Capillary Force-Induced Glue-Free Printing of Ag Nanoparticle Arrays for Highly Sensitive SERS Substrates

Background

Template-assisted assembly places nanoparticles into ordered arrays, but the arrays must then be transferred from the template to a flat target substrate. Earlier transfer processes relied on surface chemical treatment or glue materials, which the authors note may be detrimental for bioanalysis and biosensor applications. For SERS, randomly distributed metal nanoparticles limit reproducibility and quantitative analysis, and periodic nanoparticle arrays offer a route to more reproducible signals.

Design and approach

The template is an anisotropic buckling structure: a 70 nm-thick polystyrene (PS) layer is transferred onto a 2 mm-thick prestrained poly(dimethylsiloxane) (PDMS) substrate, and releasing the strain wrinkles the PS layer with a wavelength set by the degree of prestrain. Ag nanoparticles synthesized by the polyol process (average diameter 82 nm, standard deviation 16 nm) are dip-coated onto the plasma-treated template at a withdrawal speed of 0.012 mm/min, and stronger capillary forces in the troughs confine the particles there. For printing, 10 μL of water is placed on a hydrophilic silicon (Si) wafer, the filled template is laid on the droplet and removed after overnight drying, so that water, rather than glue, releases the particles onto the substrate.

How it was evaluated

The buckling templates and transferred arrays were characterized by atomic force microscopy (AFM), optical microscopy and scanning electron microscopy (SEM), and measured wavelengths were compared with finite deformation buckling theory. Transfer was compared between a hydrophilic Si target (water contact angle θ < 10°) and a hydrophobic Si target coated with a dodecyltrichlorosilane (DTS) monolayer (θ = 106°), and against a glue-based transfer using a poly(ε-caprolactone) (PCL) layer. SERS was tested with R6G, Nile blue A (NBA) and 1,2-bis(4-pyridyl)ethylene (BPE) after 24 h immersion in solutions from 10⁻³ to 10⁻¹² M, using a 633 nm He-Ne laser at 20 μW for 1 s.

Key results

At a prestrain of 35%, the buckling structure had a mean wavelength of 3.39 μm, comparable to the theoretical value of 3.89 μm, and the wavelength decreased as the prestrain increased. Templates made with prestrains of 20%, 35% and 50% gave Ag nanoparticle line arrays with widths of 4.33, 3.58 and 3.19 μm, respectively.

On the hydrophilic target, the transferred line arrays had a uniform, periodic width and a high concentration of Ag nanoparticles without glue, and arrays were also obtained over a macroscopic area. On the DTS-coated hydrophobic target, most of the Ag nanoparticles were not transferred. The authors attribute the transfer to convective flow of water toward the evaporating meniscus and conclude that it depends mainly on the surface wettability of the target substrate.

The arrays showed clear main Raman peaks of R6G at 10⁻⁹ and 10⁻¹² M; the ratio of the 612 cm⁻¹ peak intensity to the peak-free background (I/I0) was 12.78 and 9.27 at these two concentrations. Main peaks of NBA and BPE agreed with literature values, and arrays transferred onto hydrophobic substrates gave a significantly lower SERS response.

With PCL as a glue layer, most main R6G peaks at 10⁻⁶ M were not observed. The authors attribute this to a much lower nanoparticle density after adhesion-based transfer and to burial of the Ag nanoparticles in the PCL layer during annealing.

How it connects to other research

The related review by Langer and colleagues summarizes the state of SERS and the work still needed before it can be used routinely in analysis and commercial products; this study addresses one part of that problem, the ordered and glue-free placement of Ag nanoparticles, and evaluates it with probe molecules.