Low-cost and high-sensitivity label-free sensing chips

Surface plasmon resonance (SPR) sensing is a real-time and label-free detection technique that measures biomolecular interactions on a gold surface. It has the potential to benefit numerous important fields, including medical diagnostics, environmental monitoring, and food safety. The most common method to induce SPR on the surface is to use an optical prism known as the Kretschmann configuration. Based on this SPR excitation technique, commercial instruments enable real-time and label-free measurements of biomolecular binding affinity. In addition to the prism-coupling method, the SPR can be excited using metallic nanostructures. Biosensors based on extraordinary transmission of periodic nanohole or nanoslit arrays in gold already have been proposed ^{1, 2}.

Compared to the prism-based SPR sensors, gold nanostructures benefit from having a small detection volume and normal light incidence. They provide a feasible way to achieve chip-based and label-free detection for modern DNA and protein microarrays. However, high-throughput and low-cost fabrication techniques are the main issues that need to be addressed. The majority of the fabrication techniques for metallic nanostructures use focused ion beam (FIB) milling or electron-beam lithography (EBL) combined with a dry etching method or a lift-off process. However, these writing techniques are not suitable for mass production as the processes are expensive and time-consuming. In addition, gold films made by vacuum deposition lead to higher surface plasmon propagation loss due to the scattering of electrons at gold grain boundaries.

Figure 1. (a) A chamber with temperature and pressure control for nano pattern transfer from a gold coated silicon template to a plastic substrate. (b) An optical image of a large-area of gold nanoslits on the polycarbonate film. (c) A scanning electron microscope image of nanoslits on a polycarbonate film with 70nm slit width and 500nm period, both of which match the template pattern.

We developed a thermal annealing-assisted, template-stripping method to make large-area and high-quality gold nanostructures on plastic films.³ In this method, the plastic film is heated above the glass-transition temperature. We also used a polyethylene terephthalate (PET) thin film as a sealing film. Nitrogen gas was introduced into the chamber to produce high pressure on the PET film. It pressed the gold-coated silicon template and the plastic film with large-area uniformity (see Figure 1a). Because gold film has poor adhesion to the silicon template and high adhesion to the softened plastic film, the gold nanostructures were completely transferred to plastic film after peeling off from the template. Figures 1b and 1c show the optical image of the large-area sensor and the scanning electron microscope (SEM) image of the gold nanoslits on the plastic film. The atomic force microscopy images show the root mean square (RMS) roughness values are 1.50nm for the gold film on silicon template, 1.90nm on plastic substrate, 2.31nm on glass substrate, and only 0.890nm for our proposed template-stripped gold film. By comparing the RMS values of the gold film before and after thermal-annealing template-stripping, we observed a great improvement on both the surface flatness and grain size.

The smooth gold surface greatly improves the resonant quality of surface plasmon polaritons. As examined from the transmission spectra, the resonant profile is not a Lorentzian, but more like a Fano resonance with a coupling of the Bloch wave surface plasmon polariton (BW-SPP) in periodic nanostructures and localized surface plasmon resonance (LSPR) in the nanoslit gap. We found that the spectral line width of the Fano resonance decreased with the slit width reduction, and the narrowest line width is 6nm. The sensors can reach an intensity sensitivity up to 10367%/refractive index unit (RIU). This value is much higher than those of previous nanoslit and nanohole arrays fabricated by the EBL and FIB methods.

In the intensity-detection method, the signal is taken at a single wavelength. It has a higher noise compared to angular detection. The spectral intensities near the Fano resonance are very sensitive to surface environmental changes. If we integrate all the changes of spectral intensities near the Fano resonances, the signal-to-noise (S/N) ratio of the sensors can be efficiently improved.

We demonstrated the improvement of S/N ratio by using dual-period nanogrid structures with two distinct Fano resonances on the x and y axes. Using the spectral-integration method to analyze the Fano resonance changes under different refractive index media, we obtained five times the S/N ratio enhancement.⁴ Using the thermal annealing-assisted template stripping method and the dual-period nanogrids, we achieved a ∼3×10⁻⁶ RIU detection limit. Such a detection resolution is comparable with commercial SPR machines using a complicated angular detection method. Nevertheless, our sensors take advantages of a large-area, prism-free, simple detection system and they have high-throughput and a low fabrication cost. In the future, we plan to develop novel biochips with highly sensitive Fano resonances for point-of-care applications.