SPIE Startup Challenge 2015 Founding Partner - JENOPTIK Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:

SPIE Photonics West 2017 | Register Today

SPIE Defense + Commercial Sensing 2017 | Call for Papers

Get Down (loaded) - SPIE Journals OPEN ACCESS


Print PageEmail PageView PDF

Micro/Nano Lithography

Self-aligned quadruple patterning to meet requirements for fins with high density

Repeated plasma deposition and etching steps enable the patterning of fins with the potential to meet requirements of N7 and N5 technologies for profile, depth, uniformity, and pitch walk.
14 May 2016, SPIE Newsroom. DOI: 10.1117/2.1201604.006378

Over recent decades, continuous reductions in the scale of field-effect transistors in accordance with Moore's law, which states that the number of transistors in an integrated circuit doubles every two years, have enabled continuous increases in device performance and transistor density.1–3 Currently, state-of-the-art devices are based on structural elements with dimensions of 7nm or even 5nm (N7/N5). The highest-resolution patterns required for N7/N5 devices are silicon fins with a pitch of 18–28nm and metal layers with a pitch of 24–32nm. These dimensions far exceed the resolution attainable with 193 immersion (193i) lithography. Extreme UV lithography might be an alternative process for the formation of lines and spaces, but is expensive and not entirely ready for use in production.4

Purchase SPIE Field Guide to Optical LithographyTo overcome the limitations of lithography, multiple patterning methods—litho-etch or self-aligned multiple patterning—were used in the last four stages of device miniaturization based on nodes of 10–28nm (N10–N28).5, 6 To achieve the specifications for fins in N7/N5 devices, we need a self-aligned quadruple patterning (SAQP) method that provides a critical dimension (CD) of about 7nm, a CD uniformity (CDU) and pitch walk of 0.5nm (3 sigma), and a line width roughness (LWR) and line edge roughness (LER) of 1.4 and 1.2nm, respectively.

We have developed a low-cost SAQP method that has the potential to meet these requirements for fins. We started with a 193i lithography pattern with a pitch of 90nm and lines and spaces of 40 and 50nm, respectively, which we transferred onto a mandrel. Then we deposited silicon dioxide (SiO2) spacers by atomic layer deposition (ALD) and etched them by a fluorine-based plasma. We removed the mandrel using an oxygen plasma. After cleaning the pattern with dilute hydrofluoric acid, we deposited a second set of SiO2 spacers by ALD and etched them by a fluorine-based plasma. This time, we removed the mandrel with a chlorine-based plasma. We transferred the quadruple pattern into a silicon nitride pad and bulk crystalline silicon. The plasma-enhanced ALD spacers were deposited by ASM. Figure 1 shows a simulation of the patterning process produced using Coventor software.

Figure 1. Simulation images of the stages of self-aligned quadruple patterning (SAQP) obtained using Coventor SEMulator3D software show, from left to right: patterning of the first core (brown) onto a mandrel (green); deposition of silicon dioxide (SiO2) (light blue) by atomic layer deposition (ALD); etching of the first spacers; etching of the mandrel to produce the second core; further deposition of SiO2by ALD; and etching of the second spacers and silicon nitride pad (dark blue). The scale bars represent 30nm.

In this SAQP method, each patterning step gives a CDU value in the sub-nanometer range (3 sigma). As well as controlling the CDU, we developed plasma etching processes that generate vertical profiles, paying special attention to the mandrels. Figure 2 shows transmission electron microscopy (TEM) cross-sectional images of each SAQP step until the nitride pad etching step. After we completed the fin patterning, our preliminary results were as follows: a fin top CD of 7nm (measured by TEM); a fin CDU of 0.6nm (3 sigma); an average pitch walk of 0.5nm with a 3 sigma of 0.8nm; and a fin height of 115nm with a straighter profile on the top 50nm (see Figure 3).

Figure 2. Transmission electron microscopy (TEM) images of the stages of SAQP show, from left to right: patterning of the first core onto a mandrel; deposition of SiO2 by ALD; etching of the first spacers; etching of the mandrel to produce the second core; further deposition of SiO2by ALD; and etching of the second spacers and silicon nitride pad.

Figure 3. From left to right: top view of pattern obtained by critical dimension scanning electron microscopy (CDSEM) after SAQP, including etching and wet cleaning by hydrofluoric acid (HF); contour map and analysis of the pitch walk over a complete wafer based on CDSEM data; and TEM images of fins after etching and wet cleaning by HF. The sample preparation for TEM generates voids that distort the structure of the fins.

In contrast to the sub-nanometer fin CDU and pitch walk, the exacting specifications for LER and LWR were more challenging to attain. After lithography, we carried out a roughness analysis of 131 images across a wafer of 300mm. The average values of LWR and LER were 4.2 and 2.9nm, respectively, which confirms that the line edges were uncorrelated (see Figure 4, left). After SAQP, a roughness analysis revealed that the LWR was about 1.2nm and the LER about 2.2nm, which suggests that the line edges were correlated. After etching of fins, analysis of power spectral density (PSD) showed that the LER was lower than the LWR at high frequency (10−1 to 10−1.7nm−1). However, the major contribution to the high LER is found at middle and low frequencies (10−1.7 to 10−3.5nm−1), which suggests that the lines became wavy at low frequencies.

Figure 4. Left: Analysis of power spectral density (PSD) after lithography shows that the line edge roughness (LER) was lower than the line width roughness (LWR). Right: Analysis of PSD after SAQP shows that the LER was lower than the LWR in the high-frequency region and higher in the middle- and low-frequency regions.

Note that one of the most important factors in achieving high performance in SAQP is the reliance on measurement techniques, which need to have sufficient resolution to determine the roughness of small features. We have optimized the settings of CD scanning electron microscopy, such as averaging, field of view, and pixel size, for the determination of LER and LWR.7 We have also demonstrated the interpretation and feed-forward of data from scatterometry to improve the control of pitch walk.8

The performance of SAQP that we have achieved is quite close to the specifications of the technology. However, the control of certain parameters needs to be improved, such as the control of LWR and LER at the photoresist level and during the etching steps. We have identified key methods that promise to enable improvements in LER (soft-plasma treatment of the photoresist), pitch walk (optimizing plasma-enhanced ALD), and CDU and profile (optimizing the etching process). Simulations using Coventor software will be used to test our assumptions in order to attain the target specifications.

Efraín Altamirano-Sánchez, Zheng Tao, Anil Gunay-Demirkol, Gian Lorusso, Toby Hopf, Jean-Luc Everaert
Leuven, Belgium
William Clark
Paris, France
Vassilios Constantoudis
Institute of Microelectronics
National Center of Scientific Research ‘Demokritos’
Paraskevi, Greece
Daniel Sobieski, Fung Suong Ou, David Hellin
Lam Research
Leuven, Belgium

1. M. Bohr, The evolution of scaling from the homogeneous era to the heterogeneous era, IEEE IEDM Tech. Digest , p. 1-6, 2011.
2. C. Auth, C. Allen, A. Blattner, D. Bergstrom, M. Brazier, M. Bost, M. Buehler, et al., A 22nm high performance and low-power CMOS technology featuring fully-depleted tri-gate transistors, self-aligned contacts, and high density MIM capacitors, IEEE VLSI Tech. Digest, p. 131-132, 2012.
3. S. Natarajan, M. Agostinelli, S. Akbar, M. Bost, A. Bowonder, V. Chikarmane, S. Chouksey, et al., A 14nm logic technology featuring 2nd generation FinFET transistors, air-gapped interconnects, self-aligned double patterning, and a 0.0588μm2 SRAM cell size, IEEE IEDM Tech. Digest, p. 71-73, 2014.
4. K. Xu, L. Souriau, D. Hellin, J. Versluijs, P. Wong, D. Vangoidsenhoven, N. Vandenbroeck, et al., Key contributors for improvement of line width roughness, line edge roughness, and critical dimension uniformity: 15 nm half-pitch patterning with extreme ultraviolet and self-aligned double patterning, J. Micro/Nanolith. MEMS MOEMS 12, p. 041302, 2013.
5. S. Yamauchi, A. Hara, K. Oyama, S. Natori, M. Yamato, H. Yaegashi, Extendibility of self-aligned type multiple patterning for further scaling, Proc. SPIE 8682, p. 86821D, 2013. doi:10.1117/12.2011953
6. Y. Chen, Q. Cheng, W. Kang, Technological merits, process complexity, and cost analysis of self-aligned multiple patterning, Proc. SPIE 8326, p. 832620, 2012. doi:10.1117/12.916490
7. G. F. Lorusso, O. Inoue, T. Ohashi, E. Altamirano-Sánchez, V. Constantoudis, S. Koshihara, Line width roughness accuracy analysis during pattern transfer in self-aligned quadruple patterning process, Proc. SPIE 9778, p. 97780V, 2016. doi:10.1117/12.2218863
8. A. Gunay-Demirkol, E. Altamirano-Sánchez, S. Heraud, S. Godny, A.-L. Charley, P. Leray, R. Urenski, et al., Innovative scatterometry approach for self-aligned quadruple patterning (SAQP) process control, Proc. SPIE 9778, p. 977807, 2016. doi:10.1117/12.2220287