Pattern Roughness Analyses in Advanced Lithography: Power Spectral Density and Autocorrelation

Extreme ultra-violet lithography (EUVL) has been prepared for next generation lithography for several years. We could get sub-22 nm line and space (L/S) pattern using EUVL, but there are still some problems such as roughness, sensitivity, and resolution. According to 2007 ITRS roadmap, line edge roughness (LER) has to be below 1.9 nm to get a 22 nm node, but it is too difficult to control line width roughness (LWR) because line width is determined by not only the post exposure bake (PEB) time, temperature and acid diffusion length, but also the component and size of the resist. A new method is suggested to reduce the roughness. The surface roughness can be smoothed by applying the resist reflow process (RRP) for the developed resist. We made resist profile which has surface roughness by applying exposure, PEB and development process for line and space pattern. The surface roughness is calculated by changing parameters such as the protected ratio of resin. The PEB time is also varied. We compared difference between 1:1 L/S and 1:3 L/S pattern for 22 nm. Developed resist baked above the glass transition temperature will flow and the surface will be smoothed. As a result, LER and LWR will be much smaller after RRP. The result shows that the decreasing ratio of LER due to RRP is larger when initial LER is large. We believe that current ~ 5 nm LWR can be smoothed to ~ 1 nm by using RRP after develop.

The semiconductor community is well aware of the challenges that exist in developing lithographic methods that can pattern features at sub-20 nm periodic feature spacing (pitch, L0). Optical lithography already utilizes complex multiple patterning schemes to overcome diffraction limitations at 193-nm exposure wavelengths, and the delayed insertion of EUV lithography will likely require the use of multiple patterning or other assistive processes to further reduce the achievable feature sizes. An alternative to these techniques employs the directed self-assembly (DSA) of block copolymers. Block copolymers (BCPs) can naturally micro-phase separate into morphologies such as lamellae, cylinders, spheres, and gyroids at length scales down to sub-10 nm dimensions. Using the ability of BCPs to micro-phase separate in conjunction with alignment methods such as graphoepitaxy and chemoepitaxy to produce well-ordered structures, a process referred to as DSA, offers a possible method for producing sub-20 nm features in conjunction with optical patterning processes at greatly reduced cost and complexity. One of the many challenges in implementing line-space type DSA processes is the lack of methods for effective modulation and tuning of the pattern pitch (L0) produced by a given BCP. Previous studies have shown that blending homopolymer into the BCP thin films can allow for tuning of both: (1) L0 to be larger than that provided naturally by the BCP's molecular weight (MW) and (2) the relative size line-space size ratio. However, this tuning ability comes at the expense of increased line edge roughness (LER) and line width roughness (LWR). It has also been shown that either higher or lower MW BCP can be blended into a primary BCP in order to modulate and tune the pattern pitch produced from the BCP mixture, but the effects of this BCP blending on pattern LER and LWR have not been explored or reported in detail. In this study, coarse-grained molecular dynamics simulations of BCP DSA on a chemoepitaxial underlayer were implemented to characterize the impacts that blending controlled amounts of two different MW BCPs together have on DSA pattern LER and LWR. The blends shown here had LER and LWR values as much as 20% higher than those of pure, monodisperse BCPs; however, reducing the MW difference between the 2 BCPs could reduce this effect.

The variation of device characteristics is a challenge to present and future large-scale integrations. One of the origins of the variation is line width roughness (LWR). To facilitate the efforts to cope with LWR, we developed a method to accurately characterize LWR basing on the analysis of power spectral densities (PSDs). Because experimental PSDs are intrinsically discrete, we derive simple analytic formulas of the discrete PSDs by assuming that the autocorrelation function (ACF) exponentially decays with distance. The PSDs calculated by using the formulas agree excellently with experimentally obtained PSDs of photoresist LWR. From the result we find that the photoresist LWR of this study has a standard deviation of 2.5 nm and an exponentially-decaying autocorrelation function with a correlation length of 35 nm. Although the experimental PSDs inevitably contain a component produced by scanning-electron-microscope (SEM)-image noise, the two components of LWR and noise are separately determined by the method of this study. Due to this feature, the resultant variance of LWR is independent of the noise intensity. However, it is still important to reduce the noise component in order to maintain the accuracy of analysis especially in the case of LWR that has an unknown functional form of PSD. This is because even the PSDs that have different functional forms sometimes look alike in the presence of large noise. To reduce the noise effect, it is effective to average the SEM images perpendicularly to fine lines before edge detections. The procedure does not reduce the variance unlike averaging along the lines. The method of this study is applicable not only to LWR but also to other cases as far as the ACF exponentially decays with distance, or equivalently the spectral line shape is Lorentzian. Accordingly, it forms the basis for spectral analysis of most experimental results.

The resist materials are evaluated using their resolution, line edge roughness (LER), and sensitivity. However, the evaluation of chemically amplified resists is tricky because of the trade-off relationships between resolution, LER, and sensitivity. In this study, we investigated a chemically amplified resist with a fullerene matrix by analyzing the dose-pitch matrices of line width and LER. The effective quencher concentration of the fullerene resist obtained by the analysis was higher than those of typical polymer-type resists. This suggests that the quantum efficiency of acid generation in the fullerene matrix is slightly lower than those of polymer-type resists. The effective reaction radius was 0.06 nm, which was smaller than those of polymer-type resists. The proportionality constant between LER and the chemical gradient of the fullerene resist was smaller than those of polymer-type resists, probably owing to its molecular size.

High numerical-aperture (NA) extreme ultraviolet lithography (EUVL) system has been highly desired [1-3] and is now under construction at imec-ASML high-NA laboratory in Veldhoven. However, there are still many challenges to realize high-volume manufacturing (HVM) by high-NA EUVL; Line edge roughness (LER) mitigation is one of the key elements. In our previous research, we studied how normalized image log slope (NILS) and resist film thickness (FT) affect resist LER by exposure at NA 0.33 on NXE:3400 and S-Litho EUV [4] . However, mask absorber/tone, exposure dose and postexposure bake (PEB) temperature were not focused on for decreasing LER. They have the potential to mitigate resist roughness as illumination and resist thickness also have. Unbiased LER (uLER) values of metal oxide resists (MOR) were experimentally measured on half pitch 14 nm line and space (L/S) under several mask absorber/tonality, dose and PEB temperature. Per mask absorber/tone the NILS was varied by using different illumination shapes. Low-n masks exhibited higher NILS for the same illumination shape, which resulted in lower resist LER on wafer. It was also found that mask tone can contribute to LER mitigation. According to detailed investigations about mask roughness, mask error enhancement factor (MEEF) and flare by CD-SEM measurements and S-Litho EUV, the most likely reason for LER difference is flare increase given by mask tone change. From dose and PEB temperature variation study, dose was apparently more dominant to resist roughness than PEB temperature.

The need to characterize line edge and line width roughness in patterns with sub‐50 nm critical dimensions challenges existing platforms based on electron microscopy and optical scatterometry. The development of x‐ray based metrology platforms provides a potential route to characterize a variety of parameters related to line edge roughness by analyzing the diffracted intensity from a periodic array of test patterns. In this study, data from a series of photoresist line/space patterns featuring programmed line width roughness are measured by critical dimension small angle x‐ray scattering (CD‐SAXS). For samples with designed periodic roughness, CD‐SAXS provides the wavelength and amplitude of the periodic roughness through satellite diffraction peaks. For real world applications, the rate of decay of intensity, termed an effective “Debye‐Waller” factor in CD‐SAXS, provides an overall measure of the defects of the patterns. CD‐SAXS data are compared to values obtained from critical dimension scanning electron microscopy (CD‐SEM). Correlations between the techniques exist, however significant differences are observed for the current samples. A tapered cross sectional profile provides a likely explanation for the observed differences between CD‐SEM and CD‐SAXS measurements.

Nanoscale resist topography such as line edge roughness (LER) or line width roughness (LWR) is the most serious concern in sub-100 nm fabrication. Although many factors have been reported to affect LER, the generation mechanism of LER is still unclear. It is well known that the slope of image contrast correlates to the degree of roughness. However, significant LER is sill observed in chemically amplified resists for electron beam (EB) lithography, which can produce a steeper slope of image contrast than photolithography. To make clear a cause of LER, the distribution of protons and counter anions generated in chemically amplified EB resists was investigated. It was found that counter anion distribution is significantly different from proton distribution. Counter anions are inhomogeneously distributed outside a relatively smooth edge of proton distribution. This is caused by the fact that acid generators can react with low energy (~ 0 eV) electrons. The inhomodeneous distribution of counter anions outside proton distribution is considered to contribute to LER formation in chemically amplified resists for EB lithography.

MAPDST (4-(methacryloyloxy)phenyl)dimethylsulfoniumtrifluoromethanesulfonate)) based resist analogous are reported to pattern high-resolution nano features (20 nm) under wide range of lithography tools including electron beam lithography (EBL), extreme ultraviolet lithography (EUVL), 193 nm immersion lithography etc. However, these resists have not yet patterned lower node features, especially at sub-15 nm regime with ultra-low line edge roughness (LER) and line width roughness (LWR). One of the methods to improve the resolution of a resist is the structural modification. Towards this, we have developed two new hybrid copolymer resists i.e MAPDST-co-ADSM and MAPDST-co-TPMA by the copolymerization of radiation sensitive organic MAPDST with hybrid tin monomers ADSM and TPMA (ADSM = acetyldibutylstannyl methacrylate; TPMA = triphenyl tin methacrylate) for high-resolution EBL applications. The developed resists were studied for their sub-15 nm line patterns with low LER and LWR features. Various line features, starting from 30-12 nm with different line/space (L/S to L/10S) characteristics were studied at various e-beam doses 200- 1200 μC/cm2. Isolated 12 nm line features have been achieved with the resist MAPDST-co-ADSM at a dose 1200 μC/cm2. Meanwhile, the MAPDST-co-TPMA resist patterned 15 nm features at the dose 700 μC/cm2. The estimated sensitivity and contrast of resists MAPDST-co-ADSM and MAPDST-co-TPMA were 1.60; 450 μC/cm2 and 1.55; 380 μC/cm2 respectively. Similarly, the computed LER and LWR parameters exhibited by the resists MAPDST-co-ADSM and MAPDST-co-TPMA for sub-30 nm features were 0.99; 1.22 and 1.8; 3 nm respectively. The e-beam studies revealed a resolution enhancement of the hybrid resists at 12 nm regime as compared to the neat poly(MAPDST) resist (where the resist resolution was 20 nm) indicating improvements in the lithography properties of these resists.

We will give an account of our investigation on structure property relationships of amines with regards to line width roughness (LWR) and line edge roughness (LER) of a 193 nm alicyclic-acrylate resist. Specifically, we have looked at basicity, molar volume and logD as factors which may have an influence of roughness of 80 nm 1:1 L/S features. For relatively hydrophobic amines (Log D > -1), the lower the hydrophilicity at acidic pH the greater the LER and LWR becomes. Specifically, in this range of Log D, more hydrophobic larger amines, with higher basicity, tend to give worse L/S feature roughness. For amines which are more hydrophilic, the relationship becomes more complex with some amines giving a lower LER while others do not. This appears to be predicated on a delicate balance between basicity, hydrophilicy and size.

Directed self-assembly (DSA) applying chemical epitaxy is one of the promising lithographic solutions for next generation semiconductor device manufacturing. We introduced Fingerprint Edge Roughness (FER) as an index to evaluate edge roughness of non-guided lamella finger print pattern, and found its correlation with the Line Edge Roughness (LER) of the lines assembled on the chemical guiding patterns. In this work, we have evaluated both FER and LER at each process steps of the LiNe DSA flow utilizing PS-b-PMMA block copolymers (BCP) assembled on chemical template wafers fabricated with Focus Exposure Matrix (FEM). As a result, we found the followings. (1) Line widths and space distances of the DSA patterns slightly differ to each other depending on their relative position against the chemical guide patterns. Appropriate condition that all lines are in the same dimensions exists, but the condition is not always same for the spaces. (2) LER and LWR (Line Width Roughness) of DSA patterns neither depend on width nor LER of the guide patterns. (3) LWR of DSA patterns are proportional to the width roughness of fingerprint pattern. (4) FER is influenced not only by the BCP formulation, but also by its film thickness. We introduced new methods to optimize the BCP formulation and process conditions by using FER measurement and local CD valuation measurement. Publisher’s Note: This paper, originally published on 2 April 2014, was replaced with a corrected/revised version on 14 May 2014. If you downloaded the original PDF but are unable to access the revision, please contact SPIE Digital Library Customer Service for assistance.

Linewidth Roughness (LWR) remains a difficult challenge for improvement in all resist materials. In previous work we focused on the impact of key components of LWR by analyzing the Power Spectral Density (PSD) curves which can be obtained using Fractilia’s MetroLER computational software. [1] By measuring the unbiased PSD (with SEM image noise removed), accurate assessment of PSD(0) (the low-frequency limit of the PSD) and correlation length (the length scale of the transition from white to correlated noise) is possible. We showed there was an important relationship between ArF resist frequency components and LWR through lithographic process (before and after a resist trim step) as a function of resist formulation. In this paper we will study how key frequency components such as PSD(0) and correlation length change as we vary basic resist properties such as diffusion. The impact of aerial image on LWR and its frequency components will also be studied with particular attention to how correlation length affects LWR as feature size decreases. We will also look at the impact of diffusion or resist blur on PSD(0) as a function of aerial image Normalized Image Log-Slope (NILS). Understanding the relationship between PSD(0) and correlation length and how to manipulate these variables to minimize LWR for different features is crucial for more rapid LWR improvement at different nodes. [1] Charlotte Cutler, et al., “Roughness power spectral density as a function of resist parameters and its impact through process,” Proc. SPIE 10587, Optical Microlithography XXXI, 1058707 (23 March 2018).

The need to characterize line edge and line width roughness in patterns with sub-50 nm critical dimension challenges existing platforms based on electron microscopy and optical scatterometry. The development of x-ray based metrology platforms provides a potential route to characterize a variety of parameters related to line edge roughness by analyzing the diffracted intensity from a periodic array of test patterns. In this study, data from a series of photoresist line/space patterns featuring programmed line width roughness measured by critical dimension small angle x-ray scattering (CDSAXS) is presented. For samples with periodic roughness, CD-SAXS provides the wavelength and amplitude of the periodic roughness through satellite diffraction peaks. In addition, the rate of decay of intensity, termed an effective "Debye-Waller" factor, as a function of scattering vector provides a measure of the fluctuation in line volume. CDSAXS data are compared to analogous values obtained from critical dimension scanning electron microscopy (CDSEM). Correlations between the techniques exist, however significant differences are observed for the current samples. Calibrated atomic force microscopy (C-AFM) data reveal large fluctuations in both line height and line width, providing a potential explanation for the observed disparity between CD-SEM and CD-SAXS.

Objective. This study explored the feasibility of detecting hidden muscle activity in surface electromyogram (EMG) baseline. Approach. Power spectral density (PSD) analysis and multi-scale entropy (MSE) analysis were used. Both analyses were applied to computer simulations of surface EMG baseline with the presence (representing activity data) or absence (representing reference data) of hidden muscle activity, as well as surface electrode array EMG baseline recordings of healthy control and amyotrophic lateral sclerosis (ALS) subjects. Main results. Although the simulated reference data and the activity data yielded no distinguishable difference in the time domain, they demonstrated a significant difference in the frequency and signal complexity domains with the PSD and MSE analyses. For a comparison using pooled data, such a difference was also observed when the PSD and MSE analyses were applied to surface electrode array EMG baseline recordings of healthy control and ALS subjects, which demonstrated no distinguishable difference in the time domain. Compared with the PSD analysis, the MSE analysis appeared to be more sensitive for detecting the difference in surface EMG baselines between the two groups. Significance. The findings implied the presence of a hidden muscle activity in surface EMG baseline recordings from the ALS subjects. To promote the presented analysis as a useful diagnostic or investigatory tool, future studies are necessary to assess the pathophysiological nature or origins of the hidden muscle activity, as well as the baseline difference at the individual subject level.

For the first time, process impact on line-edge roughness (LER) and line-width roughness (LWR) in a back-end-of-line (BEOL) self-aligned quadruple patterning (SAQP) flow has been systematically investigated through predictive 3D virtual fabrication. This frequency dependent LER study shows that both deposition and etching effectively reduce high frequency LER, while deposition is much more effective in reducing low frequency LER. Spacer-assisted patterning technology reduces LWR significantly by creating correlated edges, and further LWR improvement can be achieved by optimizing individual process effects on LER. Our study provides a guideline for the understanding and optimization of LER and LWR in advanced technology nodes.

High‐k spacer and gate insulator materials have been exhaustively studied nowadays for the enhancement of electrostatic control and reduction of short‐channel effects in scaled devices. The work presents a high‐performance and charge trap tolerant FinFET module at 10 nm gate length. Dual layer gate insulator (inner low‐k and outer high‐k) introduces to reduce charge trapping from the channel and outside into the gate oxide. It reduces the gate leakage current by 51.6% compared to conventional FinFET. Further, they demonstrate single charge trapping (SCT) induce effects and proposed optimised high‐k spacer width of the SCT tolerant design. SCT analysis is presented in different high‐k spacer materials and back‐gate voltages. Process variation sources such as line edge roughness and line width roughness are also analysed for the circuit design.