ARTICLE pubs.acs.org/Langmuir
Microcontact Printing of Monodiamond Nanoparticles: An Effective Route to Patterned Diamond Structure Fabrication Hao Zhuang,† Bo Song,‡ Thorsten Staedler,† and Xin Jiang*,† †
Institute of Materials Engineering and ‡Department of Physical Chemistry 1, University of Siegen, 57076 Siegen, Germany
bS Supporting Information ABSTRACT: By combining microcontact printing with a nanodiamond seeding technique, a precise micrometer-sized chemical vapor deposition (CVD) diamond pattern have been obtained. On the basis of the guidance of basic theoretical calculations, monodisperse detonation nanodiamonds (DNDs) were chosen as an “ink” material and oxidized poly(dimethylsiloxane) (PDMS) was selected to serve as a stamp because it features a higher interaction energy with the DNDs compared to that of the original PDMS. The adsorption kinetics shows an approximately exponential law with a maximum surface DND density of 3.4 1010 cm2 after 20 min. To achieve a high transfer ratio of DNDs from the PDMS stamp to a silicon surface, a thin layer of poly(methyl methacrylate) (PMMA) was spin coated onto the substrates. A microwave plasma chemical vapor deposition system was used to synthesize the CVD diamond on the seeded substrate areas. Precise diamond patterns with a low expansion ratio (3.6%) were successfully prepared after 1.5 h of deposition. Further increases in the deposition time typically lead to a high expansion rate (∼0.8 μm/h). The general pattern shape, however, did not show any significant change. Compared with conventional diamond pattern deposition methods, the technique described here offers the advantages of being simple, inexpensive, damage-free, and highly compatible, rendering it attractive for a broad variety of industrial applications.
’ INTRODUCTION Synthetic chemical vapor deposition (CVD) diamond is well known for its exceptional physical and mechanical properties.13 It has been considered to be an excellent candidate for micro/ nanoelectromechanical systems,4 biosensors,5,6 field-emission devices,7,8 and other applications. Among these microelectronic applications, the patterning of diamond films is possibly the most potent strategy in the context of miniaturization of size and in the improvement of the performance of the corresponding devices. However, because of its low chemical reactivity, diamond cannot be patterned easily by conventional microfabrication processes. Typically, two paths are available to obtain patterned diamond surfaces: (i) dry etching by strong energetic species such as lasers, ion beams, and reactive plasma9,10 and (ii) selective area deposition (SAD) utilizing diamond nucleation differences on different surface areas to grow corresponding diamond patterns.1119 With regard to the high costs along with potential surface damage associated with dry etching processes,9,15,20 the SAD method appears to be very attractive. Unfortunately, conventional SAD normally involves a significant number of substrate processing steps to either remove diamond nuclei (seeds) from undesired surface areas14 or to deposit other materials (such as Pt, Si3N4, etc.), which yield a high diamond nucleation density, onto the desired areas.19 These processes not only add complexity and uncertainty to the production but also give rise to the risks of damaging the substrate and/or future device failure.14,15 To address these issues, Fox et al. developed a method to print r 2011 American Chemical Society
diamond seed patterns directly onto unmodified substrates via an inkjet printing technique. This strategy minimizes the risk of substrate damage and at the same time was found to be highly compatible with most conventional substrates.16 Nevertheless, because of the low resolution (25 μm) of the commercial inkjet technology, this method is able to provide minimal diamond feature sizes of only about 70 μm, which limits its applicability in the context of the fabrication of fine diamond structures.12,13,16 Until now, despite its technological importance, a simple, inexpensive, high-resolution SAD method has not been reported. In this article, following up on the idea of selective area printing of diamond patterns, we describe a new method that combines microcontact printing (μCP) with a nanodiamond particle seeding technique in order to fabricate fine diamond pattern structure. We chose this approach because (1) the μCP technique is a well-recognized and important tool for the selective patterning of surfaces on the microscale and submicroscale; (2) it has been widely used in patterning organic solvents,21,22 nanoparticles,21,2327 polymers,28 DNA,29 proteins,30,31 and cells32 onto different substrates; and (3) the commercial availability of different sizes of nanodiamond particles allows for a wide choice of seeds. The general concept of μCP is straightforward and analogous to ordinary stamping that Received: June 28, 2011 Revised: August 25, 2011 Published: August 25, 2011 11981
dx.doi.org/10.1021/la2024428 | Langmuir 2011, 27, 11981–11989
Langmuir
ARTICLE
Figure 1. (a) Calculation of the packing density vs particle size in a densely packed particle layer. (b) Calculated electrostatic interaction energy between DNDs and the PDMS surface.
we know from everyday life. The key issues in the μCP process are how to ink the stamp and how to print the pattern onto the substrate of interest. In this article, we report the fabrication of well-defined high-resolution diamond patterns by a μCP process. The key steps leading to our successful attempt, such as the selection of the most suitable diamond nanoparticles, the highdensity inking of the stamp with these particles, and the high transfer ratio of diamond particles onto the substrate, have initially been predicted by a theoretical calculation and have finally been confirmed by zeta potential, atomic force microscopy (AFM), and scanning electron microscopy (SEM) analyses. Compared with the existing methods mentioned above, the technique introduced here offers several distinct advantages: (i) it does not damage the substrate and is highly compatible with a large variety of potential substrate materials; (ii) it does not require any lithography procedures; (iii) it features high resolution, which is dependent only on the resolution of the corresponding PDMS stamp, and in principle it can reach values below the 100 nm mark;27 (iv) it is inexpensive, simple, and highly reproducible; and (v) the materials as well as the processing itself are not dangerous. Basic Considerations for Successful μCP. For μCP printing, it is essential to choose a suitable “ink” material. In our case, an appropriate diamond dispersion represents the ink. Its basic requirement is to induce a sufficient nucleation density for diamond growth. The nucleation density not only influences the final film morphology but also determines the minimum time it takes to synthesize a closed film without any pinholes. To achieve a high nucleation density on the substrate after the printing process, the initial density of adsorbed diamond particles on the stamp itself needs to be as high as possible. As shown in Figure 1a, according to a very basic calculation of the dependence of the maximum particle density on the particle size in a hexagonal close-packed geometry, smaller diamond particles will result in a higher packing density. Therefore, the smallest commercially available diamond particles—detonation nanodiamonds (DNDs, 5 nm, see Figure S1 in the Supporting Information)—which are processed by high-energy ball milling to break the aggregations, are chosen as the ink material. However, it is reported that the actual adsorption density of DNDs on the surface is significantly lower (1011 cm2 is reported on the Si substrate33) than its theoretical value (4 1012 cm2). The reason behind this finding lies in the electrostatic repulsion between the individual particles and their size distribution in the
dispersion. Nevertheless, the achievable values are still more than sufficient to ensure the synthesis of closed diamond films in a short time. The later requires seeding densities on the order of 1010cm2.34 The next step, after the identification of a suitable ink, is to answer the question of how to transfer the ink effectively to the poly(dimethylsiloxane) (PDMS) stamp. This inking step determines the maximum density of the final printed DNDs on the substrate surface and is strongly related to the surface properties of the PDMS stamp. The original PDMS surface is covered with OSi(CH3)2O groups featuring a hydrophobic nature.35 By utilizing oxygen plasma, it is possible to oxidize such a surface, which is a very popular strategy in PDMS surface modification. After oxidation, the surface becomes hydrophilic because of the conversion of Si(CH3)2O groups into OnSi(OH)4 n groups.35 On the basis of the different surface properties, the adsorption behavior of DNDs on the original and oxidized PDMS surfaces will be different. To shed some light on these differences and to direct our experiments, the classical DerjaguinLandauVerweyOverbeek (DLVO) theory was employed to calculate the interaction energy between DNDs and the two surfaces. In the DLVO theory, the interaction energy between colloidal particles and a planar surface (GTOT) is composed of the van der Waals interaction energy (GLW) and the double-layer interaction energy (GEL). This can be expressed as GTOT = GLW + GEL.36 The van der Waals interaction energy is given by36 GLW ðzÞ ¼
Ar 6z
ð1Þ
in which A is the Hamaker constant for the system of DNDs/ water/PDMS, r is the radius of the DNDs, and z is the distance between the DNDs and the PDMS surface. In the present study, the differences in the van der Waal interaction energies between DNDs/original PDMS and DNDs/oxidized PDMS systems are derived from the differences in the Hamaker constants of original PDMS and oxidized PDMS. According to Israelachvili’s theory, the Hamaker constant of the surface can be calculated from its Lifshitzvan der Waals surface tension component (γLW).36 However, the γLW values of PDMS were found to be similar irrespective of the method of surface modification,37,38 suggesting identical Hamaker constants for the DNDs/water/original PDMS and DNDs/water/oxidized PDMS systems, which in turn results in the same Lifshitzvan der Waals interaction forces in 11982
dx.doi.org/10.1021/la2024428 |Langmuir 2011, 27, 11981–11989
Langmuir
ARTICLE
both cases. As a consequence, the differences in the total interaction energy of both systems are mainly determined by the double-layer interaction energy, which is given by 2 kT Yp Ys expð kzÞ ð2Þ GEL ðzÞ ¼ 4πεr ε0 r e in which εr is the dielectric constant of water, ε0 is the vacuum permittivity, r is the radius of the DNDs, k is the Boltzmann constant, T is the absolute temperature, e is the unit charge of an electron, k is the inverse Debye length, z is the distance between a particle and the PDMS surface, and Yp and Ys are the effective reduced surface potentials of isolated DND and isolated PDMS surfaces (original or oxidized), respectively.39 In the case of distilled water, the pH is measured to be 5.2 because of the absorption of CO2 and the Debye length (k1) is 120 nm. Yp and Ys can be determined by39 eζ 8 tanh 4kT Yp ¼ ð3Þ ( " # )1=2 2kr þ 1 eζ 2 1 þ 1 tanh 4kT ðkr þ 1Þ2
eζ Ys ¼ 4 tanh 4kT
ð4Þ
The ζ potential of the DNDs was measured experimentally and shows a value of +48.2 mV, which reveals a positively charged surface owing to the formation of nonfreezing hydration shell on the surfaces of DNDs.40 The ζ potentials of the two PDMS surfaces were taken from the literature. Here, the corresponding values were found to be around 15 mV for the original PDMS surface and 49 mV for the oxidized PDMS surface.41 In accordance with the literature, the higher zeta potential value of the oxidized PDMS surface can also be observed on the basis of its high hydrophilicity (water contact angle of
