Open Access
Add to BibSonomyAbstract
Diffractive mask-aligner lithography allows printing structures that have a sub-micrometer resolution by using non-contact mode. For such a purpose, masks are often designed to operate with monochromatic linearly polarized light, which is obtained by placing a spectral filter and a polarizer in the beam path. We propose here a mask design that includes a wire-grid polarizer (WGP) on the top side of a photo-mask and a diffractive element on the bottom one to print a 350 nm period grating by using a classical mask-aligner in proximity exposure mode. Linearly polarizing locally an unpolarized incident beam is only possible by using a WGP on the top side of the mask. This configuration opens the possibility to use different linear polarization orientation on a single mask and allows to print high resolution structures with different orientation within one exposure.
© 2015 Optical Society of America
1. Introduction
Sub-micrometer period diffraction gratings have a number of important applications in optics and photonics, such as pulse compression gratings [1], spectroscopy in the EUV and soft X-ray spectral ranges [2], enhancement of the solar cell efficiency [3], filter based on plasmonic resonance [4], and polarizers for visible [5] or UV light [6] also known as Wire Grid Polarizer (WGP). The two main methods used to generate such structures are by interference and e-beam lithography. The first one is limited in terms of structure geometry and the second is time consuming. Mask-aligner lithography is a widely used technique for the mass production of micro and nano-structured components due to its relatively fast processing time, its flexibility in term of printable geometrical structures and its cost efficiency, specially if non-contact mode is used between mask and wafer. High-pressure mercury lamps are the typical light sources that equip most mask-aligners. Those sources emit unpolarized light with a discrete spectrum, with lines down to the UV range. Recently, the use of a binary phase mask in a mask aligner has permitted to print sub-micron structures (250 nm period grating) in proximity mode by using the i-line wavelength (λ = 365nm) of the mercury lamp spectrum [7]. The printing of structures of dimensions close to the theoretical limits is possible by using the relatively long depth of focus of the high resolution interferogram (periodic intensity distribution) generated by the overlapping of the two diffracted orders generated by the mask. For the printing of such pattern, the diffractive mask used is designed to work with monochromatic and linearly polarized light, which is obtained by placing a spectral filter and a polarizer in the beam path. This approach has been already implemented to transfer sub-micrometer structures by proximity exposures with normal [8] and oblique illumination [9] and are often referred as near field holography (NFH). The purpose of this article is to demonstrate that a double-sided mask can linearly polarize an incoming light beam by including a wire grid polarizer on the top side of the mask, while using the bottom side to generate a sub-micrometer interference pattern to write a 350 nm period grating in proximity mode by placing an appropriate diffractive structure there. The possibility to locally linearly polarize an unpolarized incident beam opens the way for printing more complex nano-structures (i.e. with different grating lines orientation) within one exposure step.
2. Mask design
2.1 Phase mask design
The diffractive element (a binary phase-mask) that generates the high resolution intensity distribution is located on the bottom side of the mask. The transfer of the diffraction grating during the lithography step is performed by using the two-beam interference lithography principle. The periodic sinusoidal intensity distribution is generated by the overlapping of the + 1st and −1st transmitted diffraction order generated by a diffraction grating used as a phase-mask. Illuminated in normal incidence, a diffraction grating transmits 3 diffraction orders (−1st, 0th and + 1st) if the period p of the phase-mask satisfies the relation:
In the Eq. (1), λ represents the wavelength of the light source in the transmission medium. By canceling the intensity of the 0th transmitted order to a fraction of ± 1st ones, the resulting interferogram behind the mask is the result of a two-beam interference, which generates a sinusoidal intensity distribution with a period that is half of the mask one. For our application where the transmission medium is air, a period of 700 nm for the phase-mask has been chosen, which satisfies the relation (1) for a wavelength of 365 nm. The choice of this period that is close to the upper limit supports a comparable simple phase-mask’s fabrication [10]. If the phase-mask period goes closer to the wavelength, a high refractive index layer must be used [7,11,12], and thus add steps in the mask fabrication. The angle θ ± 1 of the + and – 1st transmitted order regarding to the normal of the grating is given by the grating equation, and simplified (due to the normal incidence illumination) to the resultant relation:
The period Λ of a sinusoidal intensity distribution resulting of two beams interfering with an angle 2θ between them is related to the latter by the relation:
In the relation (3), n represents the refractive index of the medium (air for our application), which lead to a final interferogram period that is half of the mask one in this configuration.
As introduced above, the high-resolution interference pattern is generated only if the sole ± 1st orders are transmitted, which means that the intensity of the 0th transmitted order must be reduced to a fraction of the ± 1st. This can be performed by choosing the appropriate combination of grating ridge width (expressed in duty cycle of the period) and grating depth. The computation of the optical efficiencies versus the parameter listed above is performed by using a RCWA (Rigorous Coupled Wave Analysis). The intensity versus the grating depth and the grating ridge is calculated for a monochromatic (λ = 365nm) and TE polarized beam which illuminates at normal incidence the grating from the fused silica substrate (n = 1.475) to air. According to the etching recipe developed, an angle of the 1° of the grating ridge is included in the calculation to be as close as possible to fabricated structure. The results are displayed as color map in the left part of the Fig. 1, and show that the +/−1st transmitted order efficiency (η ± 1, sub-window in the upper-left part) can reach a value of 45% of the incident beam intensity and the 0th order (η0, sub-window in the middle-left part) can have an efficiency below 0.1%. For an easier and better understanding and to choose the best parameters, the extinction ratio η ± 1/η0 is displayed in lower-left part of Fig. 1.
Fig. 1 Result of the modeling of the diffraction grating efficiencies. Top left: ± 1st transmitted order efficiency vs grating depth and duty cycle. Middle left: 0th transmitted order efficiency vs grating depth and duty cycle. Bottom left: extinction ratio η ± 1st/η0th vs grating depth and duty cycle [log scale]. Right: Optical intensity distribution under the phase-mask for an optimized phase-mask, the right part shows the different refractive indexes as color tones. The sub-windows displays this distribution at an arbitrary plane to show the contrast.
The extinction ratio is an important parameter in grating modeling for phase-mask fabrication, because it is directly related to the quality of the interferogram, i.e. its intensity contrast. Simply, strong 0th order intensity would give rise to an extra sinusoidal modulation with a period identical with the one of the phase mask, and regularly distributed in the interferogram, which would have concluded in fine, in the printing of a grating with a period equal to that of the phase mask. The mapping of the extinction ratio according the grating depth and the width permits to find the optimum parameters (depth = 430 nm and duty cycle = 224/700 = 0,32 at the bottom of the grooves), and also give the tolerance in the grating fabrication. For the best configuration (red area in the bottom-left part of Fig. 1), the expected diffraction intensities for the ± 1st and 0th orders are 43.2% et 0.004% respectively of the incident intensity, which give rise to a perfect 350 nm periodic intensity distribution behind the phase mask (Fig. 1-right).
2.2 Polarizer design
A wire gird polarizer (WGP) is a nano-optical device consisting of a high aspect ratio metallic grating. The period of this grating is well below the wavelength of the incident light to suppress any propagating diffraction orders except the 0th one. Therefore, fabrication is challenging, especially in the UV wavelength range but WGPs have many beneficial properties such as large acceptance angle and work over a large spectral band. It is, so far the optimal component to efficiently obtain linear polarized light with a wavelength in the UV range and allows integration into a mask. In some particular applications, a local variation of the direction of the polarizer may be required; which cannot be achieved by conventional solutions.
Classically aluminum is utilized as material for wire grid polarizers [13] due to its theoretically outstanding optical performance. However, under exposure to air, heat, and UV light aluminum forms a thick oxide layer [14] degrading the optical performance of the polarizer. To ensure durability in harsh production environment, iridium is utilized as an alternative grating material [6,15].
The elements are basically characterized by the transmittances of light depending on the orientation of the electric field vector to the ridges (Fig. 2). Light polarized parallel to the ridges (TE - transverse electric) is mostly blocked or reflected while light polarized orthogonal to the ridges (TM - transverse magnetic) is mostly transmitted. The suppression of TE polarized light is described by the extinction ratio, the quotient of the transmittances of TM and TE polarized light ER = ηTM/ηTE. The design of such element puts several problems; first, transmittance and extinction ratio are contradictory and a compromise must be found. Second, optical design and fabrication processes are strongly interlinked [6,15,16]. Here, a period of 100nm which is well below the operating wavelength is chosen. The optical performance is simulated by means of RCWA algorithm and shown in Fig. 2. At the targeted wavelength of 365 nm, a ridge width of 22 nm and a height of 150 nm predict a transmittance of TM polarized light of 80% and an extinction ratio of 120.
Fig. 2 Left: Schematic view of the wire grid polarizer. The arrows show the electric field vector direction for TM (red) and TE (green) polarization. Right: Simulated transmittance of TM polarized light and extinction ratio for an iridium wire gird polarizer with a period of 100 nm a ridge width of 22 nm and a height of 150 nm
3. Mask fabrication
3.1 Phase-mask fabrication
The phase-mask is the first element to be fabricated because it is a monolithic device, and it is more resistant and easier to protect in succeeding process steps than the wire grid polarizer. The bottom side of a six inch mask blank is coated by a chromium layer that will be used as hard etching mask and by an e-beam sensitive resist. An electron beam exposure (Vistec SB350 OS) is performed to transfer the grating pattern into the resist. After development, the uncovered chromium mask is etched leaving openings on the fused silica surface, and the latter are etched by dry etching until the optimized grating depth is reached, i.e. 430 nm. Before the remaining chromium and resist material is stripped away the exact profile of the phase-mask is imaged by means of Scanning Electron Microscope coupled with a Focused Ion Beam (SEM-FIB) Fig. 3.
Fig. 3 SEM-FIB micrograph of the phase-mask’s profile. The grating structure has been locally over-coated by a thin layer of platinum for the purpose of preparation of the FIB-cut.
Finally, the remaining chromium etching mask is removed and the phase-mask is optically characterized. The optical characterization consists in measuring the intensity of the transmitted orders with a laser diode (λ = 375 nm) having an operating wavelength close to i-line. The best phase-mask exhibits an efficiency of +/−1st order of 40.6% of the incident beam intensity while the 0th order efficiency is 1.07%. Those values are close to the theoretical ones calculated from the parameters given by the SEM-FIB micrograph at such wavelength (respectively 40.5% and 1.02%) and allow efficiencies close to the ones expected at 365 nm.
3.2 Polarizer fabrication
Subsequent to the manufacturing of the phase mask, the wire gird polarizer is fabricated on the back side of the substrate. Therefore, a double patterning process is utilized [16]. First, an initial layer stack is created by spin coating of 150 nm polymer, ion beam deposition of 15 nm chromium and spin coating of 100 nm electron beam resist (Fig. 4_a). In the next step the resist is patterned by electron beam lithography utilizing a cell projection method for a structure period of 200 nm. Thereby, a reasonable writing time of 3 h for about 30 cm2 is enabled [17]. Subsequently, the chromium hard mask and the polymer are structured by ion beam etching and chlorine based reactive ion beam etching respectively (Fig. 4_b). Following, 3 nm alumina is deposited by atomic layer deposition (ALD). This thin alumina layer serves as protection of the polymer grating against the subsequent high temperature ALD process. In this process step 22 nm of iridium are deposited (Fig. 4_c). Finally, the material from the horizontal surfaces and the polymer are removed by means of ion beam etching and oxygen based reactive ion beam etching respectively (Fig. 4_d), leading to the desired metal stripe geometry with a period of 100 nm.
Fig. 4 Schematic fabrication of a wire grid polarizer. a) Initial layer stack b) Etched template grating c) Iridium coating c) Final polarizer.
4. Optical characterization of the wire grid polarizer
In order to allow for an independent characterization of the wire grid polarizer, a part of the element is fabricated without a grating on the opposite side as a reference. The measurement of the transmittance of TM polarized light and the extinction ratio versus the wavelength is performed by means of a LAMBDA 950 UV/Vis/NIR Spectrophotometer.
Figure 5 shows the measured wavelength dependent optical performance of the reference polarizer. The dashed line is located at the working wavelength of 365 nm, the sub-table displays the transmittance of TM and TE polarized light as well as the extinction ratio. The transmittance of TM polarized light of about 75% is close to the predicted value of 80%, while the achieved extinction ratio of about 74 is below the predicted value (see Fig. 2). The origin of this deviation is a discrepancy between geometry and complex refractive index of the fabricated structure compared to the idealized modelling [16]. However, the achieved suppression of undesired TE polarized light is sufficient to not degrade the contrast of the intensity distribution by adding an offset. Furthermore, the high overall transmittance allows a short exposure time.
Fig. 5 Transmittance of the TM polarized light and TM/TE Extinction Ratio (ER) [dashed] versus wavelength of the reference polarizer. The vertical dotted line marks the operating wavelength and the table summarized the values at this wavelength.
The last characterization step consists in mapping the intensity of the 0th and ± 1st transmitted order under TE and TM illumination over the mask surface to give information on the uniformity of the performances over the mask. The results are presented in the form of intensity color maps in the cases of TM polarized incident beam (Fig. 6), and represent the optical performances at 375 nm of the completed mask i.e. a binary phase mask on the bottom side and the WGP on the top side.
Fig. 6 Maps of the mask showing the transmission through the wire grid polarizer and the phase-mask in percentage of the incident intensity versus the position of the mask. Left: for 0th order with TM polarized light. Right: for ± 1st order with TM polarized light.
The transmission efficiencies are measured for each transmitted diffracted order after a linearly polarized incident beam has passed through the WGP and the phase mask. Our demonstrator is composed of seven elements (WGP + Phase mask; square area of 20*20 mm2) plus an extra smaller field comprising only a WGP for an independent optical characterization of this sole element. Figure 6-left represents the 0th order transmitted intensity, the laser source and the detector are placed face to face and normally to the surfaces of the mask. Figure 6-right is the result of the averaged transmitted intensity when the detector is placed to measure the intensity of + and - 1st transmitted orders. For both cases the source is TM polarized, i.e. grating lines of the WGP are oriented perpendicular compared to the E-field vector direction. This configuration corresponds to the optimal case because the WGP transmits the polarization perpendicular to its grating line direction, so the grating lines of the phase mask placed at the bottom part are oriented perpendicularly to the WGP ones. The label embedded shows the transmission efficiency for the sole WGP and is in agreement with the characterization performed with the spectrometer and discussed above: more than 76% of the incident TM intensity passed through the WGP. Concerning the quality of the test fields, excepted one for whom the fabrication has clearly failed, their optical performances (0th transmitted order below 1% and ± 1st over 30% of the incident intensity) are good, and uniform. The origin of the deficient field found after the optical characterization comes from the phase-mask side. It can be explained by a strong variation of the duty cycle of the phase-mask. This strong 0th order efficiency can come from a local e-beam resist or hard mask thickness variation or eventually from an unexpected e-beam current variation during the writing phase, which change the dose and then the duty cycle of the written grating. The small efficiency variations visible for the ± 1st orders transmission can be explained, due to their linear variation through the mask, by an un-uniform coating process (Cr) or etching step (polymer, Cr or Ir) if the sample is not located exactly in the middle of the coating or etching chamber.
5. Experimental results
The mask-aligner (MA8 Gen3 from Süss MicroTec) is set to satisfy the modeling condition: an i-line filter (band pass filter centered on 367 nm +/− 10 nm from the company Schott) is placed in the beam path. The illumination setup (referred as MO Exposure Optics and fully described in [18], Voelkel et al.) has to be adapted, and an aperture plate (a 6 mm wide slit) is placed before the second Köhler integrator. The slit is used to restrict the original set of illumination angles of ± 2.9° in the mask plane to 0.26° and satisfy as much as possible the 0° incident illumination angle. The choice of the slit width is a compromise between minimizing the set of angles close to 0° (by decreasing the width) and maximizing the light intensity in the mask plane (by increasing the width). A description of this approach and a detailed description of the way to choose the optimal slit width can be found in [7], Bourgin et al..
Once the illumination of the mask-aligner is set, a 4 inch silicon wafer coated with a photoresist layer (AZ nLof 2070) on a layer of Bottom Anti-Reflection Coating (AZ barli II) is placed at a distance of 10µm under the mask. Once the exposure performed (exposure time: 20s), the sample is dipped into the developer (AZ MIF 726) to reveal the structure seen in Fig. 7.
The transfer of the photoresist grating lines into the BARC layer and then into the silicon substrate has been performed by means of Reactive Ion Etching (RIE). A Focused Ion Beam (FIB)-SEM micrograph allows to shows the etched grating profile (Fig. 8).
Fig. 8 SEM micrograph of a 350 nm period grating printed by the double sided mask transferred into the silicon wafer by RIE. The Sub window shows the grating profile obtained by FIB.
Figure 7 and 8 show that a 350 m period grating can be successfully transferred by the double sided mask into a photoresist layer and into a silicon substrate. Nevertheless, the Fig. 7 shows that the resist grating shows a significant line-edge roughness. It can be explained by the fact that the period and the critical dimensions of the transferred structures are at the limit of the resist resolution. The Fig. 8 shows that 143 nm line width have been successfully transferred into the silicon substrate, which is well below the resolution expected by most i-line photoresists. Some improvements in resist, lithographic steps, and etching process are nevertheless needed to suppress the transfer of the line edge roughness into the silicon visible in Fig. 8 as well as the roughness at the bottom of the etched grating. The sub-window included in the Fig. 8 shows a slight variation in the duty cycle for every second transferred grating ridge. This can come from an over-modulation generated by a too strong 0th transmitted order, if the grating parameters of the mask differ from the optimum ones. The 0th transmitted order intensity, measured just below 1% is still too much and must be further reduced by spending more effort on the mask fabrication. To a lesser extent, adjustments in the lithographic exposures conditions can also improve the quality of the final structure.
6. Conclusion
The mask design presented here has shown that it is possible to obtain sufficiently linear polarized light from an unpolarized UV incident beam by including a wire grid polarizer on the upper side of a photo-mask, and generate a high resolution intensity distribution pattern under the latter by using a phase-mask on its bottom side. This approach is often referred as Source-Mask Optimization method. The transfer of the intensity distribution into the photoresist layer, and into the substrate by etching, shows that such an approach can be used for the mass production of high resolution structures, using proximity lithography and a relatively small modification needed to the illumination setup due to the integration of the polarizer directly on the mask.
The polarizer fabrication is still a very difficult task. The WGP and the phase-mask performance can strongly decrease if the fabricated element specifications differ from the modeled one (Fig. 6). The homogeneity across the mask surface has to be improved to transfer homogeneous structures on a large surface. The improvement of the lithographic fabrication tools, as well as the design of the polarizer will help to improve the masks quality. Another source for the degradation of the mask efficiency is the back to front alignment of phase mask and the polarizer. In the case of, only one grating direction has been used, a lateral shift of the structure according to their optimal position does not degrade the efficiency of the mask. It has been measured to be in the range of 100 µm in one direction and 400 µm in the perpendicular one. The orientation between the direction of the phase-mask lines and the polarizer lines, which must be 90° is very sensitive. A tilt will allow the propagation of polarization that is not suited for the phase-mask, in other words, the phase-mask will be illuminated with some residual TM polarization. TM polarization is not suitable for phase-mask lithography because the mask is not well optimized for this polarization and the electric field vectors of the ± 1st orders are not collinear and thus not fully interfere. A constant intensity will be added to the interferogram generated which decrease the contrast of the latter. The tilt measured between the orientation of the grating lines from there optimal value for the double-sided mask presented here has been measured to 0,3°. No significant degradation of the interference pattern has been observed, because of the small value and the non-linearity of the resist, which compensate such a minor misalignment. It is possible to improve such values by adding some alignment marks if a more critical design is needed.
Including the polarizer on the mask itself not only eases the implementation of the high resolution mask in any conventional mask-aligner for the fabrication of sub-micron structures, but opens the possibility to use different polarization directions on one mask, which is not possible if a polarizer is placed in the beam path. By using different polarizer/phase-mask orientation, some more complicated pattern can be fabricated like sub-micrometer period radial and circular gratings or holograms for e.g. counterfeiting purposes. Because the whole structure to be printed is present on one mask, only one exposure step without any lateral substrate-to-wafer alignment is needed where multiple exposure (and different masks) and further lateral alignment steps would be needed with classical shadow printing based mask-aligner lithography.
Acknowledgments
The work presented here is supported by the German Ministry of Science and Education in the frame of the ultra-optics project “Fertigungstechnologien für hoch entwickelte Mikro- und Nanooptiken” (FZK: 03Z1HN32). Financial support by the German Research Foundation DFG, Emmy Noether program SZ 253/1-1 is also acknowledged. The authors want also to thank the colleagues of the CMN group at the Fraunhofer institute IOF and Institute of Applied Physics Jena involved in this work.
References and links
1. D. Strickland and G. Mourou, “Compression of amplified chirped optical pulses,” Opt. Commun. 56(3), 219–221 (1985). [CrossRef]
2. M. Ahn, R. K. Heilmann, and M. L. Schattenburg, “Fabrication of 200 nm period blazed transmission gratings on silicon-on-insulator wafers,” J. Vac. Sci. Technol. B 26, 2179 (2008).
3. Y. M. Song, J. S. Yu, and Y. T. Lee, “Antireflective submicrometer gratings on thin-film silicon solar cells for light-absorption enhancement,” Opt. Lett. 35(3), 276–278 (2010). [CrossRef] [PubMed]
4. Y. Ye, Y. Zhou, and L. Chen, “Color filter based on a two-dimensional submicrometer metal grating,” Appl. Opt. 48(27), 5035–5039 (2009). [PubMed]
5. L. Chen, J. J. Wang, F. Walters, X. Deng, M. Buonanno, S. Tai, and X. Liu, “Large flexible nanowire grid visible polarizer made by nanoimprint lithography,” Appl. Phys. Lett. 90(6), 063111 (2007). [CrossRef]
6. T. Weber, T. Käsebier, E.-B. Kley, and A. Tünnermann, “Broadband iridium wire grid polarizer for UV applications,” Opt. Lett. 36(4), 445–447 (2011). [CrossRef] [PubMed]
7. Y. Bourgin, T. Käsebier, and U. D. Zeitner, “250 nm period grating transferred by proximity i-line mask-aligner lithography,” Opt. Lett. 39(6), 1665–1668 (2014). [CrossRef] [PubMed]
8. T. J. Suleski, B. Baggett, W. F. Delaney, C. Koehler, and E. G. Johnson, “Fabrication of high-spatial-frequency gratings through computer-generated near-field holography,” Opt. Lett. 24(9), 602–604 (1999). [CrossRef] [PubMed]
9. D. M. Tennant, K. Feder, K. F. Dreyer, R. P. Gnall, T. L. Koch, U. Koren, B. I. Miller, and M. G. Young, “Phase grating masks for photonic integrated circuits fabricated by e-beam writing and dry etching: Challenges to commercial applications,” Microelectron. Eng. 27(1-4), 427–434 (1995). [CrossRef]
10. E. Gamet, A. V. Tishchenko, and O. Parriaux, “Cancellation of the zeroth order in a phase mask by mode interplay in a high index contrast binary grating,” Appl. Opt. 46(27), 6719–6726 (2007). [CrossRef] [PubMed]
11. P. Laakkonen, M. Kuittinen, and J. Turunen, “Coated phase masks for proximity printing of Bragg gratings,” Opt. Commun. 192(3-6), 153–159 (2001). [CrossRef]
12. Y. Bourgin, Y. Jourlin, O. Parriaux, A. Talneau, S. Tonchev, C. Veillas, P. Karvinen, N. Passilly, A. R. Md Zain, R. M. De La Rue, J. Van Erps, and D. Troadec, “100 nm period grating by high-index phase-mask immersion lithography,” Opt. Express 18(10), 10557–10566 (2010). [CrossRef] [PubMed]
13. J. J. Wang, F. Walters, X. Liu, P. Sciortino, and X. Deng, “High-performance, large area, deep ultraviolet to infrared polarizers based on 40 nm line/78 nm space nanowire grids,” Appl. Phys. Lett. 90(6), 061104 (2007). [CrossRef]
14. N. Cabrera and N. F. Mott, “Theory of the oxidation of metals,” Rep. Prog. Phys. 12(1), 163–184 (1949). [CrossRef]
15. T. Weber, T. Käsebier, A. Szeghalmi, M. Knez, E.-B. Kley, and A. Tünnermann, “Iridium wire grid polarizer fabricated using atomic layer deposition,” Nanoscale Res. Lett. 6(1), 558 (2011). [CrossRef] [PubMed]
16. T. Siefke, D. Lehr, T. Weber, D. Voigt, E.-B. Kley, and A. Tünnermann, “Fabrication influences on deep-ultraviolet tungsten wire grid polarizers manufactured by double patterning,” Opt. Lett. 39(22), 6434–6437 (2014). [CrossRef] [PubMed]
17. E.-B. Kley, H. Schmidt, U. Zeitner, M. Banasch, and B. Schnabel, “Enhanced e-beam pattern writing for nano-optics based on character projection,” Proc. SPIE 8352, 83520M (2012).
18. R. Voelkel, U. Vogler, A. Bich, P. Pernet, K. J. Weible, M. Hornung, R. Zoberbier, E. Cullmann, L. Stuerzebecher, T. Harzendorf, and U. D. Zeitner, “Advanced mask aligner lithography: new illumination system,” Opt. Express 18(20), 20968–20978 (2010). [CrossRef] [PubMed]
