By utilizing a reflective plasmonic slab, it is demonstrated numerically and experimentally in this paper deep sub-wavelength imaging lithography for nano characters with about 50nm line width and dense lines with 32nm half pitch resolution (about 1/12 wavelength). Compared with the control experiment without reflective plasmonic slab, resolution and fidelity of imaged resist patterns are remarkably improved especially for isolated nano features. Further numerical simulations show that near field optical proximity corrections help to improve imaging fidelity of two dimensional nano patterns.
© 2013 OSA
The resolution of conventional photolithography is generally restricted by about half of wavelength of illuminating light due to the diffraction limit. This occurs because evanescent waves carrying subwavelength information make no contribution to the far-field imaging process . In recent years, a great number of near-field photolithography methods were developed and demonstrated both theoretically and experimentally, including evanescent near-field optical lithography [2,3], surface plasmon interference nanolithography technique [4,5], superlens [6–8] etc.
However, these near field lithography methods are usually constrained in practical applications due to the poor quality of resist patterns characterized with shallow profiles, low contrast and great aberrations compared with masks. The decaying feature of evanescent waves is believed to be the dominating reason for this problem. In 2005, D. B. Shao, and S. C. Chen presented numerical and experimental demonstrations of surface plasmons (SPs) excitation assisted nanoscale near field photolithography by introducing a Titanium (Ti) shield slab below the photoresist [9, 10]. It showed that evanescent waves amplification in reflection manner by appropriate metal shield help to enhance the imaging performance in the near field . Arnold M. D. and Blaikie R. J. proposed a general theory with rigorous analysis for attaining improved imaging by plasmonic reflective slabs . Xu et al reported a metal-cladding superlens structure for localizing SPs and projecting deep-subwavelength patterns . In 2011, Charles W. Holzwarth et al. reported demonstrative experiments of increased process latitude in absorbance-modulated lithography via a silver plasmonic reflector . Recently, the plasmonic reflector concept is further employed in variant types of hyperlens [14, 15].
In this paper, we present an experimental demonstration of plasmonic slab imaging lithography with evanescent wave amplification in reflection mode. The experiment is performed in a similar lithography configuration of Ref . The difference lies in the enhanced amplification of evanescent waves, improved resolution and fidelity of imaging, and reduced deterioration effect of SP interference by applying appropriate plasmonic reflector design. Two dimensional nano patterns with critical dimensions (about 30nm~50nm) far below the wavelength are well imaged in lithography resist. Additionally, it was found that the fidelity of resist patterns can be further improved with near field optical proximity correction (OPC) method.
2. Design and simulation of reflective plasmonic slabs
Figure 1 presents the schematic of the proposed sub-wavelength imaging lithography structure with a reflective plasmonic slab. Photo resist layer is sandwiched between the chromium mask and the silver plasmonic slab. The illuminating light comes from an i-line mercury lamp exposure system with central wavelength of 365 nm. The thickness of the Cr mask and the photo resist (PR) are 50 nm. The thickness of the silver plasmonic slab is 100nm, greatly larger than the skin depth of light in silver. The relative permittivity for Cr, photo resist, and silver are , (ALLRESIST GMBH, Strausberg), and  at 365nm, respectively. The Cr mask is milled with a transparent pattern of nano characters ‘OPEN’ with 36nm line width. The total length in the y direction of each character is about 500nm. The width of the character ‘O’ is about 420 nm and the other characters are about 320 nm.
Before the design and imaging simulation, it is helpful to calculate and analyze the optical transfer function (OTF) of the structure. For simplicity of analysis, the imaging transfer process is approximated as a virtual light source in photo resist at a distance away from the metal reflector, as shown in the inset of Fig. 2(b). For the imaging plane close to the PR-metal interface, the optical transfer function for the plasmonic slab structure is defined as 
For P-polarized light illumination, the reflection at the PR-metal interface isEq. (1) is further approximated for large () asEq. (3) has a large value when the condition is satisfied and is negiable. Therefore, the evanescent components of large are amplified in the dielectric layer and image quality could be improved.
The above analyses could be well seen in Fig. 2, where the reflection and optical transfer functions for titanium and silver reflector at two different PR thickness d are plotted. We can see that the great enhancement of reflection for silver slab compared with the Ti slab which exhibits large imaginary part of permittivity −2.3667 + 5.6444i . For the OTF curves, both Ti and Ag give amplification of evanescent waves and the latter one is much better for components with larger kx, promising an improved resolution and fidelity of imaging. Also the smaller thickness of photo resist contributes to improve the imaging transfer ability of evanescent waves by enhancing the coupling effect between the light through the mask and the reflector.
Shown in Fig. 3 are the simulated results of imaging nano ‘OPEN’ characters by a reflective silver plasmonic slab as sketched in Fig. 1. The simulation is performed by the finite element method with Comsol Mulitphysics 4.0. Figures 3(a) and 3(b) present the calculated electric field intensity distribution at 30nm below the Cr-PR interface for light illumination with linear polarization in the x and y direction, respectively. In this case, only the lines not in the direction of the polarization are imaged, due to the prohibition of light transmission through metallic slit apertures and absence of SP excitation without proper light polarization. The sum of the electric field distributions of Figs. 3(a) and 3(b) is plotted in Fig. 3(c), as a simulation for light illumination with natural polarization in our experiments. The image closely resembles the pattern of the mask. For comparison, the calculated results without the silver slab structure are shown in Fig. 3(d) as a control case. The image shows great distortions in comparison with that of Fig. 3(c).
Shown in Figs. 3(e) and 3(f) are the light distribution in the y-z plane at the position depicted in Figs. 3(c) and 3(d). And Figs. 3(g) and 3(h) plot the light distributions at variant z position. We can see that the narrow width of the line (36nm) leads to a strong electromagnetic coupling with SP excited at the interface of Ag and photo resist. The image distribution in the photo resist layer with plasmonic slab shows the line width of about 45nm at any thickness of photo resist, which is much more uniform than that without reflector. The intensity contrast of images defined as at 10nm and 30nm below the Cr-PR reach about 0.98 and 0.91, respectively. Even at the bottom of the PR (50nm thickness) the intensity contrast is larger than 0.8. However, for the structure without reflector, the intensity contrasts at 10nm, 30nm, and 50nm below the Cr/PR interface are 0.66, 0.3, and 0.1, respectively, showing a rapid decrease of imaging property. For positive photoresist, the required intensity contrast is usually larger than 0.3 . Therefore, the effective imaging depth for Fig. 3(f) is less than 30nm. And the reflective plasmonic slab structure gives an imaging depth of 50nm and even larger. This effect could be simply attributed to the superposition of the evanescent waves emitted from masks and the enhanced reflected light, as predicted in Fig. 2(b). It is worth to note that the i-line mercury lamp illumination has finite line width and it should be taken into account in the simulation. Figure 4 plots the measured spectrum of the light illumination in our experiment and show a FWHM line width of about 10nm. The real part permittivity of silver varies from −2.2 to −2.8 in this range, while the imaginary part changes slightly around 0.25 . The simulated imaging results show that the broadband of i-line do not deliver apparent deterioration of images, as shown in the insets of Fig. 4. But some background of light appears and reduces the contrast for large wavelength components. So it is possible to improve the imaging lithography performance by using a narrower band filter.
3. Imaging Lithography by a reflective plasmonic slab
To perform the experiment, lithography masks with nano patterns were fabricated. First, 50 nm thick Cr layer was deposited on a mask substrate of fused silica using E-beam evaporation and measured by Inficon quartz crystal microbalance. Nano transparent patterns like ‘OPEN’ and dense lines were then milled on the Cr layer by Focused Ion Beam (FEI Nova 200 NanoLab, @30kV Accelerating Voltage). The plasmonic slab was fabricated in experiment by depositing 110nm thickness Ag film on a fused silica substrate with Magnetron Sputtering RF power 500W, the deposition rate 2.2 nm/s and the chamber pressure ~3.0 × 10−4Pa. Upon it was spun with about 50 nm thick positive photoresist made by 3000rpm and diluted AR-P 3170 (diluted by ALLRESIST GMBH, Strausberg, 40nm@4000rpm, 26% PAC) to record the near-field images. After 2 min of prebake of photo resist at 100°C hotplate, the substrate with reflective slab was physically contacted with the mask with the help of air pressure (~0.3MPa), and then flood-exposed under an i-line (365nm) mercury lamp illuminating system. The exposure with uniform flux of 1.3mW/cm2 and exposure time ranging from 15s to 20s were applied in the experiment. The post-exposure substrate was developed by diluted AR 300-35 (ALLRESIST GMBH, Strausberg) with de-ionized water with the ratio of 1:1 for 18s at room temperature. ALL the SEM images of resist patterns were taken with S-4800 (HITACHI) cold field emission scanning electron microscope at 5kV acceleration voltage. The samples were sputtered with several nanometers Au prior to imaging to reduce charging effects.
Compared with the mask pattern in Fig. 5(a), the recorded resist pattern in Fig. 5(b) shows good fidelity and about 50nm line width by applying the reflective plasmonic slab and appropriate exposure time 16s. With the same photoresist thickness, mask, exposure dose and development time, the control experiment result without reflector in Fig. 5(c) gives a greatly deteriorated ‘OPEN’ resist images with line width larger than 100nm. This demonstrated experimentally the evanescent waves amplification and contribution to the imaging process with the help of reflective plasmonic slab, as illustrated numerically in Fig. 3. In our experiment, it is also found that the reflective plasmonic slab configuration contributes to process latitude increase for variant exposure time, as shown in Fig. 5(d) and the similar effect reported in Ref . Although the line width of resist pattern can be further narrowed to 45nm with shorter exposure time (15s), the line edge roughness (LER) of this pattern is apparent. On the other hand, increase of exposure time would deliver wider line width of resist pattern but not destruct the imaging fidelity. As another proof of the imaging fidelity, Fig. 6(a) presents the AFM measured result of the sample in Fig. 5(b). The measured thickness of the pattern is about 35nm in Fig. 6(b). The thickness decrease is mainly affected by the resist loss in the developing process and the AFM probe’s rounding effect would deliver some error as well.
As predicted by the OTF of Fig. 2(b) in the last section, the imaging transfer ability by the silver plasmonic reflector could be extended for kx ~4k0, corresponding to about 45nm half pitch resolution. Inspired by this point, lithography with dense lines is performed in experiment. Shown in Fig. 7(a) is the reflective plasmonic slab lithography result for dense lines array pattern with about 55nm half pitch. The same resist and exposure, developing conditions as Fig. 5(b) are employed here. We can see that the dense lines could be well distinguished in the imaging lithography. But the quality and fidelity of resist pattern are greatly reduced, compared with that of Fig. 5(b) and the rings circling the dense lines in Fig. 7(a). It is believed that the limited resolution of the diluted ARP-3170 photo resist brings the resist quality deterioration especially for neighboring dense lines. For the un-diluted ARP-3170, it is suggested by ALLRESIST GMBH Company for i-line lithography with half pitch not smaller than 100nm. Imaging with denser lines could be realized by reducing the resist thickness as predicted by Fig. 2(b). This is demonstrated by the result of Fig. 6(c) with 32nm half pitch and about 30nm thickness photo resist (diluted AR-P 3170 by ALLRESIST GMBH, Strausberg, 30nm@4000rpm, 26% PAC). The exposure time is 20s and the other conditions are the same as Fig. 5(b).
4. Near field optical proximity and some discussion for reflective plasmonic slab lithography
In spite of the assistance of plasmonic reflector, some slight aberrations of images are encountered. For instance, the line end shortening occurs in character ‘P’ and ‘N’ and the corner rounding effect in ‘P’, ‘E’, and ‘N’, as observed in Fig. 3(c). These effects could also be found in experiment results of Fig. 5(b). These effects are mainly attributed to the limited resolution of plasmonic slab imaging and could be referred to the near field optical proximity. The simple and direct way to relieve this problem is optical proximity corrected mask design. Figures 8(a) and 8(b) present the simple corrected mask nano ‘OPEN’ pattern and the calculated image light distribution with modified mask. Compared with Fig. 3(c), the line end shortening and the corner rounding effects are improved. It is believed that further complicated and careful masks design  with adjacent grooves, lines etc. would bring considerably improved fidelity of imaging.
One issue about SP interference would arise as we concern the defects associated with SP imaging. For instance, the SP interference ring resist pattern is observed in the plasmonic reflector lithography in Ref . But no apparent interference evidence around the nano pattern images is found in our simulations and experiments. This occurs because that, unlike the Ti reflector in Ref , the excited SP of the thick silver slab exhibits large wave vector and large loss, which prohibits the propagation of SP in the long range and reduces the deleterious effect of interference on the photoresist pattern. As an illustrative calculation, the SP propagation wavevector is (2.39 + 3.89i)k0 for Ag-PR semi-infinite interface and (1.63 + 0.36i)k0 for Ti-PR interface at 365nm wavelength. The permittivity of Ti and Ag are −2.3667 + 5.6444i and −2.4 + 0.2488i. This is of somewhat interesting that silver with smaller imaginary part of permittivity brings strongly localized SP with great propagation loss. On the other hand, the interference of SP plays a great impact role for dense lines imaging. For the same reason, the SP interference is strongly restricted below the dense nano lines pattern, which promises the localized imaging effect instead of interference pattern over an extended region. In addition, we also notify that some interference mode distributions inside images do exist in our simulation, like the apparent light intensity ripples inside the slit pattern images for vertical lines of “P”, “E” and “N” in Figs. 3(a) and 3(c). This effect, we believe, could not be simply identified as the near field optical proximity effect due to the limited resolution, but is the superimposed SP modes interference along the slit direction and localized in the patterned mask-PR-Ag region. Similar discussion is given in Ref . This localized interference behavior could be modified by the method of mask correction. But it seems require much more complicated corrections for a good performance.
In conclusion, plasmonic lithography with reflective silver slab is demonstrated numerically and experimentally in this paper. Two dimensional nano characters resist patterns with about 50nm line width and dense lines with 32nm half pitch are observed in the lithography experiment. Evanescent waves amplification and reflective plasmonic slab were experimentally proved to contribute greatly to the imaging process, with considerably improved resolution, contrast and effective depth of exposure in comparison with control experiment. This work is believed to help solve the low contrast and shallow working depth problems encountered in the near field lithography, optical storage, tweezing optics etc.
This work was supported by 973 Program of China (No. 2011CB301800) and the Chinese Nature Science Grant (61138002, 61177013). The authors thank Dr. Zongwei Xu and Prof. Fengzhou Fang for their kind efforts in fabricating masks.
References and Links
2. R. J. Blaikie, M. M. Alkaisi, S. J. McNab, D. R. S. Cumming, R. Cheung, and D. G. Hasko, “Nanolithography using optical contact exposure in the evanescent near field,” Microelectron. Eng. 46(1-4), 85–88 (1999). [CrossRef]
3. M. M. Alkaisi, R. J. Blaikie, S. J. McNab, R. Cheung, and D. R. S. Cumming, “Sub-diffraction-limited patterning using evanescent near-field optical lithography,” Appl. Phys. Lett. 75(22), 3560–3562 (1999). [CrossRef]
4. X. Luo and T. Ishihara, “Surface plasmon resonant interference nanolithography technique,” Appl. Phys. Lett. 84(23), 4780–4782 (2004). [CrossRef]
9. D. B. Shao and S. C. Chen, “Surface-plasmon-assisted nanoscale photolithography by polarized light,” Appl. Phys. Lett. 86(25), 253107 (2005). [CrossRef]
12. T. Xu, L. Fang, J. Ma, B. Zeng, Y. Liu, J. Cui, C. Wang, Q. Feng, and X. Luo, “Localizing surface plasmons with a metal-cladding superlens for projecting deep-subwavelength patterns,” Appl. Phys. B 97(1), 175–179 (2009). [CrossRef]
13. C. W. Holzwarth, J. E. Foulkes, and R. J. Blaikie, “Increased process latitude in absorbance-modulated lithography via a plasmonic reflector,” Opt. Express 19(18), 17790–17798 (2011). [CrossRef] [PubMed]
14. G. Ren, C. Wang, G. Yi, X. Tao, and X. Luo, “Subwavelength demagnification imaging and lithography using hyperlens with a plasmonic reflector layer,” Plasmonics 8(2), 1065–1072 (2013). [CrossRef]
15. Jianjie Dong, Juan Liu, Xingxing Zhao, and Peng Liu, “A super lens system for demagnification imaging beyond the diffraction limit,” Plasmonics (2013). [CrossRef]
16. P. B. Johnson and R. W. Christy, “Optical constants of the noble metal,” Phys. Rev. B 6(12), 4370–4379 (1972). [CrossRef]
17. V. Intaraprasonk, Z. Yu, and S. Fan, “Image transfer with subwavelength resolution to metal–dielectric interface,” J. Opt. Soc. Am. B 28(5), 1335–1338 (2011). [CrossRef]
18. Marvin J. Weber, Handbook of Optical Materials (CRC Press, 2003).
20. B. B. Zeng, L. Pan, L. Liu, L. Fang, C. Wang, and X. Luo, “Improved near field lithography by surface plasmon resonance in groove-patterned masks,” J. Opt. A, Pure Appl. Opt. 11(12), 125003 (2009). [CrossRef]
21. D. Shao and S. Chen, “Surface plasmon assisted contact scheme nanoscale photolithography,” J. Vac. Sci. Technol. B 26(1), 227–231 (2008). [CrossRef]