Highly aligned arrays of high aspect ratio barium titanate nanowires via hydrothermal synthesis

Nanowires have recently garnered significant research interest as attractive structures for various applications such as accelerometers,^1,2 gas sensors,^3–5 electrochemical biosensors,^6,7 power harvesters,^8,9 and dye-sensitized solar cells.^10,11 The performance, efficiency, and sensitivity of these nanowire-based devices have been improved over bulk structure-based devices due to their high surface-to-volume ratios, high compliance, high aspect ratio, and tunable electron transport. The enhancements realized by these 1-D structures have driven research efforts to investigate ways to synthesize nanowires of various sizes, compositions, properties, and alignment. Different synthesis routes to grow free-stranding nanowires have been developed, but the difficulty lies in creating arrays of nanowires that propagate normal to a substrate's surface. There are a vast number of bottom-up synthesis routes to achieve these arrays such as vapor-phase depositions,¹² electrodeposition,¹³ vapor-liquid-solid technique,^14,15 and template-assisted growth.¹⁶ The techniques utilizing epitaxial growth mechanisms typically produce the most highly aligned and uniform nanowires. However, these techniques routinely use harsh reaction parameters, specifically high temperatures and vacuum environments, and require specialized equipment that is not applicable for low cost, scalable production.

Consequently, hydrothermal reaction methods have been applied in synthesizing vertically aligned nanowire arrays due to their relatively low reaction temperatures, low cost, and scalability.^10,17 The trade-off, however, for this technique is that the degree of alignment of the nanowire array is not as high as compared to previously developed epitaxial growth methods due to the reaction being a competitive growth process. Vertically aligned nanowire arrays in a hydrothermal reaction are the result of a competitive growth mechanism that involves nucleation and growth occurring in random directions off the substrate surface. As the nanowires grow, they intersect neighboring nanowires causing growth to terminate, thus allowing only the nanowires oriented normal to the substrate surface to continue to grow throughout the reaction. Therefore, the crystal growth mechanism results in a lower degree of alignment and a lower density of nanowires compared to epitaxial growth methods. These low degrees of orientation are especially apparent as the length of the nanowires is increased and wicking of the nanowires tips occurs. Recently, epitaxial growths have been demonstrated for zinc oxide (ZnO) using hydrothermal reactions with less harsh reaction parameters.¹⁸ These hydrothermal epitaxial growths have been demonstrated for only short ZnO nanowires. In this research, a hydrothermal epitaxial growth is developed that produces highly aligned arrays of high aspect ratio barium titanate (BTO) nanowires. Previous research on textured piezoelectric films have shown increases in piezoelectric properties with increasing degree of alignment.^19–26 These property enhancements approach the values seen in single crystal samples. Therefore, the increased alignment shown in this research will be advantageous in future integration in microelectromechanical systems and nanoelectromechanical devices.

In this letter, a route to synthesize highly aligned, high aspect ratio BTO nanowires is developed via a hydrothermal synthesis technique. The competitive growth process in the hydrothermal reaction is combined with an epitaxial growth style nucleation by utilizing an underlying textured film on the substrate. This textured film allows the competitive growth process to have initial nucleation starting in a preferred orientation rather than random orientations. This not only results in more aligned growth through the thickness of the sample but also it results in a denser array of nanowires.

The TiO₂ textured film was synthesized by first obtaining a titanium foil (Alfa Aesar, 99.5%) with a thickness of 0.25 mm that was cleaned by sonication for 30 min in a solution of acetone, isopropanol, and water with a volumetric ratio of 1:1:1. This foil was then placed in a sealed Teflon tube in an autoclave and underwent a two-step hydrothermal reaction to produce a TiO₂ textured film. Both reaction steps used 20 ml of a growth solution consisting of a volumetric ratio of 20:20:1.25 of hydrochloric acid (Fisher, 35%), DI water, and titanium tetrachloride (Alfa Aesar, 99.0%), respectively. The first step was reacted at 180 °C for 1 h, and then a second step used a refreshed solution and was reacted at 180 °C for 2 h. Sodium titanate (NTO) was then hydrothermally grown on the TiO₂ using 20 ml of a 12 M sodium hydroxide (EMD, 99%) solution heated to 210 °C for 16 h in a sealed Teflon tube in an autoclave.^27,28 A final hydrothermal reaction used a 0.05 M barium hydroxide octahydrate (Sigma Aldrich, 98%) solution reacted at 210 °C for 12 h to completely convert the NTO to BTO via an ion diffusion mechanism.²⁹ To compare the effect of the TiO₂ textured film on the degree of orientation of the BTO nanowire arrays, the same NTO growth and BTO conversion reactions were performed on titanium foil that was oxidized at 750 °C for 8 h in air before the reaction.^1,27,28

The improvement in degree of orientation was qualitatively investigated using a scanning electron microscope (SEM, TESCAN VEGA3 LM), as shown in Fig. 1. The substrates used as the nucleation sites for NTO growth are illustrated in Figs. 1(a) and 1(d). The oxidation layer on the titanium foil in Fig. 1(a) that resulted from heating the titanium foil in air at 750 °C for 8 h shows no texturing, thus leading to a strictly competitive growth process during the hydrothermal growth of NTO. As shown in the cross-sectional and plan view SEM images in Figs. 1(a)–1(c), it is clearly illustrated that when growing on the oxidized titanium foil the nanowires have a low degree of orientation and the nanowire tips wick together. This previously established growth process on the oxide layer provided the most aligned NTO growth to date.²⁸ To improve upon this NTO growth procedure, a textured TiO₂ film was used instead of an oxide layer. As shown in Figs. 1(d)–1(f), the nanowires grown on this textured film have a higher growth density and are well-oriented thus preventing the nanowires tips from wicking together.

Using these nanowires as templates, highly aligned BTO nanowires were created by converting the NTO during a hydrothermal reaction, as detailed previously. The resulting highly aligned BTO nanowires are shown in the SEM image in Fig. 2(a) along with an atomic force microscopy (AFM, Park Systems XE-70) scan of the nanowire tips in Fig. 2(b). As seen in the SEM image, the morphology of the nanowires is preserved during the conversion process. The AFM shows that the highly aligned nature from this growth method allows the top facets of the nanowires to be probed, which was previously unattainable in the aligned sample grown on oxidized titanium foil. By contacting one of these top facets and applying a bipolar voltage, the phase change was captured (Fig. 2(c)) and revealed a 180° phase shift thus confirming the presence of ferroelectric behavior. Bipolar voltage applied at higher voltages also resulted in nanowire tip displacements, which is illustrated in the butterfly loop in Fig. 2(d). This butterfly loop revealed a d₃₃ piezoelectric coefficient of 32.7 pm V⁻¹. X-ray diffraction (XRD, PANalytical X'Pert Powder) patterns were acquired in order to verify the conversion of HTO to BTO. These patterns in Fig. 3 illustrate the characteristic peaks of BTO as identified by JCPDS 05-0626 in both the highly aligned and aligned cases. The characteristic peaks for NTO (JCPDS 73-1398) exist in both cases prior to the conversion reaction. The extinction of these NTO peaks and appearance of characteristic BTO peaks indicates full conversion of the nanowires in the highly aligned growth case. The XRD pattern of the nanowires grown on the oxidation layer, however, still shows some remnant NTO peaks after conversion which indicates incomplete conversion. Therefore, the hydrothermal conversion procedure converted the NTO to BTO, while preserving the nanowire morphology and alignment. This alignment is typically quantified by using the Lotgering degree of orientation, which compares the relative intensities of reflections present in a coupled θ–2θ powder diffraction pattern.³⁰ However, this technique provides differing values for degrees of orientation as a function of the reflection used in the calculation thus resulting in misleading orientation values, as demonstrated by Jones et al.³¹ Therefore, in order to quantify the degree of orientation, a commonly employed technique for measuring the orientation in semi-crystalline polymers called the Herman's Orientation Factor (HOF) was utilized in this study.^32,33

SEM images of the NTO nanowires were acquired at a 30° tilt using the same beam intensity, operating voltage, and magnification for both the aligned and highly aligned scenarios, as shown in Figs. 4(a) and 4(d). After converting these NTO nanowires to BTO, similar SEM images were acquired for both growth scenarios. The aligned BTO nanowires are shown in Fig. 4(g) and the highly aligned BTO nanowires are shown in Fig. 4(j). Using identical SEM imaging parameters allows for an accurate comparison of degree of orientation. These SEM images were analyzed using a 2-dimenstional Fast Fourier transform (FFT) to determine the HOF value. HOF is defined as a number between −0.5 and 1 relating the degree of orientation of a line structure of interest to an arbitrary reference line. In relation to the reference direction, the values of −0.5 and 1 represent perfect perpendicular and parallel alignment, respectively. A perfectly random orientation is given a value of 0. In this analysis, the grayscale SEM images were converted to the frequency domain from the spatial domain. The intensity of this mapped frequency domain indicates the distribution of nanowires alignment in the SEM images. The FFT maps for the two different cases of NTO nanowires are shown in Figs. 4(b) and 4(e), and the BTO nanowire FFT maps are shown in Figs. 4(h) and 4(k). For these maps, a high intensity symmetrical circle at the center of the map would indicate a perfectly randomly distributed nanowire orientation.³⁴ Elongation in one axis forming an ellipse represents preferred orientation in the nanowires. As shown, there is clearly more elongation in the FFT intensity maps in Figs. 4(e) and 4(k) as compared to Figs. 4(b) and 4(h), thus indicating that the nanowires grown from the textured substrate are more highly aligned than the ones grown off the non-textured oxidation layer. This matches the observation seen previously in Fig. 1.

Utilizing the FFT intensity maps, the HOF can be calculated by first creating intensity-azimuthal angle curves through azimuthal averaging. The azimuthal angle, $ϕ$⁠, is illustrated in the FFT intensity maps in Figs. 4(b), 4(e), 4(h), and 4(k). This angle is defined as the angle between the orientation reference axis and the component of interest's direction. Upon defining this angle, the HOF can be calculated using the following equation:

The angular brackets are a spatial average that account for the spatial distribution of the degree of orientation. This average is defined as follows:

where I($ϕ$) is the intensity profile of anisotropy.^35–37 Intensity-azimuthal angle profiles were taken from the FFT intensity maps at various distances from the origin along quarter-circle projections. A representative quarter-circle projection path is illustrated in Figs. 4(b), 4(e), 4(h), and 4(k), and the resulting intensity profiles from selecting multiple quarter-circle paths are presented in Figs. 4(c), 4(f), 4(i), and 4(l). Seven profiles were taken from each FFT map within close proximity of the origin as to avoid the low intensity background noise seen at a significant distance from the origin. Although these seven profiles showed the same trend, they were averaged as outlined above in order to produce a single HOF value for each growth scenario. For all four FFT maps, the same seven profile distances from the origin were used to create the intensity-azimuthal angle profile graphs. Processing the nanowire images under identical conditions gives an accurate comparison of the difference in degree of alignment. The HOF values were 0.3465 and 0.5476 for the aligned and highly aligned NTO samples, respectively. This results in an overall improvement of 58% in degree of alignment by utilizing a textured TiO₂ substrate as compared to a non-textured TiO₂ substrate. This improved alignment is preserved during the BTO conversion procedure. For the aligned and highly aligned BTO samples, the HOF values were 0.3065 and 0.4899, respectively, which equates to a 60% improvement in alignment as a result of growing from a textured substrate. Therefore, this HOF analysis shows that the nanowire alignment is significantly improved by synthesizing the nanowires on a textured TiO₂ film as compared to a non-textured TiO₂ oxidation layer, and the nanowire alignment is maintained when converting the NTO to BTO.

This research developed a synthesis methodology to produce high aspect ratio, highly aligned ferroelectric nanowire arrays utilizing versatile hydrothermal reactions. By using a textured TiO₂ film as opposed to a non-textured oxide layer to grow NTO, the degree of orientation of the nanowires was increased by 58% according to HOF calculations. The hydrothermal parameters used to convert the NTO to BTO maintained the nanowire alignment and resulted in a 60% improvement in orientation. This approach for synthesizing highly aligned, high aspect ratio ferroelectric nanowire arrays could see future implementation in high-performance microelectromechanical and nanoelectromechanical systems used in power harvesting and sensing applications.

The authors thank Dr. B. L. Lee and the Air Force Office of Scientific Research for the support for this research.