Manipulating the Morphology of Silver Nanoparticles with Local Plasmon‐Mediated Control

The local morphology of silver nanoparticle films on polydimethylsiloxane can be continuously tuned by illumination and subsequent exposure to water. Illumination causes a plasmon-mediated increase of adhesion. The lower the adhesion, the more water “development” increases the film coalescence. Therefore, a photolithography-like resist-free process is found that allows controlling silver nanoparticles. Plasmonic resonances allow for an extraordinarily strong interaction of silver nanoparticles (Ag NPs) with visible light,1 enabling a plethora of applications to be developed, e.g., surface-enhanced Raman spectroscopy,2 biological detection,3 or light trapping in solar cells.4 A facile control of the spacing of plasmonic NPs and hence the surface plasmon localization would allow local fine tuning between either large propagation length or increased field enhancement.5 For example, efficient plasmonic concentrators may apply localization tuning in order to collect electromagnetic power with delocalized surface plasmons and focus it into localized plasmon hot spots. The resulting intensity enhancement in such hot spots is a key to various plasmonic applications.6 We have discovered a novel effect, allowing for the facile photolithography-like continuous tuning of a moisture-promoted coalescence of Ag NP films at well-defined positions even on large-area substrates. We will show that the final particle distance is controlled by plasmon-mediated photooxidation of the interface between a polydimethylsiloxane (PDMS) substrate and the Ag NPs, which can ultimately even freeze their initial size and shape. Our results clearly show that a key to this new technique is the weak adhesion of Ag NPs on PDMS, which furthermore paves the way for the subsequent transfer of the modified plasmonic structures to any substrate of arbitrary shape. The preparation of films and devices on PDMS has been the object of tremendous interest during the past few years7 as elastically stretchable (opto)electronics conforms to curved surfaces,8 textiles,9 and even living tissue, allowing realization of new quality neuroprosthetic interfaces.10 The low surface free energy11 of around 23 mJ m−2 and weak adhesion of solids render PDMS the ideal stamp material for contact printing even to curved substrates.12 However, for thin-film deposition, poor adhesion is typically considered undesirable and thus adhesion promoters are used.13 Here, we intentionally omit the application of adhesion promoters and harvest the particularly low adhesion of silver on PDMS. The resulting weak interaction between film and substrate supports the formation of islands that occurs at film mass thicknesses below the percolation threshold. Typically, the size, shape, and distance of such nanoparticle (NP) islands and consequently their optical properties critically depend on parameters that cannot be controlled locally, like the mass thickness or the post-deposition annealing temperature.14 The reported technique, however, “lithographically” tunes the morphology of Ag NP films on PDMS. As the optical properties of island films depend strongly on their morphology, this modification can be easily seen with the naked eye, which is illustrated by Figure 1. After visible light exposure of the NP film through a simple shadow mask, the pattern of the mask is transferred to the film. Interestingly, the pattern appears slowly in the silver island film when left in air after exposure. We term this second process development in analogy to standard photolithography. Visible light exposure alone, even in oxygen-containing atmosphere, does not influence the optical properties of the films and does not lead to the pattern formation (see Figure 1b). Therefore, the results presented here must not be mistaken with recent studies about photooxidation of Ag NPs observed after illumination at large intensities (five orders of magnitude higher compared with that of our exposure).15 On the contrary, in our case only the areas that were kept in darkness alter during development. For technological application, a development in liquid water can favorably be used as it takes only a few minutes (see real-time video, Supporting Information for 45 °C) or even seconds at higher temperatures (70 °C). For spectroscopic studies of this phenomenon, we have used the more sluggish development in air with controlled humidity. As shown in Figure 1, films that have not been exposed to visible light show a color change from reddish to gray when stored in air (22 °C, 42% RH). Figure 2 shows the transmission (T) and reflection (R) spectra of such films during that color change. Over time, the transmission spectrum between the plasma wavelength at around 330 and 1100 nm gradually changes from a V-like shape into an L-like shape. The corresponding absorption spectra (1-R-T), shown in Figure 2b, show an increasing absorption at long wavelengths. Thermal annealing studies of island films have shown that the width of the absorption spectra is determined by inter-particle interactions.16 Thermal annealing leads to a de-wetting of the silver film and results in an increased separation of the Ag NPs. At the same time, a blue shift and narrowing of the absorption spectrum is found.17 It is essential to note, that in our case, the absorption change of pristine samples stored in humid air shows the very inverse behavior. It is, therefore, straightforward to conclude that in humid air the silver islands coalesce, so that an increased inter-particle coupling is enabled. The effect of particle coalescence is not observed during the storage of the samples in pure nitrogen. This indicates an atmospheric phenomenon and the need for a spectroscopic analysis under controlled atmosphere. Note, thermal annealing of our samples leads to the expected spectral changes (Figure S1, Supporting Information) and therefore reversibly recovers the influence of humid air to some extent. Figure 3 shows the time dependence of the transmission of green (530 nm) and infrared light (1000 nm), respectively. Initially, the sample chamber is purged with pure nitrogen (99.999%), whereas it is opened to the ambience (22 °C, 42% RH) afterward. Evidently, the spectral changes observed in Figure 2 are directly related to the change of the environment. Further experiments using dry air (shown in Figure S2, Supporting Information) clearly show that the observed color change of silver island films on PDMS is caused by water. This can also be confirmed by placing a drop of water on top of a sample [Figure S3 and also shown in the video (Supporting Information)]. It is essential to stress, that in contrast to recent reports,18 water is found to modify the Ag NP morphology even in the absence of an adjacent semiconductor as well as in complete darkness. Comparing the color change of the samples in humid atmosphere to the colors of the patterns in Figure 1, it can be concluded that the unexposed area has changed due to the storage in humid air (as explained by Figure 2), while the areas exposed to visible light have remained unaltered. This suggests that visible light exposure may substantially affect the described moisture sensitivity. In order to verify this assumption, the samples were exposed (20 min, 13.8 mW cm−2, λ > 330 nm) in different ambience and then developed in humid air. After light exposure in vacuum or in pure nitrogen, development has the same effect of particle coalescence as if of the samples that had not been exposed to light at all (Figure 2). However, for exposure in dry air or pure oxygen, a different behavior is found as shown upon subsequent development (Figure 4). While, for the unexposed area, a coalescence of the Ag NPs leads to the expected shift from V-like to L-like shape of the transmission spectrum, the silver particles in the exposed area apparently remain practically unaffected by the development. The exposure with visible light in oxygen-containing atmosphere apparently prevents the modification of the Ag NP film on PDMS upon interaction with water. As the development happens after the light exposure is already finished the exposed areas must have some kind of memory of the process of exposure. Note, at low humidity levels, this process of development can even take several days. On the other hand, optical excitations within the silver particles, plasmons, charge carriers, and ions will typically not exist for these long time scales after the exposure is finished. To understand how illumination can permanently prevent the morphology change in water, it is essential to note that the application of an adhesion promoter for silver [self-assembling monolayer of MPTMS ((3-mercaptopropyl)-trimethoxy-silane)] before film deposition completely removes the sensitivity of the Ag NP film to water (Figure S4, Supporting Information). Therefore, the described morphology change in water relies on the weak adhesion of the Ag NPs on PDMS. In view of the above results, we conclude that the light exposure locally increases the adhesion of a Ag NP on the PDMS substrate. This process may “freeze” the original morphology, while the NP films in the unexposed areas coalesce in the presence of moisture. The increased adhesion after light exposure is evidenced by the finding that the silver on unexposed areas is removed during ultrasonic bath, while the areas of Ag NP exposed to light remain on the substrate (shown in Figure S5, Supporting Information). The increased adhesion of the Ag NPs is caused by photooxidation at the interface between the particles and the PDMS. As the effect of adhesion increase is not observed on glass substrates, but on PDMS only, photooxidation only of the Ag NPs alone does not explain our findings. Photooxidation is usually found to require much higher light intensities than that used in the present work. In addition, photooxidation leads to an immediate change of the optical properties15 that was not found in our experiments (Figure 1b). On the other hand, PDMS is known to be hardened by oxidation,19 which improves adhesion.20 In fact, PDMS cross-linking is increased by photoaging without significant change of its optical properties.21 However, pure PDMS sheets (without NP films on top) are found to remain unaltered after light exposure under the conditions used in our paper. This is directly evidenced by the observation that Ag NP films deposited on substrates, which have undergone this previous light exposure in oxygen show the same Ag NP coalescence in humid atmosphere as Ag NP films deposited on pristine PDMS. Therefore, we assume that the excitation of particle plasmon resonances in the Ag NPs must mediate the photooxidation. In fact, Tanabe and Tatsuma22 have recently shown that localized surface plasmon resonances can cause photooxidation at the interface between Ag NPs and substrate. The suggested model for the described observations is summarized by the schematic shown in Figure 5 together with an atomic force image at the corresponding edge between an exposed and an unexposed area. In the unexposed area (right), during development the loosely adhering silver islands are coalescing in the presence of moisture as indicated by the reduced roughness and the characteristic change of the spectroscopic signature (Figure 4). In the exposed area (left), the plasmon-mediated photooxidation at the interface between Ag NPs and PDMS (red circles) increases the adhesion (Figure S5, Supporting Information), which “freezes” the silver islands and prevents their coalescence in moist atmosphere. The microscopic transition shown in the atomic force image (tapping phase) of Figure 5 is remarkably sharp. Single small particles show a diameter of around 50 nm, some larger clusters of particles are also observed. A topography image of the transition is shown in Figure S6 (Supporting Information). The exposed area exhibits a roughness of 21.3 nm (rms), while the unexposed area is getting smoother during development: after development in air (42% RH, 22 °C), a roughness of 0.9 nm (rms) is found that is even further reduced upon development in liquid water (0.73 nm rms, not shown). In contrast to standard lithography not only is a binary exposed/unexposed distinction (as shown in Figure 5) possible but also continuous local tuning of the particle separation through the exposure dose (Figure S7, Supporting Information). The local definition of different doses may be realized with masks of varying transmittance or by consecutive exposure through different standard lithography masks. After development in water or humid atmosphere on each substrate position, a well-controlled island coalescence is created. In this way, a local control of surface plasmon localization is achieved in a photo-resist free lithographic process of unsurpassed simplicity. Once modified on top of the PDMS “stamp,” the film can be transferred to any substrate and incorporated into arbitrary device stacks (see Figure S8, Supporting Information). In summary, we have discovered a new kind of photolithography that allows us to locally control the morphology of silver island films. The underlying mechanism is unique for the weak adhesion of the Ag NP films on PDMS. When developed in moist ambient, the Ag NPs coalesce. The critical dependence of this coalescence on the film adhesion paves the way for a lithographic control on the nanoscale. When exposed to light in oxidizing atmosphere, the large local plasmon intensity allows photooxidation at the interface between Ag NPs and the PDMS, resulting in improved local adhesion. Photooxidation of the Ag NPs has been ruled out as the sole explanation for this mechanism. Above a critical exposure dose around 16 J cm−2, coalescence in moisture is completely prevented. Smaller doses, however, facilitate fine tuning of the coalescence and the optical properties. We believe that this new approach will enable plasmonic devices beyond reach of state of the art technology. The PDMS substrates were prepared by curing a 1:10 mixture of curing agent and PDMS precursor (Sylgard 184) on a polished stainless steel plate for 1 h at 80 °C. The cured PDMS foils of around 0.5 mm thickness were removed from the plate and cut into pieces of 17 mm2 × 17 mm2 size. Silver films of 12-nm mass thickness according to quartz crystal monitoring were then thermally deposited in vacuum (10−7 mbar) at a rate of 0.1–0.5 Å s−1. We have deposited the silver films without any additional treatment or adhesion promoter directly on the top surface that was facing to air during the curing. After deposition, the films were kept in darkness and under nitrogen atmosphere. For the photolithographic exposure (20 min) of the samples, a mercury arc lamp (13.8 mW cm−2) with optical filter removing ultraviolet light (λ < 330 nm) was used. In order to study the processes during exposure and development, the transmission (T) and reflection (R) spectra were determined under controlled atmosphere. As light source and detector, a white deuterium halogen lamp (DH-2000-BAL, 800 μW cm−2, OceanOptics) and a spectrometer with a range from 186 to 1041 nm (USB 2000+XR1-ES) were used, respectively. The film topography was investigated with an atomic force microscope (diInnova from Bruker) in tapping mode. Silicon cantilevers between 215 and 235 μm in length, resonance frequency of around 84 kHz, and tip height of 15–20 μm were used. The authors are grateful to S. Adamczyk from the Department of Macromolecular Chemistry (University of Wuppertal, Germany) for carrying out the AFM measurements. P.G. acknowledges funding by the Emmy-Noether-Programm of the DFG (Deutsche Forschungsgemeinschaft). As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.