Journal of Materials Science: Materials in Electronics

, Volume 18, Issue 12, pp 1191–1195

Nano patterning of cone dots and nano constrictions of negative e-beam resist for single electron transistor fabrication

Authors

    • School of Microelectronic EngineeringNorthern Malaysia University College of Engineering (KUKUM)
  • U. Hashim
    • School of Microelectronic EngineeringNorthern Malaysia University College of Engineering (KUKUM)
  • Z. A. Z. Jamal
    • School of Microelectronic EngineeringNorthern Malaysia University College of Engineering (KUKUM)
Article

DOI: 10.1007/s10854-007-9318-1

Cite this article as:
Madnasri, S.B., Hashim, U. & Jamal, Z.A.Z. J Mater Sci: Mater Electron (2007) 18: 1191. doi:10.1007/s10854-007-9318-1

Abstract

We present an optimization of nano dot of negative tone e-beam resist which is a very important step in single electron transistor fabrication process. The optimum design of dot and nano constriction plays a significant role in determining optimum etching resolution and single electron transistor performance. In this research, we have optimized nano dot and nano constriction dimensions of resist by controlling some parameters, such as e-beam dose, spin speed, pre-bake time and image development time. However, a nano constriction design variety of 120–200 nm in width was carried out to reach the optimum design. In this paper, the fabrication process of cone nano dots using e-beam lithography with considering proximity effect is reported. As nano constriction design decreased, cone nano dot changed to pyramid nano dot and the compression effect on the dot also significantly increased as well.

1 Introduction

Electron beam lithography (EBL) has been expected as a candidate to form very fine pit or dot array for patterned media [1]. The ultimate resolution of EBL is not set by the resolution of electron-optical system and electron scattering, but by the resist resolution, the subsequent process and a combination of (a) the de-localization of exposure process as determined by the range of the Coulomb interaction between electrons and resist molecules, (b) the straggling of secondary electrons into resist, (c) the molecular structure of resist and (d) the molecular dynamics of development process [2]. There are several limitations of the resist material and process which recently have been developed for nano mask fabrication, such as resolution, sensitivity or line edge roughness [3, 4]. For that reason, the continuously development of electron-beam resist systems used for nano mask fabrication has been still focused on improving sensitivity, resolution, and etching resistance of resist materials [2].

In this research, nano dot for single electron transistor (SET) fabrication be positioned between two nano constrictions which each has design width narrower than dot diameter and expected applicable for the etch process. Nano constrictions of such device were designed by calculating proximity effect of EBL and the optimum resolution of resist [5]. Therefore, for tunnel junction formation, these nano constrictions need precisely resist patterning technique. In this study, the nano dot design diameter of 50 nm was adopted of the minimum area step size of e-beam equipment and their nano constriction designs range from 120 to 200 nm. In both electron and optical lithographic tools, the line width depends on the focus plane for the beam which is normally adjusted on the substrate holder [6]. Theoretically, a SET necessitates nano constriction width that is smaller than quantum dot diameter, because nano constriction is required as electron transport barrier for SET operation [7, 8]. The design width of nano constriction depends on the needs of the successive fabrication process either linked to source and drain mask or spaced of both. For this structure type, the uniformity and reproducibility of the dot dimension and the periodicity of the whole area are crucial parameters. Therewith, to achieve the wanted SET dimension, high resolution of resist material and process must be combined with high dimensional control of the feature shape, orientation and spacing [9]. Because the proximity effect becomes a serious problem, its correction factor is generally implemented by using exposure dose adjustment, pattern shape adjustment, equalization of background dose, or multilayer techniques [8].

2 Experimental set up

The nano dot and nano constriction arrays in this study were fabricated using \( \langle {\text{100}}\rangle \)p-type substrates, with diameter of approximately 525 μm and resistivity of 1–10 ohm-cm. All the samples made in square 15 × 15 mm form were to ease in adjusting the stage position when focusing SEM image so that the design pattern on the developed resist was easily observable. To remove organic and inorganic contaminants, the wafer was first cleaned using standard CRA cleaning and dried. The dried substrates were then heated up to 200 °C for 30 min using conduction hotplate JB-TEK Honeywall and cooled down to room temperature. Next, negative tone ma-N 2403 resists were spun on the silicon substrates using spinner model WS-400B-GNPP/UTE/10 K that was set up on spin speed 3,000 rpm for 30 s. The deposited resist was pre-baked on the hot plate JB-TEK Honeywall at temperature 90 °C for 120 s to improve the film adhesion on the substrates [10]. Subsequently, the uniform resist was cooled until room temperature and then, resist film was currently ready to be exposed using electron beam lithography system.

The masks of source, dot and drain were designed using offline GDSII Editor software (ELPHY PLUS version 4.0) by considering proximity effect, optimum resolution of resist and e-beam equipment. In addition, the optimum resolution of ma-N 2403 negative resist is 50 nm and minimum area step size of version 4.0 Raith software of EBL is 40 nm, however, nano dot and nano constriction designs were devised in the range of this equipment resolution. The nano constriction width variety was defined to optimize design width that was available for further processes in SET fabrication by considering proximity effect. As a preliminary result of this work, the resist pattern width of SEM micrograph image was observed wider of about 80 nm than that of GDSII design. By considering that result, the 80 nm nearly width deviation was used as a calculation basic to agree on the nano constriction design width.

Afterwards, the resist coated substrate was patterned using SEM-based EBL. Typical conditions for this exposure use accelerating voltage of 20 kV, spot size of 45, field size of 200 μm, microscope magnification of 450× and e-beam dose of 190 μAs/cm2. To define a dot, a higher dose factor, which was usually set up more than one in scale, was necessarily arranged to increase the adhesion of exposed resist of the dot. After e-beam exposure, the exposed resist was developed in ma-D 532 solution for 35 s and dipped in DI-water for 5 min to stop image development. The developed nanostructures were characterized by using scanning electron microscopy (JEOL SEM 6460) and atomic force microscopy (AFM SPI Probe Station 3800N, SPA Sound Proof Housing).

3 Results and discussions

The possibility of single-dot patterning in the negative e-beam resist ma-N 2403 with a minimal diameter 134 nm is demonstrated. A set of two nano constrictions that each separates dot with source and drain resist mask are subsequently described.

In Fig. 1a we found that nano constriction that breaks up source with dot is 228 nm in width, while another one is 244 nm in width. It means that the cone dots have been already successfully fabricated as revealed in Figs. 1b and c. The surface profile of AFM image as we see in Fig. 1b shows that cone nano-dots are higher than source or drain surfaces and indeed, those surfaces are not smooth.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-007-9318-1/MediaObjects/10854_2007_9318_Fig1_HTML.gif
Fig. 1

(a) SEM micrograph of masks of source, dot and drain with two nano constrictions. (b) AFM image of resist pattern of one cone dot. (c) AFM image of two cone dot arrays

Figure 1b shows some cone dots of AFM characterization. In this figure, nano-dots seem lower than sources and drains, even those surfaces of sources and drains seem flatter than those of Fig. 1c. The dots are very regular and the reproducibility is good over the whole desired area [9]. If we compare Figs. 1b and c then we found height difference between dots compared with source and drain mask. This result is accused by non uniformities of resist thickness and e-beam dose. The top resist profile area of AFM image is narrower than the bottom part such that this led to side area of SEM image appears lighter.

In Fig. 2a, a circle dot lies between source and drain mask that each separated by nano constriction. The source and drain masks were defined in dimension 500 × 500 nm and the dot diameter design was 50 nm. The widths of left and right nano constriction of the dot are 70 and 94 nm respectively. Meanwhile, a specific dot diameter in Fig. 2a is 168 nm. Figure 2c shows an SEM micrograph of a dot-array pattern fabricated using EBL; uniformly shaped 120-nm-wide dots were fabricated. As nano constriction design width of negative tone resist decreases, the cone dot of negative tone resist changes to pyramid dot and the compression effect on the dot increases as well. The compression effect of the source and drain on the dot led to the change of the dot dimension from round into ellipse. In addition, the longer ellipse dot radius perpendicular to the compression direction is observable in Fig. 2b. The positions of source and drain edges relatively do not change although have the width difference of nano constriction. In SEM micrograph image, the edge roughness of the pattern was not as sharp as the design.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-007-9318-1/MediaObjects/10854_2007_9318_Fig2_HTML.jpg
Fig. 2

SEM micrograph of nano constriction design of (a) 200 nm (b) 180 nm (c) 160 nm (d) 140 nm (e) 120 nm in width

The possibility of single-dot patterning in the negative e-beam ma-N 2403 resist with a minimal diameter 130 nm is demonstrated. Figure 3a shows two arrays of cone nano dot and nano space that come into sight very narrow. As nano constriction design decreasing from 200 to 120 nm, cone nano dots array changes to pyramid dots array as well, and even changes to rectangular dots array as observed in Fig. 4. The peak heights of cone nano dots are within 10 and 200 nm range. Unlikely, as compared to Figs. 3a, b, c, d and e are better uniformity and roughness of resist thickness.
https://static-content.springer.com/image/art%3A10.1007%2Fs10854-007-9318-1/MediaObjects/10854_2007_9318_Fig3_HTML.gif
Fig. 3

AFM micograph of nano constriction design of (a) 200 nm and very sharp cone dot lazes in the center between source and drain mask that separated by two nano constrictions (b) 180 nm (c) 160 nm (d) 140 nm (e) 120 nm in width

https://static-content.springer.com/image/art%3A10.1007%2Fs10854-007-9318-1/MediaObjects/10854_2007_9318_Fig4_HTML.gif
Fig. 4

AFM micrograph of resist pattern for source, dot and drain

One of the major factors limiting the fabrication of high-density structures by EBL is proximity effect, caused by forward scattered electrons in the volume of the resist and backscattered electrons from the substrate [2]. The other difficulties are the higher-density structures will require the longer image development time and we frequently also found the case like the faster removal of single dot compared with source and drain resist mask. It is clear that, an optimum high-density structure is considered necessary for controlling the image development time.

4 Conclusion

In conclusion, we have optimized resist patterns of cone nano dot and nano constriction for SET fabrication by using EBL. By adjusting the exposure conditions, we successfully fabricated a resist pattern of source, dot and drain in dimension as following: source mask of 548 nm in width, dot diameter of 134 nm and drain mask width of 582 nm. The dot shape change from cone to pyramid with decreasing the nano constriction width has significantly increased the compression effect on the dot.

Acknowledgements

The authors would like to acknowledge the support of IRPA Project, Malaysian Ministry of Science, Technology and Innovation(Grant No. IRPA 09-02-15-0000-SR0013/06-060) and Graduate Assistance (GA) Program of Northern Malaysia University College of Engineering (KUKUM).

Copyright information

© Springer Science+Business Media, LLC 2007