The photoacid diffusion length is a critical issue for extreme ultraviolet (EUV) lithography because it governs the critical
dimension (CD), line-edge-roughness (LER), and line-width-roughness (LWR) of photoresist materials. Laboratorybased
experimental methods that complement full lithographic testing would enable a rapid screening of materials and
process conditions. This paper provides an approach to characterize the photoacid diffusion length by applying a bilayer
stack technique. The method involves quantitative measurements of the deprotection kinetics as well as film thickness at
each process step: radiation exposure, post-exposure bake, and development. Analogous to a contrast curve, by
comparing the film thickness of the bilayer before and after development, the photoacid diffusion length was deduced in
a commercial EUV photoresist and compared to EUV lithography. Further, by combining the experiments with kinetics
modeling, the measured photoacid diffusion length was predicted. Lastly, based upon the measured kinetics parameters,
a criterion was developed that next-generation resists must meet to achieve a 16 nm photoacid diffusion length. These
guidelines are discussed in terms of correlations and contributions from the photoacid and resist properties. In particular,
the trapping kinetics of the photoacid provides a route to reduce LER and the CD at low dose.
The polydispersity in the degree of functionalization for two calix[4]resorcinarenes was determined by
measuring, quantitatively, their molecular mass distribution with matrix-assisted laser desorption/ionization time of
flight mass spectrometry. A mathematical method for polydisperse materials is described that creates a calibration curve
to correct the ion intensities in the mass spectrum to give a more reliable molecular mass distribution. One
calix[4]resorcinarene was found to give accurate molecular mass distributions with little correction, while another
having a very similar molecular structure was found to exhibit strong over counting of the oligomers having a high degree of functionalization.
It has been recently postulated that sub-22 nm photolithography with polymeric photoresists has reached a materials design barrier due to its large molecular mass and distribution. In this argument, the "pixel" size, which is related to the smallest molecular unit, determines the feature fidelity and resolution of the lithographic process. This hypothesis remains unproven, but molecular glass photoresists can provide a test because they can share similar chemical functionality to polymer resists, but with low molecular mass and a monodisperse molecular mass distribution. The low molecular mass leads to the smaller pixel size compared to the radius of gyration of the polymer photoresist. In this work, we compare the deprotection reaction-diffusion kinetics of a common photoacid generator in a polymer and molecular glass resist with similar resist chemistry to elucidate effects of molecular architecture on photoresist performance. We determine the mechanism of reaction, photoacid trapping behavior, and diffusivity by measuring and comparing the reaction kinetics parameters as a function of temperature and exposure dose. These results permit an analysis of the latent image formation which is a crucial factor in resolution and line-edge roughness. Further, knowledge of the reaction-diffusion parameters of each type of resist provides a quantitative approach to predict line-space features, crucial for design for resolution-enhancement features.
We have examined four molecular glasses (MGs) which are candidates for EUV photoresist formulations. These
derivatized glasses, and their unprotected precursors, were investigated by both proton and 13C solid state NMR
techniques in the bulk state as pure materials and as mixtures with 5 or 10 % by mass of the photoacid generator (PAG),
triphenyl sulfonium perfluorobutanesulfonate. The 13C techniques gave information about crystallinity, purity, and the
presence of the PAG. This paper characterizes the intimacy of mixing of the PAG and the MGs using proton spin
diffusion methods. Phase separation of the PAG into PAG-rich larger domains was never seen; the PAG was always
finely distributed. A maximum diameter for any PAG clustered into spherical domains was estimated to be 3.8 nm,
which is too small to reflect thermodynamic incompatibility as the driving force during relatively slow removal of
solvent. Hence, PAG blended samples are deduced to be thermodynamically compatible, with differential solubility in
the preparation solvent the most likely candidate for producing the significant inhomogeneities in PAG concentration
observed in a few samples. For one of the unprotected crystalline calix[4]resorcinarenes precursor materials, the
solvent, N-methyl 2-pyrrolidinone (NMP) was used. The resulting solid was crystalline with a segregation of isomers,
one of which formed a solid adduct with a 1:1 molecular ratio with NMP. Qualitatively, the strong NMP affinity for the
calix[4]resorcinarenes is also evident in a) the immobility of the NMP, b) the fact that the 14N quadrupolar interaction
changes when NMP goes from the crystalline, unprotected host to a glassy, protected host, and c) that NMP tends to
remain as a significant residue. Only the underivatized materials display crystallinity implying that the mixing of the
PAG with any derivatized MG is not restricted by crystallization, at least not before the post-exposure bake step. As a
final note, very strong hydrogen bonds exist in three underivatized materials which is reduced or eliminated with partial
protection with t-BOC.
Current extreme ultraviolet (EUV) photoresist materials do not yet meet performance requirements on exposure-dose sensitivity, line-width roughness, and resolution. In order to quantify how these trade-offs are related to the materials properties of the resist and processing conditions, advanced measurements and fundamental studies are required that consider EUV-resist specific problems. In this paper, we focus on the correlations between the latent image and developed image in EUV exposed line/space features. The latent images of isolated lines produced by EUV lithography are characterized by atomic force microscopy through the change in topology caused by change in film thickness that occurs upon deprotection. The resulting latent-image deprotection gradient (DGL), based on line cross-sections, and latent-image line-width roughness (LWRL) provide metrics and insight into ways to optimize the lithographic process. The results from a model poly(hydroxystyrene-co-tert-butylacrylate) resist and a model calix[4]resorcinarene molecular glass type resist show the general applicability of the metric before development.
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.
Current extreme ultraviolet (EUV) photoresist materials do not yet meet requirements on exposure-dose sensitivity,
line-width roughness (LWR), and resolution. Fundamental studies are required to quantify the trade-offs in materials
properties and processing steps for EUV photoresist specific problems such as high photoacid generator (PAG) loadings
and the use of very thin films. Furthermore, new processing strategies such as changes in the developer strength and
composition may enable increased resolution. In this work, model photoresists are used to investigate the influence of
photoacid generator loading and developer strength on EUV lithographically printed images. Measurements of line
width roughness and developed line-space patterns were performed and highlight a combined PAG loading and
developer strength dependence that reduce LWR in a non-optimized photoresist.
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.
The controlling factors in the formation of the compositional heterogeneity at the deprotection front were
investigated using 3D computer simulation. The results illustrate that the chemical composition fluctuation (CCF)
formed by the photoresist deprotection reaction is an important factor contributing to the line-edge-roughness (LER) in
addition to the deprotection gradient (DG) of the reaction front. The magnitude of the chemical composition fluctuation
and the deprotection gradient are found to depend on the ratio of the deprotection reaction rate constant to diffusion
coefficient (kP/D) and the number of hoping step (n) With this new finding, the influence on LER from various
process/material parameters such as dose/contrast, diffusivity, and reactivity can all be understood through their effects
on kP/D and n.
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.
More demanding requirements are being made of photoresist materials for fabrication of nanostructures as the feature critical dimensions (CD) decrease. For extreme ultraviolet (EUV) resists, control of line width roughness (LWR) and high resist sensitivity are key requirements for their success. The observed LWR and CD values result from many factors in interdependent processing steps. One of these factors is the deprotection interface formed during the post-exposure bake (PEB) step. We use model EUV photoresist polymers to systematically address the influence of exposure-dose on the spatial evolution of the deprotection reaction at a model line edge for fixed PEB time using neutron reflectivity. The bilayer consists of an acid feeder layer containing photoacid generator (PAG) and a model photoresist polymer, poly(hydroxystyrene-co-tert-butylacrylate) with perdeuterated t-butyl protecting group. The deuterium labeling allows the protection profile to be measured with nanometer resolution. The evolution of two length scales that contribute to the compositional profile is discussed.
An understanding of acid diffusion-reaction in chemically amplified photoresists during the post-exposure bake (PEB) is critical for both critical dimension (CD) and line edge roughness (LER) control. Despite its importance, there remains insufficient understanding of the diffusion-reaction process. This is due in part to the complex interplay between diffusion and reaction where the deprotection of the resin modifies the local acid diffusivity which in turn changes the rate of deprotection. Here, we report the direct measurement of the reaction diffusion front at a model line edge from neutron reflectivity and Fourier transform infrared spectroscopy measurements. The photoacid generator size influences the reaction extent and breath of the deprotection profile. A larger photoacid results in a sharper deprotection profile and a shorter reaction length. Under the same post-exposure bake time and temperature, the smaller photoacid leads to a much broader deprotection profile. These measurements illustrate the complexity of the reaction-diffusion process.
A correlation between polymer molecular structure and acid catalyzed reaction kinetics is demonstrated by a photoresist copolymer with an acid-labile and a non-reactive monomer. The acid catalyzed deprotection kinetics depend significantly on the composition of the non-reactive comonomer in the polymer chain. The apparent reaction rate constant decreases monotonically with increasing non-reactive comonomer composition. The phenomena are interpreted as the reduction of diffusivity of photoacid in the polymer matrix from a hydrogen-bonding interaction with the polar group in the inert comonomer. In addition, hydrogen-bonding interactions between the photoacid and the reaction product, primarily methacrylic acid, can account for the acid loss or trapping effect observed by various researchers.
The dissolution of partially deprotected chemically amplified photoresists is the final step in printing lithographic features. Since this process step can be tuned independently from the design of the photoresist chemistry, measurements of the dissolution behavior may provide needed insights towards improving line-edge roughness. We have studied the dissolution behavior of a model 193-nm photoresist, poly (methyladamantyl methacrylate), as a function of deprotection extent and developer strength. The kinetics of the dissolution process is followed using the quartz crystal microbalance technique. These photoresist films exhibit strong swelling without dissolution over a significant range of deprotection levels. At larger extents of deprotection, we observe a combination of swelling with dissolution. Additionally, we find that the degree of film swelling decreases with tetramethylammonium hydroxide developer concentration. These studies provide the insight needed to better understand the fundamentals of the dissolution of the photoresist - a key step in lithographic process.
A depth profile of the base developer counterion concentration within thin photoresist films was measured in-situ using contrast variant specular neutron reflectivity to characterize the initial swelling stage of the film dissolution. We find a substantial counterion depletion near the substrate and an enrichment near the periphery of the film extending into the solution. These observations challenge our understanding of the charge distribution in photoresist and polyelectrolyte films and are important for understanding film dissolution in medical and technological applications.
The emergence of immersion lithography for the extension of current lithography tools requires a fundamental understanding of the interactions between the photoresist and the immersion liquid such as water. Neutron reflectometry was used to measure the water concentration depth profile within immersed photoresist films. The bulk of the films swelled to the equilibrium water concentration. However a gradient in water concentration was observed near the polymer/substrate interface. Dependent on the relative hydrophilicity of the polymer and the substrate, either a depletion or excess of water was observed at the interface. Using HMDS treated silicon wafers as the substrate results in approximately 17% water by volume at the interface. The interfacial concentration decreases (or increases) to the bulk water solubility limit approximately 40 Å from the substrate. As the total film thickness approaches this length scale, the substrate induced concentration gradients lead to a film thickness dependent swelling; enhanced or suppressed swelling is witnessed for the excess or depleted interfacial concentrations, respectively. Variation of the substrate surface energy allows for tuning of the interfacial water concentration, ranging from 30% to less than 1% water by volume.
The emergence of immersion lithography as a potential alternative for the extension of current lithography tools requires a fundamental understanding of the interactions between the photoresist and an immersion liquid such as water. The water concentration depth profile within the immersed photoresist films is measured with neutron reflectometry. The polymer/substrate interface affects both the water concentration near the interface and the surface morphology of the film. Immersed films are not stable (adhesive failure) over the course of hours when supported on a silicon wafer with a native oxide surface, but are stable when the substrate is first treated with hexamethyldisilazane (HMDS). The bulk of the polymer films swells to the equilibrium water concentration, however, a gradient in water concentration is observed near the polymer/HMDS substrate interface with a concentration of approximately 17% by volume fraction and extending up to 50 Å into the film. Thus, polymers that absorb more than this amount exhibit depletion near the interface, whereas polymers that absorb less exhibit a water excess layer. These concentration gradients extend approximately 50 Å away from the interface into the film. As the total film thickness approaches this length scale, the substrate-induced concentration gradients lead to a film-thickness-dependent swelling; enhanced or suppressed swelling is witnessed for the excess or depleted interfacial concentrations, respectively. The substrate also influences the surface morphology of immersed thin films. The film surface is smooth for the HMDS-treated substrate, but pin-hole defects with an average radius of 19±9 nm are formed in the films supported on the native oxide substrates.
A variety of experimental evidence suggests that positive-tone chemically amplified photoresists have an intrinsic bias that might limit resolution during high-volume lithographic processing. If this is true, the implications for the semiconductor industry require careful consideration. The design concept of chemical amplification is based on generation of a chemically stable catalytic species in exposed regions of the photoresist film. The catalytic action of the photoproducts on the photoresist polymer causes a change in the dissolution rate in the irradiated regions of the film. Formation of a stable catalyst species is required for chemical amplification, but it has long been recognized that catalyst migration can produce a difference between the initial distribution of exposure energy and the final distribution of photoproducts. This difference, known as diffusion bias, depends on the photoresist chemistry and processing conditions. Diffusion bias is insensitive to exposure conditions, but it is possible to reduce catalyst migration through changes to resist formulation such as increasing the size of the catalyst molecule or processing conditions such as reducing the post exposure bake temperature. Another common approach to limiting diffusion bias is to incorporate base additives into the photoresist formulation to scavenge diffusing acid catalyst. All of these approaches to reducing catalyst migration generally reduce the catalytic efficiency of each photoproduct and therefore increase the total exposure dose required to pattern the film. Increases in required exposure dosage reduce the throughput of the exposure tools and can reduce the profitability of the manufacturing process. In this paper we present experimental results that are suggestive of an intrinsic photoresist bias. This diffusion bias sets a minimum resolution limit for chemically amplified resist systems that can be improved at the cost of reduced throughput and productivity.
Organic polar solvent (1-butanol) versus aqueous base (tetramethylammonium hydroxide, (TMAH)) development quality are distinguished by neutral versus charged polymer (polyelectrolyte) dissolution behavior of photoresist bilayers on silicon substrates comprising poly(4-hydroxystyrene) and poly(4-tert-butoxycarbonyloxystyrene), PHOSt and PBOCSt, respectively. This model line-edge was broadened by photoacid catalyzed deprotection to a final interfacial width of 35.7 Å and subjected to different developers. 1-butanol develops with an increased penetration depth than aqueous base development consistent with an increased solubility of the protected containing component in the organic solvent. This enhanced dissolution with the polar solvent results in an increased surface roughness of 73 Å, whereas the development with TMAH at concentrations between (0.1 to 1.1) M1 leads to surface roughness between (4.5 to 14.4) Å, as measured by atomic force microscopy. These measurements suggest that the elimination of resist swelling, in the presence of a protection gradient, is a viable strategy to reduce roughness and control critical dimensions. The influence of added salt to developer solutions was also examined by developing the model bilayer. A decrease in surface roughness from (10 to 6.5) Å was observed between (0 to 0.70) M KCl in 0.26 M TMAH.
The performance of chemically amplified photoresists, including next generation thin film 157 nm fluorinated copolymers and blends, is affected by such phenomena as polymer/substrate and polymer/air interfacial (surface energy) effects, blend miscibility, small molecule diffusion in thin films, permeability of airborne contaminants, and interactions with products from the deprotection reaction. Using near edge x-ray absorption fine structure (NEXAFS) spectroscopy, it is possible to simultaneously probe the surface and bulk chemistry of chemically amplified photoresists to determine possible causes of pattern degradation, including post exposure delay induced material failure, blend component and small molecule diffusion/segregation to the photoresist surface, and interactions between components of the photoresist formulation and developer. The surface and bulk chemistry of model photoresists were analyzed in the NEXAFS vacuum chamber, equipped with in situ processing capabilities for exposure, controlled dosing of a model contaminant gas (NMP or water vapor), and heating, to quantify component segregation and identify surface phenomena that may be responsible for pattern degradation. For model 157 nm blend films, it was found that there is segregation of one component to the surface of the photoresist film, in excess of the composition of that component in the blend. For polymer blends the more hydrophobic or lower surface tension species will typically wet the film surface when heated in air. Segregation of photo-acid generator has also been demonstrated and the effect of reducing film thickness investigated. As photoresist film thickness continually decreases and the photoresists become increasingly sensitive to environmental contaminants, the interfacial and surface regions dominate the behavior of the material and it is crucial to understand both their physical and chemical nature.
Neutron and x-ray reflectivity measurements quantify the non-uniform distribution of water within poly(4-tert-butoxycarbonyloxystyrene) (PBOCSt) and poly(4-hydroxystyrene) (PHOSt) thin films on silicon wafer substrates. Two contrasting surface treatments were examined, silicon oxide, representing a hydrophilic interface and a trimethylsilane primed surface, representing a hydrophobic interface. The distribution of water in the films was sensitive to the surface preparation and photoresist relative hydrophilicity. Depending upon the water contact angle of the substrate in comparison to the polymer film, an excess of water near the interface occurs when the substrate is more hydrophilic than the photoresist. Likewise, interfacial depletion results when the photoresist is more hydrophilic than the substrate. These non-uniform water distributions occurs within (50 ± 10)Å of the photoresist/substrate interface. The water concentration in this interfacial region appears to be independent of the photoresist properties, but is strongly dependent upon the substrate surface energy.
We demonstrate that poly(4-hydroxystyrene) and (5, 15, and 20) % tert-butoxycarboxy protected copolymers are polyelectrolytes when dissolved in aqueous base solutions. The polyelectrolyte effect is quantified through the observation of a correlation peak, measured with small-angle neutron scattering. Polyelectrolyte effects are weakened with added salts and excess base. These studies emphasize that salt additives screen the electrostatic interactions, while pH leads to the ionization of the chain. Solvent quality is quantified and the chain configurations are measured in the limit of high ionic strength. It is speculated that the developer-resist interactions will play an important role in development-induced roughness, hence these equilibrium solution measurements can serve a predictive function for future photoresists dissolution models incorporating solvent quality as a parameter.
A series of experiments are presented to demonstrate thin film confinement effects on the diffusive properties in poly(tert-butoxycarboxystyrene) (PBOCSt). Bilayer diffusion couple measurements reveal that as the thickness of a PBOCSt film is decreased, the kinetics of the deprotection reaction-front propagation (a process involving both the diffusion and reaction of photochemically activated acidic protons) are dramatically hindered. Incoherent neutron scattering measurements suggest that this retardation can be traced to a suppression of local fast relaxations (200 MHz or faster) native to the PBOCSt polymer. The reduced mobility in the thin PBOCSt films is further confirmed with moisture vapor uptake studies performed on a quartz crystal microbalance (QCM). As the film thickness drops below 500 Angstrom there is a strong reduction in the diffusivity of water in the film. In total, these are the first evidences suggesting that the deviations in lithographic performance with decreasing film thickness observed with the bilayer experiments can are due to changes in mobility, not reactivity, within a chemically amplified resist.
The form and magnitude of line edge roughness (LER) is increasingly important in semiconductor processing due to continued reductions in feature sizes. While a large body of work connects processing factors to LER magnitude, the spatial dependence of LER is needed to provide a more complete description. The distribution of deprotection within the resist is represented as a collection deprotection paths created by individual photoacid generators (PAGs). In the limit of dilute PAG concentration, the form and size of the average deprotection path is measured using Small Angle Neutron Scattering (SANS) for a model photoresist polymer and PAG mixture. The heterogeneity of the deprotection volume produces “fuzzy blobs”. The shape of these blobs is compared to the form of LER at a idealized sidewall. The sidewall morphology is consistent with models of spatially random etching up to a cutoff length scale. The cutoff length scale is ≈ 5 times the size of a single deprotection volume, suggesting that collective phenomena are responsible for observed LER.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.