Research Article www.acsami.org
Hydrothermal Synthesis and Processing of Barium Titanate Nanoparticles Embedded in Polymer Films Michael D. Toomey,† Kai Gao,† Gamini P. Mendis,† Elliott B. Slamovich,† and John A. Howarter*,†,‡ †
School of Materials Engineering, Purdue University, 701 W. Stadium Avenue, West Lafayette, Indiana 47907, United States Division of Environmental and Ecological Engineering, Purdue University, 500 Central Drive, West Lafayette, Indiana 47907, United States
Downloaded via UNIV OF WINNIPEG on January 23, 2019 at 14:44:55 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
ABSTRACT: Barium titanate nanoparticles embedded in flexible polymer films were synthesized using hydrothermal processing methods. The resulting films were characterized with respect to material composition, size distribution of nanoparticles, and spatial location of particles within the polymer film. Synthesis conditions were varied based on the mechanical properties of the polymer films, ratio of polymer to barium titanate precursors, and length of aging time between initial formulations of the solution to final processing of nanoparticles. Block copolymers of poly(styrene-co-maleic anhydride) (SMAh) were used to spatially separate titanium precursors based on specific chemical interactions with the maleic anhydride moiety. However, the glassy nature of this copolymer restricted mobility of the titanium precursors during hydrothermal processing. The addition of rubbery butadiene moieties, through mixing of the SMAh with poly(styrenebutadiene-styrene) (SBS) copolymer, increased the nanoparticle dispersion as a result of greater diffusivity of the titanium precursor via higher mobility of the polymer matrix. Additionally, an aminosilane was used as a means to retard cross-linking in polymer−metalorganic solutions, as the titanium precursor molecules were shown to react and form networks prior to hydrothermal processing. By adding small amounts of competing aminosilane, excessive cross-linking was prevented without significantly impacting the quality and composition of the final barium titanate nanoparticles. X-ray diffraction and X-ray photoelectron spectroscopy were used to verify nanoparticle compositions. Particle sizes within the polymer films were measured to be 108 ± 5 nm, 100 ± 6 nm, and 60 ± 5 nm under different synthetic conditions using electron microscopy. Flexibility of the films was assessed through measurement of the glass transition temperature using dynamic mechanical analysis. Dielectric permittivity was measured using an impedance analyzer. KEYWORDS: functional nanocomposite, polymer thin film, barium titanate, dielectric nanoparticle, hydrothermal processing
■
INTRODUCTION The use of barium titanate (BaTiO3) nanoparticles as a functional additive in polymer nanocomposites has been heavily researched due to its favorable dielectric,1−4 ferroelectric,5,6 and piezoelectric7,8 properties which offers improvements in performance of devices such as high power multilayer capacitors9 and transistors,10 by increasing dielectric loss, and enhancing flexibility.11 Traditionally, the fabrication of a BaTiO3 nanocomposite system is done by casting a polymer solution or a polymer melt loaded with presynthesized BaTiO3 nanoparticles.3,4,12,13 In situ sol−gel processing is another method of fabricating BaTiO3 nanocomposites which has been utilized.14−19 This method involves casting films of precursorcontaining polymer solution, and reacting them hydrothermally in a hydroxide solution to precipitate particles from the precursor sites.20−23 Although both methods have been successful in fabricating dielectric nanocomposite films, one problem with these methods is that the BaTiO3 particles have a tendency to aggregate during processing, which leads to inconsistencies in material performance.13,25−29 In the traditional solution/melt © 2015 American Chemical Society
processing method, powder aggregation occurs, unless dispersion agents are added.26−29 For in situ sol−gel processing, poor solubility of the precursor in the polymer matrix can contribute to particle clustering. Additionally, gelation can occur as polymer−precursor solution ages, further impacting the sensitivity of these systems to processing protocols.24 In this work, we attempt to fabricate a flexible polymer nanocomposite film with BaTiO3 nanoparticles by in situ processing and hydrothermal synthesis. We hypothesize that, matching the solubility of the titanium precursor with the polymer composition, we can improve the dispersion of BaTiO3 nanoparticles in the final film by reducing the propensity for the precursors to aggregate. Additionally, the use of block copolymers which can phase segregate will further discourage particle aggregation. The addition of a monofunctional silane as a processing aid will retard the gelation of the Received: October 27, 2015 Accepted: December 3, 2015 Published: December 3, 2015 28640
DOI: 10.1021/acsami.5b10282 ACS Appl. Mater. Interfaces 2015, 7, 28640−28646
Research Article
ACS Applied Materials & Interfaces Scheme 1. TBT Reacts with the Carbonyl Structures of the Maleic Anhydridea
a
Further reaction between the alkyl ether and other maleic anhydride moieties creates interchain crosslinks and forms a gel.
Scheme 2. Stabilization of the TBT Molecule Occurs through Reaction with APDESa
a
This silanization hinders crosslinking in closed conditions, increasing shelf life of the TBT solution.
Table 1. Composition of Polymer Solutions samples
TBT (wt %)
SMAh (wt %)
SBS (wt %)
APDES (wt %)
toluene (wt %)
2P 3P 4P
10 5 5
10 5 5
0 5 5
0 0 4.3
80 85 80.7
prior inverstigations;24 this short time to gelation for the TBT precursor limits the possibility of solution processing thin films in large volumes under neat conditions. To address this processing limitation, functional silanes (APDES or IPDCS) were added to retard crosslinking, therefore, increasing processability. For this work, most systems used APDES exclusively; IPDCS is specifically noted when it was the functional silane additive. Scheme 2 shows the anticipated reaction between APDES and TBT, leading to the increase of time until cross-linking. By hindering network formation, an intermediate solution can be used during a wider processing window without concern of unmitigated cross-linking prior to the final hydrothermal reactions. Three thin film systems were created and will be referred to as 2P, 3P, and 4P, as shown in Table 1. The two-component metalorganic precursor system (2P) was created as a solution containing SMAh and TBT in toluene. The three-component metalorganic system (3P) was created as a solution containing SMAh, SBS, and TBT in toluene. SBS was added as a means to improve flexibility in the cast film. The fourcomponent metalorganic precursor system (4P) was created as a solution containing SMAh, SBS, TBT, and APDES in toluene. The ratio between TBT and silane was chosen so that the number of silane molecules present in the solution would be enough to react with approximately half of the TBT bonding sites, in order to impede network formation in the polymer solution. The 2P and 3P solutions were created by first dissolving the polymers in toluene over 24 h, then quickly adding TBT. The 4P solution was prepared by dissolving polymer in toluene over 24 h, and slowly adding a solution of TBT, APDES, and toluene, dropwise into the polymer solution. Whenever the APDES was added too quickly, or was not first diluted in toluene, gelation in the film occurred quickly in local regions of the solution. By diluting the TBT−APDES solution and adding it to the polymer solution dropwise, this rapid gelation was avoided. After the polymer solutions were made, they were spin-cast, using a Laurell model WS-650HZB-32NPP (Laurell Corporation, New Wales,
solution, by impeding the ability of the titanium precursor to cross-link with itself.30,31 Chemical composition of the nanocomposite will be determined by X-ray diffraction and X-ray photoelectron spectroscopy. Particle size and morphology will be measured using scanning electron microscopy. Bulk film mechanical and dielectric properties will be measured using a dynamic mechanical analyzer, and an impedance analyzer, respectively.
■
EXPERIMENTAL METHODS
Poly(styrene-co-maleic anhydride) (SMAh, Mw = 224 000 g/mol, 7 wt % maleic anhydride), poly(styrene-butadiene-styrene) (SBS, Mw = 140 000 g/mol, DPstyrene = 200 per block, DPbutadiene = 1800 per block), titanium butoxide (TBT, 97% pure), toluene (99% pure), sodium hydroxide (NaOH), and barium hydroxide octahydrate (Ba(OH)2· 8H2O, ≥98% pure) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Aminopropyl dimethyl ethoxysilane (APDES, ≥95% pure) and 3-isocyanatopropyl dimethyl chlorosilane (IPDCS) were purchased from Gelest Chemical Company (Morrisville, PA, USA). All chemicals were used without further purification. ̀ The synthetic method used in this work to create BaTiO3 nanocomposite films is modified from a previously reported method.24 In this method (Scheme 1), TBT precursor binds with the maleic anhydride portion of the SMAh, one ligand of the TBT molecule is removed, and the anhydride group opens, binding to the TBT molecule. Because of this newly formed bond, steric separation between TBT molecules bound to maleic anhydride may occur, encouraging homogeneous nanoparticle formation and dispersion throughout processing. Cross-linking of the neat polymer−TBT solution occurred within 2 h of mixing due to the quaternary ligand structure of the molecule, as compared to 24−48 h using other titanium precursors, for example, titanium diisopropoxide bis(ethylacetoacetate) (TIBE), as reported in 28641
DOI: 10.1021/acsami.5b10282 ACS Appl. Mater. Interfaces 2015, 7, 28640−28646
Research Article
ACS Applied Materials & Interfaces PA, USA) on single crystal Si substrates (Wafer World, West Palm Beach, FL, USA) at 2000 rpm for 60 s. After spin-casting, the thin film samples were reacted hydrothermally in 1 M Ba(OH)2 at 80−90 °C for 2 h. To ensure similar conditions, samples were placed in the Ba(OH)2 bath 10 min after being spin-cast. After the hydrothermal reaction, samples were rinsed with ethanol, followed by a pH 10 rinse in diluted NaOH solution, and then rinsed again in ethanol to remove barium carbonate (BaCO3) adsorbed to the film surface. One sample from each treatment was Au−Pd sputtered for scanning electron microscope (SEM) imaging. In order to obtain particle size distribution, homogeneity, film thickness, and elemental surface composition of the post-hydrothermally reacted films, several characterization methods were utilized. X-ray photoelectron spectroscopy (XPS) was performed in order to attain a quantitative elemental composition using a Kratos X-ray photoelectron spectrometer (Kratos Analytical, Manchester, U.K.). Xray diffraction (XRD) was performed on the films to verify the precipitation of BaTiO3. XRD scans were taken using Cu−Kα radiation, a scanning window of 30−60° 2θ, and a scan time of 30 min using a Bruker D8 Discover Diffractometer (Bruker, Madison, WI, USA). Sputtered samples were imaged in cross section and plane view with respect to the spin-casting direction, using a XL40 SEM (JEOL, Peabody, MA, USA) at varying accelerating voltages and spot sizes. Film surfaces were imaged both before and after hydrothermal processing in 1 M Ba(OH)2 solution. Particle size analysis was performed using ImageJ software. Film flexibility was assessed by measuring the glass transition temperature using a Q800 dynamic mechanical analyzer (DMA) (TA Instruments, Schaumburg, IL, USA). DMA was performed at a testing frequency of 1 Hz, oscillation amplitude of 10 μm, and a static force of 0.01 N. In order to provide a comparative assessment of the dielectric properties, dielectric permittivity measurements were made of tape cast 2P, 3P, and 4P films before and after hydrothermal reaction. Dielectric permittivity was measured using an Agilent 4291B RF Impedance Analyzer (Agilent Technologies, Santa Clara, CA, USA).
and 55.5° represent the (110), (111), (200), (210), and (211) families of planes for BaTiO3, respectively. The peaks align with results attained by other research, and confirm the conversion of Ba(OH)2 to BaTiO3.32,33 The small peak shifts and peak broadening are due to hydroxyl ion impurities from the hydrothermal processing method.34 In order to further support results from XRD analysis, XPS was performed to analyze the surface chemistry of all presoak and postsoak thin films by examining the barium and titanium signals from the XPS spectra. The spectra were normalized to the postsoak Ti 2p 3/2 peak so that relative intensities of the various compounds can be compared. In 2P, 3P, and 4P films, BaTiO3 was formed between the presoak and the postsoak conditions, as can be clearly observed in the high-resolution spectra for the barium (binding energy 770−805 eV) and titanium (binding energy 450−475 eV). The peaks at binding energies 780 and 795 eV in the postsoak films indicate the evolution of a chemical species containing barium, as shown in Figure 2a. The Ti spectra of the postsoak samples (Figure 2b) exhibit a downward shift in binding energy from 458.9 eV to a new binding energy of 458.2 eV. This binding energy falls in the range of values reported for BaTiO3.35 The ratios of Ti to Ba in 2P, 3P, and 4P postsoak films were 1.25, 1.10, and 0.86, respectively, as shown in Table 2. This ratio near 1:1 further suggests the formation of BaTiO3 during hydrothermal reaction. In the 2P and 3P samples, where less Ba is present than Ti, it is assumed that not all of the Ti-rich precursor was converted into BaTiO3. In the case where less Ti is present than Ba (4P), it is assumed that a small amount of BaCO3 remains adsorbed to the surface after washing. The atomic ratios, Ti and Ba binding energies, when combined with XRD, are good evidence that the hydrothermal processing resulted in the formation of BaTiO3. Impedance analysis was used to find the dielectric permittivity of the nanocomposite films at 10 MHz. The 3P and 4P films had dielectric permittivity values of 2.3 and 2.6, respectively, which are in agreement with previous results;20 the 2P films were not able to be measured as they were too brittle to be handled postsynthesis. This result further indicates the need for selecting polymers with both chemical and mechanical compatibility for processing of functional nanocomposites. DMA was used to measure the glass transition temperature of the 2P, 3P, and 4P films for the verification of the improved flexibility in the SBS added systems. All films were first hydrothermally reacted before analysis, with exception to the 2P films, which became too brittle to test without failure due to cracking. This was to be as expected, given the glassy nature of SMAh, and as a result, 2P films were tested in their presoak condition. Through the observation of the loss tangent, the 2P films showed a glass transition temperature of approximately 155 °C. This high glass transition temperature explains the brittle behavior of the 2P system at room temperature. The addition of SBS in the 3P and 4P films led to a significant shift of glass transition temperature from 155 °C to approximately −70 °C. This means that, at room temperature, the film behaves in a flexible manner, since it is well above its glass transition temperature. This result supports the idea that the blending of a glassy SMAh polymer with a more flexible SBS polymer increases the flexibility of the film. SEM micrographs were taken to examine the morphology of the films before and after hydrothermal treatment. Figures 3a, 4a,c, and 5a show SEM images of 2P, 3P, and 4P films, respectively, before reacting in Ba(OH)2 solution. All films
■
RESULTS AND DISCUSSION XRD analysis was performed to verify that BaTiO3 nanoparticles had formed as a result of the exposure of the films to the hydrothermal reaction with Ba(OH)2. Figure 1 shows XRD scans from 2P, 3P, and 4P films, before and after hydrothermal processing. Peaks located at approximately 31.5, 38.5, 44.5, 51,
Figure 1. Formation of BaTiO3 in the films can be seen through the XRD spectra comparison between presoak and postsoak conditions of 2P, 3P, and 4P films. Peak shifting is due to hydroxyl ions from the hydrothermal synthesis, which causes lattice dilation. 28642
DOI: 10.1021/acsami.5b10282 ACS Appl. Mater. Interfaces 2015, 7, 28640−28646
Research Article
ACS Applied Materials & Interfaces
Figure 2. Formation of BaTiO3 can be seen in the XPS spectra of (a) Ba and (b) Ti in pre- and postsoak film samples. The enhanced Ba signal in the postsoaked samples and the shift to higher Ti binding energies are indications of the formation of BaTiO3.
clustering. Because the SMAh is below its glass transition temperature36 of 120 °C, the polymer is glassy and, therefore, the diffusion of the titanium precursors and barium atoms is limited in the 2P films, resulting in particle formation, which is dictated by the copolymer morphology. It is believed that the significant oversaturation of titanium precursors in the system described by Collins and Slamovich is such that the copolymer morphology does not significantly impact the formation of BaTiO3 at the surface. 3P films showed both large- and small-scale segregation due to polymer immiscibility. Figure 4a,b shows large-scale segregation in the form of a second phase region with an ellipsoidal morphology of sizes ranging 10−30 μm. Although there is microscale phase segregation, the particle distribution remains homogeneous across the film (Figure 4b). The smallscale segregation can be seen in Figure 4c in the form of light regions with approximate diameters of 200 nm. This nanoscale segregation is believed to be phase separation of immiscible blocks of the polymer. In some washing circumstances, particles can be preferentially removed from one phase (Figure 4e). Upon addition of the TBT to the 2P and 3P systems, gelation occurred within 2 h; however, with the addition of APDES (or IPDCS), the 4P system exhibited an extended time until cross-linking. The addition of APDES to the solution is
Table 2. Atomic Ratios and Binding Energies of Ba and Ti As Determined by XPS samples
Ti:Ba atomic ratio
Ti 2p binding energies (eV)
Ba 3d binding energies (eV)
2P 3P 4P
1.25 1.10 0.86
459, 465 459, 465 459, 465
780, 795 780, 795 780, 795
showed light cracking, which may be due to rapid drying caused by solvent evaporation during spin-casting. Additionally, there were no particles observed at the surface of the films prior to hydrothermal reaction. Figures 3b, 4b,d, and 5b show SEM images of 2P, 3P, and 4P films, respectively, after reacting in Ba(OH)2 solution. The cracks in the 3P and 4P films are no longer present after the Ba(OH)2 bath, likely due to annealing of the polymer during hydrothermal treatment at elevated temperature. Cracking is still visible in the postsoak 2P film, which suggests that the hydrothermal reaction is occurring while the film is still in its glassy state. Upon hydrothermal treatment, BaTiO3 particles were formed on the surface of the polymer films. In 2P films, BaTiO3 surface formation occurred in clusters with particle diameters of 108 ± 5 nm. This differs from results reported by Collins and Slamovich,24 which showed BaTiO3 nanoparticles without
Figure 3. Microstructures of the 2P film (a) before and (b) after the reaction with Ba(OH)2 are seen in these SEM micrographs. 28643
DOI: 10.1021/acsami.5b10282 ACS Appl. Mater. Interfaces 2015, 7, 28640−28646
Research Article
ACS Applied Materials & Interfaces
Figure 4. Morphology of 3P films shown: (a) and (c) before reaction; (b), (d), and (e) after reaction; and (f) cross section, after reaction. Both cracking and micro- and macroscale phase separation are visible in the polymer films (a) and (c). The homogeneous dispersion of BaTiO3 particles is visible on the surface of the films (b) and (d) after hydrothermal reaction, but is not visible in the cross section (f). BaTiO3-poor phases can be seen in some formulations of the 3P film (e).
likely to slow down the gelation reaction of the TBT by occupying cross-link sites with more stable ligands. The 4P film formed homogeneously dispersed particles (Figure 5b) with an average diameter of 60 ± 5 nm on the surface. This decrease in particle size suggests that freshly made solutions, with the addition of APDES, lead to the precipitation of smaller BaTiO3 particles, given constant reaction time, temperature, and concentration of Ba(OH)2 solution. BaTiO3
used in dielectric applications exhibits a critical particle size, where a shift from the desired tetragonal structure to an undesirable cubic structure occurs in BaTiO3.37−39 There is some variability in the reported critical size; however, most critical size values fall in the range of 20−100 nm (c.f. Table 3). We speculate that, by manipulating the composition of the 3P4P precursor solution, the particle size and distribution can be 28644
DOI: 10.1021/acsami.5b10282 ACS Appl. Mater. Interfaces 2015, 7, 28640−28646
Research Article
ACS Applied Materials & Interfaces
Figure 5. Microstructures of the 4P films in the (a) presoak, (b) postsoak, and (c) 3 week aged conditions can be seen in these SEM micrographs. Evidence of phase separation can be seen in (a) and (c); however, (b) exhibits a homogeneous dispersion of BaTiO3 particles.
locations along the length of the substrate. Also, tearing occurred in the films containing SBS, despite the use of liquid nitrogen to bring the polymer below its glass transition temperature; this suggests that SBS increases ductility and molecular mobility in the film.
Table 3. Particle Size Distributions of Hydrothermally Reacted Particles samples 2P 3P 4P 4P-3weeks
average particle size (nm) 108 100 60 158
± ± ± ±
5 6 5 10
additional observations
■
clustered nanoparticle formation homogeneous dispersion homogeneous dispersion inhomogeneity, increased nanoparticle size
CONCLUSIONS Polymer films embedded with BaTiO3 nanoparticles were hydrothermally synthesized using combinations of SMAh, SBS, TBT, and silane. Dielectric permittivity was found to be in agreement with previously reported work, with similar precursor loading conditions. Mechanical testing of the nanocomposite films displayed that the addition of SBS lowered the glass transition temperature below the processing temperature. The influence of chemistry and mechanical properties of the polymer matrix on particle size and homogeneity was measured in each film. The formation of BaTiO3 was verified by XRD and XPS, and further confirmed by SEM. The SMAh films precipitated BaTiO3 in a clustered formation, which is attributed to the glassy nature of SMAh under reaction conditions. The rubbery contribution of the added SBS promoted the formation of homogeneously dispersed BaTiO3 nanoparticles. The addition of APDES to the SBS−SMAh films retarded cross-linking and decreased the nanoparticle size while maintaining the dispersion of the BaTiO3 particles. Upon aging, the polymer systems with silane processing agents exhibited a larger particle size and phase separation. The plasticization of the glassy SMAh film by the SBS allowed for increased molecular mobility at the film surface and enabled homogeneous dispersion of BaTiO3 nanoparticles, which led to a more optimal hydrothermal synthesis. These processing enhancements enable the fabrication of flexible films with homogeneously dispersed BaTiO3 nanoparticles.
modified, while maintaining the dispersion of the BaTiO3 phase in the film. Samples of the 4P films were fabricated from solutions aged for 3 weeks to examine solution stability. Figure 5c shows the SEM image of the 4P solution cast after 3 weeks aging. No other solution could be cast after aging, due to cross-linking. As can be seen in Figure 5c, precipitation of BaTiO3 still occurred in the film with an increased particle size of 158 ± 10 nm; however, the particle dispersion was less homogeneous. Ostwald ripening may be the cause of the larger nanoparticle size in the aged 4P films, due to the migration of the unbound TBT as the solution ages. The success of APDES in hindering network formation allows for thermodynamically unstable regions of immiscible TBT to reduce surface free energy via Ostwald ripening. This ripening leads to the precipitation of fewer, but larger, BaTiO3 nanoparticles after aging. The 3P and 4P films show a large amount of precipitation at the surface. The 3P sample precipitated homogeneously distributed particles (Figure 4d) with an average diameter of 100 ± 6 nm on the surface. The addition of SBS, in the 3P and 4P film systems, is believed to have plasticized the SMAh, which allows for better mobility of the Ti components. In the microstructure of the 4P sample (Figure 5a), macroscale phase separation was not clearly visible, but the sample exhibited a pitted morphology. Cross sections of the hydrothermally treated BaTiO3 films were examined to determine the particle distribution in the film. Figure 4f shows a cross section of the 3P film after hydrothermal reaction which contained no discernible particles beneath the surface. This image is representative of all of the film cross sections, regardless of composition. This indicates a surface limited reaction, where Ba(OH)2 is unable to diffuse past the surface layer. This is in agreement with other reports of similar systems,24 where surface particle sizes were over an order of magnitude larger than particle sizes located anywhere else in the film. The 2P film (not shown) exhibited some separation from the substrate due to brittle fracture in a few
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Fax: (765) 494-1204. Phone: (765) 496-3103. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS M.D.T. and G.P.M. were supported in this work through the NSF-IGERT: Sustainable Electronics (Grant Number 1144843). K.G. was supported through the Purdue Research Foundation. 28645
DOI: 10.1021/acsami.5b10282 ACS Appl. Mater. Interfaces 2015, 7, 28640−28646
Research Article
ACS Applied Materials & Interfaces
■
(20) Collins, D. E.; Slamovich, E. B. Growth of BaTiO3 in Hydrothermally Derived (
