Rapid Formation of Self‐Supporting Polydimethylsiloxane Sheets with Periodic Clusters of Embedded Nickel Nanoparticles

The direct and rapid formation of a precise pattern of metallic nanoparticles (NPs) supported and/or embedded in a flexible polymeric substrate is not easy to achive. However, the development of simpler and more reliable procedures is still highly desirable. This paper presents an innovative technique, simple, cheap, and robust, for producing a self‐supported sheet of polydimethylsiloxane (PDMS) embedding periodic arrays of clusters of nickel nanoparticles (NiNPs). The method uses the pyroelectric effect in a periodically poled lithium niobate (PPLN) crystal for producing a surface charge template able to address the patterning of the NPs by applying a simple thermal stimulation. The key advantages are rapidity, single‐step, and electrode‐free operation. The reliability of the technique is demonstrated for different geometries that are called here “dots” and “grid” and for three different periods 50, 200, and 400 µm. The resulting sheets are attractive for both their flexibility and magnetic properties that can be used for detection, entrapment, and/or patterning of micro‐ and nanoparticles in various fields such as microfluidics and biomedicine.

the properties of PDMS. Alternative techniques include soft lithography, [13,26,27] where the fabrication of a separate mold is required for the pattern transferring, which also increases the complexity of the whole fabrication process. More recently also inkjet printing techniques have been presented in literature for printing NP-based inks [28,29] on the surface of PDMS substrates, but they still present different limitations due to nozzleclogging drawbacks.
In case of embedding procedures, the NPs are used as fillers in order to investigate the properties of the composite material by varying typically the surface functionalization, the shape and the concentration. [30][31][32] In particular, there is a field of research that focuses the attention on the fabrication of aligned structures of NPs embedded in PDMS substrates, in order to achieve composite materials with anisotropic properties. The most used means for aligning metal NPs in a polymer matrix is an external magnetic field, [33,34] especially in case of nickel (Ni) NPs. However, nowadays the availability of microtools for handling and separating micro-and nanoparticles in very small volumes remains a challenge and represents a key aspect in many fields such as nanotechnology or biomedicine. Therefore, the development of easier and more reliable procedures is still highly desirable for patterning metallic NPs embedded in elastomer sheets.
Here we demonstrate for the first time a simple, inexpensive and robust technique based on a pyroelectric charge template for producing free-standing PDMS sheets which embed periodic arrays of nickel NPs (NiNPs) clusters. The particles are dispersed randomly in the elastomer base and spin coated onto a lithium niobate (LN) crystal sample microengineered with an array of reversed ferroelectric domains. We use the pyroelectric effect for arranging the particles according to the template defined by the LN domain pattern through a rapid, single-step and electrode-free modality. The PDMS composite with the microarray of NiNPs clusters is peeled-off from the crystal very easily and the same crystal can be reused several times for producing different sheets, allowing us to avoid tedious replications of chemical and/or physical procedures. The high versatility and robustness of the technique allows us to fabricate microarrays with different periods ranging from 400 down to 50 µm. Moreover, we show how, after fabrication, the clusters can be magnetized easily, thus obtaining a freestanding sheet with microarrays exhibiting magnetic properties. The unique features of the proposed method will enable the transfer of different and heterogeneous devices (magnetic, spintronic, and photonic) into functional (flexible and stretchable) and complex polymer substrates. Moreover, the flexible magnetic microarray could be directly integrated into already existing stretchable electronic systems, to realize smart hybrid magneto-electronic devices with the functionality to sense and to respond to a magnetic field.

Results and Discussion
We used c-cut wafers of LN with pyroelectric properties (see the Experimental Section). We fabricated PPLN crystals by electric field poling (see the Experimental Section for details) and Figure 1a shows the typical optical microscope image.
The hexagonal areas correspond to the ferroelectric domains with polarization reversed respect to the surrounding regions. Figure 1b shows the 3D schematic view of the PPLN crystal with the original c− face (i.e., before poling) positioned up and therefore with hexagons exhibiting c+ polarity. Conversely, Figure 1c shows the schematic view of the PPLN crystal with the original c+ face up and therefore with hexagons with c− polarity. At equilibrium conditions, these polarization charges are compensated by surface screening charges of opposite sign (not shown in the schemes) producing a neutral system. For example, in case of the hexagons with c+ polarity www.advmatinterfaces.de (see Figure 1b) the polarization charge (+) into the hexagons is compensated by negative screening charges (−) provided by the environment, and the vice versa occurs in case of the hexagons with c− polarity (see Figure 1c).
We fabricated three PPLN crystals which differ each other only for the period of the hexagonal domains: 1) 400 µm, 2) 200 µm, and 3) 50 µm. We called them P400, P200, and P50, for simplicity, and we investigated the reliability of the technique by using these three crystals under different conditions, as summarized in Table 1.
We prepared a suspension of PDMS with NiNPs, that we call PDMS/Ni, according to the procedure illustrated in the Experimental Section. Each of the three crystals was used for a first experiment with the suspension spin coated on the original c− face and, after solvent cleaning, for a second experiment with the suspension on the original c+ face. The crystals P400 and P50 experienced only the positive ΔT while in case of P200 we studied the effect of both positive and negative ΔT (see Table 1). Figure 2 shows the schematic view of the fabrication steps.
The hexagonal areas in the PPLN crystal correspond to the ferroelectric domains with polarization orientation reversed respect to the surrounding regions. It is noteworthy that they are not relief structures like the scheme makes them to appear. After standard solvent cleaning, the PPLN crystal was spincoated at 7000 RPM with a thin film of the PDMS/Ni suspension (see Figure 2a) and then subjected to the two types of experiments: 1) heating and 2) cooling. In case (1) the sample was initially at room temperature and then heated up to 170 °C on a digital hotplate for about 2 min (see Figure 2b). In case (2), starting from room temperature, the sample was cooled down to −20 °C in a refrigerator for about 5 min (see Figure 2c).
Depending on the specific set of conditions illustrated in Table 1 the final pattern can be of two types that we call here "dots" or "grid." They are both periodic with the period corresponding to that of the hexagons underneath. The dots pattern consists of Ni clusters that, at microscopic scale, fill completely the hexagonal regions, assuming with high fidelity their shape. In other words, the edges of the Ni clusters correspond perfectly to those of the hexagonal domains. Conversely, the grid exhibits the complementary distribution, with NiNPs filling the regions outside the hexagons. In the heating experiment, we can have simultaneously the thermal crosslinking of PDMS if the crosslink agent is added in the suspension. Since the crosslinking occurs at 170 °C, this feature is possible only in the   www.advmatinterfaces.de heating experiment. After the crosslinking, the PDMS becomes solid, so that the NiNPs pattern remains frozen into the PDMS layer that is then peeled-off the crystal easily. As a result, we obtain a freestanding PDMS sheet reinforced with a periodic pattern of Ni clusters (see Figure 2d). In order to observe in real time the formation of the NiNPs clusters, we used a conventional Peltier cell placed under the crystal on the stage of an upright optical microscope. In this way, the crystal experienced a rapid temperature variation that is positive in the heating experiment (1) and negative in the cooling experiment (2), both able to generate the pyroelectric effect (see the Experimental Section). At this stage, in about 10 s, the NiNPs appear to move rapidly in the liquid PDMS matrix until forming a well-defined periodic pattern which follows the geometrical template of the PPLN crystal. Movies S1 and S2 (Supporting Information) were captured under the optical microscope and show the formation of the dot pattern, as an example, in the case of P200, with PDMS/Ni spin coated on the original c− face (i.e., hexagons with c+ polarity), positive ΔT, and with two different concentrations of Ni. It is noteworthy how the displacement of the particles is so clear that the regions on the hexagons are completely full of Ni, while the surrounding ones are transparent with no residue of particles appreciable at microscopic scale. Figure 3 shows the optical microscope images of the PDMS sheets, reinforced with different Ni patterns obtained by using the conditions illustrated in Table 1, while Figure 4 shows the schematic view of the pyroelectric charge template arising after positive and negative ΔT. In particular, the images in Figure 3a,c,e show the dots pattern obtained in case of the original c− face (e.g., hexagons with c+ polarity) for P400, P200, and P50, respectively, when using the positive ΔT. The temperature rise induces the pyroelectric effect, which decreases the net polarization inside the crystal. Correspondingly, the external screening charges loose partially the compensation from the polarization charge and become excessive. This means that the regions outside the hexagons exhibit a positive excess charge and vice versa occurs for the regions inside the hexagons, as reported schematically in the insets of Figure 3a,c,e. The Ni clusters follow with high fidelity the shape of the hexagons underneath, thus generating a pattern with the same periodicity of the hexagons. The results are reproducible for all of the three periods (400, 200, and 50 µm). Figure 3b,d,f shows the images corresponding still to positive ΔT but in case of the original c+ face (e.g., hexagons with c− polarity) for P400, P200, and P50, respectively. As reported in the insets of Figure 3b,d,f, the charge exposed on the face c+ will be negative and the charge exposed inside the hexagon c− will be positive. In this case the clusters form a grid surrounding the hexagons of the PPLN crystal, again following with high fidelity the periodicity of the PPLN crystal. Also in this case, the process exhibits high reproducibility for all of the three periods (400, 200, and 50 µm). Figure 3g,h shows the images corresponding to P200 treated by a negative ΔT with PDMS on the original c− and c+ face, respectively. In the case of negative ΔT, the increasing of the polarization charge leads to an imbalance with the screening charges on the surface, so that on the face c− (hexagons c+) a net negative charge arises and the opposite occurs on the c+ face.
The ferroelectric polarization decreases when experiencing the positive ΔT, thus leading to a temporary excess of screening charges (see the Experimental Section). This means that, in case of the original c+ face, the positive ΔT generates a temporary excess of charge (+) in the areas of hexagons and of charge (−) in the surroundings. In other words, a pyroelectric charge template arises on the surface of the PPLN crystal. This is an alternation of charge (+) and charge (−) that leads to a pyroelectric field distribution which follows the geometry of the reversed domains and with the key innovation of being free from conductive coatings and external high voltage generators. The opposite charge distribution occurs in case of the original c− face, namely with hexagons exhibiting an excess of charge (−) and vice versa in the surroundings. Conversely, in case of negative ΔT, the polarization charge increases in the ferroelectric domains, leading to a temporary imbalance between the polarization charge and screening charges (see the pyroelectric effect in the Experimental Section). As a consequence, on the original c+ face the pyroelectric charge template consists of hexagons with an excess of charge (−) and the surroundings with an excess of charge (+). The vice versa occurs in case of original c− face. As a further example of the reliability of the technique even in case of other geometries, the heating experiment was performed on particular PPLN geometries obtained through twice electric field poling (TEFP). [50] The TEFP approach represents a tool for the fabrication of multiscale PPLN samples. This technique enables the formation of larger hexagonal domains with satellite ones (size below 10 µm). Figure 5a shows the optical microscope image of a typical PPLN crystal after the TEFP process. The double structure, with small satellite domains surrounding the larger primary domains, is visible. Figure 5b,c shows the images corresponding to TEFP PPLN treated by a positive ΔT with PDMS/Ni on the original face c− and c+, respectively.
The NiNPs were centrifugated before mixing with PDMS matrix and this process conferred a positive surface charge to the particles due to the triboelectric effect, [51][52][53] while the PDMS is a nonpolar fluid with low dielectric constant and weak viscosity. In the presence of the pyroelectric charge template of the PPLN crystal, the Ni/PDMS suspension can be considered as an electrorheological suspension. [54] In particular, the motion and redistribution of the NPs in the PDMS layer is determined by the net force F expressed by the following general formula [55,56] EP DEP which is the combination of two prevalent forces acting on the NPs: the electrophoretic (EP) force F EP and the dielectrophoretic (DEP) force F DEP . The first term describes the Coulombic interaction between the net charge q of the NPs and the homogeneous electric field E generated by the pyroelectric charges on the areas of the hexagons and of the surroundings, in other words far away from the hexagon edges. The second term arises from the interaction of the dipole moment p induced in the NPs by the gradient ∇E of the electric field E developing across the borders of the hexagons, where the net charge on the surface jumps between (+) and (−). Approximating the shape of the NPs to a sphere, the net F DEP can be quantified by the following general formula [57] www.advmatinterfaces.de is known as the Clausius-Mossotti factor, [57] r is the radius of the NPs, ε m is the real permittivity of the medium (PDMS), and p ε * and m ε * are the complex permittivities of the particles (Ni) and of the medium, respectively. A general complex permittivity is given by ε* = ε − j(σ/ω), where ε is the real permittivity, σ is the conductivity, j 1 = − , and ω is the angular frequency of the applied field. In this work, the electric field is steady, which corresponds to a regime with ω → 0, and therefore meaning that the polarizability is governed by the conductivities. The conductivity σ p of Ni (≈1.4 × 10 7 S m −1 ) is much higher than that σ m of PDMS (≈2.5 × 10 −14 S m −1 ). Therefore, the real part of f CM (see Equation (4)) is positive and, as a consequence, the F DEP is positive as well. This means that the NiNPs are attracted by the regions with high-field gradient, such as the hexagon boarders. Moreover, a dipole-dipole interaction occurs between the NPs. [58] This causes neighboring particles to attract each other and align themselves forming a network of headto-head connections that leads to the aggregation of chain-like structures oriented along the field lines. In summary, in the regions far from the hexagon boarders the electric field is homogeneous, so that the F EP is prevalent and involves a simple Coulombic interaction which brings the particles close to (attractive force) or far away (repulsive force) from that region, depending on the sign of the polarities. This force is responsible for the clear filling or emptying of the hexagons that we observe in the dots and grid structures fabricated in this work (see Figure 3). At the same time, in the regions across the hexagon boarders we have a strong field gradient, so that the F DEP is prevalent and the NiNPs are attracted by these boarders forming the abovementioned chain-like structures oriented along the field lines. This effect is visible only at higher magnifications, such as in Figure 3a,b,g.
After creating the structures shown in Figure 3 we used a scalpel for detaching gently the cross-linked PDMS layer from the crystal, obtaining free-standing elastomer sheets reinforced with various periodic Ni structures. Figure 6a shows the scanning electron microscope (SEM) image of the NiNPs (that appear quite monodispersed and of polyedrical shape with an average size of about 500 nm) while Figure 6b-h shows some  www.advmatinterfaces.de Figure 6. SEM images of a) the NiNPs (a drop of alcoholic suspension was cast on a grid and dried a room temperature); b) sheet obtained by using the crystal P50/c−; c) cross section of sample P50 generating the dots pattern at high magnification in correspondence of a single cluster; d) cross section of sample P50 generating the dots pattern at a lower magnification under sheet bending; e,f) surface of the sheet obtained by using the crystal P200 producing the dots pattern, under two different magnifications; g,h) the surface of the sheet obtained by using the crystal P200 producing the grid pattern, under two different magnifications; i) ImageJ analysis on the image in (e) relating to the distribution of the particles in the PDMS sheet; j) ImageJ analysis on the image in (g) relating to the distribution of the particles in the PDMS sheet; k,l) large view images captured with Axio Zoom microscope of a typical free-standing sheet with the dots pattern. The red lines in (h) show the chain-like structure formed across the hexagon edges by the dipole-dipole interaction.

www.advmatinterfaces.de
SEM images of the elastomer sheets under different points of view. In particular, keeping in mind the summary of crystals and conditions illustrated in Table 1, Figure 6b-d refers to the sheet obtained by using the crystal P50 generating the dots pattern. The image in Figure 6b shows the surface of the sheet and how the NiNPs form well defined clusters in correspondence of the hexagons, even at submicron resolution. The cross-section in Figure 6c was captured at high magnification in correspondence of a single cluster. It shows that the particles are all aggregated in proximity of the PDMS surface that was in contact with the crystal, filling a depth of about 6 µm. Even at submicron resolution no particles are dispersed far from the hexagons, thus demonstrating the efficiency in particles capturing by the pyroelectric field. The image in Figure 6d is another view of the cross-section at a lower magnification in order to demonstrate the stability of the structure even under sheet bending. Figure 6e,f shows the surface of the sheet obtained by using the crystal P200 producing the dots pattern, under two different magnifications. Also, in this case of 200 µm period, the aggregation of the particles into the hexagonal region is clear at submicron resolution with no particles observable outside the hexagons. We analyzed the image in Figure 6e by ImageJ (open-source image processing program developed at the National Institutes of Health (NIH)), in order to evaluate the average distribution of the particles in the PDMS sheet and the Figure 6i shows the resulting values. On average, 95% of the particles form the hexagonal clusters, leaving only 5% in the surroundings, thus demonstrating the high efficiency of the pyroelectric charge template in displacing the particles according to the dots pattern. Figure 6g,h shows the surface of the sheet obtained by using the crystal P200 producing the grid pattern, under two different magnifications. Also, for the grid structure the particles appear to aggregate into a well-defined pattern, leaving the hexagons well empty with no residual particles. This is confirmed by the quantitative data in Figure 6j where we show the results of the ImageJ analysis on the image in Figure 6g. In average, only the 3% of the particles are located inside the hexagonal regions and 97% fill completely the surroundings. The red lines in Figure 6h show clearly the chain-like structure formed across the hexagon edges by the dipole-dipole interaction mentioned previously. Figure 6k,l shows two large view images of a typical freestanding sheet with the dots pattern, standing on the stage and supported by a couple of manual tweezers, respectively. The structure is well defined and repeatable over an area of about 340 mm 2 and exhibits a good stability even under slight bending.
After peeling-off the crystal, the reinforced elastomer sheets were subjected to a magnetization process by approaching a permanent magnet (iron boron neodymium) for about 1 min, in order to have a magnetic field gradient generated around the clusters of nickel. Movie S3 (Supporting Information) shows a typical reinforced sheet prior magnetization, while Movie S4 (Supporting Information) shows the same sheet that, after magnetization, retains a residual magnetic field capable of reacting to the proximity of a ferrous material such as the metal tweezers.

Conclusion
We have presented a direct and rapid approach for forming free-standing elastomer sheets reinforced by periodic clusters of NiNPs. We demonstrated the reliability of the technique in the case of two different patterns, dots and grid, and for three different periods (50,200, and 400 µm) and geometries. The technique is free from time consuming and expensive mold-based and/or lithography-based procedures, thus appearing very promising for large scale and cost-effective production of elastomers with periodic reinforcement. It makes use of the pyroelectric charge template generated onto the surface of a PPLN crystal by a simple thermal treatment, without using conductive coatings and high voltage generators. The technique is very rapid and robust, and the Ni patterns follow with high fidelity the geometry of the PPLN crystal that can be reused many times. Finally, the magnetization process demonstrates that the technique could open the route to the development of flexible magnetic microdevices to be used for a wide variety of new applications in the fields of sensing and recording, for example for sorting and/or capturing magnetic species at microscale. Anyway, the ability of the proposed strategy in forming reliable polymer membrane embedding period clusters of metallic nanoparticles opens the route for its useful exploitation in many advanced fields of material sciences, microfluidics, biotechnology and medicine.

Experimental Section
Lithium Niobate: The LN crystals were bought from Crystal Technology Inc. in the form of both sides polished 500 µm thick c-cut 3-inch wafers and were cut into (2 × 2) cm 2 sized samples by a standard diamond saw. It is well known that, at equilibrium, the spontaneous polarization P s of a c-cut LN crystal is fully compensated by screening charges from the environment, thus producing an electrically neutral system. According to the pyroelectric effect, a temperature change ΔT causes a variation ΔPs ∝ ΔT, which, neglecting the losses, builds up uncompensated surface charges, corresponding to a charge density σ ∝ Pc ΔT, where Pc is the pyroelectric coefficient. [59] In other words, the temperature variation perturbs the equilibrium state and generates uncompensated charges that, as a consequence, produce a high electric field on the crystal surface. [39] Electric Field Poling: The LN crystal samples were subjected to electric field poling in order to fabricate (2 × 2) cm 2 sized PPLN samples with a square array of ferroelectric domains with opposite orientation and hexagonal shape. The PPLN were obtained by standard electric field poling onto photoresist patterned samples. Three square arrays of hexagons were considered: period 400 µm, period 200 µm, and period 50 µm. [60,61] PDMS/Ni Suspension: The NiNPs were purchased from Sigma Aldrich powder, <1 µm, 99.8% trace metals basis. Their diameter falls in the range 300-500 nm. They were cleaned via multiple cycles of centrifugation and, after removal of supernatant, they were precipitated and dried in a vacuum desiccator for 24 h. After that they were monodispersed in fresh nonpolar PDMS (Sylgard 184, Dow Corning Corp) at a concentration of 12% (w/w) and uniformed by magnetic stirring and ultrasound treatments for 5 h. The PDMS was purchased in the form of two components, a prepolymer, and a cross-linker with a percentage of curing agent 1:10.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.