are usually used in a CT scan to highlight the area of interest such as tumor tissues. BaSO₄ is administered orally as a suspension to delineate features in the gastrointestinal tract using CT imaging. Despite the prevalence of the use of barium sulfate suspensions, this is the only example of barium as a CT contrast agent due to the high toxicity of Ba²⁺. Studies have shown that biocompatible NPs containing high Z elements present options for extending resolution and sensitivity of CT techniques.^65,66 In this context, we explore the proposition of using BT NPs as intravenously administered cancer targeting CT contrast agents based upon three justifications derived from this first study: first, the HU value of BT NPs is comparable to that of barium sulfate; second, proper surface-modification of BT NPs can largely enhance their dispersibility and biocompatibility; third, the size of post-modified BT NPs can be optimized for prolonged blood circulation and tumor accumulation.

Readi-Cat^® 2 is an FDA-proved trademark for orally administered barium sulfate suspensions. Using a micro-CT scanner, we compared the contrasts in HU value between BT NPs (citrate coated and unmodified) at various concentrations and Readi-Cat^® 2 (Figure 8). As shown, for the unmodified BT NPs in ethanol, the 20 mg/mL and 30 mg/mL samples exhibited at 165.6 HU and 406.8 HU, respectively. Readi-Cat^® 2 with a concentration of BaSO₄ at 21 mg/mL exhibited at 374.2 HU. As the unmodified BT was dispersed in ethanol instead of water for the CT scan, comparison of the above results needs to be normalized. The CT number of pure ethanol was previously reported at around −230 HU.⁶⁷ Thus, the CT numbers of the ethanol dispersed BT reported here are expected to increase by approximately 230 HU if the solvent was replaced by water. As a result, the 20 mg/mL unmodified BT NPs should exhibit at ~395 HU in water, compared to Readi-Cat^® 2 at 374.2 HU. For the BT@Citrate NPs in water, the higher concentrations at 40 and 50 mg/mL exhibited slightly less HU (287.9 and 289.7) than Readi-Cat (373.2). As discussed previously, the decreased HU number of the BT@Citrate NPs compared to pristine BT NPs is likely due to the significant loss of surface Ba²⁺ during the citric acid treatment (Figure S13, Supporting Information), leading to a lower molarity of Ba present, yet unaccounted for by the diameter of the NP. It is important to note that the CT numbers of both the BT@Citrate and unmodified BT NPs reported here are significantly greater than those of soft tissues, typically ranging from 20 to 100 HU.

Figure 8.

Micro-CT scans and HU values of BT@Citrate in water (a) and unmodified BT in ethanol (b) at various concentrations compared with Readi-Cat^® 2. Note that the percent composition of Ba in BaSO₄ by mass (58.84%) is almost the same as that in BaTiO₃ (58.89%). (c) HU value of the unmodified BT or BT@Citrate NPs as a function of particle concentration.

Encouraged by the comparable CT contrasts between BT@Citrate NPs and commercial barium sulfate suspensions, we predicted that these NPs were also able to show CT contrast enhancement in animals. It has been demonstrated that CT26 tumor exhibits highly leaky vasculature and its uptake of non-specific NPs is dominated by passive targeting via the enhanced permeability and retention (EPR) effect.⁶⁸ Hence CT26 was used as the tumor model in this study. We performed in vivo micro-CT imaging on both healthy mice and CT26 tumor xenografted mice after intravenous tail injection of BT@Citrate NPs. After the injection of NPs, 4 of 10 mice studied had mild diarrhea within the first 20–30 minutes. No other signs of toxicity were observed from the physical appearance of the mice for up to 24 hours. The CT images were taken every 10 minutes in the first hour and at 24 hours. As shown in Figure 9, the contrast of the bladder (highlighted by red arrow) of the healthy mouse (top) was higher than that of the surrounding soft tissues in all the scans. This result confirmed that highly localized BT NPs were able to enhance the CT contrast of the region. Increasing CT contrast was observed in the bladder for up to 24 hours, indicating that the NPs were cleared through the renal system and accumulated in the bladder. In the CT26 xenografted mouse in Figure 9 (bottom), no apparent difference in CT contrast in the tumors (highlighted by red arrow) was observed over 24 hours. This result suggested that the accumulation of BT in the tumor was not high enough for CT contrast enhancement compared to the surrounding tissue.

Figure 9.

In vivo Micro-CT imaging of mice injected with BT@Citrate NPs. Longitudinal micro-CT scans of healthy mouse (top) and CT26 tumor-bearing mouse (bottom) were taken up to 24 hours following the intravenous tail vein injection of BT@Citrate NPs.

To further investigate the biodistribution of the BT NPs, elemental analysis was performed employing inductively coupled plasma optical emission spectrometry (ICP-OES). Mouse organs (liver, spleen, and lung) and tumor were removed at 30 minutes, 1 hour, and 24 hours (2 mice per timepoint) after the injection of NPs for elemental analysis. Results revealed that most of the injected NPs was accumulated in the liver and spleen, and a small amount was accumulated in the lung (Figure 10). For the healthy mice, uptake in the spleen and liver followed a significant increase from 30 minutes at 30–40% ID/g to 60 minutes at over 65% ID/g (Figure 10 a). For the tumor-bearing mice, faster initial uptake in the spleen and liver was observed (Figure 10 b). It reached over 50% ID/g in the first 30 minutes and increased to around 60% ID/g at 1 hour. Interestingly, uptake in the lung of tumor-bearing mice was significantly higher than that of healthy mice within the first hour. One possible explanation to the faster initial uptake in the spleen, liver, and lung for the tumor-bearing mice was a faster blood circulation compared to the healthy mice. Uptake in the tumor was also observed with 3.1% ID/g at 1 hour and 4.2% ID/g at 24 hours of injection. This level of accumulation, however, was not high enough to show an enhancement of CT contrast in the tumor (Figure 9). The low tumor uptake suggested that, as expected, the BT@Citrate NPs could not maintain a blood circulation long enough to reach the tumor due to the weak surface coating of citrates via electrostatic interaction. The surface citrate molecules were likely stripped immediately after the NPs entrance to the bloodstream, causing NP aggregation and the formation of protein corona. This type of corona formation is usually caused by an immune response and the type of proteins in the corona (e.g. opsonin) tend to interact strongly with macrophages of the reticuloendothelial system (RES), leading to fast blood clearance and accumulation in the liver and spleen.^69–71 The BT@Citrate NPs demonstrated the feasibility of using surface-modified BT NPs as CT contrast agents. Given the advantage of the silica coating and PEGylation on the BT@SiO₂-PEG NPs, we anticipate that these NPs will have longer blood circulation, comparable CT contrast, and are overall a more robust system for in vivo CT imaging studies.

Figure 10.

Biodistribution of BT up to 24 hours after BT@Citrate injection in healthy mice (A) and CT26 xenografted mice (B). The BT accumulation in different organs were measured and quantified by ICP-OES. The biodistribution profile here is shown in percentage injected dose per tissue mass, %ID/g. Highest accumulation was observed in spleen and liver. Results shown are mean ± SD, n = 2.

4. CONCLUSIONS

In summary, we have presented three different routes of surface modifications to enhance the aqueous dispersibility and biocompatibility of BT NPs. Our first method utilized a molecular adsorption method to produce BT@Citrate NPs. This method significantly improved the aqueous dispersibility of BT NPs compared to non-functionalized particles. The non-covalent citrate-NP interaction was somewhat easily disrupted by the ionic strength under physiological conditions, which eventually caused NP aggregation. The stability of the NPs was further improved by silica coating in the other two methods. The second modification utilized a modified Stöber method to produce BTNA@SiO₂. This method produced tunable silica shell thicknesses (20–100 nm) with tunable core sizes and morphologies. These were larger, stable nanoparticle structures that still may serve to be functional but were above our initial target size. The last multi-step modification produced covalently PEGylated BT@ SiO₂ NPs with single core/shell structures. This is a multistep process that involves first an addition of an intermediate surface ligand to allow NP dispersion in a non-polar solvent, second a reverse-microemulsion technique to develop the silica layer, and third an amine functionalization for the final PEG coupling. Both last two methods produced protective silica shells that allowed BT NPs to be stable in saline solutions maintaining a moderate dispersibility. The last method optimized the hydrodynamic particle size (~50 nm) and provided the surface tunability to engineer additional functionalities. Both unmodified BT and BT@Citrate NPs exhibited high CT contrast compared to the commercially available BaSO₄ suspension. The in vivo CT imaging study of the showed the feasibility of using modified BT NPs as contrast agents.