The UV-SCIL fabrication process was developed and optimized to improve the quality of the nanostructures on the hard substrate transferred with substrate conformal imprint lithography (SCIL) technology. In particular, the key steps such as coating imprint resist, exposure time and etching time were investigated thoroughly. The experiment’s results illustrate that imprint resist could well serve as an etching mask for the dry etching process without oxygen plasma. The optimized etching condition and SCIL technology could also be used to transfer nanostructures on different substrates for metal nanostructured biosensors or nanophotonics.
I. INTRODUCTION
With the development of micro/nanofabrication technology, the semiconductor industry has been pushing for the development of high-precision nanoscale lithography to manufacture ever-smaller nanostructures or higher-density integrated circuits (ICs) on silicon wafer.1 Moreover, the ability to fabricate structures from the micro- to nanoscale with a high degree of precision in a wide variety of materials is of crucial importance to the advancement of micro- and nanotechnology and the nanosciences.2,3 Fabrication technologies with a high precision and sufficient resolution have also been applied in other research fields such as nanophotonics, micro- and nanofluidics, and chip-based sensors. However, critical issues such as the resolution, reliability, spin-speed and overlay accuracy of the imprint resist need to be overcome. Therefore, the development of new lithography methodologies has become extremely urgent in meeting the demand for such technology, and also for industrial nanofabrication.
Substrate conformal imprint lithography (SCIL) is an innovative full wafer scale soft lithography technique,3 which is based on the transfer of patterns from flexible polydimethylsiloxane (PDMS) stamps into ultraviolet (UV) light curing resists or sol-gel materials cured by the diffusion of solvents into the PDMS stamps with nanostructures.4 It also is an emerging high-resolution parallel patterning method, mainly aimed at fields in which electron beam and high-end photolithography are costly and do not provide sufficient resolution at reasonable throughput.4–6 SCIL bridges the gap with nanoimprint lithography (NIL) using rigid stamps for the best resolution and soft stamps for large-area patterning. It is considered a way of satisfying the demand for the fabrication of next-generation chips with even smaller feature sizes. SCIL technology has potential applications such as nanophotonics, which require the transfer of the dense nanopatterns into various substrate materials using the imprint resist structure as a mask for the etching process. Therefore, the masking capabilities of UV-SCIL resist for different etching processes are of great interest to potential users. However, the thickness of the residual resist in SCIL technology and the etching condition of the residual resist are the two key barriers to widely extending the SCIL technology application. The optimized fabrication process, which optimizes the fabrication process and guarantees the figure of the nanostructure, is described here.
II. EXPERIMENT
The SCIL processes were carried out on SUSS MA6 (Fig. 1). The processes include composite working stamps consisting of a glass carrier with patterned rubber, which were replicated from the original master pattern. The fabrication processes of the PDMS nanoimprint stamp are described in ref. 4 In particular, the in-plane rigidness of the glass carrier avoids lateral stamp deformation caused by vacuum fixing on the stamp holder. Moreover, the flexibility in the out-of-plane direction of the thin glass and PDMS allows a conformal imprint over large areas.
Photograph of MA6 mask aligner with SCIL upgrade tool: (a) 6-inch PDMS master; (b) holder for wafer; (c) SCIL device mounted on MA6.
Photograph of MA6 mask aligner with SCIL upgrade tool: (a) 6-inch PDMS master; (b) holder for wafer; (c) SCIL device mounted on MA6.
The surface of the silicon wafer was not coated with the adhesive layer. Before spin-coating imprint resist, the surface of the silicon substrates was treated with oxygen plasma. The imprint resist used in the experiments is AMONIL-MMS4, which is UV-curable (365 nm) resist and has a material viscosity of 50 mPa⋅s. After coating imprint resist (MMS4), the wafer was baked on a heat plate at 115∘ C for 30 seconds. SCIL includes the following processes: loading a stamp, loading the substrate coated with imprint resist (MMS4), activating the slide vacuum, imprinting and exposure (dose: 2 J/cm2). By applying low pressure at room temperature, the nanostructure features were formed within a few seconds due to the low viscosity of the imprint resist. The imprint resist was hardened via UV-light (365nm) through the back of the template. Finally, the substrate and template were separated by releasing the pressure.
III. RESULTS AND DISCUSSION
A. Imprinting by UV-SCIL
To reduce the thickness of the residual resist at the concave after imprinting, the spin-coating condition of AMONIL-MMS4 imprint resist was optimized by changing the spinning time and acceleration. Detailed information on the spinning condition is shown in Tab. I. The change of the thickness of MMS4 imprint resist with the spinning speed is depicted in Fig.2, which shows that the most important way to imprint a perfect nanostructure is to increase the spinning speed, thus reducing the thickness of the imprint resist. When the spinning speed reached 5000 rpm, the thickness of the MMS4 imprint resist was 161 nm.
(a) The thickness of MMS4 imprint resist changed according to the spinning speed; (b) the difference in the thickness of MMS4 imprint resist with a spinning speed of 4000 rpm.
(a) The thickness of MMS4 imprint resist changed according to the spinning speed; (b) the difference in the thickness of MMS4 imprint resist with a spinning speed of 4000 rpm.
The imprinted wafer and the SEM of the imprinted nanostructure are shown in Figs.3 and 4, respectively. The imprinted area reached ∼20 cm2 (Fig.3), which is bigger than that of electron beam lithography (EBL) fabrication. All these images show that the nanostructures can be transferred accurately by UV-SCIL technology with imprint resist (AMONIL-MMS4). The images also illustrate that the quality of the pattern transferred is at least comparable to the quality achieved with EBL.
The fabrication results: (a) the results on 4-inch silicon wafer; (b) the results on 2-inch sapphire wafer.
The fabrication results: (a) the results on 4-inch silicon wafer; (b) the results on 2-inch sapphire wafer.
SEM of the imprinted nanostructures. (a) top view of the imprinted nanostructures; (b) the shape of one nano-dot; (c) cross-section of the imprinted structures; (d) SME of nanostructures with tilt of 15 degrees.
SEM of the imprinted nanostructures. (a) top view of the imprinted nanostructures; (b) the shape of one nano-dot; (c) cross-section of the imprinted structures; (d) SME of nanostructures with tilt of 15 degrees.
The representative thickness of the residual resist was 47 nm and the ratio of the remaining resist to thickness of MMS4 was ∼0.23. The thickness of the residual resist reduced as the thickness of MMS4 reduced.
B. Etching
The nanostructured resist on the silicon substrate acted as an etching mask for the dry etching process. The remaining resist should be etched before taking the following fabrication steps. The etching of the residual resist was an important process in transferring precisely the nanostructures on the substrate; however, on the nanoscale the etching gas, etching time and ICP power will affect the final nanostructure. Therefore, an optimized etching condition is necessary for nanostructure fabrication. In the experiments different etching gases were used to etch the residual resist to obtain the ideal nanostructure on the substrate, such as the following: Cl and O (Fig. 5); Ar and Cl (Fig. 6); Ar and O (Fig. 7); Ar, Cl and O (Fig. 8); and, Ar and Cl (Fig. 9). In the etching process the ICP (Inductively Coupled Plasma) power, RF (Radio Frequency) power and pressure were also key parameters in etching the residual resist. The etching results of different etching conditions are shown in Figs. 5 to 9.
The etching results with a mixture of three etching gases and different flow rates.
The etching results with a mixture of three etching gases and different flow rates.
The etching results with etching gases Ar and Cl successively and different flow rates.
The etching results with etching gases Ar and Cl successively and different flow rates.
In the first estimation of the etching rate of the residual resist, unstructured resist layers are exposed to different etching processes. Imprinted wafers are etched after removal of the residual resist layer by etching gases. If the etching rate is quick, the residual layer will be etched, and the silicon wafer and nanostructured resist will also be etched. The nanostructures on the substrate were characterized by scanning electron microscope (SEM).
Fig. 5 shows that the etching rate was so quick that the substrate was etched. Moreover, the nanostructured imprint resist was also etched. The result could not be used to lift off the nanoplasmonic biosensor sensing elements or other nanophotonic devices.
To find good etching gas mixtures, different flow rates and mixture ratios of Ar and Cl were used to etch the residual resist. Fig. 6 shows the good etched results, which were obtained with flow rates of Ar and Cl of 20 sccm and 5 sccm, respectively. However, about 12 nm residual resist still existed. Using the reference, removing the residual resist in traditional lithography, the mixture of Ar and O was used to etch the residual resist process. Fig. 7 shows that the nanostructured resist was also etched, similar to the results in Fig. 5.
The results of Figs. 5, 7 and 8 show that the nanostructured resist was etched if O acts as an etching gas when ICP power ranges from 400 w to 700 w.
These etched results show that the best etching condition (Fig. 9) is when the etching gases are a mixture of Ar (20 sccm) and Cl (5 sccm) and the etching time is 60 s. Thus, the etching rate and selectivity of the resist (AMONIL-MMS4) can be calculated by comparing the depth of the etched structures and the thickness of residual resist.
IV. CONCLUSION
This work introduces an optimized UV-SCIL technology for AMONIL-MMS4 imprint resist. The spinning speed, exposure time and etching time of the imprint resist were investigated, and it was demonstrated that this imprint resist could well serve as an etching mask for dry etching processes without oxygen plasma. Moreover, the best etching and spinning coating conditions were obtained. Therefore, the optimized etching condition and SCIL technology can be used to transfer nanostructure on glass, quartz and silicon wafer for fabricating a metal nanostructure for use in nanoplasmonic biosensors. This technology could reduce the cost of fabrication and meet the need for the mass-production of localized surface plasmon resonance biosensors.
ACKNOWLEDGMENTS
The project was supported by funds from the following: the Major State Basic Research Development Programme of China (973 Programme: 2011CB933102 and 2011CB933203), the National Natural Science Foundation of China (Key Programme: 61335010), the National Natural Science Foundation of China (61378058), the Science and Technology Research Funding of the State Cultural Relics Bureau (20110135) and the 985 Project and the Independent Scientific Research of MUC (10301-01402904).