This PDF file contains the front matter associated with SPIE Proceedings Volume 6519, including the Title Page, Copyright information, Table of Contents, Introduction, and the Conference Committee listing.

Chemically amplified photoresists are likely to remain the primary imaging materials for the semiconductor industry. As feature sizes decrease to dimensions comparable to the characteristic size of the molecules in the photoresist, a significant challenge lies in identifying the ultimate resolution limit of these materials. To address this challenge, we investigated model photoresist materials with high resolution measurements to examine the effect of individual factors among interdependent process steps on line-edge roughness (LER). Using a bilayer film sample geometry, we measured the internal deprotection interface with nanometer resolution as a function of photoacid size, initial resist copolymer composition, and amine base quencher by neutron reflectivity and infrared spectroscopy. After development, we found that the resist chemistry and additives can play an important role in LER through its influence on acid diffusion. However, these model experiments suggest that there is a limit in LER even with an idealized exposure image contrast and decreases in the width of the reaction-diffusion front. However, there may be opportunities to further decrease LER during development by tuning the response of the photoresist to the developer solution.

Patterning technology is entering the nanomaterials era. Breakthrough advances in the basic sciences over the last twenty years are catalyzing novel material and assembly options. In the near future, these options may warrant consideration for fabricating advanced information processing technologies. However, material technology advances alone are not sufficient to induce changes and chemical substitution in manufacturing. In fact, manufacturing technology will change only when no other option exists. Therefore, the concurrent trend in lithographic challenges is noteworthy. Recent revisions of the International Technology Roadmap for Semiconductors (ITRS) indicate that it is becoming increasingly difficult for mainstream lithographic technologies to satisfy projected ITRS dimensional scaling requirements, shown in Table 1.

Immersion lithography has placed a number of additional performance criteria on already stressed resist materials. Much work over the past few years has shown that controlling the water-resist interface is critical to enabling high scan rates (i.e. throughput) while minimizing film pulling and PAG extraction (i.e. defectivity). Protective topcoat polymers were developed to control the aforementioned interfacial properties and emerged as key enablers of 193 nm immersion lithography. Achieving the delicate balance between the low surface energies required for high water contact angles (generally achieved via the incorporation of fluorinated groups) and the base solubility required for topcoat removal is challenging. More recently, additional strategies using fluoropolymer materials to control the water-resist interface have been developed to afford topcoat-free resist systems. In our explorations of fluoroalcohol-based topcoat materials, we have discovered a number of structure-property relationships of which advantage can be taken to tailor the interfacial properties of these fluorinated materials. This paper will address the effect of structure on immersion specific properties such as water contact angle, aqueous base contact angle, and dissolution rate.

Recently it is known well that blending hydrophobic additives into conventional resist polymer drastically improve its film surface hydrophobicity. So we thought that this approach was one of candidates and most promising to achieve a non-topcoat resist process for immersion lithography. And it would be able to maintain original resist performance because only a small amount of additives were added into conventional resist. Then we have investigated hydrophobic polymers for use as additives of non-topcoat resists. We have newly successfully synthesized various new highly fluorinated monomers by our peculiar fluorination process. We found that some specific methacrylate ,which have perfluorinated cyclic structure, showed excellent hydrophobicity. The other hydrophobic candidates is our fluoropolymer, FUGU, which had already developed, having partially fluorinated monocyclic structure. However its hydrophobicity is insufficient due to presence of acidic hydroxyl group which act as dissolution unit into the developer. To improve the its hydrophobicity, we protected all or a part of its hydroxyl group. The protected FUGU polymer provide good hydrophobicity whose sliding angle (S.A.) and receding angle (R.A.) were 7 degree and 90 degree respectively , compared to original polymer, FUGU. In this paper, we describe a characteristics and evaluation of these our hydrophobic polymers to apply to additives for non-topcoat resists. We have optimized these polymers to apply to additive for conventional resist. As a result, various kinds of additives were obtained. For example, some of them dissolve in developer due to the presence of alkali soluble group in the polymer, the others are soluble in developer after deprotection reaction by post exposure bake. We call the former one is 'top-coat type', the latter is 'resist type'. Two type additives were investigated to give the hydrophobicity and to depress the leaching amount to conventional resist.

Over a period of last several years 193 nm immersion lithography from a remote and unlikely possibility gradually became a reality in many fabrication facilities across the globe and solid candidate for high volume manufacturing for the next generation technology node. It is being widely understood in the industry that top-coatless resist approach is a desirable final stage of the immersion process development. However creating low-defect high performance top-coatless resist materials requires understanding of the fundamental material properties of the top layer, responsible for leaching suppression, immersion fluid meniscus stability, and in this way enabling high speed low-defect scanning. While a lot of progress has been made in implementing specific top coat materials into the process flow, clear understanding effects of the top coat properties on the lithographic conditions and printing capability is still lacking. This paper will discuss top coat materials design, properties and functional characteristics in application to novel fluoroalcohol polymer-based immersion top coat. We have used our fluoroalcohol based-series designs (titled MVP top coat materials further on in the paper) as a test vehicle for establishing correlations between top coat performance and its physical and chemical properties including hydrophobicity, molecular weight/dispersity etc. Effects of polymer-solvent interactions on the contact angle and characteristics of the top coat material are explored, providing valuable understanding transferable to design of new generation top coats and top-coatless materials. Our resultant new designs demonstrated excellent lithographic performance, profiles and low leaching levels with commercially available resist and high receding contact angles, comparable to the commercial top coat materials.

The technology of 193nm immersion lithography has been progressing rapidly toward half-pitch 45 nm generation device manufacturing. However, some intrinsic issues, the photoacid leaching and the watermark defect have remained in the immersion process. Most of approaches to overcome them were the introduction of cover coating materials (top coat) onto the resist film. Recently, we have established the non-top coat resist using novel two kinds of materials, a low leaching PAG (PhotoAcid Generator) and a surface modifier. The hydrophobic photoacid generated from the low leaching PAG decomposes by heating, and the acid migration changes to reduce the line-width roughness (LWR). The surface modifier behaves as builded-topcoat by the distribution around resist surface and enhances the surface hydrophobicity. Herein we propose the concept of novel PAG and surface modifier for immersion lithography.

A preliminary Quantitative Structure Property Relationship (QSPR) model for predicting the refractive index of small molecules and polymers at 193 nm is presented. Although at this stage the model is only semiquantitative we have found it useful for screening databases of commercially-available compounds for high refractive index targets to include in our program of synthesis of high refractive index resist polymers. These resists are targeted for use in 2^nd and 3^rd generation 193 nm immersion lithography. Using this methodology a range of targets were identified and synthesized via free radical polymerization. Novel resist polymers were also synthesized via Michael addition polymerization. Preliminary dose to clear experiments identified a number of promising candidates for incorporation into high refractive index resist materials. Furthermore, we have demonstrated imaging of a high index resist using water-based 193 nm immersion lithography.

A series of experiments were designed to probe the interaction between second generation High Index Liquids (HIL, n=1.65) and the resist stack. Three different second-generation high index liquids were tested in five experiments: measurement of the contact angle of the liquid with the resist surface; leaching of Photo-Acid Generator (PAG) into the liquid; residue analysis of droplets evaporated from the resist surface; impact of liquid soaking on resist profiles; and imaging through high-index liquids at 72nm pitch. The selected liquids were the main candidates from two potential vendors. In parallel, tests have also been done for water. The tests show that one of the main differences between highindex liquids and water is their much smaller contact angles on the organic photoresist films. This contact angle can be influenced by a topcoat, but currently seen contact angles may force a new immersion hood concept. Imaging was not affected strongly by the high-index liquids. For some liquids, low evaporation rates and a tendency to leave residue on resist were observed, which may require a dedicated liquid removal strategy to reduce defectivity.

High-refractive-index fluids (HIFs) are being considered to replace water as the immersion fluid in next generation 193nm immersion scanner. At SPIE 2006, we have demonstrated the attractive optical properties and good imaging performance for our HIF candidates, HIL-001 and HIL-002. In this paper, we will describe our latest results on the remaining issues for the practical application of HIF candidates, as well as introduce 3^rd generation fluids for the further extension of ArF immersion lithography. In order to improve the fluid transparency, we have tried two approaches. One is the improvement of transparency for HIL-001 based on a refining technology and the other is to develop a novel HIF candidate by using computational chemistry, which is named HIL-203. By passing through a suitable refining unit, HIL-001 can reach a transmittance of >99%/mm, which is as high as water. This new purification method can be applied to an on-site reuse system. It was also found that the refining unit was very effective to eliminate the impurities coming from the photo-degradation of HIL, chemical substances contamination under the air exposure, and leaching of resist components such as photo-acid generator or quencher. We have developed a new fluid for 3^rd generation immersion fluids. It had a higher refractive index than that of HIL-001 or HIL-203; however, it still falls short of our target value. Additionally, by using a novel design concept, we have developed a top-coat with high refractive index for HIL immersion lithography, which gave an appropriate contact angle for scanning exposure.

The demand for even denser semiconductor devices is driving efforts to reduce pattern dimensions in semiconductor lithography. It has been found that 193-nm immersion lithography technology can achieve smaller patterns without having to modify the infrastructure technology of existing state-of-the-art 193-nm dry lithography. This has made 193-nm immersion lithography a promising technology for next-generation mass production processes. It is now under full-scale development and is about to enter a commercial stage applicable to mass production. In 193-nm immersion lithography, the space between the optical projection system and silicon wafer is filled with liquid thereby immersing the resist film in de-ionized water during exposure. This generates a number of concerns, such as the penetration of moisture into the resist, the leaching of resist components into deionized water, and the presence of residual moisture, all of which can lead to defects that can affect post-processing. It has been reported; however, that rinse processing before and after exposure can be effective in reducing such defects [1]. Also, the trend toward finer patterns has resulted in large aspect ratios that can lead to pattern collapse, but the application of a surfactant has been found to ease this problem. Controlling the drying process after rinsing has also been found to reduce residue adhering to the resist. The stability of process data is considered to be a crucial factor in the adapting of 193-nm immersion lithography to mass production. In this report, we obtain long-term data on defects and critical dimension (CD), examine this data for process stability, and discuss the applicability of 193-nm immersion lithography to mass production.

Volume production of 55nm node device using Nikon immersion lithography tool, S609B, has started. One very important issue for volume production immersion lithography is defectivity control. The mainstream resist process for the initial stage of immersion is to use a topcoat, because chip manufacturers can continue to use conventional resist whose lithographic properties are well characterized. On the other hand, because of its simplicity and cost effectiveness, a topcoat-less resist process is an attractive candidate for the second stage of immersion lithography. Intensive study and development of topcoat-less resist is being done by resist vendors. One unexpected benefit of topcoat-less resist is it is much easier to increase hydrophobicity than developer soluble topcoat. For defectivity analysis of both topcoat and topcoat-less process, a dedicated immersion cluster comprised of a volume production immersion exposure tool, S609B, with NA of 1.07 and a resist coater-developer LITHIUS i+ (TEL) is used. Excellent defectivity data obtained using various materials and processes are shown in this paper. Understanding the root cause of defects is very important to control them. Defect source analysis (DSA) revealed particles cause many types of defects. Interesting correlations between particles and other topcoat defects and resultant pattern defects are shown in this paper. In the immersion resist process, wafer edge condition is one key issue. Using topcoat-less resist, wafer edge condition and its defectivity dependency is discussed and analyzed.

With immersion lithography approaching the insertion in production, watermarks remain as one of the main concerns for immersion specific defects. They require special attention because of their size and associated high kill-ratio, and their increasing occurrence at higher scan speeds. IMEC has been working to understand the underlying mechanism of why remaining water droplets cause these defects. This work focuses on water uptake measurements and how this parameter correlates to watermark defectivity. Ellipsometric Porosimetry (EP) is used to measure the water uptake tendencies of resist and top coat materials and stacks thereof, and investigate what parameters are affecting it. The influence of material and process parameters and the presence of a top coat on water uptake by the resist are evaluated. In parallel, the quartz crystal microbalance (QCM) technique has been used as an alternative option to measure the water uptake. Though a one-to-one comparison between the results is not straightforward, the main trends are identical for both techniques. No perfect correlation of watermark defectivity with water uptake has been found in this study. Nevertheless, the results show a tendency towards higher watermark sensitivity with higher water uptake by the film. It is recognized that the total watermark defectivity is most probably a complex interplay of different parameters with water uptake being only one of them.

The combination of immersion lithography and reticle enhancement techniques (RETs) has extended 193nm lithography into the 45nm node and possibly beyond. In order to fulfill the tight pitch and small critical dimension requirements of these future technology nodes, the performance of 193nm resist materials needs to further improve. In this paper, a high performance 193nm photoresist system based on fluorosulfonamide (FSM) is designed and developed. The FSM group has good transparency at 193nm. Compared to the commonly used hexafluoroalcohol (HFA) group, the trifluoromethyl sulfonamide (TFSM) functionality has a lower pKa value and contains less fluorine atoms. Polymers containing the TFSM functionality have exhibited improved dissolution properties and better etch resistance than their HFA counterparts. Resists based on the FSM-containing polymers have shown superior lithographic performance for line, trench and contact hole levels under the 45nm node exposure conditions. In addition, FSM resists have also demonstrated excellent bright field and dark field compatibility and thereby make it possible to use one resist for both bright field and dark field level applications. The structure, property and lithographic performance of the FSM resist system are reported.

The dominant current 193 nm photoresist platform is based on adamantane derivatives. This paper reports on the use of derivatives of diamantane, the next higher homolog of adamantane, in the diamondoid series, as monomers in photoresists. Due to their low Ohnishi number and incremental structural parameter (ISP), such molecules are expected to enhance dry etch stability when incorporated into polymers for resist applications. Starting from the diamantane parent, cleavable and non-cleavable acrylate/methacrylate derivatives of diamantane were obtained using similar chemical steps as for adamantane derivatization. This paper reports on the lithographic and etch performance obtained with a number of diamantane-containing monomers, such as 9-hydroxy-4-diamantyl methacrylate (HDiMA), 2-ethyl-2- diamantyl methacrylate (EDiMA), and 2-methyl-2-diamantyl methacrylate (MDiMA). The etch advantage, dry and wet lithographic performance of some of the polymers obtained from these diamantane-containing polymers are discussed.

Combining optical and electron beam exposures on the same wafer level is an attractive approach for extending the usefulness of current generation optical tools. This technique requires high-performance hybrid resists that perform equally well with optical and e-beam tools. In this paper Rohm and Haas EPIC^TM 2340, a 193-nm chemically amplified photoresist, is used in a hybrid exposure role. The e-beam tool was used to pattern 45 nm half-pitch features and a 193- nm immersion stepper was used to pattern 60-nm half-pitch features in the same resist layer. The effects of processing parameters and delay times were investigated.

We have investigated relationship between immersion topcoats and water, and between topcoats and ArF resist films for the use in ArF immersion lithography from the stand point of the work of adhesion characteristics. We evaluated surface free energy of topcoat films and resist films each from the contact angle measurement. From values of measured free energy, we obtained work of adhesion between topcoats and water, and between topcoats and resist films. In addition, we calculated an interfacial free energy between topcoats and resist films, which is related to the interface stability. As a result of evaluation of the interfacial free energy of four different kinds of topcoats, the topcoat which has lower surface free energy was found to have lower work of adhesion between topcoat and water, and lower interfacial free energy between a topcoat and a resist. These results indicate that the topcoat which has lower surface free energy has more less interaction between water and topcoat and stabile interface between a topcoat and a resist film.

We propose a non chemically-amplified positive-tone photoresist based on photolysis of o-nitrobenzyl phenol ether (NBP). The increase in the amount of the phenolic hydroxyl group just after the exposure to the i-line propagation light is observed via IR spectroscopy. Using near-field lithography (NFL) combined with the NBP, we form half-pitch (hp) 32 nm line and space (L/S) patterns with lower line edge roughness (LER) than those of a chemically amplified resist (CAR). The high-resolution feature of the NBP is attributed to the photoreaction system without the acid diffusion, which is inherently involved in CARs, although the NBP requires six times as much exposure dose as the CAR does. A Hp 32 nm L/S patterns with 10 nm depths are successfully transferred to the 100 nm thick bottom-layer resist through the tri-layer resist process. Hp 22 nm L/S patterns with 10 nm depths are also fabricated on the top portion of a single-layer of NBP.

Concentration gradients of photoacid generator through the thickness of the photoresist film can profoundly affect the material's performance. To engineer the acid concentration through resist thickness, we have developed a new type of resist adhesion promoting layer that incorporates photo acid generator chemistry. These adhesion promoting photo acid generators, called as a class "APPAG" enhance acid concentration at interface between the resist and the substrate. We will provide an overview on the preparation and characterization of two siloxane based APPAG materials along with a performance comparison of commercial DUV, EUV and E-beam photoresists on APPAG. Nonaflate analog (APPAG 6) with shorter acid diffusion length was found to have a mild impact on 250nm node DUV lithography. However the triflate analog (APPAG 9), owing to the larger acid diffusion length, was shown to provide a greater influence. APPAG 9 was found to give nearly a 50% improvement in depth of focus. For EUV lithography, both APPAG 6 and APPAG 9 will be shown to substantially improve performance envelope for 100nm dense lines and spaces and at reduced post exposure bake (PEB) temperatures. This indicates that this approach can be used to gain margin at reduced PEB which is desirable to minimize thermally driven diffusion effects. Thus the materials represent an important new approach to extending photoresist performance margins.

We will discuss our approach towards a second generation radiation sensitive developable bottom antireflective coating (DBARC's) for 193 nm. We will show imaging results (1:1 L/S features down to 140 nm) for some first generation implant resist material based upon a fluorinated resins and also show relative implant resistance of these first generation fluorinated resists towards As implantation (15 KeV at 5x10¹⁵ dose with 20 x 10^-4 amp). Also, discussed will be a second generation of implant resists based on a non-fluorinated resins. Surprisingly, we found that the nonfluorinated materials gave better implant resistance (~2-3 X10¹¹ atoms/cm²) despite the higher atomic number of fluorine compared to hydrogen in the fluorinated implant materials (~2-5X10¹² atoms/cm²). Finally, we will give an update on the lithographic performance of this second generation of implant resists.

New challenges face ArF bottom antireflection coatings (BARCs) with the implementation of high NA lithography and the concurrent increase use of spin-on hard masks. To achieve superior reflectivity control with high NA at least two semi-transparent ARC layers, with distinct optical indices, are necessary to effectively lower substrate reflectivity through a full range of incident angles. To achieve successful pattern transfer, these layers in conjunction with the organic resist, should be stacked with an alternating elemental composition to amplify vertical resolution during etch. This will circumvent the inherent low etch resistance of ArF resist and the decreasing film thicknesses that accompanies increasing NA. Thus, incorporating hard mask properties and antireflection properties in the same two layer system facilitates pattern transfer as a whole rather than just enhancing lithography. As with any material expected to exhibit multiple roles there is a delicate balance between optimizing materials with respect to one of its roles while not impairing its other roles. We will discuss some of these conflicts and present Si-BARCs and carbon rich underlayers which aim to balance these conflicts. In this paper we will explore simulations aimed at finding the best film thicknesses and optical indices, etch rate selectivity, and lithographic performance of high silicon content and high carbon content BARC materials designed to meet the demands of both high NA lithography and trilayer processing.

Multi-layer stack application will be established at manufacturing process beyond 65nm node. Multi-layer stack application means laminated layer like a Photoresist, Si contained hard mask, and Carbon hard mask on substrate. Multilayer stack application can be solved anti-reflection at hyper NA (that means more than 1 numerical aperture) and less etching resistance in thinner film resist, so Multi-layer stack application is required for 193nm immersion lithography process. And criteria of our material in Multi-stack application are spin-on and drain compatible type. In this report, we will discuss about Spin-on Si-contained hard mask and spin-on carbon hard mask criteria, our experiment and results to solve issue. Spin-on Si-contained hard mask is required 3-factors that is unti-reflection from substrate at hyper NA conditions, resist matching, and higher etching resistance. It is general that higher Si-content ratio in based polymer can't be matched with current 193nm photoresist. But Lower Si-contained Hard Mask can't be resisted by dry etching. In this report, we will discuss about our material approach for good resist matching (no footing issue) without reduction of Si-content ratio, pattern transfer ability by dry etching, and reflectivity simulation results at Hyper-NA condition. LWR issue after dry etching is key factor of Multi-layer stack application. We estimate that composition of based polymer in carbon hard mask material and film density in carbon Hard Mask is relative to LWR issue after dry etching. In this report, we will discus about our material approach for less LWR issue after dry etching.

As the feature sizes of integrated circuits shrink, thinner photoresist coating should be used in order to avoid high aspect ratio which can cause pattern collapse. Especially for 193 nm lithography, photoresist coating is too thin to subsequent etching step. One of the solutions to this problem is using hardmasks which have good etch selectivity to adjacent layers. In this paper, silicon-based anti-reflective spin-on hardmasks (Si-SOH) are described. One of the major problems of silicon based polymers in the hardmask compositions is poor storage stability because silanol group is reactive enough to condense each other, which can instigate molecular weight increase to yield gel-type particles. The storage stability of our hardmask materials have been improved by thermodynamically controlled synthesis and reactive mask strategy. Especially the reactive masked silanol groups can take part in crosslinking reaction under the process conditions without additional deprotection step. Although this strategy could encounter intermixing problems with other layers, we can produce silicon-based hardmasks without any deleterious effects. These hardmasks show antireflective properties and great etch selectivity to both photoresists and organic hardmasks (C-SOH).

While much work has been done in the design of photo resist for EUV lithography, these materials have typically been optimized for so called "standard developer" i.e., 2.38% tetra methyl ammonium hydroxide. However we felt that it would be reasonable to consider specifically the developer as opposed to the resist design. Indeed it has been suggested that the polarity and cation size in developer are important positive tone resist performance. It is our hypothesis that a base that could wet and penetrate faster into partially deprotected resist could result in a faster photo speed, and thus make more process margin available for resist design; for example a slower system incorporating higher quencher loadings. Additionally, we sought to probe the effects of solvent polarity with varying amounts of non-aqueous solvent additive. By reorganization of the nascent solvent shell with the non aqueous additives, we sought to perturb the development kinetics and thus change the resist's performance envelope by accelerating photo speed and potentially increasing contrast. This approach has been applied to non chemically amplified resist to good effect. In the three positive tones EUV and a 193nm photo resist was evaluated with the prototype developers we found that the performance was profoundly impacted by these two probes (i.e. solvent polarity and cation hydrophobicity).

The lithography prognosticator of the early 1980's declared the end of optics for sub- 0.5&mgr;m imaging. However, significant improvements in optics, photoresist and mask technology continued through the mercury lamp lines (436, 405 & 365nm) and into laser bands of 248nm and to 193nm. As each wavelength matured, innovative optical solutions and further improvements in photoresist technology have demonstrated that extending imaging resolution is possible thus further reducing k₁. Several author have recently discussed manufacturing imaging solutions for sub-0.3k₁ and the integration challenges. Our industry will continue to focus on the most cost effective solution. What continues to motivate lithographers to discover new and innovative lithography solutions? The answer is cost. Recent publications have demonstrated sub 0.30 k₁ imaging. The development of new tooling, masks and even photoresist platforms impacts cost. The switch from KrF to ArF imaging materials has a significant impact on process integration. This paper will focus on the need to increase the refractive index of ArF photoresist systems to enhance process capability for ultra-high NA's that are near the limitation of the immersion fluid. Data will be presented demonstrating the impact of higher refractive index photoresist systems have on the further extension of ArF Immersion. Advanced RET's will be incorporated to further explore improvements in critical imaging levels along with dominant mask effects.

The chemically-amplified resists have been exposed by hyper-NA 193nm immersion and EUV lithography. Patterns with 45nm half-pitch and below are investigated for process windows and line-edge roughness. Although the 193nm immersion and EUV lithography have totally different optics, an overlap of the resolution capability is clearly observed around 45nm half-pitches. Both lithographic processes show comparable process windows for 45nm dense lines. The 193i resist better responds to its aerial image than that of the EUV resist. Although the EUV tool has the resolution capability down to 20nm half-pitch, immature resist process limits the current resolution to 35nm half-pitch.

Scaling of designs to the 45nm or future nodes presents challenges for KrF lithography. The purpose of this work was to explore several aspects of ArF lithography for implant layers. A comparison of dark loss seen in a KrF resist and TARC system to that seen in an ArF system showed significant differences. While the KrF resist yielded dark loss that varied with CD and pitch, the ArF resist showed very little dark loss and no significant variation through the design space. ArF resist were observed to have marginal adhesion to various substrates. Improvements in adhesion performance were shown by pre-treating the substrate with various processes, of which an ozone clean provided the best results. Optimization of the HMDS priming conditions also improved adhesion, and it was observed that the HMDS reaction proceeds at different rates on different subsatrates, which is particularly important for implant layers where the resist must adhere to both Si and SiO₂. The effect of ArF resist profile with varying reflectivity swing position is shown, and some investigation into reflectivity optimization techniques was performed. Low-index ArF TARC was shown to reduce the CD variation over polysilicon topography, and wet developable BARC was demonstrated to provide consistent profiles on both Si and SiO₂ substrates. Finally, a comparison of ArF and KrF resists after As implant indicates that the ArF resist showed similar shrinkage performance to the KrF resist.

Time-of-flight secondary-ion mass spectrometry (TOF-SIMS) imaging is demonstrated as a sensitive qualitative method for characterizing surface acid concentrations and accompanying chemical changes at resist surfaces. We show its utility in analyzing the 'chemical flare' phenomenon associated with some chemically amplified photoresists. Two commercial 193 nm photoresists were studied: 'photoresist A' displays lithographic defects linked to chemical flare at die edges; 'photoresist B' does not. TOF-SIMS imaging of the surface of 'photoresist A' following exposure and post-exposure bake (PEB) reveals that die edge defects are well correlated with pronounced emanation of surface acid concentrations from, and blocking group depletion beyond, the die edge. Both photoresists also exhibit longer-range surface acidification that is not well correlated with lithographic effects. A plurality of evidence leads us to infer that photoacid migration from exposed to unexposed regions underlies the lithographic defects observed in 'photoresist A.'

Limits to the lithography process window restrict the scaling of critical IC features such as holes (contact, via) and trenches (required for interconnects and double patterning applications). To overcome this problem, contacts or trenches can be oversized during the exposure, followed by the application of a shrink technique. In this work, a novel shrink process utilizing plasma-assisted polymer deposition is demonstrated: a polymer is deposited on the top and sidewalls of photoresist by alternating deposition and etch steps, reducing the dimension of the lithography pattern in a controlled way. Hence very small patterns can be defined with wide process latitudes. This approach is generic and has been applied to both contacts and trenches. The feasibility of the plasma-assisted shrink technique was evaluated through extensive SEM inspections after lithography, after shrink, and after etch, as well as through electrical evaluations.

In this paper, the use of a single Figure-of-Merit to judge resist performance with respect to line width roughness, resolution and sizing dose is proposed and evaluated. Chemically amplified photoresists used in advanced lithography nodes need to fulfill stringent requirements for a considerable number of resist and process characteristics. Along with resolution, line width roughness and resist sensitivity are important examples where the specifications have become very tight. Previously, it has been shown that resolution, line width roughness and resist sensitivity are fundamentally interdependent. Hence, when evaluating or optimizing resist performance it is very important to take these three characteristics into consideration simultaneously. We propose to combine these characteristics in a single photoresist Figure of Merit K_LUP. This Figure of Merit, which is determined from sizing dose, imaging wavelength, exposure latitude, acid diffusion length, line width roughness and pitch allows for a direct comparison of very different resist formulations independent of the exposure tool used. Thus, K_LUP has great potential to assist in evaluating resist performance for the next lithography nodes, for both ArF and for EUV wavelengths.

A simple analytical model for line-edge roughness in chemically amplified resists is derived from an accounting of stochastic fluctuations of photon ("shot noise") and acid number densities. Statistics from this counting exercise are applied to a region defined by the effective acid diffusion length; these statistics are then modulated by the slope of the image intensity to produce a value for LER. The model produces the familiar dependence of LER on aerial image (more specifically on latent image) and dose also seen in many other models and data. The model is then applied to the special case of interference imaging, for which the aerial image is a simple, known analytic function. The resulting expression is compared to experimental data at both relatively large half-pitches, shot with 257nm, and sub-50nm half-pitches shot with 13.5nm and hyper-NA 193nm. The model captures the primary scaling trends seen at the larger length scales, however at the sub-50nm problems arise. It appears that additional effects not covered by counting photons and acids are becoming increasingly important as length scales drop below about 50nm. These additional effects will require increased attention in order to improve LER in lockstep with diminishing CD and pitch.

We have developed an improved AFM-based technique to measure intrinsic material roughness (IMR) after base development. We have found that similar results can be obtained by measuring the film roughness with a fixed develop time and variable dose to that of the previously reported interrupted development method, and thus a simple contrast curve can yield information on the innate material roughness of the exposed resists. It was found that the IMR is dependent on the PAG and the polymer employed in the resist. The IMR of the resist is also strongly dependent on the bake conditions, with increasing IMR at higher bake temperatures. Several PAGs have been identified that result in significantly lower material roughness and thus the potential for significantly reduced line width roughness in resist imaging. Evidence is presented that PAG segregation during the bake steps is responsible for increased IMR in exposed resists, presumably by increasing the dissolution rate inhomogeneity on a nano-scale level. It is also shown that the effects of PAG segregation can be mitigated by the choice of PAG and polymer for the resist.

Physically-based photoresist models, such as those in PROLITH, have been very successful in describing photolithography from a continuum standpoint. These models allow engineers to accurately predict the final resist CD on the wafer and to analyze process robustness. However, as the critical dimension continues to shrink, yield-limiting phenomena are observed that are related to the molecular nature and reaction kinetics of photoresist materials. An example of these phenomena is line-edge roughness (LER). In this paper, the origin of LER is hypothesized to be caused by fluctuations occurring in the initial position of the reactants, fluctuations during the exposure process (shot noise) and fluctuations occurring during thermally-induced reaction-diffusion (post-exposure bake). We have developed a lattice-based mechanistic simulator to better understand the stochastic nature of reactant initial position, the exposure step, the importance of the discrete nature of the reactants, the coupling to the deprotection kinetics and the deep complexity evident in the diffusion-limited acid-quencher reaction.

With the continued shrink of integrated circuit fabrication groundrules, the achievement of good critical dimension (CD) control becomes more and more dependent on optical proximity correction (OPC). Current simulation capabilities involve a first principle aerial image simulation algorithm, such as the transmission cross coefficient (TCC) algorithm, and a resist model, which captures the dynamics of the chemical amplification and the developing process. In the past few years, it has been found that the key photolithographic parameters, such as, the exposure latitude (EL) and the mask error factor (MEF) for the dense features can be very accurately simulated by the algorithm in which the latent image are made by simple Gaussian diffusion of the aerial image. However, more detailed comparison between the simulation and experiment in isolated features, or two-dimensional features indicates that the current modeling algorithm is still not good enough. This may have resulted in the fact that even advanced model based OPC may require hundreds, even more than a thousand experimental data points for model building. In addition, the model made with such endeavor does not usually extend well beyond the minimum groundrule, which can cause sub-groundrule test structures to fail. We have studied one process parameter of the photoresists, the MEF, and we found that a single Gaussian can not explain well the CD behavior and it seems that such difference varies with different resists. The effective diffusion length of some resists are found to vary with line to space ratio within a dense pitch.

The spatial distribution of polymer photoresist and deuterium labeled developer highlights a fraction of material at a model line edge that swells, but does not dissolve. This residual swelling fraction remains swollen during both the in situ development and rinse steps uncovering that the final lithographic feature is resolved by a collapse mechanism during the drying step. We demonstrate that contrast variant neutron reflectivity provides a general method to probe the nanometer resolved in situ development and rinse process step.

Recent experimental results indicate that current resists lack the ability to simultaneously meet the 2005 International Roadmap for Semiconductors (ITRS) goals for Resolution, Line Edge Roughness (LER) and Sensitivity (RLS). This RLS tradeoff has also been demonstrated through modeling work. Here we use a model to explore the impact on the RLS tradeoff of anisotropic acid diffusion and increased quantum yield. We show that both these effects can significantly improve the RLS tradeoff.

Etch resistance and post etch roughness of ArF photoresists still remain some of the critical issues during process integration for sub-100nm technology nodes. Compared to phenyl-containing KrF polymers, methacrylate polymers commonly used for ArF lithography show weak bulk etch resistance in addition to a highly damaged surface after standard etch processes. Counter to the photoresist, the etch rates of BARC are required to be very fast to prevent degradation of the photoresist before the image has been transferred to the substrate. There are a number of etch models in the literature which attempt to describe the correlation between polymer structure and blanket etch rates. Ohnishi Parameter and Ring Parameter are the most common etch models correlating atomic and structural trends in the resist polymer and etch rates. These etch models have been tested in two ways: systematically changing the composition of a terpolymer and using polymers with different functional groups. By comparing the etch rates of this large series of polymer structures it was found that these etch models were not sufficient in describing the relationship between the atomic or structural trends in polymer with etch rates. New etch models that describe the structure property relationship and etch rate trends have been developed. These new models show a better correlation with the observed experimental results. Finally, new polymers have been developed, for both ArF and BARC applications. These materials show a significant improvement in term of etch properties.

Characteristic matrix method was applied to study the anomalous Quartz-Crystal-Microbalance (QCM) data such as sequential peaks of resonance frequency and impedance. The formation of swelling layer was found to cause the broad peak of resonance frequency at the beginning of development. An impedance peak followed the frequency peak when the rigidity of the swelling layer decreased to ca. 10⁵ Pa due to absorption of the developer. Constant thickness of a swelling layer makes double sets of sequential peaks in frequency and impedance. Continuous changes in swelling layer thickness caused a single set of peaks during development. A fast data-acquisition system for QCM was set up and showed the sampling time of 1 kHz with the accuracy of 0.5 nm. It gives good and precise information for resist of large dissolution rate and interface region between resist and substrate.

We have applied chemical force microscopy (CFM) to probe the surface roughness of partially developed model resist materials in order to understand the fundamental materials properties of the resists leading to line edge roughness (LER). CFM is capable of providing simultaneous information about surface topography and chemical heterogeneity of partially developed resist films. We have used CFM to study ESCAP type resists that are used in 248 nm and extreme ultraviolet (EUV) lithography. We observe changes in both the innate material roughness and chemical heterogeneity of the resist with the introduction of photoacid generator (PAG) and with exposure and post exposure bake (PEB). We find several mechanisms by which chemical heterogeneity can contribute to increasing the innate material roughness of the resist.

A general approach to characterize compositional heterogeneity in polymer thin films using Fourier transform infrared (FTIR) spectroscopy has been demonstrated Polymer films with varying degrees of heterogeneity were prepared using a model chemically amplified photoresist where a photoacid catalyzed reaction-diffusion process results in the formation of methacrylic acid (MAA)-rich domains. Within these domains, the carboxylic acid groups dimerize through hydrogen bonding. FTIR measurements of the relative fraction of hydrogen-bonded versus free carboxylic groups are used to quantify the degree of compositional heterogeneity. It was shown that the degree of the spatial heterogeneity varies with changes in the deprotection level and initial copolymer composition. The degree of heterogeneity is small at very low and very high deprotection level and maximize when the deprotection level is around 0.25. Increased non-reactive comonomer content decreases the degree of heterogeneity by reducing the hydrogen bonding efficiency.

We have developed an AFM-based technique to measure intrinsic material roughness (IMR) after base development. Employing this technique we have deconstructed the resist into component parts and have shown that PAG is a major contributor to intrinsic material roughness. When PAG is exposed and thermal polymer deprotection is allowed to occur increased levels of IMR are present. The IMR of the resist is strongly dependent on the bake conditions, with increasing IMR at higher bake temperatures. This leads to the suspicion that the resist glass transition temperature (T_g) may be responsible for the changes in the level of IMR observed with both different PAGs, polymers and bake temperatures. We have measured the T_g in a series of model resists, both exposed and unexposed, and show the effect of changes in resist glass transition as a function of exposure dose and not the level of polymer deprotection. The T_g of the resists does not decrease with exposure or bake as may be expected, but instead is either unchanged or slightly increases. The change in T_g occurs due to exposure only with subsequent bake steps not affecting the resist T_g.

It is generally believed that the chemically amplified reaction between photo-generated acid catalyst and acid labile group of polymer during post-exposure bake (PEB) process plays a critical role for the reduction of line width roughness (LWR) in ArF lithography. In this work, we revealed experimentally how large the chemically amplified reaction affects LWR, and developed a new resist system with low LWR. Aerial image contrast dependence on LWR revealed that the innate LWR in a conventional ArF photoresist, which is independent of the aerial image contrast, was 5 nm. Surface roughness of a non-patterned resist film at half-exposed area, which was well correlated with LWR, was measured by AFM. The surface roughness increased from 1.7 nm to 10.8 nm during PEB process. The half-exposed area was baked and again dissolved into organic solution, and spin-coated on Si wafer, and then developed with 2.38 % TMAH solution. The recoated half-exposed resist film caused a 60 % reduction on the surface roughness. It revealed that uniform distribution of deblocked polymer was important factor for roughness reduction. HPLC analysis indicated that distribution of acidic group formulation in the polymer was gradually extended with increasing exposure dose. A Resist system that suppresses the chemically amplified reaction successfully reduced LWR from 6.5 nm to 4.8 nm.

We describe methods to determine transfer functions for line edge roughness (LER) from the photoresist pattern through the etch process into the underlying substrate. Both image fading techniques and more conventional focus-exposure matrix methods may be employed to determine the dependence of photoresist LER on the image-log-slope (ILS) or resist-edge-log-slope (RELS) of the aerial image. Post-etch LER measurements in polysilicon are similarly correlated to the ILS used to pattern the resist. From these two relationships, a transfer function may be derived to quantify the magnitude of LER that transfers into the polysilicon underlayer from the photoresist.¹ A second transfer function may be derived from power spectral density (PSD) analysis of LER. This approach is desirable based on observations of pronounced etch smoothing of roughness in specific spatial frequency ranges. Smoothing functions and signal averaging of large numbers of line edges are required to partially compensate for large uncertainties in fast-Fourier transform derived PSDs of single line edges. An alternative and promising approach is to derive transfer functions from PSDs estimated using autoregressive algorithms.

Line Width Roughness (LWR) is the random variation of MOS gate length along the gate width. LWR is undesirable because it degrades drive current (Ion), increases off-current (Ioff), and causes a random variation of device parameters across a die. Previously, it was determined that LWR did not impact Intel's 130 nm process devices. As device sizes shrink, the sensitivity to LWR increases, so the amount of LWR that can be tolerated in future generations needs to be re-assessed. In this paper we will present the experimental results of the effects of LWR on Intel's 65 nm process. It was found that both nominal drive current and its variation degrade with increased LWR. Additionally, Ioff increased exponentially with increased LWR. In order to maintain less than 2% degradation in Ion from LWR, the 3-Sigma % LWR should be less than 10% of the nominal final check critical dimension (FCCD). Thus, for future generations, LWR needs to scale as gate lengths decrease or else any potential benefits in increased drive current would be offset by large amounts of leakage.

UV curable resist formulations for nanoimprint must satisfy criteria for cure rate, volatility, viscosity, cohesion of the cured material and release from the template in addition to being successfully imprintable. We describe an investigation of the properties of a series of formulations comprising polyhedral oligomeric silsesquioxane and selected diluents as candidates for imprintable dielectrics. Although all have low viscosity and volatility and are successfully imprinted, significant variations in cure rate, mechanical and adhesion properties with resist composition are found. The trends observed are not all predictable from the literature, indicating that formulation optimization for this application requires a focus on the fundamentals of both materials and processes.

In order to enable design of a negative-tone polyphenol resist using polarity-change reaction, five resist compounds (3M6C-MBSA-BLs) with different number of functional group of &ggr;-hydroxycarboxyl acid were prepared and evaluated by EB lithography. The resist using mono-protected compound (3M6C-MBSA-BL1a) showed 40-nm hp resolution at an improved dose of 52 &mgr;C/cm² probably due to removal of a non-protected polyphenol while the sensitivity of the resist using a compound of protected ratio of 1.1 on average with distribution of different protected ratio was 72 &mgr;C/cm². For evaluation of the di-protected compound based resist, a di-protected polyphenol was synthesized by a newly developed synthetic route of 3-steps reaction, which is well-suited for mass production. The resist using di-protected compound (3M6C-MBSA-BL2b) also showed 40-nm hp resolution at a dose of 40 &mgr;C/cm², which was faster than that of mono-protected resist. Fundamental EUV lithographic evaluation of the resist using 3M6C-MBSA-BL2b by an EUV open frame exposure tool (EUVES-7000) gave its estimated optimum sensitivity of 7 mJ/cm² and a proof of fine development behavior without any swelling.

Substantially improved photoresist material designs, which can provide higher photosensitivity and precise critical dimension and edge roughness control, will be required to enable the application of next generation lithography technology to the production of future sub-65 nm node IC device generations. The development and characterization of novel material platforms that solve the aforementioned basic problems with chemically amplified resists (CARs) is essential and is already one of the major subjects of modern lithography research. In that regard, we have pursued development of a variety of 193 nm and EUV CARs that contain photoacid generator (PAG) units covalently bonded directly to the resin polymer backbone. However, the detailed structure-property relationships that result from this direct attachment of the PAG functional group to the polymer have previously not yet been rigorously characterized. In this work, the lithographic properties of a polymer-bound PAG CAR (GBLMA-co-EAMA-co-F4-MBS.TPS) and its blended-PAG analog resist (GBLMA-co-EAMA blend F4-IBBS.TPS) were studied and compared. The direct incorporation of PAG functionality into the resist polymer, where the resulting photoacid remains bound to the polymer, showed improved photosensitivity, resolution, and lower LER as compared with the analogous blended-PAG resist. The improved resolution and LER were expected due to the restricted photoacid diffusion and uniform PAG distribution provided by direct incorporation of the PAG into the polymer backbone to make a single-component resist material. The ability to load higher levels of PAG into the resist provided by this PAG incorporation into the polymer, as compared to the low PAG concentrations attainable by traditional blending approaches, overcomes the sensitivity loss that should result from reduction in photoacid diffusivity and concomitant smaller acid-catalytic chain lengths. In fact, the polymer-bound PAG resist achieves a faster photospeed than the blended-PAG analog material under DUV radiation in the case of the materials reported here while still providing all of the aforementioned improvements such as the improved line edge roughness.

A new series of anionic photoacid generators (PAGs), and corresponding polymers were prepared. The thermostability of PAG bound polymers was superior to PAG blend polymers. PAG incorporated into the polymer main chain showed improved resolution when compared with the PAG blend polymers. This was demonstrated by Extreme Ultraviolet lithography (EUVL) results: the fluorine PAG bound polymer resist gave 45 nm (1:1), 35 nm (1:2), 30 nm (1:3) and 20 nm (1:4) Line/Space as well as the 50 nm (1:1),30 nm (1:2) elbow patterns.

We report the formation of robust organosilicate line patterns of ~20nm half-pitch on surfaces from the self-assembled lamellar phase of a diblock copolymer of polystyrene and poly(ethylene oxide), PS-b-PEO, and an oligomeric organosilicate precursor mixtures. We could control the orientation and alignment of microdomains of this hybrid to the same degree of the thin films of organic block copolymers. By controlling the surface energy of substrates using dense organosilicate, the perpendicular orientation of lamellae to the surface was achieved. Topographic prepatterns were generated by E-beam lithography and used for alignment of the line patterns from lamellar phase. Upon removing the organic component (i.e. PS-b-PEO) by thermal treatment, the organosilicate microdomains remain as periodic line patterns with global alignment on surfaces. This method gives well-aligned silicon-containing line patterns with sublithographic length scales on surface. The self-assembled organosilicate line patterns were successfully transferred into underlying silicon substrate using anisotropic plasma etching.

Gray scale lithography GSL was implemented using an aqueous developing photosensitive polyimide, HD8820. Silicon carbide high kilo-volt power diodes and transistors fabrication required a tapered barrier for dopant ion implantation and annealing at elevated temperatures for formation of junction edge terminations. The GSL photomask was made by electron beam lithography 32-bit gray scale intensity profile exposure of a silver doped HEBS glass that is sensitive to electron beams but not to ultra-violet UV radiation. When the photosensitive polyimide HD8820 was exposed through the gray scale photomask in a UV lithography system, the desired polyimide tapered profile was obtained. The polyimide patterns were subsequently converted to amorphous graphite at 650°C in a nitrogen furnace. Boron and aluminum high temperature ion implantations at several energies and doses were performed and the resulting dopant profiles, characterized using secondary ion mass spectroscopy SIMS, were found to correlate well with computer simulations. The GSL tapered photosensitive polyimide can therefore serve as an effective ion implantation barrier to obtain precisely graded dopant profiles. Micro-optics components such as gratings and Fresnel micro-lenses were also successfully fabricated using gray scale lithography of the photosensitive polyimide.

A novel method for performing top surface imaging (TSI) on a single polymeric photoresist film using area selective atomic layer deposition (ALD) is presented. In this method, exposure of the polymer thin film creates reactive hydroxyl sites on the film surface in the exposed areas that subsequently act as nucleation and growth sites for deposition of metal oxide features using a chemically selective ALD process. It is demonstrated that titanium isopropoxide and water can be used as ALD precursors in conjunction with a chemically amplified photoresist film, formulated using a protected polymer (poly-tertbutylmethacrylate (PtBMA)) and a photoacid generator (triphenylsulfonium tris(perfluoromethanesulfonyl) methide), to successfully perform such an area selective ALD TSI process. Using this material set and methodology, micron scale photoresist features are defined, metal oxide patterns are produced, and these patterns have been transferred through the polymer film via plasma etching. A comparison is also made between the silylation chemical contrast obtained from earlier TSI methods and the metal oxide deposition contrast obtained using the ALD TSI method. It is shown that the ALD TSI chemical contrast for metal oxide deposition can be much higher than the corresponding silylation-based silicon incorporation contrast This result translates into a unique feature of the ALD TSI process which is it can be achieved without requiring the descum etch that is commonly needed with other TSI methods based on silylation.

The development of spin-on photodefinable dielectrics for use in microelectronic device fabrication and packaging has traditionally focused mainly on the use of soluble polymer precursors because the desired final dielectric polymer structure is generally insoluble in any appropriate casting solvent. One major drawback to this approach however is that high temperature processes (often >300 °C) after imaging and development are generally required to convert the imaged precursor polymer into the desired low dielectric constant material. These high processing temperatures make such materials impractical for a variety of applications. The goal of the work reported in this paper was to overcome this problem by developing new low dielectric constant polymers that can be formulated into photodefinable materials and processed at lower temperatures. In this work the use of a novel hexafluoroisopropanol (HFA)-substituted diamine to synthesize novel polybenzoxazines is reported. Whereas polybenzoxazoles form a five membered ring by situating an alcohol ortho to an amide, the benzoxazine rings in the polymer backbone described in this work are formed by the dehydration of a hexafluoroalcohol (HFA) group situated ortho to the amide to form a six membered ring. Of this general class of new polybenzoxazine amide alcohol precursor polymers, it will be shown that selected polymers exhibit good solubility in developer solutions and can be formulated into photosensitive compositions by addition of a DNQ inhibitor. Polybenzoxazine film properties including dielectric constant, thermal expansion coefficient, glass transition temperature, water absorption and dissolution rate have been measured. In particular, the dielectric constant for polybenzoxazine is reported to be as low as 2.2. In contrast to polybenzoxazoles which are known to cyclize at temperatures well above 280 °C, the new polymers reported here can be cyclized at temperatures as low as 210 °C. This significantly lower thermal cyclization temperature greatly alleviates many problems with other photodefinable dielectric polymers such as thermal stress build-up during curing and also allows these materials to be integrated with a wider variety of materials that would not survive the significantly higher processing temperature required with most current polyimides and polybenzoxazoles.