EG2605 Undergraduate Research Opportunities Program Large Scale Nano Fabrication via Proton Lithography Using Metallic Stencils Tan Chuan Fu 1, Jeroen Anton van Kan 2, Pattabiraman Santhana Raman 2, Yao Yong 2 1 Engineering Science Programme, National University of Singapore 2Centre for Ion Beam Applications, Department of Physics, National University of Singapore, 19260 Singapore, Singapore ABSTRACT In this project, the development and fabrication process of apertures in a Ni stencil for proton beam will be investigated. The apertures serve to focus a proton beam to different beam width to preserve a high beam phase space density of the beam (Saminathan.S. 2011). This is to increase the resulting beam intensity which is desired for lithography applications. The nickel stencil of the aperture structure is created using UV lithography on positive AR-p photoresist, following by Ni plating. Simulation of stopping range of the proton beam is done with SRIM (Stopping and Range of Ions in Matter) to determine the minimum thickness of the stencil. 1. Introduction Proton beam writing poses several advantages over traditional electron lithography process. It allows for deeper penetration in materials and travel in an almost straight path. This feature allows the fabrication of three-dimensional, high aspect ratio structures with vertical, smooth sidewalls and low line-edge roughness (J.A. van Kan et al, 2012). This project involves the creation and optimisation of apertures in nickel stencil mask to improve the effectiveness of proton beam lithography. Prior to the fabrication, a UV lithography mask is created for the layout and design of the nickel stencil. In determining the minimum thickness of the stencil, the stopping range of the protons in the nickel is simulated. The first step of the fabrication process involves magnetron sputtering of Cu and Cr seed layers on a silicon substrate. Spin coated with AR-p 3210 positive photoresist, the layout of the structure can be finely created using UV lithography. After development, electroplating of nickel is done to create the stencil. The final product is then obtained by etching the Cu seed layer.
2. Design and Fabrication of UV mask Firstly, a mask for the UV lithography is created using a μpg 405nm laser writer and an AZ1518 chrome mask. AZ 1518 Cr mask is a 3 inch soda lime transparent to i-line (365nm) wave length light with a pre coated thin layer of Cr and a 500nm thick layer of positive photoresist AZ 1518. Adobe Photoshop was used to create the design which will be projected onto the resist through UV exposure. For the design writing process with the laser writer, 80% of 30mW power was used which is optimised to give the highest resolution. This is followed by the development of the photoresist by immersing it in the diluted AZ developer (1 : 4 ratio with DI water) for 1 minute in order to remove the exposed AZ 1518 resist. Next, the mask was rinsed with DI water thoroughly and dried using an air gun. Then the mask was immersed in a Cr etchant for 2 minutes to remove the Cr layer in the exposed area and then again rinsed with DI water and dried with air gun. Finally, the mask was left to immerse in acetone to remove the remaining resist, rinsed with DI water and dried with air gun before it was ready to be used. 2.2. Aperture Design for UV mask Fig. 1 Schematics of a stencil design Fig. 2 Design for the UV mask w1 (um) w2 (um) w3 (um) 1 st row 10 25 50 2 nd row 25 50 100 3 rd row 50 100 250 4 th row 100 250 500 5 th row 250 500 1000 6 th row 10 100 1000 Table 1 Aperture sizes (um) corresponding to the respective rows of the UV mask design in Figure 2 Figure.1 shows the array of rectangular stencil pattern, containing different aperture sizes, written onto the soda lime mask coated with thin layers of Cr and AZ 1518 resist. Each nickel stencil contains three different aperture sizes of increasing radius. The rows correspond to a different series of aperture sizes as shown in table 1. The two columns are a repeat pattern of each other.
The rectangular array of aperture pattern is designed for positive photoresist. The white area of the design appears as transparent on the Cr mask and the black area represents the Cr coating on the mask which prevents the UV light from passing through. 3. Simulation Proton in nickel Minimum thickness of the nickel mask was determined using SRIM, with the range of the photon beam energy varying from 2keV to 2.5MeV. From the simulated results, 20um thick nickel will able to stop the trajectory of 2.25MeV proton. 25 Projected Range (um) 20 15 10 5 0 0 0.5 1 1.5 2 2.5 Ion Energy (MeV) Fig. 3 Graph of Projected Range (um) against Energy of Proton (MeV) in nickel There are other factors that need to be taken into consideration in determining the thickness of the stencil. Generally, a thicker stencil will be favoured, as it will be more rigid and easier to handle. In addition, stress on the nickel from plating imperfections and etching process will not induce as much deflection on the stencil itself, making it flatter and thus more desirable as a beam limiting aperture. On the other hand, there is a limit on the aspect ratio of the pillars on the photoresist. A thick stencil mask would require the pillars to have high aspect ratio between its height and the diameter of the pillar. This will increase the tendency of collapse of the pillars in during the developing process. Thus the choice of aperture size and photoresist used need to be considered in determining the thickness of the mask as well. 4. Fabrication The fabrication process of the stencils involves several key steps, listed in figure 3. Firstly, the substrate was prepared by sputtering of Cr and Cu seed layers onto a piece of silicon wafer. This was followed by spin coating of AR-p positive photoresist and UV lithography using the photo mask which has been made. The exposed AR-p is then developed by immersing in AR300-26 solution. The resulting substrate was then electroplated with nickel, followed by removing of the residual photoresist. The final stencil was obtained by etching away the Cu seed layer.
1) Magnetron sputtering of seed layers Cr and Cu seed layer Si substrate 2) Spin coating of AR-p photoresist AR-p Cr and Cu seed layer Si substrate 3) UV lithography Cr UV mask 4) Development of AR-p 5) Electroplating of nickel 6) Removal of residual AR-p Nickel layer 7) Etching of Cu seed layer Fig. 4 Process of fabrication of free standing stencil mask 4.1 Magnetron sputtering of seed layer Cr (5nm) and Cu(5nm) seed layers are sputtered on a silicon wafer.the Cr layer was added to enhance the adhesion of the Cu layer to the Si substrate. The Cu layer provides the conductivity required for the electroplating. Moreover, the Cu layer can also be easily etched using sodium persulfate without damaging the electroplated Ni layer. The thickness of the Cu seed layer can be varied, keeping in mind that a thicker Cu seed layer will increase the roughness of the surface of the mask but decrease the duration of the etching process.
4.2 Spin coating of AR-P photoresist layer. A positive photoresist AR-P 3210 was used as it able to provide good resolution in thick resist layer up to 40um due to its high viscosity, allowing it to generate vertical profiles with high edge steepness (Allresist n.d.). The photoresist was spin coated on substrate using a three step process. Low starting spin speed (30 s) is followed by a main spin speed of 900-1700 rpm, for 2 minutes 15 seconds. Marginal beads are reduced by a final spin rotation at 800 rpm for 5 seconds. The thickness of the spin coated layer with different main spin speed is shown in Table 2 and Fig. 4. Main spin speed (rpm) Thickness of spin coated layer (um) 900 23 1100 20 1300 18 1500 16 1700 15 Table 2 Thickness of spin coated AR-p layer with different main spin speed Thickness of layer (um) 26 24 22 20 18 16 14 12 10 700 1200 1700 2200 Main spin speed (rpm) Fig. 5 Graph of thickness of spin coated AR-p layer (um) against main spin speed (rpm) The resulting spin coat layer was prebaked at 95 o C for 10 minutes to reduce the remaining solvent content. This is to improve the resist adhesion to the substrate and avoid mask contamination and/or sticking to the mask during UV lithography (MicroChemical, 2010). 4.3 UV lithography With the photoresist layer, UV lithography was then used with the made photo mask to create the aperture array pattern on the photoresist. The exposed AR-p was then developed with AR300-26 developer diluted in 1:3 ratio for 1min followed by a DI water rinse and dried with an air gun. There is a need to minimise the developing duration to prevent the pillar from deform or collapse. Post-development baking of 110 C improves adhesion and resistance of the structure during the electroplating and developing process (MicroChemical, 2010). Aperature sizes of 100um, 250um and 500um corresponding to the 4 th row of the photomask were made and the sizes in each step were measured with a micro-ruler and compared. Fig. 6.1 Optical 98um diameter AR-p pillar Fig. 6.2 Optical 248um diameter AR-p pillar Fig. 6.3 Optical 497um diameter AR-p pillar
Figure 5 shows three of the AR-p pillars of 23um high created in this project. The 100um pillar gives the lowest aspect ratio of 4.35 which is relatively safe from collapse. 4.4 Electroplating of Nickel and Removal of residual photoresist With the AR-p pillars, nickel plating can then be used to form the core of the metallic stencil. A 20um Ni layer was electroplated using rate of 0.5um/min with 1.18A for 48 min. The rate of electroplating should be kept low to maintain proper adhesion between the nickel layer to silicon substrate. The residual AR-p was then removed by acetone in a ultrasonic bath for 1 min. Fig. 7.1 Optical 99um diameter aperture in nickel on Si substrate 4.5 Etching of Cu seed layer Fig. 7.2 Optical 250um diameter aperture in nickel on Si substrate Fig. 7.3 Optical 498um diameter aperture in nickel on Si substrate The last step of the fabrication process is the etching of Cu seed layer to lift off the nickel stencil from the substrate. This is done immersing it in a solution of 60 grams of sodium persulfate mixed with 40ml dl water. The solution is kept at 45 0 C until the nickel stencil has been separated from the silicon substrate. The nickel stencil was then rinsed with dl water and dried using an air gun. Fig. 8.1 Optical 99um diameter aperture in free standing nickel stencil Fig. 8.2 Optical 250um diameter aperture in free standing nickel stencil Fig. 8.3 Optical 498um diameter aperture in free standing nickel stencil
5. Proton Beam Lithography The fabricated free standing nickel stencil mask was used for proton beam lithography on a 200nm PMMA coated silicon substrate. The lithography was conducted with 500 kev protons from 45 0 beamline of the Proton Beam Writing facility in Centre for Ion Beam Application. After 10 minutes of exposure, the PMMA was developed with 7:1 ratio of isopropyl alcohol and deionised water and dried with N 2 air gun. Fig. 9.1 Optical 99um diameter aperture in PMMA Fig. 9.2 Optical 250um diameter aperture in PMMA Fig. 9.3 Optical 498um diameter aperture in PMMA By comparing the dimension of the apertures of the nickel mask and PMMA, it can be seen that proton beam lithography provides a highly coherent pattern transfer. The unexposed regions were well shielded from the proton beam by the 20um thick nickel stencil as simulated in SRIM. 6. Conclusion The design and fabrication of the Ni stencils of varying aperture sizes has been studied and conducted. The beam limiting aperture will serve its function to increase the intensity of the proton beam for improvements in proton beam lithography application. In addition to the points that have been mentioned, there can be much more development in the optimisation and characterisation of the quality of the end stencil mask. For example, stress from uneven electroplating process and the side wall verticality of the aperture can be studied and improved.
References Saminathan. S. (2011). Theory of ion beam transport. (Doctoral dissertation). Retrieved from University of Groningen. (Accession Order No. 978-90-367-5192-6) J.A. van Kan, P.G. Shao, P. Molter, M. Saumer, A.A. Bettiol, T. Osipowicz, F. Watt, (2005) Fabrication of a free standing resolution standard for focusing MeV ion beams to sub 30 nm dimensions, Nucl Instrum Meth B, 231 (2005) 170 175. Yao Yong, M. W. van Mourik, P. S. Raman, J. A. van Kan (Nov 2012). Improved beam spot measurements in the 2nd generation proton beam writing system Nuclear Instruments and Methods in Physics Research Section B, 306 (2013) 265-270 J. A. van Kan, P. Malar, and Armin Baysic de Vera, The second generation Singapore high resolution proton beam writing facility, Review of Scientific Instruments 83, 02B902 (2012) Allresist (n.d.) AR-P3200. Retrieved from http://www.allresist.de/wenglish/produkte/photoresists/0020.php MicroChemical (2010), Exposure of photoresist. Retrieved from http://www.microchemicals.eu/technical_information/exposure_photoresist.pdf MicroChemical (2010), Baking steps in photoresist processing. Retrieved from http://www.microchemicals.eu/technical_information/photoresist_baking.pdf