In situ synthesis of metal/polymer nanocomposite thin films on glass substrates by using highly cross-linked polymer matrices with tailorable ion exchange capabilities

Abstract

We describe the use of ion-exchangeable, highly cross-linked polymer films as a matrix for the synthesis of nanocomposite thin films and patterns with embedded Cu and Ni nanoparticles. We have successfully controlled the size of metal nanoparticles and thus optical and magnetic properties of the films by varying the initial amount of doped metal ions.

---

Inorganic nanoparticles/polymer nanocomposites with compositionally tailorable features have been intensively investigated because they have markedly unique properties based on a combination of the quantum size effect of embedded metal nanoparticles and the chemically tailorable structures of polymer matrices.^1,2 Utilization of these composites in the form of thin films is particularly important for application in optical and electrical devices. Methods for deliberately controlling the size, concentration, and dispersibility of the nanoparticle phase in the composite films are necessary to optimize the emergent physical properties.^3–5 Of these parameters, the realization of a homogeneous distribution of metal nanoparticles is often difficult to control, and as a result, recent research has focused on control of the synthesis conditions to favor the formation of uniform composite films by a variety of deposition techniques, such as vacuum deposition^6,7 and coating from a mixed solution containing polymers and preformed metal nanoparticles.^8,9 However, it can be difficult to control all of the structural parameters of the composite films fabricated with these methods, especially when compared to the in situ synthesis of metal nanoparticles directly in ion-doped polymer precursors, which can be used to generate a variety of inorganic nanoparticles in polymer films in a highly controllable manner.^10–14 Although this in situ approach allows synthesis of several organic films containing uniformly-dispersed metal nanoparticles, the approach has been mostly applied to synthesize noble metal nanoparticles such as gold,^10,11 silver,^13,14 and platinum.¹² Herein, we report a design for the in situ synthesis of polymer thin films containing Cu and Ni nanoparticles using metal ion-doped precursor thin films on a glass substrate. In situ formation of these transition metal nanoparticles could be achieved for the first time using a highly cross-linked, ion-exchangeable polymer matrix and annealing of the ion-doped precursors in a hydrogen atmosphere at relatively high temperatures. This in situ bottom-up approach is highly tailorable because it allows for control of the amount of metal ions in the precursor films and thus the size of the resultant metal nanoparticles, which provides a structure-based means to control the physical properties of the composite thin films deposited on the substrates.

This process relies on the synthesis of ion exchangeable polymers with polymerizable moieties, deposition of the polymers on a glass substrate and curing through photoinduced cross-linking, and doping of metallic ions by ion exchange followed by reduction of the ions using hydrogen gas (Fig. 1). The structural characteristics required of the polymer matrices used for the synthesis of composite films in the present approach are (1) thermal stability for the reduction of doped ions at high temperature, (2) strong adhesion of the film to the substrate, and (3) ability to control the amount of metallic ions in the film. We have previously reported that Cu and Ni ions in a polyimide matrix can be reduced with hydrogen gas at ca. 250 °C (ref. 15) and 300 °C,¹⁶ respectively. We demonstrate herein that photoinduced cross-linking of the polymer matrix not only increases the thermal stability of the matrix and allows the reduction of doped transition metal ions to form metal nanoparticles without degradation of the matrix, but causes strong adhesion between the polymer matrix and glass substrate modified with photosensitive functional groups. In addition, varying the fraction of ion exchangeable groups (carboxylic acid groups) enables the amount of metal ion doping to be controlled; therefore, the generation of nuclei and growth of metal nanoparticles can be controlled, which determines the size of the nanoparticles.

Fig. 1 Schematic diagram of the synthesis of an ion-exchangeable polymer precursor and nanocomposite thin film with corresponding infrared spectra of the samples obtained at each step. (A) PMAA. (B) Modified PMAA obtained after treatment with CMS and CX. (C) Cross-linked, modified PMAA obtained after spin coating onto a glass substrate and UV irradiation. The decrease in the intensity of the peak assigned to C=C stretching vibration is indicated by an arrow. (D) Cross-linked polymer film after ion doping through ion exchange reaction. (E) Nanocomposite thin film obtained after annealing of ion-doped precursor in a hydrogen atmosphere at 330 °C for 10 min.

In a typical synthesis, poly(methacrylic acid) (PMAA) was first treated with 4-chloromethylstyrene (CMS) and α-chloro-p-xylene (CX) in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to introduce vinyl and methyl groups into the PMAA backbone (Fig. 1A and B). CMS was used to introduce photopolymerization capability into PMAA. Various amounts of CX were introduced to vary the amount of carboxylic acid groups, while maintaining a constant amount of CMS; lower amounts of CX induced greater ion exchange capabilities. Glass substrates modified with 3-aminopropyltriethoxysilane (APTMS) were treated with 4-vinyl-benzoyl chloride to form vinyl functional groups on the substrates. After spin coating a solution of modified PMAA in dimethylformamide (DMF) onto the glass substrate, the thin films (ca. 200 nm in thickness) obtained were irradiated with UV light to induce photopolymerization. Successful photopolymerization was achieved after UV irradiation for 3 h, as confirmed by attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy (Fig. 1C). A peak assigned to C=C stretching (1650 cm⁻¹) disappeared after UV irradiation (indicated by the arrow in Fig. 1C), and the films became insoluble in solvents, thus providing ion exchangeable, cross-linked polymer thin films anchored onto glass substrates.

To provide the polymer matrix with thermal stability sufficient for the high temperature hydrogen-induced reduction of metallic ions, at least 20% of the initial carboxylic acid groups must be bound with CMS (cross-linking sites). The highly cross-linked polymer matrix obtained after photoinduced cross-linking was stable under heat treatment, even at 350 °C (Fig. S1†). The ion exchange reaction to introduce Cu and Ni ions was achieved using aqueous copper chloride (CuCl₂) or nickel chloride (NiCl₂) solution after pretreatment of the films with sodium carbonate. The introduction of metallic ions is evidenced by the appearance of a C=O stretching vibration of carboxylate anions at 1540 cm⁻¹ (Fig. 1D). The use of sodium carbonate is critical because divalent metal cations are not easily exchanged into carboxylic acids in the highly cross-linked polymer matrix; therefore, sodium ions are first introduced into the films to form carboxylate anions, which can initiate ion exchange reactions.

The amount of doped metallic ions was determined by inductively coupled plasma (ICP) spectroscopy measurements and the results were compared with those for carboxylic acid groups contained in the polymer matrix. For polymer films with 50% carboxylic acid groups (the fraction of residual carboxylic acid groups based on the initial number of carboxylic acids in PMAA), the average amount of both Cu and Ni ions that can adsorb into 200 nm thick polymer films on a 1 × 1 cm² glass slide was determined to be ca. 20.5 nmol. This is half the amount of the initially adsorbed monovalent sodium ions and carboxylic acid groups estimated from the synthesis conditions (42.7 nmol), which indicates complete ion exchange of divalent Cu and Ni ions for all of the carboxylic acid groups. For polymer films with 40 and 30% carboxylic acid groups, approximately 16.5 and 13.0 nmol of metallic ions can adsorb into the films, respectively, which demonstrates systematic control of the amount of metallic ions adsorbed into the polymer matrix (Table 1).

Table 1 Amount of doped metallic ions and carboxylic acid groups in the precursor thin films

Sample	Fraction of residual COOH^a	Amount of COOH^b/nmol	Metal	Amount of ions^c/nmol	Mean particle size/nm
a The value is on the basis of initial amount of carboxylic acid groups in PMAA.b The value is estimated from the amount of sodium ions in the films.c The amount for 200 nm thick films on glass substrate (1 × 1 cm^2).
1	30%	25.0	Cu	13.0	22.6
			Ni	12.7	19.1
2	40%	32.2	Cu	16.5	12.3
			Ni	15.9	9.5
3	50%	42.7	Cu	20.5	6.1
			Ni	20.1	4.5

---

The effect of ion loading on the size of the Cu and Ni nanoparticles formed after annealing the ion-doped films in a hydrogen atmosphere at 250 and 330 °C was investigated for Cu and Ni ion-doped films, respectively. The microstructures of films deposited on thin carbon layers on copper grids were observed directly using transmission electron microscopy (TEM) (Fig. 2A–F). TEM analysis of the formed nanoparticles revealed mean sizes of 6.1–22.6 nm and 4.5–19.1 nm for the Cu and Ni nanoparticles, respectively, which was dependent on the initial amount of doped metallic ions; a lower amount of ions resulted in much larger nanoparticles (Table 1). To the best of our knowledge, this is the first successful synthesis of Cu and Ni nanoparticles with sizes over 20 nm that have been grown in situ in a polymer matrix. The large difference in particle size with changes in the initial ion concentrations is associated with the nanoparticle growth mechanism. Initial cluster formation occurs during annealing around 200 and 250 °C for Cu and Ni ions bound to carboxylate anions, respectively, as well as during the heating process at the beginning of annealing above these threshold temperatures. It has been reported that the size of metal nanoparticles formed in situ from metallic ions in a solid matrix is dependent on the initial number of nuclei formed at the early stage of reduction.¹⁷ Consistent with the literature, fewer nanoparticles nucleate at lower ion concentrations, which results in the formation of small numbers of larger nanoparticles (Fig. 2A and D). These nanoparticles also have greater interparticle separation distances, which prevents direct migration and coalescence within the highly cross-linked polymer matrix. When the ion concentration is higher, multiple cluster nucleation can occur with small interparticle separation distances. The nanoparticle formation process includes the protonation of carboxylate anions in addition to the reduction of metallic ions to form metal nanoparticles, which results in the formation of carboxylic acid groups that simultaneously undergo dehydration reactions to generate acid anhydrides at higher temperature. This is confirmed by the appearance of C=O stretching vibrations of the acid anhydride at 1800 and 1760 cm⁻¹ (Fig. 1E). These reactions lead to additional cross-linking of the polymer matrix, the degree of which is greater for films with larger amounts of carboxylic acid groups. The effect could also inhibit further nanoparticle growth and result in the formation of smaller nanoparticles at higher ion concentrations (Fig. 2C and F). Therefore, the growth of metal nanoparticles can be kinetically controlled, by which the final particle size is determined by the cluster concentration formed during the initial reduction stage and is finally limited by the number of total ion loadings. Moreover, since Cu and Ni are reported to be respectively non-reactive and reactive toward polar functional groups in the polymer backbone,¹⁸ the effect of surface passivation of the metal nanoparticles with the surrounding polymer matrix can play a role in the size evolution process and may result in smaller Ni nanoparticles than Cu nanoparticles.

Fig. 2 (A–F) TEM images of the nanocomposite films on carbon-coated TEM grids, obtained after annealing of ion-doped precursors in a hydrogen atmosphere at 250 °C for Cu (A–C) and 330 °C for Ni nanoparticles (D–F). The images in (A–F) are taken from samples with different ion concentrations, shown as samples (A–F) in Table 1. Scale bars: 100 nm. (G) Vis absorption spectra of Cu nanoparticle/polymer composite films. Spectra (a–c) correspond to those for the samples shown in images (A–C), respectively. (H) FMR signals of Ni nanoparticle/polymer composite films. The values of θ = 0° and θ = 90° were assigned to the external magnetic field in the directions parallel and perpendicular to the film surface, respectively. Signals (d–f) correspond to those for the samples shown in images (D–F), respectively.

The physical properties of the composite films were characterized using UV-vis and ferromagnetic resonance (FMR) spectroscopy for Cu and Ni nanoparticles, respectively, because the position and breadth of surface plasmon resonance (SPR) absorption and FMR absorption for Cu and Ni nanoparticles are important parameters to evaluate the quality of the composite films. The significant effect of the nanoparticle size on these properties is shown in Fig. 2G and F. As expected with increasing Cu nanoparticle size, the SPR maximum is red-shifted from 550 nm for the smaller (4.5 nm) nanoparticles to 600 nm for larger (22.6 nm) nanoparticles (Fig. 2G). A brilliant red-colored film was obtained for larger nanoparticles, and the nanoparticles were very stable after a month without distinct color change of the film, which indicates preservation of the embedded Cu nanoparticles from oxidation in the air, presumably due to effective surface passivation of the Cu nanoparticles with the matrix polymer and the highly cross-linked nature of the matrix. For Ni nanoparticles, the effect of particle size on the intensity of the FMR signals was evident, where larger sized particles resulted in larger FMR peaks (Fig. 2H). In addition, the FMR signal for the film with larger Ni nanoparticles was not dependent on the incident angle of the external magnetic field, because the larger interparticle distance between Ni nanoparticles, due to the lower amount of doped ions, resulting in a non-interacting composite film (Fig. 2H, spectrum d). In contrast, the FMR signals for the film with smaller nanoparticles show an angular dependence of the resonance field, i.e., around 2950 Oe for θ = 0°, and 3250 Oe for θ = 90° (Fig. 2H, spectrum f). This angular dependence may be caused by magnetic interaction between Ni nanoparticles with relatively smaller interparticle distances, due to a higher amount of doped ions, which is consistent with our previous work for Ni nanoparticles embedded in a polyimide matrix.¹⁹ These results demonstrate that the present in situ approach allows fabrication of metal polymer nanocomposite thin films on a substrate with the ability to control the composite microstructures and thus the properties of the films.

Photopolymerization of the polymer matrix also allows microscale patterning of the composite films (Fig. 3). Microscale patterns of composite films containing Cu and Ni nanoparticles were fabricated on glass substrates by irradiation with UV light through a photomask placed on the spin-coated films, and rinsing with chloroform followed by ion-doping and subsequent annealing treatment (Fig. 3A and B, respectively). Atomic force microscope (AFM) images confirms successful fabrication of composite film patterns with thickness of ca. 200 nm (inset in Fig. 3A and B). The pattern feature size could be controlled from 2 μm (Fig. 3C and D) up to many centimeters, depending on the resolution of the photomask used for UV photopatterning, which would be useful for the fabrication of optical and magnetic devices.

Fig. 3 Optical microscope images for micropatterns of composite films containing Cu (A and C) and Ni nanoparticles (B and D) on glass substrates. Insets in (A and B) are atomic force microscopy (AFM) topographic images. Insets in (C and D) are TEM images of the composite patterns. Scale bars: 20 μm (A and B) and 5 μm (C and D).

In conclusion, we have reported a method that uses ion-doped, highly cross-linked, synthetic polymer thin films as precursors for the synthesis of metal/polymer nanocomposite films on a substrate. Photopolymerizable units are introduced as cross-linking sites into the PMAA backbone and substrate surface, so that the reduction of Cu and Ni ions can be achieved with hydrogen gas to yield homogeneous nanocomposite thin films containing metal nanoparticles. In addition, control of the amount of ion exchange groups in the polymer matrix enables the size of the embedded metal nanoparticles to be determined, whereby the SPR and FMR responses of the Cu and Ni nanoparticles, respectively, can be easily tailored. This photopolymerization technique has the potential for use to facilitate the fabrication of optical and magnetic composite nanodevices.

Notes and references

1. K. Y. Chun, Y. Oh, J. Rho, J. H. Ahn, Y. J. Kim, H. R. Choi and S. Baik, Nat. Nanotechnol., 2010, 5, 853–857.
2. G. V. Ramesh, S. Porel and T. P. Radhakrishman, Chem. Soc. Rev., 2009, 38, 2646–2656.
3. S. Maity, J. R. Bochinski and L. I. Clarke, Adv. Funct. Mater., 2012, 22, 5259–5270.
4. C. Vijayakumar, B. Balan, A. Saeki, T. Tsuda, S. Kuwabata and S. Seki, J. Phys. Chem. C, 2012, 116, 17343–17350.
5. A. Devadoss, A. M. Spehar-Déleze, D. A. Tanner, P. Bertroncello, R. Marthi, T. E. Keyes and R. J. Forster, Langmuir, 2010, 26, 2130–2135.
6. M. Elbahri, M. K. Hedayati, V. S. K. Chakravadhanula, M. Jamali, T. Strunkus, V. Zaporojtchenko and F. Faupel, Adv. Mater., 2011, 23, 1993–1997.
7. H. Takele, H. Greve, C. Pochstein and V. Zaporojtchenko, Nanotechnology, 2006, 17, 3499–3505.
8. D. Y. Choi, H. W. Kang, H. J. Sung and S. S. Kim, Nanoscale, 2013, 5, 977–983.
9. R. Mukherjee, S. Das, A. Das, S. K. Sharma, A. K. Raychaudhuri and A. Sharma, ACS Nano, 2010, 4, 3709–3724.
10. K. R. Berry Jr, A. G. Russell, P. A. Blake and D. Keith Roper, Nanotechnology, 2012, 23, 375703.
11. J. Dhar and S. Patil, ACS Appl. Mater. Interface, 2012, 4, 1803–1812.
12. S. W. Lee, J. R. Kumpfer, P. A. Lin, G. Li, X. P. A. Gao, S. J. Rowan and R. M. Sankaran, Macromolecules, 2012, 45, 8201–8210.
13. N. H. Mack, J. A. Bailey, S. K. Doorn, C. A. Chen, H. M. Gau, P. Xu, D. J. Williams, E. A. Akhadov and H. L. Wang, Langmuir, 2011, 27, 4979–4985.
14. Y. Mao, H. P. Xu, H. Zhao, W. Z. Yuan, A. Qin, Y. Yu, M. Faisal, Z. Xiao, J. Z. Sun and B. Z. Tang, J. Mater. Chem., 2011, 21, 13627–13633.
15. K. Akamatsu, H. Shinkai, S. Ikeda, S. Adachi, H. Nawafune and S. Tomita, J. Am. Chem. Soc., 2005, 127, 7980–7981.
16. K. Akamatsu, M. Fujii, T. Tsuruoka, S. Nakano, T. Murashima and H. Nawafune, J. Phys. Chem. C, 2012, 116, 17947–17954.
17. A. Simo, J. Polte, N. Pfänder, U. Vainio, F. Emmerling and K. Rademann, J. Am. Chem. Soc., 2012, 134, 18824–18833.
18. F. Faupel, R. Willecke and A. Thran, Mater. Sci. Eng., R, 1998, 22, 1–55.
19. S. Tomita, P. E. Jonsson, K. Akamatsu, H. Nawafune and H. Takayama, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 174432, 1–6.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra46166a

---