Home
Search
Collections
Journals
About
Contact us
My IOPscience
Optical polymers with tunable refractive index for nanoimprint technologies
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 505301 (http://iopscience.iop.org/0957-4484/25/50/505301) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 153.96.116.4 This content was downloaded on 26/11/2014 at 21:16
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 505301 (10pp)
doi:10.1088/0957-4484/25/50/505301
Optical polymers with tunable refractive index for nanoimprint technologies J Landwehr1, R Fader1,3, M Rumler1,3, M Rommel1, A J Bauer1,3, L Frey1,2,3, B Simon4,5, B Fodor4,5, P Petrik4, A Schiener6, B Winter7 and E Spiecker7 1
Fraunhofer Institute for Integrated Systems and Device Technology (IISB), Erlangen, 91058, Germany Chair of Electron Devices, University of Erlangen-Nuremberg, Erlangen, 91058, Germany 3 Erlangen Graduate School in Advanced Optical Technologies (SAOT), Erlangen, 91052, Germany 4 Research Institute for Technical Physics and Materials Science, Budapest, 1121, Hungary 5 Faculty of Science, University of Pécs, Pécs, 7624, Hungary 6 Chair of Crystallography and Structural Physics, University of Erlangen-Nuremberg, Erlangen, 91058, Germany 7 Center for Nanoanalysis and Electron Microscopy (CENEM), University of Erlangen-Nuremberg, Erlangen, 91058, Germany 2
E-mail:
[email protected] Received 13 August 2014, revised 24 September 2014 Accepted for publication 3 October 2014 Published 26 November 2014 Abstract
In order to realize a versatile high throughput production of micro-optical elements, UV-curable polymer composites containing titanium dioxide nanoparticles were prepared and characterized. The composites are based on an industrial prototype epoxy polymer. Titanium dioxide nanoparticles smaller than 10 nm were synthesized by the nonaqueous sol method and in situ sterically stabilized by three different organic surfactants. The composites exhibit high transparency. Distinct alteration of optical transmission properties for visible light and near IR wavelength range could be avoided by adaption of the stabilizing organic surfactant. Most importantly, the refractive index (RI) of the composites that depends on the fraction of incorporated inorganic nanoparticles could be directly tuned. E.g. the RI at a wavelength of 635 nm of a composite containing 23 wt% titanium dioxide nanoparticles is increased to 1.626, with respect to a value of 1.542 for the pure polymer. Furthermore, it could be demonstrated that the prepared inorganic–organic nanocomposites are well suited for the direct fabrication of lowcost micro-optical elements by nanoimprint lithography. A low response of the optical composite properties to temperature treatment up to 220 °C with a shrinkage of only about 4% ensures its application for integrated micro-optical elements in industrial production. S Online supplementary data available from stacks.iop.org/NANO/25/505301/mmedia Keywords: substrate conformal imprint lithography, composite materials, UV polymers, TiO2 nanoparticles, refractive index matching (Some figures may appear in colour only in the online journal) filters or amplitude modifiers, featuring structures with dimensions in the wavelength range as well as micro-optical elements (such as waveguides and diffractive lenses) are crucial for the upcoming technology of electrical-optical circuit boards [6] and lab-on-chip devices [7–9]. All the above mentioned devices and elements can be realized by organic or inorganic materials. In recent years, huge effort has been put in the development of organic and inorganic high refractive
1. Introduction Structures with dimensions in the wavelength range [1] of light not only enable the optical design of anti-reflective coatings [2], which improve the efficiency of light emitting diodes [3] or solar cells [4] but can also be considered as base elements in the emerging field of plasmonics [5]. In addition, passive photonic devices such as ring resonators, wavelength 0957-4484/14/505301+10$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
Nanotechnology 25 (2014) 505301
J Landwehr et al
index (RI) materials [10], as they are highly promising for the production of new photonic devices and optical elements. Inorganic materials offer outstanding high RIs (n = 2–5) but lack mechanical flexibility. They usually exhibit high absorption coefficients in the visible range [11] and require expensive multistep processing [12]. A possibility to lower production costs of optical elements based on inorganic highRI materials is the application of imprint lithography. This structuring method is highly desirable because it offers very accurate single-step fabrication of three-dimensional structures, and in parallel, enables high throughput fabrication of photonic devices over large areas [13, 14] at low cost. In addition, it is possible to build up stacked multilayer structures by sequential patterning using alignment technology [15, 16]. Although a lot of progress was achieved on singlestep patterning of inorganic sol-gels using imprint lithography [17], e.g. for titanium dioxide, the high material shrinkage [18, 19] upon temperature exposure (T > 200 °C) has to be considered a drawback for industrial production of integrated devices where reflow soldering at approximately 220 °C is essential. Organic polymers exhibit a relatively low RI (n < 2), but in comparison to inorganic materials, they have the advantage of low weight and cost, simple processing, and high mechanical impact resistance. As organic materials, e.g. polymers, in general exhibit only small shrinkage upon curing, the combination of imprint lithography with high RI organic materials promises a strong technological impact for the production of photonic devices and micro-optical elements. The imprint technology substrate conformal imprint lithography (SCIL) is especially suited for large-area patterning. It provides micro- and nano-sized structuring of inorganic and organic materials in one single step, down to a resolution of 10 nm over an area of up to 180 cm2 [20]. Moreover, the ability to adjust the RI is a powerful feature of organic materials [10], with photonic devices being highly sensitive to RI steps as small as 0.01. The chemical monomer modification of organic polymers provides stepwise tunable high RI polymers [21, 22]. A similar stepwise RI tuning of inorganic materials (e.g. silicon nitride), on the other hand, could only be realized by specialized chemical vapor deposition processing [23]. A continuously tunable RI is unique to inorganic–organic nanocomposite materials that are based on organic polymers. Such continuously tunable materials are highly desirable for the production of optical elements. However, the introduction of high RI inorganic building blocks into organic polymers has to fulfil the requirements of high transparency while avoiding phase separation and the inhibition of the polymer curing reaction. This work presents the detailed fabrication and characterization of UV-curable inorganic–organic nanocomposites, including the essential step of particle synthesis. The composites meet above requirements, especially a tunable RI, by incorporation of titanium dioxide (TiO2) nanoparticles in an organic matrix. Moreover, such composites could be directly structured using SCIL. The polymer matrix is based on KATIOBOND OM VE 110707® by DELO Industrial Adhesives, as it has been proven to be perfectly suited for the
SCIL process [24]. TiO2 was selected as low-cost, nontoxic inorganic component. The anatase crystal structure, which is the thermodynamically favoured crystal structure for nanoparticles fabricated by the applied low reaction temperature synthesis [25], exhibits a high RI of 2.56 at a wavelength of 589 nm [26]. Although other groups report on the use of commercially available TiO2 nanoparticles in combination with different matrix polymers [27], a proper particle synthesis and composite preparation had to be found to obtain homogeneous composites based on the high performance KATIOBOND polymer. As TiO2 is widely used in many applications, a wide range of synthetic approaches has evolved for the fabrication of TiO2 nanomaterials [28]. The nonaqueous sol synthesis based on organic solvents was chosen over an aqueous pathway because in a subsequent composite preparation, solvent residues from the synthesis must not inhibit the polymer curing reaction. Furthermore, the synthesis provides easy access to in situ ligand-stabilized nanoparticles with a narrow size distribution and high crystallinity at low reaction temperatures [25]. This allows a detailed study on the influence of different organic particle stabilizers on the resulting composite quality in terms of optical and thermal characteristics.
2. Experimental The synthesis of stabilized titanium dioxide (TiO2) nanoparticles was conducted according to a published method [29] using titanium tetrachloride as a titanium oxide precursor to react with the organic solvent benzyl alcohol at low temperatures. The organic surfactants 4-tert-butylcatechol (TBC), trimethoxy(7-octen-1-yl)silane (TMOS) or trioctylphosphine oxide (TOPO) were dissolved in benzyl alcohol before titanium tetrachloride was added dropwise to the solution. After a short aging time, ranging from 2 h to 1 day at temperatures below 100 °C, the stabilized particles were extracted from the reaction mixture and directly dispersed in the organic solvent tetrahydrofuran (THF). More details about the particle synthesis can be found in the supporting information. The light absorption of THF particle solutions was measured by UV–Vis spectroscopy at a weight concentration of 250 μg ml−1. The measurement was performed with the instrument UV-1602 (Shimadzu, Kyōto, KYT, Japan) over a spectral range from 190 nm to 1000 nm. The solvent contained additional stabilizer (1 mM). UV-transparent silica glass cuvettes were used. The size of the stabilized nanoparticles was determined by dynamic light scattering. Measurements were carried out using the instrument Delsa™Nano HC (Beckman Coulter, Brea, CA, USA), which uses two laser diodes with a wavelength of λ = 658 nm with each 30 mW output power at a fixed scattering angle of 165.7° and a correlator with 440 channels. The provided data is based on the arithmetic mean value of the result of three independent measurements, which were each based on 70 repetitive measurements. Data analysis was conducted by the respective software according to ISO 22412. The electric field autocorrelation functions were analyzed using the Levenberg– 2
Nanotechnology 25 (2014) 505301
J Landwehr et al
tip radius of curvature of less than 8 nm. Measurements were conducted in tapping mode covering an area of 50 × 50 μm2 at two different sample sites, close to the centre of the sample. In addition, the composite layers were analyzed by high-resolution transmission electron microscopy (HRTEM) and highangle annular dark-field (HAADF) scanning transmission electron microscopy (STEM). Rectangular, cross-sectional samples of the composite layers coated on a Si-wafer substrate were prepared by grinding, polishing and subsequent Ar+-ion milling. HRTEM and HAADF STEM imaging were carried out using a FEI Titan3 80–300 microscope operated at 300 kV. The composite samples were cooled using an LN2 sample holder (Gatan, Inc.) to minimize beam damage to the sample in STEM mode. To image the crystalline TiO2 nanoparticles without a polymer matrix HRTEM imaging was performed. Therefore, the particles dissolved in THF were directly dispersed onto TEM sample copper grids with an ultrathin carbon film (less than 3 nm thick) supported by a lacy carbon film (Ted Pella, Inc.). Furthermore, thin layers of composites were optically characterized by transmission measurements. Samples were prepared on quartz substrates by spin–coating. Sample size was 2 × 2 cm2. Solvent–free composites were used to coat 3 μm thick composite layers that contained 10 wt% titanium dioxide nanoparticles. The transmission was measured in transmitted light/bright-field mode using a microscope equipped with a spectrometer (ZEISS, Jena, Germany) with a resolution of 0.8 nm for an effective wavelength range from 400 to 850 nm. With the measuring setup exhibiting wavelength-dependent oscillations in the transmitted intensity, the data was smoothed by the Savitzky–Golay method [33]. Temperature exposure was realized by using a hotplate at ambient air conditions. For the presented data samples were heated to 150 °C at a heating rate of 2 K s−1, followed by a 50 s holding time, subsequently heated to 220 °C at 1.5 K s−1 followed by a 50 s holding time and a cooling rate of 6 K s−1 to room temperature. The final structuring of the prepared inorganic–organic composites was realized by UV-SCIL. The imprinting was performed in a clean room with a MA/BA8 Gen3 mask aligner (SUSS MicroTech AG, Garching, Germany), which was equipped with the respective SCIL imprint tooling. Imprinting was done on 100 mm silicon wafers. Composites were prepared solvent free and contained 10 wt% stabilized nanoparticles. The polymer and composite layers were spin– coated on the silicon substrate and subsequently imprinted using a multilayer polydimethylsiloxane (PDMS) stamp [20]. UV-light induced curing was conducted at 365 nm. Curing time was 2 min.
Marquardt method. Samples were measured at a particle concentration of 27.5 g l−1 in THF at 20 °C. In the case of TBC stabilized particles, the solvent contained additional stabilizer (10 mM). Particle crystallinity was evaluated by powder X-ray diffraction, while diffraction patterns were measured using the instrument X’Pert Powder (PANalytical, Almelo, Netherlands) equipped with an X’Celerator detector. Measurements were carried out using copper K-alpha radiation (0.15418 nm) at 40 kV/35 mA in a Bragg–Brentano theta–theta configuration on a corundum single crystal. Particles were dried in an oven overnight at 100 °C. The crystallite sizes were determined by using the Scherrer equation, which was applied with a form factor of 1. Spectroscopic ellipsometry was applied to determine the RI of thin composite layers. Samples were prepared on silicon substrates by spin–coating. Sample size was 2 × 2 cm2. For thin layers with a thickness from 100 to 300 nm, the composite was diluted by THF to a solvent-to-polymer ratio ranging from 8 to 10. Thick micrometre range layers were coated using solvent-free composites, which were prepared by removal of the solvent under vacuum after blending the particle solutions with the polymer. The ellipsometric measurements on which the presented data on the RI are based were performed using a SOPRA SE-5 spectroscopic ellipsometer (Semilab, Budapest, Hungary) in a spectral range of 190 nm–1700 nm at an incident angle of 75.15°. The analyzer was set to an angle between 20° and 45° in mirror analyzer mode. The evaluation of measurement data was conducted by using the program ‘Film Wizard’ from ‘Scientific Computing International’. The layer model for data fitting was based on the assumption of a single top layer treating the composite layers as a homogeneous material consisting of isolated Lorentz oscillators [30]. The number of oscillators was set to 1. Fitting range was 300–1700 nm. The influence of varying the layer model is described elsewhere for a similar composite system [31]. In this case, the fit quality was not significantly improved when including a surface roughness or an absorption in the model, which reveals high composite layer quality. Depolarization was measured directly by rotating compensator ellipsometry. Measurements were performed with a rotating compensator spectroscopic ellipsometer M-2000DI (J A Woollam Co., Lincoln, NE, USA) at angles of incidence of 70°. Sample mapping, in order to determine the thickness uniformity, was performed by the M-2000DI, as well. Composite layer dielectric function was then described with the Cauchy dispersion formula by fitting only the parameters A and B, parameter C was set to 0 [31]. Fitting range was 420–1690 nm. Refractive indices determined by the instrument SOPRA SE-5 and the respective SCI modelling, agreed with the results by using the instrument M-2000DI and applying the respective Cauchy evaluation. The composite layers prepared for RI characterization were analyzed by atomic force microscopy (AFM), as well. Measurements were performed by a Dimension Icon with Nanoscope V Controller (BRUKER, Bellerica, USA) equipped with a closed-loop XYZ tip-scanner, using NCH tips (Nanoworld, Neuchâtel, Switzerland), which exhibit a typical
3. Results and discussions 3.1. Synthesis of stabilized titanium dioxide nanoparticles
For the realization of a homogeneous inorganic–organic nanocomposite with high transparency, the inorganic component TiO2 is introduced as stabilized nanoparticles into the 3
Nanotechnology 25 (2014) 505301
J Landwehr et al
incorporating polymer matrix. The particles are in situ stabilized during the synthesis reaction by organic surfactants, enabling solubility in an organic solvent. For the needs of the composite preparation, the solvent has to be compatible with the incorporating polymer as well. Foremost, a steric stabilization is crucial to avoid phase separation and hence agglomeration in the targeted composite. At the same time, the steric surfactants must not affect the polymer curing reaction or induce light absorption in interaction with the semiconducting TiO2. Besides, the fresh synthesis of stabilized particles was essential in order to obtain homogeneous composites. The three different organic surfactants 4-tertButylcatechol (TBC), Trimethoxy(7-octen-1-yl)silane (TMOS) or Trioctylphosphine oxide (TOPO) were effectively tested to promote the stabilization in the polymer matrix while making the particles soluble in the organic solvent tetrahydrofuran (THF).
Figure 1. Normalized UV–Vis measurement of TiO2 nanoparticles stabilized by TMOS, TOPO or TBC. Table 1. Reaction parameters and respective hydrodynamic particle
3.1.1. Particle-solvent interaction. The freshly prepared and
radius of titanium dioxide particles synthesized by the benzyl alcohol route using different organic surfactants for in situ stabilization.
still wet particles had to be directly dissolved after synthesis in order to avoid particle agglomeration during the drying process. The reaction solvent benzyl alcohol and the organic surfactants both promote charge-balancing reactions on the particle surfaces. This leads to low surface charging of the particles, which favours the agglomeration of dried particles. The stability of the particles in the organic solvent, however, was high if the organic stabilizer saturated the particle surfaces. Adding a small amount of the respective organic surfactant to the solvent (1–10 mM) ensured long-term stability of the solutions over weeks, in the case of TMOSand TOPO-stabilized particles, and months for TBC stabilized particles. The resulting particle solutions in THF formed without sonication treatment, and the particles could not be separated mechanically from the stable solutions by ultracentrifugation up to 20 000 rpm. Although all three particle types yielded transparent solutions, the colouring differed. Particles stabilized by TBC formed deep red solutions, whereas particles stabilized by TOPO or TMOS yielded colourless solutions. The optical light absorption characteristics of the particle solutions were determined by UV–Vis spectroscopy. The respective normalized absorption measurements are shown in figure 1. The solutions containing TBC stabilized particles exhibited an absorption shoulder ranging from 300 nm to 600 nm, next to the intense absorption of TiO2 for wavelengths smaller than 300 nm (figure 1). The light absorption of TBC-stabilized particles between 300 nm and 600 nm induces the visible complementary colour, which could be observed in the red colouring of the particle solution. The absorption intensity, and thus the colouring of the particle solution, was stronger with higher concentrations of the stabilizer TBC. According to Niederberger et al [29] the absorption is caused by the surfactant-to-metal charge–transfer interaction between the organic surfactant TBC and the surface metal atoms of the particles. By introducing the non-aromatic surfactants TMOS and TOPO as in situ stabilizing agents the charge–transfer induced absorption could be avoided. The respective
Organic surfactant
Reaction parameters
TBC TOPO
1 day @ 70 °C 2 h @ 100 °C 2 h @ 80 °C 2 h @ 100 °C
TMOS
Hydrodynamic diameter [nm] 7.2 ± 1.4 13 ± 2.6 1.8 ± 0.3 1.4 ± 0.1 8.5 ± 1.5
solutions of stabilized particles yielded colourless solutions and exhibited no significant light absorption from 380 nm to 1000 nm (figure 1). This targeted stabilizer design ensures the application of the desired composite polymer for the fabrication of optical elements for a broad wavelength range. 3.1.2. Particle size distribution and crystallinity. In general,
TBC-stabilized particles exhibited a high degree of monodispersity, with the particle size showing low sensitivity to reaction temperature and Ti-to-surfactant ratio. Dynamic light scattering (DLS) analysis was used to determine the average particle size in THF solution as 7.2 ± 1.4 nm (table 1). Similar high monodispersity could not be achieved for TMOS and TOPO stabilized particles. The synthesis of TMOS and TOPO stabilized particles yielded particle solutions that contained a high number fraction of small particles only about 2 nm in size, but, in addition, a certain fraction of larger particles. The observed polydispersity could be explained by the short aging time of only 2 h, compared to the 1 day aging time of TBC-stabilized particle synthesis. A prolonged aging time of TMOS- and TOPO-stabilized particles in order to increase their monodispersity actually decreased the particle solubility. In case of TMOS-stabilized particles, DLS measurements in THF determined 1.4 ± 0.1 nm small particles to coexist with larger particles of 8.5 ± 1.5 nm in size. It should be noted, 4
Nanotechnology 25 (2014) 505301
J Landwehr et al
size, in case of TOPO A particles smaller than 2 nm, represents the lower limit for Bragg peak generation, even in case of high crystallinity. Attached in the supporting information an HRTEM image of TOPO A particles (figure S4) is given, which reveals the actually high crystallinity of the particles. The full width at half maximum (FWHM) analysis of all available peak reflexes by applying the Scherrer equation yielded an average crystallite size of 6.4 ± 1.9 nm for TMOS particles, 6.9 ± 1.3 nm for TOPO B particles and 4.3 ± 0.6 nm in case of TBC particles. Thus, crystallite sizes from XRD roughly match the particle sizes in solution analyzed by DLS, if the thickness of the stabilizing organic particle shell is subtracted, and the particle sizes observed by HRTEM. The good agreement indicates that the particles mainly consist of a single crystalline domain. A spherical particle shape, which is implied by DLS measurements, could be confirmed by TEM and XRD measurements with FWHM deviations for peak reflexes at different theta exhibiting only minor deviations.
Figure 2. X-ray powder diffractogram of TiO2 nanoparticles stabilized by different organic surfactants. The indicated diffraction peaks refer to anatase lattice planes. Baseline correction was applied to the data.
3.2. Preparation and characterization of inorganic–organic nanocomposites
however, that each defined particle size exhibited a narrow size distribution with a standard deviation of 18% maximum (table 1). In case of TOPO-stabilized particles, different particle sizes of either 1.8 ± 0.3 nm or 13 ± 2.6 nm could be favoured by adjusting the reaction temperature to either 80 °C or 100 °C. The different TOPO-stabilized particle types are denoted in the following as TOPO A (1.8 ± 0.3 nm) and TOPO B (13.0 ± 2.6 nm) particles. Particles stabilized by TMOS or TBC are referred to as TMOS and TBC particles. The simple nonaqueous sol synthesis based on benzyl alcohol not only provided in situ stabilized nanoparticles that exhibited small sizes and a narrow size distribution, but also showed crystallinity at low reaction temperatures (⩽100°C). X-ray powder diffraction (XRD) revealed a high crystallinity of the synthesized TiO2 particles, with regard to the small particle size (figure 2). This finding is supported by highresolution transmission electron microscopy (HRTEM) measurements of particle solutions, attached in the supporting information (figures S1–S3). The most distinct crystallinity was determined for TOPO B and TMOS particles. Rietveld analysis determined the TiO2 crystal structure as pure anatase. The characteristic anatase (101) and (200) lattice plane peaks could be identified for TBC particles as well, which exhibited less crystallinity in total. In case of TOPO A particles, no distinct crystallinity can be ascertained in the XRD measurement, but two weak peaks referring to the (101) lattice plane of the anatase phase at 2Θ = 25° and a minor peak at 2Θ = 67°, which cannot be assigned to anatase titania, is present. The HRTEM image in the supporting information (figure S4) of TOPO A particles reveals the actually high crystallinity of the particles. It supports the lack of a clear crystal structure showing indications for the presence of both rutile and anatase atom arrangements. The observed differences or lack of crystallinity measured by XRD could be ascribed to lattice plane stacking errors or other crystalline defects that are also visible in the HRTEM image (figure S4). The small particle
For the preparation of an inorganic–organic composite polymer, the ex situ synthesis method [10] is applied. THF solutions of synthesized TMOS, TOPO A and TBC particles (i.e., TiO2 nanoparticles stabilized by the organic surfactants TMOS, TOPO and TBC) were blended with the matrix polymer KATIOBOND OM VE 110707 followed by the removal of the volatile solvent. The solvent-free composite can be re-diluted with THF for needs of further processing such as spin–coating. In case of TOPO-stabilized particles, the particle type TOPO A was favoured over TOPO B due to the smaller mean particle size. The epoxy polymer KATIOBOND OM VE 110707 mainly consists of three different monomers, namely 3-ethyl-3-[(2-ethylhexyloxy)methyl]oxetane, Bisphenol-A-epichlorohydrin and 7-oxabicyclo[4.1.0] hept-3-ylmethyl 7-oxabicyclo[4.1.0]heptane-3-carboxylate. For complete composition and structural information, see the supporting information, table S1. Composites containing TOPO A particles were prepared with particle weight fractions of up to 10 wt%, whereas TMOS particle composites contained up to 20 wt% and TBC particle composites even up to 23 wt%. 3.2.1. Optical transmission characterization. The stabilized
particles formed transparent composites with the incorporating epoxy matrix. No inhibition of the polymer UV curing reaction at 365 nm is observed for any of the three organic surfactants, in case of TBC particle composites even up to particle weight fractions of 23 wt%. A minor increase in curing time is induced by the band-gap absorption of the semiconducting titanium dioxide. The optical transmission property of the different composites containing 10 wt% TMOS, TOPO A or TBC particles was determined by optical spectroscopy (figure 3(a)). The data represents the relative light transmission of cured composite layers (3 μm) in reference to the pure 5
Nanotechnology 25 (2014) 505301
J Landwehr et al
chosen to simulate a typical industrial reflow solder process with a maximum peak temperature of 220 °C for 50 s, including heating and cooling ramps [34]. The data in figure 3(b) represents the relative light transmission of the composite layers after heat exposure in reference to the light transmission of the untreated composites. Respective data of the pure polymer is included in the graph. The heat exposure did not alter the light transmission characteristic of the pure polymer at all. A minor decrease in the relative transmission after heat exposure at wavelengths smaller than 600 nm could be observed for composites containing TMOS or TBC particles, with a maximum of 7% increase of absorbance at 400 nm for TBC particle composites. No distinct changes, however, were found for the composite containing TOPO A particles. Thus, all three particle types exerted little influence on the characteristic of the matrix polymer. In addition, in case of prolonged temperature exposure to 220 °C (>3 min) or temperatures as high as 300 °C an equivalent altering of the transmission characteristics at short wavelengths was observed for the pure polymer as well (not shown here). The altering in transmission of all three composites for respective prolonged or high-temperature treatments was even more pronounced. Thus, the stabilized anatase TiO2 nanoparticles catalyse the temperature-induced degradation of the epoxy matrix polymer. Nevertheless, at standard reflow conditions (50 s, 220 °C) with a decrease in transmission of 7% maximum at 400 nm the effect of temperature-induced degradation is weak in general and in particular for the case of TMOS or TOPO A particles.
Figure 3. Transmission property of 3 μm thick layers of inorganic–
organic composites and the pure incorporating matrix polymer. (a) Relative transmission of composites containing 10 wt% TBC, TMOS or TOPO A particles, i.e. TiO2 stabilized by TBC, TMOS or TOPO, in reference to the transmission of the pure matrix polymer. (b) Relative transmission of the composites and the pure matrix polymer after heat exposure to 220 °C for 50 s in reference to nontreated layers.
3.2.2. Refractive index measurement. In order to determine
the impact of the inorganic nanoparticle fraction on the RI of the polymer, cured composite layers were measured by spectroscopic ellipsometry [32]. Depolarization was observed to the same extent for the pure polymer and the inorganic– organic composite layers. The inorganic particle fraction was not associated with the degree and existence of depolarization, but distinct depolarization occurred for wavelengths, and multiples higher in energy, that matched the layer thickness (figure 4(a)). Although the present study cannot provide adequate clarification on the correlation of depolarization on the layer thickness, which remains a focus of future studies, the occurrence of depolarization peaks were avoided for the relevant wavelength range (i.e. 420 nm– 1690 nm) by targeting a layer thickness from 100 nm to 300 nm (figure 4(b)). Thus, composite layer thicknesses of a few hundred nanometres guaranteed the most reliable measurement results and were applied for the characterization of the impact of the inorganic nanoparticle fraction on the RI. The increase of the composite RI with the particle-weight fraction is exemplarily illustrated in figure 5 for the composite containing TBC stabilized TiO2 nanoparticles, covering a wavelength range from 350 nm to 1700 nm. The measurement refers to re-evaluated published data [35]. In reference to the RI of the pure polymer, the RI of the composite is consecutively increased over the whole wavelength range
polymer. Whereas the incorporation of TMOS and TOPO A particles induced only a negligible decrease in light transmission in the polymer at 400 nm, a distinct decrease ranging from 400–600 nm could be observed for TBC particles. In general, the relative light transmission characteristics of cured composites did not differ from the light transmission characteristics of the respective THF particle solutions (figure 1). Thus, the interaction of the stabilized particles with the incorporating polymer did not result in additional light absorption effects. As a consequence, an incorporation of TOPO A or TMOS particles by the epoxy polymer KATIOBOND OM VE 110707 does not alter the high transparency of the matrix polymer in the visible light and near IR range, whereas composites containing TBC particles are suited for applications at wavelengths larger than 600 nm. For an application as integrated optical elements within an industrial production, a composite polymer has to withstand high-temperature exposure. The influence of heat exposure on the light transmission characteristic of the composites is shown in figure 3(b). The heat exposure was 6
Nanotechnology 25 (2014) 505301
J Landwehr et al
Figure 4. Depolarization and the mean square difference (MSE–
Figure 6. (a) RI at a wavelength of 635 nm of KATIOBOND OM
mean square error) between the measured and calculated Ψ and Δ values summed for all wavelengths and averaged, in order to determine the RI of thin layers, in dependence of the lower end of the wavelength range used for the fit, λcut. The data refers to (a) a cured 660 nm thick pure polymer layer of KATIOBOND OM VE 110707 and (b) an inorganic–organic composite layer with a thickness of 206 nm containing 4 wt% TOPO A particles.
VE 110707 composite containing TMOS, TOPO or TBC stabilized TiO2 nanoparticles in weight fractions ranging from 0% (pure organic resist) to 23%. Layer thickness ranges from 100–300 nm. (b) Respective slope that represents the increase of the RI at the wavelength of 635 nm in dependence on the particle weight fraction.
with increasing particle-weight fraction. A weight fraction of 23 wt% TBC particles increased the RI at a wavelength of 635 nm from initially 1.542 to 1.626. The RI at 635 nm of all three different KATIOBOND OM VE 110707 composites containing TMOS, TOPO A or TBC particles with weight fractions from 3% to 23% is shown in figure 6(a). It must be stated that depolarization due to different layer thicknesses of the measured composite samples induced different absolute RI values of the respectively thick prepared pure polymer layers, as well. Thus, the RI of the pure polymer (0 wt% particle fraction) differs in the graph. However, the standard deviation of the sample thickness within a test series was below 27%, with reference to sample thicknesses of 122 ± 22 nm (TBC), 237 ± 26 nm (TMOS) and 235 ± 23 nm (TOPO). In a first approximation, a linear RI increase is induced by any of the stabilized nanoparticles. For reasons of comparability, the respective slopes that represent the relative increase of th`e RI at a wavelength of 635 nm in dependence of the particle weight fraction are summarized in figure 6(b). Considering the standard deviation error, the slopes for TMOS (2.80*10−3 ± 4.2*10−4 wt %−1), TOPO A
Figure 5. RI of KATIOBOND OM VE 110707 composite contain-
ing TBC stabilized TiO2 nanoparticles in weight fractions ranging from 0% (pure organic resist) to 23% with sample thickness of about 120 nm measured by spectroscopic ellipsometry. 7
Nanotechnology 25 (2014) 505301
J Landwehr et al
(3.12*10−3 ± 3.0*10−4 wt%−1) and TBC (3.57*10−3 ± 3.6*−4 wt%−1) particles highly match. Thus, the influence of the organic stabilizing surfactants on the RI is negligible as expected, as the total particle volume fraction of the investigated composites is low. The particle volume percent vol%(x) correlates with the actual weight percent wt%(x) according to ⎡ vol%(x) = 100 ⎣ 100 wt%(x)−1 − 1
(
) ( ρ ρ ) + 1⎤⎦ x y
−1
average diameter of approximately 8 nm together with particles with an average diameter of approximately 1–2 nm exist isolated in the polymer matrix (figure 7(b)). The stabilization of particles as small as 1–2 nm by the incorporating matrix polymer endorses the good interaction of the organic particle stabilizer with the polymer matrix, which favours homogeneity and thus the optical quality of the composite. Moreover, the size distribution of TMOS stabilized nanoparticles within the composite match the distribution of freshly prepared particles in the THF solution used for the composite preparation. DLS measurements determined 1.4 nm small particles to coexist with 8.5 nm big particles in average (table 1). The high agreement of the particle sizes indicates a proper DLS data evaluation and an optimum interaction between TMOS stabilized TiO2 particles and the matrix epoxy polymer. Find attached in the supporting information HRTEM images of TOPO A (figure S4), TMOS (figure S7) and TBC particles (figure S8) incorporated in the polymer matrix.
−1
, (1)
where ρx is the particle density and ρy the density of the incorporating polymer. Assuming the particle density to be ρx = 4.32 g cm−3, equal to pure anatase TiO2 particles by neglecting the stabilizing organic surfactants, and a measured polymer density of ρy = 1 g cm−3, the maximum particle weight fraction of 23 wt% equates to a volume fraction of only 6.5 vol%. The bulk RI of stabilized TiO2 particles could be estimated by extrapolation of the composite RI in dependence of the particle volume fraction. 100 vol% yields bulk RIs of 2.57 ± 0.14 (TMOS), 2.79 ± 0.14 (TOPO A) and 2.83 ± 0.1 (TBC) at 635 nm. XRD analysis (figure 2) determined TMOS and TBC stabilized particles to crystalline anatase TiO2. A similar increase in the composite RI induced by TOPO A particles point at an anatase crystallite structure of TOPO A particles as well, as the crystallinity was clearly verified by HRTEM measurements (figure S4). The good correlation between the estimated bulk RIs and the RI of anatase TiO2 (n = 2.56 at 589 nm) attests to a high-quality layer modelling for fitting the ellipsometric measurement data.
3.2.4. Single-step composite structuring with UV-SCIL.
Substrate conformal imprint lithography (SCIL) was applied to demonstrate high throughput, single-step structuring of the inorganic–organic nanocomposites that contained stabilized TiO2 nanoparticles and to determine the shrinkage of imprinted structures upon temperature exposure. Representative for the fabrication of optical elements waveguide-like grating structures in the micrometre range were fabricated with composites containing up to 10 wt% TMOS, TOPO A and TBC particles. The fabrication of nanoscale structures is shown elsewhere, respectively [34]. Figure 8 shows a SEM image of imprinted waveguide structures using SCIL and a composite with 10 wt% TOPO A particles, representative for the different composites. Structural accuracy was high for all three composites, similar to the pure polymer [24]. Shrinkage upon UV-light induced curing was negligible for both the pure polymer and the composites. The incorporation of 10 wt% stabilized particles, however, alters the processing parameters in terms of viscosity and UV curing time. Layer thickness, after spin– coating for sample preparation, was increased by 100% maximum compared to the pure polymer in case of TBC particle composites, due to an increased viscosity. The exposure time of the UV-induced composite curing reaction increased from 20 s to 2 min at maximum compared to the pure polymer. The increase in exposure time, however, became negligible for layer thicknesses in the nanometre range. The temperature stability in terms of structural shrinkage was investigated for gratings with a width of 7 μm. Cured structures were exposed to a temperature ramp with a holding time of 50 s at 150 °C and the maximum temperature of 220 °C, simulating a typical industrial reflow solder process [34]. Shrinkage of 4% at maximum was equally observed from structural cross-sections for both the pure epoxy polymer KATIOBOND OM VE 110707 and the respective nanocomposites, which contained 10 wt% TBC particles.
Microscopic characterization. The thickness uniformity of 1 × 1 cm−2 large areas for up to 2 μm thick composite layers was determined above 95% by ellipsometric mapping measurements. The full mapping results for a composite layer containing 10 wt% TOPO A particles are exemplarily attached in the supporting information (figure S5). Atomic force microscopy (AFM) topography measurements confirmed the high surface quality at a microscopic scale. The maximum topographic differences over an area of 50 × 50 μm2 of layers with a thickness ranging from 100 nm to 300 nm were smaller than 15 nm with a root mean square surface roughness
