Daiki Fujioka*a,
Shingo Ikedab,
Kensuke Akamatsuc,
Hidemi Nawafunec and
Kazuo Kojimaa
aDepartment of Applied Chemistry, College of Life Sciences, Ritsumeikan University, 1-1-1 Noji-Higashi, Kusatsu-City, Shiga 525-8577, Japan. E-mail: rc007040@ed.ritsumei.ac.jp; Tel: +81-77-561-2780
bElectronic Materials Research Division, Osaka Research Institute of Industrial Science and Technology, 1-6-50 Morinomiya, Joto-ku, Osaka 536-8553, Japan
cDepartment of Nanobiochemistry, Frontiers of Innovative Research in Science and Technology (FIRST), Konan University, 7-1-20 Minatojimaminami, Chuo-ku, Kobe 650-0047, Japan
First published on 22nd February 2019
Nickel-nanoparticle-containing polyimide composite films were prepared by liquid-phase reduction of Ni2+ ions with potassium borohydride (KBH4). The nanoparticles were amorphous with diameters of approximately 10–20 nm, depending on the KBH4 concentration and reduction temperature. At high KBH4 concentrations, the nanoparticles appeared to contain various nickel boride species. The number of nanoparticles and Ni content both increased upon repeated adsorption/reduction of Ni2+ ions, where the particle growth was inhibited by the rigid polymer chain and the formation of smaller particles was favored.
Composite materials based on nanoparticles and polymers can be prepared via various methods, such as dispersing or dissolving the components in an appropriate solvent and pouring the resulting solution into a mold or onto a substrate followed by drying and solidification.13–15 However, these methods provide limited control over the spatial distribution of the particles, and close attention must also be paid to the dispersibility and solubility of the nanoparticles and polymer in the solvent. Sputtering and vacuum vapor deposition methods may also be used to introduce nanoparticles into polymer thin films synthesized by plasma polymerization,5 although these techniques require large-scale and expensive apparatuses and substantial amounts of energy, which leads to high manufacturing costs.
In situ synthesis methods, in which the metal nanoparticles are precipitated inside the polymer, have also been developed. In this approach, metal ions introduced into the polymer are reduced by chemical or thermal treatment to afford metal nanoparticles.16–18 The spatial distribution of the metal nanoparticles in the composite material can be controlled by uniformly dispersing the metal ions throughout the polymer matrix19–21 with minimal material loss and energy wastage.
Various studies on dispersing metal nanoparticles in polyimide resins have been conducted.19,22 A method for forming highly dispersed metal nanoparticles inside a polyimide resin was proposed by the Nawafune–Akamatsu group. First, the polyimide resin was immersed in an alkaline solution to generate carboxyl groups for cation exchange, allowing a precursor resin substituted with the metal ions to be prepared. This precursor resin was then heated under a hydrogen atmosphere to reduce the metal ions.23,24
One advantage of this method is that the thickness of the surface layer formed by the nanoparticles, amount of metal ions, particle size, interparticle distance, and microstructure can be controlled via parameters such as the concentration of the alkaline solution, treatment temperature, and immersion time. Furthermore, the thermal reduction under a hydrogen atmosphere causes the polyamic acid resulting from hydrolysis of the imide rings to revert to polyimide through a dehydration reaction. Consequently, although the polyimide resin retains its physical properties, it is difficult to introduce additional metal ions to obtain new particles since the cation-exchange groups are lost during thermal reduction.
In this study, we report on the convenient preparation of a nickel nanoparticle (NiNP)-dispersed composite polymer using a high-concentration solution of potassium borohydride as the reducing agent.
The process used to prepare the NiNP-containing polymer was as follows: (1) introduction of carboxyl groups into the polyimide resin for cation exchange by surface modification using potassium hydroxide, (2) preparation of a precursor resin in which the cation-exchange groups were substituted with nickel ions, and (3) precipitation of the nickel by liquid-phase reduction using aqueous KBH4 solution.
We also report on the influence of the reducing agent concentration and reduction temperature on the microstructure of the metal nanoparticle–polyimide composite, and the effects of the repeated adsorption/reduction of metal ions.
To investigate the effect of repeated reduction on the morphology of the NiNPs, the ion adsorption/reduction treatments were repeated for three, five, or seven cycles at a temperature of 30 °C.
These two absorption bands are not present in Fig. 1(b). However, an absorption band at 1550 cm−1 and a broad absorption band at 1500–1700 cm−1 can be observed; the former corresponds to the N–H bending vibration of the amide bonds formed by hydrolysis of the imide ring, while the latter is considered to originate from the stretching vibration of carboxyl groups complexed with potassium ions. This is in good agreement with a previous report,25 indicating the formation of cation-exchange groups on the film surface upon KOH treatment.
The nickel ions formed metallic clusters that afforded a weak peak on the high-angle side of the XRD patterns.
Fig. 3 shows the results of quantitative analysis of nickel and boron ions in the samples after the Ni2+ reduction treatment. It was found that the amount of deposited Ni did not change although the reducing agent concentration was different. The average Ni content in the precursor films after the introduction of Ni2+ ions was 1.48 μmol cm−2. Residual Ni2+ ions inside the resin were removed by immersing the film in aqueous CH3COOH solution. In other words, if unreduced Ni2+ ions were present, the amount of deposited Ni should be lower than the initially introduced amount. Therefore, it is considered that all of the introduced Ni2+ ions were deposited as metallic nickel inside the polyimide resin.
The boron content in the samples was found to increase with increasing reducing agent concentration. Assuming that all of the boron atoms were mixed with the deposited nickel atoms, 6–8 wt% of boron should be contained in the deposited nickel. It has been reported that metallic nickel films produced by electroless plating are amorphous when the boron content exceeds approximately 3 wt%.26–30 It has also been reported that Ni nanoparticles become amorphous when phosphorus31–33 or boron34–38 is contained in the nanoparticles. Furthermore, when boron-containing nickel films are heated at 300–400 °C, nickel borides are formed. As shown in Fig. 2, no diffraction patterns corresponding to nickel borides were observed for the samples annealed at 300 °C. Therefore, it is considered that the detected boron content originated not from the deposited nickel but from reducing agent residue inside the modified layer.
Fig. 4 shows the cross-sectional TEM images, electron diffraction patterns, and Ni particle size distributions of the Ni-NPs/PI films prior to annealing. The electron diffraction patterns show halo patterns. This is the same tendency as the result shown by the XRD patterns in Fig. S1.† The nanoparticles are amorphous. The particle size distributions were obtained by counting the particles in the area (600 × 450 nm2) shown in Fig. 4. The mean diameters of the nanoparticles in the samples reduced with 0.06, 0.12, and 0.20 mol dm−3 KBH4 solutions were 19 ± 5, 14 ± 3, and 13 ± 4 nm, respectively. Therefore, the average particle size decreased with increasing KBH4 concentration. This is because nucleation predominates over nucleus growth and the rate of nucleation would be expected to increase with increasing KBH4 concentration. In general, particle deposition begins with a nucleus that initially acts as a seed for the growing particle before increasing in size. However, at low KBH4 concentrations, the BH4− ion does not diffuse deeply into the resin. Therefore, particle nucleation preferentially occurs near the resin surface where the reducing agent can penetrate. It is considered that this nucleus acts as a reduction centre to which Ni2+ ions diffuse, resulting in the growth of Ni particles. As the KBH4 concentration increases, the reducing agent can penetrate deeper into the reformed layer, such that the nucleus precipitates inside the reformed resin layer.
Fig. 4 Cross-sectional TEM images and electron diffraction patterns of the Ni-NPs/PI films (B)–(D) shown in Fig. S1.† Particle size distributions of Ni nanoparticles are also shown. |
In addition, when the nucleation rate is higher, nucleation prevails over nanoparticle growth. Although the amount of Ni ions initially introduced was the same, the Ni was preferentially consumed for nucleation, and thus nanoparticle growth was suppressed. Therefore, the particle size decreased with increasing reducing agent concentration.
Fig. 5 shows the cross-sectional TEM images and EDS elemental analysis results for the Ni-NPs/PI films prior to annealing. The EDS elemental analysis of Ni was performed during TEM observation using a 15 nm electron beam and the same measurement time (25 s). The EDS spectra revealed a higher Ni abundance for the nanoparticles compared with the area lacking nanoparticles. No lattice images of nanoparticles were observed in high-resolution TEM images. An FFT image of the particle a supports this result. Therefore, based on the XRD, EDS, and high-resolution TEM results, the formed nanoparticles are considered to be composed of amorphous Ni metal. Moreover, small amounts of Ni were detected in the area of the resin surrounding the particles, possibly indicating Ni clusters that could not be observed by TEM.
Nanoparticles with an average size of 3–12 nm were observed in the samples reduced at temperatures of 30, 40, and 50 °C. Fig. S2† shows cross-sectional TEM image and particle size distributions of a Ni-NPs/PI film reduced at 50 °C. The electron diffraction and EDS results revealed that the formed NiNPs consisted of metallic nickel with an fcc structure. The average particle size of the nanoparticles in the sample reduced at 30 °C was 11 nm, and smaller particles with a size of approximately 4 nm were also formed in the samples reduced at 40 and 50 °C. The number of these smaller particles increased with increasing reduction temperature, and the number of particles around 10 nm also increased and their sizes tended to decrease. Therefore, a low reduction temperature promoted nucleus growth, resulting in increased particle size. Conversely, a high temperature increased the rate of the reduction reaction and led to faster nucleation, thereby increasing nucleation progresses relative to nucleus growth and leading to an increased number of particles and decreased particle size.
Fig. 6 shows that the Ni content increased with an increasing number of adsorption/reduction cycles, confirming the feasibility of increasing the amount of Ni in the polyimide resin by repeated adsorption/reduction treatment.
Fig. S4† shows cross-sectional TEM images of the samples subjected to repeated adsorption/reduction of Ni2+ ions. The formation of nanoparticles with an average particle size of 10–30 nm was confirmed in all of the samples. The surfaces of the resins subjected to three, five, and seven adsorption/reduction cycles were not covered with films, but were instead occupied by a large amount of deposited nanoparticles. In addition, the particles were dispersed throughout the reformed layer, and tended to be deposited in the deepest part of the reformed layer. Furthermore, comparison of the samples subjected to three, five, and seven adsorption/reduction cycles revealed that large particles were dispersed throughout the entire reformed layer in the sample subjected to seven cycles.
As shown by the particle size distributions in Fig. S4,† the average particle size was approximately 28 nm for the three samples. To obtain these distributions, all of the particles in the entire region (1.6 × 1.2 μm2) of the TEM images were counted. This average particle size of 28 nm is more than double that observed for the sample subjected to a single adsorption/reduction cycle. Furthermore, increasing the number of cycles from three to seven had little effect on the number of particles with a size of 30 nm or more and no change in the total number of particles was observed. This finding seems to contradict the results shown in Fig. 6, where the Ni content was found to increase with an increasing number of cycles. This suggests that an increased number of cycles resulted in an increased number of particles with diameters smaller than those shown in Fig. S4.†
Fig. 7 shows high-resolution cross-sectional TEM images and the number of particles per unit area for the samples subjected to repeated adsorption/reduction of Ni2+ ions; the arrows in Fig. 7(a)–(c) indicate the presence of numerous particles with sizes of several nanometers or less. However, since these particles are small in size and located inside the resin, the contrast of the particle image is not clear. Since it was difficult to precisely specify the particle size, the number of identified particles was counted in the whole region (250 × 150 nm2) shown in the enlarged view. The number of particles per unit area is shown in Fig. 7(b). The results revealed that the number of particles increased with an increasing number of cycles. This is consistent with the tendency shown in Fig. 6 for the amount of Ni. It is presumed that the rigid polymer network inhibited the growth of particles with sizes of 30 nm or greater; therefore, the number of large particles remained approximately constant while the number of small particles increased.
Fig. 7 High resolution cross sectional TEM images of the samples prepared by repeated Ni2+ ion adsorption/reduction treatments with for (a) three, (b) five, and (c) seven cycles. The counts in (d) were obtained for each area of 250 × 150 nm2 in (a)–(c). The counts for one repetition number in (d) were obtained from the particle size distribution for an area of 250 × 150 nm2 in the area d in Fig. 5. |
Fig. S5† presents electron diffraction patterns of the samples. The sample subjected to three adsorption/reduction cycles exhibited a halo pattern, whereas the samples subjected to five or seven cycles afforded spot-like diffraction images similar to the image observed for the polycrystalline phase. However, it was difficult to interpret the diffraction images owing to the large number of spots present. Based on the ICDD data, the XRD patterns of Ni with the fcc structure, NiB, Ni2B, and Ni3B were compared, as shown in Fig. S6.† Numerous peaks corresponding to nickel borides were observed on both the low- and high-diffraction-angle sides of the Ni(111) and Ni(200) peaks. As shown in Fig. S5(c) and (d),† diffraction spots were also observed at positions closer to the Ni(111) and Ni(200) planes of fcc Ni. The above results indicate that nickel boride nanoparticles were present in parts of the deposited material. The formation of nickel borides with high crystallinity is probably attributable to the facile incorporation of boron atoms into the precipitates upon prolonged exposure to high concentrations of KBH4.
Nanoparticles with an average size of 3–12 nm were formed in the samples reduced at temperatures of 30, 40, or 50 °C. The number of these small particles increased with increasing temperature, and the number of particles around 10 nm also increased and their sizes tended to decrease. Higher temperatures led to an increased nucleation rate and favoured nucleation rather than nucleus growth, resulting in an increased number of particles and decreased particle size.
Samples subjected to three, five, or seven adsorption/reduction cycles were found to possess an average particle size of approximately 28 nm. The surfaces of the resins subjected to these adsorption/reduction cycles were not covered with films, but were instead occupied by a large amount of deposited nanoparticles. The number of particles increased with an increasing number of cycles. This is consistent with the tendency observed for the Ni content with an increasing number of cycles. It is considered that the rigid polymer network inhibited the growth of particles with sizes of 30 nm or greater, causing the number of large particles to remain approximately constant while the number of smaller particles increased. It was therefore possible to increase both the number of particles and Ni content by performing repeated adsorption/reduction cycles. Electron diffraction results and ICDD data indicated that nickel boride nanoparticles with high crystallinity were present in parts of the deposited material. The formation of these nickel borides is probably attributable to the facile incorporation of boron atoms into the precipitates upon prolonged exposure to high concentrations of KBH4.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9ra00182d |
This journal is © The Royal Society of Chemistry 2019 |