This PDF file contains the front matter associated with SPIE Proceedings Volume 9048, including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

As extreme ultraviolet lithography (EUVL) prepares for its insertion into the high-volume manufacturing phase, many challenges still remain to be addressed. Among several issues, development of EUV resists with tight specifications of sensitivity (dose), resolution (HP) and line-edge roughness (LER) is required. Chemically-amplified resists (CARs) have been the major paradigm in the development of EUV resists, although several alternatives, such as molecular resists and inorganic resists, are also under development. Here we present a comparative study of the performance of CARs using the PSI’s EUV interference lithography tool, which can achieve patterning down to 7 nm HP. Also the current status of EUV resist availability towards 11 nm HP technology nodes is discussed. We show resolution down to 12 nm HP with CARs. Nevertheless, for patterning below 18 nm HP, the resolution is achieved at the expanse of sensitivity and LER. The global trend of decreasing sensitivity with increasing LER is valid across the different resists. This trade-off between resolution, LER, and sensitivity (i.e. RLS trade-off) is mainly dominated by the acid diffusion blur and remains a challenge. In addition, pattern collapse becomes a significant problem with increasing resolution. This can be partly overcome by the reducing the resist thickness, which leads to an increase in LER. Therefore, a new trade-off between pattern-collapse limited resolution and LER emerges. These two trade-offs make the progress in EUV resist development increasingly difficult.

Recently, both PSI¹ and ASML² illustrated champion EUVL resolution using slow, non-chemically amplified inorganic resists. However, the requirements for EUVL manufacturing require simultaneous delivery of high resolution, good sensitivity, and low line edge/width roughness (LER/LWR) on commercial grade hardware. As a result, we believe that new classes of materials should be explored and understood. This paper focuses on our efforts to assess metal oxide based nanoparticles as novel EUV resists³. Various spectroscopic techniques were used to probe the patterning mechanism of these materials. EUV exposure data is presented to investigate the feasibility of employing inorganic materials as viable EUV resists.

The suppression of stochastic effects such as line edge roughness (LER) and stochastic defect generation is required for the realization of high volume production of semiconductor devices with 11 nm critical dimension using extreme ultraviolet (EUV) lithography. In this study, the effect of the molecular weight and protection ratio of the resist polymer on LER and stochastic defect generation was investigated. By increasing the molecular weight, LER and the probability for the stochastic defect generation were decreased. The negative effect caused by the increase of molecular weight is the increase of the minimum block size for dissolution in the development process. By increasing the protection ratio, the similar effect was expected. However, LER and the probability for the stochastic defect generation were not significantly decreased by increasing the protection ratio because of the decrease of quantum efficiency of acid generation caused by the protection of proton sources. The development of protecting group that does not affect the quantum efficiency is essential to decrease the stochastic effects.

We have shown that the dissolution properties can be successfully modified to improve the line/space profile and LWR of a low diffusion EUV CA resist. The surface roughness is a function of hot spots in the nominally unexposed regions of the resist material. We conjecture that the photoacid hot spots are formed due to DC flare present in the optical train of the exposure system. We also have shown that the PAGs can be further improved for out-of-band radiation (OOB) response. The improvement can be as much as 557% for 193nm exposure, and 838% by 248nm exposure. The improved OOB response leads to better contact hole performance. We also shared our continued improvement in resist witness plate performance with the majority of our resists passing for carbon growth, and all samples passing for non-cleanables. There does appear to be a site-to-site bias which we attribute to differences between e-beam and EUV exposure and/or substrate working distance from the source. Lastly, we show outstanding lithographic process window for 24 nm contact arrays on an NXE 3300 stepper as well as 15 nm half pitch lines and spaces on the PSI interferometric tool.

Recently in the R&D area DRAM has shrunk to 1X nm, at the same time patterning technology has been one of the major challenges on 1X nm DRAM. Less than 20nm line and space and less than 30nm contact hole patterning are basically needed for 1X nm DRAM. Currently ArF immersion extension such as DPT (Double Patterning Technology), SADP (Self-Aligned Double Patterning) and SAQP (Self-Aligned Quadruple Patterning) shows robust patterning performance relative to EUV/DSA and become established process as a baseline for 1Xnm DRAM. But cost of ownership and process complexity of DPT/SADP can be the big burden for volume production. Furthermore too many DPT/SADP can make DRAM shrink meaningless. In spite of current issues on DPT/SADP, EUV source power has been the most critical issue so far. And now source power issue is pushing development of high sensitive EUV resist and related process. In this paper, author will compare EUV PTD and NTD in view of image contrast and swelling, also evaluate patterning performance between EUV PTD and NTD, and finally describe current status and issues of EUV NTD.

EUV stochastic effects are generally studied [1-7] from the point of view of Line-Edge-Roughness (LER), Line-Width-Roughness (LWR), Local CDU (LCDU), i.e. from the point of view of the CD-Variability of the printed patterns. In this paper we will look at what happens when this variability gets worse and turns into (random) locally failing patterns, such as missing contacts or locally closed trenches. In doing so, these failures contribute to setting the effective printability limit for the experimental conditions and the type of structures that are being considered. We will first discuss the way in which we have tried to quantify the amount of locally failing structures, i.e. the metrics we have adopted to describe them. Next we describe how the amount of locally failing structures depends on some pattern-related and experimental condition-related parameters. From this we concluded that – as in the case of local variability – the amount of local failure depends on the exposure dose used (the well-known photon shot noise effects) as well as on mechanisms that originate in the resist or the process. Although we have not been able to identify what this resist- or materials component is exactly, we do have some indications that point to the development step, but we will also discuss other potential contributors.

Experimental local CD uniformity (LCDU) of the dense contact-hole (CH) array pattern is statistically decomposed into stochastic noise, mask component, and metrology factor. Each component are compared quantitatively, and traced after etching to find how much improvement can be achieved by smoothing. Etch CDU gain factor is defined as the differential of etch CD by resist CD, and used to estimate etch CDU on resist CDU. Stochastic noise has influenced on not only LCDU but also local placement error (LPE) of each contact-hole. This LPE is also decomposed into its constituents in the same statistical way. As a result, stochastic noise is found to be the most dominant factor on LCDU and LPE. Etch LCDU is well expected by Etch Gain factor, but LPE seems to be kept same after etching. Fingerprints are derived from the repeating component and the boundary size for excluding proximity effect in analysis is investigated.

Laser produced plasma (LPP) light sources have been developed as the primary approach for EUV scanner imaging of circuit features in sub-20nm devices in high volume manufacturing (HVM). This paper provides a review of development progress and readiness status for the LPP extreme-ultra-violet (EUV) source. We present the latest performance results from second generation sources, including Prepulse operation for high power, collector protection for long lifetime and low cost of ownership, and dose stability for high yield. Increased EUV power is provided by a more powerful drive laser and the use of Prepulse operation for higher conversion efficiciency. Advanced automation and controls have been developed to provide the power and energy stability performance required during production fab operation. We will also discuss lifetesting of the collector in Prepulse mode and show the ability of the debris mitigation systems to keep the collector multi-layer coating free from damage and maintain high reflectivity.

Since 2002, we have been developing a CO₂-Sn-LPP EUV light source, the most promising solution as the 13.5 nm high power (>200 W) light source for HVM EUV lithography. Because of its high efficiency, power scalability and spatial freedom around plasma. Our group has proposed several unique original technologies; 1) CO₂ laser driven Sn plasma generation, 2) Double laser pulse shooting for higher Sn ionization rate and higher CE. 3) Sn debris mitigation with a magnetic field, 4) Hybrid CO₂ laser system that is scalable with a combination of a short pulse oscillator and commercial cw-CO₂ amplifiers. 5) High efficient out of band light reduction with grating structured C1 mirror. In past paper we demonstrated in small size (2Hz) experimental device, this experiment shoed the advantage of combining a laser beam at a wavelength of the CO₂ laser system with Sn plasma to achieve high CE>4.7% (in maximum) from driver laser pulse energy to EUV in-band energy ¹). In this paper we report the further updated results from last paper. (1) 20um droplets at 100kHz operation was successfully ejected by downsized nozzle and demonstrated dramatical improvement of debris on the collector mirror. We have been developing extension of high CE operation condition at 20kHz range, We have reported component technology progress of EUV light source system. (2)New generation collector mirror with IR reduction technology is equipped in mirror maker. (3)20kW CO₂ laser amplifier system is demonstrated cooperate with Mitsubishi electric. (4) We develop new Proto #2 EUV LPP source system and demonstrated 200W EUV plasma power (43W EUV clean power at I/F ) at 100kHz operation was confirmed. (5) High conversion efficiency (CE) of 3.9% at 20kHz operation was confirmed in using pico-second pre-pulse laser. (6)Improvement of CO₂ laser power from 8kW to 12kW is now on going by installation of new pre-amplifier. (7)Power-up scenario of HVM source is reported, target shipment of first customer beta LPP light source unit is 2015.

In this paper, we provide a detailed review of development of a highly-efficient high-power nanosecond pulse CO₂ laser using transverse-flow radio-frequency (RF)-exited laser amplifiers, which is for extreme ultraviolet (EUV) light source of a next generation of nano-lithography. High-density excited high-power transverse-flow CO₂ lasers were designed and built for the application of laser produced plasma (LPP) EUV source. We carried out an amplification test of the transverse-flow CO₂ laser seeded by a nanosecond pulse CO₂ laser. A four-amplifier system generated an average output power of 21 kW with an electrical input power of 400 kW for discharges. The electrical-to-optical efficiency was 5.2%. The input pulse laser had an average power of 46 W, the repetition rate was 100 kHz, and the pulse duration was 15 ns. The transverse-flow CO₂ laser has strong points in high gain and availability of the multi-fold optical path. A highly-efficient amplification was experimentally proved characterized by transverse-flow CO₂ laser amplifiers even with low-power seed. A transverse-flow CO₂ laser is a promising candidate for an amplifier in the LPP EUV light source.

Plasma based radiation sources optimized to emit 13.5 nm Extreme UV radiation also produce a significant amount of light at longer wavelengths. This so called out-of-band (OoB) radiation is detrimental for the imaging capabilities of an EUV lithographic imaging system, particularly the ultraviolet (UV) parts of the light (λ=100-400 nm). To suppress these wavelengths while maintaining the high efficiency of the mirror for EUV light, several methods have been developed, including phase-shift gratings (PsG) and anti-reflection layers (SPE layer). Both methods have achieved a suppression factor of 10 - 30 around the target wavelength. To achieve a full band suppression effect with a minimum loss of EUV light, a new scheme based on surface pyramid structures was developed. An average suppression of more than 10 times was achieved with a relative EUV efficiency of 82.2% by using the Si pyramids structure (compared to a flat multilayer (ML)). Recently, we have successfully produced a pyramid structure consisting of multilayers which greatly improves the relative EUV efficiency to 94.2%.

For full commercialization, extreme ultraviolet lithography (EUVL) technology requires the availability of EUV mask blanks that are free of defects. This remains one of the main impediments to the implementation of EUV at the 22 nm node and beyond. Consensus is building that a few small defects can be mitigated during mask patterning, but defects over 100 nm (SiO2 equivalent) in size are considered potential “killer” defects or defects large enough that the mask blank would not be usable. The current defect performance of the ion beam sputter deposition (IBD) tool will be discussed and the progress achieved to date in the reduction of large size defects will be summarized, including a description of the main sources of defects and their composition.

EUV mask blank substrates will be subject to extraordinarily demanding requirements upon flatness, smoothness and absence of residual defects. To date, no combination of available surface preparation techniques has been able to produce essentially perfect substrates with zero residual defect populations. A critical problem yet to be resolved involves small numbers of nanoscale divots and scratches which are generated by the operations used to meet smoothness requirements. A new non-contact surface sputtering technique known as accelerated neutral atom beam (ANAB) shows promise for mitigating the divot and scratch defects without increasing surface roughness and without altering flatness and planarity. This paper describes a mask blank substrate study which has been conducted to demonstrate the ANAB defect mitigation capability.

In EUV Lithography, an absence of promising candidate of EUV pellicle demands new requirements of EUV mask cleaning which satisfy the cleaning durability and removal efficiency of the various contaminations from accumulated EUV exposure. It is known that the cleaning with UV radiation is effective method of variety of contaminants from surface, while it reduces durability of Ru capping layer. To meet the expectation of EUV mask lifetime, it is essential to understand the mechanism of Ru damage. In this paper, we investigate dominant source of Ru damage using cleaning method with UV radiation. Based on the mechanism, we investigate several candidates of capping to increase the tolerance from the cycled UV cleaning. In addition, we study durability difference depending on the deposition method of Ru capping. From these studies, it enables to suggest proper capping material, stack and cleaning process.

Multiple challenges, including the availability of a reliable high power source, defect free mask, and proper resist material, have forced extreme ultraviolet (EUV) lithography to be considered for sub-10 nm half-pitch nodes. Therefore, techniques such as phase shift masks (PSMs) or high numerical aperture (NA) lithography might be considered. Such techniques require thin EUV absorber materials to be optimized to reduce EUV mask shadowing effects. Despite the challenges in dry etching of Ni and finding proper chemistries with a high etch selectivity to suitable capping materials, we decided to examine the chemical stability of Ni for existing mask cleaning chemistries. Ni, after Ag, has the highest absorption in EUV light at λ = 13.5 nm, which makes it a proper candidate—in pure form or in mixing with other elements—for thin absorber film. Depending on the composition of the final material, proper integration schemes will be developed. We studied Ni stability in commonly used mask cleaning processes based on ammonium hydroxide/ hydrogen peroxide (APM) and water mixtures. Ni films deposited with an ion beam deposition technique with a thickness of 35 nm are sufficient to totally absorb EUV light at λ = 13.5 nm. Multiple cleanings of these Ni films resulted in Ni oxidation— confirmed by time-of-flight secondary ion mass spectroscopy (TOF-SIMS) analysis as NiO with thickness about 1.5 nm. Furthermore, Ni oxidation processes are self-limiting and oxide layer thickness did not increase with a further cleaning. A three minute exposure to sulfuric acid/hydrogen peroxide mixture (SPM) can remove NiO and Ni totally. To protect Ni film from etching by SPM chemistry a 3 nm Si capping was used on top of Ni film. However, Si capping was removed by APM chemistry and could not protect Ni film against SPM chemistry. TiO₂ may be a very good capping layer for EUV optics but it is not suitable for EUV mask blanks and will be removed by APM chemistries.

One critical area for EUV lithography is the development of appropriate mask repair strategies. To this end, we have explored etching repair strategies for nickel absorber layers and focused electron beam induced deposition of ruthenium capping layers. Nickel has higher EUV absorption than the standard TaN absorber layer and thus thinner films and improved optical quality can be realized. A thin (2.5 nm) ruthenium film is commonly used as a protective capping layer on the Mo-Si EUV multi-layer mirror which mechanically and chemically protects the multi-layers during standard mask-making procedures. The gas field ion (GFIS) microscope was used to investigate helium and neon ion beam induced etching (IBIE) of nickel as a candidate technique for EUV lithography mask editing. No discernable nickel etching was observed for helium, however transmission electron microscopy (TEM) revealed subsurface damage to the underlying Mo-Si multilayers. Subsequently, neon beam induced etching at 30 keV was investigated and successfully removed the 50 nm nickel absorber film. TEM imaging also revealed subsurface damage in the underlying Mo-Si multilayer. Two damage regimes were apparent, namely: 1) beam induced mixing of the Mo-Si layers and 2) nanobubble formation. Monte Carlo simulations were performed and the observed damage regimes were correlated to: 1) the nuclear energy loss and 2) a critical implant concentration. Electron beam induced deposition (EBID) was explored to deposit ruthenium capping/protective layers. Several ruthenium precursors were screened and so far liquid bis(ethylcyclopentyldienyl)ruthenium(II) was successful. The purity of the as-deposited nanodeposits was estimated to be 10% Ru and 90% C. We demonstrate a new chemically assisted electron beam purification process to remove carbon by-products and show that high-fidelity nanoscale ruthenium repairs can be realized.

The force of adhesion of 50 nm diameter diamond-like carbon sphere probes to three quartz substrates was measured using an atomic force microscope. The force of adhesion was measured prior to cleaning, within 10 minutes after cleaning, after storage in an N₂-purged cabinet, and after storage in an N2-purged vacuum oven. The evaluated cleaning recipes were SC1-like, SPM-like, and HF-based, each followed by ultra-pure deionized water (UPW) rinse and spin drying. The measurements were conducted in a Class 100 clean room at approximately 50% relative humidity. In addition, contact angle measurements were made on three additional quartz substrates using UPW before cleaning, after cleaning, and throughout N₂ storage. The adhesion force increased after cleaning as compared to the pre-cleaned state, continued to increase until reaching a maximum after 5 days of N₂ storage, and then decreased after 26 days for all three substrates. One substrate was then stored in a vacuum oven for 3 days, and the adhesion force decreased to 46% of the pre-cleaned state. The contact angle was reduced from over 30° before cleaning to 0° immediately after cleaning. During subsequent N₂ storage, the contact angle increased to 5° or greater after 18 hours for the substrate cleaned with the HF-based recipe and after 15 days for the substrates cleaned by the SC1-like and SPM-like recipes.

In our previous work, various techniques were used to confirm the contamination deposits on the sidewall of extreme ultraviolet (EUV) mask absorbers [1-2]. In order to further understand the effects of contamination topography on mask absorbing features, direct measurements of contaminated features is needed. In this work, we investigated the contamination topography using cross-section transmission electron microscope (TEM) image analysis on four different masks. TEM specimens of contaminated features from silicon and ruthenium capped EUV masks were prepared using a focused ion beam (FIB). We conducted the contamination experiment with three different exposure sources including EUV, out-of-band, and electron induced processes. Thickness measurements from each contamination experiment were provided. Shadowing effect and geometric analysis on the contamination topography is also discussed.

Particle contamination in ultra-pure water (UPW) and chemicals will eventually end up on the surface of a wafer and may result in killer defects. To improve the semiconductor processing yield in sub-10 nm half pitch nodes, it is necessary to control particle defectivity. In a systematic study of all major techniques for particle detection, counting, and sizing in solutions, we have shown that there is a gap in the required particle metrology which needs to be addressed by the industry. To reduce particles in solutions and improve filter retention for sub-10 nm particles with very low densities (<10 particles/mL), liquid particle counters that are able to detect small particles at low densities are required. Non-volatile residues in chemicals and UPW can result in nanoparticles. Measuring absolute non-volatile residues in UPW with concentrations in the ppb range is a challenge. However, by using energy-dispersive spectroscopy (EDS) analysis through transmission electron microscopy (TEM) of non-volatile residues we found silica both in dissolved and colloidal particle form which is present in one of the cleanest UPW that we tested. A particle capture/release technique was developed at SEMATECH which is able to collect particles from UPW and release them in a controlled manner. Using this system we showed sub-10 nm particles are present in UPW. In addition to colloidal silica, agglomerated carbon containing particles were also found in UPW.

In this paper we introduce new source-mask co-optimization (SMO) capabilities for EUV with specific support of the details of imaging with NXE:33×0 scanners. New algorithms have been developed that fully exploit the adjustability of the light distribution inside the NXE:33×0 flexible illuminator, FlexPupil. The fast NXE M3D+ model accurately predicts the reflective 3D mask effects and enables novel pupil symmetries and mask defocus optimization. This mitigates the H-V bias, Bossung tilt, and pattern shift caused by shadowing and non-telecentricity, and reduces the sensitivity to flare. New pupil optimization flows will be shown. The optimized pupils are fully compliant with NXE:33×0 scanner specifications. We will demonstrate enhanced imaging performance of this NXE specific SMO on 7 nm node logic cut masks and show benefits up to 20% improved CD uniformity, and a reduction in the maximum pattern shifts.

EUV lithography (EUVL) is the most promising technology to extend the resolution limit, and is expected to be used if the enough source power is delivered and mask defect mitigation method is developed. However, even in that case, the number of EUV steps will be restricted by its high cost, and ArF immersion will still take a major role in the chip manufacturing. Therefore, it is important to check and improve the mix-match overlay (MMO) between EUV and ArF immersion steps. In this paper, we evaluate EUV MMO with ArF immersion system by comparing with dedicated chuck overlay (DCO). The major contributors on MMO are random and field component from overlay analysis. MMO is expected to be below 3nm by applying 18para CPE^TM(correction per exposure) and RegC^TM(Registraion error correction). We consider High oder CPE^TM need to be developed for further improvement.

Though scaling of source power is still the biggest challenge in EUV lithography (EUVL) technology era, CD and overlay controls for transistor‟s requirement are also precondition of adopting EUVL in mass production. Two kinds of contributors are identified as risks for CDU and Overlay: Infrared (IR) and deep ultraviolet (DUV) out of band (OOB) radiations from laser produced plasma (LPP) EUV source. IR from plasma generating CO2 laser that causes optics heating and wafer overlay error is well suppressed by introducing grating on collector to diffract IR off the optical axis and is the effect has been confirmed by operation of pre-production tool (NXE3100). EUV and DUV OOB which are reflected from mask black boarder (BB) are root causes of EUV-specific CD error at the boundaries of exposed shots which would result in the problem of CDU out of spec unless sufficiently suppressed. Therefore, control of DUV OOB reflection from the mask BB is one of the key technologies that must be developed prior to EUV mass production. In this paper, quantitative assessment on the advantage and the disadvantage of potential OOB solutions will be discussed. EUV and DUV OOB impacts on wafer CDs are measured from NXE3100 & NXE3300 experiments. Significant increase of DUV OOB impact on CD from NXE3300 compared with NXE3100 is observed. There are three ways of technology being developed to suppress DUV OOB: spectral purity filter (SPF) as a scanner solution, multi-layer etching as a solution on mask, and resist top-coating as a process solution. PROs and CONs of on-scanner, on-mask, and on-resist solution for the mass production of EUV lithography will be discussed.

As EUV Lithography (EUVL) continues to evolve, it offers a possible solution to the problems of additional masks and lithography steps that drive up the cost and complexity of 193i multiple patterning. EUVL requires a non-telecentric reflective optical system for operation. This requirement causes EUV specific effects such as shadowing. The absorber physically shadows the reflective multilayer (ML) on an EUV reticle resulting in pattern fidelity degradation. To reduce this degradation, a thinner absorber may help. Yet, as the absorber thickness decreases, reflectivity increases in the ‘dark’ region around the image field, resulting in a loss of contrast. The region around the edge of the die on the mask of unpatterned absorber material deposited on top of ML, known as the image border, is also susceptible to undesirable reflections in an ideally dark region. For EUVL to be enabled for high-volume manufacturing (HVM), reticle masking (REMA) blades are used to shield light from the image border to allow for the printing of densely spaced die. When die are printed densely, the image border of each neighboring die will overlap with the edge of a given die resulting in an increase of dose that overexposes features at the edge of the field. This effect is convolved with a fingerprint from the edge of the REMA blades. This phenomenon will be referred to as a field edge effect. One such mitigation strategy that has been investigated to reduce the field edge effect is to fully remove the ML along the image border to ensure that no actinic-EUV radiation can be reflected onto neighboring die. This has proven to suppress the effect, but residual out-of-band radiation still provides additional dose to features near the image border, especially in the corners where three neighboring fields overlap. Measurements of dense contact holes (CHs) have been made along the image border with and without a ML-etched border at IMEC in collaboration with Micron using the ASML NXE:3100. The implementation of these measurements allow for further mitigation, i.e., compensation by OPC. Mentor Graphics’ Calibre software uses the scanner’s point spread function and convolves it with the mask layout to generate a flare map. It also has the capability to add additional dose to the image border which can be optimized to fit the experimental data. This includes the transition region between the image field and border that results in a linear rolloff of dose due to partial shadowing of the REMA blades. By applying this flaremap that accounts for neighboring die to the already calibrated optical and resist models, OPC can now be enabled to compensate for field edge effects. This study has two goals. First, we will show that OPC can be used to compensate both for field edge effects with and without a etched ML border. The second is to investigate the limitations that exist for OPC in the areas altered by neighboring die. This will predict when a process to mitigate the field edge effect is needed to enable EUV HVM.

Defect avoidance methods are likely to play a key role in overcoming the challenge of mask blank defects in EUV lithography. In this work, we propose a novel EUV mask defect avoidance method. It is the first approach that allows exploring all the degrees of freedom available for defect avoidance (pattern shift, rotation and mask floorplanning). We model the defect avoidance problem as a global, non-convex optimization problem and then solve it using a combination of random walk and gradient descent. For a 8nm polysilicon layer of an ARM Cortex M0 layout, our method achieves 60% point better mask yield compared to prior art in defect avoidance for a 40-defect mask.

We verify image fidelity after mask 3D aware-OPC (using Mentor Graphics Domain Decomposition Method) and quantify pattern placement error (PPE) on wafer. First we show experimental pattern fidelity improvement of DDM-OPCed 2D-images of logic devices in 10 nm technology node with the latest NXE3300B EUV exposure tool. We then compare pattern fidelity in aerial images after DDM-OPC to aerial images using rigorous simulation of electric and magnetic field. Finally we quantify PPE in resist images with modeled 1D layouts after a perfect OPC. The perfect OPC corrects optical proximity effect, azimuthal angle through slit, and lens aberration. The forecasted PPE residual error after perfect OPC is 0.21 nm (x) and 0.76 nm (y) that can be attributed to uncorrectable components of wafer defocus and mask flatness. For modeling and compensation of pattern placement error, a new metrology method should be developed.

In this paper we discuss the EUV OPC modeling challenges and potential solutions, as well as OPC integration requirements to support the forthcoming application of EUV lithography. 10-nm-node OPC modeling is considered as an example. Wafer and mask process data were collected for calibration and verification patterns, to understand the mask making error/OPC model interactions. Several factors, including compact mask topography modeling impact, were analyzed by means of rigorous simulations and model fitting. This was performed on a large-scale data set, to ensure accurate characterization of the OPC modeling strategies, using a large number of patterns.

The EUV mask infrastructure is of key importance for a successful introduction of EUV lithography into volume production. In particular, for the production of defect free masks, actinic review of potential defect sites is required. To realize such an actinic review tool, Zeiss and the SEMATECH EUVL Mask Infrastructure consortium started a development programme for an EUV aerial image metrology system (AIMS™ EUV). In this paper, we discuss the status of the on-going system integration and show first results from the first light tests of the prototype tool.

The SEMATECH High Numerical Aperture Actinic Reticle Review Project (SHARP) is a synchrotron-based extreme ultraviolet (EUV) microscope dedicated to photomask research. SHARP has been operational and serving users since June, 2013, and in eight months, SHARP has recorded over 71,000 high-resolution images. Exposure times are 5 to 8 seconds, and 8 or more through-focus series can be collected per hour at positions spanning the entire mask surface. SHARP’s lossless coherence-control illuminator and variable numerical aperture (NA) enable researchers to emulate the imaging properties of both current and future EUV lithography tools. SHARP’s performance continues to improve over time due to tool learning and upgraded capabilities, described here. Within a centered, 3-μm square image region, we demonstrate an illumination power stability above 99%, and an average uniformity of 98.4%. Demonstrations of through-focus imaging with various illumination coherence settings highlight the capabilities of SHARP.

The framework and the current status of a newly developed PEM pattern inspection system are presented. A die-to-die defect detection sensitivity of the developing system is investigated. A programmed defect mask was used for demonstrating the performance of the system. Defect images were obtained as difference images by comparing PEM images with-defects to the PEM images without-defects. The image-processing system was also developed for die-to-die inspection. A targeted inspection throughput of 19-hour inspection per mask with 16nm pixel size for image capture was attained. Captured image of 28 nm intrusion defect in hp 64 nm L/S pattern was used for detection. The defect is clearly identified by the image processing. But several false defects are also detected. To improve the defect detection sensitivity to reach the targeted level of achieving a higher than 10 S/N value at 16 nm defect size, by applying a higher current density and a developed inspection algorithm adjustment is, currently an on-going program.

In this paper, we address a new inspection method which provides in-focus inspection capability and higher defect sensitivity compared with conventional mask inspection methods. In the Zernike phase contrast microscope, an added phase shift to background wave combines with the phase of bump and pit defects to achieve higher contrast at focus. If we use a centralized apodization to half the lens radius to further reduce the intensity of the phase-shifted background wave, the signal strength can be improved up to 6-fold of its original value. Simulation results further show that this apodization for a typical EUV mask power spectral density results in the noise decreasing in absolute level similar to the clear field reference signal. Thus large improvements in signal to noise ratios are possible with the Zernike phase contrast microscope type systems for EUV mask inspection applications.

Publisher's Note: This paper, originally published on April 17, 2014, was withdrawn at the author's request on May 25, 2016.

EUV Lithography is aimed to be inserted into mainstream production for sub-20nm pattern fabrication. Unlike conventional optical lithography, frequent defectivity monitors (adders, repeaters etc.) are required in EUV lithography. Due to sub-20nm pattern and defect dimensions e-beam inspection of critical pattern areas is essential for yield monitor. In previous work we showed sub-10nm defect detection sensitivity¹ on patterned resist wafers. In this work we report 8-10× improvement in scan rates of etched patterns compared to resist patterns without loss in defect detection sensitivity. We observed good etch transfer of sub-10nm resist features. A combination of smart scan strategies with improved etched pattern scan rates can further improve throughput of e-beam inspection. An EUV programmed defect mask with Line/Space, Contact patterns was used to evaluate printability of defects and defect detection (Die-Die and Die-Database) capability of the e-beam inspection tool. Defect inspection tool parameters such as averaging, threshold value were varied to assess its detection capability and were compared to previously obtained results on resist patterns.

High-radiance EUV source is needed for actinic mask inspection applications. LDP source for a lithography application was found to be also able to provide sufficient radiance for mask inspection purpose. Since the plasma size of LDP is properly larger than LPP, not only radiance but also power is suitable for mask inspection applications. Operating condition such as discharge pulse energy, discharge frequency and laser parameter have been tuned to maximize radiance. Introduction of new techniques and several modifications to LDP source have brought radiance level to 180 W/mm²/sr at plasma (or 130 W/mm²/sr as clean-photon radiance). The LDP source is operated at moderate power level in order to ensure sufficient component lifetime and reliability. The first lifetime test done at 10 kHz resulted in 6.5 Gpulse without failure. Debris mitigation system has been successfully installed showing optical transmission as high as 71 %.

Photon sources for extreme ultraviolet Lithography (EUVL) are still facing challenges in required performance for high volume manufacture. Currently EUVL community has focused the research and developments on the dual-pulse laser produced plasma (LPP) devices with mass-limited targets. Such complex systems require extensive optimization to enhance the conversion efficiency (CE) and components lifetime and such optimization requires significant experimental and costly efforts.
We continued to enhance our state-of-the art HEIGHTS package to analyze and optimize LPP sources and to make projections and realistic predictions of near future powerful devices. HEIGHTS package includes full 3-D detail description of all integrated physical processes involved in LPP devices. The models continued to be upgraded and well benchmarked in each interaction physics phase of plasma evolution for EUV and BEUV production.
We simulated LPP sources using small droplets as the targets and evaluated the requirements for optimization of these sources in dependence on laser wavelength. We also simulated the targets as distributed fragments resulting from the intense pre-pulse laser energy deposition. Additionally, we simulated vapor/plasma mixture created by pre-pulse laser with comparatively low intensity. We studied mass dependence, laser parameters efficiency, optimization of EUV (13.5 nm) and BEUV (6.7 nm) radiation output, and atomic and ionic debris generation to predict potential damage to the optical collection system from energetic debris and the requirements for mitigating systems to reduce debris fluence. Our modeling and simulation included all phases of laser target evolution: from laser/droplet interaction, energy deposition, target vaporization and fragmentation, ionization, plasma hydrodynamic expansion, thermal and radiation energy redistribution, and EUV/BEUV photons collection as well as detail mapping of photons source location and size. Modeling results were benchmarked against recent experimental studies for the in-band photons production and for debris and ions generation for both EUV and BEUV systems.

The suppression of outgassing from extreme ultraviolet (EUV) resist needs to be addressed for realizing EUV lithography (EUVL) because outgassing is likely the main contributor to the contamination of mirror optics in EUV scanners, which results in reflectivity loss. Resist outgassing causes two types of contamination: cleanable contamination, involving hydrocarbon contaminants, and noncleanable contamination, involving noncarbon components. The relation of cleanable contamination between EUV- and electron beam (EB)-based evaluations is linear. However, the relation of noncleanable contamination is not clear. In this study, we investigated the contribution of EUV resist components to noncleanable contamination using different photoacid generator components. The cleanability of noncleanable elements (sulfur, iodine, chlorine, and bromine) in contamination films was measured and compared for the EUV- and EB-based outgas testers. The result suggested that the chlorine and bromine contaminants were completely removed after cleaning. On the other hand, sulfur and iodine remained even after cleaning. This suggested that the careful use of iodine in resist materials is necessary due to its high photoabsorption and low cleanability. In addition, the cleaning rate and noncleanability of contaminants in the EUV-based test were larger than in the EB-based test. This suggests that the contamination film in the EUV-based test is more porous than that in the EB-based test.

EUV lithography is a technology enabling next generation electronic devices, but issues with photoresist sensitivity, resolution and line edge roughness as well as tool downtime and throughput remain. As part of the industry's efforts to address these problems we have worked with resist suppliers to quantify the relative contamination rate of a variety of resists on EUV multilayer mirror analogues following ASML approved protocols. Here we present results of our ongoing program to better understand the effect of process parameters such as dose and resist thickness on the contamination rate of ruthenium coated witness plates, additionally we present results from a study on the effectiveness of hydrogen cleaning.

EUV photoresists are considered as a potential source of optics contamination, since they introduce irradiation-induced outgassing in the EUV vacuum environment. Therefore, before these resists can be used on e.g. ASML NXE:3100 or NXE:3300, they need to be tested in dedicated equipment according to a well-defined procedure, which is based on exposing a witness sample (WS) in the vicinity of a simultaneously exposed resist as it outgasses. Different system infrastructures are used at multiple sites (e.g. NIST, CNSE, Sematech, EIDEC, and imec) and were calibrated to each other by a detailed test plan. Despite this detailed tool qualifications, a first round robin comparison of identical materials showed inconsistent outgas test results, and required further investigation by a second round robin. Since the resist exposure mode is different at the various locations (some sites are using EUV photons while others use E-gun electrons), this difference has always a point of concern for variability of test results. In this work we compare the outgas test results from EUV photon and electron exposure using the resist materials of the second round robin. Since the imec outgas tester allows both exposure methods on the resist, a within-system comparison is possible and showed limited variation between photon and electron exposure mode. Therefore the system-to-system variability amongst the different outgas test sites is expected to be related to other parameters than the electron/photon exposure mode. Initial work showed that the variability might be related to temperature, E-gun emission excursion, and/or residual outgassing scaled by different wafer areas at the different sites.

Previous investigations on the patterning mechanism of nanoparticle photoresists provided insight into ligand displacement exerting a controlling influence on dissolution behavior of nanoparticles in organic developers. Nanoparticle core-ligand interaction which dictates ligand displacement would ultimately translate to the sensitivity of the photoresist. The current study investigates enhancement of resist sensitivity via altering the core-ligand interaction of the nanoparticle, which further emphasizes our proposed patterning hypothesis.

Extreme ultraviolet (EUV) lithography is a candidate for the manufacturing of semiconductor devices at the 22 nm half pitch node and below. EUV lithography requires high performance resist with limited outgassing property. The key challenge for EUV resist is the simultaneous requirement of ultrahigh resolution (R), low line edge roughness (L) and high sensitivity (S) for lines and spaces (LS) features. To achieve high resist sensitivity EUV resist absorbance should be increased. Resin containing fluorine atom is one of the most attractive methods to improve absorbance level of EUV resist because the fluorine atom absorbs EUV light strongly. However, resist hydrophobicity (or high contact angle) also increase due to presence of fluorine atoms in the resist polymer. It is difficult to rinse high CA resist during development process so the resist containing polymer with fluorine atom may produce additional defects. In this paper, we will report the relationship between line edge roughness and acid diffusion length. We will also show the method to diminish defects caused by high contact angle (CA) resist. We achieved good resolution and LER improvement by controlling acid diffusion length. Moreover, we found the relationship of the number of defects and the structure of the monomers containing fluorine units.

Polymers with a different Tg and activation energy were prepared to clarify influences of acid diffusion on resolution at 15 nm half-pitch (hp) and 14 nm hp using a EUV micro-field exposure tool (MET) at LBNL. Resolution on such a narrow pattern was limited by collapse and pinching. Clear relationship between pinching numbers and polymer Tg indicates that acid diffusion is one of major contributors on the pinching. In addition, polymers with a low thermal activation energy (Ea) on deprotection were effective for reducing pinching. This is probably originated from its high chemically amplification character even in low post-exposure bake (PEB) temperature to obtain both large chemical contrast and short acid diffusion. On the other hand, a good correlation between a cleanable outgassing amount and Ea indicates trade-off relationship between outgassing and resolution. Advantages of n-butyl acetate (nBA) developer have been investigated in viewpoint of dissolution uniformity. Surface roughness of a non-patterned resist film at half-exposed area, which was well correlated with LWR, was measured by AFM as indicator of uniformity in development process. To avoid any differences in resist chemistry other than development process, cross linking negative tone resist was applied for this study. The surface roughness obtained by nBA, which is conventional negative-tone imaging (NTI) developer, was 32 % lower than that obtained by 2.38 % TMAH solution. NTI resist system with a nBA developer and optimized resist reduced LWR from 4.8 nm to 3.0 nm in comparison with conventional positive tone resist with a 2.38 % TMAH developer. In addition, advantage on semi-dense trench patterning was well defined. New EUV sensitizer with 1.15 times higher EUV absorption resulted in 1.15 times higher acid yield by EUV exposure. Lithography performance of the new EUV sensitizer has been investigated by MET at SEMATECH Albany. Sensitivity was indeed improved from 20 mJ/cm² to 17 mJ/cm² according to the acid yield increase, but resolution was significantly degraded.

The uniformity of acid generator distribution and the length of acid diffusion are serious problems in the development of resist materials used for the 16nm node and below. Anion-bound polymers in which the anion part of onium salts is polymerized have attracted much attention for solving these problems. In this study, the reaction mechanism of an anion-bound polymer in cyclohexanone was clarified using pulse radiolysis. The design of an efficient electron and hole transfer system is essential to the enhancement of resist performance.

Gray-scale e-beam lithography has been performed to match the EUV and e-beam aerial image log slope for studying shot noise fundamentals in the two mechanisms through line-edge roughness (LER) measurements for 50 nm lines and spaces patterned on a leading chemically amplified EUV resist. The measured e-beam exposure latitude decreased from 0.4 with binary patterning to 0.28 with gray-scale e-beam exposure designed to match the EUV incident image profile, closely matching the EUV exposure latitude of 0.26. Calculations of absorption statistics with EUV and e-beam suggest that the shot noise with e-beam patterning is expected to be 10% larger than the shot noise with EUV patterning. However, despite the matched image gradients and close to identical absorbed quanta predictions, the e-beam patterned LER is 2.5× larger than the EUV patterned LER.

The NIST Extreme Ultraviolet (EUV) Reflectometry Facility was designed in the 1990s to accommodate the largest multilayer optics envisioned at that time. However, with increasing power requirements for an EUV scanner, source collection optics have grown larger and more steeply curved than the original design would allow. To accommodate these changes, the mechanical and operational parameters of the facility have been upgraded. To access the entire surface of a larger optic, an auxiliary off-axis rotation stage has been installed allowing an increase in maximum optic size from 350 mm to 450 mm. Likewise, to deal with the deeper sags and steeper slopes of these optics, we have had to significantly expand our data analysis capabilities. In order to make these measurements, the incident radiation is reflected out of the vertical plane, allowing for measurements of effectively unpolarized radiation, an advantage for EUV lithography optics such as source collectors.

The first NXE3300B systems have been qualified and shipped to customers. The NXE:3300B is ASML’s third generation EUV system and has an NA of 0.33. It succeeds the NXE:3100 system (NA of 0.25), which has allowed customers to gain valuable EUV experience. Good overlay and imaging performance has been shown on the NXE:3300B system in line with 22nm device requirements. Full wafer CDU performance of <1.5nm for 22nm dense and iso lines at a dose of ~16mJ/cm2 has been achieved. Matched machine overlay (NXE to immersion) of around 3.5nm has been demonstrated on multiple systems. Dense lines have been exposed down to 13nm half pitch, and contact holes down to 17nm half pitch. 10nm node Metal-1 layers have been exposed with a DOF of 120nm, and using single spacer assisted double patterning flow a resolution of 9nm has been achieved.

Source power is the major challenge to overcome in order to achieve cost-effectiveness in EUV and enable introduction into High Volume Manufacturing. With the development of the MOPA+prepulse operation of the source, steps in power have been made, and with automated control the sources have been prepared to be used in a preproduction fab environment.

Flexible pupil formation is under development for the NXE:3300B which will extend the usage of the system in HVM, and the resolution for the full system performance can be extended to 16nm. Further improvements in defectivity performance have been made, while in parallel full-scale pellicles are being developed.

In this paper we will discuss the current NXE:3300B performance, its future enhancements and the recent progress in EUV source performance.

In last year’s report, we discussed the design and requirements of the optical projection module (Projection Optics Box [POB]) for the 0.5-NA Micro-field Exposure Tool (MET5) and the resulting challenges. Over the course of this past year, we have completed and fully qualified the metrology of individual mirrors. All surface figure errors have been measured over seven orders of magnitude with spatial periods ranging from the full clear aperture down to 10 nm. The reproducibility of the full aperture tests measures 16 pm RMS for the M1 test and 17 pm for the M2 test with a target of 30 pm for both tests. Furthermore, we achieved excellent results on scatter and flare: For scatter, both mirrors perform about a factor of two below specification. For flare, the larger M2 mirror performs well within and the smaller M1 mirror about a factor of two below specification. In addition, we have developed processes for correcting surface figure errors for both mirrors and have successfully demonstrated high-reflectivity coatings on pathfinder mirrors. Further, we have achieved significant goals with respect to the design, assembly, metrology and alignment of the projection module. This paper reviews this progress and describes the next step in the ambitious MET5 POB development program.

EUV lithography is expected to be introduced in volume manufacturing at the 10-nm and 7-nm node. Especially in these first EUV nodes, critical layer patterning will be balanced with the use of ArF immersion. As a consequence a good overlay and placement matching between both lithography methods becomes an enabling factor for EUV. In this paper we present an integral method to optimize critical layer patterning across the EUV and ArF scanner platform, such that good overlay and device pattern placement is achieved. It is discussed that besides classical overlay control methods, also the optimization of the ArF and EUV imaging steps is needed. Best matching is achieved by applying high-order field-to-field overlay corrections for both imaging and overlay. The lithography architecture we build for these higher order corrections connects the dynamic scanner actuators with the angle resolved scatterometer via a separate computational application server. Improvements of CD uniformity are based on source mask optimization for EUV combined with CD optimization using freeform intra-field dose actuator in the immersion scanner.

In support of the Extreme Ultraviolet Lithography (EUVL) roadmap, a SEMATECH/CNSE joint program is underway to produce multiple EUVL (wavelength of 13.5 nm) R&D photolithography tools. The 0.5 NA projection optic magnification (5X), track length and mechanical interfaces match the currently installed 0.3 NA micro-field exposure tools (MET) projection optic [1] [2] [3]. Therefore, significant changes to the current tool platforms and other adjacent modules are not necessary. However, many of the existing systems do need upgrades to achieve the anticipated smaller exposure feature sizes [4]. To date we have made considerable progress in the production of the first of the two-mirror 0.5 NA projection optics for EUVL [5]. With a measured transmitted wave front error of less than 1 nm root mean square (RMS) over its 30 μm × 200 μm image field, lithography modeling shows that a predicted resolution of ≤12 nm and an ultimate resolution of 8 nm (with extreme dipole illumination) will be possible.

This paper will present an update from the 0.5 NA EUVL program. We will detail the more significant activities that are being undertaken to upgrade the MET and discuss expected performance.

As EUV approaches high volume manufacturing, reticle defectivity becomes an even more relevant topic for further investigation. Current baseline strategy for EUV defectivity management is to design, build and maintain a clean system without pellicle. In order to secure reticle front side particle adders to an acceptable level for high volume manufacturing, EUV pellicle is being actively investigated. Last year ASML reported on our initial EUV pellicle feasibility. In this paper, we will update on our progress since then. We will also provide an update to pellicle requirements published last year. Further, we present experimental results showing the viability and challenges of potential EUV pellicle materials, including, material properties, imaging capability, scalability and manufacturability.

High numerical aperture (high-NA) extreme ultraviolet (EUV) is one option to enable a higher resolution than EUV can achieve with single patterning. An industry effort to achieve consensus on the key parameters of high-NA EUV is described. At high-NA, three-dimensional (3D) mask effects cause a loss of contrast in the image that is recovered by increasing the scanner de-magnification. This leads to a tradeoff between wafer field and mask size that has considerable impact on mask cost and scanner cost of ownership.

EUV technology has steadily progressed over the years including the introduction of a pre-production NXE:3100 scanner that has enabled EUV process development to advance one step closer to production. We have carried out the integration with 20/14nm metal layer design rules converting double patterning with ArF immersion process to EUV with a single patterning solution utilizing a NXE3100 exposure tool. The exercise through the integration of a mature test chip with an EUV level has allowed us to have early assessment of the process challenges and new workflow required to enable EUV to the mass production stage. Utilizing the NXE3100 in IMEC, we have developed an OPC model and a lithography process to support 20/14nm node EUV wafer integration of a metal layer in conjunction with immersion ArF. This allows early assessment of mix-and-match overlay for EUV to immersion system that is critical for EUV insertion strategy as well as further understanding of the litho process, OPC, and mask defect control specific to EUV single patterning. Through this work we have demonstrated high wafer yields on a 20nm test vehicle utilizing single EUV Metal layer along with additional ArF immersion levels. We were able to successfully demonstrate low mask defectivity and good via chain and open/short electrical yield. This paper summarize the learning cycles from mask defect mitigation and mix machine overlay through post metal CMP wafer integration highlighting the key accomplishments and future challenges.

Traditionally, semiconductor density scaling has been supported by optical lithography. The ability of the exposure tools to provide shorter exposure wavelengths or higher numerical apertures have allowed optical lithography be on the forefront of dimensional scaling for the semiconductor industry. Unfortunately, the roadmap for lithography is currently at a juncture of a major paradigm shift. EUV Lithography is steadily maturing but not fully ready to be inserted into HVM. Unfortunately, there are no alternative litho candidates on the horizon that can take over from 193nm. As a result, it is important to look into the insertion point of EUV that would be ideal for the industry from an economical perspective. This paper details the benefit observed by such a transition. Furthermore, it looks into such detail with an EUV throughput sensitivity study.

In EUV lithography (EUVL), the most critical issue has been low intensity of the EUV light source. Light-source intensity and resist sensitivity have a complementary relationship. Therefore, the sensitization of EUV resist is very important to compensate the low intensity of the EUV light source. However, dramatically improving the resist sensitivity of chemically amplified resist (CAR) is very difficult because of the resolution/line-width roughness/sensitivity (RLS) trade-off. Therefore, we propose a very new process: high resist sensitization by the combination lithography of EUV or EB pattern exposure with UV flood exposure (PF combination lithography) of photosensitized chemically amplified resist (PS-CAR). The combination lithography of EB pattern exposure with UV flood exposure achieved a sensitivity enhancement of more than a factor of 10 with respect to conventional EB single-exposure lithography, without loss in space resolution of line/space of 75 nm¹. The breakthrough of RLS trade-off by PF combination lithography of PS-CAR is explained based on RLS trade-off simulation method2 and acid generation mechanism of PF combination lithography of PS-CAR. One of the problems of the PF combination lithography of PS-CAR is postexposure delay (PED) effects suffered from airborne contamination. The present paper describes forming of 75 nm contact hole and PED effects in the PF combination lithography of PS-CAR. The sensitization of EUV resists by new method accelerates EUVL implementation.

We demonstrate the use of Prism Computational Sciences software models to understand the fundamental physical processes in EUVL (Extreme Ultraviolet Lithography) plasma. HELIOS-CR is a 1-D radiation-magnetohydrodynamics code used to simulate the dynamic evolution of laser-produced plasmas (LPP) such as Sn or Xe plasmas used often in EUVL. The results of HELIOS-CR simulations can be post-processed using multidimensional spectral analysis code SPECT3D to generate images and spectra that include instrumental effects, and therefore can be directly compared with experimental measurements being made to generate EUVL. The SPECT3D package computes filtered and monochromatic images, and streaked, time-integrated, and time-gated spectra based on 1-D, 2-D, or 3-D radiation-hydrodynamics results. Simulated images and spectra can be computed with instrumental effects included (e.g., spectrometer resolution, time gating, filtering) in order to facilitate comparisons with experimental data. We present results which demonstrate impact of various parameters (laser power - pre-pulse and main pulse, spot size, pre-pulse and main pulse delay, etc.) on efficiency in EUVL, and shows possible improvement areas to achieve higher efficiency.

To protect an EUV mask from contamination, a pellicle can be used. However, the pellicle membrane must be very thin due to EUV absorption. As a result, a pellicle support structure is needed to avoid deflection of the membrane by gravity. Previous authors have shown that such a structure would produce a non-uniform intensity distribution on the wafer. In this paper, we use simulation to re-examine the issue. The results show that, when coherent illumination is used, a pellicle support structure would have an undesirable effect on the aerial image. However, we also show that, when partially coherent illumination is used, the intensity non-uniformity caused by the pellicle support structure can be effectively smoothed out, resulting in a perfectly acceptable aerial image.

Laser produced plasma (LPP) light sources for extreme ultraviolet (EUV) lithography currently has been extensively studied. Most of the studies are based on CO₂ laser induced plasma from mass limited tin targets. In this work, a droplet dispenser that produces uniform droplets size of about 150μm was established. A pulsed TEA-CO₂ laser and a Nd: YAG laser irradiated the droplets producing plasma respectively to get EUV emission. An X-ray Spectrometer and EUV photodiodes were used to collect the spectra and EUV radiation. The different EUV spectral composition and angular distribution of EUV emission from plasmas induced by the CO₂ and Nd: YAG laser were studied.

The design, synthesis and characterization of non-chemically amplified negative tone electron-beam and EUV resists based on the inclusion of a radiation sensitive sulfonium functional group are outlined.. MAPDST (4-(methacryloyloxy phenyldimethylsulfoniumtriflate) and MANTMS (1-(4-(methacryloyloxy)naphthalen-1-yl)tetrahydro-1H thiopheniumtrifluoromethane sulfonate) monomers each containing the sulfonium group underwent homo- and copolymerizations using free radical polymerization with 2,2'-azobisisobutyronitrile (AIBN) initiator. These resist materials were evaluated by EB lithography using 20 keV electron beam and EUV lithography to obtain sub-20 nm line patterns. These features were optimized ranging from resist coating, pre-exposure bake, exposure to e-beam, postexposure bake, development and imaging. Our investigation showed that these newly synthesized resists are potential viable candidates for EUV lithography based on their ability to form flaw free thin films < 50nm, sensitivity, resolution and LER control.

The half-tone phase shift mask (PSM) has been suggested for better imaging performances like image contrast, NILS and H-V bias compared to the binary mask (BIM) in EUV lithography. In this paper, we measured imaging performance of a fabricated half-tone attenuated PSM with Coherent Scattering Microscopy (CSM) and the results were compared with simulation data obtained by EM-suite tool. We prepared a half-tone attenuated PSM which has 12.7% reflectivity and 180° phase shift with absorber stack of 16.5mn-thick TaN absorber and 24nm-thick Mo phase shifter. With CSM, an actinic inspection tool, we measured the imaging properties of PSM. The diffraction efficiencies of BIM were measured as 31%, 36%, and 44% for 88 nm, 100 nm, and 128 nm mask CD, respectively, while those of PSM were measured as 45%, 62%, and 81%. Also the aerial image at wafer level obtained by CSM with high volume manufacturing tool’s (HVM) illumination condition (NA=0.33, σ=0.9) showed higher image contrast and NILS with phase shift effect. And the measured data were consistent with the simulation data.

It is expected that EUV resists must simultaneously pattern 20-nm half-pitch and below, with an LWR of <1.8 nm, and a sensitivity of 5–20 mJ/cm². In order to make a resist perform optimally, new resist chemistry is required. One such approach being investigated by us is the development of polymeric non-CAR negative photo resists for sub 16 nm technology which is directly sensitive to radiation without utilizing the concept of chemical amplification (CARs). These resist designs are accomplished by homopolymers which are prepared from monomers containing sulfonium groups. We have achieved 20 nm patterns by e-beam lithography using this system. Here we will discuss in detail process parameters such as: spinning conditions for film thicknesses <50 nm and resulting surface topographies, baking regimes, exposure conditions and protocols on sensitivity, contrast, resolution and LER/LWR. Etch resistance data on these thin films will also be provided. Our results are aimed to provide a clear understanding of how these critical steps in the lithographic imaging process will affect extendibility of the non-CAR resist concept to sub 20 nanoscale features. Photodynamics and EUV exposure data will be covered.

With current progress in exposure source power, novel resist materials, and post processing techniques, EUV is getting closer to the production environment. As reported continuously, SEMATECH established cycles of learning program. The data generated from the program has been utilized to measure current state of the art of EUV photoresist for production or pilot line use. Thanks to SEMATECH core and associate members’ attention to the project, numerous EUV samples have been tested and they were based on the best performing EUV resists from associate members. This year we completed the evaluations for under-layers, lines and spaces, and contact holes. We also applied track based techniques to drive both low line edge roughness control and enlarge the process window with techniques such as FIRM^TM and track based smoothing process. In this paper we will discuss about the results from cycles of learning test and show post-processing results of the three best line and space resists when combined with different FIRM^TM materials.

A major challenge for extreme ultraviolet lithography (EUVL) is avoiding defects in the fabrication of multilayered (ML) mask blanks. Substrate defects and adders during ML coating are responsible for ML defects which causes changes on phase and amplitude of EUV light. ML defects must be identified by inspection prior to absorber patterning in order to reduce the effects of ML defects via covering them with patterns to permit the use of fewer ML defect blanks. Fiducial marks (FMs) on ML blanks can be used for mask alignment and to accurately and precisely determine the locations of ML defects. In this study, we fabricated an FM mask by resist exposure using an e-beam writer and etching. Then, we inspected FMs and ML defects with an EUV actinic full-field mask blank inspection tool developed by EIDEC-LaserTec (LT ABI). Next, we evaluated the ML defect location accuracy on the mask based on FMs of several line depths by deriving center position of FMs and defects with Lorentz, Gaussian fitting and center-of-mass calculation. Here, we explain the estimation of defect location accuracy using FMs and the LT ABI, and discuss the defect numbers which can be covered by absorber patterns. Fewer than 19 defects per blank should be required for EUV blanks to cover ML defects with patterns.

In this study, influence of geometric features of phase defect on EUV optical images, such as dark field images(ABI tool) and bright field images (Exposure tool), was studied with experiment and simulation. It is confirmed that ABI signal intensity has a direct relationship not only with defect volume but also with geometric features of the phase defect. A new method for the making of phase defect models for simulation purpose was attempted. With this new method, the simulated ABI signal intensity exhibited a good proportionality relationship with the experimentally obtained ABI signal intensity. This method is believed to be effective in the estimation of ABI signal intensity with simulation. These results lead to the conclusion that it is important to take into consideration the geometric features of phase defects when a relationship between ABI signal intensity and phase defect feature is to be examined. Also, the relationship between ABI signal intensity and impact of defect on wafer was studied with simulation. In this study, many geometric types of phase defects were used, and relationships between these defects and their impacts on wafers were studied. As a result, it was confirmed that the geometric features did not strongly affect the relationship between the ABI signal intensity and the defects’ impacts on wafer however the ABI signal intensity is found to have a close relationship with impact of defect on wafer. These results lead to the conclusion that information about ABI signal intensity is important not only for the detection of phase defect but also for estimating the impact of phase defects on wafers.

In this paper, the latest advances in the field of extreme ultra violet (EUV) light sources for application in counter-facing plasma focus devices are presented. EUV emission, plasma and electrical properties under two pulses operation are reported. Using this new plasma focus system, the total amount of supply material and the energy cost of the plasma source could be reduced. The physical behavior during a two-pulse experiment with a high repetition rate (1 kHz) is explained. Continuous operation for future practical use in advanced lithography systems is also investigated.

The effect of phase defect on extreme ultraviolet (EUV) lithography was examined using an EUV microscope. A test mask containing periodic absorber line patterns and programmed pit phase defects embedded in a multilayer-coated mask blank was prepared, and the mask patterns were observed by the EUV microscope developed by Tohoku University and constructed at the site of a beam line of the New SUBARU of the University of Hyogo. The half pitches of the absorber patterns were 64 nm and 44 nm at mask which corresponded to 16 nm and 11 nm device generations. The programmed defects included not only square-shape defects but also rectangular-shape defects with different orientations. When a phase defect was located between two adjacent absorber patterns, then the observation image intensity of the absorber lines and spaces (L/S) patterns varied, and the impact of a phase defect was predicted as an intensity variation of bright space image. Phase defect location dependency and defect shape dependency of the observation image intensity were examined. The effectiveness of the EUV microscope to predict the phase defect impacts was confirmed.

Influence of phase defect structures on Extreme Ultraviolet (EUV) microscope images were examined to predict the inclination angle dependency of the phase defect impact on wafers since the phase defect does not always propagate in a vertical direction from an EUV mask substrate surface through the multilayer. Two types of the programmed phase defect mask were prepared. One is an EUV blank with programmed bump phase defects to analyze the inclination angle of the phase defects. The other is a programmed phase defect EUV mask that consists of bump phase defects of 80-nmwide and 2.4-nm-high with half-pitch 88 nm lines-and-spaces pattern. The positions of the phase defects relative to the absorber lines were designed to be shifted accordingly. Transmission electron microscope observation revealed that the bump phase defects starting from the substrate surface propagated through the multilayer toward the center of the substrate as they reach the top surface of the multilayer. At the distances of 0 and 66 mm from the substrate center, the inclination angles varied from 0 to 4 degrees. The impacts of the inclination angles on EUV microscope images were significant even though the positions of the phase defect relative to the absorber line, as measured by a scanning prove microscope, were same.

EUV lithography (EUVL) is well known to be a strong candidate for next generation, single exposure, sub-30nm half-pitch lithography. Much progress relevant to EUVL has been reported for a decade, however, many issues continue to challenge implementation for volume production.^[1,2] On the other hand, it seems that the coat develop track process remains very similar and in many aspects returns to KrF or ArF dry process fundamentals, but in practice 26-32nm pitch patterning coat develop track process also has challenges with EUV resist. As access to EUV lithography exposures has become more readily available over the last five (5) years, several challenges and accomplishments in track processing have been reported, such as the improvement of ultra-thin film coating, CD uniformity, defectivity, line width roughness (LWR), and so on.^[3,4,5,6] The coat-develop track process has evolved along with novel materials and metrology capability improvements. By coating ultra-thin under layers and resist films and by controlling resist dissolution, the SOKUDO DUO coat develop track system at IMEC (Leuven, Belgium), with ASML NXE3100 exposure, has been used to demonstrate improved CD uniformity, LWR, and defect control. Additionally, we will show the latest lithographic results obtained by novel processing approaches in EUV coat develop track system.

Over the past several years, people have accomplished a great deal of developing the Extreme-ultraviolet lithography (EUVL) technologies. However, several problems which disturb the mass-production of EUVL still remain. One of the problems is the defect control. In order to protect the mask from defect, the usage of the pellicle is essential. However the transmission loss caused by contamination can lead to the pattern error. Therefore it is necessary to find the acceptable thickness of the contamination layer that would cause the image error. The protection ability of the pellicle in terms of critical dimension variation is studied. Our study indicated that the process latitude difference is small enough to ignore whether the pellicle is used or not. In addition the protection ability of pellicle is good against the case of conformal contamination in terms of CD variation.

Resists, underlayers, and new rinse processes were evaluated for negative tone development (NTD) using extreme ultraviolet (EUV) lithography. The most recently developed resists show resolution and sensitivity improvements. High remaining-film thickness was also achieved for better etching resistance. The underlayers smoothed the line width roughness (LWR) and prevented pattern collapse. In addition, the proposed NTD-compatible rinse process further assisted to prevent pattern collapse. The best NTD performance at EIDEC till date was achieved: 22 nm line and space (L/S) resolution, 5.4 nm LWR, and 16.8 mJ/cm² sensitivity with annular illumination for a small-field exposure tool (SFET). Furthermore, an ultimate resolution of 17 nm L/S was achieved with x-dipole illumination of SFET. The lithographic performance of the best NTD resist is comparable to the typical positive tone development resist.

We report the effect of the mesh support for the EUV pellicle on the wafer pattern image. The intensity distribution passing through the meshed pellicle was simulated with a partially coherent EUV beam showing that its non-uniformity and the CD uniformity are increased with the mesh width. In order to reduce a non-uniformity of the intensity distribution and CD uniformity, the mesh width should be narrower and the height becomes smaller as well. Thus, the image deformation on the wafer due to the mesh can be avoided by optimizing the mesh structure and thus the pellicle with the mesh support can be used for the EUV lithography.

Berliner Glas is a privately owned, mid-sized manufacturer of precision opto-mechanics in Germany. One specialty of Berliner Glas is the design and production of high performance vacuum and electrostatic wafer chucks. Driven by the need of lithography and inspection for smaller overlay values, we pursue the production of an ideally flat wafer chuck. An ideally flat wafer chuck holds a wafer with a completely flat backside and without lateral distortion within the wafer surface.

Key parameters in influencing the wafer chucks effective flatness are thermal performance and thermal management, roughness of the surface, choice of materials and the contact area between wafer and wafer chuck. In this presentation we would like to focus on the contact area. Usually this is decreased as much as possible to avoid sticking effects and the chance of trapped particles between the chuck surface and the backside of the wafer. This can be realized with a pin structure on the chuck surface. Making the pins smaller and moving pins further apart from each other makes the contact area ever smaller but also adds new challenges to achieve a flat and undistorted wafer on the chuck. We would like to address methods of designing and evaluating such a pin structure.

This involves not only the capability to simulate the ideal pattern of pins on the chuck’s surface, for which we will present 2D and 3D simulation results. As well, we would like to share first results of our functional models. Finally, measurement capability has to be ensured, which means improving and further development of Fizeau flatness test interferometers.

Extreme ultraviolet (EUV) lithography is one of the most promising technologies for 11 nm node. In this paper, a six-mirror objective system with a higher numerical aperture (NA) 0.5 and a central obscuration was designed with grouping design method. Some key issues about grouping design and control of obscuration were discussed in detail. Design result shows that the size of obscuration is smaller than 30% radius of the pupil and the composite Root-Mean-Square (RMS) wavefront error can reach 0.029 λ (λ=13.5 nm) in a 13 mm×1 mm ring field. Design of this six-mirror objective system provides a potential solution for 11 nm node of EUV lithography.

We simulated the process parameters of a pattern formation reaction that included during-the-exposure and post exposure bake (PEB) processes using an originally developed simulator. From the simulation results, the relationship between process parameters of pattern formation reaction and quencher concentration has been clarified. Moreover, we simulated the present target process parameters of extreme ultraviolet (EUV) resist for breaking the RLS trade-off. In this simulation, the process parameters were calculated from lithographic results (sensitivity, LWR, and CD) using real SEM images. This methodology was used to determine the process parameters required to break the RLS trade-off to obtain the required lithographic target of the EUV resist. We simulated the present lithography performance target using the process parameters of pattern formation reactions. These simulation results showed that a large reaction radius is necessary to break the RLS trade-off. Furthermore, we confirmed that increasing the PEB temperature leads to an improvement in the reaction radius. However, there is a discrepancy between the target radius and the controllable range of reaction radius that can be obtained by varying the PEB temperature.

Extreme-UV lithography (EUVL) is the most promising candidate of next generation technology for hp20nm node device manufacturing and beyond, however there are many critical issues to solve in the light source, tools, mask and photo resist. Regarding the development of a photo resist, it is necessary for high volume manufacturing (HVM) to improve LWR, resolution limit and sensitivity. Additionally, concerning about deterioration of a patterning performance by Out-of-Band (OoB) light existing in the EUV light, and contamination problem of exposure tool due to the photo resist outgassing are the key issues which have to be resolved toward HVM by EUVL. Especially, the outgassing problem can become a significant issue for fine patterning under high dose exposure condition. This paper proposes the novel solution for these critical issues with the application of a top coat material which is named OBPL (Outgassing & out-of-Band Protection Layer) on photo resist. The key characteristics of OBPL material are to have a role in protection against the OoB adverse effect to keep up the photo resist performance, to suppress the outgassing from photo resist as a barrier layer and to enhance the lithographic performance such as photo resist profile and process window. In designing the OBPL material, the optical property needs having not only the high absorbance of DUV (Deep-UV) light in OoB range but also high transmittance for 13.5nm wavelength to prevent the sensitivity loss. Furthermore, it is found that the polymer backbone affects the outgassing barrier capability in previous evaluation. Based on these investigations, a state-of-the-art OBPL achieves quite a positive lithographic result with sufficient OoB absorption and outgassing suppression. Moreover, this material has applicability to all types of photo resist including NTD (Negative-Tone Development) process. This paper describes the result of the feasibility study for OBPL and the lithography performance with EUV full field scanner.

At the Laboratory for Energy Conversion (LEC), ETH Zurich, droplet-based LPP-EUV light sources have been developed since 2007. The main LPP source is ALPS II, which is fully operational since more than one year. The facility is an engineering test stand for long-term e↵ect studies. In order to improve the debris mitigation techniques, it is essential to investigate the droplet plasma dynamics in time and space. Recently a new diagnostic tool based on a multiple array of motorized Langmuir probes has been constructed for this purpose. The detector has been used to map the angular and radial distribution of the ion and electron dynamics around the droplet target. In this paper, some of the experimental results obtained with the new detector are reported. The angular and radial distribution of the ion flux and kinetic energy of the droplet plasma reveals an anisotropic expansion of the ions in terms of kinetic energy and amount of ion charge around the droplet target. These results have been obtained during continuous source operation and for the first time on droplet-based laser produced plasmas.

At the Laboratory for Energy Conversion, ETH Zurich a new tin droplet-based laser-produced plasma source with application in EUV lithography is operational since Q3 2013. The EUV source ALPS II is equipped with a large capacity droplet dispenser and a high power (kW), high repetition rate (>6 kHz) Nd:YAG laser. The new source should address the requirements of high volume manufacturing for different inspection and metrology applications found in EUV lithography. The average source brightness is equal to 350 W/mm²sr. Individual droplet tracking in time and space, which is coupled to a droplet positioning and triggering system helps to increase the pulse-to-pulse EUV emission stability of the source. The lateral droplet stability is on the order of 10-15% of the droplet diameter. The individual droplet triggering yields deviations between the laser trigger and the droplet passage time at the irradiation site of less than 1 us, even for large droplet timing fluctuations (>5%). The in-band EUV radiation is measured with an energy monitor, which is coupled to a fast analog hardware-based integrator. The pulse-to-pulse EUV energy stability for high stability data equals 3% (σ). In the case of window-averaged (0.1 s) data, the EUV stability equals 0.86% (σ). Low stability data is also reported. The large brightness of the presented LPP-based light source can be tuned to adjust the EUV light stability that is required by the inspection tool.

EUV lithography is likely more sensitive to drift from thermal and degradation effects than optical counterparts. We have developed an automated approach to photoresist image-based aberration metrology. The approach uses binary or phase mask targets and iterative simulation based solutions to retrieve an aberrated pupil function. It is well known that a partially coherent source both allows for the diffraction information of smaller features to be collected by the condenser system, and introduces pupil averaging. In general, smaller features are more sensitive to aberrations than larger features, so there is a trade-off between target sensitivity and printability. Therefore, metrology targets using this technique must be optimized for maximum sensitivity with each illumination system. This study examines aberration metrology target optimization and suggests an optimization scheme for use with any source. Interrogation of both low and high order aberrations is considered. High order aberration terms are interrogated using two separate fitting algorithms. While the optimized targets do show the lowest RMS error under the test conditions, a desirable RMS error is not achieved by either high order interrogation scheme. The implementation of a previously developed algorithm for image-based aberration metrology is used to support this work.

As EUV lithography attempts to outperform other lithographical methods to the sub-14 nm node, the demand for a larger NA traditionally dominates the drive for scaling. There are, however, many challenges to overcome in order to accomplish this [1]. Due to the reflective optics in EUV systems, angular effects of oblique illumination, and non-zero chief ray angle at the objective (CRAO), must be carefully considered and will need to be well understood if high-NA EUV is to be successful. This study investigates impact on of the bias between horizontal and vertical feature CD, image placement error and NILS. Effects of sidewall absorber angle, absorption coefficient (k) and absorber thickness are observed through pitch with various source shapes in an EUV lithography system.

Influence of phase defect on printed images of mask pattern was experimentally investigated by printing contact hole (CH) pattern of hp 32 nm on wafer. And the experimental results were compared with the simulation results. A test mask prepared for this experiment contained programmed phase defects of 92 nm ~ 34.8 nm in width and of around 0.68 nm ~ 1.65 nm in depth. The defects were arrayed in a way such that the pitch of the array would differ from the pitch of the absorber contact hole pitch. Therefore, the phase defects were placed at different positions relative to those of the CH patterns. Mask patterns were printed on wafer using an exposure tool NXE3100 with a numerical aperture (NA) of 0.25 and a reduction of 4X. To evaluate the printed patterns affected by the phase defects, circular illumination was employed. The incident angle of mask illumination chief ray was 6 degrees. The printed CH patterns were measured by SEM. An influence of resolution limit of the resist pattern did seem to appear in this experiment, to be a quantitative difference between the simulation and experimental results, the relative location dependence was quite noticeable and the effect of a phase defect was mitigated by covering the defect with an absorber pattern.

The minimum target specificatons of EUV resist material are the resolution < 30nm half pitch C/H, CDU < 3.0nm, and sensitivity < 20mJ. The major pending issue of EUV resist is how to simultaneously achieve high sensitivity, high resolution and low CD Uniformity (CDU). Thus, we have studied that which factors such as acid diffusion, solvents, polymer platform and film density etc are affecting to improve CDU, sensitivity and resolution. Especially, CDU and sensitivity are the main issues among above these performances. With the results of these experiments, we could determine polymer blend PAG as polymer platform for EUV resist material. We have also researched polymer to improve the sensitivity and CDU with variation of molecular weight, poly dispersity and monomer feed ratio. Additionally, we have studied the effects of resist solvents and film density. And we have measured the outgas of our EUV resist. In this paper, we will discuss the results of these studies obtained by EUV tools of SEMATECH.

193 nm inspection for various defect types on top of the extreme-ultraviolet (EUV) mask is studied. The antireflection coating (ARC) is tried to enhance the defect inspection. However, adding ARC is not helpful to increase the sensitivity. Thus, 2 nm TaBO generally used for preventing the oxidation is mainly used. The aerial image deformation caused by the defect is compared to that of the defect free mask. Peak intensity difference is quantized and the sensitivity that is comparable to the ITRS defect inspection limit is chosen. The inspection criterion for typical defect types of extrusion, intrusion, pindot and pinhole is compared.

The Critical Dimension (CD) uniformity due to the defect on the Extreme-Ultraviolet (EUV) pellicle is reported. Based on computational simulation of the aerial images for different defect size on the wafer, it is found that the size of the defect should be smaller than 2 μm for the CD uniformity of 0.1 nm. The aerial image for the different defect materials, sulfur and ruthenium, are also simulated showing that the CD uniformity does not have a noticeable dependence on the different defect materials. However, the CD uniformity is worsened with the mesh structure due to its shadow and the much smaller defects size, less than 2 μm, can be allowed.

The carbon contamination growth (CG) on the witness samples by resist outgassing during exposure were evaluated for the model EUV resist samples having different protecting groups for chemical amplification. Four kinds of different protecting groups were chosen to compare the effects of difference in activation energy for de-protection, the molecular size and polarity of de-protected unit on CG. The residual gas analysis (RGA) measurements were also performed for all samples. Those results were compared between EUV irradiation and e-beam irradiation. On the contrary to the original expectation, it was found that the dependence of the activation energy on CG was small. From the results of RGA, it was confirmed that the size of the protecting group does not also simply correlate with the outgassing amount or CG. In the sample with relatively bigger protecting group we found larger outgassing amount than that with smaller protecting group. The smallest outgassing amount and CG were given by the sample which has the polar de-protecting unit. It is indicating that if there is the interaction between the outgassing molecules and the resist film components, the escaping of the molecules from the resist film out to the vacuum is restricted, resulting in the small outgassing and small CG. All of those features were same in EUV and e-beam irradiation.

EUV masks have replicated multilayer roughness from the substrate or the deposition process which cause line edge roughness (LER) during imaging. We have developed a model, based on the assumption that the roughness is small, that is able to analytically calculate the LER and LER Power Spectral Density (PSD) for any illumination source, defocus, and pitch. We evaluated the model for typical mask roughness values and varied illumination and other parameters to determine how the roughness induced LER behaves under different imaging conditions.

For the introduction of EUV lithography, development of high performance EUV resists is of key importance. This development involves studies into resist sensitivity, resolving power and pattern uniformity. We have used a sub-nanometer-sized 30 keV helium ion beam to expose chemically amplified (CAR) EUV resists. There are similarities in the response of resists to He⁺ ions and EUV photons: both excite Secondary Electrons with similar energy distributions.The weak backscattering of the He⁺ ions results in ultra-low proximity effects. This fact enables the exposure of dense and detailed patterns by focused He⁺ ion beams without the need for proximity correction. This paper presents contact holes and lines at 40-nm pitch in an EUV CAR resist. We have used resist sensitivity, contrast, resolution (CD) and pattern fidelity (LCDU, LWR and dose-to-print) as metrics for a comparison of SHIBL with EUVL. We show that Scanning Helium Ion Beam Lithography (SHIBL) can be a useful and economically attractive technology to (pre-)screen novel EUV resists prior to their final performance evaluation in an EUV scanner.

Particle free handling of EUV reticles is a major concern in industry. For reaching economically feasible yield levels, it is reported that Particle-per-Reticle-Pass (PRP) levels should be better than 0.0001 for particles larger than 18 nm. Such cleanliness levels are yet to be reported for current reticle handling systems. A reticle handler was built based on a modular concept with three uniform linked base frames. In the first stage of the project a dual pod loading unit, two exchange units for opening inner pods and a reticle flip unit are installed on the base frames. In the near future improvements on cleanliness will be tested and particle detection equipment will be integrated. The system will act as a testing platform for clean handling technology for industry.

Patterning uncertainty in EUV lithography arises from each lithographic component: the source, the photomask, the optical system, and the photoresist. All contribute to line roughness and contact disuniformity. In extreme cases, feature variability can result in patterning failures such as line microbridging or random missing contact holes. Historically, redundant contact holes (or vias) were placed to overcome the effects of a missing contact. Due to the aggressive CD shrink of feature size, the use of redundant contacts has been progressively decreased. For some types of devices, almost every contact of the billions found on the chip must be electrically active in order for the device to function. In such scenario, lithographic printing failures may cause catastrophic loss of yield, considering that closed contacts can hardly be corrected by smoothing techniques or etching. In this paper, the minimum contact CD which prints without failure – the contact hole printability limit – is studied for 54nm and 44nm pitch dense arrays. We find that the same resist may show dramatically different printability limits depending upon sizing dose and illumination conditions. This analysis will be implemented to estimate, through simulation-assisted experiments, the required exposure dose and aerial image to safely print sub-30nm contact holes.

This paper describes the development of a laser-produced-plasma (LPP) extreme-ultraviolet (EUV) source for advanced lithography applications in high volume manufacturing. EUV lithography is expected to succeed 193nm immersion double patterning technology for sub- 20nm critical layer patterning. In this paper we discuss the most recent results from high power testing on our development systems targeted at the 250W configuration, and describe the requirements and technical challenges related to successful implementation of these technologies. Subsystem performance will be shown including Conversion Efficiency (CE), dose control, collector protection and out-of-band (OOB) radiation measurements. This presentation reviews the experimental results obtained on systems with a focus on the topics most critical for a 250W HVM LPP source.

The feasibility of wafer-plane measurements of EUV mask surface roughness has been analyzed through stochastic resist simulations at various defocus conditions, for mask surface roughness values ranging between 50 pm and 500 pm rms. With partial coherence of 0.5, NA of 0.25, defocus of 100 nm and mask surface roughness of 50 pm rms, 1.3% of the total resist LER is contributed by the mask surface roughness induced aerial image phase roughness, while 39.1% of the total LER contribution comes from the absorbed photon image. 31.4% of the LER contribution is from the acid image and 27.9% is attributable to the quencher image at the end of the PEB reaction/diffusion processes. For surface roughness values of interest ranging between 50 pm and 150 pm rms, partial coherence of 0.5 and 100 nm defocus, the sensitivity of wafer plane aerial image LER to mask surface roughness is 9.5 nm/nm-rms, while the resist LER sensitivity is 2.9 nm/nm-rms. With hypothetical scaling of the resist parameters, the resist LER sensitivity to mask surface roughness increases to 6 nm/nm-rms.

In EUV lithography, the short wavelength of the light makes the topography of the mask stand out as three dimensional objects rather than thin masks. This generally requires use of a rigorous scattering simulator to calculate the diffracted orders of a mask in order to explain experimental results. In contrast, for optical proximity correction we cannot afford such detailed calculations and we would like to replace such detailed simulations with faster methods that give similar results. In this paper, we discuss observations we made during our printing experiments on a 0.33 NA EUV projection system. In order to extend the process window for non-nested trenches we introduced clear assist features. We observed strong tilt of Bossung curves and best focus shifts for certain pitches. These shifts can be explained by a phase difference between main and assist feature. This effect is very similar for both horizontal and vertical trenches, and it depends strongly on the illumination of the mask. We find that the best focus shift can be minimized for certain assist pitches and illumination conditions, but a general solution for random pitches appears not obvious.

We present a novel approach for wavefront sensing based on scanning diffraction imaging suitable for high-NA optics inspection, where common metrology techniques show limitations. This approach employs ptychography, whereby a well-characterized object is scanned at the focus of the aberrated test optic, and the resulting scat tered light is captured on a CCD. Under the Fresnel approximation, the diffraction patterns are processed in an iterative algorithm to reconstruct the test optic aberrations. We discuss the applicability of this wavefront metrology, present numerical simulations that validate the reconstruction, and show first experimental results from an optical prototype.

We present an update of the AIS wavefront sensor, a diagnostic sensor set for insertion in the upgraded 0.5 NA SEMATECH Albany and Berkeley METs. AIS works by using offset monopole illumination to probe localized regions of the test optic pupil. Variations in curvature manifest as focus shifts, which are measured using a photodiode- based grating-on- grating contrast monitor, and the wavefront aberrations are reconstructed using a least-squares approach. We present results from an optical prototype of AIS demonstrating an accuracy of better than λ/30 rms for Zernike polynomials Z₄ through Z₁₀. We also discuss integration strategies and requirements as well as specifications on system alignment.

EUV is an ongoing industry challenge to adopt due to its current throughput limitations. The approach to improve throughput has primarily been through a significant focus on source power which has been a continuing challenge for the industry. The subject of this paper is to review and investigate the application of SADP (Self aligned double patterning) as a speed enhancing technique for EUV processing. A process with the potential of running a 16 nm self-aligned final etched pattern in less than 10mJ exposure range is proposed. Many of the current challenges with shot noise and resolution change significantly when SADP is used in conjunction with EUV. In particular, the resolution challenge for a 16nm HP final pattern type image changes to 32nm as an initial pattern requirement for the patterned CD.

With this larger CD starting point, the burden of shot noise changes significantly and the ability for higher speed resist formulations to be used is enabled. Further resist candidates that may have not met the resolution requirements for EUV can also be evaluated. This implies a completely different operational set-point for EUV resist chemistry where the relaxation of both LER and CD together combined, give the resist formulation space a new target when EUV is used as a SADP tool. Post processing mitigation of LWR is needed to attain the performance of the final 16nm half pitch target pattern to align with the industry needs.

If the original process flow at an 85W projected source power would run in the 50WPH range, then the flow proposed here would run in the <120WPH range. Although it is a double patterning technology, the proposed process still only requires a single pass through the EUV tool, This speed benefit can be used to offset the added costs associated with the double patterning process. This flow can then be shown to be an enabling approach for many EUV applications.

Laser-produced plasma (LPP) sources for extreme ultraviolet lithography (EUVL) systems utilize CO₂ lasers operating with wavelength 10.6μm. Since multilayer-coated optics have high reflectivity for this infrared radiation (IR), a significant and detrimental amount of IR is passed through the EUVL system. One method to remove the IR from the system is to utilize a binary diffraction grating. When this grating is applied directly to the surface of the primary collector optic of the source, the majority of the IR is diverted outside the radius of the exit aperture at the intermediate focus (IF). This paper will report details on the performance of a full size (410mm diameter) Demonstration Collector utilizing IR rejection (IRR) technology with the capability to produce over 125X suppression of IR, equaling the performance of a IR spectral filter. Additional details will be reported on the technology development and use of a glassy smoothing layer to enable high EUV performance, a weighted average multilayer reflectance of 50.9% for unpolarized EUV radiation.

Lithographic CD printing variability can be easily captured with a CDU measurement, however delineating the most significant sources causing the variability is challenging. In EUV lithography, the resist, reticle, metrology methodology, and stochastics are examples of factors that influence printing variability. Determining the most significant sources of variability in contact hole and via patterning is particularly interesting because the variability can be measured as a function of two tethered dimensions. Contact hole (CH) variability has a direct impact on device performance while via variability affects metal area scaling and design. By studying sources of variability opportunities for improving device performance and scaling can be identified. In this paper, we will examine sources of contact patterning variability in EUV lithography comprehensively using various EUV exposure tools as well as simulation methods. We will present a benchmark of current state of the art materials and patterning methods with the goal of assessing contact hole printability at the limit of 0.33 NA EUV lithography.

A fast rigorous model is developed for the simulation of mask diffraction spectrum in EUV lithography. It combines a modified thin mask model and an equivalent layer method and provides an analytical expression of the diffraction spectrum of mask. Based on this model, we propose a theoretical analysis of the mask shadowing effect. Mathematical expressions for the best mask (object space) focus position and for the required correction of mask pattern size are derived. When the mask focus is positioned in the equivalent plane of the multilayer, the amount of pattern shift is reduced. When the mask pattern size is corrected using the derived formula, taking a space pattern with the target CD of 22 nm as an example, the imaging CD bias between different oriented features is below 0.3 nm.

To evaluate defects on extreme ultraviolet (EUV) masks at the blank state of manufacturing, we developed a micro coherent EUV scatterometry microscope (micro-CSM). The illumination source is coherent EUV light with a 230-nm focus diameter on the defect using a Fresnel zoneplate. This system directly observes the reflection and scattering signals from a phase defect. The scattering distributions of 30-nm-wide defects were anisotropic due to interference with speckle from multilayer scattering. Thus, printability of the defects would depend on the defect position in the multilayer.

Inorganic resists are of considerable interest for advanced lithography at the nanoscale due to the potential for high resolution, low line width roughness (LWR), and high sensitivity. Historically inorganic resists suffered from low sensitivity, however approaches have been identified to increase sensitivity while maintaining high contrast. An aqueous precursor of Hf(OH)_4-2x-2y(O₂)_x(SO₄)_y·qH₂O (HafSOx) has been demonstrated with excellent sensitivity to EUV and electrons, while still obtaining high resolution and low LWR. In this work, we characterize both HafSOx precursor solutions and spin-coated thin films using high-resolution transmission electron microscopy (HR-TEM) with energy-dispersive X-ray spectroscopy (EDS) elemental analysis. HR-TEM of precursor solutions drop cast onto TEM grids confirmed the presence of nanoscale particles. HR-TEM cross sectional images showed that spin-coated HafSOx films are initially uniform in appearance and composition for thin (12 nm) films, however thicker (30 nm) films display segregation of species leading to multilayer structures. Regardless of film thickness, extended exposure to the high energy TEM electron beam induces significant migration of oxygen species to the Si interface. These species result in the formation of SiOx layers that increase in thickness with an increase in TEM electron beam dose. Sulfate is also very mobile in the films and likely assists in the significant condensation exhibited in completely processed films.

Availability of defect-free masks is considered to be a critical issue for enabling extreme ultraviolet lithography (EUVL) as the next generation technology. Since completely defect-free masks will be hard to achieve, it is essential to have a good understanding of the defect printability as well as the fundamental aspects of a defect that result in the defects being printed. In this work, the native mask blank defects were characterized using atomic force microscopy (AFM) and cross-section transmission electron microscopy (TEM), and the defect printability of the characterized native mask defects was evaluated using finite-difference time-domain (FDTD) simulations. The simulation results were compared with the through-focus aerial images obtained at the SEMATECH Actinic Inspection Tool (AIT) at Lawrence Berkeley National Lab (LBNL) for the characterized defects. There was a reasonable agreement between the through-focus FDTD simulation results and the AIT results. To model the Mo/Si multilayer growth over the native defects, which served as the input for the FDTD simulations, a level-set technique was used to predict the evolution of the multilayer disruption over the defect. Unlike other models that assume a constant flux of atoms (of materials to be deposited) coming from a single direction, this model took into account the direction and incident fluxes of the materials to be deposited, as well as the rotation of the mask substrate, to accurately simulate the actual deposition conditions. The modeled multilayer growth was compared with the cross-section TEM images, and a good agreement was observed between them.

Mask blank defectivity remains a challenge in Extreme Ultraviolet (EUV) lithography. One of the mitigation strategies has been to identify the source of particles causing defects in the mask blank deposition tools. Vacuum components like valves, valve seals, stages, filters, etc. could be a possible source of particles in the tools. Therefore, it is necessary to quantify the amount of particles generated by the vacuum components. This feedback to the supplier can be used to help make vacuum components that shed fewer particles. We show results from a valve and nanoparticle particle test system at the College of Nanoscale Science and Engineering (CNSE) in collaboration with SEMATECH. The setup consists of a condensation particle counter (CPC), which can detect particles between 10 nm – 3 um, and a scanning mobility particle sizer (SMPS), which can provide the size distribution of the particles between 10 nm – 280 nm. We show results from testing two different types of 300 mm valves and compare the particle counts per cycle detected by the CPC for both. Moreover, choosing the best operating parameters of the valve can reduce the number of defects generated. We will present the optimized operating parameters. Selection of appropriate valve seal materials for plasma environments can also be crucial for reducing their degradation.