The escalating cost for Next Generation Lithography (NGL) tools is driven in part by the need for complex sources and optics. The cost for a single NGL tool could exceed $50M in the next few years, a prohibitive number for many companies. As a result, several researchers are looking at low cost alternative methods for printing sub-100 nm features. In the mid-1990s, several resarech groups started investigating different methods for imprinting small features. Many of these methods, although very effective at printing small features across an entire wafer, are limited in their ability to do precise overlay. In 1999, Willson and Sreenivasan discovered that imprinting could be done at low pressures and at room temperatures by using low viscosity UV curable monomers. The technology is typically referred to as Step and Flash Imprint Lithography. The use of a quartz template enabled the photocuring process to occur and also opened up the potential for optical alignment of teh wafer and template. This paper traces the development of nanoimprint lithography and addresses the issues that must be solved if this type of technology is to be applied to high-density silicon integrated circuitry.

With the realization of the α-tool, ASML is progressing with the pre-commercialization phase of its EUVL development. We report on the progress in the development of several key modules of the α-tool, including the source, wafer stage and reticle stage, wafer handling, baseframe, and optics modules. We demonstrate that the focus sensor meets its vacuum requirements, and that both stages after limited servo optimization approach the required scanning performance. A particle detection system has been built for the qualification of the reticle handling module, and preliminary results show that 50nm particles can be detected. The optics lifetime program showed substantial progress by utilizing caplayers to MoSi samples in order to suppress oxidation caused by H₂O molecules under EUV illumination: a suppression ≥ 100x is achieved, compared to uncapped MoSi.

While interferometry is routinely used for the characterization and alignment of lithographic optics, the ultimate measure of performance for these optical systems is the transfer of an image or pattern into photoresist. Simple yet flexible exposure systems play an important role in this task because they allow complex system-dependent effects to be isolated from the printing results. This enables the most direct lithography evalaution of the optical system under investigation. To address tehse issues for commercial-class EUV optics, a synchrotron-based programmable illuminator exposrue station has been implemented at Lawrence Berkeley National Laboratory (the Advanced Light Source). As previously presented, this static microfield exposure system has been used to lithography characterize a 4-mirror optical system designed for the EUV Engineering Test Stand (ETS) prototype stepper. Based on the lithographic characterization, here we present a detailed performance analysis of the 0.1-NA ETS Set-2 optic. Operation of the static printing system with the Set-2 optic yielded approximately 330 exposed wafers, where each wafer contains one or more focus-exposure matrices. A wide variety of parameters were studied includign, among others, illumination conditions, resist thickness, and mask tone. Here we present a subset of this data in terms of process-window results. The resutls demonstrate a depth of focus (DOF) approximately 2μm for isolated 70-nm line features, 1 μm for nested 70-nm line features, and 0.5μm for 70-nm contacts on 270-nm pitch.

Full-field imaging with a developmental projection optic box (POB 1) was successfully demonstrated in the alpha tool Engineering Test Stand (ETS) last year. Since then, numerous improvements, including laser power for the laser-produced plasma (LPP) source, stages, sensors, and control system have been made. The LPP has been upgraded from the 40 W LPP cluster jet source used for initial demonstration of full-field imaging to a high-power (1500 W) LPP source with a liquid Xe spray jet. Scanned lithography at various laser drive powers of >500 W has been demonstrated with virtually identical lithographic performance.

Early production EUV exposure tools may have difficulty achieving flare requirements in the 5-6% range for the 32nm technology node. In this case, flare compensation may be needed to achieve the necessary CD control budget for production. This paper explores both experimentally as well as computationally wafer CD compensation though mask CD resizing so that proper CD control across the exposure field can be maintained. Experimental resist data collected on POB#2 of the Engineering Test Stand (ETS) suggest that even a simple linear CD compensation model can signifantly improve CD contorl in the presence of flare variation. Extending a similar concpet to a hypothetical full-field 25×33 mm² mask area containgin 20 nm gate CDs shwos taht CD compensation, while computationally demanding, can be realized through a convolution approach of a 1×1 mm² mask area using a non-uniform adaptive grid.

A new generation of 0.3 numerical aperture prototype EUV optical systems is now being produced to provide an opportunity for early learning at 20-nm feature size. Achieving diffraction limited performance from these two-mirror, annular projection optics poses a challenge for every aspect of the fabrication process, including final alignment and interferometric qualification. A new phase-shifting point diffraction interferometer will be used at Lawrence Berkeley National Laboratory for the measurement and alignment of the MET optic at EUV wavelengths. Using the previous generation of prototype EUV optical systems developed for lithography research, with numerical apertures up to 0.1, EUV interferometers have demonstrated RMS accuracy levels in the 40-70 pm range. Relative to the previous generation of prototype EUV optics, the threefold increase to 0.3 NA in the image-side numerical aperture presents several challenges for the extension of ultra-high-accuracy.

We have developed a high numerical aperture (NA) small-field exposure system (HiNA) for EUV exposure process development. NA of projection optics of EUV exposure tools for 45-nm node lithography is expected to be around 0.25, which is higher than that previously expected (0.1). HiNA has compatible illumination system, which can be switched to partial coherent illumination and coherent illumination by changing some optical elements. Coherent illumination system was prepared for a high contrast imaging but the uniformity of intensity is less than that of partial coherent illumination. A reflected-type fly*fs-eye element was adopted for partial coherent illumination, which can provide uniformity of both coherency and intensity simultaneously. The coherency of the partial coherent illumination is 0.8. HiNA projection optics consists of two aspheric mirrors, with the NA and the imaging field of 0.3 and 0.3×0.5mm², respectively. We fabricated two sets of projection-optics. Although the wavefront error of set-1 optics was 7nmRMS, that of set-2 optics was improved to 1.9nmRMS, which was measured with a point diffraction interferometer (PDI) using He-Ne laser. The wavefront error of the set-2 optics was improved by using a new mirror mount mechanism. The mount system consists of several board springs made of super invar in order to minimize the deformation of mirrors by mounting stress. The projection optics of the set-2 has a remote controlled mirror adjustment mechanism which has five degrees of freedom (X,Y,Z,X-Tilt and Y-Tilt). The position of the concave secondary mirror was adjusted precisely with measuring the wavefront error using PDI.

Static and scanned images of 100nm dense features for a developmental set of l/14 optics (projection optics box # 1, POB 1) in the Engineering Test Stand (ETS) were successfully obtained with various LPP source powers last year. The ETS with POB1 has been used to understand initial system performance and lithographic learning. Since then, numerous system upgrades have been made to improve ETS lithographic performance to meet or exceed the original design objectives. The most important upgrade is the replacement of POB 1 with an improved projection optics system, POB2, having lower figure error (l/20 rms wavefront error) and lower flare. Both projection optics boxes are a four-mirror design with a 0.1 numerical aperture. Scanned 70-nm dense features have been successfully printed using POB2. Aerial image contrast measurements have been made using the resist clearing method. The results are in good agreement to previous POB2 aerial image contrast measurements at the subfield exposure station (SES) at Lawrence Berkeley National Laboratory. For small features the results deviate from the modeling predictions due to the inherent resolution limit of the resist. The intrinsic flare of POB2 was also characterized. The experimental results were in excellent agreement with modeling predictions. As predicted, the flare in POB2 is less than 20% for 2μm features, which is two times lower than the flare in POB1. EUV flare is much easier to compensate for than its DUV counterpart due to its greater degree of uniformity and predictability. The lithographic learning obtained from the ETS will be used in the development of EUV High Volume Manufacturing tools. This paper describes the ETS tool ETS tool setup, both static and scanned, that was required after the installation of POB2. The paper will also describe the lithographic characterization of POB2 in the ETS and cmpare those results to the lithographic results obtained last year with POB1.

We demonstrate direct flare measurements of 4-mirror projection optics in the Engineering Test Stand (ETS) using a conventional resist clearing method (the Kirk method). Two extreme UV lithographic projection optics, one with higher flare than the other, have been characterized and the results compared. The measured results have also been compared to analytical calculations based on measured mirror roughness and the extended point spread function. Full-field flare across the 24 mm field width has been measured, and we have verified that flare is constant across the field for EUV lithography as predicted. Horizontal (H) and vertical (V) flare bias has been observed and the cause of the H-V flare bias has been investigated. The main cause has been identified to be anisotropic mirror polishing. Simulations with the 2D Power Spectral density function have confirmed the experimental results.

The paper describes recent progress on the development of an EUV source based on a hollow cathode triggered gas discharge (HCT). The principle of operation has been described in previous publications. When operated with Xe, a repetition frequency up to 4 kHz, conversion efficiency of 0.55% inband radiation in 2π and a pinch size below 3mm in length was demonstrated. Today's requirements on a commercial EUV source for volume production of wafers still exceed the current performance by large factors both in terms of output power and life time. This paper will discuss the roadmap to high power and will also show elements of the way to extended life time. Particular focus will be put onto the physical limits of Xe as radiator and the advantages of using Sn instead. It will be demonstrated that the spectral efficiency of Sn is a factor of 3 higher than Xe.

Semiconductor chip manufacturers are expecting to use extreme UV lithography for production in 2009. EUV tools require high power, brilliant light sources at 13.5 nm with collector optics producing 120 W average power at entrance of the illuminator system. Today the power and lifetime of the EUV light source are considered as the most critical issue for EUV lithography. The present paper gives an update of the development status of EUV light sources at XTREME technologies, a joint venture of Lambda Physik AG, Goettingen, and Jenoptik LOS GmbH, Jena, Germany. Results on both laser produced plasma (LPP) and gas discharge produced plasma (GDPP), the two major technologies in EUV sources, are given. The LPP EUV sources use xenon-jet target systems and pulsed lasers with 400 W average power at 10 kHz developed at XTREME technologies. The maximum conversion efficiency form laser power into EUV in-band power is 0.75% into 2π solid angle. With 300 W laser average power at 3300 Hz repetition rate up to 1.5 W EUV radiation is generated at 13.5 nm. After a collector of 5 sr this corresponds to 0.6 W in intermediate focus without spectral purity filter and 0.5 W in intermediate focus with spectral purity filter. The direct generation of the EUV emitting plasma from electrical discharges is much simpler than LPP because the electrical energy has not to be converted into laser radiation before plasma excitation. XTREME technologies' Xenon GDPP EUV sources use the Z-pinch principle with efficient sliding discharge pre-ionization. The plasma pinch size and the available emission angle have been matched to the etendue of the optical system of 2-3 mm2 sr, i.e. no additional etendue related loss reduces the usable EUV power from the source. In continuous operation at 1000 Hz the GDPP sources emit 50W into 2π solid angle are obtained from the Z-pinch sources. Spatial and temporal emission stability of the EUV sources is in the range of a few percent. Debris shields for EUV sources have been developed which give improvement of the collector optics lifetime by several orders of magnitude.

A radiation-magnetohydrodynamic (MHD) model of gas discharges has been developed to accelerate the development of compact, intense sources of EUV radiation for microlithography. The model is an MHD numerical simulation with atomic and radiation physics. The plasma evolution is simulated with the MHRDR (Magneto-Hydro-Radiative-Dynamic-Research) 2D, three-temperature, MHD computer code. The MHD results are postprocessed with a code that calculates the 13.4-nm EUV radiation output from Xe ions, based on a detailed collisional-radiative atomic kinetics model. Modeling of a dense plasma focus discharge, in a Xe-He gas mixture, has been initiated with this new tool.

A new direct discharge source of 13.5nm radiation addresses the heat load problem by creating the plasma remote from all surfaces. The plasma is initially formed at the intersection of many pulsed xenon beamlets. Further heating is then applied via a high current pulse to induce efficient radiation from Xe¹⁰⁺ ions. The plasma is compact, with a single pulse FWHM diameter of 0.7mm and length of 3mm. It is positionally stable, as illustrated by re-imaging onto a fluorescent screen sensitive to EUV and time-integrating over 250 pulses. In this mode the averaged FWHM is 0.9mm. The conversion efficiency from stored electrical energy to radiation within 2π sterad and 2% bandwidth at 13.5nm is currently 0.55%, using xenon. Power is delivered to the plasma by a solid state-switched modulator operated at a stored energy of 25J of which 10J is dissipated in the plasma plus circuit, and 15J is recovered. The EUV output in 2% bandwidth at 13.5nm is 9mJ/sterad. Repetition rate scaling of the star pinch EUV source to 1kHz there is negligible electrode erosion at 10⁶ pulses. This is possible because the cathode for the main heating discharge is distributed into 24-fold parallel hollow cathodes, with a combined operational surface aera of approximately 20cm². The anode is similarly distributed. The walls facing the plasma are 22mm distant from it and when scaled to 6kHz will see a heat load of less than 1kWcm^-2. The cathode electrode is then expected to receive a heat load of less than 500W cm^-2. The plasma is expected to clear between pulses and be reproducible at frequencies up to at least 10kHz, at which rate the usable EUV power available at a second focus, assuming colleciton in 2sterad, is predicted to be more than 80W. The star pinch has properties that favor long life and appears to scale to the 50-100W powers needed for high throughput lithography.

Debris-free generation of a tin plasma was demonstrated in the cavity-confined configuration. Narrow band emission at 13.7-nm was observed in an emission spectrum of a cavity confined tin plasma. The spectral efficiency was as high as 12% and we found the conversion efficiency could reach 6%/2π str ultimately while lots of works are required to achieve this value. We also confirmed a magnetic field has some effect of stopping a plasma.

We have previously demonstrated (Proc. SPIE Vol. 4688, 595, 2002; JM3 Vol. 2, Nr. 1, 34, 2003) a depth of focus of several millimeters in masked ion beam lithography (MIBL). This work illustrates the use of this capability to pattern concave and convex spherical substrates 25 mm in diameter and 5.5 mm deep. Features as small as 175nm were printed across a convex substrate with 75 keV He⁺ ions where the mask-to-wafer gap varied from 1 mm on the center to 6.5 mm on the edge. We also demonstrate, for the first time, the fabrication of 800 nm features in a 4 nm thick Cr coating by masked ion beam direct structuring (MIBS), where 45 keV Argon ions sputter the film through a stencil mask. We demonstrate electronic adjustment of the radial beam profile of the incident ion beam to compensate for the continuously changing angle between the beam and the spherical substrate, thus achieving constant exposure dose (for MIBL) or sputtering rate (for MIBS) a 50 × 50 mm² field.

We present a new methodology to fabricate surface structures based on coated magnetic particles and charged polymers. We control the surface architecture via a nanolithographic approach, Dip-Pen Nanolithography, which utilizes Atomic Force Microscopy (AFM) to delivery molecular "inks" on soid substrates. This process results in surfaces with a pre-programmed architecture composed of well-characterized negatively and postively charged polymeric regions. Charged magnetic particles coated via the Layer-by-Layer method can be assembled onto the pre-programmed surface blue-print using electrostatic interactions. The possibility of incorporating our templates in simple device strategies is currently being explored.

An electron beam lithography method for printing and plating sub-50 nm isolated trenches with a high aspect ratio has been developed for the nanofabrication of magnetic thin film heads. To eliminate the issues of resist footing and resist residue in the narrow trench process, we put a thin dissolution layer of polymethylglutarimide (PMGI) as an undercoat layer between a seed layer and a resist layer. The undercoat dissolution layer competely cleared off the seed layer by the developer solution such that the sides of the narrow trench are vertical, particularly at the bottom of the narrow trench, thus facilitating plating the narrow trench with a high magnetic moment material. In this work, the narrow trenches were electroplated with both 1.0T NiFe and 1.8T CoNiFe. Three key issues in our trench process will be discussed here, including: 1) critieria for the selection of the undercoat dissolution layer materials; 2) processing conditions control , e.g. the thickness and the bake temperature of the dissolution layer to achieve vertical and smooth sidewalls; and 3) PEB delay on the narrow trench CD control, pattern degeneration, and the results from the resist top coat (RTC) experiments. With our new narrow trench process, we demonstrated the capability of fabricating narrow electrodeposited magnetic write structures with a CD of 35 nm in 0.35 μm resist (AR=10:1) and a CD of 30 nm in 0.25 μm resist (AR=8:1).

Molecular Imprints, Inc. (MII) has developed the ImprioTM 100, which is the first commercial step and repeat imprint lithography system with field-to-field alignment. This system is designed to implement the UV curable nano-replication capability of the Step and FlashTM Imprint Lithography (S-FILTM) process. To-date, the Imprio 100 system has demonstrated: 1) Full 200 mm wafer coverage with lithographically useful patterning; 2) Full wafer residual thickness control to enable practical etching (thickness variation < 50 nm, 3 sigma); 3) Field edge control compatible with 50 um kerf regions. 4) Multi-day CD uniformity measured on an analytical SEM < 2 nm, 3 sigma with no process adjustments; 5) Etch pattern transfer including break-through etch of residual material, followed by a bi-layer etch through thick planarization layers; 6) Initial level-to-level alignment target acquisition with accuracy of better than 100 nm. 7) Low air borne particle counts in tool microenvironment consistent with Class 0.1 while imprinting.

Step and Flash Imprint Lithography (SFIL) is one of several new nano-imprint techniques being actively developed. While SFIL has been shown to be capable of sub-30 nm resolution, critical dimension (CD) control of imprinted features must be demonstrated if SFIL Is to become a viable and production worthy lithography technique. In the current study, a Molecular Imprints Imprio-100 system was used to imprint resolution patterns on 200 mm wafers. A characterization of critical dimension uniformity over the all-quartz template was done and compared to the same features printed on wafers. This analysis was performed for 100, 80, 50, and 30 nm features in three ways: over a single die using 64 sites arrayed across a 21 mm field, from field-to-field for 37 die across a single wafer, and from wafter-to-wafer for six wafers. Results show that CD's transfer from template to wafer with a slight positive bias which is greatest for 50 and 30 nm line sizes. Feature profiles studies. Despite this, the maximum calculated component of process variation from the SFIL process itself was calculated to be only 6 nm.

Recent work on Step and Flash Imprint Lithography (SFIL) has been focused on process and materials fundamentals and demonstration of resolution capability. Etch transfer rpocesses have been developed that are capable of transferring imprinted images though 150 nm of residual etch barrier, yielding sub 50 nm lines with aspect ratios greater than 8:1. A model has been developed for the photoinitiated, free radical curing of the acrylate etch barrier materials that have been used in the SFIL process. This model includes the effects of oxygen transport on the kinetics of the reaction and yields a deeper understanding of the importance of oxygen inhibition, and the resulting impact of that process on throughput and defect generation. This understanding has motivated investigation of etch barrier materials such as vinyl ethers that are cured by a cationic mechanism, which does not exhibit these same effects. Initial work on statistical defect analysis has is reported and it does not reveal pathological trends.

To utilize the potential of nanoimprint lithography (NIL) you need polymers, which give relief patterns with good thermal and etch resistance, a necessity for subsequent process steps. Thermoplastic polymers with high thermal stability require high imprint temperatures. Such temperatures can cause polymer degradation and problems with pattern transfer due to the different coefficients of thermal expansion of substrate, polymer and stamp. The characteristics and benefits of two types of cross-linking prepolymers with low glass transition temperature (Tg) for nanoimprinting are described. They are soluble in organic solvents and their solutions can be processed like those of poly (methyl methacrylate) (PMMA). The imprinted patterns receive high thermal and mechanical stability through cross-linking polymerization and exhibit high plasma etch resistance. The course of the polymerization was investigated to determine the appropriate conditions for the imprint process. In thermally cross-linking allyl polymers, the cross-linking occurs during imprinting. Process time and temperature depend on the polymerization rate. Volume shrinkage during the polymerization does not adversely affect imprinting. Photochemically cross-linking epoxy-based polymers permit imprint temperatures below 100°C and short imprint times. Tg of the prepolymer determines the imprint temperature. The cross-linking reaction and structural stabilization is performed after imprinting. SEM images demonstrate the realization of the cross-linking polymer approach. Isolated lines down to 50 nm width confirm the successful application of the polymers.

One of the key questions concerning the concept of a system for hot embossing lithography is whether or not it should provide for imprinting under vacuum. We have performed experiments comparing the embossing in vacuum and in atmospheric pressure in a semi-automated imprint system. The stamps used were fully patterned, 10 cm diameter with pattern sizes ranging from 400 nm to 100 μm. It turned out that vacuum enhances the large area uniformity of the imprint by avoiding an air cushion remaining between stamp and sample during automated contact after a non-contact assembly and alignment step. Lower molecular weight polymers turned out to be more sensitive to uniformity deviation than higher molecular weight materials. Detailed analysis showed that defects typically found for relatively high processing temperatures, caused by overheated compressed air, remaining solvent in the polymer layer or even beginning polymer decomposition could be reduced substantially under vacuum embossing conditions, where the excess volume of the polymer is evacuated and free to accommodate gaseous constituents. The best result with complete cavity filling and negligible defects was obtained for imprint of a 99 kg/mol polymer at 200°C and 50 bar under vacuum. Residual layers measured across the diameter of the sample were 44.5 nm ± 9.8 nm. The non-uniformity of the residual layer is a result of the locally different pattern sizes and pattern densities of the stamp, typical for all mechanical patterning processes.

The enhanced demands for substrate materials for next-generation optics and masks have initiated detailed investigations on Zerodur as a proposed EUVL substrate material with focus on thermal expansion behavior and surface roughness. As a result of specifically adjusted process parameters, the coefficient of thermal expansion (CTE) was tailored to be a minimum at 22.5°C. Laboratory samples of Zerodur exhibit a CTE < 5 ppb/K corresponding to the lowest expasnion class of the SEMI standard P37 (19 to 25°C) for EUV mask blanks. By further variation of process parameters, the position of zero crossing, e.g. at 30°C, can be varied, revealing an attractive attribute feature of Zerodur. A new dilatometer type was mounted in 2002 with first operatinoal results revealing an improved reproducibility of ~1ppb/K in the temperature range of 0 to 50°C. A series of CTE measurements with a small block of Zerodur provides information on CTE homogeneity on a cm-scale: No CTE variation was observed within the error of measurements for a block exhibiting ± 3.5*10^-6 vairtion in refractinve index. CTE variation can cause surface deformations during changing temperature conditions. A first setup of Fizeau-Interferometer with a current resolution of 0.3 nm rms was used to record surface deformation of Zerodur due to elevated temperature was determined to be lower than the current resolution. Both methods to analyze the CTE homogeneity of Zerodur lead to the result of CTE variation below 1 ppb/K, still identifying today's need to improve metrology further. Final finishing of EUV substrates may increase roughness of super-polished surfaces significantly. Using appropriate processes a to surface roughness < 0.25 nm rms under production conditions can be achieved after final finishing of Zerodur. As an improved Zerodur-type material, recent achievements of material development demonstrate the optimization of glass-ceramic composition to nearly meeting the specification of surface roughness after a standard finishing process. These results are regarded as a promising milestone to develop an optimized glass-ceramic material providnig adjusted thermal expansion behavior and surface processability according to the specific demands of EUV technology.

Corning Incorporated is improving material and metrology in order to meet the requirements for both EUVL optics and photomask substrate applications. The EUV optics requirements present a unique challenge to the lens designer. The temperature of each optic in the system can experience a different thermal profile based upon the geometry of the element and the intensity of the beam at each element location. This places a need on the optical material for small variation in the coefficient of thermal expansion (CTE) uniformity and the ability to achieve targeted optimum zero CTE cross-over temperatures. This paper addresses Corning’s ability to target specified CTE values as well as discusses a new metrology tool for measuring CTE variations within the glass. Past data suggested that index variation within the material were related to CTE variations. This correlation was investigated with the results presented here. This preliminary work suggests a new metrology tool with the capability of non-destructively measuring peak to valley (P-V) CTE variations to within 70 parts per trillion per degree Kelvin (ppt/K) at possible spatial frequencies in the micron range on thick or thin samples. This technique is vital for certifying photomasks and will be the foundation needed to reduce CTE variations in photomasks and optics to targeted values of less than 1 ppb/K for future EUVL needs.

The reflectance stability of multilayer coatings for extreme ultraviolet lithography (EUVL) in a commercial tool environment is of utmost importance to ensure continuous exposures with minimum maintenance cost. We have made substantial progress in designing the protective capping layer coatings, understanding their performance and estimating their lifetimes based on accelerated electron beam and EUV exposure studies. Our current capping layer coatings have about 40 times longer lifetimes than Si-capped multilayer optics. Nevertheless, the lifetime of current Ru-capped multilayers is too short to satisfy commercial tool requirements and further improvements are essential.

We fabricated molybdenum (Mo)/silicon (Si) multilayers with a low-pressure rotary magnet cathode (RMC) sputtering system under several deposition conditions and investigated their properties. We obtained a high extreme UV (EUV) reflectivity of approximately 71% as measured in Mo/Si multilayers sputtered with low-pressure xenon (Xe) gas. We confirmed that the multilayers sputtered with Xe gas exhibited higher reflectivity than those sputtered with argon (Ar) gas, and that the multilayers sputtered at a lower pressure exhibited higher reflecitivty than those sputtered at a higher pressure. From trans mission electron microscope (TEM) cross-sectional images of the high-reflectivity Mo/Si multilayers, we observed thinner interdiffusion layers between the Mo and Si layers than those in Mo/Si multilayer where a 50-pair multilayer with a Γ ratio (the fractional thickness ratio of a Mo layer to the total thickness of a Mo layer and a Si layer) of 0.35 with compressive stress was stacked upon a 30-pair multilaeyr with a Γ ratio of 0.7 with tensile stress.

An experimental extreme UV (EUV) interferometer (EEI) using an undulator light source was designed and constructed for the purpose of developing wavefront measurement technology with the exposure wavelength of the projection optics of EUV lithography systems. EEI has the capability of performing five different EUV wavefront metrology methods.

The development of EUV lithography is critically based on the availability of suitable metrology equipment. To meet the industries requirements the Physikalisch-Technische Bundesanstalt (PTB) recently has installed a new EUV reflectometer at the electron storage ring BESSY II. The new reflectometer is designed for at-wavelength metrology of full-size EUVL optics. Samples with a maximum weight of 50 kg and a diameter of up to 550 mm can be investigated. Besides wavelength and angle scans also the measurement of bi-directional scattering is possible within the full sample surface. Convex and concave shaped surfaces are allowed. Not only a single mirror of the projection optics but also up to five masks can be mounted simultaneously. For future lithography production tools the requirements for the optics and masks are very stringent. The homogeneity of the multilayer reflectivity across the surface and the wavelength matching of the peak reflection become even stronger than today. To meet the increasing demands not only regarding the sample size but also regarding the accuracy of the measurements the operation of the beamline was further optimized. Diffuse scattered light limits the uncertainty in the peak reflectance. A total relative uncertainty of 0.14% is achieved with a reproducibility of 0.07%. The uncertainty in the center wavelength is mainly given by the uncertainty for the reference wavelength of the Kr 3d5/2-5p resonance. The reduction of all other sources of uncertainty results in a total uncertainty of 0.014% in the center wavelength with a reproducibility of 0.008%. We present a detailed description of the EUV reflectometer and discuss the optimized beamline conditions with the different sources of uncertainties. The results are illustrated by recent measurements.

It is well known that molybdenum and silicon is the combination with the highest EUV reflectivity of two-component multilayers in the wavelength range lambda=12.5-20nm. Using the magnetron sputter deposition method multilayers with reflectivities of typically 69% can be prepared. A further increase to R=70% was demonstrated by the introduction of tiny barrier layers (C and/or B4C) between the Mo and Si layers, which reduce the interdiffusion of both chemical elements. However, still higher reflectivities are desirable for the use of the multilayers as reflectors for EUV lithography. From model considerations we have concluded that the replacement of the Mo absorber layer by a multi-component layer consisting of two or three layers could result in an EUV reflectivity increase of up to 0.5% compared to the pure Mo/Si system assuming sharp interfaces without any roughness. Particularly Ag and Ru are promising candidates as additional elements within the absorber. Therefore we have systematically changed the thicknesses of the individual layers under the boundary condition of a fixed period thickness of dp=(6.90±0.05)nm. Microstructure and optical parameters of the multilayers have been investigated by HRTEM, X-ray diffractometry and Cu-Ka reflectometry. The most promising multilayers have also been characterized by EUV reflectometry.

Mo/Si multilayer mirrors with as well an increased as a reduced bandwidth in their spectral and angular reflectance have been designed and deposited for the wavelength about 13.5 nm. For the broadband mirrors, a non-periodic multilayer design based on the thickness optimization of each layer by a stochastic method is compared to a desing that consists of 3 different monoperiodic stacks. In addition, narrowband multilayer mirrors with a significantly reduced bandwidth based on high order reflection have been designed and fabricated. A near-normal peak reflectivity of more than 20% and 30% were achieved for the broadband mirros in the wavelength range from 13 nm to 15 nm and the incidence angles from 0° to 20°, respectively. The decrease of FWHM for Mo/Si mirrors from 0.5 nm to 0.07 nm was shown for suggested narrowband design. Both the increase and the reduction of the reflection bandwidth are unavoidably connected with a decrease of peak reflectivity. Therefore, the application of such mirrors involves areas where a maximum peak reflectivity is not required, e.g. in EUV spectroscopy, radiation filtration and for the characterization of EUV sources. Furthermore, the use of such mirrors in combination with a broadband plasma source will result in a higher integral reflectivity.

We are developing an ultra high precision Soft X-ray telescope. The design of the telescope is a normal incident one for 13.5 nm band using Mo/Si multilayers. Two ideas are introduced. One is the optical measurement system in order to monitor the prevision of the optics system. The other is the adaptive optics system with a deformable mirror. Using an x-ray optical separation filter, we can always monitor the deformation of the optics by optical light. With this information, we can control the deformable mirror to compensate the system deformation as a closed loop system. We confirmed that the absolute precision of the wave front sensor was less than 3 nm rms. The preicison of the deformable mirror was roughly 5 nm rms. The shape of the primary mirror was an off-axis paraboloide with an effective diameter of 80mm. This primary mirror was coated by Mo/Si multilayers. The reflectivity of the primary mirror at 13.5 nm was rnaging from 30 to 50%. The x-ray optical separation filter was made from Zr with a thicknness of ~170nm. The transmission of the filter for low energy x-ray was measured and was roughly 50% at 13.5nm.

Rigorous electromagnetic scattering simulation is used to characterize mask diffraction of various EUVL mask structures. Planar multilayer masks with three kinds of absorber stacks and new subtractive (binary, att-PSM, alt-PSM) etched multilayer masks are studied for lithography performance for line and space features in the 45 nm node. The simulation results for process window show that the combined depth of focus of the Cr/SiO₂ absorber stack masks is about 320 nm maximum with sigma of 0.8 when mask CD biases are applied. The biased etched binary mask has a higher DOF of 360 nm at sigma = 0.6. The line width variations caused by the absorber thickness changes are shown to be about 2 nm due the interference effects and image placement errors (IPE) are 4-7 nm with a 0.4 nm variation. The consideration of a full 2% EUV illumination bandwidth decreases the interference effects by 50%. The etched binary mask is not affected by such phenomenon and provides better contrast. The H-V bias and IPE results of absorber stacks show a linear dependence of absorber thickness. The etched binary multilayer mask has less H-V bias and IPE due to its smaller effective thickness. A simple alternating PSM structure created by the subtractive multilayer mask patterning technique is also proposed and it shows a larger 722 nm DOF.

Phase Shift Masks (PSM) for Extreme Ultraviolet Lithography (EUVL) have the potential for extending the lithographic capability of EUVL beyond the 45-nm node. Typical PSM structures, such as for attenuated PSMs (Att-PSMs), are similar to those of binary masks in the sense that patterned structures of one or more layers of absorber (attenuator) are constructed on the EUV multilayer mirror to provide the correct amount of attenuation and phase shift. However, another type of PSM implemented by etching into the Mo/Si multilayers, rather than by adding lithographic structures on top of the Mo/Si multilayers (additive approach) can provide the required phase shift for both attenuated and hard PSMs. One of anticipated technical challenges, i.e. terminating ethcing at a specific depth with good surface uniformity can be sovled by employing an etch stop layer (ESL) embedded at a target depth inside the multilayer. In designing PSMs using this subtractive fabrication technique, the position and thickness of the ESL should be optimized, so that optical function of the multilayer substrate with embedded ESL should be same or close to when it does not have any embedded layer. According to simulation, the print bias for PSMs by etching into the multilayer stack to create the phase shift is smallest and near ideal compared to other types of PSMs or binary masks fabricated by conventional methods. The increase of depth of focus by 25-75% for contacts using an attenuated PSM and by 50-100% for lines using an ideal har dPSM is another lithographic advantage as well. The design and method of fabricating PSMs by etchign intothe multilayers is described, which include the optimziation of the thinkness and dpeth of the embedded layer. Experimental results of the multilayer etch process demonstrate initial feasibility of the subtractive approach to fabricating EUV PSMs.

The development of defect-free reticle blanks is an important challenge facing the commercialization of extreme ultraviolet lithography (EUVL). The basis of an EUVL reticle are mask blanks consisting of a substrate and a reflective Mo/Si multilayer. Defects on the substrate or defects introduced during multilayer deposition can result in critical phase and amplitude defects. Amplitude- or phase-defect repair techniques are being developed with the goal to repair many of these defects. In this paper we discuss the selection of a capping layer for amplitude-defect repair, and report on experimental results of the reflectance variation over the amplitude-defect repair zone for different capping layers. Our results suggest that carbon and silicon carbide are the leading candidates for capping layer materials. We further performed a quantitative assessment of the yield improvement due to defect repair. We found that amplitude- and phase-defect repair have the potential to significantly improve mask blank yield, and that yield can be maximized by increasing the number of Mo/Si bilayers.

Fabrication of EUVL masks requires formation of both a repair buffer layer and an EUV absorber layer on top of a molybdenum/silicon (Mo/Si) multilayer coated mask blank. Alteration of the Mo/Si multilayer during etch, repair or cleaning of the EUVL mask can be detrimental to the reflectivity and thus the functionality of the final mask. IBM’s Next Generation Lithography (NGL) group has reported on EUVL mask fabrication based on an absorber of low stress chromium (Cr) and a buffer layer of silicon dioxide (SiO₂). Due to poor etch selectivity between SiO₂ and the underlying silicon capping layer, the finished masks had non-uniform and reduced EUV reflectivity after processing. This led to the development of an alternative absorber stack combination of an absorber layer of low stress TaN_x on a buffer layer of low stress Cr. This paper describes the improved reflectivity uniformity of this type of mask along with several aspects of mask quality, such as CD control and image placement.

The use of a reflective mask in extreme ultraviolet (EUV) lithography requires illumination of the mask at a small but obliquely incident angle. This results in an inherent asymmetry in the diffraction pattern from the mask. This asymmetry results in a shift of the image at the wafer. The image shift depends on the mask structure and the exposure condition. The thicknesses of the patterned mask layers, specifically, the absorber and anti-reflection layer (ARL), are particularly important for the image formation. A rigorous approach to simulating these effects is required. In this paper, a fast frequency-domain method for simulating EUV masks with the underlying multilayer is used for the study of various mask variation effects on image placement as well as linewidth, for nested 35 nm and 25 nm lines. Results indicate that EUV radiation reflected from the mask before entering the multilayer cannot be neglected. Consequently, it is of utmost importance to maintain better than ±1 nm control of absorber thickness and ARL thickness over the entire mask, in order to minimize the CD impact of variations in reflectance swing.

Material selection and lifetime issues for extreme ultraviolet (EUV) lithography are of critical importance to the success of this technology for commercial applications. This paper reviews current trends in production and use of plasma-facing electrodes, insulators, and wall materials for EUV type sources. Ideal candidate materials should be able to: withstand high thermal shock from the short pulsed plasma; withstand high thermal loads without structural failure; reduce debris generation during discharge; and be machined accurately. We reviewed the literature on current and proposed fusion plasma-facing materials as well as current experience with plasma gun and other simulation devices. Both fusion and EUV source materials involve issues of surface erosion by particle sputtering and heat-induced evaporation/melting. These materials are either bare structural materials or surface coatings. EUV materials can be divided into four categories: wall, electrode, optical, and insulator materials. For electric discharge sources, all four types are required, whereas laser-produced plasma EUV sources do not require electrode and insulator materials. Several types of candidate alloy and other materials and methods of manufacture are recommended for each component of EUV lithography light sources.

Magnetic field shield for laser produced plasma (LPP) had been investigated. The interaction between expanding LPP and magnetic field is not described by Lorentz force, but, may be described by magneto hydro dynamics. When a magnetic field strength of 0.6T was placed between LPP and a faraday cup, attainment ratio of plasma to a faraday cup was decrease to 20%. The attainment ratio was decreased from 0.4 to 0.25 with varying the distance between the plasma and the magnetic field from 10 mm to 70 mm. And, it was observed that plasma detoured around a magnetic field.

The development of suitable radiation sources is a major challenge for extreme ultraviolet lithography (EUVL). For the optimization of these sources and for the determination of the parameters needed for the system design and the system integration these sources have to be characterized in terms of the absolute in-band power, the spectral distribution in the EUV spectral region and the out-band spectral regions, the spatial distribution of the emitting volume and the angular distribution of the emission. Also the source debris has to be investigated. Therefore, JENOPTIK Mikrotechnik GmbH is co-operating with the Laser Laboratorium Goettingen, the Physikalisch-Technische Bundesanstalt (PTB) and the AIXUV GmbH in developing ready-for-use metrology tools for EUVL source characterization and optimization. The set of the tools employed for EUV-source characterization is presented in detail as well as concepts of for calibration and measurement procedures.

In the paper a newly developed compact laser plasma EUV source is presented. The source is based on the double-stream gas puff target approach. The targets are formed by pulsed injection of high-Z gas (xenon) into a hollow stream of low-Z gas (helium) using the valve system composed of two electromagnetic valves and equipped with the double-nozzle setup. The outer stream of gas confines the inner stream improving the gas puff target characteristics (higher density of high-Z gas at longer distance from the nozzle output). It causes efficient absorption of laser energy in a plasma and strong EUV production. The source has been developed in the frame of the EUV sources development project under the MEDEA+ program.

Through simulation we examine the predicted resist images for 50- and 80-nm contacts in the presence of three sources of fluctuations: those associated with random photon absorption sites, those associated with dose statistics, and those associated with resist chemistry. Photon absorption simulations were done using simple optics theory or Monte Carlo techniques. These absorption maps were converted into an initial map of photoacid concentration and a realistic PEB model was used to obtain the developable latent image. A new dissolution model was used to estimate the final resist profile, and image analysis routines were used to extract the centroid position jitter, diameter fluctuations, and edge roughness statistics. The simulations show that each source of fluctuation can be significant, and we propose an experiment that will detect the presence of shot noise and partially separate the three fluctuation sources.

The relationships between polymer molecular weight, surface roughness measured by Atomic Force Microscopy (AFM), and EUV line edge roughness (LER), were studied in four separate rounds of experiments. In Round 1, EUV-2D (XP98248B) was prepared with seven levels of added base. These seven resists were patterned using EUV lithography; the LER was determined using 100 nm dense lines. The LER of the seven resist dramatically decreases with increasing level of base. These LER results were compared with the surface roughness of these resists after development for unexposed and DUV (248 nm) exposed surfaces. In Rounds 2-4, we evaluated three sets of EUV-2D type resists prepared with polymers having Mw of 2.9, 4.9, 6.1, 9.1, 16.1, and 33.5 Kg/mole. EUV LER and surface roughness were determined for each resist. In Round 2, the polymers were substituted into the EUV-2D resist matrix with no other formulation changes. In Round 3, the PAG level was decreased with increasing polymer Mw, to obtain a constant unexposed fill thickness loss (UFTL) for all six resists. In Round 4, both PAG level and base level were modified to yield six resists with similar sensitivity and EFTL. These experiments have led to conclusion about the impact of polymer molecular weight on imaging LER and AFM surface roughness, as well as elucidating the relationship between all three.

The EUV Engineering Test Stand (ETS) is a full field, alpha class Extreme Ultraviolet Lithography (EUVL) tool that has demonstrated the printing of 70 nm resolution scanned images. The tool employs Mo/Si multilayer optics that reflect EUV radiation (13.4nm / 92.5eV) with ~67% peak reflectance per optic. For good reflectivity, many (greater than or equal to 40)Mo/Si layers must be present. Consequently, processes such as plasma induced multilayer erosion, which reduces the number of bilayer pairs on plasma facing optics, need to be understood. Since most materials readily absorb EUV photons, it is important to prevent contamination of mirror surfaces with EUV absorbing material. Contamination can occur by EUV photons “cracking” hydrocarbons or other species absorbed on the optical surfaces. The first ETS condenser component, referred to as C1, is coated with Mo/Si multilayers. Data collected from Mo/Si witness plates placed at the C1 position indicate erosion, using the Xe Laser Produced Plasma (LPP) spray jet, of 1 bilayer per ~15 million shots. Preliminary experiments with a filament jet yielded a significantly higher erosion rate. In the spray jet studies, erosion was found to depend sensitively on the composition of the residual background environment. Addition of low levels, ~7x10^-7 Torr, of H₂O to the vacuum background produced oxidation of the Si cap, and significantly slowed spray jet induced erosion. Operation of the plasma changed the environment in the Illuminator Chamber from oxidizing to carbonizing, thereby changing the nature of the contamination found environment at the C3 optic which does not view the plasma directly (and therefore does not erode). The change in environment is attributed to plasma induced outgassing of fluorocarbons in the Illuminator. Due to the non zero conductance between the Illuminator and Main Chambers, fluorocarbons were also found in the Main Chamber during Xe LPP operation. RGA data are presented that document the effect. In the presence of such outgassing, Carbon deposition rates were measured for the C3, and P.O. Box optics. For C3, a C deposition rate of 3 angstrom / 10 million shots was found, while for the PO Box, a C deposition rate of 0.02 angstrom / 10 million shots was found from the data. All data was acquired with no attempt to mitigate C deposition with gas phase additives such as O₂.

Extreme ultraviolet (EUV)-induced oxidation of silicon-capped, [Mo/Si] multilayer mirrors in the presence of background levels of water vapor is recognized as one of the most serious threats to multilayer lifetime since oxidation of the top silicon layer is an irreversible process. The current work directly compares the oxidation on a silicon-capped, [Mo/Si] multilayers caused by EUV photons with the oxidation caused by 1 keV electrons in the presence of the same water vapor environment (2 x 10^-6 Torr). Similar, 4 nm, silicon-capped, [Mo/Si] multilayer mirror samples were exposed to photons (95.3 eV) + water vapor at the ALS, LBNL, and also to a 1 keV electron beam + water vapor in separate experimental systems. The results of this work showed that the oxidation produced by ~1 µA of e-beam current was found to be equivalent to that produced by ~1 mW of EUV exposure. These results will help allow the use of 1 keV electrons beams, instead of EUV photons, to perform environmental testing of multilayers in a low-pressure water environment and to more accurately determine projected mirror lifetimes based on the electron beam exposures.

Carbon deposition in EUVL is known to occur when optical surfaces in a hydrocarbon environment are exposed to EUV light. Carbon contamination on EUV optical elements affects both the absorption and phase of the reflected light. Because the carbon deposition alters the phase structure of the reflected EUV light it effectively alters the figure of these optics and, thus, the aberrations as well. Absorption by deposited carbon not only reduces throughput but also leads to apodisation of the pupil, which in turn affects imaging performance.

Recent studies have been conducted to investigate the use of atomic hydrogen as an in-situ contamination removal method for EUV optics. In these experiments, a commercial source was used to produce atomic hydrogen by thermal dissociation of molecular hydrogen using a hot filament. Samples for these experiments consisted of silicon wafers coated with sputtered carbon, Mo/Si optics with EUV-induced carbon, and bare Si-capped and Ru-B4C-capped Mo/Si optics. Samples were exposed to an atomic hydrogen source at a distance of 200 - 500 mm downstream and angles between 0-90° with respect to the source. Carbon removal rates and optic oxidation rates were measured using Auger electron spectroscopy depth profiling. In addition, at-wavelength peak reflectance (13.4 nm) was measured using the EUV reflectometer at the Advanced Light Source. Data from these experiments show carbon removal rates up to 20 Å/hr for sputtered carbon and 40 Å/hr for EUV deposited carbon at a distance of 200 mm downstream. The cleaning rate was also observed to be a strong function of distance and angular position. Experiments have also shown that the carbon etch rate can be increased by a factor of 4 by channeling atomic hydrogen through quartz tubes in order to direct the atomic hydrogen to the optic surface. Atomic hydrogen exposures of bare optic samples show a small risk in reflectivity degradation after extended periods. Extended exposures (up to 20 hours) of bare Si-capped Mo/Si optics show a 1.2% loss (absolute) in reflectivity while the Ru-B₄C-capped Mo/Si optics show a loss on the order of 0.5%. In order to investigate the source of this reflectivity degradation, optic samples were exposed to atomic deuterium and analyzed using low energy ion scattering direct recoil spectroscopy to determine any reactions of the hydrogen with the multilayer stack. Overall, the results show that the risk of over-etching with atomic hydrogen is much less than previous studies using RF discharge cleaning while providing cleaning rates suitable for EUV lithography operations.

Low-defect mask blanks remain a key technical challenge to Extreme Ultraviolet Lithography (EUVL). The mask blank is ion-beam sputter-coated with an 81-layer Mo/Si multilayer stack for high reflectance at l = 13.4nm. The current mask coating process can achieve a median added defect level of 0.05 defects/cm2 (12 added defects 90nm or larger on a 200mm Si-wafer test substrate), but this must be reduced by about a factor of 10 to meet mask cost requirements for EUVL. To further reduce the particle defect level, we have studied pathways for particle transport, using test particles and particles native to the coating process, and combined the results into a computational model of particle transport in an ion-beam sputter system. At process pressure, gas drag is negligible for particles above 100nm, so particles travel ballistically until they hit a surface. Bounce from chamber walls allows particles to reach all surfaces in the chamber if they have initial velocities above ~100m/s. The ion beam has sufficient momentum to entrain slower particles and accelerate them toward the sputter target, where some can bounce to the substrate. The model shows preliminary agreement with experimental defect distributions on witness wafers at various positions within the coating chamber.

As the opportunities for experimental studies are still limited, a predictive simulation os EUV lithography is very important for a better understanding of the technology. One of the most critical issues in modeling of EUV lithography is the description of the mask. Typical absorber heights in the range between 80 and 100nm are more than 5 times larger than the wavelength of the used EUV radiation. Therefore, it is virtually impossible to perform parameter studies for 3D EUV masks, such as arrays of contacts or posts, with nowadays standard computers by straightforward application of finite-difference time-domain (FDTD) algorithms, which are used for the rigorous electromagnetic field simulatin of optical masks. This paper discusses the application of field decompsition techniques for an efficient simulation of 3D EUV-masks with FDTD algorithms. Comparisons with full 3D simulations are used to evaluate the accuracy and the performance of the proposed approach. The application of the new QUASI 3D rigorous electromagnetic field simulation for EUV masks reduces memory requirements and computing time by a factor of at least 100. The implemented simulation appraohc is applied for a first exploration of mask induced imaging artifacts such as placement errors, telecentricity errors, Bossung asymmetries, and focus shifts for 3D EUV masks.

EUV lithography at 13nm wavelengths will require the use of reflective multi-layer mask substrates with a patterned absorbing top layer. As the height of the absorber is many times the wavelength of the exposure radiation, EUV masks are true topographical entities. Therefore, scalar Frauenhofer diffraction simulations are not completely valid and accurate. Rigorous solution of Maxwell's equations in the topography is necessary to accurately predict many of the lithographic effects of these masks. Additionally, the interactions of the non-vertical radiation reflectance from the multilayer stack with the absorber causes further non-intuitive lithographic results which need to be understood and optimized. To further the rigorous understanding of EUV mask effects, we have extended existing 2D and 3D rigorous simulators to model EUV lithography diffraction. These simulators, METRO and METROPOLE-3D respectively, use the waveguide method for fast and accurate computation of topographical mask structures. Modifications of these simulators enable an order of magnitude speedup in the calculation of results containing large numbers of EUV multilayers, improve mask dimension modeling accuracy and enable accurate modeling of the non-normal EUV source illumination through a range of pitch values. Rigorous illumination and multilayer defect analysis results will also be shown. Additional comparisons to recently published rigorous EUV simulations results and impressive runtime results on standard desktop workstations are presented. The results show that for EUV lithography the best focus position and the symmetry of the aerial image intensity profile with respect to best focus is affected by the pattern pitch.

Electron Projection Lithography (EPL) has a high potential for applicability beyond the ITRS 65 nm node, especially for contacts and gate layers. The concept of synchronization control of the Nikon EB stepper is explained. The reticle stage and the wafer stage are servo controlled to target positions individually. The residual stage position errors are compensated by the electron beam deflection control. The electron beam deflection is feed forward controlled using predicted stage position data from a subsystem called “Filter/Predictor”. The performance of the stage position prediction of the Filter/Predictor is described. This paper also reports the performance of the first EB stepper tool, the NSR-EB1A, during its preliminary adjustment phase. Dynamic scanning and stitching exposure, which requires synchronization of both the beam deflection motion and the stage scanning motion, was realized. Dynamic resolution of 100 nm and dynamic subfield stitching accuracy of 25 nm (3sigma) were obtained, and further improvement is expected.

In Electron-beam Projection Lithography (EPL), achieving the requirements for the ITRS 45 nm roadmap node will require decreasing simultaneously both the beam blur from the Coulomb interaction effects and the geometrical aberrations from their present values. Because next generation lithography tools are required to have both high resolution and throughput, the Coulomb effect becoems more of an issue. In this paper, we propose a novel concept to effectively decrease the Coulomb effect. Based on this new concept we develop an EPL electron optical system in which not only the Coulomb effect but also the geometrical aberrations are greatly reduced. We report on the properties of this new EPL optical system.

Minimizing mask-level distortions is critical to ensuring the success of electron projection lithography (EPL) in the sub-65-nm regime. Previous research has demonstrated the importance of controlling the stress in the patterned stencil membranes to minimize image placement distortions. Low-stress, 100-mm diameter EPL mask blanks have been patterned with a layout that simulates the effects of the cross-mask and intra-subfield pattern density gradients found in a realistic circuit design. Extensive IP measurements were made to illustrate how local subfield correction schemes can be used to reduce all mask-level distortions (regardless of pattern type) to less than 15 nm (3s). Combining membrane stress control with the use of repeatable and identical reticle chucking is expected to reduce EPL mask-level distortions to the values that will be needed for the 65-nm design node.

A new mask inspection system using transmission electron beam (EB) technology is being developed to detect defects on electron projection lithography (EPL) masks with design rules of 65-nm and below. In our new system, we use the transmission EB image, multi-line Time Delayed Integration (TDI)-CCD camera, and stage scanning. The transmission EB image can detect defects hidden in stencil patterns with a very high resolution. The scanning system, combined with a high-speed image acquisition system, enables high throughput. A mganified image of the mask stencil pattern is formed on an image sensor by the electron-optic systme. The image of the stencil patterns is captured by a multi-line TDI-CCD camera and the image data is transferred to an image-processing unit composed of multi high-speed image processors. To capture an EB image, a special TDI-CCD camera was developed. The typical pixel size is 50-nm on the mask. To achieve the goal of high-speed image processing, multi-fiber cable is used to transfer the image data from the camera to the image-processor. Each image processor shares the inspection area in parallel, and performs a die-to-database comparison. Typical inspection time for a 200-nm EPL stencil mask is 4 to 5 hours. A defect of 50-nm was observed by the developing system. This inspection system will be completed in March 2004.

A novel chemical procedure is described that presents, for the first time, a dissolvable template (as opposed to a hard or soft template) for printing. The method, called Molecular Transfer Lithography (MxL), is a pseudo-maskless approach to high-throughput fabrication of devices over large areas with feature sizes that can extend below 100 nm. The MxL procedure utilizes a chemical process that replicates the surface topography of a master pattern into soluble polymeric templates. The template is dissolvable in water but not most organic materials. This property of differential solubility is useful for transferring topography or functional materials onto substrate surfaces. The template is generally chemically dissolved at the conclusion of the pattern formation process. The paper presents a variety of printing strategies and results including lines at 53 nm, three-dimensional patterns, 250 nm contact hole levels, diffraction and optical structures and 200 mm wafer printing. The technique is also useful for planarization. Multi-level printing is achieved by integration of the MxL process and dissolvable templates with standard contact aligners.

We study mirror based pattern generation systems to provide an understanding of how they can be operated in an analog mode to meet the quasi-continuous sizing and placement requirements of optical lithography. Both tilting mirrors and piston-motion mirrors are examined. The aerial images are compared with those generated by simple binary masks. The effect of grayscaling, used to place and size features, on image quality is measured. Normalized image log slope (NILS) is used as the measure of image quality. Tilting mirrors used in grayscale mode provide image quality comparable to binary masks, and piston mirrors are somewhat better.

With the objective to ramp-up 65 nm CMOS production in early 2005, preliminary works have to start today to develop the basic technological in order to be correctly prepared. In the absence of commercial advanced 193 nm scanners compatible with these aggressive design rules, electron beam technology was employed for the realization of a first 6-T SRAM cell of a size of 0.69 μm². This paper highlights the work performed to integrate E-beam lithography in this first 65 nm CMOS process flow.

An easy way to pattern 65nm CD target, when optical lithography technology is not available, is to use an Electron Beam Direct Write tool (EBDW), which is well known for its high resolution patterning potentials, with the drawback of a very low throughput. Emerging techniques of electron projection lithography also propose the same patterning capability with enhanced throughput. One of the most crucial issues, when dealing with integration, is the overlay capability of the systems. This paper exposes the studies made on the overlay capability issue of the LEICA EBDW installed in STMicroelectronics (STM) production plant in Crolles (France) and proves our tool is ready to support the 65nm node technology development.

Resolution Enhancement Technologies (RETs) have extended the life of optical lithography well into the regimes of k₁=0.3. Although there are a number of RTEs widely in use now, all involve some degree of optical and process correction (OPC). This puts additional strain on the data handling and management capabilities for process calibration and mask preparation. Alternative lithogrpahic technologies, such as Extreme UV (EUV), electron projection lithography (EPL), direct-write maskless lithography (MLL), and even nanoimprint techniques have been proposed and are under rapid development. All these alternatives create patterns usinga wavelength (or the dimension of the writing tool itself) smaller than the desired feature. Since these should confortably increase k1 back to the values enjoyed many years ago, it is generally assumed that the issues associated with OPC and the data management for RETs will go away. This is not true. Although these will no longer have "optical proximity" effects to compensate, each of these processes introduces its own signature of distortions, which in turn will require compensation and the associated data management. In this paper, we will inventory the state of development for each of these technologies, and outline the expected compensations and requirements palced on data management that are associated with the adoption of the technology. These effects include electron proximity effects and pattern stitching for EPL, Flare induced proximity effects for EUV, gray scale mapping for various maskless techniques, and even local process biasing for various nanoimprint approaches.

The two LEEPL beta-tools were completed in earlier 2002 and have been evaluated for the performance. 50nm CH patterns and 70nm L/S patterns are attained and the CD uniformity of 70nm L/S patterns with 37 shots on a 200mm wafer is under 4nm with the LEEPL beta-tools. In addition, it is proven that the fine tuning deflector can correct a mask and a wafer distortion by giving a minute angle to the electron beam. The mask distortion with respect to orthogonality and magnification is decreased on a wafer by 1/5. By means of this fine tuning deflector, Mix & Match accuracy with any other lithography tools will be better and difficulty of 1X stencil mask fabrication wil be easier. Referring to the data which has been obtained with the evaluation of the LEEPL beta-tools, the first LEEPL production tool dubbed "LEEPL-3000" is under construction to realize the satisfactory ability for 65nm node device fabrication. The shipping of the first LEEPL-3000 is scheduled in earlier 2003 and it is earlier than any other Next Generation Lithography technologies.

Imaging capabilities of low-energy electron-beam proximity-projection lithography (LEEPL) are discussed focusing mainly on the hole patterns for chemically amplified resist. LEEPL needs a multi-layer process with a resist layer less than 100 nm thick. To achieve the imaging performance of the 65nm node, we optimized intermediate spin-on-glass layer and top-layer resist, which were selected carefully. 80 nm hole patterns were achieved with 10% exposure latitude, and current imaging position and 45 nm node positions were investigated using σQBP. σQBP was improved from 64.5 nm to 48.9 nm.

By using the Near Field in Proximity X-ray Lithography (PXL), the technique is demonstrated that extends beyond a resolution of 25 nm print featuer size with 2:1 pitch to line width. "Demagnification by bias" of clear mask features is positively used in Fresnel diffraction together with multiple exposures of sharp peaks. Exposures are performed without lenses or mirrors between mask and wafer, and the "demagnification" is achieved in the selectable range 1X to 9X. Pitch is kept small by multiple, stepped exposures of sharp, intense, image peaks followed by single development. Low pitch nested lines are demonstrated. The optical field is kep compact at the mask. Since the mask-wafer gap scales as the square of the mask feature size, mask feature sizes and mask-wafer gaps are comparatively large. Because the features are themselves larger, the masks are more easily manufactured. Meanwhile exposure times, for development levels high on sharp peaks, are short, and there are further benefits including defect reduction. Many devices, including batches of microprocessors, have been demonstrated previously by traditional 1X PXL which is the most mature of the Next Generation Lithographies and which is now further extended. For 2D Near field patterning, temporal and spatial incoherence at the Critical Condition are used to show, not only that peculiarities in the aerial pattern, such as "ripple" and "bright spots", can be virtually eliminated, but also that there is an optimum demagnification, around 3X, in the Fresnel diffraction, where the contrast is highest. At this demagnification, patterns of various dimensions can be printed using various and appropriate demagnifications.

In this paper, an EUV lamp development based on high repetition rate compact capillary discharge devices is presented. The first lamp, named CAPELLA (Capillary EUV Lamp for Lithography Approach) emits up to 1 W in-band and has been selected as the illumination source of the first experimental European exposure tool (BEL). An industrial version of CAPELLA named PROXIMA has recently been operated and is the first commercial product of a new spin-off of GREMI called IGETECH. The design of a 10 kHz capillary discharge likely to emit up to 10 W in-band and labelled as ATλAS (AT wavelength Alternative Sources) is finally documented.

The increased demand for high throughput lithography with Front To Back-side Alignment (FTBA) capability has led ASML to develop the FTBA functionality within its current platform. This option is named 3DAlign and aims at an FTBA overlay of 500 nm. To relate a position on the back side of a substrate to its front side, two back-side marks are projected from the back to the front side by two optical branches. In this way, the images of the back-side marks have the proper orientation and magnification to be aligned by the standard ASML alignment hardware. The metrology challenge is to calculate the back-side mark positions in front-side (exposure) coordinates and to compensate for the systematic errors introduced by the optical branches. An absolute relation between the front- and back-side positions on the substrate is obtained by calibrating the system with a special 3DAlign calibration wafer. This is a wafer that has marks on both sides with a known offset. In this paper, the basic ideas and algorithms for the FTBA metrology are discussed.

An EUV reflectometer, based on a laser-produced plasma (LPP) light source, has been developed for characterization of EUV lithography systems. The reflectometer consists of the LPP light source, a prefocusing toroidal mirror, a grating monochromator, a polarizer, a beam intensity monitor, a refocusing toroidal mirror and a sample stage. The LPP light source is driven by a Nd:YAG laser; the laser beam is focused onto a copper tape target. A debris mitigation system that uses a rotating shutter was developed. Higher-orders formthe grating monochromator were suppressed to less than 0.2% of incident beam intensity by total reflection of three grazing incidence mirros. In order to compensate for beam intensity instability, a beam intensity monitor using a grating beamsplitter was installed between the refocusing mirror and the sample. Beam intensity instability can be corrected to less than 0.1% by using the beam intensity monitor.

utilizing a laser-based source for the generation of 13nm radiation. The EUV plasma is produced by focusing a Nd:YAG laser into a pulsed xenon or oxygen gas jet. The alternate use of these two target gases accomplishes either an intense broad-band (Xe) or a less intense narrow-band line emission (O2) at 13nm. Different types of nozzles were tested in order to optimize the emitted radiation with respect to maximum EUV intensities, small source dimensions and pulse-to-pulse stabilities. The investigation of these crucial source parameters was performed with specially designed EUV pinhole cameras, utilizing evaluation algorithms developed for standardized laser beam characterization. In addition, a rotatable pinhole camera was developed which allows spatially and angular resolved monitoring of the soft X-ray emission characteristics. With the help of this camera a strong angular dependence of the EUV intensity was found, indicating reabsorption of the EUV radiation in the surrounding gas. The results were compared with Rayleigh scattering measurements for visualization of the target gas density.

With the development of EUV lithography there is an increasing need for high-accuracy at-wavelength metrology. In particular, there is an urgent need for metrology at optical components like mirrors or masks close to the production line. Sources for metrology have to fit different demands on EUV power and spectral shape than sources for steppers systems. We present the results of the radiometric characterization of a laser produced plasma (LPP)-source, newly developed at Max-Born-Institute Berlin for use in an EUV reflectometer. It is operated with a high-power pointing-stabilized laser beam (energy per pulse up to 700 mJ, 10 ns pulse duration, < ± 25 μrad pointing stability) at 532 nm which is focussed on a rotating Au target cylinder. The incident angle of the laser beam is set to 63°, the detecting angle 55° to the target normal. The source has been characterized regarding spectral photon flux, source size and source point stability. Two independently calibrated instruments, an imaging spectrometer and a double multilayer tool for in-band power measurements were used to obtain highly reliable quantitative values for the EUV emission of the Au-LPP source. Both instruments were calibrated by Physikalisch-Technische Bundesanstalt in its radiometry laboratory at the electron storage ring BESSY II. We obtained a source size of 30 μm by 50 μm (2s horizontal by vertical) and a stability of better than 2s=5 μm horizontally and 2s=9 μm vertically. A spectral photon flux of 1*10e14 /(s sr 0.1 nm) at 13.4 nm at a laser pulse energy of 630 mJ is obtained. The shot-to-shot stability of the source is about 5% (1s) for laser pulse energies above 200 mJ. For pulse energies between 200 mJ and 700 mJ, there is a linear relation between laser pulse energy and EUV output. The spectrum shows a flat continuos emission in the EUV spectral range, which is important for wavelength scanning reflectometry. High stability in total flux and spectral shape of the plasma emission as well as low debris was only obtained using a new target position for each shot. There is also a trade off between source size and EUV power. For a slightly defocused laser, an increase in EUV power up to a factor of two is obtained, while the source size also increases by about a factor of two. It is shown that an Au-LPP source provides spectrally flat reproducible emission with sufficient power at low debris conditions for the operation of a laboratory based EUV reflectometer.

An image formation technique based on the Geometrical Theory of Diffraction was presented last conference. The technique is a scalar technique and is applicable to infinitely thin and perfectly conducting mask. We explore in this paper the extension of the technique to 1D Extreme Ultra-Violet(EUV) Lithography mask, taking into consideration both the material property and the topography of the mask. Vectorial nature of light is incorporated in the treatment. Results obtained are promising and encouraging. Computation time is relatively much shorter and the technique could simulate irradiance profile for any illumination angle. The technique is simple and elegant and lends understanding to image formation. We conclude that the asymmetry-through-focus characteristic usually found in EUV and Phase Mask imaging is an imaging phenomenon. We also conclude that corrections for proximity effect and pattern infidelity will be needed when EUV Lithography is introduced at the 32 nm node, assuming a system NA of 0.25. Lastly, for a partially coherent illumination, it appears necessary to compute the irradiance corresponding to each illumination point individually.

The International Technology Roadmap for Semiconductors requires improvements in resolution for each lithographic node. In essence, all sources of distortion in the chip fabrication process must be minimized to meet the stringent error budgets for the sub-90-nm nodes. These include the thermal distortions of the device wafer caused by energy deposition during exposure. Absorbed energy from the beam produces temperature increases and structural displacements in the wafer, which directly contribute to pattern placement errors and image blur. In this research, the thermomechanical response of the device wafer was investigated and compared for 193-nm lithography and EUV lithography. Thermal and structural finite element (FE) models were developed to numerically simulate the exposure process for both types of tools. The three-dimensional FE models include the full wafer and chuck to identify the time-dependent response. For verification purposes, the FE models were benchmarked against an analytical test case. Since the thermomechanical response is relatively sensitive to exposure energy and wafer chucking, parametric studies were performed to illustrate the effects of resist sensitivity, backside contact conductance, and effective boundary conditions. Results for both 193-nm lithography and EUV lithography are presented.

Discharge produced plasma (DPP) devices have been proposed as a light source for EUV lithography. A key challenge for DPP is achieving sufficient brightness to support the throughput requirements of exposure tools for high-volume manufacturing lithography. To simulate the environment of the EUV source and optimize the output of the source, an integrated model is being developed to describe the hydrodynamic and optical processes that occur in DPP devices. The model includes both plasma evolution and magnetohydrodynamic processes as well as detailed photon radiation transport. The total variation diminishing scheme in the Lax-Friedrich formulation for the description of magnetic compression and diffusion in a cylindrical geometry is used. Several models are being developed for opacity calculations; a collisional radiation equilibrium model, a self-consistent field model with Auger processes, and a non-stationary kinetic model. Radiation transport for both continuum and lines with detailed spectral profiles are taken into account. The developed models are integrated into the HEIGHTS-EUV computer simulation package. Preliminary results of a numerical simulation of xenon gas hydrodynamics and line radiation output are presented for various plasma conditions.

The collection efficiency is one of the most important characteristics of a source in an EUV lithography system. It is defined as the fraction of the total in-band EUV radiation power which can actually be picked up and used in the subsequent total optical path of an EUV projection system down to the wafer. It is not a property of the source alone, but is determined both by the details of the optical system with its etendue limitations and by the geometrical intensity distribution of source plasma. Until now, the source performance had to be evaluated by ray-tracing calculations of the complete optical lithographic system including the source, which were quite time-consuming. Therefore, a simple and fast, but still reliable method for evaluation of collection efficiency has been developed which is based on taking pinhole images of the source plasma with a CCD camera. From these, a source intensity distribution is constructed, which is used as input to a subsequent simple projection algorithm. By taking into account radiation from an "allowed" source volume only, which is constructed beforehand by rigorous ray tracing, the actual collected power and the collection efficiency is calculated. Comparison with detailed raytracing model calculations of the complete optical system indicate the good accuracy of the method. It can be used in a simple way for source optimization and enables reliable evaluation and specification of source performance in EUV projection systems.

We succeeded in generating a strong narrowband peak at 13.7 nm in a cavity confined Sn plasma. Fraction of the energy within 2% bandwidth at 13.7 nm against the total radiation spectrum was 11.3%. In our experiment, the plasma is generated not on a solid plate but in a cavity as described below. A YAG laser pulse ablates the surface of a concave structure Sn target in order to supply the material for plasma generation. The next laser pulse with 1064nm wavelength heats the ablated material to generate a high temperature plasma. The ablation YAG laser is focused to 600μm diameter with a flux of 10 J/cm². The heating YAG laser is focused to 100μm diameter at a delay time of 30 ns after the laser ablation. Emission spectra are observed using a grazing incidence Hitachi flat-field grating and a back side illuminated CCD detector. Fraction of the energy at 13.7nm within 2% bandwidth in the whole radiation energy was 11.3%. However, the intensity of the spectrum peak was about 1/10 against the plane Sn target LPP source. When nano (less than 200 nm diameter) particle SnO₂ deposited on a 100nm-thick Si3N4 membrane were irradiated, a sharp peak was observed. The intensity of the spectral peak of the nano particle SnO₂ target LPP source was as high as that of the plane Sn target LPP source. The EUV energy within 2% bandwidth at around 13.7 nm to the whole radiation energy of the nano particle SnO₂ target LPP source was 7.4%.

Laser plasma light source using double pulses laser irradiation and through-hole method is proposed as a mass-limited target srouce for extreme UV (EUV) radiation. After minimum necessary material is supplied using the ablation laser from a solid target, only ablated material is irradiated with the heating laser to produce a high-temperature plasma, and EUV radiation is extracted passing through the hole formed in the solid target. Fundamental concept of this scheme, EUV radiation and great reduction of particle debris were experimentally confirmed.

In order to protect a multilayer mirror from sputtering or ion implantation, high-energy ions ejected from EUV source plasma are to be blocked. We propose use of a laser-produced plasma as an ultra-fast shutter. Ion signal form an ion-source plasma dropped abruptly by two orders of magnitude to a noise level after the shutter plasma generation. The stopping effect for the high-energy ions was observed to reduce as the distance of shutter plasma expansion increases, but the suppression of ions below detection level was observed up to 10 mm. We concluded that reduction of ion signal was caused by in-take of the source plasma flow into the stream of the shutter plasma.

An EUV radiation source development with discharge-produced plasma (DPP) has been started in Gotenba branch of EUVA Hiratsuka R&D center. For the early stage of the development, fundamental characteristics of DPP including current, voltage, EUV energy and spectrum in EUV region were studied. A capillary of which the inner diameter was 2.3 mm, and Xe gas were used as a source to be expected in-band EUV radiation from magnetically-compressed Z-pinch plasma. An all-solid-state magnetic pulse compression generator was employed, which can deliver the current of 14 kA into the capillary load with the rise time of approximately 500 ns. In-band EUV energy and spectroscopic measurements were carried out. It was found that the in-band EUV energy increased with increasing the current amplitude and/or pressure of filled Xe gas. The highest in-band EUV energy obtained was 8 mJ per unit solid angle.

The Extreme UV Lithography System Development Association (EUVA) was established in Japan in May 2002 and is supported by the Ministry of Economy, Trade and Industry (METI). EUVA started the light soruce development in September 2002. This development is done by the assocaition members Gigaphoton, Ushio, Komatsu, Canon, Nikon, the National Institute of Advanced Industrial Sciecne and Technology (AIST) and several Japanese universities. The target of the four-year project is the development of a EUV light source with 10W clean focus point power. For the end of the fiscal year 2003 the development of a 4W EUV light source (clean focus point power) is planned. Both, Laser-Produced-Plasma (LPP) and Discharge-Produced-Plasma (DPP) EUV light sources are investigated at first. Our group at the EUVA Hiratsuka R&D Center is working on LPP sources. We are currently focusing on the development of a driver laser and a liquid Xenon plasma target. The laser is a Nd:YAG MOPA (Master Oscillator and Power Amplifier) system oscillating at 1064 nm. Average power, repetition rate and pulse duration of the laser system are 500 Watt, 10 kHa and 30nsec, respectively. The Xenon liquefication system operates at a maximum pressure of 5MPa and a temperature range between 160 K and 190 K. The pressure inside the vacuum chamber is below 0.1Pa during system operation. This paper presents the current status of the EUV system component development as well as first experimental results of generated EUV radiation.

Generation of extreme ultraviolet (EUV) radiation from solid targets is studied and a compact EUV source for small-scale lithographic applications and EUV metrology is development. This source is based on a transfer of conventional x-ray tube technology into the EUV spectral range. As in an ordinary x-ray tube, electrons are generated by a tungesten filament and accelerated in a high-voltage electric field towards a solid target. In the demonstrated "EUV tube" beryllium and silicon targets are used to generate radiation at 11.4 nm and 13.5 nm, respectively. The absolute converstion efficiencies into EUV photons are measured. At 13.5 nm an EUV power of 34μW or 2×10¹² photon/s (in 2% bandwidth and a solid angle of 2π sr) is demonstrated. Prospects for a further power scaling of the EUV source are discussed.

In Extreme Ultraviolet lithography, the electromagnetic modeling of the mask allows to determine the influence of the mask structure on the electromagnetic field. That makes it possible for example to analyze the influence of a defect within the multi-layer stack. This paper describes a modeling method of the EUV mask based on the Rayleigh assumptions1. These hypotheses lead to a more restrictive validity domain than rigorous methods like modal methods or Finite Difference Time Domain (FDTD), but is shown in this paper to be usable adapted for EUV mask simulation. Furthermore the simulations are less costly in memory resources and in computing time.

The defect printability for two types of patterns isolated lines and contact holes was studied through a finite-difference time-domain (FDTD) computer simulation approach at the EUV wavelength of 13.5 nm. A three-dimensional FDTD solver was employed for aerial image analysis to assess the overall acceptable defect size for both opaque as well as clear phase defects for the 32 nm technology node.

An isolated local non-flatness model was devised to investigate the influence of a non-flat mask on pattern shift. The causes of pattern shift were divided into two types: out-of-plane displacement (OPD) of the mask, and the global height variation (GHV) resulting from OPD. GHV induces a pattern shift that depends on the incident angle of the EUV beam. Reducing the incident angle was found to mitigate pattern shift on a wafer. Although the pattern shift due to OPD is negligible when the P-V flatness is more than 100 mm, the effect of GHV is not negligible. To keep the pattern shift on a mask below 12 nm, the upper limit on the non-flatness of the mask is a P-V flatness of less than 100 nm and a period of more than 300 mm. In addition, the global slope must be less than 1 μrad for an EUV mask substrate.

The printing of line patterns using an attenuated phase shift mask (PSM) and an alternating PSM was examined through simulations. The attenuated PSM consisted of a bilayer structure, in which multiple interference provides a large latitude for appropriate thicknesses of the absorber and buffer layers. The attenuated PSM provides greater depth of focus (DOF) for sparse lines. The alternating PSM has better aerial-image contrast not only for dense lines but also for isolated lines. An additive structure is proposed for the alternating PSM, which does not require any etching of the glass substrate. Given an appropriate stack of materials, this structure provides a phase shift of 180 degrees and the same reflectance simultaneously. In this study, the printability of 18-nm-wide lines was investigated for an attenuated PSM with annular illumination and an alternating PSM with normal illumination with a small σ of 0.3.

The characteristics of a defect counting tool, COMNET< based on scattering light measurement, is presented. This prototype supports the development of defect-free EUV blanks. Thanks to new improvements, it becomes possible to detect PSL particles having a diameter as low as 100 nm, with a video CCD camera, on silicon substrates or on EUV blanks. To reach this sensitivity, one of the enhancements consists in a laser irradiation close to 65°. The present configuration and the use of a CCD camera, with a variable exposure time, should lead to the detection of 80 nm PSL particles deposited on silicon substrate. This extrapolation is based on experimental results and on a simple model. To detect 100 nm particle and smaller particles, it is essential to reduce the level of stray light and to increase the signal to noise ratio. In our application, the stray light essentially comes from three sources: the noise induced by the roughness of the sample, the Rayleigh scattering of the atmosphere, and the stray light in the room. The restrictions induced by these phenomena are described in some detail. All the improvements are not only available for the characterization of silicon substrates but also for transparent blank substrates and for EUV mask blanks. The additional noise induced by the tranparent substrate is analyzed. The defects, whatever the compoent sizes and the component shapes can be detected. A cross characterization achieved with a commercial tool on silicon substate is reported. Counting measurements performed on EUV blanks are shown. Furthermore, a more explicit definition of added defects is proposed.

Although Mo/Si multilayers are now widely used in EUV lithography development programs, multilayer induced substrate stress continues to be a major issue. Standard stress values of -350 to -450 MPa, reported for Mo/Si systems produced by magnetron sputtering, induce an intolerable deformation of the surface figure of EUV optical components. Stress in e-beam deposited Mo/Si multilayers has not been reported before. At FOM Rijnhuizen, an extensive stress mitigation program has been carried out on multilayers produced by e-beam deposition and medium energy ion polishing. The stress in our standard, high reflectance Mo/Si multilayers is less than -200 MPa. Although e-beam deposition apparently halves the typical stress values obtained by sputter deposition, it is still above the allowable limit for the first lithographic system, the so-called Alpha Tool. To further reduce stress, the influence of the Mo fraction, the number of periods and the multilayer period or d-spacing has been investigated. Varying the Molybdenum fraction in e-beam deposited multilayers results in a similar dependency as reported for magnetron sputtered coatings, though at strongly reduced absolute values. Furthermore, variation of the d-spacing has a small influence on stress. The number of periods however, has no influence on the stress value in the range from 20 to 50 periods. Applying stress mitigation techniques based on adjustment of the Mo fraction, a high reflectance of above 69% at near normal incidence at 13.5 nm has been obtained for multilayers with a stress value of only -33 MPa. This has been achieved by using Mo and Si only. This stress value is sufficiently low to enable the first generations of EUVL optics.

Image fidelity is one of the fundamental requirements in lithography and it is becoming more important as feature sizes shrink below 90 nm. Image distortion depends on the mask deformation caused by the intrinsic stress in the film-substrate system. To develop an understanding of stress generation and to control film quality, measuring film stress is essential. In recent years, research laboratories and industry have increasingly adopted indirect methods for determining film stress. All of these methods are based on the measurement of substrate deformation, and the film stress is calculated from the substrate curvature by the local application of Stoney’s equation. When the two principal stresses at each point in the film plane are not equal to each other and their distribution is not uniform, the local application of Stoney’s equation is invalid. Even though the accuracy of the measurement may be high, the stress determined may not be. An alternative technique based on numerical analysis has been developed. The limitations of using Stoney’s equation and the new stress measurement technique are discussed in this paper.

A theoretical analysis of the capability of Extreme Ultra-violet Lithography (EUVL) to print contacts in the range of 30-50 nm with acceptable yield is presented. We study the problem from the viewpoint of two issues: the effects of shot noise and the available process window. Shot noise is modeled as a simple statistical fluctuation of the number of photons absorbed in a given contact. The process windows for various size contacts are simulated for a representative 0.25-NA EUV tool. We find that EUVL should be able to print 30 nm contacts with a reasonable exposure latitude, and we also conclude that the effects of shot noise will not significantly decrease yield at this exposure latitude for a resist with a dose-to-clear value of ~5 mJ/cm2. A dose-to-clear value becomes ~2 mJ/cm2 for 50 nm contacts. In addition, a comparison at the same node shows that 50-nm contacts have a greater process window than 30-nm lines.

In this work simulation parameters are developed for Shipley EUV-2D photoresist under exposure at 13.4nm. Baseline parameter values are determined from theory and experiment. The simulation parameters were tuned from these values using a commercial automatic parameter optimisation software to match simulation results to experimental lithographic data generated using the ETS Set-2 projection optics in the subfield exposure station (SES). In an attempt to maximise parameter accuracy the experimental data set used included 4 different feature sizes and known non-idealities of the exposure set-up were accounted for (mask errors, lens aberrations and metrology bias). The resulting model described the experimental data very well with only a low level of residual error.

By using a spatial frequency doubling method, our 10x Schwarzschild optic can print high-contrast features at 50 nm with low line-edge roughness (LER). In this paper, we also present new techniques for evaluating photoresist at EUV wavelengths using our system. One method is used to determine the ultimate resolution of a resist through linewidth vs. dose measurements. Another is to investigate line-edge roughness properties by varying the aerial image contrast of a pattern. A novel filtering method is proposed that would allow multiple contrasts to be printed in a single exposure. This is achieved by varying the duty cycle and line/space transmission levels of the object grating. Since this is a single exposure technique it would allow for mroe controlled contrast tests when evaluating resists.

A single-layer resist process for a technology nodes at or below 65nm utilizing a novel supercritical dry technique and Electron-beam Projection Lithography (EPL) technology is discussed. EPL is inhernelty advantageous in imaging sub-65nm geometries with high aspect ratios. Pattern collapse of these high aspect ratio (resist) structures, however, is a critical and limiting issue. By employing our novel supercritical carbon dioxide (SCCO2) dry technique, 70nm and 60nm lines and spaces patterns with a resist thickness of 250nm, whose aspect ratio is 3.5 and 4.2 respectively, have been successfully demonstrated without resist pattern collapse.

In this study, we investigated resist pattern collapse during the resist development process. We evaluated the effect of a simple improvement such as rinse-liquid sequencing and rinsing using surfactants. First, we controlled the wafer spinning speed during the rinse-liquid flow step to reduce liquid flow shock. Using this approach, we obtained a 110-nm L/S (line and space) structure with no pattern collapse. However, this technique has only a small effect on preventing pattern collapse with sub-100-nm devices. By using a rinse process with a surfactant, we could control pattern collapse with 100-nm L/S or smaller patterns. Finally, we have succeeded in controlling pattern collapse of 70-nm L/S patterns (aspects ratio of 4.6) using a surfactant during the rinse process. These two simple methods are a significant improvement over conventional rinse processes. These process improvements are available for 90-nm (and smaller) design rules and are applicable for a single layer resists.

Resist heating is one of the largest contributors to critical dimension (CD) distortion in electron beam photomask fabrication. Previous methods for reducing CD variation caused by resist heating include lower beam currents, increased delays between electron flashes, and writing in multiple passes. However, all these methods lower mask writing throughput. This leads to increased mask writing cost, which is increasingly becoming a major limiting factor to semiconductor industry productivity. In this work, we investigate a new degree of freedom for mitigating CD distortion caused by resist heating. By optimizing the sequence in which subfields are being written, it is possible to reduce CD variability caused by resist heating, without significantly increasing the mask writing time.

Cationic graft polymerization lithography is a variation of top surface imaging schemes. This technique uses a spin-coated, inert polymer film as a photoacid generator carrier. UV exposure is used to generate acid in the top surface of the film. A vapor-phase reaction between the generated acid and a silicon-containing monomer occurs in the exposed areas. The silicon-containing polymer formed, or grafted, on the surface is used as an oxygen etch mask for subsequent pattern transfer through the underlying film. A modular approach can be employed in material design, allowing optimization of characteristics of each component. A key criterion is introduced by the interaction between the transfer layer and the graft monomer. The solubility of the monomer into the inert polymer layer influences the growth behavior, and should be minimized to prevent background silylation and potential swelling. The solubility of the monomer in the inert polymer is characterized by measuring the equilibrium sorption of the vapor into the polymer. Solubility behavior can also be estimated from group contribution theories. These estimates guide the rational design of materials for this lithography process. Based on this analysis method, a new monomer, bis(vinyloxymethyl)dimethylsilane, has been designed and tested. Its sorption into a typical polymer layer has been characterized experimentally. Kinetic growth rate data have been obtained on a quartz crystal microbalance system, and preliminary imaging results using 248 nm exposure are presented.

EPL (Electron Projection Lithography) is one of viable candidates for the NGL (Next Generation Lithography) used for the 65nm technology node and below. EPL uses the membrane mask with grillage structure to reinforce the membrane area. The NGL poses stringent requirements to mask accuracy. The most severe problem of membrane mask is IP (Image Placement) accuracy. In this study, distortion characteristics of an EPL stencil mask are analyzed by FEM (Finite Element Method). Its dependence on the grillage structure and the distribution of pattern density is presented. The distortion of the wet-etched wide grillage is negligible compared to the required IP accuracy. However, the distortion of the dry-etched narrow grillage cannot be negligible, which exceeds 5nm. The accumulation of the distortion will deteriorate the IP accuracy. The non-uniform distribution causes much larger distortion in membrane than the uniform one, which is quite unacceptable for the 65nm technology node and below. Improvement of the IP accuracy by the correction using an EPL optical system is also evaluated. The EPL optical system can reduce the distortion caused by the uniform distribution. However, the large distortion caused by the non-uniform distribution cannot be corrected. EPL requires its own design rules regulating the distribution of pattern density from the view point of IP accuracy.

We investigated the defect printability of hole patterns in electron projection lithograpy (EPL) using a diamond reticle with a programmed defect pattern. The reticle was fabricated by NTT-AT and wafer exposure was performemd using Nikon's EB projection experimental column. We simulated the defect printability to udnerstand in greater detail. We found that the mask error enhancement factor (MEF) of the size shift defect category exceeded the value of one and was degraded by the amount of beam blur. On the other hand, the printability of the dot defect category was lower than the shift category. In particular, pint hole defects smaller than 100 nm were not printed. However, the defect types of under size shift, truncation, edge intrusion, and corner intrusion (they decreased the opening area), actually increased the defect size because the defect was too small for hole patterns to print. In general, the defect printability of hole patterns depends on the beam blur, and the printed error size at the hole patterns getting larger than the line patterns. We have to pay clsoer attention to the hole pattern defect than to the line patterns.

In Electron Projection Lithography (EPL) that is desigend for 65nm production tool, proximity effect correction (PEC) is an important issue for an accurate feature size control. Reticle resizing meethod is adopted for its correction. High controllability (ΔCD within ±10%) of critical dimension (CD) is required after proximity effect correction. For estimating the CD controllability, we have evaluated its dependency on exposure dose and beam blur in resizing method for the first time in EPL using Nikon's EPL experimental column (acceleration voltage=100kV, magnification=1/4, sub field size=250×250μm). Evaluating patterns were the target size of 100nm isolated line and twin lines. Beam blur was controlled by changing focus and was measured by Aerial Image Sensor (AIS). Groups of different biased patterns were located at different distances from large pattern on the wafer respectively. As a result, CD variation by proximity effect was 35-40 nm for 100nm-isolated. Under our recommended condition that resizing range in puls side was equal to that in minus side (±20%), blur latitude and dose latitude satisfied our CD uniformity budget and mask enhanced error factor (MEF) was around 1, then reticle fabrication CD controllability from budget requirement was satisfied. Therefore it was shown that proposed proximity effect correction method achieved high CD controllability.

For position measurements of the EPL reticle, a new concept reticle holder is proposed. This holder clamps the same surface during measurement as during exposure in Nikon's EPL tool, the EB Stepper. Thus the holder reproduces the deformation caused by clamping in a metrology tool with that in the EB Stepper. Investigation by simulation is described. Furthermore, an experimental holder based on this concept was manufactured, and the deformation of a 200 mm EPL reticle was measured. The experimental results and simulation results show an advantage of this method.

We are now at a major junction in lithography where non-optical lithographies, such as Electron Projection Lithography (EPL) [1], are being introduced. The mask used in EPL is a non-transparent silicon substrate with a thin silicon (~2m) membrane with openings for electrons to pass through acting as a scatterer. This must be inspected as defects may cause printable defects. Initial mask inspection work has used SEM inspection to find these defects. However, we have historically used optical mask inspection tools, utilising wavelengths at or above what we are using for imaging, to qualify masks. This technology has been increasingly difficult to sustain as we have moved from imaging using mercury lamp based sources to pulsed excimer laser based sources that are not very suited to the inspection imaging. Indeed, review of defects found has moved from optical microscopes to SEM based tools. Inspection tools have also evolved, with the first SEM based mask inspection tools being developed to find the smallest defects, however these have the penalty of very low throughput. We will show the potential of using optical systems for the transmissive inspection of these EPL masks. The high potential of existing tools will be shown together with the need for a next generation of inspection tools. We will show that simulations indicate that an inspection source with 193nm wavelength would be required for the detection of 50nm defects on a mask used to print 70nm dense lines. It will also be shown how the position of the defect within the membrane greatly influences detection as well as the implications of moving to a thinner silicon membrane.

This article presents the results of a collaborative effort between Molecular Imprints, Inc. (MII) and Photronics, Inc. to develop a baseline process for fabricating Step and Flash Imprint Lithography (S-FIL) templates that are compatible with lithography tools being developed by MII. S-FIL is a replication technique with sub-50nm resolution capability that has the potential to lead to a low cost, high throughput process. Template fabrication results and S-FIL patterning results on 200mm wafers are presented.

Step and Flash Imprint Lithography (SFIL) is an alternative lithography technique that enables patterning of sub-100 nm features at a cost that has the potential to be substantially lower than either conventional projection lithography or proposed next generation lithography techniques. SFIL is a molding process that transfers the topography of a rigid transparent template using a low-viscosity, UV-curable organosilicon solution at room temperature and with minimal applied pressure. Employing SFIL technology we have successfully patterned areas of high and low density, semi-dense and isolated lines down to 20 nm, and demonstrated the capability of layer-to-layer alignment. We have also confirmed the use of SFIL to produce functional optical devices including a micropolarizer array consisting of orthogonal 100 nm titanium lines and spaces fabricated using a metal lift-off process. This paper presents a demonstration of the SFIL technique for the patterning of the gate level in a functional MOSFET device.

In the application of the high-energy electron-beam (e-beam) lithogrpahy to the multi-level interconnect metal; the backscattered electron from the heavy metal previously patterned in lower levels on the substrate causes a significant proximity effect. We estimated the "inter-level" proximity effect in the e-beam exposure with the accelerating voltage of 50kV on some multi-level interconnect metal structures which consist in aluminum wiring and tungsten plugs. It was found that the backscattering range and the backscattering energy ratio to the incident energy depend not only on the density and thickness of metal but also on the distance between the resist and the heavy metal plugs. In this paper, a novel proximity effect correction algorithm is proposed, where the exposing patterns are divided into some classes according to the metal structure, the total backscattering energy deposited in the resist is expressed by the sum of the backscattering energy from each structural class, and the exposrue dose is modulated by the function of the total backscattering energy.

Electron beam direct writing (EBDW) system is at the head of systems fabricating circuit patterns by maskless. But the throughput of EBDW is very poor beause very large number of electron beam (EB) shots are requested for exposure of whole patterns on a wafer. We had proposed methods of reduction of the number of EB shots with Character Projection (CP) and designing the best devicve pattern for CP-EBDW to fabricate logic devices such as ASIC or SoC device. Though the method is effective to Front-End-Of-Line (FEOL) layers of cell based logic deviec, Back-End-Of-Line (BEOL) layers cannot be exposed by the method with small number of characters and EB shots. Now, we will propose methods for appropriate CP exposure and data processign for patterns in BEOL layers. By the methods, each BEOL layer in a typical logic device cna be exposed with throughputs about 6 to 8 wafers/h, with a Low-energy-EBDW system produced by e-BEAM Corporation, named "EBIS".

In order to realize SoC (System on a Chip) fabrication at low cost with quick-TAT (Turn-Around-Time) we have proposed a maskless lithography (ML2) strategy, a low-energy electron-beam direct writing (LEEBDW) system with a common character projection (CP) aperture. This paper presents a status report on our proof-of-concept (POC) system. We have developed a compact EB column consisting small electrostatic lenses and deflectors. The experimental results for our POC system indicated that the patterns corresponding to 50nm-node logic devices can be obtained with CP exposure at the incident energy of 5 keV. The technique to reduce the raw process time using a SEM function of LEEBDW system is also reported.

The number of lithographic applications that require ultra-thick layers of resist (>50 um) is rapidly increasing. One field in which this requirement is growing is in the field of MEMS (Micro-Electro-Mechanical-Systems) and MOEMS (Micro-Optical-Electro-Mechanical-Systems). A second is in the realm of wafer-level, advanced chip scale packaging1. For both of these fields, ultra-thick, as well as high aspect ratio, photolithography is an absolute requirement for device processing. The process optimization required to obtain high aspect ratio structures in these ultra-thick photoresist films is extremely challenging. Alignment accuracy is a critical factor in this process. Because of the thick films involved, the alignment keys between mask and wafer could be separated by more than a hundred microns. Standard microscope and objectives have limited depth of focus. Once resist thickness and separation gap exceeds the depth of focus, alignment accuracy suffers from image quality between substrate and mask. Various special microscope options have been developed for alignment of very thick resist. However, most of these processes suffered from one failing or another. This paper addresses a new method to perform true large gap alignment for these applications. True large gap alignment systems overcome these limitations by using identical focal planes for mask and wafer alignment key. Both alignment keys are focused clearly with resist thickness and separation gap of up to 500 um. This paper will present detailed experimental results of alignment accuracy obtained at different thicknesses of photoresist in a contact/proximity production aligner system.

In this communication, we presented a novel approach to reduce critical dimension of resist features. The proposed method involves two process steps: resist pattern formation and critical dimension reduction. This concept was demonstrated to produce sub-100 nm resist features using negative-tone chemically amplified resist, NEB22A2 and Hitachi-900D electron beam lithography system. Experimental results indicate that the degree of critical dimension reduction can be controlled by the dose of flood electron-beam exposure and second wet developer strength. The theoretical results, based on Monte Carlo simulations, were found to be in qualitatively agreement with our experimental observations.

We have implemented traditional CD-SEM metrology complimented with the 3D imaging capability of the VERASEM 3D CD-SEM from Applied Materials. 3D imaging is performed by tilting the SEM beam to capture images at two unique angles. Reconstruction of these images allows for the determination of resist thickness and sidewall angle at the same point the critical dimension, CD, is measured. These three output parameters provide the user with automated multi-metric lithographic process control. We have used these techniques to characterize e-beam lithography of isolated lines in ~0.6μm of negative resist at CDs between ~50 and ~100 nm. The flexibility of our e-beam lithography system allows us to expose an array of identical features with 30 distinct dose values over a small area of a wafer. We have characterized the resist CD and thickness as a function of small incremental decreases in dose. As the dose decreases so does the CD of the isolated resist line at a rate of ~1 nm per 1μC/cm² of area exposure. At a nominally high dose where the isolated line CD is ~100 nm the resist is measured by 3D imaging to be close to full thickness. The main observation is that the resist thickness erodes at a rate of ~5 nm in height per every 1nm decrease in CD down to the resolution limit of 50-60 nm. As the dose is further lowered the resist is then completely washed away. This subtle but significant loss in resist etch mask integrity could not have been observed by traditional top-down CD-SEM metrology alone. This also demonstrates the tilt capability of the VERASEM 3D to measure very thin resist films of ~100 nm. Additionally, we have successfully used this methodology to characterize this effect as a function of isolated line length from ~0.5-2.0μm, and resist thickness from ~0.25-0.6 μm. The CD is strongly correlated with the total isolated line length due to the e-beam proximity effect, while the resist erosion rate remains fairly constant. The resist erosion rate is also similar for the resist films regardless of initial thickness. However, we also confirm the trend that identical area doses produce larger CDs for thicker resist films with some subtle effects for the thinner films.

The discharge pumped excimer laser is a gas laser providing ultra violet (UV) radiation with well defined spectral, temporal and spatial properties. The fast development of excimer lasers in recent years has succeeded in designing very compact, table-top and turn-key systems delivering up to 20 W of radiation at 248 nm, 10 W at 193 nm and 2 W at 157 nm with repetition rates up to 2000 Hz (1, 5). Due to their short emission wavelength and compactness they are continuously replacing other light sources, like lamps and ion lasers, in applications as metrology, inspection, direct writing and material testing. Spatial and temporal beam properties of compact excimer lasers are very suitable to be utilized as illumination source in these applications. The compact excimer laser is combining the advantages of both, lamp and laser sources. It displays low temporal and spatial coherence, but has a narrow spectral emission range of a few hundred pm. The beam area is approximately 1/2 cm², the divergence is in the order of 1 mrad. Variation of beam position and beam direction are negligible for most illumination applications. Compact excimer lasers are easy to integrate in measurement and inspection systems. Typically their footprint area is 0.25 m2. The power consumption is less than 1 kW, enabling single phase electrical supply and air cooling. State-of-the-art compact excimer lasers are compliant to all relevant SEMI regulations. The laser optics exceeds the life time of the laser tube, thus no optics cleaning and exchange is necessary in a whole life time of a laser tube of a few billion pulses (6).

The use of backscatter electron detection in a wafer alignment system has been investigated. For certain types of wafer processing such an alignment system might show improved process robustness compared to optical sensors. This expectation is based on the principle that the alignment signal generated by backscattered electrons is formed by probing the volume of the alignment mark rather than its surface. This paper presents both simulations and experiment results on the viability and potential physical limitations of this alignment method. Physical properties of the backscatter electron alignment system are discussed. The results confirm that for some semiconductor wafer processes, this concept lives up to the expectations. The impact of electron beam energy, shot noise and alignment mark surface roughness on process-induced alignment shift and aligned position repeatability is investigated. In addition, the sensitivity to magnetic fields and mechanical vibrations has been investigated. The theoretically predicted relation between repeatability and illumination dose, due to shot noise has been experimentally confirmed. The results show that backscatter electron alignment is a promising alignment method, although a few issues remain unresolved.

In the world of micro- Lithography, several options exist for obtaining features below the 100nm level. Options include a variety of methods which range from additional process steps in etch, multilayer resist systems, or expensive throughput limited direct write E-beam systems. Each comes with a handful of trade offs in uniformity, repeatability and cost. Collimated (LASER) Plasma Lithography (CPL), on the other hand offers a full field exposure with minimal process intervention to obtain resolution below the 100nm barrier. CPL, uses a membrane 1x proximity mask and a collimated light source with energy peaking at 11 A°. By using a mask, an entire 22mm x 22mm field (30mm x 30mm with the next generation) can be exposed at once regardless of chip density, removing any throughput concerns as well as placement, stitching and typical E-beam machine flaw defects. Collimation, provides a predictable flux of energy to ensure minimal global divergence and energy level variation. Energy at 11 A°, allows for a high level of uniformity and penetration within the resist, without introducing resolution compromising scattering or standing wave effects. This Paper will demonstrate the capabilities of CPL as well as the advantages over traditional lithography in obtaining features below 100nm. We will also depict process techniques which take full advantage of improvements in CAR, and experiments which suggest reduction possibilities through variables in mask fabrication.

Progress toward a new method for performing design optimization of second-generation illuminator optics is described. The method involves constructing a function that approximates the figure of merit over a finite volume of solution space. This approximating function can be evaluated much more efficiently than the actual merit function, and can also be used to compute analytic values for the first and second derivatives. Optimization, subject to the imposition of nonlinear constraints, is achieved by incorporating the approximating function into an existing damped least squares optimizer. When this work is complete, the smoothing nature of the approximation will permit the optimization of noisy merit functions (e.g., Monte Carlo illumination simulations using limited numbers of rays).

This paper presents first results to assess the feasibility of a cleaning strategy for EUV production tools. The EUV experiments were performed in a dedicated UHV contamination chamber connected to the DLW20 dipole beam line at the PTB laboratories at BESSY II in Berlin. An in-situ reflectometry system has been implemented inside the contamination chamber to allow for real-time detection of mirror reflection changes. The reproducibility of the in-situ reflectometry system has proven to be about 0.03%, allowing the measurement of reflection changes below 0.1%. Cleaning cycles were performed at producation tool power density levels, i,e,, 10-30 mW/mm² broadband radiation, on capped Mo/Si mirror samples. Carbon was deposited intentionally at ypical hydrocarbon pressures in the 10^-8 mbar regime. Cleaning was done at background levels of hydrocarbons and 10^-4 mbar molecular oxygen. First results show that cleaning of a mirror at high power densities and typical tool conditions can restore the reflection close to its initial value.

The timely development of a Next-Generation Lithography depends upon its progress in many technology-dependent areas. Common to all high-throughput systems is the requirement of strict mask distortion control. Thus, in support of the Nikon electron-beam projection lithography tool program, three-dimensional finite element (FE) models have been developed to simulate the transient thermal and structural response of a 200-mm prototype stencil mask during electron beam exposure. Due to the relative size of pattern features, equivalent modeling techniques were employed for computational expedience. Equivalent thermal properties (conductivity and emissivity) have been calculated for perforated membranes as a function of pattern void fraction. Resulting temperature distributions were used as input for the FE structural models to characterize and quantify the local displacement fields. The structural models also utilized equivalent material properties (elastic modulus, shear modulus and Poisson's ratio). Support conditions corresponded to electrostatic chucking with four symmetrically located pad regions. The FE simulations predicted that under typical exposure conditions, the localized thermal distortions within the individual subfields are all less than 1.0 nm, which is well below the allotted error budget.

In this paper, we report on evaluation of a high e-beam sensitive CA resist, FEP171 for EUVL mask fabrication. This resist exhibits a PEB temperature sensitivity of ~1nm/°C and 8.3 nm (3 σ) CD uniformity across an EUVL mask patterned with 200 nm dense features using a 100 keV e-beam exposure system. This resist also showed a very high resolution and excellent exposure latitude. Dense line/space features down to 60 nm have been delineated in this resist. This paper will discuss the lithographic performance of this resist and compare it with that of UV-III CA resist and ZEP 520 non-CA resist.

Challenging requirements in optical and BioMEMS application with high aspect ratios of microstructures in access of 20 and smallest structure details of less than 1 μm have motivated this work to irmpove x-ray mask fabrication. Several approaches to pattern an intermediate x-ray mask with the gold absorber thickness of 1.6-2.2μm using a 1μm thick silicon nitride membrane have been explored. E-beam lithography is employed for primary patterning and experimental results show that high energy (100keV) e-beam lithography is a very promising approach. So-called working x-ray mask can be fabricated from intermediate x-ray mask through x-ray lithography. More than 10μm thick PMMA x-ray resist has been coated on the silicon nitride membrane by multi-coating process without crack. First exposure results indicate that adhesion and stability of sub-micrometer structures wiht these heights is critical. In order to overcome these problems a novel approach has been proposed by coating resist on both sides of the silicon nitride membrane and simultaneous patterning of both sides using x-rays. First successful experimental results have been achieved for proving the feasibility.

PREUVE is the French federative program aimed at developing knowledge concerning the critical issues of EUV Lithography. The project supported by the French ministry of industry is one of the most important project in the field of nano-technologies in France. PREUVE is focused on different relevant aspects of the EUV lithography. Different kinds of EUV sources at 13.5nm for the purposes of lithography and metrology have been studied. Also reflective optics and multilayer coatings for the illuminating and the projection cameras have been realized. A complete development of EUV masks including substrates, multilayer coatings and reticules has been realized. Relevant metrologies for reflectivity measurements and defects detection have been developed. Finally the construction of an experimental exposure bench with high magnification (x10) and very high numerical aperture (0.32) has been pursued for process development. A summary of the results obtained in the frame of the project is presented hereafter.

An advanced Monte Carlo model and software were developed to simulate electron scattering in electron beam lithography and signal formation in scanning electron microscopy at a new level of accuracy required for lithography and metrology. The model involves generation of fast secondary and slow secondary electrons, as well as generation of volume plasmons, and electron transfer between layers with regard to the difference between work functions of layers. To track SEM detector channel, the geometry of a detector and its energy transfer function were taken into account. This advanced model was used to simulate electron trajectories, deposited energy, signal from electron detector and images in SEM. Examples of simulations are presented for electron spectra, energy deposition in 50 kV maskmaking, and signals from patterned wafers in SEM.

We describe studies of the debris produced from a high-repetition-rate laser plasma EUVL source based on the mass-limited target concept. In particular, we are developing mass-limited target designs based on complex targets containing tin. Comprehensive analysis of witness-plate detection techniques can reveal many interesting details of the interaction regime, and the impact of the debris. These techniques include Optical Microscopy, Scanning Electron Microscopy, Atomic Force Microscopy, X-ray Photoelectron Spectroscopy, Auger Electron Spectroscopy, and Auger Electron Microscopy. We also describe developments of the repeller field concept of debris inhibition. This technique uses electrostatic fields to reduce the flux of plasma ions impinging on the EUV collimating optics. Here, the first measurements of debris mitigation of a tin-doped target are described, and comparisons with earlier measurements of the impact of repeller fields on ion emission from a mass-limited water-droplet target are made.

One of the key leverage factors in determining the viability of laser-plasma sources for EUVL is the conversion efficiency of laser light to EUV emission in the 13-nm region. We describe experiments and theoretical calculations on a mass-limited laser target design using tin that offers high conversion efficiency.

A new sampling device and collection method has been developed to provide an accurate and convenient means for collecting ultra low levels of condensable organic species. The sample collection and analytical method are optimized to separate, quantify, and identify individual organic components at a detection limit of approximately 0.001 μg/m³, representing a hundred fold improvement over conventional sampling methods. Given the potential threats that these contaminants pose to deep UV lithography optics, the work represents a new step in understanding the long-term risks to production 193nm exposure tools. Devices deployed at an Texas Instruments manufacturing facility have generated data that proves that this sampling device and method is capable of measuring contamination at levels never achieved before. This data will be used to develop correlation of optics degradation with different classes of airborne molecular contamination.

A commercially viable light source for EUV lithography has to meet the large set of requirements of a High Volume Manufacturing (HVM) lithography tool. High optical output power, high-repetition rate, long component lifetime, good source stability, and low debris generation are among the most important parameters. The EUV source, being developed at Cymer, Inc. is a discharge produced plasma source in a dense plasma focus (DPF) configuration. Promising results with Xe as a working gas were demonstrated earlier. To scale the DPF parameters to levels required for HVM our efforts are concentrated on the following areas: (1) thermal engineering of the electrodes utilizing direct water cooling techniques; (2) development of improved pulsed power systems for > 4 kHz operation; (3) study of erosion mechanisms for plasma facing components; (4) development of efficient debris mitigation techniques and debris shields; (5) studies of plasma generation and evolution with emphasis on improving conversion efficiency and source stability; (6) development of EUV metrology techniques and instrumentation for measurements of source size; and (7) development of an optimized collector optic matched to our source parameters. In this paper, we will present results from each of these key areas. The total in-band EUV output energy now approaches 60 mJ/pulse into 2πsr and the conversion efficiency has been increased to near 0.5 %. Routine operation at 4 kHz in burst-mode, and continuous operation at 1 kHz has been demonstrated. Improved at-wavelength source metrology now allows a determination of EUV source size utilizing imaging, and monitoring of key features of the spectrum on a pulse-to-pulse basis. With effective suppression of debris generated from the anode by several orders of magnitude, UV/EUV-catalyzed carbon growth now presents the limit in producing a clean source.

The goal of this paper is to discuss the technical and economic drivers of Moore’s Law, with special emphasis on their interdependence. These drivers can be classified as “push” (technology improvements push us down the Moore’s Law path) and “pull” (economic incentives create increasing chip production volumes which drive the technology learning curve). In fact, Moore’s Law can be considered as an instance of general learning curve theory, which places a special importance on the role of increasing chip volumes on the slope of Moore’s Law. After a general review of the history of Moore’s Law, its relevance today is described in terms of transistor shrinks rather than increasing functions per chip. The challenges required for the continuation of Moore’s Law are presented.